Efficient electrochemical reduction of nitrate to nitrogen on tin cathode at very high cathodic potentials

Article abstract of DOI:10.1016/j.electacta.2006.07.034

The electrochemical reduction of nitrate on tin cathode at very high cathodic potentials was studied in 0.1 M K₂SO₄, 0.05 M KNO₃ electrolyte. A high rate of nitrate reduction (0.206 mmol min^-1 cm^-2) and a high selectivity (%S) of nitrogen (92%) was obtained at -2.9 V versus Ag/AgCl. The main by-products were ammonia (8%) and nitrite (<0.02%). Small amounts of N₂O and traces of NO were also detected. As the cathodic potential increases, the %S of nitrogen increases, while that of ammonia displays a maximum at -2.2 V. The %S of nitrite decreases from 65% at -1.8 V to <0.02% at -2.4 V. The kinetic analysis indicated that the formation of nitrogen and ammonia proceeds through the intermediate nitrite. The reduction follows first order kinetics for both nitrate and nitrite at more cathodic potentials than -2.4 V, while at less negative potentials the kinetics is more complicated. The %Faradaic efficiency (%FE) of the reduction at -2.9 V was about 60% initially and decreased to 22% at 40 min. A cathodic corrosion of tin was observed, which was more intensive in the absence of nitrate. At potentials more negative than -2.4 V, small amounts of tin hydride were detected.

Full text of DOI:10.1016/j.electacta.2006.07.034

Electrochimica Acta 52 (2006) 1329–1338
Efficient electrochemical reduction of nitrate to nitrogen
on tin cathode at very high cathodic potentials
I. Katsounaros, D. Ipsakis, C. Polatides, G. Kyriacou∗
Department of Chemical Engineering, Laboratory of Inorganic Chemistry, Aristotle University of Thessaloniki, Box 462, Thessaloniki 541 24, Greece
Received 7 April 2006; received in revised form 30 June 2006; accepted 25 July 2006
Available online 6 September 2006
Abstract
The electrochemical reduction of nitrate on tin cathode at very high cathodic potentials was studied in 0.1 M K2SO4, 0.05 M KNO3 electrolyte.
A high rate of nitrate reduction (0.206 mmol min−1 cm−2) and a high selectivity (%S) of nitrogen (92%) was obtained at −2.9 V versus Ag/AgCl.
The main by-products were ammonia (8%) and nitrite (<0.02%). Small amounts of N2O and traces of NO were also detected.
As the cathodic potential increases, the %S of nitrogen increases, while that of ammonia displays a maximum at −2.2 V. The %S of nitrite
decreases from 65% at −1.8 V to <0.02% at −2.4 V. The kinetic analysis indicated that the formation of nitrogen and ammonia proceeds through
the intermediate nitrite.
The reduction follows first order kinetics for both nitrate and nitrite at more cathodic potentials than −2.4 V, while at less negative potentials the
kinetics is more complicated.
The %Faradaic efficiency (%FE) of the reduction at −2.9 V was about 60% initially and decreased to 22% at 40 min.
A cathodic corrosion of tin was observed, which was more intensive in the absence of nitrate. At potentials more negative than −2.4 V, small
amounts of tin hydride were detected.
© 2006 Elsevier Ltd. All rights reserved.
Keywords: Nitrate; Electrochemical; Reduction; Tin; Hydride; Nitrogen; Corrosion
1. Introduction
[3–24] and graphite [5,6,25,26]. It was proved that the conver-
sion of nitrate to the nontoxic nitrogen gas is difficult since it is
one of the eight possible products, NO2, NO2−, NO, N2O, N2,
NH2OH, NH3 and NH2NH2. When the supporting electrolyte is
basic or neutral, the main by-products are nitrite and ammonia;
furthermore, in acid electrolyte hydroxylamine and hydrazine
have been also detected [27]. Nitrogen gas was detected on vari-
ous metals and alloys, such as Cd (70–75%) [28], Pd–Cu (50%)
[7], Pb (22%) [4], Sn85–Cu15 (34%) [29], Pd–Sn–Au (34%)
[30], Sn–Pt (37%) [31] and Pd–Rh (94%) [32].
Nitrate-containing wastes can cause eutrophication of lakes,
rivers, bays and seas, while a high concentration of nitrate in
drinking water can cause serious health problems in humans.
The reduction of nitrate to nitrogen gas by other methods apart
from the biological, which cannot be applied in all cases, is a
challenge for the researchers in this field [1].
Electrochemical methods are ideal for removing nitrate
ions from concentrated aqueous solutions, that come out from
reverse osmosis, ion exchange membranes, neutralized nitric
acid wastes, as well as for solutions that contain other toxic com-
pounds, such as nuclear wastes, in which the biological method
cannot be applied [2].
According to Ferapontova and Fedorovich [33] the ra−te
of the reduction of −the anions of the second group (BrO4
,
BrO3−, NO3−, IO4 and others) is controlled by the simul-
taneous transfer of the electron and a proton from a pro-
z
A
ton donor through the reaction: Az + HBz + e− → HA
+
A
D
Numerous efforts have been made so far, concerning the elec-
trochemical reduction of nitrate on various metallic electrodes
Bz −1 (where HA = H2O, H3O+, NH4 and zi the charge). The
rate of the reduction depends on the cathodic potential, the struc-
ture of the double layer and the nature of the proton donor. High
reduction rates of the anions of the second group were achieved
at metals of high overpotential for hydrogen evolution at high
+
D
∗
Corresponding author. Fax: +30 2310 996196.
E-mail address: kyriakou@eng.auth.gr (G. Kyriacou).
0013-4686/$ – see front matter © 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2006.07.034

1330
I. Katsounaros et al. / Electrochimica Acta 52 (2006) 1329–1338
cathodic potentials. For example, efficient reduction of nitrate to
nitrogen on Pb [12] was carried out at cathodic potentials more
negative than −2 V, while that of bromate [34] up to −2 V versus
SCE. Even though tin alloys were studied for the reduction of
nitrate, in our knowledge, pure tin has not been studied at high
cathodic potentials.
This work includes preliminary experimental results on the
electrochemical reduction of nitrate on a tin cathode. Even
though at high cathodic potentials, where the rate of hydro-
gen evolution is high, ammonia is the predominant product
[4,27,29,32] the present work shows that at high cathodic poten-
tials, unexpectedly, themainproductontinisnitrogengashaving
both a high %S (92%) and a high rate of formation.
tions were prepared using ultra pure water from a Sation 9000
apparatus.
The tin foil (Sigma–Aldrich 99.9%) was polished with abra-
sive grade 600-mesh paper and washed quickly with ultra pure
water, before using it for the first time.
The composition of the catholyte was 0.1 M K2SO4 and
0.05 M KNO3, unless otherwise stated.
2.1. Analytical methods
At specific time intervals, samples of 0.1 mL were withdrawn
from the catholyte by a chromatography syringe and were ana-
lyzed, after the appropriate dilution. The determination of nitrate
and nitrite was performed by both standard colorimetric methods
[35] and ion chromatography (Dionex 4500i, AS9-HC column).
The determination of ammonia was performed by fluorometry
after its on-line derivatization with o-phthalaldehyde [36]. The
possible formation of hydrazine and hydroxylamine was inves-
tigated by colorimetric methods [37,38].
In the steady-state electrolytic experiments the solution was
degassed for about 30 min before applying the voltage by a
stream of He having a constant flow rate of 15 mL min−1. More-
over, the He stream withdrew the gaseous products from the cell
during the electrolysis. No significant volume loss was observed
in the catholyte due to the helium stream. The analysis of nitro-
gen and hydrogen was carried out by GC, using a molecular
sieve 5 A column 1.8 m, 0.3175 cm i.d., and a thermal conduc-
tivity detector (TCD), at a constant temperature of 80 ◦C. The
total volume of the produced nitrogen was calculated by the
equation: Vt = ʃCpR dt, where Vt is the total volume of the pro-
duced nitrogen, Cp the point concentration of nitrogen in the gas
stream and R is the flow rate of He.
2. Experimental
A Teflon cell (Fig. 1) having a total volume of 24 mL was used
in all experiments. A Nafion 117 (H+ form) cation exchange
membrane divided the cell into two equal volume compart-
ments. The heat transfer from the cell was achieved by two
aluminum plates (10 cm × 10 cm × 0.5 cm) and the experiments
were performed at ambient temperature. The compartment of the
catholyte had an extra volume of 2 mL on the top, in order to
secure that the electrolyte covered the whole surface of the cath-
ode, despite the volume removed due to the sampling. Moreover,
the cell had an extra free electrolyte volume of 2 mL on the top, to
prevent a blow-away of the electrolyte due to the Helium stream.
The tightness of the cell was achieved by Viton O-Rings. The
geometrical area of the cathode was 6 cm2 and the anode was a
platinized Pt foil (Alpha Metal) of equal area. The potential was
controlled by a Wenking POS 73 (Bank Elektronik) potentio-
stat and the reference was a saturated Ag/AgCl electrode. The
voltammograms were obtained under nitrogen atmosphere in a
conventional H type glass cell, using a tin foil having an appar-
ent geometrical area of 0.2 cm2, which was fitted in a Teflon
tube. All chemicals were reagent grade (Aldrich) and the solu-
In selected experiments, the gaseous products of the electrol-
ysis were collected into a sampling bag and the nitrogen oxides
(NO, NO2 and N2O) were analyzed by an Agilent Technologies
GC–MS.
Fig. 1. (a and b) The electrolysis cell: (P) aluminum plates, (S) silicone gasket, (M) nafion membrane, (D) dead volume, (RE) reference electrode. The distance
between the Luggin capillary and the cathode was about 1 mm.

I. Katsounaros et al. / Electrochimica Acta 52 (2006) 1329–1338
1331
The amount of ammonia which is presented in the experimen-
tal results is the sum of three parts: the catholyte, the anolyte as
well as the ammonia which was trapped by bubbling of the out-
going gas stream in a solution of 1N H2SO4 [29]. It should be
mentioned that up to 30% of the produced ammonia escaped
from the catholyte due to the He stream.
The %removal efficiency (%RE) of nitrate and the %selec-
tivity (%S) of the products were calculated by the equations:
no − n
%RE =
100%
no
ni
%Si = p
100%
no − n
Fig. 3. Slow-sweep voltammograms at a scan rate of 5 mV s−1 of solutions
containing various concentrations of nitrate in 0.1 M K2SO4.
where no is the initial mols of nitrate; n is the remaining mols
of nitrate; ni is the produced mols of compound i; p is mols of
nitrate needed for the formation of 1 mol of compound i (p = 1
for nitrite, ammonia and p = 2 for nitrogen).
Slow-sweepvoltammogramswitha5 mV/sscanratebetween
−1.0 and −2.9 V were taken to determine the electrokinetic
behaviour of nitrate. On the potential sweep in the negative
direction the current increases gradually from −1.6 V and this
increase depends on the concentration of nitrate (Fig. 3). The
current density depends on both the potential and the con-
centration of nitrate and that implies that the reaction is not
diffusion-controlled, which is in accordance with the results of
the steady-state electrolytic experiments. From the slope of the
diagram of log j versus log C (Fig. 4) the reaction order was cal-
culated to be 0.63, 0.78, 0.91 and 0.93 at −2.0, −2.3, −2.7 and
−2.9 V, respectively. A fractional reaction order was reported
previously by other workers on Pt–Pd–Ge [22], Pd–Sn–Au [30]
and Cu–Tl [42] and it was explained assuming a Temkin adsorp-
tion [5,27]. Recently, Casella and Contursi [43] reported a value
of 1.25 for the reaction order on Pd–Sn in acid medium.
3. Results
Fig. 2 shows the cyclic voltammograms obtained for the
reduction of nitrate and nitrite. In both cases a reversible peak
was observed at −1.2 and −1.15 V. This peak was not observed
in the voltammogram of the supporting electrolyte at 50 mV/s,
but it was observable at scan rates higher than 100 mV/s. The
peak was attributed to the oxidation/reduction of Sn2+ to Sn4+
(SnO2 and/or Sn(OH)4) species [39–41]. The increase of the
peak current in the presence of nitrate and nitrite was attributed
to the increased corrosion of tin, because they can be reduced,
facilitating the cathodic reaction [41]. Moreover, an increase in
the current density (cd) from −1.4 V in the case of nitrite and
from−1.65 Vinthecaseofnitratewasobservedanditisremark-
able that the cd of nitrite reduction is much higher than that of
nitrate at any potential. The more negative potential required for
the reduction of nitrate was attributed to the resonance structure
of nitrate, which makes it more stable than nitrite.
3.1. Constant potential electrolysis
The reproducibility of the batch electrolytic experiments was
tested by carrying out several experiments at −2.9 V under the
same conditions. An increase in the apparent rate constant of
the nitrate reduction from 0.052 to 0.102 min−1 was observed
by using the same electrode for three consecutive experiments
Fig. 2. Cyclic voltammograms in (1) 0.1 M K2SO4, (2) 0.1 M K2SO4/0.05 M
KNO3 and (3) 0.1 M K2SO4/0.05 M KNO2 electrolyte, at a scan rate of
Fig. 4. Logarithm of current density (in A m−2) against the logarithm of con-
centration (in mol L−1) at various cathodic potentials.
50 mV s−1
.

1332
I. Katsounaros et al. / Electrochimica Acta 52 (2006) 1329–1338
Table 1
%RE of nitrate and %S of the products in three different experiments under the same conditions
−
−
Experiment
NO3 (%RE)
NO2 (%S)
NH3 (%S)
N2 (%S)
Apparent rate constant (min−1
)
First
99.48
99.28
99.78
99.64
0.86
<0.02
<0.02
<0.02
8.62
9.11
8.10
7.6
85.54
86.01
91.75
93.80
0.052
0.077
0.102
0.104
Second
Third
Final
Potential −2.9 V vs. Ag/AgCl, electrolyte 0.1 M K2SO4/0.05 M KNO3 and electrolysis time 2.5 h.
and it remained constant within the experimental error, when
more experiments were performed (Table 1). The apparent rate
constant was calculated assuming first order kinetics for the
reduction of nitrate at this potential, as it was approximately
estimated from Fig. 4 and confirmed by the electrolytic experi-
ments. An experiment (final experiment in Table 1), which was
performed by an electrode that had been previously used for
the reduction of nitrate at all potentials mentioned in the fol-
lowing text, indicated no significant differences, as regards the
rate and the %S of the products, in comparison to that of the
third experiment. The reduction of nitrate took place rapidly
and the %RE was more than 99% in all experiments performed
at −2.9 V. An increase by a factor of 1.5 was also observed
in the total electrolysis current between the first and the third
experiment. The simultaneous increase in both the reaction rate
and the electrolysis current was attributed to the increase of the
specific area of the electrode, which was visible to the naked
eye. Furthermore, the electrode surface darkened and lost its
metallic lustre, while small particles of tin appeared in the
catholyte. Fig. 5(a and b) shows the SEM photographs of (a)
an unused and (b) an extensively used tin electrode. The great
harshness of the surface of the used electrode compared with the
smooth surface of the unused, leads to an increase of its specific
area.
to the cell for 150 min. The gaseous products of the cell were
passed through a glass tube containing 15 mL of 0.01 M silver
nitrate solution, which is a trapping agent of tin hydride [48,51].
At specific time intervals, samples were withdrawn from the
catholyte and were analyzed by atomic absorption in a graphite
furnace. The analysis of the trapping solution after electrolysis
for 180 min indicated that the concentration of tin in the presence
of nitrate was 3.5 ppm, while in the absence of nitrate 4.5 ppm.
It is worth to mention that tin was not detected in the trapping
solution when the electrolysis was carried out at potentials less
negative than −2.4 V. The presence of tin in the trapping solution
is a strong indication that the formation of gaseous tin hydride
takes place at potentials more negative than −2.4 V, contributing
to the corrosion of tin, to a small extent.
According to the Pourbaix diagram [44] at high cathodic
potentials, tin can be converted to gaseous SnH4 through the
reaction: Sn + 4H+ + 4e− → SnH4. The equilibrium potential at
pH = 0 is −1.074 V versus NHE. The potential depends upon the
pH and the logarithm of the partial pressure of SnH4, according
to the Nernst equation:
E0 = −1.074 − 0.059pH − 0.0148 log pSnH4 versus NHE
At pH > 7, the potential is calculated to be more negative than
−1.487 V versus NHE or −1.709 V versus Ag/AgCl, assuming
that the influence of the term log pSnH4 is negligible. Given
that our experiments were performed at much higher cathodic
potentials, the formation of SnH4 is possible. There are many
references on the formation of tin hydride on various cathodes
[45,46], but its formation on tin cathode was doubtable up to
1977 [47], since there was only one reference [48] where the
cathodic corrosion of tin was attributed to the formation of SnH4.
Later studies speculated the possible formation of SnH4 [49,50].
In order to investigate the observed cathodic corrosion of tin,
someexperimentswereperformedunderconditionscorrespond-
ing to that of the batch electrolysis; the electrode was left at open
circuit conditions for 30 min (as it happens during degassing
by the He) and afterwards a potential of −2.9 V was applied
Fig. 5. SEM photograph of (a) an unused and (b) an extensively used for elec-
trolysis tin cathode.

I. Katsounaros et al. / Electrochimica Acta 52 (2006) 1329–1338
1333
assumptions (a) and (c) are not in contradiction since the final
result is the formation of tin hydrides and metallic tin.
According to Salzberg and Mies [48] a small concentration of
ammonium in the catholyte limits the corrosion of tin. After that
we performed an experiment in 0.1 M K2SO4, 0.05 M KNO3
and 5 × 10−4 M (NH4)2SO4. The experimental results showed
that the concentration of tin in the catholyte was by 20 times
lower than that without ammonium, as it is depicted in Fig. 6.
Since the specific area of the tin electrode that was used for
the first time, is continuously increased during the electrolysis,
we preferred to perform all the reported experiments in the fol-
lowing text using a tin cathode which had been already used for
the reduction of nitrate (50 mM) at −2.9 V for three consecu-
tive experiments. In this case, as we pointed out, the roughness
factor of the electrode remains almost constant, even though the
corrosion of tin is still taking place.
Fig. 6. Concentration of tin in the catholyte vs. time in (ꢀ) 0.1 M K2SO4, (ꢀ)
0.1 M K2SO4/0.05 M KNO3 and (ꢁ) 0.1 M K2SO4/0.05 M KNO3 + 5 × 10−4 M
(NH4)2SO4 at −2.9 V vs. Ag/AgCl.
3.2. Reaction kinetics and influence of the potential
The concentration profile of nitrate versus time (Fig. 7)
revealed that the reaction follows first order kinetics at −2.5 V,
since the graph of the logarithm C/Co versus time (not depicted
here), which is characteristic for first order reaction kinetics, is
linear, with a correlation coefficient R2 = 0.975. On the other
hand, at −1.8 V more complicated kinetics is involved, which
is in agreement with the slow-sweeping rate voltammetric mea-
surements (Fig. 4). Moreover, when the electrolysis was per-
formed at −2.5 V in the presence of 25 mM KNO2 the apparent
rate constant was 0.059 min−1, while in the absence of nitrite
was 0.061 min−1, which denotes that the reduction of nitrate is
not inhibited in the presence of nitrite, under these conditions.
The variation of the concentrations of the initial nitrate and
the main products of electrolysis versus time is depicted in Fig. 8.
The concentration of nitrite displays a maximum at 10 min,
which is an indication that it is an intermediate product of a con-
secutive reactions mechanism, as it is previously proposed by
other workers [4,29]. The concentrations of N2 and NH3 display
Fig. 6 shows that the concentration of tin versus time in
the catholyte reaches a maximum after 60 min and afterwards
it remains approximately constant, when the solution contains
nitrate. A significantly higher concentration of tin was detected
in the absence of nitrate, which was gradually decreased from
180 ppm at 60 min to 100 ppm at 180 min. It should be noted
that, in this case, the concentration of tin in the trapping solution
was also higher, as it is noticed above. The decreased corrosion
of tin in the presence of nitrate was attributed to the decreased
coverage of the surface of the tin cathode by adsorbed hydrogen
due to the co-adsorption of nitrate.
Three different explanations have been given so far for the
cathodic disintegration of tin: (a) formation of tin hydrides
(SnH2 and SnH4). According to Salzberg and Mies [48], tin
under cathodic polarization forms initially SnH2, which is unsta-
ble and decomposes to metallic tin and SnH4 via a disproportion-
ation reaction: 2SnH2 → Sn + SnH4. The rate of disintegration
is inversely proportional to the concentration of the support-
ing electrolyte and depends also on the nature of the cation in
the order Na+ > K+ > Li+ ꢁ NH4+. (b) Cathodic incorporation
of alkali metals [52–56], which usually occurs at potentials by
1.4 V more positive than the potential of the deposition of each
alkali metal and leads to the formation of intermetallic com-
pounds between the tin cathode and the metal. The hydrogen
evolution takes place via the discharge of alkali–metal after its
chemical reaction with water, as follows [56]:
K(Sn) + H2O → Sn + Hads + K+ + OH−
The change of the mechanism of the conventional hydrogen
evolution reaction, leads to an increase of the overpotential of
hydrogen by about 0.25 V. According to this hypothesis the rate
of disintegration is expected to be proportional to the concentra-
tion of the supporting electrolyte. The latter is in contradiction
with the first assumption of hydride formation. (c) Tin under
4−
cathodic polarization gives tin polyanions of the type Sn9
,
Fig. 7. Concentration profile of nitrate vs. time at: (ꢀ) −1.8 V in 0.1 M
K2SO4/0.05 M KNO3, (ꢀ) at −2.5 V in 0.1 M K2SO4/0.05 M KNO3 and (ꢁ)
at −2.5 V in 0.1 M K2SO4/0.05 M KNO3 + 0.025 M KNO2.
which consequently react with the surrounded water molecules
to give hydrogen gas or hydrides and metallic tin [57]. The

1334
I. Katsounaros et al. / Electrochimica Acta 52 (2006) 1329–1338
(c) Two parallel redox reactions:
kf1
kf2
A + n1eꢃB, A + n2eꢃR
kb1
kb2
In the case of nitrate reduction, the mechanism is a combination
of the above mentioned systems.
The reduction of nitrate to nitrite can be described by the
scheme (a), namely:
kf1
−
NO3− + n1eꢃNO2
kb1
Since the reactions take place at high negative potentials, the
term kb1, which is the rate constant for the oxidation of nitrite
to nitrate, is negligible.
Assuming first order kinetics, the reaction rate can be
expressed as:
Fig. 8. Variation of the concentration of (ꢀ) nitrate, (ꢀ) nitrite, (ꢂ) ammonia
and(ꢂ)nitrogenvs. electrolysistimeat−2.5 V. Electrolyte0.1 MK2SO4/0.05 M
KNO3. The lines were drawn by non-linear regression based on the proposed
model.
i1 = n1Fkf1Cs
(1)
−
NO3
where the rate constant kf1 can be defined as an exponential
function of the cathodic potential, and CNs O is the surface con-
−
a sigmoidal form, which is characteristic for the final product of
the above mentioned mechanism. The rate of production of both
products increases simultaneously and this implies that these
species are produced through a common intermediate. It should
be noted that hydroxylamine and hydrazine were not detected,
but the GC–MS analysis of the gaseous products indicated small
amounts of N2O and traces of NO.
In order to verify that nitrite is an intermediate for the for-
mation of the final products, the reduction of 0.05 M KNO2 was
carried out at −2.5 V. The reduction of nitrite displays the same
characteristics as those obtained for nitrate, since the main prod-
ucts were ammonia and nitrogen. The reduction follows first
order kinetics, as in the case of nitrate.
3
centration of nitrate.
By taking into account the mass transfer, the current for the
transport of nitrate from the bulk to the surface is:
b
s
−
i1 = n1Fk
[CNO − CNO
]
−
(2)
−
L,NO3
3
3
−
where kL,NO is the mass transfer coefficient for nitrate and
3
CNb O the concentration of nitrate in the bulk solution.
−
T3he surface concentration CNs O can be eliminated, by com-
−
3
bining Eqs. (1) and (2), so the final expression for the partial
current density of the reduction of nitrate to nitrite is:
n1Fkf1
−
i1 =
[NO3
]
(3)
−
1 + (kf1/k
)
According to the above, a consecutive reaction mechanism
can describe the experimental results on tin at potentials more
negative than −2.5 V, including first order kinetics for both
nitrate and nitrite, as follows:
L,NO3
The reduction of nitrite to either nitrogen or ammonia is
described by Scott’s scheme (c) as “Two parallel redox reac-
tions”.
k
1
NO3−−→ NO−2
kf2
NO2−−→ NH3
k
2
NO2−−→ NH3
kf3
2NO2−−→ N2
k
3
2NO2−−→ N2
The rate constants kb given by Scott for the backward reactions
were also assumed to be negligible.
In terms of kinetics and mass transfer parameters, the expres-
sions for the partial current densities of the ammonia and nitro-
gen are:
Scott [58] published a paper describing the behaviour of some
electrochemical reaction schemes, where the thermodynamics,
the kinetics as well as the mass transfer were taken into account.
Scott calculated the mathematical expressions for three cases:
i2 = n2Fkf2Cs
(4)
−
NO2
(a) Redox reaction:
1
i3 = n3Fkf3Cs
(5)
−
kf1
A + n1eꢃB
kb1
NO2
2
i2
i3
b
s
−
+
= kL,NO (CNO − CNO
)
−
(6)
−
2
2
2
n2F
n3F
(b) Two redox reactions in series:
−
where kL,NO is the mass transfer coefficient for nitrite and
kf1
kf2
2s
CNb O and CNO are the concentrations of nitrite in the bulk
and on the surface, respectively.
−
−
A + n1eꢃB + n2eꢃE
2
2
kb1
kb2

I. Katsounaros et al. / Electrochimica Acta 52 (2006) 1329–1338
1335
The surface concentration of nitrite CNs O can be eliminated,
by combining Eqs. (4)–(6), so the expressions for the partial
−
kf3/Y
2
k3 =
[1 − (kf2kf3/k2
)]XY
−
L,NO2
current densities can be written as:
Finally, the differential equations are:
−
−
(kf2/Xk
)[kL,NO − (kf3/Y)]
L,NO2
−
2
i2 = n2F
[NO2
]
(7)
(8)
−
[1 − (kf2kf3/k2
)]XY
d[NO3
]
]
−
−
L,NO2
= −k1[NO3
]
dt
d[NO2
−
1
i3 = n3F
2
kf3/Y
[1 − (kf2kf3/k2
−
= +k1[NO3−] − k2[NO2−] − k3[NO2
]
−
[NO2
]
)]XY
−
dt
L,NO2
d[NH3]
−
= +k2[NO2
]
]
where:
dt
kf2
kf3
d[N2]
1
−
X = 1 +
,
Y = 1 +
= + k3[NO2
2
−
−
dt
k
k
L,NO2
L,NO2
The solutions of the differential equations using Laplace
transforms give the concentrations versus time:
−
[NO3−] = [NO3
]
e
−k t
1
initial
e−(k +k )t − e−k t
2
3
1
[NO2−] = k1[NO3
]
initial
−
k1 − k2 − k3
ꢀ
ꢁ
ꢁ
1
k1 e−(k +k )t − (k2 + k3) e−k t
k2 + k3 − k1
2
3
1
−
−
[NH3] = k2k1[NO3
]
]
1 +
initial k1(k2 + k3)
ꢀ
k1 e−(k +k )t − (k2 + k3) e−k t
k2 + k3 − k1
2
3
1
1
1
[N2] = k3k1[NO3
1 +
2
initial k1(k2 + k3)
The non-linear regression of the experimental results according
to the above equations gives k1 = 0.059 min−1, k2 = 0.089 min−1
and k3 = 0.145 min−1, at −2.5 V. The obtained values from the
proposed model fit well to the experimental data, as it is shown
in Fig. 8, where the curves depict the predictions of the model.
The apparent rate constant k1 increases as the potential
becomes more negative even at very high cathodic values, as
it is shown in Table 2, which verifies that the reduction of nitrate
to nitrite is not limited by the mass transfer, which is in accor-
dance with the voltammetric curves (Fig. 2).
Fig. 9 shows the %FE versus time at three different potentials.
The %FE was calculated assuming 2, 8 and 10 electrons for the
formation of nitrite, ammonia and nitrogen, respectively. At the
first stages of electrolysis, the %FE is high in all cases (60–75%)
and consequently decreases. The %FE at 40 min, where the %RE
of nitrate is 99% at −2.9 V, is 22%. The decrease of the %FE, as
the concentration of nitrate decreases, leads to a corresponding
increase in the %FE of hydrogen, which was also confirmed by
the GC analysis of the hydrogen in the outgoing gas stream.
The low rate of hydrogen evolution at the first electrolysis steps,
where the concentration of nitrate is high, leads to the conclusion
thattheevolutionofhydrogenontincathodeisstronglyinhibited
in the presence of nitrate.
If the partial current densities of Eqs. (3), (7) and (8) are
expressed in terms of reaction rate, the differential equations
are:
−
d[NO3
]
]
kf1
−
−
= −
[NO3
[NO3
]
]
−
−
dt
1 + (kf1/k
)
L,NO3
−
d[NO2
kf1
= +
−
dt
1 + (kf1/k
)
L,NO3
−
−
(kf2/Xk
)[kL,NO − (kf3/Y)]
L,NO2
−
2
[NO2
]
[1 − (kf2kf3/k2
)]XY
−
L,NO2
kf3/Y
−
[NO2 ]
−
[1 − (kf2kf3/k2
)]XY
−
L,NO2
−
−
[kf2/(Xk
)][kL,NO − (kf3/Y)]
d[NH3]
L,NO2
−
[NO2 ]
2
= +
[1 − (kf2kf3/k2
)]XY
dt
−
L,NO2
d[N2]
1
2
kf3/Y
−
[NO2 ]
= +
2
dt
[1 − (kf2kf3/k
)]XY
−
L,NO2
The terms before the concentrations are constant and they
can be represented as apparent rate constants, namely:
Table 2
Apparent rate constant of the reduction of nitrate vs. potential
Potential (V)
Apparent rate constant (min−1
)
kf1
k1 =
−2.3
−2.4
−2.5
−2.7
−2.9
0.0335
0.0532
0.0614
0.0787
0.1041
−
1 + (kf1/k
)
L,NO3
−
−
(kf2/Xk
)[kL,NO − (kf3/Y)]
L,NO2
2
k2 =
[1 − (kf2kf3/k2
)]XY
−
L,NO2

1336
I. Katsounaros et al. / Electrochimica Acta 52 (2006) 1329–1338
In order to clarify the influence of the change of the tempera-
ture and the pH, two additional experiments were performed (at
−2.9 V), using an electrode having an area of 2 cm2 instead of
6 cm2 (in order to have a lower total current) and the cell was
placed in a thermostated water bath at 25 ◦C. In the first exper-
iment, the supporting electrolyte was 0.1 M K2SO4 + 0.05 M
KNO3 and in the second, a buffer containing 0.1 M KHCO3 and
0.1 M K2CO3 + 0.05 M KNO3. No significant differences were
observed between the two experiments, within the experimental
error, since the %RE of nitrate was 95.9% in the unbuffered and
−
97.1% in the buffer, while the %S of NH3, N2 and NO2 were
7.0, 86.9 and 0.5% in the first experiment and 10.5, 84.9 and
0.36% in the second, respectively. These findings are reason-
able since in both cases, in this pH region, the water molecule
is the proton donor [33].
4. Discussion
Fig. 9. %Faradaic efficiency of the reduction of nitrate against electrolysis time
at (ꢁ) −1.8 V, (ꢀ) −2.5 V and (ꢀ) −2.9 V. Electrolyte 0.1 M K2SO4/0.05 M
KNO3.
The mechanism of the nitrate reduction is complicated,
because of its multi-electron nature and the presence of a large
number of intermediates. The presented data are insufficient to
obtain the detailed mechanism of the reduction. At this point,
we shall summarize the main experimental findings which are
useful for the further discussion: (a) the reduction of both nitrate
and nitrite at high negative potentials, follows first order kinet-
ics, (b) the hydrogen evolution on tin is inhibited in the presence
of nitrate as it is concluded from Fig. 9, (c) at cathodic potentials
more negative than −2.4 V tin is converted to tin hydride to a
small extent, (d) ammonia and nitrogen are produced through
a common intermediate and (e) at potentials between −2.0 and
−2.2 V ammonia is the main product, while at more negative
potentials the %S of ammonia decreases and that of nitrogen
displays a corresponding increase.
Based on the fact that the reduction follows first order kinet-
ics for both nitrate and nitrite, a second intermediate must be
involved, which is the precursor of ammonia and nitrogen for-
mation. This intermediate is probably the adsorbed nitrogen
monoxide, which was detected in our experiments and it has
been already proposed by other authors [7,22,59] so the reac-
tion scheme can be the following:
The influence of the cathodic potential on the product distri-
bution and the %RE of nitrate is shown in Fig. 10. The %RE
increases gradually and converges at 100% at more cathodic
potentials than −2.5 V. The %S of nitrogen displays an almost
linear increase between −1.8 V (15%) and −2.9 V (91.8%),
while that of ammonia, after a maximum at −2.2 V (63.9%),
decreases to 8.1% at −2.9 V. The form of the curve of ammonia
is unexpected, since, as it is well known from previous works
[4,27,29,32] the formation of ammonia, which is a hydrogenated
product of the nitrate reduction, is favoured at high cathodic
potentials. At potentials more negative than −2.4 V the sum
of the %S of ammonia and nitrogen remains constant within
the experimental error and this is an additional indication that
nitrogen and ammonia are produced through a common interme-
diate. The %S of nitrite decreases from 65% at −1.8 V to almost
0.5% at −2.4 V and then converges at <0.02% at −2.9 V. How-
ever, when the reduction of nitrate takes place in an unbuffered
electrolyte at current densities up to 120 mA cm−2 (at −2.9 V)
a significant increase in the pH in the vicinity of the cath-
ode is expected. In addition, an increase in the temperature of
about 10 ◦C was observed in the cell due to the high current.
−
−
NO3 → NO2 → [NO]
2[NO] → N2
[NO] + 5[H] → NH3 + H2O
Sn + 4[H] → SnH4
According to the scheme, nitrogen and ammonia are formed
through the surface adsorbed [NO]. At potentials between −1.8
and −2.2 V the %S of ammonia increases as a result of the
increase in the surface concentration of [H] since ammonia is
a hydrogenated product. At more negative potentials the sur-
face concentration of [H] is limited due to the formation of the
hydrides of tin (SnH2 and SnH4) [48], leading to an increase in
the %S of nitrogen.
Fig. 10. Influence of the cathodic potential on the distribution of the main prod-
ucts of the reduction of nitrate. The electrolyte was 0.1 M K2SO4/0.05 M KNO3
and the electrolysis time was 2.5 h.
If the incorporation of potassium into the tin cathode is
assumed, we can arrive at similar conclusions, since the incorpo-

I. Katsounaros et al. / Electrochimica Acta 52 (2006) 1329–1338
1337
ration also limits the hydrogen evolution reaction [54], leading
to a decreased %S of ammonia. Moreover, the fact that nitro-
gen was not produced in 0.1 M NH4Cl + 0.05 M NH4NO3 or in
0.1 M (NH4)2CO3 + 0.05 M NH4NO3 electrolyte at −2.9 V and
the rate of the reduction of nitrate was much lower than that in a
solution containing potassium under the same conditions shows
that the presence of potassium plays an important role in the
mechanism of the reduction as it was reported by Kabanov et al.
[60]. The two phenomena (hydride formation and incorporation)
may take place simultaneously and further work is required to
define the degree of participation of each one in the corrosion of
tinandinthemechanismofthereductionofnitrate. Furthermore,
the ions of Sn2+ which are formed in the solution due to the cor-
rosion of tin can play an important role in the overall process as it
has been proposed by Safonova and Petrii [61] on Pt electrodes.
It is well known that the rate of the reduction of anions is
strongly dependent on the nature of the cation and the con-
centration of the supporting electrolyte. This phenomenon was
attributed initially to double layer effects by Frumkin et al. [62].
Later studies introduced the term “cationic catalysis” in order to
extend this theory by taking into account the participation of ion
pairs or bridge assisted electron transfer between the non react-
ing cation of the supporting electrolyte (K+ or NH4+ in our case)
and the reacting anion, especially in the interfacial reaction layer
[63]. The formation of ion pairs possibly explains why nitrate
anions are not repelled by the negatively charged electrode even
at very high cathodic potentials. Further work is needed to clarify
the role of the supporting electrolyte in the reduction of nitrate.
The corrosion of tin seems to be the most serious problem
of the whole process, since the concentration of tin in the bulk
solutionishigherthantheenvironmentallyacceptablelevel. This
problem could be treated further by applying a less negative
potential after the removal of nitrate, in order to redeposit the
dissolved tin. Otherwise, according to Salzberg and Mies [48],
the presence of small amounts of some cations, such as cesium
or rubidium in the catholyte can stop the corrosion at current
densities up to 2.7 A/cm2.
[3] H. Li, D.H. Robertson, J.Q. Chambers, D.T. Hobbs, J. Electrochem. Soc.
135 (1988) 1154.
[4] J.O’M. Bockris, J. Kim, J. Appl. Electrochem. 27 (1997) 623.
[5] J.O’M. Bockris, J. Kim, J. Electrochem. Soc. 143 (1996) 3801.
[6] K. Bouzek, M. Paidar, A. Sadilkova, H. Bergmann, J. Appl. Electrochem.
31 (2001) 1185.
[7] A.C.A. de Vooys, R.A. van Santen, J.A.R. van Veen, J. Mol. Catal. A 154
(2000) 203.
[8] S. Ureta-Zanartu, C. Yanez, Electrochim. Acta 42 (1997) 1725.
[9] G. Horanyi, E.M. Rizmayer, J. Electroanal. Chem. 143 (1983) 323.
[10] G. Horanyi, E.M. Rizmayer, J. Electroanal. Chem. 188 (1985) 265.
[11] N. Chebotareva, T. Nyokong, J. Appl. Electrochem. 27 (1997) 975.
[12] H. Li, J.Q. Chambers, D.T. Hobbs, J. Appl. Electrochem. 18 (1988) 454.
[13] J.D. Genders, D. Hartsough, D. Hobbs, J. Appl. Electrochem. 26 (1996) 1.
[14] G.E. Dima, A.C.A. de Vooys, M.T.M. Koper, J. Electroanal. Chem.
554–555 (2003) 15.
[15] J.F.E. Gootzen, L. Lefferts, J.A.R. van Veen, Appl. Catal. A 188 (1999)
127.
[16] M.S. El-Deab, Electrochim. Acta 49 (2004) 1639.
[17] M. Fedurco, P. Kedzierzawski, J. Augustynski, J. Electrochem. Soc. 146
(1999) 2569.
[18] S. Kerkeni, E. Lamy-Pitara, J. Barbier, Catal. Today 75 (2002) 35.
[19] H.A. Duarte, K. Jha, J.W. Weidner, J. Appl. Electrochem. 28 (1998) 811.
[20] G. Horanyi, E.M. Rizmayer, J. Electroanal. Chem. 140 (1982) 347.
[21] O.A. Petrii, T.Y. Safonova, J. Electroanal. Chem. 331 (1992) 897.
[22] J.F.E. Gootzen, P.G.J.M. Peeters, J.M.B. Dukers, L. Lefferts, W. Visscher,
J.A.R. van Veen, J. Electroanal. Chem. 434 (1997) 171.
[23] L.J.J. Janssen, M.M.J. Pieterse, E. Barendrecht, Electrochim. Acta 22
(1977) 27.
[24] M.C.P.M. da Cunha, J.P.I. de Souza, F.C. Nart, Langmuir 16 (2000) 771.
[25] O.W.J.S. Rutten, A. van Sandwijk, G. van Weert, J. Appl. Electrochem. 29
(1999) 87.
[26] D. De, E.E. Kalu, P.P. Tarjan, J.D. Englehardt, Chem. Eng. Technol. 27
(2004) 1.
[27] D. De, J.D. Englehardt, E.E. Kalu, J. Electrochem. Soc. 147 (2000)
4573.
[28] F.J. Witt, Electrolytic Recycle Method for the Treatment of Radioactive
Nitric Acid Waste; Report ORNL-2532, Oak Ridge National Laboratory,
Oak Ridge, TN; September 19, 1958.
[29] C. Polatides, G. Kyriacou, J. Appl. Electrochem. 35 (2005) 421.
[30] K. Tada, T. Kawaguchi, K. Shimazu, J. Electroanal. Chem. 572 (2004) 93.
[31] K. Tada, K. Shimazu, J. Electroanal. Chem. 577 (2005) 303.
[32] H. Cheng, K. Scott, P.A. Christensen, Chem. Eng. J. 108 (2005) 257.
[33] E.E. Ferapontova, N.V. Fedorovich, J. Electroanal. Chem. 476 (1999) 26.
[34] B.B. Damaskin, N.V. Fedorovich, F.S. Sarbash, J. Electroanal. Chem. 135
(1982) 181.
[35] L.S. Clesceri, A.E. Greenberg, R.R. Trussell, Standard Methods for the
Examination of Water and Wastewater, 17th ed., American Public Health
Association, Washington, DC, 1989, p. 4.
In conclusion, tin at high potentials is the most effective cath-
ode, which has been reported in the literature so far for the
electrochemical reduction of nitrate to nitrogen, since it com-
bines both high %S for nitrogen and high rate of reduction, which
could result to a low value of space-time in a flow-cell, namely to
a small volume of the cathodic part of the reactor. The remaining
concentrations at −2.9 V of both nitrate (6.8 ppm) and nitrite
(<0.5 ppm) are under the maximum contaminant level (MCL)
in wastewaters and potable water [64]. The concentration of the
produced ammonia (68 ppm) is over the MCL, but this could be
treated thereafter by using a paired electrolysis reactor, where
the catholyte would be passed through the anodic compartment
of the cell to convert the ammonia via its oxidation to either
nitrate or nitrogen [65].
[36] S.S. Goyal, D.W. Rains, R.C. Huffaker, Anal. Chem. 60 (1988) 175.
[37] G.W. Watt, J.D. Chrisp, Anal. Chem. 24 (1952) 2006.
[38] F. Dias, A.S. Olojola, B. Jaselskis, Talanta 26 (1979) 47.
[39] S.S. Abd El-Rehim, F. Taha, M.B. Saleh, S.A. Mohamed, Corros. Sci. 33
(1992) 1789.
[40] S.S. Abd El-Rehim, S.M. Sayyah, M.M. El Deeb, Mater. Chem. Phys. 80
(2003) 696.
[41] S.A.M. Refaey, S.S. Abd El-Rehim, Electrochim. Acta 42 (1996) 667.
[42] I.G. Casella, M. Gatta, J. Electroanal. Chem. 568 (2004) 183.
[43] I.G. Casella, M. Contursi, J. Electroanal. Chem. 588 (2006) 147.
[44] A.J. Bard, R. Parsons, J. Jordan, Standard Potentials in Aqueous Solutions,
IUPAC (Marcel Dekker), New York, USA, 1985.
[45] E. Denkhaus, F. Beck, P. Bueschler, R. Gerhard, A. Golloch, Fresenius J.
Anal. Chem. 370 (2001) 7353.
References
[46] M. Kliskic, J. Radosevic, S. Gudic, Electrochim. Acta 48 (2003) 4167.
[47] B.N. Stirrup, N.A. Hampson, Surf. Technol. 5 (1977) 429.
[48] H.W. Salzberg, F. Mies, J. Electrochem. Soc. 105 (1958) 64.
[49] I.A. Ammar, S. Darwish, M.W. Khalil, S. El-Taher, Mater. Chem. Phys. 21
(1989) 1.
[1] A.M. Fan, V.E. Steinberg, Regul. Toxicol. Pharmacol. 23 (1996) 35.
[2] M. Paidar, K. Bouzek, L. Jelinek, Z. Matejka, WaterEnviron. Res. 76(2004)
2691.

1338
I. Katsounaros et al. / Electrochimica Acta 52 (2006) 1329–1338
[50] A.I. Marshakov, N.P. Chebotareva, Zashchita Metallov 28 (1992) 902.
[51] A.C. Sahayam, S. Natarajan, S. Gangadharan, Fresenius J. Anal. Chem.
346 (1993) 961.
[52] F. Haber, M. Sack, Z. Elektrochem. 8 (1902) 245.
[53] B.N. Kabanov, Electrochim. Acta 13 (1968) 19.
[54] B.N. Kabanov, I.I. Astakhov, I.G. Kiseleva, Electrochim. Acta 24 (1979)
167.
[59] S. Wasmus, E.J. Vasini, M. Krausa, H.T. Mishima, W. Vielstich, Elec-
trochim. Acta 39 (1994) 23.
[60] B.N. Kabanov, N.N. Tomashova, I.G. Kiseleva, Elektrokhimiya 6 (1970)
612.
[61] T.Y. Safonova, O.A. Petrii, Russ. J. Electrochem. 34 (1998) 1137.
[62] A.N. Frumkin, N.V. Nikolaeva-Fedorovich, N.P. Berezina, Kh.E. Keis, J.
Electroanal. Chem. 58 (1975) 189.
[55] I.G. Kiseleva, I.A. Avrutskaya, N.N. Tomashova, M.E. Niyazimbetov, M.Y.
Fioshin, B.N. Kabanov, Sov. Electrochem. 12 (1976) 859.
[56] A. Frumkin, V. Korshunov, I. Bagozkaya, Electrochim. Acta 15 (1970) 289.
[57] V.P. Gladyshev, J. Appl. Chem. USSR 48 (1975) 540.
[58] K. Scott, Electrochim. Acta 30 (1985) 235.
[63] R.R. Nazmutdinov, D.V. Glukhov, G.A. Tsirlina, O.A. Petrii, J. Electroanal.
Chem. 582 (2005) 118.
[64] EEC Council directive on the quality of water for human consumption
(No-80/778 off) J. EEC. 229, 1980, 11.
[65] Y. Vanlangendonck, D. Corbisier, A. van Lierde, Wat. Res. 39 (2005) 3028.