Fast, reversible optical sensing of NO2 using 5,10,15,20-tetrakis[3,4-bis(2-ethylhexyloxy)phenyl]-21H,23H-porphine assemblies

Article abstract of DOI:10.1039/b006342h

The optical absorbance spectrum of LB film assemblies of 5,10,15,20-tetrakis[3,4-bis(2-ethylhexyloxy)phenyl]-21H,23H-porphine (EHO) is highly sensitive to low concentrations of NO₂ in nitrogen. LB films prepared at much faster than conventional deposition rates (～ 1000 mm min^-1) yield t₅₀ response and recovery times of 25 s and 33 s respectively and show a sensitivity of 60% relative absorbance change (at 430 nm) for 4.4 ppm NO₂. The morphology of these films is revealed using atomic force microscopy to contain isolated micron-size domains which are themselves composed of grains of several nm in diameter. This unconventional structure leads to a useful sensing material as a result of the molecular functionality of the porphyrin coupled to the enhanced surface area of the porous film assembly. The EHO film shows a gradually diminishing optical response as its temperature is increased, resulting from the shift in the adsorption-desorption equilibrium towards desorption as expected. The UV-visible spectrum recovers fully after exposure to NO₂. The rate of recovery is slow at room temperature but can be accelerated dramatically with gentle heating (～ 350 K) for a few seconds. The concentration dependence of the optical response over the range 0.8-4.4 ppm follows a Langmuir model.

Full text of DOI:10.1039/b006342h

View Article Online / Journal Homepage / Table of Contents for this issue
Fast, reversible optical sensing of NO2 using 5,10,15,20-tetrakis[3,4-
bis(2-ethylhexyloxy)phenyl]-21H,23H-porphine assemblies{
Colin M. Dooling,a Oliver Worsfold,a Tim H. Richardson,*a Rose Tregonning,b
M. O. Vysotsky,b Chris A. Hunter,b Keizo Kato,c Kazunari Shinboc and Futao Kanekoc
aApplied Molecular Engineering Group, Department of Physics & Astronomy, University of
Shef®eld, Houns®eld Road, Shef®eld, UK S3 7RH. E-mail: t.richardson@shef®eld.ac.uk
bDepartment of Chemistry, University of Shef®eld, Dainton Building, Brook Hill, Shef®eld,
UK S3 7HF
cGraduate School of Science and Technology, Niigata University, Niigata, Japan
Received 3rd August 2000, Accepted 2nd November 2000
First published as an Advance Article on the web 14th December 2000
The optical absorbance spectrum of LB ®lm assemblies of 5,10,15,20-tetrakis[3,4-bis(2-ethylhexyloxy)phenyl]-
21H,23H-porphine (EHO) is highly sensitive to low concentrations of NO2 in nitrogen. LB ®lms prepared at
much faster than conventional deposition rates (y1000 mm min21) yield t50 response and recovery times of 25 s
and 33 s respectively and show a sensitivity of 60% relative absorbance change (at 430 nm) for 4.4 ppm NO2.
The morphology of these ®lms is revealed using atomic force microscopy to contain isolated micron-size
domains which are themselves composed of grains of several nm in diameter. This unconventional structure
leads to a useful sensing material as a result of the molecular functionality of the porphyrin coupled to the
enhanced surface area of the porous ®lm assembly. The EHO ®lm shows a gradually diminishing optical
response as its temperature is increased, resulting from the shift in the adsorption±desorption equilibrium
towards desorption as expected. The UV±visible spectrum recovers fully after exposure to NO2. The rate of
recovery is slow at room temperature but can be accelerated dramatically with gentle heating (y350 K) for a
few seconds. The concentration dependence of the optical response over the range 0.8±4.4 ppm follows a
Langmuir model.
gas response time (typically in the range 5±10 minutes for t50,
Introduction
the time taken for the measured sensor property to evolve to
50% of its ®nal value) and their incomplete recovery after gas
exposure.
The ability to detect toxic gases such as NO2, Cl2, HCl, SO2,
H2S, CO and NH3 in a wide range of environments has become
progressively important over the last decade as a result of
increased toxic gas production in the chemical industry.1 In
parallel with this has arrived ever more stringent health, safety
and environmental legislation which should ensure that toxic
substances have a reduced effect on human and environmental
well-being.2 However, this will only be con®rmed provided that
more sensitive gas sensors continue to be developed for
monitoring at lower concentration levels. This, coupled to the
economic requirements to produce ever less expensive sensors,
provides the motivation for further gas sensor research.
The most commonly used toxic gas sensors are those based
on semiconducting metal oxides3 such as tin oxide, the
niobates, the perovskites and the tungsten oxides. These
devices operate as a result of a gas-sensitive electrical
conductivity and normally require relatively high operating
temperatures (400±600 ³C). Other well-known sensor types
include liquid electrolyte fuel cells and infra-red sensors.3
Some attention has been directed towards organic materials
as candidates for gas sensors. Most notably, certain conducting
polymers4 arguably have achieved limited success in ``electronic
nose'' applications, and phthalocyanines have attracted a great
deal of interest as organic semi-conductometric sensors.5 On
most occasions, however, the least attractive features of the
majority of such organic sensors have been their relatively slow
Porphyrins6 and their derivatives have also been investigated
although not as thoroughly as phthalocyanines. Recently a
class of sulfonamidotetraphenylporphyrins have shown pro-
mise as useful sensing materials for chlorine in particular.7
Porphyrins are generally less conducting than most phthalo-
cyanines but share the common property of possessing rich
UV±visible absorption spectra owing to their highly conjugated
p-electron systems.8 The extended range of wavelengths now
available as narrow-band light sources in the form of either
light emitting diodes or laser diodes suggests that porphyrin
assemblies merit further study as potentially useful sensing
materials.
In this paper, the NO2 sensitivity of 5,10,15,20-tetrakis[3,4-
bis(2-ethylhexyloxy)phenyl]-21H,23H-porphine (EHO, Fig. 1)
has been studied in the solution and solid state.
Experimental
Synthesis of EHO
Materials. UV±Visible electronic absorption spectra (solu-
tion) were recorded on a Cary 3 Bio UV±Visible spectro-
photometer using quartz cuvettes. Fast atom bombardment
(zve FAB) mass spectra were obtained on a Fisons/VG Pro
Spec 3000 using a m-nitrobenzyl alcohol matrix. Column
chromatography was carried out using Florisil 100±200 mesh
(Aldrich) and silica gel (40±63 mm). Nuclear magnetic reso-
nance (NMR) spectroscopy was performed on a Bruker AM-
{Notes on ultra-fast LB deposition are available as supplementary
data. For direct elctronic access see http://www.rsc.org/suppdata/jm/
b0/b006342h/
392
J. Mater. Chem., 2001, 11, 392±398
This journal is # The Royal Society of Chemistry 2001
DOI: 10.1039/b006342h

View Article Online
This was then stirred under nitrogen for 10 minutes before
the addition of boron tri¯uoride±diethyl ether (0.25 mL,
2 mmol), after which stirring was continued under nitrogen
atmosphere and in the absence of light for a further 90 minutes.
2,3-Dichloro-5,6-dicyano-1,4-benzoquinone
(1.138 mg,
5 mmol) was then added to the mixture and the conditions
maintained for a further 90 minutes before the addition of
triethylamine (2.80 mL, 20 mmol). Solvent was then removed
under vacuum and the resulting solid was extracted with
methanol (326100 mL) using a soxhlet apparatus. The residue
was then puri®ed by column chromatography [Florisil, eluant
chloroform±petroleum ether (bp 40±60 ³C), 38 : 62]. The
porphyrin fractions were further puri®ed by column chroma-
tography [silica, eluant chloroform±petroleum ether (bp 40±
60 ³C), 38 : 62]. The product was then recrystallised from
chloroform±methanol to yield a purple solid (655 mg, 32%).
1H NMR 250 MHz (CDCl3): d (ppm)~8.91 (s, 8H), 7.76 (d,
Fig. 1 Chemical structure of EHO.
4
3
4H, JHH~2.0 Hz), 7.69 (d, 4H, JHH~8.2 Hz), 7.23 (d, 4H,
3JHH~8.2 Hz), 4.17 (d, 8H, 3JHH~5.8 Hz), 3.97 (d, 8H,
250 MHz or a AMX-400 MHz spectrometer as indicated. All
operated as Fourier transform machines. All spectra were
referenced to CHCl3. Melting points were performed on a
Reichter Ko¯er hot stage melting point apparatus.
Dichloromethane was supplied by Fisher and was stabilised
with amylene (0.02%). It was puri®ed and dried by distillation
over calcium hydride. Dry ethanol was obtained by distillation
3
3JHH~5.8 Hz), 1.95 (septet, 4H, JHH~6.0 Hz), 1.81 (septet,
4H, 3JHH~6.0 Hz), 1.12±1.80 (m, 64H), 0.8±1.10 (m, 48H),
22.8 (s, 2H). 13C NMR 250 MHz (CDCl3): d (ppm)~149.4,
147.5, 134.9, 127.6, 120.7, 120.1, 111.6, 71.9, 41.0, 39.9, 39.2,
31.9, 30.8, 29.8, 29.3, 25.2, 24.1, 23.2, 14.2, 11.4. FAB-MS
(zve), m/z: 1642 (100%, MHz) (Calc. for C108H158N4O8
1640.46).
UV±VIS
(CHCl3):
lmax/nm
Ê
over molecular sieves (4 A) and stored over molecular sieves
Ê
(e61023 mol21 l cm21)~427 (368), 522 (18), 556 (13), 592
(7), 651 (7). Melting point: 146 ³C.
Ê
(4 A). Triethylamine was dried by storage over 4 A molecular
sieves.
3,4 Dihydroxybenzaldehyde and 2-ethylhexyl bromide were
supplied by Lancaster. Aldrich supplied all other reagents. All
reagents were used as received except pyrrole, which was
distilled fresh on the day of use.
Langmuir isotherm measurement
The Langmuir isotherm for EHO was measured using a NIMA
1060 Langmuir trough. EHO was spread from a chloroform
solution (50 ml) of concentration 0.129 mg ml21 onto a water
subphase using a Hamilton microsyringe (pHy6.2, 296 K).
After evaporation of the solvent, the ¯oating Langmuir ®lm
was compressed at a rate of 15 cm2 min21 (y30% min21).
Synthesis
The synthesis and structure of 5,10,15,20-tetrakis[3,4-bis(2-
ethylhexyloxy)phenyl]-21H,23H-porphine (EHO) is outlined
below.
Ultra-fast deposition technique
3,4-Bis(2-ethylhexyloxy)benzaldehyde. 2-Ethylhexyl
bro-
Conventionally, the Langmuir±Blodgett deposition process is
considered to be a relatively slow coating process and indeed,
this is one of the reasons it has remained commercially
unattractive. Although there have been attempts to deposit
conventional fatty acid materials at high transfer rates,9 very
seldom have linear deposition rates greater than 50 mm min21
been adopted for more exotic LB ®lm-forming materials. In
this work, very high deposition rates of 1000 mm min21 were
used for the porphyrin under study. Hydrophobically treated
glass substrates (BDH Super Premium microscope slides
treated with 1,1,1,3,3,3-hexamethyldisilazane) were used, and
the gain control of the constant-perimeter Langmuir trough
(Joyce±Loebl mini-trough) was set at its maximum value to
allow the compression barrier to respond as quickly as possible
to the slight surface pressure decrease that occurred during
transfer. This is particularly important when the transfer rate is
high.
mide (5.60 mL, 0.031 mol) was added dropwise to a stirred
solution of 3,4-dihydroxybenzaldehyde (2.07 g, 0.015 mol) and
potassium carbonate (6.20 g, 0.045 mol) in butanone (60 mL)
under conditions of re¯ux. These conditions were maintained
for a period of four days after which the reaction mixture was
®ltered through Celite and the solvent was removed under
vacuum to yield a dark brown oil. This was then re-dissolved in
ether (50 mL), washed with 1 M NaOH (1650 mL), H2O
(1650 mL) and brine (1650 mL) and then dried over Na2SO4.
The solution was again ®ltered and the solvent was removed
under vacuum before puri®cation by column chromatography
[silica, eluant petroleum ether (bp 40±60 ³C)±ethyl acetate, 9 : 1]
to give the product as a mixture of diastereoisomers in the form
of a pale yellow oil (4.49 g, 83%).
1H NMR 250 MHz (CDCl3): d (ppm)~9.81 (s, 1H), 7.39
3
4
(dd, 1H, JHH~7.9 Hz, JHH~1.8 Hz), 7.37 (s, 1H), 6.92 (d,
1H, 3JHH~7.9 Hz), 3.91 (t, 4H, 3JHH~6.1 Hz), 1.76 (septet,
2H, 3JHH~4.3 Hz), 1.2±1.6 (m, 16H), 0.8±1.0 (m, 12H). 13C
NMR 250 MHz (CDCl3): d (ppm)~200.8, 190.9, 155.0, 151.9,
149.8, 149.6, 142.5, 130.8, 129.8, 127.2, 126.6, 123.8, 122.9,
112.7, 111.8, 111.4, 110.4, 71.5, 71.3, 60.3, 39.6, 39.5, 39.4, 39.2,
33.6, 30.9, 30.6, 30.5, 29.1, 28.8, 23.9, 23.8, 23.7, 23.0, 22.8,
22.6, 20.9, 19.4, 14.0, 13.9, 11.2, 11.0, 8.3. HR-FAB-MS, m/z:
363.2890 (Mz) (Calc. for C23H38O3 363.2899).
Atomic force microscopy
The atomic force microscopy (AFM) was carried out at room
temperature in air by tapping mode using a Nanoscope IIIa
(Digital Instruments) microscope.
Gas testing
5,10,15,20-Tetrakis[3,4-bis(2-ethylhexyloxy)phenyl]-
The thin ®lm samples were sliced into small segments
approximately 1 cm62.5 cm and placed in a purpose-built
gas testing chamber (Fig. 2). This consisted of a gas inlet and an
outlet, a Peltier heating±cooling stage and housings for ®bre
optic cables supplying the probe light beam from the tungsten
21H,23H-porphine. In the absence of light 3,4-bis(2-ethylhex-
yloxy)benzaldehyde (1813 mg, 5 mmol) was dissolved in a
mixture of dry dichloromethane (500 mL), dry ethanol
(3.25 mL) and freshly distilled pyrrole (347 mL, 5 mmol).
J. Mater. Chem., 2001, 11, 392±398
393

View Article Online
Fig. 2 Schematic diagram of gas testing apparatus.
white light source and delivering the light transmitted through
the sample to the optical detector. Two gas cylinders were used,
one containing 4.4 ppm (parts per million) NO2 (precisely
mixed by BOC, UK) and the other containing dry nitrogen.
The ¯ow rates of these gases were moderated by two computer-
controlled Tylan FC-260 mass ¯ow controllers in order that the
gas concentration after subsequent mixing could be precisely
selected in the range 0.1±4.4 ppm. In normal operation, the
exposure of the sample to the NO2 occurred at room
temperature (y293 K) and the recovery phase (dry nitrogen
only) occurred at elevated temperature (353 K). These condi-
tions were used in all experiments unless otherwise stated.
Optical measurements were performed using a World
Precision Instruments ``Spectromate'' spectrophotometer.
This instrument utilises a multi-channel photodiode array
and can therefore enable spectra across the entire visible region
to be collected in a short period of time. In this study, data were
collected typically every 2±4 s during exposure of the sample to
the toxic gas stream and to the dry nitrogen gas during the
recovery cycle. In addition to the spectra being recorded every
few seconds, the instrument would also display changes in the
optical absorbance (at selected wavelengths) as a function of
time, thus allowing the kinetics of the adsorption and
desorption processes to be followed.
19 mN m21. Extrapolation of the former quasi-solid region of
the isotherm to zero surface pressure yields an area per
molecule of y125 A . This value most probably represents an
Ê 2
orientation in which the porphyrin ring faces down onto the
water surface with the substituted alkyl groups protruding
from the plane of the water surface. Modelling (using CPK
space-®lled models) of such an orientation would yield an area
Ê 2
per molecule of y150 A ; the difference in the two values
occurs most probably as a result of either a degree of
aggregation within the monolayer (i.e. non-monolayer aggre-
gate regions within a monolayer host) or a tilted porphyrin
orientation in which the plane of the porphyrin core is tilted
with respect to the plane of the water surface. Normally,
thickness studies could be used to verify the orientation of the
molecules within a Langmuir ®lm. Here, however, as a result of
the ultra-fast deposition technique employed, the resulting
inhomogeneous, perforated ®lm structure does not provide a
well-de®ned ®lm thickness to be used in conjunction with
isotherm measurements. In common with conventional LB
®lms deposited more slowly (typically 0.1±10 mm min21), LB
®lms produced via ultra-fast deposition do grow in a layer-by-
layer fashion as evidenced by the gradual linear build up in
optical absorbance as a function of number of passages of the
substrate through the ¯oating Langmuir ®lm (ESI). The
average optical absorbance per monolayer is y0.025 which
is consistent with monolayer values obtained previously for
porphyrins and phthalocyanines by other authors.7,8 In this
measurement, a relatively large light beam width (y2 mm) has
been used in order to average over many EHO domains.
Results and discussion
Surface pressure±area isotherm
Fig. 3 shows the surface pressure±area isotherm of EHO. A
quasi-solid region of the isotherm is observed over the surface
pressure range 8±18 mN m21 before a slight kink appears prior
to the onset of another incompressible phase occurring above
Atomic force microscopy
Fig. 4 is an atomic force microscope image of a 20 monolayer
EHO ®lm deposited at 1000 mm min21 onto an optically ¯at
glass substrate. The image shows a disordered array of
``islands'' of EHO distributed over the surface. The underlying
substrate can be observed between some of these islands (and
indeed has been imaged separately in addition) and indicates
that the topology of the EHO assembly does not arise simply as
a result of the substrate surface roughness. It is thought that the
mechanism for the formation of these islands arises from
dewetting of the substrate ®lm during transfer from the water
surface.
Traditionally, researchers have de®ned useful LB ®lms to be
highly uniform, ordered and relatively dense molecular
assemblies. However, for certain applications these features
are actually detrimental to performance. One such application
is the main topic of this paper, namely gas sensing. In this case,
Fig. 3 Surface pressure±area isotherm for EHO.
394
J. Mater. Chem., 2001, 11, 392±398

View Article Online
Fig. 4 Atomic force microscope image of a 10 excursion LB ®lm of EHO deposited at 1000 mm min21
.
a relatively porous macrostructure is required in order to
promote ingress and egress of the analyte gas into and from the
thin ®lm. This also enhances the surface area of the sensing
®lm. Thin ®lm gas sensing is primarily characterised by a fast
surface adsorption process (often followed by a much slower
diffusion into the bulk) and therefore modi®cations to the
macrostructure that augment the surface area are advanta-
geous.
several hours after switching off the gas stream. It should be
noted that there are no bathochromic or hypsochromic shifts in
the wavelengths of any of the three main spectral bands present
(425 nm, 465 nm and 700 nm) during the exposure. This is not
surprising since such shifts are normally indicative of a change
in the aggregation state of the solute molecules. In the dilute
solution (%0.1 mg ml21) however, little or no aggregation
exists and therefore no such shifts are expected.
Analysis of this image has shown that the height of the EHO
domains within the LB assembly is signi®cantly larger than the
expected height for 20 monolayers. One might have expected a
thickness per transferred layer of around 1.3 nm, but a height
pro®le study of the image shown reveals that the typical
domain height is around y55 nm. The transfer ratio
(corresponding to the ratio of the area of the ¯oating
monolayer to the area of the substrate drawn through the
air±water interface) is close to unity. Although a qualitative
approach, the average domain height is y2.7 nm per
transferred monolayer (55 nm per 20 ``layers'') and observation
of the image suggests that the surface coverage is approxi-
mately 50%. This supports the idea that material absent from
certain uncovered areas of the substrate has been ``drawn into''
the domains of EHO during or just after deposition. The EHO
domains are typically around 0.860.8 mm in dimension
although the distribution of domain size is wide; some of
these coalesce together to form larger features. The surface
Fig. 6 shows the temporal evolution of the absorbance
spectrum of a nominally 6 monolayer transferred ®lm (note
Fig. 5 The time evolution of the UV±visible spectrum of a chloroform
solution of EHO during exposure to 4.4 ppm NO2 gas (time interval is
10 s).
Ê
roughness of each domain is typically y4±12 A and that of the
Ê
substrate is around 3±4 A. Each domain is composed of small
``grains'' of characteristic lateral dimension y30 nm. The
growth of these grains and the formation of the EHO domains
is a major topic of further investigation.
UV±visible spectroscopy: gas-sensitive optical properties
The UV±visible absorbance spectra of a solution of EHO (with
a chloroform solvent) through which an NO2 gas stream
(y4 ppm) was bubbled are shown in Fig. 5. Before any
exposure to NO2, the characteristic porphyrin spectrum is
observed with an intense Soret band appearing at y425 nm
along with weaker Q-bands in the region 500±700 nm. As the
bubbling proceeds,
a dramatic change in the spectrum
develops, involving the weakening of the Soret band intensity
and the simultaneous growth of a band around 465 nm which is
coupled to a further band at 700 nm. Such spectral changes
could indicate protonation of the porphyrin or oxidation. The
absorbance spectrum reverses (not shown) to its original form
Fig. 6 The time evolution of the UV±visible spectrum of an LB ®lm (3
excursions) of EHO during exposure to 4.4 ppm NO2 gas (time interval
is 4 s).
J. Mater. Chem., 2001, 11, 392±398
395

View Article Online
Table 1 The t50 and t90 values for response and recovery phases for an
EHO LB ®lm
t50/s
t90/s
Response phase
Recovery phase
25
33
190
73
(353 K) can be seen from Fig. 7 which shows the absorbance of
the Soret band (430 nm) as a function of time. The decay of the
Soret band intensity during the gas exposure has been found to
®t an Elovichian kinetics model11 most readily. This model
states that the rate of coverage, q, of the ®lm surface (dq/dt) is
given by eqn. (1):
dq=dt~ae{bq
(1)
where a and b are constants. This equation leads to an equation
for the coverage of the surface itself being given by eqn. (2):
q~(1=b)ln(t)zk
(2)
where k is also a constant. Thus, the probability of adsorption
of an NO2 gas molecule decreases exponentially as a function
of the number of NO2 molecules already adsorbed. Fig. 8
shows the plot of the decrease in the Soret band intensity (this is
proportional to the coverage) versus ln (t). The relationship is
linear until high values of t indicating that the Elovich model
®ts the surface adsorption process well. The deviation from
linearity at high t results from the slow diffusion process in
which NO2 diffuses into the buried layers of the ®lm.
Fig. 7 Absorbance of the Soret band (430 nm) of an LB ®lm of EHO (6
excursions) as a function of time during exposure to 4.4 ppm NO2.
that this refers to a sample which has made 3 transits through
the ¯oating monolayer in both directions). The initial pre-
exposure spectrum (highest intensity of the Soret band) is
signi®cantly broader than that of the solution (Fig. 5); its
FWHM is y45 nm compared with y17 nm for the solution.
This is attributed to the crystal ®eld effect for the relatively
closely packed porphyrin molecules in the thin ®lm assembly.
The peak wavelength of the Soret band for the solid ®lm is
bathochromically shifted by y7 nm with respect to the
solution, and this shift further develops during the gas
exposure. Similarly, the lower wavelength band forming as a
result of the gas exposure develops at 480 nm compared to
465 nm for the solution. These bathochromic shifts are again
attributed to the aggregation state of the porphyrin in the thin
®lm.10 The band at 700 nm also evolves in the solid state,
although it is relatively weak and very broad. It should be
noted that no sensitivity to NO2 gas was observed for a solution
cast ®lm prepared by simply allowing droplets of the chloro-
form solution to fall onto the substrate.
Of speci®c use here, however, are the parameters, t50 (t90)
(the time taken for the Soret band intensity to fall to 50% (90%)
of its ®nal saturated value) used to characterise the speed of
response and recovery. Table 1 indicates these values for both
the response phase and the recovery phase. The t50 for the
response phase at 293 K is relatively fast compared to many
other reports of organic gas-sensing materials at room
temperature12 and is of some interest commercially. Typically,
many phthalocyanines have yielded t50 response times around
5±10 minutes. The improved response time of EHO assemblies
prepared by ultra-fast deposition is attributed to their
enhanced surface area compared to conventional LB ®lms.
The response of gas sensors as a function of gas concentra-
tion is important as it indicates the range of concentration over
which the response is linear and yields the sensitivity. Fig. 9
represents a sequence of gas exposure±recovery cycles per-
formed for different concentrations of NO2 gas over the range
0.88±4.4 ppm. For all concentrations, the temperature of the
sample was 293 K for the exposure phases and 353 K for each
recovery phase. Heating and cooling between these tempera-
tures using the Peltier device took only y20 s. The ®gure
clearly shows that the response increases with increasing
concentration. Fig. 10 (inset) shows the magnitude of the
optical response (total change in Soret band absorbance during
The kinetics of the gas exposure (293 K) and recovery
Fig. 9 A sequence of gas exposure (NO2)±recovery (N2) cycles
performed over the concentration range 0.88±4.4 ppm.
Fig. 8 Elovich plot for an EHO LB ®lm.
396
J. Mater. Chem., 2001, 11, 392±398

View Article Online
Fig. 12 Temperature dependence of the recovery phase over the range
323±353 K. The temperature for the exposure phase is 293 K in all
cases.
Fig. 10 Langmuir adsorption plot for an EHO LB ®lm. (Inset: Change
in Soret band absorbance (for saturation) as a function of NO2 gas
concentration.)
temperature of 353 K. It can be seen that the optical response
to the gas exposure decreases rapidly as the sample temperature
increases. This is expected since the dynamic equilibrium
position between adsorption and desorption of NO2 is shifted
towards desorption as the temperature rises.
Fig. 12 shows the Soret band absorbance as a function of
time for a sequence of exposure±recovery cycles in which the
temperature during the gas exposure was maintained at 293 K,
but the temperature of the recovery phase was varied over the
range 323±353 K. It can be clearly seen that the rate of recovery
is accelerated as the sample temperature increases; again, this
arises as a result of the shift in the dynamic equilibrium
position in favour of desorption. The speci®c form of the
recovery process shows several distinct stages that probably
represent desorption from surface or bulk sites possessing
different NO2 binding strengths. Such variation in binding
strength may arise due to surface heterogeneity. This result,
combined with the temperature dependence of the optical
response during NO2 exposure, indicates that the optimum
operating conditions for sensors based on EHO are to cycle
between 293 K and 353 K for the exposure and recovery cycles
respectively. In this way, fast response and recovery rates can
be achieved.
Fig. 11 A sequence of exposure cycles performed over the temperature
range 293±323 K. The temperature for the recovery phase is 353 K in
all cases.
a gas exposure cycle) as a function of the gas concentration.
This curve is commonly observed in cases in which Langmuir
adsorption occurs. Langmuir's adsorption theory13 suggests
that a plot of c/DAbs versus c (where DAbs is the change in
Soret band absorbance and c is the gas concentration) should
yield a straight line if the assumptions on which it is based are
valid. These assumptions are that the activation energy for
adsorption is the same for all binding sites in the thin ®lm
assembly, that there is no motion of adsorbed NO2 molecules
over the surface and that NO2 molecules striking a surface site
that is already occupied do not adsorb. This latter assumption
limits the adsorption of NO2 onto a thin organic ®lm surface to
a single monolayer. Fig. 10 shows that a highly linear
relationship is indeed obtained. Even though the Langmuir
model ignores lateral interactions between adsorbed NO2
molecules, surface mobility, surface heterogeneity and the
possibility of adlayers{ thicker than one monolayer, it never-
theless provides a basic understanding of the adsorption
process for thin ®lm assemblies of EHO and NO2. Other more
complex models such as the Braunauer±Emmett±Teller
model13 allow multilayer adsorption to occur and will be
investigated in future studies in order to observe whether an
improved theoretical understanding of the experimental data
results.
Summary and conclusions
The optical absorbance spectrum of thin ®lm assemblies
containing EHO has been shown to be highly sensitive to low
concentrations of NO2. Thin ®lm assemblies prepared using an
ultra-fast LB deposition method have yielded a t50 response
time of y20 s. Relative absorbance changes of around 60% for
4.4 ppm NO2 have been found. The morphology of these
assemblies has been studied using atomic force microscopy; the
deposited ®lms contain isolated domains which themselves
contains grains of around 30 nm diameter. Future work will
focus on studying the sensitivity of EHO assemblies to other
toxic gases such as Cl2 and HCl, in order to examine its
selectivity. Further kinetics modelling over
a range of
temperature will also be performed to further develop our
understanding of the adsorption and desorption processes
active in this sensing system. In this way, activation energies for
the adsorption and desorption processes will be obtained.
The magnitude of the optical absorbance change of the Soret
band, and the rate of response and recovery, are also
temperature dependent parameters. Fig. 11 shows a sequence
of exposure cycles performed over a range of temperatures. In
all cases, the recovery phase is exacted at an elevated
Acknowledgement
The authors would like to thank former members of the group
for their initial involvement in this gas sensing project,
particularly Dr Chris George and Dr Ginny Smith. We
would also like to thank TQ Environmental plc for their
interest in this work and EPSRC for the provision of a PhD
studentship.
{An adlayer is the layer (or layers) of material deposited during one
decomposition cycle.
J. Mater. Chem., 2001, 11, 392±398
397

View Article Online
and A. M. d'A, Rocha-Gonsalves Thin Solid Films 1998, 327±329,
3151.
References
8
9
C. D. George, T. Richardson, M. E. Hofton, C. M. Vale,
M. G. M. Neves and J. A. S. Cavaleiro, Mater. Sci. Eng. C, 1999,
8±9, 559.
I. R. Peterson, G. J. Russell and G. G. Roberts, Thin Solid Films,
1983, 109(4), 371.
1
W. Gopel, J. Hesse and J. N. Zemel, Sensors: A Comprehensive
Survey. Volume 2. Chemical and Biochemical Sensors, 1991, VCH,
Weinheim, Part 1, p. 34.
R. Das and P. D. Blanc, Toxicol. Industrial Health, 1993, 9,
439.
P. T. Moseley, J. O. W. Norris and D. E. Williams, Techniques and
mechanisms in gas sensing, Adam Hilger, Bristol, 1991.
M. C. Lonergan, E. J. Severin, B. J. Doleman, S. A. Beaber,
R. H. Grubb and N. S. Lewis, Chem. Mater., 1996, 8(9), 2298.
S. Capone, S. Mongelli, R. Rella, P. Siciliano and L. Valli,
Langmuir, 1999, 15(5), 1748.
L. R. Milgrom, The Colours of Life, Oxford University Press,
Oxford, 1997.
T. Richardson, V. C. Smith, R. A. W. Johnstone, A. J. F. N. Sobral
2
3
4
5
6
7
10 The following paper O. Worsfold, C. M. Dooling, T. H. Richardson,
M. O. Vysotsky, R. Tregonning, C. A. Hunter and C. Malins,
J. Mater. Chem., in the press, DOI:10.1039/b006344o.
11 M. C. Manchado, J. M. Guil and A. R. Paniego, J. Chem. Soc.,
Faraday Trans., 1989, 85(7), 1775.
12 A. K. Ray, M. J. Cook, S. C. Thorpe and S. Mukhopadyay, Phys.
Status Solidi A, 1993, 140(2), K85.
13 J. B. Hudson, Surface Science: An Introduction, Wiley, Chichester,
1998.
398
J. Mater. Chem., 2001, 11, 392±398