A surface effect allows HNO/NO discrimination by a cobalt porphyrin bound to gold

Article abstract of DOI:10.1021/ic1007022

Nitroxyl (HNO) is a small short-lived molecule for which it has been suggested that it could be produced, under certain cofactors conditions, by nitric oxide (NO) synthases. Biologically relevant targets of HNO are heme proteins, thiols, molecular oxygen, NO, and HNO itself. Given the overlap of the targets and reactivity between NO and HNO, it is very difficult to discriminate their physiopathological role conclusively, and accurate discrimination between them still remains critical for interpretation of the ongoing research in this field. The high reactivity and stability of cobalt(II) porphyrins toward NO and the easy and efficient way of covalently joining porphyrins to electrodes through S-Au bonds prompted us to test cobalt(II) 5,10,15,20-tetrakis[3-(p- acetylthiopropoxy)phenyl]porphyrin [Co(P)], as a possible candidate for the electrochemical discrimination of both species. For this purpose, first, we studied the reaction between NO, NO donors, and commonly used HNO donors, with Co^II(P) and Co^III(P). Second, we covalently attached Co^II(P) to gold electrodes and characterized its redox and structural properties by electrochemical techniques as well as scanning tunneling microscopy, X-ray photoelectron spectroscopy, and solid-state density functional theory calculations. Finally, we studied electrochemically the NO and HNO donor reactions with the electrode-bound Co(P). Our results show that Co(P) is positioned over the gold surface in a lying-down configuration, and a surface effect is observed that decreases the Co^III(P) (but not Co ^III(P)NO^-) redox potential by 0.4 V. Using this information and when the potential is fixed to values that oxidize Co ^III(P)NO^- (0.8 V vs SCE), HNO can be detected by amperometric techniques. Under these conditions, Co(P) is able to discriminate between HNO and NO donors, reacting with the former in a fast, efficient, and selective manner with concomitant formation of the Co^III(P)NO ^- complex, while it is inert or reacts very slowly with NO donors.

Full text of DOI:10.1021/ic1007022

Inorg. Chem. 2010, 49, 6955–6966 6955
DOI: 10.1021/ic1007022
A Surface Effect Allows HNO/NO Discrimination by a Cobalt Porphyrin
Bound to Gold
†
‡
‡
†
†
ꢀ
ꢀ
ꢀ
Sebastian A. Suarez, Mariano H. Fonticelli, Aldo A. Rubert, Ezequiel de la Llave, Damian Scherlis,
‡
Roberto C. Salvarezza, Marcelo A. Martı,†,§ and Fabio Doctorovich*,†
´
†
´
ꢀ
´
´
´
Departamento de Quımica Inorganica, Analıtica y Quımica Fısica/INQUIMAE-CONICET, Facultad de
ꢀ
Ciencias Exactas y Naturales, Universidad de Buenos Aires, Ciudad Universitaria, Pabellon II, Buenos Aires
§
´
C1428EHA, Argentina, Departamento de Quımica Biologica, Facultad de Ciencias Exactas y Naturales,
Universidad de Buenos Aires, Ciudad Universitaria, Pab. II (1428), Buenos Aires, Argentina, and ‡Instituto de
´
ꢀ
Investigaciones Fisicoquımicas, Teoricas y Aplicadas (INIFTA), Universidad Nacional de La Plata-CONICET,
Sucursal 4 Casilla de Correo 16, 1900 La Plata, Argentina
Received April 12, 2010
Nitroxyl (HNO) is a small short-lived molecule for which it has been suggested that it could be produced, under certain
cofactors conditions, by nitric oxide (NO) synthases. Biologically relevant targets of HNO are heme proteins, thiols,
molecular oxygen, NO, and HNO itself. Given the overlap of the targets and reactivity between NO and HNO, it is very
difficult to discriminate their physiopathological role conclusively, and accurate discrimination between them still remains
critical for interpretation of the ongoing research in this field. The high reactivity and stability of cobalt(II) porphyrins toward
NO and the easy and efficient way of covalently joining porphyrins to electrodes through S-Au bonds prompted us to test
cobalt(II) 5,10,15,20-tetrakis[3-(p-acetylthiopropoxy)phenyl]porphyrin [Co(P)], as a possible candidate for the electro-
chemical discrimination of both species. For this purpose, first, we studied the reaction between NO, NO donors, and
commonly used HNO donors, with CoII(P) and CoIII(P). Second, we covalently attached CoII(P) to gold electrodes and
characterized its redox and structural properties by electrochemical techniques as well as scanning tunneling microscopy,
X-ray photoelectron spectroscopy, and solid-state density functional theory calculations. Finally, we studied electro-
chemically the NO and HNO donor reactions with the electrode-bound Co(P). Ourresults show that Co(P) is positioned
over the gold surface in a lying-down configuration, and a surface effect is observed that decreases the CoIII(P)
(but not CoIII(P)NO-) redox potential by 0.4 V. Using this information and when the potential is fixed to values that
oxidize CoIII(P)NO- (0.8 V vs SCE), HNO can be detected by amperometric techniques. Under these conditions, Co(P) is
able to discriminate between HNO and NO donors, reacting with the former in a fast, efficient, and selective manner with
concomitant formation of the CoIII(P)NO- complex, while it is inert or reacts very slowly with NO donors.
Introduction
proteins, the NO synthases.4-6 Furthermore, several studies
have proven distinct pharmacological effects for NO and
HNO donors and have shown that the lifetime of HNO in
living tissues is probably much longer than previously as-
sumed. This could imply the coexistence of HNO and NO in
certain tissues.7-9 The biologically relevant targets of HNO
are mainly heme proteins and thiols.10,11 HNO is also known
Nitroxyl (HNO), the one-electron reduction product of
nitric oxide (NO), is a small short-lived molecule that has
been intensively investigated in the past decade because of its
singular chemical and biological properties.1-3 More specifi-
cally, the reactions of HNO with synthetic metalloporphyrin
complexes containing divalent transition metals have been an
intense research area in recent years. From the physiological
perspective, encouraging studies suggest that HNO can be
produced under certain cofactor conditions by NO-producing
(4) Adak, S.; Wang, Q.; Stuehr, D. J. J. Biol. Chem. 2000, 275, 33554–
33561.
(5) Hobbs, A. J.; Fukuto, J. M.; Ignarro, L. J. Proc. Natl. Acad. Sci. 1994,
91, 10992–10996.
(6) Schmidt, H. H. H. W.; Hofmann, H.; Schindler, U.; Shutenko, Z. S.;
Cunningham, D. D.; Feelisch, M. Proc. Natl. Acad. Sci. 1996, 93, 14492–
14497.
*To whom correspondence should be addressed. E-mail: doctorovich@
qi.fcen.uba.ar. Fax: 54-11-4576-3341.
(7) Pagliaro, P. Life Sci. 2003, 73, 2137–2149.
(1) Farmer, P. J.; Sulc, F. J. Inorg. Biochem. 2005, 99, 166–184.
(2) Miranda, K. M. Coord. Chem. Rev. 2005, 249, 433–455.
(3) Switzer, C. H.; Flores-Santana, W.; Mancardi, D.; Donzelli, S.;
Basudhar, D.; Ridnour, L. A.; Miranda, K. M.; Fukuto, J. M.; Paolocci,
N.; Wink, D. A. Biochim. Biophys. Acta 2009, 1787, 835–840.
(8) Wink, D. A.; Miranda, K. M.; Katori, T.; Mancardi, D.; Thomas,
D. D.; Ridnour, L.; Espey, M. G.; Feelisch, M.; Colton, C. A.; Fukuto, J. M.;
Pagliaro, P.; Kass, D. A.; Paolocci, N. Am. J. Physiol. Heart Circ. Physiol.
2003, 285, 2264–2276.
r
2010 American Chemical Society
Published on Web 07/06/2010
pubs.acs.org/IC

ꢀ
Suarez et al.
6956 Inorganic Chemistry, Vol. 49, No. 15, 2010
to react with molecular oxygen, NO, and HNO itself.12 These
reactions are probably the main sinks of HNO in physiolo-
gical media.9 Given the overlap of the molecular targets and
reactivity between NO and HNO, it is very difficult to
discriminate their physiopathological role conclusively. Cy-
steine is used in some HNO/NO blocking experiments, but
discrimination of both species requires characterization of
the final products and cannot be done in situ.13,14 Clearly,
unequivocal discrimination of HNO and NO still remains
critical for interpretation of the ongoing research in this field.
As stated above, a number of studies of the HNO reactivity
have been performed with metalloporphyrin complexes, gen-
erally iron porphyrins, commonly used as heme group models.
NO is able to react fast with ferrous and also ferric porphyr-
ins, yielding the corresponding {FeNO}6/7 complexes accord-
ing to the Enermark and Feltham notation,15-17 while HNO
reacts with isolated ferric porphyrins and globins to yield
{FeNO}7 complexes.18-21 Interestingly, HNO reacts also
with ferrous myoglobin and other globins, yielding stable
{FeHNO}8 complexes.22,23 The promiscuity of iron porphy-
rins as well as the reductive nitrosylation reaction that
produces {FeNO}7 complexes from ferric porphyrins makes
them a bad choice for HNO/NO discrimination. However,
manganese(III) porphyrins suffer reductive nitrosylation
much slower than their iron(III) analogues, allowing one to
discriminateHNO from NO in solution.24 On the basis of this
data, Dobmeier et al.25 designed a Xerogel optical sensor film
for quantitative detection of HNO with an estimated dy-
namic range of 24-290 nM. However, because of the fact
that the UV-vis measurements are done in a wavelength
range in which all heme proteins strongly absorb, this method
cannot be used for most in vitro or in vivo studies. Lippard
et al. have developed a fluorescence-based system that uses a
copper(II) metallopolymer and is able to discriminate HNO
over NO, although it needs a high concentration of HNO
donor (ca. 0.3 mM).26
Although not studied as much as their iron analogues,
cobalt porphyrins also gained prominent attention for the
binding and activation of NO. The substitution of iron by
cobalt in heme proteins and synthetic porphyrins is an active
research field, with synthetic chemistry and biological per-
spectives. For example, (TPP)Co(NO) (TPP = tetraphenyl-
porphyrin) has been explored as an isoelectronic model for
oxygenated protoheme,27 and the NO adducts of cobalt-
substituted myoglobin and hemoglobin have been character-
ized by electron paramagnetic resonance and UV-vis.28,29
Several kinetic studies have been done on the NO reaction
with cobalt(II) and cobalt(III) porphyrins.1,30 The specific
association rate constants of NO with cobalt(II) porphyrins
were estimated to be kon = 2.0 ꢀ 109 M-1 s-1, while the low
value of koff = 1.5 ꢀ 10-4 M-1 s-1 shows the stability of the
CoIINO complex, which is actually better represented as a
CoIIINO- compound.30 Interestingly, the association rate of
NO with cobalt(III) porphyrins is notably smaller.30 Last, the
redox properties of the NO- (nitroxyl anion) cobalt porphy-
rin complexes (TPP)Co(NO) were also studied in organic
solvents, such as dichloromethane. Stable complexes of
[(TPP)Co(NO)]- and [(TPP)Co(NO)]þ could be obtained
at the electrode surfaces,31 and redox potentials for CoII-
( p-OCH3)TPP have been reported in several solvents.32,33
Metalloporphyrins have also been widely studied for their
use in technical applications such as gas sensors34,35 and cat-
alysts36-38 when coupled to a surface. Ordered monolayers of
porphyrins and metalloporphyrins on inert metal surfaces
have been used as convenient model systems for the study of
the reactivity and electronic structure of these devices.39-43
These types of thin-film systems have been used to study
direct metalation39,42,43 and ligand coordination reactions.41
Interestingly, a recent X-ray photoelectron spectroscopy
(XPS) study suggested direct electron transfer from a silver
surface to the Co ion from an adsorbed cobalt porphyrin,
which could be modulated by coordination of NO to the
(9) Miranda, K. M.; Paolocci, N.; Katori, T.; Thomas, D. D.; Ford, E.;
Bartberger, M. D.; Espey, M. G.; Kass, D. A.; Feelisch, M.; Fukuto, J. M.;
Wink, D. A. Proc. Natl. Acad. Sci. 2003, 100, 9196–9201.
(10) Doyle, M.; Mahapatro, S. N.; Broene, R. D.; Guy, J. K. J. Am.
Chem. Soc. 1988, 110, 593–599.
(11) Miranda, K. M.; Nims, R. W.; Thomas, D. D.; Espey, M. G.; Citrin,
D.; Bartberger, M. D.; Paolocci, N.; Fukuto, J. M.; Feelisch, M.; Wink,
D. A. J. Inorg. Biochem. 2003, 93, 52–60.
(12) Shafirovich, V.; Lymar, S. V. Proc. Natl. Acad. Sci. 2002, 99, 7340–
7345.
(13) Pino, R. Z.; Feelisch, M. Biochem. Biophys. Res. Commun. 1994, 201,
54–62.
(14) Shoeman, D. W.; Shirota, F. N.; DeMaster, E. G.; Nagasawa, H. T.
Alcohol 2000, 20, 55–59.
(15) Enemark, J. H.; Feltham, R. D. J. Am. Chem. Soc. 1974, 96, 5002–
5004.
(16) Westcott, B. L.; Enemark, J. H. Inorganic Electronic Structure and
Spectroscopy; John Wiley & Sons: New York, 1999; Vol. 2.
(17) Laverman, L. E.; Ford, P. C. J. Am. Chem. Soc. 2001, 123, 11614–
11622.
(26) Tennyson, A.; Do, L.; Smith, R.; Lippard, S. J. Polyhedron 2007, 26,
4625–4630.
(27) Scheidt, W. R.; Hoard, J. L. J. Am. Chem. Soc. 1973, 95, 8281–8288.
(28) Hori, H.; Ikeda-Saito, M.; Leigh, J. S.; Yonetani, T. Biochemistry
1982, 21, 1431–1437.
(29) Yu, N.; Thompson, H.; Mizukami, H.; Gersonde, K. Eur. J.
Biochem. 1986, 159, 129–132.
(30) Roncaroli, F.; van Eldik, R. J. Am. Chem. Soc. 2006, 128, 8042–8053.
(31) Kelly, S.; Lancon, D.; Kadish, K. M. Inorg. Chem. 1984, 23, 1451–
1458.
(32) Walker, F. A. J. Am. Chem. Soc. 1970, 92, 4235–4244.
(33) Walker, F. A.; Beroiz, D.; Kadish, K. M. J. Am. Chem. Soc. 1976, 98,
3484–3489.
(34) Guillaud, G.; Simon, J.; Germain, J. P. Coord. Chem. Rev. 1998,
178-180(2), 1433–1484.
(35) Rakow, N. A.; Suslick, K. S. Nature 2000, 406, 710–713.
(36) Mochida, I.; Suetsugu, K.; Fujitsu, H.; Takeshita, K. J. Phys. Chem.
1983, 87, 1524–1529.
(18) Bari, S. E.; Marti, M. A.; Amorebieta, V. T.; Estrin, D. A.;
Doctorovich, F. J. Am. Chem. Soc. 2003, 125, 15272–15273.
ꢀ
(19) Suarez, S. A.; Marti, M. A.; De Biase, P. M.; Estrin, D. A.; Bari, S. E.;
Doctorovich, F. Polyhedron 2007, 26, 4673–4679.
(37) Brule, E.; de Miguel, Y. R. Org. Biomol. Chem. 2006, 4, 599–609.
(38) Zampronio, E.; Gotardo, M.; Assis, M. D.; Oliveira, H. P. Catal.
Lett. 2005, 104, 53–56.
(39) Gottfried, J. M.; Flechtner, K.; Kretschmann, A.; Lukasczyk, T.;
Steinruck, H.-P. J. Am. Chem. Soc. 2006, 128, 5644–5645.
(40) Barlow, D. E.; Scudiero, L.; Hipps, K. W. Langmuir 2004, 20, 4413–
4421.
(20) Bazylinski, D. A.; Hollocher, T. C. Inorg. Chem. 1985, 24, 4285–4288.
(21) Bazylinski, D. A.; Hollocher, T. C. J. Am. Chem. Soc. 1985, 107,
7982–7986.
(22) Kumar, M. R.; Pervitsky, D.; Chen, L.; Poulos, T.; Kundu, S.;
ꢀ
Hargrove, M. S.; Rivera, E. J.; Diaz, A.; Colon, J. L.; Farmer, P. J.
Biochemistry 2009, 48, 5018–5025.
(23) Sulc, F.; Immoos, C. E.; Pervitsky, D.; Farmer, P. J. J. Am. Chem.
Soc. 2004, 126, 1096–1101.
(24) Marti, M. A.; Bari, S. E.; Estrin, D. A.; Doctorovich, F. J. Am. Chem.
Soc. 2005, 127, 4680–4684.
(41) Williams, F.; Vauhgan, O.; Knox, K. J.; Bampos, N.; Lambert, R. M.
Chem. Commun. 2004, 1688.
(42) Kretschmann, A.; Walz, M.; Flechtner, K.; Steinruck, H. P.;
Gottfried, J. M. Chem. Commun. 2007, 568.
(25) Dobmeier, K.; Riccio, D.; Schoenfisch, M. Anal. Chem. 2008, 80,
1247–1254.
(43) Auwarter, W.; Weber-Bargioni, A.; Brink, S.; Riemann, A.; Schiffrin,
A.; Ruben, M.; Barth, J. V. ChemPhysChem 2007, 8, 250–254.

Article
Inorganic Chemistry, Vol. 49, No. 15, 2010 6957
reaction with NO donors (i.e., NO(g), N-nitrosomelatonin).60,61
Second, we covalently attached Co(P) to gold electrodes
and characterized its redox and structural properties by
electrochemical techniques as well as scanning tunneling
microscopy (STM), XPS, and solid-state density functional
theory (DFT) calculations. Finally, we studied electrochemi-
cally the NO and HNO donor reactions with the electrode-
bound Co(P).
Materials and Methods
Chemicals. Co(P) was purchased from Frontier Scientific and
used as received. To study the reaction with HNO, we used AS
and TSHA as HNO donors. AS and TSHA were synthesized
according to published literature procedures.62 NO gas was
generated by the dropwise addition of a solution containing
280 mg of NaNO2 in 8 mL of degassed water, to a solution of
Figure 1. Co(P) adsorbed on a surface.
metal center.44,45 The most successful methods for surface-
molecule linking are based on the establishment of Au-S
bonds, and a large number of porphyrin monomers bearing
thiols have been prepared and attached to gold electrodes by
this method.46-52 In order to avoid the problems associated
with disulfide formation in solution, the thiol is best handled
in a protected form (for example, with an S-acetyl protecting
group),53,54 that undergoes in situ cleavage on the surface.48
This technique is easier than evaporation deposition and does
not require the use of a quartz microbalance to obtain a single
layer.55
1.2 g of FeSO4 7H2O and 0.7 mL of H2SO4 in 15 mL of degassed
3
water (see the Supporting Information, reaction scheme SI1).
The produced NO was passed through a column of KOH pellets
to remove higher nitrogen oxides and through a water scrubbing
column prior to use. All these reactions were carried out in an
anaerobic atmosphere. N-Nitrosomelatonin was prepared as in
ref 61. NaNO2 was obtained from Anedra. Ferrocene, N(Bu)4-
PF6, and KNO3 were purchased from Sigma-Aldrich. The re-
maining chemicals were of analytical grade and were used without
further purification. Ethanol, dichloromethane (CH2Cl2), and
other organic solvents used were spectroscopic grade. The water
used was of Milli-Q grade, and nitrogen and argon of high purity
were used.
The high reactivity of cobalt(II) porphyrins toward NO,
plus the stability of cobalt(III) porphyrins and the easy and
efficient way of covalently joining porphyrins to electrodes,
prompted us to test the cobalt(II) 5,10,15,20-tetrakis[3-
( p-acetylthiopropoxy)phenyl]porphyrin [Co(P)], shown
in Figure 1, as a possible candidate for the electrochemical
discrimination of both species. For this purpose, first, we
studied the reaction between the commonly used HNO donors
sodium trioxodinitrate (Na2N2O3, Angeli’s salt, AS)56,57 or
toluenesulfohydroxamic acid (TSHA), an analogue of
Piloty’s acid,58,59 with CoII(P) and CoIII(P). We also studied the
Instruments. Cyclic voltammetry and differential pulse vol-
tammetry (DPV) were carried out with a TEQ 03 potentiostat.
For voltammetry in nonaqueous media (CH2Cl2), a three-
electrode system was used, consisting of two platinum electrodes
and a working glassy carbon electrode (GCE), and the support-
ing electrolyte (SE) was 0.1 M N(Bu)4PF6. The potential was
measured against a ferrocene pseudoreference and converted to
standard calomel electrode (SCE) potentials by using E(SCE) =
E(ferrocene) þ 0.120 V.
Cyclic voltammetries for the cobalt porphyrin modified
electrodes or for the bare gold electrodes used as controls were
performed by using the corresponding electrode as the working
electrode. In an aqueous solution, a KCl-saturated Ag0/AgCl
electrode was used as the reference electrode and a platinum wire
as the counter electrode (SEs were KNO3 and KClO4, 0.1 M, pH
6). For the corresponding measurements in organic solvent, two
platinum electrodes were used. CH2Cl2 was distilled from
appropriate drying agents (CaH2) under nitrogen just prior to
use. Deaeration of all solutions was accomplished by passing a
stream of high-purity argon through the solution for 10 min and
maintaining a blanket of inert gas over the solution while
making the measurements. The inert gas was saturated with
the appropriate solvent before entering the cell to minimize
evaporation of the solvent from the cell. Solutions in the cell
were changed periodically to avoid contamination from enter-
ing the cell via Ag/AgCl because Cl- anions may interfere.
The midpoint potential (E1/2) values were measured as that
potential lying midway between the oxidation and reduction
peaks for a given couple. Aqueous experiments were carried out
at room temperature (∼25 °C). CH2Cl2 experiments were
carried out at room temperature (cathodic waves) and -60 °C
(anodic waves). In the last case, an acetone-ethanol slush bath
cooled with liquid nitrogen was used to maintain the tempera-
ture. In a water solution, the potential was measured against a
(44) Lukasczyk, T.; Flechtner, K.; Merte, L. R.; Jux, N.; Maier, F.;
Gottfried, J. M.; Steinruck, H.-P. J. Phys. Chem. C 2007, 111, 3090–3098.
(45) Flechtner, K.; Kretschmann, A.; Steinruck, H.-P.; Gottfried, J. M.
J. Am. Chem. Soc. 2007, 129, 12110–12111.
(46) Akiyama, T.; Imahori, H.; Sakata, Y. Chem. Lett. 1994, 23, 1447–
1450.
(47) Shimazu, K.; Takechi, M.; Fujii, H.; Suzuki, M.; Saiki, H.;
Yoshimura, T.; Uosaki, K. Thin Solid Films 1996, 273, 250–253.
(48) Gryko, D. T.; Clausen, C.; Lindsey, S. L. J. Org. Chem. 1999, 64,
8635–8647.
(49) Jagessar, R.; Tour., J. M. Org. Lett. 2000, 2, 111–113.
(50) Kondo, T.; Ito, T.; Nomura, S.; Uosaki, K. Thin Solid Films 1996,
284-285, 652–656.
(51) Postlethwaite, T. A.; Hutchison, J. E.; Hathcock, K. W.; Murray,
R. W. Langmuir 1995, 11, 4109–4116.
(52) Simpson, T. R. E.; Revell, D. J.; Cook, M. J.; Russell, D. A.
Langmuir 1997, 13, 460–464.
(53) Lindsey, S. J.; Prathapan, S.; Johnson, E. T.; Wagner, W. R.
Tetrahedron 1994, 50, 8941–8968.
(54) Ravikanth, M.; Strachan, J. P.; Li, F.; Lindsey, J. Tetrahedron 1998,
54, 7721–7734.
(55) Scudiero, L.; Barlow, D. E.; Mazur, U.; Hipps, K. W. J. Am. Chem.
Soc. 2001, 123, 4073–4080.
(56) Bonner, F. T.; Degani, H.; Akhtar, M. J. J. Am. Chem. Soc. 1981,
103, 3739–3742.
(57) Gratzel, M. T. S.; Henglein, A. Ber. Bunsenges. Phys. Chem. 1970, 74,
1003.
(58) Bonner, F. T.; Ko, Y. Inorg. Chem. 1992, 31, 2514–2519.
(59) Seel, F. C. B. Z. Anorg. Allg. Chem. 1972, 394, 187–196.
(60) Turjanski, A. G.; Leonik, F.; Estrin, D. A.; Rosenstein, R. E.;
Doctorovich, F. J. Am. Chem. Soc. 2000, 122, 10468–10469.
(61) De Biase, P.; Turjanski, A. G.; Estrin, D. A.; Doctorovich, F. J. Org.
Chem. 2005, 70, 5790–8.
(62) Hughes, M. N.; Cammack, R. Nitric Oxide, Part C 1999, 301, 279–
287.

ꢀ
Suarez et al.
6958 Inorganic Chemistry, Vol. 49, No. 15, 2010
Ag/AgCl reference and converted to SCE potentials by using
E(SCE) = E(Ag/AgCl) þ 0.045 V.
of concentrated HCl. The gold surface was immersed in the
resulting solution overnight. Finally, the sample was rinsed with
CH2Cl2, followed by washing with Milli-Q water, and finally
dried with an argon stream.
UV-vis spectra were recorded with a HP 8453 spectrometer
in a septum-sealed 1-cm-path-length quartz cell.
Multilayers of the protected thiol-functionalized Co(P) on
Au(111) were prepared by placing a droplet of the 0.1 mM
solution of porphyrin in CH2Cl2 onto a freshly cleaned gold
surface. After all the solvent had evaporated, the sample was
measured by XPS without any rinsing procedure at all.
Cystamine Adsorption on a Co(P)-Modified Surface. After
porphyrin adsorption, some samples were modified with cysta-
mine. This was carried out by immersing the Co(P)-modified
electrodes in a 0.1 M cystamine aqueous solution for 24 h.
Finally, the electrodes were rinsed with water, followed by
washing with Milli-Q water, and finally dried with an argon
stream.
Electrode Preparation. Before each experiment, the surface of
the gold electrode was first polished with alumina powder
(diameter, 0.3 and 0.05 μm) and rinsed with pure water. Then
this electrode was cleaned in a Piranha solution [2:1 H2SO4/
H2O2 (30% aqueous)] for 5 min. After copious rinsing with
Milli-Q water, the gold electrode was electrochemically cleaned
by potential cycling in 0.5 M H2SO4 in the potential range of
0 and 1.5 V vs Ag0/AgCl until a typical cyclic voltammogram
of clean gold was obtained. After having been rinsed with
distilled water, ethanol and dried, the electrode was ready to
be modified.
Synthesis of CoIII(P)NO-. NO reacts directly with the four-
coordinate CoII(P) at room temperature to produce the corre-
Electrochemical Measurements. The voltammograms of
Co(P) were performed before and after the addition of HNO
(or NO) donors. The donor solutions were prepared as described
above (see section Reactions of CoIII(P) and CoII(P) with HNO
and NO Donors). When Co(P) is electrochemically driven to the
CoIII state, the CoIII/CoII couple is clearly observed before the
addition of an HNO donor. In order to detect the HNO reaction
with Co(P), the current was recorded as a function of time
(direct-current amperometry), at a fixed potential of 0.8 V,
which sets Co(P) in the CoIII state, as described in the results.
XPS. Co(P)- and “Co(P) þ cystamine”-modified Au(111)
samples were analyzed by XPS using a Mg KR radiation source
(XR50, Specs GmbH) and an hemispherical energy analyzer
(PHOIBOS 100, Specs GmbH). Previously, two-point calibra-
tion of the energy scale was performed using sputter-cleaned
gold [Au 4f7/2; binding energy (BE) = 84.00 eV] and clean
copper (Cu 2p3/2; BE = 933.67 eV) samples. The element ratios
were calculated by measurement of the areas corrected by its
relative sensibility factors. The reported spectral data were col-
lected with the best possible relation between resolution and
sample damage. For the S 2p region spectral deconvolution, a
Shirley-type background, and a combination of Lorentzian and
Gaussian functions were used. The full width at half-maximum
(fwhm) was fixed at 1.1 eV, and the spin-orbit doublet separa-
tion of the S 2p signal was set to 1.2 eV. The BEs and peak areas
were optimized to achieve the best adjustment.
sponding five-coordinate CoIII(P)NO- in 75% yield. CoIII
-
(P)NO- was prepared by bubbling NO to 5 mg of CoII(P)
dissolved in 3 mL of a degassed 1:1 CH2Cl2/MeOH solution.
During the course of the addition of NO gas (usually 10-
20 min), the color of the reaction mixture changes from purple
to red-purple. The reaction was followed by UV-vis spectros-
copy until quantitative conversion. The solvent was evaporated
and the solid dried in vacuo; the CoIII(P)NO- compound is
purple in the solid state. This compound has been characterized
by IR (Figure SI1 in the Supporting Information), UV-vis
spectroscopy, and microanalysis. Experimental (calcd): C,
62.3% (62.48%); H, 5.0% (4.92%); N, 5.7% (5.69%); S,
10.3% (10.42%). The IR spectrum of the neutral porphyrin
shows a strong band at the 1679 cm-1 (KBr) region which is
assignable to νNO (Figure SI1 in the Supporting Information).
This value is similar to that reported previously for (TPP)-
Co(NO) (1689 cm-1, KBr)27 and is close to the range of NO
frequencies reported for square-pyramidal N-substituted sali-
cylideaminate complexes of cobalt containing a bent NO ligand
(1645-1675 cm-1, Nujol mull).
Reactions of CoIII(P) and CoII(P) with HNO and NO Donors.
All experiments with CoIII(P) were done at 25 °C in a 0.1 M
phosphate buffer, containing 10-4 M ethylenediaminetetra-
acetic acid to avoid HNO catalytic decomposition by CuII and
other divalent ions. All experiments with CoII(P) were done at
25 °C in CH2Cl2. Porphyrin concentrations were 5 ꢀ 10-6 M,
unless otherwise noted. All manipulations were performed
under an argon atmosphere using septa; solutions were purged
for at least 20 min with argon prior to use. All reactions were
unaffected by irradiation of the sample with the light source of
the spectrometer. AS and TSHA solutions were freshly prepared
for each set of measurements and kept on ice, and the concen-
tration was checked before each measurement by UV-vis,
following absorbance.59,63 TSHA was dissolved in 96% ethanol
and added to the porphyrin solutions to a final ethanol/water
ratio smaller than 1%. NO was produced by using N-nitro-
somelatonin and a short light flash, as described in previous
works from our group.61 The quantity of NO produced per flash
was measured by UV-vis spectroscopy, following N-nitro-
somelatonin decomposition as previously described.60 The re-
action rates were measured by following the absorbance of the
porphyrins at the peak positions of the Soret and/or Q bands.
All reactions were pursued until no more spectroscopic changes
were observed, unless otherwise stated. The reaction times
ranged from 60 s to 2 h.
STM Characterization. For STM imaging, cobalt porphyrins
were adsorbed onto preferred oriented Au(111) substrates (from
Arrandee’s). After annealing for 10 min with a hydrogen flame,
these gold substrates exhibit atomically smooth (111) terraces
separated by steps, mono- and diatomic in height, as observed
by STM. STM imaging was made in air in the constant current
mode with a Nanoscope IIE microscope from Digital Instru-
ments (Santa Barbara, CA) and using commercial Pt-Ir tips.
Typical tunneling currents, bias voltages, and scan rates were 0.5
nA, 600 mV, and 10-15 Hz, respectively. Calibrations of the
piezoscanner for the x-y plane were carried out by imaging
highly oriented pyrolytic graphite.64 Steps present in the Au-
(111) surface of 0.24 nm height65,66 were used to calibrate the
piezotube in the z direction. Image analysis was carried out using
the microscope software, version 5.30r3.sr3 from Digital Instru-
ments. For size distribution analysis, the two-dimensional iso-
tropic power spectral density (PSD) after image high-pass
filtering was used.
DFT calculations in periodic boundary conditions were per-
formed using the Quantum Espresso code,67 which is based on
the pseudopotential approximation to represent the ion-electron
Immobilization of Co(P) and CoIII(P)NO- on a Gold Surface.
For porphyrin adsorption, a solution of 0.5 mg of Co(P) or
CoIII(P)NO- in 4 mL CH2Cl2 was prepared (0.1 mM). Re-
moval of the S-acetyl groups was performed by adding drops
(64) Pong, W.-T.; Bendall, J.; Durkan, C. Surf. Sci. 2007, 601, 498–509.
(65) Martin, H.; Vericat, C.; Andreasen, G.; Hernandez Creus, A.; Vela,
M. E.; Salvarezza, R. C. Langmuir 2001, 17, 2334–2339.
(66) Bunge, E.; Nichols, R. J.; Roelfs, B.; Meyer, H.; Baumgartel, H.
Langmuir 1996, 12, 3060–3066.
(63) Hughes, M. N.; Wimbledon, P. E. J. Chem. Soc., Dalton Trans. 1976,
703–707.
(67) http://www.quantum-espresso.org/.

Article
Inorganic Chemistry, Vol. 49, No. 15, 2010 6959
interactions and plane-wave basis sets to expand the Kohn-
Sham orbitals. Ultrasoft-type pseudopotentials68 were adopted,
together with the PBE formalism, to compute the exchange-
correlation term.69 An energy cutoff of 25 and 200 Ry was used
to expand the electronic wave functions and the charge density,
respectively. The Au(111) surface was modeled as an infinite
slab of Au atoms truncated at the (111) geometry, in supercells
of dimensions 15.04 ꢀ 11.57 ꢀ 15.04 A3 (Figure SI2 in the Sup-
porting Information). Given the large size of the supercell,
required to accommodate the porphyrin molecule, the repre-
sentation of the surface was limited to two layers of Au atoms,
and the sampling of the Brillouin zone was done with a 2 ꢀ 2 ꢀ 1
Monkhorst-Pack grid.70 All atomic coordinates were fully
relaxed, with the exception of those belonging to the second
gold layer, which was frozen on its bulk structure.
Results and Discussion
The results are organized as follows: First, the reactions of
Co(P) with HNO and NO are characterized in solution, and
electrochemical measurements of Co(P) and CoIII(P)NO-
are performed in organic media. Second, Co(P) and CoIII-
(P)NO- are covalently bound to gold electrodes, and elec-
trochemical measurements of the modified electrodes are
performed. Third, the modified Co(P)-bound electrodes [on
Au(111) substrates] are characterized by STM, XPS, and
DFT calculations. Finally, the Co(P)-modified gold electrode
reactions are studied with NO and HNO to analyze the
possibility for the detection and discrimination of both
species.
HNO and NO Reactions with Co(P). The initial char-
acterization of the NO and HNO reactions with Co(P)
was performed in a CH2Cl2 solution using UV-vis spec-
troscopy. The first reaction was performed by adding NO
gas to a CoII(P) solution. The spectral changes observed
in CH2Cl2 in the Soret and Q bands are shown in Figure
2A, which are in agreement with those reported for other
cobalt porphyrins.30 The conversion of CoII(P) to CoIII-
(P)NO- is evidenced after 5 min in the spectrum by a 20 nm
shift in the Soret band to higher wavelengths (Figure 2A)
and is in agreement with previous data for cobalt por-
phyrins.71 In a similar time scale, no spectral changes were
observed for CoII(P) in the presence of the HNO donor
TSHA.
Figure 2. (A) UV-vis spectra of the CoII(P) blue line (ε = 2.4 ꢀ 105
M-1 cm-1; λ max = 415 nm), the [CoIII(P)]þ red line (ε = 8.0 ꢀ 104 M-1
cm-1; λ max = 433 nm), and the CoIII(P)NO- green line (ε = 4.9 ꢀ 104
M-1 cm-1; λ max = 429 nm) in CH2Cl2. (B) Spectral changes observed for
the reaction of CoIII(P) with TSHA to yield CoIII(P)NO-. Inset: 433 nm
peak absorbance vs time plot for the reaction, measured at room
temperature in ethanol.
Further confirmation of the CoIII(P)NO- species ob-
tained by the reaction of CoIII(P) with TSHA was obtai-
ned by isolation and IR measurement of the reaction
product, which shows a new narrow band at 1679 cm-1
assigned to NO stretching (Figure SI1 in the Supporting
Information). Previous results reported by Roncaroli and
van Eldik,30 plus the observed NO stretching band located
in a region consistent with NO-, allow us to unequivocally
assign this compound as CoIII(P)NO- instead of CoII-
(P)NO•, which would be the expected assignment by
analogy to FeNO porphyrins. The results clearly show
that in solution CoII(P) reacts with NO, and not with
HNO, while CoIII(P) reacts with HNO and not with NO,
as expected.
Cyclic Voltammetry Studies in a CH2Cl2 Solution. Prior
to the electrochemical measurements of the covalent
bound Co(P) togoldelectrodes, wecharacterized the redox
properties of Co(P) and CoIII(P)NO- in solution and
compared the results with previous data reported for
other cobalt porphyrins. Figure 3 shows cyclic voltammo-
grams of both Co(P) and CoIII(P)NO- in CH2C12/0.1 M
To study the reaction with CoIII(P), we first oxidized
CoII(P) using the one-electron oxidant ferrocinium in
ethanol as the solvent [CoIII(P) is barely soluble in CH2Cl2].
After oxidation, excess TSHA was added to CoIII(P) in an
organic solution under an inert atmosphere, spectral
changes corresponding to the conversion of CoIII(P) to
CoIII(P)NO- were slowly observed (Figure 2B). The half
life (t1/2) of the reaction under our experimental conditions
is 12 min and can be accelerated to less than 1 min by the
addition of an organic base (DBU = 1,8-diazabicyclo-
[5.4.0]undec-7-ene, 5 μL), which accelerates TSHA decom-
position in an organic solvent by deprotonation, as shown
previously.24 In a similar time scale, no spectral changes
were observed for CoIII(P) in the presence of NO gas or NO
donor.
(68) Vanderbilt, D. Phys. Rev. B 1990, 41, 7892.
(69) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77,
3865–3868.
(70) Monkhorst, H.; Pack, J. Phys. Rev. B 1976, 13, 5188.
(71) Whitten, D. G.; Baker, E. W.; Corwin, A. H. J. Org. Chem. 2002, 28,
2363–2368.

ꢀ
Suarez et al.
6960 Inorganic Chemistry, Vol. 49, No. 15, 2010
Figure 4. Cyclic voltammograms of Co(P)-modified gold electrodes in a
0.1 M KNO3/water solution (black line) and Co(P) in a CH2C12/0.1 M
N(Bu)4PF6 solution (blue line). Potential values are referred to SCE (scan
rate = 0.1 V s-1).
second. The reduction waves, in solution, are further
apart by 340 mV
Figure 3. Cyclic voltammograms of Co(P) (left panels A1 and A2) and
CoIII(P)NO- (right panels B1 and B2) in a CH2C12/0.1 M N(Bu)4PF6
solution (scan rate = 0.1 V s-1). The potential (V) is reported against
SCE. In A1 and B1, the scan is performed toward positive potentials (mea-
sured at room temperature), while in A2 and B2, the scan is performed
toward negative potentials (measured at -60 °C).
Porphyrin Attachment and Cyclic Voltammogram Stud-
ies of Co(P) Bound to Gold Electrodes. Direct confirma-
tion of Co(P) adsorption was performed by electro-
chemical measurements of the “Co(P) electrodes” after
cleaning of the surfaces with abundant water. XPS data
presented in the next section give further insight regarding
the presence of Co(P) on gold substrates. Figure 4 shows
cyclic voltammograms of the Co(P)-modified gold elec-
trode in a 0.1 M KNO3 aqueous solution as compared to
Co(P) in a CH2Cl2 solution [Co(P) is not soluble in water].
Figure 4 shows that a reversible oxidation process is
observed within an E1/2 value of about 0.4 V. The reaction
corresponds to the CoIII/CoII couple (reaction 3), and
surprisingly the value is shifted ca. 400 mV to lower
potentials compared to values in solution. These results
show that Co(P) adsorption on the gold electrode facil-
itates CoII oxidation. The same results were obtained
for corresponding electrochemical measurements in a
CH2Cl2 solution. All resulting computed redox potentials
(in solution and adsorbed) are summarized in Table 1.
To confirm that the process corresponds to reduction
of the electrode-bound porphyrin, we measured the peak
intensities of the process as a function of the scan rate.
The resulting plot, shown in Figure SI3 in the Supporting
Information, clearly demonstrates that the process corre-
sponds to electrode-bound species.
N(Bu)4PF6. Measurements were performed at both room
(20 °C) and low temperature (-60 °C).72
As shown in the left panel of Figure 3, five redox
couples are found for Co(P), starting from the CoII(P)
redox state, corresponding to the following reactions:
The first three reactions (panel A1) correspond to one-
electron oxidations, while the waves shown in A2 corre-
spond to two one-electron reductions. For CoIII(P)NO-,
four waves are observed in the right panels of Figure 3 (B1
and B2), corresponding to the following reactions:
We then turned to measure HNO-bound Co(P), i.e.,
CoIII(P)NO-. Using the same technique as that for Co(P),
CoIII(P)NO- (prepared as described above using NO gas)
was covalently attached to gold electrodes and character-
ized electrochemically. Cyclic voltammograms of “CoIII
-
(P)NO- electrodes” in a 0.1 M KNO3 solution show a
clear peak corresponding to one-electron oxidation (the
inverse of reaction 7) at 0.83 V vs SCE (Figure SI4A in the
Supporting Information). In this case, the shift due to the
goldsurface effectismuchsmallerthanbefore, only60mV.
Again, as expected, the intensities show a linear depen-
dence on the scan rate, confirming that the process corre-
sponds to a signal arising from adsorbed species (Figure
SI4B in the Supporting Information). Similar results were
obtained for the measurements in organic media. All of the
resulting E1/2 values are summarized in Table 1.
The results are consistent with previous measurements
of other cobalt porphyrins.31 As described by Richter-
Addo et al.,72 the nitroxyl porphyrin undergoes two
oxidations and two reductions. The E1/2 separation for
the first oxidation processes for CoII(P) [“Co(P)”] and
CoIII(P)NO- [“Co(P)NO”, zero global charge] in solu-
tion are only 100 mV apart
and even less for the
(72) Richter-Addo, G. B.; Hodge, S. J.; Yi, G.-B.; Khan, M. A.; Ma, T.;
Van Caemelbecke, E.; Guo, N.; Kadish, K. M. Inorg. Chem. 1996, 35, 6530–
6538.

Article
Inorganic Chemistry, Vol. 49, No. 15, 2010 6961
Table 1. Measured Redox Potentials for Co(P) and CoIII(P)NO-
a Reaction number. b A three-electrode system was used, consisting of two platinum electrodes and a working GCE (scan rate = 0.1 V s-1). Cathodic waves were measured
at room temperature and anodic waves at -60 °C. c The cobalt porphyrin modified electrodes were used as working electrodes. In aqueous solution, a KCl-saturated Ag0/
AgCl electrode was used as the reference electrode and a platinum wire as the counter electrode. For the corresponding measurements in an organic solvent, two platinum
electrodes were used (scan rate = 0.1 V s-1, measured at room temperature).
Table 2. Elemental Ratios for the Co(P) and “Co(P) þ Cystamine” Layers on
during transfer to the UHV chamber. The N/Co ratio,
slightly higher than the stoichiometric value for Co(P),
Gold
indicates some demetalation of loaded porphyrin. For
both samples, the N/S ratio is closely above 1, in agree-
ment with the stoichiometric value. These results can be
explained assuming that most of the sulfur moieties have
been anchored to the gold surface. Therefore, N/S values
are greater than 1 because of the signal attenuation of
sulfur, which is directly bound to gold and deeper than
nitrogen.
Figure 5A shows the presence of N 1s and Co 2p regions
for Co(P) on Au(111). A Co 2p3/2 peak was found at 780.3
eV. The N 1s reveals a main peak (centered at 398.4 eV)
and a broad feature at high BE (401.25 eV). Small amounts
of metal-free porphyrins could also contribute to this
feature, in agreement with some demetalation of Co(P).
In addition, contaminants from handling the sample in
air might be present.
In the following, the absolute value of the BE of S 2p3/2
is employed to obtain information about the chemical
bonding of the sulfur head to the gold surface. It is worth
noting that the S 2p3/2 core level peak of alkanethiol self-
assembled monolayers on gold can be decomposed into
three different components: C1 at 161 eV, C2 at 162 eV,
and C3 at 163-164 eV.75 The C1 component is associated
with atomically adsorbed sulfur species.76 Component C2
is related to sulfur chemisorbed on the metal surface
through a thiolate bond, while component C3 has been
assigned to both unbound thiol and disulfide species.77-79
The spectral deconvolution of the S 2p region (Figure 5B,
N/S N/Au N/C N/O N/Co
Co(P)
Co(P) on Au(111)
1
1.2
0.06
1
4
5.1
7.7
0.05 0.04 0.14
0.10 0.05 0.15
“Co(P) þ cystamine” on Au(111) 1.2
The results from Table 1 show that, due to adsorption
of the cobalt porphyrin on the gold surface, a 400 mV
decrease in the CoIII/CoII couple is observed, and there-
fore a significant difference in the porphyrin redox po-
tential for the corresponding process is found between the
free and NO-bound porphyrins. The resulting first oxida-
tion process occurs at ≈0.4 V for CoIII/CoII and at 0.83 V
for CoIII(P)NO/CoIII(P)NO--modified electrodes in aqu-
eous media. The same results are observed in an organic
solvent, showing that this is not a solvent-related but a
surface-related effect. Interestingly, the other two redox
processes (2 and 4) are not affected by the surface. The
origin of the surface effect will be studied further using
DFT methods as shown later.
XPS of Co(P) and “Co(P) þ Cystamine” Surfaces. In
this section, the XPS analysis results of protected Co(P)
and “Co(P) þ cystamine” on preferred oriented Au(111)
substrates are described. The cystamine was used to avoid
the reaction as a possible source of a spurious signal in the
bare electrode (see below). Table 2 shows the elemental
ratios for Au 4f, Co 2p, C 1s, O 1s, and S 2p, in reference to
N 1s for the Co(P) (multilayer) and “Co(P) þ cystamine”
samples. The intensity N/C and N/O ratios exceed the
exact molecular stoichiometry. The source of extra oxy-
gen is probably molecular oxygen that can adsorb to the
Co center of the molecule.73,74 Carbon arises not only
from the organic molecules but also from contamination
(75) Vericat, C.; Vela, M. E.; Benitez, G. A.; Gago, J. A. M.; Torrelles, X.;
Salvarezza, R. C. J. Phys.: Condens. Matter 2006, 18, R867–R900.
(76) Yang, Y. W.; Fan, L. J. Langmuir 2002, 18, 1157–1164.
(77) Heister, K.; Zharnikov, M.; Grunze, M.; Johansson, L. S.; Ulman, A.
Langmuir 2000, 17, 8–11.
(78) Zhong, C.-J.; Brush, R. C.; Anderegg, J.; Porter, M. D. Langmuir
1998, 15, 518–525.
(73) Berner, S.; Lidbaum, H.; Ledung, G.; Ahlund, J.; Nilson, K.;
Schiessling, J.; Gelius, U.; Backvall, J.-E.; Puglia, C.; Oscarsson, S. Appl.
€
(79) Ishida, T.; Choi, N.; Mizutani, W.; Tokumoto, H.; Kojima, I.;
Azehara, H.; Hokari, H.; Akiba, U.; Fujihira, M. Langmuir 1999, 15,
6799–6806.
Surf. Sci. 2007, 253, 7540–7548.
(74) Yoshimoto, S.; Suto, K.; Tada, A.; Kobayashi, N.; Itaya, K. J. Am.
Chem. Soc. 2004, 126, 8020–8027.

ꢀ
Suarez et al.
6962 Inorganic Chemistry, Vol. 49, No. 15, 2010
Figure 5. XPS spectra of (A) the Co 2p and N 1s regions, (B) the S 2p
region for Co(P) on Au(111), and (C) S 2p regions for “Co(P) þ
cystamine” on Au(111). The S 2p regions are shown as two peak decon-
volutions for both samples.
C) shows two components of BEs at 161.9-162.1 eV (C2)
and 163.3-163.4 eV (C3). The C2 component corresponds
to a p doublet of thiolates bound to gold, while the C3
contribution is related to free thiol, thioacetyl functions,
and disufides. Similar results were obtained for acetyl-
protected porphyrin adsorption from solution73,80 and
under ultrahigh-vacuum conditions.81 It is worth noting
that it is not clear which fraction of the C3 component is
related to any of the species that contribute to this signal.
Whatever was the origin of this component, it is clear that
it seems to be about half of the total sulfur amount. Then,
it might be claimed that Co(P) could be adsorbed with its
porphyrin ring both parallel to the surface or near to its
surface normal. In this sense, Berner et al. carried out XPS
measurements at different takeoff angles, establishing that
the thioacetyl or free thiols are located in the outermost parts
of the surface.73,80 If this is true for the current system, which
makes sense, the bound sulfur amount should be estimated
considering some attenuation. Then, it is reasonable to
suppose that the number of sulfur species attached to the
substrate is greater than that corresponding to C3 related
species. Under this circumstance, the lying-down configuration
Figure 6. (A) 100 nm ꢀ 100 nm STM image of a Co(P)-modified
Au(111) substrate. (B) Cross-sectional profile corresponding to the black
trace on Figure 6A. (C) Cross-sectional profile corresponding to the green
trace on Figure 6A.
would be, on average, favored. We found further support in
favor of molecules having their porphyrin ring parallel to
the surface by carrying out STM imaging of Co(P)-
modified surfaces (see below).
XPS analysis indicates an increase in the S/Au and N/Au
intensity ratios after cystamine adsorption. Considering
that during cystamine adsorption each molecule gives rise
to two thiols after breakage of the S-S82 bond, this result
shows that free gold sites are probably blocked. No oxi-
dized sulfur species (S 2p BE > 166 eV) such as sulfonate
were detected by XPS on any of the analyzed samples.
Characterization of a Co(P)-Bound Gold Surface by
STM. We also characterized the covalently bound Co(P)
at the molecular level by STM imaging. This is a key point
regarding the chemical activity of immobilized molecules
because it strongly depends on the porphyrin overlayer
structure and on the adsorption geometry of individual
(80) Berner, S.; Biela, S.; Ledung, G.; Gogoll, A.; Backvall, J.; Puglia, C.;
Oscarsson, S. J. Catal. 2006, 244, 86–91.
(81) Turner, M.; Vaughan, O. P. H.; Kyriakou, G.; Watson, D. J.;
Scherer, L. J.; Davidson, G. J. E.; Sanders, J. K. M.; Lambert, R. M.
J. Am. Chem. Soc. 2009, 131, 1910–1914.
(82) Wirde, M.; Gelius, U.; Nyholm, L. Langmuir 1999, 15, 6370–6378.

Article
Inorganic Chemistry, Vol. 49, No. 15, 2010 6963
Figure 8. Structures of Co(P) and CoIII(P)NO- on the Au(111) surface
optimized by DFT calculations.
These 2.9-nm-sized spots are consistent with similar
cobalt porphyrins chemisorbed on gold surfaces.85 This
also suggests that the molecules are in the lying-down
configuration.
Figure 7. (A) PSD plot obtained by analysis of the image on panel “B”.
Notice that both the power and wavelength are indicated on logarithmic
scales. (B) 150 nm ꢀ 150 nm images of Co(P) species adsorbed on a
Au(111) terrace, after high-pass filtering.
From XPS and STM analysis of the modified surfaces,
it can be inferred that most of the Co(P) molecules are
adsorbed in the lying-down configuration by multiple-
linker binding to the gold surface, with some molecules in
a second layer. The strong adsorbate-surface interaction
explains the random distribution of porphyrins through
the Au(111) terraces. Consequently, considering that a
disordered Co(P) adlayer had been grown, vacant gold
sites can be occupied by adsorbates having a smal-
ler lateral size. These findings would also be expected to
be valid in the case of porphyrin immobilized on the
polycrystalline electrodes used in the electrochemical
experiments.
To compute the surface coverage, reductive desorp-
tion86,87 of the thiolated porphyrins was performed in
basic media (0.1 M KOH). From the cathodic peak in
the voltammogram (Figure SI7 in the Supporting Infor-
mation), the charge density involved in thiol reduction
was measured. The charge density value allows estima-
tion of the number of porphyrin molecules present on the
gold surface, assuming one electron per thiolate bond.
The resulting calculation yields a surface coverage of 25%
(see the Supporting Information for more details on the
calculation). This value is similar, within experimental
error, to the one obtained using STM. We have also
verified from voltammetric runs that there was no sig-
nificant adlayer degradation after potential excursion to
1.0 V in a 0.1 M KNO3 aqueous solution.
molecules. Typical STM images of the Co(P)-covered
Au(111) surfaces are shown in Figure 6A. A step edge
between two Au(111) terraces across the image is visible
near the left-down corner (parallel to the black line
indicated as “B”). Also, vacancy gold islands, which
are frequently observed on thiol-modified surfaces, in the
atomically smooth terraces of the Au(111) substrate can
be clearly distinguished. The substrate is decorated by a
large number of randomly distributed bright spots, which
are related to porphyrin adsorption. Figure 6 shows that
there are two types of elements, the most frequently
observed are 0.42 ( 0.11 nm in height (cross section,
Figure 6B), while the larger and brighter ones exhibit
0.87 ( 0.08 nm height (cross section, Figure 6C; see also
the height histogram included in Figure SI5 in the Sup-
porting Information). The 0.4 nm heights are in good
agreement with previously reported data about molecular
contrast in STM images of metal-modified porphy-
rins74,83 and phthalocyanines.84 The brighter spots with
heights of about 0.9 nm can be assigned to agglomerates
in a second layer. In fact, it has been shown that porphy-
rins are able to form bilayers.74,83 If the molecules were ori-
ented with their porphyrin plane near the surface normal,
they should show an apparent height of about 3 nm.
Furthermore, this supposition is still valid for the brightest
spots (associated porphyrin). The possibility of having
porphyrin aggregates, in which not all of the mercapto
functions are covalently bound to the substrate, is consis-
tent with the XPS results, which show contributions of
acetyl-protected thiols, unbound thiols, or disulfides to the
sulfur signal. Repetitive scanning with the STM tip did not
remove the molecules, indicating that they are strongly
bound to the gold surface and also that the second layer is
firmly attached to the first one.
Electronic Structure Calculations of Co(P) and CoIII-
(P)NO-. Model structures of the Co(P) and CoIII(P)NO-
complexes resting on the Au(111) surface were optimized
by using DFT calculations. Geometry relaxations were
started from equivalent configurations for both systems,
with the porphyrin macrocycles lying about 3.7 A above
the surface. Interestingly, the optimizations led to the
final structures depicted in Figure 8: the average lengths
from the first gold layer to the plane of the molecule
turned out to be 3.79 and 4.43 A for the free and
nitrosylated porphyrins, respectively. The separation ob-
tained for the Co(P) model is in reasonable agreement
with the value observed from the STM adsorbates, of
0.42 ( 0.11 nm. The Co atoms sits between two Au atoms
on a bridge site.
Regarding the porphyrin orientation with respect to the
surface, the apparent porphyrin size in the x-y plane was
analyzed. Two-dimensional PSD analysis of a 150 nm ꢀ
150 nm image, after high-pass filtering, shows a broad
distribution of spots with average size 2.9 nm (Figure 7A).
(83) Yoshimoto, S.; Suto, K.; Honda, Y.; Itaya, K. Chem. Phys. 2005,
The atomic charges resulting from a Lowdin popula-
tion analysis are shown in Table 3. In Co(P), the charge on
319, 147–158.
(84) Hipps, K. W.; Lu, X.; Wang, X. D.; Mazur, U. J. Phys. Chem. 1996,
100, 11207.
(85) Berner, S.; Ledung, G.; Ahlund, J.; Nilson, K.; Schiessling, J.; Gelius,
U.; Backvall, J.-E.; Puglia, C.; Oscarsson, S. Appl. Surf. Sci. 2007, 253, 7540–
7548.
(86) Walczak, M. M.; Popenoe, D. D.; Deinhammer, R. S.; Lamp, B. D.;
Chung, C.; Porter, M. D. Langmuir 1991, 7, 2687–2693.
(87) Yang, F.; Wilde, C. P.; Morin, M. Langmuir 1996, 12, 6570–6577.

ꢀ
Suarez et al.
6964 Inorganic Chemistry, Vol. 49, No. 15, 2010
the metal becomes less positive upon binding. In contrast,
the cobalt charge increases slightly in CoIII(P)NO- when
the molecule approaches the gold surface. This behavior
is indicative of the different nature of the Co-Au inter-
action in the presence and in the absence of NO. At the
same time, some electron density seems to flow away from
the porphyrin N atoms in Co(P), while the atomic popu-
lations of the equatorial N atoms are hardly affected
when NO is bound. We can conclude that NO coordina-
tion separates the porphyrin from the gold surface and
diminishes the charge transfer from the surface to the Co
center, consistent with what was previously reported by
Flechtner et al. for a gold surface.45
shown in Figure 9A. When Co(P) is electrochemically
driven to the CoIII state, the CoIII/CoII couple is observed
before the addition of AS (blue line, 0.4 V). Under the
same conditions [oxidized Co(P)], adding AS results in a
shift of the voltammogram corresponding to CoIII
-
(P)NO- (green line, 0.83 V). This clearly shows the in
situ reaction of CoIII(P) with HNO. To verify that the
shift is due to a surface effect and that it is not affected by
the SE, the same experience was performed but with 0.1 M
KClO4 (pH 6). In turn, because the redox event observed
was pseudoreversible (green line in Figure 9A), the appa-
rent potentials were measured more carefully by DPV
techniques (Figure 9B).
HNO and NO Detection and Discrimination Using Co-
(P)-Modified Gold Electrodes. As mentioned in the intro-
duction, the main aim of this study is to develop a device
that is able to detect and discriminate HNO from NO,
using Co(P). To achieve this goal, the Co(P)-modified
electrode must be able to react with HNO but not with
NO (or any other similar species) and yield an electro-
chemical response in the presence of HNO. Alternatively,
an NO-selective detector must react with NO but not with
HNO. Given the known reactivity in solution, which
shows that CoII(P) reacts fast with NO but not with
HNO while the opposite is observed for CoIII(P), we
tested both reactions for the Co(P)-modified gold elec-
trodes. The reactions were monitored electrochemically
using the same protocol as that mentioned above. For
example, to analyze the CoIII(P) reaction with HNO or
NO, a voltammogram of Co(P) was performed before
and after the addition of an HNO (or NO) donor, as
Using similar procedures, the rest of the reactions
(listed below) were studied. As expected, the results show
that CoIII(P) reacts efficiently with HNO, while it does
not with NO donors or NO gas. This is evidenced by the
lack of electrochemical signals (current intensity) when
the porphyrin is maintained in the CoIII state, at a pot-
ential compatible with the CoIII(P)NO/CoIII(P)NO- cou-
ple (about 0.8 V), and the current intensity is monitored
after the addition of NO donors. On the other hand, NO
reacts rapidly with CoII(P)-modified electrodes while
HNO does not; i.e., the CoIII(P)NO/CoIII(P)NO- couple
is observed after the addition of an NO donor to a Co(P)-
bound electrode maintained in the CoII(P) state, while no
change is observed when HNO donors are used. The
resulting behavior of the modified electrodes is summar-
ized in Scheme 1.
A very important issue for selective HNO detection in
the presence of NO is the fact that NO is able to react with
bare gold electrodes at 0.8 V. Therefore, it is necessary to
avoid this reaction as a possible source of spurious
signals. To study the NO response on the Co(P)-modified
electrodes, we measured the current intensities produced
by a 3 μM NO solution on (a) bare gold electrodes (green
curve on Figure 10), (b) Co(P)-modified electrodes (red
Table 3. Atomic Charges According to a Lowdin Population Analysis, for the
Relaxed Structures of Co(P) and CoIII(P)NO- in a Vacuum (isolated) and on the
Au(111) Surface (Adsorbed)a
Scheme 1
P
P
a
Np and NO stand for the charges summed on the porphyrin N atoms and
on the HNO atoms, respectively.
Figure 9. InsituHNO reactionwith adsorbed CoIII(P). Potential values are referredtoSCE (scan rate = 0.1 Vs-1). (A) Cyclicvoltammograms ofCoIII(P)
before(blueline) andafter(greenline) the addition of0.7mMASina 0.1 M KNO3 aqueous solution. (B) DPVofCoIII(P) afterthe additionof0.7 mMASin
a 0.1 M KClO4 aqueous solution. Pulse amplitude: 30 mV. Pulse width: 25 ms. Pulse period: 40 ms.

Article
Inorganic Chemistry, Vol. 49, No. 15, 2010 6965
Figure 11. Intensity vs time plot for Co(P)-modified electrodes. The
arrow indicates the addition of AS.
previous observations, CoIII(P)NO-, which under the
described conditions is oxidized to CoIII(P)NO. The
CoIII(P)NO complex releases the NO ligand to yield
CoIII(P), which allows the catalytic cycle to start again.
The measured current corresponds to oxidation of
CoIII(P)NO- at 0.8 V, as is nicely shown in Figure 11,
which is detected only a few seconds after AS addition. It
should be noted that a fast equilibrium between the
continuous HNO production by AS decomposition and
HNO dimerization occurs, yielding a steady-state HNO
concentration that reacts with CoIII(P), as is further
explained below.
Figure 10. Current intensity for 3 μM NO on bare gold electrodes
(green), CoIII(P)-modified electrodes (red), and “CoIII(P) þ cystamine”-
modified electrodes (black). The current was detected by fixing the
electrode potential at 0.8 V vs SCE.
Scheme 2
As can be observed from Figure 11, the obtained
current intensity increases after the addition of the
HNO donor, in about 2-3 min. After this initial signal
increase, the current starts to decay slowly because of the
expected AS decomposition,20 with a half-life in the range
of AS to HNO decomposition under these conditions
(t1/2 = 15 min, at pH 7 and room temperature).20 The
coincidence of the current signal decay times with AS
spontaneous decomposition decay to yield HNO strongly
suggests that Co(P) is reacting with the released HNO as
expected. To analyze the quantitative response of the
electrode upon AS addition, we measured the current
intensity obtained by the addition of different initial AS
concentrations at 0.8 V. The minimum HNO detection
limit of the electrode under our experimental conditions
is ca. 1 nM, corresponding to a current of 3.65 nA. The
concentration of HNO was estimated as described in the
Supporting Information.
We also performed the same set of experiments without
stirring of the solution, although in this case the HNO
reaction with the electrode is limited by diffusion and the
intense signal obtained after AS addition falls off rapidly
after an initial burst of reaction, possibly because of limi-
ted HNO diffusion to the surface. These results are shown
in Figure SI9 in the Supporting Information. Finally, to
corroborate the selectivity of the method, the current of
the same Co(P) electrode (in the same conditions) was
measured after the addition of an NO donor. As expected,
no increase in the current intensity was observed because
of the lack of reaction of CoIII(P) with NO, showing the
HNO selectivity of the presented method.
curve on Figure 10), and (c) “Co(P) þ cystamine”-mod-
ified gold electrodes (black curve on Figure 10). NO was
produced by using N-nitrosomelatonin and a short light
flash.55 The quantity of NO produced per flash was
measured by UV-vis spectroscopy. The signal was re-
corded by fixing the potential at 0.8 V and measuring the
current versus time after NO addition.
As shown in Figure 10, a considerable current intensity
is observed because of the presence of 3 μM NO in solu-
tion when bare gold electrodes are used as working elec-
trodes. The Co(P) porphyrin partially blocks the signal
(the remaining signal is ca. 50% of that for the bare
electrode), although because of the 25% coverage of the
gold surface by Co(P), the signal is still significant. To
completely block the signal coming from the direct inter-
action of NO with gold sites, the Co(P)-modified electro-
des were subsequently modified with cystamine. As ex-
pected, for these electrodes, the current intensity dropped
to less than 5% of the original one, proving that the
reaction of NO with the electrode is almost completely
avoided. Finally, we verified that cystamine does not
interfere with the CoIII(P)NO- electrochemical signal,
as shown in Figure SI8 in the Supporting Information.
HNO Detection. In order to detect HNO, the following
reaction scheme was designed (Scheme 2), on the basis
of the previous electrochemical data. According to
Scheme 2, in order to selectively detect HNO, the rest-
ing-state electrode potential is set to 0.8 V. At this value,
Co(P) is stable in the CoIII(P) state and no current flow is
observed. The reaction with HNO yields, according to
Conclusions
Redox Potential Shift for the CoIII/CoII Gold-Adsorbed
Porphyrin. One of the most interesting points observed in

ꢀ
Suarez et al.
6966 Inorganic Chemistry, Vol. 49, No. 15, 2010
this work is the 400 mV shift of the redox potential for the
CoIII/CoII redox couple when comparing the cobalt por-
phyrin in solution with the adsorbed species, while only a
behavior of Co(P) is maintained when the porphyrin is
covalently attached to gold electrodes, allowing sensitive
electrochemical measurements to be performed. These mea-
surements are aided by a ca. -0.4 V shift of the CoIII/CoII
redox couple due to interaction with the surface, which
allows us to work at a potential in which reaction 7 takes
place but reaction 3 does not.
very small shift is observed for the nitrosylated CoIII
-
(P)NO/CoIII(P)NO- couple. In an interesting work by
Lukasczyk et al., by using XPS, evidence of electron
transfer from a silver surface to the Co metal center of
an electrostatically adsorbed porphyrin was obtained.44
They also showed that the above-mentioned Co-Ag
interaction is suppressed by axial coordination of NO to
the cobalt porphyrin, and the authors suggested that this
effect results from competition between the two axial ligands
(NO and the silver surface, considered as a “ligand”).45
Electronic coupling between the Co center of a porphyrin
and a metal surface has also been observed by using
STM.40,51,52,88 Our electrochemical measurements are in
agreement with a surface-to-porphyrin charge donation
and the suppression of this effect by NO coordination. In
the noncoordinated porphyrin, charge donation from the
surface significantly stabilizes the CoIII state, lowering the
CoIII/CoII redox potential compared to the value in
solution. The difference of about 0.6 A in the surface
separation obtained by DFT calculations for the free and
nitrosylated porphyrins is a direct evidence of the role of
NO in the weakening of the Co-Au interaction. In the
particular case of CoIII(P)NO-, where the ligand pulls the
metal out of the plane, the distance between the slab and
the Co atom is 4.64 A, too large to expect any significant
coupling between the cobalt and the surface. NO coordi-
nation suppresses most of the charge-transfer effect, and
therefore no significant change in the redox potential is
observed between CoIII(P)NO- in solution and the ad-
sorbed species. The decrease of 0.05 e- in the calculated
atomic charge of the Co metal center upon adsorption of
CoII(P) supports the notion of electron transfer from the
surface to the metal center. This is accompanied by an
increase of the charge on the porphyrin N atoms, which
could be explained as a response to the increase in the
electronic population on the cobalt, eventually depolar-
izing the Co-Np bond. This effect is not observed in the
presence of NO, in which case the change in the cobalt
atomic charge is not negative but positive, suggesting that
charge transfer from the surface is virtually suppressed.
HNO and NO Discrimination with Co(P)-Modified Gold
Electrodes. As mentioned in the Introduction, discriminat-
ing NO and HNO is a difficult task and it has not been
achieved so far in biological samples. Metalloporphyrins
are good candidates for the synthesis of gas sensors, and
iron porphyrins are known to react with NO. The main
difficulty for using iron porphyrins in HNO/NO discrimi-
nation is the fact that both the ferric and ferrous states react
with both species. Co(P), instead, is selective for either NO
or HNO depending on their redox state. We have shown
that Co(P)-modified gold electrodes are capable of reacting
efficiently with NO only when Co(P) is in the CoII state,
while in the CoIII state, only reaction with HNO is possible.
This behavior possibly relies on the fact that only the
CoIII(P)NO- nitroxyl species seem to be stable. In contrast,
FeII(P)NO, FeIII(P)NO, and also FeII(P)HNO species have
been shown to be stable.23 Moreover, the discriminative
Our results show that Co(P)-modified electrodes can
react with NO/HNO to yield CoIII(P)NO--modified elec-
trodes and that both CoII(P) and CoIII(P)NO- have at
least one electron reduction couple and one electron oxi-
dation couple accessible in aqueous media. Being able to
use the modified electrodes in aqueous media is a key issue
when developing a selective HNO sensor to be used in
biological experiments.
Selective HNO Detection. On the basis of the reactive
and electrochemical properties of the Co(P)-modified gold
electrodes, we designed a reaction scheme (Scheme 2) for its
use in HNO detection. Our results show that the “Co(P) þ
cystamine”-modified electrode responds to HNO, showing
a current intensity after HNO donor addition because of
oxidation of CoIII(P)NO-. These results make a clear point
for the future development of an HNO sensor by using the
present strategy for its use in biological media. Once devel-
oped, our system could be used directly on the sample in situ
without the need for subsequent UV-vis characterization.
Adlayer’s robustness of the present device has been
verified by exposing the modified surfaces to mild condi-
tions, such as UHV environments, air exposition, and a
wide electrochemical potential range. STM surface ana-
lysis shows that Co(P) ligands are randomly linked to the
Au(111) surface in a lying-down configuration as one- or
two-molecule aggregates.
Future work will allow the study of the long-term
stability of the electrodes, their quantitative response to
HNO, and also their response to other sources of spurious
signals such as nitrite and other nitrogen reactive species.
Also important is to test the electrode response in aerobic
media because, although working under a nitrogen atmo-
sphere in biological in vitro studies is possible, the role of
oxygen in HNO chemistry must also be taken into account.
Acknowledgment. This work was financially supported
by the University of Buenos Aires, Agencia Nacional
ꢀ
ꢀ
de Promocion Cientıfica y Tecnologica (PICT 05-32980
´
and 06-2396), and CONICET. We also thank Professor
F. Battaglini for assistance in the electrochemical
measurements.
Supporting Information Available: Synthesis of NO gas (NO)
and CoIII(P)NO-, supercell of the DFT calculation, measure-
ment of the electrode effective surface area, measurement of the
Co(P) gold electrode surface coverage, cyclic voltammograms of
CoIII(P)NO- in organic media, height distribution analysis
from STM data, effect of cystamine coverage of the Co(P)
electrode on the CoIII(P)NO- oxidation signal, estimation of
the HNO concentration, selective detection without stirring,
and nine supplementary figures. This material is available free of
charge via the Internet at http://pubs.acs.org.
(88) Zak, J.; Yuan, H.; Ho, M.; Woo, K. L.; Porter, M. D. Langmuir 1993,
9, 2772–2774.