Synthesis and characterization from Raman spectroscopy of pristine, potassium-doped and rubidium-doped fullerenes C60/C70

HAL Id: jpa-00246465

https://hal.archives-ouvertes.fr/jpa-00246465

Submitted on 1 Jan 1992

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.

Synthesis and characterization from Raman spectroscopy of pristine, potassium-doped and

rubidium-doped fullerenes C60/C70

A. Zahab, J. Sauvajol, L. Firlej, R. Aznar, P. Bernier

To cite this version:

A. Zahab, J. Sauvajol, L. Firlej, R. Aznar, P. Bernier. Synthesis and characterization from Ra-

man spectroscopy of pristine, potassium-doped and rubidium-doped fullerenes C60/C70. Journal de

Physique I, EDP Sciences, 1992, 2 (1), pp.7-13. �10.1051/jp1:1992119�. �jpa-00246465�

Classification Physics Abstracts

63.20 78.30

Short Communication

Synthesis and characterization from Raman spectroscopy

of pristine, potassium-doped and rubidium-doped

fuflerenes C60/Cm

A.

Zahab,

J-L-

Sauvajol,

L.

Firlej~

R. Aznar and P. Bemier

Groupe de Dynamique des Phases Condensdes

(*),

Universitd Montpellier II, Sdences et Tech- nique du Languedoc, 34095 Montpellier Cedex 05, France

(Received

7 October 1991, ^revised 12 November 1991, accepted 13 November

1991)

RAsumA. La prdparation et la s4par&tion du m61ange

C60/Cm

sont d6crites. Une _com-

paraison entre les spectres ^{Raman de} C60 et de C60 dop4 _au potassium et _au rubidium est

pr4sent6e et &nalys6e. L'eflet du recuit _sur la st&bilis&tion de la phase la plus conductrice _est mise _en 4vidence _par diffusion Ram&n _sur le C60 dop4 _au rubidium.

Abstract. The preparation and separation of the C60/C~O ^mixture are described. A _com-

parison between the Raman spectra ^of C60, Potassium-doped C60 and rubidium-doped C60 is

presented and analysed. From Raman experiments the role of annealing in the stabilization of the highly conducting phase is evidenced in rubidium-doped C60.

Raman Spectroscopy iS known to be _a

powerful

tool _to

Study

the

inhomogeneity

and disor- der in intercalated

graphite

and

conducting polymers

and to follow the Stuctural

changes

_upon

intercalation of _a

dopant

in all these materials.

Especially

the shift and

broadening

of the lla-

man lines of the host material and the _appearence of _new llaman lines under

doping

have been

extensively investigated

and

explained

in terms of

charge

transfer between the host material and the

dopant [1,2].

^{Late last} _year, ^Kratschmer et al. [3, 4] ^have

reported

that fullerenes

Cn,

_new

allotropes

of carbon could be

prepared

and isolated in

macroscopic quantities.

The

recent

discovery

of

superconductivity

in C60

doped

with alkali metals [5-8] ^{has introduced} a new

family

of three-dimensional molecular conductors and

opened

_{a very} active research _area.

A

variety

of

spectroscopic

studies have been

performed

to

explore

the

properties

of

M~C60-type

materials

(M

₌

K, Rb, Cs...) [9-ll].

(*) ^U-A. 233.

8 JOURNAL DE PHYSIQUE I N°I

In this _{paper we} present

preliminary

results obtained

by

llaman spectroscopy on C60

j

K~C60

and

Rb~C60 (with

_~

= 3 and _{~ >}

3).

In this last

compound

the role of

annealing

in the stabilization of the

highly conducting phase

is

analysed.

The Fullerene

production technique

which _we have used is based _on the _arc

procedure

de- scribed in [3]. ^The stainless-steel water-cooled collection

cylinder

of10 of volume is evacuated

(r- ^10~~ Torr)

and washed _many times with He gas.

During

the _process the

dynamic

He pres-

sure within the

cylinder

is maintained at 150 Torr

by controling

_gas flow and

pumping

_power.

The electric _arc is created between _two

graphite

^rods

(99.99il

of

purity,

^natural

isotope

_com-

position)

of 6 _mrn

diameter,

fixed _on water-cooled electrodes connected to _a

high

_power dc

current source. The

positive

electrode is

mobile;

the

graphite

rod fixed to it

(18

cm

long)

is

totally

sublimated

during

^the _process. The rod _on the

negative,

^{fixed electrode}

(5

cm

long)

remains

intact;

_a

volcanc-shaped deposit

_{grows on} it. Sublimation of

graphite

with dc current

(I

₌ 100

A)

instead of

previously reported

_ac current aparatus ^offers at least _one

advantage.

The

stability

of the _arc

(and

the

yield

of the

process) flepends strongly

_on fluctuations of the distance between the nearest

points

of the rods. These fluctuations decrease

significantly

when the

graphite deposit

_grows

only

on one of the electrodes. The _arc is observed

through

_a

glass

window in the

cylinder

and maintained

by manually moving

the

positive

electrode. A _screen is installed all around the _arc in order _to protect ^the soot

deposited

_on the cold walls of the

collection

cylinder

from the

degrading

action of UV radiation of the _arc

[12].

^{In usual}

working

conditions

(He

_{pressure r-} ¹⁵⁰

Torr,

dc

voltage

25 V and 3

graphite

rods consumed

during

20-30

minutes)

the

yield

of _our system

(the

ratio of the _mass of

graphite

used to the _mass of soot

obtained)

is 0.6. This

proportion

is not affected

by

the _use of static instead of dynamic

pressure of He in the reactor.

The soot _was collected after sublimation

directly

into _a filter and extracted with

boiling

toluene

during

4 hours. The dark-red solution obtained _was dried in _a rotary evaporator

yielding

_a black

powder containing

the mixture ofC60 ^and

Cm

The residual _amounts of solvent

were removed

by heating

the

sample

at 200°C

during

2 hours under _a

dynamic

_vacuum of

10~~

_{Torr [13]. The}

sample composition

_was determined

by

~~C NMR solid-state spectroscopy before

doping [14].

^The

composition

of the

pristing samples

is m 99il

C60/C~O.

^The

Cm/C~o

molecular ratio _was estimated _at

1/4.

In order _to obtain

doped fullerenes,

_{a vapor}

phase doping

_process of

C60/C~O powder by

alkali metals

(potassium

and

rubidium)

has been used

[15].

^The

C60/C~O powder

is enclosed in _a

glass

tube that contained the alkali-metal. This

glass

cell is inside _a temperature

gradient

furnace. The temperature ^{of the}

C60/C~O powder

^{is monitored} to be _a few

degrees

above the sublimation temperature ^{of the involved alkali metal. All the}

doped _C60/C~O samples

^used in this

study

have been

doped

with this

procedure. Inspite

of

that,

the

pristine samples

^have a

significant

amount of Cm> in the

following

and in order _to

clarify

the

notation~

the

C60/Cm doped samples

will be called:

M~C60

_{with M}

=

K,

Rb.

Roman spectra ^of C60 thin films

on

suprasil

slide and KBr

plates

have been

performed

at room temperature

(on

the _one hand

exposed

to air and _on the other hand under

helium)

and at low temperature ^T ₌ ^{7 K. Raman} measurements _on K3C60

(highly conducting phase [16])

and

Rb~C60 powder

in _a closed

glass-cell

have been carried out at _room temperature. Raman spectra _were recorded

using

_a standard

"Coderg

T800"

triple

monochromator spectrometer.

The instrumental linewidth _was fixed _at 6

cm~~

_{The 5145}

1

_{line of}

an argon ion laser _was used _as the

light

_source, the _{power on} the

sample being

limited to below 50 mW and the

incident beam defocussed. The scattered

light

_was collected and focussed _on the spectrometer slit in _a

back-scattering

arrangement.

The Raman spectrum

performed

_{on a}

C60/Cm

thin film

sample

in the

frequency

_range 200

cm~~-1700

cm~~ is

presented

in

figure

I. The spectrum shows all the features

previously

described

by

Bethune et al. [17] ^{and the}

intensity

ratio of the 1469 cm~~ line

(C60, Ag mode)

and 1568 cm~~

(Cm, Ag mode)

is in agreement ^with our evaluation from NMR data of _a

Cm /C~o ,molecular

ratio of about

1/4.

_~

C60/C70 _on supresll slide

I

5145

~

T z 300 K, on air

c

E

t4 t4 tC

o

200 400 600 BOO lo00 1200 1400 1600

Raman Shift

(cm-1)

Fig.

I. Raman spectrum ^{of thin film of} C60 and Cm deposited _on suprasil substrate.

+~ Undoped C60/C70 _on KBr

~

l~ _Upper

curve: T

= 295 K

Lower curve: T

= lo K

£

Ql C

~

W

l000 1200 1400 1600

Raman Shift (cm-1)

Fig. 2. Raman spectra ^for _a

C60/Cm

thin film

on KBr plate at room temperature

(T

₌ 295 K, lower

curve)

and low temperature

(T

= 7 1(, _upper

curve).

In

figure

2 the llaman spectra

performed

_{on a} thin film _on KBr at _room temperature and low temperature are shown. C60 exhibits _an order-disorder

phase

transition at around T ₌ 250 K [18]. Nevertheless _no

changes

of the Raman spectrum with temperature are

observed~ especially

10 JOURNAL DE PHYSIQUE I N°1

the 1469

cm~~

and 1569

cm~~

_lines

are not

significantly shifted, only

the

double-peak

structure of the 1569

cm~~

_band is _more

clearly

resolved _at low temperature. ^In _summary, we observe the _same number of lines in the Raman spectra ^{of the} room temperature

phase (one

molecule in the

primitive cell)

and low temperature

phase (four

^molecules in the

primitive cell)

and all

these lines do _not shift with temperature. ^{A first}

explanation

of these results is that the values of the force constants for the involved vibrations

are close in the _room- and

low-temperature phases.

In other

words,

^{the role} of the

crystalline

field

(and

its evolution with

temperature)

iri the vibrational

dynamics

in the C60

crystalline phases

could _not be evidenced from these llaman

experiments.

'~ T₌ 300 K

~~$~~

~~~$~~#©

_C60/C70

1000 1200 1400 1600

Raman Shift

(cm-1)

Fig. 3. Raman spectra for _a

C60/Cm

thin film _on suprasil slide (upper

curve)

and for _{a supercon-}

ducting K3C60 Powder sample

(lower curve).

In

figure

3 _a

comparison

between the Raman spectra

performed

_{on a}

C60/Cm sample (upper curve)

and _a K3C60

sample (lower curve)

in the

frequency

_range

[1000 cm~~-1700 cm~~]

^is

given. Upon doping

_a

general

decrease of the llaman

intensity

h observed. All the lines

assigned

to the vibrational modes of C60 and

Cm

vanish in the

K3C60 Phase.

This _means that both the C60 ^and

Cm

parts of the

sample

_are

doped.

The most

striking

features of Raman spectrum of the K3C60

conducting phase

_{are a} strong

peak

which appears ^at 1445 cm~~ _and

a broad and weak band

pointed

out at 1583 cm~~. The downshift of the strong ^{line of the} llaman spectra ^from 1468

cm~~

in C60 to 1445 cm~~ in

K3C60

is in agreement ^{with the result} of Haddon and coworkers [9]. ^The appearance of the 1445 cm~~ line in the Raman spectrum ^is the

"signature"

^of the

growing K3C60 conducting phase

_upon

doping.

This line is

significantly

broadened with respect to the width of the 1468 cm~~ line. This result _may be

certainly

related to a structural disorder induced _upon

doping.

The behaviour of the 1468 cm~~ line _can be

compared

with those evidenced in

graphite

intercalation

compounds [I]

^and

charge

^transfer salts

[19].

^{Haddon and coworkers} suggest ^that this behaviour is

analogous

to that observed for the

intralayer

Raman mode at 1582 cm~~ _in the

stage-I

donor

compounds

^of

graphite (KC8j Rbcs

^and

CsC8).

In these

graphite

interca-

lation

compounds

the 1582 cm~~

intralayer

mode is shifted to 1547 cm~~ _and broadened [20];

these features have been

assigned

to _a Fano _resonance between _an electronic Raman

scattering

continuum and the

intralayer

mode. With

regard

to the

shape

of the Raman

peak

this _prc-

cess here _can be ruled out. In the other stage-n ^donor

compounds (n

>

2)

one

expected

to observe two

peaks,

_one at the

position

of the

intralayer

^{mode and} _one

displaced

in

frequency (upshifted

with respect to the

intralayer mode) [I].

The 1448 cm~~ line does not exhibit _any of these features. On the other hand it is known that in _some

charge

transfer salts the

energies

of

symmetric

modes of host molecule _are sensitive to the amount of

charge

_on the molecule.

For instance the C=C

stretching

mode in

TCNQ depends linearly

_on the

degree

of

charge

transfer _on

TCNQ

salts

[19].

^{It is found that} a 60 cm~l downshift of the C=C

streching

mode

corresponds

to _a

complete charge

transfer. The behaviour of the 1448 cm~l line is close to the

one

previously

described for

charge

transfer salts and

suggests

^{that the} shift is related _to the

charge

transfer between the host and the

dopant.

In

agreement

^{with this} statement Haddon and coworkers [9] have shown that in the

highly doped phase

this

peak

_moves to 1430 cm~l

The

assignment

of the 1583 cm~l line is not

completely

clear _at this time. It is

important

to note its

opposite

behaviour

(upshift)

with respect to the 1448 cm~l _line

one. In other words

an

explanation

in terms of

charge

^transfer

analogous

with that

previously

discussed _seems to be ruled _out. The

frequency

of this line is closer to the strongest ^{Raman line observed in}

graphite

and

amorphous

carbon.

Inspite

of this agreement ^the

high-stability

of the C60 and Cm molecules allows _{us to} leave _out the

development

of

amorphous

carbon _or

graphite

_upon

doping.

In consequence the 1583 cm~~ line in

K3C60

cannot be

explained by

this _way. At last the attribution of this line to _an

oxydation

_process

occuring during

the

doping

_process cannot be

neglected.

Roman

experiments

from

Rb~C60 samples

have also been

performed.

It is to our

knowledge

the first time that Raman measurements have been made _on this kind of

compound.

The

Rqman

lines observed in these spectra were

assigned

with reference with the

previous

discussion

about the Raman spectra in C60 ^and

K3C60 samples.

In this

study

the role of

annealing

in the stabilization of the

conducting phase

has been

analyzed.

'~ ^RbC60

~

Upper cum»: aller annealing

'4 _Lower

curve: before annealing

1000 1200 1400 1600

Raman

Shift (cm.1)

Fig. 4. Roman spectra given by_a RbxC6o sample before

annealing (lower curve)

and after annealing

at 420°C

(upper curve).

In

figure

4 _are

displaye#

the Raman spectra ^{obtained in}

Rb~C60 sample

before

annealing

(Fig. 4,

lower

curve)

and after

annealing (Fig.

4, upper

curve).

The best conditions of

12 JOURNAL DE PHYSIQUE I N°1

annealing

have been established from [22]. ^The

annealing

temperature is close to

420°C,

and the

annealing

duration is four

days.

I)

Before

annealing

and in the

frequency

_range of interest

[1000

^cm~~-1700

cm~~]

six fines

are observed _at: 1235

cm~~ (weak),

1385

cm~~ (weak),1435 ^cm~~ (strong~

1450

cm~~ (weak)~

1472 cm~~

(medium)

and 1575

cm~~ (weak).

^The most intense line

(1435 cm~~)

can be

assigned

to the

corresponding Ag

mode

(1468 cm~~)

in reduced form of C60. The

position

of this line is lower than those observed in the

K3C60 highly conducting phase (Ref.

_{[9] and} this

work).

In agreement ^{with Haddon and coworkers} [9] we claim that this line

corresponds

to

a vibrational mode in _a

highly doped RbzC6o

state

(insulating phase)

with _{~ >} 3. Indeed in the

highly doped Kz>~C60

state the

analogous

line _was observed at 1430 cm~~ [9]. ^{The lines} at 1235

crn~~,

1472

cm~~

and 1575 cm~~ _are

assigned

to the amount of

undoped

C60 in the

unannealed

sample.

2)

After the

annealing

of the

sample, only

_one Raman

peak

at 1450 cm~~ _{appears in the} spectrum ^with a

significant intensity.

From

comparison

with the Raman spectrum ⁱⁿ K3C60

(Ref.

_[9] and this

work)

we claim that the _appearance of this line is the Raman

"signature"

of the

growing

of the

Rb3C60 highly conducting phase.

It is

interesting

to note that the

peak frequency

is not

dopant dependent.

The _same behaviour _was observed in

n-doped polyacetylene

[2]. ^{A shoulder} at 1435 cm~~ _{is also observed and attributed to the amount of the}

highly doped phase (~

_>

3)

^{in the annealed}

sample.

These results

emphasize

the role of

annealing

in the stabilization of the

highly conducting phase.

NMR results _are in agreement with these conclusions

[21].

Roman

experiments

_on

highly conducting phase

of

K3C60

and

Rb3C60

_{as a} function of the temperature are in _progress. The aim of these

experiments

is to evidence the "Raman

signature"

of the

superconducting

state which _appears in these

compounds

below 18 K.

References

[I] ^SOLIN S-A-, Pbysica 899

(1980)

443.

[2]^{A review of the features of the Raman} spectra of doped

conducting

polymers can be found in the

Proceedings

of the International Conference _on Science and

Technology

of Synthetic Metals, published in: Synth. Met. 27-28

(1988) (Santa

Fe,

1988);

Synth. Met. 41-43

(199i) (Tubingen, 1990).

[3]I~RATSCHMER W-, FOSTIROPOULOS l~. and fIUFFMAN D-R-, Chem. Phys- Lett- 170

(1990)

167.

[4] I~RATSCHMER W., LAMB L-D-, FOSTIROPOULOS l~. and fIUFFMAN D-R-, Nature 347

(1990)

354.

[5] ^HEBARD A.F.~ ROSSEJNSKY M-J-, HADDON R.C., MURPHY D.W.~ GLARUM S.H.~ PALSTRA T.T.M., RAMIREZ A.P. and Koft£AN A.R.~ Nature

(submitted).

[6]HOLCZER I(., KLEIN O-j fIUANG S--M., I(ANNER R-B-, FU K--I., WHETTEN R-L- and DIEDERICH F-, Science

(1991)

submitted.

[7]ROSSEINSKY M-I-, RAMIREZ A-P-j GLARUM S-H-j MURPHY D-W-, HADDON R-C-, fIEBARD A-F-, PALSTRA T-T-M-j KOR£AN A-R-, ZAHURAK S-M- and MAKHIJA A-V-, Phys- Rev.

Lett. 66

(1991)

2830.

[8] TANIGAKJ I(-, EBBESEN T-W-j SAITO S-j MIZUKI I-, TSAJ I-S-, KUBO l~-, KUROSHJMA S., Nature 352

(1991)

222.

[9] ^HADDON R-C-, HEBARD A-F-, RossEiNsI<Y M.J., MURPHY D-W-, DucLos S.J., LYONS K-B-, MILLER B-, ROSAMILIA I-M-, FLEMING R-M-, KOR£AN A-R-, GLARUM S-H-, MAKHIJA A-V-,

MULLER A-J-, EICK R-H-, ZAHURAK S-M-, TYCKO R., DABBAGH G. and THIEL F-A-, Nature 350

(1991)

320.

[10] ^WERTHEIM G.K., RowE J.E., BUCHANAN D.N.E., CHABAN E-E-, HEBARD A-F-, KORrAN A.R., MAKHIJA A-V- and fIADDON R-C-, Science 252

(1991)

141?.

Ill]

TYCKO R., DABBAGH G., ROSSEINSKY M-I-, MURPHY D-W-, FLEMING R-M-, RAMIREZ A-P- and TULLY J-C-, submitted to Science

(1991).

[12] ^TAYLOR R-, PARSON I.P., AVENT A-G-, RANNARD S-P-, DENNIS T-I-, HARE I-P-, I~ROTO H-W- and WALTON D-R-M-, Nature 351

(1991)

277.

[13] ^ZAHAB A., FIRLEJ L., BERNIER P. and AZNAR R.,

(In preparation).

[14] ^ZAHAB A., FIRLEJ L., BERNIER P- and AZNAR R.,

(In preparation).

[15] BERNIER P-j ZAHAB A., FIRLEJ L. and AZNAR R-,

(In preparation).

[16] ^The superconductivity of this sample _w&s confirmed by the measurements of magnetic suscepti- bility (AUBAN-SENZIER P., to be

published).

[17] ^BETHUNE D-S-, MEIJER G-, TANG W-C- and ROSEN H-I-, Chem. Phys. Lent- 174

(1990)

219-

[l8]

fIEINEY P-A-, FISCHER I-E-, MCGHIE A-R-, ROAfANOW W-I-, DENENSTEIN A-M-, MCCAULEY I-P- lr-, SMITH III A-B- and COX D-E-, Phys. Rev- Lett. 66

(1991)

2911-

[l9]

TOKUMOTO M-, I~OSHIGUZA N-, MURATA K-, I~INOSHITA N-, ANZAI fl- and ISHIGURO T-j Mol.

Cryst. Liq. Cryst. 85

(1982)195.

[20] ^NEMANICH R-J-, SOLIN S-A- and GU%RARD D., Pbys. Rev. 816

(1977)

2965.

[2i]

ZAHAB A., FIRLEJ L-j ^{BERNIER P. and AZNAR} R.,

(In preparation).

[22] ^FISCHER I-E.,

(private communication).