of 9
Editorial Note:
This manuscript has been previously reviewed at another journal that is not
operating a transparent peer review scheme. This document only contains reviewer comments and
rebuttal letters for versions considered at Nature Communications.
REVIEWER COMMENTS
Reviewer #1 (Remarks to the Author):
In this work, the authors provided an interesting and novel way to improve the working efficiency of
interfacial solar steamers. They designed and fabricated hydrogel membranes with cone-based
surface microstructures. The resulting materials show the unique combination of solar steam
generation and fog harvesting abilities and achieved clean water collection around the clock. The
experimental measurements presented in the manuscript demonstrated high performance of these
microstructured hydrogel membranes for both functions. The mechanism studies in this work have
been significantly improved after revisions. The concerns from previous referees have been well
addressed and the working mechanisms are clearly explained in the current version. I believe that
the effects of surface microstructures on surface temperature and vapor distributions revealed in
this study are very useful and of great interest to other researchers in the field. Moreover, I am
excited to see that the delicate hydrogel microstructures are realized through scalable 3D printers.
This fact, together with the outdoor test results presented in this work, indicates the potential of this
technology for practical applications. This revised manuscript can be accepted for publication in
Nature Communications. A minor suggestion is for the authors to include in the introduction some
closely relevant reviews on hydrogels for water harvesting technologies to reflect the state of the
art, such as Chem. Rev. 120, 7642 (2020); Nature Rev. Mater. 5, 388 (2020); ACS Mater. Lett. 2, 671
(2020).
Reviewer #2 (Remarks to the Author):
The authors have addressed part of the questions raised by reviewers. However, the all-day water
collection rate needs to be further calculated and justified carefully. I am not convinced of the
details of experiments. The reasons are as follows:
(1) In their reply, the collection rate was calculated based on the area of gel samples. However, I
notice that they arrange several small samples rather than a single large one and the small samples
are not tightly packed, as shown in Fig. 5a and Fig. S31. If the spacing of the arrangement is essential,
it is more reasonable to calculate the collection rate based on the overall area. Or the single larger
samples should be fabricated for performance testing, which will make the data more credible.
(2) They reported that the water vaporization enthalpy is significantly reduced to 1000 kJ kg-1, which
is even 300 kJ kg-1 lower than the enthalpy reported by Yu, G. H., et al. [Nat. Nanotech. 13, 489-495
(2018)] How to explain the further decrease of enthalpy? Whether it’s the micro-trees morphology
or the different pore size or something else? And the differential scanning calorimetric (DSC)
measurement needs to be supplemented to prove the reduced evaporation enthalpy.
(3) As shown in Fig. 5e, the condensation device is equipped with a transparent plastic cover and
without forced heat transfer. This device results in high optical loss caused by condensates and
limited heat transfer, which have adverse effects on condensation [Energy & Environmental Science
11, 1510-1519]. Especially, in their design, the middle part of covering is horizontal, which impede
the collection of water and further result in higher optical loss.
(4) I think the first question of reviewer #4 has not been addressed. The author thinks that the fog
collection processes only occur on the gel surface and will not wash out the contaminants in the gel.
However, because of high concentrations of pollutants in the absorber, diffusion of pollutants into
fog water will occur. The inductively coupled plasma spectroscopy of fog water should be
supplemented.
Point
-
to
-
point Responses to Referees’ comments
Manuscript Title: All
-
day Fresh Water Harvesting by Microstructured Hydrogel Membranes
Manuscript ID:
NCOMMS
-
21
-
04497
-
T
Reviewer #1 (Remarks to the Author
):
In this work, the authors provided an interesting
and novel way to improve the working efficiency
of interfacial solar steamers. They designed and fabricated hydrogel membranes with cone
-
based
surface microstructures. The resulting materials show the unique combination of solar steam
generation and fog ha
rvesting abilities and achieved clean water collection around the clock. The
experimental measurements presented in the manuscript demonstrated high performance of
these microstructured hydrogel membranes for both functions. The mechanism studies in this
w
ork have been significantly improved after revisions. The concerns from previous referees have
been well addressed and the working mechanisms are clearly explained in the current version. I
believe that the effects of surface microstructures on surface tem
perature and vapor
distributions revealed in this study are very useful and of great interest to other researchers in
the field. Moreover, I am excited to see that the delicate
hydrogel microstructures are realized
through scalable 3D printers. This fact,
together with the outdoor test results presented in this
work, indicates the potential of this technology for practical applications. This revised manuscript
can be accepted for publication in Nature Communications. A minor suggestion is for the authors
to
include in the introduction some closely relevant reviews on hydrogels for water harvesting
technologies to reflect the state of the art, such as Chem. Rev. 120, 7642 (2020); Nature Rev.
Mater. 5, 388 (2020); ACS Mater. Lett. 2, 671 (2020).
Response:
We thank the referee for the positive comments. We have
cited
the
se important
references in
the Introduction section.
Reviewer #2 (Remarks to the Author):
The authors have addressed part of the questions raised by reviewers. However, the all
-
day water
co
llection rate needs to be further calculated and justified carefully. I am not convinced of the
details of experiments. The reasons are as follows:
(1) In their reply, the collection rate was calculated based on the area of gel samples. However, I
notice t
hat they arrange several small samples rather than a single large one and the small
samples are not tightly packed, as shown in Fig. 5a and Fig. S31. If the spacing of the arrangement
is essential, it is more reasonable to calculate the collection rate bas
ed on the overall area. Or
the single larger samples should be fabricated for performance testing, which will make the data
more credible.
Response:
We thank the referee for the
constructive
comments.
Due
to the limitation of
the
3D printer, we
are unable to fabricate
a single
la
rge gel sample
(more than 50 cm
2
)
to fit in the rooftop tests.
However, we had studied the fog collection and solar steam generation performance of gel
membranes with different
areas
(Fig. S11 and Fig. S19)
. The results show that the
area
of
membrane has li
ttle effects on either function, which indicates that the performance of gel
m
embranes will not be affected whether
they are tightly packed or not.
In the rooftop tests, w
e used a passive polyurethane foa
m, which has a water evaporation
rate close to free water and a low fog collection rate, to
tightly
support
and fix
our gel samples.
We also
conduct
ed
a
control experiment during rooftop tests
to
subtract
the contribution from
th
e
PU foam. The processed re
sults reflect the
working performance
of microstructured gel
membranes
in the rooftop tests
, which is the focus of our study.
This
set
-
up also provides the
flexibility for people to arrange gel membranes in different
ways
(Fig. 5a),
thus
fit
ting
their water
containers in house or rooftop devices made by themselves.
We also agree with the referee that the collection rate based on the
total
water surface
area
should be provided to show the overall efficiency of our rooftop prototype.
The results
are
shown in Figure R1 below. We can see that the daily water collection rate is much lower
because
about half of the water surface was covered by
the passive PU foam.
As a comparison, t
he PU
foam
itself
showed a daily water collection rate of 1.0 to 2.2 L
/
m
2
in our control experiment
.
The
collection capability of rooftop device can be improved by assembling more PVA/PPy gel
membranes.
We have added the data in Section
22 of Supplementary Information.
Fig
. R1
Daily water collection per square meter of
total water surface
during rooftop tests. Red:
water collected during daytime (8 am to 8 pm); blue: water collected during nighttime (8 pm to
next day’s 8 am).
(2) They reported that the water vaporization enthalpy is significantly reduced to 1000 kJ
kg
-
1,
which is even 300 kJ kg
-
1 lower than the enthalpy reported by Yu, G. H., et al. [Nat. Nanotech.
13, 489
-
495 (2018)] How to explain the further decrease of enthalpy? Whether it’s the micro
-
trees morphology or the different pore size or something else?
And the differential scanning
calorimetric (DSC) measurement needs to be supplemented to prove the reduced evaporation
enthalpy.
Response:
We thank the referee for bringing up this important point
. We actually
had
noticed this decrease
of enthalpy and di
scussed with
other researchers including
the authors of Nat. Nanotech paper
about this enthalpy difference. We found that this was
most possibly
caused by the different
hydrolysis degree of PVA used in our study. The PVA
reported
in
the
Nat. Nanotech. pape
r had a
hydrolysis degree
of ~88% and the PVA used in our study had a hydrolysis degree of ~97% (the
hydrolysis degree of our PVA had been specifically
mentioned
in the
E
xperimental
S
ection
). PVA
with
a
higher hydrolysis degree contains more
OH groups on their polymeric chains, which could
bind more water molecules and reduce the enthalpy.
Similar results have been found in other
studies. In the
Nat. Nanotech. paper
1
,
a
control
gel
sample with a higher
PVA concentration
showed
a
n
enthalpy around 1000 kJ/kg.
Zhou
,
X.
,
et al.
2
reported that by adding
small
amounts
of chitosan (which contains more
OH and
NH
2
groups than PVA), the
water
evaporation
enthalpy of PVA/PPy gel system can be tuned from
~
1400 to
~
8
00
kJ/kg.
Our measurements
proved that the enthalpy was not affected by the surface microstructures (Fig.
S
21). Moreover,
the porous microstructure of PVA/PPy gel was not affected by the hydrolysis degree of PVA since
the molecular weight of PVA, concent
rations of crosslinkers, and
concentrations
of PVA and PPy
remained
same.
As suggested by the referee, we used DSC to measure the water vap
orization enthalpy in
our gel
, as shown in Figure R2 below. T
he gel
sample
was
placed in an open Al crucible
and
me
asured with
a
linear heating rate
of
5 K/min, under
a
n
itrogen flow (20 mL/min),
in the
temperature range from
2
0 to
1
8
0 °C. The effectiv
e specific heat capacity
was
calculated by
comparing the heat flow of measu
red gels with that of the
standard sapphire
sample
.
We
firstly
validated our measurements by conducting the tests on free water.
The measured enthalpy
of
free
water is 24
24
kJ/kg
, which is very close to the theoretical value of 24
50
kJ/kg
. The
measured
water vaporization
enthalpy in our hybrid gel i
s 1
73
5
kJ/kg
.
As comparisons, t
he
gel sample
reported in Nat. Nanotech. paper
1
has a DSC measured enthalpy of 1919 kJ/kg
(which
has
an
enthalpy of ~1300 kJ/kg in evaporation experiment)
and its control sample with a higher PVA
concentration has a measured
enthalpy of 1765 kJ/kg
.
As mentioned above, th
is
control sample
also
showed a
n
enthalpy around 1000 kJ/kg
in the evaporation experiment.
Note that the
enth
alpy values calculated from DSC
are higher than those tested in evaporation expe
riment
s
,
since the DSC test and
evaporation test present a full dehydration and s
lightly dehydration
processes,
respectively
1
.
Our DSC measurements confirmed the results from our evaporation
tests.
We
have
added the DSC results and related discussions in Section
15 of Supplementary
Information.
We
want to further stress on
that
t
he water evaporation rate of gel
evaporator
s is
determined by several factors including porous network, light absorption ability,
water content,
water vaporization enthalpy,
work
ing conditions (environmental temperature, humidity, etc.)
surface structures
, etc
.
Our study
focused on
the effects of
surface microstructures
and provided
meaningful information about regulating the surface temperature and vapor flow through
surface micr
ostructure design
.
Fig
. R2
DSC measurements on free water and PVA/PPy hybrid gel.
(3) As shown in Fig. 5e, the condensation device is equipped with a transparent plastic cover and
without forced heat transfer. This device results in high optical loss
caused by condensates and
limited heat transfer, which have adverse effects on condensation [Energy & Environmental
Science 11, 1510
-
1519]. Especially, in their design, the middle part of covering is horizontal, which
impede the collection of water and fu
rther result in higher optical loss.
Response:
We
agree with the referee’s comments on our floating prototype, which has
a
lower working
efficiency (a d
aily water collection rate of ~
19
L/
m
2
) due to the problems pointed out by the
referee.
However, our
study
focused on the design
of
micro
-
structured gel membranes
as
interfacial solar steam generation materials
and
understanding their
intrinsic
properties
, but not
on the optimization of
the whole
water collection
system
.
We showed this
simple
hand
-
made
p
rototype
to prove the concept of
a floating
water collection device
with day and night
working
modes
.
It demonstrated that the solar steam generation and fog harvesting functions could be
coupled in one water collection device.
We understand that t
he
floating device can be improved in many ways
to realize the full
potential of our gel membranes
, as well as other high
-
performance interfacial solar evaporators
.
For example
s
,
a
semi
-
spherical cover with high sunlight transparency can be installed. New
ope
ning
and closing
mechanisms can be applied to the cover.
Remote control function
s
can be
used to remotely switch the working modes. A
cooling part can be equipped to promote the
condensation
.
However,
we realize that
these works
require
efforts from optic
al, mechanical,
thermal, electrical, and electronic engineering
and
thus
are beyond the focus of this study
.
Development of water condensation and collection
systems
for large
-
scale applications
can be
a
very interesting topic for future studies
.
We
have
added
more
perspectives on floating device
improvement in the Discussion Section.
(4) I think the first question of reviewer #4 has not been addressed. The author thinks that the
fog collection processes only occur on the gel surface and will not wash ou
t the contaminants in
the gel. However, because of high concentrations of pollutants in the absorber, diffusion of
pollutants into fog water will occur. The inductively coupled plasma spectroscopy of fog water
should be supplemented.
Response:
We thank the referee for the
suggestion to assess the quality of collected fog water
.
We are sorry
that we couldn’t find a plasma spectroscopy for tests during the Covid
-
19 pandemic. To address
the referee’s concerns, we designed a control experiment in la
b.
To mimic a gel sample which is
possibly contaminated during solar evaporation, w
e used a PVA/PPy gel sample to evaporate
NaCl solution
with
a very high
salinity of 100
under one Sun irradiation for 6
h. Then we
conducted fog collection test on this gel
sample
in lab for
6
h (the set
-
up can be found in Section
4 of Supplementary Information). The artificial fog was generated by adding
dilute
NaCl solution
into the humidifier
(so we can
measure
the salinity by
conductance tests
, as described in Section
16
of Supplementary Information)
.
The salinity of
collected
fog
water
was measured and
compared to that of water in the humidifier.
This solar evaporation and fog collection cycle was
repeated for 10 times using the same gel sample. The salinity of water coll
ected during each fog
c
ollection
process
is
recorded
in the Table below. We can see that even
when
we used highly
concentrated saline water during solar evaporation, the salinity of water collected during fog
collection was not increased.
The results
demon
strate
that the
collected
fog water is not
contaminated by the gel evaporator
.
This can be explained by
thes
e
reasons:
1.
The
hydrogel evaporator
s
show an antifouling
property and long
-
term stability.
T
he salts or pollutants
don’t accumulate in the gel membrane
but mainly
remain in the original solution
3, 4
.
2.
The fog collection
process
on our micro
-
structured
gel membrane is fast.
T
he
cycle of fog droplets nucleation followed by their transport, growth,
and eventual drainag
e of the large water drops
takes
an average period of ~20 s
and t
he fog
droplets keep moving on the surface of gel membrane
(Fig. 3a
,
Fig.
S
6,
and
Movie
S
1).
In such
short time, the diffusion of
low
-
concentration
pollutants in the gel matrix into the quickly moving
fog droplets on the gel surface is difficult to occur.
3.
In all fog collection
processes
(in lab, rooftop
test, or in floating device), the gel membranes are tilted to facilitate the droplets transport
(Fig.
S
5a, Fig.
S
26, and Fig. 5f
)
and they
have
n
o
contact with the bottom saline water. Thus it’s not
possible for the pollutants to diffuse from the bottom water to the fog droplets on gel top
surfaces through gel membranes.
We have added
the experimental results and related discussions in Section 16 of
Supplementary Information.
Table R1.
The calculated salinity of water during in lab fog collection test.
Fog collection test #
Average salinity of collected
water (‰)
Average salinity of wa
ter in
humidifier (‰)
1
0.00126
0.00119
2
0.00148
0.00122
3
0.00137
0.001
3
8
4
0.00119
0.00121
5
0.00122
0.00135
6
0.00130
0.00117
7
0.00134
0.00129
8
0.00125
0.00109
9
0.00113
0.001
3
6
10
0.00117
0.00125
References
1.
Zhao, F. et al. Highly efficient solar vapour generation via hierarchically nanostructured
gels. Nat. Nanotech. 13, 489
-
495 (2018).
2.
Zhou, X., Zhao, F., Guo, Y., Rosenberger, B. & Yu, G. Architecting highly hydratable
polymer networks to tune the water
state for solar water purification. Sci. Adv. 5,
eaaw5484 (2019).
3.
Zhou, X., Zhao, F., Guo, Y., Zhang, Y. & Yu, G. A hydrogel
-
based antifouling solar
evaporator for highly efficient water desalination. Energy Environ. Sci. 11, 1985
-
1992
(2018).
4.
Guo, Y. et a
l. Biomass
-
Derived Hybrid Hydrogel Evaporators for Cost
-
Effective Solar Water
Purification. Adv. Mater. 32, 1907061
(2020).
REVIEWERS' COMMENTS
Reviewer #2 (Remarks to the Author):
All the issues have been addressed. It can now be published.