Thermochemical and Theoretical Study of Some Quinoxaline 1,4-Dioxides and of Pyrazine 1,4-Dioxide

Article abstract of DOI:10.1021/jo962149s

The following standard molar enthalpies of formation in the gaseous state at 298.15 K were determined from the enthalpies of combustion of the crystalline solids, and their enthalpies of sublimation and the mean (N-O) bond dissociation enthalpies were derived. Δ_fH_m⁰(g); 〈D(N-O)〉 (kJ mol^-1): quinoxaline 1,4-dioxide, 227.1 ± 2.4; 255.8 ± 2.0; 2-methylquinoxaline 1,4-dioxide, 169.9 ± 7.2; 268.3 ± 4.9; 2-methyl-3-acetylquinoxaline 1,4-dioxide, 33.1 ± 5.0; 251.6 ± 4.2; 2-phenyl-3-benzoylquinoxaline 1,4-dioxide, 355.2 ± 7.1; 227.3 ± 5.4; 2-methyl-3-carbomethoxyquinoxaline 1,4-dioxide, -148.7 ± 3.2; 242.3 ± 3.9; pyrazine 1,4-dioxide, 186.5 ± 1.9; 254.0 ± 2.3; 2-methyl-5-pyrazinecarboxylic acid, -213.6 ± 1.7. Unconstrained geometry optimizations by ab initio calculations showed the effect of steric hindrance on changes in extended delocalizations and were in accord with the trends in the mean bond dissociation enthalpies.

Full text of DOI:10.1021/jo962149s

3722
J . Org. Chem. 1997, 62, 3722-3726
Th er m och em ica l a n d Th eor etica l Stu d y of Som e Qu in oxa lin e
1,4-Dioxid es a n d of P yr a zin e 1,4-Dioxid e
W. E. Acree, J r.,*,† J oyce R. Powell,† Sheryl A. Tucker,† Maria D. M. C. Ribeiro da Silva,‡
M. Agostinha R. Matos,‡ J . M. Gonc¸alves,‡ L. M. N. B. F. Santos,‡ V. M. F. Morais,§ and
G. Pilcher¶
Department of Chemistry, University of North Texas, Denton, Texas, 76203-0068, Departamento de
Quı´mica, Faculdade de Cieˆncias, Universidade do Porto, P-4150, Porto, Portugal, Instituto de Cieˆncias
Biome´dicas Abel Salazar, Universidade do Porto, P-4050 Porto, Portugal, and Department of Chemistry,
University of Manchester, Manchester M13 9PL, U.K.
Received November 18, 1996X
The following standard molar enthalpies of formation in the gaseous state at 298.15 K were
determined from the enthalpies of combustion of the crystalline solids, and their enthalpies of
sublimation and the mean (N-O) bond dissociation enthalpies were derived. ∆fH0 (g); D(N-O)
m
(kJ mol-1): quinoxaline 1,4-dioxide, 227.1 ( 2.4; 255.8 ( 2.0; 2-methylquinoxaline 1,4-dioxide, 169.9
( 7.2; 268.3 ( 4.9; 2-methyl-3-acetylquinoxaline 1,4-dioxide, 33.1 ( 5.0; 251.6 ( 4.2; 2-phenyl-3-
benzoylquinoxaline 1,4-dioxide, 355.2 ( 7.1; 227.3 ( 5.4; 2-methyl-3-carbomethoxyquinoxaline 1,4-
dioxide, -148.7 ( 3.2; 242.3 ( 3.9; pyrazine 1,4-dioxide, 186.5 ( 1.9; 254.0 ( 2.3; 2-methyl-5-
pyrazinecarboxylic acid, -213.6 ( 1.7. Unconstrained geometry optimizations by ab initio
calculations showed the effect of steric hindrance on changes in extended delocalizations and were
in accord with the trends in the mean bond dissociation enthalpies.
Sch em e 1
In tr od u ction
A reaction scheme has been employed by Holm and his
colleagues,1,2 to order various oxygenated species, includ-
ing N-oxides, in terms of their abilities to transfer oxygen
atoms in molybdenum(IV) to molybdenum(VI) conver-
sions, by considering the enthalpy of the reaction
RN(g) + 1/2O2 f RN+ - O-
These particular transfers are important in the bio-
physical chemistry of molybdenum complexes, and the
order of decreasing enthalpy of reaction provides a
method for predicting the feasibility of such conversions.
Limited enthalpies of formation have restricted the scale
to a “scarce” selected compounds. The dissociation en-
thalpy of the (N+-O-) dative covalent bond shows large
variations depending on the molecular environment in
the immediate vicinity of the bond. The present paper
reports the mean molar (N+-O-) bond dissociation en-
thalpies, D(N-O) , in pyrazine di-N-oxide and in the
quinoxaline di-N-oxide derivatives shown in Scheme 1.
The standard molar enthalpies of formation of the
compounds 1-7 in the gaseous state were derived from
the enthalpies of combustion of the crystalline solids and
the enthalpies of sublimation. To derive the mean (N+-
O-) bond dissociation enthalpies, the enthalpies of forma-
tion of the corresponding quinoxaline derivatives and of
pyrazine are required. Unfortunately, quinoxaline de-
rivatives are so difficult to obtain in a sufficiently pure
condition for precise combustion measurements that only
two experimental values are available: for quinoxaline3
and for 2,3-dimethylquinoxaline.4 The measurement of
2-methyl-5-pyrazinecarboxylic acid was undertaken to
facilitate the thermochemical estimation of the required
enthalpies of formation of the quinoxaline derivatives.
To assist in the interpretation of the trends in the
mean (N+-O-) dissociation enthalpies, ab initio calcula-
tions were made to investigate the geometric structures
of some of these molecules.
Exp er im en ta l Section
† University of North Texas.
Ma ter ia ls. Quinoxaline 1,4-dioxide (1) was prepared by
(sodium tungstate + hydrogen peroxide) oxidation of quinoxa-
line as described by Kobayashi et al.5 2-Methylquinoxaline
‡ Departamento de Qu´ımica, Universidade do Porto.
§ Instituto de Cieˆncias Biome´dicas, Universidade do Porto.
University of Manchester.
X Abstract published in Advance ACS Abstracts, May 15, 1997.
(1) Holm, R. H. Chem. Rev. 1987, 87, 1401.
(2) Caradonna, J . P.; Reddy, P. R.; Holm, R. H. J . Am. Chem. Soc.
1988, 110, 2139.
(3) Ribeiro da Silva, M. A. V.; Matos, M. A. R.; Morais, V. M. F. J .
Chem. Soc., Faraday Trans. 1995, 91, 1907.
(4) Ribeiro da Silva, M. A. V.; Morais, V. M. F.; Matos, M. A. R.;
Rio, C. M. A.; Piedade, C. M. G. S. Struct. Chem. 1996, 7, 329.
(5) Kobayashi, Y.; Kumadaki, I.; Sato, H.; Sekine, Y.; Hara, T. Chem.
Pharm. Bull. 1974, 22, 2097.
S0022-3263(96)02149-4 CCC: $14.00 © 1997 American Chemical Society

Quinoxaline and Pyrazine 1,4-Dioxides
J . Org. Chem., Vol. 62, No. 11, 1997 3723
1,4-dioxide (2) was obtained from Maybridge Chemical Ltd.
2-Methyl-3-acetylquinoxaline 1,4-dioxide (3) was prepared by
dissolving benzofurazan 1-oxide and pentane-2,4-dione in
trimethylamine and stirring the resulting solution for 24 h at
ambient temperature; the yellow precipitate was collected by
vacuum filtration. 2-Methyl-3-carbomethoxyquinoxaline 1,4-
dioxide (4) and 2-phenyl-3-benzoylquinoxaline 1,4-dioxide (5)
were prepared in a similar fashion from benzofurazan 1-oxide
plus methyl acetoacetate or dibenzoylmethane.6,7 Pyrazine 1,4-
dioxide (6) was prepared by the (sodium tungstate + hydrogen
peroxide) oxidation of pyrazine. 2-Methyl-5-pyrazinecarboxylic
acid (7) was a gift from Agro-Pharmaceutical Division of
Nippon Soda Co. (Tokyo, J apan). All seven compounds were
further purified by four crystallizations from methanol.
Elemental analyses were in good agreement with expected
values. Mass fractions for 1, C8H6N2O2: found, C, 0.5932; H,
0.0374; N, 0.1722; calculated, C, 0.5926; H, 0.0373; N, 0.1728.
For 2, C9H8N2O2: found, C, 0.6149; H, 0.0451; N, 0.1593;
calculated, C, 0.6136; H, 0.0458; N, 0.1590. For 3,
C11H10N2O3: found, C, 0.6060; H, 0.0468; N, 0.1275; calculated,
C, 0.6055; H, 0.0462; N, 0.1284. For 4, C11H10N2O4: found,
C, 0.5651; H, 0.0418; N, 0.1202; calculated, C, 0.5641; H,
0.0430; N, 0.1196. For 5, C21H14N2O3: found, C, 0.7378; H,
0.0398; N, 0.0833; calculated, C, 0.7368; H, 0.0412; N, 0.0818.
For 6, C4H4N2O2: found, C, 0.4290; H, 0.0355; N, 0.2506;
calculated, C, 0.4286; H, 0.0360; N, 0.2499. For 7, C6H6N2O2:
found, C, 0.5215; H, 0.0429; N, 0.2031; calculated, C, 0.5217;
H, 0.0438; N, 0.2028.
Hexadecane was used as an auxiliary combustion aid for
some measurements, and its standard specific energy of
combustion, -∆cu0, was measured separately. In the combus-
tion of compounds 3 and 6, for the hexadecane used, -∆cu0/J
g-1 ) 47164.7 ( 2.7, where the uncertainty is the standard
deviation of the mean, in close agreement with the value of
Fraser and Prosen8 measured for a sample of 99.96 mol %
purity, -∆cu0/J g-1 ) 47155.0 ( 3.8. For compound 5, a
hexadecane sample from a different source was used, for which
-∆cu0/J g-1 ) 47076.7 ( 3.3. The chemical composition of an
auxiliary combustion aid should be known approximately, but
the specific energy of combustion must be known precisely.
The difference of ca. 0.2% between the two hexadecane
samples, due to slight difference in purity, will not produce
systematic error.
MPa with 1 cm3 of water added to the bomb. The electrical
energy of ignition was determined from the change in potential
difference across a capacitor when discharged through the
platinum ignition wire. For the cotton thread fuse, empirical
12
formula CH1.686O0.843, -∆cu0 ) 16250 J g-1
.
Corrections for
nitric acid formation were based on -59.7 kJ mol-1 for the
molar energy of formation of 0.1 mol dm-3 HNO3(aq) from N2,
O2, and H2O(l).13 Corrections for carbon formation were based
on-∆cu0) 33 kJ g-1
.
For each compound, at 298.15 K, (∂u/
12
∂p)T was assumed to be -0.2 J g-1 MPa-1, a value typical for
most organic solids. For each compound, -∆cu0 was calculated
by the procedure given by Hubbard et al.12 The relative atomic
masses used were those recommended by the IUPAC Com-
mission.14 The amount of substance used in each experiment
was determined from the total mass of carbon dioxide produced
after allowance for that from the cotton thread fuse, hexa-
decane, and that lost due to carbon formation. The average
ratios of the mass of carbon dioxide produced by the samples
to that calculated from its mass with uncertainties of twice
the standard deviation of the mean value were as follows: 1,
0.9995 ( 0.0001; 2, 1.0001 ( 0.0003; 3, 1.0003 ( 0.0013; 4,
1.0000 ( 0.0003; 5, 0.9980 ( 0.0003; 6, 1.0004 ( 0.0004; 7,
0.9995 ( 0.0003.
En th a lp ies of Su blim a tion . The enthalpies of sublima-
tion of 1, 2, 4, 6, and 7 were measured using the “vacuum-
sublimation” drop microcalorimetric method.15 Samples, about
5 mg, of each compound, contained in a thin glass capillary
tube sealed at one end were dropped at room temperature into
the hot reaction vessel in the Calvet High Temperature
Microcalorimeter (SETARAM HT1000) and then removed from
the hot zone by vacuum sublimation. The observed enthalpies
of sublimation [H0m(g, T K) - H0m(cr, 298.15 K)], were cor-
rected to 298.15 K using ∆T29K8.15KHm0 (g) estimated by a group
method based on the values of Stull et al.16 The micro-
calorimeter was calibrated in situ for these measurements
using the reported enthalpy of sublimation of naphthalene.17
For compounds 3 and 5, the Knudsen effusion method as
described by Burkinshaw and Mortimer18 was used. The vapor
effusing from a Knudsen cell was allowed to condense on a
quartz crystal positioned above the effusion hole, and changes
in frequency of oscillation of the quartz plate were proportional
to the mass condensed. From the Knudsen formula,
1/2
m˘ 2πRT
Com bu stion Ca lor im etr y. The standard molar enthal-
p )
(1)
pies of combustion, ∆cH0 , of 1, 4, and 7 were measured using
(
)
a
M
m
the static-bomb calorimeter in Manchester9 whereas the
remaining compounds (2, 3, 5, and 6) were measured using
the static-bomb calorimeter in Porto.10,11 The energy equiva-
lents of the calorimeters, ꢀ(calor), were determined using the
combustion of benzoic acid: in Manchester with the NBS SRM
where m˘ is the rate of mass loss, a the effective hole area,
and M the molar mass of the effusing vapor, as the rate of
change in frequency v˘ is proportional to the rate of mass loss,
then
39i for which under standard bomb conditions -∆cu0/J g-1
)
1/2
26434 ( 3, for measurements 1, 7, from 12 calibrations,
ꢀ(calor)) 15525.40 ( 0.47 J K-1, but for measurement 4, from
8 calibrations, ꢀ(calor) ) 15531.70 ( 0.68 J K-1 for an average
mass of water added to the calorimeter of 2897.0 g: in Porto
with Bureau of Analysed Samples CRM 190p for which
-∆cu0/J g-1 ) 26431.8 ( 3.7, and from 8 calibrations, ꢀ(calor)
) 15911.2 ( 1.5 J K-1 for an average mass of water added to
the calorimeter of 3119.6 g. Both calorimeters and experi-
mental techniques were tested for systematic errors by making
independent measurements of the energy of combustion of
3-methylcatechol with the results: Manchester, -∆cu0/J g-1
) 28216.9 ( 1.9; Porto, -∆cu0/J g-1 ) 28217.7 ( 0.9.11 The
experimental conditions were similar for both calorimeters.
Samples were ignited at 298.150 ( 0.001 K in oxygen at 3.04
2πR
ln p ) ln(v˘ T1/2) + ln
(2)
2
(
)
a M
and hence the standard molar enthalpy of sublimation,
∆cgrH0m, can be derived from the slope of ln(v˘ T1/2) versus T-1 by
applying the integrated form of Clausius-Clapeyron equation.
The measurements were made in Porto using an Edwards 306
vacuum coating unit providing a minimum pressure of 7 ×
10-5 Pa, fitted with an FTM3 film thickness monitor incorpo-
rating a quartz crystal oscillator (6 MHz) but with some
modifications from the apparatus described by Burkinshaw
and Mortimer.18 The cylindrical Knudsen cell was made of
(12) Hubbard, W. N.; Scott, D. W.; Waddington, G. In Experimental
Thermochemistry; Rossini, F. D.; Ed.; Interscience: New York, 1956.
(13) The NBS Tables of Chemical Thermodynamic Properties. J .
Phys. Chem. Ref. Data 1982, 11.
(6) Terrian, D. L.; Houghtaling, M. A.; Ames, J . R. J . Chem. Educ.
1992, 69, 589.
(7) Issidorides, C. H.; Haddadin, M. J . J . Org. Chem. 1966, 31, 4067.
(8) Fraser, F. M.; Prosen, E. J . J . Res. Natl Bur. Stand. 1955, 55,
329.
(9) Bickerton, J .; Pilcher, G.; Al-Takhin, G. J . Chem. Thermodyn.
1984, 16, 373.
(10) Ribeiro da Silva, M. D. M. C.; Ribeiro da Silva, M. A. V.; Pilcher,
G. J . Chem. Thermodyn. 1984, 16, 1149.
(11) Ribeiro da Silva, M. A. V.; Ribeiro da Silva, M. D. M. C.; Pilcher,
G. Rev. Port. Qu’ım. 1984, 26, 163.
(14) IUPAC. J . Phys. Chem. Ref. Data 1993, 22, 1571.
(15) Adedeji, F. A.; Brown, D. L. S.; Connor, J . A.; Leung, M.; Paz-
Andrade, M. I.; Skinner, H. A. J . Organomet. Chem. 1975, 97, 221.
(16) Stull, D. R.; Westrum, E. F.; Sinke, G. C. The Chemical
Thermodynamics of Organic Compounds; Wiley: New York, 1969.
(17) de Kruif, C. G.; Kuipers, T.; van Mittenburg, J . C.; Schaake, R.
C. F.; Stevens, G. J . Chem. Thermodyn. 1981, 13, 1081.
(18) Burkinshaw, P. M.; Mortimer, C. T. J . Chem. Soc., Dalton
Trans. 1984, 75.

3724 J . Org. Chem., Vol. 62, No. 11, 1997
Ta ble 1. Typ ica l Com bu stion Mea su r em en ts a t T ) 298.15 K (p0 ) 0.1 MP a )
Acree et al.
1
2
3
4
5
6
7
m(CO2, total)/g
m(cpd)/g
m(hexadecane)/g
m(fuse)/g
1.73197
0.79509
1.69077
0.74963
1.39041
0.37903
0.17469
0.00389
1.14196
2.08642
1.00740
1.92553
0.43525
0.23931
0.00394
1.54790
1.50739
0.50788
0.22536
0.00554
1.30839
1.94323
1.01423
0.00343
1.30972
15525.4
12.89
0.0
20349.7
66.9
0.00334
1.27713
15911.2
15.32
0.1
20339.7
62.0
0.00257
1.52885
15531.7
13.41
-2.6
23748.4
72.7
0.00288
1.38235
15525.4
13.36
0.0
21478.9
82.0
∆Tad/K
ꢀ(calor)/J K-1
ꢀf/J K-1
15911.2
15911.2
15911.2
16.05
-0.1
18186.8
30.6
16.15
-0.1
24652.2
27.5
0.0
1.1
12.6
11265.8
64.0
15.89
-0.1
20837.7
60.8
0.0
0.6
12.3
10629.0
90.0
∆m(H2O)/g
-∆U(IBP)/J
∆U(HNO3)/J
∆U(carbon)/J
∆U(ignition)/J
∆UΣ/J
0.0
1.1
13.2
0.0
1.0
15.2
0.0
1.0
9.4
0.0
1.1
16.9
3.30
1.1
14.5
-m ∆cu0(hexadecane)/J
-m ∆cu0(fuse)/J
-∆cu0(cpd)/J g-1
8239.2
63.2
25972.6
55.8
25423.3
54.2
26957.6
41.7
23443.6
46.8
21039.5
30516.5
19779.5
Ta ble 2. Va lu es of -∆cu 0/J g-1 a t 298.15 K
1
2
3
4
5
6
7
25423.3
25424.2
25423.3
25425.1
25420.1
25411.9
26933.0
26000.4
26022.6
26027.6
25956.1
26003.5
26027.2
26004.4
25972.6
23443.6
23448.2
23454.7
23444.9
23453.7
23448.0
30505.2
30516.3
30509.2
30545.9
30535.9
30528.5
19762.8
21039.5
21035.2
21032.0
21033.6
21035.5
21028.5
26964.3
26933.0
26924.5
26957.6
26982.6
19811.7
19786.0
19786.8
19809.0
19769.2
19779.5
- ∆cu0 /J g-1
23448.9 ( 1.8
25421.3 ( 2.0
26949.2 ( 9.2
26001.8 ( 9.1
30523.5 ( 6.5
19786.4 ( 7.0
21034.1 ( 1.5
Ta ble 3. Der ived Mola r Va lu es (k J m ol-1) a t T ) 298.15 K
-∆cU0 (cr)
-∆cU0 (cr)
∆fU0 (cr)
∆gcrU0m
∆fU0 (g)
m
m
m
m
1
2
3
4
5
6
7
4122.0 ( 0.9
4747.8 ( 3.4
5673.9 ( 4.2
5492.0 ( 1.2
10449.8 ( 5.2
2217.8 ( 1.6
2905.3 ( 0.5
4120.7 ( 0.9
4747.8 ( 3.4
5673.9 ( 4.2
5490.8 ( 1.2
10452.3 ( 5.2
2215.3 ( 1.6
2904.1 ( 0.5
115.1 ( 1.4
62.9 ( 3.6
112.0 ( 1.9
107.0 ( 6.2
117.0 ( 2.4
118.3 ( 2.6
167.4 ( 4.0
116.9 ( 0.8
100.9 ( 1.5
227.1 ( 2.4
169.9 ( 7.2
33.1 ( 5.0
-148.7 ( 3.2
355.2 ( 7.1
186.5 ( 1.9
-213.6 ( 1.7
-83.9 ( 4.4
-267.0 ( 1.9
187.8 ( 5.9
69.6 ( 1.7
-314.5 ( 0.9
steel, external diameter 13 mm, internal diameter 6 mm, and
depth 8 mm. The lid with an effusion hole diameter of 0.8
mm was screwed on to the cell. The complete cell was screwed
into an electrically heated cylindrical steel block, and the
temperature was controlled to (0.1 K by a Eurotherm 815
proportional integral derivative controller. Temperatures were
measured by a Labfacility P100/0620 platinum sensing detec-
tor screwed into the effusion cell. Frequency changes in the
quartz crystal oscillator circuit were measured with a Philips
Frequency Counter PM6685 connected to the FTM3 thickness
monitor. The equipment was tested with several compounds
of known standard molar enthalpy of sublimation.
tainties in calibration and in the values of the auxiliary
quantities used. To derive ∆fH0 from ∆cH0 , the values
m
m
∆fH0 (H2O, l)/kJ mol-1
)
-285.83
(
0.04 and
m
∆fHm0 (CO2, g)/kJ mol-1 ) -393.51 ( 0.1319 were used.
Discu ssion
The mean (N+-O-) bond dissociation enthalpy is one-
half of the enthalpy of the gaseous reaction
+
+
O--N-R-N-O- f N-R-N + 2O
(3)
Resu lts
requiring ∆fHm0 (O, g)/kJ mol-1 ) 249.18 ( 0.1019 and
Results for a typical combustion experiment on each
compound are given in Table 1; ∆m(H2O) is the deviation
of the mass of water added to the calorimeter from the
average mass assigned to ꢀ(calor). Samples were ignited
at 298.15 K so that ∆U(IBP) ) -{ꢀ(calor) + cp(H2O,
l)∆m(H2O) + ꢀf}∆Tad + ∆U(ign); ∆UΣ is the correction to
the standard state, and the remaining terms are as
previously defined.12 The individual values of ∆cu0 with
the mean and its standard deviation are given in Table
2. Table 3 lists the derived standard molar enthalpies
of combustion, ∆cH0 , and formation, ∆fH0 , in the crys-
∆fH0 (g) for N-R-N : the only experimental values for
m
the latter are for quinoxaline,3 2,3-dimethylquinoxaline4
and pyrazine;20 the remaining required values may be
estimated using group additivity methods. It appears
that the increment in ∆fH0 (g) for substitution into the
m
pyrazine ring is the same as for benzene: e.g., for
2-methyl-5-pyrazinecarboxilic acid an estimate using
∆fH0 (g) values for (pyrazine + toluene + benzoic acid -
m
2 benzene) gives -212.8 ( 3.0 kJ mol-1 agreeing with
the experimental value of-213.6 ( 1.7 kJ mol-1, and for
2,3-dimethylquinoxaline, an estimate by (quinoxaline +
m
m
talline and gaseous states. In accordance with normal
thermochemical practice, the uncertainties assigned to
the derived molar enthalpies are twice the overall
standard deviations of the mean and include the uncer-
(19) Cox, J . D.; Wagman, D. D.; Medvedev, V. A. CODATA Key
Values for Thermodynamics; Hemisphere: New York, 1989.
(20) Tjebbes, J . Acta Chem. Scand. 1962, 16, 916.

Quinoxaline and Pyrazine 1,4-Dioxides
J . Org. Chem., Vol. 62, No. 11, 1997 3725
Ta ble 4. Mea n Bon d Dissocia tion En th a lp ies
Sch em e 2
2 toluene - 2 benzene) gives -175.9 ( 4.4 kJ mol-1 in
agreement with the experimental value 172.9 ( 3.0 kJ
mol-1 4 Using the following ∆fHm0 (g)/kJ mol-1 values:21
.
C6H6, 82.6 ( 0.7; C6H5CH3, 50.4 ( 0.6; C6H5COCH3,
-86.7 ( 1.6; C6H5CO2CH3, -287.9 ( 2.4; C6H5C6H5, 181.4
(
2.0; C6H5COC6H5, 54.9 ( 4.4, the estimated
∆fH0m(g)/kJ mol-1 values for the quinoxaline derivatives
given in Table 4 were derived using this assumption with
an additional uncertainty of (5.0 kJ mol-1 but with no
allowance made for possible steric strain. Table 4 shows
that D(N-O) in quinoxaline di-N-oxide and in pyrazine
di-N-oxide are the same but there is a reduction for
compounds 4 and 5 which may be partially due to the
neglect of possible steric hindrance in the estimation of
initio theory, leading at the same time to the harmonic
vibrational frequencies. Similar calculations for 5 and
9 would be very time consuming because of the large
number of degrees of freedom involved and were not
made, but we believe the optimized geometries for these
molecules are also true minima. The quinoxaline moiety
of all these molecules was found to be planar, and the
global loss of planarity of 4, 5, 8, and 9 results from the
spatial orientation of the substituents with respect to the
quinoxaline planar substructure.
∆fH0 (g) for the quinoxaline derivatives. These com-
m
pounds can be ordered using the reaction scheme of Holm
et al.,1,2 in terms of their abilities to transfer oxygen
atoms in the Mo(IV) to Mo(VI) conversion by considering
one-half of the enthalpy of the gaseous reaction,
In 4 the methyl substituent, which is expected to rotate
easily, assumes an orientation so that two of its H atoms
are located almost symmetrically with respect to the
quinoxaline plane while the third, Hm, atom is oriented
so that its distance from O2 is as large as possible, the
Nj-Ck-Cl-Hm dihedral angle being 173°. For 8 the
methyl substituent assumes a symmetrical orientation
relative to the quinoxaline plane but the in-plane H atom
is oriented so that its distance from Nj is minimum, Nj-
Ck-Cl-Hm dihedral angle being now -3°. The other
substituent in 4 shows similar behavior rotating so that
the Od atom lies far from the O1 atom of the quinoxaline
di-N-oxide. The Na-Cb-Cc-Od dihedral angle is -129°
whereas for 8 this angle assumes a value of -164°
indicating that the Cb-Cc-Od-Oe group is closer to
planarity with the quinoxaline plane: complete planarity
would correspond to a dihedral angle of -180°.
These conformations of the two molecules suggest the
occurrence of steric hindrance and that due to the
presence of the O atoms in the di-N-oxide is greatly
attenuated in the molecule with the O atoms removed.
Steric hindrance can destabilize 4 with respect to 8, and
since this destabilization does not occur in the nonsub-
stituted quinoxaline derivatives, this would account for
a decrease in D(N-O) in 4 relative to 1.
quinoxaline derivative + O2 f
quinoxaline di-N-oxide derivative
giving: 5 (21.9 ( 5.4 kJ mol-1); 4 (6.9 ( 3.9 kJ mol-1); 3
(-2.4 ( 4.2 kJ mol-1); 6 (-4.8 ( 1.2 kJ mol-1); 1 (-6.6 (
2.4 kJ mol-1); 2 (-19.1 ( 4.9 kJ mol-1). As expected,
the more highly strained N-oxide derivatives are the
more powerful oxidizing agents. If such strain exists,
then the strain in the corresponding di-N-oxides will be
greater; hence, some of these structures were investi-
gated by ab initio calculations.
Ab In itio Ca lcu la tion s
Unconstrained geometry optimizations were performed
for the five molecules shown in Scheme 2. All calcula-
tions were carried out at the ab initio restricted Hartree-
Fock level using STO-3G minimal basis sets and the
American version of the program GAMESS.22,23 More
rigorous calculations are at present precluded by the
large number of atoms in these molecules. The resulting
optimized geometries of 1, 4, and 7 were confirmed as
true minima through calculation and diagonalization of
the respective Hessian matrices, at the same level of ab
The results of the geometry optimization of 5 and 9
suggest a similar behavior. In both cases the substituent
benzene rings are not coplanar with the quinoxaline
moiety. The dihedral angle Na-Cb-Cc-Od is -106° for
5 and -125° for 9 while the dihedral angle Ne-Cf-Cg-
Ci assumes values of -124° and -137° respectively for 5
(21) Pedley, J . B. Thermochemical Data and Structures of Organic
Compounds, vol. I; Thermodynamics Research Center: College Station,
Texas, 1994.
(22) Dupuis, M.; Spangler, D.; Wendoloski, J . J . NRCC Software
Catalog 1980, 1, Program QG10.
(23) Schmidt, M. W.; Boatz, J . A.; Baldridge, K. K.; Koseki, S.;
Gordon, M. S.; Elbert, S. T.; Lam, B. QCPE Bull. 1987, 7, 115.

3726 J . Org. Chem., Vol. 62, No. 11, 1997
Acree et al.
and 9. The two substituents assume more planar con-
formations, with respect to the quinoxaline moiety plane,
when the O atoms are removed which is compatible with
increased steric hindrance due to the presence of the O
atoms in the di-N-oxide molecules.
The fact that these molecules become more planar upon
breaking of the N-O bonds may have other conse-
quences. If the molecules were completely planar the
extended delocalization could occur between the Cc-Od
double bonds and the quinoxaline ring and between the
benzene rings of the substituents and the quinoxaline
rings.
completely reliable. The natural localized molecular
orbitals (NLMO) describing the πN -C bond has non-
a
b
negligible contributions from the pπ atomic orbital of atom
Cc of 1.2%, 0.7%, 0.6%, and 0.3% respectively for 8, 4, 9,
and 5, while the contribution of the pπ atomic orbital of
atom Od assumes values of 0.2%, 0.1%, 1.2%, and 1.1%
for the same molecules. On the other hand, the NLMO
describing the πN -C bond shows contributions from the
e
f
pπ atomic orbital of atom Cg of 0.2% and 0.1% for 9 and
5, respectively: for the Ci pπ atomic orbital these contri-
butions are 0.3% and 0.2% for 9 and 5, respectively.
These delocalization effects can thus be depicted as a
charge transfer from the highly occupied bond orbitals
into adjacent unoccupied antibonding orbitals and their
importance can be more quantitatively characterized
through a second-order perturbative treatment of the
Fock matrix in the NBO basis, which gives the energy
lowering associated with such interactions. By this type
of calculation we found that the “donor-acceptor” interac-
tions which represent the main contributions to the
stabilization of 4 and 8 occur between the πN -C bond
Natural bond orbital (NBO) analysis24,25 has been
performed for each compound. NBO analysis provides a
description of bonding structure in terms of highly
occupied localized single and double bonds and lone pairs
which are compatible with the electronic density result-
ing from the Hartree-Fock wave function. This descrip-
tion corresponds closely to the classical Lewis picture of
chemical bonding. In systems showing electronic de-
localization no single localized description can account
for all the Hartree-Fock electronic density, and so a
convenient measure of delocalization is the fraction of the
total electronic density not described by the localized
structures; we found that the non-Lewis electronic den-
sity amounts to 2.4%, 2.7%, 2.0%, 2.3%, 2.7%, and 2.8%
of the total electronic density, for compounds quinoxaline,
1, 8, 4, 9, and 5, respectively. These fractions will appear
as small populations of the corresponding antibonding
orbitals which are formally unoccupied when delocaliza-
tion does not occur. Thus, for compounds 8 and 4 the
a
b
orbital with the antibonding π*Cc-O bond orbital and are
d
associated with energy lowerings of 32.6 and 84.1 kJ
mol-1, respectively. Similarly, the same nominal interac-
tions for 5 and 9 produce energy lowerings of 5.0 and 32.2
kJ mol-1, respectively. For 5 and 9, we found that
another important interaction occurs between the bond-
ing πN -C bond orbital with the π*Ci-C bond orbital is
e
f
g
associated with energy lowerings of 11.3 and 25.9 kJ
mol-1, respectively. These different stabilization energies
through extended delocalizations can be ascribed to the
different conformations and are in qualitative accord with
the differences in the D(N-O) values.
antibonding πC*c-O bond orbitals show populations of 0.18
d
and 0.16 electrons, respectively, whereas for compounds
9 and 5 the occupations of the same antibonding orbitals
are 0.09 and 0.08 electrons, respectively. The occurrence
of such bonding-antibonding stabilizing interactions
implies that any attempt to describe the bonding struc-
ture through localized bonds and lone pairs will not be
Ack n ow led gm en t. Authors from Porto University
thank Centro de Investigac¸a˜o em Qu´ımica (Unidade de
I&D no. 81/94, J NICT) for financial support. Thanks
are also due to Cla´udia M. V. Vaz for some measure-
ments on compound 6.
(24) Foster, J . P.; Weinhold, F. J . Am. Chem. Soc. 1980, 102, 7211.
(25) Reed, A. E.; Weinhold, F. J . Chem. Phys. 1983, 78, 4066.
J O962149S