Tetrapodal amidoxime ligands I. Coordination isomerism due to self-complementary dimerization of a pyramidal cobalt(III) coordination module [original]

Dalton

Transactions

Dynamic Article Links

Cite this:

Dalton Trans.

, 2012, 41, 2477

www.rsc.org/dalton PAPER

Tetrapodal amidoxime ligands I. Coordination isomerism due to

self-complementary dimerization of a pyramidal cobalt(III) coordination

module†‡

John P. Boyd,* Elisabeth Irran and Andreas Grohmann*

Received 29th September 2011, Accepted 21st November 2011

DOI: 10.1039/c2dt11847e

A bis-m-amidoximato-bridged cobalt(III) dimer obtained with a new tetrapodal ligand possesses

interesting structural parameters as a consequence of intramolecular hydrogen bonding intentionally

built into the complex. Its synthesis and properties are described. The new ligand type combines

attributes of two previously described ligand classes: It binds a metal ion in a tetrapodal pentadentate

fashion and forms a pseudomacrocycle through hydrogen bonds, characteristic of chelating oxime

ligands. Coordination isomerism, which is a consequence of dimer formation, has been analyzed by

means of X-ray crystallography and carbon-13 nuclear magnetic resonance spectroscopy.

Introduction

The application of effects arising from peripheral hydrogen

bonding in transition metal complexes is of growing importance.

Such interactions are instrumental in bioinorganic chemistry and

catalysis, where secondary bonding can create a “coordination

pocket”, tune metal as well as substrate reactivity, stabilize reactive

intermediates, and facilitate intra- and intermolecular proton

transfer reactions.1,2 Two limiting cases of intramolecular hydro-

gen bonding may be distinguished: H-bonding between a metal-

coordinated species and the ligand periphery,3,4 and H-bonding

within the ligand periphery.5Additionally, hydrogen bonds have

been used to incorporate transition metal complexes into host

structures, such as cyclodextrins,6and to tune the strength of

metal ion extractants.7Methodologies to create ligands that

add hydrogen bond-mediated properties to known coordination

geometries thus constitute a useful addition to the coordination

chemist’s toolbox.

Our approach aims at linking equatorial oxime donors via an ax-

ial N-heterocycle, thereby generating a tetrapodal pentadentate N5

ligand which forms a square pyramidal coordination environment.

The capping ligand thus obtained is reinforced by two hydrogen

bonds at the basis of the pyramid. This combination is attractive

for at least two reasons: Firstly, pentadentate tetrapodands enforce

a donor pattern that may, for reasons of kinetic lability, otherwise

be unavailable for ﬁrst row transition metal complexes;8secondly,

dioxime ligands can form strong hydrogen bonds resulting in pseu-

domacrocyclic arrangements. This involves deprotonation, which

Institut f¨

ur Chemie, Technische Universit¨

at Berlin, Straße des 17. Juni 135,

10623, Berlin, Germany

† Electronic supplementary information (ESI) available. CCDC reference

numbers 825623–825624. For ESI and crystallographic data in CIF or

other electronic format see DOI: 10.1039/c2dt11847e

‡ Dedicated to Martin A. Bennett

in effect compensates positive charge brought into the compound

upon metal ion coordination (formation of an “inner salt”).

The connectivity of the oxime functions in a system of our design

necessarily differs from that of the well-known glyoxime (i.e. 1,2-

dioxime) ligands and related pseudomacrocyclic ligands, which

have also been used as bridged pairs, especially for coenzyme B12

modelling (“cobaloximes”, see Fig. 1, 1a,1c).9While 1,3-dioxime

(including malonamidoxime, i.e., 1,3-diamidoxime; 1e) ligands

form hydrogen bonds that are energetically10 and geometrically11

similar to those found in 1,2-dioxime complexes, they lack conju-

gation and therefore show decreased p-backbonding character.

Amidoximes (e.g. malonamidoxime12) are easily prepared from

nitriles by reaction with hydroxylamine (Tiemann reaction).13 By

varying a nucleophilic aromatic substitution methodology that

was established in the context of intramolecular [4 + 2] Diels–

Alder/retro-Diels–Alder ring transformations,14 we gained access

to a pyrimidine-derived nitrile precursor and from it obtained the

prototype of a new ligand class (Fig. 1, 4). We describe the synthesis

of the ligand and the properties of its cobalt(III) complex, and show

that the new type of complex possesses features characteristic of

both the hydrogen bond-forming dioxime ligands (Fig. 1, 1a,1c;

1e) and ligands of the tetrapodal pentadentate topology (2,3).15,16

Experimental

Research chemicals were obtained from commercial suppliers and

used without further puriﬁcation. THF was saturated with dry

dinitrogen, dried over potassium hydroxide for four weeks and

used without further puriﬁcation. Potassium hydroxide (5 wt%)

was dissolved in ethanol and the solution stored for several

weeks to decompose 2-butanone present as a denaturing agent;

the solvent was then distilled in vacuo. Routine NMR spec-

tra were measured on a Bruker ARX 200 spectrometer, while

the 13C{1H}-spectrum of 11 was measured on a Bruker AV

This journal is ©The Royal Society of Chemistry 2012 Dalton Trans., 2012, 41, 2477–2485 | 2477

Published on 05 January 2012. Downloaded by TU Berlin - Universitaetsbibl on 31/03/2016 08:28:14.

View Article Online

/ Journal Homepage

/ Table of Contents for this issue

Fig. 1 “Capped” (1a,17 1b18) and “bridged”(1c,91d19) pseudomacrocyclic and macrocyclic ligands; conceptual combination of the malonamidoxime pair

(1e11) and tetrapodal pentadentate ligands (2,15 316) in ligand 4.

400 spectrometer (cf. Fig. 4). EI-MS measurements were con-

ducted using a Finnegan MAT95S instrument. ESI-MS source

spectra and collision induced dissociation (CID) spectra were

acquired on a Thermo LTQ Orbitrap XL FT-spectrometer

(XCalibur software). Combustion analysis was performed using a

Thermo Finnigan Flash 1112 apparatus (EAGER 300 software).

IR spectra were measured with KBr disks on a Perkin–Elmer Spec-

trum 100 Series FT-IR spectrometer. A PAR263A potentiostat was

used for electrochemical data acquisition. The quality of iR-drop

compensation was analyzed using the RL0 algorithm.20 Melting

and decomposition points were determined on a Koﬂer apparatus

and calibrated against azobenzene, acetanilide and phenacetine,

as applicable. X-ray diffraction: Diffraction data were obtained

on an Oxford Diffraction Xcalibur S instrument equipped with

a Sapphire CCD-detector using graphite-monochromated Mo-

Karadiation with l=0.71073 A

˚. Suitable crystals were attached

to glass ﬁbers using perﬂuoropolyalkylether oil and transferred

to a goniometer where they were cooled to 150 K for data

collection. Software packages used: CrysAlis CCD for data

collection, CrysAlis Pro (for cell reﬁnement and data reduction).

Empirical absorption corrections (multiscan based on symmetry-

related measurements) were applied using CrysAlis Pro software.21

Space groups were determined from systematic absences and

conﬁrmed by the successful solution of the structure. Structures

were solved by direct methods using the program Sir2004.22

Subsequent full-matrix least-squares reﬁnement of Fo2data was

carried out using SHELXL-97.23 All non-hydrogen atoms were

reﬁned anisotropically. Treatment of hydrogen atoms: Carbon-

bound hydrogen atoms were placed in positions of optimized ge-

ometry, and isotropic displacement parameters were tied to those

of the corresponding carrier atoms. Oxygen- and nitrogen-bound

hydrogen atoms were located in the difference Fourier map and

freely reﬁned. Hydrogen atoms belonging to water molecules were

reﬁned using distance and angle restraints. Thermal displacement

parameters of hydrogen atoms were tied to those of their carrier

atoms. Supplementary crystallographic data can be obtained free

of charge from The Cambridge Crystallographic Data Centre,

CCDC 825623 (10) and CCDC 825624 (11). Molecular graphics

were created using Ortep 3v2.24

Methylmalonamide (6)

2-Pyridone (1 g, 1 mol%)25 was dissolved in 30% aqueous ammonia

(400 mL), diethyl methylmalonate (5) added (200 mL, 1.16 mol,

202 g), and the biphasic reaction mixture stirred vigorously

overnight. A ﬁne colourless precipitate was collected on a Buchner

funnel, washed with ethanol, and dried overnight in air. Colourless

solid (80 g, 60%).

Mp. 209–211 ◦C (lit.26 210 ◦C). Anal. found: C, 41.67, H, 7.11,

N, 24.19%, calcd for C4H8N2O2: C, 41.37, H, 6.94, N, 24.12%. 1H-

NMR (d6-DMSO): 7.23 (s br, 2 H, NH), 6.98 (s br, 2 H, NH), 3.06

(q, 3J(HH) =7.2Hz,1H,CH),1.55(d,3J(HH) =7.2 Hz, 3 H,

CH3); 13C-NMR (d6-DMSO): 172.4 (CO), 46.7 (CH), 15.0 (CH3);

EI-MS (70 eV, 120 ◦C): 116 (M+,~3%), 99 (M -NH3,~2%), 73

(M -HOCN, 100%), 55 (M -HOCN/CO, 20%), 44 (CONH2+,

50%); IR: 1674 cm-1(CO).

Methylmalononitrile (7)

The compound was prepared according to a literature procedure26

by reactive vacuum dehydration/distillation of methylmalon-

amide (6) over phosphorous pentoxide at 15 mm Hg/250 ◦C.

Colourless solid, max. yield 65%.

Mp. 35 ◦C (lit.27 32–36 ◦C). Anal. found: C, 59.82, H, 4.55, N,

35.18%, calcd for C4H4N2: C, 59.99, H, 5.03, N. 34.98%. 1H-NMR

(CDCl3): 3.80 (q, 3J(HH) =7.3 Hz, 1H, CH), 1.79 (d, 3J(HH) =7.3

Hz, 3 H, CH3); 13C-NMR (CDCl3): 113.0 (CN), 16.7 (CH), 16.5

(CH3); EI-MS (70 eV, 29 ◦C): 79 (M–1, 40%), 53 (M -HCN, 95%),

28 (HCNH+, 100%); IR: 2262 cm-1(CN) (lit.27 2255–2260 cm-1).

2478 |Dalton Trans., 2012, 41, 2477–2485 This journal is ©The Royal Society of Chemistry 2012

Published on 05 January 2012. Downloaded by TU Berlin - Universitaetsbibl on 31/03/2016 08:28:14.

View Article Online

2,4-Dichloropyrimidine (9)

In a 250 mL ﬂask, uracil (5.0 g, 45 mmol) and tetraethyl

ammonium chloride (1.1 g, 15 mol%) were suspended in phos-

phorous oxychloride (10 mL), and heated to 120 ◦Cfor1h

(CAUTION: vigorous evolution of hydrogen chloride gas). The

mixture was subsequently allowed to cool, while stirring was

continued overnight. The brown solution was poured on crushed

ice (100 g) and stirred until a yellowish solid appeared. The solid

was collected at 0 ◦C by ﬁltration (CAUTION: The hydrolysis

of phosphorous oxychloride at 0 ◦C is incomplete; a strongly

exothermic reaction occurs in the aqueous ﬁltrate upon warming)

and washed with ice water, dissolved in chloroform (80 mL), dried

over sodium sulfate and ﬁltered through neutral alumina (II–

III). After volatiles had been removed in vacuo, the residual oil

crystallized upon cooling. Yellowish rods (2.4 g, 36%).

Mp. 61 ◦C (lit.28 61–62 ◦C). Anal. found: C, 32.89, H, 1.37, N,

18.88%, calcd for C4H2N2Cl2: C, 32.25, H, 1.35, N, 18.80%. 1H-

NMR (CDCl3): 8.53 (d, 3J=5.3 Hz, 1 H), 7.35 (d, 3J=5.3Hz,1H);

13C-NMR (CDCl3): 162.7 and 161.0 (CCl), 160.0 and 120.3 (CH);

HR-EI-MS (370 K) M+calcd: 147.95950 m/z; found: 147.95919

m/z.

2,2¢-(Pyrimidine-2,4-diyl)bis(2-methylmalononitrile) (10)

Under an atmosphere of dry dinitrogen, commercial lithium

hydride (640 mg, 80 mmol) was suspended in dry THF (20 mL) at

20 ◦C with stirring, in a reaction ﬂask on a water bath. A solution

of methylmalononitrile (7) (6.57 g, 82 mmol) in THF (15 mL)

was added during 15 min, which resulted in hydrogen evolution.

The temperature was maintained between 25 and 35 ◦C during

addition. After the evolution of hydrogen had ceased, a slightly

yellow solution had formed. The water bath was replaced with

an oil bath, and a solution of 2,4-dichloropyrimidine (9) (5.97 g,

40 mmol) in THF (5–10 mL) was added rapidly. The exothermic

reaction warmed the mixture to 70 ◦C instantaneously, and the

oil bath was then heated to hold the reaction mixture at reﬂux

temperature for an additional 30 min. A precipitate formed in the

course of the reaction. The mixture was allowed to cool to room

temperature and then quenched with water (10 mL) to dissolve

precipitated lithium chloride. Concentrated aqueous ammonium

chloride solution (20 mL) and dichloromethane (50 mL) were

added subsequently. The phases were separated, and the aqueous

phase extracted with additional dichloromethane (20 mL). The

combined extracts were dried over sodium sulfate, ﬁltered over

neutral alumina (II–III), and volatiles were removed in vacuo.The

remaining oil (8.9 g, 94%) crystallized rapidly. Melting point and

elemental analysis data (below) were determined for this material,

and it was also used in the subsequent synthetic steps without

further puriﬁcation. The material can be recrystallized from a ten-

fold amount of benzene to yield 10 as a colourless crystalline solid

(cf. Table 1).

Mp. (unrecrystallized material): 112 ◦C; Anal. found (unrecrys-

tallized material): C, 60.13, H, 3.47, N, 35.25%, calcd for C12H8N6:

C, 61.01, H, 3.41, N, 35.58%. 1H-NMR (CDCl3): 9.12 (d, 3J=5.2

Hz, 1 H, CH), 7.89 (d, 3J=5.3Hz,1H,CH),2.29(s,3H,

CH3), 2.26 (s, 3 H, CH3); 13C-NMR (CDCl3): 162.8 and 161.9

(quaternary aromatic C), 161.6 and 118.1 (CH), 113.7 and 113.2

(CN), 40.3 and 37.9 (quaternary aliphatic C), 26.4 and 26.3 (CH3);

HR-EI-MS (370 K): M+calcd: 236.08104 m/z; found: 236.07989

m/z; IR: 2260 cm-1(CN).

2,2¢-(Pyrimidine-2,4-diyl)bis(N¢1,N¢3-dihydroxy-2-

methylmalonamidoxime) (4)

Unrecrystallized nitrile 10 (7.91 g, 33 mmol) was crushed into

small pieces and suspended in ethanol (50 mL; puriﬁed to remove

2-butanone). Aqueous hydroxylamine (50 wt%, 10 mL, 166 mmol)

was added. The nitrile began to dissolve after a brief initiation

period, and an exothermic reaction brought the mixture to reﬂux.

A precipitate formed. The mixture was left to cool and stirred

for an additional 24 h at room temperature. The precipitate was

collected, washed with a small amount of ethanol and dried in air.

Very ﬁne colourless powder (10 g, 85%), decomposes at ~190 ◦C.

Table 1 Summary of crystallogaphic data

Empirical formula C12H8N6(10)C

32 H62 Co2F12 N20 O28 (11·12H2O)

Formula weight 236.34 1520.88

T/K 150(2) 150(2)

l/A

˚0.71073 0.71073

Crystal system Monoclinic Triclinic

Space group P21/nP1

a/A

˚,b/A

˚,c/A

˚6.6056(2), 10.9046(4), 16.8884(5) 9.2213(4), 12.9948(6), 13.4621(8)

a(◦), b(◦), g(◦) 90, 97.859(3), 90 69.118(5), 83.997(4), 73.435(4)

U/A

˚31205.07(7) 1444.63(12)

Z41

Dc/Mg m-31.302 1.748

m/mm-10.086 0.716

F(000) 488 780

Crystal size/mm 0.60 ¥0.33 ¥0.30 0.40 ¥0.22 ¥0.11

Hrange/◦3.07 to 25.00 3.28 to 25.00

Range h,k,l-7to7,-12 to 12, -17 to 20 -10 to 10, -12 to 15, -11 to 15

Independent reﬂections 2107 [R(int) =0.0185] 5065 [R(int) =0.0292]

Data/restraints/parameters 2107/16/181 5065/110/535

Goodness-of-ﬁt on F21.045 0.973

Final Rindices [I>2s(I)] R1=0.0349, wR2=0.0760 R1=0.0389, wR2=0.0892

Rindices (all data) R1=0.0473, wR2=0.0800 R1=0.0534, wR2=0.0928

Largest diff. peak and hole/e A

˚-30.148 and -0.146 0.555 and -0.327

This journal is ©The Royal Society of Chemistry 2012 Dalton Trans., 2012, 41, 2477–2485 | 2479

Published on 05 January 2012. Downloaded by TU Berlin - Universitaetsbibl on 31/03/2016 08:28:14.

View Article Online

Anal. found: C, 40.03, H, 5.73, N, 37.27%, calcd for

C12H20N10O4: C, 39.13, H, 5.47, N, 38.03%. 1H-NMR (d6-DMSO):

9.36 (s br, 2 H, OH), 9.23 (s br, 2 H, OH), 8.65 (d, 3J(HH) =

5.3Hz,1H,CH),7.31(d,3J(HH) =5.3Hz,1H,CH),5.32

(s br, 4 H, NH), 5.23 (s br, 4 H, NH), 1.71 (s, 3 H, CH3),

1.65 (s, 3 H, CH3); selected 2D 1H-1H NOESY cross peaks:

7.31/1.65 (pyrimidine-C5-H/methyl), 1.65/5.32 (methyl/amino),

5.32/9.36 (amino/hydroxyl); 13C-NMR (d6-DMSO): 169.67 and

168.21 (quaternary aromatic C), 157.26 (CH), 155.18 and 154.53

(quaternary oxime C), 119.21(CH), 54.30 and 52.74 (quaternary

aliphatic C), 22.11 and 21.89 (CH3); HR-ESI-MS (RT): MH+

calcd: 369.17418 m/z; found: 369.17334 m/z; IR: 1676 and 1657

(C Noxime), 947 and 935 (N–O).

{CoIII

2[4–H]2}(OCOCF3)4(11)

In air, solid 4(830 mg, 2.25 mmol) and solid cobalt(II) acetate

tetrahydrate (560 mg, 2.25 mmol) were dissolved in a stirred

mixture of triﬂuoroacetic acid (2 mL) and ethanol (10 mL) at

room temperature. The brown solution that formed was stirred

for 1 h and then poured on cold diethylether (40 mL, 5 ◦C). An

orange–brown solid was collected by ﬁltration, washed with cold

diethylether (30 mL), and dried in vacuo (690 mg, 47%). Additional

product can be obtained by slow evaporation of the ﬁltrate. The

exchange of acetate for triﬂuoroacetate proceeds to approx. 93%,

as estimated from the resonance of remaining acetate in the 1H-

NMR spectrum. 11 decomposes at 150 ◦C.

Anal. found: C, 29.70; H, 2.65; N, 20.88%, calcd for C32 H38

Co2F12 N20 O16: C, 29.46; H, 2.94; N, 21.47%. HR-ESI MS

(MeOH; m/z;%): [CoIII2(4-2H)2]2+ (425.08(3); 100%), {[CoIII(4

-2H)]2(OCOCF3)}+(963.15(3); 10%), [CoIII2(4-2H)(4-3H)]+

(849.15(9); 5%), {CoIII(4-H)(OCOCF3)}+(539.07(6); <1%);

1H-NMR (d6-DMSO):9.70(m,1H,OH),9.09(d,3J(HH) =

6Hz,1H,CH),7.85(d,3J(HH) =6 Hz, 1 H, CH), 8.20,

7.54, 7.35, 7.08 (four s br, 8 H, NH), 2.12, 2.08 (two s br, 6 H,

CH3). The compound was dissolved in absolute triﬂuoroacetic

acid, dried in vacuo and crystallized by slow evaporation of a

highly dilute aqueous solution at 5 ◦C to obtain single crystals

of the dodecahydrate (11·12H2O) suitable for X-ray diffraction

measurements (cf. Table 1).

Results and discussion

Ligand and complex synthesis

Methylmalononitrile (7) was prepared from 5in a sequence of

ester ammonolysis and dehydration.26 The published procedure

for the preparation of methylmalondiamide (6) was modiﬁed in

two ways: 2-Pyridone was used as a bifunctional catalyst,25 and

aqueous ammonia as the solvent instead of a mixture of methanol

and liquid ammonia. This decreases the yield (from 99%, as

published, to 60%), but the reaction time is shortened considerably

(from several weeks to 12 h). Uracil (8) is converted to 2,4-

dichloropyrimidine (9) by the action of phosphorous oxychloride.

We modiﬁed the available procedures28 (see Experimental section)

by using tetraalkylammonium chlorides instead of aniline deriva-

tives as additives. Speciﬁcally, the reaction product contained less

of a lachrymatory impurity when the exposure of uracil to hot

phosphorous oxychloride was limited to 1 h.

Chlorodiazines are highly reactive with respect to nucleophilic

aromatic substitution of the Cl substituents by deprotonated

geminal dinitriles.14 In our case, the reaction of the lithium salt

of methylmalononitrile with 2,4-dichloropyrimidine (9)inTHF

produces tetranitrile 10. The lithium salt of 7is conveniently

prepared at temperatures above 15 ◦C by the reaction with

commercial lithium hydride. The reaction is relatively insensitive

to moisture which can be attributed to the low pKaof geminal

dinitriles,29 their high nucleophilicity30 and the essential inertness

of commercial LiH.31 10 is converted to the corresponding

tetraamidoxime (4) by reaction with aqueous hydroxylamine in

ethanol (Fig. 2). Although the addition of hydroxylamine to this

nitrile is an exothermic reaction, the four-fold addition goes to

near completion only upon prolonged exposure of the tetranitrile

to hydroxylamine solution at room temperature, owing to the low

solubility of the higher addition products. Combustion analysis

and ESI-MS indicate that three-fold addition products are present

in the compound as minor impurities. 1H-NMR spectroscopy

(d6-DMSO) shows well-resolved signals of the magnetically in-

equivalent methyl, hydroxyl and amide groups (see Supporting

information†). Ligand 4is sparingly soluble in common laboratory

solvents and may be stored in air for extended periods of time

without decomposition. It decomposes with evolution of gas

when heated above 190 ◦C. Although no adverse behaviour was

observed in the course of our study, it should be noted that oximes

generally pose a ﬁre and explosion hazard. Attempts to prepare

cobalt(II) compounds of 4under conditions otherwise known to

produce cobaloximes failed, and we propose reasons for this in

the following. In air, however, a rapid reaction occurs between

cobalt(II)acetate tetrahydrate and 4, and a cobalt(III) compound is

formed. In the presence of excess triﬂuoroacetic acid, the carboxy-

late counter ions are exchanged, and 11 is obtained which forms

single crystals of its dodecahydrate upon slow evaporation of water

from the aqueous solution at 5 ◦C(11·12H2O, cf. Fig. 3 and Fig. 6).

Fig. 2 i) NH4OH, cat. 2-pyridone, rt, 24 h, 60%; ii) P2O5on sand, 250 ◦C, 15 torr, 65%; iii) POCl3,NEt

4Cl, 120 ◦C, 1 h, 36%; iv) LiH, THF, reﬂux,

45 min, 94%; v) NH2OH, EtOH, 24 h, 85%.

2480 |Dalton Trans., 2012, 41, 2477–2485 This journal is ©The Royal Society of Chemistry 2012

Published on 05 January 2012. Downloaded by TU Berlin - Universitaetsbibl on 31/03/2016 08:28:14.

View Article Online

Fig. 3 Three different views of the disordered structure of the formal tetracation in 11·12H2O (ellipsoids at the 50% probability level; cf. Fig. 5; primed

symbols refer to equivalent positions 1 -x,1-y,1-z).

Oxidation pathways

The ligand does not dissolve in the reaction mixture when its com-

plexation is attempted in the absence of oxygen. When micromolar

amounts of 4and cobaltous acetate are mixed in a 1 : 1 molar

ratio at high dilution in the absence of additional triﬂuoroacetic

acid in air, an instantaneous reaction occurs, producing an orange

solution. ESI mass spectra of this crude mixture are dominated by

an ion of the composition [CoIII2(4-2H)2]2+ which is also dominant

in ESI spectra of the puriﬁed compound. Additionally, a trinuclear

cluster [Co3(O2)(4-2H)2]+accompanied by a protonated dication

[Co3(O2)(4-2H)(4-H)]2+ is observed which is not present in

the spectra of material prepared in the presence of triﬂuoroacetic

acid. CID (collision-induced dissociation) experiments with the

monocation reveal interesting behaviour, as only the simultaneous

loss of two water molecules is observed at high collision energies,

whereas a signal corresponding to the loss of a single water

molecule is not observed. It is therefore unlikely that 4ﬁrst

coordinates Co(II) in square-pyramidal fashion and then reacts

with molecular oxygen. Rather, the data points to a complex acid-

promoted oxidation mechanism involving cluster compounds that

are ﬁnally trapped in the form of the very stable “quasi-cryptand”

dimer described in the following. The cobalt(III) compound shows

an irreversible reduction peak in its cyclic voltammogram (dry

DMF, 298 K, 50–800 mV s-1, glassy carbon working electrode) at

voltages more negative than –2.7 V vs. ferrocene. This unexpected

recalcitrance with respect to stabilization of cobalt(II)canbe

explained by the lack of p-backbonding in a,g-dioximes and the

rigidity of the dimer which cannot easily accommodate an increase

in ionic radius.

Structural analysis (part 1)

X-ray structural analysis revealed a complex of the composition

[CoIII2(4–H)2](OCOCF3)4·12H2O(11·12H2O). The dinuclear tetra-

cation contains two capping ligands, each of which is deprotonated

at one of its hydroxyl groups, forming a bis-m-amidoximato-

bridged core (Fig. 3). The orientation of the pyrimidine heterocycle

(the uncoordinated nitrogen atom may be in one of two different

positions) could not be assigned unambiguously on the basis of

the diffraction data. In P1

, the asymmetric unit contains one

disordered half of the dimer that is related to the other half by

a centre of symmetry. A similar disorder was encountered during

reﬁnement of the crystal structure of the tetranitrile precursor

10 (see Supporting information†). Both possible orientations

(with respect to one cobalt ion) produced unexceptional thermal

ellipsoids in P1aswellasinP1

¯but, in the difference Fourier

map, had residual electron density in the vicinity of the atom

assigned as carbon (cf. Fig. 4). Although the structure, as discussed

in the following, represents an average, the thermal ellipsoids

are unexceptional. Disorder in 11 can be explained by invoking

the presence of four stereoisomeric forms. Of these, two are

centrosymmetric (cf. Fig. 4, B and C), and the remaining two

are a pair of enantiomers (cf. Fig. 4, A and A*). In the following,

we discuss three limiting cases, i)–iii); i) Only one enantiomer,

A or A*, is present (spontaneous racemate resolution). Then, a

solution in P1

¯necessarily leads to a diagnosis of pseudosymmetry;

ii) both A and A* are present. In this case, a solution in P1

¯leads

to a diagnosis of pseudosymmetry and disorder; the disorder,

however, could only be solved in the lower-symmetric space group

P1; iii) all four stereoisomeric forms are present, such that the

observed pseudosymmetry contains both point groups, C1and

Ci. Then, the observed stereoisomerism cannot be resolved in

a centrosymmetric space group due to the presence of non-

centrosymmetric enantiomers. For this reason, the site occupancy

factors of the isomers were reﬁned in P1 for comparison. The chiral

forms A and A* (Fig. 4) converged to approximately 75% of the

whole, indicating isomerism. By restraining the occupation of the

1- and 5-positions of the pyrimidine ring to 50 : 50 (N : CH) each, a

disordered solution in P1

¯was obtained as a compromise, because

reﬁnement in P1 was not feasible without strongly restraining the

atomic displacement parameters. Within this solution, identical

CH distances (within estimated standard deviations) were deter-

mined for both disordered positions, whereas H atoms collapsed

on their carrier atoms, with unrealistic distances, when the

disorder was ignored. This solution is compatible with all limiting

cases i)–iii).

This journal is ©The Royal Society of Chemistry 2012 Dalton Trans., 2012, 41, 2477–2485 | 2481

Published on 05 January 2012. Downloaded by TU Berlin - Universitaetsbibl on 31/03/2016 08:28:14.

View Article Online

Fig. 4 Electron density maps of the pyrimidine plane in 10 and 11·12H2O; (a) experimental electron density; (b/c) residual density resulting from ignored

disorder; (d) minor ﬂuctuation in the difference density maps of the solutions incorporating disorder, with peaks occurring at bonding critical points

(note the different scales). The position of the cobalt ion (residual electron density) can be seen in ﬁgure (a) (at right). (e) Top: Possible stereoisomers in

11 (mdenotes a mirror plane, ia centre of symmetry and A* an enantiomer). Below: Symmetry interpretation of the 13C{1H}-NMR (100 MHz) signals

assigned to amidoxime groups of 11 in D2O(cf. Fig. 3).

Coordination isomerism and 13C{1H}-NMR (in D2O)

Unlike ligand 4, ligands of the tetrapodal pentadentate class usu-

ally have C2v point group symmetry, which renders all equatorial

donor functions magnetically equivalent. In ligand 4,the“true”

point group is Cs, the only symmetry element other than identity

being a mirror plane which coincides with the pyrimidine ring

plane, i.e., the two-fold rotation axis, as found in C2v, is missing.

However, the breaking of symmetry in 4occurs not in the ﬁrst coor-

dination sphere but in the ligand periphery, as all equatorial donor

functions are amidoxime groups. Therefore, we may call 4“pseudo-

C2v symmetric”, i.e. dimerization can still occur with monomers

having their pyrimidine rings in every possible orientation (cf.

Fig. 4). For a truely C2v symmetric ligand, dimerization of the

present kind would result in a dimer belonging to point group

Ci. This means that in each monomeric motif, all four amidoxime

functions would have a slightly different chemical environment,

but each environment would occur twice per molecule, as a

consequence of inversion symmetry. Consequently, this point

group occurs only in two out of four conceivable cases for a

“pseudo-C2v symmetric”, in actual fact Cssymmetric, ligand.

These two cases are composed of two different isomers that have

different chemical environments. The other two cases show C1

point group symmetry and are a pair of enantiomers, such that

all eight amidoxime functions have slightly different chemical en-

vironments. Taking into consideration the previous explanations

for an isomeric mixture of four different compounds, 16 different

13C{1H}-resonances belonging to amidoxime functions are to be

expected, four for each of the two Cidimers and eight for each

enantiomer, indistinguishable from one another in achiral media.

This behaviour can indeed be observed over a chemical shift

range of 7 ppm (cf. Fig. 4). The 16 observed signals can easily be

ordered into four groups, based on their chemical shifts, line widths

andﬁttedintegrals(±10% maximum integration deviation). The

four groups correspond to four different chemical environments

expected for the simpliﬁed case of a single molecule in the Cipoint

group. Within each group there are two signals of comparable

intensity. In the context of the previous rationale, these can

be assigned to the C1forms, since eight magnetically different

amidoxime groups are to be expected in the enantiomers, which

have identity (E) as the only symmetry element. Moreover, each

of the two signals of comparable intensity within each of the

four groups resembles another signal in terms of chemical shift,

that consequently belongs to one of the Cisymmetric forms.

This observation can be elegantly explained, as each half of the

enantiomers resembles the orientation in one of the Ciisomers.

2482 |Dalton Trans., 2012, 41, 2477–2485 This journal is ©The Royal Society of Chemistry 2012

Published on 05 January 2012. Downloaded by TU Berlin - Universitaetsbibl on 31/03/2016 08:28:14.

View Article Online

Further, this hypothesis is consistent with the chemical shift

difference observed in the Cs-symmetric ligand 4. In conclusion,

the appearance of the amidoxime resonances in the 13C{1H}-

spectrum of 11 is fully consistent with the picture of isomerism as

deduced from the X-ray diffraction data. The distribution of these

isomers is not fully probabilistic, as in this case all 16 signals would

be of equal intensity due to equal population of the four forms.

Structural analysis (part 2)

In the structure of 11 (Fig. 3), the apical pyrimidine donor N4 oc-

cupies the position trans to the m-oximato oxygen atom. The cation

charge is compensated by four triﬂuoroacetate counterions which

form intermolecular hydrogen bonds to the cation’s hydroxyl and

amino functions. The monomeric motif is self-complementary; in

the dimer, the metal ions are coordinatively saturated in reciprocal

fashion by deprotonated hydroxyl groups of the other half of the

molecule.

Most notably, the structure of 11 is reminiscent of cryptate

complexes, with the two cobalt(III) ions completely shielded from

their surroundings as a consequence of dimerization. The geom-

etry which results from dimerization is best described in terms

of a least-squares planes analysis. In the asymmetric octahedral

motif, three best planes can be deﬁned by the coordinates of N2,

N5, N6, N7 (plane 1), N2, O3¢, N6, N4 (plane 2) and O3¢,N5,

N4, N7 (plane 3) (cf. Fig. 5). The cobalt ion surrounded by these

atoms deviates from the planes by 0.0683(11) A

˚, 0.0578(12) A

and 0.0184(11) A

˚, respectively, while the interplanar angles are

86.63(06)◦(planes 1/2), 88.41(06)◦(planes 2/3), and 89.85(06)◦

(planes 3/1). This characterizes a slightly distorted octahedral

coordination environment around cobalt. The symmetry-related

second cobalt ion (by inversion) is displaced from the planes

by 2.7209(24) A

˚, 0.8478(30) A

˚and 2.6908(21) A

˚, respectively,

which deﬁnes a “slipped cofacial dimer”. This term was coined

previously32 and indicates that there are pairs of parallel faces in

two octahedra (in the present case due to inversion symmetry)

while the centres of the octahedra are displaced (“slipped”) with

respect to one another.

Two other planes can be deﬁned by the atomic coordinates of

N2, Co1, O3¢and N2, O3, O3¢,N2¢to further characterize the out-

of-plane displacement of the bis-m-oximato bridge (see Fig. 5). The

angle deﬁned by these planes is 43.21(10)◦. The bridging motif

Table 2 Selected interatomic distances (A

˚) and angles (◦)for11·12 H2O

Co1 ◊◊◊Co1¢3.812(1) N2–O3–Co1¢a119.33(14)

Co1–N2–O3 122.14(15)

Co1–N2 1.907(2) O3¢–Co1–N2 90.93(8)

N2–O3 1.385(3) Co1–N2–O3–Co1¢67.9(2)

O3¢–Co1 1.919(2) N4–Co1–O3¢173.73(9)

Co1–N4 1.909(2) N4–Co1–N2 89.39(9)

Co1–N6 1.911(2) N4–Co1–N6 88.61(9)

Co1–N5 1.919(2) N4–Co1–N5 86.71(9)

Co1–N7 1.913(2) N4–Co1–N7 87.11(9)

O8 ◊◊◊O3 2.726(3) N5–Co1–N2–O3 45.7(2)

O10 ◊◊◊O11 2.587(4) O3–H8–O8 162(4)

O3¢◊◊◊O10 2.848(2) O11–H10–O10 169(4)

Intermolecular distances:

O11 ◊◊◊O31 2.490(1) N13 ◊◊◊O400 2.915(1)

N14 ◊◊◊O32 2.902(1) O400 ◊◊◊O500 2.840(1)

N14 ◊◊◊O800 2.928(1) O500 ◊◊◊O800 2.717(2)

aThe angle N2–O3–Co1¢deﬁnes the parameter a, which has been used to

characterize bis-m-oximato bridging (see text).33 Primed symbols refer to

equivalent positions 1 -x,1-y,1-z.

can be clearly distinguished from the type of out-of-plane bis-

m-oximato bridging found in copper glyoxime compounds where

the relation is approximately rectangular.33 Additional interatomic

distances and angles are given in Table 2.

Intramolecular H-bonding in the solid state

The hydrogen bonds in 11 are longer than in nickel dimethylgly-

oxime (O ◊◊◊O: 2.45 A

˚34) and not equivalent. However, there is no

doubt that hydrogen bonding is present as the O ◊◊◊O interatomic

distances are signiﬁcantly smaller than the sum of the van der

Waals radii (2 ¥1.52 A

˚35 vs. 2.73 and 2.59 A

˚, respectively; see

Table 2). Geometric variations of hydrogen bonding param-

eters can arise due to secondary interactions with water of

crystallization,36 but in the present case the hydrogen bonding

pattern formed by the oxime functions is rather characteristic

of a bis-m-oximato bridged dimer. The hydrogen bond formed

between the bridging oxygen atom and the adjacent hydroxyl

group has a larger interatomic distance (“central” O3 ◊◊◊O8)

than the hydrogen bond formed between the non-coordinating

hydroxyl functions (“peripheral” O11 ◊◊◊O10). This behaviour is

also observed in copper dimethylglyoxime37 and can be explained

by a decreased hydrogen bond donor capability of the coordinated

oxygen atom (O3¢). Further, as there is a short contact between

Fig. 5 Simpliﬁed representation of the coordination octahedra and H-bonding in [CoIII2(4–H)2](OCOCF3)4·12H2O] (outer coordination sphere omitted

for clarity). A) Intramolecular H-bonding pattern in the dimer. B) Intermolecular H-bonding pattern in the asymmetric unit (this motif, when doubled,

produces the structure of the whole molecule).

Published on 05 January 2012. Downloaded by TU Berlin - Universitaetsbibl on 31/03/2016 08:28:14.

View Article Online

Fig. 6 Schematic representation of cation structures and hydrogen bond motifs in 11·12H2O.

O3¢and O10 (cf. Table2andFig.5),thepresenceofbifurcatedH-

bonds involving the hydroxyl group O10–H10 and oxygen atoms

O11/O3¢(and symmetry equivalents) cannot be ruled out. Such

hydrogen bonding may provide an additional link between the two

halves of the molecule.

Intermolecular H-bonding in the solid state

The presence of amidoxime functions makes intermolecular

hydrogen bonding in 11·12H2O diverse. Two of the four triﬂu-

oroacetate anions per formula unit form strong hydrogen bonds

to the hydroxyl functions serving as acceptors for the peripheral

macrocyclic intramolecular hydrogen bonds (cf. Fig. 5 B, Fig. 6).

The short contact indicates that triﬂuoroacetate may, alternatively,

be interpreted as triﬂuoroacetic acid (TFA) acting as a hydrogen

bond donor (instead of triﬂuoroacetate acting as an acceptor).

Based on this argument, the compound can be reformulated

as a triﬂuoroacetic acid adduct of the composition {[CoIII(4-

2H)]2[OCOCF3]2·2TFA}. This explains why the dication [CoIII(4

-2H)]22+ is observed in ESI mass spectra as the dominant signal,

whereas a tetracation is not observed (see Experimental section).

The other two triﬂuoroacetate anions form weak hydrogen bonds

to an amino function (not shown in Fig. 5). Six of the twelve water

molecules of crystallization form two ‘hydrogen bond arches’,

related by inversion symmetry, which connect the amino functions

of the methylmalonamidoxime residue that does not contain the

bridging m-oximato function (cf. Fig. 5 A/B, Table 2). The residual

six water molecules form hydrogen bonds with the weakly bound

triﬂuoroacetate ions. There are no hydrogen bonds involving the

non-coordinating nitrogen atoms of the pyrimidine ring.

Conclusions

Compound 4binds to cobalt(III) ions in tetrapodal pentadentate

fashion, with formation of a dinuclear complex containing a six-

membered out-of-plane bis-m-amidoximato metallacycle and a

macrobicyclic structure based on hydrogen bonds. This unex-

pected dimerization results in efﬁcient shielding of the central

metal ions. The inﬂuence of coordination isomerism, which

derives from the use of a pyrimidine core (as opposed to an

azine of higher symmetry38), on the overall geometry of the

complex cation is subtle. The presence of triﬂuoroacetic acid

during complexation does not impede the single deprotonation

of the capping ligand. Overall, dimerization produces a formal

tetracation [Co2(4-H)2]4+, with both strong and weak hydrogen

bonds to its triﬂuoroacetate counterions in the solid state. In

ESI mass spectra, the predominant ion is the dication [Co2(4-

2H)2]2+ which is a result of further deprotonation of the ligand.

Mononuclear ions are observed but their intensity is negligible (cf.

Experimental section). We compare the structure of the dimer to

that of copper glyoxime dimers, as there appear to be no literature

examples of related bis-m-oximato-bridged cobalt complexes of

dioximes. An explanation for their absence might be that, in

cobaloximes, signiﬁcant out-of-plane placement of the hydroxyl

functions is regularly accompanied by dislocation of the metal

ion in the opposite direction.39 Yet, since malonamidoxime itself,

which can be viewed as the parent compound of the present

ligand, shows out-of-plane location of the hydroxyl function in

combination with in-plane location of the metal ion in certain

metal complexes,11 it can be assumed that this particular motif is

prone to form in cobalt(III) compounds of 1,3-dioxime ligands.

In summary, amidoxime ligands of this kind indeed combine

certain properties of tetrapodal pentadentate ligands and macro-

cyclic oxime ligands. However, while amidoxime functions are

known to act as versatile bridging donors,40 the encapsulation of

the metal ions upon dimerization in 11 is an example of synergism.

In effect, compound 4can be considered an “open prison ligand”41

or a “pseudo-cryptand”, as the shielding depends on the self-

complementarity of the mononuclear ion-ligand fragment rather

than on the ligand alone.

Future studies will focus on the exploration of other

transition metal complexes of tetrapodal amidoxime ligands,

their reactivity, and ligand variations in order to prevent

dimerization/coordination isomerism as observed in the present

case. The potential of such complexes in electrocatalytic water

splitting will be studied.42,43

Published on 05 January 2012. Downloaded by TU Berlin - Universitaetsbibl on 31/03/2016 08:28:14.

View Article Online

Acknowledgements

Support by the Deutsche Forschungsgemeinschaft (GR 1247/7-

1) is gratefully acknowledged. Thomas Hamﬂer and Manfred

Detlaff are thanked for technical assistance, and Dr Josef Seiffert,

Dr Gerald H¨

orner, Dr Maria Schlangen and Professor Aris

Chatzidimitriou-Dreismann are thanked for helpful discussions.

References

1 D. J. Wesylenko, C. Ganesamoorthy, J. Borau-Garcia and C. P.

Berlinguette, Chem. Commun., 2011, 47, 4249–4251.

2(a)A.Fihri,V.Artero,M.Razavet,C.Baffert,W.LeiblandM.

Fontecave, Angew. Chem., 2008, 120(3), 574–577, (Angew. Chem., Int.

Ed., 2008, 47, 564); (b) P.-A. Jacques, V. Artero, J. P´

ecaut and M.

Fontecave, Proc. Natl. Acad. Sci. U. S. A., 2009, 106(49), 20627–20632.

3 H. M. Colquhoun, J. F. Stoddart and D. J. Williams, Angew. Chem.,

1986, 98(6), 483–503, (Angew. Chem., Int. Ed. Engl., 1986, 25, 487).

4 R. L. Shook and A. S. Borovik, Inorg. Chem., 2010, 49(8), 3646–3660.

5 D. Natale and J. C. Mareque-Rivas, Chem. Commun., 2008, 425–437.

6 L. Caron, H. Bricout, S. Tilloy, A. Ponchel, D. Landy, S. Fourmentin

and E. Monﬂier, Adv. Synth. Catal., 2004, 346(12), 1449–1456.

7(a) C. Squires, C. W. Baxter, J. Campbell, L. F. Lindoy, H. McNab, A.

Parkin,S.Parsons,P.A.Tasker,G.WeiandD.J.White,Dalton Trans.,

2006, 2026–2034; (b) R. S. Forgan, B. D. Roach, P. A. Wood, F. J. White,

J. Campbell, D. K. Henderson, E. Kamenetzky, F. E. McAllister, S.

Parsons, E. Pidcock, P. Richardson, R. M. Swart and P. A. Tasker,

Inorg. Chem., 2011, 50(10), 4515–4522.

8 A. Grohmann, Dalton Trans., 2010, 39, 1432–1440.

9(a)J.R´

etey, Helv. Chim. Acta, 1971, 54(8), 2747–2752; (b)M.W.Bartlett

and J. D. Dunitz, Helv. Chim. Acta, 1971, 54(8), 2753–2756; (c)W.

Pannhorst, Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem.,

1977, 33, 2384–2388.

10 I. Ruff and K. Burger, Acta Chim. Acad. Sci. Hung., 1966, 47, 143–155.

11 H. Endres and N. Kadkodai-Elyaderani, Z. Naturforsch., 1982, 37b,

1255–1265.

12 H. Schmidtmann, Ber. Dtsch. Chem. Ges., 1896, 29(2), 1168–1175.

13 F. Eloy and R. Lenaers, Chem. Rev., 1962, 62(2), 155–183.

14 (a) D. A. de Bie, A. Ostrowicz, G. Geurtsen and H. C. van der Plas,

Tetrahedron, 1988, 44(10), 2977–2983; (b)A.E.Frissen,A.T.M

Marcelis, G. Geurtsen, D. A. de Bie and H. C. van der Plas, Tetrahedron,

1989, 45(16), 5151–5162.

15 S. Takano, Y. Yano and W. Tagaki, Chem. Lett., 1981, 1177.

16 C. Zimmermann, F. W. Heinemann and A. Grohmann, Eur. J. Inorg.

Chem., 2001, (2), 547–555.

17 (a) A. Gerli, M. Sabat and L. G. Marzilli, J. Am. Chem. Soc., 1992,

114, 6711–6718; (b) S. Kiani, R. J. Staples and A. B. Packard, Acta

Crystallogr., Sect. C: Cryst. Struct. Commun., 2002, 58, m593–m595.

18 A. Thibon, J. England, M. Martinho, V. G. Young, J. R. Frisch, R.

Guillot, J. J. Girerd, E. M¨

unck, L. Que and F. Banse, Angew. Chem.,

2008, 120(37), 7172–7175, (Angew. Chem., Int. Ed., 2008, 47, 7064).

19 (a) M. Achmatowicz, L. S. Hegedus and S. David, J. Org. Chem., 2003,

68(20), 7661–7666; (b)A.L.Reiff,E.M.Garcia-Frutos,J.M.Gil,O.

P. Anderson and L. Hegedus, Inorg. Chem., 2005, 44(25), 9162–9174.

20 (a) M. Grenness and K. B. Oldham, Anal. Chem., 1972, 44(7), 1121–

1129; (b) A. M. Bond, K. B. Oldham and G. A. Snook, Anal. Chem.,

2000, 72(15), 3492–3496.

21 Oxford Diffraction, CrysAlis CCD and CrysAlisPro, Version

1.171.33.66, Abingdon, Oxford, UK, Oxford Diffraction Ltd, 2008.

22 M. C. Burla, R. Caliandro, M. Camalli, B. Carrozzini, G. L. Cascarano,

L. De Caro, C. Giacovazzo, G. Polidori and R. Spagna,, J. Appl.

Crystallogr., 2005, 38, 381–388.

23 G. M. Sheldrick, SHELXL-97, Program for reﬁnement of crystal

structures, University of G¨

ottingen, Germany, 1997.

24 L. J. Farrugia, J. Appl. Crystallogr., 1997, 30, 565.

25 L. Wang and H. Zipse, Liebigs Ann., 1996, (10), 1501–1509.

26 T. Matsumoto, D. Yoshida and H. Tomita, Agric. Biol. Chem., 1981,

45(9), 2031–2035.

27 (a) F. T. Oakes and N. J. Leonard, J. Org. Chem., 1985, 50, 4986–

4989; (b) J. Pakusch, H.-D. Beckhaus and C. R¨

uchardt, Chem. Ber.,

1991, 124, 1191–1198; (c) E. Diez-Barra, A. de la Hoz, A. Moreno

and P. Sanchez-Verdu, J. Chem. Soc., Perkin Trans. 1, 1991, 2589–

2592.

28 (a) G. E. Hilbert and T. B. Johnson, J. Am. Chem. Soc., 1930, 52(3),

1152–1157; (b) N. Whittaker and T. S. G. Jones, J. Chem. Soc., 1951,

1565–1570.

29 W. S. Matthews, J. E. Bares, J. E. Bartmess, F. G. Bordwell, F.

J. Cornforth, G. E. Drucker, Z. Margolin, R. J. McCallum, G. J.

McCollum and N. R. Vanier, J. Am. Chem. Soc., 1975, 97(24), 7006–

7014.

30 (a)T.BugandH.Mayr,J. Am. Chem. Soc., 2003, 125, 12980–12986;

(b) H. Mayr and A. R. Oﬁal, J. Phys. Org. Chem., 2008, 21, 584–

595.

31 T. Ohkuma, S. Hashiguchi and R. Noyori, J. Org. Chem., 1994, 59(1),

217–221.

32 (a) Y. Kobuke and H. Miyaji, J. Am. Chem. Soc., 1994, 116, 4111–4112;

(b) S. Morikawa, C. Ikeda, K. Ogawa and Y. Kobuke, Lett. Org. Chem.,

2004, 1, 6–11; (c) H. Nakagawa, K. Ogawa, A. Satake and Y. Kobuke,

Chem. Commun., 2006, 1560–1562.

33 (a) B. Cervera, R. Ruiz, F. Lloret, M. Julve, J. Cano, J. Faus, C. Bois and

J. Mrozinski, J. Chem. Soc., Dalton Trans., 1997, (3), 395–401; (b)R.

Ruiz, F. Lloret, M. Julve, J. Faus, M. C. Mu˜

noz and X. Solans, Inorg.

Chim. Acta, 1998, 268, 263–269.

34 D.-X. Li, D.-J. Xu and Y.-Z. Xu, Acta Crystallogr., Sect. E: Struct. Rep.

Online, 2003, 59, m1094–1095.

35 (a) A. Bondi, J. Phys. Chem., 1964, 68(3), 441–451; (b) S. S. Batsanov,

Inorg. Mater., 2001, 37(9), 871–885.

36 (a) J. M. Rubin-Preminger and U. Englert, Acta Crystallogr., Sect. E:

Struct. Rep. Online, 2006, 62, m2659–m2660; (b) J. M. Rubin-Preminger

and U. Englert, Acta Crystallogr., Sect. E: Struct. Rep. Online, 2006,

62, m2972–m2973.

37 (a) E. Frasson, R. Bardi and S. Bezzi, Acta Crystallogr., 1959, 12, 201–

205; (b) R. E. Rundle and C. V. Banks, J. Phys. Chem., 1963, 67(2),

508–509.

38 P. J. Steel, Acc. Chem. Res., 2005, 38, 243–250.

39 D. J. A. De Ridder, E. Zangrando and H.-B. B¨

urgi, J. Mol. Struct.,

1996, 374, 63–83.

40 S.-G. Roh, A. Proust, F. Robert and P. Gouzerh,, J. Cluster Sci., 1996,

7(4), 593–627.

41 G. A. Lawrance and P. G. Lye, Comments Inorg. Chem., 1994, 15(5–6),

339–368.

42 P. Du, K. Knowles and R. Eisenberg, J. Am. Chem. Soc., 2008, 130(38),

12576–12577.

43 P.-A. Jacques, V. Artero, J. P´

ecaut and M. Fontecave, Proc. Natl. Acad.

Sci. U. S. A., 2009, 106(49), 20627–20632.

Published on 05 January 2012. Downloaded by TU Berlin - Universitaetsbibl on 31/03/2016 08:28:14.

View Article Online