Published on Web 02/23/2005

Charge Separation and Catalytic Activity of Ag@TiO2

Core-Shell Composite Clusters under UV-Irradiation
Tsutomu Hirakawa†,§ and Prashant V. Kamat*,†,‡
Contribution from the Radiation Laboratory and Department of Chemical and Biomolecular
Engineering, UniVersity of Notre Dame, Notre Dame, Indiana 46556-0579
Received November 23, 2004; E-mail: pkamat@nd.edu

Abstract: Photocatalytic properties of Ag@TiO2 composite clusters have been investigated using steady
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
Downloaded via UNIV OF LOUISIANA AT LAFAYETTE on January 28, 2019 at 20:43:18 (UTC).

state and laser pulse excitations. Photoexcitation of TiO2 shell results in accumulation of the electrons in
the Ag core as evidenced from the shift in the surface plasmon band from 460 to 420 nm. The stored
electrons are discharged when an electron acceptor such as O2, thionine, or C60 is introduced into the
system. Charge equilibration with redox couple such as C60/C60•- shows the ability of these core shell
structures to carry out photocatalytic reduction reactions. The charge separation, charge storage, and
interfacial charge-transfer steps that follow excitation of the TiO2 shell are discussed.

Introduction effective, result in exposing both metal to reactants and the

Noble metal nanoparticles such as Ag and Au in the nanosize surrounding medium. Corrosion or dissolution of the noble metal
domain (<10 nm) exhibit unusual catalytic, electric, and optical particles during the operation of a photocatalytic reaction is
properties.1-12 For example, Au nanoparticles of 3-8 nm likely to limit the use of noble metal such as Ag and Au.27-29
diameter have been shown to tune the catalytic properties of A better synthetic design can significantly improve the catalytic
TiO2.13-15 In an earlier study, we have shown that metal performance of oxide-metal composite. Silica has been widely
nanoparticles deposited on TiO2 nanostructures undergo Fermi employed as a shell to protect Ag and Au nanoparticles and
level equilibration following the UV-excitation and enhance the stabilize them against chemical corrosion. Preparation and
efficiency of charge-transfer process.16,17 In most of the catalytic characterization of Au@SiO2 and Ag@SiO2 core-shell cluster
studies metal nanoparticles are dispersed on an oxide surface have been reported for metal coreoxide shell clusters.30-35
(See for example, refs 18-26). Such a catalyst structure, though Significant advances have been made in recent years to design
metal core-semiconductor shell clusters.30-32,36-39 Yet, the
†Radiation Laboratory.
‡
efforts to utilize these core-shell structures as photocatalysts
Department of Chemical and Biomolecular Engineering.
§ Present Address: National Institute of Advanced Industrial Science and in the light energy conversion systems (e.g., photoelectrochemi-
Technology (AIST), Tsukuba, Japan. cal cells, hydrogen production etc.) are limited. In this context,
(1) Henglein, A. Chem. ReV. 1989, 89, 1861. it is important to elucidate the influence of the metal core on
(2) Mulvaney, P. Langmuir 1996, 12, 788.
(3) Pileni, M. P. New J. Chem. 1998, 693. the photocatalytic properties of outer TiO2 shell (Scheme 1).
(4) Templeton, A. C.; Wuelfing, W. P.; Murray, R. W. Acc. Chem. Res. 2000, The obvious questions that need to be addressed include, how
33, 27.
(5) El-Sayed, M. A. Acc. Chem. Res. 2001, 34, 257. does the photoinduced charge separation in TiO2 is influenced
(6) Toshima, N.; Yonezawa, T. New J. Chem. 1998, 22, 1179.
(7) Brust, M.; Kiely, C. Colloid Surf. A 2002, 202, 175. (24) Averitt, R. D.; Westcott, S. L.; Oldenburg, S. J.; Lee, T. R.; Halas, N. J.
(8) Kelly, K. L.; Coronado, E.; Zhao, L. L.; Schatz, G. C. J. Phys. Chem. B Langmuir 1998, 14, 5396.
2003, 107, 668. (25) Cozzoli, P. D.; Fanizza, E.; Comparelli, R.; Curri, M. L.; Agostiano, A.;
(9) Kamat, P. V. J. Phys. Chem. B 2002, 106, 7729. Laub, D. J. Phys. Chem. B 2004, 108, 9623.
(10) Adams, D.; Brus, L.; Chidsey, C. E. D.; Creager, S.; Cruetz, C.; Kagan, C. (26) Kamat, P. V.; Flumiani, M.; Dawson, A. Colloids Surf., A. 2002, 202, 269.
R.; Kamat, P. V.; Lieberman, M.; Lindsay, S.; Marcus, R. A.; Metzger, R. (27) Subramanian, V.; Wolf, E.; Kamat, P. V. J. Phys. Chem. B 2001, 105,
M.; Michel-Beyerle, M. E.; Miller, J. R.; Newton, M. D.; Rolison, D. R.; 11439.
Sankey, O.; Schanze, K. S.; Yardley, J.; Zhu, X. J. Phys. Chem. B 2003, (28) Subramanian, V.; Wolf, E. E.; Kamat, P. V. Langmuir 2003, 19, 469.
107, 6668. (29) Lahiri, D.; Subramanian, V.; T.Shibata; Wolf, E. E.; Bunker, B. A.; Kamat,
(11) Kamat, P. V. Pure Appl. Chem. 2002, 74, 1693. P. V. J. Appl. Phys. 2003, 93, 2575.
(12) George Thomas, K.; Kamat, P. V. Acc. Chem. Res. 2003, 36, 888. (30) Liz-Marzan, L. M.; Mulvaney, P. J. Phys. Chem. B 2003, 107, 7312.
(13) Haruta, M. Catal. Today 1997, 36, 153. (31) Ung, T.; Liz-Marzan, L. M.; Mulvaney, P. Langmuir 1998, 14, 3740.
(14) Yang, Z. X.; Wu, R. Q.; Goodman, D. W. Phys. ReV. B 2000, 6, 14066. (32) Caruso, F.; Spasova, M.; Saigueirino-Maceira, V.; Liz-Marzan, L. M. AdV.
(15) Valden, M.; Lai, X.; Goodman, D. W. Science 1998, 281, 1647. Mater. 2001, 13, 1090.
(16) Jakob, M.; Levanon, H.; Kamat, P. V. Nano Lett. 2003, 3, 353. (33) Mulvaney, P.; Liz-Marzan, L. M.; Giersig, M.; Ung, T. J. Mater. Chem.
(17) Subramanian, V.; Wolf, E. E.; Kamat, P. V. J. Am. Chem. Soc. 2004, 126, 2000, 10, 1259.
4943. (34) Ung, T.; Liz-Marzan, L. M.; Mulvaney, P. J. Phys. Chem. B 1999, 103,
(18) Herrmann, J. M.; Disdier, J.; Pichat, P. J. Phys. Chem. 1986, 90, 6028. 6770.
(19) Anpo, M.; Chiba, K.; Tomonari, M.; Coluccia, S.; Che, M.; Fox, M. A. (35) Dick, K.; Dhanasekaran, T.; Zhang, Z.; Meisel, D. J. Am. Chem. Soc. 2002,
Bull. Chem. Soc. Jpn. 1991, 64, 543. 124, 2312.
(20) Baba, R.; Konda, R.; Fujishima, A.; Honda, K. Chem. Lett. 1986, 10. (36) Zhong, C. J.; Maye, M. M. AdV. Mater. 2001, 13, 1507.
(21) Tada, H.; Teranishi, K.; Ito, S.; Kobayashi, H.; Kitagawa, S. Langmuir (37) Hardikar, V.; Matijevic, E. J. Colloid Interface Sci 2000, 221, 133.
2000, 16, 6077. (38) Oldenburg, S. J.; Averitt, R. D.; Westcott, S. L.; Halas, N. J. Chem. Phys.
(22) Ohko, Y.; Tatsuma, T.; Fujishima, A. J. Phys. Chem. B 2001, 105, 10016. Lett. 1998, 288, 243.
(23) Dawson, A.; Kamat, P. V. J. Phys. Chem. B 2001, 105, 960. (39) Oldfield, G.; Ung, T.; Mulvaney, P. AdV. Mater. 2000, 12, 1519.
3928 9 J. AM. CHEM. SOC. 2005, 127, 3928-3934 10.1021/ja042925a CCC: $30.25 © 2005 American Chemical Society
Ag@TiO2 Core ARTICLES

Scheme 1 Scheme 2

by the metal core? How does the charge equilibration occur

between the metal and semiconductor following the band gap
excitation? Can these core-shell structures be superior catalysts? heating was stopped and the suspension was stirred until it cooled to
room temperature.
To address these questions we have now prepared Ag@TiO2,
The procedure described above was also employed for preparing
Ag@SiO2, and TiO2 colloids in DMF/ethanol medium and their silica capped Ag particles (viz., Ag@SiO2). Active silica was added to
behavior under UV-excitation are compared. In a preliminary the reaction mixture instead of TTAEIP. The details of the method
communication we reported the ability of the Ag@TiO2 clusters can be found elsewhere.31 The cluster suspension of Ag@TiO2 and
to store electrons under UV-irradiation and discharge them on Ag@SiO2 was centrifuged and resuspended in ethanol solution. The
demand in the dark.40 The factors that control the charge procedure was repeated at least 3-times to minimize the content of water
separation and photocatalytic properties of core-shell nano- and DMF in the suspension.
structures are presented in this paper. A better understanding TiO2 Colloids. TiO2 colloidal suspension was prepared by the
of the charge transfer properties is likely to pave the way to procedure similar to the one used for Ag@TiO2 colloids except the
develop improved photocatalysts for light energy conversion. step of AgNO3 addition. The slow hydrolysis of TTEAIP produced
transparent sol of TiO2.
Experimental Section Laser Flash Photolysis. Experiment of nano-second laser flash
photolysis was performed with 308-nm laser pulses from Lambda
Materials and Measurements. Titanium-(triethanolaminato) iso- Physik excimer laser system (Laser pulse width is 10 ns, intensity was
propoxide (N((CH2)2O)3TiOCH(CH3)2) TTEAIP and dimethyl form- ∼10 mJ/pulse). Unless otherwise specified, all the experiments were
amide, DMF were purchased from Aldrich. AgNO3 (Fisher ACS grade) performed under N2 purging condition.
was used as received. Active silica (Na2O(SiO2)3-5, 27 wt % SiO2) Steady-state photolysis experiments were conducted by photolyzing
was obtained from Aldrich. Care should be taken to handle toxic N2-purged solution with UV-visible light (250 W xenon lamp). A
solvents such as DMF and hygroscopic materials such as TTEAIP. CuSO4 filter was introduced in the path of the light beam to cutoff
The absorption spectra were recorded using UV-vis spectro- light below the wavelengths of 300 nm.
photometer (Varian CARY 50 Bio UV-vis Spectrophotometer and
SHIMADZU UV-3101 PC UV-vis-NIR Scanning Spectrophotom- Results and Discussion
eter). Transmission electron microscopy (TEM) was carried out with
Synthesis of Ag@TiO2 Cluster Suspension. Metal core-
JEOL TEM-100SX Electron Microscope and HITACHI H-600 Electron
Microscope. Particle size and shape were analyzed with image
oxide shell structures were prepared by one pot synthesis that
magnification. involved reduction of metal ions and hydrolysis of Titanium-
Ag@TiO2 and Ag@SiO2 Colloids. For synthesizing Ag@TiO2 shell (triethanolaminato)isopropoxide in dimethylformamide (DMF).41
cluster, we modified the procedure of reduction of metal ions in The solvent, DMF plays an important role of reducing the Ag+
dimethylformamide.41 Desired concentration of TTEAIP (8.3 mM, ions first, followed by the slow hydrolysis of TTEAIP to form
unless otherwise specified) was prepared in 2-propanol. Two mL of a shell around the metal core. Increasing the amount of DMF
15 mM AgNO3 solution was mixed with 18 mL of TTEAIP solution. increases the primary step of reduction rate of Ag+ ion to Ag0.
Ten mL of DMF was then added into TTEAIP-Ag solution. The It is important that the reduction rate of Ag+ ion is greater than
concentration of Ag+ and TTEAIP in the reaction mixture was 1 mM the rate of formation of TiO2 shell. After several initial attempts
and 5 mM, respectively. This volume ratio of DMF and i-PrOH was we have found that the experimental conditions that employ
optimized by carrying out several batch preparations. When the amount
33% DMF in i-PrOH yields stable suspension of core-shell
of DMF was too little or when i-PrOH was excluded, aggregation of
clusters is observed. The volume ratio of DMF and i-PrOH hence is
particles. As Ag+ ions are reduced by DMF to form small metal
an important factor in the preparation of the Ag@TiO2 clusters. The particles, they quickly interact with the amine groups of
solution was stirred first for 15 min at room temperature and then TTEAIP. The condensation polymerization of TTEAIP slowly
refluxed with continued stirring. With continued heating of the solution, progresses on the surface of Ag particles to yield TiO2 shell.
the color slowly changed from colorless to light brown. After 90 min, Formation of core-shell cluster is illustrated in Scheme 2.
the color of the suspension turned to dark brown. At this point, the The TEM images of two different sets of cluster preparations
are presented in Figure 1. The concentrations of TTEAIP in
(40) Hirakawa, T.; Kamat, P. V. Langmuir 2004, 20, 5645. these two sets of synthesis were 0.03 and 5 mM respectively
(41) Pastoriza-Santos, I.; Koktysh, D. S.; Mamedov, A. A.; Giersig, M.; Kotov,
N. A.; Liz-Marzan, L. M. Langmuir 2000, 16, 2731. while maintaining the concentration of AgNO3 at 1 mM. In both

J. AM. CHEM. SOC. 9 VOL. 127, NO. 11, 2005 3929

ARTICLES Hirakawa and Kamat

Figure 1. Transmission electron micrographs of Ag@TiO2 colloids which

were prepared using the composition of (A) 5 mM TiO2 and 1 mM Ag and
(B) 5 mM TiO2 and 1 mM Ag. (C) Absorption spectra of colloidal (a)
Ag@TiO2 (b) Ag@SiO2 and (c) TiO2 suspension in ethanol.

cases the Ag core was similar with particle diameter of 3-4

nm. The TEM image shows a majority of dark images of Ag
core in this size regime. All these particles have a thin capping
of TiO2 shell of thickness in the range of 1.5-3 nm. The core-
shell structure is clearly evident in the few larger fully grown
particles (30-65 nm) that coexist with smaller size particles
(Figure 1A).
Capping of TiO2 shell on the Ag core was confirmed by
checking the stability in an acidic solution (HNO3). The Ag
cluster, stabilized by citric acid, is readily dissolved in acidic
solution (pH ) 2). Ag@TiO2 on the other hand is quite stable
in HNO3 solution even when the TiO2 shell was thin. If the
formation of TiO2 clusters in DMF was independent such that
both clusters are formed separately or in the form of a TiO2/Ag
sandwich structure, we would have observed dissolution of silver
clusters. The stability test in acidic solution asserts the argument
that the TiO2 shell on the Ag core is uniform and provides the
protection against acid induced corrosion.
One approach to increase the shell thickness is to increase
the concentration of the precursor of TiO2 (TTAEIP in the
present case). Similar approach has been widely used for
preparing silica capped metal nanoparticles.31 Thus, for Ag@TiO2
colloids prepared using higher TTAEIP concentrations we would
have expected to see a thicker TiO2 shell. However, we observe
the formation of independent TiO2 clusters with smaller size
Ag@TiO2 clusters of core diameter of <2 nm (Figure 1B).
Unlike silica capped Ag clusters, we cannot grow thicker TiO2
shell by increasing the concentration of TTAEIP.42 The hy-
Figure 2. Absorption spectra recorded following UV-irradiation of various
drolysis of TTEAIP thus, competes with Ag+ reduction and
colloidal suspensions in ethanol. The direction of the arrow indicates the
yields TiO2 clusters with and without Ag core at high TTEAIP spectral shift observed during the UV-irradiation. (A) Ag@TiO2 colloidal
concentration. These results show the necessity for optimizing suspension. The difference absorption spectrum corresponding to the
the concentration of TTEAIP for achieving uniform capping of absorption changes is shown in the inset. (B) Ag@SiO2 colloidal suspension.
The inset shows TEM image of Ag@SiO2. (C) TiO2 colloidal suspension.
Ag nanocore. The inset shows the change in absorbance at 690 nm with UV-irradiation.
The particle concentration was estimated by assuming uniform
distribution of Ag core particle and an average particle size of strong absorption in the UV). The plasmon absorption band of
5.12 and 3.65 nm as obtained from Figure 1A and B. The the small Ag particles prepared using borohydride reduction is
average number of particles in these two preparations were around 380 nm. The surface plasmon absorption band of
estimated to be 2.4 × 1014 and 6.3 × 1014 per liter, respectively. Ag@SiO2 and Ag@TiO2 particles is significantly red-shifted
Figure 2 shows the absorption spectra of Ag@TiO2, Ag@SiO2 and the maximum is seen at 410 and 480 nm respectively. It is
and TiO2 colloids before and after UV-irradiation. Ag@SiO2 evident that the red shift in the plasmon absorption seen in the
and Ag@TiO2 exhibit strong absorption in the visible. This core shell particle is dependent on the type of the oxide shell.
visible absorption arises from the surface plasmon band of Ag As shown earlier,2,30,41,43 the high dielectric constant of the TiO2
core and it is strongly influenced by the oxide shell. (Where as shell causes a red shift in the plasmon absorption of the silver
SiO2 itself has no absorption in the UV and visible, TiO2 exhibits core.
(42) Faster capping of Ag nuclei by the amine moieties is likely to suppress the (43) Kreibig, U.; Vollmer, M. Optical Properties of Metal Clusters; Springer:
growth of Ag nanoparticles at higher TTAEIP concentration Berlin, 1995.

3930 J. AM. CHEM. SOC. 9 VOL. 127, NO. 11, 2005

Ag@TiO2 Core ARTICLES

The peak shift is brought about by the refractive index of conclude that the changes in the absorption seen in Ag@TiO2
the surrounding medium. In core-shell structure, the plasmon arise from the UV excitation of the shell and not the silver core.
peak position of the metal core can be related using the The photoactive TiO2 shell thus, plays an important role of
expression 144 absorbing incident photons and injecting electrons into the silver
core. The processes that lead to storing of electrons in the Ag
λ ) {λP[∞ + 2n2EtOH + 2g(n2TiO2 - n2EtOH)/3]}1/2 (1) core are summarized below (reactions 3-5).

where n is refractive index of the surrounding medium, ∞ is TiO2 f TiO2(e + h) (3)

high frequency of the core metal, g is the volume fraction of
shell layer, λ is the estimated peak position of metal core, and TiO2(h) + ethanol f products (4)
λP is bulk plasma wavelength as represented in eq 2
TiO2(e) + Ag f TiO2 + Ag(e) (5)
λP ) [4 π2 c2 meff 0/N e2]1/2 (2)
The photoactivity of TiO2 colloids was separately checked by
where the meff is effective mass of the free electron of the metal, carrying out UV-irradiation of TiO2 colloids in deaerated
and N is electron density of metal core. ethanol. Under UV excitation, TiO2 undergoes charge separation
For Ag cluster with no medium effect we expect λP to be (reaction 1) followed by charge recombination and interfacial
around 136.3 nm. When the Ag cluster was dispersed in water charge-transfer processes. As the photogenerated holes are
or ethanol the λp is observed around 390 nm. Since nTiO2 (2.5) scavenged by ethanol, the electrons accumulate within the TiO2
for shell is much higher than nEtOH (1.359), the plasmon particles. The electron storage in TiO2 particles is marked by
absorption shows a large red shift. With increasing thickness the blue coloration with characteristic broad absorption in the
of TiO2 shell the g factor in expression (1) becomes close to red region (Figure 2C). As shown earlier,17,48 this absorption
unity and the red shift of plasmon absorption attains a maximum arises from electron trapping at Ti4+ sites. The trapped electrons
value for the Ag@TiO2 cluster. On the basis of the expression are long-lived in N2 purged suspensions and are readily extracted
1, we expect the value of λ to be around 463 nm. In the present when needed, e.g., by transfer of electrons to acceptor molecules
experiment, we observe rather broad plasmon absorption with such as thionine or C60. Since metal particles such as silver and
a maximum around 480 nm and this red-shifted maximum is gold with a favorable Fermi level (EF )0.4 V) are good electron
close to the value expected from the theoretical prediction. acceptors, we expect a facile electron transfer from excited TiO2.
Factors such as scattering effects and adsorbed chemical species This property of charge equilibration between semiconductor
are likely to contribute to the small discrepancy between and metal nanoparticles has been independently confirmed by
estimated and experimentally observed plasmon absorption peak. the addition of metal colloids to preirradiated TiO2 colloids in
Steady-State Photolysis. Figure 2A shows the changes in ethanol.16,17 The disappearance of the blue coloration (or
the absorption spectrum following the UV-irradiation of Ag@TiO2 absorbance at 600 nm) in this study has been used to determine
colloids suspended in deaerated ethanol. Before subjecting the the quantitative transfer of electrons to metal particles. The
Ag@TiO2 colloids in ethanol to UV irradiation, the plasmon spectra recorded in Figure 2A do not indicate absorption in the
absorption peak is seen at 480 nm. As these colloids are red-IR region similar to the increase observed in Figure 2C.
subjected to UV irradiation we observe a blue shift in the The lack of blue coloration in UV-irradiated Ag@TiO2 colloids
plasmon absorption band. After irradiating with UV-light for indicates that the electrons fail to accumulate in the TiO2 shell,
∼30 min, the absorption shift attains a plateau with a surface instead they are transferred to the silver core. The only spectral
plasmon absorption peak at 420 nm. The shift of 60 nm in the change we observe in this system is the shift in the plasmon
plasmon absorption reflects increased electron density in the absorption arising from the electron accumulation in the metal
Ag core during photoirradiation. Since TiO2 undergoes charge core.
separation under UV-irradiation, the photogenerated electrons When the UV-excitation of Ag@TiO2 is carried out in the
are transferred to Ag nano-core as the two systems undergo presence of an electron acceptor such as thionine dye or oxygen,
charge equilibration. The transfer of electrons from the excited we do not observe any shift in the plasmon absorption of silver
semiconductor to the metal is an important aspect that dictates core. As the photogenerated electrons are scavenged by the
the overall energetics of the composite and hence the efficiency acceptor molecules, electrons fail to accumulate in the Ag core.
of photocatalytic reduction process.17,45-47 These observations parallel the results reported for Ag/ZnO47
To confirm the participation of TiO2 shell in transferring and Au@SnO239 nanoclusters and ascertain the argument that
electrons to Ag core, we compared the photoresponse with that the shift in the plasmon absorption arises from electron storage
of Ag@SiO2 particles. When the UV-irradiation experiments in the Ag core.
were repeated with the Ag@SiO2 colloids in deaerated ethanol, We checked the reproducibility of charging and discharging
we fail to see any changes in the plasmon absorption (Figure of Ag@TiO2 system by repeated cycles of UV-irradiation of
2B). The plasmon absorption band with a maximum at 410 nm deaerated suspension followed by exposure to air. Figure 3 show
remained unperturbed during the UV-irradiation. By comparing the reproducibility of plasmon absorption peak response to the
the experimental results in Figure 2, parts A and B, we can UV-irradiation and air exposure in dark. The plasmon absorption
(44) Templeton, A. C.; Pietron, J. J.; Murray, R. W.; Mulvaney, P. J. Phys.
band shifts from 470 to 420 nm during 1 min UV irradiation of
Chem. B 2000, 104, 564 deaerated Ag@TiO2 suspension. The plasmon absorption regains
(45) Shanghavi, B.; Kamat, P. V. J. Phys. Chem. B 1997, 101, 7675. the original spectral features when the stored electrons are
(46) Subramanian, V.; Wolf, E. E.; Kamat, P. V. J. Phys. Chem. B 2003, 107,
7479.
(47) Wood, A.; Giersig, M.; Mulvaney, P. J. Phys. Chem. B 2001, 105, 8810. (48) Kamat, P. V.; Bedja, I.; Hotchandani, S. J. Phys. Chem. 1994, 98, 9137.

J. AM. CHEM. SOC. 9 VOL. 127, NO. 11, 2005 3931

ARTICLES Hirakawa and Kamat

Figure 3. Response of plasmon absorbance peak to electron storage

following the UV-irradiation of deaerated Ag@TiO2 colloidal suspension
in ethanol and dark discharge in air.
discharged in dark by exposing to air. We can repeat the
photoinduced charging and dark discharge cycles repeatedly and
reproduce the plasmon absorption response to stored electrons
(Figure 3). The suspension was deaerated for 20 min before
UV irradiation of each charging cycle. No such absorption
changes in the plasmon absorption are seen for Ag@SiO2 system
under similar experimental conditions.
A similar but less pronounced effect was noted for Au capped
with SnO2 particles.39 The electrons were injected chemically Figure 4. (A) Controlled discharge of stored electrons from Ag@TiO2
using a reductant, NaBH4. The electrons injected into SnO2 are suspension (pre-UV-irradiated under deaerated conditions) using known
transferred to Au core until the two systems attain equilibration. amounts of thionine dye. Inset shows the plasmon absorption peak position
with increasing concentration of thionine. (B) Dependence of plasmon
However, the dark discharge of electrons takes several hours absorption peak and number of stored electrons on the time of UV-
to take into effect. This is attributed to the fact that the irradiation. (The number of stored electrons was calculated by titrating with
conduction band of SnO2 is around 0 V versus NHE and is thionine dye)
slightly more positive than oxygen reduction potential. On the
other hand TiO2 conduction band is around -0.5 V vs NHE at Known amounts of concentrated thionine solution (degassed)
pH 7 and is energetically capable of transferring electrons to was injected in small increments into the UV-irradiated Ag@TiO2
O2. It should be noted that both the semiconductor shell and suspension. The absorption spectrum was recorded after each
the metal core undergo Fermi level equilibration to attain an addition of thionine. Since the reduced dye does not have any
energy level close to the conduction band of the semiconductor. absorption in the visible, we only observe a shift in the plasmon
In the presence of a redox couple the composite system further absorption band as the electrons are titrated using thionine. The
attains equilibration by transferring excess charges into the process of electron transfer to the dye molecules continues until
acceptor molecules. By comparing the discharge response of most of the stored electrons are discharged. The presence of
Ag@TiO2 and Au@SnO2 to oxygen, we can conclude that any unreduced thionine is marked by the appearance of 600
Ag@TiO2 clusters are energetically superior catalyst for pro- nm absorption band. The endpoint of titration is thus attained
moting reduction process. when we observe the appearance of the 600 nm band. Figure
Estimation of Electrons Stored in Ag Core. Since the 4A shows the changes in the absorption spectrum following
electrons stored in the Ag@TIO2 colloids can equilibrate with the addition of thionine. The plasmon absorption shift and stored
the redox couple in solution, it is possible to carry out a redox electrons were estimated at different irradiation times by titration
titration and obtain quantitative information on the stored of the stored electrons with thionine. The dependence of the
electrons. The dye thionine with absorption at 600 nm ( ) plasmon shift and the number of stored electrons versus the
60 000 M-1 cm-1) is a good electron acceptor that initially UV-irradiation time is shown in Figure 4B. Saturation in electron
produces a one electron reduction product, semithionine (Reac- storage is seen as we extend the UV-irradiation for a longer
tion 6). The semithionine (TH•-) immediately disproportionates time.
to produce a two electron reduction product, leuco dye (TH2-), As discussed in the earlier section, we observe a red-shift in
with no absorption in the visible (reaction 7). the plasmon absorption band with increased addition of thionine
to UV-irradiated Ag@SiO2 suspension. After attaining the
2TH + Ag@TiO2(2e-) f 2TH•- + Ag@TiO2 (6) endpoint a small peak corresponding to thionine at 600 nm is
seen. The plasmon shift corresponding to the discharge of the
2TH•- f TH + TH2- (7) stored electrons in the Ag@TiO2 cluster can thus be related to
The disappearance of one thionine molecule thus, represents the concentration of thionine added to discharge the electrons
transfer of two electrons. The reduction of thionine thus serves (inset, Figure 4A). From the slope of this linear plot, we estimate
as the basis for titrating electrons stored in the Ag@TiO2 that 0.42 µM electrons cause 1 nm shift in the plasmon
clusters. absorption in a 2.1 × 1017 particles/liter Ag@TiO2 suspension.
3932 J. AM. CHEM. SOC. 9 VOL. 127, NO. 11, 2005
Ag@TiO2 Core ARTICLES

Scheme 3

spectrum recorded in steady-state photolysis as shown in the

inset of Figure 2A. This observation indicates that the absorption
change observed in steady-state photolysis is a spontaneous
process that follows the excitation of the TiO2 shell. The
transient absorption-time profile (Figure 5B) shows that most
of the charge transfer from excited TiO2 to Ag is completed
within the laser pulse duration of few nanoseconds. A slow
growth that contributes about 10% of the signal arises from the
charge equilibration process which extends up to 150 µs (Figure
5C). The absorption in the red region shows the residual
Figure 5. (A) Difference absorption spectra of deaerated colloidal electrons trapped within the TiO2 nanoparticles. These are excess
Ag@TiO2 (O) and TiO2 (]) suspension recorded following 308 nm laser trapped electrons that survive within the TiO2 shell following
pulse excitation. (B) Representative absorption-time profiles at 420 nm.
(C) 540 nm and (D) 700 nm (experimental conditions are the same as in charge equilibration. Figure 5D compares the transient absorp-
Figure 5A). tion profiles at 700 nm. The maximum absorbance at 700 nm
is lower for Ag@TiO2 than the TiO2 colloids as a fraction of
On the basis of the net shift observed in the plasmon band, we the electrons are transferred to Ag core immediately following
expect a maximum storage of about 66 electrons per Ag@TiO2 the laser pulse excitation.
core@shell particle. The ability of Ag@TiO2 to store large Photocatalytic Activity of Ag@TiO2 Particles. We com-
number of electrons during UV-irradiation shows the importance pared the photocatalytic activity of the Ag@TiO2 with that of
of such composite structures in light energy conversion by TiO2 colloids by carrying out reduction of C60 following 308
storing electrons during photoirradiation and delivering these nm laser pulse excitation (Scheme 3). In our earlier studies, we
electrons on demand. The capacity of electron storage is have shown that C60 is an excellent probe to study the interfacial
determined by the size of the metal core and its ability to electron transfer in colloidal semiconductor systems.53 The
undergo charge equilibration with the TiO2 shell. Once this formation of C60•- with characteristic absorption in the IR region
maximum storage limit is attained, electron-hole recombination (1075 nm) can be conveniently used to obtain quantitative
in the TiO2 shell dominates. information on the electron-transfer yield. Figure 6(A) shows
Transient Absorption Studies using Laser Flash Photoly- the difference absorbance spectrum obtained after 308 nm laser
sis. The photoinduced charge-transfer events were probed using pulse excitation of Ag@TiO2 colloids in the presence of C60,
nanosecond laser flash photolysis. The transient absorption The absorption maximum at 1075 nm indicates formation of
spectra recorded following 308 nm laser pulse excitation of C60 anion, C60•-, as the excited Ag@TiO2 transfer electrons to
deaerated colloidal TiO2 and Ag@TiO2 suspension are shown C60. As can be seen from the inset of Figure 6A the C60-
in Figure 5 (A). The spectra were recorded 1 µs after laser pulse formation is completed in 100µs. The slow formation of C60•-
excitation. The TiO2 colloids exhibit broad absorption in the is the indication of the time required to attain charge equilibra-
400-800 nm region corresponding to the trapping of electrons tion between C60 and Ag@TiO2 system (Since C60 also absorbs
at Ti4+ sites. The Ti3+ centers formed as a result of electron at the excitation wavelength, we observe triplet C60 formation
trapping have been characterized in earlier studies.49-52 The (absorption maximum at 740 nm) along with the C60•- forma-
Ag@TiO2 clusters exhibit strong bleaching in the plasmon tion. The excited C60 is practically inert under present experi-
absorption region in addition to the absorption in the red. The mental conditions as it does not interact with the TiO2 surface.)
absorption band at 420 nm and bleaching at 540 nm is the result The maximum absorbance recorded 100 µs after laser pulse
of shift in the plasmon band following the electron storage. It excitation was used to determine the relative yield of electron-
is interesting to note that the spectral feature of the transient transfer process in TiO2 and Ag@TiO2 colloids. Figure 6B
spectrum (spectrum b in Figure 5A closely match the difference shows the photocatalytic reduction yield as a function of the
(49) Gopidas, K. R.; Bohorquez, M.; Kamat, P. V. J. Phys. Chem. 1990, 94,
concentration of added C60 for the two systems. The electron
6435. transfer yield increased initially with increasing concentration
(50) Bahnemann, D.; Henglein, A.; Lilie, J.; Spanhel, L. J. Phys. Chem. 1984,
88, 709.
of C60 and reaching a plateau at C60 concentration >30 µM.
(51) Howe, R. F.; Graetzel, M. J. Phys. Chem. 1985, 89, 4495.
(52) Rajh, T.; Ostafin, A. E.; Micic, O. I.; Tiede, D. M.; Thurnauer, M. C. J. (53) Kamat, P. V.; Gevaert, M.; Vinodgopal, K. J. Phys. Chem. B 1997, 101,
Phys. Chem. 1996, 100, 4538. 4422.

J. AM. CHEM. SOC. 9 VOL. 127, NO. 11, 2005 3933

ARTICLES Hirakawa and Kamat

In our earlier study of Au-TiO2 sandwich structure (i.e., two

particles in contact with each other in a coupled geometry), we
observed a higher catalytic activity compared to TiO2. The
increased photocatalytic reduction was attributed to the shift of
Fermi level to more negative potentials. Since C60/C60•- couple
equilibrates with the stored charges within the composite, the
yield of C60•- serves as a measure of the apparent Fermi-level.
The lower yield of C60•- observed in the present experiments
suggests that the apparent Fermi-level of the Ag@TiO2 com-
posite is lower than TiO2 alone. The charge equilibration
between the metal core and semiconductor shell, thus causes
the apparent Fermi-level less negative than neat TiO2 cluster.
Although metal core-semiconductor shell structures are quite
efficient for storing photogenerated electrons, there ability to
catalyze a reduction process is limited. The results presented
here highlight the importance of designing semiconductor-metal
composite nanostructures for light energy harvesting applica-
tions.

Conclusions

We have demonstrated the photoinduced charging and dark

discharging of electrons in a silver core-semiconductor shell
structure. The shift in surface plasmon band serves as a measure
Figure 6. (A) Difference absorption spectra recorded 0.1 µs (O) and 15
µs (b) after 308 nm laser pulse excitation of deaerated Ag@TiO2 colloidal
to determine the number of electrons stored in the metal core.
suspension in ethanol containing 32 µM C60. Inset shows the growth of The charge equilibration between the metal and semiconductor
C60•- as monitored from the absorption at 1075 nm. (B) Yield of C60•- (as plays an important role in dictating the overall energetics of
monitored from the absorbance at 1075 nm, )16000M-1cm-1) with the composite. These metal core-semiconductor shell composite
increasing concentration of C60 present in colloidal Ag@TiO2 suspension
(deaerated with N2, Ex., 308 nm). clusters are photocatalytically active and are useful to promote
light induced electron-transfer reactions. Exploring the catalytic
The electron transfer from excited TiO2 and Ag@TiO2 to C60 activity of such composite structures could pave the way for
proceeds until the two systems attain redox equilibration. designing novel light harvesting systems.
Although we expected to see a higher catalytic activity for the
Acknowledgment. The research described here was supported
Ag@TiO2 clusters, the experimental results show an opposite
by the Office of Basic Energy Science of the U.S. Department
trend. The photocatalytic reduction efficiency of Ag@TiO2 is
of Energy. This is contribution number NDRL 4574 from the
lower than that of TiO2. These results suggest that part of the
Notre Dame Radiation Laboratory.
photogenerated electrons reside within the Ag@TiO2 composite
as it attains equilibration with C60•-. JA042925A

3934 J. AM. CHEM. SOC. 9 VOL. 127, NO. 11, 2005