AXISYMMETRIC DROP SHAPE ANALYSIS (ADSA) by · PDF file2.3 Double Injection Capability ......
Transcript of AXISYMMETRIC DROP SHAPE ANALYSIS (ADSA) by · PDF file2.3 Double Injection Capability ......
AXISYMMETRIC DROP SHAPE ANALYSIS (ADSA)AND LUNG SURFACTANT
by
Sameh Mossaad Iskander Saad
A thesis submitted in conformity with the requirementsfor the degree of Doctor of Philosophy
Graduate Department of Mechanical and Industrial EngineeringUniversity of Toronto
Copyright c© 2011 by Sameh Mossaad Iskander Saad
AXISYMMETRIC DROP SHAPE ANALYSIS (ADSA)
AND LUNG SURFACTANT
Sameh Mossaad Iskander Saad
Doctor of Philosophy
Graduate Department of Mechanical and Industrial Engineering
University of Toronto
2011
Abstract
The objective of this thesis was to further develop a methodology for surface tension
measurement called Axisymmetric Drop Shape Analysis (ADSA) and to adapt it to
studies of lung surfactants, i.e. the material that coats and facilitates the functioning
of the lungs of all mammals. The key property of a functioning lung surfactant is its
surface tension, which can reach extremely low values below 1 mJ/m2. Such values are
difficult to measure; but a certain configuration of ADSA, using a constrained sessile
drop (ADSA–CSD), is capable of performing such measurements.
Clinically, lung surfactant films can be altered from both sides, i.e. from the airspace
as well as from the bulk liquid phase that carries the film. Therefore, being able to
access the interface from both sides is important. Here, ADSA–CSD was redesigned to
be used as a micro film balance allowing access to the interface from both gas- and liquid-
side. This allows deposition from the gas side as well as complete exchange of the bulk
liquid phase. The new design was used to study lung surfactant inhibition and inhibition
reversal.
A dynamic compression-relaxation model (CRM) was developed to describe the me-
chanical properties of lung surfactant films by investigating the response of surface tension
to changes in surface area. The model evaluates the quality of lung surfactant prepara-
tions – beyond the minimum surface tension value – and calculates the film properties,
i.e. elasticity, adsorption and relaxation, independent of the compression protocol.
ii
The accuracy of the surface tension measurement can depend on drop size. A detailed
analysis of drop shapes and accuracy of measured surface tension values was performed
using a shape parameter concept. Based on this analysis, the design of ADSA–CSD was
optimized to facilitate more accurate measurements. The validity analysis was further
extended to the more conventional pendant drop setup (ADSA–PD) in which accuracy
near ±0.01 mJ/m2 is now possible.
An overall upgrade of both hardware and software of ADSA–CSD, together with
extensive numerical work, is described and applied to facilitate a more efficient operation.
Finally, it is noted that the ADSA–CSD setup developed here can be used for a wide
range of colloid and surface chemical applications.
iii
†
“I can do all things through Christ who strengthens me.”
(Philippians 4:13)
To my Lord, God and Savior JESUS CHRIST.
To Marie, my parents and my siblings.
To my beloved homeland, Egypt.
iv
Acknowledgements
“Trust in the LORD with all your heart, and lean not on your own understanding;
in all your ways acknowledge Him, and He shall direct your paths.” (Proverbs 3:5-6).
I would like to express my sincere gratitude to my supervisor Professor A. Wilhelm
Neumann to whom I owe a great debt of gratitude for his patience, inspiration and
support. I would also like to thank my co-supervisor Professor Edgar J. Acosta for his
advice and support. I wish to thank my supervisory committee, Professor Michael L.
Hair and Professor Markus Bussmann, for their priceless suggestions, continuous support
and encouragement.
The experimental assistance of Ms. Zdenka Policova was invaluable and much appre-
ciated. Additional thanks to Dr. Robert David for his technical and fruitful discussions
and to my colleague Ali Kalantarian for his friendship and help. I wish to thank the
summer and co-op students: Wojciech Gryc, Regina Park, Simon Lee, Andrew Dang,
Maria Quadir and Carl Cheng, for their effort. I wish also to thank everyone in the
Laboratory of Applied Surface Thermodynamics.
I am grateful to the Provincial Government of Ontario for its financial support through
the Ontario Graduate Scholarship (OGS). I am also grateful to the University of Toronto
for awarding me an Open Fellowship and to the Department of Mechanical and Industrial
Engineering for several Teaching Assistant positions, which provided not only financial
support but also an important learning experience. I would like to thank BLES Biochem-
icals Inc. for the generous donation of the BLES samples. Part of this work is supported
by a grant from the Natural Sciences and Engineering Research Council (NSERC) of
Canada (Grant No. 8278) and part is supported by a grant from the Canadian Institutes
of Health Research (CIHR) (MOP-11 38037).
Finally, I would also like to extend a sincere thanks to all my friends, whose names
will not be written here for risk of leaving someone out, for making my sojourn in Canada
a profitable and enjoyable one. Most importantly I would like to thank my parents for
always offering support and love.
v
Contents
1 Introduction 11.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.1 Evaluation of Surface Tension . . . . . . . . . . . . . . . . . . . . 41.2.2 Pulmonary Surfactants . . . . . . . . . . . . . . . . . . . . . . . . 15
1.3 Scope and Purpose of Thesis . . . . . . . . . . . . . . . . . . . . . . . . . 16
2 ADSA-CSD as a Micro Film Balance 192.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192.2 Deposition Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.2.1 Experimental Details . . . . . . . . . . . . . . . . . . . . . . . . . 232.2.2 Results for Pure Monolayers . . . . . . . . . . . . . . . . . . . . . 302.2.3 Discussion for Pure Monolayers . . . . . . . . . . . . . . . . . . . 372.2.4 Results for Mixed Monolayers . . . . . . . . . . . . . . . . . . . . 432.2.5 Discussion for Mixed Monolayers . . . . . . . . . . . . . . . . . . 48
2.3 Double Injection Capability . . . . . . . . . . . . . . . . . . . . . . . . . 522.3.1 Experimental Details . . . . . . . . . . . . . . . . . . . . . . . . . 532.3.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 592.3.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
3 Dynamic Surface Tension Model 763.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 763.2 Previous Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
3.2.1 Diffusion Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 803.2.2 Kinetic Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 813.2.3 Adsorption-Limited Model . . . . . . . . . . . . . . . . . . . . . . 83
3.3 Development of Compression–Relaxation Model . . . . . . . . . . . . . . 853.3.1 Adsorption Process (Spreading) . . . . . . . . . . . . . . . . . . . 853.3.2 Desorption Process (Relaxation) . . . . . . . . . . . . . . . . . . . 863.3.3 Elasticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 873.3.4 Collapse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 873.3.5 The Integrated Model . . . . . . . . . . . . . . . . . . . . . . . . 88
3.4 CRM Applied to Sample Experiments . . . . . . . . . . . . . . . . . . . . 893.4.1 Experimental Details . . . . . . . . . . . . . . . . . . . . . . . . . 893.4.2 Computational Details . . . . . . . . . . . . . . . . . . . . . . . . 91
vi
3.4.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 933.4.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
3.5 Effect of Concentration, Compression Ratio and Compression Rate . . . 1043.5.1 Experimental Details . . . . . . . . . . . . . . . . . . . . . . . . . 1053.5.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1063.5.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1213.5.4 Range of Insensitivity . . . . . . . . . . . . . . . . . . . . . . . . . 126
4 Accuracy and Shape Parameter 1304.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1304.2 Drop Shapes and Shape Parameter . . . . . . . . . . . . . . . . . . . . . 137
4.2.1 Dimensional Analysis of Axisymmetric Drop Shapes . . . . . . . . 1374.2.2 Shape Parameter . . . . . . . . . . . . . . . . . . . . . . . . . . . 1404.2.3 Sample Constrained Sessile Drops . . . . . . . . . . . . . . . . . . 1414.2.4 Experimental Details . . . . . . . . . . . . . . . . . . . . . . . . . 1484.2.5 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . 151
4.3 Pendant Drop Shapes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1584.3.1 Experimental Details . . . . . . . . . . . . . . . . . . . . . . . . . 1704.3.2 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . 171
4.4 Universality of Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
5 Summary of Progress and Outlook 185
Bibliography 191
Appendix A Commercial Lung Surfactants 219A.1 Natural Surfactant Extracts . . . . . . . . . . . . . . . . . . . . . . . . . 219
A.1.1 Alveofact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221A.1.2 BLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224A.1.3 Infasurf . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224A.1.4 Curosurf . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225A.1.5 Surfacten . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225A.1.6 Survanta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
A.2 Synthetic Surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226A.2.1 ALEC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228A.2.2 Exosurf . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228A.2.3 Surfaxin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228A.2.4 Venticute . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
vii
List of Tables
2.1 Experimental parameters used for different monolayer types . . . . . . . 262.2 Collapse pressure (mJ/m2) from different runs of different monolayer types
and the average value along with the standard deviation . . . . . . . . . 352.3 Ultimate collapse pressure (mJ/m2) and film elasticity (mJ/m2) at room
temperature of 24oC from different runs of different DPPC:DPPG mixtureratios and the average value along with the standard deviation. . . . . . 45
3.1 Summary of the dynamic parameters (mean and standard deviation) cal-culated from the compression-relaxation model for different preparationsat 20% compression and 3 seconds per cycle. . . . . . . . . . . . . . . . 98
3.2 Summary of the dynamic parameters for BLES at 37C and 100% R.H. atdifferent conditions. Asterisks show range of insensitivity (explained in fig-ure 3.17): * range of insensitivity bigger than 0.2, ** range of insensitivitybigger than 0.5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
3.3 Summary of the range of insensitivity of CRM to each of the dynamicparameter (explained in figure 3.17) for BLES at 37C and 100% R.H. atdifferent conditions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
4.1 Summary of results of constrained sessile drop experiments using differentliquids and different holder sizes. . . . . . . . . . . . . . . . . . . . . . . 154
4.2 Summary of results of pendant drop experiments using different liquidsand different holder sizes. . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
viii
List of Figures
1.1 Schematic diagram of a Langmuir-Wilhelmy Balance (LWB). . . . . . . 51.2 Schematic diagram of a Pulsating Bubble Surfactometer (PBS). . . . . . 81.3 Schematic diagram of a Captive Bubble Surfactometer (CBS). . . . . . . 91.4 Schematic diagram of different constellations of Axisymmetric Drop Shape
Analysis (ADSA). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.1 Schematic diagram of ADSA–CSD experimental setup. . . . . . . . . . . 242.2 The change of surface tension with time after the deposition of octadecanol
in heptane/ethanol solution. . . . . . . . . . . . . . . . . . . . . . . . . 252.3 The process of solution deposition on a constrained sessile drop: (a) the
drop at the full size (∼80µl) before deposition; (b) the drop at half size(∼40µl) before deposition; (c) solution deposition (5µl) where more thanone solution drop falls from the needle over the apex; (d) part of the solu-tion is still attached to the needle; (e) the needle is moved closer withoutpenetrating the interface; (f) the sessile drop after complete deposition; (g)expansion of the sessile drop back to the full size right after deposition;(h) the sessile drop after complete solvent evaporation (∼2min). . . . . 27
2.4 The change of surface tension, drop surface area, drop volume and radiusof curvature at the apex with time of a constrained sessile drop of anoctadecanol monolayer. . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.5 The change of surface pressure with area per molecule of a constrainedsessile drop of an octadecanol monolayer: (a) the full isotherm for differentruns; (b) the isotherm of five different runs in the vicinity of the collapsepressure (only two points per second shown). . . . . . . . . . . . . . . . 31
2.6 The change of surface pressure with area per molecule of a constrainedsessile drop of a DPPC monolayer: (a) the full isotherm; (b) the isothermin the vicinity of the collapse pressure (only two points per second shown). 32
2.7 The change of surface tension and drop surface area with time of a con-strained sessile drop of a DPPC monolayer: (a) in the vicinity of initialepisodes of collapse; (b) in the vicinity of the ultimate collapse pressure(raw data acquired at the rate of ten images per second shown). . . . . 34
2.8 The change of surface pressure with area per molecule of a constrainedsessile drop of a DPPG monolayer: (a) the full isotherm; (b) the isothermin the vicinity of the collapse pressure (only two points per second shown). 36
2.9 The change of surface pressure with area per molecule of a constrainedsessile drop of a DPPC monolayer for different runs. . . . . . . . . . . . 43
ix
2.10 The change of surface pressure with area per molecule of a constrained ses-sile drop of mixed DPPC/DPPG monolayers for different mixture ratios:(a) 100:0; (b) 90:10; (c) 80:20; (d) 70:30; (e) 60:40; (f) 50:50; (g) 0:100. . 44
2.11 The change of surface pressure with area per molecule on successive com-pressions and expansions of a constrained sessile drop of mixed DPPC/DPPGmonolayers for different mixture ratios: (a) 100:0; (b) 90:10; (c) 80:20. . 47
2.12 Schematic diagram of double injection ADSA–CSD experimental setupand an illustration of the exchange process. . . . . . . . . . . . . . . . . 55
2.13 Detailed engineering drawing of double injection ADSA–CSD. . . . . . . 56
2.14 Change in surface tension with time and with relative surface area duringdynamic cycling: (a) 2.0 mg/ml BLES and 0.2 mg/ml protasan (basicpreparation); (b) After replacing the bulk phase with salt solution. Allcompressions/expansions are performed at 3 s/cycle periodicity in humidair (100% R.H.). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
2.15 Change in surface tension with time and with relative surface area duringdynamic cycling: (a) 2.0 mg/ml BLES and 0.2 mg/ml protasan (basicpreparation); (b) After injecting 0.15 ml/ml serum into the bulk phase;(c) 2.0 mg/ml BLES and 0.2 mg/ml protasan and 0.15 ml/ml serum mixedin a separate single injection experiment. All compressions/expansions areperformed at 3 s/cycle periodicity in humid air (100% R.H.). . . . . . . . 61
2.16 Minimum surface tension (end of compression) as a function of protasanconcentration for different batches of BLES (2.0 mg/ml). All compressions(20%) are performed at 3 s/cycle periodicity in humid air (100% R.H.). 62
2.17 Change in surface tension with time and with relative surface area duringdynamic cycling: (a) After injecting 0.30 ml/ml serum into the bulk phaseof 2.0 mg/ml BLES and 0.2 mg/ml protasan; (b) 2.0 mg/ml BLES and0.2 mg/ml protasan and 0.30 ml/ml serum mixed in a separate singleinjection experiment. All compressions/expansions are performed at 3s/cycle periodicity in humid air (100% R.H.). . . . . . . . . . . . . . . . 63
2.18 Change in surface tension with time and with relative surface area duringdynamic cycling: (a) After injecting 40 mg/ml albumin into the bulk phaseof 2.0 mg/ml BLES and 0.2 mg/ml protasan; (b) 2.0 mg/ml BLES and0.2 mg/ml protasan and 40 mg/ml albumin mixed in a separate singleinjection experiment. All compressions/expansions are performed at 3s/cycle periodicity in humid air (100% R.H.). . . . . . . . . . . . . . . . 64
2.19 Change in surface tension with time and with relative surface area duringdynamic cycling: (a) After injecting 3.0 mg/ml fibrinogen into the bulkphase of 2.0 mg/ml BLES and 0.2 mg/ml protasan; (b) 2.0 mg/ml BLESand 0.2 mg/ml protasan and 3.0 mg/ml fibrinogen mixed in a separatesingle injection experiment. All compressions/expansions are performedat 3 s/cycle periodicity in humid air (100% R.H.). . . . . . . . . . . . . 65
x
2.20 Change in surface tension with time and with relative surface area duringdynamic cycling: (a) After injecting 0.284 mg/ml cholesterol into the bulkphase of 2.0 mg/ml BLES and 0.2 mg/ml protasan; (b) 2.0 mg/ml BLESand 0.2 mg/ml protasan and 0.284 mg/ml cholesterol mixed in a separatesingle injection experiment. All compressions/expansions are performedat 3 s/cycle periodicity in humid air (100% R.H.). . . . . . . . . . . . . 66
2.21 Comparison of minimum surface tension for different inhibitors injectedor mixed into a basic preparation of 2.0 mg/ml BLES and 0.2 mg/mlprotasan. The dashed line shows the minimum surface tension of thebasic preparation. All compressions (20%) are performed at 3 s/cycleperiodicity in humid air (100% R.H.). . . . . . . . . . . . . . . . . . . . 67
2.22 Comparison of elasticity of compression for different inhibitors injectedor mixed into a basic preparation of 2.0 mg/ml BLES and 0.2 mg/mlprotasan. The dashed line shows the elasticity of compression of the basicpreparation. All compressions (20%) are performed at 3 s/cycle periodicityin humid air (100% R.H.). . . . . . . . . . . . . . . . . . . . . . . . . . . 68
2.23 Change in surface tension with time and with relative surface area duringdynamic cycling: (a) 0.01 ml/ml serum; (b) After injecting 2.0 mg/mlBLES and 0.2 mg/ml protasan into the bulk phase. All compressions/ex-pansions are performed at 3 s/cycle periodicity in humid air (100% R.H.).. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.1 The change of surface tension and surface area with time of one cycle for2.0 mg/ml BLES at wet conditions, 37C, and 20% compression ratio witha periodicity of 3 seconds per cycle. . . . . . . . . . . . . . . . . . . . . 78
3.2 The response of surface tension to a trapezoidal change in surface areausing the new compression-relaxation model for a good (solid line) anda bad (dashed line) system: (a) the trapezoidal change of surface areawith time; (b) the response of surface tension with time for both systems;(c) the change of surface tension with the relative surface area for bothsystems. The good system (solid line) is generated using parameters: εc= 150 mJ/m2, εe = 150 mJ/m2, kr = 0.1 s−1, ka = 5 s−1. The bad system(dashed line) is generated using parameters: εc = 70 mJ/m2, εe = 70mJ/m2, kr = 0.3 s−1, ka = 0.1 s−1. . . . . . . . . . . . . . . . . . . . . . 90
3.3 The change of surface tension with the relative surface area of one cyclefor four different preparations of diluted and concentrated BLES in humid(H) and dry (D) conditions. The compression-relaxation model (CRM)predictions is compared to the experimental points. . . . . . . . . . . . . 94
3.4 The change of surface tension with time of one cycle for four differentpreparations of diluted and concentrated BLES in humid (H) and dry(D) conditions. The compression-relaxation model (CRM) predictions iscompared to the adsorption-limited model (ALM), the diffusion model(DM), and the experimental points. . . . . . . . . . . . . . . . . . . . . 95
xi
3.5 A special case of collapse at high surface tension of diluted BLES mixedwith albumin in humid (H) condition: (a) the change of surface tensionwith the relative surface area of one cycle; (b) the change of surface tensionand surface area with time of one cycle. CRM: compression-relaxationmodel, ALM: adsorption-limited model, DM: diffusion model. . . . . . . 97
3.6 Comparison between the goodness of fit (mean and standard deviation) ofdifferent models. CRM: compression-relaxation model, ALM: adsorption-limited model, DM: diffusion model. . . . . . . . . . . . . . . . . . . . . . 99
3.7 Comparison between the predicted value of elasticity (mean and standarddeviation) using different models. CRM: compression-relaxation model,ALM: adsorption-limited model, DM: diffusion model. . . . . . . . . . . . 100
3.8 The minimum surface tension, γmin, measured for four different concen-trations of BLES (2, 8, 15, 27 mg/ml) at different compression ratios(10%, 20%, 30%) and different cycling conditions (1, 3, 9 s/cycle). Theerror bars indicate the standard deviation between runs repeated underthe same conditions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
3.9 The change in surface tension with time and with relative surface areaduring dynamic cycling of BLES at 10% compression and different con-centrations and periodicities: (a) 2 mg/ml and 1 s/cyc; (b) 2 mg/ml and9 s/cyc; (c) 27 mg/ml and 1 s/cyc; (d) 27 mg/ml and 9 s/cyc. Sym-bols show the measured surface tension values and lines show the valuesobtained from CRM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
3.10 The change in surface tension with time and with relative surface areaduring dynamic cycling of BLES at 30% compression and different con-centrations and periodicities: (a) 2 mg/ml and 1 s/cyc; (b) 2 mg/ml and9 s/cyc; (c) 27 mg/ml and 1 s/cyc; (d) 27 mg/ml and 9 s/cyc. Sym-bols show the measured surface tension values and lines show the valuesobtained from CRM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
3.11 The goodness of fit for CRM measured for four different concentrationsof BLES (2, 8, 15, 27 mg/ml) at different compression ratios (10%, 20%,30%) and different cycling conditions (1, 3, 9 s/cycle). The error barsindicate the standard deviation between runs repeated under the sameconditions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
3.12 The elasticity of compression, εc, calculated from CRM for four differentconcentrations of BLES (2, 8, 15, 27 mg/ml) at different compressionratios (10%, 20%, 30%) and different cycling conditions (1, 3, 9 s/cycle).The asterisks show the range of insensitivity (explained in figure 3.17):* indicate points with range of insensitivity bigger than 0.2, ** indicatepoints with range of insensitivity bigger than 0.5. . . . . . . . . . . . . . 116
3.13 The relaxation coefficient, kr, calculated from CRM for four different con-centrations of BLES (2, 8, 15, 27 mg/ml) at different compression ratios(10%, 20%, 30%) and different cycling conditions (1, 3, 9 s/cycle). Theasterisks show the range of insensitivity (explained in figure 3.17): * indi-cate points with range of insensitivity bigger than 0.2, ** indicate pointswith range of insensitivity bigger than 0.5. . . . . . . . . . . . . . . . . . 117
xii
3.14 The elasticity of expansion, εe, calculated from CRM for four differentconcentrations of BLES (2, 8, 15, 27 mg/ml) at different compressionratios (10%, 20%, 30%) and different cycling conditions (1, 3, 9 s/cycle).The asterisks show the range of insensitivity (explained in figure 3.17):* indicate points with range of insensitivity bigger than 0.2, ** indicatepoints with range of insensitivity bigger than 0.5. . . . . . . . . . . . . . 118
3.15 The adsorption coefficient, ka, calculated from CRM for four differentconcentrations of BLES (2, 8, 15, 27 mg/ml) at different compressionratios (10%, 20%, 30%) and different cycling conditions (1, 3, 9 s/cycle).The asterisks show the range of insensitivity (explained in figure 3.17):* indicate points with range of insensitivity bigger than 0.2, ** indicatepoints with range of insensitivity bigger than 0.5. . . . . . . . . . . . . . 119
3.16 Contour lines of the range of insensitivity for each of the parameters: (a)εc, (b) εe, (c) kr, (d) ka, at different periodicities, compression ratios andconcentrations. The calculation of the range of insensitivity is explainedin figure 3.17. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
3.17 Calculation of the range of insensitivity for the elasticity of the compressionεc and the adsorption coefficient ka of one cycle for 2.0 mg/ml BLES at20% compression ratio with a periodicity of 9 seconds per cycle: (a) Thechange of goodness of fit with the change of εc around the calculated valuewhile keeping the rest of the parameters (εe, kr, ka) constant. (b) Thechange of goodness of fit with the change of ka around the calculatedvalue while keeping the rest of the parameters (εc, εe, kr) constant. (c)The relative change of sensitivity with the relative change of εc. (d) Therelative change of sensitivity with the relative change of ka. The range ofinsensitivity is the range of percentage change of the parameter in whichthe sensitivity is below 1. . . . . . . . . . . . . . . . . . . . . . . . . . . 127
4.1 The change of surface tension with drop volume of a constrained sessiledrop of OMCTS formed on top of a holder with outer diameter of 7.86 mm.131
4.2 Schematic of projected areas of a pendant drop profile and a circle. . . . 135
4.3 The change of the apex curvature of a constrained sessile drop with surfacetension and drop volume for two different holder sizes: (a) outer radius of3 mm, and outer radius of (b) 6 mm. . . . . . . . . . . . . . . . . . . . 142
4.4 The change of the dimensionless apex curvature of a constrained sessiledrop with Bond number and dimensionless drop volume. The same rela-tionship can be obtained using any holder size. . . . . . . . . . . . . . . 144
4.5 The shape of a constrained sessile drop for different holder sizes and surfacetension values. Bond number is indicated on top of every figure. Theopen symbols indicate a drop profile with a small volume, while the solidsymbols symbols that with a large volume. The lines indicate the bestcircle that fit the profile. . . . . . . . . . . . . . . . . . . . . . . . . . . 145
4.6 The change of the shape parameter of a constrained sessile drop withBond number and dimensionless drop volume. The same relationship canbe obtained using any holder size. . . . . . . . . . . . . . . . . . . . . . 147
xiii
4.7 Schematic diagram of ADSA–CSD experimental setup. . . . . . . . . . . 149
4.8 The change of surface tension, drop surface area, drop volume, and thecurvature at drop apex with time of a constrained sessile drop of OMCTSformed on top of a holder with outer diameter of 7.86 mm. . . . . . . . 150
4.9 The error in surface tension measured by ADSA as a function of the shapeparameter for different drop sizes of a constrained sessile drop of OMCTSformed on top of a holder with outer diameter of 7.86 mm. For a maximumerror of ±0.1 mJ/m2, the critical shape parameter is 0.018. . . . . . . . 152
4.10 The error in surface tension measured by ADSA as a function of the shapeparameter for different drop sizes of a constrained sessile drop of: (a)OMCTS formed on top of a holder with outer diameter of 6.55 mm, and(b) DEP formed on top of a holder with outer diameter of 7.86 mm. Fora maximum error of ±0.1 mJ/m2, the critical shape parameters are 0.026and 0.027, respectively. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
4.11 The critical shape parameter as a function of the Bond number. Theexperimental points were fitted to a linear function (continuous line). . . 156
4.12 The change of the apex curvature of a pendant drop with surface tensionand drop volume for two different holder sizes: (a) outer radius of 1.5 mm;(b) outer radius of 3 mm. . . . . . . . . . . . . . . . . . . . . . . . . . . 160
4.13 The change of the dimensionless apex curvature, bd, of a pendant dropwith Bond number, Bo, and dimensionless drop volume, Vd. The Bondnumber, Bo = ∆ρgR2
h/γ, is the ratio of body forces (gravitational) tosurface tension forces using the holder’s radius as the length scale. . . . . 161
4.14 The shape of a pendant drop for different holder sizes and surface tensionvalues. Bond number is indicated on top of every figure. The symbolsindicate the drop profile, and the lines indicate the best circle that fit theprofile. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
4.15 The change of the shape parameter of a pendant drop with Bond numberand dimensionless drop volume. . . . . . . . . . . . . . . . . . . . . . . 164
4.16 The change of (a) the maximum dimensionless drop volume, and (b) theapparent contact angle of a pendant drop with Bond number. . . . . . . 166
4.17 The shape of a pendant drop with the maximum dimensionless drop vol-ume for different Bond numbers. The symbols indicate the drop profile,and the lines indicate the best circle that fit the profile. . . . . . . . . . 168
4.18 The change of the shape parameter of pendant drops with the maximumdimensionless drop volume with Bond number. . . . . . . . . . . . . . . 169
4.19 Schematic diagram of ADSA–PD experimental setup. . . . . . . . . . . 170
4.20 The change of surface tension, drop surface area, drop volume, and thecurvature at drop apex with time of a pendant drop of hexadecane formedusing a holder with outer diameter of 6.25 mm. . . . . . . . . . . . . . . 172
4.21 The error in surface tension measured by ADSA as a function of the shapeparameter for different drop sizes of a pendant drop of hexadecane formedusing a holder with outer diameter of 6.25 mm. For a maximum error of±0.1 mJ/m2, the critical shape parameter is 0.04. . . . . . . . . . . . . 173
xiv
4.22 The error in surface tension measured by ADSA as a function of the shapeparameter for different drop sizes of a pendant drop of: (a) hexadecaneformed using a holder with outer diameter of 4.24 mm, and (b) DEPformed using a holder with outer diameter of 4.24 mm. For a maximumerror of ±0.1 mJ/m2, the critical shape parameters are 0.052 and 0.039,respectively. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
4.23 The critical shape parameter as a function of the Bond number. Theexperimental points were fitted to a linear function (continuous line). . . 177
4.24 The change of surface tension with time of a pendant drop of hexadecaneformed using a holder with outer diameter of 6.25 mm. The calculationsare repeated using different software options; ADSA: the default optionsusing Canny for edge detection and Levenberg-Marquardt for optimiza-tion; ADSASUS: SUSAN is used instead of Canny; ADSANM : Nelder-Mead is used instead of Levenberg-Marquardt; TIFA: theoretical imagefitting analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
4.25 The error in measured surface tension as a function of the shape parameterfor different drop sizes of a pendant drop of hexadecane formed using aholder with outer diameter of 6.25 mm. The calculations are repeatedusing different software options; ADSA: the default options using Cannyfor edge detection and Levenberg-Marquardt for optimization; ADSASUS:SUSAN is used instead of Canny; ADSANM : Nelder-Mead is used insteadof Levenberg-Marquardt; TIFA: theoretical image fitting analysis. . . . . 183
A.1 General classification of commercial lung extracts. . . . . . . . . . . . . 220A.2 Extraction procedure for surfactants extract of lung lavage. . . . . . . . 221A.3 Extraction procedure for surfactants extract of lung tissue. . . . . . . . 222A.4 Composition of natural surfactants extracts. . . . . . . . . . . . . . . . . 223A.5 Composition of synthetic surfactants. . . . . . . . . . . . . . . . . . . . 227
xv
Chapter 1
Introduction
1.1 Overview
The significance of interfacial and capillarity phenomena in surface science has been
increasingly recognized in recent years. These phenomena occur whenever a liquid is
in contact with another fluid or solid. Common examples are menisci, liquid drops
formed in either air or another liquid, and thin films such as soap bubbles. Surface
phenomena play an important role in many engineering applications and technologies
including biotechnology, surface engineering, and micro/nanotechnology.
In the present context, the goal is to adapt a methodology for surface tension measure-
ment called Axisymmetric Drop Shape Analysis (ADSA) to a biomedical development in
the area of lung physiology, specifically lung surfactant, i.e., the material that coats and
facilitates the functioning of the lungs of all mammals. A brief overview of lung surfactant
is given first followed by a discussion of surface tension measurement techniques.
Lung surfactants are mixtures of lipids and proteins which form a thin surface film or
a monolayer at the alveolar liquid-air interface. The key property of a functioning lung
surfactant is its surface tension. This film modulates the surface tension of the lung,
lowering the normal air-water surface tension of approximately 70 mJ/m2 to extremely
1
Chapter 1. Introduction 2
low values below 1 mJ/m2. By lowering and varying the surface tension during breathing,
lung surfactant reduces the work needed for respiration and stabilizes the alveoli against
collapse and overexpansion [1]. The extremely low surface tension values are difficult to
measure; but certain techniques are capable of performing such measurements.
Deficiency of the lung surfactant, due to injury or lack of development in premature
babies, is a relatively frequent cause of death. The lack of effective surfactant in pre-
mature babies results in neonatal respiratory distress syndrome (nRDS). This is usually
treated clinically by surfactant replacement therapy using surfactant preparations such as
Bovine Lipid Extract Surfactant (BLES). Acute respiratory distress syndrome (ARDS) is
a frequent, life-threatening disease in which a significant increase in alveolar surface ten-
sion has been observed. However, in this case, it is known that poor performance of the
surfactants used for ARDS patients is due to inhibition or inactivation of the surfactant
by plasma proteins and lipids.
Drop shape techniques, based on the shape of a pendant drop, sessile drop or captive
bubble, are extensively used for surface tension measurement. Early efforts in the analysis
started by understanding the shapes of pendant drops and predicting the shapes for a
given surface tension value [2]. Later, certain dimensions of the shape were tabulated
along with the corresponding surface tension value. Consequently, the surface tension can
be calculated through interpolation using these tables [3]. A more sophisticated method
was later developed based on comparing a number of selected and measured points on the
drop periphery to calculated drop shapes and hence determine the surface tension [4, 5].
An alternative method uses a semi-empirical equation to approximate surface tension
from the total height and maximum diameter of a sessile drop or captive bubble [6].
The shape of the drop/bubble depends on the balance between gravity and surface
tension. The balance between surface tension and external forces, e.g. gravity, is reflected
mathematically in the Laplace equation of capillarity. When gravitational and surface
tension effects are comparable, then, in principle, one can determine the surface tension
Chapter 1. Introduction 3
from an analysis of the shape of the drop/bubble. The surface tension tends to round the
drop, whereas gravity deforms it, i.e. gravity elongates a pendant drop or flattens a sessile
drop. Whenever the surface tension effect is much higher than the gravitational effect, the
shape tends to become spherical in the case of both pendant and sessile drops/bubbles.
Theoretically, each drop shape corresponds to a certain surface tension value.
Axisymmetric Drop Shape Analysis (ADSA) [7–11] is a drop shape technique now
extensively used for surface tension and contact angle measurement. ADSA, based on
the shape of liquid/fluid interfaces, is complex but is adaptable to a variety of experi-
mental circumstances including pendant drops, sessile drops and bubbles. Briefly, ADSA
matches the drop/bubble profile extracted from experimental images to a theoretical
Laplacian curve for known surface tension values using a nonlinear optimization pro-
cedure [7–9]. An objective function is used to evaluate the discrepancy between the
theoretical Laplacian curve and the actual profile. This objective function is the sum
of the squares of the normal distances between the measured points (i.e. experimental
curve) and the calculated curve [9]. The optimization procedure minimizes the objective
function and, hence, finds the surface tension value corresponding to the extracted profile
from experimental images.
In this thesis, the goal is to adapt ADSA to biomedical lung surfactant development.
Certain configurations of ADSA are capable of measuring the extremely low surface ten-
sion values which are observed. Although these configurations are capable and most
suitable for surface tension measurement of lung surfactants, they had unexplored po-
tential to allow for more extensive evaluation of surface properties. Three main needs
arise in this area: First, access to the interface through the gas and liquid phases is
needed to evaluate the role that lipids, gases and inhibitors play in the surface activity
of lung surfactants. Second, a method to identify a set of characteristic properties of
the surfactant preparations is needed in a way that they are independent of compression
conditions and that can be used to compare the surface activity of different formula-
Chapter 1. Introduction 4
tions. Finally, given that the surface tension of lung surfactants changes in a dynamic
experiment, it is necessary to confirm the accuracy of the surface tension measurements
obtained via ADSA. Therefore, three main areas are developed in this thesis in response
to these needs: First, the hardware is redesigned to allow complete access to the interface
from both the air and liquid sides. Second, a detailed analysis of the dynamic surface
tension measurements is used to evaluate the intrinsic properties of the lung surfactant
film. These film properties are elasticity, relaxation and adsorption. Finally, the accuracy
of the surface tension measurements is analyzed and has been re-evaluated. Based on
this analysis, the design of certain configurations of ADSA can be optimized to facilitate
more accurate surface tension measurements.
1.2 Background
In this section a detailed background is provided for the different techniques of surface
tension measurements and natural lung surfactant. In Section 1.2.1, various in vitro tech-
niques for the evaluation of surface tension of lung surfactants are discussed. Section 1.2.2
provides more background on the composition of natural lung surfactant.
1.2.1 Evaluation of Surface Tension
Various methods and techniques have been developed for evaluating the surface activity
of lung surfactants. These methods usually belong to one of three categories: in vitro,
in situ, and in vivo techniques. In vivo techniques involve experimentation done on
the living lung of an animal. In contrast to the surface tension as an output from in
vitro and in situ techniques, the main outputs of in vivo techniques are usually the lung
compliance (the ability of the lungs to stretch due to a change in volume as a function of
an applied change in pressure) and blood oxygenation (the partial pressure of oxygen in
blood sample). Different animal models have been developed to evaluate the efficiency
Chapter 1. Introduction 5
Figure 1.1: Schematic diagram of a Langmuir-Wilhelmy Balance (LWB).
of a lung surfactant preparation on premature and adult animals. In situ techniques are
used to estimate the alveolar surface tension in excised lungs, i.e. in its natural setting
without moving it to some special medium. In situ techniques are usually considered as
an intermediate between in vivo and in vitro.
The focus of this thesis is the in vitro techniques for the evaluation of surface tension
of lung surfactants. These are the most commonly used techniques for assessing their
performance. In such techniques, the surface properties (such as the minimum surface
tension and adsorption rates at a specific cycling conditions) of the lung surfactants are
assessed by measuring the surface tension in a controlled environment experiment. In
this section, various in vitro techniques are discussed.
1.2.1.1 Langmuir-Wilhelmy Balance
In a Langmuir-Wilhelmy Balance (LWB) [12–17] an insoluble surfactant monolayer is
spread on top of a liquid subphase filling a trough, and then compressed by means of
hydrophobic barriers on the surface. During compression, surface pressure is typically
monitored by a Wilhelmy plate system as shown in figure 1.1.
Chapter 1. Introduction 6
The net force acting on the Wilhelmy slide is the sum of forces from surface tension,
gravity and buoyancy; this net force can be calculated as follows:
F = 2γ(W + T ) cos θ + (WTL)ρsg − (WTh)ρwg (1.1)
where W , T , and L are the width, thickness and length of the slide, h is the wetted
height of the slide, γ is the surface tension of the interface, θ is the contact angle, g is the
gravitational acceleration, ρs is the density of the slide material, and ρw is the density of
the subphase.
More commonly, a modified expression is used giving the difference in force (∆F ) be-
tween the case of the slide in pure subphase (no surfactant) and the case where surfactant
is present at the interface,
∆F = 2(γo − γ)(W + T ) cos θ (1.2)
where γo is the surface tension of the pure subphase and γ is the surface tension of
subphase with surfactant at the interface. This modified expression can also be expressed
in terms of surface pressure,
π = γo − γ =∆F
2(W + T ) cos θ(1.3)
Equation 1.3 is usually simplified by assuming the the slide thickness T is negligible
compared to the slide width W , and by assuming complete wetting so that the contact
angle is zero, i.e. cos θ = 1.
The Langmuir film balance has been widely used to study the π-A isotherm of many
spread monolayers including octadecanol [18], dipalmitoyl-phosphatidyl-choline (DPPC)
[19, 20], and dipalmitoyl-phosphatidyl-glycerol (DPPG) [21, 22]. It has been also used
for adsorbed films [14, 17]. The main advantages of this method are accurate control
Chapter 1. Introduction 7
of molecular area, easy integration with microscopy techniques such as Brewster Angle
Microscopy (BAM) and Atomic Force Microscopy (AFM), and wide commercial avail-
ability.
However, the Langmuir film balance has some limitations. First, it requires a rel-
atively large amount of liquid sample, usually no less than several tens of milliliters.
Second, it only allows relatively slow compression-expansions approximately 5 minutes
per cycle [1] as fast cycling creates waves at the air-water interface, which interfere with
the surface tension measurement. This limitation becomes significant in the study of
lung surfactant where experiments should mimic the frequency of the normal breathing,
i.e. human respiration action, at high rates of approximately 1.5 to 3 seconds per cycle
[1].
The third limitation is film leakage. Since lung surfactant and phospholipid films can
reach very low surface tensions (e.g. < 5 mJ/m2) under dynamic compression, the liquid
subphase is able to wet even very hydrophobic materials such as Teflon, the material out
of which most film balances are constructed. Film leakage may occur both at the trough
walls and at the barriers, either above the water level (at the air-solid interfaces) or below
the water level (at the liquid-solid interfaces, in the case of soluble surfactants). Leakage
at the air-solid interfaces can be reduced using tightly fitted barriers [23] or continuous
Teflon ribbons [14, 24]. Leakage at the liquid-solid interfaces can be reduced by priming
the Teflon walls with an alcoholic solution of lanthanum-chloride and long-chain satu-
rated phosphatidyl-choline [16]. A fourth limitation is the difficulty of controlling the
environmental conditions (temperature and humidity) during the experimentation [25].
1.2.1.2 Pulsating Bubble Surfactometer
In a Pulsating Bubble Surfactometer (PBS) [26–32] an air bubble is formed in a surfactant
suspension by drawing atmospheric air through a capillary tube as shown in figure 1.2.
After adsorption of surfactant, the bubble is pulsated between two set positions. The
Chapter 1. Introduction 8
Figure 1.2: Schematic diagram of a Pulsating Bubble Surfactometer (PBS).
bubble radius is usually monitored, while the pressure in the subphase is measured by
a pressure transducer. The surface tension, γ, is calculated from the measured values
of the pressure drop, ∆P , and the bubble radius, R, using the Laplace equation and
assuming a spherical shape,
γ =(∆P )R
2. (1.4)
This method has several advantages: it is a simple setup and easy to use, it requires
only a small volume of subphase, and it is available commercially. PBS allows fast cycling
to mimic the physiological conditions. A further development of a bulk phase exchange
system for PBS facilitates the study of the effect of a specific inhibitor on a preformed
surfactant film [33]. This system has been used to characterize the inhibition mechanisms
of different inhibitors [33–35].
However, film leakage from the bubble surface into the capillary tube at low surface
tensions is again a common complication [36]. Other limitations are the inaccurate
calculation of low surface tension values due to the spherical shape assumption [27], and
Chapter 1. Introduction 9
Figure 1.3: Schematic diagram of a Captive Bubble Surfactometer (CBS).
the difficulty of controlling the environmental conditions (specially humidity) which are
essential in lung surfactant studies.
1.2.1.3 Captive Bubble Surfactometer
In a Captive Bubble Surfactometer (CBS) [28, 37–43] an air bubble is formed in a closed
chamber filled with surfactant suspension. The air bubble floats against a hydrophobic
ceiling coated with 1% agarose gel as shown in figure 1.3. After film formation (spreading
or adsorption), the bubble can be dynamically compressed by changing the chamber
pressure.
In this case, the bubble is not expected to be spherical due to the relatively large
size compared to PBS. The shape of the interface is controlled by the balance between
surface tension forces and gravity as given by the Laplace equation,
γ(1
R1
+1
R2
) =2γ
R0
+ (∆ρ)gz (1.5)
where R1 and R2 are the principal radii of curvature at a point on the interface, R0 is
Chapter 1. Introduction 10
the symmetric radius of curvature at the apex point of the interface, ∆ρ is the density
difference across the interface, g is the local gravitational acceleration, and z is the
vertical distance between the apex point and the point on the interface. Therefore, the
surface tension value, γ, can be determined from the shape of the interface. In CBS, the
height-to-diameter ratio of the bubble is used to calculate surface tension, bubble area,
and volume [40].
The CBS has all the advantages of the PBS setup and solves the problem of leakage
because the bubble is not in contact with any solid surfaces. Interpretation of CBS
does not assume a spherical shape. Nevertheless, the captive bubble film balance has
some limitations. First, any study that requires spreading of a monolayer on the bubble
interface is not trivial. Second, the small volume of the air bubble complicates the
evaporation of the spreading solvent in the study of spread films; the non-evaporated
portion of the solvent will influence the performance of the monolayer at the interface
[44, 45]. Also, controlling the humidity in a captive bubble is relatively difficult and it
is known that such environmental conditions are vital and have significant effect on lung
surfactant films [46, 47]. Another limitation is that it is inoperable at the high surfactant
concentrations used clinically due to visual opaqueness of lung surfactant preparations
at high concentrations (> 3 mg/ml) [48]. Recent experiments use sub-phase spreading
of highly concentrated suspensions near the interface [49]; however, the validity of that
approach is yet to be determined particularly in the presence of water-soluble/dispersable
surfactant inhibitors.
1.2.1.4 ADSA–CB
Axisymmetric Drop Shape Analysis (ADSA) has been used for measuring surface tension
and contact angles [7–9]. Currently, there exist three constellations of ADSA for surface
tension measurements: a captive bubble (ADSA–CB), a pendant drop (ADSA–PD) and
a constrained sessile drop (ADSA–CSD) as shown in figure 1.4. There is also the un-
Chapter 1. Introduction 11
Fig
ure
1.4:
Sch
emat
icdia
gram
ofdiff
eren
tco
nst
ella
tion
sof
Axis
ym
met
ric
Dro
pShap
eA
nal
ysi
s(A
DSA
).
Chapter 1. Introduction 12
constrained sessile drop constellation (ADSA–SD) that is used mainly for contact angle
measurements [50–52]. As mentioned earlier, ADSA software extracts an experimental
drop profile from a digitized drop/bubble image and uses an optimization procedure to
match the extracted profile to a theoretical one calculated from the Laplace equation
of capillarity. The surface tension of the bubble/drop is then calculated based on the
matched theoretical profile. Other outputs of ADSA are drop volume, drop surface area,
contact angle (if applicable), and radius of curvature at the drop/bubble apex.
The first embodiment of ADSA working in conjunction with a Captive Bubble (ADSA–
CB) [41, 46, 47, 53–56] has a very similar setup compared to CBS as shown in figure 1.4.
However, ADSA–CB uses the more accurate and automatic ADSA software for deter-
mining the surface tension, surface area and volume from a bubble image, rather than
just using the height-to-diameter ratio of the bubble needed in CBS. The technique has
now been widely used for soluble films [46, 47], and also as a film balance, especially for
DPPC [44, 45, 57, 58].
ADSA–CB has the same advantages as CBS but provides a greater accuracy in the
calculation of the surface tension and faster data acquisition and processing. On the other
hand, ADSA–CB has the same limitations as CBS: the setup is not trivial especially for
spread films, it is inoperable at the high surfactant concentrations (> 3 mg/ml) used
clinically as the solutions are usually opaque, and controlling the humidity is relatively
difficult.
1.2.1.5 ADSA–PD
A second embodiment of ADSA working in conjunction with a Pendant Drop (ADSA–
PD) [59–63] has also been developed. In this method, a pendant drop of surfactant
suspension is formed at the end of a teflon capillary tube as shown in figure 1.4. Surface
tension, surface area and drop volume are calculated by ADSA. Alternatively, a small
flat circular holder (inverted pedestal) with a sharp-knife edge is used instead of the
Chapter 1. Introduction 13
capillary [59, 64]. This holder was introduced originally in the constrained sessile drop
constellation as will be explained later, and was later used in ADSA–PD for studying
surfactant films at the interface to prevent the film from spreading on to the solid surface
at low surface tension. In addition to the apparent advantages of simplicity and flexibility,
the pendant drop method is believed to have very high accuracy [59].
ADSA–PD has been used extensively to study the π-A isotherm of several monolayers
including octadecanol [59, 65], Dipalmitoyl-phosphatidyl-choline (DPPC) [62, 66–69],
and Dipalmitoyl-phosphatidyl-glycerol (DPPG) [70]. In this method, a pendant drop is
formed at the end of a teflon capillary tube and a known amount of insoluble substance
is spread onto its surface. The π − A isotherms are generated by changing the drop
volume in a controlled manner causing a decrease in the drop surface area and hence
compression of the film. ADSA–PD is able to achieve high rates of compression and
hence has advantages over a Langmuir film balance for certain studies. Nevertheless, the
method still has some limitations. The deposition procedure used in the pendant drop
constellation is not straightforward. This deposition can be performed in one of two ways;
in both strategies, the surfactant solution is delivered through a micro syringe needle. In
one strategy, the needle is brought as close as possible to the side of the hanging drop;
but apparently, penetration of the needle into the bulk phase was inevitable [67] and the
molecules are deposited into the bulk phase and it is uncertain whether they all diffuse
to the surface. In the alternative procedure, the needle is used to deposit the surfactant
solution on the teflon capillary above the pendant drop. The solution then flows down
over the drop surface. Remaining surfactant molecules on the teflon capillary can lead
to overestimation of the number of molecules deposited [68].
ADSA–PD was further developed to be used as a penetration film balance using an
arrangement of two co-axial capillaries connected to two branches of a microinjector [60].
This system allows the bulk phase exchange after forming a stable film and was used to
study the interfacial hydrolysis of mixed phospholipid monolayers by porcine pancreatic
Chapter 1. Introduction 14
phospholipase A2 [60]. Although ADSA–PD eliminates the surfactant concentration
limitation, it can suffer from film leakage at low surface tensions [59]. Another limitation
at low surface tension values is that although gravity assures a well deformed drop shape,
it also tends to detach the drop from the capillary holder in the case of big drops [71].
1.2.1.6 ADSA–CSD
A third embodiment of ADSA working in conjunction with a Constrained Sessile Drop
(ADSA–CSD) [48, 72–75] is now extensively used for determining the dynamic surface
tension of lung surfactants. In this setup, a sessile drop is formed on top of a small
flat pedestal (holder) with a circular sharp knife edge preventing uncontrolled spreading
and hence film leakage (figure 1.4). The sharp edge acts as a barrier to spreading. The
mechanism of this barrier is readily understood in terms of what may be called Young’s
inequality, a generalization of the Young equation: If the three phase line approaches
an edge, Young’s equation will be obeyed. When reaching the edge, motion of the three
phase line will cease and the contact line will not move until the drop has become much
larger. During this time, the apparent contact angle increases. When the contact angle
becomes so large that it becomes equal to the Young angle on the adjacent surface, motion
of the three phase line will continue, and, from the point of view of the surface tension
measurement, undesirable processes occur, such as film spreading. However, there will be
a large range of drop sizes over which the volume of the drop can be changed at virtually
constant drop diameter, offering an opportunity for film balance type of work. Hence,
the apparent contact angle will change with the drop volume.
This holder was introduced originally for the study of density and surface tension of
polymer melts [73]. Due to the presence of the sharp edge, ADSA–CSD proved to be
very suitable for a wide range of surface tension measurements including very low values,
down to below 1 mJ/m2, that are difficult to measure with other setups. ADSA–CSD
is used to perform dynamic cycling by successive compression/expansion of the drop via
Chapter 1. Introduction 15
programmed cycling of the drop volume. This is usually done through several cycles
(normally 20) with prescribed periodicity and compression ratio (percentage of surface
area reduction). In such experiments, a large amount of data is collected including the
surface tension as a function of time as well as the change of the drop surface area and
volume.
ADSA–CSD has been used so far only for the study of films adsorbed from solution
(soluble films). Matters such as the effect of humidity on film formation [25], and the
development of polymeric additives for clinical lung surfactants [76] have been considered.
ADSA–CSD studies were performed at the high rates of compression which mimic human
breathing and which are not achievable using a Langmuir setup. A specially-designed
environmental control chamber for ADSA–CSD facilitates the control of gas composition,
temperature and humidity to mimic the physiologically relevant conditions.
ADSA–CSD is believed to be free of all restrictions and limitations of other conven-
tional methods. The main advantages are that it is very simple to setup and is easy
to use and only small liquid volumes (microliter drops) are required to perform very
accurate surface tension measurements. In the context of lung surfactants, ADSA–CSD
is capable of simulating physiological conditions in the lungs, i.e. the cycling rate to
simulate breathing, environmental and temperature conditions. However, ADSA–CSD
has no concentration limitation of surfactant suspension. Other unexplored potentials:
it is possible to perform high frequency cycling and it is possible to access the air/liquid
interface from both sides to allow spreading and bulk replacement. Although such poten-
tial would be useful, especially in the lung surfactant context, it has not been explored
so far.
1.2.2 Pulmonary Surfactants
Pulmonary surfactants are mixtures of lipids and proteins forming a monolayer at the
alveolar liquid-air interface. Pulmonary surfactant consists of about 90% lipids and about
Chapter 1. Introduction 16
10% proteins. Four surfactant apoproteins are present in native surfactants, called SP-A,
SP-B, SP-C, and SP-D. Approximately 10-20% of the lipids are neutral, and the rest (80-
90%) are phospholipids (PL). Phosphatidylcholine (PC) is the predominant PL species
(about 80%) followed by phosphatidylglycerol (PG) which comprises more than 10% of
PL. Dipalmitoylphosphatidylcholine (DPPC) comprises about 50% of PC, i.e. about
40% of the total PL. The other main lipid component is dipalmitoylphosphatidylglycerol
(DPPG) which comprises almost 10-40% of PG [1, 36, 77, 78]. Native human lung
surfactant is near the high end of these percentages. [78]
A brief review of the commercial exogenous lung extracts is also provided in Ap-
pendix A. Commercial lung extracts are divided into two main categories according to
the source of the preparation: natural lung surfactants which contain exogenous lung
extracts from animals and synthetic lung surfactants which do not contain any natural
extracts. In this thesis, a clinical exogenous lung surfactant called Bovine Lipid Extract
Surfactant (BLES) is chosen as a model lung surfactant. BLES is made from lung sur-
factant lavaged from cows and is produced locally in Canada. A detailed list of other
clinically used lung surfactants worldwide is given Appendix A. It is noted that DP-
PC/PG mixtures are used as the main component of many clinical synthetic exogenous
lung surfactants, e.g. ALEC (artificial lung expanding compound) [79–81], Surfaxin (KL4
or lucinactant) [82–84], and Venticute (Recombinant SP-C or lusupultide) [85–87].
1.3 Scope and Purpose of Thesis
The objective of this thesis is to further develop and automate ADSA–CSD to allow fur-
ther studies of lung surfactants. As discussed in Section 1.2.1, ADSA–CSD can be applied
– in principle – to lung surfactant studies at relevant physiological conditions which in-
clude appropriate temperature, relative humidity, gas composition, pollution, surfactant
concentration, respiration periodicity, and compression ratio. Until now, ADSA–CSD
Chapter 1. Introduction 17
has been used only for the study of films adsorbed from solution (soluble films) and
matters such as the effect of humidity on film formation [25], and the development of
polymeric additives for clinical lung surfactants [76] have also been considered. However,
ADSA–CSD had some unexplored and undiscovered potential which will be addressed
here. The use of ADSA as a film balance was so far done using the pendant drop constel-
lation and there were considerable experimental difficulties [59] that can be avoided in a
constrained sessile drop setup. In this thesis, ADSA–CSD is redesigned to allow access
to the air-liquid interface from either side.
In clinical reality, lung surfactant films can be altered from both the airspace and the
bulk liquid phase that carries the film. Contaminations (pollution) affect the surfactant
interface from the air side and treatment (surfactant therapy for premature babies) is
instilled from the air side, while leakage of plasma and blood proteins usually occurs from
the liquid side, causing severe lung injury. Such inhibitors degrade the surface activity
of the surfactant by impairing the adsorption and the surface tension lowering abilities
of the surfactant. Therefore, accessing the lung surfactant interface from both sides is a
necessary feature in a general in vitro technique used for the assessment of the surface
activity of lung surfactant. In Chapter 2, ADSA–CSD is redesigned to be used as a micro
film balance allowing access to the interface from both the gas-side and the liquid-side.
This allows deposition from the gas side as well as complete exchange of the bulk liquid
phase. The key feature of the design is the insertion of a second capillary into the bulk
of the drop to facilitate addition or removal of a secondary liquid. The new design is
used to study lung surfactant inhibition resistance (subphase injection of inhibitor under
an uninhibited film) and inhibition reversal (subphase injection of surfactant under an
inhibited film). The access from the air-side is illustrated by a series of experiments of
the collapse surface pressure of spread pure monolayers of DPPC and DPPG (the main
constituents of lung surfactants).
To asses the performance of surfactant preparations subject to dynamic compres-
Chapter 1. Introduction 18
sion/expansion, typically only the minimum surface tension at the end of compression is
reported. While the minimum surface tension is useful for a preliminary assessment, it
must be realized that it depends on the compression protocol [46]. Therefore, it is neces-
sary to concentrate on the properties of the film itself such as its elasticity, its formation
(adsorption) and its stability (or relaxation tendency) in addition to, if not in place of,
the minimum surface tension. In Chapter 3, a dynamic compression-relaxation model
(CRM) is developed to describe the mechanical properties of lung surfactant films by
investigating the response of surface tension to changes in surface area. The model treats
the compression/expansion with the conventional definition of dilatational elasticity of
insoluble surfactants, and treats the relaxation and adsorption of the film as first order
processes driven by the excess surface free energy of the film. The model is used to eval-
uate the quality of lung surfactant preparations based on calculating the film properties
independent of the compression protocol used.
The accuracy of the determined mechanical properties of lung surfactant films depends
on the accuracy of the measured response of surface tension to changes in surface area.
Therefore, high accuracy surface tension measurements are necessary in such studies. It
is well documented that the accuracy of the surface tension determined by ADSA can
depend on drop size, unless all drops of different sizes are well deformed and far from
spherical [64]. A detailed analysis of constrained sessile drop shapes and the accuracy
of the measured surface tension values is developed using a shape parameter concept in
Chapter 4. Dimensional analysis is used to describe similarity in constrained sessile drop
shapes and to express the problem using appropriate dimensionless groups. Based on this
analysis, the design of ADSA–CSD is optimized to facilitate more accurate surface tension
measurements. The validity analysis is further extended to the pendant drop (ADSA–
PD). This setup is the most frequently used and most accurate drop shape technique for
surface tension measurement outside the area of lung surfactants.
Chapter 2
ADSA-CSD as Micro Film Balance∗
2.1 Introduction
ADSA–CSD has been extensively used for the study of lung surfactants [91]. Matters such
as the effect of humidity on film formation [25], and the choice of polymeric additives for
clinical lung surfactants [76] have been considered. ADSA–CSD studies were performed
at high rates of compression (to mimic human breathing in lung surfactant studies). A
specially-designed environmental control chamber for ADSA–CSD facilitates the control
of gas composition, temperature and humidity to mimic the physiologically relevant con-
ditions [25]. ADSA–CSD is believed to be free of all restrictions and limitations of other
methods as mentioned in detail in Chapter 1.
However, certain studies cannot be readily performed using the current ADSA–CSD
setup. In the lung surfactant context, certain insoluble molecules such as DPPC and
DPPG are closely related to the surface activity of pulmonary surfactant as shown in
Chapter 1. The study of these insoluble molecules is essential in understanding the
surface properties of lung surfactants. Studying lung surfactant inhibition is also an
important aspect. Such studies are introduced here by removing the bulk phase, after
∗Portions of this chapter were previously published in Refs. [88–90], reproduced with permission.
19
Chapter 2. ADSA-CSD as a Micro Film Balance 20
forming a sessile drop of a basic lung surfactant preparation, and exchanging it for one
containing different inhibitors. This mimics the leakage of plasma and blood proteins into
the alveolar spaces altering the surface activity of lung surfactant in a phenomenon called
surfactant inhibition. These studies cannot be performed using the current ADSA–CSD
setup.
In response to these needs, two main capabilities are added to the ADSA–CSD setup.
The first one is the deposition capability allowing deposition of molecules onto the air/wa-
ter interface from the air side. This allows the study of the collapse of deposited films
through the study of pure monolayers as well as mixed monolayers. DPPC and DPPG
monolayers are studied here to characterize the role of such molecules in maintaining
stable film properties and surface activity of lung surfactant preparations. The second
capability is the double injection, facilitating the complete exchange of the subphase of
a spread or adsorbed film from the liquid side. This feature allows certain studies rele-
vant to lung surfactant research that cannot be readily performed by other means. The
development will be illustrated through studies concerning lung surfactant inhibition.
The redesigned ADSA–CSD is a versatile tool for surface tension measurements that
has a wide range of applications outside the scope of the lung surfactant research area.
For example, the deposition capability allows for the study of properties of insoluble films
such as film relaxation and collapse and interfacial dilational elasticity. The double in-
jection capability allows studies on the interaction between a monolayer and a surfactant
dissolved in the subphase.
2.2 Deposition Capability
The collapse pressure is the highest surface pressure to which a monolayer can be com-
pressed without detectable expulsion of molecules into a new phase [92]. Experimen-
tally, it was found that many monolayers can be compressed to pressures considerably
Chapter 2. ADSA-CSD as a Micro Film Balance 21
higher than their equilibrium spreading pressure [1, 92, 93]. Eventually, however, it was
found impossible to increase the surface pressure further. If compression continues, the
monolayer collapses ejecting surfactant molecules into one or more surface or subsurface
collapse structures or phases [1, 93]. Monolayer collapse is accompanied by a change in
slope or by a horizontal plateau in the surface pressure/area π-A isotherm. The surface
pressure, π, can be computed by calculating the difference between the surface tension
of the pure liquid (without an insoluble monolayer spread at the interface), γo, and the
surface tension with the monolayer spread at the interface, γ, i.e. π = γo − γ.
These isotherms are measured with a surface balance or a film balance. Film balances
have been used to study monolayer properties for different substances, such as the collapse
pressure. The collapse pressure is used as an indication of the monolayer stability; the
higher the collapse pressure, the more stable the monolayer. The first film balance was
developed by Langmuir [12] and modified subsequently by others [92, 93]. The Langmuir
film balance suffers from three main limitations as indicated in detail in Chapter 1. First,
it requires a relatively large amount of liquid sample. Second, it only allows relatively
slow compression-expansions. Third, it suffers from film leakage especially at very low
surface tensions (very high surface pressures). These limitations become significant in
the study of lung surfactant.
The pendant drop and the captive bubble techniques have also been used to study
the π-A isotherm of several monolayers as shown in Chapter 1. Nevertheless, both
techniques have one or more of the following limitations when used as as film balance.
In both techniques, the deposition procedure to spread a monolayer on the bubble/drop
interface is neither straightforward nor trivial. In the pendant drop technique, film
leakage usually occurs at low surface tensions (high surface pressures); and gravity tends
to detach the drop from the capillary holder in the case of big drops [71]. In the captive
bubble technique, controlling the humidity is a difficult matter. More details regarding
these limitations were given in Chapter 1.
Chapter 2. ADSA-CSD as a Micro Film Balance 22
ADSA–CSD has been used so far only for the study of films adsorbed from solution
(soluble films). Here, the usage of ADSA–CSD as a film balance is illustrated. The fact
that the drop is sitting on a pedestal makes the deposition near the drop apex much
easier compared to the case of the pendant drop. Another feature is that gravity will
cause the drop to be well deformed at very low surface tension values; this guarantees
high accuracy of ADSA calculations, especially in the vicinity of the collapse pressure,
where very low surface tension values exist. Using ADSA–CSD, the collapse region will
be studied carefully to gain insight into the collapse details of three different monolayers:
octadecanol, dipalmitoyl-phosphatidyl-choline (DPPC), and dipalmitoyl-phosphatidyl-
glycerol (DPPG).
The reason for choosing DPPC and DPPG is their relevance to pulmonary surfactant
studies as shown in Chapter 1. In the literature, a DPPC/DPPG mixture ratio of 80:20
was used in several studies to mimic natural lung surfactant lipids [94–98]. This ratio
is usually selected because it is similar to the ratio found in a variety of mammalian
lung surfactant extracts. Other studies have used a ratio of 70:30 [99, 100]. Most of
these studies (using a ratio of 80:20 or 70:30) have focused on investigating the effect of
surfactant proteins (SP-A, SP-B, and SP-C) on the properties of a model lung surfactant
composed of only DPPC and DPPG at the specified ratio. Properties at the very high
surface pressures which occur in the lungs were not investigated in any of these studies.
DPPC/PG mixtures are also used as the main component of many clinical synthetic
exogenous lung surfactants, e.g. ALEC (artificial lung expanding compound) [79–81],
Surfaxin (KL4 or lucinactant) [82–84], and Venticute (Recombinant SP-C or lusupultide)
[85–87]. In ALEC, DPPC (70%) is the main component for lowering surface tension,
while egg PG (30%) is used to improve adsorption and spreading [79].
There appears to be no systematic study to date to study the effect of increasing
the DPPG content in a DPPC/DPPG mixture on the ultimate collapse pressure. Here,
ADSA–CSD is also used to study spread monolayers of mixed DPPC/DPPG films. Com-
Chapter 2. ADSA-CSD as a Micro Film Balance 23
pression isotherms were obtained for different DPPC/DPPG mixture ratios with special
focus on the ultimate collapse pressure, film elasticity and film hysteresis for every mix-
ture ratio.
2.2.1 Experimental Details
2.2.1.1 Materials and Methodology
1-Octadecanol was purchased from Sigma-Aldrich (Fluka) Co. (cat. 74720) with a purity
≥99.0%; dipalmitoyl-phosphatidyl-choline (DPPC) from Sigma-Aldrich Co. (cat. P-
4329) with a purity ≥99.0%; and dipalmitoyl-phosphatidyl-glycerol [ammonium salt]
(DPPG) from Sigma-Aldrich Co. (cat. P-5650) with a purity ∼99.0%. All amphiphilic
substances were used without further purification.
Heptane, ethanol, chloroform and methanol were used as spreading agents. Heptane
was purchased from EMD Co. (cat. HX0082-1) with a purity 99.6% (99.95% saturated
C7 hydrocarbons); ethanol from Commercial Alcohols Inc. (cat. UN-1170); chloroform
from Sigma-Aldrich Co. (cat. 36,692-7) with a purity ≥99.8%; and methanol from
Sigma-Aldrich Co. (cat. 27,047-4) with a purity 99.93%.
Ocatadecanol was dissolved in a heptane/ethanol (9:1 v/v) mixture to form a stock
solution with a concentration of 1.2270 mg/ml and then diluted in heptane only to a
concentration of 0.0258 mg/ml. DPPC was dissolved in a heptane/ethanol (9:1 v/v)
mixture to form a stock of concentrated solution and then diluted in a heptane/ethanol
(9:1 v/v) mixture to different concentrations in the range of 0.0025 to 0.0250 mg/ml.
DPPG was dissolved in a chloroform/methanol (9:1 v/v) mixture to form a stock of
concentrated solution and then diluted in a heptane/ethanol (9:1 v/v) mixture to different
concentrations in the range of 0.0025 to 0.0250 mg/ml. The stock and dilute solutions
were always freshly prepared on the day of experiment and subsequently mixed (in the
case of the mixed monolayers study) to give the required mixture DPPC/DPPG ratio
Chapter 2. ADSA-CSD as a Micro Film Balance 24
Diff
user
Light Source
Microscope CCD
Stepping MotorSyringe
Glass Cuvette
Figure 2.1: Schematic diagram of ADSA–CSD experimental setup.
and used in the experiments. The water used as the subphase in the experiments is
demineralized and glass distilled (pH ' 5); the ionic strength is almost negligible.
A schematic diagram of the experimental setup for ADSA–CSD is shown in figure 2.1.
As shown, a constrained sessile drop is formed on a stainless steel pedestal with an outside
diameter of 6.2 mm and an inside diameter of 1.5 mm. A Cohu 4800 monochrome camera
is mounted on a Wild-Apozoom 1:6 microscope. The video signal of the constrained
sessile drop is transmitted to a digital video processor, which performs the frame grabbing
and the digitization of the image.
A computer is used to acquire images from the image processor and to perform image
analysis and computation. The rate of image acquisition for the present experiments is
ten images per second. ADSA software extracts an experimental drop profile from the
digitized image and uses an optimization procedure to match the extracted profile to a
theoretical one calculated from the Laplace equation of capillarity. The surface tension of
the constrained sessile drop is then calculated based on the matched theoretical profile.
Other outputs of ADSA are radius of curvature, drop volume and drop surface area. The
fact that we can measure changes in surface area makes this methodology suited for film
balance experiments.
The volume of the drop is controlled by a syringe connected to a stepper motor.
By moving the syringe plunger, changing the volume changes the surface area of the
constrained sessile drop. Although the surface area is changed indirectly, this change is
Chapter 2. ADSA-CSD as a Micro Film Balance 25
0 20 40 60 80 100 12040
45
50
55
60
65
70
75
Time, t (s)
Sur
face
Ten
sion
, γ (
mJ/
m2 )
Figure 2.2: The change of surface tension with time after the deposition of octadecanolin heptane/ethanol solution.
essentially linear for the relatively small changes in volume [59].
2.2.1.2 Experimental Procedure
A drop of pure water was formed on the stainless steel pedestal which is connected to
the syringe and the stepper motor. A few images were acquired and analyzed by ADSA,
providing the surface tension of the pure water subphase, γo. All experiments were
performed at room temperature of 24oC.
Using a Hamilton microliter syringe with 0.01% accuracy and a micro-manipulator, a
known amount (typically 5 µl) of the diluted surfactant solution was deposited on top of
the sessile drop. After the deposition, complete evaporation of the spreading agent takes
place in approximately 60 seconds. This is shown is figure 2.2, where the surface tension
Chapter 2. ADSA-CSD as a Micro Film Balance 26
Table 2.1: Experimental parameters used for different monolayer types
MonolayerVolume Concentration 10−14 × No. Area Rate Area Rate
(µl) (mg/ml) of Molecules (cm2/min) (A2/molecule·min)
Octadecanol 5.0 0.0258 2.87 0.19 6.65DPPC 5.0 0.0250 1.03 0.19 18.94DPPG 5.0 0.0250 1.02 0.19 19.47
of pure water decreases after deposition of octadecanol in heptane/ethanol solution and
then drifts back up again as the spreading agent evaporates. In the actual experiments,
3 minutes are allowed to elapse to ensure complete evaporation of the spreading agent
before the actual compression is started.
A crucial part of the experiment is the deposition of the insoluble surfactant onto the
drop surface. The constrained sessile drop constellation used here makes the deposition
procedure much easier compared to deposition methods in the pendant drop constellation.
A detailed description of the deposition method used in the current experiments is shown
in figure 2.3. Increasing the size of the drop after the deposition promotes complete
spreading of a monolayer over the whole surface of the drop. After measuring the surface
tension of a pure water drop at full size, the size of the drop is decreased from the full
size (∼80µl) to half size (∼40µl) before deposition. Then the solution deposition (5µl) is
performed from the syringe needle over the water drop apex. Two or more solution drops
fall from the needle onto the drop, causing more than one minimum in the surface tension
in figure 2.2. In some cases, part of the solution dose remains attached to the needle as a
pendant drop; in such cases the needle is moved closer to the drop without penetrating
the interface. The sessile drop is expanded back to full size right after deposition.
At this point, the constrained sessile drop carries an insoluble monolayer and is then
covered with a glass cuvette to isolate the drop from the environment and to prevent
contamination. The surface area was then decreased by reducing the drop volume with
the motor-controlled syringe. For the first study of pure monolayers, table 2.1 summa-
Chapter 2. ADSA-CSD as a Micro Film Balance 27
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
Figure 2.3: The process of solution deposition on a constrained sessile drop: (a) thedrop at the full size (∼80µl) before deposition; (b) the drop at half size (∼40µl) beforedeposition; (c) solution deposition (5µl) where more than one solution drop falls fromthe needle over the apex; (d) part of the solution is still attached to the needle; (e)the needle is moved closer without penetrating the interface; (f) the sessile drop aftercomplete deposition; (g) expansion of the sessile drop back to the full size right afterdeposition; (h) the sessile drop after complete solvent evaporation (∼2min).
Chapter 2. ADSA-CSD as a Micro Film Balance 28
rizes the amounts of deposition as well as the compression rates for different surfactant
types calculated based on equations 2.1 and 2.2 below. For the second study of mixed
monolayers, the deposition amount for all mixtures used is 5.5 µl volume with a mix-
ture concentration of 0.025 mg/ml. The compression rate is fixed at 0.19 cm2/min for
all experiments. During the compression, images were acquired and saved for further
analysis.
All acquired images were analyzed by ADSA, providing the surface tension value,
γ, the drop surface area, A, the drop volume, V , and the radius of curvature at the
apex, Rc. Figure 2.4 shows a typical result of ADSA for the deposition of octadecanol
in heptane/ethanol solution on a water drop. The number of molecules deposited can be
calculated
Molecules =CVdAvMw
(2.1)
where C is the concentration of the dilute solution used, Vd is the volume used in the
deposition, Av is Avogadro’s number, and Mw is the molecular weight of the specific
surfactant. The change of the surface area with time is used to calculate the rate of
molecular area compression
Area
Molecules×Minutes=
(1
Molecules
)(Area
Minutes
)(2.2)
The parameters of all experiments are summarized in Table 2.1. The value of the surface
pressure, π, can be readily computed by calculating the difference between γo and γ, i.e.
π = γo − γ (2.3)
For the mixed monolayer study, there was an interest to see the effect of mixing
DPPC and DPPG on the film elasticity. The dilatational elasticity, ε, is defined as the
magnitude of surface tension reduction (or surface pressure increase) for a given reduction
Chapter 2. ADSA-CSD as a Micro Film Balance 29
0
20
40
60
80
Sur
face
Ten
sion
, γ (
mJ/
m2 )
0.4
0.5
0.6
0.7
0.8
Are
a, A
(cm
2 )
0
20
40
60
80
Vol
ume,
V (
*10−
3 cm
3 )
0 50 100 1500.5
1
1.5
2
2.5
3
Rad
ius
of C
urva
ture
, Rc (
cm)
Time, t (s)
Figure 2.4: The change of surface tension, drop surface area, drop volume and radius ofcurvature at the apex with time of a constrained sessile drop of an octadecanol monolayer.
Chapter 2. ADSA-CSD as a Micro Film Balance 30
of the relative surface area of the film. The film elasticity can be calculated from the
change in surface pressure resulting from a small change in drop surface area:
ε =
∣∣∣∣ dπ
d lnA
∣∣∣∣ =
∣∣∣∣Ad(γo − γ)
dA
∣∣∣∣ (2.4)
where π is the surface pressure, γ and γo are the surface tension values in the presence
and absence of a phospholipid film, respectively, and A is the drop surface area.
2.2.2 Results for Pure Monolayers
Figure 2.5 shows the change of surface pressure with area per molecule for an octadecanol
monolayer at a compression speed of 0.19 cm2/min (6.65 A2/molecule·min). Figure 2.5(a)
shows two runs, illustrating reproducibility. Although ten images were acquired every
second, here only two points per second are shown to enhance graph readability. This
applies for figures 2.5, 2.6 and 2.8. However, figure 2.7 shows the raw data acquired
from ADSA–CSD at the rate of ten images per second. As shown in figure 2.5(a), the
surface pressure starts near 0 mJ/m2 and reaches approximately 61 mJ/m2 as the film
is compressed. Further compression did not raise the surface pressure. A plateau is
observed indicating collapse of the monolayer at this limiting surface pressure. Further
inspection shows that there is a change in the isotherm slope (in the domain of rapid
increasing area, i.e. the liquid condensed phase) starting well below the limiting collapse
pressure. Figure 2.5(b) shows the isotherm of five different runs in the vicinity of the
collapse pressure, where point A indicates the point where the slope of the isotherms
starts to change.
Figure 2.6 shows the change of surface pressure with area per molecule for a DPPC
monolayer with a compression speed of 0.19 cm2/min (18.94 A2/molecule·min). Fig-
ure 2.6(a) shows a typical DPPC isotherm; the surface pressure starts at approximately
5 mJ/m2 and reaches about 70 mJ/m2 as the film is compressed. A plateau is observed
Chapter 2. ADSA-CSD as a Micro Film Balance 31
10 15 20 25 300
10
20
30
40
50
60
70
Area per Molecule, Am
(Å2/molecule)
Sur
face
Pre
ssur
e, π
(m
J/m
2 )
Run 1Run 2
(a)
16.5 17 17.5
50
55
60
Run 1
A
16.5 17 17.5
Run 2
A
16.5 17 17.5
Run 3
A
16.5 17 17.5
Run 4
A
16.5 17 17.5
Run 5
A
Area per Molecule, Am
(Å2/molecule)
Sur
face
Pre
ssur
e, π
(m
J/m
2 )
(b)
Figure 2.5: The change of surface pressure with area per molecule of a constrainedsessile drop of an octadecanol monolayer: (a) the full isotherm for different runs; (b) theisotherm of five different runs in the vicinity of the collapse pressure (only two points persecond shown).
Chapter 2. ADSA-CSD as a Micro Film Balance 32
25 30 35 40 45 50 55 60 65 70 750
10
20
30
40
50
60
70
Area per Molecule, Am
(Å2/molecule)
Sur
face
Pre
ssur
e, π
(m
J/m
2 )
(a)
25 30 35 4045
50
55
60
65
70
Area per Molecule, Am
(Å2/molecule)
Sur
face
Pre
ssur
e, π
(m
J/m
2 )
A
(b)
Figure 2.6: The change of surface pressure with area per molecule of a constrained sessiledrop of a DPPC monolayer: (a) the full isotherm; (b) the isotherm in the vicinity of thecollapse pressure (only two points per second shown).
Chapter 2. ADSA-CSD as a Micro Film Balance 33
indicating the collapse of the monolayer at this very high limiting surface pressure (very
low surface tension, approximately 2 mJ/m2). Such low surface tension was not achiev-
able using a pendant drop [62, 66–69], where gravity will detach the drop at very low
surface tension values. Further inspection shows that there is a change in the isotherm
slope starting at pressures well below the limiting collapse pressure. This is illustrated
in figure 2.6(b) which shows the isotherm in the vicinity of the collapse pressure, where
point A (' 52.8 mJ/m2) indicates the point where the isotherm slope changes.
As shown in figure 2.6(b), the isotherm shows a continuous change starting from
point A till the horizontal plateau indicated by the dotted line. To further understand
the isotherm behaviour in this region, we show in figure 2.7 the change of surface tension
and drop surface area with time in two key regions; the first is the vicinity of point A
shown in figure 2.7(a) and the other is the vicinity of the isotherm horizontal plateau
shown in figure 2.7(b).
In the first region, figure 2.7(a), the surface tension shows episodes of collapse in
terms of sudden changes. The curve before the time window shown is quite smooth.
These changes are in the range of 2 mJ/m2, while it was shown [101] that the error of
the measurement is less than 0.1 mJ/m2. These episodes of collapse are accompanied
by a slight decrease in the drop surface area. The reason is that for the same drop
volume, when the surface tension increases at a collapse episode, the drop becomes more
spherical, decreasing the drop surface area at a given volume.
In the second region, figure 2.7(b), the surface tension is almost constant at a very low
value corresponding to the horizontal plateau in the surface pressure isotherm. The most
interesting part in this figure is the change of drop surface area with time at this very
low surface tension. In this region, further compression cannot increase the surface pres-
sure anymore indicating an ultimate collapse pressure. However, due to the continuous
compression, the drop surface area shows sudden drops. This suggests that molecules are
leaving the interface to the bulk causing a sudden drop in the surface area. This aspect
Chapter 2. ADSA-CSD as a Micro Film Balance 34
20
24
28
32
Sur
face
Ten
sion
, γ (
mJ/
m2 )
115 116 117 118 119 120 121 122 123 124 125
0.35
0.36
0.37
Are
a, A
(cm
2 )
Time, t (s)
(a)
4
8
12
16
Sur
face
Ten
sion
, γ (
mJ/
m2 )
145 146 147 148 149 150 151 152 153 154 155
0.28
0.29
0.30
Are
a, A
(cm
2 )
Time, t (s)
(b)
Figure 2.7: The change of surface tension and drop surface area with time of a constrainedsessile drop of a DPPC monolayer: (a) in the vicinity of initial episodes of collapse; (b)in the vicinity of the ultimate collapse pressure (raw data acquired at the rate of tenimages per second shown).
Chapter 2. ADSA-CSD as a Micro Film Balance 35
Table 2.2: Collapse pressure (mJ/m2) from different runs of different monolayer typesand the average value along with the standard deviation
Monolayer Run 1 Run 2 Run 3 Run 4 Run 5 AverageStandardDeviation
Initial Episode of Collapse
Octadecanol 54.8 53.5 54.5 55.7 54.1 54.5 0.37DPPC 52.8 48.5 44.8 44.2 - 47.9 1.98DPPG 57.8 58.5 59.7 57.4 - 58.4 0.51
Ultimate Collapse Pressure
Octadecanol 61.5 62.1 61.9 60.7 60.3 61.3 0.34DPPC 70.5 70.2 70.1 70.0 - 70.2 0.12DPPG 59.9 58.2 60.7 57.2 - 59.0 0.80
will be treated in more detail in the discussion section.
Figure 2.8 shows the change of surface pressure with area per molecule for a con-
strained sessile drop of a DPPG monolayer with a compression speed of 0.19 cm2/min
(19.47 A2/molecule·min). Figure 2.8(a) shows a typical DPPG isotherm; the surface
pressure starts at approximately 3 mJ/m2 and reaches about 59 mJ/m2 as the film is
compressed. A plateau is observed indicating the collapse of the monolayer at this lim-
iting surface pressure. In contrast to octadecanol and DPPC, DPPG did not show a
sudden change in the isotherm slope before the ultimate collapse pressure. Figure 2.8(b)
shows the isotherm in the vicinity of the collapse pressure, where point A (' 58.5 mJ/m2)
indicates the first point of the change of the isotherm slope.
To illustrate the reproducibility of the results, every experiment was repeated 4 or
5 times. Table 2.2 summarizes the values of the first episode of collapse as well as the
ultimate collapse pressure from different runs for the different monolayers (Octadecanol,
DPPC and DPPG) and the average values along with the standard deviation.
Chapter 2. ADSA-CSD as a Micro Film Balance 36
25 30 35 40 45 50 55 60 65 70 750
10
20
30
40
50
60
70
Area per Molecule, Am
(Å2/molecule)
Sur
face
Pre
ssur
e, π
(m
J/m
2 )
(a)
33 34 35 36 37 38 39 40 41 4250
52
54
56
58
60
62
Area per Molecule, Am
(Å2/molecule)
Sur
face
Pre
ssur
e, π
(m
J/m
2 )
A
(b)
Figure 2.8: The change of surface pressure with area per molecule of a constrained sessiledrop of a DPPG monolayer: (a) the full isotherm; (b) the isotherm in the vicinity of thecollapse pressure (only two points per second shown).
Chapter 2. ADSA-CSD as a Micro Film Balance 37
2.2.3 Discussion for Pure Monolayers
2.2.3.1 Problem of leakage
The results of compressing an octadecanol monolayer are in good agreement with pre-
vious experiments with the pendant drop constellation [59, 65]. The new methodology
proves to be comparable with existing methods with respect to accuracy and reproducibil-
ity. ADSA–CSD shares certain advantages with ADSA–PD over Langmuir film balances.
One of the key advantages is the ability to use higher compression rates, and just as the
Langmuir balance, there is no lower limit for the rate of compression and expansion. An-
other potential strength is that environmental control (contamination, humidity, pressure
and temperature) is straightforward and easy to integrate with the current setup; this
will be a necessity, e.g. for lung surfactant studies [25, 76]. Other advantages are that
only small quantities of liquid and insoluble surfactants are required, and that a very
simple setup is needed and could be integrated with microscopy systems, e.g. BAM or
AFM, for further studies. Compared to ADSA–PD, ADSA–CSD has an easier deposition
procedure.
However, the main advantage here of ADSA–CSD over ADSA–PD and Langmuir
balance is that leakage is eliminated completely. Leakage in the pendant drop can affect
the manifestation of the collapse pressure as a horizontal plateau in the surface pres-
sure/area isotherm. For example, previous results using a pendant drop [65] have shown
the presence of the horizontal plateau for some octadecanol isotherms, while a sudden
decrease in the surface pressure after a certain point was shown for higher compression
speeds. This point was originally considered as the collapse pressure. However, the ef-
fect was recently attributed to monolayer leakage on the teflon capillary supporting the
pendant drop [88]. In fact, all of the studies performed using a pendant drop with a
teflon capillary [59, 62, 65–70] could not achieve any surface pressures larger than about
62 mJ/m2. Although leakage in pendant drops can be eliminated by using a sharp edge
Chapter 2. ADSA-CSD as a Micro Film Balance 38
holder as will be discussed in Chapter 4, the problem with low surface tension values is
partially due to the fact that the gravity forces are more pronounced at these low surface
tension values (high surface pressure values) and would act to detach the drop from the
holder. The current constrained sessile drop setup is free of any of these shortcomings.
It has the ability to measure very low surface tension values as shown in the DPPC re-
sults, and it features much easier deposition procedure and leak-proof design. A further
merit of ADSA–CSD is the ability of performing high rates of compression, e.g. for lung
surfactant studies [25, 76], see Chapter 3.
2.2.3.2 Onset of collapse versus ultimate collapse pressure
Disagreements between investigators about collapse pressure has been considerable [65].
Some consider the surface pressure at which the isotherm slope starts to change to be
the starting point of ejection of molecules from the interface as the collapse pressure in
agreement with the definition given in the introduction [92]. Others would consider a
horizontal plateau in the isotherm as the collapse pressure, while an intermediate point
in between those two case is sometimes also considered [1].
The results presented here show a distinctive difference between the onset of collapse
and the ultimate collapse pressure (ultimate strength) of these films. For example, com-
pression of a DPPC monolayer causes the isotherm to change its slope at a certain point
as shown in figure 2.6(b). This change in the isotherm slope reflects the existence of
some interaction between molecules at the interface causing some episodes of collapse.
This point is called here the onset of collapse. Upon further compression, the mono-
layer adapts and sustains higher surface pressure until reaching a maximum or ultimate
strength or an ultimate collapse pressure. Thus, a collapse region can be defined to start
with an initial episode of collapse (onset of collapse) and to end when the ultimate (or
maximum) collapse pressure is reached, manifested in a horizontal plateau in the surface
pressure/area isotherm. From an applied perspective, e.g. for lung surfactant, we are
Chapter 2. ADSA-CSD as a Micro Film Balance 39
mostly interested in knowing if the film can sustain high surface pressures (e.g. > 67
mJ/m2), i.e. surface tensions that prevent lung collapse.
Whereas a Langmuir film balance controls the surface area and measures the change
in surface pressure, ADSA–CSD measures the change of both surface area and surface
tension with time caused by the externally imposed change of drop volume. This fea-
ture allows ADSA–CSD to provide a detailed study of the collapse region, as shown in
figure 2.7 for DPPC. Figure 2.7(a) shows the change of surface tension and drop surface
area with time in the vicinity of the onset of collapse, while figure 2.7(b) focusses more
on the vicinity of the ultimate collapse pressure. Episodes of collapse occur as indicated
by the change of surface tension in the former; however, the surface tension is almost
constant in the latter. The change of drop surface area provides more information in
both figures.
In figure 2.7(a), the episodes of collapse are accompanied by some irregularities in the
surface area. In figure 2.7(b), while the surface tension is constant, the surface area shows
sudden decreases at some points. The sudden decrease in area suggests squeeze out of
some molecules from the interface. The decrease in surface area can be used to estimate
the size of regions and the number of molecules that are being pushed into the subphase.
For example, from figure 2.7(b) at time 148 seconds the area suddenly drops from 0.300
cm2 to 0.295 cm2. Since the surface tension is essentially constant, this suggests that a
region (or regions) of accumulative area equal to approximately 0.005 cm2 or 5×1013 A2
have been pushed out to the subphase. The minimum area per DPPC molecule is known
[1] to be in the range of 35 to 40 A2/molecule; this gives an estimate of the number of
molecules that have to leave the interface as in the order of 1012 molecules, i.e. 1% of
the total material deposited. Such information might be of specific interest, e.g. for lung
surfactant systems. Obviously, such information cannot be obtained by a Langmuir film
balance.
Chapter 2. ADSA-CSD as a Micro Film Balance 40
2.2.3.3 Comparison of the three systems
Table 2.2 summarizes the values obtained for onset of collapse and ultimate collapse
pressure for the three systems studied here: octadecanol, DPPC and DPPG. Octadecanol
and DPPC show a distinctive difference between the onset of collapse and the ultimate
collapse pressure, while both values are very close for DPPG. In fact, some of the DPPG
runs (run 2 and 4) did not show any sudden change in the isotherm slope until the
ultimate collapse pressure was reached, i.e. there was no separate distinctive onset of
collapse for these runs, as shown in figure 2.8. The values for the ultimate collapse
pressure were well reproducible for all three systems. However, the onset of collapse is
harder to reproduce especially for the DPPC monolayer as reflected in the high value of
the standard deviation.
Octadecanol was found to have an ultimate collapse pressure of 61.3 mJ/m2, DPPC
of 70.2 mJ/m2, and DPPG of 59.0 mJ/m2. The superior ability of DPPC to sustain such
a high pressure may help to understand why it is the main component of lung surfactant.
The difference between the ultimate collapse pressure of these three systems is believed
to be due to several factors, including the molecular structure, the chain length, and
the properties of the head group. DPPC [C40H80NO8P] has two aliphatic chains with
a large choline head group, compared to octadecanol molecules [CH3(CH2)17OH] with a
single alkyl chain and only a small head group. DPPG [C38H75O10P·NH3] has a similar
molecular structure to DPPC; however, unlike DPPC which is zwitterionic, DPPG is
anionic. Further discussion of the collapse of each system is given below.
Results of Brewster Angle Microscopy (BAM) of octadecanol monolayers have shown
that under the usual spreading conditions, irregularly shaped islands of condensed phase
are distributed in a continuous gaseous phase [65]. Upon compression, these islands even-
tually touch each other and start to coalescence and rearrange. With increasing surface
pressure, the condensed phase becomes a continous phase with irregular areas of the
residual gaseous phase [65]. However, results presented here suggest that the interaction
Chapter 2. ADSA-CSD as a Micro Film Balance 41
between these islands starts earlier than the maximum surface pressure achieved as indi-
cated by point A in figure 2.5(b). After a small transition period, the monolayer reaches
the “ultimate” collapse pressure. At this area, the islands of the condensed phase fill the
interface completely. Compressing the monolayer further does not yield any increase in
the surface pressure, but allows for more rearrangements at the interface. It is observed
that the surface pressure measured here is kept constant by further compression and does
not fall to lower values as suggested by other studies [65].
DPPC was chosen here for two reasons; it is the main component of lung surfactant
and it can achieve very low surface tension (very high surface pressure) values upon
compression. This is shown in the results of the collapse pressure values presented here
which are in good agreement with values in the literature [14, 20]. Upon compression,
DPPC monolayer forms flexible domains in the coexistence region of low density and
condensed phases. The ability of DPPC to adjust to an area reduction is attributed to
the change in the orientation of the hydrophobic tails of the DPPC molecule [102–104].
When the average area per molecule for the polar head decreases, the tilt angle of the
tail also decreases and the orientation becomes more vertical. However, the large head
group prevents the complete erection of the chains. These processes are responsible for
some of the irregularities in the vicinity of the collapse area as shown in figure 2.6(b).
Comparing figures 2.5(a) and 2.6(a), the results confirm that the range of area per
molecule over which the compression occurs is broader for DPPC than for octadecanol.
Because of the higher flexibility of the domains, a DPPC monolayer may be compressed
over a larger area per molecule range. This was attributed to the ordering of the DPPC
tails towards vertical orientation which requires more energy, extending the range of area
per molecule over which the compression occurs [66].
DPPG is a lipid found in lung surfactant where it represents 5-10% of the total
phospholipid content. Unlike DPPC (the main lipid in lung surfactant), DPPG is ionic
and forms charged monolayers on aqueous electrolyte subphases. The collapse pressure
Chapter 2. ADSA-CSD as a Micro Film Balance 42
depends on the nature of the subphase used. The values reported here are in good
agreement with previous results, where DPPG was deposited on an air/water interface
[21]. Other studies show a different collapse pressure value when DPPG was deposited
onto subphases composed of solutions of sodium bicarbonate with varying NaCl salt
concentrations [22].
Apart from the collapse pressure value, the DPPG isotherm has a similar shape to that
of DPPC. However, figure 2.8(b) shows no signs of collapse prior to reaching the collapse
pressure value. The change of the isotherm slope occurs at the collapse pressure as shown
in figure 2.8(b). Further compression causes squeeze out from the DPPG monolayer. This
squeezed out material organizes as aggregates in the subphase decreasing the surface
pressure [22]. More compression will cause the surface pressure to increase again until
another portion of the material is squeezed out. This would explain the irregularities in
the isotherm in the vicinity of the collapse pressure as shown in figure 2.8(b). Studies
using Brewster Angle Microscopy (BAM) and Atomic Force Microscopy (AFM) reveal
that the squeezed out aggregates diffuse away from the monolayer at low ionic strength
and remain in the subphase; at high ionic strength, they remain attached to the monolayer
and reincorporate again into the monolayer when the film is expanded [22].
In conclusion, the use of a constrained sessile drop constellation (ADSA–CSD) as a
film balance is illustrated. It was shown that it has certain advantages over conventional
methods. It is a very simple setup and only small quantities of liquid are required. The
ability to measure very low surface tension values, easier deposition procedure and leak-
proof design makes the constrained sessile drop constellation a better choice than the
pendant drop for studies of both soluble or insoluble monolayers.
The fact that ADSA–CSD measures the change of both surface area and surface
tension with time allows a detailed study of the collapse region. Changes in surface area
in the vicinity of the ultimate collapse pressure allow an estimate of the size of regions
and the number of molecules that are being pushed into the subphase. Such information
Chapter 2. ADSA-CSD as a Micro Film Balance 43
20 30 40 50 60 700
10
20
30
40
50
60
70
Area per Molecule, Am (Å2/molecule)
Sur
face
Pre
ssur
e, π
(m
J/m
2 )
Run aRun bRun c
Figure 2.9: The change of surface pressure with area per molecule of a constrained sessiledrop of a DPPC monolayer for different runs.
is of specific interest for lung surfactant systems and cannot be obtained by a Langmuir
film balance.
The results of compression isotherms for different systems show distinctive differences
between the onset of collapse and the ultimate collapse pressure (ultimate strength) of
these films. ADSA–CSD allows detailed study of this collapse region.
2.2.4 Results for Mixed Monolayers
Figure 2.9 shows the change of surface pressure with area per molecule for a DPPC
monolayer at a compression speed of 0.19 cm2/min (19 A2/molecule·min). The figure
shows three different runs, illustrating reproducibility. The surface pressure starts at
Chapter 2. ADSA-CSD as a Micro Film Balance 44
25 30 3545
50
55
60
65
70
75DPPC:DPPG
100:0(a)
Sur
face
Pre
ssur
e, π
(m
J/m
2 )
25 30 35
DPPC:DPPG90:10
(b)
25 30 35
DPPC:DPPG80:20
(c)
25 30 35
DPPC:DPPG70:30
(d)
Area per Molecule, Am (Å2/molecule)
25 30 35
DPPC:DPPG60:40
(e)
25 30 35
DPPC:DPPG50:50
(f)
25 30 35
DPPC:DPPG0:100
(g)
Figure 2.10: The change of surface pressure with area per molecule of a constrainedsessile drop of mixed DPPC/DPPG monolayers for different mixture ratios: (a) 100:0;(b) 90:10; (c) 80:20; (d) 70:30; (e) 60:40; (f) 50:50; (g) 0:100.
approximately 5 mJ/m2 and reaches about 70 mJ/m2 as the film is compressed. A
plateau is observed indicating the collapse of the monolayer at this ultimate collapse
surface pressure (very low surface tension, approximately 2 mJ/m2). Further inspection
shows that there is a slight change (not readily apparent in the figure) in the isotherm
slope starting at pressures round 48 mJ/m2, well below the ultimate collapse pressure.
This point was denoted in the previous section as the onset of collapse.
The change of surface pressure with area per molecule of a constrained sessile drop of
mixed DPPC/DPPG monolayers for different mixture ratios are compared in Figure 2.10
in the vicinity of the collapse pressure. The ratios studied here ranges from 100 to 50%
DPPC (with 10% increments) in the mixture in addition to pure DPPG. It can be seen
from the figure that for ratios 90:10 and 80:20, the DPPC/DPPG mixture can achieve a
very high ultimate collapse pressure, similar to pure DPPC. When the ratio of DPPG in
the mixture is increased, the ultimate collapse pressure is decreased slightly for the ratios
70:30 and 60:40. Upon further increase of DPPG (50% or more), the ultimate collapse
pressure is decreased to approximately 60 mJ/m2, similar to pure DPPG.
Every experiment for a specific mixture ratio was repeated at least 4 times. The
reproducibility of the results shown in Figure 2.9 for the case of pure DPPC is not
Chapter 2. ADSA-CSD as a Micro Film Balance 45
Table 2.3: Ultimate collapse pressure (mJ/m2) and film elasticity (mJ/m2) at roomtemperature of 24oC from different runs of different DPPC:DPPG mixture ratios andthe average value along with the standard deviation.
DPPC:DPPG Run 1 Run 2 Run 3 Run 4 AverageStandardDeviation
Ultimate Collapse Pressure
100:0 70.7±0.03 70.6±0.07 70.6±0.14 70.7±0.04 70.6 0.0290:10 70.5±0.11 70.4±0.25 70.5±0.24 70.2±0.13 70.4 0.0580:20 70.9±0.24 70.5±0.29 70.8±0.26 70.7±0.21 70.7 0.0770:30 68.7±0.23 69.0±0.41 68.1±0.44 67.4±0.82 68.3 0.2860:40 68.6±0.80 69.9±0.61 70.0±0.36 67.2±0.58 68.9 0.5050:50 60.8±1.39 60.4±1.10 60.2±0.99 61.0±1.90 60.6 0.150:100 60.8±1.12 59.6±1.68 60.0±1.42 59.2±1.82 59.9 0.25
Film elasticity at π = 50 mJ/m2
100:0 227.5±1.4 236.4±2.0 225.2±0.7 221.5±1.7 227.6 2.490:10 243.7±1.0 209.4±5.3 257.5±4.4 255.7±4.0 241.6 8.480:20 212.4±1.5 214.1±1.6 203.5±2.7 207.8±1.0 209.4 1.870:30 209.3±1.7 224.8±2.8 198.9±1.1 263.0±2.1 224.0 10.660:40 191.9±1.7 214.8±1.6 208.8±1.0 224.7±2.3 210.1 5.250:50 199.2±1.4 250.4±3.1 168.5±1.3 227.3±1.2 211.3 13.40:100 774.6±31.8 755.3±38.2 717.8±18.7 736.4±59.9 746.0 9.2
Chapter 2. ADSA-CSD as a Micro Film Balance 46
always guaranteed. A summary of calculated ultimate collapse pressure for different
mixture ratios is shown in Table 2.3 along with an average value and standard deviation.
To illustrate the reproducibility of the results for different mixture ratios, Table 2.3 shows
the ultimate collapse pressure values obtained for all runs of every mixture ratio. The
table shows that adding a limited amount of DPPG (up to 20%) to DPPC monolayer does
not alter the ultimate collapse pressure. Increasing the DPPG content in the mixture
(30% or 40%) shows a transition region, where the ultimate collapse pressure is decreased
to values ranging between 66 and 70 mJ/m2. The variations of values obtained for the
ultimate collapse pressure in this regions are not small, indicating a low reproducibility.
However, further increase of DPPG content in the mixture (50% and beyond) causes a
sharp decrease in the ultimate collapse pressure to values close to the pure DPPG. It is
noted that the reproducibility in this region is much better than that of the transition
region.
Further insight into the differences between the mixture ratios can be gained by
calculating the surface dilatational elasticity from the measured isotherms. For reason of
comparison, elasticity values are calculated at the same surface pressure, i.e. 50 mJ/m2,
for all the measured isotherms. This value is chosen for two reasons; first, it is below
the ultimate collapse pressure for the isotherms in this study. Second, this value is close
to the equilibrium surface pressure for lung surfactant systems, thus eliminating any
effects of relaxation or adsorption that could alter the calculated value of film elasticity.
A summary of calculated film elasticity values for different mixture ratios is shown in
Table 2.3 along with an average value and standard deviation. It can be seen that for
all ratios studied here, the elasticity values are around 220 mJ/m2, very close to pure
DPPC. Increasing the DPPG content in the mixture does not increase the film elasticity.
However, the pure DPPG film has a much higher elasticity of approximately 750 mJ/m2,
i.e. a better film property.
Thus far, experiments were performed by only compressing the film once and record-
Chapter 2. ADSA-CSD as a Micro Film Balance 47
15 20 25 30 35 40 45 5045
50
55
60
65
70
75
Area per Molecule, Am (Å2/molecule)
Sur
face
Pre
ssur
e, π
(m
J/m
2 )
DPPC:DPPG100:0
1st comp2nd comp
3rd comp4th comp
(a)
15 20 25 30 35 40 45 5045
50
55
60
65
70
75
Area per Molecule, Am (Å2/molecule)
Sur
face
Pre
ssur
e, π
(m
J/m
2 )
DPPC:DPPG90:10
1st comp2nd comp
3rd comp4th comp
(b)
15 20 25 30 35 40 45 5045
50
55
60
65
70
75
Area per Molecule, Am (Å2/molecule)
Sur
face
Pre
ssur
e, π
(m
J/m
2 )
DPPC:DPPG80:20
1st comp2nd comp
3rd comp4th comp
(c)
Figure 2.11: The change of surface pressure with area per molecule on successive com-pressions and expansions of a constrained sessile drop of mixed DPPC/DPPG monolayersfor different mixture ratios: (a) 100:0; (b) 90:10; (c) 80:20.
Chapter 2. ADSA-CSD as a Micro Film Balance 48
ing the change of surface pressure. A different set of experiments were performed, where
after the full compression of the monolayer (different mixture ratios), the monolayer was
then expanded and compressed again. This compression/expansion cycle was then re-
peated, i.e. 3 compression/expansion cycles for each mixture ratio. Figure 2.11 shows
the different compression/expansion cycles for three different mixture ratios, i.e. pure
DPPC, 90:10 and 80:20 DPPC/DPPG. These ratios were selected because all of them
reach the same high ultimate collapse pressure shown in Table 2.3. The second com-
pression of pure DPPC was slightly shifted to the left compared to the first compression
as shown in Figure 2.11(a). This indicated that recompression of pure DPPC causes a
small loss of molecules between the first and second compression. For the case of 10%
DPPG (Figure 2.11(b)), after the first compression, the shape of the compression/ex-
pansion isotherms is very similar to the pure DPPC isotherm. This suggests that during
the first compression, most of the DPPG was ejected from the interface leaving a pure
DPPC monolayer for the following compression/expansion cycles. However, increasing
the DPPG content to 20% (Figure 2.11(c)) shows a change in the shape of the cycle.
After the first compression, the shape of the subsequent compressions did not change.
Every compression in this case shows a slight shift to the left. It is noted that the shape
here is different from that of pure DPPC for all 4 compressions.
2.2.5 Discussion for Mixed Monolayers
2.2.5.1 Ultimate Collapse Pressure of Mixed Monolayers
As the main component of lung surfactant, DPPC is known to achieve very low surface
tension (very high surface pressure) values upon compression. This is shown in the results
of the collapse pressure values for pure DPPC presented here in Figure 2.9 and Table 2.3.
These results are in good agreement with values in the literature [14, 20, 88]. Unlike
DPPC, the collapse pressure of DPPG depends on the nature of the subphase used. The
Chapter 2. ADSA-CSD as a Micro Film Balance 49
values for pure DPPG reported here in Table 2.3 are in good agreement with previous
results, where DPPG was deposited on an air/water interface [21, 70, 88].
Results presented in Table 2.3 show that adding a limited amount of DPPG (up to
20%) to DPPC does not alter the ability of the monolayer to reach a very high ultimate
collapse pressure. With further increase of the amount of DPPG in the mixture, the
ultimate collapse pressure goes through a transition region followed by a decrease in the
ultimate collapse pressure to values close to that of the pure DPPG (for ratios of 50% and
beyond). These results suggest that an increased amount of DPPG (beyond 30%) would
adversely affect the ability of the mixture to reach high surface pressure (low surface
tension) values.
2.2.5.2 Film Elasticity of Mixed Monolayers
In addition to the high collapse pressure, a good monolayer would also have a high
elasticity value. The elasticity can be calculated from the slope of a π − A isotherm as
shown in Table 2.3. The monolayer elasticity is a measure of the film resistance (change
in the film surface pressure) to a change in the area per molecule. The results show that
the monolayer elasticity calculated at 50 mJ/m2 surface pressure is around 220 mJ/m2
for pure DPPC, and around 750 mJ/m2 for pure DPPG. The value for pure DPPC is in
good agreement with previous results [105], and the value for pure DPPG is in agreement
with values calculated from previously measured isotherms [88]. Some of the calculated
elasticity values measured from previously measured isotherms [70, 106] are somewhat
lower (between 350 and 600 mJ/m2) than the value measured here, but still higher
than that of pure DPPC. However, the resolution in the current data in the vicinity of
50 mJ/m2 surface pressure is much better than in other techniques. As elasticity is a
function of compression rate, it should be kept in mind that the compression rate for all
experiments was fixed at 0.19 cm2/min, as indicated earlier. The relatively large error for
DPPG elasticity in Table 2.3 is due to the fact that the experimental isotherms are very
Chapter 2. ADSA-CSD as a Micro Film Balance 50
steep, i.e. approaching a slope of infinity. All DPPC/DPPG mixture ratios studied here
(90 to 50% DPPC) have elasticity values very close to pure DPPC. It is clear that pure
DPPG films have better elasticity than DPPC films. However, increasing the DPPG
content in the mixture does not improve the film elasticity.
2.2.5.3 Repeated Compression Cycles
In the first set of experiments, after spreading, the monolayer was continuously com-
pressed until collapse, at which point the experiment ended. In the second set of exper-
iments, the monolayer was spread, then compressed until collapse occurred. The mono-
layer was then expanded and compressed again. Three compression/expansion cycles
were performed for every experiment. The difference between the first and the following
compressions usually gives an indication of the number of molecules lost between com-
pressions. Three DPPC/DPPG mixture ratios are compared in Figure 2.11. For the case
of pure DPPC (Figure 2.11(a)), the results suggest that some molecules were not able to
reincorporate at the interface after the first compression. The successive compressions
did not show significant change. For the case of 10% DPPG in the mixture, the shape
of the compression/expansion cycles after the first compression is essentially the same
as that of pure DPPC as shown in Figure 2.11(b). This suggests that the 10% DPPG
content was squeezed out from the interface during the first compression and was not able
to reincorporate back. The interface after the first compression is probably composed of
DPPC only. On the other hand, increasing the DPPG content to 20% (Figure 2.11(c))
shows a change in the shape of the compression/expansion cycles. Every compression in
this case shows a slight shift to the left, indicating a continuous ejection of molecules. It
is noted that the shape here is different from that of pure DPPC for all compressions.
This suggest that, at this mixture ratio, DPPG is still present at the interface at the
high ultimate collapse pressure.
Chapter 2. ADSA-CSD as a Micro Film Balance 51
2.2.5.4 Implications for Lung Surfactant Development
Conceptually, it is possible to define qualitatively a set of surface properties that are nec-
essary for the respiratory function of lung surfactants. These surface properties include
the ability to adsorb to the air/water interface, the ability to reach very low surface ten-
sion values during dynamic cycling, the ability to reach such low surface tension values
with minimum amount of area reduction (high elasticity), and the ability to respread
at the interface during expansion [1]. No single component of the complex compound
of lung surfactant has all these necessary surface properties. However, different compo-
nents interact to achieve the required system behaviour. As shown in the literature, and
reported here as well, DPPC alone (zwitterionic phospholipid) can reach extremely high
surface pressures (low surface tensions below 2 mJ/m2). However, DPPC adsorbs very
slowly to the air/water interface [1, 36], has a relatively low elasticity compared to DPPG
as shown in Table 2.3, and respreads poorly in compression/expansion cycles as shown in
Figure 2.11(a). Fluid phospholipids, neutral lipids and surfactant proteins (SP-A, SP-B,
and SP-C) cannot reach high surface pressures but have a significant effect on improving
adsorption and respreading [61, 62, 107].
The presence of anionic phospholipids, e.g. DPPG, in lung surfactant is well estab-
lished; however, their quantitative contribution to surface activity is less clear. Results
presented here suggest that an increased content of DPPG will adversely affect the sur-
face tension lowering ability of DPPC. However, it is also shown that DPPG has better
film elasticity and respreading properties in compression/expansion cycles. It is widely
believed that their association with one or more surfactant proteins may make a con-
tribution to the surface activity [96–100, 108–115]. As mentioned before a DPPC/PG
mixture has been repeatedly used to mimic natural lung surfactant [94, 95], and has also
been used as the main component of many clinical synthetic exogenous lung surfactants,
e.g. ALEC [79], Surfaxin [82], and Venticute [85].
In conclusion, results here suggest that DPPG improves the respreading properties
Chapter 2. ADSA-CSD as a Micro Film Balance 52
of the mixture and may improve the film elasticity. The only drawback of increasing the
content of DPPG is the adverse effect on the ultimate collapse pressure. However, the
low collapse pressure of DPPG is limited only to pure water as subphase. Other studies
showed that DPPG can in fact achieve very high collapse pressure values (similar to pure
DPPC) when deposited onto subphases composed of solutions of sodium bicarbonate
mixed with NaCl [22]. This leads to the expectation that DPPG may play an important
role in the future of lung surfactant development.
2.3 Double Injection Capability
Lung surfactant inhibition or inactivation is a phenomenon caused by processes that
degrade the surface activity of the surfactant [1]. Direct interaction with endogenous in-
hibitors is a main reason for such phenomena. Some inhibitors can impair the adsorption
and the surface tension lowering abilities of the surfactant. Examples of such inhibitors
include plasma and blood proteins (such as albumin, fibrinogen, and hemoglobin) [116–
122], unsaturated cell membrane phospholipids [121, 123], lysophospholipids [34, 124],
fluid free fatty acids (such as oleic acid) [35, 125], meconium [126, 127], and cholesterol
when present at high concentrations [128, 129]. Other inhibitors, usually released during
lung inflammation or injury, can chemically react with and degrade the functional compo-
nents of the surfactant. Examples include lytic enzymes (proteases and phospholipases)
[33, 130] and reactive oxidant species [131, 132].
The surface activity of lung surfactants can be assessed with various methods as
shown in Chapter 1. Most of the methods suffer from one or more of the following
limitations: inability to perform fast cycling, film leakage, uncontrolled environmental
conditions, inaccurate calculation of low surface tension values, inoperability at the high
surfactant concentrations. These limitations must be removed in the in vitro study of
lung surfactants that should mimic the conditions of normal breathing. ADSA–CSD
Chapter 2. ADSA-CSD as a Micro Film Balance 53
is believed to be free of all restrictions and limitations of other conventional methods
mentioned before. However, it is not possible yet to use ADSA–CSD as a penetration
film balance that would allow the access to the interface from the liquid side.
The purpose of this section is to present a modified design for ADSA–CSD featuring
a secondary injection system facilitating access to the interface from the liquid side.
The study of the effect of a specific inhibitor on a preformed lung surfactant film is a
good illustration of this capability. After forming a sessile drop of a basic surfactant
preparation, the bulk phase can be exchanged and different inhibitors such as serum,
albumin, fibrinogen, and cholesterol can be introduced. Results are compared, below,
with the case where the inhibitor and the surfactant preparation are premixed and co-
adsorb. This modified design can be used to evaluate lung surfactant inhibition, resistance
and reversal under completely controlled physiological conditions. The modified design
can also be used for bulk liquid sampling and on-site mixing.
2.3.1 Experimental Details
2.3.1.1 Materials and Methodology
The lung surfactant used, Bovine Lipid Extract Surfactant (BLES), was provided by
BLES Biochemicals Inc. (London, Ontario, Canada) at a concentration of 27 mg/ml
and was used without further purification. BLES was divided into 1 ml glass vials in Ar
atmosphere and stored at -20 C. On the day of the experiment, one vial was maintained
at 37.5 C water bath for one hour before diluting it to 2.0 mg/ml by a salt solution of
0.6% NaCl and 1.5 mM CaCl2. The pH value of the diluted BLES preparation ranged
from 5 to 6.
A potential lung surfactant additive, protasan (UP CL 213, chitosan chloride, deacety-
lation 75-90%), was purchased from NovaMatrix (Norway). More details about protasan
can be found elsewhere [133, 134]. Protasan was dissolved in the NaCl/CaCl2 salt solu-
Chapter 2. ADSA-CSD as a Micro Film Balance 54
tion on the day of the experiment and then diluted to the final concentration (ranging
from 0.05 to 0.50 mg/ml according to the required preparation) and mixed with 2.0
mg/ml BLES suspension. The pH of these BLES-protasan mixtures ranged from 5 to
5.5.
Here, a preparation of BLES and protasan is chosen as the basic preparation, i.e.
a basis for comparison. Through accumulated experience in the Laboratory of Applied
Surface Thermodynamics, it was observed that the minimum surface tension of BLES
alone, without any additives, varies significantly from batch to batch [91], at least at
low BLES concentrations. The minimum surface tension, i.e. the lowest surface tension
obtainable upon 20% area reduction, varied from 5.5 mJ/m2 to 20.4 mJ/m2 in 100%
relative humidity [91]. Parenthetically, it is noted that typically these minimum surface
tensions are much lower at non-physiological lower humidities. Therefore, in order to
study whether a certain inhibitor has the capability of raising the surface tension above
the physiologically interesting range of 1 to 5 mJ/m2, we need a standard enhancing
additive to guarantee an initial low surface tension, prior to any challenge of the surfactant
film by an inhibitor. The choice of additive here is protasan (chitosan chloride). A similar
procedure was described previously for the case of chitosan as a surfactant additive [91].
Serum (bovine, cat. B8655), albumin (from bovine serum, fatty acid free, globulin
free, ≥99%, cat. A0281), fibrinogen (from bovine plasma, type I-S, 65-85% protein, cat.
F8630), and cholesterol (≥99%, cat. C8667) were purchased from Sigma-Aldrich Co.
Serum was diluted in the NaCl/CaCl2 salt solution to a final concentration, ranging from
0.01 to 0.30 ml/ml according to the required preparation. Albumin was dissolved in the
NaCl/CaCl2 salt solution to a concentration of 40.0 mg/ml. Fibrinogen was dissolved
in a salt solution of 0.85% NaCl and 1.5 mM CaCl2 and adjusted to a concentration
of 3.0 mg/ml. Cholesterol was dissolved in 10 mg/ml ethanol and diluted to a final
concentration of 0.284 mg/ml. The pH value of the diluted serum preparations was near
7, while the pH value of the other inhibitor preparations ranged from 5 to 6.
Chapter 2. ADSA-CSD as a Micro Film Balance 55
Figure 2.12: Schematic diagram of double injection ADSA–CSD experimental setup andan illustration of the exchange process.
2.3.1.2 Experimental Procedure
The design for a modified ADSA–CSD including a secondary injection was developed
and manufactured. The details of the design and operation of the original ADSA–CSD
were described elsewhere [25, 46, 48]. A schematic diagram of the new design is shown
in figure 2.12 and a detailed engineering drawing is shown in figure 2.13. The modified
design features a capillary needle fitted concentrically in the opening of the main pedestal
(holder) to facilitate secondary liquid injection. The capillary needle is connected to a
separate syringe with a dedicated motor equipped with a controller.
The details of the modified ADSA–CSD are as follows: the primary fluid (lung sur-
factant preparation) is delivered from the first syringe to form a sessile drop on a circular
horizontal stainless steel pedestal. The secondary fluid (inhibitor) is injected from the
second syringe into the preformed sessile drop through a capillary needle. The base of
the pedestal is designed to accommodate a reservoir, which is used to ensure a constant
Chapter 2. ADSA-CSD as a Micro Film Balance 56
Figure 2.13: Detailed engineering drawing of double injection ADSA–CSD.
Chapter 2. ADSA-CSD as a Micro Film Balance 57
temperature.
During the experiment, the setup is enclosed in an environmental control chamber
that facilitates the control of gas composition and temperature. The humidity inside the
chamber was kept constant at 100% relative humidity at 37 C. Two stepping motors
(controller 18705/6, Oriel Instruments, Stratford, CT) were used, one for each fluid, to
facilitate the fluid volume injection/extraction and the compression/expansion during the
dynamic cycling part of the experiment. A CCD camera (model 4815-5000, Cohu Corp.,
Poway, CA) mounted on a microscope (type 400076, Wild Heerburgg, Switzerland) was
used to acquire images throughout the experiment at a rate of 20 images per second.
The images were digitized using a digital video frame grabber (Snapper-8, Active Silicon
Ltd., Iver, UK) and stored in a workstation (SunBlade 1500, Sun Microsystems, Santa
Clara, CA) for further analysis by ADSA.
The experimental procedure for a double injection experiment is as follows: First,
the capillary needle is filled with the secondary fluid (inhibitor). Then, the primary fluid
(lung surfactant) is used to form a sessile drop on the pedestal. The surface tension
is tracked until the film reaches the equilibrium surface tension (within 180 seconds).
Thereafter, dynamic cycling is performed by successive compression/expansion of the
drop. This is normally done through 20 cycles with a periodicity of 3 seconds per cycle
and a compression ratio (percentage of surface area reduction) of 20% to simulate normal
human breathing conditions [1, 25]. Unless stated otherwise, compression conditions in
all results in this work are the same (20% surface area reduction at 3 s/cycle periodicity)
in humid air (100% R.H.).
At this point, the volume replacement procedure begins. The secondary fluid is in-
jected continuously through the inner capillary needle. During the injection, the primary
fluid is simultaneously and continuously withdrawn using the other syringe and motor
controller. The volume exchanged is 5 times the drop volume to ensure a complete bulk
replacement of the primary fluid by the secondary fluid [60]. This minimum exchange
Chapter 2. ADSA-CSD as a Micro Film Balance 58
volume depends on flow dynamics as shown earlier [135]. To corroborate the recommen-
dations of Cabrerizo-Vılchez et al. [60], the volume exchange process was tested using
two miscible liquids with the same density but with different interfacial tensions, namely
dimethylsulfoxide (DMSO) and 1-chlorobenzene (CB). It was found that complete bulk
replacement can be ensured after only 2 volumes exchange for the specific dynamic flow
condition used here.
The volume replacement is performed slowly (120 seconds) to avoid film disturbance.
After completing this step, the drop is left for 180 seconds to reach equilibrium again.
Finally, the dynamic cycling part is repeated (20 cycles, 3 seconds per cycle, 20% com-
pression). The results from the two dynamic cycling parts (before and after injection)
are further analyzed and compared to evaluate the effect of injecting the secondary fluid
(usually an inhibitor) under a preformed film of the primary fluid (the basic preparation).
One of the main parameters that reflects the effect of injection is the minimum surface
tension, γmin, obtained at the end of compression in each cycle during the two dynamic
cycling parts. The reported values here are those averaged over the 20 cycles during a
specific dynamic part along with the standard deviation. The measured value of minimum
surface tension depends on the compression conditions. Another important parameter is
the dilatational elasticity as defined in equation 2.4. The elasticity can be calculated in
a similar fashion to equation 2.4 but written in terms of surface tension [25, 45, 47, 62,
107, 136],
ε =
∣∣∣∣ dγ
d lnA
∣∣∣∣ (2.5)
where γ is the surface tension of the film at any time, and A is the drop surface area
at any time. The change of surface tension with the relative surface area during the
compression part of every cycle is fitted to a fourth order polynomial and the slope is
evaluated at the midpoint of the compression [91]. The calculated slope and the value
of the relative surface area are used to calculate the dilatational elasticity according to
equation 2.5.
Chapter 2. ADSA-CSD as a Micro Film Balance 59
10
20
30
40
50
Su
rfa
ce
Te
nsio
n, γ
(m
J/m
2) (a) BLES and Protasan only
180 190 200
10
20
30
40
50
Time, t (s)
Su
rfa
ce
Te
nsio
n, γ
(m
J/m
2) (b) After Bulk Replacement
0.8 0.9 1
Relative Surface Area, Ar
Figure 2.14: Change in surface tension with time and with relative surface area duringdynamic cycling: (a) 2.0 mg/ml BLES and 0.2 mg/ml protasan (basic preparation);(b) After replacing the bulk phase with salt solution. All compressions/expansions areperformed at 3 s/cycle periodicity in humid air (100% R.H.).
2.3.2 Results
Figure 2.14 shows a solvent exchange experiment. The change of surface tension, γ,
with time during the first few cycles in the dynamic cycling of a basic preparation of 2.0
mg/ml BLES and 0.2 mg/ml protasan is shown in figure 2.14(a). The figure also shows
the change of surface tension with relative surface area, Ar, for one of the cycles (the
fifth cycle). Figure 2.14(b) shows the surface tension response for the same compression
conditions but after replacing the bulk with salt solution only (bulk wash-out). The
minimum surface tension after bulk replacement did not change. Both cases have the
same γ−Ar response with similar elasticity. The consistency of the surface activity after
bulk replacement suggests effectively irreversible adsorption. This is consistent with the
Chapter 2. ADSA-CSD as a Micro Film Balance 60
spreading of lung surfactant vesicles described in the literature [137–139].
Figure 2.15(a) shows the change of surface tension, γ, with time during the first few
cycles in the dynamic cycling of a basic preparation of 2.0 mg/ml BLES and 0.2 mg/ml
protasan. The rest of the 20 cycles did not show a significant difference from the first few.
The figure also shows the change of surface tension with relative surface area, Ar, for one
of the cycles (the fifth cycle). The open circles show the change during the compression
part of the cycle, while the solid circles give the surface tension during the expansion
part. The minimum surface tension is 2.9 ±0.3 mJ/m2. Figure 2.15(b) shows the surface
tension response for the same compression conditions but after injecting 0.15 ml/ml
serum into the bulk of the basic preparation. The minimum surface tension after the
injection is 3.5 ±0.6 mJ/m2. Comparing figures 2.15(a) and 2.15(b) show that injecting
serum under a preformed film of BLES and protasan has little effect on the minimum
surface tension. However, the γ − Ar isotherms show a significant change in the cycle
hysteresis and also in the slope of the compression part of the cycle.
Figure 2.15(c) shows the change of surface tension with time and with relative surface
area for the mixture of the inhibitor (0.15 ml/ml serum) and the basic preparation (2.0
mg/ml BLES and 0.2 mg/ml protasan) in a separate single injection experiment. In
this experiment, the mixture is used to form a sessile drop on the pedestal. Then, the
surface tension is tracked until the film reaches the equilibrium surface tension (within 180
seconds). After that, dynamic cycling is performed under standard conditions [25]. The
minimum surface tension for the mixture is 15.6 ±0.3 mJ/m2, which differs significantly
from the value achieved after the injection in the double injection experiment. The γ−Ar
isotherms show that the cycle hysteresis and the slope of the compression is also different
from figures 2.15(a) and 2.15(b).
The choice of the basic preparation in that experiment needs further comment. To
determine a suitable concentration of protasan to be added to BLES, different concen-
trations were tested and the minimum surface tension obtained for two different BLES
Chapter 2. ADSA-CSD as a Micro Film Balance 61
10
20
30
40
50
Surf
ace T
ensio
n, γ
(m
J/m
2) (a) BLES and Protasan only
10
20
30
40
50
Surf
ace T
ensio
n, γ
(m
J/m
2) (b) After Bulk Replacement
180 190 200
10
20
30
40
50
Time, t (s)
Surf
ace T
ensio
n, γ
(m
J/m
2) (c) Mixed with Inhibitor
0.8 0.9 1
Relative Surface Area, Ar
Figure 2.15: Change in surface tension with time and with relative surface area duringdynamic cycling: (a) 2.0 mg/ml BLES and 0.2 mg/ml protasan (basic preparation); (b)After injecting 0.15 ml/ml serum into the bulk phase; (c) 2.0 mg/ml BLES and 0.2mg/ml protasan and 0.15 ml/ml serum mixed in a separate single injection experiment.All compressions/expansions are performed at 3 s/cycle periodicity in humid air (100%R.H.).
Chapter 2. ADSA-CSD as a Micro Film Balance 62
0 0.1 0.2 0.3 0.4 0.50
5
10
15
Min
imu
m S
urf
ace
Te
nsio
n, γ
min (
mJ/m
2)
Protasan Conentration (mg/ml)
BLES 2.0 mg/ml (batch 1)
BLES 2.0 mg/ml (batch 2)
Figure 2.16: Minimum surface tension (end of compression) as a function of protasanconcentration for different batches of BLES (2.0 mg/ml). All compressions (20%) areperformed at 3 s/cycle periodicity in humid air (100% R.H.).
Chapter 2. ADSA-CSD as a Micro Film Balance 63
10
20
30
40
50
Su
rfa
ce
Te
nsio
n, γ
(m
J/m
2) (a) After Bulk Replacement
180 190 200
10
20
30
40
50
Time, t (s)
Su
rfa
ce
Te
nsio
n, γ
(m
J/m
2) (b) Mixed with Inhibitor
0.8 0.9 1
Relative Surface Area, Ar
Figure 2.17: Change in surface tension with time and with relative surface area duringdynamic cycling: (a) After injecting 0.30 ml/ml serum into the bulk phase of 2.0 mg/mlBLES and 0.2 mg/ml protasan; (b) 2.0 mg/ml BLES and 0.2 mg/ml protasan and 0.30ml/ml serum mixed in a separate single injection experiment. All compressions/expan-sions are performed at 3 s/cycle periodicity in humid air (100% R.H.).
batches (2.0 mg/ml) was plotted in figure 2.16. These values were obtained during the
dynamic cycling of single injection experiments. It is clear that the minimum surface ten-
sion of BLES alone, without any protasan, differs in these two batches from 5.5 mJ/m2
to 11.0 mJ/m2. It is also clear that adding a small concentration of protasan reduces the
minimum surface tension of both BLES batches; this is similar to the effect of chitosan
on different BLES batches [91]. In view of this figure, a preparation of 2.0 mg/ml BLES
and 0.2 mg/ml protasan is used as the basic preparation.
Figure 2.17(a) shows the change of surface tension, γ, with time during the first few
cycles in the dynamic cycling after injecting a higher concentration of serum, i.e. 0.30
ml/ml, into the bulk of the basic preparation (2.0 mg/ml BLES and 0.2 mg/ml protasan).
Chapter 2. ADSA-CSD as a Micro Film Balance 64
10
20
30
40
50
Su
rfa
ce
Te
nsio
n, γ
(m
J/m
2) (a) After Bulk Replacement
180 190 200
10
20
30
40
50
Time, t (s)
Su
rfa
ce
Te
nsio
n, γ
(m
J/m
2) (b) Mixed with Inhibitor
0.8 0.9 1
Relative Surface Area, Ar
Figure 2.18: Change in surface tension with time and with relative surface area duringdynamic cycling: (a) After injecting 40 mg/ml albumin into the bulk phase of 2.0 mg/mlBLES and 0.2 mg/ml protasan; (b) 2.0 mg/ml BLES and 0.2 mg/ml protasan and 40mg/ml albumin mixed in a separate single injection experiment. All compressions/ex-pansions are performed at 3 s/cycle periodicity in humid air (100% R.H.).
The figure also shows the change of surface tension with relative surface area, Ar, for
one of the cycles. Figure 2.17(b) shows the change of surface tension with time and with
relative surface area for the mixture of the inhibitor (0.30 ml/ml serum) and the basic
preparation (2.0 mg/ml BLES and 0.2 mg/ml protasan) in a separate single injection
experiment. Figures 2.18, 2.19, and 2.20 show similar results for different inhibitors:
40.0 mg/ml albumin, 3.0 mg/ml fibrinogen, and 0.28 mg/ml cholesterol, respectively.
Results for injecting an inhibitor under a preformed film of the basic preparation and
mixing the inhibitor with the preparation (letting them adsorb together to form a film)
show a similar pattern for a wide range of inhibitors.
Results for albumin (figure 2.18) show that mixing with the basic preparation causes
Chapter 2. ADSA-CSD as a Micro Film Balance 65
10
20
30
40
50
Su
rfa
ce
Te
nsio
n, γ
(m
J/m
2) (a) After Bulk Replacement
180 190 200
10
20
30
40
50
Time, t (s)
Su
rfa
ce
Te
nsio
n, γ
(m
J/m
2) (b) Mixed with Inhibitor
0.8 0.9 1
Relative Surface Area, Ar
Figure 2.19: Change in surface tension with time and with relative surface area duringdynamic cycling: (a) After injecting 3.0 mg/ml fibrinogen into the bulk phase of 2.0mg/ml BLES and 0.2 mg/ml protasan; (b) 2.0 mg/ml BLES and 0.2 mg/ml protasanand 3.0 mg/ml fibrinogen mixed in a separate single injection experiment. All compres-sions/expansions are performed at 3 s/cycle periodicity in humid air (100% R.H.).
an increase in the minimum surface tension and a decrease in film elasticity during
compression. On the other hand, the injection of albumin under a preformed surfactant
film did not change the minimum surface tension. However, injecting albumin caused an
increase in the film elasticity during compression, leading to a minimum surface tension
at a lower compression ratio, i.e. at approximately 12%, compared to the standard 20%.
Further compression caused film collapse (no further decrease in surface tension) leading
to film hysteresis as shown in figure 2.18(a).
Results for fibrinogen (figure 2.19) show a dramatic increase in the minimum surface
tension and a corresponding decrease in film elasticity during compression when fibrino-
gen was mixed with the basic preparation. However, when fibrinogen is injected under
Chapter 2. ADSA-CSD as a Micro Film Balance 66
10
20
30
40
50
Su
rfa
ce
Te
nsio
n, γ
(m
J/m
2) (a) After Bulk Replacement
180 190 200
10
20
30
40
50
Time, t (s)
Su
rfa
ce
Te
nsio
n, γ
(m
J/m
2) (b) Mixed with Inhibitor
0.8 0.9 1
Relative Surface Area, Ar
Figure 2.20: Change in surface tension with time and with relative surface area duringdynamic cycling: (a) After injecting 0.284 mg/ml cholesterol into the bulk phase of 2.0mg/ml BLES and 0.2 mg/ml protasan; (b) 2.0 mg/ml BLES and 0.2 mg/ml protasanand 0.284 mg/ml cholesterol mixed in a separate single injection experiment. All com-pressions/expansions are performed at 3 s/cycle periodicity in humid air (100% R.H.).
a preformed film, the minimum surface tension in the first cycle was very close to that
of the basic preparation. Starting with the second cycle, the minimum surface tension
increased slightly and stayed constant during the rest of the 20 cycles. In this case, it
can be seen that the film elasticity during compression has decreased as well.
Results for cholesterol (figure 2.20) show an increase in the minimum surface tension
and a decrease in film elasticity during compression when cholesterol was mixed with
the basic preparation. On the other hand, when injected under a preformed surfactant
film, cholesterol does not tend to change the minimum surface tension, and in fact tends
to improve the film elasticity during compression. This can be seen from the increased
slope of the γ−Ar isotherm shown in figure 2.20(a) when compared to the corresponding
Chapter 2. ADSA-CSD as a Micro Film Balance 67
0
10
20
30
40
Min
imu
m S
urf
ace
Te
nsio
n, γ
min (
mJ/m
2)
Serum0.15ml/ml
Serum0.30ml/ml
Albumin40mg/ml
Fibrinogen3.0mg/ml
Cholesterol0.284mg/ml
After Bulk Replacement
Mixed with Inhibitor
Figure 2.21: Comparison of minimum surface tension for different inhibitors injected ormixed into a basic preparation of 2.0 mg/ml BLES and 0.2 mg/ml protasan. The dashedline shows the minimum surface tension of the basic preparation. All compressions (20%)are performed at 3 s/cycle periodicity in humid air (100% R.H.).
isotherm of the basic preparation shown in figure 2.15(a).
A comparison of minimum surface tension obtained during dynamic cycling for dif-
ferent inhibitors injected or mixed into the basic preparation of 2.0 mg/ml BLES and 0.2
mg/ml protasan is shown in figure 2.21. For all the inhibitors tested here, the minimum
surface tension of the basic preparation (before injection) did not change significantly
after injecting the inhibitor. However, when the same inhibitors are mixed with the
basic preparation and adsorb together to form a film, the minimum surface tension is
significantly higher as shown in figure 2.21.
The corresponding comparison for the dilatational elasticity, ε, obtained during dy-
namic cycling for different inhibitors injected as well as mixed into the basic preparation
Chapter 2. ADSA-CSD as a Micro Film Balance 68
0
50
100
150
200
Ela
sticity o
f C
om
pre
ssio
n,
εcom
(m
J/m
2)
Serum0.15ml/ml
Serum0.30ml/ml
Albumin40mg/ml
Fibrinogen3.0mg/ml
Cholesterol0.284mg/ml
After Bulk Replacement
Mixed with Inhibitor
Figure 2.22: Comparison of elasticity of compression for different inhibitors injected ormixed into a basic preparation of 2.0 mg/ml BLES and 0.2 mg/ml protasan. The dashedline shows the elasticity of compression of the basic preparation. All compressions (20%)are performed at 3 s/cycle periodicity in humid air (100% R.H.).
Chapter 2. ADSA-CSD as a Micro Film Balance 69
10
20
30
40
50
Su
rfa
ce
Te
nsio
n, γ
(m
J/m
2) (a) Serum only
180 200 220
10
20
30
40
50
Time, t (s)
Su
rfa
ce
Te
nsio
n, γ
(m
J/m
2) (b) After Bulk Replacement
0.8 0.9 1
Relative Surface Area, Ar
Figure 2.23: Change in surface tension with time and with relative surface area duringdynamic cycling: (a) 0.01 ml/ml serum; (b) After injecting 2.0 mg/ml BLES and 0.2mg/ml protasan into the bulk phase. All compressions/expansions are performed at 3s/cycle periodicity in humid air (100% R.H.).
of 2.0 mg/ml BLES and 0.2 mg/ml protasan is shown in figure 2.22. The trends here are
similar to the comparison of the minimum surface tension. For the cases of serum and
albumin, the elasticity of the film of the basic preparation did not change significantly
after injecting the inhibitor. However, for the case of fibrinogen, the injection caused the
elasticity to decrease to reach a value close to that obtained from the mixture. Interest-
ingly, cholesterol apparently increased the elasticity (improved the film properties) after
being injected under the preformed film of the basic preparation.
So far, only the effect of injecting an inhibitor under a preformed film of an active
basic preparation has been studied. Another important aspect that can be studied using
this new technique is inhibition reversal, i.e. the ability to restore the normal surface
activity of an inhibited lung surfactant film. An experiment illustrating this concept is
Chapter 2. ADSA-CSD as a Micro Film Balance 70
shown in figure 2.23. In this experiment, a film was originally formed using 0.01 ml/ml
serum only (as the basic preparation in this case). Changes of surface tension with time
and with relative surface area during the dynamic cycling are shown in figure 2.23(a);
the minimum surface tension obtained in this case is 38.0 ±0.2 mJ/m2. A preparation
of 2.0 mg/ml BLES and 0.2 mg/ml protasan is subsequently injected into the preformed
film of serum, and the change of surface tension with time during the dynamic cycling
(after the injection) is shown in figure 2.23(b). This figure shows 12 cycles out of a total
of 20 cycles performed, the rest of the cycles did not show significant further changes.
The change of surface tension with relative surface area is shown for the twelfth cycle.
The minimum surface tension obtained after 12 cycles is 5.7±0.6 mJ/m2. This effectively
shows that a film of dilute serum can be replaced at least partially by a preparation of
BLES and protasan.
2.3.3 Discussion
2.3.3.1 Methodology Evaluation
In order to manipulate and intervene with a lung surfactant film in vitro, it is desirable
to have access to the film from both the liquid and the air side. In ADSA–CSD, access
from the air side is quite easy, as shown in Section 2.2. A deposition procedure to
spread films on the interface was developed and used. But access from the liquid side is
also possible and the necessary development has been described above. As the present
development does not affect any core parts of ADSA–CSD, the results have the same
reliability as those of previous studies [25, 48, 76, 88, 91]. The vehicle chosen here to
illustrate capabilities of bulk substrate manipulation is the resistance of existing lung
surfactant films against inhibitors and, briefly, to document the possibility of reversing
inhibition, i.e. the removal of inhibitors from the surface film.
A key feature of ADSA–CSD is the ability to use relevant physiological conditions in
Chapter 2. ADSA-CSD as a Micro Film Balance 71
terms of compression rates and environmental control (contamination, humidity, pressure
and temperature). Other conventional methods also have some of these features, but
typically not all. For example, a hypophase exchange system was used in conjunction
with a Pulsating Bubble Surfactometer to study the effect of albumin as an inhibitor on
a preformed lung surfactant film [34]. However, film leakage from the bubble surface into
the capillary tube at low surface tension values is a common complication [36]. Other
limitations are the inaccurate calculation of low surface tension values due to spherical
shape assumption [27], and the difficulty of controlling the environmental conditions.
2.3.3.2 Inhibition Study: Effect of Film Formation
The leakage of plasma and blood proteins into the alveolar spaces is a common cause of
severe lung injury. The results presented here show a distinctive difference between the
inhibition of a specific inhibitor when mixed with and when injected under a preformed
surfactant film. Generally, all the inhibitors studied here (serum, albumin, fibrinogen,
and cholesterol) were not able to penetrate the film or affect the lowering surface tension
ability of the chosen basic preparation film (BLES and protasan), while all of them can
alter the surface activity when mixed with the preparation. Figure 2.21 compares the
minimum surface tension achieved in both cases for every inhibitor. On the other hand,
changes in the elasticity of the film during compression (i.e. the inverse of film com-
pressibility) gives useful insight into the effect of every inhibitor. Some of the inhibitors
studied here have altered the film compressibility without showing any changes in the
minimum surface tension measured during the dynamic cycling. Detailed discussion for
every inhibitor is given below.
Two concentrations of serum were studied here, 0.15 and 0.30 ml/ml. They show a
similar pattern when mixed with the basic preparation (BLES and protasan). The min-
imum surface tension values measured for both concentrations was significantly higher
and the elasticity of film compression was much lower than that for the basic preparation
Chapter 2. ADSA-CSD as a Micro Film Balance 72
alone. On the other hand, when the bulk phase of the basic preparation was replaced
with serum (under a preformed film), the minimum surface tension did not change signif-
icantly. It can be seen from figures 2.15(b) and 2.17(a) that the first cycle after the bulk
replacement can go to a very low surface tension value (indicating a good film at this
point). Starting with the second cycle, the minimum surface tension increased slightly.
Comparison of the γ −Ar isotherms in the same two figures with figure 2.15(a) suggests
that having serum under a preformed lung surfactant film does not change the minimum
surface tension but it does increase the hysteresis as shown in the figures. Obviously,
this increase in film hysteresis must be due to change in film properties other than the
minimum surface tension. Such properties include the film elasticity during the compres-
sion part as shown in figure 2.22. Another important property is the film relaxation at
the end of the compression, where further decrease in film area causes an increase in the
surface tension. It is apparent that the characterization of inhibition mechanisms will
require other surface properties in addition to the minimum surface tension; this will be
discussed further in Chapter 3.
Results obtained here from premixing albumin and BLES shows the inhibition of the
lung surfactant when it is to co-adsorb with albumin as shown in figure 2.18(b). The
minimum surface tension was increased and the elasticity was decreased. However, for
the case of introducing albumin under a preformed film of lung surfactant, results here
suggest that the dynamic surface tension lowering ability of the film was not altered.
Similar results were obtained before for a lower range of albumin concentrations (3 to 6
mg/ml) than used here using the hypophase exchange system with PBS [34]. However,
since we have here complete information about the change of surface tension and area
during the cycling, we are able to report that the elasticity of the film has changed as
shown in figure 2.18(a). In this figure, the minimum surface tension was reached at a
lower degree of area reduction. This suggests a slight increase in the film elasticity in
this case.
Chapter 2. ADSA-CSD as a Micro Film Balance 73
Results obtained here suggest that fibrinogen is more potent than other inhibitors
when mixed with lung surfactant preparations as shown in figure 2.21, causing a very
significant increase in the minimum surface tension during the dynamic cycling. The
severe inhibition effect of fibrinogen has been reported in previous studies [116, 117, 120,
122, 140]. However, results show that injecting fibrinogen under a preformed film of
lung surfactant did not significantly change the minimum surface tension. As shown
in figure 2.19(a), the first cycle after the bulk replacement can go to a very low surface
tension value, but the minimum surface tension increased slightly starting from the second
cycle. It is important to note that fibrinogen reduces the film elasticity much more
severely than other inhibitors when injected under a preformed surfactant film as shown
in figure 2.22.
Similar to the other inhibitors, cholesterol increases the minimum surface tension
when mixed with lung surfactant preparation. Results show that the effect of the same
concentration of cholesterol is different for the case of introducing the cholesterol to the
subphase of a preformed film versus premixing with the lung surfactant preparation prior
to film formation. When injected under a preformed film, cholesterol did not significantly
change the minimum surface tension and it actually improved the film elasticity as shown
in figures 2.21 and 2.22. Cholesterol, which is removed from complete lung surfactant
in the preparation of BLES, is one of the main neutral lipids in lung surfactant with
a concentration of around 10% of the surfactant lipids [141, 142]. Until recently, lung
surfactant inhibition has been attributed to factors other than cholesterol. However, it
was reported that elevated levels of cholesterol have the ability to impair and inhibit lung
surfactant, affecting its ability of dynamic surface tension lowering [128, 129, 143–145].
2.3.3.3 Inhibition Reversal
The methodology developed above allows study of inhibition reversal, i.e. the restoration
of the normal surface activity of an inhibited lung surfactant film. This is illustrated in
Chapter 2. ADSA-CSD as a Micro Film Balance 74
figure 2.23. While a mixture of BLES and protasan has been used to form the initial
film so far, in this experiment a film was originally formed using serum only as the basic
preparation. The serum was allowed to adsorb first to form a film at the interface and the
dynamic surface tension was measured as shown in figure 2.23(a). Then the serum bulk
was replaced by a preparation of BLES and protasan and the dynamic surface tension
was measured again as shown in figure 2.23(b). The first few cycles are close to those
before bulk replacement. Starting from the sixth cycle the minimum surface tension
started to decrease from 38.0 ±0.2 mJ/m2 to 5.7±0.6 mJ/m2. Beyond the twelfth cycle
there was no further significant change in the minimum surface tension. The changes
from cycles 6 to 12 indicate that the preparation of BLES and protasan was able to
replace serum components at the interface, effectively reversing the serum inhibition. A
low concentration of serum (0.01 ml/ml) was used in this experiment just to illustrate
the capacity of the new methodology proposed here. More detailed investigations can be
readily performed.
In a recent study, spreading experiments using a Langmuir trough showed that serum
(with concentrations as low as 0.0017 ml/ml) alone adsorbs to the air-water interface
[118]. When a clinical lung surfactant (Curosurf) was spread on top of the adsorbed
serum film, only minimal decreases in surface tension were observed. This is similar to
what is seen in figure 2.23 where the equilibrium surface tension did not change before
and after the bulk replacement. However, in that previous study, only adsorption time
could be investigated, not the dynamic cycling.
In conclusion, a novel methodology for studying adsorbed or spread surfactant films
before and after replacing one bulk subphase with another possibly also containing sur-
face active material was presented. This technique allowed lung surfactant films to be
formed by adsorption from a surfactant subphase and to be studied on a new subphase
containing inhibitor molecules rather than surfactant. Such studies were performed at
physiologically relevant conditions, mimicking the leakage of plasma and blood proteins
Chapter 2. ADSA-CSD as a Micro Film Balance 75
into the alveolar spaces. The main advantage of the new method, in addition to com-
plete leakage elimination and controlled physiologically relevant conditions inherent in
ADSA–CSD, is the ability to exchange the subphase after forming adsorbed or spread
films.
Chapter 3
Dynamic Surface Tension Model∗
3.1 Introduction
A key property of lung surfactants is their dynamic surface tension response to film com-
pression and expansion. To assess the performance of surfactant preparations subject
to dynamic compression/expansion, typically only the minimum surface tension at the
end of compression is reported. While the minimum surface tension is useful for a pre-
liminary assessment, it has to be realized that it depends on the method of compression
[46]. Increasing the extent of compression or the speed of compression may significantly
change the value of the minimum surface tension achieved [25]. In fact, this makes the
comparison of different literature values difficult in the case of different compression pro-
tocols. Therefore, there is a need to concentrate on the properties of the film itself such
as the elasticity and the relaxation as proposed in the literature [25, 147] rather than
simply reporting the minimum surface tension achieved. In this chapter, a new approach
to evaluate the quality of lung surfactant preparations is developed – beyond the γmin
value as the only quantitative characteristic – based on determining the film properties
for any compression protocol.
∗A portion of this chapter was previously published in Ref. [146], reproduced with permission.
76
Chapter 3. Dynamic Surface Tension Model 77
In ADSA–CSD experiments, a large amount of data is collected including the surface
tension as a function of time as well as the change of the drop surface area and vol-
ume. Such results can provide valuable information about various properties of the film.
ADSA–CSD is used to perform dynamic cycling by successive compression/expansion of
the drop via programmed cycling of the drop volume. This is normally done through 20
cycles with a periodicity of 3 seconds per cycle and a compression ratio (percentage of
surface area reduction) of 20% to simulate normal human breathing.
A typical experimental result for 2.0 mg/ml BLES at 100% relative humidity and
37C is shown in figure 3.1. Dynamic cycling starts after an equilibrium surface tension
value due to adsorption is reached. The response of surface tension for a trapezoidal
change of drop surface area is shown. Parenthetically, the trapezoidal perturbation of
the drop volume is a consequence of the backlash of the stepping motor controlling the
syringe. The fluid mechanics rounds off the corner of the trapezoidal perturbation as
shown in the figure. In region A, the drop volume is decreased causing a decrease in
drop surface area and a corresponding decrease in surface tension due to compression.
In region B, the drop volume is kept constant while the motor is reversing its direction;
therefore, the surface area is almost constant. The surface tension in this region starts to
increase towards the equilibrium value. In region C, the drop volume is increased causing
an increase in drop surface area and a corresponding increase in surface tension due to
expansion. In region D, the drop volume is kept constant while the motor is reversing its
direction to start a new cycle; therefore, the surface area is almost constant. The surface
tension in this region starts to decrease towards the equilibrium value again.
The response of surface tension for every part of the trapezoidal area change is useful
to understand different film properties. For regions A and C in figure 3.1, the response
of surface tension to a continuous decrease (A) or increase (C) in surface area gives an
indication of the dilatational elasticity of the surfactant film. For regions B and D in
figure 3.1, the response of surface tension to a constant surface area gives an indication
Chapter 3. Dynamic Surface Tension Model 78
Figure 3.1: The change of surface tension and surface area with time of one cycle for 2.0mg/ml BLES at wet conditions, 37C, and 20% compression ratio with a periodicity of3 seconds per cycle.
Chapter 3. Dynamic Surface Tension Model 79
of the interfacial adsorption/desorption of the surfactant film. If the surface tension is
above the equilibrium value, adsorption is dominant (region D) and promotes a decrease
in the surface tension towards the equilibrium value. However, if the surface tension is
below the equilibrium value, desorption is dominant (region B) and promotes an increase
in the surface tension towards the equilibrium value.
It is important to note that near the end of region A, there are two simultaneous
effects: elasticity (due to the continuous decrease in surface area) and desorption (because
the surface tension is below the equilibrium value). Similarly, near the end of region C,
there are two simultaneous effects: elasticity (due to the continuous increase in surface
area) and adsorption (because the surface tension is above the equilibrium value).
It is necessary to quantify the dynamic surface tension responses in regions A, B, C,
and D through adequate models that can reflect elasticity of the lung surfactant film as
well as adsorption and desorption at the interface. Such phenomena have been studied,
separately, by different groups. The most relevant studies are reviewed here. However,
for the case of pulmonary surfactant films, it is expected that more than one of these
phenomena are acting at any given time. The combined effect can be evaluated by simple
addition of individual effects. This idea is introduced and evaluated in this chapter.
The simplest possible model is considered here to explain the changes in the sur-
face tension during the dynamic cycling based on three main well known processes:
adsorption/spreading, desorption/relaxation, and elasticity/compressibility. The proce-
dure developed in this chapter is to evaluate the quality of lung surfactant preparations
– beyond the γmin value as the only quantitative characteristic – based on calculating the
film properties independent of the compression protocol used. The results show that the
proposed models can explain the performance of different lung surfactant preparations
based on the chosen effective dynamic parameter.
The structure of this Chapter is as follows: Previous dynamic models are reviewed
and summarized in Section 3.2. A proposed model called Compression–Relaxation Model
Chapter 3. Dynamic Surface Tension Model 80
(CRM) is introduced and developed in detail in Section 3.3. Sample experiments are used
with the previous models and the new developed CRM in Section 3.4 for comparison pur-
poses. The computational details of CRM and how to use the model with experimental
data are explained in the same section. Section 3.5 shows a detailed study of the dynamic
surface tension (using CRM) of a lung surfactant preparation, BLES, at concentrations,
compression ratios and compression rates relevant to current exogenous surfactant ther-
apies. A detailed sensitivity study of CRM is also presented in the same section.
3.2 Previous Models
There are two approaches to describe the dynamics of adsorption at liquid/air interfaces.
The diffusion controlled model assumes that the diffusional transport of interfacially
active molecules through the bulk is the rate limiting process and that the adsorption
is instantaneous. In the kinetic model, the adsorption and desorption steps are the rate
limiting processes.
3.2.1 Diffusion Model
One of the available diffusion models assumes that the surface tension relaxation is dom-
inated by a translational-diffusion mechanism [136]. According to this mechanism, the
surfactants diffuse through the subphase to reach the air/water interface, so that the
magnitude of surface tension reduction (∆γ) is:
∆γ1(t) =Ωεo2ωo
exp (2ωot) erfc(√
2ωot)
+Ωεo√t√
2πωo− Ωεo
2ωo0 < t ≤ t1 (3.1a)
∆γ2(t) = ∆γ1(t)−∆γ1(t− t1) t1 < t ≤ t2 (3.1b)
∆γ3(t) = ∆γ2(t)−∆γ2(t− t2) t2 < t ≤ t3 (3.1c)
∆γ4(t) = ∆γ3(t)−∆γ3(t− t3) t > t3 (3.1d)
Chapter 3. Dynamic Surface Tension Model 81
where Ω is the area constant, εo is the elasticity constant, and ωo is the diffusion constant:
Ω =d lnA
dt=
1
t1ln
(1− ∆A
Ao
), εo = − dγ
d ln Γ, ωo =
D
2
(dc
dΓ
)2
(3.2)
where Γ and c are the surface and bulk concentrations, and D is the diffusion coefficient.
This surface tension response is based on a trapezoidal change in surface area where t1,
t2 and t3 are the limits of different time domains associated with this change as shown
in regions A, B, C, and D in figure 3.1.
The model parameters εo and ωo (defined above) are surface thermodynamic proper-
ties. Their values, obtained from fitting different surfactants, depend on the concentra-
tion c and on the surface activity dγ/dc. This model was compared to experiments using
a variety of soluble surfactants: n-dodecyl-dimethyl-phosphine oxide (DC12PO), sodium
bis(2-ethylhexyl)-sulfo-succinate (DESS), and n-octadecyl-trimethyl-ammonium bromide
(STAB). Good agreement was obtained between experimental data and the theoretical
curves for the selected soluble surfactants [136]. Other models were also developed for
soluble surfactants [148, 149]. However, the diffusion model may not be the most ap-
propriate for insoluble surfactant systems, like lung surfactants, where the exchange of
matter with sub-surface material controls the dynamic of these systems [150].
3.2.2 Kinetic Model
As mentioned above, other models exist beside the diffusion-controlled one to describe
the adsorption kinetics and exchange of matter. Most of them are based on the Langmuir
hypothesis about the nature of the adsorbed layer considering that there is an equilibrium
between the subsurface and the bulk of the solution. The rate of adsorption can be written
as the difference between adsorption and desorption as follows [151]:
dΓ
dt= k1c
(1− Γ
Γ∗
)− k2
Γ
Γ∗ (3.3)
Chapter 3. Dynamic Surface Tension Model 82
where Γ is the surface concentration, Γ∗ is the maximum equilibrium surface concentra-
tion, c is the bulk concentration, k1 is the adsorption coefficient, and k2 is the desorption
coefficient.
A model for an adsorption layer relaxation was derived [151] by considering the time-
dependent square pulse area A(t). In this model, the interfacial tension response can be
calculated as:
∆γ1(t) = εo∆A
Aoexp (−Kt) 0 < t ≤ t1 (3.4a)
∆γ2(t) = ∆γ1(t)−∆γ1(t− t1) t > t1 (3.4b)
where t1 is time range of the square pulse cycle while the area is compressed, ∆A/Ao is
the relative area change, εo is the elasticity coefficient and K is the sorption coefficient
(Langmuir Isotherm).
εo = − dγ
d ln Γ= RTΓ∗ Γ/Γ∗
1− Γ/Γ∗ , K =k1c
Γ∗ +k2Γ∗ (3.5)
Equation 3.4 can be used to interpret experimental data by the use of a fitting proce-
dure. It is important to note that equation 3.4 suggests that adsorption and relaxation
are first order processes. The resulting values for εo and K are obtained independently
of any knowledge of the parameters of the equilibrium adsorption isotherm. On the
other hand, both parameters contain specific constants of the adsorption isotherm and
are therefore of interest also from this point of view.
The results of relaxation experiments of human albumin (HA) adsorbed at the aque-
ous solution/air interfaces [152] were discussed on the basis of the diffusion model as
shown in equation 3.1. The resulting diffusion coefficients were found to be two to three
orders of magnitude higher than the physically expected value. It was concluded that
the adsorption layer of human albumin at the aqueous solution/air interface shows a dy-
Chapter 3. Dynamic Surface Tension Model 83
namic behaviour similar to that of surfactants, while relaxations after area disturbances
are reversible and controlled by a mechanism other than diffusion.
The same experimental data were interpreted using both models [151]: the diffusion
model (equation 3.1) and the kinetic model (equation 3.4). It was shown that the kinetic
model yields values of the parameters which are rather consistent (considering different
cycles of the same run) while those obtained from the diffusion model are more scattered.
3.2.3 Adsorption-Limited Model
In another model, transport of surfactant to the interface is assumed to be adsorption-
limited as opposed to diffusion-limited [153]. Several assumptions were adopted; one
is that there exists a maximum surfactant concentration to which the interface can be
dynamically packed, and when interfacial area is reduced from this point, surfactant is
squeezed out from the surface. Another assumption is that transport of surfactant to the
interface is adsorption rather than diffusion limited. When the interfacial concentration
of surfactant is less than its maximum equilibrium value, adsorption and desorption are
controlled by Langmuir kinetics, while for interfacial surfactant concentrations between
the maximum equilibrium value and the maximum packing concentration, the surfactant
is frozen in the surface and can neither adsorb nor desorb. The kinetics of adsorption
and desorption were defined in this model by three interfacial surface concentration (Γ)
regimes:
d(AΓ)
dt= A [k1C(Γ∗ − Γ)− k2Γ] Γ < Γ∗ (3.6a)
d(AΓ)
dt= 0 Γ∗ < Γ < Γmax (3.6b)
d(AΓ)
dt= −Γmax
dA
dtΓ = Γmax (3.6c)
where k1 is the adsorption coefficient, k2 is the desorption coefficient, C is the bulk
concentration, Γ∗ is the maximum equilibrium surface concentration, and Γmax is the
Chapter 3. Dynamic Surface Tension Model 84
maximum dynamic surface concentration (collapse).
The relationship between the normalized interfacial concentration, Γ/Γ∗, and the
surface tension, γ, is the compression isotherm. The compression isotherm is modeled
as two straight lines that meet at Γ = Γ∗, the slopes for both regions are m1 and m2.
Therefore, surface tension can be calculated as follows:
γ = γo −m1Γ
Γ∗Γ
Γ∗ < 1.0 (3.7a)
γ = γ∗ −m2
(Γ
Γ∗ − 1
)ΓmaxΓ∗ ≥
Γ
Γ∗ ≥ 1.0 (3.7b)
where γo is the water surface tension, γ∗ is the equilibrium surface tension, m1 is the
isotherm slope for γ > γ∗, and m2 is the isotherm slope for γmin < γ < γ∗.
This model [153] was used to characterize the dynamic behaviour of lung surfactant
fractions [154], containing: (1) phospholipids (PL) only, (2) PL and hydrophobic apopro-
teins (HA), (3) PL and HA and surfactant protein A (SP-A), and (4) PL and HA and
SP-A and neutral lipids (NL). The results were compared to those of native surfactant
and dipalmitoyl-phosphatidyl-choline (DPPC). The effect of individual components on
the surface activity (minimum surface tension, adsorption, film compressibility) of the
mixture was evaluated by comparing the experimental results to the model predictions.
It was noted that the model produces predictions that are qualitatively but not quanti-
tatively close to the experimental results for most of the cases.
The adsorption-limited model [153] was later extended to characterize the initial tran-
sient behaviour (that occurs upon initiation of pulsation, before steady-state was reached)
by including the effects of diffusion in the bulk phase and comparing the results to both
transient and steady-state oscillatory data [155]. It was concluded that at steady-state
cycling conditions, diffusion-limited transport is most relevant. While there are a num-
ber of situations in which the extended model gives significantly improved predictions,
steady-state oscillatory data at high bulk concentration are still adequately modeled by
Chapter 3. Dynamic Surface Tension Model 85
an adsorption-limited model, which is computationally more efficient. When compared
to previous experimental data [154], both models yield results that are very close to each
other but are far from the experimental results.
3.3 Development of Compression–Relaxation Model
As inferred from the dynamic experiments, there are four main factors that affect the
response of the dynamic surface tension: adsorption or spreading, desorption or relax-
ation, elasticity during compression and elasticity during expansion. In the proposed
model, the adsorption and relaxation rates depend on how far the surface tension is from
the equilibrium value while the elasticity for both compression and expansion depends
on changes in the surface area of the interface. The main feature is to allow the surface
tension to change due to simultaneous effects of one or more of these parameters.
3.3.1 Adsorption Process (Spreading)
Based on the Langmuir hypothesis, the rate of adsorption can be written as the difference
between adsorption and desorption as shown in equation 3.6a. As mentioned before, Γ∗ is
the maximum equilibrium interfacial concentration, i.e. the limiting value of Γeq reached
as c (concentration) is increased. Also, Γeq can be calculated as [153]
ΓeqΓ∗ =
k1c
k1c+ k2(3.8)
Note that equation 3.6a predicts adsorption for surface concentration less than the
equilibrium concentration, Γeq, and desorption for concentrations greater than this. There-
fore, for surface concentration less than Γeq (surface tension greater than γeq) and for a
given concentration and constant surface area, that equation can be simplified to
dΓ
dt= k(Γeq − Γ) (3.9)
Chapter 3. Dynamic Surface Tension Model 86
where k is a coefficient for the dynamic adsorption and desorption. This indicates that
both rates of adsorption/spreading and desorption/relaxation are proportional to the
difference between the surface coverage at any time and the equilibrium surface coverage.
However, since the mechanisms of adsorption and desorption for pulmonary surfactant
compounds are different, we use here two coefficients, one for adsorption and another for
desorption.
Assuming a linear relationship between the surface concentration and the surface
pressure in the region above the equilibrium surface tension (equivalent to the assump-
tions behind equation 3.7 in the adsorption-limited model [153]), equation 3.9 can be
further simplified to
dγ
dt= ka(γeq − γ) for γ ≥ γeq (3.10)
where ka is the adsorption coefficient.
It is important to note that the above equations were derived for a single component
system. It is assumed here that it is also valid for multi-component systems, where
the coefficient now has to be understood as an effective constant for the system, not any
specific material. The comparison between experimental results and the developed model
will show whether this assumption is reasonable.
3.3.2 Desorption Process (Relaxation)
As mentioned earlier, equation 3.6a predicts adsorption for surface concentration less
than the equilibrium concentration, Γeq, and desorption for concentrations greater than
this. Therefore, for surface concentration greater than Γeq (surface tension less than γeq)
and constant surface area, a simplified desorption relation similar to that of equation 3.10
can be used,
dγ
dt= kr(γeq − γ) for γ ≤ γeq (3.11)
where kr is the desorption or relaxation coefficient.
Chapter 3. Dynamic Surface Tension Model 87
Non-equilibrium surface pressures occurring during dynamic compression are believed
to indicate the formation of metastable surface states within phospholipid films [1]. When
the compression is stopped (the surface area is kept constant), the dynamic surface
pressure relaxes towards the equilibrium values. Relaxation has also been studied [156–
158] in terms of a nucleation mechanism [159–161] through which a similar expression to
equation 3.11 can be derived. After compression beyond collapse, the dynamic surface
pressure becomes more stable and remains in the vicinity of the dynamic collapse value
over an extended time scale [162].
3.3.3 Elasticity
The dilational interfacial elasticity is defined by [136, 151],
εo = − dγ
d lnA(3.12)
which is equivalent to
dγ
dt= −εo
(1
A
dA
dt
)(3.13)
If we consider different elasticities for compression and expansion, the above equation
becomes,
dγ
dt=
εc
(1
A
dA
dt
)for
dA
dt≤ 0
εe
(1
A
dA
dt
)for
dA
dt≥ 0
(3.14)
3.3.4 Collapse
It is well known that many surfactant layers at the interface can be compressed to pres-
sures considerably higher than their equilibrium spreading pressure [1, 92, 93]. As dis-
cussed in Chapter 2, if compression continues beyond an ultimate collapse pressure, the
surfactant layers collapse ejecting surfactant molecules into one or more surface or subsur-
Chapter 3. Dynamic Surface Tension Model 88
face collapse structures or phases [1, 93]. Therefore, once the ultimate collapse pressure,
i.e. the minimum surface tension, is reached, the surface tension can not be further
decreased by more compression. This can be formulated as follows
dγ
dt= 0 for γ ≤ γmin (3.15)
3.3.5 The Integrated Model
Combining equations 3.10, 3.11, 3.14, and 3.15 and allowing the surface tension to change
due to simultaneous effects of relaxation/adsorption and elasticity, we obtain:
dγ
dt=
dγ1dt
+dγ2dt
if γ ≥ γmin
0 if γ ≤ γmin
(3.16a)
where
dγ1dt
=
ka (γeq − γ) if γ ≥ γeq
kr (γeq − γ) if γ ≤ γeq
(3.16b)
dγ2dt
=
εc
(1
A
dA
dt
)ifdA
dt≤ 0
εe
(1
A
dA
dt
)ifdA
dt≥ 0
(3.16c)
According to this model better lung surfactant preparations can be identified as having
faster adsorption rate (higher ka), slower relaxation rate (lower kr), and higher elasticity
of compression and expansion (higher εc and εe). Figure 3.2 shows a comparison between
two hypothetical dynamic surface tension responses, where one has desired dynamic
properties and the other does not. The figure shows schematics based on an assumed,
i.e. exact, trapezoidal change in surface area. However, it is important to note that the
proposed model does not assume any specific perturbation waveform and can be applied
Chapter 3. Dynamic Surface Tension Model 89
to various perturbations including trapezoidal and sinusoidal.
These dynamic parameters are actually very similar to those used in the previous
models. The elasticities of compression and expansion, εc and εe, are defined in the same
fashion as the elasticity constant, εo, in both the diffusion model [136] and the kinetic
model [151]. The adsorption coefficient, ka, is similar to that used in the adsorption-
limited model [153]. Thus, the equations involving these parameters are not novel, but
it is the simultaneous solution for these parameters that makes this model unique. The
model introduces the possibility of simultaneous compression and relaxation effects. It
is important to note that the calculations of these parameters in the current model are
based directly on the surface tension rather than the surface concentration. This allows
the direct use of ADSA output (changes in surface tension and surface area with time)
in the fitting procedure to estimate the film properties.
3.4 CRM Applied to Sample Experiments
3.4.1 Experimental Details
Details about the experimental setup and procedure were given in Chapter 2. During the
experiment, the setup is enclosed in an environmental control chamber that facilitates the
control of gas composition and temperature. The humidity inside the chamber was kept
constant (wet at 100% relative humidity or dry < 20% relative humidity) at 37C. The
dynamic cycling experiments were carried out by periodically injecting/withdrawing liq-
uid from the drop. After reaching the equilibrium surface tension value, dynamic cycling
is started. The dynamic cycling consists of four sub-stages: compression, relaxation,
expansion and re-adsorption, as indicated in the discussion of figure 3.1. Throughout
Section 3.4, the compression conditions were: compression periodicity of 3 s/cycle, and
compression ratio (fraction of surface area reduction) of 20% to simulate normal breathing
conditions [1, 163].
Chapter 3. Dynamic Surface Tension Model 90
Surf
ace A
rea, A
Time, t
(a)
Su
rfa
ce
Te
nsio
n, γ
Time, t
(b)
Su
rfa
ce
Te
nsio
n, γ
Relative Surface Area, Ar
(c)
Figure 3.2: The response of surface tension to a trapezoidal change in surface area usingthe new compression-relaxation model for a good (solid line) and a bad (dashed line)system: (a) the trapezoidal change of surface area with time; (b) the response of surfacetension with time for both systems; (c) the change of surface tension with the relativesurface area for both systems. The good system (solid line) is generated using parameters:εc = 150 mJ/m2, εe = 150 mJ/m2, kr = 0.1 s−1, ka = 5 s−1. The bad system (dashedline) is generated using parameters: εc = 70 mJ/m2, εe = 70 mJ/m2, kr = 0.3 s−1, ka =0.1 s−1.
Chapter 3. Dynamic Surface Tension Model 91
It was shown in recent studies [25, 46, 47] that BLES films are less stable in humid
environments than in dry ones and tend to have lower surface activity compared to BLES
films in dry air. To further investigate the effect off humidity, five formulations were used
for model comparisons: 0.5 mg/ml BLES (humid air) [25, 91], 0.5 mg/ml BLES (dry air)
[25, 91], 2.0 mg/ml BLES (humid air) [91], 2.0 mg/ml BLES (dry air) [91], and 0.5 mg/ml
BLES mixed with 2.5 mg/ml Albumin (humid air) [47]. The last formulation is chosen
to illustrate the effect of a known inhibitor on the surface activity and film properties of
BLES.
3.4.2 Computational Details
The proposed compression/relaxation model, as summarized in equation 3.16, is able
to predict the surface tension response of a surfactant preparation for a given change
in surface area with time, assuming an a priori knowledge of the above mentioned four
dynamic parameters: adsorption coefficient (ka), desorption coefficient (kr), elasticity of
compression (εc), and elasticity of expansion (εe). The numerical procedure for determin-
ing these four parameters requires three steps: First, the differential equation 3.16 must
be solved using numerical integration to generate a surface tension response. Second,
initial guesses for the four parameters are needed; the experimental results are analyzed
to calculate approximate values (initial guesses) for these parameters. Third, an opti-
mization procedure is required to compare the calculated surface tension values with the
experimentally measured values, and modifying the parameter values until a matched
surface tension response is obtained. These three steps are explained in detail below.
3.4.2.1 Numerical Integration
Equation 3.16 is considered as a set of ordinary differential equations and can be solved
using an initial value problem solver for given values of the four dynamic parameters.
A variable order Adams-Bashforth-Moulton PECE solver [164] is implemented for the
Chapter 3. Dynamic Surface Tension Model 92
numerical integration. This method is considered more efficient than the Runge-Kutta
solver at stringent tolerances and when the ordinary differential equations set is particu-
larly expensive in computation time. This solver is a multi-step solver; it normally needs
the solutions at four preceding time points to compute the solution at a specific point in
time. The first point is usually the first surface tension value measured experimentally.
The following three points are usually calculated using the Runge-Kutta solver and then
the Adams-Bashforth-Moulton solver is used afterwards.
3.4.2.2 Initial Guess
As mentioned earlier, an initial guess for the dynamic parameters is required for the
curve fitting module. Simple methods previously used to calculate dilatational elasticity,
adsorption, and relaxation rates in the literature [25, 47, 91] are used to estimate initial
values.
For the elasticity of compression and expansion, the change of surface tension with the
relative surface area during the compression/expansion part of every cycle is fitted to a
fourth order polynomial and the slope is evaluated at the midpoint [91]. The calculated
slope and the value of the relative surface area are used to calculate the dilatational
elasticity according to equation 3.12. This is similar to what was used in Chapter 2.
For the adsorption and desorption coefficients, the surface tension increase/decrease
with time (region B and D in figure 3.1) is fitted to a fourth order polynomial and the
slope is evaluated at the midpoint [91]. The calculated slope and the difference between
the surface tension value and the equilibrium surface tension value are used to calculate
the adsorption and desorption coefficients according to equations 3.10 and 3.11.
3.4.2.3 Curve Fitting (Optimization Procedure)
Once the initial guesses for the dynamic parameters are calculated, the numerical in-
tegration is performed and the calculated surface tension response is compared to the
Chapter 3. Dynamic Surface Tension Model 93
experimental values using a curve fitting module (optimization procedure). Multidimen-
sional unconstrained nonlinear minimization (Nelder-Mead) is used to find the minimum
of a fitting residual function, < (see equation 3.17 below), using the simplex derivative-
free search method [165]. This is a direct search method that does not use numerical or
analytic gradients. At the end of this module, the four dynamic parameters are obtained
from the experimental result of a specific cycle. The same procedure is applied for differ-
ent cycles (typically 20 cycles) and the values reported below are the average value and
the standard deviation for a specific preparation.
3.4.3 Results
The proposed compression-relaxation model for dynamic characterization is implemented
using the above mentioned numerical details. Four preparations (high and low concentra-
tion of BLES with humid and dry conditions) were selected to illustrate the significance
of different parameters. Figure 3.3 shows the change of surface tension with the relative
surface area of one cycle for these preparations. The compression/expansion cycles of
surface layers are commonly used for qualitative and quantitative data analysis as pro-
posed in the literature [166]. The model predictions fit most of the experimental points
well. Figure 3.4 shows the change of surface tension and surface area with time for the
same four preparations. In the same figure, the compression-relaxation model is com-
pared to other models, i.e. the adsorption-limited model and the diffusion model. The
kinetic model was considered but can not be compared because it was designed to work
only with square pulse changes in surface area, which is not the case for the experiments
considered here. The results in figure 3.4 show that the proposed compression-relaxation
model has better quantitative agreement with experimental results than other models.
A special case is shown in figure 3.5, where the film has collapsed at a high surface
tension due to the presence of albumin in BLES. The proposed compression-relaxation
model shows again a better fit with the experimental points compared to other models.
Chapter 3. Dynamic Surface Tension Model 94
0.8 0.9 10
10
20
30
40BLES 0.5 mg/ml (H)
Relative Surface Area, Ar
Surf
ace T
ensio
n, γ
(m
J/m
2)
Exp.
CRM
(a)
0.8 0.9 10
10
20
30
40BLES 0.5 mg/ml (D)
Relative Surface Area, Ar
Surf
ace T
ensio
n, γ
(m
J/m
2)
Exp.
CRM
(b)
0.8 0.9 10
10
20
30
40BLES 2.0 mg/ml (H)
Relative Surface Area, Ar
Surf
ace T
ensio
n, γ
(m
J/m
2)
Exp.
CRM
(c)
0.8 0.9 10
10
20
30
40BLES 2.0 mg/ml (D)
Relative Surface Area, Ar
Surf
ace T
ensio
n, γ
(m
J/m
2)
Exp.
CRM
(d)
Figure 3.3: The change of surface tension with the relative surface area of one cycle forfour different preparations of diluted and concentrated BLES in humid (H) and dry (D)conditions. The compression-relaxation model (CRM) predictions is compared to theexperimental points.
Chapter 3. Dynamic Surface Tension Model 95
211 212 213 2140
10
20
30
40BLES 0.5 mg/ml (H)
Time, t (s)
Surf
ace T
ensio
n, γ
(m
J/m
2)
Exp.
CRM
ALM
DM
(a)
211 212 213 2140
10
20
30
40BLES 0.5 mg/ml (D)
Time, t (s)
Surf
ace T
ensio
n, γ
(m
J/m
2)
Exp.
CRM
ALM
DM
(b)
211 212 213 2140
10
20
30
40BLES 2.0 mg/ml (H)
Time, t (s)
Surf
ace T
ensio
n, γ
(m
J/m
2)
Exp.
CRM
ALM
DM
(c)
211 212 213 2140
10
20
30
40BLES 2.0 mg/ml (D)
Time, t (s)
Surf
ace T
ensio
n, γ
(m
J/m
2)
Exp.
CRM
ALM
DM
(d)
Figure 3.4: The change of surface tension with time of one cycle for four different prepa-rations of diluted and concentrated BLES in humid (H) and dry (D) conditions. Thecompression-relaxation model (CRM) predictions is compared to the adsorption-limitedmodel (ALM), the diffusion model (DM), and the experimental points.
Chapter 3. Dynamic Surface Tension Model 96
A summary of the four dynamic parameters calculated from the compression-relaxation
model for the different five preparations is shown in table 3.1. The third column of
the table shows the values of elasticity of compression calculated from an elementary
procedure used previously in the literature [91] and in Chapter 2. The values in this
table will be discussed in detail in the next section.
Figure 3.6 compares the values of the fitting residual function from different models
as an indication of the goodness of fit. The fitting residual function is defined as:
< =√∑
|γe − γm|2 (3.17)
where γe is the experimentally measured surface tension value and γm is the predicted
surface tension value from the respective model. Five cases (the four cases of diluted
and concentrated BLES in humid and dry conditions shown in figure 3.3 and the case
of adding albumin shown in figure 3.5) were considered here. It is apparent that the
compression-relaxation model gives the best fit for all five cases. The adsorption-limited
model performed very well in the second case (film collapse at low surface tension with
minimum relaxation) but performed poorly in the third case (relaxation process with no
collapse).
Figure 3.7 shows a comparison between the calculated values of elasticity between
different models. In the proposed compression-relaxation model, two different values
were considered (one for compression and another for expansion). It is noted that the
difference between calculated elasticity values are not very big except for the diffusion
model in two of the five cases.
Chapter 3. Dynamic Surface Tension Model 97
211 212 213 2140
10
20
30
40BLES 0.5 mg/ml & Albumin 2.5 mg/ml (H)
Time, t (s)
Su
rfa
ce
Te
nsio
n, γ
(m
J/m
2)
Exp.
CRM
ALM
DM
(a)
0.8 0.9 10
10
20
30
40BLES 0.5 mg/ml & Albumin 2.5 mg/ml (H)
Relative Surface Area, Ar
Su
rfa
ce
Te
nsio
n, γ
(m
J/m
2)
Exp.
CRM
(b)
Figure 3.5: A special case of collapse at high surface tension of diluted BLES mixedwith albumin in humid (H) condition: (a) the change of surface tension with the relativesurface area of one cycle; (b) the change of surface tension and surface area with time ofone cycle. CRM: compression-relaxation model, ALM: adsorption-limited model, DM:diffusion model.
Chapter 3. Dynamic Surface Tension Model 98
Tab
le3.
1:Sum
mar
yof
the
dynam
icpar
amet
ers
(mea
nan
dst
andar
ddev
iati
on)
calc
ula
ted
from
the
com
pre
ssio
n-r
elax
atio
nm
odel
for
diff
eren
tpre
par
atio
ns
at20
%co
mpre
ssio
nan
d3
seco
nds
per
cycl
e.
Pre
par
atio
n
Ref
.[9
1]C
RM
Hum
idit
yE
last
icit
yof
Ela
stic
ity
ofE
last
icit
yof
Rel
axat
ion
Adso
rpti
onC
ondit
ion
Com
pre
ssio
nC
ompre
ssio
nE
xpan
sion
Coeffi
cien
tC
oeffi
cien
tε c
(mJ/m
2)
ε c(m
J/m
2)
ε e(m
J/m
2)
kr
(s−1)
ka
(s−1)
BL
ES
0.5
mg/
ml
Hum
id99
.8±
2.9
125.
1±3.
815
7.8±
2.5
0.54
7±0.
018
2.47
4±0.
264
BL
ES
0.5
mg/
ml
Dry
105.
9±3.
811
2.7±
1.4
120.
0±3.
10.
001±
1e-4
2.78
3±0.
380
BL
ES
2.0
mg/
ml
Hum
id49
.1±
8.7
126.
3±2.
813
6.0±
15.3
3.75
1±0.
184
4.99
1±0.
570
BL
ES
2.0
mg/
ml
Dry
108.
9±5.
312
3.1±
2.1
129.
9±1.
40.
006±
0.00
10.
712±
0.12
6
BL
ES
0.5
mg/
ml
&H
um
id10
.3±
2.6
72.4±
10.6
77.6±
3.0
2.50
1±0.
400
1.55
9±0.
052
Alb
um
in2.
5m
g/m
l
*E
xp
erim
enta
lre
sult
suse
dher
ew
ere
take
nfr
omth
elite
ratu
re[2
5,47
,91
].
Chapter 3. Dynamic Surface Tension Model 99
0
10
20
30
40
50
Fittin
g R
esid
ua
l, ℜ
(m
J/m
2)
BLES (H)0.5 mg/ml
BLES (D)0.5 mg/ml
BLES (H)2.0 mg/ml
BLES (D)2.0 mg/ml
BLES 0.5 mg/ml (H)Albumin 2.5 mg/ml
CRM
ALM
DM
Figure 3.6: Comparison between the goodness of fit (mean and standard deviation) ofdifferent models. CRM: compression-relaxation model, ALM: adsorption-limited model,DM: diffusion model.
Chapter 3. Dynamic Surface Tension Model 100
0
100
200
300
400
500
600
Ela
sticity, ε
(m
J/m
2)
BLES (H)0.5 mg/ml
BLES (D)0.5 mg/ml
BLES (H)2.0 mg/ml
BLES (D)2.0 mg/ml
BLES 0.5 mg/ml (H)Albumin 2.5 mg/ml
CRM (comp)
CRM (exp)
ALM
DM
Figure 3.7: Comparison between the predicted value of elasticity (mean and standard de-viation) using different models. CRM: compression-relaxation model, ALM: adsorption-limited model, DM: diffusion model.
Chapter 3. Dynamic Surface Tension Model 101
3.4.4 Discussion
3.4.4.1 Comparison of Models
The compression-relaxation model presented here shows better fitting than other models
for the five experimental cases shown here. The model can quantify the dynamic pa-
rameters that affect the response of the dynamic surface tension: adsorption coefficient,
desorption or relaxation coefficient, and dilatational elasticity of compression and expan-
sion. The main feature in this model is that more than one process is occurring at any
given time (compared to different regions of specific mechanism for each). This is evident,
for example, in the experimental result of humid concentrated BLES, figure 3.3(c), where
the surface tension starts to increase near the end of the compression stage indicating a
relaxation process occurring during the change in area not only when the area is constant.
This suggests a combined and simultaneous effect of relaxation and elasticity. A similar
effect is observed near the end of the expansion stage of the same experiment, where
the surface tension starts to decrease, suggesting that surfactant adsorption is already
occurring during expansion.
The diffusion model did not show good agreement with most of the experimental
results. The main assumption in that model is that the diffusion of surfactant molecules
through the subphase to reach the interface is the dominant physical phenomenon. In
this model, the diffusion coefficient is the same for diffusion to or from the interface.
This would apply to soluble surfactants like DC12PO, DESS and STAB [136]. However,
pulmonary surfactants are insoluble surfactants suspended as vesicles or other aggregates
and spread at the interface forming possibly multilayers [137–139, 167].
The adsorption-limited model did show a reasonable agreement with experimental
results for some of the cases. However, for humid diluted and humid concentrated BLES,
the adsorption-limited model was not able to capture the desorption/relaxation phe-
nomenon. According to this model, the surfactant can only collapse, i.e. be squeezed
Chapter 3. Dynamic Surface Tension Model 102
out, but can neither adsorb nor desorb below the minimum equilibrium surface tension,
γ∗. The better agreement with experimental results for dry diluted and concentrated
BLES is likely due to the fact that relaxation is not pronounced in these cases.
3.4.4.2 Calculated Dynamic Parameters
Besides the minimum surface tension value, the calculated dynamic properties can be
considered a measure of the quality of the lung surfactant preparation. For example,
with respect to the calculated elasticity value for all five preparations shown in figure 3.7
and table 3.1, results suggest that there is not much variation in the elasticity values
calculated for humid and dry, diluted and concentrated BLES. However, when albumin
is added to BLES, the elasticity value is markedly decreased from approximately 125 to
75 mJ/m2, indicating surfactant inactivation or inhibition. This conclusion agrees with
findings in the literature [34, 76, 116, 117, 122, 168–170] that albumin poisons surface
activity of lung surfactant preparations.
For most of the preparations studied, the values obtained for the elasticities of com-
pression and expansion are quite close except for the case of diluted BLES at high hu-
midity. The difference between the values of elasticity calculated during compression and
expansion may suggest that the film composition during compression is not the same as
during expansion, which is consistent with the hypothesis that some less surface active
material is squeezed out of the film [1, 171, 172]. Such changes in the film composition
may happen during the course of compression or just at the end of the compression.
These possibilities should be further investigated. A hydration-film fluidization mech-
anism has been proposed to explain the effect of humidity on the film elasticity [25].
However, this effect is minimal when the compression periodicity is short enough for the
hydration effects not to take place(∼3 s/cycle) and the elasticity remains almost constant
[25].
The desorption (relaxation) coefficient calculated here is also an important quality
Chapter 3. Dynamic Surface Tension Model 103
measure for lung surfactant preparations. Comparing the calculated values for diluted
and concentrated BLES in table 3.1, it is clear that the relaxation coefficient increases
with the increase of concentration for both humid and dry conditions. However, there is
a distinctive effect of humidity on the calculated value of the relaxation coefficient. For
both diluted and concentrated BLES, the dry condition causes a more stable preparation;
the relaxation coefficient is two orders of magnitude smaller than in the case of humid
conditions. This indicates that little or no significant relaxation is experienced by BLES
films in dry air as suggested in the literature on the basis of the hydration-film fluidization
mechanism [25]. Comparing the diluted humid BLES to the same preparation mixed with
albumin, the increase of the relaxation coefficient is another indication of lung surfactant
inhibition.
In summary, the dynamic compression-relaxation model is used to describe the me-
chanical properties of insoluble pulmonary surfactant films by investigating the response
of surface tension to changes in surface area. This model introduces the possibility of
simultaneous compression and relaxation effects. The model is able to fit different exper-
imental results of diluted and concentrated BLES preparation evaluated in humid and
dry air, as well as BLES and albumin mixtures. The calculated dynamic parameters
confirm the known effect of albumin on changing the elasticity and desorption of BLES
preparations. The effect of humidity on the desorption process is illustrated by the huge
increase in the relaxation coefficient in humid conditions, indicating that the dry condi-
tion causes a more stable surfactant films. Since the humid conditions are physiologically
relevant to lung surfactant studies, the effect of humidity introduces some doubts and
questions regarding the optimistic results of lung surfactant films in dry conditions in
the literature.
Chapter 3. Dynamic Surface Tension Model 104
3.5 Effect of Concentration, Compression Ratio and
Compression Rate
In this section, ADSA–CSD is used to measure the dynamic surface tension of a lung sur-
factant preparation, BLES, at concentrations, compression ratios and compression rates
relevant to current exogenous surfactant therapies. The results are analyzed using CRM
to evaluate the effect of surfactant concentration, compression ratio and compression rate
on the surface activity and dynamic properties (elasticity, adsorption and relaxation) of
BLES preparations.
As indicated in Chapter 1 exogenous surfactant replacement therapy, in which either
synthetic or modified natural pulmonary surfactant (extracted from bovine or porcine
sources) is delivered into the patients’ lungs, has been established as a standard therapeu-
tic intervention for patients with nRDS [173]. Surfactant therapy has shown only limited
therapeutic effect on ARDS patients [174–177]. In these therapies, high concentrations of
the exogenous surfactant are commonly used to reach an effective dose, i.e. the surfactant
dose to recover the mechanical properties of the lungs. The dose is usually in the range
of 8 mg/ml [1, 178, 179]. Some other formulations, e.g. bovine lipid extracted surfactant
(BLES) preparations, are used in surfactant replacement therapy with larger doses of 27
mg/ml [25]. However, there is a lack of in vitro studies for these high concentrations
due to limitations of methodologies used. For these studies, the maximum surfactant
concentration is usually restricted to no more than 3 mg/ml. This restriction arises from
optical limitations since surfactant suspensions become murky and eventually opaque at
increased concentrations.
Although there is promising evidence for surfactant therapy in ARDS, the effec-
tiveness of high doses of commercial surfactant preparations has been inconsistent in
clinical trials [177, 180–184]. However, surfactant concentration is only one strategy
to address the problem. High frequency low tidal volume ventilation, henceforth sim-
Chapter 3. Dynamic Surface Tension Model 105
ply referred to as high frequency ventilation (HFV) has been shown to reduce, even if
marginally, the mortality of ARDS [181, 185–189]. To simulate normal breathing, cy-
cling frequencies of 0.1 - 0.5 Hz and a reduction in surface area to near 20% are necessary
[1, 25, 36, 172, 190, 191]. However, for high frequency ventilation it is necessary to
produce cycling frequencies higher than 1 Hz [192].
As indicated in Chapter 1, ADSA–CSD is capable of measuring surface tension of
lung surfactants at a wide range of concentrations, compressions ratios and frequencies.
In fact, in contrast to other methodologies, there is no upper limit to the surfactant
concentration that can be used in ADSA–CSD. In this section, ADSA–CSD is used in
conjunction with CRM to measure the effective dynamic properties of BLES at condi-
tions relevant to conventional (high concentration) and high frequency surfactant thera-
pies. Results show that CRM is generally sensitive to all the parameters, i.e., elasticity,
adsorption and relaxation, at any concentration at physiological conditions of 20% com-
pression and 3 seconds per cycle. The model is also very sensitive to all the parameters
at any concentration at low compression ratios and high frequency cycling. These later
conditions are relevant to the clinical practice of high frequency ventilation.
3.5.1 Experimental Details
Details about preparation of BLES, the experimental setup and procedures are given in
Chapter 2. During the experiment, the setup is enclosed in an environmental control
chamber that facilitates the control of gas composition and temperature. The humidity
inside the chamber was kept constant at 100% relative humidity at 37C. Drop images
were acquired throughout the experiment at a rate of 20 images per second for cycling
speeds of 3 or 9 seconds per cycle and at a rate of 30 images per second for a cycling
speed of 1 second per cycle.
The experimental procedure was as follows: First, the sessile drop was quickly formed
(within 0.5 s) on the pedestal. The drop was then left undisturbed to allow adsorption
Chapter 3. Dynamic Surface Tension Model 106
of the lung surfactant film. The surface tension was tracked until the film reaches the
equilibrium surface tension (within 180 seconds for all cases). Thereafter, dynamic cycling
was performed by successive compression/expansion of the drop. This is normally done
through 20 cycles with the required periodicity (1, 3, or 9 seconds per cycle) and the
required compression ratio (percentage of surface area reduction) of 10%, 20%, or 30%.
During every experimental run, images were acquired, stored and later analyzed with
ADSA. All experiments were reproduced at least in triplicate. A typical experimental
result for 2.0 mg/ml BLES at 100% relative humidity, 20% compression, 3 seconds per
cycle and 37C is shown in figure 3.1 showing an example of ADSA outputs as a function
of cycling time. It can be seen that cycles repeat very well. Using the change of surface
tension and area with time during dynamic cycling, dynamic parameters are calculated
using the compression-relaxation model (CRM).
3.5.2 Results
The dynamic cycling for a specific concentration is performed by successive compression/-
expansion of the drop with the prescribed periodicity and the preselected compression
ratio (percentage of surface area reduction). In this section, 36 experiments are reported:
4 concentrations (2, 8, 15, or 27 mg/ml), 3 periodicities (1, 3, or 9 seconds per cycle),
and 3 compression ratios (10%, 20%, or 30%); each experimental run contains 20 cycles
and is repeated at least 3 times. In total, 108 runs were performed.
During every experimental run, images are acquired continuously at a rate of 30
images per second (for periodicity of 1 second per cycle) or 20 images per second (for
periodicities of 3 and 9 seconds per cycle). A typical run contains approximately 600
images (20 cycles, 1 second per cycle and 30 images per second), 1200 images (20 cycles, 3
second per cycle and 20 images per second) or 3600 images (20 cycles, 9 second per cycle
and 20 images per second). Thus, results reported here rely on almost 200,000 individ-
ual surface tension measurements, involving the analysis of that number of drop images.
Chapter 3. Dynamic Surface Tension Model 107
These images are stored and later analyzed with ADSA. For every image, ADSA calcu-
lates the surface tension, the surface area, and the drop volume as shown in figure 3.1.
The reproducibility between cycles is apparent from figure 3.1.
To deal with such a huge number of data, an overall upgrade of the numerical work
was performed to facilitate the efficient operation and post-analysis functions. After
running all the acquired images through ADSA, an output file was created; this file
needs some careful data analysis. The objectives of the data analysis are editing the
ADSA output file, plotting the change of the surface tension, area and volume with time,
and some preliminary analysis. This includes separating the compression and expansion
parts of every cycle, plotting the change in the surface tension with the relative area
for every half cycle and calculating minimum surface tension values in the cycling part.
These tasks were so far performed manually consuming more than a complete day to
produce the required plots and calculations for a cycling experiment. A Matlab code was
developed to perform the data analysis and plotting for ADSA output files rapidly and
efficiently. The code was designed to process more than one ADSA output file (batch
mode for processing different experimental runs, so all input data are preselected by the
user through script parameters). The outputs from this code are two plots that are saved
in the same folder of the ADSA output files, and an output structure that contains:
the alphabetic run indicator, compression ratio, minimum surface tension mean value,
minimum surface tension standard deviation, maximum surface tension mean value and
maximum surface tension standard for the specified run. The code was also extended to
include the calculated dynamic parameters from CRM.
The minimum value for surface tension for every cycle can be obtained by inspection
from the output of ADSA. A typical run includes 20 values for the minimum surface
tension. The reported minimum surface tension value is the average value within a par-
ticular run. For a specific condition (concentration, compression ratio, and compression
rate), at least three runs were performed. From every run, an average minimum surface
Chapter 3. Dynamic Surface Tension Model 108
0 5 10 15 20 25 300
5
10
15
20
BLES Concentration (mg/ml)
γm
in (
mJ/m
2)
10% compression
1 s/cyc
3 s/cyc
9 s/cyc
(a)
0 5 10 15 20 25 300
5
10
15
20
BLES Concentration (mg/ml)
γm
in (
mJ/m
2)
20% compression
(b)
0 5 10 15 20 25 300
5
10
15
20
BLES Concentration (mg/ml)
γm
in (
mJ/m
2)
30% compression
(c)
Figure 3.8: The minimum surface tension, γmin, measured for four different concentra-tions of BLES (2, 8, 15, 27 mg/ml) at different compression ratios (10%, 20%, 30%)and different cycling conditions (1, 3, 9 s/cycle). The error bars indicate the standarddeviation between runs repeated under the same conditions.
Chapter 3. Dynamic Surface Tension Model 109
tension is reported in figure 3.8.
Figure 3.8 shows the average minimum surface tension, γmin, measured for four dif-
ferent concentrations of BLES (2, 8, 15, 27 mg/ml) at different compression ratios (10%,
20%, 30%) and different cycling conditions (1, 3, 9 s/cycle). The standard deviation
indicated by the error bars shows good reproducibility between runs repeated under the
same conditions.
At higher concentration, the minimum surface tension is sensitive to changes in pe-
riodicity only at small compression ratios. At 10% compression, the minimum surface
tension decreases with the increase of concentration and with the increase with cycling
rate. For 20% compression and above, the minimum surface tension decreases in the
range of 2 mg/ml to 8 mg/ml but stays essentially constant at higher concentrations.
However, for concentrations above 8 mg/ml, there is no significant change in minimum
surface tension. In clinical practice, high concentration BLES of 27 mg/ml is used [1, 178].
and it is believed to be diluted after delivery to lungs to a concentration close to 8 mg/ml
[179, 193–199].
Generally, increasing the extent of compression produces lower minimum surface ten-
sion values. This agrees with previous investigations of BLES at 0.5 and 5 mg/ml and
cycling speeds of 3 and 10 seconds per cycle [25]. At high concentrations, the speed of
cycling does not have a significant effect on the minimum surface tension. The decrease
of minimum surface tension with the increase of the speed of compression at low concen-
trations agrees with previous investigations in the literature of BLES 0.5 mg/ml and 20%
compression [25]. This dependence of minimum surface tension on speed is probably due
to the fact that the speed of compression seems to be fast enough to prevent, in part,
the hydration of the surfactant film [25].
The surface tension measurements of BLES at these high concentrations are now pos-
sible due to the features of the ADSA–CSD methodology as mentioned before. It is also
important to note that ADSA–CSD studies can be performed readily at high rates of
Chapter 3. Dynamic Surface Tension Model 110
compression (to mimic human breathing at 3 seconds per cycle). In this study, higher
rates are also reported (up to 1 second per cycle or 1 Hz). Even higher rates of com-
pression will become achievable after hardware upgrades for both the liquid control and
image acquisition systems. It is noted that flow conditions and viscosity may introduce
errors in the measurement of surface tension via drop/shape methods when the cycling
frequencies are 10 Hz or larger [27, 200–202]. Therefore, an upper limit of 10 Hz exists
for the high frequency studies.
Using the output of ADSA for every experimental run, the change of surface tension
and surface area with time are analyzed using CRM following the procedure explained
in section 3.4.2. As mentioned earlier, a typical experimental run consists of 20 cycles.
From every cycle, the four dynamic parameters of CRM are evaluated, i.e. elasticity of
compression (εc), elasticity of expansion (εe), desorption/relaxation coefficient (kr), and
adsorption coefficient (ka). An indication of the goodness of fit is also calculated for
every cycle. The goodness of fit is defined below in equation 3.18, indicating the relative
error between the predicted surface tension response from CRM and the experimentally
measured surface tension response. Smaller values for the goodness of fit indicate a better
fit.
Figures 3.9 and 3.10 show sample experimental results of the change in surface ten-
sion with time and with relative surface area during dynamic cycling. Figure 3.9 shows
results of BLES concentrations 2 and 27 mg/ml at 10% compression and periodicities of
1 and 9 seconds per cycle. Figure 3.10 shows corresponding results at 30% compression.
The figures illustrate the quality, or “goodness of fit” of CRM to the experimental data.
In these figures, the symbols represent the measured experimental results and the con-
tinuous lines represent the best-fit CRM curves, the four parameters (εc, εe, kr, and ka)
determined from the experimental points having been used for the calculation of CRM
curves. It can be seen that the calculated curves match the experimental points well.
Generally, results shown in figures 3.9 and 3.10 show that CRM fits almost all exper-
Chapter 3. Dynamic Surface Tension Model 111
0 0.2 0.4 0.6 0.8 10
10
20
30
40
Surf
ace T
ensio
n, γ
(m
J/m
2)
Normalized Time, tn
0.9 0.95 1
Relative Surface Area, Ar
(a) BLES 2 mg/ml − 10% comp − 1 s/cycle
0 0.2 0.4 0.6 0.8 10
10
20
30
40
Surf
ace T
ensio
n, γ
(m
J/m
2)
Normalized Time, tn
0.9 0.95 1
Relative Surface Area, Ar
(b) BLES 2 mg/ml − 10% comp − 9 s/cycle
0 0.2 0.4 0.6 0.8 10
10
20
30
40
Surf
ace T
ensio
n, γ
(m
J/m
2)
Normalized Time, tn
0.9 0.95 1
Relative Surface Area, Ar
(c) BLES 27 mg/ml − 10% comp − 1 s/cycle
0 0.2 0.4 0.6 0.8 10
10
20
30
40
Surf
ace T
ensio
n, γ
(m
J/m
2)
Normalized Time, tn
0.9 0.95 1
Relative Surface Area, Ar
(d) BLES 27 mg/ml − 10% comp − 9 s/cycle
Figure 3.9: The change in surface tension with time and with relative surface area duringdynamic cycling of BLES at 10% compression and different concentrations and period-icities: (a) 2 mg/ml and 1 s/cyc; (b) 2 mg/ml and 9 s/cyc; (c) 27 mg/ml and 1 s/cyc;(d) 27 mg/ml and 9 s/cyc. Symbols show the measured surface tension values and linesshow the values obtained from CRM.
Chapter 3. Dynamic Surface Tension Model 112
0 0.2 0.4 0.6 0.8 10
10
20
30
40S
urf
ace T
en
sio
n, γ
(m
J/m
2)
Normalized Time, tn
0.7 0.75 0.8 0.85 0.9 0.95 1
Relative Surface Area, Ar
(a) BLES 2 mg/ml − 30% comp − 1 s/cycle
0 0.2 0.4 0.6 0.8 10
10
20
30
40
Surf
ace T
ensio
n, γ
(m
J/m
2)
Normalized Time, tn
0.7 0.75 0.8 0.85 0.9 0.95 1
Relative Surface Area, Ar
(b) BLES 2 mg/ml − 30% comp − 9 s/cycle
0 0.2 0.4 0.6 0.8 10
10
20
30
40
Surf
ace T
ensio
n, γ
(m
J/m
2)
Normalized Time, tn
0.7 0.75 0.8 0.85 0.9 0.95 1
Relative Surface Area, Ar
(c) BLES 27 mg/ml − 30% comp − 1 s/cycle
0 0.2 0.4 0.6 0.8 10
10
20
30
40
Surf
ace T
ensio
n, γ
(m
J/m
2)
Normalized Time, tn
0.7 0.75 0.8 0.85 0.9 0.95 1
Relative Surface Area, Ar
(d) BLES 27 mg/ml − 30% comp − 9 s/cycle
Figure 3.10: The change in surface tension with time and with relative surface areaduring dynamic cycling of BLES at 30% compression and different concentrations andperiodicities: (a) 2 mg/ml and 1 s/cyc; (b) 2 mg/ml and 9 s/cyc; (c) 27 mg/ml and 1s/cyc; (d) 27 mg/ml and 9 s/cyc. Symbols show the measured surface tension values andlines show the values obtained from CRM.
Chapter 3. Dynamic Surface Tension Model 113
imental conditions (high and low concentrations, cycling rates and compression ratios)
very well. It can be concluded that the model using only four adjustable parameters is
capable of capturing the main features of the dynamic cycling for different conditions.
However, there is a slight inconsistency between the model and the experimental results
at the high surface tension range in some cases, e.g. figures 3.10(b) and 3.10(d). This
disagreement only appears at the highest compression ratio (30%) and the lowest com-
pression rates (9 seconds per cycle). In these cases, the compression is slow, and hence
the details of the adsorption and elasticity effects are pronounced near the highest sur-
face tension values. As the area is increased, the surface tension increases due to the
film elasticity; however, if the surface tension exceeds the equilibrium value, the surface
tension tends to decrease at the same time due to adsorption to the interface. Since
the compression is slow in these cases, the surface tension shows some fluctuations near
the highest surface tension range. The adsorption in such cases is much quicker than
the speed of compression, causing these fluctuations to be very pronounced. Such low
compression rates are not really relevant clinically; however, this lack of agreement is
considered in more detail below.
The same procedure of extracting the dynamic parameters from every cycle is applied
for all 20 cycles in every experimental run. An effective value for each of the parameters
is formed by the multi-variable optimization for every cycle. 20 values for each of the four
dynamic parameters and the goodness of fit are calculated for every run. From every run,
average values are calculated. For a specific condition (concentration, compression ratio,
and compression rate), at least three runs were performed. For each of the 36 different
experimental conditions considered, the values of the goodness of fit and the four dynamic
parameters reported below in figures 3.11 to 3.15 are the average values of the parameter
plotted on the ordinate across repeated runs and the standard deviation. The average
values of the four dynamic parameters for different conditions are also summarized in
table 3.2.
Chapter 3. Dynamic Surface Tension Model 114
0 5 10 15 20 25 300
0.5
1
1.5
BLES Concentration (mg/ml)
GF
(−
)
10% compression
1 s/cyc
3 s/cyc
9 s/cyc
(a)
0 5 10 15 20 25 300
0.5
1
1.5
BLES Concentration (mg/ml)
GF
(−
)
20% compression
(b)
0 5 10 15 20 25 300
0.5
1
1.5
BLES Concentration (mg/ml)
GF
(−
)
30% compression
(c)
Figure 3.11: The goodness of fit for CRM measured for four different concentrations ofBLES (2, 8, 15, 27 mg/ml) at different compression ratios (10%, 20%, 30%) and differentcycling conditions (1, 3, 9 s/cycle). The error bars indicate the standard deviationbetween runs repeated under the same conditions.
Chapter 3. Dynamic Surface Tension Model 115
The visual judgement of a good fit between experimental points and calculated curves
can be quantified by means of a “goodness of fit”. Figure 3.11 compares the values of the
goodness of fit from CRM calculated at different experimental conditions. The goodness
of fit is defined as:
GF =
√∑∣∣∣∣γe − γmγe
∣∣∣∣2 (3.18)
where γe is the experimentally measured surface tension value and γm is the predicted
surface tension value from CRM. Smaller values for the goodness of fit indicate a better
fit. It turns out that CRM fits almost all experimental conditions (high and low concen-
trations, cycling rates and compression ratios) very well except for a few points at higher
compression ratios. The best results for the goodness of fit are at any concentration
at low compression ratios and high frequency cycling. These conditions are relevant to
clinical practice of high frequency ventilation. The goodness of fit gives an indication of
the conditions where CRM works best. This point of the applicability of the model to
certain conditions is considered in more detail below.
Figures 3.12 and 3.14 show the calculated elasticity of compression, εc, and elasticity
of expansion, εe, obtained at different experimental conditions. Figure 3.13 and 3.15
show the calculated relaxation coefficient, kr, and adsorption coefficient, ka, obtained at
different experimental conditions. A detailed inspection of every parameter will follow
below. In these four figures, the values reported are the average values across repeated
runs and the standard deviation. Asterisks are used in these figures to highlight points
where CRM is less sensitive to one or more of the parameters. The quantification of the
sensitivity of the model is described in section 3.5.3.2.
Chapter 3. Dynamic Surface Tension Model 116
0 5 10 15 20 25 300
50
100
150
200
BLES Concentration (mg/ml)
ε c(m
J/m2 )
10% compression
1 s/cyc3 s/cyc9 s/cyc
****
(a)
0 5 10 15 20 25 300
50
100
150
200
BLES Concentration (mg/ml)ε
c (
mJ/m
2)
20% compression
(b)
0 5 10 15 20 25 300
50
100
150
200
BLES Concentration (mg/ml)
ε c(m
J/m2 )
30% compression
**
*
(c)
Figure 3.12: The elasticity of compression, εc, calculated from CRM for four differentconcentrations of BLES (2, 8, 15, 27 mg/ml) at different compression ratios (10%, 20%,30%) and different cycling conditions (1, 3, 9 s/cycle). The asterisks show the rangeof insensitivity (explained in figure 3.17): * indicate points with range of insensitivitybigger than 0.2, ** indicate points with range of insensitivity bigger than 0.5.
Chapter 3. Dynamic Surface Tension Model 117
0 5 10 15 20 25 300
1
2
3
4
5
BLES Concentration (mg/ml)
k r(s−1 )
10% compression
1 s/cyc3 s/cyc9 s/cyc
**
**
(a)
0 5 10 15 20 25 300
1
2
3
4
5
BLES Concentration (mg/ml)k
r (s
−1)
20% compression
(b)
0 5 10 15 20 25 300
1
2
3
4
5
BLES Concentration (mg/ml)
k r(s−1 )
30% compression
****
(c)
Figure 3.13: The relaxation coefficient, kr, calculated from CRM for four different con-centrations of BLES (2, 8, 15, 27 mg/ml) at different compression ratios (10%, 20%,30%) and different cycling conditions (1, 3, 9 s/cycle). The asterisks show the rangeof insensitivity (explained in figure 3.17): * indicate points with range of insensitivitybigger than 0.2, ** indicate points with range of insensitivity bigger than 0.5.
Chapter 3. Dynamic Surface Tension Model 118
0 5 10 15 20 25 300
50
100
150
200
BLES Concentration (mg/ml)
εe (
mJ/m
2)
10% compression
1 s/cyc
3 s/cyc
9 s/cyc
(a)
0 5 10 15 20 25 300
50
100
150
200
BLES Concentration (mg/ml)ε e
(mJ/
m2 )
20% compression
* * **
(b)
0 5 10 15 20 25 300
50
100
150
200
BLES Concentration (mg/ml)
ε e(m
J/m2 )
30% compression
*
**
*
** **
**
******
(c)
Figure 3.14: The elasticity of expansion, εe, calculated from CRM for four differentconcentrations of BLES (2, 8, 15, 27 mg/ml) at different compression ratios (10%, 20%,30%) and different cycling conditions (1, 3, 9 s/cycle). The asterisks show the rangeof insensitivity (explained in figure 3.17): * indicate points with range of insensitivitybigger than 0.2, ** indicate points with range of insensitivity bigger than 0.5.
Chapter 3. Dynamic Surface Tension Model 119
0 5 10 15 20 25 300
2
4
6
8
10
BLES Concentration (mg/ml)
k a(s−1 )
10% compression
1 s/cyc3 s/cyc9 s/cyc
*
* * *
(a)
0 5 10 15 20 25 300
2
4
6
8
10
BLES Concentration (mg/ml)k a
(s−1 )
20% compression
*
**
*
****
**
**
**
**
**
(b)
0 5 10 15 20 25 300
2
4
6
8
10
BLES Concentration (mg/ml)
k a(s−1 )
30% compression
**
*
**
**
****
**
****
**
(c)
Figure 3.15: The adsorption coefficient, ka, calculated from CRM for four different con-centrations of BLES (2, 8, 15, 27 mg/ml) at different compression ratios (10%, 20%,30%) and different cycling conditions (1, 3, 9 s/cycle). The asterisks show the rangeof insensitivity (explained in figure 3.17): * indicate points with range of insensitivitybigger than 0.2, ** indicate points with range of insensitivity bigger than 0.5.
Chapter 3. Dynamic Surface Tension Model 120
Table 3.2: Summary of the dynamic parameters for BLES at 37C and 100% R.H. atdifferent conditions. Asterisks show range of insensitivity (explained in figure 3.17): *range of insensitivity bigger than 0.2, ** range of insensitivity bigger than 0.5.
Elasticity of Elasticity of Relaxation AdsorptionConc. Comp. Freq. Compression Expansion Coefficient Coefficient
(mg/ml) (%) (s/cyc) εc (mJ/m2) εe (mJ/m2) kr (s−1) ka (s−1)
2
101 145.4±2.8 144.4±3.2 3.36±0.12 2.51±0.273 151.6±3.8 162.7±5.1 0.94±0.12 1.11±0.099 103.3±5.6 131.8±4.2 0.90±0.08 0.93±0.02
201 137.6±1.3 151.8±3.5 1.91±0.19 5.36±0.463 134.3±2.8 134.9±2.7 1.61±0.15 * 2.10±0.199 93.9±23.6 93.5±7.9 0.84±0.10 ** 1.07±0.08
301 102.9±1.4 114.5±2.7 3.18±0.10 4.98±0.303 113.5±2.6 106.9±2.4 1.84±0.05 2.07±0.329 126.6±3.8 106.5±4.2 1.31±0.03 * 1.02±0.07
8
101 158.6±2.6 158.8±1.9 1.08±0.07 * 1.21±0.033 151.3±5.5 147.4±5.6 2.33±0.17 0.99±0.059 ** 147.6±6.9 153.3±3.8 ** 2.56±0.08 * 0.63±0.04
201 129.8±1.8 141.4±1.9 1.96±0.22 * 4.74±0.333 112.3±4.8 * 160.9±6.8 1.49±0.04 ** 4.11±0.339 144.8±2.3 * 132.6±3.4 0.55±0.02 ** 1.31±0.04
301 109.6±1.0 * 117.8±4.0 2.10±0.19 ** 9.43±0.743 136.6±3.0 ** 96.3±4.5 2.42±0.28 ** 3.84±0.249 148.7±3.6 ** 95.1±1.3 1.03±0.05 ** 1.38±0.03
15
101 155.4±2.3 168.0±3.8 2.45±0.34 2.45±0.373 167.9±8.1 182.1±5.7 3.75±0.61 2.66±0.139 ** 115.0±3.3 165.6±3.7 ** 2.03±0.07 * 0.91±0.03
201 130.4±2.3 141.4±2.0 1.61±0.27 ** 6.02±1.033 130.7±1.6 144.6±2.3 0.36±0.04 ** 1.77±0.549 135.4±2.4 * 126.5±1.7 0.34±0.01 ** 1.46±0.07
301 * 116.2±4.8 ** 98.9±4.8 ** 2.87±0.32 ** 6.95±0.533 114.8±3.7 * 152.2±8.6 1.25±0.08 ** 5.97±0.489 174.3±4.8 ** 58.2±1.9 1.41±0.05 ** 0.94±0.02
27
101 162.8±1.7 177.5±4.4 1.04±0.14 * 1.83±0.543 161.1±3.0 185.1±2.6 0.91±0.12 2.10±0.369 143.9±1.9 163.9±2.2 0.87±0.07 0.75±0.04
201 142.7±2.2 152.6±8.3 1.28±0.20 ** 4.32±1.173 139.4±1.3 * 149.0±3.6 1.09±0.10 ** 3.44±0.109 149.3±1.8 * 128.8±4.2 0.66±0.02 ** 1.27±0.03
301 ** 143.0±17.7 ** 117.8±8.2 ** 3.44±1.59 ** 8.32±1.083 131.5±3.0 ** 113.4±5.8 1.37±0.13 ** 4.31±0.329 153.2±3.3 ** 99.8±4.9 0.87±0.02 ** 1.35±0.14
Chapter 3. Dynamic Surface Tension Model 121
3.5.3 Discussion
3.5.3.1 Dynamic CRM Parameters
The change of the elasticity of compression, εc, with concentration, compression ratio
and compression rate is shown in figure 3.12. It is expected that the elasticity is an
intrinsic property of the lung surfactant preparation that should not change with different
experimental conditions. The elasticity of compression did indeed remain almost the same
with concentrations above 8 mg/ml. There is also no significant change with compression
ratio or compression rate. This agrees with previous studies on BLES with concentrations
0.5 and 2 mg/ml at 20% compression and 3 seconds per cycle [146]. In previous studies of
DPPC spread films at compression ratios between 0.1 and 30% and frequencies 0.02, 0.05
and 0.1 Hz, the dilatational elasticities were found to be almost constant for compression
ratios higher than 1% and for all frequencies studied [203]. Spread films of of n-dodecyl
dimethyl phosphine oxide (DC12PO), at compression ratios in the range of 2% to 10%
and frequencies in the range of 0.015 to 0.57 Hz, did not show significant change in the
dilatational elasticities [166].
Elasticity of compression of BLES was previously evaluated using an elementary pro-
cedure [25]. The surface tension-relative area curve during the compression stage was
fitted to a fourth order polynomial equation, and the first derivative was used to calculate
the dilatational elasticity at the midpoint of the compression. Values obtained from that
method are in the range of 100 mJ/m2; these results are lower, but essentially in the same
order of magnitude as results presented here. However, it is important to note that the
main assumption in that procedure [25] is that the elasticity is the only active parameter
during the film compression and that any effects of relaxation during the compression
process are negligible.
Other studies of DPPC/SP-B, DPPC/SP-C, DPPC/SP-B/SP-C spread films did not
show any dependence of the elasticity on the compression speed [61, 62, 107, 147]. These
Chapter 3. Dynamic Surface Tension Model 122
results support the concept that more flexible film structures are formed in the presence
of surfactant proteins [204]. It can be concluded that the elasticity parameter in CRM
is a property of the lung surfactant preparation. In fact, the addition of an inhibitor,
e.g. serum, albumin, fibrinogen, or excess cholesterol, to lung surfactant preparations is
known to significantly reduce the elasticity of compression of BLES [90].
The change of the relaxation coefficient, kr, with concentration, compression ratio
and compression rate is shown in figure 3.13. There is some variability but not a pro-
nounced dependence on surfactant concentration. There is no significant dependence on
concentrations at all compression ratios or compression rates except for a few points at
the 10% compression. There is a small increase in the relaxation coefficient when the
compression ratio is increased from 20% to 30%. This agrees with published results for
BLES [25]. The relaxation coefficient seems to depend also on the speed of compression;
the higher the frequency, the higher the relaxation rate. This agrees with other studies
of BLES [25] and DC12PO [166].
The change of the elasticity of expansion, εe, with concentration, compression ratio
and compression rate is shown in figure 3.14. Values are very close to the elasticity of
compression. Generally, there is no significant dependence on concentration, compression
ratio or compression rate. However, the sensitivity of the model to changes in εe of
most experiments at 30% compression is low. This leads to less reliable values at these
conditions. More details regarding the sensitivity of the model are given in section 3.5.3.2.
The change of the adsorption coefficient, ka, with concentration, compression ra-
tio and compression rate is shown in figure 3.15. There is some variability but not a
pronounced dependence on surfactant concentration. However, this variability is repro-
ducible as indicated by the error bars in the figure. The adsorption coefficient seems to
increase with the increase of the compression ratio and compression rate. However, the
sensitivity of most points at 20% and 30% compression is very low. This leads to less
reliable ka values at these conditions. More details regarding the sensitivity of the model
Chapter 3. Dynamic Surface Tension Model 123
are given in section 3.5.3.2.
3.5.3.2 Sensitivity of CRM
For some experiments reported here, it is noticed that the fitting of CRM to the ex-
perimental results is not sensitive to one or more of the four dynamic parameters. To
further investigate this point, a detailed sensitivity study was performed. The details
of calculating the “range of insensitivity” for each parameter is given in section 3.5.4.
Briefly, it is attempted to identify for each parameter the range in which CRM is not
highly sensitive. The sensitivity of the model (in terms of the goodness of fit) to changes
in the respective parameter is calculated as the percentage change of the goodness of fit
divided by the percentage change of the parameter. For example, a sensitivity of 1 means
that a relative change of, say 20%, of a specific parameter yields a relative change of the
same amount, 20% in this case, of the goodness of fit. A sensitivity cut-off level of 1 is
chosen to differentiate between sensitive and insensitive values. The range of percentage
change of the parameter in which the sensitivity of the model is below 1 is called here
the “range of insensitivity” for a specific parameter. The procedure is applied for each
of the four parameters of CRM and the range of insensitivity is calculated for each in
all experiments. A summary of these calculations is shown in figure 3.16 and also in
table 3.3.
In this figure, the contour lines show the range of insensitivity for each parameter.
Higher values indicate that the model is less sensitive to a wider range for this specific
parameter. It can be seen that there are similarities between the insensitivity contours of
both elasticity of compression and relaxation coefficient and also between both elasticity
of expansion and adsorption coefficient.
From these contours, experimental conditions can be identified where CRM is less
sensitive to certain parameters. Such information can be very useful as recommendations
can be made to what experimental conditions are to be used if there is an interest in
Chapter 3. Dynamic Surface Tension Model 124
5 15 25
10
20
30
0.010.05
0.2
0.5
Co
mp
resssio
n R
atio
(%
)
Concentration (mg/ml)
1 s/cycle
5 15 25
0.01
0.05
Concentration (mg/ml)
3 s/cycle
(a) Elasticity of Compression, εc (mJ/m
2)
5 15 25
0.01
0.0
1
0.05
0.0
5 0.2
0.5
1
Concentration (mg/ml)
9 s/cycle
5 15 25
10
20
30
0.01
0.0
1
0.05
0.20.5
1
1.5
Co
mp
resssio
n R
atio
(%
)
Concentration (mg/ml)
1 s/cycle
5 15 25
0.01
0.05
0.0
5
0.2
0.2
0.5
0.5
1
Concentration (mg/ml)
3 s/cycle
(b) Elasticity of Expansion, εe (mJ/m
2)
5 15 25
0.2
0.2
0.5
1
1.5
Concentration (mg/ml)
9 s/cycle
5 15 25
10
20
30
0.010.05
0.2
0.5
Co
mp
resssio
n R
atio
(%
)
Concentration (mg/ml)
1 s/cycle
5 15 25
0.01
0.01
0.05
Concentration (mg/ml)
3 s/cycle
(c) Relaxation Coefficient, kr (s
−1)
5 15 25
0.010.05
0.20.5
1
Concentration (mg/ml)
9 s/cycle
5 15 25
10
20
30
0.0
5
0.2
0.2 0.5
1
1.5
Co
mp
resssio
n R
atio
(%
)
Concentration (mg/ml)
1 s/cycle
5 15 25
0.05
0.2 0.5
1
Concentration (mg/ml)
3 s/cycle
(d) Adsorption Coefficient, ka (s
−1)
5 15 25
0.5
1
1
Concentration (mg/ml)
9 s/cycle
Figure 3.16: Contour lines of the range of insensitivity for each of the parameters: (a)εc, (b) εe, (c) kr, (d) ka, at different periodicities, compression ratios and concentrations.The calculation of the range of insensitivity is explained in figure 3.17.
Chapter 3. Dynamic Surface Tension Model 125
Table 3.3: Summary of the range of insensitivity of CRM to each of the dynamic param-eter (explained in figure 3.17) for BLES at 37C and 100% R.H. at different conditions.
Elasticity of Elasticity of Relaxation AdsorptionConc. Comp. Freq. Compression Expansion Coefficient Coefficient
(mg/ml) (%) (s/cyc) εc (mJ/m2) εe (mJ/m2) kr (s−1) ka (s−1)
2
101 0.00 0.00 0.00 0.003 0.00 0.00 0.00 0.019 0.02 0.01 0.09 0.09
201 0.00 0.00 0.00 0.003 0.00 0.03 0.00 0.219 0.01 0.17 0.02 0.57
301 0.00 0.00 0.00 0.123 0.00 0.02 0.00 0.119 0.00 0.12 0.00 0.21
8
101 0.00 0.00 0.00 0.273 0.00 0.00 0.00 0.009 0.67 0.07 1.21 0.22
201 0.00 0.00 0.00 0.213 0.00 0.20 0.02 0.739 0.00 0.23 0.00 1.10
301 0.00 0.40 0.00 1.303 0.00 1.38 0.00 1.289 0.00 1.41 0.00 1.25
15
101 0.00 0.00 0.00 0.023 0.08 0.00 0.09 0.089 1.18 0.17 1.23 0.28
201 0.00 0.02 0.01 0.623 0.00 0.03 0.06 1.279 0.00 0.22 0.00 1.20
301 0.46 1.47 0.53 1.553 0.00 0.32 0.00 1.139 0.00 1.30 0.00 1.30
27
101 0.00 0.00 0.00 0.243 0.00 0.00 0.00 0.099 0.00 0.00 0.07 0.17
201 0.00 0.08 0.08 1.333 0.00 0.24 0.00 1.119 0.00 0.21 0.00 1.06
301 0.69 1.62 0.99 1.643 0.00 0.81 0.01 1.319 0.00 1.79 0.01 1.42
Chapter 3. Dynamic Surface Tension Model 126
a specific parameter. For example, the elasticity of compression, εc, and the relaxation
coefficient, kr, are less sensitive at high concentrations, high compression ratios, and high
cycling rates. The fit is also less sensitive to εc and kr at low compression ratios and
low cycling rates; at these conditions the parameters εe and ka are more sensitive. The
elasticity of expansion, εe, and the adsorption coefficient, ka, are less sensitive at high
concentrations and high compression ratios.
It is important to note that CRM is generally sensitive to all the parameters at any
concentration at physiological conditions of 20% compression and 3 seconds per cycle.
The model is also sensitive to all the parameters at any concentration at low compression
ratios and high frequency cycling. These conditions are relevant to conventional and high
frequency ventilation.
Generally, the effect of the concentration on the four dynamic parameters is minimal
beyond 8 mg/ml. This is important for clinical therapeutic practice where it is believed
that increased concentration will improve the surfactant properties. As shown here,
increasing surfactant concentration does improve the dynamic properties only in the low
concentration range below 8 mg/ml. Although the dynamic properties of BLES-only
preparations do not improve significantly with increasing BLES concentration, the same
might not be true in the presence of inhibitors. For example, it was shown before [146]
that when albumin is added to BLES, the elasticity value is markedly decreased indicating
surfactant inactivation or inhibition. Using ADSA–CSD in combination with CRM, the
effect of relevant additives and/or inhibitors could be evaluated at appropriate conditions
of concentration, compression rate and compression ratio.
3.5.4 Range of Insensitivity
This section explains the calculations of the range of insensitivity for each parameter of
CRM. First, the change of goodness of fit (defined in equation 3.18) is recalculated after
changing the respective parameter around the calculated value while keeping the rest
Chapter 3. Dynamic Surface Tension Model 127
0 20 40 60 80 100 120 140 160 180 2000
1
2
Go
od
ne
ss o
f F
it,
GF
(−
)
Elasticity of Compression, εc (mJ/m
2)
(a)
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 20
1
2
Go
od
ne
ss o
f F
it,
GF
(−
)
Adsorption Coefficient, ka (s
−1)
(b)
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 20
2
4
6
range ofinsensitivity
Percentage change in εc, ∆ε
c/ε
c
Sensitiv
ity o
f ε
c, (∆
GF
/GF
)/(∆
εc/ε
c)
(c)
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 20
2
4
6
range ofinsensitivity
Percentage change in ka, ∆k
a/k
a
Se
nsitiv
ity o
f k
a,
(∆G
F/G
F)/
(∆k
a/k
a)
(d)
Figure 3.17: Calculation of the range of insensitivity for the elasticity of the compressionεc and the adsorption coefficient ka of one cycle for 2.0 mg/ml BLES at 20% compressionratio with a periodicity of 9 seconds per cycle: (a) The change of goodness of fit withthe change of εc around the calculated value while keeping the rest of the parameters(εe, kr, ka) constant. (b) The change of goodness of fit with the change of ka aroundthe calculated value while keeping the rest of the parameters (εc, εe, kr) constant. (c)The relative change of sensitivity with the relative change of εc. (d) The relative changeof sensitivity with the relative change of ka. The range of insensitivity is the range ofpercentage change of the parameter in which the sensitivity is below 1.
Chapter 3. Dynamic Surface Tension Model 128
of the parameters constant. Two examples are shown in figures 3.17(a) and 3.17(b) for
one of the cycles of 2.0 mg/ml BLES at 20% compression ratio with a periodicity of 9
seconds per cycle. In figure 3.17(a), the change of goodness of fit is plotted versus the
change of one of the parameters (εc) around the calculated value while keeping the rest
of the parameters (εe, kr, ka) constant. In figure 3.17(b), the change of goodness of fit
is plotted versus the change of one of the parameters (ka) around the calculated value
while keeping the rest of the parameters (εc, εe, kr) constant.
The second step is to calculate the sensitivity of the model (in terms of the goodness
of fit) to changes in the respective parameter by calculating the percentage change of the
goodness of fit divided by the percentage change of the parameter. For example, for the
case of the adsorption coefficient, ka, the sensitivity of the model with respect to this
parameter is
S =
∣∣∣∣∆GF/GF∆ka/ka
∣∣∣∣ (3.19)
Samples are shown in figures 3.17(c) and 3.17(d) for the same experiment. In these
figure, the relative change of sensitivity is plotted versus the relative change of one of the
parameters (εc or ka).
The third step is to define the “range of insensitivity” based on the sensitivity cal-
culations. A cut-off level of 1 is chosen to differentiate between sensitive and insensitive
values. If the change in the respective parameter yields a sensitivity more than 1, the
model is considered sensitive to this parameter at this change level, and vice versa. Hence,
the range of insensitivity for this parameter is defined as the range of percentage change
of the parameter in which the sensitivity of the model is below 1.
In our example of figure 3.17(c), CRM is found to be insensitive to small changes in
this parameter, εc, in the range of roughly 98% to 102% of the calculated value for εc.
This means that the range of insensitivity of the model to the parameter εc is found to
be 0.04 in this experiment as shown in the figure. In figure 3.17(d), CRM is found to be
insensitive to small changes in this parameter, ka, in the range of roughly 82% to 133%
Chapter 3. Dynamic Surface Tension Model 129
of the calculated value for ka. This means that the range of insensitivity of the model to
the parameter ka is found to be 0.51 in this experiment as shown in the figure.
The same procedure is applied for each of the four parameters of CRM and the
range of insensitivity is calculated for each in all the experiments. A summary of these
calculations is shown in figure 3.16 and also in table 3.3.
Chapter 4
Accuracy and Shape Parameter∗
4.1 Introduction
In Chapters 2 and 3 little is known about the accuracy of the surface tension measure-
ments. This is particularly critical in Chapter 3, because the accuracy of the determined
mechanical properties of lung surfactant films depends on the accuracy of the measured
response of surface tension to changes in surface area. Therefore, high accuracy surface
tension measurements are necessary in such studies.
As indicated in Chapter 1, drop shape techniques for surface tension measurement
are based on the shape of a pendant drop, sessile drop or captive bubble. Generally, the
shape of the drop/bubble depends on the balance between gravity and surface tension.
When gravitational and surface tension effects are comparable, then, in principle, one can
determine the surface tension from an analysis of the shape of the drop/bubble. Whenever
the surface tension effect is much larger than the gravitational effect, the shape tends to
become close to spherical in the case of pendant and sessile drops/bubbles. Theoretically,
each drop shape corresponds to a certain surface tension value. For well deformed shapes,
a slight change in surface tension causes a significant change in the shape. However, for
∗Portions of this chapter were previously published in Refs. [205, 206], reproduced with permission.
130
Chapter 4. Accuracy and Shape Parameter 131
0 10 20 30 40 50 60 70 8017
18
19
20
21
22
γ (
mJ/m
2)
V (µl)
Figure 4.1: The change of surface tension with drop volume of a constrained sessile dropof OMCTS formed on top of a holder with outer diameter of 7.86 mm.
nearly spherical drop/bubble shapes, a significant change in surface tension causes only
a slight change in the shape. Thus, the sensitivity of surface tension to changes in drop
shape is low.
Despite the general success and flexibility of ADSA, it is well-documented that, as the
drop/bubble shape becomes close to spherical, the performance of all drop shape tech-
niques, including ADSA, deteriorates [64]. The same problem has also been observed by
other researchers who are using numerical optimizations for the measurement of interfa-
cial tensions [207–209]. In their study, the surface tension values calculated from large
pendant drops of pure liquids are calculated consistently and accurately. However, as the
volume of the drop decreases, the surface tension value deviates from the true value.
An illustration of such limitation is shown in figure 4.1, where the surface tension
of a pure liquid, OMCTS, is measured in the ADSA–CSD setup while changing the
drop volume. The symbols represent the actual measurement, while the straight line
represents the literature value. It is clear that the surface tension is correct for the
Chapter 4. Accuracy and Shape Parameter 132
large drops. However, as the drop volume decreases, the surface tension deviates from
the true value. The reason for choosing pure liquids for accuracy studies is that there
exists a unique constant value of surface tension for each liquid for all drop sizes; any
experimental deviation from that value is attributed to errors or limitations of ADSA. Of
course, this becomes more complicated in the case of surfactants, where surface tension
changes surface compression and compression rate as shown in Chapters 2 and 3. In
fact, two BLES constrained sessile drops of the same concentration and same volume
might have very different surface tension values. A serious implication would be clear
in case slopes of the surface tension with drop size (specifically drop surface area) are
to be calculated. Such slopes are required, for example, to calculate the elasticity of
lung surfactant films as shown in Chapter 2 (equation 2.5). A systematic error in the
calculations of these slopes should be expected if the surface tension calculated from
ADSA deviates systematically (with positive or negative slopes) from the correct value as
shown in figure 4.1. Therefore, an independent measure of the quality of the measurement
is needed to evaluate the validity of ADSA results using pure liquids. In the literature,
there are two different approaches for this problem.
The first approach to quantify the accuracy of the measurements uses a quantita-
tive criterion called “shape parameter” that was introduced to quantify the meaning of
“well-deformed” drops and “close-to-spherical” drops [64]. The shape parameter is a
dimensionless parameter that expresses quantitatively the difference in shape between a
given experimental drop and a spherical shape. A simple approach was used to calculate
the difference between the projected area of the drop and a circle with radius R0 (the
radius of curvature at the apex of the drop). The mathematical formulation of the shape
parameter was
Ps =
∣∣∣∣2π∫0
rθ∫0
rdrdθ − πR20
∣∣∣∣2π∫0
rθ∫0
rdrdθ
(4.1)
Chapter 4. Accuracy and Shape Parameter 133
Experimentally, the volume of a sessile/pendant drop of pure liquids with known surface
tension values was changed from the smallest drop size for which the numerical scheme
converges to the largest drop size that a constellation can hold. The surface tension
values obtained from ADSA for these drop sizes were compared with the literature value
and a shape parameter value was calculated for each drop image. Typically, the error in
the surface tension measurement is presented as a function of the shape parameter. A
critical shape parameter was defined based on the minimum value of the shape parameter
that guarantees an error of less than, say, ±0.1 mJ/m2, or any other relative error. The
critical shape parameter determines the range of applicability of ADSA. In other words,
for a drop shape with a shape parameter larger than the critical value, ADSA performs
within an accuracy of ±0.1 mJ/m2 [64]. It was found that the critical shape parameter
does not depend on Bond number (defined as ∆ρgR20/γ using R0 as the characteristic
length scale in that study) [64].
The second approach reported is a sensitivity test of pendant drops and liquid bridges
to measure the interfacial tension [210]. The sensitivity was defined as the change in the
projected area of the drop profile due to a small change (5%) in the surface tension.
This analysis was applied to synthetic images with pre-defined surface tension values. It
was found that the sensitivity of both pendant drops and liquid bridges decreases as the
Bond number decreases [210]. The Bond number here was defined using the diameter
of the needle used for the pendant drop case or the diameter of the circular holder used
for the liquid bridge case as the characteristic length scale. It has to be noted that such
sensitivity results will depend on the size of the drop. In the same study, a different
approach was used when dealing with experimental images. Briefly, the images of liquid
drops of known surface tension values were analyzed using a drop shape technique called
TIFA (theoretical image fitting analysis) that fits whole drop images rather than drop
profile lines [211, 212]. The surface tension results were compared to the known value
for the liquid. A minimum volume was defined based on the minimum value of the drop
Chapter 4. Accuracy and Shape Parameter 134
volume that guarantees an error of less than 1% in the surface tension measurement.
The minimum volume was found to increase for both configurations when the Bond
number decreases [210]. One concern with using this approach is that the output of
the drop shape technique for surface tension measurement is required to evaluate the
quality of the results obtained. The shape parameter, on the other hand, predicts the
appropriateness of the subsequent analysis solely based on the acquired drop image.
The idea of comparing a drop shape to a circle (a two dimensional projection of
a sphere) used earlier [64] is due to the fact that the shape of sessile/pendant drops
or bubbles will be spherical at zero gravity (Bond number of zero). Of course, the
sensitivity of any drop shape technique near the zero gravity shape is very limited and
approaches zero. Hence comparing the shape of a drop/bubble with the zero gravity
shape is an appropriate strategy to evaluate the quality of shape techniques. Therefore,
the approach of using a shape parameter [64], i.e. comparing to the case where the
sensitivity is effectively zero, is often preferable to using a local sensitivity measure [210]
that will change with the drop size. The shape parameter approach is also useful in
predicting, a priori, the quality of the measurement, i.e. without using the output of the
drop shape technique, and hence can be used to design surface tension setups.
For example, when studying lung surfactants using the ADSA–CSD constellation,
the surface tension depends both on time and rate of liquid surface area change. Surface
area change in turn depends on drop volume change. There are no pure liquids available
to mimic the surface tension response of such systems, e.g. near 1 mJ/m2, i.e. a range
critical in lung surfactant studies. The shape parameter approach can be used as an
experimental guideline to design surface tension setups; if the range of surface tensions
to be studied is known, a suitable holder size for the ADSA–CSD setup can be readily
selected.
In this chapter, a modified approach to the shape parameter is proposed. The previ-
ous definition of shape parameter is based on extracting the value of R0 (the radius of
Chapter 4. Accuracy and Shape Parameter 135
Figure 4.2: Schematic of projected areas of a pendant drop profile and a circle.
curvature at the apex of the drop) from the experimental images and comparing projected
area of the drop to the circle of that radius R0. There are four points that suggest a re-
working of this approach. First, this circle (zero gravity shape) may not be the best fit to
the experimental profile. Better overall fits may well be obtainable with circles of slightly
incorrect R0. Second, comparing projected areas might not be the best way to compare
the drop profile and the best fit circle. The drop profile and the circle might have similar
projected areas and completely different shapes as can be seen from figure 4.2. A more
appropriate comparison can be achieved by minimizing the normal distances between
the profile and the circle. Third, when dealing with theoretical profiles generated by the
Laplace equation for a given value of surface tension and curvature at the drop/bubble
apex, experience has shown that the shape parameter values are usually overestimated
specially for very small drop volumes. The fourth point is that one would expect that the
shape parameter should depend on the Bond number (dimensionless number expressing
the ratio of body forces to surface tension forces) contrary to previous findings [64] but
in agreement with the above sensitivity test [210]. It is suspected that using R0 as the
length scale in the previous shape parameter study [64] is not appropriate.
To address these concerns a new definition for the shape parameter is proposed here,
in which the experimental drop shape is compared to the circle that fits best all the points
Chapter 4. Accuracy and Shape Parameter 136
of the drop profile. The best fit circle is calculated by minimizing the normal distances
between the drop profile and the circle. Therefore, the drop profile can be compared to
the best fit zero gravity shape (circle in this case); such a circle is not restricted to a
specific radius, R0. Thus the search process for the best fit zero gravity shape will include
all possible variables, in a similar fashion to how ADSA searches for the best fit profile.
This definition will eliminate any dependence on the radius of curvature of the apex. In
this case, the circle can have any radius – not only R0 as previously considered [64] – to
fit all profile points.
As mentioned earlier, the circle (sphere, in three dimensions) is the zero gravity shape
for pendant and sessile drops/bubbles. These zero gravity shapes can be obtained under
different conditions. Examples are very low gravity, very high surface tension values,
and interfaces between fluids with similar densities. However for other configurations,
e.g. liquid bridges or sessile drops formed at the end of a capillary, the shape at zero
gravity is not spherical but will depend on the new contact conditions. These drops at
zero gravity may assume different shapes including cylindrical, nodoidal, unduloidal and
catenoidal. The definition of the shape parameter can be readily extended – in principle
– to these configurations by comparing the drop shape to the best fit zero gravity shape
that is appropriate to a specific configuration.
In this chapter, a simple strategy to construct the shape parameter is applied to
ADSA–CSD in Section 4.2, because of the need for it in Chapters 2 and 3. To under-
stand the relationship between the constrained sessile drop shape and the accuracy of
the surface tension measurement, dimensional analysis is used to describe similarity in
sessile drop shapes and to express the problem using appropriate dimensionless groups.
The proposed shape parameter (in dimensionless form) is found to depend only on two
dimensionless groups: the dimensionless volume (drop volume normalized by the cube of
the holder radius) and the Bond Number (using the holder’s radius as the length scale).
A set of experiments were performed with pure liquids to illustrate the change of
Chapter 4. Accuracy and Shape Parameter 137
the critical shape parameter with the Bond number for the two different sessile/pendant
setups. To obtain a reasonable range of Bond numbers several pure liquids with known
surface tensions and several holder sizes were used as will be shown below. In every
experiment, the drop volume is varied and the error in the surface tension measurement
is presented as a function of the shape parameter following the same procedure as in
literature [64]. A critical shape parameter is defined as the minimum value of the shape
parameter that guarantees an error below a stipulated error limit. It will be shown
below that the critical shape parameter depends only on the Bond number. Based on
the analysis presented here, specific experimental design parameters are recommended
for accurate surface tension measurements using constrained sessile drops.
Since it turned out that the very same analysis is applicable to the pendant drops,
and since the pendant drop constellation is now the most frequently used surface tension
methodology, the analysis for the shape parameter was extended to the pendant drop
situation in Section 4.3. The universality of the results is examined is Section 4.4.
4.2 Drop Shapes and Shape Parameter
4.2.1 Dimensional Analysis of Axisymmetric Drop Shapes
Five, and only five, independent parameters affect the shape of the constrained sessile
drop: the surface tension of the drop, γ, the density difference between the drop and the
surrounding fluid, ∆ρ, the gravitational acceleration, g, the volume of the drop, V , and
the diameter of the holder, Rh. As discussed in Chapter 1, the apparent contact angle
neither affects nor defines the drop shape in this constrained sessile drop setup. The drop
takes a specific shape depending on only those five independent parameters so that the
drop assumes a specific curvature at the apex, b = 1/R0. Hence, it stands to reason that
b =1
R0
= F(Rh, V, γ,∆ρ, g) (4.2)
Chapter 4. Accuracy and Shape Parameter 138
where R0 is the radius of curvature at the apex of the drop. Examples of this relationship
are shown below in Section 4.2.3.
Using standard dimensional analysis techniques [213], the parameters of equation 4.2
can be simplified to a set of dimensionless groups. The standard procedure can be
summarized briefly as follows: First, equation 4.2 can be rewritten as a relation between
parameters in the form
f(b, Rh, V, γ,∆ρ, g) = 0 (4.3)
There are 3 fundamental physical units in this equation: mass m, length l, and time t,
while there are 6 (n) dimensional variables: b [l−1], Rh [l1], V [l3], γ [m1 ·t−2], ∆ρ [m1 ·l−3],
and g [l1 · t−2]. The units of the dimensional quantities are indicated in brackets. The
dimensions of the variables in equation 4.3 can be arranged in the form of the following
dimensional matrix:
b Rh V γ ∆ρ g
m 0 0 0 1 1 0
l −1 1 3 0 −3 1
t 0 0 0 −2 0 −2
(4.4)
The rank of this dimensional matrix is r = 3. Thus, there are only n−r = 3 independent
dimensionless combinations (parameters), denoted π1, π2, π3, and equation 4.3 can be
re-expressed as
f(π1, π2, π3) = 0 (4.5)
At this point, any 3 (=r) of the variables can be selected as “repeating variables”,
that will be repeated in all of the dimensionless parameters. In this case, we choose
Rh (a characteristic length scale), γ, and g as the repeating variables. Although other
choices would result in a different set of dimensionless parameters, other complete sets
can be obtained by combining the ones already computed. Therefore, any choice of the
repeating variables is satisfactory.
Chapter 4. Accuracy and Shape Parameter 139
Each dimensionless parameter is formed by combining the three repeating variables
with one of the remaining variables. For example, let the first dimensional product be
taken as
π1 = b ·Rah · γb · gc (4.6)
The exponents a, b, and c are obtained from the requirement that π1 is dimensionless.
This requires:
m0 · l0 · t0 = (l−1) · (l1)a · (m1 · t−2)b · (l1 · t−2)c (4.7)
By equating the indices, we obtain a = 1, b = 0, c = 0, so that
π1 = b ·R1h · γ0 · g0 = b ·Rh =
Rh
R0
= bd (4.8)
A similar procedure gives:
π2 = V ·Rah · γb · gc =
V
R3h
= Vd (4.9)
π3 = ∆ρ ·Rah · γb · gc =
∆ρgR2h
γ= Bo (4.10)
Therefore the dimensionless representation of equation 4.2 is of the form
bd =Rh
R0
= F(V
R3h
,∆ρgR2
h
γ) = F(Vd, Bo) (4.11)
Equation 4.11 shows that the dimensionless curvature, bd, (the apex curvature normal-
ized by the holder’s radius) of the drop apex depends only on two dimensionless groups.
The first one is the dimensionless volume, Vd, which is the drop volume normalized by
the cube of the holder’s radius. The second one is effectively the Bond Number, Bo,
using the holder’s radius as the length scale for this case. Examples of this relationship
are shown below in Section 4.2.3.
Chapter 4. Accuracy and Shape Parameter 140
4.2.2 Shape Parameter
The proposed definition of the shape parameter for the ADSA–CSD setup is a comparison
of the drop profile with the circle that best fits all the points of the drop profile. Such
definition will eliminate any dependency on the radius of curvature of the apex. In this
case, the circle can have any radius – not only R0 as previously used – to fit all profile
points.
A simple and computationally inexpensive algebraic fit [214] can be used to find the
best fit circle for a number of points (i.e., digitized drop image). Here, the sum of squares
of algebraic distances is minimized
F(a, b, R) =n∑i=1
[(xi − a)2 + (zi − b)2 −R2
]2=
n∑i=1
(ei + Axi +Bzi + C)2 (4.12)
where a is the x-axis location for the circle center point, b is the z-axis location for
the circle center point, R is the circle radius, ei = x2i + z2i , A = −2a, B = −2b, and
C = a2 +b2−R2. Now differentiating F with respect to A, B, C yields a system of linear
equations that can be solved to yield A, B, C, and hence a, b, R can be computed. Other
algebraic fit methods exist to calculate the circle that best fits a list of points [215, 216].
Such methods were also tested in this context, and the results of the calculated shape
parameter yielded the same result.
Once the best fit circle is determined, the normal distances between every point in
the drop profile and the best fit circle can be simply calculated and used to define the
shape parameter as
P =
√√√√ 1
n
n∑i=1
[√(xi − a)2 + (zi − b)2 −R
]2(4.13)
To eliminate the effect of the size of the drop, it is desirable to define the shape pa-
rameter in dimensionless form. A non-dimensional form can be calculated by normalizing
Chapter 4. Accuracy and Shape Parameter 141
the value calculated above with the radius of the best fit circle
Pc =1
R
√√√√ 1
n
n∑i=1
[√(xi − a)2 + (zi − b)2 −R
]2(4.14)
The main conclusion that can be drawn from the previous section is that all drop
shapes depend on only two dimensionless groups, i.e. the dimensionless drop volume
and the Bond number. Since the shape parameter depends only on the drop shape, we
postulate that the shape parameter, in its dimensionless form as defined in equation 4.14,
depends on the same two dimensionless groups, in a similar fashion as equation 4.11
Pc = F(V
R3h
,∆ρgR2
h
γ) = F(Vd, Bo) (4.15)
This means that every drop shape, and it’s associated shape parameter, depends only
on Vd and Bo. Similar shapes will have similar shape parameters. Examples of this
relationship are shown below in Section 4.2.3.
This should not be understood to mean that Pc is equivalent to bd. Pc is the shape
parameter as calculated from the digitized drop profile using equation 4.14. Pc is an
indication of the difference in shape between a given experimental drop profile and the
best fit circle. On the other hand, bd is the dimensionless curvature at the drop apex and
is used here as an indication of the drop shape.
4.2.3 Sample Constrained Sessile Drops
The relationship between the curvature of the drop apex and the independent parameters
affecting the shape of the constrained sessile drop as shown in equation 4.2 is further
illustrated here. Figure 4.3 shows the change of the curvature at the constrained sessile
drop apex with the drop volume for different holder sizes (outside radius of 3 and 6 mm)
and surface tension values (18, 36, and 72 mJ/m2). These results were obtained using
Chapter 4. Accuracy and Shape Parameter 142
0 50 100 150 2000
0.5
1
1.5
2
2.5
3
Volume, V (µl)
Cu
rva
ture
, b
= 1
/R0 (
cm
−1)
Holder Size: Rh = 3 mm
γ = 72 mJ/m2
γ = 36 mJ/m2
γ = 18 mJ/m2
(a)
0 100 200 300 400 500 600 7000
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Volume, V (µl)
Cu
rva
ture
, b
= 1
/R0 (
cm
−1)
Holder Size: Rh = 6 mm
γ = 72 mJ/m2
γ = 36 mJ/m2
γ = 18 mJ/m2
(b)
Figure 4.3: The change of the apex curvature of a constrained sessile drop with surfacetension and drop volume for two different holder sizes: (a) outer radius of 3 mm, andouter radius of (b) 6 mm.
Chapter 4. Accuracy and Shape Parameter 143
the Axisymmetric Liquid Fluid Interface (ALFI) program [9], that predicts the shape of
a drop/bubble through numerical integration of the Laplace equation. The inputs for
the ALFI program are the apex curvature, b, and the capillary constant, c = ∆ρg/γ.
The outputs are the generated drop profile, drop surface area, drop volume, V , and the
contact angle at the end of the drop profile. The end condition for the integration is the
radius of the holder, Rh.
For the same holder size and the same surface tension value, as the drop volume
increases, the curvature at the drop apex increases initially, then decreases when passing
through the 90o contact angle. The same curvature can be reached twice on every curve,
one time for a small drop volume and another for a large drop volume. There exists a
maximum drop volume and a maximum curvature at the drop apex for a given holder
size and surface tension value. The maximum curvature at the drop apex corresponds
to the case of the 90 contact angle. For the same holder size, as the surface tension
increases, the maximum volume and the maximum apex curvature increase. As the
holder size increases, the range of possible drop volumes increases and the range of
possible curvatures at the drop apex decreases.
The dimensional analysis in Section 4.2.1 (equation 4.11) shows that the dimensionless
curvature, bd, of the constrained sessile drop apex depends only on two dimensionless
groups: the dimensionless volume, Vd, and the Bond Number, Bo. Figure 4.4 shows the
predicted change of the dimensionless curvature at the constrained sessile drop apex with
the dimensionless drop volume for different Bond number values (1.22, 2.44, and 4.89).
This figure is considered a general map of the relationship of the three dimensionless
groups for the constrained sessile drop problem. The effect of the holder size is effectively
integrated inside every one of the dimensionless groups.
For the same Bond number value, as the dimensionless drop volume increases, the
dimensionless curvature increases initially, then decreases. The same dimensionless cur-
vature can be obtained twice on every curve, one time for a small dimensionless drop
Chapter 4. Accuracy and Shape Parameter 144
0 1 2 3 4 5 6 7 80
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Dimensionless Volume, V/Rh
3
Dim
en
sio
nle
ss C
urv
atu
re,
Rh/R
0
Bo = 1.22
Bo = 2.44
Bo = 4.89
Figure 4.4: The change of the dimensionless apex curvature of a constrained sessiledrop with Bond number and dimensionless drop volume. The same relationship can beobtained using any holder size.
volume and another for a large dimensionless drop volume. There exists a maximum
dimensionless drop volume and a maximum dimensionless curvature at the drop apex for
a given Bond number value. As the Bond number decreases, the maximum dimensionless
volume and the maximum dimensionless apex curvature increase.
To further investigate similarity in drop shapes, figure 4.5 shows the predicted drop
profile shapes for different conditions. Figure 4.5(a) shows the shape of the constrained
sessile drop generated using ALFI for the case of surface tension of 6 mJ/m2, holder
radius of 3 mm, and apex curvature of 0.51 cm−1. This is equivalent to a Bond number
of 14.67 and a dimensionless curvature of 0.153. For the given curvature at the apex, there
exist two drop volumes that can satisfy these conditions. The dimensionless volumes for
the two drops are 0.412 and 1.598. The large volume corresponds to the case where
the maximum drop diameter is larger than the holder diameter. This case is indicated
with solid symbols in the figure. The other case of small volume is indicated with open
Chapter 4. Accuracy and Shape Parameter 145
−0.5 0 0.5
−0.2
−0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
Rh = 3 mm, γ = 6 mJ/m
2, Bo = 14.67
x (cm)
z (
cm
)
(a)
−1.5 −1 −0.5 0 0.5 1 1.5
−0.5
0
0.5
1
1.5
Rh = 9 mm, γ = 54 mJ/m
2, Bo = 14.67
x (cm)
z (
cm
)
(b)
−0.5 0 0.5
−0.2
−0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
Rh = 3 mm, γ = 216 mJ/m
2, Bo = 0.41
x (cm)
z (
cm
)
(c)
−0.08 −0.06 −0.04 −0.02 0 0.02 0.04 0.06 0.08
−0.04
−0.02
0
0.02
0.04
0.06
0.08
Rh = 0.5 mm, γ = 6 mJ/m
2, Bo = 0.41
x (cm)
z (
cm
)
(d)
Figure 4.5: The shape of a constrained sessile drop for different holder sizes and surfacetension values. Bond number is indicated on top of every figure. The open symbolsindicate a drop profile with a small volume, while the solid symbols symbols that with alarge volume. The lines indicate the best circle that fit the profile.
Chapter 4. Accuracy and Shape Parameter 146
symbols. The continuous lines indicate the best fit circles that are fitted to all the drop
profile points; these are the circles used in the definition of the shape parameter as
shown in equation 4.14. It is clear that the large volume drop is well deformed, but the
small volume can be fitted to a circle and hence will not allow accurate surface tension
measurement with ADSA.
A case of the very same Bond number is considered in figure 4.5(b). In this case, the
drop shape is generated for a surface tension of 54 mJ/m2, holder radius of 9 mm, and
apex curvature of 0.17 cm−1. This is equivalent to the same Bond number of 14.67 and
the same dimensionless curvature of 0.153. For the given curvature at the apex, there
exist two drop volumes that can satisfy these conditions. The dimensionless volumes for
the two drops are exactly the same as before, i.e. 0.412 and 1.598. It is clear that the
drop shapes in figure 4.5(b) are similar to those in figure 4.5(a). The only difference is
the scaling factor. Therefore, it is expected that the shape parameter should be exactly
the same for both cases. Furthermore, ADSA should treat them similarly, i.e. ADSA will
process the well deformed drops (solid symbol) satisfactorily and will not work accurately
with the small volume drops (open symbols).
A higher surface tension case is considered in figure 4.5(c). In this case, the drop
shape is considered for the case of surface tension of 216 mJ/m2, holder radius of 3
mm, and apex curvature of 2.1 cm−1. This is equivalent to a Bond number of 0.41
and a dimensionless curvature of 0.63. For the given curvature at the apex, there exist
two drop volumes that satisfy these conditions. The dimensionless volumes for the two
drops are 0.613 and 8.734. It is clear that both drops can be fitted well to a circle and
hence will not work accurately with ADSA. A similar Bond number case is considered
in figure 4.5(d). In this case, surface tension is 6 mJ/m2, holder radius is 0.5 mm, and
apex curvature is 12.6 cm−1. This is equivalent to the same Bond number of 0.41 and
the same dimensionless curvature of 0.63. The dimensionless volumes for the two drops
are exactly the same as before, i.e. 0.613 and 8.734. It is clear that the drop shapes in
Chapter 4. Accuracy and Shape Parameter 147
0 1 2 3 4 5 6 7 80
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
Dimensionless Volume, V/Rh
3
Sh
ape
Pa
ram
ete
r, P
c
Bo = 1.22
Bo = 2.44
Bo = 4.89
Figure 4.6: The change of the shape parameter of a constrained sessile drop with Bondnumber and dimensionless drop volume. The same relationship can be obtained usingany holder size.
figure 4.5(d) are similar to those in figure 4.5(c). The only difference is the scaling factor.
Following the similarity between equations 4.11 and 4.15, a map similar to that of
figure 4.4 can be readily derived. Figure 4.6 shows the change of the dimensionless shape
parameter with the dimensionless drop volume for different Bond number values (1.22,
2.44, and 4.89). For the same Bond number value, as the dimensionless drop volume
increases, the value for Pc increases indicating more deformed drops. There exists a
maximum value for Pc for a given Bond number value. According to this definition of
the shape parameter, the minimum value for Pc is always zero. This indicates that very
small drops are not appropriate for surface tension measurement even for drops with small
surface tension values on large holders. This figure is considered a general map of the
relationship of the three dimensionless groups for the constrained sessile drop problem.
The same relationship can be obtained using any holder size.
It can be observed that forming the biggest drop possible on a specific holder size is
Chapter 4. Accuracy and Shape Parameter 148
not always sufficient to obtain accurate results from ADSA. This is clear from the large
volume drops at low Bond number values as shown in figures 4.5(c) and 4.5(d). Although
the drop volume is large, the drop shape is still very close to a sphere preventing ADSA
from calculating accurate results. As the Bond number increases, the drop shape becomes
more deformed at large volume. Nevertheless, even at high Bond number values, the drop
volume has to be large enough to produce a well deformed drop.
At this point, it is appropriate to assume that at any Bond number, there is a critical
shape parameter above which the drop shape is considered well deformed. This critical
shape parameter corresponds to a critical volume that the drop has to exceed to be well
deformed. This critical shape parameter is expected to increase with the decrease in
Bond number. To further investigate this point, a set of experiments were performed to
illustrate the change of the critical shape parameter with the Bond number as illustrated
below.
4.2.4 Experimental Details
4.2.4.1 Materials and Methodology
Octamethylcyclotetrasiloxane (OMCTS) [C8H24O4Si4] was purchased from Sigma-Aldrich
Co. (cat. 74811) with purity ≥99.0%, Diethyl phthalate (DEP) [C12H14O4] from Sigma-
Aldrich Co. (cat. 524972) with purity ∼99.5%, and Water from Sigma-Aldrich Co. (cat.
95286). These three liquids were chosen because they have adequate vapour pressure and
cover a wide range of surface tension. OMCTS, DEP, and water have surface tensions of
approximately 18, 36, and 72 mJ/m2, respectively.
A schematic diagram of the experimental setup for ADSA–CSD is shown in Figure 4.7.
Details of the setup and how ADSA works were given before in Chapter 2. Two holder
sizes were used in this study, outer diameters of 6.55 and 7.86 mm. To ensure consistency,
the quality of the acquired images (or more specifically, the number of the pixels of the
Chapter 4. Accuracy and Shape Parameter 149
Diff
user
Light Source
Microscope CCD
Stepping MotorSyringe
Glass Cuvette
Figure 4.7: Schematic diagram of ADSA–CSD experimental setup.
extracted drop profile) was maintained to be as constant as possible for all experiments.
This is done by using different magnifications for different holder sizes. If the same
magnification were used for all holders, the drops on the smaller holders would have fewer
pixels at the interface and the calculated surface tension values would be less accurate.
4.2.4.2 Experimental Procedure
A large drop of the liquid (roughly 80 µl) was formed on the stainless steel pedestal which
is connected to the syringe and the stepper motor. All experiments were performed at
room temperature of 24oC. The constrained sessile drop is covered with a glass cuvette
to isolate the drop from the environment and to prevent contamination. Several images
were acquired and analyzed by ADSA, providing the surface tension of the pure liquid.
The drop volume was then decreased very slowly with the motor-controlled syringe and
then increased again. The rate of liquid withdrawal and addition was fixed at 0.6 µl/s,
for all experiments. During the volume reduction and enlargement, images were acquired
at the rate of two images per second and saved for further analysis.
All acquired images were analyzed by ADSA, providing a value for the surface tension,
γ, the drop surface area, A, the drop volume, V , and the curvature at the apex, b.
Figure 4.8 shows a typical result of ADSA for a constrained sessile drop of OMCTS
formed on top of a holder with outer diameter of 7.86 mm. This is the same experiment
shown previously in figure 4.1. As mentioned earlier, there exists a unique constant
value of surface tension for each liquid, any experimental deviation from that value is
Chapter 4. Accuracy and Shape Parameter 150
17
18
19
20
21
22
γ (
mJ/m
2)
0.4
0.5
0.6
0.7
0.8
A (
cm
2)
0
20
40
60
80
V (
µl)
0 50 100 150 200 250 3000.4
0.6
0.8
1
1.2
1.4
b (
cm
−1)
Time, t (s)
Figure 4.8: The change of surface tension, drop surface area, drop volume, and thecurvature at drop apex with time of a constrained sessile drop of OMCTS formed on topof a holder with outer diameter of 7.86 mm.
Chapter 4. Accuracy and Shape Parameter 151
attributed to errors or limitations of ADSA.
4.2.5 Results and Discussion
The surface tension values of large drops (at the beginning and at the end of every run) are
constant, i.e. they do not change with drop size, and they agree with the literature value.
However, values calculated from small drops are not constant and show irregular trends,
i.e. not just random errors, as shown in figure 4.8. All acquired images were analyzed
to calculate the shape parameter values according to equation 4.14. To calculate the
critical shape parameter for every experiment, the procedure proposed previously [64]
was followed. Briefly, the shape parameter values of every run are plotted against the
calculated surface tension for this run. The critical shape parameter is the value below
which the error in the calculated surface tension value exceeds a specific limit, usually
set at ±0.1 mJ/m2.
Figure 4.9 shows the error in surface tension measured by ADSA as a function of
the shape parameter for different drop sizes of OMCTS formed on a holder with outer
diameter of 7.86 mm. This case is equivalent to a Bond number of 7.91. For a maximum
error of ±0.1 mJ/m2, the critical shape parameter is 0.018. It is to be noted that for
large drop sizes (large shape parameter values), the measured surface tension is accurate
and consistent. It is also important to note that the overall error in surface tension for
this case did not exceed ±1 mJ/m2 for almost all drops in this run, except for very small
values of shape parameter below 0.005 where the drop is very small and is essentially a
spherical cap.
It was noted in figure 4.9 that there is a specific trend for the direction (positive or
negative) of the error in the surface tension measurements. Similar trends were also noted
in other cases. Figure 4.10(a) shows the error in surface tension of OMCTS formed on
a holder with outer diameter of 6.55 mm, while figure 4.10(b) shows the error in surface
tension of DEP formed on a holder with outer diameter of 7.86 mm. Both figures show
Chapter 4. Accuracy and Shape Parameter 152
0 0.005 0.01 0.015 0.02 0.025 0.03−1
−0.5
0
0.5
1
1.5
2
2.5
3
Shape Parameter, Pc
Err
or
in S
urf
ace
Te
nsio
n, ε
(m
J/m
2)
Figure 4.9: The error in surface tension measured by ADSA as a function of the shapeparameter for different drop sizes of a constrained sessile drop of OMCTS formed on topof a holder with outer diameter of 7.86 mm. For a maximum error of ±0.1 mJ/m2, thecritical shape parameter is 0.018.
similar – but not exactly the same – trends to that of figure 4.9. To further clarify this,
a series of synthetic profiles were generated from ALFI at the same conditions (same
surface tension, holder size, drop volumes) and were analyzed with ADSA. It was noted
that the error in this case has no tendency to take positive or negative values, i.e. there
are no systematic deviations from the correct value. Therefore, it can be concluded that
the systematic deviations observed experimentally is not an artifact of the optimization
procedure used. The deviations might be caused by the quality of the images and the
related optical conditions.
The same analysis was applied to experiments using the other two liquids, diethyl
phthalate (DEP) and water, using the two holders with outer diameters of 6.55 and 7.86
mm. All the results are summarized in Table 4.1. The first two columns of the table
list the holder size and the liquid used for every experiment. The third column lists
the presumably correct surface tension values as found in the literature and/or previous
Chapter 4. Accuracy and Shape Parameter 153
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035−1
−0.5
0
0.5
1
1.5
2
2.5
3
Shape Parameter, Pc
Err
or
in S
urf
ace T
ensio
n, ε
(m
J/m
2)
(a)
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04−1
−0.5
0
0.5
1
1.5
2
2.5
3
Shape Parameter, Pc
Err
or
in S
urf
ace
Te
nsio
n, ε
(m
J/m
2)
(b)
Figure 4.10: The error in surface tension measured by ADSA as a function of the shapeparameter for different drop sizes of a constrained sessile drop of: (a) OMCTS formed ontop of a holder with outer diameter of 6.55 mm, and (b) DEP formed on top of a holderwith outer diameter of 7.86 mm. For a maximum error of ±0.1 mJ/m2, the critical shapeparameters are 0.026 and 0.027, respectively.
Chapter 4. Accuracy and Shape Parameter 154
Tab
le4.
1:Sum
mar
yof
resu
lts
ofco
nst
rain
edse
ssile
dro
pex
per
imen
tsusi
ng
diff
eren
tliquid
san
ddiff
eren
thol
der
size
s.
Hol
der
Liq
uid
γ(m
J/m
2)
Bo
γmeasured
(mJ/m
2)
Pcrit
(mm
)[2
17–2
19]
Run
1R
un
2R
un
3R
un
1R
un
2R
un
3
4.20
Wat
er72
.39
0.60
71.0
23±
0.01
170
.992±
0.01
271
.024±
0.01
0-
--
6.55
Wat
er72
.39
1.46
72.3
61±
0.01
572
.321±
0.01
772
.353±
0.01
60.
036
0.03
40.
035
7.86
Wat
er72
.39
2.10
72.3
54±
0.01
772
.313±
0.00
972
.323±
0.02
10.
035
0.03
40.
034
6.55
DE
P36
.95
3.18
36.8
61±
0.01
336
.865±
0.01
336
.878±
0.01
20.
032
0.03
30.
032
7.86
DE
P37
.95
4.58
36.9
36±
0.00
636
.913±
0.00
836
.900±
0.00
60.
029
0.02
70.
029
6.55
OM
CT
S18
.20
5.41
18.1
10±
0.00
918
.137±
0.00
618
.126±
0.00
60.
029
0.02
60.
028
7.86
OM
CT
S18
.20
7.91
18.1
77±
0.00
618
.206±
0.00
418
.189±
0.00
50.
018
0.01
80.
018
Chapter 4. Accuracy and Shape Parameter 155
work [217–219]. The fourth column lists the Bond number for every experiment; a range
of Bond numbers between 0.60 and 7.91 is covered in this study. The following three
columns contain the calculated surface tension value (from the largest drop sizes) and
the accuracy of the calculation; more than 60 drops were used in the calculation of the
average value and the standard deviation. The last three columns contain the measured
critical shape parameter for every experimental run. These values were measured in a
similar procedure used in figure 4.9.
Inspection of table 4.1 shows that for all systems, except the first one listed, there is
excellent agreement from run to run, as well as with the standard values of column 3. It
is also noted that the differences from run to run could be due to ambient temperature
changes of the order of 0.1oC. The consistency from run to run and the accuracy of each
run are generally less for the first system, and the agreement with the standard values is
much poorer, out of the stipulated error range of ±0.1 mJ/m2. A critical shape parameter
does not exist for these cases. The reason of this difference in pattern is discussed in
detail below.
Figure 4.11 shows the change of the critical shape parameter as a function of the Bond
number for all the experiments of this study. In other words, the critical shape parameter
values for every run obtained in a similar fashion to figure 4.9 and summarized in the
last three columns of table 4.1 are used to calculate an average critical shape parameter
value for every Bond number as shown in figure 4.11.
Therefore, for a specific liquid, i.e. a specific surface tension value, and a specific
holder size, there exists a critical shape parameter value below which the error in the
calculated surface tension value exceeds a specific limit. If we use the same liquid with
a larger holder size, the drop shapes become more deformed and the calculated critical
shape parameter value is lower. A similar case exists when a different liquid with a
smaller surface tension value is used with the same holder size. The drop shapes become
more deformed and the calculated critical shape parameter value is lower. A larger holder
Chapter 4. Accuracy and Shape Parameter 156
1 2 3 4 5 6 7 8 9
0.02
0.03
0.04
Bond Number, Bo
Critical S
ha
pe
Pa
ram
ete
r, P
crit
Experiments Results
Best Fit (Linear)
Figure 4.11: The critical shape parameter as a function of the Bond number. Theexperimental points were fitted to a linear function (continuous line).
or a liquid with a smaller surface tension value leads to a larger Bond number (using the
holder’s radius as the length scale).
As expected, figure 4.11 shows that the critical shape parameter decreases with in-
creasing Bond number. The critical shape parameter is equivalent to a critical dimen-
sionless volume (the drop volume normalized by the cube of the holder’s radius) that
the drop has to exceed to allow for accurate measurement of the surface tension. This
can also be understood from figure 4.6 which defines the relationship between the shape
parameter, dimensionless volume and Bond number. The critical shape parameter is
one point on every curve below which the drop shape is indistinguishable from the zero
gravity shape (circle in this case). For high Bond numbers (a low surface tension liquid
and/or a large holder), it is expected that small drop volumes are sufficient to have well
deformed shapes. On the other hand, for low Bond numbers (a high surface tension
liquid and/or a small holder), it is expected that very large drop volumes are needed to
make sure that the drop is well deformed.
Chapter 4. Accuracy and Shape Parameter 157
It is expected that as the Bond number increases, the critical shape parameter will
decrease approaching the value of zero as the Bond number approaches infinity. On
the other hand, as the Bond number decreases and approaches zero, the critical shape
parameter will increase dramatically. However, it was found in this study that critical
shape parameters can not be established for Bond numbers below one, i.e. when the
diameter of holder becomes small and approaches zero. This is due to the fact that the
largest drop that can be formed on a given holder size is not big enough to calculate an
accurate surface tension value.
The relationship between the critical shape parameter and the Bond number can be
considered as a universal guide. Therefore, it was attempted to fit these points in this
range with a straight line as shown in the figure. Of course, on a wider range, it is
expected to have an exponential form. However, using a straight line in the experimental
range here seems sufficient. The best fit line is plotted in figure 4.11 as a continuous line.
Such a relationship is only applicable for a desired error of ±0.1 mJ/m2 in the measured
surface tension. Changing this value for the desired error to a higher or a lower value
would change the location of this curve downwards or upwards, respectively. However,
such change in the stipulated error will not have any effect on the shape of the curve.
The relation between the critical shape parameter and the Bond number should be
universal for this specific constrained sessile drop constellation. To further illustrate this
point, an additional experiment was performed using a fourth liquid, hexadecane (with
surface tension of 27 mJ/m2), and the holder with outer diameter of 7.86 mm. The
Bond number is 4.3 for this experiment, and the critical shape parameter is expected to
be 0.0331 from figure 8. The hexadecane experiment was repeated three times and the
measured critical shape parameter was found to be 0.034±0.0014. This value agrees with
the expected value and fits well with the other points shown in figure 4.11.
The relationship between the critical shape parameter and the Bond number can be
used in the design of a constrained sessile drop experiment for surface tension measure-
Chapter 4. Accuracy and Shape Parameter 158
ment. Through the above relationship, we know that a high Bond number is always
preferable and will facilitate more accurate surface tension measurement. A high Bond
number is equivalent to a small surface tension value and/or a large holder size. If a
rough approximation of the surface tension is known, the right holder size can be easily
selected to produce accurate measurement.
For example, based on the calculations presented here and assuming a density differ-
ence of 1000 kg/m3 and gravity of 9.80 m/s2, designing an ADSA–CSD experiment to
measure surface tension values in the range of 0.5 to 38 mJ/m2 would require a minimum
holder diameter of 4 mm to ensure a maximum error of ±0.1 mJ/m2 throughout. Such
range of surface tension values is of interest in the case of lung surfactant studies, or
compressed soluble or insoluble films. Based on this analysis, results of Chapter 3 can
be confirmed to be within the ±0.1 mJ/m2 error range. It has to be noted that using a
smaller holder for this case will cause a larger error near the high end of this surface ten-
sion range. Generally, it is advisable to make the drop as large as possible to guarantee
accurate results especially near the high end of this range. Another example is designing
an experiment to measure surface tension in the range of 15 to 72 mJ/m2, e.g. common
liquids, which would require a minimum holder diameter of 6 mm. Based on this analysis,
results of Chapter 2 can be confirmed to be within the ±0.1 mJ/m2 error range. These
rough examples serve as a guide for selecting the right holder size to produce accurate
measurement, if a rough approximation of the surface tension is known. It is important
to emphasize that it is the largest surface tension values that dictate a minimum holder
size.
4.3 Pendant Drop Shapes
In the pendant drop setup, a drop is formed at the end of a needle. Alternatively, a
small flat circular holder (inverted pedestal) with a sharp-knife edge is used instead of
Chapter 4. Accuracy and Shape Parameter 159
the capillary [59, 64] as indicated in Chapter 1. The dimensional analysis of Section 4.2.1
developed for the constrained sessile drop case is still applicable to the pendant drop
case because the same five independent parameters affects the shape of both the pendant
drop as well as the constrained sessile drop (Section 4.2.1). The definition of the shape
parameter presented in Section 4.2.2 can be also used for the pendant drop case, since
they both share the same zero gravity shape (sphere). This would not be appropriate,
e.g., in the case of a liquid bridge.
This section follows the same logic as that of Section 4.2. The relationship between
the curvature of the drop apex and the independent parameters affecting the shape as
shown in equation 4.2 is further illustrated here for the pendant drop. The change of the
curvature, b, at the pendant drop apex with the drop volume for different holder sizes
(outside diameter of 3 and 6 mm) and surface tension values (18, 36, and 72 mJ/m2) is
shown in figure 4.12. These results were obtained using ALFI as mentioned before in the
case of the constrained sessile drop.
For a given holder size and given surface tension value, as the drop volume increases,
the curvature at the drop apex increases initially, then decreases in some cases. There
exists a largest drop volume, V , and a largest apex curvature, b, for a given holder size
and a given surface tension value. Drop profiles can not be generated beyond these values.
This will be discussed in more detail below. For a given holder size, when a liquid with
a higher surface tension value is used, the largest volume increases and the largest apex
curvature decreases. For larger holders, the range of possible drop volumes is larger and
the range of possible curvatures at the drop apex is smaller.
The dimensional analysis shown in Section 4.2.1 (equation 4.11) shows that the di-
mensionless curvature, bd, of the pendant/sessile drop apex depends only on two dimen-
sionless groups: the dimensionless volume, Vd, and the Bond Number, Bo. The predicted
change of the dimensionless curvature, bd, of a pendant drop with the dimensionless drop
volume for different Bond number values (1.22, 2.44, and 4.89) is shown in figure 4.13.
Chapter 4. Accuracy and Shape Parameter 160
0 10 20 30 40 500
1
2
3
4
5
6
7
8
9
10
Volume, V (µl)
Cu
rva
ture
, b
= 1
/R0 (
cm
−1)
Holder Size: Rh = 1.5 mm
γ = 72 mJ/m2
γ = 36 mJ/m2
γ = 18 mJ/m2
(a)
0 20 40 60 80 100 1200
1
2
3
4
5
6
7
8
Volume, V (µl)
Curv
atu
re, b =
1/R
0 (
cm
−1)
Holder Size: Rh = 3 mm
γ = 72 mJ/m2
γ = 36 mJ/m2
γ = 18 mJ/m2
(b)
Figure 4.12: The change of the apex curvature of a pendant drop with surface tensionand drop volume for two different holder sizes: (a) outer radius of 1.5 mm; (b) outerradius of 3 mm.
Chapter 4. Accuracy and Shape Parameter 161
0 1 2 3 4 50
0.5
1
1.5
2
2.5
Dimensionless Volume, Vd = V/R
h
3
Dim
en
sio
nle
ss C
urv
atu
re,
bd =
Rh/R
0
Bo = 1.22
Bo = 2.44
Bo = 4.89
Figure 4.13: The change of the dimensionless apex curvature, bd, of a pendant dropwith Bond number, Bo, and dimensionless drop volume, Vd. The Bond number, Bo =∆ρgR2
h/γ, is the ratio of body forces (gravitational) to surface tension forces using theholder’s radius as the length scale.
Chapter 4. Accuracy and Shape Parameter 162
These Bond numbers correspond to the surface tension values of 72, 36, and 18 mJ/m2
using a holder with radius of 3 mm (same cases considered in figure 4.12(b)). Figure 4.13
is considered a general map of the relationship of the three dimensionless groups for the
pendant drop problem (similar to figure 4.4 for the constrained sessile drop case). The
effect of the holder size is integrated inside every one of the dimensionless groups. For
example, a Bond number of 1.22 is equivalent to the two arbitrary chosen cases of: (1)
Rh = 1.5 mm and γ = 18 mJ/m2, shown in figure 4.12(a), and (2) Rh = 3 mm and γ =
72 mJ/m2, shown in figure 4.12(b).
It can be seen from figure 4.13 that there exists a largest dimensionless drop volume,
Vd, and a largest dimensionless curvature, bd, for a given Bond number value. Drop
profiles can not be generated beyond these values. This will be discussed in more detail
below. As the Bond number increases (i.e. a larger holder and/or a liquid with smaller
surface tension value is used), the largest dimensionless volume decreases and the largest
dimensionless apex curvature increases.
Figure 4.14 shows four examples of the predicted drop profile shapes for different
conditions. Figure 4.14(a) shows a pendant drop with γ of 10.0 mJ/m2, Rh of 1.0 mm,
b of 13.2 cm−1, Bo of 1.0, bd of 1.32, and Vd of 5.26. Figure 4.14(b) shows a pendant
drop with γ of 40.0 mJ/m2, Rh of 2.0 mm, and b of 6.6 cm−1, and hence has exactly
the same Bo of 1.0, bd of 1.32, and Vd of 5.26 as figure 4.14(a). The continuous lines
indicate the best fit circle that are fitted to all the drop profile points; this circle is used
in the definition of the shape parameter as shown in equation 4.14. It is apparent that
both drops are well deformed. The only difference is the scaling factor. Therefore, ADSA
should treat them similarly, i.e. ADSA is expected to process these well deformed drops
satisfactorily.
A lower surface tension case is considered in figure 4.14(c). In this case, the pendant
drop shape is considered for the case of γ of 1.0 mJ/m2, Rh of 1.0 mm, b of 25.3 cm−1,
Bo of 10.0, bd of 2.53, and Vd of 0.60. Figure 4.14(d) shows a pendant drop with γ of
Chapter 4. Accuracy and Shape Parameter 163
−0.2 −0.1 0 0.1 0.2−0.3
−0.25
−0.2
−0.15
−0.1
−0.05
0
0.05
0.1
Rh = 1.00 mm, γ = 10 mJ/m
2, Bo = 1.00
x (cm)
z (
cm
)
(a)
−0.5 0 0.5−0.6
−0.5
−0.4
−0.3
−0.2
−0.1
0
0.1
0.2
Rh = 2.00 mm, γ = 40 mJ/m
2, Bo = 1.00
x (cm)
z (
cm
)
(b)
−0.2 −0.1 0 0.1 0.2−0.3
−0.25
−0.2
−0.15
−0.1
−0.05
0
0.05
0.1
Rh = 1.00 mm, γ = 1 mJ/m
2, Bo = 10.00
x (cm)
z (
cm
)
(c)
−0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8
−0.8
−0.6
−0.4
−0.2
0
0.2
Rh = 3.16 mm, γ = 10 mJ/m
2, Bo = 10.00
x (cm)
z (
cm
)
(d)
Figure 4.14: The shape of a pendant drop for different holder sizes and surface tensionvalues. Bond number is indicated on top of every figure. The symbols indicate the dropprofile, and the lines indicate the best circle that fit the profile.
Chapter 4. Accuracy and Shape Parameter 164
0 1 2 3 4 50
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
Dimensionless Volume, Vd = V/R
h
3
Sh
ape
Pa
ram
ete
r, P
c
Bo = 1.22
Bo = 2.44
Bo = 4.89
Figure 4.15: The change of the shape parameter of a pendant drop with Bond numberand dimensionless drop volume.
10.0 mJ/m2, Rh of 3.16 mm, and b of 8.0 cm−1, and hence has exactly the same Bo
of 10.0, bd of 2.53, and Vd of 0.60 as figure 4.14(c). It is clear that both drop can be
fitted to a circle fairly closely and hence will not work accurately with ADSA. It is noted
that constrained sessile drops at high Bond numbers are expected to be well deformed as
shown in figures 4.5(a) and 4.5(b); however, it seems that this is not true for the pendant
drop case as shown in figures 4.14(c) and 4.14(d) where drops at high Bond numbers are
very close to spherical. This point is discussed later in more detail.
Following the similarity between equations 4.11 and 4.15, a map similar – but not
equivalent – to that of figure 4.13 can be readily derived. Figure 4.15 shows the change
of the dimensionless shape parameter with the dimensionless drop volume for different
Bond number values (1.22, 2.44, and 4.89). For the same Bond number value, as the
dimensionless drop volume increases, the value for Pc increases indicating more deformed
drops. There exists a largest dimensionless volume and a largest value for Pc for a given
Bond number value. According to this definition of the shape parameter, the minimum
Chapter 4. Accuracy and Shape Parameter 165
value for Pc is always zero, similar to the constrained sessile drop case. This indicates
that very small drops, with small dimensionless volumes, are not appropriate for surface
tension measurement even for drops with small surface tension values on large holders.
Figure 4.15 is a general map of the relationship of the three dimensionless groups for the
pendant drop problem.
The fact that there exists a largest dimensionless volume for a given Bond number
value is logical and expected: The pendant drop may be successively increased in size by
adding small amounts of liquid until, at some limiting volume, the drop becomes unstable
and a large portion falls away. The growth and the general conditions that lead to the
instability of a pendant drop were studied before by Padday and Pitt [71]. In that study,
the case of the volume-radius limited pendant drop is equivalent to the case considered
here where the volume of the drop is controlled and the holder used has a fixed radius
that is completely wetted.
Padday and Pitt [71] studied the criteria of critical stability of axisymmetric menisci
in a gravity field and presented a useful form of the critical volume data of the volume-
radius limited pendant drop in terms of V/X3 and X/k, where V is the drop volume, X
corresponds to the radius of the holder, and k is a form of the capillary constant defined
as (γ/∆ρg)12 . Therefore, V/X3 is equivalent to the dimensionless volume Vd considered
here, and X/k is basically the square root of the Bond number Bo. The relationship
between the largest (i.e. critical [71] or maximum allowable) dimensionless volume and
the Bond number is recalculated here using ALFI and plotted in figure 4.16(a). Each
point in figure 4.16(a) is equivalent to the end point of the appropriate curve in figure 4.13.
The analysis was repeated for a wide range of Bond numbers ranging from 10−4 to 102.
The points are in excellent agreement with those presented by Padday and Pitt [71].
The maximum dimensionless volume as a function of Bond number is shown in fig-
ure 4.16(a). It is clear that as the Bond number increases, the maximum dimensionless
volume decreases. As an aside, it is important to note that the drop volume falling off
Chapter 4. Accuracy and Shape Parameter 166
10−4
10−2
100
102
100
101
102
103
104
105
Bond Number, Bo
Ma
xim
um
Dim
en
sio
nle
ss V
olu
me
, V
m
(a)
10−4
10−2
100
102
0
20
40
60
80
100
120
Bond Number, Bo
Ap
pa
ren
t C
on
tact
An
gle
, θ (
de
g.)
(b)
Figure 4.16: The change of (a) the maximum dimensionless drop volume, and (b) theapparent contact angle of a pendant drop with Bond number.
Chapter 4. Accuracy and Shape Parameter 167
is always less than the critical volume, first because experimental vibration promotes
instability at an earlier stage of drop formation, and second because a small part of the
drop remains attached to the needle/holder.
The change of the apparent contact angle of the liquid at the holder with Bond
number, at the maximum dimensionless volume, is shown in figure 4.16(b). This figure
indicates that for small Bond number values (i.e., a small holder and/or a liquid with a
large surface tension value), the apparent contact angle is 90 and the liquid is held just
above the narrowest point of the neck of the profile. As the Bond number becomes larger
(i.e., a larger holder and/or a liquid with a smaller surface tension value), the contact
angle at which critical conditions are reached decreases to zero at a Bond number of
10.338. The zero contact angle is the limit for any pendant drop experiment. Further
increase in Bond number (i.e., a larger holder and/or a liquid with a smaller surface
tension value) results in the shape of the critically stable profile taking a configuration
where the holder is not completely wetted. This is the form of a drop as it breaks away
from a large orifice such as that of a bath tap. This case is outside the scope of this
study.
To further understand the different shapes of critically stable profiles, figure 4.17
shows the pendant drop shapes with the maximum dimensionless volume at different
Bond number (0.01, 0.1, 1, and 10). For very small Bond numbers, the apparent contact
angle is around 90 and the maximum dimensionless volume is very large as expected from
figure 4.16. However, the drop shape is very close to spherical. Therefore, it is expected
that ADSA will not work satisfactorily in this range. As the Bond number increases, the
maximum dimensionless volume and the apparent contact angle decrease as expected
from figure 4.16 and the shape of the drop becomes more deformed. Nevertheless, at the
high end of the Bond number range, the apparent contact angle approaches zero and the
drop shape becomes close to spherical again. Thus, it is expected that ADSA will also
not work satisfactorily in this range.
Chapter 4. Accuracy and Shape Parameter 168
−1 −0.5 0 0.5 1
−1.8
−1.6
−1.4
−1.2
−1
−0.8
−0.6
−0.4
−0.2
0
Bo = 0.01, V/Rh
3 = 530.994, P
c = 0.100
x (cm)
z (
cm
)
(a)
−0.6 −0.4 −0.2 0 0.2 0.4 0.6
−0.8
−0.7
−0.6
−0.5
−0.4
−0.3
−0.2
−0.1
0
0.1
Bo = 0.10, V/Rh
3 = 47.179, P
c = 0.134
x (cm)
z (
cm
)
(b)
−0.6 −0.4 −0.2 0 0.2 0.4 0.6
−0.8
−0.7
−0.6
−0.5
−0.4
−0.3
−0.2
−0.1
0
0.1
Bo = 1.00, V/Rh
3 = 5.262, P
c = 0.165
x (cm)
z (
cm
)
(c)
−0.6 −0.4 −0.2 0 0.2 0.4 0.6
−0.8
−0.7
−0.6
−0.5
−0.4
−0.3
−0.2
−0.1
0
0.1
Bo = 10.00, V/Rh
3 = 0.599, P
c = 0.097
x (cm)
z (
cm
)
(d)
Figure 4.17: The shape of a pendant drop with the maximum dimensionless drop volumefor different Bond numbers. The symbols indicate the drop profile, and the lines indicatethe best circle that fit the profile.
Chapter 4. Accuracy and Shape Parameter 169
10−4
10−2
100
102
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
Bond Number, Bo
Sh
ap
e P
ara
me
ter
of
Ma
xim
um
Vo
lum
e,
P c
Figure 4.18: The change of the shape parameter of pendant drops with the maximumdimensionless drop volume with Bond number.
The shape parameter can be calculated using equation 4.14 for the pendant drop
shapes with the maximum dimensionless volume at different Bond numbers. The shape
parameter for the largest stable pendant drop as a function of Bond number is shown in
figure 4.18. The curve ends at Bond number of 10.338 which is equivalent to zero contact
angle. It is clear that the shape parameter is small for very small values as well as large
values of Bond number for reasons discussed above. This shows that for the pendant
drop case very small values as well as large values of Bond number are not recommended
for accurate surface tension measurements.
At this point, it is clear that at any Bond number, there is a critical shape parameter
above which the drop shape is considered well deformed. To further investigate this point,
experiments were performed to illustrate the change of the critical shape parameter with
the Bond number as illustrated below.
Chapter 4. Accuracy and Shape Parameter 170
Diffuser
LightSource
Microscope CCD
SteppingMotor
Syringe
Glass Cuvette
Figure 4.19: Schematic diagram of ADSA–PD experimental setup.
4.3.1 Experimental Details
4.3.1.1 Materials and Methodology
For the pendant drop experiments decamethylcyclopentasiloxane (DMCPS) was used
rather than OMCTS, mainly due to availability. DEP and water were used again, and a
fourth liquid (hexadecane) was added to provide more experiments at a different Bond
number. DMCPS [C10H30O5Si5] was purchased from Sigma-Aldrich Co. (cat. 444278)
with purity 97.0% and hexadecane [C16H34] from Sigma-Aldrich Co. (cat. 296317) with
purity ≥99.0%. These four liquids have reasonably low vapour pressure and cover a broad
range of surface tension. DMCPS, hexadecane, DEP, and water have surface tensions of
approximately 18, 27, 36, and 72 mJ/m2, respectively.
Figure 4.19 shows a schematic diagram of the experimental setup for ADSA–PD. A
flat circular holder is used to suspend the pendant drop. Except for the pendant drop
chamber, the setup is exactly the same as that used for the constrained sessile drop
study in Section 4.2.4.1. Three holder sizes were used in this study with outer diameters
of 3.08, 4.24 and 6.25 mm. Two holders were used for all four liquids. The smallest
holder was used only with one liquid, DMCPS. It is noted that holder sizes different from
those used in the constrained sessile drop study were used here. This is mainly due to
availability of the holders. Although the same holder radius can be used for both pendant
and constrained sessile drop setups, the holder length requirement is different for every
setup. Hence, different sets of holders were used in the two types of experiments.
Chapter 4. Accuracy and Shape Parameter 171
4.3.1.2 Experimental Procedure
All experiments were performed at room temperature of 24oC. It is important to note
that the temperature was controlled within one degree. More accurate temperature
control requires more sophisticated equipment. Since it is not the purpose of this study
to produce literature values of surface tension, such tight control of temperature is not
needed here. The experimental procedure used here was exactly the same as in the
constrained sessile drop case as shown in Section 4.2.4.2.
In total 9 different experiments were performed to cover a wide range of Bond num-
bers. Every experiment was repeated at least three times (three runs) to confirm the
reproducibility; thus, a total of 27 runs were performed in this study. Each run lasted
approximately 200 seconds, implying that approximately 400 images were acquired in
each run. Hence the study is relying on approximately 10,000 individual experimental
surface tension measurements.
4.3.2 Results and Discussion
The results obtained here are qualitatively similar to the ones obtained for the constrained
sessile drop study. Figure 4.20 shows a typical result of ADSA for a pendant drop
of hexadecane formed at the end of a holder with outer diameter of 6.25 mm. The
figure shows a non-random pattern, and hence lack of consistency with the expectation
of a constant surface tension value. The surface tension values of large drops (at the
beginning and at the end of every run) are constant, i.e. they do not change with drop
size, and they agree with the literature value. However, values calculated from small
drops are not constant and show irregular trends, i.e. not just random errors, as shown
in figure 4.20. All acquired images were analyzed and the the shape parameter values
were calculated according to equation 4.14. To calculate the critical shape parameter for
every experiment, the procedure proposed previously [64] was followed. Briefly, the shape
Chapter 4. Accuracy and Shape Parameter 172
Figure 4.20: The change of surface tension, drop surface area, drop volume, and thecurvature at drop apex with time of a pendant drop of hexadecane formed using a holderwith outer diameter of 6.25 mm.
Chapter 4. Accuracy and Shape Parameter 173
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14−0.5
−0.4
−0.3
−0.2
−0.1
0
0.1
0.2
0.3
0.4
0.5
Shape Parameter, Pc
Err
or
in S
urf
ace
Te
nsio
n, ε
(m
J/m
2)
Figure 4.21: The error in surface tension measured by ADSA as a function of the shapeparameter for different drop sizes of a pendant drop of hexadecane formed using a holderwith outer diameter of 6.25 mm. For a maximum error of ±0.1 mJ/m2, the critical shapeparameter is 0.04.
parameter values of every run are plotted against the calculated surface tension for the
run. The critical shape parameter is the value below which the error in the calculated
surface tension value exceeds a specific limit, usually set at ±0.1 mJ/m2.
The error in surface tension measured as a function of the shape parameter is shown
in figure 4.21 for different drop sizes of hexadecane formed with a holder with outer
diameter of 6.25 mm. This case is equivalent to a Bond number of 2.72. For a maximum
error of ±0.1 mJ/m2, the critical shape parameter is 0.04. It is important to note that
this is a conservative value as seen from the figure. It is to be noted that the measured
surface tension is accurate and consistent with an accuracy less than ±0.01 mJ/m2 for
well deformed drops in this run, i.e. for large drop sizes with large shape parameter
values. It is also important to note that the overall error in surface tension for this case
did not exceed ±0.2 mJ/m2 for almost all drops in this run, except for very small values
of shape parameter below 0.01 where the drop is very small and is essentially a spherical
Chapter 4. Accuracy and Shape Parameter 174
cap.
As in the constrained sessile drop case, it was noted in figure 4.21 that there is a
specific trend for the direction (positive or negative) of the error in the surface tension
measurements in the pendant drop case. Such trend is repeated when decreasing or
increasing the drop volume. Similar trends were also noted in other cases. Figure 4.22(a)
shows the error in surface tension of hexadecane formed on a holder with outer diameter
of 4.24 mm, while figure 4.22(b) shows the error in surface tension of DEP formed on a
holder with outer diameter of 4.24 mm. Both figures show similar – but not exactly the
same – trends to that of figure 4.9. Generally, the errors in the pendant drop cases are
much smaller than in the constrained sessile drop cases shown in Section 4.2.
The same analysis was applied to experiments using the other three liquids, de-
camethylcyclopentasiloxane (DMCPS), diethyl phthalate (DEP) and water, using dif-
ferent holders. Following the same procedure shown above, a critical shape parameter
value was calculated from every run. All the results are summarized in Table 4.2. The
first two columns of the table list the holder size and the liquid used for every experi-
ment. The third column lists the presumably correct surface tension values as found in
the literature and/or previous work [217–219]. The fourth column lists the Bond number
for every experiment; a range of Bond numbers between 0.61 and 5.02 is covered in this
study. The following three columns contain the calculated surface tension value (from
the largest drop sizes) and the accuracy of the calculation; more than 60 drops were used
in the calculation of the average value and the standard deviation in each run. The last
three columns contain the measured critical shape parameter for every experimental run.
These values were measured in a procedure similar to that used in figure 4.21.
Inspection of table 4.2 shows that for all systems, except the two listed first and last,
respectively, there is excellent agreement from run to run, as well as with the standard
values of column 3. It is also noted that the differences from run to run could be due
to ambient temperature changes of the order of 0.1oC. The consistency from run to run
Chapter 4. Accuracy and Shape Parameter 175
0 0.02 0.04 0.06 0.08 0.1 0.12−0.5
−0.4
−0.3
−0.2
−0.1
0
0.1
0.2
0.3
0.4
0.5
Shape Parameter, Pc
Err
or
in S
urf
ace
Te
nsio
n, ε
(m
J/m
2)
(a)
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16−0.5
−0.4
−0.3
−0.2
−0.1
0
0.1
0.2
0.3
0.4
0.5
Shape Parameter, Pc
Err
or
in S
urf
ace
Te
nsio
n, ε
(m
J/m
2)
(b)
Figure 4.22: The error in surface tension measured by ADSA as a function of the shapeparameter for different drop sizes of a pendant drop of: (a) hexadecane formed using aholder with outer diameter of 4.24 mm, and (b) DEP formed using a holder with outerdiameter of 4.24 mm. For a maximum error of ±0.1 mJ/m2, the critical shape parametersare 0.052 and 0.039, respectively.
Chapter 4. Accuracy and Shape Parameter 176
Tab
le4.
2:Sum
mar
yof
resu
lts
ofp
endan
tdro
pex
per
imen
tsusi
ng
diff
eren
tliquid
san
ddiff
eren
thol
der
size
s.
Hol
der
Liq
uid
γ(m
J/m
2)
Bo
γmeasured
(mJ/m
2)
Pcrit
(mm
)[2
17–2
19]
Run
1R
un
2R
un
3R
un
1R
un
2R
un
3
4.24
Wat
er72
.39
0.61
72.0
33±
0.03
772
.016±
0.02
671
.964±
0.02
7-
--
3.08
DM
CP
S18
.26
1.22
18.2
82±
0.00
418
.287±
0.00
418
.285±
0.00
20.
055
0.04
80.
045
4.24
HE
X27
.17
1.25
27.1
73±
0.01
127
.167±
0.01
227
.152±
0.01
40.
052
0.04
50.
052
6.25
Wat
er72
.39
1.32
72.3
99±
0.01
372
.411±
0.01
072
.391±
0.01
40.
045
0.04
40.
050
4.24
DE
P36
.95
1.33
36.9
62±
0.01
336
.935±
0.00
536
.935±
0.01
10.
043
0.04
60.
039
4.24
DM
CP
S18
.26
2.31
18.3
60±
0.00
518
.302±
0.00
518
.296±
0.01
00.
040
0.04
00.
034
6.25
HE
X27
.17
2.72
27.1
94±
0.00
627
.202±
0.00
427
.201±
0.00
80.
038
0.04
00.
042
6.25
DE
P36
.95
2.90
37.0
35±
0.01
037
.046±
0.00
837
.041±
0.01
40.
035
0.03
70.
037
6.25
DM
CP
S18
.26
5.02
17.8
79±
0.00
218
.060±
0.04
518
.086±
0.03
0-
--
Chapter 4. Accuracy and Shape Parameter 177
1 1.5 2 2.5 3 3.50.03
0.04
0.05
Bond Number, Bo
Critica
l S
ha
pe
Pa
ram
ete
r, P
crit
Experimental Results
Best Fit (Linear)
Figure 4.23: The critical shape parameter as a function of the Bond number. Theexperimental points were fitted to a linear function (continuous line).
and the accuracy of each run are generally less for the first and last systems, and the
agreement with the standard values is much poorer, out of the stipulated error range of
±0.1 mJ/m2. A critical shape parameter does not exist for these cases. The reason of
this difference in pattern is discussed in detail below.
The change of the critical shape parameter as a function of the Bond number is
shown in figure 4.23 for all the experiments of this study. In other words, the critical
shape parameter values for every run obtained in a similar fashion to figure 4.21 and
summarized in the last three columns of table 4.2 are used to calculate an average critical
shape parameter value for every Bond number as shown in figure 4.23. As expected,
figure 4.23 shows that the critical shape parameter depends only on Bond number. The
critical shape parameter is equivalent to a critical dimensionless volume (the drop volume
normalized by the cube of the holder’s radius) that the drop has to exceed to allow for
accurate measurement of the surface tension. This can also be understood from figure
4.15 which defines the relationship between the shape parameter, dimensionless volume
Chapter 4. Accuracy and Shape Parameter 178
and Bond number. The critical shape parameter is one point on every curve below which
the drop shape is indistinguishable from the zero gravity shape (circle in this case).
Similar to the constrained sessile drop case, for Bond numbers below one, the largest
pendant drop that can be formed with a given holder size is not big enough to calculate
an accurate surface tension value. Therefore the critical shape parameters at these con-
ditions, i.e. when the diameter of holder becomes small and approaches zero, cannot be
established. Contrary to the constrained sessile drop case, this is also the case for Bond
numbers above five, i.e. when the diameter of the holder becomes too large or surface
tension becomes too small. This is due to the fact that the largest pendant drop volume
that can be formed is not big enough to produce sufficiently non-spherical drops at high
Bond number in the pendant drop case. This is in agreement with what was expected
from figures 4.16 to 4.18.
The relationship between the critical shape parameter and the Bond number can
be used in the design of a pendant drop experiment. Through the above analysis, we
know that either very low or very high Bond numbers are not suitable and will not allow
accurate surface tension measurement. However, for the constrained sessile drop case,
we know that only very low Bond numbers are not suitable. A low Bond number is
equivalent to a large surface tension value and/or a small holder size, while a high Bond
number is equivalent to a small surface tension value and/or a large holder size. At low
Bond numbers (a liquid with large surface tension value used with a small holder size),
the pendant drop shape is very close to a sphere (low shape parameter). At high Bond
number (a liquid with small surface tension value used with a large holder size), the
maximum pendant drop volume is very small and the shape becomes close to a sphere
again (low shape parameter). If a rough approximation of the surface tension is known,
a suitable holder size can be easily selected to produce accurate measurement. In other
words, designing an ADSA–PD experiment to measure a specific surface tension value
would require a holder diameter between a minimum and a maximum value to ensure a
Chapter 4. Accuracy and Shape Parameter 179
maximum error below a specified error tolerance. This would also require that the volume
of the pendant drop be large enough so that the calculated shape parameter exceeds the
appropriate critical shape parameter as shown in figure 4.23.
For example, based on the calculations presented here and assuming a density dif-
ference of 1000 kg/m3 and gravity of 9.80 m/s2, holder diameters between 2.5 and 5.5
mm are recommended for a surface tension value of 15 mJ/m2 to ensure a maximum
error of ±0.1 mJ/m2 for large drop sizes. Holder diameters between 4.0 and 8.5 mm are
recommended for a surface tension value of 38 mJ/m2. Holder diameters between 5.5
and 12.0 mm are recommended for a surface tension value of 72 mJ/m2. Generally, it
is advised to make the drop as large as possible so that the calculated shape parameter
exceeds the appropriate critical shape parameter to guarantee accurate results. These
rough examples serve as a guide for selecting the right holder size to produce accurate
measurement, if a rough approximation of the surface tension is known.
It can be concluded that the use of a usually thin needle or capillary for pendant
drop experiments is generally not an optimal choice. No single holder is universally
useable. However, a holder size of 5 to 6 mm diameter would be a good choice for
most common liquids, if ultimate accuracy is not required. But if ultimate accuracy is
required, the choice of the support for a pendant drop experiment for surface tension
measurements is extremely important. The size of a suitable holder depends critically
on the Bond number, specifically the density difference across the drop interface and the
surface tension. Neither holders that are too small or too large are appropriate. It can
be seen from table 4.2 that an accuracy near ±0.01 mJ/m2 can be achieved readily.
4.4 Universality of Results
It is believed that the close-to-spherical limitation does exist for ADSA (independent of
the details of the numerical modules) as well as other drop shape techniques. To further
Chapter 4. Accuracy and Shape Parameter 180
investigate this, it is important to test the same results using different modules inside
ADSA and other drop shape techniques as well.
The structure of ADSA consists of three main modules. The first module generates
the theoretical profiles by numerical integration of the Laplace equation (Axisymmetric
Liquid Fluid Interface) “ALFI”. The fifth- and sixth-order Runge-Kutta-Verner pair is
used here for the numerical integration [9]. The second module is the image processing,
which extracts the experimental profiles of drops. Here, the Canny edge detector is used
for the image processing followed by image distortion correction (using a grid) and image
enhancement to sub-pixel resolution [8, 220]. The third module is the optimization
procedure to find the best fit of the theoretical Laplacian curves to the experimental
profile. The Levenberg-Marquardt method is used here for the numerical optimization
[7, 9]. In addition to these three modules, there has been one guiding principle: the input
must be the simplest conceptually possible, e.g. there must be no need for identifying
special points on the drop profile, such as the apex.
The three modules are freely exchangeable against other modules designed for the
same purpose. Such a study with different modules was recently performed [218, 219],
for a different drop constellation, a sessile drop with a capillary protruding into the
drop, ADSA – No Apex (ADSA–NA). To ascertain the general validity of the results in
table 4.2, a similar study was performed here for the pendant drop constellation.
The first step is to test the effect of a different image processing technique. The
Canny edge detector was replaced by a robust non-gradient edge detector called SUSAN
[221]. SUSAN is a non-gradient edge detection operator; it does not depend on intensity
gradients in the digitized image. This was followed by the standard ADSA procedure.
The results of this step “ADSASUS” for the case of a pendant drop of hexadecane formed
using a holder with outer diameter of 6.25 mm are shown in figures 4.24(b) and 4.25.
The computation time was less than one second per image on a computer with a 2.66
GHz CPU and 3.00 GB of RAM. This is similar to that of the standard ADSA.
Chapter 4. Accuracy and Shape Parameter 181
Next, the effect of a different numerical optimization technique is considered. The
Levenberg-Marquardt method was replaced by the non-gradient Nelder-Mead simplex
method [222]. Compared to the gradient methods, Nelder-Mead is computationally eas-
ier to implement since it does not require a derivative of the objective function, but
it consumes more computer time. The results of this step “ADSANM” for the case of
the same pendant drop of hexadecane formed with the same holder are shown in fig-
ures 4.24(c) and 4.25. The computation time was roughly 3 to 4 seconds per image.
The third step is to test a different drop shape technique called theoretical image
fitting analysis “TIFA” [211, 223] that uses a completely different approach from ADSA.
The difference between the strategy of ADSA and TIFA is that, in TIFA, the interfacial
properties are calculated by fitting the whole theoretical image – not the drop profile
only – to the experimental image of the drop. The theoretical image is a black and
white image of the drop generated by numerically solving the Laplace equation. An
error function measures the pixel-by-pixel difference between the gradient of the whole
theoretical and experimental images. The interfacial properties are found by fitting the
gradient of the theoretical image to the gradient of the experimental one by minimizing
the error function. The remarkable feature of TIFA is that it operates without using edge
detection algorithms. In fact, image analysis is tied to the optimization process in TIFA,
and it is not a separate module as in ADSA. The results of this step, using TIFA for
the case of the same pendant drop of hexadecane formed with the same holder are also
shown in figures 4.24(d) and 4.25. The computation time was approximately 72 seconds
per image (approximately 8 hours for this run).
Figure 4.24 shows the comparison between the surface tension values of a pendant
drop of hexadecane formed using a holder with outer diameter of 6.25 mm calculated using
the standard ADSA algorithm, the modified algorithms (ADSASUS and ADSANM), and
the TIFA algorithm. The calculated surface tension values for the large drops are very
consistent for all four algorithms, with values of 27.202±0.004 mJ/m2, 27.208±0.004
Chapter 4. Accuracy and Shape Parameter 182
26
27
28
29
γ (m
J/m
2)
γ = 27.202±0.004 mJ/m2
(a)
26
27
28
29
γ (m
J/m
2)
γ = 27.208±0.004 mJ/m2
(b)
26
27
28
29
γ (m
J/m
2)
γ = 27.201±0.004 mJ/m2
(c)
0 50 100 150 20024
26
28
30
γ (m
J/m
2)
Time, t (s)
γ = 27.194±0.005 mJ/m2
(d)
ADSA
ADSASUS
ADSANM
TIFA
Figure 4.24: The change of surface tension with time of a pendant drop of hexadecaneformed using a holder with outer diameter of 6.25 mm. The calculations are repeatedusing different software options; ADSA: the default options using Canny for edge detec-tion and Levenberg-Marquardt for optimization; ADSASUS: SUSAN is used instead ofCanny; ADSANM : Nelder-Mead is used instead of Levenberg-Marquardt; TIFA: theoret-ical image fitting analysis.
Chapter 4. Accuracy and Shape Parameter 183
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14−0.5
−0.4
−0.3
−0.2
−0.1
0
0.1
0.2
0.3
0.4
0.5
Shape Parameter, Pc
Err
or
in S
urf
ace
Te
nsio
n, ε
(m
J/m
2)
ADSA
ADSASUS
ADSANM
TIFA
Figure 4.25: The error in measured surface tension as a function of the shape parameterfor different drop sizes of a pendant drop of hexadecane formed using a holder with outerdiameter of 6.25 mm. The calculations are repeated using different software options;ADSA: the default options using Canny for edge detection and Levenberg-Marquardt foroptimization; ADSASUS: SUSAN is used instead of Canny; ADSANM : Nelder-Mead isused instead of Levenberg-Marquardt; TIFA: theoretical image fitting analysis.
mJ/m2, 27.201±0.004 mJ/m2, and 27.194±0.005 mJ/m2, respectively. For the small
drops, all algorithms fail to predict the correct value for the surface tension. Interestingly,
the ADSA algorithms are very consistent from one ADSA version to the next, even for
small drops. However, the TIFA algorithms show larger deviation from the correct surface
tension value in the case of small drops.
The error in surface tension measured by these four different algorithms as a function
of the shape parameter is shown in figure 4.25 for the same run. The original and modified
ADSA algorithms are very consistent and show very similar trends. It is important to
note that the overall error in surface tension for this case did not exceed ±0.2 mJ/m2 for
almost all drops in this run, except for very small values of shape parameter below 0.01
where the drop is very small and is essentially a spherical cap. The TIFA algorithm shows
Chapter 4. Accuracy and Shape Parameter 184
similar results for large drops (large shape parameter) but deviates dramatically for small
drops (small shape parameter). For a maximum error of ±0.1 mJ/m2, a conservative
critical shape parameter value is 0.04 in this case.
Stepping well back, it is apparent that, for drops of a reasonable size, high quality
surface tension results are readily obtainable from drop shape techniques. While the
choice of optimization method has a considerable effect on computation time and neces-
sary accuracy of the initial guess, it does not have an appreciable effect on the surface
tension result, see specifically figures 4.24(a) and 4.24(c). Less expected is the finding
that edge detection also does not have a significant effect; the standard ADSA using
Levenberg-Marquardt and Canny (figure 4.24(a)) or alternatively Levenberg-Marquardt
and SUSAN, i.e. a non-gradient edge detection strategy (figure 4.24(b)), agree well. A
method totally outside the ADSA scheme, TIFA, also yields the same results, for large
enough drops.
Finally, the results show that all techniques are outside the stipulated error range of
±0.1 mJ/m2 when the drop or bubble approaches spherical shapes. Using non-gradient
edge detection or non-gradient optimization modules did not change the results. It was
noted that TIFA techniques work well for well deformed drops but yield erroneous results
(worse than ADSA) for close-to-spherical drops. This is in agreement with other studies
in the literature [219].
Chapter 5
Summary of Progress and Outlook
The ADSA–CSD configuration has been further developed to produce a highly versa-
tile methodology for surface tension measurements. ADSA–CSD was used to probe the
behaviour of films adsorbed or spread at air-water interfaces. During these studies it
became clear that the accuracy of the data needed to be quantified and measured. In
Chapter 4, a derivation of a versatile assessment of the accuracy of surface tension mea-
surement is described. This represents a fundamental improvement which can be applied
to other shape dependant configurations.
The accuracy of any drop shape technique will depend on the deviation of an exper-
imental drop shape from the zero gravity shape, i.e. the spherical shape in the case of
sessile/pendant drops considered here. This difference between the two shapes can be
expressed quantitatively as a shape parameter. By comparing shape parameters with
the deviations of calculated surface tensions for different drop sizes of liquids of known
surface tension, a critical shape parameter was established, below which surface tension
measurements will exceed a specified error tolerance. The key element of this proce-
dure is a dimensional analysis used to describe similarity in constrained sessile drop and
pendant drop shapes. The shape parameter concept was used in Chapter 4 to quantify
the accuracy of both the constrained sessile drop (ADSA–CSD) and the pendant drop
185
Chapter 5. Summary of Progress and Outlook 186
(ADSA–PD) constellations. It was shown that an accuracy near ±0.01 mJ/m2 can be
achieved readily for well deformed drops.
It is known that the pendant drop constellation yields the most accurate results among
the drop shape methods. However, in the case of low surface tension values, gravity tends
to detach the drop from the capillary holder before it is large enough [71], making surface
tension measurement impossible. For such low surface tensions, the constrained sessile
drop constellation is more suitable. Such low surface tension values (near 1 mJ/m2) are
very common in lung surfactant studies. The analysis presented in Chapter 4 can be
used in the design of constrained sessile drop and pendant drop experiments for surface
tension measurement. If a rough approximation of the surface tension is known, a holder
of suitable size for a specified accuracy can be easily selected. The size of a suitable
holder depends critically on the Bond number, which in turn depends on the density
difference across the drop interface and the surface tension.
It was shown in Chapter 4 that the use of different edge detectors or optimization
techniques is not expected to improve the accuracy of the surface tension measurements.
If further progress on accuracy is needed, it is recommended to look further into the
quality of the images obtained. Specifically the optics of the whole system, lighting
and refraction, and possibly image magnification are to be considered. It is doubtful if
accuracy can be enhanced much but possibly the range of critical shape parameter and
the working range of Bond numbers for different constellations might be improved (cf.
tables 4.1 and 4.2). A suggested strategy would be to repeat the pure liquids experiments
and shape parameter analysis done in Chapter 4 with different optics. Any such work
would need extreme temperature control.
Using the high accuracy now obtained, the pendant drop method can be used for
many applications. Thus, more accurate temperature dependant surface tension mea-
surement data could be produced. For example, as temperature coefficients of surface
tension are generally in the order of 0.1 mJ/m2·K, using a method with accuracy of ±0.1
Chapter 5. Summary of Progress and Outlook 187
mJ/m2 requires measurements at a much wider temperature range than is needed in a
method with accuracy of ±0.01 mJ/m2. The pendant drop technique has the potential
of being used to determine the effect of temperature on surface tension with very high
accuracy. Most of the present literature data for temperature dependant surface tension
measurement were produced a long time ago with methods such as the capillary rise,
the maximum bubble pressure, or the drop weight [224]. These methods have limita-
tions of which ADSA–PD is free [93, 225]. The capillary rise method requires a very
sophisticated setup, the maximum bubble pressure method assumes a spherical bubble,
and the drop weight method relies heavily on an empirically determined correction fac-
tor [93, 225]. Therefore, there is considerable potential in drop shape techniques due
to the compactness and complete isolation of the experimental setup, allowing for var-
ious experimental conditions ranging from high pressure and high temperature to high
and ultra-high vacuum. The only problem was uncertainty about accuracy which was
removed in Chapter 4.
The shape parameter analysis presented here is also applicable – in principle – to other
configurations such as unconstrained sessile drops, captive bubbles, liquid bridges, and
liquid lenses. However, the comparison with a circle will only apply to sessile/pendant
drops and bubbles. For other configurations, the drop shape must be compared to the
respective zero gravity shape for each case. These drops at zero gravity may assume
different shapes including cylindrical, nodoidal, unduloidal and catenoidal. The radius
of the drop holder can still be used for normalization in the case of a liquid bridge. It
can be replaced by the radius of the capillary in the case of pendant drop or the contact
radius in the cases of unconstrained sessile drop or liquid lens. For the liquid bridge
case, the radius of the upper holder will be a new parameter that has to be considered in
defining the shape. For the unconstrained sessile drop and liquid lens cases, the diameter
of the pedestal (holder) does not exist. The contact angle and contact radius will play
an important role in defining the shape of the drop for these cases.
Chapter 5. Summary of Progress and Outlook 188
In the lung surfactant context, the performance of surfactant preparations were as-
sessed using the dynamic surface tension response to film compression and expansion.
Prior to this work, typically only the minimum surface tension values at the end of com-
pression were reported. Several distinctive physical surface effects are expected to play a
role in the process of film compression and expansion: elasticity, adsorption and desorp-
tion. So far only the elasticity during compression had been considered by determining
the slope of the surface tension versus surface area plot. There are two concerns with
this procedure. First, as seen in Chapter 4, the accuracy of the surface tension and
indeed the surface tension value itself determined by ADSA can depend on drop size,
unless all drop sizes are associated with a sufficiently large shape parameter. If such a
trend of erroneous surface tension occurs, the error in the elasticity as a slope of surface
tension versus area can be quite large. The results of Chapter 4 show how to avoid
these errors. Second, relying on the slope of the surface tension response to the surface
area changes as a strategy to determine elasticity assumes that there is no simultaneous
change of surface concentration, say during compression. This assumption is at variance
with known experimental facts of squeeze out, relaxation and spreading. This difficulty is
removed by the model described in Chapter 3 by allowing the surface tension to change
due to simultaneous effects of one or more of these dynamic phenomena. A dynamic
compression-relaxation model (CRM) was developed and used to describe the mechan-
ical dynamic properties (elasticity, relaxation and adsorption) of insoluble pulmonary
surfactant films by investigating the response of surface tension to changes in surface
area. Multi-variable optimization was introduced here to calculate the dynamic param-
eters from experimental results of dynamic surface tension measurements. A detailed
sensitivity study for CRM was also performed in Chapter 3.
Combined with the unique capabilities of ADSA–CSD, CRM was used to study the
effect of humidity, concentration, compression ratio and frequency of a lung surfactant
preparation, BLES. The highest concentration used was 27 mg/ml, similar to that used
Chapter 5. Summary of Progress and Outlook 189
clinically. Low surface tension at such high concentrations are very difficult to measure
using other methods as discussed in Chapter 1. While the dynamic properties of BLES-
only preparations do not improve significantly with increasing BLES concentration, the
same might not be true in the presence of inhibitors. Using ADSA–CSD, CRM studies of
the dynamic properties of lung surfactant preparations can be extended to more complex
systems containing inhibitors.
There is an urgent need to be able to access the air/liquid interface from both sides.
This need arises from clinical reality: contaminations (pollution) affect the surfactant
interface from the air side and therapeutical surfactant preparations are instilled to the
lungs from the air side, while leakage of plasma and blood proteins can occur from the
liquid side causing severe lung injury. Therefore, accessing the lung surfactant interface
from both sides is a desirable feature in any in vitro technique. The ADSA–CSD setup
was further developed in Chapter 2 to allow for such studies.
Two main capabilities were added to the ADSA–CSD setup. The first one was the
deposition capability which allows spreading of material onto the air/liquid interface from
the air side. This permits the study of the collapse of deposited films through the study of
pure monolayers as well as mixed monolayers. DPPC and DPPG monolayers, which are
the main components of most commercial lung surfactant preparations (Appendix A),
were studied to characterize the role of such molecules in maintaining stable film prop-
erties and surface activity of lung surfactant preparations. The second capability added
was the double injection which facilitates the complete exchange of the subphase of a
spread or adsorbed film from the liquid side. The development was illustrated through
inhibition studies using serum, albumin, fibrinogen, and cholesterol. Evaluation of lung
surfactant inhibition, resistance and reversal under completely controlled physiological
conditions is illustrated in Chapter 2. It is noted that the redesigned ADSA–CSD is
a versatile tool for surface tension measurements that has a wide range of applications
outside the scope of the lung surfactant research area as well.
Chapter 5. Summary of Progress and Outlook 190
Clinical lung surfactant therapies currently use mechanical ventilation, and more
recently High Frequency (more than 1 Hz) Ventilation (HFV). However, little is known
about the role of surfactants in high frequency ventilation. The present ADSA–CSD
setup is not capable of simulating the high frequency cycling of more than 1 Hz due to
mechanical limitations of the servo-mechanism controlling the syringe. However, high
frequency studies can be readily facilitated in ADSA–CSD by straightforward hardware
upgrades. Two main items are needed: a piezo-actuator with a dedicated computer-
programmable controller to facilitate the low amplitude high frequency oscillations of
the drop volume, and a high speed camera. Both are readily available commercially.
The future high frequency ADSA–CSD can be used in conjunction with CRM to char-
acterize the surface activity of relevant lung surfactant preparations at higher frequency,
and to study the changes of the dynamic properties (elasticity, relaxation, and adsorp-
tion) at high rates of film compression and expansion. A high frequency ADSA–CSD
methodology and the ability to characterize dynamic interfacial properties using CRM
at high frequency will have great potential for a wide range of colloid and interfacial
applications, beyond lung surfactants.
Bibliography
[1] R. H. Notter, Lung surfactants: Basic science and clinical applications, MarcelDekker, New York, 2000.
[2] F. Bashforth, J. C. Adams, An attempt to test the theories of capillary action,Cambridge University Press, London, England, 1883.
[3] S. Hartland, R. W. Hartley, Axisymmetric fluid-liquid interfaces, Elsevier ScientificPub. Co., Amsterdam, 1976.
[4] C. Maze, G. Burnet, A non-linear regression method for calculating surface tensionand contact angle from the shape of a sessile drop, Surface Science 13 (2) (1969)451–470.
[5] C. Maze, G. Burnet, Modifications of a non-linear regression technique used tocalculate surface tension from sessile drops, Surface Science 24 (1) (1971) 335–342.
[6] J. D. Malcolm, C. D. Elliott, Interfacial tension from height and diameter of asingle profile drop or captive bubble, Canadian Journal of Chemical Engineering58 (1980) 151–153.
[7] Y. Rotenberg, L. Boruvka, A. W. Neumann, Determination of surface tension andcontact angle from the shapes of axisymmetric fluid interfaces, Journal of Colloidand Interface Science 93 (1) (1983) 169–183.
[8] P. Cheng, D. Li, L. Boruvka, Y. Rotenberg, A. W. Neumann, Automation of Ax-isymmetric Drop Shape Analysis for measurements of interfacial tensions and con-tact angles, Colloids and Surfaces 43 (2-3) (1990) 151–167.
[9] O. I. del Rıo, A. W. Neumann, Axisymmetric Drop Shape Analysis: Computa-tional methods for the measurement of interfacial properties from the shape anddimensions of pendant and sessile drops, Journal of Colloid and Interface Science196 (2) (1997) 136–147.
[10] A. W. Neumann, J. K. Spelt, Applied Surface Thermodynamics, Marcel Dekker,New York, 1996.
[11] A. W. Neumann, R. David, Y. Zuo, Applied Surface Thermodynamics, 2nd Edition,CRC Press, 2010.
191
BIBLIOGRAPHY 192
[12] I. Langmuir, The constitution and fundamental properties of solids and liquids: II.Liquids, Journal of the American Chemical Society 39 (9) (1917) 1848–1906.
[13] J. G. Turcotte, A. M. Sacco, J. M. Steim, S. A. Tabak, R. H. Notter, Chemical syn-thesis and surface properties of an analog of the pulmonary surfactant dipalmitoylphosphatidylcholine, Biochimica et Biophysica Acta - Lipids and Lipid Metabolism488 (2) (1977) 235–248.
[14] S. A. Tabak, R. H. Notter, Modified technique for dynamic surface pressure andrelaxation measurements at the air-water interface, Review of Scientific Instruments48 (9) (1977) 1196–1201.
[15] R. H. Notter, S. A. Tabak, R. D. Mavis, Surface-properties of binary-mixtures ofsome pulmonary surfactant components, Journal of Lipid Research 21 (1) (1980)10–22.
[16] J. Goerke, J. Gonzales, Temperature dependence of dipalmitoyl phosphatidyl-choline monolayer stability, Journal of Applied Physiology Respiratory Environ-mental and Exercise Physiology 51 (5) (1981) 1108–1114.
[17] P. Dynarowicz-Latka, A. Dhanabalan, O. N. Oliveira Jr., Modern physicochemicalresearch on langmuir monolayers, Advances in Colloid and Interface Science 91 (2)(2001) 221–293.
[18] S. H. Myrick, E. I. Franses, Effect of chain length on equilibrium and dynamicsurface tension of spread monolayers of aqueous alcohols, Colloids and Surfaces A:Physicochemical and Engineering Aspects 143 (2-3) (1998) 503–515.
[19] K. B. Chen, C. H. Chang, Y. M. Yang, J. R. Maa, Monolayer collapse behaviorof dipalmitoyl phosphatidylcholine with normal long-chain alcohols at the air/wa-ter interface, Colloids and Surfaces A: Physicochemical and Engineering Aspects216 (1-3) (2003) 45–53.
[20] B. Gzyl, M. Paluch, Monolayers of lipids at the water-air interface, Progress inColloid and Polymer Science 126 (2004) 60–63.
[21] D. Grigoriev, R. Miller, R. Wustneck, N. Wustneck, U. Pison, H. Mohwald, A novelmethod to evaluate the phase transition thermodynamics of langmuir monolayers.application to DPPG monolayers affected by subphase composition, Journal ofPhysical Chemistry B 107 (51) (2003) 14283–14288.
[22] T. F. Alig, H. E. Warriner, L. Lee, J. A. N. Zasadzinski, Electrostatic barrier torecovery of dipalmitoylphosphatidylglycerol monolayers after collapse, BiophysicalJournal 86 (2) (2004) 897–904.
[23] E. S. Brown, R. P. Johnson, J. A. Clements, Pulmonary surface tension, Journalof Applied Physiology 14 (5) (1959) 717–720.
[24] J. A. Clements, R. F. Hustead, R. P. Johnson, I. Gribetz, Pulmonary surface tensionand alveolar stability, Journal of Applied Physiology 16 (3) (1961) 444–450.
BIBLIOGRAPHY 193
[25] E. J. Acosta, R. Gitiafroz, Y. Y. Zuo, Z. Policova, P. N. Cox, M. L. Hair, A. W.Neumann, Effect of humidity on lung surfactant films subjected to dynamic com-pression/expansion cycles, Respiratory Physiology and Neurobiology 155 (3) (2007)255–267.
[26] G. Enhorning, Pulsating bubble technique for evaluating pulmonary surfactant,Journal of Applied Physiology Respiratory Environmental and Exercise Physiology43 (2) (1977) 198–203.
[27] S. B. Hall, M. S. Bermel, Y. T. Ko, H. J. Palmer, G. Enhorning, R. H. Notter,Approximations in the measurement of surface tension on the oscillating bubblesurfactometer, Journal of Applied Physiology 75 (1) (1993) 468–477.
[28] G. Putz, J. Goerke, J. A. Clements, Surface activity of rabbit pulmonary surfactantsubfractions at different concentrations in a captive bubble, Journal of AppliedPhysiology 77 (2) (1994) 597–605.
[29] M. Hallman, K. Bry, U. Lappalainen, A mechanism of nitric oxide-induced surfac-tant dysfunction, Journal of Applied Physiology 80 (6) (1996) 2035–2043.
[30] S. L. Seurynck, N. J. Brown, C. W. Wu, K. W. Germino, E. K. Kohlmeir, E. P.Ingenito, M. R. Glucksberg, A. E. Barren, M. Johnson, Optical monitoring ofbubble size and shape in a pulsating bubble surfactometer, Journal of AppliedPhysiology 99 (2) (2005) 624–633.
[31] K. W. Lu, H. W. Taeusch, Combined effects of polymers and KL4 peptide onsurface activity of pulmonary surfactant lipids, Biochimica et Biophysica Acta -Biomembranes 1798 (6) (2010) 1129–1134.
[32] P. I. Reddy, A. M. Al-Jumaily, G. T. Bold, Dynamic surface tension of naturalsurfactant extract under superimposed oscillations, Journal of Biomechanics 44 (1)(2011) 156–163.
[33] G. Enhorning, B. Shumel, L. Keicher, J. Sokolowski, B. A. Holm, Phospholipasesintroduced into the hypophase affect the surfactant film outlining a bubble, Journalof Applied Physiology 73 (3) (1992) 941–945.
[34] B. A. Holm, Z. Wang, R. H. Notter, Multiple mechanisms of lung surfactant inhi-bition, Pediatric Research 46 (1) (1999) 85–93.
[35] S. B. Hall, R. Z. Lu, A. R. Venkitaraman, R. W. Hyde, R. H. Notter, Inhibitionof pulmonary surfactant by oleic acid: Mechanisms and characteristics, Journal ofApplied Physiology 72 (5) (1992) 1708–1716.
[36] Y. Y. Zuo, R. A. W. Veldhuizen, A. W. Neumann, N. O. Petersen, F. Possmayer,Current perspectives in pulmonary surfactant: Inhibition, enhancement and evalu-ation, Biochimica et Biophysica Acta - Biomembranes 1778 (10) (2008) 1947–1977.
BIBLIOGRAPHY 194
[37] S. Schurch, H. Bachofen, J. Goerke, F. Possmayer, A captive bubble method re-produces the in situ behavior of lung surfactant monolayers, Journal of AppliedPhysiology 67 (6) (1989) 2389–2396.
[38] S. Schurch, M. Lee, P. Gehr, Pulmonary surfactant: Surface properties and functionof alveolar and airway surfactant, Pure and Applied Chemistry 64 (1992) 1747–1750.
[39] G. Putz, J. Goerke, H. W. Taeusch, J. A. Clements, Comparison of captive andpulsating bubble surfactometers with use of lung surfactants, Journal of AppliedPhysiology 76 (4) (1994) 1425–1431.
[40] W. M. Schoel, S. Schurch, J. Goerke, The captive bubble method for the eval-uation of pulmonary surfactant: Surface tension, area, and volume calculations,Biochimica et Biophysica Acta - General Subjects 1200 (3) (1994) 281–290.
[41] R. M. Prokop, A. Jyoti, M. Eslamian, A. Garg, M. Mihaila, O. I. del Rıo, S. S. Sus-nar, Z. Policova, A. W. Neumann, A study of captive bubbles with AxisymmetricDrop Shape Analysis, Colloids and Surfaces A: Physicochemical and EngineeringAspects 131 (1-3) (1998) 231–247.
[42] G. Putz, M. Walch, M. V. Eijk, H. P. Haagsman, A spreading technique for formingfilm in a captive bubble, Biophysical Journal 75 (5) (1998) 2229–2239.
[43] J. R. Codd, S. Schurch, C. B. Daniels, S. Orgeig, Torpor-associated fluctuations insurfactant activity in gould’s wattled bat, Biochimica et Biophysica Acta - Molec-ular and Cell Biology of Lipids 1580 (1) (2002) 57–66.
[44] N. Wustneck, R. Wustneck, D. Vollhardt, R. Miller, U. Pison, The influence ofspreading solvent traces in the atmosphere on surface tension measurements byusing a micro-film balance and the captive bubble method, Materials Science andEngineering C 8-9 (1999) 57–64.
[45] N. Wustneck, R. Wustneck, V. B. Fainerman, U. Pison, R. Miller, Investigation ofover-compressed spread L-dipalmitoyl phosphatidylcholine films and the influenceof solvent vapour in the gas phase on π-A isotherms measured by using the cap-tive bubble technique, Colloids and Surfaces A: Physicochemical and EngineeringAspects 164 (2-3) (2000) 267–278.
[46] Y. Y. Zuo, R. Gitiafroz, E. J. Acosta, Z. Policova, P. N. Cox, M. L. Hair, A. W.Neumann, Effect of humidity on the adsorption kinetics of lung surfactant at air-water interfaces, Langmuir 21 (23) (2005) 10593–10601.
[47] Y. Y. Zuo, E. J. Acosta, Z. Policova, P. N. Cox, M. L. Hair, A. W. Neumann, Effectof humidity on the stability of lung surfactant films adsorbed at air-water interfaces,Biochimica et Biophysica Acta - Biomembranes 1758 (10) (2006) 1609–1620.
[48] L. M. Y. Yu, J. J. Lu, Y. W. Chan, A. Ng, L. Zhang, M. Hoorfar, Z. Policova,K. Grundke, A. W. Neumann, Constrained sessile drop as a new configuration
BIBLIOGRAPHY 195
to measure low surface tension in lung surfactant systems, Journal of AppliedPhysiology 97 (2) (2004) 704–715.
[49] D. Schurch, O. L. Ospina, A. Cruz, J. Perez-Gil, Combined and independent actionof proteins SP-B and SP-C in the surface behavior and mechanical stability ofpulmonary surfactant films, Biophysical Journal 99 (10) (2010) 3290–3299.
[50] H. Tavana, R. Gitiafroz, M. L. Hair, A. W. Neumann, Determination of solidsurface tension from contact angles: The role of shape and size of liquid molecules,Journal of Adhesion 80 (8) (2004) 705–725.
[51] H. Tavana, C. N. C. Lam, K. Grundke, P. Friedel, D. Y. Kwok, M. L. Hair, A. W.Neumann, Contact angle measurements with liquids consisting of bulky molecules,Journal of Colloid and Interface Science 279 (2) (2004) 493–502.
[52] H. Tavana, N. Petong, A. Hennig, K. Grundke, A. W. Neumann, Contact anglesand coating film thickness, Journal of Adhesion 81 (1) (2005) 29–39.
[53] S. Schurch, H. Bachofen, F. Possmayer, Complexity in Structure and Function ofthe Lung, Vol. 121, Marcel Dekker, New York, 1998, Ch. Alveolar lining layer:Functions, composition, structures, pp. 35–73.
[54] Y. Y. Zuo, M. Ding, D. Li, A. W. Neumann, Further development of AxisymmetricDrop Shape Analysis - Captive Bubble for pulmonary surfactant related studies,Biochimica et Biophysica Acta - General Subjects 1675 (1-3) (2004) 12–20.
[55] Y. Y. Zuo, M. Ding, A. Bateni, M. Hoorfar, A. W. Neumann, Improvement of in-terfacial tension measurement using a captive bubble in conjunction with Axisym-metric Drop Shape Analysis (ADSA), Colloids and Surfaces A: Physicochemicaland Engineering Aspects 250 (1-3 SPEC. ISS.) (2004) 233–246.
[56] Y. Y. Zuo, F. Possmayer, How does pulmonary surfactant reduce surface tensionto very low values, Journal of Applied Physiology 102 (5) (2007) 1733–1734.
[57] J. M. Crane, G. Putz, S. B. Hall, Persistence of phase coexistence in disatu-rated phosphatidylcholine monolayers at high surface pressures, Biophysical Jour-nal 77 (6) (1999) 3134–3143.
[58] J. M. Crane, S. B. Hall, Rapid compression transforms interracial monolayers ofpulmonary surfactant, Biophysical Journal 80 (4) (2001) 1863–1872.
[59] D. Y. Kwok, D. Vollhardt, R. Miller, D. Li, A. W. Neumann, Axisymmetric DropShape Analysis as a film balance, Colloids and Surfaces A: Physicochemical andEngineering Aspects 88 (1) (1994) 51–58.
[60] M. A. Cabrerizo-Vılchez, H. A. Wege, J. A. Holgado-Terriza, A. W. Neumann,Axisymmetric Drop Shape Analysis as penetration Langmuir balance, Review ofScientific Instruments 70 (5) (1999) 2438–2444.
BIBLIOGRAPHY 196
[61] R. Wustneck, N. Wustneck, B. Moser, V. Karageorgieva, U. Pison, Surface dilata-tional behavior of pulmonary surfactant components spread on the surface of apendant drop: I. Dipalmitoyl phosphatidylcholine and surfactant protein C, Lang-muir 18 (4) (2002) 1119–1124.
[62] R. Wustneck, N. Wustneck, B. Moser, U. Pison, Surface dilatational behavior ofpulmonary surfactant components spread on the surface of a pendant drop: II.Dipalmitoyl phosphatidylcholine and surfactant protein B, Langmuir 18 (4) (2002)1125–1130.
[63] J. J. Lu, L. M. Y. Yu, W. W. Y. Cheung, Z. Policova, D. Li, M. L. Hair, A. W.Neumann, The effect of concentration on the bulk adsorption of bovine lipid extractsurfactant, Colloids and Surfaces B: Biointerfaces 29 (2-3) (2003) 119–130.
[64] M. Hoorfar, M. A. Kurz, A. W. Neumann, Evaluation of the surface tension mea-surement of Axisymmetric Drop Shape Analysis (ADSA) using a shape parameter,Colloids and Surfaces A: Physicochemical and Engineering Aspects 260 (1-3) (2005)277–285.
[65] D. Y. Kwok, B. Tadros, H. Deol, D. Vollhardt, R. Miller, M. A. Cabrerizo-Vılchez,A. W. Neumann, Axisymmetric Drop Shape Analysis as a film balance: Rate depen-dence of the collapse pressure and molecular area at close packing of 1-octadecanolmonolayers, Langmuir 12 (7) (1996) 1851–1859.
[66] A. Jyoti, R. M. Prokop, J. B. Li, D. Vollhardt, D. Y. Kwok, R. Miller, H. Mohwald,A. W. Neumann, An investigation of the compression rate dependence on the sur-face pressure-surface area isotherm for a dipalmitoyl phosphatidylcholine monolayerat the air/water interface, Colloids and Surfaces A: Physicochemical and Engineer-ing Aspects 116 (1-2) (1996) 173–180.
[67] A. Jyoti, R. M. Prokop, A. W. Neumann, Manifestation of the liquid-expanded/liquid-condensed phase transition of a dipalmitoylphosphatidylcholinemonolayer at the air-water interface, Colloids and Surfaces B: Biointerfaces 8 (3)(1997) 115–124.
[68] H. A. Wege, J. A. Holgado-Terriza, M. J. Galvez-Ruiz, M. A. Cabrerizo-Vılchez,Development of a new Langmuir-type pendant-drop film balance, Colloids andSurfaces A: Physicochemical and Engineering Aspects 12 (3-6) (1999) 339–349.
[69] R. Wustneck, N. Wustneck, D. O. Grigoriev, U. Pison, R. Miller, Stress relaxationbehaviour of dipalmitoyl phosphatidylcholine monolayers spread on the surface ofa pendant drop, Colloids and Surfaces B: Biointerfaces 15 (3-4) (1999) 275–288.
[70] R. Wustneck, P. Enders, N. Wustneck, U. Pison, R. Miller, D. Vollhardt, Sur-face dilational behaviour of spread dipalmitoyl phosphatidyl glycerol monolayers,PhysChemComm 2 (1999) 1–11.
[71] J. F. Padday, A. R. Pitt, The stability of axisymmetric menisci, PhilosophicalTransactions of the Royal Society of London A: Mathematical and Physical Sciences275 (1253) (1973) 490–528.
BIBLIOGRAPHY 197
[72] F. Wildeboer-Venema, Influence of nitrogen, oxygen, air and alveolar gas uponsurface tension of lung surfactant, Respiration Physiology 58 (1) (1984) 1–14.
[73] M. Wulf, S. Michel, K. Grundke, O. I. del Rıo, D. Y. Kwok, A. W. Neumann,Simultaneous determination of surface tension and density of polymer melts us-ing Axisymmetric Drop Shape Analysis, Journal of Colloid and Interface Science210 (1) (1999) 172–181.
[74] F. Gerber, M. P. Krafft, T. F. Vandamme, M. Goldmann, P. Fontaine, Fluidizationof a dipalmitoyl phosphatidylcholine monolayer by fluorocarbon gases: Potentialuse in lung surfactant therapy, Biophysical Journal 90 (9) (2006) 3184–3192.
[75] F. Gerber, M. P. Krafft, T. F. Vandamme, The detrimental effect of serum albuminon the re-spreading of a dipalmitoyl phosphatidylcholine langmuir monolayer iscounteracted by a fluorocarbon gas, Biochimica et Biophysica Acta - Biomembranes1768 (3) (2007) 490–494.
[76] Y. Y. Zuo, H. Alolabi, A. Shafiei, N. Kang, Z. Policova, P. N. Cox, E. J. Acosta,M. L. Hair, A. W. Neumann, Chitosan enhances the in vitro surface activity ofdilute lung surfactant preparations and resists albumin-induced inactivation, Pedi-atric Research 60 (2) (2006) 125–130.
[77] R. Veldhuizen, K. Nag, S. Orgeig, F. Possmayer, The role of lipids in pulmonarysurfactant, Biochimica et Biophysica Acta - Molecular Basis of Disease 1408 (2-3)(1998) 90–108.
[78] C. J. Lang, A. D. Postle, S. Orgeig, F. Possmayer, W. Bernhard, A. K. Panda, K. D.Jurgens, W. K. Milsom, K. Nag, C. B. Daniels, Dipalmitoylphosphatidylcholine isnot the major surfactant phospholipid species in all mammals, American Journalof Physiology - Regulatory Integrative and Comparative Physiology 289 (5 58-5)(2005) R1426R1439.
[79] A. D. Bangham, C. J. Morley, M. C. Phillips, The physical properties of an effectivelung surfactant, Biochimica et Biophysica Acta 573 (3) (1979) 552–556.
[80] A. Wilkinson, P. A. Jenkins, J. A. Jeffrey, Two controlled trials of dry artificialsurfactant: Early effects and later outcome in babies with surfactant deficiency,Lancet 2 (8450) (1985) 287–291.
[81] C. J. Morley, Ten centre trial of artificial surfactant (artificial lung expanding com-pound) in very premature babies, British Medical Journal 294 (6578) (1987) 991–996.
[82] C. G. Cochrane, S. D. Revak, Pulmonary surfactant protein B (SP-B): Structure-function relationships, Science 254 (5031) (1991) 566–568.
[83] F. R. Moya, J. Gadzinowski, E. Bancalari, V. Salinas, B. Kopelman, A. Bancalari,M. K. Kornacka, T. A. Merritt, R. Segal, C. J. Schaber, H. Tsai, J. Massaro,
BIBLIOGRAPHY 198
R. D’Agostino, A multicenter, randomized, masked, comparison trial of lucinac-tant, colfosceril palmitate, and beractant for the prevention of respiratory distresssyndrome among very preterm infants, Pediatrics 115 (4) (2005) 1018–1029.
[84] M. Saleem, M. C. Meyer, D. Breitenstein, H. J. Galla, The surfactant peptideKL4 in lipid monolayers: phase behavior, topography, and chemical distribution,Journal of Biological Chemistry 283 (8) (2008) 5195–5207.
[85] B. J. Benson, Genetically engineered human pulmonary surfactant, Clinics in Peri-natology 20 (4) (1993) 791–811.
[86] R. G. Spragg, J. F. Lewis, H. D. Walmrath, J. Johannigman, G. Bellingan, P. F.Laterre, M. C. Witte, G. A. Richards, G. Rippin, F. Rathgeb, D. Hafner, F. J. H.Taut, W. Seeger, Effect of recombinant surfactant protein C-based surfactant onthe acute respiratory distress syndrome, New England Journal of Medicine 351 (9)(2004) 884–892+947.
[87] P. Markart, C. Ruppert, M. Wygrecka, T. Colaris, B. Dahal, D. Walmrath, H. Har-bach, J. Wilhelm, W. Seeger, R. Schmidt, A. Guenther, Patients with ARDS showimprovement but not normalisation of alveolar surface activity with surfactanttreatment: Putative role of neutral lipids, Thorax 62 (7) (2007) 588–594.
[88] S. M. I. Saad, Z. Policova, E. J. Acosta, A. W. Neumann, Axisymmetric DropShape Analysis - Constrained Sessile Drop (ADSA-CSD): A film balance techniquefor high collapse pressures, Langmuir 24 (19) (2008) 10843–10850.
[89] S. M. I. Saad, Z. Policova, E. J. Acosta, M. L. Hair, A. W. Neumann, MixedDPPC/DPPG monolayers at very high film compression, Langmuir 25 (18) (2009)10907–10912.
[90] S. M. I. Saad, Z. Policova, A. Dang, E. J. Acosta, M. L. Hair, A. W. Neumann,A double injection ADSA-CSD methodology for lung surfactant inhibition andreversal studies, Colloids and Surfaces B: Biointerfaces 73 (2) (2009) 365–375.
[91] N. Kang, Z. Policova, G. Bankian, M. L. Hair, Y. Y. Zuo, A. W. Neumann, E. J.Acosta, Interaction between chitosan and bovine lung extract surfactants, Biochim-ica et Biophysica Acta - Biomembranes 1778 (1) (2008) 291–302.
[92] G. L. Gaines Jr., Insoluble monolayers at liquid-gas interfaces, Interscience Pub-lishers, New York, 1966.
[93] A. W. Adamson, Physical chemistry of surfaces, Wiley, New York, 1990.
[94] A. V. Nahmen, M. Schenk, M. Sieber, M. Amrein, The structure of a model pul-monary surfactant as revealed by scanning force microscopy, Biophysical Journal72 (1) (1997) 463–469.
[95] A. V. Nahmen, A. Post, H. J. Galla, M. Sieber, The phase behavior of lipid mono-layers containing pulmonary surfactant protein C studied by fluorescence light mi-croscopy, European Biophysics Journal 26 (5) (1997) 359–369.
BIBLIOGRAPHY 199
[96] S. Krol, M. Ross, M. Sieber, S. Kunneke, H. J. Galla, A. Janshoff, Formation ofthree-dimensional protein-lipid aggregates in monolayer films induced by surfactantprotein B, Biophysical Journal 79 (2) (2000) 904–918.
[97] R. V. Diemel, M. M. E. Snel, A. J. Waring, F. J. Walther, L. M. G. V. Golde,G. Putz, H. P. Haagsman, J. J. Batenburg, Multilayer formation upon compressionof surfactant monolayers depends on protein concentration as well as lipid compo-sition: An atomic force microscopy study, Journal of Biological Chemistry 277 (24)(2002) 21179–21188.
[98] L. Wang, P. Cai, H. J. Galla, H. He, C. R. Flach, R. Mendelsohn, Monolayer-multilayer transitions in a lung surfactant model: IR reflection-absorption spec-troscopy and atomic force microscopy, European Biophysics Journal 34 (3) (2005)243–254.
[99] S. A. Roberts, I. W. Kellaway, K. M. G. Taylor, B. Warburton, K. Peters, Combinedsurface pressure-interfacial shear rheology studies of the interaction of proteins withspread phospholipid monolayers at the air-water interface, International Journal ofPharmaceutics 300 (1-2) (2005) 48–55.
[100] M. R. Morrow, S. Temple, J. Stewart, K. M. W. Keoughy, Comparison of DPPC andDPPG environments in pulmonary surfactant models, Biophysical Journal 93 (1)(2007) 164–175.
[101] M. Hoorfar, A. W. Neumann, Recent progress in Axisymmetric Drop Shape Anal-ysis (ADSA), Advances in Colloid and Interface Science 121 (1-3) (2006) 25–49.
[102] J. Katsaras, D. S. C. Yang, R. M. Epand, Fatty-acid chain tilt angles and direc-tions in dipalmitoyl phosphatidylcholine bilayers, Biophysical Journal 63 (4) (1992)1170–1175.
[103] J. F. Nagle, S. Tristram-Nagle, Structure of lipid bilayers, Biochimica et BiophysicaActa - Reviews on Biomembranes 1469 (3) (2000) 159–195.
[104] G. Ma, H. C. Allen, DPPC langmuir monolayer at the air-water interface: Probingthe tail and head groups by vibrational sum frequency generation spectroscopy,Langmuir 22 (12) (2006) 5341–5349.
[105] A. L. Caro, M. R. R. Nino, J. M. R. Patino, The effect of pH on structural,topographical, and rheological characteristics of β-casein-DPPC mixed monolayersspread at the air-water interface, Colloids and Surfaces A: Physicochemical andEngineering Aspects 332 (2-3) (2009) 180–191.
[106] C. Ohe, Y. Ida, S. Matsumoto, T. Sasaki, Y. Goto, M. Noi, T. Tsurumaru, K. Itoh,Investigations of polymyxin B-phospholipid interactions by vibrational sum fre-quency generation spectroscopy, Journal of Physical Chemistry B 108 (46) (2004)18081–18087.
BIBLIOGRAPHY 200
[107] N. Wustneck, R. Wustneck, U. Pison, Surface dilatational behavior of pulmonarysurfactant components spread on the surface of a captive bubble: III. Dipalmi-toyl phosphatidylcholine, surfactant protein C, and surfactant protein B, Langmuir19 (18) (2003) 7521–7527.
[108] J. M. Brockman, Z. Wang, R. H. Notter, R. A. Dluhy, Effect of hydrophobic surfac-tant proteins SP-B and SP-C on binary phospholipid monolayers: II. Infrared exter-nal reflectance-absorption spectroscopy, Biophysical Journal 84 (1) (2003) 326–340.
[109] A. Cruz, L. Vazquez, M. Velez, J. Perez-Gil, Effect of pulmonary surfactant proteinSP-B on the micro- and nanostructure of phospholipid films, Biophysical Journal86 (1 I) (2004) 308–320.
[110] D. Y. Takamoto, M. M. Lipp, A. V. Nahmen, K. Y. C. Lee, A. J. Waring, J. A.Zasadzinski, Interaction of lung surfactant proteins with anionic phospholipids,Biophysical Journal 81 (1) (2001) 153–169.
[111] E. P. Ingenito, R. Mora, L. Mark, Pivotal role of anionic phospholipids in deter-mining dynamic behavior of lung surfactant, American Journal of Respiratory andCritical Care Medicine 161 (3) (2000) 831–838.
[112] K. Rodriguez-Capote, K. Nag, S. Schurch, F. Possmayer, Surfactant protein inter-actions with neutral and acidic phospholipid films, American Journal of Physiology- Lung Cellular and Molecular Physiology 281 (1 25-1) (2001) L231–L242.
[113] Z. Wang, O. Gurel, S. Weinbach, R. H. Notter, Primary importance of zwitterionicover anionic phospholipids in the surface-active function of calf lung surfactantextract, American Journal of Respiratory and Critical Care Medicine 156 (4 PARTI) (1997) 1049–1057.
[114] S. H. Yu, F. Possmayer, Adsorption, compression and stability of surface filmsfrom natural, lipid extract and reconstituted pulmonary surfactants, Biochimica etBiophysica Acta - Lipids and Lipid Metabolism 1167 (3) (1993) 264–271.
[115] R. H. Notter, S. A. Tabak, S. Holcomb, R. D. Mavis, Postcollapse dynamic surfacepressure relaxation in binary surface films containing dipalmitoyl phosphatidyl-choline, Journal of Colloid and Interface Science 74 (2) (1980) 370–377.
[116] T. Fuchimukai, T. Fujiwara, A. Takahashi, G. Enhorning, Artificial pulmonarysurfactant inhibited by proteins, Journal of Applied Physiology 62 (2) (1987) 429–437.
[117] S. A. Tabak, R. H. Notter, Effect of plasma proteins on the dynamic π-A charac-teristics of saturated phospholipid films, Journal of Colloid and Interface Science59 (2) (1977) 293–300.
[118] H. W. Taeusch, J. B. De-La-Serna, J. Perez-Gil, C. Alonso, J. A. Zasadzinski, Inac-tivation of pulmonary surfactant due to serum-inhibited adsorption and reversal byhydrophilic polymers: Experimental, Biophysical Journal 89 (3) (2005) 1769–1779.
BIBLIOGRAPHY 201
[119] H. E. Warriner, J. Ding, A. J. Waring, J. A. Zasadzinski, A concentration-dependent mechanism by which serum albumin inactivates replacement lung sur-factants, Biophysical Journal 82 (2) (2002) 835–842.
[120] W. Seeger, A. Gunther, C. Thede, Differential sensitivity to fibrinogen inhibition ofSP-C- vs. SP-B-based surfactants, American Journal of Physiology - Lung Cellularand Molecular Physiology 262 (3) (1992) L286–L291.
[121] B. A. Holm, R. H. Notter, Effects of hemoglobin and cell membrane lipids onpulmonary surfactant activity, Journal of Applied Physiology 63 (4) (1987) 1434–1442.
[122] B. A. Holm, G. Enhorning, R. H. Notter, A biophysical mechanism by which plasmaproteins inhibit lung surfactant activity, Chemistry and Physics of Lipids 49 (1-2)(1988) 49–55.
[123] Z. Wang, R. H. Notter, Additivity of protein and nonprotein inhibitors of lungsurfactant activity, American Journal of Respiratory and Critical Care Medicine158 (1) (1998) 28–35.
[124] A. M. Cockshutt, F. Possmayer, Lysophosphatidylcholine sensitizes lipid extractsof pulmonary surfactant to inhibition by serum proteins, Biochimica et BiophysicaActa - Lipids and Lipid Metabolism 1086 (1) (1991) 63–71.
[125] K. G. Davidson, A. D. Bersten, H. A. Barr, K. D. Dowling, T. E. Nicholas,I. R. Doyle, Lung function, permeability, and surfactant composition in oleic acid-induced acute lung injury in rats, American Journal of Physiology - Lung Cellularand Molecular Physiology 279 (6 23-6) (2000) L1091–L1102.
[126] D. A. Clark, G. F. Nieman, J. E. Thompson, Surfactant displacement by meconiumfree fatty acids: An alternative explanation for atelectasis in meconium aspirationsyndrome, Journal of Pediatrics 110 (5) (1987) 765–770.
[127] T. Gross, E. Zmora, Y. Levi-Kalisman, O. Regev, A. Herman, Lung-surfactant-meconium interaction: In vitro study in bulk and at the air-solution interface,Langmuir 22 (7) (2006) 3243–3250.
[128] L. Gunasekara, S. Schurch, W. M. Schoel, K. Nag, Z. Leonenko, M. Haufs, M. Am-rein, Pulmonary surfactant function is abolished by an elevated proportion ofcholesterol, Biochimica et Biophysica Acta - Molecular and Cell Biology of Lipids1737 (1) (2005) 27–35.
[129] Z. Leonenko, S. Gill, S. Baoukina, L. Monticelli, J. Doehner, L. Gunasekara,F. Felderer, M. Rodenstein, L. M. Eng, M. Amrein, An elevated level of cholesterolimpairs self-assembly of pulmonary surfactant into a functional film, BiophysicalJournal 93 (2) (2007) 674–683.
[130] B. A. Holm, L. Keicher, M. Liu, J. Sokolowski, G. Enhorning, Inhibition of pul-monary surfactant function by phospholipases, Journal of Applied Physiology 71 (1)(1991) 317–321.
BIBLIOGRAPHY 202
[131] K. Rodriguez-Capote, D. Manzanares, T. Haines, F. Possmayer, Reactive oxygenspecies inactivation of surfactant involves structural and functional alterations tosurfactant proteins SP-B and SP-C, Biophysical Journal 90 (8) (2006) 2808–2821.
[132] S. A. Shelley, Oxidant-induced alterations of lung surfactant system, Journal of theFlorida Medical Association 81 (1) (1994) 49–51.
[133] A. Grenha, C. I. Grainger, L. A. Dailey, B. Seijo, G. P. Martin, C. Remunan-Lopez,B. Forbes, Chitosan nanoparticles are compatible with respiratory epithelial cellsin vitro, European Journal of Pharmaceutical Sciences 31 (2) (2007) 73–84.
[134] A. Grenha, C. Remunan-Lopez, E. L. S. Carvalho, B. Seijo, Microspheres contain-ing lipid/chitosan nanoparticles complexes for pulmonary delivery of therapeuticproteins, European Journal of Pharmaceutics and Biopharmaceutics 69 (1) (2008)83–93.
[135] J. K. Ferri, N. Gorevski, C. Kotsmar, M. E. Leser, R. Miller, Desorption kinetics ofsurfactants at fluid interfaces by novel coaxial capillary pendant drop experiments,Colloids and Surfaces A: Physicochemical and Engineering Aspects 319 (1-3) (2008)13–20.
[136] G. Loglio, U. Tesei, N. D. Innocenti, R. Miller, R. Cini, Non-equilibrium surfacethermodynamics: Measurement of transient dynamic surface tension for fluid-fluidinterfaces by the trapezoidal pulse technique, Colloids and Surfaces 57 (2) (1991)335–342.
[137] R. W. Walters, R. R. Jenq, S. B. Hall, Distinct steps in the adsorption of pulmonarysurfactant to an air-liquid interface, Biophysical Journal 78 (1) (2000) 257–266.
[138] X. Wen, E. I. Franses, Adsorption of bovine serum albumin at the air/water inter-face and its effect on the formation of DPPC surface film, Colloids and Surfaces A:Physicochemical and Engineering Aspects 190 (3) (2001) 319–332.
[139] S. Schurch, R. Qanbar, H. Bachofen, F. Possmayer, The surface-associated surfac-tant reservoir in the alveolar lining, Biology of the Neonate 67 (SUPPL. 1) (1995)61–76.
[140] W. Seeger, G. Stohr, H. R. D. Wolf, H. Neuhof, Alteration of surfactant functiondue to protein leakage: Special interaction with fibrin monomer, Journal of AppliedPhysiology 58 (2) (1985) 326–338.
[141] S. Orgeig, C. B. Daniels, The roles of cholesterol in pulmonary surfactant: In-sights from comparative and evolutionary studies, Comparative Biochemistry andPhysiology - A Molecular and Integrative Physiology 129 (1) (2001) 75–89.
[142] C. Casals, J. Arias-Diaz, F. Valino, A. Saenz, C. Garcia, J. L. Balibrea, E. Vara,Surfactant strengthens the inhibitory effect of C-reactive protein on human lungmacrophage cytokine release, American Journal of Physiology - Lung Cellular andMolecular Physiology 284 (3 28-3) (2003) L466L472.
BIBLIOGRAPHY 203
[143] E. Keating, L. Rahman, J. Francis, A. Petersen, F. Possmayer, R. Veldhuizen, N. O.Petersen, Effect of cholesterol on the biophysical and physiological properties of aclinical pulmonary surfactant, Biophysical Journal 93 (4) (2007) 1391–1401.
[144] Z. Leonenko, E. Finot, V. Vassiliev, M. Amrein, Effect of cholesterol on the phys-ical properties of pulmonary surfactant films: Atomic force measurements study,Ultramicroscopy 106 (8-9) (2006) 687–694.
[145] R. V. Diemel, M. M. E. Snel, L. M. G. V. Golde, G. Putz, H. P. Haagsman, J. J.Batenburg, Effects of cholesterol on surface activity and surface topography ofspread surfactant films, Biochemistry 41 (50) (2002) 15007–15016.
[146] S. M. I. Saad, A. W. Neumann, E. J. Acosta, A dynamic compression-relaxationmodel for lung surfactants, Colloids and Surfaces A: Physicochemical and Engi-neering Aspects 354 (1-3) (2010) 34–44.
[147] R. Wustneck, J. Perez-Gil, N. Wustneck, A. Cruz, V. B. Fainerman, U. Pison,Interfacial properties of pulmonary surfactant layers, Advances in Colloid and In-terface Science 117 (1-3) (2005) 33–58.
[148] V. B. Fainerman, R. Miller, H. Mohwald, General relationships of the adsorptionbehavior of surfactants at the water/air interface, Journal of Physical ChemistryB 106 (4) (2002) 809–819.
[149] F. Ravera, M. Ferrari, L. Liggieri, Modelling of dilational visco-elasticity of ad-sorbed layers with multiple kinetic processes, Colloids and Surfaces A: Physico-chemical and Engineering Aspects 282-283 (2006) 210–216.
[150] S. C. Biswas, S. B. Rananavare, S. B. Hall, Differential effects of lysophosphatidyl-choline on the adsorption of phospholipids to an air/water interface, BiophysicalJournal 92 (2) (2007) 493–501.
[151] R. Miller, P. Joos, V. B. Fainerman, Dynamic surface and interfacial tensions ofsurfactant and polymer solutions, Advances in Colloid and Interface Science 49(1994) 249–302.
[152] R. Miller, Z. Policova, R. V. Sedev, A. W. Neumann, Relaxation behaviour ofhuman albumin adsorbed at the solution/air interface, Colloids and Surfaces A:Physicochemical and Engineering Aspects 76 (1993) 179–185.
[153] D. R. Otis, E. P. Ingenito, R. D. Kamm, M. Johnson, Dynamic surface tensionof surfactant TA: Experiments and theory, Journal of Applied Physiology 77 (6)(1994) 2681–2688.
[154] E. P. Ingenito, L. Mark, J. Morris, Biophysical characterization and modeling oflung surfactant components, Journal of Applied Physiology 86 (5) (1999) 1702–1714.
[155] J. Morris, E. P. Ingenito, L. Mark, Dynamic behavior of lung surfactant, Journalof Biomechanical Engineering 123 (1) (2001) 106–113.
BIBLIOGRAPHY 204
[156] D. Vollhardt, U. Retter, Nucleation in insoluble monolayers: I. Nucleation andgrowth model for relaxation of metastable monolayers, Journal of Physical Chem-istry 95 (9) (1991) 3723–3727.
[157] D. Vollhardt, U. Retter, S. Siegel, Nucleation in insoluble monolayers: II. Con-stant pressure relaxation of octadecanoic acid monolayers, Thin Solid Films 199 (1)(1991) 189–199.
[158] W. Yan, S. B. Hall, Distribution of coexisting solid and fluid phases alters thekinetics of collapse from phospholipid monolayers, Journal of Physical ChemistryB 110 (44) (2006) 22064–22070.
[159] M. Avrami, Kinetics of phase change: I. General theory, Journal of ChemicalPhysics 7 (12) (1939) 1103–1112.
[160] M. Avrami, Kinetics of phase change: II. Transformation-time relations for randomdistribution of nuclei, Journal of Chemical Physics 8 (2) (1940) 212–224.
[161] M. Avrami, Kinetics of phase change: III. Granulation, phase change, and mi-crostructure, Journal of Chemical Physics 9 (2) (1941) 177–184.
[162] S. A. Tabak, R. H. Notter, J. S. Ultman, S. M. Dinh, Relaxation effects in thesurface pressure behavior of dipalmitoyl lecithin, Journal of Colloid and InterfaceScience 60 (1) (1977) 117–125.
[163] S. Schurch, H. Bachofen, F. Possmayer, Surface activity in situ, in vivo, and inthe captive bubble surfactometer, Comparative Biochemistry and Physiology - AMolecular and Integrative Physiology 129 (1) (2001) 195–207.
[164] L. F. Shampine, M. W. Reichelt, The MATLAB ode suite, Society for Industrialand Applied Mathematics Journal of Scientific Computing 18 (1) (1997) 1–22.
[165] J. C. Lagarias, J. A. Reeds, M. H. Wright, P. E. Wright, Convergence propertiesof the Nelder-Mead simplex method in low dimensions, Society for Industrial andApplied Mathematics Journal on Optimization 9 (1) (1999) 112–147.
[166] G. Loglio, R. Miller, A. Stortini, U. Tesei, N. D. Innocenti, R. Cini, Non-equilibriumproperties of fluid interfaces - aperiodic diffusion-controlled regime: II. Experi-ments, Colloids and Surfaces A: Physicochemical and Engineering Aspects 95 (1)(1995) 63–68.
[167] H. Bachofen, U. Gerber, P. Gehr, M. Amrein, S. Schurch, Structures of pulmonarysurfactant films adsorbed to an air-liquid interface in vitro, Biochimica et Biophys-ica Acta - Biomembranes 1720 (1-2) (2005) 59–72.
[168] M. Larsson, T. Nylander, K. M. W. Keough, K. Nag, An X-ray diffraction studyof alterations in bovine lung surfactant bilayer structures induced by albumin,Chemistry and Physics of Lipids 144 (2) (2006) 137–145.
BIBLIOGRAPHY 205
[169] J. J. Lu, W. W. Y. Cheung, L. M. Y. Yu, Z. Policova, D. Li, M. L. Hair, A. W.Neumann, The effect of dextran to restore the activity of pulmonary surfactantinhibited by albumin, Respiratory Physiology and Neurobiology 130 (2) (2002)169–179.
[170] B. A. Holm, R. H. Notter, J. N. Finkelstein, Surface property changes from interac-tions of albumin with natural lung surfactant and extracted lung lipids, Chemistryand Physics of Lipids 38 (3) (1985) 287–298.
[171] J. Goerke, Pulmonary surfactant: Functions and molecular composition, Biochim-ica et Biophysica Acta - Molecular Basis of Disease 1408 (2-3) (1998) 79–89.
[172] Y. Y. Zuo, A. W. Neumann, Pulmonary surfactant and its in vitro assessmentusing Axisymmetric Drop Shape Analysis (ADSA): A review, Tenside SurfactantsDetergents 42 (3) (2005) 126–147.
[173] W. A. Engle, A. R. Stark, D. H. Adamkin, D. G. Batton, E. F. Bell, V. K. Bhutani,S. E. Denson, G. I. Martin, K. L. Watterberg, Surfactant-replacement therapy forrespiratory distress in the preterm and term neonate, Pediatrics 121 (2) (2008)419–432.
[174] I. Frerking, A. Gunther, W. Seeger, U. Pison, Pulmonary surfactant: Functions,abnormalities and therapeutic options, Intensive Care Medicine 27 (11) (2001)1699–1717.
[175] J. F. Lewis, R. Veldhuizen, The role of exogenous surfactant in the treatment ofacute lung injury, Annual Review of Physiology 65 (2003) 613–642.
[176] J. J. Haitsma, P. J. Papadakos, B. Lachmann, Surfactant therapy for acute lunginjury/acute respiratory distress syndrome, Current Opinion in Critical Care 10 (1)(2004) 18–22.
[177] J. F. Lewis, R. A. W. Veldhuizen, The future of surfactant therapy during AL-I/ARDS, Seminars in Respiratory and Critical Care Medicine 27 (4) (2006) 377–388.
[178] H. W. Taeusch, Treatment of acute (adult) respiratory distress syndrome: The holygrail of surfactant therapy, Biology of the Neonate 77 (SUPPL. 1) (2000) 2–8.
[179] A. H. Jobe, Surfactant in the perinatal period, Early Human Development 29 (1-3)(1992) 57–62.
[180] C. S. Calfee, M. A. Matthay, Nonventilatory treatments for acute lung injury andARDS, Chest 131 (3) (2007) 913–920.
[181] R. Jain, A. DalNogare, Pharmacological therapy for acute respiratory distress syn-drome, Mayo Clinic Proceedings 81 (2) (2006) 205–212.
[182] K. J. Bosma, J. F. Lewis, Emerging therapies for treatment of acute lung injuryand acute respiratory distress syndrome, Expert Opinion on Emerging Drugs 12 (3)(2007) 461–477.
BIBLIOGRAPHY 206
[183] M. Tidswell, Nonventilatory interventions in ALI/ARDS: Recent work, Journal ofIntensive Care Medicine 23 (1) (2008) 70–72.
[184] R. G. Spragg, J. F. Lewis, W. Wurst, D. Hafner, R. P. Baughman, M. D. Wewers,J. J. Marsh, Treatment of acute respiratory distress syndrome with recombinantsurfactant protein C surfactant, American Journal of Respiratory and Critical CareMedicine 167 (11) (2003) 1562–1566.
[185] M. B. P. Amato, C. S. V. Barbas, D. M. Medeiros, R. B. Magaldi, G. D. P. P. Schet-tino, G. Lorenzi-Filho, R. A. Kairalla, D. Deheinzelin, C. Munoz, R. Oliveira, T. Y.Takagaki, C. R. R. Carvalho, Effect of a protective-ventilation strategy on mortal-ity in the acute respiratory distress syndrome, New England Journal of Medicine338 (6) (1998) 347–354.
[186] R. G. Brower, M. A. Matthay, A. Morris, D. Schoenfeld, B. T. Thompson,A. Wheeler, Ventilation with lower tidal volumes as compared with traditionaltidal volumes for acute lung injury and the acute respiratory distress syndrome,New England Journal of Medicine 342 (18) (2000) 1301–1308.
[187] R. G. Brower, P. N. Lanken, N. MacIntyre, M. A. Matthay, A. Morris, M. An-cukiewicz, D. Schoenfeld, B. T. Thompson, Higher versus lower positive end-expiratory pressures in patients with the acute respiratory distress syndrome, NewEngland Journal of Medicine 351 (4) (2004) 327–336.
[188] A. J. Aspros, C. G. Coto, J. F. Lewis, R. A. W. Veldhuizen, High-frequency oscil-lation and surfactant treatment in an acid aspiration model, Canadian Journal ofPhysiology and Pharmacology 88 (1) (2010) 14–20.
[189] X. Wen, K. C. McGinnis, E. I. Franses, Unusually low dynamic surface tensions ofaqueous solutions of sodium myristate, Colloids and Surfaces A: Physicochemicaland Engineering Aspects 143 (2-3) (1998) 371–380.
[190] T. D. Graff, D. W. Benson, Systemic and pulmonary changes with inhaled humidatmospheres: Clinical application, Anesthesiology 30 (2) (1969) 199–207.
[191] R. Williams, N. Rankin, T. Smith, D. Galler, P. Seakins, Relationship between thehumidity and temperature of inspired gas and the function of the airway mucosa,Critical Care Medicine 24 (11) (1996) 1920–1929.
[192] J. J. Pillow, High-frequency oscillatory ventilation: Mechanisms of gas exchangeand lung mechanics, Critical Care Medicine 33 (3 SUPPL.).
[193] A. Pettenazzo, A. H. Jobe, M. Ikegami, E. Rider, S. R. Seidner, T. Yamada, Cumu-lative effects of repeated surfactant treatments in the rabbit, Experimental LungResearch 16 (2) (1990) 131–143.
[194] D. Hafner, P. G. Germann, D. Hauschke, Effects of lung surfactant factor (LSF)treatment on gas exchange and histopathological changes in an animal model ofadult respiratory distress syndrome (ARDS): Comparison of recombinant LSF withbovine LSF, Pulmonary Pharmacology 7 (5) (1994) 319–332.
BIBLIOGRAPHY 207
[195] D. Hafner, R. Beume, U. Kilian, G. Krasznai, B. Lachmann, Dose-response compar-isons of five lung surfactant factor (LSF) preparations in an animal model of adultrespiratory distress syndrome (ARDS), British Journal of Pharmacology 115 (3)(1995) 451–458.
[196] J. F. Lewis, B. Tabor, M. Ikegami, A. H. Jobe, M. Joseph, D. Absolom, Lung func-tion and surfactant distribution in saline-lavaged sheep given instilled vs. nebulizedsurfactant, Journal of Applied Physiology 74 (3) (1993) 1256–1264.
[197] J. F. Lewis, In vivo studies of aerosolized exogenous surfactant, Aerosol Scienceand Technology 22 (4) (1995) 354–363.
[198] J. F. Lewis, L. A. McCaig, D. Hafner, R. G. Spragg, R. Veldhuizen, C. Kerr, Dosingand delivery of a recombinant surfactant in lung-injured adult sheep, AmericanJournal of Respiratory and Critical Care Medicine 159 (3) (1999) 741–747.
[199] J. C. Meister, V. Balaraman, T. Ku, A. DeSilva, S. Sood, C. F. T. Uyehara, D. A.Person, D. Easa, Lavage administration of dilute recombinant surfactant in acutelung injury in piglets, Pediatric Research 47 (2) (2000) 240–245.
[200] C. H. Chang, E. I. Franses, An analysis of the factors affecting dynamic tensionmeasurements with the pulsating bubble surfactometer, Journal of Colloid andInterface Science 164 (1) (1994) 107–113.
[201] Y. C. Liao, O. A. Basaran, E. I. Franses, Hydrodynamic effects on the oscillations ofsupported bubbles: Implications for accurate measurements of surface properties,Colloids and Surfaces A: Physicochemical and Engineering Aspects 250 (1-3 SPEC.ISS.) (2004) 367–384.
[202] Y. C. Liao, O. A. Basaran, E. I. Franses, Effects of dynamic surface tension andfluid flow on the oscillations of a supported bubble, Colloids and Surfaces A: Physic-ochemical and Engineering Aspects 282-283 (2006) 183–202.
[203] A. L. Caro, M. R. R. Nimo, J. M. R. Patino, The effect of pH on surface dilata-tional and shear properties of phospholipid monolayers, Colloids and Surfaces A:Physicochemical and Engineering Aspects 327 (1-3) (2008) 79–89.
[204] J. Perez-Gil, K. M. W. Keough, Interfacial properties of surfactant proteins,Biochimica et Biophysica Acta - Molecular Basis of Disease 1408 (2-3) (1998) 203–217.
[205] S. M. I. Saad, Z. Policova, E. J. Acosta, A. W. Neumann, Range of validity ofdrop shape techniques for surface tension measurement, Langmuir 26 (17) (2010)14004–14013.
[206] S. M. I. Saad, Z. Policova, A. W. Neumann, Design and accuracy of pendant dropmethods for surface tension measurement, Colloids and Surfaces A: Physicochemi-cal and Engineering Aspects 384 (1-3) (2011) 442–452.
BIBLIOGRAPHY 208
[207] S. Y. Lin, L. J. Chen, J. W. Xyu, W. J. Wang, An examination on the accuracyof interfacial tension measurement from pendant drop profiles, Langmuir 11 (10)(1995) 4159–4166.
[208] S. Y. Lin, W. J. Wang, L. W. Lin, L. J. Chen, Systematic effects of bubble volumeon the surface tension measured by pendant bubble profiles, Colloids and SurfacesA: Physicochemical and Engineering Aspects 114 (1996) 31–39.
[209] A. T. Morita, D. J. Carastan, N. R. Demarquette, Influence of drop volume onsurface tension evaluated using the pendant drop method, Colloid and PolymerScience 280 (9) (2002) 857–864.
[210] C. Ferrera, J. M. Montanero, M. G. Cabezas, An analysis of the sensitivity ofpendant drops and liquid bridges to measure the interfacial tension, MeasurementScience and Technology 18 (12) (2007) 3713–3723.
[211] M. G. Cabezas, A. Bateni, J. M. Montanero, A. W. Neumann, A new method ofimage processing in the analysis of axisymmetric drop shapes, Colloids and SurfacesA: Physicochemical and Engineering Aspects 255 (1-3) (2005) 193–200.
[212] M. G. Cabezas, A. Bateni, J. M. Montanero, A. W. Neumann, Determination ofsurface tension and contact angle from the shapes of axisymmetric fluid interfaceswithout use of apex coordinates, Langmuir 22 (24) (2006) 10053–10060.
[213] E. Buckingham, On physically similar systems: Illustrations of the use of dimen-sional equations, Physical Review 4 (4) (1914) 345–376.
[214] I. Kasa, A circle fitting procedure and its error analysis, IEEE Transactions onInstrumentation and Measurement IM-25 (1) (1976) 8–14.
[215] V. Pratt, Direct least-squares fitting of algebraic surface, Association for Comput-ing Machinery: Computer Graphics 21 (4) (1987) 145–152.
[216] G. Taubin, Estimation of planar curves, surfaces, and nonplanar space curves de-fined by implicit equations with applications to edge and range image segmentation,IEEE Transactions on Pattern Analysis and Machine Intelligence 13 (11) (1991)1115–1138.
[217] Design Institute for Physical Properties sponsored by AIChE (Ed.), DIPPR Project801 - Full Version, Design Institute for Physical Property Data/AIChE, 2009.
[218] A. Kalantarian, R. David, A. W. Neumann, Methodology for high accuracy contactangle measurement, Langmuir 25 (24) (2009) 14146–14154.
[219] A. Kalantarian, R. David, J. Chen, A. W. Neumann, Simultaneous measurementof contact angle and surface tension using Axisymmetric Drop Shape Analysis - NoApex (ADSA-NA), Langmuir 27 (7) (2011) 3485–3495.
[220] Y. Y. Zuo, C. Do, A. W. Neumann, Automatic measurement of surface tensionfrom noisy images using a component labeling method, Colloids and Surfaces A:Physicochemical and Engineering Aspects 299 (1-3) (2007) 109–116.
BIBLIOGRAPHY 209
[221] S. M. Smith, J. M. Brady, SUSAN: A new approach to low level image processing,International Journal of Computer Vision 23 (1) (1997) 45–78.
[222] J. Nelder, R. Mead, A simplex method for function minimization, Computer Jour-nal 7 (1965) 308–313.
[223] M. G. Cabezas, A. Bateni, J. M. Montanero, A. W. Neumann, A new drop-shapemethodology for surface tension measurement, Applied Surface Science 238 (1-4SPEC. ISS.) (2004) 480–484.
[224] J. J. Jasper, The surface tension of pure liquid compounds, Journal of Physical andChemical Reference Data 1 (4) (1972) 841–1009.
[225] M. Hoorfar, M. A. Kurz, Z. Policova, M. L. Hair, A. W. Neumann, Do polysac-charides such as dextran and their monomers really increase the surface tension ofwater, Langmuir 22 (1) (2006) 52–56.
[226] P. A. Dargaville, J. F. Mills, Surfactant therapy for meconium aspiration syndrome:Current status, Drugs 65 (18) (2005) 2569–2591.
[227] H. L. Halliday, Overview of clinical trials comparing natural and synthetic surfac-tants, Biology of the Neonate 67 (SUPPL. 1) (1995) 32–47.
[228] F. Moya, A. Maturana, Animal-derived surfactants versus past and current syn-thetic surfactants: Current status, Clinics in Perinatology 34 (1) (2007) 145–177.
[229] J. V. Been, L. J. I. Zimmermann, What’s new in surfactant: A clinical view onrecent developments in neonatology and paediatrics, European Journal of Pediatrics166 (9) (2007) 889–899.
[230] W. Seeger, C. Grube, A. Gunther, R. Schmidt, Surfactant inhibition by plasmaproteins: Differential sensitivity of various surfactant preparations, European Res-piratory Journal 6 (7) (1993) 971–977.
[231] W. Bernhard, J. Mottaghian, A. Gebert, G. A. Rau, H. V. der Hardt, C. F. Poets,Commercial versus native surfactants: Surface activity, molecular components, andthe effect of calcium, American Journal of Respiratory and Critical Care Medicine162 (4 I) (2000) 1524–1533.
[232] K. Y. C. Lee, A. Gopal, A. V. Nahmen, J. A. Zasadzinski, J. Majewski, G. S. Smith,P. B. Howes, K. Kjaer, Influence of palmitic acid and hexadecanol on the phasetransition temperature and molecular packing of dipalmitoylphosphatidyl-cholinemonolayers at the air-water interface, Journal of Chemical Physics 116 (2) (2002)774–783.
[233] M. Rudiger, A. Tolle, W. Meier, B. Rustow, Naturally derived commercial surfac-tants differ in composition of surfactant lipids and in surface viscosity, AmericanJournal of Physiology - Lung Cellular and Molecular Physiology 288 (2 32-2) (2005)L379L383.
BIBLIOGRAPHY 210
[234] B. Robertson, H. L. Halliday, Principles of surfactant replacement, Biochimica etBiophysica Acta - Molecular Basis of Disease 1408 (2-3) (1998) 346–361.
[235] O. Blanco, J. Perez-Gil, Biochemical and pharmacological differences betweenpreparations of exogenous natural surfactant used to treat respiratory distress syn-drome: Role of the different components in an efficient pulmonary surfactant, Eu-ropean Journal of Pharmacology 568 (1-3) (2007) 1–15.
[236] L. Gortner, U. Bernsau, H. H. Hellwege, G. Hieronimi, G. Jorch, H. L. Reiter, Amulticenter randomized controlled clinical trial of bovine surfactant for preventionof respiratory distress syndrome, Lung 168 (SUPPL.) (1990) 864–869.
[237] L. Gortner, F. Pohlandt, B. Disse, E. Weller, Effects of bovine surfactant in prema-ture lambs after intra-tracheal application, European Journal of Pediatrics 149 (4)(1990) 280–283.
[238] L. Gortner, U. Bernsau, H. H. Hellwege, G. Hieronimi, G. Jorch, H. L. Reiter, Doesprophylactic use of bovine surfactant change drug utilization in very prematureinfants during neonatal period, Developmental Pharmacology and Therapeutics16 (1) (1991) 1–6.
[239] P. Bartmann, U. Bamberger, F. Pohlandt, L. Gortner, Immunogenicity and im-munomodulatory activity of bovine surfactant (SF-RI 1), Acta Paediatrica - Inter-national Journal of Paediatrics 81 (5) (1992) 383–388.
[240] L. Gortner, P. Bartmann, F. Pohlandt, U. Bernsau, F. Porz, H. H. Hellwege, R. C.Seitz, G. Hieronimi, C. Bremer, G. Jorch, Early treatment of respiratory distresssyndrome with bovine surfactant in very preterm infants: A multicenter controlledclinical trial, Pediatric Pulmonology 14 (1) (1992) 4–9.
[241] J. A. Smyth, I. L. Metcalfe, P. Duffty, Hyaline membrane disease treated withbovine surfactant, Pediatrics 71 (6) (1983) 913–917.
[242] G. Enhorning, A. T. Shennan, F. Possmayer, Prevention of neonatal respiratorydistress syndrome by tracheal instillation of surfactant: A randomized clinical trial,Pediatrics 76 (2) (1985) 145–153.
[243] M. S. Dunn, A. T. Shennan, F. Possmayer, Single- versus multiple-dose surfac-tant replacement therapy in neonates of 30 to 36 weeks’ gestation with respiratorydistress syndrome, Pediatrics 86 (4) (1990) 564–571.
[244] M. S. Dunn, A. T. Shennan, D. Zayack, F. Possmayer, Bovine surfactant replace-ment therapy in neonates of less than 30 weeks’ gestation: A randomized controlledtrial of prophylaxis versus treatment, Pediatrics 87 (3) (1991) 377–386.
[245] M. S. Kwong, E. A. Egan, R. H. Notter, D. L. Shapiro, Double-blind clinical trialof calf lung surfactant extract for the prevention of hyaline membrane disease inextremely premature infants, Pediatrics 76 (4) (1985) 585–592.
BIBLIOGRAPHY 211
[246] D. L. Shapiro, R. H. Notter, F. C. Morin III, Double-blind, randomized trial ofa calf lung surfactant extract administered at birth to very premature infants forprevention of respiratory distress syndrome, Pediatrics 76 (4) (1985) 593–599.
[247] J. M. Davis, K. Veness-Meehan, R. H. Notter, V. K. Bhutani, J. W. Kendig, D. L.Shapiro, Changes in pulmonary mechanics after the administration of surfactantto infants with respiratory distress syndrome, New England Journal of Medicine319 (8) (1988) 476–479.
[248] J. W. Kendig, R. H. Notter, C. Cox, J. L. Aschner, S. Benn, R. M. Bernstein,K. Hendricks-Munoz, W. M. Maniscalco, L. A. Metlay, D. L. Phelps, R. A. Sinkin,B. P. Wood, D. L. Shapiro, Surfactant replacement therapy at birth final analysis ofa clinical trial and comparisons with similar trials, Pediatrics 82 (5) (1988) 756–762.
[249] J. W. Kendig, R. H. Notter, D. C. Cox, L. J. Reubens, J. M. Davis, W. M. Manis-calco, R. A. Sinkin, A. Bartoletti, H. S. Dweck, M. J. Horgan, H. Risemberg, D. L.Phelps, D. L. Shapiro, A comparison of surfactant as immediate prophylaxis and asrescue therapy in newborns of less than 30 weeks’ gestation, New England Journalof Medicine 324 (13) (1991) 865–871.
[250] R. L. Auten, R. H. Notter, J. W. Kendig, J. M. Davis, D. L. Shapiro, Surfactanttreatment of full-term newborns with respiratory failure, Pediatrics 87 (1) (1991)101–107.
[251] J. Kattwinkel, B. T. Bloom, P. M. Delmore, C. L. Davis, E. E. Farrell, H. Friss,A. L. Jung, K. King, D. Mueller, Prophylactic administration of calf lung surfactantextract is more effective than early treatment of respiratory distress syndrome inneonates of 29 through 32 weeks’ gestation, Pediatrics 92 (1) (1993) 90–98.
[252] M. L. Hudak, E. E. Farrell, A. A. Rosenberg, A. L. Jung, R. L. Auten, D. J. Durand,M. J. Horgan, S. Buckwald, M. R. Belcastro, P. K. Donohue, V. Carrion, W. W.Maniscalco, M. J. Balsan, B. A. Torres, R. R. Miller, R. D. Jansen, J. E. Graeber,K. M. Laskay, E. J. Matteson, A multicenter randomized, masked comparisontrial of natural versus synthetic surfactant for the treatment of respiratory distresssyndrome, Journal of Pediatrics 128 (3) (1996) 396–406.
[253] D. F. Willson, J. H. Jiao, L. A. Bauman, A. Zaritsky, H. Craft, K. Dockery, D. Con-rad, H. Dalton, Calf’s lung surfactant extract in acute hypoxemic respiratory failurein children, Critical Care Medicine 24 (8) (1996) 1316–1322.
[254] M. L. Hudak, D. J. Martin, E. A. Egan, E. J. Matteson, J. Cummings, A. L. Jung,L. V. Kimberlin, R. L. Auten, A. A. Rosenberg, J. M. Asselin, M. R. Belcastro,P. K. Donohue, C. R. Hamm Jr., R. D. Jansen, A. S. Brody, M. M. Riddlesberger,P. Montgomery, A multicenter randomized masked comparison trial of syntheticsurfactant versus calf lung surfactant extract in the prevention of neonatal respi-ratory distress syndrome, Pediatrics 100 (1) (1997) 39–50.
[255] B. T. Bloom, J. Kattwinkel, R. T. Hall, P. M. Delmore, E. A. Egan, J. R. Trout,M. H. Malloy, D. R. Brown, I. R. Holzman, C. H. Coghill, W. A. Carlo, A. K.
BIBLIOGRAPHY 212
Pramanik, M. A. McCaffree, P. L. Toubas, S. Laudert, L. L. Gratny, K. B. Weath-erstone, J. H. Seguin, L. D. Willett, G. R. Gutcher, D. H. Mueller, W. H. Topper,Comparison of Infasurf (calf lung surfactant extract) to Survanta (beractant) inthe treatment and prevention of respiratory distress syndrome, Pediatrics 100 (1)(1997) 31–38.
[256] J. W. Kendig, R. M. Ryan, R. A. Sinkin, W. M. Maniscalco, R. H. Notter, R. Guil-let, C. Cox, H. S. Dweck, M. J. Horgan, L. J. Reubens, H. Risemberg, D. L. Phelps,Comparison of two strategies for surfactant prophylaxis in very premature infants:A multicenter randomized trial, Pediatrics 101 (6) (1998) 1006–1012.
[257] D. F. Willson, A. Zaritsky, L. A. Bauman, K. Dockery, R. L. James, D. Conrad,H. Craft, W. E. Novotny, E. A. Egan, H. Dalton, Instillation of calf lung surfactantextract (calfactant) is beneficial in pediatric acute hypoxemic respiratory failure,Critical Care Medicine 27 (1) (1999) 188–195.
[258] G. Noack, P. Berggren, T. Curstedt, G. Grossmann, P. Herin, W. Mortensson,R. Nilsson, B. Robertson, Severe neonatal respiratory distress syndrome treatedwith the isolated phospholipid fraction of natural surfactant, Acta PaediatricaScandinavica 76 (5) (1987) 697–705.
[259] B. Robertson, G. Noack, J. Kok, J. Koppe, L. V. Sonderen, H. L. Halliday, G. Mc-Clure, B. McCord, M. Reid, A. Okken, C. P. Speer, W. Schroter, B. Andreason,N. Svenningsen, J. P. Relier, H. Walti, G. Bevilacqua, E. V. Cosmi, L. Gaioni,Surfactant replacement therapy for severe neonatal respiratory distress syndrome:An international randomized clinical trial, Pediatrics 82 (5) (1988) 683–691.
[260] B. Robertson, Factors influencing the clinical response to surfactant replacementtherapy in babies with severe respiratory distress syndrome, European Journal ofPediatrics 150 (6) (1991) 433–439.
[261] J. Egberts, J. P. D. Winter, G. Sedin, M. J. K. D. Kleine, U. Broberger, F. V. Bel,T. Curstedt, B. Robertson, Comparison of prophylaxis and rescue treatment withcurosurf in neonates less than 30 weeks’ gestation: A randomized trial, Pediatrics92 (6) (1993) 768–774.
[262] M. Rollins, J. Jenkins, R. Tubman, C. Corkey, D. Wilson, Comparison of clinicalresponses to natural and synthetic surfactants, Journal of Perinatal Medicine 21 (5)(1993) 341–347.
[263] C. P. Speer, O. Gefeller, P. Groneck, E. Laufkotter, C. Roll, L. Hanssler, I. Harms,E. Herting, H. Boenisch, J. Windeler, B. Robertson, Randomised clinical trial oftwo treatment regimens of natural surfactant preparations in neonatal respiratorydistress syndrome, Archives of Disease in Childhood 72 (1 SUPPL.) (1995) F8–F13.
[264] G. Bevilacqua, S. Parmigiani, B. Robertson, G. Caramia, P. Catalani, F. Chi-appe, G. Rinaldi, R. Magaldi, F. Pantarotto, G. Spennati, S. Calo, G. F. Perotti,L. Gaioni, G. Compagnoni, E. Corbella, V. Tripodi, S. Grano, N. Cassata, G. Sul-liotti, L. Gambini, P. Gancia, P. Serrao, A. Nicolo, G. Bonacini, C. Romagnoli,
BIBLIOGRAPHY 213
M. T. Gandolfo, G. D. Nisi, A. Mazza, F. Uxa, P. Monici-Preti, F. Gardini, Pro-phylaxis of respiratory distress syndrome by treatment with modified porcine sur-factant at birth: A multicentre prospective randomized trial, Journal of PerinatalMedicine 24 (6) (1996) 609–620.
[265] H. W. Taeusch, K. M. W. Keough, M. Williams, Characterization of bovine sur-factant for infants with respiratory distress syndrome, Pediatrics 77 (4) (1986)572–581.
[266] T. N. K. Raju, D. Vidyasagar, R. Bhat, Double-blind controlled trial of single-dose treatment with bovine surfactant in severe hyaline membrane disease, Lancet1 (8534) (1987) 651–655.
[267] J. D. Gitlin, R. F. Soll, R. B. Parad, Randomized controlled trial of exogenoussurfactant for the treatment of hyaline membrane disease, Pediatrics 79 (1) (1987)31–37.
[268] M. Konishi, T. Fujiwara, T. Naito, Y. Takeuchi, Y. Ogawa, K. Inukai, M. Fujimura,M. Nakamura, T. Hashimoto, Surfactant replacement therapy in neonatal respira-tory distress syndrome - a multi-centre, randomized clinical trial: Comparison ofhigh- versus low-dose of surfactant TA, European Journal of Pediatrics 147 (1)(1988) 20–25.
[269] T. Fujiwara, M. Konishi, S. Chida, K. Okuyama, Y. Ogawa, Y. Takeuchi,H. Nishida, H. Kito, M. Fujimura, H. Nakamura, T. Hashimoto, Surfactant re-placement therapy with a single postventilatory dose of a reconstituted bovinesurfactant in preterm neonates with respiratory distress syndrome: Final analysisof a multicenter, double-blind, randomized trial and comparison with similar trials,Pediatrics 86 (5) (1990) 753–764.
[270] M. Konishi, T. Fujiwara, S. Chida, H. Maeta, S. Shimada, T. Kasai, Y. Fuji,Y. Murakami, A prospective, randomized trial of early versus late administrationof a single dose of surfactant-TA, Early Human Development 29 (1-3) (1992) 275–282.
[271] J. D. Horbar, R. F. Soll, J. M. Sutherland, U. Kotagal, A. G. S. Philip, D. L.Kessler, G. A. Little, W. H. Edwards, D. Vidyasagar, T. N. K. Raju, A. H. Jobe,M. Ikegami, M. D. Mullett, D. Z. Myerberg, T. L. McAuliffe, J. F. Lucey, A mul-ticenter randomized, placebo-controlled trial of surfactant therapy for respiratorydistress syndrome, New England Journal of Medicine 320 (15) (1989) 959–965.
[272] J. D. Horbar, R. F. Soll, H. Schachinger, G. Kewitz, H. T. Versmold, W. Lindner,G. Duc, D. Mieth, O. Linderkamp, E. P. Zilow, P. Lemburg, V. V. Loewenich,M. Brand, I. Minoli, G. Moro, K. P. Riegel, R. Roos, L. Weiss, J. F. Lucey, Aeuropean multicenter randomized controlled trial of single dose surfactant therapyfor idiopathic respiratory distress syndrome, European Journal of Pediatrics 149 (6)(1990) 416–423.
BIBLIOGRAPHY 214
[273] R. F. Soll, R. E. Hoekstra, J. J. Fangman, A. J. Corbet, J. M. Adams, L. S. James,K. Schulze, W. Oh, J. D. Roberts Jr., J. P. Dorst, S. S. Kramer, A. J. Gold, E. M.Zola, J. D. Horbar, T. L. McAuliffe, J. F. Lucey, Multicenter trial of single-dosemodified bovine surfactant extract (Survanta) for prevention of respiratory distresssyndrome, Pediatrics 85 (6) (1990) 1092–1102.
[274] R. E. Hoekstra, J. C. Jackson, T. F. Myers, I. D. Frantz III, M. E. Stern, W. F.Powers, M. Maurer, J. R. Raye, S. T. Carrier, J. H. Gunkel, A. J. Gold, Improvedneonatal survival following multiple doses of bovine surfactant in very prematureneonates at risk for respiratory distress syndrome, Pediatrics 88 (1) (1991) 10–18.
[275] E. A. Liechty, E. F. Donovan, D. Purohit, J. Gilhooly, B. Feldman, A. Noguchi,S. E. Denson, S. S. Sehgal, I. Gross, D. Stevens, M. Ikegami, R. D. Zachman, S. T.Carrier, J. H. Gunkel, A. J. Gold, Reduction of neonatal mortality after multipledoses of bovine surfactant in low birth weight neonates with respiratory distresssyndrome, Pediatrics 88 (1) (1991) 19–28.
[276] J. D. Horbar, L. L. Wright, R. F. Soll, E. C. Wright, A. A. Fanaroff, S. B. Korones,S. Shankaran, W. Oh, B. D. Fletcher, C. R. Bauer, J. E. Tyson, J. A. Lemons, E. F.Donovan, B. J. Stoll, D. D. Stevenson, L. A. Papile, J. Philips III, A multicenterrandomized trial comparing two surfactants for the reatment of neonatal respiratorydistress syndrome, Journal of Pediatrics 123 (5) (1993) 757–766.
[277] S. S. Sehgal, C. K. Ewing, T. Richards, H. W. Taeusch, Modified bovine surfactant(Survanta) versus a protein-free surfactant (Exosurf) in the treatment of respiratorydistress syndrome in preterm infants: A pilot study, Journal of the National MedicalAssociation 86 (1) (1994) 46–52.
[278] A. Hamvas, F. Sessions-Cole, D. E. DeMello, M. Moxley, J. A. Whitsett, H. R.Colten, L. M. Nogee, Surfactant protein B deficiency: Antenatal diagnosis andprospective treatment with surfactant replacement, Journal of Pediatrics 125 (3)(1994) 356–361.
[279] R. F. Soll, A multicenter, randomized trial comparing synthetic surfactant withmodified bovine surfactant extract in the treatment of neonatal respiratory distresssyndrome, Pediatrics 97 (1) (1996) 1–6.
[280] B. R. B. Robertson, T. Curstedt, J. Johansson, Prospects for a new syntheticsurfactant, Acta Biomedica de l’Ateneo Parmense 71 (SUPPL. 1) (2000) 409–412.
[281] C. W. Wu, S. L. Seurynck, K. Y. C. Lee, A. E. Barron, Helical peptoid mimics oflung surfactant protein C, Chemistry and Biology 10 (11) (2003) 1057–1063.
[282] S. L. Seurynck-Servoss, M. T. Dohm, A. E. Barron, Effects of including an N-terminal insertion region and arginine-mimetic side chains in helical peptoid ana-logues of lung surfactant protein B, Biochemistry 45 (39) (2006) 11809–11818.
[283] J. Johansson, M. Gustafson, M. Palmblad, S. Zaltash, B. Robertson, T. Curstedt,Pulmonary surfactant: Emerging protein analogues, BioDrugs 11 (2) (1999) 71–77.
BIBLIOGRAPHY 215
[284] T. Curstedt, J. Johansson, New synthetic surfactants: Basic science, Biology of theNeonate 87 (4) (2005) 332–337.
[285] R. H. Pfister, R. F. Soll, New synthetic surfactants: The next generation, Biologyof the Neonate 87 (4) (2005) 338–344.
[286] T. Curstedt, J. Johansson, New synthetic surfactants: How and when, Biology ofthe Neonate 89 (4) (2006) 336–339.
[287] F. J. Walther, A. J. Waring, M. A. Sherman, J. A. Zasadzinski, L. M. Gordon,Hydrophobic surfactant proteins and their analogues, Neonatology 91 (4) (2007)303–310.
[288] I. Mingarro, D. Lukovic, M. Vilar, J. Perez-Gil, Synthetic pulmonary surfactantpreparations: New developments and future trends, Current Medicinal Chemistry15 (4) (2008) 393–403.
[289] C. J. Morley, A. D. Bangham, N. Miller, J. A. Davis, Dry artificial lung surfactantand its effect on very premature babies, Lancet 1 (8211) (1981) 64–68.
[290] A. D. Milner, H. Vyas, I. E. Hopkin, Effects of artificial surfactant on lung functionand blood gases in idiopathic respiratory distress syndrome, Archives of Disease inChildhood 58 (6) (1983) 458–460.
[291] D. J. Durand, R. I. Clyman, M. A. Heymann, Effects of a protein-free, syntheticsurfactant on survival and pulmonary function in preterm lambs, Journal of Pedi-atrics 107 (5) (1985) 775–780.
[292] C. Bose, A. J. Corbet, G. Bose, J. Garica-Prats, L. Lombardy, D. Wold, D. Donlon,W. Long, Improved outcome at 28 days of age for very low birth weight infantstreated with a single dose of a synthetic surfactant, Journal of Pediatrics 117 (6)(1990) 947–953.
[293] W. Long, A. J. Corbet, R. Cotton, S. Courtney, G. McGuiness, D. Walter, J. Watts,J. A. Smyth, H. Bard, V. Chernick, D. Stevenson, S. Goldman, F. Walther,M. Mammel, S. Boros, T. Thompson, R. Bucciarelli, D. Burchfield, R. Guthrie,A controlled trial of synthetic surfactant in infants weighing 1250 g or more withrespiratory distress syndrome, New England Journal of Medicine 325 (24) (1991)1696–1703.
[294] W. Long, T. Thompson, H. Sundell, R. Schumacher, F. Volberg, R. Guthrie,D. Stevenson, S. Goldman, F. Walther, S. Boros, F. Mammel, R. Bucciarelli,D. Burchfield, A. J. Corbet, G. McGuiness, M. Mullett, R. Cotton, R. Vaughan,M. LeBlanc, Effects of two rescue doses of a synthetic surfactant on mortality rateand survival without bronchopulmonary dysplasia in 700- and 1350-gram infantswith respiratory distress syndrome, Journal of Pediatrics 118 (4 I) (1991) 595–605.
[295] R. H. Phibbs, R. A. Ballard, J. A. Clements, D. C. Heilbron, C. S. Phibbs, M. A.Schlueter, S. H. Sniderman, W. H. Tooley, A. Wakeley, Initial clinical trial of
BIBLIOGRAPHY 216
Exosurf, a protein-free synthetic surfactant, for the prophylaxis and early treatmentof hyaline membrane disease, Pediatrics 88 (1) (1991) 1–9.
[296] A. J. Corbet, R. Bucciarelli, S. Goldman, M. Mammel, D. Wold, W. Long, De-creased mortality rate among small premature infants treated at birth with a singledose of synthetic surfactant: A multicenter controlled trial, Journal of Pediatrics118 (2) (1991) 277–284.
[297] I. Y. Haddad, J. P. Crow, P. Hu, Y. Yaozu, J. Beckman, S. Matalon, Concurrentgeneration of nitric oxide and superoxide damages surfactant protein A, AmericanJournal of Physiology 267 (3 part 1) (1994) L242–L249.
[298] S. A. Rooney, S. L. Young, C. R. Mendelson, Molecular and cellular processing oflung surfactant, Federation of American Societies for Experimental Biology Journal8 (12) (1994) 957–967.
[299] M. L. Choukroun, B. Llanas, H. Apere, M. Fayon, R. I. Galperine, H. Guenard,J. L. Demarquez, Pulmonary mechanics in ventilated preterm infants with respi-ratory distress syndrome after exogenous surfactant administration: A comparisonbetween two surfactant preparations, Pediatric Pulmonology 18 (5) (1994) 273–278.
[300] A. Anzueto, R. P. Baughman, K. K. Guntupalli, J. G. Weg, H. P. Wiedemann, A. A.Raventos, F. Lemaire, W. Long, D. S. Zaccardelli, E. N. Pattishall, Aerosolizedsurfactant in adults with sepsis-induced acute respiratory distress syndrome, NewEngland Journal of Medicine 334 (22) (1996) 1417–1421.
[301] I. W. Levin, Interactions of model human pulmonary surfactants with a mixedphospholipid bilayer assembly: Raman spectroscopic studies, Biochemistry 32 (32)(1993) 8228–8238.
[302] C. G. Cochrane, S. D. Ravak, Protein-phospholipid interactions in pulmonary sur-factant, Chest 105 (3 SUPPL.) (1994) 57S–62S.
[303] T. A. Merritt, A. Kheiter, C. G. Cochrane, Positive end-expiratory pressure duringKL4 surfactant instillation enhances intrapulmonary distribution in a simian modelof respiratory distress syndrome, Pediatric Research 38 (2) (1995) 211–217.
[304] C. G. Cochrane, S. D. Revak, T. A. Merritt, G. P. Heldt, M. Hallman, M. D.Cunningham, D. Easa, A. Pramanik, D. K. Edwards, M. S. Alberts, The efficacyand safety of KL4-surfactant in preterm infants with respiratory distress syndrome,American Journal of Respiratory and Critical Care Medicine 153 (1) (1996) 404–410.
[305] S. D. Revak, T. A. Merritt, C. G. Cochrane, G. P. Heldt, M. S. Alberts, D. W.Anderson, A. Kheiter, Efficacy of synthetic peptide-containing surfactant in thetreatment of respiratory distress syndrome in preterm infant rhesus monkeys, Pe-diatric Research 39 (4 I) (1996) 715–724.
BIBLIOGRAPHY 217
[306] J. Ma, S. Koppoenol, H. Yu, G. Zografi, Effects of a cationic and hydrophobicpeptide, KL4, on model lung surfactant lipid monolayers, Biophysical Journal 74 (4)(1998) 1899–1907.
[307] F. J. Walther, J. M. Hernandez-Juviel, R. Bruni, A. J. Waring, Protein compositionof synthetic surfactant affects gas exchange in surfactant-deficient rats, PediatricResearch 43 (5) (1998) 666–673.
[308] V. Balaraman, J. Meister, T. L. Ku, S. L. Sood, E. Tam, J. Killeen, C. F. T.Uyehara, E. A. Egan, D. Easa, Lavage administration of dilute surfactants afteracute lung injury in neonatal piglets, American Journal of Respiratory and CriticalCare Medicine 158 (1) (1998) 12–17.
[309] C. G. Cochrane, S. D. Revak, T. A. Merritt, I. U. Schraufstatter, R. C. Hoch,C. Henderson, S. Andersson, H. Takamori, Z. G. Oades, Bronchoalveolar lavagewith KL4-surfactant in models of meconium aspiration syndrome, Pediatric Re-search 44 (5) (1998) 705–715.
[310] G. Nilsson, M. Gustafsson, G. Vandenbussche, E. Veldhuizen, W. J. Griffiths,J. Sjovall, H. P. Haagsman, J. M. Ruysschaert, B. Robertson, T. Curstedt, J. Jo-hansson, Synthetic peptide-containing surfactants: Evaluation of transmembraneversus amphipathic helices and surfactant protein C poly-valyl to poly-leucyl sub-stitution, European Journal of Biochemistry 255 (1) (1998) 116–124.
[311] T. E. Wiswell, R. M. Smith, L. B. Katz, L. Mastroianni, D. Y. Wong, D. Willms,S. Heard, M. Wilson, R. D. Hite, A. Anzueto, S. D. Revak, C. G. Cochrane,Bronchopulmonary segmental lavage with surfaxin (KL4-surfactant) for acute res-piratory distress syndrome, American Journal of Respiratory and Critical CareMedicine 160 (4) (1999) 1188–1195.
[312] S. K. Sinha, T. Lacaze-Masmonteil, A. V. I. Soler, T. E. Wiswell, J. Gadzinowski,J. Hajdu, G. Bernstein, M. Sanchez-Luna, R. Segal, C. J. Schaber, J. Massaro,R. D’Agostino, A multicenter, randomized, controlled trial of lucinactant versusporactant alfa among very premature infants at high risk for respiratory distresssyndrome, Pediatrics 115 (4) (2005) 1030–1038.
[313] A. Saenz, O. Canadas, L. A. Bagatolli, M. E. Johnson, C. Casals, Physical proper-ties and surface activity of surfactant-like membranes containing the cationic andhydrophobic peptide KL4, Federation of European Biochemical Societies Journal273 (11) (2006) 2515–2527.
[314] A. Saenz, O. Canadas, L. A. Bagatolli, F. Sanchez-Barbero, M. E. Johnson,C. Casals, Effect of surfactant protein A on the physical properties and surfaceactivity of KL4-surfactant, Biophysical Journal 92 (2) (2007) 482–492.
[315] S. Hawgood, A. Ogawa, K. Yukitake, M. A. Schlueter, C. Brown, T. White,D. Buckley, D. Lesikar, B. J. Benson, Lung function in premature rabbits treatedwith recombinant human surfactant protein-C, American Journal of Respiratoryand Critical Care Medicine 154 (2) (1996) 484–490.
BIBLIOGRAPHY 218
[316] A. J. Davis, A. H. Jobe, D. Hafner, M. Ikegami, Lung function in premature lambsand rabbits treated with a recombinant SP-C surfactant, American Journal ofRespiratory and Critical Care Medicine 157 (2) (1998) 553–559.
[317] D. Hafner, P. G. Germann, D. Hauschke, Comparison of rSP-C surfactant withnatural and synthetic surfactants after late treatment in a rat model of the acuterespiratory distress syndrome, British Journal of Pharmacology 124 (6) (1998)1083–1090.
[318] D. Hafner, P. G. Germann, D. Hauschke, Effects of rSP-C surfactant on oxygenationand histology in a rat-lung- lavage model of acute lung injury, American Journalof Respiratory and Critical Care Medicine 158 (1) (1998) 270–278.
[319] M. Ikegami, A. H. Jobe, Surfactant protein-C in ventilated premature lamb lung,Pediatric Research 44 (6) (1998) 860–864.
[320] R. G. Spragg, D. Levin, ARDS and the search for meaningful subgroups, IntensiveCare Medicine 26 (7) (2000) 835–837.
Note: All references starting with Ref. 226 are only used in Appendix A.
Appendix A
Commercial Lung Surfactants∗
Several classification systems are used for commercial exogenous lung extracts [1, 174,
175, 226]. Figure A.1 divide different commercial lung extracts into two main categories
according to the source of the preparation. The first category is natural lung surfactants
which contains endogenous lung extracts from animals. The second category is synthetic
lung surfactants which does not contain any natural extracts. Generally, preclinical an-
imal testing and clinical trials suggest that the performance of natural animal-derived
preparations is much better than synthetic preparations [227–229]. However, batch to
batch variations, low content of surfactant proteins and high cost are among the limita-
tions of natural animal-derived preparations [230, 231].
A.1 Natural Surfactant Extracts
Natural surfactants [230–234] can also be classified into three categories. The first con-
tains exogenous surfactants extracted by animal lung lavage (Alveofact, BLES, Infasurf).
The general extraction procedure is summarized in figure A.2. The second group contains
exogenous surfactants processed from animal lung tissue (Curosurf). The third group is
∗References used here are contained in the bibliography list at the end of the body of the thesis.
219
Appendix A. Commercial Lung Surfactants 220
Commercial Lung Extracts
Ani
mal
Lung
La
vage
BLES
Alveofact
Infasurf
Natural Synthetic
Surfacten
Curosurf
Survanta
Exosurf
ALEC
Surfaxin
Venticute
Ani
mal
Lung
Ti
ssue
Supp
lem
ente
d Lu
ng T
issue
Prot
ein
Free
Sim
plifi
ed
Pept
ides
Prot
ein
Ana
logs
Figure A.1: General classification of commercial lung extracts.
Appendix A. Commercial Lung Surfactants 221
Intact Bovine LungsIntact Cow Lungs Intact Calf Lungs
BLESAlveofact Infasurf
BLSE CLSE
Endogenous Lung Surfactant
Endogenous Lung Surfactant
Endogenous Lung Surfactant
Bronchoalveolar Lavage (with 0.15 M NaCl)
cells removed by low-speed centrifugation (150 - 200 g)
pelleted by high-speed centrifugation (10000 - 12000 g)
extract with Chloroform-Methanol (remove SP-A)
extract with Acetone (deplete Cholesterol and NL)
Bronchoalveolar Lavage (with 0.15 M NaCl)
cells removed by low-speed centrifugation (150 - 200 g)
pelleted by high-speed centrifugation (10000 - 12000 g)
Bronchoalveolar Lavage (with 0.15 M NaCl)
cells removed by low-speed centrifugation (150 - 200 g)
pelleted by high-speed centrifugation (10000 - 12000 g)
extract with Chloroform-Methanol (remove SP-A)
extract with Chloroform-Methanol (remove SP-A)
Figure A.2: Extraction procedure for surfactants extract of lung lavage.
similar to the second one; however, these surfactants are supplemented with synthetic
additives (Surfacten, Survanta). The general extraction procedure of second and third
categories is summarized in figure A.3. A comparison of the composition of the six nat-
ural surfactant extracts is shown in figure A.4. More details about natural surfactants
can be found in a recent review [235].
A.1.1 Alveofact
Alveofact (SF-RI 1 or bovactant) is a clinical exogenous lung surfactant obtained from
bronchoalveolar lavage of bovine lung surfactant [236–240]. The lavage is pelleted by
centrifugation, followed by extraction with chloroform methanol to remove hydrophillic
surfactant protein (see figure A.2). It contains weight distribution of 99% phospholipids
Appendix A. Commercial Lung Surfactants 222
Bovine Lung TissuePorcine Lung Tissue Bovine Lung Tissue
SurfactenCurosurf Survanta
Lung Tissue Extract Lung Tissue Extract Lung Tissue Extract
homogenize or mince (in 0.15 M NaCl)
differential centrifugation and flotation steps
extract with ethyl acetate (reduce neutral lipids)
extract with Chloroform-Methanol (remove ethyl actetate)
supplement with synthetic DPPC,
Palmitic Acid, and Tripalmitin
homogenize or mince (in 0.15 M NaCl)
wash and centrifuge (1,000 - 3,000 g)
extract with Chloroform-Methanol (2:1)
supplement with synthetic DPPC,
Palmitic Acid, and Tripalmitin
liquid-gel affinity chromatography
homogenize or mince (in 0.15 M NaCl)
differential centrifugation and flotation steps
extract with ethyl acetate (reduce neutral lipids)
extract with Chloroform-Methanol (remove ethyl actetate)
Figure A.3: Extraction procedure for surfactants extract of lung tissue.
Appendix A. Commercial Lung Surfactants 223
PL/NL: 95%
Cholesterol: 4%SP−B/SP−C: 1%
(a) Alveofact
PC: 78%
Other PL: 21%
SP−B/SP−C: 1%
(b) BLES
PC: 77%
PG: 6%
PI/PS: 4%
PE/SPH: 5%
Other PL: 2%Cholesterol/NL: 5%SP−B/SP−C: 1%
(c) Infasurf
PC: 70%
PE/SPH: 17%
Other PL: 12%SP−B/SP−C: 1%
(d) Curosurf
PL: 84%
Palmitic Acid: 8%
Tripalmitin: 7%SP−B/SP−C: 1%
(e) Surfacten
PL: 84%
Palmitic Acid: 8%
Tripalmitin: 7%SP−B/SP−C: 1%
(f) Survanta
Figure A.4: Composition of natural surfactants extracts.
Appendix A. Commercial Lung Surfactants 224
(PL) and neutral lipids (NL) [that include 4% cholesterol], and 1% hydrophobic surfactant
proteins SP-B and SP-C as shown in figure A.4(a). It is formed clinically as a sterile
suspension in 0.15 M NaCl at 45 mg/ml with dosing amount of 1.2 ml/kg. Alveofact
is produced by Thomae GmbH, Biberach, Germany; and Boehringer Ingelheim Co.,
Ingelheim, Germany.
A.1.2 BLES
BLES (bovine lipid extract surfactant) is a clinical exogenous lung surfactant made from
of lung surfactant lavaged from cows [241–244]. The lavage is pelleted by centrifugation,
followed by extraction with chloroform methanol to give BLSE (bovine lung surfactant
extract). Additional extraction with acetone is used to remove cholesterol and other
neutral lipids (see figure A.2). It contains weight distribution of 98-99% phospholipids
(PL) [that include 79% phophophatidyl-choline (PC)], and 1% hydrophobic surfactant
proteins SP-B and SP-C as shown in figure A.4(b). It is formed clinically as a sterile
suspension in 0.15 M NaCl and 1.5 mM CaCl2 at 25 mg/ml with dosing amount of 5
ml/kg (135 mg/kg), 1-2 doses every 8 h. BLES is produced by BLES Biochemicals Inc.,
London, Ontario, Canada.
A.1.3 Infasurf
Infasurf (CLSE or calfactant) is a clinical exogenous lung surfactant obtained from calves
lung lavage (CLSE) [245–257]. The lavage is pelleted by centrifugation at 10,000 - 12,000
×g, followed by extraction with 2:1 chloroform methanol to give CLSE or calf lung
surfactant extract (see figure A.2). It contains weight distribution of 93% phospholipids
(PL) [that include 83% phophophatidyl-choline (PC), 6% phosphatidyl-glycerol (PG), 4%
phosphatidyl-inositol (PI) and phosphatidyl-serine (PS), 3% phosphatidyl-ethanolamine
(PE), and 2% sphinogomyelin (SPH)], 5% cholesterol and neutral lipids (NL), and 1.5%
hydrophobic surfactant proteins SP-B and SP-C as shown in figure A.4(c). Desaturated
Appendix A. Commercial Lung Surfactants 225
PC is approximately 50% of the total phopholipids. It is formed clinically as a heat-
sterilized suspension in 0.15 M NaCl at 35 mg/ml with dosing amount of 3 ml/kg (105
mg/kg), 1-4 doses every 6-12 h. Infasurf is produced by ONY Inc., Amherst, New York,
USA (for Forest Pharmaceuticals Inc., St. Louis, Missouri, USA).
A.1.4 Curosurf
Curosurf (poractant alpha) is a clinical exogenous lung surfactant obtained from minced
porcine lung tissue [258–264]. The process include washing, centrifugation at 1,000 - 3,000
×g, followed by extraction with 2:1 chloroform methanol, and liquid-gel affinity chro-
matography (see figure A.3). It contains weight distribution of 93% phospholipids (PL)
[that include 67-75% phophophatidyl-choline (PC), 12-22% phosphatidyl-ethanolamine
(PE) and sphinogomyelin (SPH)], and 1% hydrophobic surfactant proteins SP-B and
SP-C as shown in figure A.4(d). It is formed clinically as a suspension sterilized by high
pressure filter system (0.45 and 0.2 µm filters), dried or lypohilized and stored in vials
then suspended in saline by sonication at 80 mg/ml prior to use with dosing amount
of 2.5 ml/kg (200 mg/kg) to 1.25 ml/kg (100 mg/kg). Curosurf is produced by Chiesi
Farmaceutici, Parma, Italy.
A.1.5 Surfacten
Surfacten (Surfactant TA) is a clinical exogenous lung surfactant made by organic sol-
vent extraction of finely ground bovine lung tissue supplemented with synthetic DPPC,
palmitic acid, and tripalmitin [265–270]. The lung tissue goes through several centrifuga-
tion and floatation processes, extraction with ethyl acetate to reduce neutral lipids, and
additional extraction with chloroform methanol to remove ethyl acetate. The extract
is then supplemented with the aforementioned synthetic additives (see figure A.3). The
final product contains weight distribution of 84% phospholipids (PL), 8% palmitic acid,
7% tripalmitin, and 1% hydrophobic surfactant proteins SP-B and SP-C as shown in
Appendix A. Commercial Lung Surfactants 226
figure A.4(e). It is formed clinically as a suspension sterilized by high pressure filter,
lypohilized and stored in vials then suspended in sterile 0.15 M NaCl at 30 mg/ml prior
to use with dosing amount of 4 ml/kg (100 mg/kg), 1-4 doses every 6 h. Surfacten is
produced by Mitsubishi Tanabe Pharma Corporation (formerly Tokyo Tanabe Co. Ltd.),
Tokyo, Japan.
A.1.6 Survanta
Survanata (beractant) is a clinical exogenous lung surfactant made by organic solvent
extraction of processed and supplemented bovine lung tissue [230, 255, 263, 271–279]. It
is prepared by similar methods and identical synthetic additives as in Surfactant TA (see
figure A.3). The final product contains hydrophobic surfactant proteins slightly less than
1% (include a small amount of SP-B) as shown in figure A.4(f). It is formed clinically as
an autoclave-sterilized suspension in 0.15 M NaCl at 25 mg/ml with dosing amount of 4
ml/kg (100 mg/kg), 1-4 doses every 6 h. Survanta is produced by Abbott Laboratories,
Abbott Park, Illinois, USA.
A.2 Synthetic Surfactants
Synthetic Surfactants [280–282] can also be classified into three categories. The first
contains synthetic exogenous surfactants with no proteins (ALEC, Exosurf). The sec-
ond group contains synthetic exogenous surfactants with synthetic simplified peptides
(Surfaxin). The third group contains synthetic exogenous surfactants with recombinant
human apoproteins (Venticute). A comparison of the composition of the four synthetic
surfactants is shown in figure A.5. More details about synthetic surfactants can be found
in recent reviews [280, 283–288].
Appendix A. Commercial Lung Surfactants 227
DPPC: 70%
egg PG: 30%
(a) ALEC
DPPC: 85%
Hexadecanol: 9%
Tyloxapol: 6%
(b) Exosurf
DPPC: 64%
POPG: 21%
Palmitic Acid: 13%
Sinapultide (KL4): 2%
(c) Surfaxin
DPPC: 65%
POPG: 28%
Palmitic Acid: 5%rSP−C: 2%
(d) Venticute
Figure A.5: Composition of synthetic surfactants.
Appendix A. Commercial Lung Surfactants 228
A.2.1 ALEC
ALEC (artificial lung expanding compound) is a clinical exogenous lung surfactant con-
taining synthetic dipalmitoyl-phosphatidyl-choline (DPPC) and egg phosphatidyl-glycerol
(PG) in a 7:3 molar ratio [79–81, 289, 290] as shown in figure A.5(a). DPPC is the main
component for lowering surface tension, while egg PG is to improve adsorption and
spreading. It is formed clinically as a fine white powder mixed (below 30C) with sterile
0.15 M NaCl. It is produced by Britannia Pharmaceuticals Ltd., Redhill, Surry, United
Kingdom. ALEC was withdrawn from the market few years ago.
A.2.2 Exosurf
Exosurf (colfosceril palmitate) is a clinical exogenous lung surfactant containing synthetic
dipalmitoyl-phosphatidyl-choline (DPPC), hexadecanol, and tyloxapol in a 1:0.111:0.075
weight ratio [252, 254, 279, 291–300] as shown in figure A.5(b). DPPC is the main
component for lowering surface tension, while the nonionic detergent and the C16:0
alchohol hexadeconal are to improve adsorption and spreading. It is formed clinically
as a sterile lyophilized powder in vaccum-sealed vials with 0.1 M NaCl then suspended
in 8 ml of distilled water prior to use with dosing amount of 5 ml/kg (67.5 mg/kg), 1-4
doses every 12 h. It is produced by GlaxoSmithKline (formerly Burroughs Wellcome),
Research Triangle Park, North Carolina, USA. Exosurf is no longer available in USA and
Canada.
A.2.3 Surfaxin
Surfaxin (KL4 or lucinactant) is a clinical exogenous lung surfactant containing syn-
thetic dipalmitoyl-phosphatidyl-choline (DPPC), palmitoyl-oleoyl-phosphatidyl-glycerol
(POPG), palmitic acid, and a 21 amino acid synthetic hydrophobic peptide with repeated
sequences of one lysine (K) and four leucine (L) residues [82–84, 301–314]. DPPC and
Appendix A. Commercial Lung Surfactants 229
POPG are present in 3:1 weight ratio. Palmitic acid is present at 15% by weight relative
to phospholipids, while the hydrophobic peptide (known as KL4 or sinapultide) is present
at 3% by weight as shown in figure A.5(c). It is produced by Discovery Laboratories Inc.,
Warrington, Pennsylvania, USA. Recently, an aerosolized version of Surfaxin was devel-
oped by the same producer called Aerosurf. Surfaxin and Aerosurf are investigational
drug products, and are not approved for use yet.
A.2.4 Venticute
Venticute (Recombinant SP-C or lusupultide) is a clinical exogenous lung surfactant con-
taining synthetic dipalmitoyl-phosphatidyl-choline (DPPC), palmitoyl-oleoyl-phosphatidyl-
glycerol (POPG), palmitic acid, and human sequence or modified human sequence recom-
binant surfactant apoprotein (rSP-C) [85–87, 184, 198, 315–320]. DPPC and POPG are
present in 7:3 weight ratio. Palmitic acid is present at 5% by weight relative to phospho-
lipids, while the rSP-C protein is present at 2% by weight as shown in figure A.5(d). It is
produced by Altana Pharma AG (formerly Byk Gulden), Konstanz, Germany. Venticute
is an investigational drug product, and is not approved for use yet.