photocatalytic h2 production from ethanol–water mixtures over pt/tio2 and au/tio2 photocatalysts:...

13
ORIGINAL PAPER Photocatalytic H 2 Production from Ethanol–Water Mixtures Over Pt/TiO 2 and Au/TiO 2 Photocatalysts: A Comparative Study Vedran Jovic Zakiya H. N. Al-Azri Wan-Ting Chen Dongxiao Sun-Waterhouse H. Idriss Geoffrey I. N. Waterhouse Published online: 15 June 2013 Ó Springer Science+Business Media New York 2013 Abstract Pt/TiO 2 (Pt loadings 0–4 wt%) and Au/TiO 2 (Au loadings 0–4 wt%) photocatalysts were synthesized, char- acterized and tested for H 2 production from ethanol–water mixtures (80 vol% ethanol, 20 vol% H 2 O) under UV exci- tation. Average metal nanoparticle sizes determined by TEM were 1–3 nm for Pt in the Pt/TiO 2 photocatalysts and 5–7 nm for Au in the Au/TiO 2 photocatalysts. Au/TiO 2 showed an intense localized surface plasmon resonance feature at *570 nm, typical for metallic Au nanoparticles of diameter *5 nm supported on TiO 2 . X-ray photoelectron spectros- copy and X-ray diffraction analyses established that Pt and Au were present in metallic form on the TiO 2 support. X-ray fluorescence revealed close accord between nominal and actual Pt and Au loadings. The Au/TiO 2 and Pt/TiO 2 phot- ocatalysts both displayed very high activities for H 2 pro- duction under UV irradiation, with the Au/TiO 2 samples affording slightly superior rates of H 2 production at most metal loadings. The 2 wt% Au/TiO 2 and 1 wt% Pt/TiO 2 photocatalysts showed the highest H 2 production rates (32–34 mmol g -1 h -1 ). Photoluminescence studies confirmed that Pt and Au nanoparticles positively enhance the photocatalytic properties of P25 TiO 2 for H 2 production by acting as electron acceptors and thereby suppressing electron–hole pair recombination in TiO 2 . Keywords Photocatalysis Hydrogen production Titania Platinum Gold Ethanol 1 Introduction The development of a sustainable hydrogen economy, in which H 2 will replace fossil fuels for electricity generation, hinges on the discovery of clean, low cost and sustainable technologies for H 2 manufacture, distribution and storage. Current hydrogen production is based primarily around steam methane reforming, which is an energy intensive process with a significant carbon footprint [1, 2]. Alternative H 2 manufacturing technologies, which preferably use renewable H 2 sources such as water or biofuels and harness solar, wind, geothermal or hydroelectric energy to facilitate H 2 manufacture, must be found. Light from the Sun is by far the most abundant source of energy on Earth, though pres- ently \ 0.05 % of the total power used by humans annually (15,000 GW) is generated using the Sun’s solar energy (excluding solar heating and biomass combustion) [3]. Solar H 2 production is expected to be the cornerstone of future H 2 economies, yet direct conversion of solar energy to H 2 remains economically and technically challenging. Semiconductor photocatalysis is one of the most prom- ising technologies for harvesting sunlight and generating energy carriers such as H 2 from renewable sources (water and biofuels) [1, 47]. Excitation of a semiconductor photocatalyst, such as TiO 2 , with photons of appropriate energy produces electron–hole (e - –h ? ) pairs which may Electronic supplementary material The online version of this article (doi:10.1007/s11244-013-0080-8) contains supplementary material, which is available to authorized users. V. Jovic Z. H. N. Al-Azri W.-T. Chen D. Sun-Waterhouse G. I. N. Waterhouse (&) School of Chemical Sciences, The University of Auckland, Private Bag 92019, Auckland, New Zealand e-mail: [email protected] H. Idriss (&) SABIC Research Centres, Riyadh, Saudi Arabia e-mail: [email protected] H. Idriss KAUST, Thuwal, Saudi Arabia 123 Top Catal (2013) 56:1139–1151 DOI 10.1007/s11244-013-0080-8

Upload: independent

Post on 04-Nov-2023

0 views

Category:

Documents


0 download

TRANSCRIPT

ORIGINAL PAPER

Photocatalytic H2 Production from Ethanol–Water Mixtures OverPt/TiO2 and Au/TiO2 Photocatalysts: A Comparative Study

Vedran Jovic • Zakiya H. N. Al-Azri •

Wan-Ting Chen • Dongxiao Sun-Waterhouse •

H. Idriss • Geoffrey I. N. Waterhouse

Published online: 15 June 2013

� Springer Science+Business Media New York 2013

Abstract Pt/TiO2 (Pt loadings 0–4 wt%) and Au/TiO2 (Au

loadings 0–4 wt%) photocatalysts were synthesized, char-

acterized and tested for H2 production from ethanol–water

mixtures (80 vol% ethanol, 20 vol% H2O) under UV exci-

tation. Average metal nanoparticle sizes determined by TEM

were 1–3 nm for Pt in the Pt/TiO2 photocatalysts and 5–7 nm

for Au in the Au/TiO2 photocatalysts. Au/TiO2 showed an

intense localized surface plasmon resonance feature at

*570 nm, typical for metallic Au nanoparticles of diameter

*5 nm supported on TiO2. X-ray photoelectron spectros-

copy and X-ray diffraction analyses established that Pt and

Au were present in metallic form on the TiO2 support. X-ray

fluorescence revealed close accord between nominal and

actual Pt and Au loadings. The Au/TiO2 and Pt/TiO2 phot-

ocatalysts both displayed very high activities for H2 pro-

duction under UV irradiation, with the Au/TiO2 samples

affording slightly superior rates of H2 production at most

metal loadings. The 2 wt% Au/TiO2 and 1 wt% Pt/TiO2

photocatalysts showed the highest H2 production

rates (32–34 mmol g-1 h-1). Photoluminescence studies

confirmed that Pt and Au nanoparticles positively enhance

the photocatalytic properties of P25 TiO2 for H2 production

by acting as electron acceptors and thereby suppressing

electron–hole pair recombination in TiO2.

Keywords Photocatalysis � Hydrogen production �Titania � Platinum � Gold � Ethanol

1 Introduction

The development of a sustainable hydrogen economy, in

which H2 will replace fossil fuels for electricity generation,

hinges on the discovery of clean, low cost and sustainable

technologies for H2 manufacture, distribution and storage.

Current hydrogen production is based primarily around

steam methane reforming, which is an energy intensive

process with a significant carbon footprint [1, 2]. Alternative

H2 manufacturing technologies, which preferably use

renewable H2 sources such as water or biofuels and harness

solar, wind, geothermal or hydroelectric energy to facilitate

H2 manufacture, must be found. Light from the Sun is by far

the most abundant source of energy on Earth, though pres-

ently\0.05 % of the total power used by humans annually

(15,000 GW) is generated using the Sun’s solar energy

(excluding solar heating and biomass combustion) [3]. Solar

H2 production is expected to be the cornerstone of future H2

economies, yet direct conversion of solar energy to H2

remains economically and technically challenging.

Semiconductor photocatalysis is one of the most prom-

ising technologies for harvesting sunlight and generating

energy carriers such as H2 from renewable sources (water

and biofuels) [1, 4–7]. Excitation of a semiconductor

photocatalyst, such as TiO2, with photons of appropriate

energy produces electron–hole (e-–h?) pairs which may

Electronic supplementary material The online version of thisarticle (doi:10.1007/s11244-013-0080-8) contains supplementarymaterial, which is available to authorized users.

V. Jovic � Z. H. N. Al-Azri � W.-T. Chen �D. Sun-Waterhouse � G. I. N. Waterhouse (&)

School of Chemical Sciences, The University of Auckland,

Private Bag 92019, Auckland, New Zealand

e-mail: [email protected]

H. Idriss (&)

SABIC Research Centres, Riyadh, Saudi Arabia

e-mail: [email protected]

H. Idriss

KAUST, Thuwal, Saudi Arabia

123

Top Catal (2013) 56:1139–1151

DOI 10.1007/s11244-013-0080-8

either recombine or react with adsorbed species. To realize

efficient H2 production from water or biofuels, the bottom

of the conduction band of the semiconductor should be

more negative than the redox potential of the H2O/H2

couple (0.00 V vs. NHE) and the top of the valence band of

the semiconductor should be more positive than redox

potential of the O2/H2O couple (1.23 V vs. NHE) [8, 9].

TiO2 satisfies both these criteria, and is typically modified

with electron accepting co-catalysts (typically noble metals

such as Pd, Pt or Au; graphene or CNTs or semiconductors

such as CuO) to provide active sites for H2 generation and

to suppress electron–hole pair recombination in TiO2 [4, 8–

14]. Over the last decade, a large number of papers have

been published relating to photocatalytic H2 production

over TiO2-based systems, yet many crucial aspects of

reaction mechanisms remain poorly understood. Presently

it is very difficult to compare the work of different groups,

and to gauge the relative merits of different co-catalysts for

promoting H2 production, due to differences in catalyst

formulation (e.g. co-catalyst loading and semiconductor

support) and experimental conditions used to evaluate

photocatalyst performance. Comparative studies of differ-

ent photocatalysts under standardized testing conditions

would help understanding the role of each component on

the reaction.

Noble metals with suitable work functions, such as plat-

inum (U = 5.39 eV) or gold (U = 5.31 eV), when highly

dispersed as nanoparticles over TiO2 surfaces, have strong

ability to activate TiO2 photocatalysts towards H2 produc-

tion [10, 15, 16]. The apparent Fermi level for supported Pt or

Au nanoparticles on TiO2 sits ideally between the bottom of

the conduction band of TiO2 (-500 mV vs. NHE) and above

the H2O/H2 couple (0.00 V vs. NHE). Au/TiO2 is -270 mV

relative to NHE for Au particles of 5 nm diameter [17, 18].

Accordingly, electrons photoexcited in TiO2 will migrate to

the semiconductor surface and on to the supported Pt or Au

nanoparticles, thereby reducing e-–h? pair recombination

and enhancing H2 production (metallic Pt or Au co-catalysts

are the active centers for H2 production). The activity of Pt/

TiO2 or Au/TiO2 photocatalysts generally depends on three

key parameters. (a) the polymorphic nature of the TiO2

support (anatase, rutile, brookite or combinations thereof),

(b) the nature and loading of the noble metal particles, and

(c) the surface structure of the noble metal-TiO2 interface.

Murdoch et al. recently demonstrated that photocatalysts

comprising gold nanoparticles (diameter 3–12 nm) depos-

ited on nanocrystalline anatase TiO2 supports, are extremely

active for H2 production from ethanol–water mixtures under

UV irradiation, and about 100 times more active than Au

nanoparticles deposited on nanocrystalline rutile TiO2 sup-

ports [15]. Liquid slurry photoreaction studies produced

hydrogen at a rate of 2 L/kgcat. min on a 2 wt% Au/anatase

TiO2 catalyst under UV irradiation of comparable intensity

to that provided by the Sun [11]. Putting this into perspective,

a 1 kW PEM fuel cell needs 15 L of H2 per min, hence Au/

TiO2 photocatalysts are promising first generation photo-

catalysts for hydrogen production from renewable biofuels.

Waterhouse et al. observed that photocatalytic activity of

Au/TiO2 photocatalysts for H2 production from bioethanol

could be further increased when Degussa P25 TiO2 (85 %

anatase and 15 % rutile) was the support phase instead of

nanocrystalline anatase, on account of synergistic electron

transfer between the rutile and anatase phases following

photoexcitation [19–24]. Available data suggests that the

activity of Au/TiO2 photocatalysts for H2 production is

dependent on the Au loading, but largely independent of

the Au nanoparticle size (for Au particles in the size range

3–12 nm). This contrasts dark reactions over Au–TiO2 cat-

alysts (e.g. low temperature CO oxidation or hydrogenation

reactions), where strong relationships between Au particle

size and catalytic performance have been established [25].

Studies by other groups have also demonstrated the high

catalytic activity of Pt/TiO2 materials for H2 production [10,

26–28]. However, confusion exists about the relative activ-

ities of Au and Pt co-catalysts in activating TiO2 for H2

production (especially from biofuels), motivating the present

investigation.

This study aimed to systematically evaluate and com-

pare the photocatalytic activity of Pt/TiO2 (0–4 wt% Pt)

and Au/TiO2 (0–4 wt% Au) photocatalysts for H2 pro-

duction from ethanol–water mixtures under UV irradiation.

The photocatalysts were prepared using the archetypal

semiconductor photocatalyst, Degussa P25 TiO2, to facil-

itate comparison with the prior work of other groups. The

synthesis of the photocatalysts, catalyst pretreatments

before testing, and the photocatalytic tests were standard-

ized as much as possible to allow the relative activities of

the Pt and Au co-catalysts in promoting H2 production to

be ascertained. These results could then be used in the

rational design of new and improved photocatalysts for H2

production from biofuels, such as ethanol.

2 Experimental Section

2.1 Materials

Hexachloroplatinic acid hexahydrate (H2PtCl6�6H2O,

98 %), Tetrachloroauric acid trihydrate (HAuCl4�3H2O,

C99 %), urea (C99.5 %) and absolute ethanol (Puriss,

C99.5 %) were all obtained from Sigma-Aldrich and used

without further purification. Degussa P25 TiO2 (85 wt%

anatase, 15 wt% rutile, C99.5 %) was obtained from a

local supplier. All solutions were prepared using milli-Q

water (18.2 MX cm resistivity).

1140 Top Catal (2013) 56:1139–1151

123

2.2 Au/TiO2 Photocatalyst Preparation

Gold nanoparticles (1–4 wt%) were deposited on Degussa

P25 TiO2 using the deposition–precipitation with urea

method described by Zanella et al., with some modifica-

tions [29]. Briefly, HAuCl4�3H2O (1.654 g) was dissolved

in milli-Q water (1 L) to give a gold stock solution

([Au3?] = 4.2 9 10-3 M). Gold stock solution (25, 50, 75

or 100 mL for 1, 2, 3 or 4 wt% Au/TiO2 photocatalysts,

respectively), milli Q water (175, 150, 125 or 100 mL,

respectively) and urea (5.04 g) were added to a glass Schott

bottle and vigorously stirred. Degussa P25 TiO2 (2 g) was

then added, and the resulting suspension thermostated at

80 �C in a dark room for 8 h. After 8 h, the yellow Au(III)

impregnated TiO2 powders were collected by vacuum fil-

tration, washed repeatedly with milli-Q, and then air dried

at 70 �C overnight. The samples were then calcined at

300 �C for 2 h to thermally reduce surface Au(III) species

on TiO2 to metallic gold. The final Au/TiO2 powders were

all purple in colour.

2.3 Pt/TiO2 Photocatalyst Preparation

Platinum nanoparticles (1–4 wt%) were deposited on De-

gussa P25 TiO2 using the deposition–precipitation with

urea method. Briefly, H2PtCl6�6H2O (2.175 g) was dis-

solved in milli-Q water (1 L) to give a platinum stock

solution ([Pt4?] = 4.2 9 10-3 M). All other procedures

were identical to those described above for preparation of

the Au/TiO2 photocatalysts, except that here the platinum

stock solution was used instead of the gold stock solution.

After calcination at 300 �C for 2 h, the Pt/TiO2 photocat-

alysts were additionally treated under a H2/N2 flow

(10 vol% H2, 100 mL min-1) at 350 �C for 2 h to reduce

adsorbed Pt(IV) or Pt(II) species to metallic form. The

reduced samples were all grey in colour [30, 31].

2.4 Characterization

TEM images were taken on a JEOL 2012F TEM operating

at 200 kV, equipped with an Oxford Instruments ISIS

energy-dispersive X-ray spectroscopy system for qualita-

tive analysis of elements present in each sample. Powders

samples were dispersed in isopropanol and then several

drops of the resulting dispersion placed on holey carbon

coated copper TEM grids for analysis.

Powder XRD patterns were collected using Philips PW-

1130 diffractometer, equipped with a Cu anode X-ray tube

and a curved-graphite filter monochromator. Data was

taken from 2h = 2–100� (0.02�, 2� min-1) using Cu Ka X-

rays (k = 1.5418 A). The rutile:anatase ratio in the sam-

ples was determined according to the method described by

Fu et al. [32].

%Rutile =1

ðA/R)� 0:884þ1½ � � 100

where A is the peak area for the anatase (101) reflection at

2h = 25.3�, and R is the peak area for the rutile (110)

reflection at 2h = 27.4�.

N2 physisorption isotherms were determined at liquid

nitrogen temperature (-195 �C) using a Micromeritics

Tristar 3000 instrument. Specific surfaces areas were cal-

culated from the N2 adsorption data according to the

Brunauer–Emmett–Teller (BET) method using P/Po values

in the range 0.05–0.2. Cumulative pore volumes and pore

diameters were calculated from the adsorption isotherms

by the Barrett–Joyner–Halenda (BJH) method. Samples

were degassed at 100 �C under vacuum for 1 h prior to the

N2 physisorption measurements.

UV–Vis absorbance spectra were collected over the

wavelength range 250–900 nm on a Shimadzu UV-2101PC

scanning spectrophotometer fitted with a Shimadzu ISR-

260 integrating sphere attachment. BaSO4 powder was

used as a reference.

XPS data was collected using a Kratos Axis UltraDLD

equipped with a hemispherical electron energy analyzer and

an analysis chamber of base pressure *1910-9 Torr.

Spectra were excited using monochromatic Al Ka X-rays

(1486.69 eV), with the X-ray source operating at 100 W.

Sample were gently pressed into thin pellets of *0.1 mm

thickness for the analyses. A charge neutralisation system

was used to alleviate sample charge build up, resulting in a

shift of *3 eV to lower binding energy. Survey scans were

collected at a pass energy of 80 eV over the binding energy

range 1200–0 eV, whilst core level scans were collected with

a pass energy of 20 eV. The spectra were calibrated against

the C 1s signal at 285 eV from adventitious hydrocarbons.

Photoluminescence measurements were performed in air

at 25 �C using a Perkin-Elmer LS-55 Luminescence

Spectrometer. Spectra were excited at 310 nm and photo-

luminescence spectra were recorded over a range of

330–600 nm using a standard photomultiplier. A 290 nm

cutoff filter was used.

2.5 Photocatalytic Testing

Photocatalytic hydrogen production tests on the Pt/TiO2 and

Au/TiO2 photocatalysts were carried out in a tubular Pyrex

reactor (100 mL volume). Photocatalyst (0.0065 mg) was

placed in the reactor and evacuated under a nitrogen flow for

30 min to remove oxygen. Ethanol (15 mL) and milli-Q

water (3.75 mL) were then injected into the reactor through a

rubber septum and the mixture stirred continuously in the

dark for 1 h. The reactor was then exposed to UV light,

supplied from a Spectraline model SB-1000P/F lamp

Top Catal (2013) 56:1139–1151 1141

123

(200 W, 365 nm) at a distance of 5 cm from the reactor. The

photon flux at the sample was *6.5 mW cm-2 (the UV flux

from the Sun is *5 mW cm-2). Hydrogen evolution was

monitored by taking gas head space samples (1 mL) at

20 min intervals and injecting these into a Shimadzu GC

2014 equipped with a TCD detector and Carboxen-1010 plot

capillary column (L 9 I.D. 30 m 9 0.53 mm, average

thickness 30 lm). H2 produced through photoreaction was

quantified against an external calibration curve. Photocata-

lytic tests for each sample were carried out in triplicate.

3 Results and Discussion

3.1 Photocatalyst Characterization

Figure 1 shows TEM images of the 1–4 wt% Pt/TiO2 phot-

ocatalysts after H2 treatment at 350 �C for 2 h. For all

samples, supported Pt nanoparticles of size \3 nm can be

seen in the images. No attempt was made to determine the

exact Pt nanoparticle size distributions, as the size limit for

detection of metal nanoparticles by TEM is *1 nm. Pt

loadings determined by XRF for the Pt/TiO2 photocatalysts

(Table 1) were in excellent agreement with the nominal

loadings of 1–4 wt%. Powder XRD patterns for the Pt/TiO2

samples (Online Resource Figure 1) contained only peaks

characteristic for anatase and rutile in the P25 TiO2 support.

For all samples, the anatase:rutile weight ratio was calcu-

lated to be *5:1. The fact that no peaks characteristic for Pt

were observed by XRD, even at high Pt loadings (4 wt%), is

attributed to the very small size of the supported Pt clusters.

Extreme line broadening due to quantum size effects is

expected for Pt nanoparticles of size 1–3 nm. Also, Pt

nanoparticles of diameter 1.5 nm contain only 39 atoms and

hence would not be expected to have a well-developed f.c.c.

lattice that could be probed by XRD. XPS data below pro-

vides strong evidence for the presence of metallic Pt in these

samples. UV–Vis absorption spectra for the Pt/TiO2 samples

(Online Resource Figure 2) showed strong absorption below

400 nm due to the TiO2 support, as well as intense absorption

across the entire visible spectrum due to Pt interband tran-

sitions from Pt 5d ? 6sp and within the 6sp band [33].

TEM images for the 1–4 wt% Au/TiO2 photocatalysts are

shown in Figure 2. Au nanoparticles of size range 1–12 nm

Fig. 1 TEM images of a 1 wt%

Pt/TiO2, b 2 wt% Pt/TiO2,

c 3 wt% Pt/TiO2, and d 4 wt%

Pt/TiO2 photocatalysts. The Pt

nanoparticles are of size 1–3 nm

and highly dispersed over the

TiO2 support

1142 Top Catal (2013) 56:1139–1151

123

are seen in the images, located preferentially at the interface

between two TiO2 crystallites. The Au nanoparticles dis-

played pseudo-spherical morphologies with metal-support

contact angles[90�, evidence for a relatively weak metal-

support interaction. On increasing the Au loading, the

average Au particle size did not increase significantly.

Figure 3a–d shows Au particle size distributions for the

1–4 wt% Au/TiO2 samples. The mean Au size (d/nm),

standard deviation and number of particles counted are

shown in the figure. It should be noted that Au/TiO2 catalyst

pretreatment (either calcination at 300 �C or H2 exposure at

350 �C) did not affect the size distribution of Au nanopar-

ticles. XRD patterns for the 1–4 wt% Au/TiO2 photocata-

lysts (Online Resource Figures 3 and 4) were dominated by

reflections characteristic of nanocrystalline anatase and

rutile in the P25 TiO2 support. The weight fraction of anatase

and rutile in the samples was 5:1 for all samples. The Au/

TiO2 samples showed additional broad and weak XRD fea-

tures that intensified with nominal Au loading, which were

readily assigned to metallic gold nanoparticles with f.c.c.

structure on the TiO2 support. Online Resource Figure 4

shows an expanded view of the XRD patterns between

2h = 35–55�, where the development of Au(111) and

Au(200) reflections with increasing gold loading can be seen.

Identification of metallic gold by XRD for the Au/TiO2

photocatalysts agrees qualitatively with TEM and UV–Vis

absorbance data above for the same samples (Figs. 1, 3e), as

well as XPS data presented below. Figure 3e shows UV–Vis

absorbance spectra for the Au/TiO2 photocatalysts. The

samples have a broad absorption *570 nm due to the Au

nanoparticle LSPR, as well as absorption across the whole

visible spectrum because of Au interband transitions from

Au 5d ? 6sp and within the 6sp band [33–36]. The intensity

of the Au LSPR signal increased linearly with Au loading

(Online Resource Figure 5). The position and intensity of the

Au LSPR signal depends on Au particle size, shape and

dielectric constant of the support [33, 34]. Kimura et al.

reported that the LSPR for Au nanoparticles of size 5 nm on

anatase TiO2 is *570 nm [33], in good accordance with the

results of the present study (anatase is the dominant TiO2

polymorph in P25 TiO2). The asymmetry on the long

wavelength side of the Au/Anatase LSPR feature may be due

to Au/rutile, the LSPR for which is predicted to be *40 nm

higher than that of Au/Anatase LSPR at the same Au nano-

particle size [33]. Plots of (ahm)1/2 versus E, calculated from

the spectra of Fig. 3e, are shown in Fig. 3f. The extrapolated

band gap energy for P25 TiO2 is 3.15 eV, in good agreement

with literature reports [37, 38]. As the Au loadings were

increased, the titania absorption edge showed a distinctive

red shift. Extrapolated bandgap energies were 3.01, 2.95,

2.85 and 2.80 eV for the 1, 2, 3 and 4 wt% Au/TiO2 samples,

respectively. Comparable shifts with increasing gold loading

have been observed by other groups [36, 38, 39].Ta

ble

1S

um

mar

yo

fth

ep

hy

sica

l,ch

emic

alan

dp

ho

toca

taly

tic

pro

per

ties

of

the

0–

4w

t%P

t/T

iO2

and

Au

/TiO

2p

ho

toca

taly

sts

Sam

ple

Mea

nP

to

rA

u

par

ticl

esi

ze,

d,

(nm

)

Pt

or

Au

con

ten

t

by

XR

F(w

t%)

Ato

mo

rw

eig

ht

%b

yX

PS

BE

Tsu

rfac

e

area

(m2

g-

1)

BJH

cum

ula

tiv

e

po

rev

olu

me

(cm

3g

-1)

BJH

aver

age

po

red

iam

eter

(nm

)

H2

pro

du

ctio

n

rate

(mm

ol

g-

1h

-1)

H2

pro

du

ctio

n

rate

(mm

ol

m-

2h

-1)

Ti

OC

M=

Pt

or

Au

(M/T

i)

P2

5T

iO2

–0

.02

5.2

52

.12

2.7

–4

9.6

0.1

45

14

.61

.26

0.0

25

1w

t%P

t/T

iO2

\2

1.0

24

.75

0.9

23

.90

.5(0

.02

0)

48

.10

.36

02

7.8

31

.70

.65

9

2w

t%P

t/T

iO2

\2

1.8

23

.55

0.7

25

.30

.5(0

.02

1)

48

.40

.35

02

7.6

28

.70

.59

4

3w

t%P

t/T

iO2

\2

2.9

25

.05

0.6

23

.80

.6(0

.02

4)

47

.00

.36

62

8.0

27

.50

.58

4

4w

t%P

t/T

iO2

\2

4.1

25

.35

0.8

23

.20

.7(0

.02

8)

48

.20

.36

02

8.5

27

.50

.57

0

Pt

foil

––

3.4

26

.96

9.7

––

––

1w

t%A

u/T

iO2

4.8

0.9

23

.05

1.8

25

.00

.2(0

.00

9)

48

.10

.39

02

9.2

32

.50

.67

5

2w

t%A

u/T

iO2

6.4

2.0

23

.64

9.5

26

.50

.4(0

.01

7)

47

.10

.32

02

5.8

33

.40

.70

9

3w

t%A

u/T

iO2

6.6

2.8

25

.05

1.2

23

.20

.6(0

.02

4)

46

.50

.35

02

8.6

31

.50

.67

6

4w

t%A

u/T

iO2

4.8

4.1

25

.15

0.9

23

.20

.8(0

.03

2)

46

.50

.31

12

6.6

24

.70

.53

2

Au

foil

––

––

–1

00

.0

Top Catal (2013) 56:1139–1151 1143

123

XPS was used to probe the oxidation states of platinum

and gold at the surface of the Pt/TiO2 and Au/TiO2 phot-

ocatalysts, respectively. Figure 4a shows Pt 4f XPS spectra

for the Pt/TiO2 photocatalysts. The samples show peaks at

71.3 and 74.7 eV, respectively, in a 4:3 area ratio, which

indicate the presence of metallic Pt (Pt 4f7/2 and Pt 4f5/2,

respectively). As expected, the intensity of these peaks

increased with an increase in the nominal Pt loading of the

samples. Pt 4f peaks for metallic Pt are inherently asym-

metric, so we cannot completely exclude a minor contri-

bution from Pt(OH)2 or PtO (Pt 4f7/2 = 74.0 eV) or PtO2

(Pt 4f7/2 = 74.8 eV) species. However, on the basis of the

XPS data we can conclude that metallic Pt is the dominant

Pt species on the surface of the Pt/TiO2 photocatalysts. Pt

loadings determined by quantitative XPS analysis were

higher than those determined by XRF (Table 1), which is

not unexpected as XPS is a surface-analytical technique.

Figure 4b shows Au 4f XPS spectra for the Au/TiO2

photocatalysts, which contain features at 83.6 and 87.3 eV

in a 4:3 peak area ratio. These features intensify linearly on

increasing the nominal Au loading, and are assigned to the

Au 4f7/2 and Au 4f5/2 peaks of supported metallic gold

nanoparticles. The binding energies of these features are

*0.4 eV lower than those determined for a metallic gold

foil (Au 4f7/2 = 84.0 eV, and Au 4f7/2 = 87.7 eV), con-

sistent with the finding of other groups [11, 15, 37]; this is

attributed to the shift in the Au Fermi level towards the

conduction band of TiO2 thus affecting Au 4f as well as Ti

2p lines [40]. The XPS data eliminates the possibility that

amorphous Au2O3 (Au 4f7/2 = 86.9 eV) or other Au3?-

containing species, like Au(OH)3, coexist with the Au

nanoparticles on the TiO2 surface. Au loadings determined

by quantitative XPS analysis were higher than those

determined by XRF (Table 1).

N2 physisorption data for Degussa P25 TiO2 and the Au/

TiO2 and Pt/TiO2 photocatalysts is summarized in Table 1.

The BET specific surface areas for all the metal-containing

photocatalysts were similar (46–48 m2 g-1) and slightly

lower than that of the P25 TiO2 support (49.6 m2 g-1). The

BJH cumulative pore volume and BJH average pore

diameters of all the metal-containing photocatalysts were

also similar (0.31–0.39 cm3 g-1 and 25.8–29.2 nm,

respectively) and largely independent of metal loading.

Compared to data collected for the Degussa P25 TiO2

support, the BJH cumulative pore volume and BJH average

pore diameter of samples increased after noble metal

deposition, which is simply explained by Au and Pt

nanoparticles blocking micropores in the TiO2 support.

3.2 Photoluminescence Measurements

Photocatalytic H2 production relies on using excited elec-

trons to reduce hydrogen ions and at the same time holes

are trapped by a sacrificial agent (as well as water mole-

cules) resulting in oxidation. In competition with this

e-–h? recombination occurs. In the bulk and at the surface

a fraction of these e- and h? is lost via three routes:

radiative recombination, Auger recombination and

Fig. 2 TEM images of a 1 wt%

Au/TiO2, b 2 wt% Au/TiO2,

c 3 wt% Au/TiO2; and d 4 wt%

Au/TiO2 photocatalysts. The Au

nanoparticles appear as dark

spots in the TEM images. For

all samples, the Au nanoparticle

size is *3–8 nm. Selected area

diffraction patterns confirm the

presence of nanocrystalline

anatase, rutile and Au in the

samples

1144 Top Catal (2013) 56:1139–1151

123

Schockely Read Hall recombination (defects). The energy

released during the e-–h? recombination in the form of

light gives rise to photoluminescence [41–43]. Therefore

the higher the luminescence signal the lower is the

expected catalytic activity. There are methods of estimat-

ing the e-–h? life time from the attenuation of the PL

Au Particle Size (nm)

0 2 4 6 8 10 12

Pe

rce

ntag

e (

%)

0

5

10

15

20

25

d = 4.8 nm s = 1.9 nm cts. = 129

Au Particle Size (nm)

0 2 4 6 8 10 12

Per

cent

age

(%)

0

5

10

15

20

25

d = 6.4 nm s = 2.1 nm cts. = 119

Au Particle Size (nm)

0 2 4 6 8 10 12

Pe

rce

ntag

e (

%)

0

5

10

15

20

25

d = 6.6 nmσ = 2.1 nm

cts. = 91

Au Particle Size (nm)

0 2 4 6 8 10 12

Pe

rce

ntag

e (

%)

0

5

10

15

20

25

(a) (b)

(c) (d)

1 wt.% Au 2 wt.% Au

3 wt.% Au 4 wt.% Au d = 4.8 nm

σ = 1.9 nm cts. = 129

Wavelength (nm)

300 400 500 600 700 800 900

Ab

sorb

ance

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6P25 TiO2

1 wt.% Au2 wt.% Au3 wt.% Au4 wt.% Au

(e)

Energy (eV)

1.5 2.0 2.5 3.0 3.5 4.0

( αh ν

)1/2 /

(eV

1/2cm

-1/2)

0.0

0.5

1.0

1.5

2.0

2.5

3.0P25 TiO2

1 wt.% Au2 wt.% Au3 wt.% Au4 wt.% Au

(f)

Fig. 3 Au nanoparticle size distributions for a 1 wt% Au/TiO2,

b 2 wt% Au/TiO2, c 3 wt% Au/TiO2; and d 4 wt% Au/TiO2

photocatalysts, e UV–Vis absorbance spectra for the Au–TiO2

photocatalysts, where Au nanoparticles gives rise to an intense LSPR

feature at *570 nm, f corresponding (ahv)1/2 versus E plots. The

electronic absorption edge shifts to shorter wavelengths with

increasing Au loading

Top Catal (2013) 56:1139–1151 1145

123

luminescence signal but they require considerable cali-

bration that is beyond the objective of the work [43]. Au

and Pt nanoparticles are postulated to enhance the photo-

catalytic activity of TiO2 for H2 production by acting as

‘‘electron sinks’’ and suppressing electron–hole pair

recombination in TiO2 [44]. To further study this, photo-

luminescence measurements were carried out on P25 TiO2

and the Pt/TiO2 and Au/TiO2 photocatalysts (Fig. 5). P25

TiO2 gave a very intense photoluminescence signal, indi-

cating that electron–hole pair recombination occurred

rapidly following photoexcitation in the absence of a noble

metal co-catalyst. An intense feature is seen at *3.1 eV,

which corresponds to direct (X1b ? X1a, 3.45 eV and

X1b ? X2b, 3.59 eV) and phonon-assisted indirect transi-

tions (X1b ? C3, 3.19 eV; C1b ? X2b 3.05 eV; and

C1b ? X1a 2.91 eV) in anatase TiO2 (possible contribu-

tions by rutile are masked by the large excess of anatase in

P25 TiO2) [45, 46]. Weaker features at longer wavelengths

(lower energy) are due to shallow traps, (VO) ? C3,

involving oxygen defects [45]. Figure 6 shows a partial

energy level diagram for anatase TiO2 indicating these

allowed direct and indirect transitions (X denotes the edge,

and C the center, of the Brillouin Zone) [45–47]. Deposi-

tion of Pt or Au nanoparticles on the surface of the TiO2

support strongly attenuated the photoluminescence signal

of P25 TiO2, with the signal becoming progressively

weaker with increasing metal loading (Fig. 5). These

results confirm that Au or Pt nanoparticles suppress elec-

tron–hole pair recombination in TiO2 following photoex-

citation, by acting as ‘‘electron sinks’’ and in doing so

increasing the number of charge carriers (h? or e-) avail-

able for photoreactions on TiO2 surfaces. Online Resource

Figure 6 shows a plot of the normalized PL intensity versus

metal loading for the Pt/TiO2 and Au/TiO2 photocatalysts.

At a metal loading of 1 wt%, the abilities of the Pt and Au

nanoparticles to suppress electron hole-pair recombination

are similar, but Au becomes more effective than Pt at high

metal loadings. Interestingly, the photoluminescence signal

maximum for the Pt/TiO2 and Au/TiO2 photocatalysts were

slightly blue-shifted relative to that of P25 TiO2. A similar

photoluminescence blue-shift was reported by Nakajima

et al. for Pt-loaded TiO2 powders [48, 49]. Nakajima et al.

argued that following initial electron transfer from TiO2 to

Pt, further transfer of photoinduced electrons to Pt was

inhibited, and the additional photoinduced electrons could

then move in a restricted region in the TiO2, leading to a

blue shift in the PL. An alternative explanation we offer is

that supported metal nanoparticles selectively suppress

phonon-assisted indirect transitions (X1b ? C3, 3.19 eV;

C1b ? X2b 3.05 eV; and C1b ? X1a 2.91 eV) in anatase

relative to the other transitions, which would blue-shift the

photoluminescence maximum accordingly. The energies of

the direct (X1b ? X1a, 3.45 eV and X1b ? X2b, 3.59 eV)

and shallow trap transitions (VO) ? C3 were largely

unaffected by the presence of Pt or Au nanoparticles.

3.3 Photocatalytic Hydrogen Production Tests

The photocatalytic activity of Degussa P25 TiO2, and the

Pt/TiO2 and Au/TiO2 photocatalysts, for H2 production

from ethanol–water mixtures (80 vol% ethanol, 20 vol%

Binding energy (eV)68707274767880

Inte

nsity

(ar

bitr

ary

units

)Pt 4f7/2

Pt 4f5/2

3 wt.% Pt

2 wt.% Pt

1 wt.% Pt

0 wt.% Pt

Pt foil

4 wt.% Pt

(a)

Binding Energy (eV)8284868890

Inte

nsity

(ar

bitr

ary

units

)

0 wt.% Au

1 wt.% Au

2 wt.% Au

3 wt.% Au

4 wt.% Au

Au foil

(b) Au 4f7/2

Au 4f5/2

Fig. 4 Core level a Pt 4f XPS spectra for the Pt/TiO2 photocatalysts

and b Au 4f XPS spectra for the Au/TiO2 photocatalysts. The Pt 4f7/2

and Au 4f7/2 binding energies are consistent with the metallic form of

Pt and Au, respectively. The peak areas of the Pt 4f and Au 4f signals

increase almost linearly with nominal metal loading

1146 Top Catal (2013) 56:1139–1151

123

H2O) were evaluated under UV light of intensity compa-

rable to that present in Sunlight. The mechanism of this

reaction has been the subject of numerous studies and will

not be discussed in detail here [11, 50]. However, the key

overall steps are as follows. TiO2 absorbs photons with

energy greater than the electronic band gap of TiO2 (Eg

*3.1–3.2 eV), resulting in the formation of electron–hole

pairs (e-–h?). Photoexcited holes (h?) in the valence band

of TiO2 oxidise ethanol (CH3CH2OH ? 2 h? ? CH3-

CHO ? 2H?) and H2O (H2O ? 2 h? ? 0.5O2 ? 2H?),

whilst photoexcited electrons in the conduction band of

TiO2 reduce H? to H2 (2H? ? 2e- ? H2) or H2O to H2

(2H2O ? 4e- ? H2 ? 2OH-). Figure 7a shows a plot of

H2 production (mmol g-1) versus time (min) for Degussa

P25 TiO2 and the 1–4 wt% Pt/TiO2 photocatalysts. Fig-

ure 7b shows corresponding plots for the Au/TiO2 photo-

catalysts. The Degussa P25 TiO2 support, in the absence of

a co-catalyst, exhibited minimal H2 production activity

(1.26 mmol g-1 h-1). The H2 production activity of De-

gussa P25 TiO2 increased dramatically after the deposition

of Pt or Au co-catalysts. For both sets of photocatalysts,

linear H2 production rates were observed with UV expo-

sure (we have extended these tests over longer periods up

to 24 h and the same linearity was observed). Photocata-

lytic H2 production rates for all samples are plotted in

Fig. 8. For the Pt/TiO2 photocatalysts, the 1 wt% Pt sample

gave the highest H2 production rate (31.7 mmol g-1 h-1).

At higher loadings, the activity decreased progressively.

For the Au/TiO2 photocatalysts, a 2 wt% Au loading

appears optimal, affording a H2 production rate of

33.4 mmol g-1 h-1. All of the samples possessed similar

specific surface areas and pore size distributions (Table 1),

so H2 production rates normalized against photocatalyst

surface area followed the same trend. The optimum metal

loading can be rationalised in terms of (1) suppression of

electron–hole pair recombination at the TiO2 surface due to

electron trapping by metal nanoparticles; (2) loss of TiO2

surface for ethanol and water adsorption with increasing

metal loading. It is noteworthy to indicate that a similar

loading of Au on Anatase has been seen to give the highest

hydrogen production activity from water with 0.1 vol% of

methanol [51].

The decrease in the photoluminescence spectra in the

case of Pt/TiO2 and Au/TiO2 cannot be strictly correlated

with the photocatalytic activity. The quantum yield of a

photocatalytic reaction is a function of two main param-

eters: the rate constant of charge transfer and the rate

constant of electron–hole recombination [8, 52]. The

decrease in photoluminescence on addition of Pt or Au

cocatalysts is clear and may partly be linked to increasing

the life time of charge carriers. In simple terms the

presence of impurities and/or defects results in changing

the wave vector directions of the excited electrons upon

relaxation within the conduction band thus increases their

life time. While the rutile band gap is direct (allowing for

fast electron–hole recombination) the anatase band gap is

indirect (although still under debate) and the charge car-

rier life time in anatase is orders of magnitude higher than

in rutile [52, 53]. As stated in the introduction, the pres-

ence of both phases has been seen to considerably

increase the photocatalytic activity of TiO2. Various

mechanisms have been proposed to explain the synergistic

effect of both phases on the photoreaction, and these have

been recently summarized [53, 54]. The TEM images of

Fig. 2 showed that most Au particles are in between the

rutile and anatase particles and this might allow for the

Energy (eV)2.02.22.42.62.83.03.23.43.63.8

Abs

orba

nce

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

PL

Inte

nsity

0

200

400

600

800

1000

1200

1400

1600

P25 TiO2 (abs.)

P25 TiO2

1 wt.% Pt2 wt.% Pt3 wt.% Pt4 wt.% Pt

2.34

eV

2.56

eV

2.70

eV

2.80

eV

2.91

eV

3.05

eV

3.19

eV

3.45

eV

3.59

eV(a)

Energy (eV)

2.02.22.42.62.83.03.23.43.63.8

Abs

orba

nce

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4P

L In

tens

ity

0

200

400

600

800

1000

1200

1400

1600

P25 TiO2 (abs.)

P25 TiO2

1 wt.% Au2 wt.% Au3 wt.% Au4 wt.% Au

2.34

eV

2.56

eV

2.70

eV

2.80

eV

2.91

eV

3.05

eV

3.19

eV

3.45

eV

3.59

eV(b)

Fig. 5 Photoluminescence spectra for a 0–4 wt% Pt/TiO2 photocat-

alysts and b 0–4 wt% Au/TiO2 photocatalysts. The photolumines-

cence signal of P25 TiO2 is progressively suppressed with increasing

metal loading. Vertical dotted lines show the characteristic transitions

of anatase TiO2

Top Catal (2013) 56:1139–1151 1147

123

charge transfer to occur more efficiently. Therefore the

presence of Au particles (and by analogy Pt particles)

might have a role in increasing the life time of excited

electrons. Alternatively, one could simply attribute the

decrease of the photoreaction rate with increasing Au

loading to the decrease of available surface sites (three

phases together: anatase, rutile and gold particles) because

statistically increasing the % of one phase (Au in this

case) may, after a given threshold, decrease the number of

reaction sites (three-phase requirement). We will make an

argument that this is not the case. From Table 1, it can be

seen that the at.% Au detected by XPS increased linearly

with Au loading (i.e. it doubles with doubling the amount

of gold). Knowing the mean particle diameter of the Au

particles (from TEM) and number of surface Ti sites per

unit area in TiO2 the fraction of surface Ti atoms (and O

atoms) that have been covered by Au particles can be

estimated. The geometric surface area of Au particle of

5 nm size is 7.85 9 10-17 m2 (and its cross sectional area

is about 2 9 10-17 m2); a 5 nm particle of Au contains

about 3000 atoms of Au. Because the escape depth of the

XPS Au4f is relatively high (about 2.5 nm) it is reason-

able to assume that a large fraction of all Au atoms are

seen by XPS (assuming perfectly spherical particles).

Therefore the at.% Au values determined by XPS can be

taken as a true count of the number of the number of Au

atoms. The TiO2 unit cell is about 20 9 10-20 m2 in area,

therefore the number of TiO2 unit cells that is covered by

Au particles can be estimated from total number of Au

particles and their cross section area: 6.6 9 1012 particles

of Au/m2 9 100 unit cells of TiO2 = 6.6 9 1014 unit

cells covered). In 1 m2 there is about 0.23 9 1019 TiO2

unit cells. Therefore the argument of geometrical cover-

age reducing the photoreaction rate at high metal loadings

cannot be made. The reason is therefore purely electronic.

It was recognized early on that the photoreaction for

hydrogen production from alcohol and water is very

sensitive to the % of noble metals on the semiconductor,

with the optimal metal loading depending on the metal

and the support [54, 55]. One of the arguments invoked is

that the distance between the metal particles affects the

electron–hole recombination rate [55, 56]. In the absence

of more fundamental studies on this, it is not possible to

give a definitive explanation without speculation. How-

ever, changes in the electric field can cause changes in the

electron–hole recombination rate [57] and it is not

unreasonable to consider local changes in the electric field

due to changes in the distance separating the noble metal

particles [56]. Table 2 presents H2 production rates nor-

malized against the noble metal/Ti XPS ratio as well as

by the number of noble metal atoms. It is clear that rates

cannot be normalized the same way thermal catalytic

reactions are normalized and the present numbers are

merely given to show the trends.

Transition E (eV) λ(nm)

X2b ↔ X1b 3.59 345.4

X1a ↔ X1b 3.45 359.4

Γ3

Γ3

Γ3

↔ X1b 3.19 388.7

X2b↔ Γ1b

X1a ↔ Γ1b

3.05 406.6

2.91 426.1

↔ (VO)4 2.80 442.9

↔ (VO)3 2.70 459.3

↔ (VO)2 2.56 484.4

↔ (VO)1 2.34 529.93

1b

1b

1a2b

(VO)

E

(VO)(VO)3

(VO)

X

Γ

Γ3

Γ3

Fig. 6 Partial energy diagram

for anatase TiO2. Possible

transitions are highlighted

1148 Top Catal (2013) 56:1139–1151

123

The contribution of the plasmon effect to the photore-

action is also worth consideration. While the Au LSPR

increases linearly with increasing the number of Au parti-

cles (i.e. the Au loading), the photoreaction rate does not.

One simple explanation would be that the plasmonic effect

of Au into the reaction is small compared to that of the

semiconductor. Another explanation is that under the

present study (UV excitation only) only a fraction of the

excitation source will have the exact frequency needed to

excite Au particles efficiently. However the same argument

as above can be invoked which is related to the consider-

able enhancement of the electric field due to Au (or Pt)

particles [57]. Therefore a balance has to be found between

the increase in the rate due to the plasmonic effect and the

decrease in the rate due to the increasing local electric field.

The slightly lower activity of the Pt based catalysts

could be attributed to the non-rectifying Ohmic contact

formed at the interface between Pt and TiO2 particles

aiding the rapid dispersion of electrons to the surrounding

medium [17]. The rectifying nature of the Schottky barrier

in the Au/TiO2, manifests in a stronger ability of Au

nanoparticles to prevent charge recombination (evidenced

in the photoluminescence measurements of Fig. 5). Alter-

natively, metal nanoparticle size may play an important

role, with Au nanoparticles of average size 4–7 nm (Fig. 2)

offering superior electron scavenging properties compared

to the 1–3 nm Pt nanoparticles (Fig. 1). The energy sepa-

ration between the Fermi level of supported Pt or Au

nanoparticles and the bottom of the conduction band of

TiO2 becomes larger with increasing metal nanoparticle

size in the range 1–10 nm, facilitating electron capture and

H2 production as the metal nanoparticle size increases.

Conversely, for Pt and Au nanoparticles \2 nm, a loss of

metallic character could decrease electron transfer effi-

ciency and photocatalytic activity.

(a)

Time (min)

0 100 200 300 400

H2 P

rodu

ced

(mm

ol g

-1)

0

50

100

150

200

250P25 TiO2

1 wt.% Pt2 wt.% Pt3 wt.% Pt4 wt.% PtRegression

Time (min)

0 100 200 300 400

H2 P

rodu

ced

(mm

ol g

-1)

0

50

100

150

200

250P25 TiO2

1 wt.% Au2 wt.% Au3 wt.% Au4 wt.% AuRegression

(b)

Fig. 7 H2 evolution vs. UV exposure time plots for a Pt/TiO2

photocatalysts and b Au/TiO2 photocatalysts in ethanol–water

mixtures (80 vol% ethanol, 20 vol% H2O). Platinum and gold

nanoparticles dramatically enhance the activity of TiO2 for H2

production. No deactivation of the samples was noticeable over

extended periods

0 w

t.% M

etal

1 w

t.% M

etal

2 w

t.% M

etal

3 w

t.% M

etal

4 w

t.% M

etal

Rat

e of

H2 P

rodu

ctio

n (m

mol

g-1

h-1)

0

10

20

30

40P25 TiO2

Pt/TiO2

Au/TiO2

Fig. 8 Summarized H2 production rates over Pt/TiO2 and Au/TiO2

photocatalysts. The Au/TiO2 samples showed higher rates of H2

production up to 4 wt% loading

Top Catal (2013) 56:1139–1151 1149

123

4 Conclusions

Pt/TiO2 and Au/TiO2 photocatalysts, based on a Degussa

P25 TiO2 support (anatase 85 wt%, rutile 15 wt%), exhibit

very high photocatalytic activities for H2 production from

ethanol-water mixtures under UV excitation. Metallic Pt or

Au nanoparticles on the surface of TiO2 suppress electron–

hole pair recombination in TiO2 by acting as electron

acceptors, and serve as active sites for H2 production.

Under the applied testing conditions, the photocatalytic H2

production activities of Au/TiO2 photocatalysts were

marginally superior to those of the corresponding Pt/TiO2

photocatalysts at metal loadings in the range 1–3 wt%. The

high photocatalytic activities and stabilities reported here

for the Pt/TiO2 photocatalysts (31.7 mmol g-1 h-1 for

the 1 wt% Pt sample) and Au/TiO2 photocatalysts (33.4

mmol g-1 h-1 for the 2 wt% Au/TiO2 sample) suggest that

these are promising first generation materials for solar

hydrogen production from biofuels.

Acknowledgments We gratefully acknowledge funding support

from the University of Auckland, the Australian Institute for Nuclear

Science and Technology (AINSE) and the MacDiarmid Institute for

Advanced Materials and Nanotechnology.

References

1. Chiarello GL, Selli E (2010) Recent Pat Eng 3:155–169

2. Navarro RM, Pena MA, Fierro JL (2007) Chem Rev

10:3952–3991

3. Waterhouse GIN, Idriss H (2009) In: Vayssieres L (ed) On solar

hydrogen and nanotechnology. Wiley, Chichester

4. Ni M, Leung MKH, Leung DYC, Sumathy K (2007) Renew

Sustain Energy Rev 3:401–425

5. Matsuoka M, Kitano M, Takeuchi M, Tsujimaru K, Anpo M,

Thomas JM (2007) Catal Today 1–2:51–61

6. Shimura K, Yoshida H (2011) Energy Environ Sci 7:2467–2481

7. Rajeshwar K (2007) J Appl Electrochem 7:765–787

8. Linsebigler AL, Lu G, Yates JT (1995) Chem Rev 3:735–758

9. Fujishima A, Rao TN, Tryk DA (2000) J Photochem Photobiol C

1:1–21

10. Bamwenda GR, Tsubota S, Nakamura T, Haruta M (1995) J

Photochem Photobiol A 2:177–189

11. Nadeem MA, Murdoch M, Waterhouse GIN, Metson JB, Keane

MA, Llorca J, Idriss H (2010) J Photochem Photobiol A

2–3:250–255

12. Yu J, Hai Y, Jaroniec M (2011) J Colloid Interface Sci 1:223–228

13. Cheng P, Yang Z, Wang H, Cheng W, Chen M, Shangguan W,

Ding G (2012) Int J Hydrogen Energy 3:2224–2230

14. Park HG, Holt JK (2010) Energy Environ Sci 8:1028–1036

15. Murdoch M, Waterhouse GIN, Nadeem MA, Metson JB, Keane

MA, Howe RF, Llorca J, Idriss H (2011) Nat Chem 6:489–492

16. Chiarello GL, Selli E, Forni L (2008) Appl Catal B 1–2:332–339

17. Subramanian V, Wolf EE, Kamat PV (2004) J Am Chem Soc

15:4943–4950

18. Kamat PV (2008) J Phys Chem C 48:18737–18753

19. Waterhouse GIN, Murdoch M, Llorca J, Idriss H (2012) Int J

Nanotechnol 1:113–120

20. Bickley RI, Gonzalez-Carreno T, Lees JS, Palmisano L, Tilley

RJD (1991) J Solid State Chem 1:178–190

21. Datye AK, Riegel G, Bolton JR, Huang M, Prairie MR (1995) J

Solid State Chem 1:236–239

22. Ohno T, Sarukawa K, Tokieda K, Matsumura M (2001) J Catal

1:82–86

23. Ohtani B, Prieto-Mahaney OO, Li D, Abe R (2010) J Photochem

Photobiol A 2–3:179–182

24. Hurum DC, Agrios AG, Gray KA, Rajh T, Thurnauer MC (2003)

J Phys Chem B 19:4545–4549

25. Haruta M (1997) Catal Surv Jpn 1:61–73

26. Pichat P, Mozzanega M-N, Courbon H (1987) J Chem Soc Farad

Trans 1(3):697–704

27. Lin CH, Lee CH, Chao JH, Kuo CY, Cheng YC, Huang WN,

Chang HW, Huang YM, Shih MK (2004) Catal Lett 1:61–66

28. Gallo A, Montini T, Marelli M, Minguzzi A, Gombac V, Psaro R,

Fornasiero P, Dal Santo V (2012) ChemSusChem 9:1800–1811

29. Zanella R, Giorgio S, Henry CR, Louis C (2002) J Phys Chem B

31:7634–7642

30. Alonso F, Riente P, Rodrıguez-Reinoso F, Ruiz-Martınez J,

Sepulveda-Escribano A, Yus M (2008) J Catal 1:113–118

31. Huang B-S, Chang F-Y, Wey M-Y (2010) Int J Hydrogen Energy

15:7699–7705

32. Fu X, Clark LA, Yang Q, Anderson MA (1996) Environ Sci

Technol 2:647–653

33. Kimura K, Naya S-i, Jin-nouchi Y, Tada H (2012) J Phys Chem C

12:7111–7117

Table 2 Normalised H2 production rates for the Pt/TiO2 and Au/TiO2 photocatalysts

Catalyst x wt%

M/TiO2

Rate mmol

m-2 h-1XPS (M/Ti)

M = Pt or Au

Rate (in mmol m-2 h-1)/

(M/Ti) M = Pt or Au

Rate 1020

molecules m-2 h-1Rate (103 molecules

Matom-1 h-1)a M = Pt or Au

1 wt% Pt 0.659 0.020 33.0 3.97 7.94

2 wt% Pt 0.594 0.021 28.3 3.58 7.16

3 wt% Pt 0.584 0.024 24.3 3.52 5.87

4 wt% Pt 0.570 0.028 20.4 3.43 4.90

1 wt% Au 0.675 0.009 75.0 4.06 20.30

2 wt% Au 0.709 0.017 41.7 4.27 10.68

3 wt% Au 0.676 0.024 28.2 4.07 6.78

4 wt% Au 0.532 0.032 16.6 3.20 4.00

a Rate is in 1020 molecules/m2 h divided by the number of metals in 1 m2 assuming 1019 atoms per m2. For example for 1 wt% Pt the rate of

3.97 9 1020 molecules of H2 m-2 h-1 is divided by 0.5/100 9 1019 = 7940 molecules Matom-1 h-1. 0.5/100 is taken from Table 1

1150 Top Catal (2013) 56:1139–1151

123

34. Kowalska E, Mahaney OOP, Abe R, Ohtani B (2010) Phys Chem

Chem Phys 10:2344–2355

35. Chen J-J, Wu JCS, Wu PC, Tsai DP (2010) J Phys Chem C

1:210–216

36. Tanaka A, Sakaguchi S, Hashimoto K, Kominami H (2012) ACS

Catal 1:79–85

37. Rosseler O, Shankar MV, Du MK-L, Schmidlin L, Keller N,

Keller V (2010) J Catal 1:179–190

38. Li XZ, Li FB (2001) Environ Sci Technol 11:2381–2387

39. Cardenas-Lizana F, Gomez-Quero S, Idriss H, Keane MA (2009)

J Catal 2:223–234

40. Howard A, Clark DNS, Mitchell CEJ, Egdell RG, Dhanak VR

(2002) Surf Sci 3:210–224

41. Idriss H, Barteau MA (1992) J Phys Chem 96:3382–3388

42. Jing L, Qu Y, Wang B, Li S, Jiang B, Yang L, Fu W, Fu H, Sun J

(2006) Sol Energy Mater Sol Cells 90:1773–1787

43. Vazquez LN Master in photonics Experimental setup fopr carrier

lifetime measurement based on photoluminescence response.

Escole Tecnica Superior, Universtat Politecnica de Catalunya

(UPC)

44. Lunawat P, Kumar R, Gupta N (2008) Catal Lett 3–4:226–233

45. Serpone N, Lawless D, Khairutdinov R (1995) J Phys Chem

45:16646–16654

46. Abazovic ND, Comor MI, Dramicanin MD, Jovanovic DJ,

Ahrenkiel SP, Nedeljkovic JM (2006) J Phys Chem B

50:25366–25370

47. Daude N, Gout C, Jouanin C (1977) Phys Rev B 6:3229–3235

48. Nakajima H, Mori T, Watanabe M (2004) J Appl Phys 1:925–927

49. Nakajima H, Mori T (2006) Physica B 376–377:820–822

50. Sakata T, Kawai T (1981) Chem Phys Lett 2:341–344

51. Bowker M, Millard L, Greaves J, James D, Soares J (2004) Gold

Bull 37:170–173

52. Xu M, Gao Y, Moreno EM, Kunst M, Muhler M, Wang Y, Idriss

H, Woll C (2011) Phys Rev Lett 13:138302

53. Connelly K, Wahab AK, Idriss H (2012) Mater Renew Sustain

Energy 1:1–12

54. Bowker M (2011) Green Chem 9:2235–2246

55. Pichat P, Herrmann JM, Disdier J, Mozzanega MN, Courbon H

(1984) Stud Surf Sci Catal 19:319–326

56. Nardelli MB, Rapcewicz K, Bernholc J (1997) Appl Phys Lett

21:3135–3137

57. Linic S, Christopher P, Ingram DB (2011) Nat Mater 12:911–921

Top Catal (2013) 56:1139–1151 1151

123