A differential, membrane-based nanocalorimeter for general specific heat studies of very small samples, ranging from 0.5 mg to sub-μg in mass, is described. The calorimeter operates over the temperature range from above room temperature down to 0.5 K. It consists of a pair of cells, each of which is a stack of heaters and thermometer in the center of a silicon nitride membrane, in total giving a background heat capacity less than 100 nJ/K at 300 K, decreasing to 10 pJ/K at 1 K. The device has several distinctive features: (i) The resistive thermometer, made of a Ge(1 - x)Au(x) alloy, displays a high dimensionless sensitivity ∣dlnR∕dlnT∣ ≳ 1 over the entire temperature range. (ii) The sample is placed in direct contact with the thermometer, which is allowed to self-heat. The thermometer can thus be operated at high dc current to increase the resolution. (iii) Data are acquired with a set of eight synchronized lock-in amplifiers measuring dc, 1st and 2nd harmonic signals of heaters and thermometer. This gives high resolution and allows continuous output adjustments without additional noise. (iv) Absolute accuracy is achieved via a variable-frequency-fixed-phase technique in which the measurement frequency is automatically adjusted during the measurements to account for the temperature variation of the sample heat capacity and the device thermal conductance. The performance of the calorimeter is illustrated by studying the heat capacity of a small Au sample and the specific heat of a 2.6 μg piece of superconducting Pb in various magnetic fields.

To achieve accurate results in nanocalorimetry a detailed analysis and understanding of the behavior of the calorimetric system is required. There are especially two system-related aspects that should be taken in consideration: the properties of the empty cell and the effect of the thermal link between sample and cell. Here we study these two aspects for a membrane-based system where heater and thermometer are both in good contact with each other and the center of the membrane. Practical, analytical expressions for describing the frequency dependence of heat capacity, thermal conductance, and temperature oscillation of the system are formulated and compared with measurements and numerical simulations. We finally discuss the experimental conditions for an optimal working frequency, where high resolution and good absolute accuracy are combined.

No abstract