Resistivity studies on different variants of -(BEDT-TTF) 2 Cu[N(CN) 2 ]Br :

1. Resistivity studies on different variants of -(BEDT-TTF)2Cu[N(CN)2]Br : evidence for disorder and/or defect-induced inelastic scattering contributions Ch. Strack, C. Akinci, B. Wolf, M. Lang J. Schreuer, L. Wiehl E. Uhrig, W. Aßmus J.W. Goethe-Universität Frankfurt Forschergruppe 412 Collaborations: J. Müller*, J. Wykhoff, F. Steglich MPI Dresden, *Florida State Univ. J. Wosnitza High-field Laboratory, Dresden D. Schweitzer Univ. Stuttgart J.A. Schlueter Argonne NL

2. -(ET)2X : resistivity & phase diagram 500 bar cf. Limelette et al. ‘03 paramagnetic semiconductor anomalous (bad) metal T(K) 25 paramagnetic metal 10 afm insulator superconductor p Kanoda ’97 Kanoda ’97 Cu(NCS)2 X = Cu[N(CN)2]Cl Cu[N(CN)2]Br “Metallic state is characterized by 3 transport regimes“ : - high T: semiconducting - intermediate T: anomalous (bad) metal - low T:   AT2

3. Resistance maximum • order-disorder transition of ethylene endgroups • crossover from localized small polaron to coherent • large polaron behavior • consequence of strong electron correlations Parker et al., ’89 Kund et al., ’93 Tanatar et al.‘99 Wang et al., ‘00 Merino et al., ’00 -(ET)2Cu[N(CN)2]Cl under pressure DMFT for the metallic side of the Mott transition Limelette et al., ‘03 Limelette et al., ‘03 Toyota et al., ‘91 Montgomery et al., ’99 Schweitzer • magnetic scattering by Cu2+

4. Sample-to-sample variations  (300K) 67 84 89 158 -(BEDT-TTF)2Cu[N(CN)2]Br “high-resistance“ (HR) “low-resistance“ (LR) Crystals from two preparation routes (using TCE and THF & EG as solvent): - strongly sample-dependent (T) profiles ! - with metallic- and semiconducting-type behaviors !

5. Sample characterization Electron probe microanalysis: identical chemical compositions High-resolution X-ray : identical structure parameters; same occupation factors for ethylene ordering (69  2)% @ 300 K and (92  2)% @ 100 K Low cooling rate 0.1 K/min : minimizes disorder due to ethylene ordering at Tg = 77 K • cf. Wosnitza ’99: • Shubnikov de Haas oscillations • on a “LR“ crystal • LR crystals of superior quality compared to standard HR crystals RRR = 193 TD = (2  0.2) K

6. Superconductivity & hydrostatic pressure effect on Tc using He-gas pressure  (300K) “LR“ 10% - 90% width: pressure coefficient: TcLR = 0.2 K -(2.6  0.2) K kbar-1 TcHR = 0.4 K -(2.4  0.2) K kbar-1 “HR“ “LR“ vs “HR“ : almost identical Tc and Tc/p • rules out: effects of anomalous thermal contraction or internal strain magnetic scattering centres

7. resistivity maximum under pressure   (300K) (300K) “LR“ “HR“ bar bar bar -45 %/kbar -89 %/kbar - 85 %/kbar -82 %/kbar - 64 %/kbar -45 %/kbar -180 %/kbar -120 %/kbar -360 %/kbar -250 %/kbar • additional scattering contribution - strongly pronounced at intermediate temperatures • - huge pressure dependence • - reduced, though finite, in the LR crystal

8. HR vs LR crystals • Crystals differ in their real structure, i.e., the concentration and character of defects and/or disorder, which • - strongly affects the inelastic scattering contribution, i.e., the • dynamics (e.g., by influencing certain phonon modes), • - is characterized by a huge pressure dependence • - but is irrelevant for superconductivity

9. related materials in-plane -(ET)2Cu(NCS)2 THF A. Ugawa et al., PRB 38, 5122 (1988)

10. Low-T normal state -(BEDT-TTF)2Cu[N(CN)2]Br  (cm)  (cm) 28 K T0 23 K T0 “HR“ “LR“ • = 0 + AT2 for T  T0 but with strongly sample-dependent A and T0! cf. Fermi liquid: AFL (m*)2  (TF*)-2  AFL(x)(TF*(x))2  c = const. ! for TF*  T0 : cHR  3cLR  origin of   AT2 different from coherent Fermi liquid excitations !

11. Anomalous metallic range 67 84  89 (300K) T* 158 T* Tglass NMR relaxation rate sound velocity T* T* Mayaffre et al. ’94, Kawamoto et al. ’95, DeSoto et al. ‘95 X = Cu(NCS)2 Frikach et al. ’00, Shimizu et al. ‘00 X = Cu[N(CN)2]Br “incipient divergence of el“ “opening of a pseudo-gap“

12. thermal expansion T(K) anomalous metal paramagnetic insulator T* 40 25 paramagnetic metal 10 afm Insulator Superconductor p X = Cu[N(CN)2]Br X = Cu[N(CN)2]Cl T* Tglass Tglass  (10-6K-1)  (10-6K-1) TN Tc v/v, (T1T)-1 l/l  10-10 J. Müller, M.L, et al., ‘02

13. T*  T* T* anomaly -(ET)2Cu[N(CN)2]Br -(ET)2Cu(NCS)2 T* T*  =  - b M.L., J. Müller, et al., ‘02 anomaly atT*: - reminiscent of 2nd-order phase transition - of similar size and width for X = Cu(NCS)2

14. Transport vs. thermodynamic properties X = Cu[N(CN)2]Br d  (10-6K-1) dT  (a.u.) a T* Tc d/dT Tg T(K) d/dT T = T*: coincidence of sharp anomalies in (T) and (T) - not expected for a “crossover“ behavior (cf. Kondo effect T  1.25T )! - indicating a phase transition !?

15. Summary • differently prepared single crystals of -(ET)2Cu[N(CN)2]Br show: • strongly sample-dependent (T) profiles (cf. X = Cu(NCS)2) • but identical structure parameters and chemical compositions •  additional inelastic scattering contributions is • - caused by real structure effects, i.e., defects and/or disorder • - strongly prounced at intermediate temperatures 20 ... 200 K • - strongly pressure dependent • semiconducting behavior at T > 90 K is not an intrinsic property ! • origin of “  AT2 “ is different from coherent FL excitations ! • sample-independent distinct (T) anomaly at T* = 40 K which coincides with pronounced anomalies in (T), (T1T)-1, v/v, ... and remains sharp for X = Cu(NCS)2 • - not expected for a crossover ! • - but rather indicates a phase transition ! • Ch. Strack et al., Phys. Rev B 72, 054511 (2005)

16. cf. 2D Fermi liquids Kadowaki – Woods scaling: A/2 = a0 A -(ET)2Cu[N(CN)2]Br Sr2RuO4 (Ac) Ain-plane Sr2RuO4: m*/m  3.6 Sr2RuO4 (Aab) Y. Maeno et al., ‘97 a0 = 110-5 cm/(mJ/Kmol)2 Cf. M. Dressel et al., ‘97 a0/25

17. anisotropy =  Buravov et al., ‘92

18. anisotropy

19. T0: in-plane vs out-of-plane

20. T(K) anomalous metal paramagnetic insulator T* 40 25 paramagnetic metal 10 afm Insulator Superconductor p X = Cu(NCS)2 x 5 460 bar same scale Ch. Strack et al., 2004