Return to search

Mango (Mangifera indica L.) kernel fat : fatty acid profile, oxidative stability and development of fourier transform near infrared (FT-NIR) spectroscopy calibration models

Thesis (MSc Food Sc )--Stellenbosch University, 2002. / ENGLISH ABSTRACT: The oxidative stability of crude, cold-pressed mango kernel fat (MKF) was
determined over a period of 240 days using the peroxide value (PV), conjugated diene
value (CD) and p-anisidine value (AV) tests. The changes in fatty acid profile were
monitored with gas chromatography and the oxidative status of MKF effectively
predicted by FT-NIR spectroscopy. Results obtained from the different methods
complemented each other and indicated the stable character of mango kernel fat
against oxidative deterioration.
The fatty acid profile constituted palmitic acid (CI6:0; 8.43%), stearic acid
(CI8:0; 34.98%), oleic acid (CI8:1 cis; 48.05%), linoleic acid (CI8:2; 6.60%) and
arachidic acid (C20:0; 1.73%). Trace amounts of C16:1 (0.56%), C18:1 trans
(0.25%), C18:3 (0.43%), C20:1 (0.25%) and C22:0 (0.40%) were also found.
The freshly pressed MKF had a peroxide value of 2.7 meq.kg", CD value of
0.07% and an AV of 2.2 mmol.kg", After 40 days of storage, the peroxide values of
MKF stored with and without exposure to a limited amount of oxygen at 5, 15,25 and
40°C increased to 5 meq.kg" and 4 meq.kg" respectively. Emulsification of MKF
had a stabilising effect (maximum PV = 2.8 meq.kg'), while exposure to UV light
had a catalysing effect (maximum PV = 5 meq.kg'). These maximum values,
decreased after 40 days. The CD values of MKF samples stored with and without
exposure to oxygen at 5, 15,25 and 40°C increased to 0.18% and 0.16%, respectively
at day 40. The CD values of samples exposed to light increased to 0_20% and the
emulsified samples showed similar values to that of the MKF samples not exposed to
oxygen. The conjugated diene values remained stable after day 40. The p-anisidine
values of the MKF samples both stored with and without exposure to oxygen at 5, 15,
25 and 40°C varied between 0.5 and 5 mmol.kg". The weak correlation to the
measurement of nonanal, as well as the low levels of 2-alkenals produced by the
MKF, resulted in these low and sometimes non-linear values.
The peroxide, conjugated diene and p-anisidine values obtained for MKF
stored at 25°C over 240 days were low due to the low content of polyunsaturated fatty
acids in MKF. This compared favourably with the higher values attained for
sunflower, canola and olive oil, which are all rich in polyunsaturated fatty acids.
The minimal changes observed in the fatty acid profile of mango kernel fat
indicated the stability of the saturated fatty acids (CI6:0 and CI8:0) and oleic acid. In addition, the instability of linoleic and linolenic acids was evident due to oxidative
deterioration. A decrease of 7.41% and 12.80% was observed between day 0 and 240
for the C18:2/C16:0 and C18:2/C18:0 ratios respectively.
The prediction of the oxidative status of the MKF samples by near infrared
spectroscopy were possible after the development of calibration models from a total
data set of 300 samples of which one-third was used for independent validation.
Principle component analysis (PCA) indicated classification at 0, 40 and the
remaining (80 - 240) days. The best calibration model for PV yielded a SEP
(standard error of prediction) of 0.46 meq.kg", correlation coefficient (r) of 0.95, bias
of 0.02 and a root mean square error of prediction (RMSEP) of 0.46 meq.kg". The
CD calibration model had a correlation coefficient of 0.89, SEP of 0.01 %, bias of
0.001 and RMSEP of 0.01% when developed on a data set with no pre-processing
applied. The AV calibration had a SEP of 0.32 mmol.kg", bias of 0.03, RMSEP of
0.32 mmol.kg" and rof 0.93. The C18:2 and C18:3 models were built using partial
least squares (PLS) regression and the values obtained for SEP were 0.31% and
0.054%, RMSEP 0.32% and 0.05%, bias 0.05 and 0.01 and correlation coefficcients
were 0.82 and 0.54 respectively. The calibrations for CI8:1, C18:0 and C16:0 yielded
weaker correlations. Good correlations were obtained when calibrating the
CI8:2/CI6:0 and C18:2/CI8:0 ratios. / AFRIKAANSE OPSOMMING: Die oksidatiewe stabiliteit van ru, koud-geperste mango kern vet (MKV) (Mangifera
indica L.) is oor 'n periode van 240 dae bepaal deur gebruik te maak van die
peroksiedwaarde (PV), gekonjugeerde dieen waarde (CD) en p-anisidien waarde (AV)
toetse. Die veranderinge in die vetsuurprofiel is gemonitor deur gaschromatografie en
die oksidatiewe status van MKV is akkuraat voorspel word deur Fourier transformasie
naby infrarooi (FT-NIR) spektroskopie. Die resultate van die verskillende
toetsmetodes komplementeer mekaar goed en dui die stabiliteit van mango kern vet
teen oksidatiewe verval aan.
Die vetsuurprofiel is saamgestel uit palimitiensuur (C16:0; 8.43%),
steariensuur (C18:0; 34.98%), oleïensuur (C18:1 cis; 48.05%), linoleïensuur (C18:2;
6.60%) en aragiedsuur (20:0; 1.73%). Spoorhoeveelhede C16:1 (0.56%), C18:1 trans
(0.25%), C18:3 (0.43%), C20:1 (0.25%) en C22:0 (0.40%) is ook geïdentifiseer.
Die vars geperste MKF het 'n peroksiedwaarde van 2.7 meq.kg", 'n CD
waarde van 0.07% en 'n AV waarde van 2.2 mmol.kg" getoon. Na afloop van 40 dae
opbergingsperiode by 5, 15, 25 en 40°C het die PV van MKV met 'n beperkte
blootstelling aan suurstof na 5 meq.kg" vermeerder, terwyl die waardes van monsters
sonder suurstofblootstelling na 4 meq.kg" vermeerder het. Emulsifisering van MKV
het 'n stabiliserende effek (maksimum PV = 2.8 meq.kg") terwyl blootstelling aan
ultraviolet (UV) lig 'n kataliserende effek (maksimum PV = meq.kgl ) op oksidasie
gehad het. Hierdie maksimum waardes het na 40 dae afgeneem. Die CD waardes van
MKF monsters opgeberg by 5, 15, 25 en 40°C en met beperkte blootstelling aan
suurstof het vermeerder tot 0.18% terwyl die monsters sonder suurstofblootstelling by
bogenoemde temperature vermeerder het tot 0.16% na 40 dae. Die gekonjugeerde
dieen waardes van die monsters blootgestel aan UV lig het vermeerder tot 0.20%;
terwyl die geëmulsifiseerde monsters waardes soortgelyk aan die MKV monsters
sonder blootstelling aan suurstof getoon het. Gekonjugeerde dieen waardes het
gestabiliseer vanaf dag 40. Die p-anisidienwaardes van MKV monsters opgeberg by
temperature van 5,15, 25 en 40°C, met en sonder blootstelling aan suurstof, het
varieer tussen 0.5 en 5 mmol.kg". Die swak korrellasie tussen nonanal produksie en
p-anisidienwaardes, sowel as die klein hoeveelhede 2-alkenale geproduseer, was
verantwoordelik vir hierdie lae en nie linêere waardes. Die peroksied, gekonjugeerde dieen en p-anisidienwaardes wat verkry is nadat MKV
by 25°C in 240 opgeberg is, was laag weens die klein persentasie poli-onversadigde
vetsure teenwoordig in die vet. Dit vergelyk goed met die hoë waardes wat verkry is
vir sonneblom-, canola- en olyfolie wat almal ryk aan poli-onversadigde vetsure is.
Die minimale veranderinge in die vetsuurprofiel van MKF dui op die
stabiliserende invloed van versadigde vetsure (C16:0 en C18:0) en oleïensuur. Die
onstabiliteit van linoleïen- en lineensuur duidelik uit hierdie vetsure se oksidatiewe
verval. 'n Afname van 7.41% en 12.80% is waargeneem tussen dae 0 en 240 vir die
C18:2/C16:0 en C18:2/C18:0 verhoudings, onderskeidelik.
Die voorspelling van die oksidatiewe status van die MKF monsters met behulp
van FT-NIR spektroskopie was moontlik deur die ontwikkeling van kalibrasie
modelle. 'n Totale datastel van 300 monsters, waarvan ongeveer 'n derde vir
validasie aangewend is, is gebruik vir die kalibrasiemodelle. Met behulp van PCA
(hoojkomponent analise) kon drie klassifiseerbare groepe by 0, 40 en 80-240 dae
onderskei word. Die beste kalibrasiemodel vir PV het 'n standaardfout van
voorspelling (SEP) van 0.46 meq.kg", 'n korrellasiekoëffisient (r) van 0.95, 'n
oorhelling van 0.02 en 'n standaardfout van voorspelling (RMSEP) van 0.46 meq.kg"
gehad. Die CD kalibrasiemodel (geen voorafverwerking) het 'n r van 0.89, SEP van
0.01% oorhelling van 0.001 en RMSEP van 0.01% gehad. Die AV kalibrasie het 'n
SEP van 0.32 mmol.kg', oorhelling van 0.03, RMSEP van 0.32 mmol.kg" en r van
0.93 gehad. Die C18:2 en C18:3 modelle is saamgestel deur PLS (partial least
squares) regressie. Waardes verkry vir C18:2 en C18:3 was onderskeidelik: SEP
0.32% en 0.05%, RMSEP, 0.32% en 0.05%, oorhelling 0.05 en 0.01 en r 0.82 en 0.54.
In die geval van C18:1, C18:0 en C16:0, het die kalibrasies swakker statistiek
korrellasies getoon. Goeie korrellasies is verkry tydens kalibrasie vir die C18:2/C16:0
en C18:2/C18:0 verhoudings.

Identiferoai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:sun/oai:scholar.sun.ac.za:10019.1/52925
Date12 1900
CreatorsSchoeman, Mathilda Elizabeth
ContributorsManley, M., Stellenbosch University. Faculty of AgriSciences. Dept. of Food Science.
PublisherStellenbosch : Stellenbosch University
Source SetsSouth African National ETD Portal
Languageen_ZA
Detected LanguageUnknown
TypeThesis
Format255 pages : illustrations
RightsStellenbosch University

Page generated in 0.003 seconds