Stereochemical analysis of deuterated alkyl chains by MS/MS

Article abstract of DOI:10.1021/jo9912247

Vicinally deuterated sec-alkyl phenyl ethers, CH₃(CH₂)(m)CH(OPh)CHD(CH₂)(n)CH₃, display significant differences in mass spectra between threo and erythro stereoisomers. MS/MS experiments, in which parent ions of a single mass are selected and their fragmentation patterns subsequently measured, show that alkene expulsion represents virtually the only decomposition pathway. Two types of MS/MS experiment are reported: mass- analyzed ion kinetic energy (MIKE) spectroscopy of metastable ions and collisionally activated decomposition (CAD) of stable ions. The expulsion of a deuterated alkene from a monodeuterated precursor yields ionized phenol, PhOH·⁺ (m/z 94). The expulsion of an undeuterated alkene yields PhOD·⁺ (m/z 95). Without exception, the ratios (PhOD·⁺/PhOH·⁺) from precursors in the threo series have values greater than their diastereomers in the erythro series. The ratio of ratios, r = PhOD·⁺/PhOH·⁺ for the threo divided by PhOD·⁺/PhOH·⁺ for the erythro, has a value of 1.2 for the 2- phenoxy-3-deuteriobutanes and larger values for all of the higher homologues up through the monodeuterated phenoxyoctanes (m + n = 4). The highest degree of stereoselectivity, r = 5.8, is measured for 3-phenoxy-4-deuteriohexane. Experiments with multiply deuterated analogues show that alkene elimination is highly regioselective, unlike the corresponding decompositions of ionized sec-alcohols or their acetates. The fact that a large fraction of ionized sec-alkyl phenyl ethers undergo stereospecific syn-elimination means that mass spectrometry has a useful capacity to distinguish one isotopically labeled diastereomer from another.

Full text of DOI:10.1021/jo9912247

J . Org. Chem. 2000, 65, 381-387
381
Ster eoch em ica l An a lysis of Deu ter a ted Alk yl Ch a in s by MS/MS
J . P. Morizur,† M. H. Taphanel,† Philip S. Mayer,‡ and Thomas Hellman Morton*,‡,§
COS, UMR CNRS 172, Universite´ Paris VI, boıˆte 45, 4 place J ussieu, 75252 Paris Cedex 05, France,
Department of Chemistry, University of CaliforniasRiverside, Riverside, California 92521-0403, and
DCMR, URA CNRS 1307, Ecole Polytechnique, 91128 Palaiseau, France
Received J uly 30, 1999
Vicinally deuterated sec-alkyl phenyl ethers, CH3(CH2)mCH(OPh)CHD(CH2)nCH3, display significant
differences in mass spectra between threo and erythro stereoisomers. MS/MS experiments, in which
parent ions of a single mass are selected and their fragmentation patterns subsequently measured,
show that alkene expulsion represents virtually the only decomposition pathway. Two types of
MS/MS experiment are reported: mass-analyzed ion kinetic energy (MIKE) spectroscopy of
metastable ions and collisionally activated decomposition (CAD) of stable ions. The expulsion of a
deuterated alkene from a monodeuterated precursor yields ionized phenol, PhOH•+ (m/z 94). The
expulsion of an undeuterated alkene yields PhOD•+ (m/z 95). Without exception, the ratios (PhOD•+/
PhOH•+) from precursors in the threo series have values greater than their diastereomers in the
erythro series. The ratio of ratios, r ) PhOD•+/PhOH•+ for the threo divided by PhOD•+/PhOH•+ for
the erythro, has a value of 1.2 for the 2-phenoxy-3-deuteriobutanes and larger values for all of the
higher homologues up through the monodeuterated phenoxyoctanes (m + n ) 4). The highest degree
of stereoselectivity, r ) 5.8, is measured for 3-phenoxy-4-deuteriohexane. Experiments with multiply
deuterated analogues show that alkene elimination is highly regioselective, unlike the corresponding
decompositions of ionized sec-alcohols or their acetates. The fact that a large fraction of ionized
sec-alkyl phenyl ethers undergo stereospecific syn-elimination means that mass spectrometry has
a useful capacity to distinguish one isotopically labeled diastereomer from another.
Mass spectrometry has been known for more than 30
years to have the capacity to distinguish between acyclic
diastereomers, especially (though not exclusively) when
there is just one pair of stereogenic centers that are
adjacent to one another.1 As yet, no systematic interpre-
tation has allowed this phenomenon to be employed in
the deduction of unknown stereochemistries. This paper
examines a homologous series of monodeuterated sec-
phenoxy-n-alkanes, CH3(CH2)mCH(OPh)CHD(CH2)nCH3.
In these compounds, the abundance of relative fragment
ions can be used to tell one diastereomer from another.
In 1988, we demonstrated that both threo- and erythro-
2-phenoxy-3-deuteriobutanes (1t, e) manifest small but
reproducible differences in the expulsion of butene from
their molecular ions.2 Recently, we published a study of
a half-dozen homologues, showing that alkene expulsion
exhibits even greater sensitivity to stereochemistry for
homologues with longer chain lengths.3 While the terms
threo and erythro are not, strictly speaking, the correct
designations for all of the homologues of 1, we shall
continue to use the symbols t and e to indicate the series
to which a given diastereomer belongs. Here, we report
the results for seven additional homologues, completing
the chain-length series from m + n ) 0 to m + n ) 4
(compounds 1-15 in Chart 1).
Exp er im en ta l Section
Metastable ion decompositions were studied by mass-
analyzed ion kinetic energy (MIKE) spectroscopy on reverse
Nier-J ohnson (B-E configuration) double-focusing mass spec-
trometers at UCR and at the Ecole Polytechnique in Palaiseau.
The relative ion abundancies from the former instrument were
measured by fitting the observed peaks as Gaussians using
the multipeak fit function of IGOR Pro 3.03 software, while
those from the latter instrument were measured as relative
peak areas. Ion ratios differed slightly between the two
instruments: for example, CH3CHOPhCD2CH3 gave a m/z 94:
m/z 95 ratio of 2.33 (SD ) 0.015) on the former and 2.02 (SD
) 0.13) on the latter. GC/MS/MS experiments (70 eV EI) were
performed using a quadrupole ion trap (QIT) coupled to a gas
chromatograph equipped with a 30 m DB5 fused-silica capil-
lary column (0.25 mm i.d., 0.25 mm film thickness).3
Deuterated compounds were prepared by standard methods.
Samples of RCH(OPh)CD2CH3 were synthesized by reacting
the carboxylic acids, RCH(OPh)COOH, with excess methyl-
lithium.4 This was followed by reduction of the ketone product
with LiAlD4 to the vic-phenoxy alcohol, which was dehydrox-
ylated as described below. CH3CH(OPh)CH2CD3 and CH3CH2-
CH(OPh)CD2CD3 were prepared and purified as previously
described.5 Vicinally perdeuterated compounds were prepared
by repetitive exchange of the corresponding ketones with D2O,
followed by reduction, conversion to the tosylate, and SN2
displacement with sodium phenoxide in refluxing THF. Single
diastereomers (of both pure compounds and mixtures of
positional isomers) were synthesized from the corresponding
† Universite´ Paris VI.
‡ University of CaliforniasRiverside.
§ Ecole Polytechnique.
(1) (a) Audier, H. E.; Felkin, H.; Fetizon, M.; Vetter, W. Bull. Soc.
Chim. Fr. 1965, 3236-3238. For reviews of this subject, see: (b) Green,
M. M. Top. Stereochem. 1976, 9, 35-110. (c) Turecek, F. Collect. Czech.
Chem. Commun. 1987, 52, 1928-1984. (d) Turecek, F.; Splitter, J . S.
Applications of Mass Spectrometry to Organic Stereochemistry; VCH
Publishers: New York, 1994; Chapter 4.
(2) Kondrat, R. W.; Morton, T. H. Org. Mass Spectrom. 1988, 23,
555-557.
(3) Taphanel, M. H.; Morizur, J . P.; Leblanc, D.; Borchardt, D.;
Morton, T. H. Anal. Chem. 1997, 69, 4191-4196.
(4) Rubottom, G. R.; Kim, C.-w. J . Org. Chem. 1983, 48, 1550-1552.
(5) Traeger, J . C.; Luna, A.; Tortajada, J .; Morton, T. H. J . Phys.
Chem. A 1999, 103, 2348-2358.
10.1021/jo9912247 CCC: $19.00 © 2000 American Chemical Society
Published on Web 12/29/1999

382 J . Org. Chem., Vol. 65, No. 2, 2000
Morizur et al.
Ch a r t 1. Str u ctu r es of th e Mon od eu ter a ted P h en oxy n -Alk a n es (u p th r ou gh octa n e) Th a t P ossess Tw o
Vicin a l Ster eogen ic Cen ter s
alkenes by the procedure exemplified below for the stereoiso-
mers of 15 (with the exception of the mixture of erythro-13
and erythro-14, which was synthesized from trans-3-octene via
epoxidation, reduction with LiAlD4, conversion to the bromide,
and displacement with sodium phenoxide).
4-P h en oxyocta n e-5-d 1 (15). A quantity of 1.12 g (10.0
mmol) of trans-4-octene (Sigma) was epoxidized with m-CPBA,
and the crude, sweet-smelling epoxide was sealed in a Pyrex
tube with 1.5 g of a 1:3 mixture of sodium phenoxide/phenol
and heated for 18 h at 130 °C in an oven. The reaction mixture
was then taken up in water with 30 mL of ether, and the
ethereal layer was separated and washed with 6 × 25 mL
portions of 20% aqueous NaOH to remove traces of unreacted
phenol. After washing with saturated aqueous NaCl and the
removal of the solvent, the resulting vic-phenoxy alcohol (1.30
g) was converted to the tosylate without any further purifica-
tion; the tosylate reduced with excess LiAlD4 in refluxing THF
to yield 0.23 g of 15t (bp 143-148°, 20 Torr), after distillation
under aspirator pressure. Starting with 0.48 g (4.3 mmol) of
cis-4-octene (Wiley Organics), the same procedure afforded 0.51
g (2.5 mmol) of 15e (bp 144-148°, 20 Torr, 57% overall yield),
after distillation.

Deuterated Alkyl Chains
Resu lts
J . Org. Chem., Vol. 65, No. 2, 2000 383
The approach described here examines the relative
intensities of two mass spectrometric fragments that
differ by 1 amu. Because the molecular ion decomposition
is not 100% stereospecific, 13C-natural abundance inter-
feres with the measurement of ion abundance ratios.
Therefore, the best results are obtained when the dis-
sociation of a single isotopic variant is examined. Many
modern mass spectrometers accomplish this by MS/MS,
that is, selecting a single m/z value in the first stage of
mass separation and looking at subsequent decomposi-
tions in a second stage. Instrumentally, this can be
performed by either using two (or more) mass spectrom-
eters in tandem or using a single mass spectrometer that
has the ability to store ions of a single m/z value and then
to analyze its fragmentation pattern in the same region
of space.
Both methods will be discussed below. MIKE spectros-
copy represents a tandem technique, wherein the m/z
value of the parent ion is selected by a magnetic (B) sector
and the fragment masses are analyzed later by an
electrostatic analyzer (ESA). The quadrupole ion trap
(QIT) exemplifies a storage instrument, where parent
ions are trapped in an appropriately tailored electric field.
Advantages of QIT include the ease with which it can be
interfaced with a GC and also the comparative insensi-
tivity of its resolution to the kinetic energy of the
fragment ions.
An ion can be given enough energy to dissociate at the
time of its creation or subsequently. On one hand,
metastable ions are those that form with enough internal
energy to decompose, but do so on a time scale longer
than that of mass analysis. Such ions pass unchanged
through the first stage of selection, and a portion of them
fragment prior to the second stage. The dissociations that
occur between the two stages of mass selection can be
specifically examined by a variety of methods. In the
present investigation, MIKE spectroscopy was used to
examine metastable ion decompositions.
On the other hand, ions can be activated between the
two stages of mass analysis. There are many ways to do
this. Here, we have made use of energetic collisions with
inert gas atoms. Collisionally activated dissociation
(CAD) in the present study was performed in the QIT,
with sample introduction via a GC interface. Some of the
helium carrier gas leaked into the mass spectrometer,
where it served as the collision gas. MIKE spectroscopy
can also be used for CAD studies, but we found that the
resulting peaks (whose shape was sensitive to the kinetic
energy that was released during dissociation) from higher
mass parent ions became too poorly resolved to perform
useful quantitation.
Metastable ions ordinarily have lower internal energies
than those that have been activated after ionization.
Consequently, their MIKE spectra often exhibit a smaller
variety of decomposition pathways than their corre-
sponding CAD spectra. Therefore, we shall start by
presenting data from the metastable ion decompositions.
Because some degree of regioselectivity is needed in order
to observe stereochemical discrimination to a useful
extent, questions of regiochemistry are dealt with first.
Regioselectivity. Before dealing with the issue of
stereoselectivity, it is important to probe the degree to
which an elimination reaction involves the stereogenic
center of interest. Figure 1 exhibits a MIKE spectrum of
F igu r e 1. Metastable ion decomposition pattern for ionized
4-octanol-3,3,5,5-d4 following 70 eV of electron ionization (EI)
with 8 kV of accelerating voltage. All fragments gm/z 65 are
shown.
the higher mass ions from vicinally perdeuterated 4-oc-
tanol (m/z g 65, which constitute >95% of the fragment
ion intensity). As can be seen, elimination of water takes
place, but it is a minor process. The principal dissociation
pathways are the formation of pentenyl cations and the
loss of alkyl radicals (either propyl or butyl). All of these
processes involve isotopic scrambling; hence, there are
multiple peaks corresponding to each decomposition of
the d4-alcohol. In particular, loss of water shows up as
three peaks (M - H2O, M - HOD, and M - D2O, in the
proportions 0.21:0.61:0.18), and this elimination cannot
be considered regioselective to any useful extent. The loss
of water from ionized alcohols has long been known to
exhibit poor regioselectivity.6 The present results show
that unimolecular hydrogen transposition (possibly by
reversible formation of distonic ions) creates an ad-
ditional complication.
The mass spectrum of the acetate ester of CH3CH2CD2-
CHOHCD2CH2CH2CH3 shows fewer peaks. The elimina-
tion of acetic acid represents the predominant dissocia-
tion of the molecular ion, representing 75-80% of the
fragment ion intensity in the metastable ion decomposi-
tions that we observe in the MIKE spectrum. This is
homologous to the stereoselective elimination that Green
and Vogt examined in detail for sec-butyl acetate,7 for
which they propose a six-member cyclic transition state
analogous to thermal syn-elimination from neutral ac-
etates.8 In the 4-octyl system, however, the elimination
is not regioselective. Losses of CH3COOD and losses of
CH3COOH (m/z 115 and 116) have the same abundances,
whereas a six-member cyclic transition state involving
the carbonyl would demand exclusively the former (un-
less, of course, the ion rearranges beforehand). The
(6) Derrick, P. J .; Burlingame, A. L. Acc. Chem. Res. 1974, 7, 328-
333.
(7) Green, M. M.; McCluskey, R. J .; Vogt, J . J . Am. Chem. Soc. 1982,
104, 2262-2269.
(8) (a) DePuy, C. H.; King, R. B. Chem. Rev. 1960, 60, 431-457. (b)
Skell, P. S.; Hall, W. L. J . Am. Chem. Soc. 1964, 86, 1557-1558.

384 J . Org. Chem., Vol. 65, No. 2, 2000
Morizur et al.
Ta ble 1. P h OD•+/P h OH•+ In ten sity Ra tios fr om Meta sta ble Ion Decom p osition s of Sec-p h en oxy-n -a lk a n e
Molecu la r Ion sa
vic-monodeuterated
dideuterated
1b
0.43
0.30
CH3CHOPhCD2H
0.27
c
1eb
6t
CH3CHOPhCD2CH2CH3
CH3CHOPhCD2CH3
CH3CHOPhCH2CD2CH3
CH3CD2CHOPhCH2CH3
0.61
0.26 (0.35d)
0.07 (0.10d)
0.32 (0.36d)
0.07 (0.10d)
2.3 (2.0d)
0.04
c
6e
15t
15e
0.60 (0.65d)
0.54 (0.55d)
0.44 (0.52d)
CH3CH2CHOPhCD2CH2CH3
CH3CD2CHOPhCH2CH2CH3
gem-monodeuterated
b
CH3CDOPhCH2CH3
0.03
0.04
trideuterated
c
CH3CDOPhCH2CH2CH3
CD3CHOPhCH3
0.58 (0.60d)
0.15 (0.16d)
0.03
0.10
0.15
CH3CH2CDOPhCH2CH3
0.03 (0.02d)
0.02 (0.025d)
0.03
CD3CHOPhCH2CH3
CH3CHOPhCH2CD3
CD3CHOPhCH2CH2CH3
CH3CH2CDOPhCH2CH2CH3
CH3CH2CDOPhCH2(CH2)2CH3
CH3(CH2)2CDOPhCH2(CH2)2CH3
vic-tetradeuterated
CH3CD2CHOPhCD2CH3
CH3CD2CHOPhCD2CH2CH3
CH3CD2CHOPhCD2(CH2)3CH3
CH3CH2CD2CHOPhCD2(CH2)2CH3
c
0.035 (0.025d)
CD3CHOPhCH2(CH2)3CH3
pentadeuterated
28 (20d)
25 (17d)
8 (10d)
CD3CHOPhCD2CH3
20
0.65
CH3CH2CHOPhCD2CD3
hexadeuterated
11 (11d)
CD3CHOPhCD3
>100d
37
CD3CDOPhCD2CH3
a
Measured by MIKE spectroscopy (70 eV EI, 8 kV accelerating voltage). Except where otherwise noted, reported values are those
measured on the 2-sector instrument at UCR. Reference 2. c Reference 11. Measured on the 2-sector instrument at Palaiseau.
b
d
mechanisms for water elimination from the ionized
alcohol and acetic acid elimination from the ionized
acetate remain unexplored, but it is apparent that these
are not good candidates for the stereoselective mass
spectrometry of long-chain compounds.
phenoxybutane, d4-3-phenoxypentane, d4-3-phenoxyhex-
ane, and d4-3-phenoxyoctane) also show high regioselec-
tivity. Likewise, the ratios for other geminally monodeu-
terated compounds in the present study (3-d1-3-phenoxy-
pentane, 3-d1-3-phenoxyhexane, and 3-d1-3-phenoxyhep-
tane) differ only slightly from that for 4-d1-4-phenoxy-
octane. The evidence demonstrates that regioselectivity
does not depend greatly upon chain length.
To probe regioselectivity in greater detail, we have
examined the effects of perdeuterating one methyl (1,1,1-
d3-2-phenoxypropane, 1,1,1-d3-2-phenoxybutane, 4,4,4-d3-
2-phenoxybutane, or 1,1,1-d3-2-phenoxyheptane) or one
methylene group (2,2-d2-2-phenoxybutane, 2,2-d2-3-phe-
noxypentane, 2,2-d2-3-phenoxyhexane, or 4,4-d2-3-phe-
noxyhexane). Results of MIKE spectra of deuterated
analogues (including previously published data) are
summarized in Table 1. The degree of regioselectivity
that we observe mirrors that which Audier and co-
workers demonstrated in their pioneering work a decade
ago for 2-phenoxypentanes.11 A methyl group adjacent
to the functionalized methylene makes a small contribu-
tion (roughly 15% of the transferred hydrogen, if isotope
effects are ignored), while a more distant methyl group
contributes far less.
In contrast to the fragmentation patterns of the alcohol
and acetate, the MIKE spectrum of the phenyl ether of
d4-4-octanol is dominated by a single peak. Elimination
from the ionized phenyl ether gives a neutral hydrocar-
bon fragment, and the positive charge remains with the
functional group, which departs in the form of ionized
phenol. This is consistent with the Audier-Stevenson
rule,9 which postulates that the electric charge reposes
on the fragment with the lower ionization energy (IE).
Since phenol has a lower IE than most disubstituted
alkenes (but one that is greater than or equal to the IE
of tri- or tetrasubstituted alkenes with g8 carbons),10 the
sparseness of the MIKE spectrum of CH3CH2CD2-
CHOPhCD2CH2CH2CH3 is actually quite informative:
the absence of any C8-ions indicates that rearrangement
has not taken place to give branched structures that can
undergo elimination. PhOH•+ and PhOD•+ together con-
stitute 95% of the total fragment ion current. All the
other phenoxy-n-alkanes studied here show even higher
proportions of phenol•+ in their MIKE spectra. By con-
Regioselectivities for the vicinally tetradeuterated
compounds are compared in Table 1. We suppose that
two decomposition pathways are available:5 one that
involves exclusive transfer of vicinal hydrogen versus a
nonselective pathway in which all the nonmethyl hydro-
gens randomize completely. To a first approximation, we
neglect any transfer from methyl groups. Under these
assumptions, the PhOD•+/PhOH•+ intensity ratio of 28
for CH3CD2CHOPhCD2CH3 represents a lower degree of
regioselectivity than the ratio of 25 for CH3CD2CHOPh-
CD2CH2CH3. In the former, the nonselective pathway
scrambles 4 D’s with 1 H, while in the latter, the non-
selective pathway randomizes 4 D’s with 3 H’s. In this
simplified model, each observed PhOD•+/PhOH•+ inten-
sity ratio should be weighted by the proportion of H’s
among the randomizing atoms: 1/5 for a d4-pentyl, 3/7
for a d4-hexyl, and 7/11 for a d4-octyl. On the basis of
trast, the MIKE spectrum of a branched isomer, (CH3)3-
CCH2CH(CH3)CH2OPh, exhibits both PhOH•+ and C8H16
,
•+
in a ratio of about 5:2, as would be expected on the basis
of the Audier-Stevenson rule.
Alkene expulsion exhibits a high degree of regioselec-
tivity: PhOD•+ is 11 times more abundant than PhOH•+
in the MIKE spectrum of CH3CH2CD2CHOPhCD2CH2-
CH2CH3. The formation of a small amount of PhOD•+
from the geminally monodeuterated analogue, 4-d1-4-
1
phenoxyoctane (approximately /30 the intensity of Ph-
OH•+), shows that a minor elimination pathway occurs
via hydrogen transfer from the carbon to which oxygen
is initially attached.
The PhOD•+/PhOH•+ intensity ratios for other vicinally
perdeuterated homologues that we have looked at (d5-2-
(9) Audier, H. E. Org. Mass Spectrom. 1970, 2, 283-298.
(10) Lias, S. G.; Bartmess, J . E.; Liebman, J . F.; Holmes, J . L.; Levin,
R. D.; Mallard, W. G. J . Phys. Chem. Ref. Data 1988, 17, Supplement
1.
(11) Audier, H. E.; Sozzi, G.; Mourgues, P.; Milliet, A. Org. Mass
Spectrom. 1987, 22, 746-747.

Deuterated Alkyl Chains
J . Org. Chem., Vol. 65, No. 2, 2000 385
Ta ble 2. Regio- a n d Ster eoch em ica l Dep en d en ce in th e
QIT/CAD Sp ectr a in Ter m s of Obser ved P h OD•+/P h OH•+
In ten sity Ra tios fr om GC/MS/MS Exp er im en tsa
Sch em e 1
CD3CHOPhCH2CH3
CH3CHOPhCD2CH3
0.13b
2.0
CH3CDOPhCH2CH3
CH3CHOPhCH2CD3
0.07
0.01
CD3CHOPhCD2CH3
1t
2t
3t
4t
5t
6t
7t
13
CD3CDOPhCD2CH3
1e
2e
3e
4e
5e
6e
7e
18
0.44b
0.61
0.34
0.59
0.29
0.41
0.55b
0.37b
0.36b
0.39b
1.34b
0.66b
0.63
0.63
0.28
0.36b
0.17
0.08
0.26
0.12
0.07
0.21b
0.21b
0.14b
0.14b
0.47b
0.39b
0.29
0.25
0.15
8t
9t
8e
9e
10t
11t
12t
13t
14t
15t
10e
11e
12e
13e
14e
15e
Sch em e 2
a
Experiments in which all other ions were ejected except the
molecular ion, followed by collisionally activated decomposition
b
(CAD) by collisions with neutral helium. Reference 2.
this rationale, the weighted regioselectivities become 5.6
for 3-pentyl, 10.7 for 3-hexyl, 5.1 for 3-octyl, and 7.0 for
4-octyl. Consequently, 3-hexyl gives the greatest degree
of regioselectivity. This is borne out by the observation
that the 3-d1-3-hexyl analogue gives the lowest PhOD•+/
PhOH•+ intensity ratio of all of the geminally monodeu-
terated compounds examined.
than their isomers in which the deuterium resides in a
longer alkyl group (r ) 5.8 for 6, 2.6 for 9, 2.1 for 13).
Ster eoselectivity. In every example studied here, the
isomer corresponding to the threo series exhibits a larger
PhOD•+/PhOH•+ intensity ratio than does its diastereo-
mer from the erythro series. This may be seen by
comparison of the CAD spectra listed in Table 2.
Discu ssion
The method described here is based on reactivity
differences between diastereomers. A mechanistic inter-
pretation complements the empirical correlation. The
general trend can be rationalized in terms of syn-
elimination, in which expulsion of a neutral trans-alkene
is favored over the cis-alkene, as Scheme 1 represents.
Density functional calculations on the simplest sec-alkyl
Regioselectivity in the CAD experiments (as exempli-
fied by the tabulated results for 2-phenoxybutanes) does
not vary greatly from what is observed in the MIKE
spectra.
i
It is tempting to try to correlate regioselectivity and
stereoselectivity in these elimination reactions, but the
data do not support such an inference. Consider the ratio
of ratios, r equals the PhOD•+/PhOH•+ ratio from the
threo divided by the PhOD•+/PhOH•+ ratio from the
erythro, as a measure of the degree of stereoselectivity.
3-Phenoxyhexane shows the highest degree of regiose-
lectivity in Table 1, and its 4-d1-analogues, 6, show the
highest values of r (3.7 in the MIKES, 5.8 in the CAD).
3-Phenoxypentane also displays a high degree of regio-
selectivity, and it turns out that the PhOD•+/PhOH•+
ratios from CAD on the ionized 2-d1-3-phenoxypentanes,
3, are not significantly different from those for 6. Yet the
PhOD•+/PhOH•+ ratios in both cases do differ from those
for the respective 2-phenoxy-3-deuterated analogues 2
and 4. In other words, it is not possible to establish a
clear-cut correlation between regioselectivity and stereo-
selectivity, because the 2-d1-3-phenoxyhexanes (4) display
some of the lowest values of r.
phenyl ether (2-phenoxypropane, PrOPh) indicate that
alkene expulsion passes through a four-member cyclic
transition state,5 a view that is corroborated by the
following experimental data: (1) the calculations predict
an energy barrier for the decomposition of iPrOPh•+ that
agrees with the observed lower bound;12 (2) the fragment
ion has the structure of a phenol radical cation rather
than a cyclohexadienone;3,12 and (3) the high degree of
regioselectivity is described here.
The contrast with alkene expulsions from neutral
acetates (which pass through six-member cyclic transi-
tion states)8 and with ionized alcohols or acetates (which
do not display high regioselectivity, as exemplified by
Figure 1) suggests that ionized alkyl phenyl ethers
constitute a special category. The peculiarity of this
category derives, in part, from the low hydrogen atom
affinities (HA) of their radical cations. The thermody-
namic cycle in Scheme 2 summarizes how to calculate
i
the HA of PrOPh from its proton affinity (PA, which is
-∆H for putting H+ onto a gaseous molecule) and its
ionization energy (IE), as well as the ionization energy
of a hydrogen atom (IE ) 313.5 kcal mol-1).10 The
We cannot discern a systematic variation with chain-
length. The 2-phenoxybutanes (1) display the lowest
degree of stereoselectivity (r ) 1.2). Beyond that, how-
ever, the degree of stereoselectivity shows no regular
increase with the number of carbons in the chain. The
only trend appears to be that, among the 3-phenoxyal-
kanes, compounds with d1-ethyls display lower degrees
of stereoselectivity (r ) 2.4 for 5, 1.8 for 8, 1.7 for 12)
i
ionization energy for PrOPh13 is 183.5 ( 1 kcal mol-1
(12) Weddle, G. H.; Dunbar, R. C.; Song, K.; Morton, T. H. J . Am.
Chem. Soc. 1995, 117, 2573-2580.
(13) Traeger, J . C.; Morton, T. H. J . Am. Chem. Soc. 1996, 118,
9661-9668.

386 J . Org. Chem., Vol. 65, No. 2, 2000
Morizur et al.
hydrogen from carbon to oxygen. Because an ionized
phenoxy group has such a low HA, formation of a distonic
ion, as represented on the left of Figure 2, should be
endothermic (unlike the formation of distonic ions from
ionized aliphatic ethers17). The other limiting case,
represented by the solid curve to the right, corresponds
to dissociation of the C-O bond to form an ion-neutral
complex. Photoionization studies of deuterated sec-alkyl
phenyl ethers5 have been interpreted as showing that
20-50% of the molecular ions decompose via this mech-
anism, depending on internal energy. However, a complex-
mediated pathway that proceeds via a classical sec-alkyl
carbocation cannot display stereochemical discrimination,
because (by definition) the cation must be able to rotate
freely within the complex.18,19
The [s-R+ PhO•] ion-neutral complex does not cor-
respond to a local minimum on the potential energy
surface, because there is no energetic barrier to its
collapse back to s-ROPh•+. Nevertheless, such ion-
neutral complexes behave as bona fide intermediates
with nonzero lifetimes, because the degrees of freedom
liberated by breaking a covalent bond stabilize them
entropically. As noted above, the majority of ions do not
decompose via ion-neutral complexes. Published density
functional calculations5 predict that the favored pathway
does not correspond to either of the solid curves in Figure
2 but instead passes through a four-member cyclic
transition state (the dashed curve, which represents a
concerted pathway in Figure 2). This accounts for the
observed diastereoselectivity of vicinal alkene expulsion,
although the experimental data do not rule out a distonic
intermediate unequivocally.
The reaction coordinate diagram in Figure 2 prompts
the question as to how energetic perturbations affect the
reaction coordinate. What happens when the HA of the
aryloxy group is further decreased (such as by putting
on a p-methoxy group, which ought to lower the IE by
about 10 kcal mol-1)8 or is increased (such as by replacing
the phenoxy with a pyridyloxy)? Thus far, the variations
that have been studied have exerted too drastic of an
effect, and it has not proven possible to explore their
consequences for diastereoselection. We observe that, on
one hand, a p-methoxy substituent leads predominantly
to the homolytic fragmentation shown in eq 1.19 On the
other hand, ionized sec-alkyl pyridyl ethers undergo
double-hydrogen transfer,20 as exemplified by ionized
4-iso-propoxypyridine in eq 2.21 In the former instance,
no hydrogen transfer takes place, while in the latter
instance, two hydrogens move from the side chain to the
aryloxy. Neither case is suitable for exploring stereo-
selectivity.
F igu r e 2. Three-dimensional reaction coordinate diagram
illustrating plausible mechanisms for expulsion of alkene from
an ionized sec-alkyl phenyl ether (s-ROPh•+) in the gas phase.
(the same as for longer homologues up to 2-phenoxyoc-
tane5). The PA of iPrOPh is calculated to be ∼2 kcal mol-1
greater than that of coumaran (measured experimentally
as 204 kcal mol-1),14 with the para-carbon as the most
basic site. It is not easy to measure the PA of the oxygen
of iPrOPh experimentally, but it is calculated to be ∼7.5
kcal mol-1 less basic than coumaran. On that basis, the
i
HA of the oxygen of ionized PrOPh is 66.5 kcal mol-1
,
while the HA of the para-carbon is 76 kcal mol-1. The
oxygen HA is at least 25 kcal mol-1 lower than the
weakest C-H bond dissociation energy in an unbranched,
neutral alkyl ether.15
By contrast, the same calculation on ionized ethyl
10
acetate (IE ) 230.5 kcal mol-1, PA ) 200.5 kcal mol-1
)
gives an HA for the carbonyl oxygen of 117.5 kcal mol-1
.
This is comparable to the bond dissociation energy of
water and is much higher than the bond dissociation
energy of any sp3 C-H bond. Hence, an ionized acetate
ester can easily abstract a hydrogen atom from an alkyl
chain to form a distonic ion. Following the reactivity/
selectivity principle, one should not be surprised that
long-chain sec-alkyl acetates do not appear to discrimi-
nate among methylene hydrogens in their mass spectro-
metric decompositions.
Alkene expulsion from ionized sec-alkyl phenyl ethers
(s-ROPh•+) can take place via three pathways, as the
3-dimensional reaction coordinate diagram (sometimes
called a More O’Ferrall J encks plot16) in Figure 2
portrays. This diagram depicts two limiting cases by
means of solid curves, depending upon whether C-O
dissociation occurs prior to or after the transfer of a
(1)
There may well exist patterns of substituting the
phenoxy group that will enhance discrimination between
(17) (a) Morton, T. H.; Beauchamp, J . L. J . Am. Chem. Soc. 1975,
97, 2355-2362. (b) Morton, T. H.; Beauchamp, J . L. J . Am. Chem. Soc.
1977, 99, 1288.
(14) Audier, H. E.; Berthomieu, D.; Leblanc, D.; Morton, T. H. Int.
J . Mass Spectrom. Ion Processes 1998, 175, 133-147.
(15) Block, D. A.; Armstrong, D. A.; Rauk, A. J . Phys. Chem. A 1999,
103, 3562-3568.
(16) Lowry, T. H.; Richardson, K. S. Mechanism and Theory in
Organic Chemistry, 3rd ed.; Harper & Row: New York, 1987; pp 211-
229.
(18) Morton, T. H. Org. Mass Spectrom. 1992, 27, 353-368.
(19) McAdoo, D. J .; Morton, T. H. Acc. Chem. Res. 1993, 26, 295-
302.
(20) Biermann, H. W.; Freeman, W. P.; Morton, T. H. J . Am. Chem.
Soc. 1982, 104, 2307-2308.
(21) Molenaar-Langeveld, T. A.; Gremmen, C.; Ingemann, S.; Nib-
bering, N. M. M. Int. J . Mass Spectrom., in press.

Deuterated Alkyl Chains
J . Org. Chem., Vol. 65, No. 2, 2000 387
stereoisomers from the erythro series. Vicinal syn-
elimination expels trans-alkenes preferentially. The con-
sistency of this difference in fragment ion intensity ratios
provides a general method for discriminating between
diastereomers.
(2)
threo and erythro, but they have yet to be discovered.
The effect of the position of deuteration on the degree of
stereoselectivity also remains to be understood. The
correlation between regiochemistry and stereoselectivity
is by no means obvious. The pair of dideuterated 3-phe-
noxyhexanes in Table 1 exhibit PhOD•+/PhOH•+ intensity
ratios that are nearly the same. While there is little
regioselective discrimination between methylenes at
positions 2 and 4, the stereoselectivity between mono-
deuterated t and e isomers in position 2 (compounds 5),
r ) 2.4, is dramatically less than that observed between
t and e isomers in position 4 (compounds 6), r ) 5.8.
Conformational equilibria may play an important role,
but that issue has only begun to be addressed in simpler
alkyl phenyl ethers.22
Ack n ow led gm en t. This work is dedicated to the
memory of The Honorable George E. Brown, J r., (1920-
1999), Chairman of the House Committee on Science,
Space and Technology, 1991-1994. This work was
supported by NSF grant CHE 9983610 and by the
CNRS. The authors are grateful to Henri Audier,
Director of the Laboratoire des Me´canismes Re´action-
nels of the Ecole Polytechnique, Palaiseau, where much
of this work was performed.
Su p p or tin g In for m a tion Ava ila ble: Source GC mass
spectra (70 eV) and infrared C-D stretching frequencies of
compounds 2-15.This material is available free of charge via
the Internet at http://pubs.acs.org.
Con clu sion s
J O9912247
All unbranched sec-alkyl phenyl ethers with a single
deuterium vicinal to the phenoxy group exhibit more
expulsion of undeuterated alkene (to form PhOD•+) in the
mass spectra of the threo series than their respective
(22) (a) Song, K.; van Eijk, A.; Shaler, T. A.; Morton, T. H. J . Am.
Chem. Soc. 1994, 116, 4455-4460. (b) Kohler, B. E.; Morton, T. H.;
Nguyen, V.; Shaler, T. A. J . Phys. Chem. A 1999, 103, 2302-2309.