“This document is the Accepted Manuscript version of a Published Work that appeared in final form in Journal of
the American Chemical Society, copyright © American Chemical Society after peer review and technical editing
by the publisher. To access the final edited and published work see [insert ACS Articles on Request author- directed link to Published Work, see http://pubs.acs.org/doi/abs/10.1021/ja501489h

Efficient and Limiting Reactions in Aqueous Light-Induced Hydrogen
Evolution System using Molecular Catalysts and Quantum Dots
Carolina Gimbert-Suriñach,†,‡ Josep Albero,†, ‡ Thibaut Stoll,† Jérôme Fortage, § Marie-Noëlle
Collomb,§ Alain Deronzier,§ Emilio Palomares*,†,║ and Antoni Llobet*,†,┴
†

Institute of Chemical Research of Catalonia (ICIQ), Avinguda Països Catalans 16 – 43007 Tarragona, Spain.

§

Université Joseph Fourier Grenoble 1 / CNRS, Département de Chimie Moléculaire, UMR 5250, Institut de Chimie
Moléculaire de Grenoble, FR-CNRS-2607, Laboratoire de Chimie Inorganique Rédox, BP 53 – 38041 Grenoble cedex
9, France.
║

Institució Catalana de Recerca i Estudis Avançats (ICREA), Passeig Lluís Companys 23 – 08010 Barcelona, Spain.

┴

Departament de Química, Universitat Autònoma de Barcelona (UAB) – 08193 Cerdanyola del Vallès, Barcelona,
Spain.
ABSTRACT: Hydrogen produced from water and solar energy holds much promise for decreasing the fossil fuel dependence. It has recently been proven that the use of quantum dots as light harvesters in combination with catalysts is a valuable strategy to obtain photo-generated hydrogen. However, the light to hydrogen conversion efficiency of these systems
is reported to be lower than 40 %. The low conversion efficiency is mainly due to losses occurring at the different interfacial charge transfer reactions taking place in the multi-component system during illumination. In this work we have analyzed all the involved reactions in the hydrogen evolution catalysis of a model system composed of CdTe quantum dots, a
molecular cobalt catalyst and Vitamin C as sacrificial electron donor. The results demonstrate that the electron transfer
from the quantum dots to the catalyst occurs fast enough and efficiently (nanosecond timescale), while the back electron
transfer and catalysis are much slower (millisecond and microsecond timescales). Further improvements of the photodriven proton reduction should focus on the catalytic rate enhancement, which should be at least in the hundreds of
nanoseconds timescale.

Introduction
Efficient and cost-effective production of hydrogen gas from
water using sunlight as the ultimate energy source is one of the
major challenges for researchers in the 21st century. In recent
years, promising results have been reported using semiconductor materials for the photo-driven hydrogen production.1-10
When a semiconductor material is irradiated by light with
energy higher than its optical band-gap (BG), an electron of
the valence band (VB) is promoted to the conduction band
(CB) leaving a hole in the VB. Subsequently, the electrons and
holes migrate to the surface of the semiconductor where they
can be used for the reduction of protons and the corresponding
oxidation process, respectively (black arrows in Figure 1).
However, a major drawback of these systems is that the semiconductor photocatalytic activity is limited by undesired radiative and non-radiative recombination processes that occur in
the semiconductor crystal lattice competing directly with the
charge migration and the catalytic process (red arrows in Figure 1).
One strategy to improve the photocatalytic activity of semiconductors is to use a co-catalyst such as metallic platinum or

rhodium, often incorporated in the crystal lattice of the semiconductor as “dopants”. The co-catalyst not only lowers the
activation energy for proton reduction but also acts as an electron trap and favors charge separation. The overall result is an
increase of the driving force towards the formation of hydrogen, reducing the kinetic competition with the charge recombination processes.8,9 On the other hand, the so-called molecular approaches for hydrogen production have been attracting
much interest since the hydrogen production yields can be
substantially higher due to the versatility and tunability offered
by the coordinating ligands. These systems, analogously to the
photosynthetic systems in plants, algae and some bacteria,
separate the light absorption structural unit from the molecular
catalytic site.7,11
A recent elegant approach that combines both strategies mentioned before has been recently reported.3,10 They describe the
photogeneration of hydrogen in water by employing semiconductor CdSe nanocrystals as light harvesting materials,12-14
with the subsequent electron transfer to a nickel catalyst for
the concomitant reduction of protons to hydrogen. Further
examples of molecular approaches using quantum dots as

photosensitizers in pure water involve the use of hydrogenases
and their functional mimics.2,9,15-18 Yet, although the catalytic
activity of some of these examples is remarkably high (up to
hundreds of thousands of turnover numbers), the reported
quantum yields are still below 40 % even when using monochromatic light. In this context, a deeper knowledge of the
kinetics of both charge transfer and bond formation-breaking
are the keys to understand the limitations of photo-driven
hydrogen evolving systems based on molecular approaches.1921
Herein, we take advantage of a highly active system composed of quantum dots and a cobalt molecular catalyst to study
the kinetics involved in the overall photoinduced catalytic
process. We selected the components based on the following
requirements: (1) a light harvesting unit that absorbs in the
visible light region, (2) a molecular water reduction catalyst
with a well-defined structure, (3) the whole system has to
work in purely aqueous conditions.

Figure 1. Schematic representation of photo-driven hydrogen
generation at a semiconductor. Black Arrows: Light excitation,
charge migration (electrons = , holes = ⊕) and catalytic reactions. Red Arrows: photoluminescence and charge recombination.
CB = Conduction Band, VB = Valence Band, BG = Band-Gap.
S/Sox = substrate/oxidized substrate. Cat1 = catalyst for water
reduction. Cat2 = catalyst for substrate oxidation.

Water soluble CdTe quantum dots (CdTe QDs) are excellent
photosensitizing candidates as they have high extinction coefficients and offer the intrinsic advantages of semiconductorbased nanocrystals such as high photoluminescence quantum
yields and quantum confinement effects.12-14 As catalyst, we
selected the macrocyclic cobalt complex Co(III)-1 in Chart 1
because of its great stability and activity for hydrogen production in water.22,23 Our results show that the catalytic activity of
Co(III)-1 is superior to that of the cobaloxime-type catalyst
Co(III)-2, that has also been used as proton reduction catalyst
in water (Chart 1).24 An aqueous equimolar mixture of ascorbic acid/sodium ascorbate (H2A/NaHA) was chosen as both
buffer and sacrificial electron donor to trap the photogenerated
holes in the CdTe QDs.
CHART 1. Components of the catalytic system studied in this work.a

a

QD = quantum dot, Co(III)-1 and Co(III)-2 = hydrogen
evolving catalysts, H2A = ascorbic acid.
A detailed study of the thermodynamics and kinetics of the
system based on the components depicted in Chart 1, has
allowed us to build a complete energetics-kinetics scheme of
the proton reduction catalysis and identify what are the major
advantages and limitations in this kind of photo-driven hydrogen evolution processes.
Results
Electron transfer from CdTe QDs to the cobalt catalyst.
Upon light absorption by the CdTe QDs (abbreviated as “QD”
in the text), the excited electron in the conduction band (CB)
of the quantum dot is transferred to the cobalt catalyst so that
it can start the catalytic proton reduction sequence (Scheme 1).
This step is an oxidative quenching of the excited QD (QD*)
by the cobalt catalyst. Two consecutive electron transfer reactions are necessary to generate the Co(I) active species from
complexes Co(III)-1 or Co(III)-2.13,14,25,26 The reduction potentials for the couples Epc = Co(III)/Co(II) and E1/2 =
Co(II)/Co(I) were measured and are given in Scheme 1 (see
also Figure S1). We measured the energy of the CB of our
CdTe QDs and obtained an approximate value of –1.2 V vs
NHE (Figure S2) and therefore both electron transfer reactions
are thermodynamically favorable. In the case of complex 1,
the initial oxidation state of the cobalt at the beginning of the
catalysis is Co(II) since the Co(III)-1 species is reduced by
H2A (see UV-Vis spectra in Figure S3, supporting information).20 Nevertheless, the oxidized species Co(III)-1 can be
formed during hydrogen evolution turnover (Scheme 1).
Scheme 1. Relevant processes towards photo-driven
hydrogen evolutiona

Light Absorption

QD*

QD
QD* + Co(III)

1st ET

fast

QD+ + Co(II)

Co(III)-1 (≈ 10 ns) at the same concentration (compare black
and green traces in Figure 3a), indicating that the second
charge transfer reaction in Scheme 1 is kinetically more favored than the first one.

Co(III)-1 +0.35 V
E pc for Co(III)/Co(II)

Co(III)-2 -0.10 V

1st Back-ET

QD+ + Co(II)

2nd ET

QD* + Co(II)

slow
fast

QD + Co(III)
QD+ + Co(I)

Co(III)-1 -0.61 V
E 1/2 for Co(II)/Co(I)

2nd Back-ET

Hydrogen
Evolution

Co(III)-2 -0.69 V

QD+ + Co(I)

+

Co(I) + H

slow

Co(III)
H

QD + Co(II)
H2 + Co(III)

H+
1

/2 H2 + Co(II)

a

Black arrows: light absorption, electron transfer (ET) and
hydrogen evolution. Red arrows: back electron transfer reactions (back-ET). Reduction potentials Epc = Co(III)/Co(II) and
thermodynamic reduction potentials E1/2 = Co(II)/Co(I) are
given vs NHE.

The first electron transfer yield from the QDs CB to complex
Co(III)-1 or Co(III)-2 have been investigated by means of
steady-state fluorescence quenching experiments of binary
mixtures of the quantum dots and the molecular catalyst (Figure 2a and Figures S4-S5 in the supporting information). At
the same concentration, the electron transfer yield observed in
the case of complex Co(III)-1 (>90%) is significantly higher
than that of complex Co(III)-2 (50%) as illustrated in Figure
2a (compare black and red traces). Nevertheless, by increasing
the amount of Co(III)-2 considerably, it was possible to reach
electron transfer yields higher than 95 % (Figure S5). These
results indicate that the first electron transfer in Scheme 1 is
favorable for both complexes Co(III)-1 and Co(III)-2.
In order to perform the same study for the second electron
transfer, a reduced form of complex Co(III)-1, that is, Co(II)-1
was generated electrochemically and added to colloidal QDs
under nitrogen to avoid oxidation to Co(III). As depicted in
Figure 2 quenching of the QDs photoluminescence was higher
than 95% at a concentration of [Co(II)-1] = 15 µM, that is, 7.5
equivalents of Co(II) per mole of QDs. Unfortunately, it was
not possible to perform this study for the electro-generated
reduced form of complex Co(III)-2 due to the instability of the
Co(II)-2 species during long experiment times.
Furthermore, the QDs emission lifetime was analyzed in order
to evaluate the first and second electron transfer kinetics from
the QDs excited states to the different cobalt species (Figure
3). The first electron transfer is comparable and in the nanosecond (ns) timescale for both complexes Co(III)-1 and
Co(III)-2 but at high concentration ([Co(III)] > 10 µM) the
electron transfer becomes faster for Co(III)-1, while that of
Co(III)-2 remains unchanged (compare black and red traces in
Figure 3a). On the other hand, the second electron transfer to
the Co(II)-1 species resulted to be faster (< 5 ns) than that of

Figure 2. (a) Quenching emission yield as function of complex
concentration. (b) Steady state photoluminescence of colloidal
CdTe QDs upon addition of increasing amounts of Co(II)-1. The
samples were excited at λex = 405 nm. [CdTe QDs] = 2 µM.28

Figure 3. (a) Electron transfer lifetime as function of the cobalt
complex and H2A concentration. The electron transfer lifetime
was calculated following a reported methodology.27 (b) Time
resolved photoluminescence of colloidal CdTe QDs upon addition
of increasing amounts of Co(II)-1. The samples were excited at
λex = 405 nm and monitored at λem = 600 nm. [CdTe QDs] = 2
µM.28

Electron transfer from H2A to CdTe QDs. Two other charge
transfer reactions that are essential for the catalysis are (i) hole
transfer from the excited QD (QD*) to H2A, and (ii) reduction
of oxidized QD (QD+) by H2A, regenerating the initial QD. In
both cases, H2A injects an electron into the QDs valence band
(VB) which possesses a hole. The oxidation potential of H2A
is E1/2 = +0.36 V vs NHE11 and therefore the electron transfer
to the QDs VB (ca. 0.94 V vs NHE, Figure S2) is thermodynamically favored for both processes. Figure S6 in the supporting information, shows the photoluminescence quenching
of a QDs solution upon addition of H2A, which is due to hole
transfer from QD* to H2A. It reaches values of electron injection yield higher than 95 % at [H2A] = 300 µM, that is, 150
equivalent of H2A per mole of QDs. The kinetics of the process has been calculated and is comparable to that of the electron transfer from QD* to Co(II)-1 (compare blue and green
traces in Figure 3a). It is worth noticing that higher amounts of
H2A than Co(II)-1 are needed to completely quench the nanocrystals emission (150 and 10 equivalents, respectively).
However, the hydrogen evolution catalytic experiments are
performed using large excess of the hole scavenger (>1,500
equivalents, see below) and the kinetics of the hole transfer
from QD* to H2A is expected to be much faster in such conditions.

Back-electron transfer from cobalt catalyst to CdTe QDs.
All the results presented above show that electron transfer
from the QDs to the cobalt center and hole transfer from the
QDs to the H2A are both kinetically and thermodynamically
favored. However, undesirable back-electron transfer reactions
can take place at the same time and need to be taken into account when the whole system is assessed (red arrows in
Scheme 1). We have studied these recombination processes
using Laser Transient Absorption Spectroscopy (L-TAS).
The transient spectra of a QDs solution containing Co(III)-1 or
Co(III)-2, monitored at 10 microseconds (µs) after excitation
show a negative signal that matches the steady state absorption
of the QDs and therefore it can be assigned to the 1S ground
state bleaching of the nanocrystals (Figure S7). A positive
signal expanding from 620 nm to 900 nm also appears and is
similar to that observed in recent works for species derived
from analogous CdTe QDs.29,30 The transient absorption spectrum obtained by adding the reduced Co(II)-1 species to a QDs
solution did not present any significant change compared to
those obtained when Co(III)-1 or Co(III)-2 were added. On the
other hand, no signal was found at the studied time range (µsms) in the absence of cobalt complex, indicating that the transient signal arises from the photo-induced charge transfer
between the quantum dots and the cobalt species (Figure S7).
Taking into account that the reduced Co(II)-1 species derived
from the first charge transfer process doesn´t significantly
absorb beyond 650 nm (Figure S3) and also the fact that the
extinction coefficient of the QDs are at least ten times higher
than that of the cobalt species, we can attribute the positive
signal in the transient spectra to the oxidized QD (QD+).
The transient decays of the QD+ were recorded and fitted to a
power law function, equation (1), obtaining recombination
half-lifetimes τ1/2 = 9.6 and 5.1 milliseconds (ms) for Co(III)-1
and Co(III)-2, respectively (Figure S7). This result indicates
the formation of long-lived QD+ after first electron transfer. In
other words, the back electron transfer from the generated
Co(II) species to the oxidized quantum dots is slow, that is, in
the millisecond timescale.

f (t) = A·t τ

(1)

Analogous transient decay experiments using a mixture of
QDs and Co(II)-1 presented long half-lifetime (τ1/2= 4.5 ms)
for the QD+ species and showed identical recombination dynamics alike Co(III)-1 (compare green and black traces in
Figure 4a).
We have already pointed out the importance of the electron
transfer from the sacrificial electron donor H2A to the QDs
VB. Transient decay experiments were done with mixtures
containing QDs, Co(II)-1 and H2A, before and after addition
of H2A (Figure 4b). It was found that the decay of the new
formed transient signal becomes one order of magnitude faster
after addition of H2A, that is, τ1/2 =0.45 ms vs τ1/2= 4.5 ms,
indicative for a second process taking place in the mixture and
therefore supporting fast electron injection to the QDs VB by
H2A.

1 and in the dark and we did not observe the formation of
significant amount of hydrogen. Interestingly, analogous
catalytic tests using complex Co(III)-2 did not produce
significant amount of hydrogen as shown in Figure S9 in the
supporting information. The low stability of cobaloxime-type
catalysts under acidic aqueous conditions is well known and
could explain why the catalysis using Co(III)-2 was
unsuccessful.
The quantum yield (QY) as defined in equation (2) was calculated to be 10 % for the best experiment (see Figure 5 and
Figure S10 in the supporting information). This value is similar to those obtained with related photocatalytic systems that
use cobalt molecular catalysts in water 32-34 or quantum dots
with other metallic catalysts.3,11

Figure 4. Transient absorption decays of colloidal CdTe QDs (a)
upon addition of 30 µM of Co(III)-1 or Co(II)-1. (b) upon addition of 30 µM of Co(II)-1 and 250 µM of H2A. The samples were
excited at λex=430 nm and the decays were monitored at λprobe =
675 nm. [CdTe QDs] = 2 µM.28 The orange, red and purple lines
in the plots indicate the experimental points’ trend that was obtained by fitting the data to the power law function indicated in
equation (1).

Hydrogen evolution catalysis. The rate of the catalytic proton
reduction process using Co(III)-1 has been measured electrochemically by using the foot of the wave method developed by
Savéant31 obtaining a value in the microsecond timescale,
which is in agreement with other reported values in the literature (Figure S8).26 These kinetics values are compatible to all
the kinetic processes described previously and therefore encouraged us to carry out bulk experiments to evaluate the
potential of the system.
Figure 5 shows the profile of hydrogen generation upon time
using the hybrid CdTe QDs/Co(III)-1 system. The rate of the
catalysis depends on the concentration of the H2A/NaHA
buffer and under optimum conditions it reaches 650 TONCo in
1.5 h with a maximum turnover frequency (TOFCo,max) of 9
min–1 (green trace, Figure 5). This TOFCo,max value is one of
the highest reported using quantum dots as photosensitizers
and a molecular catalyst in water that has not been generated
in situ, together with two recent reports that use CdSe QDs
and diiron or nickel catalysts.9,10 Other related systems with
high catalytic activity are active only in organic solvents19 or
they use inorganic salts as precursors of the catalyst.3,8 Blank
experiments were performed in the absence of catalyst Co(III)-

Figure 5. Photocatalytic hydrogen production experiments were
performed using an aqueous mixture of [CdTe QDs] = 5.9 × 10–5
M, [Co(III)-1] = 7.5 × 10–5 M and [H2A/NaHA] = 0.1 M (▲), 0.6
M (■) or 1.2 M (●), pH = 4.1, volume = 1.5 mL, under 1 sun
irradiation (λ > 400nm). Inset: Expanded graph (0-1.6h). Turnover Number (TONCo) = mol H2/mol cobalt, average of two different replicates (within 5-10 % error, see also Figure S9). Values of
quantum yield (QY) for each experiment are given at maximum
conversion time and were calculated as QY = [2 × (molecules of
H2) ] / [photons absorbed × ∆t × Area], see also Figure S10.

(2)
Discussion
From the combination of experiments reported in the previous
section a general picture emerges of how the system “QDsMolecular Catalyst” manages to carry out the photo-induced
reduction of water to hydrogen. The detailed description of all
these experiments allowed the generation of a schematic
summary of the kinetics and thermodynamics of all the significant reactions involved in this complex system, which is
offered in Scheme 2.
The first requirement is the presence of an efficient light harvesting material such as the CdTe QDs, that are known to
interact with light very fast, that is, in the femtosecond time
scale (1, Scheme 2), generating a powerful reducing agent; its
CB is situated at ca. E = –1.2 V vs NHE. Such a low redox

potential will enable a large thermodynamic driving force with
regards to the reduction of most relevant molecular hydrogen
evolving catalysts.11,35,36 However, the latter reaction will only
proceed if the excited QD (QD*) is sufficiently long lived.
Under our system conditions we found this excited state to be
of ca. 20 nanoseconds (6, Scheme 2). Therefore for a successful system, it is imperative that the reactions coupled to this
excited state are not only thermodynamically favored but also
faster than 20 ns.
We have chosen the cobalt complexes 1 and 2, in oxidation
state +3 as proton reduction catalysts. To be able to generate
hydrogen they need to be reduced all the way to Co(I) via a
sequential two one-electron transfer processes as indicated in
Scheme 1, where their relevant redox potentials are also indicated. In all cases the thermodynamic driving force of the
electron transfers to generate the corresponding Co(I) catalytically active species is huge; e.g. ∆E = –1.55 V and –0.59 V for
the Co(III/II) and the Co(II/I) couples respectively for 1. To
evaluate the kinetics of these processes we carried out photoluminesce quenching experiments giving life times of 10 and 5
nanoseconds for the reactions indicated in arrows 2 and 3 in
Scheme 2, respectively.
Scheme 2. Energetics and kinetics scheme for the
photocatalytic hydrogen evolving system based on
CdTe QDs, cobalt molecular catalyst Co(III)-1 and
ascorbic acid/sodium ascorbate (H2A/NaHA).a

a

A = oxidized ascorbate. Black arrows indicate favorable
processes towards efficient and fast catalysis while red arrows illustrate the undesired photoluminescence and backelectron reactions. In parenthesis, the kinetics timescales are
given in milliseconds (ms), microseconds (μs) or nanoseconds (ns).

It is worth noticing that these experiments do not differentiate
between hole transfer or electron transfer. Cyclic voltammetry
experiments of Co(III)-1 and QDs show that the oxidation of
Co(III)-1 by the QDs is not thermodynamically possible and
therefore, the first process can unequivocally be assigned to
electron transfer from the QDs CB to the cobalt catalyst (Figures S1 and S2 in the supporting information). On the other
hand, assigning the second charge transfer from excited QD
(QD*) to Co(II) was not trivial as it could well be due to hole
transfer from the QDs VB to Co(II) to give the Co(III) species
and the hypothetical reduced QD (QD–). If this was the case,
the newly formed Co(III) species would be reduced back to
Co(II) by the QD– within a few nanoseconds as just commented above. L-TAS experiments discussed in detailed below

showed the formation of a long lived transient species (milliseconds). If hole transfer from QD* to Co(II) took place, this
would be faster than our L-TAS instrument response (microseconds), and no L-TAS signal would be found. Therefore, the
second charge transfer process can be attributed to the second
electron transfer from QD* to Co(II) to give Co(I) (3 in
Scheme 2). Interestingly, the reaction between QD* and
Co(II), with lower driving force, has faster kinetics than the
reaction between QD* and Co(III) (5 vs 10 ns, respectively).
This can be attributed to the electronic structure of Co(III) that
has a low spin (dπ)6 electronic configuration and therefore the
coming electron has to be accommodated in a (dπ) orbital
generating a high spin d7 complex. An additional reason why
the first reduction of Co(III) is slower might be due to the
huge driving force of this reaction (∆E = –1.55 V). In a recent
work, the dependence of the photo-catalytic hydrogen evolution rate on the ∆E using CdSe nanocrystals of different sizes
has been studied. It was shown that in the range of ∆E from –
0.5 V to –0.8 V the system works under normal Marcus theory
regime.37 In the present case, ∆E for the first electron transfer
is twice the lower limit of the range that they study and therefore it is plausible to think that we are in the Marcus inverted
region, where the first electron transfer with higher driving
force is expected to have a slower rate.
The large thermodynamic driving force for the desired forward
reaction is expected to generate slower back electron transfer
reactions (back-ET) to the QD*, however, recombination
reactions could still potentially occur with the oxidized QD+.
This is a crucial point especially if the back-ET to QD+ is fast,
in which case the cobalt catalyst will not be long lived enough
to be able to interact with protons and generate hydrogen.
L-TAS experiments were used to calculate the back-ET kinetics indicated by arrows 7 and 8 in Scheme 2, and we obtained
values of 9.6 and 4.5 milliseconds, respectively. The L-TAS
experiment was repeated in the presence of H2A showing a
decrease of the lifetime of the transient species by one order of
magnitude. This result is very important because it shows that
the electron transfer from H2A to the QDs VB is more efficient
than that of the Co(I) complex to QD+ and thus prevents a
non-productive Co-centered reaction. Furthermore, we could
also show that H2A is capable of reductively quench the QD*
by replenishing its VB in the nanosecond time scale thus
avoiding the presence of the oxidizing QD+ species (5,
Scheme 2). It is remarkable that these recombination processes
from reduced cobalt catalyst to oxidized quantum dot are 5-6
orders of magnitude slower (milliseconds) than the measured
photo-induced electron transfer dynamics from QD to Co
catalyst (nanoseconds).
Electrochemical experiments in water, under comparable
working conditions, show that the lifetime of the catalytic
cycle that generates hydrogen occurs in the microsecond time
scale, indicating that this system has a good match with the
reactions discussed above (4, Scheme 2). Indeed upon shining
light to our QDs-Co catalyst system in the presence of
H2A/NaHA generates an impressive amount of H2 with a
TOFCo,max of 9 min–1 that is among the highest ever reported in
water. The ratio of hole scavenger/QD used for the hydrogen
evolution experiments is one order of magnitude higher than
that used for the photophysical experiments (>1,500 vs 150,
respectively) and therefore the electron transfer from the H2A
to the QDs VB is faster under catalytic conditions, hindering

the formation of QD+ and favoring the proton reduction process.
The overall picture we have managed to generate for the photocatalytically induced reduction of water shown in Scheme 2,
also allowed us to identify the limitations and the potential
improvements of the present system. Thus, the main efforts
need to be directed towards the minimization of unproductive
reactions shown in red, in Scheme 2.
Ideally it would be desirable to be able to lower the rate of the
QD* recombination reaction (6, Scheme 2) so that more light
efficient systems could be generated. Alternatively, catalysts
that would react faster with QD* (2 and 3, Scheme 2) would
also increase the overall performance of the catalysis.
Another important improvement to the system would be the
use of complexes capable of catalytically cycling faster than
the microsecond time scale, avoiding unwanted competing
recombination reactions such as the one shown in arrows 7
and 8 in Scheme 2.
As just discussed, the molecular catalyst plays one of the
crucial roles in these systems since its interactions with the
QD and with protons determine the success of the photocatalytic reaction. Indeed, it is surprising to see that when replacing catalyst 1 by 2 under the same conditions, the system
practically did not yield any hydrogen, due to either catalyst
decomposition in water22,23,32 or slow electron transfer processes.21 The macrocyclic nature of the ligand in 1 not only
makes the catalyst more robust but also could contribute to the
stabilization of the Co(I) species that is formed during catalyst
turnover. In this context, molecular catalysts have the potential
to dramatically improve the performance of this type of systems given the large versatility of ligands that can be used to
tune the electronic and structural properties at the metal center.11,35,36
Conclusion
In conclusion, we have managed to fully identify for the first
time the thermodynamics and kinetics involved in a hybrid
system made of CdTe QDs material and a molecular cobalt
catalyst, that coupled together are an efficient light induced
proton reduction system in water. This in turn allowed us to
identify which reactions need to be improved in order to come
up with even better systems. Furthermore, the molecular nature of the catalysts provides a wide avenue of potential improvements given the large versatility of potential ligands that
can be used.

ASSOCIATED CONTENT
Supporting Information. Additional kinetic and spectroscopic data as well as experimental details of the synthesis of
CdTe QDs, Co(III)-1, Co(II)-1 and Co(III)-2, electrochemical
procedures and photo-chemical hydrogen evolution experiments. This material is available free of charge via the Internet at http://pubs.acs.org.

AUTHOR INFORMATION
Corresponding Author
allobet@iciq.es
epalomares@iciq.es

Author Contributions
‡These authors contributed equally.

ACKNOWLEDGMENTS
This work was supported by “La Caixa” Foundation, European Research Council starting Grant ERCstg-POLYDOT and
the LABEX program ARCANE of Grenoble (Grant N° ANR-11LABX-0003-01) (H2Photocat). EP also thanks the ICIQ,
ICREA, GenCat (Project 2009 SRD 207) and MICINN (CTQ2007-60746-BQU) for funding. CGS is grateful to GenCat for
a “Beatriu de Pinós” post-doctoral grant.

REFERENCES

(1)
Chen, X.; Shen, S.; Guo, L.; Mao, S. S. Chem. Rev. 2010, 110, 6503.
(2)
Wang, F.; Wang, W.-G.; Wang, X.-J.; Wang, H.-Y.; Tung, C.-H.; Wu, L.-Z. Angew. Chem.
Int. Ed. 2011, 50, 3193.
(3)
Han, Z.; Qiu, F.; Eisenberg, R.; Holland, P. L.; Krauss, T. D. Science 2012, 338, 1321.
(4)
Woolerton, T. W.; Sheard, S.; Chaudhary, Y. S.; Armstrong, F. A. Energy Environ. Sci.
2012, 5, 7470.
(5)
Tran, P. D.; Wong, L. H.; Barber, J.; Loo, J. S. C. Energy Environ. Sci. 2012, 5, 5902.
(6)
Zhu, H.; Song, N.; Lv, H.; Hill, C. L.; Lian, T. J. Am. Chem. Soc. 2012, 134, 11701.
(7)
Wen, F.; Li, C. Acc. Chem. Res. 2013, 46, 2355.
(8)
Li, Z.-J.; Li, X.-B.; Wang, J.-J.; Yu, S.; Li, C.-B.; Tung, C.-H.; Wu, L.-Z. Energy Environ.
Sci. 2013, 6, 465.
(9)
Li, C.-B.; Li, Z.-J.; Yu, S.; Wang, G.-X.; Wang, F.; Meng, Q.-Y.; Chen, B.; Feng, K.;
Tung, C.-H.; Wu, L.-Z. Energy Environ. Sci. 2013, 6, 2597.
(10) Das, A.; Han, Z. J.; Haghighi, M. G.; Eisenberg, R. PNAS 2013, 110, 16716.
(11) Eckenhoff, W. T.; Eisenberg, R. Dalton Trans. 2012, 41, 13004.
(12) Zhang, H.; Zhou, Z.; Yang, B.; Gao, M. J. Phys. Chem. B 2002, 107, 8.
(13) Shi, L.; Hernandez, B.; Selke, M. J. Am. Chem. Soc. 2006, 128, 6278.
(14) Tsay, J. M.; Trzoss, M.; Shi, L.; Kong, X.; Selke, M.; Jung, M. E.; Weiss, S. J. Am.
Chem. Soc. 2007, 129, 6865.
(15) Brown, K. A.; Dayal, S.; Ai, X.; Rumbles, G.; King, P. W. J. Am. Chem. Soc. 2010, 132,
9672.
(16) Brown, K. A.; Wilker, M. B.; Boehm, M.; Dukovic, G.; King, P. W. J. Am. Chem. Soc.
2012, 134, 5627.
(17) Greene, B. L.; Joseph, C. A.; Maroney, M. J.; Dyer, R. B. J. Am. Chem. Soc. 2012, 134,
11108.
(18) Nann, T.; Ibrahim, S. K.; Woi, P.-M.; Xu, S.; Ziegler, J.; Pickett, C. J. Angew. Chem. Int.
Ed. 2010, 49, 1574.
(19) Huang, J.; Mulfort, K. L.; Du, P.; Chen, L. X. J. Am. Chem. Soc. 2012, 134, 16472.
(20) Tseng, H.-W.; Wilker, M. B.; Damrauer, N. H.; Dukovic, G. J. Am. Chem. Soc. 2013, 135,
3383.
(21) Reynal, A.; Lakadamyali, F.; Gross, M. A.; Reisner, E.; Durrant, J. R. Energy Environ.
Sci. 2013,6, 3291.
(22) Varma, S.; Castillo, C. E.; Stoll, T.; Fortage, J.; Blackman, A. G.; Molton, F.; Deronzier,
A.; Collomb, M.-N. Phys. Chem. Chem. Phys. 2013, 15, 17544.
(23) McCrory, C. C. L.; Uyeda, C.; Peters, J. C. J. Am. Chem. Soc. 2012, 134, 3164.
(24) Lakadamyali, F.; Reynal, A.; Kato, M.; Durrant, J. R.; Reisner, E. Chem.-Eur. J. 2012,
18, 15464.
(25) Artero, V.; Chavarot-Kerlidou, M.; Fontecave, M. Angew. Chem. Int. Ed. 2011, 50, 7238.
(26) Leung, C.-F.; Chen, Y.-Z.; Yu, H.-Q.; Yiu, S.-M.; Ko, C.-C.; Lau, T.-C. Int. J. Hydrog.
Energy 2011, 36, 11640.
(27) Koops, S. E.; Durrant, J. R. Inorg. Chim. Acta 2008, 361, 663.
(28) Yu, W. W.; Qu, L.; Guo, W.; Peng, X. Chem. Mat. 2003, 15, 2854.
(29) Kaniyankandy, S.; Rawalekar, S.; Verma, S.; Palit, D. K.; Ghosh, H. N. Phys. Chem.
Chem. Phys. 2010, 12, 4210.
(30) Kaniyankandy, S.; Rawalekar, S.; Ghosh, H. N. J. Phys. Chem. C 2012, 116, 16271.
(31) Costentin, C.; Drouet, S.; Robert, M.; Savéant, J.-M. J. Am. Chem. Soc. 2012, 134,
11235.

(32) Guttentag, M.; Rodenberg, A.; Kopelent, R.; Probst, B.; Buchwalder, C.; Brandstätter,
M.; Hamm, P.; Alberto, R. Eur. J. Inorg. Chem. 2012, 2012, 59.
(33) Nippe, M.; Khnayzer, R. S.; Panetier, J. A.; Zee, D. Z.; Olaiya, B. S.; Head-Gordon, M.;
Chang, C. J.; Castellano, F. N.; Long, J. R. Chem. Sci. 2013, 4, 3934.
(34) Sun, Y.; Sun, J.; Long, J. R.; Yang, P.; Chang, C. J. Chem. Sci. 2013, 4, 118.
(35) Wang, M.; Chen, L.; Sun, L. Energy Environ. Sci. 2012, 5, 6763.
(36) Thoi, V. S.; Sun, Y.; Long, J. R.; Chang, C. J. Chem. Soc. Rev. 2013, 42, 2388.
(37) Zhao, J.; Holmes, M. A.; Osterloh, F. E. ACS Nano 2013, 7, 4316.

Table of Contents (TOC)

10