A Moisture-Penetrating Humidity Pump Directly Powered by One-Sun Illumination

Summary

There is broad demand for humidity control for industrial, commercial, and residential applications. Current humidity pumping technologies require intensive maintenance because of the complexity of their mechanical structures. Furthermore, indirect utilization of solar energy increases both cost and energy loss. Here, we demonstrate a new humidity pumping concept based on multilayer moisture permeable panels. Such panels, with a simple structure, may allow the penetration of moisture from indoor (adsorption) to outdoor (desorption) with little heat loss. One-sun illumination is introduced as the direct energy source. A proof-of-concept prototype is designed and established, successfully dehumidifying indoor air with the best dehumidification rate of 33.8 g⋅m⁻²⋅h⁻¹. By applying such humidity pump, the indoor latent heat load can be handled independently, without any auxiliary unit, thus consuming no electricity.

Graphical Abstract

Highlights

• •We propose a concept of moisture-permeable panel that enables moisture penetration
• •A humidity pump prototype using the moisture-permeable panel is designed
• •The humidity pump can reduce the indoor relative humidity to a medium level

Engineering; Energy Systems; Materials Science; Energy Materials

Introduction

Ambient humidity collection has become a popular topic because of its important role in water collection and air humidity adjustment, according to the different post-collection management processes. Atmospheric water harvesting, which aims to obtain clean liquid water after ambient humidity collection, has already been intensively studied (Kim et al., 2017, Mekonnen and Hoekstra, 2016, Kalmutzki et al., 2018). This contributes significantly to arid areas with potable water shortage. Humidity control, which focuses on the removal of water vapor in the air after humidity collection, also attracts dramatic attentions owing to the increasing interest on sustainable buildings (Allouhi et al., 2015, Anand et al., 2015, Grossman, 2002, Otanicar et al., 2012). In modern times, nearly 40% of the total energy consumption is associated with buildings worldwide (Zhao and Magoules, 2012, Farese, 2012, Yang et al., 2014, Costa et al., 2013). According to the United Nations, by 2050, 70% of the total population or more is trending to live in urban areas, where the density of buildings is significantly larger than countryside and people usually spend more than 80% of days indoors (Rupp et al., 2015). This makes the indoor environment control dramatically important (Yang et al., 2014), and thus increasing attention has been paid to this field. Evidence shows that more than half of energy consumption in buildings, which is equal to almost 20% of the total energy consumption, results from air-conditioning systems (Farese, 2012, Costa et al., 2013, Yang et al., 2014). Besides, temperature and humidity management is also critical in many industrial occasions (Nkwetta and Sandercock, 2016). As widely acknowledged, roughly 40% of the heating, ventilation, and air-conditioning (HVAC) load is responsible for latent heat load (dehumidification) and the other 60% for sensible heat load (Tu et al., 2017). Thus effective technologies for humidity control would be of great importance. Conventional vapor compression system, which is the most preferred dehumidification technology commercially, dehumidifies the air through the dew point method, by which the supply air is cooled down to a much lower temperature than the indoor environment needs, whereas the moisture in the air is condensed and removed (Zhang et al., 2014). However, a reheating of outlet air is essential to satisfy the temperature needs of inhabitants and thus results in large extra energy consumption. Previous researches have shown that the temperature- and humidity-independent control system, in which the sensible and latent heat loads are regulated separately, has a promising energy-saving potential (Zhang et al., 2014, Zhao et al., 2011, Zhang and Liu, 2016). Therefore different dehumidification technologies based on ambient humidity collection process have been developed to achieve the expected balance between energy consumption and indoor thermal comfort. The solid desiccant cooling system may handle the latent heat load independently and use waste heat or renewable energy, such as solar energy. The refrigerant and desiccant materials involved are mostly environment friendly. This may result in obvious energy-saving potential and positive ecological and economic effects (Daou et al., 2006, Panaras et al., 2011, Zheng et al., 2014, La et al., 2010). The rotary system, based on solid desiccant system, can handle the latent heat load continuously and more efficiently (La et al., 2010). The liquid desiccant absorption system, which utilizes the hygroscopic salt solution instead of a solid desiccant, can adjust the indoor humidity accurately and continuously (Mei and Dai, 2008, Shukla and Modi, 2017). In recent years, the desiccant-coated heat exchanger system is developing very fast, with merits of both vapor compression system and sorption-based system (Tu et al., 2017, Zheng et al., 2015a, Zheng et al., 2015b). The solid desiccant material is coated on the fins of the heat exchanger so that the latent and sensible heat loads can be handled independently and simultaneously. As the concept of sustainable building or zero energy building is receiving increasing impetus, several novel ideas on energy-saving technologies in buildings flourished regarding different aspects (Rotzetter et al., 2012, Marszal et al., 2011, Deng et al., 2014). Renewable energy resources, in most cases the solar energy, are usually the first choice (Chu and Majumdar, 2012). Thus the sorption-based dehumidification technologies become very promising choices for indoor air-conditioning in buildings.

However, the sorption-based dehumidification technologies are still not that widely applied currently. Some major drawbacks still limit the application in industries and buildings. Owing to the adsorption-desorption nature, the working process is usually intermittent (Panaras et al., 2011, La et al., 2010). The installation, especially the rotary system, although overcomes the intermittence problem, is often bulky and complex, and this inevitably increases the maintenance cost a lot (La et al., 2010). Furthermore, such systems can only utilize solar thermal energy by introducing additional energy harvesting, conversion, and transportation units and systems, which is also a major reason for bulky and complex installation (Daou et al., 2006). The direct usage of solar energy is attracting attention currently, and a large amount of research has been carried out regarding the high-efficiency usage of solar thermal energy in surface-localized heating and vapor generation (Ni et al., 2016, Wang et al., 2014, Liu et al., 2015a, Liu et al., 2015b). However, unfortunately few such researches have been related with the sorption-based humidity control technologies, which should have a large potential.

In this study, a moisture-permeable panel based on solid desiccant directly powered by one-sun illumination is proposed to serve as a passive humidity pump, which can transfer the moisture in the air from indoor to outdoor with a very simple structure. We also report a proof-of-concept prototype that successfully removes the indoor moisture to outdoor using our passive humidity pump only under one-sun illumination. We introduced both silica gel (SG)-based and SG-MIL101(Cr)-based moisture permeable panels for comparison. The MIL-101(Cr) can improve the dehumidification rate and also successfully reduce the relative humidity (RH) down to the medium level (around 50%), which fulfills the human thermal comfort requirement (Wan et al., 2009). With such technologies, the indoor latent heat load can be handled independently without consuming electricity and the HVAC systems in buildings are expected to be much more efficient.

Results

The Concept and Design of the Moisture-Permeable Panel

The moisture-permeable panel consists of three parts: the porous substrate, the desiccant layer, and the photothermal coating film (see Figure 1A). The porous substrate is introduced for both holding the desiccants and providing good thermal insulation property. The widely used desiccants, such as vermiculite or zeolite powders, are inconvenient and difficult to be shape stable, thus a matrix that may provide skeletons for desiccants to attach is essential. Besides, to prevent the heat loss of the panel, thermal insulation property should be considered. The desiccant layer is responsible for adsorbing water vapor from indoor space, transferring the water molecules from the inner to the outer surface and desorbing the attached water to outdoor space. To achieve such function, a relatively good water capture ability is expected under normal room temperature and humidity, and its desorption should happen under relatively modest temperature. The photothermal layer is introduced to provide sufficient desorption heat. Different from the traditional desorption process, the heat is neither coming from hot air, which flows over the desiccant layer surface, nor from the heat exchanger fins, which are coated by desiccant, both of which are energy-consuming processes. Instead, the needed heat is directly generated on the desiccant layer surface, which largely simplifies the installation of the structure. The photothermal layer should be made from a material that is very inert and stable considering the variable and even extreme outdoor environment.

Figure 1.

An Overview of the Moisture Permeable Panel

(A–D) (A) Schematic structures of moisture-permeable panel and expected moisture transfer path. The moisture in the indoor air passes through the porous matrix and is adsorbed by desiccant. The absorbed water molecules pass through the desiccant layer and reach the outdoor surface. The absorbed water is removed by the heat generated by CB powder under sunlight illumination. The panel structure with different desiccant layers is shown. (B) SG layer and (C) MIL101(Cr)-SG layer. (D) Photograph of the cross section of a real moisture-permeable panel.

In this research, a common thermal insulation foam, the wet foaming phenolic foam (PF), is introduced as the porous substrate due to its merits of excellent thermal insulation, low cost, easy shaping, etc. Glass and ceramic fibers are also promising alternatives, which are more rigid but suffer from greater weight and difficult shaping issues. For the desiccant layer, the SG and SG-MIL101(Cr) composite materials are chosen and investigated. Silica gel is widely applied in adsorption cooling and dehumidification fields owing to its advantages of satisfying water capture ability, non-toxic nature, easy fabrication process, and low cost (Pesaran and Mills, 1987a, Pesaran and Mills, 1987b). It plays an important role especially in rotary wheel systems and desiccant-coated heat exchanger systems (Tu et al., 2017, Ge et al., 2009, Zheng et al., 2015a). The metal organic framework MIL-101(Cr) has been investigated and applied in several cases such as carbon dioxide capture, hydrogen storage, or water adsorption (Rallapalli et al., 2013, Liu et al., 2013, Hong et al., 2009, Khutia et al., 2013). Here it is chosen as the second candidate because it also enjoys excellent water adsorption ability and stability under extreme weather conditions (hot and strong sunshine) (Ferey, 2005). Besides, it has a type V isotherm according to International Union of Pure and Applied Chemistry, and the S-shape is around medium level (see Figure S1), which is the optimum feature for solid desiccant system. The photothermal layer is made from carbon black (CB) powder with particle size around 50 nm. This very small size may guarantee good dispersion of CB powder on the panel surface. By following the ideas above, two kinds of panels, one with pure SG layer and the other with SG-MIL101(Cr) layer, have been proposed and fabricated, as shown in Figures 1B and 1C (see Figure S4, Transparent Methods section for details of MIL-101Cr fabrication and moisture-permeable panel fabrication). A cross-sectional photograph of a real panel sample is shown in Figure 1D for better understanding. The detailed fabrication process is shown in Transparent Methods. It should also be noted that two structures of desiccant layer are applied, either condensed or loose, according to the amount of desiccant in the layer. Details for each panel are listed in Table S1.

Different Desiccant Layer Structures with Different Mass Transfer Abilities

The scanning electron microscopic and transmission electron microscopic (TEM) images of panels are first given to understand the panel surface morphology (Figure 2). Figure 2A clearly shows the pure PF surface structure. The pores around the skeletons are more than 100 μm, which provides enough space for both desiccant adhesion and moisture transport. In the condensed layer (Figure 2B), most of the pores are filled with SG, and this is expected to result in a larger adsorption quantity. On the contrary, the loose layer (Figure 2C) spares more pores so that more channels can be involved in mass transfer instead of being blocked. However, the influence of less adsorption amount should be further evaluated. Figure 2D gives the details of SG-MIL101(Cr) composite desiccant. Very small MIL101(Cr) crystals are observed to be aggregated on the SG surface (and also inside the SG bulk). The TEM images provide clear appearance of the MIL101(Cr) crystals (Figures 2E and 2F). The dimension of single MIL101(Cr) crystal is around a few hundred nanometers, and very regular pore and channel structures can be seen in Figure 2F. The dynamic water adsorption and also nitrogen adsorption of the MIL101(Cr) involved are shown in Figure S2.

Figure 2.

Surface Morphologies and Water Vapor Transmission Test

Scanning electron microscopic images of (A) pure wet-foaming phenolic foam surface, (B) PF foam with condensed SG layer, (C) PF foam with loose SG layer, and (D) MIL-101(Cr) aggregates on the SG particle surface. TEM images of (E) pure MIL-101(Cr) crystals and (F) the zoomed-in area of dashed square in (E). (G) Results of permeability test. The reference sample in permeability test stands for the pure wet-foaming phenolic foam without any silica gel attachment.

To acquire detailed evidence on mass transfer, the water vapor transmission (WVT) test was carried out (see Figure S5, Transparent Methods, Water Vapor Transmission Test for experimental details). The mass loss of each kind of panel was weighed and is shown in Figure 2G, in which the slope of the line indicates the mass decrease rate proportion to WVT ability. The reference (black square) curve represents the pure PF foam sample, whereas red dots, blue triangles, and green triangles stand for panels with loose SG, condensed SG, and loose SG-MIL101(Cr) composite layers, respectively. All the mass loss curves show a linear shape. It is obvious that the impregnation of SG into the PF foam diminishes the moisture transmission ability, and the more SG there exists, the worse the mass transfer rate is. The addition of MIL-101(Cr) hardly influences the WVT ability, and the two lines of SG loose layer and SG-MIL-101(Cr) layer are mostly identical. This is attributed to the tiny particle size and very small aggregation domains of MIL-101(Cr). As presented in Figure 2D, the aggregated domains are in the range of less than 1 μm and the particles are even much smaller, which can be considered as a thin film covering the SG surface. These MIL-101Cr particles and aggregated domains barely change the pore and channel structures and thus consequently influence very little on the WVT ability. This result corresponds to our suggestion above that the amount of accessible channels and pores in the desiccant layer are considered as the dominating reasons.

Quantitatively, the specific indicators, WVT, water permeance, and permeability, can be expressed by the following equations according to the Standard Test Methods for Water Vapor Transmission of Materials:

$$\text{WVT}=\frac{G}{tA}$$ (Equation 1)

where G stands for the total mass decrease amount (in g), t is the time (h), and A is the area of the sample mouth (in m²).

$$\text{permeance}=\frac{\text{WVT}}{\Delta \text{p}}=\frac{WVT}{S\left(R_1-R_2\right)}$$ (Equation 2)

where Δp indicates the vapor pressure difference between two sides of the tested samples, S is the saturation vapor pressure under test temperature, and R₁ and R₂ are the RHs between two sides of the tested samples.

$$\text{average}\text{permeability}=\text{permeance}\times \text{thickness}$$ (Equation 3)

The sample mouth area is 9 cm² (3cmⅹ3cm), and the test environment is carefully controlled at 23°C and 50% RH. Thicknesses of all the samples are identical. Using the equations introduced above, the WVT and water permeance of each sample are calculated. The detailed results are listed in Table 1. Basically, the WVT and permeance of panels with loose SG layer are 1.5 times higher than those of panels with condensed SG layer and are almost identical to those of panels with SG-MIL101(Cr) composite desiccant layer. So it is very obvious that both loose SG and loose SG-MIL101(Cr) composite layers have good mass transfer abilities, whereas that of the condensed layer is much worse (see also Figure S3, Table S3, and Transparent Methods, The Sorption Rate Test for more information on sorption dynamics).

Table 1.

The Mass Loss Rate, WVT, Permeance and Permeability of Moisture-Permeable Panels with Different Desiccant Layer Structures

Desiccant Layer Structures	Mass Loss Rate (g⋅h−1)	WVT (g⋅h−1⋅m−2)	Permeance (g⋅h−1⋅m−2⋅pa−1)	Permeability (g⋅h−1⋅m−1⋅pa−1)
SG condensed	0.0400	44.4	0.0226	1.13ⅹ10−4
SG loose	0.0598	66.4	0.0338	1.69ⅹ10−4
SG-MIL-101(Cr)	0.0584	64.8	0.0330	1.65ⅹ10−4

The Surface Temperature (Regeneration Temperature) Enhancement by the Carbon Black Photothermal Layer

Infrared (IR) images were taken to evaluate the photothermal effect induced by the CB coating, as shown in Figure 3. Four samples with different structures and initial conditions were placed under the same light source with 1,000 W/m² intensity. The details of each sample are listed in Table 2. Photographs were taken after 0, 5, and 30 min, and all the samples (except sample D) reached equilibrium at 25°C, 70% RH working condition before they are exposed to illumination. A relatively small temperature rise of sample A (average temperature: 37.7°C after 5 min) is observed, whereas sample B has a much faster temperature increase (average temperature: 56.7°C after 5 min). This can be ascribed to the CB coating on the surface, which generates much more heat due to the photothermal effect. Sample D, which has identical structure to sample B, is pre-dried completely before taking photograph and thus resulted in an even higher surface temperature (average temperature: 58.4°C after 5 min) because water desorption usually consumes a certain amount of heat. Sample C, which has an SG-MIL101(Cr) composite desiccant layer, has a lower temperature distribution (average temperature: 47.0°C after 5 min). The dominating reason is believed to be the larger water capacity of the composite desiccant and faster water-losing rate. More detailed information can be seen in Table S2. The IR images clearly prove that the CB coating can largely increase the surface temperature. Besides, the desorption can happen under one-sun illumination condition according to the lower temperature of samples B and C, and obviously the composite desiccant desorbs water faster than the pure SG (also see Figure S2). A heat transfer model that applies the CB layer as a surface heat source is established, and the simulation also shows the same trend as Figures 2B and 2D (see Transparent Methods, Model and Simulation, and Figures S6 and S7).

Figure 3.

IR Images of Different Panels under One-Sun Illumination after a Fixed Time

(A–D) (A) Wet panel with SG layer, surface without CB coating; (B) wet panel with SG layer, surface with CB coating; (C) wet panel with MIL101-SG layer, surface with CB coating; (D) dry panel with SG layer, surface with CB coating.

Table 2.

Details of Each Sample in IR Images

Sample	a	b	c	d
Desiccant	SG	SG	SG-MIL101(Cr)	SG
CB coating	No	Yes	Yes	Yes
Dry or wet	Wet	Wet	Wet	Dry

Proof-of-Concept Prototype and Dehumidification Results

To demonstrate the feasibility and to evaluate the performance of this novel moisture pump under real working conditions, a proof-of-concept prototype applying the moisture permeable panel was established and tested. A scaled down model of a house was made with dimensions of 70 cm × 40 cm × 40 cm. Figure 4A gives the sketch of this prototype. Two windows with an area of 5 cm × 5 cm were design for MP panel installation. Movable outer shields and inner shields were introduced for shielding the sunlight or the indoor air depending on different working processes. By opening the inner shield of one MP panel, this panel absorbs water vapor from the indoor air, whereas the outer shield is closed, keeping this panel from sunlight illumination. Meanwhile, the inner shield of the other panel is closed, whereas the outer shield is opened, isolating this second panel from indoor air but under illumination from sunlight, as shown in Figure 4B. After a certain period, the shields are switched and the panel working condition is reversed. The first panel, which has absorbed a certain quantity of water, is now exposed to sunlight, whereas the other panel, which has been regenerated well, is now capturing the water vapor from indoor space, as shown in Figure 4C. By switching the working mode alternatively in a controlled period, this system absorbs indoor water vapor and loses water to outdoor simultaneously.

Figure 4.

Illustration of Proof-of-Concept Prototype for Moisture-Permeable Panel Performance Evaluation

(A) The setup of scaled down model.

(B–D) (B) and (C) present the working process. The panel under sunlight illumination is losing water, whereas the one shielded from sunlight is adsorbing water vapor from indoor air. (D) System performance of different panels: panels with loose SG layer (open cycles) and panels with SG-MIL101(Cr) layer (closed cycles). The red, blue, and black symbols represent the temperature (°C), relative humidity (%), and water content of the air (g/kg), respectively. Vertical dashed lines indicate the switching time period (20 min).

(E) The system performance of SG-MIL101(Cr) panel close to medium RH level (around 50%). The red, blue, and black symbols represent the temperature (°C), relative humidity (%), and water content of the air (g/kg), respectively.

Two identical MP panels with 5 cm × 5 cm area were installed in each window. The initial working condition inside the box was 25°C, 65% RH, and both panels reach equilibrium under such working condition. The box was tightly sealed to prevent any possible leakage, and one panel with inner shield closed was exposed to one-sun illumination (1,000 W⋅m⁻²), whereas the other panel was still covered by outer shield. The switch time period was chosen as 20 min. After the switch, the working mode of each panel was changed, and this switch repeated within the whole dehumidification process. The outside working condition was kept constant at 25°C, 70% throughout the whole experiment. Figure 4D presents the results of this prototype applying panels with both SG loose layer (open cycles) and SG-MIL-101(Cr) layer (closed cycles), and several consecutive cycles of each panel were chosen and presented. For SG loose layer panel, obvious drop of RH (open cycle, blue) as well as water content (open cycle, black) appeared in the early periods, indicating the adsorption of water vapor by the desiccant layer in the panel and the decrease of RH of the indoor space. However, as the setup kept working and the RH went down (around 60%), the RH drop became more and more subtle and finally reached a near-constant status. The indoor temperature remained almost constant, and only very small temperature variation was observed (less than 0.3°C) resulting from adsorption heat. This is associated with the thermal insulation layer because most of the adsorption heat generated in the desiccant layer spreads through the desiccant layer rather than the porous thermal insulation layer, leaving the inside temperature almost unchanged. The specific dehumidification rate can be calculated as follows:

$$\text{r}=\frac{\Delta m}{t\cdot S}\left(\text{g}\cdot {\text{m}}^{-2}\cdot {\text{h}}^{-1}\right)$$ (Equation 4)

where $\Delta m$ (g) stands for the water loss during the whole dehumidification process, t represents the time (h), and S stands for the total area of the MP panels (m²)

$$\text{m}=\text{d}\cdot m_{air}=\text{d}\cdot \text{V}\cdot \text{ρ}\left(\text{g}\right)$$ (Equation 5)

$$\text{d}=0.622\frac{RH\cdot P_S}{B-RH\cdot P_S}\left(\text{g}\cdot {\text{kg}}^{-1}\right)$$ (Equation 6)

where d is the water content (g/kg) of the air, V is the volume of the model house, $\rho$ is the density of air at the working temperature, RH is the relative humidity (%), P_s is the saturated water vapor pressure at the working temperature, and B stands for the atmospheric pressure.

For the panel with SG layer, only the first three cycles were taken into consideration. The indoor RH dropped from 65.0% to 62.2%, and the water content correspondingly varied from 12.74 to 12.29 g⋅kg⁻¹, resulting in an average dehumidification rate of r = 24.2 g⋅m⁻²⋅h⁻¹.

For the panel with SG-MIL101(Cr) layer (closed cycles), the performance was much better. During the first three cycles, the RH dropped obviously faster than that of the SG layer panel did. When the RH (closed cycles, blue) reached around 60%, the drop rate decreased to a moderate level but still kept going down regardless of this medium RH condition. The following several cycles clearly showed that the RH kept going down to a much lower value. Furthermore, Figure 4E presents a few cycles of SG-MIL101Cr panel working under much lower RH condition (around 52%, the medium RH level). Under such working condition, the indoor RH can still be steadily decreased. This is the major difference between SG-MIL101Cr layer and the SG layer. It is evident that the SG lost most of the dehumidification ability under such indoor-outdoor working conditions. However, the MIL101Cr can still have satisfactory water adsorption ability under the same condition, as described above in The Concept and Design of the Moisture-Permeable Panel. Below 60% RH, what contributes to the RH drop is only the MIL101Cr rather than SG.

To evaluate the dehumidification rate quantitatively, the process should be divided into two regions: the high-RH region (higher than 60%) and the medium-RH region (lower than 60%). At the high-RH region, the water content dropped from 12.76 to 12.13 g⋅kg⁻¹ with an average dehumidification rate of r = 33.8 g⋅m⁻²⋅h⁻¹. At the medium-RH region, the water content decreased from 12.13 to 11.59 g⋅kg⁻¹ with an average dehumidification rate of r = 15.1 g⋅m⁻²⋅h⁻¹. A brief model is established to theoretically estimate the dehumidification rate (see Transparent Methods, Dehumidification Rate).

The energy efficiency is calculated as follows:

$$\eta =\frac{H_{de}}{Q_{input}}$$ (Equation 7)

where H_de is the phase change enthalpy of removed water and Q_input is the total energy of the incident sunlight. For simplicity, we consider the humidity pump with 1 m² of panel and the working time is 1 h.

$$H_{de}=\Delta d\ast H_{water}=33.8g\ast 2513J{g}^{-1}=8.5\ast {10}^{4}J$$ (Equation 8)

$$Q_{input}=I_0\ast t={10}^{3}W{m}^{-2}\ast 3600s=3.6\ast {10}^{6}J$$ (Equation 9)

Therefore the energy efficiency $\eta$ is 2.36%. It should be noted that higher efficiency can be expected with rational improvements, e.g., new water capture material with much higher water capacity or optimized structure that allows higher permeability. See Supplemental Information for details (Transparent Methods, Energy Efficiency).

The adsorption-desorption cycle test was carried out to better understand the difference between SG panel and SG-MIL101Cr panel. The adsorption processes were carried out at different conditions: 25°C, 70% RH and 25°C, 50% RH, whereas the desorption processes were taken under one constant working condition: 25°C, 70% RH (the desorption only takes place on the outside surface, where the environment condition is considered as constant). The mass difference of each cycle represents the water transportation ability of each sample under certain conditions. The adsorption and desorption were carried out alternatively using different setups (Figures 5A and 5B) with a period of 20 min. During this process, the mass change of each sample was measured and recorded. Both samples have the same dimension of 3 cm × 3 cm. Six stable cycles were chosen and shown in Figures 5C–5F. The squares stand for SG panels, whereas the cycles represent SG-MIL101Cr panels. Under 25°C, 70% RH, both SG (Figure 5C) and SG-MIL101Cr (Figure 5D) panels have satisfactory mass differences, 0.07 and 0.084 g, respectively, which is equal to the effective amount of water transported per cycle. The better performance of the composite panel may result from the larger water uptake ability of MIL101Cr. Under 25°C, 50% RH, however, the mass differences of both samples decreased a lot, especially for SG panel (Figure 5E). Roughly 0.005 g per cycle is observed. The mass difference of SG-MIL101Cr sample (Figure 5F), although decreases, still remains at a moderate value of 0.04 g. It is very clear that under such medium RH condition as well as the photothermal-driven condition on the outside surface, the SG panel loses most of the water capture ability because the difference of water uptake amount between the indoor and outdoor working conditions is too small to provide enough positive moisture potential as the mass transfer-driven force. On the other hand, MIL101Cr can still keep an adsorption-desorption cycle with obvious water uptake amount according to its isotherms. Therefore it is evident that the SG-MIL101Cr panel not only performs better in the high-RH region but also can successfully reduce the indoor RH down to medium RH level.

Figure 5.

The Adsorption-Desorption Cycles

The experimental setup of adsorption (A) and desorption (B) processes are shown. Two different working conditions are chosen for adsorption process, whereas only one constant working condition is applied for desorption. The mass differences of the cycles of SG panel at (C) 25°C, 70% RH and (D) 25°C, 50% RH, SG-MIL101Cr panel at (D) 25°C, 70% RH and (E) 25°C, 50% RH are shown.

Discussion

Estimation of Latent Sensible Heat Load Conversion

The introduction of this solar humidity pump may realize the independent control of temperature and humidity. Usually when certain amount of moisture is adsorbed by solid desiccant, the corresponding amount of latent heat is converted into sensible in the form of adsorption heat (Maclaine-Cross and Banks, 1972). If handled improperly, most of this heat may increase the overall indoor heat load and diminish the energy conservation introduced by this solar humidity pump. A detailed mathematical modeling is beyond the scope of this article, but a brief estimation is necessary and enough to demonstrate the idea of structure design. As the dynamic adsorption curve shows (Figure S2), the porous PF layer may influence the overall WVT ability of the panel. However, the PF has such an excellent thermal insulation property that it may greatly weaken the heat loss of the panel, and therefore the little influence on mass transfer can be neglected. If the desiccant layer is in direct contact with the indoor air (Figure 6A), the heat flux to indoor (Q₁) and outdoor (Q₂) can be expressed as:

$$Q_{in}=h\cdot A\cdot \Delta t$$ (Equation 10)

$$Q_{out}=\frac{{\lambda }_{SG}}{{\delta }_{SG}}\cdot A\cdot \Delta t$$ (Equation 11)

Figure 6.

Schematic Diagram of Heat Transfer in Adsorption Process

(A and B) (A) Panel without a porous matrix layer inside. Most of the heat is transferred to indoor air via convection; (B) panel with a porous matrix layer. Most of the heat is transferred to outdoor space.

For simplicity, we consider a simple natural convection condition, where the h is around 10 W/m²K. The thermal conductivity of the SG layer is around 0.1 W/mK and the sample thickness 1 cm, thus the value of ${\text{λ}}_{\text{SG}}/{\delta }_{SG}$ is also around 10 W/m²K. So under such condition, the Q_in roughly equals Q_out. In comparison, by introducing the multilayer structure (Figure 6B), the water vapor should first go through the porous PF layer before it reaches the SG layer interface. Similarly, the adsorption heat transfer path Q'_in and Q'_out can be written as:

$$Q_{in}^{\prime }=\frac{{\lambda }_{PF}}{{\delta }_{PF}}\cdot A\cdot \Delta t$$ (Equation 12)

$$Q_{out}^{\prime }=\frac{{\lambda }_{SG}}{{\delta }_{SG}}\cdot A\cdot \Delta t$$ (Equation 13)

The thermal conductivity of pure PF layer is only around 0.03 W/mK, and in this research the thickness of PF matrix layer is identical to that of the SG layer. Therefore the value of ${\text{λ}}_{\text{PF}}/{\delta }_{PF}$ is around 3 W/m²K. Under such condition, the value of Q'_in equals 30% of Q'_out. Furthermore, the thickness of the PF layer ${\delta }_{PF}$ can be much larger than it was in this research and consequently leads to a much smaller Q'_in value and Q'_in:Q'_out ratio.

Potential Applications

A bright future for a broad application of such moisture pump is expected. Potential application situations may include commercial and residential buildings, industrial plants, and electronic devices. Such moisture pump can be easily integrated with construction materials so that a moisture-permeable wall or window is expected. Thus the indoor latent heat load can be independently handled by this humidity pump in the daytime, resulting in a higher efficiency of the cooling unit, which could only deal with the sensible heat load. Consider a department with an area of 25 m² and one inhabitant inside. The required panel area is around 5.6 m² to fully cover the moisture load (see details in Transparent Methods, Real World Application). Similar construction structures can be expected in industrial plants, but with much larger scale. The performance-price ratio of this humidity pump is the same as or even better than that of the solar photovoltaic-driven air-conditioning system (see details in Transparent Methods, Performance-Price Ratio). Another application situation is the humidity control inside some precise electronic devices. The moisture often appears as a disaster inside the precise electronic devices, which may disable the current circuit or even destroy the whole device. A common approach to remove the moisture inside the device can be achieved by drying with hot wind, but this requires the unsealing of the device, which may bring in other impurities or damage the mechanical parts during disassembly or assembly. By applying such microscale humidity pump, the moisture removal process is available without the need to disassemble the device, leaving no risk of further damage or impurity introduction. Under such circumstances, the heat source is no longer limited as sunlight, but could be any kind of surface heating technology.

Limitations of the Study

The sun illumination condition applied in the experiment is idealized. This research is a proposal of a concept and its validation rather than a systematic study of such humidity pump system. Therefore the sun illumination condition is simplified to better understand the working mechanism and the performance criteria. However, under real conditions, the solar radiation usually grazes the building facade rather than incident perpendicularly, thus solar radiation model should be then considered to better evaluate the working performance under real conditions. Such work should be carried out in a more systematic investigation of such humidity pump system, which includes the performance tests of some standard working conditions.

Besides, this humidity pump still suffers from a relatively low dehumidification rate compared with both vapor compression system and conventional solid desiccant cooling system. The relatively poor water vapor transfer ability is the dominant reason. The key to improve this issue very largely lies both on the material science, which may provide more appropriate candidate desiccants, and structure design, which could optimize the facility with aligned channels and good thermal insulation, if possible. The material fabrication and selection, as well as the structure optimization, would be a long-term work, which, when fulfilled, may largely improve the dehumidification rate of this humidity pump.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.

Published: May 31, 2019