680
Can. J. Chem. Vol. 91, 2013
Scheme 1. Green halogenation reactions of this study.
zene (100), benzonitrile (100), and nitrobenzene (100). The lack of
reactivity of thiazole and imidazole is likely due to the acidic
conditions of the reaction, which caused protonation of the
pyridine-like nitrogen (to create =NH+–), thereby deactivating the
observed reactivity of p-nitroaniline,18,22 the donor−acceptor
compound p-nitrophenol did react (entries 31 and 32), but the
sulfur in the thiophene ring is apparently not sufficiently activat-
ing in 3-methylthiophene-2-carbonitrile or 3-methylthiophene-2-
carboxaldehyde.
There are also cases where bromination will occur, while iodi-
nation will not, i.e., brominated thiophenes will react with “Br+”
but not with “I+”. As such, 2-bromo-5-iodothiophene 1-BrI is not
accessible by iodinating bromothiophene, but is accessible
through bromination of iodothiophene (entry 13) if the stoichiom-
etry is strictly kept at 1:1. Otherwise, some dibromothiophene
1-Br2 is created (entry 14).
Although peroxide is sufficient to oxidize chloride from HCl in
petroleum ether solvent,23 we saw no evidence of chlorination
using an ammonium chloride − hydrogen peroxide mix under our
conditions. Furan did not react cleanly and benzofuran was
mostly starting material even after 24 h of reaction time. The in
situ GC results appear promising for pyrrole (entries 23−26), but
the regiochemical location of the halogen could not be deter-
mined. Attempts at isolating the product(s) consistently gave an
intractable black material, presumably a form of polypyrrole,
which is unsurprising given that the first synthesis of polypyrrole
involved an oxidative coupling using hydrogen peroxide.24,25
tive species is the neutral phenol and aniline. There is no evidence
of m-substitution, even though the acidic conditions should create
a significant fraction of protonated molecules, which would be
m-directing. We conclude that the deactivating nature of the
–NH+3 and –OH+2 groups prevents reaction with any molecules on
the protonated side of the equilibrium, thereby preventing
m-substitution from occurring. These results are consistent with
those previously published.18
This reaction can be considered green because of the relative
environmental benignity of the solvent, reactants, and byprod-
ucts. There are several metrics used to measure greeness. One of
the most common is Sheldon’s E-factor, but it neglects to take into
account the nature of the materials and solvents,26 which is the
parameter we are improving upon. A more useful metric for our
purposes is the EcoScale,27 which applies “penalty points” to toxic
or flammable materials, inefficient steps in a synthesis, etc.
Table 3 shows the EcoScale calculation, which gives a result of
2−16 penalty points in favour of our green reaction over a typical
NBS reaction.
flame ionization detector for quantitation. Yields were calculated
as relative amounts compared with the total area under the GC
peaks. When necessary, the reaction mixture was spiked with a
known amount of nitrobenzene (which does not halogenate
under these conditions) to ensure that all of the substrate is ac-
counted for in the product GC. When possible, we also cross-
checked the GC ratios with 1H NMR integration ratios by isolating
the product mixtures. Peroxide was found to be an appropriate
oxidant, as there is no evidence for S,S-dioxides in any of the
thiophene reactions, nor was there any reaction with the ammo-
nium from the bromide and iodide salts. The regiochemistry of
the products was determined by comparison of GC retention
times with those of the commercially available halogenated prod-
uct(s). When unavailable, presumed products were synthesized by
literature methods and GC times compared, or the products were
identified by a combination of NMR and GC−MS. Reactions were
allowed to run for a maximum of 20 h under ambient conditions;
as a corollary to our desire to utilize a green reaction, we wanted
a simple methodology, using minimal equipment and time.
For comparison, brominations using NBS in tetrahydrofuran
were performed (in the dark) over the same length of time as the
“Br+” brominations. Selected results are given in Tables 1 and 2;
the full results are included in the supplemental material. They
show that the regioselectivity of the green reaction is similar to
that of the NBS reaction. In one case (compare entries 2 and 6), the
green halogenation method significantly outperforms the NBS
method in selectivity, if not in time.
Functional group tolerance was tested on a number of substi-
tuted thiophenes; to increase the variety of functional groups, we
also used aniline and a number of substituted phenols,21 as the
latter are more readily available and cheaper than the equivalent
the peroxide does not react with any of the electron-donating
functional groups tested, nor with halogens (with the exception of
entry 14). Substrates with electron-withdrawing groups (benzalde-
hyde, cyanobenzene, nitrobenzene, and chlorobenzene) also
showed functional group tolerance, but unfortunately, in concor-
dance with what was previously observed for other phenyl sub-
strates,18,19 the electron-withdrawing groups deactivated their
rings to substitution. Thus, several substrates did not react with
“Br+” or “I+” to any significant extent, defined herein as having
greater than 90% starting material remaining after 20 h with a 1:1
stoichiometry. They are as follows (percentage of starting material
remaining in parenthesis): 3-methylthiophene-2-carbonitrile
(100),1,3-thiazole(100),imidazole(100),benzofuran(91),chloroben-
For the GC reactions, we used a standard organic/aqueous sep-
aration workup primarily to ensure that salts, etc., did not clog up
the GC column; such a workup would not normally be considered
“green”. To show that this extraction step would not be necessary
in preparatory-scale reactions, we performed two reactions (1-Br
¡ 1-Br2 and 4 ¡ 4-Br2) on a preparatory scale with a green
workup. The procedure involved removal of the acetic acid by
rotary evaporation followed by addition of water to dissolve the
product salts and unreacted starting materials. For 1-Br2, the
product is an insoluble oil, which was isolated using a Pasteur
pipette. The nonoptimized yield was only 36%. However, when
isolating the solid product 4-Br2 by filtration, the yield is consid-
erably better at 76%. In both cases, the purity is very high (no
evidence of any impurities, starting materials, or side-products in
1
the H NMR; see Figs. S1 and S2 in the supplementary informa-
tion), indicating the potential for a truly green isolation process
once conditions are optimized.
Published by NRC Research Press