Resonance Raman Investigations of [Nife] Hydrogenase Models

Resonance Raman Investigations of [NiFe] Hydrogenase Models

THESIS

Presented in Partial Fulfillment of the Requirements for the Degree Master of Science in the Graduate School of The Ohio State University

Shelby Lee Behnke

Graduate Program in Chemistry

The Ohio State University

2016

Master's Examination Committee:

David Nagib, Committee Member

Hannah Shafaat, Advisor

Copyrighted by

Shelby Lee Behnke

2016

Abstract

Hydrogenase (H2ases) enzymes carry out bidirectional hydrogen production and oxidation reactions. To better understand the mechanism of hydrogen conversion, spectroscopic studies on small molecule mimics provide important metrics to correlate structure and function of the native enzymes. In this work, a series of molecular complexes that mimic the [NiFe] hydrogenase have been synthesized with different phosphine ligand substituents and analyzed using multiple analytical techniques, including nuclear magnetic resonance, Fourier transform infrared, and resonance Raman spectroscopies. Three model compounds have been selected as the focus of this investigation into the vibrational structure of the [NiFe] active site: [Ni(dppe)(µ-pdt)(μ-H)Fe(CO)3][BF4], [Ni(dcpe)(µ- pdt)(μ-H)Fe(CO)3][BF4], and [Ni(dppbz)(µ-pdt)(μ-H)Fe(CO)3][BF4]. (dppe= diphenylphosphinoethane, dcpe= dicyclohexylphosphinoethane, and dppbz= diphenylphosphinobenzene). These compounds have been previously synthesized and shown to exhibit high activity for proton reduction but have not been fully characterized spectroscopically to assess the solution-phase structure of the metal-hydride core. (Barton et al., 2010, JACS. 132, 14877-14885.) Specifically, resonance Raman (RR) spectroscopy was used to study the vibrational bands of each of the molecules, which were then assigned to specific normal modes of the molecule. The [Ni(dppe)(μ-pdt)(μ-H)Fe(CO)3]BF4 compound has been previously studied in great detail by RR, and this molecule was used as a template from which to understand the structures of the series. It was found that

ii perturbations of ligands on the metal center alter the molecular vibrations, particularly of the metal-hydride core. This work provides important metrics for comparison between synthetic and the natural enzyme systems, with the intention of better understanding the mechanisms of native hydrogenases and informing next-generation catalyst design.

iii

This thesis is dedicated to my husband and Revan.

Acknowledgments

I would like to thank my advisor Hannah Shafaat for all of her help and guidance. I am especially grateful to my group who are always there to lend a hand. I would also like to thank the Nagib and Turro group for their generosity and helpfulness. I would also like to acknowledge the support of the OSU Department of Chemistry and Biochemistry,

OSU Institute for Materials Research, and the National Science Foundation.

Vita

2010...... North Augusta High School

2014...... B.S. Chemistry, University of South

Carolina Aiken

2014 to present ...... Graduate Teaching Associate, Department

of Chemistry, The Ohio State University

Publications

Behnke, Shelby. Shafaat, Hannah.; Heterobimetallic Models of the [NiFe] Hydrogenases:

A Structural and Spectroscopic Comparison. Comments on Inorganic Chemistry. 2016,

36, 123-140.

Fields of Study

Major Field: Chemistry

Table of Contents

Abstract ...... ii

Acknowledgments...... v

Vita ...... vi

List of Figures ...... viii

Chapter 1: Background and Introduction ...... 1

Chapter 2: Experimental ...... 13

Chapter 3 Results and Discussion ...... 22

3.1 Synthesis……………..………...…………………………………………….22

3.2 Characterization of Ni(dppe)(μ-pdt)(μ-H)Fe(CO)3 .………………………....22

3.3 Characterization of Ni(dcpe)(μ-pdt)(μ-H)Fe(CO)3 .………………...……….33

3.4 Characterization of Ni(dppbz)(μ-pdt)(μ-H)Fe(CO)3 …..…………………….56

Chapter 4: Conclusion...... 61

References ...... 63

vii

List of Figures

Figure 1. The worldwide energy consumption from 2010 and estimated until 2050.1 ...... 1

Figure 2. Crystal structure of [NiFe] hydrogenase (PDB ID:4U9H) with active site structure highlighted...... 3

Figure 3. Proposed catalytic mechanism and discrete intermediates for hydrogen oxidation by the [NiFe] H2ases...... 5

Figure 4. [NiFe] hydrogenase compound synthesized by Darensbourg group.21 ...... 7

Figure 5. The Pohl group synthesized a [NiFe] hydrogenase mimic in 1997.29 ...... 8

Figure 6. The Schröder group was able to synthesize one of the first models with distorted geometry around the nickel site.31 ...... 8

Figure 7. General structure of the Rauchfuss-type compounds, where L is typically CO or a phosphine ligand and R is typically an alkyl or aryl group.33 ...... 9

Figure 8. The first model with redox activity on nickel center.34 ...... 10

Figure 9. Compounds from left to right: Ni(dcpe)(μ-pdt)(μ-H)Fe(CO)3, Ni(dppe)(μ- pdt)(μ-H)Fe(CO)3, and Ni(dppbz)(μ-pdt)(μ-H)Fe(CO)3. The bond lengths for M-H are listed below each compound.26,36 A crystal structure has not been determined for

Ni(dppbz)(μ-pdt)(μ-H)Fe(CO)3...... 11

Figure 10. Synthetic scheme to generate Ni(dppe)Cl2 ([1])...... 23

31 Figure 11. P-NMR of 1 (600 MHz, CDCl3)...... 24

viii

Figure 12. Synthetic scheme to generate Ni(dppe)(pdt) ([2])...... 25

31 Figure 13. P-NMR of 2 (600 MHz, CDCl3)...... 25

Figure 14. Synthetic scheme to generate Ni(dppe)(μ-pdt)Fe(CO)3 ([3])...... 26

31 Figure 15. P-NMR of 3 (600 MHz, CDCl3), the signal corresponding to Ni(dppe)(μ- pdt)Fe(CO)3 at 45.7 ppm...... 27

Figure 16. FT-IR spectra of the reaction mixture in Fig. 14 at t = 0 hrs (black), t = 3.5 hrs

(blue) and t = 6 hrs (red)...... 27

Figure 17. FT-IR spectrum of [3]. The * denotes a peak due to decomposition of the product...... 28

Figure 18. Synthetic scheme to generate Ni(dppe)(μ-pdt)(μ-H)Fe(CO)3 ([4])...... 29

31 1 Figure 19. P-NMR and (inset) H-NMR of 4 (600 MHz, CD2Cl2)...... 29

Figure 20. Solution-phase FT-IR spectrum of [4] (298 K, CD2Cl2)...... 30

Figure 21. Synthetic scheme to generate Ni(dppe)(μ-pdt)(μ-D)Fe(CO)3 ([5])...... 31

1 Figure 22. The H-NMR spectrum of 5 (600 MHz, CDCl3)...... 31

31 Figure 23. The P-NMR spectrum of 5 (600 MHz, CDCl3)...... 32

Figure 24. FT-IR of ([5])...... 32

Figure 25. Resonance Raman (514.9nm excitation, 77K) of [3] black, [4] blue, and [5] red in dichloromethane. The large solvent peaks have been removed for clarity and are indicated with *. RR spectra are normalized to the peak intensity at 217 cm-1 due to varying concentrations used...... 33

Figure 26. Synthetic scheme to generate Ni(dcpe)Cl2 ([6])...... 34

31 Figure 27. P-NMR of 6 (600 MHz, CDCl3) ([6])...... 34

Figure 28. Synthetic scheme to generate Ni(dcpe)(pdt) ([7])...... 35

31 Figure 29. P-NMR of [7] (600 MHz, CDCl3) ([7])...... 35

Figure 30. Synthetic scheme to generate ([8])...... 36

Figure 31. FT-IR of ([8])...... 36

Figure 32. Synthetic scheme to generate Ni(dcpe)(μ-pdt)(μ-I)Fe(CO)3...... 37

Figure 33. Synthetic scheme to generate Ni(dcpe)(μ-pdt)Fe(CO)3 ([9])...... 38

31 Figure 34. P-NMR of 9 (600 MHz, CDCl3), impurities due to decomposition...... 38

Figure 35. FT-IR of μ-iodo. The peak at 2076 cm-1 is due to decomposition...... 39

Figure 36. The FT-IR of [9], where * donates peaks due to decompositon...... 39

Figure 37. Synthetic scheme to generate Ni(dcpe)(μ-pdt)(μ-H)Fe(CO)3 ([10])...... 40

31 1 Figure 38. P-NMR and H-NMR (inset) of [10] (600 MHz, CD2Cl2)...... 41

Figure 39. FT-IR of [(10)]...... 41

Figure 40. UV-Vis of [9] black and [10] blue...... 42

Figure 41. Synthetic scheme to generate Ni(dcpe)(μ-pdt)(μ-D)Fe(CO)3 ([11])...... 43

1 Figure 42. The H-NMR (600 MHz, CD2Cl2) of [11]...... 44

31 Figure 43. The P-NMR (600 MHz, CD2Cl2) [11]...... 44

Figure 44. Resonance Raman (514.9nm excitation, 77K) of black [9], blue [10], and red

[11] in dichloromethane. Solvent peaks have been removed for clarity and are indicated with *. RR spectra are normalized to the peak intensity at 215 cm-1...... 46

Figure 45. Resonance Raman (514.9nm excitation, 77K) of [11] in dichloromethane without the subtracted quartz spectrum...... 46

Figure 46. Synthetic scheme to generate Ni(dppbz)Cl2 ([12])...... 47

31 Figure 47. P-NMR of [12] (600 MHz, CD2Cl2)...... 48

Figure 48. Synthetic scheme to generate Ni(dcpe)(pdt) ([13])...... 48

31 Figure 49. P-NMR of [13] (600 MHz, CD2Cl2)...... 49

Figure 50. Synthetic scheme to generate Ni(dppbz)(μ-pdt)Fe(CO)3 ([14])...... 50

31 Figure 51. P-NMR of [14] (600 MHz, CD2Cl2)...... 50

Figure 52. The -CO stretches of Ni(dppbz)(μ-pdt)(μ-I)Fe(CO)3...... 51

Figure 53. FT-IR (solution-phase) of [14]. * denotes a peak that is due to partial protonation from exposure to moisture...... 51

Figure 54. Synthetic scheme to generate Ni(dppbz)(μ-pdt)(μ-H)Fe(CO)3 ([15])...... 52

31 1 Figure 55. P-NMR of [15] (600 MHz, CD2Cl2), and the inset H-NMR (600 MHz,

CD2Cl2)...... 53

Figure 56. The carbonyl stretches of Ni(dppbz)(μ-pdt)(μ-H)Fe(CO)3, ([15])...... 53

Figure 57. UV-Vis of [14] black, [15] blue...... 54

Figure 58. Synthetic scheme to generate Ni(dppbz)(μ-pdt)(μ-D)Fe(CO)3 ([16])...... 55

1 Figure 59. H-NMR of 16 (600MHz, CD2Cl2) ...... 55

31 Figure 60. P-NMR of 16 (600 MHz, CD2Cl2)...... 56

Figure 61. Resonance Raman (514.9nm excitation, 77K) of [14] in dichloromethane. The large solvent peaks have been removed for clarity and are indicated with *...... 56

Chapter 1: Background and Introduction

In the past 50 years, the global demand for energy has increased at an alarming rate, with striking economic, political, and ecological consequences.1 An alternative energy storage medium that can be generated and consumed in a sustainable fashion is needed to replace fuels such as coal, oil, and natural gas.2

Figure 1. The worldwide energy consumption from 2010 and estimated until 2050.1

A molecule such as hydrogen is an appealing choice for a renewable fuel because it is small, carbon-free, and has the highest energy density by mass of any possible fuel.3

Hydrogen can be generated from fossil fuels using chemistry that has been well established

1 in industry. The current industrial processes used for hydrogen formation are steam reforming from hydrocarbons, electrolysis, and thermolysis. The process of steam reforming of hydrocarbons occurs at a reformer, which combines hot, high pressure steam with fossil fuels, such as natural gas. Typically, the steam methane reformer is used in industry to make hydrogen. Electrolysis is a technique that uses a direct electric current

(DC) to drive a non-spontaneous chemical reaction such as water oxidation (Eqn. 1) and proton reduction (Eqn. 2).

+ − º 4 2H2O → O2 + 4H + 4푒 E red=1.23 V (1)

+ − º 4 2H + 2푒 → H2 E red=0.00 V (2)

Typically, platinum, ruthenium, and iridium metals are used as the catalysts in electrolysis devices.5 This is a major drawback because these metals are rare and expensive.

The final technique that is typically used on an industrial scale to produce hydrogen in industrial processes is thermolysis. Thermolysis is chemical decomposition caused by extreme heat and this large input of energy is costly. There are also no natural deposits of hydrogen, other than water, because hydrogen gas naturally escapes from Earth’s atmosphere. In the United States, production of hydrogen is estimated to be a 100-billion- dollar industry; as such, it is considered a key economic commodity.6 Over 6 million tons of hydrogen were consumed for oil refining, the Haber-Bosch process, and production of methanol by the reduction of carbon monoxide7. The Haber-Bosch process is an extremely important reaction worldwide for the production of ammonia, which is the main component in fertilizer.8 While the hydrogen conversion reaction looks simple (Eq. 2), to generate and/or utilize hydrogen typically requires either complex biochemical active sites or

2 expensive and rare heavy metals. This process is not necessarily “green” but can provide transitional capacity while an infrastructure is built that develops alternative ways of generating hydrogen using wind, hydroelectric, solar, or nuclear energy.

In nature, enzymes called hydrogenases (H2ases) carry out bidirectional hydrogen production and oxidation at rates rivaling those of platinum, but using only iron and nickel at their active sites.

Figure 2. Crystal structure of [NiFe] hydrogenase (PDB ID:4U9H) with active site structure highlighted.

Like platinum, the natural H2ase enzymes operate at complete thermodynamic efficiency; however, they are sensitive to oxygen inactivation. H2ases are also large, fragile, and costly from a biosynthetic standpoint, which renders large-scale application of these systems

3 impractical. However, understanding the mechanisms of these natural systems can inform development of increasingly efficient molecular catalysts. A great deal of effort has been expended towards the study of H2ase enzymes, with biochemistry research complementing synthetic and theoretical efforts to model and characterize active site structure, spectroscopy, and reactivity. There are three classes of hydrogenase: [NiFe], [FeFe], and

[Fe]-only. All hydrogenases catalyze the uptake of H2, but only the [FeFe] and [NiFe] hydrogenases are reversible redox catalysts, driving both H2 oxidation and proton reduction.9 The [NiFe] hydrogenases represent the largest and most structurally and phylogenetically diverse class of hydrogenase enzyme with both sequence similarity and residues of the active site that are conserved.9 [NiFe] hydrogenases are found throughout the archaeal and prokaryotic kingdoms and play important roles as both hydrogen oxidizing and hydrogen producing enzymes. Across different classes, the accessory cofactors found in hydrogenase-associated subunits vary widely.10 The largest variety comes in the form of the iron-sulfur clusters in hydrogenases. These range from [4Fe-4S] to [2Fe-2S] clusters.8

Non-covalently bound flavins have also been found in [NiFe] H2ases.11 This impacts the catalytic bias and/or likely physiological role of the enzyme.11 However, despite this diversity, all the [NiFe] hydrogenases characterized to date have a common active site, shown in Figure 2, as determined by spectroscopic and functional properties.12

This active site possesses several unique features. The heterobimetallic cluster contains an iron(II) center that is terminally coordinated by two cyanide ligands and one carbonyl ligand. The iron is bridged via two cysteine thiolates to a nickel center that also contains two terminally-coordinated cysteine thiolates. In 2009, the Ogata and Lubitz

4 groups proposed a catalytic mechanism for activity based on X-ray crystallography and spectroscopic data of the [NiFe] H2ase.13 Owing to the electron-withdrawing nature of the

CN- and CO ligands, the iron site remains low-spin Fe(II) throughout the catalytic mechanism, as seen in figure 3.

Figure 3. Proposed catalytic mechanism and discrete intermediates for hydrogen oxidation by the [NiFe] H2ases.

In contrast, the nickel center cycles through the Ni(I), Ni(II), and Ni(III) oxidation states. There are two inactive states, the Ni-A, or “unready”, state and the Ni-B, or “ready”, state. These names are due to the slow reactivation kinetics of the Ni-A state compared to the rapid activation of the Ni-B state. The difference in kinetics is attributed to the different chemical nature of the bridging ligands.14 These two forms are the most oxidized states of the [NiFe] center and are activated by a one-electron reduction coupled to a proton transfer to generate the Ni-SIa state. The active forms of the enzyme are the Ni-SIa, Ni-L, Ni-C, and 5

Ni-R states. A one-electron reduction and protonation of the Ni-SIa state will yield either the Ni-C or the Ni-L state. The Ni-C state has been studied extensively with electron paramagnetic resonance (EPR) spectroscopy to determine that the Ni center is in the +3 formal oxidation state with an S=1/2 spin state and a hydride bridging between the two metals.15 In a recent study, the bond distances of the Ni-H and Fe-H were reported at 1.58

Å and 1.78 Å, respectively.16 The Ni-L state can be obtained following photolysis of the

17 18 Ni-C state. This was determined with x-ray crystallography and RR. The Ni-SIa state can be inhibited by CO, which binds to the Ni center in a bent fashion.17 The only variation that has been observed in the primary coordination sphere is the substitution of one terminal cysteine ligand by a selenocysteine.19 X-ray crystal structures of these oxidized proteins have revealed other variations including sulfinate- and sulfenate-modified

(seleno)cysteines, though the physiological relevance of these modifications has yet to be determined.19

Hydrogenases can potentially be used in bioelectrical devices or for biological hydrogen production, though most [NiFe] hydrogenases are inactivated by oxygen. The few known O2-tolerant enzymes are hydrogen-uptake enzymes, unsuitable for hydrogen production due to strong product inhibition. In contrast, the [NiFeSe] hydrogenases, where a selenocysteine is bound to the nickel, are very attractive alternatives because of their high hydrogen production activity and fast reactivation after O2 exposure; however, these enzymes are still costly to isolate or express.

Active site mimics have been used to reproduce spectral metrics or functional parameters of the native enzyme in an attempt to gain insight into the catalytic mechanism.

This approach has been highly successful for the [NiFe] H2ase; multiple groups have successfully developed compounds with varying ligand sets in order to reproduce the electronic structure of both metal centers.20–28 The first mimics were synthesized by the

Darensbourg group, closely followed by the Pohl group.21,29

Figure 4. [NiFe] hydrogenase compound synthesized by Darensbourg group.21

These mimics were able to reproduce both the heterometallic nature of the active site and key Ni-Fe bond distances. While the first protein crystal structure showed a Ni-Fe bond distance of 2.7 Å,30 the Darensbourg compound displayed a Ni-Fe distance of 3.76 Å.

Figure 5. The Pohl group synthesized a [NiFe] hydrogenase mimic in 1997.29 In contrast, the Pohl compound featured a metal-metal separation of 2.80 Å.21 A short time later, the Schröder group synthesized one of the first active site mimics containing distorted coordination around the nickel site.31

Figure 6. The Schröder group was able to synthesize one of the first models with distorted geometry around the nickel site.31

Prior to this work, the geometry around the nickel site was always observed to be square planar, with square pyramidal or octahedral coordination around the iron site.32

However, with the Schröder model, geometric rearrangement at the nickel center is seen upon coordination of the nickel precursor to the iron fragment.31 Specifically, the geometry

8 changes from square planar to near tetrahedral.31 In 2009, the first heterobimetallic model compound containing a bridging hydride was reported by the Rauchfuss group.33

Figure 7. General structure of the Rauchfuss-type compounds, where L is typically CO or a phosphine ligand and R is typically an alkyl or aryl group.33

This compound was generated by protonation of the heterobimetallic Schröder compound with an organic acid.33 Since these landmark studies, a number of other models and derivatives have been synthesized. Notably, there were no bimetallic compounds that showed redox activity at the nickel center until recently, when the Artero group and coworkers were able to synthesize a functional [NiFe] model displaying redox activity on the Ni center.34

Figure 8. The first model with redox activity on nickel center.34

Currently, the structural H2ase mimics are primarily able to catalyze H2 evolution, rather than H2 oxidation, in organic solvents. However, this reaction occurs at negative potentials, approximately -1.3 to -2.2 V vs. NHE, which translates into a 0.9 to 1.8 V overpotential.35

The discussed models are improvements over previous models, but key questions still remain, such as site of hydrogen binding and the role iron plays during catalysis. To address these questions, there has been continued investigation into synthetic systems mimicking the [NiFe] H2ase.

We have elected to investigate three model compounds: [Ni(dppe)(µ-pdt)(μ-

H)Fe(CO)3][BF4], [Ni(dcpe)(μ-pdt)(μ-H)Fe(CO)3][BF4], and [Ni(dppbz)(μ-pdt)(μ-

H)Fe(CO)3][BF4]. We have also investigated the unprotonated precursor molecules:

Ni(dppe)(µ-pdt)Fe(CO)3, Ni(dcpe)(µ-pdt)Fe(CO)3, and Ni(dppbz)(µ-pdt)Fe(CO)3.

Bond length: Bond length: Ni-H: 1.91 Å Ni-H: 1.89 Å Fe-H: 1.53 Å Fe-H: 1.49 Å

Figure 9. Compounds from left to right: Ni(dcpe)(μ-pdt)(μ-H)Fe(CO)3, Ni(dppe)(μ-pdt)(μ- H)Fe(CO)3, and Ni(dppbz)(μ-pdt)(μ-H)Fe(CO)3. The bond lengths for M-H are listed below each compound.26,36 A crystal structure has not been determined for Ni(dppbz)(μ- pdt)(μ-H)Fe(CO)3.

These compounds, previously made by the Rauchfuss group,37 have shown activity, but have not been fully characterized spectroscopically. The compounds chosen for this study can be treated as part of a series since they have similar structures, only containing variations in substituent on the nickel phosphine ligands. In this work, the above compounds were synthesized, and the structures were confirmed using nuclear magnetic resonance (NMR) and Fourier transform infrared spectroscopy (FT-IR). Following structural verification, these compounds were further characterized using resonance Raman

(RR) to study the vibrational modes, which were then assigned to specific stretches in the molecule. The [Ni(dppe)(μ-pdt)(μ-H)Fe(CO)3][BF4] compound has been previously studied by RR in great detail.38 This previous study allowed us to use this compound as a comparison to better inform on the metal-hydride cores of the other target molecules. This

11 knowledge will help determine how perturbing ligand substituents alters the structure of the metal-hydride core.

Chapter 2: Experimental

Unless otherwise noted, all reactions were performed using standard Schlenk line techniques at room temperature. Methylene chloride and toluene were distilled before use.

Solvents used in the glovebox were degassed by at least 3 freeze-pump-thaw cycles. When noted, sample preparation and purification protocols were carried out in an anaerobic glovebox (Vigor Industries, <10 ppm O2).

Spectroscopy: All NMR spectra were recorded using a Bruker Advance III HD Ascend

Wide Bore 600 MHz NMR spectrometer. All UV-Vis spectroscopy was performed on a

UV-2600 Shimadzu. The FT-IR spectra were taken using a Thermo Scientific Nicolet iS10 with an attenuated total reflectance accessory, Smart iTR, or using a Bruker Tensor 27 with an air-tight cell. All resonance Raman data was collected at 77 K with samples contained inside 4 mm EPR tubes within a liquid nitrogen finger dewar (Wilmad 50 mL Supras

Nitrogen Dewar Flask). A mixed gas Kr-Ar laser (Coherent Innova Spectrum 70-C, Laser

Innovations) provided excitation at 514.5 nm. The excitation beam was focused onto the sample using a 100 mm RFL, 90º off-axis parabolic mirror (Thorlabs). Scattered light was collected at 135º backscattering geometry by a 50 mm FD f/0.8 UV fused silica aspheric lens (Edmund Optics, 514.5 nm excitation) and focused onto the 100 μm entrance slit of a spectrograph using a 50 mm, f/4 uncoated UV fused silica lens (CVI Optics, 514.5 nm 13 excitation). Rayleigh scattering was rejected by the 514.5 nm long-pass edge filter

(Semrock RazorEdge). Raman scattered light was dispersed in an f/4.6, 0.32 m imaging spectrograph (Princeton Instruments, IsoPlane SCT 320) equipped with a 1200 gr/mm

(514.5nm) and imaged onto a Peltier-cooled CCD detector (Princeton Instruments, Pixis

100B).

Raman signal intensity was optimized and calibrated using known bands from a 1:1 v/v mixture of toluene and acetonitrile.39 The band pass and accuracy were found to be 6 cm-1 and +/- 1 cm-1, respectively. One-minute signal integrations were summed over one hour for each sample to obtain accumulated spectra of 5 mM Ni(dppe)(μ-pdt)Fe(CO)3, 5 mM Ni(dppe)(μ-pdt)(μ-H)Fe(CO)3, 1.5 mM Ni(dppe)(μ-pdt)(μ-D)Fe(CO)3, 5 mM

Ni(dcpe)(μ-pdt)Fe(CO)3, 5 mM Ni(dcpe)(μ-pdt)(μ-H)Fe(CO)3, 2 mM Ni(dcpe)(μ-pdt)(μ-

D)Fe(CO)3, and 5 mM Ni(dppbz)(μ-pdt)Fe(CO)3.The complete dppe, dcpe, and dppbz series were run in DCM and d2-DCM. Single-pixel spikes due to cosmic rays impinging upon the detector were removed manually. Features from the appropriate solvent were subtracted from the spectra of the compounds and a spline was drawn along the baseline of the resulting spectrum and subtracted to remove a broad fluorescent background.

Compounds: All compounds obtained were used without further purification. Nickel chloride hexahydrate, 1,3-propanedithiol (pdt), triethylamine (TEA), sodium methoxide

(NaOMe), and tetrafluoroboric acid diethyl ether complex were obtained from Alfa Aesar.

1,2-bis(diphenylphosphino)ethane (dppe), 1,2-bis(dicyclohexylphosphino)ethane (dcpe), d1-cholorform, d2-dichloromethane, d3-acetronitrile, and 1,2-

14 bis(diphenylphosphino)benzene (dppbz) were obtained from Acros Organics. Celite was obtained from Millipore. Deuterium oxide, iron pentacarbonyl, diiron nonacarbonyl, cobaltocene, and iodide were obtained from Fisher Scientific.

40 Ni(dppe)Cl2 ([1]): In a round bottom flask with a stir bar, NiCl2*6H2O (0.36944 g,

0.00155 mol) was added in ethanol (5 mL) to give a green suspension. While stirring, a solution of dppe (0.45334 g, 0.001138 mol) in dichloromethane was pipetted into solution, resulting in an orange precipitate. This suspension was stirred for 15 mins under N2. The product was collected via vacuum filtration and washed with ether. The orange powder

31 was collected (Yield: 0.4749g [1], 97%). P-NMR (CDCl3, 23ºC): δ 78.1 ppm.

40 Ni(dppe)(pdt) ([2]): In a round bottom flask with a stir bar, Ni(dppe)Cl2 (0.3284 g,

0.00062 mol) was suspended in toluene (1.2 mL). 1,3-propanedithiol (0.1174 mL,

0.001635 mol) was added to the suspension, followed by triethylamine (0.4415 mL,

0.00316 mol). The orange solution immediately turned red. The suspension was allowed to stir for 20 mins, after which the solid was collected by vacuum filtration and washed with 30 mL ethanol and 30 mL water. The red crystals were dissolved in a minimal volume of dichloromethane, layered with diethyl ether, and stored at -20°C for 30 min. Red crystals

31 of Ni(dppe)(pdt) were obtained and dried on a high vacuum line. P-NMR (CDCl3, 23ºC):

δ 56.1 ppm (Yield: 0.2978 g [2], 85%).

26 Ni(dppe)(μ-pdt)Fe(CO)3 ([3]): In a Schlenk flask with a stir bar, solid Ni(dppe)(pdt)

(0.20849 g, 0.00037 mol) and solid Fe2(CO)9 (0.19749 g, 0.000542 mol) were combined inside the glovebox. Dichloromethane (1.5mL) was added, and the reaction was stirred for

6 hours at room temperature. The reaction was monitored via FT-IR during this time. It was noted that the solution acquired a greenish tint when nearing completion. Upon completion, the solution was evaporated to dryness. Inside the glovebox, residual starting material was washed with MeCN until the solution became clear. The remaining compound was dissolved in a minimal amount of dichloromethane and run through a neutral alumina oxide column (activated, neutral, 60 mesh powder, S.A. 150 m2/g) to separate out decomposed material from product and unreacted precursor. A green solution containing

31 product was collected and concentrated to dryness. P-NMR (CDCl3, 23ºC): δ 45.7 ppm

FT-IR: 1935, 1993(sh), 2027 cm-1 Yield: 0.04945 g [3], 19%

26 Ni(dppe)(μ-pdt)(μ-H)Fe(CO)3 ([4]): In a Schlenk flask with a stir bar, Ni(dppe)(μ- pdt)Fe(CO)3 (0.020 g, 0.000028mol) was added to DCM (2 mL) inside of the glovebox.

HBF4*Et2O (27 μl, 0.03213 g, 0.000198 mol) was pipetted in. The green solution turned red immediately and was allowed to stir for 5 mins. The reaction was concentrated to

1 31 dryness. H-NMR (CDCl3, 23ºC): δ -3.5 P-NMR (CD2Cl2, 23ºC): δ 70.1 ppm FT-IR:

2020, 2079 cm-1 Yield: 0.01202 g [4], 60%.

26 Ni(dppe)(μ-pdt)(μ-D)Fe(CO)3([5]): In a 20 mL scintillation vial with a stir bar,

Ni(dppe)(μ-pdt)(μ-H)Fe(CO)3 (0.010g, 0.000014mol) was combined with 20 μL of D2O,

0.5mL of d1-chloroform, and 0.1mL of d3-acetonitrile inside of the glovebox. The solution

31 stirred for 16 hours. P-NMR (CD2Cl2, 23ºC): δ 70.2 ppm Yield: 0.0098 g [5], 99%

36 Ni(dcpe)Cl2 ([6]): In a round bottom flask with a stir bar, NiCl2*6H2O (0.44539 g,

0.00187 mol) was suspended in ethanol (4 mL). Solid dcpe (0.2453 g, 0.00058 mol) was added to the stirring reaction. The reaction changes from green to orange immediately and was stirred for 20 mins under N2. The precipitate was collected by vacuum filtration and

31 washed with diethyl ether. The orange solid was collected. P-NMR (CDCl3, 23ºC): δ 81.3 ppm Yield: 0.25808 g [6], 81 %

36 Ni(dcpe)(pdt) ([7]): In a round bottom flask with a stir bar, Ni(dcpe)Cl2 (0.26105 g,

0.000472 mol) was dissolved in 3 mL of DCM. While stirring, 1,3-propanedithiol (0.05114 mL, 0.0542 g, 0.000712 mol) was pipetted into the solution, followed by a solution of

CH3ONa (0.10224 g, 0.00279 mol) in 1 mL MeOH. The solution changed from orange to red immediately and was allowed to stir at room temperature for 15 mins. The solvent was then evaporated to dryness on a vacuum line, and the solid dissolved in 1 mL of DCM. The solution was vacuum-filtered through a Celite pad, and the red flow-through was collected

31 and concentrated to dryness. P-NMR (CDCl3, 23ºC): δ 73.0 ppm Yield: 0.22266 g [7],

81%

41 FeI2(CO)4 ([8]): In a Schlenk flask with a stir bar, a solution containing diethyl ether (5 mL) and Fe(CO)5 (1.62 mL, 0.0119 mol) was made inside the glovebox. I2 (3.025 g, 0.0119 mol) was dissolved in benzene (10 mL) and added to the Schlenk flask. The dark purple

17 solution was allowed to stir for 2 hours inside the glovebox at room temperature. The solution was then concentrated to dryness. FT-IR: 2038 (m), 2070 (s), 2131 (sh) cm-1 Yield:

2.6335 g [8], 97 %

36 Ni(dcpe)(μ-pdt)Fe(CO)3 ([9]): In one Schlenk flask, Ni(dcpe)(pdt) (0.01593 g, 0.000027 mol) and FeI2(CO)4 (0.0115 g, 0.000027 mol) were combined. In another Schlenk flask, cobaltocene (0.0103 g, 0.000054 mol) was added. Both flasks were cooled to -78˚ C in a dry ice/acetone bath. In the bimetallic flask, 0.7 mL of DCM was added and the formation of Ni(dcpe)(μ-pdt)(μ-I)Fe(CO)3 was checked by FT-IR. Pre-cooled DCM (0.5 mL) was added to the cobaltocene flask. The bimetallic solution containing Ni(dcpe)(μ-pdt)(μ-

I)Fe(CO)3 was transferred into the flask containing cobaltocene by cannula transfer with purified nitrogen. The reaction vessel was allowed to warm to room temperature and stirred for one hour. The solution was then removed under vacuum. Inside the glovebox, the compound was washed with acetonitrile until washings were clear, and the remaining solid was dissolved in minimal DCM. Hexanes were added to the solution to precipitate out solid, which was collected via decanting and dried on the vacuum line. FT-IR (Ni(dcpe)(μ-

-1 -1 31 pdt)(μ-I)Fe(CO)3): 2022, 2053, *2077, 2095 cm FT-IR [9]: 217, 2077 cm P-NMR

(CD2Cl2, 23ºC): δ 79.8 ppm Yield: 0.00493 g [9], 25 % *This peak is due to partial formation of [9].

36 Ni(dcpe)(μ-pdt)(μ-H)Fe(CO)3 ([10]): In a Schlenk flask with a stir bar, Ni(dcpe)(μ- pdt)Fe(CO)3 (0.002 g, 2.75µmol) in 2 mL DCM, was combined with HBF4*Et2O (18 μL,

0.00013 mol) and stirred for 5 mins inside the glovebox. The solution turned from green- brown to dark orange. The solvent was removed under vacuum. FT-IR: 2012, 2074 cm-1

1 31 H-NMR: (CD2Cl2, 23ºC): δ -2.93 P-NMR (CDCl3, 23ºC): δ 89.6 ppm Yield: 0.00064 g

[10], 32 %

36 Ni(dcpe)(μ-pdt)(μ-D)Fe(CO)3([11]): In a scintillation vial with a stir bar, Ni(dcpe)(μ- pdt)(μ-H)Fe(CO)3 (0.00064 g, 0.000879 mmol) was combined with 20 μL of D2O, 0.5mL of d-chloroform, and 0.1mL of d-acetonitrile inside of the glovebox. The solution stirred

31 for 16 hours. P-NMR (CD2Cl2, 23ºC): δ 89.6 ppm Yield: 0.000634 g [11], 99 %

36 Ni(dppbz)Cl2 ([12]): In a round bottom flask with a stir bar, NiCl2*6H2O (0.12342 g,

0.000519 mol) was combined with 3 mL ethanol under an N2 atmosphere. While stirring, a solution of dppbz (0.22484 g, 0.0050 mol) in 3 mL DCM was added. The suspension was stirred for 15 min, collected by vacuum filtration, and washed with diethyl ether. 31P-NMR

(CD2Cl2, 23ºC): δ 57.6 ppm Yield: 0.14555 g [12], 81%

36 Ni(dppbz)(pdt) ([13]): In a round bottom flask with a stir bar, a solution of Ni(dppbz)Cl2

(0.16186 g, 0.00028 mol) in 3 mL DCM was combined with 1, 3-propanedithiol (0.0285 mL, 0.03021 g, 0.000397 mol) followed by triethylamine (0.1557 mL, 0.11296 g, 0.00112 mol). The solution changed from orange to dark purple. The solution was allowed to stir for 20 min, then 25 mL of diethyl ether was added to precipitate out the product. The solid

31 was collected by vacuum filtration and washed with diethyl ether. P-NMR (CD2Cl2,

23ºC): δ 61.5 ppm Yield: 0.13200 g [13], 95 %

36 Ni(dppbz)(μ-pdt)Fe(CO)3 ([14]): In one Schlenk flask, Ni(dppbz)(pdt) (0.07498 g,

0.00011 mol) and FeI2(CO)4 (0.05124 g, 0.00012 mol) were combined. In another Schlenk flask, cobaltocene (0.1044 g, 0.000552 mol) was added. Both flasks were cooled to -78˚ C in a dry ice/acetone bath. In the bimetallic flask, 0.7 mL of DCM was added, and the formation of Ni(dppbz)(μ-pdt)(μ-I)Fe(CO)3 was checked by FT-IR. Pre-cooled DCM (0.5 mL) was added to the cobaltocene flask. The bimetallic solution containing Ni(dppbz)(μ- pdt)(μ-I)Fe(CO)3 was transferred into the flask containing cobaltocene by cannula transfer with purified nitrogen. The reaction vessel warmed to room temperature and stirred for one hour. The solution was then removed under vacuum. Inside the glovebox, the compound was dissolved in the minimal amount of DCM and run through a silica column to collect unreacted precursors. The solvent was concentrated to 1 mL DCM, hexanes were added to the solution to precipitate out solid, which was collected via decanting and dried on the

31 vacuum line. P-NMR (CD2Cl2, 23ºC): δ 59.5 ppm Yield: 0.0.0108 g [14], 12% FT-IR

-1 -1 (Ni(dppbz)(μ-pdt)(μ-I)Fe(CO)3): 2023, 2047, 2092 cm FT-IR[14]: 1940, 2030 cm

36 Ni(dppbz)(μ-pdt)(μ-H)Fe(CO)3 ([15]): In a Schlenk flask with a stir bar, Ni(dppbz)(μ- pdt)Fe(CO)3 (0.0270 g, 0.000033 mol) in 2 mL DCM was combined with HBF4*Et2O (45

μL, 0.05355 g, 0.00033 mol) and stirred for 5 min inside the glovebox. The solution turned from green-brown to dark orange. The solvent was removed under vacuum. 1H-NMR

31 (CD2Cl2, 23ºC): δ -3.4 ppm P-NMR (CD2Cl2, 23ºC): δ 68.5 ppm Yield: 0.00865 g [15],

32% FT-IR: 2022, 2081 cm-1

36 Ni(dppbz)(μ-pdt)(μ-D)Fe(CO)3([16]): In a scintillation vial with a stir bar,

Ni(dppbz)(μ-pdt)(μ-H)Fe(CO)3 (0.0082 g, .00994 mmol) was combined with 20 μL of

D2O, 0.5mL of d1-chloroform, and 0.1mL of d3-acetonitrile inside of the glovebox. The

31 solution stirred for 16 hours. P-NMR (CD2Cl2, 23ºC): δ 68.8 ppm Yield: 0.00857 g [16],

99 %

Chapter 3 Results and Discussion

3.1 Synthesis

In this work, a series of molecular complexes that mimic the [NiFe] hydrogenase have been synthesized with different phosphine ligand substituents and analyzed with multiple spectroscopic techniques, such as nuclear magnetic resonance, Fourier transform infrared spectroscopy, and resonance Raman. Synthesizing the Ni(diphosphine)(μ-pdt)(μ-

H)Fe(CO)3 series is advantageous because the dppe variant has been studied in great detail previously.38 This work laid an important foundation that can be applied to the metal centers of the Ni(dcpe)(μ-pdt)(μ-H)Fe(CO)3 and the Ni(dppbz)(μ-pdt)(μ-H)Fe(CO)3 compounds.

3.2 Characterization of Ni(dppe)(μ-pdt)(μ-H)Fe(CO)3

In order to make the bimetallic complex, each metal fragment must be separately synthesized. The precursors are easily generated in relatively high yields. The first step is to attach the desired phosphine ligand to the nickel center; which provides flexibility for tuning the electronic properties of the active site. The first step is performed by suspending

22 hydrated nickel(II) chloride in ethanol. While stirring, the desired phosphine ligand is added. The nickel phosphine complex is rapidly formed, displacing the water ligands from the metal center, and the product crashes out of solution. The solid product is then easily collected by vacuum filtration. This process works well for all three ligand sets. A schematic diagram for the synthesis of 1 can be seen in Figure 10. Compound 1 was synthesized and the structure verified through 31P-NMR spectroscopy (Figure 11). The peak at 78.1 ppm corresponds to the phosphine ligand on the nickel center.26

Stir, RT 25 mins

Figure 10. Synthetic scheme to generate Ni(dppe)Cl2 ([1]).

31 Figure 11. P-NMR of 1 (600 MHz, CDCl3).

The nickel center requires additional modification for attachment to the iron center.

For the target compounds in this project, a propanedithiolate (pdt) bridge was used.

Deprotonation of the parent thiol is accomplished by the addition of two equivalents of triethylamine, a mild base; coordination to the metal center occurs rapidly following deprotonation. The resulting products, NiII(pdt)(diphosphine), are square planar and in this case diamagnetic, which is illustrated in Figure 12. 31P-NMR was also used to verify the structure of compound 2, as seen below. The peak at 56.1 ppm corresponds to the diphenylphosphinoethane peak (Figure 13).26

2 equiv.

Toluene

RT stir 15 mins

Figure 12. Synthetic scheme to generate Ni(dppe)(pdt) ([2]).

31 Figure 13. P-NMR of 2 (600 MHz, CDCl3).

The iron fragment is synthesized separately from the nickel center. In the case of the bimetallic reaction with Ni(dppe)(pdt), the commercially-available di-iron 25 nonacarbonyl starting material is used. (Figure 14) A considerable amount of undesirable byproducts are formed during the reaction, which must be removed through multiple purification steps. Compound 3 can be quantified by NMR as well as FT-IR spectroscopy, which monitors the carbonyl ligands on the iron center (Figure 15). The reaction progress is monitored by FT-IR, and after six hours of stirring at room temperature, the reaction was deemed complete (Figure 16). However, the FT-IR spectrum of the –CO ligands of 3 is more complicated than expected probably due to sample decomposition, as this compound is found to be moisture- and air- sensitive, and is unstable for long periods even under anaerobic conditions. The stretches at 1935, 1993 (sh), and 2027 cm-1 correspond to the three carbonyl ligands of 3 (Figure 17).26

6 hr stir

DCM

Figure 14. Synthetic scheme to generate Ni(dppe)(μ-pdt)Fe(CO)3 ([3]).

31 Figure 15. P-NMR of 3 (600 MHz, CDCl3), the signal corresponding to Ni(dppe)(μ- pdt)Fe(CO)3 at 45.7 ppm.

time 0 time 3.5 hours time 6 hours

% Transmittance % 2027

20 1981 1933

1800 2000 2200 2400 Wavenumber

Figure 16. FT-IR spectra of the reaction mixture in Fig. 14 at t = 0 hrs (black), t = 3.5 hrs (blue) and t = 6 hrs (red).

100

* 70

% Transmittance % 60

1993 2027

50 1935 * 1800 1900 2000 2100 2200 Wavenumber

Figure 17. FT-IR spectrum of [3]. The * denotes a peak due to decomposition of the product. At this stage, all three compounds are then protonated with a strong non- coordinating acid. HBF4·Et2O is added to a stirring solution of each bimetallic compound in order to generate the bridging hydride (Figure 18). This hydride is easily seen in NMR

(Figure 19). After protonation, the compounds are now formally Ni(II)-Fe(II) species. The band at -3.5 ppm corresponds to the hydride.26 Because the metals are both oxidized at this point, compound 4 is more oxygen- and moisture-tolerant than 3.

3.5 equiv. stir 5 mins

DCM

Figure 18. Synthetic scheme to generate Ni(dppe)(μ-pdt)(μ-H)Fe(CO)3 ([4]).

31 1 Figure 19. P-NMR and (inset) H-NMR of 4 (600 MHz, CD2Cl2).

Hydrides typically appear at a wide range of negative values due to excited states and strong spin orbit coupling.42 The stretches at 2020 and 2079 cm-1 correspond to the three

26 carbonyl ligands of [Ni(dppe)(μ-pdt)(μ-H)Fe(CO)3][BF4] (Figure 20). 29

90 2020

% Transmittance % 88 2079

1900 2000 2100 Wavenumber

Figure 20. Solution-phase FT-IR spectrum of [4] (298 K, CD2Cl2).

A D2O exchange was performed in order to clearly identify the metal-hydride vibrational bands. This exchange requires a 16-hour stir in D2O (0.55 mmol) which is represented in Figure 21. While these compounds are moderately oxygen-tolerant, they remain semi-sensitive to moisture, and thus care must be taken to avoid decomposition. As shown in Figure 22, evidence for hydride exchange is seen in the 1H-NMR spectrum, which now lacks a peak at -3.5 ppm.38 The 31P-NMR spectrum (Figure 23) shows that the compound remains unchanged from 4. Further evidence supporting the formation of 5 can

30 be seen in the FT-IR spectrum (Figure 24), which shows bands at 2020 and 2079 cm-1 corresponding to the carbonyl ligands in Ni(dppe)(μ-pdt)(μ-D)Fe(CO)3 ([5]).

Stir 16 hrs

Figure 21. Synthetic scheme to generate Ni(dppe)(μ-pdt)(μ-D)Fe(CO)3 ([5]).

1 Figure 22. The H-NMR spectrum of 5 (600 MHz, CDCl3).

31 Figure 23. The P-NMR spectrum of 5 (600 MHz, CDCl3).

100

90 % transmittance %

2018 2079

1900 2000 2100 2200 Wavenumbers

Figure 24. FT-IR of ([5]).

Resonance Raman spectroscopy was performed on 3, 4, and 5 in DCM and d2-

DCM. Figure 25 compares the vibrational spectra of each of these compounds. The labeled peaks correspond to vibrations that are notably different between the compounds. Bands at

450, 492, 522, 617, and 1028 cm-1 remain relatively unchanged between 3 and 4. In 4, additional broad bands can be seen at 954, 1476, and 1533 cm-1, respectively. Bands at

1957 and 2028 cm-1 represent the –CO ligands in 3.38 In 4, the –CO ligand vibrations seen to shift to 2020 and 2082 cm-1 from compound 3. The spectrum of 5 also contains peaks at

2080 and 2020 cm-1, which represent –CO ligands.

2028

* 1957

954

1476

1533

2020

2082

2080

2020 Normalized Raman Intensity Raman Normalized

500 1000 1500 2000 -1 Raman Shift (cm )

3.3 Characterization of Ni(dcpe)(μ-pdt)(μ-H)Fe(CO)3

As before, the nickel precursor compounds are synthesized from stirring the reactants at room temperature (Figure 26). Isolation of the product is performed by precipitation with excess solvent, and the purification accomplished by recrystallization.

The synthesis of compound 6 was verified with NMR spectroscopy (Figure 27).

20 min stir

ethanol

Figure 26. Synthetic scheme to generate Ni(dcpe)Cl2 ([6]).

31 Figure 27. P-NMR of 6 (600 MHz, CDCl3) ([6]). 34

Similarly, the thiolate-bound precursor [7] was synthesized as in [2]. The reaction scheme can be seen in Figure 28, with the 31P NMR spectrum shown in Figure 29.

15 min stir

DCM/MeOH

Figure 28. Synthetic scheme to generate Ni(dcpe)(pdt) ([7]).

31 Figure 29. P-NMR of [7] (600 MHz, CDCl3) ([7]).

The iron center of the bimetallic complex was synthesized separately, as shown in

Figure 30. FT-IR was used to verify the carbonyl stretches of 8. The peaks at 2038, 2070

-1 41 , 2131 cm correspond to the carbonyl stretches of FeI2(CO)4.

2 hr stir

Benzene, N2

Figure 30. Synthetic scheme to generate ([8]).

40 2131 % transmittance %

2038 2070

1800 1900 2000 2100 2200 2300 wavenumbers

Figure 31. FT-IR of ([8]).

A different synthetic protocol is required to generate the bimetallic [NiFe] compounds with dcpe and dppbz ligands. In this case, the bimetallic compound proceeds

36 via a μ-I intermediate, as shown in Figure 33. The FT-IR spectrum of the bimetallic μ-I intermediate shows bands at 2022, 2053, 2077, and 2095 cm-1 (Figure 35).36 This intermediate forms rapidly after mixing of FeI2(CO)4 and Ni(diphosphine)(pdt). In order to remove the bridging iodide ligand, 3 equivalents of cobaltocene are added at while in a dry ice/acetone bath. An excess of cobaltocene was needed due to solubility issues in solution. This yields a forest green colored product which, in all cases, yields a Ni(I)-Fe(I) bimetallic compound (Figure 32). The FT-IR spectrum of compound 9 shows bands at

2017 and 2077 cm-1, and a single peak in the 31P-NMR spectrum at 80.1 ppm (Figure 34).36

This reaction was attempted with the commercially available iron starting material,

Fe2(CO)9, but the synthesis failed to produce the desired bimetallic product.

Cannula transfer

77 K

Figure 32. Synthetic scheme to generate Ni(dcpe)(μ-pdt)(μ-I)Fe(CO)3.

Cannula transfer

77 K

Figure 33. Synthetic scheme to generate Ni(dcpe)(μ-pdt)Fe(CO)3 ([9]).

31 Figure 34. P-NMR of 9 (600 MHz, CDCl3), impurities due to decomposition.

200

100

% transmittance %

2076

2053 2022

-100 2095

2000 2050 2100 2150 Wavenumbers

Figure 35. FT-IR of μ-iodo. The peak at 2076 cm-1 is due to decomposition.

Figure 36. The FT-IR of [9], where * donates peaks due to decomposition.

The addition of a strong, non-coordinating acid to 9 yields the hydride complex,

10, as before. This reaction stirs for 5 mins in an anaerobic environment until the solvent was removed under vacuum. The 31P-NMR spectrum of 10 shows a single band at

89.6ppm, while the hydride signal is evident in the 1H-NMR spectrum at -2.93 ppm (Figure

38). The FT-IR spectrum of 10 shows carbonyl bands at 2011 and 2074 cm-1 (Figure 39).36

5 min stir

DCM

Figure 37. Synthetic scheme to generate Ni(dcpe)(μ-pdt)(μ-H)Fe(CO)3 ([10]).

31 1 Figure 38. P-NMR and H-NMR (inset) of [10] (600 MHz, CD2Cl2).

0.3

0.2 2074 2011

0.1 Absorabance

0.0

-0.1

2000 2100 2200 Wavenumbers

Figure 39. FT-IR of [(10)]. 41

UV-Vis absorption data was collected to determine an ideal wavelength for performing resonance Raman spectroscopy on these compounds (Figure 40). Due to the broad band centered at 525 nm, an excitation wavelength of 514.5 nm was chosen to study

9-11 using RR spectroscopy, other wavelengths were used but due to fluorescence in the samples 514.5 nm was determined to be ideal.

Figure 40. UV-Vis of [9] black and [10] blue.

Bands from the metal-hydride core can be identified through isotopic exchange experiments. To replace the hydride bridge of 10 with a deuterium, an exchange reaction

42 is carried out with D2O (Figure 41). The proton signal in NMR spectrum corresponding to the hydride is seen to disappear after stirring for sixteen hours at room temperature under an anaerobic atmosphere. The phosphine ligand signal in Figure 42 can be seen at 89.6 ppm, which is unchanged from species 10 in the 31P-NMR, suggesting the compound remains intact.

16 hr stir

Figure 41. Synthetic scheme to generate Ni(dcpe)(μ-pdt)(μ-D)Fe(CO)3 ([11]).

1 Figure 42. The H-NMR (600 MHz, CD2Cl2) of [11].

31 Figure 43. The P-NMR (600 MHz, CD2Cl2) [11].

Resonance Raman was performed on 9, 10, and 11 in DCM and d2-DCM. Below,

Figure 44 shows a comparison between the vibrational spectra of each of the compounds.

The peaks that are labeled indicate vibrations that differ between each spectra. The spectrum of 9 has vibrations at 1942 and 2013 cm-1 corresponding to –CO vibrations.

Compound 10 contains peaks at 939, 1002, and 1032 cm-1 that are not observed in compound 9. The spectrum of 10 has vibrations at 2015 and 2080 cm-1 corresponding to –

CO vibrations. Also in 10, there are broad peaks at 842 and 1256 cm-1. Another peak that corresponds to the M-H bond is buried under the oxygen stretch at 1570 cm-1. This peak can be seen in Figure 45, which shows the spectrum of 11 directly subtracted from 10, without correcting for solvent or quartz bands. It is likely that this band represents the Fe-

H stretch, while the lower frequency bands around 1000 cm-1 reflect the Ni-H stretching motions.

2013 1942 4 * *

2080 2015

-2 2081

-4 Normalized Raman Intensity Raman Normalized

-6

500 1000 1500 2000 -1 Raman Shift (cm )

Figure 44. Resonance Raman (514.9nm excitation, 77K) of black [9], blue [10], and red [11] in dichloromethane. Solvent peaks have been removed for clarity and are indicated with *. RR spectra are normalized to the peak intensity at 215 cm-1.

3 600x10

400

200 1571

-200 Normalized Raman Intensity Raman Normalized

1200 1400 1600 1800 2000 -1 Raman Shift (cm )

Figure 45. Resonance Raman (514.9nm excitation, 77K) of [11] in dichloromethane without the subtracted quartz spectrum.

While compound 11 was synthesized in order to clearly identify the M-H vibrations, the

RR data suggests that the hydride in this sample was not fully exchanged with the deuteride due to the lack of distinct differences in peaks between the spectra of 10 and 11.

3.4 Characterization of Ni(dppbz)(μ-pdt)(μ-H)Fe(CO)3

As before, compounds 12 and 13 were synthesized from stirring the reactants at room temperature with addition of the appropriate ligands to yield the desired products, as seen in Figure 46. Isolation of the product is performed by recrystallization. The products

31 were verified with P-NMR spectroscopy on 12 (600 MHz, CD2Cl2), where the signal

31 corresponding to Ni(dppbz)Cl2 is at 57.6 ppm (Figure 47). The P-NMR of 13 (600 MHz,

CD2Cl2), contains the single peak corresponding to Ni(dppbz)(pdt) at 61.5 ppm, (Figure

49).

15 min stir

Ethanol/ DCM

Figure 46. Synthetic scheme to generate Ni(dppbz)Cl2 ([12]).

31 Figure 47. P-NMR of [12] (600 MHz, CD2Cl2).

20 min stir

DCM

Figure 48. Synthetic scheme to generate Ni(dcpe)(pdt) ([13]).

31 Figure 49. P-NMR of [13] (600 MHz, CD2Cl2).

The reaction between 8 and 13 ultimately yields the bimetallic compound 14

(Figure 50). The structure of 14 was verified with NMR (Figure 51) and FT-IR, showing

-1 peaks at 59.0 ppm and 1940 and 2029 cm . The reaction of the FeI2(CO)4 complex with

13 and excess CoCp2, as stated previously, is required to obtain product. As before, this

+ reaction proceeds via a μ-iodo cation intermediate; [Ni(dppbz)(μ-pdt)(μ-I)Fe(CO)3] , which can be monitored using FT-IR spectroscopy. The –CO bands that are visible in the intermediate are approximately 20 cm-1 higher than the hydride complex –CO bands

(Figure 52). Purification of 14 requires running the solution through a plug of celite and precipitating out the desired product with hexanes. The FT-IR spectrum of 14 as well as the μ-I intermediate are shown below. The –CO stretches of 14 are are seen at 1940 and

2030 cm-1, (Figure 53). 49

Cannula transfer, Cp2Co

77K, DCM

Figure 50. Synthetic scheme to generate Ni(dppbz)(μ-pdt)Fe(CO)3 ([14]).

The 31P-NMR spectrum of 10 shows a single band at 59.0 ppm corresponding to the phosphine signal of Ni(dppbz)(μ-pdt)Fe(CO)3.

31 Figure 51. P-NMR of [14] (600 MHz, CD2Cl2).

% Transmittance % 2047 80 2023

75 2092 1900 2000 2100 2200 Wavenumber

Figure 52. The -CO stretches of Ni(dppbz)(μ-pdt)(μ-I)Fe(CO)3.

110

105

100

95 *

1940 *

% transmittance % 90 2029

1900 2000 2100 wavenumbers

Figure 53. FT-IR (solution-phase) of [14]. * denotes a peak that is due to partial protonation from exposure to moisture.

The desired compound 15 was then synthesized by the addition of excess

HBF4*Et2O (Figure 54) and stirred under an anaerobic environment and verified with

NMR (Figure 55) and FT-IR spectroscopy (Figure 56). The 31P-NMR spectrum of 15 shows a single band at 68.0 ppm (Figure 55), while the hydride signal is evident in the 1H-

NMR spectrum at -3.4 ppm (Figure 55 inset).

5 min stir

DCM

Figure 54. Synthetic scheme to generate Ni(dppbz)(μ-pdt)(μ-H)Fe(CO)3 ([15]).

31 1 Figure 55. P-NMR of [15] (600 MHz, CD2Cl2), and the inset H-NMR (600 MHz, CD2Cl2).

The –CO bands are seen to shift to higher wavenumbers from the unprotonated species, at

2022 and 2081 cm-1.36

2081

2022 % Transmittance %

1950 2000 2050 2100 2150 2200 2250 Wavenumber

Figure 56. The carbonyl stretches of Ni(dppbz)(μ-pdt)(μ-H)Fe(CO)3, ([15]). 53

UV-Vis absorption data was collected to determine an ideal wavelength for performing resonance Raman spectroscopy on these compounds (Figure 57). As before, a broad absorption band centered at 520 nm was observed, suggesting an excitation wavelength of 514.5 nm could be used to study 14-16 using RR spectroscopy.

Figure 57. UV-Vis of [14] black, [15] blue.

In order to better study the metal-hydride structure, isotope exchange experiments were performed. To replace the bridging hydride with deuterium, 15 was stirred with D2O

54 under an inert atmosphere for sixteen hours (Figure 58). The lack of signal in the 1H-NMR at -3.4 ppm indicates that the hydride was exchanged with deuterium (Figure 59). The phosphine ligand signal in Figure 60 can be seen at 68.8 ppm, which is unchanged from species 15 in the 31P-NMR, suggesting the compound remains intact.

16 hr stir

DCM

Figure 58. Synthetic scheme to generate Ni(dppbz)(μ-pdt)(μ-D)Fe(CO)3 ([16]).

1 Figure 59. H-NMR of 16 (600MHz, CD2Cl2). 55

31 Figure 60. P-NMR of 16 (600 MHz, CD2Cl2).

3 100x10

50 2032

* 0

-50 Normalized Raman Intensity Raman Normalized

500 1000 1500 2000 -1 Raman Shift (cm )

Figure 61. Resonance Raman (514.9nm excitation, 77K) of [14] in dichloromethane. The large solvent peaks have been removed for clarity and are indicated with *.

The RR spectrum in Figure 61 only contains data from compound 14 due to issues with the samples. Due to high fluorescence in multiple solvents, spectra of compounds 15 and 16 were unable to be collected. The solvents were restricted due to solubility issues, but DCM, DMSO, THF, d2-DCM, and benzene were tested. Multiple wavelengths were also tested with no improvement in fluorescence. Notable peaks in the RR spectrum of compound [14] are the –CO vibration at 2032 cm-1 and the range 200-600 cm-1 which is dominated by metal-metal and metal-ligand derived normal modes.

3.2 Spectroscopy

Compounds 3, 4, and 5 were investigated with RR spectroscopy in order to replicate previous work done on these compounds (Figure 25).38 High-intensity, low frequency bands between 200 and 600 cm-1 dominate the RR spectra of both 3 and 4. These bands correspond to metal-metal and metal-ligand-derived normal modes.38 The lack of symmetry in the molecules lead to approximately 50 predicted modes and 20 bands observed in the spectral region for both 4 and 5.38 Compound 4 also contains bands at 954,

1473, and 1535 cm-1, which are absent in compound 5, and have been attributed to metal- hydride vibrations. Three bands at 1001, 1028, and 1106 cm-1 are observed in both compounds 3 and 4. Also, bands at 1556 and 1589 cm-1 are present in both compounds 3 and 4. These bands correspond to ring stretches. There are slight shifts (a few wavenumbers) seen between 3 and 4 at 490, 587, and 617 cm-1, where compound 3 is slightly upshifted. The carbonyl stretches of each compound can be seen in both RR and

FTIR spectra (Figure 25). Compound 3 exhibits CO vibrations at 1962 and 2028 cm-1.

Carbonyl vibrations can be seen at 2023 and 2083 cm-1 for 4. The peak at 2076 cm-1 corresponds to the carbonyl vibration of 5. The same number of carbonyl vibrations should be seen in 4 and 5, but due to low concentrations of 5 a carbonyl band is not observable.

The carbonyl bands for 3 are downshifted from 4. This is due to the increased amount of backbonding occurring in the compound 3, which is also caused by increased electron availability on the metal.

Compounds 9, 10, and 11 were also investigated using RR, as seen in Figure 25.

The peaks at 1942 and 2013 cm-1 correspond to the carbonyl modes of 9. Compound 10 contains bands at 939, 1002, and 1032 cm-1 that are not observed in compound 9. Carbonyl modes can be seen at 2015 and 2080 cm-1 for compound 10. A vibration at 1570 cm-1, which corresponds to the iron-hydride is buried under the broad oxygen stretch.

Unfortunately, due to the large contribution of quartz in the spectra, this feature is masked during data analysis. Data showing the direct subtraction of 11 from 10 can be seen in

Figure 44, where the M-H peak can be observed at 1570 cm-1.

Compound 13 was investigated with RR, but, unfortunately, we were unable to collect RR spectra of compounds 14 and 15, despite attempting experiments with a wide range of solvents, including DCM, d2-DCM, DMSO, DMF, THF, and benzene. All of these samples showed high amounts of fluorescence that masked all Raman bands. This could be due to intrinsic fluorescence, incomplete glassing, or aggregation of the sample. The spectrum of compound 13 is shown in Figure 60. Notable peaks in the RR spectrum of this compound are the -CO stretching vibration at 2032 cm-1 and the bands in the 200-600 cm-

1 region which, as before, are dominated by metal-metal and metal-ligand vibrations. There should be more bands from this ligand than the others. There should have been two distinct

–CO vibrations even in the precursor compound 13, due to the three carbonyl ligands on the iron, but due to high fluorescence one peak is obscured by the background.

The M-H vibrations seen in 4 can be assigned to vibrational normal modes.38 The peak at 1533 cm-1 corresponds to the Fe-H vibration and the peak at 954 cm-1 corresponds to the Ni-H vibration. The corresponding vibrations can be assigned for compound 10 where the peak at 1571 cm-1 corresponds to the Fe-H vibration. These peaks are very broad peaks that are indicative of a metal-hydride vibration. The spectrum for 14 cannot be interpreted due to high fluorescence in the sample. The metal-hydride vibrations in 4 are downshifted from 10. This could be due to the metal-hydride distances lengthening in the crystal structure of 10 from the bond distances seen in 4, as seen in Figure 9. The Ni-H and Fe-H bond lengths are respectively 1.89 Å and 1.49 Å in the crystal structure of 4.26,36

The bond lengths of Ni-H and Fe-H are respectively, 1.91 Å and 1.53 Å in the crystal structure of 10.36 In x-ray crystallography the metal-hydride bond lengths are always underestimated.43 Due to the small amount of change in the bond length and inherent limitations in resolution for the crystal structure, it is possible that the vibrational data shows subtle differences that cannot be seen in the crystal structure. The observed Fe-H vibration in 10 indicates that the actual Fe-H bond distance is shorter than observed in the x-ray crystal structure. The bond distances of the Ni-R state of the protein, Ni-H (1.58 Å) and Fe-H (1.78 Å)16, also show that the asymmetry of the models is opposite that seen in

59 the protein. This divergence in structure is reflected in function, as the protein performs nickel-centered chemistry, while the models perform iron-centered chemistry.

Compounds 5 and 11 gave lower quality RR spectra due to the inability to fully exchange the hydride before sample decomposition and phosphine ligands dissociation.

One possible route to synthesize better samples may be attained by adding a higher concentration of sample and exchanging with a smaller amount of D2O for a longer period of time. The smaller amount of deuterium should allow for less sample decomposition, while the longer exchange time should ensure a full exchange with the lower ratio of deuterium.

The synthesis of these [NiFe] hydrogenase models will provide greater understanding of the powerful chemistry performed by the [NiFe] H2ases. The different phosphine ligands used give insight into how the ligands affect the metal-hydride core, which can be related to the Ni-R state in the [NiFe] hydrogenase.

Chapter 4: Conclusion

The [NiFe] H2ase mimics [Ni(dppe)(μ-pdt)(μ-H)Fe(CO)3][BF4], [Ni(dcpe)(μ- pdt)(μ-H)Fe(CO)3][BF4], and [Ni(dppbz)(μ-pdt)(μ-H)Fe(CO)3][BF4] were synthesized and studied using resonance Raman spectroscopy. [Ni(dppe)(μ-pdt)(μ-H)Fe(CO)3][BF4] had been previously investigated, and the insight from those studies on the metal-hydride core can be applied to studies of [Ni(dcpe)(μ-pdt)(μ-H)Fe(CO)3][BF4] and [Ni(dppbz)(μ-pdt)(μ-

H)Fe(CO)3][BF4]. For example, it was found that compound 4 shows a Ni-H vibration at

954 cm-1 and a Fe-H vibration at 1533 cm-1.

The series of the Ni(dcpe)(μ-pdt) Fe(CO)3 compounds with and without a bridging hydride were also investigated. The metal-hydride vibrations could be clearly observed because of isotopic substitution experiments. In this series, the Ni-H vibration was observed at 844 cm-1 and the Fe-H vibration at 1571 cm-1. The spectrum of Ni(dppbz)(μ- pdt) Fe(CO)3 spectra was collected, as verified by observation of correct frequencies for the carbonyl bands, but, due to fluorescence, the spectra of [Ni(dppbz)(μ-pdt)(μ-

H)Fe(CO)3][BF4]and Ni(dppbz)(μ-pdt)(μ-D)Fe(CO)3 could not be collected. Overall, this study has provided insight into how the different ligand variants affect the metal-hydride core. The iron-hydride vibrations seen in the RR of 4 are lower raman shift than 10. This could be due to the iron-hydride distances of compound 10 actually being shorter than the bond lengths seen in the crystal structure, as seen in Figure 9. This subtle difference 61 provides insight into the metal-hydride structure that is not fully detailed in the crystal structure and can give insight in to how to better model the Ni-R state in the protein.

Compounds 4 and 10 are used as a model for the Ni-R state in [NiFe] H2ases, and thus enable predictions about analogous vibrations in the active site of the enzyme.

References

(1) (2013) DOE International Energy Agency. DOE International Energy Energy Outlook 2013; 2013.

(2) Tard, C.; Pickett, C. J. Structural and Functional Analogues of the Active Sites of the [Fe]-, [NiFe]-, and [FeFe]-Hydrogenases. Chem. Rev. 2009, 109 (6), 2245– 2274.

(3) Hydrogen Basics. U.S. Department of Energy May 10, 2016.

(4) CRC Handbook of Chemistry and Physics, 87th Ed Editor-in-Chief: David R. Lide (National Institute of Standards and Technology). CRC Press/Taylor and Francis Group: Boca Raton, FL. 2006. 2608 Pp. 139.95. ISBN 0-8493-0487-3. J. Am. Chem. Soc. 2007, 129 (3), 724–724.

(5) Dau, H.; Limberg, C.; Reier, T.; Risch, M.; Roggan, S.; Strasser, P. The Mechanism of Water Oxidation: From Electrolysis via Homogeneous to Biological Catalysis. ChemCatChem 2010, 2 (7), 724–761.

(6) Breakthrough in Hydrogen Fuel Production could Revolutionize Alternative Energy Market. http://vtnews.vt.edu/content/vtnews_vt_edu/en/articles/2013/04/040413-cals- hydrogen.html (accessed Oct 8, 2016).

(7) The Impact of Increased Use of Hydrogen on Petroleum Consumption and Carbon Dioxide Emissions. DOE International Energy September 2008.

(8) Erisman, J. W.; Sutton, M. A.; Galloway, J.; Klimont, Z.; Winiwarter, W. How a Century of Ammonia Synthesis Changed the World. Nat. Geosci. 2008, 1 (10), 636–639. 63

(9) Fontecilla-Camps, J. C.; Volbeda, A.; Cavazza, C.; Nicolet, Y. Structure/Function Relationships of [NiFe]- and [FeFe]-Hydrogenases. Chem. Rev. 2007, 107 (10), 4273–4303.

(10) McIntosh, C. L.; Germer, F.; Schulz, R.; Appel, J.; Jones, A. K. The [NiFe]- Hydrogenase of the Cyanobacterium Synechocystis Sp. PCC 6803 Works Bidirectionally with a Bias to H2 Production. J. Am. Chem. Soc. 2011, 133 (29), 11308–11319.

(11) Horch, M.; Lauterbach, L.; Lenz, O.; Hildebrandt, P.; Zebger, I. NAD(H)-Coupled Hydrogen Cycling – Structure–function Relationships of Bidirectional [NiFe] Hydrogenases. FEBS Lett. 2012, 586 (5), 545–556.

(12) Shafaat, H. S.; Rüdiger, O.; Ogata, H.; Lubitz, W. [NiFe] Hydrogenases: A Common Active Site for Hydrogen Metabolism under Diverse Conditions. Biochim. Biophys. Acta BBA - Bioenerg. 2013, 1827 (8–9), 986–1002.

(13) Ogata, H.; Lubitz, W.; Higuchi, Y. [NiFe] Hydrogenases: Structural and Spectroscopic Studies of the Reaction Mechanism. Dalton Trans. 2009, 37, 7577- 7587.

(14) Volbeda, A.; Martin, L.; Cavazza, C.; Matho, M.; Faber, B. W.; Roseboom, W.; Albracht, S. P. J.; Garcin, E.; Rousset, M.; Fontecilla-Camps, J. C. Structural Differences between the Ready and Unready Oxidized States of [NiFe] Hydrogenases. JBIC Publ. Soc. Biol. Inorg. Chem. 2005, 10 (3), 239–249.

(15) Foerster, S.; Stein, M.; Brecht, M.; Ogata, H.; Higuchi, Y.; Lubitz, W. Single Crystal EPR Studies of the Reduced Active Site of [NiFe] Hydrogenase from Desulfovibrio Vulgaris Miyazaki F. J. Am. Chem. Soc. 2003, 125 (1), 83–93.

(16) Ogata, H.; Nishikawa, K.; Lubitz, W. Hydrogens Detected by Subatomic Resolution Protein Crystallography in a [NiFe] Hydrogenase. Nature 2015, 520 (7548), 571–574.

(17) Lubitz, W.; Ogata, H.; Rüdiger, O.; Reijerse, E. Hydrogenases. Chem. Rev. 2014, 114 (8), 4081–4148. 64

(18) Ogata, H.; Mizoguchi, Y.; Mizuno, N.; Miki, K.; Adachi, S.; Yasuoka, N.; Yagi, T.; Yamauchi, O.; Hirota, S.; Higuchi, Y. Structural Studies of the Carbon Monoxide Complex of [NiFe]hydrogenase from Desulfovibrio Vulgaris Miyazaki F: Suggestion for the Initial Activation Site for Dihydrogen. J. Am. Chem. Soc. 2002, 124 (39), 11628–11635.

(19) Marques, M. C.; Coelho, R.; Pereira, I. A. C.; Matias, P. M. Redox State- Dependent Changes in the Crystal Structure of [NiFeSe] Hydrogenase from Desulfovibrio Vulgaris Hildenborough. Int. J. Hydrog. Energy 2013, 38 (21), 8664–8682.

(20) Rauchfuss, T. B.; Contakes, S. M.; Hsu, S. C. N.; Reynolds, M. A.; Wilson, S. R. The Influence of Cyanide on the Carbonylation of Iron(II): Synthesis of Fe−SR−CN−CO Centers Related to the Hydrogenase Active Sites. J. Am. Chem. Soc. 2001, 123 (28), 6933–6934.

(21) Lai, C.-H.; Reibenspies, J. H.; Darensbourg, M. Y. Thiolate Bridged Nickel–Iron Complexes Containing Both Iron(0) and Iron(II) Carbonyls. Angew. Chem. Int. Ed. Engl. 1996, 35 (20), 2390–2393.

(22) Brazzolotto, D.; Gennari, M.; Queyriaux, N.; Simmons, T. R.; Pécaut, J.; Demeshko, S.; Meyer, F.; Orio, M.; Artero, V.; Duboc, C. Nickel-Centred Proton Reduction Catalysis in a Model of [NiFe] Hydrogenase. Nat. Chem. 2016, 8 (11), 1054-1060.

(23) Sellmann, D.; Geipel, F.; Heinemann, F. W. (NEt4)2[Fe(CN)2(CO)(′S3′)]: An Iron Thiolate Complex Modeling the [Fe(CN)2(CO)(S−Cys)2] Site of [NiFe] Hydrogenase Centers. Chem. – Eur. J. 2002, 8 (4), 958–966.

(24) Wombwell, C.; Reisner, E. Synthetic Active Site Model of the [NiFeSe] Hydrogenase. Chem. – Eur. J. 2015, 21 (22), 8096–8104.

(25) Barton, B. E. Hydrogen Production from Model Complexes of the [FeFe]-and [NiFe]-Hydrogenase Active Sites, University of Illinois at Urbana-Champaign, 2011. 65

(26) Barton, B. E.; Rauchfuss, T. B. Hydride-Containing Models for the Active Site of the Nickel−Iron Hydrogenases. J. Am. Chem. Soc. 2010, 132 (42), 14877–14885.

(27) Darensbourg, D. J.; Reibenspies, J. H.; Lai, C.-H.; Lee, W.-Z.; Darensbourg, M. Y. Analysis of an Organometallic Iron Site Model for the Heterodimetallic Unit of [NiFe]Hydrogenase. J. Am. Chem. Soc. 1997, 119 (33), 7903–7904.

(28) Manor, B. Incorporation of Cyanide Ligands into Models of [FeFe]-and [NiFe]- Hydrogenase, University of Illinois at Urbana-Champaign, 2015.

(29) Osterloh, F.; Saak, W.; Haase, D.; Pohl, S. Synthesis, X-Ray Structure and Electrochemical Characterization Of A Binuclear Thiolate Bridged Ni–Fe–nitrosyl Complex, Related to the Active Site of [NiFe] Hydrogenase. Chem Commun 1997, 10, 979–980.

(30) Volbeda, A.; Charon, M.-H.; Piras, C.; Hatchikian, E. C.; Frey, M.; Fontecilla- Camps, J. C. Crystal Structure of the Nickel–iron Hydrogenase from Desulfovibrio Gigas. Nature 1995, 373 (6515), 580–587.

(31) Zhu, W.; Marr, A. C.; Wang, Q.; Neese, F.; Spencer, D. J. E.; Blake, A. J.; Cooke, P. A.; Wilson, C.; Schröder, M. Modulation of the Electronic Structure and the Ni- Fe Distance in Heterobimetallic Models for the Active Site in [NiFe]hydrogenase. Proc. Natl. Acad. Sci. U. S. A. 2005, 102 (51), 18280–18285.

(32) Kaur-Ghumaan, S.; Stein, M. [NiFe] Hydrogenases: How Close Do Structural and Functional Mimics Approach the Active Site? Dalton Trans. 2014, 43 (25), 9392- 9405.

(33) Barton, B. E.; Whaley, C. M.; Rauchfuss, T. B.; Gray, D. L. Nickel−Iron Dithiolato Hydrides Relevant to the [NiFe]-Hydrogenase Active Site. J. Am. Chem. Soc. 2009, 131 (20), 6942–6943.

(34) Brazzolotto, D.; Gennari, M.; Queyriaux, N.; Simmons, T. R.; Pécaut, J.; Demeshko, S.; Meyer, F.; Orio, M.; Artero, V.; Duboc, C. Nickel-Centred Proton

Reduction Catalysis in a Model of [NiFe] Hydrogenase. Nat. Chem. 2016, 1054- 1060.

(35) Hu, X.; Brunschwig, B. S.; Peters, J. C. Electrocatalytic Hydrogen Evolution at Low Overpotentials by Cobalt Macrocyclic Glyoxime and Tetraimine Complexes. J. Am. Chem. Soc. 2007, 129 (29), 8988–8998.

(36) Carroll, M. E.; Barton, B. E.; Gray, D. L.; Mack, A. E.; Rauchfuss, T. B. Active- Site Models for the Nickel-Iron Hydrogenases: Effects of Ligands on Reactivity and Catalytic Properties. Inorg. Chem. 2011, 50 (19), 9554–9563.

(37) Mack, A. E. Modeling The Active Site Of [NiFe]-Hydrogenase, University of Illinois at Urbana-Champaign, 2010.

(38) Shafaat, H. S.; Weber, K.; Petrenko, T.; Neese, F.; Lubitz, W. Key Hydride Vibrational Modes in [NiFe] Hydrogenase Model Compounds Studied by Resonance Raman Spectroscopy and Density Functional Calculations. Inorg. Chem. 2012, 51 (21), 11787–11797.

(39) McCreery, R. L. Raman Spectroscopy for Chemical Analysis; Chemical analysis; John Wiley & Sons: New York, 2000, 157.

(40) Booth, G.; Chatt, J. Some Complexes of Ditertiary Phosphines with Nickel(II) and Nickel(III). J. Chem. Soc. 1965, No. 0, 3238–3241.

(41) Hieber, W.; Bader, G. Reaktionen Und Derivate Des Eisencarbonyls, II.: Neuartige Kohlenoxyd-Verbindungen von Eisenhalogeniden. Berichte Dtsch. Chem. Ges. B Ser. 1928, 61 (8), 1717–1722.

(42) Buckingham, A. D.; Stephens, P. J. 528. Proton Chemical Shifts in the Nuclear Magnetic Resonance Spectra of Transition-Metal Hydrides: Octahedral Complexes. J. Chem. Soc. 1964, No. 0, 2747–2759.

(43) Ricci, J. S.; Koetzle, T. F.; Bautista, M. T.; Hofstede, T. M.; Morris, R. H.; Sawyer, J. F. Single-Crystal X-Ray and Neutron Diffraction Studies of an η2- 67

Dihydrogen Transition-Metal Complex: Trans-[Fe(η2- H2)(H)(PPh2CH2CH2PPh2)2]BPh4. J. Am. Chem. Soc. 1989, 111 (24), 8823–8827.