High-Throughput Screening of Earth-Abundant Water Reduction Catalysts toward Photocatalytic Hydrogen Evolution

Catalysts toward Photocatalytic Hydrogen Evolution

Article

High-Throughput Screening of Earth-Abundant Water Reduction

Rachel N. Motz, Eric M. Lopato, Timothy U. Connell, and Stefan Bernhard*

Cite This: Inorg. Chem. 2021, 60, 774 −781

Read Online

ACCESS

Metrics & More

Article Recommendations

sı Supporting Information

ABSTRACT: Noble-metal photosensitizers and water reduction co-catalysts (WRCs) still

present the highest activity in homogeneous photocatalytic hydrogen production. The

search for earth-abundant alternatives is usually limited by the time required to screen new

catalyst combinations; however, here, we utilize newly designed and developed high-

throughput photoreactors for the parallel synthesis of novel WRCs and colorimetric

screening of hydrogen evolution. This unique approach allowed rapid optimization of

photocatalytic water reduction using the organic photosensitizer Eosin Y and the

archetypal cobaloxime WRC [Co(GL1) pyCl], where GL1 is dimethylglyoxime and py is

pyridine. Subsequent combinatorial synthesis generated 646 unique cobalt complexes of

the type [Co(LL) pyCl], where LL is a bidentate ligand, that identiﬁed promising new

WRC candidates for hydrogen production. Density functional theory (DFT) calculations

performed on such cobaloxime derivative complexes demonstrated that reactivity depends

on hydride aﬃnity. Alkyl-substituted glyoximes were necessary for hydrogen production and showed increased activity when paired

with ligands containing strong hydrogen-bond donors.

INTRODUCTION

dyes, such as Eosin Y (EY) (Figure 1a), are signiﬁcantly

■

cheaper than metal-based phosphors and also exhibit photo-

Society is faced with an increasing demand for energy, leading

to a push for sustainable or renewable alternatives. Despite the

increasing implementation of electricity generated via solar,

hydro, and wind power, chemical fuels continue to rely on

carbon-based sources. Storage and transport of these feed-

stocks is critical in sectors like transportation and manufactur-

ing, but their use also leads to substantial emissions; 66% of

global greenhouse gas emissions result from industrial

processes, powering transportation, or producing heat and

1,12

physics amenable to both synthetic photochemistry

water reduction.

and

13−16

Co-catalysts with the highest turnover

also contain noble metals, especially colloidal platinum and

palladium, but they can also be replaced with earth-abundant

alternatives, particularly molecular complexes of cobalt, iron,

10,17

and nickel.

Speciﬁcally, there is a great variety in reported

cobalt complexes with ligand derivatives that include 2,2′-

bipyridine, glyoxime, dithiolene, macrocycles, and Schiﬀ

electricity. Hydrogen is an emerging alternative to replace

bases. The facile redox chemistry that allows cobalt to

fossil fuels; in addition to its high caloriﬁc value and sole

transition between its +3 and +1 oxidation states in single-

electron steps enables a variety of potential mechanistic

pathways; water is reduced either by monomolecular

protonation of cobalt(III or II) hydride or by a bimolecular

combustion product of water, it is also an important feedstock

in oil reﬁning and ammonia ﬁxation. It is therefore

unfortunate that hydrogen production is currently sustained

primarily by steam reforming, which is extremely carbon-

intensive: every metric ton of hydrogen produced results in

pathway, where two cobalt(III) hydrides react to evolve H

17−20

(

Figure 1b).

Multiple proposed mechanisms for this

5,20−30

almost seven metric tons of carbon dioxide.

catalyst moiety are still debated in the literature.

Photocatalytic water reduction presents a carbon-emission-

Computational work has also provided support for the

particular order of reaction intermediates and shed light on

free method for generating hydrogen. This homogeneous

process requires only three molecular components: a photo-

sensitizer to capture light and convert it to chemical energy, a

water reduction co-catalyst (WRC) to drive the redox reaction,

and an organic sacriﬁcial donor to supply electrons. The

majority of reported photocatalytic systems utilize noble-metal

photosensitizers, commonly made from iridium(III), rhenium-

potential rate-limiting steps for electrocatalytic water reduction

21,31

involving cobalt species.

Cobaloxime complexes that

Received: September 18, 2020

Published: January 7, 2021

−10

(

I), rhodium(III), ruthenium(II), and platinum(II).

While

these typically demonstrate high activity and eﬃciency, the

rarity, expense, and toxicity of noble metals necessitate the

search for more sustainable alternatives. Luminescent organic

2021 American Chemical Society

Inorg. Chem. 2021, 60, 774−781

Inorganic Chemistry

developed (Figure 2 ) in our laboratory: the organic dye Eosin Y

Article

computational costs and do not diminish the need for large

experimental datasets.

Expanding on our recent eﬀort to design more aﬀordable

high-throughput experimentation for photocatalytic water

43,44

reduction,

here, we report the ﬁrst application of parallel

synthesis and screening in a homogeneous catalytic system that

contains no noble-metal components. Newly designed and

developed photoreactors allowed rapid optimization of varied

reaction conditions including concentration of each compo-

nent, water fraction, and solution pH. Earth-abundant WRCs

based on the cobaloxime scaﬀold were synthesized in situ,

avoiding the need for traditional batch synthesis and enabling

the generation of a structurally diverse combinatorial library

with tunable activity. Computational modeling allowed the

exploration of the structure−activity relationship for bis-

glyoxime complexes and to determine the energy changes of

reaction steps that dictate catalytic activity. Hydrogen

production across this library was monitored using a

chemoselective, colorimetric tape and identiﬁed promising

new WRC candidates with higher reactivity than the parent

[

Co(GL1) pyCl] complex.

EXPERIMENTAL SECTION

Synthesis of Glyoxime Derivate Ligands. All reagents and

■

solvents were commercially sourced and used without further

puriﬁcation. Nuclear magnetic resonance (NMR) spectra were

obtained using a 500 Bruker Avance III or a 500 Bruker Avance

Figure 1. (a) Structure of Eosin Y in its neutral and photoactive

anionic form. (b) Reported mechanisms for light-driven water

reduction using Eosin Y and cobalt co-catalyst. Note that an oxidative

quenching pathway for Eosin Y involving Co species is often also

discussed. (c) Structure of cobaloxime water reduction co-catalyst.

Neo spectrometer ( H, 500 MHz; C, 125.8 MHz); spectra were

referenced to residual solvent peaks. Ligands GL11, NH1, NH2,

NH3, and NH4 and cobalt complexes [Co(GL1) pyCl] and

[

Co(GL1BF ) pyCl] were synthesized according to previously

2 2

reported procedures. Detailed synthetic procedures and character-

ization are provided in the Supporting Information.

Optimization Reactions. Homogeneous photocatalytic water

contain two bidentate dimethylglyoximes (GL1) and an axial

monodentate ligand (commonly pyridine, py), particularly

reduction reactions were performed in photoreactors designed and

[

Co(GL1) pyCl] and the diﬂuoroborylated derivative [Co-

(

GL1BF ) pyCl] (Figure 1c), are robust water reduction co-

catalysts known to evolve hydrogen in combination with

,32

15,16,30

noble-metal

and organic dye

photosensitizers. Some

variations of the glyoxime ligand structure have also been

tested including a formal tetradentate diimine-dioxime

6,33

architecture generated by alkyl bridging

and bioinspired

tethering of the glyoxime and photosensitizer moieties.

−

Optimization of a single reaction (2H + 2e → H ) has

predicated extensive ad hoc screening of new WRC

0,17

candidates;

however, the large variety in reported assay

conditions makes comparison of promising candidates

diﬃcult. Parallelized, high-throughput experimentation pro-

vides a standardized platform that enables the optimization of

multiple reaction parameters and rapid combinatorial screen-

ing for novel catalysts. These advantages, particularly along

with big data analysis, drive the continued uptake of a high-

throughput approach in ﬁelds as diverse as biocatalytic activity,

Figure 2. (a) White LED strips and camera attached to a plexiglass

cage for monitoring reaction progress; (b) 108 reaction vials (1 mL)

covered in H -sensitive colorimetric tape sealed with silicone and

36−38

plexiglass; (c) two 100 W blue LED chips (440 ± 10 nm).

pharmaceutical synthesis, and toxicology.

Combinatorial

synthesis has been applied to the discovery of new

heterogeneous photocatalysts including TiO nanoparticles

(EY) served as a photosensitizer in a mixture of 2-ethoxyethanol and

water with a large excess of triethanolamine (TEOA) as sacriﬁcial

reductant. Reactions were irradiated for at least 16 h using blue light-

emitting diodes (LEDs) (440 ± 10 nm), and progress was monitored

via time lapse photography (10 min intervals) of a commercially

available colorimetric tape sensitive to hydrogen. Reaction vials

contained a total solution volume of 650 μL. The change in color of

the hydrogen-sensitive tape was digitized using a Mathematica script

and plotted vs reaction time. The cobalt-based WRC [Co-

and conjugated organic polymers, but it is still limited by the

need for costly automated equipment and characterization

platforms. Computational screening does not require expensive

experimental infrastructure, and theoretical analyses of large

calculated property libraries have identiﬁed promising

candidates of nitride, oxynitride, and oxide photocatalysts for

41,42

heterogeneous water splitting.

These simulations, while

16,24,25

providing some predictive capability, still incur inherent

(GL1BF ) pyCl], selected for its stability and performance,

Inorg. Chem. 2021, 60, 774−781

Inorganic Chemistry

Article

Figure 3. (a) Optimal conditions for high-throughput photocatalytic water reduction to hydrogen gas; surface plots of maximum rates for (b) water

fraction vs WRC and (c) EY vs WRC; (d) series of WRC concentrations at water fractions demonstrating the dependence of incubation time on

the WRC content; and (e) pH dependence of hydrogen evolution, with error bars representing the standard deviation of four diﬀerent WRC

concentrations, and quadratic ﬁt R > 0.99.

was used to determine optimal reaction conditions, including water

fraction, concentration of each component, and pH. Optimal reaction

conditions were determined by the calculation of the maximum rate

of hydrogen production (Figure 3a).

Parallelized Catalyst Screening. Novel WRC complexes were

synthesized in situ and screened via our photoreactors. Appropriate

equivalents of cobalt chloride, ligand, and pyridine solutions in 2-

ethoxyethanol (total volume, 520 μL) were sequentially added in

aliquots to reaction vials and allowed to stand for 10 min.

Subsequently, EY in solution with 30% TEOA in water (130 μL)

was added via a pipette and photocatalytic reactions were monitored

by photography every 10 min for a total of 1000 min.

S2). The incubation time before hydrogen was detected ranged

from 200 to 900 min depending on sample conditions (see the

Supporting Information for details). The shortest incubation

times were observed at a relatively low WRC concentration

(

0.4−0.8 mM) regardless of the water fraction (Figure 3d),

and we speculate that this might represent the time required to

generate the active catalyst species (Figure 1b). Given the

complexity of the surfaces observed in these experiments

(Figure 3b,c), various combinations of proposed mechanisms

likely contribute to the activity diﬀerently based on reaction

conditions. The correlation between hydrogen production and

water-to-cobalt ratio revealed a similar parabolic relationship,

with minimal incubation times at ratios of 10 000−20 000

RESULTS AND DISCUSSION

■

(

Figure S2). Overall, limited hydrogen was produced at both

Optimization of Photocatalytic Water Reduction.

Homogeneous water reduction typically requires organic

solvents to ensure catalyst solubility but can result in a low

large and small molar equivalents of water; we rationalize the

lower activity at small molar ratios by the decreased likelihood

of catalyst/substrate interaction. Nevertheless, hydrogen was

produced even when the water content was increased up to

water content. The system developed here utilized a solvent

mixture of ethoxyethanol and water; ethoxyethanol served as a

proxy for ethanol, in which cobaloximes are typically

synthesized, but with a higher boiling point to avoid detection

issues. Although EY demonstrates solvatochromism, the

absorption of EY in water vs that in ethanol diﬀers

0.5, demonstrating a reduced reliance on organic solvents that

is more representative of commercially relevant conditions.

Water fraction was ﬁxed at 0.2 in further experiments as it

provided the highest activity across the broadest WRC

concentration range.

The WRC exhibited an eﬀective working range between 0.2

and 2 mM, assuming a ﬁxed EY concentration, and reached

peak activity at approximately 1 mM. Outside this range, little

to no hydrogen production was detected. Hydrogen produced

at low catalyst concentrations (<0.2 mM) is likely below the

system’s limit of detection. Conversely, at high concentrations

(2 mM), highly colored intermediates may limit light

2,53

minimally.

content and WRC concentration on hydrogen production

Figures 3b and S1). Lower concentrations of WRC performed

First, we tested the correlation between water

(

better with a smaller water fraction, and higher concentrations

were similarly optimal with a greater water content. Although

not the only factor responsible for strong performance,

conditions with the highest performance centered around a

catalyst/water ratio of approximately 1:7000 molecules (Figure

Inorg. Chem. 2021, 60, 774−781

Inorganic Chemistry

Article

Table 1. Representative Ligands (Grouped by Class) Screened in Heteroleptic [Co(LL) pyCl] Complexes, with Maximum H

Production in Micromoles for Reactions Run at Optimal Conditions (1 equiv CoCl , 2 equiv LL, 1 equiv py) for 1000 min

See the Supporting Information for a complete list of ligands used in this study and a comprehensive list of experimental results.

absorption and increase the likelihood of known bimolecular

that the optimized reaction parameters here are for the parent

complex. Some other Co WRCs are known to operate under

diﬀerent ideal conditions, particularly at lower pH.

Using the optimized reaction conditions, a uniform plate was

tested as a means of determining the well-to-well error of the

degradation pathways.

We next compared hydrogen

production as a relationship between WRC and EY

concentrations (Figures 3c and S3). Hydrogen formation

demonstrated a considerably higher dependence on WRC

rather than EY concentration, suggesting that turnover at the

cobalt co-catalyst was slower than photosensitization. Optimal

EY and WRC concentrations of 0.6 and 1.2 mM, respectively,

were selected at the greatest rate averaged across nonzero

concentrations of the inverse component.

detection system. The average generated H across a 96-well

plate was 16.0 ± 1.3 μ mol, corresponding to a well-to-well

error of 8.2% (Figure S5 ). Control reactions conﬁrmed that

photosensitizer, WRC, and light were all necessary to produce

hydrogen (Figure S6 ). Mercury poisoning tests in both

screening scale and 10× screening scale conﬁrmed that this

reaction proceeds without the formation or signi ﬁ cant catalytic

contribution of cobalt colloids (Figures S8 and S9 and Table

S3).

We also expected pH to aﬀect the performance of both the

WRC and the photosensitizer. The operation of cobaloxime

WRCs is pH-dependent, and the ﬂuorescence quantum yield

and lifetime of EY are the highest above pH 4 (EY pK =

1,55

2−

1,12

(

.8).

Indeed, the deprotonated, quinoid form of EY

Ligand Screening. We began screening for new molecular

cobalt WRCs using our optimized high-throughput platform.

Importantly, our reaction parameters are optimized for

homoleptic cobaloxime complexes and alternative Co WRCs

are known to perform better in diﬀerent solvents and at lower

pH. As other studies have varied the axial monodentate ligand

Figure 1a) is suggested to be the photoactive form.

Hydrogen evolution reactions were performed at varied pH,

modiﬁed by the addition of potassium hydroxide or hydro-

chloric acid, and displayed activity independent of WRC

concentration (Figures 3e and S4 ). Above an upper pH limit of

29,56,57

∼

10, no hydrogen was produced, while below pH 7.5, a

substantial precipitate, assigned to the protonated reductant

TEOA, pK = 7.76), formed. Small amounts of generated

in cobaloxime WRCs,

we instead focused on exploring

the eﬀect of equatorial ligands other than the common

,15,16,30,32,58,59

(

dimethylglyoxime.

Heteroleptic complexes of

hydrogen at pH 7.5 are consistent with previous reports that

the type [Co(LL) pyCl] were prepared in situ by combining

protonation of TEOA makes it an ineﬀective electron

aliquots of stock solutions containing bidentate ligands (LL),

6,32

donor.

The optimal calculated pH, 8.4, was only slightly

CoCl and pyridine as the consistent axial monodentate ligand.

lower than the inherent pH of prepared solutions (∼8.9 due to

the addition of TEOA) and was consequently unaltered to ease

high-throughput reaction preparation. It is notable, however,

We initially screened 46 diﬀerent bidentate ligands grouped

into discreet classes, including glyoximes (GL), bipyridines and

phenanthrolines (BP), 8-hydroxyquinolines (HQ), nitrogen

Inorg. Chem. 2021, 60, 774−781

Inorganic Chemistry

Article

heterocycles with hydrogen-bond donors (NH), and miscella-

neous ligands not readily sorted into another class (MC).

Perhaps unsurprisingly, glyoximes were signiﬁcantly more

active than any other ligand, producing the most hydrogen

over 1000 min (Tables 1 and S1). Alkyl functionalization, such

as that in the archetypal dimethylglyoxime (GL1), was

important for hydrogen production, while conjugated glyoxime

derivatives and most other ligands generated, at best, only trace

amounts of hydrogen.

stoichiometric corrections. This measure of hydride aﬃnity

was then compared to the hydrogen produced via the bis-

glyoxime complexes. An exponential curve was then ﬁtted to

the data (R = 0.92), revealing that more negative hydride

aﬃnity also results in greater hydrogen production (Figure 5).

Hydride Affinity = ΔH (Co(III)H(GL) py)

−

(ΔH (Co(I)(GL) py) + ΔH (H ))

). All complexes containing 2,2 ′-bipyridine (BP1 ) produced

Selected cobalt WRCs synthesized in batch were assessed for

hydrogen production by gas chromatography (Table S2) to

conﬁrm the validity of results obtained using complexes

synthesized in situ. A comparison of these results with high-

throughput screening revealed a close correlation between the

two techniques. For instance, the WRC containing the

deprotonated glyoxime ligand GL2 displayed 86% of the

activity compared to the analogous GL1 complex in batch and

displayed 90% activity in the high-throughput screen.

Homoleptic tris-complexes of BP1 and GL1 are reported as

17,54

successful WRCs;

complexes were being made in situ rather than the desired

Co(LL) pyCl] geometry. In situ prepared WRCs incorporat-

we also investigated whether these

[

ing either 2 or 3 equiv of GL1, GL2, and BP1 (and no

pyridine) were run in a large number of replicates and

compared with previously synthesized [Co(BP1) ]Cl (Figure

Figure 5. Hydride binding aﬃnity as determined through DFT

compared to catalytic production as measured through the parallel

reaction system. The ﬁtted curve is of the form y = a × b and

presents a correlation coeﬃcient of 0.925.

Knowing that glyoximes were necessary to achieve hydrogen

production, we next screened heteroleptic complexes of the

type [Co(GL)(LL)pyCl], that is, combining all 46 ligands with

each of the 16 glyoximes in an equimolar ratio (Figure 6).

Once again, signiﬁcant amounts of hydrogen were produced

only when the WRC contained at least one alkyl-substituted

glyoxime ligand (GL1, GL2, GL4, and GL6). Furthermore, in

most cases where the in situ WRC contained at least 1 equiv of

the four active glyoximes, hydrogen was produced independent

of the other ligand; only selected ligands (GL11, GL12, MC3,

and MC4) completely poisoned the WRC activity. We were

mindful that our high-throughput approach was not able to

discern the catalytically active species; kinetic or thermody-

namic factors may instead lead to appreciable amounts of

[Co(GL) pyCl] or [Co(LL) pyCl] rather than the desired

Figure 4. Hydrogen production over time of select cobalt WRCs.

Each trace represents the mean of 12 replicates. Inset: standard error

represented by the shaded region surrounding each trace. All samples

except for [Co(BP1) ]Cl were prepared in situ according to high-

throughput screening protocols.

negligible hydrogen, suggesting that diﬀerences in our

experimental design rendered this WRC inactive for hydrogen

production. While [Co(GL1) ] also produced hydrogen (13.4

[Co(GL)(LL)pyCl]. Our initial ligand screen revealed that

glyoximes were critical for hydrogen production, and we

consequently limited our hit selection to ligand combinations

that produced >50% of the hydrogen generated using the

μmol), reduced volumes compared to the heteroleptic complex

conﬁrmed that we were generating the expected WRC

(

[Co(GL1) pyCl]) under high-throughput conditions.

To further understand the diﬀerentiation in the function of

homoleptic glyoxime complexes, we performed DFT on several

known reaction intermediates using the B3LYP functional and

corresponding [Co(GL) pyCl] complex (Tables 1 and S1 ).

This ensured that we did not mistakenly assign activity

aﬀorded by a smaller amount of [Co(GL) pyCl] present in

-31G(d,p) basis set. The intermediates Co(II)H O(GL) py,

solution (up to a maximum of 0.5 equiv of expected cobalt

WRC).

Co(II)(GL) py, Co(III)H(GL) py, and Co(II)H (GL) py

were explored for all tested glyoximes expected to form a

ﬁve-member coordinating ring. It is notable that in these

calculations, doubly reduced Co(I)H O(GL) py does not

Heteroleptic complexes that contained 8-hydroxyquinolines

(HQ) achieved greater hydrogen evolution than this 50%

benchmark (of the corresponding bis-glyoxime WRC) when

combined with the active glyoxime ligands (GL1, GL2, GL4,

and GL6) except for the nitro-substituted derivative (HQ3).

The sulfoxyl-containing HQ2 was particularly active, generat-

ing 16.3 μmol over the tested time when combined with GL1,

which compares well to the 15.6 μmol observed with the bis-

converge, suggesting that water is not a stable ligand, and thus

reduction opens an active species for Co(I). This is consistent

with the literature. The free-energy change was estimated

according to eq 1 for the binding of a hydride, using the heat of

formation determined for each complex and any needed

Inorg. Chem. 2021, 60, 774−781

Inorganic Chemistry

The Supporting Information is available free of charge at

Article

moderate activity (14.4 and 13.7 μmol H , respectively) in

combination with select alkyl glyoximes.

Replicate reactions of [Co(GL1)(NH4)pyCl] and [Co-

GL1)(HQ2)pyCl] conﬁrmed that these ligand combinations

(

reproducibly evolved signiﬁcant amounts of hydrogen (Figure

S7 ). Interestingly, all well-performing ligands (above the 50%

benchmark) contain functional groups that serve as hydrogen-

bond donors. The catalytic cycle of cobaloximes is complex,

and both mono- and bimolecular pathways are reported;

however, it is plausible that hydrogen bonding in equatorial

glyoximes facilitates the proton-coupled electron transfer

required to form the cobalt hydride intermediate critical for

hydrogen evolution. Our results strengthen the argument

that hydrogen bonding plays an important role in facilitating

water reduction in cobalt-based WRCs. Electron donating axial

ligands, such as substituted pyridines and imidazoles, in bis-

glyoxime complexes also improve catalytic activity, rationalized

by an increased electron density and basicity of the protonated

50,55

intermediate.

The improved performance observed specif-

ically with monodentate imidazoles as the axial ligand is

notable in comparison to the nitrogen heterocycles present in

our NH ligand series, further strengthening the argument that

hydrogen bonding plays an important role in facilitating water

reduction in cobalt-based WRCs.

CONCLUSIONS

■

High-throughput experimentation enabled the rapid design

and optimization of a homogeneous photocatalytic water

reduction method without any noble-metal components. The

cobalt-based WRC [Co(GL1BF ) pyCl] demonstrated an

optimal catalyst/substrate ratio of 1:7000 molecules of water

and was eﬀective across a range of concentrations (0.2−2

mM). An initial screen of 46 diﬀerent coordinating ligands for

in situ prepared WRCs revealed that alkyl-substituted

glyoximes were critical for hydrogen production in our parallel

screening conditions. DFT simulations demonstrated a

connection between the hydride aﬃnity and the catalytic

performance for these bis-glyoxime species. A larger sub-

sequent combinatorial screen of each glyoxime derivative

paired with all other ligands (to give 616 distinct co-catalysts of

the type [Co(GL)(LL)pyCl]) identiﬁed several promising

ligand classes all containing functional groups that may act as

hydrogen-bond donors. Two distinct catalysts, [Co(GL1)-

(

HQ2)pyCl] and [Co(GL1)(NH4)pyCl], produced greater

amounts of hydrogen than the archetypal bis-glyoxime WRC

Co(GL1) pyCl]. Isolation and characterization of these

[

promising candidates, along with their mechanism for

producing hydrogen, are currently being investigated.

Figure 6. Combinatorial screen of [Co(GL)(LL)pyCl] co-catalysts

for maximum hydrogen production in micromoles for reactions run at

ASSOCIATED CONTENT

■

optimal conditions (1 equiv CoCl , 1 equiv GL, 1 equiv LL, 1 equiv

sı Supporting Information

py) for 1000 min.

https://pubs.acs.org/doi/10.1021/acs.inorgchem.0c02790.

Reagents used, synthetic procedures, and character-

ization for glyoxime derivative ligands and fully

synthesized cobalt glyoxime complexes; further descrip-

tion of photoreactor design and use; detailed exper-

imental information and results for optimization experi-

ments; results of control testing and uniform reactions;

full results from homoleptic ligand screening and

associated ligand structures; and experimental details

GL1 analogue (Figure 6). Ligands with nitrogen-containing

heterocycles capable of acting as hydrogen-bond donors also

performed above the benchmark in almost all cases when

combined with GL1, GL2, and GL4. The in situ formed

complex [Co(GL1)(NH4)pyCl] evolved the most hydrogen

of all of the screened compounds, 16.4 μmol (Figure 6). The

primary amine-containing ligands 2,3-diaminonaphthalene

(

MC5) and 8-aminoquinoline (MC6) also demonstrated

Inorg. Chem. 2021, 60, 774−781

Inorganic Chemistry

and gas chromatography analysis of larger-scale experi-

(10) Wang, M.; Na, Y.; Gorlov, M.; Sun, L. Light-driven hydrogen

Article

production catalysed by transition metal complexes in homogeneous

systems. Dalton Trans. 2009, 6458−6467.

ments (PDF)

Heteroleptic screening data (XLSX )

(11) Majek, M.; Filace, F.; von Wangelin, A. J. On the mechanism of

photocatalytic reactions with eosin Y. Beilstein J. Org. Chem. 2014, 10,

AUTHOR INFORMATION

Corresponding Author

Stefan Bernhard − Department of Chemistry, Carnegie Mellon

University, Pittsburgh, Pennsylvania 15213, United States;

orcid.org/0000-0002-8033-1453; Email: bern@cmu.edu

81−989.

■

(12) Majek, M.; von Wangelin, A. J. Mechanistic Perspectives on

Organic Photoredox Catalysis for Aromatic Substitutions. Acc. Chem.

Res. 2016, 49, 2316−2327.

13) Misawa, H.; Sakuragi, H.; Usui, Y.; Tokumaru, K. Photo-

sensitizing Action of Eosin y for Visible Light Induced Hydrogen

Evolution from Water. Chem. Lett. 1983 , 12 , 1021 −1024.

Authors

14) Li, D.; Wehrung, J.-F.; Zhao, Y. Gold nanoparticle-catalysed

Rachel N. Motz − Department of Chemistry, Carnegie Mellon

University, Pittsburgh, Pennsylvania 15213, United States

Eric M. Lopato − Department of Chemistry, Carnegie Mellon

University, Pittsburgh, Pennsylvania 15213, United States

Timothy U. Connell − Department of Chemistry, Carnegie

Mellon University, Pittsburgh, Pennsylvania 15213, United

States; orcid.org/0000-0002-6142-3854

photosensitized water reduction for hydrogen generation. J. Mater.

Chem. A 2015, 3, 5176−5182.

(15) Lazarides, T.; McCormick, T.; Du, P.; Luo, G.; Lindley, B.;

Eisenberg, R. Making Hydrogen from Water Using a Homogeneous

System Without Noble Metals. J. Am. Chem. Soc. 2009, 131, 9192−

9194.

(16) Wang, Z. Y.; Rao, H.; Deng, M. F.; Fan, Y. T.; Hou, H. W.

Photocatalytic H ₂ generation based on noble-metal-free binuclear

cobalt complexes using visible-light. Phys. Chem. Chem. Phys. 2013,

Complete contact information is available at:

https://pubs.acs.org/10.1021/acs.inorgchem.0c02790

17) Du, P.; Eisenberg, R. Catalysts made of earth-abundant

5, 16665−16671.

Notes

elements (Co, Ni, Fe) for water splitting: Recent progress and future

The authors declare no competing ﬁnancial interest.

challenges. Energy Environ. Sci. 2012, 5, 6012−6021.

(18) Call, A.; Franco, F.; Kandoth, N.; Fernandez, S.; Gonzalez-

ACKNOWLEDGMENTS

S.B. gratefully acknowledges National Science Foundation

NSF) funding CHE 1764353 for this work. Seed funding for

■

(

Bejar, M.; Perez-Prieto, J.; Luis, J. M.; Lloret-Fillol, J. Understanding

light-driven H evolution through the electronic tuning of amino-

pyridine cobalt complexes. Chem. Sci. 2018, 9, 2609 −2619.

19) Li, C.-B.; Gong, P.; Yang, Y.; Wang, H.-Y. Cobalt(II)−Salen

the development of the high-throughput reactors came from

Carnegie Mellon University’s Kavcic-Moura Fund and

Manufacturing Futures Initiative. E.M.L. was supported by

the Steinbrenner Institute Graduate Fellowship and the ARCS

Foundation Pittsburgh Chapter. T.U.C. was supported by a

Fulbright Future Scholarship. Seed code used to generate the

image analysis program was provided by Tomasz Kowalewski.

Complexes for Photocatalytic Hydrogen Production in Noble Metal-

Free Molecular Systems. Catal. Lett. 2018, 148, 3158−3164.

(20) Dempsey, J. L.; Brunschwig, B. S.; Winkler, J. R.; Gray, H. B.

Hydrogen Evolution Catalyzed by Cobaloximes. Acc. Chem. Res. 2009,

2, 1995−2004.

21) Solis, B. H.; Hammes-Schiffer, S. Theoretical analysis of

mechanistic pathways for hydrogen evolution catalyzed by cobalox-

imes. Inorg. Chem. 2011, 50, 11252−11262.

REFERENCES

(22) Muckerman, J. T.; Fujita, E. Theoretical studies of the

mechanism of catalytic hydrogen production by a cobaloxime. Chem.

Commun. 2011, 47, 12456−12458.

■

(

1) Climate Change 2014: Mitigation of Climate Change. Contribution

of Working Group III to the Fifth Assessment Report of the

Intergovernmental Panel on Climate Change; IPCC: Cambridge, UK

and New York, NY, 2014.

(23) Artero, V.; Saveant, J. M. Toward the Rational Benchmarking of

Homogeneous H -Evolving Catalysts. Energy Environ. Sci. 2014, 7,

(

808−3814.

Fontecave, M.; Field, M. J.; Artero, V. A computational study of the

2) Ogden, J. M. Prospects For Building A Hydrogen Energy

Infrastructure. Annu. Rev. Energy Environ. 1999, 24, 227−279.

3) Technical Support Document For Hydrogen Production: Proposed

24) Bhattacharjee, A.; Andreiadis, E. S.; Chavarot-Kerlidou, M.;

mechanism of hydrogen evolution by cobalt(diimine-dioxime)

catalysts. Chem. − Eur. J. 2013, 19, 15166−15174.

(

Rule For Mandatory Reporting of Greenhouse Gases; Oﬃce of Air and

Radiation, US EPA, 2008.

(25) Dempsey, J. L.; Winkler, J. R.; Gray, H. B. Mechanism of H

4) McDaniel, N. D.; Bernhard, S. Solar fuels: thermodynamics,

candidates, tactics, and figures of merit. Dalton Trans. 2010, 39,

0021−10030.

5) Tachibana, Y.; Vayssieres, L.; Durrant, J. R. Artificial photosyn-

thesis for solar water-splitting. Nat. Photonics 2012, 6, 511−518.

6) Curtin, P. N.; Tinker, L. L.; Burgess, C. M.; Cline, E. D.;

Evolution from a Photogenerated Hydridocobaloxime. J. Am. Chem.

Soc. 2010, 132, 16774−16776.

(26) Jacques, P.-A.; Artero, V.; Pecaut, J.; Fontecave, M. Cobalt and

nickel diimine-dioxime complexes as molecular electrocatalysts for

hydrogen evolution with low overvoltages. Proc. Natl. Acad. Sci. U.S.A.

2009, 106, 20627−20632.

(27) Chen, J.; Sit, P. H. Density Functional Theory and Car-

Parrinello Molecular Dynamics Study of the Hydrogen-Producing

Bernhard, S. Structure-activity correlations among iridium(III)

photosensitizers in a robust water-reducing system. Inorg. Chem.

7) Fihri, A.; Artero, V.; Pereira, A.; Fontecave, M. Efficient H ₂ -

009, 48, 10498−10506.

Mechanism of the Co(dmgBF ) and Co(dmgH) Cobaloxime

2 2 2

producing photocatalytic systems based on cyclometalated iridium-

and tricarbonylrhenium-diimine photosensitizers and cobaloxime

catalysts. Dalton Trans. 2008, 5567−5569.

Complexes in Acetonitrile-Water Solvent. J. Phys. Chem. A 2017,

121, 3515−3525.

(28) Jiang, Y.-K.; Liu, J.-H. DFT studies of cobalt hydride

(

8) Chirdon, D. N.; Transue, W. J.; Kagalwala, H. N.; Kaur, A.;

intermediate on cobaloxime-catalyzed H evolution pathways. Int. J.

Maurer, A. B.; Pintauer, T.; Bernhard, S. [Ir(N ∧N ∧N)(C ∧N)L] : a

new family of luminophores combining tunability and enhanced

photostability. Inorg. Chem. 2014, 53, 1487−1499.

Quantum Chem. 2012, 112, 2541−2546.

(29) Razavet, M.; Artero, V.; Fontecave, M. Proton Electroreduction

Catalyzed by Cobaloximes: Functional Models for Hydrogenases.

Inorg. Chem. 2005, 44, 4786−4795.

(9) Mills, I. N.; Kagalwala, H. N.; Chirdon, D. N.; Brooks, A. C.;

Bernhard, S. New Ir(III) 4,4 ′-dicyano-2,2 ′-bipyridine photosensitizers

for solar fuel generation. Polyhedron 2014, 82, 104−108.

(30) Zhang, P.; Wang, M.; Dong, J.; Li, X.; Wang, F.; Wu, L.; Sun, L.

Photocatalytic Hydrogen Production from Water by Noble-Metal-

Inorg. Chem. 2021, 60, 774−781

Inorganic Chemistry

Free Molecular Catalyst Systems Containing Rose Bengal and the

Article

Cobaloximes of BFx-Bridged Oxime Ligands. J. Phys. Chem. C 2010,

thiocyanate-free ruthenium(II) sensitizers with different pyrid-2-yl

tetrazolate ligands for dye-sensitized solar cells. Dalton Trans. 2015,

44, 11788−11796.

14, 15868−15874.

31) Solis, B. H.; Hammes-Schiffer, S. Substituent effects on cobalt

diglyoxime catalysts for hydrogen evolution. J. Am. Chem. Soc. 2011,

33, 19036−19039.

32) Du, P.; Knowles, K.; Eisenberg, R. A Homogeneous System for

Y.; Luo, Y. Discovery of a Teraryl Oxazolidinone Compound (S)-N-

(50) Yang, T.; Chen, G.; Sang, Z.; Liu, Y.; Yang, X.; Chang, Y.;

Long, H.; Ang, W.; Tang, J.; Wang, Z.; Li, G.; Yang, S.; Zhang, J.; Wei,

((3-(3-Fluoro-4-(4-(pyridin-2-yl)-1H-pyrazol-1-yl)phenyl)-2-oxooxa-

zolidin-5 -yl)methyl)acetamide Phosphate as a Novel Antimicrobial

Agent with Enhanced Safety Profile and Efficacies. J. Med. Chem.

2015, 58, 6389−6409.

the Photogeneration of Hydrogen from Water Based on a

Platinum(II) Terpyridyl Acetylide Chromophore and a Molecular

Cobalt Catalyst. J. Am. Chem. Soc. 2008, 130, 12576−12577.

(33) Kwok, C. L.; Cheng, S. C.; Ho, P. Y.; Yiu, S. M.; Man, W. L.;

(51) Schrauzer, G. N. Bis(dimethylglyoximato)Cobalt Complexes.

In Inorganic Syntheses; Jolly, W. L., Ed.; McGraw-Hill Book Company,

Inc.: New York, NY, 1968; Vol. XI, pp 61−70.

Au, V. K.; Tsang, P. K.; Leung, C. F.; Ko, C. C.; Robert, M. Precious-

metal free photocatalytic production of an NADH analogue using

cobalt diimine-dioxime catalysts under both aqueous and organic

conditions. Chem. Commun. 2020, 7491−7494.

(52) Hadjmohammadi, M. R.; Chaichi, M. J.; Yousefpour, M.

Solvatochromism Effect of Different Solvents on UV-Vis Spectra of

Flouresceine and its Derivatives. Iran. J. Chem. Chem. Eng. 2008, 27,

9−14.

(34) Dolui, D.; Khandelwal, S.; Majumder, P.; Dutta, A. The odyssey

of cobaloximes for catalytic H production and their recent revival

(53) Chakraborty, M.; Panda, A. K. Spectral behaviour of eosin Y in

with enzyme-inspired design. Chem. Commun. 2020 , 56 , 8166 −8181.

different solvents and aqueous surfactant media. Spectrochim. Acta,

Part A 2011, 81, 458−465.

(35) Cao, S.; Piao, L. Considerations for a More Accurate Evaluation

Method for Photocatalytic Water Splitting. Angew. Chem., Int. Ed.

020, 59, 18312−18320.

36) Cawse, J. N. Experimental Strategies for Combinatorial and

High-Throughput Materials Development. Acc. Chem. Res. 2001, 34,

13−221.

37) Kumar, R. A.; Clark, D. S. High-throughput screening of

(54) Goldsmith, J. I.; Hudson, W. R.; Lowry, M. S.; Anderson, T. H.;

Bernhard, S. Discovery and High-Throughput Screening of

Heteroleptic Iridium Complexes for Photoinduced Hydrogen

Production. J. Am. Chem. Soc. 2005, 127, 7502 −7510.

(55) Slyusareva, E. A.; Gerasimova, M. A. pH-Dependence of the

Absorption and Fluorescent Properties of Fluorone Dyes in Aqueous

biocatalytic activity: applications in drug discovery. Curr. Opin. Chem.

Biol. 2006, 10, 162−168.

Solutions. Russ. Phys. J. 2014, 56, 1370−1377.

(56) Panagiotopoulos, A.; Ladomenou, K.; Sun, D.; Artero, V.;

(38) Krska, S. W.; DiRocco, D. A.; Dreher, S. D.; Shevlin, M. The

Coutsolelos, A. G. Photochemical hydrogen production and

cobaloximes: the influence of the cobalt axial N-ligand on the system

stability. Dalton Trans. 2016, 45, 6732 −6738.

Evolution of Chemical High-Throughput Experimentation To

Address Challenging Problems in Pharmaceutical Synthesis. Acc.

Chem. Res. 2017, 50, 2976−2985.

(57) Wakerley, D. W.; Reisner, E. Development and understanding

(39) Nursam, N. M.; Wang, X.; Caruso, R. A. High-Throughput

of cobaloxime activity through electrochemical molecular catalyst

Synthesis and Screening of Titania-Based Photocatalysts. ACS Comb.

screening. Phys. Chem. Chem. Phys. 2014 , 16 , 5739 −5746.

Sci. 2015, 17, 548−569.

(58) Connolly, P.; Espenson, J. H. Cobalt-Catalyzed Evolution of

(40) Bai, Y.; Wilbraham, L.; Slater, B. J.; Zwijnenburg, M. A.; Sprick,

Molecular Hydrogen. Inorg. Chem. 1986 , 25 , 2684 −2688.

R. S.; Cooper, A. I. Accelerated Discovery of Organic Polymer

Photocatalysts for Hydrogen Evolution from Water through the

Integration of Experiment and Theory. J. Am. Chem. Soc. 2019, 141,

(59) Bakac, A.; Espenson, J. H. Unimolecular and Bimolecular

Homolytic Reactions of Organochromium and Organocobalt

Complexes. Kinetics and Equilibria. J. Am. Chem. Soc. 1984, 106,

41) Sawada, K.; Nakajima, T. High-throughput screening of

063−9071.

(

197−5202.

60) Bhattacharjee, A.; Chavarot-Kerlidou, M.; Dempsey, J. L.; Gray,

perovskite oxynitride and oxide materials for visible-light photo-

catalysis. APL Mater. 2018, 6, No. 101103.

H. B.; Fujita, E.; Muckerman, J. T.; Fontecave, M.; Artero, V.;

Arantes, G. M.; Field, M. J. Theoretical modeling of low-energy

electronic absorption bands in reduced cobaloximes. ChemPhysChem

(42) Wu, Y.; Lazic, P.; Hautier, G.; Persson, K.; Ceder, G. First

principles high throughput screening of oxynitrides for water-splitting

2014, 15, 2951−2958.

photocatalysts. Energy Environ. Sci. 2013, 6, 157−168.

(43) Lopato, E. M.; Eikey, E. A.; Simon, Z. C.; Back, S.; Tran, K.;

Lewis, J.; Kowalewski, J. F.; Yazdi, S.; Kitchin, J. R.; Ulissi, Z. W.;

Millstone, J. E.; Bernhard, S. Parallelized Screening of Characterized

and DFT-Modeled Bimetallic Colloidal Cocatalysts for Photocatalytic

Hydrogen Evolution. ACS Catal. 2020, 10, 4244−4252.

(44) Song, W.; Lopato, E. M.; Bernhard, S.; Salvador, P. A.; Rohrer,

G. S. High-throughput measurement of the influence of pH on

hydrogen production from BaTiO /TiO core/shell photocatalysts.

Appl. Catal., B 2020, 269, No. 118750.

45) Renom-Carrasco, M.; Lefort, L. Ligand libraries for high

throughput screening of homogeneous catalysts. Chem. Soc. Rev. 2018,

7, 5038−5060.

46) Chugunova, E.; Samsonov, V.; Akylbekov, N.; Mazhukin, D.

Synthesis of 2H-benzimidazole 1,3-dioxides, separase inhibitors, by

reaction of o-benzoquinone dioximes with ketones. Tetrahedron 2017,

47) Bookser, B. C. 2-Benzyloxymethyl-5-(tributylstannyl)tetrazole.

3, 3986−3992.

A reagent for the preparation of 5-aryl- and 5-heteroaryl-1H-tetrazoles

via the Stille reaction. Tetrahedron Lett. 2000, 41 , 2805 −2809.

48) Lan, N. M.; Burgard, R.; Wentrup, C. Rearrangement of 2-

Quinolyl- and 1-Isoquinolylcarbenes to Naphthylnitrenes. J. Org.

Chem. 2004, 69, 2033−2036.

(49) Colombo, A.; Dragonetti, C.; Magni, M.; Meroni, D.; Ugo, R.;

Marotta, G.; Grazia Lobello, M.; Salvatori, P.; De Angelis, F. New