Polymerisable di- and triesters from Tall Oil Fatty Acids and related compounds

Article abstract of DOI:10.1039/c3gc37071b

Tall Oil Fatty Acids, a low value side product from the paper industry containing mainly oleic and linoleic acids, are used for producing the polyester precursor, dimethyl 1,19-nonadecanedioate by methoxycarbonylation in the presence of [Pd₂(dba)₃], 1,2- bis(ditertiarybutylphosphinomethyl)benzene and methanesulfonic acid in methanol. The methoxycarbonylation of methyl linoleate has been used to identify other products formed and approaches to their minimisation have been developed. It has also been used for the production of trimethyl heptadecanetricarboxylates. Finally, conjugated unsaturated esters of different chain length (up to 16 C atoms), some of them available from plant oils, are subjected to methoxycarbonylation to give α,ω-diesters.

Full text of DOI:10.1039/c3gc37071b

Green Chemistry
View Article Online
View Journal
PAPER
Polymerisable di- and triesters from Tall Oil Fatty Acids
and related compounds†
Cite this: DOI: 10.1039/c3gc37071b
Marc R. L. Furst, Thomas Seidensticker and David J. Cole-Hamilton*
Tall Oil Fatty Acids, a low value side product from the paper industry containing mainly oleic and linoleic
acids, are used for producing the polyester precursor, dimethyl 1,19-nonadecanedioate by methoxycarbo-
nylation in the presence of [Pd2(dba)3], 1,2-bis(ditertiarybutylphosphinomethyl)benzene and methane-
sulfonic acid in methanol. The methoxycarbonylation of methyl linoleate has been used to identify other
products formed and approaches to their minimisation have been developed. It has also been used for
the production of trimethyl heptadecanetricarboxylates. Finally, conjugated unsaturated esters of
different chain length (up to 16 C atoms), some of them available from plant oils, are subjected to meth-
oxycarbonylation to give α,ω-diesters.
Received 19th December 2012,
Accepted 21st February 2013
DOI: 10.1039/c3gc37071b
www.rsc.org/greenchem
per year), which is rich in Tall Oil Fatty Acids (TOFA, mainly
oleic and linoleic acids).7 TOFA have traditionally been used in
Introduction
As petroleum feedstocks dwindle and prices increase, the need inks, rubbers or adhesives.8 In this paper, we report our
for renewable resources for producing the polymers that are studies on the synthesis of dimethyl 1,19-nonadecanedioate
ubiquitous in enhancing our lives is becoming more and more using TOFA as the feedstock. Furthermore, suspecting a possi-
obvious. Ideally, from an ecological and economic point of ble transfer hydrogenation from methanol to the unsaturated
view, the monomers required for such polymers would be dimethyl 1,19-nonadecenedioate (isomers) to form the saturated
derived from non-toxic wastes or side products from plant equivalent,2 we report our studies on this phenomenon using
based feedstocks. We1,2 and others3–6 have previously reported pure linoleic acid. We also discuss the possible formation of
the synthesis of a new long carbon chain polyester precursor, triesters and other products from methoxycarbonylation of
dimethyl 1,19-nonadecanedioate and its hydrogenation to TOFA and of linoleic acid. 2-Undecenoic acid, for which the
1,19-nonadecanediol.1,4–6 Mecking and coworkers have shown methyl ester is available from metathesis of methyl oleate with
that these precursors can be copolymerised to aliphatic poly- dimethyl maleate,9 undergoes isomerising methoxycarbonyla-
esters with properties very similar to those of polyethylene.3–5 tion to the Nylon-12,12 precursor, dimethyl 1,12-dodecane-
Our initial synthesis of dimethyl 1,19-nonadecanedioate dioate, which we also make from methyl 10-undecenoate,
employed purified methyl oleate as the feedstock2 but we sub- available from cracking castor oil.10 We show that the double
sequently showed that unpurified olive, rapeseed and sun- bond can be selectively moved 13 steps along a chain to form
flower oils, could provide much cheaper alternatives to dimethyl 1,17-heptadecanedioate from 2-hexadecenoic acid.
purified methyl oleate.1 Walther et al. employed high oleic
sunflower oil.6 All of these oils are used for food, so any
process that starts from them would be in competition with
food crops, a situation we are keen to avoid.
Experimental
The Kraft process for the production of pine wood pulp,
produces a waste side product known as Tall Oil (2 M tonnes
All reactions were performed using standard Schlenk tech-
niques. All solvents were degassed with nitrogen. Unless other-
wise stated, solvents were used as supplied and were not
previously dried. 1H and 13C nuclear magnetic resonance
(NMR) spectra were recorded at 298 K on a Bruker 300 MHz for
1H NMR and 75 MHz for 13C NMR spectrometer, using the
residual solvent peak to reference the spectra to tetramethyl-
silane at δ = 0 ppm. Elemental analyses were performed by the
Elemental Analysis Service of the London Metropolitan Univer-
sity. [Pd2(dba)3], linoleic acid, dimethyl maleate (Sigma
EaStCHEM, School of Chemistry, University of St Andrews, St Andrews, KY16 9ST
Scotland, UK. E-mail: djc@st-andrews.ac.uk; Fax: +44 (0) 1334 463 808;
Tel: +44 (0) 1334 463 805
†Electronic supplementary information (ESI) available: 1H-NMR and 13C-NMR
spectra of dimethyl 1,19-nonadecanedioate, trimethyl heptadecane-1,n,17-
trioate, dimethyl 1,12-dodecanedioate, dimethyl 1,17-heptadecanedicarboxylate,
trimethyl ethane-1,1,2-trioate, 1H-NMR of TOFA and linoleic acid, GCMS of
crude mixtures obtained from methoxycarbonylation reactions. See DOI:
10.1039/c3gc37071b
Aldrich),
1,2-bis(ditertiarybutylphosphinomethyl)benzene
This journal is © The Royal Society of Chemistry 2013
Green Chem.

View Article Online
Paper
Green Chemistry
(Lucite International), methyl 10-undecenoate, 2-undecenoic [Pd2(dba)3] (37 mg, 0.04 mmol, 0.08 mmol Pd), 1,2-bis(diter-
acid, 2-hexadecenoic acid (Tokyo Chemical Industry), TOFA tiarybutylphosphinomethyl)benzene (158 mg, 0.4 mmol,
(Mistral Lab Chemicals) and methanesulfonic acid (Alfa Aesar) DTBPMB) and magnesium sulfate (200 mg), dry methanol
were used as supplied.
(10 mL) and TOFA (2.5 mL) were introduced via cannula. The
solution was stirred under nitrogen for 10 min and introduced
in the autoclave by cannula. The autoclave was closed and
Dimethyl 1,19-nonadecanedioate (4) from TOFA
Method 1: Under air, [Pd2(dba)3] (37 mg, 0.04 mmol, pressurised with CO (30 bar) and heated to 90 °C with mag-
0.08 mmol Pd) and 1,2-bis(ditertiarybutylphosphinomethyl)- netic stirring for 64 h. After cooling and depressurisation of
benzene (158 mg, 0.4 mmol, DTBPMB) were introduced into a CO, dichloromethane was added and the crude mixture was
hastelloy autoclave, sealed and purged with nitrogen. Metha- introduced into a round bottom flask. The solvent was
nol (10 mL), TOFA (2.5 mL) and methanesulfonic acid removed on a rotary evaporator. The crude product was passed
(0.05 mL, 0.8 mmol, MSA) were added to the autoclave by through a silica chromatography column (hexane/ethyl acetate:
cannula. The autoclave was purged three times with CO and 92/8). The desired product was obtained as a transparent oil,
the pressure was set to 30 bar. The autoclave was heated to 1.38 g, 44%). Elemental analysis: found C 66.52, H 10.07%;
90 °C for 48 h. It was then cooled to room temperature and C23H42O6 requires C 66.63, H 10.21%. 1H NMR (300 MHz;
degassed. A sample was taken for GC analysis. Method 2: An CDCl3): (mixture of isomers)
esterification of TOFA (2.5 mL) was carried out over 3 h in dry (branched)), 3.66 (s, 6H, CH3O-), 2.29 (t,
δ
=
3.67 (s, 3H, CH3O-
7.5 Hz,
J
=
refluxing methanol (10 mL) in the presence of MSA (0.05 mL) CH3O-CO-CH2-), 1.62 (s, CH3 O-CO-CH2-CH2-), 1.44 (m), 1.23
and magnesium sulfate (200 mg) under an atmosphere of dry (s, CH2 chain). 13C NMR (75 MHz; CDCl3): (mixture of isomers)
nitrogen. The solution was cooled and introduced into the δ = 51.63 (s, CH3O-), 45.68–45.53 (s, CH3O-CO-CH-(CH2-CH2-
degassed autoclave containing [Pd2(dba)3] (37 mg, 0.04 mmol, )2), 34.29–34.05 (s, CH-(CH2-CH2-)2, 32.70 (s, CH3OCO-CH2-),
0.08 mmol Pd) and DTBPMB (158 mg, 0.4 mmol) by cannula 31.93 (s, CH-(CH2-CH2-)2, 29.79–25.13 (CH2 chain), 23.01
bearing a filter. The autoclave was heated to 90 °C for 10 h, (s, CH3 O-CO-CH2-CH2-).
then cooled to room temperature, degassed and a sample was
taken for GC analysis. For both Methods 1 and 2, a suspension
Methyl n-oxooctadecanoate (1 and 2) from linoleic acid
of Pd/C 5% (170 mg, 0.08 mmol Pd) in methanol (10 mL) was Method 1. The carbonylation was carried out as described in
added by cannula. The autoclave was purged three times with Method 1 for TOFA, above but using linoleic acid in place of
H2 and the H2 pressure was set to 30 bar. The temperature was TOFA. After depressurisation of carbon monoxide, dichloro-
set to 90 °C for 20 h. After cooling, venting and opening, the methane was added and the crude mixture was introduced
remaining yellow powder was dissolved by addition of dichloro- into a round bottom flask. The solvent was removed on a
methane (20 mL) and the yellow solution was filtered through rotary evaporator. The crude product was passed through a
paper. The solvent was removed on a rotary evaporator. Metha- silica chromatography column (hexane/ethyl acetate: 92/8).
nol was added and the mixture was warmed until the products The desired product was obtained as a white solid (654 mg,
dissolved. The mixture was placed in the freezer for 30 min 25%, mixture of isomers). This fraction was then passed down
and filtered using a Büchner funnel. The desired product was a silica chromatography column using hexane/ethyl acetate
obtained as a white powder (Method 1: 0.92 g, 32%; Method 2: (90/10) and fractions containing only methyl 17-oxoocta-
1.33 g, 49%). Elemental analysis: found C 70.81, H 11.35%; decanoate (2) were collected and evaporated to dryness (yield
C21H40O4 requires C 70.74, H 11.31%. 1H NMR (300 MHz; 3%). Elemental analysis: found C 72.90, H 11.60%; C19H36O3
CDCl3): δ = 3.65 (s, 6H, CH3O-), 2.28 (t, J = 7.5 Hz, 4H, requires C 73.03, H 11.61%. 1H NMR (300 MHz, CDCl3): δ =
CH3O-COCH2-), 1.60 (quintet, J = 7.3 Hz, 4H, CH3O-CO-CH2- 3.66 (s, 3H, CH3O-), 2.41 (t, J = 7.5 Hz, 2H, CH3-CO-CH2-), 2.29
CH2-), 1.23 (s, 26H, CH2 chain). 13C NMR (75 MHz; CDCl3): δ = (t, J = 7.5, 2H, CH3O-CO-CH2-), 2.12 (s, 3H, CH3-CO-CH2-),
174.51 (s, CO), 51.60 (s, CH3O–), 34.28 (s, CH3O-CO-CH2-), 1.65–1.51 (m, 4H, -CO-CH2-CH2-), 1.24 (s, 22H, CH2 chain). 13
C
29.84–29.32 (s, CH2 chain), 25.13 (s, CH3O-CO-CH2-CH2-). NMR (75 MHz; CDCl3): δ = 209.53 (s, CO (ketone)), 174.48 (s,
These data are consistent with those from the literature.1
CO (ester)), 51.57 (s, CH3O-), 43.97 (s, CH3-CO-CH2-), 34.26 (s,
CH3O-CO-CH2-), 29.98 (s, CH3-CO-CH2-), 29.77–29.29 (s, CH2
chain), 25.10 (s, CH3O-CO-CH2-CH2-), 24.02 (s, CH3O-CO-CH2-
CH2-). These data are consistent with those in the literature.11
Method 2: The reaction was carried out as described in
Trimethyl heptadecane-1,n,17-tricarboxylate (6) from linoleic
acid
Method 1: The carbonylation was carried out as described in
Method 1, above but using linoleic acid in place of TOFA, the Method 1 for TOFA, above but using methyl linoleate in place
reaction time being 64 h. After depressurisation of carbon of TOFA, and nitrogen (40 bar) instead of carbon monoxide.
monoxide, dichloromethane was added and the crude mixture After depressurisation of nitrogen, dichloromethane was
was introduced into a round bottom flask. The solvent was added and the crude mixture was introduced into a round
removed on a rotary evaporator. The crude product was passed bottom flask. The solvent was removed on a rotary evaporator.
through a silica chromatography column (hexane/ethyl acetate: The crude product was passed through a silica chromato-
92/8). The desired product was obtained as a transparent oil graphy column (hexane/ethyl acetate: 92/8). The desired
(635 mg, 19%). Method 2: In a Schlenk flask containing product was obtained as a white solid (1004 mg, 42%, isomers
Green Chem.
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
(1) without methyl 17-oxooctadecanoate, (2). Elemental analy- 29.78–29.55 (s, CH2 chain), 25.37 (s, CH3O-CO-CH2-CH2-).
sis: found C 73.13, H 11.70%; C19H36O3 requires C 73.03%, H These data are consistent with these from the literature.12
11.61%. 1H NMR (300 MHz; CDCl3): δ = 3.61 (s, 3H, CH3O-),
3
4
2.33 (tt, J = 7.5 Hz, J = 2.6 Hz, 4H, CH2-CO-CH2-), 2.27 (t, J =
7.5, 2H, CH3O-CO-CH2-), 1.60–1.44 (m, 6H, -CO-CH2-CH2-),
1.21 (s, 18H, CH2 chain), 0.88–0.80 (m, 3H, -CH2-CH3). 13C
NMR (75 MHz; CDCl3): δ = 211.54 (s, CO (ketone)), 174.24 (s,
CO (ester)), 51.38 (s, CH3O-), 42.79–42.49 (-CH2-CO-CH2-),
34.07 (CH3O-CO-CH2-), 31.82–31.44 (s, CH3-CH2-CH2-),
29.66–28.93 (s, CH2 chain), 24.94 (s, CH3O-CO-CH2-CH2-),
23.85 (s, -CH2-CH2-CO-CH2-CH2-), 22.64–22.37 (s, CH3-CH2-),
14.08–13.86 (s, CH3-CH2-). These data are consistent with
those in the literature.11
Dimethyl 1,17-heptadecanedioate from 2-hexadecenoic acid
Under air, [Pd2(dba)3] (18 mg, 0.0197 mmol, 0.0394 mmol Pd)
and DTBPMB (78 mg, 0.197 mmol) were introduced into a
hastelloy autoclave, sealed and purged with nitrogen. Degassed
methanol (8 mL), 2-hexadecenoic acid (1.0 g, 3.93 mmol) and
methanesulfonic acid (0.05 mL, 0.8 mmol) were added to the
autoclave by cannula. The autoclave was purged three times
with CO and the pressure was set to 30 bar. The autoclave was
heated to 90 °C for 20 h. After cooling, venting and opening,
the yellow solution was filtered through paper and the solvent
was removed on a rotary evaporator. Methanol was added, the
mixture was put in the freezer for 20 min and filtered using a
Büchner funnel. The desired product was obtained as a white
powder (0.75 g, 58%). Elemental analysis: found C 69.56, H
10.09%; C14H26O4 requires C 69.47, H 11.05%. 1H NMR
(300 MHz; CDCl3): δ 3.66 (s, 6H, CH3O-), 2.30 (t, J = 7.5 Hz, 4H,
CH3O-CO-CH2-), 1.61 (quintet, J = 7.4 Hz, 4H, CH3O-CO-CH2-
CH2-), 1.24 (broad s, 22H, CH2 chain). 13C NMR (75 MHz;
CDCl3): δ 174.64 (s, CO), 51.73 (s, CH3O-), 34.40 (s, CH3O-CO-
CH2-), 29.92–29.44 (s, CH2 chain), 25.24 (s, CH3O-CO-CH2-CH2-).
Dimethyl 1,12-dodecanedioate from 2-undecenoic acid
Under air, [Pd2(dba)3] (25 mg, 0.027 mmol, 0.054 mmol Pd)
and DTBPMB (107 mg, 0.27 mmol) were introduced into a
hastelloy autoclave, sealed and purged with nitrogen. Degassed
methanol (10 mL), 2-undecenoic acid (1 mL, 5.43 mmol) and
methanesulfonic acid (0.07 mL, 1.08 mmol) were added to the
autoclave by cannula. The autoclave was purged three times
with CO and the pressure was set to 30 bar. The autoclave was
heated to 90 °C for 44 h. After cooling, venting and opening,
the yellow solution was filtered through paper and the solvent
was removed on a rotary evaporator. Methanol was added and
the mixture was then stirred at −78 °C for 10 min and filtered
using a Büchner funnel. The desired product was obtained as
a white powder. Isolated yield (0.6 g, 43%). Elemental analysis:
found C 65.15, H 10.19%; C14H26O4 requires C 65.09, H
10.14%. 1H NMR (300 MHz; CDCl3): δ 3.66 (s, 6H, CH3O-), 2.29
(t, J = 7.5 Hz, 4H, CH3O-CO-CH2-), 1.60 (quintet, J = 7.3 Hz, 4H,
CH3 O-CO-CH2-CH2-), 1.26 (broad s, 12H, CH2 chain). 13C NMR
(75 MHz; CDCl3): δ 51.90 (s, CH3O-), 34.53 (s, CH3O-CO-CH2-),
29.78–29.55 (s, CH2 chain), 25.37 (s, CH3O-CO-CH2-CH2-)
These data are consistent with the literature.12
Trimethyl ethane-1,1,2-trioate
Under air, [Pd2(dba)3] (37 mg, 0.04 mmol, 0.08 mmol Pd) and
DTBPMB (158 mg, 0.4 mmol) were introduced into a hastelloy
autoclave, sealed and purged with nitrogen. Degassed metha-
nol (10 mL), dimethyl maleate (1 mL, 8 mmol) and methane-
sulfonic acid (0.05 mL, 0.8 mmol) were added to the autoclave
by cannula. The autoclave was purged three times with CO and
the pressure was set to 30 bar. The autoclave was heated to
90 °C for 24 h. After cooling, venting and opening, the yellow
solution was filtered through paper and the solvent removed
on a rotary evaporator. The crude mixture was passed through
a silica chromatography column (hexane/ethyl acetate: 85/15).
The desired product was obtained as a white solid. Isolated
yield (1.2 g, 73%). Elemental analysis: found C 47.13, H 5.98%;
C8H12O6 requires C 47.06, H 5.92%. 1H NMR (300 MHz;
CDCl3): δ 3.87 (t, J = 7.4 Hz, 1H, CH3-OCO-CH2-CH-), 3.77 (s,
6H, (CH3-O-CO)2CH), 3.71 (s, 3H, CH3-O-CO-CH2-), 2.95 (d, J =
7.4 Hz, 2H, CH3-O-CO-CH2-). 13C NMR (75 MHz; CDCl3): δ
171.39 (s, CH3-O-CO-CH2-), 168.90 (s, CH3-O-CO-CH-), 53.06 (s,
CH3-O-CO-CH-), 52.31 (s, CH3-O-CO-CH2-), 47.56 (s, CH-(CH3-
O-CO)2), 33.06 (s, CH3-O-CO-CH2-). These data are consistent
with those in the literature.13
Dimethyl 1,12-dodecanedioate from methyl 11-undecenoate
Under air, [Pd (dba)2] (127 mg, 0.022 mmol) and DTBPMB
(439 mg, 1.10 mmol) were introduced into a hastelloy auto-
clave, sealed and purged with nitrogen. Degassed methanol
(20 mL), methyl 11-undecenoate (5 mL, 22.2 mmol) and
methanesulfonic acid (0.3 mL, 4.6 mmol) were added to the
autoclave by cannula. The autoclave was purged three times
with CO and the pressure was set to 30 bar. The autoclave was
heated to 90 °C for 20 h. After cooling, venting and opening,
the yellow solution was filtered through paper and the solvent
removed on a rotary evaporator. Methanol was added and the
mixture was then stirred at −78 °C for 10 min and filtered
using a Büchner funnel. The desired product was obtained as
a white powder. Isolated yield (3.36 g: 58%). Elemental analy-
sis: found C 65.14, H 10.18%; C14H26O4 requires C 65.09, H
Results and discussion
Composition of the TOFA
10.14%. 1H NMR (300 MHz; CDCl3): δ 3.66 (s, 6H, CH3O-), 2.29 The TOFA were analysed using GCMS and 1H NMR techniques.
(t, J = 7.5 Hz, 4H, CH3O-CO-CH2-), 1.60 (quintet, J = 7.2 Hz, 4H, At first sight, the MS spectrum shows a single major peak with
CH3O-CO-CH2-CH2-), 1.26 (broad s, 12H, CH2 chain). 13C NMR the library offering a good match of the MS for linoleic acid
(75 MHz; CDCl3): δ 51.90 (s, CH3O-), 34.53 (s, CH3O-CO-CH2-), (m/z = 280). However, there was also a peak in the MS at m/z =
This journal is © The Royal Society of Chemistry 2013
Green Chem.

View Article Online
Paper
Green Chemistry
282 possibly arising from oleic acid. The quantitative ratio of
oleic : linoleic acids was obtained using a literature method14
1
involving H NMR analysis. This showed the TOFA to contain
72% of linoleic acid and 28% oleic acid.
The methoxycarbonylation of linoleic acid
In order to identify the products from and hence to optimise
the methoxycarbonylation of TOFA (see later) as well as to
address the question of whether or not transfer hydrogenation
of dimethyl 1,19-nonadecenedioate esters (3) to dimethyl 1,19-
nonadecanedioate (4) occurs, the methoxycarbonylation of
pure linoleic acid was carried out in methanol at 30 bar and
90 °C over 48 h using a catalyst prepared in situ from
[Pd2(dba)3] (dba = dibenzylideneacetone), 1,2-bis(ditertiary-
butylphosphinomethyl)benzene (DTBPMB) and methanesulpho-
nic acid (MSA). GC analysis of the crude product showed
several components (Fig. 1). Compound 3 was readily identi-
fied as the expected dimethyl nonadecenedioate, existing as a
mixture of esters with the double bond at different positions
in the chain. It is clear that there is little, if any of the satu-
rated diester, dimethyl nonadecanedioate (4), which has a
retention time slightly longer than that of the unsaturated di-
esters and is observed in the methoxycarbonylation products
of TOFA (Fig. 2).
Fig. 2 GC/FID of the crude mixture from the methoxycarbonylation of TOFA.
esters, with 2 being methyl 17-oxooctadecanoate and 1 being a
mixture of isomers of methyl n-oxooctadecanoate with the keto
group at different positions along the chain. Compound 5 has
not been fully identified but appears to be a mixture of methyl
methoxyoctadecanoates. Finally, compound 6 was isolated
(19% yield, 44% if the reaction was carried out in the presence
of MgSO4) and fully characterised as a mixture of trimethyl 1,
n,17-heptadecanetricarboxylates, where n indicates that the
extra carbomethoxy group is at any position along the chain.
These triesters are discussed in more detail later on, but arise
from methoxycarbonylation of the double bonds in the various
isomers of dimethyl 1,19-nonadecenedioates.
We surmised that compounds 1 and 2 might arise from
addition of water to linoleic acid or methyl linoleate followed
by isomerisation of the enol to the ketone. Possible isomerisa-
tion of the double bonds into conjugation catalysed by the Pd
complex might enhance this reaction. 1 and 2 do not appear
previously to have been synthesised by hydration of linoleic
acid or its methyl ester, but they have been formed by bacterial
oxidation of oleic acid in an aqueous environment.15 Under
the same conditions,15–19 or using metal catalysts20 linoleic
acid gives unsaturated keto esters, which can dehydrate to give
conjugated dienoic acid esters.16,17,19
In our original paper2 we suggested that dimethyl 1,19-non-
adecanedioate could be produced from methyl linoleate or
linolenate by methoxycarbonylation followed by transfer
hydrogenation from methanol. Our subsequent work on
natural oils1 as well as work on methyl linoleate by Mecking
and co-workers5 suggested that this transfer hydrogenation
does not occur to any large extent in the presence of CO. These
new results confirm that significant transfer hydrogenation
does not occur under methoxycarbonylation conditions.
In order to identify the other products formed during the
methoxycarbonylation of linoleic acid, they were separated by
column chromatography and fully characterised (see Exper-
imental and ESI†). Compounds 1 and 2 are known11 C18 keto
Drying the reagents did not reduce the amount of products
1 and 2, but water is formed from esterification of linoleic acid
by methanol and this water might then add across the methyl
linoleate. To test this, a reaction was carried out under identi-
cal conditions to those used for the methoxycarbonylation of
methyl linoleate, but under N2 in place of CO. Compounds 1
and 2 were the major products confirming that this reaction is
independent of the methoxycarbonylation. Interestingly, in
this reaction, some methyl oleate was formed, suggesting that
transfer hydrogenation is possible in the absence of CO.
The origin of the various products is shown in Scheme 1.
Methoxycarbonylation of TOFA
The methoxycarbonylation of TOFA using the same catalytic
system as was used for linoleic acid, leads to a variety of
different products (see GC in Fig. 2). Most of the major
Fig. 1 GCFID trace of the crude mixture from the methoxycarbonylation of
linoleic acid. The numbered peaks arise from compounds shown in Scheme 1.
Green Chem.
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
Scheme 1 Products formed from the methoxycarbonylation of linoleic acid or TOFA. The positions of the double bond in 3 and the mid-chain substituents in 1, 5
and 6 are variable. 2, methyl 17-oxooctadecanoate, contains the terminal methyl ketone. The identification of 5 is tentative.
Table 1 Products from the methoxycarbonylation/hydrogenation of TOFAa
1
2
3
4
5
6
M
5
10
2
2
4
1
2
28
4
26
3
26
49
33
55
3
6
1
2
10
12
13
13
M/H
E/Mb
E/M/Hc
4
a GC conversion/% (uncalibrated area of all products
M
=
methoxycarbonylation, H = hydrogenation, E = esterification with
methanol in the presence of MgSO4); b methyl oleate 11%; c methyl
stearate, 10%.
products are the same as those produced from linoleic acid,
except that dimethyl 1,19-nonadecanedioate (4) from methoxy-
carbonylation of methyl oleate is a significant component.
Hydrogenation of this mixture using Pd/C led to the desired
1,19-nonadecanedioate in very good purity and 32% isolated
yield (Table 1).
Fig. 3 GCFID trace of the crude mixture from the methoxycarbonylation of
TOFA following esterification with methanol in the presence of MgSO4. Note
that the GC conditions and retention times are slightly different from those in
Fig. 1 and 2. The identities of the shown compounds have been confirmed by
GCMS.
When the methoxycarbonylation was carried out using
syngas in place of pure CO, it led to a similar ratio between
saturated and monounsaturated diesters proving the reaction
does not need high purity CO, but works with lower (and less
expensive) CO grades. This reaction also shows that the palla-
dium catalyst does not promote hydrogenation.
addition to methyl stearate and 6, predominantly the desired
4, which was isolated in 49% yield.
Identification of triesters, 6
Knowing that compounds 1 and 2 are formed from a reac- Compound 6 gave an MS trace with a parent ion of m/z = 415.
tion between methyl linoleate and water produced during the This molecular mass corresponds to that of a triester, tri-
initial esterification of linoleic acid by methanol and that the methyl heptadecane-1,n,17-tricarboxylate (HDTC, where
n
proportion of compound 6 increases with increased reaction designates any position within the hydrocarbon chain)
time, we attempted to increase the yield of the desired 4 from (Scheme 1), which could be formed from methoxycarbonyla-
TOFA. TOFA was stirred for 3 h in methanol containing the tion of dimethyl 1,19-nonadecenedioate. Apart from reactions
amount of MSA required for the subsequent methoxycarbony- of styrene21,22 and vinyl acetate,23,24 which give predominantly
lation reaction and MgSO4. GC analysis showed that complete branched products, all other studies using Pd/DTBPMB/MSA
conversion of the acids to methyl oleate and methyl linoleate for methoxycarbonylation have shown very high selectivity to
occurred. After filtration, this solution was used for methoxy- carbonylation at the terminus of carbon chains1–6,21,25 even if
carbonylation under the standard conditions, but for 10 h in the double bond is buried deep in the chain (methyl oleate,
an attempt to reduce the amount of triester, 6, formed. GC methyl erucate)1–6 or conjugated to an ester function2,21 (see
analysis showed mainly compounds 3 and 4 with some 6 below and second carbonylation of alkynes21). However, all of
(Fig. 3). Hydrogenation of this solution (Fig. 4) gave, in the substrates we have examined have the possibility of
This journal is © The Royal Society of Chemistry 2013
Green Chem.

View Article Online
Paper
Green Chemistry
Scheme 2 Synthesis of trimethyl ethane-1,1,2-trioate from methoxycarbonyla-
tion of dimethyl maleate.
various isomers with double bonds in different positions
along the chain and that methoxycarbonylation will then place
the third ester group in a variety of positions along the chain.
In support of this, there is also more than one methoxy reso-
nance near δ 3.7 ppm. All of the NMR, MS and analytical data
are consistent with the formation of trimethyl heptadecanetri-
carboxylates, with the third methoxycarbonylation occurring at
a variety of places along the chain. The triester which has two
geminal carboxylate groups at one end (malonate) would be
expected to show NMR resonances near δ 3.3 ppm (1H) and δ
Fig. 4 GCFID trace of the crude mixture from the sequential esterification with
methanol in the presence of MgSO4, methoxycarbonylation and hydrogenation
of TOFA. Note that the GC conditions and retention times are slightly different
from those in Fig. 1 and 2. The identities of the shown compounds have been
confirmed by GCMS.
1
50 ppm (13C) for the methylene group. Although the H NMR
spectrum of the isolated triester does have very weak peaks in
the region near δ 3.3 ppm, they appear to be singlets and do
not correlate with appropriate 13C resonances. This suggests
that very little of the triester containing a geminal diester is
formed, as expected, since the amount of α,β-unsaturated
diester has been measured as only 3–5%. These triesters may
be important as cross linkers in polymerisation reactions
using naturally derived monomers.
Table 2 Main fragments obtained on GCMS of compound 6 and proposed
assignments
m/z
Fragment
415
383
355
[M]+
[M − OMe]+
[M − CO2Me]+
Further to confirm the suggestion that Pd/BDTBPMB/MSA
could catalyse the methoxycarbonylation of a trapped internal
double bond, we carried out the methoxycarbonylation of
dimethyl maleate. The reaction led efficiently to trimethyl
ethane-1,1,2-trioate (Scheme 2), together with some dimethyl
fumarate, formed by Z–E isomerism of dimethyl maleate.
Compound 5 was not clearly identified. At first sight, the
1H NMR spectrum suggests the presence of a diester bearing
an extra functional group on the carbon chain. Indeed, the
spectrum shows a resonance at δ 3.30 ppm correlated with a
13C resonance at δ 56.5 ppm (DEPT CH3), probably from a
methyl ether. The MeOCH group could then give the correlated
resonances at δ 81 ppm (13C NMR, DEPT CH) and δ 3.10 ppm
(quintet, 1H NMR), Methyl ester groups are also present (1H
NMR, δ 3.73 ppm correlating with 13C NMR, δ 52.5 ppm). The
highest mass peak in the mass spectrum is at 354, which
could be consistent with loss of methanol from dimethyl
n-methoxynonadecane-1,19-dioate Mr = 386. Such a product
could be formed by addition of methanol to dimethyl 1,19-
nonadecenedioate. The difficulty in isolating a pure sample of
this product from all the other various side-products present
in the mixture did not allowed an unequivocal identification.
isomerisation of the CvC to the C terminus where it can be
trapped by methoxycarbonylation. The one exception to this
generalisation is the carbonylation of internal alkynes, where
the first carbonylation occurs at one end of the triple bond21
because triple bond isomerisation does not occur. In the case
of dimethyl 1,19-nonadecenedioate, the double bond cannot
migrate to the end of the chain because both ends are already
esterified, so there is the possibility of adding a third ester
group wherever the double bond is in the chain.
A triester could explain the different fragments seen in the
mass spectrum (Table 2), so we isolated the proposed HDTC
(compound 6) in a microanalytically pure form from the crude
mixture (GCFID shown in Fig. 1), obtained from the methoxy-
carbonylation of linoleic acid, by column chromatography and
analysed it by 1H and 13C NMR techniques. In the DEPT
13CNMR spectrum, there are at least three resonances near δ
46 ppm, which are upside down compared with the signals
from methylene groups. These resonances correlate with 1H
NMR resonances near δ 2.3 ppm, slightly upfield from the reso-
nance arising from the methylene groups next to the terminal
–CO2Me functions (δ 2.32 ppm). These resonances are exactly
Shorter chain diesters from natural resources
where the resonances from the CH attached to a mid-chain The diesters described in this paper and our previous carbony-
–CO2Me would be expected to appear. The fact that there are at lation work on methyl oleate2 or natural oils1 all contain 19
least 3 C signals near δ 46 ppm suggest that at least 3 different carbon atoms, with the advantage that all the C atoms of the
isomers are present. Given that the methoxycarbonylation cata- oleate end up in the product. However, there is considerable
lyst is also a good double bond isomerisation catalyst, it seems interest in diesters of different chain lengths, also derived
likely that the dimethyl 1,19-nonadecenedioate will exist as from natural resources. It is well established that dimethyl
Green Chem.
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
Table 3 Methoxycarbonylation of 2-undecenoic and 2 hexadecenoic acids,
Conclusion
conversion and selectivity
Tall Oil Fatty Acids, which constitute a waste from paper manu-
facture and contain mainly oleic and linoleic acids, have been
methoxycarbonylated to a mixture of dimethyl 1,19-non-
adecanedioate and dimethyl 1,19-nonadecenedioate (mixture
of isomers). Some of the latter is converted into a mixture of
isomers of trimethyl 1,n,17-heptadecanetricarboxylate, which
may be interesting as cross-linkers in polyesters prepared from
plant based diesters. Other products include keto esters
derived from addition of water (generated in the initial esterifi-
cation of the acids by methanol) to isomerised methyl lino-
leate. In addition to helping with the identification of the
products obtained from TOFA, the methoxycarbonylation of
linoleic acid shows that esters containing two double bonds
give unsaturated diesters and triesters as major products with
no transfer hydrogenation to give saturated diesters. By pre-
esterifying the TOFA and following the methoxycarbonylation
with hydrogenation, dimethyl 1,19-nonadecanedioate, which is
potentially an important polyester precursor can be obtained
in 49% yield. Undecenoic acids or esters with the double
bond conjugated to the carbonyl group (available from meta-
thesis of methyl oleate with dimethyl maleate)9 or in the
terminal position (available from castor oil) can both be
methoxycarbonylated to dimethyl 1,12-dodecanedioate, a com-
ponent of Nylon-12,12, whilst the conjugated double bond in
2-hexadecenoic acid can be moved 13 bonds along the chain
to the least thermodynamically favoured terminal position
before methoxycarbonylation to furnish dimethyl 1,17-
heptadecanedioate.
Methyl 10-
undecenoate
2-Undecenoic
acid
2-Hexadecenoic
acid
Substrate
Conv/% (GC FID)a
Sel. linear di.
Sel. ω-2 br. di.
Sel. ω-3 br. di.
Other br. di.
83
89
6
1.5
3.5
58
84
92
4.3
1
2.6
43
79
90
3.5
1
5.5
58
Isolated yield
a Uncalibrated, based on areas; Sel = selectivity, br = branched, di =
diester.
1,18-octadecanoate can be formed by self-metathesis of methyl
oleate9,26–28 followed by hydrogenation, but in this case, half
the C atoms end up in a long chain hydrocarbon. In order to
improve the atom efficiency for obtaining shorter chain di-
esters from methyl oleate and related compounds, we have pro-
posed and demonstrated9 the metathesis of methyl oleate with
dimethyl maleate to produce dimethyl 1,11-undec-2-enedioate,
a C11 diester and methyl 2-undecenoate, an α,β-unsaturated
monoester. These products can also be formed by cross
metathesis of methyl oleate with methylacrylate.29 In order to
make the C12 diester, we attempted the methoxycarbonylation
of 2-undecenoic acid, which requires the double bond to move
8 bonds into the least thermodynamically favoured position
before being trapped there by methoxycarbonylation. Pre-
viously the maximum distance for isomerisation that has been
observed in this isomerising methoxycarbonylation chemistry
was 8 bonds (for methyl oleate1–6 or methyl erucate4,5) if the
double bond was not conjugated or 7 bonds where the double
bond was conjugated to an ester (second carbonylation of
1-octyne to give dimethyl 1,10-decanedioate).21 Gratifyingly the
reaction worked well and produced dimethyl 1,12-dodecane-
dioate with high selectivity (Table 3). The same product, which
is an important component of Nylon-12,12, was also prepared
by the methoxycarbonylation of methyl 10-undecenoate, which
is available from cracking castor oil,10 although shorter reac-
tion times could be used in this case. This reaction has pre-
viously been demonstrated using other catalysts.30 To test the
isomerising ability of the catalyst still further, 2-hexadecenoic
acid, which requires the double bond to move 13 places out of
conjugation to make the α,ω-diester, was successfully methoxy-
carbonylated to dimethyl 1,17-heptadecanedioate in high
selectivity (Table 3). These reactions are summarised in
Scheme 3.
Acknowledgements
We thank Professor F. D. Gunstone for drawing our attention
to the potential of TOFA, EaStCHEM and the University of
St. Andrews for a studentship (M. R. L. F.) and the EU Erasmus
Programme for financial support (T. S.).
Notes and references
1 M. R. L. Furst, R. Le Goff, D. Quinzler, S. Mecking,
C. H. Botting and D. J. Cole-Hamilton, Green Chem., 2012,
14, 472.
2 C. Jimenez-Rodriguez, G. R. Eastham and D. J. Cole-Hamil-
ton, Inorg. Chem. Commun., 2005, 8, 878.
3 D. Quinzler and S. Mecking, Angew. Chem., Int. Ed., 2010,
49, 4306.
4 F. Stempfle, D. Quinzler, I. Heckler and S. Mecking,
Macromolecules, 2011, 44, 4159.
5 F. Stempfle, R. Roesle and S. Mecking, ACS Symp. Ser.,
2012, 1105, 151.
6 G. Walther, J. Deutsch, A. Martin, F.-E. Baumann, D. Fridag
and A. Köckritz, ChemSusChem, 2011, 4, 1052.
Scheme 3 Methoxycarbonylation of unsaturated C11 (n = 1) and C16 (n = 6)
esters.
This journal is © The Royal Society of Chemistry 2013
Green Chem.

View Article Online
Paper
Green Chemistry
7 L. Norlin, Ullmann’s Encyclopedia of Industrial Chemistry, 20 S. Warwel, W. Pompetzki and E. A. Deckwirth, Fett Wiss.
Wiley VCH, Wienheim, 2002.
Technol.-Fat Sci. Technol., 1991, 93, 210.
8 A. Pollak, Oil Soap, 1940, 17, 87.
21 A. A. Nunez-Magro, L. M. Robb, P. J. Pogorzelec,
A. M. Z. Slawin, G. R. Eastham and D. J. Cole-Hamilton,
Chem. Sci., 2010, 1, 723.
22 H. Ooka, T. Inoue, S. Itsuno and M. Tanaka, Chem.
Commun., 2005, 1173.
23 A. J. Rucklidge, G. E. Morris and D. J. Cole-Hamilton,
Chem. Commun., 2005, 1176.
24 A. J. Rucklidge, G. E. Morris, A. M. Z. Slawin and D. J. Cole-
Hamilton, Helv. Chim. Acta, 2006, 89, 1783.
9 R. Duque, E. Oechsner, H. Clavier, F. Caijo, S. P. Nolan,
M. Mauduit and D. J. Cole-Hamilton, Green Chem., 2011,
13, 1187.
10 D. S. Ogunniyi, Bioresour. Technol., 2006, 97, 1086.
11 A. P. Tulloch, Can. J. Chem., 1977, 55, 1135.
12 N. Yamamoto, Y. Obora and Y. Ishii, J. Org. Chem., 2011,
76, 2937.
13 B. Janza and A. Studer, Org. Lett., 2006, 8, 1875.
14 G. Knothe and J. A. Kenar, Eur. J. Lipid Sci. Technol., 2004, 25 C. J. Rodriguez, D. F. Foster, G. R. Eastham and D. J. Cole-
106, 88.
Hamilton, Chem. Commun., 2004, 1720.
26 A. Behr, A. Westfechtel and J. P. Gomes, Chem. Eng.
Technol., 2008, 31, 700.
27 U. Biermann, U. Bornscheuer, M. A. R. Meier, J. O. Metzger
and H. J. Schaefer, Angew. Chem., Int. Ed., 2011, 50, 3854,
and references therein.
15 C. T. Hou, J. Am. Oil Chem. Soc., 1995, 72, 1265.
16 S. Kishino, J. Ogawa, K. Yokozeki and S. Shimizu, Biosci.,
Biotechnol., Biochem., 2011, 75, 318.
17 S. Kishino, S.-B. Park, M. Takeuchi, K. Yokozeki, S. Shimizu
and J. Ogawa, Biochem. Biophys. Res. Commun., 2011, 416,
188.
28 J. C. Mol, Green Chem., 2002, 4, 5.
18 J. A. Hudson, B. Morvan and K. N. Joblin, FEMS Microbiol. 29 A. Rybak and M. A. R. Meier, Green Chem., 2007, 9,
Lett., 1998, 169, 277. 1356.
19 J. Ogawa, S. Kishino, A. Ando, S. Sugimoto, K. Mihara and 30 D. F. Foster, R. H. Van and R. P. Tooze, WO 2008023338A1,
S. Shimizu, J. Biosci. Bioeng., 2005, 100, 355. 2008.
Green Chem.
This journal is © The Royal Society of Chemistry 2013