Mechanisms for the expulsion of propene from ionized propyl phenyl ethers in the gas phase

Article abstract of DOI:10.1021/ja9607594

Unimolecular dissociation of ionized, deuterium-substituted propyl phenyl ethers has been investigated by means of photoionization mass spectrometry as a technique for insuring that different isotopomers can be compared at the same internal energy. Elimin

Full text of DOI:10.1021/ja9607594

J. Am. Chem. Soc. 1996, 118, 9661-9668
9661
Mechanisms for the Expulsion of Propene from Ionized Propyl
Phenyl Ethers in the Gas Phase
†
,‡
John C. Traeger and Thomas Hellman Morton*
Contribution from the School of Chemistry, LaTrobe UniVersity, Bundoora, Victoria, Australia,
and Department of Chemistry, UniVersity of California, RiVerside, California 92521-0403
X
ReceiVed March 8, 1996
Abstract: Unimolecular dissociation of ionized, deuterium-substituted propyl phenyl ethers has been investigated
by means of photoionization mass spectrometry as a technique for insuring that different isotopomers can be compared
at the same internal energy. Elimination of neutral propene is the only ionic fragmentation detected in the energy
domain studied but is found to take place via competing pathways. The contrast between measured fragment ion
•
+
•+
ratios from parent ions having internal energies 0.8 eV above the dissociation barriers, HOPh :DOPh ) 1.23 for
•
+
•+
•+
PhOCH2CH2CD3 (1c) Versus HOPh :DOPh g 1.38 for PhOCH(CD3)CH3 (4b), demonstrates that 4b is not an
•+
obligatory intermediate in the dissociation of 1c . Examination of ten other deuterated analogues of n-propyl phenyl
ether permits an assignment of branching ratios. There are two ways to interpret the data using first-order kinetics:
a phenomenological analysis (which considers the contributions from different chain positions) and a mechanistic
analysis (which considers competition among four pathways). These two analyses lead to different conclusions, and
the mechanistic analysis is to be preferred. For 1c the molecular ion decomposes predominantly (>90%) via a
+
•
[
CD3CHCH3 PhO ] ion-neutral complex, with an additional 7% passing through complexes that randomize hydrogen
within the propyl cation. Only 1% of ionized 1c fragments via site-specific â-transfer from the propyl side chain.
Slightly more than 1% decomposes via site-specific γ-transfer, which can be ascribed to intervention of ion-neutral
complexes containing unrearranged, corner-protonated cyclopropanes. Correction for these alternative pathways
leads to the conclusion that the primary isotope effect for the Brønsted acid-base reaction between isopropyl cation
and phenoxy radical within ion-neutral complexes has a value kH/kD ) 1.21. Proportions of the various pathways
for ionized n-propyl phenyl ethers depend on the isotopic substitution pattern. Nearly one-sixth of PhOCD2CH2CD3
(1d) molecular ions decompose via site-specific â-transfer or γ-transfer (in a ratio of 1:1.1). The increased competition
in this d5 analogue accounts for the reported conformational dependence of its supersonic jet/REMPI mass spectrum.
Despite its ostensible simplicity, heterolysis of the C-X bond
data from intermediate regimes, those quantitative conclusions
must be viewed as approximate, at best. The present study uses
first differential photoionization efficiency curves to extract data
corresponding to the monoenergetic limit. The results for the
n-propyl phenyl ethers are compared with those for deuterated
isopropyl phenyl ethers.
+
in ions of the form CH3CH2CH2X has proven to be quite a
+
complicated reaction. Because the nascent CH3CH2CH2
1
structure does not correspond to a stable geometry, the propyl
group isomerizes in concert with departure of the leaving group
X. Isotopic labeling studies in solution reveal that 1,2-hydride
transfer to form isopropyl cation (the global minimum on the
+
Background
C3H7 potential energy surface) is not always the initial
2
rearrangement. The role of solvent in guiding the reaction is
Thirty years ago MacLeod and Djerassi described the
complete lack of regioselectivity in the expulsion of butene from
not clear. Hence it warrants exploration in the gas phase. This
paper examines C-O bond heterolysis of ionized n-propyl
4
ionized n-butyl phenyl ether in the gas phase. By separately
•
+
phenyl ether, nPrOPh to form ion-neutral complexes. Pho-
toionization mass spectrometry of nPrOPh and 11 deuterated
analogues provides a data set from which we measure branching
ratios and infer an unanticipated mechanism operating in
competition with those previously described.
labeling each carbon of the alkyl chain with deuterium, they
•+
showed that the mass spectrum exhibits a mixture of HOPh
(
•+
m/z 94) and DOPh (m/z 95), regardless of which position is
deuterated. A decade later Benoit and Harrison reported the
same phenomenon in the decomposition of ionized n-propyl
Product ratios are analyzed in terms of first-order kinetics.
This type of analysis can be rigorously applied in two limiting
cases: either where reactant molecules are monoenergetic
5
phenyl ether (nPrOPh, 1). Their analysis of the mass spectra
of PhOCD2CH2CH3 (1a), PhOCH2CD2CH3 (1b), and PhOCH2-
CH2CD3 (1c) suggested that the selection from among the R-,
â-, and γ-positions in eq 1 is nearly statistical.
(
isolated and all having the same internal energy), or where
3
they are thermalized and in contact with an infinite heat bath.
While first-order kinetics has been used in the past to interpret
(1)
†
LaTrobe University.
University of California, Riverside.
Abstract published in AdVance ACS Abstracts, September 15, 1996.
‡
X
(1) Koch, W.; Liu, B.; Schleyer, P. v. R. J. Am. Chem. Soc. 1989, 111,
In that same year (1976) two other pertinent papers appeared.
Collection and analysis of the neutral fragments expelled from
3
479-3480.
(2) (a) Lee, C. C.; Cessna, A. J.; Ko, E. C. F.; Vassie, S. J. Am. Chem.
Soc. 1973, 95, 5688-5692. (b) Vogel, P. Carbocation Chemistry; Elsevi-
(4) MacLeod, J. K.; Djerassi, C. J. Am. Chem. Soc. 1966, 88, 1840-
er: Amsterdam, 1985; pp 330-335.
1841.
(3) Klots, C. E. In Unimolecular and Bimolecular Ion-Molecule Reac-
(5) Benoit, F. M.; Harrison, A. G. Org. Mass Spectrom. 1976, 11, 599-
608.
tions; Wiley: 1994; pp 311-335.
S0002-7863(96)00759-7 CCC: $12.00 © 1996 American Chemical Society

9662 J. Am. Chem. Soc., Vol. 118, No. 40, 1996
Traeger and Morton
ionized n-butyl phenyl ether (the first experiment of its kind)6
showed that 1- and 2-butenes constitute >95% of the expelled
C4H8, and the distribution of isotopic label in the neutral
products corroborated the mass spectrometric results of
ring, either of which would proceed via â-hydrogen transfer
and would have given a ratio of R ) 0. Thus we list three
mechanistic options for propene expulsion (in order of increasing
value of the R,γ:â ratio): (i) â-elimination or 1,5-hydrogen shift
(R ) 0); (ii) hydrogen randomization (R ) 2.5); (iii) n-Pr f
+
•
iPr isomerization or intervention of an [iPr PhO ] ion-neutral
complex with no additional hydrogen transposition (R ) 5).
We must add to this list the option of predominant R-transfer,
as reported by Veith and Gross in propene expulsion from
(2)
+
12
metastable nPr2NdCH2 . This can be interpreted (based on
ab initio calculations on the stable complexes between n-propyl
MacLeod and Djerassi. With regard to nPrOPh, a Field
Ionization Kinetics study demonstrated that the lack of regi-
oselectivity in the expulsion of C3H6 is manifested at the very
1
3
cation and ammonia ) as stemming from a hydrogen-bonded
complex between the newly-formed methyl group and the
departing base, which eq 4 depicts. This mechanism thus
constitutes a fourth option: (iV) R-elimination (R ) ∞).
It would seem from this panoply of mechanisms that any
experimental result could be fit by some combination of options
i-iV. For example, one-color REMPI of indvidual conformers
-
11
•+
shortest time scales (10
could be detected.
s) on which HOPh fragment ions
7
(3)
•
+
•+
of 1b and 1d in supersonic jets gives HOPh :DOPh ratios
that correspond to values of R within experimental error of 2.5.14
While parsimony suggests option ii as a likely interpretation,
•+
•+
the jet/REMPI experiments show that the HOPh :DOPh ratio
from 1d varies significantly (ANOVA: F ) 14.00, df ) 37)
among the three conformational isomers whose (0,0)-bands were
examined. Such variation is hard to comprehend if only a single
pathway operates but is readily understood if two (or more)
pathways are in competition. In recent studies of PhSCH2CH2-
CH3 and of substituted propyl aryl ethers Nibbering and co-
In 1980 one of us proposed the intervention of ion-neutral
8
complexes to account for all of these results, as represented
for ionized n-butyl phenyl ether in eq 2. This type of
mechanism accounted not only for eq 2 but also for alkene
expulsion from a variety of other ionized primary alkyl phenyl
9
ethers. The C4H8 isomer distribution in eq 2 corresponds to a
15
workers propose just such a composite mechanism, embracing
options i-iii.
comparatively cold cation (even though the ionized precursor
was formed by 70 eV electron impact), with less methylcyclo-
propane formation than is observed from n-butyldiazonium ions
The barrier to eq 1 has recently been experimentally
1
6
10
determined to be 1.47 eV. For mass spectrometers with time
scales on the order of 1-10 µs kinetic shifts >0.5 eV are
expected. In such an apparatus one ought not to observe the
onset of fragmentation until the parent ions have internal
energies g2 eV. Consequently a conventional mass spectro-
metric experiment looks at the products of ion-neutral com-
plexes that contain g0.7 eV, more than twice the energy
in solution. Extending this interpretation to eq 1 presupposed
+
that all of the hydrogens randomize in the C3H7 cation within
+
•
a [C3H7 PhO ] ion-neutral complex, 2.
(4)
+
difference between isopropyl cation (the most stable C3H7
isomer) and corner-protonated cyclopropane.
That hypothesis was further explored in a two-color Reso-
nance Enhanced Multiphoton Ionization (REMPI) study of
1
11
PhOCD2CH2CD3 (1d) and 1b. After a naive correction for
isotope effect, the ratio of R- and γ-hydrogen transfer to
â-hydrogen transfer was found to be R ) 2.5, just as would be
predicted by the hydrogen scrambling mechanism. By contrast,
the value of the ratio in the electron impact Mass Resolved Ion
Kinetic Energy Spectrum (MIKES) was measured to be R )
(5)
The experimental results below provide evidence for contri-
bution from a previously unexamined pathway, exemplified by
eq 5. This pathway corresponds to exclusive γ-hydrogen
transfer and represents a fifth option: (V) γ-Elimination via
intervention of an ion-neutral complex containing an unrear-
ranged corner-protonated cyclopropane or via a γ-distonic
intermediate (R ) ∞).
4
.3. This latter result can easily be incorporated into the
mechanistic hypothesis: at the lower internal energies of ions
+
studied by MIKES, the C3H7 cations within the ion-neutral
complexes undergo slow transposition of hydrogen, reacting (for
the most part) as though the complex contains an isopropyl
cation that undergoes no further rearrangement. This latter
mechanism is represented in eq 3 and predicts a statistical ratio
of R ) 5. (The same ratio would obtain if the ionized n-propyl
ether simply isomerized to the ionized isopropyl ether.)
We contrast the mechanism in eq 3 with pure vicinal
elimination or 1,5-hydrogen shift from the propyl chain to the
Experimental Section
Compounds 1b and 1d were synthesized as previously reported.11
Deuterium was introduced into the other compounds using LiAlD
Compound 1a was prepared from propionic acid; 1c, 1g, and 1k from
12) Veith, H. J.; Gross, J. H. Org. Mass Spectrom. 1991, 26, 1097-
1108.
(13) Audier, H. E.; Morton, T. H. Org. Mass Spectrom. 1993, 28, 1218-
1224.
(14) Song, K.; van Eijk, A.; Shaler, T. A.; Morton, T. H. J. Am. Chem.
4
.
(
(
6) Burns, F. B.; Morton, T. H. J. Am. Chem. Soc. 1976, 98, 7308-
313.
7) Borchers, F.; Levsen, K.; Beckey, H. D. Int. J. Mass Spectrom. Ion
7
(
Phys. 1976, 21, 125-132.
(
(
(
8) Morton, T. H. J. Am. Chem. Soc. 1980, 102, 1596-1602.
Soc. 1994, 116, 4455-4460.
9) McAdoo, D. J.; Morton, T. H. Acc. Chem. Res. 1993, 26, 295-302.
10) Friedman, L.; Bayless, J. H. J. Am. Chem. Soc. 1969, 91, 1790-
(15) (a) van Amsterdam, M. W.; Ingemann, S.; Nibbering, N. M. M. J.
Mass Spectrom. 1995, 30, 43-51. (b) Matimba, H. E.; Ingemann, S.;
Nibbering, N. M. M. J. Mass Spectrom. 1996, 31, 609-622.
(16) Weddle, G.; Dunbar, R. C.; Song, K.; Morton, T. H. J. Am. Chem.
Soc. 1995, 117, 2573-2580.
1
1
794.
(
39.
11) Chronister, E. L.; Morton, T. H. J. Am. Chem. Soc. 1990, 112, 133-

Propene Expulsion from Propyl Phenyl Ethers
J. Am. Chem. Soc., Vol. 118, No. 40, 1996 9663
Experimental Results
Previous studies have investigated compounds 1a-d to assess
•+
•+
positional selectivity. From the experimental HOPh :DOPh
ratios proportions of R-, â-, and γ-transfer (expressed in terms
of the mole fractions corresponding to undeuterated 1: XR, Xâ,
and Xγ) can be estimated. Since no previous study has looked
at more than four different deuterated analogues at one time, a
uniform isotope effect has had to be assumed in order to extract
phenomenological partial relative rate factors from the data (i.e.,
to assign contributions from each position of the carbon chain).
Such a dissection also presumes that ion decay obeys first-order
kinetics, which does not adequately describe the general case
for unimolecular decay of a collection of reactive ions having
2
0,21
a distribution of internal energies.
We have enthusiastically made use of these two approxima-
tions in the past. In the present study, however, a larger number
of deuterated analogues permits determination of separate
isotope effects for positions R, â, and γ. Analysis of the
Figure 1. First differential photoionization efficiency (PIE) curves for
the molecular ion (m/z 136) and the sole fragment ion (m/z 94) from
undeuterated n-propyl phenyl ether (1).
•+
•+
HOPh :DOPh ratios as a function of ionizing energy provides
a more rigorous basis for kinetic analysis of the data. The first
3
2
-phenoxypropionic acid; 1e from propionic-2,2-d acid (Aldrich); 1f
•+
•+
differential of the HOPh :DOPh ratio as a function of photon
energy allows a comparison of monoenergetic ions. However,
as will be presented below, we believe that a phenomenological
picture (as opposed to use of a mechanistic model) leads to a
misleading interpretation of the experimental results. Figure 1
depicts the first differential photoionization energy (PIE) curves
for undeuterated 1. Onset of ionization corresponds to an
adiabatic ionization energy of IE ) 8.08 ( 0.02 eV. The
and 1j from 3-phenoxy-1,2-epoxypropane (Aldrich); 1h from phenoxy-
acetone; 1i from propionaldehyde; and 4a and 4b from 2-phenoxypro-
pionic acid. Chemical purity g98% (except for traces of solvent) was
assessed by GLC/MS analysis. Most compounds were purified by two
successive distillations at atmospheric pressure (bp 180-185° for the
n-propyl phenyl ethers; bp 170-175° for the isopropyl phenyl ethers).
As previously reported, 1d contains about 8% of d
f is estimated to contain 3% 1b and 5% 1j. From mass spectrometry
the remaining deuterated compounds were gauged to be g99 atom%
isotopomers,11 and
4
1
•+
intensity of the molecular ion current (M ) m/z 136) rises
•
+
•+
D. While the measured HOPh :DOPh ratios from 1d and 1f might
be as much as 10% too high (owing to incomplete deuteration),
experimental m/z 94:m/z 95 ratios were corrected only for 6.6% natural
monotonically with the photon energy hν until the the extent
of fragmentation becomes substantial. The value of the first
•
+
differential of M is >0 below 10.1 eV. In the domain of this
experiment m/z 94 is the only fragment ion. The fragment ion
current rises monotonically from its onset between 9.8 and 9.9
eV. As a consequence, that curve in Figure 1 is uniformly >0.
As can be seen in Figure 1, both curves exhibit crests and
troughs at energies >10.5 eV.
13
C abundance and not for incomplete deuteration.
The apparatus for measuring photoionization efficiency (PIE) curves
1
7
has been described in detail elsewhere. Briefly, a microcomputer-
controlled photoionization mass spectrometer makes use of the hydrogen
pseudocontinuum and a Seya-Namioka monochromator equipped with
1
8
a holographically ruled diffraction grating. The resolution of the
monochromator was fixed at 1.35 Å, and the absolute energy scale
was calibrated with atomic emission lines to an accuracy of better than
In photoionizing molecules with thermal energy ꢀ and
ionization energy IE, one expects the photoelectrons to be
ejected with kinetic energies ranging from zero to hν + ꢀ -
IE. The fragment ion current at a given value of hν thus
represents the decomposition of molecular ions having a
distribution of internal energies. To a good approximation,
however, the first differential of the fragment ion current as a
function of hν is proportional to the number of fragment ions
formed from molecular ions having internal energy equal to hν
0
.003 eV. All experiments were performed at ambient temperature
-
3
(297 K) with sample pressures of 10 Pa in the ion-source region.
Flight time between the ionization source and the mass selector is
estimated to be on the order of 5 µs. Experimental adiabatic 0 K
ionization energies (IEs) correspond to the first observed vibrational
progression peak in the molecular ion first differential PIE curve. All
differential PIE curves were obtained from the experimental data using
a 15-point Fourier transform filter for smoothing with the program
HORIZON (Star Blue Software, Inc.) before simple first derivatives
+
ꢀ - IE. The m/z 94 first differential curve therefore
corresponds to a plot of the efficiency of fragmentation as a
function of the internal energy of the molecular ion. Deuterated
1
9
were taken. Phenomenological and mechanistic models were fitted
•+
•+
to the experimental HOPh :DOPh ratios using the MINSQ nonlinear
least-squares program in the SCIENTIST package (Micromath, Inc.).
Branching ratios for the mechanistic model were determined at the
energy (10.35 eV) corresponding to the first minimum in the molecular
ion intensity above the onset of fragmentation in the first differential
PIE curve (cf. Figure 1). Ab initio UHF geometries were optimized
with the 6-31G** basis set using GAUSSIAN94 (Gaussian, Inc.) on
the Cray C90 at the San Diego Supercomputing Center and SPARTAN
•
+
precursors exhibit two fragment ions, HOPh (m/z 94) and
DOPh (m/z 95), and the first differential of the m/z 94:m/z 95
•
+
1
3
ratio (corrected for C natural abundance) with respect to hν
represents the relative efficiency of the two competing frag-
mentations at a given parent ion internal energy.
First differential PIE curves for HOPh :DOPh ratios from
compounds 1a and 1b are reproduced in Figure 2. The two
curves are well separated from one another, so it is apparent
that mechanism ii (complete alkyl hydrogen randomization)
cannot be the exclusive pathway at any energy in the domain
•+
•+
(Wavefunction, Inc.) software. Potential energy minima were con-
firmed by means of analytical frequency calculations.
(
17) (a) Traeger, J. C.; McLoughlin, R. G. J. Am. Chem. Soc. 1981, 103,
(for if it were, the result from 1a would have been the same as
3
647-3652. (b) Traeger, J. C.; McLoughlin, R. G.; Nicholson, A. J. C. J.
Am. Chem. Soc. 1982, 104, 5318-5322. (c) Traeger, J. C. Int. J. Mass
from 1b). The curves in Figure 2 exhibit crests and troughs
Spectrom. Ion Processes 1984, 58, 259-271.
(
18) Traeger, J. C. Rapid Commun. Mass Spectrom. 1996, 10, 119-
(20) Kondrat, R. W.; Morton, T. H. Org. Mass Spectrom. 1991, 26, 410-
415.
(21) Audier, H. E.; Berthomieu, D.; Morton, T. H. J. Org. Chem. 1995,
60, 7198-7208.
1
22.
19) Traeger, J. C.; Hudson, C. E.; McAdoo, D. J. J. Am. Soc. Mass
Spectrom. 1996, 7, 73-81.
(

9664 J. Am. Chem. Soc., Vol. 118, No. 40, 1996
Traeger and Morton
out that most of the curves have crests at 10.35 eV, and all
have troughs within (0.15 eV of 10.5 eV. Table 1 summarizes
the values of XR-Xγ extrapolated for undeuterated 1 (and the
corresponding phenomenological isotope effects) calculated at
(7)
1
0.35 eV (using either the ratio of ion currents or the first
differential at that photon energy), at the crests (using the local
maxima in the first derivative curve near 10.25 eV for 1b and
1
d in place of the ratios at 10.35 eV), and at the troughs near
•
+
•+
10.5 eV. The difference among the values thus derived
Figure 2. First differential curves of the HOPh :DOPh ratios for
the three gem-n-propyl-d ethers, 1a, 1b, and 1k. The dashed line shows
the value of the ratio R ) [first differential ratio from 1b/first differential
represents an estimate of systematic error. Since XR is less than
2
2
7
the statistical value expected for pure ii ( ) it does not seem
1
/2
ratio from 1d].
likely that pathway iV (pure R-transfer) plays any significant
role.
Table 1. Mole Fractions of Transfer from Positions R-γ
Corresponding to Undeuterated 1) and Phenomenological Isotope
To dissect the relative contributions of i, ii, iii, and V, six
more deuterated propyl phenyl ethers were studied: another d3
analogue, PhOCH2CD2CH2D (1f), as well as three d1 analogues,
PhOCH2CH2CH2D (1g), PhOCH2CHDCH3 (1h), and
PhOCHDCH2CH3 (1i). In addition two other d2 analogues were
examined, PhOCH2CHDCH2D (1j) and PhOCH2CH2CD2H
(1k). The PIE curve for 1j (not shown) exhibits a shoulder at
(
Effects (k
H
/k
D
) for Perdeuterated Positions
X
R
X
â
X
γ
H
(k /k
)
D R
(k
H
/k
D
)
â
H D γ
(k /k )
a
1
1
0.35 eV
0.18 0.18 0.65
0.18 0.16 0.65
0.15 0.20 0.65
0.20 0.17 0.62
0.75
0.87
0.68
0.82
1.18
1.05
1.43
1.10
2.25
2.33
2.36
2.02
b
0.35 eV
b
crests
troughs
b
•
+
•+
1
(
0.35 eV (HOPh :DOPh ) 4.35) and a trough at 10.55 eV
a
From ratios of ion currents measured at hν ) 10.35 eV. b From
•+
•+
HOPh :DOPh ) 3.6). Figure 5 reproduces the first dif-
first differentials of the ion currents.
ferential PIE curves for the d1 analogues.
Consider the contrast between the curves for 1a and 1k in
Figure 2. If pathways i-iii operated to the same extent for
<
10.5 eV, which do not coincide with any obvious features of
the fragment ion curve in Figure 1.
•
+
•+
both 1a and 1k, their HOPh :DOPh ratios should have been
equal. If only pathway iii were operating for 1a and only
pathway ii were operating for 1k, the ratio for the former would
Figure 3 reproduces the first differential PIE curve for the
•
+
•+
HOPh :DOPh ratio from compound 1c, while Figure 4
displays the curves for 1d and the R,R,â,â-d4 compound
PhOCD2CD2CH3 (1e). The value of ratio R (extracted from
the curves for 1b and 1d) is plotted as a dashed line in Figure
5
2
4
2
have been / ) 1.25 times that of the latter (assuming the
same proton transfer isotope effects for ii and iii). Allowing
both ii and iii to operate for 1a or 1k would diminish this ratio
of ratios, as would permitting i or iV (or both) to operate in
competition. Since the experimental ratio of ratios is 1.62 at
10.35 eV, pathway V must be invoked.
2
. As is apparent, R is >5 at 10 eV but falls to 3.6-4.3
11
(comparable to the value reported from the MIKES spectrum )
in the domain 10.1-11.0 eV. Clearly mechanism iii by itself
cannot account for the high value of R near 10 eV. A more
detailed analysis of the data shows that pathway V contributes
to varying extents, depending on the deuterium labeling pattern.
Equations 6 and 7 illustrate how phenomenological values
of XR, Xâ, and Xγ were estimated from the experimental data
by assuming an independent isotope effect for each position of
the chain. A set of five schemes (eqs 6 and 7, along with
analogous schemes for 1a,b,d) provides a basis for analyzing
ratios from the first differential curves for 1a-e. Several of
these curves exhibit pronounced crests and troughs, which do
not necessarily coincide with one another for different deuterated
analogues. This is not surprising, since kinetic shifts of different
isotopomers need not be identical; hence, one PIE curve may
be displaced from another. In substituting experimental ratios
for 1a-e into the five equations with five unknowns, one cannot
decide a priori whether data points should be selected at crests
We draw the following distinction between a phenomenologi-
cal picture and a mechanistic model to interpret these results.
A phenomenological picture uses the ten data points to assess
the relative contributions of positions R-γ as well as eight
independent primary and secondary isotope effects. A mecha-
nistic model assumes a specific set of competing pathways and
evaluates isotope effects for appropriate steps. While a
phenomenological picture gives an exact, unique solution for
the ten equations in ten unknowns, the results in Table 1 imply
that at least three-eighths of ionized 1 decomposes via V, a
conclusion that is hard to accept. Moreover, we cannot assess
from a phenomenological picture the proportions of i-iii that
accompany V. To do this requires that we go beyond a
phenomenological picture: the data from 1a-j must instead
be fitted to a composite of four pathways. However, the
additional constraint imposed by such a mechanistic model
means that observed and calculated data points may no longer
agree exactly. In examining six different mechanistic models,
we report the one that gives the best fit.
(or troughs) at energies that are close to one another, or whether
(6)
One criterion of quality of fit is the agreement between
calculated and observed values for 1k. For this compound the
•
+
•+
ratio at 10.35 eV, HOPh :DOPh ) 2.42, comes close to a
trough in the first differential curve. Phenomenological predic-
tions of the value for 1k range from 1.71 (based on ratios of
data points should all be taken at the same value of hν. It turns

Propene Expulsion from Propyl Phenyl Ethers
J. Am. Chem. Soc., Vol. 118, No. 40, 1996 9665
curves to control for the internal energy spread of photoions,22
so that first-order kinetics may appropriately be applied to
extract mechanistic information. The discussion below will
focus on photoionization at hν ) 10.35 eV, but the PIE curves
reproduced in Figures 1-5 suggest that the conclusions hold
true for the entire domain 10-11 eV. This particular value of
hν is chosen because it corresponds to the point where the first
differential PIE curve for the molecular ion manifests its first
local minimum in Figure 1. Also, most of the first differential
•+
•+
HOPh :DOPh ratios exhibit local maxima at this energy, and
here the results for ratios of ion currents do not differ greatly
from the results from the first differential curves. From the
measured IE of n-propyl phenyl ether (8.08 eV), the internal
energy of ions at hν ) 10.35 eV in the first differential curve
is 2.27 ( .03 eV, which is 10.35 - 8.08 - 1.47 ) 0.80 ( 0.05
eV above the energy barrier for propene expulsion.16
•
+
•+
Figure 3. First differential curves of the HOPh :DOPh ratios for
methyl-d n-propyl and isopropyl phenyl ethers (1c and 4b) and gem
isopropyl-d phenyl ether (4a).
•+
Relative Contributions of Competing Pathways. HOPh :
3
•+
DOPh ratios for the ten deuterated analogues 1a-j were fitted
using seven independent isotope effects applied to the competi-
tion among pathways i, ii, iii, and V. While a unique solution
is easily obtained for a phenomenological analysis of these ten
data points, a perfect fit is not easily achieved when a specific
mechanistic model is imposed. Because many more isotope
effects can be imagined than there are data, it has been necessary
to combine several of them. In the best fit, the primary isotope
effect for D-shift from a â-CHD in pathway iii (relative to
H-shift from a CH2) was set equal to the primary isotope effects
for pathways i and V. These three isotope effects are taken all
to have the same value, kH/kD ) x. Similarly, the primary
2
ion currents at 10.35 eV) and 1.86 (based on troughs near 10.5
eV in the first differential curves), to 2.34 (based on first
differentials at 10.35 eV) and 2.36 (based on crests near 10.35
eV in the first differential curves, as compared with a crest value
of 2.6 at 10.5 eV). While none of the predictions matches the
respective observed value exactly, the result from first dif-
ferentials at 10.35 eV comes closest. Therefore calculations
for mechanistic models were also performed using first dif-
ferentials at 10.35 eV.
A mechanistic model explicitly weighs the contributions of
individual pathways, but i-V are not linearly independent.
Because the preponderance of evidence suggests that iV is very
minor, it will be neglected, and only the percent contributions
of pathways i, ii, iii, and V are considered. The first differential
ratios for 1a-j at 10.35 eV can then be used to predict a value
+
isotope effects for D -transfer from deuterated propyl ions
within complexes were taken to be the same, regardless of
whether the ion scrambles (pathway ii) or not (pathway iii).
These isotope effects have the value kH/kD ) y. Secondary
isotope effects were neglected for all of these reactions. The
only secondary isotope effects explicitly considered were those
that affect the first step of pathway iii: the R-, â-, and γ-position
secondary isotope effects (treated independently of one another)
on the hydride shift that isomerizes n-propyl to isopropyl cation
•
+
•+
for 1k and also to extract a HOPh :DOPh ratio for each of
the individual pathways. The Discussion section compares
HOPh :DOPh ratios corresponding to pathway iii by itself
with the HOPh :DOPh ratios for the isopropyl-d2 phenyl ether
a and the isopropyl-d3 ether 4b. Figure 3 displays the first
•
+
•+
•+
•+
4
(e.g., the H-shift from â-CHD relative to â-CH , whose value
2
differential PIE curves for these isopropyl ethers, and the curve
for 1c is seen to be well separated from that for for 4b. The
onset of ionization for 4 corresponds to an adiabatic IE of 7.98
is represented as k /k ) z). A separate isotope effect for
H
D
D-shift from â-CD was used. Finally, the isotope effect on
2
the competition between ii and iii was explicitly considered and
(
0.02 eV. Given that the isomerization of ionized 1 to ionized
found to be large. This last isotope effect was taken to be
multiplicative (i.e., the effect for n deuteria was assumed equal
to the effect for one deuterium raised to the nth power), as were
the R- and γ-position secondary isotope effects on hydride
transfer.
4
is exothermic and that elimination of propene from ionized 4
16
has a lower barrier than from ionized 1, it is conceivable that
pathway iii operates via the intermediacy of ionized 4. As shall
be discussed below, comparison of 4a with 1k and of 4b with
1
c shows that the n-propyl and isopropyl systems manifest
Isotopic substitution is found to affect the branching among
competing pathways. The sum of the contributions of i and V
is small (e5.5%) except where position R is deuterated (1a,
different isotope effects.
Discussion
1
d, 1e, and 1i). In these latter cases the contribution of V is
This experimental investigation has three principal objec-
tives: (1) to assess relative contributions of competing pathways
to propene expulsion from nPrOPh ; (2) to gauge isotope
effects for eq 3 and compare them with the isotope effect for
propene expulsion from iPrOPh (ionized 4); and (3) to provide
an explanation for the reported conformational dependence
of HOPh :DOPh ratios in supersonic jet/REMPI mass spectra
of deuterated 1.
While some organic reactions proceed specifically via a single
mechanism, many others take place via several pathways, all
of which operate concurrently to yield the same final product.
The present data reveal that expulsion of propene from ionized
n-propyl phenyl ether falls into this second category. Our
investigation makes use of first differential ionization efficiency
greater than that of i. Compound 1d is the only instance where
the contribution of i exceeds 4%. According to the mechanistic
model undeuterated 1 should give only small proportions of i
and V: extrapolation predicts XR ) 0.29, Xâ ) 0.26, and Xγ )
•
+
•+
0
.45, considerably different from the contributions predicted
1
4
phenomenologically in Table 1.
•+
•+
The discrepancy between the phenomenological picture and
the mechanistic model signals a defect in the approximations
underlying the former. Were i, iV, and V the only pathways in
operation, the two interpretations would coincide. It would then
have been reasonable to suppose (as in eq 7) the isotope effects
(22) Hurzeler, H.; Inghram, M. G.; Morrison, J. D. J. Chem. Phys. 1958,
28, 76-82.

9666 J. Am. Chem. Soc., Vol. 118, No. 40, 1996
Traeger and Morton
greater than in ii), the net effect is opposite to what the
phenomenological model would predict. For example, from
Table 2 and y ) 1.21 we determine that 1a transfers 20% from
R and 21% from â, while 1b transfers 25% from R and 15%
from â. Nevertheless, 1e transfers 19% from R and 16% from
â, exhibiting an attenuation of R-transfer as a consequence of
â-deuteration. In conclusion, the assumptions underlying the
phenomenological picture do not coincide with conclusions
drawn from the mechanistic model. Between the two, the
mechanistic model is to be preferred.
(8)
•
+
•+
Figure 4. First differential curves of the HOPh :DOPh ratios for
n-propyl phenyl ethers having g3 deuteria with isotopic substitution
on two carbons (1d-f).
Regardless of how the experimental data are interpreted, site-
specific γ-transfer (pathway V) operates to a measurable extent.
-
1
UHF calculations set the γ-distonic ion 5 164 kJ mol
-
1
(
including a 2 kJ mol difference in zero point energy) above
nPrOPh , more than 20 kJ mol above the experimental barrier
•+
-1
5,7
height. Since previously published experiments have already
shown that the role of pathway V increases as internal energy
decreases (a result qualitatively in agreement with the data
presented here), ab initio calculations would appear to militate
against the intermediacy of 5. The tautomeric γ-distonic ion 6
-
1
is 29 kJ mol lower than 5. For 6 to be consistent with the
reported overall barrier height, the activation energy for the
•
+
intramolecular hydrogen atom abstraction 1 f 6 would have
-
1
to be <7 kJ mol . Although we have made no effort to
compute this transition state, it seems unlikely that the barrier
to hydrogen transfer could be this low. Moreover, for 6 to serve
an an intermediate it would have to undergo additional rear-
•
+
rangement to give HOPh , the structure that has been shown
•
+
•+
Figure 5. First differential curves of the HOPh :DOPh ratios for
the three n-propyl-d phenyl ethers, 1g-i.
•+ 16
to be the exclusive product of propene elimination from 1 .
Ab initio calculations place ion-neutral complex 2 slightly
below the experimental barrier for eq 1. We have previously
reported 2 as a local minimum (based on UHF normal modes
at 3-21G), but we find that the geometry achieved by conven-
tional optimization procedures at 6-31G** exhibits two negative
force constants and is predicted to collapse without a barrier to
1
1
6
Table 2. Best Fit between Model (Parameters Extracted from Data
for 1a-j) and Experiment, with Calculated Percentages of
Competing Pathways i, ii, iii, and V at 10.35 eV
•
+
•+
percent each pathway (calcd from 1a-j)b
HOPh :DOPh
obsda obsdb calcdb
i
ii
iii
v
•
+
4
. Despite that computational prediction, Figure 3 provides
1
1
1
1
1
1
1
1
1
1
1
1
1
-d
a
b
c
d
e
0
∞
∞
∞
0.5c
3.6
0.2
1.0
8.0
2.0
0.2
0.9
0.7
1.4
1.0
81.2c
68.6
17.3
6.6
5.7
13.2
5.0
16.5c
13.9
78.9
91.1
77.9
60.4
92.4
43.5
33.9
15.7
65.5
1.8c
14.0
3.5
1.2
8.6
24.3
2.4
2.8
4.2
5.3
experimental evidence that collapse does not compete effectively
against proton transfer (the second step of eq 3). The next
section provides the rationale behind this interpretation.
3.52 3.85 3.74
5.46 5.40 5.35
1.23 1.24 1.23
0.34 0.33 0.33
1.68 1.78 1.77
2.52 2.41 2.40
5.95 6.93 6.75
12.8 11.3 11.1
7.85 7.73 7.60
5.45 4.37 4.33
2.42 2.42 2.56
0
•+
•+
Isotope Effects for nPrOPh Compared to iPrOPh . The
most important derived isotope effects in this study are those
that act on pathway iii: the preference for H- versus D-shift
from the â-position (the first step of eq 3, for which x/z ) 6.5
( 0.9) and the normal isotope effect that we assign to the second
step of eq 3 (y ) 1.21 ( 0.03). These have appropriate
magnitudes for a hydride shift and a proton transfer, respectively.
The remaining isotope effects are best expressed in terms of
the relative proportions of competing pathways summarized in
Table 2. Extrapolation to a perdeuterated propyl side chain
predicts nearly complete suppression of ii.
f
g
h
i
j
k
-d
52.8
61.2
77.6
29.1
4.5c
c
c
c
c
1.2
22.2
74.3
2.3
c
c
c
4.2c
7
1.1
0.3
94.4
a
•+
•+
Ratios of HOPh to DOPh ion currents measured at hν ) 10.35
eV. b From first differentials of the ion currents. c Predicted from
parameters based on 1a-j.
for positions R and â to be independent of one another. The
mechanistic model does not make such an assumption. In the
phenomenological picture (given a normal isotope effect)
â-deuteration ought to choke off â-transfer, thereby increasing
the mole fractions of R- and γ-transfer. By contrast, the
mechanistic model concludes that a â-CD2 suppresses i (aug-
Now consider the operation of pathway iii. Strictly speaking,
this pathway represents isomerization of the n-propyl side chain
to an isopropyl without any further transposition of hydrogen.
Equation 3 portrays what we infer takes place. Alternatively,
+
rearrangement might have occurred via collapse of an [iPr
•
•+
PhO ] ion-neutral complex to 4 , followed by propene
expulsion. The value of y(k /k for proton transfer from propyl
menting V) but at the same time enhances iii at the expense of
H
D
1
3
•+
ii. Since the statistical weight of position R in iii is (which is
to phenoxy) is of particular relevance in assessing whether 4

Propene Expulsion from Propyl Phenyl Ethers
J. Am. Chem. Soc., Vol. 118, No. 40, 1996 9667
intervenes. From previous work ionized 4 is known to expel
propene more rapidly than does ionized 1 with the same internal
energy.16 While it is thus plausible for pathway iii to proceed
for the more robust conformational effect in 1d: pathways i
and V together constitute one-sixth of the fragmentation for 1d,
but less than 4% of the fragmentation for 1b. The competition
between formation of [propyl ion phenoxy radical] complexes
and these other pathways leads to a conformational sensitivity
of the mass spectra for 1d that is more pronounced than for 1b.
•+
•+
•+
via the sequence 1 f 4 f propene + PhOH , a comparison
of isotope effects demonstrates that ionized 1 does not decom-
pose in this fashion.
•+
If pathway iii did happen to pass through iPrOPh , 1k would
isomerize to 4a and 1c would isomerize to 4b. However, the
•
+
•+
HOPh :DOPh ratios for 4a and 4b at 10.35 eV, 2.68 ( 0.06
and 1.53 ( 0.01, are larger than for 1k and 1c. Correcting the
1
k and 1c results for contributions from i, ii, and V does not
•
+
•+
How might conformation affect this competition? Two
general hypotheses must be weighed. The first stipulates that
different conformers yield different distributions of excited 1•
upon multiphoton ionization. If UV photoelectron spectra of
bring their HOPh :DOPh ratios into line with those from 4a
and 4b. Since the heat of formation of 4 is 0.2 eV less than
•
+
+
•
+
that of 1 , it may be more appropriate to compare the 10.35
eV ratios from 1k and 1c with values from 4a and 4b at lower
energies. The ratio for 4a is HOPh :DOPh ) 3.03 at 10.1
eV, which declines monotonically with increasing energy over
the next 0.5 eV, as Figure 3 depicts. Below 10.4 eV, the ratio
from 4a remains significantly higher than the ratio from 1k.
For 4b the ratio has its minimum value at 10.1 eV, HOPh :
DOPh ) 1.38, which is larger than the ratio from 1c over the
domain from onset up to 10.75 eV (as can be seen from Figure
6
•
+
•+
homologues of 1 serve as a guide, many different doublet states
are accessible <13.8 eV (the energy of three 270 nm photons)
above the ground state neutral. Energy deposition by three-
photon ionization could well depend upon conformation,
particularly since the power dependence of the phenol• yield
indicates that eq 9 (where 1** stands for a superexcited neutral
+
•
+
•
+
•+ ‡
and [1 ] a vibrationally excited doublet in its lowest electronic
state) functions in parallel with the better-known ladder-
3
).
1
4
•
+
switching mechanism. If the conformation effect is simply a
matter of internal energy content, this hypothesis suggests that
other alkyl phenyl ethers should also manifest conformation-
dependent mass spectra, provided they have fragmentation
patterns at least as sensitive to internal energy as that of 1d.
Since 4 is known not to transpose hydrogens prior to
propene expulsion,16 isotope effects are calculated in a straight-
forward fashion for the isopropyl system. Competition between
•
+
•+
CH3 and CD3 is simply the HOPh :DOPh ratio from 4b and
ranges from kH/kD ) 1.38-1.62 in the domain 9.8-10.4 eV.
For the competition between CH3 and CD2H the average kH/kD
equals half the ratio from 4a and has a value of 1.35 ( 0.08 in
that domain. For isopropyl phenyl ether it is apparent that kH/
kD for H- versus D-transfer from a CD2H must differ slightly
from competition between a CH3 and a CD3. Nevertheless, the
contrast to the n-propyl case, y ) 1.21, is significant. Hence
we conclude that, although 4 would have been a kinetically
competent intermediate, pathway iii proceeds as shown in eq 3
and not via collapse of complex 2 to a covalently bonded ion.
Neither does 4 pass through complex 2 when it expels
propene, in agreement with previous work. n-Propyl and
•
+
•
+
The second hypothesis presupposes that the rotameric equi-
16
librium of ionized 1 is affected by the structure of the neutral
isopropyl parent ions eliminate via separate routes.
•+ ‡
precursor. It is hard to imagine that [1 ] decomposes before
•
+
•+
Conformational Dependence of HOPh :DOPh Ratios.
Supersonic jet/REMPI spectroscopy allows examination of the
mass spectra of individual conformational isomers of neutral
precursors. In such experiments, a cold gaseous molecule is
promoted to its first excited state (a resonant transition) and
then ionized by absorption of a second photon. When REMPI
is performed using a (0,0)-transition as the resonant absorption,
the electronically excited neutral is vibrationally cold. If the
excited state geometry does not differ greatly from that of the
ground state, then the conformation will not change in the short
time required for the subsequent ionization.
complete rotameric equilibration is achieved. The SCF elec-
•+
•+
tronic energy barrier to conversion of anti-1 to gauche-1
which have the same stability) is slightly less than we calculate
(
for the conversion of neutral, planar anti-1 to the more stable
-1
-1
neutral gauche-1 (15 kJ mol for the former versus 16 kJ mol
•+
for the latter). Assuming rapid preequilibrium of anti-1d with
gauche-1d , memory of the conformation of the neutral
•+
precursor should govern the relative proportions only for
pathways that require the propyl group to rearrange simulta-
neously with C-O bond cleavage. The coplanar C-methyl and
O-methylene bonds in the anti rotamer are better set up for eq
5
than in the gauche. Suppose a neutral precursor in the anti
conformation leads to a greater fraction of reactive anti ions.
This would predict more net transfer from the γ-position and,
(9)
•
+
•+
hence, a lower HOPh :DOPh ratio for anti-1d and a higher
ratio for anti-1b (relative to their gauche rotamers), as has been
The barrier to propene expulsion from 1•+ is so high16 that
three photons have to be absorbed at the (0,0) transition
wavelength (near 270 nm) in order to observe any fragmentation,
as eq 9 summarizes. The excitation spectrum of 1 contains
peaks from three separate conformers, in proportions that
14
reported experimentally.
ROPh•+ f [R+ PhO ] f [R′ PhO ] f HOPh
•
+
•
•+
(10)
Either explanation predicts that 1e ought to manifest an even
23
probably reflect their relative abundances at room temperature.
•+
•+
more pronounced dependence of its HOPh :DOPh ratio than
-
8
REMPI mass spectra of 1d (using 10 s laser pulses tuned to
•
+
•+
(23) Ruoff, R. S.; Klots, T. D.; Emilsson, T.; Gutowsky, H. S. J. Chem.
Phys. 1990, 93, 3142-3150.
(
0,0)-transitions) exhibit HOPh :DOPh ratios that depend
14
upon conformation. The variation among rotamers for 1b is
(
24) Winstein, S.; Holness, N. J. J. Am. Chem. Soc. 1955, 77, 5562-
not statistically significant. Table 2 provides an explanation
5578.

9668 J. Am. Chem. Soc., Vol. 118, No. 40, 1996
Traeger and Morton
has been observed for 1d. On the other hand, it also implies
heterolyses that produce stable ions (even if they fall into
shallow potential energy minima) will not show conformation-
dependent mass spectra. Equation 10 exemplifies a general case,
(2) Deuterium substitution significantly affects the proportions
of competing pathways. This phenomenon has been noted in
other gas phase reactions, to which the term “metabolic
switching” has been applied by analogy to isotope effects in
+
where R stands for a cation with the same connectivity as its
25
enzyme-catalyzed reactions.
As a consequence, a purely
+
neutral precursor, while R′ symbolizes a rearranged cation.
phenomenological picture leads to an incorrect interpretation
of the data for 1, predicting an implausibly high contribution
from site-specific γ-transfer (V). The contribution of this
pathway from mechanistic analysis, while not zero, is more
+
Here we speculate that R should endure long enough to lose
all dependence upon the geometry of its antecedents. As noted
•+
above, 1 does not fall into the category of reactions represented
by eq 10. Since the primary n-propyl cation has no discrete
consistent with the extent of cyclopropane formation reported
•
+
existence, nPrOPh passes directly to a complex containing a
rearranged cation. However, other ionized alkyl phenyl ethers
may well decompose as eq 10 represents.
from solution studies of nascent primary alkyl cations.2
(3) Isotope effects for ionized 1c and 1k are significantly
lower than for the isopropyl isomers 4a and 4b at the same
The present results provide further illustration of the analogy
between gas phase decompositions of ionized n-alkyl phenyl
ethers and solvolysis chemistry. More than 40 years ago,
Winstein and Holness noted that functionalized cyclohexanes
undergo SN1 reactions as equilibrating mixtures of conforma-
tional isomers, which they were able to lock into place by means
•+
internal energy. Consequently isomerization of nPrOPh to
•
+
iPrOPh does not occur (at least, not as the dominant
mechanism). Ionized 1 produces 2, which then expels propene
via proton transfer from the isopropyl cation to the phenoxy
radical.
24
of tert-butyl substituents. In an analogous fashion, ionization
of conformationally frozen neutrals may permit a glimpse of
rotameric effects on bond heterolyses in acyclic systems. If
ions produced by REMPI do exhibit memory effects as
described by the processes outlined above, then the presence
or absence of conformational dependence might serve as a tool
for assessing whether carbocations rearrange in concert with
the bond heterolyses that form them.
(4) R-Deuteration of 1 leads to increased participation of
pathway V, which we believe passes through ion-neutral
complex 3. Competition between this pathway and formation
•
of [propyl ion PhO ] complexes is sensitive to neutral precursor
conformation.
The mechanistic analysis described here considers branching
among four pathways (perhaps somewhat artificially). It is
plausible that some of these pathways should be combined. For
instance, ion-neutral complex 3 might be the initially formed
intermediate for both pathways iii and V, with transposition of
hydrogen simply competing with the proton transfer shown in
eq 5. Such hypotheses lead to predictions, for which experi-
mental tests are presently being devised.
Conclusions
The experiments described above lead to the following
findings:
(1) Using a mechanistic model we extract the relative
proportions of pathways i, ii, iii, and V in unimolecular
dissociations of ionized n-propyl phenyl ether (1) and its
Acknowledgment. Support for this work was provided by
the Australian Research Council and by NSF Grant CHE
+
deuterated analogues. Pathway iii, which passes through [iPr
•
PhO ] ion-neutral complexes (2), predominates for ions with
9
522604.
internal energies 0.8 eV above the dissociation barrier.
(
25) (a) Pr u¨ sse, T.; Fiedler, A.; Schwarz, H. HelV. Chim. Acta 1991, 74,
Supporting Information Available: RHF and UHF geom-
etries at 6-31G** (11 pages). See any current masthead page
for ordering and Internet access instructions.
1
1
127-1134. (b) Seemeyer, K.; Pr u¨ sse, T.; Schwarz, H. HelV. Chim. Acta
993, 76, 1632-1635. (c) Schalley, C. A.; Schr o¨ der, D.; Schwarz, H. J.
Am. Chem. Soc. 1994, 116, 11089-11097. (d) Raabe, N.; Karrass, S.;
Schwarz, H. Chem. Ber. 1995, 128, 649-650. (e) Schalley, C. A.; Schr o¨ der,
D.; Schwarz, H. Int. J. Mass Spectrom. Ion Proc. 1996, 153, 173-199.
JA9607594