The rapid growth of artificial intelligence (AI) and high-density computing workloads is placing increasing strain on conventional data-center cooling infrastructure. Although vapor-compression systems remain highly effective, they reject substantial quantities of low-grade waste heat directly to the environment, leaving it thermodynamically underutilized. This paper presents a conceptual systems-engineering architecture for a modular adsorption cooling unit designed for liquid-cooled data centers and intended to function as a supplemental bottoming cycle. The proposed system utilizes a finned flat-tube (FFT) aluminum heat exchanger coated with a 300 μm SAPO-34 (AQSOA-Z02) zeolite layer applied through a multi-stage slurry deposition and calcination process. The thin-film geometry is intended to reduce the heat- and mass-transfer limitations associated with conventional packed-bed adsorption systems. Based on thermodynamic assessment and extrapolated laboratory-scale coated-heat-exchanger data, the proposed architecture projects a theoretical specific cooling power (SCP) of approximately 400–450 W kg⁻¹ and a thermal coefficient of performance (COP) of 0.35–0.48 under optimized operating conditions, assuming preservation of laboratory-scale adsorption kinetics during scale-up. This study establishes a reproducible framework for future empirical validation while explicitly acknowledging engineering constraints including scale-up degradation, vacuum-maintenance complexity, thermomechanical fatigue, and exergy-efficiency limitations.
Yogesh Puranik (Tue,) studied this question.