Fusion Science and Technology / Volume 76 / Number 3 / April 2020 / Pages 284-290
Technical Paper / dx.doi.org/10.1080/15361055.2020.1711688
Articles are hosted by Taylor and Francis Online.
For the fuel cycles of fusion power plants, highly specialized in-line analytic systems are crucial for efficient process control, monitoring, and accountancy. One of these systems under development is infrared (IR) absorption spectroscopy of liquid hydrogen isotopologue mixtures that can be used for in-line process control and monitoring of cryogenic distillation. The main challenge of this method is the complex calibration procedure since the integral IR absorption strength is nonlinearly correlated with the isotopologue composition. Typical calibration procedures make use of well-known samples produced by mixing atomic pure samples and referenced by p-V-T-measurement. The samples are catalyzed to produce samples containing heteronuclear molecules. By this procedure, one cannot exceed the chemical equilibrium of high temperatures (mass action coefficient Kc<4). Therefore, it is not possible to produce samples with an HD, HT, or DT concentration above 50% by catalysis or natural equilibration. However, in isotope or isotopologue separation, such as in cryogenic distillation, this equilibrium will be regularly exceeded. In the case of IR absorption spectroscopy on liquid hydrogen isotopologues, additional care needs to be taken for calibration since the calibration functions are highly nonlinear. We tested our calibration in the high-purity HD regime (Kc>4) by producing a sample via cryogenic distillation and performing a cross calibration for three systems: Quadrupole mass spectrometry, Raman spectroscopy, and infrared spectroscopy. Therefore, we can also demonstrate that additional calibration points are indispensable in order to improve the systematic uncertainties below the 5% level, and a simple extrapolation from a calibration of Kc < 4 to Kc > 4 will result in a trueness and accuracy exceeding this 5% level.