Analysis of Pyrolysis Characteristics and Kinetics of Cigar Tobacco and Flue-Cured Tobacco by TG-FTIR

Three types of tobacco middle leaves were considered in this study: cigar filler tobacco (CFT), cigar wrapper tobacco (CWT) and flue-cured tobacco (FCT). CFT and CWT were cultivated in Hainan Province, China, and collected after air-curing process. FCT was cultivated in Yunnan, China, and collected after flue-curing process. After drying at 40 °C for 5 h, the samples were ground to pass through a 40-mesh sieve and then stored in a sealed container before use. The ultimate analysis of the tobacco was conducted on an Elementar Vario EL cube CHNSO analyzer (Hanau, Germany). Proximate analysis of tobacco was conducted to determine the contents of moisture, volatile matter, fixed carbon and ash by using thermogravimetric method (15). The content of cellulose, hemi-cellulose and lignin in the samples was determined according to a previous work (16). The nicotine content of the samples was determined according to the methods described in ISO 10315 (17).

Experiments were performed using a thermogravimetric analyzer (Discovery, TA Instrument, Newcastle, DE, USA) coupled with a FTIR spectrophotometer (Nicolet 8700, Thermo Electron, Waltham, MA, USA) through a transfer line. The mass change of the samples with temperature during the pyrolysis was recorded. Approximately 10 mg of the sample were placed in a platinum crucible. The experiments were performed from 40 °C to 900 °C at different heating rates (5 °C min⁻¹, 10 °C min⁻¹, 15 °C min⁻¹, and 20 °C min⁻¹) under nitrogen atmosphere (99.999% minimum purity) with a flow rate of 30 mL min⁻¹. The stainless-steel transfer pipe and the gas cell in the FTIR were both heated to a constant temperature of 210 °C to minimize secondary reactions. The resolution of the FTIR was set at 4 cm⁻¹, spectrum scan frequency was 8 times per minute, and the spectral region was from 600 cm⁻¹ to 4000 cm⁻¹.

The devolatilization index (D_i) is a comprehensive evaluation of volatile release and can reflect the reactivity of the whole pyrolysis process. A greater D_i value implies that it is more reactive. D_i is determined by the following equation (18,19,20): Di=(Rmax⋅Rmean)/(TS⋅Tmax⋅ΔT1/2) {D_i} = \left( {{R_{max }} \cdot {R_{mean}}} \right)/\left( {{T_S} \cdot {T_{max }} \cdot \Delta {T_{1/2}}} \right) where R_max is the maximum devolatilization rate, R_mean the average devolatilization rate, T_S the initial volatile release temperature, T_max is the temperature corresponding to R_max, and ΔT_1/2 is the temperature interval containing the maximum peak where R/R_max = 1/2. The devolatilization rate, R, is defined as R = dw_t/dt, where w_t is the mass percentage of the raw sample at time t. R_max and R can be obtained from the DTG curves.

For the kinetic analysis, the decomposition rate can be described by the following equation: [1] dαdt=k(T)f(α) {{d\alpha } \over {dt}}=k(T)f(\alpha ) where α is the extent of conversion, and α = (m₀−m_t)/(m₀−m_∞), m₀ is the initial mass of the sample, m_t is the mass at a given time t, and m_∞ is the final mass of the sample, T is the absolute temperature, k(T) is the reaction rate as function of the temperature and f(α) is the reaction model, which is a function of α. The reaction rate constant is typically described in accordance with the Arrhenius equation, which is a function of T and expressed as: [2] k(T)=A exp(−Ea/RT) k(T) = A\,\exp \left( { - {E_a}/RT} \right) where A is the frequency factor, E_a is the activation energy, and R is the universal gas constant (8.3145 J mol⁻¹ K⁻¹).

The heating rate β is defined as dT/dt, which is substituted in Equation [1] to obtain Equation [3]: [3] dαdT=Aβexp(−EaRT)f(α) {{d\alpha } \over {dT}} = {A \over \beta }\exp \left( { - {{{E_a}} \over {RT}}} \right)f(\alpha )

The integral form of Equation [3] is expressed as follows, [4] G(α)=∫0αdαf(α)=Aβ∫0Texp(−EaRT)dT=AEaβRP(u) G\left( \alpha \right) = \int_0^\alpha {{{d\alpha } \over {f\left( \alpha \right)}} = {A \over \beta }} \int_0^T {\exp \left( { - {{{E_a}} \over {RT}}} \right)dT = {{A{E_a}} \over {\beta R}}P\left( u \right)} where G(α) is the integral form of f(α), P(u) is an approximation, and u is defined as the equation of u = E_a/R.

The composition of different tobacco leaves is summarized in Table 1. The chemical composition of tobacco leaf is affected by factors such as genetics, agricultural practices, soil type, and curing procedures, etc. (27). Cigar tobacco and flue-cured tobacco have significant differences in fixed carbon, ash, nitrogen, oxygen and nicotine content. The nitrogen, nicotine, and ash content of cigar tobacco is higher than that of flue-cured tobacco, while the oxygen and cellulose content is on the contrary. The total content of hemicellulose, cellulose, and lignin in tobacco is from 15% to 20%.

Figure 1

Figure 2

The TG-DTG curves of CFT, CWT and FCT pyrolysis processes at the heating rate of 10 °C min⁻¹ are shown in Figure 1. The pyrolysis behavior of tobacco is very complicated, in addition to the decomposition of biopolymers such as cellulose, hemicellulose, pectin, and lignin, further comprising a large number of decompositions of non-polymer compounds. Generally, the decomposition of cigar tobacco can be divided into four stages. In stage I, the decrease in mass occurs primarily due to the evaporation of water, which accounts for almost 3% of the sample. In stage II, the thermal decomposition of hemicellulose in tobacco takes place mainly between 220 °C and 285 °C (28, 29), the peak at about 270 °C could be the maximum decomposition rate of hemicellulose. In addition, some low-temperature volatiles are released, partially due to the decomposition of sugars, pectin, and nicotine, that occurs at this stage (30, 31). The mass loss of CFT and CWT at this stage is 23% and 20%, respectively. In stage III, the temperature range and the temperature of the peak of the DTG curve are very close to the decomposition temperature of cellulose (32, 33). This indicates that stage III of the DTG curve may correspond to cellulose pyrolysis of the cigar tobacco leaves. In this stage, the thermal decomposition rate increases compared with that of stage II. In stage IV, a shoulder between 420 °C and 600 °C was observed, which could be related to the decomposition of lignin (34). Mass loss after 600 °C may include dehydrogenization and aromatization of residual carbon (3). Among the three components, hemicellulose is the easiest to be pyrolyzed, followed by cellulose, and lignin is the most difficult one. This may be due to differences in the chemical structures and natures of polymeric materials. Hemicellulose is a mixture of various monosaccharides (xylose, mannose, glucose, galactose, etc.) with a low degree of polymerization. In contrast, the cellulose molecule is a very long polymer of glucose without branches, which results in thermal degradation more difficult than in hemicellulose. Lignin is composed of three kinds of benzene-propane and is heavily cross-linked. Therefore, lignin has high thermal stability and is difficult to decompose (35). The main pyrolysis stages (stage II and stage III) of CFT and CWT contribute to a mass loss of 48.48% and 44.95%, respectively. CWT has been grown under shade treatment conditions that may be unfavorable for the accumulation of organic material in the leaves.

However, it can be seen from the DTG curves that the pyrolysis process of flue-cured tobacco is different from that of cigar tobacco. This may be caused by differences in physical structure and chemical composition. Flue-cured tobacco has an obvious mass loss peak at 200 °C, which corresponds to the decomposition of volatile compounds. The significant difference observed may be due to the difference in sugar content. According to the previous literature (27), flue-cured tobacco has a high sugar content, while in cigars - as air-cured tobacco - their sugars have been virtually depleted via metabolic activities. In addition, flue-cured tobacco has a continuous mass loss peak between 230 °C and 400 °C due to the lower hemicellulose content relative to the cellulose content, while the mass loss peak of cigar tobacco occurs at a slightly lower temperature at this stage. This result shows that the thermal decomposition temperature of biopolymers in cigar tobacco is lower than that of flue-cured tobacco, which may be due to the lower crystallinity and/or the lower degree of polymerization of the biopolymers in the cigar tobacco (36).

The TG-DTG curves of CFT, CWT and FCT pyrolysis processes at different heating rates are shown in Figure 2. The mass loss at different heating rates during the pyrolysis of samples was obtained from TG-DTG curves and is summarized in Table 3. With the rise of the heating rate, the reaction rate accelerates, and the DTG curves move toward the higher temperature. This may be due to the fact that the time required to reach the final temperature is reduced as the heating rate increases. This results in a decrease in the reaction time experienced by the sample, and thermal hysteresis occurrs during the pyrolysis process (37). As shown in Table 3, the mass loss is not strongly dependent upon the heating rate. This suggests that the total mass loss at any stage depends primarily on the basic composition of the sample (38).

The characteristic parameters of pyrolysis (T_S, T_max, R_max, R_mean, ΔT_1/2, and D_i) of CFT, CWT and FCT at varying heating rates are summarized in Table 4. At a heating rate of 10 °C min⁻¹, the devolatilization indices D_i of CFT, CWT, and FCT are 6.99 × 10⁻⁷%² °C⁻³ min⁻², 10.62 × 10⁻⁷%² °C⁻³ min⁻², and 27.3 × 10⁻⁷%² °C⁻³ min⁻², respectively. In summary, compared with CFT and CWT, FCT has a lower characteristic temperature of volatiles release, and the ΔT_1/2 is narrower. This shows that FCT is more reactive than CFT and CWT. As shown in Table 4, R_max, and D_i increase as the heating rate increases. Generally speaking, combustibles with greater D_i value imply that they are more reactive. Thus, increasing the heating rate is in favor of the release of volatile substances.

The 3D FTIR spectrum of the gases emitted during tobacco pyrolysis are shown in Figure 3. On the whole, the 3D spectra demonstrate similar gas release behavior with some slight differences. Five characteristic absorbance bands with the maximum absorption at 3565 cm⁻¹, 2357 cm⁻¹, 1749 cm⁻¹, 1504 cm⁻¹, and 667 cm⁻¹ are observed for tobacco samples. This result suggests that the gas products corresponding to these five bands are the pyrolysis products. Moreover, each characteristic FTIR peak exhibits several peaks, and the evolution with temperature corresponds to the DTG curves.

In order to identify the components of the gas products, the 2D FTIR spectra at the peak temperature of the four stages are shown in Figure 4. Based on the FTIR spectra, the result of component identification of the gas products is shown in Table 5. To avoid interference from the pyrolyzed mass, the absorbance has been normalized by dividing it by the initial mass of tobacco.

Figure 3

Figure 4

In the first stage (T1), the main volatile component is H₂O, as evidenced by the absorbance band between 3500 cm⁻¹ and 4000 cm⁻¹, which was mainly caused by the evaporation of the sample. In stage II (T2), the bands in the region from 2250 cm⁻¹ to 2500 cm⁻¹ and 580 cm⁻¹ to 700 cm⁻¹ demonstrate the release of CO₂. CO₂ is the dominant gas during tobacco pyrolysis, and is mainly produced by the cracking and reforming of carbonyl and carboxyl groups (39). Besides, a very small asymmetric absorption peak corresponding to C-O appears in the wavenumber from 1050 cm⁻¹ to 1200 cm⁻¹, indicating the existence of alcohols and phenols (40). The C=O stretching vibration in the range from 1710 cm⁻¹ to 1800 cm⁻¹ is attributed to aldehydes, ketones, and organic acids which contained the carbonyl groups. The bands in the region from 1450 cm⁻¹ to 1650 cm⁻¹ are the C=C stretching vibration and benzene skeleton vibration. The band at 966 cm⁻¹ represents NH₃. With the pyrolysis temperature increased to T3, the abovementioned absorbance bands gradually increased in intensity. The band from 2850 cm⁻¹ to 3030 cm⁻¹ is characteristic of hydrocarbon gases such as CH₄. The absorption band from 2000 cm⁻¹ to 2250 cm⁻¹ indicates the presence of CO. The spectrum obtained at T4 shows that the main components of the gases released are CO, CO₂, and water, which are caused by the carbonates inherent or generated during pyrolysis (41).

According to the widely used Lambert-Beer law, the absorbance at a specific wavenumber is linearly related to the gas concentration (44). Therefore, changes in absorbance during pyrolysis reflect changes in the concentrations of different gases. Figure 5 compares the evolution of gas products with increasing temperatures during the pyrolysis of tobacco samples. As shown, the gas products (except for CH₄ and CO) are concentrated within the temperature range from 200 °C to 400 °C. Moreover, the evolution of released CO₂ with temperature is consistent with the thermogravimetric analysis. The CO₂ released from 200 °C to 400 °C can mainly be attributed to the cracking and reforming of carboxyl groups that CWT exhibit a more intense signal in this stage, whereas above 400 °C, the polymerization reaction of solid phase coking is accompanied by the emission of CO₂ as the temperature increases.

Figure 5

Figure 6

Figure 7

The release rate of water varied with the temperature, the band below 200 °C was related to the evaporation of water. The appearance of a maximum at around 320 °C for H₂O can be attributed to the decomposition of hemicellulose and cellulose by the cracking of hydroxyl groups in the lateral chains (45). CH₄ is formed by the cleavage of methoxyl groups and methylene groups. This band shows a maximum at around 540 °C, which is associated with lignin decomposition. For CO, two peaks appear at around 330 °C and 500 °C, indicating that the formation of CO is related to the decomposition of cellulose and lignin, and the release amount of CO above 600 °C increases significantly. The release of CO is consistent with the results reported by BAKER (46), suggesting that the formation of carbon oxide occurs in two regions, a low-temperature region from 100 °C to 450 °C and a high-temperature region from 500 °C to 900 °C. The total amount of carbon oxides is independent of the heating rate, but the relative proportions of CO₂ and CO are strongly dependent on the heating rate. BAKER proposed that CO undergoes some gas phase reactions, and that CO₂ undergoes both gas phase reaction and heterogeneous reaction. The release curves of carbonyl and hydroxyl compounds were consistent in the maximum mass loss range, indicating that the formation of aldehydes and ketones from tobacco pyrolysis went along with the formation of alcohols and phenols. Aromatics were formed at temperatures above 500 °C and with a maximum at around 540 °C, suggesting that the production of aromatic compounds was related to lignin decomposition. Compared with CFT and CWT, FCT had lower decomposition intensity on CO₂, CH₄, CO, and aromatics.

To determine the activation energy of the pyrolysis of cigar tobacco and flue-cured tobacco, two different model-free isothermal conversion methods (FWO and KAS models) were used. In the FWO method, the activation energy can be determined from the plots of lg(β) versus 1/T. In the KAS method, the activation energy can be determined from the plots of lg(β/T²) versus 1/T. Figures 6 and 7 show the dependence of the activation energy on the conversion (α = 0.1–0.9) for the different tobacco leaves based on the isoconversional FWO and KAS methods, respectively. The corresponding values calculated from the slope and the coefficients of determination R² at different degrees of conversion (α) are presented in Table 6. The E_a values of CFT range from 207.4 kJ mol⁻¹ to 319.3 kJ mol⁻¹, that of CWT range from 160.4 kJ mol⁻¹ to 260.5 kJ mol⁻¹, and that of FCT range from 100.1 kJ mol⁻¹ to 192.1 kJ mol⁻¹. Importantly, the coefficients of determination R² of all curves are above 0.98 indicating that all linear plots are statistically relevant. The difference between the two methods is less than 2%, indicating the accuracy and reliability of the calculations of the FWO and KAS methods.

Figure 8

Figure 8 depicts the relationship between the activation energy E_a calculated by the FWO and KAS methods at different conversions. The kinetic analysis results show that E_a is highly dependent on α, which indicates that the pyrolysis of tobacco leaf is a complex process including parallel, competitive and consecutive reactions. Taking the results of the CFT as an example, the activation energy increased rapidly as the conversion value increased from 0.1 to 0.2, and remained approximately at about 250 kJ mol⁻¹ as α increased from 0.2 to 0.4. At this stage, the pyrolysis temperature was not particularly high, and corresponded to the drying of the cigar tobacco and the partial pyrolysis of hemicellulose. Therefore, a higher energy was required to increase the reactivity. As the pyrolysis proceeded and α increased from 0.4 to 0.6, the activation energy E_a decreased from 253.5 kJ mol⁻¹ to 219.3 kJ mol⁻¹ (FWO), this stage was mainly contributed by the thermal degradation of cellulose. Cellulose was initially pyrolyzed into active cellulose, resulting in a decrease in polymerization degree and molecular chain length. Active cellulose continued to degrade into more components with lower molecular weight and required lower activation energy (37). When α was higher than 0.6, the value of E_a increased significantly. This stage corresponded to the lignin pyrolysis in the DTG curve. Since lignin consists mainly of three forms of highly cross-linked benzene-propane (35), its thermal stability is high and the activation energy required for pyrolysis increases. At this stage, coke is produced at the same time, which also increases the activation energy value.

Generally, there are similar variations in the activation energy of the three types of tobacco. Since the conversion rate of flue-cured tobacco is from 0.1 to 0.3, corresponding to the pyrolysis of volatile components such as sugars, the activation energy gradually increases from 100.1 kJ mol⁻¹ to 121.4 kJ mol⁻¹. Comparing the activation energies during the whole process, the activation energy of FCT is relatively lower than that of CFT and CWT. Because the value of the apparent activation energy represents the minimum energy required for a chemical reaction to occur, greater apparent activation energy indicates lower pyrolytic reactivity. These results indicate that the volatile materials of FCT decomposed more easily than the components of the CFT and CWT.

Heterogeneous reactions usually involve a superposition of several elementary processes such as order-based, the diffusion, the random nucleation and nuclei growth, these reaction mechanisms G(α) are summarized in Table 2. The theoretical curves y(α) versus α were plotted using Equation [7] and are shown in Figure 9. This figure also shows the experimental curves for three types of tobacco leaf, where S1 and S2 are the curves of the sample in stage II and stage III, respectively. As shown in Figure 9, the theoretical curve for G(α) corresponding to the one-dimensional diffusion D1 is consistent with the experimental data of S1 in tobacco. Thus stage II follows one-dimensional diffusion D1, which is in accordance with the results obtained by Gao et al. (6) studying flue-cured tobacco. The experimental curve S2 of FCT matched well with the theoretical master plot of two-dimensional diffusion D2. However, the experimental curve S2 of CFT and CWT did not fit very well to any of the theoretical curves, but it was acceptably close to the theoretical curves of 1.5-order reaction F3/2 and second-order reaction F2. In order to determine the reaction mechanism further, the Coats-Redfern method was used to determine the thermal decomposition mechanism of stage III.

By combining the two mechanism functions, values for E_a, A, and correlation coefficients (R²) were obtained by the Coats-Redfern method, and are shown in Table 7. For the pyrolysis of CFT and CWT in stage III, the second-order reaction F2 had higher correlation coefficients (0.997 and 0.984) compared with the 1.5-order reaction F3/2. Simultaneously, the activation energy E_a obtained by the Coats-Redfern method was close to the activation energy (234.42 kJ mol⁻¹ and 224.65 kJ mol⁻¹) obtained by the FWO and KAS isoconversional methods. Thus, the most likely kinetic mechanism of stage III for cigar tobacco is second-order reaction F2.

Figure 9

The pyrolysis of tobacco is a complex and continuous process, which is affected by different reactions. The thermal decomposition of ligno-cellulosic biomass is initially limited by diffusion mechanisms (47). First-order and second-order reactions typically occur during pyrolysis when cellulose content is higher than the lignin content (48). In stage III, the difference of cigar tobacco and flue-cured tobacco is the temperature range, leading to a different reaction mechanism. It also shows that reaction mechanism may be highly related to the degradation behavior of the main components. In addition, the activation energy of cigar tobacco found in our experiments was higher than that reported in a study on tobacco waste (103.94 kJ mol⁻¹) (10) and tobacco dust (43.6 kJ mol⁻¹ – 227.0 kJ mol⁻¹) (13). The difference could be related to the type of species, growth condition, particle size of tobacco and the different kinetic models.