Quality & Management Category And Dr. Allen V. Astin Award for Best Paper

NACLA – Who Shot my Horse?

Dr. Malcolm Smith
WESCAN Calibration Services

Applied Category

Irradiance Responsivity Scale Realization Between 1000 nm and 2500 nm

Irradiance measuring radiometers based on InGaAs and extended-InGaAs photodiodes have been developed at the National Institute of Standards and Technology (NIST) to hold and propagate the spectral irradiance responsivity scale between 1000 nm and 2500 nm. First, the optical and electronic design considerations of these working standard radiometers will be discussed. The design of the radiometer head, including the housing and the arrangement of the optical and electronic components will be described. The photodiodes are temperature controlled to -30 degree C and -20 degree C, respectively to increase their shunt resistances. Electronic characteristics, such as photocurrent dynamic range and photocurrent-to-voltage conversion uncertainty versus photodiode shunt resistance will be discussed. After fabrication and characterization, the two radiometers were calibrated for spectral irradiance responsivity against a transfer standard Electrically-Substituted Bolometer (ESB). The calibrations were made on the new Infrared Facility for Spectral Irradiance and Radiance Responsivity Calibrations using Uniform Sources (IR-SIRCUS). The ESB is traceable to the primary standard cryogenic radiometer (HACR) through a silicon trap detector that can measure both radiant power and irradiance with 0.15 % expanded uncertainty (k=2) in the visible and near-infrared ranges. The irradiance responsivity scale realization and propagation to the transfer and working standard radiometers will be described. An irradiance responsivity tie point was also derived from an irradiance measuring InSb radiometer standard to verify the irradiance responsivity scale. The InSb radiometer was also calibrated against the HACR using another cryogenic bolometer and three different transfer standard detectors. The calibrated InGaAs and extended-InGaAs radiometers also can be used to disseminate the irradiance responsivity scale to other institutions and facilities.

Invited Category

Refractrometry Using A Helium Standard

Jack Stone
NIST

The refractive index of helium at atmospheric pressure can be calculated from first principles with a very low uncertainty, on the order of 10^-10. Furthermore, the low refractive index of helium puts minimal demands on the pressure and temperature measurements required to determine the refractive index of a given sample of helium gas. Therefore helium can serve as a practical, theory-based standard of refractive index that might be used in place of air for ultra-high accuracy interferometric length measurements. Because its index of refraction is known, helium can also be used to characterize and correct errors in a gas refractometer. We have built two refractometers based on a laser locked to the transmission maximum of a Fabry-Perot cavity, and we use helium to measure and correct pressure-induced distortions of the refractometers. (We have also characterized other sources of error in our Fabry-Perot refractometers, such as errors associated with the effect of humidity on the mirror coatings.) As a proof-of-principle of the helium-correction technique, we have used our refractometers to measure the molar refractivity of nitrogen and we find reasonable agreement with previous measurements. When our two refractometers simultaneously measure the refractive index of a common nitrogen sample, we find that the two systems agree with each other within a few parts in 10^9. The good agreement suggests that many potentially troubling sources of uncertainty can be overcome. These measurements are sufficiently encouraging that one might speculate on the possibility of developing pressure standards based on refractometry.

Theoretical Metrology

Beyond Error Bars

Alan Steele
NRC Canada

Among the many tasks facing metrology today is the conduct and analysis of measurement comparisons. At the international level, the goal is to quantify the degrees of equivalence among independent realizations of the system of units and the primary measurement techniques in each field. These key comparisons are summarized in published tables that relate each participating laboratory result to a Key Comparison Reference Value (KCRV) that is, in general, an aggregate statistical estimator such as the mean (weighted uniformly or otherwise) or the median of the quoted laboratory results (both the values and the standard deviations). We present a simple technique for displaying and interpreting the population of measurement results, and for evaluating different statistical approaches to determining when and when not to use a reference value that is calculated from the participants’ data sets. Monte Carlo methods are exploited as a fast and convenient tool to create the distributions for any candidate reference value being considered as the representative statistic for summarizing a measurement comparison without the a priori assumption of data pooling in a single normal population.