Scale-up of solar interfacial evaporation devices: advanced optical, thermal, and water management for efficient seawater desalination

Abstract

Significant progress has been made in enhancing solar interfacial evaporation (SIE) performance at the laboratory scale, however, translating these improvements to meter-scale systems suitable for practical deployment remains limited by challenges, including material scalability, thermal losses, and non-uniform water distribution. Addressing these issues is essential for the development of modular, meter-scale evaporators as baseline units for industrial-scale desalination systems. This study presents a solar multi-stage interfacial evaporation (SMIE) device with a 1 m² active area designed to address the key limitations associated with large-scale operation systematically. The device integrates: (i) a scalable photothermal absorber layer based on a Cu-CAT-1 metal–organic framework or carbon black, (ii) an inverted multi-stage configuration with optimized thermal insulation to reduce energy loss and enable latent heat recovery, and (iii) structured wicking channels engineered to maintain spatially uniform water transport. Under 1-sun illumination in a controlled laboratory setting, a 100 cm² prototype achieved freshwater production rates of 5.45 kg m⁻² h⁻¹ with deionized water and 3.90 kg m⁻² h⁻¹ with 3.5 wt% saline water. Outdoor testing of a full-scale 1 m² device yielded an average freshwater production rate of 3.5 L m⁻² h⁻¹ (32 L m⁻² day⁻¹) and an evaporation efficiency of 345%. These results confirm that the proposed SMIE design maintains high performance at increased scales and under realistic environmental conditions. A techno-economic analysis further identifies the critical role of reducing material costs, particularly the photothermal absorber and porous membrane, to enhance economic feasibility. This work provides a scalable approach for solar-driven desalination, relevant for future modular deployment in distributed and off-grid water purification applications.

---

Broader context

Solar interfacial evaporation has emerged as a promising and sustainable strategy to address the dual global challenges of water scarcity and increasing energy demand. Lab-scale devices have achieved high evaporation performance through advances in materials and structural design. However, scaling up from centimeter-scale units by connecting multiple modules introduces challenges, including increased system complexity, poor space utilization, and higher infrastructure costs. To enable practical, large-scale deployment, meter-scale (1 m²) devices are increasingly viewed as the smallest effective unit for industrial applications. In this study, we present a highly efficient and scalable solar multi-stage interfacial evaporation device with an active area of 1 m². Our design addresses critical challenges related to material scalability, energy loss, and consistent water supply, demonstrating both high performance (3.5 L m⁻² h⁻¹) and suitability for large-scale applications. By bridging the gap between laboratory innovation and practical implementation, this work provides a robust framework for advancing solar desalination technologies. Techno-economic analysis further indicates that reducing the cost of the photothermal absorption layer and porous membrane is essential to improving the economic viability and shortening the payback period of SMIE systems.

Introduction

Freshwater is essential for both ecosystems and human society, and its availability is expected to become increasingly constrained in various regions due to rising demand and climate variability.¹ Solar interfacial evaporation (SIE), which leverages thermal localization, offers high evaporation efficiency^2–5 and adaptability for decentralized water purification applications.^6–10 The efficiency of devices employing thermal localization has been improved in recent developments via multi-stage designs for condensation heat recovery,^11–13 photothermal material engineering for high solar band absorption^14–16 and low thermal band emission,^17–19 and wettability tuning for a balanced evaporation and wicking rate.^20–22 With these advancements in both materials and device engineering, evaporation efficiency can be achieved in the range of 100–684%.^23,24

However, most high-efficiency SIE devices are limited to lab-scale demonstrations^25,26 (typically <100 cm²), restricting daily freshwater output to below 10 L.²⁷ Such capacities are insufficient to meet the basic needs of a small household (∼3-person family), which may exceed 100 L d⁻¹ when considering drinking and cooking.^28,29 Although parallel assembly of multiple centimeter-scale units can, in principle, achieve higher throughput, this approach often increases system complexity, reduces space efficiency, and increases the overall cost of auxiliary infrastructure. To address practical water demands and reduce system complexity, a single-unit evaporator with an active area of approximately 1 m² is required. Such a unit is considered to be the minimal functional module for industrial-scale desalination via modular parallelization.³⁰

Despite a few recent efforts to construct large-scale evaporative systems, such as a 150 m² array assembled from 50 individual units (unit sizes: 0.6 m × 2.0 m; 1.2 m × 1.9 m), the performance of meter-scale evaporators has generally remained low, with average freshwater production rates below 0.5 L m⁻² h⁻¹.³⁰ The main limitations can be linked to: (i) lack of uniform coating with controllable and precise mass loading, (ii) uncontrollable top and side convection losses, and (iii) insufficient and non-uniform water supply for large-scale devices. Additionally, for meter-scale evaporator design, energy losses associated with the front side due to radiation and conduction from the cover to the ambient are major energy losses.^31–33 In particular, when water needs to be collected and conventionally condensed at the cover, the heat loss, due to untamed latent heat and optical blocking by the water droplets, is one significant part that limits the efficiency of the device.^34,35

The inverted device with a multi-stage configuration can well tackle this issue by directing the vapor downward to the condensation layer for efficient condensation, hence, to avoid otherwise optical blocking.^36–40 Meanwhile, the inverted multi-stage design offers opportunities for latent heat recovery by the evaporation layer at the next stage. However, when scaling up the design, the heat conduction losses from the edge of the device will be non-negligible due to the increased height of the device. These edge losses are more pronounced at higher stage numbers due to increased edge thickness. This necessitates careful consideration of the device's thermal insulation design.

Moreover, water transport becomes more complex in larger systems. For non-swelling wicking materials, water transport follows the Lucas–Washburn capillary model,^41–43 wherein the wicking distance increases only with the square root of time. This relationship leads to a spatially heterogeneous water supply under constant solar irradiation, potentially causing local dry zones and temperature hotspots.⁴⁴ Therefore, spatially adaptive water management, through material selection or the geometric design of wicking structures, is required to match water delivery with localized evaporation demands and, hence, to prevent dry zones.

Here, we demonstrate a highly efficient and large-scale solar multi-stage interfacial evaporation (SMIE) device with an active area of 1 m² (Fig. 1) that addresses the key challenges of material scalability, energy loss mitigation, and water distribution. The device incorporates: (i) scalable photothermal layers fabricated via spray-coating of either Cu-CAT-1 nanoparticles or carbon black, (ii) an inverted multi-stage design with optimized edge thermal insulation to avoid optical losses due to water droplets at the cover and recycle the latent heat, and (iii) well-designed wicking channels to mitigate the overheating problem resulting from an insufficient water supply in large-scale devices. A 10-stage prototype was constructed and evaluated under both indoor and outdoor conditions. A lab-scale device with an area of 100 cm² was developed for indoor testing. Freshwater production rates of 5.45 kg m⁻² h⁻¹ and 3.90 kg m⁻² h⁻¹ were achieved under 1-sun (1000 W m⁻²) condition with pure water and 3.5 wt% saline water, respectively. Furthermore, the full-scale device with an area of 1 m² was tested under real outdoor conditions. With an average outdoor irradiation of 651 W m⁻², the device collected 32 L of fresh water over 8 hours, i.e., 32 L m⁻² d⁻¹, achieving an average water production rate of 3.5 L m⁻² h⁻¹ (efficiency of 365%) and a peak rate of 5.1 L m⁻² h⁻¹. This 1 m² device can be used as a single unit to expand its application through parallel connection, to meet the different capacity needs of households or industries. This work offers practical design guidelines for the engineering of scalable passive solar evaporation devices with high performance.

Fig. 1 Schematics of a large-scale solar multi-stage interfacial evaporation (SMIE) device. The orange box shows the spray coating strategy for scaled-up absorption layer fabrication. The blue box shows the multi-channel water supply strategy to satisfy the requirements for a high evaporation rate. The bottom part shows the schematics of the SMIE device with detailed optical, heat, and mass transfer processes.

Results and discussion

Spray coating of Cu-CAT-1 MOF for meter-scale light absorber fabrication

Fig. 2a and b illustrate the scalable fabrication process of the Cu-CAT-1 MOF or carbon black (CB) absorption layer with high solar spectrum absorption and super-hydrophilicity for scalable solar interfacial evaporators via spray coating (Fig. 2c). For Cu-CAT-1 MOF, the Cu(OH)₂ nanowires were used as the copper source for the synthesis of Cu-CAT-1 MOF nanorods (NRs) via the hydrothermal method. Cu-CAT-1 MOF nanorods are dispersed uniformly in an anhydrous ethanol solution via ultrasonication. Subsequently, the mixed suspension solution was applied uniformly to the surface of PVA sponges, i.e., the substrate, utilizing a spray coating technique (Fig. 2b). Based on the SEM images shown in Fig. 2d, Cu-CAT-1 MOF NRs are well combined with polyvinyl alcohol (PVA) sponge by covering the surface of PVA fibers or filling the interspace of PVA fibers with the help of the anhydrous ethanol solution, not only preserving ultra-high light absorption performance (Fig. 2f) but also exhibiting excellent water transport performance (Fig. 2g).

Fig. 2 Large-scale spray coating strategy of the Cu-CAT-1 MOF based light absorber. (a)–(c) Fabrication of a Cu-CAT-1 MOF or carbon black (CB) absorption layer through spray coating and assembly processes. (d) Scanning electron microscopy (SEM) images of the Cu-CAT-1 MOF NRs, sponge and Cu-CAT-1 MOF absorption layer. (e) XRD patterns of Cu-CAT-1 MOF, Cu-CAT-1 MOF crystals and the standard peaks for Cu(OH)₂ (PDF#35-0505). (f) Absorption efficiency of the CB, Cu-CAT-1 MOF and PVA sponge within the typical solar wavelength range (250–2500 nm). (g) Water contact angle of the Cu-CAT-1 MOF, PVA sponge and CB.

By using the spray coating preparation methods, the size of the single piece light absorber is only limited by the size of the substrate (PVA sponges in this work). The size of the single piece of PVA sponges used in this work was 25 cm × 25 cm. For the subsequent form of the target 1 m² device, 16 (4 × 4) pieces of PVA sponges were pieced together using assembly (Fig. S1).

Fig. 2e shows the XRD patterns of Cu-CAT-1 MOF, further confirming the successful synthesis of the target material, with peaks that closely match the Cu-CAT-1 MOF crystals with patterns form of Cu(OH)₂. The absorption efficiency results of CB, Cu-CAT-1 MOF, and PVA sponge in the range of 250 nm to 2500 nm are shown in Fig. 2f. The average absorption efficiency of CB and Cu-CAT-1 MOF was 98.1% for CB and 97.4% with the loading concentration of 0.04 kg m⁻², respectively. Note that Cu-CAT-1 MOF showed inferior absorption in the band of 450–500 nm (leading to the dark blue color of the material) and >2000 nm.⁴⁵ The untreated PVA sponge shows the lowest absorption of 23.3%. The wettability of CB and Cu-CAT-1 MOF coated absorption layer is shown in Fig. 2g. The wettability of PVA sponge substrate is also shown here for reference. The photographs of droplets are shown for all three cases at three different times, i.e., 0 ms, 10 ms, and the time for complete absorption (15 ms for Cu-CAT-1 MOF, 70 ms for PVA sponge, and 400 ms for CB). The Cu-CAT-1 MOF shows the best performance in terms of hydrophilicity, while CB is intrinsically hydrophobic, leading to an inferior water wettability.⁴⁶ More details about the material synthesis and characterization are shown in Section S1.

Indoor test

Fig. 3a shows the SMIE device in high salinity mode. The SMIE device with low salinity mode is shown in Fig. S2. The two operation modes of the SMIE device are distinguished using an additional seawater outlet (the red dashed box in Fig. 3a), as shown in Fig. S3. The width of the additional seawater outlet is 3 cm for the lab-scale SMIE device in this work as shown in Fig. S17. The use of the additional seawater outlet (Fig. 3a) can benefit from facile brine water discharge to avoid salt precipitation in the device at the cost of higher heat losses.⁴⁷ In this study, the low salinity mode of the device was utilized to process seawater with a salinity <3.5 wt%. The schematic diagram of the SMIE device is shown in Fig. 1. Solar irradiation penetrated the top transparent sealing film (polyethylene, PE) and was absorbed by the absorption layer (10 cm × 10 cm) thereby heating the seawater in the evaporation layer (10 cm × 10 cm) to generate vapor. The seawater passively entered the evaporation layer through capillary force using regenerated cellulose tissue fibers. The generated vapor was then forced to move downwards through the micropores of a PTFE membrane (12 cm × 12 cm) into the condensation layer (10 cm × 10 cm), where it condensed. The latent heat released by condensed vapor in the condensation layer was utilized to drive further distillation processes in subsequent stages. The distilled water that accumulated in the condenser was discharged into a freshwater collection tank by gravity. To maintain the bottom temperature the same as the ambient temperature, a cooling layer was incorporated at the bottom of the SMIE device to facilitate the condensation process. More details about the construction of the lab-scale SMIE device are shown in Section S2.

Fig. 3 Laboratory test of a 100 cm² device under 1-sun condition. (a) Experimental setup of the SMIE device operating in the high salinity mode. (b) The evaporation rate and evaporation efficiency of the CB-SMIE device and the MOF-SMIE device with different stages (n). (c) The mass change and evaporation rate of the MOF-SMIE and CB-SMIE devices (n = 10; 0.1 wt%). (d) The surface temperatures of the CB and Cu-CAT-1 MOF absorption layer in dry and wet states. (e) The experimental comparison of the MOF-SMIE device (n = 10; 3.5 wt% seawater) in high and low salinity modes under 1 sun condition.

Experiments utilized a lab-scale experimental setup under 1 sun condition produced by a AAA solar simulator to evaluate the freshwater production performance of the SMIE device. The experimental processes for lab-scale SMIE device operating under indoor conditions in both low and high salinity modes are shown in Fig. S4. The effect of the number of stages of SMIE devices with different absorption layers (CB-SMIE device – 0.04 kg m⁻² and MOF-SMIE device – 0.04 kg m⁻²) on evaporation rate (r_evap) and corresponding evaporation efficiency (η_evap) is shown in Fig. 3b. The mass changes (m) of the collected freshwater were monitored (Fig. 3c), and the average freshwater production rates for various stage designs of MOF-SMIE device, i.e., 2, 4, 6, 8, and 10 stages, were determined to be 1.74 kg m⁻² h⁻¹, 2.88 kg m⁻² h⁻¹, 4.15 kg m⁻² h⁻¹, 5.10 kg m⁻² h⁻¹, and 5.45 kg m⁻² h⁻¹, respectively, as shown in Fig. 3b. Meanwhile, Fig. 3b assesses the impact of various photothermal materials (CB – 0.04 kg m⁻² and Cu-CAT-1 MOF – 0.04 kg m⁻²) on the evaporation performance. We utilized a consistent preparation method to synthesize a CB absorption layer and simultaneously characterized the materials. In contrast, r_evap of the CB-SMIE device is slightly lower (1.56 kg m⁻² h⁻¹, 2.74 kg m⁻² h⁻¹, 3.80 kg m⁻² h⁻¹, 4.20 kg m⁻² h⁻¹ and 4.45 kg m⁻² h⁻¹) than that of the MOF-SMIE device. This can be explained by the lower surface and side heat losses resulting from the lower surface temperature of the MOF-SMIE device (52 °C) than that of the CB-SMIE device (55 °C) in the wet state and n = 10 as shown in Fig. 3d and Fig. S5. More details about the energy breakdown of the MOF-SMIE device (52 °C) and the CB-SMIE device (55 °C) are shown in Fig. S6. Although the surface temperature of the MOF (77 °C) is higher than that of the CB (75 °C) in the dry state (Fig. 3d), the MOF has a lower surface temperature as an absorber in the SMIE device due to its superior water wettability mentioned above.

In addition, the effect of Cu-CAT-1 MOF loading concentration, in the range of 0–0.08 kg m⁻², on the optical absorption efficiency and evaporation rate of a 2-stage SMIE was experimentally characterized as shown in Fig. S7. The physical and SEM pictures of the SMIE absorption layer with different Cu-CAT-1 MOF loading concentrations are shown in Fig. S8 and S9. The light absorption efficiency and evaporation rate increased with the increase of Cu-CAT-1 MOF loading concentration accordingly. Further increase in loading did not significantly improve the evaporation rate after the loading concentration exceeded 0.04 kg m⁻². This is because the accumulation of Cu-CAT-1 MOF particles on the evaporator surface has a negative effect on vapor transport, although further increases in Cu-CAT-1 MOF loading concentration can improve absorption efficiency.⁴⁸ For example, the absorption efficiency and evaporation rate for the SMIE device (n = 2) with MOF loading concentration of 0.04 kg m⁻² were 92.9% and 1.75 kg m⁻² h⁻¹, respectively. For a concentration of 0.08 kg m⁻², these were 97.4% and 1.79 kg m⁻² h⁻¹, respectively. Thus, 0.04 kg m⁻² was chosen as the optimal loading concentration for Cu-CAT-1 MOF to balance cost and efficiency in this work. Similar to the MOF, the optical and evaporation performance of the CB-SMIE device did not further increase with the increase of CB loading concentration when the loading concentration exceeds 0.04 kg m⁻² (Fig. S10).

We experimentally compared the MOF-SMIE device (n = 10) using 3.5 wt% simulated seawater under both high- and low-salinity modes at 1 sun to quantify the short-term reduction in freshwater production and to highlight the long-term stability advantage of operating in high-salinity mode. As shown in Fig. 3e, the freshwater production rate of the MOF-SMIE device (n = 10; 3.5 wt% seawater) in high salinity mode (4.0 kg m⁻² h⁻¹) is lower than that in low salinity mode (5.5 kg m⁻² h⁻¹) in the first 60 minutes. In low salinity mode, the freshwater production rate gradually declined and eventually stopped after approximately 200 minutes of operation, which is attributed to salt accumulation and blockage within the pores of the porous PTFE membrane due to insufficient ion discharge. In high salinity mode, although increased brine outflow resulted in greater heat loss and a reduced volume of collected condensate, the continuous removal of concentrated saltwater effectively prevented salt accumulation in the membrane, thereby enabling stable long-term operation of the SMIE device. Under 1 sun condition, i.e., 1 kW m⁻², the SMIE device exhibited stable operation for more than 10 hours in simulated seawater without crystallization, achieving the freshwater production rate of 3.9 kg m⁻² h⁻¹ as shown in Fig. 3e. Experimental results were all obtained at a temperature of ∼25 °C and a humidity of ∼40%.

Thermal and water management

To gain insights into the impact of device design (Fig. 4a) on the water production rate of the SMIE device, a lumped parameter model was utilized for the prediction of SMIE performance and energy breakdown. More details about the model are shown in Section S3. The effect of the insulation layer's thickness (L_insulation) on evaporation rate (r_evap) and heat losses for the lab-scale (mm²) SMIE device is shown in Fig. 4b and c. In general, r_evap of the MOF-SMIE device (n = 10) increased from 4.79 kg m⁻² h⁻¹ to 5.44 kg m⁻² h⁻¹ with increasing L_insulation from 1 mm to 15 mm, as shown in Fig. 4c. This can be attributed to the linear decrease in heat convection (η_side-conv; blue bar in Fig. 4b) and heat conduction (η_side-cond; orange bar in Fig. 4b) from the side of the SMIE device. The radiation heat loss and convection heat loss from the absorption layer, i.e., η_rad and η_conv (red bar and green bar in Fig. 4c), showed a slight increase as a function of L_insulation, which can be associated with an increase in the top surface temperature (Fig. S11). To control the η_side-cond close to 5%, for the lab-scale SMIE, an insulation thickness of 5 mm is chosen (star in Fig. 4b).

Fig. 4 Impact of heat and mass transfer on the performance of the SMIE device. (a) Schematics of important parameters, the thickness of the side insulation (L_insulation) and wind speed (v) affecting the heat and mass transfer in the 1 m² SMIE device. The effect of L_insulation on side (b) and overall (c) heat losses for a lab-scale SMIE device (n = 10; v = 0.5 m s⁻¹). The effect of L_insulation on side (d) and overall (e) heat losses for a large-scale SMIE device (n = 10; v = 4.0 m s⁻¹). The diffusion distance of the SMIE device with single (f) and multi-channel (g) strategy at 2000 s. (h) The wetted area with wicking rate >1.45 kg m⁻² h⁻¹ as a function of time for both single and multiple channel strategies with different wicking channel width (L_supply).

For the large-scale SMIE device, the active area is 1 m² in this study for real-world application outdoor environment. The typical wind speed in outdoor tests (v = 4 m s⁻¹ in this work) is generally much higher than that of the indoor environment (v = 0.5 m s⁻¹). This will lead to an increase in η_conv (>25% in all cases in Fig. 4e) from the top surface of SMIE. Also, η_side-loss was increased from 14.06% (Fig. 4c) to 34.42% (Fig. 4e) when scaling the SMIE device from 100 cm² to 1 m² at L_insulation = 1 mm. This can be attributed to the reduction in radiation loss as a result of the lower top surface temperature of the large-scale SMIE device compared to the lab-scale SMIE device, as shown in Fig. S11. For a large-scale outdoor SMIE device, a minimum thickness of the insulation layer should be identified to ensure η_side-cond < 5%. Hence, the minimum L_insulation for reducing η_side-cond was increased to 10 mm, as shown in Fig. 4d, suggesting that the insulation layer has to be optimized under different operation conditions and device designs. The r_evap hence increased from 3.45 kg m⁻² h⁻¹ at L_insulation = 1 mm to 5.19 kg m⁻² h⁻¹ at L_insulation = 10 mm for a large-scale SMIE device (n = 10). The lower r_evap of the large-scale SMIE device compared to the lab-scale SMIE device is mainly due to the increased heat loss from the top surface as a result of higher wind speed as well as due to higher edge thickness as a result of multi-channel design.

Water transport management is another important aspect for optimal SMIE device designs, especially for large-scale devices where the water transport distance may exceed the wicking distance. For quantitative analysis of water transport, a finite element model was developed to describe the water transport in the SMIE device for both single-channel water supply (Fig. 4f) and multi-channel water supply (Fig. 4g). More details about the model are shown in Section S3. The numerical model was validated using experimental data, which showed a good match as long as the required distance was >20 cm (see Fig. S12). The contour plots of local water flux at 2000 s for both cases are exhibited in Fig. 4f and g. At 0 s, the whole SMIE device was in a dry condition. The dash isolines are for the water flux equal to 1.47 kg m⁻² h⁻¹ for a selected time series. 1.47 kg m⁻² h⁻¹ represents the maximum local evaporation rate without considering super thermodynamic limit phenomena in special materials, e.g., hydrogels.^49,50 Note that even at 2000 s water absorption, the area with water flux >1.47 kg m⁻² h⁻¹, defined as A_wet, was only 50%, while this area increased to ∼95% for the multi-channel water supply (Fig. 4h). Furthermore, A_wet/A as a function of time under various wicking channel widths (L_supply) for both single and multi-channel designs is shown in Fig. 4h. As expected, more channel numbers and larger L_supply are generally favorable for better water supply performance. However, for L_supply > 50 mm, further benefit from larger L_supply is minimized, indicating L_supply = 50 mm can be a reasonable design for a 1 m² SMIE device.

Outdoor test and technoeconomic analysis

To further investigate the performance of the large-scale SMIE (1 m²) device under real-world weather conditions, outdoor experiments were conducted on sunny days from July 25th to 27th, 2023. The detailed fabrication process of the large-scale device can be found in Fig. S13 and Section S2. As shown in Fig. 5a and Fig. S18, the CB-SMIE device (right) and MOF-SMIE device (left) experimental setups with high salinity mode were placed adjacent to each other on the roof of SUSTech, College of Engineering, North Building. The width of the additional seawater outlet is 0.5 m for large-scale SMIE devices. A cooling layer was incorporated at the bottom of the SMIE device to maintain the temperature of the final layer close to the ambient temperature, facilitating the condensation process.

Fig. 5 Outdoor tests of the large-scale SMIE device at SUSTech, College of Engineering, North Building on a sunny day (July 25th, 2023) and the technoeconomic analysis. (a) Experimental setup of the large-scale CB-SMIE and MOF-SMIE devices. (b) Solar radiation, temperature and water collection rate as a function of testing time. (c) The ion concentrations in raw seawater and the collected fresh water. (d) The schematic diagram of the temperature measurement point layout of the SMIE device. (e) Comparison of the evaporation rate/freshwater production rate and the area of the evaporation surface between our SMIE device and data from the literature. (f) The cost of freshwater production as a function of time, where the CB-SMIE (blue curve) and MOF-SMIE (orange curve) devices were compared with the bottled water price (red curve) and tap water price (green curve). (g) The PP as a function of evaporation rate (r_evap) and SMIE device cost (C_SMIE). (h) The cost proportion of the SMIE device, including PTFE membrane cost, absorption layer cost, and other components costs (cellulose tissue cost, PE membrane cost and other costs).

The outdoor experiment was carried out from 08 : 00 to 17 : 00 (local time). Fig. 5b shows the recorded temperature responses of each SMIE device on July 25th, 2023. The experimental results on the 26th and 27th are shown in Fig. S14. The schematic diagram of the temperature measurement point layout of the SMIE device is shown in Fig. 5d. As shown in Fig. 5b, the temperature of the absorption layer increased steadily in the first 3 hours, reaching ∼60 °C for the MOF-SMIE device and 61 °C for the CB-SMIE device. The infrared images of CB-SMIE and MOF-SMIE devices at different times are shown in Fig. S15. Although the surface temperatures of the CB-SMIE and MOF-SMIE devices were similar, the CB-SMIE device absorbed less water due to its lower water absorption capacity. Consequently, it had a lower evaporation rate and collected less condensed water compared to the MOF-SMIE device. During the 9-hour operation period, the average solar flux was measured to be 651 W m⁻². The water production results, presented in Fig. 5b, clearly demonstrate the superior performance of the MOF-SMIE device, which produced 32 L of freshwater (32 L m⁻² day⁻¹) with an average water production rate of 3.5 L m⁻² h⁻¹ (345% efficiency).

In addition, the accelerated aging tests were conducted. The CB-SMIE and MOF-SMIE (n = 10) devices were continuously exposed to a solar intensity of 5 suns for 24 hours. The absorption and evaporation performances of the CB-SMIE and MOF-SMIE devices before and after irradiation were then compared. As shown in Fig. S16, the absorption and evaporation rates of the CB-SMIE device before and after irradiation were 98.15%, 4.45 kg m⁻² h⁻¹ and 95.02%, 4.26 kg m⁻² h⁻¹, respectively, while those of the MOF-SMIE device were 97.327%, 5.45 kg m⁻² h⁻¹ and 95.88%, 5.32 kg m⁻² h⁻¹. These results demonstrate that the evaporators are capable of maintaining high efficiency and long-term operational stability. The ion concentrations of the collected water and source water were measured by ICP-OES, and the results are shown in Fig. 5c. The results demonstrate that the major ions (i.e., Na⁺, Ca²⁺, Mg²⁺, K⁺) concentrations of the collected water were all lower than 10 ppm, indicating satisfactory water quality for meeting the drinking water guideline of the World Health Organization (WHO) for these ions.⁵¹ This result demonstrates that the collected clean water is safe enough for drinking.

Fig. 5e compares the evaporation rate (r_evap)/freshwater production rate (r_fre) and effective evaporation area of our large-scale SMIE device with state-of-the-art passive solar desalination systems reported in the literature. More details about r_evap or r_fre and area of the works mentioned in Fig. 5e are shown in Table S1. Here, r_evap is represented by the solid pattern and r_fre is represented by the hollow pattern. The square point in the blue area in Fig. 5e represents the typical double-layer evaporator, the triangle point in the orange area represents the hydrogel evaporator, and the diamond point in the green area represents the multi-stage evaporator. Typical double-layer evaporators are more suitable for large-scale applications, and multi-stage evaporators are more conducive to enhancing the evaporation rate. Meanwhile, we think that SMIE devices in the meter level have greater potential for large-scale industrial applications. Our SMIE device achieves a solar-to-vapor conversion efficiency of 365% under real outdoor conditions with an active area of 1 m², significantly outperforming most reported systems considering both performance and area simultaneously. For example, Chen et al. constructed a 150 m² evaporation system, achieving an average freshwater production rate of ∼0.5 L m⁻² h⁻¹.³⁰ However, the evaporation system was formed by connecting two different sizes (0.6 m × 2.0 m; 1.2 m × 1.9 m) of meter level evaporators in series, which were similar in size to our large-scale SMIE devices (1 m × 1 m) but significantly lower in evaporation performance (3.5 L m⁻² h⁻¹ in this study).

We performed a technoeconomic analysis to assess the translational potential of the SMIE device for large-scale freshwater production. Fig. 5f shows the cost of freshwater production as a function of time, where the CB-SMIE (blue curve) and MOF-SMIE (orange curve) devices were compared with the bottled water price (red curve) and tap water price (green curve). The general approach of technoeconomic analysis has been demonstrated in our previous work.⁵² More details about the technoeconomic analysis of SMIE can be found in Section S4. We take the time when the price of fresh water equals the price of tap water as the payback (PP) period. The PP as a function of evaporation rate (r_evap) and SMIE device cost (C_SMIE) is shown in Fig. 5g. Although the MOF-SMIE device demonstrated better evaporation performance (point A), its higher cost ($1173.31) resulted in a significantly longer payback period (139.68 years) compared to the CB-SMIE device (point B; $122.29; 14.55 years). Furthermore, the cost proportion of the SMIE device was divided, including PTFE membrane cost, absorption layer cost, and other components cost (cellulose tissue cost, PE membrane cost and other costs), as shown in Fig. 5h. When MOF was used as the photothermal material for the SMIE device, the cost of the absorption layer occupied a dominant position in the C_SMIE proportion, specifically at 94.30%. When CB was used, the proportion of the absorption layer cost in C_SMIE decreased to 45.26%. The cost of the porous PTFE membrane (49.09%) will dominate the C_SMIE. Further reducing the cost of the absorption layer and the porous PTFE membrane is beneficial for shortening the PP. For example, surface carbonized wood used as the absorption layer led to lower costs, at only $ 3 per square meter.⁵³ Meanwhile, reducing the cost of the porous PTFE membrane by half can control the C_SMIE to <$ 40 per square meter, with PP = 4.76 years, as shown at point C in Fig. 5g and h.

In this work, all analyses were conducted based on the operation of a single meter-scale SMIE device. Scaling up by operating multiple meter-scale devices in parallel to form a larger system still poses certain challenges. Firstly, additional costs need to be taken into account. Additional operating costs include expenses related to seawater pretreatment, costs for monitoring, collection, maintenance, labor, and electricity. According to ref. 30, the annual operating cost is approximately $10 per year, which has a negligible impact on the payback period. The additional initial investment includes the cost of system parallelization, control systems, land leasing, and other related costs. According to ref. 30, the control system cost accounts for the main portion of the initial investment cost, ∼$100 000, aimed at achieving high precision automation for subsequent research and insights. The overall cost can be reduced by adopting more economical control systems. Moreover, as the system's operational lifespan increases, the influence of the control system costs on the payback period gradually decreases. Meanwhile, under outdoor conditions, sunlight and warm water create favorable conditions for the growth of microorganisms, such as green algae, particularly in the freshwater collection tank. To prevent this situation, it is necessary to perform microfiltration pretreatment of seawater and regularly clean and maintain the freshwater collection tank. In addition, the discharge of high-salinity brine can be harmful to the environment and should be addressed through the zero liquid discharge (ZLD) strategy.

Conclusions

Our study presents a highly efficient and scalable solar multi-stage interfacial evaporation device with an active area of 1 m², designed to address the challenges of material scalability, energy losses, and insufficient water supply in large-scale solar desalination. The device integrates an inverted multi-stage configuration, a scalable photothermal absorption layer based on either Cu-CAT-1 MOF or carbon black, enhanced thermal insulation, and structured wicking channels, resulting in improved thermal performance and freshwater yield at the meter scale. The 1 m² SMIE prototype achieved a solar-to-vapor conversion efficiency of 365% under outdoor conditions, with a freshwater production rate of 3.5 L m⁻² h⁻¹, corresponding to 32 L m⁻² day⁻¹. These findings demonstrate that the integrated design strategies effectively enable performance scaling while maintaining efficiency, thus narrowing the gap between laboratory-scale research and field-deployable desalination technologies. Our techno-economic analysis indicates that further reducing the costs of the absorption layer and porous membrane is key to shortening the payback period of the SMIE device. By demonstrating a scalable, high-efficiency solar interfacial evaporation system with validated performance under real-world conditions and economically relevant design parameters, this work provides a foundational framework for the development of modular, passive desalination technologies capable of addressing water scarcity in off-grid and resource-constrained environments.

Materials

N,N-Dimethylformamide (DMF, 99.8%) and 2, 3, 6, 7, 10, 11-hexahydroxytriphenylene (HHTP, 95%) were purchased from Sigma-Aldrich. Copper hydroxide nanowires were purchased from Zhongke Leiming (Beijing) Technology Co., Ltd. The carbon black particles were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. The cellulose sponge was commercially acquired from Meilaijie Home Furnishing Flagship Store. All materials and chemicals were used as received without further purification.

Characterization

All evaporation experiments in the laboratory were conducted under a solar simulator (Newport; Sol3A). Mass change was tested with an electronic microbalance (Mettler; ME204/02). The scanning electron microscopy (SEM) images were obtained using a FEI Nova NanoSEM 450. Absorption spectra were collected using a UV-VIS-NIR spectrophotometer (Lambda 750s). More details about the experimental instruments are shown in Table S2.

Calculation of the evaporation performance

The definitions and calculation methods for evaporation rate and efficiency can be found in our previous work.^45,54

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.

Supplementary information including supplementary figures, materials characterization, construction and performance measurements, model development, and techno-economic analysis. See DOI: https://doi.org/10.1039/d5ee01958c.

Acknowledgements

The authors acknowledge the National Natural Science Foundation of China under Grant No. 52376191 for their support. The Shenzhen Science and Technology Innovation Commission under Grant No. 20231120185819001 and KCXST20221021111207017, SUSTech High Level of Special Funds under Grant No. G03034K001, the Guangdong Basic and Applied Basic Research Foundation under Grant No. 2023A1515011595 and 2023B0303000002, and the Guangdong grant under Grant No. 2021QN02L562 are also acknowledged for their support. The SEM data were obtained using the equipment maintained by Southern University of Science and Technology Core Research Facilities. Computation in this work is supported by the Center for Computational Science and Engineering at Southern University of Science and Technology.

References

1. P. H. Gleick, Global freshwater resources: soft-path solutions for the 21st century, Science, 2003, 302(5650), 1524–1528.
2. H. Ghasemi, G. Ni and A. M. Marconnet, et al., Solar steam generation by heat localization, Nat. Commun., 2014, 5(1), 4449.
3. B. Yang, Z. Zhang and P. Liu, et al., Flatband λ-Ti3O5 towards extraordinary solar steam generation, Nature, 2023, 622(7983), 499–506.
4. N. Xu, J. Li and C. Finnerty, et al., Going beyond efficiency for solar evaporation, Nat. Water, 2023, 1(6), 494–501.
5. H. Xu, Taking solar evaporation technologies to a new era, Nat. Water, 2025, 1–2.
6. H. Li, W. Zhang and X. Liao, et al., Kirigami enabled reconfigurable three-dimensional evaporator arrays for dynamic solar tracking and high efficiency desalination, Sci. Adv., 2024, 10(26), eado1019.
7. Y. Song, S. Fang and N. Xu, et al., Solar-driven interfacial evaporation technologies for food, energy and water, Nat. Rev. Clean Technol., 2025, 1(1), 55–74.
8. B. Xu, M. Ganesan and R. K. Devi, et al., Hierarchically Promoted Light Harvesting and Management in Photothermal Solar Steam Generation, Adv. Mater., 2025, 37(5), 2406666.
9. K. Mao, Y. Zhang and S. C. Tan, Functionalizing solar-driven steam generation towards water and energy sustainability, Nat. Water, 2025, 1–13.
10. L. Zhang, Z. Xu and L. Zhao, et al., Passive, high-efficiency thermally-localized solar desalination, Energy Environ. Sci., 2021, 14(4), 1771–1793.
11. J. Gao, L. Zhang and J. You, et al., Extreme salt-resisting multistage solar distillation with thermohaline convection, Joule, 2023, 7(10), 2274–2290.
12. X. Li, X. Min and J. Li, et al., Storage and recycling of interfacial solar steam enthalpy, Joule, 2018, 2(11), 2477–2484.
13. X. Luo, L. Jiao and Y. Guo, et al., Ultrahigh freshwater production achieved by unidirectional heat transfer interfacial evaporation solar still integrated with waste heat recovery, Energy Convers. Manage., 2024, 304, 118226.
14. H. Ren, M. Tang and B. Guan, et al., Hierarchical graphene foam for efficient omnidirectional solar–thermal energy conversion, Adv. Mater., 2017, 29(38), 1702590.
15. H. Liang, Q. Liao and N. Chen, et al., Thermal efficiency of solar steam generation approaching 100% through capillary water transport, Angew. Chem., Int. Ed., 2019, 58(52), 19041–19046.
16. P. Zhang, J. Li and L. Lv, et al., Vertically aligned graphene sheets membrane for highly efficient solar thermal generation of clean water, ACS Nano, 2017, 11(5), 5087–5093.
17. A. K. Menon, I. Haechler and S. Kaur, et al., Enhanced solar evaporation using a photo-thermal umbrella for wastewater management, Nat. Sustainability, 2020, 3(2), 144–151.
18. T. A. Cooper, S. H. Zandavi and G. W. Ni, et al., Contactless steam generation and superheating under one sun illumination, Nat. Commun., 2018, 9(1), 5086.
19. S. Zheng, J. Yu and Z. Xu, Modeling and analysis of contactless solar evaporation for scalable application, Appl. Sci., 2023, 13(6), 4052.
20. Y. Zhang, T. Xiong and L. Suresh, et al., Guaranteeing complete salt rejection by channeling saline water through fluidic photothermal structure toward synergistic zero energy clean water production and in situ energy generation, ACS Energy Lett., 2020, 5(11), 3397–3404.
21. Y. Zhang, H. Zhang and T. Xiong, et al., Manipulating unidirectional fluid transportation to drive sustainable solar water extraction and brine-drenching induced energy generation, Energy Environ. Sci., 2020, 13(12), 4891–4902.
22. Y. Shao, A. Shen and N. Li, et al., Marangoni effect drives salt crystallization away from the distillation zone for large-scale continuous solar passive desalination, ACS Appl. Mater. Interfaces, 2022, 14(26), 30324–30331.
23. Z. Xu, L. Zhang and L. Zhao, et al., Ultrahigh-efficiency desalination via a thermally-localized multistage solar still, Energy Environ. Sci., 2020, 13(3), 830–839.
24. L. Zhang, Z. Xu and B. Bhatia, et al., Modeling and performance analysis of high-efficiency thermally-localized multistage solar stills, Appl. Energy, 2020, 266, 114864.
25. Y. Wang, T. Wei and Y. Wang, et al., Quasi-waffle solar distiller for durable desalination of seawater, Sci. Adv., 2024, 10(22), eadk1113.
26. Y. Wang, W. Zhao and Y. Lee, et al., Thermo-adaptive interfacial solar evaporation enhanced by dynamic water gating, Nat. Commun., 2024, 15(1), 6157.
27. S. Liu, C. Huang and Q. Huang, et al., A new carbon-black/cellulose-sponge system with water supplied by injection for enhancing solar vapor generation, J. Mater. Chem. A, 2019, 7(30), 17954–17965.
28. L. Fan, G. Liu and F. Wang, et al., Domestic water consumption under intermittent and continuous modes of water supply, Water Resour. Manage., 2014, 28, 853–865.
29. N. Xu, H. Zhang and Z. Lin, et al., A scalable fish-school inspired self-assembled particle system for solar-powered water-solute separation, Natl. Sci. Rev., 2021, 8(10), nwab065.
30. Y. Chen, L. Shen and Z. Qi, et al., Large-scale implementation of solar interfacial desalination, Nat. Sustainability, 2025, 8(2), 162–169.
31. X. Luo, J. Shi and C. Zhao, et al., The energy efficiency of interfacial solar desalination, Appl. Energy, 2021, 302, 117581.
32. S. Zheng, B. Oelckers and A. Khandelwal, et al., Interfacial solar evaporation for sustainable brine mining, Nat. Water, 2025, 1–3.
33. Y. Pu, W. Lin and X. Yao, et al., Large-scale 3D printed fouling-resistant self-floating evaporator, Nat. Commun., 2025, 16(1), 3677.
34. S. Liu and C. Huang, A stable and flexible carbon black/polyethyleneimine-bacterial cellulose photothermal membrane for high-efficiency solar vapor generation, Int. J. Energy Res., 2020, 44(11), 8904–8918.
35. X. Li, G. Ni and T. Cooper, et al., Measuring conversion efficiency of solar vapor generation, Joule, 2019, 3(8), 1798–1803.
36. H. T. Kim, L. Philip and A. McDonagh, et al., Recent advances in high-rate solar-driven interfacial evaporation, Adv. Sci., 2024, 11(26), 2401322.
37. Y. Chen, X. Luo and X. Gu, et al., The energy efficiency of multistage interfacial desalination: Comprehensive analysis and further improvement strategy, Desalination, 2023, 568, 117000.
38. S. Liu, S. Li and Q. Yang, et al., Synergistic optical and thermal management for solar water and electricity co-generation via a front-side coupling strategy, Cell Rep. Phys. Sci., 2025 DOI:10.1016/j.xcrp.2025.102720.
39. F. Nawaz, Y. Yang and Q. Zhao, et al., Can the interfacial solar vapor generation performance be really “beyond” theoretical limit?, Adv. Energy Mater., 2024, 14(22), 2400135.
40. Q. Zhao, Y. Wei and Y. Yang, et al., Architectural and real-time monitoring design of multi-stage solar still for solar water purification, Renewable Energy, 2025, 122660.
41. J. Cai, T. Jin and J. Kou, et al., Lucas–Washburn equation-based modeling of capillary-driven flow in porous systems, Langmuir, 2021, 37(5), 1623–1636.
42. A. Hamraoui and T. Nylander, Analytical approach for the Lucas–Washburn equation, J. Colloid Interface Sci., 2002, 250(2), 415–421.
43. D. I. Dimitrov, A. Milchev and K. Binder, Capillary rise in nanopores: molecular dynamics evidence for the Lucas–Washburn equation, Phys. Rev. Lett., 2007, 99(5), 054501.
44. L. Hou, S. Li and Y. Qi, et al., Advancing Efficiency in Solar-Driven Interfacial Evaporation: Strategies and Applications, ACS Nano, 2025, 19(10), 9636–9683.
45. Q. Ma, P. Yin and M. Zhao, et al., MOF-based hierarchical structures for solar-thermal clean water production, Adv. Mater., 2019, 31(17), 1808249.
46. A. Asthana, T. Maitra and R. Büchel, et al., Multifunctional superhydrophobic polymer/carbon nanocomposites: graphene, carbon nanotubes, or carbon black, ACS Appl. Mater. Interfaces, 2014, 6(11), 8859–8867.
47. S. Liu, S. Li and M. Lin, Understanding interfacial properties for enhanced solar evaporation devices: from geometrical to physical interfaces, ACS Energy Lett., 2023, 8(4), 1680–1687.
48. S. Liu, C. Huang and X. Luo, et al., High-performance solar steam generation of a paper-based carbon particle system, Appl. Therm. Eng., 2018, 142, 566–572.
49. J. Zhou, Y. Gu and P. Liu, et al., Development and evolution of the system structure for highly efficient solar steam generation from zero to three dimensions, Adv. Funct. Mater., 2019, 29(50), 1903255.
50. F. Zhao, X. Zhou and Y. Shi, et al., Highly efficient solar vapour generation via hierarchically nanostructured gels, Nat. Nanotechnol., 2018, 13(6), 489–495.
51. S. H. Frisbie, E. J. Mitchell and H. Dustin, et al., World Health Organization discontinues its drinking-water guideline for manganese, Environ. Health Perspect., 2012, 120(6), 775–778.
52. S. Liu, Q. Yang and S. Li, et al., A comprehensive review of salt rejection and mitigation strategies in solar interfacial evaporation systems, Desalination, 2024, 118507.
53. T. Li, H. Liu and X. Zhao, et al., Scalable and highly efficient mesoporous wood-based solar steam generation device: localized heat, rapid water transport, Adv. Funct. Mater., 2018, 28(16), 1707134.
54. S. Liu, C. Huang and X. Luo, et al., Performance optimization of bi-layer solar steam generation system through tuning porosity of bottom layer, Appl. Energy, 2019, 239, 504–513.

---

Footnote

† These authors contributed equally to this work.

---