Thermal oxidation reaction process and oxidation kinetics of abietic acid

Abstract

The thermal oxidation reaction process and oxidation kinetics of abietic acid were investigated by using self-designed gas–solid reaction equipment. The oxidation product and intermediates of the oxidation reaction of abietic acid were tracked by LC-MS. The results revealed a two-step oxidation reaction of abietic acid: abietic acid forms a peroxide first, followed by cracking to form hydroxyl-containing oxidized abietic acid. Both of the steps followed the pseudo-first order reaction kinetics, in which the kinetic equation of the first step is r₁ = c_A × 3.51 × 10³ × exp(−58.96 × 10³/RT), with activation energy of 58.96 kJ mol⁻¹. The kinetic equation of the second step is r₂ = c_O × 6.09 × 10⁵ × exp(−48.06 × 10³/RT) with activation energy of 48.06 kJ mol⁻¹. The kinetic equation of the total reaction is r_a = c_a × 1.12 × 10⁶ × exp(49.51 × 10³/RT) with apparent activation energy of 49.51 kJ mol⁻¹.

---

Introduction

Rosin is a renewable resource obtained by the distillation of the exudates of pines trees. It is composed mostly of abietic acid and other rosin acids, which play important roles in the aspects of everyday life, e.g., in painting,¹ adhesives² and biochemical synthesis.³ China has the world's biggest pine forest cover, and it is necessary for us to take full advantage of this resource and create economic benefits. However, due to the active structure of conjugated double bonds, resin acids and their products, such as colophony, are unstable against air, heat, and light, as well as are sensitive to mineral acids. These oxidation behavior leads to its color fastness and limit the immediate application of rosin, and these are the main reason for its reduced cost. In order to improve the antioxidant abilities and weatherability of rosin products, a series of studies have been performed, which focuses on the oxidation mechanisms, oxidation products and oxidation kinetics.

In view of the oxidation products and oxidation process, Harris⁴ assumed that the auto-oxidation of abietic acid has two paths. One starts from the addition of O₂ to the C₁₃ C₁₄ double bond. The formed peroxide cleaves into two hydroxyl fragments. The second is a substitution reaction between the reactive methylenic C₁₂ and O₂, which forms a peroxide. After cleavage, a stable hydroxyl fragment is formed. Enoki⁵ studied several oxidation products of resin acids in α-pinene solutions. Prinz⁶ identified 6 known oxidation products of abietic acid and methyl ester under different storage conditions. He also pointed out that the sensitive positions of the reactants are the C₇ and C₁₃ atoms.

Several investigations have been done regarding the kinetics of resin acids, but there are few reports on the kinetics of their thermal oxidation reactions. Ritchie⁷ reported the isomerization kinetics of L-pimaric acid and neoabietic acid in absolute ethyl alcohol, when catalyzed by a strong acid. Lawrence⁸ studied the kinetics of thermal catalyzed isomerization of abietic acid, L-pimaric acid, neoabietic acid and palustric acid at 150–200 °C. Pastorova⁹ proposed that the isomerization of abietic acid type resin acid forms a stable dehydroabietic acid. In addition, four oxidation paths were discussed. Ladero^10,11 investigated the kinetics of the esterification of rosin and polyols. Rongxiu Qin^12,13 performed a kinetic study on the room temperature oxidation of abietic acid, colophony and rosin and pointed out that the oxidation of abietic acid was a pseudo-first order reaction with an active energy of 50.29 kJ mol⁻¹. Jialing Liu¹⁴ explored the kinetics of UV-induced oxidation of colophony and obtained kinetic data from 365 nm irradiation. When applying UV spectroscopy to a kinetic study, iteration is required to eliminate the influence of the product. The main problem is that an UV absorptive intermediate will disturb iteration. It must be mentioned that the oxidation of abietic acid contains multiple steps, which are not identifiable by UV spectroscopy, and the researchers investigated the global kinetics of the oxidation reaction as a preliminary state. Therefore, it is necessary to obtain more precise and detailed data for the reaction kinetics in order to seek a more suitable method for anti-oxidation.

The aim of this study is to continue the investigations regarding the thermal oxidation reaction process and oxidation kinetics of abietic acid. More precisely, the article presents a novel micro solid-state reactor to conduct the oxidation reaction of abietic acid using a polyethylene film fixed by two aluminum sheets. In this way, abietic acid was formed as a membrane on the PE film by making experimental conditions close to the actual situation. HPLC was applied to study the oxidation kinetics of abietic acid. In addition, the intermediate and product of oxidation were tracked and detected by LC-MS. The results provide support to the knowledge of rosin and its products to solve the problem of easy oxidization during their storage and manufacture, which lead to their color fastness and economic losses.

Experimental section

Reagents and instruments

Rosin, produced in Guangxi, China; abietic acid, self-synthesized⁴ (98.1%); dehydrated ethanol; diamylamine; hydrochloric acid; and acetic acid (AR Grade). GC-MS, Shimadzu GC-MS/QP5050A. HPLC, Waters 2487UV/VIS. LC-MS, DionexUltiMate 3000/Thermo Scientific TSQ Quantum Access MAX.

Design of polyethylene film reactor

Polyethylene film was fixed between two aluminum sheets, conducting the micro oxidation reaction of abietic acid. The reactor is shown in Scheme 1.

Scheme 1 The polyethylene film reactor.

Oxidation of abietic acid

First, 0.1 g of abietic acid was dissolved in 10 mL of dehydrated ethanol. Then, 10 μL of as-prepared solution was added to PE micro reactor drop by drop,¹³ in which the film area was 3.0 × 1.7 cm². After the ethanol was fully vaporized in vacuum under room temperature, the reactor was placed in an incubator to perform oxidation, and the kinetic runs were performed between 303 and 333 K. Each sample was taken approximately every 30 min and HPLC was applied for the quantitative analysis of both abietic acid and its oxidation products by external standard method.

HPLC conditions

Mobile phase: methanol. Flow rate: 1 mL min⁻¹. Column temperature: 40 °C. Wavelength of UV detector: 241 nm/220 nm. Chromatograph column: Hypersil ODS2-C18, 5 μm, 250 nm × 4.6 mm. Injection amount: 10 μL. External standard method was applied.

LC-MS conditions

HPLC analysis was performed using the Dionex UltiMate 3000 with a binary pump, an on-line degasser, an auto-sampler and a column temperature controller. Chromatographic separations were performed on a Hypersil Gold C18 Column (10 mm × 2.1 mm, 5 μm) at 40 °C. The mobile phase consisted of methanol–water (95 : 5, v/v). The flow rate was set at 0.2 mL min⁻¹. Aliquots of 2 μL were injected into the HPLC system for analysis.

MS analysis was carried out on a Thermo Scientific TSQ Quantum Access MAX triple stage quadrupole mass spectrometer with an electrospray ionization (ESI) source running in negative-ionization mode. The typical ion source parameters were: spray voltage: 3500 V. Sheath gas pressure (N₂): 5 units. Ion transfer tube temperature: 350 °C. Collision gas (Ar): 1.5 mTorr. Q1/Q3 peak resolution: 0.7 Da. Scan width: 0.002 Da. The scan dwell time was set at 0.1 s for every channel. All data collected in centroid mode were acquired and processed using Xcalibur 2.2 software (Thermo Fisher Scientific Inc., USA).

Kinetics model I(consecutive reaction)

The chemical reaction rate (r) is known to be a function of temperature (T) and reactant concentration (c), and the differential equation is,where n is the reaction order.

As the ref. 4–6 and 15 reflected: abietic acid (A) first formed peroxide (O), followed by cracking, which formed hydroxyl-containing oxidized abietic acid (P). Thus, the reaction can be simplified to a consecutive reaction.¹⁶

Thus, the first step and second step of abietic acid oxidation can be kinetically expressed as formula (2) and (3), respectively.

r₁ = k₁c_A (2)

r₂ = −r_P = −k₂c_P (3)

where c_A and c_P represent the concentration of abietic acid and oxidation product of abietic acid in the reaction system, respectively, and k₁ and k₂ represent the rate constant of first step and second step, respectively.

Kinetic equations were derived by detecting the curve of concentration variation of abietic acid and its oxidation product vs. time, as shown in Fig. 2. Under the experimental conditions, the intermediate of the reaction cannot be quantitatively analyzed, due to the lack of absorption shown in the UV spectrum. Moreover, some other oxidant other than peroxide may be produced during the first step of oxidation; therefore, the kinetics cannot be calculated by the variation of abietic acid either. Thus, the kinetic equation of the first step was computed by theoretical calculation (formula (6)). More details will be shown below.

Kinetics model II (parallel reaction)

Prinz⁶ reported several oxidation products of abietic acid and its methyl ester under room temperature storage, indicating that peroxide and other oxidants were able to coexist during the oxidation process, as they have different reaction sites. Therefore, a possible mechanism based on a parallel reaction model was proposed, as shown below:

Thus, the kinetic equations can be presented as follows:

r₃ = k₃c_A = −k₃c_O (4)

r₄ = k₄c_A = −k₄c_P (5)

where k₃ and k₄ represent the rate constant of peroxide and other oxidant, respectively.

Similarly, the kinetic of the oxidation products are based on experimental data, while the other parallel path regarded as k₃ can be fitted by theoretical calculations.

Results and discussion

Wavelength of HPLC detection

Abietic acid shows significant absorption in ultraviolet spectroscopy because of the conjugated group. Thus, a UV spectroscope is a suitable detector for the HPLC. Moreover, it is necessary to select an appropriate detection wavenumber to analyze abietic acid along with its oxidant. The result is shown in Fig. 1, in which curve (a) is pure abietic acid, curve (b) is abietic acid after 3 hours of oxidation, curve (c) is abietic acid after 1 hour of oxidation and curve (d) is blank PE film. Curve (a) reveals the maxima absorbance of abietic acid is located at 241 nm. Curve (b) suggests that the oxide of abietic acid has no evident UV absorbance peak. Considering that both abietic acid and its oxide have similar absorbance at 220 nm, we chose this wavelength to detect abietic oxide. The UV spectrum of the oxidation intermediate is plotted as curve (c). It suggests that in range of 210–400 nm, the absorbance is almost 0. On the other hand, 200–210 nm is overlapped by solvent absorbance. Thus, it is difficult to set a wavelength to monitor the intermediate. In conclusion, we use 241 nm to detect abietic acid and 220 nm to detect the oxidation product.

Fig. 1 UV spectra of the abietic acid oxidation process. (a) Abietic acid. (b) Oxide of abietic acid. (c) Intermediate of abietic acid. (d) Blank film.

Herein, we use 180 min of oxidation at 40 °C as an example. Every 30 min, the sample was loaded into the HPLC and scanned at 241 nm (the rest of the liquid phase conditions are the described in the experimental section), as shown in Fig. 2. For clarity, only three samples scanned at 220 nm by HPLC for oxidation after 0 min, 90 min and 180 min were chosen for comparison. When the detecting wavelength is set as 241 nm, only abietic acid shows an absorbance peak (retention time is 4.2 min). When it is 220 nm, both abietic acid and the oxidation product can be detected (retention time is 3.3 min). During oxidation process, the concentration of the product increases, whereas the concentration of the product, i.e., abietic acid decreases. Thus, 241 nm can provide the total change of abietic acid in the system, as well as the apparent kinetics of the reaction. Thus, an absorbance of 220 nm can be used to study the kinetics of the formation of product.

Fig. 2 HPLC spectra under 241 nm/220 nm of the abietic acid oxidation process.

Apparent kinetics of oxidation of abietic acid

The logarithm of concentration (ln c_A) was plotted against time (t), in which the concentration of abietic acid was calculated from the working curves, y = 1.988 × 10⁵x + 5.914 × 10⁴.

The linear response of ln c_A to time suggests that the oxidation of abietic acid is a pseudo-first order reaction. The rate constant at different temperatures (k_a) were calculated.

The neperian of the rate constant of the reaction at different temperatures, in which abietic acid is oxidized, ln k_a, has a linear relationship with 1/T (T represents temperature), and it is expressed as ln k_a = −5954T⁻¹ + 13.93, R² = 0.99. The calculated apparent activation energy, E_a, is 49.51 kJ mol⁻¹, according to Arrhenius equation, which has an error range (<5%) with previous reports.¹²

Kinetics of the formation of oxidation product

When reacted at the as-stated temperatures, we used LC data collected at 220 nm to plot ln c_P–t, according to the working curves of the oxidation product y = 5.097 × 10⁴x + 5.723 × 10³. Since ln c_P has a linear relationship with t, the formation of oxidation product is a first order reaction. We acquired k₂/k₄ at different temperatures, and plotted them against T. The relationship between ln k₂ and 1/T is linear, which can be described as ln k₂ = −5780T⁻¹ + 13.32, R² = 0.99. The calculated activation energy E₂ = 48.06 kJ mol⁻¹, coefficient A₂ = exp(13.32) = 6.09 × 10⁵ min⁻¹, and rate equation is r_P = r₂ = r₄ = c_p × 6.09 × 10⁵ × exp(−48.06 × 10³/RT).

LC-MS analysis of oxidation of abietic acid

In order to explore the chemical structure information of major products in the abietic acid oxidation system, we analyzed the sample that oxidized at 60 °C for 2 hours by LC-MS.¹⁷ Negative-ionization mode was selected to run Q1 scanning on the oxidation product. The range of m/z was set as 310–400 m/z. Due to the complicated composition of the oxidation product of abietic acid, we chose three peaks that have the largest responses in the MS spectra to plot TIC. MS scanning focused on the main ion (m/z = 317) (a), fragment ion (m/z = 299), main ion (m/z = 333) (c), fragment ion (m/z = 301), main ion (m/z = 349) (b), and fragment ion (m/z = 331) to track the target ion and exclude the effects of impurities, as shown in Fig. 3. Inspired by previous studies,¹⁸ m/z = 333 can be considered a major product, since it has a much higher intensity then m/z = 317 and m/z = 349. The small peaks around m/z = 317 can be considered as the result of isomerization.

Fig. 3 Typical TIC of oxidation system of abietic acid. (a) Abietic acid. (b) Abietic acid peroxide. (c) Abietic acid oxidate.

Negative-ionization mode was selected to conduct on m/z = 317, m/z = 333 and m/z = 349. According to its molecular mass, m/z = 317 can be confirmed as abietic acid. Referring to the ion cracking mechanism involved and the ref. 19 and 20, a difference of 16 between m/z = 333 and 317 suggests a hydroxyl-containing abietic acid oxide, and a difference of 32 between m/z = 349 and m/z = 317 suggests a peroxide of abietic acid.^4–6 Thus, we assume that during the oxidation, abietic acid oxide and abietic acid peroxide are both formed at the same time. It is possible that the peroxide exists in the system as an intermediate. Comparing the TIC with the LC spectra at 220 nm, we confirmed that the 3.3 min peak in the LC spectrum is hydroxyl-containing abietic acid oxide and the peroxide has no apparent absorbance.

As discussed above, the oxidation of abietic acid contains two steps: the formation of unstable peroxide and the oxidation pyrolysis of it, which forms a hydroxyl-containing oxide.

Theoretical computation of the formation of intermediate

Based on the kinetic data of the formation of abietic acid oxide, the kinetics of formation of the intermediate can to be calculated. By combining theoretical and experimental data using kinetic model I and II, respectively, we validated our assumption on the mechanism of oxidation.

According to formula (2) and (3), Model I can be kinetically represented as:

r_A = −r₁ = −k₁c_A (6)

r_O = r₁ − r₂ = k₁c_A − k₂c_O (7)

r_P = r₂ = k₂c_O (8)

For continuous oxidation, initial parameters will be: c_A = c_A0, c_O0 = c_P0 = 0. Thus, at ideal conditions, formula (6)–(8) can be expressed by integration as follows,

c₁ = c_A = c_A0e^−k₁t (9)

in which, and c_O represents the concentration of peroxide.

Considering the parallel reaction, if the reaction is first order with respect to reactant in every path, the , where k_a represents the apparent kinetic rate constant. Thus, based on formula (4) and (5), model II can be expressed as:

c₁ = c_A = c_A0e^−k_at (12)

in which,

Fig. 4 presents the fitted curves at each temperature for the oxidation product based on model I and model II, respectively. It is evident that model I fitted the experimental data marginally better than model II. The relative residuals with reaction time for all compounds and models tested are presented in Fig. 5 at 318 K, for instance. The relative width of error bar based on model II is wider than that of model I, indicating that the error of model I is small. The goodness-of-fit parameters for both models are summarized in Table 1, showing that the Akaike information criterion (AIC) of these two models are close at 303 and 313 K, whereas the AIC of model I are much smaller than that of model II at a relatively high temperature (upon 318 K). It can be inferred from this data that model I fitted the reaction better than model II, suggesting that the oxidation process of abietic acid tends to be a consecutive reaction. The F value of model I is much higher than that of model II, indicating that model I is the best fit for the number of kinetic constants used in the model. More precisely, the coefficient of determination, R², for the rate constant of peroxide formation based on model I (k₁) and model II (k₃) ranged from 0.966–0.991 and 0.496–0.955, as shown in Table 1, indicating that model II (the parallel reaction) was inappropriate in the oxidation process of abietic acid, especially under a relative high temperature.

Fig. 4 Experimental and predicted results of concentration of oxidant vs. time profiles based on kinetic model I (A) and model II (B) at different temperatures. (a) 303 K. (b) 313 K. (c) 318 K. (d) 323 K. (e) 328 K. (f) 333 K.

Fig. 5 Relative residual concentrations of the three compounds (abietic acid, peroxide, and oxide) that represent the goodness-of-fit of model I (A) and model II (B) at 318 K.

Table 1 Goodness-of-fit parameters for model I and model II^a

T/K	Model I					Model II
	RSS	RMSE	F	AIC	R	RSS	RMSE	F	AIC	R
a Note: RSS (sum of squared residuals), RMSE (root-mean-square-deviation), F (Fisher F value), AIC (Akaike parameter).
303	1.06 × 10^−16	7.59 × 10^−18	207.46	−296.32	0.990	6.00 × 10^−19	4.29 × 10^−20	371.93	−300.32	0.955
313	6.38 × 10^−18	4.56 × 10^−19	320.84	−283.78	0.987	1.01 × 10^−17	7.21 × 10^−19	200.76	−280.57	0.920
318	2.73 × 10^−18	1.95 × 10^−19	1675.68	−289.72	0.991	5.76 × 10^−17	4.11 × 10^−18	73.69	−268.38	0.812
323	6.48 × 10^−18	4.63 × 10^−19	1617.07	−283.67	0.990	1.41 × 10^−16	1.01 × 10^−17	68.32	−262.08	0.801
328	1.19 × 10^−16	8.48 × 10^−18	621.93	−263.31	0.979	2.49 × 10^−15	1.78 × 10^−16	23.96	−242.01	0.580
333	3.96 × 10^−16	2.83 × 10^−17	348.64	−254.88	0.966	5.80 × 10^−15	4.14 × 10^−16	18.17	−236.08	0.496

---

For clarity, the fitted curves of abietic acid, peroxide and oxidant vs. time based on kinetic model I and model II at 303, 318 and 333 K are shown in Fig. 6. It is evident that under the temperature of 333 K for model I, as shown in Fig. 6(E), more amount of oxidant was produced compared with the theoretical model data. This is due to the side reaction, i.e., the hydroxyl substitution reaction at C₇ and C₁₃ caused by the temperature rising,⁶ which leads to the unsatisfactory result of using the kinetic model of a two-step continuous reaction. Moreover, the rate constant of the formation of peroxide is available by the fitting calculation, as listed in Table 2, along with k_a and k₂. Further fitting shows that the activation energy of the reaction, E₁, is 58.96 kJ mol⁻¹. The coefficient, A₁, is exp(5.86) = 3.51 × 10³ min⁻¹. Thus, the rate equation of the formation of peroxide is: r₁ = c_A × 3.51 × 10³ × exp(−58.96 × 10³/RT).

Fig. 6 Fitted curves of concentration of abietic acid ■, peroxide ▼ and oxidant ▲ vs. time based on kinetic model I and model II.

Table 2 Kinetic constant of abietic acid at different reaction temperatures

T/K	ka/min^−1	10^−6k1/min^−1	k2/min^−1
303	0.00339	0.0274	0.00305
313	0.00597	0.0481	0.00610
318	0.00811	0.0604	0.00783
323	0.01002	0.0788	0.00988
328	0.01557	0.1682	0.01473
333	0.01958	0.2180	0.01651

---

Since k₂≫k₁, peroxide formed in the reaction is transformed to epoxide immediately.²¹ Thus the first step, which forms peroxide, determines the total reaction rate. Comparing the activation energies of the two reactions,²² we found that E₁ > E₂, A₁ < A₂ and k₂ is always larger than k₁, which matches experimental results. In conclusion, the total reaction rate is determined by the first step, which is the formation of peroxide, regardless of temperature.

Conclusion

We performed trace amounts of oxidation of abietic acid in a PE membrane reactor and provided a LC analysis method. LC-MS was used to detect the intermediate and oxide. Based on what we get, we present the following conclusions:

(1) The thermal oxidation process of abietic acid was investigated by comparing the goodness of fitting of the two kinetic models to the chemical reaction, and the results reveal a consecutive reaction. In the first step, peroxide is formed, followed by further oxidation, which forms hydroxyl-containing abietic acid oxide.

(2) The kinetic equation of the first step is r₁ = c_A × 3.51 × 10³ × exp(−58.96 × 10³/RT), and activation energy is 58.96 kJ mol⁻¹. The kinetic equation of the second step is r₂ = c_O × 6.09 × 10⁵ × exp(−48.06 × 10³/RT), and activation energy is 48.06 kJ mol⁻¹. The kinetic equation of the total reaction is r_a = c_a × 1.12 × 10⁶ × exp(49.51 × 10³/RT), and apparent activation energy is 49.51 kJ mol⁻¹.

(3) E_a has an error range (<5%) with previous reported values and the experimental data has good fitting with model data in the second step. This indicates the reliability of the kinetic equation in this study.

Abbreviations

N	Reaction order
ra	Apparent chemical reaction rate
ca	Concentration of total abietic acid
T	Temperature
T	Time
ka	Apparent chemical reaction rate constant
Ea	Apparent active energy
E1	Active energy of first step
E2	Active energy of second step
r1	Chemical reaction rate (first step of consecutive reaction)
r2	Chemical reaction rate (second step of consecutive reaction)
r3	chemical reaction rate (the step of producing peroxide in parallel reaction)
r4	Chemical reaction rate (the step of producing oxide in parallel reaction)
k1	Chemical reaction rate constant (first step of consecutive reaction)
k2	Chemical reaction rate constant (second step of consecutive reaction)
k3	Chemical reaction rate constant (the step of producing peroxide in parallel reaction)
k4	Chemical reaction rate constant (the step of producing oxide in parallel reaction)
cA	Concentration of abietic acid
cO	Concentration of peroxide
cP	Concentration of oxide

Acknowledgements

We are grateful to the National Natural Science Foundation of China (11462001), the Yulin City—Guangxi University Scientific and Technological Cooperation Projects (201210505) and the Young People Fund of Guangxi Science and Technology Department (2012GXNFBA053020) for financial support.

Notes and references

1. R. Ploeger, E. René de la Rie, C. W. McGlinchey, M. Palmer, C. A. Maines and O. Chiantore, Polym. Degrad. Stab., 2014, 107.
2. E. Alvarez-Manzaneda, R. Chahboun, J. Guardia, M. Lachkar, A. Dahdouh, A. Lara and I. Messouri, Tetrahedron Lett., 2006, 47, 2577–2580.
3. B. Gigante, C. Santos, A. Silva, M. Curto, M. Nascimento, E. Pinto, M. Pedro, F. Cerqueira, M. Pinto and M. Duarte, Bioorg. Med. Chem., 2003, 11, 1631–1638.
4. G. C. Harris and T. F. Sanderson, J. Am. Chem. Soc., 1948, 70, 334–339.
5. A. Enoki, Wood Research: Bulletin of the Wood Research Institute Kyoto University, 1976, vol. 59, pp. 49–57.
6. S. Prinz, U. Müllner, J. Heilmann, K. Winkelmann, O. Sticher, E. Haslinger and A. Hüfner, J. Nat. Prod., 2002, 65, 1530–1534.
7. P. F. Ritchie and L. F. McBurney, J. Am. Chem. Soc., 1950, 72, 1197–1200.
8. R. Moore and R. V. Lawrence, J. Am. Chem. Soc., 1958, 80, 1438–1440.
9. I. Pastorova, K. Van der Berg, J. Boon and J. Verhoeven, J. Anal. Appl. Pyrolysis, 1997, 43, 41–57.
10. M. Ladero, M. de Gracia, F. Trujillo and F. Garcia-Ochoa, Chem. Eng. J., 2012, 197, 387–397.
11. M. Ladero, M. de Gracia, J. J. Tamayo, I. L. d. Ahumada, F. Trujillo and F. Garcia-Ochoa, Chem. Eng. J., 2011, 169, 319–328.
12. R.-X. Qin, P.-X. Huang, X.-M. Liu, L. Ma and Y.-L. Wu, Chem. J. Chin. Univ., 2009, 5, 024.
13. X.-M. Liu, R.-X. Qin, P.-X. Huang, J.-L. Liu, L. Ma and W.-G. Li, Acta Phys.-Chim. Sin., 2010, 26, 2115–2120.
14. J.-L. Liu, X.-M. Liu, W.-G. Li, L. Ma and F. Shen, Monatsh. Chem, 2014, 145, 209–212.
15. F. Ren, Y.-F. Zheng, X.-M. Liu, X.-Y. Yue, L. Ma, W.-G. Li, F. Lai, J.-L. Liu and W.-L. Guan, J. Mol. Struct., 2014, 307–313.
16. M. Möller and H. U. Moritz, J. Appl. Polym. Sci., 2006, 101, 4090–4097.
17. J. Cheng, X. Chen, H. Lu, Q. Chen and Y. Zhang, RSC Adv., 2014, 4, 43378–43386.
18. E. J. Hsieh, M. S. Bereman, S. Durand, G. A. Valaskovic and M. J. MacCoss, J. Am. Soc. Mass Spectrom., 2013, 24, 148–153.
19. A. Wohlfarth, K. B. Scheidweiler, X. Chen, H.-f. Liu and M. A. Huestis, Anal. Chem., 2013, 85, 3730–3738.
20. R. Liao, H. Wu, H. Deng, Y. Yu, M. Hu, H. Zhai, P. Yang, S. Zhou and W. Yi, Anal. Chem., 2013, 85, 2253–2259.
21. G. Zhu, H. Li, Y. Cao, H. Liu, X. Li, J. Chen and Q. Tang, Ind. Eng. Chem. Res., 2013, 52, 4450–4454.
22. J. Poudel and S. C. Oh, Ind. Eng. Chem. Res., 2012, 51, 4509–4514.

---