COLLOmsANDSURFACES

ELSEVIERColloids and Surfaces

A: Physicochemical and Engineering Aspects 93 (1994) 79-'5

In situ spectroscopic investigations of adsorbed surfactant andpolymer layers in aqueous and non-aqueous systems

P. Somasundaran *, S. KrishnakumarLangmuir Center for Colloid.~ & Interfaces, Henr}' Krumb School of .\-line.~. C(,[Wtlbia Unil-'ersit}', Nelv York, NY 10017,

USA

Received 8 December 1993; accepted l5 March 1994

Abstract

Adsorption of surfactants and polymers at the solid-liquid interface is used widely to modify interfacial propertiesin a variety of industrial processes such as flotation, ceramic processing, flocculation/dispersion, detergency andenhanced oil recovery. A molecular level understanding of the structure or the adsorbed layer is beneficial for improvingthese processes by manipulating the adsorbed layer. (n this paper we discuss the use or fluorescence, electron spinresonance (ESR) and Raman spectroscopy for the study of adsorbed surfactant and polymer layers in aqueous andnon-aqueous systems. For example, fluorescence studies using pyrene probe on adsorbed surfactant and polymerlayers in aqueous systems along with ESR and Raman spectroscopy reveal the role or surface aggregation andconformation of the adsorbed molecules in controlling the dispersion and wettability properties of the system. In non-aqueous systems, ESR studies using paramagnetic nitroxide probes show how the adsorption of water aIIects theconformation of adsorbed surfactant molecules and thus affects their dispersion properties.

Keywords: Electron spin res<)nance; Fluorescence; Non-aqueous dispersions; Polymer conformation;spectroscopy

Understanding the micro- and nano-structure ofthe adsorbed layer at the solid-liquid interface ishelpful for utilizing adsorption to enhance manyprocesses dependent on adsorption by designingthe adsorbed layers for optimum configurationalcharacteristics. Until recently methods of surfacecharacterization were limited to measurement ofmacroscopic properties like adsorption density,zeta potential, wettability etc. Such studies, whilebeing helpful to provide an insight into the mecha-nisms, could not yield any direct information onthe microscopic characteristics of the adsorbedspecies. Recently spectroscopic techniques havebeen developed for characterizing the adsorbedlayers of surfactants and polymers at the solid-liquid interface and for the identification of surface

Introduction

Interfaces are commonly encountered in natureand have become an integral part of all humanendeavors both scientific and technological. Of allthe interfaces known the solid-liquid interfaceis the most important owing to its involvementin several biological and industrial processes.Adsorption of surfactants and polymers at thesolid-liquid interface is discussed usually in con-nection with processes like flotation and floccula-tion as the surface modification caused by theadsorption is mainly responsible for the efficiencyof these processes [1,2].

* Corresponding author

0927-7757/94/$07.00 II:! 1994 Elsevier Science B. V. All rights reserved

SSDI0927-7757(94)02897-2

P. Somasundaran. S. Kri.~hnakumar!Colloid~ Surfaces It: Ph)'sicochem Eng. Aspects 93 (1994) 79~9580

compounds on mineral particles. In this paper theapplications of fluorescence, electron spin reso-nance (ESR) and Raman spectroscopic techniquesto probe the microstructure of the adsorbed layersof surfactants and polymers on minerals in aqueousand non-aqueous media are discussed. This paperencompasses a review of some of the past workand some of our recent attempts to delineate therole of microstructures of adsorbed surfactants andpolymers on solid surfaces.

The fluorescence mea...urements are generallycarried out by a steady state fluorescencespectrofluorometer and lifetime of fluorescence bya time-resolved fluorescence lifetime instrument[3]. The dependence of fluorescence intensity andlifetime on the physicochemical environment of thefluorescing molecule has been well documented[ 4]. Such data have been applied to micellarphotochemistry to understand the property ofmicelles [5]. We have recently adapted this tech-nique as a tool for the adsorbed layer of surfactantson solids to obtain information on the micropolar-ity. microviscosity of the probe enL'ironment and theaggregation number of the .\'urfactant at the interface.To determine the micropolarity a fluorescent mole-cule, like pyrene, which possesses a highly struc-tured fluorescence spectrum whose vibrationallines are susceptible to intensity fluctuationsbrought on by polarity changes of the medium, isused. This empirical knowledge has been found tobe of universal applicability and used widely toinvestigate the micropolarity of micelles. A prop-erly resolved fluorescence spectrum of pyrene influids has five vibrational fine structures in theregion from 370 to 400 nm (Fig. I). The intensitiesof the first (1.) and the third (13) are found to be

2. Fluorescence spectroscopy

Fluorescence emission is the radiative emissionof light by an excited molecule returning to itsground state energy level. This phenomenon bearsa wealth of information on the environment of thelight absorbing species and has been exploited fora long time for exploring the solution behavior ofsurfactants. Parameters of fundamental importancein fluorescence emission arc ( I) emission maximum(wavelength of maximum intensity), (2) quantumyield of fluorescence (emission efficiency measuredas intensity) and (3) fluorescence lifetime (time takenby the excited state to decay to lie of its initial value).

:J\

+-

'"cQJ

C.--

\,"~ ~~

35B.BB 475.8a

Wove1ength

6e3.ea

(nm)

Fig. t. Typical fluorescence spectrum of pyrene showing the five vibrational lines.

81P. Somasundaran. S. KrishnaklunarjCollo;d'i .5urfaces A: Ph}'sicochem. Eng. Aspects 93 ( 1994) 79-95

particularly sensitive to the changes in the probeenvironment. The ratio of these peaks (13/1t),sometimes referred to as the polarity parameter,changes from about 0.6 in water to a value greaterthan unity in hydrocarbon media.

2.1. Adsorption of .~odium dodecyl sulfate onalumina

by the existence of electrostatic interactionsbetween the ionic surfactant species and the oppo-sitely charged solid surface.

Region II is marked by a conspicuous increasein adsorption which is attributed to the onset ofsurfactant aggregation at the surface through lat-eral interactions between hydrocarbon chains. Thisphenomenon is referred to as hemi-micelle forma-tion [7,8]. The colloidal aggregates formed on thesurface in general between surfactant and/or poly-mer species are referred to as solloids (surfacec~) [9]. In the case of simple ionic surfact-ants they have also been called hemi-micelles,admicelles, surface micelles and surfactant self-assemblies [10].

Region III shows a decrease in the slope of theisotherm and this is ascribed to the increasingelectrostatic hindrance to the surfactant adsorptionfollowing interfacial charge reversal caused by theadsorption of the charged species. [n Region [IIand beyond, both the adsorbent species and theadsorbate are similarly charged.

Region IV and the plateau in it correspond tothe maximum surface coverage as determined by

The adsorption of sodium dodecyl sulfate (SDS)on alumina from aqueous solutions has beenextensively studied and used as a model systemhere to exemplify the different stages in the mecha-nisms of adsorption of an ionic surfactant on acharged mineral. A careful analysis of such anisotherm was first done by Somasundaran andFuerstenau and is referred to as the S-F isotherm.A typical example of an S-F isotherm is shown inFig. 2, where the adsorption of SDS on alumina atpH 6.5 under a constant salt concentration of1 0 ~ 1 kmol m ~ 3 is shown [6]. Mechanistically,

these regions may be viewed as follows.Region 1 which has a slope of unity under

constant ionic strength conditions is characterized

"Eu"'-...."0E

%0t-Q.a:0C/IQ-It

1&1

'<""-'JC/I-'>-V1&1Q00

Fig. 2. Adsorption isotherm of sodium dodecyl sulfate (SDS) on alumina at pH 6.5 in 10-1 kmol m -3 NaCI.

P .'}()nUL'iundaran. S Krishnakumar/Colloids Surfaces A: Physicochem Eng Aspects 93 ( 1994) 79-9582

as bulk methods, help to generate an overall pictureof the adsorption process. The application of spec-troscopic techniques to study these systems givesa better insight into the interior of the AlzO3/SDS solloid system. The 13/ll values for pyrenewere determined for alumina/sodium dodecylsulfate/water systems for various regions of theadsorption isotherm. The data shown in Fig. 4 are

micelle formation in the bulk or monolayer cover-age whichever is attained at the lowest surfactantconcentration; further increase in surfactant con-centration does not alter the adsorption density. Aschematic representation of adsorption by lateralinteractions is given in Fig. 3.

Classical methods based on zeta potential,adsorption density and hydrophobicity, referred to

.I~EGION rNO AGGREGATION ~~~- \

~

I ~~~ \ftEGION ~NUMBER OFAGGREGATESINCREASES-120-130 MOLECULESPER AGGREGATE

. . - . . .- -+- . - ..

REGION mSIZE OF AGGREGATESINCREASES>160 MOLECULESPER AGGREGATE

. . -. . .-+ -+- . . +-. - . -Fig. 3. Schematic representation or the correlation of surface charge and the growth of aggregates for various rcgions or th

adsorption isotherm depicted in Fig. 2.

P Somasundaran. S. Kri.\"II1/akumarjC()lloid\' Surfaces A: Physicochem. Eng. Aspects 93 (1994) 19c;,9: 83

1.5PYRENE IN ADSORBED LAYER 50S/ALUMINA

.IM HoCl, pH 6.5

SOS MICELLE 1.IM NoCt) A

WATER to.IM NoClI

-"~

;s

0 '-;' . ... 1 ... . 1 ... . 1 ... . 1

0 10-' 10-4 10-3 10-2

RESIDUAL DODECYLSULFATE. moles/liler

Fig. 4. 1 Jill Ouorescence parameter of pyrene in sodiumdodecyl sulfate (SDS) in alumina slurries.

Fig. 5. Monomer to excimer ratio (/mll.) of dinaphthyl propane(DNP) in SDS-aJumina slurries as a runl.'tion of SDS adsorptiondensity. The viscosities refer to those of ethanol-glycerolmixtures which give a similar 1m/I. ratio ror DNP.

marked by an abrupt change in local polarity ofthe probe from an aqueous environment to arelatively non-polar micelle type environment. Itis to be noted that this abrupt change occurs in aregion that is well below the critical micelle concen-tration (CMC) and coincides approximately withthe transition in the adsorption isotherm fromRegion I to Region II. In the plateau region the13/ I( value coincides with the maximum value forSDS micellar solutions, indicating completion ofaggregation on the surface. It can also be seen thatthe polarity parameter is fairly constant through-out most of Region III and above and hence seemsto be independent of surface coverage.

Information on microviscosity is obtained bystudying the excimer (excited dimer) forming capa-bilities of suitable fluorescent molecules like1,3-dinaphthyl propane (DNP). The excimer,which is a complex of a ground state and anexcited state monomer, has a characteristic emis-sion frequency. The intramolecular excimer forma-tion is a sensitive function of the microviscosity ofits neighborhood. This property, expressed as theratio of the monomer and excimer yield (Ie/1m) for1,3-dinaphthyl propane, is determined for the solu-tion and for the adsorbed layer for the variousregions of the adsorption isotherm (Fig. 5) [11].These are then compared to the Ie/1m values ofDNP in mixtures of ethanol and glycerol of knownviscosities. Based on the Ie/1m values of DNP forethanol-glycerol mixtures, a microviscosity valueof 90-120 cP is obtained for the adsorbed layer

compared to a value of 8 cP for micelles. Theconstancy of microviscosity as reported by DNPis indicative of the existence of a condensed surfac-tant assembly that holds the probe.

The dynamics of fluorescence emission of pyrenehas been previously studied in homogeneous andmicellar solutions using time-resolved Ruorescencespectrometry [12]. While the decay kinetics ofmonomer and excimer emission may be deriveddirectly for a homogeneous solution (continuousmedium), statistical methods are to be applied toarrive at similar kinetics in aggregated micellarensembles. This stems from a need to recognizethe possibility of random multiple occupancy ofthe probe in the aggregates which affects the exci-mer forming probability within the aggregate. Ifthe micellar system is viewed as groups of indivi-dual micelles with n probes, then P II' the averagenumber of probes per micelle, may be related to nby Poisson statistics through the relation

P II = nil exp( - n)/n1

This model yields the following relation for thetime dependence of monomer emission:

[mIll = [mlOI exp[ - kot + n(exp( - ket) -I)]

where ko is the reciprocal lifetime of excited pyrenein the absence of excimer formation, ke is the

S4 P. So'na.nmdaran, S. KrishnakwnarlCollo;ds Surfaces A: PhJ'sicochem. Mg. Aspects 93 (1994) 79-95

fairly constant in Region II, further adsorption inthis region can be considered to occur by theformation of more aggregates but of the same size.The transition from Region II to III correspondsto the isoelectric point of the mineral. and adsorp-tion in Region III is proposed to occur throughthe growth of existing aggregates rather than theformation of new ones due to lack of positiveadsorption sites. This is possible by the hydro-phobic interaction between the hydrocarbon tailsof the already adsorbed surfactant molecule andthe adsorbing ones. The new molecules adsorbingat the solid-liquid interface can be expected toorient with the ionic head toward the water sincethe solid particles possess a net negative chargeunder these conditions. The whole process ofadsorption has been portrayed schematically inFig. 3.

intramicellar encounter frequency of pyrene inexcited and ground states, and 1m/oJ and lm(tJrepresent the intensity of monomer emissions attime zero and time t respectively. Knowing n, onecan calculate the aggregation number N using theexpressionn = [P]/[Agg] = [P]N/([S] - [Soq])

where [P] is the total pyrene concentration, [Agg]is the concentration of the aggregates and[S] - [Soq] is the concentration of the adsorbedsurfactant.

A k.inetic analysis based on this relation of thedecay profiles of pyrene in the adsorbed layer fordifferent regions of the alumina/dodecyl sulfateadsorption isotherm [13] yielded the aggregationnumbers marked in Fig. 6 for dodecyl sulfate sol-loids. These results yield a picture of the evolutionof the adsorbed layer. The aggregates in Region IIappear to be of relatively uniform size while inRegion III there is a marked growth in the aggre-gate size. In Region II, the surface is not fullycovered and enough positive sites remain asadsorption sites. Since the aggregation number is

2.2. Effect of position of functional group onadsorbed la}'er microstructure [ 14 J

Studies of changes in the position of the sulfonateand the methyl groups on the aromatic ring of

50S/ALUMINAO.1M NoCl, pH 6.5

;f756- 258- m

19& m u166 ~

128

'n

49

iF"

N 10-9e~~.,'0E. 10-10ZQI-Go«~ 10-110cI

&oJI-cI

~ 10-121

cn-J>-U&oJg 10-130

I- OSDS ONLYASDS WITH

PYRENE. .. 1 1 1 ,." .""

)0"5 'M)-4 10-3 10-2

RESIDUAL DODECYLSULFATE. moles/liter

10-14

Fig. 6. Surfactant aggregation numbers determined for various adsorption densities (average number is indicated)

P. Somarundaran. S. Krishnakumar/Colloids Surface.\" A: Physicochem. Eng. Aspects 9J (1994) 79-95 as

alkylxylenesulfonate showed a marked effect ofsuch changes on the micelli7.ation and adsorptionat the solid-liquid interface. Fluorescence spectro-scopy was used to probe the microstructure of theadsorbed layers in these systems. Adsorption iso-therms for 4C 11 3,5-para-xylenesulfonate (Para-I),

4Cll 2.5-para-xylenesulfonate (Para-2) and 4C112,4-meta-xylenesulfonate (Meta) on alumina fromwater and the 13//1 values along various points inthe isotherm are shown in Figs. 7 and 8 respec-tively. At low adsorption densities. 13//1 is about0.6. Once the solloids form. the value goes up to

a.oo.N

0

K;~-o-c---~-c-.c

E RESIDUAL SULFONATE CONC. Kmol/m3x 105

f

6 Parol

0 Mela

0 Para2

Fig. 7. Adsorption of alkylxylencsulfonatcs on alumina.

l 12nOSORPTION DENSITY. Hol/cm x to

Fig. 8. Polarity of adsorbed layers of alkylxylenesulronates in terms of I J/ 11 as a runction of suriactant adsorption densil

... ~ -.. I- ld"wC-~ ~Ea 0.10VI O.S> 1.00 10.00 tm.m ImJ.OO

a

xNe

u"'-

0X

.>-.--rnzwc

100.

10.

1.

.001

(1)1

oot

86 P Solnaswrdaran. S. KrishnakumarlColloids Surfact's A: PhysicochrnL Eng. Aspects 93 ( 1994) 79-9$

about 1. It can also be seen that there is nodifference among the polarities of the adsorbedlayers of the three surfactants indicating that theselayers in all three cases are really compact andhence water penetration is the same in all the cases.Average aggregation numbers of the solloidsincrease from 17 to 76 with increase in adsorptionin all the three cases (Fig. 9). The aggregationnumbers of the two para-xylenesulfonates aresimilar throughout the range studied. However,at higher adsorption densities, the aggregationnumber of the nteta-xylenesulfonate is lower thanthat of the para-xylenesulfonates. This suggestshigher steric hindrance to the packing of the surfac-tant molecules in the solloids of the meta-xylenesul-fonate. Based on this evidence. the effect of changeof position of functional groups on the aromaticring of the alkylxylenesulfonates on adsorption canbe attributed to the steric constraints to the pack-ing of surfactant molecules in their aggregates.

small molecules in that the polymer adsorption isgreatly influenced by the multifunctional groupsthat it possesses [15]. This stems from the widelyvarying sizes and configurations available for thepolymer. In addition, macromolecules usually pos-sess many functional groups each having a poten-tial to adsorb on one or more given surfaces.

Among the polymeric materials, polyelectrolytesare the most important becaUse of their participa-tion in many biological processes and their utilityin processes like dispersion, flocculation, adhesionand rheology. Poly(acrylic acid) (PAA) and poly-(methacrylic acid) (PMAA) are usually chosen assimple examples of polyelectrolytes to gain insightinto the adsorption behavior of polyelectrolytes.Such an understanding is important in acquiringeffective control over the processes of coUoidalstabilization and flocculation.

Poly(acrylic acid) can exist in different confor-mations depending on the solvent, pH and ionicstrength conditions (Fig. 10) [16]. Such a flexibil-ity also influences its adsorption characteristics onsolids and in turn affects the subsequent suspensionbehavior. Using a fluorescent labelled polymer andby monitoring the extent of excimer rormation itwas shown that the polymer at the interface couldhave a stretched or coiled conformation at the

1.3. Polymer conformation in the adsorbed statE

Polymers can exist in different conformationsboth in solution and in the adsorbed state. Theadsorption of polymeric materials onto solid sur-faces can be quite different from the adsorption of

~

0

)(

~u

0X

-- v~U"';pn:. u ~~6'0 63 Ol 56

55 49

3~ 35181 17 6 Parol

r 0 Helo!I 0 Parol

If

-D

>-~

tnZWa

za

~n-o:atnaa:

A' "-J

)00 . 00

5

O".fi) 100.00

RESIDUAL SULFONATE CONC.. kmol/"'Jx to

Fig. 9. Aggregation numbers determined at different adsorption densities shown along the isotherm.

P. Somasundaran. S. KrishnakumarjCol/oids Surfaces A: Plrysicochem. ERg. ~cts 93 ( 1994) 79-95 87

Moftomer

-~ -+ + + +--+

(a)E~im.r

i I ,,--~ H 4'

~ j -~:-: :,

;; - pH 7.1 ~ ""-D ~- -".. . -.; )~O .';0 .~ ~OO 5~O

d~vel.n9t.. (tom)

v\

Fig. 10. Schematic representation of the conformation of poly(acrylic acid) at different pH values: (al pH 4.4; (b) pH 10.5.

~60(

Fig. 12. Fluorescence emission spectra or adsorlxxl polymer attwo pH values.

1») ...", pH

- (-..0 .;.,.~-~-~--

0) 10. pH

,~ L~ -A . . . .

coit,"',oluIIOft. . . .C.po"G.G/'ol"r,.~

I.w pH

~~~-8 E:. . . .coiled/od 10' b cd

100" biftdiftO

interface depending on the pH. The rationalebehind the use of this technique is the observationthat the extent of excimer formation which dependson the interaction of an excited state pyrene of thepolymer pendant group with another pyrene groupin the ground state has a direct bearing upon thepolymer conformation. This may be understoodwith reference to Fig. 11 which shows that at lowpH there is a better probability for intramolecularexcimer formation between pyrene groups resultingfrom a favorable coiled conformation. Similarly alow probability for excimer formation at high pHmay be understood as a consequence of the repul-sion between the highly ionized carboxylate groupsin the polymer and the subsequent stretching ofthe polymer chain. This difference is reflected inthe nature of their fluorescence spectra as seen inFig. 12. Also these studies have demonstrated thatthe poly(acrylic acid) adsorbed in the stretchedform on alumina at high pH was essentiallyirreversible as the conformation could not bealtered by lowering the pH subsequently (Fig. 13)[17]. In contrast. the polymer adsorbed in thecoiled form at low pH did stretch out when thepH was increased.

We have also performed a detailed investigationon the Oocculation behavior of alumina particles

'-

~':tfi

,.N~p"C _r ~-

--~: ~ -..J?. . . .PGtlioll, capoftdcd/.d,.,bed

Fig. 13. Schematic representation of the adsorption process ofpyrene-labeled poly(acrylic acid! on alumina. A. At low pHpolymer is coiled in solution which leads to B. adsorption incoiled form. C. Subsequent rdising of the pH causes someexpansion of the p.>lymer. O. Polymer at high pH in solutionis extended and E. hinds strongly to the surface in thisconformation. {fl. Subsequcntlowering of pH d(JCS not allowfor sufficient intrastrdnd interactions for coiling to occur.

High pH

Fig. II. Schematic representation or the correlation or theextent or excimer rormation and intrastrand coiling or pyrene-labeled polyCacrylic acid).

P Somasundaran. S. KrishnakumarlCo/loidr Sur/a, A: Physicochem. Eng. Aspects 9~ /994) 79-95

with and without added polymer under fixed andshifted pH conditions. The polymer conformationwas shown to be a controlling factor of the floccu-lation process in this case. The results of fluores-cence emission studies using pyrene-labeled PAAin the adsorbed state under shifted pH conditions,i.e. adsorbing the polymer at a fixed pH and thenchanging the pH to a desired value, are shown inFig. 14. It may be noted that the excimer fluores-cence emission intensity of the solidjliquid equilib-ria after adsorption at high pH values remains thesame even after changing the pH to lower values.From this observation, coupled with the aboveresults, the variation of poly(acrylic acid) confor-mation at the solid-liquid interface under shiftedpH conditions may be represented as shown inFig. 15. It may be inferred that the polymer confor-mation at a given pH may be manipulated bycontrolling the adsorption conditions.

aqueous media. 7-Dimethylamino-4-methylcouma-rin which is a relatively hydrophilic molecule wasused to probe the adsorbed layer of DAPRAL GE202 (maleic anhydride-«-olefin copolymer withboth hydrophobic and hydrophilic side-chains) atthe alumina-toluene interface. The spectrum forthe probe in the absence of DAPRAL is similar tothat obtained in water solution (maximum emis-sion at 470 om. Fig. 16). This is attributed to thehydrophilicity of the bare alumina surface. Withan increase in polymer adsorption, the maximumwavelength shifts toward the shorter wavelengthrange reaching the value for hydrocarbon solvents(390 om) at 500 mg I~ I ofDAPRAL (Fig. 17). Thisobservation supports the mechanism of polymeradsorption through the interaction between theethylene oxide chains and the hydroxyl groupson the alumina surface with the hydrocarbonside-chains dangling toward the solution and thealumina particles eventually fully covered tobecome totally hydrophobic. This is also evidencedby the fact that the alumina suspensions arestabilized significantly by the adsorption ofDAPRAL.

2.4. Polymer orientation in non-aqueous media

Fluorescence spectroscopy was also used todetect polymer orientation on particles in non-

Fig. l4. Excimer to monomer ratio I,conditions.

m for alumina with 20 ppm poly(acr acid) as a runction or final pH under changing pH

P. SomoSImIlarall. S. Krishnakumar/Co/Io;df .\'IUfaces A: Physicochelrl. Eng. A.rp«ls 93 ( 1994) 79-9j 89

r-

I", f(.Q~"", -

+--t+-T(b) mediwn pH, 5-7

~ -+ - + + ++(a) low pH, 4

--+---(c) high pH, 10

--""""~+++-++

(f) low pH, 4

Fig. 15. Schematic representation of the variation of polymer conformation at the solid liquid interface under changing pHconditions.

120 480

I

j

J

tOO

460>-.~ 80.,oS~ 60:;~2g 40;;:

,440

20

'0~i 420"1J>.

~

40000 - 400

(DAPRAL),

600 800 I(MM)

Fig. t 7. Diagram illustrating the shift in maximum emissionwavelength of coumarin at the alumina-toluene interface withadsorption of DAPRAL.

Fig. 16. Fluorescence spectrum or 7-amino-4-methylcoumarinat the alumina-toluene interface in the presence or 0 ppmDAPRAL (curve al and 100 ppm DAPRAL (curve b).

3. Electron ~pin resonance ~pectroscopy (F5R) energy gap (AE) is given by

E = hv =gBHoESR studies as applied to micellar systems rely

on the sensitivity of a free radical probe to respondto its microenvironment. Molecular species with afree electron possess intrinsic angular momentum(spins), which in an external magnetic fieldundergoes Zeeman splitting. For a system with S =1/2, two Zeeman energy levels are possible whose

The magnetic moment of the free electron issusceptible also to the secondary magnetic momentsof the nuclei and thus the Zeeman splitting will bcsuperimposed by the hyperfine splitting whichbrings about further splitting of the absorptionsignal. The hyperfine splitting pattern depends on

90 P. SOIPfIISIBtdllrIJll, S. Krishnakunwr!Colloids SurfaCGf A: PhJ'sicodlem. ug. Aspects 9) ( 1994) 79-95

the spins and the actual number of the neighboringnuclei wilh spins. If the electron is in the field of aprolon lhen the ESR spectrum would yield twolines of equal intensity, and similar interaction by anucleus with S = I, as in nitrogen. would produce atriplet of equal intensity. The line shapes of ESRsignals are subject to various relaxation processes(spin-lattice and spin-spin relaxations) occurringwithin lhe spin system as well as anisotropic effectsdue lo the differentially oriented paramagnetic cen-ters being acted upon by an external magnetic field[18]. These effects result in a broadening of theabsorption lines. Three types of ESR study can beapplied to probe surfactant microstructures-spinprobing. spin labeling and spin trapping [19]. Inthe spin probing technique. a molecule with spin isexternally added to the system, whereas in spinlabeling a spin-bearing moiety. through covalentbonding, part a part of the molecule. The spintrapping technique is mainly used for the identifica-tion of radicals produced thermally, photochemi-cally or radiolytically by trapping the radical

through chemical reactions with a spin trap (likebutyl nitroxide) and converting the radical to a freeradical which can be examined by ESR. In thisstudy, three isomeric stable free radicals 5-, 12- and16-DOXYL-stearic acids were used as the spinlabels. Tbese spin labels were co-adsorbed individu-ally on the alumina along with the main adsorbate,sodium dodecyl sulfate, and the main regions of theisotherm were investigated.

Information on micropolarity and microviscos-ity can be obtained by measuring the byperfincsplitting constant AN and the rotational correlationtime fe. The latter is a measure of the time requiredfor a complete rotation of the nitroxide radicalabout its axis. Its value can be defined as the timerequired for the nitrox.ide to rotate through anangle of one radian. The hyperfine splitting con-stants of 16-DOXYL-stearic acid measured indodecylsulfonate solloids (hemi-micelles) (15.0 G)is indicative of a less polar environment in compari-son to its value for water (16.0 G) and SDS micelles(15.6 G) (Fig. 18). Similarly microviscosities esti-

M

\&J

~t-

~

~-J\oJ(I;

U-IoCtZ0

ctI-0

Fig. 18. Comparison or ESR spectra or 16-00XYL-stearic acid in solloids. mil.-ellcs and elhanol-glycerol mixtures and corresJKmdingrotational correlation times and viscositics.

P. SomosUlldluan. S. Kris/lnakllmQrjColloids Surlace~ Ph,'sicochem. Eng. AspecIS 91 ( 1994) 79-95

mated from tc measurements and calibratedagainst tc measured in ethanol-glycerol mixturesgive reasonably high values for the solloids relaliveto the values for water. Three different microviscos-ities were obtained using different probes in thesolloids, indicating that the nitroxide group in eachcase experienced a different viscosily within thesolloid. These observations may be explained byassuming a model for the adsorption of the probein which the carboxylate group is bound to thealumina surface. Such a model would require usto attribute greater mobility for the nitroxidemoiely near the SDS-H2O interface (as in the16-DOXYL-stearic acid case) and less mobility forthe 12-DOXYL and 5-DOXYL cases [20,21]. Thiswork is lhe first reported indication of variationsin microviscosity within a surfactant solloid asestimated by any known technique.

ence or surfactant, Aerosol-OT is shown in Fig. 19.A suc("'ession or flocculated, dispersed and floccu-lated states is observed as the amount of wateradded to the suspensions is increased. ESR studieswere conducted using 7-DOXYL-stearic acid as aprobe to get structural inrormation on theadsorbed layer. Fig. 20 shows ESR spectra or7-DOXYL-stearic acid adsorbed on alumina in thepresence and absence or adsorbed surfactant. Thevalues of AI obtained in the two cases, 69 G and65 G, respectively, are significantly different fromeach other, suggesting that despite the slow motionor the probe it is possible to distinguish betweenthe probe interacting directly with the mineralsurface and that hindered in its motion by thesurfactant molecules surrounding it.

Fig. 21 shows some of the spectra oblained atdifferent water concentrations with 7-DOXYL-stea-ric acid co-adsorbed with a full monolayer ofAerosol-OT on the alumina surface. The changesobserved in the ESR line shape correspond to anincrease in probe mobility consistent with a decreasein the ordering of the probe environment. Theseobserved changes are quantified by using the con-cept of order parameter S which can be calculatedfrom the spectrum using the following equation:

{AI - [A,(meas) + C]}S = 1.66

{AI + 2[A,(meas) + C]}

3.1. Conformation of adsorbed Aerosol-OT at the

alumina-cyclohe.xane interface [22.13 J

Colloidal dispersions in non-aqueous mediahave a number or technological applications andwater is presenl in most or these cases and plays amajor role in determining the dispersion behavior.The effect of water on the stability or a colloidalsuspension or alumina in cyclohexane in the pres-

Fig. 19. Effect of water 00 the alumina suspension stability at two differ.:nl surfactant concentr4tioos: solid symbols.8.5 x 10-3 M; empty symbols, [AOT] -26.5 x 10- J M.

AOT)=

<n2

~\Q

~ =:IQM...0;:)

~-;..

92 P. Somasundaran. S. KrishlwkumarfCoJ/oids Surfaces A: Ph,.sicochem. Mg. ~clS 91 ( 1994) 79-95

;;.;;;/'1. (a)rv- .0 = 0.2v..-200

--

Wo = 1.95\If r

N ~T,.(b)v - Wo = 7.25

JOG-Fig. 20. ESR spectra of 7-DOXYL-slearic acid probe adsorbedon alumina in the absence (al and presence (bl of Aerosol-OT.

AWo = 11.1where C= 1.45-0.019[A. - A,(meas)] G. The

order parameter is usually a parameter of molecu-lar motion. S varies between 0 (low order) and I(high order). The calculated parameter plotted asa function of water concentration is shown inFig. 22. It can be seen that as more water is addedto the suspension the order parameter decreaseduntil it reached a constant value of 0.75. Theseresults are interpreted in terms of an increase inthe probe lateral diffusion within the adsorbedsurfactant layer. Such an increase in the probelate~.al diffusion is realistic. considering the struc-tural reorganization of the complex adsorbed layerwhen water is present at the interface. At low waterconcentrations, water molecules bind the carbox-ylic groups of the stearic acid molecules and thepolar groups of Aerosol aT. directly to thehydroxyl groups of the alumina surface. This bind-ing limits the ability of the probe to move withinthe adsorbed layer and is consistent with a modelof localized adsorption where the adsorbed mole-cules have limited degrees of freedom. As the wateradsorption density increases. the carboxylic groups

I-lg. 21. Effect of water on the ESR spectra or 7-DOXYL-slearic acid co-adsorbed with a monolayer or Aerosol-OT althe alumina-cyclohexane inlcrrace.

of the probe molecules interact with water mole-cules not bound directly to the hydroxyl groups oftbe mineral surface. Similar interactions betweenthe polar groups of Aerosol-OT and water mole-cules are most likely, as a result of which theadsorbed layer can become loosely bound tothe mineral surface. The molecular diffusion limitedby the binding of the surfactant molecules at lowwater concentrations thus increases markedly aswater adsorbs on particles.

4. Raman spectroscopy

Raman spectroscopy is acclaimed as a vibra-tional technique well suited for studies of aqueous

P. .'iomasundaran. ,'i. Krishnakumar;Co//oids Surfaces A: Physicochem. Eng. Aspecl.r 91 (/994) 79-9593

(/)

~~-OJECG~CG

no~~

"0~

0

o.

environments compared to IR owing to the neartransparency of the former and the ease with whichthe whole vibrational region of interest can becovered. Vibrational frequencies of molecules inthe bound state should be different from those inthe free state. Also. it is susceptible to changes in thesymmetry properties of the environment. Very fewdefinitive studies exist in the literature concerningthe Raman spectroscopy of surfactants in solution.Raman investigations of surfactant adsorbates arereported as surface-enhanced Raman studies at themetal colloid-or metal electrode-liquid interface.

The alumina--dodecylsulfate system has alsobeen probed by excited state resonance Ramanspectroscopy using tris(2.2'-bipyridyl) ruthen-ium( II) chloride, Ru( bpy)~ +, as a reporter mole-cule [24]. It has been shown that rutheniumpolypyridyl complexes serve as excellent photo-physical probes for biopolymers like nucleic acids.The excited state of Ru( bpy)~ + shows strong reso-

nance-enhanced Raman transitions when probedat 355 nm. Furthermore, it has been shown thatbinding of this ion to clay particles results insubstantial changes in the ground state transitionsof the excited state resonance Raman spectrum.

For these reasons this probe was chosen to studythe solloids formed at the alumina-water interfacewith excited state resonance Raman spectroscopy.

Raman spectra of Ru(bpy)~ + above the CMCshow frequency shifts as well as intensity changesas compared to its spectrum in water. Table I liststhe relative intensities of various lines with respectto the line at 1286 cm -( in water. These transitionscan be attributed to the perturbation of the excited

-- - --0 100 200 300

Residual Water Concentration x 103, mole/I

94 P Somasundaran, .so, KrishnakumarlCo/loids surfaces,4.' Phy.ricochem. Eng, A..rpeCIS 93 ( 1994) 79-95

state by the SDS micelles. The excited state reso-nance Raman spectrum of this probe in variousregions of the adsorption isotherm for thealumina-SDS system is shown in Fig. 23. The

spectrum of Ru(bpy)~+ on alumina in the absenceof SDS is very similar to its spectrum in water bothin terms of frequencies and relative intensities. Thistrend is continued into Region II, where the solloidaggregation process starts. In Regions III and IV,the Raman spectrum shows significant changes. Thefrequency shifts are more pronounced than in thecase of micelles. A plot of change in wavenumbersfor some of the lines in the four different regions ofthe adsorption isotherm is shown in Fig. 24. Thechanges in Raman frequency and intensity assumesubstantial significance in the transition Regions IIand III onward only. This could be due to thechange of net charge on the alumina surface frompositive to negative. The favorable net negativecharge enhances the adsorption of the probe at thesolid-liquid interface. Accordingly, no adsorptionof Ru(bpy)~ + was observed when the supernatantswere analyzed in Region I and Region II or in theabsence of SDS. These results indicate that adsorp-tion of Ru( bpy)~ + on to alumina becomes signifi-cant only close to the point of zero charge. Thetransitions at 1213, 1286 and 1428 cm-1 show sig-nificant increments and these trends clearly suggestthe probe adsorption on to the solloids. Also it was

7;00 1213 em-1A 1286elR-1

8.0 D 1428_-1

'e2(I)~-.t1&1Q.Z-.t~-.tIt:Z-.- .3:(I)

~

Fig. 24. Frequency shifts or resonance Raman lines orRu(hpYlj' as a function of SOS l.'Onccntration for analumina-SDS system.

P. Somasundaran. ,5. Krifhnakumar/CoUoidr Surfaces A: Ph}'sicochem. Eng. Aspects 93 ( /994) 79-95

seen that the 1286 peak shifted to 1281 in thepresence of solloids while it was practicallyunchanged in the SDS micelles. it may be speculatedthat Ru(bpy)~+ may be sensing different environ-ments within the SDS micelles and thealumina-SDS solloids. Implicit in these results isalso the potential of time-resolved Raman spectro-scopy as a powerful diagnostic tool to explore thesolid-liquid interface.

Ackno""cdgmenls

The authors wish to acknowledge the NationalScience Foundation, the Department of Energy,Unilever Research, Engelhard Co., NALCO andARCO for their financial support.

References

5. Conclusions

The use of various spectroscopic techniques forstudying the adsorbed layers of surfactants andpolymers in aqueous and non-aqueous systems iselucidated here. A number of new findings arereported with regard to the structure and evolutionof surfactant and polymer solloids by an integratedapproach using the classical bulk property meas-urements and the modern spectroscopic tech-niques. The aggregation number of SDS onalumina was determined at various points duringsolloid evolution and the variation of the microvis-cosity within the solloid layer determined alongthe isotherm. Fluorescence spectroscopic studiesalso indicate that steric hindrance to packing inthe adsorbed layers controls solloid formationof the alkylxylenesulfonates. The effect ofpH-dependent conformational equilibria of poly-(acrylic acid) on the adsorption process and itsimplications to flocculation have been explored tosuggest reasonable conformational structures forthe polymer solloid. Finally, the role of water incontrolling the dispersion properties of aluminadispersed in cyclohexane by increasing the lateraldiffusion of the adsorbed surfactant molecules wasdepicted using the ESR technique. [t is clear thatthere is a variety of techniques now becomingavailable to examine structures in situ at levelssmaller than ever before. With technologies nowbeing developed to examine interfaces at atomicscales and dynamics of these atoms and moleculesat pico and femto scales, opportunities will ariseto visualize processes such as aggregation andreaction between various species at interfaces asthey take place.

[1] P. Somasundaran, AIChE Symp. Ser., (1975) 71.[2] P. Somasundaran, T.W. Healy and D.W. Fuerstenau,

J. Phys. Chem., 68 (1964) 3562.[3] J.R. Lakowicz, Principles of Fluorescence Spectroscopy,

Plenum, New York, 1983.[4] J.K. Thomas, The Chemistry of Excitation at Interfaces,

American Chemical Society Monograph, 1984.[5] N.J. Turro, G.S. Cox and M.A. Paczkowski, Top. Curro

Chem., 129 (1985) 57.[6] P. Somasundaran and D.W. Fuerstenau, J. Phys. Chem.,

70 (1966) 90.[7] A.M. Gaudin and D.W. Fuerstenau, Min. Eng., 7

(1955) 66.[8] P. Somasundaran and D.W. Fuerstenau, J. Phys. Chem.,

70 (1966) 90.[9] P. Somasundaran and J.T. Kunjappu, Colloids Surfaces,

37 (1989) 245.[10] J.H. Harwell and D. Bitting, Langmuir, 3 (1987) 500.[11] P. Somasundaran, P. Chandar and NJ. Turro, Colloids

Surfaces, 20 (1986) 145.[12] J.N. Demas, Excited state lifetime measurements,

Academic Press. New York, 1983.[13] P. Chandar, P. Somasundaran and N.J. Turro, J. Colloid

Interface Sci., 117 (1987) 31.[14] A. Sivakumar and P Somasundaran, Langmuir, 10

(1994) 131.[15] H. Morawetx, Macromolecules in Solution, Wiley, New

York, 1975.[16] K.S. Arora and NJ. Turro, J. Polym. Sci, 25 (1987) 243.[17] P. Chandar, P. Somasundaran, K.C. Waterman and

NJ. Turro, Langmuir, 3 (1987) 298.[18] J.H. Wertz and J.R. Bolton, Electron Spin Resonance-

Elementary Theory and Practical Applications, Chapmanand Hall, New York, 1986.

[19] B. Ranby. in J.K. Rabek (Ed.), ESR Spectroscopy inPolymer Research, Springer-¥erlag, Berlin, 1977.

[20] K.C. Waterman, NJ. Turro, P. Chandar andP. Somasundaran, J. Phys. Chem1 90 (1986) 6829.

[21] P. Chandar, P. Somasundaran, K.C. Waterman andNJ. Turro. J. Phys. Chem.. 91 (1986) 150.

[n] C.A. Malbrel. D. Eng. Sci. Thesis, Columbia University,New York, 1991.

[23] C.A. Malbrel and P. Somasundaran, Langmuir, 8 (1992)

1285.[24] P. Somasundaran. JT. Kunjappu, C.¥. Kumar,

NJ. Turro and J.K. Barton, Langmuir, 5 (1989) 215.