Integrated extraction and catalytic upgrading of microalgae lipids in supercritical carbon dioxide

Article abstract of DOI:10.1039/c9gc00312f

Fatty acids from microalgae are attractive compounds for catalytic upgrading to chemicals, but their extraction often requires multi-step procedures and the use of various organic solvents. To relieve this bottleneck, we propose a straightforward approach of combined extraction and catalytic functionalization via olefin cross-metathesis (ethenolysis and butenolysis) in supercritical CO₂ (scCO₂). This is demonstrated for Phaeodactylum tricornutum microalgae biomass. ScCO₂ at optimum conditions (90 °C, 620 atm, ρ(CO₂) = 0.90 g mL^-1) extracted the lipids selectively and quantitatively from previously disrupted cells, while organic solvent extraction for comparison additionally extracted polar diacylglycerides and chlorophylls. In a one-pot approach, olefin cross-metathesis of the unsaturated fatty acids (FA16:1, FA18:1 and FA20:5) by alkenolysis yielded the desirable mid-chain olefin and unsaturated ester products. The product spectrum compares to alkenolysis of individual model compounds in scCO₂ as well as of separately scCO₂ extracted microalgae oil. Both these ethenolysis and butenolysis proceed with conversions of more than 81% and high selectivities to the desired products. This biorefinery approach was further illustrated by the simultaneous extraction and catalytic isomerizing alkoxycarbonylation in scCO₂.

Full text of DOI:10.1039/c9gc00312f

Green Chemistry
View Article Online
View Journal
PAPER
Integrated extraction and catalytic upgrading of
microalgae lipids in supercritical carbon dioxide†
Cite this: DOI: 10.1039/c9gc00312f
Julia Zimmerer,
Stefan Mecking
Dennis Pingen,
*
Sandra K. Hess,
Tobias Koengeter
and
Fatty acids from microalgae are attractive compounds for catalytic upgrading to chemicals, but their
extraction often requires multi-step procedures and the use of various organic solvents. To relieve this
bottleneck, we propose a straightforward approach of combined extraction and catalytic functionalization
via olefin cross-metathesis (ethenolysis and butenolysis) in supercritical CO2 (scCO2). This is demon-
strated for Phaeodactylum tricornutum microalgae biomass. ScCO2 at optimum conditions (90 °C, 620
atm, ρ(CO2) = 0.90 g mL−1) extracted the lipids selectively and quantitatively from previously disrupted
cells, while organic solvent extraction for comparison additionally extracted polar diacylglycerides and
chlorophylls. In a one-pot approach, olefin cross-metathesis of the unsaturated fatty acids (FA16:1, FA18:1
and FA20:5) by alkenolysis yielded the desirable mid-chain olefin and unsaturated ester products. The
product spectrum compares to alkenolysis of individual model compounds in scCO2 as well as of separ-
ately scCO2 extracted microalgae oil. Both these ethenolysis and butenolysis proceed with conversions of
more than 81% and high selectivities to the desired products. This biorefinery approach was further illus-
trated by the simultaneous extraction and catalytic isomerizing alkoxycarbonylation in scCO2.
Received 26th January 2019,
Accepted 1st April 2019
DOI: 10.1039/c9gc00312f
rsc.li/greenchem
environmentally harmful organic solvents. This also applies to
microalgae, where the achievable cell densities are limited. To
Introduction
Microalgae are a promising future feedstock. Unlike traditional address this problem, we pursue an integration of the biomass
crops like soy, palms or sunflowers they do not require arable extraction with the catalytic upgrading in a common solvent,
land. Also, per acre yields can be much higher. Microalgae can supercritical carbon dioxide (scCO2. Fig. 1).
be cultivated in fresh, salt or brackish water.1 Autotrophically
Advantageously, scCO2 is capable of the selective extraction
or heterotrophically grown microalgae are an established of lipids from microalgae6–11 and various catalytic reactions of
source of high value food additives and pigments. Further, the interest have been reported to be compatible with scCO2 as a
production of fuels from microalgae oils has been demon- solvent.12–16 Additionally, scCO2 is non-toxic, non-flammable
strated on a multiton scale. By comparison to fuels, a pro- and it can be easily removed. For the upgrading of palm oil,
duction of higher value chemicals would appear more sensi- alkenolysis is an established reaction.17 Considering micro-
ble.2,3 Unlike fuel production, which downgrades the plant oil algae oils as a feedstock, alkenolysis of their unconventional
feedstock to hydrocarbons, the production of chemicals can
take advantage of the particular molecular structure of the
feedstock. Notably, microalgae oils can contain unique fatty
acids with unusual chain lengths and multiple double bonds
not found in traditional plant oils.1,4,5
Yet the usage of biomass has a significant bottleneck; the
isolation of the desired substrates and extraction often
requires energy-intensive multi-step procedures and the use of
Chair of Chemical Materials Science, Department of Chemistry,
University of Konstanz, 78464 Konstanz, Germany.
E-mail: stefan.mecking@uni-konstanz.de; Fax: +49 7531 88-5152;
Tel: +49 7531 88-5151
Fig. 1 Schematic process design of extraction and catalytic valorization
of microalgae in scCO2 to mid-chain olefins, unsaturated esters and
long-chain α,ω-diesters.
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c9gc00312f
This journal is © The Royal Society of Chemistry 2019
Green Chem.

View Article Online
Paper
Green Chemistry
fatty acids (vide supra) can provide a range of useful com-
pounds, difficult to access otherwise.18 In literature ruthe-
nium-based olefin metathesis has been reported to be compa-
tible with scCO2.19–21
Results and discussion
Extraction and characterization of algae oil from
Phaeodactylum tricornutum
Fig. 3 Gas chromatogram of scCO2 extracted algae oil from
Phaeodactylum tricornutum (after transesterification with methanol for
GC analysis). Myristic acid methyl ester FA14:0 9%, palmitoleic acid
methyl ester FA16:1 47%, palmitic acid methyl ester FA16:0 25%, oleic
acid methyl ester FA18:1 8%, stearic acid methyl ester FA18:0 <1%, eico-
sapentaenoic acid methyl ester FA20:5 10%.
We employed Phaeodactylum tricornutum for our studies, as
this algae strain is robust and can contain high amounts of
unsaturated fatty acids.22 To find suitable extraction con-
ditions, 1 g of freeze-dried algae (harvested in the late station-
ary phase) was subjected to scCO2 extraction at different press-
ures and temperatures, and thus also different densities23
(Table S2 and Fig. S6 in the ESI†).
multiple unsaturated fatty acid eicosapentaenoic acid (FA20:5),
and 35% of saturated fatty acids (<1% stearic acid (FA18:0),
25% palmitic acid (FA16:0) and 9% myristic acid (FA14:0)) are
present. In total the algae oil extracted by means of scCO2 con-
tains around 4 mmol of double bonds per 1 gram of oil, as
quantified via an internal GC standard. By comparison, the oil
extracted via the modified Folch method has a similar fatty
acid profile (see Fig. S5†).
At a given density, the efficiency of extraction as reflected by
the yield of algae oil increases with increasing extraction temp-
eratures. The extraction pressure is often the technically limit-
ing factor, given by the equipment employed. Notably, for a
given pressure, the yield increases with increasing tempera-
ture, even though the scCO2 density decreases. Disruption of
the microalgae cells by ultrasonication24,25 (Fig. S4†) prior to
extraction enhanced yields by ca. 20% to 30%. Under the
optimum conditions of the range investigated, 90 °C and 621
bar corresponding to ρ = 0.9 g mL−1, 25 wt% yield of algae oil
were obtained (Fig. 2).
As a reference, an algae sample was extracted with organic
solvent using a mixture of water, methanol and chloroform
(3 : 4 : 8), as described by Folch.26 This method, which quanti-
tatively captures all fatty acid compounds present in the micro-
algae, yielded 28 wt% of extracted algae oil.
The entire composition of the extracted oils in terms of
classes of compounds present was quantified by a comprehen-
sive analysis by a combination of methods (cf. ESI†). The lipid
composition was analyzed via thin-layer chromatography
(TLC),27,28 pigments were quantified via high-performance
liquid chromatography (HPLC)29 and the content of proteins
and carbohydrates was determined via Fourier-transform infra-
red spectroscopy (FT-IR).30 Polar diacylglycerides31 and chloro-
phylls32 were found to not be extracted by the apolar scCO2
(Fig. S14 and S17†), in line with previous findings.33 Proteins
were also not detected in the investigated oils (Fig. S14†),
corresponding to a protein content below 2% (detection limit
of FT-IR). Compared to the aforementioned Folch organic
solvent extraction, extraction with scCO2 as a nonpolar solvent
is indeed more selective for the desired triacylglycerides, and
carotenoids. This is also reflected by the color of the samples
(Fig. S7†). The amount of pigments and polar diacylglycerides,
which could interfere with catalysts, are 1.1% and below the
detection limit (<5 wt% for TLC), respectively, for scCO2
extracted oil. By comparison, with Folch’s organic solvent
mixture also substantial amounts (∼10%) of phospho- and gly-
colipids (Fig. S11 and S12†) as well as chlorophylls (Fig. S17
and S18†) are extracted from the microalgae samples
(Table S8†). This accounts for the differences in yields, that is,
scCO2 extraction under the conditions studied quantitatively
extracts the triacylglycerides from the microalgae biomass.
Gas chromatographic (GC) analysis of the extracted oils
reveals similar fatty acid compositions, largely independent of
the scCO2 extraction conditions (Fig. 3). With only small vari-
ations (ca. 2%), 55% of mono-unsaturated fatty acids (47% pal-
mitoleic acid (FA16:1) and 8% oleic acid (FA18:1)), 10% of the
Fig. 2 Yields of scCO2 extraction at different pressures and tempera-
tures at constant density (0.9 g mL−1). Lower part of the bar: extraction
of freeze-dried algae, lower and upper part of the bar: extraction of
ultrasound pre-treated freeze-dried algae.
Ethenolysis of model compounds and microalgae oil
Catalytic upgrading of fatty acids from microalgae via cross-
metathesis gives access to a broad spectrum of unsaturated
Green Chem.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Green Chemistry
Paper
Scheme 1 Self- and cross-metathesis of methyl oleate with ethylene
or 2-butene.
Scheme 2 Ethenolysis products of the unsaturated fatty acids present
in algae oil from P. tricornutum, and CHD as a self-metathesis product
of FA20:5.
compounds. A general scheme of cross-metathesis of methyl
oleate with ethylene or 2-butene is shown in Scheme 1.
from FA16:1 and FA18:1 1-octene HC8:1, 1-decene HC10:1 and
The ethenolysis of unsaturated fatty acids in scCO2 as reac- methyl 9-decenoate E10:1 is formed, whereas complete etheno-
tion medium was first investigated with methyl oleate as a lysis of FA20:5 results in 1,4-pentadiene HC5:2, methyl 5-hex-
model compound which gives methyl 9-decenoate E10:1 and enoate E6:1 and 1-butene HC4:1. Yet, 1,4-pentadiene HC5:2
1-decene HC10:1 as the desired products (Scheme 1, R = H).
and 1-butene HC4:1 could not be quantified by GC analysis
Based on the results of the scCO2 extraction of algae oil and due to their low boiling point. Please note, complete ethenoly-
with respect to the limited thermal stability of the metathesis sis of FA20:5 can generate up to four equivalents of HC5:2.
catalyst, a temperature of 45 °C and a pressure of 300 bar were Furthermore, instead of cross-metathesis, up to two equiva-
chosen for all ethenolysis experiments. In accordance to Song lents of 1,4-cyclohexadiene CHD can be generated in a self-
et al.,34 1 mol% of Grubbs 1st generation catalyst was metathesis reaction of FA20:5. Its intramolecular nature in
employed. With 10 bar ethylene and total pressure of 300 bar principle favors the latter reaction.35
at 45 °C a conversion of 61% of methyl oleate was achieved
To exclude any adverse effects from additional compounds
(Table S10†). No self-metathesis products (<1%) of methyl present in small amounts in the algae oil besides the fatty
oleate (dimethyl 9-octadecendioate DE18:1 and 9-octadecene acids (e.g. carotenoids), ethenolysis of a mixture of model com-
HC18:1, Scheme 1) were observed. Due to promising results pounds (40% FA16:0 50%, FA18:1 and 10% FA20:5) resembling
with Hoveyda–Grubbs 1st generation catalyst in the ethenolysis the fatty acid composition of algae oil was investigated
of methyl oleate in dichloromethane (Table S9†), this catalyst (Table 1, column “model substrate mixture”). The conversion
was also applied in the ethenolysis with scCO2 as solvent. and selectivity for all expected reaction products were deter-
Under the same conditions as for Grubbs 1st generation cata- mined over FA16:0 as an internal standard via gas chromato-
lyst (1 mol% catalyst, 10 bar ethylene, 300 bar total pressure at graphy. The selectivity is defined as the ratio of the formed
45 °C), Hoveyda–Grubbs 1st generation catalyst gave a higher product to the theoretical maximum amount of this product at
conversion of 88% and again almost no self-metathesis pro- complete ethenolysis or in case of CHD (1,4-cyclohexadiene)
ducts were observed. Based on these results, the ethenolysis in complete self-metathesis of FA20:5 (formation of two equiva-
scCO2 was further investigated with Hoveyda–Grubbs 1st gene- lents CHD).
ration catalyst. The highest conversions of up to 88% were
In accordance to the experiments with neat methyl oleate, a
obtained using a catalyst loading of 0.5 mol% or 1 mol%, catalyst loading of 0.5 mol% Hoveyda–Grubbs 1st generation
respectively, at an ethylene pressure of 10 bar. Conversions catalyst (referring to the number of double bonds) and 10 bar
could not be improved neither by lowering the ethylene of ethylene were chosen. After 6 h, a high conversion of 86%
pressure nor by increasing the reaction time (see Table S10†).
for FA18:1 was observed (Table 1). This value is comparable to
Considering an effect on the solubility of the catalyst pre- the conversion of neat methyl oleate as single model com-
cursor, Grubbs 1st and 2nd generation catalysts have been pound. The poly-unsaturated component FA20:5 was almost
suggested to be insoluble in scCO2 (40 °C, 140 bar corresponds completely consumed after 6 h.
to 0.75 g mL−1).20 While we cannot exclude that under our con-
The selectivity for the ethenolysis products of FA18:1
ditions a dissolution of the catalyst precursor (Hoveyda– (HC10:1 and E10:1) was above 99%, as also observed for the
Grubbs 2nd generation catalyst, Grubbs 1st and 2nd generation single FA18:1 compound. The ethenolysis product E6:1 of
catalyst, respectively) is assisted by reaction with the fatty acid FA20:5 was also formed with a high selectivity of 85%.
substrate, a comparison of conversion in ethenolysis in scCO2 However, the self-metathesis of FA20:5 could not be sup-
vs. dichloromethane as a solvent (Tables S10 and S9†) suggests pressed. The selectivity for the self-metathesis product 1,4-
that in scCO2 catalyst solubility is not an issue and a substantial cyclohexadiene CHD was 84%. This suggests a higher reaction
portion or the entire amount of catalyst precursor is dissolved.
rate of the intramolecular self-metathesis leading to the for-
The ethenolysis of algae lipid feedstocks will provide a mation of CHD compared to the intermolecular cross-
more complex spectrum of unsaturated products (Scheme 2): metathesis.
This journal is © The Royal Society of Chemistry 2019
Green Chem.

View Article Online
Paper
Green Chemistry
Table 1 Ethenolysis of unsaturated fatty acids of a model substrate
mixture and scCO2 extracted algae oil with conversions of the com-
ponents and selectivities to ethenolysis products and the self-metathesis
product 1,4-cyclohexadiene (CHD)
Model
scCO2
substrate extracted
Ethenolysis of
mixturea algae oil
Composition of the initial reaction
mixture [%]
FA14:0
FA16:1
FA16:0 40
—
—
9
49
27
4
FA18:1 50
FA18:0
FA20:5 10
—
2
9
Fig. 4 Gas chromatogram of the ethenolysis of scCO2 extracted algae
oil in scCO2 (after transesterification with methanol) with Hoveyda–
Grubbs 1st generation catalyst and assignments of the ethenolysis pro-
ducts (HC8:1 1-octene, E6:1 methyl 5-hexenoate, HC10:1 1-decene,
E10:1 methyl 9-decenoate), the self-metathesis product CHD 1,4-cyclo-
hexadiene and the fatty acid esters (FA14:0 methyl myristate,
FA16:1 methyl palmitoleate, FA16:0 methyl palmitate, FA18:1 methyl
oleate). Other signals likely originate from side-products of ethyl vinyl
ether quenching.
Conversionb [%]
FA16:1
FA18:1 86
FA20:5 >99
—
81
90
92
Selectivityc [%] for
CHD
E6:1
HC8:1
HC10:1 >99
E10:1 >99
84
85
—
64
82
83
97
84
Conditions: 0.5 mol% Hoveyda–Grubbs 1st generation catalyst per
double bond, 10 bar ethylene, 300 bar CO2 (total pressure) at 45 °C,
6 h. Average of two independent experiments, cf. ESI for complete
ducts. While ethenolysis gives rise to products of mainly even-
data. a Mixture of 40% FA16:0, 50% FA18:1, 10% FA20:5. b Determined numbered hydrocarbon chains, butenolysis with 2-butene
over FA16:0 as an internal standard. c The selectivity to a product is
gives odd-numbered products. Furthermore, cross-metathesis
defined as the ratio of the product to the theoretical maximum
with internal alkenes circumvents the specific disadvantage of
amount of this product at complete ethenolysis or in case of CHD (1,4-
cyclohexadiene) complete self-metathesis of FA20:5 (determined over ethenolysis: in the cross-metathesis with ethylene an unstable
FA16:0 as an internal standard) in the gas chromatogram.
methylidene intermediate is formed which can lead to
fast decomposition of the catalyst and therefore limited
productivity.36–38
Ethenolysis of algae oil was performed applying the same
reaction conditions as for the model substrate mixture. The
selectivities and conversions were determined via gas chrom-
atography (Fig. 4) over FA16:0 as internal standard, which is,
besides FA14:0, found as saturated fatty acid in algae oil.
The conversion of the mono-unsaturated fatty acids in the
algae oil are 81% and 90% for FA16:1 and FA18:1, respectively,
and are therefore in the same range as from the ethenolysis of
the model substrate mixture. Also, the conversion of the five-
fold unsaturated fatty acid FA20:5 is again almost complete
(92%). The selectivities for the ethenolysis products of the
mono-unsaturated fatty acids are between 83% and 97% and
consistently high compared to the model substrate mixture.
The selectivity for E6:1 is 82% and equals the result of the
model substrate mixture. However, the selectivity of 64% for
CHD as self-metathesis product is lower than the values
observed with the model substrate mixture.
To establish the butenolysis in scCO2 as reaction medium,
first the butenolysis of methyl oleate as a model compound
was investigated. Hoveyda–Grubbs 2nd generation catalyst was
chosen as it gave better results in butenolysis reactions in
organic solvent compared to Hoveyda–Grubbs 1st generation
catalyst.18,39,40 Similar to the aforementioned ethenolysis in
scCO2, the butenolysis was conducted at a pressure of 300 bar
at a temperature of 45 °C.
Butenolysis of methyl oleate results in the formation of
methyl 9-undecenoate E11:1 and 2-undecene HC11:1
(cf. Scheme 1, R = Me). Conditions similar to butenolysis in di-
chloromethane were adopted (10 equivalents of 2-butene and
0.1 mol% Hoveyda–Grubbs 2nd generation catalyst),18 which
resulted in a conversion of methyl oleate of 93%. Only small
amounts of self-metathesis products (dimethyl 9-octadecendio-
ate DE18:1 and 9-octadecene HC18:1) were formed. Selectivity
for the desired butenolysis products was around 94%. The
butenolysis products were formed in a cis : trans ratio of about
20 : 80. All these results match with butenolysis of methyl
oleate in dichloromethane as solvent.18,40 Upon decreasing the
catalyst loading the selectivity was virtually unaffected whereas
the conversion dropped to 82% at a catalyst loading of
0.05 mol% and to 25% at a catalyst loading of 0.01 mol%,
Overall, ethenolysis of scCO2 extracted algae oil in scCO2
proceeds with high conversions between 81% and 92% and
high selectivities for ethenolysis products with exception of
CHD formed by intramolecular self-metathesis of FA20:5.
Butenolysis of model compounds and microalgae oil
Besides ethenolysis, the cross-metathesis with internal alkenes respectively (Table S11†). Compared to ethenolysis of methyl
such as 2-butene is a versatile tool to convert fatty acids to a oleate in scCO2, the catalyst loading required for maximum
wide spectrum of unsaturated compounds and gives access to butenolysis (0.1 mol%) is fivefold lower. As previously stated,
products with chain lengths different to the ethenolysis pro- the formation of the unstable methylidene intermediate cata-
Green Chem.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Green Chemistry
Paper
Table 2 Butenolysis of unsaturated fatty acids of a model substrate
mixture and scCO2 extracted algae oil with conversions of the com-
ponents and selectivities for butenolysis products and the self-meta-
thesis product 1,4-cyclohexadiene (CHD)
lyst species in the ethenolysis is presumably a major factor
contributing to the necessity of a higher catalyst loading. In
addition, the more reactive terminal double bond of the ethe-
nolysis products can compete with the internal one of the
starting material.
Model
scCO2
substrate extracted
Again, first the simplified model substrate mixture was
used. The following butenolysis products were expected
(Scheme 3): 2-undecene HC11:1 and methyl 9-undecenoate
E11:1 from FA18:1 and 2,5-heptadiene HC7:2, methyl 5-hep-
tenoate E7:1 and 2-pentene HC5:1 formed from FA20:5. Note
that complete butenolysis of FA20:5 can generate up to four
equivalents of HC7:2. 2-Pentene HC5:1, the smallest butenoly-
sis product of FA20:5, was not detected in GC analysis due to
its low boiling point.
The same reaction conditions as for methyl oleate were
adopted employing a catalyst loading of 0.1 mol% Hoveyda–
Grubbs 2nd generation catalyst (referring to the number of
double bonds). After two hours, GC analysis revealed almost
complete conversion for FA20:5 and a conversion of 88% for
FA18:1. Butenolysis of FA18:1 proceeded in high selectivity
(88% for E11:1 and 84% for HC11:1) whereas FA20:5 was con-
verted somewhat less selective (HC7:2 52% and E7:1 66%).
Butenolysis of
mixturea algae oil
Composition of the initial reaction
mixture [%]
FA14:0
FA16:1
—
—
9
47
25
8
FA16:0 40
FA18:1 50
FA18:0
—
1
FA20:5 10
10
Conversionb [%]
FA16:1
—
81
91
FA18:1 88
FA20:5 97
>99
Selectivityc [%] for
CHD
HC7:2
E7:1
27
52
66
—
24
55
65
85
83
82
HC9:1
HC11:1 84
E11:1 88
Conditions: 0.1 mol% Hoveyda–Grubbs 2nd generation catalyst, 10-fold
excess of 2-butene, 300 bar CO2 (total pressure) at 45 °C, 2 h. Average
Incomplete butenolysis of FA20:5, leading to two-, three- and of two independent experiments, cf. ESI for complete data. a Mixture of
40% FA16:0, 50% FA18:1, 10% FA20:5. b Determined via FA16:0 as an
four-fold unsaturated products, can decrease the selectivity for
E7:1 and HC7:2. Remarkably, the selectivity for the self-meta-
internal standard. c The selectivity to a product is defined as the ratio
of the product to the theoretical maximum amount of this product at
thesis product CHD is only 27% and thus significantly lower complete butenolysis or in case of CHD (1,4-cyclohexadiene) complete
self-metathesis of FA20:5 (determined via FA16:0 as an internal stan-
dard) in the gas chromatograms.
compared to the corresponding ethenolysis experiment. Please
note that the concentrations of 2-butene and ethylene differ.
In the butenolysis a 10-fold excess of 2-butene was applied,
whereas in the ethenolysis experiments a 4-fold excess of ethyl-
ene was used.
Based on these results, butenolysis of scCO2 extracted algae
oil (Table 2, column “scCO2 extracted algae oil” and Fig. 5) was
carried out applying the same reaction conditions (0.1 mol%
catalyst, 300 bar scCO2 at 45 °C, 10-fold excess of 2-butene).
Besides 2-nonene HC9:1, 2-undecene HC11:1 and methyl
9-undecenoate E11:1 as butenolysis products from FA16:1 and
FA18:1, respectively, methyl 5-heptenoate E7:1 and 2,5-hepta-
diene HC7:2 from FA20:5 were found. With 91% and 99% the
conversions of FA18:1 and the five-fold unsaturated fatty
FA20:5, respectively, agree with the conversions observed for
Fig. 5 Gas chromatogram of butenolysis in scCO2 of scCO2 extracted
the mixture of model compounds (Table 2). For FA16:1 a con-
algae oil (after transesterification with methanol) with Hoveyda–Grubbs
2nd generation catalyst and assignments of the butenolysis products
(HC7:2 2,5-heptadiene, HC9:1 2-nonene, E7:1 methyl 5-heptenoate,
HC11:1 2-undecene, E11:1 methyl 9-undecenoate) and the fatty acid
esters (FA14:0 methyl myristate, FA16:1 methyl palmitoleate,
FA16:0 methyl palmitate, FA18:1 methyl oleate). Other signals likely orig-
inate from side-products of ethyl vinyl ether quenching.
version of 81% was determined. The selectivities for the bute-
nolysis products of the mono-unsaturated fatty acids are high
(between 82 and 85%) and in the same range as for methyl
oleate as a single model component. Furthermore, the buteno-
Scheme 3 Butenolysis products of the unsaturated fatty acids present
lysis products of FA20:5 are formed with selectivities compar-
in algae oil from Phaeodactylum tricornutum, and CHD as a self-meta-
thesis product of FA20:5.
able to the simplified model substrate mixture.
This journal is © The Royal Society of Chemistry 2019
Green Chem.

View Article Online
Paper
Green Chemistry
All in all, in the butenolysis of scCO2 extracted algae oil conversion of the fivefold unsaturated fatty acid FA20:5 is
high conversions and selectivities were achieved comparable to higher with 88%.
those obtained in the reactions with the neat model substances.
In comparison to the two-step procedure the conversions in
this combined approach differ significantly. While the conver-
sion of the poly-unsaturated fatty acid FA20:5 is almost identi-
cal (88% for combined and 92% for two-step approach), the
mono-unsaturated fatty acids FA16:1 and FA18:1 are converted
to a much lower extent. This might be due to a slow extraction
over the duration of the experiment, such that the extracted
fatty acids are not exposed to the catalyst over the entire experi-
ment. One reason for this are different high-pressure setups.
In the combined approach the algae oil is extracted by means
of a static scCO2 batch reactor, whereas in the two-step
process the algae oil is extracted under a continuous flow.
Furthermore, reaction conditions had to be adopted to the
limited thermal stability of the Ru catalyst and were not ideal
for a maximum yield of extracted oil.
However, the selectivities in the combined approach are
comparable to the two-step procedure. The selectivities for the
ethenolysis products of the mono-unsaturated fatty acids in
the combined approach are between 60 and 88% and in the
two-step procedure between 83 and 97%. Also, for the poly-
unsaturated fatty acid FA20:5 the selectivities for the ethenoly-
sis product E6:1 (87%) and for the self-metathesis product
CHD (49%) are comparable to the extraction and separate
metathesis. All in all, the characteristic selectivity of the catalyst
is evidently not affected by components of the freeze-dried algae.
In the combined extraction and butenolysis of pre-treated
freeze-dried algae (Table 3) again lower conversions of the mono-
unsaturated fatty acids were observed. Conversions of 47% and
65% were found for FA16:1 and FA18:1, respectively. The conver-
sion of the poly-unsaturated fatty acid FA20:5 of 89% is still in
the same range as for the butenolysis two-step process.
The desired butenolysis products of the mono-unsaturated
fatty acids FA16:1 and FA18:1 are formed with selectivities
between 70 and 97%. For E7:1 a selectivity of 67% is observed.
These are again in accordance with the two-step approach.
The conformity of trends in conversions in ethenolysis and
butenolysis, respectively, in the combined approaches shows
that the combination of extraction and catalytic transform-
ation in one batch reactor is feasible but requires optimization
in terms of extraction. However, it is important to note that
this integrated one-pot approach shows comparable selectiv-
ities for the desired higher-value chemicals in relation to the
step-wise approach of extraction and catalysis, and therefore,
makes this a promising concept to overcome the typical bottle-
neck of extraction in the valorization of biomass.
Simultaneous extraction and cross-metathesis in scCO2
As demonstrated above, scCO2 is
a powerful extraction
medium for lipids of microalgae and is also a suitable reaction
medium for cross-metathesis of this algae feedstock that is
compatible with the ruthenium-based catalysts. In terms of
reducing multiple reaction steps, avoiding solvent removal and
integrating a direct valorization of the feedstock, we propose a
combined approach of extraction and cross-metathesis of
microalgae in scCO2.
The combination of extraction and ethenolysis or butenoly-
sis was performed at a scCO2 pressure of 300 bar at 45 °C. For
both catalytic transformations, 1 g of ultrasound pre-treated
freeze-dried algae were placed in a high-pressure reactor
together with the corresponding amount of the metathesis
catalyst (0.5 mol% of Hoveyda–Grubbs 1st generation catalyst
for ethenolysis and 0.1 mol% Hoveyda–Grubbs 2nd generation
catalyst for butenolysis, respectively). In case of ethenolysis, 10
bar ethylene and for butenolysis, a 10-fold excess of 2-butene
was applied.
Subjecting the freeze-dried algae to the combined extrac-
tion and ethenolysis results in a conversion of 32% and 44%
(Table 3) for FA16:1 and FA18:1, respectively. In contrast, the
Table 3 Integrated procedure of extraction and cross-metathesis of
freeze-dried algae in scCO2 with selectivities and conversions of the
unsaturated fatty acids
Combined
Combined
extraction and
ethenolysisa of
freeze-dried algae
extraction and
butenolysisb of
freeze-dried algae
Conversionc [%]
FA16:1
FA18:1
FA20:5
32
44
88
FA16:1
FA18:1
FA20:5
47
65
89
Selectivityd [%] for
CHD
49
—
87
60
88
77
CHD
HC7:2
E7:1
HC9:1
HC11:1
E11:1
36
70
67
96
97
70
HC5:2e
E6:1
HC8:1
HC10:1
E10:1
The fresh algae were ultrasonicated and freeze-dried. The composition
of fatty acids in the freeze-dried algae were assumed to be the same as
for the scCO2 extracted algae oil. The reaction mixture was analyzed via
gas chromatography after transesterification and filtration. Average of
two independent experiments, cf. ESI for complete data. a Conditions:
0.5 mol% Hoveyda–Grubbs 1st generation catalyst, 10 bar ethylene,
300 bar CO2 (total pressure) at 45 °C, 18 h. b Conditions: 0.1 mol%
Hoveyda–Grubbs 2nd generation catalyst, 10 fold-excess of 2-butene,
300 bar CO2 (total pressure) at 45 °C, 3 h. c Conversions were deter-
mined via gas chromatography via the FA16:0 present in the algae oil
as an internal standard. d The selectivity for a product is defined as the
ratio of the product to the theoretical maximum amount of this
product at complete ethenolysis or butenolysis or in case of CHD
(1,4-cyclohexadiene) complete self-metathesis of FA20:5 (determined
over FA16:0 as an internal standard) in GC. e Not detectable via GC due
to its low boiling point.
Alkoxycarbonylation
To further illustrate the concept of integrated scCO2 extraction
and functionalization of fatty acids from microalgae, isomeriz-
ing alkoxycarbonylation was studied as another catalytic trans-
formation system that enables the upgrading of algae oil into
higher value chemical intermediates. This converts an internal
double bond into a terminal ester group by reaction with
methanol and CO. Pd(II) catalysts with a sterically demanding
Green Chem.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Green Chemistry
Paper
diphosphine ligand such as [Pd(dtbpx)(OTf)2] (dtbpx = 1,2-bis
((di-tert-butylphosphino) methyl)benzene) convert unsaturated
Acknowledgements
fatty acid esters such as methyl oleate to linear long-chain D. P. gratefully acknowledges a Zukunftskolleg Marie-Curie
α,ω-diesters with high linear selectivities (>90%).41,42 As Incoming Fellowship for funding. Financial support by the
demonstrated previously, mid-chain (di-)carboxylic acid esters, Konstanz Research School of Chemical Biology through a
currently only accessible via demanding synthetic routes, can graduate fellowship is gratefully acknowledged by S. K. H.
be obtained in a two-step fully catalytic route by butenolysis Funding by the BMBF (INTEXCAT, 031B0676) is acknowledged.
and subsequent isomerizing alkoxycarbonylation without prior We thank Heiko Wagner of Christian Wilhelms’ group at the
work-up and purification.18 Thus, in order to prove the feasi- University of Leipzig for providing access and kind assistance
bility of a bio-refinery concept starting from crude microalgae with microplate FT-IR analysis.
biomass, scCO2 extraction was combined with isomerizing
alkoxycarbonylation.
Using methyl oleate as model compound, the isomerizing
alkoxycarbonylation was performed at 90 °C and 30 bar CO
Notes and references
pressurized with CO2 to a total pressure of 325 bar with
additional methanol. As a catalyst system 0.8 mol% [Pd(dtbpx)
(OTf)2] and 4 mol% (dtbpxH2)(OTf)2 was used. After 18 h, a
conversion of 29% was observed which is lower than expected
under standard conditions applying pure methanol as
solvent.41,43 This is due to the low nucleophile concentration
which is rate determining.44 Yet the high selectivity for the
linear product (92%) was not affected.
The isomerizing alkoxycarbonylation of scCO2 extracted
algae oil under identical conditions (but with a reaction time
of 96 h, affording a conversion of 50%) again, proceed with
high selectivities around 90% for both, the linear 1,19-diester
from FA18:1 and the linear 1,17-diester from FA16:1.
1 Y. Chisti, Biotechnol. Adv., 2007, 25, 294–306.
2 S. K. Hess, B. Lepetit, P. G. Kroth and S. Mecking,
Eur. J. Lipid Sci. Technol., 2018, 120, 1700152.
3 Y. Zhu, C. Romain and C. K. Williams, Nature, 2016, 540,
354.
4 C. Ratledge, OCL, 2013, 20, D602.
5 U. Biermann, U. Bornscheuer, M. A. R. Meier, J. O. Metzger
and H. J. Schäfer, Angew. Chem., Int. Ed., 2011, 50, 3854–
3871.
6 A. M. Aliev and I. M. Abdulagatov, J. Mol. Liq., 2017, 239,
96–100.
7 H. Taher, S. Al-Zuhair, A. H. Al-Marzouqi, Y. Haik, M. Farid
and S. Tariq, J. Supercrit. Fluids, 2014, 86, 57–66.
8 R. L. Mendes, J. P. Coelho, H. L. Fernandes, I. J. Marrucho,
J. M. S. Cabral, J. M. Novais and A. F. Palavra, J. Chem.
Technol. Biotechnol., 1995, 62, 53–59.
9 R. L. Mendes, B. P. Nobre, M. T. Cardoso, A. P. Pereira and
A. F. Palavra, Inorg. Chim. Acta, 2003, 356, 328–334.
10 J. Lorenzen, N. Igl, M. Tippelt, A. Stege, F. Qoura, U. Sohling
and T. Brück, Bioprocess Biosyst. Eng., 2017, 40, 911–918.
11 E. Tommasi, G. Cravotto, P. Galletti, G. Grillo, M. Mazzotti,
G. Sacchetti, C. Samorì, S. Tabasso, M. Tacchini and
E. Tagliavini, ACS Sustainable Chem. Eng., 2017, 5, 8316–
8322.
Also, in the one-pot combination of extraction and isomer-
izing alkoxycarbonylation of the lipids from freeze-dried algae,
the fatty acids were converted to the desired linear diester with
selectivities over 90%. Overall, these findings underline the
viability of this reaction in supercritical CO2.
Conclusions
An integration of biomass extraction and catalytic upgrading
in supercritical CO2 as a solvent allows for a more straight-
forward utilization of microalgae biomass as a source of 12 W. Leitner, C. R. Acad. Sci., Ser. IIc: Chim., 2000, 3, 595–600.
chemical intermediates. The desired lipids are extracted selec- 13 M. D. Lundin, A. M. Danby, G. R. Akien, T. P. Binder,
tively and converted in a one-pot approach by catalytic olefin
ethenolysis or butenolysis, respectively, to target mid-chain
D. H. Busch and B. Subramaniam, ACS Sustainable Chem.
Eng., 2015, 3, 3307–3314.
olefins and unsaturated esters. These olefin metathesis in 14 D. Guironnet, I. Göttker-Schnetmann and S. Mecking,
scCO2 proceed with high conversions and selectivities. In Macromolecules, 2009, 42, 8157–8164.
addition to its advantageous selectivity as a solvent, CO2 is 15 P. G. Jessop, T. Ikariya and R. Noyori, Chem. Rev., 1999, 99,
benign and easy to remove from the products. 475–494.
The approach of integrated extraction and catalytic upgrad- 16 W. Leitner, Acc. Chem. Res., 2002, 35, 746–756.
ing of lipids, demonstrated here for the case of olefin meta- 17 S. Chikkali and S. Mecking, Angew. Chem., Int. Ed., 2012,
thesis and isomerizing carbonylation, can help to overcome
51, 5802–5808.
the bottleneck biomass extraction represents for its utilization 18 J. Zimmerer, L. Williams, D. Pingen and S. Mecking, Green
as a feedstock, and of microalgae in particular.
Chem., 2017, 19, 4865–4870.
19 A. Fürstner, D. Koch, K. Langemann, W. Leitner and C. Six,
Angew. Chem., Int. Ed. Engl., 1997, 36, 2466–2469.
20 A. Fürstner, L. Ackermann, K. Beck, H. Hori, D. Koch,
K. Langemann, M. Liebl, C. Six and W. Leitner, J. Am.
Chem. Soc., 2001, 123, 9000–9006.
Conflicts of interest
There are no conflicts to declare.
This journal is © The Royal Society of Chemistry 2019
Green Chem.

View Article Online
Paper
Green Chemistry
21 X. Hu, M. T. Blanda, S. R. Venumbaka and P. E. Cassidy, 32 A. D. Ivahnov, T. E. Skrebets and K. G. Bogolitsyn,
Polym. Adv. Technol., 2005, 16, 146–149. Russ. J. Phys. Chem. B, 2016, 10, 1244–1247.
22 T. Tonon, D. Harvey, T. R. Larson and I. A. Graham, 33 J. A. Hyatt, J. Org. Chem., 1984, 49, 5097–5101.
Phytochemistry, 2002, 61, 15–24.
34 J. Song, M. Hou, G. Liu, J. Zhang, B. Han and G. Yang,
23 Density of neat carbon dioxide at the given temperatures
J. Phys. Chem. B, 2009, 113, 2810–2814.
and pressures were determined according to https:// 35 D. Pingen, J. Zimmerer, N. Klinkenberg and S. Mecking,
webbook.nist.gov/chemistry/fluid/. Green Chem., 2018, 20, 1874–1878.
24 C. Crampon, O. Boutin and E. Badens, Ind. Eng. Chem. Res., 36 K. A. Burdett, L. D. Harris, P. Margl, B. R. Maughon,
2011, 50, 8941–8953.
T. Mokhtar-Zadeh, P. C. Saucier and E. P. Wasserman,
25 J. M. Bermúdez Menéndez, A. Arenillas, J. Á. Menéndez
Organometallics, 2004, 23, 2027–2047.
Díaz, L. Boffa, S. Mantegna, A. Binello and G. Cravotto, 37 S. E. Lehman and K. B. Wagener, Organometallics, 2005, 24,
J. Chem. Technol. Biotechnol., 2014, 89, 1779–1784. 1477–1482.
26 J. Folch, I. Ascoli, M. Lees, J. A. Meath and F. N. LeBaron, 38 M. S. Sanford, J. A. Love and R. H. Grubbs, J. Am. Chem.
J. Biol. Chem., 1951, 191, 833–841. Soc., 2001, 123, 6543–6554.
27 A. Ventrella, L. Catucci, G. Mascolo, A. Corcelli and 39 J. Patel, J. Elaridi, W. R. Jackson, A. J. Robinson, A. K. Serelis
A. Agostiano, Biochim. Biophys. Acta, 2007, 1768, 1620–
and C. Such, Chem. Commun., 2005, 5546–5547.
1627.
40 J. Patel, S. Mujcinovic, W. R. Jackson, A. J. Robinson,
A. K. Serelis and C. Such, Green Chem., 2006, 8, 450–454.
28 T. Mock and B. M. A. Kroon, Phytochemistry, 2002, 61, 41–
´
51.
41 C. Jimenez-Rodriguez, G. R. Eastham and D. J. Cole-
29 B. Lepetit, S. Sturm, A. Rogato, A. Gruber, M. Sachse,
Hamilton, Inorg. Chem. Commun., 2005, 8, 878–881.
A. Falciatore, P. G. Kroth and J. Lavaud, Plant Physiol., 42 P. Roesle, C. J. Dürr, H. M. Möller, L. Cavallo, L. Caporaso
2013, 161, 853–865. and S. Mecking, J. Am. Chem. Soc., 2012, 134, 17696–17703.
30 H. Wagner, Z. Liu, U. K. Langner, K. Stehfest and 43 V. Goldbach, M. Krumova and S. Mecking, ACS Catal.,
C. Wilhelm, J. Biophotonics, 2010, 3, 557–566. 2018, 8, 5515–5525.
31 R. L. Mendes, A. D. Reis and A. F. Palavra, Food Chem., 44 V. Goldbach, L. Falivene, L. Caporaso, L. Cavallo and
2006, 99, 57–63. S. Mecking, ACS Catal., 2016, 6, 8229–8238.
Green Chem.
This journal is © The Royal Society of Chemistry 2019