Organic synthesis "on Water" vs "on Liquids": A comparative analysis

Article abstract of DOI:10.1021/ol500518n

Several organic reactions that are accelerated under the "on water" conditions proceed equally well or better on other "nonsolvents" or in neat conditions. Alternatively, only 1 equiv of water was necessary to obtain the same acceleration effect testifying to the fact that the water-organic compounds boundary might be unnecessary to achieve the "on water"-like reactivity.

Full text of DOI:10.1021/ol500518n

Letter
pubs.acs.org/OrgLett
Organic Synthesis “on Water” vs “on Liquids”: A Comparative
Analysis
Tal Sela and Arkadi Vigalok*
School of Chemistry, The Raymond and Beverly Sackler Faculty of Exact Sciences, Tel Aviv University, Tel Aviv 69978, Israel
S
* Supporting Information
ABSTRACT: Several organic reactions that are accelerated under the “on
water” conditions proceed equally well or better on other “nonsolvents” or in
neat conditions. Alternatively, only 1 equiv of water was necessary to obtain the
same acceleration effect testifying to the fact that the water−organic compounds
boundary might be unnecessary to achieve the “on water”-like reactivity.
ince its original introduction in 2005,1 the “on water” term
has become ubiquitous in describing organic reactions
products (esters), while water-soluble starting materials reacted
slower giving mainly the corresponding amido alcohols,
products of the reaction with water.10 Salting-out additives,
such as Na2SO4, accelerated the reaction leading to the ester
product 1, while the addition of the salting-in NaOTs resulted
in the formation of the alcohol 2 (Scheme 1).11
S
occurring in aqueous emulsions or suspensions. In many of
these reactions, rate acceleration is observed compared with the
same reactions performed in organic solutions or neat
conditions.2 The use of water could also lead to selectivity
changes.3 Nevertheless, while hydrophobic interactions that
facilitate many organic reactions in aqueous solutions have been
extensively explored,4,5 the role of the aqueous phase under the
“on water” conditions remains a mystery. Hydrogen bonding at
the water/organic compound interface, proposed by computa-
tional studies, has perhaps been the most prominent hypothesis
to explain the rate acceleration.6,7 Yet, rarely has the
“uniqueness” of water vs other nonsolvents as the reaction
medium been experimentally established. Herein, we present
our results on deciphering the “on water” effect in several
organic reactions in the context of the general “on solvent”
effect.8
Scheme 1. Salt-Dependent Selectivity in the Passerini
Reaction in- and on-Water
Although the importance of the separation between the
aqueous and organic phases in increasing the rate and
selectivity of the Passerini reaction was clearly established, the
origin of this effect remained ambiguous. Thus, we decided to
explore this reaction using a series of liquids in place of water.
Whereas the reactions were extremely sluggish when performed
homogeneously in typical organic solvents (benzene, DMSO,
or methanol), using perfluorooctane as a nonsolvent gave the
Passerini product 1 in a 94% yield after 30 min of stirring at rt
(entry 2, Table 1).12 After 10 min, the reaction on C8F18 gave
72% conversion compared with 42% (7:3 mixture of 1 and 2)
in water.13 Thus, perfluorooctane appears to be an even better
medium for this three-component reaction than water and is
comparable with 3 M Na2SO4.11 The acceleration of organic
reactions in heterogeneous mixtures with perfluorinated liquids
has previously been reported and assigned to the low cohesive
energy density, c.e.d. (and related Hildebrand parameter) of
such liquids.8,14 As water has a very high c.e.d. value,15 higher
In order to separate the genuine “on water” effect associated
with the interface boundary from other factors that can
potentially contribute to the overall “on water” reactivity, we
chose the heterogeneous reactions that are faster “on water” vs
organic solvents and (a) involve liquids, (b) proceed at rt, and
(c) do not require cosolvents or catalysts. Thus, despite the
significant “on water” acceleration, the cycloaddition of
quadricyclane and azodicarboxylates1,7c was excluded from
this study, as complex synthetic protocols, instability of one of
the reactants, and/or use of toluene as a cosolvent could all play
a role in the observed phenomenon. In all experiments, the
formal concentrations were chosen to match the typical on-
water experiments.1,2
Following the pioneering studies by Pirrung et al.,9 we
recently examined the three-component Passerini reaction that
proceeds extremely well “on water” with simple stirring at rt.
We have found that the hydrophobicity of the starting materials
plays a crucial role in the Passerini reaction “on water”: poorly
water-soluble reactants rapidly gave the Passerini reaction
Received: February 18, 2014
Published: March 24, 2014
© 2014 American Chemical Society
1964
dx.doi.org/10.1021/ol500518n | Org. Lett. 2014, 16, 1964−1967

Organic Letters
Letter
or mercury were significantly faster. For example, the reaction
between N,N-diethanolamine (6a) and glycidol (4) gave the
aminoalcohol 7a in 85% yield after 2 h when stirred on
perfluorooctane, while only a 45% conversion was obtained
after 2 h of stirring in D2O (Scheme 3, Table 2).20
Table 1. Passerini Reactions (Schemes 1 and 2) in- and on-
Liquids
a
b
entry
liquid
water
yield of 1,
%
yield of 3, %
c
d
e
1
2
3
4
5
6
7
90 (42)
44
d
C8F18
neat
94 (72)
89
85
91
5
Scheme 3. Epoxide Ring Opening with Secondary Amines
Hg
90
f
C6D6
CD3OD
1
f
0
n-C16H34
17
a
b
c
d
e
30 min. 5 min. Mixture of 1 and 2. After 10 min. D2O was used;
f
yield of the mixture of 3 and free alcohol. After 1 h.
than most solvents, it was proposed that liquids with very high
or low c.e.d. accelerate organic reactions.14 To verify this, we
used mercury as the medium for the Passerini reaction
(Caution: mercury is very toxic and must be handled in a
fume-hood using proper protecting gear). Mercury has the
highest c.e.d. values (951 cal/cm3) among the liquids.16,17 We
found that the yield of the Passerini reaction “on mercury” was
similar to those obtained on water, saturated sodium sulfate,11
or perfluoroalkanes. The yields were likely to increase if
efficient stirring was ensured, a challenging task for mercury.
Similarly, the conventional stirring was complicated for the neat
reactions, albeit the yields were only insignificantly lower than
those in the reactions on perfluoroalkanes or mercury. As we
maintained the same stirring rate throughout the experiments,
no effort was invested to further improve the yields under the
“on mercury” or neat conditions. In any event, it appears that
the observed acceleration of the Passerini reaction “on water” is
unrelated to the chemical interactions between the substrates
and water surface; efficient phase separation leading to bulking
the neat reaction played a major role in the observed effect.
Next, we reversed the situation to establish whether substrate
hydrophilicity plays a role in the observed enhanced kinetics. To
this end, we performed the Passerini reactions between the
hydrophilic water-soluble acetic acid, acetaldehyde, and 2-
furfuryl isocyanide. The reactions proceeded extremely fast and
selectively when performed on perfluoroalkanes or mercury,
but sluggishly in organic solvents or in water.
Furthermore, using hexadecane as a nonsolvent gave
significantly better yields than the reactions in water despite
the low density of the former and ensuing difficulties in efficient
stirring under these “under liquid” conditions.21 With the less
reactive hydrophobic 1-octene oxide (5), the reactions required
several days to reach significant conversion; however, stirring
on perfluoroalkanes was much more efficient than on water or
D2O (Table 2). Organic solvents (chloroform, DMSO, or
CH3CN) were the least efficient media. With the more
nucleophilic diethylamine (6b) and morpholine (6c),22 the
epoxide opening became significantly faster when water or D2O
was used but still slower than on perfluoroalkanes or neat.
Finally, when the hydrophobic solid N-(o-tolyl)piperazine was
used (6d), the “on water” reaction was noticeably faster than on
perfluoroalkanes or in neat. These starting materials are the
most similar to the originally reported system,1 thus confirming
the observed “on water” acceleration with the hydrophobic
substrates. The “on water” acceleration was also observed for
the liquid N-(phenyl)piperazine (6e) and nonaromatic N-
(cyclohexylmethyl)piperazine (6f).
When performed under the highly dilute homogeneous
conditions in a D2O solution (4 mmol/L for the limiting 1-
octene oxide),23 the reaction was extremely slow not allowing
for the kinetic studies. The results indicate that the reaction
does not proceed in the aqueous solution. Still, performing the
reaction in the moderately kosmotropic NaCl (4 M) led to
some decrease in conversion. Conversely, the reaction can be
accelerated (79% vs 69% on water) by the addition of the
surfactant sodium dodecyl sulfate (SDS),21 at concentrations
above its critical micelle concentration. As the product 8d can
be considered amphiphilic and potentially form micelles, we
also verified the possibility of the product acceleration
mechanism. When 0.5 equiv of 8d was added at the beginning
of the reaction, the conversion was lower (48% vs 69% on
water) testifying against this mechanism. Unfortunately, the
obtained information did not explain the apparent “on water”
acceleration leading us to investigate the reaction “in the
presence of water”.24,25 Surprisingly, in the presence of 1 equiv
of water, the reaction between 1-octene oxide and N-(o-
tolyl)piperazine “on perfluorooctane” gave a 67% conversion,
comparable with that obtained “on water”. Adding 3 equiv of
water further increased the conversion (77% after 17 h, Table
2).
In the latter case, mainly the analogue to 2 amido alcohol was
observed (Scheme 2, Table 1). Overall, these data demonstrate
Scheme 2. Passerini Reaction with the Hydrophilic
Reactants
that, in the truly “on solvent” Passerini reactions, water has no
distinct beneficial effect on the selectivity and yields when
compared with other nonsolvents.
The nucleophilic ring opening of epoxides with secondary
amines is another reaction that was reported to proceed very
efficiently under the heterogeneous “on water” conditions and
relatively slowly in the organic media.1,18,19 We sought to
compare this reaction in solutions vs “on liquids” at rt using
liquid substrates. We found that when both substrates were
fully miscible with water, the nucleophilic ring opening was very
slow while the same reactions performed on perfluoroalkanes
1965
dx.doi.org/10.1021/ol500518n | Org. Lett. 2014, 16, 1964−1967

Organic Letters
Letter
a
Table 2. Epoxide Ring Opening with Secondary Amines in- and on-Liquids
entry
liquid
D2O
yield of 7, %
yield of 8, %
b
b
b
1
2
3
4
5
6
7
7a 45, 7b 67, 7c 84
7a 85, 7b 75, 7c 95
7a 95, 7b, 50, 7c 94
7a 95, 7b 70, 7c 94
8a 65 (53), 8b 48 (33), 8c 47 (35)
C8F18
Hg
8a 98, 8b 31, 8c 58, 8d 36, 8e 28, 8f 50
Neat
8a 99, 8b 15, 8c 47, 8d 40, 8e 24, 8f 37
b
H2O
8d 69 (65), 8e 50, 8f 69
c
d
neat + H2O
8d 72 (73)
c
d
d
C8F18 + H2O
8d 67 (77), 8f 77
a
b
c
d
Reaction times, h: 7a−c 2, 8a 66, 8b 42, 8c 42, 8d 17 or 23, 8e,f 17. In 4 M NaCl. 1 equiv of H2O. 3 equiv of H2O.
An increase to 10 equiv did not lead to a further increase of
water” transformations, the majority of reported “on water”
accelerated reactions involved solid reactant(s).2
the product yield. Similarly, a high yield was obtained when 1
equiv of water was added to the neat reaction between 5 and
6d. Thus, only 1−3 equiv of water are sufficient to observe a
substantial increase in the product’s yield. While such
stoichiometric interactions between the organic compounds
and water molecules may involve hydrogen bonding, there was
clearly no need for the water−organic layer interface to obtain
the “on water”-like acceleration. The lack of a positive effect of
additional water, beyond the stoichiometric amounts, also
argues against such a need.
Although not comprehensive, our studies suggest that the
presence of a defined water surface is not a prerequisite for
obtaining the “on water”-like acceleration. In three different
reactions, similar or better results were obtained by mixing neat
substrates or adding nonsolvents, e.g. perfluoroalkane. The
observed behavior is consistent with high concentrations (neat
or bulked neat reactions) and, in one case, stoichiometric
interactions between some organic compounds and water
molecules responsible for the enhanced reactivity. Certainly,
more studies are necessary to better assess the generality of our
findings. Nevertheless, from the practical point of view, the “on
water” reaction protocols may have a significant advantage over
other conditions that provide similar yields. We are presently
investigating potential applications of these findings in affecting
the reactivity of organic compounds.
Ene-reactions between unactivated alkenes and azodicarbox-
ylates were reported to be accelerated “on water”.1,9b Although,
in our hands, the reactions between β-pinene (9) and isopropyl
azodicarboxylate (10) were faster than reported earlier,9b
a
clear preference for water vs CH2Cl2 or chloroform was
observed at rt (Scheme 4, Table 3). However, the reactions on
ASSOCIATED CONTENT
* Supporting Information
Scheme 4. Azodicarboxylate Addition to β-Pinene
■
S
General experimental protocols and products characterization
for all new compounds. This material is available free of charge
via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
■
Table 3. Isopropyl Azodicarboxylate Addition to β-Pinene
*E-mail: avigal@post.tau.ac.il.
Notes
a
entry
liquid
yield of 11,
%
1
2
3
4
5
6
7
8
H2O
60 (41)
58 (42)
43 (30)
46 (32)
67 (50)
71 (54)
67 (50)
66 (49)
The authors declare no competing financial interest.
4 M NaCl in H2O
CDCl3
ACKNOWLEDGMENTS
CDCl3 + 3 H2O
C6F14
■
This work was supported by Grant 1565/12 from the Bikura
(FIRST) program administered by the Israel Science
Foundation.
C6F14 + 3 H2O
neat
neat +3 H2O
a
REFERENCES
Yields are after 1 h at rt; those in parentheses are after 30 min.
■
(1) Narayan, S.; Muldoon, J.; Finn, M. G.; Fokin, V. V.; Kolb, H. C.;
Sharpless, K. B. Angew. Chem., Int. Ed. 2005, 44, 3275.
(2) For reviews, see: (a) Butler, R. N.; Coyne, A. G. Chem. Rev. 2010,
110, 6302. (b) Chanda, A.; Fokin, V. V. Chem. Rev. 2009, 109, 725.
(c) Gawande, M. B.; Bonifacio, V. D. B.; Luque, R.; Branco, P. S.;
́
Varma, R. S. Chem. Soc. Rev. 2013, 5522.
(3) (a) Breslow, R.; Maitra, U.; Rideout, D. Tetrahedron Lett. 1983,
24, 1901. (b) Lubineau, A. J. Org. Chem. 1986, 51, 2144. (c) Cativiela,
C.; García, J. I.; Mayoral, J. A.; Salvatella, L. J. Chem. Soc., Perkin Trans.
2 1994, 847.
(4) (a) Rideout, D. C.; Breslow, R. J. Am. Chem. Soc. 1980, 102, 7816.
(b) Breslow, R. Acc. Chem. Res. 1991, 24, 159.
perfluorohexane or in neat were even slightly more efficient.
The addition of water to the heterogeneous mixture on C6F14,
neat, or in solution in CDCl3 did not lead to an increase of the
reaction yield. Thus, it is likely that the acceleration observed
for the reaction between cyclohexene and solid bis(trichloro-
ethyl) azodicarboxylate on water was due to the more efficient
formation of a molten organic phase, as was proposed in ref 1.
When only liquid components are used, the ene-reaction
proceeds equally efficiently on water, neat, or on a nonsolvent,
such as perfluoroalkane. Interestingly, while clearly separable
organic liquids are considered ideal candidates for the “on
(5) (a) Otto, S.; Engberts, J. B. F. N. Org. Biomol. Chem. 2003, 1,
2809. (b) Gajewski, J. J. Acc. Chem. Res. 1997, 30, 219.
(6) Jung, Y.; Marcus, R. A. J. Am. Chem. Soc. 2007, 129, 5492.
1966
dx.doi.org/10.1021/ol500518n | Org. Lett. 2014, 16, 1964−1967

Organic Letters
Letter
(7) (a) Acevedo, O.; Armacost, K. J. Am. Chem. Soc. 2010, 132, 1966.
(b) Jung, Y.; Marcus, R. A. J. Phys.: Condens. Matter 2010, 16, 1.
(c) Mellouli, S.; Bousekkine, L.; Theberge, A. B.; Huck, W. T. S.
Angew. Chem., Int. Ed. 2012, 51, 7981. (d) Manna, A.; Kumar, A. J.
Phys. Chem. A 2013, 117, 2446. (e) Beattie, J. K.; McErlean, C. S. P.;
Phippen, C. B. W. Chem.Eur. J. 2010, 16, 8972.
(8) (a) Scherrmann, M. C.; Norsikian, S.; Lubineau, A. Adv. Org.
Synth. 2005, 1, 341. (b) Breslow, R. Phys. Org. Chem. 2006, 19, 813.
(9) (a) Pirrung, M. C.; Das Sarma, K. J. Am. Chem. Soc. 2004, 126,
444. (b) Pirrung, M. C.; Das Sarma, K. J.; Wang. J. Org. Chem. 2008,
73, 8723.
(10) Shapiro, N.; Vigalok, A. Angew. Chem., Int. Ed. 2008, 47, 2849.
(11) Sela, T.; Vigalok, A. Adv. Synth. Catal. 2012, 354, 2407.
(12) Bulking organic reactions with perfluoroalkanes have previously
been reported: (a) Crampton, M. R.; Cropper, E. L.; Gibbons, L. M.;
Millar, R. W. Green Chem. 2002, 4, 275. (b) Chambers, R. D.;
Edwards, A. R. J. Chem. Soc., Perkin Trans. 1 1997, 3623.
(13) For a more accurate comparison, throughout the studies, the
reactions were deliberately stopped at incomplete conversions. In most
cases, longer reaction times led to quantitative yields (see Supporting
Information).
(14) Myers, K. E.; Kumar, K. J. Am. Chem. Soc. 2000, 122, 12025.
(15) The importance of high c.e.d. of water in the acceleration of the
Passerini reaction has been proposed: Pirrung, M. C.; Das Sarma, K.
Tetrahedron 2005, 61, 11456.
(16) (a) Spencer, J. N.; Voigt, A. F. J. Phys. Chem. 1968, 72, 464.
(b) Vavruch, I. J. Colloid Interface Sci. 1995, 169, 249.
(17) Mercury also has the largest internal pressure value (325 cal/
cm3) although its contribution to the chemical reactivity is not well
accepted. A recent comprehensive review on liquid’s internal pressure
critically addresses this question: Marcus, Y. Chem. Rev. 2013, 113,
6536.
(18) Chanda, A.; Fokin, V. V. In Water in Organic Synthesis;
Kobayashi, S., Ed.; Stuttgart: 2013; p 697.
(19) Azizi, N.; Saidi, M. R. Org. Lett. 2005, 7, 3649.
(20) We verified that the reactions on D2O are insignificantly slower
than those on H2O to affect the overall conclusions. See also footnote
71 in ref 6.
(21) See Supporting Information for more details.
(22) Brotzel, F.; Chu, Y. C.; Mayr, H. J. Org. Chem. 2007, 72, 3679.
(23) Octene oxide solubility in D2O is ca. 9.2 mmol/L as determined
1
by the H NMR spectroscopy.
(24) (a) Ribe, S.; Wipf, P. Chem. Commun. 2001, 299. (b) Hayashi, Y.
Angew. Chem., Int. Ed. 2006, 45, 8103.
(25) Windmon, N.; Dragojlovic, V. Green Chem. Lett. Rev. 2008, 1,
155.
1967
dx.doi.org/10.1021/ol500518n | Org. Lett. 2014, 16, 1964−1967