Experimental studies on small diameter carbon dioxide evaporators for optimal Silicon Pixel Detector cooling

Since recent years the Large Hadron Collider at CERN and its experiments are the subject of upgrade programs, which are necessary to increase the foreseen collision rates and the amount of data to be gathered for the particle physics community in the future. In order to cope technologically with long operation times under high radiation levels, the LHC experiments have to upgrade or fully replace some of the most crucial detector components. To improve their resolution and read-out rate capabilities, this is especially important for the Silicon Pixel Detectors, which are installed at the very center of the large detector units at CERN. In parallel with the implementation of new and more compact detection and read-out technologies, the request for highly-effective and integrated detector cooling is getting more demanding. Due to its superior performance compared to standard refrigerants, the thermal management of many Silicon Pixel Detectors at CERN relies on boiling carbon dioxide inside compact heat exchangers of small hydraulic diameter. To allow for an optimal design of the applied cooling method and to safely operate the highly sensitive particle detectors, some of the unknowns related to carbon dioxide flow boiling in small channels must be resolved to develop new predictive methods, both for the pressure drop and heat transfer coefficient. Due to the current lack of suitable predictive models, a long-term study has been launched to create a consistent and reliable experimental data base studying the peculiarities of boiling carbon dioxide in mini- and micro-channels. By means of a new experimental setup for detector cooling R&D, various small-scale carbon dioxide evaporator layouts can be analyzed. Three dimensions of small-scale tubular evaporators in stainless steel and a multi-micro-channel cold plate embedded into silicon have been characterized for this study. By means of a parametrical characterization with high-end pressure and temperature sensors and flow visualization with a high-speed camera, results from the basic tubular single-channels can complement the findings from the multi-micro-channels and vice versa, thus creating a large and multifaceted data base. Since the thermal management of high energy physics experiments is in need for a continuous operation in the temperature range from +15 to -30 degree Celsius or even lower, the data presented focus on the influence of the saturation temperature on the two-phase pressure drop and heat transfer, whilst flow visualizations obtained for the multi-micro-channels can provide a consistent key of interpretation for the parametrical analysis. The combination of the shifting, temperature-dependent physical properties of carbon dioxide and different flow confinement conditions causes a change in the phenomenological behaviour of the flow and a transition between macro- and micro-scale flow behaviour most likely occurs within the range of test parameters. Furthermore a shift in the applicability of existing prediction methods is caused by those effects and no correlation for heat transfer and pressure drop is able to predict the experimental data and trends in the whole temperature range. A selective approach to the use of existing correlations is proposed. Thus while new comprehensive prediction models are being developed based on the data gathered for this study, at the same time some recommendations for the temperature dependent use of already existing models can be provided. This allows for an optimized design of new Silicon Pixel Detector cooling systems, today or in the very near future.