Extended Data Fig. 1: Experimental details.
From: Oscillations of the thermal conductivity in the spin-liquid state of α-RuCl3
Panel (a) shows a schematic of the mounted crystal and the applied in-plane field b. The temperatures TA, TB and TC are read off RX102A thermometers as shown. Panel (b) shows a photo of Sample 2 contacted with thick Au wires (100 μm in diameter) to thermometers and bonded to the bath by stycast. The heater and Delrin post are visible at the top right and left, respectively. Panel (c) shows the Cl octahedra enclosing Ru ions (adapted from 20). The pink and blue planes are normal to the spin axes SZ and SX, respectively. The lower sketch highlights the plane containing the ‘Z’ Ru-Ru bond and the spin axis SZ. Panel (d) displays time-traces of the temperatures TA, ⋯ , Tbath in a λyx measurement with the bath temperature fixed at 280 mK. The field is gradually increased from -13.5 to 13.5 T over 14 hours by a step-wise change of (for example) 125 mT at each step. After each step-increase, transients caused by heating (or cooling) of the spins via the magneto-caloric effect combined with eddy-current heating of the brass bath are seen in all channels. The total transverse signal ΔyT is the difference between the red and black curves (as expressed in Eq. (2), ΔyT is the sum of δy and the ‘pick-up’ of the longitudinal ΔxT caused by contact misalignment). Panel (e) shows an expanded view of 5 transient pulses bracketing H = 0 (vertical dashed line). For t < 26,000 s, the 2 effects partially cancel whereas for t > 26,000 s, they add to give large transients. Readings are recorded within the blue-shaded interval after all transients have decayed. The average over the readings gives ΔyT to a resolution of ± 3 μK. Because of systematic errors, however, the total uncertainty in measuring δy is ± 200 μK.
