Molecular Dynamics Study of Polysaccharides in Binary Solvent

ARTICLE pubs.acs.org/JPCB

Molecular Dynamics Study of Polysaccharides in Binary Solvent Mixtures of an Ionic Liquid and Water Hanbin Liu,†,‡ Kenneth L. Sale,†,‡ Blake A. Simmons,†,‡ and Seema Singh*,†,‡ † ‡

Deconstruction Division, Joint BioEnergy Institute, Emeryville, California, United States Biomass Science and Conversion Technology Department, Sandia National Laboratories, Livermore, California, United States ABSTRACT: Some ionic liquids (ILs) have great promise as effective solvents for biomass pretreatment, and there are several that have been reported that can dissolve large amounts of cellulose. The solubilized cellulose can then be recovered by addition of antisolvents, such as water or ethanol, and this regeneration process plays an important role in the subsequent enzymatic saccharification reactions and in the recovery of the ionic liquid. To date, little is known about the fundamental intermolecular interactions that drive the dissolution and subsequent regeneration of cellulose in complex mixtures of ionic liquids, water, and cellulose. To investigate these interactions, in this work, molecular dynamics (MD) simulations were carried out to study binary and ternary mixtures of the ionic liquid 1-ethyl-3-methylimidazolium acetate ([C2mim][OAc]) with water and a cellulose oligomer. Simulations of a cellulose oligomer dissolved in three concentrations of binary mixtures of [C2mim][OAc] and water were used to represent the ternary system in the dissolution phase (high [C2mim][OAc] concentration) and present during the initial phase of the regeneration step (intermediate and low [C2mim][OAc] concentrations). The MD analysis of the structure and dynamics that exist in these binary and ternary mixtures provides information on the key intermolecular interactions between cellulose and [C2mim][OAc] that lead to dissolution of cellulose and the key intermolecular interactions in the intermediate states of cellulose precipitation as a function of water content in the cellulose/IL/water system. The analysis of this intermediate state provides new insight into the molecular driving forces present in this ternary system.

’ INTRODUCTION Pretreatment measures designed to break the recalcitrance of lignocellulosic biomass are essential for the effective conversion of biomass into renewable liquid transportation fuels. Imidazolium-based ionic liquids (ILs), such as 1-ethyl-3-methyl-imidazolium acetate ([C2mim][OAc]), have been reported that dissolve significant amounts of crystalline cellulose and represent a novel class of solvents capable of overcoming the recalcitrant of microcrystalline cellulose.1 4 The dissolved cellulose can then be precipitated using either water or ethanol as an antisolvent, generating an amorphous cellulose product that can be efficiently hydrolyzed by cellulases.1 3,5 7 These early results on microcrystalline cellulose have prompted researchers to investigate ILs that can be used to pretreat and fractionate lignocellulosic biomass4,8 11 into its cellulose, hemicellulose, and lignin12 15 components, thereby enabling the cost-effective production of monomeric sugars that can be fermented into biofuels. As part of our ongoing research in biomass conversion, the work presented here addresses the interactions among the ionic liquid 1-ethyl-3methylimidazolium acetate ([C2mim][OAc]), water, and cellulose that lead first to dissolution of cellulose and then to the regeneration of cellulose from an ionic liquid aqueous solution upon dilution of the IL with water. At present, the ILs that are known to efficiently dissolve cellulose are mainly composed of imidazolium- or pyridinium-based cations r 2011 American Chemical Society

with anions, such as [OAc] , Cl , Br , and [(CH3CH2O)2PO2] ,1,2,16,17 that serve as hydrogen bond donors (for a review, see ref 18 and references therein). Previous research proposed that the basicity of the anion and its ability to accept hydrogen bonds disrupts the interchain H-bonding in cellulose.1 The ability of the anion to form hydrogen bonds with a strength greater than interchain hydrogen bonding, is a major factor in the ability of the IL salt to dissolve cellulose; however, the cation also appears to have a significant influence. Our previous simulations showed that imidazolium interacts with cellulose through hydrophobic interactions.19 NMR studies also show that the aromatic hydrogen of cellulose associated with the carbon atom between the two nitrogen atoms of the imidazolium ring forms hydrogen bonds with the oxygen atoms of the hydroxyl groups of glucose.20 With the same cation, the order of solubility of cellulose in various ILs is [(CH3CH2O)2PO2] > [OAc] > [SHCH2COO] > [HCOO] > Cl > Br = [SCN] ,1,21 which correlates well with the strength of the respective hydrogen bonds present. Upon addition of the water antisolvent, the dissolved cellulose was regenerated in amorphous and cellulose II forms, indicating that the majority of the cellulose was completely dissolved Received: December 10, 2010 Revised: April 12, 2011 Published: August 09, 2011 10251

dx.doi.org/10.1021/jp111738q | J. Phys. Chem. B 2011, 115, 10251–10258

The Journal of Physical Chemistry B in the IL solution.21,22 Although there are a significant number of studies and proposals about the mechanism of cellulose dissolution in IL, to date there have been only a few papers on the cellulose regeneration mechanism, which is the driving force for separating cellulose from IL upon the addition of an antisolvent. Mazza et al. investigated the influence of water on the dissolution of cellulose and determined the minimum amount of water (up to 21% in weight of water/[C2mim]Cl with a 2% cellulose loading at 90 °C) in the cellulose IL solution initiating cellulose precipitation.23 To better understand the fundamental interactions among [C2mim][OAc], water, and cellulose, and to provide insights for the optimization of the IL pretreatment process and the development of efficient IL recovery and recycle techniques, we investigated the molecular level interactions of binary and ternary systems of [C2mim][OAc] with cellulose and water and the factors that govern the phase behavior of ionic liquids with water and cellulose. Recently, significant attention has been devoted to the study of binary aqueous solutions of room-temperature ionic liquids (RTILs), both experimentally24 27 and computationally.28 32 The presence of water in ILs affects many of their solvent properties, and the water content is influenced by the nature of the cation and anion. Cammarata et al.33 addressed the molecular state of water in ILs using infrared spectroscopy. Their results show formation of H-bonds between water and anions, which increased in the order of [PF6] < [SbF6] < [BF4] < [(CF3SO2)2N] < [ClO4] < [CF3SO3] < [NO3] < [CF3CO2] ,33,34 and liquid-like water aggregation in the IL water mixture. These studies provide insights into the macroscopic properties of ILs in water, but it is also of interest to examine the intermolecular interactions that give rise to these bulk properties and how these interactions relate to the recovery of ILs, which is an important step to industrialize the IL pretreatment process. To this end, the current work provides some fundamental insights into the molecular level behavior of one particular IL ([C2mim][OAc]) in aqueous solutions at various concentrations. To date, experimental studies on cellulose/IL/antisolvent ternary systems have not been conclusive, and a mechanistic understanding of the interactions between water, IL, and cellulose remains unknown. Various computational methods, such as molecular dynamics simulations,19,35 COSMO-RS,36 and quantum mechanics calculations,37 have been used to study the interaction of carbohydrates with solvents, such as water and ILs. These types of simulations have been performed to study the dissolution process of biomass in ionic liquid.19,38 The conformation of soluble cellulose oligomers in water35 and in ionic liquids19 has also been investigated and highlights the key role played by the solvent on shifting the conformational preference of the oligosaccharides with respect to vacuum and crystals. In this paper, we report on molecular dynamics simulations of [C2mim][OAc]/water binary mixtures and of cellulose/[C2mim][OAc]/water ternary mixtures. MD simulation is a powerful tool for investigating these interactions as it provides a detailed analysis of interactions among individual atoms in the binary and ternary systems and how these interactions are altered by changes in the concentrations of its three components, such as the addition of water. Our analyses of the interactions between [C2mim][OAc] and water, and [C2mim][OAc] and sugar, and among [C2mim][OAc] anions and cations was aimed at answering several questions: (1) What are the structures and dynamics of the complex binary and ternary systems, (2) what are the important interactions between [C2mim][OAc] that promote dissolution of cellulose, and (3) how does water act as an antisolvent and disrupt IL cellulose interactions?

ARTICLE

Figure 1. Molecular structure of [C2mim][OAc] and glucose. C6 at the end of the alkyl side chain in the nonpolar region was used to represent the nonpolar component of the cation, C4 positioned between the nitrogen atoms of the imidazolium ring in the charged part of the cation was used to represent the polar region of the cation, and C2 of the acetate anion was used to represent the anion.

’ METHOD Molecular dynamics simulations were used to study the structure and dynamics of binary [C2mim][OAc]/water solvent systems and of ternary systems consisting of a cellulose oligomer (n = 20) dissolved in [C2mim][OAc]/water binary solvent mixtures to study the cellulose regeneration process after [C2mim][OAc] pretreatment. Molecular dynamics simulations were implemented in AMBER,39 using an all-atom force field based on the generalized AMBER force field (GAFF)40 for [C2mim][OAc]19 (Figure 1), the standard TIP3P water model41 for the water, and the carbohydrate force field, GLYCAM,42 for the cellulose oligomer. The cutoff radius of nonbonded interactions was set to 12 Å. The particle mesh Ewald summation method43 was used to calculate the electrostatic potential with periodic boundary conditions. The initial configurations were energy-minimized using the steepest descent algorithm for 1000 steps to remove any unexpected coordinate collisions. The systems were then heated for 200 ps in the NVT ensemble, during which the temperature was increased gradually to the target temperature, followed by a further 500 ps of equilibration dynamics in the NPT ensemble using a Nose Hoover constant pressure (P = 1 bar) control algorithm. Finally, 12 ns production runs were carried out in the NPT ensemble. The time step for production runs was 2 fs, and the SHAKE algorithm was employed to constrain bonds and angles involving hydrogen atoms. The last 10 ns of each trajectory was used for analysis. The simulation box sizes, types of molecules and their numbers, and related parameters for the various systems are given in Table 1. MD simulations were performed on three binary mixtures of [C2mim][OAc] and water at weight fractions of [C2mim][OAc] of 25, 33, and 50%, which equate to [C2mim][OAc] mole fractions of IL of 3, 5, and 10%, respectively (Table 1). MD simulations of a single cellulose 20-mer dissolved in [C2mim][OAc]/water binary mixtures at these same ratios of ionic liquid to water were also performed to study the interactions between the polysaccharide and the solvent system. The structural properties of the mixtures and interactions among [C2mim][OAc], water, and cellulose were examined using radial distribution functions calculated over the trajectories of 10 ns simulations. 10252

dx.doi.org/10.1021/jp111738q |J. Phys. Chem. B 2011, 115, 10251–10258

The Journal of Physical Chemistry B

ARTICLE

Table 1. Simulated Systems of Our Binary and Ternary Mixtures and Simulated Density system

waters

ion pairs

polyglucose chains (DP20)

solvent concentration

solvent concentration

simulation box size

density

(mol % of IL)

(wt % of IL)

(Å Å Å)

(g/cm 3)

1

4549

148

0

3

25

117.3 37.3 39.5

1.03

2

3709

180

0

5

33

112.4 36.1 38.2

1.05

3

1998

222

0

10

50

100.1 32.5 34.8

1.08

4

4549

148

1

3

25

125.3 36.2 38.8

1.04

5

1998

222

1

10

50

109.0 31.6 33.8

1.10

Figure 2. Radial distribution functions of the polar (a) and nonpolar (b) networks in [C2mim][OAc]/water mixtures.

It should be noted that our MD simulations are equilibrium state calculations: the polysaccharide was placed in a preequilibrated binary solvent system. Analysis of the calculated H-bond patterns and the RDF at different concentrations provide structural insight in the phase separation step, but we did not perform a full MD simulation to model the kinetic of the phase separation process. The H-bond strength was estimated by taking the dimer complex of the donor and acceptor to calculate the interaction energy using the force field parameters.

’ RESULTS AND DISCUSSION To investigate the interactions of [C2mim][OAc] with itself, water, and cellulose, we analyzed radial distribution functions for three sites present on [C2mim][OAc] (Figure 1) that were selected to represent different chemical environments of the IL: C6 at the end of the alkyl side chain in the nonpolar region to represent the nonpolar component of the cation, C4 positioned between the nitrogen atoms of the imidazolium ring in the charged part of the cation to represent the polar region of the cation, and C2 of the acetate anion to represent the anion. These three sites were carefully selected to represent each of the three chemically different environments of the [C2mim][OAc]. The coordinates of the chosen atoms are close to the center of mass for each of the three regions, and the RDFs capture the details of the spatial distribution and ordering of atoms around the regions of interest that are due to the interactions present. For example, the charged region of the cation is represented by the C4 atom between the two N atoms present in the imidazolium ring. Because the acidic hydrogen atom in the imidazolium ring is H9, which directly bonds to C4 (Figure 1), the RDF of C4 with water

or the acetate anion can be used to understand the hydrogen bond interactions between the charged region of the cation (imidazolium ring) and acetate and/or water. The polar and the nonpolar networks of [C2mim][OAc] in binary solutions as a function of [C2mim][OAc] concentration were analyzed by generating RDFs of the self-interactions of the IL. The calculations (Figure 2) indicate the existence of strong polar cation anion networks in the pure ionic liquid, with a first interaction shell centered at the sharp peak at around 4 Å, and a second shell centered at the sharp peak at around 6 Å. With the addition of water, the strong cation anion interaction network in the pure IL is disrupted, as indicated by the weaker intensity and broadening of the first peak that is shifted to around 4.2 Å. The first peak of the cation anion interaction decreases with increasing water and returns to unity at 50% ionic liquid. These results indicate that the ordered cation anion polar interaction network is disrupted by water that forms a network of interactions with the anion (Figure 3a) and with the cation (Figure 3b). The second sharp peak at ∼6 Å in Figure 2a becomes broader with increasing water content and is near unity, indicating that the IL is randomly distributed. The nonpolar nonpolar interaction network is also disrupted, as evidenced by the shift from a peak at ∼4 Å toward a more uniform distribution (Figure 2b). Overall, these results show that the increasing water concentration disorders the [C2mim][OAc] by disrupting both the polar and the nonpolar networks between cations and anions. Interactions between the water and [C2mim][OAc] were also analyzed. Radial distribution functions calculated for water oxygen atoms with the two polar sites on the ionic liquid, as well as water, are provided in Figure 3. These results show that water makes short-range interactions with all three regions of [C2mim][OAc], interacting most strongly with the acetate ion 10253



ARTICLE

(Figure 3a), which has an intense sharp peak in its g(r) at 3.8 Å. The sharp, but small, C4 O peak at 2.8 Å, primarily resulting

Figure 3. Radial distribution functions of water with different regions of IL and other water molecules in [C2mim][OAc]/water mixtures (a) with the anion, (b) with the polar region of the cation, and (c) with water.

from hydrogen-bonding interactions between the acidic hydrogen atom H9 present and the oxygen atom of water, indicates that water also has weak interactions with the polar region of the cationic imidazolium ring. Within the range of water concentrations studied (90 97% water mole fractions), increasing water (diluting [C2mim][OAc]) decreases the strength of the interaction between water and the ionic liquid, as indicated by the reduced intensity of the first water anion and water cation g(r) peaks (Figure 3a,b). The water structure is also affected by the presence of the ionic liquid. As shown in Figure 3c, the water water peaks become more refined and prominent with increasing [C2mim][OAc] concentration, indicating that the presence of [C2mim][OAc] increases the local ordering of the water network. Overall, for a binary mixture of [C2mim][OAc] and water, our calculations show that increasing the ionic liquid concentration increases the local ordering of the water water, water anion, and water cation networks. The density of the system also increases slightly with increasing ionic liquid concentration (shown in Table 1). The role of water in ionic liquids is complex and depends on the supramolecular structures present. Ionic liquids containing dissolved water may not be regarded as homogeneous solvents but have to be considered as nanostructured networks of [C2mim][OAc].28 Our simulations indicate that, as water is added to [C2mim][OAc], both water cation and water anion networks form and small water clusters are also present. Cammarata et al.33 have shown that water can form liquid-like associated aggregates when it is absorbed from the air into RTILs with relatively strong anions, such as [NO3] and [CF3CO2] , and a previous report from MD simulations also confirms the formation of a heterogeneous water network with an imidazolium-based IL with the anion [NO3] 1.28 Another MD simulation31 of [Bmim]Cl solution with water showed that, at low concentrations, water exists as small clusters and, at high concentrations, as a continuous network. It seems that a small anion, such as Cl , may disrupt the water network by breaking the water water hydrogen bonds. Our simulations suggest that, with anion [OAc] , a heterogeneous water network is formed. We found that the anion [OAc] , which is a larger anion and stronger H-bonding acceptor with water, similar to the case of [NO3] , enhanced the H-bond network of liquid water. In addition to determining the RDFs, the numbers of hydrogen bonds per water molecule in the binary mixtures at different concentrations were calculated. Water molecules can form hydrogen bonds with acetate (at the O1 and O2 positions depicted in Figure 1), with imidazolium (at H9 in Figure 1) and with other water molecules. Hydrogen bonds of O H 3 3 3 O were defined as having an O 3 3 3 O distance of less than 3.5 Å and an O H 3 3 3 O angle larger than 150°, whereas hydrogen bonds of C H 3 3 3 O were defined as having a C 3 3 3 O distance of less than 4.0 Å and a C H 3 3 3 O angle larger than 150°, as suggested by ref 44. Table 2 summarizes the average number of hydrogen bonds present over 10 ns simulations. The number of hydrogen bonds per water molecule is 0.71, 0.80, and 0.97 for [C2mim][OAc] concentrations of 25, 33, and 50%, respectively.

Table 2. Average Hydrogen Bonds for the [C2mim][OAc]/Water Binary Systemsa IL weight

total_HB

W A

W C

W W

num of water

num of IL

HB/ACT

HB/wat

25%

3223.62

721.33

30.42

2471.88

4549

148

4.873818

0.71

33%

2948.69

896.51

35.47

2016.70

3709

180

4.98063

0.80

50%

1947.71

1002.56

37.04

908.11

1998

222

4.516036

0.97

a

The total hydrogen bonds (HB) for the binary systems involving water molecules are contributed by water water (W W), water actate anion (W A), and water imidazolium cation (W C) interactions. 10254


The Journal of Physical Chemistry B The increase in hydrogen bonds with increasing [C2mim][OAc] is primarily ascribed to the increased hydrogen bonds between acetate anions and water molecules (Table 2). The calculated increase in the ordered water present as a function of increasing [C2mim][OAc] concentration is hypothesized to be due to the solvation effect of the anions and cations present (especially [OAc]). Figure 4 provides a snapshot of a binary 15% [C2mim] [OAc] and water system and depicts [OAc] forming five hydrogen bonds with water molecules and the water molecules forming small water water clusters. The strong [OAc] water

Figure 4. Snapshot of simulated binary systems.

ARTICLE

interaction seems to be the cause of increasing the extent of the ordered water present. Experimental data7 indicate that, when greater than 50 wt % water is added to cellulose dissolved in [C2mim][OAc], precipitation is clearly observed and is almost complete when 80 wt % water is added. To investigate the interactions among water, cellulose, and [C2mim] [OAc] and the changes in these interactions that potentially initiate cellulose precipitation, we simulated a polyglucose chain (n = 20) in pure [C2mim][OAc] and in binary mixtures of water and [C2mim][OAc] having weight fractions of [C2mim][OAc] of 25% and 50%. We then compared changes in the molecular level interactions between the polyglucose and [C2mim][OAc] as a function of water content by comparing the RDFs of these three systems and by counting hydrogen bonds. The RDFs of the anion (O1 and O2 atoms of acetate) and the cation (C4 atom of [C2mim]) of the ionic liquid and of water with respect to cellulose hydroxyl groups (O2, O3, and O6 of the glucose unit shown in Figure 1) are presented in Figure 5. As shown in Figure 5a, as the water content is increased, the sharp first peak near 2.8 Å increased, indicating that the probability of the anions being located in the first shell around cellulose increases, suggesting that, at lower [C2mim] [OAc] concentrations, the anions have stronger interactions with cellulose. Integration of the sugar anion RDF up to 3.5 Å shows that there are, on average, about 0.365 and 0.675 acetate anions within the 3.5 Å of each hydroxyl group of a sugar molecule in 25% and 50% water [C2mim][OAc] solutions. Considering the concentration effect, the RDF corresponds to the H-bonding strength between the

Figure 5. Radial distribution functions of hydroxyl groups of glucoses with (a) the anion, (b) the cation polar region, and (c) water in the [C2mim][OAc]/water/polysaccharide chain ternary system. 10255



ARTICLE

hydroxyl groups of cellulose and acetate. In contrast, as the water content increases, the [C2mim]+ ions are displaced outside the first solvation shell of cellulose, with a smaller peak observed at the first solvation shell at a radius of 3.3 Å, and a larger and broader peak at the second solvation shell at 5.8 Å (Figure 5b). In the ternary systems, water molecules diffuse into the surface of the sugar chain and form H-bonds with the hydroxyl groups of sugar molecules, as evidenced by the sharp peak around 3.0 Å in Figure 5c. In addition to the RDF analysis, the hydrogen bond patterns between the hydroxyl groups of the polyglucose, acetate anion, and water were analyzed and are presented in Table 3. Water can form hydrogen bonds with the sugar both as a donor and as an acceptor. At 100% [C2mim][OAc], the average number of hydrogen bonds between the cellulose and the anion is calculated to be 50.0.19,45 When the water concentration is increased to 50%, the number of hydrogen bonds calculated decreases to 31.4, and decreases to 14.8 with 75% water. The hydrogen bond counts are in a good agreement with the integral of the anion RDF, considering that the sugar chain has a degree of polymerization of 20 and each glucose unit has three free hydroxyl groups. The average number of polyglucose water hydrogen bonds (as donor) increases from 14.6 at 50% to 23.0 at 75% water, whereas the average number of glucose water hydrogen bonds (as acceptor) increases from 28.8 at 50% to 37.2 at 75% water. We hypothesize that, at higher water concentrations, the Table 3. Average Hydrogen Bonds Involving the Hydroxyl Group of Sugar and Solventa IL

sugar A

25%

14.8

23

37.2

50%

31.4

14.6

28.8

100% a

50

sugar W(donor)

sugar W(acceptor)

0

Hydrogen bonds are contributed by sugar with acetate anion (Sugar A), sugar with water while water molecules as hydrogen bond donors(sugar W(donor)), and sugar with water while water molecules as hydrogen bond acceptors(sugar W(acceptor)).

surface of the cellulose experiences a higher degree of ordering and forms a hydrogen-bonding network with the water and [OAc] present. On the basis of the changes we see in the structure of [C2mim][OAc] and of water around the cellulose 20-mer and the changes in the number of hydrogen-bonding interactions between [OAc] and cellulose, and between water and cellulose at three concentrations of the IL, we propose key intermediate steps and structures for the precipitation process of cellulose (Figure 6) as water is added to [C2mim][OAc]. These key intermediate motifs are observed in our simulation trajectory. At high concentrations of [C2mim][OAc], the anion has a very high probability of residing in a solvation shell centered at about 2.3 Å, and the anions form a large number of hydrogen bonds with the cellulose (Table 3). As water diffuses inside the first solvation shell of cellulose, the number of hydrogen bonds among water molecules and the polymer increases and the number of H-bonds between the anion and the sugar decreases (presumably because those anions forming H-bonds with the sugar are now pushed out to the first solvation shell), suggesting the formation of an anion water cellulose H-bonding network. The formation of this network displaces the cation out of the first solvation shell and leads to cellulose precipitation. In this model, phase separation of the cellulose as a function of water results from competitions between the water anion, cellulose anion, cellulose water, cellulose cation, and anion cation interactions. To determine the relative strengths of these interactions, we calculated their respective hydrogen bond strength. The glucose acetate (12 kcal/mol) and water acetate (15 kcal/mol) have similar hydrogen bond strengths, whereas the glucose glucose (4 kcal/mol) has a slightly higher H-bond strength than glucose water (3.5 kcal/mol) and water water hydrogen bonds (about 3.5 kcal/mol). Our previous simulation results, based on energy decomposition, indicated that the immidazolium cation primarily interacts with cellulose through van der Waals interactions.19 This interaction accounts for one-quarter of the total interactions between cellulose and [C2mim][OAc]. When dissolved in a pure ionic liquid, the average hydrogen-bond ratio between hydroxyl

Figure 6. Proposed key intermediate structure of cellulose precipitation pathway from ionic liquids in the presence of an antisolvent. 10256


The Journal of Physical Chemistry B groups of cellulose and [OAc] is about 1:1. Each [OAc] has a free lone pair electron in the oxygen atom that can form hydrogen bonds with other electron acceptors, such as the hydrogen atom in water molecules. As the water content increases, water molecules diffuse into the anion sugar network and the strong water anion interactions (Figure 2a, Table 2, and Figure 5) saturate the H-bonding ability of the anions. At the same time, water forms H-bonds with cellulose, acting as both a H-bond donor and a H-bond acceptor. This network pushes the previous wellbalanced cations out of the first solvation shell of cellulose. However, the cations are maintained in the second solvation shell of cellulose due to their strong electrostatic interactions with the anions present (Figure 5b).

’ CONCLUSIONS MD simulations were carried out to study the structural and dynamic properties of binary and ternary solutions of [C2mim][OAc], water, and a 20-mer of cellulose. The MD simulations demonstrate that the introduction of water over a set concentration range induces changes in the structural organization of [C2mim][OAc] and disrupts the interactions between [C2mim][OAc] and cellulose. In a binary [C2mim][OAc]/ water solution, as the concentration of water increases, the hydrogen bonding between the water and anions becomes saturated and results in the water forming liquid-like aggregates. When added to the [C2mim][OAc] and cellulose mixture, water molecules diffuse within the cellulose surface and saturate both the anion and the cellulose hydroxyl groups by forming H-bonding networks and displace the previous well-balanced cations out of the first solvation shell of cellulose. We propose that this is a key intermediate step during cellulose precipitation because of the strong ion ion interaction between the cation and the anion in the ionic liquid. These MD simulations provide a sound molecular interpretation of previous experimental results1,7 of the complex fluidphase behavior of the cellulose/IL/water system. The proposed key structural elements in the cellulose precipitation pathway enable the identification and use of the underlying key factors in the miscibility of cellulose with ionic liquids and water. This work provides information that links the molecular mechanism of [C2mim][OAc] interactions with water and cellulose to the macroscopic properties and phenomenon of cellulose precipitation by addition of water after IL pretreatment. The model of cellulose interactions with [C2mim][OAc] in the presence of water provides a framework for evaluating other novel anion cation pairs of ILs computationally that may effectively and efficiently dissolve cellulose and lignocellulose. This work provides a fundamental understanding of water as an antisolvent in regeneration processes, which is critical to research efforts to develop an optimal ionic liquid for biomass pretreatment. ’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT This work was part of the DOE Joint BioEnergy Institute (http://www.jbei.org) supported by the U.S. Department of Energy, Office of Science, Office of Biological and Environmental Research, through Contract DE-AC02-05CH11231 between

ARTICLE

Lawrence Berkeley National Laboratory and the U.S. Department of Energy. This research used resources of the National Energy Research Scientific Computing Center (NERSC). We thank Dr. Harvey Blanch for his advice and for reviewing the manuscript and Drs. Brad Holmes, Ning Sun, and Gang Cheng for their discussion on this topic.

’ REFERENCES (1) Swatloski, R. P.; Spear, S. K.; Holbrey, J. D.; Rogers, R. D. J. Am. Chem. Soc. 2002, 124, 4974. (2) Swatloski, R. P.; Rogers, R. D.; Holbrey, J. D. WO Patent 03/ 029329, 2003. (3) Dadi, A. P.; Varanasi, S.; Schall, C. A. Biotechnol. Bioeng. 2006, 95, 904. (4) Zhang, H.; Wu, J.; Zhang, J.; He, J. S. Macromolecules 2005, 38, 8272. (5) Zhu, S. D.; Wu, Y. X.; Chen, Q. M.; Yu, Z. N.; Wang, C. W.; Jin, S. W.; Ding, Y. G.; Wu, G. Green Chem. 2006, 8, 325. (6) Dadi, A. P.; Schall, C. A.; Varanasi, S. Appl. Biochem. Biotechnol. 2007, 137, 407. (7) Li, C.; Knierim, B.; Manisseri, C.; Arora, R.; Scheller, H. V.; Auer, M.; Vogel, K. P.; Simmons, B. A.; Singh, S. Bioresour. Technol. 2010, 101, 4900. (8) Barthel, S.; Heinze, T. Green Chem. 2006, 8, 301. (9) Xie, H. L.; Shi, T. J. Holzforschung 2006, 60, 509. (10) Fukaya, Y.; Hayashi, K.; Wada, M.; Ohno, H. Green Chem. 2008, 10, 44. (11) Zhao, H.; Baker, G. A.; Song, Z. Y.; Olubajo, O.; Crittle, T.; Peters, D. Green Chem. 2008, 10, 696. (12) Fort, D. A.; Remsing, R. C.; Swatloski, R. P.; Moyna, P.; Moyna, G.; Rogers, R. D. Green Chem. 2007, 9, 63. (13) Kilpelainen, I.; Xie, H.; King, A.; Granstrom, M.; Heikkinen, S.; Argyropoulos, D. S. J. Agric. Food Chem. 2007, 55, 9142. (14) Pu, Y. Q.; Jiang, N.; Ragauskas, A. J. J. Wood Chem. Technol. 2007, 27, 23. (15) Zavrel, M.; Bross, D.; Funke, M.; Buchs, J.; Spiess, A. C. Bioresour. Technol. 2009, 100, 2580. (16) El Seoud, O. A.; Heinze, T. Polysaccharides 1: Structure, Characterization and Use; Springer-Verlag: Berlin, 2005; Vol. 186, p 103. (17) El Seoud, O. A.; Koschella, A.; Fidale, L. C.; Dorn, S.; Heinze, T. Biomacromolecules 2007, 8, 2629. (18) Sun, N.; Rodriguez, H.; Rahman, M.; Rogers, R. D. Chem. Commun. (Cambridge, U.K.) 2011, 47, 1405. (19) Liu, H.; Sale, K. L.; Holmes, B. M.; Simmons, B. A.; Singh, S. J. Phys. Chem. B 2010, 114, 4293. (20) Zhang, J. M.; Zhang, H.; Wu, J.; Zhang, J.; He, J. S.; Xiang, J. F. Phys. Chem. Chem. Phys. 2010, 12, 1941. (21) Klemm, D.; Heublein, B.; Fink, H. P.; Bohn, A. Angew. Chem., Int. Ed. 2005, 44, 3358. (22) Cheng, G.; Varanasi, P.; Li, C.; Liu, H.; Melnichenko, Y. B.; Simmons, B. A.; Kent, M. S.; Singh, S. Biomacromolecules 2011, 12, 933. (23) Mazza, M.; Catana, D.-A.; Vaca-Garcia, C.; Cecutti, C. Cellulose 2009, 16, 207. (24) Mele, A.; Tran, C. D.; Lacerda, S. H. D. Angew. Chem., Int. Ed. 2003, 42, 4364. (25) Tran, C. D.; Lacerda, S. H. D.; Oliveira, D. Appl. Spectrosc. 2003, 57, 152. (26) Danten, Y.; Cabaco, M. I.; Besnard, M. J. Mol. Liq. 2010, 153, 57. (27) Xu, Y. J.; Gao, Y.; Zhang, L. Q.; Yao, J.; Wang, C. M.; Li, H. R. Sci. China, Ser. B: Chem. 2010, 53, 1561. (28) Jiang, W.; Wang, Y.; Voth, G. A. J. Phys. Chem. B 2007, 111, 4812. (29) Canongia Lopes, J. N.; Padua, A. A. J. Phys. Chem. B 2006, 110, 3330. I (30) Klahn, M.; Stu^ber, C.; Seduraman, A.; Wu, P. J. Phys. Chem. B 2010, 114, 2856. 10257



ARTICLE

(31) Hanke, C. G.; Lynden-Bell, R. M. J. Phys. Chem. B 2003, 107, 10873. (32) Lynden-Bell, R. M.; Del Popolo, M. G.; Youngs, T. G. A.; Kohanoff, J.; Hanke, C. G.; Harper, J. B.; Pinilla, C. C. Acc. Chem. Res. 2007, 40, 1138. (33) Cammarata, L.; Kazarian, S. G.; Salter, P. A.; Welton, T. Phys. Chem. Chem. Phys. 2001, 3, 5192. (34) Koddermann, T.; Wertz, C.; Heintz, A.; Ludwig, R. Angew. Chem., Int. Ed. 2006, 45, 3697. (35) Shen, T.; Langan, P.; French, A. D.; Johnson, G. P.; Gnanakaran, S. J. Am. Chem. Soc. 2009, 131, 14786. (36) Kahlen, J.; Masuch, K.; Leonhard, K. Green Chem. 2010, 12, 2172. (37) Kirschner, K. N.; Woods, R. J. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 10541. (38) Derecskei, B.; Derecskei-Kovacs, A. Mol. Simul. 2006, 32, 109. (39) Case, D. A.; Darden, T. A.; Cheatham, I. T. E.; Simmerling, C. L.; Wang, J. AMBER 9; University of California: San Francisco, CA, 2006. (40) Wang, J. M.; Wang, W.; Kollman, P. A.; Case, D. A. J. Mol. Graphics Modell. 2006, 25, 247. (41) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L. J. Chem. Phys. 1983, 79, 926. (42) Woods, R. J.; Dwek, R. A.; Edge, C. J.; Fraserreid, B. J. Phys. Chem. 1995, 99, 3832. (43) Darden, T.; York, D.; Pedersen, L. J. Chem. Phys. 1993, 98, 10089. (44) Koch, U.; Popelier, P. L. A. J. Phys. Chem. 1995, 99, 9747. (45) Liu, H.; Sale, K. L.; Holmes, B. M.; Simmons, B. A.; Singh, S. J. Phys. Chem. B 2010, 114, 4293.

10258


Molecular Dynamics Study of Polysaccharides in Binary Solvent

Recommend Documents