Oxidation of nitrite by peroxynitrous acid

Article abstract of DOI:10.1021/jp0269064

The kinetics of the oxidation of nitrite to by peroxynitrous acid at pH 5.2 is best described by the rate law k_obs = k_iso + k′[NO₂^-] + k″[NO₂^-]², in which the peroxynitrous acid isomerization rate constant k_iso = (1.10 ± 0.05) s^-1, k′ = (3.2 ± 0.1) M^-1 s^-1, and k″ = (4.2 ± 0.3) M^-2s^-1, at 25 °C. The ternary reaction may involve initial formation of an adduct between nitrite and peroxynitrite, followed by reaction with a second nitrite to form two nitrite and a nitrate. Ab initio calculations indicate that there is only a small intrinsic barrier to the net transfer of HO⁺ from peroxynitrous acid to the nitrogen atom of nitrite. A similar transfer to either of the two oxygens of nitrite produces the reactants, and would not lead to an increase in the rate of disappearance of peroxynitrous acid, as observed. The low rate constant is most likely due to stringent orientational constraints. Formal transfer of HO⁺ to ¹⁵NO₂^- results in formation of ¹⁵NO₃^-, as experimentally observed. HO⁺ transfer is common to the chemistry of peracids, a family of compounds to which peroxynitrous acid belongs.

Full text of DOI:10.1021/jp0269064

J. Phys. Chem. A 2003, 107, 1763-1769
1763
Oxidation of Nitrite by Peroxynitrous Acid
Patrick Maurer,† Chris F. Thomas,† Reinhard Kissner,† Heinz Ru1egger,† Oswald Greter,‡
Ursula Ro1thlisberger,†,§ and Willem H. Koppenol*,†
Laboratorium fu¨r Anorganische Chemie and Laboratorium fu¨r Organische Chemie,
Eidgeno¨ssische Technische Hochschule Zu¨rich, ETH Ho¨nggerberg, CH-8093 Zu¨rich, Switzerland
ReceiVed: September 4, 2002; In Final Form: December 13, 2002
The kinetics of the oxidation of nitrite to nitrate by peroxynitrous acid at pH 5.2 is best described by the rate
-
- 2
law kobs ) kiso + k′[NO2 ] + k′′[NO2 ] , in which the peroxynitrous acid isomerization rate constant kiso
)
(1.10 ( 0.05) s-1, k′ ) (3.2 ( 0.1) M-1 s-1, and k′′ ) (4.2 ( 0.3) M-2s-1, at 25 °C. The ternary reaction may
involve initial formation of an adduct between nitrite and peroxynitrite, followed by reaction with a second
nitrite to form two nitrite and a nitrate. Ab initio calculations indicate that there is only a small intrinsic
barrier to the net transfer of HO+ from peroxynitrous acid to the nitrogen atom of nitrite. A similar transfer
to either of the two oxygens of nitrite produces the reactants, and would not lead to an increase in the rate
of disappearance of peroxynitrous acid, as observed. The low rate constant is most likely due to stringent
orientational constraints. Formal transfer of HO+ to 15NO2- results in formation of 15NO3 , as experimentally
-
observed. HO+ transfer is common to the chemistry of peracids, a family of compounds to which peroxynitrous
acid belongs.
Introduction
that the oxidation of nitrite by peroxynitrous acid, eq 1, is not
fast, and, if this reaction were bimolecular, deviations from first-
order behavior in the decay of peroxynitrous acid would be
observed only at high concentrations of nitrite.
To test whether the reaction in eq 1 occurs during the
isomerization of peroxynitrous acid, we investigated the decay
of peroxynitrous acid in the presence of [15N]nitrite, which
would yield [15N]nitrate. We also studied the kinetics of
peroxynitrous acid decay at high concentrations of nitrite.
Since the suggestion by Beckman et al.1 that peroxynitrite
might be formed in vivo from the diffusion-controlled reaction
of superoxide with nitrogen monoxide,2,3 there has been a
renewed interest in the chemistry of this unstable isomer of nitric
acid.4
Peroxynitrous acid is a special peracid in that it not only
undergoes a bimolecular disproportionation common to all
members of this class of compounds,3,5-10 but also isomerizes
to nitrate.11 The isomerization is first-order in peroxynitrous acid
concentration and may occur via an intramolecular atom shift,
or as a rate-limiting homolytic cleavage into a caged radical
pair followed by fast recombination of the fragments.12
A third pathway to nitrate is possible. It is often neglected
that, with the exception of tetramethylammonium peroxy-
nitrite,13,14 peroxynitrite preparations are contaminated with
nitrite. The nitrite content can be as high as that of peroxynitrite
itself. Given a reduction potential for the ONOOH/NO2-, H2O
couple of about 1.6 V at pH 7,15 and the observation that
addition of excess nitrite to peroxynitrous acid affects its decay
rate at physiological pH,3 it is possible that some of the nitrate
formed during the isomerization could, in fact, arise from the
two-electron oxidation of nitrite. In eq 1
Experimental Section
Chemicals. Na15NO2 was obtained with 99.8% of the nitrite
present as [15N]nitrite from Cambridge Isotope Laboratories Inc.
(Cambridge, MA). Contamination by nitrate was found to be
less than 0.05% by anion exchange chromatography. Na15NO3
was prepared by ozonolysis of Na15NO2 solutions for 5 min.
The completeness of the oxidation was confirmed by anion
exchange chromatography. For the synthesis of peroxynitrite,16
nitrogen monoxide (99.5%) and potassium superoxide (reagent
grade) were obtained from Linde (Ho¨llriegelskreuth, Germany)
and Aldrich (c/o Fluka, Buchs SG, Switzerland), respectively.
Dioxygen (99.998%) and argon (99.998%) were purchased from
PanGas (Lucerne, Switzerland); all other reagents were pur-
chased from Fluka or Siegfried (Zofingen, Switzerland) in
analytical grade quality. Water was purified from deionized
water with a Millipore unit.
ONOOH + NO2- f NO3- + H+ + NO2
(1)
-
Peroxynitrite was synthesized from nitrogen monoxide and
potassium superoxide as described earlier.17 Tetramethylam-
monium peroxynitrite was prepared as described by Bohle et
al.14 Peroxynitrite stock solutions were always freshly prepared
and kept on ice and in the dark, and its concentration was
determined spectrophotometerically (ꢀ302 ) 1700 M-1cm-1).14
the nitrite concentration does not change, and the reaction would
normally go unnoticed. Considering that the rate constants for
the decay of peroxynitrous acid, obtained with a variety of
peroxynitrite preparations with and without major nitrite
contamination, all cluster around 1.2 s-1, it can be concluded
The 14NO2 impurity in the peroxynitrite was determined by
-
* Corresponding author. Telephone: 41-1-632-2875. Fax: 41-1-632-
1090. E-mail: koppenol@inorg.chem.ethz.ch.
ion chromatography after mixing an aliquot of the cold
peroxynitrite solution with 50 mM ice-cooled phosphoric acid
and subsequently adjusting to pH ≈ 7 with 50 mM sodium
hydroxide. The nitrate concentration was determined chromato-
† Laboratorium fu¨r Anorganische Chemie.
‡ Laboratorium fu¨r Organische Chemie.
§ Present address: Institute of Molecular and Biological Chemistry, Ecole
Polytechnique Fe´de´rale de Lausanne, CH-1015 Lausanne, Switzerland.
10.1021/jp0269064 CCC: $25.00 © 2003 American Chemical Society
Published on Web 02/20/2003

1764 J. Phys. Chem. A, Vol. 107, No. 11, 2003
Maurer et al.
graphically and compared to the peroxynitrite concentration
obtained spectrophotometrically. The difference between these
two values was usually about 2% and never exceeded 5%.
Nitrogen dioxide was obtained by thermal decomposition of
dinitrogen tetraoxide prepared by mixing nitrogen monoxide
with excess dioxygen at a flow rate of 2 cm3/s in a 50 cm long
glass tube, the outlet of which allowed the product gas to be
collected in an ice-cooled trap. After collection of 1-2 cm3 of
dinitrogen tetraoxide, the trap was sealed with a silicone polymer
stopper and allowed to warm to room temperature, and the seal
was punctured to permit half of the dinitrogen tetraoxide/
nitrogen dioxide mixture to escape and displace the remaining
excess dioxygen before resealing the container.
Methods. 15N NMR spectra were recorded on a Bruker
AMX-500 spectrometer with a field of 11.70 T at 50.49 MHz
(Bruker AG, Fa¨llanden, Switzerland). About 10,000 pulses per
sample were accumulated overnight. The actual number of
pulses was registered for normalization of peak integrals for
quantitative analysis. NMR sample tubes were always filled with
0.6 mL of solution. For the quantitative determination of [15N]-
nitrite and [15N]nitrate, a calibration was carried out by
measuring the peak integrals of solutions with known concentra-
tions of pure Na15NO2 and Na15NO3. Anion chromatography
was carried out on a Metrohm 732 ion separation center
equipped with a Hamilton PRP X-100 column and a conduc-
tometric detector (Metrohm, Herisau, Switzerland), with 2.5 mM
phthalate buffer at pH 4.5 as the eluent. UV/vis spectra were
recorded on a Kontron Uvikon 820 double beam instrument.
above liquid dinitrogen tetraoxide, and the syringe was then
filled with argon to enhance dissociation of the gaseous
dinitrogen tetraoxide present. The average amount of nitrogen
dioxide taken into the syringe was determined after repeated
injection of its contents into a known volume of rapidly stirred
neutral 5 mM phosphate buffer, which was subsequently
analyzed for the sum of nitrite and nitrate by ion chromatog-
raphy. The nitrogen exchange experiment was carried out by
injecting the nitrogen dioxide-argon mixture into 100 mM [15N]-
nitrite solutions at pH 5.2, and, after 20 s, the pH was brought
to 7 by addition of sodium hydroxide as in the nitrite oxidation
experiments.
Kinetics. Solutions of potassium and tetramethylammonium
peroxynitrite in 5 mM sodium hydroxide were spiked with nitrite
to concentrations ranging from 0.1 to 800 mM. These mixtures
and a nitrite-free sample were mixed with 0.2 M sodium
dihydrogen phosphate in stopped-flow spectrophotometers
(SX17MV, Applied Photophysics, Leatherhead, UK, and OLIS
RSM-16, Bogart, GA). The pH after mixing was 5.2, which
provided extensive hydronation of peroxynitrite, but not of
nitrite. In addition, experiments were carried out at pH 3.6.
During these experiments the ionic strength was kept constant
at 0.9 M, the difference made up with sodium tetraoxochlorate.
The temperature dependence of the rate of isomerization in the
presence of 100 and 600 mM nitrite was determined from 15
°C to 50 °C. The disappearance of peroxynitrous acid was
followed at 265 nm with the Applied Photophysics SX17MV,
or by following changes to the spectrum from 250 to 400 nm
with the OLIS RSM-16.
Low concentrations of [15N]nitrate could not be determined
by NMR spectroscopy and, instead, ion chromatography and
mass spectroscopy were used to determine the total amount of
nitrate and the 14N/15N ratio, respectively. Total nitrate was first
separated from nitrite and determined quantitatively by ion
chromatography. To remove the eluent electrolyte phthalate,
which interferes with the mass spectrometric determination,
copper metal powder in the presence of sulfuric acid was used
to reduce the nitrate to nitrogen monoxide, which was allowed
to react in air to form nitrogen dioxide that was collected in
neutral water to yield equal amounts of nitrite and nitrate. Ozone
was passed through this solution until all nitrite was converted
to nitrate. Completion was confirmed by ion chromatography.
Since all nitrite was converted to a single species, nitrate, the
sensitivity of the mass spectrometric measurements was doubled.
For the determination of the 14N/15N ratio, samples of the
aqueous solution were diluted 1:100 with acetonitrile and
introduced through and ionized by an electrospray inlet. Anion
mass spectra were recorded on a Finnigan TSQ-7000 mass
Simulations. Kinetics were simulated based on sets of
ordinary differential equations with a computer program that
implements a two-step implicit Runge-Kutta algorithm of
fourth order.18 This method is more stable than explicit
integration techniques, but requires more computation time
because of the calculation of the implicit variables. On the other
hand, both detection of inherent instabilities and step-width
control are easy.
Electronic Structure Calculations. To explore possible
mechanisms for a bimolecular reaction of peroxynitrous acid
and nitrite in the gas phase as well as in aqueous solution, a
series of density functional calculations were performed with
the ab initio molecular dynamics19 package CPMD.20 The same
computational setup (BLYP functional,21,22 soft norm-conserving
pseudo potentials,23 periodic boundary conditions with a basis
set of plane waves expanded up to a kinetic energy cutoff of
70 Ry) was used, as in our previous studies on peroxynitrous
acid.24,25 Gas-phase geometries were optimized in an isolated26
simple cubic cell with a box length of 9.0 Å by application of
a preconditioned conjugate gradient algorithm. Reaction energy
profiles were calculated along different possible reaction
coordinates, such as the interatomic distance between the
hydroxy oxygen of peroxynitrous acid and either the nitrogen
atom (“N-attack”) or one of the two oxygens (“O-attack”) of
NO2-. Only the cis-cis conformation of peroxynitrous acid,
which is predicted by various quantum chemical calcula-
tions,24,27-29 to be the most stable rotamer in gas phase, has
been considered.
spectrometer (ThermoFinnigan, MA). The 15NO3 signal was
-
corrected for the natural abundance of [15N]nitrogen by means
of the 14NO3 peak in the same spectrum.
-
Nitrite Oxidation and Isotope Exchange Experiments.
Peroxynitrite was mixed with 0.25 to 500 mM Na15NO2 in
alkaline solution at pH ) 12; under these conditions, the mixture
is stable. The reaction was initiated by rapid mixing via fast
simultaneous injection from syringes onto a rotating stir bar in
the reaction vessel, 5 mL of the alkaline peroxynitrite solution
with 5 mL of 0.2 M sodium dihydrogen phosphate in a
thermostated flask (25 °C); the pH of the resulting solution was
5.2-5.5, a range in which peroxynitrous acid and nitrite anion
are the dominant species. After 20 s, the pH was brought to 7
by addition of sodium hydroxide.
The setup of the system in aqueous solution was chosen in
analogy to our previous study on the conformational equilibria
of peroxynitrous acid in water.25 The addition of a single nitrite
ion to this system results in a density of 1.005 g/mL, very close
to the density of pure water or an aqueous solution of
peroxynitrous acid and nitrite. The simulation system thus
consisted of one ONOOH, one NO2-, and 52 H2O molecules
The extent of nitrogen exchange between nitrogen dioxide
and nitrite was determined as follows: a 100 mL syringe purged
with argon was used to remove about 1 mL of the vapor phase

Oxidation of Nitrite
J. Phys. Chem. A, Vol. 107, No. 11, 2003 1765
TABLE 1: Experimental and Calculated Yields (in mM) of
[15N]nitrate by the Oxidation of [15N]nitrite with
Peroxynitrous Acid at pH 5.2-5.5
initial
initial
HOONO
concn
exptl
-
-
15NO2
15NO3
yield based on
yield based
concn
yielda
k′ and k′′ (eq 2) on homolysisb
10
10
10
10
100
500
0.5
1
5
10
10
10
0.02 ( 0.02
0.03 ( 0.02
0.21 ( 0.05
0.24 ( 0.05
2.5 ( 0.3
0.01-0.02
0.03
0.13-0.15
0.26-0.30
2.30-2.60
6.83-7.21
0.12
0.23
1.08
2.12
2.49-2.51
3.03-3.06
8.3 ( 0.9
a Results are given as x ( 2σ, in which x is the average of 6-9
determinations and σ the standard deviation. b Based on 40% homolysis,
and eqs 7, 11 (k11 ) 580 M-1 s-1), 13a, and 13b. If 30% is assumed,
the yields are slightly lower, for instance, 0.12 mM in the fifth column
changes to 0.08 mM.
in a periodically repeated simple cubic cell with a box length
of 12.0 Å. Hydrogen nuclei were treated as classical particles
with the mass of the deuterium isotope. The equations of motion
were solved by means of a velocity Verlet integrator with a
time step of 7 au (0.169 fs) and a fictitious mass for the
electronic degrees of freedom of 1000 au. All simulations were
performed at constant volume and a constant temperature of
300 K. During the first ∼0.5-1.0 ps of each molecular dynamics
run, the temperature of the ions was controlled by velocity
rescaling. After this equilibration period, the run was continued
with a Nose´-Hoover thermostat30 with a frequency of 500 cm-1
coupled to the ionic degrees of freedom. Rough estimates of
the relative Helmholtz energies along the reaction coordinate
for “N-attack” were performed via thermodynamic integration.31
Average values of the constraint force at eight different N-O
distances (1.8, 2.1, 2.3, 2.5, 2.8, 3.0, 3.5, and 4.0 Å) were
obtained by constrained molecular dynamics runs for each of
these states. An additional unconstrained molecular dynamics
run was performed to determine the equilibrium N-O distance
in the product. The simulation time for each run was between
1.0 and 2.5 ps resulting in a total simulation time of ca. 15 ps
for all runs (equilibration not included). The statistical errors
of the average constraint force were estimated by calculating
the standard deviation of the running averages for windows of
17 fs. These values were also used to estimate the errors of the
relative Helmholtz energies.
Figure 1. Dependence on the nitrite concentration of the observed
rate constant for the combined nitrite oxidation and isomerization of
peroxynitrous acid, at 25 °C. (A) all nitrite concentrations, peroxynitrous
acid: 0.1 mM. Circles: pH ) 5.2, ionic strength of nitrite not constant.
Squares: pH ) 3.6, sodium tetraoxochlorate added to nitrite solutions
to achieve a constant ionic strength of 0.9 M. (B) nitrite concentrations
from 0 to 0.1 M, peroxynitrous acid, 0.1 mM; pH, 5.2. The error bars
represent two times the standard deviation. The observation wavelength
was 260 nm (ꢀHOONO ≈ 600 M-1cm-1) for all measurements.
constants of this exchange have been published. The first value
was estimated from kinetic data of the nitrogen dioxide-nitrite
electron exchange reaction by the Marcus cross relationship32
and is 0.02 M-1 s-1; the second was determined experimentally33
and is 580 M-1 s-1. If the small value were correct, the exchange
Results
•
between 15NO2- and 14NO2 would not be measurable. However,
-
Product Analysis of the Reaction of HOONO with 15NO2
.
exchange at the higher rate would contribute significantly to
the [15N]nitrate yield. Although we found no experimental
evidence for this exchange, we did include the rate constant of
580 M-1 s-1 in the kinetics simulation.
The yield of [15N]nitrate increased steadily with increasing [15N]-
nitrite concentration: a solution of 500 mM 15NO2- and 10 mM
HOONO yielded 8.3 mM 15NO3-. The dependence of the [15N]-
nitrate yield on initial [15N]nitrite and peroxynitrous acid
concentrations is shown in Table 1. The yields for 10, 100, and
500 mM [15N]nitrite with 10 mM peroxynitrous acid were
determined by 15N NMR. The results at lower ratios were
obtained with mass spectrometry.
Kinetics. The observed rate constant kobs at pH ) 5.2 for the
disappearance of 0.1 mM peroxynitrous acid in the presence of
nitrite varies from 1.1-6.25 s-1 at 0-800 mM nitrite, respec-
tively (Figure 1A). The same results were obtained at pH 3.6
and at constant ionic strength (See Supporting Information,
Tables 4 and 5, and Figure 1A). Up to 100 mM nitrite, kobs
displays an apparent linear dependence on the nitrite concentra-
tion (Figure 1B). In the presence of a high excess of nitrite,
however, the relation becomes quadratic. The best fit was
obtained with eq 2
Exchange reactions that could add to the yield of [15N]nitrate
were experimentally excluded. It was shown by 15N NMR that
15NO2 and 14NO3 do not exchange within 2 days at pH ) 7
and concentrations of 10-100 mM. However, [15N]nitrite slowly
produced [15N]nitrate when left at pH ) 5.2 for some hours
via disproportionation of [15N]nitrous acid. The most critical
-
-
problem is electron exchange between 15NO2- and 14NO2 . We
•
-
- 2
kobs ) kiso + k′[NO2 ] + k′′[NO2 ]
(2)
attempted to measure the extent of this exchange in our reaction
•
-
system by passing 14NO2 diluted with argon into 15NO2
solutions, but found no detectable amounts of 15NO2-. Two rate
in which the peroxynitrous acid isomerization rate constant kiso

1766 J. Phys. Chem. A, Vol. 107, No. 11, 2003
Maurer et al.
ref
TABLE 2: Reactions Relevant for the Oxidation of Nitrite and [15N]nitrite by Hydroxyl and Nitrogen Dioxide
reaction
no.
reaction
f NO3- + H+
rate constant
HOONO
kdirect ) 0.72-0.9 s-1
kindirect ) 0.48-0.6 s-1
k7 ) 1.1 × 1010 M-1s-1
k8 ) 1.1 × 1010 M-1s-1
k9 ) 1 × 1010 M-1s-1
k10 ) 1 × 1010 M-1s-1
k11 ) 580 M-1s-1
(5)
(6)
(7)
(8)
see text
see text
•
HOONO
f HO• + NO2
HO• + 15NO2
f 15NO2• + OH-
a
a
b
b
c
c
d
d
d
-
•
HO• + NO2
f NO2• + OH-
-
HO• + 15NO2
f 15NO3- + H+/HOO15NO
f NO3- + H+/HOONO
(9)
HO• + NO2
(10)
(11)
(12)
(13)
(14)
(15)
•
-
-
•
•
NO2
+
15NO2
f NO2 +
15NO2
•
15NO2• + NO2
f 15NO2- + NO2
k12 ) 580 M-1s-1
-
2 15NO2• + H2O
f 15NO3
+
15NO2- + 2 H+
k13 ) 6.5. x 107 M-1s-1
k14 ) 6.5 × 107 M-1s-1
k15 ) 6.5 × 107 M-1s-1
-
2 NO2• + H2O
f NO3-+ NO2- + 2 H+
NO2
+
15NO2• + H2O
f 1/215NO3- + 1/2NO3- + 1/25NO2- + 1/2NO2- + 2 H+
•
a Reference 37. b Reference 35. c Reference 33. d Reference 38, more similar values are found in the Notre Dame Radiation Laboratory database:
http://www.rcdc.nd.edu.
SCHEME 1
) (1.10 ( 0.05) s-1, k′ ) (3.2 ( 0.1) M-1 s-1, and k′′ ) (4.2
( 0.3) M-2 s-1, at 25 °C. The rate constant k′ was determined
in a temperature range from 15 to 50 °C. The activation energy
of ca. 50 kJ mol-1 for the bimolecular reaction of peroxynitrous
acid with nitrite is not very accurate, since the contribution of
the isomerization reaction had to be separated; however, it is
significantly less than Ea(kiso) ) (85 ( 1) kJ Mol-1 34
. The rate
constant k′ is small, not because of the activation energy, but
because of the low frequency factor of only about 3 × 109 M-1
s-1. The calculation of the activation energy of the reaction
corresponding to k′′ is unreliable, since the contributions of the
reactions corresponding to k′ and kiso have to be factored out.
With an already inexact k′, a numerical estimate of Ea(k′′) is
meaningless. However, it can be said that Ea(k′′) is less than or
equal to Ea(k′). The associated frequency factor is also lower
than that associated with k′.
Kinetic Simulations. All kinetic traces obtained could be
satisfactorily simulated with the experimentally found expression
for kobs, eq 2. The yield of [15N]nitrate could be simulated with
the experimental rate law, eq 2. That is, the k′ path yields one
[15N]nitrate for each reaction of [15N]nitrite with peroxynitrous
acid, and the k′′ path one [15N]nitrate per two [15N]nitrites.
A simulation with a model based on up to 40% homolysis of
peroxynitrous acid into hydroxyl and nitrogen dioxide radicals
was also attempted, but did not fit the experimental results (see
Discussion). The details of this radical model (Scheme 1) are
given in footnote 35. Rate constants for the isomerization and
Figure 2. Numbering scheme used in this work and structures of (a)
the reactants, (b) the “N-attack” product, and (c) the “O-attack” product.
Bond lengths and angles of (b) and (c) are given in Table 1.
Computational Results for the Bimolecular Reaction in
Gas Phase. Formally, an SN2-like reaction with nitrite as both
the nucleophile and the leaving group in peroxynitrous acid,
leads directly to the formation of nitric acid, if the nitrite attacks
the oxygen of the hydroxy group of peroxynitrous acid with
the nitrogen lone pair (“N-attack” pathway). Our calculations
show that this reaction can indeed occur barrierless and is
exothermic by -173 kJ/mol (Figure 3). The final product is a
hydrogen-bonded complex of nitrous acid and nitrate (Figure
2b). Characteristic bond lengths and angles are given in Table
3.
The lone pairs located on the oxygens of nitrite constitute
two alternative reactive sites for a nucleophilic attack of nitrite
on the substrate peroxynitrous acid. We label these two possible
pathways “exo” and “endo”, where “endo” denotes an attack
of the lone pair closer to O5 (Figure 2). For the “exo” pathway,
no stable adduct or product structures could be localized whereas
the “endo O-attack” occurs again spontaneously and leads to a
weakly bound, planar, and almost symmetric adduct (Figure 2c
and Table 3), stabilized by 15 kJ/mol with respect to the isolated
reactants.
for radical production vary from 0.72 to 0.90 and 0.48-0.6 s-1
,
respectively.35 All other constants in the reaction scheme were
taken from literature, eqs 5-15, and are given in Table 2. For
[15NO2-] . [14NO2-], eqs 8, 10, 12, and 14 can be neglected.
For the concentration of the hydroxyl radical, a steady-state
condition can be assumed. This can be written as
-
kindirect[HOONO] ) k7 × [HO•][15NO2 ]
(16)
•
since k9 × [15NO2 ] , k7 × [15NO2-], which is true because
[15NO2 ] , [15NO2-]. Therefore, eq 9 can also be neglected.
•
The yields of 15NO2 obtained by computer simulation of the
-
radical model and the empirical rate law are listed in Table 1,
for comparison with the experimental results.
A possible explanation for the absence of any barrier in either
the “N-attack” or the “O-attack” case is the very small electronic

Oxidation of Nitrite
J. Phys. Chem. A, Vol. 107, No. 11, 2003 1767
the more relevant pathway and only this case was considered
for the simulations in aqueous solution. In analogy to the gas-
phase results, this reaction yields a hydrogen-bonded complex
of nitrous acid and nitrate, which is stable during the time scale
of the unconstrained molecular dynamics run (2.5 ps). Relative
Helmholtz energies are calculated with respect to the simulation
for which the N2-O3 distance is maximal (4.0 Å). Compared
to this configuration the product is stabilized by 89 ( 20 kJ/
mol and an activation barrier of 6 ( 12 kJ/mol is calculated.
The resulting Helmholtz energy profile is shown in Figure 3.
Although one can assume that the difference between relative
Helmholtz energies (constant volume) and relative Gibbs
energies (constant pressure) is negligible, our calculated values
cannot be directly compared to the thermodynamic value of
-167 kJ/mol15 for the isomerization of peroxynitrous to nitric
Figure 3. Reaction profiles for the “N-attack” pathway in gas phase
(diamonds) and in aqueous solution (circles). Energies are given relative
to the two product states.
-
acid, because (i) the product in the simulation is (NO3
-
HNO2)(aq) as opposed to NO3 (aq) + H+(aq) and (ii) in the initial
TABLE 3: Bond Lengths and Angles of the Two Gas Phase
Products of the Reaction of Nitrite with Peroxynitrous Acida
state the peroxynitrous acid and nitrite are too close to be both
independently solvated. It can be expected that a separation of
the product complex into nitrous acid and nitrate and complete
independent solvation of the two species leads to further
stabilization, which would be consistent with predicted stabiliza-
tion being smaller than the thermodynamic value.
“N-attack”, Figure 2b
d(O1-N1)
“O-attack”, Figure 2c
d(O1-N1)
1.23
1.37
1.08
1.49
1.32
1.27
1.26
1.25
1.26
2.09
0.98
2.10
1.26
1.25
121.5
120.7
94.3
d(N1-O2)
d(N1-O2)
d(O2-H)
d(O2-O3)
d(H-O3)
d(O3-H)
d(O3-N2)
d(O3-O4)
d(N2-O4)
d(O4-N2)
Information about the average structure of the solute and
solvent molecules can be obtained from radial distribution
functions. Although they are noisy due to relatively short
simulation times and limited statistics (there is only one
peroxynitrous acid and nitrite molecule present in the simulation
box), some structural properties emerge clearly. The results at
the maximal separation of the two solutes agree well with the
results of a previous study on peroxynitrous acid in water,25
indicating that finite size and concentration effects (the con-
centration of the two solutes in the simulation is ≈1 M each)
are relatively small.
d(N2-O5)
d(N2-O5)
(O1-N1-O2)
(N1-O2-H)
(O2-H-O3)
(H-O3-N2)
(O3-N2-O4)
(O3-N2-O5)
(O4-N2-O5)
115.5
(O1-N1-O2)
(N1-O2-O3)
(O2-O3-H)
(H-O3-O4)
(O3-O4-N2)
(O4-N2-O5)
112.0
177.1
112.4
119.0
118.3
122.6
93.0
119.8
121.5
a Bond lengths are in ångstroms and angles in degrees. See Figure
2 for the labeling scheme of the atoms.
rearrangement that is required during the reaction. An analysis
of the Kohn-Sham orbitals of the species ONOOH, HNO3 and
NO2- reveals that all one-electron orbitals of NO2- can be found
almost unchanged in ONOOH and HNO3. The only exceptions
are one of the lone pairs of one of the nitrite oxygens, which is
not found in peroxynitrous acid, and the lone pair of the nitrite
nitrogen, which is not found in nitric acid. These orbitals are
involved in forming the new σ-bond to the hydroxy group,
which is not present in nitrite. Formally, both peroxynitrous
acid and nitric acid might be considered as nitrite anions with
an OH+ group attached to one of the lone pairs.
Computational Results for the Bimolecular Reaction in
Aqueous Solution. Since solvent effects can crucially influence
the course and the energetics of a reaction, especially when ionic
species are involved, possible bimolecular reaction pathways
were also investigated in explicit aqueous solution. Several
simulations were performed to gain insight into the effects of
solvent and to estimate relative Helmholtz energies along the
reaction path. In principle, accurate calculations of relative free
energies require extended simulation times, in which all relevant
parts of the phase space can be sampled in a statistically
significant way. Due to the relatively large computational cost
of ab initio molecular dynamics simulations, affordable sampling
times are rather limited. However, the mean constraint forces
sampled at each point of the reaction coordinate are reasonably
well converged and should allow for a first semiquantitative
estimation of the relative Helmholtz energies in aqueous
solution.
The hydrogen bonding pattern of all atoms of the reactant
molecules changes quite substantially during the course of the
reaction. The radial pair correlation function gOHw of O1 with
the hydrogens of the solvent for instance shows that at the
maximal separation of the two solutes and in the product
compound, the oxygen atom O1 is hydrophobic, while at an
intermediate N2-O3 distance, a hydrogen bond to the solvent
can be observed during 90% of the simulation time. This
indicates that the solvent cage around the peroxynitrous acid
molecule found in the initial stage collapses partially during
the course of the reaction and is then built up again around
nitrous acid in the product state. These changes in the solvation
shells require a substantial rearrangement of the surrounding
solvent, in agreement with the possible existence of a small
reaction barrier in solution.
Discussion
Our results and calculations show that there is a slow, direct
reaction between nitrite and peroxynitrous acid that becomes
obvious only at higher nitrite concentrations. Given the low rate
constant, reactions of peroxynitrite from common preparations
are not affected. However, a reaction of nitrite with the reaction
product of peroxynitrous acid and an oxidizable substrate may
have to be taken into account. One could argue that the nitrite
oxidation is mediated by radicals from peroxynitrous acid, as
other workers have proposed that peroxynitrous acid should
yield up to 40%35 hydroxyl and nitrogen dioxide radicals. If
the oxidation were effected by these radicals, and if there were
In the gas phase, the “O-attack” case yields a product that is
only weakly stabilized (by 15 kJ/mol) and thus can easily decay
back to the reactants. The case of the “N-attack” is therefore
•
complete exchange between 14NO2 and 15NO2-, then a yield
of 40% 15NO2 would be predicted; in the absence of such an
-

1768 J. Phys. Chem. A, Vol. 107, No. 11, 2003
Maurer et al.
exchange, this yield would be 20%.
benzoic acid or dihydrogenperoxotrioxosulfate, rather than to
that of peroxides. The HO+ transfer hypothesis is strongly
supported by our density functional calculations of the gas-phase
reaction of nitrite and peroxynitrous acid. These calculations
demonstrate that the direct bimolecular oxidation reaction is a
barrier-less process and occurs in SN2-like fashion with the lone
electron pair of the nitrite nitrogen as the nucleophile and a
nitrite anion as the leaving group, which, overall, results in the
formal transfer of an HO+-moiety from peroxynitrous acid to
the attacking nitrite.
HO• + 15NO2- f 15NO2 + HO- 40%
(7)
•
•
•
14NO2 + 15NO2- f 15NO2 + 14NO2- 40%
(11)
•
w 15NO2 80%
•
2
15NO2 f 15N2O4
(13a)
A similar picture is obtained from first-principles molecular
dynamics simulations in aqueous solution that indicate that there
is little or no activation energy needed for the reaction to occur
once the two reactants are close enough (ca. 2.5 Å) and in a
proper orientation ( (N2-O3-O2) ca. 120° and the nitrogen
lone-pair oriented toward O3). It should be noted that, at the
starting point of the simulation, the two reactants are only 4 Å
apart and already in the correct orientation (a (N2-O3-O2)
angle of ca. 120° and the nitrogen lone-pair oriented toward O3
turn out to be essential for the reaction to occur). Although the
values are different, we conclude that the very low activation
barrier of Ea ≈ 6 kJ/mol in the simulation is consistent with
the experimental activation energy of Ea(k′) ≈ 50 kJ/mol and,
especially, with the low frequency factor observed experimen-
tally for the following reasons. The maximal N2-O3 distance
possible in the simulation is 6 Å, corresponding to half the length
of the simulation box. An estimation of the uncertainty in the
Helmholtz activation energy if the simulations were extended
to this maximal distance yields a value as large as 30 kJ/mol,
due to the large statistical error in the average constraint forces.
If the reaction rate is determined by desolvation of the reactants
and orientational requirements, an accurate determination of the
Helmholtz activation energy is far beyond the time scale of ab
initio molecular dynamics simulations, because in this case there
will be a considerable entropic contribution, an accurate
determination of which would require unfeasible simulation
times. If, on the other hand, the reaction rate is determined by
an enthalpy barrier in the actual chemical reaction of the two
reactants, e.g., due to their electronic structure, this barrier would
be seen even within the limited time scale of our simulations.
It would seem that the small rate constant observed experi-
mentally is not due to a high activation energy, but rather to
the entropic penalty associated with the desolvation of the
reacting solutes, and the necessity of bringing them close
together in the appropriate orientation for the reaction to occur.
This hypothesis is in very good agreement with the experimen-
tally determined low-frequency factor. We can also exclude a
formal fast transfer of HO+ to one of the oxygens of nitrite for
two reasons. First, such a reaction would have led to nearly
complete [15N]nitrate formation, which was not observed.
Second, since the reactants would have formed back, namely
[15N]peroxynitrous acid and [14N]nitrite, the rate of decay of
peroxynitrous acid would have remained constant and would
not have increased as a function of the nitrite concentration.
Peracids undergo bimolecular disproportionation via a formal
HO+ transfer. Earlier we showed that peroxynitrite decay may
involve in part disproportionation via a bimolecular mechanism,
rather than via homolysis of the O-O bond.3,10 The formal
transfer of HO+ to another molecule is shown here.
-
15N2O4 + H2O f 15NO2
+
15NO3- + 2H+ 40% + 40%
(13b)
However, the first four entries of Table 1 show that the yields
of 15NO2 are substantially less than 40%, as expected for
-
homolysis plus a fast transfer of the 15N-label, and even
significantly less than 20%, expected for homolysis alone. Thus,
homolysis of the O-O bond is not a viable mechanism. A
mechanism consisting of homolysis and a direct reaction would
predict yields that at low nitrite concentrations differ even more
from those determined experimentally.
The observation that the reaction is first-order and, at high
concentrations of nitrite second-order, in nitrite supports a
mechanism that involves a direct reaction between peroxynitrous
acid and nitrite. As ternary reactions are improbable, it is more
realistic to assume that the first term (k′, eq 2) is actually
composed of an equilibrium followed by a decomposition:
HOONO + NO2- a adduct-
adduct- f NO2- + NO3- + H+
(17)
(18)
and the third-order process (k′′) is invoked under conditions of
large excess of nitrite, eq 19.
adduct- + NO2- f 2 NO2- + NO3- + H+
(19)
The structure of the adduct may be similar, but given the
small theoretical activation energy not identical, to that shown
in Figure 2a, where an attack by N2 of nitrite on O3 of
peroxynitrous acid would lead to nitrate and nitrite, reaction
18, as discussed above. It is still possible that radical reactions
go on in the background, however, they do not contribute
significantly, since the radical model [15N]nitrate yields predicted
for low [15N]nitrite concentrations are much larger than those
obtained experimentally (Table 1). When the 14NO2 / NO2
• 15
-
exchange constant k11 of 580 M-1s-1 is replaced by the older
value of 0.02 M-1s-1, as seems more likely in view of our
results, the [15N]nitrate yields in the simulation of the homolysis
model in the presence of 10 mM [15N]nitrite do not change.
Only at higher [15N]nitrite concentrations is there a contribution
from the exchange reaction, no. 11. The experiments at low
[15N]nitrite concentrations suggest that radicals from homolysis
of peroxynitrous acid contribute no more than 4-5% to the
decay of peroxynitrous acid, most probably less. Inclusion of
the direct oxidation of nitrite by peroxynitrous acid in the radical
model narrows the difference between experimental and pre-
dicted results at higher nitrite concentrations, but worsens it at
lower concentrations.
From the results obtained with large excess of nitrite it may
be inferred that peroxynitrous acid undergoes a reaction typical
for a peracid, namely the apparent transfer of HO+.8,9 This is
another hint that the chemical behavior of peroxynitrous acid
is better compared with that of other peracids such as peroxy-
Acknowledgment. We thank Dr. P. L. Bounds for helpful
discussions. Supported by the ETHZ and the SNF.
Supporting Information Available: Tables of observed rate
constants of peroxynitrous acid decomposition in presence of

Oxidation of Nitrite
J. Phys. Chem. A, Vol. 107, No. 11, 2003 1769
(25) Doclo, K.; Rothlisberger, U. J. Phys. Chem. A 2000, 104, 6464-
nitrate at variable and at constant ionic strengths. This material
is available free of charge via the Internet at http://pubs.acs.org.
6469.
(26) Hockney, R. W. In Methods in Computational Physics. AdVances
in Research and Applications; Alder, B., Fernbach, S., Rotenberg, M., Eds.;
Academic Press: New York, 1970; pp 136-211.
References and Notes
(27) McGrath, M. P.; Rowland, F. S. J. Phys. Chem. 1994, 98, 1061-
(1) Beckman, J. S.; Beckman, T. W.; Chen, J.; Marshall, P. A.;
Freeman, B. A. Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 1620-1624.
(2) Huie, R. E.; Padmaja, S. Free Radical Res. Commun. 1993, 18,
195-199.
1067.
(28) Tsai, H. H.; Hamilton, T. P.; Tsai, J. H. M.; van der Woerd, M.;
Harrison, J. G.; Jablonsky, M. J.; Beckman, J. S.; Koppenol, W. H. J. Phys.
Chem. 1996, 100, 15087-15095.
(29) Houk, K. N.; Condroski, K. R.; Pryor, W. A. J. Am. Chem. Soc.
1996, 118, 13002-13006.
(30) Tuckerman, M. E.; Parrinello, M. J. Chem. Phys. 1994, 101, 1302-
1315.
(31) Sprik, M.; Cicotti, G. J. Chem. Phys. 1998, 109, 7737-7744.
(32) Ram, M. S.; Stanbury, D. M. J. Am. Chem. Soc. 1984, 106, 8136-
8142.
(33) Stanbury, D. M.; deMaine, M. M.; Goodloe, G. J. Am. Chem. Soc.
1989, 111, 5496-5498.
(34) Padmaja, S.; Kissner, R.; Bounds, P. L.; Koppenol, W. H. HelV.
Chim. Acta 1998, 81, 1201-1206.
(3) Kissner, R.; Nauser, T.; Bugnon, P.; Lye, P. G.; Koppenol, W. H.
Chem. Res. Toxicol. 1997, 10, 1285-1292. Nauser, T.; Koppenol, W. H.
J. Phys. Chem. A 2002, 106, 4084-4086.
(4) Koppenol, W. H. Redox Rep. 2001, 6, 339-341.
(5) Goodman, J. F.; Robson, P.; Wilson, E. R. Trans. Faraday Soc.
1962, 58, 1846-1851.
(6) Goodman, J. F.; Robson, P. Trans. Faraday Soc. 1963, 59, 2871-
2875.
(7) Koubek, E.; Haggett, M. L.; Bataglia, C. J.; Ibne-Rasa, K. M.; Pyun,
H. Y.; Edwards, J. O. J. Am. Chem. Soc. 1963, 85, 2263-2268.
(8) Koubek, E.; Levey, G.; Edwards, J. O. Inorg. Chem. 1964, 3, 1331-
1332.
(9) Evans, D. F.; Upton, M. W. J. Chem. Soc., Dalton Trans. 1985,
1153.
(35) Since a scavenger for the hydroxyl radical, namely, 15NO2-, was
in large excess over HOONO in most experiments, it can be assumed that
the pathways from the free radicals to nitrate and back to peroxynitrite can
be neglected. The disappearance rate constant for peroxynitrous acid at 25
°C is about kdisapp ) 1.1-1.2 s-1, measured with KOONO and [(CH3)4]-
OONO samples, and with the assumption that hydroxyl radicals react mainly
with nitrite already present and do not recombine with nitrogen dioxide to
form peroxynitrous acid, it can be said that kdirect ) 0.6kdisapp ) 0.72 s-1
and kindirect ) 0.4kdisapp ) 0.48 s-1. If, on the other hand, we assume that,
in the absence of nitrite, recombination between hydroxyl and nitrogen
dioxide radicals take place to form peroxynitrous acid and nitrate at a 1:1
ratio (Scheme 1), then
(10) Kissner, R.; Koppenol, W. H. J. Am. Chem. Soc. 2002, 124, 234-
239.
(11) Anbar, M.; Taube, H. J. Am. Chem. Soc. 1954, 76, 6243-6247.
(12) Koppenol, W. H. In Interrelations Between Free Radicals and Metal
Ions in Life Processes; Sigel, A., Sigel, H., Eds.; Marcel Dekker: New
York, 1999; pp 597-619.
(13) Bohle, D. S.; Hansert, B.; Paulson, S. C.; Smith, B. D. J. Am. Chem.
Soc. 1994, 116, 7423-7424.
(14) Bohle, D. S.; Glassbrenner, P. A.; Hansert, B. Methods Enzymol.
1996, 269, 302-311.
(15) Koppenol, W. H.; Kissner, R. Chem. Res. Toxicol. 1998, 11, 87-
90.
(16) Systematic names: O2•-, dioxide(•1-); NO•, oxidonitrogen(•);
NO2-, dioxonitrate(1-); NO3-, trioxonitrate(1-); ONOO-, oxoperoxoni-
trate(1-). The trivial names superoxide, nitrogen monoxide, nitrite, nitrate
and peroxynitrite, respectively, are allowed. Leigh, G. J., Ed. Nomenclature
of Inorganic Chemistry; Blackwell Scientific Publications: Oxford, 1990.
Koppenol, W. H. Pure Appl. Chem. 2000, 72, 437-446.
(17) Koppenol, W. H.; Kissner, R.; Beckman, J. S. Methods Enzymol.
1996, 269, 296-302.
k
kindirect
indirect and
2
)
(3)
0.4
0.6
kdisapp ) kdirect
+
kdirect
which yields
kdisapp
0.4
kdisapp
kdirect
)
) 0.90s-1
and
kindirect
)
) 0.60 s-1 (4)
0.6
1
1 +
+
2 × 0.6
0.4
2
(18) Schwarz, H. R. Numerische Mathematik; B. G. Teubner: Stuttgart,
1988; pp 381-383 and 401-406.
(19) Car, R.; Parrinello, M. Phys. ReV. Lett. 1985, 55, 2471-2474.
(20) Hutter, J.; Ballone, P.; Bernasconi, M.; Focher, P.; Fois, E.;
Goedecker, S.; Parrinello, M.; Tuckerman, M. CPMD; Max-Plank Institut
fu¨r Festko¨rperforschung: Stuttgart, and IBM Research Laboratory: Zu¨rich,
1998.
(21) Becke, A. D. Phys. ReV. A 1988, 38, 3098-3100.
(22) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785-789.
(23) Trouiller, N.; Martins, J. L. Phys. ReV. B 1991, 43, 1993-2006.
(24) Doclo, K.; Ro¨thlisberger, U. Chem. Phys. Lett. 1998, 297, 205-
210.
and an overall decay rate constant of 1.5 s-1 for peroxynitrous acid. Experimental
support for formation of both peroxynitrous acid and nitrate after recombination of
hydroxyl and nitrogen dioxide radicals was given by Mere´nyi, G.; Lind, J.; Goldstein,
S.; Czapski, G. J. Phys. Chem. A 1999, 103, 5685-5691, in contrast to an earlier
report, where only peroxynitrous acid was reported.36
(36) Gra¨tzel, M.; Henglein, A.; Taniguchi, S. Ber. Bunsen-Ges. Phys.
Chem. 1970, 94, 292-298.
(37) Barker, G. C.; Fowles, P.; Stringer, B. Trans. Faraday Soc. 1970,
66, 1509-1519.
(38) Gra¨tzel, M.; Henglein, A.; Lilie, J.; Beck, G. Ber. Bunsen-Ges. Phys.
Chem. 1969, 73, 646-653.