Abstract
An accurate metrological investigation was performed on several units of the two de facto standards for deep-ocean temperature measurements—the SBE35 (the reference thermometer) and the SBE3 (the working thermometer) from SeaBird Scientific. Four SBE35 units were repeatedly calibrated against state-of-the-art fixed-point cells (triple point of water at 0.01 °C and melting point of gallium at 29.7646 °C), with calibration uncertainties of approximately 0.2 mK and 0.5 mK, respectively. Three SBE35 units and one SBE3 unit were calibrated in the temperature range 0 °C to 30 °C, again with sub-millikelvin calibration uncertainties, in a recently-developed water–bath calibration facility. All these calibrations evidenced (1) the deviation of each unit from its original manufacturer’s calibration (up to 1.7 mK), which were found to be inconsistent with the manufacturer’s uncertainty claims and (2) unexplained irreproducibilities, which could not be attributed to the calibration system of up to 1.5 mK. The effect of high pressures (up to 60 MPa) on the response of two SBE35 units was investigated by pressurizing the SBE35 units inside a purpose-built compact pressure enclosure. The results of the pressure investigation confirmed the existence of a small device-dependent pressure effect (approximately 0.3 mK at 60 MPa) and the need for individual temperature–pressure calibration of each SBE35 unit.