Grafted cellulose strands on the surface of silica: effect of environment on reactivity

Abstract

The design, synthesis and characterization of materials consisting of grafted poly(1 → 4-β-glucan) strands on silica is reported. The silanol-rich environment provided in these materials activates the glycosidic bond for hydrolysis under mild conditions.

---

Hydrolysis of O-glycosidic bonds in cellulose is a crucial step in the selective conversion of biomass to fuels and chemicals using aqueous phase processing. Whereas enzymes accomplish this transformation under mild conditions of pH and temperature,¹ most synthetic catalysts rely on specific acid-catalyzed mechanisms that require highly concentrated mineral acids (e.g., up to 60% to 90% sulfuric acid).² The bioinspired synthetic system of Capon is exceptional because it achieves glycosidic bond hydrolysis at a mild aqueous solution pH of 4.5. However, a major practical limitation of the Capon system is that it requires the precise positioning of a carboxylic acid functional group adjacent to the glycosidic oxygen: a molecule with carboxylic acid functionality in the ortho position has a 13 000-fold faster hydrolysis rate relative to a similar molecule where the carboxylic acid functionality is in the para position.³ In this communication, we describe a system that overcomes the rigid requirement of intramolecular acid catalyst positioning of the Capon system by essentially providing an array of acid groups in the vicinity of the glycosidic bond, and demonstrate its utility for the hydrolysis of poly(1 → 4-β-glucan) (β-Glu) strands derived from cellulose. Our approach relies on the synthesis and characterization of a new class of bioinspired materials comprising grafted β-Glu strands within an environment consisting of an ensemble of mildly acidic surface silanols in silica. These materials are unique compared with previously described cellulose–silica hybrids,⁴ because each grafted β-Glu strand is molecularly dispersed on the surface, and is not interacting with other immobilized strands (i.e., there is no clumping of β-Glu strands to form a cellulose crystal). The new materials described here also offer unique opportunities as emerging biocomposites for diagnostics (e.g., imaging and chips for high-throughput screening),⁵drug delivery,⁵enzyme immobilization within a saccharide environment,⁶ enantioselective recognition for chromatographic separations and heterogeneous catalysis,⁷ and biodegradable fibers.⁸

Grafted cellulose strands on silica schematically represented in Fig. 1 are synthesized at low (material SGL) and high (material SGH) site densities. The synthetic approach is based upon the grafting of β-cyclodextrin and calix[4]arenes to the surface of silicavia reaction of R–OH groups with surface ≡Si–Cl functionality followed by washing and drying.^9,10 Material SGH, which is synthesized by first fully hydroxylating the silica surface prior to its chlorination, represents a higher β-Glu strand grafting density whereas SGL is synthesized by chlorinating the available low hydroxyl coverage present in the Cab-o-sil silica support.

Fig. 1 Schematic illustration of grafted β-Glu strands on the surface of silica (Cab-o-sil) for materials (a) SGL and (b) SGH.

The random irreversible adsorption of circular objects in a two-dimensional plane is expected to result in a packing density of approximately 0.55.¹¹ This packing density has been previously observed when using similar surface chemistry during the grafting of calix[4]arene monomers on the surface of silica.¹² Grafted β-Glu coverages are summarized in Table 1. The measured cellobiose packing density of 0.66 in SGH is consistent with polymeric β-Glu strands grafting onto the surface randomly and irreversibly. Distances between grafted cellobiose units in Table 1 for SGH and SGL preclude significant interactions between adjacent grafted β-Glu strands. These strands are expected to consist of rigid anchoring points to the silica surface and adopt a loop and tail configuration,^13,14 such that between anchoring points the free looping strand can hydrogen bond with silanols on the silica surface, as schematically shown in Fig. 1.

Table 1 Grafted cellulose strands on SiO₂ and cellulose material characteristics

Property	SG0	SGL	SGH	Avicel	Amorphous cellulose
a Represents average distance between cellobiose units as measured using thermal gravimetric analysis, see ESI.^1SG0 was synthesized without chlorination to silica surface. b Broad. c Distance between adjacent cellobiose units in crystalline cellulose is 1.034 nm as measured on the same strand and 0.78 nm as measured between adjacent strands.^15 d Values in parentheses represent the BET surface area of the respective control material, which consists of a similarly treated silica except in the absence of cellulose. e Values in parentheses represent the nondimensional cellobiose packing density.
Grafted poly(1 → 4-β-glucan) surface coverage on SiO2 [cellobiose/nm^2]^a^,^e	0	0.43 ± 0.02 (0.35)	0.82 ± 0.03 (0.66)	1.24^c (1.0)	N.D.
Adjacent cellobiose unit distance^a/nm	N.D.	1.50 ± 0.02	1.10 ± 0.02	0.90^c	N.D.
^13C CP/MAS NMR signals/ppm	N.D.	104.8,^b 84.2,^b 75.4, 62.3^b	105.2,^b 89.1, 86.6, 75.0, 72.6, 71.9, 64.5, 62.8^b	105.3, 89.1, 84.1,^b 75.3, 72.6, 71.7, 65.5, 62.6^b	104.8, 83.6,^b 74.8,^b 62.8^b
BET surface area^d/m^2 g^−1	200 ± 6	184 ± 6 (193)	155 ± 6 (184)	N.D.	N.D.

---

Solid-state ¹³C CP/MAS NMR spectroscopy provides characterization of the hydrogen bonding environment of grafted β-Glu strands, as shown in Fig. 2. The spectra of controls demonstrate resonances at 89 ppm for C4 and 65 ppm for C6 for microcrystalline cellulose, Avicel. Resonances for amorphous cellulose are observed to be upfield-shifted at 84 ppm for C4 and 63 ppm for C6.^16,17 The SGL spectrum is similar to that of amorphous cellulose whereas the SGH spectrum resembles crystalline cellulose. In general, polymer adsorption on a surface occurs when the enthalpic energy contribution arising from interaction with the surface overcomes the entropy loss due to a reduced number of polymer conformations in the adsorbed state.^13,18,19 The crystalline-like ¹³C CP/MAS NMR spectrum of SGH in Fig. 2a must therefore be due to strong enthalpic interactions with the silica surface viahydrogen bonding, which are enabled by the high density of anchor points along a strand. The silica surface acts essentially as a surrogate to other β-Glu strands present in crystalline cellulose, to enforce a strong hydrogen bonding environment surrounding grafted β-Glu strands in SGH. Conversely, the lower density of anchor points along a β-Glu strand in SGL results in a higher conformational degree of freedom and leads to a more flexible and open looping configuration for a grafted strand. This configuration is responsible for the relatively low surface packing density of 0.35 for SGL. It also leads to a less pronounced degree of hydrogen bonding with the silica surface relative to SGH, as exhibited by the amorphous-like ¹³C NMR spectrum in Fig. 2c. In both SGL and SGH, the NMR spectrum is a reflection of the tendency for grafted β-Glu strands to hydrogen bond to silica surface silanols.

Fig. 2 ¹³C CP/MAS NMR spectra of materials consisting of grafted cellulose strands on SiO₂ and cellulose: (a) SGH, (b) Avicel (microcrystalline cellulose), (c) SGL, and (d) amorphous cellulose (Avicel precipitated in MeOH from LiCl/DMAc solution and subsequently ball-milled).

Data from N₂ physisorption at 77 K are used to further investigate interactions between grafted β-Glu strands and silica in SGH and SGL. This technique is sensitive to the degree of penetration of grafted β-Glu strands into confined pores, which are known to consist of “ditches”²⁰ between intersecting dense silica spheres in Cab-o-sil. This “ditch” region is qualitatively illustrated in Fig. 1. The ability of a grafted β-Glu strand to penetrate a “ditch” in SGH and SGL is described by the N₂ physisorption data in Fig. 3. The inset to Fig. 3 represents the subtraction isotherm of both SGL and SGH, from its respective control material, which is synthesized in the absence of cellulose. The subtraction isotherm in the regime of low P/P° values (P/P° < 0.001) is attributed to the smallest micropore region of the “ditch” and is the same for both SGL and SGH. This signifies that the smallest micropores in Cab-o-sil are equally inaccessible for β-Glu strand adsorption in both materials, presumably because the confinement there is too entropically unfavorable. The constancy of the subtraction isotherm for SGL for P/P° < 0.1 means that micropores in general are largely inaccessible to grafted β-Glu strands in SGL. For SGH, a significant fraction of micropores (corresponding approximately to 0.001 < P/P° < 0.1) are accessible to grafted β-Glu strands, as evidenced by the increase in the subtraction isotherm with relative pressure, which arises due to the ability of grafted β-Glu strands to block micropores. The observed differences in the subtraction isotherms between materials SGL and SGH are the result of the higher enthalpic energy driving force for adsorption into a pore for SGH, as discussed above within the context of the ¹³C CP/MAS NMR spectroscopic data. This driving force is also reflected by the decrease in the BET surface area relative to a control material, which is summarized in Table 1 to be (9 ± 6) m² g⁻¹ and (23 ± 6) m² g⁻¹ for SGL and SGH, respectively.

Fig. 3 Nitrogen adsorption/desorption isotherms at 77 K for SGL, SGH and their respective control materials, which consist of supports that were treated identically, except in the absence of cellulose. Inset: subtraction isotherm (using adsorption branch) between respective control material and either SGL (i) or SGH (ii). Control materials were synthesized following the same synthesis steps pertaining to either SGL or SGH, with the exception that cellulose was excluded from the synthesis.

The observed hydrogen bonding between surface silanols on silica and grafted β-Glu strands may activate glycosidic bonds in the strands towards hydrolysis, by the same intramolecular general acid catalysis mechanism invoked by Capon.³ The catalytic effect of surface silanols on hydrolysis of β-Glu strands is investigated at pH 4 HCl aqueous solution and 105 °C (autogeneous pressure) so that conditions are similar to those used by Capon. These mild conditions are known to show negligible specific acid catalyzed hydrolysis of glycosidic bonds in cellobiose, which is completely soluble in water and has no accessibility limitations.²¹Fig. 4 represents the relative amounts of soluble β-Glu fragments measured after 5 h of reaction using HPLC for SGL, SGH and a control comprising a mechanical mixture of amorphous cellulose and SiO₂. Depending on the particular β-Glu fragment released (retention time), material SGL results in 5-fold to 150-fold more β-Glu fragments released during hydrolysis than the control, and SGH results in up to 25-fold more β-Glu fragments released during hydrolysis than the control (see ESI‡). These data demonstrating increased rate of hydrolysis are all the more compelling when considering the grafted nature of β-Glu strands in SGL. Only those fragments that result from two cuts between anchor points on the surface are measurable viaHPLC. This requirement explains the lower measured activity for SGH, despite its greater degree of interaction with surface silanols, since the probability of achieving two cuts between the significantly smaller number of repeating cellobiose units located in between anchor points in SGH is much smaller. The activation of grafted glycosidic bonds by the mild acidity of silanols on silica (average pK_a ≈ 7)²² in these materials must arise by a mechanism in which silanols act as an array of weak acid localized sites.^23,24 This role of the silanols is akin to the precisely positioned carboxylic acid functional group in the intramolecular system of Capon.³ This conceptual approach has recently inspired the design of a catalytic system for the hydrolysis of glycosidic bonds in aqueous solution, and further studies will be reported in due course.²⁵

Fig. 4 Soluble β-Glu fragments released during reaction at 105 °C at pH 4 in aqueous solution for (a) SGL, (b) SGH and (c) a mechanical mixture of precipitated cellulose with Cab-o-sil (A-Cell + SiO₂).

Notes and references

1. C. S. Rye and S. G. Withers, Curr. Opin. Chem. Biol., 2000, 4, 573–580.
2. W. A. Farone and J. E. Cuzens, US Patent 556277, USA, 1996.
3. B. Capon, Tetrahedron Lett., 1963, 14, 911–913.
4. S. Sequeira, D. V. Evtuguin, I. Portugal and A. P. Esculcas, Mater. Sci. Eng., C, 2007, 27, 172–179.
5. G. Coullerez, P. H. Seeberger and M. Textor, Macromol. Biosci., 2006, 6, 634–647.
6. K. Loos, V. von Braunmuhl, R. Stadler, K. Landfester and H. W. Spiess, Macromol. Rapid Commun., 1997, 18, 927–938.
7. D. Lubda, K. Cabrera, K. Nakanishi and W. Lindner, Anal. Bioanal. Chem., 2003, 377, 892–901.
8. H. Zhang, Z. G. Wang, Z. N. Zhang, J. Wu, J. Zhang and J. S. He, Adv. Mater., 2007, 19, 698–704.
9. A. Katz, P. Da Costa, A. C. P. Lam and J. M. Notestein, Chem. Mater., 2002, 14, 3364–3368.
10. S. Shiraishi, M. Komiyama and H. Hirai, Bull. Chem. Soc. Jpn., 1986, 59, 507–510.
11. J. Feder, J. Theor. Biol., 1980, 87, 237–254.
12. J. M. Notestein, E. Iglesia and A. Katz, J. Am. Chem. Soc., 2004, 126, 16478–16486.
13. R. R. Netz and D. Andelman, Phys. Rep., 2003, 380, 1–95.
14. E. Tadd, A. Zeno, M. Zubris, N. Dan and R. Tannenbaum, Macromolecules, 2003, 36, 6497–6502.
15. A. C. O'Sullivan, Cellulose, 1997, 4, 173–207.
16. S. Park, D. K. Johnson, C. I. Ishizawa, P. A. Parilla and M. F. Davis, Cellulose, 2009, 16, 641–647.
17. P. T. Larsson, K. Wickholm and T. Iversen, Carbohydr. Res., 1997, 302, 19–25.
18. P. G. De Gennes, Macromolecules, 1981, 14, 1637–1644.
19. A. Chremos, E. Glynos, V. Koutsos and P. J. Camp, Soft Matter, 2009, 5, 637–645.
20. C. J. Gommes, P. Ravikovitch and A. Neimark, J. Colloid Interface Sci., 2007, 314, 415–421.
21. O. Bobleter, Prog. Polym. Sci., 1994, 19, 797–841.
22. M. L. Hair and W. Hertl, J. Phys. Chem., 1970, 74, 91–94.
23. J. D. Bass, A. Solovyov, A. J. Pascall and A. Katz, J. Am. Chem. Soc., 2006, 128, 3737–3747.
24. S. L. Hruby and B. H. Shanks, J. Catal., 2009, 263, 181–188.
25. A. Charmot and A. Katz, J. Catal., 2010 DOI:10.1016/j.jcat.2010.08.006 , in Press.

---

Footnotes

† This article is part of the ‘Emerging Investigators’ themed issue for ChemComm.

‡ Electronic supplementary information (ESI) available: Experimental section, Fig. S1 – Fig. S3., supplementary discussion and appendix of calculations. See DOI: 10.1039/c0cc02105a

---