Pyrazolate-based cobalt(II)-containing metal-organic frameworks in heterogeneous catalytic oxidation reactions: Elucidating the role of entatic states for biomimetic oxidation processes

FULL PAPER

DOI: 10.1002/chem.201003173

Pyrazolate-Based Cobalt(II)-Containing Metal–Organic Frameworks in

Heterogeneous Catalytic Oxidation Reactions: Elucidating the Role of

Entatic States for Biomimetic Oxidation Processes

Markus Tonigold,^[a]Ying Lu,^[a]Andreas Mavrandonakis,^[b]Angela Puls,^[c]

Reiner Staudt,^[d]Jens Mçllmer,^[e]Joachim Sauer,^[b]and Dirk Volkmer*^[a]

Abstract: Crystal structures of two

metal–organic frameworks (MFU-1

and MFU-2) are presented, both of

which contain redox-active Co^IIcentres

coordinated by linear 1,4-bis[(3,5-dime-

thyl)pyrazol-4-yl] ligands. In contrast to

many MOFs reported previously, these

compounds show excellent stability

against hydrolytic decomposition. Cata-

lytic turnover is achieved in oxidation

reactions by employing tert-butyl hy-

droperoxide and the solid catalysts are

easily recovered from the reaction mix-

ture. Whereas heterogeneous catalysis

is unambiguously demonstrated for

MFU-1, MFU-2 shows catalytic activity

due to slow metal leaching, emphasis-

ing the need for a deeper understand-

ing of structure–reactivity relationships

in the future design of redox-active

metal–organic frameworks. Mechanistic

details for oxidation reactions employ-

ing tert-butyl hydroperoxide are stud-

ied by UV/Vis and IR spectroscopy

and XRPD measurements. The catalyt-

ic process accompanying changes of

redox states and structural changes

were investigated by means of cobalt

K-edge X-ray absorption spectroscopy.

To probe the putative binding modes

of molecular oxygen, the isosteric heats

of adsorption of O₂were determined

and compared with models from DFT

calculations. The stabilities of the

frameworks in an oxygen atmosphere

as a reactive gas were examined by

temperature-programmed

oxidation

(TPO). Solution impregnation of

MFU-1 with a co-catalyst (N-hydrox-

yphthalimide) led to NHPI@MFU-1,

which oxidised a range of organic sub-

strates under ambient conditions by

employing molecular oxygen from air.

The catalytic reaction involved a bio-

mimetic reaction cascade based on free

radicals. The concept of an entatic

state of the cobalt centres is proposed

and its relevance for sustained catalytic

activity is briefly discussed.

Keywords: cobalt · heterogeneous

catalysis · oxygen · thermodynam-

ics · metal–organic frameworks

Introduction

gas or liquid adsorption and separation, molecular recogni-

tion or catalysis.^[1,2]By combining bi- or multifunctional li-

gands (predominantly polycarboxylate ligands) and (transi-

tion) metal ions, moderately robust MOFs can be prepared,

of which 1,4-benzenedicarboxylate (bdc) represents a versa-

tile linker, for instance, leading to the porous frameworks

Porous metal–organic frameworks (MOFs) constitute a rap-

idly emerging class of multifunctional hybrid materials that

might be useful for diverse technical applications, such as

MOF-5 ([Zn₄O

AHCTUNGTRENNUNG

(bdc)₃]),^[3]and MIL-101 ([Cr₃O

ACHUTNGTRENNUG CAHTUNGTREN(NUGN OH,F,H₂O)₃-

[a] M. Tonigold, Dr. Y. Lu, Prof. Dr. D. Volkmer

Universitꢀt Augsburg, Institut fꢁr Physik

Lehrstuhl fꢁr Festkçrperchemie

of coordination polymers that exhibit permanent porosity

upon solvent removal.

Universitꢀtsstrasse 1, 86159 Augsburg (Germany)

Fax : (+49)821-598-5955

Developing MOFs into heterogeneous catalysts could

yield several principal advantages, such as enhanced catalyst

stability due to the spatial separation of single catalytic sites

in the framework.^[5]MOFs can possess high porosities in the

absence of any inaccessible bulk volume (dead volume),^[2]

and most evidently their pore size(s), and thus their sub-

strate shape and size selectivity, can be systematically tail-

ored by employing different organic linkers.^[6]MOF proto-

types, such as the recently described MFU-4^[7a]or magnesi-

um formates,^[2,7b–d]possessing narrow and structurally well-

defined pores, can be used to separate gas mixtures contain-

ing molecules with very similar shapes and kinetic diameters

(e.g., separation of H₂from O₂/N₂).^[2,7a,c]On the other hand,

the pores in MOF-5, for instance, are sufficiently large to

E-mail: dirk.volkmer@physik.uni-augsburg.de

[b] Dr. A. Mavrandonakis, Prof. Dr. J. Sauer

Institut fꢁr Chemie, Humboldt-Universitꢀt zu Berlin

Unter den Linden 6, 10099 Berlin (Germany)

[c] Dr. A. Puls

Rubotherm GmbH, Universitꢀtsstrasse 142

44799 Bochum (Germany)

[d] Prof. Dr. R. Staudt

Hochschule Offenburg, Badstrasse 24

77652 Offenburg (Germany)

[e] J. Mçllmer

Institut fꢁr nichtklassische Chemie e.V. (INC)

Universitꢀt Leipzig, Permoserstrasse 15, 04318 Leipzig (Germany)

Supporting information for this article is available on the WWW

under http://dx.doi.org/10.1002/chem.201003173.

Chem. Eur. J. 2011, 17, 8671 – 8695

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

8671

allow almost free diffusion of small molecules, in contrast to

most zeolites.^[8]In short, MOFs are interesting candidates

for catalytic applications, with the advantages of a heteroge-

neous reaction course (i.e., simple catalyst recovery), while

minimising the adverse effects of slow molecular diffusion,

limiting reactant transport in microporous materials.

aluminium MOFs were active catalysts for the selective oxi-

dation of xanthene to xanthone; however, the question as to

whether or not catalytic activity might be due to leaching of

metal ions from the nodes of the MOFs into solution was

not addressed.^[19]Corma et al. have recently described the

catalytic oxidation of tetralin with molecular oxygen,^[20]

using the Co^II-containing MOF ZIF-9^[21]as a catalyst. They

achieved 23% conversion (tetralone was the main product)

after 30 h. An induction period of about 10 h existed, in

which no tetralin conversion was observed. This catalyst

shows no metal leaching; however, the molecular dimen-

sions of tetralin are approx. 0.46ꢃ0.65 nm, whereas the pore

aperture in ZIF-9 is only about 0.30 nm,^[22]thus the intra-

pore diffusion of tetralin in ZIF-9 might be strongly hin-

dered, which might indicate that the catalytic oxidation is

primarily confined to the particle surface. However, it was

found by several groups that ZIF-8 could accommodate

larger molecules than it would be expected, based on the

pore size calculated from crystallographic data;^[23]a phe-

nomenon that was already known from zeolites.^[24]On the

other hand, it was demonstrated that only the surface of

ZIF-8 operated in the transesterification of large molecules.

To obtain hydrolytically stable cobalt(II)-based MOFs,

the soft Lewis acid cobalt(II) should be coordinated to soft

Lewis base atoms, such as N-heterocyclic aromatic ligands

(e.g., pyridine or pyrazole ligands). The expected increase in

thermodynamic stability is already gleaned by simply re-

garding the complete exchange of water against ammonia in

tetraaquacobalt(II) (pK=5.5).^[25]According to the well-

known Irving–Williams series,^[26]the thermodynamic stabili-

ty of complexes should increase, if late 3d transition-metal

ions (soft Lewis acids, such as Co^II, Ni^IIor Cu^II) are coordi-

nated to soft Lewis base atoms, such as N-heterocyclic aro-

matic ligands (e.g., pyridine or pyrazole ligands). The CSD

database revealed two structures based on 3,5-dimethylpyra-

zole (3,5-dmpz), which were suitable as SBUs for the con-

struction of MOFs (Figure 1). These prototypic compounds

Three conceptually different approaches have been used

to gain catalytically active MOFs: introducing catalytic

metal centres at the nodes of the constituting secondary

building units (SBUs), attaching catalytically active func-

tional groups or coordination units at the MOF struts (i.e.,

the organic linkers), and embedding nanosized metal clus-

ters within the pores of the MOF. Finely dispersed, catalyti-

cally active metal (e.g., Cu, Pd, Pt, Au) or metal oxides

inside the MOF cavities can be obtained by depositing vola-

tile organometallic precursor complexes^[9]in the open MOF

cavities from the gas phase or from solution,^[10]enabling cat-

alysis by MOF-encapsulated clusters. Pd@MIL-101, for in-

stance, shows similar activity in Heck reactions as commer-

cial catalysts^[11]and CO oxidation is feasible with Au@ZIF-

8^[12]or Pd@MIL-101.^[10a]However, sintering or agglomera-

tion of the metal nanoparticles occurs during catalysis in

some cases^[13]and it seems difficult to achieve a homogene-

ous distribution of nanoparticles filling the internal MOF

pores—as opposed to the likewise simple deposition of

nanoparticles on the external surfaces of the MOF crys-

tals.^[14]

The integration of functional groups or metal coordina-

tion units into the organic linkers that constitute the struts

of the framework is an attractive and straightforward ap-

proach to MOF-based heterogeneous catalysts. In several

studies, metalloporphyrins,^[15]Schiff base^[5]or binaphthyl

complexes^[16]have been used as ligands. The problems relat-

ed with this approach are most apparent for metalloporphy-

ACHTUNGTRENNUNGrin-based systems: first, porphyrinic MOFs featuring large

open pores are highly susceptible to collapsing upon solvent

removal; second, it is difficult to prevent saturation of the

“free” coordination sites of the metal porphyrin, which thus

become part of the framework itself; third, attempts to in-

corporate free-base porphyrins as struts (which would then

be available for post-synthetic metallation) are generally

frustrated by the tendency of the porphyrin ligand to scav-

enge and coordinate metal ions present in the initial MOF

synthesis.

To date, the use of catalytically active metal centres posi-

tioned at the nodes of the MOF constituting SBUs has been

predominantly explored for Lewis acid catalysis,^{[1d, 17]}where-

as examples of redox catalysis are rare. MOFs with copper

metal centres, for instance, can catalyse the oxidative cou-

pling of 2,6-dimethylphenol to form poly(1,4-phenylene

are the metal complexes [Co₄O

analogue of “basic zinc acetate”, [Zn₄O

typic building unit of MOF-5-type frameworks, and [Co^II-

ACHTUNGTRENNUNG

AHCTUNGTRENNUNG

(3,5-dmpz)]₁,^[28]an infinite one-dimensional chain of tetra-

hedrally coordinated Co^IIatoms. By replacing 3,5-dmpz with

the linear bridging ligand 1,4-bis[(3,5-dimethyl)pyrazol-4-yl]-

benzene (H₂-bdpb), which has two pyrazole groups,^[29]the

two MOFs MFU-1^[30,31]and MFU-2 are derived from [Co₄O-

ACHUTNGERNNUG CAHTUGNTREN(NGUN 3,5-dmpz)]₁, respectively, as outlined

(3,5-dmpz)₆] and [Co^II

in Figure 1. By adjusting the reaction conditions (see the Ex-

perimental Section), both phase-pure MFU-1 and MFU-2

can be obtained in good yields by treating the ligand H₂-

bdpb with a suitable Co^IIsalt under solvothermal conditions.

To follow up on our previous communication on the cata-

lytic activity of MFU-1 employing tert-butyl hydroperox-

ide,^[31]we present herein a detailed crystallographic struc-

ture analysis of both MFU-1 and MFU-2. Their catalytic ac-

tivities are compared with respect to their structural features

and special emphasis is placed on elucidation of structural

stability and metal leaching under catalytic reaction condi-

ꢀ

ether). The C O/C C oxidative coupling selectivities are

comparable to those obtained for other homo- and hetero-

ꢀ

geneous catalysts (up to 90% C O/C C coupling under op-

timised reaction conditions), and display good substrate-to-

catalyst ratios as well as short reaction times.^[18]By using hy-

drogen peroxide and MIL-101 as catalysts, selective sulfoxi-

dation of aryl sulfides becomes feasible.^[4]Iron, copper or

8672

www.chemeurj.org

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2011, 17, 8671 – 8695

Pyrazolate-Based Cobalt(II)-Containing Frameworks

FULL PAPER

the same porosity (as proven by

argon BET measurements),

albeit the crystals obtained

from microwave synthesis are

significantly smaller and more

uniform in size.^[33]The final

crystallographic data and de-

tails of the refinement are sum-

marised in Table 1.

Crystallography of MFU-1:

Crystallographic data sets for

MFU-1 were obtained from dif-

ferent crystal specimens suita-

ble for single-crystal X-ray dif-

fraction by employing both

Cu_Kaand Mo_Karadiation. To

remove occluded DMF solvent

molecules, the suspension of

crystals was heated at reflux

multiple times with chloroform

and the crystals were kept for

at least 12 h in high vacuum

Figure 1. a) Formal derivation of MOF-5 ([Zn₄O

G

(OAc)₆])^[32](bdc=

ACHTUNGTRENNUNG

1,4-benzenedicarboxylate). b) Similar construction of MFU-1 [Co₄OAHCTUNGTRENNUNG

ACHTUNGTRENNUNG

plex [Co^II₄O(3,5-dmpz)₆]^[27](H₂-bdpb=1,4-bis[(3,5-dimethyl)-pyrazol-4-yl]benzene). c) Derivation of MFU-2

from catenapyrazolato cobalt(II).^[28]The MOFs are generated from the corresponding SBUs by connecting the

points of extension (indicated by grey dots) with 1,4-phenylene units.

tions. The local coordination environment and the oxidation

state of the Co centres are investigated by means of X-ray

absorption spectroscopy at the Co K edge and changes oc-

curring upon catalysis are analysed, revealing details of the

presumed catalytic reaction. To elucidate the interaction of

both frameworks with molecular oxygen, the isosteric heats

of O₂adsorption are determined. Moreover, both com-

pounds are studied by temperature-programmed oxidation

(TPO). First results from DFT calculations are presented to

support our conclusions drawn from structural and spectro-

scopic studies, gleaning further details about the activation

of oxidants by pyrazolate-based MOFs containing cobalt(II).

As a preliminary summary, activation of molecular oxygen

by MFU-1 is feasible (in contrast to MFU-2), thus enabling

heterogeneous catalytic oxidation reactions under mild reac-

tion conditions. Table 7 at the end of this article summarises

the key data of the results presented in the following sec-

tions.

(p<10^ꢀ4mbar). The crystallographic data from six inde-

pendent measurements uniformly showed a large residual

Table 1. Crystallographic data for MFU-1 and MFU-2.

MFU-1

MFU-2

formula

C₄₈H₄₈Co₄N₁₂O [Co₄O-

C₁₆H₁₆CoN₄[Co-

A

ACTHGNURTENUN(NG C₁₆H₁₆N₄)]

M_r

1044.70

cubic

P43m (no. 215)

323.26

crystal system

space group

a [ꢄ]

b [ꢄ]

c [ꢄ]

tetragonal

P4₂/ncm (no. 138)

18.6070(8)

7.1635(5)

90

¯

15.7904(2)

a [8]

90

b [8]

90

g [8]

90

V [ꢄ³]

Z

3937.13(9)

1

2480.2(2)

4

F

(000)

536

668

0.866

1.54184

5.409

]

0.441

0.71073

0.429

m [mm^ꢀ1

]

crystal size [mm] 0.073ꢃ0.071ꢃ0.070

0.054ꢃ0.052ꢃ0.016

120(2)

T [K]

120(2)

3.65–29.34

ꢀ20/21

ꢀ21/18

ꢀ20/19

17696

1975

Results and Discussion

V range [8]

min/max h

min/max k

min/max l

collected reflns

unique reflns

reflns with

I>2s(I)

3.36–58.88

ꢀ20/11

ꢀ17/20

ꢀ7/5

Single crystals of MFU-1 and MFU-2 were grown slowly

under solvothermal conditions by treating the ligand H₂-

bdpb with cobalt chloride or cobalt nitrate, respectively, in

DMF at 1208C (see the Experimental Section). For a more

efficient bulk synthesis of phase-pure compounds, we em-

ployed a microwave system (focused microwave irradiation

with 300 W output power at a frequency of 2.46 GHz),

which reduced the reaction time required from several days

to a few minutes and significantly increased the yield. Based

on results from X-ray powder diffraction (see the Support-

ing Information), these microcrystals are structurally identi-

cal to those obtained from solvothermal synthesis and show

4583

922

444

1295

params/restraints 52/1

61/6

0.942

0.0856

GoF on F²

R₁(on F,

0.961

0.0529

I>2s(I))

wR₂(on F², all

data)

0.1390

0.2274

max/min D1

0.551/ꢀ0.332

1.086/ꢀ0.911

[eꢄ^ꢀ3

]

Chem. Eur. J. 2011, 17, 8671 – 8695

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

8673

D. Volkmer et al.

electron density in the void regions of the crystal lattice.

The residual electron density was quite diffuse owing to the

strong positional disorder of residual DMF molecules oc-

cluded in the lattice voids of MFU-1. Thus, all atom coordi-

nates gleaned from the first structure solution proved to be

unstable in subsequent refinement cycles and collapsed to

chemically unreasonable structures in all attempts.

PLATON/SQUEEZE^[34]was employed to correct the initial

X-ray reflection data set for those regions containing highly

disordered components. A potential solvent volume of

2528 ꢄ³(corresponding to approximately 20 DMF mole-

cules),^[35]containing 241 electrons was found. A further in-

terpretation of the electron density maxima found in the

void regions of MFU-1, which might indicate partial inter-

penetration, is given in the Supporting Information.

Structure solutions of all data sets invariably suggested

¯

space group P43m, although subsequent refinement cycles

clearly showed the presence of unresolved electron density

maxima in close proximity to the {Co₄O} units of the frame-

work. Initial attempts to assign this residual electron density

to m₃-bridging ligand species, such as O^2ꢀor OH^ꢀ, failed.

The refinement was neither stable (even if the occupancy

factors of this crystallographic site were freely refined), nor

could we find further support from spectroscopic studies in

favour of this structural model. Nevertheless, based on theo-

retical models of the MFU-1 crystal lattice, the unassigned

electron density and the observed disorder phenomena can

be explained in terms of 1) structural linkage isomerism,

2) rotational disorder of interconnecting 1,4-phenylene units

Figure 2. Crystal packing diagram of MFU-1. {CoON₃} coordination units

are represented as polyhedra. The phenylene moieties of the bridging

bdpb ligands occupy two alternative (symmetry equivalent) positions

with equal probability, one of which is displayed in the graphics by thin

black lines. Approximately 20% of all {Co₄OACTHNUTRGEN(UNG dmpz)₆} units are placed at

and 3) chiral distortion of the {Co₄OACHTNUGTRNEUNG(3,5-dmpz)₆} SBUs,

structurally inverted positions, for which the central oxygen atoms act as

centres of symmetry. (Only major disorder components are shown; hy-

drogen atoms are omitted for clarity.)

leading to a deviation of the real structure of MFU-1 from

the idealised model, represented by the highly symmetric so-

¯

lution in space group P43m, as shown in Figure 2.

In the chosen (achiral) space group, the phenylene moiet-

ies of the bridging bdpb ligand occupy two alternative posi-

tions with 50% probability each. The bdpb ligands and

{Co₄O} units are linked into a cubic network (Figure 2) of

low density (1_calcd=0.44 gcm^ꢀ3). The structure of MFU-1 is

similar to that of MOF-5, which has a CaB₆-type framework

topology. It encloses (pseudo)octahedrally shaped {Co₄O-

ACHTUNGTRENNUNG(3,5-dmpz)₆} SBUs reminiscent of the {Zn₄OCAHTUNGTREN(NUGN CO₂)₆} SBUs

Figure 3. An ORTEP representation and the numbering scheme of the

asymmetric crystallographic unit of MFU-1. Atoms of the anisotropically

refined main component (O1, Co1, N1, C1–C5) are represented by ther-

mal displacement ellipsoids connected with bold (filled) lines, represent-

ing bonds. The minor disorder component (Co1B, N1B, C1B, C2B; 20%

probability) is represented by open circles and broken lines. Additionally,

symmetry generated atoms (from both disorder components) are con-

nected by thin lines.

in MOF-5 and phenylene rings representing the edges of the

cubic six-connected CaB₆net. The framework has three-di-

mensional intersecting channels that encompass almost

spherical voids with a volume of about 2528 ꢄ³each. The

square windows of MFU-1 are apertures of 9.0 ꢄ (distance

between opposing phenylene units regarding the van der

Waals radii).^[22]A structurally related MOF network con-

structed from m₄-oxo-bridged tetrazinc units and 3,3’,5,5’-tet-

ramethyl-4,4’-bipyrazolate linkers has recently been report-

ed.^[36]

As previously mentioned, all structure solutions and re-

finements (even those performed in the space group P₁)

gave rise to a strong residual electron density, the maxima

of which had the same distance to the central oxygen atom

as the cobalt ions (Figure 3). This electron density can be

explained in terms of two different {Co₄O} tetrahedra that

are virtually superimposed by inversion: the oxygen atom

(O1) representing the centre of symmetry, thus leads to an

observed pseudo-cubic arrangement of electron density

maxima assigned to the positions of cobalt atoms (Co1 and

Co1B), as shown in Figure 3. The refinement reveals that

20% of the {Co₄O} tetrahedra in MFU-1 are inverted

through their central oxygen atom as a statistically distribut-

ed minor disorder component.

8674

www.chemeurj.org

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2011, 17, 8671 – 8695

Pyrazolate-Based Cobalt(II)-Containing Frameworks

FULL PAPER

As an illustration of the linkage isomerism, an excerpt of

a lattice plane of MFU-1 is shown in Figure 4a, in which the

foremost {Co₄OACHTUNGTRENNUNG(3,5-dmpz)₆} SBU is symbolised by an open

the experimental finding that 20% of the {Co₄OACTHNGUTERNUN(G 3,5-dmpz)₆}

units are statistically inverted might be as follows: There are

two highly symmetrical space groups in which the MFU-1

¯

octahedron. This unit can be inverted (by employing the

central oxygen atom as the symmetry centre), as shown in

Figure 4b, and the altered unit can be re-inserted into the

crystal lattice without introducing any kinds of steric repul-

sion or strain. Thus, at first approximation there should be

no energetic differences between the different lattice iso-

mers emerging from structural differences (neglecting en-

tropic factors, however).

structure might be represented, namely, P43m or F43m. In

¯

the space group setting of P43m, all SBUs are symmetry re-

lated by simple translation. This arrangement is possible

only if the pyrazole ring systems at both ends of the same

bridging ligand are orthogonal to each other (Figure 5, top).

All X-ray data sets recorded from MFU-1 single crystals

so far indicate a strong preference for a uniform orientation

of SBUs such that the (occasional) occurrence of inverted

units (the refined ratio between both components is 0.8:0.2)

leads to a relatively strong residual electron density that is

ascribed to the cobalt centres located at the alternative posi-

tions of the minor component. A possible explanation for

Figure 5. Different arrangements of adjacent {Co₄OACHTNUTRGNEU(GN 3,5-dmpz)₆} SBUs

¯

forced by the space group symmetry settings in the space group P43m

¯

(top) and F43m (bottom). {CoN₃O} coordination units are drawn as poly-

hedra.

¯

In the space group F43m (the one most often used to repre-

sent MOF-5), on the other hand, adjacent SBUs cannot be

brought to congruence by a simple translation. Adjacent

SBUs connected to the same linker have to be inverted

before translation and thus 50% of the {Co₄OACTHUNTRGNEU(GN 3,5-dmpz)₆}

¯

nodes in the space group setting F43m are inverted, necessi-

tating, however, a parallel arrangement of pyrazole rings at

both ends of the bridging ligand, as shown in Figure 5

(bottom).

By assuming that the energetic differences between both

conformers of the bridging ligand are negligible, a statistical-

ly equally weighted distribution should be observed. Thus, a

1:1 occurrence of both conformers would be equivalent to

the fact that inversion should occur for 25% of the {Co₄O-

AHCTUNGTRE(GNNUN 3,5-dmpz)₆} units; this comes close to the 20% of inversion

disorder experimentally determined for the structure solu-

¯

tion in space group P43m. Nevertheless, it should be kept in

mind that the shape and packing density of occluded solvent

molecules may have a strong influence on this conforma-

tional disorder phenomenon.

¯

A refinement in the space group F43m (by using a unit

cell with doubled cell constants) revealed a statistical disor-

der of the inverted {Co₄OACTHNUTRGNE(UNG dmpz)₆} node as well. By using

this space group, the main disorder compound also displays

adjacent {Co₄O} units that are symmetry related by simple

translation. This finding further supports a statistical disor-

der of the {Co₄O} units, showing that no higher crystallo-

graphic order is present. However, not all atoms of the

minor disorder compound can be found in the space group

Figure 4. Structural lattice isomerism of {Co₄OACTHNUTRGNEUGN(3,5-dmpz)₆} SBUs. a) Ad-

jacent SBUs are symmetry related by simple translations. b) Inversion

(by employing the central oxygen atom as the symmetry centre) of a

single {Co₄O

on the electron density distribution in MFU-1, the statistical linkage iso-

merism is present at approximately 20% of the {Co₄O(3,5-dmpz)₆} SBUs.

ACHTUNGTRENNUNG(3,5-dmpz)₆} SBU symbolised by an open octahedron. Based

¯

G

F43m.

Chem. Eur. J. 2011, 17, 8671 – 8695

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

8675

D. Volkmer et al.

Apart from the structural disorder brought into the MFU-

1 crystal lattice by linkage isomerism, the finally converged

structure refinement showed some additional unusual crys-

tallographic properties, such as physically doubtful aniso-

tropic displacement parameters for the methyl substituents

of the pyrazolate moieties (atom C1, Figure 3). A closer in-

spection revealed that this artefact was presumably due to

gate the conformation of the pyrazole rings around the

{Co₄O} core. Two different initial conformations were taken

into account, corresponding to a local T_dand T symmetry

around the SBU of MFU-1. However, when building a peri-

odic model of the crystal structure, the symmetry of the cell

has to be reduced to T because the phenylene moieties can

occupy only one of the two symmetry equivalent positions

(see Figure 2).

the fact that the {Co₄OACHTNUGTRNEUNG(3,5-dmpz)₆} SBUs were chirally dis-

torted (T rather than T_dpoint group symmetry) with both

possible enantiomers equally distributed in the crystal lattice

(Figure 6). The electron density thus represents an equally

The DFT +D results show that the conformation of the

pyrazole rings around the {Co₄O} is flexible. Optimization

of the two initial conformations results always in structures

with a chiral distortion. The energy minimum for the inter-

planar f angle (as defined in Figure 6) found for the infinite

network is between 1.9 and 4.68 (Table 2), corresponding to

local T symmetry, but since the energy changes connected

with these angle changes (less than 0.4 kJmol^ꢀ1) are within

the numerical accuracy limits of the calculations, we cannot

rule out that the minimum is at 08. This suggests that, even

at a temperature around 100 K, MFU-1 can adopt various

possible conformations. The flatness of the potential ex-

plains the observed elongation of the thermal ellipsoid in

the single-crystal X-ray structure analysis. Thus, the long-

range crystallographic order of MFU-1 seems to be appro-

priately represented in the setting of the achiral space group

Figure 6. a) Ball-and-stick model of the complete and fully T_dsymmetri-

cal {Co₄OACHTUNGTRENNUNG(pz₆)} SBU of MFU-1. A chiral distortion of this unit is charac-

terised by the interplanar angle f between the two least-squares planes

through the atom centres O(1)-Co(1)-Co(1)’ (P¹) and N(1)-C(1)-C(3)-

C(2)-C(3)’-C(1)’-(N(1)’ (P²). Simplified representations of the SBU. In

the model showing T_dsymmetry (b), the interplanar angle f is 08, where-

as the T-symmetric chiral SBU (c) is characterised by a finite angular f

value.

¯

P43m.

Crystallography of MFU-2: The crystallographic structure

analysis of MFU-2 is based on solvent-free crystals. MFU-2

crystallises in the tetragonal crystal system in space group

P4₂/ncm (no. 138). The asymmetric unit of the unit cell is

displayed in Figure 7. The phenylene moieties of the bridg-

ing bdpb ligands occupy two alternative positions with 87

(C5) and 13% (C5B) probability.

weighted superposition of two different enantiomers that

show significant structural deviations from T_dpoint group

symmetry. The chiral distortion of the {Co₄OACTHUNTRGNEU(GN 3,5-dmpz)₆}

SBUs of MFU-1 can be rationalised in terms of the steric re-

pulsion between methyl substituents of the 3,5-dmpz ligands.

As Figure 6c demonstrates, the distortion should lead to de-

viation of atomic positions of the coordinated pyrazolate

moieties from idealised T_dsymmetry (which itself is im-

¯

posed by the symmetry settings of space group P43m). This

chiral distortion is apparent in the crystal structure of

[Co₄OACHTUNGTRENNUNG

(3,5-dmpz)₆],^[27]in which the idealised T_dsymmetry of

each individual and discrete coordination unit is reduced to

chiral T point group symmetry (Table 2).

DFT calculations augmented by a semi-empirical disper-

sion term (denoted as DFT+D) were performed to investi-

Figure 7. ORTEP representation and numbering scheme of the asymmet-

ric unit present in crystalline MFU-2. The main component is drawn with

solid lines as bonds and front ellipses for the atoms, whereas the minor

disorder component (C5B, 13% probability) is drawn with bold broken

lines as bonds and open ellipses as atoms. Symmetry-related atoms are

connected by thin lines.

Table 2. Lattice parameters a=b=c [ꢀ], selected bond lengths [ꢄ] and

angles [8] in MFU-1 and [Co₄O

MFU-1 (exptl) MFU-1 (DFT+D)^[a]

(3,5-dmpz)₆].^[27]

ACHTUNGTRENNUNG

(3,5-dmpz)₆]^[27]

ACHUTNGREN[NUG Co₄OACHTUNGTRENNUGN

a=b=c

15.790

1.95

1.98

15.804

1.96

1.97

109.5

97.6

118.3

1.9–4.6

ꢀ

Co O

1.94–1.95

1.99–2.02

108.1–111.8 (Ø 109.5)

96.6–100.2 (Ø 98.3)

114.1–122.1 (Ø 118.0)

0.1–6.4 (Ø 3.9)

The framework contains quadratic tunnels running

through the crystal, consisting of 1D chains of cobalt(II)

ions bridged by dianionic bdpb ligands that extend along the

c axis of the crystal lattice. A similar structure constructed

from 1D chains of cobalt(II) ions bridged by linear tetrazo-

late based linkers has been recently reported.^[37]Each Co^II

centre is coordinated by four nitrogen donor atoms stem-

ꢀ

Co N

Co-O-Co 109.5

N-Co-O

N-Co-N

f

98.2

118.0

0.0

[a] DFT+D=DFT calculations augmented by a semi-empirical disper-

sion term.

8676

www.chemeurj.org

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2011, 17, 8671 – 8695

Pyrazolate-Based Cobalt(II)-Containing Frameworks

FULL PAPER

ming from different ligands. The local {CoN₄} coordination

tern of the coordinating pyrazolate ligands leads to marked

structural differences. Looking down the chain of cobalt(II)

ions in MFU-3, the set of planes running through adjacent

pyrazole ring systems are strictly orthogonal to each other.

In MFU-2, on the other hand, the least-squares planes

through opposite pyrazole rings of the same coordination

unit display a torsional distortion as a result of the steric in-

teractions between adjacent methyl substituents, thus lead-

ing to an interplanar angle of

ꢀ

site adopts D_2dpoint group symmetry, with all Co N bond

distances being equal (1.992(5) ꢄ), and two out of six N-Co-

N’ angles being slightly narrowed (106.6(3)8, as opposed to

four widened angles (110.9(1)8). Adjacent cobalt(II) centres

are bridged by bdpb ligands to form 6-membered {Co₂N₄}

rings with a Co···Co distance of 3.5818(3) ꢄ resulting in a

PtS-type framework topology.^[38]Figure 8 shows the quadrat-

298 (Figure 9, as opposed to 08

in MFU-3). As a result of steri-

cally less-encumbered coordina-

tion geometry, MFU-3 shows a

pronounced framework flexibil-

ity and thus a breathing effect,

as revealed by a formal type IV

gas sorption isotherm, whereas

MFU-2 is a rigid network dis-

playing a type I gas sorption

isotherm.

Thermal stability: Investiga-

tions on the thermal stabilities

of MFU-1 and MFU-2 were

performed by combining ther-

mogravimetric analysis (TGA;

samples exposed to flowing ni-

trogen atmosphere) and varia-

ble-temperature X-ray powder

diffraction (VTXRPD) (sam-

ples kept in static air). Both un-

treated (as-synthesised) samples

and samples suspended in

CH₂Cl₂for 24 h, followed by

drying in vacuum, were exam-

ined by TGA. Additionally,

temperature-programmed oxi-

dation (TPO) was carried out

Figure 8. Crystal-packing diagram of MFU-2. {CoN₄} coordination units are represented as polyhedra. Hydro-

gen atoms are omitted for clarity. Only the major rotational disorder component out of two alternative posi-

tions of the benzene moieties is shown.

by employing O₂gas. For the

as-synthesised samples, the

TGA of MFU-1 showed a loss

of occluded solvent molecules

ic channels of 18.61 ꢄ (lengths of the diagonal) in MFU-2

running along the c axis. The square channels in MFU-2

have apertures of 6.4 ꢄ (distance between opposed phenyl

rings regarding the van der Waals radii of 1.7 ꢄ for the

carbon atoms).^[22]

in the temperature range 75–2258C to yield the solvent-free

crystal form (weight loss: 11.9%; expected 12.3% for com-

plete loss of DMF from MFU-1·2DMF). Whereas the for-

PLATON/SQUEEZE^[34]was used to further analyse the

void region of MFU-2 (but no correction of the scattering

intensities was performed). A potential solvent volume of

1164 ꢄ³was found (corresponding to approximately 8 DMF

molecules).^[35]

The structure of MFU-2 is related to that of MFU-3,^[39]

which possesses the same structural motif of cross-linked in-

finite 1D chains of cobalt(II) ions, the bridging ligands, how-

ever, are composed of simple pyrazolate rather than 3,5-di-

methylpyrazolate moieties. The different substitution pat-

Figure 9. Distortion of the pyrazolate units, characterised by the interpla-

nar angle f between the two least-squares planes through the cobalt

atoms of two adjacent Co chains (P¹) and the pyrazolate rings (P²).

Chem. Eur. J. 2011, 17, 8671 – 8695

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

8677

D. Volkmer et al.

mulation MFU-1·2DMF is in agreement with the observed

weight loss and the elemental analyses, the solvent-accessi-

ble voids in the crystal lattice could, in principle, take up to

as many as 20 DMF molecules per formula unit. Whereas a

slight reduction of solvent molecules in the voids of MFU-1

might be due to filtering, such an enormous difference be-

tween observed and expected weight loss might hint that

partial interpenetration occurs. The BET surface area and

pore volume in gas sorption experiments are lower than

those calculated from the crystallographic structure as well,

further supporting a partial interpenetration of the frame-

work.

No further weight loss occurred below 3408C; ligand mol-

ecules from the framework were degraded in the tempera-

ture range 340–11008C (weight loss: 44.1%; expected

66.5% for complete loss of bdpb). Upon suspending MFU-1

in CH₂Cl₂for 24 h, followed by drying in vacuum, the ini-

tially occluded DMF molecules were removed completely,

which was indicated by a residual weight loss of 2.9% in the

temperature range 30–1958C, with no further weight loss oc-

curring below 3408C. Accordingly, the VTXRPD patterns

indicate that the MFU-1 framework was stable up to 2708C

in static air (see the Supporting Information). TGA of

MFU-2 indicated a loss of solvent molecules up to 2608C

(weight loss: 21.8%; expected 18.5% for complete loss of

DMF from MFU-2·DMF), and ligand molecules from the

framework were degraded in the temperature range 3408C-

4708C (weight loss: 54.2%; expected 66.6%). Whereas the

formulation MFU-2·DMF is consistent with the composition

gleaned from elemental analysis, the void volume of MFU-2

generally would be able to accommodate two DMF mole-

cules per unit cell, indicating a partial loss of DMF mole-

cules during the workup procedure. The VTXRPD patterns

indicated that the framework of MFU-2 was stable up to

3008C (see the Supporting Information). Furthermore, these

results showed that both MOF frameworks were stable

upon complete removal of solvent molecules.

treated at 1408C in oxygen showed changes in the IR spec-

trum (additional bands around 1600–1700 cm^ꢀ1, indicating

the presence of carboxylate and/or amide groups), and only

an amorphous product in the XRD powder patterns (see

the Supporting Information). No spectral or structural

changes were discernible for MFU-2 in TPO experiments

upon heating the sample up to 2308C.

Oxygen isosteric heat of adsorption: To investigate the inter-

action of both framework compounds with molecular

oxygen, the isosteric heat of adsorption for different oxygen

loadings on MFU-1 and MFU-2 was examined, employing a

magnetic suspension balance. Prior to sorption experiments

all samples were pre-treated at 2008C for 2 h to remove re-

sidual solvent molecules (observed weight loss: 2%).

For a correct calculation of the isosteric heat of adsorp-

tion from the measured adsorption isotherm, a minimum of

at least three isotherms are necessary.^[41]Therefore, oxygen

adsorption isotherms for MFU-1 and MFU-2 were measured

gravimetrically at 298, 323 and 343 K and for pressures up

to 4 MPa. The executed loadings ranging from 0.01–

1 mmolg^ꢀ1are equivalent to 0.04–4 molecules of oxygen per

Co atom. All isotherms are displayed in Figure 10 for MFU-

1 and Figure 11 for MFU-2.

To probe thermal stabilities and redox behaviour of both

frameworks in reactive gas atmospheres, TPO experiments

were performed (Supporting Information). In all experi-

ments, fresh samples of MFU-1 and MFU-2 were heated at

reflux with CH₂Cl₂prior to drying in high vacuum. The sam-

ples were then kept at 2008C for 2 h to remove residual sol-

vent molecules (observed weight loss: 2%). By using TPO

in diluted oxygen (1000 ppm of oxygen in helium), MFU-1

showed reaction with oxygen starting at 1008C and a release

of CO₂and H₂O (indicative of decomposition) at 1408C and

above. In contrast to MFU-1, MFU-2 showed no sign of re-

action with oxygen or decomposition below 2308C, indicat-

ing a higher stability (and thus a lower reactivity) towards

molecular oxygen. Moreover, MFU-1 showed a colour

change from blue to green in an oxygen atmosphere after

reaching a temperature of approximately 1408C, which was

due to an additional absorption band occurring at 410 nm.

This colour change might be tentatively ascribed to the oxi-

dation of Co^IIto Co^IIIions under retention of the tetrahe-

dral coordination environment.^[40]Furthermore, the sample

&

*

)

Figure 10. Oxygen adsorption isotherms of MFU-1 at 298 ( ), 323 (

~

and 343 K ( ) (lines are from fitted virial equation).

Beginning at 1208C (393 K), MFU-1 in oxygen shows a

very slow but steady decrease of mass without reaching

equilibrium even after 72 h. X-ray powder diffraction indi-

cates that CoO and Co₃O₄are obtained as final products.

Thus, temperatures higher than 343 K were avoided. We at-

tribute this steady weight loss to the activation of molecular

oxygen at higher temperatures, which subsequently leads to

oxidation of the organic part of the framework.

The isosteric heat of adsorption, q_st, was determined by

fitting a virial-type equation to the adsorption isotherms.

The ln(p) values for a given amount adsorbed (n) were cal-

culated from the linear regressions determined from the

virial equation analysis by using Equation (1):^[42]

8678

www.chemeurj.org

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2011, 17, 8671 – 8695

Pyrazolate-Based Cobalt(II)-Containing Frameworks

FULL PAPER

&

*

)

Figure 13. Isosteric heat of adsorption of O₂on MFU-2 (8.5(9) kJmol^ꢀ1

Figure 11. Oxygen adsorption isotherms of MFU-2 at 298 ( ), 323 (

~

s

and 343 K ( ) (lines are from fitted virial equation).

!

*

^

at n !0) at 298–323 ( ), 298–343 ( ) and 323–343 K ( ). The average

*

values are also shown ( ).

lnðp=nÞ ¼ A₀þ A₁n þ A₂n²þ A₃n³þ . . .

ð1Þ

0.5 MPa, the amount of oxygen adsorbed is lower for MFU-

1 when directly compared with MFU-2.

For an analysis of the adsorbate–adsorbent interaction,

the K_Hvalue at low surface coverage is an ideal parameter.

in which p is the pressure, n is the amount of adsorbed

oxygen and A₀, A₁and so forth are virial coefficients. A₀is

related to adsorbate–adsorbent interactions and at low sur-

face coverage the Henryꢅs Law constant (K_H) is equal to 1/

exp(A₀).^[42]The q_stvalues were subsequently determined by

using a modified version of the Clausius–Clapeyron equa-

tion [Eq. (2)]:

Table 3. Parameters of the virial equations and K_Hfor MFU-1 and

MFU-2.

T [K] A₃

A₂

A₁

A₀

K_H[mmolMPa^ꢀ1g^ꢀ1

]

MFU-1

ꢀ

ꢁ

ꢀ

ꢁ

298

323

343

MFU-2

298

0.3207

0.5004

0.3820

ꢀ0.8145 1.2916 ꢀ0.3561 1.4278

ꢀ1.1645 1.4714 ꢀ0.0334 1.0340

ꢀq_st

@lnp

@T

¼

ð2Þ

RT²

ꢀ0.9224 1.3236

0.2279 0.7962

N

ꢀ0.0018 0.0261

0.0783

0.0200 0.9802

0.2603 0.7708

0.4678 0.6264

The q_stvalues for MFU-1 and MFU-2 are shown in Fig-

323

343

0.0282

0.0133

ꢀ0.0920 0.2227

ꢀ0.0299 0.1933

ures 12 and 13, respectively. From the gravimetric oxygen

adsorption isotherms, we thus derive an isosteric heat of ad-

sorption of (ꢀ11.1ꢁ0.9) kJmol^ꢀ1for MFU-1 and (ꢀ8.5ꢁ

0.9) kJmol^ꢀ1for MFU-2, throughout the entire range of

coverage explored. On the other hand, at pressures above

Table 3 contains all fitting parameters of the virial equation,

including K_H. At each temperature a higher K_Hwas ob-

served for MFU-1 than MFU-2, indicating a stronger bind-

ing of oxygen with the binding sites of MFU-1, correspond-

ing with the higher isosteric heat of adsorption.

Whereas MFU-1 (DH_iso=ꢀ11.1 kJmol^ꢀ1) showed a slight-

ly increased isosteric heat of adsorption than MFU-2

(DH_iso=ꢀ8.5 kJmol^ꢀ1), somewhat larger values of DH_iso

were found for Co-exchanged zeolites and a Co-based MOF

(ꢀ15.1 up to ꢀ18.5 kJmol^ꢀ1).^[43]The range of observed

values clearly indicated the presence of van der Waals com-

plexes and their slight variations may have indicated differ-

ent degrees of shielding of the Co^IIbinding sites by the or-

ganic linkers.

Despite the implied formation of a van der Waals com-

plex at the indicated temperatures, MFU-1 showed a meas-

urable reaction with oxygen above 1208C in TPO and

during the measurement of the gravimetric oxygen adsorp-

tion isotherms. In contrast, MFU-2, according to TPO meas-

urements, seems to be inert up to a temperature of at least

2308C (see the Supporting Information). A possible explan-

Figure 12. Isosteric heat of adsorption of O₂on MFU-1 (11.1(9) kJmol^ꢀ1

s

!

*

^

at n !0) at 298–323 ( ), 298–343 ( ) and 323–343 K ( ). The average

&

values are also shown ( ).

Chem. Eur. J. 2011, 17, 8671 – 8695

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

8679

D. Volkmer et al.

ation for these different reactivities towards molecular

oxygen is that the cobalt(II) ions in both frameworks are

sterically shielded by the organic linkers and thus poorly ac-

cessible to coordinating molecules. However, whereas the

cobalt centres in MFU-2 are completely shielded, two po-

tentially accessible coordination sites in MFU-1 should still

exist, as gleaned from space-filling models (Figure 14). The

same behaviour was observed when the sample was kept in

pure tert-butyl isonitrile (see the Supporting Information).

These experimental observations are in sharp contrast to

other Co^II-based MOFs with more accessible metal sites,^[43]

which leads us to conclude that coordinative binding of li-

gands in MFU-1 becomes largely suppressed owing to steric

shielding of the cobalt centres and symmetry constraints im-

posed by the tetranuclear coordination units.

Figure 15 shows the DFT+D optimised structures of a

van der Waals and an h²-superoxo complex of O₂with

MFU-1. For the van der Waals complex, the predicted heats

of adsorption^[46]at 298 K are ꢀ7.7 (PBE+D) and

Figure 14. Space-filling models of characteristic coordination units in

MFU-1 (left) and MFU-2 (right), illustrating the steric shielding of the

metal ions by the ligand.

a site of MFU-1 (Figure 14) may be considered to be the

“natural” coordination site of the cobalt(II) centres because

binding of a monodentate ligand would lead to a change of

coordination state from tetrahedral to trigonal bipyramidal,

which, according to a ligand-field theoretical approximation,

should lead to an energy gain.^[44]

However, as can be seen from Figure 14, the a sites in

MFU-1 are completely sterically shielded by the close-

packed methyl substituents and a widening of the cobalt co-

ordination shell would thus require a huge distortion of the

{Co₄OACHTUNGTRENNUNG(3,5-dmpz)₆} SBU, which might still be feasible for a

molecular complex, but owing to symmetry restraints will be

almost impossible if the same coordination unit is embedded

in a 3D crystalline framework. Apart from a sites, each

SBU in MFU-1 displays four easily accessible b sites, which

might serve the purpose of ligand coordination. However, if

binding of a ligand occurred in a symmetrical fashion, that

is, if the ligand coordinated to three cobalt(II) centres in m₃-

bridging mode, the individual coordination environments of

the cobalt centres would become largely distorted. This

would constitute an energetically unfavourable situation,

that is, a strained or “entatic state”.^[45]As an alternative to

this energetically unfavourable situation, a rapid outer-

sphere electron transfer might take place from the {Co₄O}

cluster core towards an oxidising substrate, alleviating the

problem of direct coordination. To examine the putative

free coordination sites in MFU-1, the compound was treated

with test molecules that were commonly considered to be

“strong” ligands in cobalt(II) coordination chemistry. In es-

sence, the UV/Vis spectrum of MFU-1 showed no change if

the compound was exposed to CO atmosphere at room tem-

perature; similarly, treating a MFU-1 suspension with a so-

lution of cyanide ions in methanol (using tetraethylammoni-

um cyanide) gave no indication of spectral changes and the

Figure 15. Structure of the van der Waals (top) and superoxo complexes

(bottom) of O₂with a metal SBU obtained by DFT (PBE+D) for the

periodic MFU-1 structure.

8680

www.chemeurj.org

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2011, 17, 8671 – 8695

Pyrazolate-Based Cobalt(II)-Containing Frameworks

FULL PAPER

ꢀ8.9 kJmol^ꢀ1(B3LYP+D), close to the experimental isoste-

ric heat of adsorption ((11.1ꢁ0.9) kJmol^ꢀ1). Whereas the

predicted stabilities of the van der Waals complex, employ-

ing different functionals, are in quantitative agreement, they

are controversial for the h²-superoxo–Co^IIIcomplex. The

PBE+D functional predicts it to be the most stable com-

those of microporous zeolites and aluminophosphates, but

significantly lower than theoretically predicted values, or the

corresponding experimental values of MOF-5 and other

structurally similar MOF compounds (600–3400 m²g^ꢀ1).^[50]

The huge discrepancy between experimental and theoretical

values of surface area and pore volume for MFU-1 might be

attributed to the presence of partially filled sections within

the real crystal structure of MFU-1. It could be speculated

that a partial filling of voids with a structural identical

framework (i.e., framework interpenetration) occurred.

However, clear-cut evidence from crystallographic studies

was hampered by the many structural factors by which the

real structure of MFU-1 differed from its idealised model,

as discussed in the previous section. In addition, recent stud-

ies on the related framework MOF-5 revealed a minor

phase consisting of doubly interpenetrated MOF-5 networks.

In this case, the presence of lattice interpenetration changed

the symmetry from cubic to trigonal and explained the peak

splitting observed in the powder XRD patterns, as well as

the reduced surface areas and pore volumes.^[51]

plex with DHACHTUNGTRENNUNG

B3LYP+D estimate is +64.9 kJmol^ꢀ1. Presently there are

no means to decide which functional is right in its predic-

tion; this is frequently the case with multinuclear transition-

metal compounds.^[47]The B3LYP +D result would be in

agreement with the experimental observation that below

1208C only a van der Waals complex is observed, whereas

the PBE+D result implies that formation of the h²-super-

oxo–Co^IIIcomplex is prevented by an energy barrier. A de-

tailed outline of the DFT calculations will be given else-

where.^[46]

BET/gas sorption: Both MOF compounds exhibit perma-

nent porosity, which has been confirmed by argon gas sorp-

tion. Prior to measurements, the samples were heated at

reflux several times in CH₂Cl₂to exchange the DMF mole-

cules with the more volatile solvent. The samples were sub-

sequently heated at 1508C for 30 h in high vacuum. The

sorption isotherms of MFU-1 and MFU-2 obtained with

argon gas reveal type I sorption behaviour for both substan-

ces, which is characteristic of microporous solids (see the

Supporting Information). The adsorption data was analysed

in the appropriate pressure range^[48]to give BET surface

areas of 1525(51) m²g^ꢀ1[49]for MFU-1 obtained by solvother-

mal synthesis and 1485(50) m²g^ꢀ1for MFU-1 from micro-

wave synthesis (see the Supporting Information). By using a

sample from microwave synthesis, the BET surface was re-

duced to 1018(33) m²g^ꢀ1 after catalysis, employing tert-butyl

hydroperoxide as described below. For MFU-2, BET surface

areas of 1474(50) m²g^ꢀ1for samples from solvothermal syn-

thesis and 1461(50) m²g^ꢀ1for samples from microwave syn-

thesis were obtained. The BET surface area of MFU-2 ex-

amined in sorption cycles after catalysis became negligible

(ca. 3 m²g^ꢀ1). Theoretical values of the specific surface area

derived from (idealised) crystal lattice models were obtained

from a Monte Carlo integration technique, in which a probe

molecule was “rolled” over the internal crystal surface.^[48]

Values obtained for a probe 3.686 ꢄ in diameter were

4117 m²g^ꢀ1for MFU-1 and 1447 m²g^ꢀ1for MFU-2; the

latter corresponding well with the average surface areas de-

termined for MFU-2 (1467 m²g^ꢀ1). The pore volumes were

evaluated by using the Dubinin–Radushkevich (DR) equa-

tion and compared to the theoretical values obtained from

the crystallographic data (by using PLATON/SQUEEZE).

The obtained pore volumes were 0.57 mLg^ꢀ1for MFU-1 ob-

tained by solvothermal synthesis and 0.56 mLg^ꢀ1for MFU-1

from microwave synthesis (expected: 1.49 mLg^ꢀ1), and

0.55 mLg^ꢀ1for MFU-2 obtained by solvothermal synthesis

and 0.53 mLg^ꢀ1for MFU-2 from microwave synthesis (ex-

pected: 0.55 mLg^ꢀ1). The values of micropore surface areas

and volumes for MFU-1 are still high, when compared to

UV/Vis spectroscopy: UV/Vis spectra of both compounds

are in complete agreement with the coordination environ-

ment revealed by X-ray single-crystal structure analysis (see

the Supporting Information): each Co^IIcentre is placed in a

distorted tetrahedral coordination environment consisting of

four aromatic N donors stemming from pyrazolate ligands

in the case of MFU-2, and three N donors and a single dia-

nionic oxo ligand in the case of MFU-1. The UV/Vis diffuse

reflectance spectra (see the Supporting Information) of both

compounds showed several similar absorption bands in the

UV region, which corresponded to intra

li

E

p!p* transitions (see the Supporting Information). In the

visible region, MFU-1 had a broad absorption band centred

at 610 nm (16393 cm^ꢀ1), which could be assigned to the

4

spin-allowed d–d transition ⁴A₂(F)! T₂(P) of tetrahedral

Co^IIions.^[52]Additional spectroscopic features included a

weak shoulder at 550 nm (18182 cm^ꢀ1), and an absorption

band in the near-infrared (NIR) region at 1000 nm

4

(10000 cm^ꢀ1) assigned to the ⁴A₂(F)! T₂(F) transition. A

band present at 270 nm (37040 cm^ꢀ1) might be ascribed to a

Co²⁺–O^2ꢀcharge transfer.^{[40, 53]}MFU-2 possessed broad ab-

sorption bands centred at 596 nm (16779 cm^ꢀ1) due to the

4

⁴A₂(F)! T₂(P) of tetrahedral Co^IIions and at >1100 nm (>

4

9100 cm^ꢀ1) for the A₂(F)! T₂(F) transition. The observed

spectral features for both MFU-1 and MFU-2 correspond

well with literature values of tetrahedral Co^IIcomplexes.^[52]

After catalysis employing tert-butyl hydroperoxide, an ad-

ditional, intensive band appeared at 410 nm (24390 cm^ꢀ1) in

the UV/Vis spectrum of MFU-1. This single absorption

band was characteristic of tetrahedral Co^IIIcomplexes,^[40]

and had been previously attributed to a ligand-to-metal

charge-transfer transition^[53a]from an oxygen atom to Co^III.

A tetrahedral coordination environment is rarely observed

for cobalt

of dodecatungstate as tetrahedral ligand^[53a]or for a cobalt-

AHCTUNGTREG(NNUN III) imido complex with a scorpionate (=hydrotris(pyrazol-

ACHTUNGTREN(NUNG III) atoms and is found, for instance, in the case

Chem. Eur. J. 2011, 17, 8671 – 8695

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

8681

D. Volkmer et al.

1-yl)borato) ligand.^[54]On the contrary, the more common

octahedral Co^IIIcomplexes show two bands with nearly

1

identical intensity at 300–450 nm for the A_1g! T_2gtransi-

1

tion and at 450–850 nm for the A_1g! T_1gtransition.^[52]In

contrast, the electronic spectrum of MFU-2 examined after

catalysis remained virtually unchanged under similar condi-

tions.

Catalytic activity of MFU-2 employing tert-butyl hydroper-

oxide: The stability of a catalyst against hydrolytic decom-

position is a necessary pre-condition for many oxidation

processes, since water is often obtained as a side product.

The hydrolytic stability of MFU-1 was already reported in a

previous communication.^[31]The stability of MFU-2 against

hydrolysis was investigated by several experiments. First,

crystals of MFU-2 were suspended in ethanol (employed as

a test protic solvent) for 96 h, after which time no discerni-

ble changes occurred in the UV/Vis spectrum or the XRPD

pattern (see the Supporting Information). This characteristic

remained unaffected when increasing amounts of H₂O were

added to the crystal suspension (up to 30 vol% H₂O). In

long-term experiments we could not find any signs of degra-

dation when MFU-2 crystals were stored under ambient

conditions (>6 months), which was in sharp contrast to the

moisture sensitivity of many zinc terephthalate type MOFs.

To compare catalytic activities of both frameworks, our

previous report about the liquid-phase oxidation of cyclo-

hexene by tert-butyl hydroperoxide in the presence of MFU-

1^[31]was complemented by similar investigations conducted

for MFU-2, which are described in the following section.

For MFU-1, we showed that the heterogeneous catalytic re-

action took place inside the pores of the MOF lattice.^[31]X-

ray powder diffraction data for the recovered catalyst gave

no signs of framework decomposition. During reaction with

tert-butyl hydroperoxide, microcrystals of MFU-1 underwent

a slow colour change from blue to green and the BET sur-

face of MFU-1 was found to become slightly reduced from

1485 to 1018 m²g^ꢀ1after the first catalytic run.

The experimental data for the liquid-phase oxidation of

cyclohexene by MFU-2, employing tert-butyl hydroperoxide

as the oxidant, demonstrated that oxidation of cyclohexene

was fast and multiple turnover was achieved (Figure 16),

similar to the results reported previously for MFU-1.

Whereas the oxidation reaction that occurred in the absence

of catalyst was almost negligible (ca. 1% conversion after

12 h), the maximum cyclohexene conversion after 22 h for

MFU-2 was 16% (TON=27), and the main reaction prod-

ucts were tert-butyl-2-cyclohexenyl-1-peroxide, followed by

2-cyclohexen-1-one and cyclohexene oxide, as revealed by

combined gas chromatographic and mass spectrometric

product analysis. The conversion achieved by using MFU-2

was slightly lower than that of MFU-1 under the same ex-

perimental conditions (MFU-1: 27.5% cyclohexene conver-

sion after 22 h, TON=27). The product distributions in both

cases, however, were almost identical.

Figure 16. Yield [%] versus time [h] curves for cyclohexene oxidation

with MFU-2 as the catalyst. Reaction conditions: cyclohexene (16 mmol),

tBuOOH (8 mmol), 1,2,4-trichlorobenzene (2 mmol; internal standard),

&

MFU-1 or MFU-2 (0.095 mmol based on cobalt centres), T=708C.

:

~

*

total conversion, : tert-butyl-2-cyclohexenyl-1-peroxide, : 2-cyclohex-

!

en-1-one, : cyclohexene oxide.

pared with MFU-1 after the reaction with tert-butyl hydro-

peroxide and cyclohexene. For MFU-2, no colour change

was observed during catalysis. Though the crystal morpholo-

gy was retained, the X-ray powder diffraction pattern of the

recovered catalyst showed an amorphous product (Support-

ing Information). In accordance with this finding, BET anal-

ysis revealed a loss of porosity (drop from 1675 m²g^ꢀ1to ca.

3 m²g^ꢀ1). These changes for MFU-2 indicated decomposi-

tion of the catalyst in contrast to the characteristics of

MFU-1, as exemplified in the following paragraphs.

Metal-leaching behaviour: To confirm the heterogeneous

nature of the catalytic test reactions, we performed hot-fil-

tration experiments, since it was known that even very small

amounts of metal ions leaching into solution could be the

single source of the observed catalytic activity with metal-

substituted zeolite catalyst.^[55]

Typical catalytic test runs were conducted as previously

described. Upon reaching a substrate conversion of approxi-

mately 50% (which took about 2 h for MFU-1 and 4.5 h for

MFU-2), the catalyst particles were removed from the hot

solution by isothermal filtration. The filtrate obtained from

the test run containing MFU-1 showed no significant cata-

lytic conversion upon catalyst removal, indicating that con-

tributions of soluble species towards catalytic activity were

negligible (Figure 17). In contrast to this, the filtrate from a

reaction mixture containing MFU-2 shows only a slightly di-

minished catalytic activity, as can be seen in Figure 17. The

different behaviour of both MOFs in hot filtration experi-

ments indicates that MFU-1 behaves as a truly heterogene-

ous catalyst, whereas the catalytic activity of MFU-2 is

mostly due to soluble metal complexes leaching from the

framework under catalytic conditions. This conclusion has

support from complementary structural investigations, which

Despite the seemingly similar catalytic properties of both

frameworks, MFU-2 showed marked differences when com-

8682

www.chemeurj.org

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2011, 17, 8671 – 8695

Pyrazolate-Based Cobalt(II)-Containing Frameworks

FULL PAPER

808C, the catalytic activity of MFU-1 increases remarkably,

but in this case slight metal leaching is observed in hot-fil-

tration experiments.

X-ray absorption spectroscopic investigations: To investigate

structural changes of the coordination and the redox states

of the cobalt centres of the MOFs before and after catalysis,

extended X-ray absorption fine structure (EXAFS) and X-

ray absorption near-edge structure (XANES) measurements

were performed at the cobalt K edge. Possible changes (if

any) in the catalytic centres (MOF nodes) could be unde-

tected by XRD measurements if no long-range crystallo-

graphic order was present and might have been character-

ised by EXAFS and XANES measurements as well.

In EXAFS, an electron is displaced from its shell (or com-

pletely from the absorbing atom) and interacts with its

neighbouring atoms. Depending on the wavelength (energy)

of the electron and the interatomic distances, constructive

or destructive interference of the electron waves can occur,

thus EXAFS provides information about the local atomic

environment around the absorbing atom. Close to the edge

energy of the absorbing atom, valence-shell electron transi-

tions have to be considered, and thus the information from

XANES spectroscopy is complementary to UV/Vis spectros-

copy, since the observed features are due to valence-shell

electron transitions and are thus influenced by the number,

coordination geometry and chemical nature of the surround-

ing ligands.

Each measurement was repeated at least three times and

the average of the measured data was taken. The Co K-edge

EXAFS spectra were analysed by using the standard

Athena and Artemis FEFF XAFS analysis codes,^[56]fitting

to models derived from the single-crystal structure data. The

fitted parameters (path lengths, Debye–Waller factors and

energy shifts) are given in the Supporting Information. The

raw and fitted Fourier transforms of the k³-weighted c(k)

spectra at the Co K edge are shown in Figure 18, and the

obtained atomic distances before and after catalysis are

given in Table 5 and compared with the values from the

crystal structure analyses. A comparison of the XANES

spectra of MFU-1 before and after catalysis is given in

Figure 19 below, indicating only a slight increase in the aver-

age oxidation state after catalysis.

Figure 17. Conversion [%] versus time [h] curves for cyclohexene oxida-

&

*

tion with MFU-1 ( ) or MFU-2 ( ) as the catalyst and after removing

the catalyst from suspension after 2 (&) or 4.5 h (*).

demonstrate a loss of porosity and crystallinity for MFU-2

samples under catalytic conditions.

All filtrates from experiments described above were addi-

tionally analysed by atomic absorption spectroscopy (AAS).

These studies yielded a surprising result: the amount of

“free” Co^IIions in the filtrate was below 1.1ꢃ10^ꢀ3% for

both test series (and thus beyond the lower detection limit

of the AAS instrument.). Whereas for MFU-1 this finding is

in accordance with the heterogeneous nature of the catalyst,

the unexpectedly low concentration of cobalt ions in the

case of MFU-2 indicates that the homogenously dissolved

species liberated from MFU-2 possess a high catalytic activi-

ty.

Additional AAS experiments indicated that leaching of

Co species from the MOFs was accelerated with an increas-

ing amount of tert-butyl hydroperoxide. In the absence of

cyclohexene, both frameworks decompose quickly, leading

to a nearly quantitative release of soluble Co ions into the

solution. Leaching can be vastly reduced by increasing the

amount of cyclohexene, which indicates that the equilibrium

between cobalt ions coordinated to the solid frameworks

and those forming soluble species, is shifted in favour of the

MOFs by a careful adjustment of the cyclohexene/tert-butyl

hydroperoxide ratio (Table 4). Apart from this, the reaction

temperature plays a crucial role in leaching behaviour: At

For MFU-1, differences are discernible for the EXAFS

spectra before and after catalysis (Figure 18): While the

main features are retained, slight differences in the Fourier-

transformed spectrum above 2 ꢄ can be seen. The region

from 0 to 2 ꢄ corresponds to the first coordination shell and

is, within statistical accuracy, identical in both cases, reveal-

ing that the fourfold coordination of Co atoms remains un-

changed.

Table 4. Concentrations of cobalt ions determined by AAS (after a reac-

tion time of 24 h).^[a]

Ratio tBuOOH/cyclohexene

Leached Co ions [%]

MFU-1

MFU-2

1:0

3:1

50.12

0.011

52.30

0.004

To derive a model for the changes in the [Co₄OACHTUNGTRENNUNG(3,5-

dmpz)₆] coordination unit of MFU-1 upon catalysis, differ-

ent structures were tested, of which the model displayed in

Figure 18 still gave the best fit. Differences between both

EXAFS fitting curves shown in Figure 18 are thus expressed

in terms of slightly changing distances and Debye–Waller

1:1 or 1:2

<0.001

[a] Reaction conditions: cyclohexene (0, 2.7, 8 or 16 mmol, as indicated),

tBuOOH (8 mmol), 1,2,4-trichlorobenzene (2 mmol) (as an internal stan-

dard), MFU-1 or MFU-2 (0.095 mmol based on cobalt centres), T=

708C.

Chem. Eur. J. 2011, 17, 8671 – 8695

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

8683

D. Volkmer et al.

Table 5. Half of the path length R as obtained by the fit to the EXAFS

data, corresponding to the bond distance (for single-scattering contribu-

tions), for MFU-1 before and after catalysis, compared with the distances

from single crystal X-ray structure analysis. The path labelling refers to

the coordination unit shown in Figure 18.

Scattering path Before catalysis

[ꢄ]

After catalysis

[ꢄ]

Crystal struct

[ꢄ]

ACHTUNGTRENNUNGure

ꢀ

Co_abO_{1_1}

2.00(3)

1.97(1)

2.88(2)

3.07(2)

3.18(2)

3.38(4)

4.01(12)

4.09(12)

4.08(11)

1.87(1)

2.00(1)

2.83(1)

3.20(2)

3.15(3)

3.38(3)

4.04(5)

4.12(5)

4.19(4)

1.95

1.98

2.88

3.07

3.18

3.53

4.05

4.20

4.22

ꢀ

Co_abN_{1_1}

ꢀ

Co_abN_{1_2}

ꢀ

Co_abC_{1_1}

ꢀ

Co_abCo_{1_1}

ꢀ

Co_abC_{3_1}

ꢀ

Co_abC_{1_2}

ꢀ

Co_abC_{2_1}

Co_ab-N_{1_1}-C_{2_1}

N_{1_1}

-

ꢀ

Co_abN_{1_3}

4.29(8)

5.32(9)

5.66(6)

5.71(6)

4.23(2)

5.13(4)

5.38(4)

5.42(4)

4.37

5.35

5.62

5.63

ꢀ

Co_abC_{3_2}

ꢀ

Co_abC_{4_1}

ꢀ

Co_abC_{1_3}

before catalysis (Table 5), but is still in a value range in

favour of a cobalt(II) centre in a tetrahedral coordination

ꢀ

site. Moreover, the structural changes relating to the Co N

distances are negligible. Two scenarios are possible: either

only negligible geometrical changes occur equally on all

{Co₄O} clusters or only a small fraction of the clusters in

MFU-1 are changed during the course of catalysis.

To complement EXAFS data for MFU-1, XANES spectra

were examined, providing specific information about the va-

lence state(s) of the cobalt ions in MFU-1 before and after

catalysis. Figure 19 shows the XANES spectra of MFU-1

before and after catalysis. The shapes of both XANES spec-

tra are very similar and it can therefore be concluded that

the local coordination environments for the huge majority

of cobalt centres are tetrahedral and thus identical. The

most significant features of the XANES spectra are charac-

Figure 18. Top: Magnitude (bold line) and imaginary part (thin line) of

the k³-weighted Fourier transformation versus the radial distance R from

the absorbing atom of the EXAFS spectra of MFU-1 before catalysis

(top, k range=1–12, fitted from 1.0–5.0 ꢄ with 17 variables for 27 inde-

pendent points, R=0.0004 and reduced c²=82) and after catalysis

(bottom, k range=1–14, fitted from 1.0–5.0 ꢄ with 15 variables for 33 in-

dependent points, R=0.0026 and reduced c²=90). The raw data and the

fitting curve are shown as solid and dashed lines, respectively. Bottom:

Stick representation and labelling of the cluster used for EXAFS data

analysis. The absorbing Co atom is highlighted as a filled sphere.

factors. Alternative structural models incorporating addi-

tional oxygen atoms coordinated to the Co centres gave

higher R and c²values (poorer fits) and gave unreasonably

huge shifts for some distances (in particular, for the Co···Co

distances).

ꢀ

Literature values of Co O distances for cobalt ions in tet-

rahedral coordination sites are 1.93(6) ꢄ for cobalt(II) and

1.81(3) ꢄ for cobaltACHTUNGTRENNUNG

(III).^{[27, 57]}For MFU-1 after catalysis, the

ꢀ

Co O distance is 1.87(1) ꢄ, according to the EXAFS fitting

data, which is lower than the original value (2.00(3) ꢄ)

Figure 19. Normalised Co K-edge XANES spectra. P=pre-edge peak,

A=peak edge; W=white line.

8684

www.chemeurj.org

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2011, 17, 8671 – 8695

Pyrazolate-Based Cobalt(II)-Containing Frameworks

FULL PAPER

terised by the pre-edge peak (P in Figure 19), the absorption

edge (A in Figure 19) and the white line (W in Figure 19).

The pre-edge structures at the Co K edges (P) usually fall

into the energy range between 7709 and 7712 eV (here:

7710.5 eV) and can be ascribed to quadrupole 1s–3d and/or

to modifications of dipole transition probability due to hy-

bridisation between 3d and 4p states (since the dipole 1s–3d

transition is dipole forbidden).^[58]When the metal is located

in a non-centrosymmetric site, such as in a tetrahedral com-

plex, extensive molecular orbital mixing of the metal 3d or-

bitals with the metal 4p orbitals occurs, thus making the 1s

to 3d–4p mixed orbital transition angular momentum al-

lowed. This mixing is known to be more extensive in tetra-

hedral than in (pseudo)octahedral complexes and increases

as the coordination number decreases. Perfectly octahedral

(centrosymmetric) complexes are expected to have no pre-

edge features in the XANES spectrum because both A_1g!

T_2gand A_1g!E_gelectronic transitions are Laporte forbid-

den.^[59]Thus, for both MFU-1 samples, a non-centrosymmet-

ric ligand environment around the cobalt ions is most proba-

ble. Comparing the samples before and after catalysis, the

pre-edge peaks are nearly identical, and only a slight de-

crease in intensities occurs for the MFU-1 sample after cat-

alysis. The first strong absorption (A) occurs as a shoulder

on the low-energy side, as clearly indicated by a distinct

peak in the first-derivative plots of all XANES spectra. Be-

cause of its high intensity, shoulder A must be assigned to a

metal 1s to 4p transition; a dipole-allowed transition. In the

absence of any “shake-down” contribution, the energy of

shoulder A, and hence, the absorption edge energies (ener-

gies of the three metal 4p orbitals) are 7716.1 and 7716.5 eV

before and after catalysis, respectively. A shift in the edge

energies usually indicates a change in the oxidation state.

The peak edge should be shifted to higher energy for cobalt

centres of larger oxidation state; a systematic trend that is

often utilised to estimate the oxidation state of cobalt,^[60]

whereby a linear dependence of the chemical shift on the

average valence can be assumed.^[61]Examination of refer-

ence samples (elemental cobalt, Co₃O₄and Co^IIIpyrazo-

late)^[62]also show this linear relationship (a shift of 3.69 eV

per oxidation state occurs, see the Supporting Information

for more information), from which an oxidation state of

+2.08(7) can be extrapolated for MFU-1 after catalysis. As

already seen from the two spectra in Figure 19, this change

is quite small (too small to be accurately quantified), but

discernible, providing further support for our previous state-

ments according to which only a small fraction of {Co₄O}

units are affected after catalytic runs. Finally, a slight change

in the coordination geometry around cobalt is also indicated

by the decrease in the intensity of the white line (W) that

shifts from 7729.4 to 7729.9 eV. These findings from

XANES are in agreement with the changes observed in the

bond lengths (EXAFS) and the results from UV/Vis. In con-

clusion, the most likely explanation for the observed spec-

tral and structural changes occurring in MFU-1 upon cataly-

sis is that just a small fraction of cobalt centres/coordination

units is affected by the catalytic reaction conditions, for

which a change in the valence state from (II) to (III) seems

most likely. The presence of cobaltACHTUNTRGNEUNG(III) centres is corrobo-

rated by UV/Vis and XPS^[31]experiments, the exact amount

of which is below a concentration level that would allow a

more quantitative determination and thus preventing us

from making a detailed structural characterisation of the

cobalt species that appear under catalytic conditions.

Whereas EXAFS curves of MFU-1 show clear differences

in samples before and after catalysis (especially above 2 ꢄ

in the Fourier-transformed spectra), the coordination envi-

ronments and valence states of the cobalt ions in MFU-2

are completely preserved under catalytic conditions (see the

Supporting Information). Since XRPD investigations reveal

a loss of crystallinity and porosity for this framework under

catalytic conditions, the results from X-ray absorption spec-

troscopy (XAS) investigations appear surprising at first

glance. However, according to AAS investigations under

catalytic conditions, only a tiny amount of soluble cobalt

species leaches into solution, which might be explained by a

collapse of the porous framework due to the removed

cobalt ions, composed of MFU-2-like fragments.

The excellent fits to the EXAFS data for both MFU-1

and MFU-2 for the samples before catalysis support the val-

idity of the structure obtained from single-crystal X-ray

structure analysis. In particular, no distinct cobalt-containing

species are observable, suggesting that the catalytic activity

is due to the metal sites in these two MOFs (and not due to

further unknown cobalt-containing impurities or side prod-

ucts in the voids of the network). After employing MFU-1

in catalysis, slight changes in the EXAFS and XANES spec-

tra are observed.

When summarising results from spectroscopic and struc-

tural investigations obtained so far, the following interpreta-

tion seems permissible in light of catalytic properties: al-

though tetrahedrally coordinated cobalt(II) centres are pres-

ent in both frameworks, their metal centres are remarkably

resistant against oxidation (yielding cobaltACTHNURTGNEU(GN III)) and they

show no tendency to accept further ligands in their first co-

ordination spheres. The stabilisation of the cobalt(II) oxida-

tion state seems to be an intrinsic property of these frame-

works. Usually, cobaltACTHNUGRTENUNG(III) peroxo complexes of, for exam-

ple, N,N’-ethylenebis(salicylimine) (salen) or porphyrin com-

plexes are kinetically inert due to their low-spin d⁶configu-

ration, thus a large number of such complexes have been

isolated and characterised at room temperature.^[63]However,

due to steric hindrance and constraints upon the possible

distortion of individual coordination environments of the

cobalt centres by the 3D network, coordination of an addi-

tional ligand would constitute an energetically unfavourable

situation, and thus a strained or entatic state as described

above. A similar stabilisation is also seen in structurally re-

lated tetrahedral cobalt of tridentate scorpionate ligands

comprising bulky substituents at the pyrazolate moieties.^[64]

An octahedral coordination environment is also disfavoured

for these complexes due to steric reasons and complexes

with peroxides or oxygen can be only isolated in rare

cases^[64a]at low temperatures (<-508C for peroxo complexes

Chem. Eur. J. 2011, 17, 8671 – 8695

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

8685

D. Volkmer et al.

C

3 ðCH₃Þ₃COOH ! 2 ðCH₃Þ₃COO þ ðCH₃Þ₃COH þ H₂O

and <ꢀ10 to 308C for dioxygen complexes), and the reac-

tive intermediates usually oxidise the pyrazolate ligands in a

subsequent reaction at higher temperatures.

ð3Þ

Principally, diffusion limitation might be another reason

why only a minor part of the metal centres participates to

the oxidation process.^[65]However, in case of oxidation reac-

tions involving tert-butyl hydroperoxide and cyclohexene,

previous experiments employing microcrystals of different

sizes showed that diffusion limitation was present to some

extent, but it did not exert a major influence on the experi-

mental reaction rates in MFU-1.^[31]

A further hint toward a reaction path involving free radi-

cals is given when the oxidation reaction is performed in

CH₂Cl₂, as opposed to the solvent-free method that is gener-

ally applied. In this case, 3-chlorocyclohexene appears as an

additional byproduct. Moreover, radical trap experiments

confirm the formation of tert-butyl peroxy radicals and cy-

clohexenyl radicals throughout the course of the catalytic re-

action (see the Supporting Information). In consideration of

the homolytic bond dissociation energies of the involved

substrates and products,^[67]a detailed catalytic cycle (in anal-

ogy to earlier reports for similar reactions)^{[63c, 66]}for the for-

mation of tert-butyl peroxy radicals catalysed by MFU-1 can

be proposed (Scheme 1).

Mechanism of tert-butyl hydroperoxide activation: The fact

that tert-butyl-2-cyclohexenyl-1-peroxide is the main reac-

tion product suggests a reaction pathway in which freely dif-

fusing peroxy radicals are generated by cleavage of tert-

butyl hydroperoxide following the route of a Haber–Weiss-

type decomposition at the Co^IImetal centres (Scheme 1).^[63c,

In the first step, tert-butyl hydroperoxide is expected to

coordinate to the cobalt centres in MFU-1 (Scheme 1, K_eq).

66]

ꢀ

The net reaction equation can be formulated as Equa-

Thermodynamic considerations suggest that the weak O O

tion (3):

bond of tert-butyl hydroperoxide (rather than the considera-

II

ꢀ

bly stronger O H bond) is cleaved by the aid of the Co

ions in MFU-1 (Scheme 1(i)). Thus, tert-butoxy radicals and

hydroxyl radicals might be created.^[67]The resulting tert-

butoxy radical might exchange a hydrogen atom with anoth-

er tert-butyl hydroperoxide molecule (DHꢂꢀ71 kJmol^ꢀ1

from the bond dissociation energies) in a fast subsequent re-

action (k=1.3ꢃ10⁸m^ꢀ1s^ꢀ1for the comparable reaction of cu-

myloxy radicals with tert-butyl hydroperoxide).^[68]Conse-

quently, a tert-butyl hydroperoxy radical and tert-butanol

would be created. A minor fraction of the tert-butoxy radi-

cals might also react directly with cyclohexene to yield a 3-

cyclohexenyl radical (k=5.7ꢃ10⁶m^ꢀ1s^ꢀ1).^[69]The hydroxy

radicals created in the first step, or hydroxy anions if elec-

tron transfer occurs from Co^II, most likely react with anoth-

er tert-butyl hydroperoxide molecule under formation of

water (Scheme 1, K’_eq). The resulting cobalt hydroperoxo

ꢀ

species could react further through fission of either the O

ꢀ

O or the Co O bond. However, the coordinative bonds be-

tween the peroxo ligands and the Co^IIIcentres are probably

very weak, owing to the steric shielding of the metal ions

and the rigid coordination geometry imposed by the crystal-

line framework. This assumption is supported by the low

isosteric heat of adsorption of molecular oxygen and the un-

changed IR spectra after a catalytic run employing tert-butyl

hydroperoxide (see the Supporting Information). This spe-

cial situation presumably leads to formation of tert-butyl

peroxy radicals (by fission of the cobalt–oxygen bond (as

proposed in Scheme 1(ii)) rather than to homolytic fission of

the oxygen–oxygen bond of the coordinated peroxo ligand.

After recovering the initial catalyst by removing the result-

ing tert-butyl hydroperoxy radical from the Co atom in the

framework, the catalytic cycle starts over again. In a subse-

quent step, the formed tert-butylperoxy radicals would react

with cyclohexene to yield the observed products (Schemes 1

and 2). Two pathways are possible for the subsequent reac-

tion of tert-butyl peroxy radicals with cyclohexene. First, the

abstraction of a hydrogen atom from the allylic position of

Scheme 1. Top: Mechanism for the formation of tert-butylperoxy radicals

catalysed by the cobalt(II) centres in MFU-1. Their further reaction with

cyclohexene, forming the main product, is shown. Bottom: Homolytic

ꢀ1 [63c, 66]

ꢀ

O H and C H bond dissociation energies [kJmol

]

of the educts,

products and intermediates involved in the catalytic formation of tert-bu-

tylperoxy radicals by MFU-1.

8686

www.chemeurj.org

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2011, 17, 8671 – 8695

Pyrazolate-Based Cobalt(II)-Containing Frameworks

FULL PAPER

and tend to decompose (recovering the initial Co^IIspecies

of the catalyst, whilst liberating the peroxy radicals).

Mechanism of O₂activation: Due to its triplet electronic

ground state, the spontaneous oxidation of organic sub-

strates by molecular oxygen is often kinetically unfavoura-

ble, since its direct utilisation in reactions with singlet-state

molecules is restricted by the Wigner spin conservation

rule.^[74]The diradical character of the free paramagnetic

oxygen molecule thus leads to preferred reactions with radi-

cals or paramagnetic metal ions.^[75]The most numerous syn-

thetic oxygen complexes are known for cobalt(II) com-

pounds,^[75]thus oxidation reactions employing MFU-1

should not only be possible when employing tert-butyl hy-

droperoxide as an oxidant, but also with molecular oxygen.

MFU-1 indeed shows a slow but constant reaction with

molecular oxygen at T>1208C, which is accompanied by a

colour change from blue to green. A similar oxidation be-

haviour is not seen for MFU-2, which shows no sign of reac-

tion with oxygen or decomposition below 2308C. However,

the interactions between O₂and MFU-1 below 1208C are

weak, as expressed by an isosteric heat of adsorption of

about only 11.1 kJmol^ꢀ1, which is probably due to the effec-

tive steric shielding of the cobalt centres. The electronic in-

teractions between dioxygen and cobalt centres in MFU-1

are clearly too weak to activate oxygen, since all attempts in

our hands of gas-phase reactions between molecular oxygen

and different hydrocarbons in the presence of MFU-1 have

failed so far, showing no or extremely low yield of the an-

ticipated products.^[76]

In recent years, N-hydroxyphthalimide (NHPI, Scheme 3)

has been recognised as a valuable co-catalyst (acting as an

electron-transfer mediator) for the aerobic oxidation of or-

ganic compounds under mild conditions. The use of coupled

catalytic systems with electron-transfer mediators usually fa-

cilitates the procedures by transporting the electrons from

the catalyst to the oxidant along a low-energy pathway,

thereby increasing the efficiency of the oxidation and thus

complementing direct oxidation reactions. As a result of

similarities with biological systems, this can be dubbed a

biomimetic approach.^[77]NHPI, as a co-catalyst, is trans-

ferred in combination with transition-metal complexes^[78,81]

into PINO by molecular oxygen (Scheme 3). NHPI is regen-

erated by abstracting a hydrogen atom from the substrate.

In the case of MFU-1, a putative mechanism includes the

chemisorption and activation of molecular oxygen by one of

the cobalt(II) centres (Scheme 3, K_eq). In a subsequent step,

the activated oxygen may abstract a hydrogen atom from

Scheme 2. Formation of the observed products through the reaction of

tert-butyl peroxy radicals with cyclohexene.

cyclohexene, and second the addition of the radical to the

double bond. Whereas the second case is usually observed,

for instance, in the case of carbon-centred radicals,^[70]for

tert-butyl peroxy radicals predominantly (ꢂ98%) the ab-

straction of an allylic proton is observed.^[71]Radical combi-

nation of the resulting 3-cyclohexenyl radicals with another

tert-butyl peroxy radical would yield the observed main

product, tert-butyl-2-cyclohexenyl-1-peroxide. The byproduct

2-cyclohexen-1-one is most likely formed by thermal decom-

position of the main product.^[72]The addition of tert-butyl

peroxy radicals to the double bond as a minor reaction fol-

lowed by cleavage of a tert-butoxy radical should lead to cy-

clohexene oxide, which explains the formation of the other

byproduct observed.

Only the asymmetric product of the radical recombination

between tert-butyl peroxy and 3-cyclohexenyl radicals is

found. No symmetric product, for example, 3,3’-bicyclohex-

ene, can be observed from the recombination of two 3-cyclo-

hexenyl radicals. Such a preference of the asymmetric prod-

uct in radical recombination reactions can be explained by

the Ingold–Fischer “persistent radical effect”.^[73]Cross-cou-

pling occurs when one radical is “transient” (undergoing

fast reactions) and the other is “persistent” (longevity by,

for example, steric crowding), and both are generated at

comparable reaction rates. Thus, a steady state involving

high concentrations of persistent and low concentrations of

transient radicals occurs, leading to preferred cross-coupling.

The Ingold–Fischer persistent radical effect was observed

earlier for the homogenously catalysed and uncatalysed re-

actions between persistent tert-butyl peroxy radicals and

transient cyclohexene radicals.^[66d]

ꢀ

NHPI (Scheme 3, K’_eq). Co O bond fission would then re-

cover the catalyst.

Another advantage of the reaction cascade employing

PINO radicals is their higher reactivity than that of peroxy

ꢀ

or oxy radicals. Since O H bond dissociation energies for

The fact that no substantial change in the IR, XANES or

EXAFS spectra is observed after multiple turnovers indi-

cates that the Co^IIIspecies are energetically unfavourable in-

termediates, which thus occur only in small concentrations

both NHPI and tert-butyl hydroperoxide are almost identical

(ꢂ370 kJmol^ꢀ1),^[79]the increased reactivity has to thus be

explained by a “polar effect”. The presence of two carbonyl

groups is responsible for a pronounced electrophilic charac-

Chem. Eur. J. 2011, 17, 8671 – 8695

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

8687

D. Volkmer et al.

tivities were qualitatively in

agreement with the expected

behaviour for oxidation reac-

tions performed by the NHPI/

PINO system. The second-

order rate constants for the hy-

drogen abstraction from differ-

ent substrates by PINO (ex-

pressed per active hydrogen

atom) at 258C in acetonitrile

were^[81b]k=5.05m^ꢀ1s^ꢀ1for cy-

clohexene, k=4.52m^ꢀ1s^ꢀ1for

cyclohexanol, k=1.12m^ꢀ1s^ꢀ1for

ethylbenzene

and

k=

0.0039m^ꢀ1s^ꢀ1for cyclohexane,

thus the obtained reactivities in

our experiments were ex-

plained. When monitoring the

time course of the oxidation re-

action of cyclohexene in detail

(Table 6, entry 3), further mech-

anistic insights became appar-

ent. During the first 8 h of the

reaction, only low conversions

Scheme 3. Formation of phthalimide-N-oxyl radicals (PINO) from N-hydroxyphthalimide (NHPI) involving

cobalt(II) centres.

ter of PINO, lowering the energy of the transition state

(Scheme 4) and thus increasing its reactivity.^[80]

MFU-1 (1 mol%) was capable of catalysing the oxidation

of cyclohexene, ethylbenzene or cyclohexanol by using at-

mospheric oxygen at close to ambient conditions (i.e., 358C,

Table 6) when NHPI (10 mol%) was added. No conversion

was observed in the absence of either components (MFU-1

or NHPI). Due to the poor sol-

Scheme 4. Transition state for the reaction of PINO with cyclohexene.^[81]

ubility of NHPI in most organic

Table 6. Oxidation of hydrocarbons by molecular oxygen catalysed by MFU-1 and NHPI for 24 h.^[a]

solvents, the catalytic test reac-

tions were carried out in aceto-

nitrile.^[81]

Entry T [8C] Educt

Products and yields [%]

Conversion [%]

Under mild reaction condi-

tions, cyclohexanol (Table 6,

entry 1) was nearly completely

transformed to cyclohexanone

after 24 h, whereas secondary

allylic or benzylic substrates

were converted to peroxides or

ketones, respectively (Table 6,

entries 3–5). Higher tempera-

tures facilitated the further de-

composition of the peroxides

(compare entries 4 and 5,

1

2

3

35

75

35

!

94

95

1.5

2

26.6

+

5.6

+

0.8 35

4

5

35

45

!

5.1

1.7

+

1.2

7

ꢀ

14.5

18

Table 6). Non-activated C H

bonds, such as those from cyclo-

[a] Reaction conditions (if not stated otherwise): atmospheric oxygen, NHPI (50 mg, 0.31 mmol), MFU-1

(15 mg, 0.038 mmol based on cobalt centres), 1,2,4-trichlorobenzene (150 mL, 1.2 mmol; internal standard),

educt (3.0 mmol; cyclohexene (300 mL), ethylbenzene (365 mL), cyclohexanol (270 mg) or cyclohexane

(300 mL)), acetonitrile (5 mL), T=358C, 24 h. The peroxides were identified by quantitative reduction to the

corresponding alcohols by using triphenylphosphine and identified by their changed retention time in GC-

hexane

(Table 6,

entry 2),

showed only sluggish reaction

even at elevated temperatures

(yielding 1.5% cyclohexanone

after 24 h). The observed reac- MS.^[82]

8688

www.chemeurj.org

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2011, 17, 8671 – 8695

Pyrazolate-Based Cobalt(II)-Containing Frameworks

FULL PAPER

were observed (yielding ꢂ0.5% of 3-cyclohexenylperoxide),

whereas after this induction period, the conversion rate in-

creased rapidly (Figure 20). When the solid catalyst was fil-

tered off from the reaction mixture after 16 h, the filtrate

in autoxidation reactions of cyclohexene or cyclopentene

catalysed by rhodium, iridium or platinum complexes.^[83]

Figure 20. Conversion [%] and yields [%] for the reaction of cyclohexene

and oxygen at 358C in the presence of catalytic amounts of NHPI and

MFU-1 versus time. The dotted lines and empty symbols represent the

time course of the reaction upon removal of the MFU-1 catalyst by filtra-

tion after 16 h.

showed a drastically decreased catalytic activity, thus dem-

onstrating that the reaction was heterogeneous in nature in

this case as well. The remaining reactivity of the filtrate

might be due to the spontaneous (non-catalysed) decompo-

sition of 2-cyclohexenylperoxide and its reaction with NHPI.

After a reaction time of 16 h, the MFU-1 catalyst did not

change according to XRPD and IR analysis, and UV/Vis

spectra indicated once again the partial formation of cobalt-

According to Equations (4) to (6), the PINO radicals ab-

stract a hydrogen atom from cyclohexene, which subse-

quently reacts with molecular oxygen to yield 2-cyclohexe-

nylperoxide. The net reaction is thus formulated according

to Equations (4)–(10). Whereas at the initial stage of the

catalytic reaction, PINO radicals are predominantly ob-

tained from the reaction of NHPI with molecular oxygen

(activated by cobalt centres), the former radicals are pro-

duced more efficiently at a later stage by catalytic decompo-

sition of the reaction product 2-cyclohexenylperoxide.

ACHTUNGTRENNUNG(III) for this reaction. However, long reaction times (greater

than 32 h) exerted a detrimental effect upon the catalyst,

which slowly decomposed to form an amorphous product.

The decomposition of the framework in the presence of

PINO was probably accelerated by the products developed

during the reaction, in particular, due to the formation of a

huge excess of 3-cyclohexenylperoxide.

Heterogeneous catalytic oxidation by using NHPI@MFU-1

and molecular oxygen: The advantage of heterogeneous cat-

alysts lies in the ease of their recovery from reaction mix-

tures, allowing the reaction products to be obtained without

additional purification steps. Additionally, the uniform pores

of MFU-1 provide a means of size- or shape-selective cata-

lytic oxidation reactions. These technical advantages are

lessened by the use of a solution of NHPI in acetonitrile,

due to the high costs of the solvent and the continuous loss

of the co-catalyst upon filtration. Consequently, the immobi-

lisation of both catalytic compounds, cobalt(II) and NHPI,

would be preferable. This approach was successful for cobal-

t(II) salen complexes and NHPI derivates; both were cova-

lently bound to a mesoporous silica carrier.^[84]To immobilise

NHPI in the pores of MFU-1, MFU-1 (15 mg) were sus-

pended for 3 d in a solution of NHPI (50 mg) in acetonitrile

(5 mL) (the same stochiometry as in the catalysis reactions

described above). Afterwards, the MFU-1 material hosting

the co-catalyst (termed hereafter NHPI@MFU-1) was fil-

tered off and dried in vacuo. Indeed, NHPI@MFU-1 cata-

The reason for the initial inhibition observed in the catal-

ysis reaction was presumably due to autocatalysis of the re-

action by the formed 2-cyclohexenylperoxide (see

Scheme 3). Whereas initially the reaction was driven by the

(slow) activation of oxygen on MFU-1 [Eqs. (4)–(7)], later

on 2-cyclohexenyloxo and -peroxo radicals were formed

during the course of the reaction, which appeared to be acti-

vated much faster at the cobalt(II) centres of MFU-1

[Eqs. (8)–(10)]. Accordingly, when catalytic amounts of tert-

butyl hydroperoxide (1 mol%) were added at the beginning

of the reaction, no inhibition was observed, yielding 19% of

2-cyclohexenylperoxide after 3 h. This appears to be due to

efficient activation of peroxides by MFU-1, when directly

compared with the corresponding activation of molecular

oxygen. During the course of the reaction, the peroxy radi-

cals should thus be mainly in charge of hydrogen abstraction

from NHPI. A similar participation of peroxides is also seen

Chem. Eur. J. 2011, 17, 8671 – 8695

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

8689

D. Volkmer et al.

lysed the oxidation of cyclohexene at 358C without any ad-

ditional solvent (Figure 21, filled symbols and solid lines).^[85]

Moreover, the activity of this heterogeneous system was

contrasting metal-leaching behaviour is observed for both

compounds under catalytic conditions: Whereas MFU-1 per-

forms as truly heterogeneous and robust catalyst, MFU-2

(albeit showing a similar catalytic activity) suffers from

metal leaching into solution. This clearly demonstrates the

importance of conducting metal-leaching tests in catalytic

investigations. The ability of the networks to adsorb and ac-

tivate molecular oxygen was examined by measuring the iso-

steric heat of adsorption and by TPO experiments, revealing

a higher oxygen affinity of MFU-1 than MFU-2. DFT calcu-

lations show that a superoxo complex might form between

the {Co₄O} nodes in MFU-1 and molecular oxygen, which

displays a hugely distorted coordination polyhedron for the

Co^IIatom involved in coordination, leading us to propose a

geometrically and energetically strained catalytic intermedi-

ate, that is, an entatic state of the catalyst. However, the ac-

tivation energy for the formation of this superoxo complex

is high, since below 1208C, complex formation cannot be

observed for MFU-1. Once formed, and in the absence of

any easily oxidisable substrate, this highly reactive entatic

species leads to rapid oxidation of the framework in multi-

ple subsequent steps. At temperatures below 1208C, oxygen

shows only weak interactions with the cobalt centre of

MFU-1, and the resulting adsorbed O₂species can be de-

scribed as a van der Waals complex. Despite the low coordi-

nation number, the cobalt centres show no tendency to open

up their coordination spheres, which can be explained in

terms of the steric shielding of the cobalt centres by the

methyl substituents of the coordinated pyrazolate moieties

and by symmetry constraints imposed by the 3D periodic ar-

rangement of the SBUs; the latter represents a different

kind of strained (i.e., entatic) state of the catalyst. The low

coordination ability of the cobalt centres represents a dis-

tinct advantage in catalytic oxidation reactions, since the for-

mation of stable cobalt complexes with oxygen-centred li-

gands is thus avoided, which might otherwise constitute a

“thermodynamic sink” and thus a potential dead end of a

catalytic cycle.^[75]Preventing formation of a coordinative

bond between O₂and cobalt centres leads to the observed

low isosteric heats of adsorption ((11.1ꢁ0.9) kJmol^ꢀ1for

MFU-1 and (8.5ꢁ0.9) kJmol^ꢀ1for MFU-2). Owing to the

weak binding of O₂at low temperatures, we were therefore

not able to characterise any structurally well-defined inter-

mediate species. By employing tert-butyl hydroperoxide as

an oxidant, the {Co₄O} core of MFU-1 showed small geo-

metrical changes under catalytic conditions, which we as-

cribed to an increase in the oxidation state from cobalt(II)

Figure 21. Conversions [%] and yields [%] versus time for the reaction of

cyclohexene with molecular oxygen at 358C. Filled symbols and solid

lines: solvent-free catalysis employing NHPI@MFU-1; Open symbols and

dotted lines: catalysis employing MFU-1 suspended in NHPI-containing

acetonitrile.

only slightly reduced in comparison with the system contain-

ing

a homogeneous solution of NHPI in acetonitrile

(Figure 21, open symbols and dotted lines).

Thus, the oxidation of cyclohexene (and presumably a

huge range of other organic substrates) can be carried out

as a suspension of NHPI@MFU-1 in the organic substrate,

thus avoiding the use of additional solvent, simplifying

workup procedures and reducing the amount of waste. Com-

pared with the usually drastic conditions employed for the

oxidation of cyclohexene (high oxygen pressure and high

temperatures),^[81a,86]the reaction can be carried out by using

NHPI@MFU-1 under more cost-saving and environmental-

friendly conditions (358C, atmospheric oxygen).

Conclusion

We have presented the solid-state structures of two pyrazo-

late-based MOFs (MFU-1 and MFU-2) of cobalt(II) centres

connected by the linear linker 1,4-bis[(3,5-dimethyl)pyrazol-

4-yl]benzene, which can be regarded as an N-heterocyclic

derivative of terephthalic acid. MFU-1 constitutes a structur-

al analogue of the Zn-containing (redox inactive) MOF-5,

featuring oxo-centred tetranuclear {Co₄O} cores, whereas

MFU-2 is composed of cross-linked 1D cobalt(II) chains.

The results from UV/Vis spectroscopy and XAS are in com-

plete accordance with the structural data obtained from

single-crystal X-ray analyses, demonstrating phase purity

and the absence of other impurities containing cobalt(II).

Both MOF compounds were compared in a catalytic test re-

action by using cyclohexene as the substrate and tert-butyl

hydroperoxide as the oxidant. Although Co^IIare found in

structurally similar (tetrahedral) coordination environments,

to cobaltACTHNUTRGENUG(N III) (based on data from XAS and UV/Vis investi-

gations), involving, however, only a small fraction of the

total metal centres. A summary of the most significant re-

sults is given in Table 7.

In terms of catalytic applications, the activation of molec-

ular oxygen by MFU-1 becomes feasible by taking advant-

age of a co-catalyst (NHPI), which can be immobilised in

the framework, leading to the composite catalyst NHPI@M-

FU-1. The immobilised NHPI serves as an electron-transfer

mediator, and thus, heterogeneous oxidation reactions em-

8690

www.chemeurj.org

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2011, 17, 8671 – 8695

Pyrazolate-Based Cobalt(II)-Containing Frameworks

FULL PAPER

Table 7. Summary of characteristic data for MFU-1 and MFU-2.

[a]

Pore

aperture

BET surface [m²g^ꢀ1

]

Pore

volume

After catalysis

Isosteric

Metal

Decomposition temper-

heat of ad- leaching ature^[c][8C]

sorption

A

[b]

[ꢄ]

before

catalysis

theoretical after

catalysis

N

]

VTXPRD EXAFS

XANES: Co

oxidation

state

[kJmol^ꢀ1

]

in catal- TGA VTXRPD TPO

ysis

MFU- 9.0

1

1485

(1525)

4117

1447

1018

0.56

(1.49)

unchanged slight geo-

metrical

2.08(7)

ꢀ11.1

ꢀ8.5

no

340

270

300

120

230

changes

MFU- 6.4

2

1477

(1675)

3

0.53

(0.55)

amorphous unchanged 2.00(7)

yes

[a] For samples from microwave syntheses; values for samples from solvothermal syntheses in parentheses. Theoretical values obtained by the program

developed by Dꢁren et al.^[48][b] For samples from microwave syntheses; theoretically expected from PLATON/SQUEEZE in parentheses. [c] Variations

are due to the different sensitivities and conditions used for the different methods. TGA gives the temperature at which the ligand is released in an inert

nitrogen atmosphere, whereas VTXRPD gives the temperature of phase transitions of the sample in static air. The TPO measurements give significantly

lower decomposition temperatures, since decomposition products are already registered at slow decomposition rates due to the higher sensitivity of the

MS. In addition, the gas flow rate has an influence on the decomposition temperature. Under static conditions, the presence of oxygen leads to totally

different decomposition temperatures in comparison to a flowing atmosphere. In TA and VTXRPD experiments, a decomposition product has to reach

a certain proportion to be detected by these methods. Thus, TPO with MS is the most sensitive method for investigating even small decomposition pro-

cesses.

ploying molecular oxygen can be performed at near-ambient

conditions.

Vis diffuse reflectance spectra (DRS) were recorded on an Analytik Jena

Specord 50 UV/Vis spectrometer in the range 190–1100 nm. Argon gas

sorption isotherms were measured with a Quantachrome Autosorb-I

ASI-CP-8 instrument. Prior to measurements, the samples of CH₂Cl₂-ex-

changed MFU-1 and MFU-2 were heated at 1508C for 30 h in high

vacuum to remove the occluded solvent molecules. Argon sorption ex-

periments were performed at 77 K in the range 1.00ꢃ10^ꢀ4ꢃp/p₀ꢃ1.00.

Gas chromatographic analyses were carried out with a Carlo Erba GC

8380 gas chromatographic instrument with a hydrogen flame ionization

detector. GC-MS data were recorded on a Varian GC 3800-Varian

Saturn 2000 MS instrument equipped with a WLD/ECD detector and a

Chrompack column (CP-5842, 25 m long). EXAFS and XANES meas-

urements were conducted at RT by using the double crystal monochro-

mator (SiGe (111) graded crystals, E/DE=4200) at the beamline KMC-2

of the electron storage ring at the Berliner Elektronenspeicherring-Ge-

sellschaft fꢁr Synchrotronstrahlung m.b.H. (BESSY II, Berlin, Germa-

ny).^[88]The XANES and EXAFS spectra were recorded at the Co K

edge; energy calibration was performed with cobalt metal foil. Measure-

ments were performed in transmission mode by using ionisation cham-

bers, additionally the fluorescence was recorded by using a fluorescence

yield detector (Si-PIN photodiode). AAS was performed on a Perkin–

Elmer 4100ZL Zeeman atomic absorption spectrometer. 30 mL aliquots

of the sample solution were dispensed automatically into the atomizer.

The standard additions method was used for calibration: wavelength:

242.5 nm, slit: 0.2 nm, pretreatment at 14008C for 30 s, atomization at

The pronounced differences in catalytic behaviour dis-

played by the structurally similar compounds MFU-1 and

MFU-2 indicate that little is known about structure–reactivi-

ty relationships in redox-active MOFs in general. Future re-

search may target the reactions leading to oxidative degra-

dation of the frameworks, as well as the influence of the

chemical nature of coordinate cobalt species at the external

boundary of the MOF microcrystals. Finally, by employing

refined and more sensitive spectroscopic techniques, the

structural characterisation of reactive MOF intermediates

might become feasible, serving as an information source for

detailed theoretical studies.

Experimental Section

Experimental methods and equipment used for investigations: tert-Butyl

hydroperoxide (70% H₂O solution) was dried over anhydrous MgSO₄.

ACTHNUTRGNEUNG

tert-Butyl-2-cyclohexenyl-1-peroxide^[87]and cobalt(III) pyrazolate^[62]were

synthesised according to literature procedures. Other materials were ob-

tained from commercial suppliers and used without further purification.

Hydrothermal syntheses of the MOFs were carried out on a TH20 tem-

perature-programmable heating block by HLC Biotec and microwave-as-

sisted syntheses were performed by using a CEM, Discover S-Class mi-

crowave system. FTIR spectra were recorded of KBr pellets in the range

4000–100 cm^ꢀ1on a Bruker IFS FTIR spectrometer. Elemental analyses

(C, H, N) were carried out on a Perkin–Elmer 2400 elemental analyzer.

25008C for 4 s, modifier: Mg

ACHTNUGTERN(NUGN NO₃)₂(15 mg), diluent: 0.2% HNO₃, charac-

teristic mass: 17.0 pg, sensitivity: 20.0 mgL^ꢀ1, LOD calculated to 3 mgL^ꢀ1

.

Isosteric heats of adsorption: oxygen adsorption isotherms were recorded

in a magnetic suspension balance (Rubotherm, Germany)^[89]that could

be operated up to 50 MPa. A scheme of the balance, including tempera-

ture and pressure sensors, and various gas supplies is shown in Figure S1

in the Supporting Information. Two different pressure transducers (New-

port Omega, US) were used in a range from vacuum up to 5 MPa with

an accuracy of 0.05%. In a typical experiment, a stainless-steel sample

holder was filled with about 0.1–0.2 g of MFU-1 or MFU-2, respectively.

The balance was evacuated for 12 h at 573 K and 0.3 Pa until constant

mass was achieved. For measuring the sorption capacity, the gas was

dosed into the balance chamber to elevated pressures. Equilibrium was

achieved within 2 h, which is characterised by constant weight and pres-

sure over a time range of approximately 15 min. The temperature was

kept constant with an accuracy of ꢁ0.1 K for each measurement. Addi-

tionally, for each isotherm, a buoyancy correction was used to calculate

the surface excess mass from the measured values. A detailed description

of this procedure can be found elsewhere.^[90]For the determination of the

density of oxygen from pressure and temperature, the program FLUID-

Thermogravimetric analysis (TGA) was performed with

a TGA/

SDTA851 Mettler Toledo analyzer in the temperature range 30–8008C in

a flow of nitrogen at a heating rate of 108Cmin^ꢀ1. Two powder X-ray dif-

fraction methods were used to characterise MFU-1 and MFU-2. Varia-

ble-temperature X-ray powder diffraction (VTXRPD) patterns were

measured by using a diffractometer equipped with an Anton Paar HTK

1200N heating chamber (q-q mode, Cu_Karadiation, l=1.5406 ꢄ) under

static air. X-ray powder diffraction data for phase analysis were collected

on a PANalytical XꢅPert PRO diffractometer (q-q mode, Cu_Karadiation,

l=1.5406 ꢄ). Powder diffraction patterns were scanned over the angular

range 2q=5.0–50.08 in steps of 0.03348, measured for 250 s per step (for

all samples except microwave-synthesised MFU-1 at 10 s per step). UV/

Chem. Eur. J. 2011, 17, 8671 – 8695

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

8691

D. Volkmer et al.

CAL was used.^[91]Temperature programmed oxidation (TPO) and pulse

chemisorption were performed on a Rubotherm/Bel Japan, BELCAT-B,

instrument equipped with a quadrupolar mass spectrometer with a ther-

mal conductivity detector (TCD) sensor. The TPO experiments were car-

ried out in the temperature range up to 2508C for MFU-1 and 3008C for

MFU-2 in 1000 ppm O₂in He. All experiments were monitored with a

quadrupolar mass spectrometer (GAM 400, InProcessInstruments, Ger-

many).

Catalytic oxidation of cyclohexene by employing tert-butyl hydroperox-

ide: MFU-1 or MFU-2 (0.095 mmol based on cobalt centres) was added

to a solution of cyclohexene (16 mmol), tBuOOH (8 mmol) and 1,2,4-tri-

chlorobenzene (2 mmol; internal standard) at RT. The reaction mixture

was stirred at 708C. Aliquots of about 50 mL were removed after time in-

tervals indicated in the main text. Each sample was diluted with CH₂Cl₂

(1 mL) and filtered through a 0.25 mm Acrodisc nylon filter. Then, the

sample was analysed by gas chromatography. For investigations on cata-

lyst activities in subsequent multiple runs, the catalysts were separated

from the reaction mixture by centrifugation and rinsed twice with CH₂Cl₂

before reuse. To investigate metal leaching from the catalyst, the hot re-

action mixture was filtered at the reaction temperature when approxi-

mately 50% of the final conversion achieved in a previous run was

reached (after 2 h in the case of MFU-1 and after 4.5 h in the case of

MFU-2). To investigate the influence of crystal size and catalyst loading

on reaction rates, different amounts of MFU-1 from solvothermal or mi-

crowave syntheses were reacted as described above. The initial rates

were determined at low substrate conversion, after reaction times of 20–

40 min. MFU-1 catalyst (0.1 mmol) was reacted with cyclohexene

(16 mmol) and tert-butyl hydroperoxide (8 mmol). The reaction was fil-

tered hot after 24 h, evaporated to dryness, the residue was dissolved in

0.2% HNO₃and afterwards analysed by AAS. The solution contained

less than 3 mgL^ꢀ1Co, showing that less than 1.1ꢃ10^ꢀ3% of the catalyst

decomposed (for full decomposition of the catalyst, the solution should

contain 330 mgL^ꢀ1of Co). All yields and conversions were based on cy-

clohexene. Conversions and turnover numbers (TONs) were calculated

from the equations: TON=100%ꢃ[sum of products]/[amount of Co],

conversion=100%ꢃ[sum of products]/[initial cyclohexene] for catalysis

employing tert-butyl hydroperoxide (since educt and tert-butanol show

the same retention times), otherwise conversion=100%ꢃ[consumed

educt]/[initial educt].

Solvothermal synthesis of MFU-1: H₂-bdpb (10 mg, 0.037 mmol) was dis-

solved in hot (100–1208C) DMF (4 mL) with the aid of an ultrasonic

bath. After cooling to RT, CoCl₂·6H₂O (40 mg, 0.168 mmol) was dis-

solved in the resulting solution and a 1m aqueous solution of HCl

(0.01 mL, 0.01 mmol) was added. Then, the solution was placed in a heat-

ing tube (10 mL) that had been previously treated with a solution of

(CH₃)₂SiCl₂in chloroform (5% by weight). The heating tube was sealed

and heated at a constant rate of 0.28Cmin^ꢀ1to 1208C and kept at this

temperature for 4 d. After hot filtration, MFU-1 was washed three times

with DMF (20 mL), CH₂Cl₂(10 mL), and suspended afterwards three

times in CH₂Cl₂(3 mL) for 24 h (followed by centrifugation and decanta-

tion each time), and was subsequently dried in vacuum for 6 h at RT

prior to further investigations to yield blue cubic crystals of phase-pure

MFU-1 (9.0 mg, 8.6 mmol, 70% based on ligand H₂-bdpb). IR (KBr):

3395 (br), 2924 (m), 2855 (w), 1679 (m), 1573 (m), 1493 (m), 1429 (s),

1374 (m), 1332 (m), 1286 (m), 1216 (w), 1108 (m), 1049 (s), 1013 (s), 981

(w), 846 (m), 754 (s), 665 (w), 617 (w), 582 (w), 524 (m), 490 cm^ꢀ1(s); el-

emental analysis calcd (%) for Co₄O

ACHTUNGTREN(GUN C₁₆H₁₆N₄)·2CAHTUNGTRENN(GUN C₃H₇NO): C 54.44, H

5.25, N 16.47, found: C 51.60, H 5.87, N 16.10.

Bigger single crystals of MFU-1 that are more suitable for single-crystal

X-ray structure analysis (edge lengths of 50–80 mm) were obtained when

H₂-bdpb (15 mg, 0.056 mmol) was dissolved in hot DMF as described

above and CoACHTUNGTRENNUNG(NO₃)₂·6H₂O (100 mg, 0.34 mmol) was added to the solu-

Catalytic oxidation of hydrocarbons with MFU-1 by employing molecular

oxygen: NHPI (0.31 mmol), MFU-1 (0.038 mmol based on cobalt cen-

tres) and 1,2,4-trichlorobenzene (1.2 mmol; internal standard) were

added to a solution of a hydrocarbon test substrate (3.0 mmol, either cy-

clohexene (300 mL), ethylbenzene (365 mL), cyclohexanol (270 mg) or cy-

clohexane (300 mL)) in acetonitrile (5 mL), and the reaction mixture was

stirred at 358C for 24 h under atmospheric oxygen by using a reflux con-

denser to avoid the escape of volatile components. Aliquots of about

50 mL were removed after time intervals indicated in the main text. Each

sample was diluted with CH₂Cl₂(1 mL) and filtered through a 0.25 mm

Acrodisc nylon filter. Then, the sample was analysed by gas chromatogra-

phy. To investigate metal leaching from the catalyst, the hot reaction mix-

ture was filtered at a reaction temperature when approximately 50% of

the final conversion achieved in a previous run was reached (after 16 h).

tion when it had been cooled to RT. The solution then was heated imme-

diately to 1208C, at which temperature it was kept for 3 d. This proce-

dure yielded MFU-1-enriched material, which was contaminated with

MFU-2.

Microwave-assisted solvothermal synthesis of MFU-1: H₂-bdpb (96 mg,

0.36 mmol) was dissolved in hot DMF (16 mL, 100–1208C) aided by an

ultrasonic bath. After being cooled to RT, CoACTHNUGTRNEUNG(NO₃)₂·6H₂O (138 mg,

0.47 mmol) was dissolved in the resulting solution and an aqueous solu-

tion of 1m HCl (0.01 mL, 0.01 mmol) was added. Then, the solution was

placed in a Pyrex sample tube (35 mL). The tube was sealed and heated

for 2 min to 1508C by employing a microwave synthesiser (CEM Discov-

er S) at 300 W, after which time, phase-pure MFU-1 (113 mg, 0.11 mmol,

90% based on H₂-bdpb) was formed after being worked up as described

above.

For the immobilization of NHPI, MFU-1 (0.038 mmol based on cobalt

centres) was suspended for 3 d in a solution of NHPI (0.31 mmol) in ace-

tonitrile (5 mL; same stochiometry as in catalysis reaction). Afterwards,

the modified MFU-1 (NHPI@MFU-1) was filtered, washed with acetoni-

trile and dried in vacuum. The modified catalyst NHPI@MFU-1 was used

in the catalysis reaction suspended in cyclohexene (300 mL, 3.0 mmol; as

the educt), together with 1,2,4-trichlorobenzene (150 mL, 1.2 mmol; inter-

nal standard) and tert-butyl hydroperoxide (0.15 mmol; initiator to pre-

vent the otherwise observed initiation time). The mixture was stirred

under atmospheric oxygen at 358C by using a reflux condenser to avoid

the escape of volatile components.

Solvothermal synthesis of MFU-2: H₂-bdpb (15 mg, 0.056 mmol) was dis-

solved in hot DMF (4 mL, 100–1208C) aided by an ultrasonic bath. After

being cooled to RT, CoACTHNUTRGNEUNG(NO₃)₂·6H₂O (100 mg, 0.344 mmol) was added

and dissolved, and the resulting solution was placed in a heating tube

(10 mL). The heating tube was sealed and heated at a constant rate of

0.158Cmin^ꢀ1to 1208C. After 10 h, the purple product was filtered when

still hot and washed three times with DMF (20 mL) and CH₂Cl₂(10 mL).

Afterwards, MFU-2 was suspended three times in CH₂Cl₂(3 mL) for 24 h

(followed by centrifugation and decantation each time), and was subse-

quently dried in vacuum for 6 h at RT prior to further investigations to

give MFU-2 (15.5 mg, 0.039 mmol, 70% based on H₂-bdpb). IR (KBr):

3432 (br), 3068 (m), 3026 (m), 2926 (m), 1574 (s), 1493 (m), 1427 (s),

1377 (m), 1319 (m), 1285 (m), 1123 (m), 1047 (s), 1013 (s), 984 (w), 847

(m), 789 (w), 677 (w), 614 (w), 581 (m), 487 (m), 437 cm^ꢀ1(w); elemental

Crystal structure determination of MFU-1 and MFU-2: Intensity data for

MFU-1 and MFU-2 were collected at 120 K on a Oxford Diffraction, Su-

perNova A diffractometer (4-circle kappa goniometer, dual wavelength

(Mo,Cu) microfocus X-ray sources, 135 mm CCD detector) using Mo_Ka

radiation (l=0.7107 ꢄ) for MFU-1 and Cu_Karadiation (l=1.5418 ꢄ) for

MFU-2. Selected crystal, measurement and refinement data are given in

Table 1. The data were corrected for Lorentz and polarisation effects and

a cylindrical absorption correction^[92]was applied for MFU-2. Programs

used for data collection: Bruker XSCANS,^[93]data reduction: Bruker

SHELXTL,^[94]structure solution: SHELXS-97,^[95]structure refinement:

SHELXL-97,^[96]molecular graphics: Diamond V3.1.,^[97]Mercury V1.5.^[98]

Structures were solved by using direct methods and refined by full-matrix

analysis calcd (%) for Co

found: C 57.7, H 5.5, N 17.2.

Microwave-assisted solvothermal synthesis of MFU-2: In a Pyrex sample

tube (35 mL), H₂-bdpb (96 mg, 0.40 mmol) and Co(NO₃)₂·6H₂O (104 mg,

ACHTUNGTRNE(GUN C₁₆H₁₆N₄)·ACHUTGNTREN(NUGN C₃H₇NO): C 57.6, H 5.9, N 17.7;

AHCTUNGTRENNUNG

0.36 mmol) were suspended at RT in DMF (1 mL). The tube was sealed

and heated for 8 min at 1208C by a microwave synthesiser (CEM Discov-

er S) at 300 W, filtered when still hot and washed with hot DMF (3ꢃ

20 mL) and CH₂Cl₂(10 mL) to obtain phase-pure MFU-2 (142 mg,

0.36 mmol, 90% based on H₂-bdpb).

8692

www.chemeurj.org

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2011, 17, 8671 – 8695

Pyrazolate-Based Cobalt(II)-Containing Frameworks

FULL PAPER

2

least squares on jFj . Disordered atoms were calculated isotropic in case

[3] O. M. Yaghi, H. Li, M. Eddaoudi, M. OꢅKeeffe, Nature 1999 , 402 ,

27 6 –2 79.

[4] Y. K. Hwang, D.-Y. Hong, J.-S. Chang, H. Seo, M. Yoon, J. Kim,

S. H. Jhung, C. Serre, G. Fꢇrey, Appl. Catal. A 2009 , 358 , 24 9 –2 53.

[5] S.-H. Cho, B. Ma, S. T. Nguyen, J. T. Hupp, T. E. Albrecht-Schmitt,

Chem. Commun. 2006 , 256 3 –2 565.

of non-positive U_eq; all others with anisotropic displacement parameters.

Hydrogen atoms were placed in geometrically calculated positions. See

the Supporting Information for details. CCDC-796315 (MFU-1) and

796314 (MFU-2) contain the supplementary crystallographic data for this

paper. These data can be obtained free of charge from The Cambridge

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

[6] M. Eddaoudi, J. Kim, N. Rosi, D. Vodak, J. Wachter, M. OꢅKeefe,

O. M. Yaghi, Science 2002 , 295 , 46 9 –4 72.

DFT calculations: DFT calculations with periodic boundary conditions

were carried out by using the gradient corrected Perdew–Burke–Ern-

zerhofer (PBE)^[99]functional employing the VASP^[100,101]program pack-

age. Dispersion interactions were added semi-empirically^[102]following

the Grimme^[103]scheme. Core electrons were described within the projec-

tor augmented wave scheme. An [Ar] core was used for Co, whereas the

4s and 3d electrons were treated explicitly. For C, N and O, the 2s and 2p

electrons were treated explicitly. The valence electrons were described by

a plane wave basis set with an energy cutoff of 800 eV along with the ap-

propriate PAW potentials. The core radii for Co, O, C, N and H were

2.30, 1.52, 1.50, 1.50 and 1.10 au, respectively.

[7] a) S. Biswas, M. Grzywa, H. P. Nayek, S. Dehnen, I. Senkovska, S.

Kaskel, D. Volkmer, Dalton Trans. 2009 , 648 7 –6 495; b) M. Dinca,

J. R. Long, J. Am. Chem. Soc. 2005, 127, 9376–9377; c) T. K. Maji,

G. Mostafa, H. Ch. Chang, S. Kitagawa, Chem. Commun. 2005 ,

243 6 –2 438; d) L. Pan, B. Parker, X. Huang, D. H. Olson, J. Y. Lee,

J. Li, J. Am. Chem. Soc. 2006 , 128 , 418 0 –4 181.

[8] a) F. Stallmach, S. Grçger, V. Kꢁnzel, J. Kꢀrger, O. M. Yaghi, M.

Hesse, U. Mꢁller, Angew. Chem. 2006 , 118 , 217 7 –2 181; b) S. Amir-

jalayer, M. Tafipolsky, R. Schmid, Angew. Chem. 2007 , 119 , 46 7 –

470.

[9] S. Hermes, M.-K. Schrçter, R. Schmid, L. Khodeir, M. Muhler, A.

Tissler, R. W. Fischer, R. A. Fischer, Angew. Chem. 2005 , 117 ,

639 4 –6 397; Angew. Chem. Int. Ed. 2005 , 44 , 623 7 –6 241.

[10] a) M. S. El-Shall, V. Abdelsayed, A. E. R. S. Khder, H. M. A.

Hassan, H. M. El-Kaderi, T. E. Reich, J. Mater. Chem. 2009 , 19 ,

762 5 –7 631; b) M. Sabo, A. Henschel, H. Frçde, E. Klemm, S.

Kaskel, J. Mater. Chem. 2007 , 17 , 382 7 –3 832.

[11] Y. K. Hwang, D.-Y. Hong, J.-S. Chang, S. H. Jhung, Y.-K. Seo, J.

Kim, A. Vimont, M. Daturi, C. Serre, G. Fꢇrey, Angew. Chem.

2008 , 120 , 421 2 –4 216; Angew. Chem. Int. Ed. 2008 , 47 , 414 4 –4 148.

[12] H.-L. Jiang, B. Liu, T. Akita, M. Haruta, H. Sakurai, Q. Xu, J. Am.

Chem. Soc. 2009 , 131 , 1130 2 –1 1303.

As a first step, the cell parameters of MFU-1 were computed by applying

symmetry restrictions with respect to the T point group. The fractional

coordinates were relaxed until the forces were smaller than 0.005 eVꢄ^ꢀ1

.

The MFU-1 was considered to be in the high spin state, that is, 12 un-

paired electrons (3 on each Co atom).

For the dioxygen adsorption on MFU-1, calculations on periodic struc-

tures were done by using an energy cutoff of 400 eV. In these calcula-

tions, the cell parameters were kept constant and only the ionic positions

were allowed to relax. The adsorption energies were calculated by DE=

E

N

U

tributions at 298 K. The vibrational frequencies were calculated numeri-

cally by displacing each ion by a step of 0.015 ꢄ in all three cartesian di-

rections. A more detailed computational study is in preparation.

[13] M. Mꢁller, S. Hermes, K. Kꢀhler, M. W. E. van den Berg, M.

Muhler, R. A. Fischer, Chem. Mater. 2008, 20, 4576–4587.

[14] F. Schrçder, S. Henke, X. Zhang, R. A. Fischer, Eur. J. Inorg.

Chem. 2009 , 313 1 –3 140.

Details on the crystallographic data, experimental methods and equip-

ment used for investigations are given in the Supporting Information.

[15] K. S. Suslick, P. Bhyrappa, J.-H. Chou, M. E. Kosal, S. Nakagaki,

D. W. Smithenry, S. R. Wilson, Acc. Chem. Res. 2005 , 38 , 28 3 –2 91.

[16] C. D. Wu, A. Hu, L. Zhang, W. Lin, J. Am. Chem. Soc. 2005 , 127 ,

894 0 –8 941.

Acknowledgements

[17] a) L. Alaerts, E. Seguin, H. Poelman, F. Thibault-Starzyk, P. A.

Jacobs, D. E. De Vos, Chem. Eur. J. 2006 , 12 , 735 3 –7 363; b) K.

Schlichte, T. Kratzke, S. Kaskel, Microporous Mesoporous Mater.

2004 , 73 , 8 1 –8 8; c) A. Vimont, J. M. Goupil, J. C. Lavalley, M.

Daturi, S. Surble, C. Serre, F. Millange, G. Fꢇrey, N. Audebrand, J.

Am. Chem. Soc. 2006 , 128 , 321 8 –3 227; d) P. Horcajada, S. Surble,

C. Serre, D. Y. Hong, Y. K. Seo, J. S. Chang, J. M. Greneche, I. Mar-

giolaki, G. Fꢇrey, Chem. Commun. 2007 , 282 0 –2 822.

[18] B. Xiao, H. Hou, Y. Fan, J. Organomet. Chem. 2007 , 692 , 201 4 –

2020.

This work was supported by the German Research Foundation (DFG)

within Priority Program 1362. We want to thank BESSY for the granted

synchrotron beamtime, especially Prof. Erko at BESSY for fruitful dis-

cussions as well as scientific and technical support. Measurement of

single-crystal X-ray diffraction data by Oxford Diffraction (now part of

Varian), is highly appreciated, in particular, immense support from Dr.

Oliver Presly. Dr. Volker Hagen (Rubokat) is acknowledged for his sup-

port on the interpretation and evaluation of TPO and isosteric heat of

adsorption. We are grateful to Prof. Armando J. L. Pombeiro for suggest-

ing to us the usage of NHPI/PINO. The calculations were carried out at

the HLRN supercomputer centre. M.T. is grateful to the Landesgraduier-

tenfçrderung Baden-Wꢁrttemberg for financial support.

[19] A. Dhakshinamoorthy, M. Alvaro, H. Garcia, J. Catal. 2009 , 267 ,

1–4.

[20] F. X. Llabrꢇs i Xamena, O. Casanova, R. G. Tailleur, H. Garcia, A.

Corma, J. Catal. 2008, 255, 220–227.

[21] K. S. Park, Z. Ni, A. P. Cote, J. Y. Choi, R. D. Huang, F. J. Uribe-

Romo, H. K. Chae, M. OꢅKeeffe, O. M. Yaghi, Proc. Natl. Acad.

Sci. USA 2006 , 103 , 1018 6 –1 0191.

[22] The van der Waals radius of C (0.17 nm) was used to calculate the

width of the apertures: A. Bondi, J. Phys. Chem. 1964 , 68 , 44 1 –

451.

[23] a) S. A. Moggach, T. D. Bennett, A. K. Cheetham, Angew. Chem.

2009 , 121 , 722 1 –7 223; b) G. Fꢇrey, C. Serre, Chem. Soc. Rev. 2009,

38, 1380–1399.

[24] E. L. Wu, G. R. Landolt, A. W. Chester, Stud. Surf. Sci. Catal. 1986 ,

28 , 54 7 –5 54.

[1] For recent reviews on different aspects of MOF applications, see:

a) L. J. Murray, M. Dinca, J. R. Long, Chem. Soc. Rev. 2009 , 38 ,

129 4 –1 314; b) J. R. Li, R. J. Kuppler, H.-C. Zhou, Chem. Soc. Rev.

2009 , 38 , 147 7 –1 504; c) A. Corma, H. Garcꢆa, F. X. Llabrꢇs i

Xamena, Chem. Rev. 2010 , 110 , 460 6 –4 655; d) J. Y. Lee, O. K.

Farha, J. Roberts, K. A. Scheidt, S. T. Nguyen, Joseph T. Hupp,

Chem. Soc. Rev. 2009 , 38 , 145 0 –1 459; e) D. J. Tranchemontagne,

J. L. Mendoza-Cortꢇs, M. OꢅKeeffe, O. M. Yaghi, Chem. Soc. Rev.

2009 , 38 , 125 7 –1 283; f) M. Eddaoudi, D. B. Moler, H. Li, B. Chen,

T. M. Reineke, M. OꢅKeeffe, O. M. Yaghi, Acc. Chem. Res. 2001 ,

34 , 31 9 –3 30; g) G. Fꢇrey, C. Mellot-Draznieks, C. Serre, F. Mill-

ange, Acc. Chem. Res. 2005, 38, 217–225; h) U. Mꢁller, M. Schu-

bert, F. Teich, H. Puetter, K. Schierle-Arndt, J. Pastrꢇ, J. Mater.

Chem. 2006, 16, 626–636.

[25] A. F. Hollemann, E. Wiberg, Lehrbuch der Anorganischen Chemie,

Vol. 101, de Gruyter, Berlin, 1995, p. 1220.

[26] H. Irving, R. J. P. Williams, J. Chem. Soc. 1953 , 319 2 –3 210.

[27] ACHTUNTRGNEUNG(m₄-Oxo)hexakis(m₂-3,5-dimethylpyrazolato-N,N’)tetracobalt(II)

[2] A. U. Czaja, N. Trukhan, U. Mꢁller, Chem. Soc. Rev. 2009 , 38 ,

128 4 –1 293.

(CSD ID: LESNUS): M. K. Ehlert, S. J. Rettig, A. Storr, R. C.

Thompson, J. Trotter, Acta Crystallogr. C 1994 , 50 , 102 3 –1 026.

Chem. Eur. J. 2011, 17, 8671 – 8695

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

8693

D. Volkmer et al.

[28] Catena

[bis(m₂-pyrazolato)cobalt(II)] (CSD-ID: AQOPEB): N.

[52] a) A. B. P. Lever, Inorganic Electronic Spectroscopy, Elsevier, Am-

sterdam, 1968, Chapter 9, p. 323; b) L. Poul, N. Jouini, F. Fiꢇvet,

Chem. Mater. 2000 , 12 , 312 3 –3 132.

[53] a) K. Nomiya, R. Kobayashi, M. Miwa, Bull. Chem. Soc. Jpn. 1983 ,

56 , 227 2 –2 275; b) J. H. Ashley, P. C. H. Mitchell, J. Chem. Soc. A

1968 , 282 1 –2 827.

[54] D. T. Shay, G. P. A. Yap, L. N. Zakhaov, A. L. Rheingold, K. H.

Theopold, Angew. Chem. 2005 , 117 , 153 2 –1 534; Angew. Chem. Int.

Ed. 2005 , 44 , 150 8 –1 510.

[55] R. A. Sheldon, M. Wallau, I. W. C. E. Arends, U. Schuchard, Acc.

Chem. Res. 1998 , 31 , 48 5 –4 93.

Masciocchi, G. A. Ardizzoia, S. Brenna, G. LaMonica, A. Maspero,

S. Galli, A. Sironi, Inorg. Chem. 2002 , 41 , 608 0 –6 089.

[29] F. Ramirez, S. B. Bhatia, A. V. Patwardhan, C. P. Smith, J. Org.

Chem. 1967 , 32 , 354 7 –3 553.

[30] MFU is the acronym for Metal–Organic Framework Ulm Universi-

ty.

[31] M. Tonigold, Y. Lu, B. Bredenkçtter, B. Rieger, S. Bahnmꢁller, J.

Hitzbleck, G. Langstein, D. Volkmer, Angew. Chem. 2009 , 121 ,

768 2 –7 687; Angew. Chem. Int. Ed. 2009 , 48 , 754 6 –7 550.

[32] Hexakis(m₂-acetato-O,O’)ACHTUNTRGNE(UNG m₄-oxo)tetrazinc (CSD-ID: ZNOXAC):

H. Koyama, Y. Saito, Bull. Chem. Soc. Jpn. 1954 , 27 , 11 2 –1 14.

[33] Similar results were recently reported on the microwave-assisted

synthesis of a range of zinc terephthalate based MOFs: Z. Ni, R. I.

Masel, J. Am. Chem. Soc. 2006 , 128 , 1239 4 –1 2395.

[56] a) B. Ravel, M. Newville, J. Synchrotron Radiat. 2005 , 12 , 53 7 –5 41;

b) M. Newville, J. Synchrotron Radiat. 2001 , 8 , 32 2 –3 24; c) B.

Ravel, J. Synchrotron Radiat. 2001 , 8 , 31 4 –3 16; d) J. J. Rehr, R. C.

Albers, Rev. Mod. Phys. 2000 , 72 , 62 1 –6 54.

[34] A. L. Spek, J. Appl. Crystallogr. 2003 , 36 , 7 –1 3.

[57] The expected distances for tetrahedral Co ions in the case of Co^II

[35] For DMF, a molecular volume of 128 ꢄ³was calculated from the

molar mass (73.1 gmol^ꢀ1) and the density at room temperature

(0.95 gcm^ꢀ3). The void volume found by PLATON/SQUEEZE was

divided by this value to obtain the approximate number of solvent

molecules per elemental cell.

[36] L. Hou, Y.-Y. Lin, X.-M. Chen, Inorg. Chem. 2008 , 47 , 134 6 –1 351.

[37] W. Ouellette, A. V. Prosvirin, K. Whitenack, K. R. Dunbar, J. Zu-

bieta, Angew. Chem. 2009 , 121 , 217 4 –2 177; Angew. Chem. Int. Ed.

2009 , 48 , 214 0 –2 143.

[38] The cobalt(II) ions resemble the tetrahedrally coordinated S atoms

in PtS, whereas the organic linker (with its four nitrogen atoms as

points of extension) displays a rectangular SBU that resembles the

square planar Pt atoms in PtS.

[39] a) Y. Lu, M. Tonigold, B. Bredenkçtter, D. Volkmer, J. Hitzbleck,

G. Langstein, Z. Anorg. Allg. Chem. 2008 , 634 , 241 1 –2 417; b) H. J.

Choi, M. Dinca, J. R. Long, J. Am. Chem. Soc. 2008 , 130 , 784 8 –

7850.

[40] a) Y. Shimura, R. Tsuchida, Bull. Chem. Soc. Jap. 1957 , 30 , 50 2 –

505; b) M. Ali, S. K. Saha, P. Banerjee, J. Chem. Soc. Dalton Trans.

1991 , 230 5 –2 309.

[41] J. Moellmer, E. B. Celer, R. Luebke, A. J. Cairns, R. Staudt, M. Ed-

daoudi, M. Thommes, Microporous Mesoporous Mater. 2010 , 129 ,

34 5 –3 53.

[42] F. Rouquerol, J. Rouquerol, K. Sing, Adsorption by Powders and

Porous Solids: Principles, Methodology and Applications, Academ-

ic Press, London, 1999.

[43] S. Ma, H.-C. Zhou, J. Am. Chem. Soc. 2006 , 128 , 1173 4 –1 1735.

[44] a) H. A. Goodwin, Top. Curr. Chem. 2004, 234, 23–47; b) T.

Nowlin, K. Cohn, Inorg. Chem. 1972 , 11 , 56 0 –5 63; c) L. Sacconi,

M. Ciampolini, G. P. Speroni, Inorg. Chem. 1965 , 4 , 111 6 –1 119;

d) M. Ciampolini, N. Nardi, Inorg. Chem. 1966 , 5 , 4 1 –4 4; e) M. J.

Norgett, L. M. Venanzi, Inorg. Chim. Acta 1968 , 2 , 10 7 –1 10; f) F.

Mani, C. Mealli, Inorg. Chim. Acta 1981 , 54 , L77L79.

are 1.93(6) and 2.00(7) ꢄ for Co O and Co N, respectively and in

ꢀ

the case of Co^IIIare 1.81(3) and 1.89(5) ꢄ for Co O and Co N, re-

spectively. CCDC codes for four-coordinate (tetrahedral) Co^II:

a) V. Divjakovic, V. Leovac, Cryst. Struct. Commun. 1978, 7, 689–

692 (CSD entry AAMTCO); b) P. E. Kruger, N. Martin, M. Nieu-

wenhuyzen, J. Chem. Soc. Dalton Trans. 2001 , 196 6 –1 970 (CSD

entry ABODUQ); c) P. G. Lacroix, F. Averseng, I. Malfant, K. Na-

katani, Inorg. Chim. Acta 2004 , 357 , 382 5 –3 835 (CSD entry

ABUWEA); d) J. Du, L. Han, Y. Cui, J. Li, Y. Li, W.-H. Sun, Aust.

J. Chem. 2003 , 56 , 70 3 –7 06 (CSD entry AJEBEW); e) U. Schroder,

L. Beyer, R. Richter, J. Angulo-Cornejo, M. Castillo-Montoya, M.

Lino-Pacheco, Inorg. Chim. Acta 2003 , 353 , 5 9 –6 7 (CSD entry

AJILAG); f) L.-J. Zhang, J.-Q. Xu, Z. Shi, X.-L. Zhao, T.-G. Wang,

J. Solid State Chem. 2003 , 173 , 3 2 –3 9 (CSD entry AKAREJ);

g) S.-M. Peng, Y.-N. Lin, Acta Crystallogr. C 1986 , 42 , 172 5 –1 731

(CSD entry FASHOW); h) F. A. Cotton, L. M. Daniels, L. R. Fal-

vello, J. H. Matonic, C. A. Murillo, X. Wang, H. Zhou, Inorg.

Chim. Acta 1997 , 266 , 9 1 –1 02 (CSD entry NEKXAC). Four-coor-

dinate Co^III: i) X. Dai, P. Kapoor, T. H. Warren, J. Am. Chem. Soc.

2004 , 126 , 479 8 –4 799 (CSD entries ABATIH, ABAVAB); j) C. A.

Jimenez, J. B. Belmar, J. Alderete, F. S. Delgado, M. Lopez-Rodri-

guez, O. Pena, M. Julve, C. Ruiz-Perez, Dalton Trans. 2007, 2135–

2144 (CSD entry CIDVIV); k) S. W. Gordon-Wylie, B. L. Claus,

C. P. Horwitz, Y. Leychkis, J. M. Workman, A. J. Marzec, G. R.

Clark, C. E. F. Rickard, B. J. Conklin, S. Sellers, G. T. Yee, T. J. Col-

lins, Chem. Eur. J. 1998 , 4 , 217 3 –2 181 (CSD entry FAJTAL); l) J.

Chakraborty, R. K. B. Singh, B. Samanta, C. R. Choudhury, S. K.

Dey, P. Talukder, M. J. Borah, S. Mitra, Z. Naturforsch. B 2006, 61,

1209–1216 (CSD entry VESMEM).

ꢀ

[58] O. Haas, R. P. W. J. Struis, J. M. McBreen, J. Solid State Chem.

2004 , 177 , 100 0 –1 010.

[59] a) A. L. Roe, D. J. Schneider, R. J. Mayer, J. W. Pyrz, J. Windom,

L. Que, J. Am. Chem. Soc. 1984 , 106 , 167 6 –1 678; b) C. R. Randall,

L. Shu, Y.-M. Chiou, K. S. Hagen, M. Ito, N. Kitajima, R. J. Lachi-

cotte, Y. Zang, L. Que, Inorg. Chem. 1995 , 34 , 103 6 –1 039; c) V.

Briois, C. Cartier dit Moulin, M. Momenteau, P. Maillard, J. Zar-

embowitch, E. Dartyge, A. Fontaine, G. Tourillon, P. Thuꢇry, M.

Verdaguer, J. Chim. Phys. 1989, 86, 1623–1634.

[45] a) B. L. Vallee, R. J. P. Williams, Proc. Natl. Acad. Sci. USA 1968 ,

59 , 49 8 –5 05; b) M. W. G. De Bolster, Pure Appl. Chem. 1997 , 69 ,

125 1 –1 303.

[46] A. Mavrandonakis, J. Sauer, unpublished results.

[47] S. Ye, F. Neese, Curr. Opin. Chem. Biol. 2009 , 13 , 8 9 –9 8.

[48] a) T. Dꢁren, F. Millange, G. Fꢇrey, K. S. Walton, R. Q. Snurr, J.

Phys. Chem. C 2007, 111, 15350–15356; b) K. S. Walton, R. Q.

Snurr, J. Am. Chem. Soc. 2007 , 129 , 855 2 –8 556.

[49] The error estimate was based on the weighing error.

[50] The reported surface areas for MOF-5 substantially vary in the

range 600–3400 m²g^ꢀ1; see, for example: a) B. Panella, M. Hirscher,

Adv. Mater. 2005 , 17 , 53 8 –5 41; b) Y. Li, R. T. Yang, J. Am. Chem.

Soc. 2006 , 128 , 72 6 –7 27; c) Y. Li, R. T. Yang, J. Am. Chem. Soc.

2006 , 128 , 813 6 –8 137; d) B. Panella, M. Hirscher, H. Pꢁtter, U.

Mꢁller, Adv. Funct. Mater. 2006 , 16 , 52 0 –5 24; e) L. Huang, H.

Wang, J. Chen, Z. Wang, J. Sun, D. Zhao, Y. Yan, Microporous

Mesoporous Mater. 2003 , 58 , 10 5 –1 14.

[60] a) L. Barbey, N. Nguyen, V. Caignaert, F. Studer, B. Raveau, J.

Solid State Chem. 1994 , 112 , 148.

[61] a) M. Croft, D. Sills, M. Greenblatt, C. Lee, S.-W. Cheong, K. V.

Ramanujachary, D. Tran, Phys. Rev. B 1997 , 55 , 8726; b) M. C.

Sꢈnchez, J. Garcꢆa, J. Blasco, G. Subꢆas, J. Pꢇrez-Cacho, Phys. Rev.

B 2002, 65, 144409; c) M. C. Sꢈnchez, G. Subꢆas, J. Garcꢆa, J.

Blasco, Phys. Rev. B 2004, 69, 184415; d) A. H. de Vries, L. Hozoi,

R. Broer, Int. J. Quantum Chem. 2003 , 91 , 57.

[62] [Co(pz)₃] (pz=pyrazolate) contains Co^IIIions in an octahedral co-

ordination environment: see ref. [28].

[63] a) C. Giannotti, C. Fontaine, A. Chiaroni, C. Riche, J. Organomet.

Chem. 1976 , 113 , 5 7 –6 5; b) A. Nishinaga, H. Tomita, K. Nishizawa,

T. Matsuura, S. Ooi, K. Hirotsu, Chem. Soc. Dalton Trans. 1981 ,

150 4 –1 514; c) L. Saussine, E. Brazi, A. Robine, H. Mimoun, J.

Fischer, R. Weiss, J. Am. Chem. Soc. 1985 , 107 , 353 4 –3 540; d) F. A.

[51] J. Hafizovic, M. Bjørgen, U. Olsbye, P. D. C. Dietzel, S. Bordiga, C.

Prestipino, C. Lamberti, K. P. Lillerud, J. Am. Chem. Soc. 2007 ,

129 , 361 2 –3 620.

8694

www.chemeurj.org

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2011, 17, 8671 – 8695

Pyrazolate-Based Cobalt(II)-Containing Frameworks

FULL PAPER

Chavez, C. V. Nguyen, M. M. Olmstead, P. K. Mascharak, Inorg.

Chem. 1996 , 35 , 628 2 –6 291.

2008C only in the case of CO, which, according to XRD measure-

ments, was due to the formation of Co₃O₄. The liquid-phase oxida-

tion of cyclohexene did not show a measurable conversion after

72 h at 4 bar oxygen pressure at 508C in the presence of MFU-1.

[77] J. Piera, J.-E. Bꢀckvall, Angew. Chem. 2008 , 120 , 355 8 –3 576;

Angew. Chem. Int. Ed. 2008 , 47 , 350 6 –3 523.

[78] a) Y. Ishii, S. Sakaguchi, T. Iwahama, J. Org. Chem. 1996 , 61 , 452 0 –

4526; b) Y. Yoshino, Y. Hayashi, T. Iwahama, S. Sakaguchi, Y. Ishii,

J. Org. Chem. 1997 , 62 , 681 0 –6 813.

[79] E. Lemaire, A. Rassat, Tetrahedron Lett. 1964 , 5 , 224 5 –2 248.

[80] R. Amorati, M. Lucarini, V. Mugnaini, G. F. Peduli, F. Minisci, F.

Recupero, F. Fontana, P. Astolfi, L. Greci, J. Org. Chem. 2003 , 68 ,

174 7 –1 754.

[81] a) S. Coseri, Catal. Rev. 2009 , 51 , 21 8 –2 92; b) F. Recupero, C.

Punta, Chem. Rev. 2007 , 107 , 380 0 –3 842; c) Y. Ishii, s. Sakaguchi,

T. Iwahama, Adv. Synth. Catal. 2001 , 343 , 39 3 –4 27.

[82] L. Horner, W. Jurgleit, Justus Liebigs Ann. Chem. 1955 , 591 , 13 8 –

152.

[83] a) V. P. Kurkov, J. Z. Pasky, J. B. Lavigne, J. Am. Chem. Soc. 1968 ,

90 , 474 3 –4 744; b) R. A. Budnik, J. K. Kochi, J. Org. Chem. 1976 ,

41 , 138 4 –1 389.

[84] F. Rajabi, J. H. Clark, B. Karimi, D. J. Macquarrie, Org. Biomol.

Chem. 2005 , 3 , 72 5 –7 26.

[85] Reaction conditions: cyclohexene (300 mL), 1,2,4-trichlorobenzene

(150 mL, 1.2 mmol; internal standard), MFU-1 (15 mg), tert-butyl

hydroperoxide (15 mL, 1 mol%; initiator), atmospheric oxygen, T=

358C.

[86] G. W. Parshall, S. D. Ittel in Homogenous Catalysis: The Applica-

tions and Chemistry of Catalysis by Soluble Transition Metal Com-

plexes, 2nd ed., Wiley, New York, 1992, pp. 255–261.

[87] M. A. Warpehoski, B. Chabaud, K. B. Sharpless, J. Org. Chem.

1982 , 47 , 289 7 –2 900.

[88] a) A. Erko, I. Packe, C. Hellwig, M. Fieber-Erdmann, O. Pawlitzki,

M. Veldkamp, W. Gudat, AIP Conf. Proc. 2000, 521, 415–418;

b) A. Erko, I. Packe, W. Gudat, N. Abrosimov, A. Firsov, SPIE

2001, 4145, 122–128.

[89] F. Dreisbach, H. W. Lçsch, J. Therm. Anal. Calorim. 2000 , 62 , 51 5 –

521.

[90] J. U. Keller, R. Staudt, Gas Adsorption Equilibria: Experimental

Methods and Adsorption Isotherms, Springer, New York, 2004.

[91] a) R. Span, W. Wagner, Int. J. Thermophys. 2003 , 24 , 1 –3 9; b) R .

Span, W. Wagner, Int. J. Thermophys. 2003 , 24 , 4 1 –1 09; c) R. Span,

W. Wagner, Int. J. Thermophys. 2003 , 24 , 11 1 –1 62.

[92] C. W. Dwiggins, Jr., Acta Crystallogr. A 1975, 31, 146–148.

[93] Bruker, XSCANS, Version 2.2, Bruker Analytical X-ray Systems,

Madison, WI, 1996.

[64] a) S. Hikichi, H. Komatsuzaki, M. Akita, Y. Moro-oka, J. Am.

Chem. Soc. 1998 , 120 , 469 9 –4 710; b) S. Hikichi, M. Yoshizawa, Y.

Sasakura, H. Komatsuzaki, Y. Moro-oka, M. Akita, Chem. Eur. J.

2001 , 7 , 501 1 –5 028; c) O. M. Reinaud, G. P. A. Yap, A. L. Rhein-

gold, K. H. Theopold, Angew. Chem. 1995 , 107 , 217 1 –2 173;

Angew. Chem. Int. Ed. Engl. 1995 , 34 , 205 1 –2 052; d) S. Hikichi, H.

Komatsuzaki, N. Kitajima, M. Akita, M. Mukai, T. Kitagawa, Y.

Moro-oka, Inorg. Chem. 1997 , 36 , 26 6 –2 67; e) J. W. Egan Jr. , B. S.

Haggerty, A. L. Rheingold, S. C. Sendlinger, K. H. Theopold, J.

Am. Chem. Soc. 1990 , 112 , 244 5 –2 446; f) S. Hikichi, M. Yoshizawa,

Y. Sasakura, M. Akita, Y. Moro-oka, J. Am. Chem. Soc. 1998 , 120 ,

1056 7 –1 0568; g) C. J. Cramer, W. B. Tolman, K. H. Theopold,

A. L. Rheingold, Proc. Natl. Acad. Sci. USA 2003 , 100 , 363 5 –3 640.

[65] a) W. J. J. Stevens, V. Meynen, E. Bruijn, O. I. Lebedev, G. Van

Tendeloo, P. Cool, E. F. Vansant, Microporous Mesoporous Mater.

2008 , 110 , 7 7 –8 5; b) J. C. Groen, T. Bach, U. Ziese, A. M. Pau-

laime-van Donk, K. P. de Jong, J. A. Moulijn, J. Pꢇrez-Ramirez, J.

Am. Chem. Soc. 2005 , 127 , 1079 2 –1 0793; c) J. C. Groen, S. Abellꢉ,

L. A. Villaescusa, J. Pꢇrez-Ramirez, Microporous Mesoporous

Mater. 2008 , 114 , 9 3 –1 02; d) J. C. Groen, J. A. Moulijn, J. Pꢇrez-

Ramirez, J. Mater. Chem. 2006 , 16 , 212 1 –2 131; e) J. C. Groen, J. A.

Moulijn, J. Pꢇrez-Ramirez, J. Ind. Eng. Chem. Res. 2007 , 46 , 419 3 –

4201.

[66] a) F. A. Chavez, J. M. Rowland, M. M. Olmstead, P. K. Mascharak,

J. Am. Chem. Soc. 1998 , 120 , 901 5 –9 027; b) F. A. Chavez, J. A.

Briones, M. M. Olmstead, P. K. Mascharak, Inorg. Chem. 1999 , 38 ,

160 3 –1 608; c) R. A. Sheldon, J. K. Kochi, Metal-Catalyzed Oxida-

tions of Organic Compounds, Academic Press, New York, 1981;

d) A. Bravo, H.-R. Bjørsvik, F. Fontana, L. Liguori, F. Minesci, J.

Org. Chem. 1997 , 62 , 384 9 –3 857; e) S. J. Blanksby, G. B. Ellison,

Acc. Chem. Res. 2003 , 36 , 25 5 –2 63; f) S. P. Heneghan, S. W.

Benson, Int. J. Chem. Kinet. 1983 , 15 , 81 5 –8 22.

[67] a) T. R. Rizzo, B. D. Cannon, E. S. McGinley, F. F. Crim, Laser

Chem. 1983 , 2 , 32 1 –3 33; b) A. C. Baldwin in The Chemistry of

Functional Groups, Peroxides (Ed.: S. Patai), Wiley, New York,

1983, p. 279; c) G. Dayma, P. A. Glaude, R. Fournet, F. Battin-Le-

clerc, Int. J. Chem. Kinet. 2003 , 35 , 27 3 –2 85; d) D. de B. Darwent,

Bond Dissociation Energies in Simple Molecules, NBSDS-NBS, 31,

GPO, Washington, DC, 1970.

[68] Measured at 298 K in benzene: D. V. Avila, K. U. Ingold, J. Lusz-

tyk, W. A. Green, D. R. Procopio, J. Am. Chem. Soc. 1995 , 117 ,

292 9 –2 930.

[69] a) M. V. Encina, M. Rivera, E. A. Lissi, J. Polym. Sci. Polym.

Chem. Ed. 1978 , 16 , 170 9 –1 717; b) H. Paul, R. D. Small, J. C.

Scaiano, J. Am. Chem. Soc. 1978 , 100 , 452 0 –4 527; c) M. V. Encinas,

J. C. Scaiano, J. Am. Chem. Soc. 1981 , 103 , 639 3 –6 397; d) J. H. B.

Chenier, S. B. Tong, J. A. Howard, Can. J. Chem. 1978 , 56 , 304 7 –

3053.

[70] S. E. Bottle, W. K. Busfield, I. D. Jenkins, J. Chem. Soc. Perkin

Trans. 2 1992 , 214 5 –2 150.

[71] W. K. Busfield, I. D. Grice, I. D. Jenkins, S. H. Tang, Aust. J. Chem.

1991 , 44 , 140 7 –1 415.

[72] T. W. Campbell, G. M. Coppinger, J. Am. Chem. Soc. 1951 , 73 ,

178 8 –1 789.

[73] a) I. W. C. E. Arends, K. U. Ingold, D. D. M. Wayner, J. Am. Chem.

Soc. 1995 , 117 , 471 0 –4 711; b) H. Fischer, J. Am. Chem. Soc. 1986 ,

108 , 392 5 –3 927; c) B. E. Daikh, R. G. Finke, J. Am. Chem. Soc.

1992 , 114 , 293 8 –2 943.

[74] L. Salem, Electrons in Chemical Reactions: First Principles, Wiley,

New York, 1982.

[75] G. Henrici-Olivꢇ, S. Olivꢇ, Angew. Chem. 1974, 86, 1–12; Angew.

Chem. Int. Ed. Engl. 1974, 13, 29–38.

[94] SHELXTLV6.12 2000, Bruker Analytical X-Ray Systems, Madi-

son, WI.

[95] G. M. Sheldrick, Acta Crystallogr. A 1990 , 46 , 46 7 –4 73.

[96] G. M. Sheldrick, SHELXL-97, Program for X-ray Crystal Structure

Refinement, University of Gçttingen, Gçttingen (Germany), 1997.

[97] Diamond–Crystal and Molecular Structure Visualization, Crystal

Impact, K. Brandenburg & H. Putz GbR, Bonn (Germany).

[98] Mercury: visualization and analysis of crystal structures, C. F.

Macrae, P. R. Edgington, P. McCabe, E. Pidcock, G. P. Shields, R.

Taylor, M. Towler, J. van de Streek, J. Appl. Crystallogr. 2006 , 39 ,

45 3 –4 57.

[99] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996 , 77 ,

3865.

[100] G. Kresse, J. Furthmꢁller, Comput. Mater. Sci. 1996 , 6 , 1 5 –5 0.

[101] G. Kresse, J. Furthmuller, Phys. Rev. B 1996 , 54 , 1116 9 –1 1186.

[102] T. Kerber, M. Sierka, J. Sauer, J. Comput. Chem. 2008 , 29 , 208 8 –

2097.

[103] S. Grimme, J. Comput. Chem. 2006 , 27 , 178 7 –1 799.

[76] No reaction was observed between molecular oxygen and ethene,

propene or CO as educts in gas-phase reactions in the presence of

MFU-1 at different reaction conditions (1–6 bar, flow rate 4–

20 mLmin^ꢀ1, 90–2008C). An exothermic reaction was observed at

Received: November 4, 2010

Revised: February 23, 2011

Published online: June 17, 2011

Chem. Eur. J. 2011, 17, 8671 – 8695

ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

8695

Article Doi

Pyrazolate-based cobalt(II)-containing metal-organic frameworks in heterogeneous catalytic oxidation reactions: Elucidating the role of entatic states for biomimetic oxidation processes

DOI: 10.1002/chem.201003173

Source and publish data:

Authors:

Article abstract of DOI:10.1002/chem.201003173

Full text of DOI:10.1002/chem.201003173

Products guided by the article

R&D Labs maybe for 108-94-1

Relevant to this article

Hot Product