Synthesis, solid state structure, and solution behaviour of the lighter lanthanide bis(trimethylsilyl)amido chlorides, [Ln{N(SiMe3)2}2(THF)(μ- Cl)]2 (Ln = Ce, Nd)

Article abstract of DOI:10.1139/v00-036

The synthesis of [Ln{N(SiMe₃)₂}₂(THF)(μ-Cl)]₂ (Ln = Ce, 1; Nd, 2) by reaction of sodium bis(trimethylsilyl)amide (2 equiv.) with LnCl₃ is reported. The same complexes were also isolated from the ligand redistribution reactions of Ln[N(SiMe₃)₂]₃ and LnCl₃ (2:1 ratio) in THF at 80°C. The crystal structure of 2, determined by X-ray diffraction, revealed a centrosymmetric dimer with bridging chlorides and pentacoordinate metal centres. ¹H NMR studies show that the solid state structure is not maintained in solution. NMR evidence for the presence of Ln[N(SiMe₃)₂]₃ and two other bis(trimethylsilyl)amide containing species, presumably Ln[N(SiMe₃)₂]Cl₂(THF)(x) and Ln[N(SiMe₃)₂]₂Cl(THF)(y), is presented.

Full text of DOI:10.1139/v00-036

454
Synthesis, solid state structure, and solution
behaviour of the lighter lanthanide
bis(trimethylsilyl)amido chlorides,
[Ln{N(SiMe3)2}2(THF)(µ-Cl)]2 (Ln = Ce, Nd)
David J. Berg and Roland A.L. Gendron
Abstract: The synthesis of [Ln{N(SiMe3)2}2(THF)(µ-Cl)]2 (Ln = Ce, 1; Nd, 2) by reaction of sodium
bis(trimethylsilyl)amide (2 equiv.) with LnCl3 is reported. The same complexes were also isolated from the ligand re-
distribution reactions of Ln[N(SiMe3)2]3 and LnCl3 (2:1 ratio) in THF at 80°C. The crystal structure of 2, determined
1
by X-ray diffraction, revealed a centrosymmetric dimer with bridging chlorides and pentacoordinate metal centres. H
NMR studies show that the solid state structure is not maintained in solution. NMR evidence for the presence of
Ln[N(SiMe3)2]3 and two other bis(trimethylsilyl)amide containing species, presumably Ln[N(SiMe3)2]Cl2(THF)x and
Ln[N(SiMe3)2]2Cl(THF)y, is presented.
Key words: lanthanide, amide, neodymium, cerium, crystal structure, X-ray, nuclear magnetic resonance, redistribution.
Résumé : On a effectué la synthèse du [Ln{N(SiMe3)2}2(THF)(µ-Cl)]2 (Ln = Ce, 1; Nd, 2) par réaction du
bis(triméthylsilyl)amide (2 équiv.alents) avec du LnCl3. On a aussi isolé les mêmes complexes à partir de réactions de
ligands de Ln[N(SiMe3)2]3 et LiCl3 (rapport 2:1) dans le THF à 80°C. La structure cristalline du produit 2, déterminée
par diffraction des rayons X, met en évidence la présence d’un dimère centrosymétrique comportant des atomes de
1
chlore agissant comme pont et des centres métalliques pentacoordinés. Les études de RMN du H montrent que la
structure solide n’est pas maintenue en solution. On rapporte des données de RMN qui mettent en évidence la présence
de Ln(N(SiMe3)3]3 et de deux autres espèces contenant du bis(triméthylsilyl)amide, probablement du
Ln[N(SiMe3)2]Cl2(THF)x et Ln[N(SiMe3)2]Cl(THF)y.
Mots clés : lanthanide, amide, néodyme, cérium, structure cristalline, rayons X, résonance magnétique nucléaire,
redistribution.
[Traduit par la Rédaction] Berg and Gendron 458
Introduction
are attractive alternative precursors because the amido group
is quite basic and the bis(trimethylsilyl)amine that is formed
is noncoordinating and volatile, making its removal straight-
forward.
The incorporation of anionic ligands into lanthanide com-
plexes is usually achieved by metathetical reactions between
a lanthanide halide or alkoxide and a salt of the ligand
(eqs. [1], [2]). While often useful, there are many instances
when pure products can not be obtained by this method due
to incomplete substitution of the leaving group or formation
of an anionic “ate” complexes (an addition reaction, eq. [3])
(for some representative examples see ref. (1)). A proton-
olysis reaction between an acidic, protonated form of the in-
coming ligand and strongly basic anionic leaving groups is
an attractive alternative in many cases, particularly when the
leaving group has little propensity to coordinate to the prod-
uct complex (eq. [4]). Bis(trimethylsilyl)amido complexes
[1]
3 M+L– + LnCl3
→ LnL3 + 3 M+Cl– 9
(M = Li, Na, K)
[2]
[3]
3 M+L– + Ln(OAr)3
→
LnL3 + 3 MOAr 9
(Ar = 2,6-t-Bu2C6H3)
3 M+L– + LnCl3 → [LnL3Cl]– M+ + 2 M+Cl– 9
(M = Li, Na, K)
′
[4]
3 HL′ + LnL3
→ LnL3 + 3 HL
(L = N(SiMe3)2)
Received December 8, 1999. Published on the NRC Research
Press website on March 31, 2000.
The synthesis of mixed ligand complexes of the lanthanides
by methathetical reactions is even more challenging because
controlled substitution of the leaving groups is difficult and
ligand redistribution often occurs under these conditions.
(2) The successful isolation of mixed bis(trimethlsilyl)amido
chloride complexes [Ln{N(SiMe3)2}2(THF)(µ-Cl)]2 (Ln = Eu,
Gd, Yb) by Aspinall et al. (3) using metathesis is noteworthy
D.J. Berg1 and R.A.L. Gendron. Department of Chemistry,
University of Victoria, P.O. Box 3065, Victoria, BC V8W
3V6, Canada.
1Author to whom correspondence may be addressed.
Telephone: (250) 721-7161. Fax: (250) 721-7147.
e-mail: DJBERG@UVIC.CA
Can. J. Chem. 78: 454–458 (2000)
© 2000 NRC Canada

Berg and Gendron
455
Table 2. Selected bond distances and angles for 2a.
Table 1. Crystallographic data for 2.
Distances (Å)
Formula
C16H44N2OSi4ClNd
Nd(1)—Cl(1)
Nd(1)—Cl(1)′
Nd(1)—O(1)
Nd(1)—N(1)
Nd(1)—N(2)
N(1)—Si(1)
N(1)—Si(2)
N(2)—Si(3)
N(2)—Si(4)
2.791(3)
Fw (dimer)
Cryst syst
Space group
a (Å)
b (Å)
c (Å)
β (deg)
V (Å3)
Z
ρ (calcd) (g cm–3)
µ (cm–1)
1145.14
2.854(3)
2.496(7)
2.300(8)
2.335(8)
1.702(9)
1.732(9)
1.705(9)
1.707(9)
monoclinic
P 21/n (no. 14)
15.210(2)
12.074(1)
16.621(3)
112.62(1)
2817.6(4)
2 (dimers)
1.35
21.1
Mo Kα, 0.7107
ambient
45.0
3667
Angles (deg)
Cl(1)–Nd(1)–Cl(1)′
Cl(1)–Nd(1)–O(1)
Cl(1)–Nd(1)–N(1)
Cl(1)–Nd(1)–N(2)
Cl(1)′–Nd(1)–O(1)
Cl(1)′–Nd(1)–N(1)
Cl(1)′–Nd(1)–N(2)
O(1)–Nd(1)–N(1)
O(1)–Nd(1)–N(2)
N(1)–Nd(1)–N(2)
Nd(1)–Cl(1)–Nd(1)′
Nd(1)–N(1)–Si(1)
Nd(1)–N(1)–Si(2)
Nd(1)–N(2)–Si(3)
Nd(1)–N(2)–Si(4)
Si(1)–N(1)–Si(2)
Si(3)–N(2)–Si(4)
73.66(10)
144.1(2)
99.0(2)
108.2(2)
76.2(2)
107.0(2)
136.7(2)
108.2(3)
81.3(3)
115.1(3)
106.3(1)
122.4(5)
115.9(4)
129.2(5)
110.2(4)
121.5(5)
120.6(5)
Radiation, λ (Å)
T
2θmax (deg)
No. obsd reflcns
No. of unique reflcns
2880
0.047, 0.049
a
b
R , Rw
aR = Σ( |Fo | – | Fc | )/Σ| Fo|.
bRw = [Σw( |Fo | – | Fc |2)/Σw( |Fo |)2]1/2
.
because the products are suitable precursors to other mixed
ligand complexes by protonolysis of the two silylamido
groups. However, these authors did not attempt to isolate the
analogous complexes for the lighter lanthanides because
they assumed that these complexes would be too
coordinatively unsaturated to exist as simple dimers. In this
contribution we demonstrate that bis(amido) complexes of
these metals are accessible both by metathesis or redistribu-
tion reactions.
aEstimated standard deviations are given in parentheses.
products were very easy to separate. It is noteworthy that
pure crystalline 1 and 2 inevitably produced some of the
tris(amido) product during recrystallization supporting facile
Results and discussion
Synthesis
The
1
redistribution in solution. This fact was confirmed by H
bis(trimethylsilyl)amido
chloride
complexes
NMR (vide infra) and by the successful synthesis of 1 and 2
from LnCl3 and Ln[N(SiMe3)2]3 by redistribution (eq. [6]).
The yields obtained by redistribution were comparable to
those obtained by metathesis.
[Ln{N(SiMe3)2}2(THF)(µ-Cl)]2 (Ln = Ce, 1; Ln = Nd, 2)
were initially prepared by stirring two equiv. of sodium
bis(trimethylsilyl)amide with the appropriate trichloride in
THF for 48 h, followed by filtration to remove NaCl, re-
moval of the solvent and recrystallization of the residue
from hexane at –30°C (eq. [5]). In agreement with earlier
observations (2, 3), complexes 1 and 2 were more intensely
coloured (1, canary yellow; 2, lilac) and more soluble than
their tris(amido) analogues, Ln[N(SiMe3)2]3. The crystals
are also easily distinguishable from the tris(amido) com-
plexes by their form: both 1 and 2 are isolated as well-
formed prisms, while the tris(amido) complexes of all the
lanthanides form feathery needles.
Solid-state structure of [Nd{N(SiMe3)2}2(THF)(µ-Cl)]2 (2)
The solid state structure of the neodymium complex 2 was
determined by X-ray crystallography. Crystallographic data
for 2 is summarized in Table 1, selected bond distances and
angles are given in Table 2, and the structure is shown in
Fig. 1. The complex is dimeric in the solid state and consists
of two five-coordinate metal centres bridged by chloride
ions. The geometry at the metal is approximately trigonal
bipyramidal with one chloride and two amido ligands defin-
ing the equatorial plane. The observed structure is
isostructural with those previously reported for the Gd and
Yb analogues (3). The metal–ligand bond distances in 2 are,
Small amounts of the tris(amido) complexes were always
isolated along with 1 and 2 during recrystallization but these
© 2000 NRC Canada

456
Can. J. Chem. Vol. 78, 2000
Fig. 1. ZORTEP plot (30% probability) of 2.
Fig. 2. Plot of chemical shift (ppm) vs. 1/T (× 1000 K–1) for 1.
Assignments: Ce[N(SiMe3)2]Cl2(THF)x (A), Ce[N(SiMe3)2]3 (B),
Ce[N(SiMe3)2]2Cl(THF)y (C), α-CH2THF (D), and β-CH2THF (E).
on average, 0.06 and 0.14 Å larger than those found in the
Gd and Yb complexes, respectively in keeping with the dif-
ference in ionic radii for these metals (0.04 and 0.11 Å, re-
spectively) (4). The only significant differences between this
structure and those for Gd and Yb is that the Ln–N–Si an-
gles span a much wider range (110.2(4)–129.2(5)° vs.
117.4(3)–123.5(3)°) and the bridging chlorides display
greater asymmetry (∆ = r(Ln-Cl) – r(Ln-Cl′) = 0.063(4)
(Nd), 0.025(6) (Gd), 0.023(6) (Yb) Å) in complex 2. The
greater asymmetry in the chloride bridges is surprising be-
cause 2 is clearly the least sterically crowded of the three
structures. A comparison of the Ln—Cl distances reveals
that the shorter Nd—Cl distance in 2 is consistent with that
predicted from the earlier Gd and Yb structures (predicted:
2.78(1) Å from both the Yb and Gd structures; observed:
2.791(3) Å). (4) In contrast, the longer Nd—Cl distance in 2
is significantly greater than expected based on the Gd and
Yb structures (predicted: 2.80(1) Å; observed: 2.854(3) Å).
Fig. 3. Plot of chemical shift (ppm) vs. 1/T (× 1000 K–1) for 2.
Assignments: Nd[N(SiMe3)2]Cl2(THF)x (A), Nd[N(SiMe3)2]3 (B),
Nd[N(SiMe3)2]2Cl(THF)y (C), α-CH2THF (D) and β-CH2THF (E).
Solution behaviour
The solution behaviour of 1 and 2, monitored in d5-
1
bromobenzene by variable temperature H NMR spectros-
copy (Figs. 2 and 3), clearly indicates that the solid state
structure is not maintained in solution. For both complexes,
five resonances were observed throughout the temperature
range studied (20–125°C). The two smaller resonances (D
and E) were established as THF signals by exchange with
d8-THF. One of the larger resonances (B) was established as
due to Ln[N(SiMe3)2]3 by addition of a large excess of the
authentic tris(amido) complex. The remaining two reso-
nances are tentatively assigned as the mono (A) and
bis(amido) (C) complexes, Ln[N(SiMe3)2]Cl2(THF)x and
Ln[N(SiMe3)2]2Cl(THF)y, respectively on the basis of inte-
gration. It is not possible to know the state of aggregation of
these two complexes in solution or the extent to which they
are solvated by THF, although it is extremely unlikely that
either exists as a base-free monomer. If it is assumed that
these are the only species present then a plausible solution
equilibrium is given in eq. [7].2
The silylamide integration ratio of Ln[N(SiMe3)2]3 (B)
relative to Ln[N(SiMe3)2]Cl2(THF)x (A) remains approximately
2 LnCl3 may also form but this product cannot be observed by NMR and there was no visual evidence for its formation.
© 2000 NRC Canada

Berg and Gendron
457
Fig. 4. Plot of chemical shift (ppm) vs. 1/T (× 1000 K–1) for the
Ln[N(SiMe3)2]3 resonances of 1(A) and 2 (B) and pure
Ln[N(SiMe3)2]3 (Ln = Ce (×), Nd (+)).
ligand systems by protonolysis reactions. However, in cases
where the reaction proceeds rapidly and irreversibly (for ex-
ample, giving an insoluble product), it is useful to remember
that this kinetic product may be derived from any of the
three amido species present in solution.
Experimental
General procedures
All manipulations were carried out under an argon atmo-
sphere, with the rigorous exclusion of oxygen and water, us-
ing standard glovebox (Braun MB150-GII) or Schlenk
techniques. Tetrahydrofuran (THF) and hexane were dried
by distillation from sodium benzophenone ketyl under argon
immediately prior to use. Ln[N(SiMe3)2]3 (Ln = Ce, Nd)
were prepared according to the literature (5).
1
Variable temperature H NMR spectra were recorded on a
Bruker WM-250 MHz spectrometer. Spectra were recorded
in C6D5Br dried over 4 Å molecular sieves for one week, us-
ing 5 mm tubes fitted with a teflon valve (Brunfeldt). The
sample was allowed to equilibrate for 15 min at each tem-
perature prior to recording the spectrum. Melting points
were recorded using a Reichert hot stage and are not cor-
rected. Elemental analyses were performed by Canadian
Microanalytical, Delta, B.C. Mass spectra were recorded on
a Kratos Concept H spectrometer using an electron impact
(70 eV) source.
3:1 at all temperatures which is what would be expected
based on eq. [7].
Preparation of [Ln{N(SiMe3)2}2(THF)(µ-Cl)]2
In
a
typical preparation,
a
solution of sodium
bis(trimethylsilyl)amide (2 equiv.) was added to a stirred
suspension of LnCl3 under argon at room temperature. The
solution was stirred for 48 h and the solvent was removed
under reduced pressure. The solid residue was extracted into
hexane, filtered through a Celite pad to remove NaCl, and
crystallized in the glovebox at –30°C. Generally this resulted
in a large mass of block-like crystals of the bis(amide) sur-
rounded by fine needles of Ln[N(SiMe3)2]3. The smaller
tris(amide) crystals can usually be removed with the mother
liquor by a Pasteur pipette. Although the larger bis(amide)
crystals are more soluble, they are slow to redissolve so
washing with cold hexane is an ideal way to wash off the
smaller tris(amide) needles.
Resonances A–C for 1 follow nearly linear behaviour
throughout the temperature range sampled. This is consistent
with each resonance being due to a single well-defined spe-
cies. The resonances due to the THF (D–E) do show curva-
ture over this temperature range. The temperature
dependence of the amido resonances for 2 show greater cur-
vature although the resonance due to Nd[N(SiMe3)2]3
(Fig. 3, B) is close to linear. The temperature dependence of
the Ln[N(SiMe3)2]3 resonances for 1 and 2 are compared
with those for separate samples of pure Ln[N(SiMe3)2]3 in
Fig. 4. The close similarity between these plots confirms the
assignments made in Fig. 2 and 3.3 In addition, since the res-
onances in Fig. 4 are essentially unperturbed from those of
pure Ln[N(SiMe3)2]3, the forward and reverse reaction rates
in eq. [7] must be slow on the NMR timescale over the en-
tire temperature range studied.
Alternatively, the complexes were prepared in comparable
yield by heating
a
THF solution of two equiv.
Ln[N(SiMe3)2]3 with one equiv. of LnCl3 in a sealed Kontes
flask at 80°C overnight. All solids dissolved on heating and
remained in solution after the reaction mixture was cooled
back to room temperature. The THF solvent was removed
under reduced pressure and the residue was recrystallized
from hexane in the glove box as described above. Com-
plexes 1 and 2 are readily accessible on a 10 g scale by ei-
ther preparative route.
The results presented here indicate that it is possible to
isolate pure solid bis(amido) chloride complexes such as 1
and 2 for the lighter lanthanides. However, in solution ligand
redistribution occurs readily so that mono and tris(amido)
species are also present. This does not necessarily preclude
the use of these complexes as precursors to other mixed
[Ce{N(SiMe3)2}2(THF)(µ-Cl)]2 (1)
Canary yellow prisms. Yield: 65%, mp. 151–153°C. Anal.
calcd. for C32H88N4O2Si8Cl2Ce2: C 33.81, H 7.80, N 4.93%;
found: C 33.39, H 7.94, N 4.78%.
3 The small difference in chemical shift between pure Ln[N(SiMe3)2]3 and that formed in the equilibrium of eq. [7] is likely a result of differ-
ences in the concentrations of the paramagnetic ions, which will exhibit differences in bulk susceptibility.
© 2000 NRC Canada

458
Can. J. Chem. Vol. 78, 2000
[Nd{N(SiMe3)2}2(THF)(µ-Cl)]2 (2)
tacts shorter than 3.5 Å were observed. The structural per-
spective plots were drawn using ZORTEP.6 Complete tables
of bond distances, angles, and anisotropic temperature pa-
rameters have been deposited as supplementary material.7
Deep lilac prisms. Yield: 81%, mp. 147–149°C.
MS(EI): m/z 499 (10%, Nd{N(SiMe3)2}2Cl+), 464 (100%,
+
Nd{N(SiMe3)2}2 ), 449 (10%, [Nd{N(SiMe3)2}2 – CH3]+),
434 (10%, [Nd{N(SiMe3)2}2 – 2 CH3]+), 419 (10%,
[Nd{N(SiMe3)2}2 – 3 CH3]+), 302 (10%, Nd{N(SiMe3)2}+)
amu. Anal. calcd. for C32H88N4O2Si8Cl2Nd2: C 33.56, H
7.75, N 4.89%; found: C 33.21, H 7.59, N 4.71%.
Acknowledgements
We thank Mrs. C. Greenwood for assistance in recording
the NMR spectra and Dr. David McGillivray for obtaining
the mass spectra. This research was supported by the Natural
Sciences and Engineering Research Council of Canada
(NSERC) and a University of Victoria Internal Research
Grant (to D.J.B.).
X-ray crystallographic studies
Crystallographic data for 2 is summarized in Table 1.
Crystals of 2 (0.5 × 0.5 × 0.5 mm) were loaded into glass
capillaries in the glove box and subsequently flame sealed.
The crystals were transferred to
a Nonius CAD4F
diffractometer equipped with MoKα radiation. The unit cell
of 2 was refined using 24 reflections in the 2θ range of 38–
42. An experimental density was not obtained because of the
air sensitivity of the compound. Six standard reflections,
measured periodically during data collection (9, 0, 7; 7, 0, 1;
8, 9, 0; 0, 7, 1; 0, 0, 8; 3, 5, 7) showed a 10% decline in in-
tensity during data collection. Intensity measurements were
collected over one quarter of the sphere. After the usual data
reduction procedures, including an absorption correction
(max/min transmission: 0.9939, 0.8507) according to a mea-
sured PSI scan, the structure was solved by Patterson meth-
ods.4 The refinements minimized w((Fo – Fc)2) and proceeded
normally using teXsan98.5 Criteria for inclusion of reflec-
tions were I > 3.5σ(I). The weighting scheme was deter-
mined by counting statistics using w = 1/(σ2(F) + 0.001F2).
Convergence was satisfactory: max shift/esd < 0.01; the
largest positive and negative peaks in the final difference
map were 1.64 and –0.97 e/Å3, respectively. A total of 227
parameters were refined. Hydrogen atoms were placed in
calculated positions. No non-hydrogen intermolecular con-
References
1. (a) S.A. Cotton, F.A. Hart, M.B. Hursthouse, and A.J. Welch.
J. Chem. Soc. Chem. Commun. 1225 (1979); (b) H. Schumann
and J. Müller. Angew. Chem. Int. Ed. Engl. 17, 276 (1978);
(c) A.L. Wayda and W.J. Evans. J. Am. Chem. Soc. 100, 7119
(1978); (d) H. Schumann. Angew. Chem. Int. Ed. Engl. 23,
474 (1984); (e) P.N. Hazin, J.W. Bruno, and G.K. Schulte.
Organometallics. 9, 416 (1990); (f) H. Schumann, P.R. Lee,
and J. Loebel. Chem. Ber. 122, 1897 (1989); (g) J. Holton,
M.F. Lappert, D.G. Ballard, R. Pearse, J.L. Atwood, and W.E.
Hunter. J. Chem. Soc. Dalton Trans. 4554 (1979).
2. T.D. Tilley, A. Zalkin, R.A. Andersen, and D.H. Templeton.
Inorg. Chem. 20, 551 (1981).
3. H.C. Aspinall, D.C. Bradley, M.B. Hursthouse, K.D. Sales,
N.P.C. Walker, and B. Hussain. J. Chem. Soc. Dalton Trans.
623 (1989).
4. R.D. Shannon. Acta Crystallogr. Sect. A: Cryst. Phys. Diffr.
Theor. Gen. Crystallogr. A32, 751 (1976).
5. D.C. Bradley, J.S. Ghotra, and F.A. Hart. J. Chem. Soc. Dalton
Trans. 1021 (1973).
4 P.T. Beurskens, G. Admiraal, G. Beurskens, W.P. Bosman, R. De Gelder, R. Israel, and J.M.M. Smits. DIRDIF94: The DIRDIF94 Program
System. Technical Report of the Crystallography Laboratory, University of Nijmegen, The Netherlands. 1994.
5 Molecular Structure Corporation. teXsan for Windows Version 1.05. The Woodlands, Texas. 1998.
6 L. Zsolnai. ZORTEP. University of Heidelberg, Germany. 1995.
7 Copies of material on deposit may be purchased from the Depository of Unpublished Data, Document Delivery, CISTI, National Research
Council of Canada, Ottawa, ON K1A 0S2, Canada.
© 2000 NRC Canada