Synthesis and characterisation of tetramethylpiperidinyloxide (TEMPO) complexes of group 13 metal hydrides

Article abstract of DOI:10.1039/b702641b

Attempts have been made to prepare low oxidation state group 13 hydride complexes by reacting the metal trihydride-Lewis base adducts, [MH ₃(quin)] (quin = quinuclidine, M = Al, Ga or In), with reducing or hydrogen abstraction reagents. The 1:

Full text of DOI:10.1039/b702641b

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/njc | New Journal of Chemistry
Synthesis and characterisation of tetramethylpiperidinyloxide (TEMPO)
complexes of group 13 metal hydrides
Cameron Jones*ab and Richard P. Roseab
Received (in Durham, UK) 20th February 2007, Accepted 18th April 2007
First published as an Advance Article on the web 14th May 2007
DOI: 10.1039/b702641b
Attempts have been made to prepare low oxidation state group 13 hydride complexes by reacting
the metal trihydride–Lewis base adducts, [MH3(quin)] (quin = quinuclidine, M = Al, Ga or In),
with reducing or hydrogen abstraction reagents. The 1 : 1 reactions with 2,20,6,60-
tetramethylpiperidinyl oxide (TEMPO) yield the first structurally characterised nitroxide–group 13
hydride complexes, [MH2(quin)(TEMPO)] (M = Al or Ga). The 2 : 1 reaction with [AlH3(quin)]
affords [AlH(quin)(TEMPO)2] which has also been structurally authenticated. The mechanisms of
these reactions are discussed.
complexes with reductants or radical hydrogen abstraction
reagents. Of the latter, 2,20,6,60-tetramethylpiperidinyl oxide
Introduction
The chemistry of aluminium and gallium trihydrides and their
(TEMPO) offered promise as it has been shown to homolyti-
Lewis base adducts (L ꢀ MH3, M = Al or Ga, L = Lewis base)
cally cleave other p-block metal–hydrogen bonds (e.g. as in
has been extensively explored.1 These compounds have found
Bun3SnH or Ph3GeH) via hydrogen transfer to give metal–
metal bonded species (e.g. Bun3SnSnBun3).9 In addition, its
numerous applications as, for example, chemo- and stereo-
selective reductants in organic synthesis,2 and as precursors for
reactivity towards low oxidation state p-block complexes has
been previously examined.10 In this report, we detail the
materials (e.g. metal films and III/V semiconductors) impor-
tant to the microelectronics industry.3 In recent years, we have
reactivity of group 13 metal trihydride adducts towards
been examining an extension of this field to the stabilisation
TEMPO and other reagents. The reactions with TEMPO did
and further reactivity of indium trihydride complexes, with
not lead to low oxidation state metal hydride complexes, but
marked success.4 We have found that the chemistry of indium
instead to the first examples of nitroxide-group 13 metal
trihydride adducts can significantly differ from that of their
hydride species.
aluminium and gallium counterparts. This is perhaps best
illustrated by our recent report of the controlled decomposi-
Results and discussion
tion of [InH3(quin)] (quin = quinuclidine) in the presence of
LiBr, which led to the first mixed oxidation state indium sub-
halide complex, [In5Br8(quin)4][(quin)2H].5
The metal trihydride adducts chosen for this study were
[MH3(quin)], M = Al, Ga and In, which all possess good
thermal stability relative to other Lewis base adducts. A
number of methods for the preparation of low oxidation state
metal hydride complexes were attempted. First, reductions of
diethyl ether solutions of each adduct at ꢁ20 1C with sodium
naphthalenide in 1 : 1 stoichiometries led only to complete
decomposition of the adducts. Similarly, UV irradiation (l =
254 nm) of a mixture of each complex with Re2(CO)10 (gen-
erating ꢂRe(CO)5 radicals) in toluene at ꢁ60 1C afforded
intractable mixtures of products, free of M–H bonds (as
determined by IR spectroscopy). These decomposition pro-
cesses are not due solely to UV irradiation of the solutions, as
repeating the irradiations in the absence of Re2(CO)10 under
identical conditions led to no reactions.
In contrast, the fascinating chemistry of aluminium and
gallium sub-halide complexes and clusters (e.g. [Al22Cl20-
(THF)12]) has been well developed, largely by the group of
Schnockel.6 Such complexes have been mainly accessed
¨
through the controlled decomposition of metastable metal(I)
halide complexes, [{MX(L)}n] (M = Al or Ga, X = halide, L
= amine, phosphine or ether). It occurred to us that there are
no examples of sub-valent group 13 hydride complexes,
[MxHy(L)z], but if these could be isolated their chemistry
would be of great interest. It is of note that, to the best of
our knowledge, there are only two structurally characterised
complexes that could be considered to contain formally low
oxidation state group 13 metal hydride fragments, viz.
[{(DAB)Ga}2GaH2][K(tmeda)2] (DAB = {(2,6-C6H3Pri2)-
NCH}2)7 and [(Ar0Ga)(m-GaH)(m-PH2)2(GaAr0)] (Ar0 = 2,6-
C6H3(2,4,6-C6H2Pri3)2).8
Attention then shifted to the use of TEMPO, a stable
nitroxide radical, which we proposed could effect controlled
hydrogen radical abstraction from the metal hydride adducts.
This seemed reasonable in light of the aforementioned report
on hydrogen abstraction from Bun3SnH by TEMPO.9 How-
ever, the 1 : 1 reactions of [MH3(quin)] (M = Al or Ga) with
TEMPO in toluene at ꢁ78 1C followed by slow warming to
20 1C gave low to moderate isolated yields of the nitroxide
anion–metal hydride complexes, 1 and 2 (Scheme 1). The
One potential route to low oxidation state group 13 metal
hydride complexes is via the reactions of metal trihydride
a School of Chemistry, Monash University, P.O. Box 23, Victoria,
3800, Australia. E-mail: cameron.jones@monash.edu.au
b School of Chemistry, Main Building, Cardiff University, Cardiff,
UK CF10 3AT
ꢃc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007
1484 | New J. Chem., 2007, 31, 1484–1487

View Article Online
by initial attack of TEMPO at the metal centre of
Sm(C5Me5)3.11 It is of note that reactions of TEMPO with
group 13 trihalides do not proceed similarly and yield para-
magnetic adducts of the type [MX3(TEMPO)].12
The fact that no gallium counterpart of 3 is formed in the
2 : 1 reaction of TEMPO with [GaH3(quin)] can be explained
by the lower Lewis acidity of gallium hydride complexes
relative to aluminium hydrides.1 This would make attack of
TEMPO at the hindered gallium centre of 2 less likely than
attack at the aluminium centre of 1, which ultimately gives 3.
It is noteworthy that 3 does not react with excess TEMPO at
ambient temperature.
The ambient temperature NMR spectra of the diamagnetic
complexes 1–3 all display significant peak broadening. This
could be due to a combination of processes including rotation
of the TEMPO ligand about the N–O and/or M–O bonds of
the complexes, and/or inversion about the N-centre of the
TEMPO ligand(s). Whatever the process in operation is,
variable temperature NMR studies on 1–3 did not lead to
any significant resolution of their spectra. In addition, no
1
hydride resonances were observed in the H NMR spectra of
the aluminium complexes due to quadrupolar broadening of
these signals by the metal (27Al, 100%, I = 5/2). In contrast,
1
the H NMR spectrum of 2 does exhibit a hydride resonance
(d 5.22 ppm) in the normal range,1 but this is very broad due to
the quadrupolar nature of both naturally occurring gallium
isotopes (69Ga, 60.1%, I = 3/2; 71Ga, 39.9%, I = 3/2). The
Scheme 1 Reagents and conditions: (i) TEMPO, toluene, ꢁ1/2 H2; (ii)
2 TEMPO, toluene, ꢁH2.
M–H stretching absorptions for 1 (1782 cmꢁ1), 2 (1850 cmꢁ1
)
and 3 (1819 cmꢁ1) are all broad and all occur at higher
related reaction with [InH3(quin)] led to indium metal deposi-
tion above ꢁ15 1C, probably because the In analogue of 1 and
2 is thermally unstable above that temperature. The only other
isolated product from this reaction was free quinuclidine.
When the reaction with [AlH3(quin)] was repeated with 2
equivalents of TEMPO, a moderate yield of the bis(nitroxide)
complex, 3, resulted. The corresponding reaction with [GaH3-
(quin)] gave the 1 : 1 complex, 2, and unreacted TEMPO as the
only identifiable products, whilst the reaction with
[InH3(quin)] led, again, to indium metal deposition.
frequencies than their metal trihydride precursors,
[MH3(quin)] M = Al (1720 cmꢁ1 13 or Ga (1810 cmꢁ1 14
)
)
(for comparison: M = In, 1642 cmꢁ1)4. This can be explained
by the greater electron withdrawing ability of the TEMPO
anion than the hydride ligand, which leads to a reduction in
the polarity of the M–H bonds of the complexes with increas-
ing TEMPO substitution.
The X-ray crystal structures of 1–3 (Fig. 1 and 2) were
determined and represent the first for nitroxide complexes of
group 13 metal hydrides. Moreover, there has been only one
previously structurally characterised TEMPO–group 13 com-
plex, viz. [B(TEMPO)3].15 The structure of 2 is isomorphous
with that of 1 (both compounds sit on a mirror plane) so only
the molecular structure of the latter is depicted in Fig. 1
(relevant metrical data for 2 are included in the caption).
The hydride ligands of all complexes were located from
difference maps and refined isotropically without restraints.
This allowed the coordination geometry of each to be assigned
as distorted tetrahedral with surprisingly acute O–M–N angles
(1 94.04(13)1, 2 90.8(2)1, 3 97.51 (mean)). The M–O, M–N and
It seems likely that the mechanisms of the 1 : 1 reactions do
not involve direct hydrogen transfer from the metal hydride to
TEMPO, as this would generate TEMPOH (as in its reaction
with Bun3SnH) and would require a second equivalent of
TEMPO to form the observed products. Another possibility
exists, whereby one molecule of [MH3(quin)] reacts with
TEMPO to generate TEMPOH and this, in effect, acts as an
acid in a subsequent reaction with a second molecule of
[MH3(quin)] to give 1 or 2 and dihydrogen. This would,
however, necessitate the formation of M(II) hydrides, and/or
their decomposition products (e.g. elemental M), in the pri-
mary step, no evidence for which was observed. Instead, we
believe TEMPO acts as a Lewis base and attacks the metal
centres, leading to homolytic cleavage of one M–H bond and
subsequent dihydrogen formation (that these M–H cleavages
occur via a bimolecular process involving M–H–M bridged
intermediates1 cannot be ruled out). Indeed, gas evolution
from all reaction mixtures was observed in the temperature
range ꢁ50 to ꢁ20 1C. This can be directly compared to the
reaction of TEMPO with Sm(C5Me5)3 which has been re-
ported to yield a dimeric form of Sm(TEMPO)3 and (C5Me5)2
M–H distances are all in the normal ranges,16 whilst the N–O
P
distances and pyramidal TEMPO N-centres ( angles: 1
330.41, 2 326.61, 3 332.61 (mean)) are indicative of fully
reduced nitroxide ligands.
Conclusion
In summary, the first examples of nitroxide–group 13 hydride
complexes have been prepared in the reactions of TEMPO
with quinuclidine adducts of aluminium trihydride or gallium
ꢃc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007
New J. Chem., 2007, 31, 1484–1487 | 1485

View Article Online
Eclipse 300 plus spectrometer and were referenced to the
residual 1H or 13C resonances of the solvent used. Mass spectra
were obtained from the EPSRC National Mass Spectrometry
Service at Swansea University. IR spectra were recorded using
a Nicolet 510 FT-IR spectrometer as Nujol mulls between
NaCl plates. Reproducible microanalyses on all compounds
could not be obtained due to their extreme air and moisture
sensitivity. Melting points were determined in sealed glass
capillaries under argon, and are uncorrected. [AlH3(quin)]13
and [GaH3(quin)]14 were synthesised by literature procedures,
whilst all other chemicals were obtained from commercial
sources and used as supplied.
Fig.
1 Molecular structure of 1 (ellipsoids shown at the 25%
probability level). Selected bond lengths (A) and angles (1):
Al(1)–O(1) 1.754(3), Al(1)–N(1) 2.003(3), Al(1)–H(1) 1.59(3),
O(1)–N(2) 1.456(4); O(1)–Al(1)–N(1) 94.04(13), O(1)–Al(1)–H(1)
117.7(10), N(1)–Al(1)–H(1) 103.2(11), N(2)–O(1)–Al(1) 113.4(2),
H(1)–Al(1)–H(1)0 115.4(9); Selected bond lengths (A) and angles (1)
for 2: Ga(1)–O(1) 1.850(5), Ga(1)–N(1) 2.078(7), Ga(1)–H(1) 1.43(6),
O(1)–N(2) 1.447(8); O(1)–Ga(1)–N(1) 90.8(2), O(1)–Ga(1)–H(1) 117(2),
N(1)–Ga(1)–H(1) 103(2), N(2)–O(1)–Ga(1) 112.7(4), H(1)–Ga(1)–H(1)0
Preparation of [AlH2(quin)(TEMPO)] 1
To a solution of [AlH3(quin)] (0.17 g, 1.25 mmol) in toluene
(15 cm3) at ꢁ78 1C was added a solution of TEMPO (0.19 g,
1.25 mmol) in toluene (15 cm3) over 5 min. The resultant
solution was slowly warmed to room temperature and stirred
overnight to yield a pale yellow solution. Volatiles were
removed in vacuo and the residue extracted with hexane
(10 cm3). Filtration, concentration and cooling of the extract
to ꢁ30 1C overnight yielded colourless crystals of 1 (0.05 g,
15%) Mp 97–100 1C. 1H NMR (400 MHz, C6D6, 298 K):
d 0.92 (br s, 12H, CH3), 1.26–1.53 (br m, 6H, TEMPO CH2),
1.58 (br, 6H, quin CH2), 1.78 (br, 1H, quin CH), 2.47 (br,
6H, quin NCH2), AlH2 resonance not observed; 13C NMR
(75 MHz, C6D6, 298 K): d 18.3 (TEMPO CH2), 20.0
(CH), 22.3 (CH3), 24.5 (quin CH2), 40.6 (TEMPO CH2),
46.9 (NCH2), 59.7 (NCMe2); IR n/cmꢁ1 (Nujol): 1782
(AlH); MS/EI m/z: 157 [TEMPOH+, 78%], 141 [TEMPOH+
ꢁ O, 100%].
0
118.6(9). Symmetry operation: x, ꢁy + 1/2, z.
trihydride. The mechanisms of these reactions probably in-
volve nucleophilic attack of TEMPO at the metal centre and
subsequent cleavage of M–H bonds. We are continuing our
efforts to prepare low oxidation state group 13 hydride com-
plexes and will report on this in due course.
Experimental
General considerations
All manipulations were carried out using standard Schlenk and
glove box techniques under an atmosphere of high purity
argon. Hexane and toluene were distilled over potassium then
freeze–thaw degassed prior to use. 1H and 13C{1H} NMR
spectra were recorded on either a Bruker DXP400 or a Jeol
Preparation of [GaH2(quin)(TEMPO)] 2
To a solution of [GaH3(quin)] (0.39 g, 2.15 mmol) in toluene
(15 cm3) at ꢁ78 1C was added a solution of TEMPO (0.34 g,
2.15 mmol) in toluene (15 cm3) over 5 min. The resultant
solution was warmed to room temperature and stirred over-
night to yield a pale yellow solution. Volatiles were removed
in vacuo and the residue extracted with hexane (10 cm3).
Filtration, concentration and cooling to ꢁ30 1C overnight
yielded colourless crystals of 2 (0.17 g, 23%) Mp 110–112 1C.
1H NMR (400 MHz, C6D6, 298 K): d 0.72 (br s, 12H, CH3),
1.23–1.77 (br m, 6H, TEMPO CH2), 1.49 (br, 6H, quin CH2),
1.78 (br, 1H, quin CH), 2.42 (br, 6H, quin NCH2), 5.22 (br,
2H, GaH2); 13C NMR (75 MHz, C6D6, 298 K): d 17.8
(TEMPO CH2), 20.0 (CH), 24.4 (CH3), 25.2 (quin CH2),
40.3 (TEMPO CH2), 47.6 (NCH2), 59.2 (NCMe2); IR
n/cmꢁ1 (Nujol): 1850 (GaH); MS/EI m/z: 338 [M+ ꢁ H,
100%], 157 [TEMPOH+, 71%].
Preparation of [AlH(quin)(TEMPO)2] 3
To a solution of [AlH3(quin)] (0.31 g, 2.16 mmol) in toluene
(15 cm3) at ꢁ78 1C was added a solution of TEMPO (0.68 g,
4.33 mmol) in toluene (15 cm3) over 5 min. The resultant
solution was slowly warmed to room temperature and stirred
overnight to yield a pale yellow solution. Volatiles were
removed in vacuo and the residue extracted with hexane
(10 cm3). Filtration, concentration and cooling of the extract
Fig. 2 Molecular structure of 3 (ellipsoids shown at the 25% prob-
ability level). Selected bond lengths (A) and angles (1): Al(1)–O(1)
1.7423(15), Al(1)–O(2) 1.7535(15), Al(1)–N(3) 2.0215(17), Al(1)–H(1)
1.55(2), O(1)–N(1) 1.4602(19), O(2)–N(2) 1.453(2); O(1)–Al(1)–O(2)
119.89(7), O(1)–Al(1)–N(3) 101.38(7), O(2)–Al(1)–N(3) 93.68(7), O(1)–
Al(1)–H(1) 116.5(8), O(2)–Al(1)–H(1) 115.8(8), N(3)– Al(1)– H(1)
103.1(8), N(1)–O(1)–Al(1) 121.37(11), N(2)–O(2)–Al(1) 120.48(10).
ꢃc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007
1486 | New J. Chem., 2007, 31, 1484–1487

View Article Online
Table 1 Crystal data for compounds 1–3
Compound
1
2
3
Empirical formula
Mr
T/K
Crystal system
Space group
a/A
b/A
c/A
C16H33AlN2O
296.42
150(2)
Monoclinic
P21/m
6.3500(13)
10.747(2)
13.239(3)
92.69(3)
902.5(3)
2
1.091
0.112
328
6069
1674
0.0913
C16H33GaN2O
339.16
150(2)
Monoclinic
P21/m
6.3500(13)
10.729(2)
13.273(3)
92.61(3)
903.3(3)
2
1.247
1.522
364
4659
1672
0.1353
C25H50AlN3O2
451.66
150(2)
Monoclinic
P21/c
12.707(3)
11.963(2)
17.660(4)
95.14(3)
2673.8(9)
4
1.122
0.100
1000
18772
6103
0.0763
b/1
V/A3
Z
Dc/Mg mꢁ3
m(Mo-Ka)/mmꢁ1
F(000)
No. reflns collected
No. independent reflns
Rint
Final R indices (I 4 2s(I))
R1 = 0.0679
wR2 = 0.1507
R1 = 0.0897
wR2 = 0.2239
R1 = 0.0599
wR2 = 0.1197
to ꢁ30 1C overnight yielded colourless crystals of 3 (0.38 g,
References
1
39%) Mp 159–162 1C. H NMR (400 MHz, C6D6, 298 K): d
1 See, for example: (a) S. Aldridge and A. J. Downs, Chem. Rev.,
2001, 101, 3305; (b) C. Jones, G. A. Koutsantonis and C. L.
Raston, Polyhedron, 1993, 12, 1829; (c) M. G. Gardiner and C.
L. Raston, Coord. Chem. Rev., 1997, 166, 1; (d) A. J. Downs and C.
R. Pulham, Chem. Soc. Rev., 1994, 175, and references therein.
2 C. L. Raston, A. F. H. Siu, C. J. Tranter and D. J. Young,
Tetrahedron Lett., 1994, 35, 5915, and references therein.
3 J. A. Jegier and W. L. Gladfelter, Coord. Chem. Rev., 2000,
206–207, 631, and references therein.
4 C. Jones, Chem. Commun., 2001, 2293, and references therein.
5 (a) M. L. Cole, C. Jones and M. Kloth, Inorg. Chem., 2005, 44,
4909; (b) For an excellent review of sub-valent indium complexes,
see also: J. A. J. Pardoe and A. J. Downs, Chem. Rev., 2007, 107, 2.
0.69 (br s, 24H, CH3), 1.30–1.61 (br. m, 12H, TEMPO CH2),
1.51 (br, 6H, quin CH2), 1.72 (br, 1H, quin CH), 2.48 (br, 6H,
quin NCH2), AlH resonance not observed; 13C NMR (75
MHz, C6D6, 298 K): d 18.4 (TEMPO CH2), 20.4 (CH), 22.1
(CH3), 24.6 (quin CH2), 40.1 (TEMPO CH2), 42.0 (NCH2),
59.7 (NCMe2); IR n/cmꢁ1 (Nujol): 1819 (AlH); MS/EI m/z:
157 [TEMPOH+, 100%].
X-Ray crystallography
6 See, for example: (a) A. Schnepf and H.-G. Schnockel, Angew.
¨
in Molecular Clusters of the Main Group Elements, eds. M. Driess
and H. Noth, Wiley-VCH, Weinheim, 2004, pp. 126–168, and
¨
references therein.
7 R. J. Baker, C. Jones, P. C. Junk and M. Kloth, Angew. Chem., Int.
Ed., 2004, 43, 3852.
8 X.-W. Li, P. Wei, B. C. Beck, Y. Xie, H. F. Schaefer III, J. Su and
G. H. Robinson, Chem. Commun., 2000, 453.
¨
Chem., Int. Ed., 2002, 41, 3532; (b) G. Linti and H.-G. Schnockel,
Crystals of 1–3 suitable for X-ray structural determination
were mounted in silicone oil. Crystallographic measurements
were made using a Nonius Kappa CCD diffractometer. The
structures were solved by direct methods and refined on F2 by
full matrix least squares (SHELX97)17 using all unique data.
All non-hydrogen atoms are anisotropic with non-hydridic
H-atoms included in calculated positions (riding model).
Hydride ligands were located from difference maps and refined
isotropically. Crystal data, details of data collections and
refinement are given in Table 1.
9 M. Lucarini, E. Marchesi, G. F. Pedulli and C. Chatilialoglu, J.
Org. Chem., 1998, 63, 1687.
10 See, for example: (a) G. H. Spikes, Y. Peng, J. C. Fettinger, J.
Steiner and P. P. Power, Chem. Commun., 2005, 6041; (b) A. K.
Naka, N. J. Hill and R. West, Organometallics, 2004, 23, 6330.
11 W. J. Evans, J. M. Perotti, R. J. Doedens and J. W. Ziller, Chem.
Commun., 2001, 2326.
CCDC reference numbers 645534–645536.
For crystallographic data in CIF or other electronic format
see DOI: 10.1039/b702641b
12 See, for example: C. Hambly and J. B. Raynor, J. Chem. Soc.,
Dalton Trans., 1974, 604.
13 J. L. Atwood, K. W. Butz, M. G. Gardiner, C. Jones, G. A.
Koutsantonis, C. L. Raston and K. D. Robinson, Inorg. Chem.,
1993, 32, 3482.
14 J. L. Atwood, S. G. Bott, F. M. Elms, C. Jones and C. L. Raston,
Inorg. Chem., 1991, 30, 3792.
Acknowledgements
15 M. Ambrecht, W. Maringgele, A. Meller, M. Noltemeyer and G.
M. Sheldrick, Z. Naturforsch., Teil B, 1985, 40, 1113.
16 As determined from a survey of the Cambridge Crystallographic
Database, February, 2007.
We thank the Australian Research Council (professorial fel-
lowship for C. J.) and the EPSRC (studentship for R. P. R.)
for financial support. The EPSRC Mass Spectrometry Service
at Swansea University is also thanked.
17 G. M. Sheldrick, SHELX-97, University of Gottingen, 1997.
¨
ꢃc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007
New J. Chem., 2007, 31, 1484–1487 | 1487