Effect of HBr formation on mercury oxidation via CaBr₂ addition to coal during combustion

Abstract

Adding CaBr₂ to coal to enhance elemental mercury (Hg⁰) oxidation during combustion has been an effective mercury control technology, but the added CaBr₂ may increase levels of noxious Br₂ or HBr gases in flue gas. Temperature-programmed decomposition (TPD) experiments were conducted to verify the effect of CaBr₂ addition on Hg⁰ oxidation. The results indicated that the amount of Hg⁰ released initially decreased with increasing amounts of CaBr₂ additive and then held steady. The optimal amount of additive was 200 μg g⁻¹. CaBr₂ addition effectively oxidized Hg⁰ released at relatively low temperatures only. The generation of HBr was confirmed by mass spectrometry. The formation of HBr occurred over a temperature range of 250 °C to 400 °C, and the HBr concentration first increased and then remained stable as levels of CaBr₂ additive were increased in coal. The maximum concentration of HBr was 18 ppm and corresponded to 200 μg g⁻¹ CaBr₂. Further analysis indicated a strong, negative linear correlation between the amount of Hg⁰ released and the HBr concentration in flue gas. Based on these findings and previous studies, the possible mechanism of oxidation of Hg⁰ by CaBr₂ was analyzed.

---

1 Introduction

Mercury and its compounds are important air contaminants due to their toxicity, persistence and bioaccumulation in the environment.¹ Coal combustion is a predominant source of mercury emission because coal is a primary source of energy in China. According to the National Bureau of Statistics of China, coal used for electricity generation amounted to 1785 Mt at the end of 2012, accounting for approximately 50.6% of the national total coal consumption. Mercury in coal-fired flue gas exists in three forms: elemental mercury (Hg⁰), oxidized mercury (Hg²⁺) and particulate-bound mercury (Hg^p).² Mercury speciation has an important effect on its control. According to recent studies, existing air pollution control devices (APCDs) have shown a degree of co-benefit mercury control.^3–6 Hg^p can be effectively captured by particulate matter control devices.^7,8 Hg²⁺ can be easily captured in a wet flue gas desulfurization (WFGD) system due to its water solubility.^9–11 Elemental mercury (Hg⁰), however, is difficult to remove by APCDs because of its high volatility and water insolubility.¹² Thus, enhancing the oxidation of Hg⁰ to Hg²⁺ allows for simultaneous removal by existing APCDs and is a promising method for mercury control.^4,13 However, the simultaneous removal efficiency of mercury is highly variable due to differing mercury speciation in flue gas. Mercury speciation is greatly influenced by halogen content in coal.¹⁴ Chlorine is the most abundant halogen element in coal. A higher chlorine content in coal leads to a higher mass fraction of Hg²⁺ in total gaseous mercury, resulting in a higher mercury removal efficiency. For low chlorine content coal, addition of a halogen salt is required to enhance Hg⁰ oxidation.¹⁵

Adding CaBr₂ to coal has been an effective method of oxidizing Hg⁰ during the combustion process in power plants and has been widely investigated in flue gas from actual coal-fired power plants. The results of a series of spot tests conducted by the U.S. Government Accountability Office (GAO) at 14 different power plants showed that addition of 25–300 μg g⁻¹ CaBr₂ to coal resulted in the oxidation of more than 90% of Hg⁰ in flue gas. Berry et al.¹⁶ conducted a site test at Plant Miller to determine the effectiveness of CaBr₂ on mercury oxidation. The results showed that the mercury oxidation efficiency increased from 60% to 90% with addition of 5 μg g⁻¹ bromine to coal. The results of a mercury control test at a 300 MW coal-fired power plant indicated that the mass fraction of Hg²⁺ increased significantly from 35% to 90% in total gaseous mercury when the bromine additive ratio was 20 μg g⁻¹.¹⁷ These studies demonstrate that CaBr₂ addition to coal is an effective way to enhance Hg⁰. However, there are some side effects and potential areas of concern with this method. It is widely known that mercury in coal exists in various occurrence modes: HgS, HgO, HgSO₄, silicate-bound Hg, etc. It is unclear if the effect of CaBr₂ addition on Hg⁰ release differs between occurrence modes. Furthermore, the addition of CaBr₂ may generate HBr or Br₂ in flue gas. Only few related field test reports were available due to the testing restrictions. It was observed that the flue gas bromine content increased with increasing calcium bromide injection by Berry et al.¹⁶ HBr or Br₂ are strong, corrosive acidic gases. Zhuang et al.¹⁸ studied the corrosion characteristics of iron in bromine-containing flue gas. The results showed that the iron was corroded by HBr both at high and low temperatures. Moreover, according to previous studies,^19,20 the halogen gas HCl inhibited the denitrification performance of SCR catalyst (V₂O₅–WO₃/TiO₂). A similar inhibition may also occur with HBr because chlorine and bromine are elements in the same family. In addition, bromine species seriously damage the atmospheric ozone layer.^21,22 Because of these side effects, it is necessary to determine if adding CaBr₂ to coal generates HBr or Br₂ in flue gas. Few relevant reports are available. Therefore, it is essential to study the generation of HBr and Br₂ resulting from CaBr₂ addition to coal and the relationship between their concentration and the amount of CaBr₂ additive.

In this study, an experimental instrument was designed to investigate the effect of CaBr₂ addition on Hg⁰ release from coal and the formation of HBr and Br₂ in flue gas. Temperature-programmed decomposition (TPD) was used to simulate the coal combustion process and a dynamic Hg⁰ release profile was obtained via a coupled, online Hg analyzer. Generation of HBr and Br₂ were monitored by mass spectrometry. The results indicated that the addition of CaBr₂ significantly decreased the amount of Hg⁰ released while increasing HBr concentration in flue gas. Remarkably, the effect of CaBr₂ addition showed a discrepancy in Hg⁰ release between different occurrence modes. Finally, the mechanism of mercury oxidation by CaBr₂ was analyzed. The results of this study can provide a reference for the application of this method of mercury control in power plants.

2 Experimental

2.1. Experiment setup

A bench-scale system consisting of a fixed bed reactor, a temperature control device, a simulated gas mixer, a mercury monitoring system, a flue gas monitoring system and an exhaust cleaning device was constructed for this study (Fig. 1). All individual gas components were supplied from cylinders and were precisely controlled by mass flow controllers (MFCs). The fixed bed reactor was a quartz tube (4 mm i.d.) with a sieve plate (80–100 mesh) in the middle for loading experimental material, and it was placed in a temperature-programmed electric heating furnace. The mercury concentration was monitored with a Lumex RA-915M mercury online analyzer. The mercury analyzer used Zeeman atomic absorption spectroscopy to measure Hg⁰ concentration down to a detection limit of 2 ng m⁻³. The concentration of HBr and Br₂ were measured by a Hiden quantitative gas analysis mass spectrometer (QGA) with a detection limit of 1 ppm. An activated carbon adsorption tube was placed after the mass spectrometer for tail gas adsorbing.

Fig. 1 Schematic diagram of the experimental system.

2.2. Samples and methods

Three types of coal (lignite, anthracite and bituminous coal) were studied. The coal was pretreated by crushing and grinding. Because CaBr₂ was added in trace amounts that were difficult to weigh accurately by an analytical balance and to obtain a more homogeneously mixed sample, CaBr₂ was added in solution form. CaBr₂ was first dissolved in deionized water to form a 2 × 10⁻⁴ g mL⁻¹ solution and then added to coal samples by mixing, ultrasonic concussing for 30 min, vacuum rotary evaporating at 60 °C for 1 h, drying at 105 °C for 3 h and grinding to powder. Coal samples with 0, 100, 200, 300 and 400 μg g⁻¹ CaBr₂ were prepared. The samples (50 mg) were accurately weighed and placed in the quartz tube, swept by a gas mixture comprised of 20 mL min⁻¹ O₂ and 80 mL min⁻¹ N₂ for 30 min, then heated by the electric heating furnace from 50 °C to 900 °C (TPD process) at a rate of 30 °C min⁻¹. With the increasing of temperature, the mercury in the coal samples started to release through thermal decomposition and the Hg⁰ concentration was recorded online by RA915M to obtain the TPD profiles. For each CaBr₂ addition, the experiment was repeated at least three times and took the average to reduce the measurement error. Prior to the experiment, the total mercury content (Hg^t) was measured by Zeeman atomic absorption spectroscopy with the aid of the solid attachment (PYRO-915⁺). The coal sample was heated to 800 °C in the solid attachment, releasing all of the mercury in the form of Hg⁰, which was then directly measured at that temperature. Each coal sample was measured for three times. The analysis method is in accordance with EPA Method 7473 for solid samples.

Because it has been widely accepted that CaBr₂ addition is capable of oxidizing Hg⁰ into Hg²⁺ and considering Lumex RA-915M is able to detect Hg⁰ only, the amount of Hg⁰ released was used as an index to reflect the effect of CaBr₂ addition on Hg⁰ oxidation according to the equation Hg^t = Hg⁰ + Hg²⁺, where Hg^t (μg kg⁻¹) is the total mercury content in coal samples and Hg⁰ (μg kg⁻¹) and Hg²⁺ (μg kg⁻¹) are the Hg⁰ and Hg²⁺ released, respectively, during the TPD experiment. The amount of Hg⁰ released from the coal sample was calculated by integrating the TPD curve.²³ The concentration of HBr and Br₂ was measured by a quantitative gas analysis mass spectrometer. A calibration equation was obtained by a linear regression of the plot of gas concentration versus the detected signal value. The regression coefficient was more than 0.999.

3 Results and discussion

3.1. Characterization of coal samples

To ensure the use of representative samples, elemental analysis was conducted before the experiment to provide essential details about the coal samples. The results are summarized in Table 1. Elemental composition varied by coal type. All of the coal samples used in this experiment were low-sulfur coal. Chlorine and mercury content were compared to other results reported in the literature. Previous studies have indicated that the mercury content in Chinese coal ranges from 20 μg kg⁻¹ to 520 μg kg⁻¹, with average mercury concentrations of 190 μg kg⁻¹ (ref. 24 and 25) and 170 μg kg⁻¹.¹⁵ The mercury content of coal samples in this study were in a reasonable range: 200 ± 3 μg kg⁻¹, 103 ± 8 μg kg⁻¹ and 124 ± 3 μg kg⁻¹ in lignite, anthracite and bituminous coal, respectively. The chlorine content was 143 μg g⁻¹ in lignite, 151 μg g⁻¹ in anthracite, 466 μg g⁻¹ in bituminous coal. These values were comparable to 260 μg g⁻¹, the average chlorine content in Chinese coal as reported by Zhang et al.¹⁵

Table 1 Characterization results of coal samples

Samples	Aad^a (%)	Elemental analysis
		Cad (%)	Had (%)	Oad (%)	Nad (%)	Sad (%)	Cl (μg g^−1)	Hg (μg kg^−1)
a ad refers to the air dried basis in industrial analysis of coal.
Lignite	55.84	23.59	2.79	6.87	0.82	0.18	143	200 ± 3
Anthracite	13.76	65.58	4.08	6.26	0.84	0.90	151	103 ± 8
Bituminous coal	16.41	70.42	4.16	6.75	0.68	0.59	466	124 ± 3

---

3.2. Hg⁰ release from coal samples without CaBr₂ addition

Fig. 2 shows the amount of Hg⁰ released from three coal samples during the TPD experiment. Approximately 70–90% of mercury in coal was released in form of Hg⁰. The Hg⁰ release ratios (ratio of Hg⁰ released to the total mercury content in coal) were 92%, 70% and 71% in lignite, anthracite and bituminous coal, respectively. The different Hg⁰ release ratios were related to chlorine content in coal. It has been widely shown that high halogen content is conducive to Hg⁰ oxidation during coal combustion.^5,15,26,27 Among the three coal samples, the chlorine content was highest in bituminous coal, which had the lowest Hg⁰ release ratio in this experiment. The chlorine content was the lowest in lignite (143 μg g⁻¹), which had the highest Hg⁰ release ratio.

Fig. 2 Total mercury content in coal and Hg⁰ release during TPD process without CaBr₂ addition.

The TPD curves of the three coal samples are shown in Fig. 3. The temperatures of Hg⁰ release were distinctly different due to varying mercury occurrence modes in coal. For bituminous coal, there was only one main Hg⁰ release peak over the temperature range over 200–350 °C, with a spark at 318 °C and for anthracite the peak appeared at 375 °C over the temperature range 300 °C to 400 °C. The lignite had three peaks at 306 °C, 386 °C and 784 °C. The assignment of peaks was carried out based on the different thermal decomposition temperatures of different mercury species. According to Luo et al.²⁸ and M. Rumayor et al.,²⁹ HgS decomposes as shown in eqn (1) over the temperature range 230 °C to 350 °C, with a spike at 310 °C. The peak at approximately 400 °C corresponded to the decomposition of HgO (ref. 30) shown in eqn (2).

HgS (s) → Hg (g) + ½S₂ (1)

HgO (s) → Hg (g) + ½O₂ (2)

Fig. 3 TPD profiles of Hg⁰ release of three coal samples.

The peak appearing over 750 °C to 850 °C was attributed to silicate-bound Hg.²³ In summary, HgS and HgO were the primary occurrence modes in bituminous coal and anthracite, respectively, while lignite contained HgS, HgO and silicate-bound Hg mercury occurrence forms.

3.3. Hg⁰ release from coal samples with CaBr₂ addition

Lignite was used in this study because it contains multiple mercury occurrence modes and has a low chlorine content. Fig. 4 shows the amount of Hg⁰ released from lignite samples with different amounts of CaBr₂ additive. Initially, the amount of Hg⁰ released decreased with increasing amounts of CaBr₂ addition. The amount of Hg⁰ released decreased by 40.4% with 200 μg g⁻¹ of CaBr₂ additive, then remained unchanged despite increasing the amount of additive. These findings agree with previous studies^14–17,31 that found the oxidation of Hg⁰ was enhanced during combustion by addition of CaBr₂ to coal. The Hg⁰ oxidation efficiency reported here was lower than values reported in the literature from other field tests. This is likely due to the use of highly pure simulated flue gas under laboratory conditions, as opposed to real flue gas which contains a certain amount of fly ash and has a retention effect on Hg⁰.^32,33

Fig. 4 Hg⁰ release with different CaBr₂ additive amount in coal.

To further investigate the effect of CaBr₂ additives on Hg⁰ release from coal, TPD curves of Hg⁰ release at different levels of CaBr₂ addition were analyzed. As shown in Fig. 5, the position of Hg⁰ release peaks remained unchanged with different amounts of CaBr₂ additive. This result indicated that CaBr₂ addition did not change the mercury occurrence modes in coal. However, calculating the amount of Hg⁰ released at different temperatures by separating and integrating Hg⁰ release peaks indicated the effect of CaBr₂ on different mercury occurrence modes was distinct. As shown in Fig. 6, during the thermal decomposition of HgS (306 °C), the amount of Hg⁰ released generally decreased with increasing amounts of CaBr₂ in coal. During the thermal decomposition of HgO (386 °C), the amount of Hg⁰ released decreased when the amount of CaBr₂ additive was less than 300 μg g⁻¹ and increased with larger amounts of CaBr₂ additive. During the thermal decomposition of silicate-bound Hg (784 °C), however, the amount of CaBr₂ additive had no effect on the amount of Hg⁰ released. These results suggest that CaBr₂ was only effectively oxidized Hg⁰ released at relatively low temperatures.

Fig. 5 TPD profiles of Hg⁰ release with different CaBr₂ additive amount.

Fig. 6 Hg⁰ release at different temperature with different CaBr₂ addition.

3.4. The release of HBr and Br₂ during TPD experiment

The results of monitoring HBr and Br₂ concentration in flue gas during TPD are summarized in Fig. 7. The results confirmed the generation of HBr in coal-fired flue gas due to the CaBr₂ additive in coal, but Br₂ was not detected. HBr was possibly generated by the hydrolysis of CaBr₂ via eqn (3).^34,35 Based on previous study,³⁵ the water in eqn (3) was probably generated by hydrogen containing organisms in coal during the TPD process through eqn (4).

CaBr₂ + H₂O → CaO + HBr (3)

C_nH_mOH + O₂ → H₂O + CO₂ (4)

Fig. 7 TPD profiles of HBr with different CaBr₂ addition.

The formation of HBr occurred over the temperature range 250 °C to 400 °C, reaching a maximum concentration at approximately 280 °C. Remarkably, the temperature range was similar to the thermal decomposition of HgS and HgO. As shown in Fig. 6, Hg⁰ released over that temperature range was significantly decreased. These findings may explain why CaBr₂ only effectively oxidized Hg⁰ released at relatively low temperatures and also suggest that the amount of Hg⁰ released is strongly related to the HBr concentration in flue gas. The HBr concentration in flue gas increased with increasing amounts of CaBr₂ additive up to 18 ppm, which was detected at the 200 μg g⁻¹ CaBr₂ maximum. Interestingly, the HBr concentration no longer increased beyond this amount of CaBr₂ additive. This phenomenon was also observed by Berry et al.¹⁶ during the spot test at Plant Miller. In previous studies, CaO has been shown to fix bromine during coal combustion.^36,37 It is possible that when HBr concentration was high, it may be adsorbed by CaO.

3.5. Analysis of oxidation mechanism of Hg⁰ by CaBr₂ addition

Based on previous theories and the results of this work, the oxidation mechanism of Hg⁰ by CaBr₂ addition was analyzed. It is currently held that bromine species can oxidize Hg⁰ in flue gas through homogeneous reactions.^38–42 Because HBr is in a reduced state and cannot directly oxidize Hg⁰, the Hg⁰–bromine reactions must occur with a more reactive form of bromine such as atomic bromine (Br˙) or molecular bromine (Br₂). Ogg et al.⁴³ concluded that the mechanism involves the reactions (Hg + Br₂ → HgBr + Br˙) and (HgBr + Br˙ → HgBr₂ + Br˙). Goodsite et al.⁴⁴ proposed by computational chemistry that the mechanism involves the reactions (Hg + Br˙ → HgBr˙) and (HgBr˙ + Br˙ → HgBr₂). Because no Br₂ was detected in this study, it is more likely that the Hg⁰ was oxidized by Br˙, although Br˙ was not detected because of the detection limit of the instrument used. In this study, HBr was the primary bromine species detected in flue gas. Additional analysis indicated that there was a strong, negative linear correlation between the amount of Hg⁰ released and the HBr concentration in flue gas (Fig. 8). The correlation coefficient (R²) was 0.9149, strongly suggesting that the Br˙ was generated from HBr.

Fig. 8 Dependence of Hg⁰ release on HBr concentration in flue gas.

4 Conclusions

Oxidation of Hg⁰ was enhanced by adding CaBr₂ to coal. A 40.4% reduction rate of Hg⁰ release was observed with the addition of 200 μg g⁻¹ CaBr₂ to coal and release of Hg⁰ remained constant with larger amounts of CaBr₂ additive. HgS, HgO and silicate-bound Hg are three mercury occurrence modes found in lignite. This study confirmed the generation of HBr from the hydrolysis of CaBr₂ in flue gas over the temperature range of 250 °C to 400 °C, while no Br₂ was detected during the experiment. Based on the similar temperature ranges obtained during the thermal decomposition of HgS and HgO, it was determined that CaBr₂ addition only effectively oxidizes Hg⁰ released at relatively low temperatures. The HBr concentration in flue gas initially increased with increasing amounts of CaBr₂ additive in coal but did not rise with larger additive amounts. The oxidation of Hg⁰ by CaBr₂ may occur through a homogeneous oxidation reaction between Hg⁰ and Br˙. The strong, negative linear correlation between the amount of Hg⁰ released and the HBr concentration in flue gas suggested that Br˙ is likely generated from HBr. It is worth mentioning that this work was performed in the atmosphere only containing N₂ and O₂. The effect of other flue gas components such as SO₂, NO or HCl will be researched in our next work. Further studies of the generated HBr and its impact needs to be conducted to comprehensively evaluate CaBr₂ addition to coal as a mercury control technology.

Acknowledgements

This work was supported by the Special Research Funding for Public Benefit Industries from the National Ministry of Environmental Protection (201309018), the National Basic Research Program (973) of China (No. 2013CB430005).

Notes and references

1. S. X. Wang, L. Zhang, L. Wang, Q. R. Wu, F. Y. Wang and J. M. Hao, Front. Environ. Sci. Eng., 2014, 8, 631–649.
2. Y. J. Wang, Y. F. Duan, L. G. Yang, C. S. Zhao, X. L. Shen, M. Y. Zhang, Y. Q. Zhu and C. H. Chen, Fuel Process. Technol., 2009, 90, 643–651.
3. Z. J. Zhou, X. W. Liu, B. Zhao, Z. G. Chen, H. Z. Shao, L. L. Wang and M. H. Xu, Fuel Process. Technol., 2015, 131, 99–108.
4. W. Chengli, C. Yan, D. Zhongbing, C. Chinmin, L. Hanxu and P. Weiping, J. Environ. Sci., 2010, 22, 277–282.
5. L. Chen, Y. F. Duan, Y. Q. Zhuo, L. G. Yang, L. Zhang, X. H. Yang, Q. Yao, Y. M. Jiang and X. C. Xu, Fuel, 2007, 86, 603–610.
6. M. Rallo, M. A. Lopez-Anton, R. Meij, R. Perry and M. M. Maroto-Valer, J. Hazard. Mater., 2010, 174, 28–33.
7. G. E. Dunham, R. A. DeWall and C. L. Senior, Fuel Process. Technol., 2003, 82, 197–213.
8. L. Zhang, Y. Q. Zhuo, L. Chen, X. C. Xu and C. H. Chen, Fuel Process. Technol., 2008, 89, 1033–1040.
9. M. Diaz-Somoano, S. Unterberger and K. R. G. Hein, Fuel Process. Technol., 2007, 88, 259–263.
10. S. X. Wang, L. Zhang, G. H. Li, Y. Wu, J. M. Hao, N. Pirrone, F. Sprovieri and M. P. Ancora, Atmos. Chem. Phys., 2010, 10, 1183–1192.
11. R. Ochoa-Gonzalez, P. Cordoba, M. Diaz-Somoano, O. Font, M. A. Lopez-Anton, C. Leiva, M. R. Martinez-Tarazona, X. Querol, C. F. Pereira, A. Tomas, P. Gomez and P. Mesado, Chemosphere, 2011, 85, 565–570.
12. A. A. Presto and E. J. Granite, Environ. Sci. Technol., 2006, 40, 5601–5609.
13. J. Xiang, P. Y. Wang, S. Su, L. Q. Zhang, F. Cao, Z. J. Sun, X. Xiao, L. S. Sun and S. Hu, Fuel Process. Technol., 2015, 135, 168–173.
14. M. P. Ancora, L. Zhang, S. X. Wang, J. Schreifels and J. M. Hao, J. Environ. Sci., 2015, 33, 125–134.
15. L. Zhang, S. X. Wang, Y. Meng and J. M. Hao, Environ. Sci. Technol., 2012, 46, 6385–6392.
16. M. Berry, Air Quality VI Conference, Crystal Gateway Marriott, Arlington, Washington, DC, September, 2007.
17. X. Shi, W. Wen, Z. Xue, X. Liu, F. Chai, G. Zhi, Y. Ren, J. Ma and W. Gao, J. Chin. Soc. Power Eng., 2014, 34, 482–486.
18. Y. Zhuang, C. M. Chen, R. Timpe and J. Pavlish, Fuel, 2009, 88, 1692–1697.
19. L. Lisi, G. Lasorella, S. Malloggi and G. Russo, Appl. Catal., B, 2004, 50, 251–258.
20. J. Yang, Q. Yang, J. Sun, Q. C. Liu, D. Zhao, W. Gao and L. Liu, Catal. Commun., 2015, 59, 78–82.
21. A. S. Mahajan, H. Oetjen, J. D. Lee, A. Saiz-Lopez, G. B. McFiggans and J. M. C. Plane, Atmos. Environ., 2009, 43, 3811–3818.
22. Y. C. Gao, M. X. Sun, X. W. Wu, Y. D. Liu, Y. Q. Guo and J. Wu, Atmos. Environ., 2010, 44, 4298–4302.
23. S. Q. Guo, J. L. Yang and Z. Y. Liu, Energy Fuels, 2009, 23, 4817–4821.
24. D. G. Streets, J. M. Hao, Y. Wu, J. K. Jiang, M. Chan, H. Z. Tian and X. B. Feng, Atmos. Environ., 2005, 39, 7789–7806.
25. L. G. Zheng, G. J. Liu and C. L. Chou, Sci. Total Environ., 2007, 384, 374–383.
26. Y. W. Tan, R. Mortazavi, B. Dureau and M. A. Douglas, Fuel, 2004, 83, 2229–2236.
27. Y. Zhuang, J. S. Thompson, C. J. Zygarlicke and J. H. Pavlish, Fuel, 2007, 86, 2351–2359.
28. G. Q. Luo, H. Yao, M. H. Xu, R. Gupta and Z. H. Xu, Proc. Combust. Inst., 2011, 33, 2763–2769.
29. M. Rumayor, M. Diaz-Somoano, M. A. Lopez-Anton and M. R. Martinez-Tarazona, Talanta, 2013, 114, 318–322.
30. B. V. L'Vov, Thermochim. Acta, 1999, 333, 21–26.
31. J. Ma, H. Yao, G. Luo, X. Fang, W. Liu and M. Xu, J. Eng. Thermophys., 2010, 31, 1407–1410.
32. M. A. Lopez-Anton, M. Diaz-Somoano and M. R. Martinez-Tarazona, Energy Fuels, 2007, 21, 99–103.
33. M. A. Lopez-Anton, M. Diaz-Somoano and M. R. Martinez-Tarazona, Ind. Eng. Chem. Res., 2007, 46, 927–931.
34. M. Zhang, P. Wang, Y. Dong, H. Sui and D. Xiao, Chem. Eng. J., 2014, 253, 243–250.
35. B. X. Peng and D. S. Wu, Fuel, 2015, 157, 82–86.
36. Y. Mochizuki, K. Sugawara and Y. Enda, Energy Fuels, 2009, 23, 4502–4506.
37. L. Bartonova and Z. Klika, J. Environ. Sci., 2014, 26, 1429–1436.
38. Y. Cao, Z. Gao, J. Zhu, Q. Wang, Y. Huang, C. Chiu, B. Parker, P. Chu and W. P. Pan, Environ. Sci. Technol., 2008, 42, 256–261.
39. S.-H. Liu, N.-Q. Yan, Z.-R. Liu, Z. Qu, P. Wang, S.-G. Chang and C. Miller, Environ. Sci. Technol., 2007, 41, 1405–1412.
40. Z. Qu, N. Q. Yan, P. Liu, Y. P. Chi and J. Jia, Environ. Sci. Technol., 2009, 43, 8610–8615.
41. I. Preciado, T. Young and G. Silcox, Energy Fuels, 2014, 28, 1255–1261.
42. I. Auzmendi-Murua, A. Castillo and J. W. Bozzelli, J. Phys. Chem. A, 2014, 118, 2959–2975.
43. R. A. Ogg, H. C. Martin and P. A. Leighton, J. Am. Chem. Soc., 1936, 58, 1922–1924.
44. M. E. Goodsite, J. M. C. Plane and H. Skov, Environ. Sci. Technol., 2004, 38, 1772–1776.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra11468g

---