Colloids and Surfaces A: Physicochem. Eng. Aspects 270–271 (2005) 323–328

A scaling analysis of�-casein monolayers at liquid–fluid interfaces

Julia Maldonado-Valderrama, M. Jose Galvez-Ruiz,Antonio Martin-Rodriguez, Miguel A. Cabrerizo-Vılchez∗

Grupo de Fısica de Fluidos y Biocoloides, Departament of Applied Physics, University of Granada,C/Fuentenueva s/n, 18071 Granada, Spain

Available online 18 October 2005

Abstract

The experimental results recently provided by a novel application of the pendant drop technique to the formation of protein monolayersat liquid–fluid interfaces are further analysed on the basis of scaling arguments. Specifically,�-casein monolayers at the air–water and thetetradecane–water interface are directly compared and the structural differences inferred by the�–A isotherms are analysed in more detail in termsof the static elasticity modulus. In addition, the model developed by Leclerc et al. based on multiblock theory is applied to the experimental data atboth interfaces [E. Leclerc, M. Daoud, Macromolecules 30 (1997) 293; V. Anguie-Beghin, E. Leclerc, M. Daoud, R.J. Douillard, Colloid InterfaceSci. 214 (1999) 143]. In this manner, the experimental differences are corroborated and subsequently quantified with scaling arguments. Finally,

anciesial ten

k ofless

f theation

ementpar-

ate ispects-urehs. Itless

nter-ationteinacesentlyuresat a

the theoretical analysis is extended to the application to�-casein monolayers at various oil interfaces in order to shed light on the discrepreported on the structural configuration attained by this protein at different oil interfaces. As a result, a correlation between the interfacsionof the liquid and the interfacial structure of the protein is probed.© 2005 Elsevier B.V. All rights reserved.

Keywords: Monolayer;�-Casein; Oil phase; Scaling exponent; Interfacial tension

1. Introduction

Proteins play an extremely important role in the productionand stabilization of many products. Their amphiphilic naturemakes them surface active so that they adsorb at different kindsof interfaces[3]. Two types of interfaces may be considered,namely penetrable and impenetrable. The former is found whenthe molecule is attracted by the interface between two solvents,whereas the latter involves, for example a solid surface. Thislast type of interface has been extensively studied so far, and hasimportant application in colloidal stabilization for instance. Inrelation to the penetrable interface, i.e. liquid–fluid interface, ithas a large potential on emulsions and remains nowadays ratherunknown. The structure attained by proteins at fluid interfacesis hence, a question of increasing relevance in order to under-stand the involved mechanisms with the aim of rationalisingtheir technological use.

The monolayer technique appears as a useful tool in the clar-ification of these processes and has been extensively used in the

∗ Corresponding author.E-mail address: mcabre@ugr.es (M.A. Cabrerizo-Vılchez).

study of proteins at the air–water interface[4,5]. However, thedifficulty added by a liquid interface results in a certain lacworks regarding protein monolayers at oil interfaces. Evenstudies in the literature, provide a quantitative analysis omonolayer behaviour, by means, for example, of the applicof a model.

The results presented by other authors appear in agreregarding the behaviour of some model globular proteins. Inticular, it seems to be accepted that a further unfolded stattained by serum albumin at different oil interfaces with resto that at the air–water interface[6–8]. Conversely, certain dicrepancies appear regarding�-casein when changing the natof the non-polar phase[7,9,10]. This is a model protein witimportant applications in the stabilisation of dairy producthas a very flexible structure in solution and it looks more orlike a “naturally unfolded protein”.

In order to deepen the understanding of this protein’s ifacial behaviour, previous works were devoted to the adaptof the Pendant Drop Technique to the formation of promonolayers at both, the air–water and the oil–water interf[11,12]. Nonetheless, those results were shown independand merely quantitatively analysed. At this point, two featmust be highlighted. On one hand, taking into account th

0927-7757/$ – see front matter © 2005 Elsevier B.V. All rights reserved.doi:10.1016/j.colsurfa.2005.09.002

324 J. Maldonado-Valderrama et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 270–271 (2005) 323–328

direct comparison between the interfacial structure of�-caseinat both interfaces was not shown before it was worth the perfor-mance of an explicit comparative analysis. On the other hand,in order to shed light on the discrepancies encountered in the lit-erature, a theoretical model to further interpret the experimentaldata has been searched.

Since proteins are polymers, they should share some surfaceproperties with other copolymers[10]. A polymer model hasbeen developed recently to describe their surface properties asthose of a multi-block copolymer, i.e. a polymer made of alter-nating hydrophobic and hydrophilic sequences[2]. In this frame,the modelling of the structure and properties of protein adsorbedlayers are tackled using the scaling law approach for polymers.This model has been applied to many experimental systems andappears very useful in the interpretation of the data enlighteningthe interfacial structure.

Recently, Douillard et al. have reviewed the interfacial prop-erties of many proteins on the basis of polymer thermodynamics[10]. In this work, the advantages of using scaling arguments inthe analysis of protein layers are illustrated and important con-clusions arise, once again, especially at the air–water interface.

Accordingly, the main goal of this work is to analysethe experimental data of�-casein at the air–water and thetetradecane–water interfaces formerly obtained[11,12], byusing the polymer approach[1,2]. Furthermore, the theoreti-cal analysis is subsequently applied to the experimental datao samp nal-ya -c ally,t turea the-o nciei

2

2

ceb iquidi er inp press ngea ssurT orcep struct d ast of ai

|

agep has

difference that occurs between dA and dπ. When there is noexchange of material with the adjoining bulk solution, i.e. whenΓ × A is constant and the time of deformation is very large orvery short compared to the time of rearrangements within thelayer,ε is called static elasticity modulus and it can be deducedfrom the equilibrium relationship between interfacial pressureand interfacial area[16]:

ε =(

dγ

d lnA

)Teq

(2)

This relationship provides structural information of the inter-face and has been widely used in the study of protein monolayers[4,5].

2.2. Polymer adsorption

The model, based on multi-block theory, is described in detailin [1], subsequently extended to proteins at fluid interfaces in[2] and further applied in very different experimental conditionsin [10,17,18]. Hence, merely some important features will bebriefly pointed out below.

In the current theory of adsorption of proteins at liquid–fluidinterfaces, the polymer chain is assumed to be made ofNdiblocks of two sequences, A and B consisting ofZA andZ monomers, respectively. The monomers of the A andB A ish oft ,f tweent n thec hilicb ea hicha ares enceo ent,a nter-f andl

lingw opics ith thec nceb omesa ty oft ing tot backt hentl

medt nota r, thej ocksa o notc ts a

btained by other authors who performed studies with therotein but at different oil interfaces. In particular, we have ased the isotherms of�-casein at the toluene–water interface[7]nd at the triolein–water interface[10,13] in which discrepanies in the behaviour at the oil interface were found. Finhe role of the nature of the interface on the interfacial structtained by the protein is discussed in detail in view of theretical considerations and an explanation to the discrepa

s proposed.

. Theoretical background

.1. Static elasticity modulus

The interfacial pressure,π, of a protein layer is the differenetween the surface energy per unit area of the liquid–l

nterface,γ0 and the surface energy measured with the laylace,γ. The response of the adsorbed molecules to comion and expansion of the interface is represented by a charea and measured due to a variation of the interfacial prehis change reflects, therefore, the magnitude of lateral fresents in the monolayer and can be used to provide

ural information. The surface elasticity modulus is definehe increase in interfacial tension for an increase in areanterface element[14,15]:

ε| =∣∣∣∣(

dγ

d lnA

)T

∣∣∣∣ =√

ε20 + (ηdω)2 (1)

The elasticity modulus is a complex quantity with a storart (elastic) and a loss part (viscous) resulting from a p

e

s

-ine.s-

n

e

Bsequences have different chemical nature, namely

ydrophobic while B is hydrophilic. The chemical structurehe polymer is represented by the ratioα = ZA/ZB. In particularor �-casein a very good agreement has been found behe experimental pattern and the theoretical predictions iase of a polymer where the hydrophobic and the hydroplocks have the same order of length[2,17,18]. Let us assumn interface between two immiscible solvents 1 and 2, wre different for both kinds of monomers. Specifically, theyelectively better for A and B, respectively. Thus, each sequf the diblock has a tendency to place itself in its better solvnd this is the reason why the polymer locates at the i

ace. At this point, one can distinguish between gas–liquidiquid–liquid interfaces.

On one hand, at the liquid–liquid interface we are deaith a penetrable interface. When the diblock is in an isotrtate, the centres of masses of each sequence coincide wentre of gravity of the macromolecule. Due to the differeetween the qualities of the solvents, as the polymer becdsorbed, this is no longer true and the centres of gravi

he sequences are on different sides on the interface. Owhe polymeric nature, a restoring force tends to bring themo each other. Localization at the interface occurs either whe number of sequences (N) or the number of monomers (Z) isarge, or else when selectivity is high[1,10].

On the other hand, at the air–water interface, it is assuhat the liquid is a good solvent for the B sequence; air is

good solvent for any of the sequences. In this manneunctions between the hydrophobic and the hydrophilic blre supposed to stick strictly to the interface and the blocks dross the interface[2]. Once adsorbed, the copolymer adop

J. Maldonado-Valderrama et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 270–271 (2005) 323–328 325

2D conformation varying with the interfacial concentration andsuch a definition implies that each block has a fractal structure.

The thermodynamic model allows to provide a descriptionof the variation of the interfacial pressure as a function of theinterfacial coverage[1,2].

π ∼= kBTΓ y (3)

wherekB is the Boltzman constant,T the absolute temperature,y an exponent characteristic of the regime of the interface andof the fractal dimension of the polymer at the beginning of itsadsorption at the interface andΓ the surface concentration ofpolymer, defined as:

Γ = m

A(4)

wherem is the mass of the polymer adsorbed at the interface ofareaA. The key parameter of the model,y, can be also obtainedfrom the elasticity modulus by applying the definition givenabove[1,5,17,18]:

ε = yπ (5)

Both approaches will be used in the treatment of the data.As the interfacial concentration increases, the polymer

molecules and the blocks overlap. The interface enters in suc-cessive semi-diluted regimes whose nature is determined as afunction of the ratioα defined above and of the polymer interfa-c eso boths

arr n oft ssives ondst ingw oft einsf

s isc heots g ina n, i.e.h ei eoww

3

c wati ainep aredc eve

Fig. 1. π–A isotherms of spread�-casein monolayers at the air–water interface(solid symbols)[11] and at the tetradecane–water interface (open symbols)[12].

a direct comparison between them has never been made so far.Accordingly, details about the experimental systems and proce-dure can de found therein and, hence, in this work we will stickto the differences arising from an explicit comparison of bothexperimental data. The significant differences observed betweenthe isotherms shown inFig. 1suggest that the nature of the inter-face has substantially affected the interfacial structure adoptedby the protein at each interface. Let us analyse in detail theseexperimental findings.

Two important differences arise from simple examinationof the isotherms. On one hand, the appearance of a hysteresisphenomenon at the oil interface in which a rather less repro-ducibility is found between compression and expansion of theinterface. On the other hand, the isotherm appears displaced tohigher molecular areas at the oil–water interface. Relating thislast feature, Murray et al. encountered a similar displacementtowards higher molecular areas with monolayers formed at thesame non-polar phase (tetradecane) but with other model pro-teins[6]. Nevertheless, studies performed with the same proteinbut at different oil phases show a contrary behaviour in which theisotherm appears displaced to slightly lower effective areas[7].Although these apparent discrepancies will be analysed later infurther detail on the basis of the theoretical considerations, somepreliminary arguments are pointed out as possible explanationsof this behaviour.

Firstly, it can be assumed that the oil phase is a better sol-v acidr e oilp c-t rfacer umptae teinf tionb tein[ kesi oteint ordert s put

ial concentration. In the case of�-casein (intermediate valuf α), the phase diagram is influenced by the properties ofides of the interface.

The numerical value ofy obtained for each of the lineegions found in the experimental data, provides informatiohe structure of the polymer at the interface in those succeemi-diluted regimes. The first semi-dilute regime correspo the 2D overlap of the polymer formed by a long self-avoidalk of blocks (y1 > 3, depending on the fractal dimension

he polymer). This regime is probably not relevant for protormed only of a limited number of blocks[10].

Thus, the first significant semi-dilute regime for proteinharacterised by the exponenty2 and should correspond to tverlap of either 2D or 3D blocks. Ify2 = 1, it would indicatehe overlapping 3D blocks, while a value ofy2 ≥ 3 would be aign of the overlapping of 2D blocks at the interface, beinddition a measure of the overall compactness of the proteiigher as the protein swells[4,5]. The second significant regim

s characterised by the value ofy3. In a similar approach, a valuf y3 = 0 would suggest thickness growth of a quasi-gel,y3 = 1ould imply the formation of quasi-brushes and, finally,y3 ≥ 3ould again indicate the overlapping of 2D blocks[10].

. Results and discussion

Fig. 1 shows the typical isotherms obtained for spread�-asein monolayers at the air–water and the tetradecane–nterface. These experimental data were independently obtreviously;�-casein monolayers at the air–water interfaceescribed in[11] while [12] is devoted to the formation of�-asein monolayers at the tetradecane–water interface. How

erd

r,

ent than air for the more hydrophobic parts of the amino-esidues in the protein, which might in fact penetrate into thhase[6,19]. This partial solubility may imply a major intera

ion between the protein monolayer and the tetradecane inteesponsible for the hysteresis phenomena. Secondly, Segt al. have pointed out that the higher affinity of the pro

or oil interfaces can be originated by a cohesive interacetween oil molecules and hydrophobic parts of the pro

20]. Finally, the marked hydrophobic character of the oil, mat very repulsive to the water molecules and hence the prends to spread as much as possible within the interface ino hinder these unfavourable interaction. All these feature

326 J. Maldonado-Valderrama et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 270–271 (2005) 323–328

Fig. 2. Static elasticity modulus of spread�-casein monolayers at the air–waterinterface (dashed line)[11] and at the tetradecane–water interface (solid line)[12].

together might explain the experimental data in the followingway.

At low interfacial coverage, the latter argument would resultin a higher effective area per molecule at the oil interface con-sistent with the displacement of the isotherm. Subsequently, asthe interface is compressed the cohesive interaction between thehydrophobic segments and the oil molecules, along with thepartial solubility into the tetradecane would cause the hysteresis phenomena. Rather similar arguments were also employein the interpretation of the adsorption behaviour of�-casein atthese same two interfaces[9].

In order to further interpret the experimental data shown inFig. 1, the surface elasticity modulus was directly derived fromthe expansion isotherms by applying the definition given in Section 2. Fig. 2 shows the elastic behaviour of�-casein at theair–water interface and the tetradecane–water interface.

In agreement with similar studies performed with�-caseinmonolayers at the air–water interface, the elastic behaviour of�-casein shows two maxima at this interface[4,5]. Even though thetwo maxima encountered at the tetradecane–water have alreabeen described in detail in[12], it was worth the performance ofan explicit comparative analysis between the elastic behaviouof �-casein at both interfaces to shed light on the interfacialconfiguration of the protein at the tetradecane interface in viewof the discrepancies with literature data mentioned above.

A maximum in ε is related to a configurational transitioniC psei s frot acenp imao rfacs onalt arisb

ards ing

to Cicuta et al. this fact accounts for a higher energetic cost ofthe transition at the oil interface[5]. This statement might berelated to the cohesive interaction between oil molecules andhydrophobic parts of the protein. This interaction could result ina certain resistance of the protein segments to collapse into theadjacent phase that may perhaps be responsible for the higherenergetic cost found on the curves.

Regarding the second peak, appearing at higher interfacialcoverage, two differences arise between both interfaces; themaximum appears to some extent higher in magnitude but some-what narrower at the tetradecane–water interface. Let us furtherinterpret this feature. On one hand, a wider maximum indicatesthat the conformational transition extends over a wider rangeof interfacial pressure[5]. Thus, the narrower shape at the oilinterface suggests that the formation of loops is favoured at thisinterface. Once again, an attractive interaction seems to appearbetween the oil molecules and the segments of�-casein locatedat the interface in agreement with Segumta et al. findings[20].On the other hand, according with Cicuta et al., differences inthe high of the maximum are related to the length of the loopscollapsing into the subphase. The higher maximum found atthe tetradecane interface as compared to the air–water interface,thus accounts for a further unfolded state attained by�-casein atthe tetradecane–water interface in agreement with the isothermdisplacement. Furthermore, this behaviour concord with the factthat increasing polarity of the adjacent phase improves the sol-v ni ngerl

thes het of themTar ntsy ppli-c entso itym con-s n, then tenceo t thea f thet Let usa

TV .ar

yy

n the interfacial structure of the protein[4,5]. According toicuta et al.[5], the first peak corresponds to the tail colla

nto the adjacent phase, whereas the second peak arisehe collapse of a second region of the molecule into the adjhase that may well be the loop region. The similar maxbserved at the air–water and the tetradecane–water inteuggests that�-casein undertakes two analogue configuratiransitions at both interfaces. However, specific differencesetween interfaces by individual analysis of each peak.

With respect to the first peak, it appears displaced towlightly higherπ at the tetradecane–water interface. Accord

-d

-

dy

r

mt

e,

e

s

ation of the protein[19], facilitating in this way the penetrationto the oil phase and resulting again in the formation of looops at this interface.

According to the plan exposed in the introduction,tructure adopted by�-casein at the air–water and tetradecane–water interface was also analysed by meansodel developed by Leclerc et al.[1] and Anguie-Beghin[2].

o this end, both data shown in this work, theπ–A isothermsnd the static elasticity modulus, were fitted to Eqs.(3) and (5),espectively.Table 1shows the values obtained for expone2 andy3 at the regions encountered in the curves by the aation of both methods. It is worth noting that the exponbtained from theπ–A isotherms and from the static elasticodulus provided similar exponents and the final exponent

idered was the mean value of both approaches. In additioumerical values of the parameters point towards the exisf two well defined regions encountered by the protein air–water, as expected but also, unexpectedly in terms o

heoretical premises, at the tetradecane–water interface.nalyse both interfaces independently.

able 1alues ofy2 andy3 of the equation of state obtained by the application of Eqs(3)nd (5)to theπ–A isotherms (Fig. 1) and the static elasticity modulus (Fig. 2),espectively, for the air–water and the oil–water interfaces

Air–water Tetradecane–water

Fig. 1 Fig. 2 Fig. 1 Fig. 2

2 5.2± 0.3 5.7± 0.4 3.3± 0.5 4.1± 0.3

3 1.8± 0.2 1.6± 0.4 1.6± 0.3 2.0± 0.3

J. Maldonado-Valderrama et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 270–271 (2005) 323–328 327

On one hand, at the air–water interface, the valuey2 = 5.5agrees well with other scaling analysis found in the literature atthe same pH value that have been collected by Douillard et al.[10]. Likewise, and as described in the cited work, this valuecorresponds to a rather large fractal dimension of the moleculedue to a partial collapse of the blocks caused by non-covalentinteractions between the monomers. In fact, further evidencefor this argument was given by lower values ofy2 reported for�-casein under conditions in which a swelling of the protein isexpected such as the presence of a denaturating agent[17], anincrease of the temperature[18] and a retreat from the isoelectricpoint[5]. Subsequently, as the interfacial coverage increases theprotein enters in the following regime characterised by the valueof y3. The numerical value,y3 ∼1.7, prove how the overlap of2D pancakes develops in a kind of 3D melt formation of thehydrophobic blocks that increases its thickness with increasingconcentration as pointed out by Douillard et al.[10].

On the other hand, the presence of the tetradecane inter-face considerably modifies these predictions. The valuey2 ∼3.7obtained at the oil interface for the region of low interfacial cov-erage, corresponds to a lower fractal dimension than the obtainedat the air–water interface. In particular, this value approaches thetheoretical value for a self-avoiding walk 2D polymer (y = 3).Moreover, this further unravelling of the protein attained at thetetradecane interface as compared to the air–water interface isconsistent with the displacement of the isotherm inFig. 1 andd a fut watei or the en-t el layet ion ot to tho

al-y froma them 1at ordet l waa ereno owsu ment edf s of

TV t oili

ATTT

�-casein at the toluene–water interface[7], the data offered bySegumpta et al. at the triolein–water interface[10,13]along withthe values obtained at the tetradecane–water interface previouslydescribed.

The authors expose the differences referred to above, in termsof the interpretation of the experimental data as done in the anal-ysis ofFig. 1. Douillard et al. points out the fact that�-caseinat the triolein–water interface seems to retain a 3D structure allover the pressure range considered. Concretely, analysis of theexponentsy2 andy3 obtained by the application of the model(Eq.(3)) to the experimental curves indicate overlapping of 3Dblocks occurs at lower interfacial coverage while the formationof a quasi-brush structure dominates at higher interfacial con-centrations corroborating the experimental predictions[10,13].Rather similar results were obtained from the analysis of thedata at the toluene–water interface. In the same line, Grahamand Phillips report a further compact structure attained by�-casein at the toluene–water interface as compared to that at theair–water interface[7]. The interpretation of the values obtainedfor exponentsy2 andy3 suggests that 3D configuration of theprotein is less clearly noticeable at this interface. However, theappearance of 2D domains seems even more improbable in viewof the results. Nevertheless, the fact that the regions describedby exponentsy2 andy3 are formed by similar structural units andthat these are nearly three-dimensional, seems to be analogue atthe toluene and the triolein interfaces.

sat-i dis-c thesefi gnif-i tein.T thed wea r tod n-n oteina sion[ cesc am-i eent thep

ateri longwu ids,w ase,r is nota highit mayw nter-f lueso v-e tionsw entw ater

escribed above Therefore, the protein seems to adopther extended interfacial configuration at the tetradecane–nterface and the theoretical treatment provides evidence fxperimental findings. In addition, at higher interfacial concrations, the formation of quasi-brushes (y3 ∼ 1.8) dominates thayer behaviour and the formation of 3D aggregates in theakes place. This result once again supports the interpretathe experimental curves that suggested the penetration inil phase at higher interfacial coverage.

Although the information provided by the theoretical ansis of the curves corroborates the experimental findings,theoretic point of view this is not as clear. In particular,odel proposes that exponentsy2 andy3 cannot be larger thant liquid interfaces and the value reported fory2 in Table 1at the

etradecane interface clearly deviates from this premise. Ino deepen the understanding of this behaviour, the modelso applied to results obtained for the same protein at diffil interfaces found in the literature. Moreover, this study alls clarifying the apparent disagreement found in the experi

al data shown therein.Table 2contains the parameters obtainrom the analysis of the data given by Graham and Phillip

able 2alues ofy2 andy3 of the equation of state at the air–water and at differen

nterfaces[7,10,13]and its respective interfacial tension at 25◦C [22]

y2 y3 γ (mJ/m2)

ir–water 5.5± 0.4 1.7± 0.3 72.6etradecane–water 3.7± 0.4 1.8± 0.3 53oluene–water 2 2 35riolein–water 1.1 1.1 14.6

r-r

e

rfe

rst

-

Therefore, the application of the model provides a verysfactory agreement and moreover, an explanation to therepancies encountered with other authors. Furthermore,ndings strongly suggest that the nature of the interface sicantly affects the interfacial structure attained by the prohe issue arising from this observation is the way in whichifferent kinds of oil could modify the protein structure. Ifdmit that the unravelling of the protein proceeds in ordeecrease the surface energy[21], one could search for a coection between the interfacial structure attained by the prnd the nature of the interface in terms of its interfacial ten

9]. Table 2shows the interfacial tensions of the three interfaonsidered[22] and the expected relationship arises from exnation of the values. An evident correlation appears betwhe value of the interfacial tension and the unravelling ofrotein at low interfacial coverage (y2).

On one hand, it can be appreciated how the triolein–wnterface has an extremely low interfacial tension. Hence, aith the preceding argument, the structure attained by�-caseinpon contact with the interface formed by these two liquould proceed with a preferred penetration into the oil ph

ather than spreading at the interface as the surface energys elevated. This feature might also explain the extremely

nterfacial coverage values reported in[20] that contrast withhe slight reduction of the interfacial pressure attained, andell suggest a complete transfer of the molecule into the oil i

ace. Relating the toluene–water interface, in view of the vabtained for exponentsy2 andy3, a certain increase in the unralling of the protein is predicted by the theoretical consideraith respect to the triolein–water interface. In clear agreemith this conclusion, the interfacial tension of the toluene–w

328 J. Maldonado-Valderrama et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 270–271 (2005) 323–328

interface appears higher than the triolein–water interfacial ten-sion. Accordingly, it would result in a similar increase in theunravelling of the protein at the toluene–water interface. Finally,the even higher interfacial tension of the tetradecane–water inter-face might well result in a higher unravelling of the protein atthis interface. In this manner, the discrepancies reported in theliterature for the unfolding degree of�-casein at different oilinterfaces seem to have a satisfactory explanation.

On the other hand, concerning the unexpectedly higher valueobtained fory2 at the tetradecane–water interface, one can finda possible explanation to this feature by following similar argu-ments. In fact, the interfacial tension of the tetradecane–waterinterface is, not only higher than the other two liquid interfacesconsidered, but it approaches the surface tension of pure water.Consequently, this feature might be related to the appearance oftwo well defined regions as well as 2D pancakes in the interfaciallayer formed at the tetradecane–water interface, similarly to thatat the air–water interface. Therefore, summing up, as the inter-facial pressure increases, the protein adopts a further unfoldedstructure and it approaches the behaviour of proteins at gaseousinterfaces.

4. Conclusions

A detailed comparative study of the interfacial behaviour of� intefi truct encea er; ac nd thn n thl prot mena ctlyf medt yersu acesr xpon ep ichc ddi-t a o� ed tob r. Asa termo as ak een

the interfacial tension of each of the liquid and the interfacialstructure of the protein. Briefly, the picture of the molecule,at least at the high interfacial coverage regime, would be thatof a polymer with alternation of hydrophobic and hydrophilicblocks with a conformation determined by the nature of the adja-cent liquids. Further interaction of the protein with the oil phaseis encountered as the interfacial tension increases. Moreover,the protein interfacial behaviour resembles to that the air–waterinterface as the oil/water interfacial tension increases.

Acknowledgements

Financial support from “Ministerio de Ciencia y Tecnologıa,Plan Nacional de Investigacion Cientıfica. Desarrollo e Inno-vacion Tecnologica (I + D + I)”, Projects MAT2001-2834-C02-01, AGL2001-3483-C02-02 and AGL2004-0153/ALI are grate-fully acknowledged.

References

[1] E. Leclerc, M. Daoud, Macromolecules 30 (1997) 293.[2] V. Anguie-Beghin, E. Leclerc, M. Daoud, R.J. Douillard, Colloid Inter-

face Sci. 214 (1999) 143.[3] E. Dickinson, Food Hydrocolloids 17 (2003) 25.[4] M. Mellema, F.A. Clark, F.A. Husband, A.R. Mackie, Langmuir 14

(1998) 1753.[5] P. Cicuta, I. Hopkinson, J. Chem. Phys. 114 (2001) 8659.

.08

[ e

[ .

[ ,

[ 407.[ ds.),

8, p.

[ .[[ g-

[ ,

[ 95.[ 6991.[ 9167.[ istry

-casein monolayers at air–water and tetradecane–waterace has been performed. Concretely, the analysis of theπ–Asotherms suggests important differences in the interfacial sure adopted by the protein at each interface. These differre explained in terms of several factors acting altogethohesive interaction between the hydrophobic segments aon-polar phase, the improved salvation of these parts i

iquid phase and a further unfolded state attained by theein at the tetradecane–water interface. These same argullow us to interpret the static elasticity modulus derived dire

rom the isotherm. Moreover, the theoretical analysis perforo the experimental curves indicates that both interfacial landergoes two analogue different regimes at the two interfespectively. The values obtained for the characteristic eents of this theory,y2 andy3, provide a detailed picture of throcesses undertaken by the protein at both interfaces whompletely consistent with the experimental findings. In aion, the model is subsequently applied to literature dat-casein monolayers at different oil interfaces that seeme not in agreement with those firstly analyzed in this paperesult, these discrepancies are theoretically elucidated inf the nature of the oil in which the interfacial tension appearsey factor. Moreover, a relationship is directly observed betw

r-

-s

ee-ts

,-

is

f

s

[6] B.S. Murray, P.V. Nelson, Langmuir 12 (1996) 5973.[7] D.E. Graham, M.C. Philips, Colloid Interface Sci. 70 (1978) 403.[8] R. Douillard, Colloid Surf. A 91 (1994) 113.[9] J. Maldonado-Valderrama, A. Martın-Molina, M.J. Galvez-Ruiz, A

Martın-Rodriguez, M.A. Cabrerizo-Vılchez, J. Phys. Chem. B 1(2004) 12940.

10] R. Douillard, M. Daoud, V. Anguie-Beghin, Curr. Opin. Colloid InterfacSci. 8 (2003) 380.

11] J. Maldonado-Valderrama, H.A. Wege, M.A. Rodrıguez-Valverde, M.JGalvez-Ruiz, M.A. Cabrerizo-Vılchez, Langmuir 19 (2003) 8436.

12] J. Maldonado-Valderrama, A. Martın-Rodriguez, M.J. Galvez-RuizM.A. Cabrerizo-Vılchez, Langmuir 20 (2004) 6093.

13] T. Segumpta, S. Damodaran, J. Colloid Interface Sci. 206 (1998)14] J. Benjamins, E.H. Lucassen-Reynders, in: D. Mobius, R. Miller (E

Proteins at Liquid Interfaces, first ed., Elsevier, Amdsterdam, 199303 (Chapter 9).

15] E.H. Lucassen-Reynders, Adv. Colloid Interface Sci. 2 (1970) 34716] D. Langevin, Adv. Colloid Interface Sci. 88 (2000) 209.17] A. Hambardzumyan, V. Anguie-Beghin, M. Daoud, R. Douillard, Lan

muir 20 (2004) 756.18] A. Hambardzumyan, V. Anguie-Beghin, I. Panaıotov, R. Douillard

Langmuir 19 (2003) 72.19] R. Wustneck, B. Moser, G. Muschiolic, Colloids Surf. B. 15 (1999) 320] T. Segumpta, L. Razumovsky, S. Damodaran, Langmuir 15 (1999)21] R.E. Anderson, V.S. Pande, C.J. Radke, Chem. Phys. 112 (2000)22] R.C. Weast, S.M. Selby, J.W. Long, I. Sunshine, Handbook of Chem

and Physics, 55th ed., CRC Press Inc., Cleveland, OH, 1974.