Solvated Structure of Cellulose in a Phosphonate-Based Ionic Liquid

Aug 18, 2017 - Chemistry is quantum computing's killer app. In an IBM lab in Yorktown Heights, N.Y., a circuit board sits at the bottom of a large tub...
0 downloads 11 Views 2MB Size
Download PDF
Article pubs.acs.org/Macromolecules

Solvated Structure of Cellulose in a Phosphonate-Based Ionic Liquid Kazu Hirosawa,† Kenta Fujii,*,‡ Kei Hashimoto,§ and Mitsuhiro Shibayama*,† †

Institute for Solid State Physics, The University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa, Chiba 277-8581, Japan Graduate School of Sciences and Technology for Innovation, Yamaguchi University, 2-16-1 Tokiwadai, Ube, Yamaguchi 755-8611, Japan § Department of Chemistry and Biotechnology, Yokohama National University, 79-5 Tokiwadai, Hodogaya-ku, Yokohama 240-8501, Japan ‡

S Supporting Information *

ABSTRACT: We investigated the solvated structure of cellulose in a phosphonate-based ionic liquid (IL) solution utilizing scattering experiments and all-atom molecular dynamics (MD) simulations. Based on the high-energy Xray total scattering experiment and MD simulations, a predominant interaction between cellulose and the IL was established, i.e., hydrogen bonding between the IL anion species and hydroxyl groups of cellulose. In addition, it was found that intramolecular hydrogen bonds existed within cellulose molecules, even when dissolved in the IL. Furthermore, the conformation of cellulose chains in the IL was investigated by a small-angle X-ray scattering experiment. As a result, it was found that cellulose molecules were dispersed at the molecular level and existed as rigid-rod-like polymers because of the intramolecular hydrogen bonds within the cellulose molecules. In dynamic light scattering experiments, a speckle pattern was observed for concentrated cellulose solutions. This indicated the existence of a physical-gel-like frozen inhomogeneity.

■

INTRODUCTION Ionic liquids (ILs) are molten salts having melting temperatures around room temperature. As ILs are highly ionic, they construct unique solvation environment for solutes. Because of these unique features, ILs are able to dissolve various compounds, namely metal ions,1 organic compounds,2,3 and biopolymers.4,5 Importantly, several kinds of ILs dissolve cellulose. Native cellulose molecules are insoluble in water and conventional organic solvents due to their multiple intraand intermolecular hydrogen-bonding networks. Hence, the dissolution process of cellulose into solvents has been a bottleneck in the processing of cellulose. In 2002, Rogers et al. reported that an IL, 1-butyl-3-methylimidazolium chloride ([C4mIm][Cl]), dissolves cellulose at 100 °C.6 Until now, several cellulose-dissolving ILs have been reported.7−10 For example, Ohno et al. reported that a phosphonate-based IL, 1ethyl-3-methylimidazolium methylphosphonate [C 2mIm][CH3(H)PO3] (Scheme 1a), is able to dissolve cellulose at room temperature.8 Because these ILs are capable of dissolving cellulose under mild conditions, they allow for improvement of energy efficiency in cellulose processing. Recently, it was reported that one can extract glucose from lignocellulose in IL solutions by using a hydrolysis reaction.11,12 It indicates that ILs are promising solvents for biomass processing of cellulose. In addition, fabrication of polysaccharide ion gel and films using ILs has been reported. Thus, ILs are also expected as novel media for biorenewable material processing.13 To develop ILs that have higher cellulose-dissolving abilities, or for optimizing cellulose processing conditions, understanding the solvated structure of cellulose in ILs is essential. © XXXX American Chemical Society

Scheme 1. Chemical Structures of (a) 1-Ethyl-3methylimidzolium Methylphosphonate ([C2mIm][CH3(H)PO3]), (b) Cellobiose, and (c) Cellulose

From a viewpoint of microscopic interactions between cellulose (or its monomer unit, glucose) and ILs, several investigations have been performed by utilizing solvatochromic analysis,8 NMR relaxation measurement,14 and neutron diffraction experiments.15,16 On the basis of such structural studies, it is Received: June 1, 2017 Revised: July 29, 2017

A

DOI: 10.1021/acs.macromol.7b01138 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

scattering intensity profile, Icoh(q).25,26 The experimental X-ray structure factor, Sexp(q), per stoichiometric volume and the radial distribution function, Gexp(r), were calculated as follows:27

found out that solubility of cellulose in ILs strongly depends on the property of IL anion species, especially its hydrogen bond accepting ability. It was also reported that the dissolution of cellulose in an IL 1-ethyl-3-methylimidazolium acetate ([C2mIm][CH3COO]) is exothermal,17 meaning that the interactions between the IL and cellulose are enthalpically favored. To investigate the chain conformation of cellulose, small-angle X-ray scattering (SAXS) and small-angle neutron scattering studies were performed on cellulose in pure IL ([C2mIm][CH3COO]) and IL containing a cosolvent (DMF/ [C2mIm][CH3COO] mixture of the molar ratio of 9:1).18,19 Despite these reports, the chain characteristics of cellulose in IL solutions have not been adequately studied. Recently, we carried out the structural analysis of glucose in the phosphonate-type IL [C2mIm][CH3(H)PO3] by means of high-energy X-ray total scattering (HEXTS) experiments and all-atom molecular dynamics (MD) simulations.20 We characterized the microscopic solvation structure of glucose and found that negatively charged oxygen atoms in [CH3(H)PO3]− group are hydrogen-bonded to the hydroxyl groups of glucose. However, the solvated structure of cellulose at the mesoscopic scale (conformation, dispersed state, and dynamics of the cellulose chains) is still unclear. In this study, we performed structural analysis on both (1) the microscopic solvation structure and (2) the mesoscopicsolvated structure of cellulose in pure [C2mIm][CH3(H)PO3]. In the microscopic solvation structure study, we chose cellobiose (Scheme 1b) as a model unit to approximate the microscopic structure of cellulose and utilized HEXTS experiment and MD simulations to evaluate the molecular interactions between cellobiose and [C2mIm][CH3(H)PO3]. From the mesoscopic viewpoint, we performed SAXS and light scattering experiments on cellulose in [C2mIm][CH3(H)PO3] solutions to investigate the conformation, entangled structure, and dynamics of the cellulose chains. Eventually, we obtained structural information on cellulose dissolved in [C2mIm][CH3(H)PO3] at the molecular level.

■

S exp(q) =

Icoh(q) − ∑ nifi (q)2 {∑ nifi (q)}2

Gexp(r ) − 1 =

1 2π 2rρ0

∫0

qmax

(1)

q[S exp(q) − 1] sin(qr )

sin(qπ /qmax ) qπ /qmax

dq

(2) Here, ni is the number of atom i per stoichiometric volume. f i and ρ0 represent the atomic scattering factor of atom i and the number density of atoms, respectively. As shown in eq 2, the Lorch window function was used in the inverse Fourier transformation procedure.28 MD Simulations. An all-atom MD simulation was performed for an NTP ensemble (298 K and 1 atm) in a cubic cell using the Materials Explorer 5.0 program (Fujitsu). The composition of cellobiose molecules and IL ions was set to be consistent with the sample solution measured in the HEXTS experiment: 70 cellobiose molecules were mixed into 250 ion pairs of [C2mIm][CH3(H)PO3]. Similar to our previous reports, the Canongia Lopes and Padua (CLaP) and the optimized potentials for liquid simulations all-atom (OPLS-AA) force fields were chosen for IL ions and cellobiose, respectively.29−33 The detailed procedures of our MD simulations have been described elsewhere.1,34−36 At first, the system was equilibrated for 1.5 ns with 0.2 fs intervals. Subsequently, trajectories of atoms were collected every 0.1 ps during 1.5−2.0 ns. The mass density value obtained from the MD simulations (1.2593 g cm−3) showed good agreement with the experimental value (1.2699 g cm−3). For comparison of the result of MD simulation with HEXTS experiment, we calculated the X-ray weighted structure factor, SMD(q), based on the trajectories of atoms as follows: S MD(q) =

∑i ∑j wij(q) {∑k (nk fk (q)/N )}2

∫0

rmax

4πr 2ρ0 (gijMD(r ) − 1)

sin(qr ) dr + 1 qr

(3)

⎧ ni(nj − 1)f (q)f (q)/N (N − 1) (i = j) i j ⎪ wij(q) ≡ ⎨ ⎪ 2ninjfi (q)f j (q)/N 2 (i ≠ j) ⎩

EXPERIMENTAL SECTION

(4)

where ni and N are the number of i atoms and the total number of atoms in the simulation box. gijMD(r) is the atom−atom pair correlation function between atoms i and j. The X-ray radial distribution function GMD(r) was evaluated from the calculated SMD(q) by the inverse Fourier transform, which is shown in eq 2. SAXS Experiments. SAXS experiments were conducted at the BL03XU beamline of SPring-8 (JASRI, Japan). All measurements were carried out at room temperature. To suppress parasitic scattering, we employed a vacuum chamber setting in which the sample environment and the vacuum chamber were directly connected without any window materials.37 Sample solutions were injected into a homemade planar cell (thickness: 2 mm) and annealed at 50 °C for 4 h to eliminate the effect of shear stress induced during the injection process. A pair of borosilicate cover glasses (thickness: 30 μm) were used as optical windows for the sample cell. The X-ray wavelength was 1.00 Å. A PILATUS (Dectris) was used as the area detector. The irradiation time was 30 s for all the sample solutions. The observed X-ray scattering intensities were normalized to the absolute intensity using glassy carbon (a secondary standard sample for intensity calibration). To obtain the scattering intensity function from cellulose molecules I(q), the scattering intensity of pure [C2mIm][CH3(H)PO3], IIL(q), was subtracted from that of the sample solutions, Isample(q). For quantitative discussion about the solvated cellulose structure, we carried out curve-fitting analysis with a model function of core− shell cylindrical scatterers as follows:38

Materials. [C2mIm][CH3(H)PO3] was synthesized according to the procedures reported previously.8 The purity of the synthesized material was confirmed by 1H NMR spectroscopy. The water content of the synthesized [C2mIm][CH3(H)PO3] was measured to be 300 ppm by Karl−Fischer titration. D-(+)-Cellobiose was purchased from Aldrich. Co. “Cellulose powder C” (a form of microcrystalline cellulose) was purchased from Advantec. Co. Kuroda et al. estimated the molecular weight of the cellulose material to be approximately 105 using liquid chromatography.21 Cellulose powder was dried at 90 °C for 24 h in vacuo prior to use. Cellulose in [C2mIm][CH3(H)PO3] solutions were prepared as follows: First, cellulose powder was dispersed into [C2mIm][CH3(H)PO3] by stirring with a spatula at room temperature. Subsequently, the mixtures were heated at 50 °C for 5 h under reduced pressure to obtain clear viscose solutions. The absence of undissolved cellulose powder granules was confirmed via optical microscopy (see section S1, Supporting Information). HEXTS Experiments. HEXTS measurements were carried out using BL04B2 beamline at SPring-8 (Japan Synchrotron Radiation Research Institute, Japan).22,23 Primary beam was monochromatized by using a Si(220) monochromator, and X-ray of 61.6 keV was used as the incident beam in our experiment. The HEXTS measurement was performed for Cellobiose 30 wt % in [C2mIm][CH3(H)PO3] solution at room temperature. The observed X-ray scattering spectrum was corrected for absorption and polarization.24 After these correction, the Compton scattering was subtracted, and we obtained a coherent

I(q) = ρPL(q)PCS(q)i(q) + Aq−2 B

(5) DOI: 10.1021/acs.macromol.7b01138 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules ⎡ 2 PL(q) = L2⎢ ⎣ qL

qL

∫0

4 sin 2(qL /2) ⎤ sin t ⎥ dt − t (qL)2 ⎦

⎡ 2J (qR core) PCS(q) = ⎢(bcore − bshell)(πR core 2) 1 + (bshell − bsolv ) qR core ⎣ × (πR out

2

2J (qR out) ⎤ ⎥ ) 1 qR out ⎦

2

We adopted a decoupling approximation in which the form factor of the core−shell cylinder is given as a product of a longitudinal, PL(q), and a cross-section parts, PCS(q).39 The number density of the core− shell cylinder, ρ, was calculated using the actual volume fraction of the sample solutions. bcore, bshell, and bsolv are the scattering length densities (SLD) of core, shell, and bulk solvent, respectively. During the present fitting analysis, both bcore and bsolv were fixed to the calculated SLD values for pure glucose and neat [C2mIm][CH3(H)PO3], respectively. Here, the SLDs were calculated using the mass density and the atomic scattering factors. The parameter L, Rout, and Rcore are the length, the outer radius, and the core radius of the core−shell cylinder, respectively. J1(x) is the first-order Bessel function. The power-law term Aq−2 was added to reproduce the excess scattering in the low-q region, where A is a constant. The validity and physical meanings of the model function will be discussed later. The interparticle interference between the core−shell cylinders is represented by the structure factor i(q), which is empirically developed from the polymer reference interaction site model as follows:40−43

i(q) =

c(q) =

β′ =

1 1 + β′c(q)PL(q)/L2

Figure 1. Radial distribution function, G(r), of 30 wt % cellobiose in [C2mIm][CH3(H)PO3] solution, as a form of r2[G(r) − 1], obtained from the inverse Fourier transform of the X-ray structure factor, S(q). Solid black dots and solid red lines show the G(r) obtained from the HEXTS experiment and MD simulations, respectively.

obtained from MD simulations, GMD(r), is shown in Figure 1 as a solid red line. As shown, Gexp(r) exhibited a high-frequency oscillation which was absent in GMD(r). Such a “ripple” pattern is an inevitable artifact which originates from experimental fluctuation in the X-ray structure factor, Sexp(q) (see Figure S2, Supporting Information). Such a fluctuation in Sexp(q) is transformed to the high-frequency oscillation in the experimental G(r) via inverse Fourier transform procedure (eq 2). Except for the artifact, the MD result reproduced the experimental results well. This agreement between the HEXTS experiment and the MD simulations indicated that the trajectory of molecules obtained from the present MD simulations was valid. Subsequently, we calculated the partial radial distribution functions, GMDpartial(r)s, which are related to the intermolecular interactions between cellobiose and [C2mIm][CH3(H)PO3]. Figure 2a shows a partial radial distribution function, r2[GMDinter(r) − 1], that contains all of the intermolecular correlations in the 30 wt % cellobiose in [C2mIm][CH3(H)PO3] solution. The partial radial distribution functions for cellobiose−anion and cellobiose−cation interactions (r2[GMDcello‑An(r) − 1] and r2[GMDcello‑Cat(r) − 1]) are also shown in Figure 2b. As shown in Figure 2a, the nearest

(6)

3[sin(2q(R out + lD)) − 2q(R out + lD) cos(2q(R out + lD))] [2q(R out + lD)]3

(

5

(1 + 2(B + C)2 ) + 2D 1 + B + 4 C

B = πR out 2Lρ ,

(1 − B − C)4 C=

4 πR out 3ρ , 3

D=

) −1

1 πR outL2ρ 2

Light Scattering Experiments. Light scattering measurements were performed using an SLS/DLS compact goniometer (ALV, Langen, 5022F-PCC-MS) coupled with a photon correlator. A He−Ne laser (wavelength, λ = 632.8 nm; 22 mW) was used as the incident beam. The scattering angle was fixed to be 90°, and temperature was controlled to be 25.0 °C. Each measurement required 30 s. The solvent IL, [C2mIm][CH3(H)PO3], was filtered with a PTFE filter (pore size: 0.2 μm) prior to use. Sample solutions were placed into test tubes and annealed at 50 °C for 4 h in vacuo. The position dependence of a time-averaged light scattering intensity, ⟨I⟩p, was measured by changing the relative position of the beam spot by rotating the test tube.

■

RESULTS AND DISCUSSION Microscopic Solvation Structure of Cellulose. The Xray weighted radial distribution function, Gexp(r), of 30 wt % cellobiose in [C2mIm][CH3(H)PO3] solution obtained from the HEXTS experiment is shown in Figure 1 as a form of r2[Gexp(r) − 1]. As established in our previous study,20 the peaks at r = 1.5 and 2.5 Å can be assigned to intramolecular correlations within IL ions and solutes. However, it is assumed that the intermolecular correlation components overlapped with the intramolecular correlations at r > 3.0 Å. Thus, it is difficult to discuss intermolecular interactions only from the experimental data. To evaluate the intermolecular correlation components and intramolecular ones separately, we performed all-atom MD simulations. The radial distribution function

Figure 2. Partial radial distribution functions, GMD(r)s, calculated from MD simulations for (a) total intermolecular correlations and (b) cellobiose−[CH3(H)PO3]− correlations and cellobiose−[C2mIm]+ correlations. C

DOI: 10.1021/acs.macromol.7b01138 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules intermolecular correlation is represented by a peak at r = 2.6 Å in the r2[GMDinter(r) − 1]. Figure 2b clearly shows that a sharp peak at r = 2.6 Å exists in the r2[GMDcello‑An(r) − 1] but not in the r2[GMDcello‑Cat(r) − 1]. This indicated that the observed peak in the r2[GMDinter(r) − 1] corresponds to the cellobiose− anion interactions. The same tendency was shown in our previous study on glucose in the [C2mIm][CH3(H)PO3] system.20 To discuss the local interactions more in detail, we evaluated the atom−atom pair correlation functions, gMDatom−atom(r)s. The calculation procedure is described in section S3 of the Supporting Information. At first, we focused on the intermolecular cellobiose−anion interactions, especially on the intermolecular hydrogen bonding. Figure 3 shows the

Figure 4. Atom−atom pair correlation functions, gMDHO−OC(r)s, for the intramolecular hydrogen bonds of (a) Oc···HO−OE and (b) OC··· HO−OI within cellobiose molecules.

reported for the glucose system, here too, the OC···HO−OE type hydrogen bond (Figure 4a bonds) can be found in [C2mIm][CH3(H)PO3] solutions.20 On the contrary, the OC··· HO−OI type hydrogen bond (Figure 4b bonds) is a new observation. We estimated the abundance ratio of the intramolecular hydrogen bonds, (b), to the intermolecular hydrogen bonds, O A···H O−O I, between cellobiose and [CH3(H)PO3]− from the viewpoint of the coordination number (see section S3 of the Supporting Information). The abundance ratio was found to be close to 1:1, indicating that a significant proportion of OI−HO bonds formed the intramolecular hydrogen bond of (b). We expected that the intramolecular hydrogen bond suppresses the twisting motion of glucose units. In such a condition, the flexibility of the cellulose chain, which is composed of cellobiose units, is expected to be suppressed in [C 2 mIm][CH 3 (H)PO 3 ] solutions. Mesoscopic Structure of the Cellulose Chains. To investigate the solvated structure of cellulose at the mesoscopic scale, a SAXS experiment was conducted for cellulose in [C2mIm][CH3(H)PO3]. Figure 5 shows the I(q) obtained for the cellulose in [C2mIm][CH3(H)PO3] solutions of various volume fractions, φ. In the field of polymer physics, it is well established that the scattering profile of dilute solution of the Gaussian chain is approximated by the Debye function. The

Figure 3. Atom−atom pair correlation functions, gMDOA−OX(r)s, for intermolecular hydrogen bonds between hydroxyl groups in cellobiose and the negatively charged oxygen atoms of [CH3(H)PO3]−.

gMDOA−OX(r) which denotes the atom−atom correlation between OA within the [CH3(H)PO3]− and the O atoms (X = H, I, or E) of the hydroxyl groups within cellobiose (the indices are shown in Figure 2). According to our previous report, the gMDOA−OX(r)s for glucose in [C2mIm][CH3(H)PO3] systems exhibit an obvious peak at r = 2.6 Å and reflected the hydrogen bond of OA···HO−OX.20 This atom−atom correlation at r = 2.6 Å in the gMDOA−OX(r)s mainly contributed to the observed first peak in the r2[GMDinter(r) − 1]. As mentioned in the Introduction, native cellulose molecules form multiple inter- and intramolecular hydrogen-bonding networks in the crystalline state. The results shown in Figure 3 indicate that the intermolecular hydrogen bonds between cellulose are disrupted by the formation of hydrogen bonds between cellobiose and [CH3(H)PO3]−. Here, it should be noted that there was a significant difference in the intensity of the first peaks of the gMDOA−OX(r)s. The differences can be ascribed to the intramolecular hydrogen bonds within cellobiose molecules, as discussed below. Figure 4 shows the gMDHO−OC(r)s for the intramolecular hydrogen bonds of (a) Oc···HO−OE and (b) OC···HO−OI within cellobiose molecules. The gMDHO−OC(r)s for the intramolecular hydrogen bonds of (a) and (b) exhibited a peak at r = 1.4 Å and r = 1.5 Å, respectively. This indicated that the intramolecular hydrogen bonds within the cellulose molecules existed, even when dissolved in [C2mIm][CH3(H)PO3]. The intramolecular hydrogen bonds prevented the hydroxyl groups of OE−HO and OI−HO from forming an intermolecular hydrogen bond with [CH3(H)PO3]−. Thus, we deduced that the difference observed for the intensity of the first peaks of the gMDOA−OX(r)s (Figure 3) originated from the existence of intramolecular hydrogen bonds. As we have already

Figure 5. SAXS profiles obtained for cellulose in [C2mIm][CH3(H)PO3] solutions with various volume fractions, φ = 0.0076−0.047 (symbols). The solid lines show the calculated scattering curve based on eq 5. D

DOI: 10.1021/acs.macromol.7b01138 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Debye function exhibits power-law behavior, I(q) ∼ q−2.44,45 The wormlike chain also shows the I(q) ∼ q−2 behavior when q−1 is sufficiently larger than the Kuhn segment length.46 However, in the present study, the SAXS profiles for dilute solutions (φ < 0.038) exhibited a power-law behavior of I(q) ∼ q−1, not I(q) ∼ q−2. The I(q) ∼ q−1 behavior corresponds to a characteristic scattering pattern of rod-like scatterers,38 indicating that the cellulose chains existed as rod-like polymers in [C2mIm][CH3(H)PO3]. The exponent of I(q) decreased with increasing φ. It indicated that contribution of the interchain interference factor, i(q), to I(q) became significant with increasing concentration. In addition, in the low-q region (q < 0.02 Å−1), a large upturn emerged in the SAXS profiles for concentrated solutions (φ ≥ 0038), and the intensity of the upturn increased with increasing φ. Here, the excess scattering exhibits a power-law behavior of I(q) ∼ q−2. We assumed that this excess scattering was an indication of aggregates formation. From such a viewpoint, the exponent of the excess scattering reflects fractal nature of aggregates formed by rod-like cellulose chain. The fractal dimension of aggregates formed by rod-like particles is reported to be 1.81−2.26.47 Hence, we conclude that the observed fractal dimension of the cellulose aggregates in our study (∼2) is consistent with the previously reported value for rod-like particle aggregates. For a more quantitative discussion, we performed a curve fitting analysis on the SAXS profiles. At first, we examined a model function for solid cylindrical scatterers without a shell layer. However, the radius of the solid cylinder obtained from the curve fitting was unrealistically small, and we could not reproduce the experimental scattering profiles sufficiently. We conjectured that the cylinder model did not work well at high-q region because complicated atomistic correlations were not negligible in the region. Thus, we adopted the “core−shell” cylinder model to reproduce the scattering pattern from cellulose molecules surrounded by IL ions. Such an approximation has been proposed in a characterization of semiflexible polymers in solution.48 Eventually, the experimental SAXS profiles were fitted by eq 5, indicative of a scattering curve for core−shell cylindrical scatterers with interparticle interference. Here, the Debye length, lD, was fixed at 0 because the core−shell cylinders, which approximate cellulose molecules, are not charged. The obtained fitting curves are shown in Figure 5 (solid lines). As shown, the experimental SAXS profiles were successfully reproduced by the model function, including their absolute scattering intensities. The Rcore and Rout, obtained as fitting parameters, are shown in Figure 6a. As shown, the Rcore was independent of φ, and its value was almost constant at ∼5 Å, which is close to the radius of a single cellulose molecule. The Rout was also independent of φ, and its value was ∼7 Å. In the previous section, we showed the partial radial distribution function for cellobiose−anion and cellobiose−cation interactions, i.e., r2[GMDCello‑An(r) − 1] and r2[GMDCello‑Cat − 1] (see Figure 2b). Both functions converged to 0 when r > 7 Å. This means that an outer limit of the first coordination shell around cellulose molecules is r = 7 Å. Considering this, the obtained value of Rout = 7 Å is expected to be valid. These results indicate that cellulose molecules are dispersed in [C2mIm][CH3(H)PO3] at the molecular level, and they exist as rigid-rod-like polymer chains. Such a rigid conformation is ascribed to the intramolecular hydrogen bonds, as discussed in the previous section. Such hydrogen bonds restrict the thermal fluctuations of the cellulose chain segments, resulting in rigid-rod-like conformation for this

Figure 6. Obtained parameters (a) Rout (blue circles) and Rcore (red squares) and (b) bshell from the curve-fitting analysis using the core− shell cylinder model function (eq 5) on the SAXS profiles of cellulose in [C2mIm][CH3(H)PO3] solutions. The dashed lines in (b) are the calculated values of bcore and bsolv.

system. The scattering length density in the shell layer, bshell, is also shown in Figure 6b. bshell was lower than bsolv. As shown in Figure 2b, r2[GMDcello‑X(r) − 1]s (X = anion or cation), which represent cellobiose−IL ions correlations, were lower than 0 in the range of r < 7 Å. It indicates that the number density of ions in the solvation shell was lower than that of bulk solvent. Thus, we conclude that bshell obtained from the curve fitting for SAXS profiles is consistent with the results of MD simulations. Very recently, two papers about solvated structure of cellulose in ILs, [C4mIm][Cl]49 and [C2mIm][CH3COO],50 have been published. Both of the authors utilized SAXS measurements to investigate the chain conformation of cellulose in the ILs. Napso et al. reported that cellulose molecules are dispersed at the molecular level and exist as rodlike polymers in [C2mIm][CH3COO], which is almost the same conclusion with ours.50 On the other hand, Jiang et al. reported that cellulose molecules exhibit relatively flexible conformation in [C4mIm][Cl].49 Such conflicting results can be explained from a viewpoint of the size of IL anion. The sizes of [CH3(H)PO3]− (our study) and CH3COO− are larger than Cl−. It indicates that Cl− can more easily disrupt intramolecular hydrogen bonds within cellulose chains. Actually, it was reported that the intramolecular hydrogen bonds are disrupted in [C4mIm][Cl],51 but they remain in [C2mIm][CH3COO]52 E

DOI: 10.1021/acs.macromol.7b01138 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

emergence of frozen inhomogeneity in the concentrated cellulose solutions was correlated to aggregates formation. Figure 8a shows the time correlation function at a given measurement point, gp(2)(τ) − 1, of light scattering intensity

based on MD simulations. Thus, it is indicated that the conformation of cellulose molecules in the ILs is strongly related to the degree of disruption of the hydrogen bonds. To investigate the spatial inhomogeneity of the cellulose solutions, we measured the position (p) dependence of the time-averaged light scattering intensity, ⟨I⟩p. Figure 7 shows p-

Figure 8. (a) Representative time correlation functions of light scattering intensity at a given measurement point, gp(2)(τ) − 1, obtained from cellulose in [C2mIm][CH3(H)PO3] solutions of various φ. (b) φ dependence of the correlation length, ξ, estimated from the fast mode decay time based on the Einstein−Stokes equation (eq 9). Figure 7. Position (p) dependence of time-averaged scattering intensity of light ⟨I⟩p from cellulose in [C2mIm][CH3(H)PO3] solutions, φ = 0.0076−0.047.

obtained from cellulose in [C2mIm][CH3(H)PO3] solutions of various φ. Here, gp(2)(τ) − 1 is defined as follows: gp(2)(τ ) − 1 =

dependence of ⟨I⟩p measured for cellulose in [C2mIm][CH3(H)PO3] solutions with various φ. Dilute solution (φ < 0.038) did not show any fluctuations for ⟨I⟩p with p. This indicated that the dilute cellulose solutions were homogeneous, like typical polymer solutions. On the contrary, the ⟨I⟩p measured for concentrated cellulose solutions (φ ≥ 0038) demonstrated an obvious p dependence, i.e., a speckle pattern. In general, a speckle pattern reflects the existence of frozen density fluctuations in the scattering media, which is characteristic of polymer gels or glasses.53 Hence, it is supposed that the translational diffusion of cellulose chains was restricted by pseudo-cross-links originating from strong entanglement between rigid cellulose chains in the concentrated solutions. In addition, the threshold φ for the speckle pattern emergence (φ = 0.038) corresponded to that of the upturn emergence in the SAXS low-q region profiles (Figure 5). This means that the

⟨I(t + τ )I(t )⟩p ⟨I(t )2 ⟩p

(7)

where I(t) is the light scattering intensity. For all of the solution examined here, the g(2) p (τ) − 1 showed two types of decay modes. The apparent diffusion coefficients, DA,fast,p and DA,slow,p, corresponding to the observed decay modes were estimated by a curve-fitting procedure using the following model function: gp(2)(τ ) − 1 = βσ12 × [A exp(−DA,fast, pq2τ ) + (1 − A) exp(−DA,slow, pq2τ )]2 (8)

where σI2 is related to fraction of dynamic component in ⟨I⟩p and β is the coherence factor. The fitting curves are shown in Figure 8a as solid lines. As shown, the experimental g(2) p (τ) − 1 was successfully reproduced by the model function given in eq F

DOI: 10.1021/acs.macromol.7b01138 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

was clarified that intermolecular hydrogen bonds between cellulose molecules are disrupted in [C2mIm][CH3(H)PO3] solutions; thus, cellulose molecules are dispersed at the molecular level. In addition, cellulose molecules exist as rodlike polymers because of the intramolecular hydrogen bonds within cellulose chains in [C2mIm][CH3(H)PO3]. Furthermore, in concentrated regions, rod-like cellulose molecules are strongly entangled, and cellulose in [C2mIm][CH3(H)PO3] solutions have frozen inhomogeneities that arise from pseudocross-links between cellulose molecules.

8. It is supposed that the slow mode originates from the diffusion of small amounts of aggregates, which is often observed in polymer solutions. Hence, we do not dwell on the properties of the slow mode at this time. We deduced that the fast mode reflects the collective diffusion of cellulose chain segments and estimated the correlation length as follows: ξ=

kBT 6πηDfast

(9)

where kB, T, and η are the Boltzmann constant, absolute temperature, and the viscosity of [C2mIm][CH3(H)PO3], respectively. The Stokes−Einstein type equation is an established one that denotes the relationship between the cooperative diffusion coefficient and the correlation length of semidilute polymer solutions and polymer gels.54 One can derive the equation from general scaling concepts, and thus it is applicable to stiff-polymer solutions. The true diffusion coefficient for the collective diffusion, Dfast, was estimated using a partial-heterodyne method, which was established in our previous study (see section S4 of the Supporting Information for a detailed procedure).55 The φ dependence of the estimated ξ is shown in Figure 8b. The power-law behavior of ξ ∼ φ−0.49±0.05 was observed. Based on the scaling concepts for semidilute polymer solutions, ξ scaled as ξ ∼ φμ, where the exponent μ is related to the fractal dimension of the polymer chains, dp: μ ∼ −1/(3 − dp).56 Hence, μ was easily predicted to be 1.0, 0.75, and 0.5 for the ideal chain (dp = 2), real chain (dp = 5/3), and rod-like polymer (dp = 1). In the present study, the obtained value, μ = 0.49 ± 0.05, is close to the predicted value for a rod-like polymer. This result indicated that the cellulose chains existed as rigid-rod-like polymers in [C2mIm][CH3(H)PO3], which was consistent with the SAXS results.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.macromol.7b01138. Figures S1−S4 (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail [email protected] (K.F.). *E-mail [email protected] (M.S.). ORCID

Kenta Fujii: 0000-0003-0057-1295 Mitsuhiro Shibayama: 0000-0002-8683-5070 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This study has been financially supported by Grants-in-Aid for Scientific Research from the MEXT (No. 16H02277 and 25248027 to M.S.). K.H. was granted by the Program for Leading Graduate Schools (MERIT), the Japan Society for the Promotion of Science. HEXTS experiments were conducted at BL04B2 of SPring-8, Japan Synchrotron Radiation Research Institute (JASRI) (2011A1434 and 2012B1502). The SAXS experiment was performed at the Frontier Soft Matter Beamline (FSBL; BL03XU), SPring-8, Hyogo, Japan, with the assistance of Atsushi Izumi, Sumitomo Bakelite, Co., Ltd. (Proposal No. 2016B7260). This study was supported by Photon and Quantum Basic Research Coordinated Development Program by MEXT Grant No. 13004017.

■

CONCLUSIONS A picture for the concentrated cellulose solutions in [C2mIm][CH3(H)PO3] obtained from HEXTS, MD simulations, SAXS, and light scattering experiments is shown in Figure 9. As shown

■

REFERENCES

(1) Fujii, K.; Hamano, H.; Doi, H.; Song, X.; Tsuzuki, S.; Hayamizu, H.; Seki, S.; Kameda, Y.; Dokko, K.; Watanabe, M.; Umebayashi, Y. Unusual Li+ Ion Solvation Structure in Bis(fluorosulnonyl)amide Based Ionic Liquid. J. Phys. Chem. C 2013, 117, 19314−19324. (2) Huddleston, J. G.; Willauer, H. D.; Swatloski, R. P.; Visser, A. E.; Rogers, R. D. Room Temperature Ionic Liquids as Novel Media for ’Clean’ Liquid-Liquid Extraction. Chem. Commun. 1998, 16, 1765− 1766. (3) Chiappe, C.; Pieraccini, D. Ionic Liquids: Solvent Properties and Organic Reactivity. J. Phys. Org. Chem. 2005, 18, 275−297. (4) Madeira Lau, R.; Van Rantwijk, F.; Seddon, K. R.; Sheldon, R. A. Lipase-Catalyzed Reactions in Ionic Liquids. Org. Lett. 2000, 2 (26), 4189−4191. (5) van Rantwijk, F.; Sheldon, R. A. Biocatalysis in Ionic Liquids. Chem. Rev. 2007, 107, 2757−2785. (6) Swatloski, R. P.; Spear, S. K.; Holbrey, J. D.; Rogers, R. D. Dissolution of Cellulose with Ionic Liquids. J. Am. Chem. Soc. 2002, 124, 4974−4975.

Figure 9. Picture of the solvated structure of cellulose solutions in [C2mIm][CH3(H)PO3] at the (a) microscopic scale and (b) mesoscopic scale.

in Figure 9a, the microscopic solvation structure was investigated using HEXTS experiments and MD simulations. It was clarified that (1) negatively charged oxygen atoms in [CH3(H)PO3]− form hydrogen bonds to the hydroxyl groups of cellulose and (2) intramolecular hydrogen bonds between the adjacent glucose segments within cellulose molecules exist even when dissolved in [C2mIm][CH3(H)PO3]. The mesoscopic structure of cellulose chains in [C2mIm][CH3(H)PO3] was investigated by SAXS and light scattering experiments. It G

DOI: 10.1021/acs.macromol.7b01138 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

(25) Cromer, D. T. Compton Scattering Factors for Aspherical Free Atoms. J. Chem. Phys. 1969, 50, 4857−4859. (26) Hubbell, J. H.; Veigele, W. J.; Briggs, E. A.; Brown, R. T.; Cromer, D. T.; Howerton, R. J. Atomic Form Factors, Incoherent Scattering Functions, and Photon Scattering Cross Sections. J. Phys. Chem. Ref. Data 1975, 4 (3), 471. (27) Pings, C. J.; Waser, J. Analysis of Scattering Data for Mixtures of Amorphous Solids or Liquids. J. Chem. Phys. 1968, 48, 3016−3018. (28) Lorch, E. Neutron Diffraction by Germania, Silica and Radiation-damaged Silica Glasses. J. Phys. C: Solid State Phys. 1969, 2, 229−237. (29) Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K. M.; Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.; Caldwell, J. W.; Kollman, P. A. A Second Generation Force Field for the Simulation of Proteins, Nucleic Acids, and Organic Molecules. J. Am. Chem. Soc. 1995, 117, 5179−5197. (30) Damm, W.; Frontera, A.; Tirado-Rives, J.; Jorgensen, W. L. OPLS All-Atom Force Field for Carbohydrates. J. Comput. Chem. 1997, 18 (16), 1955−1970. (31) Canongia Lopes, J. N.; Deschamps, J.; Pádua, A. A. H. Modeling Ionic Liquids Using a Systematic All-Atom Force Field. J. Phys. Chem. B 2004, 108, 2038−2047. (32) Canongia Lopes, J. N.; Costa Gomes, M. F.; Pádua, A. A. H. Nonpolar, Polar, and Associating Solutes in Ionic Liquids. J. Phys. Chem. B 2006, 110, 16816−16818. (33) Murzyn, K.; Bratek, M.; Pasenkiewicz-Gierula, M. Refined OPLS All-Atom Force Field Parameters for n-Pentadecane, Methyl Acetate, and Dimethyl Phosphate. J. Phys. Chem. B 2013, 117 (51), 16388−16396. (34) Fujii, K.; Soejima, Y.; Kyoshoin, Y.; Fukuda, S.; Kanzaki, R.; Umebayashi, Y.; Yamaguchi, T.; Ishiguro, S.; Takamuku, T. Liquid Structure of Room-temperature Ionic Liquid, 1-Ethyl-3-methylimidazolium Bis-(trifluoromethanesulfonyl) imide. J. Phys. Chem. B 2008, 112, 4329−4336. (35) Fujii, K.; Kanzaki, R.; Takamuku, T.; Kameda, Y.; Kohara, S.; Kanakubo, M.; Shibayama, M.; Ishiguro, S.; Umebayashi, Y. Experimental Evidences for Molecular Origin of Low-Q peak in Neutron/X-ray Scattering of 1-Alkyl-3-methylimidazolium Bis(trifluoromethanesulfonyl)amide Ionic Liquids. J. Chem. Phys. 2011, 135, 244502. (36) 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. (37) Masunaga, H.; Sakurai, K.; Akiba, I.; Ito, K.; Takata, M. Accurate Measurement of Intrinsic Scattering from Window Materials by Use of a Vacuum Camera. J. Appl. Crystallogr. 2013, 46, 577−579. (38) Livsey, I. Neutron Scattering from Concentric Cylinders. Intraparticle Interference Function and Radius of Gyration. J. Chem. Soc., Faraday Trans. 2 1987, 83 (8), 1445−1452. (39) Pedersen, J. S.; Schurtenberger, P. Cross-Section Structure of Cylindrical and Polymer-Like Micelles from Small-Angle Scattering Data. I. Test of Analysis Methods. J. Appl. Crystallogr. 1996, 29 (6), 646−661. (40) Pedersen, J. S.; Schurtenberger, P. Scattering Functions of Semiflexible Polymers with and without Excluded Volume Effects. Macromolecules 1996, 29, 7602−7612. (41) Pedersen, J. S.; Schurtenberger, P. Scattering Functions of Semidilute Solutions of Polymers in a Good Solvent. J. Polym. Sci., Part B: Polym. Phys. 2004, 42 (17), 3081−3094. (42) Pedersen, J. S. Small-angle Scattering from Surfactants and Block Copolymer Micelles. In Soft Matter Characterization; Pecora, P., Borsali, R., Eds.; Springer-Verlag: New York, 2008; Vol. 1, pp 192− 227. (43) Schweizer, K. S.; Yethiraj, A. Polymer reference interaction site model theory: New molecular closures for phase separating fluids and alloys. J. Chem. Phys. 1993, 98, 9053−9079. (44) Debye, P.; Bueche, A. M. Light Scattering by Concentrated Polymer Solutions. J. Chem. Phys. 1950, 18, 1423−1425.

(7) Fukaya, Y.; Sugimoto, A.; Ohno, H. Superior Solubility of Polysaccharides in Low Viscosity, Polar, and Halogen-Free 1,3Dialkylimidazolium Formates. Biomacromolecules 2006, 7 (12), 3295− 3297. (8) Fukaya, Y.; Hayashi, K.; Wada, M.; Ohno, H. Cellulose Dissolution with Polar Ionic Liquids under Mild Conditions: Required Factors for Anions. Green Chem. 2008, 10, 44−46. (9) Sun, N.; Rahman, M.; Qin, Y.; Maxim, M. L.; Rodriguez, H.; Rogers, R. D. Complete Dissolution and Partial Delignification of Wood in the Ionic Liquid 1-Ethyl-3-methylimidazolium Acetate. Green Chem. 2009, 11 (5), 646−655. (10) Abe, M.; Fukaya, Y.; Ohno, H. Extraction of Polysaccharides from Bran with Phosphonate or Phosphinate-derived Ionic Liquids under Short Mixing Time and Low Temperature. Green Chem. 2010, 12, 1274−1280. (11) Kuroda, K.; Miyamura, K.; Satria, H.; Takada, K.; Ninomiya, K.; Takahashi, K. Hydrolysis of Cellulose Using an Acidic and Hydrophobic Ionic Liquid and Subsequent Separation of Glucose Aqueous Solution from the Ionic Liquid and 5-(Hydroxymethyl)furfural. ACS Sustainable Chem. Eng. 2016, 4 (6), 3352−3356. (12) Kuroda, K.; Inoue, K.; Miyamura, K.; Satria, H.; Takada, K.; Ninomiya, K.; Takahashi, K. Efficient Hydrolysis of Lignocellulose by Acidic Ionic Liquids under Low-Toxic Condition to Microorganisms. Catalysts 2017, 7 (4), 108. (13) Takada, A.; Kadokawa, J. Fabrication and Characterization of Polysaccharide Ion Gels with Ionic Liquids and Their Further Conversion into Value-Added Sustainable Materials. Biomolecules 2015, 5 (1), 244. (14) 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 13C and 35/37Cl NMR Relaxation Study on Model Systems. Chem. Commun. 2006, 12, 1271−1273. (15) Youngs, T. G. A.; Hardacre, C.; Holbrey, J. D. Glucose Solvation by the Ionic Liquid 1,3-Dimethylimidazolium Chloride: A Simulation Study. J. Phys. Chem. B 2007, 111 (49), 13765−13774. (16) 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 1-Ethyl-3-methylimidazolium Acetate. Chem. Sci. 2011, 2 (8), 1594− 1605. (17) Andanson, J. M.; Padua, A. A. H.; Costa Gomes, M. F. Thermodynamics of Cellulose Dissolution in an Imidazolium Acetate Ionic Liquid. Chem. Commun. 2015, 51 (21), 4485−4487. (18) 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. (19) Endo, T.; Hosomi, S.; Fujii, S.; Ninomiya, K.; Takahashi, K. Nano-Structural Investigation on Cellulose Highly Dissolved in Ionic Liquid: A Small Angle X-ray Scattering Study. Molecules 2017, 22 (1), 178. (20) Hirosawa, K.; Fujii, K.; Hashimoto, K.; Umebayashi, Y.; Shibayama, M. Microscopic Solvation Structure of Glucose in 1-Ethyl3-methylimidazolium Methylphosphonate Ionic Liquid. J. Phys. Chem. B 2015, 119 (20), 6262−6270. (21) Kuroda, K.; Fukaya, Y.; Ohno, H. Direct HPILC Analysis of Cellulose Depolymerisation in Ionic Liquids. Anal. Methods 2013, 5 (13), 3172−3176. (22) Kohara, S.; Suzuya, K.; Kashihara, Y.; Matsumoto, N.; Umesaki, N.; Sakai, I. A Horizontal Two-axis Diffractometer for High-energy Xray Diffraction Using Synchrotron Radiation on Bending Magnet Beamline BL04B2 at SPring-8. Nucl. Instrum. Methods Phys. Res., Sect. A 2001, 467−468, 1030−1033. (23) Isshiki, M.; Ohishi, Y.; Goto, S.; Takeshita, K.; Ishikawa, T. High-energy X-ray Diffraction Deamline: BL04B2 at SPring-8. Nucl. Instrum. Methods Phys. Res., Sect. A 2001, 467−468, 663−666. (24) Sakai, S. KEK Report; National Laboratory for High Energy Physics: Tsukuba, Japan, 1990; Vol. 90-16. H

DOI: 10.1021/acs.macromol.7b01138 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules (45) Debye, P. Molecular-weight Determination by Light Scattering. J. Phys. Colloid Chem. 1947, 51, 18−32. (46) des Cloizeaux, J. Form Factor of an Infinite Kratky-Porod Chain. Macromolecules 1973, 6 (3), 403−407. (47) Mohraz, A.; Moler, D. B.; Ziff, R. M.; Solomon, M. J. Effect of Monomer Geometry on the Fractal Structure of Colloidal Rod Aggregates. Phys. Rev. Lett. 2004, 92 (15), 155503. (48) Arakawa, S.; Terao, K.; Kitamura, S.; Sato, T. Conformational Change of an Amylose Derivative in Chiral Solvents: Amylose Tris(nbutylcarbamate) in Ethyl Lactates. Polym. Chem. 2012, 3 (2), 472− 478. (49) Jiang, X.; Kitamura, S.; Sato, T.; Terao, K. Chain Dimensions and Stiffness of Cellulosic and Amylosic Chains in an Ionic Liquid: Cellulose, Amylose, and an Amylose Carbamate in BmimCl. Macromolecules 2017, 50 (10), 3979−3984. (50) 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. (51) Mostofian, B.; Cheng, X.; Smith, J. C. Replica-Exchange Molecular Dynamics Simulations of Cellulose Solvated in Water and in the Ionic Liquid 1-Butyl-3-Methylimidazolium Chloride. J. Phys. Chem. B 2014, 118 (38), 11037−11049. (52) Payal, R. S.; Balasubramanian, S. Dissolution of Cellulose in Ionic Liquids: an Ab Initio Molecular Dynamics Simulation Study. Phys. Chem. Chem. Phys. 2014, 16 (33), 17458−17465. (53) Shibayama, M. Spatial Inhomogeneity and Dynamic Fluctuations of Polymer Gels. Macromol. Chem. Phys. 1998, 199, 1−30. (54) de Gennes, P. G. Scaling Concepts in Polymer Physics; Cornell University Press: Ithaca, NY, 1979. (55) Shibayama, M. Universality and Specificity of Polymer Gels Viewed by Scattering Methods. Bull. Chem. Soc. Jpn. 2006, 79, 1799− 1819. (56) Daoud, M.; Cotton, J. P.; Farnoux, B.; Jannink, G.; Sarma, G.; Benoit, H.; Duplessix, R.; Picot, C.; de Gennes, P. G. Solutions of Flexible Polymers. Neutron Experiments and Interpretation. Macromolecules 1975, 8, 804−818.

I

DOI: 10.1021/acs.macromol.7b01138 Macromolecules XXXX, XXX, XXX−XXX