Design and synthesis of a dinucleating ligand system with varying terminal donor functions that provides no bridging donor and its application to the synthesis of a series of FeIII-μ-O-FeIII complexes

View Article Online

View Journal

Dalton

Transactions

Accepted Manuscript

This article can be cited before page numbers have been issued, to do this please use: J. B. Strautmann,

S. Dammers, T. Limpke, J. Pathier, T. Zimmermann, S. Walleck, G. Heinze-Brückner, A. Stammler, H.

Bögge and T. Glaser, Dalton Trans., 2016, DOI: 10.1039/C5DT03711E.

This is an Accepted Manuscript, which has been through the

Royal Society of Chemistry peer review process and has been

accepted for publication.

Accepted Manuscripts are published online shortly after

acceptance, before technical editing, formatting and proof reading.

Using this free service, authors can make their results available

to the community, in citable form, before we publish the edited

article. We will replace this Accepted Manuscript with the edited

and formatted Advance Article as soon as it is available.

You can find more information about Accepted Manuscripts in the

Information for Authors .

Please note that technical editing may introduce minor changes

to the text and/or graphics, which may alter content. The journal’s

standard T erms & Conditions and the Ethical guidelines still

apply. In no event shall the Royal Society of Chemistry be held

responsible for any errors or omissions in this Accepted Manuscript

or any consequences arising from the use of any information it

contains.

www.rsc.org/dalton

Page 1 of 42

Dalton Transactions

View Article Online

DOI: 10.1039/C5DT03711E

Design and Synthesis of a Dinucleating Ligand System with Varying

Terminal Donor Functions that Provides no Bridging Donor and its

Application to the Synthesis of a Series of Fe^III-ꢀ-O-Fe^IIIComplexes

Julia Bernhardette Hildegard Strautmann, Susanne Dammers, Thomas Limpke, Janine

Parthier, Thomas Zimmermann, Stephan Walleck, Gabriele Heinze-Brückner, Anja

Stammler, Hartmut Bögge, Thorsten Glaser*

Lehrstuhl für Anorganische Chemie I, Fakultät für Chemie, Universität Bielefeld, Universitätsstr.

25, Dꢀ33615 Bielefeld, Germany.

www.uniꢀbielefeld.de/chemie/ac1chair/

* Corresponding author. EꢀMail address: thorsten.glaser@uniꢀbielefeld.de

Abstract

Based on a rational ligand design for stabilizing highꢀvalent {Fe(ꢀꢀO)₂Fe} cores, a new family of

dinucleating bis(tetradentate) ligands with varying terminal donor functions has been developed:

2

.

redoxꢀinert biomimetic carboxylates in H₄julia, pyridines in susan, and phenolates in H₄hilde^Me

Based on a retrosynthetic analysis, the ligands were synthesized and used for the preparation

of their diferric complexes [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]•6H₂O, [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]ꢀ

Me₂

•7H₂O, [(julia){Fe(DMSO)(µꢀO)Fe(DMSO)}]•3DMSO, [(hilde^Me){Fe(µꢀO)Fe}]•CH₂Cl₂, [(hilde )ꢀ

2

{FeCl}₂]•2CH₂Cl₂, [(susan){FeCl(µꢀO)FeCl}]Cl₂•2H₂O, [(susan){FeCl(µꢀO)FeCl_0.75(OCH₃)_0.25}]ꢀ

(ClO₄)₂•0.5MeOH, and [(susan){FeCl(µꢀO)FeCl}](ClO₄)₂•0.5EtOH, which were characterized by

singleꢀcrystal Xꢀray diffraction, FTIR, UVꢀVisꢀNIR, Mössbauer, magnetic, and electrochemical

measurements. The strongly electronꢀdonating phenolates afford fiveꢀcoordination, while the

carboxylates and pyridines lead to sixꢀcoordination. The analysis of the ligand conformations

demonstrates a strong flexibility of the ligand backbone in the complexes. The different

hydrogenꢀbonding in the secondary coordination sphere of [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]

influences the CꢀO, C=O, and FeꢀO bond lengths and is reflected in the FTIR spectra. The

physical properties of the central {Fe(

governed by the differences in terminal ligands – Fe^IIIbonds: strongly covalent

phenolates, less covalent ꢀdonation with carboxylates, and ꢀacceptation with pyridines. Thus,

ꢀꢀO)Fe} core (dꢀd,

ꢀ

ꢀoxo→Fe^IIICT,

ν

_as(FeꢀOꢀFe), J) are

πꢀdonation with

π

[(susan){FeCl(µꢀO)FeCl}]²⁺is oxidized at 1.48 V vs Fc⁺/Fc, which is shifted to 1.14 V vs Fc⁺/Fc

by methanolate substitution, while [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}] is oxidized ≤ 1 V vs Fc⁺/Fc.

1

Dalton Transactions

Page 2 of 42

View Article Online

DOI: 10.1039/C5DT03711E

+

[(hilde^Me){Fe(µꢀO)Fe}] is oxidized at 0.36 V vs Fc /Fc to a phenoxyl radical. The catalytic

2

oxidation of cyclohexane with TONs up to 39.5 and 27.0 for [(susan){FeCl(µꢀO)FeCl}]²⁺and

[(hilde^Me){Fe(µꢀO)Fe}], respectively, indicates the potential to form oxidizing intermediates.

2

Introduction

Nature frequently employs metal ions (mainly Fe and Cu) in the active sites of enzymes to store

oxidation and/or oxygenation equivalents of dioxygen during catalytic cycles.¹A notable

example is the enzyme soluble methanemonooxygenase (sMMO), which has a dinuclear diiron

active site for which a highꢀvalent diiron(IV) intermediate (Q) is proposed to store the oxidation

equivalents of dioxygen.^2,3The modelling of the structure and the reactivity of this active site

and its intermediates is a field of intense research in bioinorganic chemistry.^4,5–7The group of

Que has provided large insight into the formation^8,9,10of dinuclear highꢀvalent intermediates and

their applications¹¹in the oxidation and oxygenation of substrates. Using alkylated tpaꢀ

derivatives (Scheme 1a),¹⁰a highꢀvalent dinuclear complex with a central {Fe^III( ꢀO)₂Fe^IV}³⁺core

ꢀ

was obtained and characterized including singleꢀcrystal Xꢀray diffraction.⁹Later, linear

tetradentate ligands like BPMCN (Scheme 1a) have been employed, which correspond formally

to a substitution of a pyridine by a tertꢀamine,¹²which is more electronꢀdonating than pyridine

and has no πꢀacceptor capabilities.

Scheme 1. Ligand design and ligands used.

In our efforts to further stabilize such highꢀvalent dinuclear iron complexes we aim to increase

the electronꢀdonating ability of the ligand.⁵Therefore, we formally substituted the remaining

pyridine donors of BPMCN by strong

σꢀ and πꢀdonating phenolate donors resulting in the

ligands of the first generation H₂L^R(Scheme 1a)¹³and synthesized mononuclear¹⁴and

2

Page 3 of 42

Dalton Transactions

View Article Online

DOI: 10.1039/C5DT03711E

dinuclear^15,16complexes. The

ꢀ

ꢀoxo diiron complexes [(L^tꢀBu²)Fe^III( ꢀO)Fe^III(L^tꢀBu²)], [(L^Me²)Fe^III(

ꢀ ꢀꢀ

Cl

III

III Cl

2

O)Fe^III(L^Me²)], and [(L )Fe (

ꢀꢀO)Fe (L )] are squareꢀpyramidally coordinated with the ligand

2

adopting a transꢀconformation. [(L^tꢀBu²)Fe^III( ꢀO)Fe^III(L^tꢀBu²)] can be oxidized at +0.27 and +0.44 V

ꢀ

+

tꢀBu

III

vs Fc /Fc resulting in the monoꢀphenoxyl radical complex [( L ²)Fe (

ꢀ

ꢀO)Fe^III(L^tꢀBu²)]⁺and the

•

tꢀBu

III

tꢀBu 2+

diꢀphenoxyl radical complex [( L ²)Fe (

ꢀ

ꢀO)Fe ( L ²)] , respectively.¹⁶While these

•

oxidations are reversible on a CV timescale, both oxidized complexes decay into mononuclear

fragments by cleavage of an Fe^IIIꢀO bond with halfꢀlives τ_1/2at ꢀ40 °C of ~27 min for the

monocation and ~6 h for the dication.¹⁶

tꢀBu

III

tꢀBu 2+ 16

Although the diꢀphenoxyl radical complex [( L ²)Fe (

ꢀ

ꢀO)Fe ( L ²)]

accumulates the

•

same number of oxidation equivalents as a highꢀvalent {Fe^IV(

ꢀ

ꢀO)₂Fe^IV}⁴⁺species postulated for

Q,²we intend to improve our ligands in order (i) to enforce metalꢀcentered oxidations leading to

Fe^IVinstead of ligandꢀcentered oxidations and (ii) ꢀ more importantly ꢀ to prevent the decay of

the dinuclear complexes upon oxidation. Based on the results obtained from the iron complexes

of the first generation summarized above, our design for ligands of the second generation has

taken the following considerations into account:

(1) The generation of an {Fe(

ꢀꢀO)₂Fe} core requires that a tetradentate ligand leaves two

coordination sites cis to each other open for a bisꢀ

ꢀꢀoxo bridge. Although the tetradentate

ligands H₂L^Rof linear topology may coordinate in a cisꢀconformation, they prefer a transꢀ

conformation. As a tetradentate ligand of tripodal topology forces the two open coordination

sites to be cis to each other, a tripodal topology is preferred ((1) in Scheme 1b).

(2) To suppress dissociation of the dinuclear complexes either in their diferric forms or upon

oxidation into mononuclear species, the two tetradentate tripodal ligand compartments

should be covalently linked ((2) in Scheme 1b). This covalent linkage must not have a

bridging donor (e.g. phenolate or alcoxide) that would prevent the formation of a more

chargeꢀdonating (ꢀꢀO)₂bridge.

(3) The organic spacers between the donor atoms must be selected in order to obtain stable

complexes of sufficient flexibility for the intended core transformations. Thus, fiveꢀ

membered chelate rings are preferred to sixꢀmembered ones. Phenylene spacers would

provide some delocalization of the lone pairs of the donor atom onto the aromatic ring

reducing its chargeꢀdonation ability. Thus, ethylene spacers are indicated ((3) in Scheme

1b). The feasibility of this basic structural motive has been corroborated by simple modeling

attempts in addition to the success of Wieghardt’s dinucleating ligand¹⁷based on two 1,4,7ꢀ

triazacyclononane rings that are covalently linked by an ethylene spacer.

(4) Donor atoms X bearing more than one organic substituent should be tertꢀamines. The

remaining terminal donor atoms X bearing only one organic substituent should constitute

3

Dalton Transactions

Page 4 of 42

View Article Online

DOI: 10.1039/C5DT03711E

strongly σꢀ and πꢀelectron donating groups. Biomimetic carboxylates should be well suited

in order to obtain a redoxꢀinnocent ligand ((4) in Scheme 1b).

Thus, this ligand design leads to the bis(tetradentate) dinucleating ligand H₄julia (Scheme 1c).

Considering the stabilization of highꢀvalent {Fe(ꢀꢀO)₂Fe} cores by the group of Que using

substituted tpa ligands, the ligand susan with terminal pyridine donors (Scheme 1c) also seems

to be a reasonable target as a dinucleating tpa version (although susan bears two more electron

donating tertꢀamine donors in comparison to two pyridine donors in tpa ligands). Moreover, in

order to make comparisons to the ligands and complexes of the first generation, we envision a

ligand of the second generation with terminal phenolate donors (H₄hilde^Me², Scheme 1c),

although this ligand provides some sixꢀmembered chelate rings.

These new ligands belongs to the class of dinucleating ligands that was introduced in the early

1970s by Robson et al.¹⁸and Kida et al.¹⁹. Most of the dinucleating ligands reported provide a

bridging phenolate or alcoholate.²⁰The success of tpaꢀderived ligands in biomimetic chemistry

lead to the development of dinucleating ligands with two tpaꢀlike compartments^6,7,21and related

ligands systems²²that do not provide a bridging donor. In this respect, it should be noted that

during our efforts the group of Schindler reported a ligand related to susan with a central

propylene instead of the ethylene spacer to stabilize peroxoꢀbridged Cu^IIcomplexes.²³

2

Herein, we describe the synthesis of the ligands of the second generation H₄julia, susan, and

H₄hilde^Meand their dinuclear diiron complexes. Their structural, spectroscopic, magnetic,

2

electrochemical, and catalytic properties are presented and differences in these properties are

explained by to the variations of the terminal donor groups. The first application using H₄julia

ꢀ

affording [(julia){Fe^III(H₂O)( ꢀO)Fe^III(H₂O)}]•6H₂O has recently been communicated.²⁴

Experimental Section

Preparations of Compounds.

Solvents and starting materials were of the highest commercially available purity and used as

received except for CH₂Cl₂, THF, and diisopropylethylamine, which were dried according to

standard procedures.²⁵All reactions were performed using standard Schlenkꢀconditions under

an argon atmosphere. The synthesis of 1 were performed by using modified literature

procedure.²⁶The preparation of [(julia){Fe^III(H₂O)( ꢀO)Fe^III(H₂O}]•6H₂O and [(julia){Fe(DMSO)(µꢀ

ꢀ

O)Fe(DMSO)}] were performed according to the published procedure.²⁴The assignments of the

NMR resonances are based on 2D COSY, HMBC, and HMQC measurements. The numbering

scheme used is provided with the NMR spectra in the Supporting Information (Figs. S1ꢀS10).

4,7-Dimethyl-1,1,10,10-tetra(tert-butyloxycarboxymethyl)-1,4,7,10-tetraazadecane (3).

A

solution of tertꢀbutyl bromoacetate 2 (6.76 g, 34.7 mmol) in CH₂Cl₂(60 mL) was added dropwise

at 0 °C to a solution of tetramine 1 (1.51 g, 8.66 mmol) and diisopropylethylamine (7.28 g, 56.3

4

Page 5 of 42

Dalton Transactions

View Article Online

DOI: 10.1039/C5DT03711E

mmol). After stirring this reaction mixture for 16 h at room temperature, all volutiles were

removed in vacuum causing the continuous deposition of diisopropylethylammonium bromide.

The residue was suspended in ethyl acetate (50 mL) and filtered. The filtrate was concentrated

to 5 mL and purified by column chromatography (basic Al₂O₃, ethyl acetate). After removing the

solvent, the residue was dissolved in pentane (15 mL) and washed with NaOH (pH = 9, 3×10

mL). The aqueous phases were extracted with pentane. The combined organic phases were

dried with MgSO₄and the solvent was removed under vacuum resulting in a yellow oil that was

stored under an argon atmosphere. Yield: 1.89 g (35%). ¹H NMR (500.25 MHz, CDCl₃)

δ

/ ppm

3

= 3.43 (s, 8H; H5), 2.80 (t, J = 7.5 Hz, 4H; H3), 2.51 (t, J = 7.5 Hz, 4H; H2), 2.47 (s, 4H, H1),

2.22 (s, 6H; H4), 1.42 (s, 36H; H8). ¹³C NMR (125.80 MHz, CDCl₃)

δ

/ ppm = 170.8 (C6), 80.9

ꢀ1

~

(C7), 56.8 (C2), 56.4 (C5), 56.0 (C1), 51.7 (C3), 42.9 (C4), 28.3 (C8). IR (KBr):

ν

/ cm = 2977

s, 2934 s, 2844 m, 2802 m, 1734 s, 1458 m, 1392 m, 1368 s, 1287 m, 1253 m, 1219 m, 1149 s,

1068 w, 1035 w, 992 w, 938 w, 9915 w, 853 w, 778 w, 752 w. MALDIꢀTOF MS: m/z 631.5 [M +

H]⁺. Found: C, 60.49; H, 9.87; N, 8.62. Calc. for C₃₂H_62.5N₄O_8.25(3•0.25H₂O): C, 60.49; H, 9.92;

N, 8.82%.

K₄julia. Solid KOH (329 mg, 5.86 mmol) was added to a solution of tetraester 3 (811 mg, 1.29

mmol) in a mixture of MeOH (11.5 mL) and H₂O (1.3 mL). After heating the reaction solution at

reflux for two days, all volatiles were removed in vacuum. EtOH was added dropwise to the

residue until a homogenous oil was formed, which allowed for stirring. Addition of CH₃CN (5 mL)

causes the deposition of a hydroscopic offꢀwhite powder that was stored under an argon

1

atmosphere. Yield: 433 mg (60%). H NMR (500.25 MHz, D₂O, chemical shifts referenced

relative to EtOH as internal standard (

δ (CH₃) = 1.17 ppm)) δ / ppm = 3.22 (s, 8H; H5), 2.75ꢀ

2.72 (m, 4H; H3), 2.59ꢀ2.56 (m, 4H; H2), 2.57 (s, 4H, H1), 2.23 (s, 6H; H4). ¹³C NMR (125.78

MHz, D₂O, chemical shifts referenced relative to EtOH as internal standard (

ppm))

δ

(CH₃) = 17.47

~

δ

/ ppm= 180.1 (C6), 59.2 (C5), 54.8 (C2), 54.3 (C1), 51.6 (C3), 42.1 (C4). IR (KBr):

ν

/

cm^ꢀ1= 3398 s br, 2963 m, 2810 m, 1587 s, 1404 s, 1331 m, 1256 w, 1134 w, 1063 w, 1017 w,

995 w, 920 w, 899 w, 837 w, 772 w, 718 m, 642 w. ESIꢀMS m/z 597.0 [K₅julia]⁺, 559.1

[HK₄julia]⁺. Found: C 30.76; H 4.89; N 8.62. Calc. for C₁₆H_32.2N₄O_11.2

K_4.2(K₄julia•3H₂O•0.2KOH):

C 30.79; H 5.20; N 8.98%.

H₄julia•4HCl. A mixture of the tetraester 3 (1.89 g, 3.00 mmol) and concentrated hydrochloric

acid (9 mL) was stirred at room temperature for two days. Removal of all volatiles in vacuum

1

resulted in a white solid. Yield: 1.65 g (100%). H NMR (500.15 MHz, D₂O, chemical shifts

referenced relative to acetone as internal standard (

3.77 (s, 8H; H5), 3.47 (t, ³J = 5.8 Hz, 4H; H2), 3.31 (t, ³J = 5.8 Hz, 4H; H3), 3.03 (s, 6H; H4). ¹³C

NMR (125.78 MHz, D₂O, chemical shifts referenced relative to acetone as internal standard (

(CH₃) = 30.89 ppm)) / ppm = 175.1 (C6), 56.6 (C5), 56.1 (C2), 51.4 (C1), 50.5 (C3), 41.4 (C4).

δ = 2.22 ppm)) δ / ppm = 3.81 (s, 4H; H1),

δ

5

Dalton Transactions

Page 6 of 42

View Article Online

DOI: 10.1039/C5DT03711E

IR (KBr):

ν

/ cm^ꢀ1= 3368 m br, 2967 s, 2860 w, 2652 w, 2581 m, 2527 s, 2495 s, 2470 s, 2452

~

s, 1765 s, 1731 vs, 1477 m, 1466 m, 1439 m, 1422 s, 1215 s, 1158 s, 1099 w, 1039 w, 956 w,

907 w, 851 w, 714 w, 670 w, 568 w, 510 w. HRꢀESIꢀMS (CH₃CN) for C₁₆H₃₁N₄O₈[H₅julia]⁺: m/z

calc.: 407.21364, found: 407.21396. Found: C 32.49; H 6.54; N 8.79. Calc. for C₁₆H₃₄N₄O₈Cl₄

(H₄julia•4HCl•2.5H₂O): C 32.17; H 6.58; N 9.38 %.

susan. Solid NaBH(OAc)₃(154 g, 726 mmol, 6.1 equivalents) was added to a solution of

tetramine 1 (20.7 g, 119 mmol, 1 equivalent) and aldehyde 4 (48 mL, 504 mmol, 4.20

equivalents) in dichlorethane (500 mL). After stirring the reaction mixture for 6 h, the reaction

was quenched by addition of aqueous NaOH solution (2 M, 800 mL). The layers were separated

and the aqueous phase was extracted with CH₂Cl₂(6 × 200 mL). The combined organic

fractions were washed with brine (80 mL) and dried (Na₂SO₄). Removal of solvent under

reduced pressure afforded the crude product, which was purified by column chromatography

(basic aluminium oxide, THF) providing the pure ligand as a yellow oil that was stored under an

argon atmosphere. Yield: 51.9 g (81%). ¹H NMR (500.05 MHz, CDCl₃)

δ

/ ppm = 8.47 (dd, ³J =

4.9 Hz, ⁴J = 1.8 Hz, 4H; H10), 7.59 (ddd, ³J = 7.6 Hz, ³J = 7.6 Hz, ⁴J = 1.8 Hz, 4H; H8), 7.48 (dd,

³J = 7.8 Hz, ⁴J = 1.3 Hz, 4H; H7), 7.08 (ddd, ³J = 4.9 Hz, ³J = 7.4, ⁴J = 1.3 Hz, 4H; H9), 3.80 (s,

8H; H5), 2.64 (dd, ³J = 8.8 Hz, ³J = 5.7 Hz, 4H; H4), 2.50 (dd, ³J = 8.5 Hz, ³J = 5.9 Hz, 4H; H3),

2.34 (s, 4H; H1), 2.09 (s, 6H; H2). ¹³C NMR (125.75 MHz, CDCl₃)

δ / ppm = 159.8 (C6), 149.0

(C10), 136.4 (C8), 122.9 (C7), 122.0 (C9), 60.9 (C5), 56.1 (C1), 55.9 (C3), 52.1 (C4), 43.0 (C2).

/ cm^ꢀ1= 3379 w, 3053 m, 3009 m, 2945 m, 2810 s, 1589 s, 1570 s, 1472 s, 1433 s,

~

IR (KBr):

ν

1362 m, 1148 w, 1117 m, 1094 w, 1047 m, 995 m, 762 s. ESIꢀMS (MeOH) m/z 539.4 [M + H]⁺,

561.4 [M + Na]⁺Found: C 70.96; H 7.87; N 20.63. Calc. for C₃₂H_42.4N₈O_0.2(susan•0.2H₂O): C

70.87;H 7.88; N 20.66%.

H₄hilde^Me². A solution of tetramine 1 (400 mg, 2.30 mmol, 1 equivalent) in MeOH (8 mL) was

added at 0 °C to a solution of formaldehyde (37% in H₂O, 1.02 g, 12.6 mmol, 5.5 equivalents) in

MeOH (4 mL) and H₂O (4 mL). After stirring for 5 minutes, a solution of phenol 5 (1.12 g,

9.17 mmol, 4 equivalents) in MeOH (8 mL) was added dropwise at 0 °C. After heating the

reaction mixture for 2 days at reflux, during which a white precipitate formed, this was filtered

1

off, washed with cold MeOH and dried under reduced pressure. Yield: 1.59 g (97%). H NMR

(500.25 MHz, CDCl₃)

1.6 Hz, 4H; H11), 3.53 (s, 8H; H5), 2.76 (s, 4H; H1), 2.61 (m, 4H; H2), 2.56 (m, 4H; H3), 2.25 (s,

6H; H4), 2.18 (12H; H13), 2.16 (12H; H12). ¹³C NMR (125.75 MHz, CDCl₃)

/ ppm = 152.5

δ

/ ppm = 9.08 (br s, 4H; OH), 6.83 (d, ⁴J = 1.6 Hz, 4H; H9), 6.63 (d, ⁴J =

δ

(C7), 131.3 (C9), 128.4 (C11), 127.6 (C10), 125.3 (C8), 121.6 (C6), 56.4 (C5), 55.3 (C2), 54.4

/ cm^ꢀ1= 3400 ꢀ 2500 m br, 3011

~

(C1), 49.2 (C3), 41.7 (C4), 20.5 (C13), 16.3 (C12). IR (KBr):

ν

m, 2962 m, 2912 s, 2855 s, 2826 m, 2733 w, 1612 m, 1487 s, 1452 m, 1439 m, 1425 w, 1375

m, 1360 m, 1337 w, 1306 m, 1275 m, 1256 m, 1221 s, 1157 m, 1142 m, 1113 m, 1094 w, 1030

6

Page 7 of 42

Dalton Transactions

View Article Online

DOI: 10.1039/C5DT03711E

m, 997 m, 951 w, 922 w, 860 m, 820 w, 770 m, 743 w, 675 w, 660 w, 573 w. ESIꢀMS (CH₂Cl₂)

m/z 511.5 [M + H]⁺. Found: C 74.38; H 8.84; N 7.84. Calc. for C₄₄H₆₂N₄O₄(H₂hilde^Me²): C 74.33;

H 8.79; N 7.88%.

[(julia){Fe(OH₂)(µ-O)Fe(OH₂)}]•7H₂O. A solution of K₄julia (80 mg, 0.14 mmol) in MeOH (10

mL) was treated with solid Fe(ClO₄)₃•10H₂O (153 mg, 0.286 mmol) resulting in a brownishꢀ

yellow suspension. The precipitate was filtered off, redissolved in a mixture of acetone (10 mL)

and H₂O (6 mL), and kept at ꢀ20 °C. This causes the deposition of a few red crystals suited for a

ꢀ1

~

singleꢀcrystal Xꢀray diffraction study. Yield: 4 mg (4%). IR (KBr):

ν

/ cm = 3443 s br, 3270 m

br, 2978 m, 2928 m, 2911 m, 1645 s, 1620 s, 1454 w, 1387 s, 1331 m, 1317 m, 1261 w, 1236

w, 1192 w, 1150 w, 1115 w, 1090 w, 1003 w, 970 w, 957 w, 914 m, 905 m, 882 s, 837 m, 822

m, 777 w, 760 w, 731 w, 640 w, 594 w, 569 w, 503 w.

[(susan){FeCl(µ-O)FeCl}]Cl₂. Under waterꢀ and dioxygenꢀfree conditions, a solution of susan

(324 mg, 0.601 mmol, 1 equiv.) in THF (22 mL) was added to a suspension of FeCl₂(168 mg,

1.32 mmol, 2.2 equiv.) in EtOH (5 mL) resulting in an orange suspension. After stirring for 2 h at

room temperature, the solvent was removed under reduced pressure. The orange residue was

dissolved in CH₃CN (100 mL) under aerobic conditions, which resulted in a slow color change

from orange to dark green. After 2 days, diffusion of Et₂O resulted in the deposition of greenishꢀ

ꢀ1

~

black crystals. Yield: 272 mg (56%). IR (KBr):

ν

/ cm = 3422 s, 3339 m, 3244 m, 3071 w, 3032

w, 2905 w, 2872 w, 1603 s, 1568 m, 1476 s, 1447 m, 1431 s, 1292 s, 1016 s, 943 s, 814 s, 772

s. MSꢀMALDIꢀTOF: m/z = 368.0 [(susan){FeCl(µꢀO)FeCl}]²⁺. Found: C 45.34; H 5.37; N 13.36.

Calc. for [(susan){FeCl(µꢀO)FeCl}]Cl₂•2H₂O, C₃₂H₄₆N₈O₃Fe₂Cl₂: C 45.52; H 5.49; N 13.27%.

[(susan){FeCl(µ-O)FeCl_0.75(OCH₃)_0.25}](ClO₄)₂. A solution of susan (143 mg, 0.232 mmol, 1

equiv.) in acetone (20 mL) was added to a solution of Fe(ClO₄)₂•6H₂O (193 mg, 0.532 mmol,

2.00 equiv.) and TBACl (198 mg, 0.712 mmol, 2.68 equiv.) in MeOH (20 mL) resulting in a

yellow solution. Addition of triethylamine (0.3 mL, 2.15 mmol, 8.10 equiv.) causes a color

change to redꢀbrown. Diffusion of MTBE resulted in the deposition of deepꢀbrown crystals,

which were filtered off, washed three times with MTBE, and dried under reduced pressure.

ꢀ1

~

Yield: 140 mg (60%). IR (KBr):

ν

/ cm = 2927 w, 2871 w, 2790 w, 1608 m, 1473 m, 1433 m,

1296 w, 1093 s, 1017 m, 947 w, 817 s, 765 m, 623 s. ESIꢀMS (CH₃CN): m/z = 368.1

[(susan){FeCl(µꢀO)FeCl}]²⁺. Found: C 41.21; H 4.82; N 11.66. Calc. for [(susan){FeCl(µꢀ

O)FeCl_0.75(OCH₃)_0.25}](ClO₄)₂•0.5MeOH, C_32.25H_44.25N₈O_9.25Fe₂Cl_3.75: C 41.35; H 4.85; N 11.78%.

[(susan){FeCl(µ-O)FeCl}](ClO₄)₂. A solution of susan (125 mg, 0.232 mmol, 1 equiv.) in

acetone (20 mL) was added to a solution of Fe(ClO₄)₂•6H₂O (169 mg, 0.466 mmol, 2.01 equiv.)

and TBACl (146 mg, 0.525 mmol, 2.26 equiv.) in EtOH (20 mL) resulting in an orangeꢀbrown

solution. Addition of triethylamine (0.2 mL, 1.43 mmol, 6.18 equiv.) causes a color change to

brown. Diffusion of MTBE resulted in the deposition of deepꢀbrown crystals, which were filtered

7

Dalton Transactions

Page 8 of 42

View Article Online

DOI: 10.1039/C5DT03711E

off, washed three times with MTBE, and dried under reduced pressure. Yield: 123 mg (55%). IR

ꢀ1

~

(KBr):

ν

/ cm = 2927 w, 2875 w, 1608 m, 1473 m, 1433 m, 1296 w, 1095 s, 1016 m, 947 w,

822 s, 765 m, 6243 s. ESIꢀMS (CH₃CN): m/z = 368.2 [(susan){FeCl(µꢀO)FeCl}]²⁺. Found: C

40.70; H 4.66; N 11.66. Calc. for [(susan){FeCl(µꢀO)FeCl}](ClO₄)₂•0.5MeOH, C₃₂H₄₂N₈Fe₂Cl₄: C

41.05; H 4.52; N 11.97%.

[(hilde^Me){Fe(

µ-O)Fe}]. A solution of Fe(ClO₄)₃•10H₂O (118 mg, 0.222 mmol) in EtOH (5 mL)

2

was added to a suspension of the ligand H₄hilde^Me(79 mg, 0.111 mmol) in EtOH (5 mL)

resulting in a blue solution. Dropwise addition of a solution of triethylamine (112 mg, 1.11 mmol)

in EtOH (5 mL) leads to a red suspension, which was heated at reflux for 45 minutes. The

precipitate was filtered off and dried under reduced pressure. Yield: 54 mg (59%). Redꢀbrown

crystals suitable for a singleꢀcrystal Xꢀray diffraction study were obtained from slow diffusion of

2

ꢀ1

~

Et₂O into a CH₂Cl₂solution. IR (KBr):

ν

/ cm = 3437 w br, 3001 w, 2966 w, 2943 w, 2910 m,

2854 m, 2725 w, 1610 w, 1474 s, 1439 m, 1369 w, 1309 s, 1267 s, 1221 m, 1198 w, 1161 m,

1101 m, 1034 w, 1007 w, 989 w, 947 w, 928 w, 854 s, 795 s, 764 m, 746 w, 727 w, 691 w, 669

+

m, 602 w, 550 w, 534 w, 500 w. MSꢀMALDIꢀTOF: m/z = 835.1{[(hilde^Me){Fe(µꢀO)Fe}] + H} .

2

Found: C 58.96; H 6.73; N 6.11. Calc. for [(hilde^Me{Fe(µꢀO)Fe}]•CH₂Cl₂, C₄₅H₆₀N₄O₅Fe₂Cl₂: C

2

58.78; H 6.58; N 6.09%.

[(hilde^Me){FeCl}₂]. Slow diffusion of Et₂O via a long glass fitting (NS14/5, 30 cm) including a

2

glass frit into a solution of [(hilde^Me){Fe(µꢀO)Fe}] (144 mg, 0.157 mmol) in CH₂Cl₂(275 mL)

2

provided red crystals suited for a singleꢀcrystal Xꢀray diffraction study. Yield: 54 mg (32%). IR

ꢀ1

~

(KBr):

ν

/ cm = 3447 w br, 2997 w, 2918 m, 2860 w, 1639 m, 1472 s, 1439 m, 1373 w, 1308

m, 1252 s, 1221 w, 1161 m, 1099 m, 1012 w, 949 w, 932 w, 868 s, 826 m, 754 w, 608 m, 561

+

Me₂

+

w. MSꢀMALDIꢀTOF: m/z = 711.6 {H₅hilde^Me} , 764.5 {H₃(hilde )Fe} . Found: C 54.41; H 6.83;

2

N 5.66. Calc. for [(hilde^Me{FeCl}₂]•0.75CH₂Cl₂•2H₂O, C_44.75H_63.5N₄O₆Fe₂Cl_3.5: C 54.33; H 6.46; N

2

5.66%.

Catalytic C-H Hydroxylation.

The oxygenation reactions of cyclohexane (CyH) were carried out under air (unless stated

otherwise) as tripleꢀ or doubleꢀmeasurements in 25 mL roundꢀbottomed flasks at room

temperature. In a typical reaction, the catalyst (5 ꢀmol) was dissolved in CH₃CN (5 mL) and

CyH was added to the solution. The resulting emulsion was stirred for 15 minutes, when

homogeneity was reached. Then base (if applicable) and oxidant were added, the flask was

closed and the mixture was stirred for 18 h. An aqueous solution of Na₂S₂O₃was added to

quench the reaction and an exactly defined amount of cyclopentanone was added as an internal

standard. Due to the poor solubility of Na₂S₂O₃in CH₃CN, often a phase separation was

observed. The organic phase was filtered through a short silica column. The aqueous phase

was extracted with EtOAc (3 mL), which was used to elude the column. A small volume of the

8

Page 9 of 42

Dalton Transactions

View Article Online

DOI: 10.1039/C5DT03711E

resulting clear, homogenous organic solution was diluted with CH₃CN and used for GCꢀMS

analysis. GC analysis was carried out using a Shimadzu GC 2010 (Rxiꢀ624 Sil Ms fused silica

column, 30 m

× 0.25 mm inside diameter, 0.14 ꢁm film thickness) with helium carrier gas. The

injection volume was 0.5 ꢁL. The oven temperature was increased stepwise after injection. First

the temperature was held at 50 °C for 2 min, then it was raised to 120 °C at a rate of 20 °C min^ꢀ

¹. This temperature was held for 4 min. Afterwards, the temperature was raised to 200 °C at a

rate of 25 °C min^ꢀ1, where it was held for 2 min. The products were identified using a

Shimadzu GCMSꢀQP2010S. Both continuous scanning of m/z for qualitative analysis and

singleꢀion monitoring (SIM) of the five biggest ionization peaks for quantitative analysis were

used. The products were identified by comparison with commercially available compounds.

Calibration measurements were performed using the same protocol as described above. The

calibration curves used for quantification were determined with values of R²of

0.9996/0.9997/0.9998/0.9984

caprolactone.

for

cyclohexanol/cyclohexanone/chlorocyclohexane/εꢀ

X-Ray Crystallography.

The singleꢀcrystals were coated with oil and measured at low temperature on a Bruker Kappa

APEX II diffractometer (four circle goniometer with 4K CCD detector, MoꢀK radiation, focussing

α

graphite monochromator; ωꢀ and φꢀscans for [(susan){FeCl(µꢀO)FeCl}]Cl₂•2H₂O, [(susan)ꢀ

{FeCl(µꢀO)FeCl_0.75(OCH₃)_0.25}](ClO₄)₂•0.5MeOH, and [(susan){FeCl(µꢀO)FeCl}](ClO₄)₂•0.5EtOH)

or on a Bruker AXS SMART diffractometer (three circle goniometer with 1K CCD detector, Moꢀ

radiation, graphite monochromator for [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]•7H₂O, [(hilde^Me){Fe(µꢀ

2

K

α

O)Fe}]•CH₂Cl₂, and [(hilde^Me){FeCl}₂]•2CH₂Cl₂). Empirical absorption corrections using equivaꢀ

2

lent reflections were performed with the programs SADABS 2.10²⁷([(julia){Fe(OH₂)(µꢀO)Feꢀ

Me₂

(OH₂)}]•7H₂O, [(hilde^Me){Fe(µꢀO)Fe}]•CH₂Cl₂, [(susan){FeCl(µꢀO)FeCl}]Cl₂•2H₂O, [(hilde )ꢀ

2

{FeCl}₂]•2CH₂Cl₂), SADABS 2008/1²⁷([(susan){FeCl(µꢀO)FeCl_0.75(OCH₃)_0.25}](ClO₄)₂•0.5MeOH),

and TWINABS 2012/1 ([(susan){FeCl(µꢀO)FeCl}](ClO₄)₂•0.5EtOH). The structures were solved

with the program SHELXSꢀ97²⁸and refined using SHELXLꢀ97.²⁸Crystal data and details

concerning data collection and structure refinements are given in Table 1. [(julia){Fe(OH₂)(µꢀ

O)Fe(OH₂)}]•7H₂O, which crystallizes in a nonꢀcentrosymmetric space group, was refined as a

racemic twin. In [(susan){FeCl(µꢀO)FeCl_0.75(OCH₃)_0.25}](ClO₄)₂•0.5MeOH) a disorder of a Cl^ꢀ

ligand with an MeO^ꢀligand at the atom Fe2 was refined in a ratio of 0.75/0.25. The solvent

MeOH molecule is disordered around a crystallographic center of inversion. [(susan){FeCl(µꢀ

O)FeCl}](ClO₄)₂•0.5EtOH was refined as a nonꢀmerohedral twin (domain size ratio = 0.54/0.46).

The second domain was rotated by 180^oabout ꢀꢁ^*. The solvent EtOH molecule is disordered

around a crystallographic center of inversion. The supplementary crystallographic data for this

paper can be obtained free of charge from The Cambridge Crystallographic Data Centre (12

9

Dalton Transactions

Page 10 of 42

View Article Online

DOI: 10.1039/C5DT03711E

Union Road, Cambridge, CB21EZ) via www.ccdc.cam.ac.uk/data_request/cif on quoting the

CCDC numbers from Table 1.

Other Physical Measurements.

Infrared spectra (400ꢀ4000 cm^ꢀ1) of solid samples were recorded on a Shimadzu FTIR 8300 or a

Bruker Vertex 70 as KBr disks. ESI mass spectra were recorded on a Bruker Esquire 3000 ion

trap mass spectrometer equipped with a standard ESI source or on a Bruker APEX III Fourier

transform ion cyclotron resonance mass spectrometer. MALDI TOF mass spectra were

1

recorded with a PE Biosystems Voyager^TMDE instrument. H and ¹³C NMR spectra were

measured on a Bruker DRX500 or a Bruker Avance III 300 spectrometer using the solvent as an

internal standard. UVꢀVisꢀNIR absorption spectra of solutions were measured on a Shimadzu

UVꢀ3101 PC spectrophotometer at ambient temperatures. Cyclic and squareꢀwave

voltammograms (CVs and SWs) were measured either by use of an EG&G

potentiostat/galvanostat 273A or by use of a VersaStat potentiostat on Arꢀflushed solutions

containing 0.2 M [NBu₄]PF₆as supporting electrolyte in a conventional electrochemical cell. The

counter electrode was a platinum wire and the reference electrode was Ag/0.01 M

AgNO₃/CH₃CN unless otherwise noted. Different working electrodes (GC, Pt, Au) were used.

The potentials are referenced versus the ferrocenium/ferrocene (Fc⁺/Fc) couple used as an

internal standard. SWs have been recorded with typical frequencies of 15 to 50 Hz with a GC

working electrode. CVs were routinely measured with scan rates of 100 or 200 mV s^ꢀ1. In order

to probe the timeꢀscale of an ECE mechanism, CVs with fast scan rates of 1ꢀ10 V s^ꢀ1using a Pt

microelectrode have been recorded. Magnetic susceptibility data were measured on powdered

samples in the temperature range 2ꢀ300 K by using a SQUID magnetometer (Quantum Design

MPMS XLꢀ7 EC) with a field of 1.0 T. The experimental data were corrected for underlying

diamagnetism by use of tabulated Pascal’s constants. The susceptibility data of the diferric

complexes were analysed on the basis of the usual spinꢀHamiltonian description for the

electronic ground state by using the simulation package julX.²⁹The Hamilton operator was:

r r

ˆ ˆ

v

ˆ

H =−2JS S +

g

ꢀ

_BS⋅B

(1)

∑

1

2

i

i=1,2

where J is the exchange coupling constant and g is the average electronic g value. Magnetic

moments were obtained from numerically generated derivatives of the eigenvalues of eq. 1, and

summed up over 16 field orientations along a 16ꢀpoint Lebedev grid to account for the powder

distribution of the sample. ⁵⁷Fe Mössbauer spectra were recorded on an alternating constantꢀ

acceleration spectrometer. The minimal lineꢀwidth was 0.24 mm s^ꢀ1fullꢀwidth at halfꢀheight. The

sample temperature was maintained constant in a bath cryostat (Wissel MBBCꢀHE0106).

⁵⁷Co/Rh was used as the radiation source. Isomer shifts were determined relative to

αꢀiron at

room temperature.

10

Page 11 of 42

Dalton Transactions

View Article Online

DOI: 10.1039/C5DT03711E

Results and Discussion

Ligand Synthesis

The three ligands were synthesized from tetramine 1 (Scheme 2), which was already described

in the literature.²⁶Two different synthetic routes for the synthesis of H₄julia starting from the

tetramine 1 seem feasible. Primary amines can be twice cyanomethylated under acidic

conditions,³⁰which is part of one synthesis of EDTA.³¹This route was followed by

Schwarzenbach et al.³²Due to the high toxicity of free HCN, which is formed by using cyanide

under acidic conditions, we preferred an alternative route to the ligand H₄julia. The reaction of

tetramine 1 with ester 2 (Scheme 2) resulted in the tetraester 3. The tetraester 3 can be cleaved

under alkaline conditions with KOH resulting in the salt K₄julia or under acidic conditions using

HCl yielding the protonated version H₄julia•4HCl (Scheme 2). The ligand susan was obtained by

the reductive amination³³of four equivalents of aldehyde 4 with tetramine 1 using NaBH(OAc)₃.

2

In analogy to the ligands of the first generation H₂L^R,¹⁴the ligand H₄hilde^Me²was synthesized by

a Mannich reaction of tetramine 1 with four equivalents of phenol 5.

Scheme 2. Synthesis of the ligands of the second generation.

Complex Synthesis

The reaction of H₄julia•4HCl with FeCl₃•6H₂O and NEt₃in water provided red singleꢀcrystals of

the hexahydrate [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]•6H₂O.²⁴On the other hand, the reaction of

K₄julia with Fe(ClO₄)₃•10H₂O in methanol provided a precipitate, which was redissolved in an

acetoneꢀwater mixture. Keeping this solution at ꢀ20 °C resulted in the deposition of a few red

singleꢀcrystals of the heptahydrate [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]•7H₂O. The complex (either

hydrate) is soluble in H₂O, MeOH, and DMF but not in other organic solvents. Dissolution in

11

Dalton Transactions

Page 12 of 42

View Article Online

DOI: 10.1039/C5DT03711E

DMSO at 60 °C followed by slow cooling to ambient temperature led to the deposition of red

crystals of [(julia){Fe(DMSO)(µꢀO)Fe(DMSO)}]•3DMSO, i. e. the coordinated water molecules

were substituted by DMSO molecules.²⁴

The reaction of susan with FeCl₂in THF under moistureꢀ and dioxygenꢀfree conditions provided,

after removal of the solvent, an orange solid, which was redissolved in acetonitrile under

ambient conditions.³⁴A slow color change to dark green was observed probably due to slow

oxidation by dioxygen. Diffusion of Et₂O provided singleꢀcrystals of [(susan){FeCl(µꢀ

O)FeCl}]Cl₂•2H₂O. As chloride counter anions may disturb electrochemical measurements as

well as chemical oxidation and oxygenation reactions, we intended to avoid chloride as

counterion. Thus, susan was reacted with Fe(ClO₄)₂•6H₂O and a slight excess of TBACl in

MeOH. Addition of NEt₃resulted in a redꢀbrown solution and diffusion of MTBE led to the

deposition of singleꢀcrystals that analyzed by singleꢀcrystal Xꢀray diffraction as [(susan){FeCl(µꢀ

O)FeCl_0.75(OCH₃)_0.25}](ClO₄)₂•0.5MeOH. Using the same stoichiometry reproducibly resulted in

singleꢀcrystals of this composition. One of the coordinated chloride positions is disordered with a

coordinated methoxy group in a 3:1 ratio. The reaction in EtOH provided deep brown singleꢀ

crystals of [(susan){FeCl(µꢀO)FeCl}](ClO₄)₂•0.5EtOH, which are not only soluble in H₂O and

MeOH, but also in organic solvents as CH₃CN and acetone.

16

In analogy to the synthesis of [(L^Me²)Fe^III(

ꢀ

ꢀO)Fe^III(L^Me²)], H₄hilde^Mewas reacted with

2

Fe(ClO₄)₃•10H₂O and NEt₃in EtOH affording the complex [(hilde^Me){Fe(µꢀO)Fe}] as a

2

microcrystalline solid, which is soluble in organic solvents as DMF, CH₂Cl₂, CHCl₃, THF, and

acetone, but not in H₂O, MeOH, or CH₃CN. Singleꢀcrystals of [(hilde^Me){Fe(µꢀO)Fe}]•CH₂Cl₂

2

were obtained by diffusion of Et₂O into a CH₂Cl₂solution. Interestingly, using a solution of

[(hilde^Me){Fe(µꢀO)Fe}] in CH₂Cl₂of the same concentration but slowing down the diffusion of

2

Et₂O results, after a longer time period, in singleꢀcrystals of [(hilde^Me²){FeCl}₂]•2CH₂Cl₂. The only

35

chloride source in this reaction is the solvent CH₂Cl₂indicating the decomposition of CH₂Cl₂

induced by the complex [(hilde^Me){Fe(µꢀO)Fe}].

2

Structural Characterization

Singleꢀcrystals of [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]•6H₂O,²⁴[(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]•7H₂O,

[(julia){Fe(DMSO)(µꢀO)Fe(DMSO)}]•3DMSO,²⁴

[(hilde^Me){Fe(µꢀO)Fe}]•CH₂Cl₂,

[(hilde^Me)ꢀ

2

{FeCl}₂]•2CH₂Cl₂, [(susan){FeCl(µꢀO)FeCl}]Cl₂•2H₂O, [(susan){FeCl(µꢀO)FeCl_0.75(OCH₃)_0.25}]ꢀ

(ClO₄)₂•0.5MeOH, and [(susan){FeCl(µꢀO)FeCl}](ClO₄)₂•0.5EtOH were analyzed by singleꢀ

crystal Xꢀray diffraction. The molecular structures (Fig. 1+2+S11) show, that all three ligands act

as intended as bis(tetradentate) donors.

The tetraꢀanion (julia)^4ꢀcoordinates a central {Fe^III( ꢀO)Fe^III}⁴⁺core resulting in neutral

ꢀ

complexes. The iron ions are sixꢀcoordinate by coordination of water or DMSO. The asymmetric

units in [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]•6H₂O and [(julia){Fe(DMSO)(µꢀO)Fe(DMSO)}]•3DMSO

12

Page 13 of 42

Dalton Transactions

View Article Online

DOI: 10.1039/C5DT03711E

consist of one complex, while only half of the complex is in the asymmetric unit of

[(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]•7H₂O, the second half is generated by a crystallographic C₂axis.

Me 4ꢀ

III

In [(hilde^Me){Fe(µꢀO)Fe}], the tetraꢀanion (hilde ²) also coordinates a central {Fe (

core but the iron ions are only fiveꢀcoordinate. The coordination environment is trigonalꢀ

bipyramidal with a strong distortion towards squareꢀpyramidal ( = 0.57). The asymmetric unit

ꢀ

ꢀO)Fe^III}⁴⁺

2

τ

consists of one half of the molecule, the other half is generated by a crystallographic C₂axis.

The crystal structure of [(hilde^Me){Fe(µꢀO)Fe}]•CH₂Cl₂contains a CH₂Cl₂molecule of

2

crystallization that is oriented by a hydrophobic binding pocket built from four methylbenzene

ꢀ

units below the {Fe^III( ꢀO)Fe^III}⁴⁺core (Fig. 3aꢀc). Remarkably, the CꢀH bonds of the CH₂Cl₂

molecule point directly to the redoxꢀactive central core. Thus, this crystal structure illustrates the

ability of this complex to preorganize anticipated substrates for CꢀH activation in a hydrophobic

binding pocket.

In [(hilde^Me){FeCl}₂], the iron ions are also fiveꢀcoordinate with a terminal chloride and the

2

coordination environment is slightly more towards trigonalꢀbipyramidal (τ = 0.68). There is also a

crystallographically imposed C₂symmetry but the CH₂Cl₂molecules of crystallization are placed

in hydrophobic channels in the crystal instead of in a molecular binding pocket.

ꢀ

In the complexes with the ligand susan, the neutral ligand susan coordinates a {Fe^III( ꢀO)Fe^III}⁴⁺

core. The sixꢀcoordinate iron ions are further coordinated with a chloride ion resulting in

dicationic complexes. The complexes in the two perchlorate salts exhibit no crystallographically

imposed symmetry, while in the chloride salt there is a crystallographically imposed C₂

symmetry. In the crystal structure of [(susan){FeCl(µꢀO)FeCl}]Cl₂•H₂O the two chloride

counteranions form a rhomb with two water molecules by OH•••Cl hydrogen bonds (Fig. 3d). In

[(susan){FeCl(µꢀO)FeCl_0.75(OCH₃)_0.25}](ClO₄)₂, a disorder at the sixth coordination site of Fe2

could be resolved with 75 % occupation of chloride (Cl2) and 25 % occupation of methanolate

(O98, C98). Although this disorder was identified first during the refinement of the diffraction

data, the methanolate coordination is clearly visible by the appearance of the characteristic lowꢀ

energy ν

(CꢀH) stretch of the coordinated methanolate at 2790 cm^ꢀ1, which is absent in the FTIR

spectra of [(susan){FeCl(µꢀO)FeCl}]Cl₂and [(susan){FeCl(µꢀO)FeCl}](ClO₄)₂(Fig. S13b).

Full lists of interatomic distances and angles are provided in Tables S1ꢀS6, while selected

interatomic distances and angles together with their mean values are summarized in Table 2.

The bond distances in the first coordination spheres are all in accordance with expected values.

The short FeꢀOPh distances of 1.84 Å in [(hilde^Me²){FeCl}₂] demonstrate the strong covalency of

the Fe^IIIꢀOPh bond. Substitution of the chlorides by a stronger electronꢀdonating

ꢀꢀoxo donor in

[(hilde^Me){Fe(µꢀO)Fe}] increases the electron density at the iron ions leading to an increase of

the FeꢀOPh bond lengths to 1.88 Å. The longer FeꢀO^carbbonds of 2.02 ꢀ 2.03 Å in the (julia)^4ꢀ

complexes reflect not only the lower covalency of an FeꢀO^carbbond vs an FeꢀOPh bond but also

2

13

Dalton Transactions

Page 14 of 42

View Article Online

DOI: 10.1039/C5DT03711E

the increase of the coordination number. The FeꢀN^pybond lengths of the pyridines in

[(susan){FeCl(µꢀO)FeCl}]²⁺trans to the

ꢀoxo ligand are slightly elongated with 2.23 Å

compared to the FeꢀN^pybond lengths of the pyridines cis to the

ꢀoxo ligand at 2.16 ꢀ 2.17 Å.

ꢀ

Hydrogen-Bonding in the Crystal Structures of [(julia){Fe(OH₂)(

[(julia){Fe(OH₂)( -O)Fe(OH₂)}]•7H₂O

µ-O)Fe(OH₂)}]•6H₂O and

µ

The complexes [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}] are embedded with hydrogenꢀbonded water

molecules in both crystal structures (Fig. 4a+b+S12). For [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]•6H₂O,

even the hydrogen atoms of the lattice water molecules were found in the difference Fourier

map and refined independently without constraints due to the high quality of the data set.

Hydrogen bonds in the secondary coordination sphere of the active sites in metalloproteins and

model complexes are known to influence the electronic structure and reactivities.^36–41The

coordinated water molecules are involved in two hydrogen bonds as hydrogen bond donors.

One of the two coordinated carboxylate oxygen atoms per iron centre is a hydrogen bond

acceptor, while the other is not involved in hydrogen bonding. The nonꢀcoordinating carboxylate

oxygen atoms are involved in one hydrogen bond except O66 in [(julia){Fe(OH₂)(µꢀ

O)Fe(OH₂)}]•6H₂O, which is involved in two hydrogen bonds (Fig. 4a). This difference leads to

slight but significant variations in the bond lengths. The C=O bond length d(C66ꢀO66) =

1.2453(11) Å is increased while the CꢀO bond length d(C66ꢀO42) = 1.2731(11) Å is decreased

̅

ꢂ(CꢀO) =

in comparison to the mean values of the other carboxylates:

ꢂ

(C=O) = 1.230 Å and

1.289 Å. These variations are reflected in the FTIR spectra (Fig. S13a). The complexes of

(julia)^4ꢀexhibit strong absorptions in the range 1625ꢀ1640 cm^ꢀ1and prominent bands of lower

ꢀ

intensity in the range 1360ꢀ1390 cm^ꢀ1(Table 3), which are characteristic for the

ν

_as(CO₂) and

ꢀ

ν

_s(CO₂) stretches, respectively.⁴²In addition, [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]•6H₂O exhibits well

resolved absorptions at 1591 and 1399 cm^ꢀ1of lower intensity, which have no counterparts in

the other complexes. These two additional bands correspond to the carboxylate with two

hydrogenꢀbonds (weakened C=O bond and a strengthened CꢀO bond), while the stronger

bands belong to the carboxylates with one hydrogen bond.

These variations also affect the FeꢀO^carbbond. The carboxylate donors with one hydrogen bond

to the coordinated oxygen donor have bond lengths of d(Fe1ꢀO1) = 2.0366(7) Å and d(Fe2ꢀ

O41) = 2.0307(7) Å. These are longer compared to d(Fe1ꢀO2) = 1.9994(7) Å of the carboxylate

with no hydrogen bond to the coordinated oxygen donor. A special case is the Fe2ꢀO42 bond

with 2.0273(7) Å, which possesses no hydrogen bond to the coordinated oxygen donor. The

expected shortening of the FeꢀO bond is compensated by the two hydrogen bonds to the

uncoordinated oxygen donor (O66), which weakens the C=O bond, strengthens the CꢀO bond,

and thus weakens the FeꢀO^carbbond. Thus, subtle changes of hydrogen bonding to the second

coordination sphere are able to influence the bond distances of the first coordination sphere.

14

Page 15 of 42

Dalton Transactions

View Article Online

DOI: 10.1039/C5DT03711E

Analysis of Ligand Conformations

Although the three ligands are of the same topology, they coordinate the central {Fe(µꢀO)Fe}

core with different conformations. In order to analyse these variations, we have chosen the

three C₂symmetric complexes in crystals of [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]•7H₂O,

[(hilde^Me){Fe(µꢀO)Fe}]•CH₂Cl₂, and [(susan){FeCl(µꢀO)FeCl}]Cl₂•2H₂O as representative

2

examples. A mononucleating tripodal ligand can adopt only one conformation in an octahedrally

coordinated complex with the two open coordination sites in cis position. The covalent

connection of two tripodal ligand units with the topology of our ligands of the second generation

results in various potential conformations, from which only two have been observed for the

cornerꢀsharing bisoctahedra described herein (Fig. 4c). In the cis conformation, the two open

coordination sites X are in a cis orientation relative to the Feꢀ

nitrogen donors of the tetramine backbone to be in one plane with the Feꢀ

idealized cornerꢀsharing bisoctahedron; the tripodal nitrogen donors N^tare trans to the

bridge. In the trans conformation, the two open coordination sites X are in a trans orientation

relative to the Feꢀ ꢀoxo bridge and

ꢀOꢀFe vector. The tripodal nitrogen donors N^tare cis to the

two of the terminal donors T are trans to the ꢀoxo bridge. This trans conformation (highest

ꢀ

ꢀOꢀFe vector. This forces the four

ꢀOꢀFe vector in an

ꢀoxo

ꢀ

possible symmetry C₂) is intrinsically chiral, while the cis conformation (highest possible

symmetry C_2v) is not. The ligand (julia)^4ꢀcoordinates in the cis conformation, while the ligand

susan coordinates in the trans conformation. It is apparent that two pyridine donors in the susan

complexes are perfectly oriented for

πꢀπ interactions at distances of 3.4ꢀ3.7 Å (Fig. 2). This

might lead to an energetic preference for the trans conformation in the susan complexes.

Although the complex [(hilde^Me){Fe(µꢀO)Fe}] possesses only fiveꢀcoordinate metal ions, its

2

topology is related to the cis conformation.

The molecular structures presented here are distorted from an idealized cornerꢀsharing

bisoctahedron. Most obviously, the FeꢀOꢀFe units are not linear but bent with the ꢀꢀoxo donor

pointing away from the bridging ethylene diamine unit. This bending is most pronounced in the

case of the fiveꢀcoordinate iron atoms in [(hilde^Me){Fe(µꢀO)Fe}] with an angle of 160.8° (Table

2

2). Moreover, the two octahedra are rotated along the FeꢀFe vector from an idealized eclipsed

conformation towards a staggered conformation (Fig. 5 right). This rotation can be

parameterized by several parameters (Table 2): (i) the torsion angle

angle between the best planes formed by the basal coordination planes perpendicular to the

FeꢀN^cvector, and (iii) the angle

between the vectors through the two donor atoms

perpendicular to the FeꢀN^cand FeꢀO^oxovectors. In an ideal bisoctahedron that is only rotated

θ

(N^cꢀFeꢀFeꢀN^c), (ii) the

α

β

along the FeꢀFe vector, all three angles would have the same value. The torsion angle

θ is in

the range 40ꢀ44° for the (julia)^4ꢀand susan complexes but 60° in [(hilde^Me){Fe(µꢀO)Fe}].

2

However, the angles

α

and

β

are larger than

θ

in the (julia)^4ꢀcomplexes, while

α

and

β

are

15

Dalton Transactions

Page 16 of 42

View Article Online

DOI: 10.1039/C5DT03711E

in the susan complexes. The strong variations in these angles in [(hilde^Me){Fe(µꢀ

2

smaller than

θ

O)Fe}] reflect the trigonal bipyramidal coordination mode.

These different coordination modes of the ligands of the second generation imply a strong

flexibility of the ligand backbone. This is strikingly demonstrated by the relative orientation of the

CꢀC bond of the central ethylene diamine unit to the FeꢀFe vector (Fig. 5 left), quantified by

angles of only 4.0° in [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}] but 79.6° in [(hilde^Me){Fe(µꢀO)Fe}] and

2

70.6° in [(susan){FeCl(µꢀO)FeCl}]Cl₂.

IR Spectroscopy

Of special interest in the FTIR spectra (Fig. S13a) are the

was established that _as(FeꢀOꢀFe) correlates with the FeꢀOꢀFe angle in the range 180ꢀ120°.

However, only a minor variation of _as(FeꢀOꢀFe) has been described for the variation of the Feꢀ

OꢀFe angle varying from 180ꢀ160° observed here.⁴³Table 3 and Fig. S13c show the

ν_as(FeꢀOꢀFe) stretches (Table 3). It

ν

̅

ν_as(FeꢀOꢀFe) on the mean distance ꢂ(FeꢀO). Interestingly, the terminal donors

dependence of

have a stronger impact on

ligand (julia)^4ꢀpossess the energetically highest

̅

ꢂ(FeꢀO) bond length. The complexes of the

ν

_as(FeꢀOꢀFe) than the

_as(FeꢀOꢀFe) stretches, which decrease slightly

̅

ꢂ(FeꢀO). The ν_as(FeꢀOꢀFe) stretches of the complexes of the ligand

ν

in energy with increasing

susan are energetically lower than those of the ligand (julia)^4ꢀand are in the range of the fiveꢀ

coordinate complexes of the ligands of the first generation. In addition to the influence of the

terminal ligands on

ν_as(FeꢀOꢀFe), the coordination polyhedron also has an impact. The

ν_as(FeꢀOꢀ

Fe) stretch of [(L^Me²)Fe^III(

ꢀ

ꢀO)Fe^III(L^Me²)] with a squareꢀpyramidal coordination environment is at

817 cm^ꢀ1and decreases to 795 cm^ꢀ1in [(hilde^Me){Fe(µꢀO)Fe}], which exhibits a strong distortion

2

towards trigonalꢀbipyramidal.

Magnetic Properties

The temperatureꢀdependence of the effective magnetic moments,

ꢀ_eff, of the ꢀꢀoxoꢀdiferric

complexes show the wellꢀknown behaviour of strongly antiferromagnetically coupled S₁=S₂=5/2

spin systems (Fig. S14). Several magnetoꢀstructural correlations were proposed for the

ꢀ

dependence of J with structural parameters in {Fe^III( ꢀO)Fe^III}⁴⁺complexes.⁴⁴Table 3 compares

the J values with selected structural parameters but a simple correlation is not apparent. The

sixꢀcoordinate complexes of susan and (julia)^4ꢀare more strongly coupled than those of the fiveꢀ

Me 4ꢀ

R

2ꢀ

Ph

2

coordinate complexes of (hilde ) and (L ) This is attributed to the strong FeꢀO covalency

of the latter that is compensated by a lower FeꢀO^oxocovalency, which is reflected in

_as(FeꢀOꢀ

(FeꢀO). This might be due to the differences in the coordination number. The

(FeꢀO) in [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}] than in [(julia){Fe(DMSO)(µꢀO)Fe(DMSO)}] is

consistent with a higher _as(FeꢀOꢀFe) and a stronger exchange coupling J. Another interesting

comparison is that of [(hilde^Me){Fe(µꢀO)Fe}] with its first generation analog [(L ²)Fe (

ν

̅

ꢂ

Fe) but not in

̅

ꢂ

shorter

ν

Me

III

2

ꢀꢀ

16

Page 17 of 42

Dalton Transactions

View Article Online

DOI: 10.1039/C5DT03711E

O)Fe^III(L^Me²)]. The latter has a higher

ν

_as(FeꢀOꢀFe) correlated with a stronger coupling J,

̅

ꢂ(FeꢀO) distances are almost the same. This might be attributed to the difference

although the

in the coordination geometry: squareꢀpyramidal in the first generation and trigonalꢀbipyramidal

in the second generation complex. This demonstrates that the specific electronic structures of

the h.s. Fe^IIIions imposed by the non(

ꢀꢀoxo) ligands influence J and are not necessarily

̅

ꢂ(FeꢀꢀꢀO).

reflected in

Mössbauer Spectroscopy

⁵⁷Fe Mössbauer spectra (Fig. S15, Table 3) of the

ꢀꢀoxo complexes consist of sharp quadrupole

doublets, while the linewidth of [(hilde^Me){FeCl}₂] is broadened, consistent with a slower

2

electronic relaxation due to the absence of an efficient spinꢀspin relaxation pathway. It should

be noted that the large quadrupole splitting of 1.14 mm s^ꢀ1for a complex without a

reflects the high covalency of the Fe^IIIꢀOPh bonds.¹⁴The isomer shift of the {Fe^III(

complexes decreases slightly from (julia)^4ꢀto susan and (hilde^Me²)^4ꢀ. This is attributed to an

ꢀꢀoxo bridge

δ

ꢀ

ꢀO)Fe^III}⁴⁺

increase of the overall covalency. The lower covalency of a chloride vs a

ꢀꢀoxo ligand is

ꢀ1

demonstrated by the decrease from 0.47 mm s^ꢀ1in [(hilde^Me){FeCl}₂] to 0.44 mm s in

2

[(hilde^Me){Fe(µꢀO)Fe}].

2

ꢀ

ꢀOxo diferric complexes are known for large quadrupole splittings |

strong contribution of the small dianion O^2ꢀion at short distance to the electric field gradient

(highly covalent E_Qin

ꢀOꢀFe bond: anisotropic covalency).⁴⁵The large quadrupole splittings

the (julia)^4ꢀcomplexes reflect the anisotropic covalency imposed by the strong

ꢀOꢀFe bond,

which dominates the electric field gradient. The more covalent ꢀbackbonding pyridine donors in

ꢁE_Q| originating from the

ꢀ



ꢁ



ꢀ

π

the susan complexes perpendicular to the FeꢀOꢀFe unit increases the electron density in this

direction. This reduces the strong axial electric field gradient and is reflected in the slightly lower

quadrupole splittings in the susan complexes. This stronger covalency perpendicular to the Feꢀ

OꢀFe unit accompanied by lower quadrupole splittings is even more pronounced in the

complexes of the first generation. It is of interest to note the high quadrupole splitting of 1.81

ꢀ1

Me

III

mm s^ꢀ1for [(hilde^Me){Fe(µꢀO)Fe}] compared to 1.23 mm s in [(L ²)Fe ( ꢀO)Fe^III(L^Me²)]. This

ꢀ

2

reflects strong differences in the electronic structure of the trigonalꢀbipyramidal coordination

sphere compared to the squareꢀpyramidal as has already been apparent in the IR and magnetic

properties.

Electronic Absorption Spectroscopy

Only the spectra of the three representative complexes [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}],

[(susan){FeCl(µꢀO)FeCl}](ClO₄)₂, and [(hilde^Me){Fe(µꢀO)Fe}] will be compared and analyzed in

2

more detail (Fig. 6a). The comparison to the spectra of the ligands allows to assign the strong

feature at 39450 cm^ꢀ1in [(susan){FeCl(µꢀO)FeCl}]²⁺to

π→π* transitions of the pyridine donors of

17

Dalton Transactions

Page 18 of 42

View Article Online

DOI: 10.1039/C5DT03711E

ꢀ1

susan and the features in [(hilde^Me){Fe(µꢀO)Fe}] at 34900 cm and above 40000 cm to

π→

π*

2

transitions of the phenolic chromophores of (hilde^Me²)^4ꢀ. The band at 36600 cm^ꢀ1in

[(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}] decreases in intensity upon acidic titration.²⁴As protonation of a

coordinated carboxylate is most feasible, this band is assigned to have a strong carboxylate

LMCT character in accordance to reports on related complexes.^43,46The bands at 22800 cm^ꢀ1

and 31400 cm^ꢀ1in [(hilde^Me){Fe(µꢀO)Fe}] have mainly p_π→d_πand p_π→d_σphenolate LMCT

2

*

character, respectively, while also some weaker ꢀ

ꢀoxo LMCT transitions contribute.¹⁶The low

energy of these phenolate LMCT bands reflects the high covalency of the Fe^IIIꢀOPh bonds.

For [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}] and [(susan){FeCl(µꢀO)FeCl}]²⁺characteristic lowꢀenergy

features are resolved (Fig. 6b+c and Table 4).^43,47,48We follow the assignment of Solomon et al.

on [(HEDTA)₂{Fe^III( ꢀO)Fe^III}]^2ꢀ.⁴⁸The electronic absorption spectrum of this complex with a nonꢀ

ꢀ

coordinated protonated carboxylate arm closely resembles that of [(julia){Fe(OH₂)(µꢀ

O)Fe(OH₂)}]. The lowꢀenergy bands at 13130 cm^ꢀ1and 18960 cm^ꢀ1in [(julia){Fe(OH₂)(µꢀ

O)Fe(OH₂)}] and at 10540 cm^ꢀ1and 17340 cm^ꢀ1in [(susan){FeCl(µꢀO)FeCl}]²⁺correspond to

6

⁶A₁→⁴T₁and A₁→⁴T₂LF transitions, respectively. The energy of this transition decreases with

increasing LF strengths. The lower energies in [(susan){FeCl(µꢀO)FeCl}]²⁺reflect the stronger

LF by the pyridine donors (πꢀacceptors) compared to the carboxylates (π donors) in

[(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]. The sharp peaks at 21220 cm^ꢀ1and 20670 cm^ꢀ1of

[(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}] and [(susan){FeCl(µꢀO)FeCl}]²⁺, respectively, are spinꢀflip

transitions (⁶A₁→⁴A₁,⁴E) that are independent of LF strength but demonstrate variations in e^ꢀꢀe^ꢀ

repulsion. The lower energy in [(susan){FeCl(µꢀO)FeCl}]²⁺indicates a reduced e^ꢀꢀe^ꢀrepulsion

and thus a higher covalency for the pyridine than the carboxylate donors.

The HEDTA^3ꢀcomplex exhibits the ⁴T₁and ⁴T₂transitions at lower energies of 11700 and 18500

cm^ꢀ1, respectively, demonstrating a stronger LF than [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]. This is

consistent with two more σꢀdonating amines instead of the σꢀ and πꢀdonating carboxylates and

4

water. The A₁,⁴E transition at 21220 cm^ꢀ1in [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}] compared to 20800

cm^ꢀ1in the HDETA^3ꢀcomplex implies a stronger electron repulsion and thus lower covalency for

[(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}].

The energies of the oxo→Fe^IIICT transitions of [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}] closely resemble

those of the HEDTA^3ꢀcomplex.⁴⁸Interestingly, the oxo→Fe^IIICT transitions of [(susan){FeCl(µꢀ

O)FeCl}]²⁺are shifted by 2000ꢀ3000 cm^ꢀ1to lower energy (Table 4). This is consistent with the

πꢀacceptor character of pyridine that decreases the energies of the d_xzand d_yzacceptor orbitals

for these oxo→Fe^IIICT transitions. In summary, the electronic absorption spectra of [(julia)ꢀ

{Fe(OH₂)(µꢀO)Fe(OH₂)}] and [(susan){FeCl(µꢀO)FeCl}]²⁺reflect their differences in electronic

structure. The pyridine donors strengthen the LF and the covalency of FeꢀN^Pybonds by

πꢀback

18

Page 19 of 42

Dalton Transactions

View Article Online

DOI: 10.1039/C5DT03711E

donation and lower the energy of the d_xzand d_yzorbitals in contrast to the carboxylate donors,

that have less covalent FeꢀO^carbbonds, a weaker LF, and d_xzand d_yzorbitals of higher energies.

Electrochemistry

The pHꢀdependent electrochemistry of [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}] in water (pK_avalues 7.7

± 0.3 and 11.4 ± 0.4) was already published.²⁴The neutral form at pH = 5.9 exhibits two

irreversible reductions with peak potentials at ꢀ0.25 and ꢀ0.56 V vs NHE. The first reduction

process is shifted cathodically to ꢀ0.92 V vs NHE in the monoꢀdeprotonated species at pH = 9.8

consistent with the increased negative charge. Most interesting, the doublyꢀdeprotonated

species at pH = 12.8 exhibits one anodically shifted almost reversible electron transfer wave at

E^1/2= ꢀ0.75 V vs NHE. Based on overall chargeꢀpotential relations using the Born equation,⁴⁹

this redox couple was assigned to be Fe^IVFe^III/Fe^IIIFe^III.²⁴This assignment was corroborated by

the appearance of an anisotropic S_t= 1/2 EPR spectrum by preparation of the doublyꢀ

deprotonated species under aerobic conditions, which is not observed under dioxygenꢀfree

conditions.²⁴The observed gꢀtensor can only be explained by spinꢀprojection techniques with an

antiferromagnetically coupled Fe^IVh.s. / Fe^IIIh.s. spin system.²⁴This indicates an air oxidation to

a coupled Fe^IVFe^IIIspecies.

In order to compare the redox properties of [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}] to those of

[(susan){FeCl(µꢀO)FeCl}]²⁺

and

[(hilde^Me){Fe(µꢀO)Fe}],

the

electrochemistry

of

2

[(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}] was studied in a nonꢀaqueous solvent. Due to its insolubility in

CH₃CN, CH₂Cl₂, or THF, we performed the experiments on MeOH solutions (Fig. 7a). In

accordance with the measurements in H₂O,²⁴there are two kinetically hindered reduction

processes. Interestingly, the oxidative scan exhibits no electron transfer waves corresponding to

an oxidation of the complex but a strong oxidative catalytic current starting at ~ 1 V (Fig. 7a).

Although the oxidation to Fe^IVFe^IIIis masked by this catalytic current, it cannot be at higher

potentials. Thus, the oxidation must be at ≤1 V.

[(hilde^Me){Fe(µꢀO)Fe}] in CH₂Cl₂solution exhibits irreversible oxidations at 0.36 V and 0.79 V

associated with reductive waves at ꢀ0.75 V and ꢀ1.34 V, respectively (Fig. 7b). An irreversible

reduction is observed at ꢀ1.67 V associated with an oxidative wave at ꢀ1.11 V. Lowering the

temperature to ꢀ40 °C does not provide any reversibility of the redox waves. In analogy to the

dinuclear complexes of the first generation,¹⁶the oxidations are assigned to ligandꢀcentered

oxidations leading to coordinated phenoxylꢀradicals while the reduction is assigned to a metalꢀ

centered reduction to Fe^II.

2

Changing the solvent to THF does not only stabilize the oxidized species (E^p= 0.23 and 0.47

ox

V), but also provide some cathodic current in the reverse scan. This indicates a lifetime of a few

seconds of the oxidized species as manifested by experiments with faster sweep rates.

Interestingly, the irreversible reduction at ꢀ0.82 V associated with the oxidation at 0.23 V obtains

19

Dalton Transactions

Page 20 of 42

View Article Online

DOI: 10.1039/C5DT03711E

+

an anodic current in the reverse scan. This implies that the oxidized [(hilde^Me){Fe(µꢀO)Fe}] with

a coordinated phenoxyl radical rearranges to a new transient species with a redox potential

shifted cathodically by ~ 1 V.

2

Cyclic voltammograms of [(susan){FeCl(µꢀO)FeCl}]Cl₂(Fig. 8a), [(susan){FeCl(µꢀO)Feꢀ

Cl_0.75(OCH₃)_0.25}](ClO₄)₂(Fig. 8b), and [(susan){FeCl(µꢀO)FeCl}](ClO₄)₂(Fig. 8c) exhibit

irreversible oxidations around ~1.5 V and chemically reversible reductions at E_1/2= ꢀ0.73 V.

Additionally, [(susan){FeCl(µꢀO)FeCl}]²⁺and [(susan){FeCl(µꢀO)FeCl_0.75(OCH₃)_0.25}]²⁺show a

quasireversible second reduction at E_1/2= ꢀ0.95 V, which does not appear in [(susan){FeCl(µꢀ

O)FeCl}]Cl₂. The nonꢀcoordinated Cl^ꢀis oxidized at 0.63 V. The absence of this wave in the

ꢀ

cyclic voltammograms of the complexes with ClO₄counteranions indicates that there is no

significant exchange of coordinated Cl^ꢀby solvent molecules.

The waves around ~1.5 V and ꢀ0.73 V are intrinsic features of [(susan){FeCl(µꢀO)FeCl}]²⁺.

Considering the pyridine donors of susan, we assign the oxidative wave around ~1.5 V to be

metalꢀcentered leading to Fe^IV. The subsequent structural rearrangements lead to associated

reductive waves at 0.52 and 0.25 V. On the other hand, the reduction at E_1/2= ꢀ0.73 V is also

assigned to a metalꢀcentered process.

Interestingly, the cyclic voltammogram of [(susan){FeCl(µꢀO)FeCl_0.75(OCH₃)_0.25}]²⁺exhibits an

additional oxidative peak at 1.14 V of lower current, which has to be assigned to the oxidation of

the minority species [(susan){FeCl(µꢀO)Fe(OCH₃}]²⁺. Thus, the substitution of a coordinated Cl^ꢀ

by the stronger electron donating OMe^ꢀresults in a stabilization of the Fe^IVspecies by 390 mV!

Catalytic C-H activation of cyclohexane

In order to test the reactivity of the diferric complexes in oxygenation catalysis, the catalytic

oxygenation of cyclohexane using several oxygenation agents was investigated (Table 4). The

protocol that has been adapted from published procedures (typical conditions:

CyH:oxidant:catalyst 900:150:1 with a catalyst concentration of 0.6ꢀ0.9 mM).⁵⁰Due to its

insolubility in CH₃CN, the reactions with [(hilde^Me){Fe(µꢀO)Fe}] were performed in CH₂Cl₂, while

2

[(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}] could not be tested due to its insolubility in either CH₃CN or

CH₂Cl₂. For reference, reactions were also conducted without catalyst (entries 31ꢀ37), as e.g.

for mꢀCPBA significant turnovers were reported in the literature.⁵¹Besides cyclohexanol

(CyOH), significant amounts of cyclohexanone (CyO) were observed, which can be formed by

disproportionation of CyOO• radicals⁵²or additional oxidation of CyOH. Therefore, we tested the

oxidation of CyOH to CyO for further reference. Without catalyst, no oxidation is observed under

O₂with and without tꢀBuOOH. In the presence of NEt₃and tꢀBuOOH, 2ꢀ3 % oxidation is

observed. With Fe(ClO₄)₃•10H₂O as catalyst a TON of 65 was observed for the oxidation of

CyOH to CyO. In the presence of NEt₃, the TON decreased to 18, which we attribute to the

visible formation of rust. On the other hand, when [(susan){FeCl(µꢀO)FeCl}](ClO₄)₂is used as

20

Page 21 of 42

Dalton Transactions

View Article Online

DOI: 10.1039/C5DT03711E

catalyst, the TON increased from 58 to 87 by using 10 eq. NEt₃. Thus, the latter is a good

catalyst for the oxidation of CyOH to CyO.

oxidation of CyO.

εꢀCaprolactone arises from the BaeyerꢀVilliger

[(susan){FeCl(µꢀO)FeCl}](ClO₄)₂catalyzes the CꢀH activation of CyH with all tested oxidants, but

by different amounts. Under air, the best TONs were obtained using AcOOH (22.0, entry 3)

followed by mꢀCPBA (19.5, entry 5) and tꢀBuOOH (16.8, entry 7), while usage of H₂O₂results in

only moderate TONs of 4.5 (entry 1). Interestingly, when the reactions were carried out under

Ar, the ratios CyOH/(CyO+lact) were almost unaffected, whereas the TONs obtained by usage

of AcOOH (entry 4) and tꢀBuOOH (entry 8) were significantly inꢀ and decreased, respectively.

For further comparisons, we have used Fe(ClO₄)₃•10H₂O as catalyst (entries 24ꢀ30).⁵³

Remarkably, this results in TONs up to ≈30 for tꢀBuOOH (entries 27+28). Thus,

Fe(ClO₄)₃•10H₂O is a better catalyst than [(susan){FeCl(µꢀO)FeCl}](ClO₄)₂for the oxidants tꢀ

BuOOH and H₂O₂.

In order to convert tꢀBuOOH into its better coordinating anion, we performed the catalytic

reaction also in the presence of 10 equivalents NEt₃. Under these conditions, the TON for

[(susan){FeCl(µꢀO)FeCl}](ClO₄)₂increases to 39.5 (entry 9). On the other hand, the TON of

Fe(ClO₄)₃•10H₂O decreases to 0.3 (entry 29), which we again attribute to the visible formation of

rust.

Complex [(hilde^Me){Fe(µꢀO)Fe}] showed only catalytic behaviour using mꢀCPBA. Remarkably,

2

the equivalents of CyOH formed increase from 4.0 under air (entry 18) to 24.3 under Ar with a

selectivity ratio CyOH/(CyO+lact) of 25.9 (entry 21). Furthermore, significant amounts of CyOH

with an also large CyOH/(CyO+lact) ratio are already formed after a few seconds. Thus,

[(hilde^Me){Fe(µꢀO)Fe}] is a fast and efficient catalyst for the hydroxylation of CyH with mꢀCPBA

2

as oxidant but the yields are lower than for [(susan){FeCl(µꢀO)FeCl}](ClO₄)₂. This indicates a

lower stability of [(hilde^Me){Fe(µꢀO)Fe}] under the catalytic conditions. This is in accordance to

2

the results of the phenolic complexes of the first generation that oxidation to a phenoxyl radical

complex results in a cleavage of the ꢀꢀoxo bridge. Such a behaviour has not yet been observed

[(susan){FeCl(µꢀO)FeCl}](ClO₄)₂.

Conclusions

Based on the results of our complexes of the first generation using the ligands H₂L^R2that (i) the

lowꢀpotential reversible oxidations are ligandꢀcentered and (ii) the oxidized dinuclear complexes

decay into mononuclear fragments, we performed a rational design for the ligand of the second

generation H₄julia to prevent this decay and to ensure metalꢀbased oxidations. Variation of the

terminal ligands results in a new ligand system including susan and H₄hilde^Me², which were used

for the synthesis of their ꢀꢀoxo diferric complexes.

21

Dalton Transactions

Page 22 of 42

View Article Online

DOI: 10.1039/C5DT03711E

[(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}] could be crystallized in form of the two hydrates

[(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]•6H₂O and [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]•7H₂O. In both

hydrates, [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}] is embedded in a water matrix forming hydrogen bonds

to the first and second coordination sphere, which differ for the two hydrates. This allows to

study the influence of hydrogen bonds on the first and second coordination sphere, which are

considered as a handle to tune the activeꢀsite properties in metalloenzymes and their

models.^{36,37,39–41}The FeꢀO^carbbond length of a carboxylate not involved in hydrogen bonds is

1.9994(7) Å. One hydrogen bond to the coordinated carboxylate Oꢀdonor increase the FeꢀO^carb

bond lengths to 2.0366(7) Å and 2.0307(7) Å. Interestingly, two hydrogen bonds to the

uncoordinated carboxylate oxygen atom lead to almost the same increase to 2.0273(7) Å. This

remote effect is transmitted by a lengthening of the C=O bond that strengthens the CꢀO bond,

which results in the observed lengthening of the FeꢀO^carbbond. This perturbation is reflected by

a shift of ν ν

_as(CO₂) by 43 cm^ꢀ1to lower energy and of _s(CO₂) by 28 cm^ꢀ1to higher energy.

ꢀ

The dioxygen activation cycle in metalloenzymes and also in models requires a certain flexibility

of the active site. The analysis of the molecular structures of the complexes of the second

generation demonstrates a strong flexibility of the ligand backbone. Especially the ethylene

spacer bridging the two ligand compartments proves to be highly flexible allowing for a variation

in the metalꢀmetal separation and in the opening angle on the opposite side for potential

substrate binding.

The intended ability of the complexes to selectively activate and hydroxylate CꢀH bonds requires

the approach of a CꢀH bond of the substrate to the oxidized metalꢀoxygen core. Besides an

accessibility to this central core, an affinity for a catalystꢀsubstrate complex is needed in analogy

to an enzymeꢀsubstrate complex. This is demonstrated by a CH₂Cl₂molecule that is weakly

bound by vanꢀderꢀWaals interactions in a hydrophobic binding pocket of [(hilde^Me){Fe(µꢀO)Fe}].

2

Like in a ‘snapshot’ of the anticipated mechanism, the CꢀH bonds are oriented to the central

core.

The electronic structures of the complexes are investigated by means of FTIR, UVꢀVisꢀNIR, and

Mössbauer spectroscopies as well as magnetic and electrochemical measurements to

determine the influence of the terminal ligands. Interestingly, the ν_as(FeꢀOꢀFe) stretching mode

and the exchange coupling constant J are influenced by the specific electronic features of the

terminal ligands and do not correlate ꢀ as was observed earlier⁴³ꢀ with the FeꢀO distance and

the FeꢀOꢀFe angle of the Fe^III(

The quadrupole splitting | E_Q| is also a sensitive probe for differences in the bonding situation.

Generally, ꢀoxo diferric complexes are known for large | E_Q| due to a strong electric field

gradient originating from the anisotropic covalency imposed by the highly covalent

bond.⁴⁵The complexes of (julia)^4ꢀexhibit the largest |

E_Q| due to the one highly covalent

ꢀ

ꢀO)Fe^IIIcore.

ꢁ

ꢀ

ꢁ

ꢀ

ꢀOꢀFe

ꢁ

ꢀ

22

Page 23 of 42

Dalton Transactions

View Article Online

DOI: 10.1039/C5DT03711E

bond. This is reduced in the susan complexes due to the stronger covalency of the pyridine

relative to the carboxylate, which results in more electron density in this direction thus

attenuating the electric field gradient. Interestingly, [(hilde^Me){Fe(µꢀO)Fe}] has a much larger

2

|ꢁE_Q| than the first generation complexes. This reflects the strong influence of the coordination

environment (trigonalꢀbipyramidal vs squareꢀpyramidal).

The covalency differences of the terminal donors are also reflected in the UVꢀVisꢀNIR spectra.

ꢀ1

[(hilde^Me){Fe(µꢀO)Fe}] exhibits a lowꢀenergy LMCT of the phenolates at 22800 cm , which

2

prevents an analysis of the dꢀd transitions, while the carboxylates in [(julia){Fe(OH₂)(µꢀ

O)Fe(OH₂)}] are assigned to contribute at 36600 cm^ꢀ1. In [(susan){FeCl(µꢀO)FeCl}]²⁺, the

→ Fe^IIILMCT are shifted ~2000ꢀ3000 cm^ꢀ1to lower energy in comparison to [(julia){Fe(OH₂)(µꢀ

ꢀꢀoxo

O)Fe(OH₂)}] due to

πꢀbackbonding of the pyridine ligands that stabilizes the d_xzand d_yzacceptor

6

orbitals. This stabilization, i.e. higher LF, is also evident by the lowꢀenergy shift of the A₁→⁴T₁

6

and A₁→⁴T₂transitions. As a direct probe of covalency, the spinꢀflip transition A₁→⁴A₁,⁴E,

which only depends on electronic repulsion, decreases from 21220 cm^ꢀ1in [(julia){Fe(OH₂)(µꢀ

O)Fe(OH₂)}] to 20670 cm^ꢀ1in [(susan){FeCl(µꢀO)FeCl}]²⁺demonstrating the stronger

nephelauxeticity i.e. covalency of a pyridine than a carboxylate.

The covalency differences also influence the electrochemical properties. [(susan){FeCl(µꢀ

O)FeCl}]²⁺can only be oxidized irreversibly at 1.48 V vs Fc⁺/Fc to an Fe^IVFe^IIIspecies.

Interestingly, the unintended but reproducible synthesis of [(susan){FeCl(µꢀO)FeCl

_0.75ꢀ

(OCH₃)_0.25}]²⁺shows that the substitution of a coordinated Cl^ꢀby the stronger electron donating

OMe^ꢀshifts the oxidation to 1.14 V vs Fc⁺/Fc, i.e. the Fe^IVspecies is stabilized by 390 mV.

Going from the

means going from the dicationic [(susan){FeCl(µꢀO)FeCl}]²⁺to the neutral [(julia){Fe(OH₂)(µꢀ

O)Fe(OH₂)}], decreases the potential for oxidation to an Fe^IVFe^IIIspecies to 1 V vs Fc⁺/Fc (the

potential is masked by the onset of an electrocatalytic current). This corroborates our initial

πꢀaccepting pyridine donors to the πꢀdonating carboxylate donors, which also

≤

hypothesis in the ligand design to substitute the

πꢀaccepting pyridine donors to πꢀdonating

carboxylate donors.

In analogy to the complexes of the first generation, [(hilde^Me){Fe(µꢀO)Fe}] in CH₂Cl₂solution

2

can be irreversibly oxidized at 0.36 V and 0.79 V vs Fc⁺/Fc leading to coordinated phenoxylꢀ

radicals. However, in THF solution, in which the complexes of the first generation decay into

+

mononuclear fragments, the oxidized [(hilde^Me){Fe(µꢀO)Fe}] rearranges to a new species with

2

a redox potential shifted cathodically by ~ 1 V.

Without optimization of the catalytic conditions, turnover numbers for the oxidation of

cyclohexane up to 39.5 and 27.0 for [(susan){FeCl(µꢀO)FeCl}]²⁺and [(hilde^Me){Fe(µꢀO)Fe}],

2

respectively, were obtained. [(susan){FeCl(µꢀO)FeCl}]²⁺is also a good catalyst for the oxication

23

Dalton Transactions

Page 24 of 42

View Article Online

DOI: 10.1039/C5DT03711E

of CyOH to CyO. Thus, the selectivity between CyOH and CyO is low. Interestingly, a strong

selectivity for CyOH has been obtained using [(hilde^Me){Fe(µꢀO)Fe}] under argon.

2

Acknowledgments

This work was supported by the Fonds der Chemischen Industrie, the DFG (Gl295/3ꢀ1 and

Gl295/3ꢀ2), and Bielefeld University. J.B.H.S. and S.W. also gratefully acknowledge the Fonds

der Chemischen Industrie for doctoral fellowships. We are grateful to Dr. E. Bothe (MaxꢀPlanckꢀ

Institut für Bioanorganische Chemie) for assistance with the fast electrochemical

measurements.

Graphical Abstract

Based on our results of the ligands and complexes of the first generation to stabilize highꢀvalent

diiron species, we have developed in a rational design approach the ligands of the second

generation H₄julia, susan, and H₄hilde^Me². The strong influence of the terminal donors on the

properties of their ꢀꢀoxo diferric complexes is presented.

24

Page 25 of 42

Dalton Transactions

Table 1. Crystal data and refinement parameters.

[(hilde^Me){Fe(ꢁꢀ

[(susan){FeCl(ꢁꢀ

[(susan){FeCl(ꢁꢀO)FeCl_0.75

ꢀ

[(susan){FeCl(ꢁꢀ

[(hilde^Me){FeCl}₂]ꢀ

2

identification code

[(julia){Fe(OH₂)(ꢁꢀ

O)Fe(OH₂)}]•7H₂O

O)Fe}]•CH₂Cl₂

O)FeCl}]Cl₂•2H₂O

(OCH₃)_0.25}](ClO₄)₂•0.5MeOH O)FeCl}](ClO₄)₂•0.5EtOH •2CH₂Cl₂

empirical formula

formula weight

crystal system

space group

a [Å]

C₁₆H₄₄Fe₂N₄O₁₈

692.25

C₄₅H₆₀Cl₂Fe₂N₄O₅

919.57

C₃₂H₄₆Cl₄Fe₂N₈O₃

844.27

C_32.75H_44.75Cl_3.75Fe₂N₈O_9.75

C₃₃H₄₅Cl₄Fe₂N₈O_9.5

959.27

C₄₆H₆₂Cl₆Fe₂N₄O₄

1059.40

951.15

triclinic

ꢃ

P1

monoclinic

C2

monoclinic

C2/c

monoclinic

C2/c

triclinic

ꢃ

P1

monoclinic

P2₁/c

19.6588(12)

7.7101(5)

10.0166(6)

90.00

17.5005(9)

11.7785(6)

21.6508(11)

90.00

15.9941(10)

15.4161(10)

15.6472(10)

90.00

8.1738(8)

13.2744(13)

19.1664(19)

83.257(5)

81.601(5)

72.579(4)

1956.9(3)

2

8.1714(4)

13.1963(6)

19.2643(10)

83.133(2)

81.364(2)

72.765(2)

1955.33(17)

2

8.5710(5)

9.9468(6)

28.6201(16)

90.00

b [Å]

c [Å]

α

β

γ

[°]

106.5190(10)

90.00

97.6970(10)

90.00

109.858(3)

90.00

95.3430(10)

90.00

[°]

V [Å³]

1455.56(16)

2

4422.7(4)

4

3628.7(4)

4

2429.4(2)

2

Z

ρ

ꢀ

[g cm^ꢀ3]

[mm^ꢀ1]

1.579

1.381

1.545

1.614

1.629

1.448

1.079

0.826

1.140

1.063

1.081

0.973

F(000)

728.0

1936.0

1752.0

982.0

990.0

1104.0

crystal size [mm³]

0.18 × 0.10 × 0.06

173(2)

0.11 × 0.10 × 0.02

183(2)

0.23 × 0.11 × 0.10

100(2)

0.35 × 0.25 × 0.13

100(2)

0.26 × 0.10 × 0.05

100(2)

0.40 × 0.10 × 0.06

183(2)

T [K]

4.24 to 52.00

3.80 to 53.98

5.20 to 54.00

5.58 to 54.00

5.40 to 54.00

2.86 to 54.00

2Θ range [°]

hkl ranges

ꢀ23 ≤ h ≤ 24, ꢀ9 ≤ k ≤ 9, ꢀ22 ≤ h ≤ 20, ꢀ15 ≤ k ≤ 11, ꢀ20 ≤ h ≤ 20, ꢀ19 ≤ k ≤ 19, ꢀ10 ≤ h ≤ 10, ꢀ16 ≤ k ≤ 16,

ꢀ10 ≤ h ≤ 10, ꢀ16

≤ k ≤ 16, 0 ≤ l ≤ 24

75009

ꢀ10 ≤ h ≤ 9, ꢀ12 ≤ k ≤ 12,

ꢀ36 ≤ l ≤ 35

13502

ꢀ10 ≤ l ≤ 12

ꢀ26 ≤ l ≤ 27

ꢀ19 ≤ l ≤ 19

ꢀ24 ≤ l ≤ 24

collected refl.

4085

12835

30537

92127

unique refl., R_int

observed refl. (I>2

completeness

2624 ,0.0214

2597

4802, 0.0458

3405

3960, 0.0233

3567

8497, 0.0267

7793

8055, 0.0302

7683

5154, 0.0486

3973

σ(I))

0.994

0.993

0.998

0.994

0.965

0.971

absorption correction

data/restraints/parameters

goodnessꢀofꢀfit on F²

multiꢀscan

2624/1/183

1.246

multiꢀscan

4802/0/268

1.039

multiꢀscan

3960/0/233

1.011

multiꢀscan

8497/0/530

1.056

multiꢀscan

8055/0/508

1.140

multiꢀscan

5154/0/285

1.028

0.0450, 0.1100

0.0456, 0.1102

0.57/ꢀ0.70

1024404

0.0524, 0.1101

0.0863, 0.1248

0.52/ꢀ0.65

1024402

0.0255, 0.0661

0.0301, 0.0689

0.80/ꢀ0.33

1024399

0.0243, 0.0627

0.0278, 0.0652

0.38/ꢀ0.39

1024400

0.0310, 0.0952

0.0341, 0.0983

0.67/ꢀ0.41

1024401

0.0498, 0.1337

0.0676, 0.1455

1.09/ꢀ0.67

1024403

R1, wR2 (I>2σ(I))

R1, wR2 (all data)

largest peak/hole [e Å^ꢀ3]

CCDC numbers

25

Dalton Transactions

Page 26 of 42

Table 2. Selected structural parameters from singleꢀcrystal Xꢀray diffraction (distances in Å, angles in °).^a

[(hilde^Me){Fe(ꢁꢀ

[(hilde^Me){Fe

[(susan){FeCl(ꢁꢀ

[(susan){FeCl(ꢁꢀO)Fe

2

[(julia){Fe(OH₂)(ꢁꢀ

[(julia){Fe(OH₂)(ꢁꢀ [(julia){Fe(DMSO)(

O)Fe(OH₂)}]•6H₂O

O)Fe(OH₂)}]•7H₂O

µꢀO)Fe(DMSO)}]

O)Fe}]•CH₂Cl₂

Cl}₂]•2CH₂Cl₂

O)FeCl}]Cl₂•2H₂O O)FeCl_0.75(OCH₃)_0.25

}

Cl}](ClO₄)₂•0.5EtOH

•3DMSO

](ClO₄)₂•0.5MeOH

Fe1ꢀO3

Fe2ꢀO3

Fe1ꢀN1

Fe2ꢀN41

Fe1ꢀN2

Fe2ꢀN42

Fe1ꢀO1

Fe2ꢀO41

Fe1ꢀO2

Fe2ꢀO42

Fe1ꢀN3

1.7692(7)

1.7686(7)

2.2117(8)

2.1989(8)

2.2324(8)

2.2357(8)

2.0366(7)

2.0307(7)

1.9994(7)

2.0273(7)

1.7769(8)

2.190(4)

2.223(4)

2.033(3)

2.032(3)

1.783(3)

1.788(3)

2.228(4)

2.223(4)

2.257(4)

2.247(4)

2.027(3)

2.015(3)

2.028(3)

2.020(3)

1.7858(7)

2.242(3)

2.294(3)

1.877(2)

1.892(2)

1.7932(3)

2.2330(13)

2.2019(13)

1.7786(10)

1.7974(10)

2.2614(13)

2.2501(14)

2.2052(13)

2.2204(13)

1.780(2)

1.798(2)

2.255(3)

2.249(3)

2.201(3)

2.217(3)

2.223(2)

2.220(3)

1.862(2)

1.833(2)

2.1741(14)

2.2278(13)

2.1653(12)

2.1608(13)

2.2559(13)

2.2011(13)

2.157(3)

2.155(3)

2.259(3)

2.202(3)

Fe2ꢀN43

Fe1ꢀN4

Fe2ꢀN44

Fe1ꢀO81

Fe2ꢀO82

Fe1ꢀCl1

Fe2ꢀCl2

Fe2ꢀO98

Fe1•••Fe2/Fe1ʹ

Fe1ꢀO3ꢀFe2/Fe1ʹ

FeꢀO^oxo

2.1421(7)

2.0991(7)

2.085(3)

2.100(3)

2.094(3)

2.3202(10)

6.3799(8)

2.3099(4)

2.2982(5)

2.3311(11)

1.745(6)

3.5693(4)

172.98(7)

1.788

2.2961(10)

2.3166(11)

3.5156(2)

167.16(4)

1.769

3.5216(11)

164.6(3)

1.777

3.5409(11)

165.1(2)

1.786

3.5215(9)

160.78(19)

1.786

3.5765(5)

171.47(9)

1.793

3.5721(7)

173.51(15)

1.789

FeꢀN^c

2.21

2.23

2.02

2.19

2.22

2.03

2.23

2.25

2.02

2.24

2.29

1.88

2.22

1.84

2.23

2.20

2.26

2.21

2.25

2.21

FeꢀN^t

FeꢀO

FeꢀN^cis

2.17

2.23

2.31

3.58

44.52

21.49

20.82

2.16

2.23

2.31

3.57

39.65

15.58

15.15

2.16

2.23

2.31

3.57

39.47

FeꢀN^trans

FeꢀX

2.12

3.52

43.32

57.77

55.72

2.09

3.52

40.67

54.98

52.36

2.10

3.54

40.38

52.95

50.00

2.32

6.38

180.00

Fe•••Fe

3.52

59.95

53.90

θ

(N^cꢀFeꢀFeꢀN^c)

α ^b

16.03

β ^c

32.25

15.51

a

Symmetry transformations used to generate equivalent atoms: ʹ ꢀx+2,y,ꢀz+1 for [(julia){Fe(OH₂)(ꢁꢀO)Fe(OH₂)}]•7H₂O, ʹ ꢀx+1,y,ꢀz+1/2 for [(hilde^Me){Fe(ꢁꢀ

2

O)Fe}]•CH₂Cl₂, ʹ ꢀx,y,ꢀz+3/2 for [(susan){FeCl(ꢁꢀO)FeCl}]Cl₂•2H₂O

^bThe angle

^cThe angle

α

β

is the angle between the best planes through the four donor atoms coordinating each Fe ion in the plane perpendicular to the FeꢀN^cvector.

is the angle between the vectors through the two donor atoms perpendicular for each Fe ion to the FeꢀN^cand FeꢀO^oxovectors.

26

Page 27 of 42

Dalton Transactions

Table 3. Structural, magnetic, and spectroscopic properties of {Fe^III(

ꢀ

ꢀO)Fe^III} complexes with ligands of the second generation and

for comparison of the first generation.

ꢀ

̅

J

ꢂ(FeꢀO)

∠(FeOFe)

ν

_as(FeOFe)

/ cm^ꢀ1

ν

_as(CO₂)/ꢀ

δ

|

ꢁ

E|_Q

Γ

Ref.

/ cm^ꢀ1

/ Å

/ °

/ mm s^ꢀ1/ mm s^ꢀ1

/ mm s^ꢀ1

ꢀ

ν

_s(CO₂)/ cm^ꢀ1

1634(1591)/

1371(1399) ^a

1626/1387

[(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]•6H₂O

1.769

167.2

889

ꢀ108

0.47

2.04

0.30

this work

[(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]•7H₂O

[(julia){Fe(DMSO)(µꢀO)Fe(DMSO)}]

[(susan){FeCl(µꢀO)FeCl}]Cl₂

1.777

1.786

1.793

1.788

1.789

1.786

164.6

165.1

171.5

172.0

173.5

160.8

882

864

814

817

816

795

this work

1637/1362

ꢀ102

ꢀ106

ꢀ105

ꢀ108

ꢀ93.5

0.47

0.46

0.44

0.47

0.44

0.42

1.91

1.58

1.65

1.63

1.81

1.14

1.22

1.23

1.01

0.27

0.31

0.32

0.31

0.33

0.38

0.27

0.30

0.31

[(susan){FeCl(µꢀO)FeCl_0.75(OCH₃)_0.25}](ClO₄)₂

[(susan){FeCl(µꢀO)FeCl}](ClO₄)₂

[(hilde^Me){Fe(µꢀO)Fe}]

2

[(hilde^Me){FeCl}₂]

2

this work

III

16

[(L^tꢀBu)Fe (

ꢀ

ꢀO)Fe^III(L^tꢀBu)]

1.782

1.783

1.772

180

810

817

829

ꢀ89

ꢀ98

ꢀ99

2

16

III

[(L^Me)Fe (

ꢀ

ꢀO)Fe^III(L^Me)]

2

III

2

b

[(L^Cl)Fe (

ꢀ

ꢀO)Fe^III(L^Cl)]

166.1

2

a

Values in parenthesis correspond to the bands of lower intensity for the carboxylates with two hydrogen bonds.

Data were taken for the CH₂Cl₂solvate.¹⁶

b

27

Dalton Transactions

Page 28 of 42

View Article Online

DOI: 10.1039/C5DT03711E

Table 4. Product yields (in equivalents based on catalyst) after 18 h of oxidation of cyclohexane.

catalyzed at 25 °C by [(susan){FeCl(ꢁꢀO)FeCl}](ClO₄)₂(X) or Fe(ClO₄)₃ꢂ10H₂O (Z) in CH₃CN or

[(hilde^Me){Fe(ꢁꢀO)Fe}] (Y) in CH₂Cl₂. Reactions were carried out (unless stated otherwise) under air in

a stoichiometry of 900 cyclohexane, 150 oxidant, and 1 catalyst.

CyOH

entry

cat

oxidant

others

CyOH

CyO

lact^a

CyCl

sum

CyO+lact

1.9

1

2

X

Y

H₂O₂

2.9

3.0

1.5

7.2

10.4

7.6

6.4

7.9

5.2

19.8

4.7

1.2

0.0

0.4

0.0

tr.^b

0.0

0.4

0.2

0.0

0.3

0.0

0.2

0.4

0.1

1.2

0.0

0.2

0.0

4.5

Ar

2.0

1.8

1.4

1.8

1.1

1.0

0.5

1.2

−

3

AcOOH

mꢀCPBA

tꢀBuOOH

H₂O₂

14.6

18.4

11.4

12.0

8.9

22.0

29.3

19.5

19.8

16.8

10.8

39.5

7.0

4

5^c

6^c

7

8

Ar

5.5

9

10 NEt₃

19.7

2.3

10

11

12

13

14

15^d

16

150 NEt₃

10 KOꢀtꢀBu

10 NEt₃

1.5

2.9

0.0

H₂O₂

135 NEt₃

0.0

−

AcOOH

tꢀBuOOH (aq)

tꢀBuOOH

(nonane)

0.5

1.0

1.2

0.0

−

0.0

0.3

10 NEt₃

0.0

17

Y

0.3

0.0

0.6

1.0

18

19^e

20^d,e

21

22^d

23^f

Y

mꢀCPBA

H₂O₂

4.0

4.4

0.0

24.3

0.1

4.4

14.3

0.1

2.6

14.1

14.0

0.1

0.2

0.0

0.1

0.8

0.2

0.1

1.4

0.4

0.7

0.6

0.0

0.3

0.1

0.2

8.3

0.1

1.7

15.6

12.0

0.2

0.4

0.0

0.1

0.9

0.3

0.1

0.0

0.2

1.1

1.8

0.0

0.6

0.0

1.4

0.0

tr.^b

0.0

4.1

2.6

0.0

1.8

0.0

0.4

0.0

9.9

9.4

0.0

27.0

0.1

5.0

22.6

0.2

5.7

29.7

26.0

0.3

0.7

0.0

0.2

1.7

0.4

0.2

1.4

0.7

2.2

1.9

−

Y

Ar

25.9

1.0

20.7

1.7

0.9

1.2

0.5

0.6

−

Y

24

Z

25

Z

AcOOH

mꢀCPBA

tꢀBuOOH

H₂O₂

26

Z

27

Z

28

Z

Ar

29

Z

10 NEt₃

20 NEt₃

30

Z

31^g,h

32^g,h

33^g,h

34^g,h

35^g,h

36^g,i

37^g,i

no

AcOOH

mꢀCPBA

tꢀBuOOH

mꢀCPBA

0.7

0.8

0.7

1.0

−

Ar

2.1

^aεꢀcaprolactone. ^btr: traces. ^c19 h reaction time. ^d10 equivalents of oxidant. ^eDropwise addition of

oxidant. ^fReaction quenched directly after addition of mꢀCPBA. ^gProduct yields calculated on

assumption of presence of 1 eq of a catalyst. ^hIn CH₃CN. ⁱIn CH₂Cl₂.

28

Page 29 of 42

Dalton Transactions

View Article Online

DOI: 10.1039/C5DT03711E

Figure 1. Molecular structures of [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}] in crystals of a)

24

[(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]•6H₂O

and b) [(julia){Fe(OH₂)(µꢀO)Fe(OH₂)}]•7H₂O, c)

Me₂

[(julia){Fe(DMSO)(µꢀO)Fe(DMSO)}],²⁴d) [(hilde^Me){Fe(µꢀO)Fe}], and e) [(hilde ){FeCl}₂].

2

Hydrogen atoms are omitted for clarity.

29

Article Doi

Design and synthesis of a dinucleating ligand system with varying terminal donor functions that provides no bridging donor and its application to the synthesis of a series of FeIII-μ-O-FeIII complexes

DOI: 10.1039/c5dt03711e

Source and publish data:

Authors:

Article abstract of DOI:10.1039/c5dt03711e

Full text of DOI:10.1039/c5dt03711e

Products guided by the article

R&D Labs maybe for 108-93-0

Relevant to this article

Hot Product