Sensors and Actuators B 107 (2005) 14–23

Sensitive optical NO sensor based on

bis [(2,9-dimethyl-1,10-phenanthroline)] copper(II) complex
H. Dacres, R. Narayanaswamy∗
Department of Instrumentation and Analytical Science, Institute of Science and Technology (UMIST), University of Manchester,
P.O. Box 88, Manchester M60 1QD, UK

Received 7 April 2004; received in revised form 13 August 2004; accepted 16 September 2004
Available online 21 December 2004

Abstract

The reaction of Cu(dmp)2 2+ with NO has been investigated for potential development into an optical fibre chemical sensor. In aqueous
solutions, a linear calibration (r = 0.9950) was obtained between 0 and 2.5 ␮M NO with a detection limit of 0.32 ␮M (∼0.0096 ppm) (s/n = 3).
The complex electrostatically immobilised in nafion produced a highly linear calibration (r = 0.999) between 0 and 3 ppm in NO gas phase with
a detection limit of 0.072 ppm (∼2.38 ␮M in solution). The response was shown to be fully reversible by immersing the films in hydroxylamine
hydrochloride solutions and a steady state response was achieved within 100 s exposure time. The nafion films were also shown to be more
selective to NO over O2 (183:1), NO2 (16:1) and CO (795:1). In solution nitrite, H2 O2 and peroxynitrite anion did not interfere.
© 2004 Elsevier B.V. All rights reserved.

Keywords: Nitric oxide detection; Absorbance; Nafion; Nitric oxide sensor

1. Introduction regulation of blood vessels, communication in the brain and

Since the identification of nitric oxide as a biological sig-
nalling molecule, the realisation that NO is involved in the
Abbreviations: ABTS, 2,2 -azinobis (3-ethylbenzthiazoline-6-sulfonic
acid); Au, absorbance units; CCD, charge coupled detector; cNOS, con-
stitutive nitric oxide synthase; DAF, diaminofluorescein; DAN, diaminon-
aphthalene; DAR, diaminorhodamine; DEANO, diethylamine NONOate;
Dmp, 2,9-dimethyl-1,10-phenanthroline; DMSO, dimethyl sulfoxide; ECR,
eriochrome cyanine R; EDTA, ethylenediaminetetraacetic acid; eNOS, en-
dothelial nitric oxide synthases; EPR, electron paramagnetic resonance spec-
troscopy; H2 O2 , hydrogen peroxide; HHCL, hydroxylamine hydrochloride;
iNOS, inducible nitric oxide synthase; KO2 , potassium superoxide; OH• , hy-
droxyl radical; MLCT, metal to ligand charge transfer; MnO2 , manganese
(IV) oxide; NEDD, N-(1-naphthyl) ethylenediamine hydrochloride; nNOS,
neuronal nitric oxide synthase; NO, nitric oxide; NO− , nitroxyl anion; NOS,
nitric oxide synthase; O2 − , superoxide anion; ONOO− , peroxynitrite anion;
R.S.D., relative standard deviation; S.D., standard deviation; sGC, soluble
guanylate cyclase; SOD, superoxide dismutase; SULF, sulfanilamide; TMD-
ABODIPY, 1,3,5,7-tetramethyl-8-(3 ,4 -diaminophenyl)-difluoroboradiaza-
s-indacene
∗ Corresponding author. Tel.: +44 161 200 4911; fax: +44 161 200 4891.
E-mail address: ramaier.narayanaswamy@manchester.ac.uk
(R. Narayanaswamy).
in immunological defence against invading organisms has
emerged [1].
NO is synthesised from l-arginine and this reaction is
catalysed by nitric oxide synthase (NOS) [2,3]. Three distinct
isoforms of NO exist, two of which are classed as constitutive
NOS (cNOS), including endothelial NOS (eNOS or NOS-III)
and neuronal NOS (nNOS or NOS-I). A third NOS species
is called inducible NOS (iNOS or NOS-II) [4–6]. The most
important distinction is that cNOS produces small amounts
of NO for short periods of time and iNOS catalyses higher
concentrations of NO for longer.
The realisation of the importance of the diverse roles en-
dogenous NO can play led to the development of a number
of techniques to monitor its concentration. This paper is con-
cerned with the development of an optical chemical sensor for
monitoring NO concentration that can be applied to biolog-
ical monitoring. Optical fibre chemical sensors have several
advantages over other conventional analytical sensors such
as electrodes [7,8]. They are not subject to electrical inter-
ference and are safer than electrochemical sensors, no refer-
ence electrode is required and they can be miniaturised for

0925-4005/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.snb.2004.09.043
 H. Dacres, R. Narayanaswamy / Sensors and Actuators B 107 (2005) 14–23
in vivo use. Other analytical techniques used for detection of
NO include electrochemistry [7,8], chromatography [9] and
electron paramagnetic resonance spectroscopy (EPR) [10].
Electrochemistry is prone to interference from other biolog-
ical components. The major drawback with chromatography
is the expense and size associated with chromatographic in-
strumentation. EPR is a sensitive and specific technique for
NO detection from a purely analytical point of view but many
of the required reagents are expensive or not commercially
available, and bulky apparatus is required.
In the development of an optical chemical sensor, a reagent
phase is primarily identified then applied to nitric oxide de-
tection in solution and this is followed by immobilising the
reagent in an appropriate polymeric support. A number of
reagents are available for the detection of NO in biologi-
cal solutions. Three of the most common colorimetric meth-
ods include the Griess assay [11,12] involving the nitrosa-
tion of the griess reagent [sulfanilamide (SULF) and N-(1-
naphthyl) ethylenediamine hydrochloride (NEDD)], the con-
version of ferrocyanide to ferricyanide [11] and the ABTS
(2,2 -azinobis (3-ethylbenzthiazoline-6-sulfonic acid)) assay
[11,13]. All three methods can simultaneously measure NO
and nitrite, which can lead to over-estimation of NO con-
centration. The oxyhemoglobin is another well-established
method for NO detection [14,15]. The major drawback of
the oxyhemoglobin assay is that oxyhemoglobin is not com-
mercially available and the preparation of oxyhemoglobin
involves oxygenation of the dithionite reduced heme protein
followed by de-salting with a separation column. Chemi-
luminescence has shown exquisite sensitivity and selectiv-
ity for NO detection [16,17]. The most widely used fluo-
rescence probes for NO detection include diaminonaphtha-
lene (DAN), diaminofluoresceins (DAFs) and diaminorho-
damines (DARs) [18]. All three fluorescent techniques are
more sensitive than the colorimetric assays available but
also have their disadvantages. DAN is poorly soluble in
solution and employs UV excitation (375 nm), which can
cause autofluorescence and serious damage to living cells
[19]. In the search for fluorescence probes, which employ
longer excitation wavelengths, DAFs and DARs were syn-
thesised. DAFs are prone to photobleaching and instability
around neutral pH [20] and although DARs are virtually un-
affected by photobleaching, they still exhibit instability at
neutral pH [21]. However, DAFs and DARs both exhibited
low detection limits for NO of 3 and 7 nM, respectively.
More recently, 1,3,5,7-tetramethyl-8-(3 ,4 -diaminophenyl)-
difluoroboradiaza-s-indacene (TMDABODIPY) was synthe-
sised to overcome these problems and was shown to be photo-
stable and pH independent over a wide range [22]. However,
on approaching physiological temperatures, TMDABODIPY
was quickly protonated, which interfered with its response to
NO.
The sol–gel encapsulation of cytochrome c in the develop-
ment of an optical fibre sensor for NO gas resulted in calibra-
tion curves with a linear range of 1–25 ppm with a detection
limit of 1 ppm [23]. Due to interference caused by NO2 , cy-
15
tochrome c , extracted and purified from the bacterium para-
coccus denitrificans, was also encapsulated in sol–gel ma-
trices and was shown to be more selective to NO over NO2
due to steric hindrance, which excluded NO2 from the heme
binding site [24]. A working range of 10–400 ␮mol l−1 NO
was achieved. An NO selective optical ratiometric sensor was
constructed using a fluorescein dye derivative (4-carboxy-
2 ,7 -difluorofluorescein-succinimidyl ester) attached to col-
loidal gold [25]. The sensor had a detection limit of 20 ␮M
NO. Fluorescein dye-labelled cytochrome c attached to col-
loidal gold [26] resulted in an increase in sensitivity to NO
and an improved detection limit of 8 ␮M. Further investi-
gations using the previous sensor design utilised the fluo-
rescently labelled heme domain of soluble guanylate cyclase
(sGC) [27]. The sensor response was measured between 0 and
0.7 mM NO concentrations with a detection limit of 1 ␮M.
One drawback of this method is the heme domain of sGC is
fragile and is only useful at room temperature for a period of
several hours.
In the identification of a suitable reagent scheme for NO
quantification, the biological redox Cu(I/II) couple is of in-
terest [28]. The Cu(II)–eriochrome cyanine R (Cu(II)–ECR)
complex has been shown to react quantitatively with NO in
aqueous solutions and with NO gas when immobilised in a
silicone rubber matrices [29]. The Cu(II) complex of 2,9-
dimethyl-1,10-phenanthroline (dmp) (Cu(dmp)2 2+ ) (Fig. 1)
can be reduced by nitric oxide in solution [30,31] and the
formation of the cuprous complex was observed at 454 nm
in the optical absorption spectrum by its characteristic metal
to ligand charge transfer (MLCT) band. The Cu(II) complex
is also pH independent between pH 5 and 9 [32] making it
effective for use at physiological pH (pH 7.4).
This paper presents the use of the Cu(II)–dmp complex
in aqueous solutions under aerobic conditions for NO de-
tection. This reagent was then used in the development of a
sensitive optical chemical sensor for NO by the electrostatic
immobilisation of the complex in nafion followed by investi-
gation of its reaction with nitric oxide gas. Nafion is a perflu-
orosulfonate resin in which hydrophilic perfluorinated ether
side chains terminate with sulfate groups, which are period-
ically attached to hydrophobic perfluoroethylene backbone
molecules. The fluorocarbon backbone provides exceptional
chemical and thermal stability while the sulfate groups are
responsible for ion exchange [33]. The sulfonate groups are
assumed to be completely dissociated thus creating a strongly
acidic medium [34].
Fig. 1. Structure of the cupric phenanthroline complex.
16 H. Dacres, R. Narayanaswamy / Sensors and Actuators B 107 (2005) 14–23
2. Experimental
2.1. Apparatus
Investigations were carried out using a commercially
available miniature fibre-optic based spectrometer (Ocean
Optics PC 1000), which utilises a small tungsten halogen
lamp as the light source and a CCD based detector for
absorbance measurements. Quartz cuvettes were used for
absorbance measurements of solutions. For immobilisation
studies, a purpose built gas blender was used to mix appropri-
ate quantities of the gases to produce known concentrations
at a controlled flow rate. All experiments were carried out
using a total flow rate of 600 ml/min unless otherwise stated
in the text. The spectrometer was used in conjunction with a
purpose built flow cell [29].
2.2. Solution studies
2.2.1. Reagents
All reagents were of analytical grade and used as pur-
chased without further purification. All solutions were pre-
pared with distilled deionised water unless otherwise stated.
2.2.2. Buffer preparation
A 0.01 M (pH 7.4) phosphate buffer solution was prepared
by mixing 0.0016 mol of potassium dihydrogen orthophos-
phate (KH2 PO4 ) (BDH) with 0.0034 mol of sodium phos-
phate dibasic (Na2 HPO4 ) (Aldrich) in water. To make buffer
solutions of varying pHs (pH 3–10), the contents of hydrion
buffer chemenvelopes (Aldrich) were added to 500 ml of dis-
tilled water.
2.2.3. Complex preparation
A 1 × 10−2 M stock solution of Cu(II) was prepared by
dissolving copper(II) sulphate pentahydrate (Aldrich) in wa-
ter. A 1 × 10−2 M dmp (Aldrich) stock solution was also
prepared in water. The two components were diluted with
phosphate buffer (0.01 M, pH 7.4) to produce the desired fi-
nal concentrations. When referring to the concentration of
the complex, this always refers to the Cu(II) concentration.
In this paper, 2,9-dimethyl-1,10-phenanthroline will also be
referred to as dmp.
2.2.4. DEANO
A 2.5 × 10−4 M diethylamine NONOate (DEANO) so-
lution was prepared in phosphate buffer and used immedi-
ately after preparation. DEANO is a solid which when dis-
solved in phosphate buffer (pH 7.4, 22–25 ◦ C) dissociates
to produce stoichiometric amounts of nitric oxide [35]. The
oxyhemoglobin assay was used to quantify NO release from
DEANO and it was demonstrated that 1 mol of NO was re-
leased after a 20 min period (22 ◦ C, pH 7.4).
2.2.5. Nitrite
A 0.01 M stock solution of sodium nitrite (Fisons) was
prepared in water by dissolving 0.069 g in 100 ml of deionised
distilled water. The appropriate amount of nitrite was added to
the reagent solutions to give a final concentration of 100 ␮M.
2.2.6. Hydrogen peroxide (H2 O2 )
A 30% (w/w) hydrogen peroxide solution was purchased
from Sigma and used as received without further purification.
From this solution, a 0.1 M stock solution was prepared in
water. This solution was prepared immediately prior to use
and added to the reagent solutions to give the desired final
concentration (100 ␮M).
2.2.7. Peroxynitrite anion (ONOO− )
The autoxidation of hydroxylamine in alkaline solutions
leads to the formation of peroxynitrite anion as its major
product [36,37]. This method was chosen for peroxynitrite
anion synthesis as it is simple, inexpensive and does not re-
quire any specialist equipment. The method is based upon
the following reaction:
NH2 O− + O2 → H2 O2 + NO− (1)
NO− + O2 → ONOO− (2)
The reaction proceeds by the attack of oxygen on a de-
protonated species yielding nitroxyl ion, which is further ox-
idised to peroxynitrite. In the presence of metal sequester-
ing agent ethylenediaminetetraacetic acid (EDTA), the per-
oxynitrite ion is stabilised. Experiments were carried out
with solutions containing 0.01 M of hydroxylamine, 0.5 M
of sodium hydroxide (NaOH) and 0.001 M of EDTA. The
solutions were bubbled with oxygen and stirred vigorously
for about 4–5 h. Following this, MnO2 (manganese(IV) ox-
ide) was added to the solution and then filtered and follow-
ing this the solution was stored in a freezer. Peroxynitrite
ion concentration was determined by UV spectrometry at
302 nm (ε = 1670 M−1 cm−1 ) [37]. The appropriate dilution
of the peroxynitrite solution was made to give a solution of
100 ␮M concentration.
2.2.8. Superoxide anion (O2 − )
Potassium superoxide (KO2 ) solutions can be used to pre-
pare solutions containing about the same concentration of
superoxide anion [38,39]. KO2 is sparingly soluble in dry
dimethyl sulfoxide (DMSO) and the use of 18-crown-6 ether
increases the solubility of KO2 allowing pale yellow solu-
tions to be prepared as long as the crown ether is in at least a
2:1 excess of KO2 . A 0.05 M solution of KO2 was prepared in
0.15 M crown ether: DMSO solution. The solution was son-
icated until both KO2 and crown ether had fully dissolved.
Superoxide anion concentration was determined by UV spec-
trometry at 250 nm (ε = 2686 ± 29 M−1 cm−1 ) [40–42]. The
appropriate dilution of the superoxide solution was made to
give a solution of 100 ␮M concentration.
Superoxide dismutase (SOD) (Sigma) from bovine blood
contained 2800 units/mg. SOD was dissolved in 10 ml of
phosphate buffer (1 mg/10 ml) and diluted to give a fi-
nal concentration of 2.5 ␮g/ml. This concentration had
 H. Dacres, R. Narayanaswamy / Sensors and Actuators B 107 (2005) 14–23 17

previously been shown to inhibit the effect of superox-

ide produced from a 50 ␮g/ml KO2 solution by 85–90%
[39].

2.2.9. Procedure
Varying amounts of DEANO were added to a 20 ␮M
Cu(dmp)2 2+ solution to give the desired final concentrations.
The stoichiometry of Cu(II): dmp was determined to be 1:2
using the molar ratio method [43]. The temperature was main-
tained at 22 ◦ C and the absorbance was read after 20 min re-
action time.
A final concentration of 100 ␮M of the appropriate inter-
ferent was added to solutions containing 20 ␮M Cu(dmp)2 2+
or 20 ␮M Cu(dmp)2 2+ with 50 ␮M NO. The procedure
was repeated three times for each interferent in both
solutions.
2.3. Immobilisation
1 × 10−2 M solutions of dmp were prepared in 5 ml of
the nafion solution (5%, w/v) and 50 ␮l of this solution was
pipetted onto a clean glass cover slip (22 mm2 , BDH) and left
to dry and spread under its own weight on a flat surface. The
dry dmp films were then placed in 10 ml of 6 mM copper(II)
sulphate solutions also containing 6 mM hydroxylamine hy-
drochloride (Aldrich) and left to react overnight. Once the
copper complex was formed in nafion the films were placed
in phosphate buffer overnight before reaction with NO gas.
Bovine plasma was purchased from Sigma and reconsti-
tuted in phosphate buffer (0.01 M, pH 7.4) and nafion films
were placed in the plasma or phosphate buffer to assess the
effect of plasma on the film response.
2.3.1. Procedure
The nafion films were exposed to a chosen concentration of
analyte gas at a flow rate of 600 ml/min and the response was
monitored as a function of time. The optical signal of the films
after gas exposure is plotted versus the concentration of the
gas to obtain calibration graphs. Results are presented as nor-
malised absorbance units (absorbance at time = t/absorbance
prior to gas exposure).
Fig. 2. Absorbance spectra of 20 ␮M Cu(dmp)2 2+ before (1) and after in-
cremental additions of NO (2) 2.5 ␮M, (3) 5 ␮M, (4) 7.5 ␮M, (5) 10 ␮M
and (6) 12.5 ␮M in phosphate buffer (pH 7.4, 0.01 M, 22 ◦ C) recorded after
20 min reaction time.
Cu(dmp)2 2+ + ROH + NO
→ Cu(dmp)2 + + RONO + H+ (3)
The mechanism involves three key steps: (1) the forma-
tion of a reversible inner-sphere adduct between Cu(II) and
NO; (2) the reaction of this adduct with ROH to give a ni-
trite derivative; (3) dissociation to the products Cu(dmp)2 + ,
RONO and H+ (Scheme 1).
When the reaction was carried out in an unbuffered aque-
ous solution, the solution pH was observed to decrease
from pH 6 to 3 in a manner consistent with acid forma-
tion [30]. No reaction was observed with pure methanol or
dichloromethane, indicating the importance of the hydrox-
ylic solvent in providing a reagent that can react with the co-
ordinated NO [30,31]. The sequence presented in Scheme 1
was proposed as the likely mechanism as it was predicted that
an outer-sphere redox pathway was unlikely due to the calcu-
lated potential for a simple electron transfer reaction between
Cu(dmp)2 2+ and NO being unfavourable (E = −0.63 V in
aqueous solutions).
3.1.2. Calibration of NO in aqueous solutions
Fig. 3 shows the calibration plot for NO in solution.
The calibration plot is linear (r = 0.9950) up to 2.5 ␮M NO

3. Results and discussion

3.1. Solution studies

3.1.1. Absorbance spectrum

Upon reaction of NO with Cu(dmp)2 2+ , the colour of
the solution changed from a pale green to a brighter yel-
low/orange. This colour change was accompanied by an in-
crease in absorbance at 454 nm (Fig. 2). This change in
absorbance is consistent with the reduction of Cu(dmp)2 2+
to Cu(dmp)+ [30,31]. The overall reaction of NO with Scheme 1. Proposed mechanism for the reduction of Cu(dmp)2 2+ by NO in
Cu(dmp)2 2+ is shown in the Eq. (3): hydroxylic solvents (S) and R is H (H2 O) or CH3 (methanol) [30].
18 H. Dacres, R. Narayanaswamy / Sensors and Actuators B 107 (2005) 14–23
Fig. 3. Absorbance (454 nm) of 20 ␮M Cu(dmp)2 2+ recorded after incuba-
tion with NO (0–2.5 ␮M), at 22 ◦ C, in 0.01 M phosphate buffer (pH 7.4) for
20 min, mean ± S.D. (n = 9).
with a detection limit of 0.32 ␮M (s/n = 3). The sensitivity
was 0.0086 absorbance units/␮M (Au/␮M). This is about
14 times more sensitive to NO than was observed with the
Cu(II)–ECR complex [29]. The detection limit of 0.32 ␮M
compares well with other colorimetric techniques, includ-
ing the Griess assay, ABTS assay and ferrocyanide assay
[11]. The oxyhemoglobin assay [14] has a lower detec-
tion limit (80 nM) than the technique presented here but
Cu(dmp)2 2+ is substantially easier to prepare. A detection
limit of 0.32 ␮M suggests that this technique will be able to
monitor NO synthesised by iNOS. The low concentrations
of NO used for calibration will also limit interference by
oxygen.
Fig. 4. (a) The effect of the various interferents on the absorbance (454 nm) of 20 ␮M Cu(dmp)2 2+ and (b) the effect of these potential interferents in solutions
containing 50 ␮M NO. The procedure was repeated three times (mean ± S.D., n = 3) and (*) represents a significant difference (P = 0.05) in the absorbance
signal compared to 20 ␮M Cu(dmp)2 2+ alone (a) or in presence of NO for (b).
3.1.3. Interferents in solution
A number of potential interferents in solution were stud-
ied. The effects of nitrite, H2 O2 , peroxynitrite anion and
superoxide anion on the absorbance of 20 ␮M Cu(dmp)2 2+
(Fig. 4(a)) and their effects on the reaction of NO with the
complex were investigated (Fig. 4(b)). These interferents
were investigated due to the fact that they could be present in
biological medium due to the reaction of NO with either O2
or superoxide [4], which under certain circumstances is also
generated by iNOS [44]. Superoxide reacts with Cu(dmp)2 2+
producing a significantly different (P = 0.05) optical signal to
Cu(dmp)2 2+ alone. This response was partially reversible us-
ing SOD, which did not show any significant effect on the
absorbance of Cu(dmp)2 2+ itself (P = 0.05). When 50 ␮M
NO was also present in the solution, both H2 O2 and su-
peroxide interfered with the reduction of Cu(dmp)2 2+ and
the superoxide interference was reversible by the addition
of SOD. It was thought that both H2 O2 and superoxide in-
terfered with the reaction by reacting with NO to produce
either hydroxyl radical (OH• ) [45] or singlet oxygen [46]
in the case of H2 O2 and peroxynitrite anion in the case of
superoxide [4]. In the latter case, it appears that peroxyni-
trite itself does not react with Cu(dmp)2 2+ (Fig. 4(a)), so
superoxide simply stops NO being available for the reaction.
In solutions, nitrite, H2 O2 and peroxynitrite do not interfere
significantly with the absorbance of Cu(dmp)2 2+ . Nitrite and
peroxynitrite do not interfere with the reaction of NO with
Cu(dmp)2 2+ .
 H. Dacres, R. Narayanaswamy / Sensors and Actuators B 107 (2005) 14–23
Fig. 5. Absorbance spectrum of dmp nafion films after incubation in
0.6 mM copper(I) sulphate (1), followed by immersion in phosphate buffer
(0.01 M, pH 7.4) overnight (2) and following NO exposure (3) (94.3 ppm at
100 ml/min).
3.2. Immobilisation studies
3.2.1. Absorbance spectrum
Initial studies attempting to immobilise the Cu(dmp)2 2+
as a complex in nafion were unsuccessful. Preparation of
dmp films in nafion followed by immersion in copper(II) sul-
phate solutions resulted in pale green films, which did not
respond to NO gas. Immersion of the dmp/nafion films in
copper(II) sulphate solutions containing equimolar amounts
of hydroxylamine hydrochloride resulted in intense yellow
films with an absorbance maximum at 450 nm (Fig. 5). The
pH of the Cu(II) solutions, containing hydroxylamine hy-
drochloride, were varied from pH 3 to 10 with maximum
intensities observed between pH 3 and 4 (Fig. 6). The pH
of the Cu(II) solutions were maintained at pH 3.62 ± 0.035
pH units (mean ± S.D., n = 3). These results are consistent
with the reduction of the Cu(II) prior to reaction with dmp
to produce the Cu(dmp)2 + complex. Prior to exposing the
films to NO gas the nafion films were immersed in phosphate
buffer (pH 7.4, 0.01 M) overnight. This resulted in an in-
crease in the absorbance intensity (Fig. 5) accompanied with
a colour change from yellow to orange. Phosphate buffer
ions probably reversibly bind to the reduced complex with-
out exerting any redox changes as inferred by the similarity
of the absorbance spectra at higher wavelengths (>600 nm)
Fig. 6. Effect of changing pH on the absorbance of dmp nafion films
immersed in copper(II) sulphate solutions containing hydroxylamine hy-
drochloride.
19
Scheme 2. Proposed mechanism for the oxidation of Cu(dmp)2 + by NO in
nafion.
[31]. However, upon exposure to NO gas, the colour of the
films changes from orange to a distinct green colour with an
accompanying decrease in absorbance at 450 nm (Fig. 5).
It appears that NO partially oxidises the nafion films,
which is confirmed by the small increase in absorbance ob-
served between 600 and 700 nm. Cu(I) and Cu(II) compounds
in acids react with NO resulting in the formation of Cu(II)
nitroxyl cation [Cu(NO− )]2+ , during which the oxidation of
Cu(I) to Cu(II) results in the formation of N2 O without any
significant formation of N2 [47]. The reduction of NO to
NO− is thermodynamically favourable with a one-electron
reduction potential of NO/NO− couple being +0.39 V [48].
Further evidence to support this mechanism comes from the
fact that SOD–[Cu(I)] reduces NO to NO− [49]. It was sug-
gested that a proton donation from SOD itself may facilitate
reduction of NO to HNO. The pKa of HNO is predicted to
be 7.2 ± 1.0, so in the acidic environment offered by nafion
it is likely that the nitroxyl cation exists as HNO [50]. The
following scheme is proposed as the most likely mechanism
for the reaction of NO with Cu(dmp)2 + in nafion (Scheme 2).
3.2.2. Kinetics
Fig. 7(a) shows that the initial rates of reaction, as noted
by the absorbance change at 450 nm, is dependent on NO
concentration and the reaction between NO and Cu(dmp)2 +
is almost complete after about 100 s. Fig. 7(b) shows the cal-
culated rate constants for each NO concentration represented
as the gradients. Fig. 8 shows a plot of these rate constants
versus NO concentration. The first-order rate constants (kobs ,
Eq. (4)) are shown to be a linear function of nitric oxide
concentration, indicating a simple first-order dependence on
[NO] (Eqs. (4) and (5)):
d[Cu(dmp)2 + ] d[Cu(dmp)2 2+ ]
− = = kobs [Cu(dmp)2 + ]
dt dt
(4)
kobs = kNO [NO] (5)
From the slope of the graph (Fig. 8), the initial reaction
rate (kNO, Eq. (5)) can be calculated as 0.0018 ppm−1 s−1
(∼0.06 M−1 s−1 ). This is approximately twelve times faster
than that observed using Cu(ECR)2 in silicone rubber [29].
3.2.3. Calibration of NO gas
A calibration graph (Fig. 9) was plotted from the data
presented in Fig. 7(a), the absorbance was measured after
100 s as the reaction appears to be completed after this expo-
sure time. A linear calibration graph was obtained (r = 0.999,
20 H. Dacres, R. Narayanaswamy / Sensors and Actuators B 107 (2005) 14–23

Fig. 7. (a) Kinetics of absorbance change (450 nm) and (b) rate of change in absorbance (450 nm) of nafion incorporating Cu(I)–dmp complex after incubation
in phosphate buffer (pH 7.4, 0.01 M) followed by exposure to NO (0–3 ppm, 600 ml/min).

0–3 ppm NO) and the detection limit was calculated to be
0.072 ppm (s/n = 3) (∼2.38 ␮M in solution). The sensitiv-
ity of the films to NO gas (0.0689 Au/ppm ∼ 0.0021 Au/␮M
in solution) was shown to be lower than that observed in
solution (0.0086 Au/␮M). The sensitivity of these nafion
films is twice that of the silicone rubber films incorporat-
ing Cu(II)–(ECR) complex and the detection limit is nearly
a sixth of that obtained with these silicone rubber films. The
detection limit compares well with the detection limits ob-
tained with other optical chemical sensors. Sol–gel encap-
sulation of cytochrome c [23] and cytochrome c extracted
from Paracoccus denitrificans [24] resulted in detection lim-
its of 1 ppm and 10 ␮mol l−1 (∼0.3 ppm) of NO, respectively.
An optical fibre NO sensor using aquocyanocobinamide re-
sulted in a detection limit of 5 ␮M (∼0.15 ppm) [51]. The
use of fluorescein derivative labelled heme proteins resulted
in detection limits ranging from 1 to 20 ␮M (0.03–0.6 ppm)
[26,27], although the response times were much faster than
those presented here ranging from 0.9 to 20 s.
3.2.4. Film regeneration
The reversibility of the film response to NO was investi-
gated using hydroxylamine hydrochloride reagent. The spec-

Fig. 9. Absorbance (450 nm) of Cu–dmp complex in nafion after immersion

in phosphate buffer followed by exposure to NO gas (0–5 ppm) in nitro-
Fig. 8. Initial rate of absorbance change (450 nm) of Cu(I)–dmp in nafion gen (total flow rate 600 ml/min) recorded after 100 s exposure, mean ± S.D.
with respect to NO concentration (0–3 ppm). (n = 3).
 H. Dacres, R. Narayanaswamy / Sensors and Actuators B 107 (2005) 14–23
Fig. 10. (a) Absorbance spectrum of Cu–dmp complex in nafion upon re-
immersion in hydroxylamine hydrochloride (HHCL) and (b) absorbance
change (450 nm) upon repeated exposure to 2 ppm NO followed by immer-
sion in hydroxylamine hydrochloride for 30 min periods.
tra presented in Fig. 10(a) indicate that the absorbance in-
tensity is almost fully restored upon immersion in hydrox-
ylamine solution for 30 min. The film was then exposed
to NO gas repeatedly, following 30 min immersion in hy-
droxylamine solutions (Fig. 10(b)). The arrows labelled NO
gas in Fig. 10(b) indicate a 5 min exposure to 2 ppm NO
gas and HHCL arrows at 300, 600, 900 and 1200 s indi-
cate a 30 min immersion time in hydroxylamine hydrochlo-
ride solution. The response of the nafion films to NO gas
shows full reversibility in hydroxylamine solution. The av-
erage absorbance of the five exposures was 0.6221 ± 0.0202
(mean ± S.D., n = 5). This gives a smaller relative standard
deviation (R.S.D.) of 3.26% compared to that obtained with
the calculated error in the absorbance intensity of five indi-
vidual films (9.57%). This suggests that more reproducible
results would be achieved using the same film than us-
ing different films for separate measurements. It was also
shown that this response was reversible even if the films
were immersed in hydroxylamine solutions for up to a
month.
3.2.5. Interferents in the immobilised phase
The next stage of the investigation into the suitability of
this reagent for NO sensing involves studying the effects of
other gases that may interfere by either reacting with the
films or by preventing reaction of NO. The gases that were
investigated included O2 , NO2 and CO. To assess how selec-
21
Fig. 11. (a) Absorbance (450 nm) of Cu–dmp complex in nafion after immer-
sion in phosphate buffer followed by exposure to CO or NO2 gas (0–8 ppm)
and (b) oxygen (0–60%) in nitrogen (total flow rate 600 ml/min) recorded
after 100 s, mean ± S.D. (n = 3).
tive the nafion films were to NO with respect to these other
gases, calibration graphs for the three gases were produced
with an exposure time of 100 s. These calibration graphs are
presented in Fig. 11(a) and (b).
CO and NO2 had little effect on the absorbance of the
nafion films (Fig. 11(a)). In fact, for both gases there was
no significant difference between the data point presented in
Fig. 11(a) as shown by an ANOVA test (P = 0.05). Comparing
the gradients of the two calibrations for CO and NO2 to that of
the NO calibration (Fig. 9), the nafion films are about 795 and
16 times more selective to NO, respectively. Fig. 11(b) is the
calibration plot for O2 and the detection limit was calculated
to be 15.20% (∼203 ppm). The nafion films were 183 times
more selective to NO gas over O2 . From these results, it can
be concluded that these nafion films respond more selectively
to NO over CO, NO2 and O2 .
The responses of the films to NO were investigated in the
presence of these interferent gases, which are represented
in Fig. 12. O2 , NO2 and CO did not interfere significantly
with the reaction of 2 ppm NO gas with the nafion films.
Films which had been prepared three months prior to this
study also did not respond to NO gas significantly differ-
ently to freshly prepared films (P = 0.05). Plasma was the
only medium, which interfered with the response of the films
to NO gas. This is probably due to plasma proteins binding
to the surface of the film and hindering the passage of NO
gas into the membrane.
22 H. Dacres, R. Narayanaswamy / Sensors and Actuators B 107 (2005) 14–23
Fig. 12. The effects of 50% O2 , 6 ppm CO, 6 ppm NO2 , ageing the film (three
months) and bovine plasma on the response of Cu–dmp complex in nafion
to 2 ppm NO gas (600 ml/min in N2 ), mean ± S.D. (n = 3), (*) indicates a
significant difference to the response of the films to NO gas (P = 0.05).
4. Conclusions
Investigations in aqueous solutions have shown that NO
reduces Cu(dmp)2 2+ to Cu(dmp)2 + . The detection limit of
NO determination using the complex was 0.32 ␮M with a
linear response (r = 0.9950) in the range 0–2.5 ␮M. Nitrite,
H2 O2 and peroxynitrite anion did not react with the complex
but superoxide does. Both, H2 O2 and superoxide, interfere
with the reaction of the complex with NO by scavenging
analyte species.
In the investigations involving the immobilisation of the
complex in nafion, the reduced complex was shown to be
oxidised by NO. A linear (r = 0.9999) calibration plot was
obtained between 0 and 3 ppm NO with a detection limit of
0.072 ppm (∼2.38 ␮M in solution) NO and response time
of 100 s. The films could be regenerated in hydroxylamine
hydrochloride and the regenerated films responded to NO
in a reproducible manner (0.6221 ± 0.0202 absorbance units,
3.26%, R.S.D., n = 5) for up to a period of one month. Aged
nafion films (three months) were shown to respond to NO in
the same manner as freshly prepared films. The nafion films
were shown to be highly selective to NO compared to O2 ,
NO2 and CO and these gases did not significantly affect the
films’ response to NO. Films immersed in bovine plasma did
not respond as well to NO gas as films immersed in phosphate
buffer solutions.
These studies have identified that the copper complex of
dmp incorporated into nafion films offers the potential of
sensitive sensing NO gas at low ppm and ppb range under
potential biological conditions. This, in turn, offers potential
for detecting NO synthesised by iNOS.
Acknowledgement
The authors acknowledge the Engineering and Physi-
cal Research Council for the Ph.D. studentship support to
H.D.
References
[1] E. Culotta, D.E. Koshland Jr., Science 258 (1992) 1862–1865.
[2] R.M.J. Palmer, D.D. Rees, D.S. Ashton, S. Moncado, Biochem. Bio-
phys. Res. Commun. 153 (3) (1988) 1251–1256.
[3] A.M. Leone, R.M.J. Palmer, R.G. Knowles, P.L. Francis, D.S. Ash-
ton, S. Moncado, J. Biol. Chem. 266 (35) (1991) 23790–23795.
[4] A. Sigel, H. Sigel, Metal Ions in Biological Systems, vol. 36, Marcel
Dekker Inc., 1999 (Chapter 17).
[5] C. Nathan, FASEB J. 6 (1992) 3051–3064.
[6] H.M. Abu-Soud, K. Ichimori, A. Presta, D.J. Stuehr, J. Biol. Chem.
275 (23) (2000) 17349–17357.
[7] K. Shibuki, Neurosci. Res. 9 (1) (1990) 69–76.
[8] T. Malinski, Z. Taha, Nature 358 (1992) 676–678.
[9] K. Funazo, M. Tanaka, T. Shono, Anal. Chim. Acta 119 (1980)
291–297.
[10] J.J.B. Kalyanaraman, J.S. Hyde, Biochem. Biophys. Res. Commun.
192 (2) (1993) 926–934.
[11] R.W. Nims, J.C. Cook, M.C. Krishna, D. Christodoulou, C.M.B.
Poore, A.M. Miles, M.B. Grisham, D.A. Wink, Methods Enzymol.
268 (1996) 93–103.
[12] L.A. Ridnour, J.E. Sim, M.A. Hayward, D.A. Wink, S.M. Martin,
G.R. Buettner, D.R. Spitz, Anal. Biochem. 281 (2000) 223–229.
[13] R.W. Nims, J.F. Darbyshire, J.E. Saavedra, D. Christodoulou, I. Han-
bauer, G.W. Cox, M.B. Grisham, F. Laval, J.A. Cook, M.C. Kr-
ishna, D.A. Wink, METHODS: A Companion to Methods Enzymol.
7 (1995) 48–54.
[14] C. Privat, F. Lantoine, F. Bedioui, E. Millanvoye van Brussell, J.
Devynck, M.-A. Devynck, Life Sci. 61 (12) (1997) 1193–1202.
[15] M.E. Murphy, E. Noack, Methods Enzymol. 233 (1994) 240–250.
[16] K. Kikuchi, T. Nagano, H. Hayakawa, Y. Hirata, M. Hirobe, Anal.
Chem. 65 (1993) 1794–1799.
[17] K. Kikuchi, T. Nagano, H. Hayakawa, Y. Hirata, M. Hirobe, J. Biol.
Chem. 265 (31) (1993) 23106–23110.
[18] T. Nagano, Luminescence 14 (1999) 283–290.
[19] A.M. Miles, D.A. Wink, J.C. Cook, M.B. Grisham, Methods Enzy-
mol. 268 (1996) 105–120.
[20] H. Kojima, N. Nakatsubo, K. Kikuchi, S. Kawahara, Y. Kirino, H.
Nagoshi, Y. Hirata, T. Nagano, Anal. Chem. 70 (1998) 2446–2453.
[21] H. Kojima, M. Hirotani, Y. Urano, K. Kikuchi, T. Higuchi, T.
Nagano, Tetrahedron Lett. 41 (2000) 69–72.
[22] X. Zhang, H. Wang, J.-S. Li, H.-S. Zang, Anal. Chim. Acta 481
(2003) 101–108.
[23] J.W. Aylott, D.J. Richardson, D.A. Russell, Chem. Mater. 9 (1997)
2261–2263.
[24] D.J. Blyth, J.W. Aylott, J.W.B. Moir, D.J. Richardson, D.A. Russell,
Analyst 124 (1999) 129–133.
[25] S.L.R. Barker, R. Kopelman, Anal. Chem. 70 (1998) 4902–4906.
[26] S.L.R. Barker, H.A. Clark, S.F. Swallen, R. Kopelman, Anal. Chem.
71 (1999) 1767–1772.
[27] S.L.R. Barker, Y. Zhao, M.A. Marletta, R. Kopelman, Anal. Chem.
71 (1999) 2071–2075.
[28] A.C.F. Gorren, E. DeBoer, R. Wever, Biochim. Biophys. Acta 916
(1987) 38–47.
[29] H. Dacres, R. Narayanaswamy, Sens. Actuators B 90 (2003)
222–229.
[30] D. Tran, P.C. Ford, Inorg. Chem. 35 (1996) 2411–2412.
[31] D. Tran, B.W. Skelton, A.H. White, L.E. Laverman, P.C. Ford, Inorg.
Chem. 37 (1998) 2505–2511.
[32] A.N. Baga, G.R.A. Johnson, N.B. Nazhat, R.A. Nazhat, Anal. Chim.
Acta 204 (1988) 349–353.
[33] S.A. John, R. Ramaraj, J. Appl. Polym. Sci. 65 (4) (1997) 777–787.
[34] D.T. Newcombe, T.J. Cardwell, R.W. Cattrall, S.D. Kolev, J. Membr.
Sci. 141 (1998) 155–164.
[35] L.K. Keefer, R.W. Nims, K.M. Davies, D.A. Wink, Methods Enzy-
mol. 268 (1996) 281–293.
 H. Dacres, R. Narayanaswamy / Sensors and Actuators B 107 (2005) 14–23
[36] M.N. Hughes, H.G. Nicklin, J. Chem. Soc. (A) (1971) 164–168.
[37] X.-F. Yang, X.-Q. Guo, Y.-B. Zhao, Talanta 57 (2002) 883–890.
[38] J.S. Valentine, A.B. Curtis, J. Amer. Chem. Soc. 97 (1) (1975)
224–226.
[39] B.R. Lokesh, M.L. Cunningham, Toxicol. Lett. 34 (1986) 75–84.
[40] J.A. Fee, P.G. Hildenbrand, FEBS Lett. 39 (1) (1974) 79–82.
[41] S. Kim, R. DiCosimo, J.S. Fillpo Jr., Anal. Chem. 51 (6) (1979)
679–681.
[42] H. Gampp, S.J. Lippard, Inorg. Chem. 22 (1983) 357–358.
[43] G.D. Christian, O’Reilly, Instrumental Analysis, second ed., Allyn
& Bacon Inc., Newton, MA, USA, 1986.
[44] Y. Xia, J.L. Zweier, Proc. Natl. Acad. Sci. U.S.A. 94 (1997)
6954–6958.
23
[45] A.J. Nappi, E. Vass, Biochim. Biophys. Acta 1380 (1998) 55–
63.
[46] A.A. Noronha-Dutra, M.M. Epperlein, N. Woolf, FEBS Lett. 321
(1) (1993) 59–62.
[47] N. Tsumori, Q. Xu, Bull. Chem. Soc. Jpn. 75 (2002) 1861–1862.
[48] R.J. Singh, N. Hogg, B. Kalyanaraman, Arch. Biochem. Biophys.
324 (2) (1995) 367–373.
[49] M.E. Murphy, H. Sies, Proc. Natl. Acad. Sci. U.S.A. 88 (1991)
10860–10864.
[50] M.D. Bartberger, J.M. Fukuto, K.N. Huok, Proc. Natl. Acad. Sci.
U.S.A. 98 (5) (2001) 2194–2198.
[51] M.K. Freeman, L.G. Bachas, Anal. Chim. Acta 256 (1992) 269–
275.