Faculteit Bio-ingenieurswetenschappen Academiejaar 2011 – 2012 Production of Cocoa Butter Equivalent through Enzymatic Acidolysis Lynn Naessens Promotor: Prof. dr. ir. Koen Dewettinck Tutor: Sheida Kadivar Masterproef voorgedragen tot het behalen van de graad van Master in de bio-ingenieurswetenschappen: Levensmiddelenwetenschappen en voeding
100
Embed
Production of Cocoa Butter Equivalent through Enzymatic ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Faculteit Bio-ingenieurswetenschappen
Academiejaar 2011 – 2012
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis
Lynn Naessens
Promotor: Prof. dr. ir. Koen Dewettinck Tutor: Sheida Kadivar
Masterproef voorgedragen tot het behalen van de graad van Master in de bio-ingenieurswetenschappen: Levensmiddelenwetenschappen en voeding
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis I
The author, promotor and tutor give permission to use this thesis for consultation and to copy parts
for personal use. Any other use falls under the copyright laws: the source must be correctly specified
when results of this thesis are used.
De auteur, promotor en tutor geven toelating om deze thesis te gebruiken voor consultatie en om
bepaalde delen te kopiëren voor persoonlijk gebruik. Elk ander gebruik valt onder het auteursrecht:
de bron moet uitdrukkelijk en correct vermeld worden als resultaten uit deze thesis worden gebruikt.
Ghent, June 2012.
The promotor The tutor
The author
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis II
Woord vooraf Deze masterproef vormt de afsluiter van 5 jaar studeren aan de faculteit bio-
ingenieurswetenschappen te Gent. De afgelopen 5 jaar waren zeker niet de makkelijkste, maar met
de steun van veel vrienden en familie ben ik ze toch heelhuids doorgekomen. Daarom wil ik zeker
een moment nemen om al deze mensen te bedanken om er voor mij te zijn, altijd te willen luisteren
en mijn gedachten af te leiden van schoolwerk wanneer ik het nodig had.
Vooraleerst wil ik een aantal mensen bedanken voor de hulp en steun tijdens de realisatie van mijn
masterproef. Mijn promotor, Prof dr. ir. Koen Dewettinck, wil ik bedanken om me de kans te geven
deze thesis in de vakgroep FTE te realiseren. Sheida, bedankt voor alle hulp en uitleg bij elk nieuw
experiment. Bedankt dat je elke keer heel snel en met veel geduld mijn vragen beantwoordde en om
zoveel tijd te spenderen in het nalezen van elk deel in deze masterproef. Alle mensen van de
vakgroep FTE wil ik bedanken om altijd klaar te staan met een woordje uitleg bij de vragen die ik had,
de hulp bij het verwerken van gegevens en waar ik bepaalde zaken kon vinden. Ook mijn vele
thesiscollega’s wil ik bedanken voor de vele praatjes tussen de experimenten in.
Ook gaat een woord van dank uit naar het bedrijf Oleon te Antwerpen. Vooral Marjan verdient de
vermelding in deze masterproef voor de tijd die ze vrijmaakte om mij het SPD proces heel geduldig
en vriendelijk uit te leggen.
De jaren op ‘het boerekot’ werden zeker aangenaam gemaakt door de vele nieuwe en ongelooflijk
leuke mensen die ik hier heb leren kennen. De toffe sfeer die we gedurende de afgelopen jaren
gecreëerd hebben, zal mij altijd bijblijven.
Mijn vrienden uit Aalter, ik wil jullie bedanken om mijn gedachten te verzetten elke keer we samen
kwamen om bij te praten of iets leuks te ondernemen.
Ook mijn ouders wil ik bedanken voor de steun die ze mij geven in alles wat ik doe en om mij de kans
te geven alles te doen wat ik maar wil.
Last but not least, mijn vriend Roberto; tijdens het realiseren van de masterproef heb je mij altijd
gesteund en geholpen waar je kon. Nu wordt het tijd dat ik wat meer aandacht aan jou besteed.
Gent, juni 2012
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis III
Table of Contents Introduction ............................................................................................................................................ 1
Literature ................................................................................................................................................ 2
1. Modification of fats and oils ............................................................................................................ 2
General conclusions .............................................................................................................................. 74
Further research ................................................................................................................................... 76
Figure 14: The % TAG (POP, POSt and StOSt) using different temperatures. ....................................... 33
Figure 15: The % TAG (POP, POSt and StOSt) and the amount of MAG + DAG formed, using different
water contents. ................................................................................................................................... 34
Figure 16: The % TAG (POP, POSt and StOSt) using different enzyme loads. ....................................... 35
Figure 17: The percentage of the TAG (POP, POSt and StOSt) for different substrate ratios. .............. 36
Figure 18: The percentage of FFA, DAG and TAGs (A) and a detail of the desired TAGs (POP, POSt and
StOSt) (B) for different ratios of glycerol added to HOSO. ................................................................ 37
Figure 19: The percentage of FFA (A) and DAG (B) for different ratios of glycerol and non-specific
enzyme added to HOSO. ..................................................................................................................... 39
Figure 20: The percentage TAG (POP, POSt and StOSt) for different ratios of glycerol and non-specific
enzyme. Method A (A), method B (B) and method C (C).................................................................... 40
Figure 21: Perturbation plot of SUS (left) and SUU (right) with A: substrate ratio (7 mol); B: enzyme
load (10%); C: water content (2%); D: temperature (65°C) and E: reaction time (6h). ...................... 43
Figure 22: Contour plot of the interaction between enzyme and temperature on % SUS (left) and %
SUU (right). ......................................................................................................................................... 44
Figure 23: Contour plot of the interaction between water and temperature on % SUS (left) and % SUU
Figure 24: Contour plot for the optimal factor levels of % SUS (left) and % SUU (right). ..................... 45
Figure 25: Scheme of the obtained fractions after fractionation using method A or B. ....................... 47
Figure 26: Percentage of SSS TAGs in SF1 (A), UUU and SUU TAGs in OF2 (B) and SUS, UUU and SUU
TAGs in SF2 (C) compared between two fractionation methods. ...................................................... 48
Figure 27: Non-isothermal crystallization and melting profile of the interesterified product (product)
and the purified product (after SPD) as measured by DSC. ................................................................ 51
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis IX
Figure 28: Non-isothermal crystallization and melting profile of the fractions after fractionation
method A as measured by DSC. .......................................................................................................... 51
Figure 29: Non-isothermal crystallization and melting profile of the SF2 fractions after fractionation
method A and B as measured by DSC. ................................................................................................ 52
Figure 30: Non-isothermal (non-tempered and tempered) SFC curve of the product, purified product,
SF2 (CBE) and CB as measured by pNMR. .......................................................................................... 53
Figure 31: Percentage of POP, POSt and StOSt (A) and SUU, SSS (B) TAGs in the CB/ CBE mixtures. .. 55
Figure 32: POP/POSt/StOSt ternary diagram showing the position of CB, vegetable fats used as CBE
and the enzymatically produced CBE (Padley et al., 1981; Smith, 2001). .......................................... 56
Figure 33: Percentage of SSS, SUS, SSU and SUU TAGs in the CB/ CBE mixtures; results obtained by
silver ion HPLC. ................................................................................................................................... 57
Figure 34: Non-isothermal crystallization and melting profile of CB and the mixtures with CBE as
measured by DSC. ............................................................................................................................... 58
Figure 35: Non-isothermal non-tempered (A) and tempered (B) SFC curve of the CBE-CB mixtures as
measured by pNMR. ........................................................................................................................... 60
Figure 36: Non-isothermal SFC curve: comparison of tempered and non-tempered CB and pure CBE
as measured by pNMR. ....................................................................................................................... 61
Figure 37: SFC melting curves indicating the hardness (A), heat resistance (B) and waxiness (C) of CB
and mixtures with CBE (Depoortere, 2011). ....................................................................................... 62
Figure 38: Isothermal diagram of the mixtures of CBE and CB. ............................................................ 63
Figure 39: Isothermal crystallization of CB at 20°C as measured by DSC. ............................................. 64
Figure 40: Isothermal crystallization at 20°C of the different ratios of CB and CBE. ............................ 65
Figure 41: Influence of Sat FA to Unsat FA and SUS to SUU TAGs ratios on aF and tind. ........................ 67
Figure 42: Isothermal crystallization at 20°C of CB and mixtures with CBE as measured by pNMR. ... 69
Figure 43: Isothermal crystallization at 20°C after 60 min as visualized by PLM: 0 (a), 20 (b), 40 (c), 60
(d), 80 (e) and 100% (f) CBE. ............................................................................................................... 71
Figure 44: Isothermal crystallization at 20°C as visualized by PLM for CB, 20, 40, 60, 80 and 100% CBE.
The microstructure is given after 24h, 2 weeks, 4 weeks and 6weeks. ............................................. 73
Figure 45: Isothermal crystallization at 20°C after 1 min as visualized by PLM: 0 (a), 20 (b), 40 (c), 60
(d), 80 (e) and 100% (f) CBE. ............................................................................................................... 86
Figure 46: Isothermal crystallization at 20°C as visualized by PLM for CB, 20, 40, 60, 80 and 100% CBE.
The microstructure is given after 1 week, 3 and 5 weeks. ................................................................. 87
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis X
List of tables Table 1: Overview of the CBAs (Depoortere, 2011). ............................................................................... 7
Table 2: Vegetable fats allowed to use as CBE in chocolate according to EU Directive 2000/36/EC. .... 8
Table 3: An overview of the enzymatic interesterification in order to produce CBEs (Depoortere,
concentrates, etc. are produced (Xu et al., 2006).
TAGs, which are modified, are called structured TAGs. Nowadays, many nutritional and functional
specific-structured triacylgycerols (SSTs) are produced. The nutritional and functional properties are
due not only to the fatty acid profile, but also to the fatty acid distribution on the glycerol backbone.
The specific characteristics of breast milk fat and cocoa butter (CB) are due to the fatty acid
distribution on the glycerol backbone. Because of the high specificity that is required to produce the
SSTs, it is impossible to apply chemical methods and only lipases with a high regiospecificity are used
(Xu, 2000).
Fractionation, hydrogenation, and interesterification are three different methods that are used to
modify fats and oils. Also blending different types of oils is a way to obtain a product with desired
characteristics.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 3
1.1 Fractionation
Fractionation is the method of physically separating high melting point fractions (stearin) from low
melting point ones (olein). Fractionation is a thermo-mechanical separation process and is
completely reversible (Kellens et al., 2007). Palm oil is the most common oil which is fractionated
quite often. Normally, palm oil can be fractionated into palm olein, palm stearin, palm mid fraction
(PMF) and so on. The separation is based on the difference in melting points of the fractions (Xu,
2000).
In the first step of the fractionation process, the fat sample should be melted completely, followed
by a slow cooling which is very gentle and without agitation in order to obtain large crystals. If a fast
cooling would be applied, it would result in small fat crystals. To obtain a good separation, the
crystals should be firm and of uniform size. During the cooling and crystallization, which is a two-step
process involving nucleation and crystal growth, the viscosity of the solution will increase (Kellens et
al., 2007). This crystallization is done in a controlled way so the fraction with the highest melting
point will crystallize first and the other fraction will still be liquid because of the lower melting point.
The liquid and solid fractions have different physical and chemical compositions and the two
fractions can be separated using filtration (Wassell et al., 2010).
There are three types of fractionation: dry fractionation, solvent fractionation and detergent
fractionation. Dry fractionation is the simplest, cheapest and a ‘green’ method because no solvents
are used and there are no losses of product. Viscosity limits the use of dry fractionation in a single
step because it restricts the degree of crystallization. That is why this method is mostly done in a
multi-step operation (Kellens et al., 2007). Solvent fractionation requires solvent and in most cases;
hexane, acetone or 2-nitro-propane is used (Bockisch, 1998). This type of fractionation is more
efficient but has higher operating costs. The other disadvantage of this method is the oil entrained
between the crystals. This can be removed by centrifugation, vacuum filtration or pressing (Salas et
al., 2011).On the other hand, the advantages of using solvents are a faster nucleation and growth of
the crystals, a lower viscosity which leads to easier filtration, a dilution of the fat that makes the heat
transfer easier and the possibility to wash the crystals repeatedly with the solvent to reduce the
amount of entrained oil (Timms, 2005). The first step in detergent fractionation, is the fractional
crystallization followed by adding water containing an aqueous detergent such as sodium lauryl
sulphate and an electrolyte (magnesium sulphate or sodium sulphate). The crystals become
dispersed in the solution and the electrolyte facilitates the agglomeration of the oil droplets during
mixing. Finally, the crystal separation is completed by centrifugation (Rajah, 1996).
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 4
1.2 Hydrogenation
To get a specific and desired melting behavior of fat blends, hydrogenation or partial hydrogenation
processes can be used. This technique is mostly applied to give firmness to margarines, plasticity and
emulsion stability to shortenings (Wassell & Young, 2007). Unsaturated fatty acids have double
bounds and are usually in the cis configuration. Hydrogenation, which adds hydrogen atoms to the
double bounds, results in a higher degree of saturation and on top of that, a more rigid structure of
the TAG and a higher melting point are obtained. But the disadvantage of the hydrogenation is that
the cis isomers can be changed into the trans ones and these trans fatty acids have negative health
effects (Wassell et al., 2010).
1.3 Interesterification
In general, it is a process that results in the rearrangement of the distribution of the fatty acids on
the glycerol backbone. Interesterification is an alternative to hydrogenation but without the
formation of trans fatty acids (Wassell & Young, 2007). It can be done in a chemical or enzymatic way,
and within or between TAGs. There are different possibilities of enzymatic interesterifications:
alcoholysis, acidolysis and ester-ester exchange. The enzymes, which are used, can be specific or
non-specific (Wassell et al., 2010).
1.4 Blending
Another way to modify fats and oils is blending oils with fully hardened ones to obtain a product with
desired physical characteristics. Blending vegetable oils from different sources is an alternative for
the hydrogenation of vegetable oils but with the right physico-chemical properties and nutritional
requirements that are demanded. Another advantage of this technique is that there is no chemical
modification. However, a disadvantage is that the blending of the right amounts of oil is often a
process of trial and error (Wassell & Young, 2007).
2. Cocoa butter
Chocolate contains many ingredients of which CB is the most important. It is the most expensive
ingredient and one third of the cost of chocolate is due to the CB (Widlak, 1999).
Some of the unique characteristics of chocolate, for instance, the viscosity and the rheological
properties depend on the crystallization of the CB. Also, the snap and the surface gloss of the
chocolate are due to the TAG composition of the CB (Widlak, 1999). CB is responsible for the
brittleness at room temperature, the cooling effect in the mouth, and for the quick and complete
melting of the chocolate around body temperature (27-33°C). However, the TAG composition of CB
can vary depending on the geographical source.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 5
2.1 Chemical properties
CB consists mainly of three fatty acids: palmitic acid (C16:0, 20 - 26%), stearic acid (C18:0, 29 - 38%),
and oleic acid (C18:1, 29 - 38%). Linoleic acid (L, C18:2, 2 - 4%) and arachidic acid (A, C20:0, 1%) are
also present in considerable amounts. More than 70% of its TAGs are symmetrical with oleic acid (O)
at the sn-2 position. The three most important TAGs are POP (21%), POSt (40%) and StOSt (27%). The
difference in amount of these fatty acids and TAGs are due to the origin of the CB (Xu, 2000; Talbot,
2009b; Smith, 2001).
2.2 Physical properties
Because of the composition of the TAGs in the CB, it can crystallize into 5 or 6 polymorphic forms,
depending on the reference. According to Van Malssen et al. (1999), when using the classification,
form V and VI are the most stable ones and these do not crystallize directly from the melt. Therefore,
a recrystallization from a metastable polymorphic form is necessary as presented in figure 1. The
desired polymorphic form of chocolate is the βV-polymorph. βVI is the most stable polymorphic form
which is typical for fat bloom (Timms, 2003). Also, the typical melting point of the CB depends on the
polymorphic form and this can range from 15 to 36°C (Huyghebaert, 1971).
Figure 1: Polymorphic transitions of CB (Van Malssen et al., 1999).
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 6
3. Cocoa butter alternatives
The discussed techniques to modify oils can also be used to produce alternatives to CB starting from
vegetable oils. As mentioned earlier, CB is an expensive ingredient and prices can be unpredictable,
also the supplies can be uncertain. Thus, for economical and technical reasons, producers are forced
to seek alternatives to replace the CB (Xu, 2000).
Vegetable fats can be used as alternatives to CB in chocolate. These replacer fats are called cocoa
butter alternatives (CBA). CBA can be divided into three categories: cocoa butter equivalents (CBE),
cocoa butter substitutes (CBS), and cocoa butter replacers (CBR). The CBAs are mostly mixtures of
different vegetable fats. The CBEs are non-lauric fats with similar physical and chemical properties as
CB. They can be mixed with the CB without changing the physical properties of it. The CBR are also
non-lauric fats with a similar fatty acid distribution but a completely different structure of TAGs to CB.
Finally, the CBS are lauric plant fats that are chemically totally different to CB but with some physical
similarities. The CBS and CBR are found in compound chocolate of which the fat phase contains other
fats than real CB, for instance in chocolate-coatings and ice-cream (Lipp & Anklam, 1998; Smith, 2001;
Stewart & Timms, 2002).
CBA should be compatible with CB in brittleness and melting behavior. The melting and
crystallization characteristics are mainly due to the TAG composition. Thus, if a substitute of CB is
produced, several aspects such as the melting behavior will be crucial. The melting behavior has to
be very similar to that of CB in order to achieve the same mouth feeling and the addition of the CBA
should not change the crystallization of the CB (Lipp & Anklam, 1998). Table 1 gives an overview of
the classification of CBAs.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 7
Table 1: Overview of the CBAs (Depoortere, 2011).
CBE CBR CBS
Origin
Illipé butter
Palm oil
Sal fat
Shea butter
Kokum butter
Mango kernel fat
Palm oil
Soybean oil
Rapeseed oil
Cottonseed oil
Palm kernel oil
Coconut oil
Processing
Hydrogenation
Fractionation
Interesterification
Hydrogenation
Fractionation
Hydrogenation
Fractionation
Interesterification
TAG composition Similar to CB
Different from CB Different from CB
Lauric acid Non lauric
Non lauric Lauric
(45-55% lauric acid)
Compatibility to CB Compatible Compatible in small
ratios
Incompatible
Crystallization
Tempering to obtain
stable polymorphic
form
Crystallize directly
from the melt in the
stable polymorphic
form
Crystallize directly
from the melt in the
stable polymorphic
form
Application
5% replacement on
total product
Compound
Compound
Compound
Remark
Cocoa butter extender
(CBEX)
Cocoa butter improver
(CBI)
High level of trans fatty
acids
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 8
3.1 Cocoa butter equivalents
One type of CBAs are CBEs which are totally compatible with CB and, for this reason, it can be mixed
with the CB without any problem. Therefore, a lot of research is done for CBEs because they have the
closest characteristics to CB (Xu, 2000). The CBEs can be divided into two groups: the cocoa butter
extenders (CBEX) and the cocoa butter improvers (CBI). CBEX are mixable with CB but not in every
ratio. The CBI have a high content of StOSt and this will increase the SFC. Due to the higher amount
of solid fat, the melting point and the hardness is increased. Chocolate with CBI has a better
resistance to softness and formation of fat bloom at higher ambient temperatures; for example in
summer or in tropical regions (Timms, 2003).
3.1.1 Legislation
The European Union has established the EU Directive 2000/36/EC relating to cocoa and chocolate
products intended for human consumption. Up to now, only 5% vegetable fats other than CB are
allowed in chocolate in some of the Member States of the European Union. These vegetable fats
should be CBEs and therefore be defined according to the technical and scientific criteria and meet
the following criteria before they can be used in chocolate (EU Directive 2000/36/EC).
They are non-lauric vegetable fats, which are rich in symmetrical monounsaturated TAGs of
the type POP, POSt and StOSt.
They are miscible in any proportion with CB and are compatible with its physical properties
(melting point and crystallization temperature, melting rate, need for tempering phase).
They are obtained only by the processes of refining and or fractionation which excludes
enzymatic modification of the TAG structure.
Table 2 gives the 6 vegetable fats that are allowed to use as a CBE in chocolate according to the EU
Directive.
Table 2: Vegetable fats allowed to use as CBE in chocolate according to EU Directive 2000/36/EC.
Name of the vegetable fat Scientific name of the plants from which the fat
can be obtained
Illipé, Borneo tallow or Tengkawang Shorea spp.
Palm oil Elaeis guineesis, Elaeis olifera
Sal Shorea robusta
Shea Butyrospermum parkii
Kokum gurgi Garcinia indica
Mango kernel Mangifera indica
As an exception, coconut oil can be used in chocolate but only for the manufacture of ice cream and
similar frozen products.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 9
If the previously mentioned vegetable fats are used in chocolate products, the consumer should be
informed correctly and objectively. They should be mentioned in the list of ingredients and the
product should be labeled with: ‘contains vegetable fats in addition to cocoa butter’ (EU Directive
2000/36/EC).
Countries outside the EU have their own regulations and these can differ from the European
legislation. For instance, it is not permitted in the United States to use vegetable fats other than CB in
chocolate, but the American legislation allows the use of it in chocolate coatings and vegetable fat
coatings. There are also countries where more than 5% of vegetable fats can be used in chocolate
but the products cannot be labeled as ‘chocolate’. (Talbot, 2009b)
3.1.2 Sources
The fats that can be used to produce CBEs are mentioned in table 2 and these are palm oil, illipé,
shea and also sal, kokum gurgi and mango kernel. In other words, these are the types of fats that are
allowed by the EU.
3.1.2.1 Palm oil
Palm oil is obtained from the flesh of the fruit of Elaeis guineensis and it is mostly produced from
trees in Malaysia or Indonesia. The fatty acid composition of palm oil and PMFs can be typical for a
specific region but it mostly consists of P and O. To make the fatty acid composition closer to CB, the
PMF can be interesterified with P or St. Palm oil can be fractionated in palm olein and palm stearin.
In addition, the content of the different TAGs depends strongly on the fraction and the used
technique to obtain that fraction but in general mainly POP and POO are found (Lipp & Anklam,
1998).
3.1.2.2 Illipé butter
Other names for Illipé fat are Borneo tallow, engkabang or tenkgawang. The fat is obtained from the
seed kernels of the Shorea stenoptera, this tree grows in Borneo, Java, Malaysia and the Philippines.
The fatty acid composition of the fat resembles somewhat CB because of the high St content. The
amount of O and St in the illipé fat is more or less equal followed by P. It has a relatively high level of
POSt and StOSt. Before using this fat, it needs to be refined (Lipp & Anklam, 1998; Storgaard, 2000).
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 10
3.1.2.3 Kokum butter
Kokum butter is also called Goa butter, it is obtained from the seed kernels of the Garcinia indica or
the Kokum tree. This is an evergreen tree that grows in the tropical forests of India. Solvent
extraction is mostly used to obtain the oil from the seeds. The butter consists of high amounts of St,
followed by O (Lipp & Anklam, 1998). When kokum is interesterified with P and/or St, it was claimed
to resemble CB very well in both the fatty acid and TAG composition, as well as the melting behavior
(Sridhar et al., 1991). The TAG composition of kokum butter consists mostly of StOSt (77%), StOO
(12%) and POSt (8%). Before using it, the fat has to be refined (Sridhar & Lakshminarayana, 1991).
3.1.2.4 Mango kernel fat
This fat is obtained from the seed kernels of the fruit of the mango tree or Mangifera indica. Solvent
extraction is necessary to release the fat because only 6 to 15% fat is present in the kernel. The most
common TAG is StOSt which accounts for 40.6% of the TAG content. To obtain higher levels of StOSt,
solvent fractionation is used and after this, a refining process of the fat is needed (Timms, 2003).
3.1.2.5 Sal fat
Other names for Sal fat are Borneo tallow or tenkgawang tallow. It is obtained from the seed kernels
of Shorea robusta which grows in Borneo, Java, Malaysia and the Philippines. In older references, this
fat is often confused with Illipé. The fatty acid composition has some resemblance to CB because of
the high amount of St. It is also high in O, followed by P. These fats resemble CB very closely due to
their fatty acid composition, of which is about 33% O, the same amount of St and about 24% P. On
top of that, Sal fat contains a considerable amount of A. This makes the most common TAGs in this
fat; StOSt (42-52%) and StOA (20%). Also Sal fat needs to be refined and fractionated before using it
as a CBE (Lipp & Anklam, 1998).
3.1.2.6 Shea butter
Shea butter is also called Karite butter or Galam butter. It is obtained from the nuts of the tree
Butyrospermum parkii which is mainly found in West Africa. The geographical origin has a big
influence on the fatty acid composition of this fat but it mainly consists of O and St. The main TAG in
Shea is StOSt, so after fractionation, this fraction is mostly used to produce CBEs. Next to this
fractionation step, refining needs to be done before it can be used as a CBE (Lipp & Anklam, 1998).
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 11
3.1.3 Production
Nowadays, for the modification of lipids in order to produce CBEs, enzymes are more used than
chemical methods. A liquid enzyme can be used or an immobilized enzyme (the enzyme is coated in a
monolayer on a solid particle). According to the legislation, the use of enzymatic produced CBEs is
not allowed within the European Union, nevertheless, a lot of research now is based on the use of
enzymes. Such enzymes or hydrolases, need a lot of water in the system for the hydrolysis of the
fatty acids from the TAGs. But in an environment with only a very small amount of water, these
enzymes can also be used to catalyze the reverse reaction; this is the so called esterification. Next to
the esterification, other reactions, called the interesterification reactions, are used to produce the
CBEs. The interesterification is the reaction between an ester and a fatty acid, an alcohol or another
ester. Different interesterification techniques are alcoholysis, acidolysis and ester-ester exchange.
The enzymatic interesterification reaction is a two step mechanism: hydrolysis and esterification
(Rozendaal & Macrae, 1997).
3.1.3.1 Hydrolysis
In nature, enzymes perform hydrolysis; this means they convert TAGs into DAGs and in the final step,
monoacylglycerols (MAGs) and free fatty acids (FFA) are formed. Thus, the hydrolysis of oils and fats
involves three steps from TAG to glycerol and FFA. In figure 2, this process is given schematically (Xu,
2003).
Figure 2: Steps of the enzymatic hydrolysis of fats and oils (Xu, 2003).
3.1.3.2 Esterification
This is the inverse reaction of hydrolysis and is only possible in an environment with a very small
amount of water, a so called micro-aqueous reaction system. In this system, the hydrolysis is
minimized while in water abundant systems, the hydrolysis is the main reaction. Esterification is
actually the simple reaction between an organic acid and an alcohol. It is the condensation of FFA on
the glycerol backbone. As can be seen in figure 3, water is one of the products that are formed during
the reaction. Therefore, it is very important to remove the water to shift the thermodynamic
equilibrium to the synthesis. The reaction can be carried out in systems using solvents or in solvent-
free systems (Rodrigues & Fernandez-Lafuente, 2010).
Figure 3: The enzymatic esterification (Xu, 2003).
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 12
The water, which is produced during the reaction, can shift the reaction equilibrium towards
hydrolysis if it is not continuously removed. On the other hand, water cannot be entirely removed
because a certain amount of water is necessary to maintain a high enzyme activity, this requires a
higher aw range while a high product yield requires aw as low as possible. One of the options to
remove water from the system is bubbling dry air or nitrogen gas in the reactor (Oh et al., 2009).
Another possibility are molecular sieves.
3.1.3.3 Alcoholysis
The alcoholysis technique is also performed using enzymes. It is the reaction between an ester and
an alcohol. Chemical alcoholysis is used in industry to produce MAG, DAG and biodiesels. Using
enzymes gives several advantages because of their high specificity. The reaction is schematically
shown in figure 4. The ester can be acylglycerols or TAGs, and the alcohol can be glycerol, methanol
or ethanol (Xu, 2003).
Figure 4: The enzymatic alcoholysis (Xu, 2003).
3.1.3.4 Enzymatic acidolysis
Another option of the interesterification method to modify fats and oils, is the enzymatic acidolysis.
This reaction involves an ester and an acid, the acid will be exchanged with another acid in the ester.
The enzymatic acidolysis is widely used for the production of CBEs. The reactions are catalyzed by sn-
1,3 specific lipases because the positional specificity is essential for the final product. This is clearly
shown in figure 5. In this figure, a sn-1,3 specific lipase is used, this means that the lipase will only
change the fatty acids on the first and third position of the TAG, in other words, the FFA (Y) will only
be implemented on position 1 and/or 3 (Esteban et al., 2011).
Figure 5: The enzymatic acidolysis between a TAG (XXX) and a FFA (Y) (Xu, 2003).
The used ester doesn’t always have to be a TAG as shown in figure 5, also DAG and MAG can be used
as ester.
A disadvantage is that the DAGs, which are formed during the reaction, can cause side reactions and
produce some by-products (Pacheco et al., 2010). Figure 6 gives an example of this problem.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 13
Figure 6: The main reactions and side reactions of the enzymatic acidolysis for a TAG (LLL) and a FFA (M and L) using a sn-1,3 specific lipase (Xu, 2003).
Several factors affect the formation of by-products and acyl migration during the reaction. These
factors are an increase in temperature, reaction time, water content and water activity. There is also
a positive correlation found between the enzyme content and the acyl migration because the carriers
of immobilized lipases induce acyl migration. This carrier can be a resin or silica (Hoy & Xu, 2001).
Acyl migration is a disadvantage when CBEs are produced, it is not wanted that the oleic acid on the
sn-2 position shifts to another position on the glycerol backbone (Rodrigues & Fernandez-Lafuente,
2010). To reduce the acyl migration, one option is to use packed enzyme bed reactors instead of the
stirred tank reactors.
3.1.3.5 Ester-ester exchange
In the product, it is also possible to have an ester-ester exchange between two TAGs. Again, a sn-1,3
specific lipase is used and the fatty acids on the positions 1 and 3 will be exchanged. This reaction is
schematically presented in figure 7 (Xu, 2003).
Figure 7: The ester-ester exchange reaction between two TAGs (XXX and YYY) with the help of a sn-1,3 lipase. X and Y are two types of fatty acids (Xu, 2003).
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 14
4. CBE production
The use of lipases has several advantages over chemical catalysts. One of those advantages is that
the enzymes produce less by-products. Other advantages are lower energy consumption and better
product control. However, one of the major benefits of using lipase for the production of CBEs is the
regiospecificity of the enzymes. Enzymes have a high specificity but this can be affected by the pH,
temperature, concentration and the reaction medium.
Most enzymes that are used are microbial lipases. These are the most attractive ones for several
reasons; they are thermostable and don’t need a co-lipase or other different specifications (Xu, 2000).
Immobilized lipases are used in plenty of applications to improve the reusability of the very
expensive enzyme and to develop its stability and selectivity. Immobilization will also decrease the
potential inhibition of the used lipase. In fact, immobilization is mostly done by adsorption, it makes
the lipase also stable during the interesterification. This is because the lipase is not soluble in organic
solvents. For the immobilization, it is very important to choose the right support material as this can
affect the activity and the stability of the immobilized lipase (Wang et al., 2006). The immobilized
enzymes can be used at higher temperatures and especially in systems with very small amounts of
water (Chopra et al., 2008).
An overview of the research that has been performed on the production of CBEs using enzymatic
interesterification is illustrated in table 3.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 15
Table 3: An overview of the enzymatic interesterification in order to produce CBEs (Depoortere, 2011).
Substrate Enzyme Reference
Mahau fat, kokum fat, dhupa fat,
sal fat, mango fat, fatty acid-
methyl ester
Lipozyme immobilized (IM) Sridhar et al., 1991; Xu, 2000
High oleic acid rapeseed oil Rhizopus arrhizus lipase,
lipozyme
Gitlesen et al., 1995
Palm oil, StStSt, stearic acid,
Stearic acid ethyl ester
Rhizomucor miehei lipase
Chinese vegetable tallow, fully
hydrogenated soybeen oil fatty
acids
Porcine pancreatic lipase Xu, 2000
Chinese vegetable tallow, stearic
acid
Porcine pancreatic lipase
PMF, stearic acid
Rhizopus arrhizus lipase
Teaseed oil, palmitic acid
Stearic acid
Porcine pancreatic lipase Xu, 2000; Wang et al., 2006
High oleate sunflower oil Lipozyme Smith, 2001
PMF and stearic acid Lipozyme Thermomyces
lanuginosis IM (TL IM)
Undurrage et al., 2001; Holm &
Cowan, 2008
Strychnos madagascariensis oil,
Trichelia emetic oil, Ximenia caffra
oil
Rhizomucor miehei lipase Khumalo et al., 2002
Refined bleached and deodorized
palm oil, fully hydrogenated
soybean oil
Lipozyme immobilized (IM) Abigor et al., 2003
Refined olive pomace oil
Porcine pancreatic lipase Ciftci et al., 2009
Palm oil
Carica papaya lipase Pinyaphong & Phutrakul, 2009
Pentadesma butyracea butter
Lipozyme TL IM Tchobo et al., 2009
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 16
The acidolysis reaction is performed using an enzyme that is sn-1,3 specific. One of the enzymes that
can be used is the lipase from Rhizomucor miehei (RM IM) or formerly known as Mucor miehei. This
enzyme is commercially available in the soluble and the immobilized form. Two important
characteristics of RM IM are the stability under diverse conditions, and the high activity. Thanks to all
of previous mentioned advantages, this enzyme has its main uses in fatty acids, oils and the
modification of fats as in the production of CBEs through enzymatic acidolysis (Rodrigues &
Fernandez-Lafuente, 2010).
RM IM is a very useful enzyme in systems where the water activity is held low. In an environment
with a low water activity, the enzyme performs active and stable; also the selectivity is greater at
lower aw (Rodrigues & Fernandez-Lafuente, 2010).
5. Product purification
The fat obtained after enzymatic acidolysis not only contains the desired SSTs but also more than 50%
consists of medium-chain (MFA) and long-chain fatty acids (LFA). Prior to the use of the product as
food, it is necessary to remove those FFA. Because of the high content of FFA, it is not easy to apply a
conventional distillation to remove them. To get rid of these FFA, short path distillation (SPD) can be
applied (Xu et al., 2002).
SPD is a thermal separation technique in which the boiling point of substances is lowered by using
high vacuum pressure. In this manner, the separation of heat-sensitive compounds, materials with a
low volatility and materials with a high molecular weight is possible (Lin & Yoo, 2009; Tovar et al.,
2011; Martins et al., 2006). Another name for SPD is molecular distillation, in which the distance
between evaporator and condensor is on the order of the average free path length of the molecules.
This system also operates under vacuum and it offers a very short residence time and a small hold up
volume (Tovar et al., 2011; Martins et al., 2006). It is a method that is frequently applied in lipid areas,
it has been used to purify MAGs, fraction PUFAs from fish oils, recover carotenoids from palm oil,
recover tocopherols, etc. (Xu et al., 2002). It is also often used to purify products that contain a lot of
MAGs and DAGs which have a strong effect on the crystallization behavior of fats (Lin & Yoo, 2009).
Important parameters that have to be considered and optimized are the temperature of the
evaporator, the feeding flow rate, the speed of the stirring roller and also the content of the FFAs;
since the method that leads to a lower FFA content will also results in a higher loss of tocopherols. An
important disadvantage of using too high temperatures, is that the amount of condensate in the
degasser pump will increase and this is not good for the performance of the distillation equipment.
High temperature has a negative effect on the amount of FFA so it is important to find the optimum
temperature (Xu et al., 2002; Martin et al., 2010).
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 17
The function of the roller in the system is to spread the feed equally on the inside surface of the
heating wall. In this way, the thickness of the film can be controlled. A faster roller speed will
improve the separation performance but more tocopherols and TAGs (due to splattering) will be lost
(Xu et al., 2002).
In the study of Xu et al. (2002), it was found that the flow rate has the most influence on the amount
of FFA and because it is strongly related to the heating capacity of the evaporator, makes it a very
essential parameter.
The process is schematically represented in figure 8. The distance between the evaporator and the
condensor is very short and a pressure drop is avoided.
The product after interesterification is brought into the feeding tank, the product goes to the
evaporator and is put as a thin layer on the inside of the wall by the stirring roller. The FFA are
evaporated and leave the equipment in the distillate receiver. A FFA trap with liquid nitrogen is
necessary to condense the FFA in order not to be sucked into the vacuum pump. The residue with
the desired part of the interesterification product is condensed and obtained by the residue receiver.
Figure 8: Process scheme for SPD (Xu et al., 2002).
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 18
6. Optimization of the reaction
The response surface methodology (RSM) is a statistical technique that is used in the investigation of
complex processes. There are only a reduced number of experiments necessary to provide enough
information to gain statistically acceptable results, which is the biggest advantage of RSM. It is a
method that is used a lot in food science research. Elibal et al. (2011) used RSM to optimize the
production of SSTs containing conjugated linolenic acid by enzymatic acidolysis. Melo Branco et al.
(2011) used RSM to model the production of SSTs from soybean oil after enzymatic acidolysis and
Shuang et al. (2009) optimized the production of SSTs by lipase-catalyzed acidolysis of soybean oil.
RSM is used to evaluate the effects of different variables in the process like reaction time, reaction
temperature, substrate ratio, enzyme load and water content, and it allows one to conclude which
variable will be the most vital (Shieh et al., 1995). It also enables the user to evaluate the effects on
the response variable(s) of multiple parameters in combination or alone. RSM can also predict the
behavior of the response variable(s) under given sets of conditions. Moreover, it is even possible to
find more than one optimum condition for the reaction due to different combinations of the
variables (Xu et al., 1998; Shekarchizadeh et al., 2009). When comparing RSM with classical one-
variable-at-a-time or full-factorial experiments, RSM performs faster and is less expensive (Shieh et
al., 1995).
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 19
Materials and methods
1. Substrates and enzyme
High Oleic Sunflower Oil (Radia 7363) and a mixture of free fatty acids (FAM) (RADIACID 0417) were
provided by Oleon company (Ertvelde, Belgium). The CB used as a reference in the experiments was
delivered by Belcolade (Erembodegem, Belgium). Lipase from Mucor miehei (RM IM) and Novozyme
435 were bought from Novozymes (Bagsvaerd, Denmark).
2. Methods
2.1 Quality of the starting oil
The quality of the starting oil was evaluated by the following methods:
2.1.1 Peroxide value (PV)
AOCS Official Method Cd 8b-90 (1996)
2.1.2 p-anisidine value (p-AV)
AOCS Official Method Cd 18-90 (1996)
2.1.3 Totox value
The Totox value, defined as 2 times the PV + p-AV, was calculated to determine the total oxidation
value (Rossell, 1994).
2.1.4 Acid value and FFA
AOCS Official Method Ca 5a-40 (1966)
2.2 Chemical composition of the starting oil
2.2.1 Fatty acid profile
AOCS Official Methods Ce 1-62 (1990) & Ce 2-66 (1989)
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 20
2.2.2 Triacylglycerol profile
2.2.2.1 TAG profile
The TAG profile was determined by using the Shimadzu HPLC in combination with an evaporative
light-scattering detector (ELSD) (Alltech-3300, Alltech Associates Inc., Lokeren, Belgium). The N2 gas
flow rate was set at 1.2 L/min, the nebulizing temperature at 45°C and the acquisition gain was 1.
The fat samples were dissolved in a mixture of 70% acetonitrile (ACN) and 30% dichloromethane
(DCM). The reversed phase C18 column (Grace-Aldrich) was used.
The mobile phase was ACN and DCM. The same elution program was used as described by Rombaut
et al. (2009). The flow was maintained at 0.72 mL/min.
2.2.2.2 Positional isomeric TAGs
Using previous method, it was not possible to separate symmetric and asymmetric TAGs. Therefore,
a second method, with a silver ion column was used to determine the TAG composition. The method
described by Macher & Holmqvist (2001) was adjusted. The TAG profile was determined with the
Shimadzu HPLC in combination with ELSD as detector. Following ELSD conditions were used: a gas
flow rate of 1.5 L/min, the nebulizing temperature of 38°C and the acquisition gain was 1.
The mobile phases were heptane and acetone, the flow rate was 1.0 mL/min. Prior to sample
injection, the column was reconditioned during 12 min at 98% heptane and 2% acetone. After
injection of the sample, the concentration of acetone was increased to 3% in 5 min and kept there
for 5 min. This was followed by a further increase to 10% in 10 min, holding it there for 5 min. Finally,
the acetone concentration was increased to 80% over 10 min. For the sample preparation, the fat
was dissolved in heptane, which was diluted to a concentration of 1 mg/mL. The samples were
analyzed in duplicate.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 21
2.3 The enzymatic acidolysis
Acidolysis reactions with lipase were carried out at different conditions. Reactions were performed in
a glass container, placed in a water bath with mechanical agitation at 300 rpm. In general, substrates
(1 mol oil + necessary quantity of FAM) and water were mixed and heated to the reaction
temperature for 20 min. The reaction started when enzyme was added. Reactions were stopped and
the interesterified oil was filtered through Wathman filter paper No 40 with vacuum to remove
enzymes.
At different time intervals, samples were drawn for analysis. Before taking the samples, the stirrer
should be turned off for 1-2 min in order to let the enzyme particles sediment. Samples were taken
from the top of the oil. Sampling was done with a 1 mL pipette and a 150 mesh metal filter. The
samples were stored in the freezer at -18°C.
The enzyme was washed with acetone to remove all fat residues and to reuse it.
The different reaction parameters that were tested, and their range, are given in table 4.
Table 4: An overview of the tested enzymatic acidolysis parameters and their range.
Parameter Tested range
Reaction time 1 to 72h
Reaction temperature 60 to 75°C
Water content 0 to 5% (based on substrate)
Enzyme load 5 to 30% (based on substrate)
Substrate ratio 1/1 to 1/7 (mol oil/ mol FAM)
2.4 Response surface methodology
The software used to optimize the interesterification reaction through RSM was Design-Expert® 8.0.2
from Stat-Ease Corporation, Minneapolis, USA. A central composite design was applied in this work.
The five factors were reaction temperature (°C), reaction time (h), substrate molar ratio
(HOSO/FAM), water content (% based on the substrate) and enzyme load (% based on the substrate).
Two responses were evaluated. The first was the amount of saturated-unsaturated-saturated (%
SUS) TAGs, mainly POP, POSt, StOSt, formed in each sample. The second one was the amount of
saturated-unsaturated-unsaturated (% SUU) TAGs, mainly POO and StOO, formed.
Using the optimized parameters given by the software, the interesterified oil was produced on a
large scale (figure 9). Larger amounts of interesterification product are necessary to perform the SPD
(see further).
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 22
Figure 9: Enzymatic acidolysis reaction on a big scale in optimized conditions.
2.5 Short path distillation Figure 10 represents the different parts of the SPD installation (VTA, Deggendorf, Germany). The
product was distilled two times to reduce the amount of FFA to an acceptable amount.
Figure 10: SPD equipment (Oleon, Belgium).
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 23
In table 5 the parameters for the different pumps and water baths are given.
Table 5: Distillation parameters in SPD.
Equipment part Condition
Feed 70°C Evaporator 200°C Residue 60°C Distillate 70°C Vacuum 0.003 mbar Wiper speed 850 rpm Pump for feed 20 Hz Pump for residue (for first distillation) 10 Hz Pump for residue (for second distillation) 18 Hz Pump for distillation 15 Hz
2.6 Fractionation
Two different solvent fractionation methods were evaluated.
The first procedure used, was the solvent fractionation described by Chong et al. (1992). This method
was based on the patented procedure from 1991 by UNILEVER PLC (European patent 0 199 580 B1).
The different steps in the fractionation procedure are given below.
1) Glyceride-hexane solution 1:10 (w/v) at 4°C for 24h
2) Filter off the precipitated fat (vacuum) at 4°C
3) Wash the crystals with hexane at 4°C
4) Evaporate the filtrate to dryness (rotavapor)
5) Filtrate-acetone solution 1:10 (w/v) at 4°C for 24h
6) Filter off the precipitated fat (vacuum) at 4°C
7) Wash the crystals with acetone at 4°C
8) Evaporate the precipitate to dryness (rotavapor)
9) Blow nitrogen gas through the product at 60°C for 4h
The second procedure used, is a method described by Chang et al. (1990) and Abigor et al. (2003).
The different steps in the fractionation procedure are given below.
1) Glyceride-acetone solution 1:10 (w/v) at 22°C for 24h
2) Filter off the precipitated fat (vacuum) at 22°C
3) Filtrate is cooled to 4°C for 4h
4) Filter off the precipitated fat (vacuum) at 4°C
5) Wash the crystals with acetone at 4°C
6) Evaporate the precipitate to dryness (rotavapor)
7) Blow nitrogen gas through the product at 60°C for 4h
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 24
2.7 Pulsed nuclear magnetic resonance (pNMR)
A Maran ultra pulsed field gradient NMR (Oxford Instruments, Tubney Woods, Abingdon, UK), 10 mm
diameter NMR tubes (Bruker, Karlsruhe, Germany) and a Water bath (Julabo, Seelbach, Germany)
were used.
The fat samples were melted in the oven for 1h at 70°C before the NMR tubes were filled with 3.5 mL
of the sample. Every sample was analyzed in triplicate.
Two different techniques were used, a non-isothermal method and an isothermal method.
2.7.1 Non-isothermal method (non-tempered)
Initially, the fat has to be melted completely to remove all crystal history. A cooling step is necessary
in the next step, so the fat is completely crystallized. Finally, the fat is held at a defined temperature
for some time to equilibrate at that temperature (Timms, 2003). This procedure is shown below.
1) Oven at 70°C for 60 min
2) Water bath at 0°C for 90 min
3) Water bath at 5°C for 60 min
4) Measure the SFC with the pNMR
5) Increase the temperature of the water bath with 5°C
6) Measure the SFC after 60 min
7) Repeat step 5 and 6 until the SFC content becomes 0%
2.7.2 Non-isothermal method (tempered)
In this method (IUPAC 2.150 serial tempered method), the samples were tempered before measuring
the SFC content by keeping them in a water bath at 26°C for 40h. This was done after step 2 in the
previous procedure. After 40h, the samples were cooled to 0°C for 90 min and the earlier described
procedure was continued from step 3.
2.7.3 Isothermal method
In the first step, the NMR tubes were placed in the oven at 70°C for 60 min in order to melt every
possible crystal present in the sample. Then, the samples were placed into the water bath of 20°C.
Measurements were done at different time intervals.
2.8 Differential scanning calorimetry (DSC)
A Q1000 Differential Scanning Calorimeter (-80 to 400°C) (TA Instruments New Castle, USA) with a
refrigerated cooling system (TA Instruments New Castle, USA) was used. An empty pan was used as a
reference. Pans were filled with 5 to 15 mg of the sample. Every sample was analyzed in triplicate.
Two procedures were used, a non-isothermal method and an isothermal method.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 25
2.8.1 Non-isothermal method
The procedure is shown below.
1) Equilibrate at 65.00°C
2) Isothermal for 10 min
3) Ramp 5.00°C/min to -20°C
4) Isothermal for 5 min
5) Ramp 5.00°C/min to 65.00°C
2.8.2 Isothermal method
The procedure is shown below.
1) Equilibrate at 65.00°C
2) Isothermal for 10 min
3) Ramp 10.00°C/min to 20°C
4) Isothermal for 230 min*
*Slow melting samples were held isothermally for 430 min.
2.9 Polarized light microscopy
The polarized light microscope (PLM) Leitz Diaplan (Pleitz, Wetzlar, Germany) was used together with
the temperature-controlled stage Linkam PE94 (Linkam Scientific Instruments, Surrey, UK) and the
Olympus color view camera (Olympus, Aartselaar, Belgium).
The samples were melted in the oven at 70°C. One drop of every sample was placed on a rest plate
with a Pasteur pipette and covered with a cover plate.
The isothermal crystallization procedure is shown below.
1) Equilibrate at 65.00°C
2) Isothermal at 65.00°C for 10 min
3) Cool to 20°C at 10.00°C/min
4) Isothermal at 20°C for 90 min
A picture of the crystals was taken at different time intervals. The samples were kept for 6 weeks at
20°C and an image of the crystallizing fat was made every week.
2.10 Statistical analysis
The data were statistically analyzed with S-plus 8_0 software package. To evaluate significant
differences between data, ANOVA was used. A multiple comparison was performed between
different ratios or contents of one of the parameters. The Tukey test was used to detect significant
differences. All significant differences were based on a 95% significance level (p=0.05). All reactions
were conducted in duplicate.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 26
Results and discussion
1. Introduction CBEs are non-lauric vegetable fats which are completely compatible with CB. Therefore, the CBE can
replace the expensive CB in chocolate. The goal of this research was to produce a CBE by enzymatic
acidolysis of the cheap and commercial available HOSO with a FAM mainly consisting of P and St.
The first part of the research deals with the characterization of the starting oil and the evaluation of
its quality. The chemical characterization consisted of analyzing the FA and TAG composition. To
evaluate the oxidative quality of HOSO, the amount of FFA, PV, p-AV and totox value were
determined.
Secondly, parameters of the enzymatic acidolysis reaction such as reaction time, reaction
temperature, water content, enzyme load and substrate ratio were optimized by RSM in order to
obtain the highest yield of POP, POSt and StOSt. Some trials were performed in order to investigate
the possibility of increasing the product yield by reducing the amount of FFA in the final product. This
was done through the use of glycerol and non-specific enzyme at some point in the enzymatic
acidolysis reaction.
The interesterified product was purified by SPD and fractionation was done in the third step to
increase the purity of the CBE.
In the fourth part of the research, the produced CBE was characterized, chemically and physically.
TAG composition was determined by HPLC and FA profile by GC. Melting and crystallization
properties were analyzed by DSC and pNMR. The produced CBE was mixed in different ratios with CB
in order to analyze the compatibility with CB. Polarized light microscopy (PLM) was applied to analyze
the crystal microstructure of these mixtures.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 27
2. Characterization of HOSO
2.1 Chemical characterization Table 6 gives the results of the different parameters tested to evaluate the oxidative quality of HOSO.
Also the FA composition of the oil is given in this table.
Table 6: Results of the quality tests and the FFA composition of HOSO. The results of the quality parameters are the average of three repetitions, the FA composition is the average of two repetitions.
HOSO
FFA (%) 0.03 ± 0.02
PV 0.32 ± 0.56
p-AV 0.64 ± 0.35
Totox 5.70
C16:0 (% P) 3.97 ± 0.08
C18:0 (% St) 2.84 ± 0.05
C18:1c (% O) 84.25 ± 0.09
C18:2 (% L) 7.76 ± 0.02
C20:0 (% A) 0.18 ± 0.02
C20:1 (%) 0.20 ± 0.02
∑others (%) 0.71 ± 0.04
Oil that is used for enzymatic interesterification, should be of good quality. A fat with a peroxide
value (PV) less than 2 is considered to be freshly produced (Gray, 1978; Robards et al., 1988;
Novozymes, 2011). The amount of peroxides formed, is related to the degree of oxidation of the oil.
Peroxides are the unstable primary oxidation products which will eventually transform into
secondary oxidation products such as aldehydes, ketones, epoxides etc. According to table 6, the
average PV of HOSO was below the maximum limit, therefore it can be considered as a good quality
oil if only the primary oxidation products were taken into account.
It is outmost important that the p-AV is held low. This value gives the amount of secondary oxidation
products present in the oil, indicative of the bad quality of the oil. For this reason, the p-AV may not
be neglected when making a conclusion about the oxidation status of the fat. According to the
protocol provided by Novozymes, the p-AV should be lower than 4 in order to prevent reactions
between the amino acids in the lipase-backbone and the secondary oxidation products. These
reactions can lead to loss of activity of the enzyme (Novozymes, 2011). HOSO met this specification
(table 6).
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 28
The totox value indicates the total oxidation of the fat sample. It uses both the PV and the p-AV
(O’Keefe & Dike, 2010). The acceptable level of the Totox value should be below 10 (Podmore, 1990)
which was the case for the starting oil.
HOSO contained an average amount of FFA of 0.03%. This low amount of FFA indicates a high quality
of the substrate.
The FA composition is given in table 6. From the table it is clear that HOSO contained a high amount
of oleic acid (C18:1). In the synthesis of a CBE, a suitable starting oil should have a high level of oleic
acid (Khumalo et al., 2002). This means that HOSO is a good substrate for CBE production.
The TAG composition is presented in table 7. HOSO consisted mainly of OOO (62.09%) and
considerable amounts of POL (7.66%), POO (11.01%) and StOO (8.43%). There were also small
amounts of LOL, LOO and StOL present.
Table 7: TAG composition of HOSO with P: palmitic acid, St: stearic acid, O: oleic acid and L: linoleic acid.
TAG HOSO
LOL 2.47 ±0.09
LOO 2.99 ±0.09
POL 7.66 ±0.10
StOL 1.35 ±0.02
OOO 62.09 ±0.87
POO 11.01 ±0.14
StOO 8.43 ±0.14
POSt ND
PPSt ND
StOSt ND
PStSt ND
*ND: not detected
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 29
2.2 Physical properties of HOSO The crystallization and melting behavior of the starting oil are measured by DSC and pNMR.
2.2.1 Non-isothermal crystallization and melting behavior as measured by DSC
The non-isothermal method was executed as was described in paragraph 2.8.1 in ‘Materials and
methods’.
In figure 11, the crystallization and melting profile of HOSO, obtained with DSC non-isothermal
method, is shown. According to figure 11, the melting curve shows a broad peak at low temperatures
(-5°C), this means that the oil was completely liquid at room temperature.
Figure 11: Non-isothermal DSC results of HOSO.
2.2.2 Solid fat content as measured by pNMR
HOSO was liquid (0% SFC) at 5°C, this is due to the high amount of triolein (OOO), which causes HOSO
to be a low melting oil.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 30
3. Enzymatic acidolysis
3.1 Optimization of the reaction conditions The effect of each parameter (reaction time, reaction temperature, water content, enzyme load and
substrate ratio) in the acidolysis reaction on the amount of desired TAGs, was analyzed separately
while the other parameters were held constant. An overview of the experimental design is illustrated
in figure 12. The selection of the best conditions was based on the percentage of POP + POSt + StOSt,
the three main TAGs in CB.
Figure 12: Process scheme to find the optimum conditions for the enzymatic acidolysis reaction.
•Time: 1, 2, 3, 4, 6, 8, 24, 30, 48, 56 and 72h
•Temperature: 70°C
•Water content: 1% (based on substrate)
•Enzyme load: 10%
•Substrate ratio : 1/7: HOSO / FAM (mol / mol)
Effect of reaction time
•Time: Best that was found in the first step
•Temperature: 60, 65, 70, 75°C
•Water content: 1%
•Enzyme load: 10%
•Substrate ratio : 1/7: HOSO / FAM (mol / mol)
Effect of reaction temperature
•Time: Best that was found in the first step
•Temperature: Best that was found in the second step
•Water content: 0%, 1%, 3%, 5% (based on substrate)
•Enzyme load: 10%
•Substrate ratio : 1/7: HOSO / FAM (mol / mol)
Effect of water content
•Time: Best that was found in the first step
•Temperature: Best that was found in the second step
•Water content: Best that was found in the third step
•Enzyme load: 5%, 10%, 15%, 20%, 25%, 30%
•Substrate ratio : 1/7: HOSO / FAM (mol / mol)
Effect of enzyme load
•Time: Best that was found in the first step
•Temperature: Best that was found in the second step
•Water content: Best that was found in the third step
•Enzyme load: Best that was found in the fourth step
The product (figure 27) gave two peaks of which A is the peak of the TAGs and peak B represents the
melting point of the FFA present in the interesterified product. The total amount of FFA was
completely melted at 60°C. The low onset temperature (-3.82°C) of peak A indicated that there were
also low melting SUU and UUU TAGs present in the product. The peak of FFA in the purified product
was almost completely eliminated and shifted towards lower temperatures (35.53°C) indicating the
presence of SSS TAGs (B2). A2 stands for the mix of SUS TAGs (CBE) and the SUU TAGs present in the
product after SPD.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 51
Figure 27: Non-isothermal crystallization and melting profile of the interesterified product (product) and the
purified product (after SPD) as measured by DSC.
Figure 28 shows the crystallization and melting graphs of each fraction obtained by fractionation
method A. In figure 29, a comparison is made between the final fraction (SF2) of the two
fractionation methods.
SF1 (method A), contained SSS TAGs resulting in a peak B1 at maximum 39.55°C. Peak A1 indicates
that a small amount of other, low melting (11.81°C), TAGs were lost during this fractionation
procedure. The OF2 (method A) contained the SUU and UUU TAGs (POO, StOO and OOO) resulting in
a peak at 7.91°C. The final fraction SF2 (method A) had a large peak around 20°C which indicated the
presence of the desired TAGs (POP, POSt and StOSt). Peak B2 (37.64°C) showed that there was still a
small amount of SSS TAGs present in the SF2 fraction.
Figure 28: Non-isothermal crystallization and melting profile of the fractions after fractionation method A as
measured by DSC.
A
B
A2
B2
A2
B2
A1 B1
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 52
When comparing the SF2 fractions of the two fractionation methods (figure 29), one can notice the
clear difference in the A2 (method A) and A2* (method B) peaks. The latter method resulted in a
higher amount of SUU TAGs (peak A2*) present in the sample, whereas mehod A resulted in a SF2
fraction closer resembling CB. This was also concluded by comparing the peak temperatures: 18.13°C
(A2) and 12.23°C (A2*), the lower temperature for peak A2* indicated that there were more SUU
TAGs left after method B. B2* (method B) showed the lower amount of POP, POSt and StOSt present
in SF2* (method B) compared to SF2 (peak A2). Consequently, based on the results obtained from
the DSC graphs, fractionation method A seems the most effective.
Figure 29: Non-isothermal crystallization and melting profile of the SF2 fractions after fractionation method A and B as measured by DSC.
4.3.2 SFC
Comparisons are made between the SFC of the interesterified product, the purified product after
SPD, different fractions of products and CB. Figure 30 shows the results of the pNMR non-isothermal
method for the following tempered and non-tempered samples: the interesterified product (product),
the product after SPD (purified product), the second stearin fraction (SF2) and CB.
The product has high contents of FFA and SSS TAGs, that is why this sample needs very high
temperature to melt completely (60°C). These results were also confirmed by DSC that showed the
complete melting of the FFA present in the product at 60°C. After SPD, the FFA were removed and
the melting point of the obtained product reduced (45°C), but this melting point remained higher
compared to CB due to the presence of SSS TAGs.
A2* B2*
C2*
A2
B2
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 53
The product after fractionation, using method A (SF2), showed a similar melting profile as CB but it
had lower SFC till 30°C. Between 30 to 40°C, the amount of SFC in SF2 is higher than CB. At 40°C both
CB and CBE melted completely. In the study of Vereecken et al. (2009), the higher SFC at high
temperatures and lower SFC at low temperatures was attributed to the increased amount of SSS
TAGs in the fat sample.
Figure 30: Non-isothermal (non-tempered and tempered) SFC curve of the product, purified product, SF2 (CBE) and CB as measured by pNMR.
0
10
20
30
40
50
60
70
80
90
100
5 10 15 20 25 30 35 40 50 60
SFC
(%
)
temperature (°C)
product tempered
product
purified product tempered
purified product
SF2 tempered
SF2
CB tempered
CB
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 54
5. Chemical composition CB/ CBE mixtures First, the FA profile of CB and CBE was determined using GC. Secondly, the produced CBE was mixed
with CB in different ratios (0, 20, 40, 60, 80 and 100% CBE). The TAG composition was determined of
each mixture.
5.1 FA profile of CB and CBE The results are shown in table 16. P, O and St are the three major fatty acids in the samples but also
considerable amounts of L (C18:2) and A (C20:0) were present. The amount of P and O in CBE was
significantly (p<0.05) higher than in CB. CBE contained significantly (p<0.05) less St compared to CB.
According to Khumalo et al. (2002), CB contains typically 34% St, 26% P and 35% O. The amount of P,
O and St in the analyzed CB was very close to the values mentioned in literature.
Table 16: FA composition of CB and CBE. The results are the average of two repetitions.
CB CBE
C16:0 (P) 27.66 ±0.13 28.62 ±0.07
C18:0 (St) 35.02 ±0.18 33.13 ±0.07
C18:1c (O) 32.48 ±0.03 34.62 ±0.05
C18:2 (L) 2.71 ±0.20 2.10 ±0.12
C20:0 (A) 0.87 ±0.12 0.37 ±0.01
∑others 0.96 ±0.01 1.04 ±0.08
5.2 TAG composition The TAG composition of the CB/ CBE mixtures was analyzed with HPLC using ACN and DCM as mobile
phases according to the procedure described by Rombaut et al. (2009). The amount of the three
main TAGs (POP, POSt and StOSt) is given in figure 31 (A). The amount of POP, POSt and StOSt in CB
was respectively 20.82, 41.57 and 33.89%. For the produced CBE, these amounts were respectively
21.52, 36.91 and 27.71%.
According to Lipp et al. (2001), a CBE contains generally a lower amount of POSt, a larger amount of
POP but a similar amount of StOSt. For the produced CBE, this statement was only confirmed in the
case of POSt. The amount of POP was similar and the amount of StOSt was significantly lower in the
produced CBE.
The amount of StOSt in CB was significantly (p<0.05) higher compared to the mixtures with CBE. The
total amount of the three main TAGs was significantly (p<0.05) higher in CB compared to the mixture
when 80% or more CBE was added.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 55
The amount of SSS and SUU TAGs in each sample is presented in figure 31 (B). The reference CB
contained 1.64% and 1.56% SUU and SSS TAGs, respectively while these amounts for CBE were 8.42%
and 4.61%. When more CBE was added, the amount of SSS and SUU in the sample increased, having
opposite effects on the melting behavior of the samples. A higher amount of SSS will increase the
melting point while a higher amount of SUU will decrease the melting point.
Figure 31: Percentage of POP, POSt and StOSt (A) and SUU, SSS (B) TAGs in the CB/ CBE mixtures.
Figure 32 gives a ternary phase diagram. The area marked with the red line shows the ratios of POP,
POSt and StOSt which still give the same tempering characteristics as CB. This means that exactly the
same ratios of POP, POSt and StOSt as CB are not necessary (Timms, 2003; Padley et al., 1981). When
plotting the ratio of POP, POSt and StOSt of the produced CBE in this ternary phase diagram, it was
noticed that it fell into the marked area and not so far from CB. This means that the produced CBE
will have the same tempering characteristics as CB (Padley et al., 1981).
0
10
20
30
40
50
60
70
80
90
100
0 20 40 60 80 100
TAG
(%
)
CBE added (%)
StOSt
POSt
POP
0
1
2
3
4
5
6
7
8
9
10
0 20 40 60 80 100
TAG
(%
)
CBE added (%)
SUU
SSS
A B
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 56
Figure 32: POP/POSt/StOSt ternary diagram showing the position of CB, vegetable fats used as CBE and the
enzymatically produced CBE (Padley et al., 1981; Smith, 2001).
Silver ion HPLC was used to determine the symmetric and asymmetric TAGs present in CB/ CBE
mixtures. Stereospecific analysis of triglycerides is important because it influences the physical
behavior (melting properties, crystallization behavior and polymorphism) of fats and oils.
The results of symmetrical and asymmetrical TAGs for different CB/ CBE mixtures are shown in figure
33. There were no SSU TAGs detected in CB while in CBE there was 2.11% of SSU TAGs present. When
the amount of CBE in the mixture was raised, the amount of SSS, SSU and SUU TAGs increased. The
opposite was noticed for the amount of SUS TAGs, which decreased when the amount of CBE was
increased. High amounts of asymmetrical TAGs in the produced CBE are undesirable. According to
Vereecken et al. (2010), asymmetrical TAGs result in a slow crystallization of the fat as they have
lower melting points compared to the analogue symmetrical TAGs.
SSU TAGs are the result of acyl migration. Several factors cause acyl migration during the reaction.
These factors are an increase in reaction temperature, reaction time, water content and type of the
enzyme.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 57
Figure 33: Percentage of SSS, SUS, SSU and SUU TAGs in the CB/ CBE mixtures; results obtained by silver ion
HPLC.
6. Physical characterization
In the next part, the physical characteristics of the CB/ CBE mixtures will be described and compared.
It is fundamental to have some insight in the different processes involved in fat crystallization to be
able to evaluate the final structure, functionality and quality of the CBE. Three levels in the structure
of the fat can be identified: a nano-, micro- and macro-scale. The primary crystallization (nano scale)
is characterized by nucleation and crystal growth and is followed by aggregation into clusters and
network formation (microscopic scale) (Dewettinck & Depypere, 2010). For instance, the snap of
chocolate depends strongly on the macroscopic properties of the CB fat network (Marangoni &
Narine, 1999).
First the non-isothermal and isothermal crystallization and melting behavior was described using DSC
and pNMR. In the second part, polarized light microscopy (PLM) was used to visualize the isothermal
crystallization.
The melting behavior of the produced CBE was compared to the melting behavior of CB, and
mixtures of CB with CBE in different ratios (0, 20, 40, 60, 80 and 100% CBE) were analyzed. The
results of DSC were compared with the ones obtained by pNMR and PLM.
0
10
20
30
40
50
60
70
80
90
100
0 20 40 60 80 100
TAG
(%
)
CBE added (%)
SUU
SSU
SUS
SSS
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 58
6.1 Non-isothermal crystallization and melting behavior
6.1.1 Non-isothermal crystallization and melting behavior as measured by DSC
Figure 34 shows the DSC graphs of the mixtures of CB and CBE in different ratios. The values of the
different parameters after integration of the peaks are given in table 17.
Figure 34: Non-isothermal crystallization and melting profile of CB and the mixtures with CBE as measured by
DSC.
Table 17: Parameters Tonset (°C), Tpeak (°C), melting heat (J/g) and width at half height (°C) of DSC melting profile (non-isothermal) for mixtures of CB and CBE.
When comparing the crystallization peaks of CB with the mixtures containing CBE, one can notice a
shoulder appearing before the main crystallization peak (indicated by A on figure 34). CB had no
shoulder in its crystallization curve. This was due to a higher amount of SSS TAGs present in the CBE.
CB contained 1.56% of SSS TAGs compared to 4.61% in the produced CBE. Especially the amount of
PPP differed a lot between CB (0.91%) and CBE (1.96%). According to Corrêa Basso et al. (2010), high
amounts of PPP increase the crystallization rate that could explain the appearance of this shoulder.
A
B
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 59
Adding less CBE meant there were less SSS TAGs present in the mixture and this resulted in a smaller
shoulder. When this shoulder was smaller, the main crystallization started earlier and the onset
temperature was higher. This can be seen in table 17 where Tonset for CB and CBE were respectively
14.05°C and 13.09°C.
Generally, the onset temperature decreased when more CBE was added. This is probably due to the
higher amount of low melting fraction (UUU and SUU TAGs) present in the CBE. Another explanation
is the higher amount of SSU TAGs present when more CBE was added. Vereecken et al. (2010) stated
that, for instance, PPO (34.5°C) had a lower melting point than POP (36°C) resulting in a faster
melting which could explain the lower Tonset when more CBE was added. Also the peak temperature
was significantly lower compared to CB when more CBE was added. This is due to a larger amount of
SUU TAGs present in the CBE.
After the main melting peak, a smaller melting peak occurred for the mixtures (indicated by B on
figure 34). This is due to a considerable amount of SSS TAGs in the CBE, this was also noticed by
Cebula & Smith (1992).
The shift of the melting peak towards lower temperatures was due to a relative higher amount of
SUU TAGs present in the sample which have lower melting points. In figure 34, a small shift of the
melting peak was noticed when more CBE was added to the CB.
6.1.2 Solid fat content as measured by pNMR
The particular melting behavior of chocolate is due to CB. Through pNMR, the SFC is determined with
the tempered and non-tempered procedure, giving information about the melting behavior of each
sample. The melting curve of CB is compared to those of mixtures of CBE with CB in different ratios.
In figure 35 (A) the results of the pNMR non-isothermal method (non-tempered) are shown when the
CBE was mixed in different ratios with CB. All mixtures had lower SFC than CB up to 30°C. The melting
profile of all mixtures was very similar to that of the CB but, at the same temperature, when the
amount of CBE increased, the SFC decreased. The mixtures were completely melted between 35 and
40°C which was similar to CB. Vereecken et al. (2010) noticed a small decrease in SFC when more SSU
TAGs were present.
During tempering of the samples, the fat crystals are transformed to the stable β-polymorphic form.
Around the tempering temperature (26°C), the unstable fat crystals (α) will melt and the more stable
crystals of polymorphic form βV remain. This polymorph causes the narrow melting profile of CB.
When other polymorphic crystals are present in the sample, a broader melting profile is obtained
(Timms, 2003). The SFC of the tempered samples is given in figure 35 (B).
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 60
Figure 35: Non-isothermal non-tempered (A) and tempered (B) SFC curve of the CBE-CB mixtures as
measured by pNMR.
In figure 36 a comparison is made between tempered and non-tempered CB and CBE.
According to Timms (2003), below the tempering temperature, a tempered sample has a lower SFC
compared to a non-tempered sample. Above this temperature, the SFC of a tempered sample is
higher. The tempering temperature in this research was 26°C. When looking at the results, one can
notice that only up to 20°C, the SFC of the tempered sample was lower than the non-tempered
sample of CB.
The same conclusion can be made for the CBE but the difference between tempered and non-
tempered CBE was not as big as for tempered and non-tempered CB. Beyond 20°C, the tempered
sample had indeed a higher SFC than the non-tempered sample resulting in a higher melting point.
0
10
20
30
40
50
60
70
80
90
100
5 10 15 20 25 30 35 40
SFC
(%
)
Temperature (°C)
0
10
20
30
40
50
60
70
80
90
100
5 10 15 20 25 30 35 40
SFC
(%
)
Temperature (°C)
A B
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 61
Figure 36: Non-isothermal SFC curve: comparison of tempered and non-tempered CB and pure CBE as
measured by pNMR.
Different zones can be distinguished in the SFC curve of the tempered samples. The SFC below room
temperature (25°C) indicates the hardness of the fat. Between 25°C and 35°C, the SFC gives an
indication of the heat resistance and if the fat has a high SFC at temperatures above body
temperature (37°C), a waxy mouth feel of the product can be noticed (Talbot, 2009b; Torbica et al.,
2006). The results of these zones of CB and the mixtures with produced CBE are given in figure 37.
As it is clear from figure (A), up to the temperature of 15°C, there was no difference in hardness for
the CB and the mixtures containing the produced CBE. At higher temperatures, the hardness
decreased when the amount of CBE increased. Regarding the heat resistance (B), when more CBE
was added, a lower heat resistance was obtained. The lower heat resistance was probably due to a
relative higher amount of SUU and SSU TAGs present in the produced CBE. SUU and SSU TAGs have
lower melting points and result in a lower SFC at the same temperature compared to CB. Finally,
from 35°C (C), the waxiness increased when more CBE was used in the mixture. The higher amount of
SFC at 35°C in the samples containing CBE, was due to the higher amount of SSS TAGs which melt at
higher temperature.
0
10
20
30
40
50
60
70
80
90
100
5 10 15 20 25 30 35 40
SFC
(%
)
Temperature (°C)
CB tempered
CB non-tempered
CBE tempered
CBE non-tempered
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 62
Figure 37: SFC melting curves indicating the hardness (A), heat resistance (B) and waxiness (C) of CB and
mixtures with CBE (Depoortere, 2011).
6.1.3 Isothermal diagram
In an isothermal diagram, the SFC of the mixtures of CB and CBE are shown as function of the ratio of
CBE in the mixture and the temperature. According to Ciftci et al. (2010), full compatible samples
should result in isotherms that are horizontal lines between the SFC of pure CB and the SFC of pure
CBE at a given temperature.
Figure 38 gives the SFC of the tempered mixtures of CB with CBE measured at different
temperatures. Due to the tempering of the samples, the β-polymorphic form of the fat crystals was
obtained.
It seems that when the temperature was increased, the produced CBE became less compatible with
CB. At a temperature between 20 to 30°C, a higher amount of CBE resulted in a lower SFC. This is due
to the higher amount of SUU TAGs present in the CBE. At 35°C, the SFC is higher when more CBE was
added to the CB because of the higher amount of SSS TAGs in the CBE. This last characteristic, and in
combination with a similar SFC at lower temperatures, is typical for cocoa butter improvers (CBI)
(Pontillon, 1998).
0
20
40
60
80
100
5 10 15 20 25
SFC
(%
)
Temperature (°C)
Hardness
0
20
40
60
80
25 30 35
SFC
(%
)
Temperature (°C)
Heat resistance
0
10
20
30
40
50
30 35 40
SFC
(%
)
Temperature (°C)
Waxiness CB
20% CBE
40% CBE
60% CBE
80% CBE
100% CBE
A
B
C
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 63
Figure 38: Isothermal diagram of the mixtures of CBE and CB.
6.2 Isothermal crystallization For the isothermal crystallization, samples were held at 20°C. Measurements were performed using
DSC and pNMR.
6.2.1 Isothermal crystallization as measured by DSC
The crystallization was performed at 20°C. At higher temperatures, the crystallization would be too
slow and at lower temperatures the fat would already crystallize during cooling up to the isothermal
temperature, making it difficult to study the isothermal crystallization.
At 20°C, a two-step crystallization process was noticed which is presented in figure 39. First α-crystals
are formed, followed by a transformation to the β’-polymorphic form during the second step (Toro-
Vazquez et al., 2005). The duration of the experiment was not long enough for the β’-polymorphs to
transform into the more stable β-polymorphs. According to Foubert (2003), this polymorphic form
can only be obtained after at least one week of storage at room temperature. The peak of the α-
polymorph cannot be integrated since it overlaps with the peak at the beginning of the isothermal
method. Therefore, only the peak of the β’-polymorphic form will be considered during integration of
the results.
0
10
20
30
40
50
60
70
80
90
100
0 20 40 60 80 100
SFC
(%
)
CBE (%)
5°C
10°C
15°C
20°C
25°C
30°C
35°C
40°C
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 64
Figure 39: Isothermal crystallization of CB at 20°C as measured by DSC.
Figure 40 gives the DSC graph of mixtures of CB and CBE using the isothermal method at 20°C. From
this figure it is clear that the crystallization at 20°C slowed down when more CBE was added. The
peak maximum will be lower when more CBE was added. Also, adding more CBE resulted in
increased tailing of the crystallization peak. This means that more time was needed for these
samples to crystallize completely. The same flat heat-flow curves, due to the very slow crystallization
of CBE, were noticed by Kerti (2001). Adding CBE also caused the crystallization to start later than
was the case for the CB.
All of these differences can be explained by the higher amount of UUU and SUU TAGs in the
produced CBE. In CB, only 1.64% of POO and StOO were present while the produced CBE still
contained 8.42% of OOO, POO and StOO. Besides, the amount of SSU TAGs may slow down the
crystallization. Especially OOO had an influence on the crystallization rate. Foubert et al. (2004b)
noticed that the β’ melting point for POO (2.5°C) and StOO (8.6°C) caused these SUU TAGs not to
crystallize at 20°C. Therefore, a slow crystallization was noticed when more SUU and SSU TAGs were
present (Vereecken et al., 2010). However, because of the high amount of SSS TAGs, co-
crystallization can occur which can make these SUU TAGs to crystallize.
α-crystallization peak β’-crystallization peak
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 65
Figure 40: Isothermal crystallization at 20°C of the different ratios of CB and CBE.
Foubert et al. (2002) developed a model which describes the kinetics of the isothermal crystallization
of fats. The kinetics of crystallization of fat is important when trying to obtain desired product
characteristics (Foubert et al., 2008). Foubert et al. (2004a) demonstrated that, compared to other
mathematical models which describe isothermal crystallization, the Foubert model shows a better fit
to the data. Moreover, the Foubert model offers flexibility towards asymmetric DSC curves. In the
Foubert equation, four parameters are important. These four parameters and the differential
equation are (Foubert, 2003):
tind (h): the time needed to obtain x% crystallization and x is chosen to be 1.
K (h-1): the rate constant
aF (J/g): the maximum amount of crystallization
n (-): the order of the reverse reaction
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 66
It is important to notice that the Foubert model was developed for single-step crystallization
processes, only the major crystallization peak is taken into account. Therefore, in this research only
the parameters of the β’-crystallization were compared (Calliauw, 2008a). The model was fitted to
the data, and the start- and endpoint of the integration was determined using the calculation
algorithm as developed by Foubert (2003). The value of parameter n is fixed at 6 in order to
determine the changes in K. This value for β’-crystallization was also found by Foubert et al. (2006)
when analyzing the two-step crystallization process of CB at 20°C. Table 18 gives the Foubert
parameters for the CB and mixtures with produced CBE.
The amount of equilibrium solid fat decreased when more CBE was added to the mixture. aF of each
mixture containing CBE, differed significantly (p<0.05) from the aF of CB. The higher aF of CB can be
explained by the higher amount of POSt (41.57%) compared to the CBE (39.91%). According to
Calliauw (2008a), the amount of POSt in a fat possitively correlates with the aF. On the other hand;
DAGs and FFA lower the aF (Calliauw et al., 2008b). Since no DAGs were detected in CBE and FFA
were removed after SPD, this could not be the reason of the decrease in aF in the CBE. aF decreased
because of the higher amount of SUU TAGs present (Foubert et al., 2004b)
tind indicates the time that was necessary to start the transformation from the α-crystals into the β’-
polymorphic form. One can notice that when more CBE was added, the tind increased, resulting in a
later transformation start. According to Chaiseri & Dimick (1995), the fat will be softer (lower SFC)
when more SUU TAGs (POO and StOO) are present, resulting in a longer induction time. Also, a
higher amount of FFA will increase the tind. The produced CBE contained 8.42% SUU compared to
1.56% in CB. The addition of more CBE resulted in a longer induction time due to the higher amount
of SUU TAGs. The lower tind of CB can be explained by the higher amount of POSt compared to CBE.
Parameter K decreased significantly (p<0.05), compared to CB, when more CBE was added. This is
probably due to the higher amount of SUU TAGs present in the CBE which decreased the rate
constant. Also a higher amount of DAGs will lower the value of K but this was not the case in this
research because of the absence of DAGs in the CBE.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 67
Foubert et al. (2004b) stated that all parameters, except K, are influenced by the ratio of saturated
(Sat FA) to unsaturated FA (Unsat FA) and the ratio of SUS to SUU TAGs. When the amount of Unsat
FA and SUU TAGs increases, aF will decrease and tind will increase. The results of these ratios for the
CB/ CBE mixtures are given in figure 41. The Sat FA are the sum of P, St and A; the Unsat FA are the
sum of O and L as analyzed by GC. The ratio of SUS to SUU TAGs and the value for aF are indicated on
the primary y-axis on the left of the graph. The ratio of Sat FA to Unsat FA and the value of tind are
indicated on the secondary y-axis on the right of the graph. When both ratios decreased, a decrease
in aF and an increase in tind was noticed. This was in accordance with the results of Foubert et al.
(2004b).
Table 18: Parameters aF (J/g), tind (h) and K (h-1
) of the Foubert model for mixtures of CB and CBE.
Sample aF (J/g) tind (h) K (h-1)
CB 72.54 ± 2.06 0.48 ± 0.78 4.63 ± 0.01
20% CBE 67.10 ± 0.85 0.49 ± 0.00 3.14 ± 0.06
40% CBE 62.37 ± 0.03 0.51 ± 0.06 2.21 ± 0.05
60% CBE 52.82 ± 0.78 0.53 ± 0.06 1.63 ± 0,10
80% CBE 50.16 ± 2.57 0.76 ± 0.57 1.31 ± 0.27
100% CBE 49.39± 0.68 1.45 ± 0.04 0.72 ± 0.16
Figure 41: Influence of Sat FA to Unsat FA and SUS to SUU TAGs ratios on aF and tind.
0
0,2
0,4
0,6
0,8
1
1,2
1,4
1,6
1,8
2
0
10
20
30
40
50
60
70
80
0 20 40 60 80 100
Sat
FA/U
nsa
t FA
an
d t
ind
(h
)
SUS/
SUU
an
d a
F (J
/g)
CBE added (%)
SUS/SUU
aF
Sat FA/Unsat FA
tind
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 68
6.2.2 Isothermal crystallization as measured by pNMR
In figure 42 the isothermal crystallization of the mixtures CB/CBE at 20°C followed by pNMR is given.
In this figure, the two-step crystallization was observed. The insert of the first 50 min (figure 42),
showed the formation of the α-crystals and the beginning of the transformation into the β’-
polymorphic form as was also noticed by Dewettinck et al. (2004). It can be concluded that the
nucleation of CB started later. In other words; the crystallization (of α-crystals) started sooner when
the amount of CBE in the mixture was increased. This was due to the higher amount of SSS TAGs in
the CBE. Possibly, the higher amount of FFA (1.32%) in the CB resulted in the slower nucleation. The
opposite was noticed for the second crystallization step in which α-crystals are transformed into the
β’-polymorph. When less CBE was added to the mixture, the increase in this second crystallization
step was steeper. Or; when more CBE was added, the transformation of α- to β’-crystals occurred
slower. This was in accordance with the results obtained by isothermal DSC and can be explained by
the higher amount of SUU TAGs present in the CBE. Ribeiro et al. (2012) stated that the longer
induction period before β’ crystallization was due to a higher amount of SUU TAGs.
After 230 min of measuring the SFC, CB reached a SFC content of 73% while pure CBE only reached a
SFC of 45%. Consequently, it can be concluded that the crystallization was slower when more CBE
was added to the mixture.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 69
Figure 42: Isothermal crystallization at 20°C of CB and mixtures with CBE as measured by pNMR.
0
10
20
30
40
50
60
70
80
0 50 100 150 200 250
SFC
(%
)
Time (min)
CB
20%
40%
60%
80%
100%
0
5
10
15
20
25
30
35
40
45
0 10 20 30 40 50
SFC
(%
)
Time (min)
CB
20%
40%
60%
80%
100%
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 70
6.3 Isothermal crystallization as visualized by PLM Polarized light microscopy (PLM) was used to visualize the isothermal crystallization at 20°C of the
mixtures of CB and CBE into crystals, crystal clusters and the formation of crystal networks. Firstly,
the beginning of isothermal crystallization was recorded. Secondly, the crystallization at 20°C was
traced during a follow-up of 6 weeks.
6.3.1 Start of isothermal crystallization at 20°C
Pictures of the crystal formation were taken after 1, 10, 30, 60 and 90 min of isothermal
crystallization at 20°C. In order to determine the relationship between microstructure and
polymorphism of CB, the temperature of 20°C is crucial. Below 20°C and depending on the duration
of the crystallization, the CB can have a granular morphology characteristic for the β’-polymorph
while higher temperatures can promote the crystallization into clusters with high polymorphic
stability (Ribeiro et al., 2012). The result after 1 min is given in figure 45 in appendix. The black
background was the liquid fat, the crystals were polarized and they appeared as the white spots in
the picture.
The pictures show that CB (a) almost didn’t show any crystal after 1 min. When more CBE was added,
more crystals in a granular appearance were noticed, corresponding to the formation of α-crystals.
This means that first the α-crystals were formed. From the isothermal pNMR results, it was clear that
the nucleation started earlier when more CBE was added. Therefore, in the samples containing more
CBE, crystals were noticed earlier compared to CB where the nucleation started later. The
explanation is the higher amount of PPP present in the CBE compared to CB. Vereecken et al. (2009)
noticed that a higher amount of PPP resulted into bigger granular crystals. This can be seen in the
picture where adding more CBE resulted into slightly larger granular crystals. Campos et al. (2010)
noticed that a higher amount of StStSt present, decreased the onset time of crystallization. The
produced CBE contained more StStSt compared to CB, which resulted into a faster formation of α-
crystals.
The result after 60 min is given in figure 43. Compared to the results after 1 min, the granular
appearance became more clear and denser although the difference was not that clear. The same
conclusion can be made when more CBE was added, the α-crystals were formed faster and in a larger
amount.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 71
Figure 43: Isothermal crystallization at 20°C after 60 min as visualized by PLM: 0 (a), 20 (b), 40 (c), 60 (d), 80
(e) and 100% (f) CBE.
After the isothermal crystallization, the samples were kept at 20°C for 6 weeks. Pictures were taken
on a regular base to evaluate the crystal formation.
6.3.2 6 week follow-up
During 6 weeks, the samples were stored at 20°C. Every week, a picture of each sample was taken to
analyze the progress in crystal formation. Figure 44 gives the results after 24h, 2, 4 and 6 weeks of
isothermal crystallization. The results after 1 week, 3 weeks and 5 weeks are given in figure 46 in
appendix.
a b
c d
e f
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 72
After 24h, a denser network of small granular crystals was formed. Also CB formed a dense granular
network of crystals. The second step in the crystallization process was noticed: the formation of α-
into β’-crystals. At 24h, this process also occurred for CB.
From the results of isothermal DSC and pNMR, it was stated that the transformation of α- into β’-
crystals occurred faster for CB. This was also noticed with PLM where a denser granular crystal
network was seen compared to the mixtures containing CBE. When more CBE was added, less crystal
formation was noticed. This was probably due to the higher amount of SUU and SSU TAGs present in
the produced CBE. These TAGs slow down the crystallization process as was stated after isothermal
DSC. The delay in polymorphic transition due to a higher amount of StStSt present in the CBE, was
explained by Campos et al. (2010).
From week 2, bigger, featherlike crystals, specific for β-crystals, were formed when 40% or less CBE
was added. Also the formation of a large amount of small crystals was noticed when 60% or more
CBE was used. After week 4, a different form of crystals was noticed when 40% or more CBE was
added compared to the CB crystals. 100% CBE still crystallized into a lot of very small crystals,
indicating that the crystallization of CBE was still going on. In the study of Vereecken et al. (2009), the
crystal network of fat samples, higher in PPP, only became denser after one month of storage at
20°C. Therefore, they stated that a higher amount of SSS TAGs lead to smaller crystals. In the
meantime, CB and 20% CBE formed large and dense crystal aggregates.
The pictures after week 6 showed the featherlike structure of the crystals when 40% or less CBE was
used. Looking at the crystals of 40% CBE, one could notice two different morphologies in the crystal
structure. The center of the big crystals had a granular structure while on the outside, featherlike
crystals surrounded the granular crystals. When 60% CBE was added, also two different forms in
which the crystals aggregated were noticed; a more featherlike part and a spiral like structure. The
crystals of the 80% CBE mixture started to grow more but in comparison with the previous samples,
there were still a lot of small crystal aggregates. The pure CBE still contained the granular and dense
structure of very small crystals.
In conclusion; crystal formation was faster for CB resulting in larger crystals aggregates. Each week,
the CB crystal network became larger and denser compared to the mixtures where CBE was added.
When more CBE was used, less crystal aggregates were formed and they were smaller than the CB
crystal network. This was another indication of the very slow crystallization of the produced CBE.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 73
Figure 44: Isothermal crystallization at 20°C as visualized by PLM for CB, 20, 40, 60, 80 and 100% CBE. The
microstructure is given after 24h, 2 weeks, 4 weeks and 6weeks.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 74
General conclusions The purpose of this study was to produce a CBE by enzymatic acidolysis starting from the cheap and
commercial available HOSO, and to compare the produced CBE with CB, chemically and physically.
In the first step, the quality of HOSO and its chemical composition, in order to be used as a source in
enzymatic acidolysis reactions, was analyzed. The oil was of good quality and contained a high
amount of OOO (62.09%). Having a high amount of TAGs with O at the sn-2 position, made this oil an
ideal source for CBE using the sn-1,3 specific RM IM and the FAM of P and St.
Secondly, the enzymatic acidolysis of HOSO with FAM was performed and the optimized value of
each parameter as obtained by RSM was: reaction time (8h), reaction temperature (65°C), water
content (1%), enzyme load (8.54%) and the substrate ratio (7.99 mol FAM: 1 mol HOSO).
Some attempts were made in order to improve the yield by adding glycerol and Novozyme 435 to the
interesterification reaction with HOSO. The result was a reduction of FFA from 65% to 9% when only
glycerol was added in a molar ratio of 1:4 (mol substrate: mol glycerol). However, the amount of DAG
increased up to 67% and the amount of desired TAGs (POP, POSt and StOSt) decreased from 20% to
13.45%. Ideally, a maximum of 2 mol glycerol can be added to reduce the FFA to approximately 36%
and 35% of DAG. Adding glycerol and the non-specific enzyme together, resulted into the same
trends of decreasing FFAs, decreasing TAGs and increasing amount of DAGs. Since it is more difficult
to remove DAGs compared to FFAs, it was concluded not to include glycerol and a non-specific
enzyme in the reaction.
The purification of the interesterified HOSO by SPD resulted in only 0.32% of FFA. Through solvent
fractionation, the amount of SUU and SSS TAGs was reduced to 8.42% and 4.61%, respectively.
Compared to CB, 1.64% SUU and 1.56% SSS, these amounts were high. The CBE contained a similar
amount of POP, and lower amounts of POSt and StOSt compared to CB. Also the FA profile was
similar to that of CB. Therefore, it can be concluded that the produced CBE will have the same
tempering characteristics as CB.
In the last part of the research, the physical characteristics of the produced CBE, when mixed in
different ratios with CB, were determined. From the results of non-isothermal DSC, a shoulder before
the main crystallization peak was noticed. This peak was due to the higher amount of SSS TAGs
present in the CBE.
The solid fat curve as measured by pNMR was similar for CBE as to CB at temperatures between 20
and 30°C. However, at lower temperatures, CBE had a lower SFC due to the higher amount of SUU
TAGs and at increasing temperatures, a higher SFC was obtained because of the larger amount of SSS
TAGs in the CBE which may result in a higher waxiness from 35°C.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 75
The two-step crystallization process was visualized by isothermal DSC and pNMR at 20°C. Due to the
higher amount of SSS TAGs present in the CBE, the nucleation into α-crystals started earlier for the
mixtures containing CBE than for CB. tind, indicating the time that was necessary to start the
transformation from the α-crystals into the β’-polymorphic form, increased when more CBE was
added to the mixture. On the other hand, the rate constant (K) decreased when more CBE was added.
These observations lead to a slow crystallization behavior of the produced CBE.
The isothermal crystallization at 20°C was also visualized by PLM during a period of six weeks. During
these weeks, it was clear that the CBE formed very quickly granular crystals (α-crystals) but this
granular network only changed very slowly into aggregates and β-crystals. The CB crystals
transformed fast into big and featherlike crystal aggregates compared to the CBE which still had very
small, granular crystals after five weeks. A different morphology of the crystals was noticed when
CBE was added to CB compared to pure CB. Adding CBE resulted in less dense crystal networks in the
shape of a spiral.
By using enzymatic acidolysis, it is possible to produce a CBE from HOSO, P and S with a chemical
composition that is very close to that of CB. However, the physical characteristics of the CBE were
different, resulting into a very slow crystallization process compared to CB. Therefore, according to
the results, when using this CBE in CB, small amounts (maximum 20%) could be applied with limited
changes to the typical physical properties of CB.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 76
Further research For the production process of CBE, it would be beneficial if methods were found to minimize the
amount of FFA left in the esterified product. One possibility is the use of ultrafiltration which is a
membrane technology. With the removal of water during the enzymatic acidolysis reaction, the yield
of the reaction can be increased. Possible methods to remove the water are adding salt hydrates or
applying pervaporation (Ghandi et al., 2000).
For further research concerning the produced CBE, it might be interesting to use the CBE in different
ratios into chocolate. It can be tested if the pure CBE has a big effect on rheological and textural
properties of chocolate.
The CBE produced by enzymatic acidolysis can be compared to commercial available CBEs in order to
study if the produced CBE shows better chemical and physical compatibility with CB. Also the major
differences between commercial available CBEs and enzymatic produced CBE can be evaluated.
Production of Cocoa Butter Equivalent through Enzymatic Acidolysis 77
References 1. Abigor R.D., Marmer W.N., Foglia T.A., Jones K.C., Diciccio R.J., Ashby R. & Uadia R.O., 2003.
Production of cocoa butter-like fats by the lipase-catalyzed interesterification of palm oil and
hydrogenated soybean oil. Journal of American Oil Chemists’ Society, 80 (12), 1193-1196.
2. Bockisch M., 1998. Fats and Oils Handbook. Illinois, AOCS Press, 838 p.
3. Calliauw G., 2008a. Molecular interactions affecting the phase composition during dry
fractionation of palm olein. PhD thesis, Ghent University, Belgium, 275p.
4. Calliauw G., Vila Ayala J., Gibon V., Wouters J., De Greyt W., Foubert I. & Dewettinck K.,
2008b. Models for FFA-removal and changes in phase behaviour of cocoa butter by packed
column steam refining. Journal of Food Engineering, 89, 274-284.
5. Campos R., Ollivon M. & Marangoni A.G., 2010. Molecular composition Dynamics and