JOURNAL

Journal ofOFOil
OILPalm
PALMResearch
RESEARCHVol.
24 (DECEMBER 2012)2012 p. 1492-1496
24 December

REDUCTION IN FREE FATTY ACIDS IN CRUDE

PALM OIL BY ENZYMATIC REMEDIATION
COWAN, D*; HOLM, H C* and YEE, H S*

ABSTRACT
Fatty acids present in crude palm oil are normally removed in processing and result in a yield loss. This
investigation studied if they could be recovered through an enzymatic condensation reaction. Free fatty acids
(FFA) present in crude palm oil at a level of 4.8%-7.2% could be converted back to triacyl glycerols utilising
a microbial esterase to perform a condensation reaction with mono and diglycerides present in the oil. The rate
of removal of fatty acids could be increased by the addition of glycerol to provide more sites for attachment.
However, the addition of glycerol resulted in an increase in the mono and diglyceride content of the oil as
the fatty acid preferentially attached to this molecule. In a glycerol free system, the reduction in fatty acids
was achieved when an immobilised form of the enzyme was used. This enzyme did not cause significant
interesterification of the palm oil, an advantage for later fractionation. An addition of a small amount of
palmitic acid (3% w/w), to the reaction gave a lower residual level of mono and diglycerides compared to a
reaction with only the intrinsic level of FFA.

Keywords: lipase, fatty acids, palm oil.

Date received: 23 April 2012; Sent for revision: 8 May 2012; Received in final form: 29 August 2012; Accepted: 26 September 2012.

INTRODUCTION but also the protective covering (exocarp) loses its

integrity and no longer protects the oil containing
Crude palm oil normally contains free fatty acids mesocarp from microbial colonisation.
which need to be removed as part of the refining Removal of these fatty acids results in a loss
process. Levels can vary according to harvesting of yield in the subsequent deacidification and
practices, including the length of time that passes deodorisation processes in the physical refining
between collection of the fruit bunches, their of the oil. Alternatively, if sodium hydroxide
handling and their processing at the extraction neutralisation is used, then the generated soaps
and refining locations. Different mechanisms will also entrain oil, producing a greater loss of oil.
for the generation of these fatty acids have been However, if the fatty acids can be re-combined with
described by Bock (2012) and include lipase activity the mono and diglycerides left in the oil as a result
coming from microorganisms that colonise the of the lipolysis, then, the possibility exists to reverse
fruit bunches and more importantly, cellular lipase the yield loss and recover the oil.
activity coming from damaged fruits. Bhattacharya and Bhattacharya (1989) evaluated
Several lipase producing microorganisms have Rihizomucour miehei lipase for the biorefining of
been isolated from palm fruit bunches and have high acid rice bran oil and observed that reduced
been shown to be capable of producing enzymes pressure and glycerol addition improved the
which can hydrolyse palm oil, Hiol et al. (1999). reduction of free fatty acids (FFA). De (2006)
Fruit damage creates a double problem because reviewed the different chemical, physical and
not only will endogenous lipase become activated enzymatic processes that could be used to deacidify
rice bran oils and concluded that in the non-
enzymatic processes, loss of yield was inevitable.
* Novozymes U.K. Ltd, 282 Chartridge Lane,
The study also compared the use of a lipase
HP5 2SG Chesham, Bucks, United Kingdom. with the autocatalytic reaction between FFA and
E-mail: dc@novozymes.com monoglycerides and concluded that the enzymatic

1492

D-Cowan.indd 1492 12/31/12 4:46 PM

 REDUCTION IN FREE FATTY ACIDS IN CRUDE PALM OIL BY ENZYMATIC REMEDIATION

reaction was more efficient and could be carried RESULTS AND DISCUSSION
out at a lower temperature. However, as the
enzymes used were predominantly 1.3 specific A preliminary experiment was carried out to
the regenerated material would be found in the examine the influence of enzyme dosage and type
oil in the form of diglycerides, which may not on FFA removal. The reduction in FFA was more
be optimum. The presence of partial glycerides complete when the non-specific lipase was used
and chlorine containing compounds has been and the speed of reduction was related to the
associated by some authors with increased levels of enzyme dosage (Table 1).
glycidyl esters and 3-MCPD (Matthaüs et al., 2011) The non-specific lipase reduced the FFA in the
therefore, the reduction in these components could oil more quickly due to the ability of this enzyme
be beneficial. However, as there are some disputes to attach a fatty acid at any position on the glycerol
about this linkage, this is not taken into account in backbone. The 1.3 specific lipase could not attach
any evaluation of the overall benefit of this process. the fatty acid at the Sn-2 position and this resulted
in a slower and incomplete re-attachment.
The influence of glycerol addition together with
MATERIALS AND METHODS the liquid formulated esterase on FFA reduction
and level of partial glycerides and triglyceride
Samples of crude palm oil were obtained from content is shown in Tables 2 and 3 respectively.
palm oil mills in Malaysia and were described as of Partial glycerides are expressed as the sum of
standard quality by the suppliers. In order to avoid monoglycerides (MG) and diglycerides (DG). The
changes in the FFA level upon storage, fresh samples observation that the addition of glycerol increases
were obtained for the individual experiments. the rate of FFA removal is in agreement with the
Palmitic acid and glycerol were of analytical grade results of Bhattacharyra and Bhattacharya (1989).
and were obtained from a local laboratory chemical When no additional glycerol was added,
supplier. Liquid and immobilised esterase from reduction in FFA was slower than with 1% or 10%
Candida antarctica B, were Novozymes standard addition, indicating that glycerol was a better
food grade products with the commercial names of acceptor of fatty acids than the partial glycerides. As
Lipozyme Calb L and Lipozyme 435, respectively. a result of this reaction, levels of partial glycerides
The 1.3 specific lipase from Rhizomucor miehei, increased when additional glycerol was included
was the Novozymes commercial product, Palatase in the reaction and this was related to the addition
20  000 litres. rate.
For the reduction of the FFA content of crude GC analysis of the triglyceride content of the oils
palm oil, 100 ml aliquots were heated to 70°C indicated that little interesterification of the oil had
and mixed with differing amounts of lipase in the occurred during the condensation of the fatty acids.
liquid or immobilised form and varying amounts In order to try and limit the increase in partial
of glycerol. The oil samples were then transferred glycerides, the condensation reaction was carried
into conical flasks and high shear mixing (24 000 out using the immobilised form of the C. antartica
rpm) was applied to each oil sample. In the case lipase. This version is completely free from glycerol
where immobilised C. antartica B enzyme was used, and as such the condensation reaction should be
this was added after the high shear mixing to avoid only between the FFA and the entrained partial
damage to the enzyme particles. The flasks were glycerides. Different dosages of the immobilised
incubated at 70°C, under vacuum and agitation enzyme were used and FFA, partial glyceride
applied via a magnetic stirrer operating at 350 rpm.
Samples were removed at intervals and heated
at 85°C for 30 min to inactivate the added enzyme TABLE 1.  REDUCTION OF % FREE FATTY ACIDS (FFA)
IN CRUDE PALM OIL FOLLOWING LIPASE ADDITION
before centrifugation at 3000 g for 15 min to AND 2% W/W GLYCEROL
separate the phases. The crude and treated oils
were analysed for FFA by titration and for mono, di Time 0.1% C. antarctica 1% C. antarctica 1.0% R. meihei
and triglycerides by gas chromatography. (hr) lipase lipase lipase
Fatty acids were determined by titration,  0 4.98 4.98 4.98
following the procedure described in the AOCS  1 3.86 1.29 4.13
Official Method F 9a-44 and the results presented  2 3.79 0.98 3.57
are the mean of two determinations. Gas  4 3.36 0.69 3.17
chromatography (GC) was performed using an  6 3.10 0.83 2.61
Agilent 7890A, fitted with a Quadrex 007-65HT,  8 2.37 0.46 NM
(65%Phenyl) methylpolysiloxane column and a
10 2.21 NM 2.38
flame ionisation detector. Results are expressed as
14 1.73 0.29 NM
area percentages of the samples as analysed.

1493

D-Cowan.indd 1493 12/31/12 4:46 PM

 JOURNAL OF OIL PALM RESEARCH 24 (DECEMBER 2012)

TABLE 2.  CHANGES IN FREE FATTY ACIDS (FFA) AND PARTIAL GLYCERIDES WITH DIFFERENT GLYCEROL
ADDITION RATES

Reaction time (hr)

Treatment
0 3 7
2% C. antarctica lipase % FFA 4.84 3.96 3.77
% MG + DG 8.80 9.00 8.60
2% C. antarctica lipase + 1% glycerol % FFA 4.84 2.82 2.23
% MG + DG 8.80 10.0 10.9
2% C. antarctica lipase + 10% glycerol % FFA 4.84 0.33 NM
% MG + DG 8.80 13.8 NM

Note: MG – monoglycerides.
DG – diglycerides.
NM – not measured.

TABLE 3.  CHANGES IN MARKER TRIGLYCERIDES TABLE 4.  INFLUENCE OF LIPASE DOSAGE ON FFA AND
FOLLOWING LIPASE REACTIONS PARTIAL GLYCERIDES LEVEL IN CRUDE PALM OIL

Triglyceride marker Treatment % FFA % MG + DG

Treatment
PPP POP POO Crude palm oil 5.01 5.97
Crude palm oil 7.13 29.09 18.72 CPO + 0.5% lipase 4.33 5.01
2% C. antartica lipase 6.85 29.60 18.58 CPO + 1.0% lipase 3.84 4.35
2% C. antartica lipase + 7.02 28.81 18.65 CPO + 2.0% lipase 3.43 3.66
1% glycerol CPO + 4.0% lipase 3.17 3.26
2% C. antartica lipase + 7.25 30.26 18.29 R2 (enzyme dosage) 0.79 0.81
10% glycerol
Note: CPO – crude palm oil. FFA – free fatty acid.
Note: P – palmitic acid. MG – monoglycerides. DG – diglycerides.
O – oleic acid.

and triglyceride contents determined after 7 hr TABLE 5.  INFLUENCE OF ENZYME DOSAGE ON
of reaction time. A reaction time of 7 hr with no TRIGLYCERIDE COMPOSITION OF CRUDE PALM OIL
intermediate sampling was chosen as previous Treatment % PPP %POP % POO
experiments, not reported here, had shown that
Crude palm oil 7.22 30.89 19.78
opening the reaction flask for sampling allowed for
CPO + 0.5% lipase 8.19 30.41 19.61
moisture uptake by the oils and hydrolysis reactions
CPO + 1.0% lipase 8.10 29.78 19.72
to re-start. The results of the FFA and partial
glyceride levels in CPO, following incubation for 7 CPO + 2.0% lipase 9.33 29.89 20.02
hr at 70°C, are shown in Table 4. CPO + 4.0% lipase 9.78 28.81 20.09
In this case increasing the enzyme dosage R2 (enzyme dosage) 0.87 0.90   0.47
resulted in the simultaneous reduction of FFA and Note: P – palmitic acid. O – oleic acid. CPO – crude palm oil.
partial glycerides, indicating that the condensation
reaction was taking place between these two
components. The reduction in FFA and partial on the resulting properties of the oil has yet to be
glycerides was correlated with the enzyme dosage established. As the triglyceride analysis is of the
over the tested range. overall areas percentage it is not possible to state
Analysis of the triglyceride content of the whether this change in composition is totally due
crude palm oil and the treated oils was made to to interesterification or to a change in composition
determine if interesterification had taken place following the re-synthesis of some triglycerides
during the condensation reaction and the results from the partial glycerides found in the oil.
for the marker components are shown in Table 5. Both lipases were able to reduce the FFA content
These results indicate that some interesterification of crude palm oil and therefore decrease losses in
had taken place which was related to the dosage of refining. When glycerol was added to the reaction,
enzyme used. However, this was not to the extent part of the FFA was attached to the glycerol,
that would have been observed using an enzyme resulting in an increase in partial glycerides. The
such as the Thermomyces lanuginosus lipase which rate of reaction between fatty acids and glycerol
will produce a randomisation of the palm oil. The was faster than that between fatty acids and
influence of this apparent limited randomisation diglycerides and therefore in the presence of glycerol

1494

D-Cowan.indd 1494 12/31/12 4:46 PM

 REDUCTION IN FREE FATTY ACIDS IN CRUDE PALM OIL BY ENZYMATIC REMEDIATION

there would be both a synthesis of diglycerides as interesterification and a reduction in the level of
well as triglycerides. This increase in the level POP. But when additional palmitic acid is present,
of diglycerides in the presence of glycerol may this reduction is largely eliminated.
be a disadvantage if the palm oil is to be used Three enzyme reactions can therefore be
directly for food purposes. However, if the oil is taking place at this time; condensation, resulting
to be used for biodiesel production, these partial in the reduction in partial glycerides and FFA;
glycerides can be readily converted by a chemical interesterification and possibly acidolysis; in which
or enzymatic methylation catalyst and an extra palmitic acid replaces some of the stearic acid in
yield obtained. Enzymatic remediation using liquid the triglycerides. Analysis of the FFA composition
lipase plus glycerol could therefore be applied of the remaining material should be carried out in
for the treatment of oils which are to be used for a further experiment to determine if acidolysis has
biodiesel or the production of oleochemicals. occurred. For practical application of this technique
When glycerol addition is not carried out, the the use of palm fatty acid distillate (PFAD) would
reaction rate is somewhat slower but no increase in supply a more balanced source of the required fatty
partial glycerides occurs. Interesterification of the acids, reflecting more accurately the composition
oils is limited suggesting that the resulting oil can of palm oil. The level of FFA and partial glycerides
be fractionated and used for food purposes. can be reduced by increasing the reaction time or
In order to maximise the reduction in partial increasing the enzyme dosage.
glycerides, one condensation reaction was carried The efficiency of water removal is also an
to compare the conversion in the presence of an important factor and as the condensation reaction
additional 3% of palmitic acid. By providing more continues, the equilibrium is more and more in
fatty acid, the reaction equilibrium between the the direction of hydrolysis. If water removal is not
partial glycerides and the triglycerides should be efficient, the FFA will not be able to re-combine
changed and the level of diglycerides reduced. with the partial glycerides and form triglycerides.
When an additional 3.0% of palmitic acid was On the laboratory scale, water removal efficiency
added to the reaction, partial glycerides decreased is affected by the volume contained within the
from 6.08% in the crude oil to 2.32% compared to reaction vessel. Increasing the amount of reactants
3.66 % in the crude oil treated only with lipase (Table decreases the surface area to volume and can be
6). By providing an excess of fatty acids the reaction expected to reduce the efficiency of water removal.
kinetics were altered towards greater condensation This was examined by comparing a condensation
with the mono and diglycerides present in the oil. reaction with 100, 280 or 500 g of crude palm oil
As the analysis method only reveals percentage and 4% w/w immobilised esterase. Following
area composition, it is not possible to conclude condensation with 4% w/w esterase the FFA level
whether these changes are a result of an in the palm oil had fallen from 4.98% to 4.06%, 3.43%
overall increase in triglycerides, or if limited and 2.41% respectively and the sum of mono and
interesterification has taken place or both. Compared diglycerides to 5.2%, 3.7% and 2.8% respectively
to esterase alone, esterase plus palmitic acid from a starting level of 6.1%. Thus, water removal
results in a relative increase in palmitic containing efficiency is also a key parameter in promoting the
triglycerides and a decrease in stearic. This would condensation reaction and needs optimisation.
be expected as palmitic acid is incorporated into Analysis of the triglyceride composition of these
the partial glycerides, raising the relative amount of different oils after condensation is shown in Table 7.
these components. However, as observed in Table 5, While overall FFA and partial glyceride content was
addition of 4% w/w esterase alone, results in some influenced by the surface area to volume ratio, it was

TABLE 6.  INFLUENCE OF ADDED FFA ON CONDENSATION REACTION

Crude palm oil Treated with lipase Treated with lipase and palmitic acid
Starting FFA level (titration)   4.98   4.98   7.80
FFA after reaction (titration)   3.43   3.43   5.59
% MAG + DAG   6.08   3.66   2.32
% PPP   7.23   9.72   9.85
% POP 29.18 25.48 28.03
% POO 20.12 19.31 19.41
% POS   4.51   4.25   3.81
% SOS   0.54   0.64   0.37
% SOO   1.76   1.67   1.57

Note: FFA – free fatty acid. DAG – diacylglycerol. S – stearic acid. MAG – monoacylglycerol. P – palmitic acid. O – oleic acid.

1495

D-Cowan.indd 1495 12/31/12 4:46 PM

 JOURNAL OF OIL PALM RESEARCH 24 (DECEMBER 2012)

TABLE 7.  INFLUENCE OF WATER REMOVAL ON FFA formulations should also be considered if the
REDUCTION reaction between glycerol and fatty acids should be
  CPO Oil amount used (g) avoided.
In determining the economy of the reaction, a
Oil/ g -   100   280   500
number of factors need to be taken into account,
%FFA (by titration   4.98   2.41   3.43   4.06
method)
including the cost of the enzyme and the operation
of the process. In addition, there is the differential
%MG + %DG   6.08   2.76   3.66   5.23
(by GC analysis) value between refined palm oil and PFAD. Re-
% PPP   7.23 10.38   9.72 10.80
using the enzyme is one means by which the cost of
conversion can be reduced. An immobilised enzyme
% POP 29.18 25.35 25.48 23.96
may be separated from a reaction by filtration and
% POS   4.51   4.20   4.25   3.63
recovered or used in a packed bed reactor. In either
% POO 20.12 19.64 19.31 17.97
case, re-use of the enzyme is possible, reducing the
% SOS   0.54   0.77   0.64   0.47 overall enzyme cost in the process.
Note: FFA – free fatty acid. DG – diglycerides. GC – gas Removing FFA from oils destined for biodiesel
chromatography. MG – monoglycerides. P – palmitic production, while at the same time rendering
acid. O – oleic acid. S – stearic acid. them capable of being converted to methyl esters
will simplify and improve the biodiesel process.
The avoidance of loss here will also improve the
not possible to see a significant connection between overall economics and sustainability of this biofuel
triglyceride content and oil amount. However, for production.
the lowest surface area to volume ratio sample the As partial glycerides have been associated with
degree of interesterification showed a tendency to the formation of 3-MCPD and glycidyl esters, their
be greater. As this would be the sample with the removal before deodorisation could help to reduce
lowest degree of re-incorporation of FFA, these the formation of these unwanted compounds.
results tend to support the hypothesis that both
condensation and interesterification can occur, and
the relative importance of these vary according to REFERENCES
the conditions of the reaction.
BHATTACHARYA, S and BHATTACHARYA, D K
(1989). J. Amer. Oil Chem. Soc. Vol. 66 No.12: 1809-
CONCLUSION AND RECOMMENDATIONS 1811.

Immobilised and soluble preparations of C. antartica BOCK (2012). http://theoilpalm.org/wp-

lipase have been applied to the condensation of fatty content/themes/oilpalm/pdf/GOFB%202-4%20
acids with partial glycerides and/or glycerol. In both Understanding%20Palm%20Oil.pdf
cases, the FFA content of the oil can be recovered
and losses considerably reduced. Interesterification COWAN, D and HOLM, H C (2008). Eur. J. Lipid Sci.
of the palm oil is effectively zero with the liquid Technol, 110: 679-691.
preparation and limited compared to T. lanuginosus
lipase with the immobilised version of the esterase. DE, B K (2006). J. Amer. Oil Chem. Soc. Vol. 83 No. 5:
C. antartica is known to have a lower activity rate in 443-448.
carrying out interesterification than the lipase from
T. lanuginosus (Cowan and Holm, 2008). HIOL, A; JONZO, M D; RUGANI, N; DRUET, D;
Enzymatic remediation could be a promising SARDA, L and COMEAU, L C (2000). Enzyme and
tool for the improvement of processing yields in Microbial Technology Vol. 26 Issues 5-6 (March 2000):
palm or other oils where fatty acids are present. 421-430.
Further studies are required to optimise the FFA
reduction while at the same time minimising MATTHÄUS, B; PUDEL, F; FEHLING, P;
interesterification or at least determining its effect VOSMANN, K and FREUDENSTEIN, A (2011). Eur.
on fractionation of the palm oil. Alternative enzyme J. Lipid Sci. Technol. Vol. 113 No. 3: 380-386.

1496

D-Cowan.indd 1496 12/31/12 4:46 PM