at SciVerse ScienceDirect

Polymer 53 (2012) 1124e1130

Contents lists available

Polymer

journal homepage: www.elsevier .com/locate/polymer

Impact of hydrogen bonding on inclusion layer of urea to cellulose: Studyof molecular dynamics simulation

Lu Cai, Yuan Liu, Haojun Liang*

CAS Key Laboratory of Soft Matter Chemistry, Department of Polymer Science and Engineering, University of Science and Technology of China, Hefei, Anhui 230026,People’s Republic of China

a r t i c l e i n f o

Article history:Received 22 March 2011Received in revised form18 November 2011Accepted 7 January 2012Available online 17 January 2012

Keywords:Molecular dynamics simulationHydrogen bonding interactionInclusion layer

* Corresponding author. Tel.: þ86 551 3600240; faxE-mail address: hjliang@ustc.edu.cn (H. Liang).

0032-3861/$ e see front matter � 2012 Elsevier Ltd.doi:10.1016/j.polymer.2012.01.008

a b s t r a c t

Molecular dynamics simulation was applied for investigating the inclusion layer of urea to a singlecellulose chain in urea-water solvent mixture. The formation of urea layer surrounding the cellulosechain, which mediated by different patterns of hydrogen bonding, was confirmed by radial distributingfunction and angular number distribution. Furthermore, the temperature properties of hydrogenbonding revealed that urea molecules are more favorable to interacting with cellulose than watermolecules at a certain temperature range. Cellulose-urea interaction energy and hydrogen bond numberwere also calculated as a function of temperature and the result of calculation further proved thehypothesis that there exists a dominative hydrogen bonding pattern between cellulose and urea, inwhich oxygen atoms of urea as proton-acceptors and hydroxyl hydrogen atoms of cellulose as proton-donors. The cellulose-urea interaction is dependent of temperature, in which hydrogen bondingbetween cellulose and urea decreased gradually with temperature increasing at temperature above265 K and it reached a plateau phase with temperature above 283 K.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

The hydrogen bonding between solvent molecules andhydrogen bonded polymer plays a very significant role in theconformational stability of the polymer and is also a very importantdriving force contributing to the dissolution of polymers in thepolymer-solvent system [1e5]. It was demonstrated that hydrogenbonding is a very important factor in protein stability, which resultsnot only from the formation of a hydrogen bond network betweenthe backbone groups of protein chains but also from the formationof many hydrogen bonds existing between the polar groups locatedon the protein chains surface and water molecules surroundingthem [6]. In the other hand, urea can be used as a kind of stabilizerfor proteins when the concentration of urea in protein solution islow enough [7,8]. However, more urea molecules in protein solu-tion would cause protein denaturation [1,2,4], in which the mostpossible reason is that the preferential adsorption of urea mole-cules onto the protein hydrophilic residues mediated by hydrogenbonding decreases both intramolecular and intermolecular inter-actions of proteins, and then leads to a swelling of the protein

: þ86 551 3601760.

All rights reserved.

[9e11]. Hydrogen bonding can further affect the solubility of thepolymer [12]. It was reported that cellulose pretreated with wateror ethanol can be dissolved in the imidazolium-based ionic liquidsand the main reason to affect cellulose solubility is the formation ofhydrogen bonding network between anion, small molecules andcellulose [3,13,14]. Now, hydrogen bonding between small mole-cules and hydrogen bonded polymer chains has attracted more andmore interest because it could reveal the mechanism of confor-mational change or dissolution of hydrogen bonded polymers.

Cellulose is one of the typical hydrogen bonded polymers whosemonomer structure is shown in Fig. 1a, but the application islimited due to its non-hydrosoluble property which is because ofhydrogen bonding between the cellulose chains [15]. Many studieshave been devoted to improve the solubility of cellulose. Now thereare some environment-safe ionic liquids have been suggested aspotential replacements of volatile organic compounds for cellulosedissolution [3,12]. And also, it has reported cellulose could be dis-solved rapidly in pre-cooled NaOH/urea aqueous solution [16].However, the mechanism of dissolution of cellulose in thesesolvents is still elusive. Zhang L. et al. suspect that hydrogenbonding between ions and cellulose and forming the NaOHhydrogen bonded cellulose-urea inclusion complexes may be oneof the major factors in dissolution of cellulose in alkali-urea solventmixture [17]. As a typical hydrogen bonding dominated organic

Fig. 1. Structures of (a) cellulose monomer and (b) urea molecule.

Fig. 2. RDF between O (A) of cellulose and oxygen atoms of urea for 12 ns samplingperiod (black curve), and for 12 ns after sampling period (red curve) at 261 K. (Forinterpretation of the references to colour in this figure legend, the reader is referred tothe web version of this article.)

L. Cai et al. / Polymer 53 (2012) 1124e1130 1125

molecule, urea whose structure is shown in Fig. 1b with thehydrogen bond properties is also considered to be the ‘host’molecules to constitutes inclusion complexes with various long-chain hydrocarbons [18e20], and becomes the denaturant orstabilizer for proteins in solvent mixtures [7,8]. Full understandingof hydrogen bonding between cellulose chain and the urea mole-cules is not only helpful to know about the mechanism of cellulosedissolution in urea-containing solvent mixture but also help us tobring on a consideration over the influence of the urea to theconformational and functional properties of polymers in manyfields. Many experimental researches have been focused oncellulose-urea system [17,21e23]. In this study, moleculardynamics (MD) simulation was applied for investigating thehydrogen bonding between cellulose and urea.

MD simulation has been carried out into the investigation ofcellulose conformation [24e27], cellulose dissolution [15,28], urea/water mixture system [11,29e31] and the typical inclusioncomplexes of urea molecules [19,32]. In this work, we apply MDsimulation to study the interaction between the single cellulosechain and urea-water solvent mixture and associated temperatureproperty. Our results could give significant insight of the hydrogenbonding behavior of polymer and small molecules, as well as itstemperature property.

2. Model and simulation method

The GROMACS 3.3.1 software package [33] is used for the energyminimization and MD simulation and analysis of MD trajectories. Itapplies the FFGMX, which is based on the GROMOS-87 force field[34] for the single cellulose chain and urea molecules, and the SPCmodel [35] for water. The partial atomic charges of all themoleculeswere assigned with the standard module of this force field, asshown in Fig. 1. The potential energy of the system is described asfollows:

V ¼X

kbðl� l0Þ2þX

kqðq�q0Þ2þX

kf½1þcosðnf�dÞ�

þXi<j

4εij

sijrij

!12

� sijrij

!6!þXi<j

qiqj4pε0rij

(1)

where l is the bond length, q is the bond angle, f is the dihedralangle, l0 is the equilibrium bond length, q0 is the equilibrium bondangle, kb, kq and kf are the force constants, n is the multiplicityfactor, and d is the phase shift. In last two terms, rij is the distancebetween atoms i and j, and ε and s are the energy and sizeparameters of the Lennard-Jones potential, respectively. An arith-metic average is used for the sigma’s, whereas a geometric averageis used for the epsilon’s, as shown in the following equation:

sij ¼12�sii þ sjj

�; εij ¼

�εiiεjj

�1=2: (2)

In the last term, qi is the charge of atom i and ε0 is the dielectricconstant of the vacuum. The last two terms are non-bondedpotential functions.

All simulations were carried out with an initial box measuring7.15 � 2.55 � 4.16 nm3. The Lennard-Jones potential truncationcutoff was set to 1 nm, and all bond lengths were constrained usingthe LINCS algorithm [36]. At a distance shorter than 1.0 nm, elec-trostatic interactions were calculated explicitly, whereas particle-mesh Ewald summation (PME) [37] was adopted for lone-rangeinteractions. All MD simulations were performed under theisothermal-isobaric, ensemble of the Berendsen method witha constant pressure of 1 atm [38]. Periodic boundary conditionswere used in energy minimizations and MD simulations. Thecellulose chain was energy-minimized in vacuum to obtain itsunbent initial conformation and then was position-restrained insucceeding MD simulations to exclude the effect of polymer chainconformation change in the solute-solvent interaction.

Every MD simulationwith a 2 fs time step was carried out for 38and 12 ns to equilibrate the entire system and sample, respectively.Coordinates and velocities of atoms were collected every 1 ps. Theradial distributing function (RDF) of the cellulose chain with ureamolecules at all the temperatures chosenwas examined to confirmthat the system was in equilibrium. As shown in Fig. 2, the super-position of RDF curves of 12 ns sample period and 12 ns after itdemonstrates system equilibration throughout the sample period.RDF curves at other temperatures (data not shown) are similar tothose obtained at 261 K. At every temperature, three random

L. Cai et al. / Polymer 53 (2012) 1124e11301126

numbers were used to ensure the reliability of our results. Thenumber of hydrogen bond and interaction energy between cellu-lose and urea were calculated respectively by g_hbond andg_energy program provided by GROMACS [33].

3. Results and discussion

In our system, seven temperatures were applied for the simu-lation: 253, 257, 261, 265, 269, 283 and 297 K. A model chain ofcellulose with 12 monomers was placed at the center of a quadratebox; the rest of the model cell was filled with 160 urea moleculesand 2058 water molecules. The mass fraction of urea was 19.76 wt%and cellulose was 4.0 wt%. The self-diffusion coefficient of water inthe system, DA, was calculated within the chosen temperaturerange and the values ranged from0.94�10�5 to 2.75�10�5 cm2 s�1

and increased linearlywith increasing temperature. Inwater, the DAof the SPC water model is 3.85 � 10�5 cm2 s�1 at 297 K and 1 atm[39]. The small difference is caused by the interaction of ureamolecules with water. Water molecules had logical self-diffusionand were linear with temperature, which indicated that theproposed system remained liquid within the temperature region.

3.1. Hydrogen bond type

Due to both of cellulose and urea are hydrogen bonded mole-cules, oxygen atoms in cellulose and urea, nitrogen atoms in ureaare all potential proton-acceptors for hydrogen bonding. In order totest the dominant way of hydrogen bonding between cellulose andurea, RDF and angular distribution probability between themwerecalculated in our study.

The RDF between oxygen atoms of cellulose and nitrogen atomsof urea are shown in Fig. 3a. The first three curves of RDF corre-sponding to the hydroxyl oxygen atoms of cellulose (O (A), O (B),and O (C), Fig.1a) have similar tendency, while the other two curvescorresponding to the backbone oxygen atoms of cellulose (O (D)and O (E), Fig. 1a) shift toward farther distances slightly. Eventhough, the first peaks of these five curves all locate between 0.25and 0.35 nm and this distance indicate the hydrogen bondinginteraction. It is clear that the hydrogen bonds of O(D)$$$HeN andO(E)$$$HeN are formed by backbone oxygen atoms of cellulose asproton-acceptors and hydrogen atoms of urea as proton-donors. Bycontrast, the hydrogen bonds of O(A)$$$H$$$N, O(B)$$$H$$$N andO(C)$$$H$$$N are relatively complicated, in which both of hydroxyloxygen atoms in cellulose and nitrogen atoms in urea could serve asproton-acceptors, accordingly, hydrogen atoms in urea andhydroxyl hydrogen atoms in cellulose could serve as proton-donors.

The RDF between hydroxyl oxygen atoms of cellulose andoxygen atoms of urea are shown in Fig. 3b. The first peak of thesethree RDF curves are also located around 0.25 nm, which implies

Fig. 3. RDF curves of (a) five oxygen atoms of cellulose with nitrogen atoms of urea and

the formation of hydrogen bonds between O(A)eH$$$O, O(B)eH$$$O and O(C)eH$$$O, in which carbonyl oxygen atoms of ureaserve as proton-acceptors and hydroxyl hydrogen atoms of cellu-lose serve as proton-donors. Furthermore, the values of first peaksin Fig. 3b approach 18, but only 4 in Fig. 3a. This result indicates thatthe hydrogen bonding interaction between cellulose and ureadepends largely on hydrogen bond of O(A)eH$$$O, O(B)eH$$$Oand O(C)eH$$$O, i.e. number of hydrogen bonds formed byoxygen atoms of cellulose or nitrogen atoms of urea as proton-acceptors only takes few proportions in the total number ofhydrogen bonds.

To further confirm the hydrogen bond of O(A)eH$$$O, O(B)eH$$$O and O(C)eH$$$O are the major pattern of hydrogen bondbetween cellulose and urea, the angular number distributionbetween cellulose backbone and the C]O bonds of urea arecalculated, taking a 0.3 nm distance cutoff which is calculatedbetween the oxygen atoms of urea and the nearest hydroxyl oxygenatom (O (A), O (B) or O (C)) of cellulose. Both of O (A) and O (B)locate in backbone of cellulose and are considered together becauseof their similar locations. The other hydroxyl oxygen is O (C), whichis linked with the side chain methylene. The peak tendency andlocation of N(q) is similar at different temperatures, so the distri-bution of N(q) at 261 K is taken as an example for discussion of thehydrogen bonding types which are only correlative with the peaktendency and location. As shown in Fig. 4a, stochastic distributionsfor two kinds of oxygen atoms were distinguished because thenumber of urea molecules surrounding them is different within the0.3 nm cutoff distance. It is clear that stochastic distributionwithout orientation correlationwill be a sine curve (blue and greenline in Fig. 4a). Due to O (C) linked with the side chain methylenecould circumrotate in all angles from 0� to 180� with a dihedralangle, the interaction between urea molecules and O (C) of cellu-lose by hydrogen bonding in any angle have equal probabilities.This deduction have been confirmed by angular number distribu-tion, in which the N(q) distribution for O (C) is consistent with thestochastic distribution (black and blue curves in Fig. 4a). Comparedwith O (C), the N(q) distribution for O (A) and O (B) shows a strongorientation correlation with the peak around 90� and this ordereddistribution of urea surrounding the oxygen atoms in cellulosebackbone might induced by the effect of steric hindrance.

For further explanation of the N(q) distribution for O (A) and O(B), the supposed scheme of angular orientation of O(A)eO, O(B)eOhydrogen bond was shown in Fig. 4b. Due to one urea moleculehave two nitrogen atoms with symmetric location, there shouldexist two equivalent-value and axial symmetric peaks with 90� asthe central axis in angular number distribution (red curve) if mosthydrogen bonds formed by oxygen atoms of cellulose or nitrogenatoms of urea as proton-acceptors. These two supposed hydrogenbonding patterns have been shown in Fig. 4b (a and g). However,from our results in angular number distribution (Fig. 4a), these two

(b) three hydroxyl oxygen atoms of cellulose with oxygen atoms of urea at 261 K.

Fig. 5. RDF curves of oxygen atoms of cellulose with water in the solvent with (blackcurves) and without urea (red curves) at 261 K. (For interpretation of the references tocolour in this figure legend, the reader is referred to the web version of this article.)

Fig. 4. (a) The curve of angular number distribution between cellulose and the C]O bonds of urea (black curve for O(C); red curve for O(A) and O(B)). The stochastic distributions ofangular number without orientation correlation were used as control (blue curve for O(C) and green curve for O(A) and O(B)). (b) The supposed scheme of angular orientation. (Forinterpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

L. Cai et al. / Polymer 53 (2012) 1124e1130 1127

peaks mentioned above do not exist. This result was also consistentwith the conclusion of RDF in Fig. 3, i.e. there is one dominanthydrogen bonding between cellulose and urea, in which oxygenatoms of urea as proton-acceptors and hydroxyl hydrogen atoms ofcellulose as proton-donors. This kind of hydrogen bonding impliesthat functions of urea to cellulose may result from its conjugateddouble-bond and other molecules with the similar structure, suchas thiourea, could also be applied as an alternative for urea topromote the dissolution of hydrogen bonded polymers [40,41].

3.2. Inclusion layer of urea to cellulose

The simulation data from RDF and angular number distributionhave, to great extent, confirmed that the interaction of oxygenatoms in urea with hydroxyl hydrogen atoms in cellulose are thedominant pattern of hydrogen bonding between cellulose and urea.However, there exists competition between cellulose-urea andcellulose-water interactions when cellulose in urea-water mixedsolution, in which cellulose chain can also interact with watermolecules to form hydrogen bond networks. In order to get a clearunderstanding of this competition, RDF curves between celluloseand water in solvent with and without urea at 261 K are calculated.As shown in Fig. 5, the first peaks in RDF curve indicate the firstwater layer. It is quite clear that a great part of water molecules infirst water layer when cellulose in water only solution, whichindicate the hydrogen bond formed by water and cellulose (redcurve). However, first peak drop significantly when cellulose inurea-water solventmixture (black curve), which indicate that manywater molecules have been pushed away from cellulose chain andreplaced by urea molecules. Combining this result with RDF data ofcellulose and urea, as shown in Fig. 3b, the first peaks with sharpprofile and high value are the abundant evidence that most ureamolecules interact with cellulose to form the first inclusion layer onthe surface of cellulose, which may induce the conformationalteration of cellulose chain and further affect the hydrophilic-hydrophobic property of cellulose in urea-water solution mixture.

3.3. Temperature property of the inclusion layer

The temperature property of interaction between cellulose andurea in urea-water solvent mixture is further investigated by RDF.As the peak tendency and location of RDF between three hydroxyloxygen atoms of cellulose and oxygen atoms of urea are almost thesame (as shown in Fig. 3b), these three kinds of hydroxyl oxygenatoms are considered together in its entirety in following discus-sion. The RDF curves of hydroxyl oxygen atoms of cellulose withoxygen atoms of urea at different temperatures are shown in

Fig. 6a, in which all the RDF curves have similar trend with theposition of peaks, but peak values vary with temperature. Firstpeaks located around 0.25 nm not only indicate hydrogen bondinginteraction between cellulose and oxygen atoms of urea as proton-acceptors, but also confirm the formation of the inclusion layer. Thesecond peaks around 0.5 nm indicate hydrogen bonding betweencellulose and hydrogen atoms of urea as proton-donors or nitrogenatoms as proton-acceptors or both of them. These two peaks havethe same changing trend with the temperature, which suggest thatthe temperature property of hydrogen bonding between celluloseand urea is independent on the types of hydrogen bonds betweenthem. In order to further clarify the temperature property of thisinteraction, the value of first peaks are shown in Fig. 6b. Below265 K, there is no correlation and no significant difference betweenthe peak values and temperature, as well as those of peaks above283 K. However, the peak value decreases gradually with temper-ature increasing from 265 to 283 K. The peak values vary withtemperature increasing are well fitted by Boltzmann function,which is shown in Fig. 6b. The g(r) values toward a certain constantwhen temperature is lower than 265 K or higher than 283 K, whichindicates that the inclusion effect of urea to cellulose was relativelystable and would not change in these temperature ranges. At thetemperature range from 265 to 283 K, g(r) values decrease with thetemperature increases indicates that the hydrogen bondingbetween cellulose and urea is weaken and the inclusion function ofurea is inhibited within this temperature range.

Fig. 6. RDF curves of (a) hydroxyl oxygen atoms of cellulose with oxygen atoms of urea, inset: the magnified first peaks and (b) the first peaks values at different temperatures, dataindicate Mean; bars, �SD (n ¼ 3).

L. Cai et al. / Polymer 53 (2012) 1124e11301128

Zhang L. et al. reported that cellulose can dissolve in pre-cooledLiOH/urea aqueous solution at 261 K [17], in which the LiOHhydrogen bonded cellulose-urea inclusion complexes was formedat low temperature. From our simulation result in Fig. 6b, it isindeed favorable for the formation of hydrogen bond at lowtemperature, which indicates that hydrogen binding betweencellulose and urea play as a key role in the inclusion complexes, andthis effect of inclusion can promote cellulose dissolution in urea-containing systems. One could speculate that temperature isa restrictive factor for the inclusion effect of urea to cellulose, inwhich hydrogen bonding between cellulose and urea play impor-tant roles during the dissolution process of cellulose. However, thisinteraction between cellulose and urea will be inhibited in sometemperature range with the temperature increasing. Therefore, theprocess of pre-cooling would be a necessary approach for thedissolution of cellulose in urea-containing solution mixture. The

Fig. 7. Interaction energy as a function of temperature between (a) cellulose and urea, (b) c(d) cellulose and solvent molecules in urea-water solvent mixture. Data indicate Mean; ba

cellulose chain will be separated after the formation of inclusionlayer and the effect of inclusion will not disappear completely withthe temperature increasing. In the previous study, it was alsoconfirmed that cellulose dissolving in solvent mixture will not beaggregated at ordinary temperature [17].

In order to further investigate the temperature properties ofcellulose-urea interaction, the interaction energies of cellulose-urea and cellulose-water in urea-water solvent mixture, as wellas that of cellulose-water in solvent without urea were calculated.As shown in Fig. 7a and b, compared with interaction energybetween cellulose and urea in solvent mixture, interaction energybetween cellulose and water is much higher at the same temper-ature, which indicates that the cellulose chain would preferentiallyinteract with urea molecules than water molecules. Moreover, theinteraction energy of cellulose-solvent (including water and urea,Fig. 7d) is also obviously smaller than cellulose-water in water

ellulose and water in urea-water solvent mixture, (c) cellulose and water in water andrs, �SD (n ¼ 3).

Table 1The number of cellulose-urea and cellulose-water hydrogen bonds in the solvent mixture and cellulose-water hydrogen bonds in water at different temperatures.

T (K) Urea-water solvent mixture Water

Cellulose-urea (N þ O)a Cellulose-urea (O)a Cellulose-water in the presence of urea Cellulose-water in the absence of urea

253 41.74 � 0.05 41.66 � 0.05 22.54 � 0.04 60.03 � 0.04257 41.46 � 0.05 41.38 � 0.05 21.81 � 0.04 61.22 � 0.04261 40.84 � 0.06 40.75 � 0.06 22.21 � 0.05 58.62 � 0.04265 41.39 � 0.06 41.30 � 0.06 22.15 � 0.04 58.51 � 0.04269 39.41 � 0.05 39.33 � 0.05 23.29 � 0.04 60.45 � 0.04283 37.22 � 0.05 37.12 � 0.05 24.77 � 0.04 58.03 � 0.04297 36.68 � 0.05 36.58 � 0.05 24.73 � 0.05 59.25 � 0.04

a N þ O means both nitrogen and oxygen atoms are acceptors, O means only oxygen atoms are acceptors.

L. Cai et al. / Polymer 53 (2012) 1124e1130 1129

(Fig. 7c) and this also demonstrates that more stable interactionbetween urea and cellulose.

For the cellulose-urea interaction shown in Fig. 7a, the energyincreases with temperature increasing from 265 to 283 K, but isindependent on temperature below 265 K or above 283 K. It isworth to mention that cellulose-water interaction in water isindependent of temperature (Fig. 7c). In addition, Idrissi A et alhave confirmed that there is also no relation between urea-waterinteraction and temperature [31], Therefore, this temperatureproperty of cellulose-urea interaction will not affected by interac-tion of cellulose-water or urea-water. In the previous study, it wasconfirmed that urea molecules can act as the “host” to form anordered inclusion channel in typical inclusion compounds and thehydrogen bonding interaction between urea and the guest polymerwould vary with temperature in a temperature range [42].Compared with those studies, the urea inclusion layer described inour system has the almost some temperature properties, eventhough there does not exist the ordered inclusion channel [43e45].

Compared with cellulose-water interaction in water (Fig. 7c),cellulose-water interaction in solution mixture are correlated withtemperature (Fig. 7b), in which energy decreases with temperatureranged from 265 to 283 K. This abnormal phenomenon indicatedthat the interaction between cellulose and urea is the dominantdriving forces in the system. Water molecules could replace urea toform hydrogen bonds with cellulose as the cellulose-urea interac-tion decrease. Both cellulose-urea and cellulose-water interactionenergy results are close to Boltzmann function cure (Fig. 7a and b).

The hydrogen bond number of cellulose with solvent moleculesat different temperatures was listed in Table 1. As mentioned inFig. 3, there are two kinds of hydrogen bond proton-acceptors:oxygen and nitrogen atoms. There is no statistically significant inthe number of hydrogen bond between both of nitrogen andoxygen atoms as acceptors (the second column) and only oxygenatoms as acceptors (the third column), which indicate thathydrogen bonding formed by nitrogen atoms as proton-acceptorsis negligible. This result gives an additional proof to the conclusionof Fig. 3, i.e. there are only two patterns of hydrogen bonds in oursystem. One is oxygen atoms of cellulose as proton-acceptors andhydrogen atoms of urea as proton-donors, the other is carbonyloxygen atoms of urea serve as proton-acceptors and hydroxylhydrogen atoms of cellulose serve as proton-donors. At eachtemperature, the hydrogen bond number surrounding the cellu-lose chain is between 60 and 65 (the sum of the second columnand the fourth column) in solvent mixture. However, the two-thirds of the hydrogen bonds are formed by cellulose-urea (thesecond column), which also indicate that urea molecules arefavored to hydrogen bonding to cellulose. The hydrogen bondnumber in solvent mixture is almost same with that in water (thefifth column), which suggest that urea can replace water mole-cules to form more stable hydrogen bonding with cellulose andwould decrease the interaction between cellulose themselves andfacilitate dissolution of cellulose in urea/water solvent mixture.

The tendency of cellulose-urea hydrogen bond number withtemperature increasing is well consistent with the RDF results andthe hydrogen bond number of cellulose-water increases withinthe temperature from 265 K to 283 K, which also agrees with theinteraction energy results.

4. Conclusions

The interaction between a single cellulose chain and ureamolecules in urea-water solvent mixture was studied by means ofmolecular dynamics simulation. There are more than 60 hydrogenbonds surrounding one cellulose chain consisting of 12 monomers,the participation of urea molecules account for two-thirds. Thissimulation results indicate that in contrast with water molecules,the cellulose chain prefers to form hydrogen bonds with ureamolecules. The more stable and stronger interaction between ureaand cellulose is favorable for the formation of inclusion layerwhich could decrease the self-interaction of cellulose chains andpromote dissolution of cellulose in urea-containing solventmixture. In the other hand, there is only one dominant way forhydrogen bond formation between cellulose and urea molecules(oxygen atoms of urea as proton-acceptors and hydroxyl hydrogenatoms of cellulose as proton-donors). The unique phenomenonindicate that functions of urea to cellulose may result from itsconjugated bond, and urea molecules would be replaced by othermolecules with the similar structure to promote the dissolution ofhydrogen bonded polymers. Furthermore, the inclusion layer isdependent of the temperature. The hydrogen bonding keep itsstability below 265 K and the hydrogen bonding interactiondecrease with increasing temperature in the range of 265e283 K,which indicate temperature is a major factor for urea molecules toform inclusion layer surrounding the cellulose chain and relativelylow temperature should be applied to make sure urea moleculescan make adequate contribution in the process of cellulosedissolution. Our simulation results is agreement with the previousexperimental results and provide some significant information tounderstand the interaction between hydrogen bonded polymersand small molecules.

Acknowledgments

This workwas supported by the Program of the National NaturalScience Foundation of China (Nos. 20874094 and 20934004),NBRPC (No. 2010CB934500), and the “Bairen” fund of CAS.

References

[1] Pace CN. Methods in Enzymology 1986;131:266e80.[2] Makhatadze GI, Privalov PL. Journal of Molecular Biology 1992;226(2):

491e505.[3] Swatloski RP, Spear SK, Holbrey JD, Rogers RD. Journal of the American

Chemical Society 2002;124(18):4974e5.[4] Zou Q, Bennion BJ, Daggett V, Murphy KP. Journal of the American Chemical

Society 2002;124(7):1192e202.

L. Cai et al. / Polymer 53 (2012) 1124e11301130

[5] Chen LJ, Qin YS, Wang XH, Zhao XJ, Wang FS. Polymer 2011;52(21):4873e80.[6] Panuszko A, Bruzdziak P, Zielkiewicz J, Wyrzykowski D, Stangret J. Journal of

Physical Chemistry B 2009;113(44):14797e809.[7] Ishida T, Rossky PJ, Castner EW. Journal of Physical Chemistry B 2004;108(45):

17583e90.[8] Chakraborty A, Sarkar M, Basak S. Journal of Colloid and Interface Science

2005;287(1):312e7.[9] Wallqvist A, Covell DG, Thirumalai D. Journal of the American Chemical

Society 1998;120(2):427e8.[10] Zou Q, Habermann-Rottinghaus SM, Murphy KP. Proteins-Structure Function

and Genetics 1998;31(2):107e15.[11] Ikeguchi M, Nakamura S, Shimizu K. Journal of the American Chemical Society

2001;123(4):677e82.[12] Liu HB, Sale KL, Holmes BM, Simmons BA, Singh S. Journal of Physical

Chemistry B 2010;114(12):4293e301.[13] Dadi AP, Varanasi S, Schall CA. Biotechnology and Bioengineering 2006;95(5):

904e10.[14] Dadi AP, Schall CA, Varanasi S. Applied Biochemistry and Biotechnology 2007;

137:407e21.[15] Bergenstrahle M, Wohlert J, Himmel ME, Brady JW. Carbohydrate Research

2010;345(14):2060e6.[16] Cai J, Zhang LN, Zhou JP, Qi HS, Chen H, Kondo T, et al. Advanced Materials

2007;19(6):821e6.[17] Cai J, Zhang LN, Chang CY, Cheng GZ, Chen XM, Chu B. ChemPhysChem 2007;

8(10):1572e9.[18] Harris KDM, Thomas JM. Journal of the Chemical Society-Faraday Transactions

1990;86(17):2985e96.[19] Harris KDM. Chemical Society Reviews 1997;26(4):279e89.[20] Harris KDM. Phase Transitions 2003;76(3):205e18.[21] Kristensen JH, Bampos N, Duer M. Physical Chemistry Chemical Physics 2004;

6(12):3175e83.[22] Lue A, Liu YT, Zhang L, Potthas A. Polymer 2011;52(17):3857e64.[23] Yang QL, Qin XZ, Zhang LN. Cellulose 2011;18(3):681e8.[24] Heiner AP, Sugiyama J, Teleman O. Carbohydrate Research 1995;273(2):

207e23.[25] Tanaka F, Okamura K. Cellulose 2005;12(3):243e52.

[26] Chen W, Lickfield GC, Yang CQ. Polymer 2004;45(21):7357e65.[27] Hardy BJ, Sarko A. Polymer 1996;37(10):1833e9.[28] Humeres E, Mascayano C, Riadi G, Gonzalez-Nilo F. Journal of Physical Organic

Chemistry 2006;19(12):896e901.[29] Bertran CA, Cirino JJV, Freitas LCG. Journal of the Brazilian Chemical Society

2002;13(2):238e44.[30] Caballero-Herrera A, Nilsson L. Journal of Molecular Structure-Theochem

2006;758(2e3):139e48.[31] Idrissi A, Cinar E, Longelin S, Damay P. Journal of Molecular Liquids 2004;

110(1e3):201e8.[32] Souaille M, Smith JC, Guillaume F. Journal of Physical Chemistry B 1997;

101(34):6753e7.[33] Van der Spoel D, Lindahl E, Hess B, Groenhof G, Mark AE, Berendsen HJC.

Journal of Computational Chemistry 2005;26(16):1701e18.[34] Berendsen HJC, Vanderspoel D, Vandrunen R. Computer Physics Communi-

cations 1995;91(1e3):43e56.[35] Berendsen HJC, Grigera JR, Straatsma TP. Journal of Physical Chemistry 1987;

91(24):6269e71.[36] Hess B, Bekker H, Berendsen HJC, Fraaije J. Journal of Computational Chem-

istry 1997;18(12):1463e72.[37] Darden T, York D, Pedersen L. Journal of Chemical Physics 1993;98(12):

10089e92.[38] Berendsen HJC, Postma JPM, van Gunsteren WF, DiNola A, Haak JR. Journal of

Chemical Physics 1984;81(8):3684e90.[39] Mahoney MW, Jorgensen WL. Journal of Chemical Physics 2001;114(1):

363e6.[40] Chen XM, Burger C, Fang DF, Ruan D, Zhang L, Hsiao BS, et al. Polymer 2006;

47(8):2839e48.[41] Dong R, Ang L, Lina Z. Polymer 2008;49(4):1027e36.[42] Brustolon M, Maniero AL, Marcomini A, Segre U. Journal of Materials Chem-

istry 1996;6(10):1723e9.[43] Oertel U, Komber H, Tenkovtsev AV, Dudkina MM, Trofimov AE, Bohme F.

Polymer 2009;50(25):5876e83.[44] Saldias C, Gargallo L, Sandoval C, Leiva A, Radic D, Caballero J, et al. Polymer

2009;50(13):2926e32.[45] Tu CW, Kuo SW, Chang FC. Polymer 2009;50(13):2958e66.