Absorption Refrigeration Using Ammonia-Water: Design, Experimentand Comparative Evaluation

Absorption Refrigeration Using Ammonia-Water: Design, Experiment and Comparative Evaluation G. Jaimes Hern´andez a aUniversidad Tecnol´ogica de San Juan del R´ıo Quer´etaro, Quer´etaro, M´exico Abstract This paper proposes, designs, builds and experimentally evaluates a small-scale solar-driven absorption refrigerator that uses ammonia (NH3) as refrigerant and water (H2O) as absorbent. The experimental campaign compares three energy-supply configurations: (i) grid connection (CFE), (ii) an off-grid photovoltaic (PV) system, and (iii) a Fresnel-based photothermal concentrator. Key metrics (generator and evaporator temperatures, coeffcient of performance (COP), stability) are presented and discussed. Major conclusions and practical recommendations for rural vaccine/medicine storage applications are provided. Key words: Solar refrigeration, absorption, ammonia-water, COP, photovoltaic, photothermal, Fresnel lens, experimental prototype. Introduction Global demand for refrigeration continues to grow, making sustainable off-grid technologies increasingly important for remote communities and for ensuring reliable medical cold–chain conditions. In terms of storage performance and system reliability, various technologies are still undergoing significant development [1–3]. This work presents an absorption refrigeration system based on the ammonia-water pair, chosen due to ammonia’s low Global Warming Potential (GWP) and its compatibility with thermally driven refrigeration cycles [4]. Solar–assisted absorption systems using the NH3–H2O working pair have been extensively studied in previous research [5,6]. Likewise, numerous studies emphasize their suitability for preserving vaccines and medical supplies in off-grid locations [7–10]. In parallel, advances in material science reveal that several graphene-based surfaces exhibit strong adsorption capabilities for molecules such as CO, CO2, and N2[11,12]. While these investigations stem from a different scientific domain, their insights into adsorption mechanisms can inform future strategies for refrigerant storage, purification, or capture. Similarly, studies involving oxiborate compounds [13–17] and trititanium pentoxide structures [18, 19] provide valuable perspectives on complex solid-state frameworks where molecular interactions can be engineered with precision. Furthermore, research on magnetic interactions and their effects on dissociated hydrogen adsorption [14, 20–30] broadens the scientific context in which functional materials may eventually contribute to innovative refrigeration, adsorption-based cooling, or energy-harvesting technologies. The structure of this paper is organized as follows. The Methodology section describes the experimental design and operational configuration of the ammonia–water absorption refrigeration prototype, including the system layout, the different energy–supply configurations, the instrumentation scheme, and the data–acquisition strategy. It also outlines the analytical procedures used to compute performance indicators and assess thermal behavior under variable operating conditions. The Results and Discussion section provides a comprehensive evaluation of the systems thermal response, performance stability, and coefficient of performance (COP) across the three operating modes, discussing both qualitative trends and quantitative results. Finally, the Conclusions section synthesizes the main contributions of this work, highlights the novelty and practical relevance of the prototype, and proposes potential directions for future optimization and real–world implementation. 1. Methodology 1.1. System Overview The prototype implements a conventional single–effect ammonia–water absorption refrigeration cycle composed of Preprint submitted to Elsevier 21 November 2025 a generator, condenser, expansion device, evaporator, and absorber. Ammonia (NH3) is used as the refrigerant due to its low saturation temperature and high latent heat, while water (H2O) serves as the absorbent [31]. In the generator, external thermal input desorbs NH3 vapor from the rich solution. The vapor subsequently condenses, expands through an expansion element, and evaporates to deliver cooling in the evaporator. Low–pressure vapor leaving the evaporator is absorbed by the weak solution in the absorber, closing the thermodynamic cycle [32]. Schematic diagrams, fabrication procedures, and material selections are detailed in [33]. All heat-exchange elements were built using locally sourced materials, including black carbon-steel tubing, aluminum finned evaporator elements, a 70 W electrical resistance heater (used both for grid and off-grid operation), and fittings compatible with ammonia exposure. The evaporator employs aluminum heat-sink structures to enhance surface area and support the required two–phase flow during phase change. 1.2. Energy Supply Configurations Three alternative energy-supply configurations were implemented to compare operational stability, efficiency, and feasibility: 1.2.1. Grid–Supplied Electrical Heating (CFE) Mains electricity from the Comisin Federal de Electricidad (CFE) powered the 70 W resistance heater and the acquisition electronics. This mode served as a baseline for stable, uninterrupted operation. 1.2.2. Off–Grid Photovoltaic (PV) System The off–grid PV configuration consisted of three 325 W polycrystalline modules connected in series, a 12/24 V MPPT/PWM charge controller, two 12 V lead-acid batteries, and a 750 W inverter. Solar energy supplied the heater and control electronics while battery storage mitigated short-term irradiance fluctuations. This configuration demonstrated the highest stability under field conditions [34]. 1.2.3. Photothermal Fresnel Concentrator A Fresnel lens (300 mm diameter, 400 mm focal length) concentrated solar radiation directly onto a heat absorber coupled to a heat pipe, transferring thermal energy to the generator without electrical conversion. Although this avoided conversion losses, it presented significant thermal variability due to fluctuating solar intensity [35]. 1.3. Instrumentation and Data Acquisition The system was fully instrumented to capture thermal behavior across key components: – One K–type thermocouple at the generator for high– temperature measurement. – Eighteen DS18B20 digital temperature sensors positioned at the condenser, evaporator, absorber, and refrigerated compartment. – One DHT11 sensor for ambient temperature and humidity. An ESP32 microcontroller, complemented by an Arduino Uno, managed data acquisition and wireless transmission to the ThingSpeak cloud platform at a sampling rate of 15 s. Firmware, wiring diagrams, and calibration procedures are provided in [33]. 1.4. Experimental Procedure and Data Analysis Temperature measurements were recorded at the generator, condenser, evaporator, and refrigerated compartment under representative environmental conditions for all three supply configurations. The following analyses were conducted: (i) Instantaneous Coefficient of Performance (COP). The COP was computed as: COP = ˙ Qevap ˙ Qgen , where ˙ Qevap is the cooling load at the evaporator and ˙ Qgen is the thermal input to the generator. Thermal estimations were obtained from temperature measurements, heater power, and component thermophysical properties [31]. Uncertainty propagation is documented in [33]. (ii) Correlation Analysis. Statistical correlations were computed among generator temperature, evaporator temperature, and COP to identify performance thresholds, thermal instability regions, and transient responses. (iii) Comparative Evaluation Across Supply Modes. Time–averaged COP, standard deviation, operational stability, and temperature–swing amplitudes were compared for the grid, PV, and photothermal configurations. All data processing, visualization, and statistical analysis were carried out using Python. 2. Results and Discussion 2.1. Thermal Behavior of the Generator and Evaporator The generator temperature exhibited distinct trends depending on the energy–supply configuration. Both the CFE grid mode and the PV system provided relatively stable and sustained thermal input, enabling consistent desorption of NH3. As generator temperature increased, desorption efficiency initially improved, producing higher vapor flow rates and promoting effective refrigeration. However, 2 beyond a practical upper threshold-identified in the prototype around 500 ◦C-performance degraded due to non– ideal heat transfer, thermal losses, increased solution imbalance, and the onset of thermodynamic inefficiencies associated with extreme temperature gradients. The evaporator temperature closely followed the dynamic behavior of the generator. Stable generator heating corresponded to steady evaporation near the design temperature, while rapid temperature swings (particularly in the photothermal mode) resulted in inconsistent evaporator cooling. These findings highlight the sensitivity of the NH3-H2O absorption cycle to fluctuations in thermal input and underline the importance of maintaining generator operation within an optimal thermal band. 2.2. Comparison of COP Across Supply Modes A systematic comparison of the three supply configurations revealed clear differences in achievable COP and operational stability: Photovoltaic system: Exhibited the highest average COP and the lowest variability. Battery buffering played a central role by smoothing out irradiance fluctuations, enabling nearly continuous heating of the generator with minimal thermal oscillation. Grid (CFE): Delivered intermediate COP values. Although the grid provides uninterrupted electrical power, the lack of thermal regulation occasionally led to overheating events at the generator, triggering a sharp decline in COP once the upper threshold was exceeded. Photothermal Fresnel concentrator: Produced the lowest mean COP and the largest variance. The absence of solar tracking caused the heat flux to fluctuate significantly, especially under partially cloudy conditions or during imperfect alignment. While peak heating was sometimes higher than in the other modes, the extreme intermittency resulted in unstable cycle operation. These results clearly indicate that not only the magnitude but also the temporal quality of the heat input governs absorption–cycle performance. 2.3. Operational Observations and Cycle Stability Field observations reinforce the quantitative trends obtained from the data: The PV configuration benefitted from electrical storage, producing highly stable generator temperatures and near– constant evaporator cooling. This translated into the most reliable COP throughout the experimental campaign. The photothermal Fresnel setup, despite its simplicity and potential for low–cost deployment, exhibited the strongest dependency on environmental variability. Without tracking, the focal point drifted rapidly, and momentary cloud cover produced immediate thermal drops that destabilized the absorption process. Grid–based operation was mechanically stable and easy to control, but thermal runaway became a risk at high generator temperatures, degrading cycle efficiency when not paired with thermostatic or power-modulation mechanisms. The thermal inertia of the working fluids played a moderating role, but was insufficient to overcome the rapid fluctuations observed in the photothermal mode. 2.4. Interpretation of Performance Differences The superior performance of the PV configuration can be attributed to two synergistic factors: (i) the steady electrical output from the battery bank, and (ii) the inherently controllable nature of resistive heating. Combined, these factors maintain the generator within a favorable operating window where the strong and weak solutions remain in balanced circulation, minimizing solution crystallization risks and maximizing evaporation efficiency. Conversely, the photothermal system lacked both tracking and thermal buffering. Direct solar concentration, while capable of delivering high peak temperatures, introduces intermittency that disrupts the delicate equilibrium necessary for stable NH3desorption and absorption. Any interruption in solar input immediately reduces generator temperature, causing vapor production to drop and lowering the COP. Grid operation illustrated a distinct limitation: although power is continuous, thermal control is not intrinsic. Without modulation, the generator overshoots optimal temperature ranges, producing excessive vapor pressure and degrading the thermodynamic balance of the absorption cycle. 2.5. Material, Safety, and Practical Considerations The use of ammonia required careful material selection due to its corrosive nature. Black steel tubing demonstrated compatibility with NH3, while aluminum was reserved for finned evaporator assemblies where thermal conductivity dominated material selection criteria. Safety considerations included leak minimization, adequate sealing of joints, and sensor placement to detect abnormal temperatures or potential system failures. For real–world applicationsparticularly for medical cold chains or rural refrigeration-robustness, safety certification, and compliance with 2–8 ◦Cstorage requirements are essential. The results indicate that, with proper thermal control and electrical storage, the prototype can feasibly achieve these standards. 3. Recommendations Based on the experimental evaluation and field observations, the following recommendations are proposed: 3 Prioritize PV + battery systems for off–grid rural deployment, as they yield the highest COP stability and are largely independent of momentary environmental fluctuations. Enhance photothermal systems with solar tracking and thermal energy storage tanks to mitigate intermittency and reduce COP variability. Incorporate thermal regulation mechanisms in grid– powered configurations to prevent overheating and maintain generator operation within the optimal thermal window. Improve generator and absorber heat-exchanger designs, potentially using advanced fin geometries, improved thermal surface treatments, or novel working–fluid enhancements to raise overall system efficiency. Conduct long–duration stability tests and evaluate costbenefit performance for real cold–chain applications, ensuring compliance with medical refrigeration standards. Details of the results, diagrams, and figures can be found in reference [33]. Since this text is a preprint, the author has chosen not to include them explicitly. 4. Conclusions This work presents a fully functional small–scale ammonia-water absorption refrigeration prototype powered by three distinct energy–supply configurations: grid electricity, an off–grid photovoltaic system with battery storage, and a photothermal Fresnel concentrator. Through extensive experimentation, the study demonstrates that the PV system offers the most stable and efficient performance, achieving the highest average COP and the lowest operational variability. The grid configuration, while reliable in terms of power availability, requires thermal regulation to prevent generator overheating. The photothermal configuration exhibited promising peak temperatures but suffered from severe fluctuations due to the lack of solar tracking and thermal buffering, resulting in reduced and inconsistent COP values. Overall, the results indicate that integrating energy buffering, thermal regulation, and improved heat– exchanger design can significantly enhance the viability of solar–driven absorption refrigeration for rural and off–grid applications. The practical recommendations derived from this work establish a clear roadmap for future prototype optimization, field deployment, and potential application in critical sectors such as vaccine and medicine refrigeration. Acknowledgements This preprint was prepared from the thesis document ′′Refrigeraci´on Solar′′ (Master in Applied Technology, Universidad Tecnol´ogica de San Juan del R´ıo Quer´etaro, April 2025) [33]. I acknowledge supervisors for the detailed experimental data and supporting materials. References [1] E. Vallejo. An hydrogen adsorption study on graphene-based surfaces with core–shell type catalysts. Carbon Letters, pages 1–10, 2023. [2] E. Vallejo. Hydrogen adsorption and dissociation, together with a multifractal transition into graphene.based surfaces. International Journal of Hydrogen Energy, 77:926–935, 2024. [3] Y. Salazar Flores and E. Vallejo. A study of hydrogen adsorption on graphene-based surfaces using copula models. International Journal of Hydrogen Energy, 170(150926):1–10, 2025. [4] Jungho Kim and Carlos Infante Ferreira. Ammonia as a refrigerant: A review of environmental and thermodynamic benefits. International Journal of Refrigeration, 84:1–13, 2017. Discusses low GWP and environmental benefits of ammonia. [5] Keith E Herold, Reinhard Radermacher, and Sanford A Klein. Absorption chillers and refrigeration systems. CRC Press, 1996. Classic reference on absorption cycles, NH3-H2O fundamentals, and solar-driven configurations. [6] Zhi Li and K Sumathy. Review of solar thermal-driven absorption cooling technologies. Renewable and Sustainable Energy Reviews, 30:667–684, 2014. Detailed overview of solar absorption systems including NH3-H2O. [7] Muhammad Saghir and Waqar Qureshi. Solar-powered refrigeration systems for vaccine storage: A review. Renewable and Sustainable Energy Reviews, 82:1012–1028, 2018. Refrigeration solutions for vaccine and medical cold chains in off-grid regions. [8] Sipho Nkomo and Philemon Chauke. Sustainable cold-chain solutions for remote medical facilities: A review. Energy for Sustainable Development, 57:165–175, 2020. Cold chain requirements and solar refrigeration for medicine and vaccine storage. [9] Bent Srensen. Solar-powered cooling for the developing worlda review. Solar Energy, 76(1):103–110, January 2004. Survey of solar absorption/adsorption for remote applications. [10] mer Kaynakli and Ilhami Horuz. Solar powered absorption cooling systems for remote areas. Renewable Energy, 30(12):2021–2031, 2005. Absorption cooling for off-grid, medicine/vaccine storage applications. [11] E. Vallejo and P. A. L´opez-P´erez. Magnetic transitions in CO adsorption on graphene–based surfaces. Physics Letters A, 425(127883):1–8, 2022. [12] E. Vallejo and P. A. L´opez-P´erez. Strong chemical adsorption of CO2and N2on a five-vacancy graphene surface. Solid State Communications, 356(114934):1–6, 2022. [13] E. Vallejo and M. Avignon. New metastable phases in an oxyborate compound obtained by an evolutionary algorithm and density functional theory. Journal of Magnetism and Magnetic Materials, 435:33–39, 2017. [14] E. Vallejo. charge and magnetic ordering a l’etude de l’ordre de spins, de charge et effects structuraux dans le modele de doubleechange: chaine lineaire, echelles de spins et mangnaites. ´edilivre Universitaire, Paris, France, 2008. [15] E. Vallejo and Avignon M. Charge and magnetic ordering in a ladder system: the Fe–Ludwigite. Revista Mexicana de F´ısica S, 53(7):1–6, 2007. [16] E. Vallejo and M. Avignon. Spin ordering in three–leg ladders in ludwigite systems. Journal of Magnetism and Magnetic Materials, 310(2):1130–1132, 2007. [17] E. Vallejo and M. Avignon. Spin and Charge Ordering in Three–Leg Ladders in Oxyborates. Physical Review Letters, 97(217203):1–4, 2006. [18] E. Vallejo and D. Olgu´ın. New metastable phases in a trititanium pentoxide compound. Materials Research Express, 2(126101):1– 10, 2015. 4 [19] D. Olgu´ın, E. Vallejo, and A. Rubio-Ponce. Ab initio study of the low–pressure phases of Ti3O5.Phys. Status Solidi B, 252(4):659–662, 2015. [20] Y. Arredondo, O. Navarro, E. Vallejo, and M. Avignon. Analysis of Spin–polaron Formation in Hund Lattices. Journal of the Korean Physical Society, 62(10):1504–1507, 2013. [21] Y. Arredondo, E. Vallejo, O. Navarro, and M. Avignon. Formation of spin-polarons in the ferromagnetic Kondo lattice model away from half-filling. J. Phys.: Condens. Matter, 24(335601):1–6, 2012. [22] O. Navarro, E. Vallejo, and M. Avignon. Spin-polarons in an exchange model. International Journal of Modern Physics B, 26(09):1–12, 2012. [23] E. Vallejo, O. Navarro, and M. Avignon. Double and super-exchange model in one-dimensional systems. Journal of Magnetism and Magnetic Materials, 322(09–12):1076–1078, 2010. [24] J. R. Su´arez, E. Vallejo, O. Navarro, and M. Avignon. Magneto– elastic phase transition in a linear chain within the double and super–exchange model. Journal of Magnetism and Magnetic Materials, 322(09–12):1064–1068, 2010. [25] E. Vallejo. Magnetoelastic effect in an exchange model. Journal of Magnetism and Magnetic Materials, 321(06):640–643, 2009. [26] J. R. Su´arez, E. Vallejo, O. Navarro, and M. Avignon. Magneto– elastic phase transitions in one–dimensional systems. J. Phys.: Condens. Matter, 21(046001):1–8, 2009. [27] J. R. Su´arez, E. Vallejo, O. Navarro, and M. Avignon. Monte Carlo study of the double and super–exchange model with lattice distortion. Journal of Physics: Conference Series, 167(012068):1–5, 2009. [28] E. Vallejo, F. L´opez-Ur´ıas, O. Navarro, and M. Avignon. Magnetic polaron structures in the one–dimensional double and super–exchange model. Solid State Communications, 149(3– 4):126–130, 2009. [29] E. Vallejo, F. L´opez-Ur´ıas, O. Navarro, and M. Avignon. Ferromagnetic polarons in the one–dimensional double and super–exchange model. Journal of Magnetism and Magnetic Materials, 320(14):e425–e427, 2008. [30] E. Vallejo. Lattice distortion in the one-dimensional double and super–exchange model. Microelectronics Journal, 39(11):1266– 1267, 2008. [31] Yunus A. Cengel and Michael A. Boles. Thermodynamics: An Engineering Approach. McGraw-Hill, New York, 8th edition, 2015. [32] J. M. Arroyo Rosa. Introduccin a la refrigeracin por absorcin, n.d. Technical monograph. [33] Gildardo Jaimes Hern´andez. Refrigeraci´on solar mediante sistema de absorci´on amoniaco-agua. Master’s thesis, Universidad Tecnol´ogica de San Juan del R´ıo, Quer´etaro, M´exico, 2025. [34] Risen Energy. Polycrystalline 325 w solar module: Technical datasheet, 2019. Product datasheet. [35] Optical Components Inc. Fresnel lens for solar concentration: Technical specifications, 2020. Datasheet for 300 mm Fresnel concentrator. 5