Letter pubs.acs.org/JPCL
Ionization of Cellobiose in Aqueous Alkali and the Mechanism of Cellulose Dissolution Erik Bialik,*,† Björn Stenqvist,† Yuan Fang,‡ Åsa Ö stlund,§ István Furó,‡ Björn Lindman,∥ Mikael Lund,*,† and Diana Bernin*,⊥ †
Division of Theoretical Chemistry, Lund University, 221 00 Lund, Sweden Division of Applied Physical Chemistry, Department of Chemistry, KTH Royal Institute of Technology, 114 00 Stockholm, Sweden § Sustainable Built Environment, SP Technical Research Institute of Sweden, 114 28 Stockholm, Sweden ∥ Division of Physical Chemistry, Lund University, 221 00 Lund, Sweden ⊥ Swedish NMR Centre, University of Gothenburg, 405 30 Göteborg, Sweden ‡
S Supporting Information *
ABSTRACT: Cellulose, one of the most abundant renewable resources, is insoluble in most common solvents but dissolves in aqueous alkali under a narrow range of conditions. To elucidate the solubilization mechanism, we performed electrophoretic NMR on cellobiose, a subunit of cellulose, showing that cellobiose acts as an acid with two dissociation steps at pH 12 and 13.5. Chemical shift differences between cellobiose in NaOH and NaCl were estimated using 2D NMR and compared to DFT shift differences upon deprotonation. The dissociation steps are the deprotonation of the hemiacetal OH group and the deprotonation of one of four OH groups on the nonreducing anhydroglucose unit. MD simulations reveal that aggregation is suppressed upon charging cellulose chains in solution. Our findings strongly suggest that cellulose is to a large extent charged in concentrated aqueous alkali, a seemingly crucial factor for solubilization. This insight, overlooked in the current literature, is important for understanding cellulose dissolution and for synthesis of new sustainable materials.
C
older literature and some industrial handbooks,4,12 cellulose swelling and dissolution in alkali is typically ascribed to deprotonation of OH groups, turning the polymer into a polyelectrolyte and causing the counterions to exert an osmotic swelling pressure.14 With some notable exceptions,15,16 this perspective is missing in the current literature. In two recent reviews,9,10 OH deprotonation is hardly mentioned in connection to alkali systems. Instead, the effect of alkali is discussed in terms of weakening of hydrogen bonds between cellulose chains and specific ion interactions. Electrophoretic NMR (eNMR)17,18 and diffusion NMR19 measurements respectively yield the electrophoretic mobility, μ, and the self-diffusion coefficient, D, of cellobiose in aqueous alkali. The effective charge, Z,20−23 of cellobiose can be obtained in unit charges via the Nernst−Einstein equation Z = μkBT/eD, where kBT is the thermal energy and e is the unit charge. The effective charge includes collected Na+ or K+ ions and can therefore be significantly smaller in magnitude than the nominal charge, Znom. Under certain circumstances not met here, can eNMR yield the nominal charge directly.24 The small experimental uncertainty is well in line with results obtained and analysis provided previously.21,24−26 Figure 1 shows the effective charge of cellobiose as a function of pH. The cellobiose solutions contain D2O and thereby their
ellulose is a fundamental structural component in plant matter and one of the most important renewable resources on earth.1 Its hierarchically ordered structure, however, makes it a difficult material to shape.2 Dissolution followed by precipitation is a viable method to create cellulosebased materials, e.g., regenerated fibers. A large number of cellulose solvents are known,3 but only a few are suitable for industrial applications. Cellulose swells in aqueous alkali, a fact that is exploited in, for instance, the Viscose process,4 during mercerisation,5 a treatment of natural fibers, and in the production of carboxymethyl cellulose.6 For a narrow range of conditions, cellulose dissolves even without any chemical modification, although the resulting solution appears to contain large aggregates.7,8 There are ongoing efforts to adapt an aqueous alkali and/or other hydroxide system as a sustainable industrial solvent system for cellulose.9−11 Success would provide an environmentally benign alternative to the Viscose process, which has a huge energy demand and generates hazardous sulfuric byproducts.4 Alternative solvents exist, notably in the Lyocell process with N-methylmorpholine-Noxide that can be recovered at high yields.12 Yet, the cheaper Viscose process still accounts for the overwhelming majority of regenerated cellulose fiber production, with 4.9 million tons produced globally in 2013, and production capacity is increasing despite the process becoming less common in developed economies.13 Pinpointing the molecular origin of the effect of alkali on cellulose is therefore crucial for developing new dissolution strategies and cellulose-based materials. In © XXXX American Chemical Society
Received: October 11, 2016 Accepted: November 22, 2016 Published: November 22, 2016 5044
DOI: 10.1021/acs.jpclett.6b02346 J. Phys. Chem. Lett. 2016, 7, 5044−5048
Letter
The Journal of Physical Chemistry Letters
obtain a particular pH. A larger amount of KOH is, however, needed to obtain the same pH because NaOH is a stronger base.29,30 Hence, an apparent cation specificity, reported previously,31,32 may appear due to the interaction between alkali and OH ions. The eNMR results raise, but cannot answer, the question of which OH group is deprotonated in the second dissociation step. Thus, we used 2D heteronuclear single quantum coherence (HSQC) NMR to measure the difference in 1H and 13C chemical shifts of cellobiose in 3 M aqueous NaCl and NaOH. Spectra are shown in the Supporting Information (SI). The chemical shifts of cellobiose and corresponding dianions with deprotonated OH groups (O−) in (i) the reducing hemiacetal and (ii) one out of four possible positions of the nonreducing ring were calculated on the DFT level. A recent benchmark study shows that this level of theory performs well for prediction of NMR chemical shifts.33 We estimated the difference in chemical shift between two closely related structures in NaOH and NaCl, which is straightforward compared to predicting chemical shifts relative to the conventional tetramethyl silane reference.34 The measured and calculated shifts are compared in Table 1 and visualized in Figure 2 for the β-anomer, which is the one most similar to cellulose. The α-anomer is summarized in the SI. The ‘weighted average’ (WA) column gives the shift difference corresponding to the mixture of deprotonated structures that best matches the measured 13C shifts in the ‘least-squares deviation’-sense. It consists of 44% ′O2, 42% ′O4, 14% ′O6 and 0% ′O3. We emphasize that these estimates are associated with an appreciable uncertainty that arises both from the fitting procedure and the intrinsic uncertainty of the calculations. Yet, it is without doubt that the deprotonation of any single OH group alone in the nonreducing unit is not consistent with our findings. Importantly, the OH group containing ′O4 is present only at the nonreducing end of cellulose, and we therefore conclude that deprotonation of ′O2 is the most important for cellulose solubilization. Deprotonation of an OH group tends to give a positive 13C chemical shift difference for the carbon atom to which it is bonded and for adjacent carbon atoms. The calculated 1H shift differences are generally negative, except for hydrogen atoms bonded to βC1, which is a deprotonated site.
Figure 1. Effective charge, Z, of cellobiose as a function of the solution pH adjusted by adding either KOH (closed circles) or NaOH (open circles). The estimated experimental uncertainty of Z is ±0.1. Dashed lines are linear fits to data below and above pH = 12.1, respectively.
acidity is characterized by their pD as obtained by commonly used pH-meters. Estimating the pH that would characterize the corresponding acidity in an H2O solution is done via the wellknown relation pH = pD − 0.4. For consistency, we present here pH values. The fact that Z decreases from 0 to significantly negative values in a limited pH range strongly favors the hypothesis that cellobiose is deprotonated in aqueous alkali. The hemiacetal OH groups of reducing sugars are typically weak acids with a pKa between 12 and 12.5.27 The first deprotonation step, signified by Z starting to decrease from around pH = 10.6, is therefore unremarkable. Further, it is less relevant to cellulose dissolution, as there is only one such group per molecule. However, Z continues to decrease with increasing pH and reaches a value of −1.5 around pH = 13.5. This shows the existence of a second deprotonation step with pKa about 1− 1.5 pH-units higher than the hemiacetal. This is consistent with reported values around pKa = 13.5 for cyclodextrins,27,28 which are better models of cellulose in the sense that they are free from reducing ends but worse in the sense that they are composed of α-glucose instead of β-glucose residues. The results do not depend on whether NaOH or KOH are used to
Table 1. 13C and 1H Chemical Shift Differences from HSQCs (bold text) between NaCl and NaOH and from DFT between βCellobiose and the Dianions Formed by Deprotonating βO1 and the Indicated Oxygen Atom on the Nonreducing Unita 13
1
C
βC1 βC2 βC3 βC4 βC5 βC6 ′C1 ′C2 ′C3 ′C4 ′C5 ′C6 a
H
NMR
WA
′O2
′O3
′O4
′O6
NMR
WA
′O2
′O3
′O4
′O6
6.2 3.4 0.8 2.5 -0.4 -0.1 1.8 1.1 3.0 1.7 2.7 1.1
7.3 2.9 1.6 2.5 −1.5 0.7 2.4 1.6 3.0 2.4 3.2 1.1
7.2 3.0 1.7 3.1 −1.4 0.8 5.0 2.6 3.1 1.2 0.2 0.2
7.9 3.2 0.8 −2.1 −0.9 0.2 −0.2 2.7 1.4 2.0 2.6 0.8
7.3 2.8 1.5 1.8 −1.6 0.6 0.3 0.8 3.7 4.2 5.0 1.5
7.5 2.7 1.5 2.5 −1.6 0.5 0.4 0.5 0.6 0.7 7.1 2.8
-0.04 -0.30 -0.13 -0.19 -0.24 -0.11 -0.15 -0.14 -0.23 -0.22 -0.23 -0.15
0.17 −0.46 −0.08 −0.12 −0.07 0.01 −0.13 −0.15 −0.40 −0.10 −0.23 −0.02
0.18 −0.43 −0.08 −0.15 −0.08 0.10 −0.17 −0.19 −0.54 −0.10 −0.06 −0.07
0.18 −0.53 −0.09 0.13 −0.06 −0.05 −0.02 −0.27 −0.21 −0.72 −0.08 −0.06
0.17 −0.47 −0.09 −0.12 −0.07 −0.09 −0.10 −0.12 −0.37 −0.09 −0.32 0.00
0.16 −0.51 −0.05 −0.06 −0.04 0.05 −0.06 −0.08 −0.05 −0.11 −0.49 0.06
See notation illustrated in Figure 2. WA is the weighted average (see text). 5045
DOI: 10.1021/acs.jpclett.6b02346 J. Phys. Chem. Lett. 2016, 7, 5044−5048
Letter
The Journal of Physical Chemistry Letters
chains in aqueous alkali. The association between cellulose molecules is quantified in terms of the average distance between anhydroglucose units (AGU), as shown in Figure 3. In
Figure 2. Visualization of the chemical shift differences from HSQC (a) and DFT (b) presented in Table 1. The maximum of the color scale is set to the maximum shift difference within each structure. Asterisks indicate those protons whose removal leads to the shift difference.
Comparison of the measured and computed shift difference for the deprotonated βC1 shows that the computed 13C shift difference is overestimated with about 1 ppm. The accuracy of shift differences for the other carbon atoms in the reducing ring is comparable to that value and their rank order is correctly given by the calculations. Notably, the negative sign of the shift difference for βC5 is reproduced. The shift difference is larger for βC4 than for βC3, despite the former being further from the βC1 O−. Hydrogen atoms bonded to a carbon atom adjacent to a carbon atom with an OH/O− group are predicted to have a negative shift difference, typically more negative than the hydrogen closest to the OH/O− group. The calculated 1H shift differences do not match the experimental shifts as well as the 13 C shift differences, but salient features are in accord with experiment. Notable are βC2, adjacent to the deprotonated hemiacetal, and ′C5, which has no OH group of its own but is adjacent to two OH-bearing carbons. An interesting feature of the calculated shift differences for both 13C and 1H is that it is not always the atoms closest to the deprotonation site that are most strongly affected. This explains that appreciable shift differences for ′C1 and ′C5 are experimentally observed even though these carbon atoms are not bonded to OH groups. Similar observations have been reported for cyclodextrins,28 but in that work the data was analyzed assuming that the shift change of a single carbon could be directly translated to the deprotonation of bonded OH groups. Our calculations suggest that, although it does not seem to significantly affect the conclusions in that work, this assumption may produce erroneous identification of deprotonation sites in close proximity. To test the effect of charge on solubility, we performed 1 μs MD simulations of neutral and O2 deprotonated cellodecaose
Figure 3. Cellulose configurations in the last frame of a 1 μs simulation for (a) neutral and (b) deprotonated cellodecaose and matrices of the average minimum distance between atoms in pairs of AGUs over 100 ns trajectory segments, for (c) neutral and (d) deprotonated cellodecaose. The AGU are arranged in sequence within each molecule, and the numbering refers to molecule index. Brighter fields correspond to closer contact between corresponding AGU; the diagonal bands correspond to intramolecular correlations.
the neutral system (Figure 3a), chains 1 and 2 associate to a sandwich-like aggregate with the rings of the two chains facing each other, within a few tens of nanoseconds. This is visible as the off-diagonal band that appears in each of the panels of Figure 3c. The third chain remained free for about 200 ns, after which it was caught by the aggregate. This can be seen from the appearance of additional off-diagonal features. After around 700−800 ns, it detaches, but then reattaches. As can be seen in Figure 3d, no persistent aggregation occurs for the deprotonated cellulose, though long-lived but ultimately transient contacts can be seen, e.g., between 100 and 200 ns. During the final 200 ns of the simulation, such contacts appear between all three chains. As can be seen from the snapshot in Figure 3b, chain 1 has detached at the end of the simulation, while contact remains between chains 2 and 3. Corresponding movies can be found in the SI. MD simulations of the neutral system show a near-complete association indicating that the cellulose model is insoluble, as it should be. As the deprotonation is the only 5046
DOI: 10.1021/acs.jpclett.6b02346 J. Phys. Chem. Lett. 2016, 7, 5044−5048
Letter
The Journal of Physical Chemistry Letters
(3) Heinze, T.; Koschella, A. Solvents Applied in the Field of Cellulose Chemistry - a Mini Review. Polim.: Cienc. Tecnol. 2005, 15, 84−90. (4) Wilkes, A. G. Regenerated Cellulose Fibres, 1st ed.; Woodings, C., Ed.; Woodhead Publishing: Cambridge, U.K., 2001; pp 37−67. (5) Klemm, D.; Heublein, B.; Fink, H.-P.; Bohn, A. Cellulose: Fascinating Biopolymer and Sustainable Raw Material. Angew. Chem., Int. Ed. 2005, 44, 3358−3393. (6) Mann, G.; Kunze, J.; Loth, F.; Fink, H. P. Cellulose Ethers with a Block-Like Distribution of the Substituents by Structure-Selective Derivatization of Cellulose. Polymer 1998, 39, 3155−3165. (7) Alves, L.; Medronho, B. F.; Antunes, F. E.; Romano, A.; Miguel, M. G.; Lindman, B. On the Role of Hydrophobic Interactions in Cellulose Dissolution and Regeneration: Colloidal Aggregates and Molecular Solutions. Colloids Surf., A 2015, 483, 257−263. (8) Schleicher, H.; Borrmeister, B. Investigations on the State of Solution of Viscose. Lenzinger Berichte 1998, 78, 7−11. (9) Budtova, T.; Navard, P. Cellulose in NaOH - Water Based Solvents: a Review. Cellulose 2016, 23, 5−55. (10) Wang, S.; Lu, A.; Zhang, L. Recent Advances in Regenerated Cellulose Materials. Prog. Polym. Sci. 2016, 53, 169−206. (11) Abe, M.; Kuroda, K.; Ohno, H. Maintenance-Free Cellulose Solvents Based on Onium Hydroxides. ACS Sustainable Chem. Eng. 2015, 3, 1771−1776. (12) White, P. Regenerated Cellulose Fibres, 1st ed.; Woodings, C., Ed.; Woodhead Publishing: Cambridge, U.K., 2001; pp 62−87. (13) Horrocks, A. R.; Anand, S. C. Handbook of Technical Textiles: Technical Textile Applications; Woodhead Publishing: Cambridge, U.K., 2016; Vol. 2. (14) Marsh, J. T.; Wood, F. C. An Introduction to the Chemistry of Cellulose; Chapman & Hall Ltd.: London, 1942. (15) Medronho, B.; Lindman, B. Competing Forces During Cellulose Dissolution: From Solvents to Mechanisms. Curr. Opin. Colloid Interface Sci. 2014, 19, 32−40. (16) Alves, L.; Medronho, B.; Antunes, F. E.; Topgaard, D.; Lindman, B. Dissolution State of Cellulose in Aqueous Systems. 1. Alkaline Solvents. Cellulose 2016, 23, 247−258. (17) Holz, M. Electrophoretic NMR. Chem. Soc. Rev. 1994, 23, 165− 174. (18) Griffiths, P. C. Electrophoretic NMR - Ions, Molecules, Mixtures, Pores and Complexes. Annu. Rep. NMR Spectrosc. 2009, 65, 139−159. (19) Price, W. S. NMR Studies of Translational Motion; Cambridge University Press: London, 2009. (20) Hallberg, F.; Furó, I.; Stilbs, P. Ion Pairing in Ethanol/Water Solution Probed by Electrophoretic and Diffusion NMR. J. Am. Chem. Soc. 2009, 131, 13900−13901. (21) Hallberg, F.; Furó, I.; Yushmanov, P. V.; Stilbs, P. Sensitive and Robust Electrophoretic NMR. Instrumentation and Experiments. J. Magn. Reson. 2008, 192, 69−77. (22) Böhme, U.; Scheler, J. Effective Charge of Polyelectrolytes as a Function of the Dielectric Constant of a Solution. J. Colloid Interface Sci. 2007, 309, 231−235. (23) Böhme, U.; Scheler, J. Hydrodynamic Size and Charge of Polyelectrolyte Complexes. J. Phys. Chem. B 2007, 111, 8348−8350. (24) Giesecke, M.; Szabó, Z.; Furó, I. The Protonation State and Binding Mode in a Metal Coordination Complex From the Charge Measured in Solution by Electrophoretic NMR. Anal. Methods 2013, 5, 1648−1651. (25) Bielejewski, M.; Giesecke, M.; Furó, I. On Electrophoretic NMR. Exploring High Conductivity Samples. J. Magn. Reson. 2014, 243, 17−24. (26) Giesecke, M.; Hallberg, F.; Fang, Y.; Stilbs, P.; Furó, I. Binding of Monovalent and Multivalent Metal Cations to Polyethylene Oxide in Methanol Probed by Electrophoretic and Diffusion NMR. J. Phys. Chem. B 2016, 120, 10358−10366. (27) Sinnott, M. Carbohydrate Chemistry and Biochemistry: Structure and Mechanism; RSC Publishing: London, 2007.
difference between the models in the two simulations, the difference in outcome suggests that the formation of a single O− group per AGU unit is sufficient to explain the solubilizing effect of alkali. There is, however, a tendency for cellulose chains to associate even when deprotonated, which explains the empirical observation of eventually gelling of alkaline cellulose solutions.35 The results presented above strongly suggest that the C2 and, to a lesser extent, C6 OH group in cellulose are deprotonated under conditions where the polymer can dissolve in aqueous alkali and that this deprotonation is sufficient to explain the solubilization. One unit charge per AGU unit (0.5 nm) is a high line charge density and, in light of the rich literature on polyelectrolyte solubility,36−38 it is unsurprising that it is sufficient to markedly affect solubility. Further measurements on larger, preferably end-protected, cellooligomers would be needed to confirm this conclusion.
■
EXPERIMENTAL METHODS For 2D NMR experiments, solutions of 0.1 M D(+)cellobiose (Sigma >98%) in D2O (Sigma-Aldrich, 99.8%), 3 M NaOH (Merck, > 99%) or 3 M NaCl (Sigma-Aldrich) were used and experiments were performed at 2.2 °C. For eNMR experiments, solutions of 0.03 up to 0.1 M D(+) cellobiose in D2O with varying NaOH or KOH (Merck, > 85%) concentrations were used. Additional details including DFT and MD, are provided in the Supporting Information.
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpclett.6b02346. Experimental details about eNMR and 2D NMR, details about computational methods, additional results from alpha-cellobiose, and HSQC spectra (PDF) Movies from MD simulations (ZIP)
■
AUTHOR INFORMATION
Corresponding Authors
*E-mail:
[email protected]. *E-mail:
[email protected]. *E-mail:
[email protected]. ORCID
Diana Bernin: 0000-0002-9611-2263 Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS Avancell, Södra Skogsägarnas research foundation, the Swedish Foundation for Strategic Research SSF, and the Swedish Research Council VR financially supported this work. The Swedish NMR Centre is acknowledged for spectrometer time and LUNARC in Lund provided computational resources.
■
REFERENCES
(1) Gross, R. A.; Kalra, B. Biodegradable Polymers for the Environment. Science 2002, 297, 803−807. (2) Fernandes, A. N.; Thomas, L. H.; Altaner, C. M.; Callow, P.; Forsyth, V. T.; Apperley, D. C.; Kennedy, C. J.; Jarvis, M. C. Nanostructure of Cellulose Microfibrils in Spruce Wood. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, E1195−E1203. 5047
DOI: 10.1021/acs.jpclett.6b02346 J. Phys. Chem. Lett. 2016, 7, 5044−5048
Letter
The Journal of Physical Chemistry Letters (28) Gaidamauskas, E.; Norkus, E.; Butkus, E.; Crans, D. C.; Grinciene, G. Deprotonation of B-Cyclodextrin in Alkaline Solutions. Carbohydr. Res. 2009, 344, 250−254. (29) Bercovici, M.; Kaigala, G. V.; Santiago, J. G. Method for Analyte Identification Using Isotachophoresis and a Fluorescent Carrier Ampholyte Assay. Anal. Chem. 2010, 82, 2134−2138. (30) Ben-Harush, K.; Wolfson, A.; Herskowitz, M. Unexpected Performance of Solid Alkaline Metal Hydroxides in Liquid Phase Oxidation of 1-Phenylethanol. Lett. Org. Chem. 2006, 3, 664−667. (31) Xiong, B.; Zhao, P.; Cai, P.; Zhang, L.; Hu, K.; Cheng, G. NMR Spectroscopic Studies on the Mechanism of Cellulose Dissolution in Alkali Solutions. Cellulose 2013, 20, 613−621. (32) Liu, Z.; Zhang, C.; Liu, R.; Zhang, W.; Kang, H.; Li, P.; Huang, Y. Dissolution of Cellobiose in the Aqueous Solutions of Chloride Salts: Hofmeister Series Consideration. Cellulose 2016, 23, 295−305. (33) Flaig, D.; Maurer, M.; Hanni, M.; Braunger, K.; Kick, L.; Thubauville, M.; Ochsenfeld, C. Benchmarking Hydrogen and Carbon NMR Chemical Shifts at HF, DFT, and MP2 Levels. J. Chem. Theory Comput. 2014, 10, 572−578. (34) Sarotti, A. M.; Pellegrinet, S. C. Application of the MultiStandard Methodology for Calculating 1H NMR Chemical Shifts. J. Org. Chem. 2012, 77, 6059−6065. (35) Roy, C.; Budtova, T.; Navard, P. Rheological Properties and Gelation of Aqueous Cellulose - NaOH Solutions. Biomacromolecules 2003, 4, 259−264. (36) Evans, D. F.; Wennerström, H. The Colloidal Domain: Where Physics, Chemistry, Biology and Technology Meet, 2nd ed.; John Wiley & Sons, Inc.: New York, 1999. (37) Osada, Y.; Kholkholov, A. Polymer Gels and Networks; Marcel Dekker: New York, 2001. (38) Piculell, L.; Lindman, B. Association and Segregation in Aqueous Polymer/Polymer, Polymer/Surfactant, and Surfactant/ Surfactant Mixtures: Similarities and Differences. Adv. Colloid Interface Sci. 1992, 41, 149−178.
5048
DOI: 10.1021/acs.jpclett.6b02346 J. Phys. Chem. Lett. 2016, 7, 5044−5048
