In-depth investigation on quantitative characterization of pyrolysis oil by ³¹P NMR

Abstract

The characterization of different heteroatom functional groups by employing ³¹P NMR has been developed for almost 30 years. In this study, an in-depth investigation of this commonly used method has been accomplished for the analysis of pyrolysis oil. Several commonly used internal standards for ³¹P NMR have been examined by in situ monitoring. The results indicated that endo-N-hydroxy-5-norbornene-2,3-dicarboximide (NHND) is not stable after a long period of storage or experiment (>12 hours), but both cyclohexanol and triphenylphosphine oxide (TPPO) can be used as internal standards if a long experiment or storage is required. The pyrolysis oil has also been investigated by both short time (16 hours) in situ monitoring and long time (14 days) ex situ monitoring. The results showed that aliphatic OH, carboxylic acids and water contents are not very stable after 2 hours, and thus a short time of preparation, storage, and experiment need to be considered to ensure a precise quantitative measurement. The decomposition products are still unclear, but some preliminary investigations for different acids, (e.g. formic acid) have been accomplished. The results indicated that the aromatic carboxylic acids (benzoic acid and vanillic acid) are more stable than formic acid and acetic acid. Interestingly, the formic acid will even decompose to some other compounds at the very beginning of the in situ monitoring test. Further characterization found that water is one of the major products for the decomposition of formic acid in the ³¹P NMR solution. As far as we know, this is the first report on such time-dependent changes when using ³¹P NMR to analyze the pyrolysis oil, and these results show that proper application of this method is essential to achieve reliable quantitative data.

---

1. Introduction

Recently, fuels converted from biomass are increasingly being viewed as promising alternative energy resources due to the growing concerns about increasing global energy consumption and the effects of carbon dioxide emissions from fossil fuels.¹ Among the various conversion technologies that are being investigated, thermochemical decomposition processes (such as pyrolysis) have been reported² as one of the promising means to produce biofuel precursors – namely pyrolysis bio-oil. Due to the complicated composition of pyrolysis oil, subsequent upgrading is necessary,³ and requires quantitative and in-depth characterization of pyrolysis oil. Unfortunately, the complex nature of pyrolysis oil brings huge barriers for traditional analytical methods, such as GC-MS, GPC, HPLC and FT-IR. GC-MS analysis has frequently been used to identify individual components in pyrolysis oil, however, only a small portion of the pyrolysis oil (∼40%) can be detected by GC, mainly due to the poor volatility.⁴ HPLC would be useful for the identification and quantification of non-volatile compounds that can't be quantified by GC, but as with GC techniques, HPLC is currently only able to quantify a small portion of pyrolysis oil.⁵ GPC can be a useful technique for determining the molecular weight distributions of pyrolysis oils, but the distributions are not quantitative, and the data can only be used for comparing similar bio-oils.⁶ Spectroscopic techniques, such as FT-IR could provide insights on whole pyrolysis oil, but its application is limited due to both the dark color and complex composition of bio-oils.⁷ Therefore, quantitative and precise characterization of whole pyrolysis oil has significant impacts for providing insights on thermochemical decomposition pathways and on evaluating upgrading processes and catalysts.

The characterization of different heteroatom functional groups by employing ³¹P NMR has been developed for almost 30 years. One of the very first reports about analysis of coal pyrolysis condensates and extracts by ³¹P NMR was published on Energy & Fuels in 1988 by Verkade's group.⁸ Five different chlorophospholane based derivatizing reagents (1–5, Fig. 1) have been examined, and various model compounds, including different phenols, alcohols, aliphatic and aromatic carboxylic acids, amines, and thiols have been derivatized with different phosphorus (P) reagents and characterized by ³¹P NMR. The reactions between different functional groups and P reagents have been summarized in Fig. 2. Detailed assignment ranges for different functional groups have also been proposed in Verkade's research. The mixtures of such derivatives were also tested and the results were shown to be quite stable and reliable. The P reagents 1, 2, and 5 have been proposed to be the suitable reagents for complicated mixtures such as coal pyrolysis condensates and extracts, as they have less overlap between different functional groups. A new P reagent (6, Fig. 1) has been evaluated, and detailed investigations on the quantitative characterization of moisture and phenols in the coal condensate,⁹ coal extracts,¹⁰ and coal liquefaction residues¹¹ have been accomplished as well. The results showed that the developed ³¹P NMR methods can successfully perform the quantitative measurement for phenols and moisture in complicated mixtures, such as coal liquid products.

Fig. 1 Phosphorus (P) reagents used in the literature.

Fig. 2 Reactions of the phosphorous reagent (TMDP) with various hydroxyl functional groups and the ³¹P NMR assignment of phosphitylated compounds.

Similar methods have been further explored by several different research groups for the characterization of biomass and biodiesel. For example, Argyropoulos' group published several review¹² and research papers on characterization of lignin^13,14 cellulose,^15,16 and poly(hydroxyalkanoate).¹⁷ P reagents 1 and 2 (Fig. 1) have been used and evaluated, and several alcohols, carboxylic acids, and substituted phenols have been phosphitylated¹⁸ by these two P reagents. The assignment ranges have been proposed for different phenols, which is very important for further assignments for the characterization of lignin. The results also indicated that ³¹P NMR provides quantitative information on the contents of different hydroxyl groups in biomass components and in the polymer. Dais' groups investigated various lipid oils, moisture and model compounds by the similar ³¹P NMR methods.^19–22 P reagent 2 (Fig. 1) was used in Dais' research, and cyclohexanol was used as internal standard. The quantitative results can be provided by employing ³¹P NMR for different fatty acids, glycerol, a mono- and diglyceride composition, and water in the lipid oils. Crestini's group^23,24 also employed the same P reagent (P reagent 2 in Fig. 1) and a similar method to explore the chemical structure for tannin. Several model compounds which can represent the hydrolysable tannins, condensed tannins, and catechin tannins have been examined and assignment ranges for tannins have been proposed. The conclusion also indicated that the ³¹P NMR method is a reliable way to provide quantitative information on the tannin structure. Ragauskas' group developed similar methods as well. The applications for ³¹P NMR characterization on lignin,^25–30 biodiesel,^31,32 and pyrolysis oils^33–39 have been well established and the results also indicated that this method can be used for quantitative investigations of various hydroxyl groups in different materials, including very complicated biomass components, and their thermochemical decomposition products. Various internal standards have been evaluated for the characterization of lignin.²⁵ There were also several other applications of ³¹P NMR methods, for example, characterization of liquefaction products from loblolly pine,⁴⁰ analysis of free hydroxylic and carboxylic groups in unsaturated polyester and alkyd resins,⁴¹ chemical derivatization for metabolite profiling of lipophilic compounds in human serum,⁴² investigation of hydrothermal upgrading products from algae,⁴³ characterization of catalytic cracking products from fast pyrolysis oils with vacuum gas oil (VGO),⁴⁴ and the study of hydrothermal upgrading of pyrolysis oil from Norwegian spruce.⁴⁵

The primary goal for this study is to further evaluate this commonly used method and provide valuable information on the quantitative measurement of various functional groups in pyrolysis bio-oil. Several commonly used internal standards for ³¹P NMR have been examined by in situ monitoring. In situ monitoring with both pyrolysis oil and several model compounds has shown that some functional groups have time-dependent changes when employing ³¹P NMR. As far as we know, this is the first report on such phenomena when using ³¹P NMR to analyze pyrolysis oil, and has implications for obtaining reliable quantitative data with this method.

2. Materials and methods

2.1 Characterization of pyrolysis oil by NMR

All liquid NMR spectral data reported in this study were recorded with a Bruker Avance/DMX 400 MHz NMR spectrometer. All the NMR data processing and plots were carried out using MestReNova v7.1.0 software's default processing template (zero-filling is 32 K) and automatic phase and baseline correction.

2.1.1 Quantitative ¹H-NMR. Quantitative ¹H-NMR was acquired with 64 transients and 10 s pulse delay by employing a standard Bruker pulse sequence “zg” at room temperature.

2.1.2 Qualitative ¹³C-NMR. Qualitative ¹³C-NMR was acquired using a standard Bruker pulse sequence (zgig), a pulse delay of 5 s and 1000 scans at room temperature with a line-broadening (LB) of 5.0 Hz.

2.1.3 Quantitative ³¹P-NMR. Quantitative ³¹P NMR was acquired after in situ derivatization of the samples (standard mixtures and pyrolysis oil) with 2-chloro-4,4,5,5-tetramethyl-1,3,2-dioxaphospholane (TMDP, 100.0 μl) in a solution of (1.6 : 1 v/v) pyridine/CDCl₃, chromium acetylacetonate (relaxation agent), and endo-N-hydroxy-5-norbornene-2,3-dicarboximide (NHND), cyclohexanol, or triphenylphosphine oxide (TPPO) as internal standards, which are used to evaluate the hydroxyl group contents of pyrolysis oil and standard samples. ³¹P NMR spectra data was acquired using an inverse gated decoupling pulse sequence, 90° pulse angle, 25 s pulse delay, and 128 scans at room temperature with a LB of 4.0 Hz. The standard error is ∼±1.2%. For the in situ monitoring experiments, the sample was scanned for 16 hours by using the same method. The bio-oil tested was produced from oak using the NREL thermochemical process development unit (TCPDU) operated in entrained flow mode at 500 °C without hot gas filtration.⁴⁶

3. Results and discussion

3.1 Comparison of three commonly used internal standards for ³¹P NMR

NHND (endo-N-hydroxy-5-norbornene-2,3-dicarboximide) and cyclohexanol are the most commonly used internal standards for ³¹P NMR analysis of pyrolysis oils. The stability of these internal standards is extremely important for quantification; however, there has been no report on the stability of these internal standards in the literature. To further evaluate these standards, the comparison of three internal standards for ³¹P NMR has been accomplished by in situ monitoring. The same ³¹P NMR sample, which contains NHND, cyclohexanol and TPPO has been monitored in the NMR spectrometer for 16 hours with spectra taken once per hour (Fig. S1†). Since the amount of P reagent – TMDP (100.0 μl) is a known value, the total amount of P atoms in the NMR sample is known, and allows for the stability of the internal standards to be evaluated.

The integration results show that the content for NHND has been decreased by ∼4% after 16 hours (Table 1) and by ∼19% after 14 days (Table 1), which indicates that NHND is not very stable for a long period of time, but for short experiments (e.g., 1–5 hours or even an overnight experiment) NHND is still stable enough for use as the internal standard. The content for cyclohexanol is relatively stable (Table 1) during the whole monitoring process (both in situ for 16 hours and ex situ for 14 days), but the differences between each data point is slightly higher than TPPO (Table 1). The reason for such phenomena is still unclear. It's possible that there was an equilibrium for the reaction between cyclohexanol and TMDP. On the basis of results in Table 1, the TPPO has been identified as the most stable and reliable internal standard for the ³¹P NMR characterization, since the differences between each data point for TPPO is less than 1%, and the content for TPPO is also very stable during the entire monitoring test. The results for the stability test indicated that all of these three commonly used internal standards are stable and reliable enough for a short period of storage and experiment. For the overnight reaction or long term storage, TPPO will be the best internal standard. Given this, TPPO was used as internal standard for all subsequent experiments in this study.

Table 1 Stability for NHND, cyclohexanol, and TPPO during 16 hours of in situ monitoring and 14 days of ex situ monitoring

Time (h)	Stability^a (mol%)
	NHND	Cyclohexanol	TPPO
a The stability is calculated as: (1 − (calculated amount of internal standard)/(actual amount of internal standard)) × 100%.b The calculated amount of internal standard is calculated as: (peak area integration for internal standard)/(peak area integration for total amount of P atoms (from −40 to 240 ppm excluded peak for TPPO)) × (P atoms in 100.0 μl of 2-chloro-4,4,5,5-tetramethyl-1,3,2-dioxaphospholane).
1	0%^b	0%	0%
2	−1%	−1%	0%
3	−1%	0%	1%
4	−1%	−1%	0%
5	−1%	−1%	0%
6	−2%	−2%	1%
7	−3%	−2%	0%
8	−3%	−2%	0%
9	−2%	−2%	−1%
10	−3%	−2%	1%
11	−2%	−1%	0%
12	−4%	−2%	1%
13	−3%	−1%	0%
14	−3%	−2%	0%
15	−3%	−1%	−1%
16	−4%	0%	1%
14 days	−19%	−1%	0%

---

3.2 Stability test for pyrolysis oil in ³¹P NMR solution

³¹P NMR has been developed for several years to provide quantitative results for biomass components and their thermochemical deconstruction products (e.g., pyrolysis bio-oil). The chemical shifts and integration regions for pyrolysis oil after derivatization with TMDP by a quantitative ³¹P NMR measurement are shown in Table S1.† In situ monitoring for a pyrolysis oil sample has been examined using ³¹P NMR and results are shown in Fig. 3 and Table 2. The time-dependent changes can be observed from the spectrum in aliphatic and acid areas. The integration results for different hydroxyl groups and water wt% for pyrolysis oil during 16 hours of in situ monitoring and 14 days of ex situ stability test have been shown in Table 2. The results indicate that the content for aliphatic OH groups has been decreased by ∼7% after 16 hours and by ∼56% after 14 days (Table 2). The content for carboxylic acid groups has been decreased by ∼46% after 16 hours and by ∼90% after 14 days (Table 2). The content for water has been decreased by ∼18% after 16 hours and by ∼60% after 14 days (Table 2). The contents for different aromatic OH groups were relatively stable during in situ monitoring, but decreased up to ∼25–66% after 14 days of storage. Since there was no precipitate during both in situ and ex situ stability tests, it's possible that the decreased contents of hydroxyl groups and water could be due to the formation of some other functional groups which have different peak ranges. There are some increasing peaks that have been shown in Fig. S2,† which may represent possible decomposition products. The contents for all the functional groups are very stable for the first two hours, but the acid groups begin to decrease significantly after the first two hours during the in situ monitoring test. Therefore, the recommended preparation and test duration should be less than 2 hours to ensure the accuracy of the quantification using this ³¹P NMR method. However, since the acid groups have been reported to be responsible for aging, corrosion, and viscosity problems⁴⁷ for pyrolysis oils, a precise quantitative characterization of acid groups is very important for evaluating the properties of pyrolysis oil. To further investigate the stability for different carboxyl functional groups, several model compounds have been used to represent different carboxylic acids in pyrolysis oils, namely formic acid, acetic acid, benzoic acid and vanillic acid. It is worth noting here that this ³¹P NMR method has recently been validated for the analysis of pyrolysis bio-oil through an inter-laboratory study.⁴⁸ Five laboratories participated, and inter-laboratory variabilities of less than 10% were achieved for aliphatic and phenolic OH groups in bio-oil. Carboxylic OH groups had a slightly higher variability, near 15%. These results are very promising, and this is the first time an NMR technique has been successfully validated by round robin for bio-oil.

Fig. 3 In situ monitoring for the stability of pyrolysis oil during ³¹P NMR with TPPO as internal standard.

Table 2 Hydroxyl group contents and water wt% for the pyrolysis oil during 16 hours of in situ and 14 days of ex situ stability tests with TPPO as internal standard

Time (h)	Aliphatic OH mmol g^−1	β-5 OH mmol g^−1	4-O-5 OH mmol g^−1	5-5 OH mmol g^−1	Guaiacyl OH mmol g^−1
1	5.4	0.8	0.5	0.4	0.6
2	5.4	0.8	0.5	0.4	0.6
3	5.4	0.8	0.5	0.4	0.6
4	5.4	0.8	0.5	0.5	0.6
5	5.3	0.8	0.5	0.4	0.6
6	5.2	0.8	0.5	0.4	0.6
7	5.2	0.7	0.5	0.4	0.6
8	5.2	0.7	0.5	0.4	0.6
9	5.1	0.7	0.5	0.4	0.6
10	5.2	0.7	0.5	0.5	0.6
11	5.1	0.7	0.5	0.4	0.6
12	5.1	0.7	0.5	0.4	0.6
13	5.1	0.7	0.5	0.4	0.6
14	5.0	0.7	0.5	0.4	0.6
15	5.0	0.7	0.5	0.4	0.6
16	5.0	0.7	0.5	0.4	0.6
14 d	2.4	0.4	0.3	0.3	0.3

---

Time (h)	Catechol OH mmol g^−1	p-Hydroxy-phenyl OH mmol g^−1	Carboxylic OH mmol g^−1	H2O wt%
1	0.3	0.1	1.3	26.8
2	0.3	0.1	1.3	26.8
3	0.3	0.1	1.2	26.8
4	0.3	0.1	1.2	26.2
5	0.3	0.1	1.0	25.9
6	0.3	0.1	1.0	25.1
7	0.3	0.1	1.0	24.8
8	0.3	0.1	0.9	24.8
9	0.3	0.1	0.9	24.0
10	0.3	0.1	0.8	23.9
11	0.3	0.1	0.8	23.1
12	0.3	0.1	0.7	22.7
13	0.3	0.1	0.7	22.6
14	0.3	0.1	0.7	22.1
15	0.3	0.1	0.7	22.1
16	0.3	0.1	0.7	22.0
14 d	0.1	0.1	0.1	11.0

---

3.3 Stability test for different acids in ³¹P NMR solution

The in situ and ex situ monitoring processes for formic acid, acetic acid, benzoic acid and vanillic acid have been accomplished and the results are shown in Table 3. The benzoic acid and vanillic acid share a very close chemical shift in the ³¹P NMR (Fig. 4, ∼136 ppm), and the integration for these two aromatic acids was counted as one group in Table 3. The results indicated that during the in situ monitoring process, aromatic acids are relatively stable compared to formic acid and acetic acid, and that acetic acid is very stable during the first 4 hours. Interestingly, the formic acid content decreased dramatically during the very first hour of in situ monitoring, and the content decreased by >90% in only 5 hours. These results can also be seen in the spectra presented in Fig. 4. Since there was not any precipitate formed during the monitoring processes, the possible reason for such a dramatic change may be caused by the formation of new compounds. There were several new peaks formed during the in situ monitoring, located at ∼147 ppm, ∼33 ppm and ∼31 ppm. In addition, the TMDP was also continually consumed during in situ monitoring, which indicated that further reaction of the products from these acids with TMDP also consumes TMDP. The total P atom content remained the same for both short term and long term monitoring, which also indicated that there was no precipitate formed, nor any other reactions which could cause any changes in the concentration of total P atoms in the ³¹P NMR sample. Since there was no prior work documenting an instability of the product between formic acid and TMDP, some preliminary investigations have been performed for formic acid in the ³¹P NMR solution here, including in situ monitoring of the reaction between formic acid and TMDP by both ¹H and ¹³C NMR, and GC-MS analysis of the final solution. To avoid possible oxidation reactions, the monitoring process was conducted in a high pressure NMR tube sealed under He gas.

Table 3 Hydroxyl group contents for the formic acid, acetic acid, benzoic acid and vanillic acid during 16 hours of in situ and 14 days of ex situ stability tests with TPPO as internal standard

Time (h)	TMDP^a	Formic acid^b	Benzoic acid and vanillic acid^b	Acetic acid^b	Total P atoms^c (%)
a The percentage calculated as: (detected amount of the TMDP)/(calculated amount for the left TMDP, which is assuming 100% reaction yield with known amount of acids mixture).b The percentage calculated as: (detected amount of the carboxylic acids)/(feeding amount of the carboxylic acids).c The percentage calculated as: (detected amount of all the P atoms)/(feeding amount of the all the P atoms, which is calculated by the known amount of TMDP).
1	88%	74%	100%	100%	100%
2	87%	54%	100%	100%	100%
3	86%	39%	100%	100%	100%
4	84%	26%	100%	100%	100%
5	83%	15%	100%	99%	100%
6	82%	8%	100%	98%	100%
7	82%	4%	100%	96%	100%
8	80%	3%	99%	88%	100%
9	80%	3%	99%	85%	100%
10	80%	3%	97%	78%	100%
11	80%	3%	97%	71%	100%
12	79%	3%	96%	66%	100%
13	79%	3%	96%	62%	100%
14	79%	3%	96%	57%	100%
15	79%	3%	95%	53%	100%
16	77%	3%	93%	49%	100%
14 d	74%	2%	82%	3%	100%

---

Fig. 4 In situ monitoring for the stability of acids mixture in ³¹P NMR with TPPO as internal standard.

The in situ monitoring ¹H NMR results (Fig. S3†) indicate that the proton in the HC=O bond in the product (compound 1 in Fig. 5) from formic acid and TMDP decreased significantly and there was only very limited amount of this peak left after 12 hours, which is consistent with the ³¹P NMR result (Fig. 6). The in situ ¹³C NMR (Fig. S4†) also showed that the product (compound 1 in Fig. 5) from formic acid and TMDP is not stable and can completely decompose. In contrast, there was a growing peak ∼6.4 ppm in the ¹H NMR, which may represent the possible decomposition product. There is a similar growing peak in the in situ ¹³C NMR (Fig. S5†), which is ∼83.5 ppm. The ³¹P NMR also indicated that after the in situ monitoring the formic acid peak was no longer present, and there was a new peak present in the aliphatic range. One of the possible decomposition pathways for formic acid in the ³¹P NMR solution has been proposed in Fig. 5. GC-MS also showed that there was a peak which has the molecular ion-radical peak at 341, which is close to the proposed decomposition product (compound 2 in Fig. 5). In addition, ³¹P NMR also indicated that water (both peaks at 132 ppm and 16 ppm, compound 3 and 4 in Fig. 5, which are the products from water and TMDP) is one of the major decomposition products. Since the contents for the P contained decomposition products are much lower than the original P content for compound 1 in Fig. 5 (product from derivatization of formic acid with TMDP), some decomposition products may not contain any OH group, or may be gaseous. The tentative decomposition pathway from formic acid to formaldehyde in the ³¹P NMR solution has also been proposed (Fig. 5).

Fig. 5 Tentative decomposition pathways for formic acid in ³¹P NMR.

Fig. 6 Comparison before (top) and after (bottom) 16 hours of in situ monitoring of formic acid in ³¹P NMR with TPPO as internal standard.

4. Conclusions

In this study, the further investigation of ³¹P NMR, a commonly used method which can provide quantitative results on the different OH groups in various materials or complicated mixtures, has been accomplished. Several commonly used internal standards for ³¹P NMR have been examined by in situ monitoring. The results indicated that NHND is not very stable after a long period of storage (>12 hours), and cyclohexanol and TPPO can be used as internal standards if a long-term experiment or extended sample storage is required. Pyrolysis oil has also been investigated by both short time (16 hours) in situ monitoring and long time (14 days) ex situ monitoring. The results showed that aliphatic OH, carboxylic acid, and water contents are not very stable after 2 hours, and thus a short time of preparation, storage, and experiment needs to be considered to ensure a precise quantitative measurement. All the aromatic OH groups are relatively stable during the short time in situ monitoring, but the contents for these functional groups were still decreased by up to 66% after 14 days. The decomposition products are still unclear, but some preliminary investigations for different acids, especially the formic acid, have been accomplished. The results indicated that, compared to formic acid and acetic acid, the aromatic carboxylic acids (benzoic acid and vanillic acid) are more stable. Interestingly, formic acid was found to decompose starting with the very first hour of in situ monitoring. Further characterization found that water is one of the major products for the decomposition of formic acid in the ³¹P NMR solution. As far as we know, this is the first report effort on such phenomena when using ³¹P NMR to analyze pyrolysis oil, and these results are extremely important for obtaining reliable quantitative data with this method.

5. Acknowledgements

This work was supported by the U.S. Department of Energy under Contract No. DE-AC36-08GO28308 with the National Renewable Energy Laboratory. Funding provided by U.S. DOE Office of Energy Efficiency and Renewable Energy Bioenergy Technologies Office. The U.S. Government retains and the publisher, by accepting the article for publication, acknowledges that the U.S. Government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this work, or allow others to do so, for U.S. Government purposes.

References

1. A. J. Ragauskas, C. K. Williams, B. H. Davison, G. Britovsek, J. Cairney, C. A. Eckert, W. J. Frederick Jr, J. P. Hallett, D. J. Leak, C. L. Liotta, J. R. Mielenz, R. Murphy, R. Templer and T. Tschaplinski, Science, 2006, 311, 484–489.
2. R. P. Anex, A. Aden, F. K. Kazi, J. Fortman, R. M. Swanson, M. M. Wright, J. A. Satrio, R. C. Brown, D. E. Daugaard, A. Platon, G. Kothandaraman, D. D. Hsu and A. Dutta, Fuel, 2010, 89(Suppl. 1), S29–S35.
3. D. A. Ruddy, J. A. Schaidle, J. R. Ferrell Iii, J. Wang, L. Moens and J. E. Hensley, Green Chem., 2014, 16, 454–490.
4. M. Garcia-Perez, A. Chaala, H. Pakdel, D. Kretschmer and C. Roy, Biomass Bioenergy, 2007, 31, 222–242.
5. D. Mohan, C. U. Pittman and P. H. Steele, Energy Fuels, 2006, 20, 848–889.
6. D. C. Elliott, A. Oasmaa, F. Preto, D. Meier and A. V. Bridgwater, Energy Fuels, 2012, 26, 3769–3776.
7. W. Mu, H. Ben, A. Ragauskas and Y. Deng, BioEnergy Res., 2013, 6, 1183–1204.
8. A. E. Wroblewski, C. Lensink, R. Markuszewski and J. G. Verkade, Energy Fuels, 1988, 2, 765–774.
9. A. E. Wroblewski, C. Lensink and J. G. Verkade, Energy Fuels, 1991, 5, 491–496.
10. A. E. Wroblewski, K. Reinartz and J. G. Verkade, Energy Fuels, 1991, 5, 786–791.
11. T. Mohan and J. G. Verkade, Energy Fuels, 1993, 7, 222–226.
12. D. S. Argyropoulos, Res. Chem. Intermed., 1995, 21, 373–395.
13. A. Granata and D. S. Argyropoulos, J. Agric. Food Chem., 1995, 43, 1538–1544.
14. A. W. King, L. Zoia, I. Filpponen, A. Olszewska, H. Xie, I. Kilpelainen and D. S. Argyropoulos, J. Agric. Food Chem., 2009, 57, 8236–8243.
15. A. W. T. King, J. Jalomaki, M. Granstrom, D. S. Argyropoulos, S. Heikkinen and I. Kilpelainen, Anal. Methods, 2010, 2, 1499–1505.
16. I. Filpponen and D. S. Argyropoulos, Ind. Eng. Chem. Res., 2008, 47, 8906–8910.
17. A. Spyros, D. S. Argyropoulos and R. H. Marchessault, Macromolecules, 1997, 30, 327–329.
18. Z.-H. Jiang, D. S. Argyropoulos and A. Granata, Magn. Reson. Chem., 1995, 33, 375–382.
19. E. Hatzakis and P. Dais, J. Agric. Food Chem., 2008, 56, 1866–1872.
20. E. Hatzakis, A. Agiomyrgianaki and P. Dais, J. Am. Oil Chem. Soc., 2009, 87, 29–34.
21. P. Dais and A. Spyros, Magn. Reson. Chem., 2007, 45, 367–377.
22. P. Dais, A. Spyros, S. Christophoridou, E. Hatzakis, G. Fragaki, A. Agiomyrgianaki, E. Salivaras, G. Siragakis, D. Daskalaki, M. Tasioula-Margari and M. Brenes, J. Agric. Food Chem., 2007, 55, 577–584.
23. F. Melone, R. Saladino, H. Lange and C. Crestini, J. Agric. Food Chem., 2013, 61, 9307–9315.
24. F. Melone, R. Saladino, H. Lange and C. Crestini, J. Agric. Food Chem., 2013, 61, 9316–9324.
25. M. Zawadzki and A. Ragauskas, Holzforschung, 2001, 55, 283–285.
26. P. M. Froass, A. J. Ragauskas and J. e. Jiang, Ind. Eng. Chem. Res., 1998, 37, 3388–3394.
27. M. Lund and A. J. Ragauskas, Appl. Microbiol. Biotechnol., 2001, 55, 699–703.
28. M. Nagy, M. Kosa, H. Theliander and A. J. Ragauskas, Green Chem., 2010, 12, 31.
29. Y. Pu, S. Cao and A. J. Ragauskas, Energy Environ. Sci., 2011, 4, 3154–3166.
30. A. Tolbert, H. Akinosho, R. Khunsupat, A. K. Naskar and A. J. Ragauskas, Biofuels, Bioprod. Biorefin., 2014, 8, 836–856.
31. M. Nagy, B. J. Kerr, C. J. Ziemer and A. J. Ragauskas, Fuel, 2009, 88, 1793–1797.
32. M. Nagy, M. Foston and A. J. Ragauskas, J. Phys. Chem. A, 2010, 114, 3883–3887.
33. H. Ben and A. J. Ragauskas, Energy Fuels, 2011, 25, 2322–2332.
34. H. Ben and A. J. Ragauskas, Energy Fuels, 2011, 25, 4662–4668.
35. M. Kosa, H. Ben, H. Theliander and A. J. Ragauskas, Green Chem., 2011, 13, 3196–3202.
36. H. Ben and A. J. Ragauskas, RSC Adv., 2012, 2, 12892–12898.
37. H. Ben and A. J. Ragauskas, ACS Sustainable Chem. Eng., 2013, 1, 316–324.
38. H. Ben and A. J. Ragauskas, Bioresour. Technol., 2013, 147, 577–584.
39. S. Pan, Y. Pu, M. Foston and A. J. Ragauskas, BioEnergy Res., 2013, 6, 24–34.
40. Y. Celikbag, B. K. Via, S. Adhikari and Y. Wu, Bioresour. Technol., 2014, 169, 808–811.
41. A. Spyros, J. Appl. Polym. Sci., 2002, 83, 1635–1642.
42. M. A. DeSilva, N. Shanaiah, G. A. Nagana Gowda, K. Rosa-Perez, B. A. Hanson and D. Raftery, Magn. Reson. Chem., 2009, 47(Suppl. 1), S74–S80.
43. B. Patel and K. Hellgardt, Environ. Prog. Sustainable Energy, 2013, 32, 1002–1012.
44. D. V. Naik, V. Kumar, B. Prasad, M. K. Poddar, B. Behera, R. Bal, O. P. Khatri, D. K. Adhikari and M. O. Garg, RSC Adv., 2015, 5, 398–409.
45. C. J. Richard, B. Patel, D. Chadwick and K. Hellgardt, Biomass Bioenergy, 2013, 56, 446–455.
46. R. M. Baldwin and C. J. Feik, Energy Fuels, 2013, 27, 3224–3238.
47. H. Ben and A. J. Ragauskas, ChemSusChem, 2012, 5, 1687–1693.
48. J. R. Ferrell III, M. V. Olarte, E. D. Christensen, A. B. Padmaperuma, R. M. Connatser, F. Stankovikj, D. Meier and V. Paasikallio, Standardization of Analytical Techniques for Pyrolysis Bio-oil: History, Challenges, and Current Status of Methods, Biofuels, Bioprod. Biorefin., 2015.

---

Footnote

† Electronic supplementary information (ESI) available: In situ monitoring data and GC-MS data. See DOI: 10.1039/c5ra23939g

---