The metallic phase of the organic conductor TMTSF -DMTCNQ stabilized by a weak irradiation disorder

HAL Id: jpa-00209477

https://hal.archives-ouvertes.fr/jpa-00209477

Submitted on 1 Jan 1982

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.

The metallic phase of the organic conductor TMTSF -DMTCNQ stabilized by a weak irradiation disorder

L. Forró, A. Jánossy, L. Zuppiroli, K. Bechgaard

To cite this version:

L. Forró, A. Jánossy, L. Zuppiroli, K. Bechgaard. The metallic phase of the organic conductor TMTSF -DMTCNQ stabilized by a weak irradiation disorder. Journal de Physique, 1982, 43 (6), pp.977-981.

�10.1051/jphys:01982004306097700�. �jpa-00209477�

The metallic phase ^{of the} organic ^conductor TMTSF -DMTCNQ ^stabilized by ^{a weak} irradiation disorder

L. Forró, ^A. Jánossy (*), ^L. Zuppiroli

Section d’Etude des Solides Irradiés, ^{Centre d’Etudes} Nucléaires de Fontenay-aux-Roses, 92260, ^France and K. Bechgaard

H.C. Oërsted Institute, Universitetsparken 5, ^{DK 2100} Copenhagen, ^Denmark (Reçu ^{le 21 octobre 1981, révisé le} 4 fevrier 1982, accepté ^le 22 fevrier 1982)

Résumé.

Lorsque le conducteur organique TMTSF-DMTCNQ ^{est soumis à une} faible irradiation aux rayons X

endommageant deux molécules pour mille (2 °/°°), ^{la phase} isolante de basse température ^est détruite par le désordre.

Des mesures du coefficient de Hall et du pouvoir thermoélectrique de cristaux purs ^et irradiés ont été effectuées

jusqu’à 2,2 ^K. Elles démontrent que l’état métallique ^{a été stabilisé} par l’irradiation et etendujusqu’aux très basses

températures.

Abstract

A low dose X-ray irradiation damaging ^two molecules per thousand (2 ~) ^{destroys the low tempera-}

ture Peierls insulating phase ^{of the} organic ^conductor TMTSF-DMTCNQ. Measurements of the Hall constant and of the thermoelectric power down ^to 2.2 K demonstrate that a metallic state is stabilized by irradiation and extended to very low temperatures.

Classification Physics ^Abstracts

61.80H - 72.15N - 72.80L

1. Introduction.

The metallic phase ^of quasi-

one-dimensional organic conductors is well known to be unstable at low temperatures. Usually ^a phase

transition occurs driven by the 3 dimensional ordering

of charge density ^{waves or in} some cases by spin density ^{waves or even} by supraconducting ^fluctua-

tions. For the last several _years the stabilization of the metallic phase ^{to all} temperatures ^{has been one} of the aims of rearch in the field of organic ^conduc-

tors [11. ^{The first} way ^to stabilize the metallic state was to apply hydrostatic pressure in TMTSF-

DMTCNQ [3] ^{and now in several} (TMTSF)2X systems [4]. ^{In the} present paper it will be shown on

the example ^of TMTSF-DMTCNQ (tetramethyl-

tetraselenafulvalene - dimethyltetracyanoquinodime - thane) that weak irradiation induced’ disorder stabi- lizes the metallic state to temperatures ^{well below} the Peierls transition of the _pure material.

Irradiation induced defects are well known to inhibit the Peierls transition in several organic ^con-

ductors [5]. ^{However the nature} of the low tempera-

ture disordered phase ^{was not clear.}

Defect concentrations of the order of 1 % give ^rise

to a conductivity activated with temperature. ^The

activation energies ^{of the order of a few meV are even}

better defined than in the Peierls distorted semi-

conducting ^{state of the} pure material. It was conjec- tured, however that the activated behaviour of the dis- ordered sample ^{does not reflect a} semi-conducting

state. The disordered material would be a quasi-one-

dimensional metal but a metal of segments ^inter-

rupted by the defects [6].

When defect concentrations much lower than 1 %

are introduced in organic conductors, ^the longitudi-

nal conductivity below the Peierls transition is observ- ed to increase and reach values typical of the metallic state [6]. ^{In the} particular ^case of TMTSF-DMTCNQ,

the conductivity ^increases by ^{two orders of} magnitude

for a defect concentration of 0.2 jo ^{at 21} K, ^{while the} anisotropy changes ^{less than a factor of two. The fact}

that the longitudinal ^and transverse conductivities vary in the same manner suggests ^{that a} change ^{in the}

number of carriers is the main ^source of the variations of the conductivities. Our aim is ^to present ^detailed data on the low temperature transport properties

of TMTSF-DMTCNQ irradiated with low doses.

2. Experimental techniques and results.

D.C.

conductivity measurements were performed by ^the

four probe method in the highly conducting ^direc-

tion of the sample.

Thermoelectric _power measurements were carried

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphys:01982004306097700

978

out in the vacuum better than 10-2 ^torr. Temperature gradient ^was generated by ^two miniature 500 Q resistances placed ^{at the ends of the} sample, ^{and moni-}

tored with chromel-alumel differential thermocouple.

The Hall effect was measured using ^an A.C. techni- que with a current in the _range of 10-400 JlA ^at

70 Hz. Temperature ^drifts cause a severe difficulty

in measuring ^Hall voltages ^{below the} phase ^transi-

tion because the non alignment resistance component varies strongly. ^{In such cases we} swept the X-axis of the X Y recorder by ^the signal ^{of a carbon in} glass-

resistor and recorded the Hall signal ^at maximum and

zero fields several times. This method enabled us to correct for temperature drifts of 50 mK.

The Hall constant was calculated from the expres- sion R H = 108 x VH d/HI (cm3/C), ^where VH(V)

is the measured Hall-voltage, d (cm) ^{is the} sample

dimension parallel ^{with the} magnetic ^field H(G).

The current I(A) ^is flowing along ^the high conducting b-axis, ^{while the} magnetic ^{field was} applied along ^a-

axis. The Hall voltage ^{was observed to be linear}

with the magnetic ^field up ^to the highest applied ^field

of 18.5 kG. The Hall constant measurements were

perfectly reproducible within the error bars plotted

of figure ^{2. We} have measured several times 3 diffe- rent crystals ^of pure and irradiated material in order to check this reproducibility.

Typical sample dimensions for D.C. conductivity

and TEP measurements were 8 x 0.1 x 0.05 mm’,

and for the Hall effect 5 ^x 0.3 ^x 0.1 mm’. All con- tacts were attached with silver paint Dupont ^4929.

The irradiation was done at room temperature with a copper tube X-ray generator. ^The correspon- dance between the defect concentrations and the X- ray doses was obtained using the results of the refe-

rence [7].

Figure ^{1 shows the} temperature dependence ^{of the}

D.C. conductivity plotted ^versus I/T for small defect concentrations. Increasing the dose of the defects in the Peierls insulating state, ^the conductivity ^increases.

For example ^the sample ^{with 0.35} % of defects con-

ducts at the liquid ^helium temperature three orders of magnitude better than the pure sample. ^{The insert}

in figure ¹ shows the logarithmic ^derivative d(log a)/d(I / T) between 10 K and 70 K. The transition at 42 K in the _pure sample ^{shifts to lower} temperatures and broadens for increasing ^defect

concentrations. For 0.2 %, ^{there is no more} sign ^of phase transition.

The results for the Hall coefficient measurements

are presented ^on figure ^{2. In the} pure sample RH

is positive ^{and small at} high temperatures and becomes

negative ^and large ^{below the} phase transition. The number of carriers involved in the transport ^mecha- nisms is of the order of 1021/cm3 ^{at room} temperature and of the order of 1017/cm3 ^{at 20 K when a} gap is

opened ^on the Fermi surface. In irradiated samples

the sign ^of RH ^{seems to} change ^at high temperatures,

even for the less irradiated samples (Fig. 3). ^{We believe}

Fig. ^1.

^D.C. conductivity ^of TMTSF-DMTCNQ ^versus 1/T ^at various level of defect concentration. Insert shows

d(log a)/d(I/T) - ^the phase transition is shifted 110 K per 1 % of defect concentration.

Fig. ^2.

Low-temperature part of the Hall coefficient of

TMTSF-DMTCNQ ^for pure and irradiated samples.

this effect to be due to a surface artefact and not to

a genuine ^bulk property ^{of the} sample. ^The crystals

were irradiated in air and we know that the reactivity

of the surface to oxygen and water is increased stron-

gly under irradiation. A thin insulating layer ^can

Fig. ^3.

High-temperature part of the Hall coefficient of

TMTSF-DMTCNQ.

add a non negligible negative ^{value of} RH ^{to the small} positive value of the sample ^itself. Electrically ^active impurities ^on the surfaces of organic [15] ^and inorga-

nic [16] semi-conductors ^are known to produce

similar anomalies. Below the phase transitions when the bulk Hall constant is large this kind of artefact,

of course, disappears.

For the defect concentration of 0.2 % ^{we have esti-}

mated the _upper limit for the Hall constant at all

temperatures : RH 8 x lO- 3 CM3 /C ^{at room}

temperature and Rn ! [ ^{1.6 x 10-1} cm3/C ^{at 2.2 K.}

In comparison ^{with the curve} of the Hall constant for the _pure sample ^the 0.2 % ^{curve is} completely

flat in the 2.2 K to 300 K range :

On the sample ^{with 1.0} % of defects two points ^were measured; ^{at room} temperature ^{and at} liquid ^helium temperature the Hall coefficients are less than +8 ^x 10- 3 cm3/C ^and ± 2.8 x 10-1 cm3/C ^res- pectively. The metallic carrier density ^is present, ^but the conductivity ^{did not} increase futher for these defect concentrations because of the reduced mobility

of the carriers.

On figure 4, the thermoelectric _power of a pure and and irradiated sample (0.2 jo) ^are plotted ^{versus the} temperature. ^{The curve} for the pure sample ^{is in} good

agreement with the results published previously by

Jacobsen et ale in reference [10]. ^{The curve for the}

irradiated sample ^{reminds us} the results of Andrieux et

al. [18] ^when the metallic state of TMTSF-DMTCNQ

was stabilized by ^a pressure of 12 kbars.

3. Discussion.

The results of figures 1, ^{2 and 4} clearly demonstrate that in contrast to the _pure

crystals, ^{the low} temperature phases of irradiated

TMTSF-DMTCNQ crystals ^exhibit high ^conducti- vities, small Hall coefficients and small thermo- electric _power. The existence of such a metallic behaviour is confirmed by two recent experiments by Forr6 et al. [19] which will be published ^{soon in a}

Fig. ^4.

Thermoelectric power of pure and irradiated

TMTSF-DMTCNQ.

paper ^on the magnetic properties ^of (TMTSF)2PF6

and TMTSF-DMTCNQ : The E.P.R. linewidth of 0.2 jo irradiated crystal has been shown to be nearly

linear from 4 K to 300 K, ^and the g ^factor nearly

constant in the whole temperature range.

The reasons for the stabilization of such a low temperature ^{metallic state} will be discussed further on, because the first part of this discussion consists in a few remarks about the transport properties ^{of the}

pure sample.

The Hall constant of the pure sample ^{at room} temperature ^was + (4.8 ± 0.9).10-3 cm3/C. Using

the formula for the low field limit for a single tight binding ^{band Hall constant} [8]

(where b is the lattice constant along ^the b-axis)

one obtains n

1.0 ^x 1021 cm- 3 density ^{of holes}

which is in a good agreement ^{with the} calculated value (8.8 ^{x 102°} CM - 3) ^{for 0.5} change ^transfer [9].

Utilizability ^{of the} single band formula shows that the contribution of the DMTCNQ chain is less than the experimental uncertainty, ^and confirms that the transport ^{is dominated} by ^the TMTSF chain in accordance with the conclusionsof Jacobsen et al. [10].

Starting ^{near 130 K} the value of RH ^{increases to}

+ (2.9 + 1.4). 10-2 cm3/C ^{at 51} K. Similar behaviour

was observed in the Hall constants of HMTSF-

TCNQ [10] ^and TTF-TCNQ [11]. Cooper et ^al.

have connected this effect with the increasing intensity

of the 2 kF ^{lines seen} by X-rays ^and interpreted ^{it as}

the sign of fluctuations towards the Peierls semi-

980

conducting ^{state. In} TMTSF-DMTCNQ the influence of these ^« resistive fluctuations » due to a gap precur-

sor in the density ^{of states can be seen in the Hall}

coefficient as well as the thermopower, D.C. resisti-

vity ^{and E.P.R. linewidth} [12].

The opening ^{of a} gap ^at 42 K is driven by ^{the TMTSF}

stack [13]. ^The gap ^on the DMTCNQ stack is smaller and the electrons of this stack dominate the trans-

port properties ^{at low} temperatures. This is the rea- son why ^the sign of the Hall coefficient changes ^near

the transition.

Let us turn back to the irradiated samples ^{and to}

the stabilization of the metallic state by ^{a weak}

disorder. The disappearance ^{of the} charge density

wave transition produces ^a high ^carrier density ^low temperature phase which exhibits all the usual metal- lic features except ^the positive slope dp/dT ^{of the} resistivity ^versus temperature ^curve. This contradic- tion is only apparent because the negative dp/dT

value is related to the one-dimensional character of this disordered phase. The metal obtained by ^irra-

diation is a metal of segments interrupted by ^the

defects and not a metal of infinite chains [6]. ^The layer-

ed compound ¹ T -TaS2 is another charge density

waves insulator in which irradiation defects have been found to produce ^{the same} strong decrease of the Hall coefficient as for TMTSF-DMTCNQ [20]. ^{But in}

the two-dimensional ^case positive values of dp/dT

have been found in the metallic disordered state [20].

There are more possibilities ^to escape defect poten- tials in two dimensions than in one.

The reasons of the smearing ^{of the} phase transitions out of the resistivity ^versus temperature ^{curves and} the stabilization of the metallic state have to be found in the pinning ^{of the} charge density ^waves by ^irradia-

tion defects. At temperatures ^{below the} phase ^tran-

sitions the charge density waves, in the _pure sample,

are ordered in three dimensions and pinned ^{on the}

lattice. The radiation induced defects fix the phases

of the charge density ^{waves in} many random points

and disturb the coherences needed for the achievement of the phase transitions. These effects have been discussed in reference [5] ^{and more} recently ^and extensively in reference [17].

After three hours of irradiation, ^{in the} crystal containing ^a concentration of 0.025 jo ^{the Hall}

constant is decreased significantly (of ^{course, the} experiment ^{at such a} low defects concentration level has been done on the same crystal ^{before and after} irradiation). This should ^mean that the random

potentials of the defects reduce the _average gap.

More precisely, ^{the more or less constant} gaps through

all the _pure samples (one gap per stack) ^are replaced by ^a distribution of _gaps in the irradiated sample.

A crude approximation ^{of such a} distribution of _gaps consists in assuming that there is no gap ^at all in a

given volume around each defect and that elsewhere the _gap is unchanged. This model is non physical ⁱⁿ

its details; nevertheless it allows a determination of

the _range of a defect. An effective medium approxi-

mation of such a distribution of metallic volumes in

an insulating matrix has been proposed ^{in reference} [5].

Applied ^to TMTSF-DMTCNQ ^it indicates that the gap has been significantly changed ^{in a volume of}

700 molecules surrounding each defect.

Even at the _very low concentration of 0.025 %

gaps ^seems also to have an effect on the mobility ^as

shown on figure ^5. The calculated value of y’ = ^{R H} alb is presented ^{in the} pure sample and the 0.025 % ^irra-

diated one. We have no serious explanation ^{of the} changes ^{in this effective} mobility ^{in such an inhomo-}

geneous sample ^{but we} think that the result is worthy

to notice.

Fig. ^5.

^Hall mobility (,uH

^{R H} Qb) in the semiconduct-

ing phase ^of TMTSF-DMTCNQ ^for 0 % ^and 0.025 % ^of

defect concentrations.

At the end of the present discussion, ^{it is} interesting

to compare the irradiation results presented ^here

above with the results of Jacobsen et ale concerning

the disordered material obtained by alloying ^the DMTCNQ stack with MeTCNQ. ^The conductivity

curve of TMTSF-DMTCNQO.75-MeTCNQO.25 ^of

reference [10] ^is very similar to the curve 0.05 % ^of figure ^{1. Thus an} irradiation defect is 500 times more

efficient in changing ^{the low} temperature ^conducti- vity ^{than a} dopant molecule in the alloy ^mentioned

above. This is a good illustration of the difference between the charge density ^waves weak >> and

« strong » pinnings. The thermoelectric _power of

figure ^{4 is} very different than the thermopower ^{of the} alloy published in reference [10] (Fig. 9). ^{The main}

effect of chemical impurities ^{is not to} change ^{the low} temperature ^Peierls gap but to change ^{the band} filling

and probably ^{to introduce} impurity ^{states in the}

Peierls _gaps.

4. Conclusion.

The first attempt ^to stabilize the metallic state in an organic ^{conductor was the} pres-

sure experiment ^on HMTSF-TCNQ [2] performed

in Orsay ^five years ago. The semi-metallic character

of this compound ^{was confirmed} by the Hall effect measurements of reference [8]. ^One year later, ^with

another batch of the same compound, ^{it was} impos-

sible to reproduce ^the experiment of reference [2] :

the crystals ^{revealed to} be less disordered and it

was impossible ^{to remove} the metal to insulator transition completely, ^{even with} pressures much

higher ^{than the 4} kbar needed in reference [2].

Crystals of both batches were irradiated at low temperatures (21 K). The « semi-metallic » batch

resistivity increased with irradiation [5] ^{while the} resistivity ^{of the «} insulating » ^{batch decreased} firstly

towards a semi-metallic state stabilized by ^disorder

and increased with further irradiation [21].

These two kind of experiment clearly ^{show that}

pressure and disorder can add their efforts to stabilize the metallic state.

In a recent paper, about metallic state and _super-

conductivity ⁱⁿ (TMTSF)2C104, Bechgaard et ^al. [22]

observed that the perchlorate ions of this radical ion salt are disordered and the _oxygen 6ccupies eight nearly equivalent positions ^{at random.} Preliminary, low-temperature X-ray ^diffuse scattering ^indicated

that no ordering ^{occurs at} least down to 15 K. This is probably ^{another case where} disorder _prevents

the S.D.W. insulating ^state and stabilizes the metal- lic state.

References

[1] WEGER, M., ^J. Physique Colloq. ³⁹ (1978) ^C6-1456.

[2] COOPER, ^{J. R.,} WEGER, M., JÉROME, D., LEFUR, D., BECHGAARD, K., BLOCH, ^A. N., CONAN, D.O., Solid State Commun. 19 (1976) ^749.

[3] ANDRIEUX, A., DUROURE, C., JÉROME, D., BECHGAARD, K., ^J. Physique ^{Lett. 40} (1979) ^L-381.

[4] PARKIN, S. S., RIBAULT, M., JÉROME, D., BECHGAARD, K., ^{Submitted to J.} Phys. ^{C :} Solid State Phys.

[5] ZUPPIROLI, L., BOUFFARD, S., ^J. Physique ⁴¹ (1980)

291.

[6] ZUPPIROLI, L., BOUFFARD, S., BECHGAARD, K., HILTY, C., MAYER, ^C. W., Phys. ^{Rev. B22} (1980) ^6035.

[7] MIHALY, G., BOUFFARD, S., ZUPPIROLI, L., ^BECH-

GAARD, K., ^J. Physique ⁴¹ (1980) ^1495.

[8] COOPER, ^J. R., WEGER, M., DELPLANQUE, G., JÉROME, D., BECHGAARD, K., ^The organic conductors and semiconductors. Proceedings of the international

conference, Siofok, Hungary, 1976, p. 363.

[9] POUGET, ^J. P., COMES, R., BECHGAARD, K., Proceedings of the ^{nato advanced} study institute, Tomar, ^Por- tugal, 1979, p. 113.

[10] JACOBSEN, ^C. S., MORTENSEN, K., ANDERSEN, ^J. R., BECHGAARD, K., Phys. ^{Rev. B 18} (1978) ^905.

[11] COOPER, ^J. R., MILJAK, M., DELPLANQUE, G., JÉROME, D., WEGER, M., FABRE, ^J. M., GIRAL, L., ^J.

Physique ³⁸ (1977) ^1087.

[12] ZUPPIROLI, L., DELHAES, P., AMIELL, J., ^{To be} published

in J. Physique, ^1982.

[13] TOMKIEWICZ, Y., ANDERSEN, ^J. R., TARANKO, ^A. R., Phys. ^{Rev. B 17} (1978) ^1579.

[14] ^{ZIMAN, J. M.,} Electrons and phonons (Calderon Press) 1962, p. 489.

[15] HEILMEIER, ^C. H., HARRISON, ^S. E., Phys. ^{Rev. 132} (1963) ^2010.

[16] RUPRECHT, M., ^Z. Naturforsch. ^13a (1958) ^1094.

[17] ZUPPIROLI, L., MUTKA, H., BOUFFARD, S., Proceedings

of the Conference on Low Dimensional Conduc- tors, Boulder (1981). ^Mol. Cryst. Liq. Crystal- logr. (82).

[18] ANDRIEUX, A., CHAIKIN, ^P. M., DUROURE, D., JÉROME, D., WEYL, C., ^J. Physique ⁴⁰ (1979)

1199.

[19] ^{FORRO, L. and} BEUNEU, F., ^{To be} published.

[20] MUTKA, H., ZUPPIROLI, L., MOLINIE, P., BOURGOIN, J. C., Phys. ^{Rev. B23} (1981) ^5030.

[21] BOUFFARD, S., Rapport CEA-R-5015 (1979).

[22] BECHGAARD, K., CARNEIRO, K., RASMUSSEN, ^F. B., OLSEN, M., RINDORF, G., JACOBSEN, ^C. S., ^PEDER-

SEN, H. J., SCOTT, ^J. C., J. Am. Chem. Soc. 103

(1981) ^2440.