You are on page 1of 8

Articles

https://doi.org/10.1038/s41565-018-0097-z

Highly efficient solar vapour generation via


hierarchically nanostructured gels
Fei Zhao1,4, Xingyi Zhou1,4, Ye Shi1, Xin Qian2, Megan Alexander1, Xinpeng Zhao2, Samantha Mendez1,
Ronggui Yang2*, Liangti Qu3* and Guihua Yu1*

Solar vapour generation is an efficient way of harvesting solar energy for the purification of polluted or saline water. However,
water evaporation suffers from either inefficient utilization of solar energy or relies on complex and expensive light-concen-
tration accessories. Here, we demonstrate a hierarchically nanostructured gel (HNG) based on polyvinyl alcohol (PVA) and
polypyrrole (PPy) that serves as an independent solar vapour generator. The converted energy can be utilized in situ to power
the vaporization of water contained in the molecular meshes of the PVA network, where water evaporation is facilitated by the
skeleton of the hydrogel. A floating HNG sample evaporated water with a record high rate of 3.2 kg m−2 h−1 via 94% solar energy
from 1 sun irradiation, and 18–23 litres of water per square metre of HNG was delivered daily when purifying brine water. These
values were achievable due to the reduced latent heat of water evaporation in the molecular mesh under natural sunlight.

S
olar energy is a promising source for boosting the evolution of network, the converted solar energy can be directly delivered to the
renewable energy technology. Despite the unmatched resource small amount of water in the molecular meshes (see 1 in Fig. 1b).
potential of solar energy, the restricted utilization efficiency PVA chains eradicate the convective heat loss of water30,31, which is
presents an enormous challenge1–3. Solar-to-thermal technologies, conventionally the main energy loss channel in solar vapour gen-
such as domestic heating4, brine desalination5–8 and power gen- erators with thermal localization effects. Additionally, the rapid
eration9–11, have been the subject of both academic research and water diffusion and capillary pumping of micron channels (see 2 in
industrialization efforts throughout the past decades. Solar vapour Fig. 1b) and internal gaps (see 3 in Fig. 1b) lead to rapid replenish-
generation that can directly transfer heat to facilitate evaporation ment of the molecular meshes through swelling of the polymeric
is an efficient way to harvest solar energy and to purify water12. network to support a sustained high rate vapour generation. A high
As solar radiation serves as the only power input for vapour gen- water evaporation rate of 3.2 kg m−2 h−1, almost two times higher
eration, a variety of materials, including ultra-black absorbers13–16, than the reported record19, was achieved under 1 sun solar radia-
plasmonic nanoparticles17–20 and thermal-concentrating ceramics21, tion with excellent stability and durability. Furthermore, the HNG
have been used to enhance the conversion efficiency. However, the exhibited a daily solar water purification yield of 20.3 l m−2 under
inefficient utilization of converted heat imposes an additional chal- natural sunlight, indicating its potential for scalable manufacturing
lenge. Among the three major heat consumption processes, that is, for practical applications in the future.
water heating, parasitic thermal loss and water vaporization, only
water vaporization is effective8. Therefore, adequate thermal man- Fabrication and characterization of the HNGs
agement in the solar vapour generation system is essential. The heat We used the in situ polymerization method for preparing the inter-
localization strategy, which restricts heat into a small amount of penetrating structured HNGs30 (see Supplementary Methods for
water in the evaporating surface, may substantially improve the effi- details). As PVA plays the role of surfactant to assist the dispersion
ciency of solar energy utilization22–24. Though it has been demon- of PPy in water, the basic building block (precursor) of HNG was a
strated that efficiency can be increased by using concentrated solar PVA–PPy cluster (Supplementary Fig. 1). After gelation, the spac-
power with high flux values (10 kW m−2 or higher)19,25–27, it would be ing among these clusters formed the internal gaps and the clusters
ideal to achieve highly efficient solar vapour generation under 1 sun formed the skeleton of the HNG with micron channels. The as-pre-
(1 kW m−2) or even weaker natural daylight. pared HNGs were black (Fig. 2a), flexible and elastic (Supplementary
Here, we report an efficient in situ energy utilization strategy Fig. 2). Scanning electron microscopy (SEM) showed that the HGNs
based on HNGs to achieve high energy conversion efficiency and, exhibited broad gaps with diameters of ~150 μ​m (highlighted by the
more importantly, high water evaporation rates under 1 sun. We red broken line in Fig. 2b) that segmented the porous structure. The
present an energy confinement strategy based on water manage- pores formed micron channels that had widths ranging from several
ment through hierarchical water pathways in a hydrogel28, which microns to tens of microns (Fig. 2c). In addition, the wall structure
consists of internal gaps, micron channels and molecular meshes showed a wrinkled internal surface (Fig. 2d), indicating shrinkage
(Fig. 1a). Furthermore, the water in the molecular meshes of HNGs of the polymeric skeleton (PVA network) during dehydration of the
have a reduced vaporization enthalpy29. To accomplish solar evapo- hydrogel (Supplementary Fig. 3). Combined with the nanostruc-
ration, a floating HNG is exposed to solar irradiation (Fig. 1b). Due tures, a three-dimensional (3D) hierarchical network of water path-
to the penetration of solar absorbers (PPy) in the polymeric gel PVA ways, such as arteries and branches, was established in the HNGs.

1
Materials Science and Engineering Program and Department of Mechanical Engineering, The University of Texas at Austin, Austin, TX, USA.
2
Department of Mechanical Engineering, University of Colorado, Boulder, CO, USA. 3Key Laboratory of Photoelectronic/Eletrophotonic Conversion
Materials, School of Chemistry and Chemical Engineering, Beijing Institute of Technology, Beijing, China. 4These authors contributed equally:
F. Zhao, X. Zhou. *e-mail: ronggui.yang@colorado.edu; lqu@bit.edu.cn; ghyu@austin.utexas.edu

Nature Nanotechnology | VOL 13 | JUNE 2018 | 489–495 | www.nature.com/naturenanotechnology 489


© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Articles Nature Nanotechnology
a Water and energy management in the HNGs
The weight ratio between water and PVA was tuned to
control the nanostructure32,36 and hence regulate the
water content and transport in the HNGs with the opti-
mized PPy/PVA ratio of 1:10 (Supplementary Figs. 6–8).
HNGs with water/PVA weight ratios of 1:0.2, 1:0.15, 1:0.1 and 1:0.075
Internal gap Micron channel Molecular mesh
were prepared (designated as HNG1, HNG2, HNG3 and HNG4,
respectively) to evaluate structure-dependent water transport.
PVA PPy Crosslinking point Given that the HNGs were completely hydrated during solar vapour
b generation, the saturated water content (denoted as Qs; Fig. 3a) and
Water transport
Confined
the half swollen time (blue curve in Fig. 3a; see Supplementary Fig.
Solar evaporation 9 for more details) of different HNG samples were measured to cal-
radiation 1
culate the water transport rate (denoted as V; red curve in Fig. 3b).
Bran
ch The results confirmed that the Qs and V were affected by the nano-
2 diffus ed
ion structured channels (Supplementary Figs. 10 and 11).
3
HNG Ar
As shown by the ultraviolet–vis (UV–vis) near infrared (NIR)
pu teria spectra in Fig. 3c, the HNGs presented excellent absorption fea-
mp l
ing
tures throughout the solar spectra with negligible optical loss. In
Bulk wat
er
addition, the HNGs showed low reflectance over a wide wave-
length range (from 250 to 2,000 nm) (Supplementary Fig. 12a).
Consequently, the light absorption of the HNGs is not sensitive to
the Qs or V values, enabling stable solar energy harvesting during
vapour generation. In addition, a ~400-μ​m-thick HNG sheet pro-
Fig. 1 | Schematic of highly efficient solar vapour generation based on vided an average transparency of ~2% (Supplementary Fig. 12b),
tailored water transport in HNGs. a, The HNG consists of hierarchical indicating that the thickness of the effective solar energy harvesting
porous structures, including internal gaps, micron channels and molecular region in HNGs is ~400 μ​m. Moreover, the Qs and V values were
meshes, wherein the solar absorber (PPy) penetrates the polymeric PVA independent of the thickness (Supplementary Fig. 13). Despite the
network of the gel. b, Schematic of a typical solar vapour generation identical energy harvesting ability of different HNGs, the utiliza-
system and the water confinement strategy. Under solar radiation, the tion of solar energy for water heating and evaporation depended
solar absorbers in the molecular meshes of the floating generator are strongly on the Qs and V values. Within 500 s, the surface of HNG1,
heated, facilitating the evaporation of water confined in the polymeric HNG2 and HNG3 could be heated to an equilibrium temperature
network (1). The water confined in the molecular mesh has a reduced (Fig. 3d). The different linear dependence between the surface
evaporation enthalpy. The evaporated water can be rapidly recovered via temperature and time represented a different energy consumption
branched water diffusion (2) and pumping (3) based on micron channels method. The sharply increasing region indicated that the heat was
and internal gaps, respectively. mainly utilized to raise the temperature, while the plateau region
indicated the tremendous energy consumption of evaporation
(known to be associated with the cooling effect). As the evaporation
As polymeric hydrogels are viscoelastic materials, the additive of water is accelerated with increasing temperature, the inflection
penetrating the skeleton alters energy storage and energy dissipa- point showed an energy balance value of water heating and evapo-
tion, which can be represented by the storage modulus (G′​) and ration. The low equilibrium temperature indicated the optimum Qs
the loss modulus (G″​), respectively32. The G′​and G″​values of a and V values for solar vapour generation. Accordingly, as the Qs and
HNG and a pure PVA gel are shown in Fig. 2e. The dynamic fre- V values increased further, the surface temperature of HNG4 rose
quency sweep experiments of both gels show a wide linear visco- gradually and continued increasing after exposure to solar radia-
elastic region. The G′​value is higher than the G″​value in each tion for 2,500 s; therefore, the converted energy was used to heat the
case, confirming their crosslinked polymeric skeleton. The lower water instead of powering evaporation. The corresponding HNG1
G′​value of the HNG indicates that it has fewer crosslinking points to HNG4 xerogels could be identically heated to ~90 °C in a short
than the PVA gel, while the lower G″​value of the HNG reveals period of time (Supplementary Fig. 14).
an efficient restriction of the slippage between polymer chains33
(discussed in detail in Supplementary Fig. 4). These results dem- Solar vapour generation under 1 sun
onstrated that the PPy chains were inserted into the molecular Under constant solar illumination (1  kW  m−2), the overall
meshes of the PVA network, indicating that the solar absorber was mass change with and without HNGs, which represented the
selectively introduced into the HNG-confined water (that is, only amount of evaporated water, was recorded once the tempera-
in the molecular meshes). ture reached steady state (pre-heating for ~500 s, see above).
The chemical composition of the HNGs was analysed by It is clear that vapour generation with the HNGs is more effi-
Fourier transform infrared (FTIR) spectroscopy. In the PVA spec- cient than pure water without HNG under 1  sun radiation
trum (red curve in Fig. 2f), the peaks located at 1,093 cm−1 can (Fig. 4a). HNG3 presented an extremely high rate of 3.2 kg m−2 h−1,
be attributed to C–O stretching, which is a characteristic peak of which was 1.3, 2.5 and 8 times that for HNG2, HNG1 and pure
PVA34. The PPy spectrum (blue curve in Fig. 2f) shows absorp- water (shown by red squares in Fig. 4b), respectively. However,
tion signals at 1,552 cm−1 and 1,045 cm−1, corresponding to the HNG4, which presented an even larger water content in the molec-
in-ring stretching of C=​C bonds in the pyrrole rings and the in- ular meshes, showed a low rate of 0.7 kg m−2 h−1. These results
plane deformation of N–H bonds, respectively35. All the charac- revealed that HNG3 has an optimized water content. Additionally,
teristic peaks of PVA and PPy were found in the FTIR spectra of the control experiments conducted under concentrated solar illu-
HNG (black curve in Fig. 2f) without shifts, thus confirm- mination also demonstrated that HNG3 presented the best per-
ing the presence of PPy in the PVA matrix. These results were formance (Supplementary Fig. 15). Note that all the experimental
in good agreement with the thermogravimetric analysis data data of solar vapour generation are calibrated to dark-condition
(Supplementary Fig. 5). evaporation data (see Supplementary Methods for details), and the

490 Nature Nanotechnology | VOL 13 | JUNE 2018 | 489–495 | www.nature.com/naturenanotechnology

© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Nature Nanotechnology Articles
a b e
106 106
PVA gel
10 5
HNG 105
10 4
104

G′ (Pa)

G′′ (Pa)
103 103
10 2
102
Internal 10 1
101
gap 100 100
1 cm 100 μm 1 10 100
Frequency (rad s–1)
c d f

Transmittance (a.u.)
HNG

PPy
Micron
channel PVA

4,000 3,000 2,000 1,000


5 μm 1 μm
Wavenumber (cm–1)

Fig. 2 | Characterization of the HNGs. a, Photograph of an as-prepared HNG. b–d, SEM images of HNGs at different magnifications showing internal
gaps (b), micron channels (c) and the wrinkled internal surface (d); the wrinkled internal surface indicates shrinkage of the polymeric skeleton during
dehydration. e, Dynamic mechanical analysis of the storage modulus (G′​) and loss modulus (G″​). The results demonstrate that PPy penetrates the
molecular meshes among the PVA network. f, FTIR spectra of a HNG, PVA and PPy showing the chemical composition of the HNGs.

a b
42
8 0.12
Half swollen time (min)

39
Saturated water

Water transport
content (g g–1)

rate (g min–1)
6
36 0.08

4
33
0.04
2 30

27 0.00
HNG1 HNG2 HNG3 HNG4 HNG1 HNG2 HNG3 HNG4
c d 50
100 1 sun

HNG1
Temperature (°C)
Absorption (%)

HNG2 40
HNG3 HNG1
90 HNG4 HNG2
AM 1.5 G HNG3
30
HNG4

80 20
500 1,000 1,500 2,000 2,500 0 500 1,000 1,500 2,000 2,500
Wavelength (nm) Time (s)

Fig. 3 | Tunable water transport and solar absorption of the HNGs. a, Saturated water content in HNGs per gram of corresponding xerogel samples. Each
error bar represents the deviation from at least 15 data points. b, Swollen behaviour from the half-saturated to the saturated state and the calculated water
transport rate showing the tunable water transport ability of the HNGs. HNG1, HNG2, HNG3 and HNG4 represent gels with water/PVA weight ratios of
1:0.2, 1:0.15, 1:0.1 and 1:0.075, respectively. c, UV-vis NIR spectra of HNG sheets ~1 mm thick. The normalized spectral solar irradiance density of air mass
1.5 global (AM 1.5 G) tilt solar spectrum is shown by the black broken line. d, The surface temperature rise of hydrogel samples relative to heating time
under 1 sun radiation. The different dependence between temperature and time of the HNGs revealed their different energy utilization abilities.

blank control experiment based on pure PVA hydrogels further where ṁ is the mass flux, hV is the vaporization enthalpy of the water
demonstrated the significance of the hierarchical structure of the in HNGs, P0 is the solar irradiation power of 1 sun (1 kW m−2), and
HNGs (Supplementary Fig. 16). The corresponding energy effi- Copt refers to the optical concentration on the absorber surface. We
ciency (η) for solar to vapour generation can then be calculated note that the vaporization enthalpy of water confined in the HNG
using the following formula16: molecular mesh is smaller than that of bulk water. Both evaporation
measurements and differential scanning calorimetry experiments
η = mh
̇ V ∕CoptP0 (1) demonstrated a reduced vaporization enthalpy compared with that

Nature Nanotechnology | VOL 13 | JUNE 2018 | 489–495 | www.nature.com/naturenanotechnology 491


© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Articles Nature Nanotechnology

a b 100 5
c
Ref.14
0 Ref.15
3
Mass change (kg m–2)

80 4

Energy efficiency (%)


Ref.16

Evaporation rate
Ref.17

Evaporation rate
–1 Ref.19

(kg m–2 h–1)

(kg m–2 h–1)


60 3
Ref.20
Water 2 Ref.21
–2 HNG1 40 2 Ref.23
HNG2 Ref.24
Current study
–3 HNG3 20 1
1
HNG4
0 0
0 1,000 2,000 3,000 HNG1 HNG2 HNG3 HNG4 40 60 80 100
Time (s) Energy efficiency (%)

d ΔT pattern e f g
0 –20 Model Temperature Water transport
70
37.6 41.3
60 + +
400 315 400 0.9
21.4 21.4
Temperature (°C)

+ +
50 21.3
+

Z (μm)

Z (μm)
21.3
40 + 200 200
HNG
30 Water Internal gap 294 0.01
Micron channel (K) (μm s–1)
20
0 0
0 2,000 4,000 Polymeric network 0 200 400 0 200 400
Time (s) X (μm) X (μm)

Fig. 4 | Vapour generation under 1 sun. a,b, The mass loss of water (a) and solar vapour generation energy efficiency (b) with corresponding evaporation
rates of different HNGs under 1 sun (1 kW m−2), with pure water as the control. Each error bar represents the deviation from at least 15 data points.
c, Comparison of HNG vapour generation performance and previous reports under 1 sun. d, Plot showing the temperature of bulk water and HNG3 under
1 sun relative to irradiation time. Insets: infrared images showing the temperature distribution after irradiation times of 0, 500 and 3,600 s. e–g, Theoretical
simulation of internal gaps, micron channels and polymeric network (corresponding to molecular meshes) modelled by red oval areas, yellow narrow bars
and blue background, respectively (e), temperature distribution (f) and water transport in a HNG (g).

of pure bulk water (Supplementary Figs. 17 and 18). To explain this Mechanism for high-efficiency solar water evaporation
reduction in vaporization enthalpy, we turned to the water cluster Figure 4d compares the surface temperature of HNG3 and the bulk
theory37,38. Water can be evaporated as either a single molecule or in water relative to irradiation time. With 1 h of solar radiation, the free-
small clusters consisting of a few to tens of molecules. When water floating HNG showed a rapid temperature increase, while the bulk
molecules are confined in the molecular mesh (Supplementary water remained at a constant temperature. The temperature of the
Fig. 19), they are more likely to escape the polymer network as HNG increased from 21.3 to 37.6 °C within 500 s, and then further
small clusters rather than individual molecules (Supplementary increased to 41.3 °C within 1 h. In contrast, the temperature of bulk
Fig. 20). As such, the water is evaporated to a state with a lower water remained at ~21.3 °C after 1 h. The insets in Fig. 4d are infrared
enthalpy change than conventional latent heat. A thermodynamic photographs of cuvettes after irradiation times of 0 s, 500 s and 1 h.
analysis was also performed to demonstrate the energy balance and These results proved that the converted energy was localized in
vaporization enthalpy reduction (Supplementary Figs. 21 and 22). the HNG. In addition, to reveal the temperature distribution in the
It should be noted that for concentrated solar irradiation, further HNG, COMSOL Multiphysics software was used to simulate a fully
molecular tailoring of HNG-based materials for balanced water hydrated HNG undergoing solar radiation. The modelled structure
transport and solar adsorption is needed to achieve optimal energy is shown in Fig. 4e, where the internal gaps, micron channels and
efficiency at various solar fluxes, despite the reduced vaporization molecular meshes were simulated by broad ovals (red, with widths
enthalpy (Supplementary Fig. 23). of 150 μ​m), narrow bars (yellow, with widths of 0.2 μ​m) and quadrate
The energy efficiency of solar vapour generation is shown in background (blue) in a two-dimensional (2D) mapping of a cross-
Fig. 4b (blue points) with error bars representing the standard section, respectively. Then, a simple heat-transfer model was used to
deviation. The energy efficiency of HNG3 reached up to 94% describe the temperature distribution in the HNG39 (Supplementary
under 1 sun. In other reported state-of-the-art systems, such a Fig. 25). The simulation results of the steady-state temperature dis-
high efficiency (greater than 90%) can only be achieved via highly tribution predicted that the maximum temperature in the HNG
concentrated solar radiation (with a flux density higher than could be 315 K at the surface of polymeric network (Fig. 4f). This
4 kW m−2) (Supplementary Fig. 24). The relatively low efficiency predicted value was close to our experimental results, therefore con-
produced by HNG1 and HNG4 further proved the importance of firming the heat confinement occurring in the HNG. Additionally,
water management, which is in agreement with the results of the a biphasic mixture model was used to simulate the water trans-
heating behavior analysis shown in Fig. 3. The energy efficiency port in the HNG40. The 2D mapping of water velocity distribution
and water evaporating rate under 1 sun were key parameters for depicted a fast water transport through the internal gaps against the
evaluating the possibility of materials to generate vapour under surface water loss (Fig. 4g and Supplementary Fig. 26). Along with
ambient conditions. Thus, the performance of the developed HNG our experimental data, the surface water loss was estimated to be
was compared with recently reported materials (Fig. 4c); the devel- 0.89 μ​m s−1. The calculated water transport rate from the bottom
oped HNG achieved record high evaporation rates with a highly (Z =​ 0, corresponding the interface of bulk water and the HNG) to
efficient energy utilization14–17,19–21,23,24. the surface was 0.9 μ​m s−1; this transport was mainly supported by

492 Nature Nanotechnology | VOL 13 | JUNE 2018 | 489–495 | www.nature.com/naturenanotechnology

© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Nature Nanotechnology Articles
a 104 b 4
3 Before desalination
10

Evaporation rate (kg m–2 h–1)


After desalination
102 3

101

Salinity (‰)
WHO
100 2 Baltic sea
10–1 EPA World sea
Dead sea
10–2 1

10–3

10–4 0
Baltic sea World sea Dead sea 0 10 20 30 40 50
Time (h)
c 106 d 4
Before desalination
5
10 After desalination

Evaporation rate (kg m–2 h–1)


3
Concentration (mg l–1)

104

103 Day 1 Day 28

Mass change (kg m )

Mass change (kg m )


0 0

–2

–2
2
–1 –1
102
–2 –2
101
1
–3 –3
100
0 3,600 0 3,600
Time (s) Time (s)
10–1 0
Na+ Mg2+ K+ Ca2+ 0 10 20 30
Time (days)

Fig. 5 | Solar desalination performance and durability of the HNGs. a, The salinities of three artificial seawater samples before and after desalination using
HNG3. The broken lines refer to the World Health Organization (WHO) and US Environmental Protection Agency (EPA) salinity standards for drinkable
water. b, Solar endurance test results of a HNG continuously exposed to 1 sun for 50 h. c, Measured concentrations of four primary ions in an actual seawater
sample before and after desalination. d, Water evaporation rates of HNGs immersed in seawater for a long time; the results of this solar desalination test
demonstrated the salt durability of the HNGs. Insets: typical solar vapour generation behaviour on exposure to seawater for 1 day and 28 days.

internal gaps, while the uniform water diffusion near the evaporat- A solar water purification prototype based on the optimized
ing surface was assisted by the micron channels. HNG3 was placed on the roof of the Engineering Teaching Center at
the University of Texas Austin campus. Here, a HNG3 sample with
Laboratory and outdoor solar desalination a diameter of ~20 cm and thickness of 2 cm (Fig. 6a) was floated in
To demonstrate the solar desalination capability of the optimized a container of brine, which was set on the upper layer of the proto-
HNG3, three brine samples with representative simulated salinities type (Fig. 6b). The evaporated water condensed on the transparent
(g of dissolved salt per kg seawater (‰))—Baltic sea (lowest salin- condenser (Fig. 6b), and the water condensate flowed to the bottom
ity, 8‰), World sea (average salinity, 35‰) and Dead sea (highest of the prototype (Fig. 6b). To demonstrate the possibility of con-
salinity, 100‰)41—were used and carefully tracked via a conduc- tinuous water purification, the water receiver of the prototype was
tivity test (Supplementary Fig. 27). After desalination, the salini- connected (Fig. 6b,c) to the brine tank and to a purified water flask
ties of the artificial brine samples were all significantly decreased (Fig. 6c); this set-up simulated the functional parts of practical solar
(by approximately four orders of magnitude; Fig. 5a), and were water purification equipment or plants. The experiment was carried
approximately two orders of magnitude below the drinking water out from 08:00 to 20:00 under natural sunlight with an average solar
standards defined by the World Health Organization (1‰) and heat flux of ~0.7 kw m−2 (Fig. 6d), and the solar irradiation, inter-
the US Environmental Protection Agency (0.5‰)42. Solar dura- nal humidity, air temperature, salt-water temperature and HNG3
bility is another important aspect of solar desalination. The water surface temperature were carefully traced to quantitatively evalu-
evaporation rate, which is independent of salinity, was maintained ate the utilization efficiency of natural sunlight. The ample sunlight
after continuous exposure for 50 h (Fig. 5b). Furthermore, a real enabled an average water purification rate of ~1.6 l m−2 h−1 (Fig. 6e,
seawater sample (from the Gulf of Mexico) was used for desalina- blue curve). As shown in Fig. 6e, the surface temperature of the
tion via the HNG. It was found that the concentrations of all four HNG sample increased to ~55 °C, indicating a weakened evapora-
primary ions (Na+, Mg2+, K+ and Ca2+) originally present in the tion cooling effect, hence reduced evaporation, compared with that
seawater were significantly reduced (Fig. 5c), and were below the in laboratory test condition. Such a phenomenon can be attributed
values typically obtained through membrane-based (10–500 mg l−1) to the saturated internal humidity of the closed system (Fig. 6e),
and thermal distillation-based (1–50 mg l−1) seawater desalination which slightly restrained water vaporization (see Supplementary
techniques42,43. As the solar evaporator needs to be in contact with Fig. 28 for details). The brine maintained a relatively low temperature
seawater for a long period of time during desalination, the salt dura- of ~30 °C (Fig. 6e), even when the internal air and condenser were
bility of the HNGs is important for practical applications. As shown heated to ~45 and ~40 °C, respectively, indicating an effective heat-
in Fig. 5d, the HNG presented a stable evaporation rate under 1 sun ing localization effect on the surface of the HNG sheet. Additionally,
irradiation on exposure to the seawater sample for up to 28 days. the data collected during non-ideal conditions (Supplementary
Such results indicate that the HNG is potentially reliable for practi- Fig. 29) indicated that the HNG enabled solar water purification
cal long-term solar desalination. during periods of low and/or varying solar irradiation, revealing

Nature Nanotechnology | VOL 13 | JUNE 2018 | 489–495 | www.nature.com/naturenanotechnology 493


© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Articles Nature Nanotechnology

a b c

Transparent condenser

Inlet (2)
HNG vapour pipe
generator
(1)

Brine

(3)
Purified Outlet pipe
water
(1) Solar vapour system
10 cm
(2) Brine tank
(3) Purified water flask

d 1.0 e 60
600
HNG surface 100
500
Irradiation (kw m–2)

Purified water (ml)

0.8 50

Temperature (°C)
Internal air
400
RH
0.6 300 40
Condenser (%)
200
0.4 100 30
Brine
40
0
0.2 20
0 1 2 3 4 5 6 7 8 9 10 11 12
0

0
:0

:0

:0

:0

:0

:0

:0
08

10

12

14

16

18

20

Time (h)
Time
f
Electrodes
Ohmic value

Water
sample
Seawater Purified water Domestic water

Fig. 6 | Outdoor solar water purification using HNGs in natural sunlight. a, Photograph of a large-scale HNG sheet. b,c Schematic of a water receiver
based on a HNG (b) for a prototype solar water purification system simulating the practical water purification equipment (c). The HNG floats on brine in a
container that is located in a holder with meshes. The vapour is condensed via the transparent condenser and flows to the bottom of the prototype, where
the purified water is stored. Brine is supplied and fresh water collected at optimum rates through the inlet pipe and outlet pipe, respectively. A sustained
water purification system was achieved under natural sunlight. d, Solar radiation recorded over time on a sunny day from 08:00 to 20:00. e, The amount
of purified water during 12 h of outdoor solar desalination (solid blue line) with careful tracing of the internal relative humidity (RH; background colour
map), and the temperature of the HNG surface (red broken line), internal air (green broken line), condenser (violet broken line) and brine (orange broken
line). f, Evaluation of water purity using a multimeter with a constant distance between electrodes. The results demonstrated a high purity of purified
water, similar to that of domestic water.

its significant potential for practical applications. The water qual- the hierarchical water pathways in the HNGs, providing expedited
ity can be directly represented by the ohmic value obtained from water absorption from bulk water and tunable water content in the
a resistance test using a multimeter at constant distance between polymeric network. The solar energy conversion is induced by the
electrodes (Fig. 6g). The real seawater, purified water and domestic light-absorbing polymer penetrating the molecular meshes of the
water (from an urban water-supply system of Austin, Texas, United HNG, which provides an efficient in situ energy utilization system
States) showed resistance values of 0.08, 1.4 and 1 MΩ​, respectively, for powering water evaporation with significantly reduced vapor-
indicating effective purification of natural seawater. ization enthalpy. These unique properties of the HNGs enabled a
solar vapour generation rate of 3.2 kg m−2 h−1, with an energy effi-
Conclusions ciency of up to ~94%. As demonstrated by a prototype solar water
As detailed above, highly efficient solar vapour generation was purification system, the HNG was able to produce fresh water with
achieved due to effective energy confinement based on tailored a daily yield of 18–23 l m−2 under natural sunlight. Apart from the
water transport in HNGs. The water distribution can be regulated by demonstrated solar water purification system, the HNG-based solar

494 Nature Nanotechnology | VOL 13 | JUNE 2018 | 489–495 | www.nature.com/naturenanotechnology

© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Nature Nanotechnology Articles
energy harvesting system may be used for many other applications, 27. Phelan, P., Taylor, R., Adrian, R., Prasher, R. & Otanicar, T. in Nanoparticle
including environmental cooling, water and moisture management, Heat Transfer Fluid Flow (eds Minkowycz, W. J., Sparrow, E. M. & Abraham,
J.) 123–142 (CRC, Boca Raton, FL, 2012).
and pollution abatement. 28. Zhao, F., Shi, Y., Pan, L. & Yu, G. Multifunctional nanostructured conductive
polymer gels: synthesis, properties, and applications. Acc. Chem. Res. 50,
Methods 1734–1743 (2017).
Methods, including statements of data availability and any asso- 29. Bellich, B., Borgogna, M., Cok, M. & Cesàro, A.Water evaporation from gel
beads. J. Therm. Anal. Calorim. 103, 81–88 (2011).
ciated accession codes and references, are available at https://doi.
30. Ma, C., Shi, Y., Pena, D. A., Peng, L. & Yu, G. Thermally responsive
org/10.1038/s41565-018-0097-z. hydrogel blends: a general drug carrier model for controlled drug release.
Angew. Chem. Int. Ed. 127, 7484–7488 (2015).
Received: 22 May 2017; Accepted: 19 February 2018; 31. Shi, Y., Ma, C., Peng, L. & Yu, G. Conductive “smart” hybrid hydrogels with
Published online: 2 April 2018 PNIPAM and nanostructured conductive polymers. Adv. Funct. Mater. 25,
1219–1225 (2015).
32. Jin, L. & Bai, R. Mechanisms of lead adsorption on chitosan/PVA hydrogel
References beads. Langmuir 18, 9765–9770 (2002).
1. Lewis, N. S. Research opportunities to advance solar energy utilization.
33. Wang, Y. et al. Dopant-enabled supramolecular approach for controlled
Science 351, aad1920 (2016).
synthesis of nanostructured conductive polymer hydrogels. Nano Lett. 15,
2. Chen, J. et al. Micro-cable structured textile for simultaneously harvesting
7736–7741 (2015).
solar and mechanical energy. Nat. Energy 1, 16138 (2016).
34. Mansur, H. S., Oréfice, R. L. & Mansur, A. A. Characterization of poly
3. Wallace, G. G. et al. Nanoelectrodes: energy conversion and storage.
(vinyl alcohol)/poly (ethylene glycol) hydrogels and PVA-derived hybrids by
Mater. Today 12, 20–27 (2009).
small-angle X-ray scattering and FTIR spectroscopy. Polymer 45,
4. Shannon, M. A. et al. Science and technology for water purification in the
7193–7202 (2004).
coming decades. Nature 452, 301310 (2008).
35. Bairi, P., Roy, B., Routh, P., Sen, K. & Nandi, A. K. Self-sustaining, fluorescent
5. Narayan, G. P. et al. The potential of solar-driven humidification–
and semi-conducting co-assembled organogel of Fmoc protected
dehumidification desalination for small-scale decentralized water production.
phenylalanine with aromatic amines. Soft Matter 8, 7436–7445 (2012).
Renew. Sustain. Energy Rev. 14, 11871201 (2010).
36. Sun, G., Li, Z., Liang, R., Weng, L. T. & Zhang, L.Super stretchable hydrogel
6. Williams, A. Solar powered water desalination heats up in Chile.
achieved by non-aggregated spherulites with diameters <​  5nm. Nat. Commun.
Water Wastewat. Int. 28, 24–28 (2013).
7, 12095 (2016).
7. Elimelech, M. & Phillip, W. A. The future of seawater desalination: energy,
37. Miyazaki, M. et al. Infrared spectroscopic evidence for protonated water
technology, and the environment. Science 333, 712717 (2011).
clusters forming nanoscale cages. Science 304, 1129–1137 (2004).
8. Li, C., Goswami, Y. & Stefanakos, E. Solar assisted sea water desalination: a
38. Fujii, A. & Kenta, M. Infrared spectroscopic studies on hydrogen-bonded
review. Renew. Sustain. Energy Rev. 19, 136163 (2013).
water networks in gas phase clusters. Int. Rev. Phys. Chem. 32,
9. Jean, J. et al. Pathways for solar photovoltaics. Energy Environ. Sci. 8,
266–307 (2013).
1200–1219 (2015).
39. Liu, Y. et al. A bioinspired, reusable, paper-based system for high-
10. Romano, M. S. et al. Carbon nanotube-reduced graphene oxide composites
performance large-scale evaporation. Adv. Mater. 27, 2768–2774 (2015).
for thermal energy harvesting applications. Adv. Mater. 25, 6602–6606 (2013).
40. Birgersson, E., Li, H. & Wu, S. Transient analysis of temperature-sensitive
11. Fraunhofer Institute for Solar Energy Systems. Photovoltaics Report (2014);
neutral hydrogels. J. Mech. Phys. Solids 56, 444–466 (2008).
https://www.ise.fraunhofer.de/content/dam/ise/de/documents/publications/
41. World Ocean Database 2013 (National Oceanic and Atmospheric
studies/Photovoltaics-Report.pdf
Administration, accessed September 2013); https://www.nodc.noaa.gov/OC5/
12. IPCC. Climate Change 2014: Synthesis Report (eds Core Writing Team,
WOD13/
Pachauri, R. K. & Meyer L. A.) (IPCC, 2015); https://www.ipcc.ch/pdf/
42. World Health Organization (WHO). Safe Drinking-Water from Desalination
assessment-report/ar5/syr/SYR_AR5_FINAL_full_wcover.pdf
(WHO, 2011); http://apps.who.int/iris/bitstream/10665/70621/1/WHO_HSE_
13. Wang, J. et al. High-performance photothermal conversion of narrow-
WSH_11.03_eng.pdf
bandgap Ti2O3 nanoparticles. Adv. Mater. 29, 1603730 (2017).
43. Surwade, S. P. et al. Water desalination using nanoporous single-layer
14. Hu, X. et al. Tailoring graphene oxide-based aerogels for efficient solar steam
graphene. Nat. Nanotech. 10, 459–464 (2015).
generation under one sun. Adv. Mater. 29, 1604031 (2017).
15. Ito, Y. et al. Multifunctional porous graphene for high-efficiency steam
generation by heat localization. Adv. Mater. 27, 4302–4307 (2015). Acknowledgements
16. Zhang, L., Tang, B., Wu, J., Li, R. & Wang, P. Hydrophobic light-to-heat G.Y. acknowledges financial support from a Sloan Research Fellowship, a Camille
conversion membranes with self-healing ability for interfacial solar heating. Dreyfus Teacher-Scholar Award, and a National Science Foundation award (NSF-
Adv. Mater. 27, 4889–4894 (2015). CMMI-1537894). Molecular dynamics simulations were performed using a Summit
17. Liu, C. et al. High-performance large-scale solar steam generation with supercomputer supported by the NSF (NSF-ACI-1532235).
nanolayers of reusable biomimetic nanoparticles. Adv. Sustain. Syst. 1,
1600013 (2017).
18. Zhou, L. et al. 3D self-assembly of aluminium nanoparticles for plasmon- Author contributions
enhanced solar desalination. Nat. Photon. 10, 393–398 (2016). G.Y. supervised the entire project. G.Y., F.Z., X.Z. and L.Q. conceived the idea
19. Bae, K. et al. Flexible thin-film black gold membranes with ultrabroadband and co-wrote the manuscript. F.Z. and X.Z. performed materials fabrication and
plasmonic nanofocusing for efficient solar vapour generation. Nat. Commun. characterization and carried out data analyses. F.Z. and Y.S. performed the numerical
6, 10103 (2015). simulations. X.Q., X.Z. and R.Y. performed the molecular dynamics simulations and
20. Zhou, L. et al. Self-assembly of highly efficient, broadband plasmonic differential scanning calorimetry measurements on the gel samples. M.A. and S.M.
absorbers for solar steam generation. Sci. Adv. 2, e1501227 (2016). assisted in experimental work. R.Y. and L.Q. assisted in the design of experiments and
21. Ni, G. et al. Steam generation under one sun enabled by a floating structure interpretation of results. All the authors discussed the results and commented on the
with thermal concentration. Nat. Energy 1, 16126 (2016). manuscript.
22. Liu, Z. et al. Extremely cost-effective and efficient solar vapor generation
under nonconcentrated illumination using thermally isolated black paper. Competing interests
Glob. Chall. 1, 1600003 (2017). The authors declare no competing interests.
23. Hadi, G. et al. Solar steam generation by heat localization. Nat. Commun. 5,
5449 (2014).
24. Li, X. et al. Graphene oxide-based efficient and scalable solar desalination Additional information
under one sun with a confined 2D water path. Proc. Natl Acad. Sci. USA 113, Supplementary information is available for this paper at https://doi.org/10.1038/
13953–13958 (2016). s41565-018-0097-z.
25. Zielinski, M. S. et al. Hollow mesoporous plasmonic nanoshells for enhanced Reprints and permissions information is available at www.nature.com/reprints.
solar vapor generation. Nano Lett. 16, 2159–2167 (2016).
26. Neumann, O. et al. Compact solar autoclave based on steam generation using Correspondence and requests for materials should be addressed to R.Y. or L.Q. or G.Y.
broadband light-harvesting nanoparticles. Proc. Natl Acad. Sci. USA 110, Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in
11677–11681 (2013). published maps and institutional affiliations.

Nature Nanotechnology | VOL 13 | JUNE 2018 | 489–495 | www.nature.com/naturenanotechnology 495


© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Articles Nature Nanotechnology

Methods measured for 1 h. The dark-condition evaporation rate was subtracted from the
Preparation of the HNGs. For preparation of the HNGs, PVA (Mw =​  15,000), solar-illuminated evaporation rate. All evaporation rates were measured after
glutaladehyde and PPy of various corresponding molar mass were dissolved in stabilization under 1 sun for 10 min.
deionized water. An appropriate amount of HCl was subsequently added. The
gelation was carried out for 2 h. The obtained PVA gel was immersed in deionized Characterization. The morphology and microstructure of the samples were
water overnight to obtain pure HNGs. The purified HNGs were frozen by liquid observed by SEM (Hitachi, S5500) operating at 5 kV. Before observation, the
nitrogen and then thawed in deionized water at 30 °C. The freeze–thaw process was HNGs were freeze dried for 24 h. The FTIR spectra of the HNGs were recorded
repeated ten times. Finally, the obtained HNG samples were freeze dried. using a FTIR spectrometer (Thermo Mattson, Infinity Gold FTIR) equipped with
a liquid nitrogen-cooled narrowband mercury cadmium telluride detector, using
an attenuated total reflection cell equipped with a Ge crystal. To understand the
Solar vapour generation experiments. The water evaporation performance mechanical properties of the HNGs, rheological experiments were performed (AR
experiments were conducted in the laboratory using a solar simulator (AbetTech, 2000EX, TA Instrument) using a parallel plate on a peltier plate in the frequency
M-LS Rev B) outputting a simulated solar flux of 1,000 W m–2 (1 sun). The solar sweep mode. Absorption spectra, transparency and reflectance were recorded
flux was measured using a thermopile (Newport, 818SL) connected to a power using a UV–vis NIR spectrometer (Cary 5000) with an integrating sphere unit
meter (Newport, 1916-R). Because the solar flux varies across the beamspot, and and automated reflectance measurement unit. The measurements were corrected
the thermopile detector is ~1 cm2 in size, the HNG was cut into small pieces with by baseline/blank correction through the control software of UV–vis NIR
a surface area of 1 cm2 to ensure an accurate power input. A HNG chip (~0.5 cm spectrometer. The concentration of ions was tracked by inductively coupled plasma
thick) was transferred and floated on pure water (or brine for the desalination mass spectrometry (Agilent 7500ce) with dilutions in 2% HNO3 to make the
tests) in a glass cuvette that was placed in the beamspot with a solar flux of 1 sun. loaded ion concentration below 10 ppm.
The mass of water loss was measured using a laboratory balance with 0.1 μ​g
resolution and calibrated to weights heavier than the total weight of the set- Data availability. The data that support the findings of this study are available
up. Before illuminating the set-up, the evaporation rate in dark conditions was from the corresponding authors upon reasonable request.

Nature Nanotechnology | www.nature.com/naturenanotechnology

© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

You might also like