Synthesis, Characterization, and Solid-State NMR Investigation of

Jun 20, 2013 - ... Margarita Darder , Bernd Wicklein , Giora Rytwo , Eduardo Ruiz-Hitzky ... Marco Carlotti , Giuseppa Gullo , Antonella Battisti , Fr...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/Langmuir

Synthesis, Characterization, and Solid-State NMR Investigation of Organically Modified Bentonites and Their Composites with LDPE Silvia Borsacchi,*,†,‡ Umayal Sudhakaran,† Marco Geppi,†,‡ Lucia Ricci,*,† Vincenzo Liuzzo,† and Giacomo Ruggeri†,§ †

Dipartimento di Chimica e Chimica Industriale, Università di Pisa, Via Risorgimento 35, 56126 Pisa, Italy INSTM Consorzio Interuniversitario Nazionale per la Scienza e Tecnologia dei Materiali, Via G. Giusti 9, 50121 Firenze, Italy § Istituto per i Processi Chimico-Fisici del Consiglio Nazionale delle Ricerche (IPCF-CNR), Via G. Moruzzi 1, I-56124 Pisa, Italy ‡

ABSTRACT: Polymer/clay nanocomposites show remarkably improved properties (mechanical properties, as well as decreased gas permeability and flammability, etc.) with respect to their microscale counterparts and pristine polymers. Due to the substantially apolar character of most of the organic polymers, natural occurring hydrophilic clays are modified into organophilic clays with consequent increase of the polymer/clay compatibility. Different strategies have been developed for the preparation of nanocomposites with improved properties, especially aimed at achieving the best dispersion of clay platelets in the polymer matrix. In this paper we present the preparation and characterization of polymer/clay nanocomposites composed of low-density polyethylene (LDPE) and natural clay, montmorillonite-containing bentonite. Two different forms of the clay have been considered: the first, a commercial organophilic bentonite (Nanofil 15), obtained by exchanging the natural cations with dimethyldioctadecylammonium (2C18) cations, and the second, obtained by performing a grafting reaction of an alkoxysilane containing a polymerizable group, 3-(trimethoxysilyl)propyl methacrylate (TSPM), onto Nanofil 15. Both the clays and LDPE/clay nanocomposites were characterized by thermal, FT-IR, and X-ray diffraction techniques. The samples were also investigated by means of 29Si, 13C, and 1H solid-state NMR, obtaining information on the structural properties of the modified clays. Moreover, by exploiting the effect of bentonite paramagnetic (Fe3+) ions on proton spin−lattice relaxation times (T1’s), useful information about the extent of the polymer−clay dispersion and their interfacial interactions could be obtained.



INTRODUCTION Polymer/clay nanocomposites are an important class of organic−inorganic composites in which clay platelets are dispersed at a nanometric level in the polymer matrix, acting as a reinforcing phase. With the addition of a very small amount of clay into the polymer matrix, if nanometric dispersion is achieved, the resulting nanocomposites exhibit substantial improvements in many physical properties, including mechanical properties (tensile modulus, flexural modulus, and strength), decreased thermal expansion coefficient, decreased gas permeability, increased swelling resistance, enhanced ionic conductivity, and flame resistance.1−5 Especially, composites of polymers with montmorillonite have received significant attention because of improvements in mechanical, thermal, electrical, and barrier properties.5−8 The occurrence of a large polymer/clay interfacial area, for which it is necessary to disperse the clay into the polymer matrix at a nanometric level, is a crucial factor for obtaining nanocomposites with the special properties mentioned above. Clays are naturally composed of hydrophilic platelets with nanometric thickness, stacked together with interplatelet galleries occupied by cations and water. In order to effectively disperse the clay into an organic polymer matrix, it is usually necessary to modify the hydrophilic character of the silicate into organophilic character, so as to © 2013 American Chemical Society

increase their mutual compatibility. This is often achieved by modifying clays by ion-exchanging their natural cations with long alkyl chains ammonium (or other headgroup) cations.8−11 Many strategies have been proposed for the preparation of polymer/clay nanocomposites, also based on polyolefins.8,12,13 Depending on different preparation conditions, various composite morphologies, sometimes coexisting, may be obtained: stacked clay platelets (tactoids) dispersed in the polymer matrix usually do not give rise to nanocomposites, while nanometric dispersions can be obtained when polymer chains intercalate within the clay interplatelet galleries or when the clay stacks are fully exfoliated throughout the polymer matrix.14 In order to promote a well-exfoliated morphology, the narrow spacing between the platelets of the clay usually has to be widened so to allow the polymer chains to penetrate within the clay platelets; moreover, it is generally necessary to functionalize both the silicate and the matrix.15 For instance, maleic anhydride grafted LLDPE (LLDPE-g-MAH) with a 0.1 wt % functionalization degree (FD) needs a quaternary ammonium salt substituted with aliphatic chains containing at Received: July 25, 2012 Revised: June 19, 2013 Published: June 20, 2013 9164

dx.doi.org/10.1021/la401686p | Langmuir 2013, 29, 9164−9172

Langmuir

Article

information about the structural and dynamic properties of nanocomposites on wide spatial (from few Å to hundreds of nm) and time (from ns to s) scales.10−13,19−22 Besides detailed structural information on the modified clay, interesting insights could be obtained onto the morphology of the polymer−clay composite, especially from the measurement and analysis of 1H spin−lattice relaxation times in the laboratory frame (T1). In bentonite, paramagnetic Fe3+ ions partly substitute Al3+ in the octahedral layer: electron−nuclear dipolar interactions become the dominant relaxation mechanisms for protons close enough to the paramagnetic centers, and the effect propagates to those farther through spin diffusion. This phenomenon, often with the support of other experimental data, has been successfully exploited for understanding the degree of dispersion of clay particles in organic matrices.13,23−29 Similarly, in this work we investigated the morphology of the composites of LDPE with either 2C18 exchanged or 2C18 exchanged and TSPM functionalized bentonite, also taking advantage of the study of analogous organically modified clays in which laponite, a synthetic smectite clay not containing paramagnetic impurities, was used in place of bentonite.18

least 18 carbon atoms, in order to ensure the formation of the exfoliated nanocomposite; shorter aliphatic chains require higher FDs. With unfunctionalized LLDPE only intercalated morphology can be obtained.16 This paper is part of a more general work concerning polyolefin−inorganic composites where the compatibility between organic and inorganic components is enhanced by directly grafting the fillers to the polymer matrix, in the perspective of obtaining materials with improved mechanical and barrier properties with respect to the pristine polymer. In particular, we attempted to develop a synthetic route for functionalizing micro- and nanosized inorganic fillers with polymerizable groups, then to disperse such modified fillers into polyolefins by melt mixing, and finally to promote their grafting onto the matrix by UV irradiation. In particular, we previously published the preparation and solid-state NMR investigation of composites based on LDPE and silica modified by reaction with an alkoxysilane containing a polymerizable group, 3-(trimethoxysilyl)propyl methacrylate (TSPM).17 In the present work, bentonite, a naturally occurring layered silicate of the smectite group, mainly containing montmorillonite, has been used for preparing nanocomposites with LDPE. The dimethyldioctadecylammonium (2C18) exchanged clay has been further modified with TSPM (Figure 1), the



EXPERIMENTAL SECTION

Materials. Bentonites are hydrated aluminosilicate minerals, composed chiefly of montmorillonite. A typical empirical formula for bentonite is (Al,Fe0.67Mg0.33)Si4O10(OH)2Na,Ca0.33. In this work, we have used a dimethyldioctadecylammonium (2C18) exchanged bentonite, called Nanofil 15 (N15), from Süd-Chemie. N15 was dried overnight at 180 °C to remove absorbed water before use. 3(Trimethoxysilyl)propyl methacrylate (TSPM, 98%, Aldrich) was used without any further purification. Photoinitiator (P1) was an alkylsubstituted α-hydroxy ketone (linear or branched alkyl with an average chain length of 12), supplied by Lamberti S.p.A.. Low-density polyethylene (LDPE, LD 158 JD from Exxon Chemical, with Mw 70.6 × 103; Mn 19.0 × 103; antioxidant Irganox 1076, 500 ± 150 ppm) was used as received, without removing the antioxidant. Maleicanhydride-functionalized low-density polyethylene (LDPE-g-MAH, Bynel 42E703 from Du Pont) was used as received. Preparations. Preparation of N15-TSPM. A 3 mL portion of 3(trimethoxysilyl)propyl methacrylate and 20 mL of an EtOH/H2O (95/5, V/V) mixture were stirred for half an hour at room temperature; then, 2 g of N15 was added. The resulting mixture was stirred at room temperature for 2 days, filtered, and washed with ethanol. The obtained solid was then extracted with CHCl3 in Soxhlet for 8 h (to remove unreacted monomer and polymerized TSPM) and then dried under vacuum for 16 h. The obtained solid was pulverized with a Retsch ball mill. Blend Preparation. Blends were prepared in a Brabender plastograph (model OHG 47055, with 30 mL mixing chamber) at 170 °C, 50 rpm, for 10 min. Polymer pellets were previously mixed together with the filler; the photoinitiator P1, when present, was added to the melt mixture. Blend compositions are summarized in Table 1.

Figure 1. Chemical structures and atom numbering of 2C18 (top) and TSPM (bottom).

methoxysilane groups of which, after hydrolysis, should react with silanols of the clay forming siloxane bonds (reaction already proved to be successful in the case of a different clay).18 The further organic modification and the photoinduced polymerization reaction of the TSPM methacrylate groups with LDPE have been exploited for preparing nanocomposites. With the aim of deeply characterizing the modified bentonite and the final composites, and in particular to obtain information about their structural properties and the nature and extent of interactions and interfaces between organic and inorganic components, we carried out thermogravimetric, FTIR, and X-ray diffraction analyses and performed several multinuclear (29Si, 13C, 1H) high- and low-resolution solidstate NMR experiments. In recent years, many characterization techniques have been exploited to unveil the structural and dynamic properties and also the nature of the interactions occurring at the interface between the organic and inorganic components.19 Among the different spectroscopic techniques, solid-state NMR is one of the most powerful, providing

Table 1. Blend Compositions (wt %) blend

LDPE

P1

B15 B16

85 84

1

N15

N15-TSPM

LDPE-g-MAH

5

10 10

5

Thin films (100−150 μm thick) were prepared by compression molding at 180 °C and 20−25 bar (with press “Campana” 20/200). UV Irradiation. UV irradiation was carried out on B16 thin films, inside a drybox chamber (O2 < 1 ppm; H2O < 2 ppm). The UV source was a conventional medium pressure mercury lamp (model HG100, 125 W) with an energy output of 5000 ± 10 μW/cm2 at 365 nm and of 3800 ± 10 μW/cm2 at 254 nm. After the irradiation, films were kept 9165

dx.doi.org/10.1021/la401686p | Langmuir 2013, 29, 9164−9172

Langmuir

Article

in the dark inside the drybox for 72 h in order to let radicals decay, and then they were drawn from the chamber still keeping them in the dark and in inert atmosphere using a desiccator. Sample Characterization. FT-IR spectra were recorded with a Perkin-Elmer Spectrum One spectrometer, inorganic fillers were finely ground together with KBr and pressed, and polymer blends were examined as thin films. X-ray diffraction analysis (XRD) was made with a Siemens diffractometer D 500/501, model Kristalloflex 810, with X-ray source Cu Kα (λ = 1.5406 Å), 2θ scanning from 1.5° to 30°, both on bentonite powder and on polymer films. Thermogravimetric analysis (TGA) on inorganic fillers was made with a Mettler Toledo TGA/SDTA 851 calorimeter from 25 to 700 °C at a scanning rate of 10 °C/min. Polymer blends were extracted with boiling xylene for 14 h in a Kumagawa extractor. Solid-State NMR. All the experiments were performed on a doublechannel Varian Infinity Plus 400 spectrometer working at 399.88, 79.44, and 100.56 MHz for 1H, 29Si, and 13C, respectively. The 13C and 29 Si high-resolution measurements were carried out on a 7.5 mm crosspolarization and magic angle spinning (CP-MAS) probe head. The 90° pulse length was 4.5 μs for 1H and 29Si, and 4 μs for 13C. In order to acquire a quantitative 29Si direct excitation (DE-MAS) spectrum, a recycle delay of 120 s was used and 2000 transients were accumulated, while nonquantitative DE-MAS spectra were acquired with a 5 s recycle delay and accumulating 8000 transients. 13C DE-MAS spectra were acquired using the depth pulse sequence for removing probe head background signal30 and a recycle delay of 2 s; 27 000 transients were accumulated. A MAS frequency of 6 kHz was used for all the spectra. The experiments for measuring proton T1 relaxation times were performed on a static probe-head. The 1H 90° pulse length was 3.3 μs. The 1H T1 relaxation time measurements were carried out in lowresolution (static) conditions by using a saturation−recovery pulse sequence with delays ranging from 0.01 ms to 8 s. For all the samples 32 scans were accumulated. Due to the presence of the paramagnetic species in the samples, which shorten the 1H T2 and thereby cause the fast decay of the FID, we combined the saturation recovery sequence with the solid−echo pulse sequence, in order to minimize the loss of signal at the initial part of the FID, using an echo delay of 10 μs. For the measurements on the clay systems (N15 and N15-TSPM), relatively poor of protons, in order to eliminate the probe head proton background signal, we applied the method proposed by Schmidt-Rohr and co-workers.31

Figure 2. FT-IR spectra of N15 (straight line) and N15-TSPM (dotted line).

Figure 3. FT-IR spectral region of hydroxyl groups of N15 (straight line) and N15-TSPM (dotted line).



spectra), ascribable to not hydrogen-bonded and hydrogenbonded bentonite OH groups, respectively, and the shoulder at 3199 cm−1 (detectable only in the spectrum of N15-TSPM), due to the stretching of hydrogen-bonded TSPM hydroxyl groups. It must be noticed that, in passing from N15 to N15TSPM, the intensity of the band at 3432 cm−1 decreases, that can be ascribed to either the involvement of a fraction of hydrogen-bonded bentonite hydroxyl groups in the reaction with TSPM or the removal of absorbed water. To estimate the amount of TSPM in the modified filler, TGA analyses were carried out on the filler before and after functionalization (Figure 4). Thermogram of N15-TSPM shows an increased weight loss percentage with respect to the pristine clay from which it is possible to estimate a TSPM amount of 1.6 wt %. A thermal stabilization (Tonset = 296.1 °C for N15-TSPM and 237.9 °C for N15) is also observed, which can be ascribed to the formation of covalent bonds between TSPM and filler.33 XRD analysis (Figure 5) indicates an interlayer distance d001 in N15-TSPM of 2.26 nm (2θ = 3.90°), shorter than that of the pristine bentonite (N15, 2.85 nm, 2θ = 3.10°). The reduction of the clay spacing after grafting with TSPM has been already observed for laponite clay and can be ascribed to the linking between different clay platelets through TSPM molecules or oligomers reacted with silanol groups at the edges of the clay particles.18

RESULTS AND DISCUSSION Bentonite Functionalization and Characterization. The functionalization of bentonite N15 was carried out through direct reaction with TSPM in EtOH/H2O solvent mixture. The mixture EtOH/H2O in neutral conditions allows the hydrolysis of alkoxysilanes to proceed, even if slowly, but the formation of siloxane is not favored.32 Therefore, we decided to use the neutral conditions for our preparation in order to hinder the self-aggregation of TSPM molecules to form, for instance, TSPM oligomers with a large molecular size that could inhibit the intercalation between clay platelets. A comparison between the FT-IR spectra of N15 and N15TSPM is reported in Figures 2 and 3. The presence of TSPM was verified by the occurrence of the signal at 1709 cm−1 in the spectrum of N15-TSPM, due to the CO stretching of the TSPM methacrylic group with the formation of hydrogen bonds between TSPM and clay Si−OH groups (in pure TSPM this stretching is observed at 1720 cm−1). Moreover, the absence of the CO stretching at 1735 cm−1 can be interpreted as a sign of the absence of polymerized TSPM (N15-TSPM was, in fact, extracted with CHCl3 in Soxhlet). By analyzing the hydroxyl spectral region of both fillers (see Figure 3), it is possible to identify three distinct O−H stretching bands: those at 3633 and 3432 cm−1 (present in both 9166

dx.doi.org/10.1021/la401686p | Langmuir 2013, 29, 9164−9172

Langmuir

Article

Figure 6. 29Si DE-MAS spectra recorded with a recycle delay of 5 s and a MAS frequency of 6 kHz of (a) N15 and (b) N15-TSPM. Spinning sidebands are marked with asterisks.

Figure 4. TGA curves of N15 and N15-TSPM.

change much in passing from N15 to N15-TSPM, the comparison between the two spectra clearly shows an increased Q4/Q3 ratio in N15-TSPM, confirming the occurrence of a reaction between TSPM and clay silanols. In the case of N15-TSPM we also recorded a quantitative 29 Si DE-MAS spectrum, which is reported in Figure 7. The

Figure 5. XRD analysis of N15 (black line) and N15-TSPM (gray line).

Solid-State NMR. A structural characterization of N15 and N15-TSPM has been carried out by recording 29Si and 13C high-resolution solid-state NMR spectra. In this case, due to the presence of Fe3+ paramagnetic ions in the clay, the use of the CP technique is not convenient.34 Indeed the interaction between the paramagnetic Fe3+ electron spins and the nuclear spins affects the nuclear relaxation times, in particular strongly shortening the 1H spin−lattice relaxation times in the rotating frame, thus dramatically reducing CP efficiency. By recording 29 Si and 13C CP spectra of N15 and N15-TSPM (here not reported), we could in fact observe a large loss in sensitivity. Therefore, 29Si and 13C spectra have been acquired through direct excitation (DE). The 29Si DE-MAS spectra of N15 and N15-TSPM, acquired with a short recycle delay of 5 s, are reported in Figure 6. In both of them two intense signals are present at about −110 and −95 ppm, which must be ascribed to fully condensed Q4 Si(OAl)(OSi)3 and Q3 Si(OAl)(OSi)2OH silicon nuclei, respectively, both present in the tetrahedral silica layers of the clay.35 In the spectrum of N15-TSPM a weak and broad signal centered at about −64 ppm can also be recognized and assigned to condensed T silicon nuclei of TSPM, which seem to be mainly T3 (Si(OSi)3R, δ ≈ −65 ppm) and, in part, T2 (Si(OSi)2(OX)R with XH or CH3, δ ≈ −56 ppm), so confirming the occurrence of a condensation reaction for TSPM. Even if these spectra are not quantitative due to the short recycle delay used (5 s), by assuming that 29Si T1’s do not

Figure 7. 29Si quantitative DE-MAS spectrum of N15-TSPM, acquired with a recycle delay of 120 s and at a MAS frequency of 6 kHz. In the inset the region of T silicon signals is highlighted. Spinning sidebands are marked with asterisks.

relative intensities of the peaks are quite different from those observed in the previously shown nonquantitative spectrum, suggesting that especially Q4 silicon nuclei have a quite long T1, as is expected considering the lack of close protons. By measuring the signal areas, through a proper spectral deconvolution procedure (also taking into account the spinning sidebands) we could estimate the relative amounts of the different silicon nuclei that resulted to be the following: 38% Q4, 60% Q3, and 2% T. Calculating a scaling factor for the Q4/ Q3 ratio in passing from the nonquantitative to the quantitative spectrum of N15-TSPM, and applying the same factor to the Q4/Q3 ratio measured from the nonquantitative spectrum of N15, it has been possible to estimate the relative amounts of Q4 and Q3 in N15, finding 33% and 67%, respectively. By 9167

dx.doi.org/10.1021/la401686p | Langmuir 2013, 29, 9164−9172

Langmuir

Article

comparing these data with those of N15-TSPM it is possible to estimate that about 6% of Q3 silicon nuclei of N15 were transformed into Q4 sites. The amount of TSPM reacted silicon nuclei is about 2% (of all the silicons of the sample), that, considering the presence of only T2 and T3 sites, corresponds to a 4−6% of TSPM reacted silanols, which must be compared with about 6% of reacted silica silanols. Considering the likely occurrence of self-condensation reactions among TSPM molecules,18 these data suggest that some clay silanols might have undergone a self-condensation process, as was previously observed for a different clay treated with the same process.18 The 13C DE-MAS spectra of N15 and N15-TSPM are shown in Figure 8. The two spectra are almost coincident, with the

Figure 8. 13C DE-MAS spectra of (a) N15 and (b) N15-TSPM, acquired at a MAS frequency of 6 kHz. Spinning sidebands are marked with asterisks, and the signal at about 110 ppm is due to incomplete probe head background signal suppression.

Figure 9. (A) Comparison of the carbonyl FT-IR spectral region of B16 before (straight line) and after (dashed line) irradiation. (B) Deconvolution of the carbonyl FT-IR spectral region of B16 after irradiation.

most intense signals being those of the quaternary ammonium cations. In particular, signals are observed at 15, 24, 31, and 52 ppm, which can be assigned to C19, C18, C3−C17, and C1 (see Figure 1 for atom numbering) of 2C18; the signal of C2, expected at about 65−70 ppm,18 could not be resolved because of its scarce intensity and line-broadening due to the proximity of nitrogen. The spectral resolution is relatively poor, which can be ascribed to a distribution of different situations experienced by the cations and/or to a shortening of 13C T2 due to the presence of the paramagnetic centers. In particular, the signal ascribed to C3−C17 appears quite broad and centered at 31 ppm. Even if TSPM signals are quite small and mostly overlapped with the cation signals, some peaks can be recognized in the spectrum of N15-TSPM, in particular at about 18 ppm (C7), 137 ppm (C6), and 168 ppm (C5) (see Figure 1 for atom numbering). Characterization of Functionalized Bentonite/LDPE Blends. In order to induce the reaction between N15-TSPM and LDPE, B16 film was irradiated with UV lamp. By comparing the FT-IR spectra of B16 before and after UV irradiation (Figure 9A), it is possible to get insights about photoinitiator and methacrylate conversions (we can see the decrease of absorption due to photoinitiator and the increase of that ascribable to polymerized TSPM). From the FT-IR spectrum of irradiated B16 (Figure 9B), four partially superimposed bands at about 1674 cm−1 (unreacted photoinitiator), 1689 cm−1 (photoinitiator byproducts), 1709 cm−1

(CO stretching of nonreacted TSPM), and 1734 cm−1 (C O stretching of reacted TSPM) are detected. The intensity ratio between the carbonyl stretching bands of reacted and unreacted TSPM increases from 0.17 (before irradiation) to 0.37 (after irradiation), similarly to what previously observed for analogous silica/LDPE blends.17 From extraction experiments 10 wt % insoluble residue in B15 and 28 wt % in B16 were detected (Table 2), suggesting in Table 2. Properties of Bentonite-Filled Blends blend

UV treatment

insoluble residue (wt %)

LDPE/LDPE-g-MAH 90/10 B15 B16

no no yes

0 10 28

both cases the occurrence of interactions between LDPE and bentonite, possibly due to polymer intercalation between clay platelets. Furthermore, in the case of B16, the much larger residue fraction could be in agreement with the establishment of chemical bonds between functionalized filler and polymer, not occurring in the case of B15. A comparison between X-ray diffractograms of B15 and B16 (before and after the UV treatment) is reported in Figure 10. From the analysis of the diffractogram of B15 an interlamellar distance d001 of 3.21 nm could be determined, higher 9168

dx.doi.org/10.1021/la401686p | Langmuir 2013, 29, 9164−9172

Langmuir

Article

but some considerations can still be drawn from the experimental relaxation times. The 1H T1 values determined for our samples using a monoexponential model, which reproduces quite well the experimental magnetization recovery curves, are reported in Table 3. Table 3. 1H T1 Relaxation Times of the Samples Investigated As Obtained from a Monoexponential Fitting of the Magnetization Recovery Curvesa

Figure 10. XRD diffractograms of sample B15 (black line), B16 before UV irradiation (gray line), and B16 after UV irradiation (light gray line).

than that of the pure organophilic clay N15 (2.85 nm), suggesting the occurrence of polymer intercalation. It can be clearly seen that the intensity of the diffraction peak in B16 (before UV irradiation) is much lower than that of B15. The peak is broadened and an interlayer distance of 3.39 nm, slightly larger than that of B15 and much larger than that of N15-TSPM, can be estimated from the position of the maximum. These data suggest a strong disordering of the clay arrangement, an average increase of the interplatelet distance, compatible with polymer intercalation, and, possibly, the exfoliation of a fraction of the clay. It can be hypothesized that these effects take place because, during melt processing of B16, the reaction of polymeric macroradicals, produced by the shear stress, promotes the grafting of N15-TSPM to the polyolefin. In the diffractogram of B16 after UV irradiation a further decrease of the relevant peak is observed. The UV irradiation promotes the formation of more radicals, deriving from the decomposition of the initiator, and consequently further increasing the above-described effects. Solid-State NMR 1H T1. In order to better investigate LDPE/ clay interactions and their relative dispersion, proton T1’s were measured and analyzed. It is well-known that 1H spin−lattice relaxation times in the laboratory frame (T1) can be exploited, in solid heterogeneous systems, either for investigating the miscibility among the heterodomains on a scale of about 100 Å, and/or to get information about the dynamics in the MHz− GHz frequency range. However, when paramagnetic centers are present, 1H T1’s are strongly shortened by the interaction between nuclear spins and unpaired electrons, and they can be peculiarly used to obtain information about the morphology of the material. Vanderhart et al. have shown how, when the paramagnetic centers are contained, as in our case, within clay layers, information such as the degree of exfoliation of the clay and its homogeneity can be derived. This requires a very accurate and complex analysis of relaxation times, combined with detailed data available from other techniques concerning the compositional and structural properties of the material.28 The analysis of 1H T1’s to derive morphological information can be very complex, since the experimental relaxation time values may be determined by a combination of many processes and properties, including motions in the fast regime, proton spin diffusion, distribution and concentration of paramagnetic centers, and so on. In our case the detailed analysis proposed by Vanderhart et al. is prevented from the lack of non-NMR data and the unavailability of a sample with known high exfoliation,

sample

T1 (ms)

N15 N15-TSPM Laponite-2C18 Laponite-TSPM LDPE B15 B16

11 6 164 209 736 670 658

a

Samples Laponite-2C18 and Laponite-TSPM, described in ref 18, are very similar to N15 and N15-TSPM, respectively, with the difference that bentonite is replaced by laponite, free from paramagnetic impurities. The uncertainties in T1 values are less than 1%.

In N15 1H T1 is very short, since all the protons (silanols and protons of the organic modifier 2C18) are very close to the surface of the clay layers, and therefore they are strongly affected by the interaction with the unpaired electron of Fe3+ species. The measured value is slightly higher than that (2.5 ms) predicted by Bourbigot et al. for the “intrinsic” relaxation time assigned to a 0.4 nm thick interfacial region.28 This is compatible with the interlayer distance of 2.85 nm obtained by XRD for N15, which, considering the typical platelet thickness of 1 nm, would correspond to a 0.9 nm gallery half-spacing. The shortening in the relaxation time observed in passing from N15 to N15-TSPM is in agreement with the reduction of the interlayer distance (from 2.85 to 2.26 nm) observed by XRD. The relaxation time measured for LDPE, 736 ms, is almost 2 orders of magnitude longer than that of N15 and N15-TSPM, and it only arises from the polymer dynamic behavior. The much longer relaxation times measured for samples with laponite, a synthetic clay similar to bentonite but free from paramagnetic impurities, confirm that the interaction with unpaired electrons of Fe3+ impurities, present within bentonite layers, is the dominant relaxation mechanism. In the cases of B15 and B16 clay/LDPE composites, T1’s of 670 and 658 ms were measured, respectively. These values are very similar and not much smaller than the value measured for pure LDPE, indicating that the shortening effect of the paramagnetic Fe3+ centers of the clay on polymer proton T1 is not very effective in the composites. In particular, on the basis of composite composition (95 wt % of polymer, with a density of approximately 1 g/cm3 and 5 wt % of clay, with a density of approximately 2 g/cm3 and a platelet thickness of about 1 nm), the volume percentages occupied by the polymer and the clay are 97.4% and 2.6%, respectively. Following Bourbigot et al.,28 we can consider a thickness of 0.4 nm for the fast-relaxing polymer layer in close contact with the clay; this polymer fraction approximately corresponds to 2.1% of the whole sample volume. Considering only the protons of the polymer, and their homogeneous density throughout the sample, 2.1% of polymeric protons are in the fast relaxing (intrinsic T1 ≈ 5 ms) layers close to the clay, and 97.9% of polymeric protons are in 9169

dx.doi.org/10.1021/la401686p | Langmuir 2013, 29, 9164−9172

Langmuir

Article

the “bulk” (intrinsic T1 = 736 ms). It is therefore possible to estimate that a complete averaging effect of spin diffusion would imply an experimental T1 of about 180 ms (inverse of the population weighted rate average). This is a limit case that would be achieved only in the case of infinitely fast spin diffusion. In the real case, considering a diffusion coefficient of the order of 0.5 nm2/ms, in the presence of a regular full exfoliation (which, considering the relative volumes occupied by clay and polymer, would correspond to gallery spacings of about 40 nm) we should rather expect a nonexponential relaxation with about one-half of all the protons having a relaxation time shorter than 180 ms, and the other longer. The experimental results obtained appear very different from this situation, ruling out the possibility of full exfoliation of the clay in both B15 and B16 composites. On the other hand, if the polymer is intercalated into the clay galleries, then only a small fraction of the polymer is found there. While protons within these galleries relax very quickly, that relaxation is not easily spread into the major portion of the polymer lying outside the intercalated “particles” because (a) average distance between these “particles” is now significantly larger than distances that can be covered by spin diffusion in a time of 1H T1 and (b) paramagnetically induced proton relaxation can only escape from the “particles” at the edges of the platelets. With platelet diameters of about 150 nm, this lateral surface area of the “particles” is very small relative to the surface area of the fully exfoliated clay. In the extreme case of complete absence of averaging, a biexponential relaxation, with a very short and a long T1, with weight of approximately 2.1% and 97.9%, respectively, should be observed. Indeed, by performing a more careful analysis, it was possible to observe that a biexponential model describes the experimental relaxation curves of both B15 and B16 slightly better than a monoexponential one. A component with a short T1 and a weight of about 1% could be revealed. It must be said that this weight value is likely to be underestimated due to the experimental difficulties in recording the very first microseconds of the FIDs to which the protons with the shortest T1 are expected to mostly contribute. In summary, these T1 results, together with those of X-ray diffraction analyses, point to an intercalated morphology of both B15 and B16 nanocomposites. This could be further confirmed by a careful analysis of the first part of the relaxation curves. In particular, in the range of delays (t) 0−50 ms, it was possible to detect, as expected in the presence of paramagnetic centers,28 a deviation from the biexponential trend. Following the model proposed by Bourbigot, Vanderhart et al.,28 the magnetization recovery curves, to which the pure LDPE curve had been subtracted, were plotted as a function of √t in the range 0.07−0.22 s1/2 (Figure 11). For both B15 and B16, at least until 0.15 s1/2, approximately linear trends are observed, as expected; both curves deviate from linearity, exhibiting a downward curvature, at √t > 0.15 s1/2. From simulated and experimental data reported in ref 28, this has been ascribed by the authors to a substantial nonexfoliation of the clay, approximately corresponding to average clay platelet distances of the order of 4 nm. It is also possible to observe that, in the linear range, the curve of B16 has a slightly larger slope than that of B15, which should correspond to a slightly larger average interplatelet distance, in agreement with the d001 values obtained from X-ray diffraction analysis.

Figure 11. First part of 1H magnetization recovery curves of B15 and B16, to which the recovery curve of pure LDPE has been subtracted, plotted as a function of the square root of the delay of the saturation recovery pulse sequence.



CONCLUSIONS In this work two different polymer/clay nanocomposites have been successfully prepared and characterized. The first is composed of low-density polyethylene (LDPE) and the commercial organophilic bentonite N15, obtained by exchanging the natural cations with dimethyldioctadecylammonium (2C18) cations. The second is composed of LDPE and the functionalized clay N15-TSPM, obtained by grafting onto N15 the alkoxysilane 3-(trimethoxysilyl)propyl methacrylate (TSPM), able to polymerize with LDPE through an UV photoinduced reaction, which should improve the dispersion of the filler in the polymer matrix. From TGA, FT-IR, X-ray diffraction, and solid-state NMR experiments it has been possible to confirm the successful grafting of TSPM onto bentonite (with a 1.6 wt % yield), mainly occurring at the clay edges, and to obtain detailed quantitative information on the different kinds of silicon sites present. N15-TSPM was successfully used for the preparation of the nanocomposite with LDPE: as clearly indicated by FT-IR and extraction experiments, the polymeric macroradicals were produced by shear stress during melt mixing, and the subsequent UV irradiation promotes the reaction between TSPM and LDPE. The combination of X-ray diffraction analysis and measurement of proton spin−lattice relaxation times allowed the morphology of the LDPE nanocomposites with N15 and N15TSPM to be investigated. Even if the samples are likely to be quite inhomogeneous, it is possible to conclude that in both composites the polymer is intercalated between clay platelets. The tethering of clay platelets together, performed by TSPM, if on one side opposes, as expected, the exfoliation of the clay, on the other side induces an increased disorder of the clay platelets stacking and an average increase of the interplatelet distance. These results are pictorially summarized in Figure 12, showing the schematic structures of the clays N15 and N15TSPM before and after their melt blending with LDPE and UV irradiation. We believe that this work confirms the importance of combining different characterization techniques, especially X9170

dx.doi.org/10.1021/la401686p | Langmuir 2013, 29, 9164−9172

Langmuir

Article

Figure 12. Scheme of the structures of N15 and N15-TSPM before and after melt blending with LDPE and the effect of UV irradiation on B16. (4) Olejniczak, S.; Kazmierski, S.; Pallathadka, P. K.; Potrzebowski, M. A review on advances of high-resolution solid state NMR spectroscopy in structural studies of polymer/clay nanocomposites. Polimery 2007, 52 (10), 713−721. (5) Schmidt, D.; Shah, D.; Giannelis, E. P. New advances in polymer/ layered silicate nanocomposites. Curr. Opin. Solid State Mater. Sci. 2002, 6, 205−212. (6) Ruiz-Hitzky, E.; Van Meerbeek, A. Handbook of Clay Science; Bergaya, F., Theng, B. K. G., Lagaly, G., Eds.; Developments in Clay Science; Elsevier Ltd.: New York, 2006; Vol. 1. (7) Stoy, W. S.; Washabaugm, F. J. Encyclopedia of Polymer Science and Engineering, 2nd ed.; Wiley: New York, 1986; Vol. 7, pp 53−73. (8) Arroyo, M.; Suarez, R. V.; Herrero, B.; Lopez-Manchado, M. A. Optimization of nanocomposites based on polypropylene/polyethylene blends and organo-bentonite. J. Mater. Chem. 2003, 13, 2915−2921. (9) Zang, Y.; Xu, W.; Qiu, D.; Chen, D.; Chen, R.; Su, S. Synthesis, characterization and thermal stability of different polystyryl quaternary ammonium surfactants and their montmorillonite complexes. Thermochim. Acta 2008, 474, 1−7. (10) Urbanczyk, L.; Hrobarikova, J.; Calberg, C.; Jerome, R.; Grandjean, J. Motional heterogeneity of intercalated species in modified clays and poly(ε-caprolactone)/clay nanocomposites. Langmuir 2006, 22, 4818−4824. (11) Mirau, P. A.; Serres, J. L.; Jacobs, D.; Garrett, P. H.; Vaia, R. A. Structure and dynamics of surfactant interfaces in organically modified clays. J. Phys. Chem. B 2008, 112, 10544−10551. (12) Lorthioir, C.; Laupretre, F.; Soulestin, J.; Lefebvre, J. M. Segmental dynamics of poly(ethylene oxide) chains in a model polymer/clay intercalated phase: solid-state NMR investigation. Macromolecules 2009, 42, 218−230. (13) Xu, B.; Leisen, J.; Beckham, H. W.; Zurayk, R. A.; Jones, E. H.; McNally, T. Evolution of clay morphology in polypropylene/ montmorillonite nanocomposites upon equibiaxial stretching: A solid-state NMR and TEM approach. Macromolecules 2009, 42, 8959−8968.

ray diffraction and the study of proton T1 by means of solidstate NMR, to gain insights into the structural features at the nanometric level of composites between polymers and clays containing paramagnetic centers.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] (S.B.); [email protected] (L.R.). Phone: +39-050-2219289 (S.B.); +39-050-2219447 (L.R.). Fax: +39-050-2219260 (S.B.); +39-050-2219260 (L.R.). Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors would like to thank Dr. F. Fallani and Mr. P. Narducci for preparation and XRD analyses of the samples. Fondazione Cassa di Risparmio di Pisa (POLOPTEL project) is acknowledged for partial financial support. S.B. wants to acknowledge INSTM for cofunding her postdoc position.



REFERENCES

(1) Ray, S. S.; Okamoto, M. Polymer/layered silicate nanocomposites: a review from preparation to processing. Prog. Polym. Sci. 2003, 28, 1539−1641. (2) Grandjean, J. Solid-state NMR study of modified clays and polymer/clay nanocomposites. Clay Mater. 2006, 41 (2), 567−586. (3) Choi, Y. S.; Ham, H. T.; Chung, I. J. Effect of monomers on the basal spacing of sodium montmorillonite and the structures of polymer-clay nanocomposites. Chem. Mater. 2004, 16, 2522−2529. 9171

dx.doi.org/10.1021/la401686p | Langmuir 2013, 29, 9164−9172

Langmuir

Article

unsaturated polyester nanocomposites. Polym. Compos. 2010, 31 (3), 482−490. (34) Forte, C.; Geppi, M.; Giamberini, S.; Ruggeri, G.; Veracini, C. A.; Mendez, B. Structure determination of clay/methyl methacrylate copolymer interlayer complexes by means of 13C solid state NMR. Polymer 1998, 39 (12), 2651−2656. (35) Rhee, C. H.; Kim, H. K.; Chang, H.; Lee, J. S. Nafion/sulfonated montmorillonite composite: A new concept electrolyte membrane for direct methanol fuel cells. Chem. Mater. 2005, 17, 1691−1697.

(14) Alexandre, M.; Dubois, P. Polymer-layered silicate nanocomposites: preparation, properties and uses of a new class of materials. Mater. Sci. Eng. 2000, R28, 1−63. (15) Kawasumi, M.; Hasegawa, N.; Kato, M.; Usuki, A.; Okada, A. Preparation and mechanical properties of polypropylene-clay hybrids. Macromolecules 1997, 30, 6333−6338. (16) Wang, K. H.; Choi, M. H.; Koo, C. M.; Choi, Y. S.; Chung, I. J. Synthesis and characterization of maleated polyethylene/clay nanocomposites. Polymer 2001, 42 (24), 9819−9826. (17) Borsacchi, S.; Geppi, M.; Veracini, C. A.; Fallani, F.; Ricci, L.; Ruggeri, G. Improving compatibility in LDPE-silica dispersions by photo-grafting reaction. Preparation and solid state NMR investigation. J. Mater. Chem. 2006, 16, 4581−4591. (18) Borsacchi, S.; Geppi, M.; Ricci, L.; Ruggeri, G.; Veracini, C. A. Interactions at the surface of organophilic-modified laponites: A multinuclear solid-state NMR study. Langmuir 2007, 23, 3953−3960. (19) Samyn, F.; Bourbigot, S.; Jama, C.; Bellayer, S.; Nazare, S.; Hull, R.; Castrovinci, A.; Fina, A.; Camino, G. Crossed characterisation of polymer-layered silicate (PLS) nanocomposite morphology: TEM, Xray diffraction, rheology and solid-state nuclear magnetic resonance measurements. Eur. Polym. J. 2008, 44, 1642−1653. (20) Brus, J.; Urbanova, M.; Kelnar, I.; Kotek, J. A solid-state NMR study of structure and segmental dynamics of semicrystalline elastomer-toughened nanocomposites. Macromolecules 2006, 39, 5400−5409. (21) Geppi, M.; Borsacchi, S.; Mollica, G.; Veracini, C. A. Applications of solid-state NMR to the study of organic/inorganic multicomponent materials. Appl. Spectrosc. Rev. 2009, 44, 1−89. (22) Geppi, M.; Borsacchi, S.; Mollica, G. Solid-state NMR of organic/inorganic multi-component materials. In Encyclopedia of Magnetic Resonance; Harris, R. K., Wasylishen, R. E., Eds.; John Wiley & Sons, Ltd.: New York, 2008. (23) Asano, A.; Shimizu, M.; Kurotsu, T. Effect of paramagnetic Fe3+ on T1H in PVA/montmorillonite-clay nanocomposites. Chem. Lett. 2004, 33 (7), 816−817. (24) Vanderhart, D. L.; Asano, A.; Gilman, J. W. Solid-state NMR investigation of paramagnetic nylon-6 clay nanocomposites. 2. Measurement of clay dispersion, crystal stratification, and stability of organic modifiers. Chem. Mater. 2001, 13, 3796−3809. (25) Vanderhart, D. L.; Asano, A.; Gilman, J. W. Solid-state NMR investigation of paramagnetic nylon-6 clay nanocomposites. 1. Crystallinity, morphology, and the direct influence of Fe3+ on nuclear spins. Chem. Mater. 2001, 13, 3781−3795. (26) Vanderhart, D. L.; Asano, A.; Gilman, J. W. NMR measurements related to clay-dispersion quality and organic-modifier stability in nylon-6/clay nanocomposites. Macromolecules 2001, 34, 3819−3822. (27) Bourbigot, S.; Vanderhart, D. L.; Gilman, J. W.; Bellayer, S.; Stretz, H.; Paul, D. R. Solid state NMR characterization and flammability of styrene-acrylonitrile copolymer montmorillonite nanocomposite. Polymer 2004, 45, 7627−7638. (28) Bourbigot, S.; Vanderhart, D. L.; Gilman, J. W.; Awad, W. H.; Davis, R. D.; Morgan, A. B.; Wilkie, C. A. Investigation of nanodispersion in polystyrene-montmorillonite nanocomposites by solid-state NMR. J. Polym. Sci., Part B: Polym. Phys. 2003, 41, 3188− 3213. (29) Jung, A.; Peter, K.; Demco, D. E.; Jehnichen, D.; Moeller, M. Nanocomposites based on layered silicates and epoxy resins: measurement of clay dispersion and exfoliation using TEM, solidstate NMR, and X-ray diffraction methods. Macromol. Chem. Phys. 2012, 213, 389−400. (30) Cory, D. G.; Ritchey, W. M. Suppression of signals from the probe in bloch decay spectra. J. Magn. Reson. 1988, 80, 128−132. (31) Chen, Q.; Hou, S. S.; Schmidt-Rohr, K. A simple scheme for probehead background suppression in one-pulse 1H NMR. Solid State Nucl. Magn. Reson. 2004, 26, 11−15. (32) Iler, R. K. The Chemistry of Silica: Solubility, Polymerization, Colloid and Surface Properties and Biochemistry; Wiley: New York, 1979. (33) Şen, S. Effect of both silane-grafted and ion-exchanged organophilic clay in structural, thermal, and mechanical properties of 9172

dx.doi.org/10.1021/la401686p | Langmuir 2013, 29, 9164−9172