Doped Framework Iron Hydroxyl Phosphate as Photocatalyst for Hydrogen Production from Water/Methanol Mixtures

Article abstract of DOI:10.1002/ejic.201500629

In the search for novel photocatalysts for hydrogen production and with the α-Fe₂O₃ photoelectrocatalyst as a recent precedent, we report herein the preparation, semiconductor properties and photocatalytic activity of metal-doped (0.1-5 wt.-% loading) iron hydroxyl phosphate (FeP). X-ray diffraction analyses of FeP samples subjected to extended photocatalytic irradiation showed the stability of this framework phosphate under photocatalytic conditions. Doping increased the photocatalytic efficiency of FeP for all dopants, with the optimal doping level between 0.1 and 1%. Under the optimized conditions (Cr at 1% doping), the photocatalytic activity of FeP reached a hydrogen production rate of 35.82 μmolg_Fe^-1 in the absence of platinum as co-catalyst. The conduction flat band potential was estimated by photocurrent measurements or impedance spectroscopy to be 0.1 eV versus NHE and the charge carrier density 2.6×10²⁰ carrierscm^-3. Transient absorption spectroscopy revealed a transient species decaying on the microsecond timescale characterized by a broad band spanning 300-750 nm. This transient was attributed to the charge-separated state. These results are promising for the development of novel photocatalytic materials based on framework metal phosphate.

Full text of DOI:10.1002/ejic.201500629

FULL PAPER
DOI:10.1002/ejic.201500629
Doped Framework Iron Hydroxyl Phosphate as
Photocatalyst for Hydrogen Production from Water/
Methanol Mixtures
[
a]
[a]
[a]
Marco Serra, Herme G. Baldovi, Mercedes Alvaro, and
Hermenegildo Garcia*[
a]
Keywords: Photocatalysis / Semiconductors / Doping / Hydrogen / Iron
–
1
In the search for novel photocatalysts for hydrogen pro- 35.82 μmolgFe in the absence of platinum as co-catalyst.
duction and with the α-Fe
precedent, we report herein the preparation, semiconductor
properties and photocatalytic activity of metal-doped (0.1–
2
O
3
photoelectrocatalyst as a recent
The conduction flat band potential was estimated by photo-
current measurements or impedance spectroscopy to be
0.1 eV versus NHE and the charge carrier density
2.6ϫ1020 carrierscm . Transient absorption spectroscopy re-
vealed a transient species decaying on the microsecond time-
scale characterized by a broad band spanning 300–750 nm.
–3
5
wt.-% loading) iron hydroxyl phosphate (FeP). X-ray dif-
fraction analyses of FeP samples subjected to extended pho-
tocatalytic irradiation showed the stability of this framework
phosphate under photocatalytic conditions. Doping in- This transient was attributed to the charge-separated state.
creased the photocatalytic efficiency of FeP for all dopants,
with the optimal doping level between 0.1 and 1%. Under
the optimized conditions (Cr at 1% doping), the photocata-
lytic activity of FeP reached a hydrogen production rate of
These results are promising for the development of novel
photocatalytic materials based on framework metal phos-
phate.
Introduction
under UV light irradiation, including availability, lack of
toxicity and high photocatalytic efficiency; however, TiO2
presents severe limitations that have proven difficult to
overcome.
Photocatalysis is fast gaining importance because, in ad-
dition to the degradation of pollutants in the gas or liquid
phase, it could also be an alternative for solar light conver-
The most important limitations of TiO as photocatalyst
2
[
1–5]
sion.
Owing to the shortage of fossil fuels, considerable
are its wide band gap and the large degree of charge recom-
effort is being focused on the development of renewable en-
ergies. Sunlight reaching the Earth’s surface is one of the
bination. For these reasons, the performance of TiO as
2
photocatalyst under sunlight illumination is far from opti-
mal and there is continuing interest in developing new pho-
tocatalysts. Also of relevance for the present work is the
primary energy resources and could provide enough energy
to cover future energy demand.[
6–9]
However, due to the low
power density of solar light, the diurnal cycle and the de-
pendence of weather, solar light as a renewable energy has
to be combined with energy vectors that accumulate and
concentrate the solar energy. One possible strategy is the
generation of hydrogen from water or the production of
discovery of the photoelectrochemical activity of α-Fe O3
2
[23–28]
powders for the generation of hydrogen from water.
Two of the main advantages of iron oxide are its lack of
toxicity and large abundance. However, the low reduction
potential of its conduction band electrons (–0.4 eV vs.
[
7,10–16]
other solar fuels based on solar light photocatalysis.
In photocatalysis, the energy of light is converted by a
[29]
NHE ) is a major drawback in photocatalysis, making the
reduction of water to H2 thermodynamically uphill. To
solid into chemical energy.[
17–22]
In typical photocatalysts,
overcome this barrier it is necessary to submit the α-Fe O3
2
the absorption of a photon leads to the generation of a
charge-separated state. The most widely used photocatalyst,
also for the production of hydrogen and solar fuels, is tita-
photoanode to a bias potential of about 1 V. In addition,
α-Fe O tends to undergo photocorrosion, leading to the
2
3
3+
dissolution of Fe during the photocatalytic reaction. In
this context it is important to evaluate other water-insoluble
iron compounds that could act as photocatalysts.
[
17,19,21]
nium dioxide.
This solid presents several advantages
[
a] Instittuto Universitario de Tecnologia Quimica, Universidad
Considering the interest in developing new photocata-
lysts, we report herein the photocatalytic activity of open-
Politécnica de Valencia,
Avenida de los Naranjos S/N, 46022 Valencia, Spain
E-mail: hgarcia@qim.upv.es
framework iron hydroxyl phosphate (FeP) for H genera-
2
www.upv.es/herme
tion from CH OH/H O that is comparable to that of com-
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201500629.
3
2
mercial TiO2 P25 under simulated sunlight irradiation.
Eur. J. Inorg. Chem. 2015, 4237–4243
4237
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Open-framework phosphates are materials that in contrast Table 1. Code, composition and photocatalytic data for the FeP
samples under study.
to bulk phosphates only contain metal–O–P linkages leav-
ing interstitial channels in which a protonated amine re-
mains entrapped compensating the net negative charge of
the lattice. As will be shown below, compared with TiO2,
Code[a]
production[b]
[μmol]
[b]
Metal [wt.-%]
H
2
2
H producton
dopant
Fe
[μmolgFe–1
]
FeP [c]
100
0.36Ϯ0.05
0.54Ϯ0.07
0.83Ϯ0.06
0.87Ϯ0.06
0.64
15Ϯ2
22Ϯ4
34Ϯ5
36Ϯ5
26.
FeP has the additional advantage of being easy to dope. We (0.25Cr) FeP
Cr:0.28
Cr:0.48
Cr:90
Cr:2.30
Cr:4.0
99.7
99.5
99.1
97.7
96.0
[
c]
(
(
(
(
0.5Cr) FeP
have found that the photocatalytic activity of FeP depends
on the nature and percentage of the dopant. Its semicon-
ductor properties were characterized by measuring the con-
[
c]
1Cr)FeP
2Cr) FeP
5Cr) FeP
0.63
0.51
26
21
duction band potential, band gap, density of charge carriers (0.25Mn) FeP Mn:0.20 99.8
and, by transient absorption spectroscopy, charge-separated (0.5Mn) FeP
Mn:0.35 99.65
Mn:1.0 99.0
Mn:1.20 98.8
Mn:3.3 96.7
Co:0.13 99.87
Co:0.27 99.73
Co:0.42 99.58
Co:0.80 99.2
0.46
0.42
0.37
0.36
0.53
0.54
0.6
0.55
19
17
15
15
22
22
25
23
(1Mn) FeP
(2Mn) FeP
(5Mn) FeP
(0.25Co) FeP
state.
Because phosphates can be prepared by hydrothermal
crystallization in the presence of a large variety of metals
in a range of concentrations resulting in doping, our find- (0.5Co) FeP
(
(
(
1Co) FeP
2Co) FeP
5Co)FeP
ings may lead to new openings in the search for novel ef-
ficient photocatalysts. There is only one report in the litera-
Co:2.1
97.9
0.38
16
ture on the use of Fe(OH)PO as a photocatalyst for the
4
[a] The empirical formula of the FeP photocatalyst is (C
2 2
N H10)-
degradation of Methylene Blue under visible-light irradia-
Fe(1–x)
M
x
(HPO (OH) H O with M = Cr, Mn or Co. [b] H pro-
4
)
2
2
2
[
30]
tion,
but until now the photocatalytic activity for
duction after 6 h irradiation with a solar simulator. [c] Average of
four independent measurements.
hydrogen generation remains unexplored. In addition, we
also show the ability of metal doping to increase the effi-
ciency of FeP as a photocatalyst.
Results and Discussion
In the initial part of our study, open-framework FeP was
synthesized by hydrothermal crystallization of an aqueous
solution of iron chloride, ethylenediamine as base and tem-
plating agent, and phosphoric acid as phosphate source at
50 °C following a reported procedure.[ In addition to
31]
1
FeP containing a single metal, a wide range of FePs con-
taining 0.25–5 wt.-% Cr, Mn or Co were also synthesized.
The possibility of doping FeP has already been considered
in the literature. Specifically, we prepared three families of
materials based on FeP containing three different dopant
elements; their main analytical and characterization data
are summarized in Table 1.
The samples based on FeP exhibit X-ray diffraction pat- Figure 1. (A) Building blocks of FeP viewed along the b axis. Red
atoms correspond to iron, green to phosphorus and blue to oxygen.
terns typical of those previously reported for one-dimen-
[
31]
Hydrogen atoms, ethylenediamine and water molecules have been
omitted for clarity. The model of the Fe–O–P(O)–O chains was
sional, open-framework iron hydroxyl phosphate.
Fig-
ure 1 shows selected XRD patterns for some of the FeP
[31,32]
obtained by using the Mercury program for the structure file.
samples as well as the structure of these materials as re- (B) XRD patterns of metal-doped samples: (a) FeP, (b) (0.25 Cr)
ported in the literature.[
31,32]
FeP, (c) (0.5 Cr) FeP, (d) (1 Cr) FeP, (e) (2 Cr) FeP and (f) (5 Cr)
FeP. For the XRD patterns of the other samples, see the Support-
ing Information.
Noteworthy is the fact that the crystallinity of FeP is
maintained with some minor variations in the peak inten-
sity in the presence of the dopant elements. The almost co-
incident cationic radii of Fe, Cr, Mn and Co, as well as their interstitial space between the chains of octahedra, forming
similar charge densities, allows their incorporation in a few interactions with the octahedra through strong electrostatic
percent into the FeP structure without noticeable variation forces and hydrogen bonds.
of the XRD pattern. The crystal structure of FeP consists
The morphologies of the FeP particles were determined
of chains of octahedra with one Fe3+ ion at the centre and by microscopy; the large size of some of the particles in this
two vertices in trans positions sharing OH groups. Phos- case allowed the use of optical microscopy. Figure 2 shows
5
+
phate anions define tetrahedra around P ions connected representative images of FeP; the material is formed from a
by two oxygen atoms to two different octahedra with one broad distribution of crystals with a long aspect ratio. Some
O atom and an OH group at the other two vertices. Ethyl- of the particles are as long as 100 μm, and others have a
enediamine, present in the diprotonated form, occupies the length of about 1 μm. These sub-millimetric crystals are
Eur. J. Inorg. Chem. 2015, 4237–4243
4238
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
formed by the aggregation of much smaller sheets. Scanning
electron microscopy revealed in better detail that the crys-
tals are in reality aggregates of a multitude of flakes with a
thickness of about 0.5 μm and a length typically greater
than 20 μm (see also Figure 2, bottom).
Figure 3. Diffuse reflectance UV/Vis absorption spectra (plotted as
the Kubelka–Munk function of the remittance, R) of metal-doped
FeP: (a) FeP, (b) (0.25 Cr) FeP, (c) (0.5 Cr) FeP, (d) (1 Cr) FeP,
(
e) (2 Cr) FeP and (f) (5 Cr) FeP. For the UV/Vis spectra of the
other doped FeP samples, see the Supporting Information.
All the FeP samples under study exhibited photocatalytic
activity for H generation under simulated solar light irradi-
2
ation in H O/CH OH. No hydrogen evolution was ob-
2
3
served for FeP in the absence of CH OH under the same
3
conditions. Figure 4 shows the evolution of hydrogen for
the undoped FeP sample. The sample exhibits an induction
period of about 1 h in which little H is produced.
2
Figure 4. Photocatalytic hydrogen production from water/methanol
mixtures using FeP as photocatalyst under simulated sunlight irra-
diation. Vertical bars represent the relative error of the measure-
–
1
ments. Reaction conditions: water/methanol (1:1, 25 mL), 1 gL
catalyst concentration.
This induction period could be related to possible modi-
fications of the oxidation state of Fe3 in FeP or to the
adsorption of hydrogen on the solid before being detectable
+
Figure 2. Top: SEM image recorded for FeP. Middle: EDS elemen- in the gas phase. However, the H adsorption on FeP at
2
tal analyses of the crystal showing the presence of Fe and P (purple:
Au; red: Fe; green: P; blue: O; grey: C). Bottom: Optical micro-
scopy image at a different magnification showing the large crystals
present in FeP.
room temperature measured independently was negligible.
It is important to mention that H is produced in the pres-
ence of a photocatalyst that does not contain any noble
metal such as Pt as co-catalyst.
2
Irradiation with monochromatic light at 300
–2
–2
Of relevance with regard to the photocatalytic activity of (0.09 mWm ) and 420 nm (5.8 mWm ) led to the genera-
the FeP samples are the optical absorption spectra, which tion of 0.54 and 0.58 μmol of hydrogen in 6 h. The fact that
are very similar regardless of the nature of the dopant and a higher light power at 420 nm results in a similar genera-
its percentage in the range 0.25–5 wt.-% (Figure 3). All the tion of hydrogen as achieved with irradiation at 300 nm is
samples exhibit two minor absorption maxima at 380 and probably a reflection of the much higher absorption at
4
20 nm, which can be attributed to the d–d transition of 300 nm of FeP (Figure 3).
3
+
Fe ions. These peaks are accompanied by an intense ab-
It should be commented that 2.4 μmol of H were pro-
2
sorption band at 280 nm, attributable to the ligand-to-metal duced with P25 TiO under the same conditions as those
2
–1
O–Fe charge-transfer band.
shown in Figure 4 at 6 h (96 μmolg ), which is about six-
Eur. J. Inorg. Chem. 2015, 4237–4243
4239
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
fold higher than that measured for FeP. It should, however, Our proposal to explain the influence of doping and the
be noted that this comparison is based on the same photo- more efficient hydrogen generation by doped semi-
catalyst weight. More accurate comparisons should be conductors is illustrated in Scheme 1. The fact that there is
based on quantum efficiencies that reflect intrinsic photo- an optimal doping level suggests that the positive effect of
catalytic activity.
the dopant increasing the efficiency of charge separation by
Concerning stability, the chemical analysis of the solu- providing interband gap trapping sites is offset by an in-
tion after 6 h of irradiation and removal of the solid catalyst crease in charge recombination efficiency caused at higher
[
21]
showed an iron content below 10 ppb, which indicates that dopant concentrations.
3
+
the possible amount of Fe leached from the photocatalyst
3
+
has to be below 0.005% of the initial Fe content.
3
+
This analytical data showing negligible Fe leaching to-
gether with the lack of apparent change in the XRD pat-
terns of the FeP samples after prolonged irradiation indi-
cate the stability of FeP as photocatalyst. The results shown
in Figure 4 have to be considered in the context of the lack
of hydrogen production in the absence of a bias potential
[
26]
of α-Fe O
and the low photostability of this oxide re- Scheme 1. Possible explanation of the influence of metal doping in
2
3
3
+
FeP on charge separation and hydrogen production. The top and
bottom blue rectangles correspond to CB and VB, respectively. The
dashed line below the CB corresponds to the interband gap state
introduced by the dopant element. (a) Charge separation; (b) mi-
sulting in leaching Fe into the solution and dissolution of
[
28]
the solid.
The influence of doping on the photocatalytic activity
for hydrogen generation is presented in Figure 5. As can be gration of conduction band electrons to the dopant level; (c) H
+
seen, for the three sets of samples the activity towards reduction.
hydrogen production is optimal when the dopant loading is
between 0.1 and 1 wt.-%, lower or higher loadings resulting
All the above data refer to irradiation with simulated so-
lar light containing about 4% of the energy in the UV re-
gion and about 42% of the energy in the visible range. Ow-
ing to the lack of photocatalytic activity of titanium oxide
under irradiation with visible light, there is continuing
interest in the development of visible-light photocatalysts.
In the present case, and considering that according to the
absorption spectra the samples absorb light in the visible
region, we were interested in determining whether the
doped FeP samples exhibit some photocatalytic activity in
visible light. To address this issue we performed a set of
photocatalytic tests using a 450 nm LED emitting light of
wavelength above 420 nm (50 mWcm–2 at a distance of
in a decrease in the photocatalytic activity. Also, the nature
of the dopant element has an influence, chromium showing
the highest photocatalytic activity. According to the data
shown in Figure 5, the amount of hydrogen can be in-
creased by a factor 2.5 by incorporating chromium in a per-
centage of about 1%.
3
cm, see the emission spectra in the Supporting Infor-
mation). All the samples exhibited under these conditions
weak but measurable photocatalytic activity. The hydrogen
generation in the presence of Cr-doped FeP under visible
light is presented in Figure 6. The results obtained indicate
Figure 5. Hydrogen production from water/methanol at 6 h irradia-
tion using doped FeP as photocatalysts under simulated sunlight
irradiation. Reaction conditions: water/methanol mixtures (1:1,
–1
2
5 mL), catalyst concentration 1 gL . Concentration of dopant el-
ement increases for each dopant from left to right: 0, 0.25, 0.5, 1,
2
and 5%.
In analogy with TiO , we propose that in the present case
2
metal doping should operate by increasing the efficiency of
charge separation. It is known that upon light absorption
and the generation of electrons in the conduction band and
holes in the valence band, ultra-fast charge recombination
is the main deactivation pathway, limiting the efficiency of
the photocatalytic reactions.[
17]
Figure 6. Hydrogen production from water/methanol mixtures
using (Cr)FeP as photocatalysts and a visible-light LED as the irra-
diation source. Reaction conditions: water/methanol (1:1, 25 mL),
Doping can introduce into
the interband gap space additional states that can act as
traps of electrons or holes, promoting charge separation
–1
1
gL catalyst concentration, 24 h. Concentration of dopant ele-
and/or increasing the lifetime of the charge-separated state. ment increases from left to right: 0, 0.25, 0.5, 1, 2 and 5%.
Eur. J. Inorg. Chem. 2015, 4237–4243
4240
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
that under simulated sunlight a considerable proportion, signal from 2.5 to 20 μs. As we will see below, quenching
about 66%, of the photocatalytic activity is due to the UV experiments revealed that this transient corresponds to the
light even though this wavelength range corresponds to a fraction of the charge-separated state of FeP materials de-
minor percentage of the total solar energy.
caying on the microsecond timescale. In accord with this,
The conduction band flat potential for FeP was esti- the three kinetic steps could correspond to a fast charge
mated by measuring the photocurrent against the bias po- recombination (regime A) followed by charge migration
tential, determining the value that changes the photo- and relocation into traps (regime B), and a final recombina-
current from positive to negative. These measurements gave tion of the free charge carriers (regime C).
a flat band potential of 0.1 eV. In addition, impedance spec-
No influence of the dopant elements either on the spectra
troscopy of thin FeP films on transparent conductive elec- or on time profiles was observed. However, we noted that
trodes (see Figures S4 and S5 in the Supporting Infor- the intensity of the signal immediately after the laser pulse
mation) allowed us to estimate values of 0.1 eV and is dependent on the nature of the dopant. This observation
2
0
–3
2
.6ϫ10 carrierscm for the flat band potential and the suggests that the dopant element may disfavour the ul-
density of charge carriers, respectively. trafast charge recombination of charge carriers on the sub-
The charge-separation state and its quenching behaviour microsecond timescale rather than the relocation or recom-
as well as the influence of metal doping were monitored bination kinetics on longer timescales because the last two
by nanosecond transient absorption spectroscopy using a factors would have led to changes in the transient spectra
266 nm laser as the excitation source. In these experiments or the transient signals. According to the proposal pre-
FeP samples were suspended in acetonitrile by sonication. sented in Scheme 1, steps a and b should occur on times-
After removal of the solid residues by decantation, the per- cales shorter than microseconds and the trapping will be
sistent FeP colloidal solutions were submitted to nano- manifest on the microsecond timescale as an increase in the
second laser flash excitation. Regardless of the presence of intensity of the initial density of the charge separation.
dopant elements, all the transient spectra were charac-
The time profile of the signal is sensitive to the presence
terized in the microsecond timescale by a continuous ab- of quenchers (Figure 7). In the present study we used oxy-
sorption spanning the whole wavelength range. As an exam- gen as conduction-band electron quencher and methanol as
ple, Figure 7 shows the transient spectra for FeP observed valence-band hole quencher. In the case of FeP, the pres-
at 1.9, 5.8, and 10.8 μs after the laser flash. The time pro- ence of oxygen increases the intensity of the signal corre-
files of the signal at different wavelengths are coincident, sponding to regimes A and C (Figure 7) and methanol in-
which suggests that it corresponds to the same transient.
creases more significantly the signal in region C (Figure 7).
This effect of the quencher increasing the intensity of the
signal without significantly affecting the kinetics of the de-
cay can be interpreted by considering geminate charge re-
combination of electron and holes as the main decay path-
way of the signal. If one quencher is present partially re-
moving one of the charge carriers, then the transient signal
can increase due to the detection of the complementary
charge that has not collapsed by recombination and there-
fore reaches the nanosecond timescale in larger concentra-
tions. The influence of oxygen and methanol on the time
profile of the signal is in agreement with the assignment of
the transient spectra to the combined absorption of elec-
trons and holes. And the transient detected on the microsec-
ond timescale being a fraction of the total charge separation
events, specifically, the population of charge-separated
states that disappear on the sub-microsecond timescale. As
an example, Figure 7 shows the time profiles of the signal
2
Figure 7. Transient absorption spectra recorded for N -purged FeP
suspended in acetonitrile (OD = 0.3 at 355 nm) at 1.9 (᭿), 5.8 (᭹)
and 10.78 μs (᭡) after 355 nm laser excitation. Inset A: Time pro-
file of the transient spectrum monitored at 400 nm. The profile
shows three kinetic regimes that have been labelled A, B and C. monitored for FeP under argon- or oxygen-purging or in
Inset B: Time profile of the signal at 420 nm for the FeP suspension
the presence of methanol. Interestingly, the time profiles for
in acetonitrile (OD = 0.31 at 355 nm) after 355 nm laser excitation
doped FeP follow the same general trend observed for un-
(
the plots a, b and c correspond to the temporal profiles after N
2
and O purging or in presence of methanol, respectively).
2
doped FeP. The most salient feature of O and CH OH
2
3
quenching in the presence of doped FeP, particularly in re-
gime C, is the considerably smaller effect than that pre-
The time profile shows three kinetic regimes (see Fig- sented in Figure 7 for FeP. Regardless of the initial value of
ure 7). The first one is a fast decay, taking place in less than ΔJ/J , at short delay times the signal for doped FeP is
o
0.15 μs and corresponding to more than 60% of the top ΔJ/ clearly influenced by the presence of the quenchers, but the
Jo. This fast decay is followed by a growth from 0.15 to signal decay at longer times is almost insensitive to the pres-
about 2.0 μs, the ΔJ/J signal increasing by about 10%. The ence of O or CH OH. This could be due to the trapping
o
2
3
third kinetic regime corresponds to the final decay of the of charge carriers at certain sites.
Eur. J. Inorg. Chem. 2015, 4237–4243
4241
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
graph equipped with a MOLSieve column and a TC detector with
argon as thee carrier gas.
Conclusions
The results presented herein illustrate the potential of
widely available FeP as photocatalysts for hydrogen pro-
duction, combining the advantages of photostability, a thin films of FeP were deposited onto an FTO electrode (“So-
Photocurrent Measurements and Determination of the Flat Band Po-
tential of the Conduction Band: For photocurrent measurements,
2
higher reduction potential of the conduction band electrons laronix” 0.8ϫ10 cm ). Deposition was performed by spreading a
than α-Fe O and easy doping during the synthesis of the paste containing the semiconductor on the FTO surface by using
2
3
the doctor blade procedure. The surface of the FTO was defined
by adhesive tape. The paste containing the semiconductor was pre-
pared in advance to the deposition of the film onto the electrode
by dispersing FeP (100 mg), acetone (1 mL) and α-terpineol
materials. Our results show that doping with metal elements
can increase the photocatalytic activity of FeP by a factor
of 2.5.
The behaviour of FeP as a semiconductor has been sup-
ported by photocurrent measurements and impedance spec-
troscopy, which have allowed the positions of the conduc-
tion and valence bands to be established and an estimation
(
1 mL). The suspension was stirred for 24 h to achieve a homogen-
eous dispersion. Then the acetone and α-terpineol were allowed to
evaporate at room temperature to obtain the final dense semicon-
ducting paste. After deposition of the paste, the adhesive tape was
of the density of carriers. The charge-separated state was removed and the semiconductor film on FTO was heated at 80 °C
detected by transient absorption spectroscopy. The tran- for 24 h to increase the mechanical adherence to the layer. Photo-
sient absorption spectrum shows a continuous absorption current measurements were carried out by using a 150 W Xenon
arc lamp as light source and a PTI model 101 monochromator.
Electrical currents were measured with an Amel 7050 potentiostat
band spanning the whole wavelength range exhibiting three
kinetic regimes and decaying completely in 20 μs. Assign-
connected electrically to the FTO electrode with a platinum wire as
ment of the transient spectrum was based on the behaviour
counter electrode. The experiments were conducted inside a quartz
of the signal in the presence of oxygen and methanol, which
cuvette (20 mL) containing an aqueous 0.1 m KCl solution as elec-
are typical electron and hole quenchers, respectively. Dop-
trolyte. The system was purged with nitrogen for at least 15 min
ant elements influence the intensity of the transient signal,
before measurement. Flat conduction band potentials were deter-
which is more intense in the presence of dopants. In ad-
mined by measuring the photocurrent by applying an increasing
dition, the interaction of the dopant elements with quench- bias from 0.3 to –0.3 Vs. (vs. Ag/AgCl) every 0.1 V (seven points).
ers renders these materials less sensitive to quenchers. The photocurrent onset was detected at 365 nm and a scan rate of
Thus, the present findings may open the way to the fur- 2 min was used, namely 1 min without irradiation followed by
min illumination of the cell.
1
ther development of a novel family of semiconductors with
promising applications as photocatalysts based on the af- The number of charge carriers was determined by using the same
fordable elements iron and phosphorus.
system connected to a frequency response analyser. The measure-
ments were taken from 0.5 to –0.1 V every 0.1 V (seven points). At
each voltage the system was scanned in the frequency range from
105 from 0.1 Hz.
Experimental Section
Transient Absorption Spectroscopy: Suspensions of FeP or doped
FeP in acetonitrile were prepared by sonicating the powder for
10 min at 150 W. The amount of solid was set to obtain an optical
density at 266 of 0.35 units. The suspensions were persistent during
the time needed for the transient absorption spectra measurements
(about 3 h) without deposition of any solid. The suspensions in
Material Preparation: Open-framework iron hydroxyl phosphate
(
FeP) was prepared as reported by Song and co-workers.[31,32]
In brief, an aqueous solution of ferric chloride, phosphoric acid
and ethylenediamine (FeCl 0.005 mol/H PO 0.04 mol/
NH CH CH NH 2.5 mL/H O 35 mL) was placed in a 200 mL
3
3
4
2
2
2
2
2
Teflon-lined autoclave using a filling factor of 75%. The autoclave acetonitrile (3 mL) were placed in Suprasil quartz cuvettes of 1ϫ
2
was heated at 150 °C for 2.5 h and after this time the reaction mix-
ture was cooled to room temperature using an ice bath. The re-
sulting precipitate was filtered, washed and dried in vacuo at room
temperature. The doped materials were prepared following the
1 cm capped with septa. The suspensions were purged with argon
or oxygen (for quenching experiments) for at least 15 min before
the measurements. For the quenching experiments with methanol,
argon was bubbled through a bottle of methanol and passed
same synthetic protocol using a mixture of ferric chloride and chro- through the FeP suspensions for at least 50 min before measure-
mium, manganese, or cobalt chloride as dopant in proportions
ranging from 0.25–5 wt.-%.
ment. Transient measurements were carried out by using the fourth
–
1
harmonic (266 nm, 20 mJpulse ) of a Nd:YAG laser (7 ns fwhp)
as the excitation source. The transient signal was recorded in trans-
mission mode using a 150 W xenon lamp as the monitoring beam
through fibre optics to collect the transmitted light. The signal from
the monochromator/photomultiplier detection system was captured
by a Tektronix 2440 digitizer and transferred to a PC computer
that controlled the experiment and provided suitable processing
and data storage capabilities.
Photocatalytic Tests for Hydrogen Production: A suspension of the
–
1
catalyst (25 mL, 1 gL ) was sonicated for 10 min and placed in a
2
closed reactor with an irradiation window of 12.56 cm and tem-
perature and pressure controllers. The reactor was placed in a ther-
mostatic bath with the temperature set to 25 °C. The suspension
was purged with an argon flow of 2 psi for 15 min prior to irradia-
tion. The photoreaction was performed from the top of the photo-
reactor by using a visible-light LED or a solar simulator (Thermo
2
Oriel 1000 W) with an irradiation spot of 100 cm , placed at a dis- Acknowledgments
tance of 10 cm. The light of the solar simulator was filtered through
an Air Mass 1.5 filter and consists approximately 4% of UV light.
The amount of hydrogen collected in the headspace of the reactor
was analysed by injecting 100 μL of the gas into a gas chromato-
Financial support by the Spanish Ministry of Economy and Com-
petitiveness (MEC) (Severo Ochoa and CTQ20212-32315) and the
Generalidad Valenciana (Prometeo 2012/014) is gratefully acknowl-
Eur. J. Inorg. Chem. 2015, 4237–4243
4242
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
edged. M. S. thanks the Spanish Consejo Superior de Investiga-
ciones Científicas (CSIC) and Technical University of Valencia for
a postgraduate scholarship.
[17] D. W. Bahnemann, Res. Chem. Intermed. 2000, 26, 207.
[
[
18] M. A. Fox, M. T. Dulay, Chem. Rev. 1993, 93, 341.
19] A. Fujishima, X. Zhang, D. A. Tryk, Surf. Sci. Rep. 2008, 63,
15.
5
[
[
20] J. M. Herrmann, Catal. Today 1999, 53, 115.
21] A. L. Linsebigler, G. Q. Lu, J. T. Yates, Chem. Rev. 1995, 95,
[
[
[
1] Y. Amao, ChemCatChem 2011, 3, 458.
735.
2] G. Centi, S. Perathoner, ChemSusChem 2010, 3, 195.
3] D. Gust, T. A. Moore, A. L. Moore, Acc. Chem. Res. 2009, 42,
[
[
[
22] A. Mills, S. LeHunte, J. Photochem. Photobiol. A: Chem. 1997,
108, 1.
1
890.
4] L. Hammarström, S. Hammes-Schiffer, Acc. Chem. Res. 2009,
2, 1859.
23] N. Beermann, L. Vayssieres, S. E. Lindquist, A. Hagfeldt, J.
Electrochem. Soc. 2000, 147, 2456.
24] U. Bjorksten, J. Moser, M. Grätzel, Chem. Mater. 1994, 6, 858.
[
4
[
[
[
[
[
5] N. Serpone, D. Lawless, R. Terzian, Solar Energy 1992, 49, 221.
[25] Y.-S. Hu, A. Kleiman-Shwarshtein, A. J. Forman, D. Hazen,
J.-N. Park, E. W. McFarland, Chem. Mater. 2008, 20, 3803.
[26] A. Kay, I. Cesar, M. Graetzel, J. Am. Chem. Soc. 2006, 128,
15714.
6] D. Abbott, Proc. IEEE 2010, 98, 42.
7] S. Dunn, Int. J. Hydrogen Energy 2002, 27, 235.
8] P. V. Kamat, J. Phys. Chem. C 2007, 111, 2834.
9] N. S. Lewis, D. G. Nocera, Proc. Natl. Acad. Sci. USA 2006,
[27] K. Sivula, F. Le Formal, M. Graetzel, ChemSusChem 2011, 4,
103, 15729.
432.
[
[
10] A. J. Bard, M. A. Fox, Acc. Chem. Res. 1995, 28, 141.
11] S. Bensaid, G. Centi, E. Garrone, S. Perathoner, G. Saracco,
ChemSusChem 2012, 5, 500.
[28] K. Sivula, R. Zboril, F. Le Formal, R. Robert, A. Weidenkaff,
J. Tucek, J. Frydrych, M. Graetzel, J. Am. Chem. Soc. 2010,
132, 7436.
[12] X. Chen, S. Shen, L. Guo, S. S. Mao, Chem. Rev. 2010, 110,
[29] M. Gratzel, Nature 2001, 414, 338.
6503.
[30] X. J. Wang, H. Pang, S. Zhao, W. Shao, B. Yan, X. Li, S. Li,
J. Chen, W. Du, ChemPhysChem 2013, 14, 2518.
[31] Y. Song, P. Y. Zavalij, N. A. Chernova, M. Suzuki, M. S. Whit-
tingham, J. Solid State Chem. 2003, 175, 63.
[32] Y. Song, P. Y. Zavalij, N. A. Chernova, M. S. Whittingham,
Chem. Mater. 2005, 17, 1139.
[
13] G. W. Crabtree, M. S. Dresselhaus, M. V. Buchanan, Phys. To-
day 2004, 57, 39.
[
14] M. Gratzel, Acc. Chem. Res. 1981, 14, 376.
15] M. Ni, M. K. H. Leung, D. Y. C. Leung, K. Sumathy, Renew-
able Sustainable Energy Rev. 2007, 11, 401.
[
[
16] J. Nowotny, C. C. Sorrell, L. R. Sheppard, T. Bak, Int. J.
Received: June 17, 2015
Hydrogen Energy 2005, 30, 521.
Published Online: August 11, 2015
Eur. J. Inorg. Chem. 2015, 4237–4243
4243
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim