Powering a CO2 Reduction Catalyst with Visible Light through Multiple Sub-picosecond Electron Transfers from a Quantum Dot

Article abstract of DOI:10.1021/jacs.7b03134

Photosensitization of molecular catalysts to reduce CO₂ to CO is a sustainable route to storable solar fuels. Crucial to the sensitization process is highly efficient transfer of redox equivalents from sensitizer to catalyst; in systems with molecular sensitizers, this transfer is often slow because it is gated by diffusion-limited collisions between sensitizer and catalyst. This article describes the photosensitization of a meso-tetraphenylporphyrin iron(III) chloride (FeTPP) catalyst by colloidal, heavy metal-free CuInS₂/ZnS quantum dots (QDs) to reduce CO₂ to CO using 450 nm light. The sensitization efficiency (turnover number per absorbed unit of photon energy) of the QD system is a factor of 18 greater than that of an analogous system with a fac-tris(2-phenylpyridine)iridium sensitizer. This high efficiency originates in ultrafast electron transfer between the QD and FeTPP, enabled by formation of QD/FeTPP complexes. Optical spectroscopy reveals that the electron-transfer processes primarily responsible for the first two sensitization steps (Fe^IIITPP ? Fe^IITPP, and Fe^IITPP ? Fe^ITPP) both occur in <200 fs.

Full text of DOI:10.1021/jacs.7b03134

Subscriber access provided by EAST TENNESSEE STATE UNIV
Article
2
Powering a CO Reduction Catalyst with Visible Light through
Multiple Sub-picosecond Electron Transfers from a Quantum Dot
Shichen Lian, Mohamad S. Kodaimati, Dmitriy S. Dolzhnikov, Raul Calzada, and Emily A. Weiss
J. Am. Chem. Soc., Just Accepted Manuscript • Publication Date (Web): 13 Jun 2017
Downloaded from http://pubs.acs.org on June 13, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 18
Journal of the American Chemical Society
1
2
3
4
5
6
Powering a CO2 Reduction Catalyst with Visible Light through Multiple Sub-
picosecond Electron Transfers from a Quantum Dot
7
8
9
Shichen Lian, Mohamad S. Kodaimati, Dmitriy S. Dolzhnikov, Raul Calzada, and Emily A.
Weiss*
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Department of Chemistry, Northwestern University, 2145 Sheridan Rd., Evanston, IL 60208ꢀ
3113
*corresponding author. Email: eꢀweiss@northwestern.edu
ABSTRACT
Photosensitization of molecular catalysts to reduce CO2 to CO is a sustainable route to
storable solar fuels. Crucial to the sensitization process is highly efficient transfer of redox
equivalents from sensitizer to catalyst; in systems with molecular sensitizers, this transfer is often
slow because it is gated by diffusionꢀlimited collisions between sensitizer and catalyst. This
article describes the photosensitization of a mesoꢀtetraphenylporphyrin iron(III) chloride
(FeTPP) catalyst by colloidal, heavy metalꢀfree CuInS2/ZnS quantum dots (QDs) to reduce CO2
to CO using 450ꢀnm light. The sensitization efficiency (turnover number per absorbed unit of
photon energy) of the QD system is a factor of 18 greater than that of an analogous system with a
facꢀtris(2ꢀphenylpyridine)iridium sensitizer. This high efficiency originates in ultrafast electron
transfer between the QD and FeTPP, enabled by formation of QD/FeTPP complexes. Optical
spectroscopy reveals that the electron transfer processes primarily responsible for the first two
sensitization steps (FeIIITPP ꢀ FeIITPP, and FeIITPP ꢀ FeITPP) both occur in <200 fs.
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 18
1
2
3
4
INTRODUCTION
5
6
7
8
Photocatalytic reduction of CO2 is a sustainable way to utilize abundant solar energy to
produce CO, a precursor for reactions that form storable H2 and liquid fuels, such as the waterꢀ
gas shift and FischerꢀTropsch processes, and reactions using organometallic complexes that run
under much milder conditions.1ꢀ3 Several molecular redox catalysts, containing precious metals
such as Re4ꢀ6 or earthꢀabundant metals such as Co7,8, Ni9,10, and Fe11,12, are capable of reducing
CO2 to CO electrochemically or photochemically. Mesoꢀtetraphenylporphyrin iron(III) chloride
(FeTPP) and its derivatives are the class of earthꢀabundant homogeneous catalysts with the best
combination of efficiency and chemical stability for the reduction of CO2 to CO, when they are
fed electrons by an electrode or photosensitizer,13ꢀ15 such as tris(bipyridine)ruthenium(II)
chloride (Ru(bpy)3), facꢀtris(2ꢀphenylpyridine)iridium (Ir(ppy)3), or organic dye.16ꢀ18 These
sensitizers, however, either contain precious metals or only absorb a significant amount of light
at 400 nm or shorter wavelengths (which only accounts for ~4% of the photons in the AM 1.5G
solar spectrum). Furthermore, without extensive organic synthesis to covalently link sensitizer
with catalyst19,20, these sensitizers do not have a mechanism for chemisorption of
photosensitizers and catalysts. The electron transfer between the two is then primarily mediated
by bimolecular collision reactions, with rates and catalytic efficiencies limited by diffusion and
concentration.14,21 A final disadvantage of these sensitizers is that adjusting their redox potentials
or absorption spectra to accommodate a certain application or catalyst, by changing the
coordination environments of metals or the size of the π systems of organics, is not
straightforward synthetically, and often results in sacrificing one desirable property – coverage
of the solar spectrum, electrochemical potential, or chemical stability – for another.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Colloidal quantum dots as photosensitizers for CO2 reduction offer solutions to all of these
problems. The heavy metalꢀfree CuInS2/ZnS core/shell QDs we use in this study have an
extinction coefficient of ε = ~70000 Mꢀ1·cmꢀ1 at 450 nm, a factor of 3 ꢀ 25 greater than that of
Ir(ppy)3 and Ru(bpy)3, Figure 1A.22ꢀ24 Quantum confinement of their electrons renders their
absorption spectrum tunable over the visible region by simply changing their diameter, and
enhances their reduction potential from ꢀ1.7 V for bulk CuInS2 25 to ~ ꢀ2.4 V vs. SCE for the QDs
we use here (diameter = 2.5 nm, Figure 1B), enough to perform all three reductions of FeTPP
required for CO2 binding and reduction. Importantly, QDs also have large surface areaꢀtoꢀ
volume ratios, and their surfaces facilitate (or can be simply functionalized to facilitate)26ꢀ31 the
2
ACS Paragon Plus Environment

Page 3 of 18
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
adsorption of molecular catalysts to form quasiꢀstatic complexes that mediate femtosecondꢀtoꢀ
picosecond rather than microsecond (diffusionꢀlimited) charge transfer between sensitizer and
catalyst.32
The two reports of catalytic systems using heavyꢀmetal containing CdS QDs as
photosensitizers for CO2 reduction (with enzyme33 and cobalt complex coꢀcatalysts34) show
promise for this strategy, but both had energy efficiencies limited by the reduction potential of
CdS QDs, which are approximately 1 V less reducing than CuInS2, and by their limited
absorption of visible light.35 Feldmann and coꢀworkers demonstrated the efficacy of a
combination of more highly reducing Cu2S nanorods with Pt nanoparticles for conversion of CO2
to CO and methane, although the 1.2 eV bandgap of Cu2S required an exceptional hole
scavenger.36
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Here we show that soluble, colloidal, heavy metalꢀfree CuInS2/ZnS QDs photosensitize the
catalytic conversion of CO2 to CO by a model catalyst, FeTPP, in DMSO, using visible light,
Scheme 1. Upon successive photoexcitations at 450 nm, the QD donates three electrons – the
first and second in <200 fs – to form the catalytically active species Fe0TPP, which binds CO2
and reduces it to CO with 84% selectivity (16% H2). The sacrificial reductant N,N,N′,N′ꢀ
tetramethylꢀpꢀphenylenediamine (TMPD) reꢀestablishes the QD’s ground state within 700 ps.
There is no detectable degradation of catalytic activity over 40 h, the longest time we monitored
the system. The sensitization efficiency of this sensitizerꢀcatalyst system is a factor of 18 higher
than the most efficient photostable homogenous system for photoreduction of CO2 to CO using
an earth abundant metal, FeTPP(OH)4 sensitized by Ir(ppy)3,14 even though Ir(ppy)3 absorbs
higherꢀenergy light (420 nm) than does the QD. Mechanistic analysis of the photosensitizerꢀ
catalyst interaction is crucial for maximizing what in electrocatalysis is the “heterogeneous rate
constant”, the rate of injection of redox equivalents into the catalyst. This type of analysis is
often complicated with, for example, Ir(ppy)3 or Ru(bpy)3, sensitizers because these complexes
form triplet excited states, undergo intramolecular charge separations, and have convoluting
groundꢀstate absorption features with the molecular coꢀcatalysts. The QDs that we use here allow
for direct measurements of the rates for photoinduced electron transfer. Specifically, timeꢀ
resolved optical spectroscopy allows us to determine the origin of the high efficiency of the QDꢀ
FeTPP system: ultrafast electron donation from the QD to FeTPP in nearꢀquantitative yield,
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 18
1
2
3
4
enabled by the formation of quasiꢀstatic
5
6
7
8
complexes of sensitizer and catalyst with no
covalent linking chemistry.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
RESULTS AND DISCUSSION
Catalytic Mixture and Mechanism. We
synthesized oleateꢀcapped CuInS2/ZnS core/shell
QDs based on a modified literature procedure37
(Figure S11) and subsequently displaced all of
the oleate ligands (as shown by NMR, see Figure
S1) with 3ꢀmercaptoꢀ1ꢀpropanol (MPO) to form
DMSOꢀsoluble QDs, the photosensitizers for the
reaction. The Supporting Information describes
our synthesis and ligand exchange procedure in
detail.
Figure 1. Characterization of the Catalyst
and Sensitizer. A) Ground state absorption
spectra of 16 µM CuInS2/ZnS core/shell QDs
(black), 8 µM FeTPP purged with CO2 (red),
and a mixture of the QDs and 0.5 eq. FeTPP
purged with CO2 (green); and PL spectrum of
the QDs (black dash), in DMSO. The blue
arrow indicates the excitation wavelength
used for the catalysis, PL, and transient
absorption experiments. B) CV of 1 mM QDs
under N2 (black), FeTPP under N2 (blue), and
FeTPP under CO2 (red), with 50 mM
tetrabutylammonium hexafluorophosphate in
DMSO, a glassy carbon working electrode,
an Ag wire reference electrode, and a Pt wire
counter electrode. The potentials have been
referenced to SCE using a ferrocene internal
standard. The discontinuities in the scans are
artifacts caused by switching the polarity of
the electrodes (shifted below 0 V when
referencing the data to Fc/Fc+).
Figure 1A shows that the broad first
excitonic peak of MPOꢀcapped CuInS2 QDs is
centered near 435 nm, and the emission peaks at
610 nm, which corresponds to a particle diameter
of 2.5 nm.22 Upon introduction of QDs, the Soret
band of FeTPP shifts to lower energy, which
suggests a change in axial ligand substitution,
specifically that the Fe center within FeTPP is
interacting with an electron donating group38,39
(probably the electronꢀenriched sulfur sites on
the surface of QDs)40. When the QDs are mixed
with FeTPP in CO2ꢀsaturated DMSO, in the
molar ratio used for the catalysis (QD:FeTPP =
1:0.02) or that used for the timeꢀresolved
spectroscopy experiments (QD:FeTPP = 1:0.5),
4
ACS Paragon Plus Environment

Page 5 of 18
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
we achieve excellent selectivity (≥95%) for photoexcitation of the QD sensitizer (as opposed to
FeTPP) at 450 nm, the excitation wavelength used for all studies, Figure 1A (green). We use a
large excess of QD sensitizer for catalysis in order to ensure that at most one FeTPP binds per
QD, and thereby maximize the probability that photoexcited electrons all go to the same catalyst.
Figure 1B shows cyclic voltammograms (CVs) of MPOꢀcapped QDs under N2, FeTPP
under N2, and FeTPP under CO2, all in DMSO. When the FeTPP sample is purged with CO2,
catalytic current is observed upon reduction of Fe+ to Fe0, which confirms that the Fe0 center
within FeTPP molecule is the catalytic site.13,14 The first observed reductive wave of the QDs (~ꢀ
2.5 V) is more negative than the catalyst standard potential, which indicates that the
photoinduced electrons within CuInS2 QDs used in this study have enough energy to reduce FeIII
to Fe0, and thereby photosensitize FeTPP.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Proposed Photocatalytic Mechanism for the Reduction of CO2 to CO via
Electron Transfer from the QD to FeTPP and Hole Transfer from the QD to TMPD.
Scheme 1 summarizes a photocatalytic mechanism for CO2 reduction by QDꢀsensitized
FeTPP, based on the previously reported electrochemical mechanism.41 During the initial
sensitization step (top blue shaded region), a QD sequentially accepts three photons, one at a
time. After each photoexcitation (to form QD*), the QD donates one electron to FeTPP and one
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 18
1
2
hole to the sacrificial reductant, TMPD, to regenerate its ground state. The FeIII center within 1
thereby consecutively accepts three electrons to become Fe0 (2), which binds CO2 to form the
catalytically active species Fe0ꢀC(+4)O2 (3). The first step of the reduction of CO2 is a protonꢀ
coupled electron transfer (PCET) from Fe0 to CO2 to form FeIꢀC(+3)OOH (4). The C(+3)OOH
moiety in 4 then accepts another electron from FeI and another proton from the surrounding
media and undergoes dehydration to form 5. Finally, a CO molecule releases from 5. In another
sensitization step, the remaining FeII center captures two electrons from a consecutively
photoexcited QD to reꢀestablish 2 and continue the catalytic cycle. We suspect that the ligand
“L” in 1 is DMSO when CO2 is not present, and is CO2 (although the exact nature of binding
mode is unclear) when it is available, because, upon addition of CO2 to FeTPP, changes in the
absorption spectrum of FeTPP, possibly indicative of axial ligand substitution, occur, see the
Figure S2.39
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We suspect, but have not proven, that the protons needed during this cycle are donated by
residual water in the DMSO (which we did not dry thoroughly) and the water formed during the
ligand exchange procedure. During the phase transfer of the QDs to DMSO, we added
benzyltrimethylammonium hydroxide (triton B) to deprotonate the added thiols (the MPO
ligands) to thiolates. This process generates water, which is then transferred to DMSO.
Catalytic Efficiency of the QD-FeTPP Mixture. The Methods section describes our
photocatalytic setup in detail. Briefly, we put QDs samples in airꢀtight vials and illuminated
these vials with a 4.5ꢀmW, 450 nm laser pointer (bandwidth = ~ 2 nm) for various amount of
time. We then sampled the head space of these vials with gas chromatography (GC) with thermal
1
conductivity detection (TCD) to monitor the production of H2 and CO. H NMR of the samples
in DMSOꢀd6 did not detect any solutionꢀphase catalytic products. Without light, the
QD/FeTPP/CO2 sample does not produce any CO or H2. After 10 hours of illumination at 450
nm, the QD/FeTPP/CO2 mixture produces both H2 and CO, whereas the QD/FeTPP/Ar mixture
produces only a trace amount of H2. With no added FeTPP, the QD/CO2 sample produces
primarily H2 and a negligible amount of CO. The FeTPP/CO2 sample (with no QDs) under the
same conditions does not produce a GCꢀdetectable amount of H2 or CO (see all GC traces in
Figure S3). Qualitative analysis of the GC data allows us to conclude that i) the detected CO is
coming from CO2, not from another carbon source such as the surface ligands or solvent, ii) the
electrons that reduce the Fe center in FeTPP originate from the photoexcited state of the QD, and
6
ACS Paragon Plus Environment

Page 7 of 18
Journal of the American Chemical Society
1
2
3
4
5
6
iii) the vast majority of CO is produced by the QD/FeTPP assembly, rather than by the QD or
FeTPP on its own.
7
8
9
Figure 2, red triangles, shows the kinetics of CO production for our reaction mixture, CuInS2
QDs with 0.02 eq. FeTPP, 1000 eq. of the hole scavenger TMPD, in DMSO purged with CO2.
The QD/FeTPP system shows no sign of decreasing catalytic activity after 40 hours of
illumination at 450 nm and 4.5 mW (Figure S8). Our goal is to demonstrate and analyze the
effectiveness of our QDꢀbased sensitization scheme for the particular FeTPP catalyst we have
chosen – that is, the efficiency with which the QD transduces light energy to reducing
equivalents, and ultimately to CO. The conventional plot of turnover number (TON =
[CO]/[catalyst]) vs. illumination time does not, however, account for the intensity or energy of
the light source (a higher photon flux, for example, will lead to more CO by creating more redox
equivalents per unit time) and is therefore not a good measure of sensitization efficiency. Instead,
we plot in Figure 2 (red triangles) the TON vs. cumulative photon energy absorbed by the
sensitizer, calculated according to eq. 1. Also
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Measurement of Catalytic Efficiency. Red triangles: Plot of the turnover number
of CO (per catalyst molecule) vs. the cumulative absorbed photon energy, upon illumination
of a mixture of 100 µM QDs (+ 1000 eq. TMPD) and 2 µM of FeTPP in CO2ꢀsaturated
DMSO. Shown for comparison are analogous data (extracted from Ref.14) for the production
of CO by a hydroxylated FeTPP catalyst, photosensitized by Ir(ppy)3 (blue squares) or CNA
(black circles), as described in the text. The Supporting Information contains the GC traces
used to construct this plot and the calculation of cumulative input photon energy from
illumination time.
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 18
1
2
3
4
5
6
7
8
9
ꢓ
ꢒ ꢔ
ꢀꢁꢂꢁꢃꢄꢅꢆꢇꢈ ꢄꢉꢊꢋꢌꢉꢈꢍ ꢎℎꢋꢅꢋꢏ ꢈꢏꢈꢌꢐꢑ ꢓ = ꢃꢆꢐℎꢅ ꢆꢏꢅꢈꢏꢊꢆꢅꢑ ꢕ
ꢘ
ꢊ ∗ ꢖꢂꢗ
ꢗ
ꢒ ꢔ
× ꢆꢃꢃꢁꢂꢆꢏꢄꢅꢆꢋꢏ ꢄꢌꢈꢄ ꢒꢖꢂ ꢔ × ꢅꢆꢂꢈ ꢊ × ꢙꢌꢄꢖꢅꢆꢋꢏ ꢋꢙ ꢎℎꢋꢅꢋꢏꢊ ꢄꢉꢊꢋꢌꢉꢈꢍ (1)
plotted in Figure 2 are the analogous datasets for two published photocatalytic reaction
mixtures: a derivatized FeTPP catalyst sensitized by Ir(ppy)3 (blue squares) and by 9ꢀ
cyanoanthracene (CNA, black circles). We chose these data for comparison because Ir(ppy)3 is
the most commonly used photosensitizer for all molecularꢀcatalyst based homogenous
photoreduction of CO2, and because FeTPP(OH)4/Ir(ppy)3 is the most efficient photostable
homogenous system for photoreduction of CO2 to CO. The combination of several other FeII and
CoII based catalysts coupled with Ir(ppy)3/Ru(bpy)3 sensitizers21,42ꢀ45 have higher turnover
frequencies, but all have turnover number vs. time plots that deviate from linearity within ten
hours of illumination, indicating some degradation or poisoning of the catalytic system. In
constructing the plots in Figure 2, we assume the TON is photonꢀlimited – that is, increasing the
power of the excitation source will increase the total amount of CO produced; we prove this
assumption for both our system and the published systems, and show our detailed calculations
using eq 1, in the Supporting Information.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The sensitization efficiency (defined as the number of CO molecules produced / J of
absorbed photon energy / catalyst molecule, i.e. the slope of the curve in Figure 2) of the QDꢀ
sensitized system is a factor of 18 larger than that of Irꢀsensitized system, and a factor of 38
larger than that of the CNAꢀsensitized system. Given the data in Figure 2, the external quantum
efficiency for conversion of photons to reducing equivalents, EQE = 2×(number of CO molecules
produced per catalyst)/(number of incident photons), is 0.01% for the QDꢀsensitized system,
0.0013% for the Irꢀsensitized system, and 0.0008% for the CNAꢀsensitized system.14 46
Importantly, the enhancement in sensitization efficiency by the QDs, relative to Ir and CNA,
is much more dramatic than what is evident from Figure 2 and the numbers we derive from it,
because in order to compare our system to the Ir and CNAꢀsensitized systems, we ignored two
important features of the published systems that overestimates their performance relative to that
of the QDꢀsensitized system. First, the catalyst used in the Irꢀ and CNAꢀsensitized systems is
(FeTPP(OH)4), whereas we use FeTPP. The hydroxyl groups of (FeTPP(OH)4) stabilize the CO2ꢀ
FeTPP species through hydrogen bonding, resulting in a decrease of the overpotential for
reduction of CO2 to CO by up to 0.4 V and an increase in the rate constant for catalysis (in their
8
ACS Paragon Plus Environment

Page 9 of 18
Journal of the American Chemical Society
1
2
electrochemical experiment) by a factor of 100.13 Second, we use a monochromatic (450ꢀnm)
laser pointer as a light source, so the energy input plotted on the xꢀaxis of Figure 2 is the only
energy we use to produce CO. In contrast, the authors of the published Ir and CNA work (and
almost all studies on the photoreduction of CO2) used a broadband light source, and the turnover
numbers plotted in Figure 2 for their systems are a result of energy absorbed by their sensitizers
over a range of wavelengths. Since we do not, however, know the intensity profile of their lamp,
we calculated the energy input plotted on the xꢀaxis of Figure 2 only from the single wavelength
at which they reported the power of their incident light, when in fact more energy was absorbed
at other wavelengths. Due to both this discrepancy and the difference in catalyst between our
experiment and the published experiment, Figure 2 only shows the lower limit of the
enhancement in sensitization efficiency by the QDs, relative to the best known photostable
sensitizers for FeTPPs. Interestingly, when we mixed Ir(ppy)3 with FeTPP (instead of the
hydroxylated FeTPP used in ref 14), we observed the formation of H2 and no CO at all (Figure
S12 in the Supporting Information) upon excitation at 405 nm. This result further proves that
QDs can photosensitize FeTPP for the reduction of CO2 much more efficiently than Ir(ppy)3
under similar illumination conditions and that production of CO using the Ir(ppy)3 sensitizer
requires the “better” (hydroxylꢀfunctionalized) FeTPP catalyst.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The catalytic selectivity for formation of CO over H2, [CO]/([CO] + [H2]), of the QD/FeTPP
system is 84%, Figure S8, slightly lower than the selectivities of the Ir(ppy)3/FeTPP(OH)4 (93%)
and CNA/FeTPP(OH)4 (100%) systems. We suspect that the H2 in our system is produced via
direct reduction of H+ by photoexcited QDs without adsorbed FeTPP, since we only used 0.02
eq. FeTPP per QD and the possible proton source includes the residual water in DMSO and
leftover water generated during ligand exchange procedure (see the Supporting Information for
details). Further optimization of the sensitizer:catalyst ratio should allow us to increase the
selectivity of our system.
Efficient photosensitization is enabled by ultrafast electron transfer (eT) from QDs to
FeTPP. Why are the QDs more efficient sensitizers for FeTPP than the Ir and CNA systems,
both in terms of quantum efficiency and energy efficiency? One immediate advantage of the
CuInS2 QD system is that it can be photoexcited at longer wavelengths than the Ir and CNA
sensitizers and still produce sufficiently reducing electrons. This feature of CuInS2 QDs is also
advantageous, in general, for minimizing side reactions that occur more frequently with higherꢀ
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 18
1
2
3
4
5
6
7
8
9
energy incident photons. The more subtle, and more important, advantage of the QDꢀsensitized
system is that, while the exceptionally low yields of PL quenching of the sensitizer in molecular
systems like Ir/FeTPP(OH)4 and CNA/FeTPP(OH)4 strongly suggest that eT in these systems
relies on collisions between freely diffusing sensitizers and catalysts to transfer electrons
between the two (Figure S9 of the Supporting Information), such that the eT is gated by
bimolecular reaction kinetics, the QDs and FeTPP form what appears to be highꢀaffinity
sensitizer/catalyst complexes. Again, the formation of these complexes is supported by the
spectral shift of the Soret band of FeTPP upon addition of QDs (Figure 1A, green vs. red).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
10
ACS Paragon Plus Environment

Page 11 of 18
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. Kinetics of Photoreduction of the Catalyst by the QD Sensitizer. A) TA spectra
of FeTPP (light redꢀdark red), or a mixture QDs and of FeIIITPP (black ꢀ blue), at various
delay times after excitation at 450 nm. FeIIITPP/CO2 itself has no transient absorptions or
bleaches in this region. B) Steadyꢀstate PL spectra of the QDs without (black) and with (red)
added FeIIITPP. The integrated PL intensity of the QDs is quenched by 45% upon addition of
0.5 eq. of FeIIITPP. C) Kinetic traces extracted at 480 nm from the TA spectra of the QDs
(black) and a mixture QDs and of FeIIITPP (red), after excitation at 450 nm. The dynamics at
these wavelengths are those of the reducing electron of the QD. D) Decay of the PL of QDs
without (black) and with (red) added FeIIITPP, monitored at 600 nm, after excitation at 450
nm. E) Kinetic traces extracted at 490 nm from the TA spectra of the QDs (black) and a
mixture QDs and of FeIITPP (red). This set of TA experiments was conducted with excitation
at 460 nm in order to selectively excite the QDs. FeIITPP does not attenuate any light at this
wavelength. F) Decay of the PL of QDs without (black) and with (red) added FeIITPP,
monitored at 600 nm, after excitation at 450 nm. Tables 1ꢀ3 list the fitting parameters for the
TA and PL data. All samples are in DMSO and purged with CO2. All mixtures of QDs and
FeTPP have 0.5 eq. FeTPP per QD. The two kinetic traces in each of (C) and (E) were
collected with the same excitation power and pumpꢀprobe overlap, so their amplitudes are
comparable at all delay times.
The consequence of the formation of these sensitizerꢀcatalyst complexes is significant
electronic coupling of the QD and FeTPP and, we show here, ultrafast reduction of the catalyst
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 18
1
2
3
4
by the QD. Upon addition of 0.5 eq. of FeTPP to the QDs, the photoluminescence (PL) intensity
5
6
7
8
of the QDs is quenched by 45%, Figure 3B, a quenching yield of 90% (quenching yield =
% ꢚꢛ ꢜꢝ ꢞꢟꢠꢡꢢꢣꢤꢡꢥ
ꢦꢚꢧꢨꢩ ꢩꢨꢪꢤꢚ ꢚꢛ ꢫꢠꢬꢜꢜ:ꢭꢮ ). Given that we know that (i) QDs reductively photocatalyze the
9
production of CO via the FeTPP catalyst (Figure 2), and (ii) eT from the QDs to FeIIITPP is
thermodynamically favorable (Figure 1B), whereas hole transfer (which would produce QDꢀꢀ
FeIVꢀTPP) and energy transfer (which requires spectral overlap between the emission of QDs and
the absorption of FeTPP) are not, we can confidently conclude that the quenching mechanism is
eT from QDs to FeTPP. The 90% yield for PL quenching indicates that eT from the QD to
FeTPP must be fast enough to outꢀcompete nearly all other decay mechanisms for the
photoexcited electron.
To measure the rate of the eT process for the initial sensitization step (FeIIITPP ꢀ FeIITPP),
we acquired transient absorption (TA) spectra of CuInS2 QDs with 0.5 eq. of FeIIITPP (blue), in
DMSO and purged with CO2, after monochromatic excitation of the system with an ultraꢀshort
pulse at 450 nm, Figure 3A. We used 0.5 eq. rather than 0.02 eq. of catalyst in this experiment
because more QDꢀcatalyst complexes afford more signal, and we are not concerned with the
yield of multiꢀelectron transfer, only the rate of the first eT step, from QD* to FeIII to form FeII.
The Supporting Information describes the features of the TA spectra of CuInS2 QDs in detail;
here we only say that since (i) the TA spectra of FeIIITPP purged with CO2, without QDs, have
no observable features in this region and time window (250 fs ꢀ 3 ns) (Figure 3A), (ii) the
ground state absorption spectra of QDs/CO2 and QDs/FeIIITPP/CO2 overlay exactly at 450 nm
(Figure 1A), and (iii) the amplitude of the ground state bleach feature (GSB, 470 nm – 510 nm)
is not sensitive to the excitonic hole46, we can use the dynamics of GSB of the QDs to selectively
monitor the dynamics of reducing electrons in the photoexcited QD. Furthermore, these
dynamics will only reflect charge separation, not charge recombination, processes involving the
QD.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3C (black) shows a kinetic trace extracted from the TA spectra of QD/CO2 samples at
a wavelength near the peak of the GSB of the QDs (480 nm). As summarized in Table S1, the
recovery of the GSB occurs with at least four distinct exponential processes, which we assign to
known excitonic processes of the QD. Figure 3C (red) shows a kinetic trace extracted from the
TA spectra of QD/FeIIITPP/CO2 sample at 480 nm. The QD/CO2 and QD/FeIIITPP/CO2 samples
were prepared using the same QD stock solution, and measured using the same excitation power
12
ACS Paragon Plus Environment

Page 13 of 18
Journal of the American Chemical Society
1
2
and pumpꢀprobe overlap. The only difference between the two samples is that one has FeIIITPP
and the other does not. Since FeIIITPP does not contribute to the TA spectrum in this region
(Figure 3A), any difference between the amplitudes or dynamics of the signals from the two
samples must be related to the response of the QD’s photoexcited electron to FeIIITPP. The
dynamics of the electron in the QD/ FeIIITPP /CO2 samples are very similar to those of the
QD/CO2 samples (see Table S1) with one major difference: the initial magnitude of the GSB (at
~200 fs after photoexcitation, the earliest time delay we can measure with our setup) is always
smaller upon addition of FeIIITPP. Given the identical conditions of the experiment, this initial
depletion of the excited state of the QD must be caused by eT from the QD to FeIIITPP within the
instrument response time (IRF~200 fs, Figure S7).47ꢀ50
This subꢀ200 fs eT process from the QD to FeIIITPP accounts for a 36% quenching of the QD
excited state when complexed with FeTPP, but recall that 45% of the PL of the QD ensemble is
quenched upon addition of FeTPP. The remaining 9% of the quenching can be accounted for
quantitatively by eT from QD* to FeIIITPP on a longer timescale (0.7 ns), which we detect by
timeꢀcorrelated single photon counting (TCSPC, Figure 3D and Table S2). We therefore
conclude that ~80% of the eT from QD* to FeIIITPP occurs within 200 fs, and the remaining
20% occurs within 0.7 ns. This dualꢀtimescale charge transfer is commonly reported in studies of
QDꢀmolecule complexes as molecules can have several binding geometries on the surface of
QDs that will result in different magnitudes of electronic coupling.32 The fact that both eT
processes occur in <1 ns supports our claim that the QD and FeIIITPP form a static complex,
even without covalent linking chemistry, because a diffusionꢀlimited process would occur no
faster than hundreds of nanoseconds at these concentrations.51
We also measured the rate of the second sensitization step, the photoreduction of FeIITPP to
FeITPP by the QDs, by mixing QDs with the catalyst preꢀreduced to FeII by cobaltocene. The
Supporting Information contains the details for this experiment, controls, and characterization of
FeIITPP. Addition of 0.5 eq. of FeIITPP again quenches the PL intensity of the QD with an
efficiency of ~90 % (Figure S4). TA (Figure 3E and Table S3) and TCSPC (Figure 3F and
Table S4) analyses of the QD/FeIITPP assembly reveals that 63% of eT from QD* to FeIITPP
occurs within 200 fs, 16% occurs in ~20 ps, and 21% occurs ~0.4 ns, respectively. Chemically
reduced FeITPP was too unstable to study eT dynamics in the QD/FeITPP assembly;52 however,
a simple Marcus theory analysis predicts that photoreduction of FeITPP should be slower than
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 18
1
2
photoreduction of FeIITPP by only a factor of 5, given the difference in reduction potentials of
FeITPP and FeIITPP, see the Supporting Information.
3
4
5
6
7
8
9
We also used TCSPC to determine that hole transfer from QDs to the sacrificial reductant
TMPD occurs in 0.7 ns, see Figure S6 and table S5.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CONCLUSIONS
In summary, colloidal CuInS2/ZnS QDs sensitize FeTPP to reduce CO2 to CO using 450ꢀnm
light with a sensitization efficiency, defined as the TON of CO per catalyst per absorbed unit of
photon energy, that is a factor of 18 greater than that of an analogous system with an Irꢀbased
organometallic sensitizer. The selectivity for CO is 84% (16% H2). The QD/FeTPP mixture is
stable after 40 hours of illumination at 4.5 mW. We have underscored, in this work, that design
of effective QDꢀsensitized catalytic systems, particularly those for multiꢀelectron transfer, must
emphasize not only the relative potentials of sensitizer and catalyst, but also their electronic
coupling. The efficiency of sensitization in our system originates in the ultrafast electron transfer
between the QD and FeTPP: the time constants for the dominant electron transfer processes
responsible for the first two sensitization steps (FeIIITPP ꢀ FeIITPP, and FeIITPP ꢀ FeITPP) are
both <200 fs. This ultrafast photoreduction is enabled by formation of a quasiꢀstatic QD/FeTPP
complex, probably driven by the interaction between Fe and sulfur on the surface of the QD. The
exact nature of this interaction will require further characterization. The ability of QDs to selfꢀ
assemble with a coꢀcatalyst is a crucial advantage of the QDꢀsensitized system over molecular
systems, where the rate of charge transfer is limited by occasional collisions between freely
diffusing species. Moreover, the methodology we present in this work – the utilization of timeꢀ
resolved optical spectroscopies to probe the dynamics of individual elementary charge transfer
step within the catalytic cycle – is critical for determining the rateꢀlimiting steps within
complicated catalytic processes.
Further optimization of our system, such as coupling the QD sensitizer with a more efficient
catalyst and tuning the sensitizer:catalyst ratio, will allow us to increase both the quantum
efficiency and selectivity of the system for CO. As outlined by Tachibana and coworkers53, the
excitation frequency (i.e., the rate of accumulating photoinduced electrons) is the bottleneck for
artificial photosynthesis. Slow rates of absorbing photons for chromophores and short lifetimes
of catalytic intermediates are two of the main reasons for low quantum yields in many
14
ACS Paragon Plus Environment

Page 15 of 18
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
photocatalytic systems. Designing a system where the chromophores quickly accumulate
photoinduced charge carriers and transport them to the catalytic centers is the key for an efficient
photocatalytic cell, and the nearꢀunity efficiency of electron transfer from the QD to FeTPP that
we achieve here is a good demonstration of this principle. It is also possible to partially substitute
or chemically dope CuInS2 to form heterostructures of CuInS2 and a proven water oxidation
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
54
catalyst such as CoO4 or BiVO4 , and thereby both eliminate sacrificial reductants and reduce
CO2 in aqueous media.
Ideally, one would modify the QD surface for highꢀaffinity adsorption of CO2 in an activated
geometry and thereby eliminate the coꢀcatalyst, as achieved by Ozin using hydrideꢀterminated Si
QDs exposed to gasꢀphase CO2.55 Direct, selective CO2 reduction by QDs is however still a work
in progress,56ꢀ62 and there is much to be learned and achieved by using QDs to deliver electrons
to a known, selective molecular catalyst.
SUPPORTING INFORMATION
Details of QD synthesis, ligand exchange procedures, catalysis setup, GC experiments, 1H NMR,
calculations of number of absorbed photons per hour, and timeꢀresolved optical spectroscopy
studies. This information is available free of charge via the internet at http://pubs.acs.org.
ACKNOWLEDGEMENT
This work was primarily supported as part of the ArgonneꢀNorthwestern Solar Energy Research
(ANSER) Center, an Energy Frontier Research Center funded by the U.S. Department of Energy
(DOE), Office of Science, Basic Energy Sciences (BES), under Award # DEꢀSC0001059
(sensitizer synthesis and catalysis studies), and by the National Science Foundation (NSF) under
Award # CHE ꢀ 1400596 (optical studies). The authors thank Dr. Marek Majewski and Prof.
Michael Wasielewski for the use of their GC instrument.
REFERENCES
(1)
(2)
Hu, Y.; Lee, C. C.; Ribbe, M. W. Science 2011, 333, 753.
Buss, J. A.; Agapie, T. Nature 2016, 529, 72.
Summerscales, O. T.; Cloke, F. G. N.; Hitchcock, P. B.; Green, J. C.; Hazari, N. Science
(3)
2006, 311, 829.
(4)
Ci, C.; Carbo, J. J.; Neumann, R.; Graaf, C. d.; Poblet, J. M. ACS Catal. 2016, 6, 6422.
15
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 18
1
2
3
4
5
6
7
8
9
(5)
Sahara, G.; Kumagai, H.; Maeda, K.; Kaeffer, N.; Artero, V.; Higashi, M.; Abe, R.;
Ishitani, O. J. Am. Chem. Soc. 2016, 138, 14152.
(6)
Nakajima, T.; Tamaki, Y.; Ueno, K.; Kato, E.; Nishikawa, T.; Ohkubo, K.; Yamazaki,
Y.; Morimoto, T.; Ishitani, O. J. Am. Chem. Soc. 2016, 138, 13818.
(7)
(8)
Zhu, G.; Li, Y.; Zhu, H.; Su, H.; Chan, S. H.; Sun, Q. ACS Catal. 2016, 6, 6294.
Gao, C.; Meng, Q.; Zhao, K.; Yin, H.; Wang, D.; Guo, J.; Zhao, S.; Chang, L.; He, M.;
Li, Q.; Zhao, H.; Huang, X.; Gao, Y.; Tang, Z. Adv. Mater. 2016, 28, 6485.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(9)
(10)
Froehlich, J. D.; Kubiak, C. P. J. Am. Chem. Soc. 2015, 137, 3565.
Thoi, V. S.; Kornienko, N.; Margarit, C. G.; Yang, P.; Chang, C. J. J. Am. Chem. Soc.
2013, 135, 14413.
(11)
(12)
(13)
(14)
(15)
(16)
Wang, D.; Huang, R.; Liu, W.; Sun, D.; Li, Z. ACS Catal. 2014, 4, 4254.
Rail, M. D.; Berben, L. A. J. Am. Chem. Soc. 2011, 133, 18577.
Costentin, C.; Drouet, S.; Robert, M.; Savéant, J.ꢀM. Science 2012, 338, 90.
Bonin, J.; Robert, M.; Routier, M. J. Am. Chem. Soc. 2014, 136, 16768.
Rao, H.; Bonin, J.; Robert, M. Chem. Commun. 2017, 53, 2830.
Kumar, B.; Llorente, M.; Froehlich, J.; Dang, T.; Sathrum, A.; Kubiak, C. P. Annu. Rev.
Phys. Chem. 2012, 63, 541.
(17)
(18)
(19)
Takeda, H.; Cometto, C.; Ishitani, O.; Robert, M. ACS Catal. 2017, 7, 70.
AriasꢀRotondo, D. M.; McCusker, J. K. Chem. Soc. Rev. 2016, 45, 5803.
Nakada, A.; Koike, K.; Nakashima, T.; Morimoto, T.; Ishitani, O. Inorg. Chem. 2015, 54,
1800.
(20)
Gholamkhass, B.; Mametsuka, H.; Koike, K.; Tanabe, T.; Furue, M.; Ishitani, O. Inorg.
Chem. 2005, 44, 2326.
(21)
Guo, Z.; Cheng, S.; Cometto, C.; AnxolabéhèreꢀMallart, E.; Ng, S.ꢀM.; Ko, C.ꢀC.; Liu,
G.; Chen, L.; Robert, M.; Lau, T.ꢀC. J. Am. Chem. Soc. 2016, 138, 9413.
(22)
(23)
Booth, M.; Brown, A. P.; Evans, S. D.; Critchley, K. Chem. Mater. 2012, 24, 2064.
Steinbacher, F. S.; Krause, R.; Hunze, A.; Winnacker, A. Phys. Status Solidi A 2012, 209,
340.
(24)
(25)
Lachish, U.; Infelta, P. P.; Grätzel, M. Chem. Phys. Lett. 1979, 62, 317.
Knowles, K. E.; Hartstein, K. H.; Kilburn, T. B.; Marchioro, A.; Nelson, H. D.; Whitham,
P. J.; Gamelin, D. R. Chem. Rev. 2016, 116, 10820.
(26)
Caldwell, M. A.; Albers, A. E.; Levy, S. C.; Pick, T. E.; Cohen, B. E.; Helms, B. A.;
Milliron, D. J. Chem. Commun. 2011, 47, 556.
(27)
135, 3383.
Tseng, H.ꢀW.; Wilker, M. B.; Damrauer, N. H.; Dukovic, G. J. Am. Chem. Soc. 2013,
(28)
(29)
Knauf, R. R.; Lennox, J. C.; Dempsey, J. L. Chem. Mater. 2016, 28, 4762.
Waiskopf, N.; BenꢀShahar, Y.; Galchenko, M.; Carmel, I.; Moshitzky, G.; Soreq, H.;
Banin, U. Nano Lett. 2016, 16, 4266.
(30)
Williams, E. S.; Major, K. J.; Tobias, A.; Woodall, D.; Morales, V.; Lippincott, C.;
Moyer, P. J.; Jones, M. J. Phys. Chem. C 2013, 117, 4227.
(31)
(32)
Su, F.; Lu, J.; Tian, Y.; Ma, X.; Gong, J. Phys. Chem. Chem. Phys. 2013, 15, 12026.
Harris, R. D.; Bettis Homan, S.; Kodaimati, M.; He, C.; Nepomnyashchii, A. B.;
Swenson, N. K.; Lian, S.; Calzada, R.; Weiss, E. A. Chem. Rev. 2016, 116, 12865.
(33) Chaudhary, Y. S.; Woolerton, T. W.; Allen, C. S.; Warner, J. H.; Pierce, E.; Ragsdale, S.
W.; Armstrong, F. A. Chem. Commun. 2012, 48, 58.
(34)
(35)
133, 10146.
(36)
Chai, Z.; Li, Q.; Xu, D. RSC Adv. 2014, 4, 44991.
MorrisꢀCohen, A. J.; Frederick, M. T.; Cass, L. C.; Weiss, E. A. J. Am. Chem. Soc. 2011,
Manzi, A.; Simon, T.; Sonnleitner, C.; Döblinger, M.; Wyrwich, R.; Stern, O.;
Stolarczyk, J. K.; Feldmann, J. J. Am. Chem. Soc. 2015, 137, 14007.
(37)
So, D.; Konstantatos, G. Chem. Mater. 2015, 27, 8424.
16
ACS Paragon Plus Environment

Page 17 of 18
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
(38)
Jin, S.; Tagliazucchi, M.; Son, H.ꢀJ.; Harris, R. D.; Aruda, K. O.; Weinberg, D. J.;
Nepomnyashchii, A. B.; Farha, O. K.; Hupp, J. T.; Weiss, E. A. J. Phys. Chem. C 2015, 119, 5195.
(39)
(40)
(41)
(42)
Praneeth, V. K. K.; Näther, C.; Peters, G.; Lehnert, N. Inorg. Chem. 2006, 45, 2795.
Peterson, M. D.; Jensen, S. C.; Weinberg, D. J.; Weiss, E. A. ACS Nano 2014, 8, 2826.
Bonin, J.; Chaussemier, M.; Robert, M.; Routier, M. ChemCatChem 2014, 6, 3200.
Chen, L.; Guo, Z.; Wei, X.ꢀG.; Gallenkamp, C.; Bonin, J.; AnxolabéhèreꢀMallart, E.;
Lau, K.ꢀC.; Lau, T.ꢀC.; Robert, M. J. Am. Chem. Soc. 2015, 137, 10918.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(43)
(44)
51, 7799.
(45)
Takeda, H.; Ohashi, K.; Sekine, A.; Ishitani, O. J. Am. Chem. Soc. 2016, 138, 4354.
Chan, S. L.ꢀF.; Lam, T. L.; Yang, C.; Yan, S.ꢀC.; Cheng, N. M. Chem. Commun. 2015,
RosasꢀHernandez, A.; Alsabeh, P. G.; Barsch, E.; Junge, H.; Ludwig, R.; Beller, M.
Chem. Commun. 2016, 52, 8393.
(46)
Li, L.; Pandey, A.; Werder, D. J.; Khanal, B. P.; Pietryga, J. M.; Klimov, V. I. J. Am.
Chem. Soc. 2011, 133, 1176.
(47)
10, 6372.
Lian, S.; Weinberg, D. J.; Harris, R. D.; Kodaimati, M. S.; Weiss, E. A. ACS Nano 2016,
(48)
Asbury, J. B.; Hao, E.; Wang, Y.; Ghosh, H. N.; Lian, T. J. Phys. Chem. B 2001, 105,
4545.
(49)
Hong, X.; Kim, J.; Shi, S.ꢀF.; Zhang, Y.; Jin, C.; Sun, Y.; Tongay, S.; Wu, J.; Zhang, Y.;
Wang, F. Nat Nano 2014, 9, 682.
(50)
Chan, W.ꢀL.; Ligges, M.; Jailaubekov, A.; Kaake, L.; MiajaꢀAvila, L.; Zhu, X. Y.
Science 2011, 334, 1541.
(51)
(52)
(53)
Knowles, K. E.; Malicki, M.; Weiss, E. A. J. Am. Chem. Soc. 2012, 134, 12470.
Grodkowski, J.; Behar, D.; Neta, P.; Hambright, P. J. Phys. Chem. A 1997, 101, 248.
Inoue, H.; Shimada, T.; Kou, Y.; Nabetani, Y.; Masui, D.; Takagi, S.; Tachibana, H.
ChemSusChem 2011, 4, 173.
(54)
Chang, X.; Wang, T.; Zhang, P.; Zhang, J.; Li, A.; Gong, J. J. Am. Chem. Soc. 2015, 137,
8356.
(55)
Sun, W.; Qian, C.; He, L.; Ghuman, K. K.; Wong, A. P. Y.; Jia, J.; Jelle, A. A.; O’Brien,
P. G.; Reyes, L. M.; Wood, T. E.; Helmy, A. S.; Mims, C. A.; Singh, C. V.; Ozin, G. A. Nat. Commun.
2016, 7, 12553.
(56)
(57)
129, 12090.
(58)
Zhou, R.; Guzman, M. I. J. Phys. Chem. C 2014, 118, 11649.
Kang, Z.; Tsang, C. H. A.; Wong, N.ꢀB.; Zhang, Z.; Lee, S.ꢀT. J. Am. Chem. Soc. 2007,
Sahu, S.; Liu, Y.; Wang, P.; Bunker, C. E.; Fernando, K. A. S.; Lewis, W. K.; Guliants,
E. A.; Yang, F.; Wang, J.; Sun, Y.ꢀP. Langmuir 2014, 30, 8631.
(59)
Fujiwara, H.; Hosokawa, H.; Murakoshi, K.; Wada, Y.; Yanagida, S.; Okada, T.;
Kobayashi, H. J. Phys. Chem. B 1997, 101, 8270.
(60)
Kauffman, D. R.; Ohodnicki, P. R.; Kail, B. W.; Matranga, C. J. Phys. Chem. Lett. 2011,
2, 2038.
(61)
(62)
Manthiram, K.; Beberwyck, B. J.; Alivisatos, A. P. J. Am. Chem. Soc. 2014, 136, 13319.
Chang, X.; Wang, T.; Gong, J. Energy Environ. Sci. 2016, 9, 2177.
17
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 18
1
2
3
4
5
6
TOC Figure
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
18
ACS Paragon Plus Environment