UPt 3 : More data after all these years

Transcription

1 UPt 3 : More data after all these years C. P. Opeil, S.J., M. J. Graf Boston College, Physics Department, Chestnut Hill, MA, USA A. de Visser University of Amsterdam, Van der Waal-Zeeman Institute, Amsterdam, The Netherlands

2 The UPt 3 Story Begins on a Friday Afternoon: CeCu 2 Si UPt 3 UBe 13

3 UPt 3 exhibits Unconventional - SC (T c ~ 0.5K): van Dijk, Physica B 186 (93) Fisher et al., PRL 62,(89).

4 UPt 3 Spin Fluctuation Behavior 1) Crystal Structure HCP 2) Thermodynamic Signature of Spin Fluctuations - Specific Heat: T 3 ln(t/t SF ) at low T, AFM correlations in Magnetism, T SF ~ 18 K, B // a-b 3) Neutron scattering [Aeppli, G. et al. PRB32(85)] - AFM and FM spin fluctuations T SF ~ 18 K UPt 3 Ferromagnetic Correlations In-plane, exist up to T~150 K B // a,b axis B // c axis Antiferromagnetic Correlations between planes, appear at T<30 K Frings et al. PRB 31 (85) c axis a-b plane u Stabilization of AFM correlations starting at ~ 30 K along c (Goldman, PRB,36 (1987)) u Low T, T 3 ln(t/t SF ) typical of SF systems u HF γ = 422 mj/k 2 mole UPt 3 Stewart et al. PRL 52 (84)

5 Using impurities to study the evolution of magnetism and superconductivity Why not pressure? Drives system away from magnetic instability, has minimal effect on superconductivity Pd (or Au) for Pt, or Th for U strengthen AFM in UPt 3 Opposite effect compared to pressure, presumably because all these substitutions slightly reduce c/a ratio in this HCP system U(Pt 1-x Pd x ) 3 is most studied x 0.005: size of ordered moment increases, but onset temperature is constant, and it is still only observable by neutron scattering x > 0.01: onset of conventional AFM with same magnetic structure, much larger ordered moment, and detectable by bulk and local probes

6 Magnetic Phase Diagram of U(Pt,Pd) 3 Þ Small-moment AF order for x 0.01 Þ Large-moment AF order for x 0.08 Þ Superconductivity suppressed at x SC ~ Keizer et al., PRB, 1999 de Visser et al., PRL, 2000

7 Phase diagram: detailed for x < x c 1.6 Temperature (K) Temperature (K) SC Weak AFM QCP? AFM Pd Concentration (at. %) M. J. Graf et al., Physica B 284-8, 1281 (2000).

8 DC Magnetization Measurements Magnetization shown to be linear up to B = 9 T T SF χ(h, T constant ) χ(h constant, T)! mol = µ 0M mol B

9 Hydrostatic pressure increases c/a Pd, Th decrease c/a (Pd shrinks lattice, Th expands it) Pd and Th provide negative chemical pressure 24 Pd Th Pressure T SF (K) 20 Doping Pressure Variation of c/a tunes magnetism Qualitative agreement: Pd (and Th) doping, Hydrostatic Pressure and the change in the c/a ratio 10 3!(c/a)/(c/a) From Opeil et al., Physica B, June 2002

10 Normal State Resistivity of U(Pt 1-x Pd x ) < T< 300 K T N Resistivity (µohm-cm) x = x = U(Pt 1-x Pd x ) 3 B = 0 Tesla Rho-0.0% Rho-0.4% Rho-0.5% Rho-0.6% Rho-0.7% Rho-0.8% Rho-1.0% Rho-1.4% Rho-2.0% Temperature (K)

11 Electronic Transport/Resistivity Direct Evidence T N vanishes abruptly between x = in resistivity, but µsr says no! 1.6 Temperature (K) Pd Concentration (at. %) data scaled for comparison x=0.050 data van Sprang (Thesis UvA, unpublished)

12 Is there any evidence for quantum critical behavior near x c? FL α = 2 NFL α 2, 3Dim - AFM α = 3/2 Look for deviations from T 2 behaviors Power law resistivity: ρ = ρ 0 + AT α 2.0 x c Exponent! x Pd Exponent reaches minimum value near 1.5 at concentration near x c α eff = dln(ρ ρ 0 )/dlnt constant at low T -- true power law From Opeil et al., Physica B (2002)

13 NFL Behavior? Disorder & Impurity Effects AFM 3D, α EFF = 3/2 A. Rosch, Physical Review Letters, 82, 4280 (1999). Opeil and Graf, Physica B (2002)

14 Can we infer a QCP from resistivity data? Evidence for QCP 1) Decrease of temperature dependent exponent α to near 1.5 2) T N 0 K at x c = ) Power law behavior of x c = consistent with a strongly disordered NFL (Rosch). No bump upon entering the FL regime. Evidence against QCP 1) Exponent for pure UPt 3 is only 1.75, varies from sample to sample 2) No perfect agreement with Temperature dependence in Rosch model 3) Disorder can cause loss of coherence of spin fluctuations α = 2, α 3/2 Riseborough, PRB 29(84): SF scattering and disorder interplay...loss of SF coherence Lui et al., PRB 61(00): NFL arises from disorder...e.g. CeRhRuSi 2 Conclusion: Evidence is ambiguous from Electronic Transport!

15 U(Pt,Pd) 3 Polycrystal (6T) SCM- Measurements graphs scaled for clarity T N vanishes between x = and 0.010

16 U(Pt,Pd) 3 Single Crystal SCM- Measurements Opeil et al., JMMM (2004). Except by neutrons, No T N for x = single crystal B = 6 T

17 NFL behavior in x c = 0.006? Known NFL/QCP behavior in CeCu 5.9 Au 0.1 Field B = 0 T Field B = 3 T NFL behavior associated with a QCP manifests itself by an upturn in the magnetic susceptibility i.e. χ(t) - βt 1/2 (3D behavior) or - β lnt (2D behavior) as T N 0 K.

18 Nothing special happens at x = Specific Heat of U(Pt 1-x Pd x ) 3

19 Specific Heat of U(Pt 1-x Pd x ) 3 Nothing special happens at x = AF QCP (3D) Divergence as T 0 K γ ~ -ln(t/t 0 ) in Specific Heat

20 Based on the SCM and specific heat data, we believe the on set of LMAF is a crossover behavior not the result of a true phase transition. Detection depends on the timescale of the probe Y. Okuno and K. Miyake, J. Phys. Soc. Japan 67 (1998).

21 Very broad transition into conventional AFM $ 2 G(t) = A 1 & 3 e!" 1 t cos # t % ( ) e!" 1 ' t Magnetic ' ) + A ( 2 G KL (t) + A 3 G KT (t) Paramagnetic or weak AFM Fractional magnetic signal = (A 1 +A 2 )/(A 1 +A 2 + A 3 ) 1% Pd Fractional Magnetic Signal x = x = x = % Th Temperature (K) Th - solid lines Pd - dashed lines Graf et al., Phys. Rev. B 68, (2003)

22 Why is the transition region so broad? Speculation: Th substitution slows down magnetic fluctuations of weak AFM Observable in µsr at 7 K Fluctuation Timescale [Sec] 1% Th mag. susc. NMR neutron specific heat ZF-µSR scattering 1% Pd Pure UPt 3 Y. Okuno and K. Miyake, J. Phys. Soc. Japan 67 (1998). time domains

23 Conclusions: 1) The change in T SF quantitatively correlates to the change c/a ratio. 2) Resistivity behavior is consistent with NFL, but not distinguishable from disorder effects. 3) No confirmation of NFL nor proposed QCP found by Cantilever Magnetometry or Specific Heat at x = NFL behavior arised from magnetic instability and SF scattering. 4) AFM ordering (LMAF) at T N is not observed for x = by bulk magnetization for U(Pt,Pd) 3 single and polycrystals. 5) We believe that the transition from SMAF to LMAF is a crossover behavior. Pd slows down the temporal fluctuations of the SMAF phase. Future Work: (Graf Dilution Refrigerator: Boston College) 1) Specific Heat and Cantilever Magnetometry (0.010 < T < 1 K) on U(Pt 1-x Pd x ) 3 0 < x < to examine the affect of doping on the new AFM ordering reported at (~ 18 mk) by Schubeth et. al. (c/t vs T) and Schöttl et al.(τ vs T). 2) Hydrostatic Pressure studies on U(Pt 1-x Pd x ) < x < single crystals to examine the possibility of re-entrant (unconventional) SC and impurity effects about the AF magnetic instability.

24 With special thanks to J. L. Smith: Happy 65 th Birthday!