Direct observation of a hydrogen atom adduct to C-5 in uracil. A neutralization-reionization mass spectrometric and ab initio study

Article abstract of DOI:10.1021/jp011349r

The elusive hydrogen atom adduct to the C-5 position in uracil was generated specifically in the gas phase, and its unimolecular dissociations were elucidated by neutralizafion-reionization mass spectrometry. Collisional electron transfer to the 5,6-dihydrouracil-6-yl cation generated the 5,6-dihydropyrimidine-2,4(1H,3H)-dion-6-yl radical (1), which was the most stable hydrogen atom adduct to C-5 in uracil. A substantial fraction of 1 was stable on the 5.1 μs time scale. The main unimolecular dissociations of 1 were specific losses of hydrogen atoms from C-5 and ting cleavages, as determined by deuterium labeling. Radical 1 did not isomefize unimolecularly to 5,6-dihydropyrimidine-2,4(1H,3H)-dion-5-yl (2). Ab initio calculations up to effective QCISD(T)/6-311+G(3df,2p) and combined density functional theory and perturbational calculations up to B3-MP2/6-311+G(3df, 2p) provided bond dissociation and transition state energies for several radical and ion dissociations that were used for assessing rate constants by RRKM theory. The observed dissociations occur to a large part from excited electronic states of uracil radicals that are formed by vertical electron capture. The adiabatic ionization energies of uracil and 1 were calculated as 9.24 and 6.70 eV, respectively.

Full text of DOI:10.1021/jp011349r

J. Phys. Chem. A 2001, 105, 8339-8351
8339
Direct Observation of a Hydrogen Atom Adduct to C-5 in Uracil. A
Neutralization-Reionization Mass Spectrometric and ab Initio Study
Erik A. Syrstad, Shetty Vivekananda, and Frantisˇek Turecˇek*
Department of Chemistry, Bagley Hall, Box 351700, UniVersity of Washington,
Seattle, Washington 98195-1700
ReceiVed: April 10, 2001; In Final Form: July 2, 2001
The elusive hydrogen atom adduct to the C-5 position in uracil was generated specifically in the gas phase,
and its unimolecular dissociations were elucidated by neutralization-reionization mass spectrometry. Collisional
electron transfer to the 5,6-dihydrouracil-6-yl cation generated the 5,6-dihydropyrimidine-2,4(1H,3H)-dion-
6-yl radical (1), which was the most stable hydrogen atom adduct to C-5 in uracil. A substantial fraction of
1 was stable on the 5.1 µs time scale. The main unimolecular dissociations of 1 were specific losses of
hydrogen atoms from C-5 and ring cleavages, as determined by deuterium labeling. Radical 1 did not isomerize
unimolecularly to 5,6-dihydropyrimidine-2,4(1H,3H)-dion-5-yl (2). Ab initio calculations up to effective
QCISD(T)/6-311+G(3df,2p) and combined density functional theory and perturbational calculations up to
B3-MP2/6-311+G(3df,2p) provided bond dissociation and transition state energies for several radical and
ion dissociations that were used for assessing rate constants by RRKM theory. The observed dissociations
occur to a large part from excited electronic states of uracil radicals that are formed by vertical electron
capture. The adiabatic ionization energies of uracil and 1 were calculated as 9.24 and 6.70 eV, respectively.
Introduction
Nucleobase residues in DNA and RNA are susceptible to
attack by energetic particles produced by radiation or chemical
processes.1 In particular, radiolysis of water produces electrons
and small radicals (H, OH, etc.) that are known to add to the
nucleobases and cause chemical modifications that are referred
to as radiation damage.1,2 Although the kinetics of radical
additions to the four DNA nucleobases have been studied in
detail, much less is known about the reactivity of the RNA
nucleobase uracil. Flossman, Westhof, and co-workers studied
the formation of uracil radicals following irradiation with 100
kV X-rays3 or upon photosensitized irradiation with UV light.4
Electron paramagnetic resonance spectra indicated the formation
of several transient species that were assigned to hydrogen atom
adducts to uracil. Von Sonntag and co-workers studied pulse
radiolysis of uracil derivatives in aqueous solution and concluded
that hydrogen atom adducts to O-4 and C-6 were formed by
protonation of the intermediate uracil anion radical.5 For the
ring numbering in uracil, see the formulas below. Two mech-
anisms for the formation of hydrogen atom adducts to nucleo-
bases have been suggested, as shown for uracil in Scheme 1.1
The first mechanism (mechanism I) assumes direct addition to
the nucleobase of a hydrogen atom produced by radiolysis of
water. The presumed sites of H atom additions are at the C-5s
C-6 double bond to produce radicals 1 or 2.6 Mechanism I is
analogous to that presumed for the addition of OH radicals to
nucleobases;7 however, additions of H atoms are much less
thoroughly documented.1,2 The second mechanism (mechanism
II) presumes addition of a thermal electron to form a uracil
anion-radical. The latter is protonated by the solvent or an acid
present in the solution to give the H atom adduct 3.8
results obtained so far provided only indirect evidence of the
existence of transient radicals through redox reactions2 and UV-
vis absorption spectra.9 Under radiolytic or photolytic conditions,
a variety of products are typically formed that are monitored
spectroscopically in the mixture.5,9,10 Structures and intrinsic
properties of nucleobase radicals are therefore not amenable to
analysis. Limited information about the structures and energetics
of several nucleobase radicals has been provided by ab initio
calculations that were carried out at low6 to medium levels of
theory.11
SCHEME 1
Although the radiation chemistry of uracil and other pyrimi-
dine nucleobases has been studied extensively, most of the
* To whom correspondence should be addressed. Tel.: (206) 685-2041.
Fax: (206) 685-3478. E-mail: turecek@chem.washington.edu.
10.1021/jp011349r CCC: $20.00 © 2001 American Chemical Society
Published on Web 08/16/2001

8340 J. Phys. Chem. A, Vol. 105, No. 36, 2001
Syrstad and Vivekananda
We now describe the targeted, specific preparation of two
pivotal uracil radicals relevant to radiolytic mechanisms I and
II.12 In the present paper, we report on the specific formation
in the gas phase of 5,6-dihydropyrimidine-2,4(1H,3H)-dion-6-
yl (1), which is the elusive intermediate of H atom addition to
C-5 in uracil. We prepared radical 1 by femtosecond collisional
electron transfer to its cationic precursor 1+, and studied the
mechanisms of radical unimolecular dissociations by neutraliza-
tion-reionization mass spectrometry13,14 with the help of deu-
terium-labeled derivatives 1a-1c. In addition, we report detailed
ab initio calculations, performed at high levels of theory, up to
effective QCISD(T)/6-311+G(3df,2p), of radical 1, several of
its isomers, dissociation products, and relevant transition states.
Excited electronic states are found to play an important role in
the unimolecular chemistry of 1 and are therefore addressed by
time-dependent density functional theory calculations.15 In an
accompanying paper, we address the chemistry of 4-hydroxy-
3,4-dihydropyrimidine-2(1H)-on-4-yl radical (3), which is the
elusive intermediate of protonation at O-4 in uracil anion-
radical.16
84, 76, 72, 60, 57), and a small amount (<5%) of a methyl
ester (m/z 104, 74, 72, 62, 60, 59, 57).
Ethyl 3-(p-toluenesulfonyloxy)pentanoate-3-d (6) was pre-
pared by tosylation of ester 5 (p-toluenesulfonyl chloride,
pyridine, 0 °C, 48 h) followed by standard workup (ice water,
extraction with chloroform, drying with MgSO4 followed by
solvent evaporation). Crude tosylate 6 was used further without
purification.
Ethyl 2-pentenoate-3-d (7). Crude tosylate 6 (34 mmol) was
dissolved in tetrahydrofuran (40 mL), cooled to -78 °C, and
potassium t-butoxide (3.8 g, 34 mmol) was added under
nitrogen. The reaction mixture was stirred at -78 °C for 2 h
and then allowed to warm to room temperature. Pentane (50
mL) was added, and the slurry was poured on ice/water (300
mL), the organic layer was separated, and the aqueous layer
was extracted with 3 × 25 mL of pentane. The pentane solution
was washed with brine, dried with magnesium sulfate and
decolorized with active charcoal. Pentane was distilled off
through a 20 cm Vigreux column, and the residue was distilled
at ca. 100 Torr to yield 4.4 g of ester 7. Mass spectrum: m/z
129 (M+•).
Experimental Part
2-Pentenoic acid-3-d (8). Ethyl 2-pentenoate-3-d (7, 4.4 g,
34 mmol) was dissolved in a mixture of methanol (25 mL) and
water (5 mL), sodium hydroxide (2.4 g, 60 mmol) was added
and the solution was refluxed for 3 h. The solvents were
evaporated and the residue was taken in water, acidified to pH
1 with hydrochloric acid, and extracted with five 25 mL portions
of dichloromethane. Workup followed by distillation gave 2.45
g (71%) of 8, bp 100 °C/20 Torr.
6-Ethyl-6-d-5,6-dihydropyrimidine-2,4(1H,3H)-dione (4c).
2-Pentenoic acid-3-d (8, 2.4 g, 24 mmol) and urea (3.65 g, 61
mmol) were heated in a stainless steel pressure vessel (40 mL
volume) at 195 °C for 2 h.17 The vessel was cooled to room
temperature, depressurized to release ammonia, and the gummy
residue was dissolved in 50 mL of warm water. The warm
solution was decolorized with active charcoal, filtered, and the
filtrate volume was reduced to 15 mL on a rotavap. Compound
4c crystallized upon standing at 4 °C. The crystals were collected
and recrystallized once from water. Yield 650 mg (19%), mp
192 °C. 70 eV mass spectrum (m/z, rel intensity): 144 (1), 143
(11), 142 (2), 115 (7), 114 (100), 113 (51), 112 (5), 72 (3), 71
(75), 70 (12), 60 (4), 59 (2), 57 (4), 56 (3), 44 (12), 43 (14), 42
(5), 41 (3). 1H NMR (300 MHz, CDCl3): 1.0 (t, 3H, -CH2CH3),
1.62 (m, 2H, -CH2CH3), 2.42 (d, 1H, H-5), 2.70 (d, 1H, H-5),
5.35 (broad s, 1H, N-H), 7.40 (broad s, 1H, N-H).
Materials. Uracil, 2-pentenoic acid, urea, and ethyl-3-
oxopentanoate (all Aldrich) were used as received. Deuterium-
labeled reagents, sodium borodeuteride (99% D, Aldrich), ND3
(Matheson, 99% D), D2O, CD3OD, and acetone-d6 (all Cam-
bridge Isotope Laboratories) were also used as received. 6-Ethyl-
5,6-dihydropyrimidine-2,4(1H,3H)-dione (4) was synthesized
according to literature,17 and recrystallized from water, yield
1
27%, mp 194 °C. H NMR (300 MHz, CDCl3): 1.0 (t, 3H,
-CH2CH3), 1.64 (m, 2H, -CH2CH3), 2.42 (dd, 1H, H-5), 2.70
(dd, 1H, H-5), 3.55 (m, 1H, H-6), 5.60 (broad s, 1H, N-H),
7.62 (broad s, 1H, N-H). 70 eV mass spectrum (m/z, rel
intensity): 142 (9), 114 (5), 113 (100), 112 (35), 71 (3), 70
(80), 58 (3), 56 (7), 43 (23), 42 (7), 41 (7).
6-Ethyl-5,6-dihydropyrimidine-2,4(1D,3D)-dione (4a). Com-
pound 4 (144 mg, 1 mmol) was dissolved in THF (13 mL) and
stirred overnight at 20 °C. Then, 11 mL of CH3OD was added
and the solution was stirred for 10 min, followed by solvent
evaporation under vacuum at room temperature. 70 eV mass
spectrum of 4a was (m/z, rel intensity) 145 (2), 144 (11), 143
(7), 142 (2), 116 (5), 115 (80), 114 (76), 113 (26), 112 (4), 72
(5), 71 (100), 70 (29), 59 (6), 58 (8), 57 (13), 56 (7), 55 (7).
6-Ethyl-5,6-dihydro-5,5-d2-pyrimidine-2,4(1H,3H)-dione
(4b). Compound 4 (160 mg, 1.1 mmol) was dissolved in 43
mL D2O containing 450 mg K2CO3. The solution was stirred
72 h at 20 °C, acidified with 37% DCl in D2O, and solvent was
evaporated to dryness in vacuo. The product was taken in
methanol, the solution was filtered, and the solvent was
evaporated in vacuo. 70 eV mass spectrum (m/z, rel intensity):
148 (4), 147 (3), 146 (3), 145 (4), 144 (9), 143 (1), 116 (14),
115 (100), 114 (11), 113 (31), 73 (8), 72 (86), 71 (8).
6-Ethyl-6-d-5,6-dihydropyrimidine-2,4(1H,3H)-dione (4c)
was prepared from ethyl 3-oxopentanoate in five steps, as
follows.
Ethyl 3-hydroxypentanoate-3-d (5). Ethyl 3-oxopentanoate
(5.0 g, 34.7 mmol) was dissolved in 25 mL of methanol and
added at 0 °C to a solution of sodium borodeuteride (0.73 g,
16.6 mmol) in methanol (25 mL). The mixture was stirred at 0
°C for 45 min, acidified with acetic acid (2 g), and the solvent
was evaporated in vacuo. The residue was extracted with ether
(4 × 25 mL), dried, and the solvent was evaporated in vacuo
to yield 4.9 g (96%) of an oily product. GC-MS showed
essentially one peak corresponding to 5 (m/z 118, 102, 90, 88,
Methods
Measurements were carried out on a tandem quadrupole
acceleration-deceleration mass spectrometer described previ-
ously.18 Electron ionization was used to generate cations 1+-
1c+ in a standard electron ionization source. Typical ionization
conditions were as follows: electron energy 70 eV, emission
current 500 µA, temperature 200-250 °C. Ion-molecule
reactions and charge exchange ionization were carried out in a
tight chemical ionization (CI) source. Typical ionization condi-
tions were as follows: electron energy 100 eV, emission current
1-2 mA, temperature 280-300 °C, ion source potential 80 V.
NH3, ND3, acetone, CD3COCD3, and 2-methylpropane, water,
and D2O were used as CI reagent gases at pressures 1.0-1.5 ×
10-4 Torr as read on an ionization gauge located at the diffusion
pump intake. Stable precursor ions were passed through a
quadrupole mass filter operated in the radio frequency-only
mode, accelerated to the total kinetic energy of 8250 eV and
neutralized in the collision cell floated at -8170 V. The

Hydrogen Atom Adduct to C-5 in Uracil
J. Phys. Chem. A, Vol. 105, No. 36, 2001 8341
precursor ion lifetimes were 30-40 µs. Dimethyl disulfide
(DMDS) was admitted to the differentially pumped collision
cell (cell I) at a pressure such as to achieve 70% transmittance
of the precursor ion beam. The ions and neutrals were allowed
to drift to a four segment conduit,19 where the ions were reflected
by the first segment floated at +250 V. The neutral flight times
in standard NRMS measurements were 5.1 µs. The fast neutral
species were reionized in a second cell (cell II) by collision
with oxygen at a pressure adjusted such as to achieve 70%
transmittance of the precursor ion beam. The ions formed in
cell II were decelerated, energy filtered, and analyzed by a
quadrupole mass filter operated at unit mass resolution. The
instrument was tuned daily to maximize the ion current of
reionized CS2+•. Typically, 40 repetitive scans were accumulated
per spectrum. Collisional activation of transient radicals (denoted
as NCR) was carried out in the differentially pumped conduit
between cell I and cell II using He at pressures to achieve 50%
precursor beam transmittance. The conduit segments were
maintained at +250 V to prevent detection of any reionized
species formed in this region.
several systems since.29,30 This resulted in error cancellation and
provided relative energies denoted as B3-PMP2 or B3-
ROMP230 as discussed below. Calculations on closed-shell
systems are marked by B3-MP2. In addition, a composite
procedure was adopted that consisted of a single-point quadratic
configuration interaction calculation,31 QCISD(T)/6-31G(d,p),
and basis set expansion up to 6-311+G(3df,2p) through PMP2
or ROMP2 single-point calculations according to eq 1:
QCISD(T)/6-311+G(3df,2p) ≈ QCISD(T)/6-31G(d,p) +
MP2/6-311+G(3df,2p) - MP2/6-31G(d,p) (1)
This level of theory is intermediate between those of the
Gaussian 2 (MP2) method32 which uses the 6-311G(d,p) basis
set in the large QCISD(T) calculation and the G2(MP2, SVP)
method33 which uses the 6-31G(d) basis set instead. We also
utilized the previous finding that restricted open-shell calcula-
tions (ROMP2) provided good stabilization energies for small
organic radicals.27 Basis set expansions to effective QCISD/6-
311+G(2d,p), QCISD(T)/6-311+G(2d,p), and QCISD(T)/6-
311+G(2df,p) were also tested for selected systems.
Franck-Condon energies in vertical neutralization and reion-
ization were taken as absolute differences between the total B3-
MP2/6-311+G(2d,p) energies of fully optimized ion or neutral
structures and those in which an electron has been added to an
optimized cation structure or subtracted from an optimized
neutral structure. No zero-point corrections were applied to the
calculated Franck-Condon energies.
Collisionally activated dissociation (CAD) spectra were
measured on a JEOL HX-110 double-focusing mass spectrom-
eter of forward geometry (the electrostatic sector E precedes
the magnet B). Ion collisions with air were monitored in the
first field-free region at pressures to achieve 70% transmittance
of the ion beam at 10 keV. The spectra were obtained by
scanning E and B simultaneously while maintaining a constant
B/E ratio (B/E linked scan).
Gradient optimizations of excited-state geometries were
performed with spin-unrestricted configuration interaction singles
(UCIS)34 using the 6-31+G(d,p) basis set. Improved energies
for excited states were obtained from UCIS and time-dependent
density functional theory15 single-point calculations using the
B3LYP hybrid functional and the larger 6-311+G(2d,p) basis
set.
RRKM calculations used Hase’s program35 that was recom-
piled for MS-DOS and run under Windows NT.36 Vibrational
state densities were obtained by direct count of quantum states
in 2 kJ mol-1 steps for internal energies up to 80-120 kJ mol-1
above the threshold. The rotational states were treated adiabati-
cally37 and the microscopic rate constants, k(E,J,K), were
Boltzmann-averaged over the thermal distribution of rotational
J and K states pertaining to the ion source temperature. Thermal
rate constants were calculated using the standard transition state
theory formulas.38 The activation energies were taken from
single-point calculations and the partition functions were
calculated from the B3LYP/6-31+G(d,p) moments of inertia
and scaled harmonic frequencies using the rigid rotor-harmonic
oscillator approximation.
Calculations
Standard ab initio and density functional theory calculations
were performed using the Gaussian 98 suite of programs.20
Geometries were optimized using Becke’s hybrid functional
(B3LYP)21 and the 6-31+G(d,p) basis set. For some species,
geometries were optimized with Hartree-Fock calculations, HF/
6-31G(d,p) and HF/6-31+G(d,p), as specified below. Spin-
unrestricted calculations (UB3LYP and UHF) were used for
open-shell systems. Spin contamination in the UB3LYP calcula-
tions was small as judged from the S2 operator expectation
values that were 0.75-0.77. The optimized structures were
characterized by harmonic frequency analysis as local minima
(all frequencies real) or first-order saddle points (one imaginary
frequency). Complete optimized structures in the Cartesian
coordinate format and total energies are available from the
correspondence author upon request. The B3LYP/6-31+G(d,p)
frequencies were scaled by 0.963 (ref 22, for other scaling
factors see ref 23), the UHF/6-31G(d,p) frequencies were scaled
by 0.893, and used to calculate zero-point vibrational energies
(ZPVE), enthalpy corrections, and partition functions. The rigid-
rotor harmonic oscillator approximation was used in all ther-
mochemical calculations. Single-point energies were calculated
at several levels of theory. In three sets of calculations, MP2-
(frozen core)24 and B3LYP energies were calculated with basis
sets of increasing size, e.g., 6-311+G(2d,p), 6-311+G(2df,p),
and 6-311+G(3df,2p). Spin contamination in the UMP2 cal-
culations was moderate for uracil radicals and transition states,
as evidenced by the spin expectation values S2 that ranged
between 0.76 and 0.91. Spin annihilation using Schlegel’s
projection method25 (PMP2)20 reduced the S2 values to 0.75-
0.76. In addition, restricted open-shell (ROMP2) calculations26
were carried out for the entire set of structures to deal with
spin contamination.27 The PMP2 and ROMP2 energies were
averaged with the B3LYP energies according to the empirical
procedure that was introduced previously14a,14b,28 and tested for
Results and Discussion
Specific generation of radicals by neutralization-reionization
mass spectrometry in the gas-phase relies on fast collisional
electron transfer to a cationic precursor. The precursor cation
is synthesized by the methods of gas-phase chemistry, acceler-
ated to a high velocity (118 625 m s-1 for 1+) and discharged
by a glancing collision with dimethyl disulfide as a thermal
electron donor (Scheme 2).19 Because the electron transfer must
occur over the distance of a few molecular diameters (5-10
Å),39 the time scale for the formation of 1 is limited to <8 fs.
This implies that the radical is initially formed with the geometry
of the precursor ion. Hence, ion 1+ must be prepared by a
reliable method and its structure must be examined carefully
by available experimental and computational methods.14

8342 J. Phys. Chem. A, Vol. 105, No. 36, 2001
Syrstad and Vivekananda
SCHEME 2
Unimolecular dissociations of 1+ proceeded mainly by
elimination of HNCO to form a C3H4NO+ ion at m/z 70
(measured 70.0295, calculated 70.0293) that dominated the
metastable-ion and CAD spectra (Table 1). Deuterium labeling
in 1a+-1c+ indicated a clean elimination of HNCO containing
the N-1 and/or N-3 nitrogens that was not accompanied by
hydrogen atom exchange between the NsH and CsH hydrogen
atoms (Table 1). Energy calculations confirmed that elimination
of HNdCdO containing C-2 and N-3 was the lowest-energy
dissociation of 1+ that required 165 kJ mol-1 to form HNd
CH-CH2-CO+ (9+) at the 0 K thermochemical threshold
(Table 2). Formation of the isomeric CH2dCH-NH-CO+
(10+) ion by elimination of HNdCdO containing C-4 and N-3
was slightly more endothermic and required 186 kJ mol-1 at
the 0 K thermochemical threshold. These ring-cleavage dis-
sociations may have activation energies above the dissociation
thresholds (see below for radical dissociations), but these were
not studied here. Loss of H atom from 1+ to form the uracil
cation-radical (11+•) was calculated to be substantially endo-
thermic (360 kJ mol-1, Table 2). The high threshold energy for
the loss of H explained why it did not occur competitively in
metastable-ion or collisionally activated dissociations.
TABLE 1: Collisionally Activated Dissociation Spectra of
Ions 1+ and 1a+-1c+
relative intensitya
relative intensitya
m/z 1+ 1a+ 1b+ 1c+
m/z
1+
1a+
1b+
1c+
114
113
112
97
0.9
0.6
45 0.9 5.7 0.3
0.7
0.2
0.2
44 1.4 4.4 5.3 3.0
43 5.7 5.0 4.8 4.8
41 3.3 3.3 2.2 2.9
40 3.0 1.9 1.0 1.5
39 1.0 0.7 0.5 0.6
38 0.7 0.7 0.8 0.6
1.1
87
0.6
Because electron ionization usually does not form ions with
a narrow distribution of internal energies, a more specific
ionization means was sought to generate 1+. The appearance
energy for the formation of 1+ from 4 was calculated by B3-
PMP2/6-311+G(2d,p). Assuming that 4 was thermalized at the
ion source temperature (523 K), the calculated appearance
energies (AE) ranged between 9.3 and 10.1 eV for the products
86
85
84
0.2
0.5
0.2
0.7
0.5
30
0.5
0.2
72
0.9 61.5b
0.5
29 0.5 5.6 1.7 4.4
28 3.8 1.5 4.9 1.4
27 1.0 1.0 1.3 0.8
26 0.7 0.8 0.5 0.4
71
0.7 59.2b
1.6 64.4b
2.7
0.8
70 65.6b
4.0
1.8
4.0
0.4
0.9
0.2
0.3
69
68
67
57
56
55
54
53
52
3.8
1.6
0.4
17
16
0.8
•
(1+ + C2H5 ) formed with internal energies corresponding to 0
0.6 0.5
0.4
and 523 K, respectively. Charge-transfer ionization with COS+•
(IE ) 11.18 eV)40 was therefore deemed to be sufficiently
exothermic to promote fast dissociation of 4+• to form 1+ that
would have a range of internal energies that was bracketed by
the ionization exothermicity, IE(COS) - AE(1+) ) 1.08-1.88
eV and the lowest dissociation threshold of the ion (vide supra).
In addition, the excess internal energy in the 1+ thus formed
(<1.9 eV) should be dissipated by collisions with COS in the
high-pressure CI source. Fully thermalized 1+ was calculated
to have 43 kJ mol-1 mean internal energy at 523 K.
15 0.7 0.5
14
0.4
0.4
0.4
0.5
0.5
0.3
0.3
0.3
a Relative to the sum of CAD ion intensities. b Dominant product
of metastable ion dissociations.
Ion Structures and Energetics. Cation 1+ was generated
efficiently by 70 eV dissociative ionization of 6-ethyl-5,6-
dihydrouracil (4, Scheme 2). Deuterium-labeled ions 1a+-1c+
were prepared accordingly using the labeled precursors 4a-
4c. Ion 1+ showed a singlet peak in a high-resolution mass
spectrum that had the correct elemental composition, C4H5N2O2
(measured 113.0348, calc. 113.0351), as established by accurate
mass measurements. Ion 1+ was further characterized by
collisionally activated dissociation (CAD) mass spectra.
The relative energies of 1+ and several other ion isomers were
calculated at a high level of theory recently.41,42 In the context
of the present work, we note that 1+ is only the fifth most stable
isomer in the family of protonated uracil cations and is 82 kJ
mol-1 less stable than N-1 protonated 2,4-dihydroxypyrimidine,
which is the most stable ion isomer.41,42 Regarding the other
relevant ion isomers, it is noteworthy that cation 2+ was unstable
TABLE 2: Ion Dissociation Energies
energya
species/reaction
1+ f 9+ + HNdCdO
basis set
B3LYP
MP2b
164
164
163
B3-MP2b
QCISDc
169
169
169
QCISD(T)c
6-31+G(d,p)
169
159
158
156
191
169
163
161
351
348
346
347
6-311+G(2d,p)
6-311+G(2df,p)
6-311+G(3df,2p)
6-31+G(d,p)
161
161
159
166
166
165
1+ f 10+ + HNdCdO
1+ f 11+• + H•
6-311+G(2d,p)
6-311+G(2df,p)
6-311+G(3df,2p)
6-31+G(d,p)
6-311+G(2d,p)
6-311+G(2df,p)
6-311+G(3df,2p)
177
178
179
173
170
170
187
188
189
183
184
186
351 (359)d
350 (357)
355 (362)
349 (353)d
348 (352)
351 (354)
358 (358)d
357 (356)
361 (361)
356 (356)d
355 (355)
360 (361)
a In units of kJ mol-1 at 0 K. b From spin-projected MP2 energies wherever it applies. c From effective energies with basis sets expanded according
to eq 1. d From restricted open-shell MP2 calculations.

Hydrogen Atom Adduct to C-5 in Uracil
J. Phys. Chem. A, Vol. 105, No. 36, 2001 8343
Figure 2. Neutralization (CH3SSCH3, 70% transmittance)-reionization
(O2, 70% transmittance) mass spectra of (a) 1a+, (b) 1c+, (c) 1b+. The
insets show expanded regions of survivor ion peaks from separate
multiscan measurements.
Figure 1. (a) Neutralization (CH3SSCH3, 70% transmittance)-reion-
ization (O2, 70% transmittance) mass spectrum of 1+. (b) Neutralization
(CH3SSCH3, 70% transmittance)-collisional activation (He, 50% trans-
mittance)- reionization (O2, 70% transmittance) mass spectrum of 1+.
predominant loss of D (Figure 2c), in line with the foregoing
results. A small fraction of loss of H (ca. 15%) from 1b+ can
be attributed to an admixture of 1,5-D2 or 3,5-D2 isotopomers
that were not distinguished by mass from 1b+. The correspond-
ing radicals were expected to eliminate H-5 rather than D-5,
because of isotope effects (see below). The specific loss of H
from the C-5 methylene group in 1 indicated exclusive formation
of the most stable uracil tautomer 11 by NR dissociations.
and isomerized spontaneously to 1+.41 It can be readily shown
that an isomerization of 1+ by intramolecular proton migration
to form a more stable cation isomer must involve at least one
four-membered transition state and therefore is likely to face a
substantial energy barrier.41 Ion 1+ is therefore kinetically well
stabilized against isomerization to more stable structures.
Formation and Dissociations of Radical 1. Collisional
neutralization of 1+ yielded a fraction of nondissociating radicals
that following reionization gave rise to abundant survivor ions
that appeared at m/z 113 in the NR mass spectrum (Figure 1a).
The dissociations observed upon NR comprised loss of H (m/z
112) and ring-cleavages producing HNCO (m/z 43) together with
the complementary C3H4NO fragment at m/z 70, C3H3NO at
m/z 69, fragments of the C2HxN and CHxN groups at m/z 38-
42 and m/z 26-28, respectively, and CO at m/z 28.
The ring cleavage yielding C3H4NO showed predominant
retention of H-1 and/or H-3 in the eliminated HNCO molecule,
as evidenced by the appropriate mass shifts of m/z 70 to m/z 71
in the complementary C3(H,D)4NO fragments (Figures 2a-2c).
The HNCO molecule lost can include the C-2 carbonyl group
in combination with the N-1-H or N-3-H groups, or the C-4
carbonyl and the N-3-H group. These were not distinguished
by the present labeling experiments. Energy arguments favor
cleavage of the C-4sC-5 and C-2sN-3 bonds to eliminate
HNCO containing N-3 and C-4, as discussed below. Deuterium
labeling further revealed that the m/z 28 fragment consisted of
CO and HNCH, where the latter contained the C-6 methine
group.
The NR spectrum of 1+ generated by charge exchange
ionization with COS+• of 4 was closely similar to the spectrum
of the ion made by 70 eV electron ionization and is not shown
separately. This indicated that effects on NR dissociations of
1+ of the precursor ion energy were negligible or that stable 1+
produced by electron ionization had internal energies close to
thermal. Collisional activation with helium of the intermediate
fast neutral species resulted in deeper dissociations that enhanced
the formation of small fragments in the NCR spectrum (Figure
1b).
The dissociation mechanisms were elucidated by deuterium
labeling. NR of the 1,3-D2-labeled ion 1a+ showed a clean
(>95%) loss of a light hydrogen from C-5 and/or C-6 (Figure
2a). These positions were distinguished in the 6-D-labeled ion
1c+ which upon NR showed clean (>99%) loss of a light
hydrogen (Figure 2b). NR of the 5,5-D2-labeled ion 1b+ showed
Dissociation Products of 1. To characterize some of the
important fragments produced by dissociation of 1 we obtained
reference NR spectra of the uracil cation radical (11+•, m/z 112),
C3H4NO+ (m/z 70) prepared by dissociative ionization of 4, and
C3H3NO+• (m/z 69) that was prepared from two different
precursors, e.g., by dissociative ionization of uracil and male-
imide by loss of HNCO and CO, respectively. The NR spectra
are shown in Figure 3a-d. NR of uracil showed an abundant
survivor ion in keeping with the stability of the neutral molecule
(Figure 3a). The major dissociation was loss of HNCO to form
C3H3NO at m/z 69. This is also an abundant dissociation of
cation-radical 11+• upon electron ionization and its occurrence

8344 J. Phys. Chem. A, Vol. 105, No. 36, 2001
Syrstad and Vivekananda
C3H3NO would be detected efficiently in the NR mass spectrum.
We did not investigate the structure of the C3H3NO+• fragment
further, although the NR dissociations were compatible with
both the HNdCH-CHdCdO+• and 3-propenelactam+• struc-
tures.
In contrast to C3H3NO, the NR spectrum of C3H4NO+ (m/z
70) showed only a weak survivor ion and a dominant formation
of CO (m/z 28) and CH2CHdNH (m/z 42) (Figure 3b). This
attested to the low stability of the C3H4NO radical in keeping
with the calculated bond dissociation energy for CH2dCH-
NH-CdO• f CH2-CHdNH• + CO that required only 7 kJ
mol-1 at the 0 K threshold. The radical energetics is discussed
in more detail later in the paper. The low stability of neutral
C3H4NO corroborates the conclusion that the prominent m/z 70
peak in the NR spectrum of 1 must be due to a stable ion formed
by a post-reionization dissociation of 1+. Conversely, however,
transient formation of CH2dCH-NH-CdO• from 1 followed
by its consecutiVe dissociations is compatible with the formation
upon NR of abundant low-mass fragments HNCO, CH2-CHd
NH•, and CO.
In summarizing the observed unimolecular dissociations of
1, loss of a hydrogen atom from C-5 was a specific dissociation
of the radical. Ring cleavages also occurred in 1 and gave rise
to small-mass fragments HNCO, CH2-CHdNH, and CO.
Dissociation Energetics of 1. To interpret the observed
dissociations of uracil radicals, we performed ab initio and
density functional calculations of the pertinent relative, activa-
tion, and dissociation energies. The energies were obtained at
several levels of theory as summarized in Table 3. The energies
obtained at the highest levels of theory, which in this work were
effective QCISD(T)/6-311+G(3df,2p) and B3-PMP2/6-311+G-
(3df,2p), are visualized in a potential energy diagram (Figure
4). At all levels of theory 1 was the most stable uracil radical,
followed by 5,6-dihydropyrimidine-2,4(1H,3H)-dion-5-yl (2),
3, 4,5-dihydro-2-hydroxy-pyrimidine-4(3H)-on-2-yl (12, Figure
Figure 3. Neutralization (CH3SSCH3, 70% transmittance)-reionization
(O2, 70% transmittance) mass spectra of dissociation products: (a)
[uracil]+•, (b) m/z 70 ion from dissociative ionization of 4, (c) m/z 69
ion from dissociative ionization of uracil, and (d) m/z 69 ion from
dissociative ionization of maleimide.
in the NR spectrum of 11+• can be due to a combination of
neutral and post-reionization dissociations. Interestingly, NR of
1+ and 11+• showed very similar [m/z 69]/[m/z 112] abundance
ratios, 0.32 and 0.31, respectively, which indicated that the
formation of the m/z 69 fragment from 1 was due to consecutive
dissociations 1 f 11 f C3H3NO+•. In addition, one can use
the relative abundances of m/z 112, m/z 69, and other ions in
the NR spectrum of 11+• to subtract the contributions of
dissociations of reionized 11+• in the NR spectrum of 1.43 These
amounted to 45% of the total NR ion intensities and cor-
responded to the fraction of 1 that dissociated by loss of H-5 to
form 11. The remaining 55% of the NR fragment ion intensities
were due to ring cleavage dissociations of 1 or post-reionization
dissociations of 1+. The contributions of the latter two processes
were roughly estimated as follows. Assuming that the relative
intensity of the m/z 70 ion in the NR spectrum (12.5%) was
due entirely to dissociations of reionized 1+ (see below), we
obtain from the corresponding CAD relative intensity of the
m/z 70 ion (60%) that 12.5/0.6 ≈ 21% of all ring fragments
originated from reionized 1+.43 The contribution of the radical
ring-cleavage dissociations was therefore 55-21 ) 34%. After
this separation of ion and radical dissociations, the branching
ratio for ring cleavages and loss of H-5 in neutral 1 is estimated
as 34/45 ≈ 0.75.
5), and 2-hydroxy-5,6-dihydropyrimidine-4(3H)-on-5-yl.12
A
number of other tautomeric radical structures were also calcu-
lated to be local energy minima, which are not discussed here.
Out of these, radicals 1, 2, and 3 were hydrogen atom adducts
to the most stable uracil tautomer 11, which were of most
interest for uracil radical chemistry.
The optimized structure of 1 was unexceptional, as shown
in Figure 5. The ring in 1 was only slightly puckered with C-5
being out of plane and the axial hydrogen atom (H-5a) pointing
nearly perpendicular to the ring. The radical site in 1 was
localized mainly at C-6, as indicated by the calculated atomic
spin density (0.92). The unpaired electron can activate cleavage
of bonds to the adjacent atoms, e.g., N-1sC-2, N-1sH-1, C-5s
H-5, and C-4sC-5. Dissociation products resulting from cleav-
age of these bonds were therefore of interest. Loss of H-5a was
the lowest-energy dissociation of 1 that required 114 kJ mol-1
at the 0 K thermochemical threshold (Figure 4, Table 3). The
dissociation involved a transition state (TS1) that was 133 kJ
mol-1 above 1 and 19 kJ mol-1 above the products. Hence,
addition of a H atom to C-5 in uracil had an activation energy
barrier. Dissociation of the N-1sH-1 bond in 1 was 177 kJ
mol-1 endothermic to form the less stable uracil tautomer 13.41,42
In addition, the dissociation must proceed through a transition
state (TS2) that was 203 kJ mol-1 above 1 and 26 kJ mol-1
above the threshold energy for the formation of 13 and H atom.
The large energy difference between TS2 and TS1 strongly
favors elimination of H-5a in perfect agreement with experiment.
A transition state (TS3) was also located for the 1,2-migration
of a hydrogen atom from C-5 to C-6 that resulted in a mildly
The NR spectra of the C3H3NO+• ions prepared by dissocia-
tive ionization of uracil and maleimide were very similar and
indicated identical ion structures or mixtures thereof (Figure
3c-d). The substantial survivor ion at m/z 69 indicated that the
C3H3NO intermediate was a stable species. This indicates that
if formed as a neutral molecule by dissociations of radical 1,

Hydrogen Atom Adduct to C-5 in Uracil
J. Phys. Chem. A, Vol. 105, No. 36, 2001 8345
TABLE 3: Dissociation and Activation Energies for Uracil Radicals 1 and 2
energya,b
reactionc
method
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)
B3LYP/6-31+G(d,p)
133
127
126d
125
125
100
102d
96
140
137
130
135
135
124
124
120
127
119
115
122
118
203
196
214
210
163
162
111
107
148
144
124
121
156
154
148
130
184
165
209
200
208
197
125
118
119
116
116
89
137
133
B3LYP/6-311+G(2d,p)
B3LYP/6-311+G(2df,p)
B3LYP/6-311+G(3df,2p)
PMP2/6-311+G(2d,p)
198
198
165
211
212
199
161
161
165
128
126
148
168
166
170
131
131
119
105
117
143
151
124
120
156
154
198
222
200
209
93
ROMP2/6-311+G(2d,p)
161
194
160
116
154
119
152
147
168
219
212
84
117
98c
96
87
PMP2/6-311+G(2df,p)
ROMP2/6-311+G(2df,p)
PMP2/6-311+G(3df,2p)
ROMP2/6-311+G(3df,2p)
QCISD/6-311+G(2d,p)
PMP2
166
162
172
168
201
195
204
199
161
155
161
155
155
153
155
154
184
182
183
181
84
114
112
117
115
92
101
97
79
120
119
156
158
130
129
162
161
229
226
227
231
88
83
122
144
145
145
144
178
179
208
210
179
178
117
119
157
156
115
116
151
152
153
154
170
167
236
235
176
177
112
112
141
141
123d
122
ROMP2
123d
QCISD/6-311+G(2df,p)
PMP2
117
118
140
140
180
180
210
211
174
173
159
160
183
184
107
107
136
137
ROMP2
QCISD/6-311+G(3df,2p)
PMP2
122
123
142
143
185
186
213
215
174
173
120
121
161
160
124
125
159
161
160
161
182
183
243
242
195
196
111
112
139
139
ROMP2
QCISD(T)/6-311+G(2d,p)
PMP2
113
114
134
135
170
170
199
200
173
171
114
115
150
149
112
113
144
146
147
148
165
165
220
219
179
179
102
102
130
131
ROMP2
QCISD(T)/6-311+G(2df,p)
PMP2
109
109
130
131
172
172
200
202
168
166
154
155
179
179
97
97
126
126
ROMP2
QCISD(T)/6-311+G(3df,2p)
PMP2
114
114
114
114d
110
113
133
133
130
127
127
129
177
178
180
203
205
205
168
166
164
116
118
112
154
153
148
121
122
120
153
154
154
154
155
139
178
178
168
227
226
211
197
198
203
101
102
103
106
100
102
128
129
126
ROMP2
B3-PMP2/6-311+G(2d,p)
B3-PMP2/6-311+G(2df,p)
B3-PMP2/6-311+G(3df,2p)
182
185
206
208
161
161
141
140
176
175
123
124
112
149
127
159
213
213
a In units of kJ mol-1 at 0 K. b From single-point energies on B3LYP/6-31+G(d,p) optimized geometries including zero-point vibrational energy
c
corrections. Reaction 1: 1 f 11 + H•. Reaction 2: 1 f TS1. Reaction 3: 1 f 13 + H•. Reaction 4: 1 f TS2. Reaction 5: 1 f TS3. Reaction
6: 1 f 14a. Reaction 7:1 f TS5. Reaction 8: 1 f 16b. Reaction 9: 1 f TS6. Reaction 10: 1 f 10 + HNCO. Reaction 11: 1 f 9 + HNCO.
Reaction 12: 1 f TS7. Reaction 13: 1 f 17 + CO. Reaction 14: 2 f 11 + H•. Reaction 15: 2 f TS4. d From single-point energies on
HF/6-31G(d,p) optimized geometries (HF/6-31+G(d,p) for transition states) and zero-point energy corrections.
tion should be kinetically preferred. This was supported by
RRKM calculations of unimolecular rate constants that were
performed using both the effective QCISD(T)/6-311+G(3df,2p)
and B3-PMP2/6-311+G(3df,2p) potential energy surfaces. In
each case, the loss of H-5 was 2-3 orders of magnitude faster
than the isomerization for reactant internal energies up to 200
kJ mol-1 above the threshold (Figure 6). Loss of a deuterium
atom from C-5 in 1b was also substantially faster than migration
to C-6, as indicated by the pertinent RRKM rate constants (not
shown). These findings were perfectly consistent with the results
of deuterium labeling that showed a clean loss of a deuterium
atom from C-5. Had isomerization to 2 occurred, H and D would
have become equivalent at the C-6 methylene in the intermediate
radical 2 and therefore would have been lost competitively.
Interestingly, the transition state for the hydrogen atom loss from
C-6 in 2 (TS4) was below TS3, so that loss of H-6 could also
Figure 4. Potential energy diagram for dissociations of 1 and 2. Roman
be expected to occur without isomerization to 1 (Figure 4).
numerals: B3-PMP2/6-311+G(3df,2p) energies. Bold numerals: Ef-
Mechanisms for Ring Opening in 1. Ring cleavage dis-
fective QCISD(T)/6-311+G(3df,2p) energies from basis set expansions
sociations of 1 were all endothermic (Figure 4). Dissociation
of the C-4sC-5 bond proceeded with concomitant rotation of
the incipient vinyl group about the N-1sC-6 bond to produce
an open-ring intermediate, anti-syn-syn-anti-2,4-diazahex-5-en-
1,3-dion-1-yl (14a, Scheme 3) that was 116 kJ mol-1 less stable
than 1 (Table 3). This reaction path was confirmed by intrinsic
reaction coordinate calculations44 from the transition state for
through PMP2 calculations.
endothermic isomerization of 1 to 2. However, the TS3 energy
(168 kJ mol-1 above 1) was substantially higher than that for
TS1 (Table 3).
The large difference in the activation energies for the loss of
H-5a and isomerization to 2 indicated that the former dissocia-

8346 J. Phys. Chem. A, Vol. 105, No. 36, 2001
Syrstad and Vivekananda
Figure 5. B3LYP/6-31+G(d,p) optimized structures of uracil radicals in ground electronic state. Bond lengths in ångstroms, bond and dihedral
angles in degrees. The structure of the 2A state of 1 was optimized with UCIS/6-31+G(d,p). The bond lengths in the 2A state that differed most from
2
those in the X state are shown as bold italics.
the ring opening (TS5). Radical 14a can isomerize by internal
rotations to produce several other conformers (Scheme 3). Note
that because there are four bonds for internal rotations in 14a,
e.g., N-3sC-4, C-2sN-3, N-1sC-2, and N-1sC-6, there are
16 theoretically possible combinations of conformers, of which
the anti-syn-syn-syn conformer is identical with 1, and the all-
syn conformer is impossible for steric reasons. We calculated
the relative energies of the syn-syn-syn-anti (14b), anti-anti-
syn-anti (14c), anti-anti-syn-syn (14d), all-anti (14e), and syn-
syn-anti-anti (14f) conformers, which all had at least one
favorable alignment of the CsO and NsH bond dipoles. The
0 K enthalpies of the conformers relative to 14a were -1, 2, 4,
-6, and -31 kJ mol-1 for 14b, 14c, 14d, 14e, and 14f,
respectively.
Because the open-ring radicals 14a-14f represent isomers of
1 and not detectable dissociation products, the energetics of their
further dissociations were also investigated. A dissociation of
the C-2sN-3 bond in 14a-d and 14f can produce anti-anti-
CH2dCH-NH-CO• (10a) and HNCO that was calculated to
require 154 kJ mol-1 from 1. syn-anti-CH2dCH-NH-CO•
(10b) was 9 kJ mol-1 less stable than the anti-anti conformer
10a, so that the formation of 10b had a correspondingly higher
threshold energy. Dissociation of 1 via 14b by loss of the C-4s
O carbonyl to form the anti-anti conformer of the N-vinyl-N′-
ureidyl radical, CH2dCH-NH-CO-NH• (15, Scheme 3),
required 197 kJ mol-1 at the 0 K thermochemical threshold and
was therefore substantially more endothermic than the loss of
HNCO.
Another mechanism for ring cleavage in 1 comprised dis-
sociation in 1 of the N-1sC-2 bond with concomitant rotation
about the C-4sC-5 and C-5sC-6 bonds to produce an open-
ring intermediate, anti-anti-syn-anti-1,5-diazahex-1-en-4,6-dion-
6-yl (16a) (Scheme 4), that was 121 kJ mol-1 less stable than
1. The (Z)-syn-syn-anti-anti-(16b) and all-anti (16c) conformers
were also stable structures whose respective 0 K enthalpies were
1 and 20 kJ mol-1 relative to 16a. A further dissociation of
16a to HNdCH-CH2-CO• (9) and HNCO required 58 kJ
mol-1 so that the product threshold energy was 178 kJ mol-1
above 1. Dissociation of 1 via 16a by loss of the C-2-O carbonyl
to yield the syn-anti conformer of the 3-imidopropionamidyl
radical (17, Scheme 4) was substantially endothermic and
required 230 kJ mol-1 at the 0 K thermochemical threshold
(Figure 4).
In addition to their inherent endothermicities, the ring
cleavages in 1 involved additional activation barriers that were
calculated as 154 (TS5) and 153 kJ mol-1 (TS6) for the
cleavages of the C-4sC-5 and N-1sC-2 bonds, respectively
(Figure 4). Hence, the activation energies for the ring cleavages

Hydrogen Atom Adduct to C-5 in Uracil
J. Phys. Chem. A, Vol. 105, No. 36, 2001 8347
SCHEME 4
The data in Table 3 (reactions 1, 3, 11, 13, and 14) showed
that B3LYP calculations mostly oVerestimated bond dissociation
energies in uracil radicals for basis sets ranging from 6-31+G-
(d,p) through 6-311+G(3df,2p). In contrast, PMP2 calculations
with these basis sets mostly underestimated bond dissociation
energies. Simple averaging of the B3LYP and PMP2 energies
(denoted B3-PMP2) thus often resulted in error cancellation
and provided excellent agreement with the benchmark dissocia-
tion energies. In addition, the fit between the B3-PMP2 and
benchmark energies showed only a weak dependence on the
basis set used, so that even the relatively inexpensive45 B3-
PMP2/6-311+G(2d,p) dissociation energies were within 4 kJ
mol-1 root-mean-square deviation of the benchmark data.
Experimental data are scarce to provide a rigorous check for
the calculated parameters. We note, however, that the calculated
adiabatic ionization energy of uracil, 9.23 and 9.25 eV from
B3-PMP2/6-311+G(2d,p) and effective QCISD(T)/6-311+G-
(3df,2p), respectively, (Table 4) reproduced very well the
experimental IEa (9.2 eV) that was extrapolated from the
photoelectron spectrum.46 Likewise, the B3-PMP2/6-311+G-
(2d,p) electron affinity of uracil (0.055 eV) compares well with
the experimental values, 0.054 eV47 and 0.093 eV.48
Figure 6. RRKM unimolecular rate constants for dissociations and
isomerizations of 1. Full circles: Dissociation of the C-5sH bond.
Empty circles: Ring-opening by cleavage of C-4sC-5 bond (k1).
Upside-down triangles: Ring closure 14a f 1 (k-1). Squares: Dis-
sociation 14a f 10a + HNCO (k2). Diamonds: Ring-opening by
cleavage of N-1sC-2 bond (k3). Triangles: Rotation in 14a f 14b.
Hexagons: Isomerization 1 f 3. The dashed line at log k ) 5.13
corresponds to the rate constant for 50% dissociation on the experi-
mental time scale.
SCHEME 3
The transition state energies calculated with B3LYP/6-31+G-
(d,p) showed very good agreement with the benchmark energies
(reactions 2, 4, 5, 7, 9, 12, and 15, Table 3). However, because
B3LYP calculations overestimated bond dissociation energies,
they severely underestimated the barriers for hydrogen atom
additions, as noted previously for other systems.49
The B3-PMP2 transition state energies were mostly in very
good agreement with the benchmark data. The largest deviation
was for TS7 (reaction 12, Table 3) which was underestimated
by B3-PMP2 by 13-16 kJ mol-1. However, the overall
agreement with the benchmark energies, including reaction 12,
was within 9 kJ mol-1 rmsd. We also note that effective QCISD
calculations consistently overestimated bond dissociation and
transition state energies by 9 kJ mol-1 rmsd. These deviations
canceled out in the calculations of activation energies for
hydrogen atom additions for which QCISD performed at the
benchmark level. It may also be noted that relative energies
from spin-projected MP2 calculations did not differ significantly
from the ROMP2-based values in B3-MP2 or effective QCISD-
(T) calculations.
in 1 were greater than that for loss of H-5a, but comparable to
that for isomerization to 3.
Comparison of Radical Dissociation and Activation Ener-
gies Calculated at Different Levels of Theory. The data in
Table 3 allowed us to compare the performance of the various
schemes used to calculate the total of 15 dissociation and
activation energies for uracil radicals. As benchmark values,
we chose the effective QCISD(T)/6-311+G(3df,2p) energies
with which the other energies, calculated at presumably lower
levels of theory, were compared. The value of such comparisons
is in identifying schemes that equal or approach the performance
of the benchmark scheme, but are computationally less demand-
ing and therefore practical for use with even larger open-shell
systems.
In summary, inexpensive B3-PMP2/6-311+G(2d,p) calcula-
tions provide very good estimates of bond dissociation and
transition state energies in uracil radicals when compared with
benchmark effective QCISD(T)/6-311+G(3df,2p) data. Other

8348 J. Phys. Chem. A, Vol. 105, No. 36, 2001
TABLE 4: Ionization and Franck-Condon Energies
Syrstad and Vivekananda
was faster than the loss of HNCO in open-ring intermediates
14a and 14b having up to 230 kJ mol-1 internal energy. Hence,
except for a fraction of highly energetic radicals, the dissociation
was the rate-determining step. RRKM calculations of the rate
constants for the dissociation of 14b (k2) and ring closure in
14a (k-1) showed k2/(k-1 + k2) values ranging between 0 and
0.94 for internal energies from TS7 up to 260 kJ mol-1 above
14a (379 kJ mol-1 above 1). Note the crossover of the log k
curves for k2 and k-1 at E ) 323 kJ mol-1 (Figure 6). Hence,
the upper bound for the branching ratio for the ring cleavage
dissociations vs loss of H is given simply by k1/kH. The
calculated RRKM rate constants indicate that, at all levels of
theory, 0 < k1/kH < 0.25 for 1 having internal energies between
that of the dissociation threshold and 380 kJ mol-1. Hence, keff/
kH < 0.25 and the loss of H-5a should predominate at all
internal energies.52
ionization
energya
Franck-Condon
energy
method
1
11
1
VNb
VIc
62
B3LYP/6-31+G(d,p)
6.96 9.21
6.92 9.21
6.93 9.22
6.92 9.22
6.66 9.25
6.61 9.33
6.70 9.34
6.66 9.41
6.72 9.35
6.68 9.43
6.70 9.14
6.74 9.23
6.77 9.24
6.63 9.14
6.68 9.23
B3LYP/6-311+G(2d,p)
B3LYP/6-311+G(2df,p)
B3LYP/6-311+G(3df,2p)
PMP2/6-311+G(2d,p)
46
45
45
54
53
52
52
52
52
63
65
ROMP2/6-311+G(2d,p)
PMP2/6-311+G(2df,p)
ROMP2/6-311+G(2df,p)
PMP2/6-311+G(3df,2p)
ROMP2/6-311+G(3df,2p)
QCISD/6-311+G(2d,p)d
QCISD/6-311+G(2df,p)d
QCISD/6-311+G(3df,2p)d
QCISD(T)/6-311+G(2d,p)d
QCISD(T)/6-311+G(2df,p)d
65
Discussion of Radical Dissociations: The Role of Excited
Electronic States. The calculated potential energy surfaces
suggested that elimination of a hydrogen atom should dominate
the unimolecular dissociations of radical 1. In the discussion
we now compare the predicted dissociations with those observed
for radicals generated by collisional electron transfer.
QCISD(T)/6-311+G(3df,2p)d 6.70 9.25
B3-PMP2/6-311+G(2d,p)
B3-PMP2/6-311+G(2df,p)
B3-PMP2/6-311+G(3df,2p)
6.79 9.23
6.81 9.28
6.82 9.28
50
49
49
64
64
The NR dissociations of 1 showed specific loss of a hydrogen
from C-5 that was not accompanied or preceded by isomeriza-
tion to 2. This was in keeping with the RRKM calculated rate
constants for these competitive reactions that predicted the loss
of H-5a to be substantially faster (Figure 6). Note that the
absence of unimolecular isomerization 1 T 2 in the gas phase
contrasts the acid-base-catalyzed isomerization of uracil radi-
cals in solution.5 Referring back to the present gas-phase data,
the calculated rate constants also indicated that ring cleavage
dissociations should not compete efficiently with the loss of
H-5a, which was in sharp contrast with NR dissociations of 1
in which ring-cleavage dissociations producing HNdCdO and
CH2CHdNH accounted for about 34% products.
a Adiabatic ionization energies at 0 K in electronvolts. b Energy
difference in units of kJ mol-1 between the vertically neutralized radical
and the optimized neutral structure. c Energy difference in units of kJ
mol-1 between the single-point energy of the cation with the optimized
radical structure and the energy of the optimized cation structure.
d Effective energies from basis set expansions (eq 1).
schemes, e.g., B3LYP and effective QCISD, showed systematic
bias in dissociation or transition state energy calculations.
Kinetics of Competing Dissociations of 1. The effective
QCISD(T) and B3-PMP2 transition state energies were used
in RRKM calculations of the unimolecular rate constants for
loss of H-5a and ring cleavage dissociations in 1. Figure 6 shows
the RRKM rate constants based on the effective QCISD(T)
energies; those based on the B3-PMP2 energies showed very
similar trends and are not discussed separately. The dissociation
to anti-anti-CH2dCH-NH-CO• (10a) and HNCO proceeds in
a minimum of three steps, e.g., 1 T 14a f 14b f 10a, where
the ring-opening is reversible. The rate of formation of 10a is
in general time-dependent,50 but can be approximated by the
time-independent Bodenstein (steady-state) formula51 through
an effective rate constant, keff ) k1k2/(k-1 + k2), where k1 is
the rate constant for ring opening by C-4sC-5 bond dissociation
in 1, k-1 is the rate constant for the reverse ring closure, and k2
is the rate constant for the dissociation of the open-ring
intermediate 14a. The branching ratio for the ring cleavage
dissociation and loss of H-5a is given by keff/kH, where kH is
the rate constant for the hydrogen atom loss. Exact evaluation
of the k2/(k-1 + k2) term to find the kinetically most faVorable
dissociation route would be a daunting task because of the
multitude of conformers in the open-ring intermediates 14a-f,
rotational barriers for their interconversion (Figure 4), and the
multitude of transition states for the loss of HNCO from the
conformers. We have investigated the potential energy profiles
for dissociations of the C-2sN-3 bonds in 14a-14f, which all
showed potential energy barriers above the dissociation threshold
corresponding to the formation of 10a. As a case in point, a
transition state (TS7) was obtained for the loss of HNCO from
14b that was 112 kJ mol-1 above 14b and 227 kJ mol-1 above
1 (Figure 4). This was significantly higher than the barrier for
intramolecular rotation about the CO-NH bond interconverting
14a and 14b, which was 41 kJ mol-1 above 14a. According to
RRKM calculations (Figure 6), the 14a T 14b interconversion
Another question related to the NR dissociations concerned
the internal energy in the radicals formed by collisional electron
transfer that drives the observed dissociations. We have shown
previously for related heterocyclic radicals14a,14b that the vibra-
tional energy in the species that were formed in the ground
electronic state was convoluted from the internal energy of the
precursor cation and the Franck-Condon energy acquired upon
vertical electron transfer. With 1, the sum of the ion mean
thermal energy, e.g., 43 kJ mol-1 for 1+ generated by charge-
exchange ionization followed thermalization at 523 K, and the
Franck-Condon energy (Table 4) results in an estimated 43 +
49 ) 92 kJ mol-1 mean vibrational energy in the radical.14a
This is substantially less than the transition state energy for the
lowest-energy dissociation (TS1, 133 kJ mol-1, Table 3). In
addition, the loss of H-5a showed ca. 37 kJ mol-1 kinetic shift
to attain rate constants allowing 50% dissociation on the 5.1 µs
time scale. The internal energy needed for a kinetically relevant
loss of H-5a can possibly be accessed in a fraction of radicals
belonging to the high-energy tail of the internal energy
distribution curve. However, a dissociation by ring-cleavage,
including the TS5 energy and an associated 43 kJ mol-1 kinetic
shift,53 would require that dissociating 1 contain >190 kJ mol-1
internal energy, which can be accessed only by an extremely
small fraction of the highest-energy radicals. Clearly, these
cannot account for the ca. 34% of ring dissociations observed.
The competition between the loss of hydrogen atom and ring
cleavage dissociations is thus incompatible with the estimated
internal energy distribution in radicals formed in the ground
electronic state.

Hydrogen Atom Adduct to C-5 in Uracil
J. Phys. Chem. A, Vol. 105, No. 36, 2001 8349
TABLE 5: Excited State Energies of Radical 1
radical
optimized
geometry
∆Ea
TD-B3LYPe
f b
TD-B3LYP
τ(µs)
TD-B3LYP
configurationc
state
CISd
CIS
CIS
CIS
TD-B3LYP
(X) 1f
A
B
C
D
E
A
B
C
D
E
A
B
C
D
E
5.26
5.64
5.72
6.03
6.36
3.66
4.95
5.04
5.08
5.42
3.84
5.00
5.03
5.13
5.38
3.21
3.76
4.37
4.51
4.73
2.31
2.47
3.40
3.63
4.01
2.02
2.60
3.42
3.60
3.94
0.002
0.0004
0.001
0.011
0.001
0.0014
0.0001
0.031
0.001
0.010
0.0013
0.031
0.000
0.0007
0.0001
0.010
0.0015
0.001
0.007
0.0003
0.012
0.001
0.000
0.0001
0.005
0.011
0.0009
0.0001
0.000
0.040
0.4
1.8
0.7
0.06
0.6
1.2
9.4
0.03
0.9
0.08
1.2
0.2
1.1
1.2
0.2
3.4
0.4
3.8
30R f 31R
mixed R,âh
mixed R,â
mixed R,â
mixed R,â
30R f 31R
mixed Rh
mixed R
30R f 33R
mixed R
30R f 31R
mixed R
mixed R
30R f 31R
30R f 32R
29â f 30â
30R f 33R
30R f 34R
30R f 31R
30R f 32R
30R f 33R
30R f 34R
30R f 35R
30R f 31R
30R f 32R
30R f 33R
30R f 34R
30R f 35R
(A) 1i
1+j
>20
17.5
0.3
0.5
3.8
20
0.03
>20
1.25
8.0
>20
30R f 33R
mixed R
0.04
a Excitation energies in units of electronvolt. b Oscillator strengths. c Dominant configurations with expansion coefficients >0.85. d From spin
unrestricted UCIS/6-311+G(2d,p) single-point calculations. e From time-dependent UB3LYP/6-311+G(2d,p) single-point calculations. f UB3LYP/
6-31+G(d,p) optimized geometry of the ground electronic state. g From time-dependent UB3LYP/6-31+G(d,p) calculations. h Configurations consisting
i
of several R and/or â electron excitations. UCIS/6-31+G(d,p) optimized geometry of the excited A state. j B3LYP/6-31+G(d,p) optimized geometry
of the cation.
It has been pointed out recently that excited electronic states
play an important role in the dissociations of radicals formed
by femtosecond collisional electron transfer.14a,14b,54 This is due
to the fact that the incoming electron can enter an unoccupied
orbital in the cation or cause excitation of a valence electron to
form the radical in an excited electronic state. Bright excited
states that have short (<ns) radiative lifetimes compared to the
µs time scale for dissociations will emit a photon and decay to
the ground electronic state without affecting the dissociation
kinetics. In contrast, photochemically dark states that have µs
or longer radiative lifetimes can dissociate on the corresponding
potential energy surfaces or undergo internal conversion through
avoided crossings or conical intersections to a vibrationally
excited lower electronic state. Such dark excited states can be
expected to affect the dissociation kinetics by opening new
dissociation channels or by funneling the electronic energy to
vibrational energy of the ground electronic state.
Excited States of 1. We have addressed the energetics of
excited electronic states in 1 by time-dependent density func-
tional theory (TD-B3LYP) and configuration interaction singles
(CIS) calculations. The former approach has been shown to
provide reliable excitation energies for several molecular
systems, whereas the energies from CIS calculations are
typically overestimated and do not converge to the ionization
limit.15 CIS calculations, combined with gradient optimizations,
were used to obtain geometries for local energy minima of the
first excited state (2A) that in turn served for single-point energy
calculations with TD-B3LYP. The calculated energies, oscillator
strengths, radiative lifetimes, and dominant configurations of
the five lowest excited states of 1 are summarized in Table 5,
Figure 7. TD-B3LYP/6-311+G(2d,p) excitation and B3-PMP2/6-
311+G(2d,p) ionization energies of 1.
in a nonrelaxed geometry, the vibrational wave packet unfolds
rapidly on the potential energy surface to distribute the internal
energy among the vibrational degrees of freedom.55 Because
the Franck-Condon energies in vertical electron capture were
low, e.g., 25 kJ mol-1 from the CIS energies of the vertically
2
2
2
and the X, A, and B states depicted in Figure 7.
The 2A through 2E excited states formed by vertical neutral-
ization of 1+ all had energies that were above those for the
transition states for the hydrogen losses or ring cleavages in
ground state 1. Hence, any of the radiationally long-lived excited
states can drive the fragmentations of ground-state 1 following
internal conversion. The calculated radiative lifetimes of the
2
formed and fully optimized A state, most of the vibrational
modes must remain in V ) 0 states.56 Therefore, the radiative
2
lifetime of radicals formed by electron capture in the A state
2
2
also depends on the transition moments from the geometry of
the corresponding potential energy minimum. Table 5 shows
that the radiative lifetime of the relaxed 2A state was comparable
to that formed by vertical electron capture, and a similar situation
was expected for the higher excited states, as well.57
vertically formed A through E states, τ ) 1/Aij, where Aij is
the Einstein coefficient for spontaneous radiative transition to
the ground state, were in the µs range, although it must be noted
that the values calculated by CIS and TD-B3LYP differed (Table
5). Following vertical electron capture to form the excited state

8350 J. Phys. Chem. A, Vol. 105, No. 36, 2001
Syrstad and Vivekananda
(15) Stratmann, R. E.; Scuseria, G. E.; Frisch, M. J. J. Chem. Phys.
1998, 109, 8218.
(16) Wolken, J. K.; Turecek, F. J. Phys. Chem. A 2001, 105, 8352.
(17) Lee, C. K.; Shim, J. Y. Bull. Korean Chem. Soc. 1991, 12, 343.
(18) Turecek, F.; Gu, M.; Shaffer, S. A. J. Am. Soc. Mass Spectrom.
1992, 3, 493.
(19) Turecek, F. Org. Mass Spectrom. 1992, 27, 1087.
(20) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A.
D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.;
Ochterski, J.;. Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.;
Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.;
Cioslowski, J.; Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz,
P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-
Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe,
M.; Gill, P. M. W.; Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J.
L.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98, ReVision
A.6, Gaussian, Inc., Pittsburgh, PA, 1998.
(21) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 1372, 5648. (b) Stephens,
P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J. Phys. Chem. 1994,
98, 11 623.
(22) Rauhut, G.; Pulay, P. J. Phys. Chem. 1995, 99, 3093.
(23) (a) Finley, J. W.; Stephens, P. J. J. Mol. Struct. (THEOCHEM)
1995, 357, 225. (b) Wong, M. W. Chem. Phys. Lett. 1996, 256, 391. (c)
Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16 502.
(24) Møller, C.; Plesset, M. S. Phys. ReV. 1934, 46, 618.
(25) (a) Schlegel, H. B. J. Chem. Phys. 1986, 84, 4530. (b) Mayer, I.
AdV. Quantum Chem. 1980, 12, 189.
Can ring-cleavage dissociations occur on the excited-state
potential energy surface? Although detailed investigation of
excited-state potential energy surfaces of 1 would be a daunting
task, an insight can be gained from the comparison of the
2
2
optimized geometries of the X and A states of 1 (Figure 5).
2
2
Compared with the X state, the A state had a planar ring (Cs
symmetry) and showed a substantial elongation of the N-1s
C-2 bond and concomitant shortening of the N-1sC-6, C-2s
N-3, and C-2sO-2 bonds. These changes upon electron
excitation are consistent with the weakening of the N-1sC-2
bond to facilitate ring cleavage.
Conclusions
The hydrogen atom adduct to the C-5 in uracil was generated
specifically in the gas phase and found to be stable on the 5.1
µs time scale. The dissociations of uracil radicals generated by
femtosecond electron capture are driven by the formation of
excited electronic states. Specific loss of an axial hydrogen from
the C-5 methylene in radical 1 was observed in accord with the
calculated transition state energies and RRKM dissociation rate
constants. In addition, abundant ring-cleavage dissociations
occurred in the radicals that resulted in expulsion of HNCO
and CO.
(26) McWeeny, R.; Diercksen, G. J. Chem. Phys. 1968, 49, 4852.
(27) Parkinson, C. J.; Mayer, P. M.; Radom, L. J. Chem. Soc. Perkin
Trans. 2 1999, 2305.
(28) Turecek, F. J. Phys. Chem. A 1998, 102, 4703.
(29) (a) Turecek, F.; Wolken, J. K. J. Phys. Chem. A 1999, 103, 1905.
(b) Turecek, F.; Carpenter, F. H. J. Chem. Soc. Perkin Trans. 2 1999, 2315.
(c) Turecek, F.; Polasek, M.; Frank, A. J.; Sadilek, M.: J. Am. Chem. Soc.
2000, 122, 2361.
(30) Rablen, P. R. J. Am. Chem. Soc. 2000, 122, 357.
(31) Pople, J. A.; Head-Gordon, M.; Raghavachari, K. J. Chem. Phys.
1987, 87, 5968.
(32) Curtiss, L. A.; Raghavachari, K.; Pople, J. A. J. Chem. Phys. 1993,
98, 1293.
Acknowledgment. Support of this work by NSF (Grants
Nos. CHE-9712570 and CHE-0090930) is gratefully acknowl-
edged. Computational facilities used in this work were supported
by NSF (Grant No. CHE-9808182) and the University of
Washington. Thanks are due to Dr. Martin Sadilek for assistance
with CAD spectra measurements.
References and Notes
(1) von Sonntag, C. In Physical and Chemical Mechanism in Molecular
Radiation Biology; Glass, W. A., Varma, M. N., Eds.; Plenum Press: New
York, 1991; pp 287-321.
(2) Steenken, S. Chem. ReV. 1989, 89, 503.
(3) Flossmann, W.; Westhof, E.; Mu¨ller, A. J. Chem. Phys. 1976, 64,
1688.
(33) (a) Curtiss, L. A.; Redfern, P. C.; Smith, B. J.; Radom, L. J. Chem.
Phys. 1996, 104, 5148. (b) Smith, B. J.; Radom, L. J. Phys. Chem. 1995,
99, 6468.
(34) Foresman, J. B.; Head-Gordon, M.; Pople, J. A.; Frisch, M. J. J.
Phys. Chem. 1992, 96, 135.
(4) Westhof, E.; Lion, Y.; van de Vorst, A. Int. J. Radiat. Biol. 1977,
32, 499.
(5) Deeble, D. J.; Das, S.; von Sonntag, C. J. Phys. Chem. 1985, 89,
5784.
(35) Zhu, L.; Hase, W. L. Quantum Chemistry Program Exchange;
Indiana University: Bloomington, Indiana, 1994. Program No. QCPE 644.
(36) Frank, A. J.; Sadilek, M.; Ferrier, J. G.; Turecek, F. J. Am. Chem.
Soc. 1997, 119, 12 343.
(6) Pullman, B.; Pullman, A. Quantum Biochemistry; Interscience
Publishers: New York, 1963.
(37) Zhu, L.; Hase, W. L. Chem. Phys. Lett. 1990, 175, 117.
(38) Levine, I. N. Physical Chemistry, 3rd ed.; McGraw-Hill: New York,
1988; p 839.
(39) Holmes, J. L. Mass Spectrom. ReV. 1989, 8, 513.
(40) NIST Standard Reference Database No. 69, February 2000 Release.
http://webbook.nist.gov/chemistry.
(41) Wolken, J. K.; Turecek, F. J. Am. Soc. Mass Spectrom. 2000, 11,
1065.
(42) Podolyan, Y.; Gorb, L.; Leszczynski, J. J. Phys. Chem. A 2000,
104, 7346.
(43) The relative intensities in the CAD spectra obtained at kiloelec-
tronvolt collision energies are insensitive to the internal energy of the ion,
see for example: Busch, K. L.; Glish, G. L.; McLuckey, S. A. Mass
Spectrometry/Mass Spectrometry, VCH Publishers: New York, 1988; pp
163-166. With uracil ions and 1+, decreasing the precursor ion transmit-
tance from 70% (17% multiple collisions) to 50% (31% multiple collisions)
changed the fragment ion relative intensities by <5%.
(44) Gonzalez, C.; Schlegel, H. B. J. Phys. Chem. 1990, 94, 5523.
(45) Typical CPU running times on IBM RS/6000 for B3LYP and MP2/
6-311+G(2d, p) single point calculations were under 3 h, as compared to
QCISD(T)/6-31G(d, p) calculations that took 36 h and required over 7 GB
scratch disk space.
(7) (a) Hayon, E.; Simic, M. J. Am. Chem. Soc. 1973, 95, 1029. (b)
Fujita, S.; Steenken, S. J. Am. Chem. Soc. 1981, 103, 2540.
(8) (a) Symons, M. C. R. J. Chem. Soc., Faraday Trans. 1 1987, 83,
1. (b) Candeias, L. P.; Steenken, S. J. Phys. Chem. 1992, 96, 937.
(9) (a) Hayon, E. J. Chem. Phys. 1969, 51, 4881. (b) Myers, L. S.;
Hollis, M. L.; Theard, L. M.; Peterson, F. C.; Warnick, A. J. Am. Chem.
Soc. 1970, 92, 2875.
(10) (a) Henriksen, T.; Snipes, W. J. Chem. Phys. 1970, 52, 1997. (b)
Box, H. C.; Budzinski, E. E. J. Chem. Phys. 1975, 62, 197. (c) Sagstuen,
E.; Hole, E. O.; Nelson, W. H.; Close, D. M. J. Phys. Chem. 1989, 93,
5974. (d) Ohta, N.; Tanaka, N.; Ito, S. J. Chem. Soc., Perkin Trans. 2 1999,
2597.
(11) (a) Krauss, M.; Osman, R. J. Phys. Chem. 1993, 97, 13515. (b)
Colson, A.-O.; Sevilla, M. D. J. Phys. Chem. 1995, 99, 13033. (c) Colson,
A.-O.; Sevilla. M. D. Int. J. Radiat. Biol. 1995, 67, 627. (d) Colson, A.-O.;
Becker, D.; Eliezer, I.; Sevilla, M. D. J. Phys. Chem. A 1997, 101, 8935.
(12) Wolken, J. K.; Syrstad, E. A.; Vivekananda, S.; Turecek, F. J. Am.
Chem. Soc. 2001, 123, 5804.
(13) For most recent reviews, see: (a) Zagorevskii, D. V.; Holmes, J.
L. Mass Spectrom. Rev. 1999, 18, 87. (b) Schalley, C. A.; Hornung, G.;
Schroder, D.; Schwarz, H. Chem. Soc. ReV. 1998, 27, 91.
(14) For applications of NRMS to the gas-phase chemistry of hetero-
cyclic radicals see: (a) Wolken, J. K.; Turecek, F. J. Am. Chem. Soc. 1999,
121, 6010. (b) Wolken, J. K.; Turecek, F. J. Phys. Chem. A, 1999, 103,
6268. (c) Nguyen, V. Q.; Turecek, F. J. Am. Chem. Soc. 1997, 119, 2280.
(d) Nguyen, V. Q.; Turecek, F. J. Mass Spectrom. 1997, 32, 55. (e) Nguyen,
V. Q.; Turecek, F. J. Mass Spectrom. 1996, 31, 1173. (f) Turecek, F. J.
Mass Spectrom. 1998, 33, 779.
(46) Yu, C.; O’Donnell, T. J.; LeBreton, P. R. J. Phys. Chem. 1981,
85, 3851.
(47) Desfrancois, C.; Abdoul-Carime, H.; Schermann, J. P. J. Chem.
Phys. 1996, 104, 7792.
(48) Hendricks, J. H.; Lyapustina, S. A.; de Clercq, H. L.; Snodgrass,
J. T.; Bowen, K. H. J. Chem. Phys. 1996, 104, 7788.
(49) (a) Nguyen, M. T.; Creve, S.; Vanquickenborne, L. G. J. Phys.
Chem. 1996, 100, 18 422. (b) Jursic, B. S. Chem. Phys. Lett. 1996, 256,

Hydrogen Atom Adduct to C-5 in Uracil
J. Phys. Chem. A, Vol. 105, No. 36, 2001 8351
603. (c) Skokov, S.; Wheeler, R. A. Chem. Phys. Lett. 1997, 271, 251. (d)
Basch, H.; Hoz, S. J. Phys. Chem. A 1997, 101, 4416.
(50) Connors, K. A. Chemical Kinetics, VCH Publishers: New York,
1990; p 101.
V. Q.; Sadilek, M.; Frank, A. J.; Ferrier, J. G.; Turecek, F. J. Phys. Chem.
A 1997, 101, 3789.
(55) See for example: Lehmann, K. K.; Pate, B. H.; Scoles, G. In Mode
SelectiVe Chemistry; Jortner, J., Levine, R. D., Pullman, B. Eds.; Kluwer:
Dordrecht, 1991; pp 17-23.
(51) Bodenstein, M. Z. Physik. Chem. 1913, 85, 329.
(52) Time-dependent branching ratios were also calculated for dissocia-
tion times t ) 1/kH and showed values between 0 and 0.20 for internal
(56) Dunbar, R. C. J. Chem. Phys. 1989, 90, 7369.
(57) This follows from the fact that as the electron is promoted to higher
excited states and becomes less tightly bound, the excited-state geometry
of the radical resembles more closely that of the ion. Hence, Franck-Condon
effects on vertical electron capture diminish when higher excited states of
the radical are formed.
energies in 1 ranging from TS7 to 360 kJ mol-1
.
(53) Lifshitz, C. Mass Spectrom. ReV. 1982, 1, 309.
(54) (a) Polasek, M.; Turecek, F. J. Am. Chem. Soc. 2000, 122, 9511.
(b) Frank, A. J.; Turecek, F. J. Phys. Chem. A 1999, 103, 5348. (c) Nguyen,