Minimizing solvent waste in catalytic reactions in highly recyclable hydrocarbon solvents

Article abstract of DOI:10.1039/d0ob00734j

This paper describes chemistry using organocatalysts in hydrocarbon solvents that minimizes solvent waste by using inexpensive, non-volatile, relatively inflammable, and easily recyclable poly(α-olefin)s (PAOs) as hydrocarbon solvents. These studies show that when substrates have limited solubility in PAO solvents, this issue can be addressed by adding a small amount of a cosolvent. Kinetic studies were also carried out and show that reactions carried out in PAOs are kinetically comparable to reactions in conventional non-recyclable hydrocarbon solvents. A range of strategies that separate and isolate products from reactions in PAOs using a polyisobutylene (PIB)-supported DMAP catalyst have been studied using four different catalytic reactions. In the most general procedure, the PAO phase containing a PIB-bound catalyst is separated from products by low energy liquid/liquid gravity separation. This can be accomplished using a minimal amount of a polar solvent. In another example, the product's low solubility leads to it precipitating during the reaction. In this case, a simple filtration recycles the PAO and a PIB-bound DMAP catalyst. We have demonstrated that the PAO phase containing a PIB bound DMAP catalyst can be recycled for at least 10 cycles without loss of activity. Our studies further showed that leaching of the PAO phase into polar solvents was orders of magnitude less than conventional hydrocarbon solvents such as heptane. The result is that the overall solvent waste generation is lower than for the same reaction carried out in conventional solvents.

Full text of DOI:10.1039/d0ob00734j

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Organic &
Biomolecular Chemistry
View Article Online
View Journal
PAPER
Minimizing solvent waste in catalytic reactions in
highly recyclable hydrocarbon solvents†
Cite this: DOI: 10.1039/d0ob00734j
Sopida Thavornpradit,
James M. Killough
and David E. Bergbreiter
*
This paper describes chemistry using organocatalysts in hydrocarbon solvents that minimizes solvent
waste by using inexpensive, non-volatile, relatively inflammable, and easily recyclable poly(α-olefin)s
(PAOs) as hydrocarbon solvents. These studies show that when substrates have limited solubility in PAO
solvents, this issue can be addressed by adding a small amount of a cosolvent. Kinetic studies were also
carried out and show that reactions carried out in PAOs are kinetically comparable to reactions in conven-
tional non-recyclable hydrocarbon solvents. A range of strategies that separate and isolate products from
reactions in PAOs using a polyisobutylene (PIB)-supported DMAP catalyst have been studied using four
different catalytic reactions. In the most general procedure, the PAO phase containing a PIB-bound cata-
lyst is separated from products by low energy liquid/liquid gravity separation. This can be accomplished
using a minimal amount of a polar solvent. In another example, the product’s low solubility leads to it pre-
cipitating during the reaction. In this case, a simple filtration recycles the PAO and a PIB-bound DMAP
catalyst. We have demonstrated that the PAO phase containing a PIB bound DMAP catalyst can be
recycled for at least 10 cycles without loss of activity. Our studies further showed that leaching of the PAO
phase into polar solvents was orders of magnitude less than conventional hydrocarbon solvents such as
heptane. The result is that the overall solvent waste generation is lower than for the same reaction carried
out in conventional solvents.
Received 8th April 2020,
Accepted 14th May 2020
DOI: 10.1039/d0ob00734j
rsc.li/obc
any net greenhouse gas generation (GHG) though there are
Introduction
GHG costs associated with these solvents’ syntheses and use.8
Solvents typically constitute the major mass fraction of any syn- Our interests in solvents are a bit different and are focused on
thetic or homogeneously catalyzed reaction and are a major developing solvents or solvent systems that are highly recyclable.
component of most synthetic processes. It is thus not surprising Highly recyclable solvents or solvent systems would in principle
that there is great interest in developing greener and more sus- minimize the need to produce more chemicals and could be a
tainable alternatives for legacy solvents in synthesis and step in what has been described as Moore’s Law for Green
catalysis.1–3 However, there is an ongoing need for new schemes Chemistry.9 Recycling would also in effect amortize GHG costs
and new media that can address this problem. Recently, con- associated with solvent synthesis.
siderable success has accrued to schemes that use surfactants
We have suggested that poly(α-olefin)s (PAOs) can be
enabling organic chemistry in aqueous media, and to schemes alternatives to hydrocarbons like heptane and hexane. PAOs
that use less conventional solvents like deep eutectic solvents are prepared on large scale from α-olefins like 1-decene and
and supercritical liquids like scCO2.4–6 Many new sustainable 1-dodecene.10–12 These α-olefins are most economically
solvents have been developed and evaluated as alternatives to formed from petrochemical ethylene though they could be bio-
conventional solvents in other chemistry. Resources like the based since they can also be prepared by renewable starting
GSK tables of sustainable solvents provide chemists with a list materials. Our initial studies looked at relatively viscous PAOs
of such solvents.7 However, most of these solvents are volatile solvents that are mixtures of structurally and diastereomeric
organic compounds that are not always easily recyclable. C40, C50, and C60 hydrocarbons formed by oligomerization of
Nonetheless, since these solvents are derived from biological 1-decene.13 Subsequent studies showed that lower viscosity
material, their combustion or degradation does not result in C20, C24, C30, and C36 PAO fractions were comparably recyclable
and that these PAO fractions can be used in SN2 chemistry, re-
cycling a hydrocarbon phase-transfer catalyst more efficiently
Department of Chemistry, Texas A&M University, P.O. Box 3012, College Station,
than heptane.14 We also showed that the flash-point of these
Texas, 77842-3012, USA. E-mail: bergbreiter@tamu.edu
PAOs allowed us to substitute them for hexane or heptane to
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
make ‘safer’ alkyllithium solutions and that products from
d0ob00734j
This journal is © The Royal Society of Chemistry 2020
Org. Biomol. Chem.

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
View Article Online
Paper
Organic & Biomolecular Chemistry
reaction mixtures containing these solvents could be isolated
without significant PAO contamination.15 We have since
shown that PAO and PAO solvent systems are useful extractants
for removing trace organics from water.16 These results
suggested to us that PAOs could also be effective in other
chemistry. The studies below explore these recyclable hydro-
carbon solvents in a series of reactions using a recyclable
analog of the nucleophilic catalyst 4-(dimethylamino)pyridine
(DMAP) and describe various strategies that minimize solvent
waste in these sorts of organocatalytic reactions.
Results and discussion
Poly(α-olefin)s (PAOs) are inexpensive hydrocarbons that are
commercially available on large scale.10–12 While their indus-
trial importance derives from their widespread use as base oils
in industrial and automotive lubricants, we have proposed that
they could have a role in green chemistry as safer recyclable
solvents. In two preliminary studies, we showed that PAOs
Scheme 1 Synthesis of a PIB-bound DMAP catalyst 5.
could be used in place of heptane in phase-transfer catalyzed Deprotection of 3 was carried out using trifluoroacetic acid to
SN2 chemistry and that PAOs in place of volatile alkanes make form the piperazine-terminated PIB oligomer (4). Finally, the
alkyllithium reagents ‘safer’.14,15
oligomer PIB 4 was allowed to react with 4-chloropyridine in a
To evaluate PAOs in a broader setting, we here explore their nucleophilic aromatic substitution reaction to form the
use in organocatalytic reactions showing how PAOs and a cata- desired DMAP-terminated PIB oligomeric catalyst (5).
lyst anchored in PAO can be recycled and separated from pro-
We envisioned two scenarios that could demonstrate the
ducts using a recyclable analog of the nucleophilic organo- utility of a PAO anchorable catalyst like 5 in a highly recyclable
catalyst 4-(dimethylamino)pyridine (DMAP) as an example. We PAO solvent system. In the first, the starting material would be
also kinetically compared the reactivity of this catalyst in soluble but the product would be insoluble in this hydro-
heptane and PAO and to reactivity in PAO containing various carbon solvent mixture. In this case, we would carry out a reac-
cosolvents.
To demonstrate the utility of PAO in catalytic reactions, we Recycling would only require the addition of fresh substrates.
chose to prepare a PAO soluble polyisobutylene (PIB)-bound
A second scenario would involve a situation in which the
tion and then separate the solid product from the solution.
version of 4-(di-methylamino)pyridine (DMAP), 5, a commonly starting materials and product would both be soluble in a PAO
used nucleophilic catalyst that has been previously attached to solvent system. In this case, we would isolate the product by
many other soluble and insoluble polymer supports.17–23 4- using a liquid/liquid biphasic separation. Based on our past
(Dimethylamino)pyridine was chosen for this study because work, we would have to form a biphasic system using either a
DMAP is an efficient catalyst for a variety of reactions24,25 and perturbant or by using a PAO-immiscible extraction solvent to
would allow us to study a variety of substrates and products isolate the product. This scenario would also require that any
that could have varying solubility in PAO. Vinyl-terminated PIB workup scheme would have to separate the product from the
was selected as a phase anchor for this organocatalyst because PAO and the catalyst 5 with minimal loss of the PAO or 5.
our prior studies showed that PAO was very effective both as a Ideally, this separation would minimize the amount of the
solvent for PIB-bound azo dyes and as a way to separate these required cosolvents.
PIB-bound azo dyes from polar organic solvents.13 This phase
selective solubility would ensure high levels of catalyst recovery using a Knoevenagel condensation of 4-chlorobenzaldehyde,
in recycling studies.
and ethyl cyanoacetate in PAO506, a fully hydrogenated C36
The recyclability of PAO and catalyst 5 was first studied
The modified PIB oligomer DMAP-terminated catalyst (5) trimer of 1-dodecene. This reaction (eqn (1)) used a 10 mol%
was synthesized starting with commercially available vinyl-ter- loading of 5
minated polyisobutylene (PIB) that had an Mn value of 1000
Da.26 This synthesis (Scheme 1) first formed a hydroxy-termi-
nated PIB oligomer (1). Compound 1 was next converted to an
ð1Þ
iodo-terminated PIB oligomer (2) by the Appel reaction. Then
an excess quantity of Boc ((CH3)3COC(vO)–) protected pipera- and was carried out for 6 h at 90 °C using 2 M toluene in
zine was dissolved in DMF and added to a heptane solution of PAO506 as a recyclable solvent system. Toluene was added so
compound 2 under thermomorphic conditions to convert 2 that the substrates would be soluble at a 0.3 M concentration.
into the Boc-piperazine-terminated PIB oligomer (3). In this example, cooling the reaction mixture to room tempera-
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2020

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
View Article Online
Organic & Biomolecular Chemistry
Paper
complete, 3 mL of acetonitrile was used to extract the product
from PAO506 phase. The PAO506 phase containing 5 was then
reused by adding fresh substrates. Ten cycles were carried out.
The average percent yield per cycle after 10 cycles was 95%.
While substrates can be highly soluble in PAO and while
product precipitation has precedent in other catalytic reactions
we have studied using heptane soluble PIB-bound
catalysts,29,30 neither case is a general phenomenon in our
experience. A more likely situation would be one where a cosol-
vent is needed to insure that the substrates are soluble in a
hydrocarbon like a PAO. In those cases, the product or pro-
ducts too are likely to be soluble and practical schemes would
be needed to separate the PAO and the PAO phase-anchored
catalyst from the product when using a PAO solvent system.
Most of the catalytic reactions we have studied using recycl-
able PIB-bound catalysts also used cosolvents. In those cases,
we typically used an equivolume mixture of heptane and a
polar solvent. A liquid/liquid phase separation separated the
heptane solution of a PIB-bound catalyst from the polar
organic solvent phase after a temperature change or pertur-
bant addition induced biphase formation.31–36 However, while
PIB-bound catalysts or other alkane soluble polymeric catalysts
typically had >99.9% phase selective solubility in those separ-
ations and could be separated and recycled, some heptane par-
titions into the polar organic solvent during the liquid/liquid
separation. The ca. 5–15% of the heptane that partitions into
the polar organic solvent phase is typically removed along with
the polar solvent at reduced pressure during product isolation.
Thus, while the heptane is only partially recyclable, the pro-
ducts are not contaminated by heptane. That prior work also
did not recycle other solvents either.
Our goal in this work was to use the low volatility and
phase separability of PAO to maximize the recyclability of a
hydrocarbon solvent and to minimize the use of other volatile
solvents. However, while 5–15% heptane contamination of the
products was not an issue in our earlier work, even small
losses of 1–2% of the PAO solvent would pose a problem as the
nonvolatile PAO would not be as removable as heptane. To
avoid the requirement for an additional purification step, we
expanded on a recent study of alkyllithium/PAO reactivity that
isolated products with minimal PAO contamination15 to
develop procedures that could usefully separate PAO, a PAO
phase-anchored catalyst, and a cosolvent from products with
minimal PAO contamination of the product and product
solution.
Fig. 1 Comparative plot for the reaction of 4-chlorobenzaldehyde and
cyanoacetate using DMAP organocatalyst 5 in PAO506 in a Knoevenagel
reaction forming a cyanoacrylate product.
ture led to self-separation of the product as a solid. This solid
product was separated from the PAO506 and PAO506-soluble
catalyst by centrifugation and decantation. Fresh substrates
were then added to the PAO506/toluene phase containing 5 to
effect a second cycle. This process of reaction/separation was
repeated for 10 cycles with no significant decrease in volume
of the PAO506 phase. Conversions in each cycle were analyzed
by 1H NMR spectroscopy and were 98%. While the product
self-separated, the precipitated product did contain traces of
PAO506. Thus, the combined precipitate from these 10 cycles
was triturated twice with 10 mL of heptane. The product so iso-
lated was characterized using 1H and 13C NMR spectroscopy
and had a melting point consistent with that in the litera-
ture.27 The average isolated yield of the isolated product for
these 10 cycles was 85%.
This experiment also allowed us to compare the activity of
catalyst 5 to a conventional DMAP catalyst in this same 2 M
toluene/PAO506 solvent mixture (Fig. 1). This comparison used a
5 mol% loading of 5 or DMAP and monitored the formation of
cyanoacrylate by 1H NMR spectroscopy over a 10 h reaction. The
results in Fig. 1 show that 5 is comparable in activity to DMAP.
To further explore the utility and recyclability of PAO sol-
vents in reactions catalyzed by 5, we examined the addition of
trimethylsilyl cyanide (TMSCN) to aldehydes in PAO506. DMAP
is one example of a commonly used catalyst for this reaction.28
The starting materials TMSCN and benzaldehyde were soluble
in PAO so a cosolvent was not required. This reaction was
carried out using 10 mol% of catalyst 5 in 10 mL of PAO506 for
2 h at ambient temperature using 0.36 M TMSCN and 0.3 M
benzaldehyde (eqn (2)). At this point, the conversion of benz-
aldehyde to products as measured by 1H NMR spectroscopy
was 100%. However, unlike the example in eqn (1), the pro-
ducts remained in solution. To isolate the product after the
reaction was
We studied two schemes to separate PAO from a PAO misci-
ble cosolvent. The first (Scheme 2A) involved a liquid/liquid
biphasic separation where we added an unfunctionalized PEG
oligomer as a perturbing cosolvent. We chose PEG for two
reasons. First, PEG (poly(ethylene glycol)) oligomers are one of
the few widely used polymeric solvents and PEG is relatively in-
expensive and nontoxic.37,38 Second, poly(alkene oxide)s and
polyolefins are immiscible. Second (Scheme 2B), we examined
the use of a PAO-immiscible extraction solvent – acetonitrile.
The idea was that a small amount of acetonitrile or acetonitrile
with 5–10 vol% water would allow us to quantitatively recycle
ð2Þ
This journal is © The Royal Society of Chemistry 2020
Org. Biomol. Chem.

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
View Article Online
Paper
Organic & Biomolecular Chemistry
PIB1000-bound catalyst like 5 too could be recyclable with the
PAO506 solvent.
To explore the viability of using PEG as a perturbing solvent
in recycling PAO506 we studied Boc protection of an amine and
a catalytic Boc protection of a phenol using 5. In the first of
these experiments, we examined the reaction of piperidine
with Boc anhydride in a reaction that was carried out at room
temperature for 2 h using 10 mL of an equivolume mixture of
PAO506 and THF or 2-methyltetrahydrofuran (MeTHF) (eqn
(3)). Then the miscible
ð3Þ
Scheme 2 Schemes A (using a PEG perturbant) and B (using aceto-
nitrile as a immiscible polar solvent) for liquid/liquid separation to
recycle a PAO phase from a product phase.
solution was perturbed by the addition of 5 mL of PEG400. The
product yields in the PEG400 phases were determined by 1H
NMR spectroscopy. Up to seven cycles could be carried out.
The results showed that the PAO506 was fully recyclable
through 7 cycles with THF as a cosolvents and through 4 cycles
with MeTHF. The product yields averaged 98% in THF and
76% in MeTHF. We believe the slightly higher yield seen in
THF versus MeTHF results from the fact that the recycled
PAO506 phase contains more MeTHF than THF and that some
product remains in the PAO506 phase.
the PAO phase. Our prior work used PAO solvents as safer sol-
vents for reactive alkyllithium reagents and we were able to
isolate products without PAO contamination using an aceto-
nitrile extraction step after removal of all the volatile solvents
at reduced pressure.15 However, those experiments typically
only used a small amount of PAO and there was no effort to
either recycle the PAO or to otherwise reduce the larger
amounts of conventional solvents that were not recycled.
The experiments below address both those issues. To con-
sider what more general schemes for PAO solvent and catalyst
recycling might work, we first examined versions of these strat-
egies looking at the separability of PAO506, a cosolvent like
toluene or tetrahydrofuran (THF), and a PIB-bound dye (as a
PIB-bound catalyst surrogate). Once a separation strategy was
established, we explored its practicality in a PIB-bound DMAP
catalyzed reaction.
The initial experiments examining PAO immiscible pertur-
bants that could be added to a mixture of PAO506 and toluene
or THF to induce biphase formation. Unfortunately, addition
of perturbants like sucrose, diethylene glycol, acetone, or ethyl-
ene diamine did not lead to phase separation of an equivo-
lume mixture of PAO506 and toluene or THF. However, a
biphase did form using PEG400 (H(OCH2CH2)8OH) or PEG2000
(H(OCH2CH2)45OH) as perturbants. In these experiments,
addition of 5 mL of PEG400 or 5 g of PEG2000 to an equivolume
mixture of PAO506 and toluene or THF led to a biphasic
mixture where the ca. 5 mL of a PAO506 phase separated. When
this experiment was carried out with a PAO506/toluene or
a PAO506/THF solution containing a PIB1000-bound dye 6
(ca. 10−3 M in dye), the perturbed reaction
We next explored a similar Boc protection reaction in
PAO506 using 5 as a catalyst with 2,6-dimethylphenol as a sub-
strate (eqn (4)).
ð4Þ
These studies involved using 10 mL of an equivolume
mixture of PAO506 and MeTHF or toluene and were carried out
1
for 3 h at room temperature using 1 mol% 5 as a catalyst. H
NMR spectroscopy indicated that the conversion of the starting
phenol was quantitative. In these reactions the product was
isolated by combining the PEG400-rich phases from 5 cycles
(ca. 7.6 mL per cycle). Water (100 mL) was added and the
resulting aqueous phase was extracted 3 times with either
diethyl ether if MeTHF was the reaction solvent or toluene if
toluene was used as the reaction solvent. The organic phase
was then extracted with once more with water and the organic
solvent was removed at reduced pressure. The average yield of
product for Boc esterifications catalysed by 5 in toluene was
86% and 87% for esterifications carried out in MeTHF.
While these initial experiments showed that we could
recycle PAO506, they used a large amount of PEG400 and other
conventional solvents. Thus, we briefly explored schemes
using recyclable PEG2000 in place of PEG400 (Scheme 3). This
experiment also used 5 as a catalyst in Boc protection of 2,6-di-
methylphenol. It recycled both the PAO506 and the PEG2000
.
mixture separated such that visually all of the dye was in the The reaction was carried out using 5 mL of PAO506 and 5 mL of
1
PAO506 phase. This qualitative observation suggested that a toluene. The reaction was complete in 2 h based on H NMR
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2020