2050
Russ.Chem.Bull., Int.Ed., Vol. 52, No. 9, September, 2003
Novakov et al.
Table 3. Effect of substituents in the Grignard reagent molecules
on the yields of ketonitriles 3a—d
compounds 4 and 5 were found to be the most energetiꢀ
cally favorable.
According to calculations, the fourꢀ and sixꢀmemꢀ
bered rings in molecules 4 and 5, respectively, are nearly
planar. The atomic charge distribution patterns (alternaꢀ
tion of negative and positive charges) for these rings are
typical of rather stable cyclic structures (see Scheme 3).
The bond orders for the newly formed C—Me and N—Mg
bonds (0.90—0.96 and 0.81—1.05, respectively) are charꢀ
acteristic of single bonds. The multiplicities of the carꢀ
bon—nitrogen bonds in molecules 4 and 5 are reduced
to 2 (the corresponding bond orders lie between 1.90
and 1.99). This means that the intermediates 4 and 5
correspond to the situation where the addition of the
Grignard reagent to dinitrile 1 is nearly completed.
Namely, one should only cleave a weak Mg—Me bond
Grignard Molar excess of
reagent Grignard reagent
Duration of
first stage/h
Yield
3a—d (%)
MeMgI
EtMgI
PrMgBr
BuMgBr
2.8
3.2
3.2
3.2
3
3
8
78
75
70
72
12
Note. The concentration of substrate 1 was 0.3 mol L–1
.
route via the fourꢀmembered ring 4. The choice of a parꢀ
ticular reaction pathway is determined by the steric hinꢀ
drances produced by reactants, by the reactant concenꢀ
trations, and by the solvent nature. The conditions for the
synthesis of alkyl(cyanoadamantyl)ketones in 70—78%
yields are substantiated.
(bond order is 0.06 and bond energy is +19.4 kJ mol–1
)
for 4 and eliminate a rather loosely bound MeMgBr (inꢀ
teraction energy is +29.4 kJ mol–1) for 5 (see Table 2) to
obtain ketimine 2.
Experimental
Thus, the reaction of dinitrile 1 with MeMgBr can
involve the formation of two cyclic intermediates, 4 and 5,
the latter being somewhat more energetically favorable.
The choice of particular reaction route seems to be deꢀ
pendent on many factors. In particular, if either molꢀ
ecule 1 or the Grignard reagent molecule produce some
steric hindrances, only one RMgBr molecule is involved
in the formation of the cyclic intermediate and the reacꢀ
tion proceeds via the fourꢀmembered ring 4. For instance,
our calculations of structures of the types 4 and 5 for
the reaction of compound 1 with sterically hindered
PhCH2MgBr showed that the structure with the fourꢀ
membered ring is by 38.3 kJ mol–1 more preferable than
the structure containing the sixꢀmembered ring. The
course of the reaction can also be strongly affected by the
reagent concentrations. For instance, an increase in the
RMgBr concentration should favor the reaction via the
sixꢀmembered ring 5.
IR spectra were obtained on a SpecordꢀM82 spectrometer
(KBr pellets) and 1H NMR spectra were recorded on a Bruker
AMꢀ300 spectrometer (300.13 MHz) in DMSOꢀd6 with Me4Si
as internal standard. GLC analyses were performed on a
Perkin—Elmer—Sigma 2S chromatograph equipped with
a flameꢀionization detector and a stainless steel column
(0.3×100 cm, 10% PPMSꢀ4, Chromatone NꢀAWꢀDMCS,
nitrogen as the carrier gas) with linear temperature programꢀ
ming in the range from 80 to 230 °C. Quantumꢀchemical calcuꢀ
lations were carried out by the MNDOꢀPM3 method10 with
full geometry optimization for all compounds involved in reꢀ
actions.
3ꢀAcetylꢀ1ꢀcyanoadamantane (3a). A reactor equipped with
a stirrer, a reflux condenser with a calcium chloride tube, and a
dropping funnel was charged with Mg (8.16 g, 0.28 mol) and
Et2O (100 mL). Then, MeI (17.4 mL, 0.28 mol) dissolved in
Et2O (50 mL) and compound 1 (18.6 g, 0.10 mol) dissolved in
THF (100 mL) were added successively from a dropping funnel
over a period of 0.5 h. The reaction mass was stirred for 3 h,
cooled, and then water (25 mL), 12% NaOH solution (40 mL),
an additional 30 mL of water and NH4Cl (5 g) was added. The
precipitate that formed was filtered and extracted with THF
(3×150 mL). The extract was evaporated and the stillage resiꢀ
due was distilled in vacuo. The yield was 15.8 g (78%), b.p.
165—168 °C (2 Torr). Recrystallization from PriOH gave the
target product of 98% (GLC monitoring), b.p. 98—102 °C.
Found (%): C, 76.62; H, 8.38; N, 6.81. C13H17NO. Calcuꢀ
Based on the results of our calculations, this can be
done with certainty using a twofold excess of the Grignard
reagent. In practice, the use of a THF—Et2O (1 : 2)
mixture and a 2.8 to 3.2 excess of the Grignard reagent
allows one to obtain the monoaddition products in
70—78% yields (Table 3). It should be noted that a reaꢀ
sonably high (up to 80%) yield of the diaddition product
in the reaction of dinitrile 1 with MeMgI can be obtained
if the molar excess of the Grignard reagent is greater
than 4.4 and the substrate concentration is at least
lated (%): C, 76.81; H, 8.43; N, 6.89. IR spectrum, ν/cm–1
:
2910, 2853 (≡C—H); 2230 (C≡N); 1704 (C=O). 1H NMR,
δ: 1.51—1.75 (m, 12 H, CH2 adamantane); 1.97 (s, 2 H,
CH adamantane); 2.10 (s, 3 H, Me).
0.1—0.2 mol L–1
.
Thus, we proposed the most probable mechanism of
the reaction of 1,3ꢀdicyanoadamantane with Grignard reꢀ
agents, involving the formation of cyclic intermediates 4
and 5. The route via the sixꢀmembered ring 5 is somewhat
more energetically preferable (by 10 kJ mol–1) than the
1ꢀCyanoꢀ3ꢀpropionyladamantane (3b) was synthesized analoꢀ
gously to compound 3a using EtI (25.6 mL, 0.32 mol) instead of
MeI. The yield was 16.2 g (75%), b.p. 179—181 °C (2 Torr).
Recrystallization from PriOH—hexane mixture (1 : 3, v/v) gave