Metal-catalyzed anaerobic disproportionation of hydroxylamine. Role of diazene and nitroxyl intermediates in the formation of N2, N 2O, NO+, and NH3

Article abstract of DOI:10.1021/ja046724i

The catalytic disproportionation of NH₂OH has been studied in anaerobic aqueous solution, pH 6-9.3, at 25.0 °C, with Na ₃[Fe(CN)₅NH₃]·3H₂O as a precursor of the catalyst, [Fe^II(CN)₅H₂O] ^3-. The oxidation products are N₂, N₂O, and NO⁺ (bound in the nitroprusside ion, NP), and NH₃ is the reduction product. The yields of N₂/N₂O increase with pH and with the concentration of NH₂OH. Fast regime conditions involve a chain process initiated by the NH₂ radical, generated upon coordination of NH₂OH to [Fe^II(CN)₅H ₂O]^3-. NH₃ and nitroxyl, HNO, are formed in this fast process, and HNO leads to the production of N₂, N ₂O, and NP. An intermediate absorbing at 440 nm is always observed, whose formation and decay depend on the medium conditions. It was identified by UV-vis, RR, and ¹⁵NMR spectroscopies as the diazene-bound [Fe ^II(CN)₅N₂H₂]^3- ion and is formed in a competitive process with the radical path, still under the fast regime. At high pH's or NH₂OH concentrations, an inhibited regime is reached, with slow production of only N₂ and NH₃. The stable red diazene-bridged [(NC)₅FeHN=NHFe(CN)₅] ^6- ion is formed at an advanced degree of NH₂OH consumption.

Full text of DOI:10.1021/ja046724i

Published on Web 09/25/2004
Metal-Catalyzed Anaerobic Disproportionation of
Hydroxylamine. Role of Diazene and Nitroxyl Intermediates in
the Formation of N2, N2O, NO+, and NH3
Graciela E. Alluisetti,† Alejandra E. Almaraz,† Valent´ın T. Amorebieta,*,†
Fabio Doctorovich,‡ and Jose´ A. Olabe*,‡
Contribution from the Departamento de Qu´ımica, Facultad de Ciencias Exactas, UniVersidad
Nacional de Mar del Plata, Funes y R. Pen˜a, Mar del Plata B7602AYL, Argentina, and
Departamento de Qu´ımica Inorga´nica, Anal´ıtica y Qu´ımica F´ısica, INQUIMAE, Facultad de
Ciencias Exactas y Naturales, UniVersidad de Buenos Aires, Pabello´n 2, Ciudad UniVersitaria,
Buenos Aires C1428EHA, Argentina
Received June 3, 2004; E-mail: amorebie@mdp.edu.ar; olabe@qi.fcen.uba.ar
Abstract: The catalytic disproportionation of NH2OH has been studied in anaerobic aqueous solution, pH
6-9.3, at 25.0 °C, with Na3[Fe(CN)5NH3]‚3H2O as a precursor of the catalyst, [FeII(CN)5H2O]3-. The oxidation
products are N2, N2O, and NO+ (bound in the nitroprusside ion, NP), and NH3 is the reduction product. The
yields of N2/N2O increase with pH and with the concentration of NH2OH. Fast regime conditions involve a
chain process initiated by the NH2 radical, generated upon coordination of NH2OH to [FeII(CN)5H2O]3-
.
NH3 and nitroxyl, HNO, are formed in this fast process, and HNO leads to the production of N2, N2O, and
NP. An intermediate absorbing at 440 nm is always observed, whose formation and decay depend on the
medium conditions. It was identified by UV-vis, RR, and 15NMR spectroscopies as the diazene-bound
[FeII(CN)5N2H2]3- ion and is formed in a competitive process with the radical path, still under the fast regime.
At high pH’s or NH2OH concentrations, an inhibited regime is reached, with slow production of only N2 and
NH3. The stable red diazene-bridged [(NC)5FeHNdNHFe(CN)5]6- ion is formed at an advanced degree of
NH2OH consumption.
as an NO donor;7 thus, although it inhibits guanylate cyclase
activity, a vasorelaxing activation is obtained in the presence
Introduction
Hydroxylamine (NH2OH) is important in the synthesis of
caprolactam, a precursor of polyamide 6, and is also employed
as a reducing agent for photographic processes and dyeing and
in the production of oximes for use in paints, pharmaceuticals,
and pesticides.1 It is formed as an intermediate during the
reversible bacterial interconversions of ammonia into nitrite and/
or nitrate in aerobic soils.2 Nitrosomonas europaea derives
energy for growth from the oxidation of ammonia to nitrite.
Two enzymes are involved: ammonia monooxygenase, a
membrane-bound enzyme that catalyses the oxidation of am-
monia in the reaction NH3 + O2 + 2e- + 2H+ f NH2OH +
H2O, and hydroxylamine oxidoreductase (HAO), a soluble
multiheme enzyme that catalyzes the reaction: NH2OH + H2O
of catalase-H2O2.6 Reactions with oxyhemoglobin and myoglo-
bin are also considered biochemically significant.6,7
In the reactions with transition metal complexes, NH2OH
behaves as a strong reductant, leading to diverse oxidized
products or as an oxidizing agent producing ammonia, depend-
ing on the metal and spectator ligands, the pH, and the medium.8
Free radicals derived from one-electron reduction or oxidation
of NH2OH (NH2 or NHOH, respectively) have been detected
as intermediates in some of these reactions.8
(4) (a) Hendrich, M. P.; Logan, M.; Andersson, K. K.; Arciero, D. M.;
Lipscomb, J. D.; Hooper, A. B. J. Am. Chem. Soc. 1994, 116, 11961-
11968. (b) Hendrich, M. P.; Petasis, D.; Arciero, D. M.; Hooper, A. B. J.
Am. Chem. Soc. 2001, 123, 2997-3005. (c) Upadhyay, A. K.; Petasis, D.
T.; Arciero, D. M.; Hooper, A. B.; Hendrich, M. P. J. Am. Chem. Soc.
2003, 125, 1738-1747. (d) Hendrich, M. P.; Upadhyay, A. K.; Riga, J.;
Arciero, D. M.; Hooper, A. B. Biochemistry 2002, 41, 4603-4611. (e)
Cabail, M. Z.; Pacheco, A. A. Inorg. Chem. 2003, 42, 270-272.
(5) (a) Allen, J. W. A.; Watmough, N. J.; Ferguson, S. J. Nat. Struct. Biol.
2000, 7, 885-888. (b) Singh, J. Biochim. Biophys. Acta 1973, 333, 28-
36. (c) Stach, P.; Einsle, O.; Schumacher, W.; Kurun, E.; Kroneck, P. M.
H. J. Inorg. Biochem. 2000, 79, 381-385. (d) Einsle, O.; Messerschmidt,
A.; Huber, R.; Kroneck, P. M. H.; Neese, F. J. Am. Chem. Soc. 2002, 124,
11737-11745. (e) Lui, S. M.; Soriano, A.; Cowan, J. A. J. Am. Chem.
Soc. 1993, 115, 10483-10486. (f) Crane, B. R.; Siegel, L. M.; Getzoff, E.
D. Biochemistry 1997, 36, 12120-12137.
f NO2 + 4e- + 5H+.3,4 NH2OH can be reduced to NH3 by
-
some heme cd1-5a,b and cytochrome c-containing nitrite reduc-
tases (NiR),5c,d as well as by some sulfite reductases.5e,f NH2-
OH is a natural product of mammalian cells and is produced
through the decomposition of nitrosothiols.6 It is widely used
† Universidad Nacional de Mar del Plata.
‡ Universidad de Buenos Aires.
(1) Buchner, W.; Schliebs, R.; Winter, O.; Buchel, K. H. Industrial Inorganic
Chemistry; VCH Veriagsgeselischaft: 1989; pp 52-55.
(2) Hughes, M. N. The Inorganic Chemistry of Biological Processes, 2nd ed;
Wiley: New York, 1981.
(6) Feelisch, M.; Stamler, J. S., Eds. Methods in Nitric Oxide Research;
Wiley: Chichester, 1996.
(7) Wang, P. G.; Xian, M.; Tang, X.; Wu, X.; Wen, Z.; Cai, T.; Janczuk, A.
J. Chem. ReV. 2002, 102, 1091-1134.
(3) Hooper, A. B. In Autotrophic Bacteria; Schlegel, H. G., Bowien, B., Eds.;
Science Tech Publishers: 1989.
(8) Wieghardt, K. AdV. Inorg. Bioinorg. Mech. 1984, 3, 213-274.
9
13432
J. AM. CHEM. SOC. 2004, 126, 13432-13442
10.1021/ja046724i CCC: $27.50 © 2004 American Chemical Society

Disproportionation of Hydroxylamine
A R T I C L E S
was also supplied with an absolute pressure transducer, MKS Baratron
622 A, and with a mechanical stirrer. The pH was controlled with a
Hanna HI9321 pHmeter, calibrated against Merck standard buffers. The
NMR studies (15N) were performed in a Bruker 500 MHz instrument,
using Ar-flushed solutions of Na3[Fe(CN)5NH3]‚3H2O in D2O (ca. 0.06
M, pH 9), after adding a great excess of 15NH2OH. IR spectra were
obtained using either a Thermo Mattson (model Genesis II) or a Nicolet
510P FTIR instrument. Resonance Raman spectra were measured with
the 457.9-nm excitation line of an Ar+-laser (Coherent Innova 400)
using an U1000, ISA spectrograph equipped with a liquid nitrogen
cooled CCD camera. The spectral resolution was 4 cm-1, and the step
width (increment per data point) was 0.53 cm-1. Accumulation times
of the spectra were 15 s. The samples were placed in a rotating cell to
avoid photodamage by the incident beam (power of ca. 30 mW). Phenol
was determined with a Konic HPLC instrument model KNK-500-A
containing a UV-vis detector.
Slow disproportionation of NH2OH in alkaline solutions gives
NH3 and N2/N2O mixtures.9 These processes are also apparent
in the reactions of NH2OH with iron porphyrins and bacteria.5f,10
It has been suggested that this reaction type only occurs in the
presence of metal ion catalysts, involving coordination of NH2-
OH to the metal.11 In general, the redox properties of coordinated
NH2OH as compared to the free ligand have not been studied
in great detail.8
We focus on the reactivity of NH2OH bound to the
[FeII(CN)5]3- fragment.12 The pentacyano(L)ferrate(II) com-
plexes are well characterized systems, and many L ligands form
moderate to strong stable complexes (Kst ) ca. 103-105 M-1).
In many cases, the pentacyano(L)ferrate(III) derivatives have
been also characterized.13 The potential for the Fe(III),Fe(II)
couple to catalyze ligand redox is intriguing. Reactive inter-
mediates may be stabilized by the [FeII(CN)5]3- fragment during
these redox processes.14 The catalytic disproportionation of NH2-
OH using nitroprusside (NP) has been considered,15,16 but the
mechanistic proposal admits a significant revision, as will be
shown here. Our main goal is to investigate the mechanism of
the catalytic disproportionation reaction(s) of NH2OH in anaero-
bic medium, with an eye toward characterizing its coordination
Procedures and Identification of Reactants and Products. At pH’s
6-8, the dissociation of NH3 from the [Fe(CN)5NH3]3- ion (λmax
)
396 nm, ꢀ ) 450 M-1 cm-1) yields quantitatively the [Fe(CN)5H2O]3-
ion in a few minutes (λmax ) 444 nm, ꢀ ) 660 M-1 cm-1).18 This
species decays slowly through a decomposition process leading to the
[Fe(CN)6]4- and Fe2+ ions by way of successive release and recombina-
tion of cyanides (t1/2 ) ca. 2 h).19 Thus, it is sufficiently long-lived,
relative to the time scale of the main reactions under study, to trap the
free ligands present in the solutions. We checked spectroscopically on
these dissociation-decomposition processes, to put them under control
in our reaction conditions, including the absence of perturbations
associated with oxygen leakage. Due to the possible catalytic role of
the ferrous ions, we verified that the distribution of products in the
experiments to be described below did not change by adding a great
excess of EDTA as a potential scavenger.18,20
ability to the iron center. We address the conditions favoring
the formation of N2 vs N2O and NO+/NO2
.
-
Experimental Section
Reagents. Na3[Fe(CN)5NH3]‚3H2O was synthesized and purified
according to literature procedures.17 NH2OH‚HCl (Merck, 99%), 15NH2-
OH‚HCl (Isotec, 99% of 15N atoms), (N2H5)2SO4, pyrazinamide, D2O,
95% thiosuccinic acid, and 99% methyl methacrylate (Aldrich) were
used as received. N2 and Ar (AGA, UAP) were used for previous
bubbling of the solutions. The buffer solutions were prepared using
reagent grade Na2B4O7‚10H2O (Merck, 99%), KH2PO4 (Merck, 99%),
and NaOH (Anwill, 99%). Deionized distilled water was used in all
the experiments.
Physical Measurements. UV-vis spectra were recorded in a UV
2101-PC Shimadzu instrument, in the range 300-900 nm. Comple-
mentary measurements in the near-infrared region (NIR), up to 1500
nm, were obtained with a Shimadzu 3101 PC instrument. Some kinetic
runs were carried out with an Applied Photophysics RX 1000 stopped
flow (SF) accessory, linked to a Hewlett-Packard 8453 diode array
spectrophotometer. Qualitative and quantitative gas production were
conducted using a thermostated homemade flow reactor (volume, 0.07
dm3) linked to a vacuum system and to an Extrel Emba II quadrupolar
mass spectrometer through a thin thermostated capillary. The reactor
The experiments were carried out under oxygen-free conditions
(bubbling with, and mixed under N2 or Ar), avoiding the formation of
the [FeIII(CN)5H2O]2- ion (λmax ) 394 nm, ꢀ ) 750 M-1 cm-1),21
potentially produced by oxidation of [FeII(CN)5H2O]3- 18
.
The reactions were studied at 25.0 °C in buffer solution (0.1 or 1 M
in phosphates for pH 6-8 or 0.1 M in Borax for pH 9.3). The detailed
procedures were as follows. For the UV-vis kinetic experiments,
Na3[Fe(CN)5NH3]‚3H2O was dissolved in a volume of buffer solution
(final concentration (0.02-3.0) × 10-3 M). A sample was put in the
quartz cell (optical path 0.1 or 1 cm) and a desired amount of NH2OH
solution was added over the range 2 × 10-5-10-1 M. Absorbance
values at fixed wavelengths (440, 520 nm) were registered as a function
of time. The kinetic runs in the SF regime were performed under
pseudo-first-order conditions in NH2OH ((2.0-5.0) × 10-3 M, I ) 1
M, NaCl; pH 7), by monitoring, at 440 nm, the decay of the absorbance
of 1.0 × 10-4 M [Fe(CN)5H2O]3-. The measured values were fitted to
ln(At - A∞) against time. The pseudo-first-order rate constants (kobs
/
s-1) were plotted against [NH2OH]0, thus obtaining a second-order rate
constant (kf/M-1 s-1). For the measurement of the gaseous products,
0.025-0.040 dm3 of a buffered solution of NH2OH at the desired
concentration (5 × 10-4 - 2 × 10-1 M) and a measured amount of
solid Na3[Fe(CN)5NH3]‚3H2O (final concentration, (1-3) × 10-3 M)
were placed separately inside the reactor. Both reactants were mixed
after evacuation, and the time evolution of the gaseous stream was
continuously monitored. The evolution of the total pressure of the
system (psys) was obtained by numerical integration of the balance
equation: (dpsys/dt) ) (dpobs/dt) + (ke × pobs2), with pobs being the
measured pressure and ke ≈ 2 × 10-5 Torr-1 min-1, the leak constant.
The products of the reaction were N2 and N2O, and quantitative
determinations required the calibration of the mass spectrometer
(9) (a) Nast, R.; Hieber, W.; Proeschel, E. Z. Anorg. Allg. Chem. 1948, 13,
339. (b) Nast, R.; Foppel, J. Z. Anorg. Allg. Chem. 1950, 263, 310.
(10) (a) Choi, I. K.; Liu, Y.; Wei, Z.; Ryan, M. D. Inorg. Chem. 1997, 36,
3113-3118. (b) Barley, M. H.; Takeuchi, K. J.; Meyer, T. J. J. Am. Chem.
Soc. 1986, 108, 5876-5885. (c) Bazylinski, D. A.; Arkowitz, R. A.;
Hollocher, T. C. Arch. Biochem. Biophys. 1987, 259, 520-526. (d)
Castignetti, D.; Hollocher, T. C. Appl. EnVironm. Microbiol. 1982, 44, 923-
928. (e) Bagchi, S. N.; Kleiner, D. Biochim. Biophys. Acta 1990, 1041,
9-13.
(11) (a) Lunak, S.; Veprek-Siska, J. Collect. Czech. Chem. Commun. 1974, 39,
41-48. (b) Lunak, S.; Veprek-Siska, J. Collect. Czech. Chem. Commun.
1974, 39, 391-395. (c) Bonner, F. T.; Akhtar, M. J. Inorg. Chem. 1981,
20, 3155-3160.
(12) Lunak, S.; Veprek-Siska, J. Collect. Czech. Chem. Commun. 1974, 39,
2719-2723.
(13) (a) Macartney, D. H. ReV. Inorg. Chem. 1988, 6, 101-151. (b) Baraldo,
L. M.; Forlano, P.; Parise, A. R.; Slep, L. D.; Olabe, J. A. Coord. Chem.
ReV. 2001, 219-221, 881-921.
(14) Olabe, J. A. AdV. Inorg. Chem. 2004, 55, 61-126.
(15) Banyai, I.; Dozsa, L.; Beck, M. T.; Gyemant, G. J. Coord. Chem. 1996,
37, 257-270.
(18) Toma, H. E. Inorg. Chim. Acta 1975, 15, 205-211.
(19) Olabe, J. A.; Zerga, H. O. Inorg. Chem. 1983, 22, 4156-4158.
(20) Bridgart, G. J.; Waters, W. A.; Wilson, I. R. J. Chem. Soc., Dalton Trans.
1973, 1582-1584.
(16) Wolfe, S. K.; Andrade, C.; Swinehart, J. H. Inorg. Chem. 1974, 13, 2567-
2572.
(17) Kenney, D. J.; Flynn, T. P.; Gallini, J. B. J. Inorg. Nucl. Chem. 1961, 20,
75-81.
(21) Espenson, J. H.; Wolenuk, S. G., Jr. Inorg. Chem. 1972, 11, 2034-2041.
9
J. AM. CHEM. SOC. VOL. 126, NO. 41, 2004 13433

A R T I C L E S
Alluisetti et al.
Figure 1. Time evolution of the total gaseous products (in moles) during the successive addition of 5 × 10-2 M NH2OH (R0 ) 20) over 2.5 × 10-3
[Fe(CN)5H2O]3- (0.025 dm3 of buffer solution), at pH 8, I ) 2.6 M (phosphates), 25 °C.
M
sensibilities. The solubility of N2O in the buffer solutions was
considered by measuring independently its Henry constant. From the
values of the calculated psys and the solubility of N2O, a corrected molar
fraction and the concentration of the gaseous products in the condensed
phase were established. We verified that the redox capacity in the
exhausted solutions was nearly equal to the one corresponding to the
initial [FeII(CN)5NH3]3-, by titration with iodine;22 in this way, the
absence of oxidized [FeIII(CN)5L]n- or reducing (NH2OH, N2H4) species
was confirmed. Ammonia, NP, and the [(NC)5FeN2H2Fe(CN)5]6- ion
were identified and quantified in these solutions. Ammonia was
extracted by distillation under N2 current; it was identified by mass
spectrometry and quantified by acid/base titration. Alternatively, a
spectral quantification was made directly on the residual solutions by
means of its reaction product with a solution of phenol and hypochlorite
ion containing NP as catalyst.23 NP was identified in the solid residue
by IR spectroscopy24 and quantified in the solutions through the reaction
with thiosuccinic acid.25 It was also indirectly confirmed in the reacting
solutions by means of the reaction with 1,2-dimethylhydrazine,26 whose
product led to a typical EPR signal of the [Fe(CN)5NO]3- radical.27
The [(NC)5FeN2H2Fe(CN)5]6- complex was identified in some of the
exhausted solutions (pH 8-9.3, in assays performed with great excess
of NH2OH) by its absorption band at 520 nm, in agreement with a
previous characterization based on gel-permeation experiments and RR
measurements.28 We report now that a product with the same absorption
features has been obtained in an independent two-step experiment: first,
a solution of the [Fe(CN)5N2H4]3- ion29 was titrated with 2 equiv of
hexacyanoferrate(III), leading to a 440-nm absorbing product, [Fe-
(CN)5N2H2]3- (see Results). This solution was mixed with an equivalent
of [FeII(CN)5NH3]3-, with the ensuing development of the 440-nm f
520-nm conversion. Control experiments showed that the decomposition
of NH2OH was very slow in the absence of [Fe(CN)5NH3]3-, with an
min-1. This value is about 1 order of magnitude smaller than the slowest
rate measured in the presence of complex.
Numerical simulations were performed by using a specific program,
based on an adaptation of the “predictor-corrector” approach developed
by Gear.30
Results
(i) Catalytic Behavior. When solutions of excess NH2OH
are mixed with [FeII(CN)5NH3]3-, the evolution of N2 and N2O
is rapidly observed, with formation of NH3 in the residual
solutions. Clearly, a disproportionation process is underway. For
a constant concentration of complex, the catalytic behavior can
be visualized by making successive additions of NH2OH to the
exhausted reaction mixtures. The parameter R0 ) {[NH2OH]0/
[Fe(CN)5NH33-]0}, where [NH2OH]0 and [Fe(CN)5NH33-]0 are
the initial concentrations, will be used further on. Figure 1 shows
the time evolution of the number of moles of total gaseous
products for R0 ) 20 and pH 8. After an induction period of a
few minutes, a plateau region is attained in approximately 25
min, with total consumption of initial NH2OH. If the same
amount of NH2OH is added again to the mixture, a fast response
(i.e., without the induction period described above) is obtained,
and the gas evolution proceeds with a similar rate profile as in
the first cycle. The system evolves in the same catalytic way if
previously oxidized [FeIII(CN)5L]n- (L ) NH3, H2O) is added
as a starter, instead of Fe(II). On the other hand, the catalytic
process is not observed if [Fe(CN)6]4- is initially used, or
through the addition of L ligands forming very stable
[FeII(CN)5L]n- complexes, such as cyanide, pyrazinamide, and
the like.13 To confirm that the [Fe(CN)5]3- fragment remains
intact along the reaction, we carried out experiments with 15NH2-
OH, which showed unequivocally only 15N-containing products.
observed rate of formation of gaseous products lower than 10-6
M
(22) Siggia, S.; Hanna, J. G. QuantitatiVe Organic Analysis Via Functional
Groups, 4th ed.; Wiley-Interscience: 1979.
(23) Koroleff, F. In Methods of Seawater Analysis; Grasshoftt, K., Ed.; Verlag
Chemie: New York, 1976; pp 126-133.
(ii) Stoichiometry. The balances of nitrogen were performed
(24) Paliani, G.; Poletti, A.; Santucci, A. J. Mol. Struct. 1971, 8, 63-74.
(25) Szacilowski, K.; Stochel, G.; Stasicka, Z.; Kisch, H. New J. Chem. 1997,
21, 893-902.
at iron concentrations of ca. 10-3 M. The transformed NH2OH
corresponds to: n0
≈ [(4 × nN O + 3 × nN )] (see Table
NH2OH
2
2
(26) Gutie´rrez, M. M.; Amorebieta, V. T.; Estiu´, G. L.; Olabe, J. A. J. Am.
Chem. Soc. 2002, 124, 10307-10319.
1), where n0
is the number of initial moles of NH2OH
NH2OH
(27) van Voorst, J. D. W.; Hemmerich, P. J. Chem. Phys. 1966, 45, 3914-
and nN O, nN are the moles of products at the end of the reaction.
2
2
3918.
We estimate the errors in the nitrogen balances around 20%
(28) Funai, I. A.; Blesa, M. A.; Olabe, J. A. Polyhedron 1989, 8, 419-426.
(29) (a) Katz, N. E.; Olabe, J. A.; Aymonino, P. J. J. Inorg. Nucl. Chem. 1977,
39, 908-910. (b) Olabe, J. A.; Gentil, L. A. Transition Met. Chem. 1983,
8, 65-69.
(30) Gear, C. W. Comun. ACM 1971, 14, 176.
9
13434 J. AM. CHEM. SOC. VOL. 126, NO. 41, 2004

Disproportionation of Hydroxylamine
A R T I C L E S
Figure 2. Successive spectra obtained during the reaction of 1.5 × 10-4 M [Fe(CN)5H2O]3- with 9 × 10-4 M NH2OH (R0 ) 6), pH 7, I ) 1.9 M
(phosphates), 25.0 °C. Time elapsed between initial and final spectra: 5 min.
Table 1. Measured Values of N2O and N2 Concentrations
Obtained in the Catalytic Decomposition Experiments, with
Variation of pH and Concentration of Initial NH2OH, at Constant
Initial [Fe(CN)5NH3]3-: (2.0-2.5) × 10-3
M
%N
yieldd
a
b
pH
[NH2OH]0
R0
[N2O]c
[N2]c
[N2]/[N2O]c
7.0 1.25 × 10-2
7.0 5.0 × 10-2
7.0 7.5 × 10-2
7.0 0.21
5
1.3 × 10-3 2.3 × 10-3
1.8
2.1
2.1
72.0
0.4
1.5
1.6
8.3
38.8
0.3
1.3
97
96
97
103
88
103
100
93
20 4.6 × 10-3 9.8 × 10-3
30 7.0 × 10-3 1.5 × 10-2
85 1.0 × 10-3 7.2 × 10-2
8.0 5.0 × 10-3
8.0 1.25 × 10-2
8.0 5.0 × 10-2
8.0 7.5 × 10-2
8.0 0.10
2
5
8.3 × 10-4 3.6 × 10-4
1.5 × 10-3 2.3 × 10-3
20 5.6 × 10-3 9.2 × 10-3
30 2.4 × 10-3 2.0 × 10-2
40 8.0 × 10-4 3.1 × 10-2
96
100
99
9.3 1.25 × 10-2
9.3 5.0 × 10-2
9.3 7.5 × 10-2
5
2.6 × 10-3 7.0 × 10-4
20 6.2 × 10-3 8.2 × 10-3
30 7.0 × 10-4 2.3 × 10-2
32.9
95
a Molar concentration (M). b R0 ) {[NH2OH]0/[Fe(CN)5NH33-]0} c Mo-
lar concentration (M) expressed in the condensed phase. d Calculated using
{[(4[N2O] + 3[N2])/[NH2OH]0]} × 100.
Figure 3. Plot of kobs (s-1) vs [NH2OH] in the reaction of 1.0 × 10-4
M
[Fe(CN)5H2O]3-, with a controlled excess of NH2OH, I ) 1 M, NaCl; pH
7 (phosphates), 25 °C.
when the reaction products are N2, N2O, NH3, and NP and below
10% when N2 and NH3 are exclusively produced (see below
for the two described regimes, fast and slow, respectively). The
results indicate that the global operating processes for the gas
evolution can be described by the following equations:
short times. These results lead to well-behaved pseudo-first-
order plots (Figure SI 1). Figure 3 displays a linear dependence
of kobs (s-1) on [NH2OH], from which a value of kf ) 190 M-1
s-1 has been obtained. On the other hand, the intense 440-nm
absorption observed for longer times in Figure 2 reflects the
buildup of another intermediate, as shown below.
4NH2OH f N2O + 2NH3 + 3H2O
3NH2OH f N2 + NH3 + 3H2O
(iv) Evidence for Oxidation of Fe(II) to Fe(III) upon
Coordination. The following results were obtained at compar-
able or slightly deficient concentrations of NH2OH with respect
to the iron complex, given that, in the presence of excess NH2-
OH, the subsequent reactivity described above hides all evidence
of the Fe(III) chromophore. Figure 4 shows the successive UV-
vis spectra obtained after mixing equimolar solutions of
[Fe(CN)5H2O]3- and NH2OH, at pH 7.
(iii) Spectroscopic and Kinetic Evidence for the Coordina-
tion of NH2OH. Figure 2 shows the successive spectra of the
solutions affording gas evolution and formation of NH3, at pH
7. Although the initial 440-nm absorption of [Fe(CN)5H2O]3-
shifts to lower wavelengths just after mixing, a new, more
intense band develops, also at 440 nm, with increasing time.
This behavior is also observed with ca. 10-3 M [Fe(CN)5H2O]3-
,
The decay at 444 nm is accompanied by the onset of a band
centered at 395 nm, with a similar intensity, along with a
pronounced bending of the baseline for wavelengths above 700
nm. This corresponds to the tail of a new emerging broad band
in the NIR region centered at ca. 1400 nm, corresponding to
maintaining R0 > 1. The initial shift is indicative of NH2OH
coordination, as the time scale is consistent with the SF kinetic
study of the primary interaction of [Fe(CN)5H2O]3- with excess
NH2OH, monitoring the decay of [Fe(CN)5H2O]3- at pH 7, for
9
J. AM. CHEM. SOC. VOL. 126, NO. 41, 2004 13435

A R T I C L E S
Alluisetti et al.
Figure 4. Successive UV-visible spectra obtained during the reaction of 8 × 10-5 M [Fe(CN)5H2O]3- with 8 × 10-5 M NH2OH, pH 7, I ) 1.9 M
(phosphates), T ) 25 °C. Inset: same as before, pH 9.3. The shoulder at 440 nm in the initial spectrum corresponds to the reactant, and the maximum at
395 nm in the final spectrum, to the Fe(III) product. The main absorptions at ca. 400 nm in the first two spectra correspond to [Fe(CN)5NH3]3-, which is
present in equilibrium at pH 9.3. Time elapsed between initial and final spectra at both pH’s, ca. 1 h.
the mixed-valent [(NC)5FeIIINCFeII(CN)4H2O]5- ion.31 This
absorption disappears with initial 10-5 M [FeII(CN)5H2O]3-, but
the 395-nm absorption remains (bridge cleavage is achieved
upon dilution, with persistence of the Fe(III)-chromophore in
the mononuclear ion).18 The inset in Figure 4 shows similar,
slower changes observed at pH 9.3. In complementary experi-
ments at pH’s 7 and 9.3, oxygen was allowed to mix with
[FeII(CN)5H2O]3- under the same conditions as with NH2OH;
remarkably equal final spectra were obtained, as shown in Figure
SI 2. The product of reaction with O2 has been early identified
6) × 10-3 M, still maintaining R0 < 1, pH 7. The spectra of
the solutions appear quite different, yielding equilibrium
mixtures of the [Fe(CN)5H2O]3- ion and the cyano-bridged
dimer {[Fe(CN)5]2H2O}6- (λmax ) 385 nm).31 However, the
changes observed during the reaction with NH2OH are similar
to those at the lowest concentrations. We observe the increase
of the product absorbance at ca. 385 nm and, again, a
pronounced bending in the baseline for wavelengths above 700
nm. In other experiments with (2-6) × 10-3 M [Fe(CN)5H2O]3-,
R0 < 1 and pH 7, we observe the evolution of N2 but not of
N2O. In the exhausted solutions ammonia and NP are also found.
This shows that, even under addition of substoichiometric initial
NH2OH, not only coordination and reduction but also oxidation
ensues.
(v) Evidence for the Formation of NH2 Radicals. When
the reaction of NH2OH with [Fe(CN)5NH3]3- is initiated in the
presence of 0.1 M methyl methacrylate, we observed the
development of a filamentous suspension, indicative of a radical-
induced polymerization.32 This was not observed in control
experiments at pH 7-8 with aerated solutions of [Fe(CN)5NH3]3-
in the absence of NH2OH or with aerated solutions of NH2OH
in the absence of [Fe(CN)5NH3]3- Assuming a reductive
cleavage of the N-O bond in NH2OH, both amino (NH2) or
hydroxyl (OH) radicals could be formed. Each of them shows
a different reactivity toward benzene.33 NH2 does not react with
benzene under our reaction conditions, but the OH radical should
react in a fast way yielding phenol. When the reactions were
performed adding 20 mM benzene, no production of phenol
was observed. Finally, a control experiment was made by mixing
equimolar solutions of [Fe(CN)6]4- with NH2OH. No radicals
were detected, and no spectral changes occurred, showing no
reaction at all.
as [FeIII(CN)5H2O]2- 18
To confirm the oxidation state of the
.
metal, we added hydrazine to the mixed [Fe(CN)5H2O]3--NH2-
OH solution. Figure SI 3 shows a fast conversion of the 395-
nm band into a new one centered at 440 nm (ꢀ > 4500 M-1
cm-1), which further evolves slowly to a red complex absorbing
at 520 nm (ꢀ > 5500 M-1 cm-1). Along with the above changes,
the decay of the NIR absorption is observed at the earliest stages.
The experiments with hydrazine are quite relevant to the present
study. As discussed below, the spectral changes associated with
the 440 f 520-nm conversions also appear as main features of
the NH2OH-disproportionation reactions, although without the
specific addition of hydrazine. Overall, the results show that
the [Fe(CN)5H2O]3--NH2OH solutions display oxidizing activ-
ity, as expected for Fe(III) species reacting with hydrazine.28
The similar, slower changes at pH 9.3 confirm that oxidation
of the metal center occurs at all the studied pH’s. The slow
decay of the 395-nm absorptions with addition of a slight excess
of NH2OH suggests that a reductive fate of the Fe(III) species
is operative.
The influence of the concentration of [Fe(CN)5H2O]3- on the
spectroscopic observations was investigated in the range (2-
(31) (a) Emschwiller, G. C. R. Acad. Sci. Paris 1967, 265, 281-284. (b)
Emschwiller, G.; Jorgensen, C. K. Chem. Phys. Lett. 1970, 5, 561-563.
(c) Souto, M. F.; Cukiernik, F. D.; Forlano, P.; Olabe, J. A. J. Coord. Chem.
2001, 54, 343-353.
(32) (a) Kochi, J. K. Free Radicals; Wiley: 1973; Vol. II. (b) Corvaja, C.;
Fischer, H.; Giacometti, G. Z. Physik. Chem. Neue Folge 1965, 45, 1-19.
(33) Neta, P.; Maruthamuthu, P.; Carton, P. M.; Fessenden, R. W. J. Phys. Chem.
1978, 82, 1875-1878.
9
13436 J. AM. CHEM. SOC. VOL. 126, NO. 41, 2004

Disproportionation of Hydroxylamine
A R T I C L E S
Figure 5. Time evolution of the gaseous products observed for the reaction
between 2.5 × 10-3 M [Fe(CN)5H2O]3- and NH2OH: 1.25 × 10-2 M, R0
) 5, N2O (0) N2 (O); 5.0 × 10-2 M, R0 ) 20, N2O (4) N2 (3); 7.5 ×
10-2 M, R0 ) 30, N2O (+) N2 (/); pH 7, I ) 1.9 M (phosphates), T ) 25
°C. Each curve reflects the accumulation of species in the reactor.
Figure 7. Absorbance at 440 nm vs time measured for the reaction between
2.5 × 10-3 M [Fe(CN)5H2O]3- and 1.25 × 10-2 M NH2OH, R0 ) 5; (O)
pH 7, (0) pH 8, (4) pH 9.3. T ) 25 °C.
and also a similar dependence of the production of N2 and N2O
with R0 as described for pH 8. The time elapsed for reaching
the plateau region is again shorter than for pH’ s 7 and 8, ca.
30 min. This behavior, comprising the onset of a fast regime,
shows that the rate increases with pH, although we do not
observe proportional changes with the concentration of H+.
At pH 9.3, we detect a new situation for R0 ) 30, not shown
in the previous figures for lower pH’s. At the early stages N2,
N2O, and NH3 were produced (rising part for the first 10 min),
but when the transformed NH2OH was around 10%, the reaction
rate diminished noticeably. After this point, only N2 and NH3
are produced, and this slow regime persists until completion.
Turning back to Table 1, it can be seen that nearly exclusive
yields of N2 were also attained at R0 ) 85 and 40 for the lower
pH’ s 7 and 8, respectively (not shown in Figures 5 and SI 3).
These values indicate that the change of regime occurs at any
pH and is poorly sensitive to the concentration of NH2OH.
Further on, we will refer to the latter behavior as the inhibited
regime.
In most of the above experiments, we observed that NP is
also produced, with a pH-dependent final concentration. At pH
7, nearly 50% of the initial complex was transformed into NP.
This fraction became smaller, down to 30% and 10% at pH’s
8.0 and 9.3, respectively. However, NP was not found under
the inhibited regime.
It is useful to compare the above results with the parallel
changes in the intense 440-nm absorption band in experiments
at ca. 10-3 M [Fe(CN)5H2O]3-. Figure 7 shows the absorbance
curves against time for R0 ) 5, at different pH’s.
Figure 6. Time evolution of the gaseous products observed for the reaction
between 2.5 × 10-3 M [Fe(CN)5H2O]3- and NH2OH: 1.25 × 10-2 M, R0
) 5, N2O (O) N2 (0); 5.0 × 10-2 M, R0 ) 20, N2O (4) N2 (3); 7.5 ×
10-2 M, R0 ) 30, N2 (+); pH 9.3, I ) 0.2 M (borates), 25 °C. For R0 ) 30,
N2 and N2O are detected at the beginning, up to the plateau region.
(vi) Influence of pH and Concentration of NH2OH in the
Distribution of Gaseous Products. Table 1 displays the
measured concentrations of N2O and N2 obtained in experiments
with different initial concentration ratios of NH2OH/complex
(R0) at different pH’s. For nearly fixed concentrations of Fe(II)
complex, the time elapsed for the complete transformation of
increasing amounts of NH2OH (between 0.0125 and 0.2 M) is
approximately constant, ca. 50 min.
Figure 5 shows the S-shaped concentration-time profiles for
the production of [N2] and [N2O] at pH 7.0. Remarkably, the
N2/N2O ratios are close to 2 in most of the experiments (also at
pH 6, Figure SI 4), with the only exception at a very high value
of R0, 85. At pH 8, Table 1 shows that more N2O is produced
for R0 ) 2, but the yield of N2 increases gradually for greater
R0’s (range 5-30; cf. also Figure SI 5 for pH 8, with a similar
shape of the curves compared to those in Figure 5). The time
needed to finish the process is now ca. 40 min, which is slightly
smaller than the above-reported value in Figure 5. Figure 6
shows the corresponding changes in [N2O] and [N2] with time
at pH 9.3. It shows a similar timing for the induction process
The rates of development of the band and the maximum
attained intensity increase with pH. In all cases, the decay is
fast and complete in 30 min. Note that the experiment at pH 7
reveals only a modest absorption. Figure 7 clearly shows the
intermediate character of the 440-nm absorbing species in the
fast regime, which evolves along with the production of N2,
N2O, NP, and NH3. In addition, a correlation was observed
between the absorbance at 440 nm and the increase in the
production of N2, when the concentration of NH2OH was
increased at constant pH and complex concentration. Figure 8
shows the time evolution of the absorbance traces at 440 and
520 nm, for R0 ) 30, pH 9.3.
9
J. AM. CHEM. SOC. VOL. 126, NO. 41, 2004 13437

A R T I C L E S
Alluisetti et al.
(b) UV-vis. The estimation of the molar absorbance, ꢀ g
4500 M-1 cm-1 was done under conditions favoring the fast
formation of the intermediate (cf. Figure 8). The absorption
properties are consistent with similar features found in diazene
complexes.35b
(c) 15N NMR. Figure 10 shows the spectrum of a reaction
mixture of excess 15NH2OH with [Fe(CN)5NH3]3- at pH ≈ 9
containing the long-lived 440-nm absorption band. Five signals
are observed at -76, -242, -281, -339 and -373 ppm,
relative to CH3NO2 as external standard.36 The most intense
one at -281 ppm is presently assigned to free NH2OH, while
the other are in principle assigned to species arising as steady-
state intermediates. The peaks at -372 and -339 ppm can be
traced to NH3 and N2H4, respectively (the latter one could also
correspond to a structurally related intermediate), and the very
weak one at -242 ppm could be due to bound NH2OH.37
Finally, the last signal at -76 ppm can be confidently assigned
to the bound diazene intermediate.37 This last compound is
expected to generate two different signals for each nitrogen
atom, unless HNdNH is coordinated in an η2 mode or there is
a fluxional coordination exchange between both nitrogen atoms.
(d) Chemical Evidence. The 440 f 520-nm conversion in
Figure 8 is clearly indicative of a mononuclear f dinuclear
transformation involving bridged diazene. The same process has
been found with hydrazine as a precursor, in oxidizing condi-
tions (see also Experimental Section and Results, section iv).28
(e) Competitive Experiments. For R0 > 40, at 25.0 °C and
pH 8-9.3, with the reactions operating under the inhibited
regime (persistence of the 440-nm absorption), we observed that
the reduction of ethylene to ethane competes with the dispro-
portionation of NH2OH to N2 and NH3. In this situation we did
not observe the development of the band centered at 520 nm.
These hydrogenation processes are typical of diazene reactiv-
ity.38
Figure 8. Absorbance vs time measured for the reaction between 2.5 ×
10-3 M [Fe(CN)5H2O]3- and 7.5 × 10-2 M NH2OH, (R0 ) 30); (O) 440
nm and (4) 520 nm; pH 9.3, I ) 0.2 M (borates), T ) 25 °C. (1) Fast
absorbance increase for the first 10-15 min (corresponds to the initial fast
production of N2 and N2O in Figure 6 for the same period). (2) Slowly
decreasing plateau at 440 nm (corresponds to a constant rate of only N2
production, up to 300-400 min; note the change in scale in Figure 6. (3)
For t ) 600 min the 520-nm band emerges and the rate of N2 production
increases (not seen in Figure 6). (4) Up to 1000 min, A440 ) A520, and the
consumption of NH2OH is ca. 50%. The rate of N2 production still increases.
(5) The production of N2 drops abruptly. Maximum intensity of the 520-
nm band. Total consumption of NH2OH.
Note that the conditions correspond to the inhibited regime
(cf. Figure 6). The buildup of the 440-nm band, rapidly attaining
high intensity, is followed by a very slow decay, compared with
the one observed for R0 ) 5 in Figure 7 at the same pH.34 It is
instructive to compare the absorbance values in Figure 8 with
the gas production measurements in Figure 6 (same conditions,
R0 ) 30), along with the degree of consumption of NH2OH.
We describe in the legend of Figure 8 the different time-
dependent features of these indicators. Both species were very
light-sensitive, although stable in the darkness. When the
measurements were performed under illumination, the N2
production behavior was similar, but the 520-nm band intensity
was smaller and decayed down to zero along with the stop in
the production of N2. Finally, the 440-520-nm conversions were
also detectable under the fast regime conditions, where the 520-
nm absorptions behaved as transients, along with the 440-nm
decays.
(vii) Characterization of Intermediates in the Solutions
with an Intense 440 nm Absorption. (a) Resonance Raman.
Figure 9 shows the spectrum in aqueous solution, after irradia-
tion close to the maximum of the long-lived 440-nm band. The
main absorption appears at 1482 cm-1, together with several
peaks in the range 2060-2100 cm-1, which are typical of
cyanide-stretching vibrations associated with Fe(II).13 Absorp-
tions at 1639, 1556, 989, and 917 cm-1 correspond to buffer
components. The peak at 1482 cm-1 can be assigned to the
NdN stretching, by comparison with absorptions measured in
other diazene complexes.35a Stronger and conclusive evidence
on diazene coordination is given by the downward shift of the
latter absorption to 1438 cm-1, when initial 15NH2OH was used.
This is in full agreement with a doubly labeled 15N-species.35a
Discussion
The reaction of NH2OH with cyanoferrates appears quite
complex. Large quantities of NH2OH are consumed by starting
with [FeII(CN)5H2O]3- leading to reduction (NH3) and oxidation
(N2, N2O, NO+) products, with the participation of free radicals.
The appearance of N2H2 as a bound intermediate is unprec-
edented in previous studies of NH2OH redox chemistry. The
time evolution and the dependence on pH, as well as on the
NH2OH concentration for the formation of products, place a
high premium on understanding the mechanism of the dispro-
portionation process. We anticipate a simplified Scheme 1,
consistent with the known chemistry of the reactive species
involved, followed by a more detailed description of the
proposed paths.
We analyze first the conditions chosen to handle the solutions.
For the [FeII(CN)5L]3- complexes (L ) H2O, NH3, NH2OH,
(35) (a) Lehnert, N.; Wiesler, B. E.; Tuczek, F.; Hennige, A.; Sellmann, D. J.
Am. Chem. Soc. 1997, 119, 8879-8888. (b) ibid. 1997, 119, 8869-8878.
(36) The proton-decoupled 15N NMR spectrum has an inherent low sensitivity,
nearly an order of magnitude less than 13C signals. Unfavorable sensitivity
conditions in our system are given by the dynamic evolution and bubble
formation. Routine measurements with 15N NMR usually require concentra-
tions in the range 100-1000 mM, which are nearly an order of magnitude
greater than those used in the present study (cf. Silverstein, R. M.; Webster,
F. X. Spectrometric Identification of Organic Compounds, 6th ed.; Wiley:
1997).
(34) The high intensity and slow decay of the 440-nm band are observed at
other R0 values for different initial concentrations of [Fe(CN)5H2O]3-. Thus,
at pH 7 and (1-2) × 10-5 M [Fe(CN)5H2O]3-, it was observed for R0 ca.
5-6.
(37) Mason, J. Chem. ReV. 1981, 81, 205-227.
(38) (a) Stanbury, D. M. Prog. Inorg. Chem. 1998, 47, 511-561. (b) Back, R.
A. ReV. Chem. Intermed. 1984, 5, 293-323. (c) Hunig, S.; Muller, H. R.;
Thier, W. Tetrahedron Lett. 1961, 11, 353-357.
9
13438 J. AM. CHEM. SOC. VOL. 126, NO. 41, 2004

Disproportionation of Hydroxylamine
A R T I C L E S
Figure 9. Resonance Raman spectrum of the solution containing the 440-nm long-lived species. See Experimental Section for the measurement conditions.
The peak at upper-right corresponds to the most intense absorption in a complementary experiment with initially labeled hydroxylamine (15N).
Figure 10. 15N NMR spectrum of the solution containing the long-lived 440-nm absorbing species. See Experimental Section for the measurement conditions.
N2H4, N2H2, etc.), full cyanide deprotonation is ensured at pH
6.0, the lowest used in the experiments.39 The relevant pKa’s
for NH3OH+ and NH2OH are 5.9 and 13.7, respectively.8 These
values could be decreased by ca. 2 pKa units upon coordina-
tion.29 We conclude that NH2OH should be the predominant
species in our reaction conditions. For the possible oxidized
intermediates [FeIII(CN)5L]2-, deprotonation of L may occur
significantly at pH’s greater than 8.40 Upon dissolution of solid
Na3[FeII(CN)5NH3]‚3H2O, reaction 1 is onset:
[FeII(CN)5NH3]3- + H2O h [FeII(CN)5H2O]3- + NH3 (1)
The values of k1 and k-1 are 1.75 × 10-2 s-1 and 365 M-1
s-1, respectively (25.0 °C, I ) 0.1 M).41 The pH range 6-8 is
(39) (a) Toma, H. E.; Malin, J. M. Inorg. Chem. 1973, 12, 1039-1045. (b)
Malin, J. M.; Koch, R. C. Inorg. Chem. 1978, 17, 752-754.
9
J. AM. CHEM. SOC. VOL. 126, NO. 41, 2004 13439

A R T I C L E S
Scheme 1
Alluisetti et al.
+
sufficiently low for trapping most of the produced NH3 as NH4
The Fe(III) ion generated in reaction 3 goes back to Fe(II)
under reductive conditions, mainly provided by the excess NH2-
OH.20 On the other hand, the production and role of the
oxidizing NH2 radicals43 in the reduction of NH2OH by
transition metal complexes has been previously documented.8
In the Results section, we emphasized on a reaction process
that switched between a fast and a slow, inhibited regime. We
interpret it as the reaction following at least two different
pathways. One produces a certain intermediate, while another
one consumes it. Then, the intermediate concentration functions
as a trigger for the change of regime. For high concentrations
the reaction attains the fast regime. Eventually the intermediate
concentration may be decreased enough, shifting the reaction
path to the slow, inhibited regime. As anticipated in Scheme 1,
we consider the NH2 radical as the active intermediate. Then,
reaction 3 is the initiation of a chain reaction, with reactions 4
and 5 operating as propagation steps.
(pKa 9.24),42 thus allowing for the regeneration of [FeII(CN)5-
H2O]3- in the reaction medium.
Spectral and kinetic evidence on the coordination of NH2-
OH have been presented. The second-order rate constant in
Figure 3 is consistent with similar values found for the
coordination of neutral ligands into [FeII(CN)5H2O]3-, such as
NH3 (see eq 1), amines, N2H4, etc..13,41
The initial wavelength shift in Figure 2 can be interpreted
through reactions 2a and 2b, considering two alternative
coordination modes of NH2OH.8,13
[FeII(CN)5H2O]3- + NH2OH h
[FeII(CN)5O(H)NH2]3- + H2O (2a)
[FeII(CN)5H2O]3- + NH2OH h
[FeII(CN)5NH2OH]3- + H2O (2b)
NH2 + NH2OH f NH3 + NHOH
(4)
(5)
The evidence on the Fe(III) oxidized product and radical
formation implies that a follow-up to coordination must occur.
We propose reaction 3:
NHOH + NH2OH f HNO + H2O + NH2
[FeII(CN)5O(H)NH2]3- + H+ f [FeIII(CN)5H2O]2- + NH2
Reactions 4 and 5 are thermodynamically allowed in our
reaction conditions.43,44 The observed induction period (see
Figures 5, 6, SI 4, and SI 5) is also indicative of this complex
kinetics.45 According to our simulations, the induction is
controlled by the slow reactions 3 and 5. Thus, the basis is set
for the fast catalytic consumption of NH2OH through the
formation of oxidized and reduced products. HNO (nitroxyl)46
appears as a plausible intermediate in our reaction conditions,
(3)
The primary interaction described by eqs 2-3 agrees with
the slower oxidation observed with increasing pH. This picture
is in contrast with the lack of substitutional or redox reactivity
of [FeII(CN)6]4- toward NH2OH. As the redox potentials of
hexa- and pentacyano-complexes are similar (ca. 0.4 V),13 it
becomes evident that iron-oxidation requires an inner-sphere
path.
(43) Stanbury, D. M. AdV. Inorg. Chem. 1989, 33, 69-138.
(44) Free energies for reactions 4 and 5 were calculated as -106 kJ mol-1 and
-0.8 kJ mol-1, respectively. The chain propagation and its precursor,
reaction 3, reveal a close analogy between hydroxylamine and hydrogen
peroxide chemistry (Fenton-type), with Fe(II) species leading to free radicals
(cf. Jones, C. W. Applications of Hydrogen Peroxide and DeriVatiVes; The
Royal Society of Chemistry: U.K., 1999; p 44).
(40) (a) Davies, G.; Garafalo, A. R. Inorg. Chim. Acta 1976, 19, L3-L4. (b)
Davies, G.; Garafalo, A. R. Inorg. Chem. 1976, 15, 1101-1106.
(41) Toma, H. E.; Batista, A. A.; Gray, H. B. J. Am. Chem. Soc. 1982, 104,
7509-7515. The formation rate constants for the coordination of neutral
ligands into [Fe(CN)5H2O]3- decrease slightly with the ionic strength, with
reported values of ca. 2 × 102 M-1 s-1 at I ) 1 M.
(42) Fasman, G. D. Handbook of Biochemistry and Molecular Biology, Physical
and Chemical Data; CRC Press: Cleveland, OH, 1976; Vol. 1.
(45) Benson, S. W. Foundations of Chemical Kinetics; McGraw-Hill Book
Co.: New York, 1960.
9
13440 J. AM. CHEM. SOC. VOL. 126, NO. 41, 2004

Disproportionation of Hydroxylamine
A R T I C L E S
a precursor of N2O and N2 according to the fast reactions 6
and 7:8
The [FeII(CN)5N2H2]3- ion traps most of the iron sites, the NH2
radical concentration has been sufficiently depleted through the
termination steps, and the reaction develops very slowly (see
Figure 6). In these conditions, bound diazene decays mainly
through reaction 9.38,51
2HNO f N2O + H2O
(6)
(7)
HNO + NH2OH f N2 + 2H2O
[FeII(CN)5N2H2]3- + [FeII(CN)5O(H)NH2]3-
f
The set of reactions 4-7 sustain the NH2OH processing and
agree with the stoichiometry found in the fast regime conditions.
The relative enrichment of N2 with an increasing concentration
of initial NH2OH is consistent with reaction 7 being favored
over reaction 6 (cf. Table 1). The production of N2O could only
be realized for R0’s > 1. No production was observed upon
treatment with substoichiometric NH2OH, given that NP and
N2 did form, and this can be traced to the rapid depletion of
NH2OH.
In parallel with the radical path, the formation and decay of
the 440-nm band has been observed (Figures 2 and 7). We have
already identified the species absorbing at this wavelength as
[FeII(CN)5N2H2]3-, and we propose that it is formed through
reaction 8:
2 [FeII(CN)5H2O]3- + N2 + NH3 (9)
Reactions 8 and 9 satisfy the equimolar stoichiometry found
for the N2 and NH3 production in the inhibited regime and also
sustain the slow molecular path for NH2OH processing antici-
pated in the scheme.
Figure 8 indicates that after the slowly decreasing plateau
corresponding to the inhibited regime, the [FeII(CN)5N2H2]3-
ion becomes the precursor of a species absorbing at 520 nm,
which we already assigned to the bridged-diazene dimer. Its
RR spectrum showed a band at 1440 cm-1 after irradiation at
515 nm,28 and this frequency is consistently lower than the one
assigned above to [FeII(CN)5N2H2]3-. Therefore, we propose
reaction 10, consistent with a specific test described in the
Experimental Section:
[FeII(CN)5NH2OH]3- + NH2OH f
[FeII(CN)5N2H2]3- + 2H2O (8)
[FeII(CN)5N2H2]3- + [FeII(CN)5H2O]3-
h
The 440-nm band was previously traced to the [FeII(CN)5-
NH2OH]3- complex.12,15,49 However, the d-d transitions in
[FeII(CN)5L]3- complexes with L ) N-binding saturated ligands,
NH3, amines and diamines, N2H4, appear all around 400 nm,
[(NC)5FeIIHNdNHFeII(CN)5]6- + H2O (10)
The dimer is quite stable although light sensitive and
decomposes giving N2 if oxygen or other oxidants are present.
The appearance of the dimer goes in parallel with a faster
evolution of N2 (see legend in Figure 8). This can be explained
because more [FeII(CN)5H2O]3- and oxidant are available as
far as NH2OH is being consumed. The [FeII(CN)5H2O]3- ion
binds competitively to the [FeII(CN)5N2H2]3- complex and
simultaneously promotes more NH2OH decomposition (through
reaction 2a) and subsequent N2 production.
with ꢀ ) ca. 500 M-1 cm-1 13
. The intense 440-nm band can be
more reasonably assigned to a metal-to-ligand charge-transfer
(MLCT) transition from Fe(II) to the vacant π* orbital in the
[FeII(CN)5N2H2]3- ion.35b There are several examples of well
characterized monodentate coordination of N2H2 in low-spin
d6 complexes.50 The [FeII(CN)5N2H2]3- ion must be moderately
inert toward dissociation of N2H2, according to its σ-π bonding
properties,35 but it should be destroyed by the oxidants existing
in the reaction pool.
To explain fully the diversified product pattern, we come
again to the radical path. As well as leading to N2 and N2O
production, HNO can be oxidized to NO+. The identification
of NP as an oxidation product at an early stage of the process
provides complementary evidence of the formation of HNO46
in the propagation chain. The possibility of obtaining NP from
NH2OH has been anticipated.52 Successive two-electron pro-
cesses have been proposed for the oxidation by high-valent
aqueous metal ions8,53 and in the enzymatic HAO-mediated
conversions of NH2OH to nitrite.4 Good stoichiometric evidence
has been found for the participation of [FeII(CN)5HNO]3- in
the full six-electron reduction of NP to NH3.26 Theoretical
calculations have been reported providing evidence on
the stability, geometry, and spectroscopy of [FeII(CN)5-
The termination reactions for the NH2 radical can be
envisaged through a secondary interaction with [FeII(CN)5L]3-
reductants (L ) H2O, N2H2, etc.) or through the less probable
(in most of our experimental conditions) self-reaction giving
hydrazine.47
Equation 8 describes a competitive reaction, regarding eq 3,
for the decay of [FeII(CN)5O(H)NH2]3- 48
and it becomes
,
predominant with increasing pH’s and NH2OH-concentration.
(46) (a) Bonner, F. T.; Dzelzkalns, L. S.; Bonucci, J. A. Inorg. Chem. 1978,
17, 2487-2494. (b) Akhtar, M. J.; Bonner, F. T.; Borer, A.; Cooke, I.;
Hughes, M. N. Inorg. Chem. 1987, 26, 4379-4382. (c) Southern, J. S.;
Hillhouse, G. L. J. Am. Chem. Soc. 1997, 119, 12406-12407. (d)
Bartberger, M. D.; Fukuto, J. M.; Houk, K. N. Proc. Natl. Acad. Sci. U.S.A.
2001, 98, 2194-2198. (e) Shafirovich, V.; Lymar, S. V. Proc. Natl. Acad.
Sci. U.S.A. 2002, 99, 7340-7345. (f) Sulc, F.; Immoos, C. E.; Pervitsky,
D.; Farmer, P. J. J. Am. Chem. Soc. 2004, 126, 1096-1101.
HNO]3- 54
.
(47) Laszlo, B.; Alfassi, Z. B.; Neta, P.; Huie, R. E. J. Phys. Chem. A 1998,
102, 8498-8504.
(48) Equation 8 comprises a condensation process for NH2OH, with possible
unknown intermediate formation. In eq 8, the competition with reaction 3
could be better appreciated if the O-bound complex would be the reactant.
However, the N-bound complex was considered a more realistic approach
to diazene formation, given that the competition with reaction 3 is still
established through reactions 2a and b.
(49) Mulvey, D.; Waters, W. A. J. Chem. Soc., Dalton Trans. 1975, 951-959.
(50) (a) Smith, M. R., III; Cheng, T. Y.; Hillhouse, G. L. J. Am. Chem. Soc.
1993, 115, 8638-8642. (b) Cheng, T. Y.; Ponce, A.; Rheingold, A. L.;
Hillhouse, G. L. Angew. Chem., Int. Ed. Engl. 1994, 33, 657-659. (c)
Marchenko, A. V.; Vedernikow, A. N.; Dye, D. F.; Pink, M.; Zaleski, J.
M.; Caulton, K. G. Inorg. Chem. 2004, 43, 351-360.
(51) Equation 9 has been written after calculating that the one-electron outer
sphere oxidation of [FeII(CN)5N2H2]3- by free NH2OH is endergonic. The
one-electron reduction potential for the couple NH2OH/NH2, OH- is -0.6
V, estimated considering the following ∆G0 values: -23.4, 192, and
-157.8 kJ mol-1 for NH2OH, NH2 radical, fand OH- ion in aqueous
solution, respectively. The potential for the [FeIII,II(CN)5N2H2]2-,3- couple
should be of ca. 0.5 V.13
(52) Caulton, K. G. Coord. Chem. ReV. 1975, 14, 317-355.
(53) Johnson, M. D.; Hornstein, B. J. Inorg. Chem. 2003, 42, 6923-6928.
(54) Gonza´lez Lebrero, M. C.; Scherlis, D. A.; Estiu´, G. L.; Olabe, J. A.; Estrin,
D. A. Inorg. Chem. 2001, 40, 4127-4133.
9
J. AM. CHEM. SOC. VOL. 126, NO. 41, 2004 13441

A R T I C L E S
Alluisetti et al.
The decreasing yield of NP in the exhausted solutions for
increasing pH’s can be traced to the incidence of the well-
established reaction of NP with NH2OH.16
initial precursor, great amounts of NH2OH can be processed,
with [FeII(CN)5H2O]3- as the cycling catalyst. Catalysis is
inhibited by strongly binding ligands such as cyanide.
Coordination of NH2OH to [FeII(CN)5H2O]3- and production
of NH2 sets the basis for establishing a fast radical path leading
to reduced and oxidized products, NH3 and nitroxyl, HNO,
respectively. The latter species is the precursor for the formation
of N2 and N2O and can be further oxidized to NP. A slow
molecular path, accomplished simultaneously with the above-
described one, is operative for high concentrations of NH2OH.
The NH2 radical concentration becomes low and the [FeII-
(CN)5N2H2]3- ion accumulates in the medium, trapping most
of the iron sites. Thus, it can be characterized by chemical and
spectroscopic tools (RR, UV-vis, 15NMR). In certain condi-
tions, the [(NC)5FeIIHNdNHFeII(CN)5]6- ion remains in the
exhausted solutions after consumption of NH2OH. The inter-
mediacy of well characterized diazene complexes is a novel
result in the redox chemistry of NH2OH.
The production of minor quantities of NP compared to the
other products leads to some resemblance with the behavior of
the HAO enzyme. The formation of NP in our reaction
conditions, although mechanistically quite significant, is not
catalytic however for the production of nitrite. Moreover, given
the strong Fe-NO+ bond in NP, the catalytic iron site becomes
partially blocked, because the NO+/NO2- conversion and release
of nitrite occur at pH’s > 10. This is in contrast with HAO,
which presumably releases nitrite at lower pH’s, thus allowing
for further coordination and processing of NH2OH.
[FeII(CN)5NO]2- + NH2OH + OH- f
[FeII(CN)5H2O]3- + N2O + H2O (11)
Our previous analysis has been done considering mononuclear
complexes, although equilibrium mixtures of [FeII(CN)5H2O]3-
and [(NC)5FeIINCFeII(CN)4H2O]6- exist in the solutions under
some of our experimental conditions.31,55 The first ion predomi-
nates at ca. (1-3) × 10-5 M, with increasing amounts of the
dimer for greater concentrations. In oxidative conditions and a
low concentration of NH2OH, mixed-valent dinuclear complexes
are produced.31 In excess of NH2OH however the mixed valent
complexes were not observed at all, suggesting a bridge cleavage
and the predominance of mononuclear species. In either case,
the coordination of substrates and subsequent reactions involving
the labile iron-aqua site in the mixed-valent dimers should be
similar to those proposed for the mononuclear ions.
Finally, we refer to the simulation results and interpretation.
A simplified mechanism has been employed, emphasizing in
the main oxidation and reduction products, with omission of
the NP and bridged-diazene formation reactions. All the possible
reactions presented explicitly or implicitly in the Discussion have
been considered, as shown in Scheme SI 6, with the values of
the rate constants extracted from the literature, as well as the
ones which were considered as fitting parameters. We choosed
pH 7 for the calculations, having in mind that the observed
behavior was similar at the different pH’s. Through this
procedure, the importance and the relative sensibility of the
reactions were studied, allowing for the establishment of a
minimum scheme. The main focusing requirements deal with
the induction periods, the distribution of products (N2, N2O,
and NH3) and the conditions for the change of regime. The
output in Figures SI 7 and 8 reproduces qualitatively the most
important focused points, namely the following: (a) At a
constant iron concentration, different initial concentrations of
NH2OH (ca. 0.025 M to 0.2 M) are processed in approximately
the same time period. (b) The distribution of gaseous products
shows enrichment in N2 with increasing concentration of initial
NH2OH, up to a critical value after which the production is
much slower. (c) The global yield of NH3 is well reproduced.56
Acknowledgment. This work was supported by the Universi-
ties of Buenos Aires and Mar del Plata, the Government
Agencies ANPCYT and CONICET, and by the Volkswagen-
Stiftung (Germany). We appreciate the aid of Professor Daniel
Murgida (University of Berlin) in performing the RR measure-
ments and the collaboration of Dr. Alejandro R. Parise with
some experiments. V.T.A., F.D., and J.A.O. are members of
the scientific staff of CONICET.
Supporting Information Available: Figures: SI 1 (Loga-
rithmic plots for the reaction of [FeII(CN)5H2O]3- with excess
NH2OH). SI 2 (Reaction of O2 with [FeII(CN)5H2O]3-, in similar
conditions as with NH2OH). SI 3 (Reaction of hydrazine with
the reaction product of [FeII(CN)5H2O]3- and NH2OH). SI 4
(Time evolution of gaseous products at pH 6). SI 5 (Time
evolution of gaseous products at pH 8). Scheme SI 6 (List of
reactions and associated rate constants employed in the simula-
tions). Figures: SI 7 (Simulation of time evolution of the
concentration of N2 and N2O at different values of R0). SI 8
(Simulation of time evolution of the concentration of [FeII-
(CN)5N2H2]3- at different values of R0). This material is
available free of charge via the Internet at http://pubs.acs.org.
Conclusions
NH2OH transforms to NH3 and diverse oxidation products
(N2, N2O, NO+) in different yields according to pH and
concentration of initial substrate. With [FeII(CN)5NH3]3- as the
(55) James, A. D.; Murray, R. S. J. Chem. Soc., Dalton Trans. 1977, 326-329.
(56) Ongoing experiments aim to better explore the transition region through a
broader variation of pH’s and concentrations of hydroxylamine. These
results will be reported in due course.
JA046724I
9
13442 J. AM. CHEM. SOC. VOL. 126, NO. 41, 2004