Triphosgene and DMAP as Mild Reagents for Chemoselective Dehydration of Tertiary Alcohols

Article abstract of DOI:10.1021/acs.orglett.9b01959

The utility of triphosgene and DMAP as mild reagents for chemoselective dehydration of tertiary alcohols is reported. Performed in dichloromethane at room temperature, this reaction is readily tolerated by a broad scope of substrates, yielding alkenes preferentially with the (E)-geometry. While formation of the Hofmann products is generally favored, a dramatic change in alkene selectivity toward the Zaitzev products is observed when the reaction is carried out in dichloroethane at reflux.

Full text of DOI:10.1021/acs.orglett.9b01959

Letter
pubs.acs.org/OrgLett
Cite This: Org. Lett. XXXX, XXX, XXX−XXX
Triphosgene and DMAP as Mild Reagents for Chemoselective
Dehydration of Tertiary Alcohols
Moshood O. Ganiu, Alexander H. Cleveland, Jarrod L. Paul, and Rendy Kartika*
Department of Chemistry, Louisiana State University, 232 Choppin Hall, Baton Rouge, Louisiana 70803, United States
S
* Supporting Information
ABSTRACT: The utility of triphosgene and DMAP as mild reagents for chemoselective dehydration of tertiary alcohols is
reported. Performed in dichloromethane at room temperature, this reaction is readily tolerated by a broad scope of substrates,
yielding alkenes preferentially with the (E)-geometry. While formation of the Hofmann products is generally favored, a dramatic
change in alkene selectivity toward the Zaitzev products is observed when the reaction is carried out in dichloroethane at reflux.
Scheme 1. Elimination of Tertiary Alcohol
ehydration of tertiary alcohols is a useful synthetic
Dreaction to produce substituted alkenes, but it remains a
formidable challenge particularly with respect to controlling
the resulting double bond selectivity. The difficulty in this
functional group interconversion can be also underscored by
the fact that the required reagents are often incompatible with
various sensitive functional groups. For instance, it is well
precedented that dehydration of tertiary alcohols can be
effectively promoted by strong Brønsted acids.1 Nonetheless,
such conditions are not generally amendable to complex
molecules synthesis, and this reaction can potentially lead to a
mixture of alkene isomers due to the propensity of the
participating carbocation intermediates to undergo structural
rearrangement.2 Other broadly used reagents include thionyl
chloride3 and phosphoryl chloride,4 but a mixture of
dehydration products could be similarly produced due to the
involvement of chlorosulfite 3a or dichlorophosphate 3b
intermediates (Scheme 1) that are reportedly subject to facile
self-ionization to carbocations. While there are several elegant
accounts on this fundamental transformation that highlight
heterogeneous,5 homogeneous,6 and enzymatic catalysis,7
arguably the most applicable method to dehydrate tertiary
alcohols in complex molecular systems can be found in the use
of the Burgess reagent.8 With respect to Burgess activation,9
the origin of alkene selectivity arises from controlling the
regioselective syn (Ei) elimination of the participating
sulfamate ester 3c.
reaction mechanism from our chlorination reactions, we
envisioned an intermediacy of ammonium carbamate ion 3d
that could, in theory, be subjected to elimination, therefore
producing the double bond. From the operational perspective,
the use of triphosgene as a dehydration reagent would be more
advantageous than SOCl2 or POCl3, as triphosgene exists as
stable, nonhygroscopic crystalline materials at room temper-
ature that can be conveniently and safely handled in typical
laboratory operations.11
Table 1 depicts our screening study. To test our hypothesis,
we employed tertiary alcohol 4 as a model substrate. As shown
in entry 1, this compound was initially subjected to conditions
pertinent to our chlorination technology, i.e. 0.5 equiv of
triphosgene and 2.0 equiv of pyridine, in dichloromethane at
In recent years, our group has developed methods to
chlorinate primary and secondary alcohols using a mixture of
triphosgene and amine bases, particularly triethylamine or
pyridine.10 Recognizing that SOCl2 or POCl3 also serve as
reagents for the chlorination of alcohols, we hypothesized that
our triphosgene-amine base chemistries could be perhaps
applied to dehydrate tertiary alcohols. Drawing a parallel
Received: June 6, 2019
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.9b01959
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Table 1. Reaction Optimization12
Scheme 2. Scope of Aromatic Groups12
temp
(°C)
conc
(M)
time
(h)
c
a
entry
base
equiv
4:5:6:7
1
2
3
4
5
6
7
8
pyridine
pyridine
pyridine
pyridine
pyridine
DMAP
DMAP
DMAP
DMAP
DMAP
DMAP
DMAP
DMAP
DMAP
DMAP
DMAP
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
5.0
1.5
1.0
0.0
2.0
reflux
rt
0
−78
rt
reflux
rt
0.5
0.5
0.5
0.5
0.1
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.2
2
2
0:57:17:26
0:58:12:30
0:63:08:29
7:66:04:23
0:61:14:25
0:80:16:4
4
24
18
18
26
22
22
28
22
24
24
24
24
0:78:13:9
0
1:76:12:11
11:70:10:09
77:18:03:03
100:00:00:00
0:80:13:07
10:58:13:19
0:41:09:50
100:00:00:00
100:00:00:00
9
−20
−40
−78
rt
rt
rt
whereas the alkene stereoselectivity in 9f eroded with electron-
deficient CF3 groups. Halogenated substituents (9d and 9e)
did not alter the ratio of (E) vs (Z) from that of the parent
phenyl group 9a;12e but longer reaction times were noted. The
utility of other aromatic rings was also investigated. While the
naphthyl substrate led to product 9g with reduced selectivity,
the thiophene variant 9h was obtained essentially only as an
(E)-isomer.
10
11
12
13
14
15
rt
rt
b c
,
16
a
Ratio of the compounds was determined by GC-MS analyses with an
The utility of our method to dehydrate cyclic tertiary
alcohols, viz. 10, was also examined. As shown in Scheme 3, we
assumption that they elicited identical response. The alkene geometry
was deduced by NOE. Triphosgene was absent from the reaction
mixture. The use of other organic bases, such as 2,6-lutidine and
DBU, led to the recovery of starting material 4.
b
c
Scheme 3. Scope of Substrates12
reflux.10b,c Interestingly, this reaction indeed produced
elimination products 5 and 6, favoring the (E)-isomer, but a
substantial amount of chlorination product 7 was also
observed. Attempts to minimize chlorination by systematically
cooling the reaction temperature and diluting the reaction
concentration were not successful (entries 2 to 5). Nonetheless
at −78 °C, the reaction exhibited much higher selectivity
toward (E)-alkene 5.
Assuming that alkyl chloride 7 was produced via an SN1
pathway that occurred competitively, we proposed that the
utility of a slightly stronger base could perhaps further favor
kinetics toward the desired elimination. As noted in entry 6,
replacement of pyridine with DMAP indeed suppressed
chlorination. The reaction temperature was then systematically
modulated to improve alkene selectivity, but the (E) vs (Z)
selectivity was relatively unchanged (entries 6−9). Interest-
ingly, substrate 4 failed to react when the temperature was
cooled below −40 °C. We also surveyed the molar amount of
DMAP (entries 12−14). While excess DMAP did not affect
the statistical distribution of products, lowering its equivalence
to 1.5 and 1.0 surprisingly led to a significant increase of
chlorination. Our control experiments are shown in entries 15
and 16, as starting material 4 was not affected when DMAP or
triphosgene was absent from the reaction mixture. Given the
reactions produced comparable results both at reflux and at
room temperature (entries 6−7), the latter was chosen as
optimized conditions due to its operational simplicity.
a
b
rr = regiomeric ratio. The reaction was performed in 2-g scale.
subjected various ring sizes to this study.13 With the exception
of a cyclobutanol-derived substrate that resulted in decom-
position, 5−8-membered starting materials could be smoothly
eliminated with triphosgene and DMAP to produce the
corresponding 1-phenylcycloalkenes 11b−11e in good yields.
Our reaction is amendable for scale-up, as product 11c was
cleanly obtained in 97% yield from 2 g of the respective
substrate. Tertiary alcohols bearing electronically diverse aryl
substituents 11f−11h and aliphatic groups 11i−11k were
We proceeded to evaluate the scope of substrates in
preparative scale reactions, commencing with tertiary alcohols
8 that were elaborated with various aromatic groups (Scheme
2).12 These initial studies revealed that substrates bearing
electron-rich anisole and dimethylamino substituents further
favored the (E)-geometry in the resulting alkenes 9b and 9c,
B
DOI: 10.1021/acs.orglett.9b01959
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
found to be tolerated. Remarkably, compounds 11i−11k were
furnished as the predominant regioisomers, rendering elimi-
nation to the methylenecyclohexane isomer relatively non-
competitive. An attempt to dehydrate 1-cyclohexyl-1-cyclo-
butanol did not yield cyclobutene. Instead, the reaction
produced tetrasubstituted alkene 11l. We also prepared and
evaluated substrates containing common protecting groups.
While TBDPS, Bz, and Bn were found to be compatible and
furnish cyclohexenes 11m−11o in high yields, TBS and MOM
ethers were surprisingly cleaved under the reaction conditions.
As shown in Scheme 4,12 our studies continued with
substrate motif 12 to identify whether regioselectivity between
examined the applicability of our method toward dehydration
of tertiary alcohols concomitantly with chlorination of these
functional groups. As depicted in Table 2, we prepared a series
Table 2. Concomitant Dehydration and Chlorination12
Scheme 4. Hofmann vs Zaitzev Selectivity12
a
rr = regiomeric ratio.
of substrates 15a−17a. To affect transformation for each
functional group, these compounds were subjected to reaction
conditions in which the molar amount of the reagents was
doubled, i.e. 1.0 equiv of triphosgene and 4.0 equiv of DMAP.
These studies revealed that chemoselective dehydration and
chlorination could be achieved in a single synthetic operation,
as cyclohexene-bearing aliphatic chlorides 15b and 16b, as well
as dichloride 17b, were generated in good yields.
Scheme 5 signifies the advantage of our method, as it
tolerates a complex molecule that bears various functional
Scheme 5. Application to a Complex Molecule12
a
Further alkene isomerization to trisubstituted endocyclic alkene 1-
(cyclohexen-1-yl)ethyl)benzene, i.e. compound 14b′, was observed.
Hofmann and Zaitsev products 13 and 14, respectively, could
be biased. Using our optimized conditions (Conditions A), we
found that tertiary alcohols bearing a sterically congested
isopropyl and cyclohexyl group predominantly generated
exomethylene products 13a and 13b. Nonetheless, the
Hofmann preference eroded with an n-octyl substituent in
13c or a phenyl group in 13d, most likely driven by π
conjugation in the resulting stilbene. Remarkably, the Zaitzev
products 14a−14d could be primarily formed simply by
changing the solvent to dichloroethane and performing the
reactions at reflux (Conditions B). Both Conditions A and B,
furthermore, yielded alkenes 14c and 14d with a strong
preference toward the (E)-geometry. The equilibrium between
the Hofmann vs Zaitzev products could be directly monitored
by subjecting substrate 12a to triphosgene-DMAP in dichloro-
ethane at two successive temperature profiles (Conditions C).
While at room temperature the reaction initially produced
exomethylene 13a, gradual isomerization to the fully
substituted alkene 14a was noted as the same crude mixture
was heated to reflux.
groups. Using montelukast methyl ester (+)-18 as an example,
exposure of this compound to triphosgene and DMAP in
dichloromethane at room temperature cleanly furnished the
corresponding dehydration product (+)-19 in 96% yield. In
this instance, the thioether, cyclopropyl, methyl ester, and
quinoline moieties were found to be inconsequential to the
reaction conditions for the elimination of the tertiary alcohol.
As proposed in Scheme 6, the mechanism for our
dehydration reaction commences with the decomposition of
triphosgene by DMAP to yield in situ generated, putative
dehydration reagent 20.14 This species subsequently activates
the tertiary alcohol moiety to pyridinium carbamate ion 21a,
leading to E2 elimination of one of the least sterically
congested β-hydrogens by excess DMAP to afford the
Hofmann products. Nonetheless, the effectiveness of DMAP
in this chemistry as opposed to pyridine or other amine bases
(Table 1) can also imply an alternative mechanism, in
particular the Ei elimination.15 The possibility for this pathway
Given that functional groups, such as primary and secondary
alcohols as well as epoxides, could be effectively chlorinated
with a mixture of triphosgene and pyridine,10b−d we then
C
DOI: 10.1021/acs.orglett.9b01959
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
the Hofmann product upon its protonation to the
carbocationic intermediate by the conjugate acid of DMAP.
In conclusion, we have developed a new synthetic reaction
for the chemoselective dehydration of tertiary alcohols using a
mixture of triphosgene and DMAP. Our chemistry exhibited
interesting reactivity behaviors, which include its preference
toward formation of (E)-alkene as well as its ability to affect
the Hofmann versus Zaitzev selectivity through the change of
solvent and reaction temperature. Further studies on the
reaction mechanism are currently ongoing in our laboratory.
Our results will be reported in due course.
Scheme 6. Proposed Reaction Mechanism
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.9b01959.
Experimental procedures and spectral data of new
compounds (PDF)
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: rkartika@lsu.edu.
ORCID
Rendy Kartika: 0000-0002-2042-2812
Notes
The authors declare no competing financial interest.
can be reasoned by a notion that activation of the carbamate
moiety by electron-rich DMAP enables resonance structures
21b ↔ 21c, thus substantially raising the HOMO en route to
the proposed internal elimination mechanism.
ACKNOWLEDGMENTS
■
This material is based upon work supported by the National
Science Foundation under CHE-1464788. Generous financial
supports from Louisiana State University are gratefully
acknowledged. A.H.C. thanks the Louisiana Board of Regents
for the Graduate Fellowship (LEQSF(2015-20)-GF-02).
Whether proceeding via E2 or Ei elimination,16 either
mechanism will overall recycle 2 molar equiv of DMAP in its
protonated state and simultaneously extrude CO2 gas as the
sole byproduct. Furthermore, the strong preference toward (E)
vs (Z) alkene geometry as described in Scheme 2 can be
explained from the minimization of destabilizing gauche
interactions in the respective reactive intermediate 22a vs
22b, preceding the elimination of the pyridinium carbamate
moiety in a syn fashion for the Ei pathway or antiperiplanar for
E2. Supported by the trend in electronic effects (i.e., electron-
rich arene groups improve (E) selectivity), the underlying
factor toward this preorganized conformational bias most likely
originates from stabilization of the σ*C−O antibonding orbital
by the π electrons, which requires orientation of the aryl
substituents that consequently render 22b unfavorable due to
steric repulsion.
REFERENCES
■
(1) Henne, A. L.; Matuszak, A. H. J. Am. Chem. Soc. 1944, 66, 1649.
(2) (a) Dubois, J. E.; Lomas, J. S.; Sagatys, D. S. Tetrahedron Lett.
1971, 12, 1349. (b) Grob, C. A.; Schiess, P. W. Angew. Chem., Int. Ed.
Engl. 1967, 6, 1.
(3) Lomas, J. S.; Sagatys, D. S.; Dubois, J. E. Tetrahedron Lett 1971,
12, 599.
(4) Sauers, R. R. J. Am. Chem. Soc. 1959, 81, 4873.
(5) (a) Frija, L. M. T.; Afonso, C. A. M. Tetrahedron 2012, 68, 7414.
(b) D'Onofrio, F.; Scettri, A. Synthesis-Stuttgart 1985, 1985, 1159.
(c) Firouzabadi, H.; Iranpoor, N.; Hazarkhani, H.; Karimi, B. Synth.
Commun. 2003, 33, 3653. (d) Hoffman, R. V.; Bishop, R. D.; Fitch, P.
M.; Hardenstein, R. J. Org. Chem. 1980, 45, 917. (e) Nishiguchi, T.;
Kamio, C. J. Chem. Soc., Perkin Trans. 1 1989, 1, 707.
(6) (a) Alvarez-Manzaneda, E. J.; Chahboun, R.; Torres, E. C.;
Alvarez, E.; Alvarez-Manzaneda, R.; Haidour, A.; Ramos, J.
Tetrahedron Lett. 2004, 45, 4453. (b) Martin, J. C.; Arhart, R. J. J.
Am. Chem. Soc. 1971, 93, 4327.
(7) (a) Demming, R. M.; Fischer, M. P.; Schmid, J.; Hauer, B. Curr.
Opin. Chem. Biol. 2018, 43, 43. (b) Nestl, B. M.; Geinitz, C.; Popa, S.;
Rizek, S.; Haselbeck, R. J.; Stephen, R.; Noble, M. A.; Fischer, M. P.;
Ralph, E. C.; Hau, H. T.; Man, H.; Omar, M.; Turkenburg, J. P.; van
Dien, S.; Culler, S. J.; Grogan, G.; Hauer, B. Nat. Chem. Biol. 2017, 13,
275.
Formation of the alkyl chloride and Zaitsev byproducts can
be accounted by the following scenarios. While less likely,17 it
is conceivable that pyridinium carbamate ion 21a (or its
chloroformate analog) can similarly undergo self-ionization to
produce a tertiary carbocation, which can be captured by
chloride ions that are liberated in abundance from triphosgene.
As observed in Table 1, this E1 pathway becomes competitive
with the use of pyridine or with the decreasing amount of
DMAP in the reaction mixture based on the increasing amount
of alkyl chloride. Naturally, the involvement of carbocationic
intermediates will also lead to the Zaitzev product. Based on
the facile conversion of 13a to 14a under equilibration
conditions (Scheme 4), we cannot rule out the possibility that
the Zaitzev product is actually produced via isomerization of
(8) Recent examples: (a) Liu, Y.; Wang, H. Chem. Commun. 2019,
55, 3544. (b) Yu, X.; Xiao, L.; Wang, Z.; Luo, T. J. Am. Chem. Soc.
̈
2019, 141, 3440. (c) Peter, D.; Bruckner, R. Eur. J. Org. Chem. 2018,
2018, 6256. (d) Feng, J.; Yu, T.; Zhang, Z.; Li, J.; Fu, S.; Chen, J.; Liu,
D
DOI: 10.1021/acs.orglett.9b01959
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
B. Chem. Commun. 2018, 54, 7665. (e) Xuan, J.; Zhu, A.; Ma, B.;
̈
Ding, H. Org. Lett. 2018, 20, 4153. (f) Peter, D.; Bruckner, R. Chem. -
pyridine (15 equiv) furnished alkene mixtures 5:6 in a 81:19 ratio
without the presence of alkyl chloride 7.
Eur. J. 2017, 23, 12104. (g) Cheng, H.; Zhang, Z.; Yao, H.; Zhang,
W.; Yu, J.; Tong, R. Angew. Chem., Int. Ed. 2017, 56, 9096.
(9) (a) Atkins, G. M.; Burgess, E. M. J. Am. Chem. Soc. 1968, 90,
4744. (b) Burgess, E. M.; Penton, H. R.; Taylor, E. A. J. Am. Chem.
Soc. 1970, 92, 5224. (c) Burgess, E. M.; Penton, H. R.; Taylor, E. A. J.
Org. Chem. 1973, 38, 26.
(10) (a) Ayala, C. E.; Villalpando, A.; Nguyen, A. L.; McCandless, G.
T.; Kartika, R. Org. Lett. 2012, 14, 3676. (b) Villalpando, A.; Ayala, C.
E.; Watson, C. B.; Kartika, R. J. Org. Chem. 2013, 78, 3989.
(c) Villalpando, A.; Saputra, M. A.; Tugwell, T. H.; Kartika, R. Chem.
Commun. 2015, 51, 15075. (d) Cleveland, A. H.; Fronczek, F. R.;
Kartika, R. J. Org. Chem. 2018, 83, 3367.
(11) (a) Eckert, H.; Forster, B. Angew. Chem., Int. Ed. Engl. 1987, 26,
894. (b) Damle, S. B. Chem. Eng. News 1993, 71, 4. (c) Cotarca, L.
Org. Process Res. Dev. 1999, 3, 377.
(12) (a) Isolated yields after column chromatography. (b) The ratio
of alkene isomers was determined by GC-MS analyses of the crude
reaction mixtures with an assumption that they elicited an identical
response. (c) The alkene geometry of the major isomer was deduced
by NOE. (d) The calculated ratio of alkene isomers may not be
reflective of the actual ratio in the crude reaction mixtures due the
possibility for some alkenes to isomerize under our GC conditions,
e.g. compounds 13c and 14c. (e) The lower isolated product yields
in some examples were due to the high volatility of the alkenes. (f)
The content of an alkyl chloride in the crude reaction mixture, as
determined by GC-MS with an assumption that they elicited an
identical response, typically ranges between 0% to 9%. The only
exception is substrate 8d, which produced 18% of the chlorination
product. See Table S-1 and GC chromatograms in the Supporting
Information for details on each substrate.
(13) During the course of our study, a single example on the
dehydration of tertiary cyclohexanol using excess triphosgene and
pyridine was reported: Liu, Y.; Yeung, Y.-Y. Org. Biomol. Chem. 2017,
15, 6478.
(14) Putative dehydration reagent 20 was proposed given the molar
equivalances of triphosgene and DMAP in our reaction. A second
addition of DMAP to this active species is also hypothetically possible.
For an analogous reaction between phosgene and pyridine: King, J.
A.; Donahue, P. E.; Smith, J. E. J. Org. Chem. 1988, 53, 6145.
(15) Ei elimination generally favors the formation of Hofmann
products: (a) Froemsdorf, D. H.; Collins, C. H.; Hammond, G. S.;
DePuy, C. H. J. Am. Chem. Soc. 1959, 81, 643. (b) Haag, W.; Pines, H.
J. Org. Chem. 1959, 24, 877. (c) DePuy, C. H.; King, R. W. Chem. Rev.
1960, 60, 431.
(16) Preliminary attempts to discern the E2 vs Ei mechanism using
diastereomerically pure tertiary alcohols 23a and 23b were not
productive. Dehydration of 23a afforded primarily exomethylene 24a
along with a lesser amount of stilbenes 24b as as mixture of alkene
isomers, which most likely have been generated via carbocationic
intermediates. The major isomer of 24b was not determined.
Substrate 23b was unreactive under the reaction conditions.
(17) (a) Dehydration of substrate 4 with catalytic TsOH (0.1
equiv) in dichloromethane (0.5 M) at reflux furnished alkenes 5 and 6
as a 62:38 mixture. Had our reaction proceeded primarily via the E1
mechanism, we believed that poorer (E) vs (Z) alkene selectivity as
well as the Zaitzev preference would have been observed. (b) For
comparison, dehydration of substrate 4 was performed with SOCl2 or
POCl3 in the presence of pyridine, see refs 3 and 4, respectively.
Treatment with SOCl2 (1.5 equiv) and pyridine (5.0 equiv) produced
a 41:8:51 mixture of products 5:6:7, whereas POCl3 (2.5 equiv) and
E
DOI: 10.1021/acs.orglett.9b01959
Org. Lett. XXXX, XXX, XXX−XXX