- Email: [email protected]
Crystalline properties and decomposition kinetics of cellulose fibers in wood pulp obtained by two pulping processes
Polymer Degradation and Stability 96 (2011) 679e685
Contents lists available at ScienceDirect
Polymer Degradation and Stability journal homepage: www.elsevier.com/locate/polydegstab
Crystalline properties and decomposition kinetics of cellulose fibers in wood pulp obtained by two pulping processes Matheus Poletto*, Vinícios Pistor, Mara Zeni, Ademir J. Zattera* Laboratory of Polymers (LPOL), Center of Exact Sciences and Technology (CCET), Caxias do Sul University (UCS), Caxias do Sul, RS, Brazil
a r t i c l e i n f o
a b s t r a c t
Article history: Received 24 May 2010 Received in revised form 3 December 2010 Accepted 8 December 2010 Available online 16 December 2010
In this study two cellulose fibers, Eucalyptus grandis (CEG) and Pinus taeda (CPT), obtained through the kraft and sulfite pulping processes, respectively, were characterized. Fourier transform infrared (FTIR) spectroscopy, X-ray diffraction (XRD) and thermogravimetric analysis (TGA) were carried out. From the XRD analysis the interplanar distance, crystallite size and crystallinity index were calculated and the degradation kinetics parameters were determined by TGA at heating rates of 5, 10, 20 and 40 C min1 using the Avrami, Flynn-Wall-Ozawa (FWO) and Criado methods. The results obtained by FTIR showed that the composition of the fibers is similar, while from the XRD analysis slight differences in the crystallinity were observed. The thermogravimetric analysis showed higher thermal stability for CPT than CEG while the values for the activation energy (Ea) were higher for CEG than CPT. The results obtained by Avrami and Criado methods showed that the degradation mechanism in the CEG samples involves a diffusion process while in the case of CPT the degradation process is a phase boundary controlled reaction. The degradation mechanisms demonstrated that the difference between thermal stability and Ea may be due to differences in the type of crystalline structure of the samples obtained through the two pulping processes. Ó 2010 Elsevier Ltd. All rights reserved.
Keywords: Eucalyptus grandis Pinus taeda Cellulose Crystallinity Kinetics Degradation
1. Introduction Cellulose is a polymer which contains crystallites and thus it has a paracrystalline morphology [1,2]. The linear cellulose molecules are linked laterally by hydrogen bonds to form linear bundles, leading to a crystalline structure [1]. Native cellulose is known to be a composite of two distinct crystalline modifications, namely Ia and Ib, whose fractions vary depending on the origin of the cellulose sample [2e4]. The Ia and Ib structures are assigned to one-chain triclinic and two-chain monoclinic unit cells, respectively [3]. These allomorphs are believed to coexist in the fibril in different ratios [4]. In addition, Ia cellulose is reported to be the dominant polymorph in bacterial and alga celluloses, while Ib cellulose is predominant in higher plants such as cotton and wood [5]. Therefore, the crystalline structure of cellulose affects the physical and mechanical properties of the cellulose fibers. The degree of crystallinity of cellulose is one of the most important crystalline structure parameters, and the rigidity of cellulose fibers increases and the flexibility decreases with an increasing ratio of crystalline to amorphous regions [1,6].
* Corresponding author. Tel.: þ55 54 3218 2108; fax: þ55 54 3218 2253. E-mail addresses: [email protected] (M. Poletto), [email protected] (A.J. Zattera). 0141-3910/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.polymdegradstab.2010.12.007
Changes occurring in the cellulose structure during pulping have been investigated in several studies [1,3,4,7]. Gümüskaya et al. [1] concluded that not only chemical agents but also the temperature and pressure in a cooking digester had major effects on the crystalline structure of cellulose in cotton linters during cooking. Hult et al. [4] investigated the organization of cellulose microfibrils in holocellulose, sulfite pulp and kraft pulp and revealed that in kraft pulp the microfibrils are more closely associated than in the sulfite pulp and holocellulose. Due to the complexity of cellulose thermal decomposition reactions, extensive research has been carried out in this area. Wu and Dollimore [8] investigated the thermal degradation behavior of natural cellulosic materials and demonstrated that the ratecontrolling mechanism involved mainly the phase boundary and probably diffusion processes. Nada and Hassan [9] investigated cellulose and some cellulose derivatives and reported activation energy values of 53e182 kJ/mol. Antal et al. [10] using dynamic thermogravimetric analysis found activation energy values of between 190 and 250 kJ/mol, depending on the type of cellulose, the heating rate and the mass of sample. Capart et al. [11] calculated kinetics parameters of microgranular cellulose using dynamic and isothermal methods in a nitrogen atmosphere and described two reactions with activation energies of 202 and 255 kJ/mol, respectively. The aim of this study was to investigate the effects of
680
M. Poletto et al. / Polymer Degradation and Stability 96 (2011) 679e685
chemical treatments on the crystalline properties and decomposition kinetics of cellulose fibers in wood pulp during two pulping processes. 2. Materials and methods 2.1. Materials Bleached sulfite cellulose fibers from Pinus taeda (CPT) were supplied by Cambará S.A (Cambará do Sul, Brazil). Bleached kraft cellulose fibers from Eucalyptus grandis (CEG) were supplied by CMPC S.A. (Guaíba, Brazil). The samples were dried at 70 C for 24 h in a vacuum oven before the tests. The average fiber particle length for CTP and CEG is around 150 mm. 2.2. Fourier transform infrared (FTIR) spectroscopy The Fourier transform infrared (FTIReNicolet IS10-Thermo Scientific) spectroscopy analysis was carried out with 32 scans, in the range of 4000 cm1 to 400 cm1, at a resolution of 4 cm1 using attenuated total reflectance (ATR), on cellulose powder under pressure on ATR crystal. 2.3. Thermogravimetric analysis (TGA) The thermogravimetric analysis (TGA50eShimadzu) was carried out under N2 atmosphere, from 25 up to 610 C. Approximately 20 mg of each sample was used. The analysis was carried out at four different heating rates (5, 10, 20 and 40 C min1). The results obtained were used to calculate the kinetics parameters. 2.4. X-ray diffraction (XRD) X-ray diffractograms were collected using a sample holder mounted on a Shimadzu diffractometer (XRD-6000), with monochromatic Cu Ka radiation (l ¼ 0.15418 nm) and the generator working at 40 kV and 30 mA. Intensities were measured in the range of 5 < 20 < 30 , typically with scan steps of 0.05 and 2 s/step (1.5 min1). Peak separations were carried out using Gaussian deconvolution. The determination coefficients (r2) were close to unity (0.9972 and 0.9947 for CEG and CPT, respectively). The dspacings were calculated using the Bragg equation [3,12], the crystallite sizes (L) were calculated using the Scherrer equation [3,12] and the crystallinity index (CrI) using the Segal method [1,3]. 2.5. Theoretical Considerations The fundamental equation used in all kinetic studies is generally described as:
da ¼ kðTÞf ðaÞ dt
(1)
where k is the rate constant and f (a) is the reaction model, a function dependent on the reaction mechanism. Eq (1) expresses the rate of conversion, da/dt, at a constant temperature as a function of the reduction in reactant concentration and the rate constant. In this study, the conversion a rate is defined as [13e16]:
a ¼
m0 mt m0 mf
(2)
where m0, mf and mt are the initial and final weights of the sample and its weight at time (t), respectively. The rate constant k is generally given by the Arrhenius equation:
Ea
kðTÞ ¼ Ae RT
(3)
where Ea is the apparent activation energy (kJ mol1), R is the gas constant (8.314 kJ mol1), A is the pre-exponential factor (min1), and T is the absolute temperature (K). The combination of Eqs. (1) and (3) gives the following relationship:
da ¼ AeEa =RT f ðaÞ dt
(4)
For a dynamic TGA process in a non-isothermal experiment, introducing the heating rate, b ¼ dT/dt, into Eq. (4), Eq. (5) is obtained as:
da ¼ dt
A Ea =RT e f ðaÞ
b
(5)
Equations (4) and (5) are the fundamental expressions of analytical methods used to calculate kinetic parameters on the basis of TGA data. 2.6. Flynn-Wall-Ozawa method In the FWO method [13,14], it is proposed that through the integration of Equation (5) and substituting (Ea/RT) with x, the following relationship can be obtained:
Za 0
da AE PðxÞ ¼ gðaÞ ¼ bR f ðaÞ
(6)
were P(x) is a function known as the Arrhenius integral that has no analytical solution but can be solved by numerical methods or using different approaches. The test method is based on integral approaches proposed by Doyle [17], in a range of logP(x)/x¼(Ea/RT) to 20 < x < 60. The integral of P(x) can be expressed simplistically as:
ln b ¼ ln
AE E 5:330 1:052 gðaÞR RT
(7)
In Eq. (7) g(a) is a function of the conversion. According to the principles of the isoconversional FWO method, it is assumed that the reaction rate at a given conversion is a function only of the temperature. Therefore, for different heating rates (b) and a given degree of conversion (a), a linear relationship is observed through a plot of log b vs. 1/T, and the apparent activation energy (Ea) is obtained from the slope of the straight line obtained [13e15]. 2.7. Criado method The degradation reaction mechanism can be determined using the Criado method [18e20] which can accurately determine the reaction mechanism in a solid reaction process, defined by a Z(a)type function:
ZðaÞ ¼
da=dt
b
pðxÞT
(8)
where x ¼ E/RT and p(x) is an approximation of the temperature integral which cannot be expressed in a simple analytical form. Paterson [19] proposed a reasonable relationship between p(x) and P(x) as shown in Eq. (9):
pðxÞ ¼ xex PðxÞ
(9)
M. Poletto et al. / Polymer Degradation and Stability 96 (2011) 679e685
681
Senum and Yang [20] proposed the fourth rational expression of P(x):
PðxÞ ¼
expðxÞ x3 þ 18x2 þ 86x þ 96 x x4 þ 20x3 þ 120x2 þ 240x þ 120
(10)
When x > 20, the error of Eq. (10) is less than 105%, which forms the basis of the analysis presented in this paper. Combining equations (1), (8) and (9) we obtain:
ZðaÞ ¼ f ðaÞgðaÞ
(11)
From equations (1), (8) and (9), the following relationship can be derived:
ZðaÞ ¼
da Ea Ea eRT PðxÞ dT R
(12)
Equation (11) is used to plot the master Z(a) vs. a curves for the different models listed in Table 1 [21,22], whereas Eq. (12) is used to represent the experimental curve. By comparing these two curves, the type of mechanism involved in the thermal degradation can be identified.
corrected kinetic constant (k0 ) through the use of the following equation:
2.8. Avrami analysis According to the Avrami model [23e25], under isothermal conditions, the degree of phase conversion (a(t)) is given by:
aðtÞ ¼ 1 ekt
n
(14)
According to Eq. (14), the parameters n and k are obtained, respectively, from the slope and the intercept of the straight line obtained from the plot of ln[-ln(1-a)] vs.lnt. Jeziorny [26] modified the Avrami equation. In Jeziorny’s analysis, the kinetic constant (k0 ) is determined in the same way as in the Avrami equation. However, a is now a function of temperature (a (T)). The corrected kinetic constant (k’) as a function of the Avrami kinetic constant (k) and the heating rate (b) is given as follows:
lnk0 ¼
. lnk b
t1=2 ¼
. 1 n ln2 k0
(16)
(13)
where k is the Avrami constant and n is the Avrami exponent which is dependent on the process dimensionalities. The parameters n and k can be determined by using the Avrami equation in the double logarithmic form:
ln½ lnð1 aðtÞÞ ¼ lnk þ nlnt
Fig. 1. IR spectra obtained for the cellulose samples studied.
(15)
The reaction half-life time (t1/2), which is defined as the time in which 50% of the reaction occurs [27], can be calculated from the
3. Results and discussions 3.1. Fourier transform infrared (FTIR) spectroscopy The IR spectra given in Fig. 1 show the composition of the cellulose samples studied. For the two samples a broad band can be observed in the region of 3600e3200 cm1, attributed to stretching of the hydroxyl (OH) groups [28,29], and 3000e2800 cm1, attributed to CH stretching [28e31]. According to Xiao [30] and Colom (a) [32], the band observed in the range of 1700e1550 cm1 may be associated with water absorption. The bands at 1430, 1365, 1335 and 1315 cm1 are attributed to CH2, in-plane CH deformation, in-plane OH bending and CH2 wagging, respectively [28e33], and the bands at 1160, 1110, 1052, 1028, 895 and 670 cm1 are associated with the asymmetric CeOeC bridge stretching, the anhydroglucose ring, C-OR stretching, CeOeC pyranose ring skeletal vibration, CH deformation and out-of-plane OH bending, respectively [28e33]. As can be observed, the spectra show the same structural characteristics. This suggests,
Table 1 Algebraic expressions for g(a) and f(a) for the most frequently used mechanisms of solid-state processes. Mechanism e Solid state process
g(a)
f(a)
A2 - Nucleation and growth (Avrami eq.1) A3 - Nucleation and growth (Avrami eq.2) A4 - Nucleation and growth (Avrami eq.3) R1 - Phase boundary controlled reaction (one-dimensional movement) R2 - Phase boundary controlled reaction (contracting area) R3 - Phase boundary controlled reaction (contracting volume) D1 - One-dimensional diffusion D2 - Two-dimensional diffusion (Valensi equation) D3 - Three-dimensional diffusion (Jander equation) D4 - Three-dimensional diffusion (GinstlingeBrounshtein equation) F1 - Random nucleation with one nucleus on the individual particle F2 - Random nucleation with two nuclei on the individual particle F3 - Random nucleation with three nuclei on the individual particle
[ln (1 a)]1/2 [ln (1 a)]1/3 [ln (1 a)]1/4
2 (1 a)[ln (1 a)]1/2 3 (1 a)[ln (1 a)]2/3 4 (1 a)[ln (1 a)]3/4 1 2 (1 a)1/2 3 (1 a)2/3 (1/2)a1 [ln (1 a)]1 (3/2)[1 (1 a)1/3]1 (1 a)2/3 (3/2)[1 (1 a)1/3]1 1a (1 a)2 (1/2)(1 a)3
a
[1ln (1 a)]1/2 [1ln (1 a)]1/3
a2
(1 a)ln (1 a) þ a [1 (1 a)1/3]2 [1 (2/3)a] (1 a)2/3 ln (1 a) 1/(1 a) 1/(1 a)2
682
M. Poletto et al. / Polymer Degradation and Stability 96 (2011) 679e685
reactions can occur, which reduce the total amount of amorphous region and therefore increase the crystallite size of CPT. 3.3. Thermogravimetric analysis (TGA)
Fig. 2. Diffractograms obtained for the cellulose samples studied.
in a preliminary qualitative analysis, that the CEG and CPT samples are similar in terms of composition. 3.2. X-ray diffraction (XRD) Fig. 2 shows the X-ray diffraction analysis of the cellulose samplesused in this study. The details for the three peaks are: peak 1 is Ib (11 0); peak 2 is Ib (110) and peak 3 is Ib (200) [4]. These peaks are separated by Gaussian deconvolution, and the d-spacings, crystallite size, crystallinity indices and Z-values were calculated. In this study, the Z-discriminant function developed by Wada and Okano [3] for the determination of the crystalline structure (monoclinic and triclinic) of cellulose in pulp samples was used. By employing discriminant analysis it is possible to categorize cellulose as type Ia or Ib [3]. The Z-value indicates whether cellulose is Ia or Ib [3]. The function which discriminates between Ia or Ib is given by Z ¼ 1693 d1-902 d2-549 d1 is the d-spacing of peak1, d2 is the d-spacing of peak 2, where Z > 0 indicates Ia and Z < 0 indicates Ib dominant type, [3]. The results are presented in Table 2. As can be seen, the Z-values for the sulfite and kraft pulps indicate that the cellulose samples are Ib dominant type. It was observed that the conversion of cellulose Ia to cellulose Ib in the sulfite pulp (CPT) is more pronounced than in the kraft pulp (CEG), given the more negative Z-value for the sulfite pulp, as can be seen in Table 2. These results suggest that the sulfite treatment degrades triclinic cellulose Ia more easily than monoclinic cellulose Ib [3]. The values for the d-spacings and crystallinity indices obtained for the two pulps were similar. The crystallinity index (CrI) shows slight differences in the crystallinity between the two cellulose samples. However, the d-spacing value for CPT in (11 0) was around 20% higher than for CEG. The increase in the crystallite size for CPT in (11 0), might be associated with a reduction in the corresponding amorphous region [2,12]. If the amorphous domains of cellulose are attacked during the pulping treatment, chain scission and peeling
The degradation of cellulose occurs through a series of complex chemical reactions. The degradation mechanism involves two competitive reactions. The first path of low activation energy leads to the production of carbonaceous residues via cellulose dehydration reactions. The second path of high activation energy, called depolymerization, initially involves a reduction in the polymerization degree of the cellulose and then the small chains undergo transglycosylation leading to tars or levoglucosan, and their subsequent decomposition [34,35]. Fig. 3 shows the TGA and DTG curves of the two cellulose samples using a heating rate of 10 C min1. A weight loss of around 5% occurs at temperatures of 25 Ce150 C for the two samples. This corresponds to the vaporization and removal of bound water in the cellulose [11,36,37]. The main decomposition step occurs in the range of 240 Ce370 C for CEG and 250 Ce375 C for CPT. In this stage the cleavage of the glycosidic linkages of cellulose reduces the polymerization degree leading to the formation of CO2, H2O and other hydrocarbon derivatives [37]. Differences in the decomposition profiles of the two cellulose samples, according to Fig. 3, indicate the slight differences in the thermal stability of the samples. The DTG peaks were centered at 353 C and 360 C for CEG and CPT, respectively, as presented in Fig. 3. The DTG curve of CPT was shifted to higher temperatures with increasing crystallite size. This behavior suggests that celluloses with higher crystallite size have higher thermal stability. Kim et al. [12] studied different types of cellulose and noted an increase in the crystallite size promoted by higher thermal stability. 3.4. Kinetics results In order to obtain the kinetics parameters studied, the conversion (a) was determined according to methods given in the literature [8,11,36]. Figs. 4 to 6 show the behavior of the CEG sample. Since the behavior is similar for the two samples, only one is presented as a model representative of the methods used. Fig. 4 shows the curves of a as a function of temperature for the different heating rates studied. As can be observed in Fig. 4, with an increase in the heating rate the degradation curves are shifted to higher temperatures, which corresponds to shorter time intervals. The Avrami plots for the CEG samples are shown in Fig. 5. From the linear fit the kinetics parameters were obtained, that is, the rate constant k0 , the Avrami exponent n and the half-life time t1/2 [27]. The results obtained are given in Table 3. It was observed from analysis of the results from the Avrami model that the k’ value increases with an increase in the heating rate, i.e., the average reaction rate increases when higher temperatures are provided in shorter time intervals [27]. The time required for 50% of the reactions to occur (t1/2) [27] decreases with an increase in the heating rate (b). Since t1/2 is dependent on k0 , it can be noted that an increase in k0 contributes to a reduction in t1/2,
Table 2 Parameters obtained from the XRD analysis for the samples studied. Sample
d (nm)
d (nm)
L (nm)
L (nm)
L (nm)
(11 0)
d (nm)
(110)
(200)
(11 0)
(110)
(200)
0.618 0.605
0.553 0.555
0.397 0.399
3.783 4.731
2.370 2.370
3.826 3.825
CEG CPT
CrI (%)
Z-values
74.9 1.2 75.5 1.2
1.532 25.345
M. Poletto et al. / Polymer Degradation and Stability 96 (2011) 679e685
Fig. 3. TGA and DTG curves obtained in the study of the cellulose degradation at a heating rate of 10 C min1.
683
Fig. 5. Representation of the results obtained from the Avrami equation for the CEG samples.
which may be related to the heat diffusion through the complex structure of cellulose. The k0 and t1/2 results for CEG and CPT were similar. The two samples also showed a value for the Avrami exponent of n z 3. For n z 3 the effect analogous to crystallization described by Avrami would be degradation controlled by diffusion in three dimensions [23e25], in agreement with reports in the literature [8]. However, there is a distinction between the two samples studied. With an increase in b the CEG sample showed a tendency toward a reduction in n, in contrast to the behavior observed for the CPT sample, i.e., with an increase in b there was an increase in the exponent n. Thus, for CEG the degradation process will be forced at the crystal vertices and, analogously to crystallization, the exponent n tends toward 2, indicating that the degradation of the crystals may occur sporadically from the nucleus [38]. This difference may also be associated with the difference between the crystallite size in the 11 0 plane, as shown in Table 2. Fig. 6 shows the values in the conversion range of 0.2e0.8 for the determination of the Ea values obtained using the method proposed by FlynneWalleOzawa (FWO) [13,14].
The linear fits obtained from the plot of b vs. 1/T [13,14] are shown in Fig. 6. The linear correlation coefficients (r) were close to unity (minimum ¼ 0.9878 and maximum ¼ 0.9998). Therefore, the Ea values were calculated through the angular coefficient obtained from the straight line fit. Table 4 shows the Ea values for the cellulose samples studied. It can be observed that both samples showed a decrease in the Ea values in the range of a ¼ (0.2e0.8) with 209.47e157.79 and 173.30e138.47 kJ mol1 for the CEG and CPT samples, respectively. This range of values is consistent with values reported in the literature [9e11]. However, as can be observed, the CEG sample had higher Ea values than the CPT samples. According to Kim et al. [12] higher crystal sizes promote an increase in the degradation temperature; consequently the CPT cellulose fiber may have a higher thermal stability than the CEG one. Also, the authors report that the activation energy is not affected by the crystallite size. Moreover, the finding that CPT had the lowest activation energy values was attributed to the thermal decomposition of sulfite cellulose being controlled by dehydration, whereas the high Ea values for CEG may indicate depolymerization of the kraft cellulose
Fig. 4. Curves of the conversion (a) as a function of temperature (T/ C) for the different heating rates studied.
Fig. 6. Data obtained from the plot of log b vs. 1/T (K1) using the FWO method for the CEG sample.
684
M. Poletto et al. / Polymer Degradation and Stability 96 (2011) 679e685
Table 3 Kinetics parameters obtained through the linear fit of the Avrami equation. Sample
b ( C.min1)
k0 (min1)
t ½ (min)
n
r
CEG
5 10 20 40
0.04 0.07 0.15 0.29
2.25 1.94 1.62 1.36
3.67 3.42 3.24 2.87
0.999 1.000 0.998 1.000
CPT
5 10 20 40
0.03 0.07 0.15 0.31
3.39 2.51 1.62 1.27
2.51 2.55 3.22 3.34
0.999 1.000 0.998 0.998
with the production of levoglucosan. Furthermore, the decrease in Ea at a values of 0.2e0.8 observed for the two cellulose samples might be the result of autocatalysis in the dehydration process [35]. Thus, a mixed depolymerization and dehydration mechanism can be considered for the two cellulose samples, independently of the cellulose treatment. Similar behavior was observed by Soares et al. [35] for cellulose powder and kraft paper. Gümüskaya et al. [1] and Ouajai and Shanks [38], describe that different pulping conditions can affect the cellulose crystallinity. Thus, by investigating the thermal stability [12], it is possible to evaluate the influence of different treatments on the cellulose crystalline structure and how it affects the degradation kinetics and degradation mechanism. The Ea results obtained using the FWO method were used to determine the degradation mechanisms proposed by Criado et al. [18]. This method uses reference theoretical curves obtained from equation (11) that are derivatives of the f(a) and g(a) functions represented in Table 1, called master curves, which are compared to experimental data for determination of the mechanism of a solidstate process [21,22]. These mechanisms represent how the solidstate degradation process occurs. The algebraic expressions that represent the theoretical mechanisms are separated into four groups, An Rn Dn and Fn, as can be seen in Table 1. These mechanisms describe processes of nuclei formation on the propagation of the degradation process; diffusion processes that are related to the heat transfer capacity along the material structure; reaction mechanisms controlled by the surface of the sample; and the random degradation of nuclei, respectively. Figs. 7 and 8 present the master curves as well as the results of the experimental data obtained. The determination of the experimental Z(a) values was carried out using a heating rate (b) of 10 C min1, and the Ea values obtained applying the FWO method were used to calculate Z(a) in equation (12). The physical meaning of the function Z(a) varies depending on the theoretical approach to the mechanisms, as shown in Table 1. The experimental data for the CEG sample in the conversion range of a ¼ 0.2e0.4 were superimposed on the D1, D2 and D3 curves and according to the literature these degradation mechanisms refer to the diffusion processes in one, two and three dimensions, respectively [18,21]. Similar results were described by
Fig. 7. Master curves and experimental data obtained using the method of Criado et al. for the CEG sample.
Wu and Dollimore [8]. Subsequently, from a ¼ 0.5 the mechanism tends toward F1, corresponding to random nucleation with one nucleus in the individual particle [18,21]. These phenomena are in agreement with the values for the Avrami exponent given in Table 3, where the samples underwent degradation by diffusion in three dimensions. The Avrami exponent, in its original derivation is associated with the dimensional crystal growth processes. In this case, it is important to mention that the physical meaning of n used in this study is related to the degradation kinetics, i.e., the crystalline structure does not change with increasing b. Thus, n refers only to the difficulty of heating transfer by varying b. Changes in the Avrami exponent from 3 to 2 (according to Table 3) for the CEG cellulose fiber may be related to a greater heat transfer difficulty in transferring heat. This is due to the type of packing between the cellulose chains, since the characteristics associated with n suggest a more ordered structural form [39] on increasing the heating rate. This factor could result in an increase in the Ea values, as shown in Table 4 for the CEG cellulose fiber. The reported characteristic of n (n z 2) suggests diffusion processes [23e25,39], and this is in agreement with the diffusion process mechanism demonstrated (D1, D2, D3) in Fig. 7. For CPT, for a values in the range of 0.3e0.7 the degradation mechanism corresponded to R1, i.e., the phase boundary-controlled
Table 4 Activation energies obtained using the Flynn-Wall-Ozawa method. CEG
CPT
a
Ea (kJ mol1)
r
Ea (kJ mol1)
r
0.2 0.3 0.4 0.5 0.6 0.7 0.8
209.47 199.82 190.29 181.75 171.66 164.36 157.79
0.9984 0.9995 0.9998 0.9998 0.9997 0.9996 0.9993
173.30 167.27 160.89 156.09 151.30 143.91 138.47
0.9991 0.9990 0.9993 0.9990 0.9994 0.9989 0.9991
Fig. 8. Master curves and experimental data obtained using the method of Criado et al. for the CPT sample.
M. Poletto et al. / Polymer Degradation and Stability 96 (2011) 679e685
reaction (one-dimensional movement) [22], in agreement with the results obtained by Wu and Dollimore [8]. The R1 mechanism is also consistent with the results obtained from the XRD analysis, where the larger crystallite size for CPT might be related to the action of temperature on the boundaries due to the greater interfacial perimeter between crystals. 4. Conclusions Two distinct types of cellulose fibers obtained using different processes were characterized. The IR spectra showed a similar composition for the two samples. The XRD analysis demonstrated that, although the samples have close crystallinity indices, the crystallite size of the CPT cellulose obtained by the sulfite process in one of its orientations (11 0) is around 20% greater than that of the CEG cellulose obtained by the kraft process, which may be associated with a reduction in the corresponding amorphous region by the sulfite process. The thermogravimetric analysis showed that the CPT degradation curve is shifted to higher temperatures in the case of CEG, suggesting that the greater thermal stability is due to the larger crystallite size. These results are in agreement with the XRD results obtained. However, higher activation energy values were observed for the CEG cellulose fiber. This observation is consistent with reports in the literature. Through the Avrami parameters it was found that there are differences between the degradation processes of the cellulose pulps. This difference is related to the packing of the cellulose chains. To better understand this behavior, results obtained through the Criado method for the CEG cellulose fiber confirmed that the degradation occurs by diffusion, D1, D2 and D3, and that the heat transfer is affected by the type of arrangement and packing of cellulose crystals obtained by the kraft process. This characteristic could explain the higher Ea values obtained for CEG in comparison to CPT. Moreover, the degradation mechanism observed for CPT applying the Criado method is the phase boundary controlled reaction (R1). This supports the fact that the greater thermal stability of CPT obtained by the sulfite process is related to the increased crystallite size. Therefore, crystallinity, crystallite size, thermal stability and cell structure were more affected during kraft pulping than under the sulfite pulping conditions. This can be explained by the fact that in kraft pulping there is a more degraded cellulose structure due to the cooking process conditions, in comparison to the sulfite process. Acknowledgment The authors are grateful to Cambará S.A and CMPC S.A. for supplying the cellulose fibers and CAPES, CNPq and FAPERGS for financial support. The authors also thank Ms. Heitor L. Ornaghi Jr. for the suggestions given in this work. References [1] Gümüskaya E, Usta M, Kirei H. The effects of various pulping conditions on crystalline structure of cellulose in cotton linters. Polym Degrad Stab 2003;81: 559e64. [2] Duchesne I, Hult EL, Molin U, Daniel G, Iversen T, Lennhon H. The influence of hemicellulose on fibril-aggregation of kraft pulp fibres as revealed by FE-SEM and CP/MAS 13C-NMR. Cellulose 2001;8:103e11. [3] Wada M, Okano T. Localization of Ia and Ib phases in algal cellulose revealed by acid treatments. Cellulose 2001;8:183e8. [4] Hult EL, Iversen T, Sugiyama J. Characterization of the supermolecular structure of cellulose in wood pulp fibres. Cellulose 2003;10:103e10.
685
[5] Sun JX, Sun XF, Zhao H, Sun RC. Isolation and characterization of cellulose from sugarcane bagasse. Polym Degrad Stab 2004;84:331e9. [6] Neagu RC, Gamstedt EK, Berthold F. Stiffness contribution of various wood fibers to composite materials. J Comp Mater 2006;40:663e99. [7] Hult EL, Larsson PT, Iversen T. Cellulose fibril aggregationean inherent property of kraft pulps. Polymer 2001;42:3309e14. [8] Wu Y, Dollimore D. Kinetic studies of thermal degradation of natural cellulosic materials. Thermochim Acta 1998;324:49e57. [9] Nada AMA, Hassan ML. Thermal behavior of cellulose and some cellulose derivatives. Polym Degrad Stab 2000;67:111e5. [10] Antal MJ, Várhegyi G, Jakab E. Cellulose pyrolysis kinetics: revisited. Ind Eng Chem Res 1998;34:1267e75. [11] Capart R, Khezami L, Burnham AK. Assessment of various kinetic models for the pyrolysis of a microgranular cellulose. Thermochim Acta 2004;417:79e89. [12] Kim UJ, Eom SH, Wada M. Thermal decomposition of native cellulose: influence on crystallite size. Polym Degrad Stab 2010;95:778e81. [13] Flynn JH, Wall LA. General treatment of the thermogravimetry of polymers. J Res Natl Bur Stand 1966;70A:487e523. [14] Ozawa T. A new method of quantitative differential thermal analysis. Bull Chem Soc Jpn 1966;39:2071e85. [15] Pistor V, Ornaghi FG, Ornaghi Jr HL, Fiorio R, Zattera AJ. Thermochim Acta 2010;510:93e6. [16] Poletto M, Dettenborn J, Pistor V, Zeni M, Zattera AJ. Materials produced from plant biomass. Part I: evaluation of thermal stability and pyrolysis of wood. Mater Res 2010;13:375e9. [17] Doyle CD. Kinetic analysis of thermogravimetric data. J Appl Polym Sci 1961;5: 285e92. [18] Criado JM, Malek J, Ortega A. Applicability of the master plots in kinetic analysis of non-isothermal data. Thermochim Acta 1989;147:377e85. [19] William LP. Computation of the exponential trap population integral of glow curve theory. J Comp Phys 1971;7:187e90. [20] Pérez-Maqueda LA, Criado JM. The accuracy of Senum and Yang’s approximations to the Arrhenius integral. J Therm Anal Calorim 2000;60:909e15. [21] Tiptipakorn S, Damrongsakkul S, Ando S, Hemvichian K, Rimdusi S. Thermal degradation behaviors of polybenzoxazine and silicon-containing polyimide blends. Polym Degrad Stab 2007;92:1265e78. [22] Núñez L, Villanueva M, Núñez MR, Rial B. Influence of an epoxy reactive diluent on the thermal degradation process of the system DGEBA n ¼ 0/1,2 DCH. J Appl Polym Sci 2004;92:1199e207. [23] Avrami M. Kinetics of phase change. I. J Chem Phys 1939;7:1103e12. [24] Avrami M. Kinetics of phase change. II. J Chem Phys 1940;8:212e24. [25] Avrami M. Kinetics of phase change. III. J Chem Phys 1941;9:177e84. [26] Jeziorny A. Parameters characterizing the kinetics of the non-isothermal crystallization of poly(ethylene terephthalate) determined by d.s.c. Polymer 1978;19:1142e4. [27] Liu TX, Mo ZS, Wang SG, Zhang HF. Nonisothermal melt and cold crystallization kinetics of poly(aryl ether ether ketone ketone). Polym Eng Sci 1997; 37:568e75. [28] Tserki V, Matzinosa P, Kokkoub S, Panayiotou C. Novel biodegradable composites based on treated lignocellulosic waste flour as filler. Part I. Surface chemical modification and characterization of waste flour. Composites: Part A 2005;36:965e74. [29] Viera RGP, Filho GR, de Assuncão RMN, Meireles CS, Vieira JG, de Oliveira GS. Synthesis and characterization of methylcellulose from sugar cane bagasse cellulose. Carbohydr Polym 2007;67:182e9. [30] Xiao B, Sun XF, Sun RC. Chemical, structural, and thermal characterizations of alkali-soluble lignins and hemicelluloses, and cellulose from maize stems, rye straw, and rice straw. Polym Degrad Stab 2001;74:307e19. [31] de Menezes AJ, Pasquini D, da Silva Curvelo AA, Gandini A. Self-reinforced composites obtained by the partial oxypropylation of cellulose fibers. 1. Characterization of the materials obtained with different types of fibers. Carbohydr Polym 2009;76:437e42. [32] Colom X, Carrasco F, Pagès P, Cañavate J. Effects of different treatments on the interface of HDPE/lignocellulosic fiber composites. Comp Sci Tech 2003;63: 161e9. [33] Colom X, Carrillo F, Nogués F, Garriga P. Structural analysis of photodegraded wood by means of FTIR spectroscopy. Polym Degrad Stab 2003;80:543e9. [34] Quiévy N, Jacquet N, Sclavons M, Deroanne C, Paquot M, Devaux J. Influence of homogenization and drying on the thermal stability of microfibrillated cellulose. Polym Degrad Stab 2010;95:306e14. [35] Soares S, Camino G, Levchik S. Comparative study of the thermal decomposition of pure cellulose and pulp paper. Polym Degrad Stab 1995;49:275e83. [36] Yao F, Wu Q, Lei Y, Guo W, Xu Y. Thermal decomposition kinetics of natural fibers: activation energy with dynamic thermogravimetric analysis. Polym Degrad Stab 2008;93:90e8. [37] Elanthikkal S, Gopalakrishnapanicker U, Varghese S, Guthrie JT. Carbohydr Polym 2010;80:852e9. [38] Ouajai S, Shanks RA. Composition, structure and thermal degradation of hemp cellulose after chemical treatments. Polym Degrad Stab 2005;89:327e35. [39] Sperling LH. Introduction to physical polymer science. New Jersey: John Wiley & Sons; 2006.
Contents lists available at ScienceDirect
Polymer Degradation and Stability journal homepage: www.elsevier.com/locate/polydegstab
Crystalline properties and decomposition kinetics of cellulose fibers in wood pulp obtained by two pulping processes Matheus Poletto*, Vinícios Pistor, Mara Zeni, Ademir J. Zattera* Laboratory of Polymers (LPOL), Center of Exact Sciences and Technology (CCET), Caxias do Sul University (UCS), Caxias do Sul, RS, Brazil
a r t i c l e i n f o
a b s t r a c t
Article history: Received 24 May 2010 Received in revised form 3 December 2010 Accepted 8 December 2010 Available online 16 December 2010
In this study two cellulose fibers, Eucalyptus grandis (CEG) and Pinus taeda (CPT), obtained through the kraft and sulfite pulping processes, respectively, were characterized. Fourier transform infrared (FTIR) spectroscopy, X-ray diffraction (XRD) and thermogravimetric analysis (TGA) were carried out. From the XRD analysis the interplanar distance, crystallite size and crystallinity index were calculated and the degradation kinetics parameters were determined by TGA at heating rates of 5, 10, 20 and 40 C min1 using the Avrami, Flynn-Wall-Ozawa (FWO) and Criado methods. The results obtained by FTIR showed that the composition of the fibers is similar, while from the XRD analysis slight differences in the crystallinity were observed. The thermogravimetric analysis showed higher thermal stability for CPT than CEG while the values for the activation energy (Ea) were higher for CEG than CPT. The results obtained by Avrami and Criado methods showed that the degradation mechanism in the CEG samples involves a diffusion process while in the case of CPT the degradation process is a phase boundary controlled reaction. The degradation mechanisms demonstrated that the difference between thermal stability and Ea may be due to differences in the type of crystalline structure of the samples obtained through the two pulping processes. Ó 2010 Elsevier Ltd. All rights reserved.
Keywords: Eucalyptus grandis Pinus taeda Cellulose Crystallinity Kinetics Degradation
1. Introduction Cellulose is a polymer which contains crystallites and thus it has a paracrystalline morphology [1,2]. The linear cellulose molecules are linked laterally by hydrogen bonds to form linear bundles, leading to a crystalline structure [1]. Native cellulose is known to be a composite of two distinct crystalline modifications, namely Ia and Ib, whose fractions vary depending on the origin of the cellulose sample [2e4]. The Ia and Ib structures are assigned to one-chain triclinic and two-chain monoclinic unit cells, respectively [3]. These allomorphs are believed to coexist in the fibril in different ratios [4]. In addition, Ia cellulose is reported to be the dominant polymorph in bacterial and alga celluloses, while Ib cellulose is predominant in higher plants such as cotton and wood [5]. Therefore, the crystalline structure of cellulose affects the physical and mechanical properties of the cellulose fibers. The degree of crystallinity of cellulose is one of the most important crystalline structure parameters, and the rigidity of cellulose fibers increases and the flexibility decreases with an increasing ratio of crystalline to amorphous regions [1,6].
* Corresponding author. Tel.: þ55 54 3218 2108; fax: þ55 54 3218 2253. E-mail addresses: [email protected] (M. Poletto), [email protected] (A.J. Zattera). 0141-3910/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.polymdegradstab.2010.12.007
Changes occurring in the cellulose structure during pulping have been investigated in several studies [1,3,4,7]. Gümüskaya et al. [1] concluded that not only chemical agents but also the temperature and pressure in a cooking digester had major effects on the crystalline structure of cellulose in cotton linters during cooking. Hult et al. [4] investigated the organization of cellulose microfibrils in holocellulose, sulfite pulp and kraft pulp and revealed that in kraft pulp the microfibrils are more closely associated than in the sulfite pulp and holocellulose. Due to the complexity of cellulose thermal decomposition reactions, extensive research has been carried out in this area. Wu and Dollimore [8] investigated the thermal degradation behavior of natural cellulosic materials and demonstrated that the ratecontrolling mechanism involved mainly the phase boundary and probably diffusion processes. Nada and Hassan [9] investigated cellulose and some cellulose derivatives and reported activation energy values of 53e182 kJ/mol. Antal et al. [10] using dynamic thermogravimetric analysis found activation energy values of between 190 and 250 kJ/mol, depending on the type of cellulose, the heating rate and the mass of sample. Capart et al. [11] calculated kinetics parameters of microgranular cellulose using dynamic and isothermal methods in a nitrogen atmosphere and described two reactions with activation energies of 202 and 255 kJ/mol, respectively. The aim of this study was to investigate the effects of
680
M. Poletto et al. / Polymer Degradation and Stability 96 (2011) 679e685
chemical treatments on the crystalline properties and decomposition kinetics of cellulose fibers in wood pulp during two pulping processes. 2. Materials and methods 2.1. Materials Bleached sulfite cellulose fibers from Pinus taeda (CPT) were supplied by Cambará S.A (Cambará do Sul, Brazil). Bleached kraft cellulose fibers from Eucalyptus grandis (CEG) were supplied by CMPC S.A. (Guaíba, Brazil). The samples were dried at 70 C for 24 h in a vacuum oven before the tests. The average fiber particle length for CTP and CEG is around 150 mm. 2.2. Fourier transform infrared (FTIR) spectroscopy The Fourier transform infrared (FTIReNicolet IS10-Thermo Scientific) spectroscopy analysis was carried out with 32 scans, in the range of 4000 cm1 to 400 cm1, at a resolution of 4 cm1 using attenuated total reflectance (ATR), on cellulose powder under pressure on ATR crystal. 2.3. Thermogravimetric analysis (TGA) The thermogravimetric analysis (TGA50eShimadzu) was carried out under N2 atmosphere, from 25 up to 610 C. Approximately 20 mg of each sample was used. The analysis was carried out at four different heating rates (5, 10, 20 and 40 C min1). The results obtained were used to calculate the kinetics parameters. 2.4. X-ray diffraction (XRD) X-ray diffractograms were collected using a sample holder mounted on a Shimadzu diffractometer (XRD-6000), with monochromatic Cu Ka radiation (l ¼ 0.15418 nm) and the generator working at 40 kV and 30 mA. Intensities were measured in the range of 5 < 20 < 30 , typically with scan steps of 0.05 and 2 s/step (1.5 min1). Peak separations were carried out using Gaussian deconvolution. The determination coefficients (r2) were close to unity (0.9972 and 0.9947 for CEG and CPT, respectively). The dspacings were calculated using the Bragg equation [3,12], the crystallite sizes (L) were calculated using the Scherrer equation [3,12] and the crystallinity index (CrI) using the Segal method [1,3]. 2.5. Theoretical Considerations The fundamental equation used in all kinetic studies is generally described as:
da ¼ kðTÞf ðaÞ dt
(1)
where k is the rate constant and f (a) is the reaction model, a function dependent on the reaction mechanism. Eq (1) expresses the rate of conversion, da/dt, at a constant temperature as a function of the reduction in reactant concentration and the rate constant. In this study, the conversion a rate is defined as [13e16]:
a ¼
m0 mt m0 mf
(2)
where m0, mf and mt are the initial and final weights of the sample and its weight at time (t), respectively. The rate constant k is generally given by the Arrhenius equation:
Ea
kðTÞ ¼ Ae RT
(3)
where Ea is the apparent activation energy (kJ mol1), R is the gas constant (8.314 kJ mol1), A is the pre-exponential factor (min1), and T is the absolute temperature (K). The combination of Eqs. (1) and (3) gives the following relationship:
da ¼ AeEa =RT f ðaÞ dt
(4)
For a dynamic TGA process in a non-isothermal experiment, introducing the heating rate, b ¼ dT/dt, into Eq. (4), Eq. (5) is obtained as:
da ¼ dt
A Ea =RT e f ðaÞ
b
(5)
Equations (4) and (5) are the fundamental expressions of analytical methods used to calculate kinetic parameters on the basis of TGA data. 2.6. Flynn-Wall-Ozawa method In the FWO method [13,14], it is proposed that through the integration of Equation (5) and substituting (Ea/RT) with x, the following relationship can be obtained:
Za 0
da AE PðxÞ ¼ gðaÞ ¼ bR f ðaÞ
(6)
were P(x) is a function known as the Arrhenius integral that has no analytical solution but can be solved by numerical methods or using different approaches. The test method is based on integral approaches proposed by Doyle [17], in a range of logP(x)/x¼(Ea/RT) to 20 < x < 60. The integral of P(x) can be expressed simplistically as:
ln b ¼ ln
AE E 5:330 1:052 gðaÞR RT
(7)
In Eq. (7) g(a) is a function of the conversion. According to the principles of the isoconversional FWO method, it is assumed that the reaction rate at a given conversion is a function only of the temperature. Therefore, for different heating rates (b) and a given degree of conversion (a), a linear relationship is observed through a plot of log b vs. 1/T, and the apparent activation energy (Ea) is obtained from the slope of the straight line obtained [13e15]. 2.7. Criado method The degradation reaction mechanism can be determined using the Criado method [18e20] which can accurately determine the reaction mechanism in a solid reaction process, defined by a Z(a)type function:
ZðaÞ ¼
da=dt
b
pðxÞT
(8)
where x ¼ E/RT and p(x) is an approximation of the temperature integral which cannot be expressed in a simple analytical form. Paterson [19] proposed a reasonable relationship between p(x) and P(x) as shown in Eq. (9):
pðxÞ ¼ xex PðxÞ
(9)
M. Poletto et al. / Polymer Degradation and Stability 96 (2011) 679e685
681
Senum and Yang [20] proposed the fourth rational expression of P(x):
PðxÞ ¼
expðxÞ x3 þ 18x2 þ 86x þ 96 x x4 þ 20x3 þ 120x2 þ 240x þ 120
(10)
When x > 20, the error of Eq. (10) is less than 105%, which forms the basis of the analysis presented in this paper. Combining equations (1), (8) and (9) we obtain:
ZðaÞ ¼ f ðaÞgðaÞ
(11)
From equations (1), (8) and (9), the following relationship can be derived:
ZðaÞ ¼
da Ea Ea eRT PðxÞ dT R
(12)
Equation (11) is used to plot the master Z(a) vs. a curves for the different models listed in Table 1 [21,22], whereas Eq. (12) is used to represent the experimental curve. By comparing these two curves, the type of mechanism involved in the thermal degradation can be identified.
corrected kinetic constant (k0 ) through the use of the following equation:
2.8. Avrami analysis According to the Avrami model [23e25], under isothermal conditions, the degree of phase conversion (a(t)) is given by:
aðtÞ ¼ 1 ekt
n
(14)
According to Eq. (14), the parameters n and k are obtained, respectively, from the slope and the intercept of the straight line obtained from the plot of ln[-ln(1-a)] vs.lnt. Jeziorny [26] modified the Avrami equation. In Jeziorny’s analysis, the kinetic constant (k0 ) is determined in the same way as in the Avrami equation. However, a is now a function of temperature (a (T)). The corrected kinetic constant (k’) as a function of the Avrami kinetic constant (k) and the heating rate (b) is given as follows:
lnk0 ¼
. lnk b
t1=2 ¼
. 1 n ln2 k0
(16)
(13)
where k is the Avrami constant and n is the Avrami exponent which is dependent on the process dimensionalities. The parameters n and k can be determined by using the Avrami equation in the double logarithmic form:
ln½ lnð1 aðtÞÞ ¼ lnk þ nlnt
Fig. 1. IR spectra obtained for the cellulose samples studied.
(15)
The reaction half-life time (t1/2), which is defined as the time in which 50% of the reaction occurs [27], can be calculated from the
3. Results and discussions 3.1. Fourier transform infrared (FTIR) spectroscopy The IR spectra given in Fig. 1 show the composition of the cellulose samples studied. For the two samples a broad band can be observed in the region of 3600e3200 cm1, attributed to stretching of the hydroxyl (OH) groups [28,29], and 3000e2800 cm1, attributed to CH stretching [28e31]. According to Xiao [30] and Colom (a) [32], the band observed in the range of 1700e1550 cm1 may be associated with water absorption. The bands at 1430, 1365, 1335 and 1315 cm1 are attributed to CH2, in-plane CH deformation, in-plane OH bending and CH2 wagging, respectively [28e33], and the bands at 1160, 1110, 1052, 1028, 895 and 670 cm1 are associated with the asymmetric CeOeC bridge stretching, the anhydroglucose ring, C-OR stretching, CeOeC pyranose ring skeletal vibration, CH deformation and out-of-plane OH bending, respectively [28e33]. As can be observed, the spectra show the same structural characteristics. This suggests,
Table 1 Algebraic expressions for g(a) and f(a) for the most frequently used mechanisms of solid-state processes. Mechanism e Solid state process
g(a)
f(a)
A2 - Nucleation and growth (Avrami eq.1) A3 - Nucleation and growth (Avrami eq.2) A4 - Nucleation and growth (Avrami eq.3) R1 - Phase boundary controlled reaction (one-dimensional movement) R2 - Phase boundary controlled reaction (contracting area) R3 - Phase boundary controlled reaction (contracting volume) D1 - One-dimensional diffusion D2 - Two-dimensional diffusion (Valensi equation) D3 - Three-dimensional diffusion (Jander equation) D4 - Three-dimensional diffusion (GinstlingeBrounshtein equation) F1 - Random nucleation with one nucleus on the individual particle F2 - Random nucleation with two nuclei on the individual particle F3 - Random nucleation with three nuclei on the individual particle
[ln (1 a)]1/2 [ln (1 a)]1/3 [ln (1 a)]1/4
2 (1 a)[ln (1 a)]1/2 3 (1 a)[ln (1 a)]2/3 4 (1 a)[ln (1 a)]3/4 1 2 (1 a)1/2 3 (1 a)2/3 (1/2)a1 [ln (1 a)]1 (3/2)[1 (1 a)1/3]1 (1 a)2/3 (3/2)[1 (1 a)1/3]1 1a (1 a)2 (1/2)(1 a)3
a
[1ln (1 a)]1/2 [1ln (1 a)]1/3
a2
(1 a)ln (1 a) þ a [1 (1 a)1/3]2 [1 (2/3)a] (1 a)2/3 ln (1 a) 1/(1 a) 1/(1 a)2
682
M. Poletto et al. / Polymer Degradation and Stability 96 (2011) 679e685
reactions can occur, which reduce the total amount of amorphous region and therefore increase the crystallite size of CPT. 3.3. Thermogravimetric analysis (TGA)
Fig. 2. Diffractograms obtained for the cellulose samples studied.
in a preliminary qualitative analysis, that the CEG and CPT samples are similar in terms of composition. 3.2. X-ray diffraction (XRD) Fig. 2 shows the X-ray diffraction analysis of the cellulose samplesused in this study. The details for the three peaks are: peak 1 is Ib (11 0); peak 2 is Ib (110) and peak 3 is Ib (200) [4]. These peaks are separated by Gaussian deconvolution, and the d-spacings, crystallite size, crystallinity indices and Z-values were calculated. In this study, the Z-discriminant function developed by Wada and Okano [3] for the determination of the crystalline structure (monoclinic and triclinic) of cellulose in pulp samples was used. By employing discriminant analysis it is possible to categorize cellulose as type Ia or Ib [3]. The Z-value indicates whether cellulose is Ia or Ib [3]. The function which discriminates between Ia or Ib is given by Z ¼ 1693 d1-902 d2-549 d1 is the d-spacing of peak1, d2 is the d-spacing of peak 2, where Z > 0 indicates Ia and Z < 0 indicates Ib dominant type, [3]. The results are presented in Table 2. As can be seen, the Z-values for the sulfite and kraft pulps indicate that the cellulose samples are Ib dominant type. It was observed that the conversion of cellulose Ia to cellulose Ib in the sulfite pulp (CPT) is more pronounced than in the kraft pulp (CEG), given the more negative Z-value for the sulfite pulp, as can be seen in Table 2. These results suggest that the sulfite treatment degrades triclinic cellulose Ia more easily than monoclinic cellulose Ib [3]. The values for the d-spacings and crystallinity indices obtained for the two pulps were similar. The crystallinity index (CrI) shows slight differences in the crystallinity between the two cellulose samples. However, the d-spacing value for CPT in (11 0) was around 20% higher than for CEG. The increase in the crystallite size for CPT in (11 0), might be associated with a reduction in the corresponding amorphous region [2,12]. If the amorphous domains of cellulose are attacked during the pulping treatment, chain scission and peeling
The degradation of cellulose occurs through a series of complex chemical reactions. The degradation mechanism involves two competitive reactions. The first path of low activation energy leads to the production of carbonaceous residues via cellulose dehydration reactions. The second path of high activation energy, called depolymerization, initially involves a reduction in the polymerization degree of the cellulose and then the small chains undergo transglycosylation leading to tars or levoglucosan, and their subsequent decomposition [34,35]. Fig. 3 shows the TGA and DTG curves of the two cellulose samples using a heating rate of 10 C min1. A weight loss of around 5% occurs at temperatures of 25 Ce150 C for the two samples. This corresponds to the vaporization and removal of bound water in the cellulose [11,36,37]. The main decomposition step occurs in the range of 240 Ce370 C for CEG and 250 Ce375 C for CPT. In this stage the cleavage of the glycosidic linkages of cellulose reduces the polymerization degree leading to the formation of CO2, H2O and other hydrocarbon derivatives [37]. Differences in the decomposition profiles of the two cellulose samples, according to Fig. 3, indicate the slight differences in the thermal stability of the samples. The DTG peaks were centered at 353 C and 360 C for CEG and CPT, respectively, as presented in Fig. 3. The DTG curve of CPT was shifted to higher temperatures with increasing crystallite size. This behavior suggests that celluloses with higher crystallite size have higher thermal stability. Kim et al. [12] studied different types of cellulose and noted an increase in the crystallite size promoted by higher thermal stability. 3.4. Kinetics results In order to obtain the kinetics parameters studied, the conversion (a) was determined according to methods given in the literature [8,11,36]. Figs. 4 to 6 show the behavior of the CEG sample. Since the behavior is similar for the two samples, only one is presented as a model representative of the methods used. Fig. 4 shows the curves of a as a function of temperature for the different heating rates studied. As can be observed in Fig. 4, with an increase in the heating rate the degradation curves are shifted to higher temperatures, which corresponds to shorter time intervals. The Avrami plots for the CEG samples are shown in Fig. 5. From the linear fit the kinetics parameters were obtained, that is, the rate constant k0 , the Avrami exponent n and the half-life time t1/2 [27]. The results obtained are given in Table 3. It was observed from analysis of the results from the Avrami model that the k’ value increases with an increase in the heating rate, i.e., the average reaction rate increases when higher temperatures are provided in shorter time intervals [27]. The time required for 50% of the reactions to occur (t1/2) [27] decreases with an increase in the heating rate (b). Since t1/2 is dependent on k0 , it can be noted that an increase in k0 contributes to a reduction in t1/2,
Table 2 Parameters obtained from the XRD analysis for the samples studied. Sample
d (nm)
d (nm)
L (nm)
L (nm)
L (nm)
(11 0)
d (nm)
(110)
(200)
(11 0)
(110)
(200)
0.618 0.605
0.553 0.555
0.397 0.399
3.783 4.731
2.370 2.370
3.826 3.825
CEG CPT
CrI (%)
Z-values
74.9 1.2 75.5 1.2
1.532 25.345
M. Poletto et al. / Polymer Degradation and Stability 96 (2011) 679e685
Fig. 3. TGA and DTG curves obtained in the study of the cellulose degradation at a heating rate of 10 C min1.
683
Fig. 5. Representation of the results obtained from the Avrami equation for the CEG samples.
which may be related to the heat diffusion through the complex structure of cellulose. The k0 and t1/2 results for CEG and CPT were similar. The two samples also showed a value for the Avrami exponent of n z 3. For n z 3 the effect analogous to crystallization described by Avrami would be degradation controlled by diffusion in three dimensions [23e25], in agreement with reports in the literature [8]. However, there is a distinction between the two samples studied. With an increase in b the CEG sample showed a tendency toward a reduction in n, in contrast to the behavior observed for the CPT sample, i.e., with an increase in b there was an increase in the exponent n. Thus, for CEG the degradation process will be forced at the crystal vertices and, analogously to crystallization, the exponent n tends toward 2, indicating that the degradation of the crystals may occur sporadically from the nucleus [38]. This difference may also be associated with the difference between the crystallite size in the 11 0 plane, as shown in Table 2. Fig. 6 shows the values in the conversion range of 0.2e0.8 for the determination of the Ea values obtained using the method proposed by FlynneWalleOzawa (FWO) [13,14].
The linear fits obtained from the plot of b vs. 1/T [13,14] are shown in Fig. 6. The linear correlation coefficients (r) were close to unity (minimum ¼ 0.9878 and maximum ¼ 0.9998). Therefore, the Ea values were calculated through the angular coefficient obtained from the straight line fit. Table 4 shows the Ea values for the cellulose samples studied. It can be observed that both samples showed a decrease in the Ea values in the range of a ¼ (0.2e0.8) with 209.47e157.79 and 173.30e138.47 kJ mol1 for the CEG and CPT samples, respectively. This range of values is consistent with values reported in the literature [9e11]. However, as can be observed, the CEG sample had higher Ea values than the CPT samples. According to Kim et al. [12] higher crystal sizes promote an increase in the degradation temperature; consequently the CPT cellulose fiber may have a higher thermal stability than the CEG one. Also, the authors report that the activation energy is not affected by the crystallite size. Moreover, the finding that CPT had the lowest activation energy values was attributed to the thermal decomposition of sulfite cellulose being controlled by dehydration, whereas the high Ea values for CEG may indicate depolymerization of the kraft cellulose
Fig. 4. Curves of the conversion (a) as a function of temperature (T/ C) for the different heating rates studied.
Fig. 6. Data obtained from the plot of log b vs. 1/T (K1) using the FWO method for the CEG sample.
684
M. Poletto et al. / Polymer Degradation and Stability 96 (2011) 679e685
Table 3 Kinetics parameters obtained through the linear fit of the Avrami equation. Sample
b ( C.min1)
k0 (min1)
t ½ (min)
n
r
CEG
5 10 20 40
0.04 0.07 0.15 0.29
2.25 1.94 1.62 1.36
3.67 3.42 3.24 2.87
0.999 1.000 0.998 1.000
CPT
5 10 20 40
0.03 0.07 0.15 0.31
3.39 2.51 1.62 1.27
2.51 2.55 3.22 3.34
0.999 1.000 0.998 0.998
with the production of levoglucosan. Furthermore, the decrease in Ea at a values of 0.2e0.8 observed for the two cellulose samples might be the result of autocatalysis in the dehydration process [35]. Thus, a mixed depolymerization and dehydration mechanism can be considered for the two cellulose samples, independently of the cellulose treatment. Similar behavior was observed by Soares et al. [35] for cellulose powder and kraft paper. Gümüskaya et al. [1] and Ouajai and Shanks [38], describe that different pulping conditions can affect the cellulose crystallinity. Thus, by investigating the thermal stability [12], it is possible to evaluate the influence of different treatments on the cellulose crystalline structure and how it affects the degradation kinetics and degradation mechanism. The Ea results obtained using the FWO method were used to determine the degradation mechanisms proposed by Criado et al. [18]. This method uses reference theoretical curves obtained from equation (11) that are derivatives of the f(a) and g(a) functions represented in Table 1, called master curves, which are compared to experimental data for determination of the mechanism of a solidstate process [21,22]. These mechanisms represent how the solidstate degradation process occurs. The algebraic expressions that represent the theoretical mechanisms are separated into four groups, An Rn Dn and Fn, as can be seen in Table 1. These mechanisms describe processes of nuclei formation on the propagation of the degradation process; diffusion processes that are related to the heat transfer capacity along the material structure; reaction mechanisms controlled by the surface of the sample; and the random degradation of nuclei, respectively. Figs. 7 and 8 present the master curves as well as the results of the experimental data obtained. The determination of the experimental Z(a) values was carried out using a heating rate (b) of 10 C min1, and the Ea values obtained applying the FWO method were used to calculate Z(a) in equation (12). The physical meaning of the function Z(a) varies depending on the theoretical approach to the mechanisms, as shown in Table 1. The experimental data for the CEG sample in the conversion range of a ¼ 0.2e0.4 were superimposed on the D1, D2 and D3 curves and according to the literature these degradation mechanisms refer to the diffusion processes in one, two and three dimensions, respectively [18,21]. Similar results were described by
Fig. 7. Master curves and experimental data obtained using the method of Criado et al. for the CEG sample.
Wu and Dollimore [8]. Subsequently, from a ¼ 0.5 the mechanism tends toward F1, corresponding to random nucleation with one nucleus in the individual particle [18,21]. These phenomena are in agreement with the values for the Avrami exponent given in Table 3, where the samples underwent degradation by diffusion in three dimensions. The Avrami exponent, in its original derivation is associated with the dimensional crystal growth processes. In this case, it is important to mention that the physical meaning of n used in this study is related to the degradation kinetics, i.e., the crystalline structure does not change with increasing b. Thus, n refers only to the difficulty of heating transfer by varying b. Changes in the Avrami exponent from 3 to 2 (according to Table 3) for the CEG cellulose fiber may be related to a greater heat transfer difficulty in transferring heat. This is due to the type of packing between the cellulose chains, since the characteristics associated with n suggest a more ordered structural form [39] on increasing the heating rate. This factor could result in an increase in the Ea values, as shown in Table 4 for the CEG cellulose fiber. The reported characteristic of n (n z 2) suggests diffusion processes [23e25,39], and this is in agreement with the diffusion process mechanism demonstrated (D1, D2, D3) in Fig. 7. For CPT, for a values in the range of 0.3e0.7 the degradation mechanism corresponded to R1, i.e., the phase boundary-controlled
Table 4 Activation energies obtained using the Flynn-Wall-Ozawa method. CEG
CPT
a
Ea (kJ mol1)
r
Ea (kJ mol1)
r
0.2 0.3 0.4 0.5 0.6 0.7 0.8
209.47 199.82 190.29 181.75 171.66 164.36 157.79
0.9984 0.9995 0.9998 0.9998 0.9997 0.9996 0.9993
173.30 167.27 160.89 156.09 151.30 143.91 138.47
0.9991 0.9990 0.9993 0.9990 0.9994 0.9989 0.9991
Fig. 8. Master curves and experimental data obtained using the method of Criado et al. for the CPT sample.
M. Poletto et al. / Polymer Degradation and Stability 96 (2011) 679e685
reaction (one-dimensional movement) [22], in agreement with the results obtained by Wu and Dollimore [8]. The R1 mechanism is also consistent with the results obtained from the XRD analysis, where the larger crystallite size for CPT might be related to the action of temperature on the boundaries due to the greater interfacial perimeter between crystals. 4. Conclusions Two distinct types of cellulose fibers obtained using different processes were characterized. The IR spectra showed a similar composition for the two samples. The XRD analysis demonstrated that, although the samples have close crystallinity indices, the crystallite size of the CPT cellulose obtained by the sulfite process in one of its orientations (11 0) is around 20% greater than that of the CEG cellulose obtained by the kraft process, which may be associated with a reduction in the corresponding amorphous region by the sulfite process. The thermogravimetric analysis showed that the CPT degradation curve is shifted to higher temperatures in the case of CEG, suggesting that the greater thermal stability is due to the larger crystallite size. These results are in agreement with the XRD results obtained. However, higher activation energy values were observed for the CEG cellulose fiber. This observation is consistent with reports in the literature. Through the Avrami parameters it was found that there are differences between the degradation processes of the cellulose pulps. This difference is related to the packing of the cellulose chains. To better understand this behavior, results obtained through the Criado method for the CEG cellulose fiber confirmed that the degradation occurs by diffusion, D1, D2 and D3, and that the heat transfer is affected by the type of arrangement and packing of cellulose crystals obtained by the kraft process. This characteristic could explain the higher Ea values obtained for CEG in comparison to CPT. Moreover, the degradation mechanism observed for CPT applying the Criado method is the phase boundary controlled reaction (R1). This supports the fact that the greater thermal stability of CPT obtained by the sulfite process is related to the increased crystallite size. Therefore, crystallinity, crystallite size, thermal stability and cell structure were more affected during kraft pulping than under the sulfite pulping conditions. This can be explained by the fact that in kraft pulping there is a more degraded cellulose structure due to the cooking process conditions, in comparison to the sulfite process. Acknowledgment The authors are grateful to Cambará S.A and CMPC S.A. for supplying the cellulose fibers and CAPES, CNPq and FAPERGS for financial support. The authors also thank Ms. Heitor L. Ornaghi Jr. for the suggestions given in this work. References [1] Gümüskaya E, Usta M, Kirei H. The effects of various pulping conditions on crystalline structure of cellulose in cotton linters. Polym Degrad Stab 2003;81: 559e64. [2] Duchesne I, Hult EL, Molin U, Daniel G, Iversen T, Lennhon H. The influence of hemicellulose on fibril-aggregation of kraft pulp fibres as revealed by FE-SEM and CP/MAS 13C-NMR. Cellulose 2001;8:103e11. [3] Wada M, Okano T. Localization of Ia and Ib phases in algal cellulose revealed by acid treatments. Cellulose 2001;8:183e8. [4] Hult EL, Iversen T, Sugiyama J. Characterization of the supermolecular structure of cellulose in wood pulp fibres. Cellulose 2003;10:103e10.
685
[5] Sun JX, Sun XF, Zhao H, Sun RC. Isolation and characterization of cellulose from sugarcane bagasse. Polym Degrad Stab 2004;84:331e9. [6] Neagu RC, Gamstedt EK, Berthold F. Stiffness contribution of various wood fibers to composite materials. J Comp Mater 2006;40:663e99. [7] Hult EL, Larsson PT, Iversen T. Cellulose fibril aggregationean inherent property of kraft pulps. Polymer 2001;42:3309e14. [8] Wu Y, Dollimore D. Kinetic studies of thermal degradation of natural cellulosic materials. Thermochim Acta 1998;324:49e57. [9] Nada AMA, Hassan ML. Thermal behavior of cellulose and some cellulose derivatives. Polym Degrad Stab 2000;67:111e5. [10] Antal MJ, Várhegyi G, Jakab E. Cellulose pyrolysis kinetics: revisited. Ind Eng Chem Res 1998;34:1267e75. [11] Capart R, Khezami L, Burnham AK. Assessment of various kinetic models for the pyrolysis of a microgranular cellulose. Thermochim Acta 2004;417:79e89. [12] Kim UJ, Eom SH, Wada M. Thermal decomposition of native cellulose: influence on crystallite size. Polym Degrad Stab 2010;95:778e81. [13] Flynn JH, Wall LA. General treatment of the thermogravimetry of polymers. J Res Natl Bur Stand 1966;70A:487e523. [14] Ozawa T. A new method of quantitative differential thermal analysis. Bull Chem Soc Jpn 1966;39:2071e85. [15] Pistor V, Ornaghi FG, Ornaghi Jr HL, Fiorio R, Zattera AJ. Thermochim Acta 2010;510:93e6. [16] Poletto M, Dettenborn J, Pistor V, Zeni M, Zattera AJ. Materials produced from plant biomass. Part I: evaluation of thermal stability and pyrolysis of wood. Mater Res 2010;13:375e9. [17] Doyle CD. Kinetic analysis of thermogravimetric data. J Appl Polym Sci 1961;5: 285e92. [18] Criado JM, Malek J, Ortega A. Applicability of the master plots in kinetic analysis of non-isothermal data. Thermochim Acta 1989;147:377e85. [19] William LP. Computation of the exponential trap population integral of glow curve theory. J Comp Phys 1971;7:187e90. [20] Pérez-Maqueda LA, Criado JM. The accuracy of Senum and Yang’s approximations to the Arrhenius integral. J Therm Anal Calorim 2000;60:909e15. [21] Tiptipakorn S, Damrongsakkul S, Ando S, Hemvichian K, Rimdusi S. Thermal degradation behaviors of polybenzoxazine and silicon-containing polyimide blends. Polym Degrad Stab 2007;92:1265e78. [22] Núñez L, Villanueva M, Núñez MR, Rial B. Influence of an epoxy reactive diluent on the thermal degradation process of the system DGEBA n ¼ 0/1,2 DCH. J Appl Polym Sci 2004;92:1199e207. [23] Avrami M. Kinetics of phase change. I. J Chem Phys 1939;7:1103e12. [24] Avrami M. Kinetics of phase change. II. J Chem Phys 1940;8:212e24. [25] Avrami M. Kinetics of phase change. III. J Chem Phys 1941;9:177e84. [26] Jeziorny A. Parameters characterizing the kinetics of the non-isothermal crystallization of poly(ethylene terephthalate) determined by d.s.c. Polymer 1978;19:1142e4. [27] Liu TX, Mo ZS, Wang SG, Zhang HF. Nonisothermal melt and cold crystallization kinetics of poly(aryl ether ether ketone ketone). Polym Eng Sci 1997; 37:568e75. [28] Tserki V, Matzinosa P, Kokkoub S, Panayiotou C. Novel biodegradable composites based on treated lignocellulosic waste flour as filler. Part I. Surface chemical modification and characterization of waste flour. Composites: Part A 2005;36:965e74. [29] Viera RGP, Filho GR, de Assuncão RMN, Meireles CS, Vieira JG, de Oliveira GS. Synthesis and characterization of methylcellulose from sugar cane bagasse cellulose. Carbohydr Polym 2007;67:182e9. [30] Xiao B, Sun XF, Sun RC. Chemical, structural, and thermal characterizations of alkali-soluble lignins and hemicelluloses, and cellulose from maize stems, rye straw, and rice straw. Polym Degrad Stab 2001;74:307e19. [31] de Menezes AJ, Pasquini D, da Silva Curvelo AA, Gandini A. Self-reinforced composites obtained by the partial oxypropylation of cellulose fibers. 1. Characterization of the materials obtained with different types of fibers. Carbohydr Polym 2009;76:437e42. [32] Colom X, Carrasco F, Pagès P, Cañavate J. Effects of different treatments on the interface of HDPE/lignocellulosic fiber composites. Comp Sci Tech 2003;63: 161e9. [33] Colom X, Carrillo F, Nogués F, Garriga P. Structural analysis of photodegraded wood by means of FTIR spectroscopy. Polym Degrad Stab 2003;80:543e9. [34] Quiévy N, Jacquet N, Sclavons M, Deroanne C, Paquot M, Devaux J. Influence of homogenization and drying on the thermal stability of microfibrillated cellulose. Polym Degrad Stab 2010;95:306e14. [35] Soares S, Camino G, Levchik S. Comparative study of the thermal decomposition of pure cellulose and pulp paper. Polym Degrad Stab 1995;49:275e83. [36] Yao F, Wu Q, Lei Y, Guo W, Xu Y. Thermal decomposition kinetics of natural fibers: activation energy with dynamic thermogravimetric analysis. Polym Degrad Stab 2008;93:90e8. [37] Elanthikkal S, Gopalakrishnapanicker U, Varghese S, Guthrie JT. Carbohydr Polym 2010;80:852e9. [38] Ouajai S, Shanks RA. Composition, structure and thermal degradation of hemp cellulose after chemical treatments. Polym Degrad Stab 2005;89:327e35. [39] Sperling LH. Introduction to physical polymer science. New Jersey: John Wiley & Sons; 2006.