Anion structure determination in the gas phase: Chemical reactivity as a probe

Article abstract of DOI:10.1021/jo961463j

In the gas phase, the discrimination between two isomeric anion structures is a challenge that requires different solutions for different applications. The anionic oxy-Cope rearrangement involves the rearrangement of an alkoxide to an isomeric enolate; the mechanistic study of such a process in the gas phase requires a simple and selective probe process. Using a flowing afterglow mass spectrometer, we have examined the utility and limitations of using chemical reactivity to discriminate between alkoxides and enolates in the gas phase. A series of alkoxides and enolates were allowed to react with three chemical probe reagents: methanol-O-d, methyl nitrite, and dimethyl disulfide. Quantitative and qualitative characterization of each probe reagent reveals the especially broad and flexible utility of dimethyl disulfide as a chemical probe. Dimethyl disulfide is a selective reagent with ambident behavior that reacts efficiently with all anions studied and fully capitalizes on the structure/reactivity differences between alkoxides and enolates. Alkoxides behave as classical "hard bases" when allowed to react with dimethyl disulfide, effecting elimination across the C-S bond, whereas enolates, "soft bases", attack at sulfur. Methyl nitrite is also a selective ambident probe reagent but, due to its particularly slow reaction with enolates, is useful only in conjunction with a more reliable probe such as dimethyl disulfide. Methanol-O-d, for a variety of reasons detailed in the paper, is unsuitable as a chemical probe reagent for the unequivocal discernment between alkoxides and enolates.

Full text of DOI:10.1021/jo961463j

9422
J . Org. Chem. 1996, 61, 9422-9429
An ion Str u ctu r e Deter m in a tion in th e Ga s P h a se: Ch em ica l
Rea ctivity a s a P r obe
J eehiun K. Lee† and J oseph J . Grabowski*,‡
Departments of Chemistry, Harvard University, Cambridge, Massachusetts 02138, and
University of Pittsburgh, Pittsburgh, Pennsylvania 15260
Received J uly 31, 1996X
In the gas phase, the discrimination between two isomeric anion structures is a challenge that
requires different solutions for different applications. The anionic oxy-Cope rearrangement involves
the rearrangement of an alkoxide to an isomeric enolate; the mechanistic study of such a process
in the gas phase requires a simple and selective probe process. Using a flowing afterglow mass
spectrometer, we have examined the utility and limitations of using chemical reactivity to
discriminate between alkoxides and enolates in the gas phase. A series of alkoxides and enolates
were allowed to react with three chemical probe reagents: methanol-O-d, methyl nitrite, and
dimethyl disulfide. Quantitative and qualitative characterization of each probe reagent reveals
the especially broad and flexible utility of dimethyl disulfide as a chemical probe. Dimethyl disulfide
is a selective reagent with ambident behavior that reacts efficiently with all anions studied and
fully capitalizes on the structure/reactivity differences between alkoxides and enolates. Alkoxides
behave as classical “hard bases” when allowed to react with dimethyl disulfide, effecting elimination
across the C-S bond, whereas enolates, “soft bases”, attack at sulfur. Methyl nitrite is also a
selective ambident probe reagent but, due to its particularly slow reaction with enolates, is useful
only in conjunction with a more reliable probe such as dimethyl disulfide. Methanol-O-d, for a
variety of reasons detailed in the paper, is unsuitable as a chemical probe reagent for the unequivocal
discernment between alkoxides and enolates.
Structure determination is the foundation upon which
mass spectrometric techniques. CID can be very infor-
mative,7 but is an energetic method which can actually
induce an ion to isomerize. Since many methods require
instrument modification, a general method of structure
determination applicable across the wide range of mass
spectrometric approaches used in organic chemistry
would ultimately be the most useful to the broadest range
of users. Such a general method is the use of chemical
reactivity to elucidate structure.8-11
Our interest in structure determination stems from our
examination of unimolecular rearrangements, and the
anionic oxy-Cope in particular.12,13 In this reaction, an
alkoxide undergoes a [3,3]-sigmatropic rearrangement to
an enolate (eq 1). In the gas phase, the study of this
advances in chemistry are made. Various structural
determination methods allow one to declare unambigu-
ously the successful synthesis of a target compound,1-3
to elucidate aspects of reaction mechanisms via product
identification,4 and to monitor changes in secondary and
tertiary protein structure.5 Many methods of elucidating
structure are well-known and used regularly, such as
NMR and IR spectroscopy for condensed-phase and solid-
phase samples. In the gas phase, methods of structural
determination are more limited; examples include various
spectroscopic techniques (PES, IR, UV-vis) and methods
using mass spectrometry.
An especially challenging aspect of gas-phase structure
determination is the structural assignment of reactive
intermediates, particularly organic anions. PES can be
experimentally difficult to execute,6 while traditional
simple mass spectrometric methods fail since ionized
structural isomers themselves possess the same mass-
to-charge ratio. One useful solution to such a dilemma
for ion structure is collision-induced dissociation (CID)
(1)
rearrangement raises a classic problem of structure
elucidation: both the starting material (alkoxide) and the
product (enolate) have the same mass-to-charge signa-
ture. We sought to use chemical reactivity as a general
method to discern between these two types of species.
Although the use of chemical probe reagents to discern
structure is not new, we sought to fully generalize a
† Harvard University.
‡ University of Pittsburgh.
X Abstract published in Advance ACS Abstracts, December 1, 1996.
(1) Holton, R. A.; Somoza, C.; Kim, H.-B.; Liang, F.; Biediger, R. J .;
Boatman, P. D.; Shindo, M.; Smith, C. C.; Kim, S.; Nadizadeh, H.;
Suzuki, Y.; Tao, C.; Vu, P.; Tang, S.; Zhang, P.; Murthi, K. K.; Gentile,
L. N.; Liu, J . H. J . Am. Chem. Soc. 1994, 116, 1597-1598.
(2) Holton, R. A.; Kim, H.-B.; Somoza, C.; Liang, F.; Biediger, R. J .;
Boatman, P. D.; Shindo, M.; Smith, C. C.; Kim, S.; Nadizadeh, H.;
Suzuki, Y.; Tao, C.; Vu, P.; Tang, S.; Zhang, P.; Murthi, K. K.; Gentile,
L. N.; Liu, J . H. J . Am. Chem. Soc. 1994, 116, 1599-1600.
(3) Nicolaou, K. C.; Yang, Z.; Liu, J . J .; Ueno, H.; Nantermet, P. G.;
Guy, R. K.; Claiborne, C. F.; Renaud, J .; Couladouros, E. A.; Paul-
vannan, K.; Sorensen, E. J . Nature 1994, 367, 630-634.
(4) Grabowski, J . J .; Zhang, L. J . Am. Chem. Soc. 1989, 111, 1193-
1203.
(7) Groenewold, G. S.; Gross, M. L. In Ionic Processes in the Gas
Phase; Almoster Ferreira, M. A., Ed.; D. Reidel Publishing Co.: Boston,
1982; pp 243-266.
(8) Andrist, A. H.; DePuy, C. H.; Squires, R. R. J . Am. Chem. Soc.
1984, 106, 845-850.
(9) Rozeboom, M. D.; Kiplinger, J . P.; Bartmess, J . E. J . Am. Chem.
Soc. 1984, 106, 1025-1029.
(10) Thomas, D. A.; Bloor, J . E.; Bartmess, J . E. J . Am. Soc. Mass
Spectrom. 1990, 1, 295-300.
(11) Chou, P. K.; Dahlke, G. D.; Kass, S. R. J . Am. Chem. Soc. 1993,
115, 315-324.
(5) Raleigh, D. P.; DeGrado, W. F. J . Am. Chem. Soc. 1992, 114,
10079-10081.
(6) Zittel, P. F.; Ellison, G. B.; ONeil, S. V.; Herbst, E.; Lineberger,
W. C.; Reinhardt, W. P. J . Am. Chem. Soc. 1976, 98, 3731-3732.
(12) Evans, D. A.; Golob, A. M. J . Am. Chem. Soc. 1975, 97, 4765-
4766.
(13) Lee, J . K.; Grabowski, J . J . Manuscript in preparation.
S0022-3263(96)01463-6 CCC: $12.00 © 1996 American Chemical Society

Anion Structure Determination in the Gas Phase
J . Org. Chem., Vol. 61, No. 26, 1996 9423
Ta ble 1. Resu lts for th e Rea ction s of An ion s w ith Meth a n ol-O-d in th e Ga s P h a se a t 298 Ka
∆H°acid (AH)
number of
protons exchanged
a
anion (A-)
(kcal mol-1
376.0
)
other products
kobs ( 1σb-d
4.8 ( 0.4
Effe
O–
0
O–
0.24
(CH3OD)n
O–
375.4
365.3
0
0
O–
4.1 ( 0.4
0.20
(CH3OD)n
O–
O–
0.30 ( 0.05
0.015
(CH3OD)n
O–
369.1
5
O–
(CH3OD)n
0.82 ( 0.07
0.041
dn
a
b
Products reported can be the result of several sequential anion-methanol collisions (see text and schemes). Gas-phase acidities
were obtained from ref 28 unless otherwise noted. c The minimum number of independent experimental determinations of the rate coefficient
(kobs) is 5. Reaction rate in units of 10-10 cm3 molecule-1 s-1
.
e This is the apparent bimolecular rate coefficient, taken at 0.3 Torr of He;
d
see text. f Reaction efficiency, kobs/kcoll, where kcoll is calculated according to ref 24.
methodology by selectively discriminating between alkox-
ides and enolates with the use of one ambident reagent.8-11
The importance of such well-characterized methodology
is made clear by identification of gas-phase anionic oxy-
Cope rearrangements which were previously believed to
not occur.9,13
NH3
-8 H2N- + H•
(2)
e
9
N2O
9
-8 O•- + N2
e
(3)
(4)
•
O
•- + CH4 f HO- + CH3
Toward the goal of characterizing probe reagents for
general use as a selective reagent for alkoxide/enolate
H2N- + (CH3)2CHOCH(CH3)2
f
discrimination, we examined a series of prototypical
alkoxides and enolates with three promising probe
reagents. The reactions of model anions propoxide,
isopropoxide, propanal enolate, and acetone enolate were
each examined with methanol-O-d (MeOD), methyl ni-
trite (MeONO), and dimethyl disulfide (MeSSMe). MeOD
was chosen because of its capability to exchange its
deuteron; MeONO and MeSSMe had shown promise in
earlier studies for displaying different reactivity with
different anions.4,14-16 Each probe reagent was studied
both quantitatively and qualitatively to assess its reli-
ability as a structural determinant, especially with
respect to its utility under the thermally-equilibrated
conditions as are routinely attained in the flowing
afterglow.
(CH3)2CHO- + CH3CHdCH2 + NH3 (5)
Methyl nitrite was prepared within 24 h of use by adding a
solution of dilute sulfuric acid and methanol to a cooled
aqueous solution of sodium nitrite and methanol.21,22 The
gaseous product was collected in a standard glass sample flask
which was cooled in an acetone/dry ice bath. The flask was
wrapped in aluminum foil to prevent photoinduced decomposi-
tion of the methyl nitrite. All other reagents were obtained
from standard commercial suppliers and were used as received.
Before use, each liquid sample was subjected to several freeze-
pump-thaw cycles to remove dissolved gases.
Each reaction of interest was examined rigorously using
both qualitative and quantitative methods. Qualitative ex-
periments were run to identify all reaction products and
involved examining the products of the reaction at progressive
reaction times. Practically, this process involves taking a
complete mass scan of the ionic contents of the flow tube at
different time points in the reaction. Quantitative examina-
tion involved kinetic experiments to measure total reaction
rate coefficients and branching ratio experiments to determine
the yields of the various product ions. Rate coefficients were
obtained by monitoring the disappearance of the reactant ion
“A-” as a function of time (distance) under pseudo-first-order
conditions (i.e., a constant concentration, or flow of the neutral
reactant “B” was maintained (eqs 6-8)), and all reported
Exp er im en ta l Section
Experiments were conducted at ambient temperature (298
K) and in 0.3 Torr of helium buffer gas in a flowing after-
glow17,18 which has been previously described.19 Amide was
generated by dissociative electron attachment to ammonia (eq
2). Hydroxide was generated by allowing the atomic oxygen
radical anion (generated by dissociative electron attachment
to N2O, eq 3) to abstract a hydrogen atom from methane (eq
4). Acetone enolate and propanal enolate were synthesized
in the flow tube by exothermic proton transfer from acetone
or propanal to amide or hydroxide. Propoxide and isopropoxide
were generated by an elimination reaction of either amide or
hydroxide with isopropyl ether (eq 5) or propyl ether.20
A
- + B
9
k8 C- + D
(6)
(7)
rate of reaction ) k[A-][B]
pseudo-first-order conditions ([B] . [A-]):
(14) DePuy, C. H.; Bierbaum, V. M.; King, G. K.; Shapiro, R. H. J .
Am. Chem. Soc. 1978, 100, 2921-2922.
rate of reaction ) k′[A-] (8)
(15) King, G. K.; Maricq, M. M.; Bierbaum, V. M.; DePuy, C. H. J .
Am. Chem. Soc. 1981, 103, 7133-7140.
results are the average of data collected over several indepen-
dent experimental days.23 We estimate the absolute accuracy
of our rate coefficients to be (20% (the precision is as indicated
in Tables 1-3); collision rate coefficients (kcoll) were calculated
according to the variational transition state theory of Su and
(16) Noest, A. J .; Nibbering, N. M. M. Adv. Mass. Spectrom. 1980,
8A, 227-237.
(17) Ferguson, E. E.; Fehsenfeld, F. C.; Schmeltekopf, A. L. In
Advances in Atomic and Molecular Physics; Bates, D. R., Estermann,
I., Eds.; Academic Press: New York, 1969; Vol. 5, pp 1-56.
(18) Graul, S. T.; Squires, R. R. Mass Spectrom. Rev. 1988, 7, 263-
358.
(19) Grabowski, J . J .; Melly, S. J . Int. J . Mass Spectrom. Ion
Processes 1987, 81, 147-164.
(20) DePuy, C. H.; Bierbaum, V. M. J . Am. Chem. Soc. 1981, 103,
5034-5038.
(21) Hartung, W. H.; Crossley, F. Org. Synth. 1943, 2, 108.
(22) Semon, W. L.; Damerell, V. R. Org. Synth. 1930, 10, 22.
(23) Guo, Y.; Grabowski, J . J . J . Am. Chem. Soc. 1991, 113, 5923-
5931.

9424 J . Org. Chem., Vol. 61, No. 26, 1996
Lee and Grabowski
Ta ble 2. Resu lts for th e Rea ction s of An ion s w ith Meth yl Nitr ite in th e Ga s P h a se a t 298 K
anion
(A-)
∆H°acid(AH)
product
ion (m/z)
observed
product
a
(kcal mol-1
376.0
)
mol formula
amountsb-d
yieldse (%)
kobs ( 1σc,f
Effg
0.11
-
O–
NO2
46
60
90
major
major
minor
2.3 ( 0.2
ONCH2O-
NO-(C3H7OH)
-
O–
375.4
365.3
369.1
NO2
46
60
90
major
major
minor
6.9 ( 0.1
0.32
ONCH2O-
NO-(C3H7OH)
-
O–
NO2
46
58
86
0.073 ( 0.02
0.531 ( 0.01
0.396 ( 0.01
7.1
52.9
40.0
0.41 ( 0.02
0.26 ( 0.01
0.019
0.012
ONChHCH3
CH3C(NO)CHO-
-
O–
ONCH2
44
46
86
0.262 ( 0.002
0.071 ( 0.001
0.667 ( 0.001
25.8
6.9
67.3
-
NO2
CH3C(O-)dCHNO
a
b
Gas-phase acidities were obtained from ref 28 unless otherwise noted. The average fractional yield, ( 1 standard deviation, based
on observation of only the m/ z values listed for that reaction. c The minimum number of independent experimental determinations of the
branching ratio is 2, and of the rate coefficient (kobs), 5. Quantitative branching ratios were not measured for the alkoxides. We suspect
d
(see ref 27) that NO- is also a product for these reactions; however, NO- will be collisionally detached before it can be detected in our
instrument. e The yield of the product ion of given molecular formula actually formed in the reaction cited (i.e., corrected to include all
naturally occurring isotopic variants). These values are calculated by taking the “observed amounts” and dividing them by the “isotope
factor” and then renormalizing. The isotope factor is the fraction of the product ion with the given molecular formula that will be detected
at the m/ z value listed (which is the predominant isotopic peak). These values were calculated from values in the Table of the Isotopes
in ref 27. The isotope factors for m/ z 44, 58, 46, and 86 are 0.9827, 0.9716, 0.9915, and 0.9586, respectively. f Reaction rate coefficient
g
in units of 10-10 cm3 molecule-1 s-1
.
Reaction efficiency, kobs/kcoll, where kcoll is calculated according to ref 24.
Ta ble 3. Resu lts for th e Rea ction s of An ion s w ith Dim eth yl Disu lfid e in th e Ga s P h a se a t 298 K
anion
(A-)
∆H°acid(AH)
product
ion (m/z)
observed
product
kobs ( 1σc,f
a
(kcal mol-1
376.0
)
mol formula
amountsb,c
yieldsd (%)
(cm3 molecule-1 s-1
)
Efff
1g
O–
CH3S-
47
93
107
0.051 ( 0.01
0.673 ( 0.03
0.276 ( 0.02
4.8
68.0
27.2
23 ( 2
CH3SCH2S-
CH3S-(C3H7OH)
O–
375.4
CH3S-
47
93
107
0.016 ( 0.002
0.735 ( 0.001
0.249 ( 0.001
1.5
74.0
24.5
22 ( 1
1g
CH3SCH2S-
CH3S-(C3H7OH)
O–
365.3
369.1
CH3S-
47
103
0.067 ( 0.007
0.933 ( 0.007
6.5
93.5
6.8 ( 0.1
10 ( 1
0.33
0.49
CH3C(SCH3)CHO-
O–
CH3S-
47
103
0.261 ( 0.02
0.739 ( 0.02
25.4
74.6
CH3C(O-)CHSCH3
a
b
Gas-phase acidities were obtained from ref 28 unless otherwise noted. The average fractional yield, ( 1 standard deviation, based
on observation of only the m/ z values listed for that reaction. c The minimum number of independent experimental determinations of the
branching ratio is 2, and of the rate coefficient (kobs), 5. The yield of the product ion of given molecular formula actually formed in the
d
reaction cited (i.e., corrected to include all naturally occurring isotopic variants). These values are calculated by taking the “observed
amounts” and dividing them by the “isotope factor” and then renormalizing. The isotope factor is the fraction of the product ion with the
given molecular formula that will be detected at the m/ z value listed (which is the predominant isotopic peak). These values were
calculated from values in the Table of the Isotopes in ref 27. The isotope factors for m/ z 44, 58, 46, and 86 are 0.9827, 0.9716, 0.9915,
and 0.9586, respectively. e Reaction rate coefficient in units of 10-10 cm3 molecule-1 s-1
according to ref 24. The numbers calculated for the efficiency of these two reactions are 1.04 for the propoxide and 1.09 for isopropoxide.
.
f Reaction efficiency, kobs/kcoll, where kcoll is calculated
g
That these ratios are calculated to be slightly greater than 1 is a reflection of the imperfections in modeling kcoll, the error in the data
which is used to calculate kcoll (dipole moments and polarizabilities), and the error in kobs
.
Chesnavich.24 Reaction efficiency, the probability of chemical
are capable of undergoing proton/deuteron exchange with
MeOD.14 The two alkoxides are found to undergo a facile
clustering reaction with MeOD to form varied-size clus-
ters of RO-(MeOD)n. The lower basicity of the alkoxides
(∼375 kcal mol-1) relative to the acidity of MeOD (381.7
kcal mol-1) renders proton transfer impossible and the
lack of “acidic” protons renders exchange improbable,
such that clustering is the only reactivity the alkoxides
display with MeOD.26 Clustering is a process that is not
uncommon in the flowing afterglow, where the relatively
high pressure of He creates thermally equilibrated condi-
tions which can stabilize ion-molecule clusters. In
measuring the kobs of a clustering reaction, the k mea-
sured (and listed in Table 1) is only an apparent bimo-
lecular rate constant, since clustering is a termolecular
reaction per collision, is determined for a given reaction as
the ratio of the observed rate coefficient to the collision rate
coefficient (Eff ) kobs/kcoll). Branching ratios were obtained by
monitoring the reactant and product ion intensities of a given
reaction at a fixed concentration of neutral reagent, under
pseudo-first-order conditions, as a function of the time (dis-
tance).25 The precision of our branching ratios is as indicated
in Tables 2 and 3; we estimate that the absolute accuracy of
our branching ratio experiment is ( 2% or less.
Resu lts
Meth a n ol-O-d . The results of the reactions of pro-
poxide, isopropoxide, propanal enolate, and acetone eno-
late with methanol-O-d are summarized in Table 1. In
the gas phase, like in the solution phase, certain anions
(24) Su, T.; Chesnavich, W. J . J . Chem. Phys. 1982, 76, 5183-5185.
(25) Grabowski, J . J .; Lum, R. C. J . Am. Chem. Soc. 1990, 112, 607-
620.
(26) DePuy, C. H.; Bierbaum, V. M. Acc. Chem. Res. 1981, 14, 146-
153.

Anion Structure Determination in the Gas Phase
J . Org. Chem., Vol. 61, No. 26, 1996 9425
F igu r e 1. Series of mass spectral scans of the reaction of acetone enolate (m/ z 57) with methanol-O-d (molecular weight ) 33).
Acetone enolate exchanges its protons for deuterons (m/ z 57 disappears, m/ z 58 and 59 grow in); the cluster of acetone enolate
to MeOD (m/ z 90) is also observed, as well as deuterated versions of this cluster.
process dependent on the amount of He present (eqs
9-11; He* represents energetic He).
scribed in footnote e.27 The alkoxides are found to react
fairly efficiently with methyl nitrite to produce mostly
NO2- (m/ z 44) and ONCH2O- (m/ z 60). A minor product
is the cluster of NO-(C3H7OH) (m/ z 90). Quantitative
branching ratios were not measured for these reactions
as the major products are the same from the two model
alkoxides and are independent of alkoxide; we speculate
that NO- is also a product which is formed, but which
undergoes collision-induced detachment under our condi-
tions (EA(NO• ) 0.026 eV)).15,28 In contrast to the
alkoxides, the enolates react with low efficiency, 2% or
less, and are a factor of 5-30 times slower than the
alkoxide reaction. Acetone enolate reacts with methyl
nitrite to form the nitrosated enolate (CH3C(O-)CHNO,
m/ z 86) as the major product ion (67%). Also formed are
(9)
(10)
(11)
Acetone enolate, however, behaves in the fashion that
makes MeOD’s potential as a gas-phase probe reagent
clear. Although clustering is a also a facile process for
acetone enolate, this ion also exchanges five protons for
deuterons (Figure 1). Note that both deuterated as well
as undeuterated forms of acetone enolate are observed
to cluster with MeOD.
-
-
ONCH2 (m/ z 44, 26%) and a minor amount of NO2
(m/ z 46, 7%). Propanal enolate is also nitrosated by
methyl nitrite (CH3C(NO)CHO-, m/ z 86, 40%), but the
major product ion formed is ONChHCH3 (m/ z 58, 53%).
Propanal enolate is about 4 kcal mol-1 less basic than
acetone enolate and shows, unexpectedly, markedly
different patterns of behavior than its sister enolate.
Unlike the alkoxides, propanal enolate does possess an
exchangeable proton; however, unlike acetone enolate,
it does not exchange it for a deuteron with MeOD in a
0.3 Torr helium environment. Rather, facile clustering
is observed to form CH3CHCHO-(MeOD)n. Another
cluster formed when propanal enolate is generated is the
complex of propanal enolate with neutral propanal,
CH3CHCHO-(CH3CH2CHO). Although acetone enolate
clusters to acetone as well, propanal enolate clusters to
propanal much more efficiently; experimentalists should
beware in that this facile clustering of aldehyde enolates
can complicate analysis. An intriguing aside is that when
exposed to MeOD, the CH3CHCHO-(CH3CH2CHO) clus-
ter readily exchanges three protons for deuterons (a
simple cluster would contain three exchangeable protons
in contrast to none for the covalent adduct).
-
A small amount of NO2 (7%) is also formed in the
propanal enolate reaction.
Dim eth yl Disu lfid e. The results of the reactions of
propoxide, isopropoxide, propanal enolate, and acetone
enolate with dimethyl disulfide are summarized in Table
3. The alkoxides react with unit efficiency with dimethyl
disulfide to yield CH3SCH2S- (m/ z 93) as the major
product. The cluster of methanethiolate to the parent
alcohol, CH3S-(C3H7OH) (m/ z 107), is also produced in
relatively high yield, as well as a small amount of
methanethiolate, CH3S- (m/ z 47). The enolates (“HA-”)
react quite efficiently (Eff ) 0.33 and 0.49) with dimethyl
disulfide, forming the thiomethoxylated enolates CH3SA-
as the major product ion (for acetone enolate, CH3C(O-)-
CHSCH3; for propanal enolate, CH3C(SCH3)CHO-; m/ z
103). Some methanethiolate (CH3S-) is also formed, in
Meth yl Nitr ite. The results of the reactions of the
model ions with methyl nitrite are collected in Table 2.
The product yields (column 6) are calculated from the
“observed amounts” and the “isotope factors”, as de-
(27) CRC Handbook of Chemistry and Physics, 73rd ed.; Lide, D.
R., Ed.; CRC: Boca Raton, FL, 1992.
(28) Lias, S. G.; Liebman, J . F.; Levin, R. D.; Kafafi, S. A. Negative
Ion Energetics: NIST Standard Reference Database 19B, 1993.

9426 J . Org. Chem., Vol. 61, No. 26, 1996
Lee and Grabowski
Ta ble 4. Resu lts for th e Rea ction s of a Ser ies of An ion s
w ith Dim eth yl Disu lfid ea
Sch em e 1. Rea ction P a th w a ys by w h ich En ola tes
Rea ct w ith Meth a n ol-O-d a
observed products
HA-
CH3S-
CH3S- (H2A)
CH3SCH2S-
CH3SA-
–O
–O
x
x
x
x
O–
x
x
x
O–
x
a
O–
“SEP” refers to separation of an ion-molecule complex.
x
x
Although not shown, some (enolate/MeOD)- cluster is formed in
these reactions as well (see text).
a
Check mark (x) indicates the presence of that product.
alkoxide studied generates CH3SCH2S- when allowed to
react with MeSSMe, but does not form thiomethoxylated
product.
greater quantities for acetone enolate (25%) than for
propanal enolate (7%).
Rea ctivity of La r ger An ion s. In order to examine
the consistency of reactivity of the probe reagents as the
size of alkoxides and enolates increased, we studied each
reagent with selected larger ions. For example, the
Discu ssion
When establishing a general methodology, certain
criteria must be met in order to ensure a standard for
reliability. For these chemical probes, we sought the
following qualities. First, the probe should yield a unique
product when allowed to react with an alkoxide or an
enolate; that is, its reaction with all alkoxides should
yield a product indicative of reaction with an alkoxide
as opposed to an enolate, and its reaction with all
enolates should yield a product indicative of reaction with
an enolate. Second, its reactions with both alkoxides and
enolates should be fast and efficient; if reactions are too
slow, then a “negative” result wherein one does not see
an expected product could mean either that the ion of
unknown structure is not of the correct form for the probe
reaction or that its reaction with the probe is too slow to
yield any observable products. Third, on a practical level,
the probe should be readily available, easy to use, and
generally applicable to other systems in addition to those
described herein.
reaction of 2-pentanone enolate (∆Hacid ) 368.6 kcal
28
mol-1
)
with MeOD yielded only cluster as product and
did not exhibit the desired exchange (eq 12). Large anion
(12)
Meth a n ol-O-d . The ability of enolates to exchange
their protons for deuterons when allowed to react with
an appropriate deuterated substrate is well-known
(Scheme 1).14 The general utility of MeOD as a probe to
discern between enolates and alkoxides, however, is
limited. First, alkoxides and enolates are both observed
to cluster; therefore, presence of an (anion/MeOD)-
cluster is inconclusive of either alkoxide or enolate
presence. Since clustering is the only behavior alkoxides
exhibit, the unequivocal assignment of structure as an
alkoxide is not possible. Second, the efficiency for the
overall enolate reactions are low. The main complication
of this type of slow reaction is that the lack of observed
exchange is inconclusive; no exchange could mean either
that the unknown structure is not an enolate or else that
the probe reaction is simply too inefficient to yield the
expected exchange product. An extension of this is the
problem displayed by propanal enolate: clustering and
exchange are probably in kinetic competition, and the
clustering is far more efficient (eq 14, pathway B). This
issue is also evidenced by the interesting observation that
CH3CHCHO-(CH3CH2CHO) cluster exchanges three pro-
1-hexene-3-oxide reacts with methyl nitrite in the ex-
pected fashion to produce NO2 and ONCH2O-, but the
-
rate of the reaction is a factor of 10 slower than that of
its three-carbon counterpart, isopropoxide (kobs ) (7.7 (
0.1) × 10-11 cm3 molecule-1 s-1, eq 13).
(13)
The results of various larger anions with dimethyl
disulfide are compiled in Table 4. All enolates react
efficiently with MeSSMe to generate CH3SA- as the
major product, and no CH3SCH2S- is formed. The large

Anion Structure Determination in the Gas Phase
J . Org. Chem., Vol. 61, No. 26, 1996 9427
The enolates behave as nucleophiles (soft bases) and
react, at least in part, by attack at the nitrogen of methyl
nitrite, followed by extrusion of CH3O- in a Claisen-type
condensation (Scheme 3). Subsequent deprotonation of
the nitrosated carbonyl compound by methoxide within
the ion-molecule complex ultimately yields the nitro-
sated enolate anion (m/ z 86 for propanal and acetone).
Alternatively, if after the Claisen-like reaction the meth-
oxide adds to the carbonyl carbon, a retro-Claisen con-
densation occurs and ONCH2- (m/ z 44) and ONChHCH3
(m/ z 58) are produced for acetone and propanal enolate,
respectively. Like alkoxides, enolates exhibit SN2-C as
well, but in minor amounts (7% for acetone and propanal
enolates).
(14)
tons for deuterons, although no exchange is observed for
CH3CHCHO-. Third, and perhaps most important, the
exchange reaction is not general for all enolates: al-
though acetone enolate exchanges five protons (compli-
cated by an equally efficient clustering reaction in 0.3
Torr of helium), propanal enolate and 2-pentanone eno-
late do not. Therefore, MeOD is an unreliable structural
probe reagent, useful in a small number of cases, all of
which would require extensive controls in order to ensure
a high degree of confidence in any results obtained.
Meth yl Nitr ite. The reactivity of enolates and alkox-
ides with methyl nitrite has been described by King15 and
Nibbering16 and their co-workers and is generalized in
reaction Schemes 2 and 3. Methyl nitrite is an ambident
reagent with several potential reaction sites, including
at carbon, at hydrogen, and at nitrogen. Its potential as
a probe reagent can be explained using classical hard-
soft acid-base (HSAB) theory: one would expect a hard
base to favor the H site (deprotonation) while a soft base
would favor an alternative site.29 Reaction at the two
different sites would yield two discrete products, indica-
tive of soft or hard base reaction. The main issue, then,
is whether alkoxides and enolates will behave in classic
HSAB fashion. Such behavior is indeed observed when
alkoxides and enolates are allowed to react with MeONO.
Alkoxides (Scheme 2) behave as hard, localized bases
which eliminate across the C-O bond of methyl nitrite.16
The resultant NO- can undergo nucleophilic addition
with the newly formed H2CdO within the complex, to
ultimately yield OdNCH2O- (m/ z 60).10 Alternatively,
the NO- can cluster to the newly formed neutral alcohol
to yield NO-(ROH). NO- itself is not observed in our
flowing afterglow, and if any is formed, it undergoes
collision-induced electron detachment under the existing
swarm conditions (P ) 0.3 Torr, T ) 300 K). In addition
to the three ion channels from the initial elimination
Alkoxides, (hard bases) react with MeONO to produce
the unique ion OdNCH2O- whereas enolates, soft bases,
produce nitrosated products.29 MeONO, however, pos-
sesses drawbacks which hinder it from being an ideal
probe. First, as in the case of MeOD, even small, simple
enolates react with MeONO with a low reaction efficiency
whereas an ideal probe would be near unit efficient in
its reactions. Even more limiting is the fact that the
efficiency decreases with increasing anion size; 1-hexene-
3-oxide reacts with an efficiency of 0.040 whereas the
propoxides react at efficiencies of 0.11 and 0.32. Second,
on a more practical level, methyl nitrite is inconvenient.
It is not readily available in pure form; it is a gas and
must be synthesized and isolated, preferably on the day
of experimentation as the compound is photoactive (BDE
[CH3O-NO] ) 41.8 kcal mol-1).27 Although methyl
nitrite can be generated in situ (the bulk of the evaporate
from a 10:1 mixture of methanol:isoamyl nitrite is CH3-
ONO),30 we have found that this method does not provide
the concentration of MeONO necessary to drive a diag-
nostic reaction to any degree of completion in a flowing
afterglow and should not be used when utilizing MeONO
as a chemical probe.9,13 Thus, while methyl nitrite neatly
capitalizes on the differences in activity between an
alkoxide and an enolate, its low efficiency and its chang-
ing reactivity properties with increasing anion size urge
caution when used for careful structural determination.
Dim eth yl Disu lfid e. The reactions of anions with
dimethyl disulfide are described in Schemes 4 and 5.4
Dimethyl disulfide, like methyl nitrite, possesses at least
reaction, alkoxides also display a substantial amount of
-
SN2 at carbon (SN2-C) to yield NO2
.
Sch em e 2. Rea ction P a th w a ys by w h ich Alk oxid es Rea ct w ith Meth yl Nitr itea
a
The ultimate ionic products of the various competing channels are shown in heavy boxes for clarity. “SEP” refers to separation;
“ADD”, addition; “E2CX”, elimination across a C-X bond, “SN2-X”, substitution at atom X.

9428 J . Org. Chem., Vol. 61, No. 26, 1996
Lee and Grabowski
Sch em e 3. Rea ction P a th w a ys by w h ich En ola tes Rea ct w ith Meth yl Nitr itea
a
The ultimate ionic products of the various competing channels are shown in heavy boxes for clarity. “SEP” refers to separation;
“ADD”, addition; “PT”, proton transfer; “SN2-X” substitution at atom X.
Sch em e 4. Rea ction P a th w a ys by w h ich
Alk oxid es Rea ct w ith Dim eth yl Disu lfid ea
The prototypical alkoxides and enolates were found to
behave as classic hard and soft bases, respectively, when
allowed to react with dimethyl disulfide. Alkoxides
eliminate across the C-S bond of dimethyl disulfide
(Scheme 4); elimination followed by separation yields
methanethiolate (CH3S-, m/ z 47) and/or the cluster of
methanethiolate with the newly formed alcohol (CH3S--
(ROH), m/ z 107 for model three-carbon alcohols). Elimi-
nation can also be followed by readdition of the meth-
anethiolate into the resultant thioformaldehyde, to ulti-
mately yield the unique product CH3SCH2S- (m/ z 93).4
Enolates, in analogy to their reactions with MeONO,
behave solely as nucleophiles, attacking a sulfur center
of dimethyl disulfide to form a complex containing the
thiomethoxylated carbonyl compound and methanethi-
olate (CH3S-). Deprotonation of the thiomethoxylated
neutral by methanethiolate within the complex, followed
by separation, yields the unique thiomethoxylated eno-
late ion ((CH3S)RCdC(O-)CH3), m/ z 103 for model three-
carbon enolates).
In summary, the prototypical alkoxides react with
dimethyl disulfide to form mostly CH3SCH2S- and no
thiomethoxylated product; in contrast, the propanal and
acetone enolates form mostly thiomethoxylated product
and do not produce CH3SCH2S-. Interestingly, no SN2-C
to generate CH3SS- is observed for any of the anions
(∆Hacid(CH3SSH) ) 343.1 kcal mol-1).31 In addition to
the formation of unique products for the different types
of reactant ions, the reactions of both types of anions
proceed at high efficiency (unit efficiency for the alkox-
ides, and 0.3-0.5 efficiency for the enolates). Dimethyl
disulfide therefore fulfills the criteria required for a
reliable, general structural probe. First, alkoxides yield
the unique product CH3SCH2S-, and enolates, thiomethox-
a
The ultimate ionic products of the various competing channels
are shown in heavy boxes for clarity. “SEP” refers to separation;
“ADD”, addition; “E2CX”, elimination across a C-X bond.
two potential reaction sites, at hydrogen and at sulfur.
Again, HSAB theory would predict that a hard base
would prefer deprotonation while a soft base would prefer
an alternative site. This potential ambident behavior of
MeSSMe, as with MeONO, renders it ideal as a choice
for further exploration as a chemical probe reagent.
(29) Pearson, R. G. Surv. Prog. Chem. 1969, 5, 1-52.
(30) Caldwell, G.; Bartmess, J . E. Org. Mass Spectrom. 1982, 17,
456.
(31) Moran, S.; Ellison, G. B. J . Phys. Chem. 1988, 92, 1794-1803.

Anion Structure Determination in the Gas Phase
J . Org. Chem., Vol. 61, No. 26, 1996 9429
Sch em e 5. Rea ction P a th w a ys by w h ich En ola tes Rea ct w ith Dim eth yl Disu lfid ea
a
The ultimate ionic products of the various competing channels are shown in heavy boxes for clarity. “SEP” refers to separation; “PT”,
proton transfer.
ylated product. This uniqueness of reactivity for alkox-
ides versus enolates is also consistent over a range of
larger alkoxides and enolates (Table 4). Second, the
reactions of both alkoxides and enolates with dimethyl
disulfide are efficient; furthermore, the efficiency stays
quite high for larger anions (Eff (hexanal enolate) )
0.22)). Third, dimethyl disulfide is practical: com-
mercially available in high purity, it can be used without
further purification and is easily handled since it is a
volatile liquid. The importance of the establishment of
dimethyl disulfide as the chemical reagent of choice for
alkoxide/enolate isomer discrimination is borne out by
our use of it to successfully probe anionic oxy-Cope
rearrangements which were heretofore believed to not
proceed in the gas phase.9,13
odology were methanol-O-d, methyl nitrite, and dimethyl
disulfide. In the course of our studies, we have found
dimethyl disulfide to be the most reliable chemical probe
reagent: not only is it easily obtained in a pure and
readily handled form, but it reacts efficiently with both
alkoxides and enolates, and its ambident nature leads
to a mode of reactivity that generates a unique, distinc-
tive product for alkoxides versus enolates. Such a
general, reliable probe can be used not only for the model
compounds described herein but for any alkoxide and for
any enolate, whenever such differentiation is needed,
whether for rearrangement studies or any other struc-
tural studies.13 Methyl nitrite in concentrated form, as
opposed to in situ generation, is useful when used in
conjunction with a more reliable probe such as dimethyl
disulfide.13 Methanol-O-d, however, is not suitable for
use as a structural probe reagent, and cannot be depend-
ably used for alkoxide-enolate discrimination.9
Con clu sion s
Gas-phase mechanism elucidation is of particular
importance in revealing intrinsic reactivity, free of
solvent and counterion effects. Mechanistic studies by
nature must rely on structure determination. Our initial
interest in the elucidation of the gas-phase anionic oxy-
Cope rearrangement led to the need for a general,
selective process to discriminate between alkoxides and
enolates. Toward this end, we examined chemical probes
which would capitalize on the differences in structure and
reactivity between alkoxides and enolates. The chemical
probes characterized in our search for a general meth-
Ack n ow led gm en t. We gratefully acknowledge sup-
port of this research through a grant farm the National
Science Foundation. J .K.L. thanks the Department of
Defense (National Defense Science and Engineering
Graduate Fellowship in Chemistry P-28231-CH-NDF)
for support and the Dow Chemical Co. for a fellowship
administered by the Organic Division of the American
Chemical Society.
J O961463J