HAL Id: hal-00895736https://hal.archives-ouvertes.fr/hal-00895736

Submitted on 1 Jan 2010

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

Effect of host particle size on the modification of powderflow behaviours for lactose monohydrate following dry

coatingQi Zhou, Brian Armstrong, Ian Larson, Peter J. Stewart, David A.V. Morton

To cite this version:Qi Zhou, Brian Armstrong, Ian Larson, Peter J. Stewart, David A.V. Morton. Effect of host particlesize on the modification of powder flow behaviours for lactose monohydrate following dry coating.Dairy Science & Technology, EDP sciences/Springer, 2010, 90 (2-3), <10.1051/dst/2009046>. <hal-00895736>

Original article

Effect of host particle size on the modificationof powder flow behaviours for lactosemonohydrate following dry coating

Qi ZHOU1, Brian ARMSTRONG

2, Ian LARSON1, Peter J. STEWART

1,David A.V. MORTON

1*

1 Drug Delivery, Disposition and Dynamics, Monash Institute of Pharmaceutical Science,Monash University, Parkville, VIC 3052, Australia

2 Freeman Technology, Malvern, Worcestershire, UK

Received 17 April 2009 – Revised 1st September 2009 – Accepted 22 September 2009

Published online 17 December 2009

Abstract – Our previous work demonstrated that the powder flowability of a cohesive lactosesample can be improved substantially using a dry coating technique. Our study reported here aimsto investigate the influence of host particle size on the modification of powder flowability followingdry coating process. Four commercial lactose monohydrate powders with different particle sizeswere coated by an intensive mechanical process or mixed using a conventional tumbling process,both with magnesium stearate. All four untreated lactose samples showed a relatively poor powderflow. After dry coating, poured and tapped densities of all the lactose samples increased, while Carrindices and Hausner ratios decreased substantially. The angle of repose values were reduced to anotable extent only for particles with a median size larger than about 7 μm after dry coating. Bothspecific energy (SE) and cohesion values of lactose samples, measured by a powder rheometersystem, decreased substantially after coating. In contrast, no apparent changes in powder flow wereevident for conventionally mixed batches, except that in the dynamic powder rheometrymeasurement, a relatively small change in SE was observed. This study demonstrated that forthe finer particles examined, cohesive forces were more influential in the powder bed after thesurface treatment and resulted in a relatively poor flow. However, for the larger powders studied, thecohesive inter-particle forces could be overcome after this dry coating, whereby satisfactory flowcould be obtained. This study indicated that the host particle size was a critical factor in influencingthe modification of cohesive powder flowability.

lactose monohydrate / dry coating / powder flow / particle size / powder rheometry

摘要 – 水合物的颗粒粒径对干法包衣粉体流动性的影响○ 前期的研究结果证明了干法包衣技术可以充分地改善一种粘性乳糖样品的粉体流动性 ○本文目的是研究主体材料颗粒大小对干法包衣粉体流动性的影响 ○四种不同颗粒大小的乳糖一水合物分别采用机械干法包衣和常规转筒法混合进行处理，两种方法都使用了硬脂酸镁作为处理材料 ○四种未处理的乳糖样品表现出非常差的流动性 ○干法包衣后，所有乳糖样品的松密度和堆积密度显著增加，而 Carr指数和 Hausner 比值大大的降低; 只有中值粒度大于 7 μm 的中等颗粒样品的休止角值显著地降低 ○包衣后乳糖样品的比能和粘合力显著地降低 ○相反，常规混合的样品在粉体流动性上没有显著地改变，只是比能值有微小的改变 ○本实验表明对于细微颗粒而言，虽然经过表面包衣改性处理，颗粒间粘合力的影响依然较大，使得粉体的流动性仍然处于相对较差的

*Corresponding author (通讯作者): david.morton@pharm.monash.edu.au

Dairy Sci. Technol. 90 (2010) 237–251© INRA, EDP Sciences, 2009DOI: 10.1051/dst/2009046

Available online at:www.dairy-journal.org

Article published by EDP Sciences

水平 ○ 然而，对粒径中等和较大的颗粒干法包衣可以克服颗粒内部之间的粘合力并且获得满意的流动性 ○本研究证明了包衣主体材料的粒径是影响粘性粉体流动性改善的主要因素之一○

乳糖一水合物 / 干法包衣 / 粉体流动性 / 颗粒粒径 / 粉末流变性状测定法

Résumé – Effet de la taille des particules de poudre de lactose monohydraté sur son enrobageà sec et sur son écoulement. Quatre poudres commerciales de lactose monohydraté ont été enro-bées à sec par mécanofusion et mélangées avec du stéarate de magnésium de façon convention-nelle dans un tambour. Avant traitement, les quatre échantillons de lactose non traité présentaientdes propriétés d’écoulement relativement faibles. Après enrobage à sec, les masses volumiques,apparente et tassée, de tous les échantillons de lactose augmentaient tandis que les indices de Carr etles ratios d’Hausner diminuaient nettement. Les plus grands changements de masse volumiqueétaient trouvés pour les échantillons de lactose ayant une taille de particules médiane entre 7 et 20 μm.L’angle de repos était réduit de façon beaucoup plus importante pour les particules ayant une taillemédiane supérieure 7 μm. L’énergie spécifique et les valeurs de cohésion des échantillons de lactosemesurées à l’aide d’un rhéomètre à poudres diminuaient substantiellement après enrobage. Aucontraire, aucun changement apparent dans l’écoulement de la poudre n’était évident pour les lotsmélangés en tambour, si ce n’est que, dans les mesures dynamiques de rhéométrie de la poudre, undéplacement limité de l’énergie spécifique était observé. Cette étude démontre que pour les petitesparticules, les forces cohésives sont influentes dans le lit de poudre, même après traitement de surfaceet résultent dans un relativement faible écoulement. Cependant, les poudres de plus grosse tailleétudiées ont démontré que les forces de cohésion inter-particulaires à l’intérieur de la poudre d’originepouvaient être surmontées et qu’un écoulement satisfaisant pouvait être obtenu après enrobage à sec.Dans ces séries de poudres de lactose, il a étémontré pour la première fois dans cette étude qu’il sembley avoir une taille de particule critique pour laquelle une amélioration fonctionnelle significative del’écoulement de la poudre peut être obtenue à l’aide de la nouvelle approche d’enrobage à sec. Larhéométrie de la poudre a été démontrée comme étant capable de détecter des différences fines dans lecomportement de la poudre à l’écoulement, que des méthodes traditionnelles n’étaient pas capablesd’obtenir.

lactose monohydraté / enrobage à sec / écoulement des poudres / taille des particules /rhéométrie des poudres

1. INTRODUCTION

Lactose is derived from milk and iswidely used in the pharmaceutical industryas an excipient due to its safe and practicalproperties. Lactose is commonly used astablet or capsule diluents and as a flowimproving carrier in interactive mixtures.Such interactive mixtures are commonlyencountered in tablets and in dry powderinhaler formulations [29]. In addition, finelactose powders can be used as agglomera-tion modifiers in oral and inhaled powders,but have poor flow properties due tothe strong inter-particle cohesive forces[2, 19, 33]. Since poor flow will introduceproblems in transportation, mixing, samplefilling, compression and packaging of

formulations, many studies have beenundertaken with the aim to improve thepowder flowability of such fine pharmaceu-tical powders [20, 28, 31, 35].

Dry coating techniques are widely usedto improve the flow of cohesive powdersby modifying their inter-particle interac-tions. In general, dry coating is simpler,cheaper, safer and more environmentfriendly than solvent-based alternatives [8].“Mechanofusion” is a term used for arecently developed dry coating approachthat has gained interest for particle and pow-der modification [26].

A number of different mechanofusionsystems are available, but in general theyconsist of a cylinder chamber and a processhead that rotate relative to each other at high

238 Q. Zhou et al.

speed to create intense shear and compres-sion of the core (host) and coating (guest)particles both via impaction with the faceof the process head and via compressionas the particles are pushed between the edgeof the head and the chamber wall. The pro-cess head will consequently break upagglomerates of the cohesive particles toexpose their surfaces as the process headrotates at high speed, a considerable amountof thermo-mechanical energy is generatedwhich coats the guest material onto theexposed surfaces of the host particles [4,25]. Although the particle interactions andkinetic energy exchanges in the mechanofu-sion process have been studied using simu-lation and modelling tools, the mechanismof mechanofusion for different materialsappears complex and is not well understood[11, 12, 26]. However, unlike general mill-ing and co-milling processes, the energyinput is more controlled because the processhead geometry, speed and gap from the wallare fixed, and the process can be tuned toencourage coating but not size reduction[24].

Mechanofusion is used in a wide varietyof applications. Its ability to engineer parti-cles is exploited to modify a range of theirphysico-chemical properties. Flowability ofsome poor-flowing powders has beenreported to be improved dramatically aftermechanofusion. For example, 5 μm poly-methylmethacrylate (PMMA) particlescoated with 10% (w/w) of 0.015 μm TiO2

particles usingmechanofusion were reportedtoflow so freely that they appeared to possessa near-zero angle of repose (AOR). In con-trast, both the original PMMA and TiO2 par-ticles had poor flow properties [36]. Similarfindings are also reportedwhenmicron-sizedPMMA is mechanofused with nano-sizedTiO2, Al2O3 or SiO2 particles [18]. Process-ing of ground polystyrene resin of 10 μmsize with carbon black via mechanofusionwas also demonstrated to produce an easilyflowing toner material of rounded shape[36].Dry coating of corn starchwith different

silica particles also showed a dramaticimprovement in powder flow reflected bylower angles of repose values [35].

In our earlier investigations, powder flowhas been improved significantly after me-chanofusion treatment of a lactose monohy-drate milled powder with a median particlesize of ~ 20 μm [37]. It has been wellknown that the particle size is a key factorin powder flowability [10, 13, 16]. Thusin the present study, the role of the particlesize of the lactose monohydrate in theimprovement of powder flow followingdry coating has been investigated. Morespecifically, the work was intended to iden-tify the minimum size in this lactose systemwhose powder flow can be improved usingthe dry coating technique. A set of fourcommercially available cohesive α-lactosemonohydrate (in short, lactose) samples,which have relative small particle sizesand poor flowability, were coated with mag-nesium stearate as an inter-particle forcemodifying agent via mechanofusion. Allsamples were also mixed with magnesiumstearate using a conventional tumblingmethod. The powder flow properties ofuntreated and treated samples were thencharacterized using both traditional methodsand a set of more recent techniques basedon powder rheometry and shear testing.

2. MATERIALS AND METHODS

2.1. Materials

Four α-lactose monohydrate sampleswere used in this study. Pharmatose® 450M(P450) and Pharmatose® 350M (P350) weredonated by DMV International, Veghel,The Netherlands. Sorbalac® 400 (S400) wasobtained from Meggle GmbH, Wasserburg,Germany and Lactohale® LH 300 (L300)from Friesland Foods Domo, Zwolle,The Netherlands. Magnesium stearateNF (MgSt) was supplied by MallinckrodtChemicals, Phillipsburg, USA and

Effect of particle coating on flow of lactose 239

propan-2-ol from Honeywell Burdick &Jackson, Muskegon, USA. All samples wereused as received.

2.2. Methods

2.2.1. Dry coating and tumblingmixing of cohesive powders

Dry coating of lactose sampleswithMgStwas carried out in a mechanofusion system(Nobilta-130, Hosokawa Micron Corpora-tion, Osaka, Japan). Prior to mechanofusiontreatment, lactose samples (~ 100 g) weremanually premixed with a level of MgSt thatwas estimated to be a minimum that could beused to provide a complete covering (2%,2%, 1% and 1% w/w for L300, S400, P450and P350 respectively) [24]. Then, themixtures were poured into the process vessel(process volume = 1 L). The mechanofu-sion processing was performed for 10 minat 5000 rpm in order to coat the MgSt ontothe host lactose particles. Cold water was cir-culated through the incorporatedwater jacketto prevent the vessel temperatures fromexceeding 25 °C.

Conventional blending of lactose with thesame content of MgSt as in the mechanofu-sion studies was also carried out using a Tur-bula® T2F mixer (Glen Mills Inc., Clifton,USA). Lactose samples (~ 20 g) and addi-tives were weighed into a glass container thatwas approximately a quarter full. Thecontainer was then fixed in the mixer andtumbled for 30 min at 72 rpm.

2.2.2. Scanning electron microscopy

Scanning electronmicroscope (Phenom™,FEI Company, Hillsboro,Oregon,USA)wasemployed to examine the morphology of thelactose samples. Each sample was pouredonto a double-sided carbon black tape thatwas mounted on a sample holder. Sampleswere sputter coated with gold using anelectrical potential of 2.0 kV at 25 mA(SCD005, BAL-TEC AG, Balzers,

Germany). Scanning electron microscopy(SEM) micrographs were taken using thein-built image capture software.

2.2.3. Powder bulk densitiesand derived indices

The poured density (ρp) was measuredby pouring the samples slowly into a10 mL measuring cylinder via a funnel ata fixed height. The tapped density (ρt) wasdetermined after 1250 taps of an automatictapper (AUTOTAP™, Quantachrome Instru-ments, Boynton Beach, USA). The tapperoperated with a 3.18 mm vertical travel ata tapping speed of 260 tap·min−1. Fourreplicates were carried out for eachmeasurement.

The Carr index (CI) [9] and the Hausnerratio (HR) [15] were calculated from thepoured density and tapped density usingthe following equations:

CI ¼ qt � qP

qt; ð1Þ

HR ¼ qt

qp� ð2Þ

2.2.4. Particle sizing

The particle size distributions of both ori-ginal and processed powder samples weremeasured by laser diffraction (Mastersizer® S,Malvern Instruments, Worcestershire, UK)using the 300 RF lens equipped with a smallvolume sample presentation unit (capacity150 mL). Approximately 500 mg of lactosepowder were sonicated in 20 mL of propan-2-ol in a water bath for 3 min prior to themeasurement. Particle size analysis of eachsample was performed using the referencerefractive index of lactose (1.533) andpropan-2-ol (1.378) [3]. The average parti-cle size distribution was calculated fromfour replicates of each sample. The averageparticle size of the primary powders was

240 Q. Zhou et al.

described by the volume median diameter(VMD).

2.2.5. Angle of repose

The AOR was measured as an indicatorof powder flowability. The measurementwas done according to the guidance inUSP 31-NF 26. A sample of ~ 50 g wascontinuously fed through a sieve (850 μmaperture size) and a glass funnel (diame-ter = 10 cm; length = 10 cm; and stemdiameter = 1 cm) to form a balanced coneon the flat circular surface of a cylinder alu-minium stand (surface diameter = 5 cm).The height of the glass funnel was adjustedto stay 2 cm above the top of the cone,which was found to be important in orderto standardize the impact of falling powderon the tip of the cone and so to obtain con-sistent results. The AOR was defined as theangle between the sides of the cone formedand the horizontal (Fig. 1):

tan a ¼ 2hR; ð3Þ

where α is the AOR, h is the height and Ris the diameter of the cone formed. Fourreadings for each sample were obtained.

2.2.6. Powder rheological propertiesand inter-particle interaction

Detailed powder flow behaviours, repre-sented by their rheological properties, werecharacterized using the Freeman FT4 rheom-eter (Freeman Technology, Worcestershire,UK). Briefly, the powders can be evaluatedby measuring their dynamic, shear and bulkproperties, using a range of attachments,whilst axial and rotational forces are mea-sured. A number of control/measurementmodes are available including position,velocity, force and torque. A more detaileddescription of the testing methodologieswas reported in the literature [13]. In thisstudy, dynamic mode and shear mode wereapplied to evaluate the powder flow proper-ties and the inter-particle interactions.

In the dynamicmode, a bladewith a diam-eter of 23.5 mm was traversed through the25 mL samples in a 25 mm diameter glassvessel with a blade tip speed of 100 mm·s−1

and a helix angle of−5°. The energy requiredtomove the blade through the powder duringan upward traverse was defined as the spe-cific energy (SE). A higher SE value gener-ally represents a poorer powder flow [13].Conditioning cycles, in which the blademovesupwards to lift the sampleswith a gen-tle shear, were used in order to minimize or

Figure 1. Diagram of the measurement of AOR (h, height; R, diameter; and α, AOR).

Effect of particle coating on flow of lactose 241

standardize the influence of the stress historyon the measurement.

In the shear mode, a shear head wasattached to the powder rheometer, and shearstress was measured with respect to the nor-mal stress for a given consolidating stress.A fuller description of the principles of shearcell testing was described by Schwedes [30].For this application, a consolidating stress of9 kPa was applied to the powder bed prior toeach test. Shear tests were then carried out atnormal stresses of 7, 6, 5, 4 and 3 kPa. Theshear stress at each normal stress wasrecorded, and yield loci were plotted. Thecohesion of each sample was evaluated asthe shear stress at zero normal stress byextrapolating the yield loci. The cohesionvalue thusprovides ameasure of the cohesiveinter-particle forceswithin thebed, and hencea higher value corresponds to a more cohe-sive powder [30].

2.2.7. Statistical analysis

The statistical analysis of the dataderived from different batches of lactosesamples was carried out using analysis ofvariance with Tukey’s post hoc analysis ata P value of 0.05 (SPSS version 15.0.0,SPSS Inc., Chicago, IL, USA).

3. RESULTS AND DISCUSSION

3.1. SEM

Representative SEM micrographs ofuntreated and processed L300 and P350are shown in Figures 2 and 3. Images ofL300 and P350 were chosen to demonstratethe different morphological properties ofrelatively smaller and larger groups in thisseries of lactose powders, respectively. Itcan be observed from Figure 2a that smallparticles of untreated batch of L300 with aparticle size of about 4 μm tend to form

larger agglomerates rather than exist asnon-agglomerated particles. Only occasion-ally could non-agglomerated particles beobserved for untreated and mixed batchesof L300. A higher proportion of non-agglomerated particles are apparent in themicrographs of the mechanofused L300,although many particles are still present asagglomerates (Fig. 2c).

For untreated and mixed P350, fine par-ticles less than about 10 μm tended to formagglomerates or are adhered on the surfaceof larger particles. In contrast, for the me-chanofused batch, most of the particles werepresent as non-agglomerated particles, andvery few fine particles can be found onthe surface of larger particles (Fig. 2f).

At the higher magnification of 3500×,flake-shaped MgSt particles appeared toadhere on the large lactose surface, whichsuggested that the tumbling mixing processwas unable to provide enough shear andenergy to spread all MgSt onto lactose sur-faces (Fig. 3e). At this high magnification, itis quite clear that untreated and mixedbatches of P350 exhibit flat smooth surfacesand sharp edges. However, the mechano-fused particles have more rounded edges.This can attribute to either attrition duringthe high-shear processing or coating mate-rial covering the sharp edges or both. Anundulating surface for the mechanofusedbatch of P350 contrasts the flat surfaces ofthe untreated batch and covers almost thewhole of the lactose particles mechanofusedwith MgSt (Fig. 3f). This suggests that theentire surface had been modified and sug-gests that MgSt may well have been coatedacross the surface of P350 particles by themechanofusion. SEM micrographs alsoshowed that the lactose powders were visu-ally better dispersed after mechanofusiontreatment as reflected by a greater propor-tion of non-agglomerated particles and lessagglomerates. Repeated measurements con-firmed that these differences in dispersionwere not due to the sample preparation.

242 Q. Zhou et al.

3.2. Particle size

Particle size distributions are shown inFigure 4. The VMD value of MgSt was7.92 ± 0.70 μm. The VMD values ofuntreated samples of L300, S400, P450and P350 were 3.86 ± 0.02, 6.49 ± 0.10,19.14 ± 0.20 and 29.38 ± 2.12 μm, respec-tively, which are significantly different fromeach other (P < 0.05). Consequently, it wasconsidered that these four grades of lactosesamples satisfy the objective to investigatethe influence of a range of relatively fine par-ticle sizes on their bulk powder flow proper-ties. There are no significant differences inVMD values between untreated and mixedbatches of P350 as well as for P450(P < 0.05). A very slight reduction inVMD values was observed for L300 andS400 after mechanofusion treatment

(P < 0.05). It was also shown that theuntreated lactose samples had very similarparticle size distribution patterns to their cor-responding mixed and mechanofusedbatches (Fig. 4).

The particle size results indicate that nosubstantial changes in particle size weredetected after mixing or mechanofusiontreatments. These findings are in accordancewith those of other reports [22]. It is sug-gested that any coating layer of MgSt onthe surface of lactose particles is too thinto be detected using a laser diffraction tech-nique compared with larger primary sizes oflactose host particles ranging from ~ 4 to30 μm. Since the mechanofusion processinvolves intensive high-shear forces, thevery slight reductions in particle sizes forL300 and S400 after mechanofusion treat-ment are possibly due to the attrition during

Figure 2. SEM micrographs at a magnification of 750× of (a) untreated L300; (b) mixed L300;(c) mechanofused L300; (d) untreated P350; (e) mixed P350; and (f) mechanofused P350.

Effect of particle coating on flow of lactose 243

the mechanofusion processing or alterna-tively this may be due to improved disper-sion and detachment of primary fineparticles [5].

3.3. Powder bulk densitiesand derived indices

Both poured and tapped densitiesincreased with the increase of VMDfor untreated lactose samples (P < 0.001)(Fig. 5). The CI and HR values of untreatedP350 that had the largest relative particle sizewere significantly lower than those ofuntreated samples of L300, S400 and P450(P < 0.05). However, it was noted that nosignificant difference was observed in CIand HR values between untreated samplesof L300, S400 and P450 (P > 0.05),although they had significantly different

particle sizes between them. All untreatedpowders have “very very poor” flow accord-ing to the classification system of Carr [9].

No significant difference in poured den-sity was observed between the untreatedand the tumbling mixed batches for lactosesamples, except P350 (P > 0.05). Tappeddensities of the untreated batches werefound to be significantly higher than thoseof their corresponding mixed batches forall four grades (P < 0.05). Significantlylower CI and HR values were observed aftertumbling mixing with MgSt for S400 andP450 (P < 0.05).

In notable contrast, both the poured den-sity and the tapped density of the mechano-fused batches for all four grades of lactosesamples were very markedly higher thanthose of their corresponding untreated ormixed batches (P < 0.001). The greatest

Figure 3. SEM micrographs at a magnification of 3500× of (a) untreated L300; (b) mixed L300;(c) mechanofused L300; (d) untreated P350; (e) mixed P350; and (f) mechanofused P350.

244 Q. Zhou et al.

Figure 4. Particle size distributions of lactose and magnesium stearate samples by laser diffraction.

Figure 5. Poured density (a), tapped density (b), CI (c) and HR (d) results of untreated (■),tumbling mixed ( ) and mechanofused ( ) lactose samples (mean ± SD, n = 4).

Effect of particle coating on flow of lactose 245

increases between untreated and mechano-fused batches in poured density were foundfor S400 followed by P450, P350 and L300.Both CI and HR values of mechanofusedbatches were dramatically lower than thoseof their corresponding untreated batchesfor all four grades of lactose samples(P < 0.001). The CI values decreased~ 32, 42, 42 and 44% after mechanofusiontreatment for L300, S400, P450 and P350,respectively.

Powder densities and their derivedparam-eters such as CI and HR values arewidely used to evaluate powder flowabilityparticularly for cohesive powders [1, 17,23, 33]. Generally, non-cohesive powderswill be packed more efficiently, and hencehave less scope for consolidation and so arenot as compressible during a consolidationprocess such as tappingasa cohesive powder.They consequently exhibit good flowability,which is a key factor during manufacturing.In contrast, cohesive powders are highlycompressible and have poor powder flow.When cohesive powders are poured into acylinder, loose powder bed structures con-taining a high air volume will be produceddue to their high inter-particle forces, espe-cially the van der Waals forces. During thetapping process, this open cohesive bedstructure can collapse significantly and rear-range its structure to reduce the voidage,withincrease in the number of particle-particlecontacts.

In general, the cohesivity of a powder isstrongly affected by particle size [10, 14].For coarse particles, the inter-particle forcesare much easier to overcome, comparedwith fine cohesive particles, and hence theyflow freely under gravitational forces. In thisstudy, untreated P350 has higher poured andtapped densities, and better flowabilitythan the powders of untreated L300, S400and P450, which contain greater propor-tions of smaller particles. Our CI andHR results indicated that flowability ofthese cohesive lactose powders should beimproved dramatically after dry coating

via mechanofusion. But given that the parti-cle size distributions of our lactose powdersdid not shift after mechanofusion treatment,these findings suggest that the cohesiveinter-particle forces among the lactose parti-cles in the powder beds are much weakerafter coating with MgSt. This is in agree-ment with earlier studies that looked at drugdetachment [7].

3.4. AOR

The AOR values of lactose samples arelisted in Figure 6. The AOR values of allfour untreated lactose samples were > 55°and can be classified as “very poor” powderflow according to Carr’s classification[6, 9]. The AOR values of mechanofusedbatches were found to be significantly lowerthan those of their corresponding untreatedbatches for all four lactose samples(P < 0.001). In contrast, no difference wasobserved in the AOR values between alluntreated batches and their correspondingtumbling mixed batches (P > 0.05). P450mechanofused demonstrated the greatestdecrease of 37.1% in AOR values thanuntreated P450, followed by P350 with adecrease of 33.7%. But the AOR values ofL300 and S400 only decreased 5.5% and7.1%, respectively, after the mechanofusionprocessing.

Angle of repose values are a commonmethod for evaluating the flow propertiesof a powder [6, 10, 27]. In general, AORvalues are dependent on the balancebetween the cohesive inter-particle forcesand the particle weight (i.e. gravitationaldetachment force). Therefore, particle sizewill play a crucial role in determining theAOR values of a powder. For typical cohe-sive fine particles (d ≈ 10 μm), cohesiveforces can be four orders of magnitudehigher than the particle’s weight [34].Hence, the AOR values will often not dis-tinguish between cohesive fine powderssince the cohesive forces will overwhelmgravitational detachment and a similar very

246 Q. Zhou et al.

tall cone will be formed for cohesive pow-ders within these different size ranges[10]. In the present study, all the originalcohesive lactose powders demonstratedvery poor flow, which was reflected by highAOR values > 55°. No meaningful differ-ences were therefore found in the AOR val-ues for all four grades of cohesive untreatedlactose powders, despite having signifi-cantly different particle sizes. In contrast,after the mechanofusion treatment, theAOR of P450 and P350 was improved dra-matically, and this reflected a flowability,which can be deemed as satisfactorily formanufacturing [6].

The improvement in AOR values wasmuch more significant for larger particleswith primary median particle sizes of20 μm and over than for smaller particleswith primary median particle sizes of7 μm and under. For this lactose powderseries, there appears to be a critical particlesize range where significant improvement inthe AOR values can be achieved by surfacemodification, which is due to the transitionwhere gravitational forces can overcomethe cohesive forces within the powder. Inthis study, a particle size between the range7 and 20 μm appears to represent such acritical size range for the transition from

cohesive to flowable powders, in the caseof our surface modification approach.

3.5. Rheological propertiesand inter-particle interaction

Rheological property results ofuntreated, mixed and mechanofused sam-ples measured by a rheometer are shownin Figures 7 and 8. The SE values ofuntreated samples decreased with anincrease in primary particle size. The SEvalue of untreated L300 with primary med-ian particle size of 3.86 μm was 10.7 ±0.8 mJ·g−1, while that of untreated P350with a primary median particle size of29.38 μm was only 5.4 ± 0.3 mJ·g−1. Therewas no apparent difference betweenuntreated batches of L300 and S400 incohesion value. However, P450 and P350,which have larger particle sizes, showedsubstantially lower cohesion values thanthose of L300 and S400.

These rheological results detected usinga powder rheometer suggest that powder co-hesivity increased with a decrease in particlesize which contributes to the poorer flowbehaviour of finer powders. These findingsare in accordance with those of previousstudies [16]. For particles with a similar

Figure 6. Angle of repose results of untreated (■), tumbling mixed ( ) and mechanofused ( )lactose samples (mean ± SD, n = 4).

Effect of particle coating on flow of lactose 247

surface structure, smaller particles have lar-ger surface area and higher interfacial sur-face energy. During powder handling,there are more contacts between smallerparticles, which results in a larger contactarea per powder mass and stronger cohesiveforces. For the powders investigated in thisstudy, they are all cohesive with relative fineparticle sizes. The apparent relationshipbetween particle size and flow behaviour

for cohesive powders is generally notobserved in CI, HR and AOR values. Thissuggests that densities and AOR methodsare not sensitive enough to distinguish dif-ferences in powder flow between the cohe-sive powders in this study. In contrast,dynamic mode powder rheometry is ableto detect subtle differences in powder flowbetween larger and smaller cohesive lactosepowders and suggests a relationship

Figure 7. SE results of lactose samples measured by a powder rheometer (mean ± SD, n = 4).

Figure 8. Cohesion results of lactose samples measured by a powder rheometer.

248 Q. Zhou et al.

between particle size and powder flowbehaviour.

Significant decreases in SE were foundfor lactose samples, except P350, after tum-bling with MgSt, and the decreases are lessfor larger particles (P < 0.05). In contrast,no apparent differences in cohesion valuewere found between untreated and mixedsamples. These results further demonstratethat dynamic rheological measurement candetect subtle changes in the powder flowbehaviours between untreated and mixedbatches, which other methods such as CI,AOR or powder rheometry in the shearmode are unable to detect (Fig. 7).

Both SE and cohesion values for all fourlactose samples decreased markedly aftermechanofusion treatment, in contrast tothe AOR values that could not differentiatethe improvement for the L300 and S400samples. It is notable from the cohesionvalues derived from the powder rheometryin its shear mode that lactose samples werevery much less cohesive after mechanofu-sion treatment with MgSt. As there are nosignificant changes in particle size, theobserved changes in cohesive forces in thepowder bed should be due to the modifica-tion of the surface chemistry and structureby means of the dry intensive coating. Ithas been reported that coating fine particleswith force control agents can result inreduction in pull-off forces between fineparticles [7, 21]. Moreover, lubrication ofdry coating of particle surface has beendemonstrated to modify the surface energyof pharmaceutical powders [22, 32]. Modi-fication of inter-particle interactions bychanging the surface chemistry via drycoating could account for our observed sub-stantial improvements in powder flow char-acteristics after dry coating with MgSt.

4. CONCLUSIONS

This work has shown substantial flowchanges in a series of lactose monohydrate

powders with different particle sizes follow-ing dry coating via the mechanofusion tech-nique. Various powder flow evaluationmethods indicated that all untreated lactosesamples exhibit very poor flow properties.Particle size was to a limited extent foundto affect some indices of powder flow ofuntreated lactose samples, with a larger sizecorresponding to a slightly improved flow.After the intensive mechanical dry coatingwith magnesium stearate, powder flow indi-ces of all samples were substantiallyimproved, while no dramatic changes wereobserved after tumbling mixing with mag-nesium stearate. The improvement in pow-der flow for lactose monohydrate samplesby mechanofusion treatment was also foundto be influenced strongly by their particlesize. Greatest changes in poured and tappeddensities were found for particles with pri-mary particle sizes in the range 7–20 μm.Angles of repose decreased dramaticallyafter mechanofusion treatment for particles> 20 μm, which is attributed to the compet-ing influence of cohesive inter-particleforces and the gravitational detachmentforces. This study suggests that the particles> 20 μm could exhibit good powder flowafter dry coating and be suitable for theuse in pharmaceutical manufacturing. Toauthors’ knowledge, this is the first studyto investigate the effect of host particle sizeon the improvement in powder flow by drycoating technique. The true mechanismof dry coating, the structure of the surfaceof these processed particles and the result-ing impact on the inter-particle inter-actions and powder flow deserve furtherinvestigation.

Powder flow and cohesivity, reflected bytheir rheological properties, are in goodagreement with those measured by tradi-tional flow measurements. Powder rheome-ter is found to be able to detect subtledifferences in powder flow behavioursbecause it can test samples over a wide rangeof stress conditions and can be a useful toolfor powder flow characterization.

Effect of particle coating on flow of lactose 249

Acknowledgements: Thanks to HosokawaMicron Corporation for their help in mechanofu-sion operations. Q.Z. acknowledges the financialsupport from Faculty of Pharmacy and Pharma-ceutical Sciences, Monash University in the formof faculty scholarship.

REFERENCES

[1] Abdullah E.C., Geldart D., The use of bulkdensity measurements as flowability indica-tors, Powder Technol. 102 (1999) 151–165.

[2] Adi H., Larson I., Chiou H., Young P., TrainiD., Stewart P., Agglomerate strength anddispersion of salmeterol xinafoate frompowder mixtures for inhalation, Pharm.Res. 23 (2006) 2556–2565.

[3] Adi H., Larson I., Stewart P., Laser diffrac-tion particle sizing of cohesive lactose pow-ders, Powder Technol. 179 (2007) 90–94.

[4] Alonso M., Satoh M., Miyanami K., Mech-anism of the combined coating – mechano-fusion processing of powders, PowderTechnol. 59 (1989) 45–52.

[5] Alvarez R.C., Ageorges H., Fauchais P.,Fournier P., Smith A., The effect of me-chanofusion process and planetary-millingon composite powder preparation: agglom-eration and fragmentation, Mater. Sci.Forum. 442 (2003) 67–72.

[6] Amidon G.E., Bergren M.S., Grant D.J.W.,Marshall K., Itai S., Physical test methodsfor powder flow characterization of pharma-ceutical materials: a review of methodsstimuli to the revision process, Pharmacope-ial Forum 25 (1999) 8298–8308.

[7] Begat P., Price R., Harris H., Morton D.A.V.,Staniforth J.N., The influence of force con-trol agents on the cohesive-adhesive balancein dry powder inhaler formulations, KONAPowder Part. J. 23 (2005) 109–121.

[8] Bose S., Bogner R.H., Solventless pharma-ceutical coating processes: a review, Pharm.Dev. Technol. 12 (2007) 115–131.

[9] Carr R.L., Evaluating flow properties ofsolids, Chem. Eng. 72 (1965) 163–168.

[10] Castellanos A., The relationship betweenattractive interparticle forces and bulkbehaviour in dry and uncharged fine pow-ders, Adv. Phys. 54 (2005) 263–376.

[11] ChenW.L., Dave R.N., Pfeffer R., Walton O.,Numerical simulation of Mechanofusion sys-tem, Powder Technol. 146 (2004) 121–136.

[12] Dave R., Chen W.L., Mujumdar A., WangW.Q., Pfeffer R., Numerical simulation ofdry particle coating processes by the discreteelement method, Adv. Powder Technol. 14(2003) 449–470.

[13] Freeman R., Measuring the flow propertiesof consolidated conditioned and aeratedpowders – A comparative study using apowder rheometer and a rotational shear cell,Powder Technol. 174 (2007) 25–33.

[14] Geldart D., Types of gas fluidization,Powder Technol. 7 (1973) 285–292.

[15] Hausner H.H., Friction conditions in a massof metal powder, Int. J. Powder Metall. 3(1967) 7–13.

[16] Hou H., Sun C.C., Quantifying effects ofparticulate properties on powder flowproperties using a ring shear tester, J. Pharm.Sci. 97 (2008) 4030–4039.

[17] Iida K., Hayakawa Y., Okamoto H., DanjoK., Leuenberger H., Evaluation of flowproperties of dry powder inhalation ofsalbutamol sulfate with lactose carrier,Chem. Pharm. Bull. 49 (2001) 1326–1330.

[18] Jiang Y.B., Matsusaka S., Masuda H.,Yokoyama T., Evaluation of flowability ofcomposite particles and powder mixtures bya vibrating capillary method, J. Chem. Eng.Jpn. 39 (2006) 14–21.

[19] Jones M.D., Price R., The influence of fineexcipient particles on the performance ofcarrier-based dry powder inhalation formu-lations, Pharm. Res. 23 (2006) 1665–1674.

[20] Kim E.H.J., Chen X.D., Pearce D., Effect ofsurface composition on the flowability ofindustrial spray-dried dairy powders, Col-loids Surf. B 46 (2005) 182–187.

[21] Kumon M., Machida S., Suzuki M., KusaiA., Yonemochi E., Terada K., Applicationand mechanism of inhalation profileimprovement of DPI formulations by me-chanofusion with magnesium stearate,Chem. Pharm. Bull. 56 (2008) 617–625.

[22] KumonM., Suzuki M., Kusai A., YonemochiE., Terada K., Novel approach to DPI carrierlactose with mechanofusion process withadditives and evaluation by IGC, Chem.Pharm. Bull. 54 (2006) 1508–1514.

[23] Li Q., Rudolph V., Weigl B., Earl A.,Interparticle van der Waals force in powderflowability and compactibility, Int. J. Pharm.280 (2004) 77–93.

[24] Morton D., Dry powder inhaler formulationscomprising surface-modified particles with

250 Q. Zhou et al.

anti-adherent additives, US Pat. No. 0127972A1, 2008.

[25] Naito M., Kondo A., Yokoyama T., Appli-cations of comminution techniques for thesurface modification of powder materials,Iron Steel Inst. Jpn. Int. 33 (1993) 915–924.

[26] Pfeffer R., Dave R.N., Wei D.G., RamlakhanM., Synthesis of engineered particulates withtailored properties using dry particle coating,Powder Technol. 117 (2001) 40–67.

[27] Pilpel N., The flow properties of magnesia,J. Pharm. Pharmacol. 16 (1964) 705–716.

[28] Raula J., Laehde A., Kauppinen E.I., Anovel gas phase method for the combinedsynthesis and coating of pharmaceuticalparticles, Pharm. Res. 25 (2008) 242–245.

[29] Rowe R.C., Sheskey P.J., Owen S., Hand-book of pharmaceutical excipients, Pharma-ceutical Press, London, UK, 2006.

[30] Schwedes J., Review on tester for measuringflow properties of bulk solids, GranularMatter 5 (2001) 1–43.

[31] Sun C.C., Improving powder flow propertiesof citric acid by crystal hydration, J. Pharm.Sci. 98 (2009) 1744–1749.

[32] Swaminathan V., Cobb J., Saracovan I.,Measurement of the surface energy of lubri-cated pharmaceutical powders by inverse gaschromatography, Int. J. Pharm. 312 (2006)158–165.

[33] Thalberg K., Lindholm D., Axelsson A.,Comparison of different flowability tests forpowders for inhalation, Powder Technol.146 (2004) 206–213.

[34] Valverde J.M., Castellanos A., Ramos A.,Watson P.K., Avalanches in fine, cohesivepowders, Phys. Rev. E 62 (2000) 6851–6860.

[35] Yang J., Sliva A., Banerjee A., Dave R.N.,Pfeffer R., Dry particle coating for improv-ing the flowability of cohesive powders,Powder Technol. 158 (2005) 21–33.

[36] Yokoyama T., Urayama K., Maito M., KatoM., Yokoyama T., The angmill mechanofu-sion system and its applications, KONAPowder Part. J. 5 (1987) 59–68.

[37] ZhouQ.,ArmstrongB.,Larson I., Stewart J.P.,Morton A.V.D., Improving powder flowproperties of a cohesive lactose monohydratepowder by intensive mechanical dry coating,J. Pharm. Sci. (2009) DOI: 10.1002/jps.21885.

Effect of particle coating on flow of lactose 251