bm8002946

3004

Biomacromolecules 2008, 9, 3004–3013

Morphological and Thermal Properties of Cellulose-Montmorillonite Nanocomposites Pierfrancesco Cerruti,* Veronica Ambrogi, Alessandro Postiglione, Jozef Rychly´, Lyda Matisova´-Rychla´, and Cosimo Carfagna Institute of Polymer Chemistry and Technology-CNR, Via Campi Flegrei 34, 80078 Pozzuoli (Na), Italy, Department of Materials and Production Engineering, University of Naples Federico II, Piazzale Tecchio 80, 80125 Naples, Italy, and Polymer Institute, Slovak Academy of Sciences, Du´bravska´ Cesta 9, SK-84236 Bratislava, Slovakia Received March 21, 2008; Revised Manuscript Received September 8, 2008

Cellulose-layered montmorillonite (MMT) nanocomposites were prepared by precipitation from N-methylmorpholine-N-oxide (NMMO)/water solutions. Two hybrid samples were obtained to investigate the influence of the reaction time on the extent of clay dispersion within the matrix. It was observed that longer contact times are needed to yield nanocomposites with a partially exfoliated morphology. The thermal and thermal oxidative properties of the hybrids, which might be of interest for fire-resistant final products, were investigated by thermogravimetry and chemiluminescence (CL). The nanocomposites exhibited increased degradation temperatures compared to plain cellulose, and the partially exfoliated sample showed the maximum stability. This result was explained in terms of hindered transfer of heat, oxygen, and degraded volatiles due to the homogeneously dispersed clay filler. Kinetic analysis of the decomposition process showed that the degradation of regenerated cellulose and cellulosebased hybrids occurred through a multistep mechanism. Moreover, the presence of nanoclay led to drastic changes in the dependence of the activation energy on the degree of degradation. CL analysis showed that longer permanence in NMMO/water solutions brought about the formation of carbonyl compounds on the polymer backbone. Moreover, MMT increased the rate of dehydration and oxidation of cellulose functional moieties. As a consequence, cellulose was found to be less stable at temperatures lower than 100 °C. Conversely, at higher temperatures, the hindering of oxygen transfer prevailed, determining an increase in thermo-oxidative stability.

Introduction Cellulose is one the most abundant natural materials, accounting for roughly 50% of plant biomass, and it is currently one of the most promising polymeric resources, being a component of paper, textile, membranes, artificial fibers, etc. Furthermore, it is a raw material which can be readily modified by several chemical reactions rendering it potentially useful for a variety of applications. Due to its structure, cellulose cannot be processed as a conventional thermoplastic polymer, and it has to be regenerated from solution prior to use. However, regeneration is hard to achieve, because of intra- and intermolecular hydrogen bonds, which hinder the dissolution of cellulose. The first and widely employed methods to dissolve cellulose were Cu(OH)2 in aqueous ammonia,1 and the viscose process in alkaline solutions by the use of carbon disulphide. The latter was rapidly commercialized due to its low costs, and it is still employed for the production of about the 80% of artificial textile fibers.2 More recently, several organic systems, such as NaOH/urea aqueous solutions, liquid ammonia/ammonium/thiocyanate, LiCl/1,3-dimethyl-2-imidazolidinone (DMI), LiCl/N,N-dimethylacetamide (DMAc), N-methylmorpholine-N-oxide (NMMO)/ water, have also been investigated for regenerated cellulose fiber production.3-5 NaOH/urea/water is one of the cheapest cellulose solvents and has a substantial low environmental impact compared to other dissolving systems. However, a low percent* To whom correspondence should be addressed. Tel.: +39-081-8675214. Fax: +39-081-8675230. E-mail: [email protected].

age of cellulose can be completely solubilized, involving particularly polymeric fractions with low viscosity-average molecular weight.6 The use of LiCl/N,N-dimethylacetamide (DMAc)7 has also become very popular, as it is a fast, easy, and reproducible method to dissolve cellulose directly. Most recently, processes using environmentally acceptable solvent systems, such as NMMO/water (Lyocell process), have been largely used because of their ability to dissolve high amount of cellulose (up to 35 wt %) and to obtain viscous, optically anisotropic solutions. The formation of liquid crystalline solutions of cellulose may have useful applications for the generation of new, advanced materials. However, high temperatures are required for the dissolution of cellulose. This represents a big concern, as the flash point of the solvent is 150 °C, which is very close to the temperature used for the dissolution of the cellulose (130 °C).8 Nanocomposites are a new class of hybrid materials characterized by an ultra fine dispersion of a phase (typically of nanometer order of magnitude) into a polymeric matrix.9-11 Due to this dispersion, these materials possess unique properties, behaving much differently than conventional composites or microcomposites, and offering new technological and economical opportunities. The first studies on nanocomposites were carried out in 1961, when Blumstein performed the polymerization of a vinyl monomer intercalated into a montmorillonite structure.12 Since then, clay-based polymer nanocomposites have emerged as a new class of materials and attracted considerable interest and investment in research and development worldwide.13,14

10.1021/bm8002946 CCC: $40.75  2008 American Chemical Society Published on Web 10/09/2008

Cellulose-Montmorillonite Nanocomposites

Of particular interest are polymer nanocomposites reinforced with organically modified layered silicate (OMLS) because of their demonstrated significant enhancement, relative to an unmodified polymer resin, of a large number of physical properties, including barrier, flammability resistance, thermal and environmental stability, solvent uptake, and rate of biodegradability of biodegradable polymers.13 The pioneering studies on OMLS nanocomposites were carried out in the early 90s, when researchers from Toyota discovered the possibility of synthesizing a nanostructure from a polymer and an organophilic silicate. Their new material, based on polyamide 6 and organomodified montmorillonite showed great improvement in mechanical and barrier properties as well as in heat distortion temperature. These improvements were obtained at very low loading of layered silicate (4% wt): the same result could be achieved by adding the 50% wt of clay microparticles.15 Montmorillonite clays have a large surface area, which provides a substantial interfacial region in the nanocomposites that is expected to lead to an enhancement in mechanical and thermal properties at very low percent addition of the filler. Several natural biodegradable polymers are attracting considerable interest in materials science research as matrices in the preparation of layered silicate nanocomposites. Preparative techniques include intercalation of polymers or prepolymers from solution, in situ intercalative polymerization method, and melt intercalation method.16 Park et al. prepared, by melt intercalation, nanocomposites based on thermoplastic starch (TPS) with one unmodified Na+ MMT and different ammonium cations modified montmorillonites. They found that the TPS/ cloisite-Na+ nanocomposites showed higher tensile strength and thermal stability, better barrier properties to water vapor than the TPS/organomodified MMT hybrids.17,18 These results were explained in terms of polar interaction between the silicate layers and the TPS, which is stronger in the case of the unmodified Na+ MMT. The same authors also employed cellulose acetate for the fabrication of hybrids by melt intercalation, using organically modified clay.19,20 However, cellulose esters can be extruded to produce cellulose nanocomposites only with the additional use of plasticizers. Intercalation of polymers from solution allows using unmodified cellulose as composite matrix. In literature, one study on cotton/clay nanocomposites is available. In that work, significant changes in decomposition temperatures compared to unfilled cotton were reported, indicating that these materials have the potential to be used in flame retardant textiles.21 The addition of the montmorillonite clay in amounts as small as 1% increased decomposition temperatures by about 40 °C. This significant increase in the thermal stability at low filler contents is a common property of synthetic polymer-clay nanocomposites. Furthermore, the author observed that the ammonium salt of dodecylamine was able to separate the silicate sheets of the montmorillonite efficiently, whereas the use of unmodified clay failed to cause its exfoliation. Actually, from the above-reported literature it turns out that the morphology of polymer/clay hybrids depends on the compatibility and interaction between polymer and silicate layer. In addition, in the case of polymers with bulky repeating units, such as piranosic rings in cellulose, the interlayer distance should be large enough to allow polymer to enter the clay galleries. In the present paper, the preparation and characterization of cellulose-based nanocomposites is described. For this purpose, microcrystalline cellulose dissolved in a 50 wt % NMMO/water

Biomacromolecules, Vol. 9, No. 11, 2008

3005

system was used. To enable cellulose to diffuse into the clay galleries, a trimethyl-stearylammonium-modified montmorillonite was used as filler. The clay modifier contains polar cationic groups, which are supposed to positively interact with cellulose and long alkyl chains that provide large interlayer distance. The study was carried out by comparing the properties of the pure regenerated cellulose with the corresponding nanocomposite containing 3 wt % of modified montmorillonite to investigate on the resulting thermal properties of the material, which might be of interest for fire-resistant final products. We were aware that high temperatures as well as long stirring times could bring about the partial degradation of the material. In fact, as soon as the polymer dissolution takes place, cellulose degradation starts occurring mainly through radical processes, accompanied by discoloration of the resulting fibers, accelerated decomposition of NMMO, or even thermal runaway reactions.22,23 Moreover, ammonium-exchanged MMT suspensions were found to be able to catalyze degradation of cellulose leading to chain scission and formation of carbonyl groups on polymer backbone.24 On the other hand, a prolonged contact time between the cellulose solution and the clay suspension is required to achieve effective polymer diffusion into the silicate layers. Based on the above-reported considerations, the influence of contact time was investigated in our study by preparing two nanocomposite samples characterized by different reaction times. Morphological, thermal and thermo-oxidative properties were determined for both pure and filled cellulose. To probe the occurrence of cellulose degradation during the regeneration process, the prepared samples were also characterized by chemiluminescence (CL) due to the outstanding sensitivity of this technique to degradation reactions involving radical species.25-27

Experimental Section Materials. For nanocomposite preparation the following materials were used: microcrystalline cellulose, average powder size 50 µm, FMC Avicel type PH-101; 50 wt % N-methylmorpholine-N-oxide aqueous solution (NMMO), Lancaster; trimethylstearylammonium salt modified montmorillonite (MMT), NANOCOR Nanomer I.28. Preparation of Regenerated Cellulose. A total of 3 g of microcrystalline cellulose and 60 g of 50 wt % NMMO aqueous solution were introduced in a round-bottom flask immersed in an oil bath kept at 130 °C under magnetic stirring at a constant nominal speed of 800 rpm. After approximately 2 h, as soon as the mixture turned brownred and transparent due to complete dissolution of cellulose, the polymer was precipitated into 200 mL of distilled water and then filtered and rinsed with water. To remove the residual NMMO, the product was washed with hot water at T ) 80 °C for 30 min, then filtered, and dried in the oven under vacuum at T ) 80 °C overnight. After being dried, regenerated cellulose was ground in a mortar to the uniform powder. Preparation of Cellulose Nanocomposites. A total of 0.090 g of organically modified montmorillonite was added to 60 g of NMMO aqueous solution at 130 °C, and stirred for 1 h to achieve a good dispersion of the clay. Successively, 3 g of cellulose was added and the cellulose/MMT composite was obtained according to the same procedure followed for the regenerated cellulose. In this way, a nanocomposite containing 3 wt % with respect to the cellulose weight was prepared. The sample thus obtained is designated as 3%MMT-A. To achieve an effective intercalation of the polymer into the clay galleries, an alternative procedure was followed, in which the solution of dissolved cellulose with montmorillonite was kept under stirring 15 additional minutes, before being precipitated into water. This nanocomposite is designated as 3%MMT-B.

3006


Cerruti et al. Table 1. Interlayer Spacing and 2θ Values for Nanomer I.28 and Cellulose-Based Nanocomposites

Figure 1. Semilogarithmic plot of the XRD spectra of the organomodified montmorillonite Nanomer I.28 and the cellulose-based nanocomposites.

Times of reaction longer than 15 min led to a vigorous bubbling of gases formed into the solution, related to thermal degradation processes of the polymer, which eventually resulted in a brownish, apparently degraded product. Characterization Techniques. X-ray Diffraction (XRD). XRD experiments were performed using a Philips PW 1710 diffractometer with a rotating anode generator and a wide-angle power goniometer. The radiation was Cu KR not filtered, with 40 kV voltage and 20 mA intensity. The scan rate was 1 °/min over a diffraction angle 2θ, ranging between 2 and 40°. Transmission Electron Microscopy (TEM). Prior to being observed, powdered samples were embedded into an epoxy resin. Ultrathin sections (with thickness of ca. 50 nm) for TEM observations were then cut under cryogenic conditions using a Leica EM FCS ultramicrotome with a diamond knife. The TEM micrographs were obtained by means of ZEISS EM 900 transmission electron microscope under an accelerated voltage of 50 kV. ThermograVimetric Analysis (TG). The thermal stability of cellulose nanocomposites was evaluated by thermogravimetry. TG measurements were carried out at 1, 5, 10, and 20 °C/min heating rates from 25 to 650 °C both in a nitrogen and in an air atmosphere (flow rate 50 mL/ min) on powdered samples (approximately 15 mg) on the Dupont Instrument model 951 thermogravimetric balance. To eliminate small amounts of water absorbed by hygroscopic materials, a heating program including a 30 min, isothermal segment at 90 °C, followed by a heating ramp up to 650 °C, was adopted. To calculate the apparent activation energy, Ea, kinetic analysis of the TG data was performed by using the classical Flynn-Wall-Ozawa isoconversional method.28,29 Chemiluminescence (CL). CL experiments were performed on the photon counting instrument Lumipol 3 manufactured at the Polymer Institute of Slovak Academy of Sciences, Bratislava, Slovakia. The instrument dark count rate was 2-4 counts/s at 40 °C. Finely ground samples (average weight 5 mg) were placed into aluminum pans, and analyzed both under nitrogen and oxygen at a gas flow of 3 L/h. Nonisothermal runs from 40 to 220 °C were performed at a heating rate of 2.5 °C/min. The intensity of luminescence was expressed as counts/s · 5 mg.

Results and Discussion X-ray Diffraction (XRD). Figure 1 shows the XRD pattern of the Nanomer I.28 along with the spectra related to 3%MMT-A and 3%MMT-B nanocomposites. Diffraction angle values and calculated interlayer spacings are listed in Table 1. The diffraction pattern of Nanomer I.28 exhibits a peak at 2θ ) 3.4°, corresponding to the mean interlayer spacing of the (001) plane for the alkylammonium modified silicate. The calculated spacing of 26.0 Å is typically found in a paraffinic arrangement.30-32

sample

2θ (°)

interlayer spacing (Å)

Nanomer I.28 3%MMT-A 3%MMT-B

3.4 3.2 6.3

26.0 27.6 14.0

It is reported in literature that, in the absence of organomodifiers, montmorillonite is characterized by an interlayer spacing comprised in the range between 9.5 and 15 Å33 depending on the interlayer ions (Ca2+, Na+, etc.). Substitution of these ions by bulky alkyl chains leads to an increase in the characteristic interlayer spacing, which can favor guest molecules to enter the silicate galleries. The XRD curve of Nanomer I.28 displays two other peaks at higher diffraction angles (2θ ) 5.1° and 7.0°, corresponding to d-spacing ) 17.3 and 12.6 Å, respectively). The basal diffraction at 2θ ) 5.1° (d-spacing ) 17.3 Å) may correspond to a fraction characterized by a different alkylammonium chain arrangement in the interlayer space, indicating that possibly lateral bilayer to inclined monolayer chain orientations are present.34 The broad peak at 7.0° may be related to the (002) reflection representing the half-length of the actual average distance (26.0 Å) between two silicate layers. In the case of the nanocomposite 3%MMT-A, the occurrence of a 1.6 Å increase in the interlayer distance (Table 1) can be observed. This indicates poor intercalation of cellulose within montmorillonite stacks, yielding a material with both microcomposite and nanocomposite morphological features. A different morphology is evidenced for 3%MMT-B, which was obtained by reacting the polymer/clay mixture for a longer time after complete dissolution of cellulose. The XRD pattern of this sample (Figure 1) shows that the peak at 2θ ) 3.4° was shifted to smaller angles (which were not accessible with our equipment), suggesting the occurrence of clay platelets exfoliation operated by the polymer. It is also worth noting that a new diffraction peak can be observed at a 2θ value of 6.3°, corresponding to a 14.0 Å spacing. This peak is likely due to the structural instability of Nanomer I.28 rather than to 002 reflection of the intercalated organosilicate, as the peak related to the 001 reflection is absent. Most probably, a small fraction of the clay structure collapsed, causing the reduction of the interlamellar gallery height, and the arrangement of the intercalating agent into a lateral monolayer orientation.34 A similar observation has been reported in literature for several polymer nanocomposite systems.35-38 Yoon et al. observed that the interlayer structure of the clay collapsed from the bilayer arrangement of alkyl chains to monolayer one during annealing of polystyrene-based nanocomposites.37 An analogous result was found by Kwon and co-workers, who related it to the presence of montmorillonite in which a partial substitution of the metal cations was accomplished by hydrogen ions deriving from acid-base reactions of ammonium ion in water.38 Tanoue et al. also observed reduction of the interlamellar gallery height of sodium MMT intercalated with dimethyl benzyl hydrogenated tallow ammonium chloride upon melt intercalation by polystyrene at 200 °C. They attributed the reduction of interlayer spacing to the decomposition of the onium intercalant, leading to its breakdown into amines and a long-chain olefins.39 In the present work, two main factors accounting for the reduction of the interlayer spacing in 3%MMT-B can be considered, namely degradation of the ammonium organomodifier and intercalation of NMMO between the clay galleries followed by layer rearrangement.



3007

Therefore, reaction times must be carefully adopted in order to overcome this issue, aiming at the optimization of the composite morphology without compromising its final properties due to degradation reactions. Transmission Electron Microscopy (TEM). TEM provides the morphological features of cellulose layered-silicate nanocomposites. Figures 3a and 3b are micrographs of the 3%MMT-A and 3%MMT-B samples, respectively. Clay platelets are visible as dark lines (some are indicated with arrows). In both cases, clay is not homogeneously distributed within the polymer matrix, as evidenced by the clear zones observable in Figure 3a,b. This suggests that a partial compatibility between polymer and clay surface exists. Figure 2. Semilogarithmic plot of the XRD spectra of Nanomer I.28, before and after treatment with a NMMO solution.

To check whether NMMO intercalation is able to alter the montmorillonite structure and to evaluate the influence of the high-temperature NMMO treatment on the structure of the organo-modified clay, the latter was mixed with a NMMO solution at 130 °C, under the same conditions used for the preparation of the composites, in the absence of cellulose. Figure 2 shows the XRD curve of Nanomer I.28, along with the spectra related to the clay kept under stirring in a 50 wt % NMMO aqueous solution at 130 °C for 60 and 120 min, respectively. This treatment brought about only slight modifications to the clay structure, even after a 2 hour mixing. The peak at 3.4° was almost unaffected, showing evidence of a barely decreased interlayer spacing. Besides, no new peaks could be detected, although the peak at 5.1° disappeared, suggesting a rearrangement of a small fraction of intercalated alkyl chains. As already observed,21 NMMO molecules have a molecular size too small to change the d-spacing between the silicate layers significantly. This result suggests that the reduction of the interlayer spacing observed for 3%MMT-B can not be attributed to the exchange of alkylammonium groups of the modified MMT for NMMO. Therefore, it is likely that the observed reduction is due to thermal degradation reactions occurred during thermal treatment. In fact, although NMMO and the organo-modifier were found to be stable in the conditions used for solution compounding, cellulose/NMMO solutions do not. Several homolytic and heterolytic side reactions yielding considerable byproduct formation in the system cellulose/NMMO/water have been described.8 Radical reactions between cellulose and primary aminyl radicals formed upon degradation of NMMO cause random scission of the polymer backbone, yielding carboxyterminated cellulose oligomers. Heterolytic cleavage of the N-O bond in NMMO also produces oxidized cellulose derivatives. These byproducts in turn act as inducers of NMMO degradation reactions, accelerating the overall reaction kinetics. Due to their higher polarity, the degradation products of cellulose can enter the clay galleries, facilitating thermal degradation of the alkylammonium modifier and reducing the interlayer spacing of the organo-clay. The above-reported results highlight the key role of the reaction time on the morphology of the resulting composites. Clearly, a complete clay intercalation or even exfoliation can be obtained only by complete dissolution of the cellulose in the NMMO/water system, which allows polymer/clay interaction to occur in a homogeneous phase. Besides, higher mixing times favor the diffusion of the macromolecules into the silicate galleries. However, homogeneous cellulose solutions in NNMO/ water can be obtained only at temperatures as high as 130 °C, where the thermal degradation rate of cellulose is not negligible.

However, by a comparison between the two images, a different morphology characterizing the two hybrids can be put into evidence. Figure 3a shows the prevalence of clay aggregates, in which the montmorillonite layers retain much of their parallel alignment and appear to stack together in large domains of about 100 nanometers. Furthermore, a small amount of individual nanometer-thick sheets is also evident. This suggests the occurrence of a mixed morphology in which regions of both exfoliated and intercalated nanostructure are present, the latter being predominant. On the other hand, Figure 3b shows the clay to be better dispersed throughout the polymer matrix. Single platelets and stacks of just a few individual layers as well as intercalated tactoids can be observed. That is, a prevailing exfoliated morphology is evidenced by TEM. This is in good agreement with the XRD spectrum (see Figure 2), which no longer exhibits the characteristic peaks of the organophilic montmorillonite. From the above-mentioned results, it can be inferred that increased reaction times yielded improved clay dispersion. Thermogravimetric Analysis (TG). Thermal stability of regenerated cellulose and cellulose-based nanocomposites was investigated by thermogravimetry under nitrogen and air flow. In Figure 4, the TG traces recorded in nitrogen at 5 °C/min for all the investigated samples are reported as an example. TG curves show that the process occurs in a single weight-loss step. The presence of the nanoclay filler leads to slight changes in thermograms. In particular, the degradation temperatures of both nanocomposite samples are increased at small weight loss values (300 °C), which imply the involvement of far superior thermal energies, are able to cause extensive depolymerization through faster cleavage of the cellulose backbone bonds, in which a radical mechanism is also operating.45-47 In this stage, the presence of the montmorillonite is not able to alter the behavior observed for the neat polymer. As far as the measurements carried out in air atmosphere are concerned, TG curves (Figure 5) show a second decomposition step due to the oxidation of the char produced during the first weight loss stage. The presence of nanoparticles also improved the cellulose thermo-oxidative stability. In particular, an increase in thermal stability by about 10 °C at a 5% weight loss (see also Table 2) was observed for both the nanocomposites compared to regenerated cellulose. The second weight loss step occurs from 300 to 500 °C and reflects the occurrence of thermal-oxidative degradation of the char. This process is slowed down for cellulose nanocomposites, suggesting an increased stability. Moreover, the partially exfoliated nanocomposite shows an increase in the temperature of the maximum decomposition rate, indicating improved stabilization against the oxidative process with respect to the intercalated composite. It is reasonable to hypothesize that the presence of the nanoclay retards the thermal transfer, and hinders the oxygen diffusion into the polymer bulk. The presence of montmorillonite also produces an increase in char formation in both the composites in comparison to the regenerated cellulose.

Kinetic analysis of the TG data recorded under air and nitrogen atmosphere for pure regenerated cellulose and its nanocomposites was performed using the isoconversional method established by Flynn-Wall-Ozawa, under the assumption of a first order degradation process. The apparent activation energy Ea was obtained from a linear fitting of log β versus 1000/T at different fractional weight losses (R), where β is the experimental heating rate and T is the absolute temperature. Figure 6 shows the thermograms obtained for the regenerated cellulose at different heating rates under nitrogen. These TG curves correspond to a single-stage decomposition reaction. However, it is noteworthy to observe that the residence time in the system has a marked effect on char formation, as lower heating rates bring about increased char yields. That is, the competition between volatilization and charring is strongly affected by the heating rate. In fact, at the slowest heating rates charring reactions are favored since the temperature remains low enough to inhibit the volatilization.48 At higher heating rates the system reaches a high temperature rapidly, and volatilization can occur quite efficiently. To investigate how the presence of clay and the reaction time affected the mechanism of the degradation process, Ea values were plotted versus R for all the prepared materials. In Figure 7, the curves relative to measurements carried out under nitrogen and air for the neat polymer and the nanocomposites are reported. Changes in Ea with the decomposition extent are observed for all the samples, under both a nitrogen and an air atmosphere, suggesting that different processes govern the degradation phenomenon at the initial and final stages.49 The overall process is thus constituted by a multistage reaction in which each singlestage reaction contributes partially to the global mechanism to a different extent, depending on the decomposition degree. It can also be observed that the presence of nanoparticles leads to drastic changes in the dependence of the activation energy on the degree of conversion. The average values of the activation energies obtained for each stage of the decomposition are summarized in Table 3. Figure 7a,b also shows that at fractional weight loss values above 60-70% the energy of activation of all the samples increases steeply because of charring processes. This is reflective of the breakdown in the assumption of the isoconversional model, caused by the variable amount of char formed as a function of heating rate, i.e., a competition between two different reaction pathways so that the reactions occurring at fixed conversion are not the same at different heating rates.



3009

Table 2. Loss Temperature of 5 wt % (T5wt%), Maximum Decomposition Rate Temperature (Tmax), and Char Yield at 600 °C (%Char yield600°C) for the Regenerated Cellulose and its Nanocompositesa40 nitrogen environment

air environment

sample

T5wt.% (°C)

Tmax (°C)

%Char yield600 °C

T5wt.% (°C)

Tmax1 (°C)

Tmax2 (°C)

%Char yield600 °C

regenerated cellulose 3%MMT-A 3%MMT-B

273.0 280.1 280.5

309.5 310.5 305.9

18.6 20.6 21.3

268.3 277.0 277.5

300.2 301.3 297.7

487.7 488.3 495.7

0.2 2.0 2.9

a

Heating rate ) 5 °C/min.

Figure 5. TG curves for the regenerated cellulose and its nanocomposites heated at 5 °C/min under air.

Figure 6. TG curves of neat regenerated cellulose under nitrogen at 1, 5, 10, and 20 °C /min.

As far as the measurements performed in nitrogen are concerned, pure cellulose shows high activation energy values in the fractional weight loss range 0-5%. Subsequently, activation energy decreases steeply and a steady value of about 134 kJ/mol is obtained. Finally, an increase in Ea is observed at a high degree of decomposition (65%), where charring starts to occur. It should be noticed that the overall dependence of Ea on R gives a curve, which shows a concave shape. The shapes of the dependence of Ea on R were identified from model data for different complex processes, comprising competing, independent, consecutive, as well as reversible reactions.50,51 It was demonstrated that concave Arrhenius dependencies are typical of complex processes containing a reversible intermediate stage followed by an irreversible one. The reversible dehydrations of crystal hydrates demonstrate these characteristic dependencies.52,53 The most commonly accepted mechanism for cellulose pyrolysis is the Boide-Shafizadeh mechanism,54,55 which describes the cellulose thermal decomposition as constituted of two different pathways, dehydration and depolymerisation, characterized by different activation energies. At relatively low

temperature and lower conversion degrees, cellulose dehydration prevails, leading to the formation of anhydrocellulose,56 with elimination of water. In addition, the anhydrocellulose can then undergo thermal scission reactions to produce a terminal hydroxyl glucosidic chain end. The latter can also undergo a dehydration reaction, producing further water molecules. It was calculated that the theoretical mass loss due to water elimination during anhydrocellulose formation is about 5%.57 Besides, a similar amount of water could be theoretically produced during the second dehydration stage, giving a value of 10% as a rough estimate of the weight loss involved in the overall dehydration processes. As shown in Figure 7, for neat cellulose a decomposition process showing an Ea of about 220 kJ/mol is predominant up to a 5% relative weight loss. This suggests a kinetic scheme in which the reversible dehydration process contributes largely to the determination of Ea values at low conversions. At this stage, Ea is limited by the sum of the activation energy of the irreversible reaction and the enthalpy of the reversible dehydration.58 At higher conversion degrees the contribution of the irreversible depolymerisation pathways prevails distinctly, and the overall Ea values (135 kJ/mol) are limited by the activation energy associated with the latter reaction. The activation energies determined in the present paper are comparable with Ea values found in the literature: Fraga et al. obtained an average value of 187 kJ/mol by Friedman’s method,59 whereas 238 kJ/mol was found by Va´rhegyi et al.43 by means of a least-squares method. The latter authors explained their result in terms of a high activation energy rate determining stage in a complex mechanism scheme. A different behavior is observed in the case of the two nanocomposites. 3%MMT-A shows lower Ea at lower conversion values, and then a slight increase is recorded, with Ea values ranging in between 180 and 200 kJ/mol up to a 50% conversion. The dependence of Ea on R is even less evident for 3%MMTB, which shows higher and constant Ea values (210 kJ/mol) throughout the degradation process prior to the increase due to charring. In the case of the nanocomposites, it is likely that the clay platelets act as a barrier to hinder the diffusion of heat and migration of water and degraded volatiles as well. This might imply that the partial contribution of dehydration reactions is significant also at higher degradation degrees, determining markedly higher activation energy values throughout the investigated range of decomposition. It should be noticed that this phenomenon is particularly evident in the case of the partially exfoliated nanocomposite. In fact, the exfoliated silicate platelets are thought to be more effective in blocking heat and volatiles than tactoids in intercalated structures. As a consequence, the activation energy of 3%MMT-A tends to be lower. However, since the tactoids still show retardant effects to heat and volatiles, the activation energy of this sample is still higher than that of the regenerated cellulose. The results obtained from the experiments performed in air atmosphere indicate that Ea increases with R over the investi-

3010


Cerruti et al.

Figure 7. Activation energy (Ea) as a function of R for the decomposition processes of neat regenerated cellulose and its nanocomposites under (a) a nitrogen and (b) an air atmosphere. Table 3. Activation Energies for Different Stages of Thermal Degradation of Cellulose and its Nanocomposites under Nitrogen and Air fractional weight loss range 0-5% 15-50% 55-70% 80-95%

Ea under air (kJ/mol)

Ea under nitrogen (kJ/mol) regenerated cellulose

3%MMT-A

3%MMT-B

regenerated cellulose

3%MMT-A

3%MMT-B

222.5 133.9 153.5

176.1 190.4 264.2

208.9 203.5 240.4

152.3 173.0 265.9 169.6

170.8 180.5 293.5 165.6

174.3 196.9 340.8 177.0

gated decomposition range and slight differences can be observed between the three samples (Figure 7b and Table 3). This probably reflects the fact that the thermal oxidation constitutes the main degradative process and it markedly affects the overall degradation kinetics. However, it is worth noting that at fractional weight losses lower than 5%, significantly higher activation energies are observed for both the nanocomposites with respect to the cellulose, suggesting that the presence of the montmorillonite causes an increase in stabilization, probably affecting the rate of the initiation of thermal oxidation. It is believed that thermal oxidation of cellulose occurs through a chain reaction mechanism close to the auto-oxidation processes proposed for thermal oxidative degradation in liquid phase.47 Free radicals were detected in photoirradiated and thermally oxidized cellulose. Their reaction with oxygen leads to the formation of peroxides and hydroperoxides. The latter can decompose homolytically to yield hydroxyl radicals, able to abstract hydrogen on polymer backbone and to further propagate thermo-oxidative chain reaction. The net result is depolymerisation and formation of oxidized compounds. Thus, any process able to retard hydroperoxide formation can lead to a partial stabilizing effect. In general, hydroperoxide decomposition shows lower values of activation energy, as detected for cellulose in air with respect to nitrogen. Therefore, the presence of montmorillonite layers is able to stabilize the materials against oxidation due to a decreased diffusion of oxygen within polymer bulk, and to a retardation of hydroperoxide formation, as suggested by higher Ea values calculated for both the nanocomposites. As the temperature increases, the limiting stage of the degradation process shifts toward mechanism initiated by random scission characterized by higher activation energies.60,61 Also in this stage, 3%MMT-B sample shows the higher average value of activation energy (196.9 kJ/mol). In the 60-80% fractional weight loss range, due to the occurrence of charring, activation energy values are not reliable, as char yield depends on heating rate. The second decomposition step distinguishable from TG thermograms under air is related to the oxidation of the char.

For all the samples this stage is characterized by steeply decreasing activation energies (Figure 7b), which might be the manifestation of the autocatalytic combustion process. Chemiluminescence Emission in Nitrogen and Air. CL experiments in inert atmosphere give differentiation among respective samples at considerably lower temperatures. The CL method may be found useful at the investigation of different free radical processes that are responsible for the light emission from the cellulose. Figure 8 shows the CL emission for the cellulose in a nitrogen atmosphere and the effect of intercalated and partially exfoliated cellulose-montmorillonite nanocomposites. In the case of pure cellulose, a first CL maximum at 140 °C and a more intense peak, centered at around 210 °C, are observed. The latter is likely related to the dehydration and subsequent cleavage processes of the polymer chain, followed by recombination of free radicals formed, which can transfer the energy released during such process to accepting molecules, mainly carbonyl compounds, present in the system due to regeneration process.62 The peak at 140 °C can be connected with the presence of peroxides, formed in the polymer as a result of the prolonged heating under air in the NMMO solution: decomposition of the latter species can cause the cleavage of

Figure 8. CL emission as a function of the temperature in nitrogen for the regenerated cellulose and cellulose-based nanocomposites.



3011

Scheme 1. Formation of Carboxylic Acids through Oxidation of Hydroxyls in Cellulose

Scheme 2. Formation of Carboxylic Acids through Oxidative Scission of the Piranosic Rings in Cellulose

the polymer chain, leading to the formation of excited carbonyls, which emit CL. The presence of clay leads to a different behavior that seems to be affected either by the morphology of the composite or by the time of residence prior to the regeneration of the cellulose. First, the peroxidic peak is shifted to lower temperature (3%MMT-A) and finally it becomes completely eliminated (3%MMT-B). As for the second peak, in the case of the intercalated nanocomposite a difference can be noticed in comparison to the original regenerated cellulose, as an increase of the area under the peak is observed. This suggests that the clay can increase the quantum yield of CL reaction in comparison to the pure polymer. The partially exfoliated nanocomposite, obtained using longer times of reaction, shows even a much more intense peak. This indicates the occurrence of degradation reactions due to longer mixing time employed to prepare the composite, which brought about the formation of carbonyl compounds on the polymer backbone. Moreover, montmorillonite polarity can favor the interaction with cellulose chain terminal hydroxyl moieties, thus increasing the rate of their dehydration and oxidation63 without significant changes of the main chain cleavage. The total cellulose luminescence intensity is greater in oxygen than in nitrogen, because under an oxygen atmosphere it can be related both to decomposition of peroxides formed during the preparation of samples (as already seen in nitrogen), and to chain-scission reactions leading to free radical intermediates, which can also be converted to peroxyl radicals, regardless of the mode of initiation.62 A great number of studies devoted to cellulose degradation was focused on its oxidation catalyzed by heat, light, or chemicals, such as periodate.47,64-68 Oxidation of hydroxyls yields carboxylic acids via aldehyde formation (Scheme 1). Another reaction that occurs in the presence of oxygen is the oxidative scission of the piranosic rings (Scheme 2). These functional groups, generated during the extensive cellulose oxidation, can lead to the formation of hydroperoxides and peracids, which act as an efficient source of excited carbonyls, resulting in light emission.69 Figure 9 shows the CL emission in oxygen as a function of the temperature, for all the prepared samples. It can be noted that, at a given temperature, the intensity is approximately by 2 orders of magnitude higher than in nitrogen. This indicates that the flushing oxygen actively participates to the formation of peroxyl radicals. It may be of interest that Whatman filter paper which is pure cellulose appears to be more stable that the regenerated cellulose. A strikingly different behavior is displayed by the CLtemperature curve for the partially exfoliated nanocomposite

when compared with original regenerated cellulose. The marked difference between the maximum temperatures of CL emission which is markedly higher than 220 °C for the 3%MMT-B sample suggests a superior level of thermo-oxidative stability attained, and confirms the results obtained by means of thermogravimetry experiments under air. Thus, in the case of regenerated samples, the partially exfoliated morphology provides increased thermo-oxidative stability, owing to the maximized dispersion of clay platelets into polymer matrix. Therefore, CL experiments suggest that the hindering of oxygen transfer within the regenerated materials, which enhances stability, is quantitatively more important than the catalytic activity ascribed to the clay, detected under an inert atmosphere. This globally leads to an increase in thermo-oxidative stability of the cellulose. This may be also exemplified by the values of the rate constants obtained from nonisothermal CL runs by the method described elsewhere (Table 4).70 Here, cellulose-based nanocomposites are less stable at lower temperature of the oxidation (100 °C) and more stable at higher temperatures, such as 200 °C. This higher temperature region is in perfect agreement with thermogravimetry data in air while no conclusion can be drawn from thermogravimetry at lower temperatures because of the lack of sensitivity.

Conclusions Cellulose-based nanocomposites were obtained by solution precipitation of microcrystalline cellulose from a 50 wt % NMMO/water system in presence of organo-modified MMT. Two composites containing 3 wt % of modified MMT were

Figure 9. CL emission as a function of the temperature in oxygen for regenerated cellulose, cellulose-based nanocomposites, and Whatman filter paper (full line is the experimental run and points denote the theoretical fit of the run from the parameters of the experimental run determined according to ref 70).

3012


Cerruti et al.

Table 4. Rate Constants of Oxidation of Different Cellulose-Based Materials at 100, 150, and 200 °C rate constant of cellulose oxidation (s-1)* 100 °C

sample Whatman filter paper Neat regenerated cellulose 3%MMT-A 3%MMT-B

150 °C

200 °C

-7

5.9 × 10 3.6 × 10-6

-5

2.1 × 10 8.5 × 10-5

4.4 × 10-4 1.8 × 10-3

5.3 × 10-6 4.1 × 10-6

1.0 × 10-4 4.8 × 10-5

1.4 × 10-3 5.9 × 10-4

* The rate constant is the average value calculated from the rate constant of faster (ionic) and slower (free radical) degradation of cellulose sample according to ref.70

prepared by varying the contact time in the regenerating solutions. The properties of the pure regenerated cellulose were compared with those of the corresponding hybrids. From X-ray diffraction and TEM analysis it was inferred that reaction time significantly affected the morphology of the resulting composites. In particular, longer reaction times allowed better dispersion of macromolecules within the silicate galleries, yielding a partially exfoliated nanocomposite. The thermal and thermal oxidative properties of the hybrids were investigated by TG and CL. The nanocomposites showed increased degradation temperatures compared to plain cellulose. In particular, the maximum stability was observed for the partially exfoliated sample, as the higher extent of clay dispersion hindered the migration of degraded volatiles and the diffusion of heat and oxygen. Kinetic analysis of the TG data was performed by means of the Flynn-Wall-Ozawa method. It resulted that the degradation of regenerated cellulose as well as cellulose-based hybrids occurred through a multistep mechanism. For plain cellulose, dehydration reactions prevailed at lower conversion degrees, contributing largely to the determination of Ea. At higher conversions, irreversible depolymerisation pathways became predominating, thus reducing Ea values. On the other hand, the presence of MMT led to drastic changes in the dependence of the activation energy on the conversion extent. More precisely, the homogeneous dispersion of MMT determined higher Ea values in nitrogen atmosphere, while improved thermal oxidative stability due to retarded peroxide formation was observed under air. Moreover, it was observed that MMT increased the rate of dehydration and oxidation of cellulose functional moieties. The rate constants of CL emission upon oxidation of cellulose-based materials showed that dehydration and oxidation were responsible for a reduced stability at temperatures lower than 100 °C. On the other hand, at higher temperatures the hindering of oxygen transfer prevailed, determining an increase in thermo-oxidative stability. Acknowledgment. The authors acknowledge the financial support from grant agency VEGA (Project Nos. 2/5109/27 and 2/6115/27) and thank Dr. Vincenzo Frezza for performing X-ray diffraction experiments.

References and Notes (1) Schweizer, E. J. Prakt. Chem. 1857, 72, 109–111. (2) Cross, C. F.; Bevan, E. J.; Beadle, C. Plastic Compound of Cellulose. US520770, 1894. (3) Potthast, A.; Rosenau, T.; Buchner, R.; Ro¨der, T.; Ebner, G.; Bruglachner, H.; Sixta, H.; Kosma, P. Cellulose 2002, 9, 41–53. (4) Takaragi, A.; Minoda, M.; Miyamoto, T.; Qing Liu, H.; Na Zhang, L. Cellulose 1999, 6, 93–102. (5) Heinze, T.; Koshella, A. Polı´meros 2005, 15, 84–90, available from //www.scielo.br/scielo.php?script)sci_arttext&pid)S010414282005000200005&lng)en&nrm)iso. (6) Zhiang, L.; Ruan, D.; Zhou, J. Ind. Eng. Chem. Res. 2001, 40, 5923– 5928.

(7) Dupont, A. L. Polymer 2003, 44, 4117–4126. (8) Rosenau, T.; Potthast, A.; Sixta, H.; Kosma, P. Prog. Polym. Sci. 2001, 26, 1763–1837. (9) Giannelis, E. P. Appl. Organomet. Chem. 1998, 12, 675–680. (10) Jordan, J.; Jacob, K. I.; Tannenbaum, R.; Sharaf, M. A.; Jasiuk, I. Mater. Sci. Eng. 2005, 93A, 1–11. (11) Hussain, F.; Hojjati, M.; Okamoto, M.; Gorga, R. E. J. Comp. Mater. 2006, 40, 1511–1575. (12) Blumstein, A. Bull. Chim. Soc. 1961, 899–905. (13) Ray, S. S; Okamoto, M. Prog. Polym. Sci. 2003, 28, 1539–1641. (14) Zeng, Q. H.; Yu, A. B.; Lu, G. Q.; Paul, D. R. J. Nanosci. Nanotechnol. 2005, 5, 1574–1592. (15) Kojima, Y.; Usuki, A.; Kawasumi, M.; Okada, A.; Fukushima, Y.; Kurauchi, T.; Kamigaito, O. J. Mater. Res. 1993, 8, 1185–1189. (16) Ray, S. S.; Bousmina, M. Prog. Mat. Sci. 2005, 50, 962–1079. (17) Park, H. M.; Li, X.; Jin, C. Z.; Park, C. Y.; Cho, W. J.; Ha, C. S. Macromol. Mater. Eng. 2002, 287, 553–558. (18) Park, H. M.; Lee, W. K.; Park, C. Y.; Cho, W. J.; Ha, C. S. J. Mater. Sci. 2003, 38, 909–915. (19) Park, H. M.; Liang, X.; Mohanty, A. K.; Misra, M.; Drzal, L. T. Macromolecules 2004, 37, 9076–9082. (20) Park, H. M.; Misra, M.; Drzal, L. T.; Mohanty, A. K. Biomacromolecules 2004, 5, 2281–2288. (21) White, L. A. J. Appl. Polym. Sci. 2004, 92, 2125–2131. (22) Rosenau, T.; Potthast, A.; Adorjan, I.; Hofinger, A.; Sixta, H.; Firgo, H.; Kosma, P. Cellulose 2002, 9, 283–291. (23) Wendler, F.; Graneβ, G.; Heinze, T. Cellulose 2005, 12, 411–422. (24) Sakran, M. A. J. Radioanal. Nucl. Chem. Lett. 1996, 213, 87–98. (25) Billingham, N. C., O’Keefe, E. S. Chemiluminescence from oxidative degradation of polymers. In International Conference on AdVances in the Stabilization and Controlled Degradation of Polymers; Patsis, A. V., Ed.; Technomic: Lancaster,1989; Vol. I, pp 1-8. (26) Zlatkevich, L. Chemiluminescence in evaluating thermal oxidative stability. In Luminescence Techniques in Solid-State Polymer Research; Zlatkevich, L., Ed.; Marcel Dekker: New York, 1989; pp 135-197. (27) Cerruti, P.; Carfagna, C.; Rychly´, J.; Matisova´-Rychla´, L. Polym. Degrad. Stab. 2003, 82, 477–485. (28) Flynn, J. H.; Wall, L. A. J. Polym. Sci., Part B: Polym. Lett. 1966, 4, 323–328. (29) Ozawa, T. Bull. Chem. Soc. Jpn. 1965, 38, 1881–1886. (30) Lagaly, G. Solid State Ionics 1986, 22, 43–51. (31) Ambrogi, V.; Silvestre, M. G.; Vito, F.; Carfagna, C.; Errico, M. E.; Mancarella, C. Mol. Cryst. Liq. Cryst. 2005, 429, 1–20. (32) Vaia, R. A.; Teukolsky, R. K.; Giannelis, E. P. Chem. Mater. 1994, 6, 1017–1022. (33) Xu, W. B.; Bao, S. P.; He, P. S. J. Appl. Polym. Sci. 2002, 84, 842– 849. (34) Triantafillidis, C. S.; LeBaron, P. C.; Pinnavaia, T. J. J. Solid State Chem. 2002, 167, 354–362. (35) Zanetti, M.; Kashiwagi, T.; Falqui, L.; Camino, G. Chem. Mater. 2002, 14, 881–887. (36) Delozier, D. M.; Orwoll, R. A.; Cahoon, J. F.; Johnston, N. J.; Smith, J. G.; Connell, J. W. Polymer 2001, 43, 813–822. (37) Yoon, J. T.; Jo, W. H.; Lee, M. S.; Ko, M. B. Polymer 2001, 42, 329–336. (38) Kwon, O. Y.; Park, K. W.; Jeong, S. Y. Bull. Korean Chem. Soc. 2001, 22, 678–684. (39) Tanoue, S.; Utracki, L. A.; Garcia-Rejon, A.; Tatiboue¨t, J.; Cole, K. C.; Kamal, M. R. Polym. Eng. Sci. 2004, 44, 1046–1060. (40) Zong, R.; Hu, Y.; Wang, S.; Song, L. Polym. Degrad. Stab. 2004, 83, 423–428. (41) Wan, C.; Tian, G.; Cui, N.; Zhang, Y; Zhang, Y. J. Appl. Polym. Sci. 2004, 92, 1521–1526. (42) Soares, S.; Camino, G.; Levchik, S. Polym. Degrad. Stab. 1995, 49, 275–283. (43) Va´rhegyi, G.; Antal, M. J.; Jakab, E.; Szabo´, P. J. Anal. Appl. Pyrolysis 1997, 42, 73–87. (44) Pouwels, A. D.; Eijkel, G. B.; Boon, J. J. J. Anal. Appl. Pyrolysis 1989, 14, 237–280. (45) Bash, A.; Lewin, M. J. Polym. Sci., Polym. Chem. Ed. 1973, 11, 3071– 3093. (46) Hon, D. N.-S. J. Polym. Sci., Polym. Chem. Ed. 1979, 17, 441–454. (47) Malesˇic, J.; Kolar, J.; Strlic, M.; Kocar, D.; Fromageot, D.; Lemaire, J.; Haillant, O. Polym. Degrad. Stab. 2005, 89, 64–69. (48) Ball, R.; McIntosh, A. C.; Brindley, J. Phys. Chem. Chem. Phys. 1999, 1, 5035–5043. (49) Vyazovkin, S.; Sbirrazzuoli, N. Macromol. Rapid Commun. 2006, 27, 1515–1532.

Cellulose-Montmorillonite Nanocomposites (50) Nuń˜ez, L.; Fraga, F.; Nuń˜ez, M. R.; Villanueva, M. Polymer 2000, 41, 4635–4641. (51) Avella, M.; Carfagna, C.; Cerruti, P.; Errico, M. E.; Gentile, G. Macromol. Symp. 2006, 234, 163–169. (52) Vyazovkin, S. V.; Lesnikovich, A. I. Thermochim. Acta 1990, 165, 273–280. (53) Vyazovkin, S. V.; Linert, W Int. J. Chem. Kinet. 1995, 27, 73–84. (54) Di Blasi, C. J. Anal. Appl. Pyrolysis 1998, 47, 43–64. (55) Statheropoulos, M.; Kyriakou, S. A. Anal. Chim. Acta 2000, 409, 203–214. (56) Bradbury, A. G. W.; Sakai, Y.; Shafizadeh, F. J. Appl. Polym. Sci. 1979, 23, 3271–3280. (57) Liu, Q.; Lu, C.; Yang, Y.; He, F.; Ling, L. J. Mol. Struct. 2005, 733, 193–202. (58) Vyazovkin, S. A Int. J. Chem. Kinet. 1996, 28, 95–101. (59) Fraga, A.; Ruseckaite, R. A.; Jimeńez, A. Polym. Test. 2005, 24, 526–534. (60) Levchik, S. V.; Balabanovich, A. I.; Levchik, G. F.; Costa, L. Fire Mater. 1997, 21, 75–83. (61) Vyazovkin, S.; Dranca, I.; Fan, X.; Advincula, R. Macromol. Rapid Commun. 2004, 25, 498–503.


3013

(62) Strlic, M.; Kolar, J.; Pihlar, B.; Rychly, J.; Matisova-Rychla, L. Eur. Polym. J. 2000, 36, 2351–2358. (63) Ouyang, C. S.; Wang, C. M. J. Electrochem. Soc. 1998, 145, 2654– 2659. (64) Schutt, B. D.; Serrano, B.; Cerro, R. L.; Abraham, M. A. Biomass Bioenergy 2002, 22, 365–375. (65) Shafizadeh, F.; Bradbury, A. G. W. J. Appl. Polym. Sci. 1979, 23, 1431–1442. (66) Łojewska, J.; Mis´kowiec, P.; Łojewski, T.; Proniewicz, L. M. Polym. Degrad. Stab. 2005, 88, 512–520. (67) Milichovsky, M.; Sopuch, T.; Richter, J. J. Appl. Polym. Sci. 2007, 106, 3641–3647. (68) Vicini, S.; Princi, E.; Luciano, G.; Franceschi, E.; Pedemonte, E.; Oldak, D.; Kaczmarek, H.; Sionkowska, A. Thermochim. Acta 2004, 418, 123–130. (69) Konoma, F.; Cai, X.; Osawa, Z. Polym. Degrad. Stab. 2000, 69, 105– 111. (70) Rychly´, J.; Matisova´-Rychla´, L.; Laza´r, M.; Slova´k, K.; Strlicˇ, M.; Kocˇar, D. J. Carbohydr. Polym. 2004, 58, 301–309.

BM8002946

bm8002946

Recommend Documents