Cellulose (2018) 25:4345–4355 https://doi.org/10.1007/s10570-018-1860-x ORIGINAL PAPER Ia to Ib mechano-conversion and amorphization in native cellulose simulated by crystal bending . . . . Pan Chen Yu Ogawa Yoshiharu Nishiyama Ahmed E. Ismail Karim Mazeau Received: 4 September 2017 / Accepted: 17 May 2018 / Published online: 5 June 2018 The Author(s) 2018 Abstract The bending of rod-like native cellulose (Ia) form to the alternating Ib one or vice versa. This crystals with degree of polymerization 40 and 160 mechanical deformation converts the Ia form pro- using molecular dynamics simulations resulted in a gressively to the Ib form, as has been experimentally deformation-induced local amorphization at the kink- observed for ultrasonication of microﬁbrils. Ib is also ing point and allomorphic interconversion between able to partially convert to Ia-like organization but this cellulose Ia and Ib in the unbent segments. The conversion is only transitory. The qualitative agree- transformation mechanism involves a longitudinal ment between the behavior of ultrasonicated microﬁb- chain slippage of the hydrogen-bonded sheets by the rils and in silico observed Ia ? Ib conversion length of one anhydroglucose residue (* 0.5 nm), suggests that shear deformation and chain slippage which alters the chain stacking from the monotonic under bending deformation is a general process when cellulose ﬁbrils experience lateral mechanical stress. Keywords Bending Chain slippage Kink Electronic supplementary material The online version of this article (https://doi.org/10.1007/s10570-018-1860-x) con- Molecular dynamics simulation Allomorphic tains supplementary material, which is available to authorized conversion users. P. Chen (&) State Key Laboratory of Pulp and Paper Engineering, Introduction South China University of Technology, 510640 Guangzhou, China In recent years, cellulose nanoﬁbers (CNF) and e-mail: email@example.com nanocrystals (CNC) have received increased interest P. Chen in materials science (Habibi et al. 2010; Grishkewich Wallenberg Wood Science Center, KTH Royal Institute of et al. 2017). These cellulose nanomaterials are Technology, 56-58 Teknikringen, SE-10044 Stockholm, obtained by disintegrating cellulosic ﬁbers in the plant Sweden cell wall through chemical or mechanochemical Y. Ogawa Y. Nishiyama K. Mazeau (&) treatments. These materials may experience various Univ. Grenoble Alpes, CNRS, CERMAV, stress conditions in industrial processes and in use that BP53, 38000 Grenoble Cedex 9, France may alter the inner structure of the cellulose nanocrys- tals through dislocations (Hidayat et al. 2012) and A. E. Ismail allomorphic transitions. Understanding the mechani- Department of Chemical and Biomedical Engineering, cal response of crystalline cellulose to various external West Virginia University, Morgantown, WV 26505, USA 123 4346 Cellulose (2018) 25:4345–4355 forces is important for fundamental understanding and the chains, which have degree of polymerization 160 for development of more effective processing. to produce a crystal of 80 nm in length, this crystal is Processing cellulose by ultrasound is a scalable identical to the one already described. process currently used in many laboratories. It is often Each crystal model was energy minimized (EM) observed that sonicated cellulose ﬁbrils have a tortu- and then equilibrated by molecular dynamics (MD). ous morphology, showing numerous kinks and can be EM was performed using the steepest descent method accompanied by an allomorphic transformation (Bri- followed by the conjugated gradient method, with the ois et al. 2013) from the triclinic cellulose Ia to the convergence criterion being a maximum force of -1 -1 monoclinic Ib form. 10 kJ mol nm . In the dynamics process, the We recently carried out molecular modeling to temperature was slowly increased from 0 to 300 K mimic a 3-point bending test of cellulose Ib crystals over 6 ns, followed by a 94 ns equilibration period. (Chen et al. 2016). Upon bending, the crystal devel- oped a sharp kink identical to that observed experi- Bending mentally for processed celluloses (Kekaelaeinen et al. 2012; Wang et al. 2012; Zeng et al. 2012; Martoia et al. The bending was carried out as described in a previous 2016). We found shear deformations are very impor- paper (Chen et al. 2016). Brieﬂy, the equilibrated tant when the bending load is applied on the crystals were rotated to align their long axes with the z- hydrophobic face, for which the bending rigidity is axis and their respective hydrophobic surfaces (the also the lowest. 110 surface for Ia and the 100 surface for Ib) parallel In this study, we applied this bending protocol to to the yz plane, as shown in Fig. 1. They were then produce kinks in molecular models of pure Ia and pure bent normal to their hydrophobic surface by displacing Ib cellulose crystals on their hydrophobic face. We bridge oxygen atoms at the center of the top surface to followed the evolution of the internal atomic structure aﬁxed x coordinate. Constraints were also applied on to investigate the molecular details of allomorphic the glycosidic oxygen atoms at the two ends of the transition and morphological changes. Results bottom layers of cellulose chains, with the x coordinate obtained on short crystals (about 20 nm in length) of constrained at the reducing end and the x and z coor- Ia and Ib will be shown ﬁrst, followed by results dinates constrained at the non-reducing end. The obtained on a long crystal (about 80 nm in length) of crystal was bent by gradually moving the x coordinate Ia. of the constrained atoms at the center of the microﬁb- ril. A stepwise displacement of -0.015 nm was applied, followed by either 1 or 3 ns of MD simula- Computational methods tion, which was enough for the energy to converge (Chen et al. 2016). Model construction Simulation setups The cellulose microﬁbril models of Ia and Ib were built from crystal structures (Nishiyama et al. All simulations have been performed with GRO- 2002, 2003) reﬁned from X-ray and neutron diffrac- MACS version 4.6 for cellulose chains with degree of tion data. The dominant hydrogen bond network, polymerization 40 and version 2016.3 for the ﬁbril pattern A, was considered. The constructed models with degree of polymerization 160 (Hess et al. 2008). consisted of 40 chains (5 9 8) of 40 residues each, The cellulose ﬁbril was modelled using the GROMOS exposing the 100 and 010 surfaces for Ib and the 110 56A force ﬁeld, which is a mostly explicit-atom carbo force ﬁeld, although carbons bonded to a single and 110 surfaces for Ia, as shown in Fig. 1. The two hydrogen atom are replaced by a CH1 united atom models have approximately square cross sections with similar dimensions of * 3.12 nm 9 * 4nm 9 * (Hansen and Huenenberger 2011) whose Lennard– Jones parameters have been subsequently optimized 21 nm. To estimate if the size of the crystals affects the results, we have also considered a long crystal (Chen et al. 2014). MD simulations were performed in the NVT ensemble using a 2 fs integration time step. initially in the Ia organization. Except for the length of The stochastic velocity-rescaling algorithm of Bussi 123 Cellulose (2018) 25:4345–4355 4347 Fig. 1 Snapshots of cellulose Ia and Ib crystals kinked at their hydrophobic surface for various bending deﬂections. Top: typical crystal model showing the nomenclature of the crystallographic planes. The three black lines passing through the crystal cross- section (one at the bottom of each of the two extremities and one on top at the center) indicate the location of the constrained atoms. Glucose residues of pairs of consecutive chains of the inner 010 (Ib) plane or the 110 (Ia) plane highlighted in red were selected to estimate the chain slippage during bending. Bottom: snapshots of the kinked crystals; labels a to h correspond to bending deﬂections where chain slippage occurred. Atoms of some residues are colored and displayed as spheres to facilitate visualization of the monoclinic and triclinic characters of the internal organization within the two linear segments et al. (2007) was used to control the temperature at where D (nm) is the distance between the centers of 300 K. Bonds involving explicit hydrogen atoms were gravity of two anhydroglucose units belonging to two constrained using the LINCS algorithm (Hess et al. consecutive chains in the bending direction (see 1997). Electrostatic and Lennard–Jones interactions Fig. 1), and a = 0.78 nm and b = 0.80 nm are the were both truncated with a cut-off radius of 1.2 nm; no respective unit cell dimensions. We obtained a rough long-range electrostatic or Lennard–Jones interactions estimate of the free energy by calculating the torsional were computed. entropy contribution from the dihedral angle distribu- tion (Chen et al. 2012). Analysis In addition, powder diffraction proﬁles were cal- culated using the Debyer software package (Wojdyr To quantify the mutual sliding of adjacent chains, we 2011), which is based on the Debye scattering deﬁned the ‘‘chain slippage,’’ D , as: slip equation (Debyer 1915) and has been used previously rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 2 2 (Nishiyama et al. 2012). a þ b D ¼ D slip 123 4348 Cellulose (2018) 25:4345–4355 XX perfect Ia and Ib crystals (Nishiyama et al. sin Qr ij IQðÞ¼ f f i j 2002, 2003), whereas it explores the gg and gt Qr ij i j orientations (x at ± 60) at the crystal surface or within the less organized zones (amorphous or para- where Q is the scattering vector, r =|r - r | is the ij i j crystalline) (Mazeau and Heux 2003). Similarly, pyran distance between atoms i and j, I(Q) is the scattering rings are in the most stable C chair conformation in intensity, and f and f are the scattering factors of 1 i j the unstressed crystals and departures from this form atoms i and j. A wavelength of 1.5418 A, correspond- clearly indicate that the structure becomes amorphous ing to the wavelength of CuKa, is used. (Mazeau and Heux 2003). Analysis of the kinked structures shows two populations of the polar angle, h, of the puckering parameters (Cremer and Pople 1975). Results The dominant population is centered around h =15, which is representative of the C classical conforma- Local structures: amorphous kinks surrounded 1 tion of the pyran, while the minor population is by crystalline straight segments centered at h =75, indicating that some pyran rings adopt the boat and skew-boat conformations. Figure 1 gives snapshots of the atomistic models of Figures 2 and 3 show that the linear segments the two cellulose crystals, showing increasingly remain highly organized for the entire range of studied ‘‘kinked’’ deformations. Figures 2 and 3 report, deﬂections. The hydroxymethyl groups keep the initial respectively, the conformational states of the hydrox- tg conformation (Fig. 2) and the pyran rings adopt ymethyl groups (Fig. 2) and the pyran rings (Fig. 3)as exclusively the C conformation (Fig. 3). The regular a function of deﬂection and of their position along the organization of the linear portions of the crystals can cellulose crystal. These two parameters are used to be visualized in the snapshots shown in Fig. 1. Near probe the local state of organization in the kinked the loading point, however, the structure becomes crystals. The hydroxymethyl group explores exclu- disorganized. The appearance of the gg and gt sively the tg orientation (x = 180) in the core of the Fig. 2 Conformational states explored by the hydroxymethyl groups as a function of deﬂection (horizontal axes) and location along the length of the crystals (vertical axes). Glucose residues are numbered from 1 to 40, starting from the non- reducing end of the crystals. Left: distribution of the gg (a:Ia, b:Ib), tg (c:Ia, d:Ib) and gt (e:Ia, f:Ib) conformations for the Ia (a, c, e) and Ib (b, d, f) crystals. Right: orientation of the O6 atom in a glucose residue in (from top to bottom) the gg, tg and gt conformations respectively. The fractions are indicated by the colors at the right of the plots 123 Cellulose (2018) 25:4345–4355 4349 Fig. 3 Distribution of the C chair conformation of the Ia the crystals (vertical axes). Glucose residues are numbered from (a) and Ib (b) crystals explored by the pyran rings as a function 1 to 40, starting from the non-reducing end of the crystals. The of deﬂection (horizontal axes) and location along the length of fractions are indicated by the colors at the right of the plots conformers of the hydroxymethyl groups, together cellobiose) larger than that of Ib, due to the increase in with the appearance of the unconventional forms of nonbonded interactions (16.6 kJ/mol cellobiose) that the pyran rings accompanying the kinking, is percep- overcompensates for the stabilization of bonded tible for deﬂections beyond 2–3 nm. The disorganized interactions (-12 kJ/mol cellobiose). That Ib is more portion of the micro-ﬁbril propagates progressively stable than Ia agrees with previous modeling simula- and symmetrically on both sides of the kink as tions, where differences of 8–10 kJ/(mol cellobiose) deﬂection increases. using empirical force-ﬁelds (Neyertz et al. 2000; Figure 4 shows the predicted wide-angle X-ray Mazeau and Heux 2003) and of 1 kJ/(mol cellobiose) powder diffractograms for Ia (Fig. 4a) and Ib from DFT calculations (Bucko et al. 2011; Li et al. (Fig. 4b), respectively, at various deﬂections. The 2011) have been reported. diffracted intensities decrease by increasing deﬂec- The energy-deﬂection curves show several discon- tion, indicating an overall decrease in crystallinity tinuities where small increases in deﬂection are induced by kinking. This feature is consistent with the accompanied by energy drops of up to 1.5 kJ/(mol previous observations about the distribution of the cellobiose). Such energy drops are speciﬁc to bending conformation of the hydroxymethyl groups and the at the hydrophobic surfaces of native cellulose crys- pyran ring puckering. tals, as they are not observed in the other directions Diffractograms predicted for Ia and Ib at various (Chen et al. 2016). For deﬂections larger than 5 nm, levels of deﬂection are superimposed in Fig. 4c. As the energy of the kinked Ia microﬁbril is lower than expected, at small deﬂections, the most intense that of its initial equilibrated crystal, suggesting that diffraction peak of Ib (200) appears at a larger 2 - h the structural rearrangements for this allomorph are value than that for Ia (110). In contrast, the two favorable, and probably irreversible, despite the diffractograms predicted from the strongly kinked Ia presence of a signiﬁcant amount of amorphous phase and Ib become almost identical, indicating that the that has a higher energy than the crystals (Mazeau and straight domains of the two kinked crystals undergo Heux 2003). In contrast, no such behavior is observed ultrastructural reorganizations and converge to a for the Ib form, where the energies of the kinked Ib similar structure. crystal are always higher than that of the initial straight form. Discontinuities in the potential energy curves Torsional entropy compensation Figure 5 shows how, taking the torsional entropy into account, the potential energy and free energy of the Ia The potential energy increase due to bending is and Ib crystals vary with deﬂection. In the initial state, accompanied by larger torsional freedom in the kinked the total energy of the Ia crystal is about 4.6 kJ/(mol region, resulting in a much smaller net increase in free 123 4350 Cellulose (2018) 25:4345–4355 Fig. 4 Simulated wide-angle X-ray powder diffraction: kinked models of Ia (a) and Ib (b) and superimposed diffractograms of the Ia and Ib models at various deﬂections (c) energy, between 2 and 4 kJ/(mol cellobiose) less than the deﬂections where energy drops, indicating that the the potential energy for cellulose Ia and Ib, respec- event is associated by longitudinal slip of the hydro- tively (Fig. 5). Thus, under ambient conditions, cel- gen-bonded sheets. The magnitude of the displace- lulose chains should kink more easily than what the ments is remarkably constant, ± 0.5 nm, which potential energy increase or the theoretical bending corresponds to the length of a glucosyl residue. In rigidity would imply. cellulose Ia and Ib, the hydrogen bonded sheets are staggered by about 0.25 nm. The main structural Chain slippage at the origin of the energy drops difference between the two allomorphs is the direction of the stagger, which is unidirectional in Ia but Figure 6 shows the variation of the chain slippage alternating in Ib. Mutual sliding of 0.5 nm thus distances (D ) with deﬂection. In the straight corresponds to local Ia ! Ib allomorphic transi- slip segments, all of the chains within a sheet behaved tions. The sliding of the hydrogen-bonded sheets can similarly. Thus D represents the longitudinal orga- be visualized by inspecting the snapshots of the kinked slip nization of consecutive sheets within the crystalline microﬁbril models. Atoms of selected glucosyl units in parts. D ﬂuctuates around a constant value, meaning the linear segments are represented by spheres in slip that the supramolecular organizations are globally Fig. 1 to facilitate visualization of the loci of the maintained in the linear segments of the microﬁbril allomorphic transitions. Figure 1b, c shows snapshots during kinking. D also shows occasional jumps at of the Ia and Ib microﬁbrils before and after the ﬁrst slip 123 Cellulose (2018) 25:4345–4355 4351 Fig. 5 Variation of the energies of Ia and Ib as a function of bending deﬂection. a Total potential energies of the Ia (red) and Ib (blue) crystals. Labels a– h refer to snapshots shown in Fig. 1. b and c variation with bending deﬂection of the potential energy (grey), torsion entropy TDS (green) and free energy (blue) of the Ib (b) and Ia (c) crystals. All energies reported in b and c are relative to the initial unkinked crystals energy drop, while Fig. 1c, e, g give snapshots after Interestingly, the 5 kJ/(mol cellobiose) energy drop each energy drop for the kinked Ia and Ib microﬁbrils. between tags f and g is due to three concerted sliding The initial organization of the cellulose chains in events on the non-reducing end side of the microﬁbril the crystal inﬂuences the number and location of the (Fig. 6). Finally, under extreme bending, at a deﬂec- allomorphic transitions (Figs. 1, 6). For the Ia system, tion of 7.5 nm (Fig. 1g right), the organization is the ﬁrst transition takes place at a deﬂection of 1.8 nm, mainly Ib-like on both sides of the kink, showing that where D between layers 3 and 4 on the reducing end the allomorphic transformation is almost complete. slip side varies suddenly by ? 0.5 nm (Fig. 5). Just after This progressive transformation from triclinic form to the transition (Fig. 1c), the reducing side of the a monoclinic form leads to the stabilization of the microﬁbril is composed of two triclinic blocks sepa- kinked state after being displaced 5 nm compared to rated by a monoclinic interface between layers 3 and 4. its initial form (Fig. 5). This structural event is repeated at each energy drop Similar relaxations by sliding exists for the Ib form, and, after each new sliding event, the total amount of with the energy drops accompanying * 0.5 nm the monoclinic form increases. This accumulation of longitudinal displacements of the sheets (Fig. 5) the Ib-like form can be observed in Fig. 1g where the resulting in local Ib ? Ia allomorphic transitions microﬁbril is strongly bent after three energy drops (Fig. 1c, e, g). The ﬁrst transition occurs at a deﬂection corresponding to six sliding events, and the vast of 2.5 nm between layers 4 and 5 on the non-reducing majority of the crystal is in Ib-like organization, (right) end of the ﬁbril, which, at tag c, is organized in except for the amorphous kink. Note that, at this two monoclinic blocks separated by a triclinic-like bending level, the non-reducing end side of the interface between layers 4 and 5). However, layers 4 microﬁbril (to the right of the kink in Fig. 1e, g) is and 5 in the Ib system undergo several successive also affected by the structural reorganization. transitions on the non-reducing side, converting the 123 4352 Cellulose (2018) 25:4345–4355 Fig. 6 Variation of the chain slippage distance (D ) between adjacent slip hydrogen-bonded sheets as a function of bending deﬂection for Ib (left) and Ia (right). Blue: reducing end, red: non-reducing end. Traces are displaced vertically by 1 nm for visibility. Layers are numbered from 1 (the layer opposite the loaded layer) to 8 (the loaded layer) interface into Ia at 2.5 nm of deﬂection, before the local structural transformation in vacuum. Exper- reverting to Ib at 3.5 nm and transforming again into imentally, the allomorphic transition and creation of Ia at 6.5 nm. This suggests that the Ib to Ia-like kinks by sonication is observed on long microﬁbrils conversion in Ib is transitory; after each new sliding but not on short fragments of crystals, and in aqueous event, the organization within the microﬁbril alter- media (Briois et al. 2013). nates between pure Ib and a Ia/Ib mixture. It is only To address the question of size effects, we have for extreme curvatures greater than approximately repeated the kinking protocol on a 80 nm long Ia 6 nm of deﬂection, that a signiﬁcant portion of the crystal, corresponding to a degree of polymerization of non-reducing side of the microﬁbril transforms into Ia. 160. A load applied at the center of the crystal We note that the reducing end side of the microﬁb- perpendicular to the hydrogen bonded layers produces ril is unaffected by bending and keeps its initial a kink similar to that observed in the short crystals. monoclinic organization. This asymmetry of the Visual inspection of the model at large deﬂections also deformation with respect to the reducing end might reveals two distinct coexisting organizational states, as come from the sight ratchet-like morphology of the structure is clearly disorganized at the kinking cellulose chains, which results in asymmetric shear point but remains highly organized on both sides of the compliance. kink. Figure 7 displays the evolution of the slippage distances as a function of deﬂection. Similar to that Effect of the crystal length observed for the short crystals, the long crystal undergoes 0.5 nm relative displacements of the To keep the computational demands of our simula- hydrogen-bonded layers, which indicates local allo- tions reasonable, we ﬁrst opted to study relatively morphic conversion. Among the seven pairs of short crystal segments to study the effect of bending on adjacent layers, four such transitions are observed in 123 Cellulose (2018) 25:4345–4355 4353 Fig. 7 Variation of the chain slippage distance (D ) between adjacent hydrogen-bonded sheets as a function of bending deﬂection for slip the long model of Ia. Layers are numbered from 1 (the layer opposite the loaded layer) to 8 (the loaded layer) the deﬂection range 15–30 nm. However, the sliding distribution of the kink angle is 60 (Usov et al. 2015), of the layers in the short crystal appears at smaller kink which corresponds to a deﬂection of 5 nm in our angles compared to the long crystal, suggesting that model. To reach this level of deﬂection, the Ia crystal sliding is easier in the short model. For example, the model underwent two transitions on the reducing end ﬁrst transition takes place at a kink angle of 11.5 for and three on the non-reducing end (Figs. 1, 6), leading the 20 nm model, whereas it appears only at 19.5 for to an accumulation of the maximal possible amount of the 80 nm model. Ib. Overall, the evolution with kinking of the mor- In contrast, the cellulose microﬁbrils from tunicate, phology, the local internal organization (amorphous or dominated by the Ib phase, remain unchanged when crystalline) as well as the allomorphic transitions are they are sonicated. Our models show that the reducing independent of size effects. This concordance of end of Ib can undergo some transitions to Ia during structural behavior between the short and long model kinking, but at the observed kink angle of 60 the two suggests that the molecular mechanism of response to straight segments surrounding the kink are entirely a lateral force by phase transition is general. monoclinic (Ib), corroborating that transitory Ib ? Ia conversions are undetected. Thermal treatment transforms Ia into Ib (Horii et al. Discussion 1987; Yamamoto et al. 1989; Debzi et al. 1991; Yamamoto and Horii 1993). The high thermal expan- Experimentally sonicated microﬁbrils showed numer- sion perpendicular to the pyranose plane upon heating ous kinks when observed in TEM, with a 20% measured by in situ X-ray diffraction suggests the decrease in crystallinity and an increase in gg and gt thermally induced phase transition is also due to conformers of the hydroxylmethyl groups as estimated mutual sliding at these hydrophobic interfaces. by solid-state CP-MAS NMR spectroscopy (Briois et al. 2013). Our models qualitatively agree with these experimental observations. The simulation also sug- Conclusions gests that crystallinity degrades only at the kinking point and the linear segments remain crystalline. We have performed molecular dynamics simulations We also reported that cellulose microﬁbrils from of cellulose crystals of length 20 nm and 80 nm to Glaucocystis, mostly in Ia, partly transformed into Ib study their behavior under bending deformations. by ultrasonication (Briois et al. 2013). Kinking the Ia Comparable to experimental observations, when bent microﬁbril model agree with this observation. It has on the hydrophobic plane, cellulose crystals tend to been recently measured that the maximum of the create kink zones with loss of crystalline order. while 123 4354 Cellulose (2018) 25:4345–4355 Grishkewich N, Mohammed N, Tang J, Tam KC (2017) Recent the two linear segments on each side of the kink advances in the application of cellulose nanocrystals. Curr remain crystalline. The creation of kinks is accompa- Opin Colloid Interface Sci 29:32–45 nied by translations of the hydrogen-bonded sheets of Habibi Y, Lucia LA, Rojas OJ (2010) Cellulose nanocrystals: the crystalline segments, corresponding to partial chemistry, self-assembly, and applications. Chem Rev 110:3479–3500 allomorphic transitions. The Ia ? Ib conversions Hansen HS, Huenenberger PH (2011) A re-optimized GRO- accumulate with increasing bending deﬂection, while MOS force ﬁeld for hexopyranose-based carbohydrates Ib ? Ia conversions are only transitory. accounting for the relative free energies of ring conform- ers, anomers, epimers, hydroxymethyl rotamers, and gly- Acknowledgments This work was supported by the cosidic linkage conformers. J Comput Chem 32:998–1032 Wallenberg Wood Science Center (WWSC) funded by the Hess B, Bekker H, Berendsen HJC, Fraaije JGEM (1997) Knut and Alice Wallenberg Foundation. This work was LINCS: a linear constraint solver for molecular simula- performed as part of the Cluster of Excellence ‘‘Tailor-Made tions. J Comput Chem 18:1463–1472 Fuels from Biomass’’, which is funded by the Excellence Hess B, Kutzner C, van der Spoel D, Lindahl E (2008) GRO- Initiative by the German federal and state government to MACS 4: algorithms for highly efﬁcient, load-balanced, promote science and research at German universities. We and scalable molecular simulation. J Chem Theory Comput acknowledge the computational resource from the 4:435–447 Centre d’Experimentaion et de Calcul Intensif (CECIC). Hidayat BJ, Felby C, Johansen KS, Thygesen LG (2012) Cel- lulose is not just cellulose: a review of dislocations as Open Access This article is distributed under the terms of the reactive sites in the enzymatic hydrolysis of cellulose Creative Commons Attribution 4.0 International License (http:// microﬁbrils. Cellulose 19:1481–1493 creativecommons.org/licenses/by/4.0/), which permits unre- Horii F, Yamamoto H, Kitamaru R, Tanahashi M, Higuchi T stricted use, distribution, and reproduction in any medium, (1987) Transformation of native cellulose crystals induced provided you give appropriate credit to the original by saturated steam at high temperatures. Macromolecules author(s) and the source, provide a link to the Creative Com- 20:2946–2949 mons license, and indicate if changes were made. Kekaelaeinen K, Illikainen M, Niinimaeki J (2012) Morpho- logical changes in never-dried kraft ﬁbers under mechan- ical shearing. Cellulose 19:879–889 Li Y, Lin M, Davenport JW (2011) Ab initio studies of cellulose I: crystal structure, intermolecular forces, and interactions References with water. J Phys Chem C 115:11533–11539 Martoia F, Dumont PJJ, Orgeas L, Belgacem MN, Putaux JL Briois B, Saito T, Petrier C, Putaux J.-L, Nishiyama Y, Heux L, (2016) On the origins of the elasticity of cellulose nanoﬁber Molina-Boisseau S (2013) Ia ? Ib transition of cellulose nanocomposites and nanopapers: a micromechanical under ultrasonic radiation. Cellulose 20:597–603 approach. RSC Adv 6:47258–47271 Bucko T, Tunega D, Angyan JG, Hafner J (2011) Ab initio study Mazeau K, Heux L (2003) Molecular dynamics simulations of of structure and interconversion of native cellulose phases. bulk native crystalline and amorphous structures of cellu- J Phys Chem A 115:10097–10105 lose. J Phys Chem B 107:2394–2403 Bussi G, Donadio D, Parrinello M (2007) Canonical sampling Neyertz S, Pizzi A, Merlin A, Maigret B, Brown D, Deglise X through velocity rescaling. J Chem Phys 126:014101/ (2000) A new all-atom force ﬁeld for crystalline cellulose I. 1–014101/7 J Appl Polym Sci 78:1939–1946 Chen P, Nishiyama Y, Mazeau K (2012) Torsional entropy at Nishiyama Y, Langan P, Chanzy H (2002) Crystal structure and the origin of the reversible temperature-induced phase hydrogen-bonding system in cellulose Ib from synchrotron transition of cellulose. Macromolecules 45:362–368 X-ray and neutron ﬁber diffraction. J Am Chem Soc Chen P, Nishiyama Y, Mazeau K (2014) Atomic partial charges 124:9074–9082 and one Lennard–Jones parameter crucial to model cellu- Nishiyama Y, Sugiyama J, Chanzy H, Langan P (2003) Crystal lose allomorphs. Cellulose 21:2207–2217 structure and hydrogen bonding system in cellulose Ia Chen P, Ogawa Y, Nishiyama Y, Ismail AE, Mazeau K (2016) from synchrotron X-ray and neutron ﬁber diffraction. J Am Linear, non-linear and plastic bending deformation of Chem Soc 125:14300–14306 cellulose nanocrystals. Phys Chem Chem Phys Nishiyama Y, Johnson GP, French AD (2012) Diffraction from 18:19880–19887 nonperiodic models of cellulose crystals. Cellulose Cremer D, Pople JA (1975) Gerneral deﬁnition of ring pucker- 19:319–336 ing coordinates. J Am Chem Soc 97(6):1354–1358 Usov I, Nystrom G, Adamcik J, Handschin S, Mezzenga R, Debye P (1915) Zerstreuung von Ro¨ntgenstrahlen. Ann Phys Schutz C, Fall A, Bergstrom L (2015) Understanding (Berlin) 46:809–823 nanocellulose chirality and structure-properties relation- Debzi EM, Chanzy H, Sugiyama J, Tekely P, Excofﬁer G (1991) ship at the single ﬁbril level. Nat Commun 6:7564 The Ia ? Ib transformation of highly crystalline cellulose Wang QQ, Zhu JY, Gleisner R, Kuster TA, Baxa U, McNeil SE by annealing in various mediums. Macromolecules (2012) Morphological development of cellulose ﬁbrils of a 24:6816–6822 bleached eucalyptus pulp by mechanical ﬁbrillation. Cel- lulose 19:1631–1643 123 Cellulose (2018) 25:4345–4355 4355 Wojdyr M (2011) Debyer, 9 Jan 2017 version. http://debyer. alkaline and acidic solutions at high temperatures. readthedocs.io and http://code.google.com/p/debyer/wiki/ Macromolecules 22:4130–4132 debyer. Accessed 1 May 2018 Zeng X, Retulainen E, Heinemann S, Fu S (2012) Fiber defor- Yamamoto H, Horii F (1993) CPMAS carbon-13 NMR analysis mations induced by different mechanical treatments and of the crystal transformation induced for Valonia cellulose their effect on zero-span strength. Nord Pulp Pap Res J by annealing at high temperatures. Macromolecules 27:335–342 26:1313–1317 Yamamoto H, Horii F, Odani H (1989) Structural changes of native cellulose crystals induced by annealing in aqueous
Cellulose – Springer Journals
Published: Jun 5, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
All the latest content is available, no embargo periods.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud