raman spectroscopic characterization of the interlayer structure of na + ...

Raman Spectroscopic Characterization of the Interlayer Structure of Na+-MontmorilloniteClay Modified by Ditetradecyl Dimethyl Ammonium Bromide

Elena A. Sagitova,†,‡ Patrice Donfack,*,‡ Kirill A. Prokhorov,† Goulnara Yu. Nikolaeva,†

Viktor A. Gerasin,§ Nadezhda D. Merekalova,§ Arnulf Materny,‡ Evgeny M. Antipov,§ andPavel P. Pashinin†

A.M. ProkhoroV General Physics Institute of RAS, A.V. TopchieV Institute of Petrochemical Synthesis of RAS,Moscow, Russia, School of Engineering and Science, Jacobs UniVersity, Bremen, Germany

ReceiVed: NoVember 14, 2008; ReVised Manuscript ReceiVed: February 18, 2009

Raman spectroscopy has been applied for the rapid and nondestructive monitoring of the interlayer structureof sodium montmorillonite (MMT) clay modified by ditetradecyl dimethyl ammonium (DDA+) bromide.This work demonstrates that a detailed analysis of Raman spectra in the fingerprint region (600-1600 cm-1),in combination with model simulation, allows one to distinguish different conformational states of DDA+ inthe interlayer space of the modified clay, namely, a liquidlike state but rich in trans conformers, disorderedconformational states, and a crystallike conformation appearing at increasing modifier content. Theseconformations differ in the angle between their alkyl chains, the relative content of trans and gauche conformersand the relative length of trans segments. The shape and width of the Raman band at 1300 cm-1 and the peakintensity ratio I1088/I1064 can be used for a qualitative analysis of the ratio of gauche/trans conformers. Theintegral intensity ratios I*1064/I*1300 and I*1300/I*705 help to determine the proportion of trans conformers andthe content of the modifier in the clay, respectively, thereof providing quantitative characterization of themodified clay (conformational reorganization and modifier content). Noteworthy, the transition from a liquidliketo crystal-like conformation is further supported by the splitting of the symmetric C-C stretching Ramanband of the trans segments within the alkyl chains at 1133 cm-1 (liquidlike conformation) into two modes at1124 and 1135 cm-1 corresponding to two parallel trans chains of nonequivalent lengths (crystal-likeconformation).

Introduction

Clays modified by alkylammonium surfactants are widelyused as fillers for polymer-clay nanocomposites,1-3 absorbentsfor the treatment of contaminated waste streams,4 as well asmodels for the study of chain aggregation in biomembranes.5

Modified clays offer “sandwich” structures in which alkylam-monium molecules can form their own organophilic layersembedded in the clay interlayer galleries (space between silicateplates of clay crystallites). The structure of the organic layersformed in the galleries strongly depends on the nature of themodifier and its content.1,6-9 The knowledge of the behavior oforganic molecules embedded in the interlayer space is of greatimportance for both scientific and industrial applications. Forinstance, as far as the fabrication of polymer-clay nanocom-posites is concerned, the modification of clay by surfactantsallows the penetration of nonpolar polymer macromolecules intothe clay interlayer space. Therefore, it is of great importance toknow the structure of the organophilic layers in the modifiedclay, which provides the best penetration of polymer macro-molecules within the interlayer galleries.2,6,9

Several experimental methods are applied to obtain structuralinformation about organophilic layers in modified clay. X-raydiffraction allows for the estimation of the height of the clayinterlayer space, which significantly changes during the modi-

fication process.1,2,6-9 On the basis of X-ray diffraction data,the types of the organophilic layers (lateral monolayer, lateralbilayer, and paraffin monolayer or paraffin bilayer) are defined.The application of differential scanning calorimetry (DSC) forstudying the structure of modified clays helps to distinguish onlybetween the amorphous and crystalline states of the surfactantmolecules in the clay interlayer space. However, neither X-raydiffraction nor DSC techniques provide any information aboutthe conformational order of embedded molecules.1,6 In order tostudy the conformations of the absorbed molecules, Fouriertransform infrared spectroscopy (FTIR)6,10 and recently Ramanspectroscopy7,8 have been applied. Raman spectroscopy offersmore advantages than IR adsorption, because neat clay (claywithout surfactants) yields a featureless Raman spectrum, whichdoes not mask the Raman spectrum of the surfactants.7,11

According to experimental results obtained by FTIR and Ramanspectroscopy,6-8,10 the molecular conformation of alkyl chainsin the organic layers of the modified clays depends on thepacking density, temperature, chain length, and content of themodifier and can vary from disordered chains containingnumerous gauche conformers to well-ordered chains in all-transconformation. However, previous Raman studies on the behaviorof alkylammonium surfactants in modified clays were limitedto very few surfactants that include dioctadecyl dimethylammonium bromide (DODAB)7 and cetyl trimethyl ammoniumbromide (CTAB).8 Additionally, these studies were restrainedto the spectral region 2750-3200 cm-1. In this region, Ramanand FTIR spectra provide a qualitative characterization of the

* To whom correspondence should be addressed. E-mail: [email protected]. Tel. +494212003569. Fax: +494212003229.

† A.M. Prokhorov General Physics Institute of RAS.‡ Jacobs University.§ A.V. Topchiev Institute of Petrochemical Synthesis of RAS.

J. Phys. Chem. B 2009, 113, 7482–74907482

10.1021/jp810050h CCC: $40.75 2009 American Chemical SocietyPublished on Web 05/04/2009

conformational order, but a quantitative analysis is ratherdifficult due to overlapping bands.

In this work, we present the first Raman spectroscopicdetermination of the interlayer structure of sodium montmoril-lonite clay modified by ditetradecyl dimethyl ammoniumbromide. Moreover, we demonstrate that Raman scattering canbe efficiently applied in the spectral region 600-1600 cm-1

for both qualitative and quantitative characterization of theconformational order of alkylammonium surfactants in modifiedclays. In fact, contrary to the previously investigated spectralregion 2750-3200 cm-1, the region 600-1600 cm-1 mostlycontains nonoverlapping Raman lines sensitive to conforma-tional reorganization of embedded molecules in the clayinterlayer space.

Experimental and Simulation Methods

Preparation of Modified Clay. To produce our samples, weused Na+-montmorillonite clay (MMT) with a cation exchangecapacity (CEC) of 95 mgeq/100 g and ditetradecyl dimethylammonium bromide (DDAB, [(CH3(CH2)13)2N+(CH3)2Br-] asmodifier. The mass contents of ditetradecyl dimethyl ammoniumions (DDA+) in the DDA+-MMT samples were 19% (0.50CEC), 26% (0.75 CEC), 33% (1.00 CEC), 42% (1.50 CEC),and 50% (2.00 CEC). The numbers in brackets give the DDA+

contents expressed through the cation exchange capacity of theclay.

Modified clay samples were synthesized by a cation-exchangereaction between equimolar suspensions of Na+-MMT anddifferent portions of DDAB solution in water. Before claymodification, the Na+-MMT powder (250 mg) was dissolvedin deionized H2O (500 mg) at room temperature. The suspensionof Na+-MMT in water was homogenized using an ultrasonicdisperser (UZDN-A, Russia) for at least 1 h. Different propor-tions of the DDAB solution (2.5 × 10-3 mol/L) in hot deionizedwater (70-80 °C) were added to the Na+-MMT suspensionsunder constant stirring with subsequent filtration and doublewashing of the filtered rest by deionized H2O. The DDA+

content in the obtained suspensions of DDA+-MMT wasdetermined after 48 h following the preparation of the DDA+-MMT suspensions, using an original method described in detailsearlier.7

For X-ray diffraction and DSC studies, the DDA+-MMTsuspensions were dried in an evacuated oven at room temper-ature for 3 h to obtain powder samples; the DDA+-MMTsuspension samples for X-ray experiments were previouslyapplied to glass slides. Finally, for Raman interrogation theobtained powder (from each suspension) was pressed during 5min under 3 MPa. In the final products all DDAB moleculeswere attached to silicate plates of the clay and were mainlylocated in the interlayer galleries in accordance with results ofsimilar manufacturing technique reported earlier.2

The Raman Experiment. For the Raman experiment we usedan Ar+ ion laser (Inova 308, Coherent Inc., U.S.A.), a Triaxsingle monochromator (TRIAX 550, Jobin Yvon, France)equipped with a liquid N2 cooled CCD detector (CCD 3500,Jobin Yvon, France), and a microscope setup equipped with anefficient high magnification and high numerical aperture objec-tive (×100, N.A. 0.95; Nikon, Japan). The second Raman setupused for reference measurements included an Ar+-Kr+ ion laser(Stabilite 2018, Spectra-Physics, U.S.A.), a double monochro-mator (U1000, Jobin Yvon, France), and a cooled photomulti-plier detector.

To record Raman spectra of the modified clay (DDA+-MMT),the Ar+ ion laser line at 488.0 nm was used. Special brassy

plates were used as sample holders. The scattered light wascollected in the backscattering geometry using the ×100objective. Signal dispersion was achieved by the TRIAXmonochromator equipped with a 1200 grooves/mm diffractiongrating, setting the slit width to 100 µm, and detected by theliquid N2 cooled CCD detector. The laser power at the sampledid not exceed 4 mW. The spectra were recorded with exposuretimes from 20 to 120 s and accumulated 2 to 10 times dependingon the sample. Spectral analysis was easily achieved using twocurve fitting routines (the second routine was not mandatorybut just performed for precision). Before fitting, a fourth orderpolynomial background was first subtracted from the experi-mental spectra. First, and only when necessary, an easycomputerized deconvolution of the spectra was achieved usingour NGLabSpec software package for automated spectralacquisition, which provides a weighted sum of Gaussian andLorentzian functions. Second, the spectra were also subjectedto a homemade program based on a nonlinear least-squaresfitting and the Levenberg-Marquardt algorithm. This programalso uses a weighted sum of Gaussian and Lorentzian functions.The results of the two ways of spectra deconvolution havecoincided within the accuracy of Raman measurements.

For the interpretation of the Raman spectra of modified clay,we also recorded reference Raman spectra of solid alkylam-monium surfactants, which include DDAB (the modifier),DODAB (CH3(CH2)17)2N+(CH3)2Br-), CTAB (CH3(CH2)15N+-(CH3)3Br-), and n-alkanes (CnH2n+2; in this work, only spectracorresponding to n ) 10, 14 will be shown). These spectra weremeasured under Ar+-Kr+ laser excitation at 472.7 nm with a90° scattering geometry. The spectral resolution for the Ramanexperiments was estimated to be approximately 2 cm-1. Thearomatic breathing mode of toluene at 1003.7 cm-1 was usedfor wavenumber calibration.

X-ray Diffraction and Differential Scanning Calorimetry(DSC). X-ray spectra (in back scattering geometry) of themodified clays (exposed to air) were recorded using a DRON-3spectrometer (Russia) with a CuKR excitation at 0.154 nm,monochromated by a graphite monocrystal, at room temperatureand a relative air humidity of 70%. Thermograms of the air-exposed modified clay were obtained by a differential scanningmicrocalorimeter (DTAS-1300, Russia) in the temperature rangefrom 10 to 180 °C above the melting point of DDAB (110 °C)but far below the melting point of clay. The heating rate wasequal to 8 °C/min. The weight of the samples for DSC wasabout 20 mg.

Mathematical Modeling. Computer-simulated modeling ofDDA+ conformations in DDAB monocrystal and in modifiedclay tactoids was carried out by molecular dynamics methodsbased on the minimization of the potential energy as follows:

EDDA+-MMT ) EMMT + EDDA+ + EMMT,DDA+ (1)

where EDDA+-MMT, EMMT, and EDDA+ are the potential energiesof DDA+-MMT, MMT tactoids, and DDA+, respectively. Theterm EMMT,DDA+ is the energy of interaction between MMT andDDA+.

Each term, EDDA+-MMT, EMMT and EDDA+ in eq 1 is describedby the sum of the harmonic bond lengths and angles terms, the3-fold torsional potentials, 6-12 van der Waals interactions, andthe Coulombic electrostatic potential. The charges of each atomwere assigned by charge equilibration method. This method isrecommended for use with the Universal force field (Rappe,A. K.; Casewit, C. J.; Colwell, K. S.; Goddard, W. A.; Skiff,W.M.; J. Am. Chem. Soc. 1992, 114, 10024). Molecular dynam-ics simulations were carried out on SGI workstation for

Raman Characterization of Na+-MMT Clay Modified by DDAB J. Phys. Chem. B, Vol. 113, No. 21, 2009 7483

100-300 ps at 375 K with time step of 0.001 ps until the totalpotential energy became nearly constant for the last 40-100ps. The potential energy of eq 1 was minimized until the root-mean square (rms) gradient became 0.1 kcal/mol/A. Further-more, five or six snapshots were extracted from each moletrajectory file during the last 40-100 ps, within which thesystem became equilibrated and the potential energies wereminimized until the rms gradient became less than 0.01 kcal/mol/A. For the minimization, we have used a combination ofmethods, starting with the steepest descent method, followedby the quasi-Newton method and ending with the truncatedNewton-Raphson minimizer.

The models for the DDA+ ion conformations in the interlayergalleries of the MMT were obtained under the assumption thatMMT crystallizes in the C2/m space group of the monocliniccrystal lattice with lattice parameters a ) 5.20 Å, b ) 9.20 Å,and c ) 10.13 Å.12 The early steps of this computation yieldednumerous configurations (minimization of eq 1), which stronglydepend on initial conditions (interlayer spacing, angles betweenalkyl chains, relative amount of trans conformers, etc.). Thenumber of optimum configurations could be reduced by applyingspecial limiting conditions, provided by our structure-sensitiveexperiments (X-ray diffraction and Raman scattering).

Results and Discussion

X-ray Diffraction, DSC, and Mathematical Modeling.Figure 1 shows X-ray spectra (Figure 1A) and DSC thermo-grams (Figure 1B) of MMT clay modified by DDAB. Theanalysis of the X-ray spectra and thermograms was similar toour previous X-ray and DSC studies of the MMT modified byDODAB and CTAB.2 More detailed principles of X-ray spectralanalysis of modified clays can be also found in earlier reports.1,10

Similar to MMT modified by DODAB or CTAB,2 withincreasing DDA+ ion content the (001) first-order basal reflec-tions (marked by the asterisks in Figure 1A) in the X-ray spectraof DDA+-MMT shifts toward smaller reflection angles. Fromthe analysis of the spectral characteristics (position, intensity,shape, etc.), the nonmarked peaks in the X-ray spectrum ofDDA+-MMT (e.g., at DDA+ content of 2.00 CEC) wereassigned to second-order basal reflections.2 The height of theclay interlayer space was calculated from the position ofthe first-order basal reflections as described previously.1,10 Thetype of the organophilic layers was determined by comparingthe interlayer space height and the length of the DDAB moleculein the monocrystal. Figure 1B shows the DSC thermograms of

DDA+-MMT at different DDA+ content; a thermogram of solidDDAB powder is also included and reveals the existence ofthree melting peaks. As it can be seen from Figure 1B, anadditional melting peak appearing at higher temperature for thehighest DDA+ (2.00 CEC) content could be due to DDABclusters in the interlayer space, since it vanishes after rinsingthe samples in water.2

X-ray diffraction and DSC data corresponding to the differentmass contents of the modifier are summarized in Table 1. Thedata from X-ray analysis refer to the height of the clay interlayerspace and are simple and straightforward; in fact, an increasein the DDA+ content leads to gradual filling of the clay interlayerspace with DDA+. This results in the growth of the spacingbetween the layers. As aforementioned, this change of interlayerspaces in modified clay with increasing surfactant content wasalready observed for MMT modified by DODAB in our previousstudy.2

Similar to our previous studies on DODAB,7 DSC measure-ments of the modified clay (DDA+-MMT) showed that at a lowDDA+ content no traces of DDA+ ions melting could beobserved. The peak at 23 °C appeared at 1.00 CEC of DDA+

content. This temperature is much less than the melting pointof solid DDAB powder (see Figure 1B). Further increase inthe modifier content slightly shifted this peak to highertemperatures. This points to the fact that the structure formedby DDA+ in the interlayer galleries of DDA+-MMT is liquidlikeat low DDA+ content less than 1.00 CEC, and becomes morecrystal-like with increasing DDA+ content. The melting peaksobserved remained however below the melting points of solidDDAB. Apparently, such a crystal-like structure is less perfectand/or the crystallites are smaller than in the case of solidDDAB.

As already mentioned in the previous section, the outcomeof computer-simulated modeling strongly depends on initialconditions. Hence, for modeling DDA+ conformations in theMMT interlayer space, we have to use the data obtained fromX-ray diffraction and Raman studies as initial magnitudes ofthe heights of the clay interlayer space and the relative amountof trans conformers, respectively. The initial magnitudes of theangle between alkyl chains of DDA+ ions in the MMT interlayerspace were considered to be equal to that found in DDABmonocrystals. The number of trans conformers was computedin two different situations regarding the minimum length of transsegments. In the first case, one trans zigzag consisted of at least4 C atoms (T-conformer), and in the second case, a trans zigzagconsisted of at least 5 C atoms (TT-conformer).The computa-tional accuracy of the number of trans conformers correspondingto different simulated conformational states was 8%. As willbe shown later, within this computational accuracy, potentialconformational states with respective proportions of transconformers, obtained with our Raman and X-ray data appliedas initial conditions were selected, which showed good agree-ment with the Raman experiment.

Figure 2 displays the most probable computer-simulatedconformations of DDA+ ions in DDAB monocrystal and in themodified clay tactoid obtained using as initial conditions theinterlayer heights provided by X-ray diffraction and the differentproportions of trans conformers provided by Raman scatteringmeasurements directly related to the conformal order of thealkylammonium surfactants in the modified clay. We found thatDDA+ ions in DDAB monocrystals adopt a conformation asdepicted in Figure 2a. This is consistent with the findingsreported in the literature about the alkylammonium ions chainconformation in DDAB13 and in DODAB14 monocrystals. For

Figure 1. X-ray spectra (A) and DSC thermograms (B) of MMT claymodified by DDAB.

7484 J. Phys. Chem. B, Vol. 113, No. 21, 2009 Sagitova et al.

the DDA+-MMT tactoids, it is possible that DDA+ ions existat any of the conformations shown in Figure 2a-d. The mostreliable conformations would however be determined by themass content of the modifier. In fact, the conformation shownin Figure 2b was selected as a potential conformational stateby reconsidering insightful information provided by the Ramanspectrum of DDA+-MMT at low DDA+ ion content as will beexplained later.

Raman Characterization in the Region (600-1600 cm-1).Raman scattering can be efficiently implemented for bothqualitative and quantitative characterization of the conforma-tional order of alkylammonium surfactants in modified clays.The assignment of the main Raman modes of DDA+-MMT andsolid DDAB in the spectral regions 600-1600 cm-1 and2750-3200 cm-1 are summarized in Table 2. The assignmentsare made based on earlier assignments of Raman spectra of solidalkylammonium surfactants15 and primary amines,16 neat clay,17

polyethylene,18-22 and n-alkanes.20-24 However the assignment

of the bands at 1064 and approximately 1130 cm-1 was doneusing the more detailed work reported by Gall et al.19 and alater study carried out by Snyder et al.24

Raman spectra of DDA+-MMT in the region 600-1600 cm-1

change dramatically with increasing DDA+ ion content. TheRaman spectra of DDA+-MMT with different mass content ofDDA+ ions and solid DDAB in this region are shown in Figure3. All spectra were normalized to the peak intensity of the CH2-twisting vibrational mode at 1300 cm-1. Spectral details of theRaman modes around 1300 cm-1 (bandwidth) and 1130 cm-1

(bandsplit), and the experimental and simulated estimations ofthe proportions of trans conformers in DDA+-MMT arecompiled in Table 3. The bandwidth of the Raman mode around1300 cm-1 and the peak intensity ratio I1088/I1064 were determineddirectly from the experimental spectra (without any decomposi-tion; a simple zoom was sufficient). The spectral details of theRaman modes around 1130 cm-1 and the respective peakintensities were determined by applying spectral fitting decom-position. The bands at 705 and 1300 cm-1 do not overlap withother lines and were fitted separately after cutting off a narrow

TABLE 1: X-ray Diffraction and DSC Measured Parameters for the Modified Clay DDA+-MMT

DDA+ ion content X-ray diffraction data

cation exchangecapacity of modified

clay (CEC)mass content

DDA+/DDA+-MMTrelative mass

ratio DDA+/MMT

height of theclay interlayer

space (nm)

type of the structureof the clay

interlayer space

DSC peak in therange from 10 to

180 °C, (°C)

0 (neat clay) 0% 0 1.25 empty layers absent0.50 19% 0.24 1.77 monolayers absent0.75 26% 0.36 2.12 bilayers absent1.00 33% 0.49 3.42-3.98 bilayers 231.50 42% 0.74 3.27-3.60 bilayers 302.00 50% 1.00 3.42 bilayers 31

Figure 2. Computer-simulated conformation of DDA+ ions in DDABmonocrystal (a) and in the interlayer space of the modified clay DDA+-MMT tactoids (a-d).

TABLE 2: Tentative Assignment of the Raman Bands Observed in Modified Clay DDA+- MMT at Different DDA+ IonContent and in Solid DDAB

Raman bands (cm-1)

DDA+-MMT at different DDA+ content

0.50 CEC 0.75 CEC 1.00 CEC 1.50 CEC 2.00 CEC solid DDAB tentative assignment

706 706 706 706 706 absent ν1(A1) of SiO4 in clay1064 1064 1064 1064 1064 1064 C-C asymmetric stretch, trans conformers

1070-1090 broadband 1088 C-C asymmetric stretch, gauche conformers1133 1133 1132 1130 1124 1124 C-C symmetric stretch, trans conformersabsent absent absent absent 1135 1135 C-C symmetric stretch, trans conformers1299 1299 1299 1300 1297 1297 CH2 twist, both trans and gauche conformers1445 1443 1445 1441 1439 1437 CH2 bend2852 2852 2853 2854 2852 2852 CH2 symmetric stretch2883 2886 2885 2888 2886 2885 CH2 asymmetric stretch

Figure 3. Raman spectra of solid DDAB (top) and of the modifiedclay DDA+-MMT with DDA+ ion content from 0.50 to 2.00 CEC.


spectral region around each of them. The band at 1300 cm-1

was deconvoluted into two lines.Qualitative Characterization of the Conformational Or-

der. In the spectral window (600-1600 cm-1), we found atleast two spectral features, which permit to qualitatively describethe conformational order of alkyl chains in DDA+-MMT.

First of all the position, shape, and width of the Raman bandaround 1300 cm-1 are influenced by the DDA+ ion content, ascan be seen in Figure 3 and in Tables 2 and 3. This band is asuperposition of two vibrations: one at 1297 cm-1 (twistingvibration of CH2 groups in trans conformers of alkyl chains)and the other one at 1305 cm-1 (twisting vibration of CH2 groupsin alkyl chains with numerous gauche conformers). The integralintensity of this band is proportional to the total content of alkylchains.18 The increasing number of gauche conformers in alkylchains is accompanied with an increasing asymmetry of the 1300cm-1 band on its high-frequency side.18 In the Raman spectraof DDA+-MMT, with increasing DDA+ ion content from 0.50up to 1.50 CEC we have observed the increasing asymmetry ofthe band at approximately 1300 cm-1 on its high-frequency sideaccompanied by a band broadening from 12 to 23 cm-1.However, with further increase of the DDA+ ion content from1.50 to 2.00 CEC, the width of this band suddenly decreasesfrom 23 to 19 cm-1 and its position shifts from 1300 to 1297cm-1 (see Figure 3, Table 3). Therefore, in the DDA+-MMT,the number of gauche conformers increases with increasingDDA+ ion content from 0.50 CEC up to the critical value of1.50 CEC and then decreases from 1.50 to 2.00 CEC.

Second, we observed that the peak intensity ratio I1088/I1064

is strongly affected by the DDA+ ion content as depicted inFigure 3. Figure 4 shows the dependence of the ratio I1088/I1064

on the DDA+ ion content in the modified clay (bottom trace);the trace at the top, included for comparison, represents a similardependence for two other bands in the spectral window(2750-3200 cm-1) and will be discussed later. The Raman bandat 1064 and the broadband at approximately 1088 cm-1 areasymmetric stretching vibrations of C-C bonds in transsegments of alkyl chains and in alkyl chains with numerousgauche conformers, respectively.15,16,18 The ratio I1088/I1064 isexpected to increase with increasing number of gaucheconformers.16,18 As shown in Figure 4, the dependence of theratio I1088/I1064 on DDA+ ion content is not monotonic. The ratioincreases in agreement with literature,16,18 but it then decreasesat a DDA+ ion content of 2.00 CEC. Therefore, it can beconcluded that the number of gauche conformers increases withincreasing DDA+ ion content from 0.50 to 1.50 CEC and

decreases from 1.50 to 2.00 CEC. This perfectly agrees withthe CH2-twisting vibration (∼1300 cm-1) band broadeningindicating the conformational rearrangement (transition fromtrans to gauche conformers) discussed above.

Quantitative Characterization of the ConformationalOrder. The number Rtrans of trans conformers in the alkyl chainswas calculated using the formula

Rtrans ) kI*1064

I*1300, k ) 1.3 (2)

where I*1064 and I*1300 are the integral intensities of the Ramanbands at 1064 and 1300 cm-1, respectively. The coefficient kwas determined from the ratio I*1064/I*1300 for solid DDAB underthe assumption that the alkyl tails of the DDA+ ions adopt theconformation shown in panel a of Figure 2 as discussedabove13,14 with 85% of trans conformers in the CH2-chains. Thenumber Rtrans of trans conformers at different DDA+ ion contentsare compiled in Table 3. Rtrans decreases with increasing DDA+

content from 0.50 to 1.50 CEC and increases from 1.50 to 2.00CEC. This perfectly agrees with the CH2-twisting vibration bandbroadening and the behavior of the ratio I1088/I1064 discussedabove. As it can be seen from Table 3, with the exception ofthe 0.50 CEC DDA+ content the values Rtrans obtained fromthe Raman experiment (calculated using eq 2) and from thesimulation are in good agreement. The deviation between theexperimental result and the theoretical estimation at 0.50 CECmodifier content could be due to an angle, between alkyl chainsof DDA+ ions in the MMT interlayer space different at this

TABLE 3: Spectral Characteristics of the Raman Modes around 1300 and 1130 cm-1 and the Experimental and SimulatedEstimations of the Relative Content of Trans Conformers in DDA+-MMT

DDA+-MMT at different DDA+ ion content

samples 0.50 CEC 0.75 CEC 1.00 CEC 1.50 CEC 2.00 CEC solid DDAB

bandwidth ∆ν (cm-1) of the Raman band at ∼1300 cm-1 12 12 17 23 19 7trans -conformations (Raman experiment) 70- 76% 60-62% 59-64% 43-48% 57-64% 85%trans - conformations

(Computer simulation with 5 C atoms in trans segments)21% (77%)a 49% 61% 44% 63% 75%

trans -conformations(Computer-simulated with 4 C atoms in trans segments)

43% (75%)a 55% 65% 52% 67% 75%

Raman band 1130 cm-1 bandsplit bandsplitband position ν (cm-1) 1133 1133 1132 1130 1124/1135 1124/1135bandwidth ∆ν (cm-1) 11 11 16 20 10/14 8/6intensity I1130 1.03 0.78 0.53 0.26 0.21/0.25 0.30/0.30

a These values were recalculated with the initial configuration considered to describe the conformational state shown in Figure 2b. Note thedifferent disposition of the alkyl chains from the situation where the initial configuration for the simulation was considered to describe theDDAB monocrystal conformation (Figure 2a).

Figure 4. Ratios of the peak intensities, I1088/I1064 (empty symbols)and Is/Ias (full symbols) vs DDA+ ion content in the DDA+-MMT. Theseratios qualitatively describe gauche/trans conformers ratios in thespectral regions 600-1600 and 2750-3200, respectively.


content from that in DDAB monocrystals. Meanwhile, in theaforementioned computer simulation model, initial values of theangle between alkyl chains of DDA+ ions in DDA+-MMT wereassumed to be equal to those in DDAB monocrystals. Therefore,the simulated values of the proportions of trans conformers at0.50 CEC (Table 3) do not represent their values in theconformational state shown by Figure 2b. In fact, for themodifier content of 0.50 CEC we repeated the simulationassuming as initial configuration the conformational state shownin Figure 2b. The recalculated values (see values with thesuperscript symbol a) of trans conformers are compiled in Table3 also. As can been seen from Table 3, the deviation betweenthe experimental results and the theoretical estimation of theproportions of trans conformers vanishes. However, the simula-tion could not a priori be performed with reasonable guessesof initial angles between alkyl chains, characteristic of thismodifier content (Figure 2b conformation is not known inadvance); only further analysis of relevant Raman fingerprintscould help to select the conformational state Figure 2b as willagain be evidenced in the following section.

The Splitting of the Line at about 1130 cm-1 versus DDA+

Ion Conformation. The vibrational mode at approximately 1130cm-1 found in the Raman spectra of molecules with alkyl chains,corresponds to the symmetric stretching vibration of the C-Cbonds in trans segments and its position depends strongly onthe length of trans segments.20

We have noticed a very interesting and unexpected resultabout the behavior of the 1130 cm-1 band in the Raman spectraof the modified clay with varying DDA+ ion content. We haveobserved this band at 1133 cm-1 in the Raman spectrum ofDDA+-MMT with 0.50 CEC ion content. Its intensity decreasesmonotonically with increasing DDA+ ion content from 0.50 to1.50 CEC, which most probably describes a decrease of transconformations within this range or a corresponding increase ingauche conformers, and perfectly agrees with previous conclu-sions concerning the conformational rearrangement. However,this band progressively broadens, its position shifts from 1133to 1130 cm-1 with increasing DDA+ ion content up to 1.50CEC, and finally it splits into two lines at 1124 and 1135 cm-1

at 2.00 CEC DDA+ ion content. We also found the same lines(1124 and 1135 cm-1) in the solid DDAB Raman spectrum(Figure 3, Tables 2 and 3). In order to explain the splitting ofthe band at about 1130 cm-1, we used Raman spectra of solidCTAB, DODAB and n-alkanes. Figure 5 depicts Raman spectraof DDA+-MMT at a DDA+ ion content of 2.00 CEC, n-alkanes,

solid DDAB, CTAB, and DODAB in the spectral window1000-1400 cm-1. All spectra are normalized to the peakintensity of the band at 1300 cm-1. Both n-alkanes and CTABmolecules have single alkyl chains, while n-alkanes have a CH3

group on each end of the chain and in CTAB one terminal CH3

group is substituted by a (CH3)3-N- group. Additionally, alkylchains of liquid n-alkanes (n ) 5 - 17) contain both transconformations and numerous gauche conformers, while solidCTAB consists of alkyl chains in ideal trans conformation. Infact, as evidenced in Figure 5 a unique and strong marker bandat about 1064 cm-1 is found in the Raman spectrum of CTABand describes ideal trans conformations, while liquid n-alkanesshow additional broad bands around 1080 cm-1 accounting fornumerous gauche conformers. The DODAB molecule is similarto the DDAB molecule in the sense that it consists of two alkylchains and has a conformation as shown in Figure 2a in thecrystalline state.13,14

The dependence of the position of the symmetric C-Cstretching vibration (approximately 1130 cm-1) as a functionof the length of the trans segments of the alkyl chains forn-alkanes and solid alkylammonium surfactants having a singlealkyl chain is plotted in Figure 6. As can be seen from Figures5 and 6, the position of this vibrational mode strongly dependson the length of trans segments in the alkyl chains but also onthe overall chemical composition of the molecule containingsuch chains. In particular, in the case of identical length of thealkyl chains, the substitution of a CH3 group (n-alkanes, dotsin Figure 6) by a (CH3)3-N group (CTAB, triangle in Figure6) results in a down shift of the symmetric stretching vibrationof C-C bonds. The splitting of the band at approximately 1130cm-1 into two lines was observed only in the Raman spectra ofsolid DODAB and DDAB, both with the characteristic molec-ular conformation shown in Figure 2a. This conformationcontains two nonequivalent alkyl trans-chains parallel to eachother. One of these trans-chains consists of 18 C atoms or 14 Catoms for DODAB or DDAB, respectively, with a (CH3)2-N-terminal group. The second segment is a CH3-(CH2)n- chainwith 14 or 10 C atoms for DODAB and DDAB, respectively.Accordingly, by analyzing the Raman band positions of differentsolid alkylammonium surfactants and n-alkanes we can assignthe lines at 1124 and 1135 cm-1 in the Raman spectra of DDA+-MMT to the symmetric C-C stretching vibrations in the transchains consisting of (CH3)2-N-(CH2)13-CH3 and -(CH2)9-CH3 segments, respectively.

It is important to point out the similarity (strong intensity)of the bands at 1064 and 1133 cm-1 due to trans segments inthe Raman spectrum of DDA+-MMT at 0.50 CEC (Figure 3)

Figure 5. Raman spectra of solid alkylammonium surfactants (DOD-AB, DDAB, and CTAB), DDA+-MMT at 2.00 CEC of DDA+ contentand liquid n-alkanes (n ) 10, 14) in the region 1000-1400 cm-1.

Figure 6. Raman shift of the symmetric stretching of C-C bonds vsthe length of trans segments of alkyl chain for n-alkanes (dots) andsolid alkylammonium surfactants (triangle).


and in the Raman spectrum of CTAB which show an ideal transconformation of alkyl chains (Figure 5). Therefore, we proposethat the splitting of the 1130 cm-1 band would not be observedin the Raman spectra of DDA+-MMT if the alkyl chains in theDDA+ ions had both identical chemical composition (sameterminal group) and length of trans-chains, namely in the caseof two -(CH2)13-CH3 trans segments with a common terminal(CH3)2-N group (Figure 2b). In this case, however, only aunique angle of rotation should have been set between the twoalkyl trans segments, which was not the case in the numericalsimulation where initial angles were considered to describe thecrystal DDAB conformation (Figure 2a). Therefore as alreadymentioned earlier, the numerical simulation repeated for themodifier content of 0.50 CEC with the hypothesis (initialconditions) that the initial angles would correspond to thesituation highlighted by the conformational state Figure 2byielded a good agreement between the simulated and the Ramanestimation of the proportions of trans conformers (see Table3). As a matter of fact, in the interlayer galleries of DDA+-MMT, the alkyl chains adopt a conformation as in Figure 2b,at low DDA+ ion content (e.g., 0.50 CEC). With increasingDDA+ ion content they try to adopt a crystal-like conformationas in Figure 2a that is characteristic for both DODAB andDDAB monocrystals. In fact these two conformations (Figure2a,b) were previously identified by Venkataraman and Vasude-van,25 who investigated the behavior of chain conformation ofDODAB and dicetyl dimethyl ammonium bromide in cadmiumthiophosphate crystal that also exhibits interlayer structures suchas in clay. However, the persistence of the broad Raman bandaround 1088 cm-1 helps to conclude that even at the highestDDA+ ion content of 2.00 CEC, a mixed conformationalstructure most probably exists in DDA+-MMT, with some ofthe DDA+ ions having a conformation like the DDAB crystalshown in Figure 2a and the others still containing chains withnumerous gauche conformers as in Figure 2c. As it has beendemonstrated for the conformations in Figure 2, panels b anda, it is important to evaluate the possibility for the existence ofthe conformations shown in Figure 2, panels c and d. Theseconformations content about 50 and 64% trans conformations,respectively, and were selected from the results of the math-ematical simulation with initial conditions (interlayer spacing,angles between alkyl chains, relative amount of trans conform-ers) provided by our structure-sensitive experiments (X-raydiffraction and Raman scattering). Regarding the distributionof trans segments in the alkyl chain, the conformation in Figure2c has a much lower order than the conformation in Figure 2d.In fact as it can be noticed from the Raman spectra in Figure 3,when DDA+ content increases from, for example, 0.75 to 1.00CEC, the Raman band at 1064 cm-1 that corresponds to chainswith trans conformation (higher order) decreases progressively,corresponding to a gradual loss of order. However, the bandaround 1088 cm-1, representing the contribution of chains withnumerous gauche conformers (lower order), increases and showsevidence of the increase of disorder. In addition the values Rtrans

of the proportions of trans conformation calculated from ourRaman results for the modifier content of 0.75 and 1.00 CECare 60-62% and 58-64%, respectively, and are comparableto the simulated values for the conformations shown in Figure2c,d, respectively. Therefore although these two conformationalstates may be considered to be less defined (especially Figure2c with the increased disorder), we suggest that both conforma-tions are probable and would correspond to DDA+ content ofalmost 0.75 CEC and at about 1.00 CEC respectively.

As long as the intensity of the Raman band at 1130 cm-1 inthe spectra of DDA+-MMT for a DDA+ ion content of 1.50CEC tends to be comparable to the intensities of the lines at1124 and 1135 cm-1 found in the Raman spectra of solidDDAB, we can suggest that in DDA+-MMT with a DDA+ ioncontent of 1.50 CEC part of the DDA+ ions already haveadopted the crystal-like conformation (Figure 2a).

Comparing Raman scattering, X-ray diffraction, and DSCdata, we show that DDA+ ions with crystal-like conformationas in Figure 2a are formed in organic bilayers, but not inmonolayers of DDA+-MMT. The conformation shown in Figure2b appears in DDA+-MMT with low ion contents, that is, whenno traces of DDA+ ions melting were observed in DSC analysis(see Table 1). Accordingly, we consider this conformation tobe liquidlike.

Raman Estimation of the DDA+ Ion Content. We havefound that the integral intensity ratio I*1300/I*705 is proportionalto the mass ratio of DDA+ ions to MMT. This is an expectedresult, since the line at 705 cm-1 observed only in the Ramanspectra of the modified clay samples (see Figure 3) is assignedto the vibrational mode of the clay structural units, namely theν1 (A1) mode for SiO4stetrahedrons in clay silicate plates.17

Therefore, the ratio I*1300/I*705 in the Raman spectra of MMT-DDA+ can be reasonably used to determine the mass contentof DDA+ ions with respect to that of MMT. Figure 7 showsthe calibration curve comparing the predicted values of theDDA+/MMT mass ratio from Raman data (ratio I*1300/I*705) andthe values predefined by sample preparation (i.e., through thecation exchange chemical reaction during clay modification) andmeasured by an original method described earlier.2 The perfectagreement and linear fit for this calibration not only reflectsthe precision of the predetermined values of the DDA+ ioncontents, but it also validates the reliability of our Ramanmeasurements as can be noticed from the perfect agreementreached above on the behavior of different spectral features.

The Spectral Region 2750-3200 cm-1. Figure 8 shows theRaman spectra of DDA+-MMT modified by different masscontent of DDA+ ions in the region (2750 - 3200 cm-1). Allspectra in this region are normalized to the peak intensity ofthe CH2 asymmetric stretching vibrational band (νas) at ap-proximately 2885 cm-1. Table 2 shows the peak positions ofthis band and those of the CH2 symmetric stretching vibrationalmode (νs) at approximately 2853 cm-1. The dependence of theintensity ratio Is/Ias of these two bands on the mass content ofDDA+ ions is shown in Figure 4.

As reported in previous works on n-alkanes andpolyethylene,21-23 the ratio Is/Ias is expected to increase withboth increasing number of gauche conformers as well as with

Figure 7. Integral intensity ratio I*1300/I*705 vs mass ratio of DDA+

ions and clay (MMT).


packing rearrangement (disordering or reordering). The transi-tion from ideal trans to ideal gauche conformation of CH2 chainsis accompanied by a shift of the νs (CH2) vibration from 2848to 2876 cm-1 and that of the νas (CH2) vibration from 2881 to2925 cm-1. As it can be seen from Figures 4 and 8 and Table2, the band position of the νas (CH2) vibration and the ratioIs/Ias depend on the DDA+ ion content. They increase withincreasing DDA+ ion content as a result of conformationalrearrangement. However, they decrease at a DDA+ ion contentof 2.00 CEC. Therefore, in agreement with previous reports21-23

the number of CH2-chains with gauche conformers increaseswith increasing DDA+ ion content from 0.50 to 1.50 CEC anddecreases from 1.50 to 2.00 CEC. This perfectly agrees withthe results obtained from the spectral region 600-1600 cm-1.However, overlapping of Raman bands in the region 2750-3200 cm-1 impedes quantitative analysis of DDA+-MMTinterlayer structure as it was well done in the 600-1600 cm-1

fingerprint region.

Conclusion

The conformational order of alkyl chains of ditetradecyldimethyl ammonium bromide embedded in the interlayer spaceof sodium montmorillonite clay has been studied. We haveshown that Raman spectra of clays modified by this dialkylam-monium surfactant, recorded in the spectral region 600-1600cm-1, are highly informative for both qualitative and quantitativeanalysis of the clay interlayer structure. Many spectral featuresdetected in this region were interpreted, and they were foundto be in a perfect agreement with each other and also to correlatewith DSC analysis.

Namely, with increasing content of the modifier up to thecritical value of 1.50 CEC, both the asymmetry of the 1300cm-1 Raman band and the peak intensity ratio of the bands at1088 and 1064 cm–1 increase. These findings directly evidencethe increase in the gauche/trans conformers ratio. However,further increase in the modifier content reverses this behavior,pointing to a drastic conformational transition. We have foundout that the total content of trans conformers can be evaluatedfrom the integral intensities ratio I*1064/I*1300 and perfectlycorrelates with the behavior of the 1300 cm-1 Raman band andthe intensity ratio I1088/I1064. Moreover, a careful interpretationof the spectral characteristics such as the peak position andintensity, and more importantly, the band splitting of the Ramanmode at approximately 1130 cm-1 permits the recognition of

two critical conformations of DDA+ ions in DDA+-MMT. Infact, a conformational transition was observed from a liquidlikeconformation consisting of two identical -(CH2)13-CH3 transsegments with a common terminal (CH3)2-N- group at lowmodifier content (0.50 CEC) to a crystal-like conformationconsisting of two nonequivalent trans segments >N-(CH2)13-CH3 and -(CH2)9-CH3, corresponding to the splitting of the1130 cm-1 band into two lines at 1124 and 1135 cm-1,respectively, at high modifier contents (g1.50 CEC). However,conjointly analyzing two experimental details, the 1130 cm-1

band behavior and the 1088 cm-1 band existing even at thehighest modifier content, we have concluded on a conforma-tional transition from a quasi pure liquidlike conformation atthe lowest modifier content, to mixed conformational structures,where part of the DDA+ ions adopt either a liquidlike or acrystal-like (modifier content g1.50 CEC) conformation, butthe other part of the DDA+ ions contain alkyl chains withnumerous gauche conformers. Moreover, the integral intensityratio I*1300/I*705 was found to be essential for the prediction ofthe relative mass content of alkylammonium ions embedded inthe interlayer space of DDA+-MMT. Here, the predicted DDA+/MMT mass ratios yielded a perfect and linear calibration fitwith the values of these ratios measured by a different methodduring sample preparation. This observation not only reflectsthe high precision of the predetermined values of the DDA+

ion content, but it also validates the reliability of the Ramanresults, as also reflected by the perfect agreement reached aboveusing the behavior of different spectral features. This is furtherevidenced by the fact that computer simulations of the DDA+

conformational states in DDA+-MMT yielded reliable modelsand showed an improved efficiency only upon the applicationof our Raman data (proportions of trans conformations atdifferent modifier content) as initial conditions of the math-ematical models.

These results demonstrate the capability of Raman spectros-copy for a rapid and nondestructive monitoring of the interlayerstructure of sodium montmorillonite clay modified by ditet-radecyl dimethyl ammonium bromide. It has the potential tobe applied for structural characterization of nanocompositematerials modified by alkylammonium surfactants.

Acknowledgment. This work is partially supported by theGrant of the President of the Russian Federation for LeadingScientific Schools (476.2008.2). E.A.S. acknowledges supportby INTAS (Project No. 06-1000014-6055). We are grateful toDr. Phedor N. Bakhov for his support in the preparation of themodified clay samples and Dr. Dmitry N. Kozlov for fruitfuldiscussions of our results.

References and Notes

(1) Alexandre, M.; Dubois, P. Mater. Sci. Eng. 2000, 28, 1–63.(2) Gerasin, V. A.; Bakhov, F. N.; Merekalova, N. D.; Korolev, Y. M.;

Fischer, H. R.; Antipov, E. M. Polym. Sci., Ser. A. 2005, 47, 954–967.(3) Schleidt, S.; Spiess, H. W.; Jeschke, G. Colloid Polym. Sci. 2006,

284, 1211–1219.(4) Boutfatit, M.; Ait-Amar, H.; McWhinnie, W. R. Desalination 2007,

206, 394–406.(5) Chen, X.; Hu, N.; Zeng, Y.; Rusling, J. F.; Yang, J. Langmuir 1999,

15, 7022–7030.(6) Vaia, R. A.; Teukolsky, R. K.; Giannelis, E. Chem. Mater. 1994,

6, 1017–1022.(7) Prokhorov, K. A.; Saguitova, E. A.; Nikolaeva, G. Y.; Kozlov,

D. N.; Pashinin, P. P.; Antipov, E. M.; Gerasin, V. A.; Bakhov, F. N.;Guseva, M. A. Laser Phys. Lett. 2005, 2, 285–291.

(8) He, H.; Frost, R. L.; Xi, Y.; Zhu, J. J. Raman Spectrosc. 2004, 35,316–323.

Figure 8. Raman spectra of solid DDAB and of DDA+-MMT withDDA+ ion content from 0.50 to 2.00 CEC in the region 2750-3200cm-1.


(9) Antipov, E. M.; Guseva, M. A.; Gerasin, V. A.; Korolev, Y. M.;Rebrov, A. V.; Fischer, H. R.; Razumovskay, I. V. Polym. Sci., Ser. A 2003,45, 1130–1139.

(10) Osman, M. A.; Seyfang, G.; Suter, U. W. J. Phys. Chem. B 2000,104, 4433–4439.

(11) Dias, P. M.; De Faria, D. L. A.; Constantino, V. R. L. J. InclusionPhenom. Macrocyclic Chem. 2000, 38, 251–266.

(12) Fermeglia, M.; Ferrone, M.; Pricl, M. Fluid Phase Equilib. 2003,212, 315–329.

(13) Okuayma, K.; Iijima, N.; Hirabayashi, K.; Kunitake, T.; Kusunoki,M. Bull. Chem. Soc. Jpn. 1988, 61, 2337–2341.

(14) Okuayma, K.; Soboi, Y.; Iijima, N.; Hirabayashi, K.; Kunitake,T.; Kajiyama, T. Bull. Chem. Soc. Jpn. 1988, 61, 1485–1490.

(15) Kalyanasundaram, K.; Thomas, J. K. J. Phys. Chem. 1976, 80,1462–1473.

(16) Amorin da Costa, A. M.; Geraldes, C. F. G. C.; Teixeira-Dias,J. J. C. J. Raman Spectrosc. 1982, 13, 56–62.

(17) Frost, R. L.; Rintoul, L. Appl. Clay Sci. 1996, 11, 171–183.

(18) Strobl, G. R.; Hagedorn, W. J. Polym. Sci., Polym. Phys. Ed. 1978,16, 1181–1193.

(19) Gall, M. J.; Hendra, P. J.; Peacock, C. J.; Cudby, M. E. A.; Wills,H. A. Spectrochim. Acta, Ser. A 1972, 28, 1485–1496.

(20) Gorelik, V. S.; Zlobina, L. I.; Sharts, O. N. Proc. SPIE 2000, 4203,66–77.

(21) Abbate, S.; Zerbi, G.; Wunder, S. L. J. Phys. Chem. 1982, 86, 3140–3149.

(22) Snyder, R. G.; Strauss, H. L.; Eliger, C. A. J. Phys. Chem. 1982,86, 5145–5150.

(23) Snyder, R. G.; Scherer, J. R.; Garber, B. P. Biochim. Biophys. Acta1980, 606, 47–53.

(24) Snyder, R. G.; Cameron, D. G.; Casal, H. L.; Compton, D. A. C.;Mantsch, H. H. Biochim. Biophys. Acta 1982, 684, 111–116.

(25) Venkataraman, N. V.; Vasudevan, S. J. Phys. Chem. B 2002, 106,7766–7773.

JP810050H


raman spectroscopic characterization of the interlayer structure of na + ...

Documents