Un ordre à longue portée de type spin-Peierls et un état à ... Zeeman et l'apparition simultanée d'un nouveau mode à un champ de résonance (Hr ~ 5 kOe).
J.
France 50
Phys.
(1989)
2727-2739
15
SEPTEMBRE
2727
1989,
Classification
Physics Abstracts 71.30
The
-
72.80L
-
74.70K
ubiquity of
-
75.30F
the
new
organic
conductor
ditetramethyl-
dithiadiselenafulvalene-hexafluorophosphate (TMDTDSF )2PF6 P. Auban (1), D. Jérome K. Bechgaard (2)
(1),
K.
Lerstrup (2),
I. Johannsen
(2),
M.
Jorgensen (2)
(1) Laboratoire de Physique des Solides (associé au CNRS), Université Paris Sud, 91405 France (2) H. C. Oersted Institute, Universitetsparken 5, DK 2100 Copenhagen, Denmark (Reçu
le 6 avril 1989, révisé le 12 mai 1989,
accepté le
1 er
and
Orsay,
juin 1989)
Nous présentons une méthode de préparation et les résultats d’une étude des de transport sous haute pression du matériau (TMDTDSF )2PF6 élaboré à partir de la nouvelle molécule di-tétraméthyldithiadisélénafulvalène. Le composé est isostructural à la série (TMTTF-TMTSF )2X. Une localisation induite par les interactions coulombiennes est observée au-dessous de 180 K sous pression atmosphérique. Un ordre à longue portée de type spin-Peierls et un état à ondes de densité de spins sont observés à 20 et 7 K respectivement. L’existence d’un état antiferromagnétique est bien établie par la disparition de la résonance électronique du mode Zeeman et l’apparition simultanée d’un nouveau mode à un champ de résonance (Hr ~ 5 kOe) largement supérieur au champ de résonance Zeeman. La dépendence angulaire de ce nouveau mode est la signature d’une résonance antiferromagnétique. On peut considérer (TMDTDSF )2PF6 comme un matériau prototype pour l’étude des différentes instabilités de basse température d’un gaz électronique quasi 1-D en fonction de la pression. Les effets de localisation et l’instabilité spin-Peierls sont supprimés sous pression. Au-delà de 20 kbar, un état à conduction de type métallique est stable à basse température. Toutefois l’absence d’état supraconducteur audessus de 0,56 K pourrait être liée au désordre structural des empilements cationiques ou à la faiblesse du couplage interchaines.
Résumé.
2014
propriétés
Abstract. Preparation, magnetic and transport properties under pressure of an organic conductor based on the new molecule di-tetramethyldithiadiselenafulvalene (TMDTDSF )2PF6 are reported. This compound is isostructural to the (TMTTF-TMTSF )2X series. At ambient pressure, Coulomb induced localization is observed below 180 K. Spin-Peierls ordering occurs at 20 K and a spin density wave state is stabilized below 7 K. The onset of an antiferromagnetic state has been conclusively established by the vanishing of the ESR spectrum at the Zeeman frequency (X-band) and the concomitant occurrence of a new resonance line at a resonance field (Hr ~ 5 kOe) far above the Zeeman field (g 2) and whose angular dependence provides the signature for antiferromagnetic resonance. Both charge localization and spin-Peierls states are suppressed under pressure. Above 20 kbar, in spite of the stabilization of the conducting state (metallic-like) at low temperatures no superconductivity can be observed down to 0.56 K. This feature could be related to the existence of structural disorder or to the weakness of the interchain coupling in the title compound. (TMDTDSF)2PF6 can be considered as a prototype material for the study of the instabilities occurring in a quasi 1-D electron gas at low temperatures via an adequate adjustment of the pressure. 2014
=
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphys:0198900500180272700
2728
Introduction. in organic conductors was first discovered in the quasi-one dimensional radical-cation salt (TMTSF)2PF6, below 1 K under a pressure of 9 kbar [1]. Below a critical pressure of about 9 kbar this organic conductor undergoes a sharp conductor to semiconductor transition at a temperature which depends on the applied hydrostatic pressure ( TN = 12 K under ambient pressure). Furthermore, NMR experiments [2] have shown that the transition towards the dielectric state is accompanied by the onset of an internal magnetic modulation. Single crystal susceptibility [3] NMR [4, 5] and antiferromagnetic resonance experiments [6] have firmly established the existence of a static spin modulation, a spin density wave (SDW) characterized by a modulation wave vector with components (2 kF, 0.246b*, 0.06c*) and amplitude 0.08 Blmolecule at 4.2 K. (TMTSF )2PF 6 belongs to a broader family of conducting materials with different anions X [7]. One of them, X C104 displays superconductivity at 1.2 K under ambient pressure [8]. Furthermore, the (TMTSF)2X series possesses an all-sulfur analog series, (TMTTF)2X, which is isomorphous to the (TMTSF)2X series. This series has been known for quite some time [9] but never triggered much interest due to the highly insulating character of the conductivity below 200 K or so [10]. The investigation of the magnetic properties of (TMTTF )2PF6 and (TMTTF)2Br has revealed the existence of a non-magnetic ground state (spin-Peierls) [11, 12] with a structural distorsion [13] and of an antiferromagnetic SDW ground state [14] in the former and latter materials below 19 and 13 K respectively. Furthermore, a cross-over between the SP and the SDW ground states is observed in (TMITF)2PF6 under pressure (around 10 kbar). The insulating SDW ground state of (TMTTF)2Br is suppressed by a pressure of 25 kbar [15]. A possibility of superconductivity below 4 K has been reported in this material under pressure [15]. The ability to span a broad variety of physical properties in the (TM’ITF)2X series either by looking at different salts with different anions or by the use of high pressure has strongly stimulated the development of the theory of these conductors. It has also been realized that unravelling the properties of various (TMTTF)2X compounds might be very helpful for the understanding of the pairing mechanism in the superconducting phase of (TMTSF )2PF6 since this phase shares a common border with the SDW ground state [16]. The freezing out of charge degrees of freedom at high temperatures together with the existence of phase transitions at low temperatures 20 K) involving the spin degrees of freedom has led theoreticians to propose an interpretation in terms of 1D physics concepts. As emphasized by Barisié and Brazovskii [17], the existence of a potential at wave-vector 4 kF has important consequences on the physics of the 1D electron gas. As noticed by Pouget [18], the two organic molecules in the unit cell are equivalent by inversion symmetry. However, the intermolecular intrastack bonds display a dimerization due to the anion 3-D environment. Consequently, the holes in the conduction band see a small anion potential with periodicity 4 kF which is superimposed to the main organic lattice potential whose periodicity is 8 kF. This additional anion potential creates small gaps à (4 kF ) at ± 2 kF in the 1-D bands and in turn makes the conduction band 1/2-filled instead of 3/4-filled as inferred from purely stoechiometry consideration [19]. The relative size of the 4 kF potential in different compounds can indeed be estimated from the existence of a slight dimerization of the organic stack which is revealed by the X-ray structure analysis. The dimerization is rather large for the all-sulfur materials (> 1 % or so) but remains relatively small or even non-existing in the allselenium series. The loss of the charge degrees of freedom below 200 K has been attributed to a 4 kF localization due to the repulsive interactions between 1-D carriers in the stacks [19]. When a band is half-filled a new scattering channel is open within the 1-D theoretical framework besides backward (gl) and forward (g2) scattering, this is the Umklapp scattering
Superconductivity
=
2729
of
carriers
from
± kF to :i: kF with the corresponding coupling constant The dimerization gap A (4 kF) is usually small in the (TMTTF )2X 93 - g 1 A (4 A:F)/Ep [17]. series. Thus, the effect of g3 «.c gl) on the electronic properties of the 1-D conductors are expected to be important only in the low temperature regime (say below room temperature around Tp). In the weak coupling limit, Tp is proportional to g3 [19]. Since the spin susceptibility of (TMTTF )2X materials is significantly enhanced over the bare Pauli value and does not show any sign of activation down to the low temperature domain, the sign of gl is positive [17]. If the inequality gl : 2 g2 + g3 if fulfilled, a gap àp opens up in the charge degrees of freedom at Tp (= Llp /7T) driven by a 4 kF Wigner localization of the carriers. Below 7p the conductivity becomes activated but the spin degrees of freedom remain unaffected by the charge localization. Whenever 7p is large, say 7p > t 1- /7r where t_L is a transverse overlap integral, 1-D physics remains applicable in the temperature domain Tp:> T:> t 1- / 7T. Consequently, the two competing correlations, the 2 KF-CDW (bond) and the 2 KF-SDW diverge according to the same power law (Tp/T). The 2 KF-CDW may couple to the lattice through the electronphonon interaction and gives rise to a 2 kF-lattice modulation observable by X-ray diffuse two
scattering [19]. Thus, if the electron-phonon interaction is large enough, the resulting 2 kF lattice modulation can be detected by the occurrence of a 1-D softening of the 2 kF phonon. This additional dimerization (of the 1-D chain of localized Heisenberg spins) is accompanied by a reduction of the spin degrees of freedom as the uniform chain of spins evolves towards a dimerized chain of spin singlets. 1-D spin-Peierls fluctuations exist in this temperature domain. These effects (2 kF lattice fluctuations and the concomitant reduction of Xs) become visible below the mean-field SP temperature esp. In (TMTTF)2PF6, TsP is about 40 K at ambient pressure. Below esp the growth of the 2 KF-SDW fluctuations is weakened by the reduction of the spin degrees of freedom and a transition towards a 3-D SP long range ordered state is achieved at Tsp via the interstack coupling between 2 kF bond CDW and lattice fluctuations. Below Tsp( , 20 K in (TMITF)2PF6) the properties of the material are characterized by a 3-D 2 kF lattice modulation and a dimerized assembly of electrons in singlet states.
of weak 4 kF Coulomb localization (for example if Tp is much smaller than room temperature) the growth of the 1-D SP fluctuations may not be large enough in the 1-D regime for the SP instability to win the competition against the SDW state. Thus, a SDW state gets stabilized instead. This situation is encountered in (TMTTF )2Br at ambient pressure (Tp -- 100 K) [20] and also in (TMTTF)2PF6 above 10 kbar [11]. The crossing-over between SP and SDW phases has been extensively studied in (TN=F)2PF6 under pressure [21]. The P-T diagram of the all-selenium compound (TMTSF)2PF6 has also shed light on the well-known cross-over between SDW and Superconductivity. However, so far the sequence of SP, SDW and conducting (superconducting) states has yet never been observed in the same compound by a proper adjustement of the pressure. Therefore, we have decided to elaborate a new material with the same structure as the (TMTTF-TMTSF )2X series but based on a new hybrid molecule, TMDTDSF, hoping to get a material behaving as a prototype for the In
case
physical properties. Material
preparation.
The TMDTDSF molecule can be viewed as the hybrid between TMTTF and TMTSF molecules since the two rings of the molecule look like TMTSF and TMTTF respectively. It is different from another di-selena-dithia molecule reported by Wudl et al. [22] as the molecule
2730
work does not contain the cis-trans mixture which is a built-in source of disorder in the stacking of hetero atoms for the compound of Wudl et al. [22]. So far, the only method of preparation for « mixed » molecules has involved the mixing of the two halves and carrying out one of the standard coupling reactions on the mixture. This procedure will of course result in a mixture of compounds, and the separation of the products is tedious and difficult as best and outright impossible at worse. We have developed a procedure in which these problems have been eliminated and in this manner prepared among other mixed or « unsymmetrical » tetrachalcogenfulvalenes-the tetramethyl-dithia-diselenafulvalene, TMDTDSF, [1] in figure 1. in
our
-
Fig.
1.
-
Preparative
scheme for TMDTDSF.
procedure is loosely based on a method described by Cava et al. [23], but the shortcomings of this method [24] have been eliminated as well, and thus our method provides exclusively the desired, mixed product. We prepared the 2-bisethylphosphonates of various 1,3-dithioles by mixing the appropriate dithiolium-iodide with 1.2 equivalent of triethylphosphite in acetonitrile at slightly elevated temperature [25]. Evaporation of the solvent provided the phosphonate ester in almost quantitative yield. The 4,5-dimethyl-1,3-dithiole-2-bisethylphosphonate [2] was treated with 1.2 eq. potassium-tert-butoxide in THF at - 72 °C, whereupon one eq. of 4,5-dimethyl-1,3-diselenole-2(1-piperidinium) hexafluorophosphate [3] was added. The temperature was allowed to rise The
2731
slowly to - 20 °C and stirring was continued until all solids had disappeared. Diethyl ether added, and the reaction mixture extracted several times with icewater to remove THF, tert-butanol, and salts. The organic phase, containing the adduct [4] was dried over magnesium sulfate, filtered and added a large excess of glacial acetic acid and stirred at room temperature for several hours. The product was filtered off, the filtrate washed with water, dried, the solvent evaporated and the remanence subjected to column chromatography (silica/toluene) to obtain further product. The total yield was typically around 25 %. Elemental analysis, NMR spectroscopy, and gas chromatography of the product revealed no signs of contamination with the symmetrical TMTTF or TMTSF. Single crystals of (TMDTDSF)2PF6 are obtained by electrolysis of solutions (10- 3 M) of the recrystallized and gradient sublimed material in THF solutions containing nwas
BU4NPF6(0.1 M). The determination of the crystal structure [26] provides a triclinic structure P1 with two donors per unit cell as for the (TMTTF-TMTSF)2X series. Unit cell parameters at T 13.374 Â, a 295 K are a 7.632 À, c 83.23 (deg), /3 7.206 Â, b 85.59 (deg), with 71.63 (deg), y interplanar distances 3.61 and 3.59 Â. These values are nearly midway between those of all-sulfur and all-selenium compounds. At first sight the structure is disordered in terms of the heteroatoms stacking. In this work we report the results of a preliminary investigation of the transport and magnetic properties of (TMDTDSF )2PF 6 between 1 bar and 23 kbar. =
=
=
=
=
=
=
Experimental. All experiments except NMR were performed on single crystals with a typical size 2 x 0.2 x 0.2 mm3. The conductivity was measured by the usual low frequency four probe technique. The production of high pressures and low temperatures was achieved by the regular pressure and temperature equipment available at Orsay. Single crystal ESR spectra were obtained in a standard microwave (9.4 GHz) equipment with Hllc * or in the b’ - c * plane. NMR experiments were performed on a 40 mg powdered sample at 45 MHz. VL We report the transport property data in figures 2 and 3. The conductivity (u (300 K) 200 fl-1 cm-1) increases almost linearly under pressure up to 23 kbar with an =
=
Fig.
2. - Pressure
dependence
of the electrical
conductivity
up to 20 kbar at
room
temperature.
2732
initial pressure coefficient of a In o- / 8P + 33 % kbar-1, figure 2 . As shown in figure 3 the of the versus profile resistivity temperature is very dependent on the applied pressure. Several in shown voltage jumps (not figure 3) are usually observed while cooling the sample below 200 K at atmospheric pressure. These jumps are presumably related to the development of microcracks in the sample or (and) at the contacts between the evaporated gold layer and the thin gold wires. Consequently, no detailed analysis of the temperature dependence can be performed on the ambient pressure resistivity data. However, resistivity data do not reveal any sharp anomaly which could be related to the onset of the metal-insulator transition, instead a shallow resistivity minimum is observed at T. - 180 K. The shallow resistivity minimum is shifted down in temperature under pressure and is fully suppressed above 9 kbar. Thus, a sharp conductor-to-semiconductor transition is observed on the resistivity in the range of 20 to 5 K depending on the applied pressure. At 23 kbar, the resistivity exhibits a metalliclike temperature dependence over the entire temperature domain with a ratio p (300 K)/p (4.2 K ) ., 100 and no sign of saturation at low temperatures. We have been able to cool the sample down to 0.56 K at the pressure of 23 kbar without any sign of =
superconductivity.
3. Temperature 16.5 kbar and 23 kbar.
Fig.
-
dependence
of the
resistivity
of
(TMDTDSF )2PF6
at ambient pressure,
The zero pressure behaviour in figure 3 is reminiscent of the temperature dependence of p in (TMTTF)2PF6 at ambient pressure [20]. A similarity is also observed between the ESR data of the mixed sulfur-selenium compound and that of the (TMTTF )2X like compounds. The ESR provides a single Lorentzian line whose characteristics are AH r--- 160 G, g - 2.0087 for Hl/c * and dH=175G, g=2.0094 for HI/b’, at 300 K. For Hl/c *, the linewidth decreases smoothly with temperature down to 70 G at 10 K (with gc*(10 K ) --. 2.0094), figure 4. Below 9 K, we have noticed a clear cut change in the shape of the EPR signal at the Zeeman field, which is the superposition of a relatively « narrow » signal ( =$ 65 G) and of another weaker signal with similar g-factor, becoming broader upon cooling below 9 K (65 G). The origin of the « narrow » signal is still unclear. If this signal is due to paramagnetic impurities following a Curie law, its weight in the ESR spectrum should be negligible above 15 K. However, the remarkable feature displayed by X-band magnetic
2733
Fig.
4.
Temperature dependence of the ESR linewidth àHp.,. (HIlc
-
*).
is the onset, below 7 K, of a new resonance line in the fiel range 4the Zeeman resonance field ouf - 3.3 kOe), figure 5. When the far from (i.e. is rotated in the plane perpendicular to the a (stacking) axis, the angular field magnetic the resonance of field, figure 5, is reminiscent of the typical « bubble » pattern dependence which is observed in antiferromagnets with the rotation axis aligned close to the « hard » magnetic axis [27]. So far, only the high field branch of the bubble has been studied, at T 4.5 K. The axis of symmetry (the easy axis) for the angular dependence is located 12.5 deg. away from the c * axis in the b’ - c * plane. The angular distance between the two maxima of the resonance field ait = 5.53 kOe, is about 20 deg. Rotating the field in the b’ - c * plane decreases the value of the resonance field down to the lowest observable resonance at 4.1 kOe. No resonance could be detected over the entire field domain when the angle exceeds - 45 deg. or + 15 deg. with respect to the c * axis. The value of the resonance resonance
experiments
5.5 kOe
=
Fig. 5. a-axis (T
-
Angular dependence
= 4.5 K ).
of the AFMR field
resonance
Hr when the sample is
rotated around the
2734
field corresponding to the central cusp anomaly (Hr = 5.4 kOe ) provides an estimate for the zero-field antiferromagnetic resonance frequency [28], f2- - 4.2 kOe which is very close to values obtained in other organic materials displaying a SDW state [29]. The spin susceptibility, figure 6, has been derived from the area under the ESR spectrum and a calibration with a (TMTSF)2PF6 single crystal of known susceptibility [30]. The proton spin-lattice measurements provide data very similar to those obtained in (TMTTF)2PF6 in the 35-20 K temperature domain [31]. However, below 20 K an exponential vanishing of Tîl is observed for the all sulfur material whereas an upturn of the relaxation rate is noticed for (TMDTDSF )2PF6 below = 18 K giving rise to a Tï1maximum at 7 K. Below about 10 K the recovery of the magnetization after the usual saturation comb becomes non-exponential. The time constant which is plotted in figure 7 corresponds to the time constant which is obtained after a long delay. The insert of figure 7 shows that the Ti1 singularity which develops at TN 7 K can be analyzed up to 15 K or so by1 a square root temperature divergence. Below 7 K we have not tried to analyse the Ti temperature dependence in details because of the strong non-exponential character of the magnetization recovery. However, we have also plotted in figure 7, around TN, the temperature dependence of the time constant corresponding to the growth of the magnetization after saturation up to M 0.63 Mo. In the limit of exponential recovery, this time constant is nothing but Tl. As shown in figure 7, the anomaly of this time constant at TN is far more pronounced than that observed in the T-dependence of the long time recovery. Non exponential recoveries have been systematically observed at the onset of aniferromagnetic ordering in organic conductors [14]. So far, no satisfying interpretation has been proposed. One possibility lies in a possible quenching of the nuclear spin diffusion process between neighbouring nuclei due to large local magnetic fields gradients in the SDW state or even in the critical region near TN. Such a decoupling is known to exist at the onset of a magnetic ground state in (TMTTF )2Br at 19 K since T2 decreases and T2 increases below the phase transition [14]. Consequently, the maximum of T11at TN instead of a divergence and the apparent saturation of In Tinear TN, which is displayed in the insert of figure 7, may be attributed to the abovediscussed quenching of spin diffusion. =
=
6. Spin susceptibility XS versus temperature deduced from ESR measurements (Holl c *). The low temperature region shows the spin-Peierls transition around 20 K. The whole temperature behaviour is presented in the insert. The absolute value of XS is derived from a calibration with the
Fig.
-
(TMTSF)2PF6 compound.
2735
Temperature dependence of the long-time tail Tl (lH) at ambient pressure showing the onset of the SDW state at 7 K. Around 7 K, the temperature dependence of the time constant corresponding to the growth of the magnetization after saturation up to M 0.63 Mo is also plotted (*). The insert shows a fit of the experimental data with a power law Ti’ - (F- T,)- "2 ( TN 7 K).
Fig.
7.
-
=
=
Discussion and conclusion.
As already emphasized, the transport data in figure 3 are very similar to those obtained in the series of all-sulfur compounds. The Coulomb-induced localization below 180 K is rapidly removed under pressure. Above 10 kbar a large regime in temperatures exhibiting metalliclike conductivity extends down to the sharp phase transition ait = 20 K and even lower temperatures at higher pressures. As for (TMTTF )2Br under very high pressure, the insulating ground state is fully suppressed above 20 kbar. The material retains then its conducting state down to the lowest temperatures. In (TMTTF)2PF6 a pressure induced stabilization of the conducting ground state was also expected under pressure but at a pressure which is beyond the possibilities of the hydrostatic pressure equipment [15]. However, the highlight of the present compound is provided by the low temperature behaviour at ambient pressure. The Tianomaly at 7 K and the square root divergence of Tiat higher temperatures both argue in favour of the establishment of a SDW state at 7 K. Similar features have been observed by NMR at the onset of a SDW state in (TMTTF )2Br [14], (TMTTF )2PF6 under 13 kbar [11] and (TMTSF)2PF6 [31]. The existence of a SDW state is indeed firmly established by the observation of a new (orientation dependent) resonance mode below 7 K. The maximum of Tiis attributed to the divergence of antiferromagnetic fluctuations at the onset of a SDW state and the square root temperature dependence is the signature of critical fluctuations in a 3-D critical regime [32]. As noticed in the previous section, the change in the EPR line shape below 9 K makes the discussion of the ESR properties more delicate at the onset of the SDW state. We believe that the broadening of the small amplitude signal can be attributed to the onset of antiferromagnetism. The line broadening is admittedly smaller than what is observed in other materials with SDW states occurring at higher temperatures [20]. In the latter compounds, the possible existence of an extrinsic impurity resonance line interferes less severely with the intrinsic resonance provided by itinerant electrons since they exhibit somewhat higher values of
2736
A
closely resembling situation
of low TN is encountered in the quenched-state of In this the existence of a SDW state was suggested by NMR data [33] case, (TMTSF)2CI04, below 7 K and confirmed by AFMR experiments at 1.6 K [34]. However, the approach to the SDW above 3.6 K is characterized by an only moderate broadening of the ESR spectrum [35]. The anomaly at 20 K in the temperature dependence of the susceptibility, figure 6, which is not accompanied by any broadening or shift of the ESR line is the signature of a SP transition affecting the spin degrees of freedom. However, this reduction remains rather modest when compared to the drastic drop of susceptibility which is observed below 19 K in (TMTTF )2?F6 where X (10 K) amounts to only X (20 K)/20 [11]. The weak reduction of the spin degrees of freedom below 20 K in the S-Se material leaves the 2 kF spin fluctuations active enough to enable the establishment of a SDW state at a lower temperature. Apparently, the BCS relation between the magnitude of the spin-Peirls gap and the ordering temperature is not followed in the present situation (otherwise there would be no spin degrees of freedom left for the establishment of antiferromagnetism below Tsp/2). We attribute the failure of usual descriptions of phase transitions to account for the development of the spin-Peirls gap below Tsp to the close competition between SP and SDW long range orders (vide-infra). Summarizing transport and magnetic data a phase diagram can be drawn for (TMDTDSF)2PF6 under pressure, figure 8. In the low pressure regime we have sketched as a guide for the eye the pressure dependence of the SP transition according to the known pressure dependence in (TMTTF)2PF6 [21].
TN.
Phase diagram Temperature-Pressure deduced from resistivity measurements under pressure Fig. 8. (*, A), ESR susceptibility (e) and NMR determination of Tl (9). The diagram shows the presence of charge localization (CL), spin-Peierls (SP) and spin density wave (SDW) states. The data points (o ) above 20 kbar indicate « symbolically » that the compound remains conducting down to the lowest -
temperature reached in this work, i.e. 0.56 K. As usual in the physics of quasi 1-D conductors, the occurrence of transitions towards 3D long range ordered states and the nature of these states are governed by the transverse coupling between 1-D correlations. When both SP and SDW 1-D channels become equally divergent at low temperatures (in the vicinity of 9 kbar) there exists a competition for the establishment of one or the other of the two mutually exclusive orders. In the very vicinity of the phase transitions the two competing instabilities can thus be treated phenomenologically
2737
by a Landau expansion of the free energy comprising symmetries as outlined in reference [21].
two order
parameters of different
where in equation (1) the coefficient of the quadratic term first vanishing determines the nature of the long range order, ASDW, Asp::> 0 and the fourth order term is positive. A prediction of the Landau expansion is the possibility for a phase transition between the two states of low symmetry [20] (via a first order transition). This transition between SP and SDW orderings is indeed observed at 7 K in the present study and provides a strong support for the description in terms of a Landau expansion. Such a phase transition had never been observed so far in the study of organic conductors. However, we have failed to detect any first order character associated with the SP - SDW transition. We attribute this failure to the smallness of the jump of the SP order parameter which is only weakly developed when the SDW state becomes stabilized at the 7 K transition. In addition, the repulsive interaction between the two order parameters is responsible for the depression of the onset of SDW at P 1 bar and for its rise under pressures between 1 bar and 10 kbar. The absence of superconducting transition down to 0.56 K when the SDW state is suppresed under pressure (P 23 kbar ) could be a consequence of the structural disorder in the packing of sulfur and selenium atoms as inferred from the structure determination [26] and from monochromatic Laue X-ray diffraction patterns [36]. In conclusion, the new material (TMDTDSF)2PF6 is a remarkable prototype material for the study of quasi 1-D conductors belonging to the quasi 1-D (TMTTF-TMTSF )2X family. Competitions between spin-Peierls and SDW on the one hand and between SDW and conducting ground states on the other hand have been clearly observed. The novelty of our investigation is the finding of a phase transition between SP and SDW long range ordered states. The rapid suppression of the SP state which is observed under pressure can be attributed to the pressure induced decrease of the dimerization gap 2l(4 kF ) as inferred from the temperature 7p at which Coulomb localization becomes important. The depression of 7p removes the opening of the pseudo-gap in the density of states at the Fermi level at 7’so, (Tsop proportional to Tp). When esp is small enough to be of the order of magnitude of the temperature TN at which long range SDW alone can establish, the stability of the SDW state wins over that of the SP state [21]. The SDW-SC competition has been attributed to the frustration of the SDW state due to large deviations to the complete nesting of the Fermi surface and to the concomitant rise of the 1-D to 3-D cross-over temperature. Both features are induced under pressure ; The absence of superconductivity raises an interesting problem since it may be due to the disordered nature of the material. Although the 3-D superconducting state of these Q-1-D conductors is very effectively suppressed by a weak disorder potential, the amount of Fermi surface nesting above 23 kbar may not be good enough to stabilize a 3-D SDW state instead. Such a behaviour is different from what is observed in (TMTSF)2CI04 when varying the cooling rate below 40 K. In this compound the existing anion disorder in rapidly cooled samples suppresses superconductivity but instead a SDW state sets in around 5 K. However, for the moment, we cannot rule out the possible existence of superconductivity below 0.56 K, since some well ordered S-Se conductors belonging to the (DMET)2X series do exhibit superconductivity in the range of 0.4-0.6 K [37]. We may also suggest another possibility for the low value (or even absence) of T, in the presently investigated compound. Within a tight binding description of the quasi 1-D band structure the transverse energy can be approximated by £.1 (k) = - 2 tb cos kb - 2 tb cos 2 kb b which takes into account the overlap between near and next near neighbour chains along the b-direction with t’ - tl/ta [38]. The stability of the =
=
2738
SDW state depends on the nesting of the Fermi surface. As long as a good nesting of the quasi 1-D Fermi surface can be preserved, the ground state is a SDW state. This situation prevails when t’ is smaller than a critical value th * which is of the order of the mean field SDW transition temperature in the limit Of tb = 0, i.e. th * = TsoDw [39]. As similar values of TSDw = 12 to 14 K are found for the (TMTSF)2PF6 salt at ambient pressure and for the present mixed S-Se compound around 10 kbar, we may infer that tb * is the same for both series of materials (t’* , 12-14 K). On the other hand, the superconducting critical temperature depends on the interchain coupling tb but not on the nesting properties (t’). Therefore, the situation at P > 20 kbar may be such that tb is actually too large for the stabilization of the SDW state but yet too small b t,,, TsDw ) for the onset of superconductivity within the attainable temperature domain. Very likely the amplitude of ta for (TMDTDSF )2PF6 lies in between the all-sulfur and all-selenium values according to the computation of overlap integrals for the (TMTTF-TMTSF )2X series [40]. The likeliness of randomly distributed disorder S-Se, as inferred from X-ray analysis [36], and the role it plays on various low temperature instabilities is still an open problem and will be the matter of future thorough investigations. The present material is particularly well adapted for future experimental investigations. 1D 2 kF lattice softening should be observable below 60 K as it is for (TMTTF )2PF6 by X-ray diffuse scattering techniques. The study of the SP - SDW transition at 7 K would also be very valuable. This compound is also a very good candidate for more thorough AFMR experiments at very low temperatures and extensive NMR studies since the presence of 77 Se nuclei (I 1/2 ) in the TMDTDSF molecule is very appropriate for spin-lattice relaxation studies. Other TMDTDSF salts with non-centrosymmetric anions are presently under =
investigation. Acknowledgments. We wish to thank J. C. Ameline for his skillful help in high pressure and low temperature techniques. We thank J. P. Pouget for pointing out reference [18]. This work has been partly supported by the ESPRIT-Basic Research Action 3121.
Note added in proof : The assumption of a SDW ground state above 5 kbar has been made by with other TMTTF and TMTSF salts. So far, no magnetic data have been obtained for this compound under pressure. The first order transition SP - SDW is not in contradiction with strong fluctuation effects when the first order character of the transition is weak.
comparison
References
[1] [2] [3] [4] [5] [6] [7]
JÉROME D., MAZAUD A., RIBAULT M. and BECHGAARD K., J. Phys. Lett. France 41 (1980) L-95. ANDRIEUX A., JÉROME D. and BECHGAARD K.,J. Phys. Lett. 42 (1981) L-87. MORTENSEN K., TOMKIEWICZ Y., SCHULTZ T. D. and ENGLER E. M., Phys. Rev. Lett. 46 (1981) 1234. TAKAHASHI T., MANIWA Y., KAWAMURA H. and SAITO G., Physica 143B (1986) 417. DELRIEU J. M., ROGER M., TOFFANO Z., WOPE MBOUGUE E. and SAINT JAMES R., Synth. Metals 19 (1987) 283. TORRANCE J. B., PEDERSEN H. J. and BECHGAARD K., Phys. Rev. Lett. 49 (1982) 881. BECHGAARD K., JACOBSEN J., MORTENSEN K., PEDERSEN H. J. and THORUP N., Solid State Commun. 33 (1980) 1119.
2739
[8] [9] [10] [11] [12] [13]
[14] [15] [16] [17] [18]
[19] [20] [21] [22]
[23] [24] [25]
[26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40]
BECHGAARD K., CARNEIRO K., OLSEN M. and RASMUSSEN F. B., Phys. Rev. Lett. 46 (1981) 852. GALIGNE J. L., LIAUTARD B., PEYTAVIN S., BRUN G., MAURIN M., FABRE J. M., TORREILLES E. and GIRAL L., Acta. Cryst. B 35 (1979) 2609. COULON C., DELHAES P., FLANDROIS S., LAGNIER R., BONJOUR E. and FABRE J. M.,J. Phys. France 43 (1982) 37. CREUZET F., BOURBONNAIS C., CARON L. G., JÉROME D. and BECHGAARD K., Synth. Metals 19 (1987) 289. MAAROUFI A., FLANDROIS S., COULON C., DELHAES P., MORAND J. P. and FILHOL G.,J. Phys. France Colloq. 44 (1983) C-3, 1091. POUGET J. P., MORET R., COMES R., BECHGAARD K., FABRE J. M. and GIRAL L., Mol. Cryst. Liq. Cryst. 79 (1982) 129. CREUZET F., TAKAHASHI T., JÉROME D. and FABRE J. M.,J. Phys. Lett. France 43 (1982) L-755. PARKIN S. S. P., CREUZET F., RIBAULT M., JÉROME D., BECHGAARD K. and FABRE J. M.,Mol. Cryst. Liq. Cryst. 79 (1982) 249. JÉROME J. and SCHULZ H. J. S., Adv. Phys. 31 (1982) 299. BARISI0106 S. and BRAZOVSKII S., in Recent Developments in Condensed Matter Physics, J. T. Devreese Ed. (Plenum, New York) 1981. POUGET J. P. in Low Dimensional Conductors and Superconductors, D. Jérome and L. G. Caron Eds. (Plenum Press) 1987, p. 17. EMERY V. J., BRUINSMA R. and BARISI0106 S., Phys. Rev. Lett. 48 (1982) 1039. COULON C., Thèse, Université de Bordeaux I (1982). CARON L. G., CREUZET F., BUTAUD P., BOURBONNAIS C., JÉROME D. and BECHGAARD K., Synth. Metals B 123 (1988) 27. WUDL F., WALSH W. M. Jr., HAUSER J. J., DISALVO F. J., NALEWAJEK D., RUPP L. W. Jr., WASZCZAK J. V., FELDER R. J., POEHLER T. O., CHAIKIN P. M. and LACOE R., J. Chem. Phys. 79 (1983) 1004. GONELLA N. C. and CAVA M. P.,J. Org. Chem. 48 (1978) 369. FABRE J. M., GIRAL L., GCUASMIA A., CHRISTAU H. J. and RIBEIL Y., Bull. Soc. Fr. Chim. (5) (1987) 823. (a) ISHISKAWA K., AKIBA K. and INAMOTO N., Tetr. Lett. 41 (1976) 3696. (b) LERSTRUP K., JOHANNSEN I. and JORGENSEN M., Synth. Metals 27 (1988) B9. THORUP N. et al., to be published. COULON C., SCOTT J. C. and LAVERSANNE R., Phys. Rev. B 33 (1986) 6235. LAVERSANNE R., Thèse, Université de Bordeaux (1987). TORRANCE J. B., J. Phys. France 44 (1983) C3-799. MILJAK M., COOPER J. R. and BECHGAARD K., J. Phys. France 44 (1983) C3-893. CREUZET F. et al., to be published. BOURBONNAIS C., STEIN P., JÉROME D. and MORADPOUR A., Phys. Rev. B 33 (1986) 7608. TAKAHASHI T., JÉROME D. and BECHGAARD K., J. Phys. Lett. France 43 (1982) L-565. WALSH W. M., WUDL F., AHARON-SHALOM E., RUPP L. W., VANDENBERG J. M., ANDRES K. and TORRANCE J. B., Phys. Rev. Lett. 49 (1982) 885. TOMI0106 S., JÉROME D., MONOD P. and BECHGAARD K., J. Phys. Lett. France 43 (1982) L-839. POUGET J. P. and MORET R., private communication. KIKUCHI K., MURATA K., HONDA Y., NAMIKI T., SAITO K., ISHIGURO T., KOBAYASHI K. and IKEMOTO I.,J. Phys. Soc. Japan 56 (1987) 3436. MONTAMBAUX G. in Low Dimensional Conductors and Superconductors, D. Jérome and L. G. Caron Eds. (Plenum Press) 1987, p. 233. YAMAJI K., Mol. Cryst. Liq. Cryst. 119 (1985) 105. DUCASSE L., ABDERRABBA M., HOARAU J., PESQUER M., GALLOIS B. and GAULTIER J.,J. Phys. C 19 (1986) 3805.
