Purdue University
Purdue e-Pubs
Birck and NCN Publications
Birck Nanotechnology Center
4-2011
Calculation of single chain cellulose elasticity using
fully atomistic modeling
Xiawa Wu
Birck Nanotechnology Center, Purdue University, [email protected]
Robert Moon
Birck Nanotechnology Center, Purdue University, [email protected]
Ashlie Martini
Purdue University, [email protected]
Follow this and additional works at: http://docs.lib.purdue.edu/nanopub
Part of the Nanoscience and Nanotechnology Commons
Wu, Xiawa; Moon, Robert; and Martini, Ashlie, "Calculation of single chain cellulose elasticity using fully atomistic modeling" (2011).
Birck and NCN Publications. Paper 1029.
http://docs.lib.purdue.edu/nanopub/1029
This document has been made available through Purdue e-Pubs, a service of the Purdue University Libraries. Please contact [email protected] for
additional information.
PEER-REVIEWED
MOLECULAR MODELING
Calculation of single chain cellulose
elasticity using fully atomistic modeling
XIAWA WU, ROBERT J. MOON,
AND
ASHLIE MARTINI
ABSTRACT: Cellulose nanocrystals, a potential base material for green nanocomposites, are ordered bundles of
cellulose chains. The properties of these chains have been studied for many years using atomic-scale modeling.
However, model predictions are difficult to interpret because of the significant dependence of predicted properties
on model details. The goal of this study is to begin to understand these dependencies. We focus on the investigation
on model cellulose chains with different lengths and having both periodic and nonperiodic boundary conditions, and
predict elasticity in the axial (chain) direction with three commonly used calculation methods. We find that chain
length, boundary conditions, and calculation method affect the magnitude of the predicted axial modulus and the
uncertainty associated with that value. Further, the axial modulus is affected by the degree to which the molecule is
strained. This result is interpreted in terms of the bonded and nonbonded contributions to potential energy, with a
focus on the breaking of hydrogen bonds during deformation.
Application: This study lays the groundwork for understanding the predictions of atomistic models and to help
transition their use from an investigative method to a predictive tool useful for cellulose-based application design.
WRPLVWLF PRGHOLQJ RI FHOOXORVH KDV EHHQ XVHG WR
FRPSOHPHQW H[SHULPHQWDO PHDVXUHPHQWV RI FHO
OXORVH QDQRFU\VWDOV E\ SURYLGLQJ WKH DWRPLF OHYHO GHWDLO
QHFHVVDU\ WR SUHGLFW VWUXFWXUDO HQHUJHWLF DQG PHFKDQLFDO
FKDUDFWHULVWLFV DQG WR JDLQ D IXQGDPHQWDO XQGHUVWDQGLQJ
RI WKH DWRPLFVFDOH RULJLQV RI WKHVH FKDUDFWHULVWLFV 6XFK
PRGHOV IUHTXHQWO\ DUH XVHG WR SUHGLFW HODVWLF SURSHUWLHV
EHFDXVH WKH VLPXODWLRQ PHWKRGV DUH UHODWLYHO\ VLPSOH DQG
UHVXOWV FDQ EH FRPSDUHG WR H[SHULPHQWDO GDWD 3UHGLFWLRQV
RI HODVWLFLW\ LQ WKH D[LDO FKDLQ GLUHFWLRQ IRU WKH OD RU ,‰
SRO\PRUSKV RI FU\VWDOOLQH FHOOXORVH KDYH EHHQ UHSRUWHG
IURP DWRPLVWLF PRGHOEDVHG VWXGLHV IRU PRUH WKDQ \HDUV
>O@ 8QIRUWXQDWHO\ TXDQWLWDWLYH FRPSDULVRQ RI HODVWLFLW\
UHVXOWV IURP PRGHOWRPRGHO RU WR H[SHULPHQWDO PHDVXUH
PHQWV KDV EHHQ GLIILFXOW EHFDXVH RI WKH VLJQLILFDQW HIIHFW
RI YDULDWLRQV LQ WKH VLPXODWLRQ PHWKRGV DWRPLF LQWHUDFWLRQ
PRGHOV DQG FRQILJXUDWLRQ RI WKH PRGHOHG VWUXFWXUHV )RU
H[DPSOH WKH D[LDO PRGXOXV SUHGLFWHG XVLQJ D PRGHO RI
[[ ,‰ XQLW FHOOV ZDV a VPDOOHU WKDQ WKDW SUHGLFWHG
XVLQJ D PRGHO RI [[ ,‰ XQLW FHOOV >@ ,Q DQRWKHU VWXG\
D GLIIHUHQFH LQ WKH LQLWLDO HTXLOLEULXP XQLW FHOO OHQJWK RI
FDXVHG D YDULDWLRQ LQ WKH D[LDO PRGXOXV SUHGLF
WLRQ >@
7KH JRDO RI WKLV SDSHU LV WR FOHDUO\ LOOXVWUDWH VRPH RI WKHVH
HIIHFWV VXFK WKDW WKH PHDQLQJ RI SUHGLFWLRQV PDGH XVLQJ DW
RPLVWLF PRGHOV FDQ EH XQGHUVWRRG PRUH IXOO\ :H UHSRUW UH
VXOWV RI D PROHFXODU G\QDPLFV DQG PROHFXODU PHFKDQLFV
EDVHG VWXG\ IRFXVLQJ RQ WKH PRVW EDVLF VWUXFWXUH D VLQJOH
FHOOXORVH PROHFXOH WR XQGHUVWDQG WKH HIIHFWV RI VRPH YDUL
DEOHV RQ PRGHO SUHGLFWLRQV
0DQ\ FKDLQV RI GLIIHUHQW OHQJWKV FRH[LVW LQ D EXON SRO\PHU
DQG VR WKH YDULDWLRQ RI VWUXFWXUH HQHUJHWLF DQG HODVWLF SURS
A
HUWLHV ZLWK FKDLQ OHQJWK LV LQVLJQLILFDQW +RZHYHU DW WKH QD
QRVFDOH SURSHUWLHV DUH OLNHO\ WR EH LQIOXHQFHG E\ VRFDOOHG HQG
HIIHFWV ZKHUH WKH PROHFXOH LV VKRUW HQRXJK WKDW WKH FRQWULEX
WLRQ WKH HQGV RI WKH FKDLQV PDNH LV FRPSDUDEOH WR WKDW RI WKH
PDLQ ERG\ RI WKH FKDLQ 7R FKDUDFWHUL]H WKLV HIIHFW ZH IR
FXVHG WKLV VWXG\ RQ PRGHO FHOOXORVH FKDLQV ZLWK GLIIHUHQW
OHQJWKV DQG KDYLQJ ERWK SHULRGLF DQG QRQSHULRGLF ERXQGDU\
FRQGLWLRQV 7KHVH PRGHOV DUH XVHG WR SUHGLFW HODVWLFLW\ LQ WKH
D[LDO GLUHFWLRQ ZLWK WKUHH FRPPRQW\ XVHG FDOFXODWLRQ PHWK
RGV 5HVXOWV DUH GLVFXVVHG LQ WHUPV RI WKH HIIHFWV RI FKDLQ
OHQJWK ERXQGDU\ FRQGLWLRQV FDOFXODWLRQ PHWKRG DQG VWUDLQ
RQ WKH PRGHOSUHGLFWHG D[LDO PRGXOXV DQG WKH XQFHUWDLQO\
DVVRFLDWHG ZLWK WKRVH SUHGLFWLRQV
METHODS
:H FUHDWHG VLQJOH FKDLQ FHOOXORVH PROHFXOHV ZLWK OHQJWKV IURP
RQH WR UHSHDW XQLWV ZKHUH RQH XQLW FRQWDLQV WZR JOXFRVH
ULQJV DQG KDV D OHQJWK RI L (Fig. 1). Figure 2 VKRZV DQ H[
DPSOH FRQVLVWLQJ RI IRXU UHSHDW XQLWV 7R FUHDWH FKDLQV RI ILQLWH
OHQJWK D VLQJOH PROHFXOH ZDV SODFHG LQ WKH PLGGOH RI D VLPX
ODWLRQ ER[ ZLWK GLPHQVLRQV ; ; n [ L c LQILQLWH
1. Schematic of single cellulose chain repeat unit, showing the
directionally of the 1 4 linkage and intramolecule hydrogen
bonding (dotted lines).
APRIL 201 1 | TAPPI JOURNAL
37
MOLECULAR MODELING
2. Snapshots of a model cellulose chain at its equilibrium
and strained lengths where n is the number of repeat units.
Atoms represented as grey (carbon), red (oxygen), and white
(hydrogen)spheres.
OHQJWK PROHFXOHV ZHUH REWDLQHG E\ DGGLQJ D FRYDOHQW ERQG
DFURVV WKH SHULRGLF ERXQGDU\ OLQNLQJ WKH KHDG FDUERQ DQG WDLO
R[\JHQ DWRPV ,Q WKLV FDVH WKH PROHFXOH LV LQILQLWHO\ ORQJ LQ
LWV D[LDO GLUHFWLRQ DQG LVRODWHG ,Q WKH RWKHU WZR GLUHFWLRQV
VXFK WKDW WKH VLPXODWLRQ ER[ LV [[ n [ L c 7KH ODWHUDO
GLPHQVLRQV RI WKH VLPXODWLRQ ER[ ZHUH FKRVHQ WR EH c WR
HQVXUH WKDW WKH PROHFXOH GLG QRW LQWHUDFW ZLWK SHULRGLF LP
DJHV RI LWVHOI GXULQJ WKH VLPXODWLRQ WDNLQJ LQWR FRQVLGHUDWLRQ
DOO SRVVLEOH EHQGLQJ URWDWLQJ DQG RWKHU PRYHPHQW RI WKH PRO
HFXOH $WRPLF LQWHUDFWLRQV ZHUH GHVFULEHG E\ WKH &203$66
IRUFH ILHOG >@ ZLWK DOO LQWHUDFWLRQ FXWRII GLVWDQFH RI c 7KLV
IRUFH ILHOG ZDV FKRVHQ EHFDXVH LW KDV EHHQ DSSOLHG VXFFHV
IXOO\ WR PRGHO FHOOXORVLF PDWHULDOV SUHYLRXVO\ >@
,QLWLDO SRVLWLRQV DQG FKDUJHV IRU HDFK DWRP ZHUH FRPSXWHG
DQG WKHQ FRQYHUWHG WR D /$0036 ODUJHVFDOH DWRPLFPROHFX
ODU PDVVLYHO\ SDUDOOHO VLPXODWRU LQSXW ILOH $OO VXEVHTXHQW VLP
XODWLRQV ZHUH SHUIRUPHG XVLQJ /$0036 >@ (DFK LQLWLDO FRQ
ILJXUDWLRQ ZDV HTXLOLEUDWHG XVLQJ PROHFXODU G\QDPLFV LQ WKH
197 HQVHPEOH FRQVWDQW QXPEHU RI DWRPV YROXPH DQG P
SHUDWXUH DW D WHPSHUDWXUH RI . IRU SV EHIRUH SURGXF
WLRQ UXQV XVLQJ PROHFXODU PHFKDQLFV 7R HYDOXDWH WKH VWDWLVWL
FDO VLJQLILFDQFH RI FDOFXODWHG UHVXOWV HDFK VLPXODWLRQ FDVH
FKDLQ OHQJWK DQG ERXQGDU\ FRQGLWLRQ ZDV UXQ WLPHV ZLWK
VDPH LQLWLDO DWRPLF VWUXFWXUH EXW GLIIHUHQW LQLWLDO YHORFLWLHV
UDQGRPL]HG
PRPHQWD
7R HQDEOH FDOFXODWLRQ RI WKH D[LDO PRGXOXV HDFK PRGHO ZDV
UHSHDWHGO\ GHIRUPHG LQ VPDOO LQFUHPHQWV LQ WKH FKDLQ GLUHF
WLRQ DV LOOXVWUDWHG LQ )LJ 7KH HTXLOLEULXP OHQJWK L RI D UH
SHDW XQLW DQG WKH FRUUHVSRQGLQJ PLQLPXP SRWHQWLDO HQHUJ\
U ZHUH FDOFXODWHG XVLQJ PROHFXODU PHFKDQLFV ZLWK WKH FRQ
MXJDWH JUDGLHQW HQHUJ\ PLQLPL]DWLRQ DOJRULWKP 3DUDPHWHUV
WKDW TXDQWLI\ FODVWLF SURSHUWLHV ZLWK DQG ZLWKRXW WKH DVVXPS
WLRQ RI FRQVWDQW DUHD KDYH EHHQ GLIIHUHQWLDWHG DV ³FKDLQ VWLII
QHVV´ DQG ³FKDLQ PRGXOXV´ UHVSHFWLYHO\ >@ +RZHYHU WKLV LV
QRW D VLJQLILFDQW LVVXH IRU D VLQJOH FKDLQ LQ D PXFK ODUJHU VLPX
ODWLRQ ER[ ZKHUH WKH FURVVVHFWLRQDO DUHD LV WDNHQ WR EH LWV RF
38
TAPPI JOURNAL | APRIL 2011
3. Equilibrium unit cell length (red circles) and potential energy
(black squares) for finite and infinite chains as functions of
chain length. Each data point represents the average value of 10
repeated simulations.
FXSLHG DUHD LQ D XQLW FHOO D YDOXH WKDW FKDQJHV OLWWOH GXULQJ WKH
GHIRUPDWLRQ 7KHUHIRUH ZH DSSOLHG VWUDLQV XS WR ,Q WKH
FHOOXORVH ,‰ VWUXFWXUH RQH XQLW FHOO FRQVLVWLQJ RI WZR PROH
FXOHV KDV D FURVVVHFWLRQDO DUHD RI c 7KXV DQ DSSUR[L
PDWH DUHD IRU D VLQJOH PROHFXOH LV KDOI RI WKLV YDOXH RU c 7KLV LV WKH YDOXH XVHG LQ WKH HODVWLFLW\ FDOFXODWLRQV :H DOVR
QRWH WKDW LQ WKH SK\VLFDO V\VWHP LQWUDPROHFXODU FRYDOHQW
ERQGV ZRXOG EUHDN DW VRPH VWUDLQ +RZHYHU IHZ IRUFH ILHOGV
LQFOXGLQJ WKH RQH XVHG KHUH FDQ DFFRXQW IRU EUHDNLQJ RI FR
YDOHQW ERQGV ,QWURGXFLQJ WKLV FDSDELOLW\ LQWR WKH PRGHO LV DQ
DFWLYH DUHD RI UHVHDUFK
,QLWLDOO\ DOO ILQLWH DQG LQILQLWH FKDLQV KDYH WKH VDPH XQLW
FHOO OHQJWK c >@ +RZHYHU WKH HTXLOLEULXP OHQJWK FDO
FXODWHG GXULQJ WKH GHIRUPDWLRQ SURFHVV ZDV IRXQG WR H[KLELW
VRPH YDULDWLRQ EHWZHHQ c DQG c (Fig. 3). 7KLV LV FRQ
VLVWHQW ZLWK D SUHYLRXV REVHUYDWLRQ WKDW DQ LVRODWHG VLQJOH FHO
OXORVH PROHFXOH ZLOO KDYH D ODUJHU HTXLOLEULXP OHQJWK WKDQ WKDW
RI D FHOOXORVH QDQRFU\VWDO >@ )LJXUH DOVR LOOXVWUDWHV WKDW WKH
SRWHQWLDO HQHUJ\ RI ILQLWH DQG LQILQLWH PROHFXOHV LV NFDO
PROJOXFRVH 7KXV ZH REVHUYH WKDW WKH HTXLOLEULXP XQLW FHOO
OHQJWK DQG SRWHQWLDO HQHUJ\ DUH QRW VWURQJ IXQFWLRQV RI ERXQG
DU\ FRQGLWLRQV RU WKH QXPEHU RI UHSHDW XQLWV ,Q DGGLWLRQ VLJ
QLILFDQW YDULDWLRQ LQ WKH GDWD IRU VKRUW FKDLQV LQGLFDWH WKDW DW
OHDVW VL[ UHSHDW XQLWV VKRXOG EH XVHG IRU FDOFXODWLRQV EDVHG RQ
4. A representative plot of potential energy (left y-axis;
normalized by 6202 kcal/mol) and corresponding number of
hydrogen bonds (right y-axis; normalized by 13) as functions
of strain. Data from a simulation of an infinite chaing with eight
repeat units.
PRGHOSUHGLFWHG OHQJWK DQG HQHUJ\ ,Q WKH ILQLWH FKDLQ FDVH
HUURU LQ VKRUW FKDLQ FDOFXODWLRQV LV WKH UHVXOW RI PROHFXOH HQG
HIIHFWV >@ DQG WKH FRQILJXUDWLRQ EHLQJ GHSHQGHQW RQ IHZHU
DWRPV 7KH ODWWHU HIIHFW LV WKH SULPDU\ VRXUFH RI HUURU RE
VHUYHG IRU LQILQLWH FKDLQV ZLWK D VPDOO QXPEHU RI UHSHDW XQLWV
IHFW RI K\GURJHQ ERQGV FDQ FDXVH SUHGLFWHG HODVWLF SURSHUWLHV
WR GHFUHDVH E\ WR >@ $QRWKHU VWXG\ IRXQG WKDW UH
PRYDO RI K\GURJHQ ERQGV UHVXOWHG LQ D GHFUHDVH RI EHWZHHQ
DQG LQ ,Į FKDLQ VWLIIQHVV ZKLOH IRU ,‰ UHPRYDO FDXVHG
HLWKHU D GHFUHDVH RU D LQFUHDVH LQ FKDLQ VWLIIQHVV GH
SHQGLQJ RQ WKH LQLWLDO DWRPLF FRRUGLQDWHV >@ ,Q WKH ORZ VWUDLQ
UHJLPH RXU REVHUYDWLRQV DUH FRQVLVWHQW ZLWK WKHVH ILQGLQJV
+RZHYHU RXU UHVXOWV DOVR LQGLFDWH WKDW WKH GHSHQGHQFH RI
HODVWLF EHKDYLRU RQ K\GURJHQ ERQGLQJ VKRXOG EH OHVV IRU HODV
WLFLW\ FDOFXODWHG DW ODUJH VWUDLQ UDWHV
7KH D[LDO PRGXOXV XQLWV 3D FDQ EH REWDLQHG XVLQJ WKUHH
PHWKRGV $OWKRXJK WKH IXQFWLRQDO IRUPV RI WKH PHWKRGV DUH
HTXLYDOHQW ZKLFK SDUDPHWHUV DUH REWDLQHG IURP VLPXODWLRQ
GLIIHUV DQG VR WKH VRXUFH RI FDOFXODWLRQ HUURU PD\ GLIIHU DV
ZHOO 1RWH WKDW DOO WKUHH PHWKRGV DUH LQ VRPH ZD\ GHSHQGHQW
RQ DQ HVWLPDWLRQ RI DUHD ZKLFK DV GLVFXVVHG SUHYLRXVO\ LV DQ
DSSUR[LPDWLRQ 7KH ILUVW PHWKRG >@ FDQ EH H[SUHVVHG DV
ZKHUH WKH IRUFH F XQLWV 1 DQG PROHFXOH HTXLOLEULXP OHQJWK
L XQLWV P DUH REWDLQHG E\ ILWWLQJ D SRO\QRPLDO WR SRWHQWLDO
HQHUJ\ U XQLWV - PROHFXOH OHQJWK L XQLWV P LV REWDLQHG
IURP WKH VLPXODWLRQ DQG WKH FURVVVHFWLRQDO DUHD A XQLWV P
IRU D VLQJOH PROHFXOH PXVW EH VSHFLILHG
7KH VHFRQG PHWKRG >@ FDQ EH H[SUHVVHG DV
RESULTS
Figure 4 VKRZV D UHSUHVHQWDWLYH SORW RI WKH FKDQJH LQ SRWHQ
WLDO HQHUJ\ ZLWK VWUDLQ EODFN VTXDUHV $V H[SHFWHG HQHUJ\
LQFUHDVHV UDSLGO\ ZLWK VWUDLQ 7KH WRWDO SRWHQWLDO HQHUJ\ LV WKH
UHVXOW RI WKH FRQWULEXWLRQV RI QRQERQGHG YDQ GHU :DDOV
&RXORPE K\GURJHQ ERQGV LQWHUDFWLRQV DQG ERQGHG LQWHUDF
WLRQV LH FRYDOHQW ERQGV )LJXUH VKRZV WKDW WKH HQHUJ\
DVVRFLDWHG ZLWK QRQERQGHG LQWHUDFWLRQV UHG FLUFOHV LV ODUJHU
WKDQ WKDW RI WKH ERQGHG LQWHUDFWLRQV JUHHQ WULDQJOHV DW VPDOO
VWUDLQV OHVV WKDQ a +RZHYHU WKH RSSRVLWH LV WUXH ZKHQ
WKH VWUDLQ LV ODUJH 7KXV WKH WRWDO SRWHQWLDO HQHUJ\ LV GRPL
QDWHG E\ QRQERQGHG LQWHUDFWLRQV DW ORZ VWUDLQ DQG ERQGHG
LQWHUDFWLRQV DW KLJK VWUDLQ 0RGHO FKDLQV RI DOO OHQJWKV DQG
ERXQGDU\ FRQGLWLRQV H[KLELWHG VLPLODU EHKDYLRU
7KH UROH RI K\GURJHQ ERQGLQJ ZDV HYDOXDWHG E\ FRXQWLQJ
WKH QXPEHU RI K\GURJHQ ERQGV SUHVHQW GXULQJ GHIRUPDWLRQ
$ K\GURJHQ ERQG LV LGHQWLILHG ZKHQ WKH K\GURJHQ ERQGLQJ
GLVWDQFH LV OHVV WKDQ RU HTXDO WR QP DQG WKH GRQRUK\GUR
JHQDFFHSWHU DQJOH LV OHVV WKDQ RU HTXDO WR ƒ $V VKRZQ LQ
)LJ K\GURJHQ ERQGLQJ GHFUHDVHV UDSLGO\ ZLWKLQ WKH ILUVW VWUDLQ 7KH K\GURJHQ ERQG LV PRGHOHG E\ D FRPELQDWLRQ RI
YDQ GHU :DDOV DQG &RXORPELF HQHUJLHV DQG VR FRQWULEXWHV WR
WKH QRQERQGHG HQHUJ\ 7KH KLJKHU QRQERQGHG HQHUJ\ DW
VPDOO VWUDLQ SULPDULO\ UHVXOWV IURP WKH H[LVWHQFH RI K\GURJHQ
ERQGV DV VXJJHVWHG IURP WKH SORW $V WKH VWUDLQ LQFUHDVHV WKH
K\GURJHQ ERQGV EUHDN YHU\ TXLFNO\ DQG FRQWULEXWH OHVV WR WKH
UHVSRQVH RI WKH PROHFXOH )RU VPDOO VWUDLQV UHPRYLQJ WKH HI
ZKHUH 8 DQG L DUH REWDLQHG IURP WKH VLPXODWLRQ DQG DQG PXVW
EH VSHFLILHG 7KH HUURU DVVRFLDWHG ZLWK WKLV PHWKRG LV GHSHQ
GHQW RQ WKH DFFXUDF\ RI WKH VSHFLILHG HTXLOLEULXP OHQJWK 7KH
HTXLOLEULXP OHQJWK RI D VLQJOH FKDLQ LV OLNHO\ WR EH JUHDWHU WKDQ
WKDW RI D FU\VWDO 7KHUHIRUH HUURU PLJKW EH LQWURGXFHG LQWR WKH
SURFHVV LI WKHUH LV D GLVFRQQHFWLRQ EHWZHHQ WKH PRGHO DQG WKH
VRXUFH RI LWV HTXLOLEULXP OHQJWK +RZHYHU LI WKH HTXLOLEULXP
OHQJWK LV FDOFXODWHG XVLQJ D ILW WR WKH U DQG L GDWD REWDLQHG
IURP WKH VLPXODWLRQ ZH ILQG WKDW WKLV PHWKRG \LHOGV WKH VDPH
UHVXOW DV WKH ILUVW PHWKRG ZLWKLQ VWDWLVWLFDO VLJQLILFDQFH (Fig.
5).
7KH WKLUG PHWKRG >@ FDQ EH H[SUHVVHG
ZKHUH VWUHVV R XQLWV 3D DQG OHQJWK L DUH REWDLQHG IURP WKH
VLPXODWLRQ L DQG PXVW EH VSHFLILHG $W ILUVW JODQFH WKLV PHWK
RG DSSHDUV WR EH WKH VLPSOHVW DQG VR ZRXOG SRWHQWLDOO\ KDYH
WKH ODVW HUURU +RZHYHU WKH VWUHVV LV FDOFXODWHG IURP WKH VR
FDOOHG SUHVVXUH YLULDO ZKLFK FRQWDLQV HUURU LQ LWVHOI )LUVW WKH
SUHVVXUH YLULDO LV D IXQFWLRQ RI DUHD VR WKDW DSSUR[LPDWHG SD
APRIL 2011| TAPPI JOURNAL
39
MOLECULAR MODELING
5. Comparison of three methods for calculating axial elasticity
in large strain. Data normalized by 122 GPa. Methods 1 (blue)
and 2 (green) yield the same result for molecules with different
lengths and boundary conditions, while method 3 (red) predicts
lower values.
UDPHWHU LV EHLQJ VSHFLILHG LPSOLFLWO\ 6HFRQG WKH YLULDO H[SUHV
VLRQ LV QRW HTXLYDOHQW WR WKH PHFKDQLFDO &DXFK\ VWUHVV >@
$V VKRZQ LQ )LJ ZH ILQG XVLQJ WKLV PHWKRG \LHOGV D[LDO
PRGXOL WKDW GLIIHU IURP WKRVH SUHGLFWHG E\ WKH ILUVW WZR PHWK
RGV E\ XS WR IRU LQILQLWH FKDLQV DQG IRU ILQLWH FKDLQV
:H WKHUHIRUH GR QRW DGYRFDWH XVH RI WKLV PHWKRG DQG ZLOO QRW
GLVFXVV LW IXUWKHU KHUH
7KH D[LDO PRGXOXV ZDV FDOFXODWHG IRU ERWK ILQLWH DQG LQIL
QLWH FKDLQV ZLWK GLIIHUHQW QXPEHUV RI UHSHDW XQLWV IURP GDWD
WDNHQ LQ WKH ODUJH DQG VPDOO VWUDLQ UHJLPHV Figures 6 DQG 7
VKRZ WKH UHVXOWV ZKHUH WKH SORWV RQ WKH WLJKW VKRZ WKH UDZ
GDWD DQG SORWV RQ WKH OHIW SURYLGH D VWDWLVWLFDO UHSUHVHQWDWLRQ
7KH VWDWLVWLFDO SORWV LOOXVWUDWH WKH LQWHUTXDUWLOH UDQJH ODUJH
UHFWDQJOH PHDQ SRLQW LQVLGH WKH UHFWDQJOH PHGLDQ OLQH
LQVLGH WKH UHFWDQJOH DQG PD[LPXP DQG PLQLPXP YDOXHV
KLJKHVW DQG ORZHVW SRLQW UHVSHFWLYHO\ 7KH GDWD DUH QRUPDO
L]HG LQ WKHVH ILJXUHV WR KLJKOLJKW WKH IDFW WKDW WKH WUHQGV DWH
QRW VSHFLILF WR WKH IRUFH ILHOG XVHG
)LJXUH LOOXVWUDWHV WKH UHVXOWV REWDLQHG IURP WKH ODUJH
VWUDLQ ! UHJLPH )RU ERWK ILQLWH DQG LQILQLWH FKDLQV WKH
XQFHUWDLQO\ LQ WKH GDWD GHFUHDVHV DV WKH FKDLQ OHQJWK LQFUHDV
HV )RU H[DPSOH WKH VWDQGDUG GHYLDWLRQ LQ WKH ILQLWH PROHFXOH
GDWD LV ZLWK VL[ UHSHDW XQLWV DQG ZLWK UHSHDW XQLWV
IRU LQILQLWH PROHFXOHV WKH VWDQGDUG GHYLDWLRQ LV ZLWK VL[
UHSHDW XQLWV ZKLOH WKDW ZLWK UHSHDW XQLWV LV FDOFXOD
WLRQV EDVHG RQ QRUPDOL]HG GDWD ,Q DGGLWLRQ WKH GDWD GR QRW
DSSHDU WR WUHQG LQ DQ\ SDUWLFXODU ZD\ IRU WKH LQILQLWH FKDLQV
ZKLOH WKH D[LDO PRGXOXV RI WKH ILQLWH FKDLQV LQFUHDVHV ZLWK
FKDLQ OHQJWK DQG DV\PSWRWHV WR WKH VDPH YDOXH DV WKDW SUH
6. Finite (top) and infinite (bottom) molecules at large strain; axial modulus normalized by 177 GPa (the highest value from
calculation). The plots on the right show the raw data with a solid line indicating convergence, and left ones provide a statistical
representation. Ten independent simulations were run for each case.
40
TAPPI JOURNAL| APRIL 2011
MOLECULAR MODELING
7. Finite (top) and infinite (bottom) molecules at small strain; axial modulus normalized by 802 GPa (the highest value from
calculation). The plots on the right showe the raw data with a solid line indicating convergence and left ones provide a statistical
representation. Ten independent simulations were run for each case.
GLFWHG IRU WKH LQILQLWH FKDLQ
)LJXUH VKRZV WKH UHVXOWV REWDLQHG IURP WKH VPDOO VWUDLQ
UHJLPH $V ZLWK WKH ODUJH VWUDLQ UHJLPH WKH GDWD XQFHU
WDLQO\ GHFUHDVHV ZLWK LQFUHDVLQJ FKDLQ OHQJWK +RZHYHU HYHQ
ZLWK WKH ORQJHVW FKDLQ WKH XQFHUWDLQW\ LV VL[ WLPHV WKDW RE
WDLQHG IURP WKH VDPH PRGHO LQ WKH ODUJH VWUDLQ UHJLPH ,Q DG
GLWLRQ WKH PDJQLWXGH RI WKH D[LDO PRGXOXV LV PXFK ODUJHU LQ
WKH VPDOO VWUDLQ UHJLPH FRQYHUJLQJ WR D YDOXH WZR WR WKUHH
WLPHV WKDW RI ODUJH VWUDLQ LQ ERWK ILQLWH DQG LQILQLWH FDVHV
7KHVH REVHUYDWLRQV FDQ EH XQGHUVWRRG LQ WHUPV RI WKH UROH
RI SRWHQWLDO HQHUJ\ DQG K\GURJHQ ERQGV $V VKRZQ LQ )LJ WKH HQHUJHWLFV RI WKH ORZ VWUDLQ UHJLPH LV GRPLQDWHG E\ QRQ
ERQGHG LQWHUDFWLRQV DVVRFLDWHG ZLWK WKH SUHVHQFH RI K\GURJHQ
ERQGV 7KHVH K\GURJHQ ERQGV DUH H[WUHPHO\ VWURQJ DQG VR
FRQWULEXWH WR D KLJK D[LDO PRGXOXV ,Q WKH KLJK VWUDLQ UHJLPH
ZKHUH PRVW K\GURJHQ ERQGV KDYH EURNHQ WKH HQHUJHWLFV LV
GRPLQDWHG E\ WKH UROH RI ERQGHG LQWHUDFWLRQV DQG ZH REVHUYH
D FRUUHVSRQGLQJO\ ORZHU D[LDO PRGXOXV
&RPSDULVRQ RI WKH D[LDO PRGXOL UHSRUWHG KHUH WR WKRVH LQ
WKH OLWHUDWXUH SRVHV D FKDOOHQJH EHFDXVH DV VKRZQ KHUH UH
VXOWV RI D PROHFXODU VLPXODWLRQ GLIIHU VLJQLILFDQWO\ ZLWK PRGHO
SDUDPHWHUV DQG FDQ EH H[SHFWHG WR YDU\ HYHQ IRU D IL[HG VHW
RI FRQGLWLRQV 7KH D[LDO PRGXOXV SUHGLFWLRQV LQ )LJV ODUJH
VWUDLQ DQG VPDOO VWUDLQ DSSURDFK *SD DQG *3D
UHVSHFWLYHO\ 0ROHFXODU VLPXODWLRQEDVHG SUHGLFWLRQV RI WKH
D[LDO PRGXOXV RI FHOOXORVH ,‰ DUH W\SLFDOO\ LQ WKH UDQJH RI *3D >@ 6R DOWKRXJK WKHUH LV FHUWDLQO\ RYHUODS EHWZHHQ
RXU UHVXOWV DQG WKRVH UHSRUWHG HOVHZKHUH LW LV GLIILFXOW WR
UHDFK GHILQLWLYH FRQFOXVLRQV ,Q WHUPV RI H[SHULPHQWDO UHVXOWV
WKH FRPSDULVRQ LV HYHQ PRUH GLIILFXOW WR PDNH )RU H[DPSOH
WKH D[LDO PRGXOXV IRU FHOOXORVH ,‰ REWDLQHG IURP [UD\ VFDWWHU
LQJ ZDV UHSRUWHG WR EH “ *3D >@ )XUWKHU WKH HODVWLF
LW\ RI D VLQJOH FKDLQ ZLOO OLNHO\ GLIIHU IURP WKDW RI D FU\VWDO EH
FDXVH RI WKH UROH RI LQWHUFKDLQ K\GURJHQ ERQGLQJ )RU
H[DPSOH ZH KDYH IRXQG WKH PHDQ D[LDO HODVWLFLW\ IRU D [[
FHOOXORVH FU\VWDO WR EH *3D ZKHUHDV WKDW RI WKH FRUUH
VSRQGLQJ HLJKW XQLW VLQJOH FKDLQ LV *3D &RPSOLFDWLQJ WKH
LVVXH IXUWKHU LV WKH FRRSHUDWLYH K\GURJHQ ERQGLQJ VXFK WKDW
RPLWWLQJ LQWHUFKDLQ K\GURJHQ ERQGLQJ DV LV QHFHVVDULO\ WKH
FDVH IRU D VLQJOH FKDLQ ZLOO DIIHFW LQWUDFKDLQ K\GURJHQ ERQG
LQJ >@ +RZHYHU ZKLOH WKH UHVXOWV VKRZQ LQ WKLV SDSHU FDQ
QRW EH GLUHFWO\ H[WUDSRODWHG WR FHOOXORVH QDQRFU\VWDOV WKH
WUHQGV REVHUYHG ERWK LQ WKH PDJQLWXGH DQG WKH XQFHUWDLQW\
RI D PRGHOSUHGLFWHG D[LDO PRGXOXV DUH JHQHUDOO\ DSSOLFDEOH
CONCLUSION
7KLV VWXG\ KLJKOLJKWV WKH HIIHFWV RI FKDLQ OHQJWK VWUDLQ UH
JLPH DQG FDOFXODWLRQ PHWKRG RQ D[LDO PRGXOXV SUHGLFWLRQ
XVLQJ DWRPLVWLF VLPXODWLRQ DQG WKH XQFHUWDLQO\ DVVRFLDWHG ZLWK
WKDW SUHGLFWLRQ $OO WKUHH RI WKHVH FRQVLGHUDWLRQV DIIHFW WKH
PRGHO SUHGLFWLRQV DQG QHHG WR EH WDNHQ LQWR FDUHIXO FRQVLGHU
DWLRQ ZKHQ XVLQJ DWRPLVWLF PRGHOLQJ RU LQWHUSUHWLQJ GDWD RE
APRIL 201 1 | TAPPI JOURNAL
41
MOLECULAR MODELING
WDLQHG IURP LW 7KH OLPLWDWLRQ RI WKLV VWXG\ LV WKDW LW ZDV SHU
IRUPHG ZLWK VLQJOH FKDLQV PRGHOHG XVLQJ RQO\ RQH IRUFH ILHOG
&203$66 +RZHYHU D FRPSUHKHQVLYH VWXG\ WKDW LQFOXGHV
FKDUDFWHUL]DWLRQ RI PXOWLSOH IRUFH ILHOGV DQG WKH HIIHFW RI
PRGHO VL]H LQFOXGLQJ FU\VWDO VWUXFWXUH LV XQGHUZD\ 7KHVH
W\SHV RI VWXGLHV DUH FULWLFDOO\ QHFHVVDU\ WR HQDEOLQJ LQWHUSUHWD
WLRQ RI PRGHO GDWD 2QO\ E\ XQGHUVWDQGLQJ KRZ ³ZKDW JRHV
LQ´ DIIHFWV ³:KDW FRPHV RXW´ FDQ DWRPLVWLF VLPXODWLRQV WUDQVL
WLRQ IURP DQ LQYHVWLJDWLYH PHWKRG WR D SUHGLFWLYH WRRO XVHIXO
IRU FHOOXORVHEDVHG DSSOLFDWLRQ GHVLJQ TJ
7.
Nishiyama, Y., Langan, P., and Chanzy, H., J. Am. Chem. Soc.
124(31): 9074(2002)
8.
Mazeau, K. and Heux, L., J. Phys. Chem. B 107(10): 2394(2003).
9.
Bergenstråhle, M., Berglund, L.A., and Mazeau, K., J. Phys. Chem. B
111(30): 9138(2007).
10. Zhou, M., Proc. R. Soc. Lond. A 459(2037): 2347(2003).
11. Diddens, I., Murphy, B., Krisch, M., et al.. Macromolecules 41(24):
9755(2008).
12. Qian, X., Mol. Simul. 34(2): 183(2008).
LITERATURE CITED
1.
Moon, R., Martini, A., Nairn, J.A., et al., Chem. Soc. Rev., in press,
2011, DOI:10.1039/C0CS00108B,.
2.
Tanaka, F. and Iwata, T., Cellulose 13(5): 509(2006).
3.
Eichhorn, S.J. and Davies, G.R., Cellulose 13(3): 291(2006).
4.
Sun, H., Ren, P., and Fried, J.R., Comput Theor. Polym. Sci. 8(1-2):
229(1998).
5.
Miyamoto, H., Umemura, M., Aoyagi, T., et al., Carbohydr. Res.
344(9): 1085(2009).
6.
Plimpton, S., J. Comput Phys. 117(1): 1(1995).
ABOUT THE AUTHORS
With the increasing availability and use of atomistic
simulations to characterize organic materials, we felt
it was important to step back and consider, for the
very simple system of a single cellulose chain, how
parameters that define a model affect its predictions.
Significant work in studying cellulose using atomistic
simulation has been conducted since the mid-1480s.
and the models have progressed in complexity, size,
and accuracy. However, results are typically reported
without considering variations due to small changes
in model parameters or simply because of the statistical nature of the models themselves. Therefore, this
work complements previous studies by providing a
reference point and hopefully setting a precedentfor
future research to report the subtle, yet important,
model details that affect predictions and the uncertainty in those predictions.
The most significant challenge in this study was
isolating individual model parameters to determine
their effects independently. Atomistic models are extremely complex and we could only hope to make
quantitative statements about the roles of model parameters by defining a problem simple enough that
individual effects could be identified, yet complex
enough to approach something experimentally measurable. We chose to focus on the axial modulus of individual cellulose chains with which we could explicitly control characteristics such as chain length and
strain, and where there is complementary (albeit not
directly comparable) experimental data available for
cellulose nanocrystals. We were very surprised to discover just how much variation was inherent in these
42 TAPPl JOURNAL | APRIL 2011
models, even when all input parameters were constant. However, along with the realization that uncertainty exists, came the understanding that we could
quantify it and so take its effect into account when reporting atomistic predictions. The critical analyses of
model predictions in this study are an important first
step, and will form the basis for more comprehensive
characterizations, including multiple force fields and
various crystal sizes and structures. This work will
help support future research in developing new engineered nonwoven materials and composites that are
based on cellulose nanoparticles.
Wu is a graduate research assistant and Martini is
assistant professor with the School of Mechanical
Engineering, Purdue University, West Lafayette, IN,
USA. Moon is a materials research engineer with the
U.S. Forest Service, Forest Products Laboratories, and
adjunct assistant professor with the School of Materials
Engineering, Purdue University, Birck Nanotechnology
Center, West Lafayette, IN, USA. Email Martini at
[email protected].