• Aucun résultat trouvé

Molecular-statistical model for the nematic phase of semiflexible dimers

N/A
N/A
Protected

Academic year: 2021

Partager "Molecular-statistical model for the nematic phase of semiflexible dimers"

Copied!
12
0
0

Texte intégral

(1)

HAL Id: jpa-00247812

https://hal.archives-ouvertes.fr/jpa-00247812

Submitted on 1 Jan 1993

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Molecular-statistical model for the nematic phase of semiflexible dimers

Eugene Terentjev, Charles Rosenblatt, Rolfe Petschek

To cite this version:

Eugene Terentjev, Charles Rosenblatt, Rolfe Petschek. Molecular-statistical model for the ne- matic phase of semiflexible dimers. Journal de Physique II, EDP Sciences, 1993, 3 (1), pp.41-51.

�10.1051/jp2:1993110�. �jpa-00247812�

(2)

J, Pbys. II France 3 (1993) 41-51 JANUARY 1993, PAGE 41

Classification

Physics Abstracts 61.308 36.20

Molecular-statistical model for the nematic phase of semwexible dimers

Eugene

M.

Terentjev,

Charles Rosenblatt and Rolfe G. Petschek

Department of Physics, Case Western Reserve, University Cleveland, OH 44106, Great-Britain

(Received

25 March 1992, accepted jn final form 23

October1992)

Abstract. A microscopic model is developed to describe properties of a nematic phase con- sisting of semiflexible dimers. The effect of the chain bonding the two mesogens is described by

the local conditional probability of their mutual orientation in terms of

a bare stiffness param- eter Q -J

EB/kT,

assuming an even number of carbon atoms in the chain. In the framework of a molecular field approximation we obtain

a complete statistical description of the nematic with expressions for order parameter, mean-field potential, free energy and phase transition

parameters, The width of the N-I transition hysteresis is in agreement with observed values.

Comparison with data on the transition temperatures in three series of nematogenic dimers enables us to obtain the quantitative dependence of the rigidity Q on the length of the

(CH2)n

chain connecting the two monoiners.

1. Introduction.

Thermotropic

main-chain

liquid crystalline polymers

and

oligomers

have been the

objects

of extensive

study

in recent years. In the nematic

phase

there is a

preferred

axis of orientation of mesogens which bound the flexible chain. For rather stiff chains this results in an

exponential

increase of the effective

persistence lengtlI

of the chain in the direction

along

the nematic direc- tor [1, 2] n.

Recently

we

developed

consistent theories of linear

elasticity

and

flexoelectricity

in

long-chain polymers

[2, 3],

taking

iiIto account all relevant interactions between monomers and

expressing corI~esponding

material constants in terms of definite molecular characteristics and the spacer

I.igidity

parameter Q.

This latter parameter is

especially important

to

investigate

when

dealing

with chains of nematic mesogens, because it deteI.i~iines the

specific regimes

of the chain behavior. When Q <

I,

one arrives at the case of

freely joiiIted

monomers, or even

completely free,

at Q

= 0. As

Q - cxJ the spacer

approaclIes

tlIe liiuit of a

rigid

rod. The intermediate

region

of Q > I defines

the most common case of seiniflexible

bonding,

when the

tendency

of

subsequent

monomers

to

align along

the same diI.ection competes with the nematic mean

field,

which acts on each

monomer

separately.

This

gives

rise to

complicated dependences

of

macroscopic properties

on

the parameters of interniononier iiIteractions and spacer

rigidity.

A very favorable

object

to

in,,estigate

these

properties

is a dimer

composed

of two nematc-

genic

monomers, where the influence of the spacer is not obscured

by

the entropy effects im-

(3)

posed by

the

long

chain [2]. The purpose of this paper is to

apply

a

general

statistical

approach

to the case of a

thermotropic

nematic

liquid crystal composed

of these

dimers,

to obtain ob- servable

macroscopic

cliaracteristics of the system

and, by comparison

with

experimental data,

to make certain conclusions about the nature of the mesogens

bonding

and estimates of the value of the

rigidity

parameter Q.

Before

proceeding,

we note that over the years countless observations of sc-called "odd- even" effects have been made

(see,

for

example

[4,

5]).

The reason for the observed behavior is

clear,

since spacers with an odd iiuniber of

i~ietlIylene (or other)

groups

impose

a considerable

equilibrium

bend between

subsequent

mesogens,

making

the

corresponding

dimer

effectively

biaxial.

Recently

Ileaton and Luckhurst [6] and

Photinos,

Samulski arid Toriumi [7]

developed

a detailed model for seI~iiflexible nematic

dimers,

based on the isomeric mechanism of spacer

flexibility.

This iuodel is based on the

assumption

of fixed tetrahedral

configuration

of bonds of each carbon atom, overall

flexibility

is due to torsional rotations of C C bonds. As a

result the statistical

properties

of the neiuatic

phase

are determined

by population

of

gauche

bonds on the spaceI. chaiiI. TlIis

model, seemingly

very

realistic,

has certain

disadvantages.

First,

the cost of such a detailed

description

of every bond

configuration

and orientation is that authors have to make

severe

approximations

in order to obtain any

practical

result. It

does not look

plausible

tlIat this model caiI be

successfully applied

for the

description

of

longer oligomers

and

polymers.

Second, tlIe isomeric mechanism of

flexibility completely

overlooks the

possibility

of a continuous

bending

of the

chain,

which is distributed between each C C bond in

equal portions.

A

straightforward

calculation with the

help

of the molecular

modelling

software

package "Sybyl"

shows that the energy of one

gauche

bond (~- 0.02

eV)

is sufficient for continuous

bending

of the all-trails

(ClI2)Io

chain for the

angle

~- 15°. Creation of the

pair

of

gancfie

bonds

requires

the same energy as the continuous

bending

of this chain for 24°.

Therefore we argue

that,

in

spite

of the

quite

detailed and

specific description,

the isomeric model of the spacer chain

flexibility

cannot be considered as a final realistic

approach, obviously

a combination of discrete torsional rotations and continuous bond deformations takes

place

in

a real system.

For these reasons we addressed the

problem

of the nematic

liquid crystalline phase

of semi- flexible dimers

applying

the

coIuputationally siiuple

continuous model of the spacer

flexibility,

which does not assume any

specific

atomic iuechanism and uses an

integrated phenomenologi-

cal

rigidity

paI~aiueter Q. In the present work we do not take into account the dimer

biaxiality and, therefore,

our results are

quantitatively applicable only

to the case of an even-numbered

spacer when the two

Iuesogenic

iuoiioiuers are

parallel

in

equilibrium.

Let us

emphasize

that the

configuration

of the spacer chain

connecting

the two

mesogenic

units is

strongly

influenced

by

these

interacting

multi-atomic aggregates, and differs from the

configuration

of the free

chain,

for

example

a teriuinal group. It seems

highly improbable

to find an odd number of

gancfie

conforinat,ions in spacers of

niesogeiiic

diniers in a nematic

phase.

In the

isotropic phase gauche

excitations are niore easy to appear, still their energy must be much

higher

than that for the free

(CII2)n

chain. We present below the

expression

that shows how the distribution of the local conditional

probability

of the two monomers orientation narrows due to the effect of the nematic mean field. The central result of this paper is the

dependence

of the bare spacer

rigidity

Q on the

length

of the

(CII2),1 chain,

obtained

by

the

analysis

of

experimental

data for dimers with even-numbered spacers. This

dependence,

Q

~-

I/n,

indicates that the continuous meclianisiu of

flexibility

niay

prevail

even in the

isotropic phase.

We will return to the discussion of this

question

below: in the

remaining

of this work we assume that the value of Q is the saute for all diiners and expect that this

simplification

does not

change qualitative

predictions.

(4)

N°i NEMATIC OF SEAIIFLEXIBLE DIAIERS 43

n ~,

u u'

u~ ,

.. 8

~ U

~~ / fi'

Fig-I-

Mutual orientation of rod-like mesogens, bounded by a semiflexible chain, in a nematic director field.

2. Mean field

theory.

We consider a dimer as a sequence of t,vo rod-like mesogens of

length

I and width

d, figure

16.

In order to be able to compare our results with observations for

corresponding

monomers we

assume that the

length

of these is

exactly

I, which includes the hard core and terminal groups;

the

length

of the dimer in

equilibriuiu

is 21, so that the spacer

actually

consists of two linked terminal

(CiI2)n/2

groups. The effective

flexibility

of such spacer

depends

on the total number

n of

CH2

units between two hard cores, but we do not consider this spacer as a separate steric

object, assuming

that its two parts are

already

included in the

shapes

of the two monomers.

Therefore the effect of the spacer is

purely

orientational in our model. It is necessary to

give

a consistent

sign

to the orientation u of each of these

rods,

because the

probable positions

of

subsequent

monomers

depend

on this

sign.

We define

P~(r r',u,u')

to be the

probability

that for an isolated diiuer the mesogen with its center of symmetry

position

r and orientation u is followed

by

another mesogen, which has the

position

r' and orientation u'. We assume that this bare bond

probability

P~ does not

depend

on parameters of other mesogens, other than the

linearly adjacent

ones. ~fe do not consider lIere the limit of

freely jointed

mesogens; it has

been shown

(see ill,

for

example)

that its

properties

are not very much different from those for

free monomers. III the

opposite liiuit,

wlIen the mesogen orientation is

strongly

determined

by

the

adjacent

mesogen on the

chain,

a reasonable

approximation

for the

bond-probability

P~ is

~~

~~~~~'~ ~

8il~

Q ~~~

~~~~

~'~~~ ~~~

where the

coupling

Q is determined

by

an effective energy of

bending

of the spaceri Q

~-

EB/kT

> I.

Equation (I)

iueans tlIat if a

long polymer

chaili of these monomers was consid-

ered,

its

isotI.opic peI.sisteiIce length (tlIe

sc-called I(uhn

segment)

would be

equal

to al.

In the nematic

phase

the oI.ientational

probability

distribution function

P(u)

for each mesc- gen is determined

by

tlIe statistical

weiglIt

of the

remaining

part of the dimer and

by

the

interaction with the iuean field in the media.

Ileducing

the

general expression

[2] to the case of

dimers,

one obtains

P(u,ii)

=

je~~l~"

~~~l/~~

/ P~(u, u')e~~l~"'~~l/~~du' (2)

(5)

where Z is the normalization

factor, U((u .n))

is the mean-field

potential,

and n b the nematic director.

Generally,

this

potential U(u, n)

has an

axially syuinietric

form on the unit

sphere

with the two

deep

wells at the

poles

(~i,

-u)

and a barrier of

height

J at the equator

(uLn).

The

simplest possible

form of U is U = Uo

J(u n)~,

an

approximation

which works

fairly

well if the

coupling

constant J is

sufficiently large.

The

analytical expression

for

J,

which

is

proportional

to the nematic order parameter, will be derived below. We note here that since the N-I transition is first

order,

there is no

region

of nematic

corresponding

to small

J/kT (actually,

in many cases

J/kT

may be considered as a

large parameter). Therefore,

the

expansion

of

expressions

like that in

equation (2)

in powers of

Legendre polynomials (like

it is done in [6], for

example) is, strictly speaking, incorrect,

and one must

always

retain the full

exponential

form.

Three

pairs

of unit vectoI.s are involved in the

equation (2):

u,

u';

u, n and

u',

n

(see Fig. I).

Substituting

the

expI.essions

for P~ and U and

using

the

relationship

cos7

=

cosbi

cosb2 + sin

bI

sin b2 cos ~, one can

diI.ectly integrate equation (2)

over the azimuthal

angle

~ to obtain

p~

~& cos2 @1

j

~

g~cos2 @2~n cos @i cos @2zo

~~

~i~ ~1~i~ ~2 ~i~ ~2d~2 ~3j Z

where lo is the Infeld function of zeI.o oI.der.

Since the mean-field

coupling

constant

Ji together

with

Q, plays

the most

important

role in the present

iuicroscopic theory,

it is desirable to derive an

analytic expression

for J

using

interaction parameteI.s in the system. One can do this

by

minimization of the free energy over the mean field

potential which, strictly speaking,

appears as an unknown function in

equation (2).

In the molecular field

approximation

for

probability

distributions the free energy of a

system of flexible diiuers contains two contributions: from the

single

molecule internal free energy and from the interaction between iuesogens which are located on the different dimers.

The influence of tlIe link between bound mesogens lIas been accounted for with the

help

of the bond

probability equation (I).

In

addition, anisotropic

mesogens interact nia an effective

pair potential

[8] which contains attractive and hard-core

repulsive

parts. The

repulsive

part contributes both to an

oI.ientatioiI-dependent

cut-off for the effective attraction and to the

packing (translational)

entI.opy of the system

(see Eq. (6) below).

In this work we assume,

following StI.aley

[9] the simplest model for the

repulsive potential

which reflects the symmetry

requiremeiIts:

the contact

step-function B(ri, uilr2,u2),

which is

equal

to

unity

outside the

restricted

regions

where the hard cores of the t,vo iuesogens would otherwise penetrate into each other. These

regions

are deteI.ruined

by

the condition that the distance between two

centers of symiuetI.y rI r? is less tlIaiI the

anisotropic

form-factor

i~~

~s d +

j

~

i(u~ i~~)2

+

(u~

r~~121

(4)

where the unit vector I.12 "

ri?/r12.

A reasonable

approximation

for the attractive part of the

pair potential

wlIiclI accounts for botlI

isotropic

and

anisotropic

contributions is

U~~~'(1,

2)

=

( [Io(uI u2)~ +12(ui u2)(ui FI2)(u2 FI2)] (5)

l'I2 l'I2

Note that in iuost known cases the

isotropic

Van der lvaals attraction is much

stronger

than

anisotropic

one, which is determined

by anisotropies

of molecular

polarizabilities, viz.,

G

~-

(5 7)

x

lo,?

'-

(1- 5)

x 10~~~

eI.gcni~.

Following

[8] ,ve define the effective

paiI. potential

for interaction of monomers "I" and "2"

as the sum of attractive part and the

packing

entropy term, which both have the

twc-particle

(6)

N°i NEMATIC OF SEAIIFLEXIBLE DIR,IERS 45

microscopic

form

(see

also

[9]):

U~~'(1, 2)

= U~~~'(1,

2)E(r12i

ui,

u2)

+

~~

(l E(r12i

ui,

u2)] (6)

nap

Here the second term differs from zero

only

in

regions

which are restricted for the mesogens

"I" and "2" due to their steric

repulsion.

The factor

(I nap),

where no is the mesogen

volume,

accounts for the correction to the

Onsager approximation

for the dense

packing,

which was discussed

by

Gelbart and Baron

[10].

The

twc-particle

interaction part of the internal energy takes the form

F~~t

=)p( £ / l§(ui,u(ri))Pk(u2,n(r2))U~~'(1,2)dridr2duidu2 (7)

j,k=1,2

where pd "

p/2

is the

density

of

dimers,

and the

probability

distributions Pk are determined

by equation (2).

The summation in

equation (7)

is carried out over mesogens located on different

molecules. The interaction of

adjacent

iuesogens on the same chain is accounted for

by

the bond

probability

P~ and contributes to the

local, single-chain

free energy. This local free energy of a clIain in a iuean field is the excess of the

logarithm

of the dimer

partition

function

(-kT

In

Z)

over the

backgI.ound

average of the mean field U in the system; it consists of the orientational eiItropy of a semiflexible dimer and the internal energy and entropy associated with its

bending:

Fjo~ = NAT In Z N

£ f U(u, ~i)Pk (u, n)du (8)

~ ~

k=1,2

The partition function of a semiflexible diiner is

Z =

f e~~~~>~~)/"~e~(~ "'~e~~(""")/~~du'du (9)

and N

=

pV

is the total number of iuesogens in the system. This

expression

reduces to the usual orientational entropy contribution Fjo~ =

pkT J

P In Pdudr in both

limiting

cases Q

- 0

and Q

- cxJ. These

correspond respectively

to

freely jointed

or unbounded mesogens, and to

rigidly

bounded mesogens wlIiclI foriu a

rigid

rod. Note that an increase of the nematic mean field causes an effective incI.ease of tlIe bare

rigidity

so that the actual

(observed)

width of the

probability

distribution of tlIe I.elative monomers orientation is: 11 cS Q +

J/kT.

It has been shown [2] tlIat the variation of the total free energy of the system Tint, + Fioc

over the mean field

potential gives

the

expression

for U,

exactly corresponding

to that for the

free monoIuer systeni:

U(iI,

ii cS pd

=1,2~ ~ (u')U~~'(u, u'; r12)dr12du' (10)

This

justifies

the use of tlIe

equation (2)

for the mesogen orientational

probability. Again using

the initial

assumption

U m Uo-

J(u.u)~,

one can

perform

both the radial

(r12)

and

angular (u) integrations

in

equation (10) by

iueans of successive saddle

point approximations

at

J/kT

> I

on each

non-analytical

step. This transforiiis the

integral equation (10)

into an

algebraic

one.

Then, in order to deterniine the mean-field

coupling

constant

J,

it is convenient to consider small

angles

91 of deviation of u from ii

(such

a liiuitation

corresponds

to the

approximation

(7)

J/kT

> I

above).

TlIen u,e are able to consider the mean field

coupling

constant to

obey

the relation J

=

)3~U/3b][~~=o.

~Ve thus obtain

J cS

)p(~)/~G

+

~(12

+

)Io)

+ ~~~

Tj

S +

)HS (11)

Yap

where the nematic order paraiueter S =

-0.5+1.5(cos~ bi)

contains the additional

dependence

on Q and T. We estimate tlIe "baI.e" Iuean-field

coupling

constant H

using typical

values for the molecular parameters.

Taking

~- 30

I,

d

~- 5

I,

p

~- 10~~

cm~~,

temperature kT ~- 4 x 10~~~ erg

and, approxiIuately,

lo " 12 =

G/5;

G

~- 3 x

10~~~ergcm~ (clarify

with the Ref. [10], for

exaniple).

The three contributions to

equation (11) give, respectively,

17 x

10~~~,

5 x 10~~~ and 10~~~ erg,

so that the total

coupling

constant H

~- 25 x kT.

If the

anisotropic

attractioiI betiveeii

mesogenic

molecules is much weaker

(Io,2

<

G),

ther-

motropic

nematic

ordeI.ing

is deterniined

priniarily by

an

isotropic

attraction modulated

by

an asymmetry of molecular

shape

(12. The first term then dominates

equation (11)

and one

can expect muclI lower values for H

~-

(8 10)

x

kT, corresponding

to a lower transition temperature, as ,ve will see belo,v.

The molecular-statistical

description

of the nematic

phase composed

of semiflexible dimers with the

only

model parameter

being

Q iuay now be considered

complete.

There are several

simple

ways to obtain the

niacroscopic,

observable parameters of the system. One

method, analogous

to that of

lklaier-Saupe,

is to describe the

branching

of the nonzero solution for the order parameter

S,

I-e-

by looking

for tlIe nontrivial solution of the

equation

S

= -0.5 +

1.5(cos~ bi)

#

1.5J/H.

This

gives

the ultimate

superheating

temperature T* of the transition.

It is also

helpful

to use the fact that the N-I transition is

weakly

first

order,

and to derive the

microscopic expressions

for the coefficients of the Landau free energy

expansion

F m

aS~+bS~+cS~+..

as functions of the iuolecular parameters. One can then obtain the

complete description

of the

phase

transition and the

low-temperature

nematic

phase.

In

particular,

the

instability

threshold of the

isotropic phase (ultimate supercooling

temperature

TI)

and the width of the first-order transition

hysteresis

T*

TI,

which both are

easily

observable

quantities,

can be deteriniiIed iiI this model.

3. Phase transition

properties.

Using equations (2)

and

((7)-(9)),

we find the

following

relations:

j

~ 2~

f

~J(1111)~/kT~fl(11 U')~J(11'.1l)~/kT

j~ ~j2

~~i ~~

fi~Il ~

°~ ~ Z

32J/kT j~~j

The Landau coefficients are determined

by isotropic

averages at J

= 0:

a(Q,

H,

ET)

=

pH

I

$

[((cos~

bI + cos~

b2)~)o (cos~ bi

+ cos~

b2)(])(13)

~~~' ~' ~~~ II ~~ ~~

~~~~°~~~~ ~ ~°~~~~~~~° ~~~~

-3((cos~

91 + cos~

b?)~)o(cos~

bi + cos~

b2)o

+

2(cos~

bi + cos~

b2)(]

(8)

N°i NEMATIC OF SEAIIFLEXIBLE DIAIERS 47

~~~'~'~~~ 2~3~~ ~~~

~~~~°~~~~ ~ ~°~~~~~~~° ~~~~

-4((cos~

bi +

cos~ b2)~)o(cos~

i

+

cos~ b2)o 3((cos~

bi +

cos~ b2)~)(

+

+12((cos~

bI + cos~

b2)~)o(cos~ bi

+

cos~ b2)( 6(cos~ bi

+

cos~ b2)(]

where

((cos~ bi

+ cos~

b2)~)o

+

) f e~

~~ ~i ~"

~?Zo(Q

sin

bi

sin

b2) (16)

o

(cos~

bi + cos~

b2)~'sin bi

sin

b2d81db2 and, clearly, (cos~ bi

+ cos~ b~)o +

2/3.

ExamiIIing equations ((12)-(16))

one can

easily

notice that in the limit Q <

I,

the coefficients

approach

the monomer form with a

=

ao(T T~'°~°.)

cS

~p$[kT 0.0445H];

b cS -8.3 x

10~~pH(H/kT)~

and c cS 5.6 x

10~~pH(H/kT)~.

These 9values

correspond quite

well to the

predictions

of a

Maier-Saupe-like theory: recalling

the value of H in

equation (I I)

and

assuming

that the

isotropic-nematic

transition temperature is of the same order of

magnitude

as the ultimate

supercooling

temperature Ti in a, we find

kTf°~°.

~-

kTflJ~°.

~- 5 x

10~~~

ergi which is of order 90 °C. In the

opposite limiti

Q - cxJ, the

integrands

in

equation (17)

take the form

exp(Q cos(bi b2))

and the Landau coefficients become a cS

(p$[kT 0.088H];

m

-3.2 x

10~~pH(H/kT)~

and c cS 4.3 x

10~~pH(H/kT)~.

These values for the monomers with doubled

length

exhibit

a shift of the transition temperature and a more

strongly

first order

transition with

corresponding

increases of the order parameter

discontinuity

and

enthalpy.

In order to describe the

crossover behavior of the system in the most

interesting region (Q

>

I),

one must calculate the

integI.als

in

equations ((12)-(16)) numerically.

From the theoretical

point

of

view,

there are three characteristic temperatures

describing

the N-I first order

plIase

traiIsition the ultimate

superheating

T* and ultimate

supercooling Ti points,

and the

equilibrium

transition

point

T~ at which the free energy of the

non-symmetric phase

becomes lower tlIan that of the

isotropic phase.

The ultimate

superheating

temperature T* at which the bifurcation of tlIe nonzero order parameter S takes

place,

can be determined very

precisely

in our statistical model. All other characteristic

points

on the temperature scale

can be referenced to T*. The difference between T* and the ultimate

supercooling temperature

TI

(the

ultimate widtlI of the weak first order transition

hysteresis)

is T* -TI =

9b~/32uoc.

The

actual transition temperature Tc lies between TI and T* T* T~ cS 0.022b~

lace

« T* TI for Q = 0 "short monomer"

case) kTm°~°

cS 0.04995H. In the

figure

2 the upper curve represents

the behavior of

Tc(Q),

which is normalized

by Tm°~°.

The lower curve on this

plot

shows the relative variation of the ultimate

supercooling

temperature TI.

One can

clearly

see the

sharp change

in both characteristic temperatures in the

region

of

Q between 2 and 5, as well as their saturation when the bond between the two mesogens

becomes very

rigid.

Since both Tc and Ti are

easily

measurable

quantities,

we show in

figure

3 the

dependence

of their ratio on the bond

rigidity

Q.

Again,

it is clear that the

region

of intermediate Q has the most dramatic influence on the

thermodynamic properties

of the

system

of semiflexible dimers.

Returning

to the definition of the bare bond

rigidity

Q in the

equation (I),

we note as an aside that the value of Q cS 3

corresponds

to the characteristic

angle

7

~- 45°

of thermal fluctuations of the dimer iii an

isotropic

inert solvent

(see Fig. I).

Recently

several series of dimers with two identical

nematogenic

monomers and spacers of

varying lengths

[12]

(CH2),1

have been

synthesized.

These

molecules,

with an even number of

(9)

2

1.8

u1

§

l.6

/

#

~

,

u1 lA

~

~

~ l.2

~1 O

fj

~

~

~~0

5 lo 15 20 25

Q

Fig.2.

Reduced transition (n /T~~°~°' solid curve) and supercooling

(Ti

/T~~°~° dashed

curve)

temperatures ns. Q

0.965

0.964

0.963

~

0.962

fl

b~ 0.961

0.96

o.959

~'~~~0

5 lo 15 20 25 30

Q

Fig.3.

The ratio Ti/T~ vs. Q

methylene units,

are

likely

to be

good

candidates which can be

reasonably

well described

by

our model. Consider, for

example,

the molecule

7l-~-N=HC-#-O-(CII?)n-O-~-CH=N-~-7l (17)

where

~

denotes a benzene

ring

and terniinal groups 7l are

CH3, C2H5,

and

C3H7. Figure

4 shows the N-I transition temperatures for even

homologues

of these series

[13].

One of the

predictions

of our model is the univeI.sal curve for the

dependence

on Q of the temperatures

Tc(Q),

cf.

figure

2.

Using

tlIis curve and the

experimental

transition temperatures it is

possible

to obtain values for Q as a funct.ion of a spacer

length.

In order to refer these temperatures

to the theoretical curve

Tc(Q),

it is convenient to normalize them

by

the

extrapolated

value of Tc for n

=

0, which, presumably, corresponds

to very

large

Q and therefore to the saturation

(10)

N°i NEAIATIC OF SEAIIFLEXIBLE DIAIERS 49

540

520 *

A

soo "

m

A

480

~

~46 +

m

440

m

A .

420

400

O 2

n

(11)

region

in

figure

2. Linear

extrapolation gives

the

following

values for

Tc(n

=

0):

for

CH3

terminal group 537 + 8

Ii;

for

C2H5

520 + 8 K and for

C3H7

529 + 8 K.

Referring

to these temperatures as the saturation values

on the curve

Tc(Q) (Fig. 2)

and

normalizing

the transition temperatures for n =

214, 6, 8,

10 and 12 to

Tc(n

=

0),

we obtain the

corresponding

values for

Q(n)

for all three materials under consideration. It is

important

to note that there is an error of

~- 1.5 $l in the calculation of reduced temperatures. This error propagates to

Q(n)

and

rapidly

increases when

approaching

the flatter

(close

to

saturation) region

of the

Tc(Q)

curve.

Figure

5 represents the variation of Q with inverse spacer

length

I

In,

and the solid line is a linear least squares fit to the function I

IQ

= An+B. For these three series at

least,

which differ

only by

their terminal groups

7l,

it is clear that this

dependence

is

universal: Q cS

(70 +10) In;

within our accuracy the coefficient B is not

distinguishable

from

zero.

4. Discussion.

The

preceeding result,

Q

~-

70/n,

can be

explained by

a

simple

argument which assumes

that each

[CH2 CH~]

unit in the even- numbered bond has an

equilibrium angle

0° and

an effective

bending

energy Ea. The

probability

of its

bending

to some

angle

7 is then p ~-

exp[(Eo/kTcos7].

If we

neglect

both correlations between

subsequent

units and the influence of terminal monomers, the total

probability

of

bending

of the all-trans

configuration

of the

spacer chain

[CH2 CII2]n/2 by

the overall

angle

7 must be P~

~-

exp[(2Eo/nkT)cos7]

=

exp[Q(u u')].

The factor 2 is needed to account for the fact that each unit contains two

methylene

groups. If we consider the

i-O-j bonds, connecting

the spacer with

mesogenic

cores as flexible units as well witlI the

corresponding bending

energy w (w >

Eo),

the

resulting

expression

for Q must be corI.ected in the

following

way:

~

nw +

4Eo'

Q 2Eo ~ w ~

~~~~~°~

~~~~

If there are in

double-gauche (g-t-g) conformations,

located somewhere on the spacer

chain,

similar arguments about non- correlated sequence of elastic units will lead to an

equation Q(n),

similar to

equation (18)

,vith two types of units with

bending energies Eo

and E21

1/Q

~-

in (6 2E2/Eo)m].

This is not the kind of

dependence

that one observes in the

fig-

ure 5.

The

proportionality represented by equation (18)

has been tested

by

direct quantum chem-

istry

calculations [14] of

bending energies

of various conformations of a free

[CH2]n

chain and

averaging

over the rotation around the cliain axis. The

dependence Din),

that one observes in the

figure 5, corresponds

to the

equation (18)

with very little deviation from

(I In) dependence.

This suggests, that in the materials [12] the concentration of double-

gauche

conformations was very low even at the transition

point.

There is a reasonable

quantitative agreement

between the

figure 5, equation (18)

and the value of the

bending

energy

Eo,

calculated with the

help

of molecular

modeling

software

Sybyl. Indeed,

if we take the molecule

OH-CH2-CH2-OH

as

representing

the unit

[-CH2 CH2-]

in the spacer and find the coefficient for the

potential

energy for

bending

U cS const Eo cos 7, this calculation

gives

Eo cS 2.2 + 0.I

eV, [14, 15].

We have obtained a similar value for the

bending

constant Eo

by

numerical calculation of energy of different

[CH2]n

chains as a function of the

angle

between the terminal C H bonds. For

temperatures of order 520 II tlIis

corresponds

to a coefficient I

IA

=

2Eo/kT

~-

100,

very close

to the

expected

mean value

I/A

= 70 from the

plot I/Q(n), figure

5; as it was

expected, 2/w

< I.

This remaI.kable

agreement

is not

IIecessarily

universal.

llepresentation

of the spacer as

a

non-correlated sequence of

symmetrical

elastic units is

obviously

too gross an

approximation

to

(12)

N°1 NEAIATIC OF SEAIIFLEXIBLE DIAIERS 51

be considered

appropriate

for all niaterials. For

example,

one can expect

significant

deviations from the linear

dependence

Q

~-

I/n

for dimers which possess

dipole

moments in the terminal groups close to the spacer. However the theoretical

predictions

of this

work,

such as

figures 2,3,

are

universali

with the parameter Q

effectively describing

the spacer as a whole. The

three series which have been used for

comparison

with our

predictions

possess an

relatively

flexible connection between the

bonding

chain and hard

mesogenic

cores

through

the ether group, therefore it

might

be

relatively

easy to rotate these cores around the

long

molecular axis. Moreover, there is an absence in the terminal mesogens of

large

transverse electric or steric

dipoles.

It is for these reasons that such a remarkable

"universality"

in

Q(n)

is found for these series of dimers.

At this

juncture

we have

found,

for three series of

dimers,

an excellent

agreement experiment

and theoretical

predictions

for spacer

energies. Nevertheless,

the

theory

makes other

predic-

tions as

well,

such as

Ti/Tc.

Tests of these results

require

further

experimentation,

which are

currently underway. Finally, although

the niodel does not

exhaustively

account for all

possible

intramolecular

interactioiIs,

it does consider tlIe

bending

of a spacer group in a self-consistent way. It thus makes an

important

contI.ibution to our

understanding

of the role of the

spacer(s)

in

polymeric

and

oligomeI.ic liquid crystals.

Ack~iowledgiiieiits.

We are

grateful

to A.C.GI.iffin for

providing

his

experimental

data for N-I transition temper- atures and to I<.NatlI foI~ useful discussions. This work was

supported through

the Advanced

Liquid Crystal Optical

MateI.ials Science and

Technology

Center

by

the State of Ohio

Depart-

ment of

Development

and board of regents and the National Science Foundation

through

grant

No.: DMR-8920147.

References

ill

Khoklilov A-R- and Semenov A-N-, Sov. Phys. Uspekhi156

(1988)

427.

[2] Petschek R-G- and Terentjev E-M-, Phys. Rev. A45

(1992)

930.

[3] Petschek R-G- and Terentjev E-M-, Phys. Rev. A45

(1992)

5775.

[4] Griffin A-C-, SullivaiI S-L- and Hughes W-E-, Liq. Gryst. 4

(1989)

667.

[5] Emsley J-W-, Luckhurst G-R-, Shilston G-S- and Sage I., Mol. Gryst. Liq. Gryst. 102

(1984)

223.

[6] Heaton N-J- and Luckhurst G-R-, fifol. Phys. 66

(1989)

65.

[7] Photinos D-J-, Samulski E-T- and Toriumi il., J. Ghem. Phys. 94

(1990) 2758;(1990)4688.

[8] Gelbart W-AI- and Beii-Shaul A., J. Ghem. Phys 77

(1982)

916.

[9] Straley J-P-, Phys. Rev. A8

(1973)

2181.

[10] Gelbart W-M- and BaI.on B. J. Ghein. Phys. GG

(1977)

207.

[11] Baran J. and Les A., Jfol. Gryst. Liq. Gryst. 54

(1979)

273.

[12] Griffin A-C- and Samulski E-T-, J. Am. Ghem. Sac. 107

(1985)

2975;

Griffin A-C- and Britt T., J. Am. Ghem. Sac. 103

(1981)

4957;

Blumstein R., Poliks M., Stickles E., Blumstein A. and Volino F., Mol. Gryst. Liq. Gryst. 129

(1985)

375.

[13] Hung R-S-L- and Griffin A-C-, private communication.

[14] Nath Il., private commuuicatioiI.

[15] Clark M., Cramer R-D- and van Opdeuboscli N., J. Camp. Ghem.10

(1989)

982.

Références

Documents relatifs

This chapter studies three dimensions of the biologists’ influence: 1) their role in the creation of new values which have come to underpin the new destination; 2) their role as

In this paper, a numerical model has been developed using the commercially-available ABAQUS software to predict the large displacement behaviour of a lightly

This sudden increase in inversion and internal rotation forces, combined with either dorsi- or plantarflexion, produces sufficient strains to rupture the ankle lateral

— Ultrasonic measurements in CBOOA near the smectic A-nematic transition show the occurrence of critical absorption and therefore support the second order nature of the

Conversely, catecholamine requirement and transfusion were less frequent in patients on conscious sedation, and duration of intensive care unit and Abbreviations: BAV, balloon

Caloric curves (classical temperature derived from experimental proton data versus thermal excitation energy) constrained at average volumes (left) and for selected ranges of

monomers, we obtain a complete statistical description of the nematic phase with expressions for order parameter, mean-field potential, phase transition parameters and Frank