вход по аккаунту


Production of Hydrogen by Electrolytic Purification of Water.

код для вставкиСкачать
Dev. Chem. Eng. Mineral Process. 14(1/2), pp. 71-84, 2006.
Production of Hydrogen by Electrolytic
Purification of Water
G. Mathieson”*,A. Langdon’ and G. Jamieson’
Departments of Chemistry and Materials and Process Engineering,
University of Waikato, Private Bag 3105, Hamilton, New Zealand
Works Filter Systems Limited, PO Box 20-271 Te Rapa, Hamilton,
New Zealand
Electrolytic purification of water is not new but has never been widely applied because of
failure to meet realistic standards of economy. A continuous process with low cost
consumables can compete with standard water treatment processes, even without energy
recovery in the form of hydrogen.
In order to extend the lower limit of feasible
conductivities a specialised cell conjguration was developed. This utilkes a novelform of
corroding anode that produces a highly eflective flocculant, in close proximity to a
cathode, separated by a simple porous membrane. The active surface of the corroding
anode is continually replenished. Flocculant clearance is achieved by high water flow
velocity in a flume. Hydrogen is produced separately from anodic products. Other
functions like electrolytic flotation, redox and sterilisation are synergistic. Because
electrical energy input dominates the operating cost there is a strong incentive to recover
the hydrogen, which is produced at about 0.02 Nm’ per mJ of water treated and about
25% eficiency.
Establishing economic means of producing hydrogen is by definition essential to use
of hydrogen as an energy carrier in a hydrogen economy. While electrolysis of water
is a well-known means of producing hydrogen, the required electrical energy has a
* Author for correspondence (gam
G.Mathieson, A . Langdon and G.Jamieson
relatively high value and some is lost in the process. If the hydrogen that is produced
is used either for heating by combustion, or to make electricity in another lossy step,
then the entire process cannot be justified on the basis of efficiency.
While using the hydrogen as a form of energy storage from low electrical tariff
periods has been considered, there is another possibility. By achieving another
process goal simultaneously, electrolysis may become economic. Electrolytically
induced flotation, flocculation and sterilisation are mechanisms that can be used to
purify water (Sixpence, 1977; Musquere, Ellingsen et al., 1983; Osasa, Nakakura, et
al., 1993; Ordonez, 1997; Tao, 1999), which would otherwise have to be carried out
by other means. The hydrogen and oxygen bubbles produced by electrolysis are small
enough to attach to fine contaminants and float them out, in analogy to traditional
dissolved air flotation systems. The corrosion of a trivalent anode produces metal ions
that rapidly hydrolyse into amorphous floc particles, in analogy to familiar chemical
flocculation. The production of oxidising agents at a non-corroding anode leads to
disinfection of the treated water.
The striking advantage of electrolytic treatment is that combined functions of
flotation, flocculation and sterilisation can be integrated in a single structure. For
example, if both bubbles and flocculant are released from an electrode, the bubbles
will rapidly float out the floc, allowing for simple separation. Moreover, there appears
to be a synergy between nucleation of bubbles and flocculation (da Rosa and Rubio,
2005). This is taken to the extreme by production of bubbles and flocculant in close
proximity at an electrode surface. Theoretical analyses, field trials and laboratory
experiments have been carried out to determine the most promising commercial
applications. Both the conductivity and pH of the water are crucial to whether or not
the processing is cost-effective. In certain conditions electrolytic water treatment is
both more effective and potentially less costly than established methods (Rohland,
1915 ; Stuart, 1946; Bonilla, 1947; Khosla, Venkatachalam, et al., 199I. ; Robinson,
2000; Tetrault, 2003A). In all cases, maximising the production efficiency of
hydrogen, i.e. lowering cell voltage by reducing cathodic overvoltage, was considered
as a sub-goal to maximising the cost-effectiveness of the entire process. From t h s
point of view, energy recovery in the form of hydrogen is seen as a bonus. Municipal
Production of Hydrogen by Electrolytic Purlfication of Water
wastewater plants that generate electricity from the natural gas fermented from the
bio-solids in wastewater can use the recovered hydrogen with little or no modification
to existing equipment. Because electrolytic processing is also applicable to relatively
small-scale operation at the scale of individual factories, there is an opportunity to
develop smaller scale hydrogen recovery and usage apparatus.
Theory and Modelling
For a simple direct current supply voltage VDc, the main sources of electrical energy
loss in an electrochemical reaction are shown as either opposing voltages, or
resistances, in Figure 1.
Figure I . The DC electrical equivalent of an electrolytic cell.
The equilibrium potentials for the anode and cathode are VAeand VC, respectively.
These cannot necessarily be considered constant with respect to current because both
the pH and temperature are subject to alteration by the passage of current. The
resistance of the supply, Rs, including metallic leads and the electrolyte are
considered to be ohmic (linear) functions of the current. This is well justified for Rs,
but as shall be explained later, the electrolyte resistance is likely to depart from ohmic
behaviour at high current densities. The anodic and cathodic overvoltages are qA and
qc. These are defined by the Butler-Volmer relation in Equation (1) and its simplified
G. Mathieson, A. Langdon and G. Jamieson
form, the current-centric Tafel equation in Equation (2) (Bockris and Reddy, 1970),
accurate when In11 > 118 mV (Riley, 2003), where n is the number of electrons
transferred by the reaction and 7 is the overpotential. These overpotentials may be
considered as equivalent to the voltages across the non-linear resistors RA and RC
respectively, and b s is reflected in Equation (3) where REis the electrolyte resistance,
Equation (2) shows that in order to get significant current to pass, an overvoltage
must be applied, yet Equation (4) shows that an overvoltage contributes directly to
energy losses. Hence all practical designs must apply some form of compromise, so
that the area of the electrodes is not too large while the energy efficiency is not too
low. The acceptable limits for either depend on the application.
Standard industrial electrochemical processes use high concentration, high
conductivity electrolytes, since this minimises the loss attributed to the term VRE in
Equation (4). For situations where the electrolyte is dilute and cannot be altered for
economic or environmental reasons, such as in wastewater, the loss due to electrolyte
resistance can be significant, even dominant, because while the electrode overvoltages
are functions ofthe logarithm of the current density, the voltage across the electrolyte
is linear or supra-linear with current.
The conductance of the electrolyte can be estimated from its conduct:ivity (a), the
area of the electrode interface ( A ) , and the average gap between the electrodes ( I ) , as
given by:
s, =u-A
.. . ( 5 )
Production of Hydrogen by Electrolytic Purification of Water
O\snaltage for
Hydrogen edutlon (V)
Voltage Drop (V)
for 10 mm gap
Electrolyte Voltage Drop (V)
for 1 mm gap
Voltage Drop (V)
for 0.1 mm gap
0 001
. . .”. .. ElectrolyteVoltage Drop (V)
Current Dendty A m“
at 10 mm gap for 6.6 M KOH
at 70 C a r d 100 Slrn
Figure 2. How electrolyte voltage drop is lowered to the order of the cathodic
overvoltage by reducing the gap; based on an iron cathode and fluid with 0.02 S m-’
conductivity like tap water (resultfor concentrated electrolyte shown for comparison).
Whle increasing the surface area offers lower losses, it also leads to an increase in
both size and capital cost. The next obvious target is to lower the distance between the
Figure 2 shows how the voltage drop across the electrolyte gap becomes dominant
if the electroIyte conductivity is reduced, and how reducing the gap size can lower the
losses. This result is based on iron cathode exchange current data (Pletcher and Walsh,
Whle Equations (1) and (2) are accurate at low to moderate current densities, at
high current densities the overvoltages can be raised by: inadequate diffusion of
dissolved reactants to the electrode, inadequate diffision of dissolved or precipitated
products from the electrode and poor clearance of gas bubbles from the electrodes.
However, if the bubbles of around 10 pm diameter that are formed are released
rapidly enough from the surface they stir the diffusion layers effectively, greatly
improving mass transport (Wendt and Kreysa, 1999).
Partial occlusion of the electrolyte gap by gas bubbles leads to an increase in
effective electrolyte resistance. Two extreme forms are explored for illustration.
(i) If a contiguous sheet of gas forms between two electrodes from bubbles that
coalesce, ths will grossly increase the effective resistance of the electrolyte even
if it does not fill the entire space, hence this is to be avoided if at all possible.
G.Mathieson, A. Langdon and G. famieson
(ii) Evenly dispersed fine bubbles will reduce the electrolyte conductivity from KO to
slightly more than predicted by the linear gas void to total volume ratio,
(Wendt and Kreysa, 1999). Theoretical estimates of the conductivity ratio are in
a narrow range defined by Equation (6), i.e. if the gas bubbles occupy half the
space then the resistance will be a little more than doubled:
For the simple case of parallel electrodes with plug flow of fluid that dominates
the motion of the bubbles, the terminal part of the flow will have the highest gas
loading. The build-up can be modelled as spatially linear so that the average level is
the same as that at the halfway point. Hence the term triangular gas void ratio,
shown in Equation (7) below, equal to half the value of the voiding at the exit point,
which can be used in place of cg in Equation (6).Given that V,,, is molar volume in m3
per mole and D, is the required faradic volumetric rate of hydrogen and oxygen
bubbles dose per fluid flow for an example of flotation treatment by simple
electrolysis of water, a low value of DIof 60000 A m*3s (1 N(Vmin) ) results in a
of about 1%. Note that this does not depend on the gap size, though for very narrow
gaps coalescence may cause the sheeting problem. Furthermore the utility in reducing
the gap size is limited by increases in fluid pressure drops and greater difficulty in
preventing contact of the electrode pairs. For example, the pressure drop in laminar
flow between parallel plates is proportional to the cube of the of the gap
(Wendt and Kreysa, 1999).
=--3 IVm
4 F Dl
(m3of hydrogen and oxygen gas per m3 of liquid)
Experimental Details
Consider an example of field work carried out at a tannery. Tannery effluent
is particularly noxious, because the chemicals used in the tanning process
are preservatives, including chromium compounds, and the pH tends to be high.
Production of Hydrogen by Electrolytic Purification of Water
Electrolytic processing is feasible because the high salt content of the effluent gives a
high electrical conductivity of around 1.5 S m-'. While research on the subject dates
back to early in the 20" century, (Rohland, 1915; Fassina, 1938) commercialisation
has not occurred, perhaps due to excessive power consumption. Other researchers
have produced promising results with rendering plant effluent (Tetrault, 2003A), but
with cell voltages of about 50 V (Tetrault, 2003B). Toward the end of 2004, Works
Filter Systems Limited was invited to trial an electrolytic effluent processor at a site
in the Eastern Waikato region of New Zealand.
The design of the field-trialled system is protected by NZ Provisional Patent
537700. This utilises a novel form of 6060 alloy aluminium as a corroding anode that
produces a highly effective flocculant, in close proximity to a steel cathode, but
separated by a simple porous nylon mesh membrane. The active surface of the
corroding anode is continually replenished. Flocculant clearance into the water is
achieved by high flow velocity in a flume. Hydrogen is produced outside this flow
and is therefore not mixed with anodic products. The effective cathode area was
about 0.4 m2, driven to current densities of up to about 100 A m-2.A smaller version
was made for laboratory work.
The composition of 6060 Aluminium alloy is shown in Table 1 and indicates that
for levels of up to 100 mglL of A1 added by corrosion, the total level of additional
metals is likely to be less than 1 mg/L with most elements well below 1 mg/L. In
particular, Cr is added at the rate of 0.0005 mg per mg of Al, which is insignificant
compared to the background level in Tannery Effluent at any reasonable level of A1
addition. Some of the more noble impurities such as Cu may not be anodically
oxidised, but released to the flow in metallic form as the bulk alloy disintegrates.
Because of the nylon membrane, the diffusion of metals from the main flow stream to
the cathode is limited and the rate of plating-out is low.
Power supplies: 1. D3800 DC power supply from Dick Smith Pty Ltd with < 1%
voltage ripple. 2. Continuously adjustable full-wave rectifier composed of an earthleakage protected 230 V single phase supply, a 4 A variac kindly supplied by Physics
Department, Waikato University, a step-down transformer, and four 50 A bridgerectifier modules in parallel on a heatsink. The output was fed to the elecrolyser via
G. Mathieson, A. Langdon and G. Jamieson
Table 1. Impurities in A1 Alloy EN A W-6060 (MA TTER Project 2001)).
Total other
Other elements
A1 remainder
Weight %
0.35 - 0.6
0.30 - 0.6
0.10 - 0.30
< = 0.15
< = 0.15
< = 0.10
< = 0.10
< = 0.10
< = 0.05
< = 0.05
> 98.35
80 A cable, and protected from short-circuit by a 100 A high rupture current (HRC)
fise. This power supply could operate continuously at 0 V to 30 V fill wave rectified
rms at 0 A to 40 A rms,and supply up to 60 A for a few minutes at up to 10 V,
limited by rectifier heat sinking. The peak voltage of this supply is approximately 1.5
times the rms value, and the 100 Hz harmonic component is significant. The
continuous variability enabled fine control of the current delivery during experimental
Electrical measurements were made in the field with a Fluke 337 clamp-meter. pH
measurements were made in reference to regular 2 or 3 point calibration.
Results and Discussion
The most surprising result in the field trials was the effectiveness of the treatment at
high pH, including very low residual aluminium and chromium levels for the pH of
the outflow (Pontius, 1990). The processing can lower the pH when the initial pH is
above 8. This would be expected where the anolyte and catholyte are separated, but a
pH drop was observed even in the whole outflow. This suggests that the normal
protective layer is continuously disrupted by electrolysis, and oxygen sourced from
outside the electrolytic system is able to corrode additional aluminium. ‘These effects
are still under investigation in the laboratory, supported by past studies (IA, Qu, et al.,
1999; Zeppenfeld, 2003).
Production of Hydrogen by Electrolytic Pur fication of Water
The current voltage characteristic was largely dependent on the conductivity and
state of the anode surface (Chen, Chen, et al., 2002). To a great extent the
characteristic of initial current with voltage was linear, suggesting that the electrolyte
and membrane dominate the total resistance. Most interestingly, in situations where
the treatment was effective, the initial current with a new anode was low, but the
current increased steadily over an hour or two to a level that could exceed double the
initial value. In cases of low salinity the current would usually decline monotonically
from the initial value and treatment was unlikely to be effective for very long, if at all.
Cell voltages from 3.5 V to 10 V were effective in treating tannery effluent. Current
doses (current per volumetric flow rate, or charge per volume) of between 1 and
5 A/(litre/min) were necessary to cause good flocculation. The low end was sufficient
for saline solutions, while the high end was necessary for heavily loaded tannery
effluent. Note that 1 A/(litre/min) corresponds faradically to about 5.59 ppm of A13+
or 10.56 ppm A1203equivalent.
Figures 3, 4 and 5 show the results from a trial on 12 May 2005. Turbidity,
measured in nephleometric turbidity units (NTU), is a good (though not linear)
indicator of the amount of suspended solids in the water. Because the quality of the
tannery effluent changes so quickly, the samples were taken over just a half hour
period, with inflows collected at either end of the time period. The figures show that
generally, the higher the dose of aluminium, the greater the reductions in outflow
turbidity, pH and chromium, though the change of the inflow also has an impact. The
cost-effectiveness of the electrolytic processing, including cost of electricity and
aluminium, is favourable compared to dosing with polyaluminium chloride (a
commercial, partly hydrolysed, flocculant).
The specialised anode has a clear advantage over other corroding anodes.
Trackmg of the water flow due to the normally inevitable uneven corrosion is
eliminated. The clearance of flocculant from the specialised anode is high, as long as
the flow speed is at least about 0.1 m s-’. This is in stark contrast to earlier models
where the floc would rapidly clog the anode and block both the flow and the current.
Hydrogen production was estimated faradically around 6.96 x lo” m3 of gas (at
STP) per m3 of fluid treated per charge dose. Moderate treatment at about
G. Mathieson, A. Langdon and G. famieson
3 N(litre/min), equal to 1.8 x lo5 C m-3, gives about 20 litres of hydrogen per cubic
metre of water treated. The efficiency of hydrogen production and the fraction of
energy that can be recovered from the electrical input are identical, and may be
estimated by the ratio of the standard potential for electrolysis of water (1.23 V) to the
cell voltage. For a cell voltage of 5 V, the efficiency is about 25%.
The purity of the hydrogen obtained depends both on the quality of the water and
the use of separating membranes. The high oxygen demand of many wastewaters will
consume at least some of the oxygen produced, but the traces of CH,,CO, C02 and
sulphur compounds that are present in mast organic wastewaters will inevitably
contaminate the hydrogen if the fluid flows along the cathode surface. This is of little
or no consequence if the fuel gas (BactogasQ - a registered trademark of Works Filter
Systems Limited) is simply combusted. Most of these contaminants are either harmful
or fatal to known fuel cell catalysts (Vennekens, 2000). For premium uses of pure
Figure 3. Comparison of electrolytic treatment on tannery efluent with
polyaluminium chloride dosing (inflow at 2:30pm: pH 8.9,850 NTU, I7250 uS/cm at
I9”C; at 3 pm: pH 8.6, 900 NTU, 16560 uS/cm at 19°C).
Production of Hydrogen by Electrolytic Purification of Water
Figure 4. Residual metals comparisonfor electrolytic treatment of tannery effluent.
Figure 5. pH data comparisonfor electrolytic treatment of tamely efluent..
G. Mathieson, A . Langdon and G.Jamieson
hydrogen, this suggests that a gas impervious, but ion-permeable membrane, be
employed and that uncontaminated liquid be used as the catholyte. While this will
restrict the treatment to that offered by the anode compartment, the only known
hnction that would be impaired is electroplating of impurities onto the cathode.
In some cases there is overall merit in having a relatively high cell voltage for a
corroding anode system, because it both enhances the pH changes in the vicinity of
the electrodes and the space-time yield, but at very high anodic overvoltage the
current fraction of anodic oxygen production increases at the expense of flocculant
production and quality. The ultimate limit for anodic overvoltage is when partly
corroded fragments of the anode are released by the disintegration of' the anode
surface. These fragments are not effective as flocculants and they greatly increase the
amount of metal added to the water.
The ideal treatment system depends on the water to be treated and the required quality
of the outflow. Some examples of the various field applications that have been tried
are listed by water type and suitable treatment combination in Table 2.
Beside iron water (Coup and Campbell, 1964), tannery effluent (Fassina, 1938)
and laundry water recycling, other trialled field applications include: landfill leachate,
municipal wastewater terminal effluent, and sludge returned from a clarifier. The
most promising of these, despite its relative difficulty, is landfill leachate (Ihara,
Shimada, et al., 2003) because of the value in making an inflow that contains
numerous pervasive and persistent toxins into an outflow that is acceptable for
discharge into stormwater drains. A novel specialised processor has been designed for
this application. A longer term strategy is to target point sources of, for example,
heavy metals and treat these streams before they enter municipal wastewater conduits.
If the treatments are successful, large amounts of impurities will be separable from
the bulk of the water. The means of separation, besides electroflotation, include
settling and filtration. In some ways, the challenge of disposal will remain, from a
much more concentrated form.
Production of Hydrogen by Electrolytic Purification of Water
Further laboratory work is continuing in order to elucidate some of the
mechanisms of treatment and establish measurable figures of merit, such as pH
change, for use in control and evaluation in applications.
Table 2. Examples of electrolytic water treatment options.
water type
Bore water with
ferrous iron at
about 20ppm.
Conductivity of
about 2000 pS
cm.' by NaCI.
Achieved by
rapid oxidation
of the iron to Fe
111 using a noncorroding anode
- no tlocculant
gas will
float out the
oxidation of
Fe II to Fe
Ill may
Chloride IS
converted to
chlorine at the
anode. That
remaining atter
Fe oxidation is
free chlorine.
Must not
I 0000 ps cm-'
Effluent with
Cr, suspended
solids and high
but variable pH.
Corrode an
aluminium or
iron anode to
gas will
float out the
but not
pH falls in
most cases.
improves the
health of
Near pure
or Iron.
Laundry water
recycling, water
that IS high In
pH and salinity
with a lot of
fibrous matter.
Flocculate after
the gross solid is
Float out
the gross
solid first.
may reduce
Sterilize after
flocculation so
that the
demand is low.
for floating/
. Corroding
for making
Electrolysis is demonstrably an economic way to purify some particularly noxious
types of water, with the added benefit of energy recovery in the form of hydrogen. By
using a specialised corroding anode (New Zealand Provisional Patent number
537700) and other proprietary water treatment apparatus including the Rutherford
Award winning Porous Ceramic Dual Media Filtration (PCDM) system, Works Filter
Systems Limited (New Zealand), and Works Filter Systems Pty (Australia) intend to
supply commercial electrolytic water treatment products.
G. Mathieson, A. Langdon and G. Jarnieson
The PhD project work on which this paper is based is supported by Technology New
Zealand and Works Filter Systems Limited. Special thanks to the Hamilton City
Council for the use of their wastewater processing sites for field trials.
Bockris, J.O.M. and A.K.N. Reddy, 1970. Modern Electrochemistry. Plenum Press, New York.
Bonilla, C.F.. 1947. Possibilities of the electronic coagulator for water treatment. Inter nndSewnge,
Chen, X . , G.Chen, et al., 2002. Investigation on the electrolysis voltage of electrocoagulation. Chem.
Eng. Sci'.,57(13), 2449-2455.
Coup, M.R. and A.G. Campbell, 1964. The Effect of Excessive Iron Intake upon the Health and
Production of Dairy Cows. NZ Journnl ofAgriculturnl Resenrch, 7(4), 624-638.
da Rosa, J.J. and 1. Rubio. 2005. The FF (flocculation-flotation) process. Miner. Eng.. 18(7), 701-707.
Fassina, L., 1938. Purification of tannery effluents. J. Am. Lenther Chem. Assoc., 33, 380.
Ihara, L,E. Shimada, et at., 2003. Treatment of landfill leachate. Absfmcts offnpers, 225th ACS
Nntionnl Meefing, New Orleans, LA, United States, March 23-27,2003, 1EC-265.
Khosla, N.K., S. Venkatachalam, et al., 1991. Pulsed electrogeneration of bubbles for i:lectroflotation.
J. Appl. Electrochem., 21(1 I), 986-90.
Lu, G.,J. Qu,et at., 1999. The electrochemical production of highly effective polyalurmnium chloride.
Wnter Research, 33(3), 807-813.
MATTER Project, The University of Liverpool, 2001. Alloy Composition Details.
Musquere, P., F. Ellingsen, et al., 1983. Electrotechnics in drinking and wastewater. Wnrer Supply,
8(Special Subject (2-3)), SS 8-1 to SS 8-25.
Ordonez, G.A., 1997. In situ electrolytic disinfection. Proceedings - water quality technology
conference, Costa Rica.
Osasa, K., H. Nakakura, et al., 1993. Treatment of colloidal waste material by electroflotation using
sacrificial electrodes. Kngnku Kogaku Ronbunshu, 19(2), 317-324.
Osipenko, V.D. and P.I. Pogorelyi, 1977. Electrocoagulation neutralization of Chromium Containing
effluent. Mernllurgist (English translnfion),21(9-10),44-45.
Pletcher, D. and F.C. Walsh, 1990. lndustriol Elecfrochemisfry,University Press, Cambridge.
Pontius, F.W., 1990. Wnkr Qualiw nnd Trentment - A hnndbook ofcommunity wntersupplies,
Riley, J . , 2003. Dynnmic Elecfrochemisfry.University of Bristol, Bristol, UK.
Robinson, V., 2000. A new technique for treatment of wastewater. Enviro 2000 Ozwn,rerOzwnste
Conference. Darling Harbour, Sydney, Australia.
Rohland, P., 1915. The action of electrolytes on eMuents containing organic waste. KislloidZeitschrifr, 16, 58-60.
Stuart, F.E., 1946. Electronic water purification; Progress report on the electronic coagulator - a new
device which gives promise of unusually speedy and effective results. Inter and Sewnge, 84(May),
Tao, Y., X.Zhang, et al., 1999. Electropurification of sulfur-containing sewage by electro-oxidizing
desulfurization. Elecfrochemicnl nnd Solid-State Letters, 2(3), 133-134.
Tetrault, A., 2003A. Electrocoagulation Treatment of Low Temperature Rendering P1,ant Wastewater
at an Australian Abattoir. EnvironzO3, NZWWA, Auckland, New Zealand.
Tetrault, A., 20038. Personal Communication. Environz03. NZWWA . Auckland, New Zealand: Cell
Voltage of EC Pacific Unit is about 50 V.
Vennekens, G . , 2000. Fuel Cell Catalyst Poisons - personal communication. G . Mathieson.
Wendt, H. and G. Kreysa, 1999. Elecfrochemicnl Engineering: Science nnd Technology in Chemicnl
nnd Other lndusrries, Springer.
Zeppenfeld, K . , 2003. Electrochemical production of aluminum hydroxide from alkaline aluminate
liquors. Aluminium (Isernhngen. Germnny), 79(9), 754-759.
Без категории
Размер файла
828 Кб
production, hydrogen, water, purification, electrolytic
Пожаловаться на содержимое документа