Synthesis and interconversions of reduced, alkali-metal supported iron-sulfur-carbonyl complexes

Article abstract of DOI:10.1039/c7dt01506b

We report the synthesis, interconversions and X-ray structures of a set of [mFe-nS]-type carbonyl clusters (where S = S^2-, S₂^2- or RS^-; m = 2-3; n = 1-2). All of the clusters have been identified and characterized by single crystal X-ray diffraction, IR and ¹³C NMR. Reduction of the parent neutral dimer [μ₂-(SPh)₂Fe₂(CO)₆] (1) with KC₈ affords an easily separable ～1-:-1 mixture of the anionic, dimeric thiolate dimer K[Fe₂(SPh)(CO)₆(μ-CO)] (2) and the dianionic, sulfido trimer [K(benzo-15-crown-5)₂]₂[Fe₃(μ₃-S)(CO)₉] (3). Oxidation of 2 with diphenyl-disulfide (Ph₂S₂) cleanly returns the starting material 1. The Ph-S bond in 1 can be cleaved to form sulfide trimer 3. Oxidation of sulfido trimer 3 with [Fc](PF₆) in the presence of S₈ cleanly affords the all-inorganic persulfide dimer [μ₂-(S)₂Fe₂(CO)₆] (4), a thermodynamically stable product. The inverse reactions to form 3 (dianion) from 4 (neutral) were not successful, and other products were obtained. For example, reduction of 4 with KC₈ afforded the mixed valence Fe(i)/Fe(ii) species [((FeI2S₂)(CO)₆)₂Fe^II]^2- (5), in which the two {Fe₂S₂(CO)₆}^2- units serve as bidendate ligands to a Fe(ii) center. Another isolated product (THF insoluble portion) was recrystallized in MeCN to afford [K(benzo-15-crown-5)₂]₂[((Fe₂S)(CO)₆)₂(μ-S)₂] (6), in which a persulfide dianion bridges two {2Fe-S} moieties (dimer of dimers). Finally, to close the interconversion loop, we converted the persulfide dimer 4 into the thiolate dimer 1 by reduction with KC₈ followed by reaction with the diphenyl iodonium salt [Ph₂I](PF₆), in modest yield. These reactions underscore the thermodynamic stability of the dimers 1 and 4, as well as the synthetic and crystallization versatility of using the crown/K⁺ counterion system for obtaining structural information on highly reduced iron-sulfur-carbonyl clusters.

Full text of DOI:10.1039/c7dt01506b

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
Synthesis and interconversions of reduced,
alkali–metal supported iron–sulfur–carbonyl
complexes†
Cite this: Dalton Trans., 2017, 46,
9163
J. Patrick Shupp,
Amber R. Rose and Michael J. Rose
*
We report the synthesis, interconversions and X-ray structures of a set of [mFe–nS]-type carbonyl clusters
(where S = S2−, S2 or RS−; m = 2–3; n = 1–2). All of the clusters have been identified and characterized
2−
by single crystal X-ray diffraction, IR and 13C NMR. Reduction of the parent neutral dimer [μ2-(SPh)2Fe2(CO)6]
(1) with KC8 affords an easily separable ∼1 : 1 mixture of the anionic, dimeric thiolate dimer K[Fe2(SPh)
(CO)6(μ-CO)] (2) and the dianionic, sulfido trimer [K(benzo-15-crown-5)2]2[Fe3(μ3-S)(CO)9] (3). Oxidation
of 2 with diphenyl-disulfide (Ph2S2) cleanly returns the starting material 1. The Ph–S bond in 1 can be
cleaved to form sulfide trimer 3. Oxidation of sulfido trimer 3 with [Fc](PF6) in the presence of S8 cleanly
affords the all-inorganic persulfide dimer [μ2-(S)2Fe2(CO)6] (4), a thermodynamically stable product. The
inverse reactions to form 3 (dianion) from 4 (neutral) were not successful, and other products were
obtained. For example, reduction of 4 with KC8 afforded the mixed valence Fe(I)/Fe(II) species [((FeI2S2)-
(CO)6)2FeII]2− (5), in which the two {Fe2S2(CO)6}2− units serve as bidendate ligands to a Fe(II) center.
Another isolated product (THF insoluble portion) was recrystallized in MeCN to afford [K(benzo-15-
crown-5)2]2[((Fe2S)(CO)6)2(μ-S)2] (6), in which a persulfide dianion bridges two {2Fe–S} moieties (dimer of
dimers). Finally, to close the interconversion loop, we converted the persulfide dimer 4 into the thiolate
dimer 1 by reduction with KC8 followed by reaction with the diphenyl iodonium salt [Ph2I](PF6), in modest
yield. These reactions underscore the thermodynamic stability of the dimers 1 and 4, as well as the syn-
thetic and crystallization versatility of using the crown/K+ counterion system for obtaining structural infor-
mation on highly reduced iron–sulfur–carbonyl clusters.
Received 26th April 2017,
Accepted 22nd June 2017
DOI: 10.1039/c7dt01506b
rsc.li/dalton
reduction of dinitrogen.17 In nitrogenase, the biogenesis of the
catalytically active M-cluster involves the transformation of
Introduction
Iron sulfur clusters are prevalent in many metalloproteins, par- cysteinyl sulfur to inorganic sulfide in the form of a [2Fe–2S]
ticularly those that perform the activation of small molecules cluster, and then later to [4Fe–4S] clusters. Recent work by
such as H2, CO2 and N2.1 Among these, the [FeFe]- Wiig et al. demonstrated that NifEN catalyzes the transfer of
hydrogenase2–5 and nitrogenase6–9 have received much atten- CH3+ derived from S-adenosyl methionine to a [4Fe–4S] sulfide
tion due to the investigation of synthetic analogues in explor- by a nucleophilic attack.18,19 Radical abstraction of protons
ing fundamental processes of electron transfer, cluster reactiv- by the adenosyl radical likely begins the process of forming
ity and substrate activation – in addition to possible uses for multiple Fe–C bonds.
industrial catalysis. While applications of [2Fe–2S]-type clus-
In one case closely related to the present work, models of
ters generally pertain to proton reduction,10–13 they have also [NiFe]-hydrogenase were prepared by installing a nucleophilic
been shown to perform additional reactions such as liberation sulfur on an [Fe2(SR)] cluster via a reaction with carbon di-
of H2 from amino-boranes14 and the reduction of molecular sulfide (CS2): specifically, reaction of (NEt3H)[(μ-SPh)(μ-CO)
oxygen.15,16 In related biomimetic work, [2Mo–6Fe–8S] clusters Fe2(CO)6] with CS2 afforded (NEt3H)[(μ-SPh)Fe2(S
̲
̲–S)
[2Sn–6S], generating
a
system that photo-catalyzes the sources provided the heterobimetallic model complexes.
Another report explored the facile modularity of the μ3-S ↔ μ2-S
sulfido motif: Tatsumi and coworkers reported the intercon-
Department of Chemistry, The University of Texas at Austin, USA.
E-mail: mrose@cm.utexas.edu
version of [4Fe–4S] ↔ [2Fe–2S] clusters supported by N(Si
(CH3)3) ligands, promoted by the absence or presence (respect-
ively) of pyridine.22 This same report delineated the redox con-
version of the [(py)2Fe2S2]0 cluster to either the [4Fe–4S]− or
†Electronic supplementary information (ESI) available. CCDC 1544150–1544153.
For ESI and crystallographic data in CIF or other electronic format see DOI:
10.1039/c7dt01506b
This journal is © The Royal Society of Chemistry 2017
Dalton Trans., 2017, 46, 9163–9171 | 9163

View Article Online
Paper
Dalton Transactions
[4Fe–4S]2− cluster upon reduction with 0.5 or 1 equiv. of
Workup part A for K[Fe2(SPh)(CO)6(μ-CO)] (2). The mixture
Na[C10H8], respectively. Notably, the clusters were isolated and was separated into the two desired components by extracting 2
crystallized as stable alkali metal adducts. as a dark brown solution into Et2O (10 × 2 mL extractions,
Generating precursors for these systems can be an arduous until the extractions are colorless); a red powder of complex 3
task, as the chemistry is highly thermodynamic in nature. remained insoluble. The combined Et2O extractions were
Clusters such as [Fe3(μ3-S)(CO)9]2− and [Fe2(μ-SR)(μ-CO)(CO)6]− allowed to stand overnight, during which time complex 2
are synthesized by “spontaneous self-assembly”23 of starting remained soluble; any precipitate (3) was removed by filtration
materials such as [Fe(CO)5] and S8, with multiple clusters and combined with the original Et2O insoluble portion. The
forming in a single reaction. Often times the desired cluster is brown Et2O soluble portion was evaporated in vacuo and
not a major product, leading to further complications in puri- washed thoroughly with pentane to afford a brownish golden
fication. To remedy this, in this work we have isolated and powder. The product was isolated as golden brown X-ray
characterized two new alkali-supported iron sulfur clusters quality crystals grown by slow diffusion of pentane into an
K[Fe2(μ-SR)(μ-CO)(CO)6] (2) and K2[Fe3(μ3-S)(CO)9] (3), which Et2O solution of the complex at −20 °C; the product was identi-
can provide a much more robust starting material for synthetic fied as the thiolate-bridged dimer K[Fe2(SPh)(CO)6(μ-CO)] (2).
modeling. These clusters also provide a unique advantage in Yield: 208 mg (50%). IR (νCO in cm−1): 2027, 1970, 1901, 1726.
the characterization of products by X-ray diffraction due to the Elem. analysis (%): calcd C 34.34, H 1.11; found C 34.25, H
utility of the alkali cation(s) and associated crown(s) ether in 1.30. 13C NMR in d3-MeCN (δ from TMS): 118.3, 127.0, 128.7,
providing convenient separation and multiple avenues of crys- 133.1, 143.7, 224.9.
tallization. Additionally, the alkali counterions are more resili-
Workup part B for [K(benzo-15-crown-5)2]2[Fe3(μ3-S)(CO)9] (3).
ent than their organic counterparts under reductive con- The remaining Et2O insoluble portion of K2[Fe3(μ3-S)(CO)9]
ditions, enabling broader exploration of reaction pathways. In (red material) was dissolved in MeCN containing excess benzo-
this work, we have explored the isolation of reduced, anionic 15-crown-5 ether, and the red crystalline product was obtained
iron–sulfur–carbonyl clusters and their reactions with carbon- via slow vapor diffusion of Et2O. The product was identified as
based electrophiles (CH2I2, CF2Br2), as well as sulfur based [K(benzo-15-crown-5)2]2[Fe3(μ3-S)(CO)9] (3). Yield: 120 mg
modifications (alkylation, de-alkylation, S–S bond making/ (43%). IR (νCO in cm−1): 1995, 1914, 1892, 1874. Elem. analysis:
breaking). Discrete interconversions among species are investi- calcd C 48.70, H 5.03; found C 49.47, H 5.35. 13C NMR in d3-
gated to understand the reactivity properties of the iron versus MeCN (δ from TMS): 15.64, 66.26, 67.71, 68.67, 69.44, 148.5,
sulfur centers in the clusters.
223.5.
Reaction of K[Fe2(SPh)(CO)6(μ-CO)] (2) with CH2I2 and
CF2Br2 to (1). In a typical reaction, 100 mg (0.202 mmol) of 2
was dissolved in 20 mL of Et2O and cooled to −20 °C and 2.0
equiv. CH2I2 or CF2Br2 were then dissolved in Et2O and added
to the chilled solution. As the reaction warmed to room temp-
erature a white precipitate formed (KI or KBr, respectively).
Experimental
Unless otherwise stated, all reactions were performed void of
oxygen with the aid of an argon atmosphere glovebox or N2
atmosphere Schlenk line. The starting materials Fe(CO)5,
PhSH, S8, Ph2IPF6, [Fc]PF6 and benzo-15-crown-5 were pur-
chased from Strem, Aldrich, Acros, or Oakwood Chemicals and
used without further purification. KC8 was prepared from a
reported synthesis.24 Deuterated MeCN was purchased from
Cambridge Isotopes and was used as received. HPLC grade
solvents were purchased from EMD, Fisher, Macron, or
J. T. Baker, and dried with the use of an alumina column
system (Pure Process Technology). Caution: Reduced iron-
carbonyl complexes can be pyrophoric and should only be
handled under inert atmosphere.
After
4 h, the solution was filtered, and the solvent
was removed in vacuo to afford a bright red oil. The oil was
dissolved in pentane, and X-ray quality crystals from both
reactions were obtained by prolonged storage of the pentane
solution of the products at −80 °C. Both CH2I2 and CF2Br2
reactions afforded crystalline samples of [μ2-(SPh)2Fe2(CO)6]
(1). Yield: 25 mg (25%).
Synthesis of [K(benzo-15-crown-5)2]2[((Fe2S2)(CO)6)2Fe] (5)
and [K(benzo-15-crown-5)2]2[((Fe2S)(CO)6)2(μ-S)2] (6) from (4).
These two clusters were isolated as the primary products in the
reduction of [μ2-(S)2Fe2(CO)6] (4). The starting material for this
reaction was synthesized from the reported procedure.25 A vial
was charged with 202 mg [μ2-(S)2Fe2(CO)6] (0.406 mmol) and
Synthesis of the iron complexes
K[Fe2(SPh)(CO)6(μ-CO)] (2) and [K(benzo-15-crown-5)2]2 dissolved in 15 mL of THF. Next, two equiv. of KC8 (109 mg,
[Fe3(μ3-S)(CO)9] (3). These two reduced clusters were isolated 811 mmol) were added to the translucent red solution at room
as separable products resulting from the same reaction. First, temperature. The slurry was allowed to stir at room temperature,
262 mg (0.526 mmol) of [μ2-(SPh)2Fe2(CO)6] (1) was dissolved and the reaction quickly changed to a dark opaque green color.
in 15 mL of THF and brought to −20 °C. Next, a portion of KC8 After 90 min, the mixture was filtered to remove the graphite
(207 mg, 1.53 mmol) was added, and the mixture was allowed byproducts. The solvent was then removed in vacuo to afford a
to warm to room temperature. After 2 h the reaction was brown-green oil. The crude reaction product was washed
filtered, and the solvent was removed in vacuo, resulting in a thoroughly with Et2O (20 × 2 mL) until the resulting washes
black oil.
were virtually colorless and a forest green powder remained.
9164 | Dalton Trans., 2017, 46, 9163–9171
This journal is © The Royal Society of Chemistry 2017

View Article Online
Dalton Transactions
Paper
Workup part A for [K(benzo-15-crown-5)2]2[((Fe2S2)(CO)6)2Fe] IR (νCO in cm−1): 1995, 1914, 1892, 1874. Elem. analysis calcd:
(5). To obtain K2[((Fe2S2)(CO)6)2Fe] (5), a small amount of C 48.70, H 5.03; found: C 49.47, H 5.35.
THF (5–10 mL) was added to the green powder; the resulting
Conversion of [Fe3(μ3-S)(CO)9] (3) to [μ2-(SPh)2Fe2(CO)6] (1).
solution was filtered and treated with an excess of benzo-15- A vial was charged with 197 mg (0.372 mmol) of the bare
crown-5 for crystallization via vapor diffusion of Et2O. potassium starting material [Fe3(μ3-S)(CO)9], and dissolved in
Successive recrystallization of the polycrystalline solid ∼15 mL THF. Once the solution was homogeneous, 125 mg
with MeCN and Et2O produced X-ray quality crystals. Yield: S2Ph2 (1.5 equiv., 0.562 mmol) was added, followed by 249 mg
53 mg (12%). IR (νCO in cm−1): 2018, 1978, 1917. Elem. ana- (0.752 mmol) [Fc]PF6 at room temperature. The reaction was
lysis calcd: C 43.1, H 4.26; found: C 43.46, H 4.65. 13C NMR in allowed to proceed for 2 h before solvent was removed by
d3-MeCN (δ from TMS): 26.17, 68.20, 68.67, 69.24, 69.97 vacuum. The desired product was extracted into pentane, and
148.79, 214.60.
Workup part B for [K(benzo-15-crown-5)2]2[((Fe2S)(CO)6)2(μ-S)2] hexanes as eluent. The first purple band was identified as
(6). To obtain 6, the remaining powder (THF insoluble) can be [Fe3S2(CO)9], common byproduct in the synthesis of
was purified by column chromatography (silica gel) with
a
dissolved in 10 mL of MeCN. If MeCN is added to the mixture [μ2-(S)2Fe2(CO)6] (4); the subsequent yellow band that followed
before extraction into THF, the aforementioned product is was not collected (ferrocene byproduct of the oxidant).
somewhat soluble, but will not precipitate out in subsequent The last red band was collected and identified as
recrystallizations. To the MeCN extract, excess benzo-15-crown- [μ2-(SPh)2Fe2(CO)6] by TLC and IR. Yield: 20 mg (7%).
5 was added, and the solution was filtered. Vapor diffusion of
Conversion of [Fe3(μ3-S)(CO)9] (3) to [μ2-(S)2Fe2(CO)6] (4).
Et2O afforded the desired product 6, which crystallizes in blue- This reaction proceeded in analogous fashion to the previous
green rods and needles. Yield: 106 mg (20%). IR (νCO in cm−1): conversion to [μ2-(SPh)2Fe2(CO)6] (1). A vial was charged with
2023, 1992, 1945, 1911. Elem. analysis calcd: C 44.41, H 4.38; 152 mg K2[Fe3(μ3-S)(CO)9] (0.287 mmol) and dissolved in
found: C 44.77, H 4.71. 13C NMR in d3-MeCN (δ from TMS): ∼15 mL THF. Once the solution was homogeneous, 21.1 mg S8
67.95, 68.36, 68.86, 69.62, 148.4, 223.2.
(0.658 mmol) was added to the reaction, and a slight darken-
Conversion of [μ2-(S)2Fe2(CO)6] (4) to [μ2-(SPh)2Fe2(CO)6] (1). ing of the solution to red-brown was observed. After
A vial was charged with 149.1 mg (0.436 mmol) of complex 4, 15 minutes, 190 mg [Fc]PF6 (0.574 mmol) was added to the
which was dissolved in 15 mL of THF. Next, 116 mg of KC8 (2 solution, and the reaction was allowed to proceed for 2 h. After
equiv.) was added to the reaction, and the solution turned removing all solvent by vacuum, the product was extracted into
green, indicating the reduction of the starting material. After hexanes, and the nonpolar products were purified by column
1 h reaction time, excess graphite was filtered from the chromatography (silica gel) using hexanes as eluent. Again, a
reaction, and 368.2 mg (0.873 mmol, 2 equiv.) of [Ph2I]PF6 was purple band was present in the crude mixture, and was identi-
added to the solution. An immediate color change to red was fied as [Fe3S2(CO)9]; the red band prior to the purple band was
observed, and after
1
h
the solvent was removed collected and identified by IR and TLC to be the desired
under vacuum. Nonpolar products were extracted into pentane product [μ2-(S)2Fe2(CO)6]. Yield: 15 mg (11%).
(5 × 2 mL) and purified by column chromatography through
silica gel with hexanes. The product was confirmed to be
complex 1 by IR. Yield: 8.2 mg (4%).
Physical methods
The 13C NMR spectra were obtained on an Agilent MR
400 MHz spectrometer, and chemical shifts are reported in
ppm referenced to TMS. Infrared spectra were recorded by
using a Bruker Alpha spectrometer with the use of a diamond
ATR crystal. Elemental analyses were performed by Midwest
Microlab. Details regarding X-ray data collection and analysis
are located in the ESI.†
Synthesis of [K(benzo-15-crown-5)2]2[Fe3S(CO)9] (3) from
K2SO3. The following is a modification of the procedure to
isolate the NEt4 salt of [Fe3(μ3-S)(CO)9] first reported by Cherng
et al.26 A Schlenk flask was filled with 40 mL MeOH and
purged with N2 for 30 min. Next, 1.59 g of KOH and 0.41 g of
K2SO3 were added under N2. Finally, 1 mL of Fe(CO)5 (1.45 g,
7.42 mmol) was added and the solution slowly turned red. The
reaction heated to reflux under N2 for 4 days. Upon cooling,
the insoluble byproducts were removed by filtration using a
fritted, airfree filter tube and the MeOH was removed in vacuo.
The crude product was dissolved in THF and filtered to remove
Results and discussion
Synthesis and interconversions
another crop of insoluble material. Following removal of THF Generally, iron–sulfur clusters generated by spontaneous self-
solvent by vacuum, the crude product was washed thoroughly assembly are not isolated as ‘bare’ alkali-supported salts due
with Et2O (20 × 2 mL) until the solvent became virtually color- to their tendency to form viscous oils. Instead, they are precipi-
less. (Note: The resulting powder can be used as a starting tated out of solution via the introduction of organic salts such
material as the bare potassium salt.) To obtain crystals for the as NEt4Cl or PPh4Br.20,23,27–31 While this is sufficient for iso-
purpose of analysis, the powder was dissolved in MeCN and an lation of the complexes, the utility of these formulations are
excess amount (about 4.2 equiv.) of benzo-15-crown-5 ether limited when harsh reductions are required to generate or
was added. Vapor diffusion of Et2O afforded dark red plates of isolate further reduced species. In this same vein, many of the
[K(benzo-15-crown-5)2]2[Fe3(μ3-S)(CO)9] (3). Yield: 206 mg (5%). anionic iron–sulfur complexes are quite reducing, causing
This journal is © The Royal Society of Chemistry 2017
Dalton Trans., 2017, 46, 9163–9171 | 9165

View Article Online
Paper
Dalton Transactions
instability on an intramolecular basis as well. Our original goal
was to synthesize a reduced [2Fe–2S] cluster that was compati-
ble with both strong reductants and electrophiles. The known
diiron dithiolate dimer [μ2-(SPh)2Fe2(CO)6] (1) was chosen as a
starting material, as the two singly valent iron centers were
deemed to be susceptible to reduction. Initial attempts to
reduce the cluster with Na/Hg amalgam afforded an intract-
able black oil containing multiple products that were insepar-
able by differential solvent extraction or crystal growth. The
reductant KC8 proved more effective, as the potassium counter-
ion allowed for the separation of the two major products by
their respective solubility in Et2O. Another advantage of KC8 lies
in its modularity, considering the wide variety of crowns and
clathrates available for K+ sequestration, most of which retain
high solubility in organic solvents. Interestingly, we found the
‘bare’ mono-potassium salt K[Fe2(μ-RS)(μ-CO)(CO)6] (2) was
soluble in Et2O, providing a facile means of obtaining pure
Scheme 1 Synthetic scheme detailing syntheses and interconversions
arising from compound 1 as reported in this work.
samples of the cluster for crystallization. However, after extrac- initial reactants. Our first attempt involved the charge-
tion into Et2O as the ‘bare’ potassium salt, subsequent addition balanced reaction of two equiv. of K[Fe2(μ-SPh)(μ-CO)(CO)6]
of benzo-15-crown-5-ether did not afford the crown-K+ sup- with the dihalo C1 sources, aiming to achieve 2 × KX elimin-
ported cluster of formulation [K(benzo-15-crown-5)][Fe2(μ-SR) ation and isolation of a corresponding neutral complex con-
2+
2+
(μ-CO)(CO)6]. This could be due to the unique position of the K taining a CH2 or CF2 fragment within cluster. Indeed, pre-
in the crystal lattice, forming a network of bonds with neighbor- cipitation of KX from Et2O was observed, and a pentane
ing COs, thiolates and phenyl carbons (vide supra, Fig. 1). The soluble red material was isolated as the product. However, sub-
Et2O insoluble portion from the extraction process described sequent crystallization from pentane at −80 °C revealed the
above was then extracted with MeCN containing excess benzo- product as the original neutral dimer, [μ2-(SPh)2Fe2(CO)6] (1).
15-crown-5-ether afforded the (crown)alkali-supported trimeric Our tentative explanation for this behavior is that the electro-
species [K(benzo-15-crown-5)2]2[Fe3(μ3-S)(CO)9] (3).
philic carbon units accepted four electrons from two Fe0-Fe0
With the novel alkali-supported iron–sulfur–carbonyl clus- clusters to regenerate two FeI–FeI dimers, with concomitant
ters in hand, we investigated their propensity to react with C1 generation of C2X4, (X = H, F). Based on the stoichiometry
electrophiles (Scheme 1, top). Due to the reported examples required to produce complex 1, it can be deduced that some
for reactions of carbonyl-containing FeS clusters with electro- Fex(CO)y byproduct is also formed, but evidence of such clus-
philic carbon sources,22,23,30 CH2I2 and CF2Br2 were chosen as ters after chromatography of the crude product mixture was
not conclusive.
To deconvolute these results, a more straightforward set of
reaction conditions were chosen to investigate the reproduci-
bility of this process (Scheme 1). Reactions containing 0.5
equiv. of diphenyl-disulfide (Ph2S2) by itself did not regenerate
the [2FeI–2(SPh)] starting material 1. However, the combi-
nation of 0.5 equiv. of Ph2S2 plus 1.0 equiv. of [Fc]PF6 did lead
to quantitative conversion of anionic 2 to neutral 1. This is
likely due to the need for redox-driven elimination of potass-
ium (as KPF6) before the ensuing coordination chemistry can
occur at the iron centers. This suggests that the coordination
environment of the iron centers is controlled by the oxidation
state of the iron complex, and not the available ligand reagents
in solution. Indeed, this was further substantiated by the gene-
ration of the dianion K2[Fe3(μ3-S)(CO)9] (3) from the reduction
of [μ2-(SPh)2Fe2(CO)6] (2) with KC8 (Scheme 1). Simply reducing
the [FeIFeI] dimer leads to cleavage of an S–C bond to generate
a new bridging sulfide species, despite phenyl thiolate being
present in solution.
During the conversion of 1 to 3, we observed that free phenyl
thiolate (PhSK) is formed as a byproduct (observed by XRD in
Fig. 1 ORTEP diagram (50% thermal ellipsoids) of K[Fe2(μ-SPh)(μ-CO)
(CO)6] (2) at 100 K. Two molecules are shown to highlight the unique
subsequent reactions as C(SPh)4). The anionic PhS− (as K+ salt)
bonding motif of the K-counterion.
has quite similar solubility properties to the dianion 3, render-
9166 | Dalton Trans., 2017, 46, 9163–9171
This journal is © The Royal Society of Chemistry 2017

View Article Online
Dalton Transactions
Paper
ing solvent separation unviable. While 3 can be purified by the complex 4 to complex 1. By adding a source of Ph+, namely
addition of crowns and subsequent recrystallization, we were [Ph2I]PF6, the nucleophilic sulfides can be arylated. This con-
unable to isolate the bare potassium complex from the version is low-yielding, likely due to the formation of 5 and 6
reduction of 1. This limited the utility of 3 as a starting material, as side products, thus preventing a clean transformation.
as the synthetic versatility derives comes from the ability to use Using PhI as a reagent to perform this conversion was un-
multiple crown ethers in subsequent crystallization attempts. successful (no reaction with the reduced species), demonstrat-
Thus, another more direct route was devised. The known ing the need for a more labile Ph–I bond.
persulfide iron carbonyl dimer [Fe2(μ-S2)(CO)6] (4) was selected
While both 4 and 3 contain only iron, sulfide and carbonyl
due to the hypothesized ‘redox-only’ ease of converting the moieties, introduction of 2–4 equivalents of KC8 to 4 paradoxi-
μ2-(S2)2− ligand in 4 to the μ3-(S)2− moiety found in 3. cally does not generate 3. This is somewhat counterintuitive,
Surprisingly, reduction of 4 (red) in THF with KC8 in an as the conversion of persulfide to sulfide appears facile,
attempt to generate 3 (also red) instead generated a green solu- whereas de-arylation of phenyl thiolate to sulfide appears
tion,32,33 ultimately affording THF soluble and insoluble por- difficult. Further attempts to perform the 4 → 3 conversion
2
tions. Treatment of the THF soluble portion with dibenzo-15- using balanced stoichiometry via addition of
/
equiv.
3
crown-5 in THF afforded a sparing quantity of red polycrystal- [Fe3(CO)12] led only to elimination of sulfur and isolation (and
line material (not suitable for single crystal XRD). crystallization) of other known reduced iron clusters, such as
Recrystallization of this material from MeCN afforded single [K2(benzo-15-c-5)4][Fe2(CO)8] and [K(benzo-15-c-5)2]2[Fe(CN)4].
crystals of the novel dianionic complex [K(dibenzo-15-crown-
Relatedly, a synthesis by Cherng et al.26 demonstrated that
5)2]2[Fe(Fe2(μ-S)2(CO)6)2] (5), in which the sulfide moieties of [NEt4][Fe3(μ3-S)(CO)9] can be generated by refluxing KOH and
two [Fe2S2(CO)6]2− units serve as two sets of bidentate ligands, Fe(CO)5 in methanol with Na2SO3 as the sulfur source.
coordinating a central Fe(II) ion (Scheme 2). There is pre- Modification of this procedure finally allowed for 3 to be iso-
cedence for the starting material 4 serving as a bidentate lated as the bare potassium salt (a versatile synthon), or crys-
sulfur chelate following reduction of the persulfide bridge to tallized in the form [K2(benzo-15-c-5)4]2[Fe3(μ3-S)(CO)9]. In the
inorganic sulfides, as evident in the complexes reported by isolation of this complex, various crown ethers were employed
Song, namely [Fe2(CO)6(μ-S2)GeCl2] and [Fe4(CO)12(μ-S4)Ge].34
to aid in the solvent separation and subsequent crystallization
The THF insoluble green portion from the above reaction of the complex. Crown ethers offer the unique ability to
was re-crystallized from MeCN to afford dark red single crys- modify the solubility of charged alkali metal complexes, as
tals. However, instead of the expected outcome of the well as lowering the solvent effects of Lewis basic solvents
reduction of the persulfide bridge into two bridging sulfides, such as THF by encapsulating the potassium ion in a saturated
the product was identified as [K(benzo-15-crown-5)2]2[((Fe2S) oxygen coordination sphere36 (Fig. 5). Generally, an increase in
(CO)6)2(μ-S)2] (6), containing a ‘dimer of dimers’ vis a vis a per- the number of benzo groups on the crown ether correlates
sulfide bridge between di-iron units. This result is evidently with lower solubility. For instance, the application of a the
not a direct conversion from the starting material. The starting alkyl-only 15-crown-5 often results in K[FeSCO] complexes that
material 4 was treated with two equiv. of KC8, likely generating are too soluble to crystallize from MeCN/Et2O mixtures.
the doubly-reduced species (green solution) in situ.35 However, However, use of benzo-15-crown-5 imparts a more ideal solubi-
during isolation and crystallization, the complex auto-oxidized lity that facilitates crystallization in standard solvents.
(perhaps via reduction of MeCN) to generate the more stable Furthermore, the addition of dibenzo-15-crown-5 rendered the
persulfide bridge between the diiron units. As of yet, crystals K[FeSCO] clusters much less soluble, and all crystallization
of the green material have not been obtained. However, this attempts resulted in only powder products.
reduction can be used to afford a transformation from
Following the eventual isolation of pure 3 from the K2SO3
route, interconversions involving oxidations of K2[Fe3(μ3-S)
(CO)9] proceeded in a straightforward manner (Scheme 1).
Addition of 1.5 equiv. of S2Ph2 followed by 2 equiv. of [Fc]PF6
afforded [μ2-(SPh)2Fe2(CO)6] (1) as a product. Substitution of
SPh2 for S2Ph2 resulted in no conversion to 1. While the atom
identity of SPh2 would indicate a more straightforward conver-
sion, the greater stability of the S–C bond compared to the S–S
bond precluded the desired conversion. The analogous conver-
sion of 3 to 4 using [Fc]PF6 and S8 was also viable, albeit in low
yield (10%). We report this reaction not as a method to
efficiently generate 4, but rather as a proof of concept for this
type of interconversion. A common byproduct that was, in fact,
separated the mixture via chromatography was the known
violet iron cluster [Fe3S2(CO)9]; this cluster is also formed
during the synthesis of 4 and is easily recognized by its bright
color, IR spectrum and distinct Rf value (Fig. 1–4).
Scheme 2 Synthetic scheme detailing syntheses and interconversions
arising from compound 4 as reported in this work.
This journal is © The Royal Society of Chemistry 2017
Dalton Trans., 2017, 46, 9163–9171 | 9167

View Article Online
Paper
Dalton Transactions
cations. They were unable to crystallize the complex
[NEt3H][Fe2(SPh)(μ-CO)(CO)6], instead growing crystals of
[PPh4][Fe2(SPh)(μ-CO)(CO)6]. Although this complex was stable
for characterization, it may not be a suitable precursor for
further reductive or reactivity studies. Song et al. were,
however, able to crystalize [NEt3H][Fe2S(p-MeC6H4)(μ-CO)
(CO)6], demonstrating the importance of both cation selection
and the electronics of the sulfur ligand. In our work, cluster 2
appears to be stable due to the position adopted by the ‘free’
(i.e. no crown) potassium cation. The K+ ion is ‘sandwiched’
among a thiolate, a phenyl unit and the O atom of several CO
ligands. The Lewis acidity and (relatively) small size of the pot-
assium ion allows it to reside much closer to the atoms on the
cluster than organic cations would allow. This likely allows a
direct mode of stabilization of the complex by delocalizing
some of the electron density localized on the sulfur and CO
moieties. In cluster 2, the closest atom to the K is the oxygen
from the single bridging CO at 2.608(4) Å, followed by the
terminal CO ligands at 2.890(3) Å. Meanwhile, the thiolato
sulfur participates in a long-range contact with the potassium
ion [S⋯K = 3.1760(19) Å] (Table 1).
Fig. 2 ORTEP diagram (50% thermal ellipsoids) of [K2(benzo-15-c-
5)4]2[Fe3(μ3-S)(CO)9] (3) at 100 K.
Fe–S bonds. The wide range of sulfur-based donors in this
set of clusters results in an overall dataset that is not amenable
to predicting (or rationalizing) any systematic trend between
iron oxidation state and Fe–S bond distance. Complex 2 shows
a high extent of uniformity in the bond lengths of atoms that
are interchangeable through chemical symmetry operations
(but not crystallographic symmetry). Both of the Fe–S distances
are experimentally identical at 2.2676(13) Å. This is not present
in any of the other clusters, even if they contain higher sym-
metries in the FeS cluster. For example, complex 3 exhibits a
3-fold rotational axis, and yet all of Fe–S bonds are dissimilar
(2.1878(8), 2.1933(8), 2.1932(8) Å). This is likely due to the steric
bulk and asymmetric location of the potassium crown com-
plexes in the crystal structure, affording a lower symmetry space
group. In the sulfide-trimer cluster 3, the two faces of the
pyramid closest to the cation exhibit slightly longer bond
lengths for the metal–sulfide bonds (2.1933(8), 2.1932(8) Å vs.
2.1878(8) Å). Also, complex 3 contains the shortest Fe–S bonds,
as would be expected due to the high level of covalency between
the diffuse S2− and Fe(0) orbitals, with lengths ranging from
2.1878(8) Å to 2.1933(8) Å. This trimeric structure also contains
the most Fe–S bonds to a single sulfur (three); the higher extent
of bonding presumably results in the shorter bond distances.
Complex 5 exhibits a notable variation in its Fe–S bond
lengths, ranging from 2.2917(14)–2.3289(15) Å. Complex 6
(bearing the persulfide bridge between {Fe2S} units) exhibits
only a very slight variation in bond lengths: the distance from
each set of iron atoms to the bridging persulfide are 2.2574(11)
Å and 2.2644(11) Å (0.007 Å difference). The bonds from the
Fig. 3 ORTEP diagram (50% thermal ellipsoids) of [K(dibenzo-15-
crown-5)2]2[Fe(Fe2(μ-S)2(CO)6)2] (5) at 100 K.
Fig. 4 ORTEP diagram (50% thermal ellipsoids) of [K(benzo-15-crown-
5)2]2[((Fe2S)(CO)6)2(μ-S)2] (6) at 100 K.
X-Ray structures
K+ counterion in 2. Due to the redox active nature of iron– irons to the sulfides lie in a larger range from 2.2875(10) Å and
sulfur–carbonyl complexes and the nucleophilic character of 2.3028(11) Å (0.0153 Å difference). Both variations are due to
the reduced iron carbonyl species, characterization of clusters the steric bulk and asymmetric placement of the potassium
by X-ray diffraction can be difficult due to their instability. For crown complexes. The disulfide bond [S–S = 2.1657(17) Å] is
example, in 2013 Song and coworkers reported a synthesis of quite similar to the previously reported value of 2.164(2) Å (as
the anion of complex 2, crystallized with various organic AsPh4 salt).39
9168 | Dalton Trans., 2017, 46, 9163–9171
This journal is © The Royal Society of Chemistry 2017

View Article Online
Dalton Transactions
Paper
Fig. 5 ORTEP diagram (50% thermal ellipsoids) of a typical [K(benzo-15-crown 5)2] complex (in this case, derived from 3). Two perspectives are
shown to demonstrate the ‘sandwich-style’ bonding motif. The saturated coordination sphere prevents interactions between the potassium ion and
Lewis basic solvents.
Fe–Fe bonds. Despite the fact that Fe(I)–Fe(I) distances are persulfide), where the Fe(I)–Fe(I) distance is 2.555 Å. As a
generally shorter than Fe(0)–Fe(0) distances, we found no result, Fe(I)–Fe(I) distances ‘bookend’ the range of Fe–Fe
reliable correlation between oxidation state and Fe–Fe bond bonds, with the remaining Fe(I)–Fe(I) and Fe(0)–Fe(0) distances
distance. For example, while the shortest Fe–Fe bond length is in the intermediate range.
the Fe(I)–Fe(I) bond in the persulfide dimer (5) at 2.4885(11) Å,
Fe–C(O) bonds. Contrary to Fe–Fe bond lengths, the CuO
the next shortest is the Fe(0)–Fe(0) bond in complex 2 at bonds across all complexes follow a distinct trend correlated to
2.5006(11) Å. The longest Fe–Fe bonds by far among the three the oxidation state of the iron centers present in each cluster,
Fe(0) atoms in the sulfide trimer 3, with lengths ranging from as well as the overall charge of the complex. The neutral Fe(I)
2.5864(6) Å to 2.5940(5) Å. Notably, these distances can be con- complexes [μ2-(SPh)2Fe2(CO)6] and [μ2-(S)2Fe2(CO)6] have the
sidered as ‘outliers’, due to the trimer structure versus the di- shortest CuO bonds (1.132 Å and 1.135 Å respectively) as
iron core in all of the other clusters. Thus, within the set of expected due to their low back-bonding capabilities.
dimer-based clusters, the longest Fe–Fe distance is found in Meanwhile, the reduced thiolate cluster 2 contains only Fe(0),
the previously reported structure of 4 (Fe2S2 dimer, where S2 = the CuO bond lengths are shorter (1.145(4) Å) than the mixed
Table 1 Selected bond lengths and ν(CO) values (highest energy feature only) for complexes 2, 3, 5, and 6 from this work, as well as the metrics for
1 and 4 as described elsewhere
Fe2(SPh)2 a (1)
K[Fe2S] (2)
K2[Fe3S] (3)
Fe2S2 b (4)
K2[(Fe2S2)2Fe] (5)
K2[(Fe2S)2(μ2-S)] (6)
Two Fe(I)
One Fe(II)
Oxidation state
Fe(I)
Fe(0)
Fe(0)
Fe(I)
Fe(I)
Fe–Fe
2.494
2.5006(11)
2.5899(6)
2.5940(5)
2.5864(6)
2.1878(8)
2.1933(8)
2.1932(8)
—
2.555
2.5010(9)
2.4885(11)
2.5225(8)
Fe–S
2.262
2.287
2.2676(13)
2.227
2.245
2.2917(14)
2.3053(16)
2.3289(15)
—
2.2875(10)
2.3028(11)
S–S
Fe–μS
—
—
—
—
2.026
—
2.1657(17)
2.2574(11)
2.2644(11)
—
—
—
1.156(5)
—
1.795(4)
2023
—
—
μCO–K
t-CO–K
S–K
—
—
—
1.132
—
1.786
2072
2.608(4)
2.890(3)
3.1760(19)
1.145(4)
1.187(6)
1.791(4)
2027
—
—
—
1.171(3)
—
1.754(3)
1995
—
—
—
1.135
—
1.793
2085
—
—
—
1.152(6)
—
1.771(5)
2018
C–O
μC–O
Fe–CO
ν(CO) in cm−1
1726
Values derived from structures previously reported by a Adeleke et al.37 b Farrugia et al.38
This journal is © The Royal Society of Chemistry 2017
Dalton Trans., 2017, 46, 9163–9171 | 9169

View Article Online
Paper
Dalton Transactions
Fe(I)/Fe(II) complex 5 (1.152(6) Å) and the all Fe(I) complex 6
(3) A number of interconversions between iron–sulfur–car-
(1.156(5) Å). Due to the charge of 2 being only 1−, while 5 and bonyl clusters are possible, including reductive de-alklyation,
6 are both 2− charge. More importantly however, the K+ in 2 is as well as re-alkylation. The diphenyl iodonium salt [Ph2I]PF6
not encapsulated by a crown complex, instead forming an is a particularly useful reagent for mild arylation of iron-
extended lattice with contacts to several CO ligands, allowing bound sulfides.
for the donation of some electron density away from the CO π*
orbitals. In fact, compared to the same anion balanced by tet- of the complexes has a predictable effect on the C–O bond
raphenylphosphonium namely, (PPh4)[(μ-SPh)(μ-CO) lengths in the supporting carbonyl ligands. Alternately, the
(4) The iron oxidation states and extent of negative charge
–
Fe2(CO)6] – reported by Song et al., cluster 2 has drastically type of sulfur donor (thiolate, sulfide), nuclearity, and the posi-
shorter terminal and bridging CO ligands (complex 2: 1.145(4) tion and identity of the ‘bare’ K+ or K+-crown counterion(s)
Å and 1.187(6) Å vs. PPh4 salt: 1.174 Å and 1.221 Å). As exhibit the dominant effect(s) on the Fe–Fe and Fe–S bond
K2[Fe3(μ3-S)(CO)9] has the shared effects of an overall 2− charge distances.
on the FeS cluster, an all Fe(0) network, and the most negative
sulfur donor in the series (S2−), the CO bond lengths are the
longest at 1.171(3) Å.
Acknowledgements
Notably, the IR data corroborates the conclusions derived
from comparing the CuO bond metrics. The two neutral This work was funded by the American Chemical Society
[2Fe–2S] clusters 1 and 4 both exhibit their highest energy CO Petroleum Research Fund (53542-DN13), the Welch
stretching frequencies at 2072 and 2085 cm−1 (respectively), Foundation (F-1822) and a Cottrell Scholar Award to MJR from
which are the highest energy values overall in this set of clus- the Research Corporation for Scientific Advancement (RCSA
ters. Interestingly, the stretching frequencies of the corres- 23640). The authors gratefully acknowledge Dr Vince Lynch for
ponding PPh4 and K+ supported [(μ-SPh)(μ-CO)Fe2(CO)6] are assistance in obtaining and interpreting X-ray diffraction data.
+
quite similar.20 However, the K+-supported cluster reported
herein exhibits shorter CO bond lengths, thus indicating
comparatively lower electron density in the CO π* orbitals.
References
Interestingly, the IR stretches of the bridging carbonyl ligands
do not reflect this trend (νCvO PPh4 : 1727 cm−1, versus K+:
1726 cm−1). This discrepancy is possibly due to the differences
for complex 2 in the crystalline state with a rigorously encapsu-
lated K+ ion, whereas the powder sample (ATR-IR) may exhibit
a less ordered clathrate; this would decrease the electron
density transferred from CO to K+. This deviation notwith-
standing, the comparative electron density at each metal
center can be reasonably inferred from either the crystal struc-
ture data or IR data. Complexes 5 and 6 exhibit an intermedi-
ate extent of back-bonding, with their highest energy CO
stretches at 2018 cm−1 and 2023 cm−1, respectively. Following
this, the trimeric, dianionic cluster 3 exhibits highest extent of
back-bonding (1995 cm−1).
1 W. Lubitz, H. Ogata, O. Rudiger and E. Reijerse, Chem.
Rev., 2014, 114, 4081–4148.
2 T. R. Simmons, G. Berggren, M. Bacchi, M. Fontecave and
V. Artero, Coord. Chem. Rev., 2014, 270–271(1), 127–150.
3 T. Liu and M. Y. Darensbourg, J. Am. Chem. Soc., 2007, 129,
7008–7009.
4 J. M. Camara and T. B. Rauchfuss, Nat. Chem., 2012, 4(1),
26–30.
5 F. Gloaguen, J. D. Lawrence, T. B. Rauchfuss, M. Bénard
and M. M. Rohmer, Inorg. Chem., 2002, 41(25), 6573–6582.
6 I. Čorić and P. L. Holland, J. Am. Chem. Soc., 2016, 138(23),
7200–7211.
+
7 G. Christou, P. K. Mascharak, W. H. Armstrong,
G. C. Papaefthymiou, R. B. Frankel and R. H. Holm, J. Am.
Chem. Soc., 1982, 8, 2820–2831.
8 J.-L. Zuo, H.-C. Zhou and R. H. Holm, Inorg. Chem., 2003,
42(15), 4624–4631.
Summary and conclusions
In closing, we have performed a closed loop of interconver-
sions on a series of reduced, alkali–metal supported iron–
sulfur–carbonyl clusters. The main conclusions are as follows:
(1) The thermodynamically stable thiolate dimer 1 and per-
9 R. P. Pesavento, C. P. Berlinguette and R. H. Holm, Inorg.
Chem., 2007, 46(2), 510–516.
10 L.-C. Song, A.-G. Zhu and Y.-Q. Guo, Dalton Trans., 2016,
45(45), 5021–5029.
sulfide dimer 4 are dominant products in this series of trans- 11 G. Qian, W. Zhong, Z. Wei, H. Wang, X. Zhiyin, L. Long and
formations, owing to the stability of the diamagnetic Fe(I)–Fe(I) X. Liu, New J. Chem., 2015, 39, 9752–9760.
bonding system and their inability to serve as bidentate 12 C. A. Tooley, S. Pazicni and E. B. Berda, Polym. Chem.,
ligands to other adventitious iron centers. 2015, 6(44), 7646–7651.
(2) The crown/K+ counterion system is superior in isolating 13 S. Roy, T.-A. D. Nguyen, L. Gan and A. K. Jones, Dalton
reduced clusters to organic counterions like NEt4 or PPh4, due Trans., 2015, 44(33), 14865–14876.
to both the reductive stability of the K+ ion and the diverse 14 A. M. Lunsford, J. H. Blank, S. Moncho, S. C. Haas,
assortment of commercially available 18-crown-6 and
15-crown-5 reagents.
S. Muhammad, E. N. Brothers, M. Y. Darensbourg and
A. A. Bengali, Inorg. Chem., 2016, 55(2), 964–973.
9170 | Dalton Trans., 2017, 46, 9163–9171
This journal is © The Royal Society of Chemistry 2017

View Article Online
Dalton Transactions
Paper
15 S. Dey, A. Rana, D. Crouthers, B. Mondal, P. K. Das, 27 P. R. Challen, S.-M. Koo, W. R. Dunham and
M. Y. Darensbourg and A. Dey, J. Am. Chem. Soc., 2014,
136(25), 8847–8850.
16 T. Liu, B. Li, M. L. Singleton, M. B. Hall and
M. Y. Darensbourg, J. Am. Chem. Soc., 2009, 131(23), 8296–
8307.
D. Coucouvanis, J. Am. Chem. Soc., 1990, 112, 2455–2456.
28 R. L. Holliday, L. C. Roof, B. Hargus, D. M. Smith,
P. T. Wood, W. T. Pennington and J. W. Kolis, Inorg. Chem.,
1995, 34(17), 4392–4401.
29 J. W. Van Hal and K. H. Whitmire, Organometallics, 1998,
17(21), 5197–5201.
17 A. Banerjee, B. D. Yuhas, E. A. Margulies, Y. Zhang,
Y. Shim, M. R. Wasielewski and M. G. Kanatzidis, J. Am. 30 F. T. Al-Ani, D. L. Hughes and C. J. Pickett, J. Organomet.
Chem. Soc., 2015, 137(5), 2030–2034. Chem., 1986, 307(2), 31–34.
18 J. A. Wiig, Y. Hu and M. W. Ribbe, Proc. Natl. Acad. 31 Y.-C. Shi and F. Gu, Chem. Commun., 2013, 49(22), 2255–
Sci. U. S. A., 2011, 108(21), 8623–8627. 2257.
19 J. A. Wiig, Y. Hu and M. W. Ribbe, Nat. Commun., 2015, 6, 32 X. Wu, K. S. Bose, E. Sinn and B. A. Averill, Organometallics,
8034. 1989, 8, 251–253.
20 L.-C. Song, L. Li, Y.-Y. Hu and H.-B. Song, J. Organomet. 33 D. E. Barber, M. Sabat, E. Sinn and B. A. Averill,
Chem., 2013, 743, 123–129. Organometallics, 1995, (7), 3229–3237.
21 L.-C. Song, H. Tan, A.-G. Zhu, Y.-Y. Hu and H. Chen, 34 D. Seyferth, R. S. Henderson and L.-C. Song,
Organometallics, 2015, 34, 1730–1741. Organometallics, 1982, 1(1), 125–133.
22 K. Tanifuji, S. Tajima, Y. Ohki and K. Tatsumi, Inorg. 35 L.-C. Song, Acc. Chem. Res., 2005, 38, 21–28.
Chem., 2016, 55(9), 4512–4518.
23 D. Coucouvanis, A. Salifoglou, M. G. Kanatzidis,
36 O. M. Williams, A. H. Cowley and M. J. Rose, Dalton Trans.,
2015, 44(29), 13017–13029.
W. R. Dunham, A. Simopoulos and A. Kostikas, Inorg. 37 J. A. Adeleke, Y.-W. Chen and L.-K. Liu, Organometallics,
Chem., 1988, 27(22), 4066–4077.
1992, 11(7), 2543–2550.
24 D. M. Ottmers, Carbon, 1966, 4(1), 125–127.
25 J. Barrett, A. Spentzos and C. Works, J. Chem. Educ., 2015,
92(4), 719–722.
38 L. J. Farrugia, C. Evans, H. M. Senn, M. M. Hanninen
and R. Sillanpaa, Organometallics, 2012, 31(7), 2559–
2570.
26 J.-J. Cherng, Y.-C. Tsai, C.-H. Ueng, G.-H. Lee, S.-M. Peng 39 K. S. Bose, E. Sinn and B. A. Averill, Organometallics, 1984,
and M. Shieh, Organometallics, 1998, 17(2), 255–261.
3(7), 1126–1128.
This journal is © The Royal Society of Chemistry 2017
Dalton Trans., 2017, 46, 9163–9171 | 9171