An imidazolium ionic compound-supported palladium complex as an efficient catalyst for Suzuki–Miyaura reactions in aqueous media

Abstract

A water soluble and efficient ionic compound-supported palladium complex was prepared. The structure and composition of the complex was characterized by FT-IR, NMR, ICP-OES and mass spectrometry. The complex exhibits high catalytic activity for Suzuki–Miyaura reactions in aqueous media. Moreover, the catalyst shows outstanding stability and reusability, and it can be recovered simply and effectively and reused six times without much activity decrease.

---

The palladium-catalyzed Suzuki–Miyaura (SM) cross-coupling reaction is an important formation of C–C bonds in synthetic organic chemistry.¹ It is a powerful tool for the construction of unsymmetrical biaryl derivatives that are structural elements of numerous natural products, agrochemicals, pharmaceuticals, and polymers.^2–5 Generally, the Suzuki–Miyaura (SM) reaction can be catalyzed by Pd–phosphine complexes.^6–9 However, the synthesis of phosphane ligands is difficult. Moreover, they are usually toxic, unrecoverable, and thermally unstable. Though a number of phosphine-free ligands for the Suzuki–Miyaura reaction had been reported to avoid the above mentioned disadvantages, the major drawbacks associated with them are the use of toxic solvents and the requirement for high temperature and long reaction time.^10–18 Therefore, the quest for the development of a robust, efficient and cost effective catalyst remains a great challenge.

From the standpoint of green chemistry, water is known to be a potential replacement for organic solvents.^19–21 The major advantages associated with water are its low cost, non-flammability and low toxicity, and there is no need to desiccate substrates prior to the reaction.²² Combining the use of water as a solvent with hydrophilic metal catalysts results in easy separation and recycling of the catalyst from the product. In general, it has been observed that neutral catalysts are not retained in water and are lost during product isolation. Therefore, an effective approach is to use metal complexes with charged groups to avoid catalyst leaching into the organic layer.²³

Presently, ionic liquid chemistry is getting into the epoch-making development, and has been attracting significant attention as alternative solvents and catalysts for chemical reactions because of low melting point, negligible vapour pressure, non-flammability, high conductivity, good solubility, and potential recyclability.²⁴ The functionalized ionic liquids formed through the incorporation of additional functional groups, has recently gained remarkable interest.²⁵ Some ligands supported to imidazolium-based ionic liquids have been reported, however, these needed to be dissolved in another ionic liquid for catalytic applications (Fig. 1).^26,27 SiO₂-supported imidazolium ionic liquid-immobilized PdEDTA has been reported, and is an efficient and reusable catalyst for the Suzuki reaction in water.²⁸ These catalysts can strongly retained, when the extraction of product was carried out.

Fig. 1 Ligand tethered to the imidazolium-based ionic liquid.

Metal complexes of Schiff bases are widely applied in the fields of synthesis and catalysis.²⁹ Although ionic liquid supported Schiff base with hexafluorophosphate as ligand can promote the Pd-catalyzed Suzuki–Miyaura coupling reaction, this reaction proceed in organic solvent and under reflux.³⁰ Based on the extensive research focused on the modification of Schiff base ligands and our previous work on the self-assembled catalyst,^31,32 we became interested in the synthesis of water soluble ionic compound-supported metal complexes. Herein, we report an efficient, water soluble and reusable, ionic compound-supported palladium complex catalyst, and investigate the catalytic properties for Suzuki–Miyaura coupling reactions in aqueous media (Fig. 2).

Fig. 2 Imidazolium ionic compound-supported palladium complex.

As illustrated in Scheme 1, the synthesis of the imidazolium ionic compound-supported Schiff base is a straightforward process starting from 5-chloromethylsalicylaldehyde, which are prepared from the concentrated hydrochloric acid, poly formaldehyde and salicylic aldehyde. 5-Chloromethylsalicylaldehyde was treated with 1-methylimidazole resulted in the imidazolium ionic compound-supported salicylic aldehyde 1 in almost quantitative yield. This intermediate can be used to implement the condensation of 1 with 4-methylaniline in ethanol to give the imidazolium ionic compound-supported Schiff base 2 in 97% yield. In the ¹H NMR spectrum of 2, a peak representing the azomethine proton (CH=N) appeared at δ 9.23, and the hydroxyl proton resonated at δ 13.20 (Fig. 3). The FT-IR spectrum of 2 exhibited absorption bands at 1630 cm⁻¹ (C=N stretch) and 3410 cm⁻¹ (OH stretch) (Fig. 4). Peaks corresponding to other protons were also shown. A peak at m/z 306.16 in the HR-MS spectrum of 2 corresponding to the [M − Cl]⁺ ion further confirmed the structure of 2.

Scheme 1 Preparation of the imidazolium ionic compound-supported palladium Schiff base complex.

Fig. 3 The comparison of ¹H NMR spectra between the ligand 2 and its palladium complex 3.

Fig. 4 Comparison between the FT-IR of imidazolium ionic compound-supported Schiff base and the imidazolium compound-supported palladium Schiff base complex.

Subsequently, we focus on the synthesis of the imidazolium ionic compound-supported palladium complex 3. Pd(OAc)₂ was added to the solution of the imidazolium ionic compound-supported Schiff base 2 in methanol, and then a pale yellow powder was obtained as the catalyst 3. After washing with solvent (methanol) thoroughly and drying, the solid remains pale yellow. However, the imidazolium ionic compound-supported palladium Schiff base complex 3 between the ligand and palladium acetate was synthesized without using any deprotonating base. The compound has a high melting point of above 200 °C.

The structure of the synthesized palladium complex 3 was characterized using different techniques such as IR, NMR and HR-MS. The ¹H NMR spectral difference between the ligands and their metal complexes is one of the important methods for characterization of complexes in the literature.^33,34 By the comparison of the ¹H NMR of imidazolium ionic compound-supported Schiff base 2and the resulting Pd complex 3, the shift of the peak for the azomethine proton to δ 8.45 and the absence of a peak (δ 13.20) for the hydroxyl proton suggested that complete complexation with the metal had occurred through the nitrogen of the C=N group and oxygen of the OH group (Fig. 3). This result is in accordance with our previous work.^31,32,35 In the FT-IR spectrum of 3, the C=N peak was shifted to 1624 cm⁻¹, and the peak in the region of 3410 cm⁻¹ (OH stretch) in the spectrum of 2 disappeared. The HR-MS spectrum of 3 showed a peak at m/z 753.1833 that corresponds to the [M + 2H − Cl]⁺ ion. In order to further confirm the structure of the catalyst, ICP-OES was used to analyzed for Pd content (found: 13.46% calcd: 13.50). All of these facts indicated that the ligand had coordinated with the metal ion through the nitrogen of the C=N group (Fig. 4).³⁶

Next, we concentrate on investigating the catalytic applications of 3 for SM reactions. In the beginning, we chose a model reaction between p-bromoanisole and phenylboronic acid. Since water solvent endows the reaction with green and safe properties,^37,38 we focused on the use of water to replace organic solvents, and the results are summarized in Table 1. It was found that carrying out the reaction at r.t. with 0.05 mol% of 3 in the presence of K₂CO₃ gave the coupled product in 75% yield (Table 1, entry 1). The use of an organic base, such as triethylamine and pyridine resulted in a poor yield of the coupled product (Table 1, entry 5, 6). After both of the amounts of the catalyst and the reaction time were optimized, the best condition is the combination of catalyst of 0.05 mol% and reaction time of 4 h. It is noteworthy that the yield of coupled product was found to be approximately 92% when EtOH was used as the reaction medium and K₂CO₃ was used as base (Table 1, entry 13). However, under the same condition, the yield of the coupled product is 78% in pure DMF within 12 h (Table 1, entry 14). When the mixture of water and DMF was used as solvent, the yield can significantly be improved. This fact shows that water play an important role in this reaction (Table 1, entry 15).

Table 1 Optimization of the reaction conditions for the Suzuki reaction catalyzed by 3^a

Entry	Solvent	Catalyst (Pd mol%)	Base	Time (h)	Yield^b (%)
a Reaction conditions: the catalyst 3, 1 mmol of p-bromoanisole, 1.2 mmol of phenylboronic acid, 2 mmol of base, 5 mL of solvent, r.t. in air.b Isolated yield.
1	H2O	0.05	K2CO3	2	75
2	H2O	0.05	KOAc	2	70
3	H2O	0.05	Na2CO3	2	54
4	H2O	0.05	NaOAc	2	40
5	H2O	0.05	TEA	2	34
6	H2O	0.05	Pyridine	2	20
7	H2O	0	K2CO3	2	0
8	H2O	0.01	K2CO3	2	61
9	H2O	0.08	K2CO3	2	85
10	H2O	0.10	K2CO3	2	92
11	H2O	0.10	K2CO3	4	98
12	H2O	0.10	K2CO3	5	98
13	EtOH	0.10	K2CO3	8	92
14	DMF	0.10	K2CO3	12	78
15	DMF/H2O (v/v, 1 : 1)	0.10	K2CO3	8	94

---

Based on the above investigations, the coupling of various aryl halides and arylboronic acids were investigated under the optimized conditions to investigate the general applicability of this protocol. As summarized in Table 2, it is found that the electronic effects of either coupling partner produced no significant impact on the reaction yields (Table 2, entries 1–10). These results suggested that the ionic compound-supported palladium complex 3 display the outstanding catalytic activities for the coupling reactions. However, ortho-substituted aryl bromides and arylboronic acids were coupled in very lower yields (Table 2, entry 11). This fact indicates that the steric hindrance of coupling partners has a definite influence on this coupling reaction. Even if the reaction time was extended, the yield was not improved. Highly sterically hindered di-ortho substituted biphenyls cannot be synthesized under the same conditions (Table 2, entry 12). In order to show the effect of the ligand in this reaction, we used Pd(OAc)₂ as catalyst instead of the Pd complex 3, but the target compound can be obtained in lower yield. The results showed that in the absence of the ligand, the reaction was sluggish and only 22% yield was obtained (Table 2, entry 13). From these studies we can predict that this catalyst can be applied to a wide range of aryl iodides and bromides.

Table 2 Suzuki cross-coupling catalyzed by 3^a

Entry	ArX	R	Time (h)	Yield^b (%)
a Reaction conditions: the catalyst 3 containing 0.10 mol% Pd, 1 mmol of aryl halide, 1.2 mmol arylboronic acid, 2 mmol of K2CO3, 5 mL of H2O, r.t. in air.b Isolated yield.c 1 mol% Pd was used.d Pd(OAc)2 used as catalyst.
1		4-Me	4	99
2		4-Cl	4	99
3		H	4	89
4		4-Me	4	97
5		4-Cl	4	96
6		2-Cl	4	96
7		4-Cl	4	88
8		H	4	98
9		2-Cl	4	97
10		4-Me	4	96
11		H	4	Trace
12		H	4	0
13		H	4	22^d
14		H	4	43(72^c)
15		4-Me	4	34(68^c)
16		H	2	98
17		4-Cl	2	96

---

Aryl chlorides are usually cheaper but less active than aryl bromides. Unfortunately, for coupling reactions of aryl chlorides, which are much more difficult to activate than aryl iodides and bromides, only low yields were obtained under the same conditions (Table 2, entries 14 and 15). When the reaction temperature was raised to 95 °C and the catalyst loading was increased to 1 mol%, the corresponding coupling products were obtained in moderate yields. However, the coupling of aryl iodides with phenylboronic acid can be completed within a shorter time (Table 2, entries 16 and 17).

In another investigation, the reusability of the catalyst was also evaluated in the coupling of methyl 4-bromobenzoate with phenylboronic acid. In detail, after the first run of the reaction was completed, ethyl acetate was directly added to extract the target compound, and the aqueous solution containing the catalyst was reused in the next cycle of the coupling reactions by mixing them with a new substrate and base. It is noteworthy that the catalytic activity was maintained at least until the six uses without much loss in activity (Fig. 5). To avoid deactivation of catalyst owing to the accumulation of potassium bromide, which makes the solution saturated, glycerol was used to wash potassium bromide after the second run was completed and the aqueous solution was concentrated under vacuum. The gradual decrease in the yield of the product during recycling could be mainly attributed to the leaching of the catalyst during extraction of the catalyst. The leaching of the catalyst could be detected by atomic absorption spectrometry (AAS) (Table 3). Pd complex species in the extraction was still the catalyst 3 confirmed by HR-MS. In order to confirm the stability of the Pd complex, the ¹H NMR of the spent catalyst after six run were check. It was found that the used Pd complex catalyst can retain its structure, indicating a good stability and recyclable applicability for the Suzuki cross-coupling reaction in aqueous media (see ESI†).

Fig. 5 Reusability of the imidazolium ionic compound-supported palladium complex in the Suzuki coupling of methyl 4-bromobenzoate with phenylboronic acid over 6 runs.

Table 3 The leaching amount of the catalyst during each recycles

Run	1	2	3	4	5	6
The leaching amount (μg)	9	10	8	11	7	8

---

Conclusions

In summary, we have synthesized a novel imidazolium compound-supported palladium Schiff base complex, and explored its catalytic activity for the Suzuki reactions in aqueous media. The catalyst showed high activity and excellent yields for these reactions. Moreover, the catalyst shows outstanding stability and reusability, and reused six times without much activity decrease. Further investigation of the detailed mechanism and application of this chemistry are in progress in our lab.

Experimental section

General information

All reagents were commercially available and used as supplied. Unless otherwise stated, analytical grade solvents and commercially available reagents were used as received. Thin layer chromatography (TLC) employed glass 0.20 mm silica gel plates. Flash chromatography columns were packed with 200–300 mesh silica gel. All new compounds were characterized by IR, ¹H NMR, ¹³C NMR and HR-MS. The IR spectra were run on a Nicolete spectrometer (KBr). The ¹H NMR and ¹³C NMR spectra were recorded on a BRUKER AVANCE III 400 MHz spectrometer. The chemical shifts (δ) were given in parts per million relative to an internal standard tetramethylsilane, high resolution mass spectra (HR-MS) were measured with Bruker MICROTOF-Q III instrument and accurate masses were reported for the molecular ion. Inductively coupled plasma atomic emission spectrometer (ICP-OES) was run on PerkinElmer Optima 7300 DV. Melting points were determined on a Perkin-Elmer differential scanning calorimeter and the thermometer was uncorrected.

The synthesis of 5-chloromethylsalicylaldehyde

To a 100 mL round-bottom flask equipped with a magnetic stirring bar was concentrated hydrochloric acid (80 mL), poly formaldehyde (4.7 g) and salicylic aldehyde (12 g). The mixture was stirred for approximately 0.5 h at room temperature. POCl₃ was slowly added within 1 h, and the mixture was stirred for 20 h to obtain the solid (16 g, 98%). The analytical sample was afforded by the recrystallization from petroleum ether. Mp: 264–268 °C; ¹H NMR (400 MHz, DMSO) δ 10.92 (s, 1H), 10.27 (s, 1H), 7.72 (d, J = 2.3 Hz, 1H), 7.58 (dd, J = 8.5, 2.4 Hz, 1H), 7.02 (d, J = 8.5 Hz, 1H), 4.75 (s, 2H).

The synthesis of the ionic compound-supported salicylic aldehyde

A mixture of 1-methylimidazole (5 mmol) and 2 (5 mol) was refluxed in ethanol for 24 h, followed by the vacuum distillation to give a pale yellow solid, which was washed with ethyl acetate to give the ionic compound-supported salicylic aldehyde 2 in almost quantitative yield. Mp: >250 °C; ¹H NMR (400 MHz, D₂O) δ 9.86 (d, J = 2.4 Hz, 1H), 8.68 (s, 1H), 7.69 (s, 1H), 7.58–7.48 (m, 1H), 7.44–7.30 (m, 2H), 6.97 (dd, J = 8.6, 2.5 Hz, 1H), 5.30 (d, J = 2.1 Hz, 2H), 3.81 (d, J = 2.6 Hz, 3H). ¹³C NMR (101 MHz, D₂O) δ 196.54, 160.27, 137.45, 136.06, 133.45, 125.63, 124.03, 122.16, 121.16, 118.19, 51.79, 35.90. HR-MS: m/z calcd for C₁₂H₁₃N₂O₂ [M − Cl]⁺: 217.0983 found: 217.0985.

The synthesis of the ionic compound-supported Schiff base

A round-bottom flask equipped with a stirring bar was charged with the ionic compound-supported salicylic aldehyde (2.0 mmol) and 4-methylaniline (2.2 mmol) in 10 mL of ethanol. The mixture was refluxed for 6 h, and cooled to room temperature to form the solid product. The crude product was purified by recrystallization from ethanol to obtain the analytical sample (yield, 97%). Mp: >300 °C; ¹H NMR (400 MHz, DMSO) δ 13.20 (s, 1H), 9.26 (d, J = 18.5 Hz, 1H), 8.93 (s, 1H), 7.89–7.67 (m, 3H), 7.50 (d, J = 6.4 Hz, 1H), 7.31 (dd, J = 19.7, 8.2 Hz, 4H), 7.04 (d, J = 8.5 Hz, 1H), 5.40 (s, 2H), 3.86 (s, 3H), 2.35 (s, 3H); ¹³C NMR (101 MHz, DMSO) δ 161.45, 160.85, 146.09, 137.01, 133.87, 132.60, 130.41, 125.88, 124.36, 122.63, 121.63, 120.14, 117.70, 51.66, 36.31, 21.07. HR-MS: m/z calcd for C₁₉H₂₀N₃O [M − Cl]⁺: 306.1612, found: 306.1640.

The synthesis of the ionic compound-supported palladium complex

A mixture of the ionic compound-supported Schiff base (2.0 mmol) and palladium acetate (224 mg, 1.0 mmol) was added in methanol (5.0 mL) and refluxed for 6 h until the product had completely precipitated. After cooling, the palladium complex was separated by filtration and recrystallized from ethanol (5 mL) (yield, 97%). Mp: 264–268 °C; ¹H NMR (400 MHz, DMSO) δ 9.10 (s, 1H), 8.45 (s, 1H), 7.72 (dt, J = 22.7, 1.8 Hz, 2H), 7.49 (dd, J = 14.0, 5.2 Hz, 3H), 7.35–7.22 (m, 3H), 6.65 (d, J = 8.8 Hz, 1H), 5.23 (s, 2H), 3.84 (s, 3H), 2.34 (s, 3H). ¹³C NMR (101 MHz, DMSO) δ 171.46, 167.42, 147.09, 138.25, 136.57, 135.65, 130.15, 124.30, 123.87, 122.52, 122.16, 119.09, 52.27, 36.30, 20.99. HR-MS: m/z calcd for C₃₈H₃₈Cl₂N₆O₂Pd [M + 2 − Cl]⁺: 753.1936, found: 753.1833. Elemental analysis calcd (%) for C₃₈H₃₈Cl₂N₆O₂Pd: C, 57.92; H, 4.86; found: C 57.85, H 4.92.

General experimental procedures for Suzuki–Miyaura couplings

In a typical experiment, the ionic compound-supported palladium complex (0.001 mmol of Pd) was added to a mixture of aryl halide (1.0 mmol), arylboronic acid (1.2 mmol) and K₂CO₃ (2.0 mmol) in water (5.0 mL), and the reaction mixture was stirred at r.t. The progress of the reaction was monitored by means of TLC. After the completion of the reaction, the reaction mixture was extracted with a mixture of ethyl acetate (3 × 5 mL). The organic layer was dried with anhydrous Na₂SO₄, and then the solvent was evaporated under reduced pressure. The resulting residue was purified by column chromatography over silica gel (mesh 200–300), using n-hexane–ethyl acetate as an eluent, to give the desired product. All of the products are known compounds, and their ¹H NMR data were identical to those reported in literature.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (No. 21462016, 21262010), Natural Science Foundation of Gansu Province (No. 1506RJZG049) and Students science and technology innovation project of Hexi University (No. 106).

Notes and references

1. L. Yin and J. Liebscher, Chem. Rev., 2007, 107, 133.
2. J. H. Clark, D. J. Macquarrie and S. J. Tavernier, Dalton Trans., 2006, 4297.
3. K. C. Nicolaou, C. N. C. Boddy, S. Brase and N. Winssinger, Angew. Chem., Int. Ed., 1999, 38, 2096.
4. O. Baudoin, M. Cesario, D. Guenard and F. Gueritte, J. Org. Chem., 2002, 67, 1199.
5. L. Pu, Chem. Rev., 1998, 98, 2405.
6. H. Firouzabadi, N. Iranpoor and M. Gholinejad, J. Organomet. Chem., 2010, 695, 2093.
7. R. Martin and S. L. Buchwald, Acc. Chem. Res., 2008, 41, 1461.
8. B. Punji, J. T. Mague and M. S. Balakrishna, Dalton Trans., 2006, 1322.
9. S. Doherty, J. G. Knight, T. H. Scanlan, M. R. J. Elsegood and W. Clegg, J. Organomet. Chem., 2002, 650, 231.
10. J. H. Kim, J. W. Kim, M. Shokouhimehr and Y. S. Lee, J. Org. Chem., 2005, 70, 6714.
11. F. Godoy, C. Segarra, M. Poyatos and E. Peris, Organometallics, 2011, 30, 684.
12. G. A. Grasa, A. C. Hillier and S. P. Nolan, Org. Lett., 2001, 3, 1077.
13. T. Mino, Y. Shirae, M. Sakamoto and T. Fujita, J. Org. Chem., 2005, 70, 2191.
14. J. M. Lu, H. Ma, S. S. Li, D. Ma and L. X. Shao, Tetrahedron, 2010, 66, 5185.
15. B. M. Manuel, H. O. Simon, A. T. Ruben, V. M. Jesus and M. M. David, Inorg. Chim. Acta, 2010, 363, 1222.
16. I. D. Kostas, F. J. Andreadaki, K. D. Dimitra, C. Prentjasb and M. A. Demertzisb, Tetrahedron Lett., 2005, 46, 1967.
17. S. A. Patil, C. M. Weng, P. C. Huang and F. E. Hong, Tetrahedron, 2009, 65, 2889.
18. M. L. Kantam, P. Srinivas, J. Yadav, P. R. Likhar and S. Bhargava, J. Org. Chem., 2009, 74, 4882.
19. A. K. Nezhad and F. Panahi, Green Chem., 2011, 13, 2408.
20. A. Modak, J. Mondal, M. Sasidharanb and A. Bhaumik, Green Chem., 2011, 13, 1317.
21. S.-L. Mao, Y. Sun, G.-A. Yu, C. Zhao, Z.-J. Han, J. Yuan, X. Zhu, Q. Yang and S.-H. Liu, Org. Biomol. Chem., 2012, 10, 9410.
22. R. N. Butler and A. G. Coyne, Chem. Rev., 2010, 110, 6302.
23. K. H. Shaughnessy, Chem. Rev., 2009, 109, 643.
24. N. V. Plechkova and K. R. Seddon, Chem. Soc. Rev., 2008, 37, 123.
25. J. P. Hallett and T. Welton, Chem. Rev., 2011, 111, 3508.
26. N. Audic, H. Clavier, M. Mauduit and J.-C. Guillemin, J. Am. Chem. Soc., 2003, 125, 9248.
27. S. Doherty, P. Goodrich, C. Hardacre, V. Parvulescu and C. Paun, Adv. Synth. Catal., 2008, 350, 295.
28. J.-F. Wei, J. Jiao, J.-J. Feng, J. Lv, X.-R. Zhang, X.-Y. Shi and Z.-G. Chen, J. Org. Chem., 2009, 74, 6283.
29. K. C. Gupta and A. K. Sutar, Coord. Chem. Rev., 2008, 252, 1420.
30. B. Li, Y.-Q. Li and J. Zheng, ARKIVOC, 2010, 9, 163.
31. J. Wang, Y. Zong, X. Wang, Y. Hu, G. Yue and Y. Pan, Green Chem., 2016, 18, 967.
32. J. Wang, Y. Zong, G. Yue, Y. Hu and X. Wang, RSC Adv., 2015, 5, 76285.
33. A. Sarkar, S. Bhattacharyya, S. K. Dey, S. Karmakar and A. Mukherjee, New J. Chem., 2014, 38, 817.
34. Y. M. Lee, E. S. Kim, H. J. Kim, H. J. Choi, Y. I. Kim, S. K. Kang and S. N. Choi, Dalton Trans., 2009, 126.
35. J. Wang, Y. Zong, S. Wei and Y. Pan, Appl. Organomet. Chem., 2014, 28, 351.
36. A. Pui, C. Policar and J.-P. Mahy, Inorg. Chim. Acta, 2007, 360, 2139.
37. A. K. Nezhad and F. Panahi, Green Chem., 2011, 13, 2408.
38. A. Modak, J. Mondal, M. Sasidharanb and A. Bhaumik, Green Chem., 2011, 13, 1317.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra19850c

‡ Xicun Wang is first communication author.

---