Mononuclear Iron(II) Dicarbonyls Derived from NNS Ligands - Structural Models Related to a Pre-Acyl Active Site of Mono-Iron (Hmd) Hydrogenase

FULL PAPER

DOI:10.1002/ejic.201403013

Mononuclear Iron(II) Dicarbonyls Derived from NNS

Ligands – Structural Models Related to a “Pre-Acyl”

Active Site of Mono-Iron (Hmd) Hydrogenase

Keren A. Thomas Muthiah,^[a]Gummadi Durgaprasad,^[a]

Zhu-Lin Xie,^[a]Owen M. Williams,^[a]Christopher Joseph,^[a]

Vincent M. Lynch,^[a]and Michael J. Rose*^[a]

Keywords: Iron / Carbonyl ligands / Enzymes / Hydrogenase / Photolysis / Schiff bases

We report the syntheses and characterization of dicarbon-

yliron complexes derived from tridentate, ortho-substituted

of each dicarbonyl cation exhibits two ν(CO) stretches at ν ≈

˜

2070 and 2030 cm^–1. Low-temperature ¹H NMR spectroscopy

Schiff base pyridine/thioether ligands (_RNNS). Metalation re- measurements of 2-CO to 5-CO in CD₃CN (–35 to 5 °C) re-

actions of _RNNS (R = CH₃, OCH₃) at low temperature (–78 °C)

with [Fe(CO)₄(Br)₂] afforded the desired complexes [(_RNNS)-

vealed sharp resonances in the diamagnetic region. Under

dark conditions, each dicarbonyl species is relatively stable

Fe(CO)₂Br]Br (2-CO_Br, 3-CO_Br). Reactions under similar con- (Ͻ10% loss of CO, 1–2 h). However, photolysis revealed

ditions with more sterically demanding ligands [R = quinol-

ine (Q), ClPh] afforded complex salts of the form [(_RNNS)-

Fe(CO)₂Br][Fe(CO)₃(Br)₃] (4-CO_Feand 5-CO_Fe, respectively).

Alternatively, the metalation of the _RNNS ligands (for all

R ϶ H) with [Fe(CO)₄(Br)₂] in Et₂O at room temperature reli-

ably affords the complex species of type [(_RNNS)Fe(CO)₂Br]-

varying extents of photostability (stability rank: R =

OMeϾMe≈QϾClPh). An examination of the structural pa-

rameters reveals that higher photostabilities correlate with

shorter Fe–C(O) bond lengths, which are induced by varia-

tion of the ortho substituent of the pyridine ring. DFT calcula-

tions along the putative photolysis pathway revealed that the

[Fe(CO)₃(Br)₃] (2-CO_Fe, 3-CO_Fe, 5-CO_Fe). The metalation re- bulky ligand substituent (in 5-CO) destabilizes the mono-

actions of _RNNS at only moderately low temperatures (–20 to

0 °C) result in the loss of CO to form the corresponding trigo-

nal-bipyramidal iron(II) dibromide species of type [(_RNNS)-

FeBr₂] (2-Br, 4-Br, 5-Br; μ_eff≈ 5.3 μ_B, S = 2). The IR spectrum

carbonyl intermediate, and this is a likely explanation for its

more rapid rate of CO photodissociation. Relevance to a pos-

sible “apo-active site” of mono-iron hydrogenase (pre-acyl

formation) is discussed.

responsible for the utilization of H₂under physiological

conditions. Overall, these enzymes have been proposed to

serve as inspiration for the synthesis of base-metal com-

plexes that could replace platinum, iridium, and rhodium

complexes in catalytic processes.

Introduction

Hydrogenases are enzymes that generate or utilize di-

hydrogen (H₂) and can be found in both bacteria and ar-

chaea. Owing to the reliance of such microorganisms on

the natural abundance of elements, the enzymes utilize only

biologically available metal ions, most notably iron and

nickel. There are three known types of hydrogenases,

[NiFe]-, [FeFe]-, and the most recently discovered mono-

[Fe]-hydrogenase, also known as the dihydrogen-forming

In contrast to the better-studied [NiFe]- and [FeFe]-

hydrogenases, [Fe]-H₂ase is not a redox enzyme: it contains

no iron–sulfur electron transport system and, to date, only

the Fe^IIoxidation state has been observed in active prepara-

tions. Not surprisingly, the active site of Hmd is unique in

biological systems, as it splits molecular hydrogen, despite

the proposition that the metal center is redox-inactive. The

enzyme performs the heterolytic cleavage of H₂, followed

by a hydride transfer to methenyltetrahydromethanopterin

(H₄MPT⁺) to form the reduced product methylene-H₄MPT.

The methanopterin substrate acts as a C₁carrier in a meth-

anogenic CO₂fixation pathway. Thus, [Fe]-hydrogenase

performs one of the key steps in the reduction of carbon

dioxide to methane.^[1,2]

methylenetetrahydromethanopterin

dehydrogenase

(Hmd).^[1,2]Mono-[Fe] H₂ase is endogenously expressed in

some strains of methanobacter such as Methanocaldococcus

jannaschii (known enzyme X-ray structure), Methano-

thermobacter marburgensis, and Methanobrevibacter smithii.

Notably, the enzyme is only significantly expressed under

nickel-limiting conditions; otherwise, [NiFe]-hydrogenase is

[a] Department of Chemistry, The University of Texas at Austin,

Austin, TX 78712, USA

As shown in Scheme 1, the active site consists of one Fe^II

center ligated to a bidentate 6-hydroxypyridylacyl moiety,

two carbonyl ligands in cis orientation, the sulfur atom

E-mail: mrose@cm.utexas.edu

http://ww w .roseresearchgroup.org

Supporting information for this article is available on the

WWW under http://dx.doi.org/10.1002/ejic.201403013.

Eur. J. Inorg. Chem. 2015, 1675–1691

1675

www.eurjic.org

FULL PAPER

from the Cys₁₇₆residue, and the solvent/H₂coordination stabilize the cis-{Fe(CO)₂} moiety. Additionally, to high-

site trans to the acyl unit. The iron–acyl unit is one of the light the structure and properties of a putative apo- or inter-

distinctive features that sets the Hmd active site apart from mediate active site, we have avoided the use of an acyl li-

all other known metalloprotein active sites. The mechanism gand to stabilize the complex.

responsible for the formation of this rare metal–carbon

A close inspection of crystallographically characterized

bond is still unknown.^[3,4]Recently, Shima and co-workers iron dicarbonyls reveals surprisingly few examples of mono-

reported an elegant ¹³C-labeling study to trace the origin of nuclear non-phosphine- and non-acyl-supported iron di-

the acyl/pyridone-containing Fe–GMP (GMP = guanosine carbonyls. Of these select examples, several complexes pos-

monophosphate) cofactor.^[5]The highly substituted apo- sess nonbridging thiolates. For example, Pickett and co-

pyridone/GMP cofactor could be constructed biosyntheti- workers reported an early structural mimic of [Fe]-H₂ase

callyinautotrophicM.marburgensisandM.smithiifrom[¹³C]- derived from the tetradentate (N₂S₂), pyridine-based dithio-

acetate and ¹³CO₂(through pyruvateǞketoallulose). The late ligand shown below. The resulting complex exhibited a

generated pyridone contains an ortho-(CH₂)CO₂H substitu- mononuclear Fe^IIcenter ligated to cis-carbonyl ligands and

ent (the acyl precursor), but the exact pathway for the re- the deprotonated dithiolate ligand.^[11]To prepare CO-re-

duction and/or decarboxylation of the acidic moiety re- leasing molecules (CORMs), Westerhausen synthesized a

mains unclear. Interestingly, growth under a ¹³CO atmo- mononuclear iron carbonyl thiolate complex derived from

sphere led to ¹³C labeling of both the metal-bound carbonyl 2 equiv. of aminothiophenol, formulated as [(L_NS)₂Fe(CO)₂],

(CϵO) ligands and the Fe–¹³C(=O)_acylunit. Additionally, which also featured the cis-dicarbonyl motif.^[17]Liaw re-

Hu and co-workers found that the incubation of a synthetic ported the structure of a closely related complex derived

Fe–C(=O)_acylcomplex with ¹³CO under select conditions from 1 equiv. of the same ligand, which generated [(L_NS)-

(dark, several days; UV, hours) resulted in exchange to form Fe{cis-(CO)₂}(CN)]^–as a five-coordinate species.^[18]

a labeled Fe–¹³C(=O)_acylunit.^[6]This suggests a plausible

N-ligated methylpyridine or Fe–(CH₂)_alkylintermediate

during the exchange, wherein the acyl unit may be (re)-

formed across an intermediate by an organometallic pro-

cess.

Scheme 1. The possible protonation states in the active site of [Fe]-

hydrogenase.

In this work, we were interested to probe such a putative

apo-active site (or intermediate thereof) before the forma-

tion of the acyl unit. Many recently reported model com-

More prevalent, however, are other sulfur-bearing li-

plexes utilize an Fe⁰or Fe^–IIsource (namely, [Fe(CO)₅] or gands such as thiocarbamates, xanthates, thioureas, thio-

Na₂[Fe(CO)₄], respectively) as the metalating agent, and pyridines, and 2-thiothiophene (2TT). These ligands are less

this is the most efficient synthetic route to iron carbonyls prone to promoting dimerization and, thus, have been used

containing an acylpyridyl unit.^[6–8]As we were interested in in mononuclear systems. For example, Hu reported a struc-

the isolation of apo/intermediate complexes without the tural [Fe]-H₂ase model derived from 2-thiopyridine (pyS),

acyl unit, Fe^IIsources were most applicable. Additionally, namely, [(pyS)Fe{cis-(CO)₂}(CH₃CO)(CN)]^–,^[10]whereas

the most likely biological source of iron carbonyl during the Liaw and co-workers structurally characterized the related

assembly of the active site is Fe^II. As such, we have utilized complexes [(pyS)Fe{cis-(CO)₂}(trans-CN)₂]^–and [(pyS)₂Fe-

[Fe(CO)₄(X)₂] (X = Br, I), as reported by others.^[9–12]Sev- {cis-(CO)₂}].^[19,20]

The

thiophene-derived

complex

eral recent reports regarding the assembly of the [FeFe] H- [(bpy)(2TT)₂Fe{cis-(CO)₂}] (bpy = 2,2-bipyridine) was also

cluster also provide inspiration for investigations regarding reported.^[21]Mononuclear carbonyls such as [(EtOCS₂)Fe-

the biogenesis of the mono-[Fe] active site.^[13–16]

{cis-(CO)₂}(CN)₂]^–(xanthate)and[(Et₂NCS₂)Fe{cis-(CO)₂}-

This research is focused on the utilization of biologically (CN)₂]^–(carbamate) may also be reliably derived from thio-

relevant ligands as found in the active site, such as N and carboxylates.^[19]The thiourea-derived complex [(L_NS)₂Fe-

S donors in addition to carbonyl ligands. That is, we wished {cis-(CO)₂}] is another closely related example.^[22]

to avoid phosphines, cyclopentadienyl (Cp) derivatives,

However, an inspection of the structure database reveals

carbenes, or carbanion-based pincer ligands to artificially no examples of Fe^IIcarbonyls derived from mixed N/S

Eur. J. Inorg. Chem. 2015, 1675–1691

1676

www.eurjic.org

FULL PAPER

donor sets containing a thioether motif; such a donor set plex, presumably facilitating the dissociation of the bromide

would offer a different strategy to include sulfur donors and ions (poor leaving group in neat CH₂Cl₂). The resulting

prevent dimerization. Thus, we devised a simple series of bromide salts could be recrystallized at low temperatures

ortho-substituted Schiff base pyridine/thioether ligands from MeCN/Et₂O or DMF/Et₂O. Notably, 3-CO_Brcrys-

[_RNNS; R = Me, OMe, quinoline (Q), ClPh] to explore the tallizes as an H₂O·MeCN solvate (even from DMF); there-

importance of the substituent to the stabilization or destabi- fore, it is stable to both coordinating and protic solvents.

lization of the cis-{Fe-(CO)₂} fragment. In this work, we Thus, it appears that a temperature-dependent, dissociative

present the synthetic routes, X-ray structures, and proper- mechanism is the main pathway of CO loss. Along the same

ties of iron(II) cis-carbonyls derived from a set of Schiff lines, reaction or recrystallization attempts above –20 °C re-

base pyridine/thioether ligands. We report the effects of sulted in the loss of CO and the isolation of the correspond-

light, temperature, and reaction conditions (solvent, ing green dibromide species [(_RNNS)Fe(Br)₂] in all cases.

stoichiometry, CO gas) in the preparation and characteriza- Notably, the isolation of the bromide salts was exclusive to

tion of the complexes.

the _MeNNS and _OMeNNS ligands. For example, under the

same metalation conditions (CH₂Cl₂/MeCN, low tempera-

ture), reactions with the bulkier _QNNS ligand afforded the

same cation, but it was instead paired with a complex anion

in the form [(_QNNS)Fe(CO)₂Br][Fe(CO)₃(Br)₃] (4-CO_Fe).

The reasons for this are discussed further below.

Results and Discussion

Syntheses and Reactivity

The same NNS-ligated iron carbonyl cations could also

be prepared at room temperature by the reactions of _RNNS

with [Fe(CO)₄(Br)₂] in Et₂O (both reactants soluble), which

led to the immediate precipitation of the crude products.

Recrystallization from MeCN/Et₂O at –40 °C afforded

orange to red crystalline samples of complex species of the

type [(_RNNS)Fe(CO)₂Br][Fe(CO)₃(Br)₃] [R = CH₃(2-

CO_Fe), OCH₃(3-CO_Fe), and Q (4-CO_Fe)]. One exception

was that the recrystallization of [(_ClPhNNS)Fe(CO)₂Br]-

[Fe(CO)₃(Br)₃] from MeCN/Et₂O was especially slow (2–4

weeks) and afforded a different complex salt, namely,

[(_ClPhNNS)Fe(CO)₂Br][Fe(Br)₄] (5-CO_FeBr). No carbonyl

products were isolated for the simplest ligand with R = H.

In the context of other synthetic approaches, Dar-

ensbourg and co-workers performed metalations of mono-

The NNS ligands used in this work were prepared from

the condensations of ortho-substituted pyridinecarbal-

dehydes with 2-(methylthio)aniline in MeOH at room tem-

perature or under mild reflux conditions. The iron carbonyl

bromide starting salt [Fe(CO)₄(Br)₂] was prepared by modi-

fication of a previously reported procedure.^[9]The overall

synthetic and reactivity scheme is summarized in Scheme 2.

The metalation of unsubstituted _HNNS with [Fe(CO)₄-

(Br)₂] was pursued initially in tetrahydrofuran (THF),

CH₂Cl₂, and MeCN under ambient conditions; this re-

sulted in the formation of violet bisligated species of the

type [(_RNNS)₂Fe](X)₂as the only isolable product (as the

–

Br^–salt or the BF₄salt after AgBF₄treatment). We next

attempted to prevent the loss of CO by performing the met-

alations at low temperature.

At low temperatures (–80 °C), the metalation of _RNNS dentate ligands (carbenes, py, PPh₃) and bidentate ligands

with [Fe(CO)₄(Br)₂] in CH₂Cl₂afforded green solutions, (bpy, 2-pyridone) with [Fe(CO)₄I₂] in hexane at room tem-

which upon the addition of MeCN afforded bright orange perature;^[23]no bisligated products were reported in these

products of the desired carbonyl complexes [(_RNNS)- cases, except for PMe₃, which afforded the complex

Fe(CO)₂Br]Br [R = CH₃(2-CO_Br), OCH₃(3-CO_Br)]. The [Fe(CO)₂(PMe₃)₂I₂] under certain conditions. These re-

addition of MeCN was required to isolate the desired com- searchers noted that dark conditions were required for the

Scheme 2. Synthetic pathways for iron species with _RNNS (R = CH₃, OCH₃, Q, PhCl) ligands derived from metalation with [Fe(CO)₄-

(Br)₂] or FeBr₂.

Eur. J. Inorg. Chem. 2015, 1675–1691

1677

www.eurjic.org

FULL PAPER

isolation of the complexes, and this is also an important ranged from brown to green and blue-green. The ligands

point in this work (vide infra).^[9]

with the least restrictive R groups (R = H, OCH₃) afforded

As we at first serendipitously obtained the dibromide only violet solutions of the bisligated dication, that is,

species that resulted from the loss of CO from the carbonyl [(_RNNS)₂Fe]²⁺. None of the isolated dibromide species

complexes, we subsequently independently synthesized could be converted into the corresponding carbonyl com-

these species in higher yields by the metalation of FeBr₂pounds under 1 atm of CO (–40 °CǞroom temp.), most

with _RNNS in THF. A selection of ligands with bulky sub- likely because of the low affinity of the high-spin Fe^IIcenter

stituents (R = CH₃, Q, ClPh) afforded dark green solutions (S = 2, vide infra) for CO. The crystal data and refinement

of the desired dibromide species (2-Br, 4-Br, and 5-Br, parameters for the carbonyl, dibromide, and bisligated com-

respectively). Crystallization from pentane or Et₂O vapor plexes are summarized in (Table 1), and their structures will

diffusion afforded single crystals of the products, which be discussed in the next section.

Table 1. Crystal data and refinement parameters for the carbonyl, dibromide, and bisligated complexes.

2-CO_Br

3-CO_Fe

5-CO_Fe

2-Br

Formula

FW

Color

C₁₈H₁₉N₃O₃SBr₂Fe

573.09

orange

C₂₁H₁₇Br₄Fe₂N₄O₆S

870.78

orange

C₂₁H₁₅Br₅ClFe₂N₂O₂S

906.11

red

C₁₄H₁₄Br₂N₂S

458.00

brown

Habit

Size [mm]

T (K)

block

0.25ϫ0.18ϫ0.11

163(2)

needle

0.23ϫ0.08ϫ0.04

153(2)

needle

0.08ϫ0.005ϫ0.002

100(3)

block

0.32ϫ0.25ϫ0.20

163(2)

λ [Å]

0.71073

orthorhombic

Pna2₁

13.133(3)

18.926(5)

8.738(2)

90

0.71073

0.77490

orthorhombic

Fdd2

29.751(6)

36.638(7)

10.333(2)

90

11263(4)

16

2.137

8.317

1.029

R1 = 0.0508

wR₂= 0.1103

R1 = 0.0801

wR₂= 0.1231

0.71073

monoclinic

P2₁/n

11.794(4)

10.086(3)

13.714(5)

90

102.913(4)

90

1590.0(9)

4

1.913

6.096

1.177

R1 = 0.0346

wR₂= 0.0728

R1 = 0.0424

wR₂= 0.0798

Lattice

Space group

a [Å]

b [Å]

c [Å]

α [°]

β [°]

triclinic

¯

P1

8.8332(3)

13.5199(4)

13.6379(5)

114.510(2)

101.668(2)

98.957(2)

1397.86(8)

2

2.069

6.870

1.010

R1 = 0.0299

wR₂= 0.0561

R1 = 0.0521

wR₂= 0.0635

90

γ [°]

V [Å³]

2171.9(10)

4

1.753

4.493

1.059

R1 = 0.0340

wR₂= 0.0665

R1 = 0.0406

wR2 = 0.0691

Z

d_calcd.[g/cm³]

μ [mm^–1]

GOF on F²

R indices [IϾ2σ(I)]

R indices (all data)

4-Br

5-Br

6-Br

1-L2

Formula

FW

Color

Habit

Size [mm]

T (K)

C₁₇H₁₄Br₂Fe N₂S

494.03

red

needle

0.29ϫ0.13ϫ0.10

153(2)

C₂₀H₁₇Br₂Cl₃FeN₂S

639.43

green

needle

0.13ϫ0.10ϫ0.07

153(2)

C₁₅H₁₆Br₂FeN₂S

472.03

green-blue

block

0.24ϫ0.09ϫ0.04

100(2)

C₅₄H₅₁N₉S₄B₄F₁₆Fe₂

1413.22

violet

parallelepiped

0.4ϫ0.4ϫ0.2

153(2)

λ [Å]

0.71073

monoclinic

P2₁/n

9.8718(6)

14.6668(7)

16.4285(11)

90

95.639(2)

90

2367.1(2)

4

1.794

4.450

1.054

0.71073

orthorhombic

P2₁2₁2₁

9.3989(7)

11.9399(8)

15.3156(11)

90

1718.7(2)

4

1.824

5.642

1.062

R1 = 0.0326

wR₂= 0.0609

R1 = 0.0406

wR₂= 0.0632

0.71073

monoclinic

Pn

8.2970(10)

19.867(2)

35.702(3)

90

95.824(3)

90

5854.6(10)

4

1.603

0.737

1.885

Lattice

Space group

a [Å]

b [Å]

c [Å]

α [°]

β [°]

triclinic

¯

P1

8.0880(4)

10.2583(5)

11.1872(5)

72.491(2)

76.713(3)

81.492(3)

858.38(7)

2

1.911

5.654

1.087

R1 = 0.0291

wR₂= 0.0634

R1 = 0.0442

wR₂= 0.0689

γ [°]

V [Å³]

Z

d_calcd.[g/cm³]

μ [mm^–1]

GOF on F²

R indices [IϾ2σ(I)]

R1 = 0.0287

wR₂= 0.0597

R1 = 0.0361

wR₂= 0.0630

R1 = 0.0957

wR₂= 0.2327

R1 = 0.1045

wR₂= 0.2353

R indices (all data)

Eur. J. Inorg. Chem. 2015, 1675–1691

1678

www.eurjic.org

FULL PAPER

X-ray Structures

ide ion; the crystallized complex is an MeCN·H₂O solvate,

which indicates the stability of the carbonyl ligands in the

presence of these solvents under the crystallization condi-

tions (aerobic, dark, hydrated MeCN, –20 °C). The iron

center exhibits a pseudo-octahedral geometry, and two

carbonyl ligands are bound in cis fashion (the more com-

monly observed orientation). In addition, the neutral NNS

chelate binds in equatorial fashion, and the remaining coor-

dination site is occupied by one bromide ligand. The Fe–

N_pybond length [2.039(3) Å] is longer than those typically

[(HNNS)₂Fe][BF₄]₂(1-L2)

The structure of the bisligated species 1-L2 is shown in

Figure 1. The two NNS ligands chelate the metal center in

an overall pseudo-octahedral geometry without any steric

repulsion. The thioether (S)CH₃group of each NNS moiety

is tilted away from the pyridyl fragment of the adjacent

NNS ligand. The complex exhibits typical bonding param-

eters for a low-spin Fe^IIspecies, including relatively short

Fe–N_pyand Fe–N_SBbond lengths [1.938(8) and 2.007(8) Å,

respectively]. The Fe–S_CHdistance is comparatively longer

3

[2.269(3) Å] but well within the normal range for the bind-

ing of neutral sulfur donors to a low-spin iron(II) center.

Figure 1. ORTEP diagram (50% thermal ellipsoids) of the dication

of the bisligated complex 1-L2. Hydrogen atoms and counterions

are omitted for clarity.

[(_MeNNS)Fe(CO)₂Br]Br·MeCN·H₂O (2-CO_Br)

Figure 2. ORTEP diagrams (50% thermal ellipsoids) of _MeNNS

complexes. Top: 2-CO_Br. Bottom: 2-Br. Hydrogen atoms are omit-

This complex (Figure 2) crystallizes as a cationic species

with the charge balance provided by an outer-sphere brom- ted for clarity.

Table 2. Selected bond lengths [Å] and angles [°] derived from the X-ray structures of the dicarbonyl, dibromide, and bisligated complexes.

2-CO_Br

3-CO_Fe

5-CO_Fe

2-Br

4-Br

5-Br

6-Br

1-L2

R¹, R²

CH₃, H

2.039(3)

1.957(3)

2.2675(11)

2.4618(8)

–

OCH₃, H

2.013(2)

1.955(3)

2.2489(8)

2.4546(6)

–

ClPh, H

2.017(10)

1.956(10)

2.257(4)

2.443(2)

–

CH₃, H

2.200(3)

2.153(3)

2.5519(12)

2.4665(8)

2.4258(10)

–

Q, H

ClPh, H

2.2376(19)

2.1262(19)

2.5793(7)

2.4576(4)

2.4421(4)

–

H, CH₃

2.166(3)

2.126(3)

2.5696(11)

2.4103(7)

2.5004(7)

–

H, H

1.938(8)

2.007(8)

2.269(3)

–

Fe–N_py

Fe–N_SB

Fe–S_Me

Fe–Br

2.196(2)

2.133(2)

2.5758(8)

2.5023(5)

2.4060(5)

–

Fe–Br

Fe–C(O)

trans SB

(Fe)C–O

trans SB

Fe–C(O)

trans Br

(Fe)C–O

trans Br

R¹···C(O)_SB

1.819(4)

1.826(3)

1.836(13)

1.138(4)

1.801(4)

1.128(4)

ca. 2.61

1.129(4)

1.801(4)

1.130(4)

ca. 2.63

1.115(16)

1.816(14)

1.108(16)

ca. 2.90

–

Eur. J. Inorg. Chem. 2015, 1675–1691

1679

www.eurjic.org

FULL PAPER

found for Fe–N_pybonds in low-spin Fe^IIcomplexes, most fashion, but twinning issues have precluded full structure

likely because of the steric effect of the 2-methylpyridine solution at this time (see connectivity structure, Figure S4).

moiety, which prevents the complete approach of the ligand

[(_PhClNNS)Fe(CO)₂Br][Fe(Br)₄] (5-CO_FeBr

)

to the metal center. The Fe–N_SB[1.957(3) Å] and Fe–S

[2.2675(11) Å] bond lengths appear unaffected by the meth-

ylpyridine unit and are within the expected range for a low-

spin Fe^IIion. The axial and equatorial carbonyl ligands ex-

hibit distinct metric parameters [Fe–C(O) = 1.801(4),

1.819(4) Å; C–O = 1.128(4), 1.138(4) Å] owing to the dif-

ferent charges of the trans ligands (Br^–vs. neutral N_py). One

notable feature is the close contact between the pyridyl CH₃

moiety and the equatorial carbonyl ligand. The modeled

pyridyl CH₃protons (not located in the density map) sug-

gest a CH₃···C(O) contact distance of ca. 2.2–2.6 Å, which

is smaller than the sum of the Van der Waals radii. As a

result, the Fe–N_pybond is elongated [2.039(3) Å] compared

to those for the other ortho substituents owing to repulsion

between the CH₃group and the carbonyl ligand (see

Table 2). The connectivity of the corresponding complex

salt [(_MeNNS)Fe(CO)₂Br][Fe(CO)₃(Br)₃] (2-CO_Fe) was also

confirmed by a preliminary single-crystal X-ray diffraction

study (Figure S1).

We also prepared an NNS iron carbonyl complex bearing

an aryl substituent at the pyridine ortho position. Owing to

the higher solubility of this complex compared with those

of the other complexes, the anion in the complex

[(_PhClNNS)Fe(CO)₂Br][Fe(CO)₃(Br)₃] (obtained initially

from a reaction in Et₂O) undergoes slow oxidation under

crystallization conditions (aerobic MeCN/Et₂O) to form the

ferric [Fe(Br)₄]^–anion and afford the isolated complex 5-

CO_FeBr(Figure 4). Under an inert atmosphere, the original

complex yielded crystals of low quality for X-ray structure

determination (Figure S5). In the structure, the chloro-

phenyl group is severely rotated (dihedral angle 75.23°) to

accommodate the binding of the intact (and nearly unper-

turbed) Fe–CO unit in the adjacent coordination position.

Despite the apparent steric pressure, the Fe–N_pybond

length [2.017(10) Å] remains comparable to those in 3-CO_Fe

and 4-CO_Fe(see Table 2) and is much shorter than that

found in proximity to the sterically impinging methylpyr-

idine unit in 2-CO_Br[Fe–N_py= 2.039(3) Å].

[(_OMeNNS)Fe(CO)₂Br][Fe(CO)₃(Br)₃] (3-CO_Fe)

The structure of the 2-methoxypyridine derivative is

shown in Figure 3. Like that of 2-CO_Br, the structure exhib-

its a pseudo-octahedral geometry with an identical arrange-

ment of the NNS, cis-CO, and bromide ligands. The Fe–

N_pybond length of 2.013(2) Å in 3-CO_Feis shorter than

that observed in the 2-methylpyridine derivative 2-CO_Br

[2.039(3) Å], as the O atom of the methoxy group is less

bulky. Interestingly, the H₃CO_py···C(O) contact is still quite

close (ca. 2.63 Å) and well within the sum of the Van der

Waals radii for carbon and oxygen (1.7 + 1.5 = 3.2 Å). The

remaining bond lengths are unremarkable and are listed in

Table 2. The connectivity of the corresponding bromide salt

[(_OMeNNS)Fe(CO)₂Br]Br (3-CO_Br) was also confirmed by a

preliminary single-crystal X-ray diffraction study (Fig-

ure S3). The analogous quinoline complex [(_QNNS)Fe-

(CO)₂Br][Fe(CO)₃Br₃] (4-CO_Fe) also crystallized in this

Figure 4. ORTEP diagram (50% thermal ellipsoids) of 5-CO_FeBr

.

Hydrogen atoms are omitted for clarity.

[(_RNNS)Fe(Br)₂] (2-Br, 4-Br, 5-Br, and 6-Br)

ORTEP diagrams of the four dibromide complexes are

either displayed alongside the corresponding dicarbonyl

species in Figure 2 or in Figure 5. Each complex is neutral

overall and exhibits a five-coordinate Fe ion bound to a

mostly planar NNS ligand frame. One notable exception is

the quinoline-based _QNNS complex, which exhibits signifi-

cant distortion of the conjugated ligand; there is a differ-

ence of 28.96° between the quinoline and arylthioether moi-

eties (cf. a 7.53° offset in 3-CO_Br). All of the bond lengths

from the Fe ion to the NNS ligand (Fe–N_py≈ 2.20 Å; Fe–

N_SB≈ 2.15 Å; Fe–S ≈ 2.65 Å) are significantly longer than

those in the corresponding carbonyl complexes owing to

the high-spin configuration of the iron center (vide infra).

The Fe–Br bond lengths are more variable, within a narrow

range from ca. 2.4 to 2.5 Å; 4-Br exhibits the longest Fe–

Br length [2.5023(5) Å], most likely because of repulsion be-

tween the infringing quinoline C–H unit and the equatorial

bromide ligands. The remaining bond lengths are unre-

markable (Table 2).

Figure 3. ORTEP diagram (50% thermal ellipsoids) of 3-CO_Fe

.

Hydrogen atoms are omitted for clarity.

Eur. J. Inorg. Chem. 2015, 1675–1691

1680

www.eurjic.org

FULL PAPER

Figure 5. ORTEP diagrams (50% thermal ellipsoids) of the dibromide complexes 5-Br, 6-Br, and 4-Br. Hydrogen atoms are omitted for

clarity.

Characterization of the Dibromides

from the more ideal trigonal-bipyramidal geometries ob-

served in 2-Br, 4-Br, and even 5-Br. Thus, it is possible that

the puckering of the _MeN_MeNS framework in 6-Br contrib-

utes to its higher observed magnetic moment.

Magnetic Susceptibility

Solid-state magnetic measurements (298 K) on the di-

bromide complexes revealed a high-spin Fe^IIcenter in all

cases. We were unable to prepare the dibromide species de-

rived from the unsubstituted _HNNS ligand, owing to its

propensity to form only the diamagnetic bisligated species

DFT Calculations for the Dibromides

To shed more light on the experimental results above, we

(even from FeBr₂in THF or toluene). Considering the com- performed DFT calculations (6-31G*/PW91) on the di-

plexes of the aldehyde-derived ligands, 2-Br and 4-Br af- bromide complexes. In each case, energy calculations with

forded both the highest and lowest μ_effvalues of 5.48 and the X-ray structure coordinates were tested for both S = 2

5.21 μ_Brespectively (for S = 2, the spin-only value, μ_SO,is and S = 1 configurations. The aldehyde-derived complexes

4.90 μ_B). The dibromide derived from the iminomethyl li- 2-Br, 4-Br, and 5-Br exhibited the S = 2 configuration as

gand, 6-Br (no corresponding carbonyl complex was iso- the lowest energy state by a wide margin of ca. 20 kcal/mol

lated), has a slightly higher value of 5.52 μ_B. Indeed, a no- (20.3, 21.7, and 20.3 kcal/mol, respectively). The ketone-de-

table steric clash between the iminomethyl moiety and the rived (i.e., iminomethyl) complex 6-Br also exhibited an S

pyridine unit (vide supra, X-ray section) breaks the conju- = 2 configuration as the ground spin state but with a

gation between the pyridyl and arylthioether moieties in the smaller energy difference (16.6 kcal/mol) between the S =

ligand framework. As a result, there is greater distortion 2 and S = 1 configurations. Overall, the DFT results are

Table 3. Properties of _RNNS Fe^IIdribromide species.

Experimental

μ_eff[μ_B] (S = 2: μ_so= 4.90 μ_B)

DFT

S = 1 [kcal/mol]

Complex

λ [nm] (ε, m^–1cm^–1)

S = 2 [kcal/mol]

2-Br

4-Br

5-Br

6-Br

460 (650)

460 (1180)

460 (1570)

460 (780)

5.48

5.21

5.24

5.52

reference (i.e., 0)

+20.3

+21.7

+20.3

+16.6

0

Eur. J. Inorg. Chem. 2015, 1675–1691

1681

www.eurjic.org

FULL PAPER

1

consistent with the experimental magnetic measurements through the loss of CO. Thus, the H NMR spectra were

(solid state, 298 K; Table 3).

recorded at low temperature.

Photolysis of CO Ligands

Solution Studies of the Dicarbonyl Complexes

We wished to quantitatively assess the photostabilities of

this series of iron carbonyls. Measurements were standard-

ized by using the same salt of each complex. Complexes 2-

CO_Feto 5-CO_Fewere prepared in MeCN at standard UV/

Vis concentrations (ca. 0.1 mM) in quartz cuvettes, and the

samples were subjected to broadband illumination from a

white-light source (100 mW/cm²). Conversion to the corre-

sponding dibromide species (i.e., the loss of 2 equiv. of CO)

was monitored by the increase in the absorption at λ =

460 nm, the characteristic absorption feature of the di-

bromide species. The changes in the absorption of 2-CO_Fe

during photolysis for 50 min are shown in Figure 7. A dark

control experiment (see Supporting Information, Figure

S11) over the same period as that used in the photolysis

experiment revealed less than 10% conversion. Of note is

the decrease in the metal-to-ligand charge-transfer (MLCT)

Light/Dark Stability

The conditions under which we isolated the NNS-ligated

iron dicarbonyls provided some insight into their stability

in solution. For example, recrystallization at low tempera-

tures (–20 °C or below) was required for the isolation of

pure samples of each complex. Otherwise, the loss of CO

led to the formation of the dibromide species. At ambient

temperatures, we observed that orange solutions of the

complexes in MeCN (or DMF to a lesser extent) in the

dark are somewhat stable over the course of minutes to tens

of minutes. The conversion to the dibromide is apparent by

a distinct color change to green.

¹H NMR Spectra

1

The H NMR spectrum of 2-CO_Fein CD₃CN at –15 °C band exhibited by 2-CO_Feat λ ≈ 380 nm, which most likely

is shown in Figure 6. The aromatic resonances and multi- emanates from d_π(Fe)Ǟπ*(CO) charge transfer. The forma-

plicities are clearly resolved, and the resonance of the Schiff tion of 2-Br is indicated by the increase in the broad absorp-

base proton (δ = 9.68 ppm) is shifted downfield compared tion band centered at λ ≈ 460 nm. We assign this band as a

to that of the free ligand (δ = 8.49 ppm). In the alkyl region, weak MLCT band (ε ≈ 650 m^–1cm^–1) from the Fe^IIcenter

the methylpyridine CH₃group (δ = 3.03 ppm) undergoes a to the π* system of the ligand. The analogous changes in

marked downfield shift relative to that of the free ligand (δ the UV/Vis absorption upon the photolysis of 3-CO_Fe, 4-

= 2.46 ppm) owing to the partial positive charge at the adja- CO_Fe, and 5-CO_Fe(as well as dark control experiments) are

cent N_pyatom upon binding to the Fe^IIcenter. In contrast, similar (Figures S12–S14).

the S–CH₃resonance (δ = 2.64 ppm) remains nearly un-

To confirm that the changes in the UV/Vis absorption

changed compared to that of the free ligand (δ = 2.63 ppm), spectrum correlated with the loss of CO ligands, we moni-

as the electron density at the thioether S donor is not tored the reaction by solution IR spectroscopy during the

greatly affected in its relatively weak metal-bonding interac- photolysis. The bromide salt of the _MeNNS complex, that

tion.

is, 2-CO_Br, was used to ensure that the cation-based CO

1

The H NMR spectra of the remaining carbonyls (2-CO stretches of [Fe(CO)₃(Br)₃]^–did not correlate with our UV/

to 5-CO, both Br and complex salts) all exhibit relatively Vis spectroscopy observations. For comparison, the solid-

sharp peaks in the diamagnetic region, consistent with an state IR spectrum of 2-CO_Bris shown in the inset of Fig-

analogous low-spin Fe^IIconfiguration (Figures S6–S10). In ure 8 (all other IR spectra are shown in Supporting Infor-

each case, orange solutions of the dicarbonyl complexes in mation, Figures S15–S19). As shown in Figure 8, the spec-

CD₃CN turned green when equilibrated to ambient tem- tra (0–30 min) show a systematic decrease in the intensities

perature, indicating the formation of the dibromide species of the CO stretches, which confirms that the changes in the

1

Figure 6. H NMR spectrum of 2-CO_Fein CD₃CN at –25 °C (500 MHz).

Eur. J. Inorg. Chem. 2015, 1675–1691

1682

www.eurjic.org

FULL PAPER

Figure 7. Changes in the UV/Vis absorption spectrum of 2-CO_Fe

during photoconversion to the corresponding Fe^IIdibromide 2-Br

by dissociation of the CO ligands. Photolysis conditions: MeCN,

100 mW/cm²AM1.5 white light, 298 K.

Figure 9. Photoconversion of the cationic dicarbonyl species (as

indicated) to the Fe^IIdibromides by dissociation of the CO ligands.

Inset: expanded view of the same plot at short time periods. Photol-

ysis conditions: MeCN, 100 mW/cm²white light, 298 K.

UV/Vis spectra correlate with the loss of the CO ligands at

Table 4. Observed rates of photoconversion from the _RNNS cat-

ionic dicarbonyl species to the Fe^IIdibromide species (MeCN,

AM1.5 source, 100 mW/cm², 298 K). Less than ca. 10% thermal

conversion was observed under dark conditions over the same time

period (ca. 60–120 min).

the metal center. Note that the feature at ν = 1620 cm^–1

˜

(ligand-based C=N stretch) remains largely unchanged.

Complex

t_1/2[min]

2-CO_Fe

3-CO_Fe

4-CO_Fe

5-CO_Fe

12.0

22.6

12.7

8.5

ylpyridine moiety, i.e., trans from N_SB) observed in 2-CO_Br

[1.819(4) Å] compared with those in the other complexes

(ca. 1.82–1.84 Å). Indeed, in agreement with this pattern is

the correlation of the kinetically fastest transformation by

5-CO_Fewith the longest Fe–C(O) bond [1.836(13) Å] adja-

cent to the ortho substituent. However, the slowest t_1/2value

[3-CO_Fe: 22.6 min, 1.826(3) Å] does not correlate with the

shortest Fe–C(O) bond adjacent to the ortho substituent [2-

CO_Br: 1.819(4) Å]. This interpretation must be further sub-

stantiated in future work.

Figure 8. Solution IR spectra recorded during the photoconversion

of 2-CO_Brto 2-Br in MeCN. Photolysis conditions: ca. 100 mM,

100 mW/cm²white light, 298 K. Inset: solid-state IR spectrum of

2-CO_Br.

The kinetic results for the photolyses of 2-CO_Feto 5-

CO_Feare summarized in Figure 9 and Table 4. The aryl-

appended 5-CO_Feexhibits the fastest CO dissociation rate

(t_1/2= 8.5 min), whereas the methyl- and quinoline-substi-

tuted complexes exhibit less sensitivity to light (t_1/2= 12.0

and 12.7 min). The methoxy-substituted 3-CO_Feexhibits

the slowest transformation to the dibromide by a wide mar-

gin (t_1/2= 22.6 min).

These results are somewhat counterintuitive in light of

the structural parameters observed among the complexes.

Upon first inspection, the methyl-substituted 2-CO appears

to exert the greatest steric impingement upon the bound

CO ligand adjacent to the ortho substituent. The repulsion

effect between the methylpyridine and carbonyl moieties in

DFT Studies of Dicarbonyls and Photolysis Intermediates

To gain insight into the varying rates of CO photolysis,

DFT and time-dependent DFT (TD-DFT) calculations

were performed on the postulated photolysis intermediates.

Of particular interest was the interplay between the ligand

structure and the arrangement of the donors about the

metal center.

Ground-State Calculations

Optimized Geometries

The synthesized complexes 2-CO and 5-CO as well as the

2-CO_Bris reflected in the elongated Fe–N_pybond length in silico complex 1-CO (R = H) were evaluated by density

[2.039(3) Å] compared to the Fe–N_pybonds in the other functional theory. Geometry optimizations of the crystal-

dicarbonyl complexes (ca. 2.0–2.2 Å). However, this phe- structure coordinates of each of these monocations pro-

nomenon appears to be effectively compensated by the ceeded smoothly at the S = 0 configuration. In general, the

overall shorter Fe–C(O) bond length (adjacent to the meth- calculations afforded structures with bonding parameters

Eur. J. Inorg. Chem. 2015, 1675–1691

1683

www.eurjic.org

FULL PAPER

closely matched (less than Ϯ0.05 Å) to those experimentally ple, the calculated symmetric stretch of ν = 2050 cm^–1for

˜

found by X-ray analysis (see Table 5). However, some trends 2-CO is lower than the experimentally observed stretch at

and deviations are noted. For example, the calculated Fe– ν = 2069 cm^–1in the solid-state IR spectrum of 2-CO. In

˜

C(O)_SB(i.e., trans from the Schiff base N atom) bond another case, the calculated asymmetric stretch for 5-CO

lengths for 2-CO and 5-CO (1.779 and 1.790 Å) are both was 2005 cm^–1, whereas the experimentally observed lowest

lower than those observed in the X-ray structures [1.819(4) ν(CO) stretch was at ν = 2025 cm^–1[solid state, attenuated

˜

and 1.836(13) Å]. The same trend is mirrored in the Fe– total reflectance (ATR) IR]. This supports a high level of

C(O)_Brdistances of 2-CO and 5-CO, as evidenced by the confidence (considering the +20Ϯ3 cm^–1correction from

shorter calculated distances of 1.761 and 1.757 Å compared the DFT to experimental values)^[24]for the predicted IR

with the X-ray parameters [1.801(4) and 1.816(14) Å]; the stretches of the in silico complex 1-CO.

latter comparison is one of the greatest discrepancies

TD-DFT and UV/Vis Absorption Calculations

(0.059 Å) observed in the calculations. The metal–ligand

bonds (i.e., the Fe–N and Fe–S bond lengths) are generally

It is first noted that the experimentally observed UV/Vis

very well replicated (Ϯ0.02 A) in the DFT-calculated struc- spectra for 2-CO to 5-CO exhibit very little variance, which

tures of 2-CO and 5-CO. The chlorophenyl-appended com- is notable for the wide range of R groups [small and large,

plex 5-CO appears to provide the greatest differences be- electron-donating (ED) and electron-withdrawing (EW)]

tween the calculated and experimental structures. This dif- implemented in the present work. This clearly suggests that

ference appears to be due to an inaccurate representation the electronic environment that leads to the

of the steric effect of the π electron cloud of the phenyl d_π(Fe)Ǟπ*(CO) MLCT (vide infra) is not greatly affected

ring (it is underestimated) in the calculations. This is also by the ligand substituents. For example, 2-CO, 3-CO, and

indicated by a slightly decreased extent of linearity along 5-CO exhibit λ_maxin the near-UV at 425Ϯ5 nm. The only

the ortho-pyridine–chlorophenyl axis in the calculated exception to this trend is the quinoline-derived complex 4-

structure (C_o-py–C_Ar–C_Cl= 176.98°) with respect to that in CO, which exhibits a slightly redshifted λ_max= 435 nm. It

the X-ray structure (C_o-py–C_Ar–C_Cl= 178.14°). Overall, we is of note that 4-CO is neither on the high side nor the low

conclude that the DFT-optimized structures are representa- side of the dissociation rates for CO photolysis; therefore,

tive of the experiment results, and that the DFT calcula- this minor λ_maxshift is not correlated with any change in

tions generate an accurate structure for the putative com- the kinetics. Overall, the lack of substituent effects on the

plex 1-CO (with no ortho substituent, R = H).

MLCT band is reasonable for two reasons: (1) the neutral

overall charge of the NNS framework does not provide

much electron density for the substituents to modulate and

(2) the ortho placement of the substituent incurs direct in-

teraction of the substituent with the {Fe(CO)₂}²⁺moiety,

which likely modulates the electronic effect of the group

away from its “textbook” Hammet σ value.

Table 5. Comparisons of the experimental (2 and 5) and DFT-cal-

culated (1, 2, and 5) bond lengths [Å], bond angles [°], and IR

bands [cm^–1] for three NNS iron dicarbonyl complexes.

Complex

1-CO

DFT

2-CO_Br2-CO

X-ray DFT

5-CO_Fe5-CO

X-ray DFT

R₁, R₂

H, H

1.975

1.976

2.283

2.442

1.778

CH₃, H CH₃, H

2.039(3) 2.026

1.957(3) 1.976

2.2675(11)2.287

2.4618(8) 2.443

1.819(4) 1.779

ClPh, H ClPh, H

2.017(10) 2.011

1.956(10) 1.972

2.257(4) 2.285

2.443(2) 2.449

1.836(13) 1.790

The calculated UV/Vis spectra for 1-CO, 2-CO, and 5-

CO (Figure 10) continue this trend. The three calculated

λ_maxvalues (no correction applied) are within a narrow

range from λ = 386.43 to 388.31 nm. Although the absolute

values of the calculated absorption maxima are blueshifted

by 40 nm, the narrow range calculated between 2-CO and

5-CO is consistent with the experimental results. For 1-CO,

the orbitals implicated in the transition are from an admix-

ture transition of [d_π(Fe)|p_π(Br)]Ǟ[σ*(Fe–CO)_1,2|σ*(Fe–L)]

(Figure 11). Notably, the σ* orbitals of both Fe–CO units

(i.e., trans from Br^–and trans from N_SB) are implicated in

the destination orbital (Figure 11, right side). This is indica-

tive that the charge-transfer band is not localized to either

{σ*(Fe–CO)} fragment and, thus, that the sequential pho-

todissociation of CO must be instead governed by the struc-

tural dynamics of the excited state. It is also important to

note that the originating orbital contains substantial d_π-

(Fe)|π*(CO) backbonding, which is the primary metal–li-

Fe–N_py

Fe–N_SB

Fe–S_Me

Fe–Br

Fe–C(O)

trans SB

(Fe)C–O

trans SB

Fe–C(O)

trans Br

(Fe)C–O

trans Br

1.160

1.763

1.163

1.138(4) 1.161

1.801(4) 1.761

1.128(4) 1.164

1.115(16) 1.160

1.816(14) 1.757

1.108(16) 1.165

ca. 2.90 ca. 2.90

R₁···C(O)_SBca. 2.60 ca. 2.61 ca. 2.60

ν(CO)₁

2056

2069

2050

2070

2052

(sym)

2005

(sym)

2014

(asym)

(sym)

2009

(asym)

ν(CO)₂

2032

2025

(asym)

Hessian calculation and IR frequencies

To confirm a true energy minimum for each complex, gand interaction that must be disrupted in the photoexcited

Hessian calculations were performed. All of the optimized state. The analogous calculations for 2-CO and 5-CO

geometries for 1-CO, 2-CO, and 5-CO afforded Hessian yielded similar results (Figures S20–21) with varying extents

outcomes with no negative frequencies. The calculated of [σ*(Fe–CO)_1,2|σ*(Fe–L)] in the destination orbital. Over-

ν(CO) values for 2-CO and 5-CO are consistently all, it is clear that the absorption processes at λ ≈ 425 nm

20Ϯ3 cm^–1lower than the experimental values. For exam- are responsible for the CO photolysis.

Eur. J. Inorg. Chem. 2015, 1675–1691

1684

www.eurjic.org

FULL PAPER

Figure 10. Calculated TD-DFT absorption spectra of 1-CO (top), 2-CO (middle), and 5-CO (bottom).

Eur. J. Inorg. Chem. 2015, 1675–1691

1685

www.eurjic.org

FULL PAPER

Figure 11. Molecular orbitals of 1-CO involved in the [d_π(Fe)|p_π(Br)]Ǟ[σ*(Fe–CO)_1,2|σ*(Fe–L)] MLCT band postulated to be responsible

for CO photolysis, as identified by TD-DFT calculations.

Photolysis Intermediates

the remaining equatorial carbonyl ligand. Additionally, the

Fe–C(O) and (Fe)C–O bond lengths are only slightly modu-

One important remaining question regards the exact lated (by –0.004 and +0.002 Å, respectively) by the loss of

mechanism (or sequence) of CO dissociation that transpires the axial CO ligand.

during the photolysis of the dicarbonyl. To investigate this

By comparison, the removal of the equatorial CO ligand

open question, we have generated DFT-optimized struc- (trans to N_SB) results in a geometry-optimized complex (in-

tures of the monocarbonyl congeners of the complexes with termediate II₁) that is quite different to that for the simple

alternate CO ligands removed (i.e., either the CO ligand removal of the CO_eqligand. First, the loss of CO_eqis less

trans to the Br^–ion or that trans from N_SB). The results of thermodynamically unfavorable (only +22.1 kcal/mol) than

the calculations for the unsubstituted complex (1-CO) are the removal of CO_ax(+34.5 kcal/mol). Second, the remain-

summarized in Figure 12. The in silico photolysis was per- ing CO_axligand remains in essentially the same position,

formed, that is, the CO ligand was removed, and the energy although the Fe–C(O) bond is shortened by 0.06 Å. Inter-

of free CϵO was then compensated (intermediate I₁). The estingly, the formerly axial Br^–ligand shifts into the vacated

removal of the axial CO ligand (trans to Br^–) is, not surpris- pseudoequatorial position. Taken together, the two in silico

ingly, thermodynamically unfavorable by +34.5 kcal/mol. generated intermediates clearly indicate a preference for a

The DFT-optimized structure of the resulting five-coordi- five-coordinate monocarbonyl in which the three NNS do-

nate monocarbonyl species remains remarkably similar to nor atoms and an exogenous ligand (CO or Br^–) comprise

that of the dicarbonyl starting compound. This is indicated the equatorial ligation. Intermediate II₁is likely lower in

by the nearly identical angle of ca. 90° (prephotolysis: energy than I₁as the CO ligand is in the apical position,

83.00°; postphotolysis: 88.61°) between the Br^–ligand and which has been observed in numerous five-coordinate Fe^II

Figure 12. DFT geometry-optimized photolysis intermediates of 1-CO generated by the removal of the axial (top) or equatorial (bottom)

CO ligand. The energies were compensated for the removed CϵO unit.

Eur. J. Inorg. Chem. 2015, 1675–1691

1686

www.eurjic.org

FULL PAPER

Figure 13. DFT geometry-optimized photolysis intermediates of 5-CO generated by the removal of axial (top) or equatorial (bottom)

CO ligands. The stated energies were compensated for the removed CϵO unit.

and Mn^Imonocarbonyls. For 2-CO (see Supporting Infor-

mation, Figure S22), an analogous pattern in I₂and II₂is

Conclusions

We summarize our findings as follows:

observed for the bond-length changes (minimal), energy

(1) Metalations of neutral NNS ligands with Fe^IIcarb-

comparisons, and Br^–ligand migration upon CO_axor CO_eq

onyl starting salts at low temperatures generally afford the

removal.

desired cationic dicarbonyl species with one bromide-lig-

ated and one outer-sphere anion.

For chlorophenyl-containing 5-CO, slightly different re-

sults were obtained (Figure 13). Although the removal of

(2) The identity of the counteranion is determined by

the axial CO ligand (+34.1 kcal/mol, II₅) is still more unfa-

slight changes in the reaction conditions such as the choice

vorable than the loss of the equatorial CO ligand

of solvent (CH₂Cl₂/MeCN vs. Et₂O) or the steric demands

of the ligand (R = CH₃vs. Q).

(+25.5 kcal/mol, I₅), one of the resulting intermediates is

quite different (Figure 13). Chiefly, the removal of the equa-

(3) Generally, reactions performed at temperature above

torial CO ligand does not result in the migration of the

–20 °C result in the loss of all CO ligands and the formation

Br^–ligand to the equatorial position. Instead, the impinging

of the simple Fe^IIdibromide species.

steric bulk of the chlorophenyl unit apparently prevents the

(4) The resulting dicarbonyls are light-sensitive and exhi-

downward movement of the Br^–ligand. This occurrence is

bit photolysis t_1/2values (100 mW/cm²) that vary from min-

utes to tens of minutes depending on the concentration,

counterion, and solvent conditions.

attributed to the long Fe–Br bond [X-ray: 2.443(2) Å; DFT:

2.449 Å] and presumably the larger ionic radius of Br

(1.82 Å) than that of C/O (ca. 1 Å). Overall, the following

(5) The DFT studies have revealed that the bulky chloro-

are postulated: (1) The experimental photolysis leads to the

phenyl group destabilizes a putative photolysis intermediate

loss of the equatorial CO ligand and a driving force towards

(monocarbonyl I₅), a likely explanation for the rapid pho-

the migration of the Br^–ligand towards the equatorial vac-

tolysis of 5-CO relative to those of the other dicarbonyls.

ancy; this affords the intermediate with the lower energy (I₁

(6) Overall, this set of iron dicarbonyls may exhibit struc-

or I₅). Then, for 5-CO, (2) the large chlorophenyl unit in

tural similarities to a putative apo-active site of mono-iron

intermediate I₅prevents the relaxation of the Br^–ligand

hydrogenase (before the formation of the acyl unit) and will

into the equatorial position and, therefore, the destabiliza-

provide a useful template for future studies into the biogen-

tion of the structure leads to a faster loss of the second CO

esis that occurs in the native enzyme.

ligand. It may also be pointed out that the 5-CO intermedi-

ate I₅is the closest calculated structure to that of the actual

final product of the photolysis, that is, the dibromide spe-

cies 5-Br. Experimentally, the result is an expedited CO

Experimental Section

photolysis rate for 5-CO as compared to those for 2-CO, 3-

CO, and 4-CO.

General Procedures and Reagents: All organic starting materials

were purchased from Acros Organics or Sigma–Aldrich and used

Eur. J. Inorg. Chem. 2015, 1675–1691

1687

www.eurjic.org

FULL PAPER

without further purification. Most NNS ligands were prepared ac-

cording to procedures reported by us (_OMeNNS)^[25]or others

(_HNNS, _MeNNS, _QNNS).^[26–28]The Fe^IIstarting salt [Fe(CO)₄-

(Br)₂] was prepared by the reaction of [Fe(CO)₅] (Strem) with Br₂

according to the published procedure^[9]and purified by recrystalli-

zation from CH₂Cl₂at –20 °C instead of sublimation. HPLC grade

solvents were purchased from EMD, Fisher, Macron, or J.T. Baker

and dried with an alumina column system (Pure Process Technol-

ogy). Deuterated solvents (CDCl₃and CD₃CN) was purchased

from Cambridge Isotopes or Acros Organics and used as received.

CD₃CN): δ = 10.18 (s, 2 H), 8.58 (d, 2 H), 8.34 (d, 2 H), 8.02 (t, 2

H), 7.82 (m, 6 H), 7.75 (t, 2 H), 7.38 (t, 2 H), 2.14 (s, 6 H) ppm.

C₂₆H₂₄B₂F₈FeN₄S₂(686.08): calcd. C 45.52, H 3.53, N 8.17; found

C 45.29, H 3.77, N 8.27.

[(_MeNNS)Fe(CO)₂Br]Br·MeCN·H₂O (2-CO_Br·MeCN·H₂O): The

recrystallized starting salt [Fe(CO)₄Br₂] (0.150 mg, 0.457 mmol)

was added to dry CH₂Cl₂(3 mL) at –78 °C in a 10 mL Schlenk

flask and stirred for 5 min in the dark. The _MeNNS ligand (0.112 g,

0.457 mmol) was dissolved in dry CH₂Cl₂(3 mL) at –25 °C and

cooled to low temperature. The cold ligand solution was transfer-

red to the [Fe(CO)₄(Br)₂] solution at –78 °C by cannulation in the

dark. Stirring for 15 min in the dark afforded a dark red solution,

to which was added MeCN (0.5 mL); the solution was kept at

–25 °C overnight. The resulting dark orange precipitate was col-

lected by filtration and washed several times with pentane. The

product was recrystallized by the slow diffusion of Et₂O into the

MeCN solution at –25 °C, and red crystals formed, yield 70 mg

1-[6-(4-Chlorophenyl)pyridin-2-yl]-N-[2-(methylthio)phenyl]methan-

imine

(_ClPhNNS):

6-(4-Chlorophenyl)-2-pyridinecarbaldehyde

(0.500 g, 2.30 mmol) was dissolved in MeOH (10 mL). Upon the

addition of a solution of 2-methylthioaniline (0.574 g, 2.30 mmol)

in MeOH (10 mL), a bright yellow hazy solution formed. The mix-

ture was heated under reflux for 2 h, during which it turned clear.

Upon cooling to room temperature, a greenish yellow precipitate

was observed, and the solvent was evaporated under reduced pres-

sure. The resulting solid was washed with Et₂O (10 mL) to remove

excess 2-methylthioaniline; finally, a yellow powder was collected

(30%). Selected IR: ν = 3011 (w), 2069 (s), 2032 (s), 1597 (s), 1484

˜

(s), 1469 (m), 1445 (m), 1433 (m), 1387 (m), 1376 (m), 1318 (m),

1278 (m), 1239 (m), 1174 (m), 796 (s), 770 (w), 676 (m br), 428 (m)

cm^–1. ¹H NMR (500 MHz, CD₃CN, –35 °C): δ = 9.69 (s, 1 H), 8.29

(d, 1 H), 8.07 (d, 2 H), 7.73 (m, 4 H), 3.01 (s, 3 H, _pyCH₃), 2.68 (s,

3 H, SCH₃) ppm; note: this spectrum was obtained in the presence

of 1 equiv. of NaClO₄to improve the solubility of the bromide salt

(Figure S7), for the analogous spectrum in the absence of NaClO₄,

see Figure S6.

by filtration, yield 0.64 g (63%). Selected IR: ν = 1592 (m), 1570

˜

(m), 1495 (m), 1426 (m), 1317 (w), 1163 (m), 1033 (s), 930 (m), 804

1

(vs), 745 (vs), 677 (w), 663 (w) cm^–1. H NMR (CDCl₃): δ = 8.67

(s, 1 H), 8.29 (d, 1 H), 8.01 (d, 2 H), 7.89 (m, 3 H), 7.79 (d, 1 H),

7.48 (m, 2 H), 7.35 (br, 1 H), 7.12 (br, 1 H), 6.73 (1 H), 2.47 (s, 3

1

H) ppm. The H NMR spectrum revealed a portion (ca. 20%) of

unreacted aldehyde, which was not separated by chromatography

or fractional crystallization; the ligand was thus used as prepared.

[(_MeNNS)Fe(CO)₂Br][Fe(CO)₃(Br)₃] (2-CO_Fe): The _MeNNS ligand

(50 mg, 0.207 mmol) in Et₂O (5 mL) was added dropwise to a clear

solution of [Fe(CO)₄(Br)₂] (67 mg, 0.207 mmol) in Et₂O (5 mL). An

immediate color change to orange-green was observed along with

the precipitation of the crude product. This mixture was stirred for

another 30 min, after which the crude product was collected by

filtration under an ambient atmosphere. The product was recrys-

tallized by the vapor diffusion of Et₂O into a MeCN solution of

the complex at –20 °C to afford orange-red crystals suitable for X-

ray diffraction studies (connectivity), yield 50 mg (30%). Selected

N-[1-(6-Methylpyridin-2-yl)ethylidene]-2-(methylthio)aniline

(

_MeN_MeNS): 2-Methylthioaniline (0.500 g, 3.69 mmol) and 2-

acetyl-6-methylpyridine (0.515 g, 3.69 mmol) were dissolved sepa-

rately in a minimal amount of acetic acid. The ketone was added

to a round-bottomed flask, followed by the aniline. A solution of

zinc chloride (0.540 g, 3.69 mmol) in acetic acid was added to the

mixture and stirred. The clear, yellow-orange mixture was heated

to reflux for 2 h with no color change. As the clear mixture cooled

to room temperature, it became darker brown-orange, and a yel-

low-orange precipitate formed. The mixture was filtered, and the

precipitate was washed with Et₂O. The precipitate was then dis-

solved in CH₂Cl₂and extracted twice with potassium oxalate solu-

tion (to remove Zn²⁺ions) and then twice more with distilled water.

The organic layer was collected and dried with sodium sulfate, and

the solvent was evaporated under reduced pressure to afford a gold-

IR: ν = 2092 (m), 2075 (s), 2033 (vs), 1411 (w), 1237 (w), 798 (m),

˜

773 (m), 621 (vs), 569 (vs), 511 (s), 442 (w) cm^–1. ¹H NMR

(500 MHz, CD₃CN, –25 °C): δ = 9.68 (s, 1 H), 8.30 (d, 1 H), 8.09

(m, 2 H), 7.92 (d, 1 H), 7.82 (t, 1 H), 7.77 (t, 1 H), 7.67 (d, 1 H),

3.03 (s, 3 H, _pyCH₃), 2.64 (s, 3 H, SCH₃) ppm (Figure 6).

[(_MeNNS)Fe(Br)₂] (2-Br): Method A:

A solution of _MeNNS

(0.192 g, 0.0793 mmol) in dry toluene (10 mL) was added to a red

solution of [Fe(CO)₄(Br)₂] (0.260 mg, 0.0793 mmol) in dry toluene

(10 mL) at room temperature. The solution quickly turned green,

accompanied by vigorous gas evolution (CO gas). The reaction

mixture was stirred for 30 min, and the precipitated green solid

was collected by filtration under a N₂atmosphere and dried under

vacuum to give a green powder, yield 0.150 g (40%). Suitable crys-

tals for X-ray diffraction were obtained by the vapor diffusion of

Et₂O into a MeCN solution of the complex. The analytical data

were consistent with those of the complex prepared by Method B.

yellow oil, yield 712 mg (75%). Selected IR: ν = 1639 (s), 1574 (s),

˜

1460 (s), 1257 (w), 1086 (m), 1070 (m), 1036 (m), 966 (w), 733 (s)

1

cm^–1. H NMR (CDCl₃): δ = 8.17 (d, 1 H), 7.61 (t, 1 H), 7.23 (m,

2 H) 7.12 (m, 2 H), 6.71 (d, 1 H), 2.61 (s, 3 H), 2.42 (s, 3 H), 2.34

(s, 3 H) ppm.

[(_HNNS)₂Fe][BF₄]₂(1-L2): At room temperature, to a stirred solu-

tion of 1 equiv. of _HNNS (100 mg, 0.438 mmol) in MeCN (10 mL)

was added THF (2 mL) containing [Fe(CO)₄(Br)₂] (118 mg,

0.438 mmol, 1 equiv.). The addition of iron(II) carbonyl immedi-

ately generated a dark violet color, which was indefinitely stable

under ambient conditions. To isolate the BF₄salt of the complex,

AgBF₄(85 mg, 0.438 mmol, 1 equiv.) in THF (2 mL) was added;

although no color change was observed, a turbid solution was gen-

erated (AgBr), and the resulting mixture was stirred overnight and

then filtered twice through Celite. The vapor diffusion of Et₂O at

room temperature afforded violet blocks suitable for X-ray diffrac-

Method B: The complex was also prepared directly from FeBr₂

in THF. Under an inert atmosphere, the _MeNNS ligand (100 mg,

0.413 mmol) was dissolved in THF (10 mL), and to this was added

THF (10 mL) containing FeBr₂(80.0 mg, 0.371 mmol). Upon the

addition of the metal salt, the solution turned dark green, and a

brown precipitate formed. After stirring overnight, the THF was

removed in vacuo to afford a brown solid, which was partially re-

dissolved in CH₂Cl₂(50 mL). This mixture was filtered, and the

tion, yield 65 mg (21%). Selected IR bands: ν = 1604 (w), 1585

˜

(w), 1477 (m), 1306 (w), 1162 (w), 1029 (vs), 916 (m), 775 (s), 765 dark green CH₂Cl₂solution was subjected to the vapor diffusion

1

(s), 750 (s), 606 (m), 516 (s), 417 (w) cm^–1. H NMR (500 MHz, in

of Et₂O at –15 °C to afford single crystals as brown blocks, yield

Eur. J. Inorg. Chem. 2015, 1675–1691

1688

www.eurjic.org

FULL PAPER

39 mg (23%). Selected IR: ν = 3011 (w), 1597 (s), 1484 (s), 1469

(w), 1586 (w), 1508 (m), 1455 (w), 1431 (w), 1415 (w), 1378 (w),

˜

(m), 1445 (m), 1433 (m), 1387 (m), 1376 (m), 1318 (m), 1278 (m), 1142 (w), 981 (m), 969 (m), 834 (s), 791 (m), 750 (s), 491 (m) cm^–1.

1239 (m), 1174 (m), 796 (s), 770 (w), 676 (m br), 428 (m) cm^–1. Solid-state magnetic susceptibility (298 K): μ_eff= 5.21 μ_B. UV/Vis

Solid-state magnetic susceptibility (298 K): μ_eff= 5.48 μ_B. UV/Vis

(MeCN): λ_max= 460 nm, ε = 1 180 m^–1cm^–1. C₁₇H₁₄Br₂FeN₂S

(MeCN): λ_max= 460 nm, ε = 650 m^–1cm^–1. C₁₄H₁₄Br₂FeN₂S (494.03): calcd. C 41.33, H 2.86, N 5.67; found C 40.59, H 2.94, N

(457.99): calcd. C 36.71, H 3.08, N 6.12; found C 36.70, H 3.13, N

6.11.

5.44.

[(_ClPhNNS)Fe(CO)₂Br][Fe(CO)₃(Br)₃] (5-CO_Fe): The _ClPhNNS iron

carbonyl complex salt was prepared according to the procedure for

2-CO_Fewith _ClPhNNS ligand (50 mg, 0.148 mmol) and iron tetra-

carbonyl bromide (48 mg, 0.148 mmol). The precipitate was an

orange powder, which showed an IR spectrum consistent with the

formulation of the complex salt [(_ClPhNNS)Fe(CO)₂Br][Fe(CO)₃-

[(_OMeNNS)₂Fe][BF₄]₂(3-L2): To a stirred MeCN solution (5 mL) of

_OMeNNS ligand (50 mg, 0.194 mmol, 1 equiv.) was added a MeCN

solution (5 mL) of [Fe(H₂O)₆][BF₄]₂(65 mg, 0.194 mmol, 1 equiv.)

at room temperature. Over the course of 5 min, the solution

changed from bright orange to blue-violet. The vapor diffusion of

Et₂O into the MeCN solution at room temperature afforded long

blue-violet needles of the product, suitable for X-ray diffraction

(Br) ]. Selected IR: ν = 2095 (m), 2067 (m), 2037 (s), 2025 (vs),

˜

3

1600 (w), 1254 (w), 1095 (m), 966 (w), 760 (m), 621 (s), 581 (s), 437

(w) cm^–1. Samples of the complex as both the [Fe(CO)₃(Br)₃]^–salt

(main product) and [Fe(Br)₄]^–salt (minor product) were obtained

upon recrystallization by the vapor diffusion of Et₂O into an

MeCN solution of the complex at –20 °C over an extended period

(1–2 weeks); this afforded very small, thermally unstable and ex-

tremely light-sensitive orange crystals suitable for synchrotron X-

(connectivity), yield 34 mg (23%). Selected IR: ν = 1607 (m), 1586

˜

(w), 1483 (s), 1283 (s), 1045 (vs), 1030 (vs), 954 (s), 794 (s), 764 (s),

1

731 (s), 618 (w), 519 (w) cm^–1. H NMR (400 MHz) in CD₃CN: δ

= 11.10 (br s, 1 H), 8.65 (d, 1 H), 8.22 (d, 2 H), 8.11 (t, 1 H), 7.93

(m, 2 H), 7.74 (t, 1 H), 7.40 (br d, 1 H) ppm. C₂₈H₂₈B₂F₈FeN₄O₂S₂

(746.14): calcd. C 45.07, H 3.78, N 7.51; found C 45.85, H 4.07, N

7.45.

ray diffraction, yield 43 mg (48%). Selected IR: ν = 2092 (m), 2070

˜

(m), 2033 (s), 2025 (s), 1614 (w), 1253 (w), 1094 (m), 967 (m), 768

[(_OMeNNS)Fe(CO)₂Br][Fe(CO)₃(Br)₃] (3-CO_Fe): The _OMeNNS iron

carbonyl complex salt was prepared according to the procedure for

2-CO_Fewith _OMeNNS ligand (50 mg, 0.194 mmol) and iron tetra-

carbonyl bromide (63 mg, 0.1935 mmol). The crude precipitate was

a green-orange powder. The pure complex was isolated upon

recrystallization by the vapor diffusion of Et₂O into a MeCN solu-

tion of the complex at –20 °C, which afforded orange-red rod-

shaped crystals suitable for single-crystal X-ray diffraction, yield

1

(m), 619 (m), 580 (s), 439 (w) cm^–1. H NMR (500 MHz, CD₃CN,

5 °C): δ = 9.79 (s, 1 H), 8.34 (m, 2 H), 8.27 (br, 1 H), 7.86 (m, 2

H), 7.75 (m, 4 H), 7.58 (d, 1 H), 7.53 (d, 1 H), 2.54 (s 3 H) ppm

(Figure S10).

[(_ClPhNNS)Fe(Br)₂] (5_ClPh-Br): Method A: A solution of _ClPhNNS

(50 mg, 0.148 mmol) in dry CH₂Cl₂(3 mL) under a N₂atmosphere

was transferred by cannulation into a solution of [Fe(CO)₄(Br)₄]

(48 mg, 0.148 mmol) in dry CH₂Cl₂and held at –78 °C. The reac-

tion mixture changed from red to orange upon the addition of the

ligand. Next, a small volume of MeCN (3 mL) was added by can-

nulation, and the mixture changed to red-orange. The resulting

mixture was then stored at –20 °C for several days to afford a small

amount of green crystals suitable for X-ray diffraction. The analyti-

cal data obtained were consistent with those of the sample prepared

by Method B.

50 mg (31%). Selected IR: ν = 2095 (m), 2081 (m), 2032 (vs), 1593

˜

(w), 1561 (w), 1483 (m), 1243 (w), 1192 (w), 797 (m), 572 (s), 518

1

(m) cm^–1. H NMR (500 MHz, CD₃CN, –5 °C): δ = 9.66 (s, 1 H),

8.32 (br, 1 H), 8.21 (br, 1 H), 7.89 (br, 2 H), 7.82 (br, 1 H), 7.75 (t,

1 H), 7.31 (d, 1 H), 4.22 (s, 3 H), 2.64 (s, 3 H) ppm (Figure S8).

[(_QNNS)Fe(CO)₂Br][Fe(CO)₃(Br)₃] (4-CO_Fe): Surprisingly, this

complex salt was prepared by same procedure (CH₂Cl₂, –78 °C)

that afforded the bromide salts 2-CO_Brand 3-CO_Br(see procedure

above) with _QNNS ligand (127 mg, 0.458 mmol) and iron tetracarb-

onyl dibromide (150 mg, 0.458 mmol). A small portion of Et₂O

(3 mL) was layered on the reaction mixture, and storage of the

solution at –25 °C afforded X-ray quality crystals of the complex

Method B: This complex was also prepared from FeBr₂in THF

according to the procedure for 2-Br with _ClPhNNS (100 mg,

0.295 mmol) in THF (10 mL) and FeBr₂(57 mg, 0.264 mmol) in

THF (10 mL); metalation afforded a light green solution. Upon

evaporation of the THF, dissolution in CH₂Cl₂(50 mL) afforded a

green solution that was subjected to the vapor diffusion of Et₂O at

–15 °C to afford X-ray quality crystals (green needles) of the prod-

salt, yield 163 mg (42%). Selected IR: ν = 2075 (s), 2035 (s), 2018

˜

(s), 1606 (w), 1586 (w), 1338 (w), 1022 (w), 822 (m), 750 (m), 616

1

(w) cm^–1. H NMR (500 MHz, CD₃CN, –35 °C): δ = 9.98 (s, 1 H),

8.85 (d, 1 H), 8.68 (d, 1 H), 8.39 (m, 2 H), 8.25 (m, 2 H), 8.01 (m,

2 H), 7.86 (m, 2 H), 2.78 (s, 3 H) ppm (Figure S9).

uct, yield 70 mg (47%). Selected IR: ν = 3035 (m), 1620 (m), 1598

˜

(m), 1589 (m), 1092 (m), 832 (s), 811 (s), 767 (s), 522 (m), 475 (m)

[(_QNNS)Fe(Br)₂] (4-Br): Method A: The dibromide complex of

_QNNS was prepared according to the procedure above, except that

the resulting CH₂Cl₂solution was brought to room temperature

and prepared by vapor diffusion with Et₂O. The solution was then

cooled to –25 °C and stored at that temperature for several days.

This process afforded green X-ray quality crystals of the product

in low yield. The analytical data were consistent with those of the

complex prepared by Method B.

cm^–1. Solid-state magnetic susceptibility (298 K): μ_eff= 5.24 μ_B.

UV/Vis (MeCN): λ_max

=

460 nm,

ε

=

1

570 m^–1cm^–1.

C₁₉H₁₅Br₂ClFeN₂S (554.51): calcd. C 41.16, H 2.73, N 5.05; found

C 40.75, H 2.69, N 4.89.

[(_MeN_MeNS)Fe(Br)₂] (6-Br): The complex was prepared from FeBr₂

in THF according to the procedure for 2-Br with _MeN_MeNS

(200 mg, 0.780 mmol) in THF (10 mL) and FeBr₂(151 mg,

0.700 mmol) in THF (10 mL); metalation afforded a dark green

solution. After evaporation of the THF, the residue was dissolved

in CH₂Cl₂(30 mL) to afford a green solution, which was subjected

to the vapor diffusion of pentane at ambient temperature to afford

X-ray-quality green-blue crystals, yield 261 mg (79%). Selected IR:

Method B: The complex was also prepared from FeBr₂in THF

according to the procedure for 2-Br with _QNNS (100 mg,

0.359 mmol) in THF (10 mL) and FeBr₂(70.0 mg, 0.325 mmol) in

THF (10 mL); metalation afforded a dark green solution. After the

evaporation of the THF, the residue was dissolved in CH₂Cl₂

(70 mL) to afford a green solution, which was subjected to the

vapor diffusion of Et₂O at ambient temperature to afford X-ray

ν = 1589 (s), 1465 (m), 1359 (m), 1007 (w), 968 (w), 799 (s), 773 (s),

˜

756 (s), 491 (w), 464 (w) cm^–1. Solid-state magnetic susceptibility

quality green needles, yield 68 mg (42%). Selected IR: ν = 1611

(298 K): μ_eff= 5.52 μ_B. UV/Vis (MeCN): λ_max= 460 nm, ε =

˜

Eur. J. Inorg. Chem. 2015, 1675–1691

1689

www.eurjic.org

FULL PAPER

780 m^–1cm^–1. C₁₅H₁₆Br₂FeN₂S (472.02): calcd. C 38.17, H 3.42, N

5.93; found C 38.28, H 3.39, N 5.91.

31G(d) basis set obtained from the EMSL Basis Set Exchange^[33,34]

with diffuse functions on all heteroatoms and the ligating carbonyl

carbon atoms. The modified basis set (m6-31G) of Mitin et al. was

used for iron,^[35]and the revised basis set (r6-311G) of McGrath

was used for bromine.^[36]TD-DFT and Hessian calculations were

performed at the same level of theory as the optimizations. All

Hessian calculations showed no imaginary frequencies, and rota-

tions/translations showed energies of less than 5 cm^–1. Graphical

manipulations were performed with ChemCraft.^[37]

1

Physical Measurements: The H NMR spectra were collected with

a Varian DirecDrive 400 MHz spectrometer, and chemical shifts

were referenced to CD₃CN. The UV/Vis absorption spectra were

obtained at 298 K with an Ocean Optics fiber optic system

equipped with a low-intensity PX-2 pulsed xenon lamp and a

USB2000-XR1-ES detector. Quantitative UV/Vis studies (ε values,

kinetic stability) were obtained with ca. 0.1 mm solutions of the

complexes in MeCN in 1 cm quartz cuvettes. Kinetic plots (photol-

ysis of CO in MeCN) were analyzed in SigmaPlot 8.0 and fitted to

the exponential equation y = y₀+ a(1 – e^–bx). The solid-state infra-

red spectra were recorded with a Bruker Alpha spectrometer

equipped with a diamond ATR crystal. The solution IR data dur-

ing the photolysis of 2-CO_Brwere obtained with the same instru-

ment equipped with a 0.1 cm path length cell with CaF₂windows;

the concentration was ca. 100 mm in MeCN; the spectra were plot-

ted in absorbance mode to maintain linearity with concentration.

Photolysis was achieved by the use of a Newport solar simulator

equipped with a 150 W xenon lamp and an AM1.5 filter; the sam-

ples were placed ca. 10 cm from the light source to achieve 1 Sun

illumination (100 mW/cm²) with broadband white light. The solid-

state magnetic susceptibilities were measured at 298 K with a John-

son Matthey Mark I instrument.

Supporting Information (see footnote on the first page of this arti-

cle): connectivity structures of alternative salts of the dicarbonyls,

¹H NMR spectra of the dicarbonyl complexes, kinetic traces of

photolyses, solid-state IR spectra of the carbonyls, and additional

schemes regarding the TD-DFT-calculated photolysis pathways.

Acknowledgments

The authors gratefully acknowledge Steve Sorey and Angela Span-

genberg for technical assistance in acquiring low-temperature

NMR spectra. For crystal data collection for 5-CO_FeBr, we thank

Professor Allen Oliver (Notre Dame) and Professor Jeanette

Krause (University of Cincinnati) for data collection through the

SCrALS (Service Crystallography at Advanced Light Source) pro-

gram at the Small Crystal Crystallography Beamline 11.3.1 at the

Advanced Light Source (ALS), Lawrence Berkeley National Labo-

ratory, which is supported by the U.S. Department of Energy, Of-

fice of Energy Sciences Materials Sciences Division (DE-AC02-

05CH11231). This research was supported by startup funds to

M. J. R. provided by the UT Austin College of Natural Sciences,

the Robert A. Welch Foundation (F-1822), and the ACS Petroleum

Research Fund (53542-DN13). DFT calculations were performed

with a computer purchased with funds from a Ralph Powe Junior

Faculty Award to M. J. R. Support from the National Science

Foundation (NSF) (NSF-REU program at UT Austin, grant

number CHE-1003947) to C. J. is acknowledged.

X-ray Diffraction Data Collection and Crystal-Structure Refine-

ment: The data were collected with a Rigaku AFC12 diffractometer

with a Saturn 724+ CCD or a Nonius Kappa CCD diffractometer

with a Bruker AXS Apex II detector, both with a graphite mono-

chromator with Mo-K_αradiation. Reduced temperatures were

maintained with an Oxford Cryostream low-temperature device.

The data reductions were performed with Rigaku Crystal Clear

version 1.40.2. The structures were solved by direct methods with

SIR973 and refined by full-matrix least-squares techniques on F²

with anisotropic displacement parameters for the non-H atoms

with SHELXL-97.4. The structure analysis was aided by the use of

the programs PLATON985 and WinGX.6. The crystallographic

data for 5-CO_FeBrwere collected through the SCrALS (Service

Crystallography at Advanced Light Source) program at the Small-

Crystal Crystallography Beamline 11.3.1 at the Advanced Light

Source (ALS), Lawrence Berkeley National Laboratory. The

hydrogen atoms attached to carbon atoms were calculated in ideal

positions with isotropic displacement parameters set to 1.2U_eqof

the attached atom (1.5U_eqfor methyl hydrogen atoms). The data

were checked for secondary extinction effects, but no corrections

were necessary. The neutral atom scattering factors and the values

used to calculate the linear absorption coefficient are from the In-

ternational Tables for X-ray Crystallography.

[1] S. Shima, R. K. Thauer, Chem. Rec. 2007, 7, 37.

[2] J. A. Wright, P. J. Turrell, C. J. Pickett, Organometallics 2010,

29, 6146.

[3] S. Shima, O. Pilak, S. Vogt, M. Schick, M. S. Stagni, W. Meyer-

Klaucke, E. Warkentin, R. K. Thauer, U. Ermler, Science 2008,

321, 25.

[4] T. Hiromoto, K. Ataka, O. Pilak, S. Vogt, M. S. Stagni, W.

Meyer-Klaucke, E. Warkentin, R. K. Thauer, S. Shima, U.

Ermler, FEBS Lett. 2009, 583, 585.

[5] M. Schick, X. Xie, K. Ataka, J. Kahnt, U. Linne, S. Shima, J.

Am. Chem. Soc. 2012, 134, 3271.

[6] D. Chen, R. Scopelliti, X. Hu, Angew. Chem. Int. Ed. 2010, 49,

7512; Angew. Chem. 2010, 122, 7674.

[7] L.-C. Song, Z.-J. Xie, M.-M. Wang, G.-Y. Zhao, H.-B. Song,

Inorg. Chem. 2012, 51, 7466.

[8] B. Hu, D. Chen, X. Hu, Chem. Eur. J. 2014, 20, 1.

[9] P. J. Turrell, J. A. Wright, J. N. T. Peck, V. S. Oganesyan, C. J.

Pickett, Angew. Chem. Int. Ed. 2010, 49, 7508.

[10] D. Chen, R. Scopelliti, X. Hu, J. Am. Chem. Soc. 2010, 132,

928.

[11] X. Wang, Z. Li, X. Zeng, Q. Luo, D. J. Evans, C. J. Pickett, X.

Liu, Chem. Commun. 2008, 3555.

CCDC-1046982 (for 2-CO_Br), -1046983 (for 3-CO_Fe), -1046984 (for

5-CO_Fe), -1046985 (for 2-Br), -1046986 (for 4-Br), -1046987 (for

5-Br), -1046988 (for 6-Br), and -1046989 (for 1-L2) contain the

supplementary crystallographic data for this paper. These data can

be obtained free of charge from The Cambridge Crystallographic

Data Centre via www.ccdc.cam.ac.uk/data_request/cif.

DFT Calculations: Dibromide Complexes: Orbital calculations,

spin configuration energies, and geometry optimizations of the

crystal structures of the iron(II) dibromide complexes 2-Br and 3-

Br to 6-Br were performed with the Firefly software package^[29]

with either a combination of the 6-31G* basis set and the pure

functional PW91^[30,31]or a TZV/B3PW91 basis set. The orbitals

were visualized with MacMolPlt.^[32]Dicarbonyl Complexes: All

geometries were optimized with the PW91 functional with Firefly

8.0.0.^[29]We used a hybrid basis set consisting of the Pople 6-

[12] T. Liu, B. Li, C. V. Popescu, A. Bilko, L. M. Perez, M. B. Hall,

M. Y. Darensbourg, Chem. Eur. J. 2010, 16, 3083.

[13] J. M. Kuchenreuther, W. K. Myers, D. L. M. Suess, T. A. Stich,

V. Pelmenschikov, S. A. Shiigi, S. P. Cramer, J. R. Swartz, R. D.

Britt, S. J. George, Science 2014, 343, 424.

[14] S. E. McGlynn, D. W. Mulder, E. M. Shepard, J. B. Broderick,

J. W. Peters, Dalton Trans. 2009, 4274.

Eur. J. Inorg. Chem. 2015, 1675–1691

1690

www.eurjic.org

FULL PAPER

[15]

[16]

[25] S. E. A. Lumsden, G. Durgraprasad, K. A. M. Thomas, M. J.

Rose, Dalton Trans. 2014, 43, 10725–10738.

J. M. Kuchenreuther, W. K. Myers, T. A. Stich, S. J. George, Y.

NejatyJahromy, J. R. Swartz, R. D. Britt, Science 2013, 342,

472.

G. Berggren, A. Adamska, C. Lambertz, T. R. Simmons, J. Es-

selborn, M. Atta, S. Gambarelli, J. M. Mouesca, E. Reijerse,

W. Lubitz, T. Happe, V. Artero, M. Fontecave, Nature 2013,

499, 66.

V. P. L. Velásquez, T. M. A. Jazzazi, A. Malassa, H. Görls, G.

Gessner, S. H. Heinemann, M. Westerhausen, Eur. J. Inorg.

Chem. 2012, 1072.

W.-F. Liaw, N.-H. Lee, C.-H. Chen, C.-M. Lee, G.-H. Lee, S.-

M. Peng, J. Am. Chem. Soc. 2000, 122, 488.

W.-F. Liaw, J.-H. Lee, H.-B. Gau, C.-H. Chen, S.-J. Jung, C.-

H. Hung, W.-Y. Chen, C.-H. Hu, G.-H. Lee, J. Am. Chem. Soc.

2002, 124, 1680.

[26] S. K. Chatterjee, S. Roy, S. K. Barman, R. C. Maji, M. M.

Olmstead, A. K. Patra, Inorg. Chem. 2012, 51, 7625.

[27] S. Roy, P. Mitra, A. K. Patra, Inorg. Chim. Acta 2011, 370, 247.

[28] M. A. Gonzalez, S. J. Carrington, N. L. Fry, J. L. Martinez,

P. K. Mascharak, Inorg. Chem. 2012, 51, 11930.

[29] A. Granovsky, Firefly, v. 7.1.G: http://classic.chem.msu.su/

gran/firefly/index.html.

[30] J. P. Perdew, J. A. Chevary, S. H. Vosko, K. A. Jackson, M. R.

Pederson, D. J. Singh, C. Fiolhais, Phys. Rev. B 1993, 48, 4978.

[31] J. P. Perdew, J. A. Chevary, S. H. Vosko, K. A. Jackson, M. R.

Pederson, D. J. Singh, C. Fiolhais, Phys. Rev. B 1992, 46, 6671.

[32] B. M. Bode, M. S. Gordon, J. Mol. Graphics Modell. 1998, 16,

133.

[17]

[18]

[19]

[33] D. Feller, J. Comput. Chem. 1996, 17, 1571–1586.

[34] K. L. Schuchardt, B. T. Didier, T. Elsethagen, L. Sun, V. Guru-

moorthi, J. Chase, J. Li, T. L. Windus, J. Chem. Inf. Model.

2007, 47, 1045–1052.

[20] W.-F. Liaw, C.-H. Chen, G.-H. Lee, S.-M. Peng, Organometal-

lics 1998, 17, 2370.

[21] W.-F. Liaw, J.-H. Lee, H.-B. Gau, C.-H. Chen, G.-H. Lee, Inorg.

Chim. Acta 2001, 322, 99.

[22] W.-S. Hong, C.-Y. Wu, C.-S. Lee, W.-S. Hwang, M.-Y. Chiang,

J. Organomet. Chem. 2004, 689, 277.

[23] B. Li, T. Liu, C. V. Popescu, A. Bilko, M. Y. Darensbourg, In-

org. Chem. 2009, 48, 11283.

[24] V. Jonas, W. Thiel, J. Chem. Phys. 1995, 102, 8474–8484.

[35] A. V. Mitin, J. Baker, P. J. Pulay, J. Chem. Phys. 2003, 118,

7775–7782.

[36] M. P. McGrath, J. Chem. Phys. 2009, 130, 176101.

[37] G. A. Zhurko, Chemcraft: http://ww w .chemcraftprog.com.

Received: October 22, 2014

Published Online: February 2, 2015

Eur. J. Inorg. Chem. 2015, 1675–1691

1691

Article Doi

Mononuclear Iron(II) Dicarbonyls Derived from NNS Ligands - Structural Models Related to a Pre-Acyl Active Site of Mono-Iron (Hmd) Hydrogenase

DOI: 10.1002/ejic.201403013

Source and publish data:

Authors:

Article abstract of DOI:10.1002/ejic.201403013

Full text of DOI:10.1002/ejic.201403013

Products guided by the article

R&D Labs maybe for 1384182-24-4

Relevant to this article

Hot Product