FTIR study of tobacco and glycerol–tobacco mixtures

Thermochimica Acta 573 (2013) 146–157

Contents lists available at ScienceDirect

Thermochimica Acta journal homepage: www.elsevier.com/locate/tca

TGA/FTIR study of tobacco and glycerol–tobacco mixtures A. Gómez-Siurana, A. Marcilla ∗ , M. Beltrán, D. Berenguer, I. Martínez-Castellanos, S. Menargues Dpto. Ingeniería Química, Universidad de Alicante, Apdo. 99, 03080 Alicante, Spain

a r t i c l e

i n f o

Article history: Received 15 May 2013 Received in revised form 23 July 2013 Accepted 9 September 2013 Available online 1 October 2013 Keywords: TGA/FTIR Tobacco Glycerol TGA analysis

a b s t r a c t In this work, the evolution with temperature of the qualitative composition of the gases evolved in the pyrolysis of glycerol, tobacco and tobacco–glycerol mixtures has been studied. The pathways for different types of compounds (i.e., water, CO, CO2 , carbonylic compounds, alkenyl or alkyl groups containing compounds, alcohols and phenols and aromatic compounds) have been established, and their relationship with the different reaction steps involved in the pyrolysis process have been suggested. The comparison among the behavior observed in the pyrolysis of tobacco, glycerol and a mixture glycerol–tobacco has permitted us to suggest possible interactions between tobacco and glycerol affecting the composition of the gases evolved. © 2013 Elsevier B.V. All rights reserved.

1. Introduction The importance and interest of the study of the pyrolysis of tobacco and the different additives used in the fabrication of commercial cigarettes in order to obtain basic information about the characteristics of the cigarette smoking process and the smoke evolved has been pointed out repeatedly [1–5]. A burning cigarette is a very complex system, where many chemical reactions and physical processes occur [1,3–7] depending on the temperature and the oxygen availability, in the combustion zone and in the pyrolysis and distillation zones. Moreover, some of the volatile compounds evolved in these processes condense and/or are filtered by the tobacco fibers when the mainstream smoke traverses the cigarette during a puff, thus remaining as eventual reactants for the combustion, pyrolysis and distillation processes when they are reached by the burning line. In this way, several attempts focused on the modeling of the complex processes involved in tobacco smoldering have been developed, where the evaporation, pyrolysis and combustion reactions have been considered together with the heat and mass transfer and fluid dynamics phenomena, in order to predict the behavior of cigarettes as well as the effect of smoldering conditions in the product concentrations [8–11], or even the diffusion processes of the gases evolved through the paper wrapper of a cigarette [12]. The knowledge of the composition of the products appearing in tobacco smoke is an important issue that has been widely studied

∗ Corresponding author. Tel.: +34 96 590 2953; fax: +34 96 590 3826. E-mail address: [email protected] (A. Marcilla). 0040-6031/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.tca.2013.09.007

by different researchers in order to obtain information about the cigarette smoke chemistry and toxicity [i.e., 7,13,14] and the relationship with the cigarette design features [15] and to design strategies to develop commercial products capable of reducing tobacco toxicity [3,16]. On the other hand, the knowledge of the composition of the products generated in tobacco pyrolysis, as well as the dynamic of the weight loss processes involved can also be interesting from the point of view of energy recovery from tobacco wastes [17–22]. Even the insecticidal, fungicidal, and bactericidal properties of pyrolysis bio-oil from tobacco have been studied [23]. The importance of performing studies of cigarette ingredients has been recognized by different authors, as well as the need for developing regulatory requirements for testing cigarette ingredients based on their pyrolytic behavior instead of on their use as food or drug additives [i.e., 1,4,24]. Several authors have reported results of studies of pyrolysis of ingredients and mixtures of ingredients added to cigarettes, carrying out the experimental runs in different pyrolysis equipments, as different reactor types [1,4], thermobalances [25,26] or in smoking machines [24,27–31]. One advantage of these studies is the possibility of using higher levels of ingredient, thus increasing the probability that effects of the ingredient itself can be detected [24]. Nevertheless, it should be taken into account that the pyrolytic behavior of an isolated ingredient could be different to that presented when the pyrolysis runs are carried out with the ingredient mixed with tobacco [1,32–34]. Glycerol is widely used as a cigarette ingredient, acting as a moisturizing and surface active agent for flavor application. It is reported to neither be mutagenic, carcinogenic nor produce adverse reproductive effects as it is generally recognized as safe [24]. Carmines and Gaworski [24] have evaluated the toxicological

A. Gómez-Siurana et al. / Thermochimica Acta 573 (2013) 146–157

3R4F cigarettes from the Reference Cigarette Program of the College of Agriculture of the University of Kentucky and glycerol (CAS number 56-81-5) with purity > 99% provided by Sigma Aldrich have been used for the TGA/FTIR study of the pyrolysis of pure components and the mixtures thereof. Table 1 shows the composition of the reference cigarettes, as reported by the supplier. Samples of tobacco, glycerol and glycerol–tobacco mixture of less than 10 mg were pyrolyzed in an N2 atmosphere (99.999% minimum purity) using a TGA Netzsch TG209 thermobalance at a heating rate of 35 ◦ C/min under a nitrogen flow of 30 mL min−1 (STP). The sample temperature was measured with a thermocouple directly attached to the crucible, i.e., very close to the sample. The heating rate was selected high enough to ensure the quality of the FTIR data obtained [41]. To ensure the measurement of the actual sample temperature, a calibration of the temperature was performed using the Curie-point transition of standard metals, as suggested by the equipment recommendations. The output of the gas from the TGA was connected to a Bruker Tensor 27 FTIR Table 1 Composition of the cigarettes of reference. 100% of mass corresponds to the sum of the masses of the different tobaccos mixed in the cigarette (i.e., Flue cured, Burley, Maryland, Oriental and Reconstituted). Blend summary

3RF4 reference cigarette

Flue cured Burley Maryland Oriental Reconstituted (Schweitzer Process) Glycerol (dry-weight basis @ 11.6% OV) Isosweet (sugar)

35.41% 21.62% 1.35% 12.07% 29.55% 2.67% 6.41%

3. Results and discussion 3.1. Analysis of TGA and FTIR results Fig. 1 shows the TGA-curves obtained for tobacco, glycerol and a glycerol–tobacco mixture (with 24% (w/w) of glycerol) showing the sample weight loss during the thermal decomposition, as well as the corresponding derivative curves (DTG curves), where the temperatures corresponding to the maximum decomposition rate (i.e., DTG-peak temperatures) can be determined. The TGA curves of several tobacco types have already been described in the bibliography [43,44], and an assignment of the different decomposition stages has been suggested. In general, the pyrolysis of tobacco is usually described in terms of its major components, cellulose, hemicellulose and lignin, which have different thermal decomposition behavior depending upon the heating rate and the presence of contaminants [20,21,39,43–45]. Hemicellulose decomposes first (220–400 ◦ C), followed by cellulose, which converts to non-condensable gas and condensable organic vapors at 0

100

80

-5

TG-Tobacco TG-Glycerol

-10

TG-mixture DTG-Tobacco

60

DTG-Glycerol

-15

DTG-Mixture 40

DTG (%/min)

2. Experimental methods

spectrometer through a heated line, as described in the bibliography [42]. The balance adapter, the transfer line and the FTIR gas cell were heated at 200 ◦ C, thus avoiding the condensation of the less volatile compounds. On the other hand, the low volumes in the thermobalance microfurnace, transfer line and gas measurement cell permit low carrier gas flowrates to be used and allow for good detection of the gases evolved in the pyrolysis process. A lag of around 15 s, equivalent to 8.8 ◦ C has been estimated between the moment when the gases go outside the TGA furnace and their arrival at the IR cell. The reference tobacco was grinded in order to avoid the heterogeneity associated with the different tobacco fibers, and the mixtures with glycerol were prepared by weighing the corresponding amounts of each material directly in the crucible, and carefully mixing with the aid of a fine needle. Each thermogravimetric run was repeated twice in order to ensure the reproducibility of results. The glycerol–tobacco mixtures were prepared by adding glycerol (around 25%, w/w) to the 3R4F grinded tobacco, directly in the TGA crucible, and mixing thoroughly with a thin needle. The total amount of glycerol is thus the sum of that added and that already contained in the reference tobacco. In order to avoid the influence of the possible changes of the homogeneity and the properties of the mixture with time, the TGA analysis were carried out immediately after each sample of mixture was prepared. The percentage of glycerol added has been chosen according the results obtained in a previous work [32] in order to enhance the effect of the glycerol on the mixture behavior.

% of remaining mass

effect of glycerol as a cigarette ingredient, and report that, under the flash pyrolysis experimental conditions used in their work, glycerol does not pyrolyze extensively, being partially transferred intact to smoke, and yielding acrolein and glycolaldehyde as pyrolysis products. Moreover, when glycerol is added to tobacco, the amount of tar and water in smoke increased while the amount of nicotine decreased, and an effect of dilution was found, in agreement with its humectant role, which increases the presence of water and glycerol in smoke. Baker and Bishop [1] reported the composition of the glycerol pyrolysate obtained in flash pyrolysis conditions simulating a burning cigarette, showing the presence of 99.8% of glycerol, whereas Baker et al. [30] found that the presence the glycerol increases the total particulate matter (TPM) in the smoke, probably caused by the direct intact transfer of glycerol to the smoke particulate phase. Paine et al. [35] studied the flash pyrolysis of glycerol by isotopic labeling, concluding that the formation of acetaldehyde and acrolein occur by unimolecular reactions, and described two competitive mechanisms for the formation of acetaldehyde from pyrolysis of glycerol, one of them involving dehydration. In this study, the formation of formaldehyde was also reported. The usefulness of TGA connected on line with different analytical techniques for the knowledge of the evolution with the temperature of the volatile products evolved in pyrolysis processes has been widely demonstrated in the bibliography [22,25,36–40]. Thus, this work is focused on the qualitative study of the evolution with the temperature of the composition of the gases evolved in the pyrolysis of glycerol, tobacco and tobacco–glycerol mixtures through TGA/FTIR with the main objective of obtaining basic information to assist in the knowledge of the influence of glycerol on the composition of the smoke evolved from cigarettes. This study will be used as a reference for the next study where the effect of different catalysts is investigated.

147

-20 20

-25

-30

0 0

100

200

300

400

500

600

700

T (ºC)

Fig. 1. TGA and DTG-curves of tobacco, glycerol and glycerol–tobacco mixture.

148

A. Gómez-Siurana et al. / Thermochimica Acta 573 (2013) 146–157 0.008

100 TG-Tobacco

0.006

TG-mixture GS-Tobacco GS-Glycerol

60

% mass

GS-Mixture 0.004

40 0.002

20

0

GS intensity /mass of sample

TG-Glycerol

80

0 0

100

200

300

400

500

600

700

T (ºC)

Fig. 2. TGA and normalized GS-curves of tobacco, glycerol and glycerol–tobacco mixture.

320–420 ◦ C. Lignin decomposes slowly, starting at temperatures as low as 160 ◦ C, and extending up to 800–900 ◦ C. In general, solid residues derive mainly from lignin and some from hemicellulose, probably due to the linkage through covalent bonds which prevent their release during pyrolysis. In a previous work [32], the TGApyrolytic behavior of different compounds widely used as tobacco additives – including glycerol – and their mixtures with tobacco was studied. The results obtained in this work are in good agreement with these reported data, and the following assignment of decomposition steps of Fig. 1 can be suggested: Tobacco: evaporation of moisture at temperatures lower than 140 ◦ C, with DTG-peak temperature at around 83 ◦ C; evaporation of glycerol and other volatile compounds in the range 140–215 ◦ C; two overlapped processes in the range 215–366 ◦ C corresponding to decomposition of hemicellulose and other polymers (276 ◦ C) and pyrolysis of cellulose and other components of the tobacco (323 ◦ C), respectively. Pyrolysis of lignin takes place typically in a wide range of temperature, yielding a very diffuse DTG-peak that, in this case could contribute to that observed at around 450 ◦ C. Dehydrogenation and aromatization of char and/or decomposition of endogenous inorganic compounds take place at around 650 ◦ C. In the following text, in the sake of simplicity, we will refer for the decomposition events at around 276 ◦ C, 323 ◦ C and 450 ◦ C, as the hemicellulose, cellulose and lignin peaks. Glycerol: a single fast weight loss step, with DTG-peak at 219 ◦ C, mainly attributed to the evaporation of glycerol. Glycerol–tobacco mixture: the DTG curve differs from that of tobacco in (a) an increase of the peak ascribed to water (despite the fact that the DTG curve of pure glycerol is almost free of such a peak), (b) a noticeable increase of the importance of a peak at around 230 ◦ C, which results from a displacement of the process related with the loss of glycerol toward higher temperatures, (c) no evidence of the peak at 276 ◦ C, which overlaps with the previous peak, and (d) slight displacement toward lower temperatures as well as some decrease of the relative importance of the last decomposition step of tobacco. As was observed elsewhere [32], these results indicate that some interaction between tobacco and glycerol could occur and consequently, the thermal behavior of the mixture cannot be obtained by merely adding the thermal behavior of tobacco and glycerol. Additional information about the process can be obtained from the Gram–Schmidt curves (GS curves) that, based on vector analysis, reconstruct the acquired interferograms, allowing the plots of the total evolved gases detected by the spectrometer to be obtained. Fig. 2 shows the GS curves obtained for the systems studied in this work. In general, as was expected, the maxima of intensity measured in the IR cell are coincident with the weight loss stages (i.e.,

the decomposition steps), and the shapes of the DTG and GS curves are very similar. However, the GS curves reveal some interesting aspects: first at all, it is worthwhile to point out the case of glycerol, which does not reflect any well-defined peak in the GS curve, but a continuous presence of compounds in the carrier gas is detected in the IR cell. This behavior is coherent with the sharp DTG-peak observed for glycerol, which was associated to its evaporation processes, if we consider that glycerol, with normal boiling point at 290 ◦ C, is continuously condensing through the TGA–FTIR transfer line, heated at 200 ◦ C, and simultaneously evaporating as a consequence of the carrier gas flow. Thus, this interpretation should be confirmed by the analysis of the evolution of the IR bands in the spectra obtained at each temperature, which should always show the glycerol spectrum. The comparison between the DTG and GS curves for tobacco also show some differences which, in this case, are related with the relative importance of the different peaks, attributable to the differences in the intensity of the IR bands corresponding to the different functional groups in the compounds evolved from the TGA furnace at each temperature. Moreover, some displacement toward higher temperatures as a consequence of the delay between the time of generation of compounds in the TGA and the arrival to the IR cell can also be observed. Finally, the GS curve corresponding to the glycerol–tobacco mixture also reflects differences in the temperature and relative importance of peaks, with respect to the DTG-curve. The interpretation of the GS curve is not as clear as that of the DTG curve, where only the weight loss must be considered, and requires the analysis of the IR spectra of the gases obtained in each step. The event which in the DTG curves of Fig. 1 was attributed to the glycerol loss (in the range of 140–220 ◦ C) is not as clear in the GS curves of Fig. 2, and two events of gas generation occur, one of them with GS-maximum at around 193 ◦ C and the other appearing as a shoulder of the main peak, at around 243 ◦ C. The assignment of such decomposition steps should be carried out from the composition of the evolved gases. Fig. 3 shows the three-dimensional (3D) diagram corresponding to the pyrolysis of tobacco, glycerol and the tobacco–glycerol mixture, where the absorbance corresponding to the vibrational modes of the different chemical bonds and functional groups of compounds appearing in the gases evolved in the TGA furnace at each temperature versus the wavenumber and versus the temperature represented. As can be seen, the general aspect of the 3D diagram corresponding to the glycerol–tobacco mixture is somewhat similar to that of tobacco; nevertheless, some differences in the relative intensity and the wavenumber of several peaks, which are reflected in the respective GS curves of Fig. 2, are significant. Crossing sections of the 3D diagrams in a parallel way to the temperature axis for the different maxima of absorbance obtained at each of the different vibrational modes (i.e., for fixed wavenumber) allows the evolution of such an IR band to be observed, whereas crossing the diagrams in a way parallel to the wavenumber axis permits the IR spectra corresponding to the gases evolved at each temperature to be obtained, thus indicating the type of the chemical bonds or functional groups present in those gases at that moment. As can be seen, different from tobacco, the 3D diagram corresponding to glycerol shows the appearance of the same IR bands during the overall TGA run, thus corroborating the previous consideration that glycerol is probably evaporated in the TGA furnace and after simultaneous condensation and evaporation in the transfer line occur, resulting in the continuous arrival of glycerol vapors to the IR cell. On the other hand, the 3D diagrams of tobacco and the glycerol–tobacco mixture show different bands and/or differences in their relative intensity depending on the temperature, thus indicating that a more detailed analysis of the spectrum of the gases evolved in each decomposition step is required, as well as a study of the evolution with the time of the intensity of the different IR bands, in order

A. Gómez-Siurana et al. / Thermochimica Acta 573 (2013) 146–157

149

Relave intensity (displaced scale)

0,004

0,003

0,002

0,001

0 4000

3500

3000

2500 2000 wavenumber (cm-1 )

1500

1000

500

Fig. 4. Normalized FTIR spectra obtained for the gas evolved at several selected temperatures in the TGA analysis of glycerol. The relative intensity has been obtained by dividing each intensity value by the mass of glycerol pyrolyzed.

obtained at the maxima and shoulders of the Gram–Schmidt thermograms for each system of Fig. 2 have been selected. In this way, Fig. 4 shows, as an example, the FTIR normalized spectra obtained for glycerol pyrolysis at several selected temperatures. Each normalized spectra has been obtained by dividing the absorbance by the initial mass of glycerol. As it can be seen, the obtained spectra are very similar, and the bands in the range of 4000–3500 cm−1 and 2000–1300 cm−1 clearly indicate the presence of water in the gas evolved in the overall range of temperature. The IR spectrum of pure glycerol shows a sharp and intense band at 1107–1028 cm−1 corresponding to C O stretching vibrations and bands in the range of 3000–2800 cm−1 corresponding to stretching C H vibrations, as well as others at 1300–1400 cm−1 and 3650 cm−1 , associated with C H and O H vibrations respectively that, in this case, will appear overlapped with the above-mentioned bands of water. According to Fig. 4, these glycerol bands could start at around 180 ◦ C (in fact, according to the 3D diagram of Fig. 3, at a slightly lower temperature), and can be more clearly observed as the temperature increases, thus corroborating the continuous presence of glycerol in the gas evolved from the TGA furnace. Moreover, as the temperature increases, the appearance of bands at 2322 and 2353 cm−1 reveals the presence of CO2 in the gas evolved, and suggests that some decomposition also occurs. Therefore, the continuous presence of water would be the consequence of dehydration processes simultaneous to glycerol evaporation, where, besides water, 0,02

to obtain information regarding the qualitative composition of the gases evolved at each temperature and on the differences in the pyrolytic behavior of the tobacco and glycerol–tobacco mixture. 3.2. Analysis of the FTIR spectra obtained at each temperature According to the bibliography [40], the appearance of a strong IR signal associated with water can hide the detection of some gases, as H2 and Cl2 , which could be evolved from tobacco pyrolysis, thus being undetectable with FTIR under the conditions used in this work. However, as has been previously mentioned, FTIR provides a useful tool for the understanding of the composition of the gases evolved at each temperature. With this objective, the FTIR spectra

Intensity (displaced scale)

0,015

Fig. 3. 3D IR spectrum of tobacco and glycerol.

0,01

0,005

0 4000

3500

3000

2500

2000

1500

1000

500

wavenumber (cm-1) Fig. 5. Normalized FTIR spectra obtained for the gas evolved at several selected temperatures in the TGA analysis of the reference tobacco. The relative intensity has been obtained by division of each intensity value by the mass of tobacco pyrolyzed.

150

A. Gómez-Siurana et al. / Thermochimica Acta 573 (2013) 146–157

(a)

(d)

(b)

(e)

(c)

(f)

Fig. 6. FTIR spectra corresponding to the gas evolved at several selected temperatures in the TGA analysis of the reference tobacco.

other non-volatile materials should also be formed that are thermally decomposed at higher temperatures, with CO2 formation. This behavior is in good agreement with the low slow weight loss observed in the TGA curve of glycerol at temperatures above 240 ◦ C (Fig. 1). Moreover, before the glycerol weight loss starts, at around 120 ◦ C, a weight loss of around 2% can be measured that, in accordance with Fig. 4, corresponds to water loss, whereas the IR bands of chemical bonds in alcohols as well as those corresponding to CO2 begin to be appreciable at higher temperatures (beyond 180 ◦ C). In accordance with the 3D diagrams shown in Fig. 3, the case of tobacco is quite different to that of glycerol, and the evolution with the temperature of the IR bands corresponding to the gas evolved from the TGA furnace (i.e., the pathway followed by the composition of the gas evolved along the TGA run) presents noticeable changes, not only in their relative intensities, but also in the wavenumber or, in other words, in the corresponding chemical

bonds. Thus, Fig. 5 shows the normalized FTIR spectra obtained at selected temperatures, chosen from the GS-curve of Fig. 2, where the relative intensity was calculated dividing by the initial mass of tobacco. As it can be seen, the appearance of maxima at different temperatures and for different bands suggests the need of a careful analysis of the evolution with the temperature of several selected bands. Thus, the assignment of the main IR bands in some selected normalized FTIR spectra has been carried out, as is shown in Fig. 6. According to Fig. 6a, at around 113 ◦ C, the spectrum is mainly dominated by the spectrum of water, with the characteristic shape of bands in the range of 4000–3500 cm−1 and 1800–1300 cm−1 , despite the fact that the bans of CO2 (2322 and 2353 cm−1 ) also appear. At 193 ◦ C (Fig. 6b), water still appears, but the bands corresponding to CO2 have increased noticeably and, moreover, the bands previously observed for glycerol (1107–1028 and 3000–2800 cm−1 ) start to be appreciable. At 288 ◦ C (Fig. 6c),

A. Gómez-Siurana et al. / Thermochimica Acta 573 (2013) 146–157

151

Relave intensity difference

but with absorbance very close to that of the CO2 band. The positive bands appearing in Fig. 7, resulting from the spectra subtraction, indicate that the relative content of the corresponding functional groups is higher in the gas evolved in from the tobacco–glycerol mixture than that from tobacco, and lower in the reverse case. The first comment on considering Fig. 7 is that the spectra resulting from subtraction are not the same as those shown in Fig. 4, corresponding to glycerol. Thus, it could be concluded that there exists some type of glycerol–tobacco interaction which has as a consequence somewhat changes in the respective thermal behavior. According to Fig. 7:

4000

3500

3000

2500

2000

1500

1000

500

wavenumber (cm-1) Fig. 7. Resulting spectra after subtracting the normalized spectrum obtained, at each temperature, from the gases evolved in the pyrolysis of the glycerol–tobacco mixture and tobacco. Before subtraction, each spectrum has been rescaled between 0 and 1. The scale of the spectrum resulted from subtraction has been displaced in order to facilitate the visualization of the result at each temperature.

the CO2 bands are still the most intense, but other bands appear: at 2187 and 2104 cm−1 – characteristics of carbon monoxide –, in the range of 1800–1710 cm−1 – corresponding to carbonyl (C O) groups in aldehydes and carboxylic acids –, and in the range of 1145–1210 cm−1 – corresponding to C O bonds in alcohols and phenols –; moreover, the shape of bands in the ranges of 4000–3500 and 2000–1300 cm−1 changes, and deviates from the typical spectrum of water, reflecting the appearance of other O H bonds, i.e., O H bonds in alcohols and other hydroxyl compounds. The normalized spectrum obtained at 338 ◦ C is very similar to that shown in Fig. 6c, corresponding to the gases evolved at 288, but at 433 ◦ C (Fig. 6d), some of the above mentioned bands have almost disappeared, as those assigned to carbon monoxide and C O bonds, whereas those corresponding to carbonyl groups and aliphatic C H bonds have increased noticeably. The normalized spectra at 456 ◦ C and 471 ◦ C are also very close (Fig. 6e), and show as the main difference with that obtained at 433 ◦ C, the noticeable increase of a band at around 3012 cm−1 , associated with the presence of CH2 groups and, perhaps, aromatics (at around 3070 cm−1 ) and once again, the incipient presence of CO. The spectrum at 496 ◦ C is very close to that of Fig. 6e, and only reflects the increase of the bands associated to CO. Finally, the spectrum obtained at 656 ◦ C (Fig. 6f) reflects the presence of CO, CO2 and water as main components in the gas evolved, which is coherent with the decomposition of endogenous carbonates [46] and/or from an ulterior pyrolysis of the organic solid residue (i.e., coke) remaining at these temperatures yielding CO, CO2 and water. Once the main characteristics of the decomposition profiles for glycerol and tobacco have been established, the pyrolytic behavior of the glycerol–tobacco mixture was analyzed. In order to obtain information about the influence of the presence of glycerol on the composition of the products of pyrolysis, the spectra obtained in the tobacco pyrolysis at selected temperatures have been subtracted from the spectra corresponding to the pyrolysis of the glycerol–tobacco mixture at the same temperature. In order to avoid the influence of the mass of sample pyrolyzed in each step and at each temperature, the spectra have been scaled between 0 and 1 before subtraction. Fig. 7 shows the results of this comparison. It must be taken into account that at temperatures higher than 150–160 ◦ C, the CO2 bands are the most intense in both cases, tobacco and mixture, with the only exception of the spectrum corresponding to the gas evolved from the mixture at around 590 ◦ C, where the most intense band appears at 3016 cm−1 ( CH2 groups),

• In the low temperatures range, the result of the spectra subtraction appears to exhibit some background noise which hinders the interpretation. However, below 170 ◦ C, no significant differences have been found between tobacco and the mixture with glycerol, thus showing that in both cases, water evaporation takes place and, thus, the normalized spectrum corresponding to the gases evolved from the TGA furnace is practically the same in both cases, i.e., the water spectrum, as Fig. 6a reflects. The only difference is that the evolution of CO2 seems to start slightly earlier in the case of tobacco, as the subtracted spectrum at 153 ◦ C reflects. • The main difference in the 160–200 ◦ C range, represented by the results at 193 ◦ C, lies in the CO2 generation, which seems to be higher in the mixture than in the tobacco of reference. As it has been pointed out previously, some pyrolysis (not evaporation) of glycerol could be starting at this step (see Fig. 4), resulting in an increase of the CO2 content in the gases evolved from the mixture. • In the range of temperatures corresponding to the main decomposition step, where hemicellulose and cellulose, among others, are decomposed (representative spectra at 288 and 338 ◦ C), no significant differences in the composition of the gases evolved have been observed as a consequence of the addition of glycerol, and only and incipient increase of the CO2 content in the gases obtained from the mixture can be observed at the temperature associated with cellulose pyrolysis. This is in agreement with the profiles shown by the tobacco and mixture GS curves of Fig. 4, and the analysis of the corresponding FTIR spectra reveals that the addition of glycerol does not affect the composition of the gases evolved in the hemicellulose and pectin decomposition step, but this increases the CO2 content in the gasses obtained in the range of temperature related with cellulose pyrolysis. • In the 400–500 ◦ C range of temperature, the gas evolved from the mixture appears to have a higher content of CO2 and alkanes and/or alkyl substituted compound content and lower amount of water. Moreover, the differences in the alkyl groups compound content increases as the temperature increases. It is interesting to observe that, in this range of temperature, the spectra obtained from glycerol (Fig. 4) showed the appearance of bands corresponding to O H, C O and C H bonds, but the spectra corresponding to the gases evolved from the mixture with tobacco only show the increase of CH3 and CH2 bands. Thus, the addition of glycerol to tobacco could result in an eventual interaction between lignin – which started the decomposition at low temperatures and has been pyrolyzing in a wide range of temperature – and other non volatile materials remaining in the TGA furnace at these temperatures as well as the non-volatile residue obtained from the fractions of glycerol that yielded CO2 in previous steps and the gases evolved from the pyrolysis are different than the sum of the gases obtained from tobacco and glycerol. • The effect of glycerol is even more noticeable in the range of 500–600 ◦ C, where the differences in the alkanes, aromatics and CO content are higher. In order to provide a clearer idea of the above-mentioned differences, Fig. 8 shows the comparison

152

A. Gómez-Siurana et al. / Thermochimica Acta 573 (2013) 146–157

• Finally, above 600 ◦ C, the main differences are the presence of more CO and CO2 and less water in the gas evolved from the tobacco–glycerol mixture. Tobacco

Relave intensity

3.3. Semiqualitative study of the evolution of the gases evolved

Glycerol-tobacco mixture

4000

3500

3000

2500 2000 wavenumber (cm-1)

1500

1000

500

Fig. 8. Comparison between the normalized spectrums obtained at 586 ◦ C from the gases evolved in the pyrolysis of the glycerol–tobacco mixture and tobacco.

between the scaled spectra of the gases evolved from the pyrolysis of a tobacco and tobacco glycerol mixture at 586 ◦ C, where the above-mentioned differences can be clearly observed. As can be seen, in this range of temperature, tobacco yields mainly water and CO2 as well as unsaturated compounds (with CH2 groups), whereas CO and other compounds, probably oxygenated compounds, are also evolved from the mixture. Specifically, the noticeable increase of bands in the range of carbonyl groups (at around 1750 cm−1 ) and CH2 groups might be in agreement with the formation of acrolein and acetaldehyde reported in the bibliography [35].

The previous analysis of the representative FTIR spectra of the gas evolved from TGA at different temperatures serves to select a series of IR band characteristics of the compounds appearing in the gases from pyrolysis at different temperatures, whose evolution with the temperature should be analyzed in order to obtain information about the correspondence between the processes involved in each degradation step and the qualitative composition of the products of pyrolysis. The bands selected in this work as well as their assignment are shown in Table 2. The assignments have been performed according to the bibliography [47], and also comparing the obtained spectra from the NIST database (http://webbook.nist.gov/chemistry/name-ser.html.es). Table 2 also shows some specific bibliographic references where similar IR band assignments in products of tobacco smoking are reported. The profiles showing the evolution with the temperature of the relative intensity of selected IR bands representative of several functional groups in the gases evolved in the pyrolysis of tobacco and the glycerol–tobacco mixture are presented in Fig. 9. According to Ahamad and Alshehri [40], The main gases released in tobacco pyrolysis under dynamic TGA conditions are H2 O, CO2 , CO, NH3 , HCN, NO, NO2 and some volatile organic compounds (low chain carbonylic compounds as aldehydes and acids, alcohols and phenols, alkanes, alkenes and aromatics). Some of these compounds can be followed through evolution with time of the

Table 2 IR bands selected for the analysis of the evolution with the temperature of the pyrolysis gases. Wavenumber (cm−1 )

Origin

Assignment

References

3640–3550 (selected: 3566)

O H Symmetrical and asymmetrical stretching

Water

[36,38,40]

3130–3070 1600–1500 900–670 (several) (selected: 3076)

Aromatic C H in plane bend Aromatic C C C ring stretch Aromatic C H out of plane bend

Aromatics and heterocycles with 2 C C as furan and furfural

[40,48]

3040–3010 1600–1680 (selected: 3016)

C H stretch in alkenes (in C C stretch

Alkenes

[40,48]

2970–2950 2880–2860 1470–1430 1380–1370 2935–2915 2865–2845 1485–1445 (selected: 2968, 2835)

Methyl C H asymmetric stretch Methyl C H symmetric stretch Methyl C H asymmetric bend Methyl C H asymmetric bend Methylene C H asym. stretch Methylene C H sym. stretch Methylene C H bend

Alkanes or alkyl substituents

[38,40,48]

2400–2224 (selected: 2361)

Asymmetrical stretching in O C O

CO2

[36,38,40]

2180–2108 (selected: 2195)

Stretching vibration in CO

CO

[36,38,40]

2250

CH2 )

HNCO

[36]

1900–1600 (selected: 1749)

Carbonyl groups

Aldehydes, ketones

[38,40]

1050–1200 (selected: 1047, 1191)

C O stretch

Alcohols, phenols

[38]

3400–3325 1130–1360

N H stretch C N stretch

[40]

966

NH3

[36]

714

HCN

[36]

A. Gómez-Siurana et al. / Thermochimica Acta 573 (2013) 146–157

intensity of representative IR bands shown in Fig. 9, where the relative absorbance (i.e., absorbance divided by the mass of sample pyrolyzed) of selected bands in the gases evolved from the tobacco and the tobacco-mixture pyrolysis versus temperature is shown. It should be taken into account that the case of glycerol is quite different, and a continuous increase of the bands representative of carbonyl groups (1749 cm−1 ) and water (3566 cm−1 ) occurs, with a maximum at around 90 ◦ C, as well as some CO2 generation above 250 ◦ C. Fig. 9a shows the evolution of the band at 3566 cm−1 , which has been selected as representative of water. As can be seen, although water is almost continuously evolved, two main episodes of water generation appear, below 160 ◦ C, related with the evaporation of adsorbed water, and the main one, with a maximum at around 335 ◦ C which could result from the overlapping of the decomposition steps of hemicellulose–pectin and cellulose, through the cleavage of the hydroxyl groups in the lateral chains of these polymers [38]. Water can also evolve from the decomposition of other substances: as an example, Kruse et al. [36] have studied the pyrolysis of polypeptides by TG/FTIR, showing that in the range 30–180 ◦ C only water was detected, probably from intra or intermolecular water cleavage. The addition of glycerol results in an (expected) increase of the water release at lower temperatures (below 190 ◦ C) and in the range of temperature corresponding to the glycerol weight loss step (at around 235 ◦ C), as well as in an unexpected increase at around 500 ◦ C, which is in agreement with the results shown in Fig. 7 which reflects noticeable changes in the pyrolytic behavior of the mixture in comparison with that of tobacco, as a consequence of the presence of glycerol.

153

Ahamad and Alshehri [40] reported that aromatics are formed at temperatures up to 500 ◦ C. Fig. 9b shows the evolution of the IR band at 3076 cm−1 , characteristic of aromatics, in the gases evolved from tobacco and mixture, and even though in the experimental conditions used in the present work for the FTIR spectra acquirement, the aromatic bands (Table 2) cannot be clearly distinguished (see Fig. 6e), the appearance of a maximum at around 530 ◦ C suggests that the generation of aromatic products could be attributed to the lignin and other compounds contained in tobacco that decomposes in this pyrolysis step, and the presence of glycerol does not contribute in a significant way to the aromatics formation. Fig. 9c shows the evolution of the IR band at 3016 cm−1 , which has been selected as representative of CH2 groups. As can be seen, this band shows a very well-defined maxima at around 530 ◦ C, that is associated with the lignin decomposition step, as aromatics. The appearance of these type of compounds could be related with alkene formation (1,3-butadiene, isoprene) but, if the characteristic carbonyl groups IR bands appear simultaneously, could also correspond with some unsaturated aldehydes, as acrolein and crotonaldehyde or even to hydrogen cyanide, if the presence of the nitrile group could be assessed. Moreover, the appearance of this type of compound seems to be only related with tobacco pyrolysis, and is not influenced noticeably by the addition of glycerol. The case of aliphatic methyl and methylene groups is very interesting. As can be seen in Fig. 9d and e, there are three maxima for the generation of these compounds – actually there is a maximum with two shoulders – but the respective temperatures and their relative importance change, and it appears that methylene groups

Fig. 9. Evolution of different gaseous products with temperature in the pyrolysis of tobacco and the glycerol–tobacco mixture. (a) Water, (b) aromatics, (c) CH2 , (e) CH3 , (f) CO2 , (g) CO, (h) CO and (i) OH.

CH2 groups, (d)

154

A. Gómez-Siurana et al. / Thermochimica Acta 573 (2013) 146–157

Fig. 9. (Continued ).

are favored at lower temperatures (maximum at around 340 ◦ C and shoulders at around 250 ◦ C and 480 ◦ C), whereas methyl groups are favored at higher temperatures (maximum at around 480 ◦ C and shoulders at around 250 ◦ C and 340 ◦ C). According to the respective temperatures, the maximum of methylene should correspond with the decomposition step associated with the pyrolysis of cellulose, whereas the maximum of CH3 groups should correspond to the step related with the pyrolysis of lignin. Therefore, it seems that, within the range of temperature of the main weight loss, the formation of CH2 groups in front of the formation of CH3 groups is favored at low temperatures, perhaps indicating the formation of longer or less branched products, whereas the inverse

tendency could occur at high temperatures. Te addition of glycerol does not noticeably affect, despite that a slight increase of methyl and methylene groups in the range of 380–590 ◦ C an a decrease at around 350 ◦ C seems to occur. Fig. 9f shows the evolution profile obtained for CO2 . As can be seen, the curves reflect in a reliable way the shape of the GS curves shown in Fig. 2, thus indicating that the formation of CO2 occurs in all the pyrolysis steps. At the temperatures of the main decomposition peaks (i.e., 290 and 340 ◦ C), the CO2 yields obtained from tobacco at both decomposition steps are similar, despite that in accordance with Fig. 2 the weight loss involved in the first one is higher than in the second one (thus, the mass of CO2 generated

A. Gómez-Siurana et al. / Thermochimica Acta 573 (2013) 146–157

per mass of decomposed solid should be lower in the second event). Nevertheless, in the case of the mixture, the yield of CO2 obtained at around 340 ◦ C is higher than at 290 ◦ C, and similar to that obtained in the absence of the added glycerol, whereas at 290 ◦ C the addition of glycerol seems to result in a decrease of the yield of CO2 . It is interesting to observe that, according to Fig. 1, the TGA of the tobacco–glycerol reflected that the decomposition step associated with the hemicellulose pyrolysis had apparently almost disappeared and overlapped with the step associated with the glycerol loss. However, the FTIR results indicate that this step still occurs, and is a well differentiated decomposition step. In the range of 230–260 ◦ C, glycerol contributes to a slight increase of the yield of CO2 , as well as temperatures higher than 600 ◦ C. Zhou et al. [38] suggested that the formation of CO2 from pyrolysis of pectin in the 200–300 ◦ C range can be attributed to cleavage of carboxylic groups in lateral chains, whereas above 300 ◦ C, it is due to the cracking of the O C O group. The evolution of CO is presented in Fig. 9g. As can be seen, a maximum appears at around 340 ◦ C that suggests that the CO formation could be more related with the cellulose pyrolysis step than with the hemicellulose pyrolysis step. Another event of CO generation appears at around 500 ◦ C, which is associated with lignin pyrolysis step. Moreover, a noticeable increase of the yield of CO can be observed at temperatures higher than 600 ◦ C, probably related with char decomposition [38]. These results are in good agreement with the data reported by Baker [1], who described the existence of two distinct formation regions for carbon oxides, a low-temperature region, at around 100–450 ◦ C and a high-temperature region, in the range 550–900 ◦ C. Moreover, in the low-temperature region, considerable amounts of water were also obtained, and a dependence of the amount of CO in the lowtemperature region and CO2 in the high-temperature region on the heating rate was observed. Baker [1] suggested that CO could undergo reactions in phase gas, whereas CO2 undergoes both, gas phase and heterogeneous reactions, the latter with the carbon char at the higher temperature. Kruse et al. [36] also reported the appearance of CO2 at around 200 ◦ C, which was attributed to decarboxylation, and CO in the range 300–450 ◦ C. The yield of CO seems not to be much affected by the addition of glycerol, although it seems to be slightly increased at temperatures higher than 400 ◦ C. Fig. 9h and i shows, respectively, the evolution with the temperature of the yield of carbonyl and hydroxyl compounds. The profiles obtained in the range where the maximum weight loss takes place are similar to that of CO (Fig. 9g), thus indicating that the aldehydes and ketones as well as the alcohol and phenol formation from tobacco pyrolysis goes along with the CO formation. Nevertheless, the increase observed for CO at higher temperatures is not observed for carbonyl nor hydroxyl compounds, showing that they are not evolved from the char decomposition. Carbonyl groups also show a maximum below 150 ◦ C which has not been observed for hydroxyl compounds or for CO and that is significantly increases in the case of the mixture thus suggesting that is clearly related with the addition of glycerol. The yields obtained in the pyrolysis steps corresponding to the maxima at 290 ◦ C and 340 ◦ C are higher for tobacco than for the mixture, revealing that the formation of these compounds in these pyrolysis steps is decreased as a consequence of the addition of glycerol. Contrarily, below 250 ◦ C and above 500 ◦ C the yield of carbonylic compounds, but not hydroxylic compounds, increases in the mixture. The study of the presence of some nitrogen compounds reported as products of tobacco pyrolysis, as NH3 , amines, HCN and HCNO [36,40], is difficult because their IR characteristic bands could be hidden by the bands corresponding to other compounds. Bands characteristics of N H and C N stretching, in the range of 3400–3325 cm−1 and 1360–1130 cm−1 , respectively, will be

155

overlapped with O H and C O stretching in water and alcohols [47]. Kruse et al. [36] identify the presence of NH3 , HCN and HNCO in the gases evolved from the pyrolysis of dipeptides by the bands at 966, 714 and 2250 cm−1 . In the present work, the band associated with HNCO has not been observed, whereas the band of NH3 shows very low intensity and the band of HCN is overlapped with other bands appearing in the lowest wavenumber zone (see Figs. 5, 6 and 8). The superposition of all the profiles shown in Fig. 9 permits us to obtain some idea about the compounds evolved in each decomposition step, and if the GS curves are also superposed, an assignment to each decomposition step can be also performed. Thus: • In the step associated to water decomposition (at around 100 ◦ C), not only has water been detected, but also some carbonyl compound generation which, considering that tobacco and tobacco–glycerol mixtures produce similar amounts, would evolve from evaporation of low molecular mass oxygenated compounds or from decomposition of glycerol or other compounds. • The decomposition step at around 200 ◦ C, which was related to the glycerol loss, is characterized by the formation of CO2 , water and carbonyl and hydroxyl compounds, which could be related with the evaporation and decomposition of glycerol and other components of the tobacco blend. • The DTG curves of Fig. 1 revealed a noticeable increase of a decomposition step at around 240 ◦ C which was clearly related with the addition of glycerol. This step is characterized by the formation of water, carbonyl and hydroxyl compounds, CO2 , and CH2 and CH3 containing compounds. CO, aromatic and CH2 compounds do not seem to be formed in significant amounts at this step. • The steps of maximum decomposition rate at around 290 ◦ C and 340 ◦ C, assigned to the hemicellulose and cellulose pyrolysis, respectively, seem to involve the formation of water, CH2 and CH3 containing compounds, CO, CO2 and carbonyl and hydroxyl compounds. An incipient evolution of alkenes and aromatic compounds would also have taken place at these steps. • In the range of 410–550 ◦ C, where the main decomposition of lignin takes place, the maxima of methyl groups, aromatics and carbon–carbon double bonds occurs. Moreover, CO and methylene groups also show secondary maxima as well as, to a lesser extent, carbonyl groups. Hydroxyl compounds and CO2 are also formed at this step. Thus, the possible formation of several hydrocarbon compounds (unsaturated, saturated and aromatic) and CO can be suggested as the main products of the lignin cracking, and short-chain and/or methyl-substituted hydrocarbons are favored. The maxima for the formation of aromatics and unsaturated compounds appear at slightly higher temperatures than the CH3 maximum, also suggesting differences in the routes and temperatures for the formation of the different compounds. • At temperatures higher than 600 ◦ C, practically only CO and CO2 are formed. The maximum observed for CO2 at this range of temperature could be attributed to the decomposition of inorganic materials, as carbonates. The pyrolysis of the residual char seems to increase as the temperature increases, with the formation of CO2 , CO and, perhaps, some aromatic compounds. 4. Conclusions The results obtained in this work allow us to establish some pathways related with the qualitative composition of the gases evolved from tobacco pyrolysis at each temperature, as well as the effect associated with the use of glycerol as a cigarette ingredient. The more significant IR bands in the obtained IR spectra have been selected and considered as representative of compounds or groups of compounds generated from the pyrolysis of tobacco

156

A. Gómez-Siurana et al. / Thermochimica Acta 573 (2013) 146–157

and glycerol, and their behavior through the TGA runs has been analyzed. In this way, at around 100 ◦ C, besides water evolution, some carbonyl compound generation has been observed, which has been attributed to the decomposition of glycerol or other low molecular mass oxygenated compounds. Water, carbonyl and hydroxyl compounds and CO2 are formed in the decomposition step associated with the glycerol loss, at around 200 ◦ C. At higher temperatures, when the main pyrolysis step takes place, the composition of the evolved gases show several changes depending on the temperature, and at around 240 ◦ C water, carbonyl and hydroxyl compounds, CO2 , and CH2 and CH3 containing compounds are evolved. As the temperature increases CO and aromatic and CH2 are formed, and in the range of 410–550 ◦ C the maxima of methyl groups, aromatics and carbon–carbon double bound compounds appear. At temperatures higher than 600 ◦ C, practically only CO and CO2 are formed. Depending on the temperature, the presence of glycerol seems to change the pathway of the evolved gases composition and, as an example, increases the formation of the following compounds: • water at low temperatures • CH2 and CH3 containing compounds at around 450 ◦ C • CO2 in the cellulose decomposition step (at around 340 ◦ C) • carbonyl and carboxyl compounds at temperatures lower than 250 ◦ C and higher than 500 ◦ C Acknowledgements Financial support for this investigation has been provided by the Spanish “Secretaría de Estado de Investigación” del Ministerio de Economía y Competitividad (MAT2011-24991) and by the Generalitat valenciana (PROMETEO/2012/015). References [1] R.R. Baker, The formation of oxides of carbon by the pyrolysis of tobacco, Beit. zur Tabakf. 8 (1) (1975) 16–27. [2] Z. Czégény, M. Blazsó, G. Várhegyi, E. Jackab, C. Liu, L. Nappi, Formation of selected toxicants from tobacco under different pyrolysis conditions, J. Anal. Appl. Pyrolysis 85 (2009) 47–53. [3] R.R. Baker, Smoke generation inside a burning cigarette: modifying combustion to develop cigarette that may be less hazardous to health, Prog. Energy Combust. Sci. 32 (2006) 373–385. [4] R.R. Baker, L.J. Bishop, The pyrolysis of non-volatile tobacco ingredients using a system that simulates cigarette combustion conditions, J. Anal. Appl. Pyrolysis 74 (2005) 145–170. [5] R.R. Baker, The generation of formaldehyde in cigarettes—overview and recent experiments, Food Chem. Toxicol. 44 (2006) 1799–1822. [6] R. Hertz, T. Streibel, Ch. Liu, K. McAdam, R. Zimmermann, Microprobe samplingphoto ionization time-of-flight mass spectrometry for in situ analysis of pyrolysis and combustion gases: examination of the thermo-chemical processes within a burning cigarette, Anal. Chim. Acta 714 (2012) 104–113. [7] C. Busch, T. Streibel, C. Liu, K.G. McAdam, R. Zimmermann, Pyrolysis and combustion of tobacco in a cigarette smoking simulator under air and nitrogen atmosphere, Anal. Bional. Chem. 103 (2012) 419–430. [8] C.H. Yun, W.J. Kim, S.C. Yi, Modelling and simulation of evaporation–pyrolysis processes of a naturally smoldering cylindrical cellulosic material rod: effect of smoldering rate on product concentrations, Ind. Eng. Chem. Res. 14 (2008) 120–130. [9] S.C. Yi, M.R. Hajaligol, Product distribution from the pyrolysis modeling of tobacco particles, J. Anal. Appl. Pyrolysis 66 (2003) 217–234. [10] A. Rostami, J. Murthy, M.R. Hajaligol, Modelling of a smoldering cigarette, J. Anal. Appl. Pyrolysis 66 (2003) 281–301. [11] M.S. Saidi, M.R. Hajaligol, F. Rasouli, Numerical simulation of a burning cigarette during puffing, J. Anal. Appl. Pyrolysis 72 (2004) 141–152. [12] A. Rostami, M.R. Hajaligol, Modelling the diffusion of carbon monoxide and other gases from the paper wrapper of a cigarette during puffing, J. Anal. Appl. Pyrolysis 66 (2003) 263–280. [13] K. Torikai, S. Yoshida, H. Takahashi, Effects of temperature, atmosphere and pH on the generation of smoke compounds during tobacco pyrolysis, Food Chem. Toxicol. 42 (2004) 1409–1417. [14] K. Torikai, Y. Uwano, T. Naakamori, W. Tarora, H. Takahashi, Study on tobacco components involved in the pyrolytic generation of selected smoke constituents, Food Chem. Toxicol. 43 (2005) 559–568.

[15] E. Gregg, C. Hill, M. Hollywood, M. Kearney, K. McAdam, D. McLaughlin, S. Purkis, M. Williams, The UK smoke constituents testing study. Summary of results and comparison with other studies, Beit. zur Tabakf. 21 (2) (2004) 117–138. [16] T.E. McGrath, A.P. Brown, N.K. Meruva, W.G. Chan, Phenolic compound formation from the low temperature pyrolysis of tobacco, J. Anal. Appl. Pyrolysis 84 (2009) 170–178. [17] S. Purwono, B. Murachman, J. Wintoko, N.E. Permatasari, D. Lidyawati, Modeling of the rate of weight loss in the pyrolysis of waste of tobacco stem for the development of charcoal briquette, Mater. Sci. Eng. B 1 (4) (2011) 479–484. [18] M.K. Akalin, S. Karagoz, Pyrolysis of tobacco residue. Part 1. Thermal, Therm. BioResour. 6 (2) (2011) 1520–1531. [19] M.K. Akalin, S. Karagoz, Pyrolysis of tobacco residue. Part 2. Catalytic, Therm. BioResour. 6 (2) (2011) 1773–1805. [20] C.R. Cardoso, M.R. Miranda, K.G. Santos, C.H. Ataíde, Determination of kinetic parameter and analytical pyrolysis of tobacco waste and sorghum bagasse, J. Anal. Appl. Pyrolysis 92 (2011) 392–400. [21] Y. Yang, T. Li, S. Jin, Y. Lin, H. Yang, Catalytic pyrolysis of tobacco rob: kinetic study and fuel gas produced, Bioresour. Technol. 102 (2011) 11027–11033. [22] R. Bassilakis, R.M. Carangelo, M.A. Wójtowicz, TG-FTIR analysis of biomass pyrolysis, Fuel 80 (2001) 1765–1786. [23] C.J. Booker, R. Bedmutha, T. Vogel, A. Gloor, R. Xu, L. Ferrante, K.K.C. Yeung, I.M. Dcott, K.L. Conn, F. Berruti, C. Briens, Experimental investigations into the insecticide, fungicidal, and bactericidal properties of pyrolysis bio-oil from tobacco leaves using a fluidized bed pilot plant, Ind. Eng. Chem. Res. 49 (2010) 10074–10079. [24] E.L. Carmines, C.L. Gaworski, Toxicological evaluation of glycerin as a cigarette ingredient, Food Chem. Toxicol. 43 (2005) 1521–1539. [25] R.R. Baker, S. Coburn, C. Liu, The pyrolytic formation of formaldehyde from sugars and tobacco, J. Anal. Appl. Pyrolysis 77 (2006) 12–21. [26] R.R. Baker, S. Coburn, C. Liu, J. Tetteh, Pyrolysis of saccharide tobacco ingredients: a TGA–FTIR investigation, J. Anal. Appl. Pyrolysis 74 (2005) 171–180. [27] C.L. Gaworski, R. Lemus-Olalde, E.L. Carmines, Toxicological evaluation of potassium sorbate added to cigarette tobacco, Food Chem. Toxicol. 48 (2008) 339–351. [28] E.L. Carmines, R. Lemus, C.L. Gaworski, Toxicological evaluation of licorice extract as a cigarette ingredient, Food Chem. Toxicol. 43 (2005) 1303– 1322. [29] R.R. Baker, J.R. Pereira da Silva, G. Smith, The effects of tobacco ingredients on smoke chemistry. Part I. Flavourings and additives, Food Chem. Toxicol. 42S (2004) S3–S37. [30] R.R. Baker, J.R. Pereira da Silva, G. Smith, The effects of tobacco ingredients on smoke chemistry. Part II. Casing ingredients, Food Chem. Toxicol. 42S (2004) S9–S52. [31] C. Liu, A. Parry, Potassium organic salts as burn additives in cigarettes, Beit. zur Tabakf. 20 (5) (2003) 341–347. [32] A. Gómez-Siurana, A. Marcilla, M. Beltran, I. Martinez, D. Berenguer, R. García-Martínez, T. Hernández-Selva, Thermogravimetric study of the pyrolysis of tobacco and several ingredients used in the fabrication of commercial cigarettes: effect of the presence of MCM-41, Thermochim. Acta 523 (2011) 161–169. [33] A. Gómez-Siurana, A. Marcilla, M. Beltran, I. Martinez, D. Berenguer, R. García-Martínez, T. Hernández-Selva, Study of the oxidative pyrolysis of tobacco–sorbitol–saccharose mixtures in the presence of MCM-41, Thermochim. Acta 530 (2012) 87–94. [34] I. Schmeltz, A. Wenger, D. Hoffmann, T.C. Tso, Chemical studies on tobaco smoke. 63. On the fate of nicotine during pyrolysis and in a burning cigarette, Agric. Food Chem. 27 (3) (1979) 608–882. [35] J.B. Paine III, Y.B. Pithawalla, J.D. Naworal, C.E. Thomas Jr., Carbohydrate pyrolysis mechanisms from isotopic labeling Part 1: the pyrolysis of glycerin: discovery of competing fragmentation mechanisms affording acetaldehyde and formaldehyde and the implications for carbohydrate pyrolysis, J. Anal. Appl. Pyrolysis 80 (2007) 297–311. [36] J. Kruse, K.U. Eckhardt, T. Regier, P. Leinweber, TF-FTIR, LC/MS, XANES and Py-FIMS to disclose the thermal decomposition pathways and aromatic N formation during dipeptide pyrolysis in a soil matrix, J. Anal. Appl. Pyrolysis 90 (2011) 164–173. [37] O. Senneca, S. Ciaravolo, A. Nunziata, Composition of the gaseous products of pyrolysis of tobacco under inert and oxidative conditions, J. Anal. Appl. Pyrolysis 79 (2007) 234–243. [38] S. Zhou, Y. Xu, C. Wang, Z. Tian, Pyrolysis behavior of pectin under the conditions that simulate cigarette smoking, J. Anal. Appl. Pyrolysis 91 (2011) 232–240. [39] W. Wang, Y. Wang, L. Yang, B. Liu, M. Lan, W. Sun, Studies on thermal behavior of reconstituted tobacco sheet, Thermochim. Acta 437 (2005) 7–11. [40] T. Ahamad, M.A. Alshehri, TG-FTIR-MS (evolved gas analysis) of bidi tobacco powder during combustion and pyrolysis, J. Hazard. Mater. 199–200 (2012) 200–208. [41] V. Berbenni, A. Marini, G. Bruni, T. Zerlia, TG/FT-IR: an analysis of the conditions affecting the combined TG/spectral response, Thermochim. Acta 258 (1995) 125–133. [42] E. Post, S. Rahner, H. Moler, A. Rager, Study of recyclable polymer automobile under coatings containing PVC using TG/FTIR, Thermochim. Acta 263 (1995) 1–6. [43] V. Oja, M.R. Hajaligol, B.E. Waymack, The vaporization of semi-volatile compounds during tobacco pyrolysis, J. Anal. Appl. Pyrolysis 76 (2006) 117–123.

A. Gómez-Siurana et al. / Thermochimica Acta 573 (2013) 146–157 [44] Y.J. Sung, Y.B. Seo, Thermogravimetric study on stem biomass of Nicotiana tabacum, Thermochim. Acta 486 (2009) 1–4. [45] R.H. Venderbosch, W. Prins, Fast pyrolysis, in: R.C. Brown (Ed.), Thermochemical Processing of Biomass: Conversion into Fuels, Chemicals and Power, John Wiley & Sons, Chichester, 2011. [46] R.R. Baker, M.T.D.L. Davis, in: Nielsen (Ed.), Tobacco Production, Chemistry and Technology, Blackwell Science, Oxford, UK, 1999, pp. 308–349 (Chapter 12).

157

[47] J. Coates, Interpretation of infrared spectra, a practical approach, in: R.A. Meyers (Ed.), Encyclopedia of Analytical Chemistry, John Wiley & Sons Ltd., Chichester, 2000, pp. 10818–10837. [48] R.K. Sharma, J.B. Wooten, V.L. Baliga, P. Martoglio-Smith, M.R. Hajaligol, Characterization of char from the pyrolysis of tobacco, Agric. Food Chem. 50 (2002) 771–783.