See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/313775429

Capillary driven electrokinetic generator for

environmental energy harvesting

Article in Materials Research Bulletin February 2017

DOI: 10.1016/j.materresbull.2017.02.022

CITATIONS READS

0 35

5 authors, including:

Changzheng Li Liu Kang

Wuhan University Wuhan University
6 PUBLICATIONS 9 CITATIONS 25 PUBLICATIONS 176 CITATIONS

SEE PROFILE SEE PROFILE

Huidong Liu Xuejiao Hu

Wuhan University Wuhan University
19 PUBLICATIONS 53 CITATIONS 59 PUBLICATIONS 782 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Flexible electret films View project

environmental energy harvesting View project

All content following this page was uploaded by Xuejiao Hu on 09 May 2017.

The user has requested enhancement of the downloaded file.

 Materials Research Bulletin 90 (2017) 8186

Contents lists available at ScienceDirect

Materials Research Bulletin

journal homepage: www.elsevier.com/locate/matresbu

Short communication

Capillary driven electrokinetic generator for environmental energy

harvesting
Changzheng Lia , Kang Liua,* , Huidong Liua , Bing Yanga , Xuejiao Hua,b,**
a
School of Power and Mechanical Engineering, Wuhan University, Wuhan, HB 430072, China
b
MOE Key Laboratory for Hydraulic Machinery Transients, Wuhan, HB 430072, China

A R T I C L E I N F O A B S T R A C T

Article history:
Received 28 January 2017 We propose a novel electrokinetic generator driven by the capillary force. By coupling the capillary with
Received in revised form 11 February 2017 water evaporation, the generator can output power as long as water evaporation exists, thus it can
Accepted 13 February 2017 harvest environment energies from waste heat, wind power to solar energy. Experimental results show
Available online 15 February 2017 that a capillary driven electrokinetic generator, with evaporation area of 4.9 cm2 and uidic channel area
of 2 mm2, can output maximum electrical voltage of
40 mV and current of 9.6 mA. The voltage and
Keywords: current can both be greatly enhanced by reducing the area of uidic channel, approaching maximum
Capillary power of 0.22 mW according to the theoretical calculations. In addition, a live-tree driven electrokinetic
Electrokinetic
generator with an output electrical power of 2.7 mW is demonstrated. These results show that the
Streaming current
capillary driven electrokinetic generator is of promising potential in self-powered systems.
Self-powered system
2017 Elsevier Ltd. All rights reserved.

1. Introduction external or internal to the channels [1518]. Because of the

Efciently producing electricity from renewable sources is a
crucial goal of current energy research. Recently, the energy
conversion in micro- and nanouidic devices have attracted
increasing attention [1]. Many studies have focused on the
electrokinetic phenomena known as streaming currents and
streaming potentials, which arise when a pressure driven uid
ows through a narrow channel [24]. Some researchers employed
porous materials to generate streaming currents [4,5], while others
concentrated on the electrical properties of single well-dened
channel [611]. A microjet structure is also proposed to improve
the efciency of energy conversion [1214]. However, the uid
ows in the mentioned devices are all driven by external pressure
force, which consumes lots of high-grade energy and cannot be
employed to supply power for self-powered systems.
Capillary force which exists inherently on curved liquid
interface is a convenient driving force for the creation of
pressure-driven ow in micro- and nanouidic channel. Passive
ow in narrow channels can be realized by utilizing menisci either
* Corresponding author.
** Corresponding author at: School of Power and Mechanical Engineering, Wuhan
University, Wuhan, HB 430072, China.
E-mail addresses: liukang@hust.edu.cn (K. Liu), xjhu@whu.edu.cn (X. Hu).
dependence of the capillary pressure on the characteristic radius of
the air/liquid interface curvature [19], systems using internal
menisci offer the ability to create very high pressure difference
along channels and have been successfully applied to single
channels and heterogeneous arrays to supply near-constant
continuous ow. Isolate internal menisci whose positions move
with time are impractical to utilize, because it cannot last for a long
time. However, static air/liquid menisci can be produced by
coupling with transpiration, which can provide sustaining capil-
lary force and thus continuous liquid ow [20].
In this paper, an electrokinetic generator driven by capillary
force using coupled capillary/transpiration effects is proposed. The
whole system is driven by the transpiration of water, which could
be induced by temperature difference, blowing or sunlight
illumination, etc. Thus, different kinds of environmental energy
can be harvested simultaneously. Experiments are carried out to
demonstrate the performance of the generator by employing
alumina nanoporous membrane as uidic channel to generate
streaming current and transpiration surface to supply capillary
force. Theoretical analysis on the output voltage and electrical
power are also presented based on a Poisson-Boltzmann descrip-
tion of electrostatic potential and a Navier-Stock description of the
ow in the nanochannel. At last, by utilizing the capillary force of
the leaves of a live tree, we demonstrate a live-tree-driven
electrokinetic generator for the rst time.

http://dx.doi.org/10.1016/j.materresbull.2017.02.022
0025-5408/ 2017 Elsevier Ltd. All rights reserved.
82 C. Li et al. / Materials Research Bulletin 90 (2017) 8186
2. Experimental details
2.1. Electrode fabrication
Thin layer of Ag and AgCl are deposited directly on the two sides
of the nanoporous alumina nanoporous membrane to work as
electrodes. The fabrication procedure is illustrated as shown in
Fig. S1. Firstly, a layer of Cr (about 15 nm) is deposited on one side of
the alumina membrane to enhance the adherence between the
membrane and the latter deposited Ag. Then the Ag layer
(
100 nm) is deposited via magnetron sputtering (MS). After that,
a silver wire is attached on the Ag lm with silver paste to work as
an extending for easier connection. AgCl layer is electrodeposited
on the surface of the silver lm by cycle voltammetry in the
solution containing 0.1 M HCl, with continuous stirring. Then, the
other side of the membrane is deposited with Cr and Ag layer (also
15 nm and 100 nm respectively) and attached to a silver wire,
following the same procedure as the rst side. Here the membrane
is processed side by side to avoid electro-deposition (ED) occurs on
both side of the membrane during AgCl deposition.
2.2. Tree generator
We fabricate a live tree generator by connecting the electroki-
netic generator with a twig. During the process, two key points
should be achieved: i) perfect sealing to ensure the liquid
continuum; ii) remove the noncondensable gas in the liquid to
avoid bubble occurrence when the liquid pressure decrease. Firstly,
the incision of the twig and all the equipment are boiled in the NaCl
solution for about an hour to remove the noncondensable gas.
Then we quickly connect the twig and the generator under the
boiled liquid. At last, the whole setup is xed up for the
measurements after temperature of solution recovers to room
temperature. During the cooling process, a thin layer of oil is
dropped on the top of the solution to avoid the noncondensable gas
dissolving into the solution again.
3. Results and discussion
Fig. 1a shows the structure of the proposed generator. The
central part is a traditional electrokinetic generator, in which
stream currents generate in the nanoporous membrane across the
uidic channel when electrolyte solution ows through it. On the
two sides of the membrane, electrodes are plated on to convert
Fig. 1. Schematic of a capillary driven electrokinetic generator. (a) The structure. (b)
Static air/liquid menisci formed in the porous membrane combined with
transpiration. (c) Inner structure of the nanoporous membrane and electrodes.
(d) SEM image of the nanoporous membrane after depositing electrodes. (e)
Schematic illustration of the streaming current generate in a single nanochannel.
chemical energy to electricity. Unlike a normal pressure driven
electrokinetic generator, the uidic ow in this work is driven by
the capillary force supplied by the air/liquid menisci (Fig. 1b)
forming in the top placed transpiration surface (porous material).
Continuous transpiration of water in the top surface ensures static
air/liquid menisci, so as to maintain a steady ow through the
channel below. Therefore, any energy that may induce water
transpiration in environment can be harvested to generate
electricity, such as wind power, solar energy, heat, etc.
In the experiments, the whole device is constructed in a
cylinder tube (diameter 2.5 cm) made of organic glass. Streaming
currents are generated in a piece of alumina nanoporous
membrane (anodic aluminum oxide) with thickness 60 mm and
surface fraction 60% (Fig. 1c). As shown in Fig. 1c, the membrane
consists of tens of thousands of vertical nanochannels with
diameter of about 200 nm. Thin layers of Ag and AgCl (
100 nm)
are plated on the two side of the membrane to work as electrodes.
Fig. 1d shows the SEM image of the membrane after depositing
electrodes. The streaming current generating in a nanochannel is
schematically illustrated in Fig. 1d. The diameter of the uidic
channel for solution to ow through is 1.6 mm. NaCl solution of
106 mol/L is prepared as working solution, so as to get the best
performance [7]. Before all the test, the electrokinetic property of
the employed alumina nanoporous membrane is evaluated by a
pressure driven experiment, as shown in Fig. S2. The pressure
difference of 0.5 MPa will generate a streaming potential of 1.4 V.
The transpiration surface (diameter of 1.5 cm) is made of the same
porous membrane as used to generate streaming currents. Fig. 2a
inset shows the photograph of the fabricated generator.
As shown in Fig. 2a, the generator is covered by the
polyethylene lm to prevent evaporation before experiments,
thus it outputs a voltage of zero. When the polyethylene lm is
removed, the output open-circuit voltage (V) increases gradually to

7 mV at ambient condition of 14  C in temperature and 35% in
relative humidity. When we increase the environmental wind
velocity to promote the evaporation, V shows rapid increase with
the wind velocity (Fig. 2b). The voltage achieves a value of 20 mV
when velocity of wind is
20 m/s. As we know, the evaporation
rate of the water can also be enhanced by increasing the
environment temperature and reducing the relative humidity.
Thus here we also exam the output performance of the generator
under different intensity of sun light irradiation and record the
induced temperature and humidity variation. As shown in Fig. 2c,
the output voltage increases with the increased temperature and
reduced humidity as we predicted. Under an environmental
temperature of 58.8  C and relative humidity of 7.7%, the generator
outputs a high open-voltage of
39 mV. Plotting the relation
between the evaporation rate and output voltage in the two test,
we nd that, the voltage has a linearly relation with evaporation
rate.
To further investigate such a relation and the underlying
mechanisms, theoretical analysis is carried out by employing the
Poisson-Boltzmann equation [21] to characterize the streaming
current and Navier-Stock equation to characterize the ow in a
single nanochannel (details in Supplementary Note). Since the
membrane consists of tens of thousands of the same channels, the
real streaming current and internal resistance is thus:
Rin = R0/n0, I = I0  n0, (1)
where R0 and I0 are the internal resistance and streaming current of
a single nanochannel, n0 is the number of the nanochannels
available in the uidic channel. The open-circuit voltage equals to
that of a single channel. The diameter of the nanochannel is taken
as 200 nm in calculation according to the size of the alumina
nanopores, and the surface charge density of the channel wall (s),
 C. Li et al. / Materials Research Bulletin 90 (2017) 8186 83

Fig. 2. Open-circuit voltage of the generator. (a) Output voltage of the generator at ambient condition of 14  C in temperature and 35% in relative humidity. Inset is the
photograph of the fabricated generator. (b) Output voltage of the generator under different wind-velocities. (c) Output voltage of the generator at environment of different
temperatures and humidity. (d) Relation between open-circuit voltage and evaporation rate of the generator. Solid line is the results of the theoretical predictions.

which is proportional to Zeta potential, is obtained by tting the
relation between the open-circuit voltages and the applied
external pressures derived in the external pressure driven
experiments (18 mC/m2, Fig. S2). The transpiration rate on the
transpiration surface is calculated as J = (ht/rcp)MA/RT(pc  p1),
where ht is the heat transfer coefcient of convection, M is the
molecule mass of water, A is the area of the transpiration surface,
and pc, =p1 are the vapor pressure on the surface of air/liquid
curvature and of the environment respectively. pc is determined by
the Kelvin equation. r, cp are the mass density and specic heat of
air. T is the environment temperature. The red line in Fig. 2d shows
the theoretical results of the open-circuit voltage, which is
consistent with the experimental results. It indicates that, the
calculated streaming current is identical with the values in
experiments.
To investigate the output electric power of the generator, we
measured the output voltage of the generator with different load
resistances. Since a traditional electrokinetic generator can be
modeled by the electrical equivalent circuit as shown in Fig. 3a
inset [8], the output voltage (voltage on the load resistance) can be
calculated as
V = IRinRL/(Rin + RL),
where I is the total streaming current of the generator, RL is the load
resistance and Rin is the internal resistance which is determined by
the surface potential of the channels and the concentration of the
solution. The streaming current, I, is proportional to the ux of the
uid ow through the channels, which equals to the transpiration
rate on the top surface in steady state. Thus, the wind velocity
determines the streaming current in the channels with certain area
of transpiration surface and constant number of nanochannels.
Fig. 3a shows the load resistance dependence of the output voltage
with different wind velocities (w). By tting the experimental
results with Eq. (2), I and R0 are obtained for each wind velocity. Rin
is found to be always 4.1 kV, and I are 2.90 mA, 4.56 mA, 5.39 mA for
5 m/s, 10 m/s, 15 m/s, respectively. Here we also compared the
internal resistance and streaming current with the experimental
results. The theoretical Rin in calculation is 1.72 kV, which is
slightly lower than the experimental value (4.1 kV), while the
streaming current is a little higher (11 mA, 16 mA, 20 mA for 5 m/s,
10 m/s, 15 m/s respectively). The accordant discrepancy may come
from the resistance of the electrodes, for the resistance of the
electrodes is also on the order of 1 kV according to experiment
measurement.
Fig. 3b shows the electric output power (P) when the wind
velocity is xed at 15 m/s. The results agree well with the
theoretical prediction using the obtained R0 and I in the tting. P
shows a peak at
4 kV where RL = R0 The maximum P is 30 nW,
which seems a little lower for practical applications. The low
output electrical power is induced by the low pressure drop
through the channel, about 4 kPa even when w = 15 m/s, as shown
in Fig. S3. Thus we can promote the driven pressure to enhance the
electrical power, by reducing the diameter of the uidic channel (d)
or increasing the area for evaporation. Reducing the uidic channel
diameter can lead to two contrary effects to P with given
(2)
environmental condition (the transpiration rate is constant): 1)
increase of driven pressure, ow velocity and thus streaming
current in a single channel; 2) reducing the number of nano-
channels. Since the electrical power in a single channel (P0) is
quadratic proportional to the streaming current, the total electrical
power, which is the sum of P0 in all channels, can thus be efciently
enhanced by reducing the uidic channel diameter. However, the
maximum driven pressure is limited by the capillary pressure that
the transpiration surface can supply.
According to YoungLaplace equation, the maximum capillary
force that a porous membrane provided is determined by the
84 C. Li et al. / Materials Research Bulletin 90 (2017) 8186

Fig. 3. Output power of the generator. (a) Load resistance dependence of the output voltage at different wind velocities. Solid line shows the tting with suitable I and R0. Inset
is the equivalent circuit of the generator with load resistance. (b) Output power of the generator with different load resistance under the wind velocity of 15 m/s. (c)
Theoretical results of the electrical power variation with different evaporation rates with uidic diameter of 10 mm, 20 mm and 30 mm, respectively. (d) Theoretical results of
the electrical power variation with uidic diameters under different wind velocities. The wind velocity is set as 1 m/s, 3 m/s and 5 m/s, respectively. The environment
condition is taken as 25  C in temperature and 73% in humidity.

diameter of the pores inside. A theoretical high capillary pressure
up to 60 MPa could be achieved when the pore diameter is 2 nm,
below which, a regular curvature cannot form because of the effect
of disjoining force [22]. It is interesting to found that, the liquid
under the menisci would maintain at negative pressure because
the pressure drop is larger than 1 atm. In fact, negative pressure of
3.3 MPa and 4 MPa has already been found in experiments by
using porous materials in earlier literatures [23,24]. Hence, much
larger capillary pressure than that in experiments is available for
such a generator. Assuming the pore diameter of the transpiration
surface is 2 nm, the maximum theoretical results for a setup as we
used in experiment are shown in Fig. 3c. P increases apparently
with the evaporation rate. While with the decrease d, P increases
accordingly in high d region, but shows a peak at low d region as
shown in Fig. 3d. Reducing d would apparently enhance the ow
rate in nanochannels as well as the driven capillary pressure. The
peak is the result of the two inverse reactions with the reducing of
d: 1) the enhancement of ow velocity and electrical power as we
discussed above; 2) the inhibition of evaporation in the transpira-
tion surface because of reduced vapor pressure above the
curvature according to Kelvin equation. In high d region, the
effect of vapor pressure decrease is negligible, but when d is small
enough, it becomes accountable and even larger than the effect of
the streaming current enhancement. It should be noted that, with
suitable d and w, the maximum power can be as high as 0.22 mW.
Promoting the area for evaporation is another effective way to
increase the output power. Here we demonstrate a live-tree-driven
electrokinetic generator for the rst time, by utilizing the water
transportation on the leaves of a live tree (Fig. 4a). As shown in
Fig. 4b, the tree generator using a Magnolia Grandiora twig with

10 leaves output an open-circuit voltage of
0.15 V, and the
variation trend of the voltage in 20 h keep in consistent with the
evaporation rate of the tree. By tting the output voltages at
different load resistances with Eq. (2), Rin and I are obtained to be
3.7 kV and 53 mA, respectively. The output current is found to be
one order of magnitude higher than that of the porous lm driven
generator above, which induces a high output power of 2.7 mW
when RL equals Rin (Fig. 4c, d). As we know, the transpiration of
plants would vary with the environment temperature and light
intensity, due to the self-thermoregulation. Thus here we also
measured the voltage variation with different environmental
temperature and light intensity. As shown in Fig. 4e, f, the output
voltage increases almost linear with the temperature, which
agrees well with the results calculated from the experiment-
obtained transpiration rates. V is also observed to increase with the
light intensity; however it tends to be saturated at high intensity.
Such a saturation behavior is believed to be caused by the
shutdown of some leave-stomas under strong light. Based on the
results and analysis above, the output of such a tree generator
varies sensitively with the physiological cycle of plants, which
could potentially works as a biological sensor.
4. Conclusions
In summary, a capillary-driven electrokinetic generator is
proposed in this study for harvesting environmental energies
including wind, solar energy and all kinds of waste heat. The
output voltage and electrical power generated is theoretically and
experiment analyzed. The theoretical maximum electrical power
output for such a generator with the transpiration surface
diameter of 1.5 cm and uidic channel diameter of 10 mm could
be as high as 0.22 mW. To increase the evaporation area, a live-tree
driven electrokinetic generator is also experimentally demon-
strated, which output an open-circuit voltage of 0.2 V and electrical
 C. Li et al. / Materials Research Bulletin 90 (2017) 8186 85

Fig. 4. Performance of the electrokinetic generator driven by a live tree. (a) Photograph of the tree generator. (b) Open-circuit voltage and evaporation rate variation in 20 h. (c,
d) Output voltage and power of the tree generator with different load resistances. (e, f) Dependence of the output voltage on the environmental temperature (e) and light
intensities (f). The calculated results in (e) and (f) are derived from the measured transpiration rates and theoretical calculation in Supplementary Note. The errors of the
measured voltages are obtained from the uctuation of the voltage in measurements, while the errors of the calculated voltages are derived from the uctuation of the
measured evaporation rate. All the tests were performed at ambient condition of
14  C in temperature and
30% in relative humidity.

power of 2.7 mW. These results demonstrate the potential of the
capillary driven electrokinetic generator in applications of
supplying power for self-powered systems.
Acknowledgment
The authors gratefully acknowledge the nancial supports
received from the National Natural Science Foundation of China
(No. 50906064 and 51606082)
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.
materresbull.2017.02.022.
References
[1] S.E. Pennathur, J.C. Eijkel, A. Berg, Energy conversion in microsystems: is there
a role for micro/nanouidics? Lab Chip 7 (2007) 12341237.
[2] J. Yang, F.Z. Lu, L.W. Kostiuk, D. Kwok, Electrokinetic microchannel battery by
means of electrokinetic and microuidic phenomena, J. Micromech. Microeng.
13 (2003) 963970.
[3] F. Munshi, S. Chakraborty, Hydroelectrical energy conversion in narrow
connements in the presence of transverse magnetic elds with electrokinetic
effects, Phys. Fluids 21 (2009) 122003.
[4] W. Olthuis, B. Schippers, J. Eijkel, A. Berg, Energy from streaming current and
potential, Sens. Actuators B-Chem. 111 (2005) 385389.
[5] M.C. Lu, S. Satyanarayana, S. Karnik, A. Majumdar, C.H. Wang, A mechanical-
electrokinetic battery using a nano-porous membrane, J. Micromech.
Microeng. 16 (2006) 667675.
[6] H. Daiguji, Y. Oka, T. Adachi, K. Shirono, Theoretical study on the efciency of
nanouidic batteries, Electrochem. Commun. 8 (2006) 17961800.
[7] F.H.J. Van der Heyden, D.J. Bonthuis, D. Stein, C. Meyer, C. Dekker, Electrokinetic
energy conversion efciency in nanouidic channels, Nano Lett. 6 (2006)
22322237.
[8] F.H.J. Van der Heyden, D.J. Bonthuis, D. Stein, C. Meyer, C. Dekker, Power
generation by pressure-driven transport of ions in nanouidic channels, Nano
Lett. 7 (2007) 10221025.
[9] Y. Ren, D. Stein, Slip-enhanced electrokinetic energy conversion in nanouidic
channels, Nanotechnology 19 (2008) 195707.
[10] F.H.J. Van der Heyden, D. Stein, C. Dekker, Streaming currents in a single
nanouidic channel, Phys. Rev. Lett. 95 (2005) 116104.
[11] D. Gillespie, High energy conversion efciency in nanouidic channels, Nano
Lett. 12 (2012) 14101416.
86 C. Li et al. / Materials Research Bulletin 90 (2017) 8186
[12] A.M. Dufn, R.J. Saykally, Electrokinetic hydrogen generation from liquid water
microjets, J. Phys. Chem. C 111 (2007) 1203112037.
[13] A.M. Dufn, R.J. Saykally, Electrokinetic power generation from liquid water
microjets, J. Phys. Chem. C 112 (2008) 1701817022.
[14] H. Daiguji, P. Yang, A.J. Sezeri, A. Majumdar, Electrochemomechanical energy
conversion in nanouidic channels, Nano Lett. 4 (2004) 23152321.
[15] D. Juncker, H. Schmid, U. Drechsler, H. Wolf, M. Wolf, M. Michel, N. Rooij, E.
Delamarche, Autonomous microuidic capillary system, Anal. Chem. 74
(2002) 61396144.
[16] L.J. Yang, T.J. Yao, Y.C. Tai, The marching velocity of the capillary meniscus in a
microchannel, J. Micromech. Microeng. 14 (2004) 220225.
[17] M. Zimmermannab, H. Schmida, P. Hunzikerb, E. Delamarche, Capillary pumps
for autonomous capillary systems, Lab Chip 7 (2007) 119125.
View publication stats
[18] E.W. Washburn, The dynamics of capillary ow, Phys. Rev. 17 (1921) 273283.
[19] G.K. Batchelor, An Introduction to Fluid Dynamics, Cambridge Univ Pr., 2000.
[20] N.S. Lynn, D.S. Dandy, Passive microuidic pumping using coupled capillary/
evaporation effects, Lab Chip 9 (2009) 34223429.
[21] S.H. Behrens, D.G. Grier, The charge of glass and silica surfaces, J. Chem. Phys.
115 (2001) 6716.
[22] Jacob N. Israelachvili, Intermolecular and Surface Forces, vol 28, Academic Pr.,
2011, pp. 303.
[23] T.D. Wheeler, A.D. Stroock, The transpiration of water at negative pressures in a
synthetic tree, Nature 455 (2008) 208212.
[24] W.D. Machin, A simple method for the generation of negative pressure in
liquids, Can. J. Chem. 76 (1998) 15781580.