Unexpected Roles of Triethanolamine in the Photochemical Reduction of CO2 to Formate by Ruthenium Complexes

Article abstract of DOI:10.1021/jacs.9b11897

A series of 4,4′-dimethyl-2,2′-bipyridyl ruthenium complexes with carbonyl ligands were prepared and studied using a combination of electrochemical and spectroscopic methods with infrared detection to provide structural information on reaction intermediates in the photochemical reduction of CO₂ to formate in acetonitrile (CH₃CN). An unsaturated 5-coordinate intermediate was characterized, and the hydride-transfer step to CO₂ from a singly reduced metal-hydride complex was observed with kinetic resolution. While triethanolamine (TEOA) was expected to act as a proton acceptor to ensure the sacrificial behavior of 1,3-dimethyl-2-phenyl-2,3-dihydro-1H-benzo[d]imidazole as an electron donor, time-resolved infrared measurements revealed that about 90% of the photogenerated one-electron reduced complexes undergo unproductive back electron transfer. Furthermore, TEOA showed the ability to capture CO₂ from CH₃CN solutions to form a zwitterionic alkylcarbonate adduct and was actively engaged in key catalytic steps such as metal-hydride formation, hydride transfer to CO₂ to form the bound formate intermediate, and dissociation of formate ion product. Collectively, the data provide an overview of the transient intermediates of Ru(II) carbonyl complexes and emphasize the importance of considering the participation of TEOA when investigating and proposing catalytic pathways.

Full text of DOI:10.1021/jacs.9b11897

Subscriber access provided by BIU Pharmacie | Faculté de Pharmacie, Université Paris V
Article
Unexpected Roles of Triethanolamine in the Photochemical
Reduction of CO2 to Formate by Ruthenium Complexes
Renato N. Sampaio, David C. Grills, Dmitry E. Polyansky, David J. Szalda, and Etsuko Fujita
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.9b11897 • Publication Date (Web): 27 Dec 2019
Downloaded from pubs.acs.org on December 29, 2019
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 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 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.
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 36
Journal of the American Chemical Society
1
2
3
4
5
Unexpected Roles of Triethanolamine in the Photochemical Reduction of CO2 to Formate
by Ruthenium Complexes
6
7
8
9
Renato N. Sampaio,†,* David C. Grills,† Dmitry E. Polyansky,† David J. Szalda,§ Etsuko Fujita†,*
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
†Chemistry Division, Brookhaven National Laboratory, Upton, NY 11973-5000, USA
§Department of Natural Science, Baruch College, CUNY, New York, NY 10010, USA
Abstract
A series of 4,4ʹ-dimethyl-2,2'-bipyridyl ruthenium complexes with carbonyl ligands were prepared and
studied using a combination of electrochemical and spectroscopic methods with infrared detection to
provide structural information on reaction intermediates in the photochemical reduction of CO2 to formate
in acetonitrile (CH3CN). An unsaturated 5-coordinate intermediate was characterized, and the hydride
transfer step to CO2 from a singly-reduced metal hydride complex was observed with kinetic resolution.
While triethanolamine (TEOA) was expected to act as a proton acceptor to ensure the sacrificial behavior
of 1,3-dimethyl-2-phenyl-2,3-dihydro-1H-benzo[d]imidazole (BIH) as an electron donor, time-resolved
infrared (TRIR) measurements revealed that about 90% of the photogenerated one-electron reduced
complexes undergo unproductive back electron-transfer. Furthermore, TEOA showed the ability to
capture CO2 from CH3CN solutions to form a Zwitterionic alkylcarbonate adduct, and was actively
engaged in key catalytic steps such as the metal hydride formation, the hydride transfer to CO2 to form
the bound formate intermediate and the dissociation of formate ion product. Collectively, the data provide
an overview on transient intermediates of Ru(II) carbonyl complexes, and emphasize the importance of
considering the participation of TEOA when investigating and proposing catalytic pathways.
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 36
1
2
3
4
INTRODUCTION
5
6
7
8
The catalytic reduction of CO2 to formate is one of the many routes in which CO2 can be converted
into high energy carbon-based products. Molecular catalysts that are capable of carrying out this reaction
have promoted a continuous momentum of research activity, which has led to an extensive catalog of
metal complexes that have been investigated over the years using electrochemical and/or photochemical
approaches.1-2 While these metal complexes are different in their composition, which controls their
physical and chemical properties, they often share a very common condition when employed in
photochemical assays for the reduction of CO2 to formate – the use of triethanolamine (TEOA).1-3
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
Light-driven CO2 reduction catalysis requires chromophores to harvest light to create excited states
that in some cases may be potent enough to transfer high-energy electrons to activate catalysts that are in
solution or covalently linked to the chromophore.1-2, 4 This process occurs in kinetic competition with
recombination pathways that often limit the overall efficiency of photoinduced electron transfer and the
following catalytic reactions. Hence, to maximize the desired reactions, sacrificial electron donors have
been routinely exploited.1-3 TEOA is oxidized irreversibly around 0.5 V vs NHE and is highly soluble in
aqueous and organic solvents, making it a suitable sacrificial reductive quencher for the excited states of
chromophores to generate strongly reducing chromophore radicals.3 In other cases, TEOA is used as a
simple Brønsted base to ensure the irreversibility of another, more potent, sacrificial electron donor, by
rapidly deprotonating the oxidized donor.1, 3 Unfortunately, much of the published research ignores the
possibility of secondary roles of TEOA in proposed catalytic mechanisms. In a few rare examples, TEOA
was found to bind to metal centers of catalysts and to assist in the capture of CO2 in the ground state;5-7
however, there have been no substantial indications of whether these findings are more general or if this
type of reactivity is a major contribution to the multiple steps of a catalytic cycle.
Chart 1. Molecular structures of the Ruthenium(II)-carbonyl complexes of the general, abbreviated

formula, Ru(CO)(X)n+ studied here. All charged complexes were isolated as the PF6 salt.
Herein, we investigate a series of Ru(II) carbonyl complexes, Chart 1, to map out the catalytic
pathways and transient intermediates in the photochemical reduction of CO2 to formate in acetonitrile
2
ACS Paragon Plus Environment

Page 3 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
(CH3CN). This class of catalysts has been studied for over 30 years, with much of the chemistry reported
by different groups,2, 8-15 and yet the investigation of their catalytic intermediates is mostly restricted to
steady-state observations with no kinetic information. The carbonyl ligand remains bound and unreactive
and serves as an infrared (IR) spectroscopic reporter to identify multiple catalyst forms and their redox
states generated by electrochemical and photochemical methods. While the results presented here establish
some general conclusions, they also uncover the big picture of photochemical assays that employ TEOA.
In CH3CN solutions, TEOA fails to efficiently perform its intended task of ensuring the sacrificial
behavior of electron donors such as 1,3-dimethyl-2-phenyl-2,3-dihydro-1H-benzo[d]imidazole (BIH).
Time-resolved infrared (TRIR) spectroscopy has revealed that the photochemically generated one-
electron reduced species mostly undergo charge recombination within a few microseconds, while only
about 10% persist on longer time scales. Unexpected findings also demonstrated that TEOA partakes in
critical supporting reactions including the zwitterionic capture of CO2, and the assistance in metal hydride
formation and formate product dissociation. TEOA also provides an alternative mechanism in which the
hydride is transferred to CO2 to form the bound formato complex, with a 106 rate enhancement compared
to in the absence of TEOA. The importance of these findings for photochemical CO2 reduction is
discussed.
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
EXPERIMENTAL METHODS
Materials and Synthesis. Acetonitrile (CH3CN, Aldrich, anhydrous, 99.8%), triethanolamine (TEOA,
Aldrich, purris. p.a., ≥ 99%), and triethylamine (TEA, Aldrich, ≥ 99%) were dried over activated
molecular sieves (Aldrich – 3 Å, beads, 8-12 mesh). Tetrabutylammonium formate salt (TBAHCOO) was
prepared according to a reported procedure16 and dried under reduced pressure. The Ru(II)-carbonyl
complexes,
[Ru(dmb)2(CO)(CO)](PF6)2,
[Ru(dmb)2(CO)(CH3CN)](PF6)2,
[Ru(dmb)2(CO)(HCOO)](PF6), and [Ru(dmb)2(CO)(H)](PF6), where dmb = 4,4-dimethyl-2,2-
bipyridine,17-20 and the chromophore [Ru(bpy-OMe)3](PF6)2 (where bpy-OMe is 4,4-dimethoxy-2,2-
bipyridine),21 were prepared according to published methods. All Ru(II)-carbonyl complexes are
abbreviated without (dmb)2, i.e., Ru(CO)(CO)2+, etc. All chemicals were stored in an inert N2-atmosphere
1
13
glove box before use. H and C NMR spectra were recorded on a Bruker Avance III spectrometer
operating at 400.16 and 100.62 MHz, respectively.
Preparation of [Ru(dmb)2(CO)(CH3CN-TEOA)](PF6)2. 50 mg of [Ru(dmb)2(CO)(CH3CN)]2+ was
dissolved in 25 mL of acetone containing 2 mL of TEOA. The solution mixture was stirred overnight at
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 36
1
2
3
4
5
6
7
8
9
30º C. The initially colorless solution turned dark orange. A precipitate, obtained by adding saturated
aqueous NH4PF6 and cold water, was collected by filtration, and dried under vacuum. Single crystals of
[Ru(dmb)2(CO)(CH3CN)](PF6)2 and [Ru(dmb)2(CO){NHC(CH3)-OC2H4)}N{(C2H4OH)2](PF6)2 (i.e.,
Ru(CO)(CH3CN-TEOA)2+) for X-ray analysis were grown from slow ether diffusion into acetone.
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
Preparation of (BIH). In a round bottom flask, 7 g (36.0 mmol) of 2-phenylbenzimidazole (Aldrich, 97%)
was dissolved in 300 mL of acetone containing 3 g (53.5 mmol, 1.5 equiv.) of potassium hydroxide (KOH,
Aldrich). To this solution, 6.15 g (43.3 mmol, 1.2 equiv.) of methyliodide (Aldrich, ReagentPlus, ≥ 99.5%)
was added and the mixture was heated at 50º C for 12 hours. Acetone was removed by rotary evaporation.
The resulting oil-like mixture was re-dissolved in 100 mL of ethyl acetate and purified by solvent
extraction in 300 mL of Milli-Q water. This process was repeated until the aqueous layer became colorless.
After removal of ethyl acetate by rotary evaporation, 2-phenyl-1-methylbenzimidazole was obtained as
an oil and confirmed by 1H-NMR (400 MHz, DMSO-d6, ): 7.86 (d, J = 6 Hz, 2H), 7.69 (d, J = 8 Hz, 1H),
7.64–7.52 (m, 4H), 7.33–7.22 (m,2H), 3.88 (s, 3H). The second methylation step was performed by
dissolving the oil product in 300 mL of ethyl acetate, after which 7.7 g (54.2 mmol, 1.5 equiv. relative to
starting material) of methyliodide was added. This solution was refluxed at 80º C for 12 hours. The
precipitate in the reaction mixture was filtered and washed with ethyl acetate multiple times. This solid,
2-phenyl-1,3-dimethylbenzimidazolium iodide, was characterized by 1H-NMR(400 MHz, DMSO-d6, ):
8.14 (dd, J = 6.4, 3.1 Hz, 2H), 7.9 (d, J = 7.6 Hz, 2H), 7.88–7.73 (m, 5H), 3.89 (s, 6H). The solid was
dissolved in 300 mL of methanol, cooled in an ice bath, and then NaBH4 (6.8 g ,179.8 mmol, 5 equiv.)
was added slowly. After stirring for approximately 30 minutes at room temperature, a white solid was
collected by filtration and washed with water. The final product, 1,3-dimethyl-2-phenyl-2,3-dihydro-1H-
benzo[d]imidazole (BIH), was recrystallized from (5:1) ethanol:water and dried under vacuum (5.7 g, 70
%), and stored in a N2-atmosphere glove box before use. 1H-NMR (400 MHz, DMSO-d6, ): 7.55 (dd, J
= 6.4, 1.9 Hz, 2H), 7.45 (dd, J = 5, 19 Hz, 3H), 6.62 (dd, J = 5.4, 3.3 Hz, 2H), 6.45 (dd, J = 5.4, 3.2 Hz,
2H), 4.87 (s, 1H), 2.48 (s, 6H).
X-ray diffraction data. Crystals of [Ru(dmb)2(CO)(CH3CN)](PF6)2 and [Ru(dmb)2(CO){NHC(CH3)-
OC2H4)}N{(C2H4)OH)}2](PF6)2, were mounted on the end of a glass fiber and the data were collected
with a Bruker Kappa Apex II diffractometer. Data collected at 296 K indicated triclinic symmetry and a
space group of
for both crystals. Crystal data and information about the data collection are provided
푃1
in the Supporting Information.
4
ACS Paragon Plus Environment

Page 5 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Electrochemistry. Cyclic voltammetry measurements were conducted using a BASi Epsilon potentiostat
in a standard three-electrode cell equipped with a glassy carbon (3 mm diameter) working electrode, a
silver (1.6 mm diameter) counter electrode, and a non-aqueous Ag/AgCl pseudo-reference electrode,
equilibrated in a 0.1 M tetrabutylammonium hexafluorophosphate (TBAPF6)/CH3CN electrolyte solution.
In a typical experiment, 1 mM of the complex of interest was dissolved in electrolyte solution prepared
with dry CH3CN (or 1.4 M TEOA/CH3CN) and 0.1 M TBAPF6 salt. All prepared solutions were saturated
with argon prior to the experiment. Cyclic voltammograms of the ferrocenium/ferrocene redox couple,
E1/2(Fc+/0), were measured before and after each experiment and used to convert the measured reduction
potentials to the normal hydrogen electrode (NHE) scale with the conversion constant, E1/2(Fc+/0) = 0.63
V vs NHE.22 Simulations of the cyclic voltammograms were carried out using the DigiSim simulation
software.23
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
Infrared spectroelectrochemical (IR-SEC) data were measured in controlled potential electrolysis mode at
fixed applied potentials. A custom-made reflective IR cell equipped with boron-doped diamond working,
platinum counter, and solid-state leakless miniature Ag/AgCl reference (EDAQ) electrodes was coupled
to a Fourier transform infrared (FTIR) spectrometer (Bruker, IFS 66/S) using a VeeMAX III Variable
Angle Specular Reflectance Accessory (Pike Technologies), and the IR absorption changes were
monitored following application of each potential.
Sodium amalgam reductions. A homemade T-shape glassware equipped with an optical cell was used.
The sodium amalgam (Na–Hg, 0.5% Na in Hg) compartment was separated from the main sample
compartment and optical cell by a frit. Purified anhydrous CH3CN was vacuum-transferred to the
glassware, and the apparatus was flame-sealed. Stepwise reductions were obtained by gradually
transferring the solution into the amalgam compartment through the frit and returning it to the main
compartment while monitoring the changes in ultraviolet-visible absorption.
Infrared spectroscopy. Steady-state mid-infrared absorption spectra were measured on a Thermo Nicolet
Nexus 670 FTIR spectrometer with a spectral resolution of 2 cm−1.
Time-resolved infrared (TRIR) spectroscopy. TRIR measurements were performed in the 1978 – 1870
cm−1 region using a CW external-cavity quantum cascade laser (EC-QCL, Daylight Solutions, 21052-
MHF-012) as the mid-infrared probe light source; or in the 2120 – 2005 cm–1 region using EC-QCL model
21049-MHF. A Nd:YAG pulsed laser (SpectraPhysics, Quanta-Ray LAB-170-10, 8-10 ns FWHM) was
used as the excitation light source tuned to 532 nm using a built-in second harmonic generator. The mid-
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 36
1
2
3
4
5
6
7
8
9
IR beam was split into reference and probe beams, with the latter passing through the IR cell. These beams
were directed onto a matched pair of fast rise time IR detectors (Kolmar Technologies, Inc., KMPV9-0.5-
J2, DC-20 MHz). The analog signals were simultaneously digitized on an oscilloscope (Teledyne LeCroy
HDO 4034, 12-bit, 350 MHz, 2.5 GS/s) and normalized to remove noise from laser intensity fluctuations.
For a more complete description of the TRIR instrumentation, the authors suggest previously reported
work in which it was applied to pulse radiolysis measurements.24 Solutions for the TRIR experiments
were prepared in an inert N2-atmosphere glove box using an airtight demountable liquid IR cell equipped
with CaF2 windows (Harrick Scientific, DLC-S25, pathlength d = 1.9 mm). TRIR experiments did not
cover the 2005 – 1978 cm–1 infrared region due to a gap between the tuning ranges of the two EC-QCLs.
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
Computational methods. Density functional theory calculations were performed using the Gaussian 09
software package.25 Ground-state geometries of the singly- and doubly-reduced complexes were
optimized using the M06 hybrid meta exchange-correlation functional,26 with the Integral Equation
Formalism Polarizable Continuum Model (IEF-PCM)27 for CH3CN as solvent. The LANL08 relativistic
effective core potential (RECP) and associated uncontracted basis set was sued to describe the Ru-
atoms,28-29 and 6-31G(d) was used for all other atoms. Transition state calculations were performed under
these same conditions. Single point energy calculations at the 6-311++G(2df,p) level were performed for
all optimized structures and transition states to evaluate the electronic energies. Frequency calculations of
the optimized structures were conducted to ensure that all local minima exhibited real frequencies,
whereas transition states displayed only one imaginary frequency. Frequency calculations were also used
to evaluate the Gibbs free energies of the local minima and transition states. For all calculations, the Ru(II)
carbonyl complexes had bpy ligands instead of dmb in order to avoid undesired negative frequencies
occurring at the methyl groups. Cartesian coordinates for optimized transition states are included in the
Supporting Information.
RESULTS
Early work has demonstrated that TEOA reacts with one of the carbonyl ligands of Ru(CO)(CO)2+ via
a nucleophilic attack on the carbonyl-carbon by a deprotonated TEOA-hydroxy group to yield a complex
with a CO-TEOA adduct coordinated to the metal, Ru(CO)(CO-TEOA)+ (Figure S1).14 This process is
analogous to the Lewis acid-base equilibrium between metal carbonyls and metallocarboxylic acids, M–
COn + OH– M–COOHn-1. The equilibrium between Ru(CO)(CO)2+ and OH– or TEOA was investigated
⇌
through titration measurements of the base while monitoring the changes in the ultraviolet-visible (UV-
6
ACS Paragon Plus Environment

Page 7 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
vis) absorbance spectrum. Equilibrium constants of 3 × 104 M-1 and 2 × 102 M-1 were obtained for OH–
and TEOA, respectively (see Figure S1 for the TEOA data). For reasons that are relevant to the discussion
in this work and to other related published results,14, 30 any reference to the Ru(CO)(CO-TEOA)+ complex
and its characterization, will hereafter imply the addition of the parent Ru(CO)(CO)2+ catalyst to a
concentrated (1.4 M) TEOA/CH3CN solution (1:4 v:v). Addition of TEOA to solutions containing other
forms of the catalyst did not cause noticeable reactivity, with the exception of the solvento complex,
Ru(CO)(CH3CN)2+, for which the coordinated solvent of Ru(CO)(CH3CN)2+ slowly underwent partial
solvolysis of the nitrile to form the bound imino ester adduct, Ru(CO)(CH3CN-TEOA)2+ (Figure S2), in
a manner similar to that previously observed for a related rhenium complex.5 This complex was later
prepared and isolated as a single crystal, allowing X-ray crystallographic confirmation of the proposed
molecular structure. Note that in this case, while the TEOA is deprotonated similar to in Ru(CO)(CO-
TEOA)+, the proton remains in the complex, bound to the N-atom of CH3CN. Crystallographic data are
discussed 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
Figure 1. Steady-state infrared absorption spectra of the Ru(II)-carbonyl complexes in the (CO) carbonyl
stretching region in CH3CN.
Infrared spectra were measured for all Ru(CO)(X)n+ complexes at frequencies that report on the
carbonyl stretching vibrations, v(CO) (Figure 1 and Table 1). Ru(CO)(CO)2+ shows two vibrational bands
centered around 2100 – 2000 cm1 associated with symmetric and asymmetric stretching modes of the cis
terminal carbonyl groups. Substitution of one carbonyl with other groups that have been identified as
intermediates in the catalytic CO2 reduction cycle led to red shifts of the remaining CO stretching
frequency due to an inductive effect on the π-electron back donation from the metal to the carbonyl group.
Molar absorption coefficients () at the v(CO) maxima were measured as 1650 ± 50 M1 cm1 for
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 36
1
2
3
4
5
6
7
8
Ru(CO)(HCOO)+, Ru(CO)(CO-TEOA)+ and Ru(CO)(H)+, while the higher energy v(CO) of
Ru(CO)(CO)2+ and the v(CO) of Ru(CO)(CH3CN)2+ displayed slightly larger values (1800 ± 50 M1
cm1).
9
Table 1. Experimentally determined carbonyl vibrations, v(CO), of the studied Ru(CO)(X)n+ complexes
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
in their resting state (RS) and respective singly- and doubly-reduced states, in CH3CN solution.
v(CO) (cm1)
Complex
RS
1st red
1970a
2nd red
1862a
Ru(CO)(CH3CN)2+
Ru(CO)(CO)2+
2008
2096/2040 2067/2008b
Ru(CO)(H)+
Ru(CO)(HCOO)+
Ru(CO)(CO-TEOA)+
1930
1976
1962
1894a
1938a
1928b
aDetermined from spectroelectrochemical methods. bObtained from TRIR experiments.
Figure 2. (a) Infrared spectra of a 0.75 M TEOA/CH3CN solution mixture measured while titrating in
13
known amounts of CO2. (b) C-NMR spectra of the indicated solutions. (c) Reaction scheme for the
formation of the CO2-TEOA zwitterion.
The direct reaction of TEOA with CO2 in CH3CN was observed by the appearance of a carbonyl
vibration at 1648 cm 1, which followed a linear dependence on the concentration of dissolved CO2
monitored at 2342 cm1 (Figure 2a). This new IR band appears in a region known for the asymmetric C=O
stretching vibrations of bicarbonate in CH3CN, at 1685 and 1646 cm1,16, 31 and the 13C-NMR spectrum
13
(Figure 2b) with added CO2 shows a resonance at 158.9 ppm typical of organic carbonates.32 The
presence of equilibrated carbonates formed due to residual water was ruled out by further addition of water
which resulted in the formation of a new resonance at 160.5 ppm consistent with inorganic carbonate.33
Based on these results, a preliminary assumption that CO2 may be reacting with TEOA to form a
zwitterionic alkylcarbonate structure, CO2-TEOA, shown in Figure 2c seems reasonable. Similar
8
ACS Paragon Plus Environment

Page 9 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
chemistry has precedent in CO2-binding organic liquids that contain an alcohol and an amidine or
guanidine base and chemically bind CO2 to form stable zwitterionic adducts.31, 34 The infrared spectra and
13C-NMR characterization of these reported adducts support the assignment established here for CO2-
TEOA.32 While tertiary amines have a relatively strong Brønsted basicity to act as proton acceptors,35 the
N-atom of TEOA alone is not basic enough to deprotonate its terminal OH groups. This process becomes
more thermodynamically favorable by considering a concerted proton-coupled nucleophilic attack of CO2
(Figure 2c). Quantitative analysis of the data in Figure 2a provided an equilibrium constant, Keq = 0.07
M1, for the zwitterionic capture of CO2 by TEOA in CH3CN (see SI for details).
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. Cyclic voltammograms for 1 mM solutions of the indicated Ru(CO)(X)n+ complexes measured
at 100 mV/s scan rate, in 0.1 M TBAPF6 CH3CN (bottom) or 1.4 M TEOA/CH3CN (top) electrolyte
solution.
(Spectro)Electrochemistry. Cyclic voltammograms for 1 mM solutions of the indicated Ru(CO)(X)n+
complexes are shown in Figure 3 in argon-saturated CH3CN with (top) or without (bottom) added 1.4 M
TEOA. At a scan rate of 100 mV s1, Ru(CO)(CO)2+ undergoes an irreversible two-electron reduction at
E1/2 = 0.91 V vs NHE.9 The solvento complex, Ru(CO)(CH3CN)2+, shows two reductions at E1/2 = 1.03
and 1.23 V vs NHE, both of which were quasi-reversible and scan rate dependent (Figure S5). Similar
scan rate dependence on the reversibility of the one- and two-electron reductions was observed for the
formato complex, Ru(CO)(HCOO)+, at E1/2 = 1.16 and 1.36 V vs NHE (Figure S6). Reversible redox
chemistry was observed for the Ru(CO)(H)+ complex in a 25-5000 mV s1 scan rate range, with two
sequential one-electron reductions appearing at E1/2 = 1.30 and 1.50 V vs NHE. In all cases, the redox
chemistry was characteristic of ligand-based reductions, and the couples with irreversible behavior for
Ru(CO)(CH3CN)2+ and Ru(CO)(HCOO)+ reported on ligand (CH3CN, or HCOO–) labilization.9, 11 The
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 36
1
2
3
4
5
6
7
8
9
Ru(CO)(CH3CN)2+ and Ru(CO)(HCOO)+ showed a third couple that is reversible with E1/2 = –1.55 V vs
NHE. All formal reduction potentials are compiled in Table 2. The redox properties of Ru(CO)(X)n+
changed substantially in the presence of TEOA (Figure 3, top). With the exception of Ru(CO)(H)+, each
complex showed an irreversible reduction followed by a reversible couple at E1/2 = 1.30 V vs NHE. This
second reduction is identical in potential to the first reduction of the hydrido complex, Ru(CO)(H)+ (see
below).
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
Table 2. Reduction potentials of the Ru(CO)(X)n+ complexes in CH3CN.
E
1/2
(V vs NHE)a
Complex
2+/+
+/0
0/–
Ru(CO)(CH3CN)2+
Ru(CO)(CO)2+
1.03b
0.91c
1.23b
< 0.91
Ru(CO)(H)+
Ru(CO)(HCOO)+
Ru(CO)0
1.30
1.1b
1.50
1.36b
1.55
b
aPotentials were converted to NHE using the conversion constant, E1/2(Fc+/0) = 0.63 V vs NHE. Quasi-reversible reduction,
cIrreversible reduction, reported value corresponds to the cathodic peak potential at 100 mV s−1 scan rate.
Simulation of the scan-rate-dependent data for the Ru(CO)(HCOO)+ complex allowed kinetic
modeling of the formate dissociation triggered by the 1st and 2nd reductions (Figure S6 and Table S5). The
mechanism was modeled with a combination of sequential electrochemical reduction and chemical steps
that accounts for the dissociation of HCOO– (see Figure S6 and Table S4 for details). The loss of formate
from the singly-reduced Ru(CO)(HCOO)+ in CH3CN is sluggish with a forward rate constant, kf = 0.07
s1 and is thermodynamically unfavorable, with Keq = 0.03 M. Attempts to resolve the kinetics for formate
loss in 1.4 M TEOA/CH3CN were unsuccessful because the data were nearly scan rate independent,
suggesting the dissociation was too fast to measure electrochemically. However, lowering the TEOA
concentration to 50 mM was found to be sufficient to decelerate the dissociation of formate, allowing
simulation of kf 1 s1 for the singly-reduced species. Under this condition, the formate dissociation at the
first electron reduction is thermodynamically allowed with Keq = 3.9, and is accelerated by up to 2 orders
of magnitude for the same reaction when compared to in the absence of TEOA.
10
ACS Paragon Plus Environment

Page 11 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Figure 4. Absorption changes during reductive spectroelectrochemical measurements in CH3CN on (a)
Ru(CO)(H)+, (b) Ru(CO)(CH3CN)2+, and (c) Ru(CO)(HCOO)+ in the presence of 0.5 M TBAHCOO (top)
or 1.4 M TEOA (bottom), all containing 0.1 M TBAPF6 as the supporting electrolyte.
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
Spectroelectrochemical data for the Ru(CO)(X)n+ complexes are shown in Figure 4. Upon reduction
of Ru(CO)(H)+ at –1.4 V vs NHE, the (CO) at 1930 cm–1 was lost concomitantly with the growth of a
new (CO) vibration at 1894 cm–1 assigned to Ru(CO)(H)0. The doubly-reduced Ru(CO)(H)– was
generated thereafter by application of –1.6 V vs NHE, with a new (CO) at 1862 cm–1. The intermediate
Ru(CO)(H)0 and initial Ru(CO)(H)+ complexes were fully recovered upon stepwise re-oxidation, while
maintaining isosbestic points consistent with those observed in the reductive steps. Reduction of the
Ru(CO)(CH3CN)2+ complex at –1.15 V vs NHE initially generated Ru(CO)(CH3CN)+ with (CO) at 1970
cm–1, followed by the subsequent appearance of another band at lower wavenumber. This resultant (CO)
vibration was characteristic of the Ru(CO)(H)+ resting state at 1930 cm–1. Spectroelectrochemical data for
Ru(CO)(HCOO)+ were collected with 0.5 M TBAHCOO to minimize the effect of HCOO– dissociation
(Figure 4c, top). At an applied potential of –1.45 V vs NHE, the initial absorption changes were consistent
with formation of the singly-reduced Ru(CO)(HCOO)0 with (CO) at 1938 cm–1. At longer equilibration
times, the singly-reduced Ru(CO)(HCOO)0 was converted to a set of two overlapping vibrations that
appeared in the 1925–1850 cm–1 region. Spectral fitting with a sum of three Voigt functions afforded
peaks at 1900 cm1 and 1870 cm1, which could not be assigned to any of the known (CO) previously
characterized for the Ru(CO)(X)n+ complexes. The final amplitudes of the new peaks at 1900 cm1 and
1870 cm1 were potential-dependent and reported on the equilibrated amounts that were generated after
reduction, but their relative amplitudes remained constant regardless of the potential applied. This latter
behavior was indicative of an equilibrium being established between these two species. While
spectroelectrochemical reduction of Ru(CO)(HCOO)+ in TBAHCOO-containing electrolyte did not
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 36
1
2
3
4
5
6
produce the Ru(CO)(H)+, experiments performed in 0.1 TBAPF6/CH3CN with or without TEOA resulted
in the immediate formation of Ru(CO)(H)+ (Figure 4c, bottom).
7
8
9
Sodium-amalgam (Na-Hg) reductions of the Ru(CO)(H)+ and Ru(CO)(CH3CN)2+ provided UV-vis
characterization of the 5-coordinated Ru(CO)0 complex (Figure S7). The redox states of each complex
were monitored by the changes in isosbestic points during the stepwise reduction. The singly- and doubly-
reduced forms of the Ru(CO)(H)+ showed clear ligand radical character with a strong band in the 450 –
600 nm region. While the IR-SEC data suggest that the doubly-reduced Ru(CO)(CH3CN)0 dissociates
CH3CN to form the 5-coordinate Ru(CO)0 which may react with a proton to form Ru(CO)(H)+, the UV-
vis spectrum obtained at the 2nd electron reduction of Ru(CO)(CH3CN)2+ by Na-Hg exhibited clear ligand
radical character with no indication of reactivity with protons, i.e. Ru(CO)0 + H+ → Ru(CO)(H)+.
Therefore, a direct comparison of these two spectra suggests that the 5-coordinate Ru(CO)0 complex was
isolated in the Na-Hg experiment due to the strictly anhydrous conditions and absence of electrolyte
(TBAPF6) salts. The third reduction of Ru(CO)(CH3CN)2+ showed increased ligand radical character with
absorption changes that were notably different from those measured for Ru(CO)(H)0. This reduction was
instead assigned to the formation of Ru(CO)–, which is consistent with the reversible reduction at E1/2 = –
1.55 V vs NHE that was common among Ru(CO)(CH3CN)2+ and Ru(CO)(HCOO)+ initial species and
therefore characterizes the reversible redox couple, Ru(CO)0/–.
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
Time-Resolved IR. Isolation of various catalyst forms that are proposed as reaction intermediates in the
reduction of CO2 allowed for their spectroscopic and electrochemical characterization. Each of these may
have an active role in the catalytic pathways through their initial resting state or reduced states. By using
transient absorption measurements, the singly-reduced states of Ru(CO)(X)n+ were photochemically
generated using [Ru(bpy-OMe)3]2+ as a chromophore, BIH as a reductive quencher, and TEOA as a base.
The excited state of [Ru(bpy-OMe)3]2+ oxidizes the BIH donor, producing a strong reductant, [Ru(bpy-
OMe)3]+, and a radical cation, BIH•+. The redox chemistry and sacrificial behavior of the electron donor,
BIH, requires a base to deprotonate BIH•+ to the BI• radical (Scheme 1).36 This radical intermediate is a
potent reductant (E1/2 = 1.43 V vs NHE; or 2.06 V vs Fc+/0),37 capable of reducing either the
chromophore or the catalyst in the dark. The final BI+ oxidized form does not take part in regeneration
reactions. Once the singly-reduced [Ru(bpy-OMe)3]+ is formed, it may transfer an electron to the catalyst,
which is the subject of interest here. The electron transfer steps are diffusion-controlled bimolecular
12
ACS Paragon Plus Environment

Page 13 of 36
Journal of the American Chemical Society
1
2
3
4
5
reactions36 and are not the focus of this work. Instead, we investigated the chemical transformations of the
catalyst after being reduced by one electron.
6
7
8
Scheme 1. Redox chemistry and acidity of BIH in CH3CN. Potentials are vs. NHE.
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
All time-resolved experiments were performed with 0.5 mM [Ru(bpy-OMe)3]2+, 5 mM Ru(CO)(X)n+,
and 0.2 M BIH in either neat CH3CN or 1.4 M TEOA/CH3CN solution, saturated with either N2 or CO2.
Under these conditions, the electron transfer to Ru(CO)(X)n+ is complete in a few hundred nanoseconds
(250 – 750 ns). In the absence of a base, BIH•+ is a reversible quencher and charge recombination between
the singly-reduced Ru(CO)(X)n-1 and BIH•+ follows second-order equal concentration kinetics. For all
Ru(CO)(X)n+ complexes studied, charge recombination occurs on the microsecond timescale with rate
constants, kcr, ranging from 0.4 – 2  1010 M1 s1, with no net photochemistry (Figure S8).
TEOA itself is a common sacrificial electron donor.3 However, in cases where a chromophore requires
a more potent reductant, such as BIH, TEOA is added to the solution solely to act as a Brønsted base, with
the assumption that it will ensure deprotonation of the oxidized radical, BIH•+ (Scheme 1). This prevailing
belief was initially assumed by us. However, our transient data revealed that TEOA is extremely
inefficient at deprotonating BIH•+ in CH3CN. Thus, photochemical reduction of the Ru(CO)(X)n+
complexes in 1.4 M TEOA/CH3CN solution did not result in persistent, irreversible, absorption changes.
Instead, a large fraction (~90%) of the transients generated after pulsed laser excitation undergo
unproductive charge recombination with BIH•+. For Ru(CO)(CO-TEOA)+, Ru(CO)(HCOO)+, and
Ru(CO)(COOH)+, the absorption changes were complex and will be examined shortly. For
Ru(CO)(CH3CN)2+, the (CO) transient absorption band displayed a 40 cm-1 shift to lower energy
consistent with one-electron reduction, i.e. Ru(CO)(CH3CN)2+ + e– → Ru(CO)(CH3CN)+ (Figure S9).
Kinetic traces that report on the concentration profile of Ru(CO)(CH3CN)+ in the presence of TEOA did
not follow second-order equal concentration kinetics, while a small, ~10%, fraction of Ru(CO)(CH3CN)+
was still present at longer timescales. The kinetic traces were instead well modeled by a set of two rate
equations that coupled the time-dependent concentration of Ru(CO)(CH3CN)+ with the slow
deprotonation of BIH•+ by TEOA (Equations 1–2),
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 36
1
2
3
4
+
[
] = ― 푘푐푟 푅푢
퐵퐼퐻
푑 푅푢
• +
[
+ ][
]
1
푑푡
5
6
7
• +
[
] = ―푘푐푟 푅푢
푑 퐵퐼퐻
• +
• +
[
+ ][
]
[
― 푘푑푝[푇퐸푂퐴] 퐵퐼퐻
]
퐵퐼퐻
2
푑푡
8
9
where kcr is the second-order rate constant for charge recombination and kdp is the rate constant for
deprotonation of BIH•+. In a 1.4 M TEOA/CH3CN solution containing 0.2 M BIH, the numerical solution
of equations 1-2 to the Ru(CO)(CH3CN)+ kinetic trace (Figure S9c) provided kdp = 4.5 × 103 M1 s1. This
result shows that the deprotonation of BIH•+ by TEOA is slow and therefore, insufficient to kinetically
compete with charge recombination. Similar results were obtained for Ru(CO)(H)+ under the same
conditions.
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 quantitative deprotonation of BIH•+ in the absence of competitive redox chemistry is an acid-base
chemical reaction that requires a proton acceptor that is at least 2 pKa units more basic than the pKa of
BIH•+. The pKa of BIH•+ was estimated to be 14.4, from thermodynamic relationships using available
values for hydricity and reduction potentials in CH3CN (see SI).37-38 The nitrogen acidity of the conjugate
acid of TEOA (TEOAH+) in CH3CN has been estimated using calorimetric methods, pKaN = 15.9.39 Here,
pKaN = 13.74 was estimated using an isodesmic relationship with the known pKa for TEAH+ (18.82)39-40
and DFT-computed free energies for the acid-base reaction, TEOA(H+) + TEA
TEOA + TEA(H+).
⇌
Regardless of the methodology employed to obtain the pKaN of TEOAH+, the acid-base equilibrium
predicts that 40 – 97% of the BIH•+ should be deprotonated. However, despite favorable thermodynamics,
the deprotonation reaction is slow compared with recombination.
The slow deprotonation reaction of BIH•+ by TEOA in CH3CN was contrasted with other proton
acceptors with known pKa values in CH3CN (Figure S10): triethylamine (pKa(TEAH+) = 18.8)39-40
,
formate (pKa(HCOOH) = 20.7)41, and 1,8-diazabicyclo[5.4.0]undec-7-ene (pKa(DBUH+) = 24.34)40. The
kinetic data obtained from the photochemical reduction of Ru(CO)(H)+ were analyzed according to
numerical solutions to equations 1-2, from which rate constants for the deprotonation of BIH•+ were
TEA
HCOO–
DBU
extracted (kdp
= 5.5  105 M1 s1; kdp
= 7.4  106 M1 s1; kdp
= 3.0  107 M1 s1). When
plotted on a logarithmic scale, the obtained rate constants exhibit a linear dependence on the pKa of the
base, consistent with the predicted behavior of Brønsted acid-base catalyzed reactions. About 1 M TEA
was sufficient to induce ~ 90% irreversibility in the photoreduction of the Ru(CO)(H)+ complex, while
the stronger bases HCOO– and DBU, only required a few tens of millimolar concentration. Indeed, a recent
study has shown that substitution of molar concentrations of TEOA with millimolar levels of DBU, more
14
ACS Paragon Plus Environment

Page 15 of 36
Journal of the American Chemical Society
1
2
3
4
5
than doubled the overall photocatalytic performance for the reduction of CO2 to CO with rhenium-
tricarbonyl type catalysts.42
6
7
8
9
The photochemical one-electron reductions of Ru(CO)(CO)2+, Ru(CO)(HCOO)+, and Ru(CO)(H)+
obtained in neat CH3CN and 1.4 M TEOA/CH3CN solutions, will be examined in the following sections.
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
TRIR of Ru(CO)(CO)2+. Photochemical reduction of Ru(CO)(CO)2+ in CH3CN in the presence of BIH
but in the absence of TEOA resulted in a bleach of the carbonyl vibration bands at 2096 and 2040 cm1
and the formation of two new v(CO) bands at 2067 and 2008 cm1 (Figure 5a). This was expected,
considering that the ligand-based reduction increases the -donation from the metal to the carbonyl
ligands. While the irreversible two-electron reduction of Ru(CO)(CO)2+ was reported to cause CO-ligand
loss from electrochemical measurements in solution,8 or cause the loss of a bipyridine ligand to form
polymeric structures,43 it is unclear from CV data whether such labilization occurs at the first, or second
discrete one-electron reduction due to the irreversible nature and closely spaced redox potentials.2
According to the data shown in Figures 5a and S8a, reversible recombination of the singly-reduced
Ru(CO)(CO)+ with BIH•+ occurred in a few tens of microseconds with no net photochemistry involving
dissociation of CO.
Figure 5. TRIR spectra of (a) Ru(CO)(CO)2+ in neat CH3CN and (b) Ru(CO)(CO-TEOA)+ in 1.4 M
TEOA/CH3CN, measured at the indicated time delays after pulsed 532 nm excitation of the [Ru(bpy-
OMe)3]2+ chromophore in the presence of 0.2 M BIH. Overlaid are spectral simulations based on a least
squares fit of known v(CO) IR spectra. (c) Simulated time-dependent concentration profiles of the
transient species present in (b) based on a least squares fit global analysis. The trace for Ru(CO)(CO-
TEOA)+ represents how much of the starting material is depleted as a function of time.
In 1.4 M TEOA/CH3CN, the equilibrated form of Ru(CO)(CO)2+ is Ru(CO)(CO-TEOA)+, with a
single v(CO) vibration at 1962 cm-1. Photochemical reduction of Ru(CO)(CO-TEOA)+ resulted in
absorption changes at early timescales consistent with the bleach of this single v(CO) vibration concurrent
15
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 36
1
2
3
4
5
6
7
8
9
with the appearance of a new absorption band at 1928 cm-1 assigned to the v(CO) of the singly-reduced
Ru(CO)(CO-TEOA)0 (Figure 5b). At time scales beyond 5 µs the absorptions were not unique to
Ru(CO)(CO-TEOA)+/0, but instead indicated the coexistence of additional intermediates. The spectra were
satisfactorily modeled by a least squares fit of the known IR spectra of the bleached-Ru(CO)(CO-TEOA)+,
singly-reduced Ru(CO)(CO-TEOA)0 and Ru(CO)(CH3CN)+ (Figure S11). Analysis of the deconvoluted
concentration profiles revealed that the singly-reduced solvento complex, Ru(CO)(CH3CN)+, is partially
formed in kinetic competition with charge recombination. Since the rate at which Ru(CO)(CH3CN)+ is
formed, k = 5.3 × 105 s-1, must coincide with that of the loss of the (CO-TEOA)– ligand, the overall ligand
exchange reaction, Ru(CO)(CO-TEOA)0 + CH3CN → Ru(CO)(CH3CN)+ + (CO-TEOA)– is a clear
indication of CO dissociation assisted by TEOA. An overview of the reactions involved in the
photochemical one-electron reduction of Ru(CO)(CO)2+ and Ru(CO)(CO-TEOA)+ is shown in Scheme
2.1.
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 2.1
TRIR of Ru(CO)(HCOO)+. The photochemical reduction of Ru(CO)(HCOO)+ in neat CH3CN solution
resulted in a v(CO) vibration that was 38 cm-1 shifted to lower energy, with a clean isosbestic point around
1960 cm-1 (Figure 6a). Consistent with data obtained from spectroelectrochemical experiments, the new
band centered at 1938 cm-1 was assigned to the v(CO) vibration of the singly-reduced formato complex,
Ru(CO)(HCOO)0. There was no net photochemistry involving the loss of HCOO– or CO on the timescale
of charge recombination. Persistent photochemical reduction of Ru(CO)(HCOO)+ was obtained by adding
50 mM TBAHCOO as a base to the CH3CN solution (Figure S12b-c). Although the presence of formate
salt may alter the equilibrium and rates that connect reactants and products in Ru(CO)(HCOO)0 + CH3CN
Ru(CO)(CH CN)+ + HCOO–, there was no evidence for the loss of formate since the concentration of
⇌
3
Ru(CO)(HCOO)0 initially formed remained unchanged on the millisecond timescale.
16
ACS Paragon Plus Environment

Page 17 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
The TRIR spectra recorded in 1.4 M TEOA/CH3CN are shown in Figure 6b. Immediately after the
reduction of Ru(CO)(HCOO)+, the singly-reduced Ru(CO)(HCOO)0 was formed. From this point
forward, the absorption changes were complex implying the coexistence of multiple transient species. The
spectral and kinetic data were adequately modeled with a weighted sum of known IR spectra of the
bleached-Ru(CO)(HCOO)+, Ru(CO)(HCOO)0 and Ru(CO)(CH3CN)+. Quantitative analysis of the time-
dependent concentration profiles revealed that Ru(CO)(CH3CN)+ is partially formed in kinetic
competition with charge recombination, and must coincide with the rate of the formate ion dissociation
(5.5  105 s–1). Thus, analogously to the electrochemical data in the presence of TEOA, the
photochemically-generated singly-reduced Ru(CO)(HCOO)0 undergoes accelerated dissociation of
formate product in the presence of TEOA. An overview of these reactions in the presence and absence of
TEOA is shown in Scheme 2.2.
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 6. TRIR spectra of Ru(CO)(HCOO)+ measured at the indicated time delays after pulsed 532 nm
excitation of the [Ru(bpy-OMe)3]2+ chromophore in the presence of 0.2 M BIH: (a) in neat CH3CN and
(b) in 1.4 M TEOA/CH3CN. Overlaid are spectral simulations based on a least squares fit of known v(CO)
IR spectra. (c) Simulated time-dependent concentration profiles of the transient species present in (b)
based on a least squares fit global analysis. The trace for Ru(CO)(HCOO)+ represents how much of the
starting material is depleted as a function of time.
Scheme 2.2.
TRIR of Ru(CO)(H)+. Similar to all previous measurements performed in neat CH3CN solution, the TRIR
spectra recorded for Ru(CO)(H)+ were characteristic of the reversible reduction to Ru(CO)(H)0 (Figure
17
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 36
1
2
3
4
5
6
7
8
9
7a). The Ru(CO)(H)+ bleach was centered at 1930 cm-1 and a lower energy v(CO) band at 1894 cm-1 that
was assigned to the singly-reduced hydrido complex, Ru(CO)(H)0, appeared. Experiments performed in
saturated CO2 (0.28 M), did not result in the expected hydride transfer reactivity to CO2 during the
timescale of charge recombination. As described earlier in this section, the addition of 50 mM of
TBAHCOO as a base was sufficient to irreversibly generate Ru(CO)(H)0 in N2-saturated solution (Figure
S13a-b). Kinetic traces monitored at 1894 cm-1 were recorded as a function of known titrated amounts of
CO2, Figure S13c-d. All kinetic traces displayed monoexponential behavior, and according to the rate law
equation written for the metal-hydride reactivity with CO2, equation 3, they were indicative of a pseudo-
first-order kinetic reaction.
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
0
[
] = ― 푘퐻푇 퐶푂2 푅푢(퐶푂)(퐻)
푑 푅푢(퐶푂)(퐻)
0
[
]
[
]
3
푑푡
Quantitative analysis of these data provided a rate constant for hydride transfer to CO2 of kHT = 9.6 M-1
s-1 to produce the singly-reduced formato complex, Ru(CO)(HCOO)0.
Figure 7. TRIR spectra of Ru(CO)(H)+ measured at the indicated time delays after pulsed 532 nm
excitation of the [Ru(bpy-OMe)3]2+ chromophore in the presence of 0.2 M BIH: (a) in neat CH3CN and
(b) in 0.7 M TEOA/CH3CN. Overlaid are spectral simulations based on a least squares fit of known v(CO)
IR spectra. (c) Simulated time-dependent concentration profiles of the transient species present in (b)
based on a least squares fit global analysis. The trace for Ru(CO)(H)+ represents how much of the starting
material is depleted as a function of time.
TRIR measurements on the photochemical reduction of Ru(CO)(H)+ in the presence of TEOA showed
unexpected complications. A slow but gradual formation of Ru(CO)(HCOO)+ was observed during the
sample preparation in the CO2 saturation step prior to the transient measurement. This complicates the
analysis of hydride transfer to CO2 because direct reduction of the existing Ru(CO)(HCOO)+ can result in
misinterpretation of the data. In order to minimize such undesired reactivity prior to the experiment, a
18
ACS Paragon Plus Environment

Page 19 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
solution containing the catalyst was mixed with a second solution containing only CO2 (0.28 M)
immediately before pulsed laser excitation using syringe pumps and gas-tight syringes. The solutions were
mixed with equal volumes such that the final concentration of CO2 was (0.14 M). It is worth noting that
the zwitterionic capture of CO2 by TEOA under these conditions resulted in a 10 mM concentration of
CO2-TEOA. The TRIR spectra measured for Ru(CO)(H)+ were recorded with a fresh aliquot every 2–3
data points. The absorption changes that follow the photochemical one-electron reduction of Ru(CO)(H)+
showed an apparent time-dependent red-shift of the bleach at 1930 cm1 concurrent with the decay of the
Ru(CO)(H)0 absorbance at 1894 cm 1 (Figure 7b).
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 2.3.
A careful global analysis of the data (Figure 7c), revealed that the singly-reduced formato complex
was formed at an observed rate of 4.1  105 s1, after which it slowly decayed to reach an equilibrium
concentration with the formation of the singly-reduced solvento complex, Ru(CO)(CH3CN)+, at a rate of
3.5  104 s1. A representative example of the overlapped transient species taken at 15 µs after pulsed laser
excitation is shown in Figure S14a. Single kinetic traces were also measured at saturation concentrations
of CO2 and CO2-TEOA (278 mM and 20 mM, respectively), at wavenumbers that predominantly reported
on the concentrations of bleached-Ru(CO)(H)+, Ru(CO)(H)0, Ru(CO)(HCOO)0, and Ru(CO)(CH3CN)+,
Figure S14b. The overall absorption changes were identical to those shown in Figure 7c, but the rate (8.2
 105 s1), and amplitude of Ru(CO)(HCOO)0 formation increased. This result demonstrates a CO2
concentration effect on the formation of the singly-reduced formato complex, from which tentative plots
of the obtained rates as a function of the CO2 and CO2-TEOA concentrations provided rate constants for
hydride transfer in the presence of TEOA to CO2 and CO2-TEOA of kHT = 2.8 × 106 M1 s1 and kHT = 4.1
× 107 M1 s1, respectively (Figure S14c-d). These rate constants are about 5–6 orders of magnitude larger
19
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 36
1
2
3
4
5
6
7
8
9
than those measured for the hydride transfer to CO2 in CH3CN solutions without TEOA, and the
implications of this result to the catalytic reduction of CO2 to formate will be more rigorously examined
in the Discussion section. An overview of the reactions involved in the photochemical one-electron
reduction of Ru(CO)(H)+ in the presence and absence of TEOA is shown in Scheme 2.3.
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
DISCUSSION
While decades of research has promoted the design of a large number of CO2 reduction catalysts, most
studies focus on the macro-observations of product selectivity and catalyst performance. More rigorous
examination of the catalytic pathways is often conducted using electrochemical and computational
methods, while only a limited number of studies succeeded in identifying transient intermediates, their
kinetics and contributions to the overall catalytic reaction. In this discussion, the electrochemical and
spectroscopic data presented earlier will be combined to provide a global view of the thermodynamics and
kinetic events in the reduction of CO2 to formate by Ru(II)-carbonyl catalysts. This includes the key steps
of formate ion dissociation, metal hydride formation and the hydride transfer to CO2 to produce the metal-
bound formate intermediate. Complementary DFT calculations were guided by the experimental results
to provide a comparative interpretation of the free energy landscape of the relevant catalytic steps. While
TEOA is customarily used in most photochemical assays, its active participation in the catalytic pathways
is often disregarded. Here, this concern will be brought into context for discussion. Though there has been
early work,2, 8 and some recent indications,44 that suggest a connection between metal-formate and
metallocarboxylic acid intermediates, the results discussed here will examine the metal hydride,
Ru(CO)(H)+, as the active catalyst and its ability to reduce CO2.
Thermodynamic considerations. Hydricity is a thermodynamic measure of the Gibbs free energy change,
–
GºH , for the heterolytic dissociation of the metal-hydride bond, and is conceptually used to quantify the
ability of a metal hydride to donate a hydride ion. Since the catalytic production of H2 and formate are
intimately connected to metal hydride intermediates, a proper knowledge of the thermodynamic hydricity
allows for a predictive estimate of product selectivity. Here, the competitive hydride reactivity for the
reduction of H+ and CO2 are illustrated in Scheme 3, where the thermodynamic hydricity and pKa of the
reaction components are combined in a diagram that allows for direct and quantitative analysis of the
overall reaction driving forces.45 Once the pKa of the proton source is defined, the diagram establishes
specific thermodynamic boundaries that regulate product selectivity and limit the choices of metal
hydrides that can operate under catalytic conditions.
20
ACS Paragon Plus Environment

Page 21 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
–
1
46
–
The hydricity of formate, GºH (HCOO ) = 44 kcal mol in CH3CN, imposes an upper limit for
exergonic reactions of metal hydrides to reduce CO2 to formate. Theoretically, the more hydridic the metal
hydride, the more facile the hydride transfer to CO2 becomes. This is reasonably true for stoichiometric
reactions in the absence of proton sources. For a metal hydride to become catalytic, protons are necessary
to reform the metal hydride at the end of the catalytic cycle, and this requirement introduces
thermodynamic boundaries that may constrain its catalytic activity. In this context, the pKa of the proton
source becomes a delimiter for the metal hydride recycling, while at the same time it establishes the driving
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
―
―
(
)
Δ퐺H = Δ퐺H ―(Δ퐺H H2 ―1.364 × p퐾 )
–
force for H2 evolution,
, where GH is the hydricity of a metal
a
2
1
–
hydride and GH (H2) = 77 kcal mol is a measure of the free energy of dissociation of H2 to form a
hydride ion and free proton in CH3CN.46 If only the hydride transfer step is considered, one may naturally
conclude that the higher the pKa of a proton source, the more difficult H2 evolution becomes, and therefore
it results in lower competitive catalysis for the reduction of CO2 to formate. However, as mentioned
earlier, the choice of a proton source must be carefully considered to allow for the protonation of the metal
center to reform the active metal hydride complex. An example of these considerations is given in Scheme
3 where the TEOA hydroxy groups act as the proton source.
Scheme 3. Free energy relationships for metal hydride reactivity to produce formate and/or H2. The
indicated pKa of the TEOA hydroxy groups limits the cyclability of the metal hydride and defines the
driving force for H2 evolution. Hydricity and pKa values for [Ru(dmb)2(CO)(H)]+ and [Ru(tpy)(bpy)(H)]+
are compared for discussion. All values are reported with two significant digits (see SI).
For reasons that will be clarified later, TEOA is considered here as a proton source and the estimated
acidity for its hydroxy groups (pKaOH = 34.4) was used to complete the diagram in Scheme 3. Calculated
hydricity values for two metal hydrides, [Ru(dmb)2(CO)(H)]+ and [Ru(tpy)(bpy)(H)]+, at three different
redox states were combined with their related pKa values and included in the diagram for discussion. The
21
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 36
1
2
3
4
5
6
7
8
9
reaction driving forces for hydride transfer to H+ and CO2 were quantified according to known free energy
relationships shown in Scheme 3.45 The hydricity value for [Ru(dmb)2(CO)(H)]+ was experimentally
1
–
determined as GH = 51 ± 0.5 kcal mol by means of direct hydride transfer and by the H2 heterolysis
method.46-47 Complementary DFT computational methods provided an estimated hydricity value of
―
∆퐺H
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
= 50.8 kcal mol1 using an isodesmic relationship, which is in good agreement with the experimental
data.48 Additional hydricity values for different redox states of the metal hydrides, the homolytic bond
cleavage of a metal-hydride bond and pKa values were determined according to free energy relationships
described by known thermochemical cycles.45-46, 49 Details on the methodologies employed are described
in the SI.
The hydricity of Ru(tpy)(bpy)(H)+ is sufficient to exergonically reduce CO2 to the formate anion
―
Δ퐺
without additional reduction of the catalyst with
) = – 5.0 kcal mol1, while this same reaction is
(HCOO
7.0 kcal mol1 endergonic for Ru(CO)(H)+. In cases where it is necessary to enhance the reactivity of
metal hydrides, ligand- or metal-based reductions have widely been employed as a means to increase
electron density at the metal hydride bond.11, 46 For instance, the singly-reduced forms of the metal
hydrides listed here become ~6.5 kcal mol1 more hydridic compared to their initial resting state.
Additional ligand-based reductions monotonically enhance the reactivity of metal hydrides towards the
reduction of CO2, though it may collaterally increase the susceptibility to competitive H2 evolution.
Although [Ru(tpy)(bpy)(H)]+ is a better hydride donor than Ru(CO)(H)+ and Ru(CO)(H)0, the
thermodynamic boundaries imposed by TEOA as the proton source impede the metal hydride to be
reformed under catalytic conditions given that the TEOA hydroxy groups are more basic than
[Ru(tpy)(bpy)(H)]+.
Activation of Ru(CO)(H)+ to the more hydridic singly-reduced Ru(CO)(H)0 is necessary to achieve
―
Δ퐺
more favorable thermodynamic driving forces for formate anion production, with
= 0.8 kcal
(HCOO
)
1
mol , while still being incompetent for H2 evolution, (H ) = 15 kcal mol1. Besides, the strong basicity
Δ퐺
2
of Ru(CO)(H)+ allows for the use of proton sources with pKa values as high as 39, thus making hydrogen
evolution even more difficult, while allowing for Ru(CO)(H)+ to be re-formed. The increased hydricity of
the doubly-reduced Ru(CO)(H)– facilitates hydride transfer to CO2,
= –5.5 kcal mol1, with
―
Δ퐺(HCOO
)
Δ퐺
predominant selectivity over H2 evolution,
= 8.5 kcal mol1 with TEOA as the proton source,
(H2)
although at the cost of an additional 0.2 V for activation.
22
ACS Paragon Plus Environment

Page 23 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
One must emphasize that despite the relations derived from Scheme 3 providing predictions for
reaction equilibria, they are purely thermodynamic and do not account for the kinetic aspects of the
reaction activation barriers, which may decisively control the final outcomes in catalytic reactions that
have multiple competitive pathways.45 Additionally, these relationships consider the overall reactions with
HCOO– and H2 as final products, yet most intermediate products that follow from a hydride transfer step
exist transiently coordinated to the metal center as M–HCOO– or M–H2. Therefore, additional free energy
corrections that account for this bond cleavage must be considered in order to more accurately report on
reaction driving forces.
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
Metal-hydride formation. Transition metal hydrides are undeniably recognized as key intermediates in
the conversion of protons to hydrogen gas or CO2 to formate. It is therefore important to consider here the
circumstances in which they are formed. In most molecular artificial photosynthetic designs for H2
evolution and the reduction of CO2, protons and electrons are used to source the metal-hydride formation,
provided that the metal has an available coordination site and appropriate pKa to allow for protonation.
The calculated pKa for Ru(CO)(H)+ (pKa = 39) translates into a conjugate base, Ru(CO)(CH3CN)0,
that is possibly sufficiently basic to deprotonate water (pKa = 38 – 41)46, 50 present in CH3CN solutions as
a proton source. In dry CH3CN conditions, reductive cyclic voltammetry experiments on
Ru(CO)(CH3CN)2+ and Ru(CO)(HCOO)+ at 100 mV/s led to ligand loss chemistry and formation of the
unsaturated 5-coordinate Ru(CO)0, with no clear electrochemical evidence that Ru(CO)(H)+ has been
generated. Although this observation may, at first, seem to contradict the results obtained from
spectroelectrochemical measurements, where application of potentials at the second electron reduction
resulted in the irreversible formation of Ru(CO)(H)+, one needs to recall that the experimental timescales
are notably different. At fixed applied potentials, an equilibrium amount of Ru(CO)0 may react with traces
of water and form Ru(CO)(H)+. Alternatively, it has been reported that ammonium salts are susceptible to
fragmentation reactions in the presence of strong bases,51 and this degradation pathway may supply
protons for the protonation of the metal. While precautions were taken to ensure that all experiments were
performed under the driest possible conditions, the formation of Ru(CO)(H)+ was always observed and
the possibility of residual water as the proton source cannot be ruled out.
Ru(CO)(H)+ metal hydride formation became more evident with the addition of TEOA. Photochemical
assays of CO2 reduction using molecular catalysts are generally conducted in the presence of high
concentrations of TEOA (1 – 1.4 M), yet much attention is given to the sacrificial quenchers, e.g. BNAH
23
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 36
1
2
3
4
5
6
7
8
9
and BIH, as the proton sources once oxidized by one electron. Cyclic voltammetry experiments conducted
in N2-saturated 1.4 M TEOA/CH3CN solutions exhibited a reduction wave at more negative potentials
that was common to all Ru(CO)(X)n+ complexes, and which matched the peak potentials for the
Ru(CO)(H)+/0 redox couple. Spectroelectrochemical measurements at potentials just negative enough to
initiate the second reduction of each complex unarguably validated the formation of Ru(CO)(H)+ as
probed by its characteristic v(CO) stretching vibration at 1930 cm-1.
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 pKa of the TEOA hydroxy groups in water (15.12) is comparable to those found for primary
alcohols such as ethanol and methanol (15.5), which is just slightly higher than the pKa of water (14).
Although the acidities of alcohols and TEOA have not been experimentally determined in CH3CN, they
are expected to behave as weak acids, with pKa values as high as those considered for water (pKa = 38 –
41)46, 50. The ability of the 5-coordinated Ru(CO)0 complex to abstract a proton from TEOA was tested by
quantifying the free energy associated with the Ru(CO)0 + TEOA
Ru(CO)(H)+ + TEOA(–H+) acid-
⇌
base reaction using computational methods, from which the TEOA-hydroxy pKa of 34.4 was estimated
relative to the calculated value of the Ru(CO)(H)+ pKa of 39. This result supports the experimental
observations that the TEOA-hydroxy groups function as a proton reservoir for the catalytic conversion of
CO2 to formate, provided that the metal-hydride has the appropriate pKa.
Organic hydride electron donors, such as BNAH and BIH, have also been considered as proton donors
in photochemical assays when in their oxidized forms, BNAH•+ and BIH•+, since their sacrificial behavior
requires the loss of a proton from their radical cationic form. In this case, two possible protonation
pathways may be considered: 1) the first pathway involves the direct protonation of the metal by BIH•+.
The difference in pKa’s between BIH•+ (14.3) and Ru(CO)(H)+ (39) validates this pathway. However, the
radical cation, BIH•+ is also an electron acceptor, and thus the direct protonation of the metal from BIH•+
is in kinetic competition with unwanted charge recombination reactions; 2) the second pathway uses an
external base functioning as the proton acceptor reservoir, carrier and eventually the proton donor. This
dual character is therefore subject to two prerequisites, i.e., the added base has to be sufficiently basic to
carry out the deprotonation of BIH•+ but at the same time it must be more acidic than the metal hydride.
Protonated-TEOA, which has a pKaN ~ 14 meets the acidic criterion but, as shown in the Results section,
it inefficiently deprotonates the oxidized BIH•+ under transient conditions. Notwithstanding that
protonated-TEOA is a better acid than the TEOA-hydroxy groups (pKaOH = 34.4), it is only formed at low
concentrations with kinetics for protonation of the metal that are diffusion-controlled. Therefore, given
24
ACS Paragon Plus Environment

Page 25 of 36
Journal of the American Chemical Society
1
2
3
4
5
the experimental evidence that TEOA is capable of sourcing the Ru(CO)(H)+ metal hydride formation, it
is reasonable to consider TEOA as the predominant proton source in photochemical assays in CH3CN.
6
7
8
9
Hydride transfer. Computational studies that detail the mechanisms by which CO2 reacts with metal
hydrides to produce formate are widely more available than experimental investigations.11, 52-59 This is
largely because of the experimental challenges inherent to the transient detection and short-term stability
of most metal hydrides. Only a few isolated cases in which a metal hydride is capable of reducing CO2 in
its resting state allowed for more systematic studies.53-60 The slow reactivity between these metal hydrides
and CO2 allowed for the time-resolved detection of the hydride transfer step using conventional UV-vis
and NMR spectroscopy or with rapid mixing stopped-flow measurements. The results gathered from these
studies marked the initial steps towards a better understanding of the variables that control the kinetics for
hydride transfer to CO2. This includes solvent effects, the presence of Lewis acid additives, and
corroboration of computationally proposed mechanisms.
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
To date, the extensive library of computational work has established two prevailing mechanisms for
which a metal hydride may react with CO2 to produce formate. The first and more frequently
communicated mechanism proposes a stepwise pathway wherein hydride transfer is initiated after the
nucleophlic attack of CO2 by the metal hydride to form a loosely H-bound formate intermediate – hydride
abstraction. At this stage, the bound formate may dissociate as the anionic product, or the H-bound
formate may undergo an inner sphere rotational rearrangement that stabilizes the commonly observed O-
bound formate product. The second proposed mechanism introduces the concept of a single concerted
pathway, in which CO2 inserts into the metal hydride bond to adopt a four-centered intermediate – CO2
insertion.52, 60 The weak interaction between the CO2-oxygen and the metal pulls electron density from
the carbon to break the thermodynamic stability of CO2, at the same time as the metal hydride bond is
electron-enriched. This concerted and beneficial interaction ultimately results in facilitated nucleophilic
attack by the metal hydride to form the O-bound formate product. Note that for both mechanistic situations
described, the O-bound formate is the immediate product of the hydride transfer to CO2, and as such, the
―
Δ퐺
thermodynamic relationships quantified from Scheme 3,
), are not adequate representations of the
(HCOO
푏표푢푛푑
―
Δ퐺
real reaction free energy,
), for the hydride transfer step. Since a metal-oxygen bond must be
(HCOO
cleaved to afford the final HCOO– anion, this free energy must be considered when determining the actual
reaction driving force according to equation 4,
푏표푢푛푑
푑푖푠푠표푐푖―푎푡푖표푛
―
―
Δ퐺
) = Δ퐺(HCOO ) ― Δ퐺(HCOO
)
4
(HCOO
25
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 26 of 36
1
2
3
4
Δ퐺푑푖푠푠표푐푖푎푡푖표푛 = ― RT ln(퐾 )
―
where
and Keq is the equilibrium constant for formate dissociation (0.03 M),
(HCOO
)
eq
5
6
7
which was obtained from scan-rate dependent electrochemical data. While the overall free energy change
–
0
Δ
to generate a HCOO anion product from reacting the singly-reduced [Ru(dmb)2(CO)(H)] with CO2 is
8
9
Δ퐺푏표푢푛푑
) for the hydride transfer step itself is about – 1.3 kcal mol–1. A similar
(HCOO
–1
―
―
퐺(HCOO
)= 0.8 kcal mol ,
analysis may be considered for [Ru(tpy)(bpy)(H)]+ if the equilibrium constant for formate dissociation
from [Ru(tpy)(bpy)(HCOO)]+ is known. This value was not measured in the present study, nor was it
available from the literature. Nonetheless, since [Ru(tpy)(bpy)(HCOO)]+ is stable and does not
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
spontaneously dissociate HCOO– without further reduction to [Ru(tpy)(bpy)(HCOO)]0, the reaction free
Δ퐺푏표푢푛푑
―
energy for the hydride transfer step,
) must be more negative than – 5.0 kcal mol–1, given that Keq
(HCOO
for formate dissociation is a very small number.
Here, the reactivity of the Ru(CO)(H)+ metal hydride towards CO2 is considered and time-resolved in
its transient singly-reduced Ru(CO)(H)0 form. To the best of our knowledge, the results disclosed here are
the first ever reported for this class of Ru(II) catalyst and possibly the first to investigate the hydride
transfer to CO2 from transiently reduced metal hydride complexes. Guided by the obtained experimental
results, computational methods were utilized to gather further information about the reaction pathways
that control the hydride transfer to CO2.
The electrochemical assessment of Ru(CO)(H)+ indicated that its singly-reduced form was stable
under the electrochemical timescales of cyclic voltammetry and controlled potential electrolysis
experiments. The photochemically generated Ru(CO)(H)0 further supported this observation by showing
the long-lasting formation of Ru(CO)(H)0 under inert N2 conditions, with no indication of side reactivity
with solvent, BIH, BIH•+ or TBAHCOO salts. Once the stability of Ru(CO)(H)0 had been verified,
controlled titrations of CO2 to the photochemically generated Ru(CO)(H)0 allowed for the quantitative
determination of the rate constant for hydride transfer to CO2, kHT = 9.6 M1 s1. This result alone does not
provide sufficient insight into the pathway undertaken for the formation of the O-bound formate product.
Theoretical calculations performed for Ru(CO)(H)0 indicated a stepwise pathway for hydride transfer,
Scheme 4. The H-bound formate intermediate is transiently formed after the nucleophilic attack of CO2
by the metal hydride, followed by rotational rearrangement to generate the O-bound formate product.
26
ACS Paragon Plus Environment

Page 27 of 36
Journal of the American Chemical Society
1
2
3
Scheme 4. DFT-computed free energies for hydride transfer to CO2.
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
Comparison of the measured rate constant for Ru(CO)(H)0 with that obtained for the
Ru(tpy)(bpy)(H)+metal hydride, kHT = 1.82  102 M1 s1,56 revealed that even though the free energy
change for Ru(CO)(H)0 to react with CO2 to generate and O-bound formate is thermodynamically less
favorable when compared to Ru(tpy)(bpy)(H)+, it does so three-orders of magnitude faster when compared
to Ru(tpy)(bpy)(H)+, for the same reaction with a similar mechanistic pathway.57, 60 This difference in
kinetic behavior is one example showing that metal hydride reactivity may not simply be determined based
on pure thermodynamic predictions, such as those outlined in the diagram in Scheme 3.
The results and discussions have so far demonstrated that TEOA is the major proton source in
photochemical assays conducted in CH3CN containing BIH or BNAH as the sacrificial electron donor. It
was also demonstrated that TEOA is capable of capturing CO2 in CH3CN solution to form the zwitterionic
alkylcarbonate adduct, CO2-TEOA, with an equilibrium constant of 0.07 M1. This translates into about
20 mM of the CO2-TEOA present in solution at saturation concentrations of free CO2. The possibility of
capturing CO2 into non-gaseous products is by itself very interesting, but combining the capture of CO2
with its reduction is strategically advantageous in terms of reducing the energy requirements to perform
both steps separately. Only a limited number of studies have reported an acceleration in the reduction of
CO2 via hydrogenation pathways in the presence of bases and alcohols, that presumably contained
carbonate, carbamate or alkylcarbonate forms of captured CO2.61-62 Yet, it remains controversial whether
the captured form of CO2 is actively being reduced, as it exists in dynamic equilibrium with free CO2 in
solution. Recently, the synergistic capture of CO2 and its reduction were more convincingly speculated
for a direct insertion of an alkylcarbonate to a Ru(II) metal-hydride catalyst that resulted in the catalytic
production of formate.63 Here, the question of whether CO2-TEOA is actively involved in the reduction
of CO2 to formate catalyzed by the Ru(CO)(H)0 metal hydride, is experimentally addressed.
27
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 28 of 36
1
2
3
4
5
6
7
8
9
When Ru(CO)(H)0 is photochemically generated in the presence of 1.4 M TEOA, the obtained rate
constant for the formation of the singly-reduced Ru(CO)(HCOO)0 is unusually enhanced compared to the
hydride transfer step measured in CH3CN solutions without TEOA, by about 5 orders of magnitude
relative to the CO2 concentration, or 6-orders of magnitude if the zwitterionic CO2-TEOA concentration
is used. It has been categorically demonstrated that the hydride transfer to CO2, for different metal
hydrides, is highly sensitive to the solvation environment, with rate enhancements that span over 2 orders
of magnitude.53, 56-57 This effect was interpreted as the transition state being stabilized in the reaction
coordinate rather than a change in mechanism in the hydride transfer step. Based on this available
comparison, it would appear somewhat exaggerated to address the 5 orders of magnitude rate enhancement
for Ru(CO)(H)0 hydride transfer to free CO2 in 1.4 M TEOA/CH3CN simply based on solvent-like effects
in which TEOA stabilizes the transition state. The large disparity in the measured rate constants for
hydride transfer may instead be indicative of a change in mechanism, in which Ru(CO)(H)0 is being
converted to Ru(CO)(HCOO)0 in the presence of TEOA. For that, one must consider the pre-activated
C=O bonds of the CO2-TEOA zwitterion.
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
Note that bending the CO2 molecule from its original linear equilibrium geometry pre-activates CO2
by making the carbon atom more electrophilic. Under this condition, hydride transfer becomes both
thermodynamically and kinetically more favorable. Unfortunately, activated CO2 does not spontaneously
exist. Although carbamates and alkylcarbonates, which includes the CO2-TEOA zwitterion, are
recognized strategies to activate the C=O bonds, the electron-enriched CO2-carbon makes it a poorer
electrophile and consequently makes the hydride transfer thermodynamically more demanding. Therefore,
the formation of Ru(CO)(HCOO)0 from the hydrido catalyst with the zwitterionic CO2-TEOA may not
proceed through a direct hydride transfer pathway. Instead, a concerted CO2-TEOA insertion/hydride
transfer pathway may provide a thermodynamically inexpensive and kinetically faster alternative for the
overall conversion of CO2 to formate. Taken altogether, the results presented herein provide strong
evidence that captured CO2 in the form of an alkylcarbonate is a promising alternative to access
thermodynamically and kinetically more favorable pathways for the reduction of CO2 to formate, provided
that the metal hydride catalyst is sufficiently hydridic and not too sterically congested to allow for the
inner sphere interaction between the metal and the captured form of CO2.
Formate dissociation. Electrochemical studies of the formato complex in CH3CN have shown that its
singly-reduced form is stable on the cyclic voltammetry timescale, while a second reduction is required
to obtain the formate ion product.11 The electrochemical and TRIR measurements presented here are in
28
ACS Paragon Plus Environment

Page 29 of 36
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
accordance with these early predictions. DFT calculations showed that complete HCOO– dissociation at
the first electron reduction of Ru(CO)(HCOO)+ is endergonic by Gº = 5.2 kcal mol1. However, an
attempt to weaken the Ru–O bond may instead result in formation of a bidentate metal-formate with the
electron delocalized through the O---Ru---O bonds, which is a transition state in the equilibrium between
two O-bound Ru(CO)(HCOO)0 isomers, Scheme 5 (left). Photochemically generated Ru(CO)(HCOO)0
showed no clear evidence for the dissociation of the bound formate within hundreds of microseconds; an
observation that agrees with the sluggish kinetics for the loss of formate obtained from the scan-rate
dependent electrochemical data.
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 5. DFT-computed free energies for formate dissociation.
Previous studies on polypyridyl-metal complexes have indicated that placing more electron density at
the metal strongly influences the stability of anion ligands.64 This context supports the electrochemical
observations of rapid HCOO– dissociation from the doubly-reduced Ru(CO)(HCOO)–. Once HCOO–
dissociates, the six coordinate Ru(CO)(CH3CN)0 may exist transiently and in thermal equilibrium with
the five-coordinate Ru(CO)0 complex in the absence of proton sources, Scheme 5 (right). In fact, the
computed endergonic free energy change, Gº = 13.8 kcal mol1, for this reaction equilibrium Ru(CO)0
+ CH CN
Ru(CO)(CH CN)0 implicates the five-coordinate Ru(CO)0 as the most probable transient
⇌
3
3
intermediate in the absence of proton sources. The experimental data gathered from electrochemical and
sodium amalgam reductions further support this computational analysis.
As described in the Results section, electrochemical reduction of Ru(CO)(HCOO)+ is irreversible at
the first reduction in the presence of TEOA, followed by formation of the metal hydride complex and its
subsequent one-electron reduction to Ru(CO)(H)0. Comparative analysis of the extracted rates for the loss
29
ACS Paragon Plus Environment