Macromolecules 2008, 41, 4777-4785 4777

Interfacial Conformations and Molecular Structure of PMMA in

PMMA/Silica Nanocomposites. Effect of High-Energy Ball Milling
J. González-Benito*,† and G. González-Gaitano‡
Technological Institute of Chemistry and Materials “AlVaro Alonso Barba”, AVda. UniVersidad 30,
Leganés 28911 Madrid, Spain, and Departamento de Quı́mica y Edafologı́a, Facultad de Ciencias,
UniVersidad de NaVarra, Pamplona 31080, NaVarra, Spain
ReceiVed February 4, 2008; ReVised Manuscript ReceiVed May 2, 2008

ABSTRACT: This work offers a new vision of the nanoparticle-polymer interface in inorganic nanoparticle
filled thermoplastic nanocomposites in terms of macromolecular structure, conformations, and interactions. In
particular, the molecular structure and interfacial conformation of PMMA in silica/PMMA nanocomposites obtained
by high-energy ball milling (HEBM) have been studied. HEBM was applied to achieve an extraordinary dispersion
of high content of silica nanoparticles within the PMMA polymer. Fourier transform infrared spectroscopy (FTIR)
was used to investigate the effect that the important shear forces imposed by the HEBM process has on the
PMMA structure and its dynamics and also the effect of the presence of silica nanoparticles inserted by HEBM
on the particular conformations of PMMA appearing at the interface. It has been demonstrated that HEBM induces
particular conformational changes on both the ester group and the backbone of the PMMA, which seem to be the
cause of subsequent appearance of specific polymer chain packing at the interface.

1. Introduction Some researchers have tried to overcome this by studying model

polymer nacomposites consisting of polymer films confined
In recent years, organic polymer-inorganic particles nano- between silica plates,9 for instance.
composites have received great interest1–9 since they can exhibit
unique properties (thermomechanical, electrical, optical, etc.). In general, two experimental results are found for the glass
However, generally, these improved or even unique properties transition temperature, Tg, of polymer thin films: (i) Tg is a
can be only achieved if there is an adequate dispersion of the function of the thickness,9,13,14,17–19 and (ii) Tg can change if
nanoparticles within the polymer matrix. Besides, it is generally polymer film is deposited on a substrate.13,19 When the thickness
accepted that to properly understand the influence of nanofillers of the polymer film decreases and favorable interactions between
in the final properties is necessary to get a perfect dispersion of the polymer and the substrate exist, an enhancement or no
nanoparticles with clear separation each other. In fact, the change in the Tg has been usually found while, when the
objective is usually to obtain the same interparticle distance with thickness decreases in a film with poor interaction polymer-
homogeneous interphases which, in the ideal case, would imply substrate, a Tg decrease is observed. In the case of favorable
a perfect surface wetting by the polymer. interactions between the polymer and the substrate, experiments
It is generally accepted that the so-called “interphase zone” carried out with neutron reflectivity have shown that there exists
plays an important role on inducing important changes on the an increase in the density gradient to the surface.20,21 All these
properties observed in polymers filled with nanoparticles. The results seem to be related with the specific conformation adopted
interphase zone is defined as the region surrounding each particle by the polymeric chains exactly at the surface of the substrate
where the polymer chain dynamics may be altered due to that allows better macromolecular packing. For example, there
intimate contact with the nanoparticles (resulting in restricted are studies pointing out the existence of conformational changes
or enhanced mobility) plus the surrounding matrix influenced in the PMMA backbone on an aluminum surface22 due to the
by that immediate layer.10 In fact, the small size of nanoparticles existence of ionic bonds generated by the hydrolysis of ester
makes the interfacial zone a so important fraction of the whole groups over the aluminum surface. However, Grohens et al.,23
material that it can be considered as one of the proper based on the previous works of O’Reilly,24 estimated the
constituents of the composite. However, specific analysis to conformational energies at the interfaces using reflected infrared
adequately study the interphase region is not an easy task, and absorption spectroscopy discussing their results in terms of
still considerable efforts should be focused on this issue. conformational changes due to the restrictions imposed by the
Bansal et al.11 have recently studied the thermal behavior of surface to the chain motion.
polymeric thin films to subsequently extrapolate their results At this point it seems opportune to take a further step by
to the thermal behavior of nanoparticles filled polymers. Their performing these studies in polymers filled with nanoparticles
experiments are based on the hypothesis that the polymer close with the objective of investigating the influence of these fillers
to the nanoparticles surface should behave in a similar way to in the macromolecular conformations and, if possible, to
that of a thin film of the same polymer deposited over a substrate understand their relation with the final properties of the “real”
with the same nature than that of the nanoparticles. However, nanocomposite material. A paramount requirement must be
several works have shown that the dynamics of a polymer fulfilled, i.e., the “perfect” dispersion of the nanoparticles within
surface may be quite different than that of the bulk,12–16 and the polymer.
therefore new experiments should be carried out to reinforce Recently, González-Benito et al.25 have used high-energy ball
the conclusions extracted from the experiments on thin films. milling (HEBM) as a method to mix a polymer with inorganic
nanoparticles in the solid state, achieving an extraordinary
* Corresponding author. E-mail: [email protected].
dispersion of nanoparticles within the polymer. However, it is
†
Technological Institute of Chemistry and Materials “Alvaro Alonso Barba”. not clear yet what is the effect that the important shear forces
‡
Universidad de Navarra. imposed to the mixture provoke in the polymeric structure and,
10.1021/ma800260k CCC: $40.75  2008 American Chemical Society
Published on Web 06/11/2008
4778 González-Benito and González-Gaitano Macromolecules, Vol. 41, No. 13, 2008

finally, what kind of adhesion is achieved between the nano- Table 1. Absorption Peaks in s-PMMA (1050-1300 cm-1)
particles and the polymer. This has led us to consider two main frequency
objectives, by using FTIR spectroscopy, to determine what is peak (cm-1) Nagai’s assignment26
the effect of the HEBM: (i) on the molecular structure of the ν1 1268 νa(C-C-O) coupled with ν(C-O)
polymeric chains of a commercial PMMA and (ii) on the ν2 1238
macromolecular conformations at the silica nanoparticles/ ν3 1192 skeletal stretching coupled with internal C-H deformation
PMMA interface. This study should provide important informa- ν4 1148
ν5 1060 planar zigzag arrangement of the backbone carbon atoms
tion to understand which kind of interactions and arrangement
between the macromolecules and nanoparticles are required to
achieve such a good nanoparticle dispersion within the polymer for each sample from 400 to 4000 cm-1 were taken at different
as that one obtained by HEBM.25 temperatures within the interval 30-150 °C every 10 °C. The
samples were heated in a SPECAC heater specifically designed for
In particular, it is well-known that an important molecular the FTIR equipment controlling the temperature with an accuracy
feature, which has been identified as contributing to certain of (1 °C. The temperature stability was achieved after 5 min of
physical properties, is the conformational energy (∆H). Probably hold on time.
because of its simplicity, infrared spectroscopy has been the 2.4. Characteristics of the PMMA Infrared Spectrum. Al-
most widely used method to measure energy differences between though still with very slight differences in the final interpretation,
conformational isomers and has been applied to polymers. there have been many authors who did their particular assignations
Because of this, the simple use of IR spectroscopy was chosen to the IR absorption bands of the infrared spectrum of PMMA. In
for this work. particular, it is generally accepted that in the range between 1050
and 1300 cm-1 there are at least four absorption peaks associated
with the ester group for syndiotactic poly(methyl methacrylate),
2. Experimental Section
s-PMMA,26,27 which do arise from specific interactions.27 The
2.1. Materials. Fumed silica nanoparticles (99.8% pure silica, different possible rotational positions (PMMA conformations) of
200 m2/g surface area, 2.2 g/cm3 density, and 14 nm diameter) were the ester group seem to yield different force constants (hence
supplied by Sigma-Aldrich and atactic poly(methyl metacrylate), different peaks), so variations on the absorbances of those peaks
PMMA (Oroglas V852T), with density 1.19 g/cm3, average number should be related with the population of the different PMMA
molecular weight Mw ) 75 300, and polydispersity index PDI ) conformations. In Table 1 Nagai’s assignation of the s-PMMA IR
1.5, was purchased from Atoglas S.A. Methyl ethyl ketone and absorption bands observed in the interval 1050- 1300 cm-1 is
tetrahydrofuran HPLC grade (Aldrich) were used as solvent of collected.26
PMMA in the light scattering and size exclusion chromatography In the event for which two absorption peaks are due to the same
experiments, respectively. vibration but the frequencies are different (the ester group can be
2.2. Sample Preparation. PMMA pellets were first grinded in found in either one of two different rotational positions, causing
a MF 10 basic Microfine grinder drive (IKA WERKE) to obtain the force constants to be slightly different), the sum of the integrated
relatively fine PMMA particles. These were mixed with the fumed absorbances for both peaks should not depend on the temperature,
silica nanoparticles in a weight proportion of 10% of silica. provided the absorbance per ester group is the same for both
Subsequently, about 20 g of mixture was introduced in a vial of locations. On the other hand, the ratio of the absorbances for the
alumina with 190 g of 20 mm diameter alumina balls. The vial two peaks should vary with temperature as the number of ester
was then hermetically closed and placed in a Pulverisette 5 Fritsch groups increases in one of the conformations (conformation i) at
apparatus where the powder was milled (400 rpm) for 10 h at room the expense of the number of ester groups in the other conformation
temperature (high-energy ball milling, HEBM). To study the effect (conformation i + 1). Therefore, a plot of ln [A(νi)/A(νi+1)] vs 1/T
of the milling time along the blending process, samples of (van’t Hoff plot) should be a straight line if the population
silica-PMMA were extracted at different milling times (1, 2, 4, 6, distribution between the two conformers is controlled by the
8, and 10 h), leaving the equipment to rest 25 min every hour of Maxwell-Boltzmann statistics. Here A(νi) and A(νi+1) represent
active milling. On the other hand, to avoid excessive heating of the absorbances of the two peaks assigned to the same vibration in
the equipment as well as of the milled powders, a subcycle of two conformations i and i + 1, respectively, and T is the absolute
milling was considered: every 6 min of active milling was followed temperature. In fact, the absorption peaks at ∼1270 and ∼1240
by 3 min of resting. For later comparisons PMMA alone (without cm-1 have been used habitually in the literature to calculate the
nanoparticles) was subjected to the same ball milling cycles. van’t Hoff conformational energies from the slopes of the van’t
The powders coming from the milling process were diluted in Hoff plots.28,29
KBr (less than 1 wt %) to subsequently prepare tablets for being Other important bands for structural studies are those at 860 cm-1
studied by FTIR spectroscopy. All the samples prepared had characteristic for the all-trans or close to the all-trans (TT)
sufficiently low amount of PMMA or silica-PMMA to satisfy the conformation, while the trans-gauche one (TG) absorbs at 843
Lambert-Beer law. cm-1 and is characteristic of the random coil.28,30
2.3. Experimental Techniques. The hydrodynamic radii of
PMMA chains were obtained by dynamic light scattering (DLS) 3. Results and Discussion
with a photon correlation spectrometer DynaPro MS/X by analyzing
the autocorrelation functions by the cumulants method using the
3.1. Effect of HEBM on PMMA. DLS permits to obtain,
Dynamics V6 software. Samples were dissolved in methyl ethyl from the intensity autocorrelation function, the size spectrum
ketone (7 mg/mL) and filtered with syringe filters of 0.02 µm of the particles present in the sample. The analysis of this
(Anotop) before the measurement. All experiments were carried function can be carried out assuming a single relaxational mode
out at 25 °C. by the cumulant method, yielding the mean size, expressed as
Size exclusion chromatography as a function of milling time (0, the hydrodynamic radius, Rh, and the width of the distribution.
4, and 10 h) was performed with a Shimadzu CTO-6A unit using The apparent radii and polydispersity, Γ, as a function of milling
a 515 HPLC pump of Waters and PLGel columns from Polymer time for PMMA are plotted in Figure 1. Before the milling the
Laboratories. Refractive index detection was performed at room mean radius is 7.8 nm, which reduces to 5.3 nm after 10 h of
temperature from 10 mg mL-1 polymer solutions in THF (20 µL treatment. At the same time, the scattered intensity diminishes
injections at a flow rate of 1 mL min-1) with a Shimadzu RID-6A following the same trend, passing from 46 to 21 kcounts/s, as
refractometer. correspond to the smaller size of the polymer (data not shown).
FTIR spectra were recorded with a FT-IR Spectrum GX (Perkin- Although DLS experiments do not allow obtaining an absolute
Elmer). Five scans and a resolution of 4 cm-1 were used. Spectra molecular weight, it is possible to estimate the relative change
Macromolecules, Vol. 41, No. 13, 2008 PMMA in PMMA/Silica Nanocomposites 4779

Figure 1. Evolution of PMMA hydrodynamic radius, Rh, and polydis-

persity, Γ, as a function of milling time (methyl ethyl ketone as solvent).

Figure 3. Absorption FTIR spectra of PMMA at 30 and 150 °C without

further treatment and subjected for 10 h to high-energy ball milling.

Figure 2. Normalized chromatograms obtained by SEC of PMMA

milled for different times.

in molar mass after milling. Thus, if a random coil model in a

thermodynamically good solvent is assumed for the PMMA (as Figure 4. Evolution with temperature of the normalized IR absorption
methyl ethyl ketone is), it can be demonstrated that M′ ) M0(R′/ associated with the cis and trans ester group vibration, νa(C-C-O).
R0)2, where M is the average molecular weight of the polymer Sample of PMMA milled for 2 h.
with hydrodynamic radius R′ and the 0 subscript stands for the
polymer without treatment. Therefore, taking the data obtained representing the extremes studied for those variables (temper-
from DLS (Figure 1), it is possible to estimate that the average ature and milling time) are plotted in Figure 3.
molecular weight decreases about 54% after 10 h of milling. Let us consider first the effect of temperature. This is reflected
These results clearly point out the chain scission that was in small variations in the relative absorption peaks (see Figure
considered in ref 25 but not confirmed experimentally there. In 3). One example of this effect can be observed in the relative
addition, SEC provides complementary information to DLS, as absorbance between the peaks centered at 1270 cm-1 (ν1) and
shown in Figure 2 where the normalized curves of the output 1242 cm-1 (ν2), which is clearly shown in Figure 4. These peaks
signal from the refractometer as a function of milling time have were assigned by Nagai26 and Willis et al.32 to the νa(C-C-O)
been plotted. Here, the PMMA peak shifts to longer elution vibration, whereas Havriliak and Roman27 pointed out that the
volumes and diminishes its height, whereas a new peak of splitting of this vibration into two components is a consequence
smaller molecular weight arises, whose contribution becomes of two rotational-isomeric states of the ester group, where the
concomitantly higher. All these results prove undoubtedly the C-OCH3 and CR-CH3 bonds are in mutual cis or trans
rupture of the polymeric chains caused by the HEBM and fully orientation arising from the internal CR-CO bond rotation
agree with those by Smith et al.,31 who studied the effect of (Figure 5).
high-energy mechanical milling on PMMA at ambient and From spectroscopic and ab initio studies of MMA, Owen et
cryogenic temperatures. al.33 and Konoka et al.34 associate each band with one of the
Another interesting parameter to be analyzed is the polydis- two conformers. The ab initio results have established that the
persity. This parameter has been included in Figure 1, where it high-frequency band (ν1) is due solely to the trans conformation,
can be observed how the distribution widens, undergoing an whereas the low-frequency component (ν2) is associated ex-
initial increase in Γ and followed by a decrease at highest milling clusively with the cis conformation. Thus, the corresponding
times. The chromatograms (Figure 2) follow the same trend, polymer peaks ν1 and ν2, must be assigned to respectively the
also. These findings are in accordance with those reported by trans and cis conformation of the ester group (Figure 5). In view
Smith et al.31 However, in our case it is more clearly evidenced of that, if equilibrium between rotational isomers is considered,
that the chromatogram peak broadens at intermediate milling the ratio of the absorbances for the two peaks must be related
times due to the appearance of a bimodal distribution, which to the conformational energy difference between the trans and
finally tends to disappear at longer milling times. cis rotational states by the van’t Hoff equation.
To see the effect of ball milling (in terms of milling time) Custom is to use curve fitting to analyze spectroscopic bands.
and the effect of temperature in the structure and conformation Although generally it gives good results, sometimes these
of the PMMA, IR absorption spectra of the polymer in the more methods usually take the presence of theoretical bands whose
sensitive region to conformational changes were collected. In existence is not completely ensured. Besides, it is necessary to
particular, as an example, four normalized FTIR spectra assume a specific shape for the bands (Lorenzian, Gaussian,
4780 González-Benito and González-Gaitano Macromolecules, Vol. 41, No. 13, 2008

Figure 5. 3D models representing schematically the (a) trans and (b)

cis rotational isomers causing the IR absorption peaks of the ester group
at 1242 and 1270 cm-1, respectively.

etc.). Because of this, it seems reasonable to try to extract

information directly from the experimental data, with minimal
baseline corrections and normalization for subsequently con-
ducting proper comparisons. In this procedure the most impor-
tant point is to follow a systematic way in the analysis of the Figure 6. Evolution with temperature of the first moment of the band
bands, selecting the same intervals for baseline corrections, and centered at 1270 cm-1 for the PMMA subjected to different milling
times (0, 1, 2, 4, 6, 8, and 10 h).
keeping exactly the same method for obtaining a particular
parameter to finally be able to compare changes in the Table 2. Values of Glass Transition Temperature, Tg, Obtained
spectroscopic bands. Indeed, these procedures might lead to from the Plots of the First Moment of the Absorption Band as a
distortions in the band shape, but if these are considered in the Function of Temperature from the Vant’ Hoff Plots, Together
same way for all the spectra to be compared, the final conclusion with the Conformational Energy Difference, ∆H, above and
should not be altered. below Tg (PMMA Samples)
The plots in Figure 4 have been obtained after considering a milling Tg (°C) from ν1 Tg (°C) from ∆H (cal mol-1) ∆H(cal mol-1)
baseline correction from 1217 to 1326 cm-1 and normalizing time (h) band shift Aν2/Aν1 ratio T < Tg T > Tg
at the maximum of the peak centered at higher energy. It can 0 115 115 177 ( 7 786 ( 39
be observed (i) a small peak shift to lower energies (about 5 1 111 114 182 ( 5 730 ( 30
2 106 103 156 ( 8 732 ( 23
cm-1 for the peak at 1270 cm-1 and about 2 cm-1 for the peak 4 100 100 140 ( 10 782 ( 25
centered at 1240 cm-1) and (ii) a relative absorbance increase 6 97 97 142 ( 10 771 ( 26
of the peak centered at higher wavenumbers when the temper- 8 94 95 144 ( 9 771 ( 24
ature increases. 10 90 90 146 ( 13 840 ( 24
Grohens et al.35 associated the shift to higher frequencies in
the νa(C-C-O) band at 1270 cm-1 to specific attractive rotation around a bond of the pendant ester group, for instance.
interactions involving the methoxy group. In our case, when If the attractive interactions are considered to be due to
PMMA alone is considered, the shift of this band only can be contribution of the methoxy groups, the only way to change
due to changes in inter- or intramolecular interactions. Therefore, their positions to reduce the attractive interactions is by rotating
the results obtained suggest that heating the sample allows the ester group around the axis which passes along the bond
breaking, at least partially, those linkages (attractive interactions) CR-COOCH3. This assumption agrees with the results of
by a simple thermal activation which might be moreover Schmidt-Rohr et al.,39 who suggested that the -relaxation in
conditioned by changes in chain conformation. The fact that PMMA consists of a 180° flip of the ester group about the
the ν1 band shifts more than the ν2 suggests that the trans C-CO bond. On the other hand, the second temperature
conformation favors the attractive interactions involving the observed as a change in the slope in Figure 6 coincides with
methoxy group with preference to the cis conformation. the R-relaxation in PMMA or the glass transition temperature,
Because of the small shift observed, about 5 cm-1 for the ν1 Tg, obtained by DSC.25 This relaxation is usually assigned to a
band, and the resolution of the spectra, 4 cm-1, the ν1(C-C-O) cooperative motion of the backbone bonds, which is the other
bands were baseline corrected with a straight line built from possible molecular movement that may reduce attractive interac-
1258 to 1298 cm-1. This helps to improve mathematically tion by putting apart the interacting groups each other.
(indirectly) the resolution by using the first moment of the In Figure 6 all milling times under study from 0 to 10 h have
absorption band, 〈ν1〉. Specifically, the absorption bands were been included. In all the cases the same behavior as a function
considered as distributions of energies prior to taking from them of temperature was observed although with some differences:
the first moment of the distribution.36 (a) the temperature at which the second thermal transition occurs
Figure 6 shows, as an example, a plot of 〈ν1〉 for the (Tg) decreases with milling time (Table 2) and (b) the slopes of
ν1(C-C-O) band as a function of temperature. As can be the plots increase with milling time.
expected from observation of Figure 4, the 〈ν1〉 decreases as On the other hand, and considering what was stated above,
temperature increases showing two slope changes, in particular result ii extracted from Figure 4 which accounts for the variation
at 55 °C (very smooth) and 115 °C, respectively, in the absence in the relative absorbance between the peaks ν1 and ν2 can be
of milling (Figure 5). The first temperature matches that of the interpreted in terms of a population change between the two
-relaxation in PMMA37,38 which suggests that the -relaxation different rotational positions (temperature-induced cis-trans
must be due to a motion activated at low energy, such as the rotational isomerization of the ester group). In particular, the
Macromolecules, Vol. 41, No. 13, 2008 PMMA in PMMA/Silica Nanocomposites 4781

Figure 7. Van’t Hoff plots in terms of absorbance ratios of the cis-trans

rotational isomers of the ester group for the pure PMMA subjected at Figure 8. Evolution of the νa(C-C-O) band as a function of milling
different milling times (0, 1, 2, 4, 6, 8, and 10 h). time for PMMA.

ratio of absorbances at 1240 and 1270 cm-1, Aν2/Aν1, decreases

(ii) the band shift is more important with temperature, and (iii)
as temperature increases. This result means taking into account
the change in the slope assigned to the Tg occurs at lower
the peaks assignation, that the increase in the temperature implies
temperatures (Table 2).
an enhancement of the trans content at expense of the cis
content. The same behavior as a function of temperature is Assuming the explanation put forward above, which accounts
observed for all the milling times. Therefore, in order to study for restrictions in the ester group motion due to specific
the effect of temperature and ball milling process in the PMMA interactions, the first evidence suggests a conformational change
macromolecular structure and conformation, these data were induced by the ball milling, which would reduce the number of
analyzed for the samples by using the van’t Hoff equation as specific interactions or would make them to be weaker. The
described in ref 40 at all milling times and plotted as shown in second evidence is concordant with the first one because the
Figure 7. (In our case the absorbance ratios, Aν2/Aν1, were weaker the interactions between ester groups the more pro-
obtained from plots as those represented in Figure 4.) nounced temperature effect on the band shift is expected. Finally,
It is clearly observed how the absorbance ratio Aν2/Aν1 the reduction of Tg was elsewhere interpreted in terms of a
changes its slope at a defined value of temperature that can be reduction in the average molecular weight of the PMMA due
associated with the glass transition temperature, Tg, considering to chain scission when the polymer is subjected to extremely
the values obtained for the same samples measured by dif- high shear forces,25,31 which completely agrees with what has
ferential scanning calorimetry (DSC).25 The Tg for each sample been observed experimentally in this work by DLS and size
(Table 2) were deduced from the intercept between the straight exclusion chromatography.
lines fitted in each part of the plots (Figure 7). Besides, Figure 3 shows that the relative absorbance of several
The conformational difference energies, ∆H, were then bands also changes with milling time; in these cases other type
calculated from linear regression analysis (Table 2) having in of representations done with different normalization criteria aid
all cases regression coefficients higher than 0.99. Two values to perceive better such changes. For instance, in the case of ν1
of ∆H are then obtained for the same ester group, the lower and ν2 bands similar plots as those shown in Figure 4 can be
one coming from data at temperatures below the Tg and the represented (Figure 8) in order to see whether the milling process
higher one obtained at temperatures above the Tg. In accordance exerts any population change between the two different rota-
with the experimental work by O’Reilly and Mosher24 and with tional positions (rotational isomerization of the ester group).
the theoretical work carried out by Vacatello and Flory41 and As can be seen in Figure 8, the ν2 band increases at the expense
Sundararajan,42 these results point out that the conformational of the ν1 one with milling time, showing a clear equilibrium
cis-trans energy difference strongly depends on the backbone displacement to the cis conformation of the ester group. This
conformation. Since backbone motions are assumed to be frozen result is in accordance, first, with the assumption of a ball
at temperatures below Tg, the conformational energies calculated milling-induced conformational change and, second, with the
at those temperatures may only refer to changes in side-chain less number or weaker specific interactions if they are favored
conformations. However, above the Tg, both the side-chain and with the trans conformation, as suggested by the band shifting
the backbone motions are allowed, and therefore the calculated in Figures 4 and 6. Furthermore, it is worth highlighting in
energies might reflect a cooperative movement.23 Figure 8 the presence of the isosbestic point, which is the most
Furthermore, a possible extended explanation can be con- realistic evidence of the equilibrium that exists between the
sidered here. At relatively low temperature (at least 30 °C) only species absorbing at ν1 and ν2.
free ester groups (isolated, having no specific interactions), The Tg’s at different milling times can be obtained from curve
without other restrictions for its movement different than the fitting of plots in Figure 7 (Table 2). As expected, a reduction
proper rotation around a single bond, are allowed to isomerize. on Tg with milling time is observed in good correspondence
On the contrary, when the temperature is high enough, these with the Tg values different from the ν1 band shifts, leading to
ester groups intra- or intermolecularly interacting each other the same conclusion. On the other hand, the similar values for
begin to rotate aided by the cooperative backbone motion conformational difference energy, ∆H, above and below Tg
occurring at temperature higher than Tg. This scheme would regardless the milling time (Table 2) suggest that the ball milling
perfectly account for the interpretation made for the ester band process does not exert any change in the dynamic mechanism
shifts observed. of the PMMA, at least within the experimental uncertainty in
Let us consider now the effect of milling time. This variable ∆H which might include small changes due to variation in weak
has important consequences on the FTIR spectra. In Figure 6 it specific interactions.
is observed that the higher the milling time is (i) for the same Another interesting observation from Figure 3 is that the width
temperature the ν1 band shifts toward slightly higher frequencies, of the carbonyl band (1733 cm-1) diminishes, and its relative
4782 González-Benito and González-Gaitano Macromolecules, Vol. 41, No. 13, 2008

absorbance clearly increases with milling time. In addition, the

width of the ν1 band also decreases.
At this point, one could think that the PMMA particles within
the KBr plates might lead to scattering phenomena producing
partial absorption at the edge of the particles. Because of this,
it is possible to think that as the particle size of the PMMA
reduces as grinding proceeds, over-absorption and other effects
might play a smaller role and the intensity relationships may
become closer to those of thin film spectra. However, the spectra
analysis presented here shows, among other things, the existence
of an unambiguous isosbestic point and clear changes pointing
to the glass transition temperatures (they perfectly match with
those obtained by DSC). Therefore, we must conclude that these
evidences do not seem to come from sample artifacts. On the
other hand, these so alleged scattering effects should have been
detected at higher wavenumbers, where the scattering is really
Figure 9. Normalized spectra at the carbonyl peak for the pure PMMA
important, and this has not been the case. and nanosilica-filled PMMA (silica-PMMA) samples milled for 10 h.
In thin film spectra, the CdO mode is the most intense in The spectra were collected at 30 °C.
the spectrum.24,43–48 However, from refs 24 and 43–48 important
variations in the absorbance ratio between the carbonyl band
and the highest peak of the group assigned to ester vibration
can be plainly observed. In particular, this ratio has two clear
dependences: one of them is the film thickness, and the other
one is the tacticity of the PMMA. Therefore, these results point
that the relative absorbances due to the vibration modes
associated with the PMMA groups are highly dependent on their
spatial disposition (conformations and configurations of the
macromolecules). It is well-known that the polymer conforma-
tion can greatly change at its surface with respect to that in the
bulk, and this may be the reason why important changes are
observed as a function of film thickness.
The shape and widths of infrared bands are an important
source of information about interactions and molecular dynam- Figure 10. Evolution of the normalized IR absorption associated with
the cis and trans ester group vibration, νa(C-C-O), with temperature
ics.49 In polymer spectroscopy, it is common to study the shapes for the sample silica-PMMA milled for 2 h.
and widths of the bands because they are related to the
distribution of local environments experienced by radiation
absorbing or emitting moieties, but the band due to a particular absorbance between the peaks centered at 1270 cm-1 (ν1) and
absorption or emission reflects the superimposition of different 1242 cm-1 (ν2) and (ii) in the ν1 shifts clearly shown in Figure
frequencies that are a result of interactions in slightly different 10 in which, by normalizing the interval corresponding the ν1
local environments.12,49,50 Therefore, the width reduction of the and ν2 bands, similar plots to those of Figure 4 are represented
carbonyl and ester bands suggests that this group is the other showing the same behavior with temperature.
counterpart responsible for the specific interactions appearing Figure 11 shows the plots for the values of 〈ν1〉 associated
in the ester group and which are reduced as a consequence of with the ν1(C-C-O) band as a function of temperature when
the milling process. Again the experimental data are pointing the PMMA is filled with nanosilica particles. As can be expected
out that the great shear forces applied to the polymer during (Figure 10), 〈ν1〉 decreases as temperature increases, showing
the mill process exert a so dramatic change in the structure and nearly two slope changes (Figure 11) as in the cases of milled
conformation of the PMMA that the orientation between PMMA without silica nanoparticles, and therefore, the same
interacting groups gives up operating. In addition to this, the interpretation can be done for these PMMA-modified systems.
increase in the relative absorption of the carbonyl group probably However, when silica nanoparticles are within the polymer after
reflects a higher change of dipole moment in its vibration if the HEBM process, two important differences arise: (i) for a
less specific interactions are affecting it. specific temperature and milling time, the values of 〈ν1〉 are
slightly lower, and (ii) the second temperature onset for the 〈ν1〉
3.2. Nanosilica-Filled PMMA. For comparison, the normal- vs T plots, Tg’s, reduces to some extent (Table 3).
ized spectra at the carbonyl peak for the pure PMMA and If the position of the ν1 band depends on the attractive
nanosilica-filled PMMA (silica-PMMA) are shown in Figure interactions involving the methoxy and carbonyl groups, which
9. These spectra were collected at 30 °C for the samples which are conditioned by changes in chain conformation as was stated
underwent 10 h of ball milling. before (they are favored by the trans conformation of the ester
In general, in the region shown in Figure 9, the FTIR spectra group), the slightly lower values of 〈ν1〉 in the silica-PMMA
are practically the same except (i) for the interval from 980 to samples may be due to lower population of trans conformers. Since
1150 cm-1, in which the contribution of the bands coming from the presence of silica nanoparticles is the only difference between
the Si-O-Si stretching modes typical of silica are clearly both kinds of samples, two possible explanations might account
observed, and (ii) for the carbonyl band which is slightly for this evidence: either specific interactions between silica nano-
narrower in the silica-PMMA sample. These effects are better particles and PMMA might induce cis conformation enhancement,
noticed in the difference spectrum (silica-PMMA-PMMA) or the presence of rigid silica nanoparticles might force slightly
shown as an inset in Figure 9. more chain scissions during the milling process, thus increasing
Let us consider now the effect of temperature on the FTIR the population of cis with respect to the trans conformers.
spectra for the samples of PMMA with silica nanoparticles. As previously shown (Figure 8), the milling process induces
Once more, this effect can be observed: (i) in the relative an increase in number of the cis conformers. Therefore, the
Macromolecules, Vol. 41, No. 13, 2008 PMMA in PMMA/Silica Nanocomposites 4783

all the milling times considered. Therefore, in the presence of

dispersed silica nanoparticles, the effect of the HEBM in the
macromolecular structure and conformation of PMMA can also
be examined by fitting the experimental data via a van’t Hoff
equation40 (Figure 12). In this case the absorbance ratios (Aν2/
Aν1) were obtained from plots as those represented in Figure
10.
For the silica-PMMA samples and as in the case of PMMA
without milling, it is clearly observed how the absorbance ratio
Aν2/Aν1 changes its slope at a well-defined value of temperature
that can be associated with the glass transition temperature.25
The Tg values for each sample (Table 3) were obtained from
the intercept between the straight lines fitted in each part of the
plots. Once more, it is possible to see that in the presence of
silica nanoparticles the values of Tg are slightly lower, indicating
that the presence of rigid ceramic nanoparticles facilitates the
polymer chain scission.
As in the case of pure PMMA, the conformational energies
were calculated from linear regression analysis, having in all
cases regression coefficients higher than 0.99 (Table 3). For
the same ester group two values of ∆H were obtained, coming
from data at temperatures below the Tg (the lower one) and
above Tg (the higher one). As in the case of PMMA alone, the
Figure 11. Evolution with temperature of the first moment of the band
results point out a strong dependence of conformational
centered at 1270 cm-1 for the silica-PMMA sample subjected to cis-trans energy difference of the ester group on the backbone
different milling times (1, 2, 4, 6, 8, and 10 h). conformation, and a similar explanation is valid in this case.
One important difference, however, can be detected in the data
Table 3. Glass Transition Temperature, Tg, Obtained from the in Table 3; i.e., the conformational energies are about 33-35%
Plots of the First Moment of the Band as a Function of higher when silica nanoparticles are within the PMMA. A
Temperature and the Van’t Hoff Plots, Together with the
Conformational Energy Difference, ∆H, above and below Tg possible explanation of this fact may be the stabilization of the
(Silica-PMMA Samples) trans conformation by means of specific interactions between
the ester moieties through the carbonyl or the methoxy groups
milling Tg (°C) from ν1 Tg (°C) from ∆H (cal mol-1) ∆H (cal mol-1)
time (h) band shift Aν2/Aν1 ratio T > Tg T > Tg and the silanols of the silica nanoparticles. However, no
experimental evidence was found to support this argument. In
1 105 105 211 ( 14 1132 ( 30
2 107 106 187 ( 7 876 ( 21
fact, the bands assigned to the carbonyl and ester groups seem
4 100 96 205 ( 11 1087 ( 17 to point out just the opposite since they narrow only to some
6 93 93 204 ( 9 1073 ( 27 extent and do not shift to the direction associated with the
8 88 87 221 ( 15 1084 ( 28 existence of attractive interactions (lower wavenumbers in the
10 85 85 187 ( 15 1005 ( 21 case of carbonyl band and higher ones in the case of ester band).
Another explanation for the increase in the conformational
second explanation seems to be more plausible since it would energy difference when the silica nanoparticles are within the
explain the slightly lower Tg’s observed for the silica-PMMA PMMA is that the HEBM allows the spreading of the polymer
samples. In fact, for a particular temperature and milling time onto the nanoparticles surface in such a way that an efficient
the values of Aν2/Aν1 ratio are higher when the silica nanopar- coating is achieved due to a better macromolecular packing.
ticles are present. This effective packing might be the cause of the polymer
On the other hand, the van’t Hoff plots in Figure 12 show constraint nearby the nanoparticles surface, yielding therefore
how the Aν2/Aν1 ratio decreases with the temperature, which a reduction of the molecular mobility. Assuming this explanation
implies, as in the case of pure PMMA, the temperature produces one might conclude that the HEBM allows obtaining a very
an increase of the trans content with respect to the cis one. The good mechanical adhesion between the silica nanoparticles and
same behavior as a function of temperature was observed for the PMMA. In fact, this good adhesion has been already
experimentally confirmed by atomic force microscopy (AFM).25
With this explanation in mind one would expect higher values
of Tg due to a restricted polymer motion. However, this evidence
was not observed by FTIR. A possible reason may be that this
technique is not sensible enough to evidence a second slope
change between the Tg and 150 °C in the plots of Figure 12.
However, the DSC experiments for the same samples25 did show
a double Tg, being one of them even higher than the typical Tg
of PMMA at 115 °C. Although in ref 25 this higher Tg was
speculated to be due to a fraction of PMMA with higher
molecular weight, it seems now to be more reasonable to assign
it to the fraction of PMMA packed more efficiently nearby the
nanoparticles. In fact, this fraction with higher Tg was esti-
mated25 in 16% by weight in the whole polymer, which would
Figure 12. Van’t Hoff plots in terms of absorbance ratios of the imply a percentage of 0.14% of constrained polymer in the KBr
cis-trans rotational isomers of the ester group for the silica-PMMA disks for the FTIR experiments. Taking into account the
sample subjected at different milling times (1, 2, 4, 6, 8, and 10 h). absorbances measured from the PMMA spectra, such a small
4784 González-Benito and González-Gaitano Macromolecules, Vol. 41, No. 13, 2008

the presence of such nanoparticles, in addition to the milling

process, helps the polymer chain to adopt a specific conforma-
tion on the nanosilica surface, which consequently would allow
the macromolecules to be more densely packed at the interface.

Conclusions
The effects of high-energy ball milling (HEBM) on the
structure and interfacial conformations of PMMA in silica/
PMMA nanocomposites have been studied. Results from DLS
and SEC evidenced that PMMA subjected to HEBM at room
temperature experiences chains scission, more extensively the
longer the milling time is. The use of FTIR spectroscopy has
shown (i) the existence of inter- and/or intramolecular specific
interactions between ester groups that are conditioned by
Figure 13. Evolution of the νa(C-C-O) band as a function of milling
time for a silica-PMMA mixture. changes in chain conformation, the trans ester one favoring such
interactions with preference to the cis conformation and disap-
pearing by thermal activation, (ii) a Tg decrease of the PMMA
when it is subjected to the HEBM process, and (iii) that the
incorporation of silica nanoparticles within the PMMA by means
of HEBM allows to achieve a good mechanical adhesion
between them, generating an interface where the PMMA chains
have a preferential backbone conformation which helps them
to be better packed and therefore with restricted mobility.

Acknowledgment. The authors gratefully acknowledge the

MEC of Spain for financial support (projects MAT2007-61607 and
CTQ2006-14933/BQU).

References and Notes

(1) Yang, M.; Dan, Y. Colloid Polym. Sci. 2005, 284, 243–250.
(2) He, J.; Li, H.; Wang, X.; Gao, Y. Eur. Polym. J. 2006, 42, 1128–
1134.
(3) Kinloch, A. J.; Mohammed, R. D.; Taylor, A. C.; Eger, C.; Sprenger,
S.; Egan, D. J. Mater. Sci. 2005, 40, 5083–5086.
(4) Zhao, H. X.; Li, R. K. Y. Polymer 2006, 47, 3207–3217.
(5) Kim, S. H.; Kwak, S. Y.; Suzuki, T. Polymer 2006, 47, 3005–3016.
(6) Li, Y. Q.; Fu, S. Y.; Mai, Y. W. Polymer 2006, 47, 2127–2132.
(7) Yang, H.; Zhang, Q.; Guo, M.; Wang, C.; Du, R. N.; Fu, Q. Polymer
2006, 47, 2106–2115.
Figure 14. Effect of milling time on a selected region of the PMMA (8) Rong, M. Z.; Zhang, M. Q.; Zheng, Y. X.; Zeng, H. M.; Walter, R.;
FTIR spectra where information about the backbone conformation can Friedrich, K. Polymer 2001, 42, 3301–3304.
be extracted. (9) Rittigstein, P.; Priestley, R. D.; Broadbelt, L. J.; Torkelson, J. M. Nat.
Mater. 2007, 6, 278–282.
(10) Ash, B. J.; Siegel, R. W.; Schadler, L. S. Macromolecules 2004, 37,
amount of constrained polymer would yield about 0.005 1358–1369.
absorbance units for the ester groups, which is below the (11) Bansal, A.; Yang, H.; Li, C.; Cho, K.; Benicewicz, B. C.; Kumar,
detection limit of the FTIR equipment. S. K.; Shadler, L. S. Nat. Mater. 2005, 4, 693–698.
(12) Turrión, S. G.; Olmos, D.; Ekizoglou, N.; González-Benito, J. Polymer
The milling time effect has been also considered in the case 2005, 46, 4023–4031.
of silica nanoparticle-filled PMMA (Figure 13). In this case the (13) Keddie, J. L.; Jones, R. A. L.; Cory, R. A. Europhys. Lett. 1994, 27,
same general observation can be made, and therefore similar 59–64.
(14) Van Zanten, J. H.; Wallace, W. E.; Wu, W. L. Phys. ReV. E 1996, 53,
conclusions than in the case of pure PMMA can be extracted. R2053–R2056.
However, when silica nanoparticles are present, the milling (15) Forrest, J. A.; Dalnoki-Veress, K.; Stevens, J. R.; Dutcher, J. R. Phys.
effect described in the case of PMMA alone seems to be more ReV. Lett. 1996, 77, 2002–2005.
pronounced since Aν2/Aν1 ratios are higher for all the milling (16) Fryer, D. S.; Peters, R. D.; Kim, E. J.; Tomaszewski, J. E.; Pablo,
J. J.; White, C. C.; Wu, W. Macromolecules 2001, 34, 5627–5634.
times. Once more, this result suggests that the presence of rigid (17) Roth, C. B.; McNerny, K. L.; Jager, W. F.; Torkelson, J. M.
ceramic particles slightly favors the polymer chain scission. Macromolecules 2007, 40, 2568–2574.
Finally, the region of the FTIR spectra where interesting (18) Ellison, C. J.; Torkelson, J. M. Nat. Mater. 2003, 2, 695–700.
information about backbone conformation can be found is shown (19) Keddie, J. L.; Jones, R. A. L.; Cory, R. A. Faraday Discuss. 1994,
98, 219–230.
in Figure 14. As mentioned in the Experimental Section, the (20) Fernandez, M. L.; Higgins, J. S.; Penforld, J.; Shackelton, C. S. Polym.
absorption band at 860 cm-1 is characteristic for the all-trans Commun. 1990, 31, 124–127.
or close to the all-trans conformation (TT), whereas the (21) Wu, W.; Majkrzak, C. F.; Satija, S. K.; Anker, J. F.; Orts, W. J.;
trans-gauche one (TG) absorbs at 843 cm-1 and is characteristic Satkowski, M.; Smith, S. D. Polymer 1992, 33, 5081–5084.
(22) Konstadinis, K.; Thakkar, B.; Chakraborty, A.; Potts, L. W.; Tannen-
of the random coil.28,30 In all the samples under study it was baum, R.; Tirrell, M.; Evans, J. F. Langmuir 1992, 8, 1307–1317.
mainly observed the absorption at 842 cm-1, pointing out that (23) Grohens, Y.; Brogly, M.; Labbe, C.; Schultz, J. Polymer 1997, 38,
the most important backbone conformation is the TG. However, 5913–5920.
only in the samples with silica nanoparticles subjected to the (24) O’Reilly, J. M.; Mosher, R. A. Macromolecules 1981, 14, 602–608.
(25) Castrillo, P. D.; Olmos, D.; Amador, D. R.; González-Benito, J. J.
milling process for more than 6 h a new peak at 874 cm-1 Colloid Interface Sci. 2007, 308, 318–324.
appears. If this peak is attributed to the existence of macro- (26) Nagai, H. J. Appl. Polym. Sci. 1962, 7, 1697–1702.
molecules with a highly trans conformation, one may think that (27) Havriliak, S.; Roman, N. Polym. Prepr. 1966, 7, 253–400.
Macromolecules, Vol. 41, No. 13, 2008 PMMA in PMMA/Silica Nanocomposites 4785

(28) Chen, J.; Zheng, G.; Xu, L.; Zhang, J.; Lu, Y.; Xue, G.; Yang, Y. Pawelzik, U.; Boeffel, C.; Spiess, H. W. Macromolecules 1994, 27,
Polymer 2001, 42, 4459–4462. 4733–4745.
(29) Grohens, Y.; Prud’homme, R. E.; Schultz, J. Macromolecules 1998, (40) Tretinikov, O. N.; Ohta, K. Macromolecules 2002, 35, 7343–7353.
31, 2545–2548. (41) Vacatello, M.; Flory, P. J. Macromolecules 1986, 19, 405–415.
(30) Spevacek, J.; Schneider, B.; Dybal, J.; Stokr, J.; Baldrian, J. J. Polym. (42) Sundararajan, P. R. Macromolecules 1986, 19, 415–421.
Sci., Part B: Polym. Phys. 1984, 22, 617–635. (43) Lade, R. J.; Morley, I. W.; May, P. W.; Rosser, K. N.; Ashfold,
(31) Smith, A. P.; Shay, J. S.; Spontak, R. J.; Balik, C. M.; Ade, H.; Smith, M. N. R. Diamond Relat. Mater. 1999, 8, 1654–1658.
S. D.; Koch, C. C. Polymer 2000, 41, 6271–6283. (44) Shin, H. S.; Lee, H.; Jun, C.; Jung, Y. M.; Kim, S. B. Vib. Spectrosc.
(32) Willis, H. A.; Zichy, V. J. I.; Hendra, P. J. Polymer 1969, 10, 737–
2005, 37, 69–76.
746.
(33) Baker, B. L.; Orgill, M. K.; Owen, N. L.; Stephenson, E. H.; Williams, (45) Shin, H. S.; Jung, Y. M.; Oh, T. Y.; Chang, T.; Kim, S. B.; Lee,
G. A.; MacDonald, J. N.; Boggs, J. E. J. Mol. Struct. 1995, 356, 95– D. H.; Noda, I. Langmuir 2002, 18, 5953–5958.
104. (46) Tretinnikov, O. N.; Ohta, K. Macromolecules 2002, 35, 7343–7353.
(34) Tsuji, T.; Ito, H.; Takeuchi, H.; Konaka, S. J. Mol. Struct. 1999, 475, (47) Tretinnikov, O. N.; Zhbankov, R. G. J. Mater. Sci. Lett. 1991, 10,
55–63. 1032–1036.
(35) Grohens, Y.; Schultz, J.; Prud’homme, R. E. Int. J. Adhes. Adhes. (48) González-Benito, J.; Koenig, J. L. Macromolecules 2002, 35, 7361–
1997, 17, 163–167. 7367.
(36) Mikes, F.; González-Benito, J.; Serrano, B.; Bravo, J.; Baselga, J. (49) Painter, P.; Sobkowiak, M.; Park, Y. Macromolecules 2007, 40, 1730–
Polymer 2002, 43, 4331–4339. 1737.
(37) Hammerschmidt, J. A.; Gladfelter, W. L. Macromolecules 1999, 32, (50) Moron, R. L.; Olmos, D.; González-Benito, J. Compos. Sci. Technol.
3360–3367. 2006, 66, 2758–2768.
(38) Bistac, S.; Schultz, J. Int. J. Adhes. Adhes. 1997, 17, 197–201.
(39) Schmidt-Rohr, K.; Kulik, A. S.; Beckham, H. W.; Ohlemacher, A.; MA800260K