10 likes | 122 Views
Nonlinear Conductivity in the FISDW states of the Organic Conductor. T. Vuletić 1,2 , P. Senzier 1 , C. Pasquier 1 , S. Tomić 2 , D. Jérome 1 contact e-mail: tvuletic@ifs.hr, jerome@lps.u-psud.fr 1 Laboratoire de Physique des Solides, Université Paris-Sud, Orsay, France
E N D
Nonlinear Conductivity in the FISDW states of the Organic Conductor T. Vuletić1,2, P. Senzier1, C. Pasquier1, S. Tomić2, D. Jérome1 contact e-mail: tvuletic@ifs.hr, jerome@lps.u-psud.fr 1 Laboratoire de Physique des Solides, Université Paris-Sud, Orsay, France 2 Institute of Physics, Zagreb, Croatia (TMTSF)2PF6 Final approach to the question whether the CONDUCTIVITY INCREASES [1] or DECREASES [2,3] under the electric field in the FISDW states of the (TMTSF)2PF6 is attempted. The experimental conditions, that is, high pressure (9 kbar), magnetic field (12T) and temperature range (400 mK – 6 K) in our study are similar to those used in the previous works and, in fact, dictated by FISDW phase diagram, [4,5]. Furthermore, we have enlarged the range of the applied electric field (1µV/cm - 100mV/cm, corresponding to 1µA to 100 mA), by combining lock-in and pulse technique resistance measurements methods. In addition, samples from different batches and with different contact configurations are investigated in order to clarify presently unclear and controversial experimental situation. From theoretical work of Virosztek and Maki [6] we infer that the sign of observed non-linear effect is direct consequence of the shape of Rxx vs. H magnetoresistance curve. This shape, i.e. sharpness of peaks and height (depth) of plateaus [7,8] in magnetoresistance are strongly sample dependent, which is further connected to the purity level of the virgin sample, which cannot be controlled by available chemical techniques. This sample dependence might offer the way-out of the current dilemma. We propose the strong sample/batch dependence as an answer to ambiguous question of nonliner conduction in FISDW. Single - particle transport Longitudinal resistance Rxx vs. T. Data is for the sample with 4 annular contacts. Curves acquired with lock-in technique in low electric field limit (10 µA) are reproduced (within measurement error) with RDC points. RDC points are result of averaging of resistances acquired in low current/ field range with pulse technique (still linear range up to 1mA). For all magnetic field values low-temperature resistance is higher than high-temperature. Only for highest fields you see the local maximum in resistance but the resistance is not substantially decreasing, although it is to be expected that it will be approaching zero for lowest temperatures (according to previous, unpublished, work of L. Balicas or N. Biškup in Orsay). Thus it is to say that this sample didn’t enter the quantized regime. In question was possibility that four annular contacts on the sample short-circuited b’ planes in the sample, thereby removing QHE. But other sample (from the same batch) with 8 contacts showing the same behaviour ruled this out as a possibility. Different magnetic fields correspond to a sequence of FISDW phases: phase 2: 11.7, 12T phase -2: 11.3T phase 3: 10.3T phase 4: 9.1T phase 5: 7.8T References Experimental observation of nonlinear conductivity in FISDW [1] W.Kang, J.R.Cooper and D.Jérome Synth. Met., 41-43, 2083, (1991) [2] T.Osada et al. Phys.Rev.Lett., 58 , 1563, (1987) [3] L. Balicas Phys.Rev.Lett., 80 , 1960, (1998) FISDW phase diagram [4] J.R.Cooper, W.Kang, P.Auban, G.Montambaux and D.Jérome Phys.Rev.Lett., 63 , 1984, (1989) [5] S.T.Hannahs, J.S.Brooks, W.Kang, L.Y.Chiang and P.M.Chaikin Phys.Rev.Lett., 63 , 1988, (1989) Theoretical prediction on nonlinear conductivity in FISDW [6] A.Virosztek and K.Maki Phys.Rev.B , 39 , 616, (1989) or Synth.Met., 29, 385, (1989) Peaks & plateaus in Rxx vs. H [7] L. Balicas, G.Kriza, F.I.B.Williams Phys.Rev.Lett., 75 , 1960, (1995) [8] J.B.Young, H.Cho, S.T.Hannahs and W.Kang, » QHE & Edge States in (TMTSF)2PF6«, I don ’t know the journal, do you? Magnetoresistance FISDW -- nonlinear transport 11.7 T phase N = 2 Longitudinal resistance R vs. +/- H. By (anti)simmetrysation one gets Rxx and Hall resistance. FISDW phase transitions are clearly visible. The onset of FISDW at 5.5T is as expected for 0.4K, together with FISDW phase transition positions [4,5]. Small weakly resolved peaks and especially plateau values of Rxx which are higher than resistance values in metallic phase (at low magnetic field and/or high temperature) have to be noted [7,8]. This is directly connected to weak 0.4K minima in Rxx vs. T curves. Of course, because they are the same points in (T, H) phase diagram. Previous figures can be used as a key which determines FISDW nonlinearity, that is are you (later in the course of experiment) going to detect increase or decrease of resistance for high electric fields. Simply put, if resistance is higher in FISDW than in metallic, then resistance has to decrease. Reversed situation gives rising resistance. Refering to Virosztek and Maki: In the non-ohmic regime where the SDW is sliding Rxx magnetoresistance recovers normal (metalic) state values. This picture is able to accomodate all previous experimental work on nonlinear conduction in FISDW state, nevermind apparent contradictions. With magnetic field stabilised and sample in one of FISDW phases, voltage is measured on the sample in relation to applied current. This gives Rxx vs. E curves for multitude of points in temperature. Figures present situation for the phase 2 at 11.7T, last phase reachable with our magnet. Low electric field data is noisy, so it is not easy to make objective quantification of low field resistance, which is in effect DC resistivity of the sample, RDC (without SDW condensate sliding) and evaluation of threshold electric field comes with large error bars. Osada et al. [2] presented figures with the same overall shape of R vs. T curves (but, with resistance increasing) : very low threshold field (below measurable limit) and saturation value of resistance for high electric field values. In figure above, right you can extrapolate all the curves to saturate at the same resistance (within measurement error). Plots for different FISDW phases show the same general behaviour with gradually decreasing values for saturation resistances for subsequent phases N= -2, 3, 4, 5. Values of resistance are in fact high temperature values of magnetoresistance outside FISDW phases. With Joule heating declined, taking into account large non-linear effects observed (100% to 300%), and as most indicating fact, saturation resistance values resembling magnetoresistance values if there were no FISDW, then it is tempting to suggest that the FISDW gap is gradually removed. All this accords to Virosztek and Maki This amounts to picture where decreasing resistance simply follows the fact that the low-temperature FISDW resistance is higher than high-temperature metallic resistance for the samples I have investigated (see fig.1.). 11.3 T phase N = -2 CONCLUSION We have shown that the features of both the collective conduction channel in the SDW ground state and the SDW low-frequency dielectric response are sample (batch) dependent. In the PF6 samples in which the SDW is pinned by impurities the electric threshold field which characterizes the sliding mechanism gradually weakens with temperature, concomitantly as the SDW dielectric relaxation gradually slows down. In contrast, in the PF6 and AsF6 samples in which the commensurability pinning prevails [5,7], ET displays the maximum below 3 K, and a critical slowing down of the SDW dielectric relaxation might be observed. It should be noted that the latter behaviour was only observed in the sample in which an extremely narrow ET peak was found. Keeping in mind that the SDW wave vector is found to be very close to commensurability [8], we propose that subtle variations of the disorder level in nominally pure samples might be responsible for the different behaviours observed.