Thermochimica Acta 522 (2011) 60–65 Contents lists available at ScienceDirect Thermochimica Acta jou rnal h omepage: www.elsevier.com/locate/tca Optical calibration for nanocalorimeter measurements P. Swaminathan a,b , B.G. Burke b , A.E. Holness b , B. Wilthan c , L. Hanssen c , T.P. Weihs a,∗ , D.A. LaVan b,∗∗ a Department of Materials Science and Engineering, Johns Hopkins University, Baltimore, MD 21218, USA b Ceramics Division, Material Measurement Laboratory, National Institute of Standards and Technology, Gaithersburg, MD 20899, USA c Optical Technology Division, Physical Measurement Laboratory, National Institute of Standards and Technology, Gaithersburg, MD 20899, USA a r t i c l e i n f o Article history: Available online 23 March 2011 Keywords: Nanocalorimetry Nanocalorimeter Platinum emissivity Thin film Melting point Calibration a b s t r a c t We describe a method for calibration of nanocalorimeters from 573 K to 873 K, using resistive heating and optical pyrometry for temperature measurement. A platinum strip suspended on a silicon nitride membrane serves as both heater and temperature sensor. The calibration described here, relating resis- tance to temperature, enables subsequent temperature measurement. Measurements of the emissivity of as-deposited and annealed platinum thin films were also performed as a function of wavelength and temperature; these measurements are needed to correct the temperature recorded using the pyrometer. The calibration was validated by measurement of the melting point of a pure aluminum film; the melting point established with a Gaussian fit to the dT/dt data agreed within 0.1% (0.7 K) at 933.5 K. The melting point established from the minimum in the dT/dt plot agreed within 0.6% (5.5 K) at 933.5 K. Finite element modeling was used to characterize the temperature distribution. Published by Elsevier B.V. 1. Introduction Nanocalorimetry refers to chip based measurement of thermal and thermodynamic phenomena in materials with mass of the order of nanograms [1–3]. We are interested in using this tech- nique to study high temperature, rapid exothermic reactions in metallic multilayer systems [4]. The individual layers in these sys- tems have thicknesses on the nanometer scale and react to form stable intermetallic compounds. During reaction, adiabatic temper- atures (maximum film temperature in the absence of heat losses) of 1773 K can be reached with self propagating heating rates as high as 10 6 K/s [5]. In order to study thermophysical changes at such high temperatures and heating rates we make use of the nanocalorimeter. The nanocalorimeter chip consists of a silicon nitride (SiN x ) membrane supporting a Pt strip that serves as both a heater and a temperature sensor [6,7]. In order to calculate the temperature dur- ing an experiment, the Pt strip must be a priori calibrated to obtain its temperature coefficient of resistance (TCR). Previously, the cali- bration has been performed by heating the entire chip in a furnace and measuring resistance [8]; that approach provides the relation- ship between mean temperature (uniform within the limits of the furnace) and mean resistance. During an experiment, a temperature ∗ Corresponding author. ∗∗ Corresponding author. Tel.: +1 301 975 6121. E-mail addresses: weihs@jhu.edu (T.P. Weihs), david.lavan@nist.gov (D.A. LaVan). gradient exists in the chip and the dominant center temperature is higher than the mean temperature. When chips were calibrated in this manner, the difference between that central temperature and the mean temperature resulted in a temperature offset (error) that could be as large as 40 K at the melting point of aluminum. While there are different approaches to reducing the gradient, great efforts would be necessary to eliminate the temperature gradient and eliminate this source of error in the nanocalorimeter measure- ments. In the calibration strategy presented here, the central tempera- ture is measured and related to the mean resistance, resulting in no offset error in the temperature. We have used finite element mod- eling to measure the temperature distribution based on the current design of the nanocalorimeter and show that the central region is the dominant temperature (the central region at maximum tem- perature is approximately 50% of the sample area) at both short times and under steady state conditions. The existence of temper- ature gradients in the sample region is not ideal, and does lead to slight peak broadening, but with this calibration procedure, it does not result in a peak offset. The calibration strategy presented here also offers advantages as it does not require the long times necessary to achieve thermal equilibrium in a furnace; in addi- tion, furnace calibration temperatures cannot exceed 620 K without unpredictable electrical shorts through the SiN x under the electri- cal contact pads [8] that limits accuracy during high temperature measurements as the TCR has to be estimated by extrapolating from lower temperature calibration [7]. Here we describe a method to directly calibrate nanocalorime- ters at high temperatures (demonstrated to 873 K) by locally 0040-6031/$ – see front matter. Published by Elsevier B.V. doi:10.1016/j.tca.2011.03.006