The allyl leaving group approach to tricoordinate silyl, germyl, and stannyl cations

Article abstract of DOI:10.1021/ja990389u

A group 14 atom bonded to three mesityl groups (2,4,6-trimethylphenyl) and to one allyl group serves as a novel precursor to tricoordinate group 14 cations, the analogues of the carbocation. The double bond of the allyl group provides an accessible reaction site that is located beyond the ortho methyl groups. Reaction of various electrophiles with the double bond releases the allyl group and leads to formation of the group 14 cations. The mesityl groups then are of sufficient steric bulk to protect the tricoordinate metal center from attack by nucleophiles. This approach is used herein with silicon, germanium, and tin as the central atom. The ²⁹Si chemical shift (δ 225) indicates full cationic character for the silicon system. The ¹¹⁹Sn chemical shift (δ 806) indicates less than full cationic character for the tin system. The positive charge for the germanium system has been assessed by examination of the aromatic ¹³C chemical shifts. These results provide the highest current cationic character for silylium and stannylium ions.

Full text of DOI:10.1021/ja990389u

J. Am. Chem. Soc. 1999, 121, 5001-5008
5001
The Allyl Leaving Group Approach to Tricoordinate Silyl, Germyl,
and Stannyl Cations
Joseph B. Lambert,* Yan Zhao, Hongwei Wu, Winston C. Tse, and Barbara Kuhlmann
Contribution from the Department of Chemistry, Northwestern UniVersity, EVanston, Illinois 60208-3113
ReceiVed February 8, 1999
Abstract: A group 14 atom bonded to three mesityl groups (2,4,6-trimethylphenyl) and to one allyl group
serves as a novel precursor to tricoordinate group 14 cations, the analogues of the carbocation. The double
bond of the allyl group provides an accessible reaction site that is located beyond the ortho methyl groups.
Reaction of various electrophiles with the double bond releases the allyl group and leads to formation of the
group 14 cations. The mesityl groups then are of sufficient steric bulk to protect the tricoordinate metal center
from attack by nucleophiles. This approach is used herein with silicon, germanium, and tin as the central
atom. The 29Si chemical shift (δ 225) indicates full cationic character for the silicon system. The 119Sn chemical
shift (δ 806) indicates less than full cationic character for the tin system. The positive charge for the germanium
system has been assessed by examination of the aromatic 13C chemical shifts. These results provide the highest
current cationic character for silylium and stannylium ions.
Although tricoordinate carbocations (R3C+) have been a
central feature of mechanistic organic chemistry for decades,
analogous species containing other positively charged heavy
atoms in place of carbon have not enjoyed a similar role. In the
particular case of the group 14 congeners (Si, Ge, Sn), the search
for stable, tricoordinate cations in the condensed phase has
extended over more than half a century.1,2 The principal
impediments have been high kinetic sensitivity of the species
to nucleophiles and low stabilization provided by π-delocalizing
substituents, compared with carbon. Inability to isolate stable
group 14 cations in the condensed phase has inhibited their
development as reactive intermediates in mechanistic schemes.
Although thermodynamic stability of the cations increases with
atoms lower in the periodic table because of their lower
electronegativity, the problems facing Ge and Sn are similar to
those of Si. We have recently reported a solution to the silyl
cation problem,3 and we discuss herein this approach within
the full context of group 14.
in 1992 for the case of stannylium cations.4 Superior anions
were reported in 1993.5 This stage of development reached its
apex with the solution of the X-ray structures of several silyl
cations by Reed’s1b and our6 groups. Our studies used the
tetrakis(pentafluorophenyl)borate anion, (C6F5)4B- (TPFPB),
and resulted in species that retained solvent as a weak fourth
coordination site.6 Reed’s studies used a series of halogenated
carboranes as anions and resulted in species that retained the
anion as a weak fourth coordination site.1b The silyl cationic
character was variously estimated by these authors as about 30-
50%, although other authors characterized the species as arenium
and halonium ions.7 Theoreticians pointed out that silyl cations
can complex with both saturated C-H bonds and even rare
gases.8 This observation seemingly doomed the existence of free
silyl cations in condensed phase, because any solvent or anion
would necessarily possess some nucleophilicity and conse-
quently would bind with silicon to some extent.
Steric effects provided the critical stabilizing factor in the
original creation of species containing multiple bonds to silicon,
such as disilenes, >SidSi<,9 to prevent dimerization and
reaction with external reagents. The idea therefore is not
particularly novel that bulky substituents might protect silylium
cations from reaction with nucleophiles. In fact, a decade ago
we tried to remove hydride from trimesitylsilane (mesityl,
Nucleophilic Isolation
Our overall strategy has been to protect the cationic species
from all nucleophiles. The development of this strategy,
however, took more than a decade and required four separate
innovations: (1) hydrocarbon solvents, (2) anions of very low
nucleophilicity, (3) bulky substituents, and (4) methodology for
creation of the cation. All four were necessary; the absence of
any one prohibited formation of a free cation. We introduced
aromatic solvents and low nucleophilicity anions in this context
(4) Lambert, J. B.; Kuhlmann, B. J. Chem. Soc., Chem. Commun. 1992,
931-932.
(5) (a) Lambert, J. B.; Zhang, S. J. Chem. Soc., Chem. Commun. 1993,
383-384. (b) Xie, Z.; Liston, D. J.; Jel´ınek, T.; Mitro, V.; Bau, R.; Reed,
C. A. J. Chem. Soc., Chem. Commun. 1993, 384-386.
(6) Lambert, J. B.; Zhang, S.; Stern, C. L.; Huffman, J. C. Science 1993,
260, 1917-1918. Lambert, J. B.; Zhang, S.; Ciro, S. M. Organometallics
1994, 13, 2430-2443.
(1) For reviews of the Si case, see (a) Lambert, J. B.; Kania, L.; Zhang,
S. Chem. ReV. 1995, 95, 1191-1201. (b) Reed, C. A. Acc. Chem. Res.
1998, 31, 325-332. (c) Lickiss, P. In The Chemistry of Organic Silicon
Compounds, 2nd ed.; Rappoport, Z., Apeloig, Y., Eds.; John Wiley &
Sons: Chichester, UK, 1998; pp 557-594.
(2) In this paper, these species are, respectively, called silylium,
germylium, and stannylium ions, according to IUPAC recommendations.
The terms silylenium, silicenium, and their Ge and Sn analogues have been
discarded and are no longer appropriate.
(7) Olah, G. A.; Rasul, G.; Li, X.-Y.; Buchholz, H. A.; Sandford, G.;
Prakash, G. K. S. Science 1993, 260, 983-984.
(8) (a) Maerker, C.; Kapp, J.; Schleyer, P. v. R. In Organosilicon
Chemistry II; Auner, N., Weis, J., Eds.; VCH: Weinheim, Germany, 1996;
pp 329-359. (b) Arshadi, M.; Johnels, D.; Edlund, U.; Ottosson, C.-H.;
Cramer, D. J. Am. Chem. Soc. 1996, 118, 5120-5131.
(9) Raabe, G.; Michl, J. In The Chemistry of Organic Silicon Compounds;
Patai, S., Rappoport, Z., Eds.; John Wiley & Sons: Chichester, UK, 1989;
pp 1015-1142.
(3) Lambert, J. B.; Zhao, Y. Angew. Chem., Int. Ed. Engl. 1997, 36,
400-401.
10.1021/ja990389u CCC: $18.00 © 1999 American Chemical Society
Published on Web 05/14/1999

5002 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Lambert et al.
Table 1. Chemical Shifts (ppm) for Mes3MCH2CH)CH2 and
abbreviated Mes, is 2,4,6-trimethylphenyl) by reaction with trityl
cation,10 the standard method at this time for trying to generate
silyl cations (the Corey method).11 The steric bulk surrounding
the hydride, however, thwarted access of the trityl cation and
led to no reaction (eq 1). We prepared trihexylsilylium TPFPB
Mes3M+ (C6F5)4B-
δ(13C)
δ(29Si) δ(119Sn) ipso ortho meta para
145.0 138.6 129.9 137.9
M
Xa
Si allyl
TPFPBb C6D6
CH3C6H5
solvent
C6D6
225.5
225.7
150.3 144.8 130.5
-
Mes3SiH + Ph3C+ X- f no reaction
(1)
c
d
1,4-(CH3)2C6H4 225.6
e
∆
C6D6
C6D6
5.3 6.2 0.6
-
Ge allyl
143.5 138.6 129.0 137.6
148.8 141.9 130.7 139.7
5.3 3.3 1.7 2.1
144.6 142.1 128.8 138.1
806 147.1 146.3 130.4 141.9
806
via the Corey method and found that the hexyl chains provided
little or no additional protection over methyl, as the 29Si chemical
shift of δ 90 in benzene confirmed complexation with an
external nucleophile.6 Movement of the chemical shift to δ 65
in mesitylene indicated that the nucleophile was solvent. Reed
and co-workers12 succeeded in making tri(tert-butyl)silylium
with a halocarborane anion via the Corey method, but its 29Si
chemical shift and X-ray structure indicated coordination with
the anion through the halogen atom. Thus tert-butyl is insuf-
ficiently bulky to prevent access of nucleophiles. There is a
further complication with alkyl side chains. As Schleyer pointed
out,8a the C-H bond of methane is capable of complexing with
silylium cations. Both rigid and flexible alkyl groups can provide
a C-H bond within binding distance to the silicon atom, so
that an agostic interaction may serve as the fourth coordination.
Thus, even if a large alkyl substituent could deflect solvent and
anion from the silicon atom, it is possible that a C-H bond
would be accessible and would provide a fourth coordination
site.
TPFPBb C6D6
e
∆
C6D6
C6D6
Sn allyl
TPFPBb C6D6
CH3C6H5
1,3,5-(CH3)3C6H3
1,2-(Cl2)C6H4
C6D6
806
806
e
∆
2.5 4.2 1.6 3.8
a X refers to the fourth group associated with Mes3M, either allyl or
the anion. b TPFPB is (C6F5)4B-. c C6D6/CH3C6H5 1/3. d C6D6/1,4-
(CH3)2C6H5 1/1. e The difference between the 13C chemical shifts of
the cation and of the allyl precursor in C6D6.
Results
The allyl derivatives were unknown, but we were able to
synthesize them by reaction of the halo trimesityl derivatives
with either the allyllithium reagent or the allyl Grignard reagent
(eq 3).
Our solution to the problem of alkyl coordination was to
revert to the mesityl group, all of whose C-H bonds are out of
the range of coordination without requiring severe angle-bending
distortion. The ortho methyl groups provide a steric shield both
above and below the plane of the silylium cation, preventing
access by external nucleophiles. A new method, however, was
needed to generate the cation, as hydride removal (eq 1, the
Corey method) is sterically prohibited.
Consequently, we used allyl as a leaving group. Although
allyl is unknown as a leaving group in carbon organic chemistry,
there is precedent in silicon organic chemistry.13 The approach
is illustrated in eq 2,
Mes3MBr/Cl + CH2dCHCH2-Li/MgX f
Mes3MCH2CHdCH2 (3)
Chlorotrimesitylsilane was prepared from trimesitylsilane. The
germanium and tin halo precursors were prepared by the reaction
of SnBr4 or GeCl4 with mesitylmagnesium bromide to produce
Mes3M-Br/Cl. The silicon and germanium derivatives were
treated with allyllithium to yield allyltrimesitylsilane (1) and
allyltrimesitylgermane (2), respectively, and the tin derivative
was treated with allylmagnesium bromide to yield allyltrimesi-
tyltin (3). The structures were proved by NMR spectroscopy
and X-ray crystallography.
Mes3MCH2CHdCH2 + E+ f
Various electrophiles were essayed in the reaction of eq 2.
We found that trityl was relatively ineffective. Our best results
were obtained by use of the solvated silyl cation, Et3Si-
,
(C6H6)+,5a,6 or the free â-silyl cation, Et3SiCH2CPh2+ 15 in either
Mes3MCH2 CHCH2E f Mes3M+ + CH2dCHCH2E (2)
+
in which M represents Si, Ge, or Sn and E+ represents an
appropriate electrophile. The double bond of the allyl group is
designed to provide a nucleophilic site outside the steric control
of the mesityl methyl groups. Reaction of the positively charged
electrophile E+ with the allyl double bond would result in a
â-silyl, -germyl, or -stannyl cation, which normally is unstable
and decomposes by expulsion of a silyl cation or its analogue.14
This step is favored by relief of steric strain, as the tetrahedral
environment around Si, Ge, or Sn is reduced to a trigonal
environment. The intention then is that, upon loss of allyl, the
three mesityl groups snap into a cage that prevents access of
external nucleophiles to the central atom and retains the
tricoordinate nature of the cation.
case with TPFPB as the anion and arenes as the solvent.
Manipulations were carried out at room temperature under
nitrogen in the protected environment of a glovebox to avoid
exposure to atmospheric water and oxygen. The resulting
solutions were stable for weeks and exhibited only single signals
in the 29Si or 119Sn spectra. Table 1 lists the chemical shifts of
the heavy atoms as a function of solvent. Figure 1 displays the
29Si spectrum of Mes3Si+ and the 119Sn spectrum of Mes3Sn+.
Product recovery was explored by reaction of Mes3Si+ or
Mes3Sn+ with Bu3SnH. Hydride transfer occurred, and Mes3-
SiH or Mes3SnH was observed as the principal product.
Ultraviolet spectra were recorded in benzene.
Discussion
(10) Lambert, J. B.; Lentz, K. T.; Kania, L. Unpublished results.
(11) Corey, J. Y. J. Am. Chem. Soc. 1975, 97, 3237-3238.
(12) Xie, Z.; Bau, R.; Benesi, A.; Reed, C. A. Organometallics 1995,
14, 3933-3941.
The first consideration was the steric accessibility of the allyl
group. Figure 2 is a stereoscopic representation of the X-ray
structure of allyltrimesitylsilane (Mes3SiCH2CHdCH2, 1). The
(13) Shade, L.; Mayr, H. Makromol. Chem. Rapid Commun. 1988, 9,
477-482. Uhlig, W. In Organosilicon Chemistry; Auner, N., Weis, J., Eds.;
VCH: Weinheim, Germany, 1994; pp 21-26.
(14) Lambert, J. B. Tetrahedron 1990, 46, 2677-2689.
(15) Lambert, J. B.; Zhao, Y. J. Am. Chem. Soc. 1996, 33, 7867-7868.

Allyl LeaVing Group Approach to Tricoordinate Cations
J. Am. Chem. Soc., Vol. 121, No. 21, 1999 5003
structured layers in the liquid.16 He found that some liquid
clathrates could be enticed to crystallize at lower temperatures,
as we also found for the solvated silyl cation Et3Si(toluene)+
TPFPB-.6 The silylium, germylium, and stannylium ions
reported herein, however, have not as yet yielded crystals,
possibly because of low lattice energy. Consequently, these ions
have been characterized by NMR spectroscopy, as described
in the next three sections.
The Silylium Ion. The chemical shift of Mes3Si+ TPFPB-
was observed at δ 225.5 in benzene. Our solvated trialkylsilyl
cations5a,6 and Reed’s similar but anion-complexed ions1b,5b
exhibited chemical shifts ranging up to δ 115. The solvent
dependence of the chemical shift indicated that the silicon of
these trialkylsilyl cations is involved in coordination with
solvent, if not anion.
Figure 1. Upper: The 79.46 MHz 29Si spectrum of trimesitylsilylium
tetrakis(pentafluorophenyl)borate in C6D6. Lower: The 149.15 MHz
119Sn spectrum of trimesitylstannylium tetrakis(pentafluorophenyl)borate
in C6D6.
The sensitivity of the chemical shift of the trimesityl species
to its nucleophilic surroundings and hence the coordination status
of silicon can be assessed by changing the solvent and the anion.
As Table 1 indicates, the chemical shift of δ 225-6 is invariant
to solvent in benzene, toluene, and p-xylene. In contrast, the
solvated trialkylsilyl cation Et3Si(arene)+ TPFPB- exhibited a
solvent-dependent chemical shift: δ 92.3 in benzene, 87.1 in
3/1 benzene/toluene, and 81.8 in pure toluene.6 These arene
solvents therefore do not penetrate the protection by the mesityl
methyl groups and do not complex with the silicon (arenes are
nucleophilic primarily on their π face, so their approach to
silicon must be with the sterically maximal face down). We
did find, however, that acetonitrile and triethylamine either were
smaller or more strongly nucleophilic, as 1/3 benzene/acetonitrile
led to a chemical shift of δ 37.0 and 1/1 benzene/triethylamine
led to a chemical shift of δ 47.1, both consistent with
tetracoordination. We supplied allyltrimesitylsilane to Reed and
Fackler for reaction with electrophiles bearing carborane anions.
With the closo-7,8,9,10,11,12-Cl6-CB11H6 anion, they measured
the chemical shift of Mes3Si+ to be δ 225.9,18 identical to the
values with the TPFPB anion. Thus the chemical shift is
invariant both to anion and to arene solvents, indicating the
absence of or the identity of coordination in every case. The
observed change in chemical shift with variation of nucleophi-
licities of the arenes with the trialkylsilyl cases, however,
militates against identical coordination and favors absence of
coordination for the solvent in the trimesityl case.
The value of δ 225 is at a considerably higher frequency
than (downfield from) any previous silicon species not contain-
ing divalent silicon or attachment to a transition metal.19 For
Me3Si+ in the gas phase, calculated values ranged from δ 351
to 418, depending on the basis set used.20 Reed and co-workers
made adjustments for substituents (iPr3Si+) and nonspecific
solvation and estimated a range of δ 220-320 for the tricoor-
dinate geometry, compared with the observed δ 115, corre-
sponding to 36-52% silylium ion character.21 If compared
directly with the gas-phase values, δ 115 represents 28-33%
silylium character.
allyl group clearly protrudes beyond the mesityl methyl groups.
Consequently and in contrast to the case of the hydride of
trimesitylsilane (Mes3SiH), the double bond of the allyl group
is accessible to attack by electrophiles.
The reaction of eq 2 occurred successfully for all three central
atoms, to deliver, respectively, the silylium ion (Mes3Si+, 4),
the germylium ion (Mes3Ge+, 5), and the stannylium ion (Mes3-
Sn+, 6), with degrees of complexation with solvent or anion
that still must be assessed. The byproducts of eq 2 were CH2-
CHdCH2SiEt3 or CH2CHdCH2CPh2CH2SiEt3, depending on
whether the solvated silyl cation or the free â-silyl carbocation
served as the electrophile.
It is noteworthy that the 29Si resonance for either byproduct
does not appear in the upper spectrum of Figure 1. The reaction
of eq 2 always produces two benzene layers (or arene layers in
general). The ionic species are found in the lower layer, and
the upper layer is benzene containing the nearly nonpolar
byproduct. This behavior is characteristic of the formation of
liquid clathrates. Such species were originally discovered with
aluminate salts of a variety of cations,16 but other large anions
have now been found to give this behavior, including borates.17
Liquid clathrates have been defined16b as “nonstoichiometric
liquid inclusion compounds [formed] upon the interaction of
aromatic molecules with [salts]”. They were characterized by
the formation of two solvent layers, the lower layer also
containing the salt and the upper layer containing pure solvent.
Our solutions look identical to the phases depicted in Figure
6.10 of ref 16b. It is likely but not proved that our materials
exist as a liquid clathrate. Moreover, the precursors (the solvated
silyl cation, Et3Si(C6H6)+, or the free â-silyl cation, Et3SiCH2-
CPh2+) also form these two layers, so they too may exist as
liquid clathrates. In each case the upper layer of solvent is
syringed off so that the actual reagent is the ion-containing lower
layer.
Alkyl groups, however, are inappropriate models for the
chemical shift of silicon attached to aryl groups. At the time of
our experiments, no theoretical calculations had been published
for the chemical shift of triarylsilylium ions. Now, however,
Atwood has suggested that the primary condition for liquid
clathrate formation is a low lattice energy. The large size of
the cation, as in the present case, permits greater numbers of
arene molecules in the clathrate. His suggested model is of
(16) (a) Atwood, J. L. In Inclusion Compounds, Atwood, J. L., Ed.;
Academic Press: London, 1984; Vol. 1, Chapter 9. (b) Atwood, J. L. In
Coordination Chemistry of Aluminum; Robinson, G. H., Ed.; VCH: New
York, 1993; pp 197-232.
(17) Bond, D. R.; Jackson, G. E.; Joa˜o, H. C.; Hofmeyer, M. N.; Modro,
T. A.; Nassimbein, L. R. J. Chem. Soc., Chem. Commun. 1989, 1910-
1911. Hunter, R.; Haueisan, R. H.; Irving, A. Angew. Chem., Int. Ed. Engl.
1994, 33, 566-568. Dioumaev, V. K.; Harrod, J. F. Organometallics 1996,
15, 3859-3867.
(18) Reed, C. A.; Fackler, N. Unpublished results.
(19) West, R.; Denk, M. Pure Appl. Chem. 1996, 68, 785-788.
Grumbine, S. K.; Tilley, T. D.; Arnold, F. P.; Rheingold, A. L. J. Am.
Chem. Soc. 1994, 116, 5495-5496. Tilley, T. D. J. Am. Chem. Soc. 1998,
120, 11184-11185.
(20) Kutzelnigg, W.; Feischer, U.; Schindler, M. NMR; Basic Principles
and Progress; Springer-Verlag: New York, 1991; Vol. 23.
(21) Xie, Z.; Manning, J.; Reed, R. W.; Mathur, R.; Boyd, P. D. W.;
Benesi, A.; Reed, C. A. J. Am. Chem. Soc. 1996, 118, 2922-2928.

5004 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Lambert et al.
Figure 2. Stereoscopic representation of the X-ray structure of allyltrimesitylsilane.
two groups22,23 have published such values. The calculations
vary with the choice of basis set and with the angle of twist for
the arene rings. The calculated angle of twist of trimesitylsily-
lium was 48°, compared with 29° for triphenylsilylium (sityl)
lacking the methyl groups.22 Despite this twisting distortion,
there still is considerable conjugation of the phenyl rings with
the empty orbital on silicon, because the calculated arene C-C
bond lengths were pronouncedly unequal and in a dactylic
pattern (long-short-short). The calculated chemical shift was δ
230.1, very favorably comparable to the experimental value of
δ 225.5. Moreover, when benzene was introduced calculationally
at an appropriate geometry for complexation but at a closest
approach distance, the value changed very little, to δ 228.6.22
The benzene ring could not get close enough to coordinate with
silicon. In contrast, the value for trimethylsilylium changed
calculationally from δ 361.6 to 80.4 on addition of benzene to
the system.22 These calculations confirm the conclusions that
trialkylsilylium ions complex with solvent but trimesitylsilylium
is an entirely free silyl cation. The calculations also confirmed
that the ortho methyl C-H bonds are too far away to complex
with the empty silicon 3p orbital.
322 for Me3Sn(FSO3). These values indicated considerable
deshielding of the tin atom from the expectation for pure
tetracoordination close to δ 0. The solvent dependence, however,
is reminiscent of the case for the solvated trialkylsilylium ions.
There are no calculations of tin shieldings comparable to those
for silicon shieldings. Arshadi et al., however, have presented
a correlation between 29Si and 119Sn chemical shifts, and a
chemical shift of δ 360 for silicon corresponds to one of about
δ 1700 for tin.26 If the correlation is valid and the relationship
is linear, then the observed 119Sn shifts of ca. δ 350 correspond
to only about 20% stannylium ion character (parallel to that
for trialkylsilylium ions generated under similar conditions when
compared with chemical shifts calculated for the gas phase).
Arshadi et al. thus concluded and we agree that the tin cations
prepared by the Japanese group and by us had low stannylium
character.26
To test the allyl leaving group method for tin, we first allowed
allyltributylstannane to react with trityl TPFPB in benzene by
analogy with eq 2. The lower layer of the presumed liquid
clathrate produced a single 119Sn peak at δ 262.8. Examination
of the upper layer by GC/MS revealed the presence of the
expected byproduct, allyltriphenylmethane. For comparison,
tributyltin hydride was allowed to react with trityl TPFPB by
analogy with eq 1. The resulting lower layer produced a 119Sn
peak at essentially the same chemical shift, δ 262.7, as for the
material from the allyl leaving group method. The two methods
thus produce the same stannyl cation, solvated by benzene and
possessing ca. 20% stannylium character, in line with our and
Sakurai’s previous observations.4,25
Experimental evidence for conjugation of the aryl rings with
silicon can be obtained by 13C NMR spectroscopy and by
ultraviolet spectroscopy. The NMR evidence is discussed below
in the context of the germylium cation, so that all systems can
be examined together. The ultraviolet spectrum had to be taken
1
in benzene solution, not an optimal medium. The strong Lb
benzene peak at 254 nm provides intense end absorption. Two
absorptions were observed beyond this end absorption, a
maximum at 304 nm and a shoulder at 370 nm. Tailing of the
latter absorption into the visible causes the dark yellow color
of the solution. The triphenylmethyl cation (trityl) exhibits two
peaks at 409 and 428 nm.24 Thus the silylium cation absorbs at
shorter wavelength than trityl. Nonetheless, its absorption close
to the visible provides further support for conjugation between
the mesityl rings and the silicon atom. It should be noted that
the crystals of Et3Si(toluene)+ TPFPB- were colorless.6
The Stannylium Ion. The strategy of an allyl leaving group
also should be applicable to the case of tin. Low nucleophilic
media have previously been used in attempts to prepare a free
stannylium ion. In 1992 we used the trityl cation to abstract
hydride from trialkylstannane (the method of eq 1) with various
solvents and anions and obtained 119Sn chemical shifts in the
range δ 150-360.4 In 1994 Kira, Oyamada, and Sakurai25a
reported similar results with a different anion. Earlier, Birchall
and Manivannan25b had reported a 119Sn chemical shift of δ
Removal of the allyl group in the case of butyl substituents
in essence is superfluous, as hydride is readily extractable. The
observed chemical shift indicates that, as expected, butyl is too
small to prevent coordination of the stannyl cation with external
nucleophiles, which appear to be solvent as judged by the
solvent dependence of the chemical shift. Consequently, bulkier
substituents again are needed. We prepared allyltrimesitylstan-
nane (δ -156) and allowed it to react with trityl TPFPB. Its
failure to react paralleled the situation with allyltrimesitylsilane
and presumably was due to the low reactivity of trityl with a
double bond. We then had recourse to the more reactive
electrophiles used in the silicon case: the solvated silyl cation,5a,6
Et3Si(C6H6)+, and the free â-silyl carbocation,15 Et3SiCH2CPh2
.
+
Both these electrophiles reacted successfully with allyltrimesi-
tylstannane to give identical lower (liquid clathrate) layers that
exhibited a single 119Sn peak at δ 806 (lower spectrum in Figure
2). These solutions were stable at room temperature for several
days.
The observed 119Sn chemical shift is at a record deshielded
position for species other than divalent tin and consequently
indicates considerable progress toward a tricoordinate species,
(22) Mu¨ller, T.; Zhao, Y.; Lambert, J. B. Organometallics 1998, 17, 278-
280.
(23) Kraka, E.; Sosa, C. P.; Gra¨fenstein, J.; Cremer, D. Chem. Phys.
Lett. 1997, 279, 9-16.
(24) Branch, G.; Walba H. J. Am. Chem. Soc. 1954, 73, 1564-000.
(25) (a) Kira, M.; Oyamada, T.; Sakurai, H. J. Organomet. Chem. 1994,
471, C4-C5. (b) Birchall, T.; Manivannan, V. J. Chem. Soc., Dalton Trans.
1985, 2671-2675.
(26) Arshadi, M.; Johnels, D.; Edlund, U. J. Chem. Soc., Chem. Commun.
1996, 1279-1280.

Allyl LeaVing Group Approach to Tricoordinate Cations
J. Am. Chem. Soc., Vol. 121, No. 21, 1999 5005
compared with the previous extreme of about δ 360. Unfortu-
nately, there are no calculations of 119Sn chemical shifts of a
free stannylium cation that can serve as a point of comparison.
It is for this reason that the linear relationship of Arshadi et
al.26 is useful. A 29Si chemical shift of δ 225 translates to a
119Sn chemical shift of about δ 1100. Two conclusions may be
drawn from this observation. (1) By strict linear proportionality,
an observed 119Sn chemical shift of δ 806 is about three-quarters
of the extreme expected for a fully tricoordinate stannylium ion.
(2) If the relationship is nonlinear, the possible range of
stannylium character is vastly expanded, including the pos-
sibilities that we have reached the tricoordinate stannylium
extreme or that we are even less than three-quarters of the way
to the extreme.
As our working model, we decided to accept the linear model,
meaning that our observed stannyl cation is about three-quarters
of the way along the continuum from tetra- to tricoordination.1b
If so, then either solvent or anion must be able to have access
to some degree to the tin atom. The longer C-Sn bonds,
compared with C-Si bonds, or perhaps the higher electrophi-
licity of the tin center could bring about the residual coordina-
tion. We examined the solvent dependence of the 119Sn chemical
shift of trimesitylstannyl TPFPB and found it to be invariant at
δ 806 for benzene, 1/2 benzene/toluene, 1/2 benzene/p-xylene,
and 1/2 benzene/1,2-dichlorobenzene. This constancy of the
chemical shift with considerable variation of the nucleophilicity
of the solvent indicates that aromatic solvents are not able to
coordinate with the tin cation. Two alternative conclusions
remain. (1) The value of δ 806 indeed corresponds to the
extreme for tricoordination. (2) A nucleophile other than solvent
is able to coordinate partially with the tin cation. The only
remaining possible nucleophile is the anion TPFPB, and the
likely nucleophilic site would be the para or (less likely) meta
fluorine atoms. Marks et al.27 in fact have observed one X-ray
structure in which the para and meta fluorine atoms chelate to
a zirconium cation.
697. This chemical shift is more shielded by about 100 ppm
than that of the trimesitylstannyl cation, so negative progress
had been achieved by attempting to bulk up the aryl substituent.
Two conclusions may be drawn. (1) Two triisopropylphenyl
rings and one phenyl ring are slightly less sterically constraining
than three mesityl groups, allowing a nucleophile such as solvent
or anion better access to the tin atom. (2) The C-H bonds of
the isopropyl groups approach more closely to the tin atom than
those of methyl on the mesityl groups, allowing agostic
interactions in the tri(isopropyl)phenyl case that were absent in
the mesityl case. Either way, we have a stannyl cation with
slightly less tricoordinate (stannylium) character than that of
trimesitylstannyl.
We also operated on the anion, arguing that replacement of
the four para fluorine atoms of TPFPB- might eliminate the
primary nucleophilic center of the anion. We took two ap-
proaches, neither successful. On one hand, we replaced 4-F with
4-Me3Si, in the hopes that the size of the group would prevent
its access to the tin atom. Although the anion synthesis was
successful, trityl tetrakis[4-(trimethylsilyl)-2,3,5,6-tetrafluo-
rophenyl]borate proved to be intractably insoluble in aromatic
solvents, so that we were not able to effect any reactions. On
the other hand, we replaced 4-F with 4-H to give trityl tetrakis-
(2,3,5,6-tetrafluorophenyl)borate. This material gave irrepro-
ducible results and may not have been pure.
Consequently, we are left with δ 806 as the high frequency
extreme for a stannyl cation. Although, according to the assumed
linear Arshadi correlation, this value corresponds to only about
75% stannylium ion character, it represents considerable
progress from previous observations of ca. δ 360. It is ironic
that the stannylium cation, which presumably has greater
thermodynamic stability than the silylium cation because of the
higher electropositivity of tin, remains a slightly more difficult
goal. Thermodynamic stability of course never was the problem
with silylium ions. The problem was high kinetic reactivity of
the silyl center, and that problem remains or may even be
heightened for the tin case, because of the higher amount of
positive charge on tin and its higher polarizability. Thus the
higher thermodynamic stability of tin is counterbalanced by its
higher electrophilicity and by its higher accessibility because
of the longer C-Sn bonds.
The ultraviolet spectrum of trimesitylstannylium TPFPB in
benzene showed clear maxima at 286 and 398 nm. The latter is
at slightly longer wavelength than the longest wavelength
silylium absorption (370 nm) and supports on one hand high
stannylium character for this species and on the other hand
greater interaction of the phenyl rings when the central atom is
tin rather than silicon, probably because the longer Sn-aryl
bonds position the ortho methyl groups further from each other,
allowing greater planarity of the aryl rings.
In light of these observations we sought to synthesize an
allyltriarylstannane system with bulkier aryl groups than mesityl.
We encountered considerable synthetic problems in synthesizing
the precursor halostannane, Ar3SnX. Reaction of SnBr4 with
2,4,6-tri-tert-butylphenyllithium (supermesityllithium) led to
introduction of only two supermesityl groups, one of which was
rearranged, as the product was dibromosupermesityl(dimethyl-
(3,5-di-tert-butylphenyl)methyl)stannane, sMesSnBr2CMe2(3,5-
tBu2C6H3), in which sMes stands for supermesityl. Reaction of
SnCl4 with 2,4,6-triisopropylphenyllithium (TipLi) also led to
the introduction of only two aryl groups, albeit unrearranged,
to form Tip2SnCl2.
As this last procedure occurred in low yield, we had recourse
to the Masamune method.28 Reaction of TipMgBr with SnBr4
led to formation of the cyclic stannoxane, di(Tip)cyclotristan-
noxane, (Tip2SnO)3. Reaction of this material with hydrogen
chloride then gave Tip2SnCl2 in good yield. It remained to
introduce a third aryl group. Unfortunately, mesityllithium failed
to react. Only the sterically smaller phenyllithium was success-
ful, giving Tip2C6H5SnCl, which successfully reacted with
allyllithium to give allyldi(Tip)phenylstannane (7), whose
structure was confirmed by X-ray crystallography.
The Germylium Ion. The only previous study of monoger-
myl cations used hydride removal from R3GeH (R alkyl) with
perchlorate as the anion and solvents other than arenes.29 The
ionic species prepared in sulfolane most likely were solvent
complexes, and the species in dichloromethane were strongly
complexed with the anion. By analogy with similar silyl and
stannyl species, these germyl materials may have had 10%
germylium character. Sekiguchi and co-workers30a have prepared
a free cyclotrigermyl cation. The new, low nucleophilicity anions
and solvents have not previously been examined in the
monogermanium context, although Sekiguchi and co-workers30b
We allowed 7 to react with Et3Si(C6D6)+ TPFPB- in order
to form the stannyl cation C6H5Tip2Sn+ TPFPB-. The chemical
shift of the only species in the lower layer was observed at δ
(27) Yang, X.; Stern, C. L.; Marks, T. J. Organometallics 1991, 10, 840-
(29) Lambert, J. B.; Schilf, W. Organometallics 1988, 7, 1659-1660.
(30) (a) Sekiguchi, A.; Tsukamoto, M.; Ichinohe, M. Science 1997, 275,
60-61. (b) Ichinohe, M.; Fukaya, N.; Sekiguchi, A. Chem. Lett. 1998,
1045-1046.
842.
(28) Masamune, S.; Sita, L.; Williams, D. J. J. Am. Chem. Soc. 1983,
105, 630-631.

5006 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Lambert et al.
Figure 3. The 100.57 MHz 13C spectrum of the aromatic region of trimesitylgermylium tetrakis(pentafluorophenyl)borate in C6D6. The triplet at
δ 128 is from solvent. The doublets centered at δ 149, 139, and 137 are from the carbons on the anion, each carbon split by fluorine.
have utilized them for their cyclotrigermyl systen. For optimal
results, we went directly to the mesityl system with the allyl
leaving group. Treatment of allyltrimesitylgermane with Et3Si-
(C6D6)+ TPFPB- in benzene produced a two-layer system
closely resembling those for the cases of silicon and tin.
Germanium lacks a sensitive and convenient nuclide for
examination of its coordination properties (the spin of 73Ge is
9/2, its natural abundance is 7.76%, and its natural sensitivity
compared with 1H is 1.09 × 10-4).31 Consequently, there is no
direct NMR method to assess the degree of germylium character
in the species produced by the allyl leaving group method with
low nucleophilic anion and solvent. When phenyl is attached
to sp2 carbon (alkene, carbonyl, carbocation, and so on),
delocalization of charge through resonance generates positive
charge in the phenyl ring. The extent of charge delocalization
is dependent on a number of factors, such as the degree of
electron deficiency at the R carbon and the angle of twist of
the phenyl ring in relation to the p orbital on the R carbon. The
resulting positive charge has a palpable effect on the 13C shifts
of the phenyl carbons, particularly the ortho and para carbons,
to which positive charge is transferred by delocalization.
Thus the 13C chemical shifts of the mesityl carbons can
provide an alternative approach to assessing the degree of
positive charge delocalization. Three factors must be taken into
account: (1) the effect of the change in heteroatom (C to Si,
Ge, Sn) at the R position to the aryl ring; (2) the effect of the
ortho and para methyl groups; and (3) finally the desideratum,
the extent of positive charge delocalization. The most straight-
forward comparison is between the starting material allyl
derivatives and the cations. This comparison automatically takes
the first two factors into consideration, as both starting material
(Mes3MCH2CHdCH2) and product (Mes3M+) have aryl methyl
groups and the central atom M in common. The effect of the
introduction of positive charge in place of allyl consequently is
for the most part isolated.
Table 1 contains the aromatic chemical shifts for these six
systems (three allyl, three cationic). Assignments were made
according to peak intensities, expected positions, and 13C-1H
coupling relationships as revealed by gHMQBC (gradiant hetero
multiple quantum bond correlation) 2D spectra. Thus the meta
carbons are essentially insensitive to all effects and remain
constantly at about δ 129-130. Moreover this peak is the
strongest in the spectrum, both because the meta carbon is the
only aromatic carbon bearing a hydrogen and because there are
two meta carbons per ring. Of the other three carbons, the ortho
is expected to be most intense because of its 2-fold presence in
each ring. The ipso carbon should appear at highest frequency,
by analogy with the observations in the series C6H5Si/Ge/
SnMe3.32 The effect of methyl in our mesityl derivatives is to
move the carbons 5-9 ppm to higher frequency (compared with
phenyl). Such a shift, however, could put the ortho carbon in
the vicinity of the ipso carbon. The gHMQBC spectra of the
allyl derivatives clearly showed the relationships between each
ring carbon and the methyl groups.
Figure 3 illustrates the case for the germanium cation. The
aromatic region also contains the triplet for C6D6 at δ 128 and
the aromatic carbons from the anion, (C6F5)4B-. We have seen
these latter carbons in many spectra, and they are easily
identified from their doubling through coupling to 19F. In the
spectrum of Figure 3, they appear as doublets centered around
δ 137, 139, and 149. The meta carbon appears at its usual
position (δ 130.7) and with the highest intensity. The ipso is
identified at the highest frequency of the remaining carbons,
the ortho by its second highest intensity, and the para by
difference.
As the data in Table 1 indicate, there are shifts (from allyl
starting material to product cation) to higher frequency for each
of the aromatic carbons, although as expected the effect is
minimal for the meta carbons (0.6-1.7 ppm). The effects at
the ipso carbon are largely the result of the change in the
(31) Harris, R. K. In NMR and the Periodic Table; Harris, R. K., Mann,
B. E., Eds.; Academic Press: London, 1978; p 7.
(32) Bullpitt, M.; Kitching, W.; Adcock, W.; Doddrell, D. J. Organomet.
Chem. 1976, 116, 161-185.

Allyl LeaVing Group Approach to Tricoordinate Cations
J. Am. Chem. Soc., Vol. 121, No. 21, 1999 5007
Product Studies (Silylium Ion). (a) 1-Allyl-1,1-diphenyl-2-(tri-
ethylsilyl)ethane. In a N2-filled glovebox, trityl TPFPB (160 mg, 0.17
mmol) was dissolved in dry toluene (1.5 mL) in a valved 5 mm NMR
tube. Addition of triethylsilane (25 mg, 0.22 mmol) produced a light
brown oil at the bottom. The colorless top phase was then taken off
with a syringe to remove the triphenylmethane byproduct. 1,1-
Diphenylethene (40 mg, 0.22 mmol) was added, and the oil phase
changed to deep green. Allyltrimesitylsilane (80 mg, 0.19 mmol) in
toluene (0.7 mL) was added and a deep red oil formed. The light orange
top phase was removed and analyzed by GC/MS. 1-Allyl-1,1-diphenyl-
electronegativity of the attached Si, Ge, or Sn on loss of allyl
and are not of interest. The effects at the ortho and para carbons
should result from the delocalization of charge on formation of
the cation. These effects are in the range 2.1-6.2 (the para
carbon for the Si case overlapped with anion carbons and could
not be observed). Cationic character for the germyl system is
comparable to those of the silyl and stannyl systems. The effects
are much smaller than that for loss of hydride in the pair Ph3-
CH f Ph3C+, in which the ortho and para carbons move to
higher frequency by about 15 ppm.33 The shorter C-Ph bond
length and the absence of the ortho methyl groups in trityl permit
a more nearly planar system and stronger resonance interactions.
2-(triethylsilyl)ethane was found: MS (EI) m/z 296 (27), 295 (M+
-
allyl, 97), 267 (81), 205 (23), 204 (100), 176 (29), 175 (45), 147 (47),
145 (24). The top layer also contained the excess of the starting
materials, including CH2dCPh2, Mes3SiCH2CHdCH2, Mes3SiH (pre-
sumably from reaction of Mes3Si+ with Et3SiH), Et3SiCH2CHPh2 (from
the reaction of Et3SiCH2C+Ph2 with Et3SiH), and other uncharacterized
materials. (b) Trimesitylsilane. The lower level was washed with dry
toluene (3 × 2 mL) to remove all the nonpolar materials. Addition of
Bu3SnH (0.14 g, 0.48 mmol) changed the red oil on the bottom to
brown. Toluene (1.5 mL) was added, and the new top layer was taken
off with a syringe and analyzed by GC/MS. Many materials formed as
Bu3SnH decomposed in the GC column. The major product, however,
was found to be trimesitylsilane: MS (EI) m/z 386 (M+, 1) 267 (28),
266 (99), 251 (54), 235 (38), 160 (35), 148 (16), 147 (95), 146 (100).
Conclusions
Treatment of allyltrimesitylsilane, -germane, and -tin with
strong electrophiles, respectively, produces silylium, germylium,
and stannylium cations. Requisite conditions include arene
solvents, low nucleophilicity anion [tetrakis(pentafluorophenyl)-
borate], and protection from atmospheric water and oxygen. The
29Si chemical shift of the silylium species (δ 225) is consistent
with pure tricoordination. The 119Sn chemical shift of the
stannylium species (δ 806) is at a record high frequency but
may be short of that expected for pure coordination. Such an
assessment could not be made for the germylium ion, but
analysis of the aryl 13C shifts on cation formation are consistent
with charge development on germanium that is comparable to
that on silicon and tin.
Tributylstannylium TPFPB Complex. (a) By Hydride Abstrac-
tion from Bu3SnH. In a N2-filled glovebox, trityl TPFPB (160 mg,
0.17 mmol) was dissolved in dry C6D6 (0.7 mL) in a valved 5 mm
NMR tube. Addition of Bu3SnH (60 mg, 0.21 mmol) produced a light
brown oil at the bottom: 13C NMR (C6D6) δ 12.9, 24.6, 26.7, 27.3;
119Sn NMR (C6D6) δ 262.7. (b) By the Allyl Method. In a N2-filled
glovebox, trityl TPFPB (160 mg, 0.17 mmol) was dissolved in dry
C6D6 (0.7 mL) in a valved 5 mm NMR tube. Addition of allyltribu-
tylstannane (63 mg, 0.19 mmol) produced a light brown oil at the
bottom: 13C NMR (C6D6) δ 12.9, 24.3, 26.7, 27.4; 119Sn NMR (C6D6)
δ 262.8. The colorless top layer was removed with a syringe and
analyzed by GC/MS. Allyltriphenylmethane (Ph3CCH2CHdCH2) was
found: MS (EI) m/z 244 (28), 243 (M+ - allyl, 84), 166 (21), 165
(100), 115 (10), 91 (11), 77 (10), 51 (12), 41 (18), 39 (18).
Bromotrimesityltin.37 To 100 mL of toluene containing SnBr4 (6.0
mL, 0.046 mol) in a 500 mL flask was added dropwise an ether solution
of mesitylmagnesium bromide (165.6 mL, 1.0 M). The reaction mixture
was allowed to reflux for 10 h. The reaction was then quenched with
H2O and 10% aqueous HBr. Dimesityltin bromide either fell out as a
dense white solid or became suspended in the organic layer. Filtration
of the organic layer removed the disubstituted product. The toluene
layer then was washed with H2O (2 × 50 mL), NaHCO3 (2 × 50 mL),
and again with H2O (2 × 50 mL). The solution was dried (MgSO4),
and the solvent was removed by rotary evaporation. Methanol was
added to the resultant yellow-orange oil to aid crystal formation by
solvent diffusion. The product was a white powder (8.2 g, 32%): mp
159-160 °C; 1H NMR (CDCl3) δ 2.23 (s, 18H), 2.25 (s, 9H), 6.78 (s,
6H); MS (EI) m/z 555 (M+ - 1, 32), 436 (100), 477 (22), 357 (13).
Experimental Section
Diethyl ether, benzene, and toluene were distilled from sodium/
benzophenone prior to use. Pentane and hexane were refluxed with
LiAlH4 and distilled directly into the reaction vessels. Deuterated
benzene was dried with Na/K alloy, vacuum-transferred to a Schlenk
tube, and placed in a nitrogen-filled glovebox. Triethylsilane was dried
with LiAlH4, distilled to a Schlenk tube, and placed in the glovebox.
NMR spectra were recorded on a Varian Unity Plus 400 (1H, 400 MHz;
13C, 100 MHz; 29Si, 79 MHz, 119Sn, 149 MHz), Gemini 300 (1H, 300
MHz, 13C, 75 MHz), or VXR-300 (1H, 300 MHz, 13C, 75 MHz)
spectrometers.
Trimesitylsilane was prepared in 52% yield by the method of
Gynane et al.34
Chlorotrimesitylsilane was prepared in 83% yield by the method
of Zigler et al.35
Allyltrimesitylsilane (1) was prepared by the method of Lambert
et al.36 The X-ray structure has been reported.36
Trimesitylsilylium Tetrakis(pentafluorophenyl)borate (TPFPB)
(4). In a N2-filled glovebox, trityl TPFPB6 (160 mg, 0.17 mmol) was
dissolved in dry C6D6 (0.7 mL) in a valved 5 mm NMR tube. Addition
of triethylsilane (25 mg, 0.22 mmol) produced two layers, the lower
one consisting of a light brown oil. The colorless top phase was taken
off with a syringe to remove the triphenylmethane byproduct. 1,1-
Diphenylethene (40 mg, 0.22 mmol) was added, and the oil phase
changed to deep green. Allyltrimesitylsilane (80 mg, 0.19 mmol) in
C6D6 (0.5 mL) was added to create two layers again. The lower layer
was a deep red oil. The light orange top phase was removed, and the
oil was examined by NMR spectroscopy: 1H NMR (C6D6) δ 1.92 (s,
9H), 2.10 (s, 18H), 6.60 (s, 6H); 13C NMR (C6D6) δ 21.5, 23.7, 130.5,
137.0 (d), 138.9 (d), 144.8, 149.1 (d), 150.3; 29Si NMR (C6D6) δ 225.5
(1JSi-C ) 70.6 Hz).
Allyltrimesitylstannane (3). Allylmagnesium bromide (12 mL,
0.012 mol) was added in excess to bromotrimesityltin (6.3 g, 0.011
mol) in toluene. After refluxing for 20 h, the reaction mixture was
quenched with H2O and 10% aqueous HBr. The toluene portion was
washed with H2O, NaHCO3, and H2O again. The solution was dried
(MgSO4), the solvent was removed by rotary evaporation, and hexane
and methanol were added to force crystallization. Recrystallization from
1
CH3OH yielded a white powder: 2.8 g, 49.2%; mp 116-118 °C; H
NMR (CDCl3) δ 2.21 (s, 18H), 2.23 (s, 9H), 2.38 (d, 2H), 4.74 (m,
2H), 5.78 (m, 1H), 6.78 (s, 6H); 13C NMR (CDCl3) δ 20.9, 25.6, 26.3,
113.6, 128.8, 136.8, 138.1, 142.1, 144.6; 119Sn NMR (CDCl3) δ -155.
Trimesitylstannylium Tetrakis(pentafluorophenyl)borate (TPF-
PB) (6). (a) From the Solvated Triethylsilyl Cation. In a N2-filled
glovebox, trityl TPFPB (160 mg, 0.17 mmol) was dissolved in dry
C6D6 (0.7 mL) in a valved 5 mm NMR tube. Addition of triethylsilane
(25 mg, 0.22 mmol) produced a light brown oil at the bottom. The
(33) Ray, G. J.; Kurland, R. J.; Colter, A. K. Tetrahedron 1971, 27, 735-
752.
(34) Gynane, M. J. S.; Lappert, M. F.; Riley, P. I.; Rivie`re, P.; Rivie`re-
Baudet, M. J. Organomet. Chem. 1980, 202, 5-12.
(35) Zigler, S. S.; Johnson, L. M.; West, R. J. Organomet. Chem. 1988,
341, 187-198.
(36) Lambert, J. B.; Stern, C. L.; Zhao, Y.; Tse, W. C.; Shawl, C. E.;
Lentz, K. T.; Kania, L. J. Organomet. Chem. 1998, 568, 21-31.
(37) Lapkin, I. I.; Sedel′nikova, V. A. Zh. Obshch. Khim. 1960, 30,
2771-2777 (Engl. Trans. 2753-2758).

5008 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Lambert et al.
colorless top phase containing the triphenylmethane byproduct was then
removed with a syringe. Allyltrimesitylstannane (98.3 mg, 0.19 mmol)
in C6D6 (0.5 mL) was added, and a deep red oil formed. The light
orange top phase was removed, and the oil was examined by NMR
spectroscopy: 1H NMR (C6D6) δ 1.95 (s, 18H), 2.01 (s, 9H), 6.63 (s,
6H); 13C NMR (C6D6) δ 21.0, 24.1, 130.4, 137.3 (d), 138.2 (d), 141.9,
146.3, 147.1, 149.0 (d); 119Sn NMR (C6D6) δ 806. (b) From the Free
â-Silyl Carbocation. In a N2-filled glovebox, trityl TPFPB (160 mg,
0.17 mmol) was dissolved in dry C6D6 (0.7 mL) in a valved 5 mm
NMR tube. Addition of triethylsilane (25 mg, 0.22 mmol) produced a
light brown oil at the bottom. The colorless top phase containing the
triphenylmethane byproduct was then removed with a syringe. 1,1-
Diphenylethene (40 mg, 0.22 mmol) was added, and the bottom oil
changed to deep green. Allyltrimesitylstannane (98.3 mg, 0.19 mmol)
in C6D6 (0.5 mL) was added, and a deep red oil formed. The light
orange top phase was removed, and the oil was examined by NMR
spectroscopy: 1H NMR (C6D6) δ 1.95 (s, 18H), 2.01 (s, 9H), 6.63 (s,
6H); 13C NMR (C6D6) δ 21.0, 24.1, 130.4, 137.3 (d), 138.2 (d), 141.9,
146.3, 147.0, 149.0 (d); 119Sn NMR (C6D6) δ 806.
Product Studies (Stannylium Ion). The red oil was washed with
dry toluene to remove all the nonpolar materials. Addition of Bu3SnH
(0.14 g, 0.48 mmol) changed the red oil on the bottom to brown.
Toluene (1.0 mL) was added. The top layer was removed with a syringe
and analyzed by GC/MS. Several products formed as Bu3SnH
decomposed in the GC column. The major product, however, was found
to be trimesitylstannane: MS (EI) m/z 477 (M+, 1), 415 (75), 359 (74),
237 (93), 207 (16), 119 (100).
NMR (C6D6) δ 23.08, 23.42, 24.52, 34.80, 45.50, 46.42, 124.3, 128.0,
130.8, 135.9, 137.1 (d), 137.8, 139.4 (d), 143.6, 149.5 (d), 152.8, 154.4,
157.1; 119Sn NMR (C6D6) δ 693.
Chlorotrimesitylgermane.38 To a N2-protected, 250 mL, round-
bottomed flask equipped with an addition funnel and a condenser and
containing GeCl4 (1.71 mL, 0.015 mol) was added mesitylmagnesium
bromide (48 mL, 0.048 mol) dropwise. The solution was allowed to
reflux overnight. The reaction solution was quenched with H2O. The
organic phase was washed with H2O and dried (MgSO4). Stripping the
solvent by rotary evaporation gave a yellow-white solid. Recrystalli-
zation from ethanol yielded 2.33 g (34%) of a white solid: mp 162-
164 °C; 1H NMR (CDCl3) δ 2.21 (s, 18H), 2.25 (s, 9H), 6.83 (s, 1H);
13C NMR (CDCl3) δ 21.06, 24.89, 129.83, 136.85, 139.45, 143.38.
Allyltrimesitylgermane (2). A 100 mL, round-bottomed flask fitted
with a rubber septum and a N2 inlet needle was charged with
allyltriphenyltin (1.43 g, 3.64 mmol) and a magnetic stirring bar.
Anhydrous diethyl ether (100 mL) was added to the flask via a syringe.
Phenyllithium (2.43 mL, 4.37 mmol) in ether and cyclohexane was
then added quickly, and a large amount of precipitate (Ph4Sn) formed
immediately. After 0.5 h, the suspension was transferred through a wide
bore cannula to an enclosed glass frit under N2. The solution was filtered
into a 250 mL flask, which had been charged with chlorotrimesitylger-
mane (1.0 g, 2.1 mmol) and a stirring bar. The red solution was stirred
at room temperature, and the reaction was monitored by 1H NMR. After
2 days, the reaction was complete. The reaction mixture was quenched
with H2O and extracted with hexane (200 mL). The organic solution
was dried (MgSO4), concentrated, and recrystallized from ethanol to
give a white solid (0.25 g, 49.2%): mp 160-163 °C; 1H NMR (CDCl3)
δ 2.12 (s, 18H), 2.26 (s, 9H), 2.48 (d, 2H), 4.88 (q, 2H), 5.65 (m, 1H),
6.84 (s, 6H); 13C NMR (CDCl3) δ 20.8, 24.6, 29.2, 115.7, 129.0, 136.8,
137.6, 138.6, 143.5. Anal. Calcd for C30H38Ge: C, 76.50; H, 8.07.
Found: C, 75.65; H, 8.07. Its X-ray structure is provided in the
Supporting Information.
Hexakis(2,4,6-triisopropylphenyl)cyclotristannoxane was prepared
in 40% yield by the method of Masamune and Sita.28
Chlorophenylbis(2,4,6-triisopropylphenyl)tin (7) was prepared in
81% yield by the method of Masamune and Sita:28 1H NMR (CDCl3)
δ 1.03 (d, 12H), 1.09 (d, 12H), 1.26 (d, 12H), 2.89 (septet, 4H), 3.11
(septet, 2H), 7.42 (m, 3H), 7.83 (m, 2H); 13C NMR (CDCl3) 24.00,
24.40, 25.09, 34.32, 37.68, 122.34, 128.74, 135.31, 138.95, 146.16,
150.99, 154.72. Its X-ray structure is included in the Supporting
Information.
Trimesitylgermylium Tetrakis(pentafluorophenyl)borate (TPF-
PB) (5). In a N2-filled glovebox, trityl TPFPB (160 mg, 0.17 mmol)
was dissolved in dry C6D6 (0.7 mL) in a valved 5 mm NMR tube.
Addition of triethylsilane (25 mg, 0.22 mmol) produced a light brown
oil at the bottom. The colorless top phase containing the triphenyl-
methane byproduct was then removed with a syringe. Allyltrimesi-
tylgermane (122.2 mg, 0.19 mmol) in C6D6 (0.5 mL) was added, and
a deep red oil formed. The light orange top phase was removed, and
the oil was examined by NMR spectroscopy: 1H NMR (C6D6) δ 2.11
(s, 18H), 2.24 (s, 9H), 6.82 (s, 6H); 13C NMR (C6D6) δ 21.4, 23.2,
Allylphenylbis(2,4,6-triisopropylphenyl)tin (8). To a stirred solu-
tion of 0.54 g (0.00083 mol) of bis(2,4,6-triisopropylphenyl)phenyltin
chloride was added slowly allylmagnesium chloride (0.54 mL, 2.0 M
in diethyl ether). After addition was complete, a white precipitate formed
immediately. The reaction was allowed to stir for 4 h. The reaction
was quenched with H2O, and the aqueous layer was extracted with
diethyl ether. The organic layer was dried (MgSO4), and the solution
was concentrated to give a yellow product. Further recrystallization
from ethanol gave 0.15 g (26%) of a white product: 1H NMR (CDCl3)
δ 0.89 (d, 12H), 0.99 (d, 12H), 1.20 (d, 12H), 2.48 (d, 2H), 2.89 (m,
6H), 4.88 (m, 2H), 6.03 (m, 1H), 6.93 (s, 4H), 7.34 (m, 2H); 13C NMR
(CDCl3) δ 24.2, 24.8, 25.0, 27.4, 34.6, 38.0, 112.8, 121.94, 128.9, 137.0,
137.5, 140.0, 142.5, 145.6, 150.1, 155.5; 119Sn NMR (CDCl3) δ -148.
Its X-ray structure is provided in the Supporting Information.
Phenylbis(2,4,6-triisopropylphenyl)stannylium Tetrakis(pentaflu-
orophenyl)borate (TPFPB). In a N2-filled glovebox, trityl TPFPB (160
mg, 0.17 mmol) was dissolved in dry C6D6 (0.7 mL) in a valved 5 mm
NMR tube. Addition of triethylsilane (25 mg, 0.22 mmol) produced a
light brown oil at the bottom. The colorless top phase containing the
triphenylmethane byproduct was then removed with a syringe. Al-
lylphenylbis(2,4,6-triisopropylphenyl)stannane (122.2 mg, 0.19 mmol)
in C6D6 (0.5 mL) was added, and a deep red oil formed. The light
orange top phase was removed, and the oil was examined by NMR
spectroscopy: 1H NMR (C6D6) δ 1.16 (d, 12H), 1.09 (d, 12H), 1.03
1
1
125, 130.7, 137.0 (d, J(CF) ) 245 Hz), 138.8 (d, J(CF) ) 244 Hz),
1
139.7, 141.9, 148.8, 149.1 (d, J ) 240 Hz).
Acknowledgment. This work was supported by the National
Science Foundation (Grant No. CHE-9725652). The authors
thank Prof. C. A. Reed for permission to quote unpublished
data and Prof. Claire Tessier for discussions about liquid
clathrates.
Supporting Information Available: Experimental details
and tables of results for the crystal structures of 2, 7, and 8.
This material is available free of charge via the Internet at
http://pubs.acs.org.
JA990389U
(38) Gynane, M. J. S.; Lappert, M. F.; Riley, P. I.; Rivie`re, P.; Rivie`re-
Bandet, M. J. Organomet. Chem. 1980, 202, 5-12.
(d, 12H), 2.1 (m, 2H), 2.7 (m, 4H), 7.04 (s, 2H), 7.1-7.3 (m, 5H); 13
C