Preparation, spectroscopic characterization, and deprotonation reactions of Si(NHR)4 (R = i-Pr, t-Bu, p-tolyl) - EPR identification of persistent radicals formed by oxidation of polyimidosilicates

Article abstract of DOI:10.1139/v05-264

Treatment of Cl₂Si(NH-t-Bu)₂ (6a) with t-BuNH ₂ in boiling toluene yields trisamino(chloro)silane ClSi(NH-t-Bu)₃ (7); formation of the tetraaminosilane Si(NH-t-Bu)₄ is not observed. The reaction of Si

Full text of DOI:10.1139/v05-264

21
Preparation, spectroscopic characterization, and
deprotonation reactions of Si(NHR)4 (R = i-Pr, t-Bu,
p-tolyl) — EPR identification of persistent radicals
formed by oxidation of polyimidosilicates
Andrea F. Armstrong, Tristram Chivers, and Jari Konu
Abstract: Treatment of Cl2Si(NH-t-Bu)2 (6a) with t-BuNH2 in boiling toluene yields trisamino(chloro)silane ClSi(NH-
t-Bu)3 (7); formation of the tetraaminosilane Si(NH-t-Bu)4 is not observed. The reaction of SiCl4 with 4 equiv. of
LiNHR produces the corresponding tetraaminosilanes Si(NHR)4 (2a, R = i-Pr; 2b, R = t-Bu; 2c, R = p-tol) in good
yields. When the sterically demanding adamantyl derivative LiHNAd is employed, only disubstitution occurs to form
Cl2Si(NHAd)2 (6b). Oxidation of the dimeric imidosilicates {Li3[Si(NR)3(NHR)]·THF}2 (3a, R = i-Pr; 3b, R = t-Bu)
with 1 mol of iodine produces the persistent radicals {Li2[Si(NR)3(NHR)]·LiI·3THF}·, which, on the basis of EPR
spectra, exist as SiN3Li3I cubes in solution. The spirocyclic tetraimidosilicate monoanion radical {[(THF)2Li(µ-
Nnaph)2Si(µ-Nnaph)2Li(THF)2]}–· (10) is formed upon oxidation of the tetralithiated species {Li4[Si(Nnaph)4]·4Et2O}
(1) and {[Li(12-crown-4)]2[(Et2O)2Li(µ-Nnaph)2Si(µ-Nnaph)2Li(Et2O)2]} (8) with iodine. The spectroscopic characteriza-
tion of hexa(amino)disiloxane (t-BuNH)3SiOSi(NH-t-Bu)3 (14) formed from the reaction of Cl3SiOSiCl3 with 6 equiv.
of LiNH-t-Bu is discussed.
Key words: imido ligands, silicate, radicals, EPR spectra, lithium.
Résumé : Le traitement du Cl2Si(NH-t-Bu)2 (6a) par de la t-BuNH2 dans du toluène bouillant conduit à la formation
de trisamino(chloro)silane, ClSi(NH-t-Bu)3 (7); on n’a pas observé de formation du tétraaminosilane, Si(NH-t-Bu)4. La
réaction du SiCl4 avec quatre équivalents de LiNHR conduit à la formation des tétraaminosilanes correspondants, Si(NHR)4
(2a, R = i-Pr; 2b, R = t-Bu; 2c, R = p-tol), avec de bons rendements. Lorsqu’on effectue la réaction avec le LiHNAd,
un dérivé de l’adamantane dont les contraintes stériques sont plus grandes, on ne se produit qu’une disubstitution
conduisant à la formation du Cl2Si(NHAd)2 (6b). L’oxydation des imidosilicates dimères {Li3[Si(NR)3(NHR)·THF}2 (3a,
R = i-Pr; 3b, R = t-Bu) avec une mole d’iode conduit à la formation des radicaux persistants {Li2[Si(NR)3-
(NHR)·LiI·3THF}· qui, sur la base de leurs spectres RPE, existeraient en solution sous la forme de cubes de SiN3Li3I.
L’oxydation par de l’iode des espèces tétralithiées {Li4[SiNnaph)4]·4Et2O} (1) et {Li[12-couronne-4]2[(Et2O)2Li(µ-
Nnaph)2Si(µ-Nnaph)2Li(Et2O)2]} (8) conduit à la formation du radical monoanion tétraimidosilicate spirocyclique
{[(THF)2Li(µ-Nnaph)2Si(µ-Nnaph)2Li(THF)2]}–· (10). On discute de la caractérisation spectroscopique de
l’hexa(amino)disiloxane (t-BuNH)3SiOSi(NH-t-Bu)3 (14) qui se forme par réaction du Cl3SiOSiCl3 avec six équivalents
de LiNH-t-Bu.
Mots clés : ligands imido, silicate, radicaux, spectre RPE, lithium.
[Traduit par la Rédaction] Armstrong et al. 28
Introduction
known; the coordination behaviour (6, 7) and radical chem-
istry (6, 7, 9, 10) of this important member of the family of
polyimido anions have been examined.
In contrast, the chemistry of the isoelectronic tetraimido-
silicate tetraanion ([Si(NR)4]4–) remains essentially un-
known; only one example, {Li4[Si(Nnaph)4]·4Et2O}x (1), has
been characterized (11). This naphthyl derivative was origi-
nally targeted because, unlike the majority of primary
amines, 1-aminonaphthalene can be doubly deprotonated
An ongoing area of research in main group chemistry is
the synthesis and characterization of imido [NR]2– (R = alkyl,
aryl) analogues of common p-block oxoanions such as sul-
fate and phosphate (1–3). The chemistry of the tetraimido-
phosphate trianion [P(NR)4]3–, formally isoelectronic2 with
orthophosphate [PO4]3–, has been extensively explored. Both
main group (4–7) and transition-metal (8) derivatives are
Received 2 August 2005. Published on the NRC Research Press Web site at http://canjchem.nrc.ca on 22 February 2006.
A.F. Armstrong, T. Chivers,1 and J. Konu. Department of Chemistry, University of Calgary, Calgary, AB T2N 1N4, Canada.
1Corresponding author (e-mail: chivers@ucalgary.ca).
2The original definition of “isoelectronic” (isosteric) compounds restricts the term to species having the same number of atoms and
the same number of electrons (I. Langmuir. J. Am. Chem. Soc. 41, 1543 (1919)). Thus, according to this definition, N3– is
isoelectronic with O2–. However, nitrides are compositionally different to oxides as a result of the different charges on the anions.
Can. J. Chem. 84: 21–28 (2006)
doi:10.1139/V05-264
© 2006 NRC Canada

22
Can. J. Chem. Vol. 84, 2006
Fig. 1. Solid-state structures of 3a, 4, and 13. The i-Pr (3a), t-Bu (4a, 13), and Ad (4b) groups are omitted from cluster N atoms for
clarity.
NH-i-Pr
Ph
NSiMe3
P
Si
Si
Li
N
N
N
N
N
N
Li
N
Li
N
N
N
N
(THF)
Li
N
Li
N
Li
N
Li
N
Li
N
Li
N
(THF)
Li
Li
Li
Li
Li
Li
N
Li
Li
Li
Si
Si
P
NH-i-Pr
NSiMe3
Ph
3a
13
4
Li2N
Et O
2
[1]
{Li4[Si(Nnaph)4] 4Et2O}
SiCl4 + 4
- 4 LiCl
(12, 13). The metathesis reaction between dilithiated amide
(Li2Nnaph) and silicon tetrachloride (SiCl4) in diethyl ether
proceeds with the elimination of lithium chloride (LiCl) to
yield 1 (eq. [1]) (11).
{Li3[P(NR)3(NSiMe3)]}2 (4a, R = t-Bu; 4b, R = Ad) with
halogens (9, 10). As the synthesis of radicals of the p-block
elements is a topical area of research (16), it will, thus, be of
interest to examine the one-electron oxidation of 1 to see if
similar behaviour is observed. Furthermore, the fact that 3a
is isostructural with 4 (Fig. 1) suggests that the oxidation of
3a may well result in the formation of cubic radicals that are
isostructural with the tetraimidophosphate radicals
{Li2[P(NR)3(NSiMe3)]·LiX·3THF}· (5a, R = t-Bu; 5b, R =
Ad; X = Cl, Br, I) described previously (9, 10).
In this contribution, we describe the synthesis and multi-
nuclear NMR spectroscopic characterization of (i) the
tetraaminosilanes Si(NHR)4 (2a, R = i-Pr; 2b, R = t-Bu; 2c,
R = p-tol), (ii) the aminochlorosilanes Cl2Si(NHAd)2 (6b)
and ClSi(NH-t-Bu)3 (7), and (iii) the hexa(amino)disiloxane
(t-BuNH)3SiOSi(NH-t-Bu)3 (14) and the tetraimidosilicate
complex {[Li(12-crown-4)]2[(Et2O)2Li(µ-Nnaph)2Si(µ-Nnaph)2-
Li(Et2O)2]} (8). We also report the characterization, by EPR
spectroscopy, of persistent radicals formed by oxidation of
tetramido- and aminotris(imido)-silicates with iodine.
Unfortunately, it has not proven possible to grow crystals
of 1 that are suitable for X-ray analysis; in consequence, the
solid-state structure of 1 remains unknown. In an effort to
structurally characterize a tetraimidosilicate, the preparation
of the isopropyl analogue of 1 was also attempted by the re-
action of tetraaminosilane Si(NH-i-Pr)4 (2a) with 4 equiv. of
n-butyllithium. Regardless of the reaction conditions, only
trilithiation was achieved, resulting in the isolation of
{Li3[Si(N-i-Pr)3(NH-i-Pr)]·THF}2 (3a) (11). This behaviour is
reminiscent of observations reported for trisamino(imino)phos-
phoranes, for which only twofold deprotonation of (i-
PrNH)3PNSiMe3 is observed under a variety of reaction con-
ditions (10). However, it should be noted that the reaction of
the tert-butyl analogue (t-BuNH)3PNSiMe3 with n-BuLi
yields the tetraimidophosphate [P(N-t-Bu)3(NSiMe3)]3– via
triple deprotonation under mild reaction conditions (5).
Thus, it is of interest to attempt the tetralithiation of Si(NH-
t-Bu)4 (2b) to determine whether a similar increase in reac-
tivity over 2a is observed.
Experimental
An examination of the extant literature (14) regarding 2b
revealed that no practical synthetic route to this compound
has been reported, though 2a is well-known and can be pre-
pared by the reaction of excess isopropylamine with SiCl4
(15). However, only a 6% yield of the desired product 2b is
recovered from the analogous reaction of SiCl4 with excess
tert-butylamine, even at elevated reaction temperatures (14).
Thus, the synthesis of 2b has been reinvestigated, as only
minimal experimental detail is available from previous work.
The isoelectronic relationship between the tetraimidosilicate
tetraanion [Si(NR)4]4– of 1 and the tetraimidophosphate
trianion [P(NR)4]3– is especially intriguing in light of the re-
markably stable radicals formed upon oxidation of
Reagents and general procedures
All experiments were carried out under an argon atmo-
sphere using standard Schlenk techniques. Toluene, diethyl
ether, and tetrahydrofuran (THF) were dried over Na–benzo-
phenone, distilled, and stored over molecular sieves prior to
use. 12-Crown-4 (98%), Cl3SiOSiCl3 (>98%), n-BuLi
(2.5 mol/L solution in hexanes), tert-butylamine, and SiCl4
(99%) were used as received from Aldrich. The lithiated
amides LiHNR (R = i-Pr, t-Bu, p-tol) were prepared by stan-
dard procedures involving the reaction of the corresponding
primary amines with n-BuLi in hexane (i-Pr, t-Bu) or Et2O
(p-tol) at –78 °C. Iodine was sublimed prior to use. Com-
© 2006 NRC Canada

Armstrong et al.
23
pounds 1 (11), 3b (11), and 13 (17) were prepared by the
methods reported in the literature.
an additional 15 h. The pale orange slurry was centrifuged;
the supernatant liquid was filtered, leaving a yellow filtrate.
The solvent was removed in vacuo, giving Si(NH-p-tol)4 as
1
a pale orange powder (0.338 g, 0.747 mmol, 78%). H NMR
Instrumentation
(C6D6) δ: 6.90, 6.71 (d of d, 4H, C6H4, 2J(1H,1H) = 76.8 Hz,
3J(1H,1H) = 8.2 Hz), 3.80 (s, 1H, NH), 2.11 (s, 3H, p-CH3).
13C{1H} NMR (C6D6) δ: 142.66 (s, C-NH), 130.22 (s, o-
CH), 128.53 (m-CH), 117.59 (C-CH3), 20.53 (CH3).
29Si{1H} NMR (C6D6) δ: –52.9 (s). Anal. calcd. for
C28H32N4Si: C 74.29, H 7.13, N 12.38; found: C 73.22, H
7.13, N 11.91.
7
1H, Li, 13C, 29Si, and 31P NMR spectra were collected on
1
a Bruker DRX-400 spectrometer with H and 13C chemical
shifts referenced internally to 1H impurities and natural
abundance 13C in the deuterated solvents, respectively; exter-
nal standards were used for the 7Li (LiCl in D2O), 29Si
(SiMe4 in CDCl3), and 31P (85% H3PO4 in D2O) spectra. All
spectra were collected at 22 °C (unless otherwise noted). IR
spectra were recorded as Nujol or Fluorolube mulls on KBr
plates using a Nicolet Nexus 470 FT IR spectrometer in the
range 4000–400 cm–1. Mass spectra were recorded on a
Bruker Esquire 3000 ESI ion trap mass spectrometer. Ele-
mental analyses were provided by the Analytical Services
Laboratory, Department of Chemistry, University of Calgary,
Calgary, Alberta and by Canadian Microanalytical Service
Ltd., Vancouver, British Columbia. EPR spectra were re-
corded at 22 °C on a Bruker EMX 113 spectrometer; spec-
tral simulations were carried out using WINEPR SimFonia;
spectra were plotted using WinEPR (18). The g values of the
radicals were calculated using the field–frequency ratio of
each sample.
Preparation of Cl2Si(NH-t-Bu)2 (6a)
A solution of SiCl4 (4.0 mL, 34.9 mmol) in diethyl ether
(20 mL) was added to a solution of t-BuNH2 (25.0 mL,
0.238 mol) in diethyl ether (150 mL) at 22 °C over 5 min,
resulting in a dense white slurry. The reaction mixture was
stirred for 16 h, then filtered through a sintered glass frit,
yielding a clear colourless filtrate. The white solid (t-
BuNH3)Cl was washed with diethyl ether (100 mL). The
solvent was removed from the filtrate in vacuo, giving
Cl2Si(NH-t-Bu)2 as
a
colourless liquid (7.024 g,
1
28.88 mmol, 83%). H NMR (C6D6) δ: 1.48 (br, NH), 1.11
(s, t-Bu). 13C{1H} NMR (C6D6) δ: 50.61 (s, CMe3), 32.63
(s, CMe3). 29Si{1H} NMR (C6D6) δ: –38.9 (s). Literature
values (from ref. 18): 1H NMR (C6D6) δ: 1.49, 1.11.
13C{1H} NMR (C6D6) δ: 50.6, 32.5.
Preparation of Si(NH-i-Pr)4 (2a)
SiCl4 (1.3 mL, 1.93 g, 11.3 mmol) was added dropwise to
a white slurry of LiHN-i-Pr (3.040 g, 46.74 mmol) in diethyl
ether (50 mL) at 0 °C. The reaction mixture was stirred for
3 h, then warmed slowly to 22 °C and stirred for an addi-
tional 15 h. The pale yellow slurry was centrifuged; the
supernatant liquid was filtered, leaving a colourless filtrate.
The solvent was removed in vacuo and the product was
washed with hexane (5 mL), giving Si(NH-i-Pr)4 as a white
Preparation of Cl2Si(NHAd)2 (6b)
SiCl4 (0.26 mL, 2.27 mmol) was added to a white slurry
of LiHNAd (0.714 g, 4.54 mmol) in toluene (30 mL) at
0 °C. The reaction mixture was allowed to warm to 22 °C
and stirring was continued for 28 h. The reaction mixture
was filtered and the solvent was removed from the filtrate in
vacuo, giving Cl2Si(NHAd)2 as a white solid (0.731 g,
1
powder (2.292 g, 8.80 mmol, 78%). H NMR (C6D6) δ: 3.22
3
1
(sept, 1H, CH3-CH, J(H,H) = 7 Hz), 1.10 (d, 6H, CH3-CH,
1.83 mmol, 81%). H NMR (C6D6) δ: 1.90 (br, 1H, CH),
3J(H,H) = 7 Hz, Me). 13C{1H} NMR (C6D6) δ: 42.65 (s,
1.79 (d, 2H, CH2-C-NH), 1.49 (t, 2H, CH-CH2-CH).
13C{1H} NMR (C6D6) δ: 51.20 (s, C-NH), 46.25 (s, CH),
36.45 (s, CH2-C-NH), 30.35 (s, CH-CH2-CH). 29Si{1H}
NMR (C6D6) δ: –38.8 (s). Anal. calcd. for C20H30N2SiCl2: C
60.13, H 8.07, N 7.01; found: C 59.85, H 8.24, N 7.13.
1
CH), 28.34 (s, CH3). Literature values (from ref. 15): H
3
NMR (C6D6) δ: 3.21 (sept, 1H, CH3-CH, J(H,H) = 6.7 Hz),
1.10 (d, 6H, CH3-CH, 3J(H,H) = 6.7 Hz, Me). 13C{1H}
NMR (C6D6) δ: 42.6 (s, CH), 28.3 (s, CH3).
Preparation of Si(NH-t-Bu)4 (2b)
Preparation of ClSi(NH-t-Bu)3 (7)
A solution of SiCl4 (0.79 mL, 6.90 mmol) in diethyl ether
(15 mL) was added to a white slurry of LiHN-t-Bu (2.191 g,
27.71 mmol) in diethyl ether (50 mL) at 0 °C. The reaction
mixture was stirred for 1 h, then warmed slowly to 22 °C
and stirred for an additional 3.5 h. The white slurry was fil-
tered, and the solvent was removed from the colorless fil-
trate in vacuo, giving Si(NH-t-Bu)4 as a white powder
tert-Butylamine (10.0 mL, 95.16 mmol) was added to a
solution of Cl2Si(NH-t-Bu)2 (3.078 g, 12.65 mmol) in tolu-
ene (20 mL) at 22 °C, resulting in a clear colorless solution.
The reaction mixture was refluxed at 113 °C for 24 h. The
resulting white slurry was cooled to 22 °C and filtered
through a sintered glass frit, resulting in a clear colorless fil-
trate. The white solid (t-BuNH3·HCl) was washed with tolu-
ene (5 mL). The solvent was removed from the filtrate in
vacuo, giving spectroscopically pure ClSi(NH-t-Bu)3 as a
colourless liquid (3.138 g, 11.21 mmol, 89%). 1H NMR
(C6D6) δ: 1.20 (s, t-Bu). 13C{1H} NMR (C6D6) δ: 49.47 (s,
CMe3), 33.28 (s, CMe3). 29Si{1H} NMR (C6D6) δ: –45.2 (s).
1
(1.830 g, 5.78 mmol, 84%). H NMR (C6D6) δ: 1.28 (s, t-
Bu). 13C{1H} NMR (C6D6) δ: 48.69 (s, CMe3), 33.99 (s,
CMe3). 29Si{1H} NMR (C6D6) δ: –52.9 (s). MS (EI): 316.6
([M]+). Anal. calcd. for C16H40N4Si: C 60.70, H 12.73, N
17.70; found: C 60.28, H 11.47, N 17.97.
Preparation of Si(NH-p-tol)4 (2c)
Preparation of {[Li(12-crown-4)]2[(Et2O)2Li(␮-
Nnaph)2Si(␮-Nnaph)2Li(Et2O)2]} (8)
n-Butyllithium (4.65 mL, 11.62 mmol) was added to a so-
lution of 1-aminonaphthalene (0.833 g, 5.817 mmol) in di-
ethyl ether (30 mL) at –78 °C and stirred for 1 h. SiCl4
SiCl4 (0.11 mL, 0.16 g, 0.96 mmol) was added dropwise
to a white slurry of LiHN-p-tol·Et2O (0.606 g, 3.723 mmol)
in diethyl ether (35 mL) at 0 °C. The reaction mixture was
stirred for 2 h, then warmed slowly to 22 °C and stirred for
© 2006 NRC Canada

24
Can. J. Chem. Vol. 84, 2006
(0.17 mL, 1.484 mmol) was added to the reaction mixture
dropwise. After 30 min, the reaction mixture was warmed to
22 °C and stirred for an additional 24 h. The resulting yel-
low slurry was then filtered, yielding a clear yellow solution.
12-Crown-4 (0.48 mL, 3.024 mmol) was added to the solu-
tion at 22 °C, resulting in a yellow slurry, which was stirred
for 1 h. The solvent was then removed in vacuo, giving 8 as
while the methyl groups resonate at lower field (33.3 vs.
32.6 ppm).
Attempts to prepare 2b by the reaction of 7 with an excess
of tert-butylamine at elevated temperatures were unsuccess-
1
ful; in all cases, no reaction could be detected by H NMR
spectroscopy. In consequence, an alternative synthetic route,
based on the method by which the trisamino(imino)phos-
phoranes (RNH)3PNSiMe3 (R = i-Pr, Cy, t-Bu, Ad) have
been prepared (5, 10) (eq. [2]), was adopted whereby SiCl4
was treated with 4 equiv. of LiHN-t-Bu at 22 °C (eq. [3]).
This reaction proceeds with the elimination of 4 equiv. of
LiCl and produces the desired compound 2b in 84% yield.
The identity of 2b was established by multinuclear NMR,
mass spectrometry (M+ at 316.6), and elemental (C, H, N)
analysis. The trends in multinuclear NMR data that were ob-
served on moving from the bisamino compound 6a to the
trisamino species 7 continue with this tetraaminosilane: the
1H NMR signal is at still lower field (1.28 ppm) as is the
CMe3 peak in the 13C NMR spectrum (34.0 ppm), while the
second 13C NMR resonance (48.7 ppm) and the 29Si NMR
signal (–52.9 ppm) appear at higher field. No trace of con-
tamination from unreacted 6a or 7 was detected in the NMR
spectra of this compound.
1
a bright yellow powder (1.489 g, 1.166 mmol, 80%). H
NMR (d8-THF) δ: 8.5 (br), 8.10 (d, 4H), 7.44 (d, 4H), 7.09
(t, 4H), 7.02 (t, 4H), 6.90 (t, 4H), 6.40 (t, 4H), 3.40 (q, 16H,
Et2O), 3.33 (s, 24H, 12-crown-4), 1.14 (t, 24H, Et2O).
13C{1H} NMR (d8-THF) δ: 162.37, 137.44, 128.61, 128.24,
127.06, 124.63, 124.19, 120.70, 108.46, 106.92, 68.82 (s,
12-crown-4), 65.52 (s, (CH3CH2)2O), 14.59 (s,
7
(CH3CH2)2O). Li NMR (d8-THF) δ: 0.63 (s). Anal. calcd.
for C72H100N4SiO12Li4: C 68.12, H 7.94, N 4.41; found: C
67.20, H 8.00, N 4.83.
Preparation of (t-BuNH)3SiOSi(NH-t-Bu)3 (14)
A solution of Cl3SiOSiCl3 (1.479 g, 5.191 mmol) in di-
ethyl ether (10 mL) was added to a white slurry of LiHN-t-
Bu (2.463 g, 31.15 mmol) in diethyl ether (40 mL) at 0 °C.
After 30 min, the white slurry was warmed to 22 °C and
stirred for an additional 15 h. The resulting pale yellow
slurry was filtered and the solvent was removed from the fil-
trate in vacuo, giving (t-BuNH)3SiOSi(NH-t-Bu)3 as a pale
yellow solid (1.770 g, 3.505 mmol, 68%). Sublimation of
the crude product at 135 °C under reduced pressure resulted
in the isolation of a white powder (1.177 g, 2.331 mmol,
45%). IR (cm–1) υ: 3412 (m, (N-H)), 1230 (vs, (Si-O-Si)).
1H NMR (C6D6) δ: 1.34 (s, t-Bu), 1.31 (s, t-Bu, minor reso-
nance) (relative intensities ca. 5:1), 1.21 (NH). 13C{1H}
NMR (C6D6) δ: 49.10 (s, CMe3), 33.84 (s, CMe3), 48.80,
33.87 (minor resonances). 29Si{1H} NMR (C6D6) δ: –54.5
(s). Anal. calcd. for C24H60N6Si2O (14): C 57.09, H 11.98, N
16.64; C28H71N7Si2O (14 + t-BuNH2): C 58.17, H 12.38, N
16.96; found: C 56.32, H 12.19, N 15.99, Cl < 0.3.
[2]
Cl3PNSiMe3 + 3LiHNR
Et2O
⎯⎯⎯→ (RNH)3PNSiMe3 + 3LiCl
0 °C
[3]
SiCl4 + 4LiHN-t-Bu
Et2O
⎯⎯⎯→ Si(NH-t-Bu)4 + 4LiCl
0 °C
The ease with which this reaction (eq. [3]) occurs indi-
cates that electronic effects rather than the steric influence of
the bulky t-BuNH2 group, as previously postulated (14), pre-
clude the formation of 2b in solvothermal reactions between
SiCl4 and t-BuNH2. The formation of LiCl presumably pro-
vides a sufficient driving force for complete aminolysis to
occur. The utility of this synthetic approach is not limited to
the tert-butyl group: the known derivatives 2a and 2c have
also been prepared in ca. 80% yields by this procedure and
identified by multinuclear NMR spectra.
Results and discussion
Preparation and NMR characterization of
Cl2Si(NHAd)2 (6b), ClSi(NH-t-Bu)3 (7), and Si(NHR)4
(2a, R = i-Pr; 2b, R = t-Bu; 2c, R = p-tol)
Following the success of this synthetic methodology for
producing 2b, it was of interest to determine whether other,
perhaps even more sterically hindered tetraaminosilanes,
could be prepared in the same manner. To that end, the syn-
thesis of “Si(NHAd)4” (2d) was attempted via the analogous
reaction of SiCl4 with lithium adamantylamide (LiHNAd).
However, the disubstituted compound 6b was isolated in
81% yield in this case. As the electron-withdrawing effects
of NHAd and NH-t-Bu groups are expected to be quite simi-
lar to one another, this failure to generate tris- or tetra-
(adamantylamino)silanes must be attributed to the greater
steric bulk of the adamantyl cages as compared with that of
the tert-butyl substituents of 2b.
With only limited information available (14) regarding
previous attempts to synthesize 2b, the reaction of SiCl4
with t-BuNH2 was reinvestigated. As reported in the litera-
ture (19), treatment of SiCl4 with an excess of tert-
butylamine at room temperature results in the aminolysis of
two silicon–chlorine bonds and the formation of 6a in 83%
yield. Further treatment of 6a with an excess of t-BuNH2 in
boiling toluene resulted in the formation of 7, a colourless
liquid, in 89% yield. While several physical properties of 7
have been determined, including the boiling point and re-
fractive index (14), no spectroscopic data for this compound
1
have been reported. The tert-butyl H NMR resonance of 7
Attempted preparation of {Li4[Si(NR)4]} (R = t-Bu, p-
tol) — Synthesis and NMR characterization of {[Li(12-
crown-4)]2[(Et2O)2Li(µ-Nnaph)2Si(µ-Nnaph)2Li(Et2O)2]} (8)
With an efficient synthetic route to Si(NH-t-Bu)4 estab-
lished, an attempt to tetralithiate this species with n-BuLi
was carried out. The two reagents were combined in hexane
(1.20 ppm) is shifted slightly downfield as compared with
that of 6a (1.11 ppm), while the 29Si NMR signal (–45.2 ppm)
of 7 appears at higher field than the corresponding resonance
of 6a (–38.9 ppm). The α-carbon signal in the 13C NMR
spectrum of 7 is shifted to higher field (49.5 vs. 50.6 ppm),
© 2006 NRC Canada

Armstrong et al.
25
1
at 22 °C. The H NMR spectrum of the product of this reac-
naph
N
naph
N
O
O
O
O
tion, a fine white powder, contained two t-Bu resonances at
1.64 and 1.48 ppm in a 1:3 ratio, as well as a broad signal
attributable to a NH fragment; unreacted n-BuLi was also
detected. These spectroscopic data are consistent with the
formation of a trilithiated species {Li3[Si(N-t-Bu)3(NH-t-
Bu)]}2 (3b) (11), which is isostructural with 3a. The reac-
tion was repeated under various conditions. In all cases, a
significant amount of unreacted n-BuLi was observed in the
1H NMR spectrum of the product, along with a broad signal
that was attributed to a NH fragment, the presence of which
was confirmed by IR spectroscopy. Reactions between 2b
and Mg-n-Bu2 under a variety of conditions also resulted in
incomplete deprotonation, as revealed by sharp N–H stretch-
ing vibrations in the IR spectra of the products.
(Et2O)2Li
Si
Li(OEt2)2
Li
2
N
naph
N
naph
8
spirocyclic geometry has been observed in the solid state for
the isoelectronic tetra(naphthylimido)phosphate {[Li(TH-
F)4][(THF)2-Li(µ-Nnaph)2P(µ-Nnaph)2Li(THF)2]} (9a) (4) and
in solution for the tert-butyl analogue {[Li(THF)4]{Li(TH-
F)2[(µ-N-t-Bu)2P(µ-N-t-Bu)2]Li(THF)2} (9b) (10).
R
N
R
N
In light of the limited reactivity of 2b toward n-BuLi, it
was decided to undertake similar reactions with Si(NH-p-
tol)4 (2c), as the increased acidity of the NH protons of
arylamino substituents compared with that of alkylamino
groups has been observed to facilitate complete lithiation in
related trisamino(thio)phosphate systems (20). The spectro-
scopic data obtained from these experiments were inconclu-
sive, but did not provide any strong evidence to support the
formation of “{Li4[Si(N-p-tol)4]}”. Based on these results, it
is proposed that fourfold deprotonation of Si(NHR)4 (R =
alkyl, aryl) is not possible using alkyl lithium reagents, as
they are not sufficiently basic to effect the removal of the
fourth proton. The synthesis of 1 requires the double
deprotonation of 1-aminonaphthalene prior to reaction with
SiCl4 (11); since H2Nnaph is the only primary amine that
can be dilithiated, the naphthyl derivative 1 may be a unique
example of a tetralithiated tetraimidosilicate.
[Li(THF)4]
(THF)2Li
P
Li(THF)2
N
R
N
R
9
Surprisingly, only one signal (0.63 ppm) was observed in
7
the Li NMR spectrum of 8 in d8-THF, indicating that the
chemically inequivalent lithium cations exchange rapidly in
solution despite the presence of 12-crown-4. This can be ra-
tionalized by taking into account the presence of four highly
nucleophilic imido groups and the large negative charge of
the tetraimidosilicate tetraanion. These two factors combine
to render chelation by the [Si(Nnaph)4]4– preferable to, or at
least competitive with, an η mode of coordination by the
crown ether. A variable-temperature Li NMR study in d8-
4
7
THF showed that the spectrum is highly dependent upon
temperature, confirming that the lithium cations are in dy-
namic equilibrium in solution. The single resonance broad-
ens significantly even at 0 °C, then resolves into several
sharp resonances at lower temperatures (–45 to –60 °C). It is
expected that further lowering of the temperature would
eventually result in two singlets in a 1:1 ratio attributed to
the N,N′-chelated lithium ions and the [Li(12-crown-4)]+ cat-
ions. However, this resolution could not be achieved above
the temperature at which the NMR solvent began to gel.
With the library of viable R groups thus curtailed, the
preparation of a crown ether complex of 1 was undertaken,
with the expectation that the resulting material would be
more crystalline than the diethyl etherate 1. The synthesis of
1 by the literature procedure (11) was modified by the addi-
tion of 2 equiv. of 12-crown-4 after filtration to remove the
LiCl by-product. The resulting bright yellow powder was
1
characterized by H and 13C NMR spectroscopy, as well as
by an elemental (C, H, N) analysis. The overlapping
1
multiplets in the aromatic region of the H NMR spectrum
of the product render it difficult to interpret. However, only
10 distinct signals were observed in the aromatic region
(106.9–162.4 ppm) of the 13C NMR spectrum, in addition to
the resonances for 12-crown-4 and diethyl ether ligands, in-
dicating the equivalence of the four Nnaph substituents.
Based on these data, this novel tetraimidosilicate complex is
thought to exist in a spirocyclic conformation {[Li(12-
crown-4)]2[(Et2O)2Li(µ-Nnaph)2Si(µ-Nnaph)2Li(Et2O)2]} (8),
with two Li+ cations sequestered by 12-crown-4; the remain-
ing two Li+ cations are each N,N′-chelated by the
naphthylimido groups and bis-solvated by two diethyl ether
ligands.
Formation and EPR characterization of persistent
polyimidosilicate radicals
Oxidation reactions of the tetraimidosilicate 1 were ex-
plored to determine whether 1 is capable of forming stable
radicals,3 in a manner similar to that reported for the
isoelectronic tetraimidophosphates {Li3[P(NR)3(NSiMe3)]}2
(4a, R = t-Bu; 4b, R = Ad) (6, 9, 10). The addition of
0.5 mol of iodine to a bright yellow solution of 1 in THF re-
sulted in a deep green-brown solution, indicating the forma-
tion of a paramagnetic species, presumably a lithiated
derivative of trianion radical [Si(Nnaph)4]3–·. An EPR spec-
trum of this solution (Fig. 2) at 22 °C displayed a single
broad resonance (ca. 15 G), with no resolved hyperfine cou-
pling (HFC) to the four nitrogen nuclei of the tetraimido-
silicate radical. Though this was initially surprising, it
Though single crystals of 8 are readily obtained from
THF–hexane solutions, they are not of sufficient quality for
X-ray analysis; thus, the proposed structure of 8 has not
been confirmed in the solid state. However, the proposed
3 In this article, a “stable” radical is one that is inherently stable as an isolated species and does not decompose under an inert atmosphere at
room temperature (16). A “persistent” radical is one that has a relatively long lifetime under the conditions that are used to generate it; radi-
cals less stable than this are termed “transient”.
© 2006 NRC Canada

26
Can. J. Chem. Vol. 84, 2006
Fig. 2. Experimental EPR spectra of iodine-oxidized 1 (top) and
10 (middle); simulated EPR spectrum of 10 (bottom).
Scheme 1. Oxidation of 3a by iodine to form two cubic radicals
12 (R = i-Pr).
NH-i-Pr
Si
Li
NH-i-Pr
Si
RN
Li
NR
Li
R
NR
Li
R
N
N
N
R
THF
Li
N
Li
+
I2
2
R
N
Li
Li
Li
Si
NH-i-Pr
R
NR
I
rect measurement of this coupling. The striking difference
between the line widths produced by the two symmetrical
spirocyclic radicals 10 (ca. 3 G) and 11 (ca. 0.5 G) is
thought to be caused by spin delocalization into the naphthyl
groups of 10, and the resultant presence of numerous
1
inequivalent H hyperfine couplings. This is not an unrea-
sonable hypothesis, since Li2Nnaph is reported to be para-
magnetic and the EPR spectrum contains sufficiently large
1H HFCCs to produce a signal that is approximately 17 G in
width (12).
While the green colour formed upon oxidation of 1 and 8
persists for more than 24 h in both instances, the EPR sig-
nals were surprisingly weak, indicating the presence of only
a small amount of paramagnetic material in the solutions.
This observation suggests that either the tetraimidosilicate
radical has only a short lifetime or that radical formation is
not quantitative, i.e., that the radical 10 is not the exclusive
product of these oxidation reactions. Attempts to isolate the
tetraimidosilicate radical were unsuccessful owing to the
limited stability of this persistent radical. Overall, the stabil-
ity of 10 is comparable to that of the isostructural neutral
radical 11 (10).
Previous work with tetraimidophosphate radicals has
shown that cubic derivatives are significantly more stable
than spirocycles such as 10 and 11 (6, 7, 9, 10). Oxidation of
the dimeric aminotris(imido)silicate 3a with iodine is ex-
pected to proceed in an analogous fashion to oxidations of
the structurally similar cluster 4. As depicted in Scheme 1,
this process can be viewed to involve disruption of the Li6N6
core of 3a, yielding two molecules of the cubane radical
{Li2[Si(N-i-Pr)3(NH-i-Pr)]·LiI·3THF}· (12). In practice, the
addition of 1 equiv. of I2 to a slurry of 3a in THF yields an
intensely yellow solution that gives rise to a somewhat stron-
ger EPR signal than do the products of the corresponding re-
actions with 1 and 8. The EPR spectrum of the reaction
mixture is primarily a seven-line pattern with multiple
smaller HFCs present as well (Fig. 3); the g value of this
radical is 2.0044. The best simulation of this spectrum was
obtained by including HFCCs to three equivalent nitrogen
atoms (8.1 G), one unique nitrogen atom (2.0 G), three lith-
3460
3470
3480
3490
[G]
3500
3510
should be noted that extensive line broadening is often ob-
served in fluxional radical systems: the variable-temperature
7Li NMR data obtained for the related crown ether complex
8 are certainly indicative of considerable solution dynamics,
and lithium exchange is expected to occur still more readily
in the absence of any crown ether.
In an attempt to resolve the nitrogen and lithium HFCs in
the EPR spectrum of the [Si(Nnaph)4]3–· radical, the crown
ether complex 8, believed to be less fluxional than 1, was
oxidized with iodine. An EPR spectrum (Fig. 2) of this dark-
green reaction mixture revealed a nine-line pattern, which is
consistent with coupling to four equivalent 14N (I = 1, 99.6%)
nuclei; the g value of this radical is 2.0037. The pronounced
“S-curve” shape of the spectrum and the broadness of the
spectral lines are attributed to unresolved HFC (21) to the
lithium cations (7Li, I = 3/2, 92.6%) in the spirocyclic radical
monoanion {[(THF)2Li(µ-Nnaph)2Si(µ-Nnaph)2Li(THF)2]}–· (10).
A good simulation of this spectrum (Fig. 2) was obtained
by using hyperfine coupling constants (HFCCs) of 3.8 G to
four nitrogen atoms and 0.7 G to two lithium cations. These
HFCCs are slightly smaller than those observed for the
isostructural tetraimidophosphate radical [(THF)2Li(µ-N-t-
Bu)2P(µ-N-t-Bu)Li(THF)2]· (11) for which the corresponding
HFCCs are 4.30 G (14N) and 1.08 G (7Li) (10). However, the
value of the 7Li HFCC for 10 should be regarded as approxi-
mate, since the broadness of the spectral lines precludes di-
naph
N
naph
N
t-Bu
N
t-Bu
N
(THF)2Li
Si
Li(THF)2
(THF)2Li
P
Li(THF)2
N
naph
N
naph
N
t-Bu
N
t-Bu
10
11
© 2006 NRC Canada

Armstrong et al.
27
Fig. 3. Experimental (top) and simulated (bottom) EPR spectra
of 12.
Preparation and characterization of (t-BuNH)3-
SiOSi(NH-t-Bu)3 (14)
As an alternative precursor to polyimidosilicate radicals,
we targetted hexa(tert-butylamino)disiloxane (t-BuNH)3-
SiOSi(NH-t-Bu)3 (14), which was synthesized by the reac-
tion of hexachlorodisiloxane with 6 equiv. of LiHN-t-Bu at
0 °C in diethyl ether (eq. [4]) and purified by sublimation.
The resultant white powder was characterized by IR and
multinuclear NMR spectra, GC–MS, and elemental analyses.
[4]
Cl3SiOSiCl3 + 6LiHN-t-Bu
Et2O
⎯⎯⎯→ (t-BuNH)3SiOSi(NH-t-Bu)3 + 6LiCl
1
The H NMR spectrum of the product exhibits a minor
resonance at 1.31 ppm in addition to the major peak at
1.34 ppm. The integrated relative intensities of these two
resonances exhibit an approximately 1:5 ratio. The 13C NMR
spectrum is consistent with this observation, showing one
major set of tert-butyl resonance (49.10 and 33.84 ppm) and
a second set of minor resonances at 48.80 and 33.87 ppm.
These minor resonances, with the same relative intensities,
persist in the NMR spectra of samples purified by either
sublimation or recrystallization. The 29Si NMR spectrum of
the product exhibits a single sharp resonance at –54.6 ppm.
Furthermore, the GC–MS revealed only a single component.
ium cations (0.9 G), and an iodine atom (0.4 G) (Fig. 3).
Thus, this EPR spectrum is consistent with the formation of
the cubic radical 12 and confirms that 3a undergoes an anal-
ogous reaction to 4 when oxidized.
1
Since the two signals in the H NMR spectrum are very
close together, the integration of the relative intensities must
be regarded as approximate, and we have considered the
possibility that the ratio of the intensities is 1:6 with the
smaller resonance attributed to a component with a single t-
BuN group. The possible candidates for the “single t-BuN”
species include t-BuNHLi, t-BuNH2, or [t-BuNH3]Cl, each
of which could form a hydrogen-bonded adduct with the O
atom of 14. The product shows no resonance in the 7Li
NMR spectrum, thus ruling out t-BuNHLi as an additional
component. Elemental analyses show that the product con-
tains insignificant amounts of Cl (<0.3%), thus excluding [t-
BuNH3]Cl as an impurity. Although we cannot count out the
possibility of one molecule of t-BuNH2 as a hydrogen-
bonded component of the product, the IR spectrum
(Fluorolube mull) shows only extremely weak, broad bands
in the 3120–3240 cm–1 region, which could be assigned to
hydrogen-bonded t-BuNH2, in addition to a sharp, medium-
intensity band at 3412 cm–1 attributed to υ(N-H) of 14. We
also note that the CHN analyses are in better agreement with
the calculated values for 14 than with the alternative (14 + t-
BuNH2). We are led to the tentative conclusion, therefore,
that the best explanation of the combined spectroscopic and
analytical data for 14 involves some sort of intermolecular
interaction in solution that leads to inequivalence of one of
the N-t-Bu groups. Unfortunately, numerous crystals of 14
obtained by recrystallization or sublimation were only
weakly diffracting, and consequently, we have been unable
to obtain a solid-state structure by X-ray crystallography. In
view of the uncertainty about the structure of 14, we have
not pursued the hexalithiation.
Like the tetraimidosilicate radical 10, the amido-
tris(imido)silicate 12 is considered to be a persistent radical,
since it decomposes in solution within 72 h.3 This is some-
what surprising, as cubic tetraimidophosphate radicals are
more stable than related spirocyclic species and they can be
isolated and structurally characterized in the solid state (6, 7,
9, 10). However, it must be noted that the isopropyl substitu-
ents of 10 provide only modest steric protection to prevent
dimerization of the radical: stable tetraimidophosphate radi-
cals occur only when bulkier tert-butyl or adamantyl groups
are in place (10). Thus, oxidation of the tert-butyl derivative
3b with 1 mol of iodine was undertaken, with the expecta-
tion that a highly persistent or stable radical would result.
Surprisingly, the paramagnetic species generated from this
reaction was found to be even less stable than 12, persisting
in solution for only a few seconds. An EPR spectrum of the
reaction mixture produced a very weak signal, consisting of
a poorly resolved seven-line pattern. While this suggests
HFC of the unpaired electron to three equivalent 14N nuclei,
consistent with the formation of a cubic radical, the experi-
mental parameters used to acquire this spectrum were not
optimal owing to the low sample concentration. Oxidation of
the related dimer {Li3[PhSi(N-t-Bu)3]}2 (13) (17), an or-
ganic derivative of a trisimidosilicate that exists as a Li6N6
hexagonal prism bicapped by the two Si–Ph units (cf. 3a in
Fig. 1), also yielded a transient radical. The lower stability
of this radical as compared with that of 12 can be attributed
to a combination of the smaller steric bulk of the phenyl
group (cf. the NH-i-Pr unit), combined with the fewer num-
ber of electronegative (i.e., N) atoms over which the un-
paired electron can be delocalized. However, no explanation
can be given at the present time as to why the oxidation of
the more sterically protected 3b leads to less stable radicals
than the corresponding reactions of the isopropyl derivative
3a.
Conclusions
The reaction of SiCl4 with lithiated primary amides
LiHNR (R = alkyl, aryl) in diethyl ether provides a versatile
© 2006 NRC Canada

28
Can. J. Chem. Vol. 84, 2006
5. A. Armstrong, T. Chivers, M. Krahn, M. Parvez, and G.
Schatte. Chem. Commun. (Cambridge), 2332 (2002).
6. A. Armstrong, T. Chivers, M. Parvez, and R.T. Boeré. Inorg.
Chem. 43, 3453 (2004).
and efficient route to tetraaminosilanes Si(NHR)4. Fourfold
deprotonation of Si(NHR)4 does not appear to be possible
using either n-BuLi or MgBu2, suggesting that the prepara-
tion of tetraimidosilicates [Si(NR)4]4– can only be achieved
if the amine is doubly deprotonated prior to reaction with
SiCl4, i.e., by the use of Li2NR. EPR studies of the products
of the one-electron oxidation of tetraimido or amido-
tris(imido)silicates with iodine indicate the formation of per-
sistent spirocyclic and cubic radicals, respectively. The
reasons for the lower stability of these polyimidosilicate rad-
icals as compared with that of the analogous tetraimido-
phosphate radicals are not yet fully understood and merit
further investigation.
7. A.F. Armstrong, T. Chivers, H.M. Tuononen, and M. Parvez.
Inorg. Chem. 44, 5778 (2005).
8. A. Armstrong, T. Chivers, M. Krahn, and M. Parvez. Can. J.
Chem. 83, 1768 (2005).
9. A. Armstrong, T. Chivers, M. Parvez, and R.T. Boeré. Angew.
Chem. Int. Ed. 43, 502, (2004).
10. A.F. Armstrong, T. Chivers, H.M. Tuononen, M. Parvez, and
R.T. Boeré. Inorg. Chem. 44, 7981 (2005).
11. J.K. Brask, T. Chivers, and M. Parvez. Inorg. Chem. 39, 2505
(2000).
12. D.A. Armstrong, D. Barr, W. Clegg, S.R. Drake, R.J. Singer,
R. Snaith, D. Stalke, and D.S. Wright. Angew. Chem. Int. Ed.
Engl. 30, 1707 (1991).
Acknowledgements
13. W.J. Grigsby, T. Hascall, J.J. Ellison, M.M. Olmstead, and P.P.
Power. Inorg. Chem. 35, 3254 (1996).
14. H. Breederveld and H.I. Waterman. Research (London), 6, 43S
(1953).
15. M. Söldner, A. Schier, and H. Schmidbaur. Inorg. Chem. 37,
510 (1998).
16. P. Power. Chem. Rev. 103, 789 (2003).
We thank the University of Calgary, the Natural Sciences
and Engineering Research Council of Canada (NSERC), and
the Academy of Finland (JK) for funding. The assistance of
J.S. Ritch and J.F. van Humbeck in obtaining the IR spec-
trum and GC–MS of 14, respectively, is gratefully acknowl-
edged.
17. D.J. Brauer, H. Bürger, and G.R. Liewald. J. Organomet.
Chem. 308, 119 (1986).
18. Bruker AXS. Versions 1.25 [computer program]; WinEPR.
Version 2.11 [computer program]. Bruker Analytische
Messtechnik GmbH, Rheinstetten, Germany. 1996.
19. V. Passarelli, G. Carta, G. Rossetto, and P. Zanella. Dalton
Trans. 413 (2003).
References
1. T. Chivers. Top. Curr. Chem. 229, 143 (2003).
2. G.M. Aspinall, M.C. Copsey, A.P. Leedham, and C.A. Russell.
Coord. Chem. Rev. 227, 217 (2002).
20. T. Chivers, M. Krahn, M. Parvez, and G. Schatte. Inorg. Chem.
3. J.K. Brask and T. Chivers. Angew. Chem. Int. Ed. 40, 3960
(2001).
4. P.R. Raithby, C.A. Russell, A. Steiner, and D.S. Wright.
Angew. Chem. Int. Ed. Engl. 36, 649 (1997).
42, 3994 (2003).
21. J.A. Weil, J.R. Bolton, and J.E. Wertz. Electron paramagnetic
resonance: Elementary theory and practical applications.
Wiley-Interscience, New York. 1994.
© 2006 NRC Canada