Photocatalytic carboxylation of organic substrates with carbon dioxide at zinc sulfide with deposited ruthenium nanoparticles

Article abstract of DOI:10.1002/cplu.201300438

Photocatalytic carboxylation of acetylacetone with carbon dioxide has been performed by using ZnS-based photocatalysts. The formation of two isomeric carboxylic acids, as proven by IR and ¹³C NMR spectroscopy, was observed. The reaction yield was enhanced after deposition of ruthenium nanoparticles on ZnS. The reaction encompasses ruthenium-mediated one-electron reduction of CO₂ to CO₂^.- with electrons from the conduction band of ZnS and one-hole oxidation of acetylacetone to the relevant radical. Coupling of photogenerated radicals leads to the formation of carboxylic acids. Generation of CO₂^.- has been confirmed by spin-trapping EPR measurements. The process described herein may find applications for the solar-light-driven green synthesis of C_n+1 carboxylic compounds from C_n substrates by utilising carbon dioxide. Dropping acid: Zinc sulfide with deposited ruthenium(0) nanoparticles photocatalyzes the carboxylation of acetylacetone with carbon dioxide as a substrate (see picture). The reaction proceeds through coupling of photogenerated R^. and CO₂^.- radicals. Copyright

Full text of DOI:10.1002/cplu.201300438

CHEMPLUSCHEM
FULL PAPERS
DOI: 10.1002/cplu.201300438
Photocatalytic Carboxylation of Organic Substrates with
Carbon Dioxide at Zinc Sulfide with Deposited Ruthenium
Nanoparticles
Tomasz Baran,[a, b] Angela Dibenedetto,[b, c] Michele Aresta,[c] Krzysztof Kruczała,[a] and
Wojciech Macyk*[a]
Dedicated to Professor Graz˙yna Stochel on the occasion of her 60th birthday
Photocatalytic carboxylation of acetylacetone with carbon di-
oxide has been performed by using ZnS-based photocatalysts.
The formation of two isomeric carboxylic acids, as proven by
IR and 13C NMR spectroscopy, was observed. The reaction yield
was enhanced after deposition of ruthenium nanoparticles on
ZnS. The reaction encompasses ruthenium-mediated one-elec-
tion band of ZnS and one-hole oxidation of acetylacetone to
the relevant radical. Coupling of photogenerated radicals leads
Cꢀ
to the formation of carboxylic acids. Generation of CO2 has
been confirmed by spin-trapping EPR measurements. The pro-
cess described herein may find applications for the solar-light-
driven green synthesis of Cn+1 carboxylic compounds from Cn
substrates by utilising carbon dioxide.
Cꢀ
tron reduction of CO2 to CO2 with electrons from the conduc-
Introduction
The conversion of carbon dioxide is a key research area. New
sustainable synthetic methodologies, both reducing the emis-
sions of CO2 into the atmosphere and saving fossil carbon re-
sources, should be developed. The latter goal is relevant to
step from a “linear to a cyclic C economy”.[1–3] Several ap-
proaches to CO2 utilisation have been developed, including
thermal, catalytic, electrochemical, photoelectrochemical, pho-
tochemical and biochemical technologies.[4,5] The photocatalyt-
ic conversion of CO2 to added-value chemicals or fuels seems
to be a particularly attractive perspective. Both homo- and het-
erogeneous photocatalysts have been used.[4] Among the
latter, inorganic semiconductors, such as TiO2, ZnO, ZnS, CdS,
GaP, SiC, ZrO2 and WO3, can be mentioned. So far, the yields of
CO2 reduction to C1 and C2 products (e.g., formate, methanol,
methane and oxalate) are not high and the selectivity is barely
controlled; nevertheless, such attempts show that the reaction
is feasible, but catalysts need to be further improved.[5–15] On
the other hand, photochemical reactions can be used to imple-
ment a new route to carboxylates, which are products with
added value with respect to both the starting organic sub-
strate and CO2. The direct functionalisation of hydrocarbons or
other organic substrates would make use of cheap and abun-
dant solar energy within a single-step carboxylation reaction
and avoid technologies based on the oxidation of aliphatic
chains (toluene to benzoic acid) or partial oxidative opening of
aromatic rings (naphthalene to benzene-o-dicarboxylic acid).
The introduction of a carboxylic group into an organic sub-
strate is not an easy process and nowadays it is realised by
using complex and harsh procedures characterised by quite
a low carbon utilisation fraction and the production of large
amounts of waste.
The photoreduction of CO2 would give a variety of energy-
rich products. Direct reduction proceeds through an uphill
one-electron transfer with the formation of the carbon dioxide
Cꢀ
radical anion. The redox potential of the CO2/CO2 couple in
water amounts to ꢀ1.9 V versus a standard hydrogen elec-
trode (SHE), which makes the process strongly unfavourable
(in dry organic solvents the situation is even worse because
the potential is ꢀ2.1 V vs. SHE).[16] This extremely low potential
is related to the large reorganisation energy between the
linear molecule and the bent radical anion. Multi-electron re-
duction processes of carbon dioxide lead to more stable prod-
ucts, and therefore, these reactions are thermodynamically pre-
ferred, but they are difficult to control because several prod-
ucts are often concurrently formed [Eqs. (1)–(5)]:[1,16]
[a] T. Baran, Dr. K. Kruczała, Prof. W. Macyk
Faculty of Chemistry, Jagiellonian University
Ingardena 3, 30-060 Krakꢀw (Poland)
E-mail: macyk@chemia.uj.edu.pl
[b] T. Baran, Prof. A. Dibenedetto
Department of Chemistry, University of Bari
Orabona 4, 70125 Bari (Italy)
CO2 þ 2 Hþ þ 2 eꢀ ! HCOOH
CO2 þ 2 Hþ þ 2 eꢀ ! CO þ H2O
CO2 þ 4 Hþ þ 4 eꢀ ! HCHO þ H2O
Eꢁ ¼ ꢀ0:61 V
Eꢁ ¼ ꢀ0:53 V
Eꢁ ¼ ꢀ0:48 V
ð1Þ
ð2Þ
ð3Þ
[c] Prof. A. Dibenedetto, Prof. M. Aresta
Interuniversity Consortium Chemical Reactivity and Catalysis (CIRCC)
via Celso Ulpiani 27, 70126 Bari (Italy)
This article is part of the “Early Career Series”. To view the complete
series, visit: http://chempluschem.org/earlycareer.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2014, 79, 708 – 715 708

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
Results and Discussion
CO2 þ 6 Hþ þ 6 eꢀ ! CH3OH þ H2O
CO2 þ 8 Hþ þ 8 eꢀ ! CH4 þ 2 H2O
Eꢁ ¼ ꢀ0:38 V
Eꢁ ¼ ꢀ0:24 V
ð4Þ
ð5Þ
Zinc sulfide materials, ZnS-A and ZnS-B, were prepared accord-
ing to published procedures by precipitation with Na2S or thio-
urea (ZnS-A and ZnS-B, respectively) from a solution of ZnSO4.
The obtained fine-crystalline powders are light grey. Such ma-
terials were selected because of known activity in carbon diox-
ide reduction, photoregeneration of NADH (NADH=nicotina-
mide adenine dinucleotide), and dehydrodimerisation of 2,5-di-
hydrofuran or cyclohexene.[11,34,35] The materials with deposited
metallic ruthenium are dark grey.
One photon can induce the transfer of one electron in a pho-
tochemical reaction, according to the Stark–Einstein law. There-
fore, the probability of a one-electron process is much higher
than that of a multiple-electron process. The energetic stability
of the carbon dioxide molecule (DG8=ꢀ394 kJmolꢀ1), howev-
er, constitutes a strong limitation to CO2 reduction and makes
this reaction highly endoergic.[17]
Diffuse reflectance spectra of the prepared materials, con-
verted by the Kubelka–Munk function, are presented in
Various organic reactions, realised in homo- and heterogene-
ous photocatalytic systems, are known.[18–21] They include cis–
trans isomerisation, substitution and cycloaddition process-
es.[22–24] The photocatalytic carboxylation of organic com-
pounds under mild conditions extends this list.[25] The basic
concept of photocatalytic carboxylation is CꢀC bond formation
Cꢀ
Figure 1. From the [F(R )E]2 versus E plots, which is the trans-
1
formation valid for direct semiconductors, the band gap ener-
gies have been determined and are summarised in Table 1.
upon the coupling of two radicals: an organic one and CO2
.
This mechanism is characteristic of type B photocatalysis,
which involves the coupling of two radicals, in contrast to
type A photocatalysis, which is characterised by the formation
of various reduction and oxidation products.[24,26] The coupling
of photogenerated radicals is an exceptional process because
a huge majority of photocatalytic CO2 reduction processes,
leading to simple C1 products, are type A reactions.[27] Kisch
et al. reported the coupling of organic radicals photogenerated
at CdS, resulting in CꢀC and CꢀN bond formation.[28] A differ-
ent mechanism of CO2 photocoupling was reported by Yanagi-
da and co-workers, who suggested that cadmium sulfide
would promote the intermolecular addition of CO2 to a C=O
bond.[29] Electrons from the excited photocatalyst were able to
reduce both carbon dioxide and an organic substrate (e.g., ar-
omatic ketones and benzyl halides), while triethylamine was
consumed as a sacrificial electron donor. In such a process,
a carboxylic acid formed as a result of radical–radical anion
coupling.
Figure 1. Transformed diffuse reflectance spectra of the synthesised photo-
catalysts.
Table 1. Band gap energies and absorption onsets determined for the
synthesised photocatalysts.
Photocatalyst
Band gap energy
Aonset
Cꢀ
Generation of CO2 requires a particularly strong reducer.
(eV; ꢃ0.02 eV)
[nm]
The commonly used TiO2 photocatalyst does not enable this
process because the bottom of its conduction band is located
at about ꢀ0.3 to 0.6 V versus SHE, depending on the sample
and pH conditions.[28] In contrast, zinc sulfide, which is a wide
band gap (3.6 eV) semiconductor, offers particularly strong re-
duction potentials, as low as ꢀ1.8 to ꢀ2.0 V versus SHE, and
its activity in photocatalytic water splitting and carbon dioxide
reduction has been demonstrated.[11,28,30–32] The low electro-
chemical stability of ZnS can be overcome in the presence of
electron donors that can efficiently pick up photogenerated
holes to prevent photocorrosion of the material.[33]
ZnS-A
0.5% Ru@ZnS-A
ZnS-B
3.47
3.57
3.64
3.65
357
347
341
339
0.5% Ru@ZnS-B
The absorption onset of the ZnS-B material is localised at l
ꢂ350 nm and is characteristic of microcrystalline zinc sulfide.
ZnS-A has a smaller band gap energy than ZnS-B. The pres-
ence of ruthenium at the surface does not influence the elec-
tronic properties significantly. Defect states, grain boundaries
and charged impurities lead to the formation of additional
electronic states within the band gap. The presence of such
states reflects in a tail in the absorption spectrum (known as
the Urbach tail), which extends to low photon energies. Appa-
rently, ZnS-A contains more defects than ZnS-B.
Herein, we describe a photocatalytic system based on zinc
sulfide that enables the carboxylation of acetylacetone (acac)
Cꢀ
to form two isomeric carboxylic acids as a result of CO2 cou-
pling to organic radicals formed upon oxidation by photogen-
erated holes.
The emission spectra of aqueous suspensions of ZnS-A and
0.5% Ru@ZnS-A measured at room temperature are shown in
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2014, 79, 708 – 715 709

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
Figure 2. Emission spectra of ZnS-A materials. Excitation at l=350 nm.
Figure 4. XRD patterns of the photocatalyst: ZnS-A, 1% Ru@ZnS-A and ZnS-
B. The XRD pattern of the standard sphalerite structure of ZnS is also shown.
Figure 2. The spectrum of ZnS-A shows four emission maxima
at l=394, 455, 481 and 526 nm. The energy of fundamental
band gap emission was determined to be 3.15 eV (Figure 2;
l=394 nm). Two emission bands at l=455 and 481 nm (2.72
and 2.58 eV) of ZnS were reported elsewhere; however, they
were attributed differently: Wang et al. explained their origin
by the existence of surface states,[36] whereas Zhu et al. attrib-
uted such bands to defect-related emissions.[37] The emission
above l=500 nm originates from some specific impurities. No
new emission band was observed after ruthenium deposition,
but the fundamental emission was significantly suppressed.
This observation may reflect efficient electron transfer from the
conduction band to ruthenium particles. The intensities of
other bands are also smaller with respect to unmodified ZnS.
Figure 3A shows the SEM image of the as-prepared 0.5%
Ru@ZnS-A photocatalyst. The material is composed of irregular
agglomerates (0.2–2.0 mm) of smaller nanoparticles. The chemi-
cal composition of the material has been examined by using
EDX. The results presented in Figure 3D prove a homogeneous
distribution of ruthenium within the material.
blende structure (sphalerite) of both ZnS-A and ZnS-B (JCPDS
no. 05-0566). ZnS-B is composed of bigger crystals than those
of nanoparticles of ZnS-A. No ruthenium-related signals could
be detected in the diffractograms.
EDX measurements confirmed the deposition of ruthenium
on the ZnS surface (Figure 5). In the case of 0.1% Ru@ZnS-A,
0.5% Ru@ZnS-A and 1.5% Ru@ZnS-A, a signal assigned to
ruthenium can be observed (19.27 keV).[38] Raman spectroscopy
excluded the formation of ruthenium(IV) oxide because bands
at n˜ =515 and 626 cmꢀ1, which are characteristic for this mate-
rial (RuꢀO stretching modes), were not found.[39]
All materials are characterised by relatively high specific sur-
face areas. ZnS-A offers a much higher surface area (105 and
99 m2 gꢀ1 for ZnS-A and 0.5% Ru@ZnS-A, respectively) than
those of the ZnS-B materials (56 and 55 m2 gꢀ1 for neat and
Ru-modified samples, respectively). Because the photocatalytic
reaction occurs at the semiconductor surface, an increase in
the specific surface area should improve the yield of the pho-
tocatalytic reaction. Interestingly, the specific surface areas of
ZnS loaded with ruthenium is slightly reduced relative to the
corresponding materials. This observation is the consequence
A detailed analysis of the composition of the photocatalysts
was performed by means of XRD, EDX and Raman spectrosco-
py. XRD analysis of the photocatalyst (Figure 4) proves the zinc
Figure 3. A) SEM image of 0.5% Ru@ZnS-A. Energy-dispersive X-ray spectros-
Figure 5. EDX results of three photocatalysts: 0.1% Ru@ZnS-A (black), 0.5%
copy (EDX) mapping of S (B), Zn (C) and Ru (D).
Ru@ZnS-A (red) and 1.5% Ru@ZnS-A (green).
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2014, 79, 708 – 715 710

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
of the larger aggregates formed in the case of ruthenium-
loaded ZnS-A and ZnS-B, as confirmed by dynamic light scat-
tering measurements: lꢂ350 and 460 nm for 0.5% Ru@ZnS-A
and 0.5% Ru@ZnS-B, respectively, compared with l=300 and
350 nm for ZnS-A and ZnS-B, respectively.
Temperature-programmed desorption (TPD) tests revealed
good adsorption of carbon dioxide on the catalyst, particularly
after ruthenium deposition. For ZnS-B, a cumulative volume of
adsorbed CO2 was estimated to be 0.09 mLgcat.ꢀ1, whereas for
0.5% Ru@ZnS-B the measured value amounted to
0.42 mLgcat.ꢀ1, although the specific surface areas of rutheni-
um-modified materials are slightly lower than those of the cor-
responding unmodified photocatalysts. Good adsorption of
CO2 is essential for efficient photoinduced interfacial electron
transfer between the catalyst and the substrate. A significant
influence of ruthenium on CO2 adsorption indicates good bind-
ing of carbon dioxide to metal nanoparticles.[40] Better CO2 ad-
sorption at 0.2% Ru@ZnS-B was also confirmed in the desorp-
tion tests (Figure 6). Desorption of carbon dioxide was signifi-
cantly amplified by UV light. A similar photodesorption effect
was previously reported for RuO2·xH2O/TiO2 materials.[41]
Figure 7. Tautomeric forms of acac and the expected carboxylation prod-
ucts.
which do not compete with acac for carboxylation reactions,
can be also considered.
Irradiation performed for 1–6 h in the presence of 0.5%
Ru@ZnS-A resulted in the formation of two isomeric carboxylic
acids, as demonstrated by IR and 13C NMR spectroscopic analy-
ses. Figure 8 presents the FTIR spectra of the reaction mixture
Figure 8. Photocatalytic carboxylation of acac in CCl4 in the presence of
0.5% Ru@ZnS-A. Top: IR spectra of the reaction mixture before (black) and
after irradiation (red); bottom: the differential spectrum.
Figure 6. Light-induced desorption of carbon dioxide from photocatalysts:
a) desorption in the dark, b) photodesorption.
before and after 6 h of irradiation. The formation of carboxylic
acids was confirmed by the following IR bands: 1) n˜ =
1710 cmꢀ1 (stretching vibration of C=O of the carboxylic
group); 2) n˜ =1259 cmꢀ1 (stretching vibration of CꢀO in CꢀOH,
coupled with an in-plane deformation vibration of OꢀH at n˜ =
1390–1440 cmꢀ1); 3) vibration of the ꢀC(=O)ꢀOꢀ group (bands
at n˜ =1550–1610 and 1300–1420 cmꢀ1). A reduction in the in-
tensity of the band assigned to stretching vibrations of the Cꢀ
H bond (n˜ =2850–2956 cmꢀ1 for aliphatic carbon) may be ex-
plained by CꢀC bond formation and a reduction in the
number of CꢀH bonds. As expected, no similar changes in the
IR spectra were detected when a solution of acac with the
photocatalyst was bubbled with argon instead of carbon diox-
ide prior to irradiation. Moreover, no carboxylic acids were de-
tected if ZnS-A was used instead of 0.5% Ru@ZnS-A.
The compound acetylacetone (acac) was selected as
a model organic substrate for carboxylation. This choice was
justified by the expected ability of acac to form radicals stabi-
lised by keto–enol tautomerism. Experimental and theoretical
studies have shown that the enolic form, with a conjugated p-
electron system and stabilised by intramolecular hydrogen
bonding, prevails in the gas phase, whereas in solution the
amount of the keto form increases with the solvent polari-
ty.[42,43] The acac substrate offers two different sites for carboxy-
lation to occur: the internal methylene (CH2) and terminal
methyl groups (CH3; Figure 7). In principle, two products, 2-
acetyl-3-oxobutanoic acid and 3,5-dioxohexanoic acid, could
be expected. Carbon tetrachloride was used as a solvent for
two reasons: it enables the use of IR spectroscopy to follow
the reaction progress (it does not absorb IR in the region at
which C=O vibrations occur), and assures good solubility of
carbon dioxide. However, other solvents (including water),
NMR spectroscopy gave further and more detailed evidence
for the formation of carboxylic acids. Figure 9 presents the
13C NMR spectrum of the reaction mixture taken after 4 h irra-
diation. Good accordance between the collected and predicted
spectra (calculations performed with ChemDraw 8.0 software;
Figure 10) proves the presence of carboxylic acids. 13C NMR
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2014, 79, 708 – 715 711

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
171 ppm, originating from the carboxyl group of 3,5-dioxohex-
anoic acid, appears in the spectrum. The molar ratio of the
two acids, which was roughly estimated through integration of
the NMR spectroscopy signals, was 2:1 2-acetyl-3-oxobutanoic
acid/3,5-dioxohexanoic acid. Although, so far, the amount of
acid formed after 4 h (max. 1%) did not allow isolation of the
carboxylic acid, the catalysts appeared to be quite stable
under the applied working conditions. No other products were
detected. The fact that the free acids are formed, rather than
their alkaline salts, which happens in other synthetic method-
ologies involving the use of bases such as sodium phenate, is
a very positive fact because it avoids post-reaction processing
that is usually necessary to convert the salts into the free acid
with the formation of large volumes of inorganic waste prod-
ucts. This is a unique feature of photochemical carboxylation
with respect to chemically driven carboxylations that use high
temperatures and lead to the formation of salts.[30] Moreover,
the formation of two isomers accounts for a different mecha-
nism with respect to the chemical route based on the extrac-
tion of a proton.
Figure 9. 13C NMR spectrum of the reaction mixture collected after 4 h of ir-
radiation. Experimental conditions: 0.5% Ru@ZnS-A suspended in a solution
of acac in CCl4 bubbled with carbon dioxide.
Materials based on ZnS-B are also active in the photocatalyt-
ic carboxylation process. Irradiation of the suspension, contain-
ing 0.5% Ru@ZnS-B and acac in CCl4, saturated with CO2, for
4 h resulted in the synthesis of carboxylic acids. Figure 11
shows IR spectra of the reaction mixture before and after irra-
Figure 10. Calculated 13C NMR spectra of 2-acetyl-3-oxobutanoic acid (top)
and 3,5-dioxohexanoic acid (bottom).
spectroscopic analysis indicates that 2-acetyl-3-oxobutanoic
acid is the main product of acac carboxylation.
The assignment of signals to every carbon of this acid
agrees with the calculated spectrum: 1) the signal at d=
208 ppm assigned to C1 of keto form and C10 of enolic form
of acac, to C2 of 3,5-dioxohexanoic acid or C2 of 2-acetyl-3-ox-
obutanoic acid; 2) 193 ppm assigned to C4 of enolic form of
acac; 3) d=183 ppm: C4 of 2-acetyl-3-oxobutanoic acid; 4) d=
171 ppm: C6 of 3,5-dioxohexanoic acid; 5) d=160 ppm: C7
and C12 of the enolic form of acac; 6) d=100 ppm: C5 of the
enol form of acac; 7) d=78 ppm: C3 of 2-acetyl-3-oxobutanoic
acid or C3 of 3,5-dioxohexanoic acid; 8) d=65 ppm: C13 of
the enolic form of acac; 9) d=58 ppm: C2 of the keto form of
acac; 10) d=30 ppm: C3 of the keto form of acac and C1 of
3,5-dioxohexanoic acid or C1 of 2-acetyl-3-oxobutanoic acid;
and 11) d=24 ppm (broad signal) assigned to C6, C8 and C9
of the enolic form of acac (numbers of atoms according to
those shown in Figure 7).
Figure 11. Photocatalytic carboxylation of acac in CCl4 in the presence of
0.5% Ru@ZnS-B. Top: IR spectra of the reaction mixture before (black) and
after irradiation (green); bottom: the differential spectrum.
diation. The appearance of characteristic bands attributed to
C=O vibrations (in the carboxylic group) was proof that the re-
action took place. The reaction did not proceed in systems
lacking any of the components (the photocatalyst, light or
carbon dioxide mixture saturated with argon), similarly to sys-
tems with ZnS-A. No degradation of acac was found in any
case.
To verify the mechanism of carboxylic acid formation, spin-
trapping measurements were performed. Figure 12 presents
the EPR spectrum of the reaction mixture, containing 0.5%
Ru@ZnS-A and 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) in 1%
methanol/water, recorded upon continuous irradiation. Metha-
nol was used to ensure consumption of photogenerated holes
and to reduce the efficiency of recombination processes. The
The signal at d=183 ppm, which is essential for identifica-
tion of the product, is assigned to the carboxyl group of 2-
acetyl-3-oxobutanoic acid. In addition, a small signal at d=
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2014, 79, 708 – 715 712

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
relative to 0.5% Ru@ZnS-A (Figure 2), and therefore, lower
photocatalytic activity of this material, the results indicate the
beneficial role of ruthenium nanoparticles in one-electron CO2
reduction processes.
Conclusion
Materials based on zinc sulfide have been synthesised, charac-
terised and tested as photocatalysts for the formation of car-
boxylic moieties from activated CꢀH bonds and CO2. IR and
13C NMR spectroscopy studies confirmed that the photocatalyt-
ic carboxylation of acac occurred in the presence of rutheni-
um-modified ZnS. Two products, namely, 2-acetyl-3-oxobuta-
noic acid and 3,5-dioxohexanoic acid, were found in the reac-
tion mixture. The formation of these products was in accord-
ance with the expected mechanism involving 1) one-electron
reduction of carbon dioxide to the carbon dioxide radical
anion, as confirmed by spin-trapping experiments; and 2) one-
hole oxidation of the organic substrate and radical generation
(Figure 14). The carboxylic acid was synthesised as a result of
Figure 12. Experimental (a) and simulated (b) EPR spectra of DMPO adducts
observed in a UV-irradiated water/methanol mixture saturated with CO2 con-
taining 0.5% Ru@ZnS-A of photocatalyst. The spectral components are as-
C
signed to the following spin adducts: c) DMPO –CO2 (aN =1.57 mT,
C
a
b-H =1.87 mT) and d) DMPO –CH2OH (aN =1.57 mT, ab-H =2.25 mT).
spectrum is a superposition of two components, assigned to
C
EPR signals originating from DMPO adducts: DMPO –CO2 (aN =
1.57 mT, ab-H =1.87 mT)[44] and DMPO –CH2OH (aN =1.57 mT, ab-
C
H =2.25 mT).[45] The EPR spectrum recorded in the presence of
0.5% Ru@ZnS-A and in the absence of methanol (Figure 13)
shows photogeneration of only one DMPO adduct, which is
characterised by the parameters: aN =1.45 mT, ab-H =1.60 mT.
[46,47]
C
This signal might be assigned to DMPO –SO3.
Sulfur-based
radicals can be formed upon catalyst photocorrosion.
The control experiment (not shown), performed under the
same conditions, but in the absence of the photocatalyst, did
not reveal generation of these radicals. EPR experiments prove
Cꢀ
therefore generation of CO2 radicals in the presence of 0.5%
Ru@ZnS-A as a result of one-electron reduction of CO2.
C
The DMPO –CO2 adduct was also observed if ZnS-A was
used instead of 0.5% Ru@ZnS-A (results not shown); however,
the EPR signal intensity was smaller. The EPR spectra analysis
revealed the following distribution of photogenerated radicals:
Figure 14. A plausible mechanism for the photocatalytic carboxylation of
acac.
C
C
C
DMPO –CO2, 19%; DMPO –CH2OH, 66%; DMPO –SO3, 15%; re-
corded in the presence of 0.5% Ru@ZnS-A, whereas in the
case of ZnS-A the distribution was 13, 55 and 32%, respective-
ly. Taking into account the higher fluorescence yields of ZnS-A
the reaction between these two radicals. The formation of the
new CꢀC bond resembled the processes promoted by CdS, as
described by Kisch et al.[28] The described reaction was enabled
by the redox potential of the conduction band edge of ZnS,
which was low enough to reduce carbon dioxide in a one-elec-
tron process. Differences in the reactivity of the two materials
ZnS-A and ZnS-B could be related to the different morpholo-
gies of the photocatalysts, as well as to their electronic proper-
ties (band gap energy, potential of the conduction band edge,
surface states, etc.). Moreover, the enhanced adsorption of CO2
at photocatalysts with deposited ruthenium(0) and the catalyt-
ic properties of ruthenium nanoparticles in electron-transfer re-
Cꢀ
actions improve the efficiency of CO2 formation, and there-
fore, also the overall yield of the reaction. The described pro-
cess may be an interesting example of a solar-light-driven
green synthesis of Cn+1 compounds utilising carbon dioxide as
a carbon source for the carboxylation of Cn species. EPR studies
Figure 13. Experimental (top) and simulated (bottom) EPR spectra of DMPO
adduct observed in a UV-irradiated aqueous suspension of 0.5% Ru@ZnS-A
saturated with CO2 (without methanol). The spin adducts are assigned to
Cꢀ
proved the possibility of generating CO2 radicals, even in
C
DMPO –SO3 (aN =1.45 mT, ab-H =1.60 mT).
water; however, further studies on the choice of the solvent,
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2014, 79, 708 – 715 713

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
BET specific surface area analysis was performed by using a Chemi-
sorb Micromeritics System. The samples were flushed with N2 for
30 min to clean their surface. The analysis was performed by using
a mixture of 30% N2 and 70% He. Analysis of the basic sites of se-
lected photocatalysts was performed by using CO2 as a probe gas.
Pulse chemisorption measurements were also used to determine
CO2 sorption isotherms. The experiment consisted of pulse chemi-
sorption followed by TPD. Before every measurement, the catalyst
was purged with a flow of N2 for 1 h. Afterwards, CO2 was flushed
by using He as a carrier gas. The flow ratio was kept at 1:1 CO2/He.
optimisation of quantum yields, as well as the extension of
such a process to non-activated CꢀH systems are in progress.
Experimental Section
ZnS-A: ZnS-A was prepared according to a literature procedure
with minor modifications.[28] Under an argon atmosphere (Schlenk
system), a solution of Na2S (0.1 mol) in distilled and deoxygenated
water (25 mL) was added dropwise to an aqueous solution of
ZnSO4·H2O (0.1 mol, 25 mL). The mixture was stirred for 24 h. After
filtration under argon, the powder was washed with H2O to neu-
trality and dried at room temperature under vacuum.
The photocatalyst (1 gLꢀ1) was suspended in the solvent (CCl4,
8 mL) and the organic substrate (0.1 molLꢀ1) was added. CO2 was
bubbled through the suspension for 15 min to remove oxygen and
saturate the solution. Photocatalytic tests were performed in
a closed quartz reactor (volume of 10 or 15 mL) by using XBO-150
or HBO-150 lamps as a light source. No additional optical filters
were applied. Samples of the solution were collected periodically,
filtered through syringe filters and analysed (IR and 13C NMR spec-
troscopy).
ZnS-B: A sample of ZnS-B was prepared with modification of the
procedure described by Kisch and co-workers.[28] Under a nitrogen
atmosphere, a solution of NaOH (10 g, 0.25 mol) in distilled and de-
gassed water (20 mL) was added dropwise to a solution of
ZnSO4·H2O (0.01 mol, 20 mL). The primarily formed solid Zn(OH)2
was dissolved to form [Zn(OH)4]2ꢀ before an aqueous solution
(40 mL) of thiourea (1.52 g, 0.02 mol) was added. The mixture was
heated to 353 K for 48 h. After separation by centrifugation, the
powder was washed with H2O to neutrality and dried in an oven
(333 K).
Samples for EPR measurements were prepared by mixing 0.5%
Ru@ZnS-A photocatalyst (0.5 mg) with water (990 mL), methanol
(10 mL) and DMPO (1 mL). The suspensions were saturated with
CO2, if not otherwise stated. The measurements were performed at
room temperature by using a Bruker Elexsys E-500 spectrometer
operating in X-band (9.8 GHz) and 100 kHz magnetic field modula-
tion equipped with super high sensitivity cavity ER 4122 SHQE. The
spectra were recorded at 2 mW microwave power, 0.1 mT modula-
tion amplitude, time constant 40.96 ms, conversion time 81.92 ms
(4 scans). EPR parameters of the spin-trap adducts were deter-
mined by a simulation procedure in EPRSIM32 software.[48] Samples
were irradiated by using an ER 203 UV irradiation system (50 W
high-pressure mercury lamp, full light).
Ru@ZnS: Ruthenium nanoparticles were deposited at the synthes-
ised materials. ZnS (1 g) was suspended in an aqueous solution of
RuCl3·xH2O (10 mL, cꢂ1, 10 or 15 mm) under an argon atmos-
phere. After stirring for 3 h, a concentrated solution of NaBH4 in
ethanol was added dropwise (in excess) under ultrasonic agitation.
The powder was filtered off and washed with water under argon
and dried at room temperature under vacuum. Dry materials were
stored under a nitrogen atmosphere. The amount of Ru loaded in-
creased with increasing concentrations of RuCl3 in the solution.
The obtained samples contained 0.5, 1.0 and 1.5 wt% of Ru, de-
pending on the applied concentration of RuCl3 solution. The
amount of deposited ruthenium was calculated by assuming 100%
yield of ruthenium reduction by NaBH4. As proof for this assump-
tion, EDX analyses of Ru performed for the solutions collected after
Ru deposition gave negative results.
Acknowledgements
We thank Dr. Piotr Pietrzyk for assistance with Raman measure-
ments. Support from the National Science Centre within the
2011/01/N/ST5/05543 grant and the European Institute of Innova-
tion and Technology, under the KIC InnoEnergy, NewMat project,
is highly acknowledged. Characterisation of the materials (TPD,
BET surface area, EDX) was performed at the University of Bari
thanks to financial support from CIRCC and the Foundation for
Polish Science (VENTURES, 2011-8/1), while mechanistic studies
are part of the project “Activation of small molecules in photoca-
talytic systems”, realised within the TEAM programme. Both VEN-
TURES and TEAM programmes, awarded by the Foundation for
Polish Science, are co-financed by a European Regional Develop-
ment Fund.
Diffuse reflectance spectra of the photocatalysts were measured by
using a UV/Vis/near-IR spectrophotometer (UV-3600, Shimadzu)
equipped with an integrating sphere (15 cm diameter). Materials
were ground with BaSO4; the formed pellets were subjected to dif-
fuse reflectance spectroscopy measurements.
Prepared materials were analysed by powder XRD (MiniFlex 600,
Rigaku) operated at a voltage of 40 kV. Solid-state Raman spectra
were recorded by using a Renishaw inVia Raman microscope. Pho-
toluminescence spectra were collected at ambient temperature by
using a FluoroLog (Horiba) spectrofluorometer. Energy dispersive
X-ray spectra were recorded on an EDX-720 (Shimadzu) spectrome-
ter operated at 5 kV voltage. 13C NMR spectra were collected on
a Bruker 600 MHz spectrometer. Hydrodynamic particle diameters
were measured by using a Zetasizer NanoZS (Malvern) instrument.
A scanning electron microscope (Vega 3, Tescan), equipped with
a LaB6 cathode and EDX detector, was operated at a voltage of
30 kV.
Keywords: carbon dioxide fixation
nanomaterials · photochemistry · semiconductors
·
CꢀC coupling
·
[1] M. Aresta, A. Dibenedetto, Dalton Trans. 2007, 2975–2992.
[2] M. Aresta, Carbon Dioxide as a Chemical Feedstock (Ed.: M. Aresta),
Wiley-VCH, Weinheim, 2010.
[3] M. Hoffmann, J. Moss, M. Baum, Dalton Trans. 2011, 40, 5151–5158.
[4] T. Yui, Y. Tamaki, K. Sekizawa, O. Ishitani, Top. Curr. Chem. 2011, 303,
151–184.
[5] S. C. Roy, O. K. Varghese, M. Paulose, C. A. Grimes, ACS Nano 2010, 4,
1259–1278.
Photodesorption tests were performed as follows: photocatalysts
(450 mg) were placed in the IR gas cuvette and purged for 10 min
with a stream of CO2, followed by a stream of argon (20 min). Af-
terwards, the signal of desorbed CO2 was measured by IR spectros-
copy (Spectrum BX, PerkinElmer).
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2014, 79, 708 – 715 714

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
[6] J. Yu, J. Jin, B. Cheng, M. Jaroniec, J. Mater. Chem. A 2014, 2, 3407–
3416.
[29] H. Fujiwara, M. Kanemoto, H. Ankyu, K. Murakoshi, Y. Wada, S. Yanagida,
J. Chem. Soc. Perkin Trans. 2 1997, 317–322.
[7] W.-J. Ong, M. M. Gui, S.-P. Chai, A. R. Mohamed, RSC Adv. 2013, 3, 4505–
4509.
[30] J. Zhang, J. Yu, Y. Zhang, Q. Li, J. R. Gong, Nano Lett. 2011, 11, 4774–
4779.
ˇ
´
[8] E. E. Barton, D. M. Rampulla, A. B. Bocarsly, J. Am. Chem. Soc. 2008, 130,
6342–6344.
[31] K. Kocꢄ, M. Reli, O. Kozꢅk, Z. Lacny, D. Plachꢅ, P. Praus, L. Obalovꢅ, Catal.
Today 2011, 176, 212–214.
[9] H. Fujiwara, H. Hosokawa, K. Murakoshi, Y. Wada, S. Yanagida, T. Okada,
H. Kobayashi, J. Phys. Chem. B 1997, 101, 8270–8278.
[10] B. Vijayan, N. M. Dimitrijevic, T. Rajh, K. Gray, J. Phys. Chem. C 2010, 114,
12994–13002.
[11] P. Johne, H. Kisch, J. Photochem. Photobiol. A 1997, 111, 223–228.
[12] J. Mao, K. Li, T. Peng, Catal. Sci. Technol. 2013, 3, 2481–2498.
[13] S. N. Habisreutinger, L. Schmidt-Mende, J. K. Stolarczyk, Angew. Chem.
2013, 125, 7516–7557; Angew. Chem. Int. Ed. 2013, 52, 7372–7408.
[14] S. Navalꢁn, A. Dhakshinamoorthy, M. ꢂlvaro, H. Garcia, ChemSusChem
2013, 6, 562–577.
[32] B. R. Eggins, P. K. J. Robertson, J. H. Stewart, E. Woods, J. Chem. Soc.
Chem. Commun. 1993, 349–350.
[33] H. Yoneyama, Catal. Today 1997, 39, 169–175.
[34] A. Dibenedetto, P. Stufano, W. Macyk, T. Baran, C. Fragale, M. Costa, M.
Aresta, ChemSusChem 2012, 5, 373–378.
[35] R. Kꢃnneth, G. Twardzik, G. Emig, H. Kisch, J. Photochem. Photobiol. A
1993, 76, 209–215.
[36] Y. Wang, L. Zhang, C. Liang, G. Wang, X. Peng, Chem. Phys. Lett. 2002,
357, 314–318.
[37] Y.-C. Zhu, Y. Bando, D.-F. Xue, Appl. Phys. Lett. 2003, 82, 1769–1771.
[38] J. Bearden, Rev. Mod. Phys. 1967, 39, 78–124.
[39] S. Y. Mar, C. S. Chen, Y. S. Huang, K. K. Tiong, Appl. Surf. Sci. 1995, 90,
497–504.
[15] H.-C. Hsu, I. Shown, H.-Y. Wei, Y.-C. Chang, H.-Y. Du, Y.-G. Lin, C.-A. Tseng,
C.-H. Wang, L.-C. Chen, Y.-C. Lin, K. H. Chen, Nanoscale 2013, 5, 262–
268.
[16] E. Fujita, Coord. Chem. Rev. 1999, 185–186, 373–384.
[17] Z. Jiang, T. Xiao, V. L. Kuznetsov, P. P. Edwards, Philos. Trans. R. Soc.
London Ser. A 2010, 368, 3343–3364.
[18] D. Ravelli, D. Dondi, M. Fagnoni, A. Albini, Chem. Soc. Rev. 2009, 38,
1999–2011.
[40] J. Mascetti, Carbon Dioxide as Chemical. Feedstock (Ed.: M. Aresta),
Wiley-VCH, Weinheim, 2010.
[41] W. Macyk, H. Kisch, Monatsh. Chem. 2007, 138, 935–940.
[42] M. M. Folkendt, B. E. Weiss-Lopez, J. P. Chauvel, N. S. True, J. Phys. Chem.
1985, 89, 3347–3352.
[19] P. Pichat, Catal. Today 1994, 19, 313–333.
[20] D. Dondi, A. Albini, M. Fagnoni, S. Protti, S. M. Carpio, Green Chem.
2009, 11, 1653–1659.
[21] D. Ravelli, A. Albini, M. Fagnoni, Chemistry 2011, 17, 572–579.
[22] H. Al-Ekabi, P. De Mayo, Tetrahedron 1986, 42, 6277–6284.
[23] S. Yanagida, K. Mizumoto, C. Pac, J. Am. Chem. Soc. 1986, 108, 647–654.
[24] H. Kisch, Angew. Chem. 2013, 125, 842–879; Angew. Chem. Int. Ed.
2013, 52, 812–847.
[43] H. Nakanishi, H. Morita, S. Nagakura, Bull. Chem. Soc. Jpn. 1977, 50,
2255–2261.
[44] F. A. Villamena, E. J. Locigno, A. Rockenbauer, C. M. Hadad, J. L. Zweier,
J. Phys. Chem. A 2006, 110, 13253–13258.
[45] B. C. Gilbert, G. R. Hodges, J. R. Lindsay Smith, P. MacFaul, P. Taylor, J.
Chem. Soc. Perkin Trans. 2 1996, 519–524.
[46] J. Kuncewicz, P. Za?a˛a˛a˛bek, K. Kruczała, K. Szaciłowski, W. Macyk, J. Phys.
Chem. C 2012, 116, 21762–21770.
[25] H. Inoue, Y. Kubo, H. Yoneyama, J. Chem. Soc. Faraday Trans. 1991, 87,
553–557.
[47] P. L. Zamora, F. A. Villamena, J. Phys. Chem. A 2012, 116, 7210–7218.
[48] T. Spałek, P. Pietrzyk, Z. Sojka, J. Chem. Inf. Model. 2005, 45, 18–29.
[26] W. Schindler, H. Kisch, J. Photochem. Photobiol. A 1997, 103, 257–264.
[27] M. Fagnoni, D. Dondi, D. Ravelli, A. Albini, Chem. Rev. 2007, 107, 2725–
2756.
Received: December 30, 2013
[28] G. Hçrner, P. Johne, R. Kꢃnneth, G. Twardzik, H. Roth, T. Clark, H. Kisch,
Chem. Eur. J. 1999, 5, 208–217.
Published online on March 18, 2014
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2014, 79, 708 – 715 715