Fast screening for hydrolysable and condensed tannins in lignocellulosic biomass
using reactive Py-GC/MS with in situ silylation
Marco Mattonai, Erika Ribechini
Department of Chemistry and Industrial Chemistry, Via G. Moruzzi 13, 56124 Pisa (Italy) Corresponding author: Erika Ribechini. e-mail: [email protected], tel: +390502219312
ABSTRACT
Rapid screening techniques for lignocellulosic biomass are required for the development of efficient conversion strategies in the field of biofuels and value-added chemicals. This is particularly true for tannins, which are highly valuable in the tannery and nutraceutical industries. In the present work, we propose a quick method based on reactive pyrolysis with in situ hexamethyldisilazane (HMDS) derivatization followed by GC/MS analysis for the qualitative determination of tannins in lignocellulosic biomass. Different pyrolysis times were used to study pyrolysis mechanisms of tannins. Reference compounds belonging to hydrolysable and condensed tannins, both in monomeric and polymeric forms, were used to select specific pyrolysis products that could be used as markers. When pyrolysis time was increased, hydrolysable tannins were found to be highly thermostable, while condensed tannins showed extensive degradation. An optimal pyrolysis time was determined to reduce the number of peaks and increase the abundances of the most characteristic components. The results were applied to assess the presence of tannins in four different biomass: oak gall nuts, hazelnut cuticles, grape seeds and pomegranate bark.
KEYWORDS
Reactive pyrolysis; Py-GC/MS; Derivatisation; Condensed tannins; Hydrolysable tannins; Biomass
1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1
INTRODUCTION
Tannins are minor components of biomass, and they can be classified into two categories: hydrolysable tannins and condensed tannins [1]. Hydrolysable tannins can also be grouped into gallotannins and ellagitannins. Tannins are considered to be high value-added chemicals thanks to their nutraceutical properties and their ability to convert animal skin into leather [2-4]. The presence of these molecules in a lignocellulosic biomass should be taken into great consideration when developing strategies for the exploitation of biomass. Therefore, rapid and efficient screening techniques are required to evaluate the composition of a biomass so that no potential for high-value chemical production is wasted.
Analytical pyrolysis coupled with gas chromatography and mass spectrometry (Py-GC/MS) is a fundamental technique for analysing a lignocellulosic biomass [5-9]. Its main advantages are small sample quantities, no sample pretreatment, and high reproducibility. Analytical pyrolysis is widely used to study the conversion of biomass into biofuels on a laboratory scale [10,11].
Most of the pyrolysis products of lignocellulose have highly polar functional groups such as hydroxyl and carboxyl. Molecules with these groups tend to tail badly on common GC stationary phases and generate asymmetric and unresolved peaks. To solve this problem, mobile-hydrogen groups can be converted into less polar ones by derivatization reactions such as methylation or silylation. The choice of derivatizing agent is crucial, as both derivatization strategies have disadvantages. Methylation is usually performed with tetramethylammonium hydroxide (TMAH), which produces a strong basic environment and can generate uninformative products [12,13]. Silylation, which can be performed with various derivatizing agents, is a slower reaction than methylation, and can lead to the formation of partially silylated compounds [14]. Reactive pyrolysis with silylation is a particularly innovative and still unexplored technique, which allows analytical pyrolysis to be performed at relatively long times, by keeping the sample in a sealed glass capsule inside the pyrolysis furnace. Once the desired pyrolysis time has elapsed, the glass capsule is crushed by a metal rod, and the pyrolysis products are sent to the GC/MS system [15]. Reactive pyrolysis with a silylating agent has been reported to improve chromatographic quality and prevent partial derivatization in the analysis of carbohydrates [16]., thanks to the This is made possible by means of a prolonged exposure of the sample to high temperature and pressure.
Analytical pyrolysis has been explored for the identification of tannins in natural products [5,17-20]. Tannins are often derivatized by thermally-assisted methylation with TMAH. This strategy could however be problematic when dealing with whole lignocellulosic materials, as some derivatized tannins could be mistaken for lignin pyrolysis products [21]. Moreover, thermally-assisted methylation promotes the hydrolysis of the ester bonds that can be found in hydrolysable tannins, increasing the number of pyrolysis
2 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 3
products [22]. Only one work in the literature appears to deal with the silylation of hydrolysable tannins, but this was performed as an ex situ sample preparation step prior to GC/MS analysis [23].
In the present work, reactive pyrolysis-GC/MS with insitu silylation was evaluated as a screening technique for detecting monomeric and polymeric tannins in biomass. Gallotannin, ellagitannins and condensed tannins model compounds were pyrolyzed in the presence of hexamethyldisilazane (HMDS) at various times to study their pyrolysis mechanisms and to overcome partial silylation. An optimal reaction time was established, and it was used to study four different biomass, which are known for their tannins content. To the best of our knowledge, this is the first work describing the behaviour of tannins in reactive pyrolysis with in situ silylation.
MATERIALS AND METHODS
Materials and samples: Gallic acid (≥ 99%, Extrasynthese, France), tannic acid (ACS, Sigma-Aldrich, USA),
ellagic acid (97%, Lancaster Synthesis, USA), punicalagin (≥ 99%, A+B mixture, Phytolab, Germany), (+)-catechin (≥ 96%, Sigma-Aldrich, USA) and procyanidin C1 (≥ 99%, Phytolab, Germany) were used as model compounds for gallotannins, ellagitannins and condensed tannins. Hexamethyldisilazane (HMDS, Sigma-Aldrich, USA) was used as derivatizing agent in all experiments. Quartz wool (Frontier Lab, Japan) was used for the analysis of reference compounds. Oak gall nuts were purchased from Kremer Pigmente (Germany). Hazelnut cuticles were provided by a hazelnut processing factory in Viterbo, Italy. Grape seeds were provided by a local seller of enological products. Pomegranate bark was collected by the authors from a tree growing in Montopoli, Pisa, Italy (43°40'26.1"N, 10°44'23.8"E).
Py-GC/MS apparatus: All analyses were performed with an EGA/Py-3030D microfurnace pyrolyser (Frontier
Lab, Japan) coupled to a 6890 gas chromatograph and a 5973 mass spectrometric detector (Agilent Technologies, USA). A PY1-1050 micro-reaction sampler was used as sample holder in all experiments. This sample holder has been described in previous publications [15,16]. The maximum pyrolysis temperature allowed by the micro-reaction sampler system is 400 °C. Higher temperatures usually generate excessive pressure in the glass capsule, leading to premature shattering.
Sample preparation: When gallic acid, tannic acid, ellagic acid and catechin were analysed, phase saturation
and peak asymmetry were observed. To solve this problem, 2% (w/w) blends of each reference compound with quartz wool were prepared. Each blend was grinded to a fine powder using a Pulverisette 23 ball mill (Fritsch, Germany) for 1 min at 50 Hz. Quartz wool was also pyrolysed alone to obtain a background pyrogram. Punicalagin and procyanidin C1 were analysed without any pretreatment, since the pyrograms of their quartz wool blends showed very small and noisy peaks. Due to the greater sample quantity, a higher
3 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 5
split ratio was used for the analysis of these two compounds (see below). Raw materials were grinded with the ball mill and analyzed without further preparation. When the quartz wool blends were analysed, the sample amount was 200 μg. When punicalagin, procyanidin and raw samples were analysed, the sample amount was 100 μg. In all cases, 3 µL of HMDS were added. Sample amounts were chosen on the basis of previous work with reactive pyrolysis made in our research group [24,25]. While a large excess of derivatizing agent can prevent the formation of partially derivatized compounds, volumes greater than 3 μL are not recommended, as they would generate excessive pressure in the glass capsule during pyrolysis. Before analysis, the capsule was put under a gentle stream of nitrogen to ensure inert atmosphere, and finally it was flame-sealed.
Experimental parameters: Pyrolysis furnace temperature was 400 °C and interface temperature was 280 °C
in all experiments. Pyrolysis times of 0.2, 0.5, 1, 2, 5 and 10 min were investigated for all samples. The GC injector was operated in split mode at a temperature of 280 °C. When punicalagin and procyanidin C1 were analysed, a split ratio of 50:1 was used, while a split ratio of 10:1 was used in all other cases. Chromatographic separation was achieved with an HP-5MS UI column (30 m x 0.25 mm, film thickness 0.25 µm, Agilent Technologies, USA) coupled to a deactivated silica precolumn (2m x 0.32 mm, Agilent Technologies, USA) and helium as carrier gas (1 mL min-1). The following temperature program was used for
the GC oven: 60 °C isothermal for 1 min; 20 °C min-1 up to 200 °C, then isothermal for 1 min; 5 °C min-1 up to
280 °C, then isothermal for 20 min. The mass spectrometer was operated in EI positive ion mode (70 eV, m/z range 50-800). The transfer line was kept at 300 °C, the ion source at 230 °C and the quadrupole analyzer at 150 °C.
Data interpretation: Pyrograms were analyzed with the Automated Mass spectra Deconvolution and
Identification Software (AMDIS, version 2.71). Pyrolysis products were identified by comparison with literature references [23], mass spectral libraries (Wiley and NIST/EPA/NIH) and by interpretation of the mass spectra. Semi-quantitative calculations were performed by integrating the chromatographic peaks and expressing the areas as percentages. Reproducibility of peak areas was evaluated by performing the analysis of reference compounds at the same pyrolysis time for three times. Coefficients of variation (CVs) were calculated by dividing the standard deviation of peak areas by the corresponding average, and were found to be lower than 10%.
RESULTS AND DISCUSSION
Gallotannins: Figure 1 shows the pyrograms of the six reference tannins after 1 minute of pyrolysis. The
chromatogram of gallic acid showed a single peak that could be attributed to the per-silylated form of the molecule (compound #7). The mass spectrum of per-silylated gallic acid is reported in Figure 2. The base
4 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 7
peak was at m/z 73, which corresponds to the Si(CH3)3+ ion and is typical of every silylated compound. The
signals at m/z 458, 443 and 281 can be attributed to M•+, [M – CH3]+ and [M – Si(CH3)4 – ·OSi(CH3)3]+,
respectively [26]. The structures for these ions are given in Figure 3. A shoulder peak at a lower retention time was observed at short pyrolysis times. This peak was attributed to the partially derivatized form of gallic acid. As shown in Figure 4, this peak rapidly disappeared when pyrolysis time was increased. No additional peak was detected at long pyrolysis times, proving that gallic acid is thermally stable under these experimental conditions.
Peaks that could not be attributed to pyrolysis products of the sample are labeled with an asterisk in Figure 1. The peak at 7.9 min can be attributed to a derivative of HMDS. This derivative is probably obtained by the reaction of HMDS with itself. This hypothesis is supported by the lower intensity in the pyrogram of punicalagin, and by its absence in the pyrogram of procyanidin C1, since in both these analyses the ratio of HMDS to reference compound was considerably lower. The peaks at 13.5 and 16.2 min belong to the TMS esters of palmitic and stearic acids, respectively.
Tannic acid was used as model compound for gallotannins. The chromatogram of tannic acid showed an intense peak of gallic acid, and two additional small peaks at 28.2 and 30.2 min. These peaks showed identical mass spectra, reported in Figure 2, and were attributed to galloyl-gallate isomers (compounds #12 and #13), consisting of two gallic acid molecules linked through an ester bond. Structures of the main fragment ions of galloyl-gallate are given in Figure 3.
The ester bond of galloyl-gallate can involve either the meta- or para- hydroxyl group, explaining the presence of two isomers. However, assigning a particular galloyl-gallate isomer to a particular peak is not straightforward. The para-isomer does not exist in nature, as the biosynthesis of tannic acid occurs with the selective formation of ester bonds on the meta- position [27,28]. Its presence in the chromatograms could be due to isomerization reactions taking place in the pyrolysis environment. The isomerization of galloyl-gallate has already been reported in the literature [29]. When pyrolysis time was increased, the relative abundances of the two peaks remained constant, suggesting that the isomerization is a quick process. If this is the case, then the relative abundances of the two isomers depend on their relative stability. The meta-isomer should be more stable than the para-meta-isomer, due to the lower steric hindrance, and therefore the larger of the two peaks could be attributed to the meta-isomer.
At short pyrolysis times, the shoulder peak belonging to partially derivatized gallic acid was also found in the pyrograms of tannic acid as well. When pyrolysis time was increased, the shoulder peak disappeared as expected, but the peaks of galloyl-gallate isomers remained unaltered. This shows that galloyl-gallate is as thermally stable as gallic acid. The thermal stability of galloyl-gallate is in apparent contradiction with the presence of gallic acid, since it could derive from the cleavage of the ester bonds. However, commercial
5 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 9
tannic acid extracts contain a mixture of polygalloyl-glucose species, with a number of gallic acid units per glucose ranging from 5 to 12 [29]. This means that some of the gallic acid units are not bonded to other gallic acid units. The thermal stability of galloyl-gallate is an important result, as the two isomers can be used as markers for gallotannins in a lignocellulosic matrix.
Ellagitannins: Ellagic acid is the monomer of ellagitannins. The pyrogram of ellagic acid showed a single
peak that was attributed to the fully derivatized form of the molecule (compound #14). The mass spectrum of ellagic acid, reported in Figure 2, showed signals at m/z 590, 575 and 487, corresponding to M•+, [M – CH3]+ and [M – CH3 – Si(CH3)4]+, respectively. The structures for these ions are given in Figure 3. No peak was
found that could be attributed to partially derivatized forms of ellagic acid at any pyrolysis time. When pyrolysis time was increased, no peaks belonging to fragmentation products were detected in the pyrograms, proving that ellagic acid is as thermally stable as gallic acid and galloyl-gallate. The results obtained for ellagic acid show that silylation is a milder derivatization technique than methylation. In fact, when thermally-assisted methylation is used, extensive hydrolysis of ellagic acid can be observed, and numerous pyrolysis products are obtained [30].
Punicalagin was chosen as a representative polymeric ellagitannin. Its pyrogram showed the peak of ellagic acid, and an additional peak at lower retention times. The mass spectrum of this peak showed a similar fragmentation pattern to that of ellagic acid, with signals at m/z 636, 621 and 533. Based on the structure of punicalagin and the m/z values in the mass spectrum, the peak was attributed to a decarboxylated derivative of ellagic acid (compound #11). The fragment ions of this species are shown in Figure 3. Without one of the carboxylic moieties, an additional hydroxyl group can be derivatized, explaining both the higher molar mass and the lower retention time. Since the formation of this species was not observed in the pyrolysis of monomeric ellagic acid, compound #11 was chosen as a marker of polymeric ellagitannins. No notable change was observed in the pyrograms of punicalagin when pyrolysis time was increased, suggesting that compound #11 is thermally stable.
Condensed tannins: The pyrogram of catechin showed a peak that could be attributed to the per-silylated
form of the molecule (compound #10). The mass spectrum of catechin is reported in Figure 2, and the structures for the main ions are reported in Figure 3. Four additional peaks were observed in the pyrograms at low retention times. Based on their mass spectra, these peaks were attributed to pyrolysis products of catechin (compound #3, #4, #5 and #6). All four pyrolysis products can be obtained from catechin by cleavage of the pyran ring, as shown in Figure 5. Some of these pyrolysis products were also found in Py-GC/MS analysis of catechin with in situ methylation [20].
No peak was found in the pyrograms that could be attributed to partially derivatized forms of catechin or its pyrolysis products. However, when pyrolysis time was increased, the relative total area of these four peaks
6 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 11
increased significantly, as shown from the results of semi-quantitative calculations reported in Figure 4. After 10 minutes of pyrolysis, the total abundance of the four pyrolysis products was comparable to that of catechin. This shows that, contrary to hydrolysable tannins, catechin is not thermally stable, and undergoes extensive degradation in these experimental conditions. Compounds #4, #5 and #6 share an alkyl-catechol skeleton, which is also shared by some characteristic products of lignin [21]. This means that they are not good markers for condensed tannins in biomass, and so the only good peak to be used as marker is the one of catechin.
Procyanidin C1 was used as model compound for polymeric condensed tannins. The pyrogram of procyanidin showed the same peaks of catechin, with two more intense peaks at low retention times. These peaks were attributed to dihydroxy-benzene and dihydroxy-toluene (compounds #1 and #2). Like compounds #4, #5 and #6, dihydroxybenzene is a bad marker for condensed tannins, since it can also be formed from the pyrolysis of carbohydrates [20]. The presence of additional pyrolysis products suggests that procyanidin C1 is even less thermostable than catechin. In addition to the new pyrolysis products at low retention times, the pyrograms of procyanidin showed two peaks at 23.4 and 24.7 min. The corresponding mass spectra, which are reported in Figure 2, showed the molecular ions at m/z 546 and 560, respectively. Based on the m/z values in the mass spectra, we attributed the chromatographic peak with molecular ion at m/z 560 to a dehydrated form of catechin (compound #9). The elimination of the hydroxyl group on the pyran ring is favored by the formation of a conjugated double bond. Moreover, the dehydration is likely to be induced by the cleavage of the C-C bond between the catechin monomers of procyanidin C1, as shown in Figure 5. The absence of the hydroxyl group on the pyran ring could explain the difference between the fragmentation patterns of this compound and per-silylated catechin. Structures for the main fragment ions are given in Figure 3. Following the same argument, we hypothesized that the peak at m/z 546 corresponds to a catechin molecule in which a CH2O molecule is lost by the pyran ring
(compound #8). The structure of this compound is reported in Figure 2 as molecular ion. The elimination of a -CH2- group generates a furan ring, in which the newly formed double bond is conjugated with both
aromatic moieties. The extended conjugation could explain the scarce fragmentation of the molecular ion in the ion source.
When pyrolysis time was increased, the total relative abundance of the pyrolysis products of catechin increased, while the intensity of the peaks of #8 and #9 decreased. This suggests that compounds #8 and #9 are thermolabile as catechin. Since the peaks of compounds #8 and #9 were not detected in the pyrogram of monomeric catechin, they were chosen as markers for polymeric condensed tannins.
Analysis of raw samples: The analysis of the model compounds at long pyrolysis times revealed that
hydrolysable tannins are highly thermostable, while condensed tannins undergo thermal degradation to produce various pyrolysis products. Free gallic acid molecules required pyrolysis times greater than 1 min to
7 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 13
achieve full derivatization. Both the presence of partially derivatized gallic acid and the thermal degradation of condensed tannins contribute to a decrease in the sensitivity of the technique, since the total amount of analyte in the sample is spread among a greater number of peaks. For this reason, an optimal pyrolysis time was established to minimize both partial derivatization and thermal degradation reactions, and 1 min was chosen as the best pyrolysis time.
Once optimization was achieved, the method was tested on four different biomass raw samples: oak gall nuts, hazelnut cuticles, grape seeds and pomegranate bark. Figure 6 shows the ion chromatograms extracted from the pyrograms of the four raw materials.
Gall nuts showed peaks of gallic acid, galloyl-gallate and ellagic acid. This was the only sample that showed signals of polymeric gallotannins. The results agree with HPLC/MS analyses performed on the same sample [31]. The cited reference also reports the presence of other phenolic acids, which were not detected in our experimental conditions. It is however worth noticing that galloyl-gallate isomers were not detected by the HPLC/MS analyses. This could be due to the high acidity of the mobile phase, which could have promoted hydrolysis of the ester bonds.
Hazelnut cuticle showed peaks of gallic acid and catechin. Although pyrolysis was performed for only 1 min, the peak of whole catechin was much lower than those of its pyrolysis products (compounds #5 and #6). There are two possible explanation for this. The first explanation is that the peaks of compounds #5 and #6 are higher due to the pyrolysis of lignin. The second possible explanation is that there are some inorganics species in the raw material. Inorganics are well-known catalysts that enhance the pyrolysis of natural organic materials [32,33].
Grape seeds proved to be the sample with greater tannins variety, showing peaks of gallic acid, ellagic acid and both monomeric and polymeric condensed tannins. The results agree with other studies in the literature [31,34], in which procyanidins with polymerization degrees up to 15 were detected. As for hazelnut cuticles, the peak of whole catechin was much smaller than the peaks of compounds #5 and #6. An additional peak of interest was detected between those of compounds #9 and #10. This peak showed the same mass spectrum of catechin, and was therefore assigned to epicatechin, which is an isomer of catechin. Finally, pomegranate bark showed peaks of gallic acid, monomeric and polymeric ellagitannins, and catechin. The peak of epicatechin was found in this sample as well. Literature results show that pomegranate peels contain a very wide variety of ellagitannins [35]. Distinction of this compounds is not feasible with the technique presented in this work, and is better performed with classical techniques such as HPLC/MS. 8 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 15
CONCLUSIONS
The present work described a rapid screening technique for tannins in whole biomass.
the pyrolytic behavior of both monomeric and polymeric tannins under reactive pyrolysis with in situ derivatization was described. Pyrolysis times in the range 0.2-10 min were observed. Hydrolysable tannins proved to be highly thermostable, while condensed tannins showed extensive degradation when pyrolysis time was increased. Moreover, short pyrolysis times caused partially derivatized compounds to be detected in the pyrograms of gallic and tannic acid. A reaction time of 1 min was chosen as optimal value to minimize both partial derivatization and thermal degradation of condensed tannins.
Hydrolysable tannins proved to be thermostable towards pyrolysis with in situ silylation, while condensed tannins showed extensive degradation. Partial derivatization was overcome at long pyrolysis times. All three polymeric referenceReference polymeric tannins generated pyrolysis products that distinguish them from the pyrolysates of their monomeric counterparts. This allowed us to establish a set of molecular markers that could be used to assess the presence of either free monomers or polymeric tannins in a biomass sample. The proper m/z signal for each marker was chosen based on the mass spectra.
Once optimized, the technique was tested on four raw materials. Extraction of the proper m/z signals allowed rapid detection of tannins to be obtained, providing results that agreed with the literature. This method is capable of detectingcan detect the presence of tannins in biomass, and to provides information on their class and polymerization. The main advantages of the method are the low amount of sample required (0.1 mg), the lack of sample pre-treatment steps, and the short analysis time (40 min). The main disadvantage, which is shared with most of the Py-GC/MS techniques, is that absolute quantitation and detailed speciation of the tannins cannot be achieved. If quantitative analysis is required, more accurate techniques such as HPLC/MS are recommended. In conclusion, we believe this analytical method provides useful information in the development of efficient conversion strategies for lignocellulosic biomass.
REFERENCES
[1] K. Khanbabaee and T. van Ree, Natural product reports, 18, (2001) 641.
[2] V. Muccilli, N. Cardullo, C. Spatafora, V. Cunsolo and C. Tringali, Food chemistry, 215, (2017) 50. [3] M.C. Figueroa-Espinoza, A. Zafimahova, P.G.M. Alvarado, E. Dubreucq and C. Poncet-Legrand, Food
chemistry, 178, (2015) 38.
[4] A. Russell, Chemical Reviews, 17, (1935) 155.
[5] M. Mattonai, D. Licursi, C. Antonetti, A.M. Raspolli Galletti and E. Ribechini, Journal of Analytical and Applied Pyrolysis, (2017).
[6] F. Chen, B. Hou, S. Chen, H. Zhang, P. Gong and A. Zhou, Renewable Energy, 114, (2017) 960.
[7] J.D. Murillo, E.A. Ware and J.J. Biernacki, Journal of Analytical and Applied Pyrolysis, 108, (2014) 234. 9 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 17
[8] A. Massaro, M.P. Colombini and E. Ribechini, Journal of Analytical and Applied Pyrolysis, 121, (2016) 205.
[9] J.C. Del Rio, Á.T. Martínez and A. Gutiérrez, Journal of Analytical and Applied Pyrolysis, 79, (2007) 33.
[10] S. Wang, G. Dai, H. Yang and Z. Luo, Progress in Energy and Combustion Science, 62, (2017) 33. [11] M.K. Akalın and S. Karagöz, TrAC Trends in Analytical Chemistry, 61, (2014) 11.
[12] J.M. Challinor, Journal of Analytical and Applied Pyrolysis, 61, (2001) 3.
[13] D. Fabbri and R. Helleur, Journal of Analytical and Applied Pyrolysis, 49, (1999) 277. [14] D. Fabbri and G. Chiavari, Analytica Chimica Acta, 449, (2001) 271.
[15] A. Hosaka, C. Watanabe, N. Teramae and H. Ohtani, Journal of Analytical and Applied Pyrolysis, 106, (2014) 160.
[16] M. Mattonai, D. Tamburini, M.P. Colombini and E. Ribechini, Analytical chemistry, 88, (2016) 9318. [17] Z. Xia, A. Singh, W. Kiratitanavit, R. Mosurkal, J. Kumar and R. Nagarajan, Thermochimica Acta, 605,
(2015) 77.
[18] J. Kaal, K.G. Nierop, P. Kraal and C.M. Preston, Organic Geochemistry, 47, (2012) 99.
[19] P.A. Case, C. Bizama, C. Segura, M.C. Wheeler, A. Berg and W.J. DeSisto, Journal of Analytical and Applied Pyrolysis, 107, (2014) 250.
[20] M. Ismayati, A. Nakagawa-izumi and H. Ohi, Journal of Wood Science, 63, (2017) 350.
[21] D. Tamburini, J.J. Łucejko, M. Zborowska, F. Modugno, E. Cantisani, M. Mamoňová and M.P. Colombini, International Biodeterioration & Biodegradation, 116, (2017) 243.
[22] C.A. Clausen, International Biodeterioration & Biodegradation, 37, (1995) 101. [23] T.-Y. Chu, C.-H. Chang, Y.-C. Liao and Y.-C. Chen, Talanta, 54, (2001) 1163.
[24] M. Mattonai, D. Tamburini, M.P. Colombini and E. Ribechini, Analytical Chemistry, 88, (2016) 9318. [25] M. Mattonai and E. Ribechini, Analytica chimica acta, 1017, (2018) 66.
[26] F.I. Schäfers and K. Herrmann, Journal of Chromatography A, 240, (1982) 387. [27] P. Arapitsas, Food chemistry, 135, (2012) 1708.
[28] R. Niemetz and G.G. Gross, Phytochemistry, 66, (2005) 2001.
[29] M. Verzele and P. Delahaye, Journal of Chromatography A, 268, (1983) 469.
[30] G.C. Galletti and P. Bocchini, Rapid communications in mass spectrometry, 9, (1995) 250. [31] A. Restivo, I. Degano, E. Ribechini and M.P. Colombini, PloS one, 9, (2014) e88762.
[32] P.R. Patwardhan, J.A. Satrio, R.C. Brown and B.H. Shanks, Bioresource technology, 101, (2010) 4646. [33] K. Wang, J. Zhang, B.H. Shanks and R.C. Brown, Applied Energy, 148, (2015) 115.
[34] C. Prieur, J. Rigaud, V. Cheynier and M. Moutounet, Phytochemistry, 36, (1994) 781. [35] S. Akhtar, T. Ismail, D. Fraternale and P. Sestili, Food Chemistry, 174, (2015) 417.
10 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 19
Table 1: Retention time of the identified species in the pyrograms of reference compounds, along with the main m/z signals in their mass spectra. The number of TMS groups for each species is reported in brackets.
# tR (min) Compound m/z 1 7.2 1,2-dihydroxybenzene (2 TMS) 254, 239, 166, 151, 136, 73 2 7.7 n-methyl-n’,n’’-dihydroxybenzene (2 TMS) 268, 253, 180, 73 3 9.4 1,3,5-trihydroxybenzene (3 TMS) 342, 327, 268, 253, 147, 133, 73 4 9.9 4-(hydroxymethyl)-1,2-dihydroxybenzenebenzene-1,2-diol (3 TMS) 356, 341, 253, 73 5 11.3 4-(hydroxyvinyl)-dihydroxybenzenebenzene-1,2-diol I (3 TMS) 368, 353, 249, 191, 73 6 11.7 4-(hydroxyvinyl)-dihydroxybenzenebenzene-1,2-diol II (3 TMS) 368, 353, 249, 191, 73 7 12.7 Gallic gallic acid (4 TMS) 458, 443, 281, 179, 147, 73 8 23.4 Dihydroxyphenyl-benzofurandiol2-(3,4-dihydroxyphenyl)benzofuran-4,6-diol (4 TMS) 546, 531, 458, 147, 73 9 24.7 Dihydroxyphenyl-benzopyrandiol 2-(3,4-dihydroxyphenyl)-benzo-2H-pyran-5,7-diol (4 TMS) 560, 545, 471, 307, 179, 73 10 25.6 Catechin catechin (5 TMS) 650, 368, 355, 280, 267, 179, 147, 73 11 26.7 Decarboxylated decarboxy-ellagic acid (5 TMS) 636, 621, 533, 73
12 28.2 m-galloylgallate (6 TMS) 754, 739, 369, 281, 179, 147, 73 13 30.2 p-galloylgallate (6 TMS) 754, 739, 369, 281, 179, 147, 73 14 33.9 Ellagic ellagic acid (4 TMS) 590, 575, 487, 73
11 323 324 325 326 21
Figure captions
Figure 1: Total ion pyrograms of the reference compounds at 1 min of pyrolysis. Peaks are numbered according to Table 1. Peaks labeled with an asterisk are interferents.
Figure 2: Mass spectra of the pyrolysis products obtained by Py-GC/MS of the reference compounds. Compounds are numbered according to Table 1. The mass spectrum of compounds #5 and #6 and of compounds #12 and #13 were identical.
Figure 3: Molecular ions for the pyrolysis products of hydrolysable and condensed tannins, and their ions obtained by fragmentation and rearrangement processes in the mass spectrometer. Compounds are numbered according to Table 1.
Figure 4: Relative peak areas as a function of pyrolysis time. Left: partially and fully-derivatised forms of gallic acid; right: catechin and its pyrolysis products.
Figure 5: Pyrolysis products of catechin and procyanidin. Compounds are numbered according to Table 1. Figure 6: Normalized extracted ion pyrograms obtained by Py-GC/MS of raw samples. Peaks are labeled according to Table 1. 12 327 328 329 330 331 332 333 334 335 336 337 338 339 340 23
