N,S-Dimethyldithiocarbamyl oxalates as precursors for determining kinetic parameters for oxyacyl radicals

Article abstract of DOI:10.1039/c4cc06132b

N,S-Dimethyldithiocarbamyl oxalates (e.g.6, 10) are novel, readily prepared precursors to alkyloxyacyl radicals 1 that are more suitable for kinetic studies than existing precursors; 10 has allowed the determination of accurate rate data for the cyclization of the butenyloxyacyl radical 5 (k_c = 1.2 × 10⁷ s^-1 at 21 °C). This journal is

Full text of DOI:10.1039/c4cc06132b

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
N,S-Dimethyldithiocarbamyl oxalates as precursors
for determining kinetic parameters for oxyacyl
radicals†
Cite this: Chem. Commun., 2014,
0, 12040
5
Received 6th August 2014,
Accepted 21st August 2014
ab
ab
Sara H. Kyne‡ and Carl H. Schiesser*
DOI: 10.1039/c4cc06132b
www.rsc.org/chemcomm
N,S-Dimethyldithiocarbamyl oxalates (e.g. 6, 10) are novel, readily
prepared precursors to alkyloxyacyl radicals 1 that are more suitable
for kinetic studies than existing precursors; 10 has allowed the deter-
mination of accurate rate data for the cyclization of the butenyloxyacyl
7
À1
radical 5 (k
c
= 1.2 Â 10 s at 21 8C).
Methods for the generation of carbon-centred free radicals
1
abound. Halides, arylselenides, aryltellurides, xanthates, pyr-
2
idinethioneoxycarbonyl (PTOC, Barton) esters and dithio-
3
carbamyl (Kim) esters, while most commonly used, form just
a subset of the chemical toolbox available to the practitioner
when considering the use of radical chemistry in synthesis.
1
Scheme 1
Despite this, there are still some classes of carbon-centred
radical for which no robust precursors have been reported.
4
Oxyacyl radicals (1) are one such class of radical. To date only diselenide contamination can skew the interpretation of the
5
6
phenylselenide (2a), phenyltelluride (2b) and PTOC oxalate kinetic data.
7
(
3) precursors have appeared in the literature (Scheme 1), and
With this in mind, we recently reported the use of high-level
these have their own specific limitations and drawbacks. For computational techniques for the determination of kinetic data for
example, studies using tellurides (2b) and PTOC oxalates (3) are the 5-exo and 6-endo modes of ring closure of a series of substituted
7
,8
4
complicated by photochemical instability, while the former acyl and oxyacyl radicals. While the calculated rate constants (k
(
c
) for
9
2a) can be contaminated with traces of diphenyl diselenide,
the cyclization of the acyl radicals (e.g. 4) were in good agreement
which for the purposes of synthesis can often present no real with available experimental data (Scheme 2), we reported that the
7
À1
disadvantage, however for those of us interested in determining oxyacyl system (5) cyclizes with a rate constant of 8 Â 10 s (80 1C)
accurate rate data, this contaminant provides some significant using G3(MP2)-RAD, well over an order of magnitude faster than
6
À1
7
challenges. Diphenyl diselenide is known to react rapidly with that reported by Newcomb (4 Â 10 s ), based on data provided
5
chain-carrying reagents such as tributyltin hydride to afford by Bachi. We argued that this discrepancy was largely the result
9
benzeneselenol (Scheme 1). This impurity reacts some three of PhSeSePh contamination during the experiments performed
orders of magnitude faster as a hydrogen atom donor than the by Bachi, a situation acknowledged by Newcomb when he
1
0
intended stannane, and even undetectable levels of diphenyl concluded that ‘‘one must be aware of the fact that generation
of PhSeH from small impurities of PhSeSePh or from photolysis
7
a
b
can be problematical’’.
ARC Centre of Excellence for Free Radical Chemistry and Biotechnology, Australia
School of Chemistry and Bio21 Molecular Science and Biotechnology Institute,
The University of Melbourne, Victoria, 3010, Australia.
In search of a more robust oxyacyl radical precursor appropriate
for kinetic studies, we explored the possibility that ‘‘Kim esters’’
derived from oxalic acid may be suitable for this purpose.
We now report that alkyl N,S-dimethyldithiocarbamyl oxalate
esters (e.g. 6) are readily prepared compounds that generate oxyacyl
radicals under photochemical conditions. In addition, these novel
precursors have allowed us to kinetically ‘‘recalibrate’’ the cyclization
3
E-mail: carlhs@unimelb.edu.au; Fax: +61 3 9347 8189; Tel: +61 3 8344 2432
Electronic supplementary information (ESI): Experimental details for the pre-
†
1
13
paration of 6 and 10. General protocol for kinetic experiments. H and C NMR
spectra of compounds 6 and 10. See DOI: 10.1039/c4cc06132b
‡
Current address: Institut Parisien de Chimie Mol ´e culaire, Universit ´e Pierre et
Marie Curie, Sorbonne Universit ´e s - UMPC Univ. Paris 6, 75005 Paris, France.
1
2040 | Chem. Commun., 2014, 50, 12040--12042
This journal is ©The Royal Society of Chemistry 2014

View Article Online
Communication
ChemComm
Scheme 2
of 5; the value of k reported herein is in excellent agreement
c
with our previously reported calculations.
N,S-Dimethyldithiocarbamyl esters (Kim esters) were developed
by Kim as more robust precursors for alkyl radicals than the
corresponding pyridinethioneoxycarbonyl (PTOC) esters reported
Scheme 4
3
by Barton a few decades ago, and we have successfully employed
reaction mixtures that were difficult to analyse by GC because of
broad overlapping signals that we ascribed to tin-based byproducts.
Consequently, an alternative kinetic paradigm was required.
To our delight, when this reaction was repeated using tert-
dodecanethiol as the source of hydrogen atom, GC analysis of
the crude reaction mixture revealed cleanly the formation of
methylbutyrolactone 11 and butenylformate 12 by comparison
with authentic samples (Scheme 4). When repeated with a ten-
fold excess of thiol (pseudo first-order conditions) in benzene
21 1C) at the concentrations listed in Table 1, ratios of [12]/[11]
consistent with radical chemistry (Scheme 4) were observed
Table 1). The relative rate constant k /k is obtained by linear
regressional analysis of the data in Table 1 (Fig. 1) and
application of the integrated rate equation (eqn (2)).
12,13
these precursors on several occasions.
Guided by this previous
work, 1-octanol was stirred in oxalyl chloride overnight. Following
removal of the solvent in vacuo, the residue, presumed to be the
chloride 7 was dissolved in dichloromethane and further reacted
3
with N-hydroxy-N,S-dimethyldithiocarbamate and a catalytic
amount of DMAP in dichloromethane to afford 6 in 94% yield
(Scheme 3). The dithiocarbamate 6 was not stable to chromato-
graphy, however, in our hands, with the correct stoichiometry
and careful workup, 6 could be isolated with excellent purity.†
When a solution of 6 and tert-dodecanethiol (RSH, 0.6 M), in
benzene, was irradiated in a Rayonet photochemical reactor
(
(
H
c
(350 nm) for one hour, we were delighted to observe the
quantitative formation of 1-octyl formate (8) by GC analysis and
direct comparison with an authentic standard; presumably 8 is
formed from the intermediate oxyacyl radical 9. This outcome
demonstrates the synthetic utility of this novel radical precursor.
We next turned our attention to the cyclization of the
2
Inspection of Fig. 1 reveals excellent linearity (R = 0.996)
and confirms the validity of eqn (2); a value of the relative rate
constant (k
H
/k
c
) of 0.59 Æ 0.02 is obtained. Determining a value
for k requires knowledge of the rate constant for hydrogen atom
c
3
-butenyloxyacyl radical 5. The dithiocarbamyl precursor 10
was prepared in identical manner to 6 from 3-butenol and
was isolated in 85% yield.† Given that the rate constant (k ) for
hydrogen transfer from tributyltin hydride to an oxyacyl radical
transfer (k ) from tert-dodecanethiol (RSH); unfortunately this
H
rate constant has not been determined for oxyacyl radicals 1.
It is well established that unlike acyl radicals, oxyacyl
radicals are not stabilised by resonance, as evidenced by IR
H
7
1
has been previously reported, initial experiments were
carried out under photochemical conditions (Rayonet) in benzene
using 10 and Bu SnH. Disappointingly, these experiments provided
1
1,14
spectroscopy and computational studies.
It is somewhat
for radicals 1
surprising therefore that the reported value of k
H
3
6
À1 À1
7
reacting with Bu
3
SnH (1.1 Â 10
M
s , 80 1C) is almost
1
5
identical to that for acyl radicals (e.g. 4). On the basis of
Table 1 Relative rate data (21 1C) for the ring closure of the butenyloxy-
acyl radical 5 in benzene. Reactions performed under pseudo first-order
conditions in tert-dodecanethiol (RSH)
a
À1
b
À1
[
RSH] (M)
[12]/[11]
k
/k
H c
(M
)
c
k (s )
7
7
7
7
7
7
0
0
1
2
2
3
.43
.75
.20
.00
.50
.00
0.27
0.41
0.64
1.09
1.46
1.76
0.63
0.55
0.53
0.55
0.58
0.59
1.1 Â 10
1.2 Â 10
1.3 Â 10
1.3 Â 10
1.2 Â 10
1.2 Â 10
a
b
Average of three experiments. Spot value calculated from eqn (2)
Scheme 3
(see text).
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 12040--12042 | 12041

View Article Online
ChemComm
Communication
report data for the ring-closure of 5, product ratios are reported
5
in neat Bu
3
SnH. Consequently, it is difficult to determine
reliable reagent concentrations and, from them, to calculate rate
constants. On the other hand, more robust data are provided for
the cyclization of the related radical 13. Using the reported
5
product ratio and average value of [Bu
3
SnH], together with
7
À1
4
our value of k
c
(2.7 Â 10 s , 80 1C) for 13, we suggest that
6
À1 À1
4.6 Â 10 M
s
H
(80 1C) is a more appropriate value for k when
tributyltin hydride reacts with oxyacyl radicals 1. This value for
H
k is well aligned with that for a primary alkyl radical reacting
6
À1 À1
18
with Bu SnH, namely 6.4 Â 10 M
s
(80 1C).
3
Fig. 1 Dependence of [12]/[11] on tert-dodecanethiol concentration
at 21 1C for the cyclization of 5 in benzene.
We thank the Australian Research Council through the
Centres of Excellence Scheme for generous financial support.
1
4
relative radical stabilities, we would expect 5 to react significantly
faster with tert-dodecanethiol than its acyl counterpart (4). CCSD(T)/
aug-cc-pVDZ calculations reported previously suggest that 5 is about
Notes and references
À1
14
4
5 kJ mol less stable than 4. Applying this same methodology to
other radicals reveals that a primary alkyl radical (e.g. ethyl) is about
§
¶
This work.
H
This is k for a primary alkyl radical (5-hexenyl) reacting with tert-
À1
5
5 kJ mol less stable than an acyl radical, while a vinylic radical butylthiol (ref. 16). While many workers have assumed that tert-
À1
dodecanethiol reacts with the same rate constant as tert-butylthiol, we
have recently verified, in independent work, that the two reagents react
with primary alkyl radicals with rate constants (k ) within experimental
H
(
e.g. ethylenyl) is calculated to be about 98 kJ mol less stable
than 4 at CCSD(T)/aug-cc-pVDZ.§ These data, in turn, suggest
that an oxyacyl radical such as 5 is more similar in reactivity to a error of each other (ref. 17).
primary alkyl radical than an acyl radical. Importantly, 5 would
1
Encyclopedia of Radicals in Chemistry, Biology and Materials, ed.
C. Chatgilialoglu and A. Studer, John Wiley & Sons Ltd, Chichester,
UK, 2012.
not be expected to react with a rate constant of an acyl radical
too slow), or a vinyl radical (too fast).
With this in mind, we suggest that 5 abstracts hydrogen
atom from tert-dodecanethiol with a similar rate constant (k )
(
2
D. H. R. Barton, D. Crich and W. B. Motherwell, Tetrahedron, 1985,
41, 3901.
H
3 S. Kim, C. J. Lim, S.-E. Song and H.-Y. Kang, Chem. Commun., 2001,
1410.
4 A. N. Hancock and C. H. Schiesser, Chem. Sci., 2014, 5, 1967.
6
À1 À1
to that of a primary alkyl radical, namely 6.8 Â 10 M
s
at
1
6
2
1 1C. ¶ With this assumption, our value of k /k leads to:
H c
5
M. D. Bachi and E. Bosch, J. Org. Chem., 1992, 57, 4696.
7
À1
6 M. A. Lucas and C. H. Schiesser, J. Org. Chem., 1998, 63, 3032.
k
c
= 1.2 Â 10 s (21 1C).
7
P. A. Simakov, F. N. Martinez, J. H. Horner and M. Newcomb,
J. Org. Chem., 1998, 63, 1226.
This value of k
c
is in excellent agreement with our calculated
8
9
M. A. Lucas and C. H. Schiesser, J. Org. Chem., 1996, 61, 5754.
D. Crich and Q. W. Yao, J. Org. Chem., 1995, 60, 84.
7
À1
4
(
G3(MP2)-RAD) value of 1.5 Â 10 s (21 1C) and significantly
7
faster than the value reported previously.
10 M. Newcomb, Tetrahedron, 1993, 49, 1151.
1
1
1
1
1 C. E. Brown, A. G. Neville, D. M. Rayner, K. U. Ingold and J. Lusztyk,
Aust. J. Chem., 1995, 48, 363.
2 M. W. Carland, R. L. Martin and C. H. Schiesser, Org. Biomol. Chem.,
2004, 2, 2612.
Not only are we delighted with the agreement between our
calculated and experimental data, this study provides further
validation of our computational methodology for determining
rate data and also suggests that our assumption regarding k
for oxyacyl radicals is approximately correct.
3 T. Fenner, J. M. White and C. H. Schiesser, Org. Biomol. Chem., 2006,
H
4, 466.
4 T. Morihovitis, C. H. Schiesser and M. A. Skidmore, J. Chem. Soc.,
Perkin Trans. 2, 1999, 2041.
c
Alternatively, using our calculated value of k , we can determine
6
À1
15 C. Chatgilialoglu, C. Ferreri, M. Lucarini, P. Pedrielli and G. F. Pedulli,
a value for k
H
of 7 Â 10 s , slightly (but not significantly) higher
Organometallics, 1995, 36, 1299.
6 M. Newcomb, A. G. Glenn and M. B. Manek, J. Org. Chem., 1989,
than the value used earlier based on the experimental value for a
primary alkyl radical (vide supra).
1
54, 4603.
1
1
7 A. H. Hancock and C. H. Schiesser, unpublished.
8 L. J. Johnson, J. Lusztyk, D. D. M. Wayner, A. N. Abeywickrema,
A. L. J. Beckwith, J. C. Scaiano and K. U. Ingold, J. Am. Chem. Soc.,
1985, 107, 4594.
Finally, we are now in a position to revisit the work of Bachi
and Bosch and provide a more robust value for kH for Bu SnH
reduction of 1 by the application of eqn (2). While the authors
3
1
2042 | Chem. Commun., 2014, 50, 12040--12042
This journal is ©The Royal Society of Chemistry 2014