J.A. Giesen, S.M. Grayson / Tetrahedron Letters 61 (2020) 152016
3
salts. This is followed by the base-catalyzed hydrolysis of 8, afford-
ing 2. The benzylation of this compound was carried out with rel-
ative ease, and the presence of residual benzyl alcohol (BnOH) from
the etherification or hydrolysis is readily removed via liquid–liquid
extraction with organic solvent from water at pH ~7. Unfortu-
nately, the other hydrolytic byproduct, benzoic acid, is exceedingly
difficult to remove via extraction, due to the relative pKa’s of ben-
zoic acid (4.20) and the AB monomer 2, determined to be 4.43
(supplemental Fig. S1). This would require purification via liquid
column chromatography (LCC) requiring an undesirable amount
of time and large volume of waste.
Another route that was considered was the direct benzylation of
bis-MPA with NaH, (2 eq.) and BnBr (1.8–2 eq.) (Scheme 4). The
NMR of the crude product shows the formation of the desired
monobenzyl ether protected compound of the benzyl ester (10).
However, without a means to control the mono-functionalization,
the reaction contains byproducts of the unprotected diol (9) and
dibenzyl ether (11), each with a benzyl ester. Ultimately, this path-
way requires LCC for purification and removes it as a viable option
for large scale synthesis due to the time intensive purification.
There are a few methods for the selective mono-benzylation of
unprotected diols [32], the most attractive is the selective mono-
benzylation with BnBr in the presences of silver(I) oxide (Ag2O)
[33,34]. In order to attempt this selective mono-benzylation the
carboxylic acid was first converted to the ethyl ester (12) via a Fis-
cher acid catalyzed esterification (Scheme 5). Bis-MPA (1) was
added to ethanol with catalytic DowexÒ resin, containing a sulfonic
acid moiety, in a Soxhlet extractor containing 4 Å molecular sieves
and heated to reflux. The resulting ester 12, was isolated in gram
quantities at high yield and purity (>95%) by removal of the cata-
lyst via filtration and subsequent evaporation of the solvent.
Attempts at further purification by chromatography or extraction
greatly reduced yield with minimal increase in purity.
The ethyl ester (12), was then selectively protected to the
monobenzyl ether of the ethyl ester (13) via a silver mediated
etherification. The Ag2O acts as a Lewis acid to both oxygens of
the two alcohol groups helping to catalyze the reaction without
the need of NaH or another base. Once the first protection occurs,
the activity of the second hydroxyl is sufficiently reduced prevent-
ing a second benzylation. There are a few aspects to consider when
contemplating this method. Firstly, the cost of silver on a large
scale required for gram-scale quantities is not feasible without
reclaiming most of the catalytic silver. This was addressed by iso-
lating the silver via centrifugation of the reaction mixture, pellet-
ing the silver, and allowing for easy isolation of the product in
the supernatant. The isolated silver pellet weighs more than the
original silver used (presumably a combination of the product
Ag4O4 as well as the starting material Ag2O). While it is doubtful
this pellet can be immediately reused, its recovery is important
in limiting the cost of the overall reaction. Once recovered it can
be reduced back to Ag2O according to literature [35]. Unfortu-
nately, the overall yield of this reaction is only about 50% if purified
by LCC, with a notable presence of the unreacted 12 and residual
BnBr in the crude product. This need for LCC might be of concern,
however, the presence of these impurities/byproducts do not affect
the subsequent hydrolysis of 13 and were easily removed post
hydrolysis via simple purification.
Rather serendipitously the ensuing hydrolysis actually reduces
the complexity by converting residual BnBr quantitatively to
BnOH, while the unreacted 12 is hydrolyzed back to 1. BnOH and
ethanol are easily removed from the reaction mixture by organic
extraction at neutral pH, due to the relatively high pKa’s. The AB
monomer 2, can then be isolated via acidification of the aqueous
solution to a pH ꢀ4 followed by organic extraction with no evi-
dence of contamination from residual 1. This is not unexpected
and undoubtedly the result of the miniscule solubility of 1 in chlo-
roform especially when compared to that of water. Ultimately, this
three-step synthesis proved be the most efficient route for produc-
ing gram quantities of 2 at high purity, confirmed by NMR. The
resulting total synthesis has an overall yield of 47%
(0.954 ꢁ ~0.5 ꢁ 0.974 = 0.465) and purification achieved by liq-
uid–liquid extraction. Most notably this pathway requires abso-
lutely no column chromatography for purification. Unfortunately,
the highly efficient (>95% yields) of the esterification to 12 and
hydrolysis (>97% yield) to 2 are hurt by the less efficient mono-
benzylation to 13.
Polymerization to linear oligomers
Once 2 was isolated a number of attempts were made to poly-
merize linear polymers via a variety of activated esterification che-
mistries analogous to those used in the dendronization (Scheme 6).
Utilizing N,N0-dicyclohexylcarbodiimide (DCC) or N-ethyl-N0-(3-
dimethylaminopropyl)carbodiimide (EDC) in the presence of cat-
alytic 4-dimethylaminopyridine (DMAP) were added to a neat
solution of 2 and allowed to react. They were monitored with
MALDI-TOF regularly to assess the DP. Variations to the rate at
which the activating agent was added to 2 were attempted, includ-
ing the reaction either by single addition, multiple aliquoted addi-
tions or slow drop-wise addition of the activating agent. This first
direct A – B (carboxylic acid/alcohol) approach yielded relatively
small oligomers and cyclics which is somewhat expected from this
type of step growth condensation polymerization.
Additional, polymerizations were attempted utilizing an initia-
tor to preventing the cyclization of small oligomers by producing a
polymer with a single reactive end. This was performed by the
incremental addition of 2 to a solution of activating agent contain-
ing of a small amount ethanol or 13 as an ‘‘initiator” for the poly-
merization, in an effort to prevent unwanted intramolecular
cyclization and thus yield higher Mn polymers. Unfortunately,
these attempts also yielded predominately small oligomers with
all samples displaying multiple polymers initiated from water,
ethanol, and DMAP as determine by matrix assisted laser desorp-
tion/ionization time-of-flight mass spectrometry (MALDI-TOF
MS) (supplemental Fig. S2). It is clearly evident from this data
and confirmed by GPC (supplemental Fig. S3) that the vast majority
of the sample is oligomers of <1 K Da. The cyclics and smaller oli-
gomers were removed via GPC fractionation, in order to isolate lar-
ger oligomers initiated from water, DMAP and ethanol (Fig. 2). The
MALDI shown in Fig. 2 was obtained from the 18.0 to 18.5 min
retention time fraction of the crude linear bis-MPA sample yielding
the orange GPC trace (supplemental Fig. S3). The linearity and low
DP was confirmed by MALDI-TOF m/z analysis, with the three sam-
ples averaging to a DP of 10, with an Mn of 2070 and Mw 2130. This
is in agreement with the GPC data (supplemental Fig. S3), with the
Scheme 4. Direct but nonselective Williamson benzyl etherification of bis-MPA. i)
BnBr (2 eq), NaH (2 eq.), THF, 2 h. ii) NaOH (0.5 M), overnight.