Cellulose Dissolution in Ionic Liquid: Ion Binding Revealed by Neutron

1 day ago - Support. Get Help · For Advertisers · Institutional Sales; Live Chat. Partners. Atypon · CHORUS · COPE · COUNTER · CrossRef · CrossCheck ...
0 downloads 0 Views 2MB Size
Download PDF
Article Cite This: Macromolecules XXXX, XXX, XXX−XXX

pubs.acs.org/Macromolecules

Cellulose Dissolution in Ionic Liquid: Ion Binding Revealed by Neutron Scattering Vikram Singh Raghuwanshi,† Yachin Cohen,*,‡ Guillaume Garnier,† Christopher J. Garvey,§ Robert A. Russell,§ Tamim Darwish,§ and Gil Garnier*,† †

Macromolecules Downloaded from pubs.acs.org by WESTERN UNIV on 09/21/18. For personal use only.

Bioresource Processing Research Institute of Australia (BioPRIA), Department of Chemical Engineering, Monash University, Clayton, Victoria 3800, Australia ‡ Department of Chemical Engineering, Technion-Israel Institute of Technology, Haifa 3200003, Israel § Australian Nuclear Science and Technology Organisation (ANSTO), New Illawarra Rd., Lucas Heights, NSW 2234, Australia S Supporting Information *

ABSTRACT: Dissolution of cellulose in 1-ethyl-3-methylimidazolium acetate (EMIMAc) ionic liquid (IL) was investigated by small-angle neutron scattering (SANS) with contrast variation. Cellulose and EMIMAc of different deuteration levels provide sufficient contrast in revealing the cellulose dissolution processes. Two experiments were performed: hydrogenated microcrystalline cellulose (MCC) was dissolved in deuterated IL (IL-D14), and deuterated bacterial cellulose (DBC) was dissolved in hydrogenated IL (IL-H14). Contrary to the expectation of high contrast between MCC and IL-D14, a dramatic reduction of the measured intensity (scattering cross section) was observed, about 1/3 of the value predicted based on the scattering length density (SLD) difference. This is attributed to the tight binding of acetate ions to the cellulose chains, which reduces the SLD difference. Measurements using small-angle X-ray scattering (SAXS) corroborate this effect by indicating increased contrast due to ion adsorption resulting in enhanced SLD difference. The experiments performed with DBC dissolution in IL-H14 suggest the presence of fractal aggregates of the dissolved cellulose, indicating lower solubility compared to the MCC. Contrast variation SANS measurements highlight tight ion binding of at least one acetate ion per anhydroglucose unit (AGU). EMIMAc is a successful cellulose solvent, as in addition to disrupting intermolecular hydrogen bonding, it imparts effective charge to the cellulose chains hindering their agglomeration in solution.

■

INTRODUCTION Cellulose is the most abundant and renewable polymer; it is used in many industrial and technological applications ranging from biomedical to nanomaterials and encompassing traditional products in textile and paper. A main hindrance for successful applications has been the lack of an efficient and environmentally benign method to dissolve cellulose chains from their tight binding in the native crystal. The discovery of ionic liquids (IL) as a solvent by Rogers and co-workers has opened novel ways to dissolve cellulose.1 However, applications of IL for cellulose are still limited due to technical and economic issues. The search for new improved solvents is of current interest. It relies, in part, on obtaining a better understanding of the dissolution mechanism.2 It has been suggested that IL ions adsorb onto the cellulose surface, disrupt the interchain hydrogen bonding network within the crystalline fibrillar structure, and release the chains into the IL matrix. Studies by molecular dynamics simulations3 and NMR spectroscopy4 have indicated that dissolution involves formation of strong association of IL anions capable of hydrogen bonding to the hydroxyl groups of cellulose. These © XXXX American Chemical Society

may be augmented by the van der Waals interactions of the conjugated ring structures in IL cations in close contact with sugar rings.5 Recent studies of cellulose/IL solutions by small-angle X-ray scattering (SAXS) provided evidence of a solvation sheath of the IL around the dissolved cellulose chain. SAXS measurements and molecular simulations of cellulose dissolved in a mixture of tetrabutylammonium acetate and dimethyl sulfoxide indicated that approximately one acetate ion binds to each anhydroglucose unit (AGU).6 Another recent SAXS study and simulation on cellulose dissolved in 1-ethyl-3-methylimmidazolium methylphosphonate came to a similar conclusion.7 The SAXS patterns in these studies were analyzed using a persistent worm-like chain model, in which a core− shell structure was evaluated for the electron density profile in a cross section perpendicular to the chain segment, indicating IL solvation of the cellulose chain.6,7 Received: July 4, 2018 Revised: September 5, 2018

A

DOI: 10.1021/acs.macromol.8b01425 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

interaction of IL ions with the cellulose’s surface during the dissolution process and the strength of this interaction. For full evaluation, both microcrystalline cellulose (MCC) dissolved in deuterated EMIMAc (IL-D14) and DBC in hydrogenated EMIMAc (IL-H14) were studied at different temperatures. Understanding cellulose solvation is important in the search of new and more readily available solvents, such as protic ionic liquids,21 and in applications of ILs as solvents for cellulose derivatization22,23 or formation of cellulose-coated emulsions.24

The concept of polymer solvation originated from studies on the importance of the hydration sheath8 formed by water molecules around natural polymers, such as denatured proteins9 or synthetic water-soluble polymers and gels,10 responsible for the unique thermodynamic phenomena termed collectively “hydrophobic effects”, including a lower critical solution temperature (LCST) phase behavior. It is much less common in organic solvents but has been recently observed in IL solutions of poly(ethylene oxide)11 and poly(benzyl methacrylate),12 the latter also exhibiting LCST behavior due to specific interactions of the ring-conjugated IL cations solvating benzyl pendant groups of the polymer chain. The loss of solvent degrees of freedom, due to specific polymer solvation interactions, results in a negative entropy of solution, which is the key to LCST behavior in solvated polymer systems. Although strong solvation interactions are important to render cellulose soluble in IL, these solutions do not exhibit LCST behavior, as opposed to hydroxide-based solutions. Despite the importance of polymer solvation, there is no direct evidence for strong and dense binding in such systems. 1-Ethyl-3-methylimidazolium acetate (EMIMAc) is considered an environmental friendly and safe solvent which is very effective in dissolving cellulose.1,13 Previous SAXS analysis of cellulose/EMIMAc solutions indicated negligible signal from aggregates,14 as was evident in excess scattering at low angles seen in other cellulose/IL systems.6,7 The SAXS patterns from semidilute cellulose/IL solution followed the scaling rules expected in good solvent conditions,14 and a rather large positive osmotic second virial coefficient evaluated by light scattering from cellulose solutions in a mixture of EMIMAc and dimethylformamide indicated strong polymer−solvent interactions.15 Because of the relatively large persistence length of the dissolved cellulose chains, the SAXS patterns from the solutions exhibit rigid-rod-like behavior at intermediate scattering vectors (q), with a characteristic q−1 power law relation of intensity to q.7,14 This allows direct evaluation of the average scattering length density (SLD) of the solvated chains from the absolute value of the scattered intensity in a readily accessible q range. The observation of the IL-solvation effect on the SLD of the dissolved cellulose chains is greatly enhanced in neutron scattering measurements using contrast variation. This may be achieved when either the polymer or solvent is deuterated, exploiting the large difference in the scattering lengths of hydrogen (H) and deuterium (D). The effect of deuteration on the SLD contrast is much more significant than any isotopic effect on the interaction between IL and cellulose, which might be expected following a report on the effect of deuteration on the acidity of carboxylic acids and phenols.16 Deuterated EMIMAc17,18 and deuterated bacterial cellulose (DBC)19,20 have already been reported in neutron scattering and reflectivity studies. However, a thorough comparison of cellulose versus IL deuteration to better visualize the dissolution mechanism has not been previously reported. Furthermore, little is known on whether there is an enrichment of the IL cation/anion at the cellulose/IL interface and the strength of the specific interaction that is responsible for the process of dissolution. The main objective of this study was to evaluate the extent of solvation of dissolved cellulose chains by EMIMAc, directly from the absolute intensity of small-angle neutron scattering (SANS) from unaggregated solutions in a broad and readily accessible q range. A second objective was to elucidate the

■

EXPERIMENTS

Materials. Microcrystalline cellulose (MCC) was purchased from Sigma-Aldrich, USA. Deuterated cellulose (DBC) was prepared by growing Gluconacetobacter xylinus (ATCC 53524) in deuterated medium as reported earlier.19,25 DBC was freeze-dried for 48 h. 1Ethyl-3-methylimidazolium acetate (EMIMAc) ionic liquid (99% purity) was purchased from Sigma-Aldrich. Deuterated EMIMAc was prepared at ANSTO’s National Deuteration Facility. The isotopic purity of the 1-ethyl-3-methylimidazolium cation was determined by mass spectrometry (MS) to be 89 ± 2% D, and the isotopic purity of the acetate anion was determined by a published method using 13C {1H, 2H} nuclear magnetic resonance (NMR) to be 86% D.26 Detailed NMR analysis is provided in the Supporting Information. ILs were heated at 90 °C under vacuum to remove water traces. Cellulose in IL. In the first experiments, MCC (C6H10O5) was mixed with the deuterated EMIMAc (IL-D14). In the second experiment, DBC (C6D7H3O5) with three exchangeable H was mixed with the hydrogenated EMIMAc (IL-H14). The cellulose powder was mixed with IL, placed in the cell (1 mm thickness) and measured at 25 and 90 °C. Small-Angle Neutron Scattering (SANS). SANS measurements were made at the QUOKKA beamline27 of the Australian Nuclear Science and Technology Organization (ANSTO, Sydney, Australia). Samples were measured in the transmission mode with the neutron wavelength, λ = 5 Å (Δλ/λ = 10%). Measurements were conducted at two different sample-to-detector distances (4 and 1.3 m) to cover an adequate range of length scales. Data reduction, converting the raw isotropic 2D scattering patterns and the measurement geometry into the familiar I(q) formalism, where q = 4π sin ϑ/λ is the scattering vector, 2ϑ the scattering angle, and λ the incident wavelength, was made using the Igor-based software.28 The scattering patterns were corrected for the detector sensitivity, the dark current of neutrons and detector noise, and the scattering due to the sample holder and then normalized to the absolute scattering cross section using the incident beam flux. Small-Angle X-ray Scattering (SAXS). SAXS measurements were performed using a Molecular Metrology SAXS/WAXS system equipped with a sealed microfocus tube (MicroMax-002+S) emitting Cu Kα radiation, two Göbel mirrors, three pinhole slits, and a generator that was powered at 45 kV and 0.9 mA. The scattering patterns were recorded by a two-dimensional position-sensitive wire detector (Gabriel) positioned 150 cm behind the sample. The solutions were sealed in thin-walled glass capillaries about 2 mm in diameter and 0.01 mm wall thickness and measured under vacuum at 25 °C. The scattered intensity I(q) was recorded at the interval 0.007 < q < 0.26 Å−1. The scattering intensity was normalized with respect to time, solid angle, primary beam intensity, capillary diameter, and transmission to yield the absolute specific scattering cross section.

■

RESULTS SANS experiments were performed to quantify cellulose solvation by EMIMAc ionic liquid (IL) in the dissolution process, utilizing the way solvation affects the scattering length density (SLD) of the dissolved cellulose chains. The SLD of hydrogenated microcrystalline cellulose (MCC) is ∼1.76 × 10−6 Å−2, and that of deuterated bacterial cellulose (DBC) is B

DOI: 10.1021/acs.macromol.8b01425 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules 7.2 × 10−6 Å−2. The three labile hydroxyl groups on the DBC chain decrease its SLD to ∼5.56 × 10−6 Å−2 having been reverted to hydrogen in the purification processes.29 The SLD of EMIMAc (IL-H14) is 1.14 × 10−6 Å−2, and that of fully deuterated IL-D14 is 6.07 × 10−6 Å−2. Table 1 gives the

ment time. Furthermore, it was not possible to obtain a statistically significant measurement at low q, requiring a 20 m setting of the detector distance from the sample. Another observation is that the scattering intensity is lower for the solution measured at 90 °C. Further information from the SANS curves was extracted by fitting the data in the experimentally accessible q range (0.01−0.3 Å−1). As shown in Figure 1, this can be fit by the power law

Table 1. Molecular Weight, Density, and Scattering Length Density (SLD) of Hydrogenated Cellulose, Deuterated Cellulose (Labile Hs), Hydrogenated EMIMAc, and Deuterated EMIMAc material

notation

Mw (g/mol)

density (g/cm3)

EMIMAc

IL_H14 IL_D14 MCC DBC

170.2 184.2 162.1 169.1

1.03 1.12 1.50 1.56

cellulose

I(q) = I0q−1 + b

where I0 is the prefactor of the power-law relation and b represents the background scattering due to the IL solvent. The q−1 relation of intensity to scattering vector is characteristic of rod-like entities and is typically observed in the range of scattering vectors (1/L) < q < (1/Rc), where L and Rc represent the length and cross-section dimensions of the rod. In this case, the prefactor, I0 is given by30

SLD (ρ ) (×10−6 Å−2) 1.14 6.07 1.76 5.56

(C8H14N2O2) (C8D14N2O2) (C6H10O5) (C6D7H3O5)

iN y I0 = jjj zzzπLA2 Δρ2 = φπAΔρ2 kV {

molecular weight, density, and the calculated SLD from the molecular formula. Table 2 presents the calculated SLD differences (contrast) Δρtheo that may be achieved theoretically between the different materials.

SLD difference Δρ (× 10−6 Å−2)

Δρ2 (×10−12 Å−4)

cellulose_H-IL_H14 cellulose_H-IL_D14 cellulose_D-IL_H14 cellulose_D-IL_D14

0.63 4.31 4.42 0.51

0.40 18.6 19.5 0.26

(2)

(N )

where V is the number of rod-like particles per unit volume, L, A, and φ are the rod length, cross-section area, and volume fraction, respectively, and Δρ is the SLD difference between the rod and its surrounding medium. Equations 1 and 2 are valid also for a rod-like structure with a nonuniform cross section, such as a core−shell cylinder, when q ≪ 1/Rc (Rc being the radius of gyration of the cross section).31 The conformation of cellulose dissolved in ionic liquids was shown to conform to a worm-like chain with a significant persistence length (Lp)6,7 so that the power law of eq 1 is followed in the relevant q range.6,7,14,15 The surprisingly low scattering cross section in the SANS patterns from the solution of 3 wt % MCC in IL-D14, hence the low value of I0, may be due to different causes. An obvious possibility is that not all the cellulose is molecularly dissolved and that the scattering from nonexfoliated crystals has a negligible contribution in this relevant q range. This possibility is discounted as MCC readily dissolves in EMIMAc into a clear solution with no visible indication of aggregates. We propose that the low scattering cross section is due to the strong binding of IL ions to the cellulose interface. As the measured data do not have a sufficient signal above the background at the high-q range, it cannot provide resolution of the crosssection structure. However, the measured power-law prefactor allows assessment of the extent of solvation by considering the average SLD of the cellulose chain with its bound IL molecules. We can consider two extreme cases. If the molar volume of the adsorbed ions is unchanged, so that they contribute to the increase in cross-section area and the volume fraction commensurate with the reduction in the contrast, then IL adsorption should not affect significantly the total scattering cross section. Obviously, this is not what the measurements indicate. On the other hand, tight binding of the ions, in a manner that does not significantly affect the cross-section area, will result in significant reduction of contrast. Although this case may be unrealistic, it can provide a lower-limit estimate to the extent of adsorbed IL ions solubilizing the cellulose chain. The values of I0 obtained by fitting eq 1 to the data in Figure 1 are used to calculate the experimental SLD difference Δρexp between the cellulose chains and the IL solvent using eq 2. Here we use constant values of φ ≈ 0.02 (calculated from 3 wt % MCC in IL-D14) and a cross-section area A of 32 Å2 as

Table 2. Theoretical SLD Difference Calculated for Hydrogenated and Deuterated Cellulose, Hydrogenated Ionic Liquid, and Deuterated Ionic Liquid cellulose in IL solvents

(1)

Hydrogenated Cellulose (MCC) Dissolved in Deuterated EMIMAc (IL-D14). Figure 1 shows the SANS curves for 3 wt % MCC dissolved in IL-D14 solvent, performed at 25 and 90 °C. The first striking observation is the low value of the absolute scattering intensity (specific scattering cross section) which results in low signal-to-noise observed in the measure-

Figure 1. SANS curves for 3 wt % hydrogenated microcrystalline cellulose (MCC) dissolved in deuterated ionic liquid (IL-D14) at 25 °C (open square) and 90 °C (open star). SANS curves fitted with the power law (eq 1) are shown by the solid lines. C

DOI: 10.1021/acs.macromol.8b01425 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules given for a chain in the cellulose I crystal structure.32 The evaluated parameters for the 3 wt % MCC/IL-D14 solution at different temperatures are given in Table 3. The Δρexp value at

using eq 2. Inasmuch as the adsorbed ions occupy a more significant volume around the dissolved cellulose chains (y > 1), a higher adsorption ratio (x) is to be expected to fit the scattering intensities measured experimentally. For example, a molar adsorption ratio of about 2:1 acetate per AGU is predicted by eq 5 for a 50% increase in the cross-section area per chain. Deuterated Bacterial Cellulose (DBC) Dissolved in Hydrogenated EMIMAc (IL-H14). SANS experiments were performed on the freeze-dried 3 wt % DBC dissolved in the hydrogenated EMIMAc (IL-H14) at 25 and 90 °C (Figure 2).

Table 3. SLD Difference Value Δρexp for 3 wt % MCC/ILD14 Obtained from Fitting the SANS Curves at Different Temperatures by the Power Law T (°C) 25 90

I0 (obtained from fitting)

Δρexp (×10−6 Å−2)

Δρtheo (×10−6 Å−2)

x (mol acetate/ AGU)

2.44

4.31

0.74

1.97

4.31

0.93

0.0012 (theo I0 = 0.0037) 0.00078

25 °C is 2.44 × 10−6 Å−2, which is 1.8 times smaller than the value calculated theoretically (Δρtheo 4.31 × 10−6 Å−2). At the higher temperature of 90 °C, Δρexp decreases to 1.97 × 10−6 Å−2, which is 2.2 times smaller than Δρtheo. The low values of Δρexp are a manifestation of the low scattering cross section measured by SANS; we propose this is due to the strong binding of IL acetate ions to the cellulose interface which renders its solubility in the ionic liquid. There is compelling evidence, by neutron diffraction with isotopic substitution, of particularly strong hydrogen bonding of the glucose hydroxyl groups to the acetate anions of EMIMAc in which it is dissolved. This is confirmed by molecular dynamics simulations.18 Thus, if we consider x acetate anions attached to one anhydroglucose unit (AGU) of cellulose, the value of x can be estimated from the ratio of Δρexp to Δρtheo as Δρexp Δρtheo

1+x =

(∑ bi)ac (∑ bi)cell ρIL

1−

−

ρIL ρcell

ρcell

Figure 2. SANS curves for 3 wt % DBC dissolved in hydrogenated ionic liquid (IL-H14) at 25 °C (open square) and 90 °C (open star). SANS curves fitted with the power law (eq 1) shown by the solid lines.

(3)

where (∑bi)cell and (∑bi)ac are the scattering lengths of one AGU of cellulose and one acetate ion, respectively. The data presented in Table 3 show that the low scattering cross section of the cellulose chains dissolved in IL results from the decrease in the SLD contrast (Δρexp) and is due to association of IL ions with the dissolved cellulose chains. Assuming tight binding of acetate anions, the lower-limit estimate of the molar ratio of bound acetate to AGU can be calculated by eq 3, using the scattering lengths of one AGU and deuterated acetate (3.15 × 10−6 and 4.49 × 10−6 Å, respectively). The molar binding ratios thus evaluated are 0.74:1 at 25 °C and 0.93:1 at 90 °C, as given in Table 3. The foregoing analysis assumes negligible change in the cross-section area of the cellulose chain upon ion adsorption. Considering that the adsorbed ions occupy a shell around the cellulose chain, such that its cross-section area increases by a factor y from A0 of the bare chain to a total cross-section area per chain of Atotal, then A v y = total ≈ 1 + x ads A0 vcell (4)

The scattering patterns from the DBC/IL-H14 solutions differ from those of the MCC/IL-D14 in two ways: the scattering intensity level at low q was higher in DBC solutions, and the temperature dependence was insignificant. Nevertheless, the patterns could be fitted by eq 1, similarly to the scattering from the MCC/IL-D14 solutions, indicating that the dissolved DBC exhibits a rigid linear structure at the length scales probed by the measured q range (0.01−0.3 Å−1)just as did the dissolved MCC. The resulting values of I0,exp and the respective evaluated experimental SLD differences Δρexp calculated by eq 2 are given in Table 4. The evaluated Δρexp at 25 and 90 °C are Table 4. Scattering Length Density (SLD) Difference for DBC in IL-H14 Obtained from I0

where vcell and vads are the molar volumes of an AGU and the adsorbed species, eq 3 can be modified as ρ (∑ bi) ÄÅ É 1 + x ∑ b ac − y ρIL ÅÅ I0,exp ÑÑÑ1/2 ( ) i cell ÅÅ ÑÑ cell ÅÅ Ñ = ρIL ÅÅ I0,theo ÑÑÑ − 1 ÅÇ ÑÖ ρcell

T (°C)

I0 (obtained from fitting)

Δρexp (×10−6 Å−2)

Δρtheo (×10−6 Å−2)

25 90

0.0029 (theo I0 = 0.0039) 0.0028

3.8 3.7

4.42 4.42

quite similar, 3.8 × 10−6 and 3.7 × 10−6 Å−2, respectively. These values are closer to the calculated SLD difference (Δρtheo = 4.42 × 10−6 Å−2), yet they are still smaller. The observation that the fitted value of I0,exp is smaller than the theoretical value I0,theo cannot be explained by ion adsorption, as in the case of MCC dissolved in IL-D14. The scattering length of an adsorbed acetate ion (1.4 × 10−6 Å) is small relative to that of a deuterated AGU (C6D7H3O5: 10.4 ×

(5)

where I0,exp is the value of the power-law prefactor obtained by fitting eq 1 and I0,theo is the calculated value of this prefactor D

DOI: 10.1021/acs.macromol.8b01425 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules 10−6 Å), yet its adsorption increases the SLD difference to the IL-H14 solvent. In this case, we need to consider that the volume fraction of molecularly dissolved cellulose chains is lower than expected from full dissolution of DBC. Sluggish dissolution of BC compared to plant cellulose has been reported.33 The large difference in the degree of polymerization of BC as compared to MCC (about 4000 and 200, respectively) can have a pronounced effect on the structure of the dissolved chains. As discussed in the following section on SAXS measurements, some chains can remain in so-called fractal aggregates, which contribute mostly to the scattering intensity in the low-q range. If we consider that only a fraction z of the dissolved DBC is in the form of individual chains, contributing to the scattering as modeled by eqs 1 and 2, then the left-hand side of eqs 3 and 4 should be multiplied by a factor z−0.5 > 1. The volume fraction of dissolved cellulose needed to account for the lower value of I0,theo (without considering ion adsorption) is lower by a factor of z = 0.74 relative to its initial value. Considering adsorption of acetate ions at the same molar ratio as evaluated for the MCC solution requires a reduction of the dissolved cellulose content by a factor of 0.59 Small-Angle X-ray Scattering (SAXS) from Solutions of Cellulose (DBC and MCC) in EMIMAc (IL-H14). SAXS experiments were performed to evaluate possible difference in the extent to which cellulose dissolves in EMIMAc as individual chains, depending on the source of cellulose (bacterial or microcrystalline). The patterns from 2 wt % cellulose solutions measured at 25 °C, shown in Figure 3,

solvent as compared to DBC, as well as the enhanced solvation effectiveness of EMIMAc for MCC.14,15 Because the most prominent feature of the SAXS pattern from the 2 wt % MCC/IL-H14 solution shown in Figure 3 is the apparent q−1 power-law relation of intensity to scattering vector in the mid to high q range, it can be evaluated using the ion-adsorption model by applying eq 3 or 4. The X-ray SLD of cellulose (13 × 10−6 Å−2) is higher than that of EMIMAc (9.4 × 10−6 Å−2) or acetate ion (value approximated as acetic acid, 9.4 × 10−6 Å−2) and scattering length ratio of acetate ion to AGU, (∑bi)ac/(∑bi)cell, is 0.36. The fit of the q−1 power law relation to the SAXS patterns from both cellulose solutions is shown in the inset of Figure 3, yielding values of I0,exp of 19.3 and 11.5 × 10−4 for MCC and DBC solutions, respectively, whereas I0,exp evaluated by eq 1 is 7.0 × 10−4. Applying eq 3 yields the lower limit of the molar adsorption ratio of acetate ion per AGU as x ≈ 0.6, which indicates the SAXS measurements support the notion of ion adsorption. Incidentally, the SAXS intensity of the DBC solution is lower by a factor of about 0.6 relative to MCC solutions, similar to that evaluated by the analysis of SANS data described above.

■

DISCUSSION Neutron scattering from polymer solutions using isotopic substitution of hydrogen with deuterium in one of the constituents is a useful method for gaining information about the conformation and structure of the dissolved polymer. This is especially effective when the solvent is deuterated, as it provides high contrast with low incoherent scattering background. It was therefore very surprising that the SANS patterns from microcrystalline cellulose (MCC) dissolved in deuterated ionic liquid EMIMAc exhibited a very low level of scattered intensity. In particular, since MCC readily dissolves in this IL yielding clear solutions, and previous SAXS and light scattering measurements gave strong evidence for full dissolution to isolated cellulose chains. The interpretation of this effect sheds light on the unique mechanism by which this ionic liquid is so effective in destructing the strong intermolecular forces within the cellulose crystal leading to its molecular dissolution. Following reported results of NMR4,18 and X-ray diffraction18 measurements, as well as molecular simulations,3,18 it is assumed that the acetate ions of the IL are the species that closely associate by hydrogen bonding to hydroxyl groups of the anhydroglucose units (AGU). The dramatic reduction in contrast between the deuterated IL and cellulose is explained by tight binding of acetate ions to cellulose, possibly by hydrogen bonding of one anion to two AGU hydroxyls.18 In the case of MCC dissolved in deuterated IL, the adsorbed ions increase the SLD of cellulose, thus reducing its contrast with the IL. This effect on the cellulose SLD implies that the molar volume of the associated ions is smaller than those in the IL solvent and highlights the strength of the cellulose−ion interaction. Thus, the cellulose dissolution mechanism can be described schematically as shown in Figure 4. As the IL surrounds the cellulose fibrils, ions adsorb to the cellulose surface, penetrate the fibrillar crystal, and disrupt interchain interactions by tight adsorption to the exfoliated chains. The tight binding of the ions result in effective charging of the dissolved chains, reducing their propensity for reassociation. The effect of temperature cannot be readily explained. Weakening of the interaction of the deuterated ion with the polymer chain with increasing temperature is expected to lower

Figure 3. SAXS patterns from 2 wt % cellulose in EMIMAc (IL-H14) after background subtraction. Open squares: MCC; open circles: DBC. Inset: high-q region displaying fits of q−1 power law (solid lines).

exhibit distinct differences between solutions of cellulose from microcrystalline or bacterial source. Intense excess scattering at low q is observed in the scattering from the DBC solution, relative to the MCC solution, with an apparent power-law relation of intensity to scattering vector with an exponent between 2 and 3. This indicates that a significant fraction of cellulose chains exist in so-called “fractal aggregates”.6,7 The scattering from MCC/EMIMAc solution, however, does not exhibit any excess scattering at the low-q region of this measurement, attesting better dissolution of MCC in this E

DOI: 10.1021/acs.macromol.8b01425 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 4. Schematic of the mechanism of cellulose dissolution in EMIMAc ionic liquid.

■

its SLD and hence increase its contrast with the IL and the measured intensity. That the opposite effect is observed may be a result of other effects, such as reduction of the solvent SLD due to its lower mass density.

Corresponding Authors

*E-mail: [email protected] (G.G.). *E-mail: [email protected] (Y.C.).

■

ORCID

CONCLUSION Dissolution of cellulose in EMIMAc ionic liquid (IL) was monitored by using contrast variation small-angle neutron scattering (SANS) measurements. Contrary to the expectations of high contrast between hydrogenated microcrystalline cellulose (MCC) and deuterated IL, a dramatic reduction of the measured intensity (scattering cross section) was observed. This was attributed to tight binding of deuterated acetate ions to the cellulose chains, the effect of which was to reduce the difference in scattering length density (SLD). The associating ions are assumed to be the acetate anions, following reported measurements by NMR and X-ray diffraction with isotopic substitution as well as molecular simulations.18 However, further studies using specific partially deuterated EMIMAc are needed to better evaluate this issue. Measurements of smallangle X-ray scattering (SAXS) corroborate this model by indicating increased contrast due to ion adsorption resulting in enhanced SLD difference, but the effect is less dramatic. Experiments were also performed with deuterated bacterial cellulose (DBC); however, the presence of fractal aggregates of the dissolved cellulose, and hence a reduced concentration of molecularly dissolved chains, hindered quantitative evaluation. These results shed light on the dissolution mechanism of cellulose in EMIMAc, highlighting its successful exfoliation of the cellulose crystals to individual chains. The tight binding of the ions interferes with the interchain association and results in effective charging of the dissolved chains maintaining their separation.

■

AUTHOR INFORMATION

Vikram Singh Raghuwanshi: 0000-0001-9524-1314 Gil Garnier: 0000-0003-3512-0056 Author Contributions

V.S.R. and Y.C. are shared first authors. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Financial support is from Australian Research Council (ARC), Australian paper, Carter Holt Harvey, Circa, Norske Skog, Orora and Visy through the Industry Transformation Research Hub Grant IH130100016. Deuteration of ionic liquid and bacterial cellulose was undertaken at the National Deuteration Facility which was partly funded by the National Collaborative Research Infrastructure Strategyan initiative of the Australian Government. Y.C. acknowledges the Israel Science Foundation, Grant No. 1164/14.

■

REFERENCES

(1) Swatloski, R. P.; Spear, S. K.; Holbrey, J. D.; Rogers, R. D. Dissolution of cellose with ionic liquids. J. Am. Chem. Soc. 2002, 124 (18), 4974−4975. (2) Yuan, X. M.; Cheng, G. From cellulose fibrils to single chains: understanding cellulose dissolution in ionic liquids. Phys. Chem. Chem. Phys. 2015, 17 (47), 31592−31607. (3) Liu, H. B.; Sale, K. L.; Holmes, B. M.; Simmons, B. A.; Singh, S. Understanding the Interactions of Cellulose with Ionic Liquids: A Molecular Dynamics Study. J. Phys. Chem. B 2010, 114 (12), 4293− 4301. (4) Remsing, R. C.; Swatloski, R. P.; Rogers, R. D.; Moyna, G. Mechanism of cellulose dissolution in the ionic liquid 1-n-butyl-3methylimidazolium chloride: a C-13 and Cl-35/37 NMR relaxation study on model systems. Chem. Commun. 2006, 12, 1271−1273. (5) Rabideau, B. D.; Agarwal, A.; Ismail, A. E. The Role of the Cation in the Solvation of Cellulose by Imidazolium-Based Ionic Liquids. J. Phys. Chem. B 2014, 118 (6), 1621−1629. (6) Idstrom, A.; Gentile, L.; Gubitosi, M.; Olsson, C.; Stenqvist, B.; Lund, M.; Bergquist, K. E.; Olsson, U.; Kohnke, T.; Bialik, E. On the

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.macromol.8b01425. Detailed nuclear magnetic resonance (NMR) spectra of protonated 1-ethyl-3-methylimidazolium acetate and deuterated 1-ethyl-3-methylimidazolium-d14 ionic liquid (PDF) F

DOI: 10.1021/acs.macromol.8b01425 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules dissolution of cellulose in tetrabutylammonium acetate/dimethyl sulfoxide: a frustrated solvent. Cellulose 2017, 24 (9), 3645−3657. (7) Hirosawa, K.; Fujii, K.; Hashimoto, K.; Shibayama, M. Solvated Structure of Cellulose in a Phosphonate-Based Ionic Liquid. Macromolecules 2017, 50 (17), 6509−6517. (8) Kauzmann, W. Some Factors in the Interpretation of Protein Denaturation. Adv. Protein Chem. 1959, 14, 1−63. (9) Baldwin, R. L. Dynamic hydration shell restores Kauzmann’s 1959 explanation of how the hydrophobic factor drives protein folding. Proc. Natl. Acad. Sci. U. S. A. 2014, 111 (36), 13052−13056. (10) Kojima, H. Studies on the phase transition of hydrogels and aqueous solutions of thermosensitive polymers. Polym. J. 2018, 50 (6), 411−418. (11) Asai, H.; Fujii, K.; Nishi, K.; Sakai, T.; Ohara, K.; Umebayashi, Y.; Shibayama, M. Solvation Structure of Poly(ethylene glycol) in Ionic Liquids Studied by High-energy X-ray Diffraction and Molecular Dynamics Simulations. Macromolecules 2013, 46 (6), 2369−2375. (12) Fujii, K.; Ueki, T.; Hashimoto, K.; Kobayashi, Y.; Kitazawa, Y.; Hirosawa, K.; Matsugami, M.; Ohara, K.; Watanabe, M.; Shibayama, M. Microscopic Structure of Solvated Poly(benzyl methacrylate) in an Imidazolium-Based Ionic Liquid: High-Energy X-ray Total Scattering and All-Atom MD Simulation Study. Macromolecules 2017, 50 (12), 4780−4786. (13) Okushita, K.; Chikayama, E.; Kikuchi, J. Solubilization Mechanism and Characterization of the Structural Change of Bacterial Cellulose in Regenerated States through Ionic Liquid Treatment. Biomacromolecules 2012, 13 (5), 1323−1330. (14) Napso, S.; Rein, D. M.; Khalfin, R.; Cohen, Y. Semidilute solution structure of cellulose in an ionic liquid and its mixture with a polar organic co-solvent studied by small-angle X-ray scattering. J. Polym. Sci., Part B: Polym. Phys. 2017, 55 (11), 888−894. (15) Rein, D. M.; Khalfin, R.; Szekely, N.; Cohen, Y. True molecular solutions of natural cellulose in the binary ionic liquid-containing solvent mixtures. Carbohydr. Polym. 2014, 112, 125−133. (16) Perrin, C. L.; Dong, Y. M. Secondary deuterium isotope effects on the acidity of carboxylic acids and phenols. J. Am. Chem. Soc. 2007, 129 (14), 4490−4497. (17) Bowron, D. T.; D’Agostino, C.; Gladden, L. F.; Hardacre, C.; Holbrey, J. D.; Lagunas, M. C.; McGregor, J.; Mantle, M. D.; Mullan, C. L.; Youngs, T. G. A. Structure and Dynamics of 1-Ethyl-3methylimidazolium Acetate via Molecular Dynamics and Neutron Diffraction. J. Phys. Chem. B 2010, 114 (23), 7760−7768. (18) Youngs, T. G. A.; Holbrey, J. D.; Mullan, C. L.; Norman, S. E.; Lagunas, M. C.; D’Agostino, C.; Mantle, M. D.; Gladden, L. F.; Bowron, D. T.; Hardacre, C. Neutron diffraction, NMR and molecular dynamics study of glucose dissolved in the ionic liquid 1ethyl-3-methylimidazolium acetate. Chem. Sci. 2011, 2 (8), 1594− 1605. (19) Su, J. L.; Raghuwanshi, V. S.; Raverty, W.; Garvey, C. J.; Holden, P. J.; Gillon, M.; Holt, S. A.; Tabor, R.; Batchelor, W.; Garnier, G. Smooth deuterated cellulose films for the visualisation of adsorbed bio-macromolecules. Sci. Rep. 2016, 6, 36119. (20) Martinez-Sanz, M.; Mikkelsen, D.; Flanagan, B. M.; Rehm, C.; de Campo, L.; Gidley, M. J.; Gilbert, E. P. Investigation of the microand nano-scale architecture of cellulose hydrogels with plant cell wall polysaccharides: A combined USANS/SANS study. Polymer 2016, 105, 449−460. (21) Meenatchi, B.; Renuga, V.; Manikandan, A. Cellulose dissolution and regeneration using various imidazolium based protic ionic liquids. J. Mol. Liq. 2017, 238, 582−588. (22) Heinze, T.; Dorn, S.; Schoebitz, M.; Liebert, T.; Koehler, S.; Meister, F. Interactions of ionic liquids with polysaccharides - 2: Cellulose. Macromol. Symp. 2008, 262, 8−22. (23) Hanabusa, H.; Izgorodina, E. I.; Suzuki, S.; Takeoka, Y.; Rikukawa, M.; Yoshizawa-Fujita, M. Cellulose-dissolving protic ionic liquids as low cost catalysts for direct transesterification reactions of cellulose. Green Chem. 2018, 20 (6), 1412−1422.

(24) Rein, D. M.; Khalfin, R.; Cohen, Y. Cellulose as a novel amphiphilic coating for oil-in-water and water-in-oil dispersions. J. Colloid Interface Sci. 2012, 386, 456−463. (25) Raghuwanshi, V. S.; Su, J. L.; Garvey, C. J.; Holt, S. A.; Holden, P. J.; Batchelor, W. J.; Garnier, G. Visualization and Quantification of IgG Antibody Adsorbed at the Cellulose-Liquid Interface. Biomacromolecules 2017, 18 (8), 2439−2445. (26) Darwish, T. A.; Yepuri, N. R.; Holden, P. J.; James, M. Quantitative analysis of deuterium using the isotopic effect on quaternary C-13 NMR chemical shifts. Anal. Chim. Acta 2016, 927, 89−98. (27) Wood, K.; Mata, J. P.; Garvey, C. J.; Wu, C. M.; Hamilton, W. A.; Abbeywick, P.; Bartlett, D.; Bartsch, F.; Baxter, P.; Booth, N.; Brown, W.; Christoforidis, J.; Clowes, D.; d’Adam, T.; Darmann, F.; Deura, M.; Harrison, S.; Hauser, N.; Horton, G.; Federici, D.; Franceschini, F.; Hanson, P.; Imamovic, E.; Imperia, P.; Jones, M.; Kennedy, S.; Kim, S.; Lam, T.; Lee, W. T.; Lesha, M.; Mannicke, D.; Noakes, T.; Olsen, S. R.; Osborn, J. C.; Penny, D.; Perry, M.; Pullen, S. A.; Robinson, R. A.; Schulz, J. C.; Xiong, N.; Gilbert, E. P. QUOKKA, the pinhole small-angle neutron scattering instrument at the OPAL Research Reactor, Australia: design, performance, operation and scientific highlights. J. Appl. Crystallogr. 2018, 51, 294−314. (28) Kline, S. R. Reduction and analysis of SANS and USANS data using IGOR Pro. J. Appl. Crystallogr. 2006, 39, 895−900. (29) Raghuwanshi, V. S.; Su, J. L.; Garvey, C. J.; Holt, S. A.; Raverty, W.; Tabor, R. F.; Holden, P. J.; Gillon, M.; Batchelor, W.; Garnier, G. Bio-deuterated cellulose thin films for enhanced contrast in neutron reflectometry. Cellulose 2017, 24 (1), 11−20. (30) Glatter, O.; Kratky, O. Small-Angle X-ray Scattering; Academic Press: New York, 1982. (31) Hjelm, R. P. The Small-Angle Approximation of X-Ray and Neutron Scatter from Rigid Rods of Non-Uniform Cross-Section and Finite Length. J. Appl. Crystallogr. 1985, 18 (Dec), 452−460. (32) Nishiyama, Y.; Langan, P.; Chanzy, H. Crystal structure and hydrogen-bonding system in cellulose 1 beta from synchrotron X-ray and neutron fiber diffraction. J. Am. Chem. Soc. 2002, 124 (31), 9074−9082. (33) Shen, X. Y.; Ji, Y. L.; Wang, D.; Yang, Q. Solubility of a High Molecular-Weight Bacterial Cellulose in Lithium Chloride/N,Ndimethylacetamide Solution. J. Macromol. Sci., Part B: Phys. 2010, 49 (5), 1012−1018.

G

DOI: 10.1021/acs.macromol.8b01425 Macromolecules XXXX, XXX, XXX−XXX