Journal of Colloid and Interface Science240,284–293 (2001)doi:10.1006/jcis.2001.7625, available online at http://www.idealibrary.com on

Aggregation Behavior of Nonionic Surfactants SynperonicA7 and A50 in Aqueous Solution

Mei Li,∗ Yahya Rharbi,∗ Mitchell A. Winnik,∗,1 and Kenneth G. Hahn, Jr.†∗Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, Ontario, Canada M5S 3H6;

and†ICI Paints Research Center, 16651 Sprague Road, Strongsville, Ohio 44136-1739

Received September 15, 2000; accepted April 13, 2001; published online June 26, 2001

The aggregation behavior of Synperonic A7 and A50, fatty alco-hol ethoxylate nonionic surfactants (CmEn) with mean polyoxyethy-lene chain lengths of 7 and 50, respectively, was investigated inaqueous solution by means of surface tension, laser light scattering,steady-state fluorescence, and fluorescence quenching experiments.Surface tension measurements at room temperature indicate thatthe critical micelle concentration of A7 is 0.1 g/L (0.22 mM) andthat of A50 is 0.17 g/L (0.07 mM). Static light scattering measure-ments reveal that the averaged aggregation number of A7 in wateris 417 and that of A50 is 23. When solutions of A7 were allowedto age over a period of months at room temperature, the micelleMw and hydrodynamic radius increased. The aggregation numbersfound from fluorescence quenching measurements, using excimerformation from 1-ethylpyrene as a probe, were about 300 for freshlyprepared solutions of A7 and about 40 for A50. We also found thatthe presence of air has a strong effect on the data obtained from thefluorescence quenching measurements. C© 2001 Academic Press

Key Words: surfactant; aggregation number; surface tension;laser; fluorescence; cmc; micelle.

mod

ee

y

l

sednd

thets

andse

i-by

icallyts in

edrals thed totheyn-redralt of

n of

y-n 157

sur-boveandhisns,sticbtaintion

INTRODUCTION

Nonionic surfactants are widely used both in the hoand in industry (1). These substances normally consistwater-insoluble hydrophobe to which an oligo(ethylene oxi(EO)x chain is attached. As the number of EO units in tchain increases, the hydrophilic–lipophilic balance is shiftedfavor of the water phase (2). In view of the importance of thmaterials, a significant number of studies over the past 20 yhave investigated the self-association of a variety of noniosurfactants. Many of the studies focused on the aggregabehavior, microstructure, and physicochemical propertiesthe micelles formed in relatively dilute solution (3–5), whereother experiments examined the periodic structures formby these surfactants at high concentrations (6–8). A varietphysical techniques have been employed to characterizephysical properties of surfactants in solution and the sizesshapes of the micelles they form. These techniques inc

1 To whom correspondence should be addressed.

,

280021-9797/01 $35.00Copyright C© 2001 by Academic PressAll rights of reproduction in any form reserved.

ef ae)

heinsears

nictionof

asedofthe

andude

surface tension, viscometry, X-ray scattering (3, 9, 10), pulfield gradient NMR (3, 11), fluorescence quenching (3, 4), aquasi-elastic light scattering (3, 6–8). It was found thatcritical micelle concentration (cmc) of nonionic surfactantends to be significantly lower than that of ionic surfactants,for surfactants of similar EOx length the cmc values decreastrongly with increasing alkyl chain length (4). The onset of mcelle formation by these surfactants is not markedly affectedthe presence of electrolytes. The cmc decreases monotonwith increasing temperature, and many nonionic surfactanwater undergo phase separation when heated (2, 12).

A number of widely used nonionic surfactants are obtainby ethoxylation of hydrocarbon alcohols obtained from natusources. The precursor alcohols are often mixtures, and thusurfactants contain a mixture of hydrophobes. There is a neeunderstand how the mixture of hydrophobic groups affectsmicellization properties of the surfactants. For example, Speronic A7 and A50 are fatty alcohol ethoxylates manufactuby ICI. They are prepared by the ethoxylation of the natuproduct Synprol, and the hydrophobe is reported to consisa mixture of 66% linear C13 chains and 34% linear C15 chains(13). The ethoxylation process gives rise to a wide distributioethoxylation chain lengths (EO)x with averagex = 7 for A7 andx = 50 for A50. However, 4.5% of Synprol remains unethoxlated and 16.5% of the surfactant molecules contain more thaEO units. Dimitrovaet al.(13) studied the phase changes of Ain the concentration range of 30–85%. They found that thefactant produced hexagonal structures at concentrations a30%. Above 55%, the system changed to a lamellar structureremained lamellar for A7 concentration up to about 85%. In tpaper, the self-assembly of A7 and A50 at low concentratioi.e. 2–10 g/L, was studied using surface tension, quasielalight scattering, and fluorescence decay measurements to oinformation about the cmc of the surfactants, their aggreganumber, and the microfluidity of the core of the micelles.

EXPERIMENTAL

Synperonic A7 and A50 (aCH3–(bCH2)x–cCH2–(OcCH2cCH2)y–OH) were supplied by ICI Americas Inc., WilmingtonDelaware, and were used as received.1H NMR measurements

4

N

nt

i

h

t

e

a

n

l

p

c

-

od

be

q.

age

,

rayfac-nsin

ionines.e-ione

Lle.henof

AGGREGATION OF NO

were carried out at 400 MHz with a Varian Unity 400 spetrometer on solutions of the surfactants in CDCl3. In thesespectra there is one three-proton signal, a triplet at 0.9 pfrom the aCH3 group in the alkyl chain substituent, one siglet at 1.2–1.4 ppm from the (bCH2)x and a broad singlet a3.2–4.0 ppm from the combined contributions of the –cCH2–Oand (OcCH2

cCH2–)y groups. The integration of the three-protosignal was used to determine the integration value of a sinproton and to calculate the values ofx and y from the spec-tra. We find thatx = 8, y = 6.5 for A7 andx = 9.2, y = 53.7for A50. 1-Ethylpyrene(EtPy), purchased from Eugene, Oregwas used as received. All water used was purified by a Mpore Milli-Q purification system. Static surface tension mesurements on Synperonic A7 and A50 were carried out witDuNouy tensiometer (DSC 70535) at room temperature (22◦C).These experiments were carried out in the dilution mode,adding increments of water to the solution in the cell after easet of measurements was complete. Each individual soluwas measured five to eight times to ensure that the results wreproducible.

Light scattering measurements.Static and dynamic lightscattering measurements were carried out at 25± 0.1◦C usinga variable angle light scattering instrument. This instrumwas equipped with a Spectra Physics He–Ne laser, Model(60 mW at 632.8 nm), as laser source and a Brookhaven DigBI 2030AT correlator. All the solutions measured are in the cocentration range of 2 to 6 g/L and were clarified by passing ethrough a 0.1-µm Waterman filter before measurement.

Static laser light scattering (LLS) measures the angularpendence of the excess absolute time-averaged scattering isity, which is related to the weight average molar massMw ofmicelles by the following equation atq Rg¿ 1 (14, 15):

KC

Rvv(θ )= 1

Mw

(1+ 1

3〈Rg〉2 q2

)+ 2A2C. [1]

HereK = 4π2n20(dn/dc)2/Nλ4 is the optical constant,Rvv(θ )

is the excess Rayleigh ratio,〈Rg〉 is the radius of gyration, andq = [4π (sinθ/2)]/λ is the scattering vector. WhenRvv(θ ) ismeasured at a series ofC and θ values, we can in principledetermineMw, Rg, and A2 by the use of a Zimm plot, whichallows both theC andθ extrapolations to be made on a singgrid (16).

The refractive index increments of A7 and A50 in water wemeasured at 25± 0.1◦C using a KMX-16 differential refrac-tometer equipped with a 632.8-nm He–Ne laser. Both samgave values of 0.13 mL/g.

In dynamic LLS, the intensity–intensity time correlation funtion G(2)(t,q) was measured.G(2)(t,q) is related to the normal-ized first-order electric field time correlation function|g(1)(t,q)|as (14, 15)

G(2)(t,q)=〈I (0,q)I (t,q)〉= A[1+ β∣∣g(1)(t,q)

∣∣2], [2]

IONIC SURFACTANTS 285

c-

pm-

ngle

on,lli-a-

a

bychionere

nt127italn-ch

de-ten-

e

re

les

-

where A is the measured baseline,β is a parameter depending on the coherence of the detector, andt is the delay time.For a polydisperse sample,|g(1)(t,q)| is related to the linewidthdistributionG(0) as

∣∣g(1)(t,q)∣∣ = 〈E(0,q)E∗(0,q)〉=

∫ ∞0

G(0)e−0t d0. [3]

G(0) can be calculated from the Laplace inversion ofG(2)(t,q)on the basis of Eqs. [2] and [3]. In our study, the cumulant methwas used to analyze the time correlation functions.

The cumulant expansion of the correlation function canexpressed as (14, 15)

ln∣∣g(1)(t)

∣∣ = −0t + 1

2!µ2t2− 1

3!µ3t3+ · · ·

= Km(0)(−t)m/m!, [4]

where Km(0) is the mth cumulant of g(1)(t) and µi =∫∞0 G(0)(0 − 0)i d0. The parameters0,µ2, andµ3 are ob-

tained by a least-squares fit of the correlation function in E[4]. The variance ofµ2/0

2 characterizes the distribution widthof G(0). The cumulant expansion is valid ift is small andG(0)is not too broadly distributed. This expansion has the advantof providing reliable information aboutG(0) in terms of0 andµ2 without prior knowledge of the form ofG(0). For a diffusiverelaxation,0 is a function of bothC andq:

0 = D0q2(1+ kdC)

(1+ f

⟨R2

g

⟩zq2). [5]

Here D0 is the translational diffusion coefficient atC→ 0 andq→ 0, f is a dimensionless constant, andkd is the diffusionsecond virial coefficient.D0 can be further converted to thehydrodynamic radius (Rh) using the Stokes–Einstein equationD0 = kBT/6πηRh, with kB the Boltzmann constant,η thesolution viscosity, andT the absolute temperature.

UV absorption measurements.All UV absorption spectrawere measured with a Hewlett–Packard 8452A diode-arspectrometer. The background was subtracted using a surtant solution free of dye as a reference. For all the solutioexamined, the contribution to the optical density by the probewater was negligible because of its much lower concentratand lower extinction coefficient at the excitation wavelengthcomparison to the probe solubilized in the surfactant micellExtinction coefficients for ethylpyrene in A7 and A50 were dtermined from UV absorption measurement at a concentratof 5 g/L for both A7 and A50, where virtually the entire probis incorporated into the micellar phase.

A small but known amount of ethylpyrene (0.2 to 0.5 mof 0.5 mM solution in acetone) was added into a small bottAcetone was evaporated under a stream of nitrogen, and tan aliquot of surfactant solution (5 mL) with a concentration

5 g/L was introduced to the bottle. The global concentration ofthe probe is around 2× 10−5 to 5× 10−5 M, much lower than

he

w

i

u

.

ws

l

rieday,theingwasounterebyinn ofthe

-),

ichas-

mi-m-ers:

re-ing

a-ntthe

anyand

thaterbe-nic

286 LI ET

its saturation limit in A7 and A50 solutions. The solutions westirred for 3 days. Then fluorescence and excitation spewere run on each sample, and excitation spectra at the monoemission wavelength (375 nm) were compared to those atexcimer emission wavelength (480 nm). When compared,two excitation spectra show no differences, implying a laof ground-state preassociation of the probe and a compdissolution of the probe in the micellar solutions (17). Tsamples were then subjected to UV absorption measuremExtinction coefficients of the different UV peaks were calclated from the slopes of the concentration dependence ofabsorbance.

Probe saturation experiments.In these experiments, an excess amount of ethylpyrene (1 mg crystal) was added into eof several centrifuge tubes, and a surfactant solution (5 mL) wa concentration of 2 to 5 g/L was added to each tube. Thelution was also stirred for 3 days. Then the excess proberemoved by centrifugation for 40 min at 15,000 rpm, and Uabsorption spectra were measured. To make sure that saturhad been reached, the solutions were returned to the centrtubes and stirred for 1 more day, and the centrifugation andmeasurement steps were repeated. When the absorption beconstant, we assumed that the solution was saturated withprobe. The solutions were also checked by excitation speat the monomer and excimer emission wavelenghts. We fono difference between the two spectra, implying the succesremoval of the excess probe crystals (17).

Steady-state fluorescence measurements.Steady-state fluo-rescence measurements were carried out with a SPEX (2Fluorolog spectrometer, simultaneously in the S&R mode1-nm steps, integrating counts for 1 s. The spectrometerused in the right-angle configuration when the optical denwas less than 1 and in the front face configuration whenoptical density exceeded 1. Emission spectra were measwith λex = 344 and excitation spectra with bothλem= 375 nm(monomer) andλem= 485 nm (excimer). The monomer anexcimer intensities of the spectra were calculated by ingrating the fluorescence spectra from 367 to 382 nm formonomer emission and from 450 to 550 nm for the excimemission.

Dynamic fluorescence measurements.To determine micelleaggregation numbers by time-resolved fluorescence quenc(TRFQ), fluorescence decay profiles of excited EtPy were msured by the single photon timing technique (18) withλex=344 nm andλem= 375 nm for EtPy. The excitation source waa coaxial flash lamp (Edinburgh Instruments Model 199F) fillwith deuterium. Data were collected to a maximum of 20,0counts in the most intense channel. In this technique, the msured decay profile is a convolution of the true decay andinstrument response function. To determine the instrumentsponse function, the mimic technique (19) was employed, usp-bis-[2-(5-phenyl-oxazyl)]benzene and 2,5-diphenyloxazo

degassed cyclohexane as the reference compounds with atime of 1.1 ns for the former and 1.04 ns for the latter.

AL.

rectramerthetheckleteents.

u-the

-achithso-as

VationfugeUVcamethe

ctrand

sful

1.2)inas

itytheured

dte-theer

hingea-

sed00ea-there-ingin

For each sample, series of dilution experiments were carout. The probe concentration was changed in the following wwhile keeping the surfactant concentration constant. Initially,decay profiles were taken for the surfactant solution containthe highest probe concentration. Then a part of the solutiontaken out from the fluorescence cell and a measured amof surfactant solution was added. The contents of the cell wagitated with a Vortex-Genie 2, Model G-560, manufacturedScientific Industries, Inc., and bubbled with nitrogen for 15 mbefore each experiment to ensure the complete redistributiothe probe between micelles and the removal of oxygen fromsystem.

Analysis of fluorescence decays.Nonexponential decay profiles were fitted to the Poisson quenching micelle model (20

I (t) = I0 exp

[t

τ0−〈n〉(1− exp(−kqt)

]. [6]

This model assumes that the excited dye (EtPy∗) and its quencher(EtPy) satisfy a Poisson distribution among the micelles, whare assumed to be monodisperse in size. The model alsosumes that the rate of exit and entrance of EtPy in thecelles and the rate of excimer dissociation are negligible copared to the quenching rate. This model has four parametthe fluorescence intensity att = 0 (I0), the unquenched EtPylifetime τ0, the pseudo-first-order quenching ratekq, which de-scribes the rate of excimer formation in the micelle, and〈n〉,the average number of EtPy molecules per micelle, which islated to the aggregation number of surfactant by the followequation:

〈n〉 = Nagg[EtPy]

([surfactant]− cmc). [7]

In practice one has only three fitting parameters, sinceτ0 isknown independently from experiments at low [EtPy]. By mesuring a series of EtPy emission decay profiles with differeprobe concentrations in surfactant solution, we can calculateaggregation number from the slope of the plot of〈n〉 versus[EtPy].

RESULTS AND DISCUSSION

Critical Micelle Concentration

The cmc of nonionic surfactants can be determined by mmethods, most commonly surface tension measurementsdye solubilization experiments (21). Meguroet al. (22) re-port a method based on keto–enol tautomerism and claimthis method is suitable for every kind of surfactant. Anothmethod is related to formation of a charge-transfer complextween 7,7,8,8-tetracyanoquinodimethane (TCNQ) and noniosurfactant micelles (23).

life-Surface tensions of the A7 and A50 solutions were mea-sured as functions of concentration at 22◦C with a DuNouy

N

-

d

da

n

ttt

s

a

1

tts

seater-

ules

ousng

re-of

nt

leshatr

AGGREGATION OF NO

FIG. 1. Surface tension of A7 (s) and A50 (h) aqueous solution as functions of surfactant concentration.

tensiometer. These solutions were prepared by sequentialtion with deionized water of solutions prepared at a conceration above the surfactant cmc. Equilibrium values of the sface tension were determined as the average of repeated reaof at least eight measurements at each concentration. We fothat A7 solutions take significantly longer (ca. 1 h) to reaequilibrium than A50 solutions. After equilibrium is reachethe surface tension is constant. The results are plotted agthe concentration in Fig. 1. The surface tension values of bA7 and A50 solutions are constant at concentrations abovecmc, but increase with increasing dilution below the cmc.Fig. 1, we observe a sharp break in the plot of surface tensiosurfactant concentration at 0.1 g/L (0.22 mM) for aqueous sotions of A7 and at 0.17 g/L (0.07 mM) for aqueous solutionsA50. The concentrations at which the breaks occur are attributo the respective critical micelle concentrations of the two sfactants. We note that A50, with the longer EOx chain, has asomewhat lower cmc value. The results also show that neithe A7 nor the A50 plots have a minimum near the cmc insurface tension plot. This type of minimum is often associawith the presence of impurities in the surfactant (24).

Static and Dynamic Light Scattering

Static and dynamic light scattering techniques are widely uin the characterization of micellar solutions of nonionic surfatants (3). For example, Triton X-100 is an octylphenol ethoxylwith an average of 9.6 EO units per molecule. Brownet al. (6)used light scattering measurements to examine Triton X-over a broad range of concentrations (0.5–35%) and overtemperature span from 10 to 45◦C. They found a significantmicellar size polydispersity at their lowest measured concentions and found that the micelle aggregation number increawith increases in both temperature and surfactant concentraPhillieset al. (7, 8) studied the same system and found that TriX-100 micelles are adequately represented as hard sphere

recently reexamined Triton X-100 solutions by dynamic ligscattering and by fluorescence quenching experiments, bu

IONIC SURFACTANTS 287

ilu-nt-ur-dingsund

ch,inst

oththeInvs

lu-ofted

ur-

herheed

edc-te

00the

tra-sedion.on. We

lower concentrations, 2–6 g/L. Although we found an increain micelle size with increasing temperature, we also found ththe micelles had a narrow size distribution at each of the tempatures examined between 5 and 25◦C. At 24.6◦C, Triton X-100has an aggregation number close to 100 surfactant molecper micelle.

In the experiments reported here, we examined aquesolutions of A7 and A50 at surfactant concentrations rangifrom 2 to 6 g/L at 25◦C. Zimm plots of A7 and A50 micelles,obtained from static light scattering measurements, are psented in Fig. 2. From Eq. [1], we see that extrapolation[KC/Rvv(θ )] to C→ 0 andq→ 0 leads toMw. The slope ofthe [KC/Rvv(θ )]C→0 versusq2 plot normally provides a valueof 〈Rg〉, whereas the slope of the [KC/Rvv(θ )]θ→0 versusCplot leads to a value of the second virial coefficientA2. Wefound A2 is −3.35× 10−5 cm3 mol/g2 for A50 and−1.69×10−4 cm3 mol/g2 for A7. Negative values ofA2 would normallyimply that there is a small attractive force between surfactamicelles in solution. The larger negative value ofA2 for A7 maysignify an increase in the aggregation number of the micelwith increased surfactant concentration. Figure 2 shows tRvv(θ ) of both A7 and A50 micelles has no significant angula

htt at

FIG. 2. Zimm plots of A7 (a) and A50 (b) micelles in aqueous solution at25◦C, whereC ranges from 2 to 6 g/L.

t

g

t

c

g

3

e

arly

rusyseithentpeakssur-asrela-Thesnce

ilib-bebethe

ion

don-

asolu-nts.rly

datants ofolu-ic

288 LI ET

FIG. 3. Mutual diffusion coefficients of A7 (fresh (j) and 3 months old (d))and A50 micelles (fresh (h) and 3 months old (s)) as a function of surfactanconcentration.

dependence, indicating that the micelles are quite small.molar mass of A7 micelles is 1.85× 105 g/mol and that of A50micelles is 5.95× 104 g/mol. Based upon the mean molar maof the A7 (Mw = 444) and A50 (Mw = 2540) molecules, wecalculate an aggregation number (Nagg) for A7 of 417, whereasfor A50 we find a much smaller number,Nagg≈ 23.

The mutual diffusion coefficients of A7 and A50 micellewere also measured over the same concentration range (FiThe D values plotted were calculated from the first cumulaof the correlation functions, and the variance ofµ2/0

2 wascalculated from the second cumulant. Figure 3 indicatesthe D values of A50 micelles have essentially no concentratdependence, whereas theD values of A7 micelles decreaswith increasing concentration. The hydrodynamic radiusA50 micelles calculated fromD0 is 3.4 nm, which is quiteclose to that of Triton X-100 micelles at 25◦C. This similarityis accidental, since Triton X-100 has a much larger aggreganumber (100 vs 23 for A50), whereas A50 has a much lonEOx chain length (54 vs 9.6 for Triton X-100). The varianof µ2/0

2 at each concentration of A50 is less than 0indicating a narrow distribution of sizes for A50 micellessolution.

In carrying out these experiments, we found an unusual aeffect for solutions of A7 in water. As the samples are allowedage, the micelles become larger andD values decrease. In Fig.we show that a solution of A7 examined approximately 3 monafter it was initially prepared hasD values nearly 25% smallethan those of freshly prepared solutions. In contrast, no schange is seen for solutions of A50. The hydrodynamic radof fresh A7 micelles is about 5.2 nm and increases to 7.2 nm a3 months. The variance ofµ2/0

2 also increased from 0.1 to 0.12implying a slight broadening of the micelle size distributioStatic light scattering measurements also indicated an incrin molar mass and aggregation number of A7 micelles overtime. This may be related to oxidation of the ethoxyethyle

chain. All the light scattering results of A7 and A50 aqueosolutions are summarized in Table 1.

AL.

The

ss

s. 3).nt

hationeof

tiongere

.1,in

ingto

thsruchiusfter,n.ase

thisne

TABLE 1LLS Results of Synperonic A7 and A50 Aqueous Solutions

Sample 〈Rh 〉(nm) µ2/02 〈Mw 〉(g/mol) 〈Nagg〉

A50 3.4 0.06 5.95× 104 23A7 5.5 0.10 1.85× 105 417A50a 3.4 0.05 6.19× 104 24A7a 8.2 0.12 3.52× 105 793

a A50 and A7 concentrated aqueous solutions (20 g/L) after storage for ne3 months.

Partitioning of EtPy in A7 and A50 Micelles

EtPy is a hydrophobic molecule with low solubility in wate(ca. 0.1µM). The solubility increases dramatically in aqueosolutions of A7 and A50, owing to the solubilization of EtPin the hydrophobic core of the surfactant micelles. In themicelles, EtPy has a typical pyrene absorption spectrum wthe maximum absorption peak at 344 nm. The two adjacpeaks appear at 328 and 314 nm. The absorbance of thesewas measured as a function of the EtPy concentration in thefactant solutions. A blank of the same solution, free of EtPy, wused as a reference. The absorbance of EtPy was calculatedtive to that at 400 nm, which was considered as the baseline.extinction coefficientε corresponding to the different UV peakwas calculated from the slope of the concentration dependeof the absorbance values. Theseε values are listed in Table 2.

Probe molecules in the water phase are in dynamic equrium with those in the micellar phase. Information about propartitioning is essential for calculating the fraction of the proin the water phase (25). This knowledge is necessary to verifyapplicability of the model used to calculate micelle aggregatnumbers.

To determine the partitioning coefficient of EtPy in A7 anA50 micelles, four solutions were prepared with surfactant ccentrations of 1, 2, 3, and 4 g/L, each saturated with EtPydescribed above. The concentrations of the probes in the stion were calculated from the UV absorbance measuremeWe found that the micellized EtPy concentration varies lineawith the surfactant concentration (Figs. 4a and b). When thewere plotted in units of moles per liter for both the surfactaand EtPy concentrations (Fig. 4b), we found that the two setdata defined a common line. This result indicates that the sbility of EtPy in the micelles is related only to the hydrophob

TABLE 2Extinction Coefficients (ε) of 1-Ethylpyrene in A7 and A50

ε(104 M−1 cm−1)

λmax(nm) A7 A50

344 4.31 3.16328 3.14 2.31

us314 1.39 1.04

N

o

l

-

a

4

u-theultsantles

nson

AGGREGATION OF NO

FIG. 4. Solubility enhancement of ethylpyrene in A7 (j) and A50 (d)aqueous solutions as a function of surfactant concentrations. (a) [Etpy] in uof mol/L, and A7 and A50 in g/L; (b) both probe and surfactants concentratiin mol/L.

groups of the micelles. In this case, a simple model of two-phpartitioning is sufficient to characterize the system.

The partitioning of EtPy between the water and micelphases can be described by the equation.

K = [probe]satm

/([probe]sat

w Csur), [8]

where [probe]sat is the saturated probe concentration in the mcellar (m) and water (w) phases, respectively,Csur (g/L) is thesurfactant concentration, andK is the equilibrium constant. Partitioning can also be described by the partition coefficientKv thattakes into account the quantities of the probe in the micellarwater phases relative to the phase volumes,

Kv = K M/vR, [9]

whereM is the mean molecular weight of the surfactant (4for A7, 2540 for A50) andvR is the molar volume of the hy-drophobe. For decane, the molar volumevR = 0.195 L/mol, and

for undecane,vR = 0.211 L/mol. Based upon the known composition of the surfactant, the averagevR for hydrophobe in A7

IONIC SURFACTANTS 289

nitsns

ase

ar

i-

nd

4

TABLE 3Partitioning of 1-Ethylpyrene into A7 and A50 Micelles

fw atCsur (g/L)a

Samples K (L/g) Kv 1 2 3 4

A7 699.2 1.59× 106 0.14 0.07 0.05 0.04A50 160.5 1.93× 105 0.62 0.31 0.21 0.16

a fw is the fraction of EtPy in the water phase, expressed as percentage.

and A50 is 0.255 L/mol. The fraction of EtPy in waterfw canthen be calculated as

fw = 1/(1+ KCsur). [10]

The equilibrium constants for EtPy in both surfactant soltions, as well as the fractions of EtPy in the water phase fordifferent surfactant solutions, are listed in Table 3. These resshow that EtPy has a strong tendency to solubilize in surfactmicelles. In every case, there is less than 1% of EtPy molecuin the water phase.

Steady-State Fluorescence

The ratio IE/IM of the excimer-to-monomer intensities insteady-state fluorescence spectra is also realted to〈n〉 and thefirst-order rate constantkq (26). As pointed out in a previouspublication (27), under our experimental conditionsIE/IM isproportional to the average number〈n〉 of EtPy per micelle:

IE

IM∝ 〈n〉 = [EtPy]

[micelle], [11]

[micelle] = Nagg

Csurf− cmc. [12]

Plots of IE/IM versus [EtPy]/(Csurf− cmc) for both A7 andA50 solutions are linear (Figs. 5 and 6). For A50 solutio(Fig. 6) the data fall on a single line, consistent with a comm

- FIG. 5. Excimer-to-monomer ratioIE/IM of EtPy as a function of the av-erage number of EtPy molecules per micelle forCA7 = 2, 5, and 10 g/L.

d

l

t

db

ig

u

nuo

i

s ofor-

aredin abe-nsved

nsthreeflu-so-

owtant

n-of

ex-les

ataanded asdforag-tion,/L),less,

ntra-the

ur-

hatas-lateatore

at-s aidesouti-lled

290 LI ET

FIG. 6. Excimer-to-monomer ratioIE/IM of EtPy as a function of averagenumber of EtPy molecules per micelle forCA50 = 2, 5, and 10 g/L.

aggregation number for the micelles. The data for each ofthree different surfactant concentrations are distributed athis line. For A7, the plot in Fig. 5 looks similar. Here, on closinspection, one notes that the data obtained at the highesfactant concentration fit a somewhat steeper slope than theobtained for the lowest surfactant concentration. This trensimilar to that found in the TRFQ experiments, describedlow, where the aggregation number atCsurf = 10 g/L was 28%higher than that at 2 g/L. The difference in slopes seen in Fis somewhat smaller.

Analysis of Fluorescence Decay Profiles

During preliminary experiments, we observed that thequenched lifetime of EtPy in A50 aqueous solution decreawith the storage time (30 to 40 days) of the solution, evethe stored solution was free of dye. At the same time, we fothat the decay profiles at low probe concentration in solutiaged much longer than 1 week were no longer exponentialFig. 7). We recall that neither dynamic nor static light scatter

FIG. 7. Fluorescence decay profiles for the EtPy emission in A7 aquesolutions (CA7 = 5 g/L). From C1 to C4, [EtPy]= 3.5, 8.6, 18.4, and 38µM.

AL.

theongersur-data

ise-

. 5

n-sed

ifndns

(seeng

experiments showed any change in A50 micelles over perioda month. This type of problem is consistent with peroxide fmation through oxidation of the EOx chain. Reproducible fluo-rescence decay results were obtained only for freshly prepA50 solutions, and these solutions were always used withday or two of their preparation. In addition, as we emphasizelow, meaningful results were obtained only for micelle solutiothat were degassed to rid the solution of most of the dissoloxygen.

Fluorescence decay profiles of fresh A7 and A50 solutiowere measured at three different probe concentrations anddifferent surfactant concentrations. Figure 7 shows typicalorescence decay profiles of EtPy emission in A7 aqueouslutions. The fluorescence decay profiles of solutions with lconcentrations of EtPy and high concentrations of surfacA7 or A50 are exponential. The unquenched EtPy lifetimesτ0

calculated from these decays were 205± 1 ns in A50 aqueoussolution and 220± 2 ns in A7 aqueous solution and are idependent of the surfactant concentration. At higher ratios[EtPy]/Csurf, the fluorescence decay profiles are no longerponential (cf. decays C1 to C4 in Fig. 7). These decay profiwere fitted to Eq. [6]. The fitting parameter〈n〉 represents theaverage number of EtPy molecules per micelle. Plots of〈n〉 vsEtPy for A7 and A50 solutions are presented in Fig. 8. The dfit a separate straight line for each surfactant concentration,each plot passes through zero. When these data are replott〈n〉 vs [EtPy]/(Csurf-cmcsurf) (Fig. 9), a master curve is obtainefor the A50 solution. We calculate an aggregation numberA50 of 40±5% at each of the three concentrations. Since thegregation number is not dependent on surfactant concentrawe infer that over this range of concentrations (2.0 to 10.0 gthe micelles form by a closed association process. Neverthethere is a discrepancy between the values ofNaggdetermined byTRFQ (Nagg= 40) and by static light scattering (Nagg= 23).Since this discrepancy is not caused by a surfactant concetion dependence of the LLS results, we are unable to explaindifferent aggregation numbers obtained by TRFQ and LLS.

For A7, the results are different. When the data for this sfactant are plotted as〈n〉 vs [EtPy]/(Csurf− cmcsurf), we obtainthree straight lines with different slopes, which indicates tthe aggregation number of A7 micelles increases with increing surfactant concentration. From these results, we calcuNagg values for A7 of 252 at 2 g/L, 306 at 5 g/L, and 32210 g/L. For A7, the TRFQ and light scattering results are mconsistent. By light scattering, we findNagg= 417 in the limitof low concentration. Here the larger size obtained by light sctering is consistent with the idea that light scattering provideweight-averaged aggregation number, whereas TRFQ prova number-averaged value. Almgren and Lofroth (28) pointthat this conclusion is valid only under certain limiting condtions. Under other conditions a more complex average—catheq average (quenching)—characterizes the micelles.

ous We stress that we found air to have a very strong effecton the TRFQ measurements described here, especially for A7

N

e

it

f

n

u

tity

xy-

lleichher,ethatn

tets-s,

tanty tor-

AGGREGATION OF NO

FIG. 8. Average number of EtPy molecules per micelle (〈n〉 =[EtPy]/[micelle] versus [EtPy] for (a)CA7 = 2, 5, and 10 g/L and (b)CA50 = 2,5, and 10 g/L).

solutions. When air was not thoroughly removed from the florescence cell, we were often unable to fit individual decprofiles to the Poisson quenching micelle model. Under thcircumstances, we obtained fits characterized by highχ2 val-ues, with a large scatter inkq values and high〈n〉 values at lowprobe concentration. This observation is quite different frothat reported by Zanaet al. (29). They examined a range osurfactant micelles, including the oligo(ehtylene oxide) surfatants C12EO6 and C12EO8, and found that the presence of aduring the TRFQ experiments had no significant effect onaggregation number. In their experiments, they measuredaggregation number for each surfactant at only one molar ccentration, which is 3 to 10 times higher than that we employfor A7, and 14 to 50 times higher than that we employedA50 solution.

Several authors have reported that exposure to air had afect on emulsion stability and surfactant aggregation (30, 3In order to appreciate the potential magnitude of the effect,note that at room temperature and normal pressure, the solity of oxygen in water at one atmosphere of air pressure is ab1.3 mM; in pentanol–dodecane (2 : 1), the oxygen concentrais about 8 mM (32). If we now assume that the oxygen solubil

in the micelle core is similar to that in the pentanol–dodeca

IONIC SURFACTANTS 291

u-ayse

mfc-rhetheon-edor

ef-1).webil-oution

mixture, we calculate that there is on average about 0.02 ogen molecule per micelle in a C12EO6 solution, whereas in A7micelles the number increases to about 0.5.

Microfluidity of Micelles

The rate of fluorescence quenching within the core of a micekq depends on two factors, the size of the micelle core, whdetermines the local concentration of the dye and the quencand the “fluidity” of the core, which determines how rapidly thdye and quencher can diffuse. Recent results have shownthe core of a Triton X-100 micelle is significantly less fluid thathat of an SDS micelle.

In Fig. 10 we plot values for the intramicellar quenching raconstantkq for EtPy obtained in each of the TRFQ experimendescribed above. Thekq values for A50 solutions are very similar for all of the different probe and surfactant concentrationwith an average value of 12.5µs−1. This value is very similarto that found for pyrene in SDS micelles. Values ofkq for A7solutions with the same surfactant concentration are consfor various probe concentrations, but exhibit a weak tendenclower values with increasing surfactant concentration. In “nomal” micelles formed by closed association,kq is independent

FIG. 9. Average number of EtPy molecules per micelle〈n〉 versus

ne[EtPy]/[micelle] for (a) CA7 = 2, 5, and 10 g/L and (b)CA50 = 2, 5, and10 g/L.

-

s

e

t

ih

w

c

s

ofant

useileas

ing

to

t of

see

3,

er,

.

rk,

k,

t-

,

292 LI ET

FIG. 10. Intramicellar quenching rate constantkq for EtPy versus[EtPy]/[micelle] forCA7 = 2(s), 5(1), and 10(h) g/L andCA50 = 2(d), 5(N),and 10(j) g/L.

of the probe or surfactant concentration, as predicted by Eq.This term is a pseudo-first-order rate constant describing theof a second-order process (EtPy∗ +EtPy) at fixed EtPy concentration. The quencher concentration is determined by the mvolume of the micelles. From this point of view, the weakdecreasing tendency ofkq with increasing A7 concentration isconsistent with the increasing aggregation number of thecelles determined from values of the parameter〈n〉. Thekq val-ues found for A7 micelles are on the order of 2.5µs−1, similarin magnitude to thekq values found for EtPy in Triton X-100 mi-celles. This similarity is likely to reflect an accidental compention of two factors. Triton X-100 micelles have a rather low cofluidity, which retards the rate of excimer formation. A7 micellhave threefold larger aggregation numbers. The similarity ofkq values may reflect a higher core fluidity in A7 leading to fasEtPy diffusion, compensated for by a larger micelle size resultin a lower local concentration of EtPy. When taking into accouthe two factors, we found (kqNagg)A7/(kqNagg)A50 = 1.5, whichmight be due to the higher microviscosity of the core in A7 thin A50 or to shape dependence of the rate of excimer format

Although viscosity is a parameter that describes the flow ofmogeneous liquids on a macro scale, one sometimes employterm “microviscosity” to describe the fluidity of microdomainlike that of the core of a micelle. For these small domains,keep in mind that the domain may be more fluid at its centhan at its edges, and the magnitude of the microviscositythen depend upon where the probe is located.

CONCLUSIONS

The aggregation behavior of the nonionic surfactant fatty alhol ethoxylates Synperonic A7 and A50 was investigated. Frsurface tension measurements, we found cmc values of 0.1(0.19 mM) for A7 and 0.17 g/L (0.07 mM) for A50. A7 formmicelles with an aggregation number close to 300, and this va

increases with surfactant concentration. It also increases wthe solutions are allowed to age at room temperature for a pe

AL.

[6].rate

eanly

mi-

a-res

theeringnt

anon.o-

s thesweterill

o-omg/L

lue

of months. Under similar conditions, the aggregation numberA50 micelles remains constant and does not vary with surfactconcentration. Here we find somewhat different values ofNagg

by light scattering (25) and by fluorescence quenching (40± 2)measurements. This difference is in the wrong direction, becathe former gives a weight-averaged aggregation number, whTRFQ gives the number-averaged aggregation number. Air wfound to have a very strong effect on fluorescence quenchmeasurements. From a comparison of the product ofkq andNagg,the microviscosity of the core of the A7 micelle was inferredbe lower than that of the A50 micelle.

ACKNOWLEDGMENTS

The authors thank ICI, ICI Canada, and NSERC Canada for their supporthis research.

REFERENCES

1. For a review of current trends in the soaps and detergents industry,Chem. Eng. News78,37 (2000).

2. Meguro, K., Ueno, M., and Esumi, K.,in “Nonionic Surfactants: Physi-cal Chemistry,” Surfactant Science Series (Martin J. Schick, Ed.), Vol. 2pp. 109–178. Dekker, New York, 1987.

3. Zana, R., Ed., “Surfactant Solutions: New Method of Investigation.” DekkNew York, 1987.

4. Jonsson, B., Lindman, B., Holmberg, K., and Kronberg, B.,in “Surfactantsand Polymers in Aqueous Solution,” pp. 33–113. Wiley, New York, 1998

5. Zachariasse, A. K.,in “Surfactants in Solution” (K. L. Mittal, Ed.), Vol. 7,p. 79. Plenum, New York, 1984.

6. Brown, W., Rymden, R., Stam, J., Almgren, M., and Svensk, G.,J. Phys.Chem.93,2512 (1989).

7. Streletzky, K., and Phillies, G. D. J.,Langmuir11,42 (1995).8. Phillies, G. D. J., Stott, J., and Ren, S. Z.,J. Phys. Chem, 97, 11563

(1993).9. Paradies, H. H.,J. Phys. Chem.84,599 (1980).

10. Kunieda, H., Shigeta, K., and Suzuki, M.,Langmuir15,3118 (1999).11. Sadaghuabum, A. S., and Khan, A.,Langmuir7, 898 (1991).12. Almgren, M., Brown, W., and Hvidt, S.,Colloid Polym. Sci.273, 2

(1995).13. Dimitrova, G. T., Tadros, Th. F., and Luckham, P. F.,Langmuir11, 1101

(1995).14. Chu, B., “Laser Light Scattering,” 2nd. ed. Academic Press, New Yo

1991.15. Berne, B. J., and Pecora, R., “Dynamic Light Scattering.” Wiley, New Yor

1976.16. Zimm, B. H.,J. Chem. Phys.16,1099 (1948).17. Winnik, F. M.,Chem. Rev.93,587 (1993).18. O’Connor, D. V., and Phillips, D., “Time-Correlated Single Photon Coun

ing.” Academic Press, New York, 1984.19. James, D. R., Demmer, D. R. M., Verrall, R. E., and Steer, R. P.,Rev. Sci.

Instrum.54,1121 (1983).20. (a) Yekta, A., Aikawa, M., Brown, W., and Turro, N.,J. Chem. Phys. Lett.

63,543 (1979); (b) Tachiya, M.,Chem. Phys. Lett.33,289; (c) Infelta, P. P.,Gratzel, M., and Thomas, J. K.,J. Phys. Chem.78,190 (1974).

21. Mukerjee, P., “Solution Chemistry of Surfactants” (K. L. Mittal, Ed.)pp. 153–174. Plenum, New York, 1979.

22. Meguro, K., Ueno, M., and Esumi, K.,in “Nonionic Surfactants, PhysicalChemistry” (Martin J. Schick, Ed.), pp. 112. Dekker, New York, 1987.

23. Deguchi, K., and Meguro, K.,J. Colloid Interface Sci.38,596 (1972).

henriod24. Tabata, Y., Ueno, M., and Meguro, K.,J., Am. Oil Chem. Soc.61, 1123

(1984).

N

,

m,

,

AGGREGATION OF NO

25. (a) Wilhelm, M., Zhao, C. L., Wang, Y., Xu, R. L., Winnik, M. A., MuraJ. L., Riess, G., and Croucher, M. D.,Macromolecules24, 1033 (1991);(b) Vorobyova, O., Yekta, A., Winnik, M. A., and Lau, W.,Macromolecules31,8998 (1998).

26. Infelta, P. R., and Gr¨atzel, M.,J. Chem. Phys.70,179 (1979).27. Rharbi, Y., Kitaev, V., and Winnik, M. A.,Langmuir15,2259 (1999).28. (a) Almgren, M., and Lofroth, J.-E.,J. Colloid Interface Sci.81,486 (1981);

(b) Almgren, M., and Lofroth, J.-E.,J. Chem. Phys.76, 2735 (1982);(c) Almgren, M.,Adv. Colloid Interface Sci.41,9 (1992).

IONIC SURFACTANTS 293

29. Alargova, R. G., Kochijashky, I. I., and Zana, R.,Langmuir14,1575 (1998).30. Karaman, M. E., Ninham, B. W., and Pashley, R. M.,J. Phys. Chem. 100,

15503 (1996).31. (a) Bunkin, N. F., Kiseleva, O. A., Lobeyev, A. V., Movchan, T. G., Ninha

B. W., and Vinogradova, O. I.,Langmuir13,3024 (1997); (b) Meagher, L.,and Craig, V. S. T.,Langmuir10, 2736 (1994); (c) Craig, V. S., NinhamB. W., and Pashley, R. M.,Nature364,317 (1993).

32. Almgren, M., Grieser, F., and Thomas, J. K.,J. Am. Chem. Soc.102,3188(1980).