Photophysical and photochemical characterisation of bacterial semiconductor cadmium sulfide particles

Article abstract of DOI:10.1039/a708742j

Klebsiella pneumoniae forms electron-dense cadmium sulfide particles (ca. 5-200 nm in diameter) on the cell surface in response to the presence of cadmium ions in the growth medium. In the current study, these 'bio-semiconductor' particles have been spectroscopically characterised using UV-VIS absorption and luminescence analysis. The spectroscopic properties observed suggest that they are similar in size and possess photoactive traits analogous to CdS systems prepared by conventional chemical methods. The optical nature of the bacterial semiconductor particles means that, in principle, they are capable of performing a variety of photoredox reactions. The reactions involving photoelectrochemical indicators such as methyl viologen (MV²⁺) and methyl orange (MO^-) are considered and, by comparing initial rates of reaction and altering reaction variables, a general mechanism of photoactivity for the cadmium sulfide 'bio-semiconductor' is proposed.

Full text of DOI:10.1039/a708742j

View Online / Journal Homepage / Table of Contents for this issue
Photophysical and photochemical characterisation of bacterial
semiconductor cadmium sulÐde particles
Peter R. Smith,a Justin D. Holmes,a David J. Richardson,b David A. Russella and
John R. Sodeaua*
a School of Chemical Sciences and b School of Biological Sciences, University of East Anglia,
Norwich, UK NR4 7T J
Klebsiella pneumoniae forms electron-dense cadmium sulÐde particles (ca. 5È200 nm in diameter) on the cell surface in response to
the presence of cadmium ions in the growth medium. In the current study, these “bio-semiconductorÏ particles have been spectro-
scopically characterised using UVÈVIS absorption and luminescence analysis. The spectroscopic properties observed suggest that
they are similar in size and possess photoactive traits analogous to CdS systems prepared by conventional chemical methods. The
optical nature of the bacterial semiconductor particles means that, in principle, they are capable of performing a variety of
photoredox reactions. The reactions involving photoelectrochemical indicators such as methyl viologen (MV2`) and methyl
orange (MO~) are considered and, by comparing initial rates of reaction and altering reaction variables, a general mechanism of
photoactivity for the cadmium sulÐde “bio-semiconductorÏ is proposed.
The bacterium Klebsiella pneumoniae is able to transform
cadmium ions into nanometer-sized cadmium sulÐde particles
which are deposited onto the cell surface. Such a biosynthetic
mechanism overcomes the toxic e†ect of any cadmium species,
which may be present in the bacteriumÏs local environment.
Aspects of this biochemical process have been reported in the
literature, including the kinetics of particle formation and
their growth characteristics.1h5 It was concluded from pre-
vious results that the large “superparticlesÏ observed by elec-
tron microscopy (EM), e.g. 200 nm, were actually built up
from smaller particle aggregates ca. 4 nm in diameter.5 In con-
ventional chemical studies these small dimension species are
termed “Q-particlesÏ and have been shown to possess distinc-
tive photochemical and photophysical properties. Therefore
the main aim of the present investigation was to compare and
contrast the electronic and photoredox properties of the bio-
particles with their well studied chemical counterparts.
There have been many reports in the literature relating to
the photoprocesses associated with colloidal semiconductor
systems.6h9 The majority are related to particles produced
and supported within chemically synthesised environments,
e.g. microemulsions, but few such investigations of biologically
generated semiconductor particles have been reported.10h13
One of the earliest reports was centred on the electronic spec-
troscopy associated with CdS “peptide-cappedÏ intracellular
particles formed within the yeasts Candida glabrata and
Schizosaccharomyces pombe.11h18
The photochemistry of bipyridinium salts, such as methyl
viologen (MV2`) and methyl orange (MO~) are well docu-
mented.22h30 The reduced forms of both MV2` and MO~
display strong absorption bands in the visible region of the
spectrum, making them easy to monitor in any investigation
of their photoredox chemistry. In fact, the viologens exist in
three main oxidation states,24,25 namely V2` H V~~ H V0.
The e†ects of irradiating chemically, synthesised CdS colloi-
dal systems in the presence of MV2` and MO~ have also
been previously reported.20,31 The reduction of MV2` ions to
MV~` radical ions can be monitored from the electronic
absorption spectrum as the radical cation is blue with absorp-
tion bands at j ca. 600 and 465 nm.
In some cases it has been reported that the MV~` radical
ion can re-form the parent MV2` by passing an electron to
another acceptor: this process is termed “electron relayÏ. For
example, in mixed systems of methyl viologen and methyl
orange with colloidal CdS an unexpected enhancement of
MO~ reduction can be explained in terms of this e†ect.29,30
The kinetics of methyl orange reduction photosensitised by
colloidal CdS have also been well documented.27h29 The rate
determining step has been established as the reduction of the
unprotonated form of the MO~ to the semi-reduced radical
HMOÕ~ via reaction (1):27,28
MO~ ] H` ] e~ ] HMO~~
(1)
(2)
(3)
HMO~~ ] H` ] e~ ] H MO~
2
Recently some of the key photobiological and spectroscopic
features associated with the K. pneumoniae system have been
published. For example, CdS bio-particles have been shown to
photoprotect the bacterium from destructive UV-A radiation
(j \ 320È400 nm)2 and to produce both oxygen- and carbon-
based radicals upon illumination with visible light.19 Further-
more it has been established that the onset of absorption
2HMO~~ ] H MO~ ] MO~
2
The subsequent reactions which lead to hydrazine formation
[reactions (2) and (3)] are much faster than reaction (1). MO~
has an absorption maximum at j \ 463 nm, whereas the
max
secondary hydrazine derivative is colourless (j \ 247 nm).
max
It is these absorption characteristics that lead to the use of
(j
) for the extracellular bio-particles of CdS lies in the
wavelength range between 440È450 nm with an absorption
methyl orange as a probe for photoredox reactions since the
reduction of the unprotonated form can be followed by moni-
toring the decrease in absorbance of the system at a wave-
length of 463 nm.26h29
onset
maximum wavelength (j ) between 400 and 410 nm.2,5,19
This absorption proÐle corresponds to a particle size of ca. 4
max
nm from a Henglein plot and an energy band gap of 2.9È3.25
eV.20,21 Preliminary accounts of CdS bio-particle photoredox
properties have also been given and this behaviour is dis-
cussed in some detail within this report because of the poten-
tial applications in light-driven electron-transfer reactions.19
In the present study, experiments to determine the CdS bio-
particlesÏ luminescence and to evaluate the extent of their pho-
toredox properties were performed. Contrasts between the
photochemical and photophysical behaviour of the bacterial
CdS particles compared to their well studied conventional
J. Chem. Soc., Faraday T rans., 1998, 94(9), 1235È1241
1235

View Online
with an initial pH of ca. 7, were set to the desired pH using
chemical counterparts should be readily apparent. Several key
factors such as the e†ect of pH on the reduction rates of the
positively charged viologen, MV2` and the negatively charged
MO~ were monitored.
either HCl (2 M) or NaOH (2 M). The individual photoredox
species, i.e. methyl viologen (MV2`, 100 lM) or methyl orange
(MO~, 20 lM) was added and the sample transferred to a
vacuum-line cell attachment.
All photoredox reactions required samples to be degassed
thoroughly before illumination to avoid secondary oxidation
of the photoredox species by molecular oxygen. Removal of
oxygen was achieved by the “freezeÈpumpÈthawÏ method,
repeated four times, using a cardiceÈacetone mixture to freeze
the sample with a pressure of 0.1 mbar. The sample was kept
under vacuum throughout the experiment and was irradiated,
between wavelengths of 350 and 700 nm, with a medium pres-
sure mercury lamp. A water Ðlter cell (10 cm pathlength) was
used between the lamp and the sample to remove IR radi-
ation. The reaction was monitored using a Hewlett Packard
HP 8452A diode-array spectrophotometer using HP 89532A
software, with absorbance measurements taken at wavelengths
averaged over a 4 nm range. An absorption reading was
recorded every 10 s, with an integration time of 1 s. The data
were processed using the “CalculationsÏ function available on
the kinetics software (HP 89532A), allowing best Ðt treatment
to be performed.
Experimental
Growth of bacterium and formation of CdS crystallites
The Class II pathogen Klebsiella pneumoniae (previously
known as Klebsiella aerogenes) NCIMB 418 was used
throughout this study. A modiÐcation of the growth medium
used by Pickett and Dean32 was used for batch culturing:
(contained in 1 l Milli-Q water) 3.2 mg FeSO É 7H O, 64 mg
4
2
MgSO É 7H O, 0.62 g KCl, 0.60 g sodium b-glycerophosphate
4
2
and 0.96 g (NH ) SO as the main components in tricine (50
4 2
4
mM) bu†ered to pH 7.6 using NaOH (2 M). Batch cultures of
K. pneumoniae were grown under aerobic conditions in sterile
conical Ñasks (250 ml) by inoculating 50 ml of the medium
with 100 ll K. pneumoniae stock suspension in the absence of
cadmium ions (“unloadedÏ). To produce the samples contain-
ing “bio-CdSÏ particles on the bacterial surface (“loadedÏ),
cadmium ions [Cd(NO ) ] were added at a concentration of 1
mM and the Ñasks were then shaken for 24 h at 310 K. At the
3 2
Experiments involving methyl viologen (100 lM) were moni-
tored at wavelengths of 598, 600 and 602 nm. A high-pass
optical Ðlter ([515 nm) was positioned over the entrance slit
of the detector to protect it from the high light intensity
emitted by the medium pressure mercury lamp. In the case of
methyl orange (20 lM), the disappearance of the oxidised form
which absorbs between j \ 462 and 466 nm was monitored. A
high-pass optical Ðlter ([400 nm) was used to protect the
detector in this instance. Studies of the competitive photore-
duction reaction using a composite solution of methyl violo-
gen (100 lM) and methyl orange (20 lM) were monitored at
two wavelengths. The formation of reduced methyl viologen
was followed by the absorption at j \ 600 nm, and the disap-
pearance of the oxidised form of methyl orange monitored at
j \ 464 nm. A high-pass Ðlter ([400 nm) was used to protect
the detector. All absorbance measurements were referenced
against growth medium (no glucose).
time of inoculation the approximate cell density was 106 cells
ml~1. After 24 h growth the cell density was ca. 109 cells
ml~1. For the loaded samples this corresponded to a CdS
concentration of approximately 0.2 mM.1,5
Electronic spectroscopy
UVÈVIS absorption measurements were taken of loaded and
unloaded bacterial cultures (3 ml, ca. 109 cells ml~1, 24 h
growth), over a pH range 3È12. The appropriate pH was
obtained by the addition of either HCl (1 M) or NaOH (1 M) to
the bacterial suspension. An Hitachi U-4001 double beam
spectrophotometer incorporating a barium sulfate (BaSO )
4
integrating sphere attachment was used to record the spectra
of the loaded samples, referenced against unloaded bacterial
samples of the appropriate pH. The integrating sphere collects
the scattered light from the turbid samples and redirects it to
the detector, hence reducing the loss of light through scat-
tering. Absorption spectra were recorded using 1 cm path-
length quartz spectroscopic cuvettes.
Results and Discussion
Electronic spectra of extracellular CdS particles
Luminescence excitation and emission spectra were record-
ed for loaded and unloaded bacterial cultures (3 ml, ca. 109
cells ml~1, 24 h growth), between a pH range 3È12 [obtained
by adjustment with either HCl (1 M) or NaOH (1 M)], and
over a range of excitation and emission wavelengths. The
luminescence characteristics were measured using a Perkin-
Elmer LS-50 luminescence spectrometer, with a front surface
accessory (FSA) attachment.
Absorption spectra. It has been previously reported that at
pH 7 the onset of absorption (j
onset
ticles of K. pneumoniae lies between the wavelength range
) for extracellular CdS par-
440È450 nm with an absorption maximum wavelength (j
)
max
around 400È410 nm.2,5,19 Therefore as a necessary prelude to
an investigation of the e†ect of pH on the photochemical and
photophysical properties of CdS bio-particles, potential modi-
Ðcations to their electronic absorption spectra by pH were
monitored.
Photoredox reactions
To harvest cells for photoredox studies loaded and unloaded
bacterial samples (250 ml) taken from batch cultures (24 h
growth) were centrifuged using a Sorval RC-5B Refrigerated
Superspeed centrifuge (10 000g for 20 min). The supernatant
was then removed and the resultant pellet washed in growth
medium (minus glucose, 10 ml). This procedure was repeated
three times before the pellet was resuspended in the growth
medium (minus glucose, 10 ml). For the loaded samples this
corresponded to a CdS concentration of ca. 4 mM.1,5 These
concentrated bacterial solutions were then sonicated using a
Jencons Sonics and Materials Inc. Vibra-Cell (VCX 400W)
with a standard horn (13 mm) and replaceable tip for 2 h
(amplitude 48%; pulse on 2.5 s, o† 1.0 s). The sonicated
samples were then diluted, approximately ten-fold
([CdS] B 0.4 mM), with growth medium (no glucose) to give
The absorption proÐle of samples loaded and unloaded
with CdS particles were recorded using a UVÈVIS spectro-
photometer incorporating an integrating cell attachment. Fig.
1 shows the absorption proÐles obtained at pH 3, 7, 10 and 12
for K. pneumoniae loaded with CdS particles (using the corre-
sponding unloaded samples as reference).
The proÐles clearly show that over a pH range between 3
and 10, j
lies between 394 and 399 nm. The j
, which is
max
onset
obtained by extrapolating the straight line region of the
absorption spectrum (between the absorption tail and the
shoulder) to the point which intersects the absorption tail, was
found to be hypsochromically shifted at low pH values, e.g. at
pH 3, j
B 475 nm and between pH 7 and 10, j
B 440
onset
onset
B 450 nm). The
nm. Increasing the pH to 12 resulted in an apparent red-shift
of j
by 18È23 nm (j B 417 nm, j
max
max
onset
an optical density (j \ 650 nm) of 1.0. These samples (4 ml),
intensity of the electronic absorption spectra also appeared to
OD
1236
J. Chem. Soc., Faraday T rans., 1998, V ol. 94

View Online
Fig. 1 Electronic absorption spectra of CdS loaded K. pneumoniae
cells at various pH values. Spectra recorded using the integrated cell
attachment, referenced against unloaded samples.
decrease with increasing pH. This observed intensity e†ect
may, in part, be due to a decrease in the turbidity in the bac-
terial suspensions as the pH increases. The clarity of the bac-
terial samples studied was at an optimum for spectroscopic
analysis at pH 12. The dependency of the electronic absorp-
tion proÐle on pH was also found to be a reversible process
between the pH range 3È12. That is, if the pH of the bacterial
solutions were raised and then lowered the absorbance trends
remained the same with no permanent change in the absorp-
tion spectrum.
Emission and excitation spectra. Luminescence studies were
performed on K. pneumoniae samples loaded and unloaded
with cadmium sulÐde particles between pH 3È12 and over a
range of excitation and emission wavelengths. In all cases a
broad feature centred at 350È380 nm was observed: such a
band is a typical proÐle measured for many biological samples
and is associated with the luminescence arising from aromatic
amino acids in proteins, e.g. tyrosine and tryptophan.33 For
both the loaded and unloaded samples, no signiÐcant di†er-
ences were detected between the luminescence emission pro-
Fig. 2 (a) Luminescence emission spectra of CdS loaded and
unloaded K. pneumoniae cells at pH 7,
j \ 270 nm, slits
ex
(excitation : emission) 5.0 : 5.0 nm. (b) Luminescence emission spectra
of loaded and unloaded K. pneumoniae cells at pH 12, j \ 270 nm,
ex
slits (excitation : emission) 5.0 : 5.0 nm.
cence is poor but over a period of days subsequent corrosion
can increase the luminescence by orders of magnitude. No
change in the intensity of luminescence of the bio-CdS par-
Ðles produced either from excitation at a wavelength (j ) of
ex
400 nm, i.e. the absorption maximum (Fig. 1), at wavelengths
above 400 nm, or at pH values below 7. Exciting the loaded
ticles (j \ 270 nm) was noticed over a period of 3È4 weeks.
ex
and unloaded samples at wavelengths close to j
possibly
This result therefore suggests that the bio-CdS particles are
max
causes biological components present in both sample types to
luminesce. This “bio-luminescenceÏ observed in both sample
types presumably saturates any signals emanating from the
bio-CdS particles. Di†erences in the emission spectra between
the loaded and unloaded samples were only found to occur
upon excitation between the wavelength range 250È320 nm
and at a pH between 10È12. These results are illustrated in
Fig. 2(a) and (b).
protected from corrosion, possibly by a biological membrane.5
A comparison can be made between the CdS particles
formed on K. pneumoniae and the “peptide-cappedÏ intracellu-
lar particles formed by the yeasts C. glabrata and S. pombe
cultured in the presence of cadmium salts.11h18 The optical
properties of the CdS crystallites formed by these yeasts are
size dependent. In C. glabrata CdS crystallites 2 nm in diam-
eter, which are stabilised by glutathione (c-Glu-Cys-Gly) pep-
tides, have been reported by Dameron and Winge16 to
At a pH of 12 and a j \ 270 nm a luminescence emission
ex
wavelength (j ) at B450 nm was observed in the loaded
luminesce at a wavelength of 430 nm (j \ 355 nm). Acidi-
em
ex
samples, which is close to the onset of absorption of the CdS
Ðcation of the “CdS-yeastÏ containing system to pH 1.5 from a
bio-particles (j
B 450 nm, pH 12). This luminescence band
pH of 11È13 was shown to be accompanied by a reduction in
the luminescence intensity associated with the CdS particles.
Such a pH e†ect was also observed with the loaded samples
K. pneumoniae investigated in this study. Although there are
obvious similarities in the luminescence properties of the two
systems, there are also subtle di†erences in the electronic
absorption characteristics. Dameron and Winge16 have
reported that acidiÐcation of CdS particles from C. glabrata
onset
was absent in the unloaded bacterial samples. By analogy with
previous studies on conventional systems, such a feature can
be assigned to excitonic luminescence originating from the
CdS particles.31,34 Weak and broad emission bands, com-
monly observed in inorganic CdS dispersions between wave-
lengths 500 and 700 nm and attributable to the recombination
of charge carriers,35 were not observed in the loaded CdS
samples. Examination of the luminescence excitation spectra
for the loaded and unloaded samples revealed that there was
little measurable di†erence between them. Any excitation
peaks which could have originated from bio-CdS particles
were unresolvable against broad and intense bio-luminescence
bands.
from pH 5 to 4 resulted in an attenuation of j
(B300 nm)
max
from 335 nm (pH 5) to
and a concomitant red shift in j
onset
370 nm (pH 4). In contrast they also observed that acidiÐca-
tion of CdS particles from S. pombe to pH of 4 from a pH of
7È5 led to an attenuation of j (B300 nm) with minimal red-
max
shifting of j
. These observations were not seen in the K.
onset
For many inorganic CdS particles the intensity of lumines-
pneumoniae CdS system; no attenuation of j
occurred upon
max
J. Chem. Soc., Faraday T rans., 1998, V ol. 94
1237

View Online
acidiÐcation and there was a red-shift of j
under basic
[reactions (4)È(7)]:37
onset
rather than acid conditions.
k1
The K. pneumoniae system is more analogous to the inorga-
nic CdS systems between the pH range 3È10, where the shape
of the electronic absorption spectra is una†ected by the pH.36
It is difficult to postulate if the intensity of absorption of the
bio-CdS samples is changing with pH as interferences due to
sample turbidity at low pH values obscure the intensities
observed. At pH 12 however, unlike with inorganic CdS
excitation
CdS ] hv ÈÈÈ e~ ] h`
(4)
(5)
(6)
(7)
k~1
luminescence
MV2` reduction
back reaction
e~ ] h` ÈÈÈ CdS ] hv
1
k2
MV2` ] e~ ÈÈÈ MV~`
k~2
MV~` ] h` ÈÈÈ MV2`
ads
In addition high energy, electrons from the photolysis light
source can corrode the semiconductor producing cadmium
metal deposits.2
The MV~` radical ion production is in equilibrium with the
back reaction whereby it combines with a positive hole in the
valence band thus reforming the methyl viologen dication. A
net rate of MV~` accumulation will only be observed if the
systems, a hypsochromic shift in j
was seen. Analogous to
max
the yeast systems mentioned above and the bio-CdS particles
investigated in this work, Uchihara et al.36 have reported that
the intensity of the luminescence emission proÐle of both Cd-
and S-rich colloids of inorganic CdS increases as the pH
increases (particles [10 nm in diameter, j \ 355 nm, j
\
ex
em
515 nm). The surface characteristics which might give rise to
these observations are discussed in the next section.
rate coefficient sequence is in the order k [ k
and k [
1
~1
2
k
.
~2
The rate of reduction is dependent on two features:37 (i) the
Photoredox reactions
rate of production of the high energy electrons, which will be
constant under similar irradiating conditions; (ii) the extent of
binding between the particle surface and the redox species.
Using the above model it can be seen why there are di†er-
ences in photoactivity at high and low pH values for the bio-
particles. The positively-charged redox species, methyl
viologen, will bind to CdS via the surface sulÐde, S2~ sites,
which will be more prevalent at pH 10 than in acidic solu-
tions. The behaviour is directly related to the fact that fewer
surface HS~ sites will be present at low concentrations of
hydrogen ions. Furthermore, in acidic conditions hydrogen
sulÐde can be produced which e†ectively removes the S2~
binding sites.
Methyl viologen. The luminescence results described above
indicate that the photoactivity of the CdS bio-particles is max-
imised and observable under conditions of high pH. Photoly-
sis experiments using a medium-pressure mercury lamp were
therefore performed using loaded and unloaded samples of K.
pneumoniae at pH 10. Studying the photoredox reactions at
pH 10 rather than at a higher pH allows a direct comparison
with conventional inorganic CdS systems, also measured at
pH 10. Reduction of MV2` (100 lM) was monitored at three
wavelengths (598, 600 and 602 nm). The results, shown graphi-
cally as absorbanceÈtime plots in Fig. 3, provide evidence that
the unloaded samples are not photoactive and that for loaded
samples ([CdS] B 0.4 mM) the formation of MV~` occurs at
an initial rate of 4.8 ] 10~9 mol dm~3 s~1. This rate is com-
parable to that obtained using nanocrystalline CdS colloids
prepared by Matsumoto et al.22 who report a MV~` pro-
duction rate of ca. 1 ] 10~9 mol dm~3 s~1 by colloidal CdS
particles with an average diameter of 4.4 nm, ranging in size
from 3 to 8 nm, at pH 10 ([CdS] \ 0.06 mM, [MV2`] \ 1
mM, [EDTA] \ 1 mM, average quantum yield of MV~` pro-
Excess surface Cd2` sites will also be more prevalent at
lower pH. These not only repel MV2` from the semicon-
ductor surface but can also act as radiationless recombination
sites thereby using electrons otherwise available for reduction.
Under basic conditions Cd2` sites will be blocked because
they are transformed into cadmium hydroxide, Cd(OH) , as
2
reported by Spanhel et al.37
duction at 5 mW irradiation (U ~`) \ 1.62%).
Methyl orange. The photoreduction of MO~ by CdS par-
ticles would be expected to be promoted at di†erent surface
sites from MV2` because it is a negatively charged species.
Measurement of its rate of reduction and comparison with the
results obtained using MV2` provides a direct method of
investigating surface binding to the bio-particles.
Initially, the e†ect of pH on the rate of photoreduction of
methyl orange (20 lM) was monitored. It has previously been
reported that the reduction potential of methyl orange is pH
dependent and varies in the following way: E0 (MO~/
HMO~) \ [0.058 pH V vs. SCE.29,30,38 The dependency of
E0 on pH is due to the formation of an intermediate semi-
reduced radical HMO~~ which readily reacts to form the
hydrazine derivative as described in reactions (2) and (3).
In the present study, at both pH 4 and 10, no photoreduc-
tion was observed for unloaded K. pneumoniae samples. Mea-
surements of the UVÈVIS spectrum at these pH values
showed that photolysis of CdS loaded samples ([CdS] B 0.4
mM) produced the reduced protonated form of methyl orange,
MV
Analogous photolysis experiments were carried out under
acidic conditions at pH 4: no reduction was observed with
either the loaded or unloaded samples.
The di†erence between the high and low pH results can be
rationalised with a model for the bio-particles that closely
resembles conventional chemical systems. Hence the K. pneu-
moniae CdS particles provide binding sites for methyl viologen
that are readily accessible but which are passivated under
acidic conditions. In general terms, the mechanism of electron-
transfer in the CdS system can be summarised as follows
H MO~. The associated disappearance rates of MO~ can be
2
calculated from the absorbanceÈtime proÐles shown in Fig.
4(a) and (b): at pH 10 the rate coefficient was 4.6 ] 10~7 mol
dm~3 s~1 whereas at pH 4 the corresponding value was
1.4 ] 10~6 mol dm~3 s~1. Therefore a thirty-fold enhance-
ment factor for the photoreduction was measured in acidic
solutions. These rates are faster than those obtained by Mills
and Green28 who have reported the rate of methyl orange
disappearance for powdered CdS (mean particle diameter [ 1
lm) and colloidal CdS (mean diameter \ 8 nm). For pow-
dered CdS the rate of disappearance was reported as
7.7 ] 10~8 mol dm~3 s~1 at pH 4.4 ([CdS] \ 10 mM,
Fig. 3 Average absorbance values (j \ 598, 600 and 602 nm) moni-
tored over time, upon irradiation of loaded ([CdS] B 0.4 mM) and
unloaded samples of K. pneumoniae containing MV2`
([MV2`] \ 100 lM) at pH 10
1238
J. Chem. Soc., Faraday T rans., 1998, V ol. 94

View Online
tion of the bio-particle surface should also be considered.
At acidic pH values, S2~ surface sites will be transformed
into surface HS~ ions and possibly H S, which may be rel-
2
eased into solution. In this situation the approach of MO~ to
the bio-particle surface will be less restricted by electrostatic
repulsion, owing to the reduced negative charge. Furthermore
the Cd2` surface sites on the CdS particle surface will not be
capped by hydroxide groups leading to an increased attrac-
tion for the MO~, and subsequent binding. Electron-transfer
and photoreduction should increase under these conditions.
Cd2` radiationless recombination sites will, of course, exist
and have a deleterious e†ect on reduction rate. However, from
the results obtained it appears that the e†ect is small in con-
trast to the increase in rate caused by the surface attraction.
The results obtained for the bio-particle system are in
agreement with those reported for chemically synthesised
semiconductor systems. Peral and Mills,29 using CdS colloids,
and Zang et al.30 investigating ZnS sols, observed enhanced
rates of reduction of methyl orange at pH 4 compared to pH
10.
Methyl orange and methyl viologen composite solutions. It
has been reported that the MV~` radical ion can re-form the
parent MV2` ion by passing an electron to another acceptor:
this process is termed “electron relayÏ.29,30 For example, in
mixed systems of MV2` and MO~ with colloidal CdS an
unexpected enhancement of MO~ reduction has been
explained in terms of this e†ect. Therefore as a Ðnal probe of
the bioparticlesÏ photoredox characteristics, the competitive
interaction between MV2` and MO~ at pH 10 at the bac-
terial surface was investigated.
CdS loaded and unloaded systems containing MO~ (20
lM), MV2` (100 lM) were prepared. UVÈVIS absorption
spectra were measured as a function of photolysis time. The
absorbance at two wavelengths was monitored: one at 464
nm, which provides a measure of the decrease in MO~ con-
centration and the other at 600 nm for quantiÐcation of MV~`
radical ion production. The addition order of the two reac-
tants to the CdS bio-particles was found, as expected, to have
no e†ect because as determined previously the redox species
bind to di†erent surface sites. The kinetic results obtained are
shown in Fig. 5. It can be seen that, initially, the methyl
orange was reduced, as indicated by the sharp decrease in the
absorbance monitored at j \ 464 nm, whilst the absorbance
at j \ 600 nm remained constant. After ca. 2000 s the decreas-
ing absorbance at j \ 464 nm ceased, the absorbance at
j \ 600 nm increased rapidly and the solution turned blue
Fig. 4 (a) Average absorbance values (j \ 462, 464 and 466 nm)
monitored over time, upon irradiation of loaded ([CdS] B 0.4 mM)
and unloaded samples of K. pneumoniae containing MO~
([MO~] \ 20 lM) at pH 10. (b) Average absorbance values (j \ 462,
464 and 466 nm) monitored over time, upon irradiation of loaded
([CdS] B 0.4 mM) and unloaded samples of K. pneumoniae containing
MO~ ([MO~] \ 0.02 mM) at pH 4.
[MO~] \ 40 lM, [EDTA] \ 50 mM, N purged solution). As
2
expected the rate of photoreduction of methyl orange by col-
loidal CdS particles was faster than with the powdered CdS,
i.e. 2.7 ] 10~7 mol dm~3 s~1 ([CdS] \ 0.5 mM, [MO~] \ 40
lM, [EDTA] \ 50 mM, N purged solution).
2
The fact that photoreduction of methyl orange occurs is a
clear indication that the CdS bio-particles possess photoac-
tive, positively charged surface sites (presumably Cd2`), which
promote MO~ binding and subsequent electron transfer. This
model is conÐrmed by consideration of the pH e†ects.
Thus at pH 10 the negatively charged S2~ surface sites
would promote repulsion of the MO~ from the bio-particle
surface. Furthermore the Cd2` sites will be converted into
cadmium hydroxide, which has two mechanistic consequences.
The Ðrst is that incoming MO~ ions will not be attracted to
these potential redox centres. Secondly, the formation of
surface Cd(OH) will reduce the number of Cd2` sites, which
2
could promote radiationless recombination of free electrons.
This e†ect will leave more electrons available for reduction.
Overall it should therefore be expected that MO~ photore-
duction by CdS particles is possible but relatively inefficient at
high pH.
From consideration of the pH dependence for the reduction
potential of methyl orange discussed above, it can be calcu-
lated that at pH 4, E0 (MO~/HMO~) \ [0.22 eV vs. SCE
compared to [0.58 eV vs. SCE at pH 10.29,30 Although the
dependence of the E0 value for MO~ on pH is a source of
much debate,29,30 it does predict that its reduction would be
more favourable at the lower pH. However at both pH 10 and
4, the conduction band potential of the semiconductor is con-
siderably more reducing than the reduction potential of
MO~.30 Therefore the e†ect of pH on the chemical composi-
Fig. 5 Absorbance values (j \ 464 and 600 nm) monitored over
time, upon irradiation of loaded samples ([CdS] B 0.4 mM) of K.
pneumoniae containing MO~ and MV2` ([MO~] \ 20 lM,
[MV2`] \ 100 lM) at pH 10
J. Chem. Soc., Faraday T rans., 1998, V ol. 94
1239

View Online
indicating the formation of MV~`. Furthermore, after ca. 2500
s the absorbance at j \ 464 nm, i.e. MO~, was then observed
to increase.
No change in either the MV2` or MO~ signal was
observed upon photolysis of unloaded samples.
A simple inspection of the individual rate coefficients mea-
sured at pH 10 for methyl viologen and methyl orange in this
study, i.e. 4.8 ] 10~9 and 4.6 ] 10~7 mol dm~3 s~1, respec-
tively ([CdS] B 0.4 mM, [MV2`] \ 100 lM, [MO~] \ 20 lM),
suggests that MO~ should be reduced preferentially, as com-
pared to MV2`, in the CdS bio-particle system. Indeed these
are the qualitative observations found in the competitive
experiments. However the mechanism of interaction between
the two species are, as discussed above, quite di†erent and
only analysis of the two-component system can provide a
quantitative picture of the chemistry.
Scheme 1 Possible electron-transfer scheme for the generation of the
radical species MV~` and HMO~~ (and subsequent H MO~
2
formation) in the presence of bacterially generated CdS. The methyl
orange anionic radical can be formed via two pathways; (a) the direct
transfer of an electron from the conduction band of CdS (k ) or (b) the
d
indirect transfer of an electron from CdS with MV2` acting as an
“electron-relayÏ (k ] k ).
in
er
As seen in Fig. 5 the methyl orange component was
observed to be reduced before the methyl viologen with a rate
coefficient, 1.0 ] 10~6 mol dm~3 s~1 ([CdS] B 0.4 mM),
which is some Ðfty times faster than in its one-component
counterpart. No reduction of MV2` occurred over the same
period of time. Such an observation is, at Ðrst sight, surprising
because MV2` was clearly shown to be reduced at pH 10 in
its one-component system. This enhancement in reduction
rate in the presence of methyl viologen could be due to the
MV2` acting as an electron relay as reported for chemically
synthesised semiconductor particles in composite solutions by
Peral and Mills29 and Zang et al.,30 although it may be
because of the e†ect of MV2` on surface charge. At over-
potentials [200 mV, which both MO~ and MV2` possess,
the reduction process is reported to be a di†usion-limited elec-
trochemical reaction.22 This means that surface characteristics
are of great importance, and the reaction could well proceed
at a faster rate in the presence of methyl viologen owing to the
MV2` binding to S2~ rich surface sites, and producing a posi-
tive electrostatic Ðeld around the semiconductor particle.30
This would allow the MO~ to bind, or come into the close
proximity of a greater number of semiconductor particles, and
hence cause an increase in electron-transfer and reduction
rate.
At pH 10, methyl orange has a reduction potential of ca.
[0.58 eV vs. SCE, compared to a non-pH dependent
reduction potential of [0.69 eV for methyl viologen. Under
ideal conditions, these data imply that methyl orange would
be preferentially photoreduced to methyl viologen with CdS
particles. It was not until the methyl orange appeared fully
reduced after 2000 s that a blue coloration was observed in
the system, and the methyl viologen became reduced at a rate
of 1.0 ] 10~8 mol dm~3 s~1 ([CdS] B 0.4 mM), faster than its
reduction rate on its own (4.8 ] 10~9 mol dm~3 s~1). The
explanation for the increased rate of reduction of methyl viol-
ogen in the system containing methyl orange may again be
due to the presence of the oppositely charged species making
it more electrostatically favourable for binding to the semicon-
ductor surface.
Conclusions
“Bio-CdSÏ particles synthesised by the bacterium K. pneumon-
iae demonstrate both optical and photoactive traits analogous
to chemically synthesised inorganic CdS systems. Optical
similarities between the biological and chemical CdS systems
are reÑected in the luminescence emission spectra obtained for
the bio-semiconductor particles; the wavelength of maximum
intensity (j ) was observed to be close to the onset of absorp-
em
tion (j
B 450 nm) and is possibly excitonic in nature. In
onset
colloidal Q-CdS particle systems, excitonic luminescence is
frequently noticed upon particle “activationÏ, achieved under
conditions of high pH and in the presence of excess cadmium
ions.31
The photochemical properties associated with the bio-
semiconductor particles establish unequivocally that they
have the potential to drive photoinduced reactions presently
being undertaken with colloidal CdS dispersions. Both posi-
tively charged, i.e. MV2`, and negatively charged species, i.e.
MO~, were reduced by the bio-CdS particles at rates compa-
rable to inorganic CdS “Q-particlesÏ; the rate of reduction of
MV2` by bacterial CdS was similar to that reported by Mat-
sumoto et al.22 for CdS Q-particles with an average diameter
of 4.4 nm. This result reÑects the non-stoichiometry of the
bio-CdS particles under di†erent pH conditions, with S2~ rich
sites allowing binding of positively charged species at high pH
values and Cd2` rich surface sites promoting binding of nega-
tively charged species to the semiconductor surface, encour-
aging electron transfer and hence photoreduction at low pH
values. The comparable reduction rate of MV2` by the
bio-CdS and inorganic Q-CdS particles also supports the
theory that the large “superparticlesÏ observed on the bacterial
cell wall ca. 200 nm in diameter are composed of amorphous
discrete quantum dots of CdS approximately 4 nm in diam-
eter.5,19
Finally upon examination of the competitive reaction
between MV2` and MO~ the rates of reduction were dis-
covered to be dependent on the surface characteristics of the
bio-particles, as seen with their inorganic counterparts. These
results coincide with, and give some credence to, the sugges-
tion that in inorganic CdS systems MV2` can act as an
“electron-relayÏ transferring electrons to surrounding MO~
molecules.
The reason for the preferential reduction of methyl orange
over methyl viologen would seem to be simply due to the less
negative reduction potential of methyl orange, compared to
methyl viologen, at pH 10.
Once the methyl orange becomes fully reduced after 2000 s,
the methyl viologen is reduced as indicated by the increase in
absorbance observed at 600 nm. The accompanying increase
in absorbance at 464 nm could be due to overlap from the
absorption band at 600 nm produced by the reduced MV2`.
Additionally, the increase in absorption at 464 nm may be due
to the re-formation of the unprotonated MO~, which absorbs
The authors are grateful to the SERC Biotechnology
Directorate/Clean Technology Unit for the award of a
research assistantship for P.R.S. and to the Wellcome Trust
for the award of a toxicology studentship for J.D.H. Addi-
tionally, we would like to thank Professor A.J. Thomson for
the use of the UVÈVIS spectrometer used in this work.
at 464 nm, as the reduced H MO~ species becomes oxidised
2
References
perhaps by substrates in the media. Scheme 1 illustrates the
reaction pathways that may be occurring in the illuminated
1
J. D. Holmes, P. R. Smith, R. Evans-Gowing, D. J. Richardson,
system containing MV2` and MO~.
D. A. Russell and J. R. Sodeau, Arch. Microbiol., 1995, 163, 143.
1240
J. Chem. Soc., Faraday T rans., 1998, V ol. 94

View Online
2
J. D. Holmes, P. R. Smith, R. Evans-Gowing, D. J. Richardson,
D. A. Russell and J. R. Sodeau, Photochem. Photobiol., 1995, 62,
1022.
H. Aiking, K. Kok, H. van Heerikhuizen and J. vanÏt Riet, Appl.
Environ. Microbiol., 1982, 44, 938.
H. Aiking, A. Stijnman, C. van Garderen, H. van Heerikhuizen
and J. vanÏt Riet, Appl. Environ. Microbiol., 1984, 47, 374.
J. D. Holmes, D. J. Richardson, S. Saed, R. Evans-Gowing, D. A.
Russell and J. R. Sodeau, Microbiology, 1997, 143, 2521.
G. Z. Jaeckel, T ech. Phys., 1926, 7, 301.
J. K. Inman, A. M. Mraz and W. A. Weyl, Solid L uminescent
Materials, Wiley, New York, 1st edn., 1948, p. 182.
R. Rossetti, S. Nakahara and L. E. Brus, J. Chem. Phys., 1983, 79,
1086.
21 S. Ogawa, F. F. Fan and A. J. Bard, J. Phys. Chem., 1995, 99,
11182.
22 H. Matsumoto, H. Uchida, T. Matsunaga, K. Tanaka, T. Sakata,
H. Mori and H. Yoneyama, J. Phys. Chem., 1994, 98, 11549.
23 Y. Nosaka and M. A. Fox, L angmuir, 1987, 3, 1147.
24 K. Komers, J. Chem. Res. (S), 1994, 293.
25 J. W. Park, J. H. Kim, B. K. Hwang and K. Park, Chem. L ett.,
1994, 2075.
3
4
5
26 G. T. Brown and J. R. Darwent, J. Phys. Chem., 1984, 88, 4955.
27 A. Mills and G. Williams, J. Chem. Soc., Faraday T rans. 1, 1987,
83, 2647.
6
7
28 A. Mills and A. Green, J. Photochem. Photobiol. A: Chem., 1991,
59, 199.
8
9
29 J. Peral and A. Mills, J. Photochem. Photobiol. A: Chem., 1993,
73, 47.
R. Rossetti, J. L. Ellison, J. M. Gibson and L. E. Brus, J. Chem.
Phys., 1984, 80, 4464.
30 L. Zang, C. Y. Liu and X. M. Ren, J. Photochem. Photobiol. A:
Chem., 1995, 85, 239.
10 R. N. Reese, R. K. Mehra, E. B. Tarbet and D. R. Winge, J. Biol.
Chem., 1988, 263, 4186.
11 R. N. Reese and D. R. Winge, J. Biol. Chem., 1988, 263, 12832.
12 C. T. Dameron, B. R. Smith and D. R. Winge, J. Biol. Chem.,
1989, 264, 17355.
31 H. Weller, Angew. Chem., Int. Ed. Engl., 1993, 32, 41.
32 W. A. Pickett and A. C. R. Dean, Microbios, 1976, 15, 79.
33 B. Kierdaszuk, I. Gryczynski, A. Modrak-Wojcik, A. Bzowska,
D. Shugar and J. R. Lakowicz, Photochem. Photobiol., 1995, 61,
319.
13 R. K. Mehra, P. Mulchindani and T. C. Hunter, Biochem.
Biophys. Res. Commun., 1994, 200, 1193.
14 C. T. Dameron, R. N. Reese, R. K. Mehra, A. R. Kortan, P. J.
Carroll, M. L. Steigerwald, L. E. Brus and D. R. Winge, Nature
(L ondon), 1989, 338, 596.
15 J. Barbas, V. Santhanagopalan, M. Blaszczynski, W. R. Ellis Jr.
and D. R. Winge, J. Inorg. Biochem., 1992, 48, 95.
16 C. T. Dameron and D. R. Winge, Inorg. Chem., 1990, 29, 1343.
17 E. Grill, E. L. Winnacker and M. H. Zenk, FEBS, 1986, 197, 115.
18 R. K. Mehra, E. B. Tarbet, W. R. Gray and D. R. Winge, Proc.
Natl. Acad. Sci. USA, 1988, 85, 8815.
19 J. D. Holmes, J. A. Farrar, D. J. Richardson, D. A. Russell and J.
R. Sodeau, Photochem. Photobiol., 1997, 65, 811.
20 A. Henglein, J. Chim. Phys., 1987, 84, 1043.
34 A. J. Ho†man, G. Mills, H. Yee and M. R. Ho†mann, J. Phys.
Chem., 1992, 96, 5546.
35 T. Ucihara, M. Matsumara, J. Ono and H. Tsubomura, J. Phys.
Chem., 1990, 94, 415.
36 T. Uchihara, T. Nakamura and A. Kinjo, J. Photochem. Photo-
biol. A: Chem., 1992, 66, 379.
37 L. Spanhel, M. Haase, H. Weller and A. Henglein, J. Am. Chem.
Soc., 1987, 109, 5649.
38 A. Mills and G. Williams, J. Chem. Soc., Faraday T rans. 1, 1989,
85, 503.
Paper 7/08742J; Received 4th December, 1997
J. Chem. Soc., Faraday T rans., 1998, V ol. 94
1241