Conduction electron spin resonance measurements on TTF-TNNQ and (TMTTF)2BF4 under hydrostatic pressure

HAL Id: jpa-00210455

https://hal.archives-ouvertes.fr/jpa-00210455

Submitted on 1 Jan 1987

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.

Conduction electron spin resonance measurements on TTF-TNNQ and (TMTTF)2BF4 under hydrostatic

pressure

L. Forró, J.R. Cooper, G. Sekretarczyk, M. Krupski, K. Kamarás

To cite this version:

L. Forró, J.R. Cooper, G. Sekretarczyk, M. Krupski, K. Kamarás. Conduction electron spin resonance

measurements on TTF-TNNQ and (TMTTF)2BF4 under hydrostatic pressure. Journal de Physique,

1987, 48 (3), pp.413-418. �10.1051/jphys:01987004803041300�. �jpa-00210455�

Conduction electron spin ^resonance measurements ^on TTF-TCNQ ^and

(TMTTF)2BF4 ^under hydrostatic pressure

L. Forró (1,2), ^{J. R.} Cooper (1), ^G. Sekretarczyk (3), ^M. Krupski (3) and K. Kamarás (4) (1) Institute of Physics ^{of the} University, Bijenicka 46, ^P.O.B. 304, ⁴¹⁰⁰¹ Zagreb, Yougoslavia

(2) CEA/IRDI/DMECN/DTech/SESI, ^{C.E.N. de} Fontenay-aux-Roses, ^{B.P. n°} 6, ⁹²²⁶⁵ Fontenay-aux-Roses Cedex, ^France

(3) Institute of Molecular Physics ^Polish Academy ^of Sciences, Smoluchowskiego ^{17/19, 60-173} Poznan, Poland

(4) Central Research Institute for Physics, ^P.O.B. 49, ^H-1515 Budapest, Hungary (Requ ^{le 3} septembre 1986, accept6 le 13 novembre 1986)

Résumé. 2014 Nous présentons ^{des mesures} de résonance paramagnétique ^sur les électrons de conduction de deux conducteurs organiques : ^un système bi-chaînes tétrathiofulvalène-tétracyanoquinodiméthane (TTF-TCNQ) ^{et une}

mono-chaîne bis tétraméthyl-tétrathiofulvalène borontétrafluoride ((TMTTF)2BF4). Dans les deux cas la raie de résonance s’élargit ^{fortement sous} l’influence de la pression ; ^{ceci est en} contradiction avec le mécanisme d’Elliott.

Quant ^au facteur g, il ^ne dépend pas de la pression, alors que la susceptibilité paramétrique ^{diminue à} température

ordinaire de - 8 ± 1 %/kbar pour TTF-TCNQ ^et de - 3 ± 1 %/kbar pour (TMTTF2BF4.)

Abstract.

Conduction electron spin ^resonance (CESR) measurements under hydrostatic pressure ^{on two} chain

organic ^conductor tetrathiofulvalene-tetracyanoquinodimethane (TTF-TCNQ) ^and single ^chain organic ^conductor

bis tetramethyl-tetrathiofulvalene borontetrafluorid ((TMTTF)2BF4) ^are presented. ^{In both cases the CESR}

linewidth increases strongly with pressure in contradiction with the Elliott-mechanism for spin relaxation. The g- factor is pressure independent ^{for both} compounds ^{while the} spin susceptibility ^decreases by - 8 ± 1 %/kbar for TTF- TCNQ and - 3 ± 1 %/kbar for (TMTTF)2BF4 ^{at room} temperature.

Classification

Physics ^Abstracts

72.30P

72.15N

1. Introduction.

Conduction electron spin ^resonance (CESR) ^measure-

ments at ambient pressure have been carried ^{out on a} number of effectively one-dimensional (lD) organic conductors, including ^TTF-TCNQ [1], ^TMTTF-TCNQ [2], (TMTSF)2PF6 [3], TMTSF-DMTCNQ [1] ^etc...

and have been very useful in elucidating their unusual

magnetic properties. ^It is indeed peculiar ^that organic

metals with scattering ^{times as low as} 10-14-10-15 s do

give ^an observable CESR signal ^{even at room} tempera-

ture. Note that very few ordinary isotropic ^{metals show}

CESR and then usually only ^at very low temperatures.

The ESR linewidth (AH) reflects the spin ^lattice

relaxation rate and in ordinary metals it is usually

determined by ^the spin-phonon interaction. However in many organic metals, mainly ^{two chain} compounds,

4H has an unusual temperature dependence : namely ^it

increases with decreasing temperature [1]. ^{Thus the} proportionality between dH and the phonon population

is not obeyed ^{in two chain} systems ^like TTF-TCNQ

although ^{it is} roughly followed in single ^chain systems like (TMTTF)2BF4 [4].

The spin susceptibility ^can also be obtained from the

intensity ^{of the CESR} signal. One of the most exciting things ^{in the} high temperature ^metallic phase ^of organic

conductors is the enhanced and temperature dependent spin susceptibility (Xs ). ^For example ^{in TTF-TCNQ at}

room temperature XS is enhanced by ^a factor of 3 to 5 times over the value calculated using ^{a bandwidth}

deduced from optical ^or thermoelectric power measurements. Furthermore this paramagnetic suscep-

tibility ^{which is} temperature independent ⁱⁿ ordinary

metals falls by ^a factor of 2.5 on cooling ^{down to 60 K} [5].

These strange features of AH and X ^{are believed to}

arise from the 1D nature of the electronic band structure but after a lot of experimental ^and theoretical work a generally accepted explanation ^{is still} missing.

Attempts ^{have been made} to account for the tempera-

ture dependence ^{of AH in terms of} interchain coupling [6] ^{or as a precursor to} the Peierls instability [1J. ^The

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphys:01987004803041300

414

two mechanisms most often cited for the T dependence

and enhancement of ys ^are a) electron-electron interac- tions which enhance XS and introduce ^a log ^T depen-

dence [7] ^and b) polaron ^{formation which} gives ^{a T} dependent narrowing of the bandwidth [8].

By application ^of high pressure ^one hopes ^to get

insight into these mechanisms since it will produce

controlled changes in the onchain and interchain trans- fer integrals ti ^and t, . It will also increase the mean free

path (decrease ^the scattering time) and alter the charge

transfer. Thus the screening of the Coulomb interac- tions may be altered. Furthermore as we have been interested in the different ESR _responses for two chain and single ^chain compounds ^{for some time, we have}

made a comparative study ^{of TTF-TCNQ and} (TMTTF)2BF4 ^under pressure. TTF-TCNQ ^has already

been measured by Berthier et al. [2]. However in their low field measurements at 40 MHz the linewidth was

comparable ^{with the} applied magnetic field, ^{no infor-}

mation could be obtained about the g-factor, ^{and there}

was also a possibility ^{that the ESR} intensity ^{could have}

been reduced because of changes in the skin depth

under pressure, giving ^an overestimate of the pressure

dependence of Xs.

Before discussing ^the present high field ESR data for

single crystals ^{let us} briefly ^{mention some} effects of pressure ^on other physical properties ^{of both com-} pounds.

TTF-TCNQ is one of the best studied organic

conductors under hydrostatic pressure. The compress-

ibility ^{of the a,} b, ^{c lattice} parameters ^{is 0.26} %, ^{0.43 %}

and 0.32 % _per kilobar respectively [9]. ^{The lattice} parameters ^have strong temperature dependence, too, and it has been emphasized that thermal contraction of the lattice between 290-65 K has approximately ^the

same effect as the application ^{of 5} kbars pressure ^at

room temperature. Calculating ^the change ⁱⁿ tj with ^b-

axis spacing [11], ^{the obtained} increase in tl was ⁱⁿ good agreement ^{with the} experimental ^{results on the}

pressure dependence ^{of the} plasma frequency, namely

d In tl/dp ^{= 2 %/kbar} [12]. ^{But the} strong increase of the conductivity ^{d In} Q /dp

^{28 %/kbar can not be} explained ^{with such a} change ^{in the one electron}

bandwidth [10]. Generally, ^{all the} Fermi-level _proper- ties exhibit anomalously large pressure dependence ⁱⁿ TTF-TCNQ, ^for example proton relaxation rate 11 T

(- ²³ %/kbar), spin susceptibility (- ⁸ %/kbar) [2], ^etc.

(TMITF)2BF4 ^{has not been so} thoroughly ^investi- gated ^under pressure. As far as we know, ^{there are no} available compressibility ^{data. A} priori this material is

expected ^{to be less} compressible ^than TTF-TCNQ because with full charge transfer from the TMTTF molecules to the anions it resembles more an ionic

crystal. However, thermal contraction and compress-

ibility measurements on a similar compound (TMTSF)2PF6 [13] give similar values to those of TTF- TCNQ. ^{The first} experiments ^on (TMTTF)2BF4 ^under hydrostatic pressure have been performed ⁱⁿ Japan ^and

show a similar strong pressure dependence ^{of the}

electrical conductivity ^{to that for} TTF-TCNQ, namely

d In uldp ^{= 19 % kbar} [14].

2. Experimental ^details.

We have performed ^our pressure and temperature dependent ^ESR measurements at X-band using ^a high

pressure apparatus, which works at pressures up ^to 5 kbars and at temperatures ^{from 80 to 350 K} [15]. ^The

apparatus ^{contains a} cylindrical ^{corundum resonator} directly coupled ^{to the} waveguide by ^a matching

corundum wedge. The pressure chamber in which the resonator is placed ^{is made of} nonmagnetic beryllium

bronze. The temperature ^{of the} system ^{was controlled} by ^a nitrogen vapour flux pumped through ^{a heat} exchanger disposed outside the _pressure chamber. In this work the pressure medium ^was petroleum ^ether.

Single crystals ^{of TTF-TCNQ and} (TNMF)2BF4

were placed ^{in the resonator in an} orientation which minimizes the skin effect and gives Lorentzian line-

shapes. ^Possible changes ^{in the} Q-factor ^{of the re-}

sonator in temperature and pressure ^were monitored

by ^the signal ^{of a} copper sulfate crystal in the orienta- tion (g -- 2.50) ^{in which it does not} disturb the sample signal (g = 2.01). The static magnetic ^{field and mic-}

rowave frequency ^{were recorded} continuously during

the experiment.

3. Results and discussion.

The first ESR parameter ^whose change ^{is obvious}

under pressure is the linewidth. Figures ^{1 and 2} display

the temperature variation of AH for TTF-TCNQ and

Fig. ^1.

CESR linewidth versus temperature ^{for a} single crystal ^{of TTF-TCNQ at 1 bar} (0) and 5 kbars (x) ^{in the}

orientation Hdc ^{11 c *.}

Fig. ^2.

CESR linewidth versus temperature ^{for a} single crystal ^of (TMTTF)2BF4 ^{at 1 bar} (0) and 5 kbars (x) ^{in the}

orientation Hdc II b * .

(TMTTF)2BF4 ^at ambient pressure and ^at 5 kbars

respectively. ^{In both cases AH increases} linearly ^with

pressure. For TTF-TCNQ AH increases by factor of 2.7 for 5 kbars and this factor is constant in the measured temperature range 80-310 K. This relative increase in AH is in good agreement ^{with the earlier low field} studies.

The linewidth in TTF-TCNQ ^at ambient _pressure does not satisfy ^the Elliott relation [16] ^for spin

relaxation which is often used for isotropic ^{metals :}

where Ag is the deviation of the g-factor ^{from the free}

electron value and Tf is the electron scattering ^rate along the chains measured in resistivity. Namely ^the

observed linewidth AH is a factor of 100 less than that

expected ^from equation (1) [17]. Furthermore, ^{as the} resistivity ^{decreases on} cooling ^{the linewidth should}

decrease too. However the actual temperature depen-

dence of AH is the opposite (Fig. 1). ^{It is} presently

believed that the very ^narrow linewidth (and ^{also its}

anomalous temperature dependence) ^{is the} consequ-

ence of the quasi-one-dimensional electronic structure.

In a strictly ^1D system ^{the Fermi} surface consists of two

points ^at kF ^{and -} kF ^{and both the forward} (q = 0) ^and

backward (q- ² kF ) scatterings ^{do not} give spin ^relax-

ation [18]. ^{Hence to account for the observed finite}

linewidth one has to include 3D effects like interchain

scattering, because every process which makes the system ^{resemble an} isotropic metal will broaden the line. In the first attempt ^{at a} mathematical formulation

Weger ^{included the} TII IT .1. ^factor in the Elliott formula

[6]. This is the ratio of onchain scattering ^{time and}

interchain tunnelling time and is a measure of the

dimensionality ^{of the} system :

T11 ^is given by ^the golden ^{rule of} perturbation theory

to be T-’ = ² iTIh 1 T .L IZ ^{Ti. Inserting this into}

equation (2) ^one gets for the linewidth :

where (1. ^{is the} interchain overlap integral.

It is tempting ^to attribute the spectacular increase of AH with pressure ⁱⁿ figure ^{1 to} the increased transverse transfer integral ^{and onchain} scattering ^time. Supposing

a pressure dependence similar for (1. ^{as for} (I, ^{it turns}

out that the decisive factor is Ti. This is not surprising

since in a single particle picture a -- w p 2 ^{r and from}

optical measurements it is known that the change ⁱⁿ plasma frequency ^{is an order of} magnitude less than the

change ⁱⁿ conductivity with pressure [10]. ^The lengthen- ing of Till by pressure changes both cr and AH by nearly

the same factor. However the origin ^{of such a} large sensitivity of the electron lifetime to pressure remains a

somewhat controversial question.

There is a striking difference in the temperature dependence ^{of d In} uldp ^{and d In} åBldp. ^{The latter is}

constant at 33 %/kbar (Fig. 1) in the measured tempera-

ture range, while the former falls from 28 % kbar at room temperature ^{to 10 %/kbar at} 100 K. The latter temperature dependence ^was attributed to the increas-

ing ^{role of} the Frohlich collective mode in the conduc-

tivity ^{as the} temperature is lowered. The extra conduc-

tivity depends very weakly ^on pressure in the incom- mensurate regime, ^so the overall pressure effect is reduced. Similarly ^{one can consider} different contribu- tions to the linewidth :

The first and second terms are the usual single particle spin relaxation by phonons ^{for momentum} changes

near to 0 and 2 kF. ^{The third term could enter as the}

temperature is lowered and collective Frohlich modes build up. However nothing ^is known about the effect of such modes on spin relaxation but if anything they

should reduce the spin flip ^rate for those electrons which are condensed in the collective state. Whether these long ^life-time fluctuating ^CDWs give ^{a contri-}

bution to the spin flip ^{rate of the} uncondensed electrons

by scattering ^{them is} questionable. However with the appearance of CDWs, ^the population of the soft 2 kF phonon modes increases thus increasing ^the spin flip scattering ^{rate, and will increase} the second term in

equation (4). Irradiation experiments have shown that there is some extra contribution to the linewidth as the temperature is lowered towards the phase ^transition [19], and that this contribution is as sensitive to defects

as the Frohlich conductivity. ^The temperature ^inde-

416

pendence ^{of AH} (5 kbars)/AH (1 bar)

^{2.7 means that}

any relaxation mechanism associated with the CDW fluctuations is either negligible ^{or has the same} pressure

dependence ^{as the} single particle mechanism which is dominant near room temperature.

At first sight ^the temperature dependence ^{of AH} (Fig. 2) ^of (TMTTF)2BF4 ^{suggests that the} original

Elliott formula (1) applies ^as is stated in the literature for most of the single ^chain compounds. However there is again ^a problem ⁱⁿ understanding why ^{the line is so}

narrow, i.e. the magnitude ^{of AH is not} consistent with

equation (1). From the pressure dependence ^{of AH it is}

obvious that the Tr ^{term alone cannot account for the}

linebroadening, ^because Tp 1, ^{as measured} in the resis-

tivity, decreases with pressure, ^and consequently ^AH

should decrease. Only ^an interchain hopping ^rate increasing ^with pressure ^can broaden the line as obser- ved. So we propose ^an empirical ^{formula for AH in} single ^chain compounds :

In (TMTTF)2BF4 the pressure dependence ^{of these} scattering ^rates compete with each other : the latter wins and AH increases with pressure, although ^this

increase is only ^{45 % for 5} kbars, ^{i.e. a factor} of 3.5 less than for TTF-TCNQ.

One could ask what is the role of the pressure

dependence ^of Ag ^{in the} linebroadening. Figure ³

Fig. ^3.

g-factor ^versus temperature ^for single crystals ^of

TTF-TCNQ and (TMTTF)2BF4 ^{at 1 bar} (0) and 5 kbars (x).

(The orientations in respect ^{to static} magnetic ^{field are the}

same as for Figs. 1 ^and 2.)

shows that Ag is pressure independent ^{for both com-} pounds. ^For (TMTTF)2BF4 ^{this is} comprehensible since g ^{is a molecular} property, ^and contracting ^the

lattice one does not expect ^to change ^molecular quan- tities. In the case of TTF-TCNQ the constancy of g

means more, since the molecular g-factors ^{of each}

chain are mixed together ^{with the} respective spin susceptibilities giving ^a common g value [20] :

It can be concluded that at room temperature ^the spin susceptibility of the donor (X p ) ^and acceptor ^chain

(X A ) changes ^{at the same rate} with pressure because the total g-factor is pressure independent. ^However

below 200 K at 5 kbars g ^{seems to} be below the ambient pressure values. Within the strong coupling

limit corresponding ^to equation (6) ^this implies ^an

increased contribution of the TCNQ chain in respect ^to the TTF chain in the total susceptibility (XT). ^{From the}

NMR measurements of Takahashi et al. [21] it is known that _XA (i.e. XTCNQ) ^{changes slope below 200 K, it}

decreases faster than at higher temperatures. ^This

change ^is usually ^{ascribed to a} precursor effect, namely

1D fluctuations towards the semi-conducting phase.

Susceptibility measurements of KCP have proved ^ex- perimentally that these fluctuations can be suppressed by pressure [22] ; ^the temperature dependent suscepti- bility ^at normal pressure becomes nearly temperature independent ^{at a} preessure of 20 kbars. From the g- factor measurements of TTF-TCNQ such a tendency

can be postulated ^{for the TCNQ} chain, ^{but measure-}

ments at higher pressures ^are needed to confirm this.

It was demonstrated above that in TTF-TCNQ the

spin susceptibility of donor and acceptor chains vary in the same way under pressure ^{at room} temperature. ^In

figure ^{4 are shown the} changes ^{in total} spin susceptibili- ty ^{for both} compounds ^{at room} temperature. ^For TTF-TCNQ d In x Idp = - 8 ± 1 % ^{kbar and for}

Fig. ^4.

Normalized spin susceptibility ^versus pressure for

(TMTTF)2BF4 ^{and TTF-TCNQ at ambient} temperature.

Different symbols ^{mark different runs.}

(TMT’rF)2BF4 ^{d In} X Idp

^{3 ± 1} %/kbar. ^{It is im-} mediately clear that an increase of 2 % in tl ^{in the}

formula for the spin susceptibility ^of noninteracting

electrons cannot account for the change ^{in X.}

The motionally ^band narrowing ^model requires ^the phonon frequency w ^{to be} comparable ^{to the bandwidth}

4 ti, ^otherwise Migdal’s ^theorem holds and the phonons

have no effect on y [23]. ^One expects ^the phonon frequencies (external ^and internal) ^{to be similar in}

TTF-TCNQ and (TMTTF)2BF4, ^{as well as their}

pressure dependence. However the difference in d In X Idp ^{of a} factor of 3 between the two compounds

tends to go against ^the explanation of the pressure

dependence of X within this model.

Let us consider the changes in X under pressure

expected if Coulomb correlations are important.

TTF-TCNQ and (TMTTF)2BF4 ^{have the same room}

temperature susceptibilities ^of 6.0 x 10-4 emu/mole.

Calculating ^{the Pauli} susceptibility ^{with a} total band- width 4 tj ^{of 0.5 eV} (which ^{is a} reasonable parameter for these compounds), ^one gets ^a value three times smaller than the measured one [5]. ^{As is shown in} figure 4 this enhancement is strongly pressure depen-

dent for TTF-TCNQ. Furthermore it is tempting ^to

attribute the strong temperature dependence of x (see Fig. 5) ^{to the} temperature dependence of the enhance- ment mechanism. Note in figure 5 ^{that the} pressure

dependence of X ^is slightly temperature dependent ⁱⁿ TTF-TCNQ ; ^{from - 8 ± 1 %/kbar at room} tempera-

ture it changes ^{to - 6 ± 2 %/kbar at 80 K.}

When treating the electron-electron interactions it is usual to use the Fourier transform of the Coulomb interactions and to distinguish between the forward

scattering (g2 ) ^which corresponds ^{to small momentum}

Fig. ^5.

Temperature dependence ^{of the} spin susceptibility

of TTF-TCNQ ^at 1 bar (0) and 5 kbars (x).

transfer q = 0, and the backward scattering (gl ) ^which displays ^the large ^momentum transfer with 2 kF. ^{In real}

space 92 corresponds ^{to a} long range interaction, ^which

is normally screened, while g1 is defined by ^{the short}

range onsite Coulomb interaction U. Calculation of the

high temperature magnetic susceptibility ^{in the} parquet

approximation [7] shows that:

where gl(T) = gll(l + g, In EFIT), g, - nF U, ^and

nF and X p ^{are the} density ^{of states} and the Pauli

susceptibility ^{for the} noninteracting electron gas.

Equation (7) holds for weak interactions, ^{that is}

U/4 t ..: ^1.

Applying these results to the spin susceptibility ^of

TTF-TCNQ separated ^{for each} chain, ^{it turns out that}

91 TCNQ ::- 61, TTF

^{With such} parameters ^the temperature

dependence and the enhancement of TTF-TCNQ come

mainly ^{from the TCNQ chain.}

For the enhancement of the factor of 3 in X ^{at room} temperatures ^one needs g, -- 0.5. This value does not

really fall within the weak coupling limit, nevertheless

equation (7) ^can be useful for a qualitative description

of the experimental results. Reduction of 9 1 ^{to 0.3}

accounts well for the observed decrease of 40 % of X ^at

5 kbars at room temperature. ^As ti depends ^{little on}

pressure, this reduction in g1 ^{means an} increased

screening ^{of U} under pressure. There could be several

reasons for this :

a) the increase in the onchain screening ;

b) interchain screening by the free electrons on the

neighbouring chains ;

c) ^excitonic screening ^of highly polarizable ^molecules [24]. ^{This can be} strongly pressure dependent ^{since the} polaron binding energy EB ^varies roughly ^like R- 4 (R ^is

the interchain distance) ^{and reduces the effective}

interaction to U - 2 EB ;

d) ^the screening ^of the short range interaction ^is

especially ^{sensitive to the} charge ^transfer (p) [25]. ^The screening is least effective near p

0.5 and the most effective near p

0.75.For TTF-TCNQ low tempera-

ture measurements have shown that _p changes ^from

0.59 at normal pressure ^to 0.615 at 5 kbars [26] ^which

should considerably ^{reduce U.}

The decreased value of g1 under pressure qualitat- ively ^{accounts for a} temperature dependence of X ^at

5 kbars weaker than at ambient pressure. However it accounts only qualitatively ^since equation (7) ^{cannot fit}

the temperature ^variation of y ^{with a} temperature

independent g1 ^at any pressure. But ^{as was} mentioned in the introduction, ^a temperature drop ^{from room}

temperature ^{to 100 K contracts the lattice as much as a} 5 kbars _pressure at room temperature, and all the

screening mechanisms discussed above can be switched

on by cooling ^the sample. (Note ^{that in TTF-TCNQ at}

normal pressure the charge transfer of 0.55 at room

418

temperature increases below 220 K to 0.59 [27], ^which

may increase the efficiency ^{of the} screening).

Accepting ^the screening ^of the onsite Coulomb interaction as the source of the variation of X ^with

pressure, it is really simple ^to explain the much weaker decrease of y ^for (TMTTF)2BF4 ^{at 5 kbars} (Fig. 4). ^The charge ^{transfer cannot} help ⁱⁿ screening ^since p

0.5 is the most unfavorable, and it does not vary with pressure. Furthermore the reduction of U ^on the TMTTF molecule via screening by metallic electrons of

neighbouring ^{chains is less} effective since the anions do not form conducting ^chains. The poor polarizability ^of

the closed shell anions also disadvantages ^{the excitonic} screening mechanism. The weaker screening ^{of the}

onsite U in (TMTTF)2BF4 is also shown in the tempera-

ture dependence of X [4]. ^{Between 100} K and 300 K it increases only by ^{50 %} (in ^{TTF-TCNQ it is 100} %), although ^{the thermal variation of the lattice is} probably

similar to TTF-TCNQ.

In conclusion, ^the application ^of hydrostatic pressures

has confirmed the importance of the interchain hopping

in the spin relaxation processes. The increase in the linewidth has the same origin ^as the increased single particle conductivity: rl is ^a strong ^{function of the} pressure. The enhancement and unusual temperature dependence of y ^{is ascribed to Coulomb} interactions,

then from the present ^work it follows that U is strongly

pressure dependent ^{in TTF-TCNQ} and less in

(TMTTF)2BF4. ^{This indicates that} changes ⁱⁿ charge

transfer with pressure ^and interchain screening ^mechan-

isms _{may be} important ^for reducing ^{the onsite Coulomb}

interaction U.

Acknowledgments.

Useful discussions with S. Barisi£ , ^A. Bjelis, ^A. Janossy

and L. Zuppiroli ^are gratefully acknowledged. ^{One of}

us (L. F.) acknowledges ^the hospitality ^{of Prof.} A.

Graja and Prof. J. Stankowski during ^his stay ^{at the} Institute of Molecular Physics ^{in Poznan.}

References

[1] TOMKIEWICZ, Y., Phys. ^{Rev. B 19} (1979) ^4038.

[2] BERTHIER, C., COOPER, ^J. R., JEROME, D., SODA, G., WEYL, C., FABRE, J. M. and GIRAL, L., Mol. Cryst. Liq. Cryst. ³² (1976) ^267.

[3] ^{PEDERSEN, H.} J., SCOTT, ^{J. C. and} BECHGAARD, K., Solid State Commun. 35 (1980) ^207.

[4] DELHAES, P., COULON, C., AMIELL, J., FLANDROIS, S., TOREILLES, E., FABRE, ^{J. M. and} GIRAL, L., Mol. Cryst. Liq. Cryst. ⁵⁰ (1979) ^43.

[5] COOPER, ^J. R., MILJAK, ^{M. and} KORIN, B., ^Chemica Scripta ¹⁷ (1987) ^79.

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

[7] BARISIC, S., BJELIS, A., ^{in :} Theoretical Aspects of

Band Structures and Electronic Properties of

Pseudo-One-Dimensional Solids, ^ed. by Komimura, ^{D. Reidel} (D. Reidel Publ. Co., Dordrecht) 1985 and references therein.

[8] ENTIN-WOHLMAN, O., KAVEH, M., GUTFREUND, H., WEGER, ^{M. and} MOTT, ^N. F., ^Philos. Mag.

B 50 (1984) ^251.

[9] ^DEBRAY, D., MILLET, R., JEROME, D., BARISIC, S., GIRAL, ^{L. and} FABRE, ^J. M., ^J. Physique ^Lett.

38 (1977) ^L-227.

[10] COOPER, ^J. R., Phys. Rev. B 19 (1979) ^2404.

[11] BERLINSKY, A. J., CAROLAN, J. F. and WEILER, L.,

Solid State Commun. 15 (1974) ^795.

[12] WELBER, B., ENGLER, ^E. M., GRANT, ^{P. M. and} SEIDEN, Bull. Am. Phys. ^{Soc. 21} (1976) ^311.

[13] MOROSIN, B., SCHRIBER, ^J. E., GREENE, ^{R. L. and} ENGLER, ^E. M., Phys. Rev. B 26 (1982) ^2660.

[14] MURATA, K., private communication.

[15] STANKOWSKI, J., GALEZENSKI, A., KRUPSKI, M., WAPLAK, S., GIERSZAL, H., ^{Rev. Sci. In-}

strum. 47 (1976) ^128.

[16] ELLIOTT, ^R. J., Phys. ^{Rev. 89} (1953) ^689.

[17] FORRÓ, L., Ph. D. thesis 1985, unpublished.

[18] YAFFET, Y., Solid State Physics, ^{V. 14, ed.} by ^H.

Ehrenreich, ^F. Seitz, D. Turnbull (Academic,

N. Y.) ^1965.

[19] SANQUER, M., BOUFFARD, ^{S. and} FORRÓ, L., ^J.

Physique ⁴⁷ (1986) ^1035.

[20] TOMKIEWICZ, Y., TARANKO, ^A. R., ENGLER, ^E. M., Phys. Rev. Lett. 37 (1976) ^1705.

[21] TAKAHASHI, T., JEROME, D., MASIN, F., FABRE, ^J.

M., GIRAL, L., ^J. Phys. ^{C 17} (1984) ^3777.

[22] JEROME, D., SODA, G., COOPER, ^J. R., FABRE, ^J.

M. and GIRAL, L., Solid State Commun. 22

(1977) ^319.

[23] MIGDAL, ^A. B., Soviet Phys. ^{JETP 7} (1958) ^996.

[24] ^{LE BLANC, O. M., J. Chem.} Phys. ⁴² (1965) 4307:

[25] MAZUMDAR, ^{S. and} BLOCH, ^A. N., Phys. ^{Rev. Lett.}

50 (1983) ^207.

[26] MEGTERT,S., COMES, R., VETTIER, C., PYNN, ^R.

and GARITO, ^A. F., Solid State Commun. 31

(1979) ^977.

[27] POUGET, ^J. P., ^Chemica Scripta ¹⁷ (1981) ^85.