19F NMR study of ligand dynamics in carboxylate-bridged dIIron(II) complexes supported by a macrocyclic ligand

Article abstract of DOI:10.1039/c5dt02138c

A series of asymmetrically carboxylate-bridged diiron(ii) complexes featuring fluorine atoms as NMR spectroscopic probes, [Fe₂(PIM)(Ar^4F-PhCO₂)₂] (10), [Fe₂(F₂PIM)(Ar^TolCO_2<

Full text of DOI:10.1039/c5dt02138c

Dalton
Transactions
View Article Online
View Journal
PAPER
19F NMR study of ligand dynamics in carboxylate-
bridged diiron(II) complexes supported by a
macrocyclic ligand†
Cite this: DOI: 10.1039/c5dt02138c
Mikael A. Minier and Stephen J. Lippard*
A series of asymmetrically carboxylate-bridged diiron(II) complexes featuring fluorine atoms as NMR
spectroscopic probes, [Fe2(PIM)(Ar4F-PhCO2)2] (10), [Fe2(F2PIM)(ArTolCO2)2] (11), and [Fe2(F2PIM)-
(Ar4F-PhCO2)2] (12), were prepared and characterized by X-ray crystallography, Mössbauer spectroscopy,
and VT 19F NMR spectroscopy. These complexes are part of a rare family of syn N-donor diiron(II) com-
pounds, [Fe2(X2PIM)(RCO2)2], that are structurally very similar to the active site of the hydroxylase enzyme
component of reduced methane monooxygenase (MMOHred). Solution characterization of these com-
plexes demonstrates that they undergo intramolecular carboxylate rearrangements, or carboxylate shifts,
Received 5th June 2015,
Accepted 26th August 2015
DOI: 10.1039/c5dt02138c
a
dynamic feature relevant to the reactivity of the diiron centers in bacterial multicomponent
www.rsc.org/dalton
monooxygenases.
Reliable strategies involve the use of sterically demanding
ligands such as m-terphenyl carboxylates or dinucleating
Introduction
Biology utilizes enzymes with carboxylate-bridged diiron ligand frameworks. A variety of binding modes are possible for
centers to catalyze difficult transformations such as methane carboxylate-bridged diiron complexes. In the solid state struc-
oxidation by methane monooxygenase (MMO).1–4 These diiron ture of [LFe2(OAc3)]·2MeCN, where L is a bis-tetradentate pyr-
centers typically have an asymmetric arrangement of ligands azolate, all carboxylate binding modes in a dinuclear system
that is conserved across the superfamily of bacterial multi- were observed within two molecules in the asymmetric unit
component monooxygenases (BMMs) (Fig. 1).5–8 In general, two in the crystals.14 Moreover, the solution structures of these
nitrogen donors from histidine bind to the diiron unit in a syn complexes may differ from those in the solid or may switch
fashion with respect to the diiron vector. The bridging carboxy- between multiple conformations. For example, 19F NMR
lates adopt different coordination modes. In the reduced state spectroscopic studies of [Fe2(Ar4F-PhCO2)4(THF)2] (Ar4F-Ph = 2,6-
of the MMO hydroxylase component (MMOHred), one carboxy- bis-4-fluorophenylphenyl) revealed an equilibrium between
late bridges in a symmetric μ–η1:η1 fashion and the other in an doubly and quadruply bridged forms, with conversion to the
asymmetric μ–η1:η2 mode. In the oxidized form of the enzyme, quadruply bridged form below −60 °C.15
MMOHox, the latter carboxylate shifts into a monodentate
Previously, our lab reported two diiron(II) complexes,
terminal position. This alteration in the carboxylate bridging [Fe2(PIM)(Ph3CCO2)2] (1) and [Fe2(PIM)(ArTolCO2)2] (2) (ArTol
=
mode, or carboxylate shift,9 is proposed to be mechanistically 2,6-bis-tolylphenyl), supported by the dinucleating macrocyclic
important based on both biological4 and synthetic model ligand, H2PIM (Scheme 2, inset).16 Together with bulky carb-
studies10,11 (Scheme 1).
oxylate ligands, the PIM2− macrocycle enforces geometries in
Efforts to replicate the chemistry of MMOH using small diiron(II) complexes featuring syn N-coordination and asym-
molecules have been reviewed.12,13 The rational synthesis of metric carboxylate bridging modes, closely resembling that in
carboxylate-bridged dinuclear metal complexes is challenging, MMOHred. Complexes 1 and 2 were characterized by X-ray crys-
owing to the propensity of these ligands to form polymers. tallography, Mössbauer spectroscopy, UV-Vis, EPR, and NMR
spectroscopy, and by cyclic voltammetry. Reaction of 2 with
AgClO4 produced the diiron(III) complex, [Fe2(μ-OH)2(ClO4)2-
(PIM)(ArTolCO2)Ag] (3), and reaction of 2 with AgSbF6 formed a
species believed to be the tetrairon(III) complex, [Fe4(μ-F)6-
Department of Chemistry, Massachusetts Institute of Technology, Cambridge,
MA 02139, USA. E-mail: lippard@mit.edu
†Electronic supplementary information (ESI) available: X-ray crystallographic
(PIM)2(ArTolCO2)2], which was not fully characterized. When 1
data, Mössbauer spectra, and NMR spectra. CCDC 1054175–1054179. For ESI
and crystallographic data in CIF or other electronic format see DOI: 10.1039/
was allowed to react with O2, three iron(III) species were
observed, two of which were identified as [Fe2(μ-OH)(PIM)-
c5dt02138c
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Fig. 1 Graphical representations of the oxidized (left) and reduced (right) MMOH active sites. The green coloring highlights a carboxylate shift in
Glu243 between the two structures.
(Ph3CCO2)3] (4) and [Fe4(μ-OH)6(PIM)2(Ph3CCO2)2] (5). The
reaction of 2 with O2 produced a mixture of [Fe2(μ-O)(PIM)-
(ArTolCO2)2] (6) and [Fe2(μ-OH)2(PIM)(ArTolCO2)2] (7), com-
plexes that resemble the resting diiron(III) state, MMOHox, of
MMO. When 6/7 further reacted with water, [Fe4(μ-OH)6-
(PIM)2(ArTolCO2)2] (8) formed.
With rare syn-N asymmetrically carboxylate-bridged diiron(II)
complexes 1–2 in hand, we sought to understand their solu-
tion dynamics by applying NMR spectroscopy. Because of their
paramagnetism, however, 1 and 2 are not well suited for such
a study. We therefore introduced fluorine atoms as 19F NMR
spectroscopic handles by modifying the macrocyclic H2PIM
ligand to create H2F2PIM, and introduced the fluorinated ter-
phenylcarboxylate, Ar4F-PhCO2H, which we used previously to
investigate the dynamics of the diiron(II) tetracarboxylate com-
Scheme 1 The carboxylate shift in diiron complexes.
plexes as mentioned above. With these ligands, we prepared
Scheme 2 Summary of chemistry reported for 1 and 2. The macrocyclic ligand, H2PIM, is depicted by colored oxygen atoms (red) and nitrogen
atoms (blue). The full ligand is depicted in the upper right corner.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
three new diiron(II) complexes, [Fe2(PIM)(Ar4F-PhCO2)2] (10), observed in the electron density, the torsion angle was deter-
[Fe2(F2PIM)(ArTolCO2)2] (11), and [Fe2(F2PIM)(Ar4F-PhCO2)2] mined by a difference Fourier analysis followed by a rigid
(12). Their solution dynamics were probed by using VT 19F group refinement. In other cases, the methyl groups were
NMR spectroscopy.
modeled to an idealized disordered methyl group, with two
sets of hydrogen atoms at 50% occupancy, rotated relative to
each other by 60°. The hydrogen atom isotropic displacement
parameters were fixed to 1.5 (methyl) or 1.2 (non-methyl)
times the U value of the atom to which they are bound. Dis-
tance similarity restraints and anisotropic displacement para-
meter restraints were placed on disordered atoms. Data
collection and refinement parameters, and other details of
refinement for individual structures can be found in the ESI.†
Ellipsoid plots and other X-ray structure graphics were
generated with Mercury CSD 3.3.
Experimental methods
General considerations
Chemicals were purchased from commercial sources and used
as received. Solvents were saturated with argon, purified by the
passage through two columns of activated alumina, and stored
over 3 Å molecular sieves in an MBraun dry box. (2-Hydroxy-5-
methylphenyl)boronic acid, (2-hydroxy-5-fluorophenyl)boronic
acid, H2PIM, ArTolCO2H, Ar4F-PhCO2H, compounds L4a, and 2,
were prepared according to published procedures.16–18 All
manipulations of air sensitive compounds were performed in
an MBraun dry box. A ThermoNicolet Avatar 360 spectrometer
was used to obtain IR spectra and the data were processed
with the OMNIC software. Melting points were obtained with a
Stanford Research Systems OptiMelt. NMR spectra were
recorded on either a 500 MHz Varian Inova spectrometer or a
300 MHz Varian Mercury spectrometer. 1H and 13C spectra
were referenced to residual solvent peaks. 19F spectra were
referenced to CFCl3 (0.00 ppm). VT-NMR between 308 and
178 K were performed on a 500 MHz Varian Inova spectro-
meter. Reversibility of the VT-NMR experiments was confirmed
by comparing initial and final spectra at room temperature.
57Fe Mössbauer spectra were obtained on a WEB Research Co.
MSI spectrometer with a 57Co source in Rh matrix. Solid
samples were pulverized and suspended in Apiezon M grease
inside a nylon sample holder and corresponding spectra were
obtained at 80 K. Isomer shift values (δ) were referenced to
metallic iron foil and spectra were fit to Lorentzian lines using
the WMOSS program.
Synthesis
Ligand synthesis
Bis(3-bromobenzyl)ether (L1).
3-Bromobenzaldehyde (9.25 g, 50 mmol) was added to 50 mL
CH2Cl2 and afterwards triethylsilane (6.4 g, 55 mmol) was
added. The reaction was cooled with an ice-bath at 0 °C. Triflic
acid (44 μL) was added carefully to the solution. After 10 min
stirring, the volume was reduced to about 20 mL and the solu-
tion was directly loaded onto a column of silica with some
bicarbonate mixed in the top sand layer to remove trace acid.
The product was eluted with hexanes to 2% ethyl acetate in
hexanes. The combined fractions were stripped to form a
colorless oil (7.8 g, 87.6%). Spectroscopic data matched that
previously reported.16 An ∼7% impurity of TES protected
benzyl alcohol was present and easily removed during purifi-
cation in the next synthetic step.
X-Ray data collection and refinement
Single crystals of H2PIM, H2F2PIM, and 10–12 were coated
with Paratone oil and mounted onto a Bruker SMART APEX
CCD X-ray diffractometer using Mo Kα radiation. Data collec-
tion was performed at 100 K and the diffractometer was con-
trolled with the APEX2 (v. 2010.1-2) software package.19 Data
reduction was performed with SAINT20 and absorption correc-
tions with SADABS.21 XPREP22 was used to determine the
space group through analysis of metric symmetry and systema-
tic absences. Initial solutions were determined using direct
methods and refinement was performed with either the
3′,3′′′-(Oxybis(methylene))bis(5-methyl-[1,1′-biphenyl]-2-ol) (L3a).
SHELXL-97 software package or SHELX-2014 using full-matrix L1 (2.98 g, 8.36 mmol), [Pd(PPh3)4] (386 mg, 334 μmol),
least squares refinement on F2.23 PLATON24 was used to check (2-hydroxy-5-methylphenyl)boronic acid (3.18 g, 20.9 mmol),
for higher symmetry. Non-hydrogen atoms were refined aniso- and K3PO4 (6.21 g, 29.3 mmol) were added to a Schlenk flask,
tropically. Hydrogen atoms were fixed at idealized positions placed under vacuum, and back-filled three times with nitro-
using a riding model except for the hydroxyl protons, which gen. Degassed water (30 mL) and THF (30 mL) were transferred
were located in the electron density maps and refined semi- into the flask via a cannula. The reaction was heated to 80 °C
free using distance restraints appropriate for 100 K. Methyl for 18 h before cooling to rt and subsequent quenching with
group hydrogen atoms were handled on a case-by-case basis. 30 mL 1 M HCl (aq). The product was extracted with 3 × 50 mL
When maxima for methyl group hydrogen atoms could be ethyl acetate, dried over Na2SO4, filtered, and stripped to
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
form an oil. Purification was achieved with column chromato- afford a dark green oil. Purification by column chromato-
graphy (hexanes to 20% ethyl acetate in hexanes), yielding L3a graphy (hexanes to 10% ethyl acetate in hexane) yielded L4b as
as a yellow solid (2.51 g, 73.0%). Spectroscopic data matched a light yellow solid (410.4 mg, 32.1%). 1H-NMR (CDCl3,
that previously reported.16
3′,3′′′-(Oxybis(methylene))bis(5-fluoro-[1,1′-biphenyl]-2-ol) (L3b).
500 MHz): δ 11.34 (s, 2H, OH), 9.86 (s, 2H, CHO), 7.62 (s, 2H,
Ar), 7.53 (d, 2H, Ar), 7.45 (m, 4H, Ar), 7.37 (dd, 2H, Ar), 7.24
(dd, 2H, Ar), 4.67 (s, 4H, CH2). 13C-NMR (CDCl3, 125 MHz): δ
195.85 (CHO), 155.58 (1JC–F = 240 Hz), 155.31 (4JC–F = 2 Hz),
138.52, 135.48 (4JC–F = 1 Hz), 132.23 (3JC–F = 7 Hz), 128.62,
128.60, 127.68, 125.14 (2JC–F = 24 Hz), 120.42 (3JC–F = 6 Hz),
117.35 (2JC–F = 22 Hz) (Ar), 72.15 (CH2). 19F-NMR (CDCl3,
3
282 MHz): δ −123.7 (t, JH–F = 8 Hz). ESI-MS(−) = 473.1
[M − H]− (Calcd = 473.12 [M − H]−), ESI-MS(+) = 513.1
[M + K]+, (Calcd = 513.09 [M + K]+). Mp = 123 °C.
H2F2PIM.
[Pd(PPh3)4] (175 mg, 385 μmol), (2-hydroxy-5-fluorophenyl)-
boronic acid (1.50 g, 9.62 mmol), and K3PO4 (2.87 g,
13.5 mmol) were added to a Schlenk flask. The system was
placed under vacuum and back-filled with nitrogen three
times. Degassed water (15 mL) and di(3-bromobenzyl) ether
(1.37 g, 3.85 mmol) dissolved in THF (15 mL) were transferred
into the flask by cannula. The system was heated to reflux and
was allowed to stir for 20 h. Water (10 mL) was added and the
product was extracted three times with 50 mL of ethyl acetate.
The combined organic solutions were dried with MgSO4, fil-
tered, and stripped to form a black oil. Column chromato-
graphy (hexanes to 20% ethyl acetate in hexanes) yielded
1
L3b as a tan solid (1.27 g, 78.5%). H-NMR (CDCl3, 500 MHz):
δ 7.47 (m, 4H, Ar), 7.40 (m, 4H, Ar), 6.94 (m, 4H, Ar), 6.89 (m,
2H, Ar) 5.15 (br s, 2H, OH), 4.65 (s, 4H, CH2). 13C-NMR (CDCl3,
75 MHz): δ 157.23 (1JC–F = 239 Hz), 148.70 (4JC–F = 2 Hz),
139.28, 136.79 (4JC–F = 2 Hz), 129.63, 129.12 (3JC–F = 8 Hz),
128.65, 128.59, 127.86, 117.12 (3JC–F = 8 Hz), 116.65 (2JC–F = 23
Hz), 115.63 (2JC–F = 23 Hz) (Ar), 72.52 (CH2). 19F-NMR (CDCl3,
282 MHz): δ −124.3 (m). Mp = 86 °C.
3,3′-Diaminodiphenyl sulfone (196 mg, 790 μmol) and 5b
(375 mg, 790 μmol) were added to a dry Schlenk flask with
175 mL of dry acetonitrile. Trifluoroacetic acid (220 μL) was
added to the mixture, which was heated to 85 °C. After 6 h, the
reaction was allowed to cool to rt, during which time a large
amount of yellow precipitate formed. The solid was collected
by filtration, washed with ether, and dried under vacuum. The
material (382 mg, 70.4%) was used without further purifi-
cation. 1H-NMR (CDCl3, 500 MHz): δ 13.18 (br s, 2H, OH), 8.62
(s, 2H, CHN), 7.92 (d, 2H, Ar), 7.83 (s, 2H, Ar), 7.76 (s, 2H, Ar),
7.59 (t, 2H, Ar), 7.44 (m, 6H, Ar), 7.36 (d, 2H, Ar), 7.22 (dd, 2H,
Ar), 7.12 (dd, 2H, Ar), 4.68 (s, 4H, CH2). 13C-NMR (CD2Cl2,
125 MHz): δ 164.35 (CHN), 156.06 (1JC–F = 237 Hz), 155.60,
149.69, 143.64, 138.52, 137.33, 132.19 (3JC–F = 7 Hz), 131.46,
130.21, 129.41, 128.84, 128.63, 126.38, 124.95, 123.56, 122.19
3′,3′′′-(Oxybis(methylene))bis(5-fluoro-2-hydroxy-[1,1′-biphenyl]-
3-carbaldehyde) (L4b).
(2JC–F = 24 Hz), 119.52 (3JC–F = 8 Hz), 117.37 (2JC–F = 23 Hz) (Ar),
3
74.07 (CH2). 19F-NMR (CD2Cl2, 470 MHz): δ −125.6 (t, JH–F
=
This reaction is extremely moisture sensitive and it is therefore 9 Hz). Mp = 184 °C.
important to rigorously dry all reagents and solvents. L3b
(1.13 g, 2.70 mmol), paraformaldehyde (1.22 g, 40.5 mmol),
Metal complex synthesis
[Fe2(PIM)(Ar4-PhCO2)2] (10). H2PIM (157.5 mg, 232 μmol) and
MgCl2 (1.03 g, 10.8 mmol), and TEA (2.26 mL, 16.2 mmol) Ar4F-PhCO2H (144.0 mg, 463 μmol) were combined in 3 mL of
were added to a Schlenk flask containing 7.5 g of 3 Å activated THF. [Fe2(mes)4] (136.5 mg, 232 μmol), pre-dissolved in 1.5 mL
sieves. Acetonitrile (125 mL) was added to the flask by THF, was injected quickly to create a dark red mixture. After
cannula. The reaction mixture was heated to reflux and stirred 3 h, the solvent was stripped, the crude residue washed with
for 3 days, quenched with 1 M HCl (ca. 125 mL), and stripped 3 × 1 mL ether, and the material was crystallized by slow
from as much acetonitrile as possible. The product was diffusion of pentane into a solution in CH2Cl2. The resulting
extracted with 3 × 150 mL of ethyl acetate. The combined red crystals were washed (2 × 1 mL pentane) and dried under
organic layers were dried with Na2SO4, filtered, and stripped to vacuum to yield 242.2 mg (74.2%) red crystalline material. Anal.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
Calcd for Fe2C80H54F4N2O9S·(C0.5H1Cl1) (10·0.5CH2Cl2): C, 10 min of reaction time, immediate purification provides gram
66.70; H, 3.82; N, 1.93. Found: C, 66.95; H, 3.61; N, 1.92. quantities of L1 in 87.6% yield. The original synthesis of
CH2Cl2 was detected in solution NMR samples of 10 in THF- H2PIM (Scheme S1†) was limited by a Negishi-coupling of L1
d8. IR (KBr): 3049 (w), 2920 (w), 2850 (w), 2336 (w), 1653 (w), and the aryl-zinc reagent of 2-bromo-4-methyl-phenol-tetra-
1604 (m), 1573 (s), 1534 (s), 1506 (s), 1433 (m), 1411 (m), 1382 hydropyran that proceeded in 35% yield.16 To improve this
(m), 1325 (m), 1308 (m), 1282 (m), 1221 (m), 1199 (m), 1091 coupling step, the aryl-zinc reagent was replaced with an aryl-
(m), 1074 (m), 979 (w), 837 (s), 790 (s), 707 (m), 685 (w), 595 boronic acid, L2, to do a Suzuki-coupling. By doing so, a pro-
(w), 539 (m), and 469 (w) cm−1. Mössbauer (polycrystalline, tecting group for the phenoxyl moiety was not needed, elimi-
apiezon M grease): δ = 1.05(2) mm s−1, ΔEQ = 2.12(2) mm s−1
ΓL/R = 0.33(2) mm s−1. Mp (dec) = 74 °C.
,
nating the next step in the main synthetic sequence. Thus, the
new synthesis required (2-hydroxy-5-methylphenyl)boronic
[Fe2(F2PIM)(ArTolCO2)2] (11). H2F2PIM (150 mg, 218 μmol) acid (L2a), which was prepared from the commercially avail-
and ArTolCO2H (132 mg, 437 μmols) were dissolved in THF able (2-methoxy-5-methylphenyl)boronic acid following a lit-
(3 mL) in a dry box. [Fe2(mes)4] (129 mg, 218 μmol) was dis- erature procedure.18 The Suzuki-coupling between L1 and L2a
solved in 1.5 mL THF and added to the mixture. After 2 h, the yielded the desired product, L3a, in 61–80% yield using [Pd-
solvent was removed and the residue washed twice with 1 mL (PPh3)4] as the catalyst. The last step of the synthesis was
of ether. The residue was then dissolved in 2 mL of CH2Cl2 unmodified. In total, the overall yield for the synthesis of
and layered with 10 mL of pentane. The resultant material H2PIM increased to 30.9% from 13.3%. H2F2PIM, the analogue
(278 mg, 91.0%) was filtered and washed with pentane. X-ray of H2PIM where fluorine substituents replace the methyl
quality crystals were obtained by slow diffusion of pentane groups, was prepared in an analogous manner in an overall
into a solution of 11 in CH2Cl2. Anal. Calcd for Fe2C82H60F2- yield of 15.1%. The solid-state structures of both H2PIM and
N2O9S·(C4H9Cl3) (11·1.5CH2Cl2·0.5C5H12): C, 66.10; H, 4.45; N, H2F2PIM were obtained and show the ligands to be pre-
1.79. Found: C, 66.18; H, 4.06; N, 1.82. CH2Cl2 and pentane organized for supporting a dinuclear complex. Crystallo-
were detected in NMR samples of 11. IR (KBr): 3054 (w), 3028 graphic details are reported in the ESI.†
(w), 2915 (w), 2857 (w), 1649 (m), 1601 (m), 1581 (s), 1527 (s),
Analogous to the synthesis of 1–2, 10–12 were synthesized
1449 (m), 1434 (m), 1403 (m), 1386 (m), 1341 (m), 1302 (m), in one pot by mixing the appropriate macrocyclic ligand, 2.0
1277 (m), 1201 (m), 1184 (s), 1149 (s), 1110 (m), 1070 (m), 1000 equiv. of carboxylic acid, and 1.0 equiv. of [Fe2(mes)4] in THF;
(m), 956 (w), 881 (w), 838 (w), 791 (s), 762 (m), 736 (m), 702 the yields ranged from 74.2–92.8%. X-ray quality crystals of
(m), 685 (m), 604 (m), 543 (m), 529 (m), and 459 (w) cm−1
.
10–12 were obtained by slow diffusion of pentane into methyl-
Mössbauer (polycrystalline, apiezon M grease): (site 1, 50%): δ ene chloride solutions of the complexes (Fig. 2). Compounds
= 0.93(2) mm s−1, ΔEQ = 2.03(2) mm s−1, ΓL/R = 0.27(2) mm s−1 10–12 feature the expected syn-N dicarboxylate-bridged diiron
(site 2, 50%): δ = 1.15(2) mm s−1, ΔEQ = 1.98(2) mm s−1, ΓL/R
=
centers. A comparison of the bond distances in the m-terphenyl-
carboxylate-bridged complexes 2 and 10–12 is provided in
0.28(2) mm s−1. Mp (dec) = 200 °C.
[Fe2(F2PIM)(Ar4-PhCO2)2] (12). The same procedure was used Table 1. There are noteworthy differences in bond distances
as that described for 10 except that H2F2PIM (159 mg, and angles of the μ–η1:η2 bridging carboxylate ligand for these
232 μmol) replaced H2PIM. Obtained were 304.7 mg (92.8%) of structures. The Fe1–O3 bond length increases from 2.342(6) Å
maroon crystals. Anal. Calcd for Fe2C78H48F2N2O9S·(CH2Cl2) in 2 to 2.348(2) Å in 12, 2.443(2) Å in 11, and 2.535(2) Å in 10.
(12·CH2Cl2): C, 63.00; H, 3.31; N, 1.81. Found: C, 63.26; H, The trend does not follow the number of methyl-to-fluoro sub-
3.36; N, 1.87. CH2Cl2 was detected in NMR samples of 12. IR stitutions in the complexes and is tentatively attributed to
(KBr): 3054 (w), 2954 (w), 2933 (w), 2859 (w), 1610 (m), 1581 packing interactions. The weakening of the Fe1–O3 interaction
(m), 1532 (s), 1510 (s), 1446 (m), 1405 (m), 1386 (m), 1321 (m), correlates with an increase in the Fe1–O4–C1 angle, which
1299 (m), 1212 (m), 1187 (m), 1152 (m), 1074 (m), 1001 (w), rises from 95° in 2 and 12 to 99° in 11 and 102° in 10. A corres-
958 (w), 840 (m), 794 (s), 698 (m), 633 (w), 542 (m), and 469 (w) ponding decrease in the Fe1–O4 distance occurs, from 2.103(2)
cm−1. Mössbauer (polycrystalline, apiezon M grease): δ = Å in 12 to 2.062(2) Å in 11 and 2.043(2) Å in 10. These corre-
1.04(2) mm s−1, ΔEQ = 2.01(2) mm s−1, ΓL/R = 0.33(2) mm s−1
Mp (dec) = 232 °C.
.
lated changes are consistent with what is predicted to occur
for varying the strength of the interaction of the dangling
oxygen to the metal center9 and demonstrate a minor carboxy-
late shift within the solid state.
In all solid-state structures to date containing PIM or F2PIM
there is a wedge-shaped conformation (curved green line,
Fig. S14†). In 2 and 10–12, the wedge provides space for the
Results and discussion
Synthesis and characterization
In our previous report, H2PIM was prepared in gram quantities m-terphenylgroup of the μ–η1:η2 bridging carboxylate, which
by an 8 step synthesis.16 Since then, a shorter route with 5 packs along the walls of the wedge (Fig. S15†). The m-terphenyl
steps was developed (Scheme 3). Instead of the original two- unit of the μ–η1:η1 bridging carboxylate is oriented perpendicu-
step procedure to produce the bis(3-bromobenzyl)ether (L1), a lar to the edge of the wedge (Fig. S15†). The configuration in 1
reductive condensation reaction25 was utilized with 3-bromo- (Fig. S15,† top left) is reversed, however, most likely as a conse-
benzaldehyde, 1.1 equiv. of Et3SiH, and 1 mol% TfOH. After quence of the shape of the carboxylate. The steric bulk of the
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Scheme 3 Synthesis of H2PIM and H2F2PIM.
Fig. 2 Solid state structures of 10–12 and cartoon of the μ–η1:η2 carboxylate bridging mode (bottom right). Ellipsoids are drawn at 50% and sol-
vents, disordered atoms, and hydrogen atoms are removed for clarity. The carbon atoms of the m-terphenyl wings are colored bright blue to help
distinguish those atoms from that of the macrocyclic ligand.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
Table 1 Comparison of select interatomic distances and angles in 2, and 10–12, in order of increasing Fe1–O3 distance. The numbering scheme
corresponds to the cartoon in Fig. 2 and does not necessarily agree with numbers assigned in the X-ray structure
Å
Fe1–Fe2
Fe1–O1
Fe1–N1
Fe1–O3
Fe1–O4
Fe1–O5
Fe2–O2
Fe2–N2
Fe2–O4
Fe2–O6
2
3.601(1)
1.895(4)
1.900(3)
1.883(3)
1.881(2)
2.044(4)
2.054(3)
2.042(3)
2.051(3)
2.342(6)
2.348(3)
2.443(2)
2.535(2)
2.091(3)
2.103(2)
2.062(2)
2.043(2)
2.048(6)
2.048(3)
2.028(2)
2.026(2)
1.888(3)
1.888(3)
1.886(3)
1.877(2)
2.038(5)
2.024(3)
2.031(3)
2.032(3)
2.029(4)
2.039(2)
2.038(2)
2.055(2)
1.998(4)
1.955(3)
1.991(2)
1.973(2)
12
11
10
3.6387(6)
3.5648(8)
3.5680(7)
°
Fe1–O4–Fe2
Fe1–O4–C1
Fe2–O4–C1
2
122.2(2)
122.9(1)
120.8(1)
121.0(1)
95.5(3)
95.4(2)
98.7(2)
102.2(2)
133.4(4)
133.8(2)
129.9(2)
132.9(2)
12
11
10
trityl groups in 1 is directed away from the macrocycle whereas
the m-terphenyl groups in the other complexes point toward
the macrocycle. The m-terphenyl groups impose a stronger
effect on the conformation of the macrocycle. These points
will be important later in our discussion of the solution
dynamics of these complexes.
Mössbauer spectra for compounds 10–12 were obtained at
80 K and compared to those for 1 and 2 (Table 2). Compounds
1–2 and 11 nicely fit to a two-site model and 10 and 12 fit to a
single site (Fig. 3 and S6†). The reason for the difference is not
clear, although 10 and 12 both contain the Ar4F-PhCO2H carb-
oxylate, unlike the carboxylates in 1 (Ph3CCO2H) and 2 and 11
(ArTolCO2H). The δ values for 10–12 range from 0.93 to
1.15 mm s−1 with ΔEQ values ranging from 1.98 to 2.12 mm
s−1, consistent with high-spin Fe(II) centers, as is the case
with 1–2.
Table 2 Mössbauer parameters for 1–2 and 10–12
Site 1
Site 2
(mm s−1
)
δ
ΔEQ
ΓL/R
δ
ΔEQ
ΓL/R
1a
2a
11
10
12
0.97
0.95
0.93
1.05
1.04
2.25
2.02
2.03
2.12
2.01
0.35
0.32
0.27
0.33
0.33
47%
37%
50%
—
1.18
1.10
1.15
—
2.33
2.04
1.98
—
0.38
0.38
0.28
—
53%
63%
50%
—
—
—
—
—
—
a Ref. 18.
BMMs, we sought to understand the carboxylate dynamics in
this system by using solution NMR spectroscopy. Standard H
1
spectroscopic NMR techniques were not possible because of
paramagnetic effects, but fluorine substitutions provide a
spectroscopic handle with 19F NMR. VT 19F NMR spectra for
10–12 were obtained.
NMR studies
In THF-d8, the 19F NMR spectrum of compound 10 shows a
single fluorine peak at −116 ppm, indicating rapid exchange
Owing to the rare asymmetric carboxylate bridging modes in
1–2 as well as the importance of the carboxylate shift in
Fig. 3 Mössbauer spectra of 11 (left) and 10 (right).
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Fig. 4 VT 19F NMR of 10 in CD2Cl2. Intramolecular processes 1 and 2 are highlighted in blue and green respectively. Minor impurities are observed
at higher temperatures.
of all four F atoms at room temperature. As the system is four very weak fluorine resonances of unknown origin were
cooled down, the single peak remains and broadens slightly observed at −111.6, −112.2, −113.1, and −119.5 ppm. These
upon reaching −70 °C. The broadening suggests that fluorine features are believed to be associated with a structurally
atom exchange was slowed down, but not enough to show similar species and are also observed in samples of 2 + 10 and
splitting on the 19F NMR timescale. We wondered whether the 12, discussed below.
coordinating capabilities of THF to the iron atoms may
increase the exchange rates of the carboxylate ligands, and occur through intermolecular or intramolecular exchange
CD2Cl2 was therefore investigated as an alternative solvent.
events, we prepared a mixture of 2 and 10 and repeated the 19
To test whether the dynamic processes observed for 10
F
At 35 °C in CD2Cl2, 10 displays a single 19F peak at NMR experiment. If intermolecular exchange of carboxylates
−113.9 ppm. As the temperature is decreased, this peak broad- occurs, a new species with a mixture of carboxylates, [Fe2(PIM)-
ens and becomes nearly flat at −10 °C (Fig. 4). Two peaks then (ArTolCO2)(Ar4F-PhCO2)], would form, which might be dis-
grow in at −109.0 ppm and −119.2 ppm, commencing at tinguishable from 10 by its 19F NMR spectrum (Scheme 4). At
−20 °C. The feature at −109.0 ppm sharpens into a peak at 35 °C, two fluorine peaks are observed at −113.0 and
−106.3 ppm at −60 °C and then starts to broaden and further −113.9 ppm (Fig. S1†), a clear indication that intermolecular
shift to −103.4 ppm at −95 °C, possibly due to increased carboxylate exchange occurs. Furthermore, the observation of
solvent viscosity at low temperature. The peak at −119.2 ppm two, instead of one, peak means that the intermolecular carb-
sharpens slightly at −30 °C and shifts to −120.3 ppm and then oxylate exchange process is slower on the NMR time scale at
broadens into the background at −50 °C. After further cooling 35 °C. Therefore, all exchange processes that we observe in our
to −60 °C, two peaks arise at −118.2 ppm and −127.8 ppm, VT experiments in CD2Cl2 are intramolecular processes. To
which sharpen and shift to −119.3 ppm and −131.3 ppm at ensure that the exchange processes observed in the VT NMR of
−80 °C before broadening at −95 °C with chemical shift values 10 were not altered by addition of 2, the spectrum was exam-
of −120.6 and −134.4 ppm. Integration of the three peaks ined down to −95 °C. The VT NMR of 2 + 10 looks the same as
(−104.6, −119.3, and −131.3 ppm) at −80 °C reveals a ratio of that as 10 alone except that the number of peaks is doubled
2 : 1 : 1, respectively. We refer to the first peak split at 35 °C at owing to the presence of [Fe2(PIM)(ArTolCO2)(Ar4F-PhCO2)]. At
−113.9 ppm as process
1 and the split of the broad −50 °C, an additional peak occurs at −107.0 ppm and at
−120.3 ppm peak at −30 °C as process 2. At 35 °C and below, −80 °C, and additional peaks are observed at −120.1 and
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
changes occurred for 12 except for slight alterations in chemi-
cal shifts (Fig. S4†). The F2PIM2− fluorine resonance in 12 is
1.6–6.5 ppm upfield relative to that in 11. The Ar4F-PhCO2 flu-
−
orine resonances in 12 are 1.2–3.2 ppm upfield of those in 10.
We can conclude that there are no significant differences
between in the exchange processes in complexes 2 and 10–12.
A model for these intermolecular processes is presented in
Scheme 5. The PIM fluorine substituents are labeled A or B
and each fluorine atom on the carboxylate ligands is assigned
the color blue, orange, green, or red. Examination of the space
filling model of any of the complexes 2 and 10–12 clearly indi-
cates that rotation of the μ–η1:η2 carboxylate would be unlikely
due to a steric clash with the PIM macrocycle. Rotation of the
μ–η1:η1 carboxylate to interchange the blue and orange
markers appears possible, although one of the two Fe–O
bonds might need to transiently dissociate in the process. The
μ–η1:η2 carboxylate could shift into a μ–η1:η1 bridging mode to
form an intermediate with approximate two-fold symmetry. If
the same carboxylate undergoes a μ–η1:η1 to μ–η1:η2 carboxy-
late shift, it would either return to the original complex or the
mirror image of the original complex. The mirror image
results in the interchange of A and B. Similar to the blue and
orange carboxylates, the green and red carboxylates may be
able to rotate in a μ–η1:η1 bridging mode and such green-red
intramolecular interchange could occur after a μ–η1:η1 to
μ–η1:η2 carboxylate shift. As discussed above, the PIM and
F2PIM macrocycles in 2 and 10–12 adopt a wedge-shaped con-
formation that encircles the diiron(II) center. The diphenyl-
sulfone unit curves to orient the wedge-shape to form a cradle
for the m-terphenyl group of the μ–η1:η2 bridging carboxylate.
In order to interchange the blue/orange and red/green pairs
(Scheme 5), the macrocyclic ligand must undergo a confor-
mational change that inverts the conformation of the sulfone,
hereafter the “sulfone inversion”. With the m-terphenyl group
of the μ–η1:η2 bridging carboxylate packed tightly with the
macrocyclic cradle, it is likely that a carboxylate shift to the
symmetric intermediate would be necessary before a sulfone
inversion, which would need to be associated with rotations of
both carboxylates. The sulfone inversion would lead to mixing
of all color markers and A and B.
From the VT NMR spectroscopic experiments, we know that
process 1 distinguishes the two carboxylates, which corres-
ponds to the sulfone inversion in our model. Process 2 differ-
entiates the two sides of either the red/green or the blue/
orange carboxylates (Scheme 5). Our data do not allow us to
conclusively assign the individual carboxylates in this
exchange process. A blue/orange exchange by rotation of the
O2C–Caryl bond appears to be sterically inhibited by sulfone
and ether units of the macrocycle projecting below the macro-
cycle plane. A similar O2C–Caryl bond rotation on the red/green
carboxylate on the canyon side of the macrocycle during a
μ–η1:η1 bridging mode intermediate appears less sterically
hindered than that of the blue/orange exchange. We therefore
tentatively assign process 2 as a blue/orange exchange. The
carboxylate exchange that gives rise to process 2 slows down
sufficiently below −95 °C to give rise to the observed NMR
Scheme 4 Depiction of the intermolecular carboxylate exchange
between
(Ar4F-PhCO2)].
2
and 10 through an intermediate, [Fe2(PIM)(ArTolCO2)-
−135.2 ppm. Integration at −80 °C, however, reveals an
approximately 4 : 1 : 1 ratio in order of decreasing chemical
shift. This observation suggests that the two carboxylate pos-
itional isomers of [Fe2(PIM)(ArTolCO2)(Ar4F-PhCO2)] are not
equivalent and that one is favored over the other. This ratio
also supports the notion that process 1 involves an exchange
between the two carboxylate positions.
Having asymmetric bridging carboxylates in these diiron(II)
complexes eliminates the 2-fold symmetry axis of the X2PIM2−
ligand and introduces chirality to these complexes. Moreover,
the two fluorine atoms in F2PIM2− are chemically inequivalent
in these complexes and could offer some information about
the intramolecular exchange processes observed in 10. To
explore this possibility, we first examined the VT 19F NMR
spectrum of 11 in CD2Cl2. At 35 °C, a single fluorine resonance
occurs at −32.4 ppm (Fig. S2†). The free ligand, H2F2PIM, has
a fluorine chemical shift at −125.6 ppm in CD2Cl2 at 25 °C.
The 93 ppm difference in chemical shift is indicative of para-
magnetic shifting and proves that iron binds to F2PIM2− in
solution. Cooling to −95 °C shifts the −32.4 ppm resonance to
38.8 ppm. The peak slowly broadens with decreasing tempera-
ture, indicating that the two fluorine atoms exchange rapidly
on the NMR time scale at all temperatures in this experiment,
but the broadening at low temperature suggests that the
exchange process slows down. Another possibility is that the
broadening is due to viscosity changes as the temperature is
lowered. The chemical shift change in this fluorine peak
follows Curie behavior (Fig. S3†).
The VT 19F NMR spectrum of 12 was also obtained to
ensure that the change to F2PIM2− from PIM2− does not alter
the exchange processes observed in 10. Indeed, no significant
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Scheme 5 Model for the intramolecular exchange processes that occurs in the family of compounds 2, 10–12. A space-filling model of 12 is pro-
vided in the top left in order to visualize the spatial orientation of the ligands.
spectral behavior, characterized by broadened peaks at −103.4 suggested coordination changes in Glu209 to be the main
and −104.9 ppm in 10 and 12, respectively. Additional broad- alteration. A comparison of diiron active sites in X-ray struc-
ening occurs below −80 °C for the upfield carboxylate at tures of MMOHred and that of the H-B complex reveal that, in
−119.3 and −131.3 ppm in 10 and may reflect hindered both cases, Glu243 bridges in a μ–η1:η2 mode, but the confor-
rotation of the 4-fluorophenyl groups on the m-terphenyl mation differs.28 The 2.9 Å resolution of the latter complex as
groups of the Ar4F-PhCO2 ligand. The data from the macro- well as the ambiguity of its oxidation state prevent a meaning-
−
cycle fluorine atoms in 11 and 12 are consistent with intercon- ful explanation, however. Nevertheless, it is clear from all
version of the two enantiomers generated by the carboxylate these observations that there is conformational flexibility in
shift remaining rapid on the NMR time scale even at −95 °C.
the coordination of Glu209 and Glu243 at the active site of
MMOHred. In the present model system, the macrocyclic
framework preorganizes the diiron(II) center to contain both
μ–η1:η1 and μ–η1:η2 bridging modes, similar to that in
MMOHred. The variable temperature 19F solution NMR spectro-
scopic study described here demonstrates that the bridging
carboxylates retain the ability to shift coordination modes
while remaining as a diiron(II) complex. Thus the diiron(II)
PIM complexes have the flexibility to accommodate structural
Comparison to MMOH
A significant change in the coordination mode of the bridging
Glu243 in MMOH occurs upon reduction of the diiron(III)
active site to the diiron(II) form.3 EXAFS26 and MCD27 studies
of MMOHred in complex with the regulatory protein, MMOB,
showed additional alterations in the primary coordination
sphere at the diiron center at the active site. The MCD study
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
changes, evoked by temperature changes, that provide a valu-
able precedent of potential relevance to the conformational
alterations observed in soluble methane monooxygenase.
8 D. A. Whittington and S. J. Lippard, J. Am. Chem. Soc.,
2001, 123, 827–838.
9 R. L. Rardin, W. B. Tolman and S. J. Lippard, New J. Chem.,
1991, 15, 417–430.
10 L. H. Do, T. Hayashi, P. Moënne-Loccoz and S. J. Lippard,
J. Am. Chem. Soc., 2010, 132, 1273–1275.
Conclusions
11 J. R. Frisch, R. McDonnell, E. V. Rybak-Akimova and
L. Que Jr., Inorg. Chem., 2013, 52, 2627–2636.
12 L. Que Jr. and W. B. Tolman, Nature, 2008, 455, 333–340.
13 L. H. Do and S. J. Lippard, J. Inorg. Biochem., 2011, 105,
1774–1785.
14 B. Burger, S. Dechert, C. Grosse, S. Demeshko and
F. Meyer, Chem. Commun., 2011, 47, 10428–10430.
15 D. Lee and S. J. Lippard, Inorg. Chem., 2002, 41, 2704–
2719.
A series of fluorine substituted analogues of 2, 10–12, were pre-
pared and characterized. From VT 19F NMR spectra of these
complexes it was possible to characterize intramolecular carb-
oxylate exchange processes for this family of syn-N asymmetri-
cally carboxylate-bridged diiron(II) complexes. The results
mimic the occurrence of carboxylate shifts observed in
MMOHred, evoked both by redox changes and complex for-
mation with the regulatory protein MMOB.
16 L. H. Do and S. J. Lippard, J. Am. Chem. Soc., 2011, 133,
10568–10581.
17 C.-T. Chen and J. S. Siegel, J. Am. Chem. Soc., 1994, 116,
5959–5960.
18 K. Vishnumurthy and A. Makriyannis, J. Comb. Chem.,
2010, 12, 664–669.
19 APEX2, APEX2, 2008-4.0, Bruker AXS, Inc, Madison, WI,
2008.
Acknowledgements
This work was supported by a grant from the National Institute
of General Medical Sciences (Grant GM 32114 to S. J. Lippard)
and the National Science Foundation Graduate Research
Fellowship Program (Grant No. 1122374 to M. A. Minier).
20 SAINT, SAINT: SAX Area-Detector Integration Program,
2008/1, University of Göttingen, Göttingen, Germany,
2008.
References
1 A. L. Feig and S. J. Lippard, Chem. Rev., 1994, 94, 759–805.
2 B. J. Wallar and J. D. Lipscomb, Chem. Rev., 1996, 96, 2625–
2658.
21 G. M. Sheldrick, SADABS: Area-Detector Absorption Correc-
tion, University of Göttingen, Göttingen, Germany, 2008.
22 XPREP, XPREP, 2008/2, Bruker AXS, Madison, WI, 2008.
3 M. Merkx, D. A. Kopp, M. H. Sazinsky, J. L. Blazyk, 23 G. M. Sheldrick, Acta Crystallogr., Sect. A: Fundam. Crystal-
J. Müller and S. J. Lippard, Angew. Chem., Int. Ed., 2001, 40,
2782–2807.
4 C. E. Tinberg and S. J. Lippard, Acc. Chem. Res., 2011, 44,
280–288.
5 A. C. Rosenzweig, C. A. Frederick, S. J. Lippard and
P. Nordlund, Nature, 1993, 366, 537–543.
6 A. C. Rosenzweig and S. J. Lippard, Acc. Chem. Res., 1994,
27, 229–236.
logr., 2008, 64, 112–122.
24 A. L. Spek, PLATON, A Multipurpose Crystallographic Tool,
Utrecht University, Utrecht, The Netherlands, 2008.
25 B. A. Gellert, N. Kahlcke, M. Feurer and S. Roth, Chem. –
Eur. J., 2011, 17, 12203–12209.
26 D. J. Rudd, M. H. Sazinsky, S. J. Lippard, B. Hedman and
K. O. Hodgson, Inorg. Chem., 2005, 44, 4546–4554.
27 N. Mitić, J. K. Schwartz, B. J. Brazeau, J. D. Lipscomb and
E. I. Solomon, Biochemistry, 2008, 47, 8386–8397.
7 A. C. Rosenzweig, P. Nordlund, P. M. Takahara,
C. A. Frederick and S. J. Lippard, Chem. Biol., 1995, 2, 409– 28 S. J. Lee, M. S. McCormick, S. J. Lippard and U.-S. Cho,
418.
Nature, 2013, 494, 380–384.
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.