Homogeneous benzoylation of cellulose in 1-allyl-3-methylimidazolium chloride: Hammett correlation, mechanism and regioselectivity

Abstract

Homogeneous benzoylation of cellulose with a series of substituted benzoyl chlorides, in which substituents varied from electron donating to electron withdrawing groups, was investigated in an ionic liquid 1-allyl-3-methylimidazolium chloride (AmimCl). The electronic effect of substituents had a considerable effect on the reaction. A plot of Hammett parameters of substituents vs. degree of substitution (DS) of resultant cellulose esters exhibited a V-shaped graph: a negative slope for negative Hammett parameters, and a positive slope for positive Hammett parameters. Such a Hammett plot indicated the mechanism of cellulose benzoylation: the reaction underwent a unimolecular ionization mechanism with a carbocationic intermediate when benzoyl chlorides were with electron-donating substituents, and a bimolecular addition–elimination mechanism with a tetrahedral intermediate when benzoyl chlorides were with electron-withdrawing substituents. In addition, ¹³C-NMR analysis showed that most of the substituted benzoyl chlorides exclusively preferred the primary hydroxyl at the C-6 position. This unusually high regioselectivity was attributed to the synergistic effect of appropriate reaction rate, moderate steric effect and reaction mechanism of benzoylation.

---

Introduction

Esterification endows unmodified cellulose with some new performances, and as a result greatly enriches the application range of cellulosic materials.¹ However, in general, it is difficult to conduct the homogeneous esterification of cellulose, because unmodified cellulose can neither be molten nor is soluble in conventional solvents. In the recent decade, the advent of some ionic liquids (ILs), which can break the hydrogen-bonding network in cellulose to dissolve cellulose excellently, provides a new and versatile platform for the utilization and homogeneous derivatization of cellulose resources.^2–4 A variety of cellulose esters, including widely-used cellulose acetate, water-soluble cellulose sulfate, and chiral-recognizable cellulose phenylcarbamates and benzoates, has been efficiently and homogeneously prepared in cellulose/ILs solutions.^5–8 More interestingly, by using ILs as the medium, some novel cellulose esters, such as pyro-pheophorbide a ester of cellulose, 3,6,9-trioxadecanoic acid ester of cellulose and cellulose dehydroabietate, were also synthesized.^9–11 The homogeneous esterification of cellulose in ILs not only provides a possibility to control degree of substitution (DS) and structure of cellulose esters, but also gives a high efficiency. However, so far, there is almost no available information about the reaction mechanism and reaction process of the homogeneous esterification of cellulose at the molecular level. In addition, except the dissolving effect of ILs, only few information on the role of other factors on the cellulose esterification in ILs is available. Recently, Zhang et al. indicated that Brönsted acidic ILs can be used as a novel catalyst for the efficient acetylation of cellulose to one-step synthesize cellulose acetate with different DS and solubility.^12,13 Such comprehensive information is indispensable to reliably control and efficiently optimize the conditions for synthesis of cellulose esters, and helps practicably prepare specific cellulose esters with well-designed structure and optimal performances.¹⁴ In physical organic chemistry, investigating the effect of substituents on the aromatic ring on the reaction rate is one of the most important methods to clarify the reaction process of esterification, alcoholysis, hydrolysis, etc.^15–17 Further, combining the experimental result with Hammett parameter of the substituent, a Hammett plot can be obtained to understand how the reaction mechanism varies as a function of the electronic change induced by different substituents.^18–20

Regioselectively substituted cellulose esters, one kind of the most important derivatives of cellulose with defined structures, show superior, versatile and adjustable physicochemical properties with different positions of functional groups around the anhydroglucose unit (AGU) in cellulose. More importantly, they are crucial to understand the structure–property relationship of cellulose derivatives.²¹ For example, compared with randomly substituted cellulose acetate propionates (CAPs), regioselectively substituted CAPs had improved optical properties.²² The chiral separation ability of regioselectively substituted cellulose esters was quite different from the corresponding non-regioselectively substituted ones with the same substituent. Some racemates were found to be more efficiently separated on regioselectively substituted ones.^23–25 However, the regioselective synthesis of cellulose esters is a complex and challenging work. The general strategy is the utilization of trityl chloride or thexyldimethylchlorosilane as the protecting group firstly, followed by esterification of cellulose, subsequent removal of the protecting group and finally chemical modification of deprotected hydroxyls. This includes a long reaction process, extra synthesis steps, unexpected side reactions, lost yield and so on.^26–29 Therefore, taking the advantage of inherent reactivity difference of cellulose hydroxyls at different positions, the directly regioselective reaction of cellulose is fascinating and exciting. Though many efforts were made to accomplish the directly regiocontrolled synthesis of cellulose esters, only limited reactions were found to achieve high regioselectivity.^30,31 In our earlier work, we demonstrated that the reaction between cellulose and benzoyl chloride in 1-allyl-3-methylimidazolium chloride (AmimCl) exhibited exclusively C-6 regioselectivity, but still there exists little knowledge about the reaction mechanism, influencing factors and general applicability.⁸

In this study, homogeneous benzoylation of cellulose was investigated systematically with ten kinds of substituted benzoyl chloride in AmimCl. Combining with Hammett parameters, the reaction mechanism of cellulose benzoylation was probed. In addition, it was surprising that many substituted benzoyl chlorides exhibited unusually high C-6 regioselectivity. The fundamental reason and applicability for the high regioselection were investigated in detail.

Experimental section

Materials

Microcrystalline cellulose (MCC, Avicel PH-101, DP = 220) with degree of polymerization (DP) of 220 was dried in vacuum at 80 °C for 24 h prior to use. Ionic liquid 1-allyl-3-methylimidazolium chloride (AmimCl) was synthesized according to our previous work.³² Benzoyl chloride, o-toluoyl chloride, m-toluoyl chloride, p-toluoyl chloride, 3,5-dimethyl benzoyl chloride, 4-tert-butyl benzoyl chloride, 4-chlorobenzoyl chloride and 4-nitrobenzoyl chloride were purchased from Alfa-Aesar. 2-Methoxybenzoyl chloride and 4-methoxybenzoyl chloride were provided by Sigma-Aldrich. All other chemicals were obtained from Beijing Chemical Reagent Company, China. All reagents were analytical grade and used without further purification.

Characterization

NMR spectra were acquired on Bruker NMR spectrometers with 16 scans for ¹H-NMR (Bruker AV400), and 10 000–18 000 scans for ¹³C-NMR (Bruker AV600 or AV400) at room temperature in DMSO-d₆. A few drops of trifluoroacetic acid-d₁ were added to shift active hydrogens to lower field area for ¹H-NMR test.

FT-IR spectra were recorded with a Bruker Tensor 27 FT-IR spectrometer from 400 to 4000 cm⁻¹.

The degree of substitution (DS) of cellulose esters was calculated directly from ¹H-NMR by the following equation³³

DS = 7I_phenyl/nI_AGU (1)

where, I_phenyl, peak integral of phenyl protons; I_AGU, peak integral of protons of anhydroglucose unit; n, the number of protons on the benzene ring of substituted benzoyl chlorides.

Preparation of cellulose/AmimCl solution

For a typical preparation of cellulose solution in AmimCl, 0.4 g cellulose was added to 9.6 g AmimCl in a three-necked flask, and the mixture of cellulose/AmimCl was mechanically stirred at 80 °C for 2 h, yielding a clear and viscous solution with 4.0 wt% of cellulose concentration.³²

Synthesis of cellulose substituted benzoates

A variety of cellulose benzoates were synthesized in ionic liquid AmimCl according to Scheme 1. The experimental details are listed in Table 1. A typical synthesis process is shown as follow.

Scheme 1 Synthesis route of various cellulose benzoates.

Table 1 Conditions and results of homogeneous benzoylation of cellulose with various substituted benzoyl chlorides in AmimCl

Samples	Substituted benzoyl chloride	Molar ratio of substituted benzoyl chloride/AGU	T (°C)	DS
				Reaction time (h)
				0.5	1.0	2.0
a Pyridine was added as the acid scavenger.b 25 wt% DMF was added into the cellulose/AmimCl solution to replace the same amount of AmimCl, and the cellulose concentration was maintained at 4.0 wt%.
CB1		3 : 1	60	0.90	1.05	1.15
CB2		3 : 1	80	1.08	1.19	1.35
CB3		5 : 1	80	1.26	1.39	1.76
CoMB1		3 : 1	60	1.31	1.49	1.67
CoMB2		3 : 1	80	1.68	1.89	2.05
CoMB3		5 : 1	80	1.97	2.08	2.36
CmMB1		3 : 1	60	0.89	1.04	1.11
CmMB2		3 : 1	80	1.14	1.28	1.47
CmMB3		5 : 1	80	1.24	1.55	1.96
CpMB		3 : 1	80	1.16	1.34	1.51
CdMB^a		3 : 1	80	0.61	0.98	1.28
CpBB^a		3 : 1	80	0.49	0.61	0.83
CpCB		3 : 1	80	1.18	1.35	1.51
CpNB1		3 : 1	80	1.52	1.71	1.92
CpNB2^b		3 : 1	80	0.91	1.03	1.13
CoTB		3 : 1	80	2.15	2.36	2.49
CpTB1		3 : 1	80	1.31	1.59	1.76
CpTB2^b		3 : 1	80	0.57	0.68	0.77

---

Cellulose (0.4 g, 2.469 mmol) was dissolved in AmimCl (9.6 g). Benzoyl chloride (1.041 g, 7.407 mmol, 3 mol mol⁻¹ AGU) was then added to the cellulose/AmimCl solution at 80 °C. After 0.5 h, the homogeneous mixture was poured slowly into 100 mL methanol to obtain a white precipitate, which was washed three times with methanol, filtered and dried. Then, the obtained solid was redissolved in DMSO, precipitated again and thoroughly washed with methanol (200 mL) to get purified product. Finally, the product was dried under vacuum at 80 °C; yield: 0.51 g (77%).

Results and discussion

Homogeneous benzoylation of cellulose in AmimCl

The defined structure of obtained cellulose benzoates has been characterized clearly by ¹H-NMR and ¹³C-NMR spectroscopy. Typical ¹H-NMR and ¹³C-NMR spectra of cellulose benzoate with DS of 0.99 are shown in Fig. 1. In ¹H-NMR spectrum, protons of cellulose backbone appear at 2.8–5.3 ppm. New peaks appeared at 7.0–8.0 ppm are the characteristic peaks of the benzoate group. In ¹³C-NMR spectrum, the peak at 165.9 ppm is attributed to the carbonyl carbon (C-7), peaks at 128–134 ppm to phenyl carbons, and peaks at 63–103 ppm to cellulose backbone carbons (C-1, C-2, C-3, C-4, C-5 and C-6). These phenomena proved that the benzoylation of cellulose in AmimCl was achieved successfully.

Fig. 1 (A) ¹H-NMR and (B) ¹³C-NMR spectra of the cellulose benzoate with DS of 0.99.

The effect of reaction parameters, including temperature, reaction time, and molar ratio of substituted benzoyl chloride/anhydroglucose unit (AGU), on the benzoylation reaction was investigated. The result is summarized in Table 1. The reaction was found to be accelerated through raising the temperature. For example, during the benzoylation of cellulose with o-toluoyl chloride, under the condition of o-toluoyl chloride/AGU molar ratio of 3 : 1, a DS of 1.67 was achieved after 2 h at 60 °C (CoMB1), whereas, as the temperature increased to 80 °C, a similar DS (DS = 1.68) was obtained after only 0.5 h (CoMB2). Moreover, the DS of products increased as the reaction time prolonged. During the reaction between cellulose and benzoyl chloride, the increase of reaction time from 0.5 h to 1 h and 2 h led to an increase in the DS from 1.26 to 1.39 and 1.76, respectively (CB3). In addition, as the molar ratio of substituted benzoyl chloride/AGU increased, the DS of product increased correspondingly. During the benzoylation of cellulose with m-toluoyl chloride, when the molar ratio increased from 3 : 1 to 5 : 1, the DS correspondingly increased from 1.47 to 1.96 at 80 °C for 2 h (CmMB2 to CmMB3). Obviously, the DS value of cellulose benzoates can be controlled by simply varying the reaction time, temperature and molar ratio of substituted benzoyl chloride/AGU.

From Table 1, it can also be seen that the substituent on the benzene ring has a great impact on the reactivity of substituted benzoyl chlorides. For most of substituted benzoyl chlorides, either electron-donating or electron-withdrawing substituents accelerated the benzoylation reaction. For example, the reactivity of 4-nitrobenzoyl chloride and 2-methoxybenzoyl chloride was much higher than that of benzoyl chloride. With the same reaction parameters (molar ratio of acylation reagent/AGU = 3 : 1, temperature = 80 °C, time = 1 h), the DS was 1.19, when benzoyl chloride was used as the acylation reagent (CB2); while the DS increased to 1.71, when 4-nitrobenzoyl chloride as the acylation reagent (CpNB1); and the DS further increased to 2.36, when 2-methoxybenzoyl chloride as the acylation reagent (CoTB). In addition, the high steric effect of substituents, i.e. tert-butyl group, was unfavorable for the esterification reaction. When cellulose reacted with 4-tert-butyl benzoyl chloride or 3,5-dimethyl benzoyl chloride without any catalysts, the reaction system turned to brown or black color within 0.5 h because of the cellulose degradation, and, finally, little amount of the product could be precipitated. Therefore, for the esterification of cellulose with 4-tert-butyl benzoyl chloride and 3,5-dimethyl benzoyl chloride, pyridine was needed to use as the catalyst and acid scavenger. Even then, the DS of the product was much lower than that of other cellulose substituted benzoates obtained under the same reaction conditions. In brief, the reactivity of substituted benzoyl chlorides is in the following order:

Hammett correlation and mechanism of cellulose benzoylation

The Hammett correlation, which is of widespread importance in physical organic chemistry, is a general method for examining change in the charge during a reaction.^16,32,34 Combining Hammett constants (σ) of substituents with DS of cellulose benzoates, a plot of σ vs. DS is obtained and shown in Fig. 2. The Hammett plot is a V-shaped graph composed of two intersecting straight lines with a negative slope for negative Hammett parameters and a positive slope for positive Hammett parameters. The unsubstituted benzoyl chloride (Hammett parameter = 0) is the turning point. Therefore, higher the absolute value of Hammett parameter, higher the reactivity of the corresponding substituted benzoyl chlorides is. In other words, the reactivity of substituted benzoyl chlorides is enhanced by both electron-withdrawing substituents (NO₂, Cl) and electron-donating substituents (CH₃, OCH₃) on the aromatic ring. Based on the above Hammett correlation, it is easy to predict the reaction rate of the unperformed benzoylation of cellulose by Hammett parameters of substituents.

Fig. 2 DS of cellulose benzoates plotted against Hammett substituent constants σ.

The Hammett correlation of cellulose benzoylation is consistent with that of the solvolysis of substituted benzoyl chlorides, of which the mechanism experiences a transfer from dissociative mechanism to associative mechanism as the sharp change of the slope in Hammett plot.^35–38 Likewise, for cellulose benzoylation in our work, V-shaped Hammett plot revealed the mechanism of cellulose benzoylation: the benzoylation reaction underwent a unimolecular ionization mechanism with a carbocationic intermediate when benzoyl chlorides were with electron-donating substituents, and a bimolecular addition–elimination mechanism with a tetrahedral intermediate when benzoyl chlorides with electron-withdrawing substituents. To be more specific, for negative Hammett parameters (electron-donating substituents), the smaller value of Hammett parameters, namely the stronger electron-donating ability of substituents, the higher reactivity of substituted benzoyl chlorides is. The electron-donating ability is beneficial to the stability of the carbocation intermediate, thus benzoyl chlorides with electron-donating substituents react with cellulose undergoing a unimolecular ionization mechanism with a carbocationic intermediate. For positive Hammett parameters (electron-withdrawing substituents), the higher value of Hammett parameters, namely the stronger electron-withdrawing ability of substituents, the higher reactivity of substituted benzoyl chlorides is. The electron-withdrawing ability is favorable to the increase of the electrophilicity of the carbonyl carbon. Thus, the benzoyl chlorides with electron-withdrawing substituents react with cellulose undergoing with a bimolecular addition–elimination pathway through a tetrahedral intermediate.

Because Hammett parameters are the response of only the electronic effect of substitutes, the ortho-substituents, which involve both electronic effect and steric effect, do not have the Hammett parameter. As a result, o-methoxybenzoyl chloride and o-toluoyl chloride are not shown in the Hammett plot (see Fig. 1). Despite these, the above reaction mechanism should be suitable for them. According to the reaction mechanism, substituted benzoyl chlorides with electron-donating groups react with cellulose undergoing a unimolecular ionization mechanism. The rate-determining step is the carbocation formation. Owing to the inductive effect and steric effect, the electron-donating group at ortho-position is more beneficial to the carbocation formation and stability than those at meta- and para-position, thus the formation rate of the intermediate and the reaction rate are faster. Namely, o-methoxybenzoyl chloride and o-toluoyl chloride should give a higher reactivity than the corresponding meta- and para-substituted benzoyl chlorides. The conjecture is completely consistent with the synthesis result, indicating that the reaction mechanism is reasonable to all of benzoyl chlorides.

Either ionization or addition–elimination mechanism, the transition state is higher polar than the original reagent. Thus, the high polarity of reaction media will accelerate the benzoylation reaction. Previous reports have shown that, in ILs, the aprotic polar co-solvents, such as DMF, DMSO and DMAc, can facilitate the dissolution of cellulose, reduce the viscosity of the solution, and change the polarity of the solvent system.^39–41 In cellulose benzoylation, DMF was chosen as the co-solvent to investigate the effect of reaction media on the reaction rate. The addition of DMF significantly decreased the reaction rate. When DMF/AmimCl (25 wt% DMF) was used as the reaction media, during the benzoylation of cellulose with 4-nitrobenzoyl chloride, the DS decreased from 1.62 to 0.91 at 80 °C for 0.5 h (CpNB1 to CpNB2). A similar phenomenon was observed during the reaction of cellulose with 4-methoxybenzoyl chloride. When DMF/AmimCl (25 wt% DMF) was used as the reaction media, the DS decreased from 1.31 to 0.57 at 80 °C for 0.5 h (CpTB1 to CpTB2). Therefore, AmimCl acts as a higher polarity solvent than DMF in cellulose benzoylation.

Regioselectivity of cellulose benzoylation in AmimCl

By simply controlling the reaction condition, a series of cellulose monobenzoates (DS around 1.0) were synthesized in AmimCl. In ¹³C-NMR spectra of cellulose benzoate, 3-methylbenzoate, 4-methylbenzoate, 3,5-dimethyl benzoate, 4-chlorobenzoate and 4-nitrobenzoate with their DS values around 1.0 (see Fig. 3), peaks of C-7, C-1, C-4 and C-6 are very sharp and well resolved. In addition, the peak of C-6 influenced by esterification of the hydroxyl at C-6 position appears at 63–65 ppm, exhibiting a downfield shift of about 3–5 ppm compared with the corresponding carbon in unmodified cellulose. The C-1 and C-4 signals appear at 102 ppm and 79 ppm, respectively, which are in accordance with the corresponding carbon in unmodified cellulose absolutely. These phenomena proved that, for most of the cellulose monobenzoates, including cellulose benzoate, 3-methylbenzoate, 4-methylbenzoate, 3,5-dimethyl benzoate, 4-chlorobenzoate and 4-nitrobenzoate, the benzoylation exclusively preferred at the primary hydroxyl group of cellulose, while the substitution at the C-2 and C-3 was few, even nonexistent.

Fig. 3 ¹³C-NMR spectra of cellulose benzoate (DS = 0.99), cellulose 3,5-dimethyl benzoate (DS = 0.94), cellulose 3-toluoyl benzoate (DS = 0.98), cellulose 4-toluoyl benzoate (DS = 1.03), cellulose 4-chlorobenzoate (DS = 1.11) and cellulose 4-nitrobenzoate (DS = 1.06).

According to ¹³C-NMR spectra, the benzoylation reaction of cellulose in AmimCl provides >90% selectivity at the C-6 position. Until today, there are few reports on the direct esterification with such unusually high C-6 regioselectivity. Edgar et al. attempted to accomplish directly the regioselective esterification of cellulose by reaction with several bulky acylation reagents, including pivaloyl chloride, adamantoyl chloride, and 2,4,6-trimethylbenzoyl chloride, in all three important classes of cellulose reaction solvents (DMAc/LiCl, DMSO/TBAF, ionic liquids).³⁰ However, the result indicated that although C-6 substitution was consistently preferred, C-2 and C-3 substitutions were always observed at significant levels even for the total DS below 1.0. Several other acylation reagents, such as trifluoroacetic anhydride, formic acid and p-tosyl chloride, have been employed also, but the best result in these studies achieved only modest positional selectivity, not high regioselectivity.^31,42–44

In general, only the steric hindrance of esterification reagents is considered to achieve high regioselectivity. However, we believe that appropriate reaction rate and mechanism are also important for the regioselectivity, via monitoring the minor discrepancy in the occurring regioselectivity with the change of substitutions in benzoyl chlorides and reaction conditions. Among all of benzoyl chlorides, unsubstituted benzoyl chloride and 4-chlorobenzoyl chloride afforded similar reactivity and steric hindrance, and gave the best regioselectivity (almost a complete C-6 selectivity). As the reactivity of benzoylation reagents increased, the regioselectivity slightly decreased. In ¹³C-NMR spectra of cellulose 4-nitrobenzoate, 3-methylbenzoate and 4-methylbenzoate, some tiny peaks appeared at the edge of the signals of C-7, C-1 or/and C-4, indicating that a slight amount of 2-OH and/or 3-OH were substituted. In addition, pyridine, which can be used as the catalyst and acid scavenger to accelerate the esterification reaction rate, led to the decrease of the regioselectivity, as shown in Fig. 4. When pyridine was used, C-7 peak broadened, and some tiny peaks appeared at the edge of signals of C-7, C-1 and C-4. It is reasonable that there is a trade-off relationship between regioselectivity and reaction rate. However, too slow reaction rate, i.e. the synthesis of 3,5-dimethyl benzoate, is not beneficial to the regioselectivity, owing to the cellulose degradation. Therefore, the appropriate reaction rate, determined by the reactivity and steric hindrance of esterification reagents and affected by the reaction conditions, is important to the regioselectivity. The reason for the low regioselectivity of cellulose reacted with pivaloyl chloride, adamantoyl chloride and 2,4,6-trimethylbenzoyl chloride may be the inappropriate reaction rate. In addition, the reaction mechanism had a certain effect on the regioselectivity also. For example, although 4-chlorobenzoyl chloride had the similar reactivity and steric effect with 4-toluoyl chloride, the regioselectivity of the former was higher than the latter due to their different reaction mechanisms. The addition–elimination mechanism of 4-chlorobenzoyl chloride needed to overcome the stronger steric hindrance than the ionization mechanism of 4-toluoyl chloride, thus 4-chlorobenzoyl chloride was more inclined to react with the primary hydroxyl group (6-OH). In conclusion, under the synergistic effect of appropriate reaction rate, moderate steric effect and mechanism of benzoyl chlorides, the benzoylation of cellulose in AmimCl achieved unusually high regioselectivity. In view of this high regioselectivity, regioselectively modified cellulose derivatives can be prepared directly even via a “one pot” process with adding acylation reagents in turn.

Fig. 4 ¹³C-NMR spectra of cellulose benzoates prepared under different reaction conditions. Spectrum (a) of cellulose benzoate with DS = 0.99 prepared at 60 °C for 20 min without catalyst; and spectrum (b) of cellulose benzoate with DS = 1.08 prepared at 60 °C for 10 min with pyridine as the acid scavenger.

Conclusions

A series of cellulose substituted benzoates were synthesized homogeneously in AmimCl. The substituent on the benzene ring had a great impact on the benzoylation. For most of substituted benzoyl chlorides, either electron-donating or electron-withdrawing substituents accelerated the benzoylation reaction. Combining Hammett constants (σ) of substituents, a V-shaped Hammett plot of DS vs. σ was obtained. The Hammett plot could be utilized to forecast the previously unperformed reactions. And it was indicated that benzoyl chlorides with electron-donating substituents reacted with cellulose by a unimolecular ionization mechanism, and benzoyl chlorides with electron-withdrawing substituents by a bimolecular addition–elimination mechanism.

The DS of cellulose benzoates was readily controlled by altering reaction temperature, reaction time, and molar ratio of benzoylation agent/AGU. In addition, in the benzoylation of cellulose, AmimCl acted as a higher polarity solvent than DMF.

Furthermore, ¹³C-NMR measurement demonstrated that most of benzoyl chlorides gave an unusually high regioselectivity (>90% selectivity) at the C-6 position of cellulose in AmimCl, which provided a new strategy for the direct synthesis of regioselective cellulose derivatives. The high regioselectivity of cellulose benzoylation in AmimCl was attributed to the synergistic effect of appropriate reaction rate, moderate steric effect and reaction mechanism.

Acknowledgements

This work was supported by the National Science Foundation of China (No. 51103167, No. 21174151, and No. 51425307).

References

1. D. Klemm, B. Heublein, H. P. Fink and A. Bohn, Angew. Chem., Int. Ed., 2005, 44, 3358–3393.
2. Y. Cao, J. Wu, J. Zhang, H. Q. Li, Y. Zhang and J. S. He, Chem. Eng. J., 2009, 147, 13–21.
3. A. Pinkert, K. N. Marsh, S. S. Pang and M. P. Staiger, Chem. Rev., 2009, 109, 6712–6728.
4. T. Liebert and T. Heinze, BioResources, 2008, 3, 576–601.
5. J. Wu, J. Zhang, H. Zhang, J. S. He, Q. Ren and M. Guo, Biomacromolecules, 2004, 5, 266–268.
6. M. Gericke, T. Liebert and T. Heinze, Macromol. Biosci., 2009, 9, 343–353.
7. S. Barthel and T. Heinze, Green Chem., 2006, 8, 301–306.
8. J. M. Zhang, J. Wu, Y. Cao, S. M. Sang, J. Zhang and J. S. He, Cellulose, 2009, 16, 299–308.
9. M. Granström, J. Kavakka, A. King, J. Majoinen, V. Makela, J. Helaja, S. Hietala, T. Virtanen, S. L. Maunu, D. S. Argyropoulos and I. Kilpelainen, Cellulose, 2008, 15, 481–488.
10. S. Dorn, A. Pfeifer, K. Schlufter and T. Heinze, Polym. Bull., 2010, 64, 845–854.
11. X. T. Xu, W. G. Duan, M. Huang and G. H. Li, Carbohydr. Res., 2011, 346, 2024–2027.
12. X. X. Zhang, W. Zhang, D. Tian, Z. H. Zhou and C. H. Lu, RSC Adv., 2013, 3, 7722–7725.
13. D. Tian, Y. Y. Han, C. H. Lu, X. X. Zhang and G. P. Yuan, Carbohydr. Polym., 2014, 113, 83–90.
14. T. Iwata, A. Fukushima, K. Okamura and J. Azuma, J. Appl. Polym. Sci., 1997, 65, 1511–1515.
15. M. Kaplan, J. Chem. Eng. Data, 1961, 6, 272–276.
16. J. P. Richard and W. P. Jencks, J. Am. Chem. Soc., 1982, 104, 4689–4691.
17. G. F. Wang, Z. Q. Zhang and L. H. Song, Green Chem., 2014, 16, 1436–1443.
18. L. P. Hammett, Chem. Rev., 1935, 17, 125–136.
19. B. D. Song and W. P. Jencks, J. Am. Chem. Soc., 1989, 111, 8470–8479.
20. E. V. Anslyn and D. A. Dougherty, Modern Physical Organic Chemistry, University Science Books, 2006, pp. 421–432.
21. S. C. Fox, B. Li, D. Q. Xu and K. J. Edgar, Biomacromolecules, 2011, 12, 1956–1972.
22. M. Kogure, H. Bekku and K. Tasaka, 2013, WO2013005716–A1.
23. S. Kondo, C. Yamamoto, M. Kamigaito and Y. Okamoto, Chem. Lett., 2008, 37, 558–559.
24. E. Francotte and T. Zhang, Analusis, 1995, 23, M13–M20.
25. G. Felix, J. Chromatogr. A, 2001, 906, 171–184.
26. B. Helferich and H. Koester, Eur. J. Inorg. Chem., 1924, 57, 587–591.
27. W. M. Hearon, G. D. Hiatt and C. R. Fordyce, J. Am. Chem. Soc., 1943, 65, 2449–2452.
28. U. Erler, D. Klemm and I. Nehls, Macromol. Rapid Commun., 1992, 13, 195–201.
29. T. Erdmenger, C. Haensch, R. Hoogenboom and U. S. Schubert, Macromol. Biosci., 2007, 7, 440–445.
30. D. Q. Xu, B. Li, C. Tate and K. J. Edgar, Cellulose, 2011, 18, 405–419.
31. M. Schnabelrauch, S. Vogt, D. Klemm, I. Nehls and B. Philipp, Angew. Makromol. Chem., 1992, 198, 155–164.
32. H. Zhang, J. Wu, J. Zhang and J. S. He, Macromolecules, 2005, 38, 8272–8277.
33. V. W. Goodlett, J. T. Doughert and H. W. Patton, J. Polym. Sci., Part A: Polym. Chem., 1971, 9, 155–161.
34. C. Hansch, A. Leo and R. W. Taft, Chem. Rev., 1991, 91, 165–195.
35. D. N. Kevill and W. F. K. Wang, J. Chem. Soc., Perkin Trans. 1, 1998, 2, 2631–2637.
36. P. Campos-Rey, C. Cabaleiro-Lago and P. Herves, J. Phys. Chem. B, 2009, 113, 11921–11927.
37. C. A. Bunton, N. D. Gillitt, M. M. Mhala, J. R. Moffatt and A. K. Yatsimirsky, Langmuir, 2000, 16, 8595–8603.
38. P. Campos-Rey, C. Cabaleiro-Lago and P. Herves, J. Phys. Chem. B, 2010, 114, 14004–14011.
39. R. Rinaldi, Chem. Commun., 2011, 47, 511–513.
40. M. Gericke, T. Liebert, O. A. El Seoud and T. Heinze, Macromol. Mater. Eng., 2011, 296, 483–493.
41. J. Vitz, T. Erdmenger, C. Haensch and U. S. Schubert, Green Chem., 2009, 11, 417–424.
42. T. Liebert, M. Schnabelrauch, D. Klemm and U. Erler, Cellulose, 1994, 1, 249–258.
43. T. Heinze and T. Liebert, Prog. Polym. Sci., 2001, 26, 1689–1762.
44. K. Rahn, M. Diamantoglou, D. Klemm, H. Berghmans and T. Heinze, Angew. Makromol. Chem., 1996, 238, 143–163.

---