Structural study of the reaction of methylzinc amino alcoholates with oxygen

Article abstract of DOI:10.1021/om100449t

Reaction of ZnMe₂ with 1,3-bis(dimethylamino)propan-2-ol (Hbdmap) in 2:1 ratio forms both [MeZn(bdmap)·ZnMe₂] ₂ (2) and [MeZn(bdmap)]₃·ZnMe₂ (3) depending on the concentration of the reaction. In the former, ZnMe₂ is coordinated to a free N-donor of the bdmap ligand and rather more loosely to the oxygen of the alkoxide. In 3, the ZnMe₂ is coordinated to two free N-donors of the bdmap ligand. 2 reacts with O₂ at low temperatures with controlled insertion into one of the Zn-C bonds of the coordinated ZnMe₂ group to form the peroxide [MeZn(bdmap)] ₂MeZnOOMe (4). 4 decomposes slowly, and the hydroxide [MeZn(bdmap)]₂MeZnOH (5) was isolated; in addition to 5, two other decomposition products have been unambiguously identified, namely, (MeZn) ₅(bdmap)₃O (6) and (MeZn)₄(bdmap) ₄ZnO (7). The formation of these species can be linked to reactions of the hydroxide (5), or its associated radical [MeZn(bdmap)] ₂MeZn(O^?)], with species such as ZnMe₂ or MeZn(bdmap), present is solution as a result of operating Schlenk equilibria. The structure of [MeZn(bdmap)]₄ (1) is also reported.

Full text of DOI:10.1021/om100449t

3318 Organometallics 2010, 29, 3318–3326
DOI: 10.1021/om100449t
Structural Study of the Reaction of Methylzinc Amino Alcoholates with
Oxygen
†
†
‡
†
€
Nathan Hollingsworth, Andrew L Johnson, Andrew Kingsley, Gabriele Kociok-Kohn,
and Kieran C Molloy*,†
†Department of Chemistry, University of Bath, Claverton Down, Bath, BA2 7AY, U.K. and
‡SAFC Hitech, Power Road, Bromborough, Wirral CH62 3QF, U.K.
Received May 10, 2010
Reaction of ZnMe2 with 1,3-bis(dimethylamino)propan-2-ol (Hbdmap) in 2:1 ratio forms both [MeZn-
(bdmap) ZnMe2]2 (2) and [MeZn(bdmap)]3 ZnMe2 (3) depending on the concentration of the reaction.
3
3
In the former, ZnMe2 is coordinated to a free N-donor of the bdmap ligand and rather more loosely to the
oxygen of the alkoxide. In 3, the ZnMe2 is coordinated to two free N-donors of the bdmap ligand. 2 reacts
with O2 at low temperatures with controlled insertion into one of the Zn-C bonds of the coordinated
ZnMe2 group to form the peroxide [MeZn(bdmap)]2MeZnOOMe (4). 4 decomposes slowly, and the
hydroxide [MeZn(bdmap)]2MeZnOH (5) was isolated; in addition to 5, two other decomposition pro-
ducts have been unambiguously identified, namely, (MeZn)5(bdmap)3O (6) and (MeZn)4(bdmap)4ZnO
(7). The formation of these species can be linked to reactions of the hydroxide (5), or its associated radical
[MeZn(bdmap)]2MeZn(O•)], with species such as ZnMe2 or MeZn(bdmap), present is solution as a result
of operating Schlenk equilibria. The structure of [MeZn(bdmap)]4 (1) is also reported.
Introduction
Reformatsky-type reactions.11 Indeed, many of these reactions
require in situ formation of a zinc peroxide in the presence of an
amino alcohol auxilary ligand,12 a system typified by that which
is the focus of this paper.
The reaction between ZnR2 (R = Me, Et) and oxygen has a
history dating back over 150 years since the early experiments of
Frankland.1 The products of this reaction have variously been
suggested as the alkoxides RZn(OR) and Zn(OR)2 or the per-
oxides RZn(OOR) and Zn(OOR)22-6 with insertion of O2 into
either one or both Zn-C bonds; more recent reports have
isolated cubanes [RZn(OR)]4 and also bis-heterocubanes
[(RZn)6Zn(OR)8] from these reactions, in which the presence
or absence of water is influential of the outcome.7 The reactions
are generally uncontrollably fast, and the interception of inter-
mediates along the reaction pathway has proved challenging.
The widely held belief concerning the mechanism of this reac-
tion is that it is radical in nature8 on the basis of the early work
of Davies, who demonstrated that the autoxidation of ZnEt2
was inhibited by a radical scavenger (galvinoxyl), though
not the first stage of the reaction.6 Furthermore, there are
numerous alkylzinc-mediated reactions involving ZnR2/air that
also clearly involve radicals,9 e.g., epoxidation of enones10 and
Alternative insights into mechanistic aspects of the ZnR2/
O2 reaction have appeared over the last six years that have
challenged the long-held assertion concerning a radical
mechanism,13,14 primarily through the work of Lewinski,
ꢀ
who was first to structurally authenticate organozinc per-
oxides arising from O2 insertion into a Zn-C bond. Thus,
reaction of either EtZn(BDI) or EtZn(azol) with O2 resulted
15
in the isolation of the peroxides [(EtOO)Zn(BDI)]2 and
[(EtOO)Zn(azol)]2[EtZn(azol)]2,16 respectively.
Further work has demonstrated the importance of an
equilibrium between four- and three-coordinate species i.e.,
R2Zn(L)2 h R2Zn(L) þ L (L = N- or O-donor) in mitigating
the rate of R-Zn/O2 insertion. So, while ZnBut2 reacts
*To whom correspondence should be addressed. E-mail: k.c.molloy@
bath.ac.uk.
(1) Seyferth, D. Organometallics 2001, 20, 2940.
(2) Frankland, E. J. Chem. Soc. 1862, 15, 363.
(3) Lissenko, A. Jahresber.-Inst. Hydrol., GSF-Forschungszent.
Umwelt Gesund. 1864, 470.
(11) Mileo, E.; Benfatti, F.; Cozzi, P. G.; Lucarini, M. Chem. Com-
mun. 2009, 469.
(12) Enders, D.; Zhu, J. J.; Raabe, G. Angew. Chem., Int. Ed. Engl.
(4) Demuth, R.; Meyer, V. Ber. Bunsen-Ges. Phys. Chem. 1890, 23, 394.
(5) Abraham, M. H. J. Chem. Soc. 1960, 4130.
(6) Davies, A. G.; Roberts, B. P. J. Chem. Soc. B 1968, 1074.
1996, 35, 1725.
ꢀ
(13) Lewinski, J.; Suwala, K.; Kubisiak, M.; Ochal, Z.; Justyniak, I.;
€
(7) Jana, S.; Berger, R. J. F.; Frohlich, R.; Pape, T.; Mitzel, N. W.
Lipkowski, J. Angew. Chem., Int. Ed. 2008, 47, 7888.
ꢀ
Inorg. Chem. 2007, 46, 4293.
(8) Encyclopaedia of Inorganic Chemistry; Grevy, J. M., Ed.; Wiley:
(14) Lewinski, J.; Sliwinski, W.; Dranka, M.; Justyniak, I.; Lipkowski,
ꢀ
J. Angew. Chem., Int. Ed. 2006, 45, 4826.
ꢀ
Chichester, 2005.
(15) Lewinski, J.; Ochal, Z.; Bojarski, E.; Tratkiewicz, E.; Justyniak,
(9) Akindele, T.; Yamada, K.-I.; Tomioka, K. Acc. Chem. Res. 2009,
42, 345.
(10) Porter, M. J.; Skidmore, J. Chem. Commun. 2000, 1215.
I.; Lipkowski, J. Angew. Chem., Int. Ed. 2003, 42, 4643.
(16) Lewinski, J.; Marciniak, W.; Lipkowski, J.; Justyniak, I. J. Am.
Chem. Soc. 2003, 125, 12698.
ꢀ
r
pubs.acs.org/Organometallics
Published on Web 07/02/2010
2010 American Chemical Society

Article
Organometallics, Vol. 29, No. 15, 2010 3319
rapidly with O2 at -78 °C in THF to afford [(THF)tBuZn-
OBut]2, the presence of methylpyridine (py-Me) requires
reaction at -45 °C, and the peroxide [(py-Me)tBuZn(OO-
But)]2 results.14 The implications of this appear to be that the
donor determines both the rate of formation of a three-coordi-
nate intermediate by ligand dissociation (from a nominally four-
coordinate precursor), and it is this coordinatively unsaturated
center that is the reactive site; Zn-C bonds at a four-coordinate
zinc are inactive to O2 insertion. Presumably, the amino alcohol
azole plays a similar role in the reaction of EtZn(azol) with O2,
which results in a product containing Zn-Et and ZnOOEt
bonds in equal proportion.16 These assertions have been sup-
-0.82 (s, 3H, CH3Zn). 13C NMR: 66.4 (CO), 66.1 (CH2N), 64.4
and 53.9 (OOCH3), 45.8 (CH3N), -18.3 (CH3Zn).
In Situ NMR Experiment. Dimethylzinc (0.2 mL, 2 M solution
in hexane, 0.4 mmol) was added to d8-toluene (0.6 mL) in a
Young’s NMR tube. Hbdmap (0.035 mL, 0.2 mmol) was then
added to this mixture in a nitrogen-filled glovebox. 1H and 13
C
NMR spectra obtained from this sample were identical to those
of pure 2. The headspace of the NMR tube was then filled with
dry O2 for 30 s, after which the O2 was replaced with N2. 1H and
13C NMR carried out on this sample were in agreement with that
1
of 4 (albeit the MeZn region of the H NMR spectrum was
greater by 6H due to the unreacted Me2Zn unit of 2). The sample
was then heated to 358 K over a 20 min period, kept at this
temperature for 15 min, then cooled to 298 K, and the spectra
were recorded down to 218 K to assess the decomposition of the
peroxide (4).
ported in two further reports by Lewınski, in which N,N-
´
chelating ligands (tBu-DAB, Pyr-pyr) have retarded the reaction
of ZnR2 with O2 to afford examples of both tri- and tetrazinc
clusters containing two ZnOOR functions.13,17
Crystallography. Experimental details relating to the single-
crystal X-ray crystallographic studies are summarized in Table 1.
For all structures, data were collected on a Nonius Kappa
CCD diffractometer at 150(2) K using Mo KR radiation (λ =
0.71073 A). Structure solution was followed by full-matrix least-
squares refinement and was performed using the WinGX-1.70
suite of programs.19 Corrections for absorption were made in all
cases. Compound 1 crystallizes along with a molecule of hexane.
For 3 there is half a molecule of toluene to every Zn complex,
which was refined isotropically. Compound 5 crystallizes with
half a molecule of toluene in the asymmetric unit, located about
a center of inversion. In this structure, H(3), bonded to O(3), was
located in the difference Fourier map and refined; however it
was restrained to an ideal bond length. Furthermore, the ligand
attached to both Zn(1) and Zn(3) in 5 was found to be disordered
in the ratio 90:10; atoms with a lower occupation factor were
refined isotropically. Compound 6 crystallizes with a molecule
of toluene in its lattice and includes a disordered bdmap ligand,
in which the CH2 and NMe2 carbons are split 70:30. For 7, there
is one molecule of toluene in the asymmetric unit. There is also
disorder in the ratio 55:45 in one of the bdmap ligands; as a
result, the bond length N(3)-C(13) was restrained.
In this respect, the validity of interpreting the inhibition of
the ZnR2/O2 reaction by radical scavengers as evidence for a
radical mechanism has been questioned,16 as these reagents
could be influencing the course of the reaction by metal
coordination, hence altering the coordination number.
Our own interest in the chemistry of zinc amino alcoho-
lates has, serendipitously, given us an entry into this area of
chemistry. We now wish to report a series of structural and
NMR studies that relate to products from the reaction of
ZnMe2 with 1,3-bis(dimethylamino)propan-2-ol (Hbdmap)
and that shed further light on the formation and evolution of
organozinc peroxides.
Experimental Section
General Procedures. Elemental analyses were performed
using an Exeter Analytical CE 440 analyzer. 1H and 13C
NMR spectra were recorded on a Bruker Advance 500 MHz
FT-NMR spectrometer as saturated solutions at room tempera-
ture, unless stated otherwise; chemical shifts are in ppm with
respect to Me4Si, and coupling constants are in hertz.
All reactions were carried out under an inert atmosphere
using standard Schlenk techniques. Solvents were dried and
degassed under an argon atmosphere over activated alumina
columns using an Innovative Technology solvent purification
system (SPS). Hbdmap18 was prepared by literature methods.
Syntheses. Synthesis of [MeZnbdmap.Me2Zn]2 (2). Dimethyl-
zinc (1 mL 2.0 M solution in hexane, 2.0 mmol) was added to
Hbdmap (0.175 mL, 1.0 mmol) at room temperature, whereby an
exothermic reaction took place and methane was evolved. On
standing, colorless cube-like crystals appeared (1.09 g, 85%, mp
76 °C). Anal. Found (calc for C20H52N4O2Zn4): C 37.7 (37.4), H
Results and Discussion
We have previously reported on the reaction of ZnR2 with
the amino alcohols Hdmae, Hbdmap and Htdmap.20
1
8.17 (8.16), N 8.58 (8.72). NMR (d8 toluene 298 K), H NMR:
Reactions of both Hdmae and Htdmap proceed smoothly to
give either tetrameric [RZn(dmae)]4 or dimeric [RZn(tdmap)]2
(R=Me, Et), respectively, as stable crystalline solids. In con-
trast, reactions with Hbdmap proved less straightforward.20 A
1:1 reaction of ZnMe2 with Hbdmap (4 mmol each reagent/
total 20 mL toluene) resulted in only partial alkane elimina-
tion, unlike reactions with Hdmae or Htdmap, in which 1 equiv
of CH4 was liberated in both instances. When an excess of
Hbdmap (2.2 equiv) was used, an oil analyzing as [MeZn-
(bdmap)] was finally isolated though NMR spectra did not
allow any definitive structural information to be determined,
while mass spectra indicated the presence of Zn2-Zn7 species
present in the oil, on the basis of isotope distribution patterns.
An aged sample of this oil has now yielded crystals of the
3.72 (m, 2H, OCH), 2.07 (dd, 4H, J = 9.0, 12.9 Hz, CH2N), 1.93
(s, 24H, CH3N), 1.75 (dd, 4H, J = 2.3, 12.9 Hz, CH2N), -0.82 (bs,
18H, CH3Zn). 13C NMR: 66.8 (CO), 65.0 (CH2N), 45.3 (CH3N),
-12.2 (CH3Zn).
Synthesis of [MeZnbdmap]2MeZnOOMe (4). Compound 2
(0.5 g, 0.78 mmol) was dissolved in hexane (1 mL) and stirred in
an atmosphere of dry O2 for 30 s, after which time the O2
atmosphere was replaced with N2. The solution was placed in a
freezer overnight, whereby crystals of 4 formed (0.38 g, 84%, mp
101 °C). Anal. Found (calc for C18H46N4O4Zn3): C 37.2 (37.4),
H 7.84 (8.01), N 9.66 (9.68). 1H NMR (d8 toluene 298 K): 3.61
(m, 2H, OCH), 3.51 and 3.44 (2 singlets, 3H, OOMe), 2.06 (m,
4H, CH2N), 1.89 (bs, 12H, CH3N), 1.85 (bs, 12H, CH3N), 1.63
(m, 4H, CH2N), -0.74 (s, 3H, CH3Zn), -0.78 (s, 3H, CH3Zn),
ꢀ
(17) Lewinski, J.; Suwala, K.; Kaczorowski, T.; Galezowski, M.;
Gryko, D. T.; Justyniak, I.; Lipkowski, J. Chem. Commun. 2009, 215.
(18) Campbell, K. N.; LaForge, R. A.; Campbell, B. K. J. Org. Chem.
1949, 14, 346.
(19) Farrugia, L. J. J. Appl. Crystallogr. 1999, 32, 837.
(20) Johnson, A. L.; Hollingsworth, N.; Molloy, K. C.; Kociok-
€
Kohn, G. Inorg. Chem. 2008, 47, 12040.

3320 Organometallics, Vol. 29, No. 15, 2010
Table 1. Crystal Data and Structure Refinement for 1-7
Hollingsworth et al.
1
2
3
4
5
6
7
formula
C
35H87-
N8O4Zn4
C20H52
N4O2Zn4
642.14
triclinic
P1
8.0653(3)
9.5037(4)
10.5151(5)
93.155(2)
112.003(2)
100.315(2)
728.58(5)
1
-
C59H140
N12O6Zn8
1636.79
monoclinic
P21/n
21.5525(3)
8.3720(1)
24.0648(4)
-
C18H46
N4O4Zn3
578.70
monoclinic
P21/c
8.2358(1)
20.3000(3)
16.5998(3)
-
C41H93
N8O6Zn6
1186.45
triclinic
P1
9.0591(2)
13.1668(3)
13.4498(3)
73.491(1)
72.678(1)
74.866(1)
1440.84(6)
1
-
C59H140
N12O8Zn10
1799.73
monoclinic
P21/n
17.9285(2)
9.4279(1)
25.3635(3)
-
C39H88
N8O5Zn5
1076.02
orthorhombic
Pbn21
9.8564(1)
20.8362(2)
25.8918(3)
-
fw
945.61
monoclinic
C2/c
18.0120(4)
15.0540(4)
36.528(1)
cryst syst
space group
a (A)
b (A)
c (A)
R (deg)
β (deg)
γ (deg)
V (A3)
Z
97.679(1)
102.068(1)
97.066(1)
99.119(1)
9815.8(4)
8
1.280
1.971
4040
4246.23(11)
2
1.280
2.264
1732
2754.19(7)
4
1.396
2.618
1216
4232.96(8)
2
1.412
2.828
1884
5317.4(1)
4
1.344
2.266
2272
F
calc (mg/m3)
1.464
3.273
1.367
2.501
μ(Mo KR) (mm-1
F(000)
cryst size (mm)
)
336
0.40 ꢀ 0.25
623
0.20 ꢀ 0.20
0.30 ꢀ 0.20
0.20 ꢀ 0.20
0.30 ꢀ 0.15
0.25 ꢀ 0.15
0.23 ꢀ 0.20
ꢀ 0.03
ꢀ 0.08
ꢀ 0.20
ꢀ 0.10
ꢀ 0.10
ꢀ 0.15
ꢀ 0.20
θ range
2.93-27.49
8171 [0.0810]
4876
0.9432, 0.5893
1.075
0.1064,0.2546
3.24-27.53
3338 [0.0705]
2672
3.75-27.46
9621 [0.0836]
6453
3.08-27.49
6294 [0.0997]
4411
3.21-27.51.
6582 [0.0532]
5359
3.04-30.02
12 330 [0.0560]
8659
3.59-27.49
11 000 [0.0717]
8728
indep reflns [R(int)]
reflns obsd (>2σ)
max., min. transmn
goodness-of-fit
final R1,
0.7719, 0.3508
1.073
0.0424, 0.1014
0.629, 0.629
1.014
0.0425, 0.0957
0.7720, 0.5022
1.015
0.0423, 0.0949
0.7880, 0.6346
1.061
0.0369, 0.0847
0.6560, 0.5221
1.021
0.0393, 0.0721
0.6600, 0.6238
1.026
0.0385, 0.0740
wR2 [I > 2σ(I )]
R1, wR2 (all data)
largest diff peak,
hole (e A3)
0.1739, 0.2878
3.210, -0.749
0.0579, 0.1092
0.792, -0.708
0.0797, 0.1093
0.587, -0.777
0.0757, 0.1085
1.523, -0.789
0.0515, 0.0922
0.790, -0.686
0.0733, 0.0821
0.500, -0.554
0.0615, 0.0815
0.598, -0.351
Flack param
-0.003(10)
long period, suggesting Schlenk-type equilibria are operating:20
2RZnðbdmapÞ h ZnR2 þ ZnðbdmapÞ2
For comparison, the corresponding cadmium species, RCd-
(bdmap), are trimeric solids.21 Furthermore, when synthe-
sizing a related cadmium compound with 1,3-bis(tert-butyl-
amino)-2-(tert-butylaminomethyl)propan-2-ol (Httbap),
[MeCd(ttbap)], a dimeric species with a molecule of CdMe2
coordinated to the pendant N-donors, i.e., [MeCd(ttbap)
3
CdMe2]2,22 was isolated, which caused us to question why we
had never seen analogous coordination chemistry among the
organozinc amino alcoholates. As a consequence of these ob-
servations, we decided to look more closely at the ZnMe2/
Hbdmap reaction, initially with regard to its concentration
dependence.
When ZnMe2 and a sample of Hbdmap that had been
handled in air were reacted in a 1:1 ratio but in a more concen-
trated manner (2 mmol each reagent/7 mL toluene) than in our
previous work, the resulting solution yielded three distinct
crystals, though resolution of the sample into analytically pure
samples proved impossible. Crystallographic analysis of se-
lected crystals showed that the reaction had resulted in a mixture
Figure 1. Asymmetric unit of 1 showing the labeling scheme used;
thermal ellipsoids are at the 40% probability level. Selected
metrical data (A and deg): Zn(1)-C(1) 2.000(10), Zn(1)-O(1)
2.005(7), Zn(1)-O(4) 1.985(7), Zn(1)-N(1) 2.207(9); C(1)-
Zn(1)-O(1) 124.7(5), C(1)-Zn(1)-O(4) 124.0(5), C(1)-Zn(1)-
N(1) 112.5(5), O(1)-Zn(1)-O(4) 102.8(3), O(1)-Zn(1)-N(1)
84.5(3), O(4)-Zn(1)-N(1) 98.7(4) .
comprising the adduct [MeZn(bdmap) ZnMe2]2 (2), the per-
3
oxide [MeZn(bdmap)]2MeZnOOMe (4), and a product likely to
be a result of peroxide decomposition, [(MeZn)5(bdmap)3O]
(6); the formation of the peroxide most probably arises from the
presence of oxygen in the Hbdmap sample used. In light of this
serendipitous observation, we have looked more carefully at the
course of this reaction.
tetramer [MeZn(bdmap)]4 (1) (Figure 1), which adopts essen-
tially the same open cubane structure as [MeZn(dmae)]4,20 in
which the additional coordination by the donor amine has
weakened some of the intermolecular O:fZn interactions.
While a 1:1 reaction of Hbdmap with ZnEt2 does give [EtZn-
(bdmap)], this is also an oil consisting of species of undeter-
First, using a degassed sample of Hbdmap and 2 equiv. of
2.0 M ZnMe2 in hexane with no additional solvent added
€
(21) Johnson, A. L.; Hollingsworth, N.; Kociok-Kohn, G.; Molloy,
K. C. Inorg. Chem. 2008, 47, 9706.
(22) Hollingsworth, N.; Molloy, K. C. Unpublished results.
mined nuclearity. Crystals of [Zn(bdmap)2 Hbdmap]2 and
3
[EtZn3(bdmap)5] crystallize from these respective oils over a

Article
Organometallics, Vol. 29, No. 15, 2010 3321
Scheme 1
afforded the adduct [MeZn(bdmap) ZnMe2]2 (2) in 85%
angles at Zn(2) are also very disparate, with the C(2)-Zn-
(2)-C(3) [145.75(19)o] being particularly open. The structure of
2 is of the form described by Yamakawa and Noyori as the key
intermediate in the addition of dialkylzinc to aldehydes, pro-
posed on the basis of ab initio MO calculations.23
3
yield. When the reaction solution is diluted (2 mmol ZnMe2
in 20 mL toluene) but retaining a 2:1 reaction stoichiometry
(ZnMe2:Hbdmap), the crystalline product again contains 2,
but a different adduct, [MeZn(bdmap)]3. ZnMe2 (3), is also
3
present. We have been unable to synthesize the latter adduct
uniquely using a 4:3 reaction stoichiometry; if the solution is
too concentrated, only 2 forms, while on significant dilution
to avoid this problem, nothing crystallizes from solution
(Scheme 1; N = NMe2):
The 1H NMR spectrum of 2 shows a single signal for the
MeZn and Me2Zn moieties, with a total integral of nine
hydrogens (Figure 4a). At 238 K, this signal resolves itself
so that both MeZn (3H) and Me2Zn (6H) are discernible
(Figure 4a, inset). The pattern of NMR coupling constants for
the CH2 protons of the bdmap ligand (two dd with J values
rationalizable using the Karplus equation and the angles
observed in the X-ray structure) is consistent with the ligand
bridging between two metal centers, as we have previously
seen with [MeCd(bdmap)]3.21
Compound 2 contains a dimeric [MeZn(bdmap)]2 core
(Figure 2) in which one arm of the amino alcohol acts as a κ2-
O,N bidentate donor to afford a four-membered Zn2O2 ring;
each zinc has tetrahedral coordination similar to that seen in
[MeZn(tdmap)]2.20 The free arm of the amino alcohol, along
with oxygen of the Zn2O2 ring acts to coordinate a further
equivalent of ZnMe2 at each end of the dimer. The coordination
sphere of the coordinated ZnMe2 unit is a highly distorted
tetrahedron, including relatively strong coordination to nitro-
gen [Zn(2)-N(2) 2.181(3) A] along with a much longer inter-
action with the ring oxygen [Zn(2)-O(1) 2.773(2) A]. The bond
The adduct [MeZn(bdmap)]3 ZnMe2 (3), which forms along
3
with 2 in more dilute solutions, incorporates a six-membered
Zn3O3 ring but bonded across the ring [Zn(2)-O(3) 2.627(2) A]
to yield two fused four-membered Zn2O2 heterocycles
(23) Yamakawa, M.; Noyori, R. J. Am. Chem. Soc. 1995, 117, 6327.

3322 Organometallics, Vol. 29, No. 15, 2010
Hollingsworth et al.
Figure 2. Asymmetric unit of 2 showing the labeling scheme used;
thermal ellipsoids are at the 50% probability level. Selected geo-
metric data (A and deg): Zn(1)-O(10) 2.102(2), Zn(2)-O(1)
2.773(2), Zn(1)-N(10) 2.142(3), Zn(2)-N(2) 2.181(3); O(1)-
Zn(1)-O(10) 90.12(8), Zn(1)-O(1)-Zn(10) 89.88(8), C(3)-
Zn(2)-C(2) 145.75(19). Symmetry operation: 2-x, 1-y, 1-z.
Figure 4. 1H NMR spectra of (a) 2 at 298 K (inset at 238 K),
(b) freshly dissolved 4 at 298 and (inset) 218 K, and (c) 4 after
heating at 90 °C overnight, at 298 and (inset) 238 K.
Figure 3. Asymmetric unit of 3 showing the labeling scheme used;
thermal ellipsoids are at the 40% probability level. Selected geomet-
ric data (A and deg): Zn(1)-O(1) 1.940(2), Zn(1)-O(3) 2.061(2),
Zn(1)-N(3) 2.140(3), Zn(2)-O(1)1.989(2), Zn(2)-O(2) 2.008(2),
Zn(2)-O(3) 2.627(2), Zn(2)-N(1) 2.296(3), Zn(3)-O(2) 1.987(2),
Zn(3)-O(3) 1.990(2), Zn(3)-N(4) 2.232(3), Zn(4)-N(5) 2.230(4),
Zn(4)-N(6) 2.233(3); O(1)-Zn(1)-O(3) 84.63(8), O(1)-Zn(2)-
O(2) 116.39(9), O(2)-Zn(3)-O(3) 89.21(8), C(4)-Zn(4)-C(5)
139.37(19), Zn(1)-O(1)-Zn(2) 115.78(10), Zn(2)-O(2)-Zn(3)
109.35(9), Zn(1)-O(3)-Zn(3) 124.08(10).
Zn(2) on the other hand is five-coordinate (ZnCO3N) as a result
of the trans-ring Zn(2)-O(3) bond. There are three different
bonding modes for the bdmap ligand. That based on O(3) is
μ3,κ3-O,N,N, bridging Zn(1) and Zn(3) via oxygen supported
by two additional N:fZn bonds, along with a third, bridging
Zn(2)-O(3) interaction. The ligand incorporating O(1) is μ2,κ2-
O,N, with now only one N:fZn supporting the Zn-O-Zn
bridge. Finally, O(2) is simply μ2 with respect to one of the four-
membered Zn2O2 rings. This structural motif was one of our
original suggestions for the structure of [MeZn(bdmap)]n based
on NMR evidence, when the nuclearity of the oil was uncer-
tain;20 the isolation of tetrameric 1, however, confirms that a
complex mixture of oligomers is present in this oil.
In contrast to 2, the ZnMe2 adducted to the [MeZn(bdmap)]3
trimer in 3 utilizes the two pendant nitrogen atoms of bdmap
which is solely μ2 with respect to the Zn-O rings [O(2)]. The
coordination sphere of Zn(4) is more regular in terms of donor
bonds [Zn-N: 2.230(4), 2.233(3) A] than the ZnMe2 adduct in 2,
though the C-Zn-C angle remains open [139.37(19)o]; clearly
there are electronic rather than steric reasons for this behavior.
It is this coordination of ZnMe2 to an oligomer [MeZn-
(bdmap)]n, as evidenced in the structures of 2 and 3, that
we believe is the reason that 1:1 reactions of ZnMe2 and
(Figure 3). This structure is unusual, particularly in comparison
with those of 4 and 5 (below), each of which have conventional
Zn3O3 rings with trans-ring Zn-O interactions > 3.4 A. Such
open six-membered rings are known for organozinc alkoxides
but are themselves relatively rare,24 and the fused nature of 3
parallels that of the cadmium species [MeCd(bdmap)]3, which
we have reported recently.21 Here, however, the preference of
cadmium for coordination numbers greater than 4 makes the
structure more expected. Both Zn(1) and Zn(3) in 3 are four-
coordinated, with a ZnCO2N coordination sphere in which the
Zn-O bonds are generally shorter [1.940(2)-2.061(2) A] than
in 2 [2.034(2), 2.102(2) A], presumably due to steric congestion
in the smaller ring. Conversely, the donor N:fZn bonds in 3
tend to be weaker [2.140(3)-2.296 A] than in 2 [2.142(3) A].
(24) Hecht, E. Z. Anorg. Allg. Chem. 2000, 626, 2223.

Article
Organometallics, Vol. 29, No. 15, 2010 3323
Hbdmap yielded only partial CH4 elimination and required
an excess of alcohol to form [MeZn(bdmap)]n.20
When pure 2 is exposed to an atmosphere of dry O2, an inser-
tion reaction occurs to form the peroxide [MeZn(bdmap)]2-
MeZnOOMe (4) (Scheme 1). The IR spectra of both 4 and
Hbdmap are rich in the region 800-900 cm-1 in which ν(O-O)
might be expected, on the basis of published data for other zinc
peroxides (813 w,9 854 m,11 868 m cm-1 10). Bands at 896 s and
811 w cm-1 , which do not coincide with bands in the spectrum
of Hbdmap (876 m, 819 m cm-1 are the closest) are plausible
candidates for this stretch. If, after exposure to the gas, the
solution is kept at -20 °C to inhibit reaction, crystals of 4 can be
isolated. The insertion of O2 into one Zn-C bond can be seen
quite clearly in the 1H NMR spectra: addition of O2 to a solution
of 2 affords qualitatively the same 1H NMR spectrum as that of
a freshly prepared solution of pure 4 after isolation. The 298 K
1H NMR spectrum of 4 (Figure 4b) shows three unequal MeZn
singlets at -0.74, -0.77, and -0.82 ppm. In addition, a major
singlet at 3.51 ppm, which we assign to the methylperoxo
(MeOOZn) moiety, has appeared along with a weaker signal
at 3.44 ppm, which we ascribe to the evolution of the peroxide
into some other, so far unidentified, product(s), as it grows with
time at the expense of methylperoxide signal. The total integral
for the three MeZn signals and the two signals at ca. 3.5 ppm
(MeOO þ decomposition products) is 9:3, as expected for 4.
The room-temperature 1H NMR data for 4 become
resolved and reveal a more complex picture at low tempera-
ture (218 K; Figure 4b, insets). The region associated with
MeZn splits into what appears to be three 1:1:1 groups, the
most intense of which (-0.51, -0.61, -0.69 ppm) must arise
from 4, as these are of an intensity that correlates them with
the intense MeOO singlet at 3.48 ppm (shifted from 3.51 ppm
at 298 K; integration, sum of signals at -0.51, -0.61, -0.69
to that at 3.48 ppm = 9:3); each of the singlets within this trio
can be assigned to one of the three unique MeZn environ-
ments in 4. Also, the growing singlet at 3.44 ppm, which we
assign to either OMe or OOMe of a decomposition product-
(s) of 4, resolves at 218 K into two singlets (3.53, 3.44 ppm)
and a weaker, broader signal at 3.56 ppm. The latter inte-
grates to 1H with respect to the weakest set of three singlets
(-0.47, -0.56, -0.72 ppm), which have a relative intensity of
9H; that is, these signals are consistent with the hydroxide
[MeZn(bdmap)]2MeZnOH (5), which we have isolated and
confirmed crystallographically (see below). On the basis of
integrals of the MeZn signals we estimate the ratio 4:5:
unidentified third (OMe) species to be ca. 70:10:20%.
Unfortunately, we have so far been unable to confirm the fate
of the methoxy group, which we suggest must also arise from
cleavage of the ZnO-OMe bond (see below), but a plausible
suggestion is a methylzinc methoxide aggregate of some sort,
such as [MeZn(bdmap)]2MeZnOMe. However, the intermedi-
ate group of 1:1:1 signals (-0.58, -0.65, -0.80 ppm) have a
total relative intensity of 9H with respect to the two signals at
(3.53, 3.44 ppm), which each integrate to 1.5H, which therefore
implies a structure in which two similar but nonequivalent
methoxy groups are present, that is, more complex structurally
than a simple Zn3 species such as [MeZn(bdmap)]2MeZnOMe.
Compound 4 is related to the trimeric [MeZn(bdmap)]3 core
of 3, with one bdmap ligand replaced by the OOMe group
(Figure 5). However, in this case (and 5, below) the six-mem-
bered, non-planar Zn3O3 ring does not incorporate further
trans-ring Zn-O bonds, and each metal center remains four-
coordinate. Of the bdmap ligands that remain on going from 3
to 4, one is μ2,κ3-O,N,N and the other μ2,κ2-O,N, as in 3, while
Figure 5. Asymmetric unit of 4 showing the labeling scheme used;
thermal ellipsoids are at the 50% probability level. Selected geo-
metric data (A and deg): Zn(1)-O(1) 1.987(2), Zn(1)-O(2)
1.979(2), Zn(1)-N(1) 2.168(3), Zn(2)-O(2) 2.028(2), Zn(2)-O(3)
2.019(3), Zn(2)-N(3) 2.138(3), Zn(3)-O(1) 1.987(2), Zn(3)-
O(3) 1.996(2), Zn(3)-N(2) 2.168(3), O(3)-O(4) 1.471(4), O(4)-
C(18) 1.377(5); O(1)-Zn(1)-O(2) 98.85(10), O(2)-Zn(2)-O(3)
111.33(10), O(1)-Zn(3)-O(3) 92.10(10), Zn(1)-O(1)-Zn(3)
129.00(12), Zn(1)-O(2)-Zn(2) 123.46(12), Zn(3)-O(3)-Zn(2),
128.74(13), O(4)-O(3)-Zn(3) 115.8(2), O(4)-O(3)-Zn(2)
104.51(18).
the bond length data for Zn-O and Zn-N (Figure 4) are very
close to those for 3. The O-O separation in the peroxide
[1.471(4) A] is similar to data reported by others for zinc
peroxides [e.g., 1.483(2),16 1.451(2) A15].
The relationship between the dimeric precursor (2) and the
trizinc final peroxide product (4) can be seen in the scheme for
the reaction sequence (Scheme 1). In 2, the weak bond between
oxygen and one coordinated ZnMe2 is easily broken, genera-
ting a reactive, three-coordinate zinc center. Oxygen, either as
MeOO• in a radical mechanism or O2 in a concerted mechanism,
is then able to react at this coordinatively unsaturated metal
center. Coordination of the R-oxygen of the peroxide to a
neighboring zinc of the strained Zn2O2 ring allows a ring
opening to generate the six-membered Zn3O3 ring seen in 4.
Finally, it would seem reasonable that in the Zn3O3 ring in 4, O,
N chelation of ZnMe2 is more crowded than in the Zn2O2 ring in
2, and hence the second coordinated ZnMe2 in 2 is released as
the peroxide (4) is formed.
Allowing a solution of 2exposed to dry O2 to stand for several
days at room temperature (rather than -20 °C) results in
formation of 4 along with more extensive decomposition
products, a process that can be replicated by heating a freshly
prepared solution of 4 for 3 h at 85 °C (Scheme 1). This causes
1
partial decomposition of the peroxide (diminution of the H
NMR signal at 3.55 ppm) while the signal at 3.44 ppm, arguably
a methoxide decomposition product, grows. However, remark-
ably, at least some of the peroxide survives this treatment even
after heating overnight (Figure 4c). The plethora of MeZn and
MeO signals observable at low temperature (Figure 4c, inset)
implies that the decomposition of the peroxide follows a number
of routes; NMR signals due to [MeZn(bdmap)]2MeZnOH (5),
(MeZn)5(bdmap)3O (6), and (MeZn)4(bdmap)4ZnO (7) (see
below) are likely to be among them. While pure 4 seems more
stable than previously reported organozinc peroxides, the pre-
sence of ZnMe2 does seem to facilitate this decomposition. A
solution containing ZnMe2 and Hbdmap (2:1) (forming 2), with
O2 added, which we have shown forms 4 in situ but with some

3324 Organometallics, Vol. 29, No. 15, 2010
Hollingsworth et al.
residual ZnMe2 eliminated from the adduct, does, when heated
overnight at 85 °C, afford complete decomposition of the
peroxide, albeit accompanied by the formation of additional
insoluble, and hence unidentified, by-products.
As discussed above, from a solution of 2 exposed to O2 and
heated at 85 °C for 3 h we have isolated three products in
addition to the peroxide (4), though the reaction beyond the
formation of 4 is not clean (Figure 4c) and none of the three
species have been characterized other than crystallographi-
cally. Furthermore, no product contains an OMe (or new
1
OOMe) function, which is implied by the change in the H
NMR spectrum of 4 around 3.5 ppm with time (Figure 4b,c).
We suggest compound 5, [MeZn(bdmap)]2MeZnOH, is
formed from homolysis of the O-O bond in 4. This decom-
position route for main group metal peroxides has not been
widely observed,25,26 but is more common among transition
metal analogues where oxidation of the metal can afford MdO
species.27 Recent work on organozinc peroxides has, however,
argued convincingly for this mode of decomposition for zinc
peroxides17,28,29 and in one case the MeO• radical generated by
homoloysis of a ZnO-OMe bond trapped by judicious choice
of co-ligand.13 We assume that, in the absence of any other
available radical, [MeZn(bdmap)]2MeZnO• scavenges H• from
the reaction solvent. It has recently been suggested that RZnO•
may abstract H• from the departing RO• followed by a 1,2-H
shift in the resultant species,29 though in our case we have no
evidence for this through the formation of formaldehyde. Of
course, in discussing the formation of 5 we cannot exclude the
possibility that it is formed from hydrolysis involving adventi-
tious moisture. However, the structural similarity of 4 and 5,
the retention of all the Me-Zn bonds during the course of the
reaction (except where oxygenation has occurred), our isola-
tion of zinc-oxo species 6 and 7 from these reactions, and the
lack of any evidence for the formation of species of type
[MeZn(OR)]4 or (MeZn)6Zn(OR)8, which are commonly seen
when ZnMe2 reacts with O2 in the presence of water,7 are all
consistent with the suggested O-O homolysis route.
The structure of 5 (Figure 6) is directly analogous to that of
its precursor 4 in terms of ligation, with only minor changes
in geometric data. Zn(1), not involved in bonding to either
OOMe in 4 or OH in 5, is essentially unchanged in the length
of its Zn-O and Zn-N bonds. Both Zn(2) and Zn(3) are
more weakly coordinated to their respective nitrogen donor
atoms in 5, due to replacement of OOMe with a less elec-
tronegative OH group, diminishing the Lewis acidity of the
associated metal. In contrast, both Zn(2) and Zn(3) strengthen
their bond to oxygen in 5 vs 4 [Zn(2)-O(3): 1.975(2) vs
2.019(3) A; Zn(3)-O(3): 1.976(2) vs 1.996(2) A], while the
angle at O(3) is significantly narrower in the hydroxide
[123.91(10)o] than the peroxide [128.74(13)o].
Figure 6. Asymmetric unit of 5 showing the labeling scheme used;
thermal ellipsoids are at the 50% probability level. Selected geo-
metric data (A and deg): Zn(1)-O(1) 1.9862(18), Zn(1)-O(2)
1.9822(17), Zn(1)-N(1) 2.173(2), Zn(2)-O(2) 2.0232(17), Zn-
(2)-O(3) 1.975(2), Zn(2)-N(3) 2.175(2), Zn(3)-O(1) 1.9877(19),
Zn(3)-O(3) 1.976(2), Zn(3)-N(2) 2.192(2) A; O(1)-Zn(1)-
O(2) 96.77(8), O(2)-Zn(2)-O(3) 108.53(8), O(1)-Zn(3)-O(3)
96.98(8), Zn(1)-O(1)-Zn(3) 126.27(9), Zn(1)-O(2)-Zn(2)
123.48(9), Zn(3)-O(3)-Zn(2) 123.91(10), H(3)-O(3)-Zn(3)
115(3), H(3)-O(3)-Zn(2) 113(3).
both 4 and 5 distort to form two fused four-membered rings
in 6 (Figure 7), as also seen in 3. This is done by forming a
bond between Zn(2) and O(1) [2.0899(19) A], which in 4 and
5 is a long, non-bonding separation [4: 3.138; 5: 3.346 A]; the
two fused Zn2O2 rings in 6 have a structural parallel in the
cadmium trimer [MeCd(bdmap)]3.21 The bdmap ligand
based on O(1) remains μ2,κ3-O,N,N bridging Zn(1) and
Zn(3), as it does in 4 and 5, but the bdmap based on O(2)
uses the free N-donor present in the two precursors to
coordinate Zn(5) of the additional Zn2 fragment of 6,
becoming μ2,κ3-O,N,N in the process. The final bdmap,
which is part of the (MeZn)2(bdmap) fragment that has been
incorporated into 6, adopts a μ2,κ2-O,N coordination, leav-
ing one Me2N donor [N(6)] uncomplexed. The two halves of
the molecule are centered on O(3), originating from the
peroxide in 4 or hydroxide in 5, which has tetrahedral μ4-
coordination. The Zn4O motif is found in basic zinc salts
30
such as Zn4(O)(O2CMe)6 and has featured in the decom-
position products of other organozinc peroxides.17,28 All the
zinc atoms in 6 are four-coordinated (as in 4 and 5), except
Zn(1), which expands its coordination number to five by
interactions with two donor nitrogen atoms [N(1), N(3)].
Related to 6 is a second decomposition product, (MeZn)4-
Zn(bdmap)4O (7) (Figure 8). This differs from 6 in having one
zinc that is devoid of methyl groups and is a fusion of two six-
membered Zn3O3 rings along a common edge. Each methylzinc
is tetrahedral at the metal with ZnCO2N, while the more Lewis
acidic Zn(5) is five-coordinated (ZnO3N2 coordination). The
Zn3O3 ring involving Zn(1), Zn(2), and Zn(5) is structurally
analogous to that in the peroxide (4) or hydroxide (5), so that
O(5) can be equated with the R-oxygen of the peroxide. The
geometry about O(5) is trigonal planar and is a rare example of
a planar μ3-OZn3 unit, others being {[pyr-2,5-Ph2]Zn(THF)2
The second product associated with decomposition of the
peroxide is the Zn5 aggregate (MeZn)5(bdmap)3O (6), a solid
that melts at room temperature. Despite the structural
relationship between 6 and its precursor 5, there is a sig-
nificant difference. The six-membered Zn3O3 rings present in
(25) Kumar, S. S.; Singh, S.; Roesky, H. W. Inorg. Chem. 2005, 44,
1199.
(26) Li, H.; Song, H. B.; Duan, L.; Cui, C.-P.; Roesky, H. W. Inorg.
Chem. 2006, 45, 1912.
(27) DiPasquale, A. G.; Kaminsky, W.; Mayer, J. M. J. Am. Chem.
Soc. 2002, 124, 14534.
ꢀ
(28) Lewinski, J.; Bury, W.; Dutkiewicz; Maurin, M.; Justyniak, I.;
Lipkowski, J. Angew. Chem., Int. Ed. 2008, 47, 573.
(29) Lewinski, J.; Koscielski, M.; Suwala, K.; Justyniak, I. Angew.
Chem., Int. Ed. 2009, 48, 7017.
ꢀ
ꢀ
(30) Hiltunen, L.; Leskela, M.; Makela, M.; Ninisto, L. Acta Chem.
Scand. A 1987, 41, 548.

Article
Organometallics, Vol. 29, No. 15, 2010 3325
Figure 8. Asymmetric unit of 7 showing the labeling scheme used;
thermal ellipsoids are at the 40% probability level. Only the major
component of the disorder in the bdmap ligand based on O3 is
shown; the molecule of solvation (toluene) has also been omitted
for clarity. Selected geometric data (A and deg): Zn(1)-O(1)
1.952(3), Zn(1)-O(2) 1.988(3), Zn(1)-N(1) 2.184(4), Zn(2)-O(1)
2.063(3), Zn(2)-O(5) 1.896(3), Zn(2)-N(7) 2.240(3), Zn(3)-O(4)
2.051(3), Zn(3)-O(5) 1.899(3), Zn(3)-N(5) 2.222(4), Zn(4)-O(3)
1.996(3), Zn(4)-O(4) 1.944(3), Zn(4)-N(4) 2.06(3), Zn(5)-O(2)
2.095(3), Zn(5)-N(3) 2.161(4), Zn(5)-O(3) 2.112(3), Zn(5)-N(2)
2.195(4), Zn(5)-O(5) 1.903(3); O(1)-Zn(1)-O(2) 97.84(11),
O(1)-Zn(2)-O(5) 115.98(12), O(4)-Zn(3)-O(5) 119.60(12),
O(3)-Zn(4)-O(4) 94.55(12), O(2)-Zn(5)-O(3) 160.36(11),
O(2)-Zn(5)-O(5) 100.17(11), O(3)-Zn(5)-O(5) 99.43(12),
N(2)-Zn(5)-N(3) 121.77(16).
Figure 7. Asymmetric unit of 6 showing the labeling scheme used;
thermal ellipsoids are at the 50% probability level. Selected geo-
metric data (A and deg): Zn(1)-O(1) 2.1839(17), Zn(1)-O(2)
2.0189(18), Zn(1)-N(1) 2.309(2), Zn(1)-N(3) 2.434(2), Zn(2)-
O(1) 2.0899(18), Zn(2)-O(2) 2.0007(16), Zn(2)-O(3) 1.9712(15),
Zn(3)-O(1) 2.0848(16), Zn(3)-O(3) 1.9539(17), Zn(3)-N(2)
2.236(2), Zn(4)-O(3) 1.9678(16), Zn(4)-O(4) 2.0262(16),
Zn(4)-N(5) 2.202(2), Zn(5)-O(3) 2.0019(15), Zn(5)-O(4)
2.0755(17), Zn(5)-N(4) 2.184(2); O(1)-Zn(1)-O(2) 77.50(6),
N(1)-Zn(1)-N(3) 96.17(9), O(1)-Zn(2)-O(2) 80.13(7), O(1)-
Zn(2)-O(3) 87.31(7), O(2)-Zn(2)-O(3) 97.36(7), O(1)-Zn(3)-
O(3) 87.91(7), O(3)-Zn(4)-O(4) 88.83(7), O(3)-Zn(5)-O(4)
86.55(6), O(4)-Zn(5)-N(4) 91.50(8), O(3)-Zn(5)-N(4)
110.30(7), Zn(1)-O(1)-Zn(2) 94.35(7), Zn(1)-O(1)-Zn(3)
118.25(7), Zn(1)-O(2)-Zn(2) 102.51(8), Zn(2)-O(3)-Zn(3)
95.96(7), Zn(2)-O(3)-Zn(4) 114.68(8), Zn(2)-O(3)-Zn(5)
108.80(7), Zn(3)-O(3)-Zn(4) 131.50(8), Zn(3)-O(3)-Zn(5)
111.06(8), Zn(4)-O(3)-Zn(5) 94.20(7), Zn(4)-O(4)-Zn(5)
90.30(7).
part of 6, (MeZn)2(bdmap), is most simply rationalized by
reaction of [(MeZn)3(bdmap)2(O•)] with half of dimeric 2 in a
process that results in loss of Me•:
- Me
f
3
ðMeZnÞ3ðbdmapÞ2O• þ MeZnðbdmapÞ ZnMe
3
2
ðMeZnÞ5ðbdmapÞ3O ð6Þ
(O)Zn[pyr-2,5-Ph2 ]}2 (Hpyr-2,5-Ph2=2,5-diphenylpyrrole),31
[Zn3{Ph2PC(H)Py}4(O)],32 [Zn6(O)2(FDCA)5(H2O)(DMF)]
(FDCA=1,10-ferrocene dicarboxylate),33 and [Zn3(O)2(H2L)-
(L)2] [H2L = N,N-bis(5-ethyl-1,3,4-thiadiazol-2-yl)-2,6-pyridi-
nedicarboxamide].34 The μ3-OZn bond lengths in 7 [1.896(3)-
1.903(3) A] are shorter than the analogous bonds involving μ4-
OZn4 in 6 [1.9539(17)-2.0019(15) A] but longer than those for
linear μ2-OZn2 [1.854(1) A] in the highly hindered compound
[(TpCum, Me)Zn]2(O) [TpCum, Me = hydrotris(3-p-cumenyl-5-
methylpyrazolyl)borate].35 The ligands fall into two groups; those
based on O(1) and O(4) are μ2,κ2-O,N, while the O(2),O(3)-
centered ligands are both μ2,κ3-O,N,N in their bonding modes.
The details of how these decomposition products are formed
are a matter for conjecture, but the structural elements that
make up 6 and 7 provide an indication. Compound 6 consists of
a six-membered (MeZn)3(bdmap)2O fragment that can be
traced back to the homolysis of the O-O bond in 4 leading
to, initially, the radical [(MeZn)3(bdmap)2(O•)], which forms
either 5 by H• abstraction or 6 by another route. The remaining
This is the general type of reaction sequence that others
have used to explain the formation of zincoxane products
from zinc peroxide intermediates. For example,13,17
Of2
t
Me2Znð BuDABÞ
t
Me2Znð BuDABÞ
t
MeðMeOOÞZnð BuDABÞ
f
- MeO , - Me
3
3
t
½MeZnðOOMeÞꢁ2½MeZnOZnMeð BuDABÞꢁ2
Of2
EtZnðPyr-pyrÞ
- EtO , - Et
3
3
½EtOOZnðPyr-pyrÞꢁ½EtZnðPyr - pyrÞꢁn
f
f
Zn4ðPyr-pyrÞ6O
However, the fact that the decomposition products have
been observed from reactions in which 4 is formed in situ
from adduct 2 (which contains excess ZnMe2) and also from
isolated 4 (which has no excess ZnMe2) means that the
radical derived from 4 must react at zinc of a second species
but not with loss of Me•, as all the zinc centers in 6 retain
methyl groups, unless 4 undergoes Schlenk equilibria which
generate ZnMe2 in situ. Thus, while the decomposition of 4
and formation of 6 do appear to be facilitated by the presence
(31) Zhu, G.; Tanski, J. M.; Parkin, G. J. Chem. Crystallogr. 2002, 32,
469.
(32) Murso, A.; Stalke, D. Dalton Trans. 2004, 2563.
(33) Kim, Y. S.; Kim, J.; Kim, D.; Chae, H. K. Chem. Lett. 2007, 36,
150.
(34) Shen, X.-Q.; Yao, H.-C.; Yang, R.; Li, Z.-J.; Zhang, H.-Y.; Wu,
B.-L.; Hou, H.-W. Polyhedron 2008, 27, 203.
(35) Ruf, M.; Vahrenkamp, H. Inorg. Chem. 1996, 35, 6571.

3326 Organometallics, Vol. 29, No. 15, 2010
Hollingsworth et al.
of ZnMe2, as noted above in the discussion of Figure 4c, it
seems likely that the decomposition also involves equilibria
of the following kind (L = bdmap):
former, the ZnMe2 is coordinated to a free N-donor of the
bdmap ligand and rather more loosely to the oxygen of the
alkoxide. 2 reacts with O2 at low temperatures with controlled
insertion into one of the Zn-C bonds of the coordinated
ZnMe2 group to form the isolable peroxide [MeZn(bdmap)]2-
MeZnOOMe (4), a reaction that supports the idea that in situ
formation of a three-coordinate zinc is key to this reaction. 4
decomposes slowly, and the hydroxide [MeZn(bdmap)]2-
MeZnOH (5), possibly arising from homolysis of the peroxide
O-O bond, is isolated. In addition to 5, two other decomposi-
tion products have been unambiguously identified, namely,
(MeZn)5(bdmap)3O (6) and (MeZn)4(bdmap)4ZnO (7). The
formation of these species can be linked to reactions of the
hydroxide (5), or its associated radical [MeZn(bdmap)]2Me-
Zn(O•)], with species such as ZnMe2 or MeZn(bdmap), present
in solution as a result of operating Schlenk equilibria.
Equilibria of the kind shown above are prevalent in this
area of organozinc chemistry, and we have previously shown
that Zn(bdmap)2 crystallizes from an aged sample of [MeZn-
(bdmap)]n, and EtZn3(bdmap)5 from [EtZn(bdmap)]3 by simi-
lar routes.20 Of the species that are plausibly present in solution,
we would expect the order of reactivity to be ZnMe2 > MeZnL
> ZnL2, based on a diminishing number of reactive Zn-Me
bonds, the monodentate nature of Me compared to bi/triden-
tate bdmap, and finally the increased coordination at zinc in
Zn(bdmap)2 (CN = 5) in comparison with the two organo-
metallics (CN=4).20 In addition, we might expect Zn(OOMe)2
to be highly reactive and arguably be, at least in part, the origin
of any insoluble end-products such as Zn(OMe)2 or ZnO.
Acknowledgment. We thank the EPSRC for a doctoral
training award (to N.H.). SAFC Hitech is thanked for a
gift of ZnMe2.
Supporting Information Available: This material is available
free of charge via the Internet at http://pubs.acs.org. Crystal-
lographic data for the structural analysis (in CIF format) have
also been deposited with the Cambridge Crystallographic Data
Center, CCDC nos. 698985-698990 and 766016 for 1-6 and 7,
respectively. Copies of this information may be obtained from
the Director, CCDC, 12 Union Road, Cambridge, CB21EZ,
UK (fax: þ44-1233-336033; e-mail: deposit@ccdc.cam.ac.uk or
www.ccdc.cam.ac.uk).
Conclusions
The reaction of ZnMe2 with Hbdmap in 2:1 ratio forms both
[MeZn(bdmap) ZnMe2]2 (2) and [MeZn(bdmap)]3 ZnMe2
(3) depending on the concentration of the reaction. In the
3
3