Coupled reactions of condensation and charge transfer. 1. Formation of olefin dimer ions in reactions with ionized aromatics. Gas-phase studies

Article abstract of DOI:10.1021/ja962635x

The toluene radical ion C₆H₅CH₃(·)⁺, generated by resonance two-photon ionization, does not react with a single isobutene molecule (i-C₄H₈) which has a significantly higher ionization potential (ΔIP = 0.42 eV). However, a reaction is observed involving two i-C₄H₈ molecules, to form the dimer ion C₈H₁₆(·)⁺. A coupled reaction of dimer formation and charge transfer to the dimer is exothermic if the product is an ionized hexene with a low IF. Correspondingly, the observed nominal second-order rate coefficients, (5-25) x 10^-12 cm³ s^-1, are enhanced by a factor of > 10⁵ over the expected value for direct endothermic charge transfer. Pressure and concentration effects suggest a sequential mechanism that proceeds through a C₆H₅CH₃·⁺(i-C₄H₈) reactive π complex. The complex can isomerize to a nonreactive CH₃C₆H₄-t-C₄H₉(·)⁺ adduct,or react with a second i-C₄H₈ molecule to form a C₆H₅CH₃·⁺-(i-C₄H₈)₂ complex, in which the olefin molecules are activated by the aromatic ion. Similar reactions are observed in the benzene/propene system with a somewhat larger ΔIP of 0.48 eV, suggesting that the charge density on the olefin in the complex is still sufficient to activate it for nucleophilic attack. However, aromatic/olefin systems with ΔIP > 0.87 eV show no olefin dimer formation. At low [i-C₄H₈] and [Ar] number densities, the rate of formation of C₈H₁₆(·)⁺ is proportional to [i-C₄H₈]²[Ar]. The corresponding fourth-order rate coefficient shows a strong negative temperature coefficient with k = 11 x 10^-42 cm⁹ s^-1 at 300 K and ₂ x 10^-42 cm⁹ s^-1 at 346 K, suggesting that the mechanism can be efficient in low-temperature industrial and interstellar environments. The direct formation of the dimer bypasses the monomer olefin cation and its consequent side-reactions, and directs the products selectively into radical ion polymerization. The products and energy relationships that apply in the gas phase are observed also in clusters.

Full text of DOI:10.1021/ja962635x

8332
J. Am. Chem. Soc. 1997, 119, 8332-8341
Coupled Reactions of Condensation and Charge Transfer.
1. Formation of Olefin Dimer Ions in Reactions with Ionized
Aromatics. Gas-Phase Studies
Michael Meot-Ner (Mautner),* Yezdi B. Pithawalla, Junling Gao, and
M. Samy El-Shall*
Contribution from the Department of Chemistry, Virginia Commonwealth UniVersity,
Richmond, Virginia 23284-2006
ReceiVed July 31, 1996X
Abstract: The toluene radical ion C6H5CH3•+, generated by resonance two-photon ionization, does not react with a
single isobutene molecule (i-C4H8) which has a significantly higher ionization potential (∆IP ) 0.42 eV). However,
a reaction is observed involving two i-C4H8 molecules, to form the dimer ion C8H16•+. A coupled reaction of dimer
formation and charge transfer to the dimer is exothermic if the product is an ionized hexene with a low IP.
Correspondingly, the observed nominal second-order rate coefficients, (5-25) × 10-12 cm3 s-1, are enhanced by a
factor of >105 over the expected value for direct endothermic charge transfer. Pressure and concentration effects
suggest a sequential mechanism that proceeds through a C6H5CH3•+(i-C4H8) reactive π complex. The complex can
•+
isomerize to a nonreactive CH3C6H4-t-C4H9•+ adduct, or react with a second i-C4H8 molecule to form a C6H5CH3
-
(i-C4H8)2 complex, in which the olefin molecules are activated by the aromatic ion. Similar reactions are observed
in the benzene/propene system with a somewhat larger ∆IP of 0.48 eV, suggesting that the charge density on the
olefin in the complex is still sufficient to activate it for nucleophilic attack. However, aromatic/olefin systems with
∆IP > 0.87 eV show no olefin dimer formation. At low [i-C4H8] and [Ar] number densities, the rate of formation
of C8H16•+ is proportional to [i-C4H8]2[Ar]. The corresponding fourth-order rate coefficient shows a strong negative
temperature coefficient with k ) 11 × 10-42 cm9 s-1 at 300 K and 2 × 10-42 cm9 s-1 at 346 K, suggesting that the
mechanism can be efficient in low-temperature industrial and interstellar environments. The direct formation of the
dimer bypasses the monomer olefin cation and its consequent side-reactions, and directs the products selectively
into radical ion polymerization. The products and energy relationships that apply in the gas phase are observed also
in clusters.
Introduction
or impact by metal atoms or ions that were reported recently.6-8
The initiation mechanisms can be best isolated and studied in
the gas phase or in clusters,9-21 and the results then applied to
the condensed phase. For example, our recent studies of the
reactions of metal ions with isobutene have lead to a novel
Ionic polymerization of olefins B can be initiated by
ionization of the olefin by various mechanisms, including charge
transfer from another radical cation A•+ to form B•+. In the
gas phase, these reactions are efficient when the ionization
potential (IP) of the olefin is comparable to or lower than that
of A, and the reaction is exothermic.
In this paper, we examine reaction systems where the opposite
is true, i.e., the ionization potential of the olefin is substantially
higher than that of the aromatic initiator. This rules out direct
charge transfer from A•+ to B. However, under favorable
conditions, a reaction with the overall stoichiometry of reaction
1 can be then energetically favorable, if the product is the ion
of a larger olefin whose IP is lower than that of A.
(4) Taylor, R. B.; Williams, F. J. Am. Chem. Soc. 1969, 91, 3728.
(5) Okamato, H.; Fueki, K.; Kuri, Z. J. Phys. Chem. 1967, 71, 3222.
(6) Vann, W.; Daly, G. M.; El-Shall, M. S. In Laser Ablation in Materials
Processing: Fundamentals and Applications; Braren, B., Dubowski, J.,
Norton, D., Eds.; MRS Symposium Proceedings Series; 1993; Vol. 285, p
593. Vann, W.; El-Shall, M. S. J. Am. Chem. Soc. 1993, 115, 4385.
(7) Deakin, L.; Den Auwer, C.; Revol, J. F.; Andrews, M. P. J. Am.
Chem. Soc. 1995, 117, 9916.
(8) El-Shall, M. S.; Slack, W. Macromolecules 1995, 28, 8546.
(9) El-Shall, M. S.; Marks, C. J. Phys. Chem. 1991, 95, 4932.
(10) El-Shall, M. S.; Schriver, K. E. J. Chem. Phys. 1991, 95, 3001.
(11) Coolbaugh, M. T.; Vaidyanathan, G.; Peifer, W. R.; Garvey, J. F.
J. Phys. Chem. 1991, 95, 8337.
A
•+ + 2B f B2•+ + A
(1)
(12) Coolbaugh, M. T.; Whitney, S. G.; Vaidyanathan, G.; Garvey, J. F.
J. Phys. Chem. 1992, 96, 9139.
(13) Castelman, A. W., Jr.; Guo, B. C. J. Am. Chem. Soc. 1992, 114,
6152.
(14) Brodbelt, J. S.; Liou, C. C.; Maleknia, S.; Lin, T. J.; Lagow, R. J.
J. Am. Chem. Soc. 1993, 115, 11069.
(15) Daly, G. M.; El-Shall, M. S. Z. Phys. D 1993, 26, S186.
(16) Daly, G. M.; El-Shall, M. S. J. Phys. Chem. 1994, 98, 696.
(17) Daly, G. M.; El-Shall, M. S. J. Phys. Chem. 1995, 99, 5283.
(18) Desai, S. R.; Feigerle, C. S.; Miller J. J. Phys. Chem. 1995, 99,
1786.
(19) Daly, G. M.; Pithawalla, Y. B.; Yu, Z.; El-Shall, M. S. Chem. Phys.
Lett. 1995, 237, 97.
Since three-body collisions are improbable, reaction 1 requires
a sequential mechanism, as will be discussed below. This
mechanism directs the products selectively into the radical cation
polymerization channel, and avoids side reactions that would
result from reactions of the monomer olefin ion.
Reaction 1 constitutes a specific initiation mechanism, in
addition to catalysis,1-3 ionizing radiation4,5 electrode processes,1-3
X Abstract published in AdVance ACS Abstracts, August 15, 1997.
(1) Kennedy, J. P.; Marechal, E. Carbocationic Polymerization; Wiley:
John and Sons: New York, 1982.
(20) Meot-Ner (Mautner), M.; Sieck, L. W.; El-Shall, M. S.; Daly, G.
M. J. Am. Chem. Soc. 1995, 117, 7737.
(2) Goethals, E. S., Ed. Cationic Polymerization and Related Processes;
Academic Press: New York, 1984.
(3) Ystenes, M. J. Catalysis 1991, 129, 383.
(21) El-Shall, M. S.; Daly, G. M.; Yu, Z.; Meot-Ner (Mautner), M. J.
Am. Chem. Soc. 1995, 117, 7744. El-Shall, M. S. Polym. Prepr. 1996, 37,
367.
S0002-7863(96)02635-2 CCC: $14.00 © 1997 American Chemical Society

Coupled Reactions of Condensation and Charge Transfer
J. Am. Chem. Soc., Vol. 119, No. 35, 1997 8333
beam enters and exits. The source is placed inside a vacuum chamber
which is also equipped with windows. The cell temperature is
monitored through two type T copper-constantan thermocouples
(Omega). Gas mixtures are prepared in a heated (>100 °C) 2 L flask
and admitted to the ion source at selected pressures via an adjustable
needle valve. The cell pressure is monitored with a 0.01-10 Torr
capacitance manometer (MKS, 1301) coupled with the gas inlet tube.
Mixtures are typically made by microliter injection of liquid samples
into the evacuated heated sample flask followed by the addition of the
bath gas (Ar in the present study).
technique for bulk polymerization by the impact of metal atoms
and ions on the liquid monomer.6,8
The present olefin condensation process is initiated by charge
transfer from an aromatic center which is ionized selectively in
a complex mixture, using resonant two-photon ionization22 high-
pressure mass spectrometry (R2PI-HPMS).23 The ionized
aromatic center can bind to an olefin molecule and induce on
it a positive charge density sufficient for nucleophilic attack by
a second olefin molecule. These reactions belong to the so far
small class of higher-order ion-molecule reactions where an
ion associates with a molecule and activates it for reaction with
a further molecule. Examples are the reactions of O3, N2O5,
and NO2 adsorbed on alkali metal cations,24,25 ion-assisted
reactions of HCl with ClONO2,26 and reactions of silicon atoms
clustered to naphthalene+.27 In particular, the present reactions
are similar to the dimerization of the fluorinated olefin C2F4 to
form C4F8 through association with and then elimination of
In the isobutene/toluene system, the R2PI of toluene was obtained
via the 00 transition at λ ) 266.76 nm. We also used two-photon
0
ionization at λ ) 258.94 nm, and the results were similar to those
obtained using the 000 resonance ionization. These photons create C6H5-
CH3•+ ions with excess energy of 0.48 or 0.74 eV, respectively, above
the IP, much lower than the excess energy required for ring opening
in ionized benzene, 3.5-5.0 eV.31-33
The laser beam is slightly focused within the center of the cell using
a quartz spherical lens (f ) 60 cm, d ) 2.54 cm). The laser output (λ
) 266.76 nm, 100-300 µJ, ∆t ) 15 ns, 10 Hz repetition rate) is
generated by an excimer (XeCl) pumped dye laser (Lambda Physik
LPX 101 and FL-3002, respectively). Coumarin 540A (Exciton) dye
laser output passes through a â-BaB2O4 crystal (CSK) cut at 52° to
generate tunable frequency-doubled output of 10-8 s pulses. The
spatially filtered ultraviolet radiation passes through the high-pressure
cell, and focusing is adjusted to minimize three photon processes (i.e.,
unimolecular fragmentation) while still providing sufficient ion current
(photon power density ∼105 W/cm2). The reactant and product ions
escape through a precision pinhole (200 µm, Melles Griot) and are
analyzed with a quadrupole mass filter.
The quadrupole mass filter (Extrel C-50, equipped with 1.9 cm
diameter rods having a resolution better than 1 amu, FWHM, in the
mass range of 1-500 amu) is mounted coaxially to the ion exit hole.
The distance from the ion exit hole to the C50 lens stack is
approximately 2 cm. The ion current from the electron multiplier is
amplified and then recorded with a 350 MHz digital oscilloscope
(LeCroy 9450).
Ion signal intensities of each ion were integrated for 40-80 s to
obtain sufficient signal intensity. The main source of error in the
measurements was possible drift in the signal intensities while all the
ions were recorded. To check and minimize this effect, each ion
intensity was recorded in ascending mass order and then recorded again
proceeding in reverse order. The replicate measurements were
compared, and the experiment was accepted only if intensities in the
replicate measurements for all ions differed by <15%. In the acceptable
experiments, the signal intensities from the replicate measurements were
summed and averaged. When measured in this manner, relative ion
intensities obtained in 4-6 replicate experiments were reproducible
within (20%, and rate coefficients (see below) were reproducible within
(30%.
CF3 ,
+ 28 but in the present case, the ion also serves as a charge
donor. There is also a basic analogy with reactions of a
hydrocarbon ion and two H2O molecules29 or an ionized
aromatic and two polar molecules23 to form protonated dimers,
where neutrals attached to the ion react with an additional
molecule to extract a proton from the hydrocarbon ion. In these
systems proton transfer to one polar molecule would be
endothermic and is not observed, but reactions with two
molecules are driven energetically by the formation of a strong
hydrogen bond of 30 kcal/mol to form a protonated dimer. The
present system is basically similar, but it results in covalent,
rather than hydrogen, bond formation. In the present system
the aromatic core ion serves as both an activator and a charge
donor, similar to an anode in electrochemical polymerization.30
The present system constitutes an extension of our studies
of polymerization in isobutene.20,21 In our first studies, polym-
erization was initiated by a full charge on the i-C4H8•+ reactant,
or by partial charge transfer to the olefin in the [C6H6‚i-C4H8]•+
complex. The ionization potentials (IPs) of the components in
this complex are similar within 0.1 eV, and therefore a charge
density of about 0.5 may be located on the olefin, which is
apparently still sufficient for activating the olefin.
A basic question arises as to how much charge density on a
molecule is still sufficient to activate it to undergo ionic type
process. The charge density can be varied in aromatic-olefin
systems with various differential IPs of the components, leading
to various degrees of charge distribution between the reactants.
In this paper we shall investigate several systems with varying
IP differences.
Results
Experimental Section
Reaction System. Time-resolved ion profiles are illustrated
in Figure 1a, and the normalized intensities are shown in Figure
1b. All ion profile measurements were replicated 2-6 times
and yielded relative ion intensities reproducible within (15%.
The main process is the decay of the reactant C6H5CH3•+ (i.e.,
The application of R2PI-HPMS has been described in detail
elsewhere.23 Briefly, the HPMS ion source is a cubic aluminum block
of about 2 cm3, fitted with quartz windows through which the laser
(22) Lubman, D. M.; Li, L.; Hager, J. W.; Wallace, S. C. In “Lasers
and Mass Spectrometry; Lubman, D. M., Ed.; Oxford University Press:
Oxford, 1990.
(23) Daly, G. M.; Meot-Ner (Mautner), M.; Pithawalla, Y. B.; El-Shall,
M. S. J. Chem. Phys. 1996, 104, 7965.
(24) Rowe, B. R.; Viggiano, A. A.; Fehsenfeld, F. C.; Fahey, D. W.;
Ferguson, E. E. J. Chem. Phys. 1982, 76, 742. Viggiano, A. A.; Deakyne,
C. A.; Dale, F.; Paulson, J. F. J. Chem. Phys. 1987, 87, 6544.
(25) Viggiano, A. A.; Deakyne, C. A.; Dale, F.; Paulson, J. F. J. Chem.
Phys. 1987, 87, 6544.
(26) Van Doren, J. M.; Viggiano, A. A.; Morris, R. A. J. Am. Chem.
Soc. 1994, 116, 6957.
T•+) ion and the formation of the dimer ion of m/z 112, i.e.,
•+
C8H16•+, which can also be denoted as i-C4H8•+(i-C4H8) or I2
.
In parallel, the adduct at m/z 148 is also formed. This adduct
can be a noncovalent π complex denoted as C6H5CH3•+(i-C4H8)
or T•+(I), or a covalent adduct tert-butyltoluene•+, i.e., CH3C6H4-
t-C4H9•+, denoted as TI•+. This is the most stable adduct isomer,
and the observed, nonreactive m/z 148 ion (and its adducts with
i-C4H8 molecules) probably corresponds to this covalent adduct.
(27) Bohme, D. K. Chem. ReV. 1992, 92, 1478.
(31) Jones, E. G.; Bhattacharya, A. K.; Tiernan, T. O. Int. J. Mass
Spectrom. Ion Phys. 1975, 14, 147.
(32) Rosenstock, H. M.; Larkins, J. T.; Walker, J. A. Int. J. Mass
Spectrom. Ion Phys. 1973, 11, 309.
(33) Beynon, J. A.; Hopkinson, J. A.; Lester, G. R. Int. J. Mass Spectrom.
Ion Phys. 1969, 2, 291.
(28) Morris, R. A.; Viggiano, A. A.; Paulson, J. F. J. Phys. Chem. 1993,
97, 6208.
(29) Sieck, L. W.; Searles, S. K. J. Chem. Phys. 1970, 53, 2601.
(30) Bhadani, S. N.; Parravano, G. In Organic Electrochemistry; Beizer,
M. M., and Lund, H., Eds.; Marcel Dekker: New York, 1983, p 995.

8334 J. Am. Chem. Soc., Vol. 119, No. 35, 1997
Meot-Ner et al.
+ I3•+ + ... and ∑TI•+ ) TI•+ + TI2•+ + ..., to clearly observe
the parallel formation of the primary products.
In several experiments, the ∑TI•+ product group continued
•+
to increase while the ∑I2 product group leveled off after the
decay of about 70% of the T•+ reactant ion, as observed in
Figure 1c after about 2.4 ms. This could indicate reactions into
the two product channels from two different isomers of T•+.
However, the low energy photoionization should not form
isomers, as noted above. Also, we observe below that the
product distributions into the two channels vary with the
concentration of isobutene, which cannot affect the photopro-
duction of various T•+ isomers.
At high laser fluences we observed toluene fragment ions at
m/z 65+ and 91+ and the tert-butyl cation (m/z 57), and their
adducts with i-C4H8 molecules. There are no plausible cross-
reactions from these even-electron ions to the radical ions of
interest, and in fact their summed normalized intensities remain
constant with reaction time. Most studies were performed at
laser fluences where these ions were negligible, and they are
not considered further.
The normalized ion intensities are used to calculate rate
coefficients as follows. The pseudo-first-order rate coefficient
for the overall reaction T•+ to products is calculated from the
decay of the reactant ion:
k ) -d ln [T•+]/dt
(2)
The corresponding nominal second-order rate coefficient for
the overall forward reaction, kf, is calculated using the number
density of the reactant I as
kf ) k/[I]
(3)
The nominal second-order rate coefficients for the two
channels are calculated from the product distributions as
•+
kf(I2•+) ) k [(
I
•+)/(
I
2
+
TI•+)]
TI•+)]
(4)
(5)
∑
∑
∑
f
2
•+
k (TI•+) ) k [( TI•+)/(
I
2
+
∑
∑
∑
f
f
Rate coefficient measurements were replicated 2-6 times and
were reproducible within (30%. The results are presented in
Table 1.
Kinetic Observations. The principal kinetic observations
can be summarized as follows.
•+
(1) The overall reaction leading from T•+ to I2 proceeds
orders of magnitude faster than expected for direct endothermic
charge transfer (∆H° ) +0.42 eV) from T•+ to produce the
monomer ion I•+ (observed nominal second-order rate coef-
ficients k ) (5-25) × 10-12 cm3 s-1 vs expected kf e kcollision
exp(-∆H°/RT) ≈ 10-17 cm-3 s-1). Analogous and somewhat
larger rate enhancement is observed in the benzene/propene
system for propene2•+ formation (∆H° ) +0.48 eV, k(obsd) )
(1-5) × 10-12 cm3 s-1, k(expected) e 10-18 cm3 s-1). In fact,
the monomer olefin ions isobutene•+ and propene•+ are not
obserVed, and the dimer ions appear to form directly through
the reactions of the aromatic ions.
Figure 1. (a) Ion time profiles in the toluene (T)/isobutene (I) system
with P(C6H5CH3) ) 0.0052 Torr (N ) 1.7 × 1013 cm-3), P(i-C4H8) )
0.00083 Torr (N ) 2.7 × 1013 cm-3), and P(Ar) ) 0.71 Torr (N ) 2.3
•+
× 1016 cm-3) at 298 K. Note the parallel formation of I2 (m/z 112)
•+
and TI•+ (m/z 148) followed by C8H14 (m/z 110) through a H2 loss
reaction (see ref 20). (b) Normalized intensities of the primary ions.
(c) Normalized intensities, with consecutive products from the primary
•+
•+
ions [TI•+ + TI2•+] and [I2 + C8H14 + I3•+] summed to show the
distribution into the primary channels.
Significantly, the formation of the monomer ion i-C4H8•+ is
(2) The adduct TI•+ and the dimer I2•+ are formed in parallel
in comparable yields (Figure 1c), although the third-body [Ar]
or the efficient third-body [C3H8], which could collisionally
•+
not observed. At later reaction times the C8H16 ion reacts
•+
with i-C4H8 to yield the unreactive ion C8H14 through H2
transfer.20 At long reaction times and high i-C4H8 concentra-
tions, small intensities of higher adducts of i-C4H8 molecules
to the product ions are also observed. In Figure 1c, the
stabilize an excited complex to produce TI•+, are in excess by
•+
factors of 200-1000 over the reactant [I] that produces I2
.
(3) The overall nominal second-order rate coefficient kf
increases with [Ar] (Figure 2), while the product ratio I2•+/TI•+
is independent of [Ar] (Figure 3a). As a result, the nominal
•+
intensities of the two products I2 and TI•+ are summed with
•+
•+
•+
their respective further products, i.e., ∑I2 ) I2 + C8H14

Coupled Reactions of Condensation and Charge Transfer
J. Am. Chem. Soc., Vol. 119, No. 35, 1997 8335
Table 1. Thermochemistrya and Kineticsb for the Formation of Olefin Dimer Ions B2+, Covalent Adducts AB+,a and Neutral Olefin Dimers
B2 in the Reactions of Aromatic Ions A+ with Olefin B
+ b
d
B2
AB+ c
-∆H° a
B2
a
a
a
a
A+
B
∆IPa
-∆H° a
k2
k3
k2
k3
-∆H° a
+
C6H6
i-C4H8
i-C4H8
0.0
0.30
26.2
re
33.7
17.0
17.0
+
C6H5CH3
16.3f
13f
5.6f
33.5f
15f
6.6f
2.0g
1.0g
4.8g
2.5g
+
C6H6
C3H6
i-C4H8
C3H6
C2H4
0.48
0.80
0.91
1.27
48.7h
7.6
38.8
31.2
2.5h
36.0h
0.81h
26.2
17.0
26.2
27.9
+
1,4-C6H4(CH3)2
nri
+
C6H5CH3
nri
nri
36.0
36.2
+
C6H6
a Units: ∆IP, eV; ∆H°, kcal/mol; k2, 10-12 cm3 s-1; k3, 10-28 cm6 s-1. Values of k3 calculated from k2 •)+ k3[Ar]. Error estimates for rate
coefficients, based on replicate measurements, (30%. b Thermochemistry for the reactions A•+ + 2B f B2 + A, with the formation of (E)-
•+
CH3CHCHCH3•+, CH3CHCH2CH2CHCH3•+, and (E)-(CH3)2CHCHCHCH(CH3)2 as plausible condensation products for the ethylene, propene,
and isobutene dimer cations, respectively.39 c Thermochemistry for the formation of covalent adducts AB•+ (benzene/isobutene, toluene/isobutene,
benzene/propene, toluene/propene, and benzene/ethylene) is calculated for the products C6H5-t-C4H9•+, 1,4-CH3C6H4(t-C4H9)•+, C6H5-i-C3H7•+, 1,4-
CH3C6H4(i-C3H7)•+, and C6H5C2H5•+, respectively.39 d Thermochemistry for the reactions A•+ + 2B f A•+ + B2, with formation of the neutral
analogues of the isomers in the footnote b. e Reaction suggested by kinetic simulations (ref 20). f Nominal rate coefficients obtained in a reaction
system of P(C6H5CH3) ) 0.0010 Torr (N ) 3.2 × 1013 cm-3), P(i-C4H8) ) 0.0016 Torr (N ) 5.2 × 1013 cm-3), and P(Ar) ) 0.69 Torr (N ) 2.2
•+
× 1016 cm-3) at 300 K. The overall nominal forward rate coefficient kf2 was 28 × 10-12, and the I2•+/TI•+ product ratio was 0.86. For the I2
channel, the fourth-order rate coefficient of k3/[I] ) 11 × 10-42 cm9 s-1 applies. g Nominal rate coefficients obtained in a reaction system of
P(C6H5CH3) ) 0.0010 Torr (N ) 2.8 × 1013 cm-3), P(i-C4H8) ) 0.0018 Torr (N ) 5.1 × 1013 cm-3), and P(Ar) ) 0.69 Torr (N ) 1.9 × 1016 cm-3
)
at 346 K. The nominal second-order rate coefficient was 6.8 × 10-12 cm3 s-1, and the product ratio I2•+/TI•+ was 0.42. For the I2 channel, the
fourth-order rate coefficient of k3/[I] ) 2.0 × 10-42 cm9 s-1 applies. The measurements were done at about constant [Ar] and [I], and the third-
order and fourth-order rate coefficients will give similar temperature coefficients. The temperature study was replicated at isobutene number
densities of 1.8 × 1013 and 6.9 × 1013 cm-3, and gave similar temperature coefficients. h Nominal rate coefficients obtained in a reaction system
•+
of P(C6H6) ) 0.00068 Torr (N ) 2.2 × 1013 cm-3), P(C3H6) ) 0.01 Torr (N ) 3.2 × 1014 cm-3), and P(Ar) ) 0.80 Torr (N ) 2.6 × 1016 cm-3
)
at 300 K. i nr denotes non-reactive systems, with k2 < 10-13cm3 s-1
.
Figure 2. Nominal second-order rate coefficients (10-12 cm3 s-1) as a function of third-body [Ar] number density. Data measured at a constant
i-C4H8 partial pressure of 0.00084 ( 0.00003 Torr (N ) (27 ( 1) × 1012 cm-3) at 298 K. Solid lines are fitted through experimental points (solid
squares), broken lines from kinetic simulation of Scheme 1 (open circles), and dotted lines from kinetic simulation of mechanism 2 (open triangles).
second-order rate coefficient kf(I2•+) for the I2•+ channel and kf
(TI•+) for the TI•+ channel both increase with [Ar] (Figure 2).
The same trends are observed in substituting [Ar] by the efficient
polyatomic third-body [C3H8] (see Appendix 2). This is unusual
in competitiVe association/transfer kinetics where the association
product usually increases relatiVe to the transfer product with
third-body pressure and efficiency.
(7) With decreasing temperature, the rate of the reaction into
both channels increases sharply, and the I2•+/TI•+ product ratio
increases (Table 1).
(8) In the reactive systems, the rate coefficients vary inversely
with ∆IP (Table 1). Olefin dimer ions are formed in reaction
systems with ∆IP e 0.42 eV, and not formed for ∆IP g 0.87
eV between the components. Decreasing dimer formation with
increasing ∆IP is observed also in preformed clusters.34
(4) The product ratio I2•+/TI•+ increases with [I] (Figure 3b).
The nominal second-order rate coefficient kf(I2•+) for the
formation of I2•+ also increases with [I], showing a kinetic order
intermediate between [I] and [I]2 dependence (Figure 4).
Discussion
Reaction Mechanism. In this section we shall show that
the kinetic features are consistent with a mechanism that
involves an additional complex along the reaction pathway,
subsequent to the usual excited complex. Kinetic considerations
•+
(5) The products I2 (including its higher adducts and
C8H14•+) and TI•+ (including its higher adducts) do not
interconvert, even at high isobutene concentrations (Figure 5).
(6) Toluene concentration has no significant effect on the
rate coefficients and product distributions (Figure 6).
(34) El-Shall, M. S.; Yu, Z. J. Am. Chem. Soc. 1996, 51, 13058.

8336 J. Am. Chem. Soc., Vol. 119, No. 35, 1997
Meot-Ner et al.
Figure 3. Product distribution ratio [I2•+]/[TI•+] (a) as a function of
third-body number density, measured under the conditions given in
the Figure 2 caption and (b) as a function of i-C4H8 number density.
Data measured at constant total pressure of 0.80 Torr (N ) 2.6 × 1016
cm-3) at 298 K. Solid lines are fitted through experimental points (solid
squares), broken lines from kinetic simulation of Scheme 1 (open
circles), and dotted lines from kinetic simulation of mechanism 2 (open
triangles).
Figure 4. Nominal second-order rate coefficients (10-12 cm3 s-1) as a
function of i-C4H8 number density. Data measured at a constant source
pressure of 0.80 Torr (N ) 2.6 × 1016 cm-3). (a) Solid lines are fitted
through experimental points (solid squares) and broken lines from
kinetic simulation of Scheme 1 (open circles). (b) Solid lines are fitted
through experimental points and dotted lines from kinetic simulations
of mechanism 2 (open triangles). (See Appendix 1b.)
concerning more simple alternative mechanisms are presented
below and in Appendix 1. The proposed mechanism is
summarized in Scheme 1(mechanism 1) (reactions 6-9).
Scheme 1
Note that the mechanism involves three different adducts
•+
between C6H5CH3 and i-C4H8. The existence of these
complexes is inferred from the kinetic observations; of course,
they all correspond to m/z 148, and are indistinguishable mass
spectrometrically. In Scheme 1, the excited complex [C6H5-
CH3•+(C4H8)]*, i.e., [T•+(I)]*, is formed in reaction 6, and it
Figure 5. Normalized ion time profiles in the toluene/isobutene system
at 298 K, at high i-C4H8 concentration, showing the absence of
interconversion between the products In and TIn at high isobutene
concentration. P(C6H5CH3) ) 0.00062 Torr (N ) 2.0 × 1013 cm-3),
P(i-C4H8) ) 0.023 Torr (N ) 7.5 × 1014 cm-3), and P(Ar) ) 1.0 Torr
(N ) 3.2 × 1016 cm-3) at 298 K.
•+
undergoes stabilization to the noncovalent complex C6H5CH3
-
•+
•+
(C4H8), i.e., T•+(I), by collisional or radiative stabilization
(reaction 7). The complex then isomerizes to the final,
presumably covalent adduct CH3C6H4-t-C4H9•+, i.e., TI•+ (reac-
tion 9). In a competitive process, the stabilized noncovalent
T•+(I) adduct also reacts with I to form I2•+, presumably through
an intermediate complex C6H5CH3•+(C4H8)2 (reaction 8).
Assuming that the formation of the stabilized noncovalent
T•+(I) is rate-controlling for the overall forward reaction, we
obtain
by the small overall efficiency, then
kf ) k6(k7[M] + k7r)/k-6
(11a)
With respect to the contributions of collisional and radiative
stabilization, i.e., k7[M] vs k7r, we note that kf is proportional
to the third-body density as observed in Figure 2, suggesting
that collisional stabilization is dominant under high-pressure
conditions. We note however that the intercept of kf vs [Ar] is
not at the origin. Therefore, radiative stabilization may contribute
kf ) k6(k7[M] + k7r)/(k-6 + k7[M] + k7r)
(10)
Here k7[M] is the rate of collisional stabilization and k7r is the
rate of radiative stabilization. If k-6 . k7[M] + k7r, as indicated

Coupled Reactions of Condensation and Charge Transfer
J. Am. Chem. Soc., Vol. 119, No. 35, 1997 8337
consistent with the observation in Figure 3b. Note that at high
[I], the relation becomes nonlinear (Figure 3b), possibly due to
excess TI•+ production by direct collisions of [T•+(I)]* with I.
The i-C4H8 molecule can serve as an efficient third body that
could stabilize and isomerize the complex directly to the
covalent adduct TI•+, bypassing the noncovalent adduct T•+(I)
(note that reaction 8 requires this adduct).
From Figures 2 and 4, it is evident that the nominal second-
order rate coefficients kf, kf(I2•+), and kf(TI•+) are actually
composite entities of a complex kinetic order with respect to
[I] and [Ar]. Note that kf is a second-order rate coefficient
obtained using eq 3, i.e., kf ) k/[I]. Figure 2 shows that at low
third-body density kf is also proportional to [Ar], and the pseudo-
first-order overall rate coefficient is therefore proportional to
[I][Ar]; i.e., the overall reaction is third-order. Third-order rate
coefficients are reported in Table 1.
Figure 4 shows that, at low [I], the nominal second-order
rate coefficient kf(I2•+) calculated from eqs 2-4 is proportional
to [I], and therefore the pseudo-first-order rate of the reaction
to produce I2•+ is proportional to [I]2. From Figure 2, it is also
proportional to [Ar] at low third-body densities. Therefore, at
low [I] and [Ar], the pseudo-first-order rate varies as [I]2[Ar]
•+
and the rate coefficients for the formation of I2 are fourth-
order with values of 11 × 10-42 cm9 s-1 at 300 K and 2 ×
10-42 cm9 s-1 at 346 K.
To further test the mechanism, we simulated Scheme 1 using
the ACUCHEM program for coupled reactions,35 using rate
coefficients as explained in Appendix 2. The model should be
able to reproduce the effects of varying [I] and [Ar] as shown
in Figures 2-4.
Figures 2-4 show that the simulations reproduce the observed
trends, at least qualitatively. Specifically, Figure 2 shows that
the overall rate coefficient kf increases with [Ar]. Referring to
Scheme 1, this is due to increasing competition of the collisional
stabilization of [T•+(I)]* with increasing [Ar], vs back-dissocia-
tion. Correspondingly, the rate coefficients for the two products,
kf(I2•+) and kf(TI•+), also increase in parallel and maintain a
constant product ratio. In both the experiment and simulations,
kf tends to level off at high [Ar], as its value would ultimately
have to become constant when reaching the collision rate.
Figure 3a shows that the product ratio remains constant with
[Ar], as k8[I] and k9 are independent of [Ar]. Figure 3b shows
that the product ratio increases with [I], as k8[I]/k9 increases
with [I]. We note a significant deviation toward increased TI•+
vs I2•+ production at high [I]. A possible reason could be that
[T•+(I)]* may be isomerized directly to the nonreactive covalent
adduct TI•+ through collision with the efficient third-body I.
Figure 4 shows that the overall rate coefficient kf increases
slightly with [I]. This results from the more efficient competi-
tion of the forward reaction channels of T•+(I) compared with
back-dissociation through (T•+I)* as k8[I] increases vs k-7. We
also observe the increase of kf(I2•+) with increasing [I]. Note
that at the same time kf(TI•+) decreases, as with increasing k8-
[I] vs constant k9, the dimer channel competes more effectively.
Also note that the simulation reproduces the nonlinear variation
of these partial rate coefficients with [I].
Figure 6. Normalized ion time profiles in the toluene/isobutene system
at low and high C6H5CH3 concentration, demonstrating the absence of
significant effects on the ion time profiles. (a) P(C6H5CH3) ) 0.00015
Torr (N ) 4.9 × 1012 cm-3), P(i-C4H8) ) 0.00035 Torr (N ) 1.1 ×
1013 cm-3), P(Ar) ) 0.72 Torr (N ) 2.3 × 1016 cm-3). (b) P(C6H5-
CH3) ) 0.0056 Torr (N ) 182.7 × 1012 cm-3), P(i-C4H8) ) 0.00068
Torr (N ) 22 × 1012 cm-3), P(Ar) ) 0.72 Torr (N ) 2.4 × 1016 cm-3
)
at 298 K. In the latter system, (C6H5CH3)2•+ is also present (about 20%
•+
of the C6H5CH3 signal intensity), which comes to equilibrium with
•+
C6H5CH3
.
significantly at low pressures, as may be expected for these large
species with many vibrational modes. Substituting an intercept
of kf ) 10-12 cm3 s-1 at [M] ) 0, and using the other rate
coefficients as described in Appendix 2a, yields k7r ) 1.4 ×
104 s-1, about an order of magnitude smaller than the rate of
collisions with the third body. However, radiative stabilization
may become dominant in low-pressure experiments such as ion
cyclotron resonance at kcollision[M] < 103 s-1, and such experi-
ments would be of interest.
Under high-pressure conditions where k7[M] > k7r, eq 11a
reduces to eq 11b, and kf increases with third-body density [M],
as is observed in Figure 2.
kf ) k6k7[M]/k-6
(11b)
Another feature of the simulations is that, with the rate
coefficients used, the concentrations of the reactive, noncovalent
[T•+(I)]* and T•+(I) adducts always remain 2-4 orders of
magnitude below the sum of the other species, consistent with
these species being transient intermediates.
Assuming that the rearrangement of the stabilized noncovalent
intermediate T•+(I) to the final covalent adduct TI•+ (reaction
9) is a unimolecular reaction at the high-pressure limit (i.e.,
pseudo-first-order), then the product ratio will be given by eq
12. The product ratio is then independent of [M], consistent
We observe a decrease in the overall rate coefficient with
increasing temperature. This can be attributed to the increasing
I2•+/TI•+ ) k8[I]/k9
with the observation in Figure 3a, and increases with [I],
(12)
(35) ACUCHEM Version 1.4, copyright National Institute of Standards
and Technology, Nov 7, 1986.

8338 J. Am. Chem. Soc., Vol. 119, No. 35, 1997
Meot-Ner et al.
rate coefficient k-6 for the back-dissociation exciting the
complex [T•+(I)]*, similar to the effect that leads to negative
temperature coefficients in ion-molecule association reactions.
However, reactions 8 and 9 may involve barriers, and k8 and k9
may increase with temperature, leading to complex overall
temperature coefficients.
complex similar to aromatic dimer ions that we investigated
previously, which for the present IP difference of 0.42 eV
typically have binding energies of 12-15 kcal/mol.36,37 Alter-
natively, the complex may be a covalent distonic radical cation
similar to those observed by Holman et al. in the reactions of
C6H6•+ with cycloalkanes.38 These species have higher energies
than the π-ionized alkylbenzene cations, and may serve a role
similar to that of a noncovalent complex.
Further support for reaction 8 of a noncovalent adduct is
obtained by observations in preformed aromatic/olefin clusters
in which the aromatic molecule was selectively photoio-
inized.21,34 The results show close similarities between the
cluster and gas phase studies, as summarized below.
(1) In benzene (B)/isobutene, where the IPs are similar, in
both studies I2•+, BI•+, and their higher adducts with I are
observed. Also observed are I•+ and its reaction product with
I, i.e., t-C4H9+. The results indicate that charge transfer from
benzene•+ to i-C4H8 occurs in both the gas phase and in
clusters.20,21
•+
For the formation of the nonreactive 1,4-CH3C6H4-t-C4H9
covalent adduct, the thermochemistry gives ∆H° ) -33.5 kcal/
mol.39 In relation to its possible back-dissociation to C6H5-
CH3•+ + i-C4H8, the observations would yield equilibrium ion
ratios of [TI•+]/[T•+] g 100. In the model reaction system used
for Figure 1 this leads to -∆G° (300) g 10.5 kcal/mol, and a
typical association entropy change of ∆S° ) -35 cal/(mol K)
gives -∆H° g 21 kcal/mol, consistent with a covalent adduct.
•+
We note that the two distinct isomers, a reactive C6H5CH3
-
•+
•+
(2) In toluene/isobutene, I2 and TI•+ are observed. How-
(i-C4H8) complex and the unreactive covalent CH3C6H4-t-C4H9
ever, I•+ and its t-C4H9•+ product are not observed in either the
adduct, suggests a barrier between them. This barrier is below
the energy of the reactants, since the temperature coefficient
for the formation of CH3C6H4-t-C4H9•+ is negative. The barrier
for the reaction of C6H5CH3•+(i-C4H8) with i-C4H8 to form
•+
gas phase or clusters, indicating that I2 is formed directly in
both cases. In all of the reactions observed, the monomer olefin
ion and its product, the protonated olefin monomer, are not
observed in either phase where the IP of the aromatic is lower
than that of the olefin.
•+
C8H16 must still be lower, possibly negligible, as indicated
by the fact that the temperature coefficient for the formation of
•+
C8H16 is even more negative.
(3) In p-xylene/isobutene, no I2•+ is observed in the gas phase.
Other Reaction Systems and the Relation between Reac-
tivity and Energetics. A main point of interest is the charge
density on the olefin component in the A+•(B) complexes, which
depends on the IP difference between the components.36,37 The
question is how large a ∆IP between the components still places
enough charge density on the olefin to allow nucleophilic attack
by another olefin molecule.
It is observed in the clusters, but in very small yield.34 Spectral
•+
shifts in the clusters indicate that I2 is formed from larger
clusters containing several I molecules, where a cooperative
process involving several olefin molecules can effectively lower
the IP of the aggregate to allow ultimate charge transfer from
the aromatic center.
(4) In benzene/propene, [(propene)n•+, n g 2] and [benzene•+-
(propene)n] adducts are observed in both the gas phase and
clusters.
To investigate this question, we examined several aromatic-
olefin combinations A•+ + B as shown in Table 1. To consider
all the possible reactions in these systems, we also present the
thermochemistry for the formation of the AB•+ adducts,
probably alkylbenzene ions, that are observed in all the systems.
Table 1 shows an inverse relation between the ∆IPs of the
components and the rate coefficients to form B2•+. As noted
(5) Olefin dimers are not observed in toluene/propene and
benzene/ethylene, either in the gas phase or in clusters. Also,
•+
no I2 is observed in the mesitylene/isobutene system in
clusters. In other words, in systems where the IP difference
between the reactants is >0.87 eV, the olefin dimer ion is not
formed in either phase.
•+
above, the formation of B2 is observed for reactants with a
∆IP < 0.42 eV, but not for larger ∆IP, even though the
The close similarities between the gas phase and the “semi-
condensed” cluster phase suggest that the same mechanism is
operative in both. This suggests that the mechanism can
possibly apply also in the condensed phase. The cluster studies
are reported in detail elsewhere.34
•+
formation of B2 would always be significantly exothermic.
Hypothetical ion-catalyzed dimerization (A•+ + 2B f A•+
+ B2) leading to the neutral olefin dimer is also exothermic for
all the reaction systems in Table 1. Its energetics are less
favorable than for the formation of the ionic olefin B2•+ dimers
in most of the systems. Exceptions are the toluene/isobutene
system where the energetics are similar and the p-xylene/
isobutene system in which an ionic dimer is less energetically
favored and is in fact not observed in the gas phase. This ion-
catalyzed process may in fact be occurring in any of the reaction
systems that we investigated, even with high efficiency, but since
it does not lead to new ionic products, we would not observe
it.
Reaction Complexes and Products. In a preceding study20
we observed the formation of C8H16•+ from i-C4H8•+ + i-C4H8.
The product was identified as a covalent ion of an octene with
•+
an IP of 8.55 ( 0.15 eV, possibly (CH3)2CHCHCHCH(CH3)2
.
It reacted with i-C4H8 by H2 transfer, yielding a nonreactive
C8H14•+. The present C8H16 ion shows similar chemistry
•+
(Figure 1b). We also note that the product ion must correspond
to a neutral with an IP < 8.82 eV (the IP of toluene), since
•+
otherwise the reaction would more likely produce C6H5CH3
We noted that the ∆IP of 0.48 eV in the benzene/propene
system makes direct endothermic charge transfer even more
strongly prohibitive than in the toluene/isobutene system, with
+ C8H16, amounting to ion-catalyzed dimerization of neutral
i-C4H8 similar to the reaction of CF3+ with two C2F4 molecules
to produce C4F8.28 These considerations are consistent with a
an expected efficiency of k/kcollision ) exp(-∆H°/RT) ) 10-9
.
product octene ion branched on the olefinic carbon atom. It is
•+
(36) Meot-Ner (Mautner), M.; Hunter, E. P.; Hamlet, P.; Field, F. H. J.
Am. Chem. Soc. 1978, 100, 1466.
(37) Meot-Ner (Mautner), M.; El-Shall, M. S. J. Am. Chem. Soc. 1986,
108, 4386.
of interest whether the dimer ion formed in the C6H5CH3
-
(i-C4H8)2 complex is the same as that formed in the reaction of
•+
i-C4H8 with i-C4H8 in the gas phase.20
(38) Holman, R. W.; Rozeboom, M. D.; Gross, M. L.; Warner, C. D.
Tetrahedron. 1986, 42, 6235. Holman, R. W.; Warner, C. D.; Hayes, R.
N.; Gross, M. L. J. Am. Chem. Soc. 1990, 112, 3362.
(39) Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Levin, R. D.; Mallard,
W. G. J. Phys. Chem. Ref. Data 1988, 17, Suppl 1.
We observe that the final, presumably covalent, CH3C6H4-
•+
t-C4H9 adduct does not interconvert to form the isobutene
•+
dimer C8H16•+. This suggests that the reactive C6H5CH3
-
(i-C4H8) adduct may be a noncovalent aromatic/olefin π

Coupled Reactions of Condensation and Charge Transfer
J. Am. Chem. Soc., Vol. 119, No. 35, 1997 8339
Furthermore, the IP of propene, 9.73 eV, is above the two-
photon energy used to ionize benzene, 9.58 eV, so that direct
charge transfer from excited (C6H6•+)* can be ruled out
energetically. Therefore, this system presents an energetically
even more clear-cut candidate to distinguish between dimer
formation through the proposed reaction 8 or through direct
charge transfer followed by dimer formation (see Appendix 1a).
We have, in fact, confirmed the direct mechanism of the
benzene/propene system by an independent method using the
selected ion flow tube (SIFT) technique. The mass-selected
C6H6•+ ion was injected into a He/C3H6 gas flow (not containing
C3H6•+). The production of (C3H6)n•+ ions was observed, with
n ) 2-6, with rate coefficients comparable to those measured
in our R2PI-HPMS method. The results will be reported
elsewhere.40
efficiency already at moderately low temperatures. At interstel-
lar temperatures below 100 K, such reactions can therefore
become effective synthesis pathways. Moreover, at such low
temperatures large clusters of monomers can attach to an ionized
aromatic and allow concerted multistep polymerization to drive
substantially endothermic charge transfer from the aromatic
surface. Low temperatures may also be useful in industrial
applications. More detailed temperature studies are of interest
to define the functional form of the temperature coefficients.
The present mechanism avoids the formation of the olefin
monomer ions and their reaction products. For example, in the
benzene/propene system we observed the exclusive formation
•+
of (propene)n with n ) 2-7. In contrast, ion-molecule
•+
+
reactions between C3H6 and C3H6 produce C3H7+, C4H7
,
C4H8•+, and C5H9+ that can further polymerize.44,45 The present
mechanism leads exclusively to the C6H12 radical dimer ion
•+
Summary, and Implications for Industrial and
Astrochemical Polymerization
channel, suggesting a useful photoinitiation method for pure
products.
We noted the similarity of the gas-phase observations to those
in preformed clusters.34 This suggests that the mechanism may
also apply in the condensed phase in common aromatic solvents
such as toluene. This can allow photoinitiation using a “solVent
as initiator approach” to eliminate chemical initiators, with
beneficial economic and environmental results.
The main observation is a coupled reaction of dimer formation
and ionization of the olefin dimer. These processes are observed
in simple gas-phase reactions of an ionized aromatic molecule
with two olefin molecules. The observed kinetic trends,
especially the pressure effects, are best reproduced by a
mechanism through a collisionally stabilized noncovalent
intermediate complex A•+(B) in which the olefin molecule is
“adsorbed” on the ionized aromatic surface, and assumes a
charge density by interaction with the aromatic ion. Upon
collision with another olefin molecule, in the resulting A•+(2B)
complex, one olefin molecule can carry sufficient charge density
to activate it for nucleophilic attack by the second olefin
molecule, resulting in covalent condensation. Formation of an
olefin with a lower IP than that of the aromatic component will
then result in full charge transfer, leading to the observed product
ion. Alternatively, the required processes in the A•+(2B)
complex may occur simultaneously. Other alternative mecha-
nisms are discussed in the Appendices. Theoretical study of
the potential energy surfaces and transition states may help to
identify the mechanism.
The observed processes add to the small group of multibody
ion-molecule reactions where attachment to an ion activates a
neutral molecule for reaction with a further neutral molecule.
The observed reactions are similar to anodic electrochemical
polymerization, where strong anionic nucleophiles added to
olefins were assumed to reduce their effective oxidation
potentials,41 although in the complex solution system this was
debated.42 The present simple gas-phase systems are clearer
examples where collaborative interaction between substrate
molecules effectively shifts their oxidation potential and allows
charge transfer from the “electrode”, which is otherwise
prohibitively endothermic.
Acknowledgment. This research is supported by the Na-
tional Science Foundation (Grant No. CHE 9311643). Acknowl-
edgment is also made to the Thomas F. and Kate Miller Jeffress
Memorial Trust (Grant No. J-302) for the partial support of this
research.
Appendix 1. Alternative Mechanisms
a. Direct Charge Transfer, Followed by Dimer Forma-
tion. This mechanism would entail charge transfer from T•+
to I through the dissociation of the reaction complex T•+(I)* to
I•+ + T, in competition with stabilization to form TI•+. We do
not observe the formation of I•+, but if formed, it could react
rapidly with T by charge transfer to form T•+ and with I to
•+
form C8H16 (and t-C4H9+). Therefore, its absence in detect-
able concentrations does not rule out conclusively its formation.
However, as noted in observation 1 above, equating the 0.42
eV endothermicity with the activation energy suggests a reaction
efficiency r ) k/kcollision ) exp(-Ea/RT) of 10-7. This calcula-
tion is based on relative IPs from evaluated accurate thermo-
chemical data.46 In comparison, the observed nominal second-
•+
order rate coefficient for the formation of I2 (and through it,
its subsequent products) under the present conditions ranges
from 5 to 25 × 10-12 cm3 s-1. The observed reaction efficiency
of (0.5-2.5) × 10-2 is about 5 orders of magnitude larger than
expected for direct charge transfer.
Other evidence against the formation of i-C4H8•+ is the lack
of formation of t-C4H9+, a known product of the reaction of
The toluene-isobutene reaction shows a significant negative
temperature dependence. The present exploratory two-point
study cannot define the functional form which may be complex
because of the multistep mechanism and because of nonexpo-
nential terms.43 As an estimate, using kf ) aTn, the second-
order rate coefficients for the formation of I2•+ and TI•+ would
give temperature coefficients as large as T-12.9 and T-7.8, and
extrapolate to the collision rate at 214 and 168 K, respectively.
This suggests that the reactions may achieve unit collision
•+
i-C4H8 with i-C4H8 under similar conditions.20
Another test of the direct charge transfer mechanism can be
performed by using high concentrations of toluene, to use T as
a scavenger for I•+. In the direct transfer mechanism, the I•+
product would react competitively with I to give I2•+ in reaction
8, and with T to regenerate T•+ by fast exothermic charge
transfer. This competing process would become more signifi-
cant with increasing [T]/[I] density ratios, and the rate of
(40) Pithawalla, Y. B.; Gao, J.; Meot-Ner (Mautner), M.; Bohme, D. K.
El-Shall, M. S. Manuscript in preparation.
(41) Manning, G.; Parker, V. D.; Adams N. R. J. Am. Chem. Soc. 1969,
91, 4584.
(42) Eberson, L.; Utley, J. H. P. In Organic Electrochemistry; Beizer,
M. M., Lund, H., Eds., Marcel Dekker: New York, 1983; p 409.
(43) Viggiano, A. A. J. Chem. Phys. 1986, 84, 244.
(44) Bowers, M. T.; Elleman, D. D.; O’Malley. R. M.; Jennings, K. R.
J. Phys. Chem. 1970, 74, 2583.
(45) Abramson, F. P.; Futrell, J. H. J. Phys. Chem. 1968, 72, 1994.
(46) Levin, R. D.; Lias, S. G. Ionization Potential and Appearance
Potential Measurements, 1971-1981; National Bureau of Standards:
Washington, DC, 1982.

8340 J. Am. Chem. Soc., Vol. 119, No. 35, 1997
Meot-Ner et al.
production of I2•+ would decrease correspondingly. On the other
hand, the production of TI•+ through collisional stabilization
of (T•+I)* should not be affected by the [T]/[I] concentration
ratio. The net effect is that the I2•+/TI•+ product ratio would
decrease with increasing [T]/[I] concentration ratio. To test this,
we performed two experiments under similar conditions but with
[T]/[I] ratios of 0.3 and 8.3, i.e., concentration ratios different
by a factor of 25. The I2•+/TI•+ product ratio remained constant
at 0.3 ( 0.1 in both experiments as shown in Figure 6. This
contraindicates an I•+ intermediate that would be trapped by T.
For these reasons, we can rule out direct charge transfer from
T•+ to I as a significant process in our system.
I2 •+/(TI)•+ ) k15[I]/(k13[M])
(18b)
The product ratio I2•+/TI•+ would be inversely dependent on
[M], which is contrary to the results in Figure 3a. However, at
high pressures of the efficient third-body I, this mechanism may
contribute to the formation of TI•+, as suggested by the decrease
of the I2•+/ TI•+ product ratio in Figure 3b.
We also note that the third-body number density [Ar] is in
large excess over [I] in our experiments. For example, in the
reaction system in Figure 1, it is in excess by a factor of 850,
and large ratios between 200 and 1000 apply in all of our
experiments. The product distribution could be explained only
if the collisional stabilization efficiency of [T+•(I)]* by [Ar] is
smaller by factors of 200-1,000 than the reaction efficiency
with I to form I2•+, i.e., with collision efficiency of Ar < 0.005,
at least an order of magnitude smaller than usual even for
monoatomic third bodies.
b. Competitive Reactions of an Excited Complex. The
first step in the reaction is the formation of an excited complex
[T•+(I)]*. Evidence for its formation and back-dissociation to
reactants is the low overall reaction efficiency, <10-2
.
The common ion-molecule mechanism would proceed
through competitive reactions of [T•+(I)]* by back-dissociation
to reactants (reaction -6), in competition with dissociation to
products and stabilization, the hypothetical reactions 13-15,
that constitute mechanism 2.
To check the third-body effect, we performed an experiment
with an efficient polyatomic collisional third body, n-C3H8, that
has many vibrational modes to absorb the internal energy of
the complex. We examined reaction systems with [Ar] or
[C3H8] )1.5 × 1016 cm-3 and [I] ) 2.8 × 1012 cm-3. We
found that, with Ar, kf was 22 × 10-12 cm3 s-1, and, with
n-C3H8, 44 × 10-12 cm3 s-1. The partial rate coefficients were,
with Ar, kf(I2•+) ) 5.8 × 10-12 cm3 s-1 and kf(TI•+) ) 17 ×
C6H5CH3•+ + C4H8 f [C6H5CH3•+(C4H8)]*
[C6H5CH3•+(C4H8)]* f C6H5CH3•+ + C4H8
(6)
(-6)
10-12 cm3 s-1, and with C3H8, kf2(I2•+) ) 13 × 10-12 cm3 s-1
[C6H5CH3•+(C4H8)]* + M f CH3C6H4C4H9•+ + M
(13)
•+
and kf(TI•+) ) 32 × 10-12 cm3 s-1. Correspondingly, the I2
/
TI•+ product ratio was with Ar 0.35 and with C3H8 0.41.
Therefore, the rate coefficients increased by about a factor of
2, and since the collision efficiency of C3H8 should be near
unity, that of Ar is then g0.5. Most importantly, the product
distribution did not change significantly with the more efficient
third body. Therefore, the product ratio cannot be attributed to
low collisional efficiency of the third-body Ar to stabilize the
excited [T•+(I)]* complex.
[C6H5CH3•+(C4H8)]* f CH3C6H4C4H9•+ + hν (14)
[C6H5CH3•+(C4H8)]* + C4H8 f C8H16•+ + C6H5CH3
(15)
Dissociation of [T•+(I)]* to I•+ + T was ruled out in the
preceding section, and for a bimolecular process it would have
To further examine mechanism 2, we performed ACUCHEM
kinetic simulations applied to the model system of Figure 1, as
we did above for Scheme 1. Rate coefficients were assigned
as described in Appendix 2. The results in Figure 2 show that
the mechanism can simulate the observed pressure effects on
kf as well as mechanism 1 (Figure 2), although it does not match
well the pressure effect on the partial rate coefficient kf(I2•+).
•+
to be replaced by a reaction of [T•+(I)]* with I to form I2
.
This reaction would bypass the second intermediate T•+(I) in
Scheme 1. The formation of the final stable TI•+ could be
through collisional or radiative stabilization of [T•+(I)]*.
Steady-state assumption on [T•+(I)]* would then yield eqs 16-
18 for the formation of the products.
In this mechanism the rate formation of I2•+ is limited by kcollision
-
kf2(I2•+) ) k6 k15[I]/(k-6 + k15[I] + k13[M] + k14) (16)
[I], and because of the much higher number density of [Ar],
the product distribution I2•+/TI•+ could be matched only by using
an unusually small collision efficiency of 0.0022 for the
stabilization of [T•+(I)]* by Ar. Furthermore, the third-body
density effect on the product distribution I2•+/TI•+ from the
simulation shows a strong inverse dependence on [Ar] as shown
in Figure 3a, as expected from the analytical solution, eq 18b,
above.
All of the kinetic evidence therefore suggests that the
competitive transfer/association directly from the excited com-
plex [T•+(I)]* as in mechanism 2 is less compatible with the
experimental observations than Scheme 1.
kf2(TI•+) ) k6(k13[M] + k14)/(k-6 + k15[I] + k13[M] + k14)
(17a)
Assuming radiative stabilization of [T•+(I)]*, i.e., if k14
.
k13[M], then k13[M] would be negligible in the denominator of
eq 17a, and the overall forward rate coefficient kf ) kf(I2•+) +
kf(TI•+) would be independent of third-body density [M], in
contrast to the results in Figure 2. The results therefore suggest
predominantly collisional stabilization under our high-pressure
conditions, although radiative stabilization may be significant
at low pressures, as discussed above. Under our conditions k13-
[M] . k14 and eq 17a reduces to eq 17b.
Appendix 2. Rate Coefficients for Kinetic Simulations
kf2(TI•+) ) k6(k13 [M])/(k-6 + k15[I] + k13[M]) (17b)
We select rate coefficients for Scheme 1 as follows. For
reactions 6, 7, and 8 we assume unit collision efficiency and a
rate coefficient of 10-9 cm3 s-1, so the rates are given by 10-9[I]
s-1 for reaction 6 and 10-9[Ar] s-1 for reaction 7. Note that
the efficiency of Ar may be <1, but the experiments with
propane as carrier gas suggest that the efficiency is g0.5 for
this reaction as discussed above. We adjusted the other rate
coefficients to simulate the model system in Figure 1, on the
The product ratio according to this mechanism would be given
by eq 18a. With collisional stabilization predominant, i.e., with
I2•+/(TI)•+ ) k15[I]/(k13[M] + k14)
(18a)
k13[M] . k14, eq 18a is reduced to eq 18b.

Coupled Reactions of Condensation and Charge Transfer
J. Am. Chem. Soc., Vol. 119, No. 35, 1997 8341
basis of the following considerations. In the model system, [I]
[T•+(I)]*, and we adjusted this to k-6 ) 6.08 × 108 s-1 to fit
the ion profiles in the model experimental system. All of the
rate coefficients were kept constant in the simulations of pressure
and concentration effects.
) 2.74 × 1013 cm-3; therefore, we use k8[I] ) 2.74 × 104 s-1
.
The parallel reaction 9 is a first-order or pseudo-first order
reaction as discussed above. The value of k9 ) 5.07 × 104 s-1
reproduces the product distribution of [I2•+]/[TI•+] ) 0.54
observed experimentally in the model system.
To select rate coefficients for mechanism 2, we chose rate
coefficients to simulate the reaction system described in Figure
1. As for Scheme 1, for k6 and k15 we apply unit collision
efficiency. The product ratio I2•+/TI•+ is given by kcollision[I]/
To assign k-7, we consider the following points. The order
of the overall reaction with respect to [I] depends on the relative
rates of the reactions of the complex T•+(I). If it is in rapid
equilibrium with the reactants through reaction -7, the con-
centration of T•+(I) would be proportional to [I], and the overall
kcollision[Ar]f, where an efficiency f ) 0.0022 must be applied
to match the observed product distribution. Note that if we
assumed that reaction 15 proceeds with collision efficiency <1,
the stabilization efficiency f would have to be proportionally
even smaller. To match the observed overall rate coefficient
kf, the required value of k-6 is 3.0 × 106 s-1. This value is
much smaller than for Scheme 1, as there the back-dissociation
competes with stabilization at the rate kcollision[Ar], while in
mechanism 2 it competes with the forward reaction determined
by kcollision[I], under conditions where [Ar] . [I]. As for Scheme
1, the rate coefficients were kept constant in the simulations of
the pressure and concentration effects.
•+
rate of formation of I2 through reaction 8 would make kf-
(I2•+) linear with [I] (i.e., the pseudo-first-order rate coefficient
•+
for formation of I2 would be proportional to [I]2). On the
other hand, if the formation of T•+(I) was rate controlling and
irreversible, with k-7 , k8[I] + k9, then kf(I2•+) would be
independent of [I], (i.e., the pseudo-first-order rate coefficient
•+
for the formation of I2 would be proportional to [I]). The
results in Figure 4 show an intermediate behavior. This suggests
that k-7 is comparable to k8[I] and k9. Therefore, we used k-7
) 2.68 × 104 s-1. Finally, the overall reaction rate is controlled
primarily by the back-dissociation of the first excited complex
JA962635X