Investigation of the mechanism of decyanation of 2,2-diphenylpropionitrile induced by LiAlH4

Article abstract of DOI:10.1002/1099-1395(200005)13:5<233::AID-POC235>3.0.CO;2-O

The LiAlH₄ reduction of 2,2-diphenylpropionitrile (4) in THF yields a mixture of the hycrocarbon 6 and the expected amine 5. Medium effects and reduction with LiAID₄ suggest a mechanism involving the attack of the hydride reagent on the cyano carbon atom followed by the fragmentation of the formed imine salt intermediate. Copyright

Full text of DOI:10.1002/1099-1395(200005)13:5<233::AID-POC235>3.0.CO;2-O

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2000; 13: 233–236
Investigation of the mechanism of decyanation of
2,2-diphenylpropionitrile induced by LiAlH4
Jean-Marc Mattalia,* Nicolas Bodineau, Jean-Claude NeÂgrel and Michel Chanon
Laboratoire AM3, Faculte des Sciences de Saint-Je roà me, Case 561, 13397 Marseille Cedex 20, France
Received 15 October 1999; revised 7 December 1999; accepted 17 December 1999
ABSTRACT: The LiAlH4 reduction of 2,2-diphenylpropionitrile (4) in THF yields a mixture of the hydrocarbon 6
and the expected amine 5. Medium effects and reduction with LiAlD4 suggest a mechanism involving the attack of the
hydride reagent on the cyano carbon atom followed by the fragmentation of the formed imine salt intermediate.
Copyright 2000 John Wiley & Sons, Ltd.
KEYWORDS: decyanation; diphenylacetonitriles; lithium aluminum hydride; medium effects; carbanions
INTRODUCTION
The reductive decyanation of nitriles has been reported
under a variety of conditions. Different mechanisms
operate depending on the reducing agent used: electron
transfer (alkali metals),1 radical chain (Bu3SnH, SmI2)2,3
and a polar mechanism (RLi, RMgX).4–6 For the
reductive decyanation of a-aminonitriles with LiAlH4,7,8
complex because its decomposition could occur after the
an SN1 mechanism with assistance of the amino group
and formation of an iminium ion intermediate was
proposed. The decyanation is also observed under more
drastic conditions with KOH,9 NaOH10 or NaNH2.11
We previously reported, together with the usual
formation of amine,12 the first cleavage of an a-
sulfonitrile induced by LiAlH4 and presented evidence
to support the participation of radical intermediates.13
The a-sulfonyl group is not compulsory to observe this
type of reactivity14 [in Ref. 14, the relative yields of
amine and hydrocarbon (23 and 58%) should be
inverted]. Black and Doyle found that the treatment of
9-allylfluorene-9-carbonitrile (1) and 2,2,4-triphenyl-
pent-3-enenitrile (2) with LiAlH4 in Et2O yields the
decyanation product together with the expected amine
(Table 1, entries 1 and 2). In contrast, no decyanation was
observed in reactions of 2,2-diphenylpent-4-enenitrile (3)
and similar compounds (Table 1, Entry 3).15
nucleophilic attack by another hydride ion. Two of the
four hydrides are readily available.16
Table 1. Reaction of nitriles 1±4 with LiAlH4
Relative yield (%)b
Entry Nitrile
Conditions
Amine Hydrocarbon
1
2
3
4
5
6
7
1
2
3
4
4
4
4
Slurry, Et2Oa
33
25
100
71
100
51
48
67
75
Slurry, Et2Oa
Slurry, Et2Oa
Slurry, THFc
29
Slurry, Et2O
Solution, THFd
Slurry, Et2O–HMPAe
(90:10, v/v)
49
52
8
9
4
4
LiAlD4, slurry, THFf
Slurry, THF,
45
69
37 (100% D)
31
degassedg
10
4
Solution, THF,
57
43
degassedd,g,h
The authors proposed an attack on the cyano group
followed by elimination of a hydrogen cyanide complex
and formation of a carbanion: the nitriles yielding the
decyanation product are usually precursors of the most
stabilized carbanions (Scheme 1). In Scheme 1, the
structure of the imine salt intermediate could be more
a
See Ref. 15.
b
1
Estimated from H NMR (200 or 400 MHz) of the crude product.
c All reductions carried out were 0.14–0.16 M in nitrile 4 using 3.5–3.8
equiv. of LiAlH4. The reaction time was 24 h.
d
Solutions of LiAlH4 prepared as described by Krishnamurthy and
Brown24 or purchased directly from Aldrich were used.
e The reaction time was 66 h, the excess of LiAlH4 was 2.2 equiv. and
43% of the nitrile did not react.
f Isolated yields.
*Correspondence to: J.-M. Mattalia, Laboratoire AM3, Faculte´ des
Sciences de Saint-Je´roˆme, Case 561, 13397 Marseille Cedex 20,
France.
g
Degassed by three or four freeze–pump–thaw cycles and using an
atmosphere of purified argon.
h
4 equiv. of LiAlH4 were used.
E-mail: jean-marc.mattalia@am3.u-3mrs.fr
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2000; 13: 233–236

234
J.-M. MATTALIA ET AL.
refluxing Et2O (clear solution), reduced 4 to 2,2-
diphenylpropionaldehyde with a 72% isolated yield.19
The reaction of nitrile 4 with 2.2 equiv. of LiAlH4 in
Et2O–HMPA (90:10, v/v) for 66 h gave similar yields of
5 and 6 (Table 1, entry 7). The conversion of 4 was about
57%. For comparison, 4 was totally consumed when the
reaction was run in Et2O with 1.5 equiv. of LiAlH4 for
42 h. Part of the crude product was treated with 2 M HCl
for 1 h in order to convert the eventual imine into 2,2-
diphenylpropionaldehyde. From GC and NMR analysis
the relative yield of aldehyde was estimated to be <1%.
The LiAlD4 reduction of nitrile 4 in THF led to the
formation of deuteriated hydrocarbon 6d (Table 1, entry
8). The 1H NMR spectrum of the crude product shows a
total absence of the quadruplet at ꢀ = 4.14 ppm. Quanti-
tative deuteriation would agree with an SN2 pathway.
Carbons bearing two or three aryl groups (Ar2CHX or
Ar3CX, X = leaving group) exhibit a great reactivity
toward the SN2 and SN1 processes.20 On the other hand,
the cyano group can act as a leaving group. We
mentioned above the reductive decyanation of a-
aminonitriles with the hydride ion via an SN1 polar
pathway.7,8 In the LiAlH4 reduction of nitrile 4, an SN2
process seems unlikely, however, since (1) the cyano
group has a poor leaving ability and (2) the attacked sp3
carbon would be sterically hindered both by the methyl
and the two phenyl substituents.
Scheme 1
Such a mechanism was also proposed for the reductive
decyanation promoted by Grignard reagents6 and by
alkali fusion.9 Phenyl or alkyl cyanides and potassium
cyanate, respectively, were trapped in these reductions.
(The participation of an SN2 reaction at the sp carbon has
also been proposed for these decyanations.17 Gregory et
al.4 suggested a four-membered transition state for the
decyanation of several 1-benzyl-1-cyano-1,2,3,4-tetrahy-
dronaphthalenes with lithium reagents). In such decyana-
tions the carbon atom bearing the cyano group is
substituted by two phenyl groups. Thus, 2,2-diphenyl-
propionitrile and 2-methyl-2-phenylpropionitrile react
with phenyldimethylsilyllithium to give the substitution
product with 17% and 77% yields, respectively.18
Although no mechanistic study was undertaken, the
authors proposed an attack on the cyano group followed
by the displacement of the cumyl anion and subsequent
silylation by the silyl cyanide.
We observed that when 2,2-diphenylpropionitrile (4)
was reduced with LiAlH4 in THF, a mixture of the amine
5 and hydrocarbon 6 was formed (Scheme 2; Table 1,
entry 4). We discuss here, in addition to the pathway
described in Scheme 1, the different mechanisms that
could be involved in this transformation taking into
account medium effects, experiments with LiAlD4 and
the effect of dioxygen.
Solvent and additive effects fit better with the
mechanism described in Scheme 1. In ethereal solvents
LiAlH4 is thought to react as an ion pair.21,22 In Et2O, the
lithium ion is less solvated than in THF (steric effects)
and could stabilize the imine salt intermediate, hence
preventing the expulsion of a carbanion. Another way to
‡
modulate the participation of Li in the reaction is to add
HMPA, a powerful dissociating and basic solvent.23 With
HMPA, the complexation of the nitrile group with the
lithium ion would be diminished and the reactivity of 4
would decrease in the absence of electrophilic assistance.
Moreover, the imine salt intermediate, not being
stabilized by the lithium cation, would see its fragmenta-
tion favored. The steric hindrance of the nitrile group of 4
probably favors the first addition of hydride over the
second, leading to an increased ability of the intermediate
to expulse a carbanion.16
Scheme 2
The observed deuterium incorporation is consistent
with an internal hydride capture: as soon as the carbanion
is formed, it could react with hydrogen cyanide (or
deuterium cyanide) in a solvent cage leading to 6d. Clear
solutions of LiAlH4 favored the decyanation over the
amine formation (Table 1, entries 6 and 10). Krishna-
murthy and Brown24 observed that clear solutions are
more reactive than slurries toward the nucleophilic
substitution of halides. They proposed that insoluble
impurities might coat the undissolved LiAlH4 and hinder
its reactivity. We suggest that some metallic impurities
could complex the imine salt intermediate and prevent
the formation of a carbanion.
RESULTS AND DISCUSSION
The polar pathway
It is surprising that the nitrile 4 exhibits such a reactivity
in THF whereas 2,2-diphenylpent-4-enenitrile (3), a very
closely related substrate, does not in Et2O. A solvent
effect seems to operate here. Indeed, when 2,2-diphenyl-
propionitrile (4) was reduced with LiAlH4 in Et2O, no
decyanation product was formed, in agreement with
Black and Doyle’s observations (Table 1, entry 5). In this
field, Nagai and Hirata, using 0.3 equiv. of LiAlH4 in
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2000; 13: 233–236

DECYANATION OF 2,2-DIPHENYLPROPIONITRILE BY LiAlH4
235
a
better reducing agent than AlH4
(EHOMO
=
1.182 eV, IP = 4.225 eV).34 We proposed such a
possibility for the LiAlH4 reduction of a gem-disubsti-
tuted cyclopropyl bromide.35 This reduction was ob-
served only under strictly anaerobic conditions, dioxygen
being an extremely efficient scavenger of the chain. We
therefore carried out the reduction under strictly
anaerobic conditions (entries 9 and 10). The yield of
hydrocarbon 6 did not significantly change, hence the
chain proposed for the disubstituted bromide does not
seem to apply here. (We have noticed that when the
reaction was performed in a Schlenk tube connected to a
balloon filled with nitrogen, up to 19% of 1,1-
diphenylethanol can be formed in the medium.36 We
verified that the alcohol was produced from dioxygen
which can diffuse into the balloon. The origin of 1,1-
diphenylethanol has not been clearly established but is
consistent with the proposed mechanism since the
reaction of carbanions with dioxygen is very fast37). In
an attempt to trap a radical intermediate, we have
introduced the hex-5-enyl group as radical probe and
reduced 2,2-diphenylhept-6-enenitrile with LiAlH4 in
THF under usual conditions. No cyclized hydrocarbon
was formed (<1%).
Scheme 3
The radical pathway
There are several precedents of reductions involving
simultaneously radical and polar pathways. Ashby and
Welder reported evidence suggesting the coexistence of
the SN2 and SET routes in the reaction of alkyl iodides
with LiAlD4.25,26 Such a proposition, applied to 4, is
described in Scheme 3. It is therefore necessary to
examine more closely the characteristics of this reaction
which could hint at a radical participation. If free radicals
were involved, we would expect some hydrogen atom
incorporation during the LiAlD4 reduction, the radical
competitively abstracting a hydrogen atom from the
metal hydride and from the solvent. Indeed, Newcomb
suggested that hydrogen abstraction from LiAlH4 is at
least one order of magnitude slower than that from
THF.27 In the reduction of 4, no hydrogen atom
incorporation was observed when only 1.3 equiv. of
LiAlD4 was used.
CONCLUSION
We suggest that the nucleophilic attack of the hydride
reagent on the nitrile group followed by the formation of
a carbanion is the major pathway for the decyanation of
2,2-diphenylpropionitrile induced by LiAlH4. From
literature data, nitriles bearing two phenyl groups on
the a-position readily give C—CN bond cleavage in the
presence of LiAlH4 but also with Grignard reagents or
sodium amide. These groups stabilize the carbanionic
intermediate and weaken the C—C bond. Such a
reactivity is reminiscent of the so-called ‘cyanophilic’
reaction which involves the reaction of tetracyano-
methane with phenolates and thiophenolates to give aryl
cyanates and thiocyanates, respectively.38,39
Recently, a SET was proposed in an anomalous
reductive decyanation of a-tosyloxy-a-(trifluoromethyl)-
phenylacetonitrile with NaBH4.28 Eberson estimated the
Á
standard potential of AlH4 /AlH4 redox couple as E°ox
=
(0.1–0.3 V) vs NHE in THF.29 House discussed
possible electron transfer between carbanions and
unsaturated carbonyl compounds in terms of the
difference in standard electrode potentials. He concluded,
that for E°red E°ox more positive than 0.4 V (equiva-
lent to DG° < 9.2 kcal mol 1) electron transfer would be
feasible.30 House’s rule, applied to our system, taking
E°ox
to the nitrile 4 would be possible if E°red of 4 was more
positive than 0.7 V. The reduction potential of
diphenylacetonitrile, a substrate closely related to 4, is
E°red
1.82 V (vs SCE, CH3CN, 1.58 V vs NHE).31
= 0.3 V, the electron transfer from the hydride ion
EXPERIMENTAL
General. Et2O, THF and HMPA were dried over sodium
and then distilled. Glassware was oven-dried prior to use.
Product mixtures were analyzed by gas chromatography
(GC) and by integration of the 1H NMR spectrum (200 or
400 MHz). GC (30 m DB1 column or 25 m BPX5
column) was performed on a Fisons 8000 Series
apparatus equipped with a flame ionization detector.
Analysis by GC: column temperature, 150°C for 3 min,
increased at 5°C min
injector temperature, 300°C; detector temperature,
280°C; vector gas, helium. Gas chromatography coupled
with mass spectrometry (GC–MS) was performed on a
=
Thus nitrile 4 is, by itself, not sufficiently oxidizing and
participation of a SET route seems unlikely. One must
underline, however, the limits of such a kind of argument
because salts effects may drastically modify the redox
relationship in the donor–acceptor couple.32,33 An
.
electron transfer could also occur from AlH3 formed
1
in situ leading to a radical chain pathway (Scheme 3). We
have calculated HOMO energies and ionization poten-
to 240°C, held for 10 min;
.
.
tials of both AlH4 /AlH4 and AlH3/AlH3 couples.
AlH3 . (EHOMO = 2.028 eV, IP = 0.385 eV) appears to be
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2000; 13: 233–236

236
J.-M. MATTALIA ET AL.
Ribermag R 10.10C system (vector gas, helium; tension
acceleration, 70 eV; injector, 300°C; interface, 300°C;
source, 150°C).
REFERENCES
1. Arapakos PG. J. Am. Chem. Soc. 1967; 89: 6794–6796.
2. Curran DP, Seong CM. Synlett 1991; 107–108.
3. Kang H-Y, Hong WS, Cho YS, Koh HY. Tetrahedron Lett. 1995;
36: 7661–7664.
General procedure for heterogeneous LiAlH4 reduc-
tions. A Schlenk tube containing LiAlH4 (Fluka, >97%,
0.483 g, 12.7 mmol) was purged with nitrogen and THF
(16.2 ml) was added. The nitrile 4 (Aldrich, 97%,
0.760 g, 3.56 mmol) in THF (7.8 ml) was then transferred
into this heterogeneous slurry. The mixture was stirred at
room temperature under a nitrogen atmosphere for 24 h.
Then, with stirring and cooling, water (0.4 ml), a 20%
solution of NaOH (0.4 ml), water (0.7 ml) and MgSO4
were added in succession. After filtration the residue was
washed with THF and the solvent carefully distilled.
Chromatography on silica gel (elution: light petroleum)
afforded 6 (27% isolated yield).40 1H NMR (CDCl3):
ꢀ = 1.63 (d, J = 7.2 Hz, 3 H), 4.14 (q, J = 7.2 Hz, 1 H),
7.10–7.34 (m, 10 H). GC–MS analysis was consistent
with data described in the literature. C14H14 (182.27):
calculated C 92.26, H, 7.74; found: C 92.30, H 7.53%.
Further elution with ethyl acetate and 20% ethanol in
ethyl acetate gave 5 (52% isolated yield). 1H NMR
(CDCl3): ꢀ = 1.53 (s, 2 H), 1.68 (s, 3H), 3.32 (s, 2 H),
7.13–7.36 (m, 10 H). GC–MS (70 eV): m/z (%) 211 (1.9)
4. Gregory GB, Johnson AL, Ripka WC. J. Org. Chem. 1990; 55:
1479–1483.
5. Kulp SS, Romanelli A. Org. Prep. Proced. Int. 1992; 24: 7–12.
6. Blicke FF, Tsao E-P. J. Am. Chem. Soc. 1953; 75: 5587–5590.
7. Rajagopalan P, Advani BG. Tetrahedron Lett. 1965; 2197–2200.
8. Chauvie`re G, Tchoubar B, Welvart Z. Bull. Soc. Chim. Fr. 1963;
1428–1433.
9. Berkoff CE, Rivard DE, Kirkpatrick D, Ives JL. Synth Commun.
1980; 10: 939–945.
10. Milart P, Wilamowski J, Sepiol JJ. Tetrahedron 1998; 54: 15643–
15656.
11. Klenk MM, Suter CM, Archer S. J. Am. Chem. Soc. 1948; 70:
3846–3850.
12. Seyden-Penne J. In Reductions by the Alumino- and Borohydrides
in Organic Synthesis (2nd edn). Wiley-VCH: New York, 1997;
149–156.
13. Mattalia J-M, Berchadsky Y, Pe´ralez E, Ne´grel J-C, Tordo P,
Chanon M. Tetrahedron Lett. 1996; 37: 4717–4720.
14. Mattalia J-M, Samat A, Chanon M. J. Chem. Soc., Perkin Trans. 1
1991; 1769–1770.
15. Black DSC, Doyle JE. Aust. J. Chem. 1978; 31: 2323–2326.
16. Brown HC, Garg CP. J. Am. Chem. Soc. 1964; 86: 1085–1089.
17. King JA, Meltzer RI, Doczi J. J. Am. Chem. Soc. 1955; 77: 2217–
2223.
18. Fleming I, Solay M, Stolwijk F. J. Organomet. Chem. 1996; 521:
121–124.
‡
(M ), 181 (17.0), 103 (14.5), 44 (27.8), 30 (100.0).
19. Nagai W, Hirata Y. J. Org. Chem. 1989; 54: 635–640.
20. March J. In Advanced Organic Chemistry. Reactions, Mechanisms
and Structure (4th edn). Wiley: New York, 1992; 339–348.
21. Loupy A, Seyden-Penne J. Tetrahedron 1980; 36: 1937–1942.
22. Krishnamurthy S. J. Org. Chem. 1980; 45: 2550–2551.
23. Lefour J-M, Loupy A. Tetrahedron 1978; 34: 2597–2605.
24. Krishnamurthy S, Brown HC. J. Org. Chem. 1982; 47: 276–280.
25. Welder CO, Ashby EC. J. Org. Chem. 1997; 62: 4829–4833.
26. Ashby EC, Welder CO. J. Org. Chem. 1997; 62: 3542–3551.
27. Newcomb M. Acta Chem. Scand., Ser. B 1990; 44: 299–310.
28. Ne´meth G, Posza´va´cz L, Bo´zsing D, Simig G. J. Fluorine Chem.
1996; 78: 87–89.
29. Eberson L. Acta Chem. Scand., Ser. B 1984; 38: 439–459.
30. House HO. Acc. Chem. Res. 1976; 9: 59–67.
31. Homma K, Yamada S. Chem. Pharm. Bull. 1997; 45: 1198–1201.
32. Astruc D. In Electron Transfer and Radical Processes in
Transition Metal Chemistry, VCH: New York, 1995; 337–346.
33. Fukuzumi S. Bull. Chem. Soc. Jpn. 1997; 70: 1–28.
34. St-Amant A., PhD Dissertation, University of Montreal, 1991.
35. Pe´ralez E, Ne´grel J-C, Chanon M. Tetrahedron Lett. 1995; 36:
6457–6460.
C15H17N (211.14): calculated C 85.26, H 8.11, N 6.63;
found C 83.39, H 8.36, N 6.65%. Under these conditions,
the conversion of nitrile 4 was always better than 95%.
The mass balance was about 95%. The agreement
between NMR analysis (relative yields 6:5 = 29:71) and
yields of isolated products was within Æ5%.
1,1-Diphenyl-1-deuterioethane (6d) and 2,2-diphenyl-
1,1-dideuteriopropanamine (5d). The nitrile 4 (0.201 g,
0.94 mmol) in THF (2 ml) was added to an heterogeneous
slurry of LiAlD4 (Fluka, D ꢀ 99%, 0.150 g, 3.57 mmol)
in THF (4.2 ml). After the usual work-up and chromato-
graphy on silica gel (light petroleum as eluent),
deuteriated hydrocarbon 6d was isolated (37% yield).
‡
GC–MS (70 eV): m/z (%) 183 (42.1) (M ), 168 (100.0),
1
166 (29.9), 153 (13.7). H NMR (CDCl3): ꢀ = 1.63 (s, 3
36. Dimitrov V, Stanchev S, Milenkov B, Nekiforov T, Demirev T.
Synthesis 1991; 228–232.
37. Chanon M, Julliard M, Santamaria J, Chanon F. New J. Chem.
1992; 16: 171–201.
38. Zefirov NS, Makhon’kov DI. Chem. Rev. 1982; 82: 615–624.
39. Zefirov NS, Chapovskaya NK, Makhon’kov DI. J. Org. Chem.
USSR 1976; 12: 608–612.
H), 7.10–7.34 (m, 10 H). Further elution with 20%
ethanol in ethyl acetate gave 5d (45% isolated yield).
‡
GC–MS (70 eV): m/z (%) 213 (2.1) (M ), 181 (40.6), 165
(23.5), 103 (17.5), 77 (15.8), 32 (100.0). 1H NMR
(CDCl3): ꢀ = 1.45 (s, 2H), 1.67 (s, 3 H), 7.05–7.36 (m, 10
H).
40. Popielarz R, Arnold DR. J. Am. Chem. Soc. 1990; 112: 3068–3082.
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2000; 13: 233–236