Comfortable, efficient, and affordable heating, ventilation, and air conditioning systems in buildings are highly desirable due to the demands of energy efficiency and environmental friendliness. Traditional vapor-compression air conditioners exhibit a lower coefficient of performance (COP) (typically 2.8–3.8) owing to the cooling-based dehumidification methods that handle both sensible and latent loads together. Temperature- and humidity-independent control or desiccant systems have been proposed to overcome these challenges; however, the COP of current desiccant systems is quite small and additional heat sources are usually needed. Here, we report on a desiccant-enhanced, direct expansion heat pump based on a water-sorbing heat exchanger with a desiccant coating that exhibits an ultrahigh COP value of more than 7 without sacrificing any comfort or compactness. The pump’s efficiency is doubled compared to that of pumps currently used in conventional room air conditioners, which is a revolutionary HVAC breakthrough. Our proposed water-sorbing heat exchanger can independently handle sensible and latent loads at the same time. The desiccants adsorb moisture almost isothermally and can be regenerated by condensation heat. This new approach opens up the possibility of achieving ultrahigh efficiency for a broad range of temperature- and humidity-control applications.
Dramatically improving energy efficiency with adequate humidity control and sufficient fresh air ventilation is a long-standing focus of heating, ventilation, and air conditioning (HVAC) systems1 because electric HVAC systems are significant contributors to energy shortages and environmental problems2,3,4, generally requiring a series of compromises among energy efficiency, thermal comfort, and utility cost. Although adequate fresh-air ventilation is necessary to ensure indoor air quality, increasing air exchange will decrease the sensible heat ratio (SHR, defined as sensible load divided by the total cooling load; air-conditioning loads typically contain 60% sensible load for cooling and 40% latent load for dehumidification). With a reduction in SHR, a traditional vapor-compression air conditioner must be operated at a lower evaporation temperature (5–7 °C), which adopts a cooling-based dehumidification method to handle both sensible and latent loads together. This results in lower cooling capacity and a lower COP (2.8–3.8, in general), and a reheating process is required to meet supply-air requirements when the SHR drops below 0.6 (which is typical of Summer nights and swing seasons when sensible gains are low)5.
To solve this problem, considerable efforts have been made to develop efficient alternative refrigeration systems such as caloric-based technologies6,7,8,9 (those using the magnetocaloric, elastocaloric, barocaloric, or electrocaloric effects) and innovative system designs such as temperature- and humidity-independent control10 (THIC). Although caloric-based technologies show great potential for energy savings, they are still far from being practically applied11. A THIC system usually consists of a thermally driven dehumidification unit (for example, a liquid or solid desiccant system) and a vapor-compression air conditioner specializing in handling the sensible heat load (such as variable-refrigerant flow technology). It was reported that THIC systems might save 25–50% of electrical consumption10,12,13,14 by adopting a higher evaporation temperature (e.g. 15–20 °C), and the corresponding electric COP increases by about 40–60% (refs 10 and 12) under different operating conditions compared to conventional HVAC systems. However, the COP (equal to latent load divided by primary thermal energy input) of current solid-desiccant systems is quite small, typically ranging from 0.5 to 1.0 (ref. 15) due to high regeneration temperature [usually in the range 60–120 °C (refs 15 and 16)]. Even though solar heat or industrial waste heat can sometimes be used, a large volume and high utility cost limits the large-scale application of solid-desiccant systems in residential buildings12.
It is quite natural to consider whether desiccant dehumidification could be used in an isothermal process, since, if adsorption heat could be easily taken away, then desiccant dehumidification will be very efficient, while desiccant regeneration can be handled by the waste heat from the air-conditioning system, for example, as condensation heat. In this case, it would be reasonable to have an evaporation temperature of 15 °C for sensible cooling and dehumidification and a condensation temperature of 45 °C for desiccant regeneration12. Then, a novel desiccant material having a sufficient water adsorption capacity difference under two-cycle conditions [such as 15 °C/80% relative humidity (RH) and 45 °C/30% RH] should be adopted. It is estimated that if the HVAC systems have 40–50% latent heat the evaporation temperature can increase from approximately 5–7 to approximately 15–17 °C and the condensation temperature can decrease from approximately 50–55 to approximately 40–45 °C, and then the COP can be nearly doubled.
Here, we report a novel concept for a desiccant-enhanced direct expansion heat pump (DDX HP) (Fig. 1a) based on the proposed water-absorbing heat exchanger (WSHE) fabricated by coating a desiccant on the surfaces of a conventional evaporator and condenser. In order to guarantee continuous operation, two of the same WSHEs will switch from evaporator (condenser) into condenser (evaporator) alternately (Fig. 1b). The sensible load is handled in the same way as before, by convection, but without overcooling (an evaporation temperature approximately 15–17 °C), and the desiccant coatings treat the latent load in a nearly isothermal way (Fig. 1c). Therefore, the processed air leaving the evaporator satisfies the requirement for supply air (Fig. 1d). In this DDX HP, the evaporation temperature rises while the condensation temperature drops compared to traditional air conditioners, because a WSHE does not need to cool the process air below its dew point to condense the moisture, and the adsorbed water evaporation strengthens the heat-dissipation capacity of the condenser. Therefore, a DDX HP shows a great potential of achieving much higher energy efficiency.
In fact, similar concepts of adsorption heat exchangers have been studied by several researchers13,17,18,19,20, but most were designed to dehumidify the air or to be used for adsorption refrigeration that focused on the mass-exchange process. To achieve the above-mentioned goals, four problems need to be solved: (1) novel desiccants that have an adequate moisture uptake capacity difference under two-cycle conditions should be invented; (2) sensible heat transfer should not be obviously weakened by the desiccant coated layer; (3) a matched control strategy with a DDX HP to achieve high efficiency at any mix of sensible and latent loads should be developed; and (4) the loss of sensible-load capacity due to operational mode switching should be minimized. This is the first time, to our knowledge, that a high-efficiency approach has been devised that inherits advantages of THIC, but also has the merits of compactness, low cost, and realization of the effective utilization of condensation heat.
Our concept of a water-adsorbing heat exchanger is illustrated in Fig. 1c, which is typically operated at 10–20 °C for cooling/adsorption and at 40–50 °C for heating/desorption. In this design, a desiccant-coated layer is adhered onto the fin surface of the base heat exchanger (here we use a fin-and-tube heat exchanger as an example) to form an integrated heat and mass exchanger. The desiccant-coated layer is based on a composite mesoporous, silica-gel-supported lithium chloride (CSGL) desiccant made from lithium chloride as the filler, owing to the appropriate deliquescence relative humidity and mesoporous silica-gel particles (50–100 mesh) as the matrix due to a large void volume. Such a composite desiccant can provide a small adsorption capacity at low relative humidity (less than 50% RH) and high adsorption capacity at intermediate relative humidity (in the range 50–80% RH) due to the combination of capillary condensation and solution absorption. Given that the evaporation temperature required for cooling often approaches the dew point and that the relative humidity on the fin surface of the condenser is generally below 30% RH, CSGL perfectly satisfies the requirement of a sufficient water adsorption capacity difference at two typical cycle conditions (for example, 15 °C/80% RH and 45 °C/30% RH). In addition, the water-borne compound adhesive used as a binder (approximately a particle radius thick) not only guarantees strong adhesion of granular desiccants to the coil, but also provides a protective coating to prevent chloride corrosion. Meanwhile, it is easier to avoid corrosion or carryover of liquid droplets by controlling the salt content21 and to prevent saturation or leakage by adjusting the duration of the moisture uptake process. In fact, the thickness of the desiccant-coated layer is the one important parameter that affects both the desiccant loading and the convection heat transfer. The optimal value that satisfies both criteria is approximately 10% of the fin space based on our experience. Details of the WSHE fabrication are given in the Methods section.
As a model system, the salt content of CSGL is 16.2 wt.%, and its isotherms are presented in Fig. 2a. It can be seen that the water content of CSGL at 15 and 45 °C under a specific relative humidity shows little difference, indicating that the moisture uptake capacity of CSGL is insensitive to temperature but mainly depends on the relative humidity. The adsorption capacity difference between a typical cooling/adsorption phase (80% RH) and a condensation/desorption phase (30% RH) is 0.34 kg kg−1, more than twice that of silica gel. The surface morphology looks similar to shattered glass and the globular silica gels are embedded in a solidified silica sol (Supplementary Fig. 1). Its thickness is estimated to be in the range 0.2–0.25 mm derived from the desiccant loading on unit area, 158.6 g m−2, considering that the density of 50~100 mesh granular silica-gel is 600~700 kg m−3. The altitude difference of the desiccant layer is about 0.05–0.15 mm indicated by variations of color shown in Fig. 2b, which also confirms the distribution of globular silica gels. The pore-size distributions of the desiccant-coated layer show double peaks around the points at 4 and 20 nm (Supplementary Fig. 2), with an average pore diameter 8.4 nm which is slightly smaller than that of mesoporous silica gel (10.4 nm). A rational assumption would be that large pores are partially filled or at least narrowed. The cumulative pore volume at 4 nm is very small, while that at 20 nm is relatively larger, which represents the solidified binder and globular silica gel, respectively. The specific surface area and pore volume are 143 m2 g−1 and 0.25 cm3 g−1, respectively. Other textural properties of CSGL film, such as thermal conductivity, specific heat capacity, and adsorption rate constant, were also measured as 5.27 W m−1 K−1, 1.2 kJ kg−1 K−1, and 1.0 × 10−3 s−1, respectively.