Convincing Evidence, Not Involving Cyclizable Radical Probes, That the Reaction of LiAlH4 with Hindered Alkyl Iodides Proceeds Predominantly by a Single Electron Transfer Pathway

Article abstract of DOI:10.1021/jo961651+

Previous workers have maintained that evidence for the radical nature of the reaction of LiAlH₄ with sterically hindered alkyl iodides is due to radical initiation by impurities followed by a halogen atom radical chain process involving the cyclizable alkyl iodide probe and that the reduction of the C-I bond actually proceeds by an S_N2 pathway. In order to resolve the validity of this explanation, 1-iodo-2,2-dimethylhexane (the saturated counterpart of the cyclizable probe), which is not capable of this halogen atom radical chain process, was allowed to react with LiAlD₄. The reduction product, 2,2-dimethylhexane, contained only 4-76% deuterium depending on the conditions of the reaction. This result is consistent with the reaction proceeding by a SET process via a radical intermediate and is inconsistent with an S_N2 pathway. We have determined the influence of the nature of the reaction on the type of reactor surface (Pyrex, Teflon, stainless steel, and quartz) used in the reaction. We have also studied the influence of AlD₃ (a byproduct in the reduction) in the mechanistic evaluation of this reaction.

Full text of DOI:10.1021/jo961651+

3542
J . Org. Chem. 1997, 62, 3542-3551
Con vin cin g Evid en ce, Not In volvin g Cycliza ble Ra d ica l P r obes,
Th a t th e Rea ction of LiAlH4 w ith Hin d er ed Alk yl Iod id es P r oceed s
P r ed om in a n tly by a Sin gle Electr on Tr a n sfer P a th w a y†,‡
E. C. Ashby* and Catherine O. Welder
School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, Georgia 30332
Received August 27, 1996X
Previous workers have maintained that evidence for the radical nature of the reaction of LiAlH4
with sterically hindered alkyl iodides is due to radical initiation by impurities followed by a halogen
atom radical chain process involving the cyclizable alkyl iodide probe and that the reduction of the
C-I bond actually proceeds by an SN2 pathway. In order to resolve the validity of this explanation,
1-iodo-2,2-dimethylhexane (the saturated counterpart of the cyclizable probe), which is not capable
of this halogen atom radical chain process, was allowed to react with LiAlD4. The reduction product,
2,2-dimethylhexane, contained only 4-76% deuterium depending on the conditions of the reaction.
This result is consistent with the reaction proceeding by a SET process via a radical intermediate
and is inconsistent with an SN2 pathway. We have determined the influence of the nature of the
reaction on the type of reactor surface (Pyrex, Teflon, stainless steel, and quartz) used in the reaction.
We have also studied the influence of AlD3 (a byproduct in the reduction) in the mechanistic
evaluation of this reaction.
In tr od u ction
nature of LAH in the reduction of sterically hindered
alkyl halides.5e The methodology used to establish the
mechanism of the reduction included the following: (1)
direct spectroscopic evidence of radical intermediates by
electron paramagnetic resonance (EPR) spectroscopy, (2)
cyclization of cyclizable alkyl halides that contain the
5-hexenyl group, (3) use of radical traps, and (4) stereo-
chemical studies. The extent of electron transfer was
studied as a function of solvent, substrate, leaving group,
and the hydride reagent.
Upon study of the reaction of unhindered primary
cyclizable probes, 6-halo-1-hexenes (X ) I, Br, Cl), with
LiAlH4, we reported that no cyclized products were
observed. Also, deuterium incorporation in the uncyc-
lized product (1-hexene) was quantitative when LiAlD4
(LAD) was used as the nucleophile. It was concluded that
either an SN2 mechanism best describes the reaction or
the radical species involved collapse to product substan-
tially faster than cyclization to the cyclopentylmethyl
radical (kcyclization ) 2.5 × 105 s-1 at 25 °C).19
Of all the nucleophiles1-18 that have been reported as
one-electron donors, only lithium aluminum hydride
(LAH) has been questioned as to its electron donor ability.
In 1984 we published an article that supports the SET
† The initial studies were communicated earlier. Ashby, E. C.;
Welder, C. O. Tetrahedron Lett. 1995, 36, 7171. Ashby, E. C.; Welder,
C. O.; Doctorovich, F. Tetrahedron Lett. 1993, 34, 7235.
‡ This paper is dedicated to Professor Ernest Eliel who introduced
E.C.A. to LiAlH4 42 years ago as a student at the University of Notre
Dame. Professor Eliel has been a role model for many, a man of
excellence in scientific investigation and teaching, and a man of
character and integrity.
X Abstract published in Advance ACS Abstracts, May 1, 1997.
(1) Kornblum, N.; Michel, R. E.; Kerber, R. C. J . Am. Chem. Soc.
1966, 88, 5660.
(2) Russell, G. A.; Danen, W. C. J . Am. Chem. Soc. 1966, 88, 5663.
(3) Kim, J .; Bunnett, J . F. J . Am. Chem. Soc. 1970, 92, 7463.
(4) (a) Ashby, E. C.; Bowers, J .; DePriest, R. Tetrahedron Lett. 1980,
21, 3541. (b) Ashby, E. C.; Lopp, I. G.; Buhler, J . D. J . Am. Chem. Soc.
1975, 97, 1964. (c) See also: Holm, T.; Crossland, I. Acta Chem. Scand.
1971, 25, 1158.
(5) (a) Ashby, E. C.; Deshpande, A. K. J . Org. Chem. 1994, 59, 3798.
(b) Ashby, E. C.; Pham, T. N.; Amrollah-Madjdabadi, A. J . Org. Chem.
1991, 56, 1596. (c) Ashby, E. C. Acc. Chem. Res. 1988, 21, 414. (d)
Ashby, E. C.; Pham, T. N. Tetrahedron Lett. 1987, 28, 3197. (e) Ashby,
E. C.; DePriest, R. N.; Goel, A. B.; Wenderoth, B.; Pham, T. N. J . Org.
Chem. 1984, 49, 3545. (f) Ashby, E. C.; Wenderoth, B.; Pham, T. N.;
Park, W. S. J . Org. Chem. 1984, 49, 4505. (g) Ashby, E. C.; DePriest,
R. N.; Goel, A. B.; Wenderoth, B.; Pham, T. N. Tetrahedron Lett. 1983,
24, 2825. (h) Ashby, E. C.; Goel, A. B. Tetrahedron Lett. 1981, 22, 1879.
(6) (a) Ashby, E. C.; Goel, A. B. Tetrahedron Lett. 1981, 22, 4783.
(b) Ashby, E. C.; Goel, A. B.; DePriest, R. N. Tetrahedron Lett. 1981,
22, 3729. (c) Ashby, E. C.; DePriest, R. N.; Goel, A. B. Tetrahedron
Lett. 1981, 22, 1763.
(7) (a) Ashby, E. C.; Su, W. Y.; Pham, T. N. Organometallics 1985,
4, 1493. (b) Ashby, E. C.; DePriest, R. N.; Su, W. Y. Organometallics
1984, 3, 1718. (c) See also Smith, G. F.; Kuivila, H. G.; Simon, R.
Sultan, L. J . Am. Chem. Soc. 1981, 103, 833.
(8) (a) Ashby, E. C.; Gurumurthy, R.; Ridlehuber, R. W. J . Org.
Chem. 1993, 58, 5832. See also (b) Santigo, A. N.; Iyer, V. S.; Adcock,
W.; Rossi, R. A. J . Org. Chem. 1988, 53, 3016. (c) Bangerter, B. W.;
Beatly, R. P.; Kouba, J . K.; Wreford, S. S. J . Org. Chem. 1977, 42,
3247.
(9) (a) Ashby, E. C.; Park, W. S.; Goel, A. B.; Su, W. Y. J . Org. Chem.
1985, 50, 5184. (b) See also Russell, G. A.; J anzen, E. G.; Strom, E. T.
J . Am. Chem. Soc. 1964, 86, 1807.
(10) (a) Ashby, E. C.; Argyropoulos, J . N. J . Org. Chem. 1986, 51,
3593. (b) Ashby, E. C.; Goel, A. B.; DePriest, R. N. J . Org. Chem. 1981,
46, 2429.
(12) (a) Ashby, E. C.; Pham, T. N. J . Org. Chem. 1987, 52, 1291.
See also: (b) Russell, G. A.; Lamson, D. W. J . Am. Chem. Soc. 1969,
91, 3967. (c) Fischer, H. J . Phys. Chem. 1969, 73, 3834. (d) Ward, H.
R.; Lawler, R. G.; Cooper, R. A. J . Am. Chem. Soc. 1969, 91, 746. (e)
Lepley, A. R.; Landau, R. L. J . Am. Chem. Soc. 1969, 91, 748.
(13) (a) Ashby, E. C.; Coleman, D. J . Org. Chem. 1987, 52, 4554. (b)
Ashby, E. C.; DePriest, R. N.; Tuncay, A.; Srivastava, S. Tetrahedron
Lett. 1982, 23, 5251.
(14) (a) Ashby, E. C.; Park, B.; Patil, G. S.; Gadru, K.; Gurumurthy,
R. J . Org. Chem. 1993, 58, 424. (b) Ashby, E. C.; Goel, A. B.; DePriest,
R. N. Tetrahedron Lett. 1981, 22, 4355.
(15) Ashby, E. C.; Wenderoth, B.; Pham, T. N.; Park, W. S. J . Org.
Chem. 1984, 49, 4505.
(16) Tolbert, L. M.; Sun, X. J .; Ashby, E. C. J . Am. Chem. Soc. 1995,
117, 2681.
(17) (a) Ashby, E. C.; Argyropoulos, J . N. J . Org. Chem. 1985, 50,
3274. (b) Ashby, E. C.; Argyropoulos, J . N. Tetrahedron Lett. 1984,
25, 7. (c) Ashby, E. C.; Park, W. S. Tetrahedron Lett. 1983, 24, 1667.
(18) For reactions involving 1-halonorbornanes with Me3Sn-, Ph2P-,
N(iPr)2-, PhS-, and the 2-nitropropyl anion, see: Ashby, E. C.; Sun,
X.; Duff, J . L. J . Org. Chem. 1994, 59, 1270. For reactions involving
cyclizable norbornene probes with Na, Mg, sodium naphthalenide, and
Bu2SnH, see: Ashby, E. C.; Pham, T. N. Tetrahedron Lett. 1984, 25,
4333.
(19) Chatgilialoglu, C.; Ingold, K. U.; Scaiano, J . C. J . Am. Chem.
Soc. 1981, 103, 7739. For kinetic parameters for several primary alkyl
radical clocks see Griller, D.; Ingold, K. U. Acc. Chem. Res. 1980, 13,
317.
(11) (a) Ashby, E. C.; Coleman, D.; Gamasa, M. J . Org. Chem. 1987,
52, 4079. (b) Ashby, E. C.; Coleman, D. T., III; Gamasa, M. P.
Tetrahedron Lett. 1983, 24, 851.
S0022-3263(96)01651-9 CCC: $14.00 © 1997 American Chemical Society

Reactions of 1-Halo-2,2-dimethylhexanes with LiAlD4
J . Org. Chem., Vol. 62, No. 11, 1997 3543
Sch em e 1
system and that the SN2 pathway is believed to be a
minor pathway.
It is important to note that the radical chain process
shown in Scheme 1, steps d, f, and g, was discussed by
us in detail in 1984.5e We reported the reaction of 6-iodo-
1-heptene (9) with LAH in 1:0.1 ratio (eq 2). This
experiment was specifically designed to uncover the
extent of the halogen atom radical chain process, and
indeed, the formation of the cyclized iodide 10 in 70%
yield provided this evidence.
In the past 10 years several reports20 have appeared
that suggest that a trace amount of radical initiator can
result in a significant yield of cyclized alkyl halide formed
via a halogen atom radical chain process and that the
cyclized alkyl halide could then react with LAH by a
conventional polar (SN2) process.20b In order to address
this possibility, we independently synthesized 1-(iodo-
methyl)-3,3-dimethylcyclopentane (7).5d If all of the
cyclized hydrocarbon is a result of the reduction of alkyl
iodide 7 in an SN2 fashion, then when LAD is substituted
for LAH as nucleophile, the following two reactions (eqs
3 and 4) should give identical results as to the ratio of 3
It was necessary to add two methyl groups to the
5-position of the 6-halo-1-hexene system in order to slow
the competing SN2 reaction and/or the competing hydro-
gen atom abstraction reaction in order to obtain evidence
for SET. Indeed, significant evidence for electron trans-
fer was observed for the reaction of 6-iodo-5,5-dimethyl-
1-hexene (1) with LAH as evidenced by formation of the
cyclized product 1,1,3-trimethylcyclopentane (3) in high
yield (eq 1). (Less cyclized product was observed with
to 3d 1
:
the corresponding bromide and no cyclized product was
observed for the corresponding chloride or tosylate.)
Studies employing LAD as the nucleophile reacting with
iodide 1 show significant protium incorporation in both
the uncyclized product 2 and the cyclized product 3.
When dicyclohexylphosphine (DCPH), a radical trapping
agent, was added to the reaction, significantly more
radical intermediate was trapped by the DCPH as
evidenced by lower deuterium incorporation in products
2 and 3.
We found that the reduction of 7 gave 99% 1,1,3-
trimethylcyclopentane containing 98% D (eq 3). There-
fore, no evidence of SET was observed in the reduction
of 7 by LAD. However, when iodide 1 was allowed to
react with LAD, 91% 1,1,3-trimethylcyclopentane was
observed, but it contained only 57% D (eqs 1 and 4).
These results show that products 3 and 3d 1 do not result
solely from 7 or else the same ratio of 3 to 3d 1 would have
been observed in eqs 3 and 4. Clearly a radical inter-
mediate is formed in eq 4 which abstracts hydrogen
atoms from a source other than LAD. These data are
contrary to the suggestion that the reduction of 1 by LAH
proceeds by an SN2 process. Also, if the radical anion 4
is initiated by an impurity in eq 4, then the same
impurity would be expected to initiate the formation of
radical anion 8 in eq 3 while it clearly does not. These
In 1991 the mechanism shown in Scheme 1 was
published5b for the reduction of 6-iodo-5,5-dimethyl-1-
hexene (1), specifying the hydrogen atom radical chain
nature of the reaction. It was proposed that, in step a,
LAH transfers one electron to alkyl iodide 1 to form the
corresponding radical anion 4, which rapidly dissociates
in step b to form the radical 5. Radical 5 can enter a
halogen atom radical chain process as shown in steps d,
-
f, and g; it can react with AlH4 (step c) in a hydrogen
atom radical chain process yielding alkene 2; or it can
cyclize (step d) to form radical 6. Cyclized radical 6 can
then undergo a hydrogen atom radical chain process (step
e) to form alkane 3. We have also reported5b that the
AlH3 radical anion is the major one-electron donor in the
(20) (a) Newcomb, M.; Curran, D. P. Acc. Chem. Res. 1988, 21, 206.
(b) Newcomb, M.; Kaplan, J .; Curran, D. P. Tetrahedron Lett. 1988,
29, 3451. (c) Park, S. U.; Chung, S. K.; Newcomb, M. J . Org. Chem.
1987, 52, 3275. (d) Newcomb, M.; Sanchez, R. M.; Kaplan, J . J . Am.
Chem. Soc. 1987, 109, 1195.

3544 J . Org. Chem., Vol. 62, No. 11, 1997
Ashby and Welder
results are inconsistent with the notion that the reaction
of 1 or 7 is initiated by impurities.
substantial amounts of rearranged products were de-
tected. He warns that adventitious impurities or side
reactions can be responsible for initiation of a radical
chain isomerization process in reactions of any alkyl
halide probe (even though he only studied unhindered
primary probes) with boron or aluminum hydride reduc-
ing agents.20c
In 1986 Newcomb and co-workers studied the rates of
cyclization of 6-substituted hex-5-en-1-yl radicals (eq 5).21
They reported that radicals are stabilized by electron
acceptors (such as CN) and donors (such as OCH3) and
that radical 12d , which incorporates both an acceptor and
a donor in the same molecule, has greater stability than
the sum of the individual components. Therefore, by
Recall that as early as 1984 we reported no evidence
of SET when the primary unhindered 6-halo-1-hexenes
(X ) I, Br, Cl) were reduced by various metal hydrides,
including LAH.5e Therefore, it was not surprising to us
when the report appeared stating that there is no
evidence of SET with primary unhindered probes 13a
and 13b. Additionally, we were surprised that an
unhindered primary probe, such as 13a , would be used
to refute the mechanism suggested for a hindered pri-
mary probe. After all, our reported mechanism of SET
was reported for hindered primary alkyl halides. Be-
cause of this improper comparison we modified the
captodative probe 13a . Insertion of the geminal dimethyl
group into probe 13a yielded the neopentyl type probe
7-iodo-6,6-dimethyl-2-methoxy-2-heptenenitrile (17), which
still incorporates the captodative system proposed by
Newcomb, but also adds the steric hindrance that is
usually necessary to observe SET in primary alkyl
iodides. This is the probe that should have been studied
in order to challenge our initial report since we did not
work with a unhindered primary system but with a
hindered neopentyl type system. When probe 17 was
allowed to react with LAH, the mechanism of the reaction
was altered significantly as compared either to probe 13a
or the original neopentyl type probe 1. As shown in eq
7, a Michael addition of hydride to probe 17 resulted in
definition, captodative stabilization exists for 12d . In
1987 the same group used the new cyclizable radical 11d
to study the reactions of unhindered primary alkyl
iodides and bromides with boron and aluminum hydride
reducing agents to determine whether the reductions
proceeded by an SN2 or an SET pathway.20c According
to Newcomb, due to the faster cyclization of radical 11d
to radical 12d over most other probe rearrangements,
“problems with trapping of free radical intermediates
before rearrangement should not arise with [11d ]”.22
More importantly, Newcomb and co-workers “believed
that cyclic radical [12d ] would be so stable that it would
be incapable of abstracting halogen from [13] in a radical
chain isomerization propagation step”22 (eq 6). Since the
halogen atom radical chain process should then be
inoperative, it was hoped that quantitation of the extent
of SET would be possible.
anion 18 which then cyclized to the six-membered ring
product 19. Therefore, probe 17 cannot be used in a
manner analogous to probe 1 because in probe 17,
reaction occurred at the R,â-unsaturated nitrile, whereas
in probe 1, reaction occurred at the carbon-iodine bond.23
(Undesirable reaction at the R,â-unsaturated nitrile also
occurred in reactions involving the uncyclized probe 13.)
Data from the present report will conclusively support
our position that the reaction of LiAlH4 with 1 is an SET
process and not an SN2 process. This is accomplished
by removing the possibility of a halogen atom radical
chain process by allowing the saturated analog of 1 to
react with LiAlD4. Thus protio product formed in the
reaction can only be the result of a radical reaction with
the solvent.
To this end, 1-iodo-2,2-dimethylhexane (20), the satu-
rated counterpart of 1, was chosen for study. Since the
product of a SET reaction with LAH would be indistin-
guishable from the SN2 product [2,2-dimethylhexane
(22)], LAD was necessarily chosen as the nucleophile in
the study of the saturated probe. Distinction between
SET and SN2 is thus possible. For a reaction involving
radical intermediates, hydrogen atoms (from solvent)
Newcomb’s stated premises are that (1) all of the alkyl
halide that is converted to radical 11d will lead to
rearranged radical 12d , (2) radical 12d cannot abstract
halogen atom from 13, and (3) the uncyclized reduction
product 2-methoxy-2-heptenenitrile (15) could not arise
from trapping of radical 11d before cyclization because
substantially slower radical probes have rearranged in
the presence of various boron and aluminum hydrides.
No radical-derived cyclized products were observed;
therefore, it was concluded that SET from the reducing
agent to the halide to give a free radical was not an
important process. Newcomb further states that in
previous mechanistic studies it is reasonable to suspect
that only limited amounts of SET occurred even when
(21) Park, S. U.; Chung, S. K.; Newcomb, M. J . Am. Chem. Soc. 1986,
108, 240.
(22) Park, S. U.; Chung, S. K.; Newcomb, M. J . Org. Chem. 1987,
52, 3276.
(23) Sun, Xiao-J ing. The Extent and Importance of Single Electron
Transfer in Organic Reactions. Ph.D. Dissertation, Georgia Institute
of Technology, 1994.

Reactions of 1-Halo-2,2-dimethylhexanes with LiAlD4
J . Org. Chem., Vol. 62, No. 11, 1997 3545
Ta ble 1. Va r ia tion of Stoich iom etr y in th e Rea ction of
1-Iod o-2,2-d im eth ylh exa n e (20) w ith LAD in THF a
Ta ble 2. Va r ia tion of Con cen tr a tion in th e Rea ction of
1-Iod o-2,2-d im eth ylh exa n e (20) w ith LAD in THF a
products
products
exp 20:LAD time, h % 20 % 22 (% D) % dimers MBb
exp
[20], M
time, h
%20
%22 (% D)
MBb
1
1:20
1
5
7
1
5
74
21
11
88
60
33
5
28
90
74
71
61
61
26
68
trace
trace
trace
100
89
93
92
93
91
92
95
97
93
93
91
92
1
2
3
0.27
0.070
0.035
22
8
20
0
0
2
100 (67)
100 (28)
86 (21)
100
100
88
82 (72)
4 (48)
33 (42)
58 (39)
84 (34)
67 (6)
7 (5)
19 (8)
22 (5)
30 (3)
31 (4)
2
1:5
a
All reactions were carried out in an Ar atmosphere glovebox
in the absence of light at room temperature and at a 20:LAD ratio
of 1:5. Reactions were carried out in used Pyrex flasks using
trace
trace
3
8
24
20
1
5
8
b
Teflon-coated stir bars. MB is material balance.
3
4
1:1
1:0.2
trace
trace
trace
trace
trace
72% D was formed. Therefore, at least 28% of the
hydrocarbon formed was derived from radical 21 that
abstracted a hydrogen atom from the solvent. Obviously,
more than a trace amount of radicals were formed even
in the absence of the halogen atom radical chain process
described in Scheme 1 (steps d, f, and g). When the
stoichiometry was reduced from a 20-fold excess of LAD
to a 5-fold excess (Table 1, experiment 2), an even lower
deuterium content was observed in the resulting hydro-
carbon (34%). The trend of an increasing ratio of non-
deuterated alkane 22 to deuterated alkane 22d 1 contin-
ued as the stoichiometry of LAD dropped further to 1:1
(experiment 3) and 1:0.2 (experiment 4) as expected.
Indeed, at the stoichiometries of 1:1 and 1:0.2, the
deuterium content of the product dropped to 3-8%,
indicating that the majority of the product (>90%) is
formed via a radical intermediate. These results are
inconsistent with an SN2 process.
24
48
a
All reactions were carried out in an Ar atmosphere glovebox
in the absence of light at room temperature and at a concentration
of 0.070 M with respect to 20. Reactions were carried out in used
Pyrex flasks using Teflon-coated stir bars. MB is material
balance.
b
and/or deuterium atoms (from LAD and its by-products)
may be abstracted. Therefore, both the nondeuterated
and the deuterated alkane could be formed (eq 8). On
Effect of Con cen tr ation on th e Reaction of 1-Iodo-
2,2-d im eth ylh exa n e (20) w ith LAD. The concentra-
tion of LAD in a reaction should affect the ratio of
nondeuterated to deuterated product. A variation in
concentration of LAD in solution automatically leads to
a variation in the ratio of deuterium atom donors (LAD
and its byproducts) to hydrogen atom donors (e.g. the
solvent). Therefore, as the concentration of LAD in
solution decreases, radical intermediates formed in the
reaction are less likely to encounter a deuterium atom
donor, and thus, the ratio of deuterated product to
nondeuterated product should decrease. Table 2 shows
the results of variation of concentration of alkyl iodide
20 on the deuterium content in the resulting hydrocar-
bon. When the initial concentration of 20 was 0.27 M,
the resulting 2,2-dimethylhexane (22) was found to
contain 67% D (experiment 1). Therefore, 33% of the
alkane 22 resulted from the trapping of radical 21 by
hydrogen atoms from THF. At the lower concentration
of 0.070 M, more efficient hydrogen atom abstraction
from the solvent was observed in that the resulting
alkane 22 contained only 28% D. Further dilution of the
reaction to 0.035 M resulted in even more hydrogen atom
abstraction from the solvent (as expected), yielding only
21% D in 22. Therefore, the predicted trend of lower
deuterium incorporation in the product at lower concen-
trations of alkyl iodide 20 was observed. These results,
of course, are consistent with the proposal that this
reaction is proceeding via the formation of a radical
intermediate.
the other hand, for an SN2 reaction, only product 22d 1 is
expected. Product 22 can be formed only through a
radical intermediate and thus provides evidence for a
SET process. Therefore, nondeuterated alkane 22 rep-
resents the minimum amount of product formed through
a radical pathway.
As will be discussed, conclusive evidence for a SET
pathway was obtained for iodide 20, even in the absence
of the halogen atom radical chain process depicted in
Scheme 1 (steps d, f, and g). The formation of nondeu-
terated 22 in yields over 90% in the reaction of 20 with
LiAlD4 provides incontrovertible evidence that the pre-
dominant reaction involves the formation of a radical
intermediate, and therefore, the product is not formed
by an SN2 process.
Resu lts a n d Discu ssion
Effect of Stoich iom etr y on th e Rea ction of LAD
w ith 1-Iod o-2,2-d im eth ylh exa n e (20). If an SET
mechanism were in effect in the reaction of alkyl iodide
20 with LAD, then a variation in the stoichiometry of
20:LAD should have an effect on the ratio of nondeuter-
ated alkane 22 to deuterated alkane 22d 1 (eq 8). Radical
21 in solution can abstract a deuterium atom from LAD
or its byproducts (e.g. AlD3), or it can abstract a hydrogen
atom from the solvent. Higher stoichiometries of LAD
increase the probability that 21 will encounter a deute-
rium atom over a hydrogen atom. Likewise, at lower
LAD stoichiometries, radical 21 has a higher probability
of encountering a THF molecule, and thus more protio
hydrocarbon should be formed.
Effect of Va r ia tion of Solven t a n d Nu cleop h ile on
th e Rea ction of 1-Iod o-2,2-d im eth ylh exa n e (20). The
experiments in Table 3 were carried out in order to show
that SET was not exclusive to THF as the solvent and to
provide evidence that the solvent was indeed the source
of hydrogen atoms which were abstracted by radicals in
solution. Experiments 1 and 2 were reported in earlier
The expected trend described above was indeed ob-
served for the reaction of 20 with LAD (Table 1). The
results of experiment 1 for a 20-fold excess of LAD show
that, after 7 h, 82% 2,2-dimethylhexane (22) containing

3546 J . Org. Chem., Vol. 62, No. 11, 1997
Ashby and Welder
Ta ble 3. Va r ia tion of Solven t in th e Rea ction s of 1-Iod o-2,2-d im eth ylh exa n e (20) w ith LAD a n d LAHa
products
material
exp
solvent
time
nuc
20:nuc
% 20
% 22 (% D)
% dimers
trace
balance
1
2
3
4
5
6
7
THF
THF
Et2O
Et2O
THF-d8
THF-d8
THF-d8
8 h
20 h
2 d
4 d
2 h
2 d
3 d
LAD
LAD
LAD
LAD
LAH
LAH
LAD
1:5
1:1
1:5
1:1
1:5
1:1
1:5
0
28
8
26
0
100 (28)
67 (6)
84 (33)
63 (8)
96 (1)
88 (15)
91 (89)
100
95
trace
trace
trace
9
92b
89b
96
0
0
100c
91
trace
a
All reactions were carried out in an Ar atmosphere glovebox in the absence of light at room temperature and at a concentration of
b
0.070 M with respect to 20. Reactions were carried out in used Pyrex flasks and Teflon-coated stir bars were employed. A trace quantity
of 5,5-dimethyl-1-hexene was observed. c 5,5-Dimethyl-1-hexene (3%) was also obtained.
Ta ble 4. Effect of Ligh t on th e Rea ction of
tables and are repeated in this table for convenience of
comparison.
1-Iod o-2,2-d im eth ylh exa n e (20) w ith LADa
products
Experiment 1, carried out in THF, may be directly
compared to experiment 3, which was carried out in
diethyl ether at the same concentration and stoichiom-
etry. Even though the LAD reaction was carried out in
two different solvents, the extent of hydrogen atom
abstraction, as noted by the deuterium incorporation in
alkane 22, was comparable with 28% D (72% hydrogen
atom abstraction) at completion in the THF reaction
(experiment 1) and 33% D (67% hydrogen atom abstrac-
tion) near completion in the diethyl ether reaction
(experiment 3). A comparison of the data in THF and
diethyl ether (experiments 2 and 4, respectively) at the
lower stoichiometry of 1:1 also shows the similarity of
the two solvents.
In order to support the solvent as the source of
hydrogen atoms, experiments were carried out in deu-
terated THF (experiments 5-7). When a 5-fold excess
of LAH was employed in a reaction with THF-d8 as
solvent, only 1% D was found in the product hydrocarbon
(experiment 5). This result is not surprising due to the
kinetic isotope effect. Not only was the strength of the
aluminum-deuterium bond of LAD weakened to an
aluminum-hydrogen bond, but also, the carbon-hydro-
gen bond of THF was strengthened to a carbon-deuterium
bond in THF-d8, making deuterium atom abstraction
from the solvent a higher energy process. However, when
the 5-fold excess of LAH employed in experiment 5 was
reduced to 1 equiv of LAH (experiment 6), the product
was found to contain 15% D. The relatively high deu-
terium incorporation (15% D) in the alkane 22 supports
deuterium atom abstraction from the solvent by a radical
intermediate which in turn indicates that THF acts as a
hydrogen atom donor in the experiments that use LAD
as the nucleophile. The high yield of dimers in experi-
ment 6 supports the formation of radical intermediates
which combined to form dimers rather than abstracting
a deuterium atom from the THF-d8 or a hydrogen atom
from LAH or its byproducts. Experiment 7 was included
to determine the deuterium content in 22 when both
solvent and nucleophile are labeled with deuterium.
Alkane 22 was found to contain 89% D. When one takes
into account isotopic purities (THF-d8 contains 99.5% D
and LiAlD4 contains 98.5% D) and the kinetic isotope
effect (kH/kD ≈7), the observed deuterium content of 89%
can be rationalized.
time,
h
% 22
%
material
exp lighting
% 20 (% D) dimers otherb balance
1
dark
1
5
8
24
0.5
1
1.5
2
88
60
33
5
4 (48)
33 (42) trace
58 (39) trace
84 (34)
37 (13) 10
92
93
91
92
95
99
92
94
3
2
mercury
lamp
46
13
1
2
3
4
4
67
69
16
18
0
71 (12) 19
a
Reactions were carried out at a concentration of 0.070 M with
respect to 20 and a RI:LAD ratio of 1:5 in an Ar atmosphere
glovebox. Experiment 1 was carried out in a used Pyrex flask at
room temperature. Experiment 2 was carried out in a Pyrex test
tube. b A total of six “other” compounds of unknown structure were
observed in the photochemical experiment, each in <1% yield all
having a m/ e of 111.
Alkyl iodides 1 and 20 vary only in the fact that 1
possesses a remote double bond while 20 does not. Since
significant evidence of SET has been reported above for
the reaction of the saturated alkyl iodide 20 with LAD,
it is again suggested that the unsaturated alkyl iodide 1
also reacts by SET since the presence of a remote double
bond in 1 is hardly cause to suggest that the reaction
mechanism changes from SET to SN2.
Effect of Ligh t on th e Rea ction of 1-Iod o-2,2-
d im eth ylh exa n e (20) w ith LAD. Photochemical exci-
tation is known to enhance the formation of radical
species in solution. In order to observe the effect of
photochemical excitation on the reaction of 20 with LAD,
an experiment was carried out in the presence of a
mercury lamp. It is important to note that, in the
absence of LAD, no reaction occurred when iodide 20 was
exposed to the lamp.
As can be seen in Table 4, the rate of the reaction of
1-iodo-2,2-dimethylhexane (20) with LAD was signifi-
cantly enhanced in the presence of the mercury lamp.
The reaction carried out in the dark (experiment 1) was
not yet complete after 24 h, while the reaction carried
out in the presence of the mercury lamp (experiment 2)
was complete within 2 h.
The deuterium incorporation in alkane 22 was much
lower for the photochemical experiment (experiment 2),
indicating more hydrogen atom abstraction in the pres-
ence of the photochemical lamp than in the corresponding
experiment carried out in the dark. Not only was the
deuterium content of the product low, but it was also
consistent throughout the reaction. Unlike any reaction
presented thus far, the deuterium content in the initial
aliquot (37% alkane containing 13% D) was within
experimental error of the deuterium content at comple-
It can be concluded from the solvent and isotope studies
that SET from LAD to 1-iodo-2,2-dimethylhexane (20)
was not exclusive to reactions carried out in THF and
that the solvent acts as a hydrogen atom source (or a
deuterium atom source for reactions in THF-d8) toward
radical species formed during the course of the reaction.

Reactions of 1-Halo-2,2-dimethylhexanes with LiAlD4
J . Org. Chem., Vol. 62, No. 11, 1997 3547
Ta ble 5. Effect of th e Byp r od u cts Alu m in u m Deu ter id e
a n d Lith iu m Iod id e in th e Rea ction of
1-Iod o-2,2-d im eth ylh exa n e (20)a
Ta ble 6. Rea ction of 1-Iod o-2,2-d im eth ylh exa n e (20)
w ith LAD in Used P yr ex F la sk sa
products
products
exp
1
time, h
%20
%22 (%D )
% dimers
MBb
exp
1
nuc
time, h
% 20
% 22 (% D) % dimers MBb
1.25
5
7
1.25
5
48
1.25
5
48
1
5
1
5
24
1
5
24
1
5
24
48
48
68
12
0
80
64
0
75
44
0
97
55
89
65
trace
90
55
9
26
72
94 (29)
6
21
100 (57)
9
35
100 (35)
4
35 (29)
4
29 (56)
86 (44)
5 (50)
36 (41)
77 (39)
4 (48)
33 (42)
84 (34)
100 (66)
90 (76)
94
84
94
86
85
100
84
79
100
101
90
93
94
90
95
91
88
92
93
LAD
1
5
8
24
48
70
24
66
1
88
60
33
5
38
33
18
trace
88
31
0
4 (48)
33 (42)
58 (39)
84 (34)
60 (5)
65
80
95
10 (49)
58 (38)
76 (33)
92
93
91
92
98
98
98c
98c
98
92
83
trace
trace
3
trace
trace
trace
trace
2
3
2
AlD3
AlH3
LAD
3
4d
4
5
5
24
3
7
a
All reactions were carried out at a concentration of 0.070 M
with respect to 20 at a RI:Nuc ratio of 1:5 in an Ar atmosphere
4
6
7
glovebox in the absence of light at room temperature in used Pyrex
trace
2
b
flasks using Teflon-coated stir bars. MB is material balance. c A
d
trace quantity of 5,5-dimethyl-1-hexene was observed. One
equivalent of lithium iodide was initially added to the reaction.
88
60
5
0
0
trace
3
92
100
90
8
9
tion of the reaction (71% alkane containing 12% D).
Dimer formation was also enhanced in the photochemical
experiment. In reactions carried out in the dark, dimers
were typically observed in trace quantities. However, as
the data from experiment 2 show, a significant amount
of alkyl iodide 20 was converted to dimeric compounds
under photochemical excitation with 19% dimer forma-
tion by the completion of the reaction. New products
were formed in the photochemical experiment that were
not observed in any other reactions studied. Six com-
pounds, each formed in trace quantity, were observed by
GC/MS, and all were found to have a m/ e of 111. Each
of the compounds was formed in low yield, and the
compounds had very similar GC retention times. Since
clean separation was not possible, probably due to the
similarity of the structures, no attempt was made at
further characterization.
Decreased reaction time, decreased deuterium incor-
poration in the 2,2-dimethylhexane, and increased dimer
formation can all be explained by the enhancement of a
radical process induced by photochemical excitement.24
Effect of Alu m in u m Deu ter id e on th e Rea ction
of 1-Iod o-2,2-d im eth ylh exa n e (20) w ith LAD. As
early as 1983 we reported evidence that the AlH3
produced in situ during the LAH reduction of alkyl
halides can also be a one-electron transfer agent.5e,g
Since aluminum deuteride, AlD3, is an expected byprod-
uct of the reaction of 20 with LAD, we considered it to
be of value to independently study the effect AlD3 has
on 20.
a
All reactions were carried out at a concentration of 0.070 M
with respect to 20 at a RI:LAD ratio of 1:5 in an Ar atmosphere
glovebox in the absence of light at room temperature using Teflon-
b
coated stir bars. MB is material balance.
Effect of Lith iu m Iod id e on th e Rea ction of
1-Iod o-2,2-d im eth ylh exa n e (20) w ith LAD. Lithium
iodide is also an expected byproduct of the reaction of
LAD with alkyl iodides. In order to determine whether
or not the lithium iodide produced in situ had an effect
on the reduction of iodide 20 by LAD, a reaction was
carried out in which 1 equiv of LiI was initially added to
the reaction mixture (Table 5, experiment 4).
As can be seen in Table 5, the reaction carried out in
the absence of added LiI (experiment 1) and the reaction
carried out in presence of added LiI (experiment 4) were
very similar except that the rate was somewhat faster
in the presence of added LiI. The deuterium content of
alkane 22 over the course of the experiments had a high
correlation, with the initial deuterium content of 48-49%
D and the final deuterium content of 33-34% D. There-
fore, it appears that the LiI byproduct has little, if any,
effect on the mechanistic course of the reaction.
Effect of Rea ctor Su r fa ces on th e Rea ction of
1-Iod o-2,2-d im eth ylh exa n e (20) w ith LAD. As can
be seen in Table 6, nine experiments carried out in
different used Pyrex flasks varied in reaction rate and/
or deuterium content of the product. After 5 h, the
remaining starting material 20 varied from 12% (experi-
ment 1) to 65% (experiment 5). More importantly, the
deuterium content in 2,2-dimethylhexane (22) varied
from 29% D at completion (experiment 1) to 76% D at
completion (experiment 9). We have been troubled by
our inability, in the present work, to obtain the kind of
consistency of data in duplicate experiments that we have
observed in the past. The only variable in this reaction
that we had not studied previously is the nature of the
reactor surface. Therefore, a study was initiated to
determine the effect of different reactor surfaces on the
reaction of 20 with LAD in order to find a vessel in which
consistent data would be available. The results of this
study are shown in Table 7.
As shown in Table 5, AlD3 reacts with 20 at a much
slower rate than does LAD under the same conditions
(experiments 1 and 2). Both the reaction of 20 with LAD
and with AlD3 gave high yields of 2,2-dimethylhexane
(22). However, the deuterium incorporation in alkane
22 was significantly lower when AlD3 was used as the
nucleophile. Once again, lower deuterium incorporation
in the product indicates more hydrogen atom abstraction
of a radical intermediate from the solvent. The results
of experiment 3 show that the reaction of 20 with AlD3
is much slower than the corresponding reaction with
AlH3.
(24) Radical species formed by photochemical excitation are not
expected to be identical to the radical species formed during thermal
reactions. See: Kropp, P. J . Acc. Chem. Res. 1984, 17, 131.
Teflon has been used as a reactor surface in reactions
in which glass has been suspected of catalyzing the

3548 J . Org. Chem., Vol. 62, No. 11, 1997
Ashby and Welder
Ta ble 7. Rea ction of 1-Iod o-2,2-d im eth ylh exa n e (20)
w ith LAD in Va r iou s Rea ction Vesselsa
Ta ble 8. Rea ction s of 1-Ha lo- a n d
1-Tosyl-2,2-d im eth ylh exa n es w ith LADa
products
products
reaction
vessel
time,
h
% 22
(% D)
exp
X
time
% RX
% 22 (% D)
% dimers
3
MBb
exp
1
% 20
% dimers MBb
92
1
2
3
I
24 h
12 d
4 d
5
53
59
84 (34)
47 (84)
27 (95)
92
100
95c
used Pyrex
1
5
8
24
1
5
8
24
1
2
5
8
24
1
5
8
24
1
5
88
60
33
5
90
74
61
9
61
37
16
8
4 (48)
33 (42)
58 (39)
84 (34)
3 (61)
18 (54)
33 (52)
77 (42)
17 (26)
24 (26)
34 (27)
41 (27)
47 (27)
4 (37)
32 (34)
44 (32)
83 (29)
5 (40)
26 (39)
39 (39)
71 (32)
78 (35)
4 (40)
Br
OTs
trace
trace
3
93
91
92
93
92
94
87
99
92
95
97
97
97
85
77
91
99
90
85
80
83
100
93
86
90
a
All reactions were carried out at a concentration of 0.070 M
with respect to RX and at a RX:LAD ratio of 1:5 in an Ar
atmosphere glovebox in the absence of light at room temperature
2
3
Teflon
b
in used Pyrex flasks. Teflon-coated stir bars were employed. MB
trace
1
is material balance. c 2,2-Dimethyl-1-hexanol (9%) was also ob-
served.
stainless steel
21
31
45
48
50
35% D by completion. A comparison of experiments 4
and 5 reveals little difference between the reactions
carried out in treated Pyrex with those carried out in new
Pyrex. The similarity of the reactions carried out in these
two vessels adds credibility to the conclusion that the
treated Pyrex and new Pyrex vessels did not interfere
with the product distribution and are appropriate for use
in mechanistic studies involving LAH/LAD.
Reactions were also carried out in quartz (experiment
6). Encouragingly, the data was not only consistent
between duplicate experiments, but it also closely re-
sembled the data obtained in treated Pyrex and new
Pyrex vessels. Importantly, the integrity of the quartz
vessel did not appear to change with repeated use.
The surface of the reaction vessel was not the only
surface that played a significant role in this reaction.
When unfiltered, heterogeneous solutions of LAH or LAD
were employed, inconsistent results were noted. Forma-
tion of dimers originating from the alkyl radical 21
ranged from <0.5-25% when undissolved solids were
present in the flask. On the other hand, when filtered,
homogeneous solutions were employed in the reaction,
dimers were consistently formed in low yield (Table 7).
Only filtered, homogeneous solutions of LAH and LAD
were used throughout the present studies.
The following generalizations, derived by combining
the above information on the involvement of reactor
surfaces, should be applied to future mechanistic stud-
ies: (1) Teflon-coated stir bars should be avoided because
Teflon was found to react with LAH, (2) reactions may
be carried out in dichlorodimethylsilane-treated Pyrex,
new Pyrex, or quartz vessels, and (3) homogeneous
solutions of LAH or LAD should be employed since the
surface of undissolved nucleophile was found to affect
product distribution.
Since the effect of the reactor surface was determined
late in this study, many of the reactions presented herein
were carried out in used Pyrex flasks. However, since
the data obtained in treated Pyrex, new Pyrex, or quartz
vessels (with no Teflon stir bar present) always gave more
evidence of radical formation than did the corresponding
reaction in used Pyrex vessels, it was concluded that the
hydrogen atom abstraction observed in used Pyrex ves-
sels represents the minimum trapping of radicals that
can occur in the reactions of LAD with the alkyl halides
studied.
0
4
5
treated Pyrex
new Pyrex
93
51
29
1
94
64
45
7
2
4
7
trace
8
1
2
5
24
48
1
5
8
0
6
quartz
96
62
33
0
31 (35)
53 (34)
88 (30)
trace
trace
2
24
a
All reactions were carried out at a concentration of 0.070 M
with respect to 20 at a RI:LAD ratio of 1:5 in an Ar atmosphere
glovebox in the absence of light at room temperature in the absence
of Teflon-coated stir bars. MB is material balance.
b
reaction. To test the integrity of Teflon as a reaction
vessel in the presence of LAH, a homogeneous solution
of LAH in THF (0.32 M) was allowed to stir in a Teflon
flask for 2 days. Ion chromatographic analysis of the
quenched basic solution revealed the presence of fluoride
ions. Therefore, LAH reacted with Teflon, and even
though data were included using a Teflon reaction vessel
(experiment 2), it was concluded that Teflon is not a
suitable vessel for reactions involving LAH. Likewise,
the use of Teflon-coated stirring bars should be avoided
when studying the mechanism of reactions involving
LAH.
A stainless steel vessel was employed so that the effect
of a metal surface could be observed (experiment 3). In
this experiment, the formation of several dimers was
observed very early in the reaction and, at the completion
of the reaction, dimer formation was found to be much
more extensive than had been previously observed. The
stainless steel surface appears to catalyze the reduction
of alkyl iodide 20, and thus, it is inappropriate as a
reaction vessel.
Reactions in used Pyrex flasks that had been treated
with dichlorodimethylsilane to deactivate the surface
were monitored as well (experiment 4). Encouragingly,
consistent results were obtained with the deuterium
content of 22 found to be 37% D in the initial aliquot (4%
reaction) and 29% D near completion. (Decreasing
deuterium content over the course of the reaction was
also observed in used Pyrex and Teflon flasks; however,
the amount of deuterium in product 22 was not consistent
between duplicate experiments.)
Effect of th e Lea vin g Gr ou p on th e Rea ction of
1-Ha lo-2,2-d im eth ylh exa n es w ith LAD. In order to
broaden the scope of this reaction, a study was carried
out on the effect of changing the leaving group from
iodide to bromide to tosylate (Table 8). Since the alkyl
bromide has a less favorable reduction potential than the
alkyl iodide, electron transfer from LAD to the alkyl
Consistent results were also obtained for reactions that
were carried out in previously unused (new) Pyrex flasks
(experiment 5). The deuterium content of alkane 22 after
5% reaction was found to be 40% D which decreased to

Reactions of 1-Halo-2,2-dimethylhexanes with LiAlD4
J . Org. Chem., Vol. 62, No. 11, 1997 3549
bromide would be less favored than for the corresponding
alkyl iodide. The tosylate of 2,2-dimethyl-1-hexanol (23)
has such an unfavorable reduction potential that it is not
expected to act as an electron acceptor, thus lessening
the possibility of the SET pathway. The tosylate should,
however, behave similarly to the iodide if an SN2 reaction
were involved since the two groups display similar
nucleofugacity. Therefore, the reaction of tosylate 23
should represent the maximum extent of deuterium
incorporation for reactions of the alkyl halide, if an SN2
pathway is in effect.
1-Bromo-2,2-dimethylhexane (24) was allowed to react
with LAD (experiment 2). Even though the reaction with
LAD was much slower than the reaction of the corre-
sponding iodide 20, once again, significant evidence for
an SET pathway was observed in that 16% of the
hydrocarbon formed in the reduction did not contain
deuterium and hence indicates the presence of a radical
precursor that abstracted a hydrogen atom from the
solvent.
As shown in experiment 3, after 4 days 59% of tosylate
23 still remained unreacted while iodide 20 was almost
completely consumed in 1 day (experiment 1). Some of
the tosylate 23 (9%) was converted to the alcohol, 2,2-
dimethyl-1-hexanol, due to the cleavage of the tosylate
by LAD. Obviously, the SN2 displacement of the tosyl
group was severely impeded by the neighboring geminal
dimethyl groups. Also, note that the 2,2-dimethylhexane
(22) that was formed had a very high deuterium content
(95% D), suggesting that little, if any, of the hydrocarbon
was derived from an SET pathway. If 20 reacts with
LiAlH4 via an SN2 pathway, then the results of the
reaction of 20 and 23 should be similar. The results are
very different.
P ossibility of Oth er Ra d ica l Ch a in P r ocesses. We
mentioned in our 1993 communication† that a radical
chain process involving THF is possible in the reaction
of 20 with LAD (eqs 9 and 10). Although we cannot
rigorously eliminate this possibility, we have not been
able to acquire evidence for its existence. All attempts
to isolate the R-iodo THF or its dehydrohalogenation
product, 1,2-dihydrofuran, required of this process, were
not successful. In addition, we found that the reduction
of 7 with LAD produced 3 containing 98% D (eq 11). If
the reaction represented by eq 11 proceeded according
to the radical chain process (eq 12), then product 3 should
have been formed with a much lower degree of deuterium
incorporation as observed for all the other cases when a
radical chain process was observed. Furthermore, radical
25 should be considerably more stable than radical 21,
due to the resonance stablization of 25 and, hence, less
likely than 21 to participate in a halogen atom radical
chain process.
observed is that the rate of formation of the product slows
down dramatically at approximatley 20% reaction. Prod-
uct does continue to be formed at a slower rate beyond
the 20% mark due to the reaction of 20 with the
byproduct, AlD3. On the basis of all of the above
observations, we conclude that a radical chain process
involving THF is not significant.
The question as to whether the reaction of 1 and/or 20
with LAH is initiated by SET, as we have held, or
impurities, as Newcomb has held, rests on the results of
a number of experiments. We have carried out reactions
of 1 with doubly recrystallized LAH, doubly distilled 1
in THF doubly distilled over NaAlH4, and found no
difference in the rate or product ratios compared to the
reactions carried out under normal conditions where the
LAH used is from a filtered solution of LAH in THF, and
both the alkyl halide and THF are singly distilled.5e Also,
addition of trace transition metals as well as experiments
involving rigorous exclusion of oxygen or addition of large
amounts of oxygen had little effect on the reactions. In
the absence of any evidence that the reaction is initiated
by impurities rather than initiated by the electron rich
-
AlH4 ion, we conclude that the reaction of sterically
hindered alkyl iodides is initiated by electron transfer
from the electron rich nucleophile AlH4- to the favorable
electron acceptor alkyl iodide followed by a hydrogen
atom radical chain process and, less likely, by a THF or
halogen atom radical chain process, to produce the
reduction product predominantly, if not entirely, via a
radical intermediate which abstracts a hydrogen atom
from the solvent. (Also see arguments presented in the
introduction relating to eqs 3 and 4.)
Previously we had reported5e that the reaction of
6-iodo-1-heptene (9) with LiAlH4 in a 1:0.1 ratio shows
evidence of a halogen atom radical chain process but not
a radical chain process involving THF (eq 2). After 70
h, the starting iodide 9 was recovered in 19% yield and
the cyclized iodide 10 was formed in 70% yield, indicating
a halogen atom radical chain process. However, the
combined hydrocarbon yield was only 11%, indicating
that a radical chain process involving the THF radical is
not involved.
P r op osed Mech a n ism of th e Rea ction of 1-Iod o-
2,2-d im eth ylh exa n e (20) w ith LAD. With all the
combined information from the preceding experiments,
the following mechanism is proposed for the reduction
of 20 by LAD (Scheme 2). It is suggested that alkyl iodide
20 accepts an electron from LAD (step a) to form the
corresponding radical anion 26 which rapidly dissociates
(step b) to form radical 21. Radical 21 abstracts a
If the THF radical (25) were involved in a radical chain
process (eqs 9 and 10) in the reaction of 20 with LAD in
a 1:0.2 ratio, then the rate of product formation should
not vary greatly beyond the 20% mark. What we
-
deuterium atom from AlD4 (step c) to form the deuter-

3550 J . Org. Chem., Vol. 62, No. 11, 1997
Ashby and Welder
Sch em e 2
a remote double bond in the terminal position of the
molecule and since the results of reactions of 20 and 1
with LAH or LAD are similar, we continue to maintain
the position that 1 reacts with LAH or LAD by an SET
process. The nature of the vessel surface was found to
have some effect on the reaction; however the results
involving the use of old Pyrex, used in these and previous
studies, compared to quartz or new Pyrex, as demon-
strated in the present studies, was not significantly
different, and hence the SET nature of the reaction
cannot be attributed to the reaction vessel surface.
Exp er im en ta l Section
Ma ter ia ls. THF and diethyl ether were distilled from
sodium benzophenone ketyl and then further distilled from
NaAlH4 prior to use in mechanistic studies. THF-d8 (Aldrich)
was distilled from NaAlH4 prior to use. Diisopropylamine and
ethyl isobutyrate were purchased from Aldrich and distilled
from CaH2. LiAlH4 and LiAlD4 were purchased from Aldrich,
allowed to stir in the desired solvent overnight, and then
filtered in an argon atmosphere glovebox to obtain homoge-
neous solutions which were titrated before use (see below).
Bromobutane and methyllithium (1.4 M in diethyl ether) were
purchased from Aldrich and used as received. EDTA (Baker)
was dried at 80-120 °C prior to use. An authentic sample of
2,2-dimethylhexane (Wiley Organics) was used in order to
verify its formation in reactions involved in the mechanistic
studies. All synthetic reactions were carried out under a
nitrogen atmosphere. All mechanistic reactions were carried
out in an argon atmosphere glovebox. Glassware was allowed
to dry for at least 2 h at 200 °C and cooled in a desiccator
prior to use. Solutions of aluminum hydride were prepared
according to literature procedure.26
In str u m en ta tion . GLC analyses were performed on a
Varian 3700 gas chromatograph with a J & W DB-5 fused
silica column (30 m, 0.25 mm i.d., 25 µm film thickness)
equipped with a FID. Mass spectra were obtained from a VG
70-SE mass spectrometer equipped with a double sector
magnetic analyzer. NMR spectra were obtained from a Varian
Gemini 300 MHz instrument with chemical shifts reported
relative to tetramethylsilane (δ 0.00). Photochemical experi-
ments were carried out in Pyrex glassware using a 450 W
Hanovia immersion lamp.
Gen er a l. Yields were determined by gas chromatography
relative to an internal standard. The following equation was
used to calculate the deuterium incorporation of a compound
from mass spectral data collected in the selected ion resonance
mode:
ated alkane 22d 1. The aluminum deuteride radical anion
reacts with the starting material 20 (step d) to form the
corresponding radical anion 26 and AlD3 in a hydrogen
atom radical chain process. Radical 21 also reacts with
a suitable hydrogen atom donor, such as THF (step e),
to form the nondeuterated alkane 22. A third fate of
primary radical 21 is rearrangement to the secondary
radical 27 by a 1,5 hydrogen atom shift.25 It is proposed
that LAD or one of its Al-D byproducts then abstracts a
hydrogen atom from radical 27 as shown (step g) to form
5,5-dimethyl-1-hexene (2) in trace yield. Any AlD3 radi-
cal anion formed in step g can then react as shown in
step d. The AlD3 formed during the reaction can also
react by SET with alkyl iodide 20 as shown in step h.
Radical anion 26 dissociates to radical 21 and iodide ion
(step i). As shown in step j, AlD3 can act as a deuterium
atom source to radical 21 forming 22d 1 and the AlD2I
radical anion. Then AlD2I radical anion can function as
a single-electron donor to alkyl iodide 20 (step k) to form
AlD2I and radical anion 26, which can dissociate as
described above.
% D ) 100{1 - [(A + B + C)/A][a/(a + b + c)]}
where A ) M, B ) M + 1, and C ) M + 2 for a sample of
natural abundance isotopes and a ) M, b ) M + 1, and c ) M
+ 2 for a sample with suspected deuterium incorporation.
(Only the molecular ion was used to calculate % D.) In order
for this equation to be valid, the M - 1 peak for a nondeuter-
ated sample must be negligible (<10% of M). A, B, and C are
ideally obtained from a sample of natural abundance isotopes,
but Benyon’s Table27 may be used for the values of A, B, and
C if necessary.
Titr a tion of LAH/LAD Solu tion s. A 0.10 mL aliquot of
a homogeneous solution of LAH was quenched in ∼30 mL of
deionized water. EDTA (10.00 mL, 0.02 M) was added to the
aluminum sample so that at least a 10% excess of EDTA was
employed. The solution was boiled for 2-3 min until the
cloudiness disappeared and then cooled to room temperature
or below. Once cool, a buffer solution (10 mL of stock solution
made from 77 g of ammonium acetate and 230 mL of glacial
Con clu sion s
The reaction of 1-iodo-2,2-dimethylhexane (20), the
saturated analog of 6-iodo-5,5-dimethyl-1-hexene (1),
with LiAlH4 and LiAlD4 has been studied in detail. The
results clearly show that the reaction of 20 with LAH or
LAD cannot be described by an SN2 process and indeed
shows all the characteristics expected for an SET process.
Since the difference between 20 and 1 is the presence of
(26) (a) Brown, H. C.; Yoon, N. M. J . Am. Chem. Soc. 1966, 88, 1464.
(b) Brown, H. C.; Krishnamurthy, S. J . Org. Chem. 1980, 45, 849.
(27) Silverstein, R. M.; Bassler, G. C.; Morrill, T. C. Spectometric
Identification of Organic Compounds, 4th ed.; J ohn Wiley & Sons: New
York, 1981.
(25) For a review on radical rearrangements, see: Freidlina, R. Kh.;
Terent’ev, A. B. Acc. Chem. Res. 1977, 10, 9.

Reactions of 1-Halo-2,2-dimethylhexanes with LiAlD4
J . Org. Chem., Vol. 62, No. 11, 1997 3551
acetic acid diluted to 1.0 L and adjusted to pH ) 4) was added.
Ethanol was then added to equal the amount of solution
already present. Diphenylthiocarbazone (dithizone) indicator,
1-2 mL (60 mg of dithizone in 250 mL of absolute ethanol
stored in a dark bottle), was also added. The excess EDTA
was then titrated with zinc sulfate (0.02 M). The endpoint
was noted by a sharp color change from blue to pink. All
calculations were based on a 1:1 metal:EDTA stoichiometry.
Red u ction of Alk yl Ha lid es. To a flask, in an argon
atmosphere glovebox at room temperature, were added the
desired alkyl halide (0.1 mmol), decane (10 µL, an internal
standard for GC), and the desired solvent. To this solution
was added a homogeneous solution of LAH or LAD in the same
solvent. The concentration of the reaction was calculated on
the basis of the alkyl halide (usually 0.070 M). The reaction
vessel was then covered in aluminum foil to shield the reaction
from light. At the desired time, an aliquot (0.1-0.2 mL) was
removed and quenched in water (0.2 mL). The organic
compounds were extracted with diethyl ether (0.2 mL), and
the organic layer was analyzed by GC and/or GC/MS.
P h otoch em ica l Red u ction of 1-Iod o-2,2-d im eth ylh ex-
a n e (20). To a Pyrex test tube in an argon atmosphere
glovebox were added 1-iodo-2,2-dimethylhexane (0.040 g, 0.17
mmol), decane (15 µL, 0.077 mmol), and THF (1.69 mL). The
test tube was capped with a rubber septum, removed from the
glovebox, and placed in the sample holder for the Hanovia
mercury lamp. An oil bubbler was added to the test tube
through the rubber septum. A homogeneous solution of LAD
in THF (1.2 M, 0.69 mL, 0.83 mmole) was added to the test
tube by syringe. The lamp was turned on. At the desired time
the lamp was turned off and an aliquot (0.2 mL) was removed
from the test tube by syringe and quenched in water (0.2 mL).
The organic compounds were extracted with diethyl ether (0.2
mL), and the organic layer was analyzed by GC and GC/MS.
Note: the mercury lamp required a 5 min cooling time before
it would turn on again. Therefore, for each 30 min of reaction,
the lamp was on for only 25 min.
yield). A solution of the resulting ethyl 2,2-dimethylhexanoate
in 40 mL of ether was added over 40 min to a suspension of
LiAlH4 (4.54 g, 120 mmol) in 100 mL of diethyl ether at 0 °C.
After 2 h, the reaction was quenched with water (50 mL), the
gelatinous residue was washed with diethyl ether (4 × 50 mL),
and the combined washes were dried over MgSO4. The product
was obtained by vacuum distillation (109-113 °C/87 mmHg)
in 73% overall yield (6.94 g: 1H NMR (CDCl3) δ 3.32 (s, 2H),
1.38-1.18 (m, 6H), 0.90 (t, 3H, J ) 6.7 Hz), 0.86 (s, 6H).
1-Br om o-2,2-d im eth ylh exa n e (24). The title compound
was prepared from 2,2-dimethyl-1-hexanol by a method analo-
gous to the previously described synthesis of 1-bromo-2,2-
dimethyl-5-hexene5e in 36% yield upon vacuum distillation
(69-70 °C/14 mmHg: 1H NMR (CDCl3) δ 3.29 (s, 2H), 1.38-
1.14 (m, 6H), 0.99 (s, 6H), 0.91 (t, 3H, J ) 7.0 Hz); MS m/ e
(relative intensity) (C8H17Br) 179 (0.2), 177 (0.2), 137 (13), 135
(13), 99 (91), 57 (100).
1-Iod o-2,2-d im eth ylh exa n e (20). The title compound was
prepared from 2,2-dimethyl-1-hexanol by a method analogous
to the previously described synthesis of 2,2-dimethyl-1-iodo-
5-hexene5e in 65% yield upon vacuum distillation (87-90 °C/
17 mmHg): 1H NMR (CDCl3) δ 3.16 (s, 2H), 1.38-1.12 (m,
6H), 1.02 (s, 6H), 0.91 (t, 3H, J ) 6.9 Hz); MS m/ e (relative
intensity) (C8H17I) 240 (3), 183 (39), 113 (95), 71 (100), 57 (99),
55 (64), 43 (82), 41 (72).
Tosyla te of 2,2-Dim eth yl-1-h exa n ol (23). To a solution
of tosyl chloride (4.01 g, 21.0 mmol) in 50 mL of pyridine cooled
in an ice bath was added 2,2-dimethyl-1-hexanol (1.025 g, 7.87
mmol). The mixture was allowed to stir for 24 h at 4 °C.
Water (50 mL), followed by 10% aqueous HCl (50 mL) and
diethyl ether (60 mL), was added to the flask. The water layer
was extracted with ether (2 × 50 mL). The combined organic
layers were washed with 60 mL of 10% HCl then 50 mL of
10% NaHCO3. The organic layer was dried over MgSO4,
filtered, and concentrated by rotary evaporation. The product
was purified by neutral alumina chromatography using pe-
troleum ether as the eluent in 76% yield (1.71 g): 1H NMR
(CDCl3) δ 7.80 (d, 2H, J ) 8.1 Hz), 7.35 (d, 2H, J ) 8.1 Hz),
3.68 (s, 2H), 2.46 (s, 3H), 1.30-1.00 (m, 6H), 0.85 (s, 6H), 0.84
(t, 3H, J ) 7.1 Hz). MS m/ e (relative intensity) (C15H24O3S)
187 (15), 157 (14), 155 (31), 112 (21), 99 (56), 98 (60), 91 (51),
57 (100).
2,2-Dim eth yl-1-h exa n ol. A solution of lithium diisopro-
pylamide (LDA) in THF was prepared as follows: to a solution
of diisopropylamine (12.3 mL, 8.88 g, 87.8 mmol) in 70 mL of
THF was added a solution of methyllithium (1.3 M in ether,
57 mL, 74 mmol) at -78 °C under N2. After 1 h a solution of
ethyl isobutyrate (9.9 mL, 8.5 g, 73 mmol) in 20 mL of THF
was added over 25 min. The solution was stirred at -78 °C
for an additional 75 min. 1-Bromobutane (7.8 mL, 10. g, 73
mmol) was then added to the flask dropwise over 20 min. The
solution was allowed to warm to room temperature. After 3.5
h the reaction was quenched with water (35 mL) and extracted
with ether (3 × 50 mL). Combined ether layers were dried
over MgSO4, filtered, concentrated, and then distilled (105-
110 °C/66 mmHg) to obtain a colorless liquid (10.3 g, 82%
Ack n ow led gm en t. We are indebted to Sarah J .
Shealy and David E. Bostwick for their mass spectrom-
etry assistance. This work was supported by the Na-
tional Science Foundation (Grant CHE-8914309) and
the Georgia Institute of Technology, School of Chemistry
and Biochemistry.
J O961651+