• Nenhum resultado encontrado

X-ray structural study of the mesophases of some cone-shaped molecules

N/A
N/A
Protected

Academic year: 2024

Share "X-ray structural study of the mesophases of some cone-shaped molecules"

Copied!
8
0
0

Texto

(1)

HAL Id: jpa-00210212

https://hal.archives-ouvertes.fr/jpa-00210212

Submitted on 1 Jan 1986

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.

X-ray structural study of the mesophases of some cone-shaped molecules

A.M. Levelut, J. Malthête, A. Collet

To cite this version:

A.M. Levelut, J. Malthête, A. Collet. X-ray structural study of the mesophases of some cone-shaped

molecules. Journal de Physique, 1986, 47 (2), pp.351-357. �10.1051/jphys:01986004702035100�. �jpa-

00210212�

(2)

X-ray structural study of the mesophases of

some

cone-shaped molecules

A. M. Levelut

(*),

J. Malthête

(+)

and A. Collet

(+)

(*)

Laboratoire de

Physique

des Solides

(**),

Université Paris-Sud, 91405 Orsay Cedex, France

(+)

Laboratoire de Chimie des Interactions Moléculaires,

Collège

de France, 11,

place

Marcelin Berthelot, 75005 Paris Cedex, France

(Reçu

le

10juillet

1985, accepté

sous forme définitive

le 14 octobre 1985)

Résumé. 2014 Grâce à la diffraction des rayons X, nous avons montré que des molécules

coniques

peuvent former des

mésophases

assez ordonnées. Parmi les quatre

mésophases

étudiées, trois sont

classiques puisque

les colonnes

forment un réseau

régulier plan,

tandis que la

quatrième

est

quasi

cristalline : seules les chaînes

paraffiniques participent

à l’ordre tridimensionnel bien

qu’elles

soient à l’état fondu.

Abstract. 2014 We have established

by X-ray

diffraction that

cone-shaped

molecules can form

relatively

well ordered

columnar

mesophases.

Four

mesophases

have been studied. Three of them have the classical two-dimensional

organization

in

parallel

columns while the fourth one shows a

quasi-crystalline

structure :

3D-ordering

takes

place

at the level of the

long paraffinic

chains which surround the conical core

despite

the fact that the chains are

in a « molten state ».

Classification

Physics

Abstracts

61.30E

1. Introduction.

Since the 1977

discovery [1]

of the first

thermotropic mesophase

of disc-like

molecules,

it has been shown that molecules with flat

rigid

cores surrounded

by

six

or

eight elongated

flexible chains can form

mesophases

in which the molecules are stacked in columns

[2].

The columns are

parallel

and form a

periodic

2D

lattice. The

paraffinic

medium which surrounds each column is in a molten state and therefore ensures

that there is no correlation across the

columns,

of the molecular

positions along

the column axis. It has been shown

recently

that

mesophases

can also be obtained

when the flat core is

replaced by

a conical one

[3, 4].

We

report

here a structural

analysis

of some of these

new

mesophases,

and we compare their

symmetries

with those of the usual columnar

mesophases

of disc-

like

compounds.

2.

Experiments.

We have studied the

mesophases

of the three hexa- esters of

cyclotricatechylene

la-c. The transition

temperatures

are

reported

in table I.

Under

optical microscope

the textures seem cha-

racteristic of columnar

mesophases [3, 4],

but the

viscosities are very

high

and the

mesophases

have a

Table I. - Transition temperatures

of

the three studied

cyclotricatechylene

hexaesters

(OC).

waxy appearance. It was

possible

to obtain

regions

with a

homogeneous

direction of extinction between crossed

polarizers

when the

mesophase

was

spread

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphys:01986004702035100

(3)

352

out on a

glass plate

with a

spatula.

We therefore used

this

technique

for the obtention of

samples

for

X-ray

diffraction

experiments.

Two kinds of

samples

were

prepared :

very thin

samples spread

on a

single crystal

of mica and thicker

samples

with a lower

degree

of

orientation which were

picked

out of a

glass plate

with a rasor blade. The

X-ray

set up is very

simple, consisting

of a monochromatic beam

(CuKa)

issued

from a double bent

graphite pyrolitic crystal,

and a

photographic plate collecting

the diffracted

X-rays.

With the

technique

described

above,

we obtained

fibre

samples

with the fibre axis

parallel

to the stretch-

ing

direction. When this axis is

perpendicular

to the

X-ray beam,

one obtains a section of the

reciprocal

space

containing

the fibre axis which

gives

all the

information. We also used a Guinier camera with CuKa radiation in order to have a better accuracy

on the lattice

spacing

values.

3.

Experimental

results.

Typically,

in a columnar

mesophase

of disc-like

molecules,

a

single

domain

gives

a diffraction

pattern

that consists of one

equatorial plane containing

all

the

Bragg spots

and two kinds of diffuse features

lying

out this

equatorial plane [5].

One

part

of this diffuse

scattering

radiation

gives

an

image

of the

organization

of the

paraffinic

chains since this

part

can be

compared

to the main diffuse

ring

seen for a

molten

paraffin.

The other

part

of the diffuse scat-

tering

is localized in a set of

reciprocal planes parallel

to the

plane containing

the

Bragg peaks,

and per-

pendicular

to the column axis. One can relate the

Bragg peaks

to the inter-column

order,

which is a two- dimensional

crystalline order,

while the diffuse scat-

tering intensity

describes the way in which the cores are stacked in a column.

In

figure 1,

we

present X-ray patterns

of

aligned samples

for three of the four studied

mesophases (the mesophase

of

compound la

and the low-tem-

perature

one of

compound 16

are

isomorphous).

Among

these

patterns,

two fit the

general

scheme of a

columnar

phase (Figs. 1,

iv and

1, v);

the third one

for

compound lc

at low

temperature

is that of an

imperfect

3D

crystal (Figs.

1 i to

1, iii).

In this last case,

the information about its

organization

can be deduced

from a

semi-quantitative analysis

of the

pattern.

3.1 MESOPHASES OF

lc. - The

diffraction

patterns

of

figure 1,

i to

1,

iii show

Bragg spots lying

on

layer

lines

perpendicular

to the column axis

(Fig. 1 A).

The width of these

spots

increases with the

length

of

scattering

vector. There are

mainly

two

independent

factors

responsible

for this

broadening :

fluctuations of

orientation,

which

spread

out the

spots

on circular

arcs of the same

angular width,

and an increase of the width

parallel

to z

(fibre axis)

with the xy compo- nents of the

scattering

vector. The

layer

lines corres-

pond

to a

periodicity along

the column axis of 20.96

A.

All the

rings

observed on the

powder

pattern cor-

respond

to the

spots

observed for the oriented

samples,

and can be indexed on an orthorhombic lattice

a = 47.07

A,

b = 31.58

A,

c = 20.96

A.

In table II

we have listed the measured and calculated lattice

spacing

values for the observed

spots.

It is clear that many

Bragg

reflections are

missing.

On each

line,

we see one group of a few

peaks

which are all in the

same range of

scattering angles.

As the index I

increases,

the

perpendicular

component increases simultane-

ously.

Therefore all the 0 0 l reflexions vanish. This fact

implies

that the

projection

of the electronic

density along

the column axis is uniform. Moreover the extinctions are consistent with the existence of three

perpendicular

helicoidal two-fold axes i.e. with the space group

P 212121.

Besides the

major layer lines,

we also observed a second set of diffuse lines

(B) parallel

to the first set.

The line of

highest intensity corresponds

to a scat-

tering

vector of

(4.82 Å) - 1

while the others show a

modulation of this

period (4.82 A)

with a wave

length

of 20.96

A.

A few weak

Bragg spots

are located on this second set of

layer

lines. Therefore the structure must be modulated in the column axis direction.

The ratio between the two

periods

is close to the

rational number

13/3.

At

last,

a broad diffuse

scattering ring (C)

indicates that

paxa,ffinic

chains are in a

liquid

state.

The

general description

of this diffraction pattern is consistent with a model in which an

assembly

of

helicoidal columns forms a 3D

imperfect crystal.

The

pitch

of the helices is 20.96

A,

and the molecules

forming

the column are

equidistant

with a

period

of 4.82

A.

This last value is

equal

to the core-to-core distance in the column of the

cyclotriveratrylene crystals [3, 6].

For a more

precise description,

we can

use our

knowledge

of the core structure obtained from

different

crystalline

structures

[6, 7]. Unfortunately

we have no information on the external crown formed

by

the

aliphatic chains, except

for the fact that an

important

statistical disorder takes

place

at this level

as is evidenced

by

the

corresponding

diffuse

scattering

at about 4.5

A.

This lack of information

concerning

the

major part

of the

compound, together

with the

problem

of the fluctuations of the lattice unit vectors,

precludes

a

quantitative analysis

of the diffraction data.

Nevertheless,

one can derive some ideas on the

organization

of the

phase

from a

comparison

of the

reciprocal image

of a helix with our

pattern.

Let us

give

the form factor -% of a helix of

pitch P,

radius R on which atoms with a

scattering

factor F

are

regularly spaced

with a

period p

in the helix axis

direction. The relation

between p

and P is 13 p =

3, p = C [8]

where

Sz

and

Sl

are the two

components

of the scat-

tering

vector S

respectively parallel

and

perpendicular

(4)

Fig.

1. 2013 X-ray diffraction patterns of

aligned samples

of 1 band

Ic :

9

(i)

and

(ii) compound 1c

at room temperature on mica,

(iii)

without substrate. A : Bragg spots, B : diffuse lines, C :

broad/

diffuse

scattering

from melted chains, M : mica Bragg spots;

*

(iv) compound 1c

at 115 OC, without substrate;

9

(v) compound led

at 70 °C on mica.

The column axis

(stretching direction)

is

parallel

to the arrows.
(5)

354

Table II. -

Assignment of

Powder

rings

in the

mesophases.

(6)

to the helix axis. 1 m n are

integers

which

obey

the

relation 1 = 13 m + 3 n and

J.

is the Bessel function of order n. If we

keep

in mind the

general properties

of Bessel

functions,

we must assume that a measurable contribution to the

intensity

comes

only

from the low orders n and low values of the variable

(corresponding

to the first extrema of each

Jn).

The sum is reduced

to one Bessel function for each value of I. The

position

of the visible

Bragg spots

on the first

layer

line I = 3

(m

=

0, n = 1) corresponds

to a helix of mean radius

9

A.

Such an helix made of

equidistant rigid

molecules

will

give

a measurable

intensity

in

Bragg spots

for

/ = 6, (m = 0, n = 2),

/ = 7

(m = 1, n = - 2),

1 = 10

(m = 1, n = - 1),

1 = 13

(m = 1,

n =

0).

The intense diffuse lines for 1 = 13 and I = 10 mask the

Bragg spots

but for 1 =

7,

we observe one

Bragg peak lying

the diffuse line. Its

position

is consistent with the same helix of radius 9

A,

and its

intensity

is weak.

According

to this

simple model,

the two

layers

1 = 6 and 1 = 7

(Fig. 2c)

should be identical. In

fact,

the

intensity

of

the

Bragg spots

is

higher

on level 1 = 6 than on level

1 = 7 and the

position

of the intense visible

peaks corresponds

to a

larger

value for

S 1.

in this level

(I

=

6).

The same

problem

rises for the

comparison

between the

layers

1 = 9 and 1 = 4 since

Bragg spots

are not

unambiguously

identified for / = 4.

Nevertheless,

the

repartition

of the

intensity

on the

two

layer

lines 1 = 6 and / = 9 is consistent with a

helicoidal structure. As no other lines

correspond

to

the same

period,

we must admit that this helix is continuous

(Fig. 2b),

and that the

apparent pitch corresponds

to 20.96

A/n.

The absence of

higher

level

layer

line is

probably

a consequence of the

imper-

fections of the

crystal (the Bragg

reflexions are

only

visible in a limited domain of

scattering

vector

[9]).

If the

apparent period

is divided

by

n, the helix

is formed with molecules or

part

of molecules with a

n-fold

symmetry

and the structure factor is propor- tional to

nJ_(2 nRS 1.)

where m is an

integer.

From

the

positions

of the maxima of

intensity,

we deduce

that there are two

components

to the 3D helicoidal

structure a first one with a two-fold

symmetry

and a

Fig.

2. - Schematic

representation

of the diffracted

intensity

versus the

scattering

vector

S(I S I = 2 sin 0

for one column

of infinite

length

with three different structures :

a)

Uniaxial stack of discs of radius of

gyration

R. The distance between two discs

is p.

The diffracted

intensity

is

proportional

to

6 S.. - ’ j 2(2 1tS 1. R). Sz and S 1.

are the two components of the

scattering

vector

respectively parallel

and

perpendicular

p

to the column axis, m is an

integer

and

Jo

the zero order Bessel function.

b)

Continuous helix of radius R and

pitch

P. The diffracted

intensity

is

proportional

to

b Sz - 1 J’(2 7rS.L

R); n is an

integer and Jn

the nth order Bessel function.

c)

Discontinuous helix of radius R and

pitch

P with

points lying

on

equidistant planes perpendicular

to the column axis

at a distance p. The relation between the

period

C, the

pitch

P

and p

is C = 13, p = 3 P. The

intensity

is

proportional

to

b S - c Y- J (2 1tS1. R)J

where l, m, n are

integers

which

verify

the relation / = 13 m + 3 n.

(In

fact

figures

4a

and 4b are described

by

these last two relations in which n

(respectively

m) is

equal

to

zero.)

Notice that we have the first maximum of each Bessel function and

neglected

the contribution of

highest

order n > 4. We have plotted 1, m (therefore in figure 2a, n = //13) and the relative value of

intensity

maxima for each layer line of the top of figures 2b and 2c. The diffraction patterns of different kinds of assemblies of

parallel

columns derive from these three basic

examples :

for a column made of i identical continuous helices obtained

by

successive rotations of 2 7c/i around their common

axis, the scattered

intensity

vanishes unless n = 0 modulo i. If the columns form a 2D array,

only

the zero

layer

line is dis- continuous, but if we have a 3D lattice, all the

layer

lines become discontinuous.

d) Schematic

representation

of the pattern of

figure

2a : we have a

superposition

of figure 2a to which we add a super-

structure of

period

P,

figure

2c, with a

large

radius R ~ 9

A,

and

figures

2b in which there are

only

two

layers corresponding

to n = 2, n = 3 with R

equal

to 4.3 and 4.9 A

respectively.

The diffuse

scattering

indication of the molten state of the chain is not shown on this schematic

drawing.

(7)

356

mean radius of 4.3

A

and a second one with a three- fold

symmetry

and a radius 4.9

A.

The existence of a

continuous helix inside a 3D

crystal

of

regularly

stacked molecules can be described as follows : In each column the same

equation

of a helix

gives

the relation between the

in-plane

coordinate and the

z coordinate of atoms, but in a

given

column the

values of z are discrete : z = wp + 6 where w is an

integer

and 6 is fixed for one

given column,

but can take all the values between 0 and p. The last feature that we have to take into account is the existence of the diffusive discs. All these discs have the same

aspect

and

correspond

to a mean

period

of 4.82

A (Fig. 2a)

with a modulation of 20.96

A.

The modulation

only

has a

longitudinal component

and therefore the helicoidal order has

disappeared

at this level. The

intensity dependence along

these

planes

agrees with the calculated one for a conical core made of the carbon atoms of the

ring (CH2CP)3 using

the structural data of references

[6]

and

[7] (Fig. 3).

The presence of

a continuous line means

that,

if the cones are

regularly

stacked in

columns,

no correlation in their

positions along

the column axis is established between

neigh- bouring

columns. Let us

put

all these observations

together

in order to have an idea of the molecular

organization :

the conical molecules are stacked in columns with a

period

of 4.82

A.

The

aliphatic

chains

of each molecule are in a disordered state, but the

paraffinic

medium is not

unformely spread

around

the core and is wounded in a helix

(Fig. 4).

Parallel helices are correlated

along

the z direction

with fluctuations of weak

amplitude.

These fluctuations

are

responsible

for the width of the

large angle Bragg

reflections. Between the helix of

paraffmic

chains

and the non-helicoidal central

part

of the

column,

the benzoate groups form a continuous helix : the orientation and the axial

positions

are related in a

column,

the orientation at a

given

level is the same for

two

equivalent

columns but the location of that

part

of the molecule is not identical for each column.

Fig.

3. - Form factor of a conical core

rotating

around its

axis calculated from the

crystal

structure data

[6, 7]

at the

level of the first diffuse

plane :

full line :

taking

into account

only

the carbon atoms of the

ring (CH2l/»3;

dashed line :

adding

to this

ring

an external shell

including

the six

-O-C

linkages.

Fig.

4. - Schematic

representation

of the helical structure of the

lc low-temperature mesophase (section

perpendicular

to the column

axis) :

a, tribenzo conical core + the six benzoate groups; b, the

paraffmic

crown

consisting

of the

six

n-dodecyl

chains.

This correlation loss on

going

from the outside of a

column towards its core may be induced either

by

the

flexibility

of the ester

linkages

or

by

the existence

of an orientational disorder of the cones

pointing

up or down.

The centre of the column is

isotropic

around the

column axis. All this

description

is derived from a

qualitative analysis

of the fibre

pattern

and this

analysis

does not

provide

any information about the

phase relationship

between the different

parts

of the helix. We can notice that weak

spots

seen on the level I = 1

(m = 1, n

= -

4)

can be due to inter-

mediate zones between the benzoate group

(R

= 4-5

A)

and 9

A. Unfortunately

the number of

parameters

that one should introduce in a

quantitative

evaluation

is too

high

in

comparison

with the

precision

of

experi-

mental data

(the imperfection

of the

crystal

is res-

ponsible

for the low resolution in wave vector and for the small number of visible

reflexions).

One

important point

about the molecular conformation

concerns the orientation of the conical cores. The

proposed

space group P

2t 2i 2i implies

that the two

columns of the unit cell are

polarized

in

opposite

directions but in fact a space group P

212 21

is not

excluded. In such a case the two

possible

orientations

are still

equally probable.

Then the orientation of the cones in one column is

probably

maintained over a

large

correlation

length,

but the orientation inside

a column is

independent

of the

position

of the column

in the unit cell. This second space group

implies

a

large longitudinal displacement

of the cores relative

to the external

part

of the columns and

probably

is the best choice for the

description

of this

organiza-

tion in which a

high degree

of disorder coexists with

a rather well defined

periodicity.

On

heating

this

mesophase,

the 3D lattice and the modulation with

a 20.96

A period disappears (Fig. 1, iv) :

a diffuse

scattered

intensity replaces

the 2nd and 3rd

layer

lines.

The helical structure therefore remains

although

the

chains are in a less ordered state and the correlation between columns is lost. The 2D columnar lattice is

hexagonal

and exhibits the same structure as that of the

Dbo

columnar

phase

of

triphenylene

ethers

[10].

(8)

3.2 MESOPHASES OF

la

AND

lb.

- Derivative 1 b forms two columnar

phases

which

only

differ

by

their

2D lattices

(Table II) :

at low

temperatures,

we observe

an

oblique

lattice with two columns per unit cell. A similar

mesophase

is seen for

compound la.

At

high temperatures compound 1 b

forms an

hexagonal

columnar

phase

with four columns per unit cell.

The diffraction

patterns

of oriented

samples

of the two

mesophases

are similar. The cores are

regularly

stacked at a distance of 4.82

A.

The maximum of

intensity

on the diffuse

plane

which is characteristic of the core

ordering

in a column is

split

out of the

symmetry

axis. In fact if we refer to the

crystal

struc-

ture of

cyclotriveratrylene [6, 7]

the

intensity depen-

dence fits with a conical core

including

the

carboxylic

groups

(Fig. 3).

In this case, since we have

only

a 2D

crystalline ordering,

we do not know

anything

about

the dielectric

properties

of the

mesophases.

The

hexagonal

lattice

(confirmed by optical

observations

[4])

with four columns per unit cell

corresponds

to

a rather

peculiar phase,

in which three columns are

equivalent

and on low

symmetry sites,

while the fourth one is on a

higher symmetry position [11, 12].

If each column is

polar (i.e.

with an

unique

orien-

tation,

up or

down,

of the conical

cores)

we should

have a ferroelectric structure of

symmetry

3 m or 6 mm.

4. Conclusion.

Some of the

mesophases

made of molecules

having

a conical core are similar to those of disc-like mole-

cules. Besides these

mesophases

an

original

structure

exists for the

low-temperature mesophase

of

1c :

the external shell of each

column,

i.e. the

paraffinic medium,

has an helical structure. The columns are

arranged

in an

imperfect

tridimensional lattice in the

same way as a set of identical infinite screws threaded

one in another with a certain looseness.

Moreover,

the chains are not in an all-trans

conformation,

but rather in a « molten state » ; the cores are

regularly

stacked in a

column,

but uncorrelated from one

column to another. This lack of correlation may ori-

ginate

from an orientational disorder due to an

inversion of the cone direction within each column

(the

half-life of a

given

conical conformer is about one

minute at 100 OC in

solution).

This

mesophase

is not

ferroelectric.

The other

mesophases

of the three studied

cyclotri-

catechylene

derivatives can be ferroelectric if no

orientational disorder of the cores

(up

and

down)

takes

place,

but

X-ray

data do not

give

any infor- mation on this

problem,

besides a confirmation of the conical

shape

of the cores.

Acknowledgments.

We are indebted to Dr. M. Cesario for

bringing

some

of her

unpublished

data to our

knowledge.

NB : Dr. Lin Lei has

brought

to our attention his

publication (in

chinese Wuli 11

(1982) 171)

in which

the

symmetry properties

of eventual

mesophases

made

of « bowl-like » molecules have been

already

dis-

cussed.

References

[1]

CHANDRASEKHAR, S., SHADASHIVA, B. K., SURESH, K. A.,

Pramana 9

(1977)

471.

[2]

DESTRADE, C., GASPAROUX, H., FOUCHER, P., NGUYEN

HUU TINH, MALTHÊTE, J., JACQUES, J., J. Chim.

Phys.

80

(1983)

137.

[3]

MALTHÊTE, J., COLLET, A., Nouv. J. Chim. 9

(1985)

151.

[4]

ZIMMERMANN, H., POUPKO, R., LUZ, Z., BILLARD, J.,

Z.

Naturforsch.

40a

(1985)

149.

[5]

LEVELUT, A. M., J. Chim.

Phys.

80

(1983)

149.

[6]

CERRINI, S., GIGLO, E., MAZZA, F., PAVEL, N. V.,

Acta

Crystallogr.

B 35

(1979)

2605.

[7]

CANCEILL, J., CESARIO, M., COLLET, A., GUILHEM, J., PASCARD, C.,

unpublished,

see reference

[3].

[8] COCHRAN, W., CRICK, F. H. C., VAND, V., Acta Crys- tallogr. 5 (1952) 581.

[9] GUINIER, A., X-Ray

diffraction

in

crystals, imperfect crystals and amorphous

bodies (W. Freeman and Co, San Francisco,

U.S.A.)

1963.

[10]

LEVELUT, A. M., J.

Physique

Lett. 40

(1979)

L-81.

[11]

LEVELUT, A. M., OSWALD, P., GHANEM, A., MALTHÊTE, J., J.

Physique

45

(1984)

745.

[12]

Yu FENG SUN, SWIFT, I., J.

Physique

45

(1984)

1801.

Referências

Documentos relacionados