Application of 13C NMR Spectroscopy and 13C-Labeled Benzylammonium Salts to the Study of Rearrangements of Ammonium Benzylides

Article abstract of DOI:10.1021/jo961231k

Ylides generated from N-(cyanomethyl)-N,N-dimethyl-N-[α-(trimethylsilyl)benzyl]ammonium chloride (4) and fluoride anion afford the products of [1,2] shift 11 and [2,3] shift 13. Formation of product 13 shows that, in the presence of water from TBAF, rearrangements and [1,3]H shift in ylide intermediates become competitive processes. The reaction of N-benzyl-N,N-dimethyl-N-[α-(trimethylsilyl)benzyl]ammonium bromide (5) and ¹³C labeled (at the benzyl carbon) salt 5* gave a mixture of 10, 14, and 15 as products of [1,4], [1,2], and [2,3] rearrangement, respectively. ¹³C NMR spectra of products derived from salt 5* exclude [1,3]H shift in ylide 9a^+-. Rearrangement of ylides generated from N-benzyl-N,N-dimethyl-N-[(dimethylphenylsilyl)methyl]ammonium bromide (6*) (enriched in ¹³C at benzyl carbon) and n-BuLi reveals that N,N-dimethyl-2-[(dimeth-ylphenylsilyl)methyl]benzylamine (20*) is not formed by a [1,4] shift but instead, via a [2,3] shift in silylmethylide followed by subsequent [1,4]Si and [1,2]H shift, as previously suggested in the literature. This mechanism is unique to some silyl-substituted ylides.

Full text of DOI:10.1021/jo961231k

452
J . Org. Chem. 1998, 63, 452-457
Ap p lica tion of 13C NMR Sp ectr oscop y a n d 13C-La beled
Ben zyla m m on iu m Sa lts to th e Stu d y of Rea r r a n gem en ts of
Am m on iu m Ben zylid es
Tadeusz Zdrojewski and Andrzej J on´czyk*
Faculty of Chemistry, Warsaw University of Technology, Koszykowa 75, 00-662 Warsaw, Poland
Received J uly 1, 1996 (Revised Manuscript Received November 21, 1997)
Ylides generated from N-(cyanomethyl)-N,N-dimethyl-N-[R-(trimethylsilyl)benzyl]ammonium chlo-
ride (4) and fluoride anion afford the products of [1,2] shift 11 and [2,3] shift 13. Formation of
product 13 shows that, in the presence of water from TBAF, rearrangements and [1,3]H shift in
ylide intermediates become competitive processes. The reaction of N-benzyl-N,N-dimethyl-N-[R-
(trimethylsilyl)benzyl]ammonium bromide (5) and 13C labeled (at the benzyl carbon) salt 5* gave
a mixture of 10, 14, and 15 as products of [1,4], [1,2], and [2,3] rearrangement, respectively. 13C
NMR spectra of products derived from salt 5* exclude [1,3]H shift in ylide 9a +-. Rearrangement
of ylides generated from N-benzyl-N,N-dimethyl-N-[(dimethylphenylsilyl)methyl]ammonium bro-
mide (6*) (enriched in 13C at benzyl carbon) and n-BuLi reveals that N,N-dimethyl-2-[(dimeth-
ylphenylsilyl)methyl]benzylamine (20*) is not formed by a [1,4] shift but instead, via a [2,3] shift
in silylmethylide followed by subsequent [1,4]Si and [1,2]H shift, as previously suggested in the
literature. This mechanism is unique to some silyl-substituted ylides.
Recently, we have shown that ylides 2+- and 3+-
,
Sch em e 1
generated from benzylammonium salts 1 by means of
suitable base-solvent systems, yield a mixture of products
2 and 3 via [2,3] (Sommelet-Hauser) and the hitherto
unknown [1,4] (“reverse” Sommelet-Hauser) rearrange-
ment, respectively (Scheme 1).1
The use of salt 1a enriched in 13C at the benzyl cyano
group allowed us to confirm this mechanism.2 However,
the salts 1 and some structurally related thienyl ana-
logues are so far the only examples of ammonium salts
that generate arylides prone to [1,4] shift.1-4 To collect
more information on this rearrangement, we investigated
ylides generated from silyl-substituted salts 4-6 and
fluoride anion or n-BuLi and applied 13C NMR technique
and labeled benzylammonium salts. This method, al-
ready shortly described by us,2 which allows us to
distinguish the paths of rearrangements of ylides, is now
developed and detailed. Regioselective generation of
ylides via desilylation of ammonium salts by means of
fluoride anion has been extensively studied.5
Resu lts a n d Discu ssion
Rea r r a n gem en ts of Ylid es Gen er a ted via Desil-
yla tion of Am m on iu m Sa lts. On the basis of the
(1) (a) J on´czyk, A.; Lipiak, D.; Sienkiewicz, K. Synlett 1991, 493.
(b) J on´czyk, A.; Lipiak, D. J . Org. Chem. 1991, 56, 6933.
(2) Zdrojewski, T.; J on´czyk, A. Tetrahedron Lett. 1995, 36, 1355.
(3) Lipiak, D. Ph.D. Dissertation, Technical University (Politech-
nika), Warsaw, 1991.
(4) A [1,4] shift has been observed in the case of some allylides: (a)
J enny, E. F.; Druey, J . Angew. Chem. 1962, 152. (b) Dietrich, W.;
Schultze, K.; Mu¨hlsta¨dt, M. J . Prakt. Chem. 1976, 318, 1008. (c)
J emison, R. W.; Laird, T.; Ollis, W. D.; Sutherland, T. O. J . Chem.
Soc., Perkin Trans. 1 1980, 1450. (d) Sugiyama, H.; Sato, Y.; Shirai,
N. Synthesis 1988, 988. (e) Honda, K.; Inoue, S.; Sato, K. J . Am. Chem.
Soc. 1990, 112, 1999. (f) Honda, K.; Inoue, S.; Sato, K. J . Org. Chem.
1992, 57, 428. (g) Hagen, J . P.; Lewis, K. D.; Lovell, S. W.; Rossi, P.;
Tezcan, A. Z. J . Org. Chem. 1995, 60, 7471. (h) Gulea-Purcarescu, M.;
About-J audet, E.; Collignon, N.; Saquet, M.; Masson, S. Tetrahedron
1996, 52, 2075. (i) J on´czyk, A.; Zdrojewski, T.; Grzywacz, P.; Balcerzak,
P. J . Chem. Soc., Perkin Trans. 1 1996, 2919.
(5) (a) Maeda, Y.; Shirai, N.; Sato, Y. J . Chem. Soc., Perkin Trans.
1 1994, 393 and references therein. (b) Kawanishi, N.; Shirai, N.; Sato,
Y.; Hatano, K.; Kurono, Y. J . Org. Chem. 1995, 60, 4272 and references
therein. (c) Zhang, Ch.; Maeda, Y.; Shirai, N.; Sato, Y. J . Chem. Soc.,
Perkin Trans. 1 1997, 25.
(6) Vedejs, E.; West, T. G. Chem. Rev. (Washington, D.C.) 1986, 86,
941.
reactivity of ylide 3+- (Scheme 1) benzylides 7+- and 9+-
,
regioselectively generated from the salts 4 and 5 by
means of fluoride anion (tetra-n-butylammonium fluo-
ride, TBAF, or cesium fluoride) in aprotic dipolar sol-
vents, were expected to undergo [1,4] rearrangement to
afford products 8 and 10, respectively (Schemes 2 and
3).
Commercial hydrated TBAF is not suitable for desily-
lation of ammonium salts because it favors equilibration
of the ylides formed.5,6 Furthermore, due to solvation of
the fluoride anion by water, TBAF‚3H2O did not desil-
ylate salt 5 effectively. Dehydration of molten TBAF‚3H2O
in vacuo usually causes partial decomposition of this
(7) Dehydrated TBAF contains up to 0.3 mol of water and some other
products: (a) Sharma, R. K.; Fry, J . L. J . Org. Chem. 1983, 48, 2112.
(b) Cox, D. P.; Terpin´ski, J .; Lawrynowicz, W. J . Org. Chem. 1984, 49,
3216.
S0022-3263(96)01231-5 CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/13/1998

Rearrangements of Ammonium Benzylides
J . Org. Chem., Vol. 63, No. 3, 1998 453
Ta ble 1. P r od u cts 11 a n d 13 F or m ed fr om th e Sa lt 4
Sch em e 2
ratio of the productsa
source
of F-
total yield
ca. (%)
entry
solvent
11
13
1
2
3
TBAF
TBAF
CsF
DMSO
HMPA
DMF
50
55
96
50
45
4
86
80
80
a
Determined by GC and 1H NMR.
Ta ble 2. P r od u cts 10, 14, a n d 15 fr om Sa lts 5 a n d 5* a n d
TBAF in DMSO
ratio of 13C/13C in products of
ratio of
products
rearrangement of 9*a +- and
9*b+- a,b
yield
entry salt (%) 10 14 15
10*
14*
15*
1
2
5
5
80c 12 35 53
22d 10 40 50
3
4
5*
5*
80c
20d 10 40 50
9
38 53
9.21
10.03
10.51
10.46
10.22
9.71
a
Ratio of the concentrations of 13C in positions marked by *
b
(Scheme 4) in products 10*, 14*, and 15* (sa/sb). Estimated error
is e10% for 10, and possibly lower in the case of products formed
with high yields. c “Anhydrous” TBAF. Reaction with CsF.
d
The same reaction carried out by using anhydrous CsF
in DMF afforded mainly product 11 (yield 80%) and only
a small amount of aminonitrile 13 (Table 1). Recently,
salt 4 (with bromide counterion) has been rearranged
under the same conditions to give only the product 11.5c
We attribute the generation of the more stable ylide
12+- (precursor of the product 13) to the presence of
water remaining in the “anhydrous” TBAF, which facili-
tates the equilibration of ylides (7+- h 12+-). Indeed,
the Karl Fischer method applied for water determination
in “anhydrous” TBAF obtained according to above cited
method8 indicated that the amount of water varies but
does not drop below ca. 10% (what corresponds to the
formula TBAF‚1.6H2O).
Sch em e 3
Reaction of salt 5 in DMSO either with “anhydrous”
TBAF, or with CsF, afforded a mixture of products 10,
14, and 15 via [1,4], [1,2], and [2,3] shifts, respectively.
A low yield was obtained using CsF, due to the poor
solubility of CsF in DMSO. On the other hand, reaction
of 5 with CsF in DMF led to formation of some unidenti-
fied products, apart from compounds 10, 14, and 15
(Scheme 3, Table 2).
material.7 Another procedure for drying this salt involves
azeotropic removal of water with benzene or benzene-
acetonitrile,8 followed by reduced pressure replacement
of these solvents with dry DMF, DMSO, or HMPA.
However, if salt 4 was allowed to react with such
“anhydrous” TBAF in DMSO or HMPA, it gave an
equimolar mixture of products 11 (via 7+-) and 13 (via
12+-), in a total yield 80-90%. These results show that
ylide 7+- prefers reaction via a [1,2] instead of a [1,4]
shift (for the reasons unknown at present) and that 7+-
equilibrates with ylide 12+-, leading to formation of
R-amino nitrile 13 (Table 1, Scheme 2) via [2,3] rear-
rangement.
Degenerate [1,3]H shift in the ylide 9+- intermediate
would have no consequence on the product distribution
from an unlabeled precursor. Therefore, to determine
whether rearrangements are preceded by [1,3]H shift in
ylide 9+-, the salt 5* 13C labeled at the benzyl carbon
was synthesized and allowed to react with TBAF or CsF
in DMSO, and the products were analyzed by 13C NMR
in a manner previously described by us.2 In all cases
studied the product mixture consisted of 10*a ,b, 14*a ,b,
and 15*a ,b (Scheme 4).
The ratio of 13C/13C in products derived from salt 5*
shows that very little if any [1,3]H shift in ylide inter-
mediate 9*a +- occurred (Table 2), due to equal acidities
of both benzylic positions. On the other hand, partial
[1,3]H shift in ylide 7+- leading to ylide 12+- is well
explained by better stabilization of the latter species
(Scheme 2, Table 1).
Structure 5 is another example of a salt that generates
ylides prone to [1,4] shift, although to only a small extent.
Rea r r a n gem en ts of Ylid es Gen er a ted via Rea c-
tion of a Ba se w ith Am m on iu m Sa lts. Our earlier
Preferential [2,3] rearrangement of cyanomethylides
is well documented, for example, in the products of
N-aryl- or N-(hetarylmethyl)-N-(cyanomethyl)-N,N-di-
alkylammonium salts and a base.9
(8) (a) Moss, R. A.; Kmiecik-Lawrynowicz, G.; Krogh-J espersen, K.
J . Org. Chem. 1986, 51, 2168. (b) Moss, R. A.; Zdrojewski, T. J . Phys.
Org. Chem. 1990, 3, 694.
(9) (a) Mander, L. N.; Turner, J . V. J . Org. Chem. 1973, 38, 2915.
(b) Kocharyan, S. T.; Karapetyan, V. E.; Babayan, A. T. Zh. Org. Khim.
1985, 21, 56. (c) Kocharyan, S. T.; Ogandzhanyan, S. M.; Razina, T.
L.; Babayan, A. T. Zh. Org. Khim. 1982, 18, 1861.

454 J . Org. Chem., Vol. 63, No. 3, 1998
Zdrojewski and J on´czyk
Sch em e 4
investigations indicated that concerted mechanisms with
six-electron aromatic transition states and suprafacial-
suprafacial characteristics are possibly involved in the
[1,4] rearrangement of ammonium benzylides.1 However,
another explanation for the formation of the products of
formal [1,4] shift of ammonium benzylides has been
proposed.10 Thus, treating salt 6 with NaNH2/NH3(liq)
or n-BuLi/THF led to the formation of product 20, as well
as 19 and 21.10 The formation of 20 was explained by
sequential [1,4]Si and [1,2]H shifts from intermediate 18,
which in turn is a result of [2,3] rearrangement of
silylmethylide 16+-
. These previous studies gave no
experimental evidence that would support the suggested
route a, 18 f 20 (Scheme 5), but they referenced the
precedence of anionic C-C shifts of silyl group during
reactions of R- and â-silyl substituted amines.11 Alter-
natively, product 20 may result via [1,4] rearrangement
of benzylide 17+-, which also appears as the precursor
of amine 21 (route b, Scheme 5).
13C labeling (*) the salt 6 at the benzyl carbon should
allow a choice between these two alternative mechanisms
(Scheme 5). Thus, salts 6 and 6* (enriched up to ca. 10%
in 13C at the benzyl carbon, Figure 1) were prepared and
allowed to react with n-BuLi following literature proce-
1
dures,10 and the product mixtures were analyzed by H
and 13C NMR.
We found that this reaction afforded the three silyl-
amines 19-21, as described, in the ratio of ca. 1:5:1, with
a yield of ca. 100%. The same reaction carried out with
6* resulted in a similar yield and product ratio. Deter-
mination of the abundance of 13C in both benzyl positions
F igu r e 1. 13C NMR spectra (aliphatic carbons region) of salts
6 and 6* (enriched up to ca. 10% in 13C at the benzyl carbon).
For full description of spectra see Experimental Section.
in 20* (by comparison with the natural abundance of 13
C
in 20) from the 13C NMR spectra (Figure 2) indicated that
all of the marker was located at the benzyl carbon R to
Si (20a * not 20b*, Scheme 5).
In this case, 13C labeling at the benzylic and/or either
of the cyano groups in 1 does not lead to unequivocal
results because identical structures are formed irrespec-
tive of [1,4] rearrangement or [2,3], then [1,4]NMe2 and
[1,2]H shifts. Assuming that the route leading to forma-
tion of 3, visualized in Scheme 6, is valid, the amino
groups in R-cyano amines should be susceptible to attack
by nucleophiles (or electron pairs), which is against the
This result was consistent with product 20 being
formed via the route suggested by Sato et al.10 (i.e., 16+-
f 18 f 20 not 17+- f 20, Scheme 5).
At this point in our investigations the question arises
whether product 3 might result from 1 via a mechanism
similar to that which forms 20 from 6 (Scheme 6).
(10) Sato, Y.; Yagi, Y.; Koto, M. J . Org. Chem. 1980, 45, 613.
(11) (a) Sato, Y.; Ban, Y.; Shirai, H. J . Organomet. Chem. 1976, 113,
115. (b) Sato, Y. Ban, Y.; Aoyama, T.; Shirai, H. J . Org. Chem. 1976,
41, 1962. (c) Sato, Y.; Toyo’oka, T.; Aoyama, I.; Shirai, H. J . Org. Chem.
1976, 41, 3559. (d) Sato, Y.; Kobayashi, Y.; Sugiura, M.; Shirai, H. J .
Org. Chem. 1978, 43, 199. (e) Himbert, G. J . Chem. Res., Synop. 1978,
104.
(12) (a) Bruylants, P. Bull. Soc. Chim Belg. 1924, 33, 467. (b) Kalir,
A.; Edery, H.; Pelah, Z.; Balderman, D.; Porath, G. J . Med. Chem. 1969,
12, 473. (c) Ahlbrecht, H.; Dollinger, H. Synthesis 1985, 743. (d)
Kudzma, L. V.; Spencer, K. H.; Anaquest, S. A. S. Tetrahedron Lett.
1988, 29, 6827. (e) Zdrojewski, T.; J on´czyk, A. Synthesis 1990, 224.
Zdrojewski, T.; J on´czyk, A. Liebigs Ann. Chem. 1993, 375.

Rearrangements of Ammonium Benzylides
J . Org. Chem., Vol. 63, No. 3, 1998 455
Sch em e 5
F igu r e 2. 13C NMR spectra (aliphatic carbons region) of the
mixture of silylamines 19-21 (upper) and 19*-21* (lower).
Signals with reported shifts are ascribed to silylamine 20 (20*),
see Experimental Section for details.
Sch em e 6
experimental facts indicating that the cyano group
exhibits such reactivity.12 Therefore, we should observe
the formation of structure 22, which, in fact, is not found.
So route a in Scheme 5 is probably unique to some silyl-
substituted ylides only.
Our data show that ylides generated from the salts of
select structures (like 1 or 5) in some base-solvent
systems undergo [1,4] rearrangement. We have also
demonstrated that 13C NMR spectroscopy of the products
formed from reactions of a base with ammonium salts
enriched in 13C at the benzyl carbon atom is a convenient
tool to study rearrangements of ammonium ylides.
column. 1H and 13C NMR spectra were measured at 200 and
50 MHz, respectively. Additionally, for carbon peak assign-
ment the attached proton test (APT)13 was applied.
13C NMR spectra of compounds of natural and enriched in
13C isotopomers populations were measured (concentration 100
mg/mL; at 1 s; decoupler profile WALZ 16; nt 10000). The
degree of enrichment in 13C isotopomers (s) was calculated
from the following equation
IA*I
s )
S(1.108%)
IS*IA
Exp er im en ta l Section
where IA, IA* and IS, IS* denote integral intensities of 13C
signals of investigated carbon A and standard one (S) in
spectra of compound of natural and enriched in 13C isoto-
pomers population, respectively, and 1.108% means natural
content of 13C isotopomers.2
Gen er a l Meth od s. Commercial TBAF‚3H2O, Ph13CO2H
(99% 13C), (chloromethyl)dimethylphenylsilane, and a 2 M
solution of n-BuLi in cyclohexane (all Aldrich) were used.
Melting points (measured with a capillary melting point
apparatus) and boiling points are uncorrected. Gas chroma-
tography (GC) analyses were performed on a HP 50+ capillary
(13) Patt, S. L.; Shoolery, J . N. J . Magn. Reson. 1982, 46, 535.

456 J . Org. Chem., Vol. 63, No. 3, 1998
Zdrojewski and J on´czyk
The 13C/13C ratios of the concentrations of markers in the
two considered positions A and B in salt 5* and products 10*,
14*, and 15* were calculated as follows
analyzed by means of GC and NMR spectroscopy. The yields
of crude products were 80-100%.
Salt 4. The product mixture consisted of equimolar amounts
of 11 and 13 (Table 1). A benzene solution of the products,
obtained after extraction, was shaken with 3% aqueous HCl
(3 × 10 mL), the phases were separated, the combined water
phases were made alkaline with solid NaHCO3 and extracted
with benzene (3 × 10 mL), and the organic extracts were
washed with brine, and dried (Na2SO4). The solvent was
evaporated, and the residue was distilled to give 11 (0.2 g,
46%): bp 80 °C (0.075 Torr); mp 56-58 °C (cyclohexane-
IA*IB
13C/13C )
IB*IA
where IA, IA* and IB, IB* denote integral intensities of 13C
signals of investigated carbons A and B in spectra of com-
pounds of natural and enriched in 13C isotopomers population,
respectively. The estimated error of this approach is 10%.
N-Cya n om et h yl-N,N-d im et h yl-N-[r-(t r im et h ylsilyl)-
ben zyl]a m m on iu m Ch lor id e (4). A mixture of [R-(dimethyl-
amino)benzyl]trimethylsilane14 (11.8 g, 56 mmol) and chloro-
acetonitrile (6.3 g, 84 mmol) was heated at ca. 40 °C for 24 h,
diluted with the mixture of benzene and acetone (2:1), and
thoroughly stirred, and the solid product was filtered to give
1
hexane); H NMR (CDCl3) δ 2.22 (s, 6H), 2.78 and 2.83 (part
AB of ABX, J AB ) 16.77, J AX ) 7.65, J BX ) 5.37 Hz, 2H), 3.54
(part X of ABX, 1H), 7.30-7.45 (m, 5H); 13C NMR (CDCl3) δ
22.9, 42.7, 66.6, 117.9, 127.6, 128.2, 128.6, 138.3. Anal. Calcd
for C11H14N2: C, 75.82; H, 8.10; N, 16.08. Found: C, 75.69;
H, 8.30; N, 16.07. The organic phase, after being shaken with
aqueous HCl, was washed with aqueous NaHCO3 and dried
(Na2SO4), the solvent was evaporated, and the residue was
distilled to give 13 (0.15 g, 35%): bp 70 °C (0.075 Torr) [lit.9c
bp 105-107 °C (5 Torr)]; 1H NMR (CDCl3) δ 2.29 (s, 6H), 2.39
(s, 3H), 4.87 (s, 1H), 7.16-7.32 (m, 3H), 7.50-7.57 (m, 1H);
13C NMR (CDCl3) δ 18.3, 41.1, 60.8, 114.9, 125.5, 128.0, 128.7,
130.7, 131.5, 137.1. An authentic sample of 13 was prepared
from 2-methylbenzaldehyde, aqueous dimethylamine, and
sodium cyanide [yield 65%, bp 79 °C (0.1 Torr)], as described
for R-(dimethylamino)phenylacetonitrile.16
1
4 (10.0 g, 63%): mp 170-173 °C; H NMR (DMSO-d6) δ 0.17
(s, 9H), 3.27 (s, 3H), 3.38 (s, 3H), 4.75 and 4.89 (AB, J ) 16.3
Hz, 2H), 4.92 (s, 1H), 7.45-7.65 (m, 5H); 13C NMR (DMSO-
d6) δ 0.0, 51.7, 52.9, 53.7, 73.7, 112.1, 129.1, 129.4, 129.8, 130.6,
131.4, 133.8. Anal. Calcd for C14H23ClN2Si: C, 59.45; H, 8.20;
N, 9.90. Found: C, 59.28; H, 7.98; N, 9.65.
N-Ben zyl-N,N-dim eth yl-N-[r-(tr im eth ylsilyl)ben zyl]am -
m on iu m Br om id e (5). A solution of [R-(dimethylamino)-
benzyl]trimethylsilane14 (2.1 g, 1 mmol) and benzyl bromide
(2.1 g, 1.2 mmol) in ethanol (3 mL) was heated at ca. 40 °C for
24 h and diluted with ethyl acetate (10 mL), and the solid
product was filtered and washed with ethyl acetate to give 5
Sa lt 5. Treating salt 5 or 5* with “anhydrous” TBAF in
DMSO afforded the mixtures of 10 (10*), 14 (14*), and 15 (15*)
(Table 2, entries 1 and 3), ratios of which were determined by
GC and 1H NMR spectroscopy. These mixtures were also
analyzed by 13C NMR to determine the 13C/13C ratios of the
concentration of marker in both benzylic positions (Scheme
4). Reference samples of the products 10, 14, and 15 were
prepared as follows.
N,N-Dim eth yl-2-ben zylben zyla m in e (10) via rearrange-
ment of ylide generated from benzhydryltrimethylammonium
iodide by means of NaNH2 in NH3(liq):17 bp 189-191 °C (33
Torr) [lit.17 bp 189-191 °C (33 Torr)]; 1H NMR (CDCl3) δ 2.29
(s, 6H), 3.42 (s, 2H), 4.25 (s, 2H), 7.18-7.40 (m, 9H); 13C NMR
(CDCl3) δ 38.3 (CH2Ph), 45.5, 62.0 (CH2N), 125.8, 126.0, 127.2,
128.2, 128.8, 130.3, 130.5, 137.2, 139.8, 141.0.
1
(3.6 g, 96%): mp 188-190 °C; H NMR (DMSO-d6) δ 0.23 (s,
9H), 2.85 (s, 3H), 2.94 (s, 3H), 4.47 and 4.67 (AB, J ) 12.3 Hz,
2H), 5.11 (s, 1H), 7.45-7.75 (m, 10H); 13C NMR (DMSO-d6) δ
0.4 (SiMe3), 49.7 (+NMe), 49.8 (+NMe), 67.4 (CH2Ph), 74.2
(CHPh), 128.1, 128.7, 129.0, 129.3, 129.4, 130.3, 131.1, 132.6,
133.4, 134.0. Anal. Calcd for C19H28BrNSi: C, 60.03; H, 7.46;
N, 3.70. Found: C, 60.11; H, 7.48; N, 3.58.
Sa lt 5*. Benzoic-carboxy-13C acid (1.0 g, 8.2 mmol) was
added to benzoic acid (10.0 g, 82 mmol), and this mixture was
converted to benzyl bromide via the following intermediates:
Ph13COCl (with SOCl2, 90%), Ph13CO2Et (with EtOH, 90%),
Ph13CH2OH (with LAH in Et2O, 96%), and Ph13CH2Br (with
triphenylphosphine dibromide,15 96%). Starting from this
benzyl bromide and [R-(dimethylamino)benzyl]trimethyl-
silane,14 the salt 5* (mp 187-190 °C) was prepared as
described above for 5.
1,2-Dip h en yl-1-(d im eth yla m in o)eth a n e (14) was ob-
tained by reduction of 1,2-diphenyl-1-(dimethylamino)ethene18
with NaBH4 in methanol: bp 140-142 °C (0.4 Torr) [lit.19 bp
5*: 13C NMR (DMSO-d6) δ 0.4 (SiMe3), 49.7 (+NMe2), 67.5
(*CH2Ph), 74.1 (CHPh), 128.0, 128.7, 128.9, 129.2, 129.3, 130.2,
131.0, 132.5, 133.3, 134.0.
142 °C (1 Torr)]; H NMR (CDCl3) δ 2.33 (s, 6H), 3.02 (dd, J
) 13.1, 3J ) 9.55 Hz, 1H), 3.38 (dd, 2J ) 13.1, 3J ) 5.0 Hz,
1H), 3.52 (dd, 3J ) 9.55, 3J ) 5.0 Hz, 1H), 7.00-7.35 (m, 10H);
13C NMR δ 39.9 (CH2), 42.9, 72.6 (CHN), 125.6, 126.9, 127.7,
127.8, 128.6, 129.2, 139.4, 139.6.
1
2
Calculated concentration of 13C in both benzyl positions:
s*CH ) 10.75% (relative to integral intensity of signals of
Ph
+NMe22) and 10.85% (relative to integral intensity of signal of
SiMe3);
2-Meth yl-N,N-d im eth ylben zh yd r yla m in e (15) was ob-
tained from 2-methylbenzhydryl chloride20
with an excess of
sCHPh ) 1.08% (relative to +NMe2) and 1.09% (relative to
SiMe3).
aqueous dimethylamine: mp 47-49 °C (lit.21
mp 46 °C); 1H
NMR (CDCl
3) δ 2.22 (s, 6H), 2.38 (s, 3H), 4.29 (s, 1H), 7.00-
Calculated ratio of 13C/13C in both benzyl positions: 9.95.
Gen er a l P r oced u r e for th e Rea ction s of 4, 5, a n d 5*
w ith TBAF in DMSO. Commercial TBAF‚3H2O (4.0 g, 12.5
mmol) was dissolved in the mixture of benzene/acetonitrile (1/1
v/v, ca. 50 mL), and water was removed with these solvents
on a rotary evaporator. The procedure was repeated twice,
and the residual sticky oil was evaporated at 40 °C/0.01 Torr
for 30 min to obtain a white semisolid containing ca. 10% water
(Karl Fischer titration) that was dissolved in dry DMSO (15
mL). This solution was placed in an ice bath, and during
stirring the salt 4, 5, or 5* (2.5 mmol) was added. The mixture
was stirred at 20-25 °C for 24 h, poured into water (100 mL),
and extracted with benzene (3 × 10 mL), the organic extracts
were washed once with water and twice with brine and dried
(Na2SO4). The solvent was evaporated, and the residue was
7.10 (m, 2H), 7.18-7.32 (m, 4H), 7.40-7.50 (m, 2H), 7.83 (d,
J ) 7.6 Hz, 1H); 13C NMR (CDCl3) δ 19.9 (CH3Ar), 44.9, 72.8
(CHN), 126.2, 126.3, 126.8, 127.0, 128.3, 128.4, 130.4, 135.4,
141.5, 142.4.
R ea ct ion of Sa lt 4 w it h CsF in DMF . Dried (180 °C,
reduced pressure) CsF (0.76 g, 5 mmol) was suspended in
freshly dried DMF (5 mL), the mixture was stirred and cooled
to ca. 0 °C, and salt 4 (0.28 g, 1 mmol) was added in one
portion. The cooling bath was removed, stirring was continued
for 24 h, and the mixture was worked up as described in the
General Procedure. The crude oil (0.20 g) was analyzed by
GC to show 86% of 11 and ca. 3% of 13 (Table 1, entry 3) and
(17) Kantor, S. W.; Hauser, C. R. J . Am. Chem. Soc. 1951, 73, 4122.
(18) Hauser, C. R.; Taylor, H. M.; Ledford, T. G. J . Am. Chem. Soc.
1960, 82, 1786.
(14) Duff, J . M.; Brook, A. G. Can. J . Chem. 1977, 55, 2589.
(15) Wijnberg, J . B. P. A.; Wiering, P. G.; Steinberg, H. Synthesis
1981, 901.
(16) Taylor, H. M.; Hauser, C. R. Organic Syntheses; Wiley: New
York, 1973; Coll. Vol. V, 437.
(19) Stewart, A. T. J r.; Hauser, C. R. J . Am. Chem. Soc. 1955, 77,
1098.
(20) Norris, J . F.; Blake, J . T. J . Am. Chem. Soc. 1928, 50, 1808.
(21) Wragg, A. H.; Stevens, T. S.; Ostle, D. M. J . Chem. Soc. 1958,
4057.

Rearrangements of Ammonium Benzylides
J . Org. Chem., Vol. 63, No. 3, 1998 457
solidified when degassed in vacuo. Recrystallization afforded
11 (0.14 g, 80%), mp 56-58 °C (cyclohexane-hexane).
Rea ction of Sa lt 5 or 5* w ith CsF in DMSO. The
suspension of dried CsF (0.76 g, 5 mmol) in dry DMSO (6 mL)
was stirred while salt 5 or 5* (0.38 g, 1 mmol) was added in
one portion. The mixture was stirred at rt for 48 h and worked
up as described in the General Procedure to give a mixture of
amines 10, 14, and 15 or 10*, 14*, and 15* (Table 2, entries 2
and 4), which were analyzed by GC (purity ca. 90%) and 1H
and 13C NMR spectroscopy in the manner already described
above.
Rea ction s of Sa lts 6 a n d 6* w ith n -Bu Li. Salt 6 was
prepared via reaction of N,N-dimethyl-N-(dimethylphenylsi-
lyl)methylamine [synthesized in turn from (chloromethyl)-
dimethylphenylsilane and dimethylamine11c] with benzyl bro-
mide by literature procedure.10 The salt 6* was obtained
similarly using benzyl bromide enriched in 13C at benzyl
carbon to ca. 10%.
1:5:1. Distillation of this mixture gave a fraction of bp 150-
155 °C (2 Torr) containing silylamines 19, 20, and 21 in the
ratio of 1:6:1 (1H NMR).
The salt 6* (1.82 g, 5 mmol) was treated with a 2 M solution
of n-BuLi in cyclohexane in the same manner as salt 6. The
crude products (yield ca. 100%) were distilled, and the fraction
of bp 150-155 °C (2 Torr) was collected and analyzed by 1H
NMR to show silylamines 19*, 20*, and 21* in the ratio of
1:8:1.
Both mixtures (19, 20, 21 and 19*, 20*, 21*) were analyzed
by 13C NMR. The large excess of amine 20 (20*) in the samples
[relative to 19 (19*) and 21 (21*)] allowed the separation of
individual resonance signals of 20 and 20*.
20: δ -2.9 (SiMe2), 22.3 (SiCH2, J Si-C ) 47.6 Hz), 45.4
(NMe2), 62.2 (CH2N), 124.0, 126.7, 127.7, 128.9, 129.3, 130.2,
133.6, 135.7, 139.0, 139.1.
20*: δ -2.9 (SiMe2), 22.3 (SiCH2, J Si-C ) 47.6 Hz), 45.4
(NMe2), 62.2 (CH2N), 124.0, 126.7, 127.7, 129.0, 129.4, 130.2,
133.6, 135.7, 139.0, 139.1.
6*: yield 90%; mp 157-159 °C; 1H NMR (CDCl3) δ 0.51 (s,
6H), 3.00 (s, 6H), 3.65 (s, 2H), 4.96 (s and d, J C-H ) 145.7 Hz,
2H, *CH2), 7.20-7.30 and 7.45-7.55 (m, 10H); 13C NMR
(CDCl3) δ -2.3 (SiMe2, J Si-C ) 55.2 Hz), 52.5 (+NMe2), 57.0
(SiCH2, s ) 1.09%, concentration of 13C relative to integral
intensity of signal of +NMe2), 71.0 (Ph*CH2, s ) 10.15%,
concentration of 13C relative to integral intensity of signal of
+NMe2), 127.6 (J Si-C ) 45.9 Hz), 128.3, 128.5, 130.1, 130.2,
133.0, 133.7.
The determined degree of enrichment in 13C (s) at both
benzyl carbons in 20* gave the following results:
sCH ) 10.31% and sCH ) 1.23% (both relative to NMe2).
Si
N
2
2
Ack n ow led gm en t. The financial support of this
work by the State Committee for Scientific Research,
Warsaw, Poland (Grant No. 2 2616 92 03), is gratefully
acknowledged.
Treating of the salt 6 with a 2 M solution of n-BuLi in
cyclohexane according to literature procedure10 led to forma-
tion of the mixture of 19, 20, and 21 (ca. 100%) in a ratio of
J O961231K