Interception of deaminatively generated benzyl carbenium ions by acetone

Article abstract of DOI:10.1021/jo991827i

Essentially free benzyl carbenium ions were generated via protonation of phenyldiazomethane with benzoic acid in acetone. Interestingly, no proton transfer occurred below -20 °C. After protonation and dediazoniation of the diazoalkane at -20 °C, the solvent was found to intercept the deaminatively generated carbocations to yield initially the corresponding O-benzyl oxonium ion and benzyl benzoate. The onium ion, however, was labile under the reaction conditions, and decomposed into a cascade of products whose concentrations as a function of time were used to trace the reaction pathway. Thus, the O-benzyl oxonium ion reacted with benzoate ion to yield (2- benzyloxy)isopropyl benzoate; subsequent decomposition of this O-benzyl-O- benzoyl ketal produced 2,2-dibenzyloxypropane (a dibenzyl ketal), 2- benzyloxypropene, and benzyl alcohol. In a related study, benzyl cations were generated via thermolyses of N-benzyl-N-nitroso-O-benzoyl hydroxylamine at 0 and -70 °C. The product distributions were found to be temperature-dependent and different from that in the PhCHN₂ + PhCO₂H case.

Full text of DOI:10.1021/jo991827i

J . Org. Chem. 2000, 65, 1825-1829
1825
In ter cep tion of Dea m in a tively Gen er a ted Ben zyl Ca r ben iu m Ion s
by Aceton e
Fenhong Song,† Ron W. Darbeau,* and Emil H. White‡
Departments of Chemistry, The J ohns Hopkins University, Baltimore, Maryland 21218,
and McNeese State University, Lake Charles, Louisiana 70609
Received November 30, 1999
Essentially free benzyl carbenium ions were generated via protonation of phenyldiazomethane with
benzoic acid in acetone. Interestingly, no proton transfer occurred below -20 °C. After protonation
and dediazoniation of the diazoalkane at -20 °C, the solvent was found to intercept the
deaminatively generated carbocations to yield initially the corresponding O-benzyl oxonium ion
and benzyl benzoate. The onium ion, however, was labile under the reaction conditions, and
decomposed into a cascade of products whose concentrations as a function of time were used to
trace the reaction pathway. Thus, the O-benzyl oxonium ion reacted with benzoate ion to yield
(2-benzyloxy)isopropyl benzoate; subsequent decomposition of this O-benzyl-O-benzoyl ketal
produced 2,2-dibenzyloxypropane (a dibenzyl ketal), 2-benzyloxypropene, and benzyl alcohol. In a
related study, benzyl cations were generated via thermolyses of N-benzyl-N-nitroso-O-benzoyl
hydroxylamine at 0 and -70 °C. The product distributions were found to be temperature-dependent
and different from that in the PhCHN2 + PhCO2H case.
In tr od u ction
ions as the latter process leads to internal collapse of the
ion-pair to form the corresponding ester. The yield of
solvent-derived products (SDPs) is increased by condi-
tions that decrease the diffusion rate of the nitrogenous
molecule,1b,2 by use of large mole fractions of highly
nucleophilic solvents,1b,d-f and through use of poorly
nucleophilic counterions.1b,d,e Alkenes can also form from
the cation-counterion reaction if the former possesses a
removable â-hydrogen (eq 1) and if the basicity of the
latter is at or above some threshold level.
Quite recently, a number of articles1 have appeared
in the literature concerning the reactivity, synthetic and
mechanistic utility, and stereochemistry of deaminatively
generated carbocations. These carbocations, formed as
part of a nitrogenous-molecule separated ion-pair (NSIP),
may be produced via several precursors1b and may
represent the most reactive carbocations accessible in the
solution phase.1b,c The high reactivity1 of deaminatively
generated carbocations is due, in part, to the low activa-
tion energy required for the loss of nitrogen (or of N2O)
that allows the cation to be formed with minimal solvent
participation.1 Additionally, the temporary screening of
the cation from its counterion by the sheer physical
presence of the nitrogenous molecule results in the
maximal positive charge at the electron-deficient center.1
These exceedingly reactive cations can be chosen so
that â-eliminations and rearrangements are not possible
and reaction with nucleophiles is their only fate. The
nucleophiles available to them are the counterion and
the solvent. Cation-solvent reactions must occur before
diffusion of the nitrogenous molecule from between the
Two previous studies1d,e have dealt with the intercep-
tion of deaminatively generated benzyl carbenium ions
by acetonitrile in the presence of counterions of low
nucleophilicity (triflate)1d and of modest nucleophilicity
(pivalate)1e (npivalate ≈ 2.7; ntriflate < 0)3. In those cases,
thermolyses of N-benzyl-N-nitrosoamides at 25 °C were
used to generate the benzyl carbenium ion. In the present
work in acetone, benzyl carbenium ions originated via
protonation of phenyldiazomethane with benzoic acid at
-20 °C (eq 2a) and via nitrosation of N-benzyl-O-benzoyl
hydroxylamine at 0 and -70 °C (eq 2b). The product
distributions from these runs were used to determine the
prevailing reaction mechanisms.
* Address correspondence to this author at McNeese State Univer-
sity.
† This author is now at The Center for Advanced Research in
Biotechnology, 9600 Gudelsky Drive, Rockville, MD 20850.
‡ This paper is dedicated to the memory of Professor Emil H. White,
1926-1999.
(1) (a) White, E. H.; Field, K. W.; Hendrickson, W. H.; Dzadzic, P.;
Roswell, D. F.; Paik, S.; Mullen, P. W. J . Am. Chem. Soc. 1992, 114,
8023. (b) Darbeau, R. W.; White, E. H.; Song, F.; Darbeau, N. R.; Chou,
J . C. J . Org. Chem. 1999, 64(16), 5966. (c) White, E. H.; Darbeau, R.
W.; Chen, Y.; Chen, D.; Chen, S. J . Org. Chem. 1996, 61, 7986. (d)
White, E. H.; De Pinto, J . T.; Polito, A. J .; Bauer, I.; Roswell, D. F. J .
Am. Chem. Soc. 1988, 110, 3708. (e) Darbeau, R. W.; White, E. H.;
Nunez, N. P.; Coit, B. C.; Daigle, M. A. J . Org. Chem. 2000, 65, 1115.
(f) Darbeau, R. W.; White, E. H. J . Org. Chem. 1997, 62, 8091. (g) Song,
F.; St. Hilaire, V. R.; White, E. H. Org. Lett. 1999, 1 (12), 1957. (h)
Wiberg, K. B.; Osterle, C. G. J . Org. Chem. 1999, 64, 7756. (i) Wiberg,
K. B.; Osterle, C. G. J . Org. Chem. 1999, 64, 7763. (j) Wiberg, K. B.;
Shobe, D. S. J . Org. Chem. 1999, 64, 7768. (k) Darbeau, R. W.; Delaney,
M. S.; Ramelow, U.; J ames, K. R. Org. Lett. 1999, 1(5), 761.
(2) The yield of SDP also rises with decreasing temperature and to
a lesser extent as the nitrogenous molecule is changed from N2 to
N2O.1b
(3) Swain, C. G.; Scott, C. B. J . Am. Chem. Soc. 1953, 75, 141.
10.1021/jo991827i CCC: $19.00 © 2000 American Chemical Society
Published on Web 03/01/2000

1826 J . Org. Chem., Vol. 65, No. 6, 2000
Song et al.
Ta ble 1. P r od u ct Distr ibu tion a fr om th e Rea ction of
P h en yld ia zom eth a n e w ith Ben zoic Acid in Aceton e
product distribution
run no.
5
9a
10
11
12
1b,c
1b,d
2e,f
3g
58
60
54
56
33
16
13
41
4
6
9
2
0
13
7
5
5
17
1
0
a
Standard deviation ∼0.5. bRuns performed in acetone-d6/
c
acetone-h6 (9:1; vol/vol) and started at -78 °C. Sample stored at
4 °C for 23 h after completion of run. dSolvent removed in vacuo
e
and residue redissolved in CDCl3. Runs performed in acetone-h6
at -20 °C; data obtained in CDCl3 after solvent removal in vacuo;
f
preferential loss and some hydrolysis were likely. Sample stored
at 4 °C for 4 days after completion of run. gRuns performed in
acetone-d6 at -20 °C.
Sch em e 1. P r op osed Mech a n ism for th e
F or m a tion of (2-Ben zyloxy)isop r op yl Ben zoa te
Resu lts a n d Discu ssion
Acidification of phenyldiazomethane (eq 2a) and nit-
rosation of N-benzyl-O-benzoyl hydroxylamine (eq 2b)
were employed here because of the high speed of these
reactions in producing the cations (vide infra). The former
reaction is completed after ∼3 h at -20 °C whereas the
latter is essentially instantaneous even at -78 °C.1b
These rapid decompositions permit spectroscopic obser-
vation of thermolabile, transient intermediates. With few
exceptions, runs were performed in acetone-d6 (99%) in
NMR tubes to allow direct 1H NMR analyses of the
products.
of 5, 16% of 9a , 13% of 12, 5% of 2,2-dibenzyloxypropane
(11), and 6% of 10 (Table 1, Schemes 1 and 2).
Th e La bility of th e On iu m Ion a n d a Ca se of Slow
P r oton Tr a n sfer . In an attempt to detect the putative
oxonium ion, 8, a solution of benzoic acid in acetone-d6/
acetone-h6 (1/9, v/v) at -78 °C was added to phenyldia-
zomethane in the same solvent system at the same
temperature. The 1H NMR spectrum showed that no
reaction had occurred in over 20 min at -78 °C. When
the mixture was warmed stepwise in 10 °C increments
(10 min per ramp and a 10 min hold time), reaction was
observed to begin only at -20 °C (vide supra) as evi-
denced by the appearance of product peaks at δ 5.42 from
benzyl benzoate and at δ 4.88 from 9a .4
On the basis of this NMR study, it appears that (1)
protonation of phenyldiazomethane with benzoic acid at
temperatures <-20 °C is quite slow and (2) the onium
ion is undetected by NMR spectroscopy at -20 °C. This
observation may arise because 8 is too labile under these
conditions or because of the slow kinetics of phenyldia-
zomethane decomposition by benzoic acid <-20 °C.
However, (3) 8 must persist long enough and be present
in a sufficient steady-state concentration to effect bimo-
lecular reactions.
Rea ction of P h en yld ia zom eth a n e w ith Ben zoic
Acid in Aceton e. Separate solutions of benzoic acid and
phenyldiazomethane in acetone-d6/acetone-h6 (9:1, vol/
vol) were allowed to equilibrate at -78 °C. The benzoic
acid solution was injected via a septum, with stirring,
into a round-bottom flask containing the solution of
phenyldiazomethane at -78 °C. No reaction occurred as
evidenced by the persistence of the red color of the
phenyldiazomethane and the absence of N2 evolution (i.e.,
effervescence). The sample was allowed to warm to -20
°C and was kept at that temperature until the reaction
ceased. The reaction products were analyzed immediately
1
after the reaction. The H NMR spectrum showed that
the reaction afforded benzyl benzoate (5; 56%), (2-
benzyloxy)isopropyl benzoate (9a ; 41%) (vide infra for
product identification), 2-benzyloxypropene (12; 1%), and
benzyl alcohol (10; 2%) (Table 1, Schemes 1 and 2). After
complete reaction, the product mixture was stored at 4
°C for 23 h; 1H NMR (acetone-d6) showed 58% of 5, 33%
of 9a , 5% of 12, and 4% of 10 (Table 1, Schemes 1 and
2). After a further 3 days at 4 °C, the solvent was removed
It may be that detection of 8 is possible at tempera-
tures well below -20 °C (vide infra), but this would be
contingent upon the successful acidification of phenyl-
1
in vacuo. The H NMR spectrum (in CDCl3) showed 60%

Interception of Benzyl Carbenium Ions by Acetone
J . Org. Chem., Vol. 65, No. 6, 2000 1827
Sch em e 2. P ossible Decom p osition P a th w a ys of
(2-Ben zyloxy)isop r op yl Ben zoa te
should be pointed out that the rate-limiting protonation
of the phenyldiazomethane would be succeeded by rapid
dediazoniation because of the facility of the loss of N2,
the inherent stability of the benzyl cation, and the
formation of the ion-pair in a polar solvent.
Id en t ifica t ion of In t er m ed ia t es a n d P r od u ct s.
Partial proof of the structure of compound 9a was
obtained by model studies involving preparation of the
analogous compound 9b by the reaction of phenyldiaz-
1
omethane with acetic acid in acetone at -20 °C. The H
NMR and IR spectra (vide infra) are consistent with the
assigned structure of 2-benzyloxyisopropyl acetate.4 Ad-
ditionally, when D2O was added to 9b, the latter hydro-
lyzed to the expected products: acetic acid, acetone, and
compounds 12 and 10.
Identification of compound 11 was based on the com-
1
parison of the H NMR signals in the product with that
of an authentic sample prepared via “trans-etherification”
of 2,2-dimethoxypropane with benzyl alcohol under reflux
in the presence of catalytic quantities of p-toluenesulfonic
acid.6a Compound 12 was identified in a similar fashion:
in this case, 12 was independently prepared by elimina-
tion of benzyl alcohol from 2,2-dibenzoxypropane with
P2O5 in the presence of quinoline at 120 °C.6b The
identities of 10 and 5 were deduced by comparison with
spectra of commercial samples and by spiking.
Th e Rea ction Mech a n ism . The formation of 9a is
believed to occur via trapping of the benzyl carbenium
ion of the NSIP by acetone to form the metastable onium
ion, 8 (Scheme 1). The latter then reacted with the
benzoate (or acetate for 9b ) counterion to produce 9
(Scheme 1).7a 2-Benzyloxypropene is believed to have
arisen from â-deprotonation of 8 by benzoate (Scheme
2),7b whereas 11 ostensibly formed from attack on 8 by
10 (Scheme 2).
diazomethane at significantly lower temperatures (pos-
sibly via triflic acid).1d The issue of the slow proton
transfer is not unusual for acid-base reactions involving
carbon bases (or carbon acids) and is well documented
in the acid decompositions of diazoalkanes that undergo
esterification with carboxylic acids.5a In these cases, the
protonation of the diazomethane is rate-limiting.5a,b In
the present study, the reaction between benzoic acid and
phenyldiazomethane would then be expected to be even
slower because of the lower basicity of phenyldiaz-
omethane (vs diazomethane)5c and the inherent low rates
of diffusion and reactivity at depressed temperatures. It
Some benzyl alcohol was probably formed by reaction
between trace moisture and the first-formed benzyl
cation. The low mole fraction of water in acetone,
however, makes this source of the alcohol, somewhat
poorly competitive with other routes (vide infra).1b How-
ever, during storage of the product mixture for 4 days at
4 °C, compound 9a evidently underwent slow decomposi-
tion to produce compounds 10, 11, and 12 (vide supra).
The implication therefore is that 9a must be in equilib-
rium with an ion pair containing an onium ion and
benzoate (4′) [although it is formed irreversibly from the
NSIP (4)].8 Hence, on standing, 9a generates a steady-
state concentration of 4′, the onium ion of which is
reversibly attacked by benzoate (to regenerate 9a ) and
reversibly attacked by water to yield a labile hemiacetal
that fragments to acetone and benzyl alcohol (Scheme
2). The onium ion can also be attacked by benzyl alcohol
to form 11 (Scheme 2). As with the first generation of
(4) (a) The structure assignment of compound 9a was based on
comparison of the chemical shifts observed in the products with the
expected (estimated) value for the methyl signal based on the following
considerations. From the chemical shifts of the methyl groups in ethyl
benzoate (δ 1.42)4b and in benzyl ethyl ether (δ 1.20)4c the ∆δ ) 0.22
presumably due to the differences in the inductive effects of the
benzyloxy and benzoate groups on the Me group. Assuming that the
effect is transferable then the Me group of 9 would be expected to be
at δ 1.76, i.e., 0.22 ppm higher than the Me group of 11 (known to be
δ 1.54). The Me peaks for 9 (δ 1.88 for 9a and δ 1.73 for 9b) are close
to the expected value along with a corresponding benzylic signal (δ
4.79 for 9a and δ 4.66 for 9b). On the basis of known values for 2,2-
dimethoxypropane [δ 3.20, C(CH3)2]4b and 1-methoxy-1-acetoxy ethane
[δ 3.30, C(CH3)2],4d an acetoxy group and a methyloxy group differ by
0.10 ppm in their effects on the chemical shift on Me. Assuming that
the acetate (similarly benzoate) and methyloxy (similarly benzyloxy)
groups would exert a similar difference in 9, a benzylic peak at δ 4.68
(the benzyloxy peak is at δ 4.58 in 11) was expected for the corre-
sponding signal in compound 9. This estimated chemical shift is also
close to the observed values (δ 4.79 for 9a and δ 4.66 for 9b). (b) The
Aldrich Library of NMR Spectra, 2nd ed.; Aldrich: Milwaukee, 1983;
Vol. 2, p 281. (c) Barleunga, J .; Alonso-Cires, L.; Campos, P. J .; Asensio,
G. Synthesis 1983, 53. (d) Pihlaja, K.; Lampi, A. Acta Chem. Scand.
1986, 40B, 196.
(6) (a) Lorette, N. B.; Howard, W. L. J . Org. Chem. 1960, 25, 521.
(b) Crocker, H. P.; Hall, R. H. J . Chem. Soc. 1954, 2052. (c) Borkovec,
A. B. J . Org. Chem. 1961, 26, 4866-4868. (d) Anderson, C.; Larson,
J .; Hallberg, A. J . Org. Chem. 1990, 55, 5757.
(7) (a) This reaction is analogous to the attack by pivalate on the
sp-hybridized carbon of the N-benzylnitrilium ion derived from
acetonitrile.1e (b) No deprotonation of the acetonitrilium ion by pivalate
to form the ketimine was observed.1e This mechanistic divergence is
probably related to the inherent differences in product stability
between the cummulated product and the monovinyl species.
(8) The onium ion in the ion-pair, 4′, is expected to be less reactive
than that produced initially during the encounter between the NSIP,
4, and acetone. Additionally, 4′ is almost surely a spectrum of intimate
and solvent-separated ion-pairs.
(5) (a) Isaacs, N. In Physical Organic Chemistry, 2nd ed.; J ohn Wiley
& Sons: New York, 1995; p 407. (b) More O’Ferrall, R. A. Adv. Phys.
Org. Chem. 1967, 5, 331. (c) We have been unable to find the pKb values
of diazoalkanes. However, resonance interaction involving the lone pair
of electrons on the benzylic group with both the adjacent diazonium
group and the aromatic ring would be expected to lower the basicity
of phenyldiazomethane below that of diazomethane. (d) Isaacs, N. In
Physical Organic Chemistry, 2nd ed.; J ohn Wiley & Sons: New York,
1995; p 239.

1828 J . Org. Chem., Vol. 65, No. 6, 2000
Song et al.
onium ions, these onium ions generated by ionization of
the N2O4-acetone reaction is not competitive at de-
pressed temperatures.
9a 8 are also apparently deprotonated to yield compound
12.
Nitr osation of N-Ben zyl-O-ben zoylh ydr oxylam in e
in Aceton e-h 6. N-Benzyl-O-benzoylhydroxylamine was
nitrosated in acetone at -70 °C, and the solvent was
removed in vacuo. The product consisted only of 5 (50%),
10 (40%), 11 (6%), and 4-hydroxy-4-methyl-2-pentanone
(4%; from acid-catalyzed aldol-type reaction of acetone).
When nitrosation was performed at 0 °C and the solvent
was removed as above, the product contained 5 and
diacetylfuroxan (13) (from the reaction of acetone with
N2O4)9 in a ratio of 6/5, respectively. Surprisingly,
compounds 10, 11, and 12 accounted for less than 10%
of the product mixture.
Su m m a r y a n d Con clu sion
Acetone intercepts deaminatively generated benzyl
cations to form an electrophilic onium ion that is com-
petitively scavenged by the benzoate counterion in an
equilibrium with (2-benzyloxy)isopropyl benzoate. Two
generations of onium ions are apparently formed: the
first originates as an ion-pair containing the benzoate ion
derived from capture of the deaminatively generated
carbocation; the second arises from simple solvolysis of
the labile (2-benzyloxy)isopropyl benzoate.
Proton transfer between benzoic acid and phenyldia-
zomethane is very slow below -20 °C presumably be-
cause of the feeble acidity and basicity of PhCO2H and
PhCHN2, respectively, as well as the low rates of diffu-
sions and reactions at depressed temperatures.
The formation of 5 from these runs is evidence of the
formation and decomposition of N-benzyl-N-nitroso-O-
benzoylhydroxylamine and the generation of the NSIP,
which on internal collapse produces the observed ester
(eq 2b).1b However, despite the intermediacy of the
deaminatively generated benzyl cation, only trace amounts
of the other intermediates and products observed in the
PhCHN2 + PhCO2H case were detected here. Thus,
although both deamination methods used in this study
successfully generate the benzyl cation, SDPs are not
readily accessible via hydroxylamine nitrosation (eq 2b).
It would appear then that in both cases a fraction of
the benzyl cations were intercepted by acetone to gener-
ate the undetectable electrophilic oxonium ion 8. In the
PhCHN2 + PhCO2H system, no acid stronger than
benzoic acid was present during the reaction. As a result,
the modestly nucleophilic benzoate ion (generated as the
conjugate base after protonation of the diazoalkane) was
available to enter into bimolecular reactions with 8 giving
the array of products observed in that case. However,
HNO3 is a side product of X-H (X ) C, heteroatom)
nitrosation via N2O4. Indeed, if moisture is present in the
system then both HNO2 and HNO3 are generated (facile
oxidation of the former by air, e.g., yields the stronger,
latter acid). Under these more strongly acidic conditions,
most of the benzoate ions presumably exist largely in the
protonated and much less nucleophilic form, i.e., as
benzoic acid and are thus less adept at intercepting the
onium ion.10 Consequently, the products 10, 11, and 12
observed in the PhCHN2 + PhCO2H system are gener-
ated here in only trace amounts because the nucleophile
from which they are formed was significantly deactivated.
When strong acids are present, the benzoate ion is
apparently converted to benzoic acid whose nucleophi-
licity is too low for competitive attack on the onium ion.
Consequently, limited yields of solvent-derived product
are observed from nitrosation of N-benzyl-O-benzoylhy-
droxylamine in acetone. The more acidic conditions
inherent in this approach, in tandem with adventitious
moisture serve to hydrolyze many of the labile benzyl
intermediates and products. Additionally, at tempera-
tures around 0 °C, large yields of the side product 13 are
observed, although this problem does not arise at de-
pressed temperatures. The hydroxylamine nitrosation
approach is therefore a poorer method of studying car-
benium ion interception in acetone.
Exp er im en ta l Section
Ma t er ia ls a n d Met h od s. All commercial reagents were
reagent grade and were used without further purification.
Spectra were recorded on 300 MHz FT-NMR, FT-IR, and UV-
vis spectrometers.
Sta bility of P h en yld ia zom eth a n e a n d N-Ben zyl-O-
ben zoylh yd r oxyla m in e: Ha n d lin g a n d Stor a ge. Phenyl-
diazomethane is labile and was prepared as needed. It
decomposes under acidic conditions and in the presence of
moisture and light. Ca u t ion ! Phenyldiazomethane should
be handled with extreme care because of its possible muta-
genicity11a and carcinogenicity (local and systemic).11b N-
Benzyl-O-benzoylhydroxylamine is comparatively stable, and
the dry solid can be stored indefinitely under nitrogen at -25
°C. Efficient fume hoods and appropriate personal protection
(chemical-resistant gloves, safety glasses, lab coat, etc.) are
recommended when handling these compounds.
Nitric acid’s role in the diminished yield of benzyl
derived products may extend beyond inhibition of their
formation through deactivation of benzoate. It may also
be involved in hydrolyses of labile benzyl species. Thus,
acid-catalyzed hydrolysis of labile 9a in particular, as well
as of other intermediates and products ostensibly re-
sulted in the significantly larger yield of benzyl alcohol
observed. Theoretically, only 1 equiv of water would be
sufficient to effect the observed hydrolyses. Typically,
∼10-4 moles of hydroxylamine in 2 mL of acetone were
used in these runs (vide infra); thus, only ∼1 mg of water
would be sufficient to produce the observed results. The
absence of the furoxan, 13, would appear to indicate that
2,2-Diben zyloxyp r op a n e (11). The title compound was
prepared by the general procedure of Lorette and Howard:6a
1H NMR (CDCl3) δ 7.38-7.25 (m, 10H), 4.58 (s, 4H), 1.54 (s,
6H); (lit.6c bp 125 °C at 1 Torr).
(9) Peterson, L. I. Tetrahedron Lett. 1966, 1727.
(10) (a) The pyridinium ion (pKa ) 5.210b) probably hydrogen bonds
with benzoate mitigating against the reaction of the latter with 8. (b)
Davies, D. T. In Aromatic Heterocyclic Chemistry, 1st ed.; Oxford
University Press: New York., 1992; p 35.
(11) (a) Lee, K.; Gold, B.; Mirvish, S. Mutat. Res. 1977, 48, 131. (b)
Preussman, R.; Stewart, B. W. In Chemical Carcinogenesis; Searle,
C., Ed.; ACS Monograph No. 182; American Chemical Society: Wash-
ington, DC, 1984; pp 643-828.

Interception of Benzyl Carbenium Ions by Acetone
J . Org. Chem., Vol. 65, No. 6, 2000 1829
2-Ben zyloxyp r op en e (12). Following the general proce-
dure of Crocker and Hall,6b 2,2-dibenzyloxypropane (60 mg,
0.25 mmol) was added to a slurry of phosphorus pentoxide (36
mg, 0.25 mmol) and quinoline (32 mg, 0.25 mmol) in a flask
connected to a U-tube. The mixture was heated to 120 °C at
0.1 Torr. A colorless liquid (0.10 mmol, ∼15 mg) was collected
in the U-tube (cooled to - 78 °C). The distillate consisted of
compound 12 [1H NMR (CDCl3) δ 4.73 (s, 2H), 3.94 (m, 2H),
1.88 (m, 3H)] and an equimolar amount of benzyl alcohol [1H
NMR (CDCl3) δ 4.70 (s, 2H)]; the aromatic signals were
obscured in an unresolved multiplet at δ 8.90-7.17 (m). The
1H NMR spectrum (CDCl3) of an analogue, benzyl vinyl ether,
has been reported:5d δ 7.4-7.3 (m), 6.60 (dd, J ) 14, 7 Hz),
4.80 (s), 4.35 (dd, J ) 14, 2 Hz), 4.12 (dd, J ) 7, 2 Hz).
Ben za ld eh yd e t osylh yd r a zon e was prepared as de-
scribed by Closs and Moss.12
P h en yld ia zom eth a n e (1). Meth od A.12 To 10 mL of
ethanol was added sodium (100 mg, 2.3 mmol). After the
reaction was over, benzylaldehyde tosylhydrazone (549 mg, 2
mmol) was added at 25 °C. After 1 h of reaction at 60 °C, the
solution became red in color. The resulting reaction mixture
was treated with 40 mL of ice-water and then extracted with
pentane (10 mL). The organic phase was separated, dried with
Na2SO4, and concentrated in vacuo (20 Torr) at 0 °C. When
the residue was distilled at 40 °C under vacuum (0.1 Torr), a
red oil (100-200 mg) was collected in a U tube cooled to -70
°C: 1H NMR (CDCl3) δ 7.30-6.90 (m, 5H), 4.93 (s, 1H) (the
method afforded 5-10% trans-stilbene as a impurity with δ
6.60). Meth od B. To 2.5 mL of methanol was added sodium
(52 mg, 2.5 mmol) previously rinsed with pentane. After
cessation of hydrogen evolution, benzylaldehyde tosylhydra-
zone (342 mg) was added. After dissolution, the solvent was
evaporated at 0.1 Torr to give a solid that was heated at 90-
100 °C with an oil bath at 0.1 Torr. A red oil (∼100 mg) was
collected at -70 °C in a U-tube. The red oil was distilled at 0
°C and 0.1 Torr to give phenyldiazomethane: 1H NMR (CDCl3)
δ 7.28 (t, 2H, J ) 8.1 Hz), 7.03 (t, 1H, J ) 8.1 Hz), 6.92 (d, 2H,
J ) 8.4 Hz), 4.93 (s, 1H). No trans-stilbene was observed by
this procedure.1b
(2-Ben zyloxyl)isop r op yl Aceta te. Acetic acid (5.7 µL, 6.0
mg, 0.10 mmol) was injected into a solution of phenyldiaz-
omethane (∼12 mg, 10 mmol) in 0.5 mL of acetone in a NMR
tube with a septum cooled to -70 °C. The mixture was shaken
immediately and then kept at -20 °C in a freezer. After ∼3 h,
the solvent and volatile products were distilled at 0.1 Torr and
25 °C over ∼25 min to give a colorless oil: 1H NMR (CDCl3)
7.36 (m, 5H), 4.66 (s, 2H), 2.03 (s, 3H), 1.73(s, 6H); IR (CDCl3)
1725, 1378 and 1360 (d), 1219, 1125 cm-1. The NMR spectrum
of the distillate showed that (2-benzyloxyl)isopropyl acetate
was also partially distilled.
Decom p osition of (2-Ben zyloxy)isop r op yl Aceta te in
th e P r esen ce of D2O. D2O (10 µL, 10 mg, 0.55 mmol) was
added to (2-benzyloxy)isopropyl acetate; after ∼4 h, the 1H
NMR spectrum showed ∼69% of decomposition to give 27% of
benzyl alcohol, 24% of 2-benzyloxylpropene, 19% of acetone,
35% of acetic acid.
10 min, no reaction was observed until at -20 °C. The 1H NMR
spectrum showed two new major peaks at δ 4.83 (s) (2-
benzyloxy)isopropyl benzoate and δ 5.38 (s) (benzyl benzoate);
50% of the phenyldiazomethane still remained. After 23 h at
4 °C, the 1H NMR spectrum showed δ 5.38 (s) (benzyl
benzoate), 4.83 (s) (2-benzyloxy)isopropyl benzoate, 4.73 (s),
4.67 (s) in a ratio of 49:27:4:3. After 3 days at 4 °C, this sample
was concentrated in vacuo (20 Torr); the 1H NMR (CDCl3)
spectrum showed that the reaction afforded 5 [60% (δ 5.37)];
9a [16%, δ 4.79 (s) and 1.88 (s)]; 12 [13%, δ 4.73 (s, 2H), 3.95
(m, 2H), 1.89 (m, 3H)]; 11 [5%, δ 4.58 (s, 4H), 1.54 (s, 6H)]; 10
(6%, δ 4.70).
In another run, phenyldiazomethane (∼10 mg, 0.082 mmol)
was added to a solution of benzoic acid (11 mg, 0.090 mmol)
in 1 mL of acetone at -80 °C. No reaction occurred within 30
min at -80 °C based on the persistence of the red color. More
benzoic acid (6.0 mg, 0.049 mmol) was added; there was still
no reaction. The sample was allowed to warm to 25 °C
gradually, and the red color disappeared. The mixture was
concentrated in vacuo: 1H NMR (CDCl3) showed 5 (63%); 9a
(4%); 12 (6%); 11 (9%); 10 (17%).
At -20 oC in Aceton e-d 6. In a typical run, a solution of
benzoic acid (10.2 mg, 8.4 mmol) in 50 µL of acetone-d6 was
injected into a solution of phenyldiazomethane (∼12 mg, 10
mmol) in 0.5 mL of acetone-d6 in an NMR tube with a septum
cooled to -70 °C. The mixture was shaken immediately and
then kept at -20 °C in a freezer. After completion of the
reaction in ∼3 h, the reaction products were analyzed by NMR.
The 1H NMR (acetone-d6) spectrum showed benzyl benzoate
(5, 56%); (2-benzyloxy)isopropyl benzoate (9, 41%); 2-benzy-
loxypropene (12, 1%); and benzyl alcohol (10, 2%).
Syn th esis a n d Decom p osition of N-Nitr oso-N-ben zyl-
O-ben zoylh yd r oxyla m in e (6) in Aceton e-h 6; a t 0 °C. To a
stirred solution of N-benzyl-O-benzoylhydroxylamine (20 mg,
20 µL, 0.088 mmol) and pyridine (9.8 mg, 0.12 mmol) in 2 mL
of acetone at 0 °C was injected gaseous dinitrogen tetroxide
(2.0 mL, 0.089 mmol) with a hypodermic syringe through a
septum. After being stirred for ∼5 min, the resulting mixture
was concentrated in vacuo (20 Torr). The 1H NMR (CDCl3)
spectrum of the residue showed that the reaction afforded 13
[δ 2.63 (s, 3H), 2.72 (s, 3H)]9 and 5 [δ 4.70 (s, 2H)] in a ratio
of 6/5, respectively, plus trace quantities amount of an
unknown at δ 5.70. At -70 °C. To a stirred solution of
N-benzyl-O-benzoylhydroxylamine (20 mg, 20 µL, 0.088 mmol)
and pyridine (9.8 mg, 0.12 mmol) in 2 mL of acetone at -70
°C was injected gaseous dinitrogen tetroxide (2.0 mL, 0.089
mmol) with a hypodermic syringe through a septum. After
being stirred for ∼5 min, the resulting mixture was concen-
trated in vacuo (20 Torr). The 1H NMR (CDCl3) spectrum of
the residue showed that the reaction afforded 5 (50%), 11 (6%),
10 (40%), and 4-hydroxy-4-methyl-2-pentanone [δ 2.62 (s, 2H),
2.17 (s, 3H), 1.25 (s, 6H), 2.5 (bs, 1H) 6%].
Dia cetylfu r oxa n (13) fr om th e Rea ction of Aceton e
w ith N2O4 a t 0 °C. Liquid N2O4 (1.5 mL, 3.93 g, 0.042 mol)
was injected into acetone (1.2 g, 0.021mol) in a flask with
stirring at 0 °C. After 10 min, the resulting mixture was
concentrated in vacuo (20 Torr) to give 1.2 g (0.066 mmol, 31%)
of diacetylfuroxan (13) in the form of a yellow oil: 1H NMR
(CDCl3) δ 2.72 (s, 3H), 2.63 (s, 3H) [lit.9 δ 2.63 (s, 3H), 2.72 (s,
3H)].
Th e Rea ction of P h en yld ia zom eth a n e w ith Ben zoic
Acid . In Aceton e-d 6. The 1H NMR spectrum of phenyldiaz-
omethane (2.5 mg, 0.02 mmol) in 0.25 mL of acetone-h6/
acetone-d6 (1/9, v/v) exhibited a CH peak at δ 5.44 (s, 1 H) at
25 °C and 5.67 (s, 1H) at -78 °C. When a solution of benzoic
acid (2.5 mg, 0.02 mmol) in 0.25 mL of acetone-h6/acetone-d6
(1/9, v/v) at -78 °C was added to this solution at -78 °C, the
1H NMR spectrum was unchanged after 20 min. When the
reaction mixture was warmed in the NMR probe in 10°
increments (over 10 min per increment) and held for a further
Ack n ow led gm en t is made to the donors of the
Petroleum Research Fund, administered by the Ameri-
can Chemical Society, and to the D. Mead J ohnson
Foundation for support of this research. The authors
also wish to thank Dr. Brown L. Murr and Dr. Alex
Nickon for their insight and advice.
(12) Closs, G. L.; Moss, R. A. J . Am. Chem. Soc. 1964, 86, 404.
J O991827I