Preparation of polypropylene based hyperbranched absorbent fibers and the study of their adsorption of CO₂

Abstract

Ring-opening copolymerization of aziridine in situ into substrates to introduce polyethyleneimine (PEI) not only required strict experimental conditions but employed the highly toxic monomer aziridine. In this paper, an effective and safe new procedure is developed for hyperbranched structure synthesis by stepwise growth using N-(2-chloroethyl)-benzaldimine as a monomer. Hyperbranched solid amine fibers for CO₂ capture were prepared through co-irradiation grafting copolymerization of polypropylene fibers with glycidyl methacrylate, followed by amination with ethylenediamine, Hoffman alkylation with N-(2-chloroethyl)-benzaldimine and then hydrolysis to remove benzaldehyde groups. It was shown that the adsorption performance of the hyperbranched solid amine fibers G2.0 and G3.0 has been greatly improved compared with first generation G1.0 fibers, and the adsorption capacities of G2.0 and G3.0 were 5.35 mmol g⁻¹ and 5.53 mmol g⁻¹ at 30 °C, respectively. The amine utilization of G2.0 fibers could reach 84.1%. These results demonstrate that the branched structure can promote the adsorption capacity and efficiency greatly due to its low mass transfer resistance of CO₂, which is more favorable than a linear amination reagent.

---

1. Introduction

Global warming is mainly attributed to an increase in atmospheric levels of CO₂ which is attributed to the unrestrained burning of fossil fuels.¹ The most effective method to reduce CO₂ emission from the burning of fossil fuels is scrubbing the flue gas stream with an amine solution. But this brings some disadvantages such as corrosion, high regeneration energy and solvent degradation.^2,3 To solve these problems, solid amine adsorbents such as amine-treated MCM-41,^4,5 SBA-15,^6,7 Zeolite 13X^8,9 and activated carbon^10,11 have been used, attracting much attention. The solid amine adsorbents can retain high adsorption capacities even in the presence of water, compared with microporous or mesoporous adsorbents without amine modification. What is more, they can also be regenerated in mild conditions. The adsorption capacity for CO₂ mainly depends on the amine loading and the pore volume of the porous substrates. However, increasing amine loading will block the pore volume of the porous substrate, thus reducing the adsorption capacity of CO₂.^6,12 So it is of great significance to develop alternative substrates or to improve the amine agents to increase amine loading without affecting adsorption performance.

Much work advocating the use of fibrous adsorbents for CO₂ adsorption has been published by our group,^13–16 demonstrating that fibers have great potential as substrates for CO₂ adsorption. Fibrous adsorbents possess the advantages of large external surface area, short transit distance, low pressure drops, and flexibility. Yang¹⁵ grafted allylamine onto polyacrylonitrile fibers and obtained an adsorption ability of 6.22 mmol g⁻¹ PAN-AF when the grafting degree was 60.0 wt%. Zhuang¹⁶ grafted glycidyl methacrylate onto polypropylene fibers, followed by reaction with triethylenetetramine to make a novel kind of solid amine-containing fibrous adsorbent and obtained an adsorption ability of 4.72 mmol g⁻¹. What is more, the adsorbent could maintain almost constant adsorption behavior within six recycles.

In 1990, Kim put forward the concept of “hyperbranch”, and this then became one of the hottest research topics in Polymer Chemistry.¹⁷ Unlike dendritic molecules, hyperbranched molecules do not require uniform structure and high symmetry, thus they are easy to prepare. The most important property is that hyperbranched molecules can have many functional groups that could both improve the hydrophilicity of the material and increase the number of adsorption sites for use as an adsorbent.^18,19 Among them, polyethyleneimine (PEI) because of its high amine content and hyperbranched structure was widely reported to modify silicon dioxide^20,21 and fibers^13,14 that are used for gas separation. The in situ ring-opening polymerization of aziridine into porous silica to introduce PEI for the synthesis hyperbranched polymers has been conventionally carried out and used for CO₂ capture. However, on the one hand, the approach has some additional disadvantages such as the use of the highly toxic monomer aziridine and strict experimental conditions.^22,23 On the other hand, Li²⁴ has studied the effect of the molecular weight of the PEI on the CO₂ capture performance of PEI-nano silica adsorbents and found that a PEI with a high molecular weight leads to unsatisfactory amine utilization because of its high steric hindrance due to the increased viscosity. Thus the development of a PEI with low molecular weight and with an accessible structure will be of benefit to improve its adsorption capacity and the amine efficiency.

In this study, an effective and safe new procedure is developed for hyperbranched structure synthesis by stepwise growth using N-(2-chloroethyl)-benzaldimine as a monomer. The hyperbranched solid amine fibers for CO₂ capture were prepared through co-irradiation grafting copolymerization, followed by amination, Hoffman alkylation and hydrolysis. The effects of the branched structure on the adsorption performance of the fiber have been investigated.

2. Experimental

2.1 Materials and reagents

Glycidyl methacrylate (GMA, Sigma-Aldrich CO) was used after the removal of the inhibitor with neutral aluminum oxide. Polypropylene (PP) fibers were provided by Xinshun Special Fiber Company (Zhongshan, China). Methanol, ethanol, triethylamine (TEA), hydrochloric acid (HCl), aluminum chloride hexahydrate (AlCl₃·6H₂O), magnesium sulfate anhydrous (MgSO₄), benzaldehyde, ammonium ferrous sulfate ((NH₄)₂Fe(SO₄)₂·6H₂O), sodium hydroxide (NaOH), potassium iodide (KI) and ethylenediamine (EDA) were all purchased from Guangzhou Chemical Reagents Factory. Diethylenetriamine (DETA) was purchased from Tianjin Fuchen Chemical Reagents Factory. 2-Chloroethylamine hydrochloride was purchased from Aladdin. Ether was provided by Kaixin Chemical Industry Co. (Hengyang City). All reagents above were analytical grade (AR).

2.2 Preparation of hyperbranched solid amine fibers

2.2.1 Synthesis of N-(2-chloroethyl)-benzaldimine (CEBI)²⁵. To a 250 mL three-neck round bottom flask was added 5.3 g (50 mmol) benzaldehyde, 5.8 g (50 mmol) 2-chloroethylamine hydrochloride, 10 g MgSO₄ and 50 mL anhydrous ether. Then a mixed solution (5.6 g (55 mmol) TEA and 50 mL anhydrous ether) was added to the reaction system dropwise over 30 min. The reaction was carried out at room temperature, and stirred for 6 hours. After reaction, the solvent was removed in vacuo to obtain N-(2-chloroethyl)-benzaldimine (CEBI).

2.2.2 Synthesis of PP-GMA. PP, GMA and DMF were added into a 250 mL conical flask, and then subjected to γ-ray irradiation. After irradiation, the grafted fiber (PP-GMA) was washed with DMF, and then dried at 60 °C for 12 h. The grafting degree of PP-GMA was calculated using eqn (1):where W₀ and W₁ are the weights (g) of the original and grafted PP fiber, respectively.

2.2.3 Synthesis of G1.0 fiber: PP-GMA-EDA. PP-GMA, AlCl₃·6H₂O and EDA were mixed and added into a 100 mL three neck flask. The reaction was carried out at 105 °C for 6 h. After reaction, the product fibers (G1.0, PP-GMA-EDA) were washed with deionized water thoroughly to remove the residual AlCl₃ and EDA, then dried at 60 °C for 12 h. The primary or secondary amine contents (n, mmol g⁻¹) of G1.0 fibers were calculated by the following equation:where E is the mass fraction of nitrogen measured by elemental analysis.

2.2.4 Synthesis of G1.5 fibers by Hoffman alkylation: PP-GMA-EDA-CEBI. 1.5 g PP-GMA-EDA (G1.0) fiber, 60 mL deionized water and 1.32 g KI were added into a 100 mL three-neck flask, then 9 g CEBI in 30 mL anhydrous methanol was slowly dropped into the reaction system. After reflux of the solution at 60 °C for 12 h the obtained fiber (G1.5, PP-GMA-EDA-CEBI) was washed with excess deionized water and anhydrous ethanol, and then dried at 60 °C for 12 h.

2.2.5 Synthesis of G2.0 fibers: PP-GMA-EDA-CEA. The bulky phenyl groups of CEBI were removed from the G1.5 fiber by mixing the PP-GMA-EDA-CEBI G1.5 Fiber with 4 M HCl (90 mL) and reacted at 30 °C for 8 h. Then the fiber was washed with 1 M NaOH and with deionized water until the solution turned neutral. And the resulting product G2.0 fiber (PP-GMA-EDA-CEA) was dried at 60 °C under vacuum. The contents of primary, secondary, tertiary amines or imine (n, mmol g⁻¹) of G2.0 fibers were calculated by the following equation:where E is the mass fraction of nitrogen measured by elemental analysis. A_i represents the different types of nitrogen peak area tested by X-ray photoelectron spectrometer.

2.2.6 Synthesis of G2.5 and G3.0 fibers. The experimental conditions used to prepare G2.5 and G3.0 fibers were the same as for the description of G1.5 and G2.0 fibers, and the calculation method for the different types of amine is as for eqn (3).

The preparation scheme of the hyperbranched solid amine fibers is presented in Scheme 1.

Scheme 1 Reaction scheme for hyperbranched adsorbent fiber preparation.

2.2.7 Synthesis of G1.0-DETA fibers. In order to evaluate the effect of the structure of amination reagent on the adsorption, G1.0-DETA fibers (PP-GMA-DETA) were synthesized through using DETA, instead of EDA, to react with G0.5 PP-GMA fibers. DETA is an amination reagent with a longer linear chain. The experimental conditions used to prepare G1.0-DETA fibers are the same as for the description of G1.0 fibers. The structure of the G1.0-DETA fiber is presented in Scheme 2.

Scheme 2 Structure of G1.0-DETA fiber.

In general description, all the modified fibers with amino groups as end groups, G1.0, G2.0 and G3.0, are described as full generation fibers; and those fibers with phenyl groups as the end groups, G0.5, G1.5, and G2.5, are described as half generation fibers.

2.3 Physical and chemical characterization

Infrared (IR) spectra (Tensor-27 spectrometer), Elemental analysis (Elementar, Vario EL), solid state ¹³C NMR analysis (AVANCE AV, Bruker) and X-ray photoelectron spectroscopy (ESCALAB 250, Thermo-VG Scientific) were used to confirm the structure of the polymers.

A thermogravimetric analyzer (Netzsch TG-209) was employed to determine the thermal stability of different full generation samples. The test was carried out from 100 °C to 600 °C using a heating rate of 10 K min⁻¹.

2.4 CO₂ adsorption experiment

1.0 g of the fiber sample was placed into a specially designed glass tube (Φ = 13 mm), followed by purging with dry N₂ to remove air and excess water. Then a mixture of gases (CO₂ : N₂ = 1 : 9 (volume ratio)) with a flow rate of 30 mL min⁻¹ was introduced. The outlet CO₂ concentration was determined by gas chromatography (Techcomp 7900) with a thermal-conductivity detector.

The adsorption amount was calculated as follows:

where Q represents the adsorption capacity (mmol CO₂ g⁻¹), t is the adsorption time (min). C_im and C_eff are the influent and effluent concentrations of CO₂ (vol%), respectively. V is the total flow rate, 30 mL min⁻¹; W and 22.4 are the weight of sample (g) and molar volume of gas (mL mmol⁻¹), respectively.

After adsorption, the fiber was regenerated by purging with N₂ with a flow rate of 30 mL min⁻¹ at 90 °C.

3. Result and discussion

3.1 Chemical characterization

Fig. 1 shows the FT-IR spectrum of grafted fibers at each stage. Compared with G0 fibers, there is a sharp band at 1736 cm⁻¹, which is attributed to the C=O stretching vibration of grafted GMA. In addition, a broad and strong absorption band at 3412 cm⁻¹ is attributed to O–H bending vibration. The intensity of this band at 3412 cm⁻¹ increased after amination because it is the overlap of O–H and N–H bending vibrations. The result above proves that EDA was successfully introduced onto the surface of the material. And it is found that there was only a band at 1542 cm⁻¹ for N–H in-plane bending vibration in the range from 1550 cm⁻¹ to 1650 cm⁻¹ for the full generation fiber (G1.0, G2.0 and G3.0), but there was another band at 1640 cm⁻¹ for C=N stretching vibration for the half generation (G0.5, G1.5 and G2.5), demonstrating the successful preparation of a hyperbranched absorbent.

Fig. 1 FT-IR spectra of G0, G0.5, G1.0, G1.5, G2.0, G2.5, G3.0 fibers.

In order to determine the type of nitrogen and whether the structure was a linear chain or a branched one, elemental analysis, X-ray photoelectron spectroscopy and ¹³C NMR spectra were carried out and are presented in Tables 1, 2 and Fig. 2, 3, respectively. Compared with G0, both the content of carbon and hydrogen of G0.5 decreased and the calculated grafting degree according to the elemental analysis was 292%, corresponding well to the result measured from weight gain (312%). After amination with EDA, the nitrogen content of G1.0 reached 6.04 mmol g⁻¹, similar to the data (6.67 mmol g⁻¹) calculated by weight gain. Furthermore, it was obvious that the nitrogen content of the full generation increased with increasing generation, proving that the monomer (CEBI) was successfully introduced onto the surface of the fibers. However, the amine number of G2.0 and G3.0 from weight gain was a little lower than that from the elemental analysis, which could be attributed to the difficulty of collection of G2.0 and G3.0 after treatment by strong acid and strong base that led to the brittleness of the fibers. Moreover, with respect to G2.0 and G3.0, incomplete hydrolysis which is confirmed by the ¹³C NMR spectra where the chemical shifts of the benzyl and imine groups appear at 130 ppm and 140 ppm respectively, gave an explanation for the lower amine content than expected. In addition, as shown in Table 2, combining the elemental analysis data and X-ray photoelectron spectroscopy,^26,27 the amount of the different types of amines was calculated. The presence of tertiary amine implicated that the branching can happen to form a hyperbranched polymer.

Table 1 Element analysis of G0, G0.5, G1.0, G2.0 and G3.0 fibers

Fibers	WG^a (g per g raw material)	Element content (wt%)			Amine content (mmol g^−1)
		C/%	H/%	N/%
a , where WG represents the weight gain (g per g raw material), Wr and Wp are the weight of the raw material and product, respectively.b The unknown quantities of Cl^− would affect the calculation if the raw material refers to G1.5. Thus for the accuracy and convenience, the raw material in this step refers to G1.0.c The raw material in this step refers to G2.0.
G0	—	84.30	13.66	0.00	—
G0.5	3.12	65.59	9.02	0.00	—
G1.0	0.25	50.88	9.46	8.46	6.04
G2.0	0.30^b	53.57	9.42	10.86	7.76
G3.0	0.22^c	52.74	9.34	12.09	8.64

---

Table 2 The amine type and content of G1.0, G2.0 and G3.0 fibers

Fibers	Amine content (mmol g^−1)
	Total	Primary	Secondary	Tertiary	Imine
G1.0	6.04	3.02	3.02	—	—
G2.0	7.76	2.58	1.89	1.89	1.39
G3.0	8.64	2.84	2.29	2.02	1.48

---

Fig. 2 ¹³C NMR spectra for G2.0 and G3.0.

Fig. 3 Typical XPS wide scan spectra for G2.0 and G3.0. The inset shows the characteristic peaks of N 1s at higher resolution.

The results of TG analysis of different generation absorbent fibers are presented in Fig. 4. In the case of PP, there was only one continuous weight loss from 310 °C to 460 °C. In the TG curve of PP-GMA, three distinct weight loss steps were observed at 200 °C, 260 °C, and 350 °C. The first weight loss could be attributed to decomposition of the homopolymer, and the second one between 260 °C and 350 °C was probably related to the degradation of GMA, which also indicated that the monomer GMA was grafted onto the surface of PP. Comparing the G1.0, G2.0 and G3.0 fibers, the thermal stability was G3.0 > G1.0 > G2.0. As is known to all, increasing chain length would reduce the thermal stability. However, G3.0 with the longest chain length showed better stability than the first two generations, overcoming this defect.

Fig. 4 The thermogravimetric curve of different generation absorbent fibers.

3.2 Adsorption property

The cumulative adsorption curves are shown in Fig. 5, and the adsorption capacity and amine utilization efficiency of G1.0, G2.0, G3.0 fibers are shown in Table 3. With the increase of fiber generation, the adsorption capacities increased due to the increase of alkyl amine content, demonstrating the importance of high amine content.

Fig. 5 The cumulative adsorption curves of G1.0, G2.0, G3.0 fibers (adsorption temperature: 30 °C; N₂: 27 mL min⁻¹; CO₂: 3 mL min⁻¹).

However, the equilibrium adsorption time increased with the thickening of the adsorption layer which would extend the diffusion time. Moreover, the efficiency of G3.0 decreased to 77.3% while G2.0’s reached 84.1%, which could be ascribed to the longer chain length which increases the steric resistance and disfavors the adsorption. This reminds us of the fact that an appropriate amination reagent which combines alkyl amine content with lower steric resistance should be chosen.

For the purpose of studying the effect of a branched structure on the CO₂ adsorption performance, DETA was employed and immobilized onto PP-GMA fiber. Since G2.0 (PP-GMA-EDA-EDA) and G1.0-DETA (PP-GMA-DETA) fibers had similar nitrogen content, the influence of amine content on the adsorption capacity was excluded based on the mechanism in the presence of water vapor.^28,29 In addition, G2.0 (PP-GMA-EDA-EDA) had a branched structure, and G1.0-DETA (PP-GMA-DETA) had a linear one. The adsorption capacities and efficiency of G2.0 and G1.0-DETA were compared. As shown in Fig. 6 and Table 3, in the early stages, the adsorption rate of G1.0-DETA was almost the same (up to 20 min) as G2.0’s, and the adsorption capacities of the two materials were 2 mmol g⁻¹ at that stage. The results may be attributed to similar numbers of primary amines on the outer layer. After 20 min, the adsorption rate of G1.0-DETA became slow on account of the higher mass transfer resistance of CO₂ into the inner layer for the molecular chain entanglement. Meanwhile, the molecular chain entanglement would also prevent the adsorption sites from reacting with CO₂. Both the adsorption capacities and the amine utilization of G2.0 fiber were much higher than G1.0-DETA. The above results demonstrate that the branched structure of G2.0 (PP-GMA-EDA-EDA) is more favorable for adsorption in the presence of water than the linear structure of G1.0-DETA.

Fig. 6 The cumulative adsorption curves of G2.0 and G1.0-DETA fibers (adsorption temperature: 30 °C; N₂: 27 mL min⁻¹; CO₂: 3 mL min⁻¹).

Table 3 The adsorption capacity and amine utilization efficiency for CO₂ of G1.0, G2.0, G3.0 and G1.0-DETA (adsorption temperature: 30 °C; N₂: 27 mL min⁻¹; CO₂: 3 mL min⁻¹)

Sample	Alkyl amine (mmol g^−1)	Adsorption capacity (mmol g^−1)	Efficiency (%)
G1.0	6.04	4.08	67.5
G2.0	6.36	5.35	84.1
G3.0	7.15	5.53	77.3
G1.0-DETA	6.64	3.59	54.1

---

Compared with other fibrous adsorbents reported (Table 4), G3.0 showed relatively high adsorption efficiency, demonstrating its potential for practical application. More interestingly, Wu¹⁴ has prepared a fibrous adsorbent (PP-GMA-PEI) through grafting GMA onto PP fiber, followed by reaction with PEI. Though PP-GMA-PEI has a similar amine content to G3.0, its adsorption capacity was far below than G3.0’s, further confirming the superiority of our route.

Table 4 Comparison of CO₂ adsorption capacity of amine functionalized fibrous sorbents

Substrate	Grafted monomer	Amine	Temperature (°C)	pCO2 (atm)	Adsorption capacity (mmol g^−1)	Ref.
a Allylamine was directly grafted onto the polyacrylonitrile fiber (PAN).b PEI was coated onto glass fiber by using epichlorohydrin as cross-linking agent.
PP	GMA	Branched	30	0.10	5.53	This work
PP	GMA	PEI	25	0.10	2.51	14
PP	Am	PEI	25	0.10	5.91	14
PP	GMA	TETA	30	0.15	4.72	16
PAN	—^a	Allylamine	22	0.15	6.22	15
Glass fiber	—^b	PEI	30	0.24	4.12	13

---

In order to evaluate the regeneration performance of G3.0 fibers, 8 cycles of adsorption–desorption were carried out, and the results are shown in Fig. 7. After 8 cycles, no significant decrease in CO₂ adsorption capacity was observed: the adsorption capacity of adsorbent was maintained above 99%. Furthermore, FT-IR spectra of fresh and regenerated G3.0 fibers are presented in Fig. 8. It is obvious that the regenerated fibers exhibit nearly the same spectra as the fresh. All these results confirmed that G3.0 fibers have great regeneration performance that is of great significance for practical application.

Fig. 7 The regeneration performance of G3.0 fiber.

Fig. 8 FT-IR spectra of fresh and regenerated G3.0 fibers.

4. Conclusion

Polypropylene based hyperbranched absorbent fibers for CO₂ capture were prepared by co-irradiation grafting copolymerization of GMA onto PP fiber, followed by amination, Hoffman alkylation and hydrolysis. The present work has shown that the adsorption performance of G2.0 and G3.0 is greatly been improved compared with G1.0, and the adsorption capacities of G2.0 and G3.0 were 5.35 mmol g⁻¹ and 5.53 mmol g⁻¹ at 30 °C, respectively. In particular, a branched structure can promote the adsorption capacity and efficiency greatly due to its low mass transfer resistance of CO₂, which is more favorable than linear amination reagent. Furthermore, hyperbranched absorbent fibers could be easily and completely regenerated under mild conditions and are stable in cyclic operations for 8 cycles.

Acknowledgements

The authors gratefully acknowledge the financial support provided by the National Natural Science Foundation of China (Grant no. 51173211, 51473187).

References

1. D. Aaron and C. Tsouris, Sep. Sci. Technol., 2005, 40, 321–348.
2. L. Espinal, D. L. Poster, W. Wong-Ng, A. J. Allen and M. L. Green, Environ. Sci. Technol., 2013, 47, 11960–11975.
3. L. Dumée, C. Scholes, G. Stevens and S. Kentish, Int. J. Greenhouse Gas Control, 2012, 10, 443–455.
4. U. Patil, A. Fihri, A. H. Emwas and V. Polshettiwar, Chem. Sci., 2012, 3, 2224–2229.
5. Y. Belmabkhout, R. Serna-Guerrero and A. Sayari, Ind. Eng. Chem. Res., 2010, 49, 359–365.
6. A. Zhao, A. Samanta, P. Sarkar and R. Gupta, Ind. Eng. Chem. Res., 2013, 52, 6480–6491.
7. S. Hao, H. Chang, Q. Xiao, Y. Zhong and W. Zhu, J. Phys. Chem. C, 2011, 115, 12873–12882.
8. F. Su, C. Lu, S. C. Kuo and W. Zeng, Energy Fuels, 2010, 24, 1441–1448.
9. D. P. Bezerra, R. S. Oliveira, R. S. Vieira, C. L. Cavalcante Jr and D. C. S. Azevedo, Adsorption, 2011, 17, 235–246.
10. M. Keramati and A. A. Ghoreyshi, Phys. E, 2014, 57, 161–168.
11. C. Zhang, W. Song, G. Sun, L. Xie, J. Wang, K. Li, C. Sun, H. Liu, C. E. Snape and T. Drage, Energy Fuels, 2013, 27, 4818–4823.
12. W. J. Son, J. S. Choi and W. S. Ahn, Microporous Mesoporous Mater., 2008, 113, 31–40.
13. P. Li, B. Ge, S. Zhang, S. Chen, Q. Zhang and Y. Zhao, Langmuir, 2008, 24, 6567–6574.
14. Q. Wu, S. Chen and H. Liu, RSC Adv., 2014, 4, 27176–27183.
15. Y. Yang, H. Li, S. Chen, Y. Zhao and Q. Li, Langmuir, 2010, 26, 13897–13902.
16. L. Zhuang, S. Chen, R. Lin and X. Xu, J. Mater. Res., 2013, 28, 2881–2889.
17. Y. H. Kim, J. Polym. Sci., Part A: Polym. Chem., 1998, 36, 1685–1698.
18. A. Sunder, M. Krämer, R. Hanselmann, R. Mülhaupt and H. Frey, Angew. Chem., Int. Ed., 1999, 38, 3552–3555.
19. J. Wang, C. Q. Li, J. Li and J. Z. Yang, Sep. Sci. Technol., 2007, 42, 2111–2120.
20. M. Badaničová and V. Zeleňák, Monatsh. Chem., 2010, 141, 677–684.
21. H. Kassab, M. Maksoud, S. Aguado, M. Pera-Titus, B. Albela and L. Bonneviot, RSC Adv., 2012, 2, 2508–2516.
22. P. Lopez-Aranguren, L. F. Vega and C. Domingo, Chem. Commun., 2013, 49, 11776–11778.
23. J. H. Drese, S. Choi, R. P. Lively, W. J. Koros, D. J. Fauth, M. L. Gray and C. W. Jones, Adv. Funct. Mater., 2009, 19, 3821–3832.
24. K. Li, J. Jiang, F. Yan, S. Tian and X. Chen, Appl. Energy, 2014, 136, 750–755.
25. Z. Yao, H. Zhang, T. Dong and D. Yan, Journal of Fudan University, 1989, 28, 34–38.
26. W. Zhang, S. Wang, J. Ji, Y. Li, G. Zhang, F. Zhang and X. Fan, Nanoscale, 2013, 5, 6030–6033.
27. Y. Le, D. Guo, B. Cheng and J. Yu, Appl. Surf. Sci., 2013, 274, 110–116.
28. D. M. D’Alessandro, B. Smit and J. R. Long, Angew. Chem., Int. Ed., 2010, 49, 6058–6082.
29. S. Choi, J. H. Drese and C. W. Jones, ChemSusChem, 2009, 2, 796–854.

---