Nitrite Reduction to Nitrous Oxide and Ammonia by TiO2 Electrons in a Colloid Solution via Consecutive One-Electron Transfer Reactions

Article abstract of DOI:10.1021/acs.jpca.6b01761

The mechanism of nitrite reduction by excess electrons on TiO₂ nanoparticles (e_TiO2^-) was studied under anaerobic conditions. TiO₂ was loaded with up to 75 electrons per particle, induced by γ-irradiation of acidic TiO₂ colloid solutions containing 2-propanol. Time-resolved kinetics and material analysis were performed, mostly at 1.66 g L^-1 TiO₂. At relatively low nitrite concentrations (R = [e_TiO2^-]_o/[nitrite]_o > 1.5), e_TiO2^- decays via two consecutive processes; at higher concentrations, only one decay step is observed. The stoichiometric ratio Δ[e_TiO2^-]/[nitrite]_o of the faster process is about 2. This process involves the one-electron reduction of nitrite, forming the nitrite radical (k₁ = (2.0 ± 0.2) × 10⁶ M^-1 s^-1), which further reacts with e_TiO2^- (k₂) in competition with its dehydration to nitric oxide (NO) (k₃). The ratios k₂/k₃ = (3.0 ± 0.5) × 10³ M^-1 and k₂ > 1 × 10⁶ M^-1 s^-1 were derived from kinetic simulations and product analysis. The major product of this process is NO. The slower stage of the kinetics involves the reduction of NO by e_TiO2^-, and the detailed mechanism of this process has been discussed in our earlier publication. The results reported in this study suggest that several intermediates, including NO and NH₂OH, are adsorbed on the titanium nanoparticles and give rise to inverse dependency of the respective reaction rates on the TiO₂ concentration. It is demonstrated that the reduction of nitrite by e_TiO2^- yields mainly N₂O and NH₃ via consecutive one-electron transfer reactions.

Full text of DOI:10.1021/acs.jpca.6b01761

Article
pubs.acs.org/JPCA
Nitrite Reduction to Nitrous Oxide and Ammonia by TiO Electrons in
2
a Colloid Solution via Consecutive One-Electron Transfer Reactions
†
†
‡
,†
Sara Goldstein, David Behar, Tijana Rajh, and Joseph Rabani*
†Institute of Chemistry and the Accelerator Laboratory, The Edmond Safra Campus, The Hebrew University of Jerusalem, Jerusalem
9
1904, Israel
‡
Center for Nanoscale Materials, Argonne National Laboratory, Argonne, Illinois 60439, United States
S Supporting Information
*
ABSTRACT: The mechanism of nitrite reduction by excess
−
electrons on TiO2 nanoparticles (eTiO2 ) was studied under
anaerobic conditions. TiO was loaded with up to 75 electrons
2
per particle, induced by γ-irradiation of acidic TiO2 colloid
solutions containing 2-propanol. Time-resolved kinetics and
−
1
material analysis were performed, mostly at 1.66 g L TiO . At
2
−
relatively low nitrite concentrations (R = [eTiO2 ] /[nitrite] > 1.5),
o
o
−
eTiO2 decays via two consecutive processes; at higher concen-
trations, only one decay step is observed. The stoichiometric ratio
−
Δ[eTiO2 ]/[nitrite] of the faster process is about 2. This process involves the one-electron reduction of nitrite, forming the
o
6
−1 −1
−
nitrite radical (k = (2.0 ± 0.2) × 10 M s ), which further reacts with e
oxide (NO) (k ). The ratios k /k = (3.0 ± 0.5) × 10 M and k > 1 × 10 M s−1 were derived from kinetic simulations and
product analysis. The major product of this process is NO. The slower stage of the kinetics involves the reduction of NO by
eTiO2 , and the detailed mechanism of this process has been discussed in our earlier publication. The results reported in this study
(k ) in competition with its dehydration to nitric
1
TiO
6
2
2
−1
3
−1
3
2
3
2
−
suggest that several intermediates, including NO and NH OH, are adsorbed on the titanium nanoparticles and give rise to inverse
2
−
dependency of the respective reaction rates on the TiO concentration. It is demonstrated that the reduction of nitrite by e
2
TiO
2
yields mainly N O and NH via consecutive one-electron transfer reactions.
2
3
270 M− s−1 at pH 2.3 for N O and NH OH, respectively,
1
INTRODUCTION
2
2
■
where under the same experimental conditions, the reductions
of N O and N are extremely slow, if they occur at all. The
question arises whether nitrite reduction by eTiO2 to N O, N ,
The application of TiO nanoparticles for decontamination of
wastewater from organic and inorganic hazardous metal ions
has become a mature research field.
2
14
2
2
−
1
−4
2 2
Most photocatalytic
works focus on oxidation reactions in aerated media. Much less
is known about the reductive processes, although the reduction
of nitrogen compounds, particularly nitrate, has attracted
and NH
processes yielding NO, HNO, and NH
or via a multielectron transfer process. The present work
demonstrates that this process yields mainly N
consecutive one-electron transfer reactions.
3
proceeds via consecutive one-electron transfer
OH as intermediates
2
5
−13
considerable attention.
The reduction of nitrate in aqueous
O and NH via
2 3
solutions involves the formation of nitrite and several nitrogen
intermediates with known physical and chemical properties,
14
including NO, NH OH, N , and N O. Previously, we have
2
2
2
EXPERIMENTAL SECTION
■
−
shown that TiO electrons (e
) are unable to reduce N and
2
TiO
2
2
Materials. Water for preparation of the solutions was
deionized and purified using a Milli-Q purification system. The
following chemicals were of analytical grade and used as
received: titanium tetrachloride, neocupreine (neoc), ascorbic
acid, phenol, sodium nitroprusside and 8-hydroxyquinoline,
hydroxylamine hydrochloride, CuCl , Na PO ·12H O,
N O, and the NO reduction under anoxic conditions yields
2
mainly NH and N O via consecutive one-electron transfer
3
2
13
reactions. Contrary to this finding, Mohamed et al. reported
that nitrate reduction by eTiO2 to NH3 proceeds via a
−
multielectron transfer process.
Relatively little research has been carried out on the time-
resolved kinetics of nitrite reduction by e
constant has been determined as 8 × 10 M s−1 at pH 2.3
using the pulse radiolysis technique. This is several orders of
magnitude higher than that reported for NO and NH OH
using the rapid-mixing stopped-flow technique, i.e., 730 and
2
3
4
2
Na C H O ·2H O, ethylenediaminetetraacetic acid (EDTA),
3 6 6 7 2
−
. The rate
hypochlorite from Sigma-Aldrich, Ar 99.999%, N 99.999%, and
TiO
−1
2
2
5
5
Received: February 21, 2016
Revised: March 29, 2016
Published: March 30, 2016
2
©
2016 American Chemical Society
2307
DOI: 10.1021/acs.jpca.6b01761
J. Phys. Chem. A 2016, 120, 2307−2312

The Journal of Physical Chemistry A
Article
N O from Maxima, Israel. The colloidal stock TiO solution
part of the collected solution was equilibrated with O -saturated
2
2
2
was prepared by slowly introducing TiCl under a stream of
water. The remaining solution in the collecting syringe was
immediately bubbled with Ar for 5 min to remove NO and
analyzed for residual nitrite. Control experiments showed an
error of <10%.
4
argon to vigorously stirred 0.1 M HCl followed by dialysis of
the resulting optically transparent colloidal solutions at pH 2.5.
Subsequently, the solution was heated for 96 h at 55 °C to
15
obtain the desired nanoparticle size.
Ammonia was assayed using a modification of the Berthelot
reaction based on indophenol formation via the reaction of
Characterization. The characterization of TiO21 nano-
4
18
particles was reported in our earlier manuscript. The
phenol with alkaline hypochlorite. Briefly, phenol (6 g) and
crystalline character of the TiO nanoparticles was investigated
sodium nitroprusside (20 mg) were dissolved in 100 mL of
2
by X-ray diffraction (XRD). The sample exhibited diffraction
patterns characteristic of 64.4% crystal phase of pure anatase
phosphate-citrate buffer at pH 12 (3 g of Na PO ·12H O, 3 g
3
4
2
of Na C H O ·2H O, 0.3 g of EDTA). Alkaline hypochlorite
3 6 6 7 2
(
(
5.7 nm particle size) and 35.6% crystal phase of pure brookite
5.2 nm particle size). The weighted average diameter was
(0.3%) was prepared in 0.57 M NaOH. Two milliliter samples
were mixed with 0.8 mL of phenol-nitroprusside and 1.2 mL of
alkaline hypochlorite. The solutions were incubated at 45 °C
for 15 min in 5 mL plastic syringes with no gas phase to avoid
determined to be 5.52 nm. Transmission electron microscopy
(
TEM) measurements show that the particles have an
elongated faceted structure with polygonal geometry. The
average short diameter was found to be 4.09 nm, and the
average longer diameter was found to be 5.57 nm. These values
are in close agreement with the weighted average diameter of
the evaporation of NH from the solution. The absorption was
3
4
−1
−1 18
monitored at 635 nm, applying ε635 = 2.3 × 10 M cm .
Despite the centrifugation, the TiO2 induced high blank
absorbance, although it did not interfere with the analysis.
Analysis of hydroxylamine was carried out using the 8-
5
.52 nm obtained using XRD. On the basis of the XRD
2
19
measurements, the average particle surface area (95.7 nm ),
hydroxyquinoline method. Briefly, 0.5 mL of 1% 8-
3
average particle volume (88.0 nm ), and average particle weight
hydroxyquinoline in ethanol and 0.5 mL of 1 M Na CO
2
3
19
(
3.4 × 10− g) were calculated assuming an ellipsoid shape and
were added to a 1 mL sample. The absorption at 708 nm was
monitored after incubation for 5 min at 95 °C and subsequent
cooling. A calibration curve was prepared with a standard
solution of hydroxylamine hydrochloride.
−3
the density of anatase TiO to be 3.9 g cm . Under the
2
experimental conditions of 0.8−1.66 g/L TiO , the number of
2
1
8
TiO particles will consequently be 2.4−4.7 × 10 particles/L.
2
−
Methods. Production of e
. γ-Radiolysis experiments
N
2
O was analyzed using a 5890 Hewlett-Packard gas
TiO
2
1
37
chromatograph equipped with a thermal conductivity detector,
were carried out at room temperature using a Cs source (6.4
−
1
−1
a 12 ft −1/8 in. Porapak Q column, and a flow of 20 mL min
Gy min ). Unless otherwise stated, TiO electrons were
2
of He as a carrier gas. The operating oven, injector, and
detector temperatures were 60, 120, and 150 °C, respectively.
The retention times of O /N and N O were 0.8 and 2.8 min,
produced upon irradiation of 10 mL syringes containing an Ar-
saturated 1.66 g/L TiO colloidal suspension, 0.011 M HCl,
2
2
2
2
and 1 M 2-propanol with no gas phase. The concentration of
−
−1
respectively. The yields of N
2
O were calculated based on a
eTiO was determined either directly using ε660 = 550 M
2
standard curve prepared by injecting known amounts of N O
−
1
14
2
cm
or indirectly via the reduction of cupric neocuprein
gas (Maxima, Israel).
(
Cu(neoc) ) to Cu(neoc)2+ using ε445(Cu(neoc) ) = 7500
2
+
+
2
2
Kinetics. Stopped-flow kinetic measurements were carried
out with the Bio SX-17MV Sequential Stopped-Flow from
Applied Photophysics with a 1 cm optical path at 25 °C. Anoxic
−
1
−1
2+
M
cm , determined when Cu(neoc)2 was reduced by
ascorbic acid in acidic solutions. Irradiation of the samples was
−4
carried out for up to 4 h, producing up to (6.2 ± 0.17) × 10
−
aqueous solutions of nitrite were mixed with eTiO2 in 1.66 g/L
M eTiO2−
.
TiO , 1 M 2-propanol, and 0.011 M HCl at a 1:1 volume ratio,
2
Flow Mixing for Product Analysis. Aqueous nitrite solutions
−
and the decay of eTiO2 was followed at 660 nm. The effect of
−
were mixed with equal amounts of eTiO2 in 1.66 g/L TiO , 1 M
2
oxygen during the kinetic measurements was minimized by
2
-propanol, and 0.011 M HCl using a flow system (SP230iw
−
successively flushing the system with eTiO2 or anoxic water.
syringe pump, World Precision Instruments). The reactant
solutions were streamed from two glass syringes in opposite
directions at the site of mixing. The mixing system was
comprised of whole-glass capillary tubing connected to the
syringes by three-way Teflon stopcocks. The mixed solution
reached the collecting glass syringe containing a small amount
of anoxic water 1.3 s after mixing so that all solutes were diluted
by a factor of 2.18. Product analysis was carried out at least 10
min after mixing to allow the termination of slow reactions.
Product Analysis. Generally, centrifugation was employed
−
The lack of considerable O reduction by e
during the
2
TiO
2
observed time range was confirmed by control experiments
with anoxic water. Each value given is an average of at least
three measurements.
Modeling of the experimental results was carried out using a
noncommercial program developed at Brookhaven National
Laboratories by Dr. H. A. Schwarz.
RESULTS
■
when precipitation of TiO occurred during analysis.
2
Stopped-Flow Kinetics Studies. At high ratios of R =
Nitrite concentration was assayed by mixing equal volumes
−
−
16
[eTiO2 ] /[nitrite] , the decay of the eTiO2 absorbance occurs
o o
of the sample and the Griess reagent prepared in HCl, and
via two predominant processes, as shown in Figure 1. The fast
decaying fraction increases at the expense of the slow decay
the absorption at 540 nm was read after 10 min. To avoid
possible interference by the reaction of HNO with 2-
2
1
7
with increasing nitrite concentration. Note that the term
propanol, the mixed reactants were added to anoxic Griess
reagent placed in the collecting syringe about 3.3 s after passing
the mixing point. The 2 s additional delay was introduced by
extending the capillary tube between the mixing point and the
collecting syringe to ensure termination of the reaction. Next,
−
“
5
nitrite” refers to the sum [HNO ] + [NO ] in the presence of
2
2
−
.5 mM HCl. In view of the fast reaction rate of e
with
TiO
2
5
nitrite, practically all of the nitrite was consumed by the end of
the fast kinetic stage. When 75 μM NO2− was reacted with 150
2
308
DOI: 10.1021/acs.jpca.6b01761
J. Phys. Chem. A 2016, 120, 2307−2312

The Journal of Physical Chemistry A
Article
Figure 3. Effect of R on the yields of NH and N O using 280 μM
3
2
Figure 1. Typical kinetic traces for the reaction of 150 μM eTiO − with
−
eTiO2 and 5 mM HCl. The solid lines were computed on the basis of
2
15, 30, 45, and 75 μM nitrite in 5.5 mM HCl monitored at 660 nm.
the proposed mechanism described in Scheme 1 and the rate constants
listed in Table 1 (see Discussion).
μM eTiO2−, a relatively low concentration of eTiO − remained at
2
production is about 10% of the initial nitrite, which is the
combined error in the measurement of [NH ] + [N O] .
the end of the slower process. This implies that the fast
reduction process of nitrite involves more than one electron. At
relatively low ratios of R, only one decay process is observed, as
3
2
Only small concentrations of NH OH (<2.6% of [nitrite] )
2
o
were observed in the narrow range of 1 < R < 2. Residual NO
becomes important at R < 2.5, and unreacted nitrite is observed
at R < 1.5 (Figure 4).
−
shown in Figure 2 for 110 and 220 μM NO2 reacting with 150
μM eTiO2−
.
Figure 4. Residual concentrations of nitrite and NO at R < 2.
Experimental conditions were the same as in Figure 3. The solid lines
were computed using the proposed mechanism described in Scheme 1
and the rate constants listed in Table 1.
Figure 2. Kinetic traces for the reaction of 150 μM eTiO − with 110 and
2
220 μM nitrite in 5.5 mM HCl monitored at 660 nm. The lower initial
absorption in the 220 μM nitrite system is due to a significant decay
during the mixing time.
DISCUSSION
■
−
At relatively low concentrations, nitrite reduction by e
in
Decreasing the TiO concentration enhances the decay rates
TiO
2
2
of both processes and increases the fraction of the fast stage at
the expense of the slow decay. This is attributed to the
anoxic acidic solutions shows two predominant kinetic
processes. The ratio between consumed eTiO2 and [nitrite]o
−
adsorption of the reaction intermediates at the TiO surface,
2
under the conditions of Figure 1 is about 2. Because nearly all
of the nitrite is consumed under these conditions, the
stoichiometry corresponds to a two-electron reduction of the
nitrite to HNO without formation of NO as an intermediate.
However, such an interpretation is in contrast with the
considerable amount of NO produced (Figure 4) as well as
resulting in an increase in their local concentrations at the
reaction site with decreasing [TiO2].
Product analysis included N O and NH , which are stable
2
3
−
toward eTiO2 as well as the residual electrons, NH OH, NO,
2
and nitrite, if they remain. The yields of N O and NH (Figure
2
3
3
) increase sharply at R > 1, become the major products at R >
, and together approach the equivalent initial concentration of
the slower kinetic stage, which takes place in the time range
previously observed for NO reduction by excess eTiO2−
.
14
The
2
nitrite at R > 3. The highest yield of N O is observed at R = 3,
2
fast process can be attributed neither solely to the one-electron
reduction of nitrite to NO nor solely to the two-electron
reduction of nitrite to HNO. Therefore, the one-electron
while the yield of NH increases steadily with R. At R = 5,
3
[
NH ] + 2[N O] accounts for 99% of the nitrite. Because
3
2
−
unreacted eTiO2 is observed at the end of the reduction process
under these experimental conditions, neither nitrite nor NO or
NH OH remains. Thus, the upper equivalent limit of N2
•
−
•2−
reduction of nitrite forming HN O or NO2 (or protonated
2
20
forms of these transients), followed by a subsequent fast one-
electron reduction to HNO in competition with dehydration to
2
2
309
DOI: 10.1021/acs.jpca.6b01761
J. Phys. Chem. A 2016, 120, 2307−2312

The Journal of Physical Chemistry A
Article
NO, must be invoked. The ratio of about 2 between reacted
computed bimolecular rate constants must be 3 orders of
magnitude lower than the respective values derived for the
distribution of the homogeneous reactants. In the absence of
better information concerning the nature of adsorption and the
adsorption−desorption rates, we found it appropriate to report
the results in terms of homogeneous reactions.
The complex mechanism shown in Scheme 1 raises the
question of whether more than one set of rate constants may fit
the experimental data. However, the simulations of both the
reaction products (Figures 3 and 4) and the time profiles
−
eTiO2 and [nitrite] is the weighted average of 1−6 electrons
o
reacting with one nitrite molecule during the fast stage. The
proposed mechanism is described in Scheme 1.
Scheme 1. Proposed Mechanism for Nitrite Reduction by
−
a
eTiO2
(
Figures 5 and 6) using the rate constants listed in Table 1
show a very good agreement with the experimental data.
Computed simulations similar to those presented in Figures 5
and 6 were obtained for various nitrite concentrations with
relatively small adjustments of the reaction rate constants.
6
−1 −1
A value of k = (2.0 ± 0.2) × 10 M
s
was derived from
1
the excellent fits of the simulations at relatively high nitrite
concentrations as shown in Figure 6. At [nitrite] = 110 and
o
−
2
20 μM, the nitrite reacts with >75% and >85% eTiO2
,
respectively. This is evident from the low yields of products
involving HNO as their precursor (Figure 3). Therefore, the
derivation of k is not significantly affected by the uncertainties
1
−
in the rate constants of competing eTiO2 reactions. The k value
1
required for best fits of the kinetic traces shown in Figure 1
a
Specific rate constants are listed in Table 1.
6
6
−1 −1
ranged between 1.7 × 10 and 2.3 × 10 M s ; k ranged
4
−
1 −1
between 700 and 1000 M s , and the ratio k /k was within
.7 × 10 and 3.3 × 10 M . The computed curves in Figures 3
2
3
The rate constants derived from the simulations that are
3
3
−1
2
required for best fits to the experimental results are presented
in Table 1. These rate constants are in terms of homogeneous
reactions, although eTiO2 is confined to the TiO nanosurface;
and 4 are based on the weighted average reaction rate constants
given in Table 1.
−
2
With the exception of k , k , and k , where only upper or
7
9
12
all of the reaction intermediates are produced at the surface,
and some further react in the same microvolume. In this case,
the reaction volume is on the order of 10 nm for one
lower values apply, in most cases changing a rate constant
affects the simulated time profiles and distribution of the
products. If the change is not too large, the initial simulated
traces and products can be restored by appropriate changes to
other rate constants that have an opposite effect. The error
limits of the rate constants in column 4 of Table 1 include both
the interrelation between the kinetic parameters and the
2
3
nanoparticle, equivalent to ∼0.5 mL of 1.6 g/L TiO , which was
2
used in most experiments. The concentrations of the reactants
and their reaction rates at the surface are evidently much higher
than the homogeneous values. Consequently, the resulting
Table 1. Reactions and Rate Constants Employed in the Kinetic Simulations
no.
reaction
average k values
comment
a
−
2−
2•2−
2.0 × 10 M s−1
6
−1
±2 × 10 M s−1
5
−1
1
2
3
4
e
e
TiO + NO → NO
2
−
2•2−
+
3.0 × 10 M s−1
8
−1
k > 1 × 10 M s−1, k /k = (3.0 ± 0.6) × 10 M
6
−1
3
−1
TiO + NO
(+3H ) → HNO + H
O
2
2
2
2
3
b
•2−
+
5
−1
k3 > 370 s−1, k /k = (3.0 ± 0.6) × 10 M
3
−1
NO2
+ (2H ) → NO + H O
1 × 10 s
2
2 3
−
+
800 M− s−1
1
±70 M−1 s−1
0 < k < 3 × 10 M s−1, k /k < 3
e
TiO + NO + (H ) → HNO
2
8 × 10 M s−1
6
−1
21
21
21
7
−1
5
6
7
HNO + HNO → N2O + H2O
HNO + NO → HN
HN + NO → N
5
5
6
6 × 10 M s−1
6
−1
±1 order of magnitude provided that k < k and k /k = 3.75
5 6 6 8
O
2
2
5 × 10 M s−1
9
−1
k ≫ k , k
O
O + HNO
2
2
2
2
7
5
6
−
+
•
14
1.6 × 10 M s−1
6
−1
4 × 10 < k < 3 × 10 M s−1
5
6
−1
8
9
1
1
1
e
TiO + HNO(+H ) → NH
2
O
2
8
•
•
1.4 × 10 M s−1
8
−1
22
NH
O + NH
O → N
+ 2H
O
k9/k10 < 10
2
2
2
2
−
•
+
8
−1
k10 > 1 × 10 M s−1
7
−1
0
1
2
eTiO2 + NH2O (+H ) → NH2OHads
1 × 10 M s−1
−
−
eTiO2 + NH OH → NH OH
1 × 10 M s−1
5
−1
3 × 10 < k < 3 × 10 M s−1
4
5
−1
2
ads
2
ads
11
−
ads−
+
1 × 10 M s−1
6
−1
k12 > k11
e
TiO + NH
OH (+2H ) → NH
+ H
O
2
2
3
2
a
The reduction intermediate of the nitrite is assigned as NO2•2−, although we do not know the nature of the reaction product.20 bThe value of k = 5
3
3
6 −1
×
10 or 1.6 × 10 s , depending on whether the nitrite is reduced at the nitrogen or oxygen atom, respectively, which was determined by pulse
20
•
−
radiolysis in a homogeneous solution. The decomposition of HN O (H on the nitrogen) is catalyzed even by moderately weak acids such as
2
6
−1
3
−1
6
−1 −1
3
−1
CH CO H, implying that k > 1.6 × 10 s . Values of k > 1 × 10 s and k > 3 × 10 M
s
fit the data, provided that k /k = 3.0 × 10 M .
3
2
3
3
2
2 3
2
310
DOI: 10.1021/acs.jpca.6b01761
J. Phys. Chem. A 2016, 120, 2307−2312

The Journal of Physical Chemistry A
Article
amount of data. The error margins of k , k , and k are also
1
4
11
relatively small. This is evident because small variations in these
rate constants introduce significant deviations in the respective
6
−1 −1
computed time profiles. The value of k = 2 × 10 M
s
1
derived from the simulations is somewhat higher than that
previously reported for one electron per TiO particle, namely
2
5
−1 −1
5
8
× 10 M
s
for the same pH. Because the redox potential
of eTiO2 becomes more negative with an increasing number of
electrons per particle, the higher value determined in this
−
23
study for 75 electrons per TiO particle is reasonable. At a given
2
concentration of nitrite, the relative magnitudes of the fast and
slow decays are determined predominantly by the competition
between reactions 2 and 3, while reaction 4 is a rate-
determining step of the slow kinetic stage. The computed
and experimental time profiles are similar over a wide range of
3
−1
k and k values, provided that k /k = (3.0 ± 0.6) × 10 M
2
3
2
3
6
−1 −1
and k > 1 × 10 M s . The rate constants of reactions 5, 6,
2
and 8 cannot be varied independently. Reactions 5 and 6 retard
the decay of the absorbance by removing HNO and producing
N O. On the other hand, reaction 8 removes both HNO and
2
−
eTiO2 , enhances the decay of the absorbance, and leads to the
formation of NH . Therefore, the freedom to choose the above
3
rate constants is limited, as the same rate constants must fit the
end products and time profiles of both the fast and the slow
stages for all nitrite concentrations. Although both reactions 5
and 6 produce N O, their effect on the kinetics is different
2
because reaction 6 gives rise to the regeneration of nitrite. The
−
only requirement concerning reaction 7 is that k ≫ k . The k
7
6
7
Figure 5. Reaction of 150 μM eTiO2 with 15 and 45 μM nitrite in 5.5
21
value derived from homogeneous solutions agrees with this
mM HCl monitored at 660 nm. The computed lines were obtained
using the mechanism shown in Scheme 1 and k = 2 × 10 M s , k3
1 × 10 s , k = 8 × 10 M s , k = 6 × 10 M s , k = 5 × 10
s , k = 1.6 × 10 M s , k = 1.4 × 10 M s , k = 1 × 10
s , k11 = 1 × 10 M s , and k = 1 × 10 M s . The rate
constants k = 3.3 × 10 and k = 8 × 10 M
the 15 μM kinetic traces, while k = 3.0 × 10 and k = 9 × 10 M
were employed for the 45 μM kinetic traces.
•
requirement. Reactions 9 and 10 compete for NH O , which is
6
−1 −1
2
1
5
−1
6
−1 −1
6
−1 −1
9
8
the precursor of NH OH. The choice of the ratio k /k < 10 is
2
9
10
=
M
M
5
6
7
−
1
1
−1
6
−1 −1
8
−1 −1
limited by the lack of significant amounts of N . A value of 1.4
2
8
9
10
8
−1 −1
22
7
−
−1
5
−1 −1
6
−1 −1
×
10 M
s
was measured in homogeneous solutions,
which is within the acceptable range. A lower limit of k > 10
12
8
2
−1 −1
s
were employed for
10
2
4
8
2
−1 −1
−1 −1
s
M
s
was determined from the kinetics. Increasing these
2
4
constants by 1−3 orders has no effect on the simulations.
NH OH was not accumulated in our system, implying that
2
5
−1 −1
k11 > 1 × 10 M s , which is considerably higher than the
value of 270 M−
1
s
−1
previously reported. The pronounced
14
difference is attributed to the formation of adsorbed NH OH
2
produced on the titanium nanoparticle by reaction 10, whereas
the rate constant 270 M−
1
s
−1
was determined for the eTiO2−
reaction with bulk NH OH. Apparently, the equilibration of
2
bulk with surface NH OH is slower than that of the reduction
2
rate. Evidently, the revision of k11 applies also to the NO
1
4
system, which we have also previously studied. Using k11
>
5
−1 −1
1
M
0 M
s
fits very well with the observed data if k = 350
4
−
1
−1
14
s . This value is about half of that previously reported,
although it is in agreement with the value of 330 M s−1
−1
derived from the NO system (see the open circles in Figure 8
of ref 14). The apparent discrepancy can be resolved if NO
Figure 6. Reaction of 150 μM eTiO − with 110 and 220 μM nitrite in
14
2
produced at the TiO surface is adsorbed; in the NO study,
2
5
.5 mM HCl monitored at 660 nm. The computed lines were obtained
the mixing involved aqueous bulk NO.
6
−1 −1
using the mechanism shown in Scheme 1 and k = 2.1 × 10 M
s
−
1
The enhanced decay rates of eTiO2 observed upon decreasing
6
−1 −1
(for 110 μM nitrite) or k = 1.95 × 10 M
s
(for 220 μM nitrite).
1
[
TiO ] suggest that at least some of the reacting species are
The other rate constants were the same for both cases, namely k = 3.0
2
2
•
2−
8
−1 −1
5
−1
2
−1 −1
6
6
5
×
10 M s , k = 1 × 10 s , k = 8 × 10 M s , k = 8 × 10
s , k = 6 × 10 M s , k = 5 × 10 M s , k = 1.6 × 10
s , k = 1.4 × 10 M s , k = 1 × 10 M s , k = 1 × 10
adsorbed to the nanoparticles. These include NO2 , HNO,
3
4
5
−1
−1
−1
−1
6
−1 −1
9
−1 −1
M
M
M
NO, and NH OH. Nitrite may be only partially adsorbed.
6
7
8
2
−1
8
−1 −1
8
−1 −1
9
10
11
Further discussion regarding this is located in the Supporting
Information.
−1
6
−1 −1
s , and k12 = 1 × 10 M s .
The rate constants of reactions 3, 5−7, and 9 listed in Table
experimental errors in the kinetics and the product analysis.
The latter is small compared to the former in view of the large
1 are for homogeneous solutions. These rate constants may be
different if adsorption to TiO is important. The simulations
2
2
311
DOI: 10.1021/acs.jpca.6b01761
J. Phys. Chem. A 2016, 120, 2307−2312

The Journal of Physical Chemistry A
Article
derive fairly accurate results for the ratios of k /k and k /k , the
(8) Rengaraj, S.; Li, X. Z. Enhanced Photocatalytic Reduction
2
3
6
8
3+
Reaction over Bi −TiO Nanoparticles in Presence of Formic Acid as
lower limit of k /k , and the upper limit of k /k . The error
2
10
9
5
6
a Hole Scavenger. Chemosphere 2007, 66, 930−938.
margin of the absolute rate constant may be much higher, as
shown in Table 1. Note that the computations are not sensitive
(9) Ergas, S. J.; Rheinheimer, D. E. Drinking Water Denitrification
Using a Membrane Bioreactor. Water Res. 2004, 38, 3225−3232.
to the value taken for k because reactions 5 and 6 are rate-
7
(10) Zhang, F.; Pi, Y.; Cui, J.; Yang, Y.; Zhang, X.; Guan, N.
determining for the reduction of HNO to N O. Determinations
2
Unexpected Selective Photocatalytic Reduction of Nitrite to Nitrogen
of the uncertainty limits are demonstrated and further discussed
in the Supporting Information.
on Silver-Doped Titanium Dioxide. J. Phys. Chem. C 2007, 111, 3756−
3
761.
(
11) Kominami, H.; Nakaseko, T.; Shimada, Y.; Furusho, A.; Inoue,
CONCLUSIONS
We have unequivocally shown the existence of a reaction path
in which nitrite reduction to N O and NH by e
proceed via the formation of NO as an intermediate. This
bypass involves consecutive one-electron reactions, including
the same reactions proposed for the NO system. The study
H.; Murakami, S.-y.; Kera, Y.; Ohtani, B. Selective Photocatalytic
Reduction of Nitrate to Nitrogen Molecules in an Aqueous Suspension
of Metal-Loaded Titanium(IV) Oxide Particles. Chem. Commun. 2005,
■
−
does not
2
3
TiO
2
2
933−2935.
(12) Kominami, H.; Furusho, A.; Murakami, S.-y.; Inoue, H.; Kera,
Y.; Ohtani, B. Effective Photocatalytic Reduction of Nitrate to
Ammonia in an Aqueous Suspension of Metal-Loaded Titanium(IV)
Oxide Particles in the Presence of Oxalic Acid. Catal. Lett. 2001, 76,
31−34.
provides evidence for adsorption of both NO and NH OH as
2
well as the intermediates NO2• and HNO to the TiO surface.
2‑
2
(
13) Mohamed, H. H.; Mendive, C. B.; Dillert, R.; Bahnemann, D.
The detailed mechanism quantitatively agrees with both
material balance and time profiles of the electron decay.
W. Kinetic and Mechanistic Investigations of Multielectron Transfer
Reactions Induced by Stored Electrons in TiO2 Nanoparticles: A
Stopped Flow Study. J. Phys. Chem. A 2011, 115, 2139−2147.
(14) Goldstein, S.; Behar, D.; Rajh, T.; Rabani, J. Nitric Oxide
Reduction to Ammonia by TiO Electrons in Colloid Solution Via
Consecutive One-Electron Transfer Steps. J. Phys. Chem. A 2015, 119,
ASSOCIATED CONTENT
Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jpca.6b01761.
■
*
S
2
2
(
760−2769.
15) Gao, R. M.; Safrany, A.; Rabani, J. Fundamental Reactions in
2
Determination of the uncertainty limits and further
discussion, effect of the k value on the deviation from
the reference trace (Figure 1S), effect of changing k and
TiO Nanocrystallite Aqueous Solutions Studied by Pulse Radiolysis.
Radiat. Phys. Chem. 2002, 65, 599−609.
1
2
(16) Green, L. C.; Wagner, D. A.; Glogowski, J.; Skipper, P. L.;
Wishnok, J. S.; Tannenbaum, S. R. Analysis of Nitrate, Nitrite, and N-
15 Labeled Nitrate in Biological Fluids. Anal. Biochem. 1982, 126,
131−138.
k3 on the computed traces (Figure 2S), and effect of
changing k , k , and k on the computed traces (Figure
5
6
8
3S) (PDF)
(17) Williams, D. L. H. Nitrosation; Cambridge University Press:
Cambridge, U.K., 1988; p 212.
AUTHOR INFORMATION
Corresponding Author
E-mail: rabani@mail.huji.ac.il.
■
(18) Scheiner, D. Determination of Ammonia and Kjeldahl Nitrogen
by Indophenol Method. Water Res. 1976, 10, 31−36.
(19) Palling, J. D.; Hollocher, T. C. Basis and Optimization of the
Indooxine Method to Determine Oximes. Microchim. Acta 1985, 86,
*
Notes
1
(
37−144.
The authors declare no competing financial interest.
20) Lymar, S. V.; Schwarz, H. A.; Czapski, G. Reactions of the
−•
2‑•
Dihydroxylamine (HNO2 ) and Hydronitrite (NO2 ) Radical Ions.
J. Phys. Chem. A 2002, 106, 7245−7250.
ACKNOWLEDGMENTS
This work was generously supported by the US−Israel BSF
under Contract 2012158.
■
(21) Lymar, S. V.; Shafirovich, V.; Poskrebyshev, G. A. One-Electron
Reduction of Aqueous Nitric Oxide: A Mechanistic Revision. Inorg.
Chem. 2005, 44, 5212−5221.
(
22) Lind, J.; Merenyi, G. Kinetic and Thermodynamic Properties of
REFERENCES
■
•
the Aminoxyl (NH O ) Radical. J. Phys. Chem. A 2006, 110, 192−197.
2
(
1) Lang, X.; Ma, W.; Chen, C.; Ji, H.; Zhao, J. Selective Aerobic
(23) Dimitrijevic, N. M.; Savic, D.; Micic, O. I.; Nozik, A. J.
Oxidation Mediated by TiO Photocatalysis. Acc. Chem. Res. 2014, 47,
2
Interfacial Electron-Transfer Equilibria and Flatband Potentials of α-
Ferric Oxide and Titanium Dioxide Colloids Studied by Pulse
Radiolysis. J. Phys. Chem. 1984, 88, 4278−83.
3
(
55−363.
2) Hoffmann, M. R.; Martin, S. T.; Choi, W.; Bahnemann, D. W.
Environmental Applications of Semiconductor Photocatalysis. Chem.
Rev. 1995, 95, 69−96.
(
3) Arbour, C.; Sharma, D. K.; Langford, C. H. Picosecond Flash
Spectroscopy of Titania Colloids with Adsorbed Dyes. J. Phys. Chem.
990, 94, 331−335.
4) Mohamed, H. H.; Bahnemann, D. W. The Role of Electron
Transfer in Photocatalysis: Fact and Fictions. Appl. Catal., B 2012,
28, 91−104.
5) Gao, R. M.; Safrany, A.; Rabani, J. Reactions of TiO Excess
1
(
1
(
2
Electron in Nanocrystallite Aqueous Solutions Studied in Pulse and
Gamma-Radiolytic Systems. Radiat. Phys. Chem. 2003, 67, 25−39.
(
6) Kudo, A.; Domen, K.; Maruya, K.-I.; Onishi, T. Photocatalytic
−
Reduction of NO3 to Form NH over Pt-TiO . Chem. Lett. 1987,
3
2
1
(
019−1022.
7) Gao, W.; Jin, R.; Chen, J.; Guan, X.; Zeng, H.; Zhang, F.; Guan,
N. Titania-Supported Bimetalic Catalysts for Photocatalytic Reduction
of Nitrate. Catal. Today 2004, 90, 331−336.
2
312
DOI: 10.1021/acs.jpca.6b01761
J. Phys. Chem. A 2016, 120, 2307−2312