Printer Friendly

Addressing the interface in polymer-clay nanocomposites by electron paramagnetic resonance spectroscopy on surfactant probes.


Nanodispersed particles of layered silicates (clays) can improve the mechanical properties, gas barrier behavior, thermal resistance, and flame retardance of polymers, even if they are added in amounts only as small as 2% (1-7). Despite substantial efforts that have produced many encouraging results, the reasons for these improvements are not fully understood. Furthermore, depending on the polymer, such nanocomposites can be produced by polymerization in the presence of the clay, by intercalation from melt, or by intercalation from solution (7, 8). Yet, it cannot be reliably predicted which of these approaches will work for a given polymer or, if several do work, which of them will lead to the best material properties. The situation is particularly complicated for hydrophobic polymers that require organically modified clays (organoclays) for nanocomposite formation. Such organoclays are synthesized by ion exchange of natural or synthetic clay minerals with cationic surfactants, so that the choice of surfactant is introduced as another variable parameter that is understood on a theoretical level only in rather broad terms and can usually be optimized only empirically (6). The interface layer of surfactant molecules mediating between the inorganic clay particles and the polymer governs the entropy and enthalpy of intercalation of the polymer between the silicate layers and mechanical coupling between the particles and polymer chains, which is responsible for the improved material properties. Insight into the relation between structure and dynamics of this layer on the one hand and properties and function of the nanocomposites on the other hand is thus of great interest (5).

Most work in this field has concentrated on techniques for characterizing bulk properties and on scattering techniques that depend on a significant degree of long range order (2-7). Local methods, such as magnetic resonance experiments, could potentially complement the information from the other techniques and could thus help to develop a microscopic picture of the interface layer. Electron paramagnetic resonance (EPR) spectroscopy with its high selectivity and sensitivity (9, 10) appears to be well suited for this purpose, but, to the best of our knowledge, it has not yet been applied to polymer-clay nanocomposites. In particular, spin probe techniques allow us to address specific sites in complex systems and can provide information on dynamics at time scales between approximately 10 ns and 1 [micro]s (11-13) and on structure at length scales between 0.5 and 10 nm (14, 15). Spin probes mimicking anionic or cationic surfactants have been applied recently to obtain rather detailed information on surfactant layers and surfactant aggregates in the context of polymers (16, 17) and of template-based synthesis of inorganic mesoporous materials (18).

In the present work we explore what information can be obtained by applying both conventional continuouswave (CW) EPR (9) and modern pulse EPR (10) techniques to such surfactant spin probes localized in the surfactant layer on organoclay particles and in polymer-clay nanocomposites. As model systems for this initial study we employ synthetic clays organically modified by a cationic surfactant and their composites with polystyrene with a narrow molecular weight distribution and varying molecular weight. We choose these systems since very similar systems have been carefully investigated in the past by other techniques (19), variability in the composition can be kept at a minimum, and intercalation of the polymer is also possible from solution, apparently leading to structures that later reorganize in the melt (20).

The paper is organized as follows. First, we describe how EPR data of spin probes are related to local dynamics and structure in the vicinity of these probes and to the spatial distribution of the probes. After the Materials and Methods section we discuss dynamics of the surfactant anchor groups in organoclays and changes in this dynamics caused by intercalation of polystyrene. For an organoclay sample we compare dynamics in the anchor group region and tail region of the surfactant. We further examine whether local elemental analysis can prove proximity of the probes to the clay particle surface and to neighbor molecules and finally consider how the spatial distribution of surfactant anchor groups is changed by going from an organoclay to an intercalated polymer-clay nanocomposite.


Reorientation of a spin probe on the time scale of the EPR measurement (~10 ps-1 [micro]s) causes changes in the spectral lineshape. as the resonance frequencies depend on orientation of the molecule in the magnetic field (Fig. 1). At X-band frequencies (9-10 GHz) the lineshape is dominated by the hyperfine coupling of the electron to the [.sup.14]N nucleus of the nitroxide group with nuclear spin I=1. For a molecule at a certain orientation this hyperfine coupling causes a splitting of the EPR line into three lines corresponding to magnetic quantum numbers [m.sub.1] = -1, 0, and 1 of the [.sup.14]N nucleus. The magnitude A of this splitting depends on orientation, as the unpaired electron is localized in a 2[p.sub.z][pi]-orbital whose lobes define the z-direction of the molecular frame (Fig. 1a). Along this z-direction, the hyperfine splitting A is maximum ([A.sub.zz] = 3.4 mT), in the plane perpendicular to it, it is minimum ([A.sub.xx] [approximately equal to] [A.sub.yy] [approximately equal to] 0.6 mT).

The simplest method for characterizing motion of the spin label is the measurement of the maximum splitting 2[A'.sub.zz] in the EPR spectrum, which corresponds to the difference of the resonance fields between the low-field maximum and the high-field minimum of the spectrum (Fig. 1c-f). This splitting depends on the rotational correlation time [[tau].sub.c] for the rotational diffusion of the probe. For simplicity we assume for the moment that rotational diffusion itself is isotropic, i.e., has the same rate about all axes of the molecular frame. In the rigid limit ([[tau].sub.c] [much greater than] 1 [micro]s), where reorientation of the spin probe during the measurement is negligible, 2[A'.sub.zz] is equal to 2[A.sub.zz] = 6.8 mT. In the fast limit ([[tau].sub.c] [much less than] 10 ps), where the correlation time is short with respect to the inverse resonance frequency, 2[A'.sub.zz] approaches the isotropic hyperfine coupling [a.sub.iso] = ([A.sub.xx] + [A.sub.yy] + [A.sub.zz])/3 = 3.1 mT. For rotational correlation times [[tau].sub.c] between 10 ps and 1 [micro]s, the spectral lineshape and 2[A'.sub.zz] depend on [[tau].sub.c] (Fig. 1b, d, e). The dependence is strongest for [[tau].sub.c][DELTA]v [approximately equal to] 1, where the total spectral width [DELTA]v is given by [DELTA]v = 2.8 * [10.sup.7] [s.sup.-1] m[T.sup.-1] * 2[A.sub.zz] = 1.9 * [10.sup.8] [s.sup.-1], corresponding to [[tau].sub.c] [approximately equal to] 5 ns. The inflection point of the dependence corresponds to 2[A'.sub.zz] [approximately equal to] 5 mT.

In soft matter such as polymers the rotational correlation time depends on temperature. The temperature at which 2[A'.sub.zz] = 5 mT is attained is termed EPR glass transition temperature, [T.sub.5mT], or [T.sub.50G]. Note that [T.sub.5mT] is a characteristic that depends not only on the free volume and dynamics of the environment but also on the size of the probe molecule. EPR spectra can thus provide estimates for the rotational correlation time of probe molecules. The method is also very sensitive to differences in the dynamics of a probe molecule in a heterogeneous material (dynamical contrast) (16, 17) or can be used in a semi-quantitative way by discussing changes in 2[A'.sub.zz] values or spectral lineshapes that occur as a result of changes in material composition. Finally, in the slow tumbling regime around [[tau].sub.c][DELTA]v [approximately equal to] 1, more sophisticated data analysis by simulating and fitting spectral lineshapes (13) can provide additional details on the reorientation process, such as information on preferred axes of rotation for nonspherical molecules or for molecules electrostatically attached to an interface, and information on microscopic order such as in surfactant layers (16).


Hyperfine coupling is not restricted to nuclei of the probe molecule that carries the electron spin. It can also occur because of dipolar interaction of the electron spin through space with nuclei in neighbor molecules and because of transfer of spin density through orbital overlap to a neighbor molecule. Such intermolecular hyperfine couplings can be used for elemental analysis in the vicinity of the probe molecule, as they are fairly short-range: the dipolar coupling decays with the inverse cube of the distance between electron and nuclear spin and the transfer of spin density decays roughly exponentially with distance. To identify the type of nucleus, the nuclear Zeeman frequency must be measured, which can be achieved by electron nuclear double resonance (ENDOR) spectroscopy or by electron spin echo envelope modulation spectroscopy (ESEEM) (10). In our case, where the spin probe molecule contains protons ([.sup.1]H), it is impossible to safely assign proton signals to neighbor molecules. To prove the vicinity of solvent molecules or of a polymer chain to the nitroxide group of the spin probe, we thus have to resort to deuterated solvents or polymers.


On the other hand, observation of signals due to nuclei of a type that is unique to a certain component of a heterogeneous system proves direct contact between this component and the nitroxide group. For example, consider [.sup.19]F nuclei in expandable fluoromica. For the strongly electronegative fluorine, already very small spin densities result in significant hyperfine couplings, so that fluorine nuclei in the neighborhood of a spin probe can be readily detected (21). On the other hand, we have demonstrated that segregation of surfactants from a fluorinated polymer results in the absence of [.sup.19]F ENDOR signals (16). Hence, proximity of the spin-carrying moiety of a probe molecule to the surface of a clay particle should be detectable by ENDOR spectroscopy. In particular, high-field ENDOR spectroscopy at W-band frequencies of ~94 GHz has been shown to simplify application of this technique considerably, as [.sup.1]H and [.sup.19]F ENDOR signals are separated in standard ENDOR experiments, while at the conventional X-band frequency, they have to be separated by more involved two-dimensional experiments (22).


Dipolar couplings between the electron spins in two probe molecules are proportional to the inverse cube of the distance between these molecules. For distances larger than 2 nm, as they are relevant for polymer-clay nanocomposites, the dipolar couplings are small compared to EPR linewidths and thus cannot directly be extracted from the lineshape. Pulse electron electron double resonance (ELDOR) techniques can be used to separate the dipolar couplings between electron spins from all the other interactions (23, 24). In particular, data from the four-pulse double electron electron resonance (DEER) experiment (25) contain the full information on the spin-to-spin pair correlation function (15, 26).

In contrast to most scattering methods, such EPR techniques provide the selectivity and sensitivity to characterize distance distributions between sites of interest, even if these sites correspond to only a small fraction of the whole material. In the case at hand, we characterize the spatial distribution of surfactant anchor groups and changes in that distribution on intercalation by the polymer. This provides information that can be compared to results on the interlayer distance obtained from wide-angle X-ray scattering (WAXS). The EPR results are complementary to the WAXS information, as the distribution of surfactant anchor groups can be analyzed even in the limit of complete exfoliation where the clay platelets are no longer parallel to each other and the WAXS peak corresponding to the interlayer distance vanishes.


Organoclays were prepared from Laponite RD and Laponite RDS (Rockwood Additives, UK) and from Somasif ME 100 (Co-Op Chemical Co., Japan) by the following general procedure: 2 g of the synthetic clay were dispersed in 200 mL water (Milli-Q) and stirred for 30 min. A 100:1 mixture (molar ratio) of hexadecyltrimethylammonium chloride (CTAC, Aldrich) and 4-(N,N-dimethyl-N-hexadecyl)ammonium-2,2,6,6-tetramethylpiperidine-1-oxyl-iodide (Cat-16, Molecular Probes, Inc., see Fig. 2a) was added as a solution in 50:50 mixture of deionized water and ethanol (Laponites: 1.76 g CTAC, 40 mL water and 40 mL ethanol; Somasif: 1.76 g CTAC, 40 mL water and 40 mL ethanol). This corresponds to a fivefold excess of surfactant with respect to the cation exchange capacity (CEC) for Laponites (CEC: 0.55 meq/g) and a threefold excess for Somasif (CEC: 0.85 meq/g). After stirring for 6 h the precipitate was separated by filtration and washed with a 1:1 mixture of hot water and ethanol until an AgN[O.sub.3] test for chloride ions in the washing liquid was negative. The material was then dried at 70[degrees]C for 24 h in vacuum. Preliminary gravimetric experiments suggest that this procedure yields organoclays with a 1.5 to 2-fold excess of surfactant molecules with respect to CEC.


Samples with the specifically designed tail-labeled surfactant 1 (Fig. 2b) were prepared analogously. Synthesis and characterization of this compound will be described elsewhere. Samples for EPR measurements were generally prepared by pressing 100 mg of organoclay or of the organoclay/polystyrene (PS) mixtures at a temperature of 433 K for 30 min at a pressure of 70 MPa in a Weber-Press (Maschinen Apparatebau GMBH). Mixtures were prepared from 75 mg PS and 25 mg organoclay, for PS we used GPC standards (Fluka) with a narrow molecular weight distribution and the following molecular weights: PS 10,000 ([M.sub.w] ~ 9500, [M.sub.w]/[M.sub.n] ~ 1.06), PS 30,000 ([M.sub.w] ~ 32,000, [M.sub.w]/[M.sub.n] ~ 1.03), and PS 100,000 ([M.sub.w] ~ 94,900, [M.sub.w]/[M.sub.n] ~ 1.06).

All EPR measurements were performed with a Bruker Elexsys 580 EPR spectrometer at X-band frequencies (9.3-9.8 GHz), except for the high-field ENDOR measurements, which were performed with a Bruker Elexsys 680 spectrometer. For variable temperature CW EPR we used a 4103 TM resonator with a glass Dewar and the Bruker ER 4111 VT temperature control unit. A microwave power of 2 mW, a modulation amplitude of 0.1 mT, and a modulation frequency of 100 kHz were employed. Where necessary, [Fe.sup.3+] background signals were eliminated by subtraction of the spectrum of organoclay without spin probes. For ESEEM and DEER experiments we used a 3-mm split-ring resonator (ER 4118X_MS3). The temperature was set to 80 K by cooling with liquid nitrogen using an Oxford cryostat and cooling system. At this temperature, background signals due to [Fe.sup.3+] impurities in Somasif are negligible in pulse EPR experiments owing to their short relaxation time. Longer relaxation times and an interference with the signal of the spin probes is observed for [Fe.sup.3+] impurities in Laponites. For ESEEM experiments, we applied the stimulated echo sequence ([pi]/2)-[tau]-([pi]/2)-T-([pi]/2)-[tau]-echo with the usual four-step phase cycling (10) to eliminate unwanted echoes. Time traces with 1024 data points and an evolution time increment [DELTA]T = 16 ns were recorded with pulse lengths of 16 ns for the [pi]/2 pulses. Four-pulse DEER experiments were performed with the sequence ([pi]/2)[.sub.vA]-[[tau].sub.1]-([pi])[.sub.vA]-t'-([pi])[.sub.vB]-([[tau].sub.1] + [[tau].sub.2]-t)-([pi])[.sub.vA]-[[tau].sub.2]-echo, (+x, -x) phase cycling of the first pulse, and averaging over 25 increments of [[tau].sub.1] ([DELTA][[tau].sub.1] = 8 ns) to suppress proton modulations (15, 25). Here, [v.sub.A] is the observer frequency (local maximum at the low-field edge of the EPR absorption spectrum) and [v.sub.B] the pump frequency (global maximum of the EPR absorption spectrum, [v.sub.B] = [v.sub.A] - 65 MHz). The time t' after the first [pi]-pulse is incremented, the pulse lengths and delays are as follows: ([pi]/2)[.sub.vA] = 16 ns, ([pi])[.sub.vA] = 32 ns, ([pi])[.sub.vB] = 12 ns, [[tau].sub.1] = 200 ns, [[tau].sub.2] = 1.6 [micro]s, [DELTA]t' = 8 ns, [t'.sub.0] = 80 ns. High-field ENDOR spectra were measured at W-band frequencies (~94.0-94.2 GHz) and at a temperature of 80 K using a Bruker ENDOR resonator using the Mims ENDOR sequence ([pi]/2)-[tau]-([pi]/2)-t-([[pi].sub.rf])-[t.sub.mix]-([pi]/2)-[tau]-echo. Pulse lengths were 140 ns for the microwave pulses and 10 [micro]s for the radiofrequency pulse, with [tau] = 360 ns, t = 260 ns and [t.sub.mix] = 4.6 [micro]s. The radiofrequency power was 50 and 200 W with waiting times of 2 and 4 ms between repetition of the experiment, respectively.

DEER data were analyzed with the home-written software DEERfit, which is freely available at


CW EPR spectra of surfactant spin probes in organoclays derived from Somasif ME 100 were measured for the pure organoclay and for mixtures of 25 wt% organoclay and 75 wt% PS (molecular weights 10,000, 30,000, and 100,000) in a temperature range from 293 to 433 K at temperature intervals of 10 K. Typical data for Cat-16 modified clay are shown in Fig. 3a. At room temperature, spectra in the absence and presence of polymer are hardly distinguishable and correspond to the rigid limit, i.e., the rotational correlation time of the nitroxide group is longer than 1 [micro]s (data not shown). As the nitroxide moiety is located in the anchor group region (see Fig. 2), this indicates that the surfactant molecules are immobilized at the surface of the clay particles at this temperature. Already at a temperature of 313 K, i.e., 60 K below the glass transition temperature of PS, differences in the spectra are readily apparent. In particular, the organoclay without polymer exhibits a small fraction of surfactant anchor groups with rotational correlation times shorter than a nanosecond (arrow). This fraction increases with increasing temperature, but it remains much smaller than the more immobile fraction, and thus cannot be assigned to excess surfactant. Note that the number of spins is proportional to the double integral of the spectrum: hence, at comparable amplitude narrow-line spectra of mobile probes correspond to a much lower number of spins than the broad-line spectra of more immobile probes. At 313 K, spectra for the more immobile fraction in the sample without polymer and in the sample with PS of molecular weight 100,000 no longer correspond to the rigid limit, as is apparent by broadening of the outer peaks and a decrease of 2[A'.sub.zz] (see dotted lines in Fig. 3a).

Surprisingly, at 313 K this fraction appears to become more mobile by intercalation with PS 100,000, but more immobile by intercalation with PS 10,000. However, this trend is reversed at higher temperatures. At the glass transition temperature of PS (373 K) and above, polymer intercalation consistently leads to decreased mobility of the major surfactant fraction. A stronger effect and a better defined mobility of this fraction is found for PS 10,000, indicating that under the conditions used in sample preparation, complete intercalation occurs for this molecular weight but not for the ten-times-larger molecular weight of PS 100,000. This interpretation is in agreement with WAXS data (not shown), in which a peak corresponding to the interlayer distance in the pure organoclay persists for the sample intercalated by PS 100,000, although a peak corresponding to longer distances also appears. In contrast, for PS 10,000 the peak for the clay interlayer distance in the organoclay almost completely vanishes. In the pure organoclay, mobility of the surfactant anchor groups at 433 K roughly corresponds to the EPR glass transition temperature (2[A'.sub.zz] [approximately equal to] 5.3 mT). i.e., to a rotational correlation time of 5 ns. Intercalation by PS 10,000 leads to an increase of [[tau].sub.c] to approximately 100 ns. For these two samples, the 2[A'.sub.zz] values of the more immobile component exhibit a clear decrease with temperature in the range between 413 and 463 K (data not shown). Substantial broadening of the high-field minimum for the composite with PS 100,000 leads to large uncertainties of 2[A'.sub.zz] values for PS 100,000 and no clear trend can be established. Nevertheless it is safe to say that the mobility of the surfactant anchor groups in this sample is intermediate between the mobilities in the two other samples. The rotational correlation time of the fast component is significantly shorter than 1 ns and generally decreases with increasing temperature. For this component, data for the composites with PS 10,000 and PS 100,000 virtually coincide, and up to temperatures of 423 K, mobility is higher in the pure organoclay than in the composites. Together with the fact that the more mobile component is not observed in composites below the glass transition temperature of polystyrene, this suggests that it stems from surfactant molecules that are in close contact with the polymer independent of the degree of intercalation. Assignment of this fraction to particle edges would be consistent with this finding.


In contrast to the anchor group-labeled surfactant probe Cat-16, the specifically designed tail-labeled surfactant probe 1 exhibits some mobility on the EPR time scale at room temperature (data not shown). At a temperature of 313 K we observe two fractions, one of which has a rotational correlation time shorter than 5 ns (Fig. 3b, arrows in left panel). The fractions of highly mobile and less mobile fraction surfactant tails are comparable. With increasing temperature (Fig. 3b, center and right panel) the highly mobile fraction increases and the less mobile fraction decreases. Furthermore, the rotational correlation time for the highly mobile fraction decreases with increasing temperature, as is apparent in a further narrowing of the lines of this fraction (see arrows in Fig. 3b). In the spectrum at 433 K, the less mobile fraction is hardly detectable. Attempts to simulate the spectrum of the highly mobile fraction by a model of isotropic rotational diffusion fail. In particular, the high-field line (inset in right panel of Fig. 3b) cannot be reproduced by assuming this kind of dynamics. The double-maximum structure of this line indicates restricted motion of the tail end of the surfactant. Given the fact that at the same temperature, dynamics of the anchor group region is much slower, this is hardly surprising. A study of the influence of polystyrene intercalation on tail end dynamics is beyond the scope of the present paper, but is planned for the near future.

In organoclays prepared from the synthetic layered silicate Laponite with a lateral particle size of only 25 nm compared to 1.5 [micro]m for Somasif, we find a much larger fraction of mobile Cat-16 surfactants (Fig. 3c). This indicates that the mobile fraction is related to the region close to the edges of the particle. Phosphonate modification of the particle edges, corresponding to the change from Laponite RD to RDS, indeed significantly changes the shape of the high-field maximum for the mobile fraction (see arrows in detail plots of Fig. 3c). The appearance of a fraction with even higher mobility than in Laponite RD (leftmost narrow feature in the central panel of Fig. 3c) might be related to edgeon attachment of surfactants to the negative pyrophosphate groups. At the same time, a fraction with intermediate mobility roughly corresponding to the mobile fraction in Laponite RD persists. Spectral changes on mixing the organoclay with polystyrene and pressing and tempering the mixture are hardly significant at a temperature of 413 K (compare central and left panel of Fig. 3c). A more detailed analysis of the temperature dependence indicates changes in the activation energy for mobilization of the most mobile component (data not shown), which in turn suggests that the Laponite particles are indeed dispersed in the polymer. Because of the more complex and harder to interpret EPR line-shape for Laponites compared to Somasif and the higher technological relevance of clay particles with a larger aspect ratio, we focus on Somasif for testing the applicability of more sophisticated pulse EPR methods.


High-field ENDOR spectra of Cat-16 in Somasif organoclay were recorded in a frequency range from 133 to 148 MHz in covering both the [.sup.19]F and the [.sup.1]H Zeeman frequencies. The spectra are clearly dominated by the [.sup.1]H line at approximately 143 MHz, but weak signals are also observed at the [.sup.19]F frequency of 134.2 to 134.5 MHz (Fig. 4). Although the [.sup.19]F signal only slightly exceeds the noise level, we are reasonably certain about the assignment, since the signal is reproducible and exhibits the expected frequency shift when changing the static field from 3.3571 to 3.3493 T. In fact, the small intensity of the signal is not unexpected, as the fluorine atoms are built into the silica framework and not exposed to the surface of the clay particles (see also Fig. 2b). Therefore, very small hyperfine couplings and inefficient transfer of spin polarization from the electron to the nuclear spin during the ENDOR experiment are anticipated.

In itself, the proximity of the surfactant anchor groups to the clay surface is hardly surprising. The experimental result is of some interest nevertheless, as it proves that the nitroxide group really probes the interface between surfactant and clay particles, and as it shows that the high-field ENDOR experiment is in principle capable of detecting such proximity. The latter finding opens up the possibility to test in future work, whether nitroxide groups in the tail region of the surfactant may be located close to the clay surface, too.


ESEEM results for shock-frozen glassy dispersions of Cat-16 in Somasif organoclay dispersed in toluene and toluene-[d.sub.8] are compared in Fig. 5. For protonated toluene (Fig. 5a, b) the spectrum is dominated by a proton peak but also exhibits a peak at the [.sup.14]N nuclear Zeeman frequency, which very probably originates from the nitrogen nucleus in the quaternary ammonium group of the surfactant probe, but might also be due to quaternary ammonium groups in neighboring CTAC molecules. The unusually narrow [.sup.14]N peak centered at the nuclear Zeeman frequency is explained by the fact that quaternary nitrogen has an almost tetrahedral point symmetry of its first coordination sphere and thus a small nuclear quadrupole coupling of about 100 kHz.


The time-domain data (Fig. 5c) and spectrum (Fig. 5d) change drastically for the dispersion in toluene-[d.sub.8]. The spectrum is now clearly dominated by the deuterium peak at approximately 2 MHz. This is clear evidence that toluene penetrates into the surfactant layer down to the anchor group region (see also Fig. 2b). Note that the surfactant layer may not be as well ordered and as dense as may be suggested by Fig. 2b. In particular, X-ray diffraction data indicate that even for rather large surfactant loadings, the long axis of the hexadecyl tails of N,N-dimethyl-N-hexadecyl ammonium ions may be closer to a parallel direction than to a perpendicular direction with respect to the surface of the clay particle (27). If that is the case, the long alkyl tails will probably change their conformation on dispersing the organoclay in organic solvents and a loose layer of surfactant molecules interdispersed with solvent may ensue. On adding polymer and evaporating the solvent, such particles might then conceivably form different interface structures than ensue from melt intercalation of polystyrene. This hypothesis can be tested in future work by comparing ESEEM data of nanocomposites of organoclay with deuterated polystyrene prepared from solution and from the melt. Preliminary analysis of the modulation depth and the decay of the modulation for toluene-[d.sub.8] by ratio analysis (28) indicates that the deuterium nuclei can approach as close as 5.5 [Angstrom] to the electron spin and that, on average, each electron spin has ten to eleven deuterium nuclei in such a close proximity. Although the coordination shell model underlying this analysis may oversimplify the situation for such a disordered system, it is safe to conclude the solvent molecules indeed penetrate into the surfactant layer down to the region of the anchor groups.



DEER time-domain data for Cat-16 in pure pressed and tempered Somasif organoclay and a pressed and tempered mixture of the same organoclay with PS 10,000 differ significantly from each other (Fig. 6). First, it is apparent that within 1.5 [micro]s the echo signal decays to 33% of its maximum value for the pure organoclay, but only to 49% of its maximum value for the 1:3 clay/polymer mixture. Such a slowdown of the decay corresponds to an increase of the mean distance between surfactant anchor groups, as expected for intercalation of the clay layers by polymer chains. Indeed, WAXS data, which will be published later, also show an increase in the interlayer distance for this sample.


Second, the shape of the decay curve clearly changes. For the pure organoclay, the curve exhibits a cusp at t = 0, corresponding to a larger fraction of spin pairs at shorter distances than would be expected for a homogeneous distribution of the surfactant anchor groups (best fit for this model: dotted line in Fig. 6a) in three dimensions. Actually, such a cusp can be expected for a two-dimensional distribution of spin pairs on a surface or for a layer structure. In the actual distance distribution (inset in Fig. 2a, best fit: solid line), distances up to approximately 4 nm are overrepresented with respect to a homogeneous distribution. In contrast, data for the nanocomposite can be nicely fitted by a homogeneous distribution in three dimensions (solid line in Fig. 6b). This indicates that the cusp is mainly due to pairs of surfactant anchor groups on neighboring clay platelets that are separated by intercalation. Alternatively, the cusp could be due to patches of surfactant on an inhomogeneously covered surface of the platelet with intercalation causing a more homogeneous distribution of the surfactant. Deciding between these two possibilities and quantifying the model in terms of changes in the surfactant distribution during intercalation and exfoliation will require further experiments with varying polymer content of the nanocomposite and possibly with varying surfactant coverage.


EPR spectroscopy on spin-labeled surfactants can potentially provide detailed information on local dynamics and structure in organoclays and clay-polymer nanocomposites. For organoclays based on synthetic expandable fluoromica we find that the anchor groups of a major part of the surfactant molecules are only weakly mobile up to temperatures well above the glass transition of polystyrene. Surprisingly, intercalation by polystyrene tends to immobilize rather than mobilize the anchor group region. Surfactant fractions with different mobility in the anchor group region are observed, with only a tiny part of much more mobile surfactants for the larger particles of expandable fluoromica (~1.5 [micro]m particle size) and comparable parts of fractions with more similar mobility for the smaller Laponite particles (~20 nm particle size). Modification of the Laponite particles by phosphonate groups influences surfactant dynamics. In expandable fluoromica organoclay we find two surfactant fractions that differ in the mobility of the hydrophobic tail ends. However, in this case the fractions are comparable already at 313 K and the more mobile fraction clearly dominates at 433 K.

Local elemental analysis by EPR can be used to detect proximity of the spin labels to the clay surface for fluoromica, which may potentially provide information on surfactant orientation with respect to the surface. Such elemental analysis also proves proximity of solvent molecules to surfactant anchor groups for organoclay dispersed in toluene. Measurements of the spatial distribution of surfactant anchor groups by the DEER experiment can detect changes in this distribution caused by polymer intercalation into the organoclay layer structure.

The present work has thus established EPR methodology for probing different aspects of the polymer-surfactant-clay interface layer. More extensive variation of system parameters and more sophisticated quantification of the data will be required to relate our preliminary findings to characteristics of the intercalation and exfoliation process and to properties of polymer-clay nanocomposites.


We thank Markus Wolkenhauer for acquiring WAXS data (to be published later) on our samples, Dariush Hinderberger for supplying the program for ratio analysis and performing this analysis for deuterium ESEEM data, and Qinghui Mao for gravimetric analysis of surfactant coverage. Financial help from DFG Schwer-punktprogramm 1051 "High-field EPR in Biology, Chemistry, and Physics" is gratefully acknowledged.

[c] 2004 Society of Plastics Engineers

Published online in Wiley InterScience (

DOI: 10.1002/pen.20104


1. Y. Kojima, A. Usuki, M. Kawasumi, A. Okada, T. Kurauchi, and O. Kamigaito, Polym. Sci. A: Polym. Chem., 31. 983 (1993).

2. E. P. Giannelis, Adv. Mater., 8, 29 (1996).

3. T. J. Pinnavaia and G. W. Beall, eds., Polymer-Clay Nanocomposites, Wiley & Sons, UK (2000).

4. R. A. Vaia and R. Krishnamoorti, eds., Polymer Nanocomposites: Synthesis, Characterization, and Modeling (ACS Symp. Ser. 804), Oxford University Press, New York (2001).

5. R. Krishnamoorti, R. A. Vaia, and E. P. Giannelis, Chem. Mater., 8, 1728 (1996).

6. A. C. Balazs, C. Singh, E. Zhulina, and Y. Lyatskaya, Acc. Chem. Res., 32, 651 (1999).

7. M. Alexandre and P. Dubois, Mater. Sci. Eng., 28, 1 (2000).

8. M. Biswas and S. S. Ray, Adv. Polym. Sci., 155, 167 (2001).

9. J. A. Weil, J. R. Bolton, and J. E. Wertz, Electron Paramagnetic Resonance, Wiley, New York (1994).

10. A. Schweiger and G. Jeschke, Principles of Pulse Electron Paramagnetic Resonance, Oxford University Press, Oxford (2001).

11. L. J. Berliner, ed., Spin Labeling: Theory and Applications, Academic Press, New York (1976).

12. L. J. Berliner, ed., Biological Magnetic Resonance, Vol. 14, Plenum, New York (1998).

13. D. J. Schneider and J. H. Freed, in Biological Magnetic Resonance, Vol. 8: Spin Labeling--Theory and Applications, Chapter 1, L. J. Berliner and J. Reuben, eds., Plenum Press, New York (1989).

14. G. Jeschke, Macromol, Rapid, Commun., 23, 227 (2002).

15. G. Jeschke, Chem. Phys. Chem., 3, 927 (2002).

16. S. E. Cramer, G. Jeschke, and H. W. Spiess, Macromol, Chem. Phys., 203, 182 (2002).

17. S. E. Cramer, G. Jeschke, and H. W. Spiess, Colloid Polym. Sci., 280, 569 (2002).

18. J. Y. Zhang, P. J. Carl, H. Zimmermann, and D. Goldfarb, J. Phys. Chem. B. 106, 5382 (2002).

19. R. A. Vaia and E. P. Giannelis, Macromolecules, 30, 7990 (1997).

20. T. M. Wu, S. F. Hsu, and J. Y. Wu, Polym. Eng. Sci., 42, 2295 (2002).

21. G. Jeschke and A. Schweiger, Chem. Phys. Lett., 246, 431 (1995).

22. G. Jeschke, Curr. Opin. Solid. St. M., 7, 181 (2003).

23. A. D. Milov, K. M. Salikhov, and M. D. Shirov, Fiz. Tverd. Tela (Leningrad), 23, 957 (1981).

24. A. D. Milov, A. G. Maryasov, and Yu, D. Tsvetkov, Appl. Magn. Reson., 15, 107 (1998).

25. M. Pannier, S. Veit, A. Godt, G. Jeschke, and H. W. Spiess, J. Mag. Res., 142, 331 (2000).

26. G. Jeschke, A. Koch, U. Jonas, and A. Godt, J. Mag. Reson., 155, 72-82 (2002).

27. J. H. Yang, Y. S. Han, J. H. Choy, and H. Tateyama, J. Mater. Chem., 11, 1305 (2001).

28. T. Ichikawa, L. Kevan, M. K. Bowman, S. A. Dikanov, and Yu. D. Tsvetkov, J. Chem. Phys., 71, 1167 (1979).


Max Planck Institute for Polymer Research

Postfach 3148, 55021 Mainz, Germany

*To whom correspondence should be addressed.

COPYRIGHT 2004 Society of Plastics Engineers, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2004 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Jeschke, G.; Panek, G.; Schleidt, S.; Jonas, U.
Publication:Polymer Engineering and Science
Date:Jun 1, 2004
Previous Article:Structure-property relationships in polymer blend nanocomposites.
Next Article:Polystyrene magadiite nanocomposites.

Related Articles
Polymerization compounding: epoxy-montmorillonite nanocomposites.
Extrusion processing of TPO nanocomposites.
Structure-property relationships in polymer blend nanocomposites.
Polystyrene magadiite nanocomposites.
Melt processing of PA-66/clay, HDPE/clay and HDPE/PA-66/clay nanocomposites.
Characteristics of polystyrene/polyethylene/clay nanocomposites prepared by ultrasound-assisted mixing process.
Polyolefin nanocomposites: formulation and development.
Morphology, thermal, and mechanical properties of vinylester resin nanocomposites with various organo-modified montmorillonites.
Conductive polypropylene/clay/polypyrrole nanocomposites.
Melt processing effects on the structure and mechanical properties of PA-6/clay nanocomposites.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters