LIPID STABILITY IN SOYFLOURS PRODUCED FROM RAW AND
PROCESSED SOYBEANS.
A THESIS PRESENTED TO THE
BY
ROSEBONSI 
BSc. (GHANA) 1997
IN PARTIAL FULFILMENT OF THE REQUIREMENT FOR THE DEGREE 
MASTER OF PHILOSOPHY IN FOODSCIENCE
JUNE 2001
Gj-365723
T P  Jf3.8-.Sfc> 6
Ubc,c. l
iCjOfiS
DECLARATION
1 hereby dcclare that with the exception o f  literature cited the information in this 
document was produced by me through research under supervision in the 
Department o f  Nutrition and Food Science, University o f  Ghana and the Food 
Research Institute, Ghana.
ROSE BONSI PROF. G. S. AYER NOR
(CANDIDATE) (SUPERVISOR)
DEDICATION
To Abena Tcnewaa and NanaYaw Baabu.
ABSTRACT
Soybean, a high protcin/oil legume is known (o contain high levels of  
polyunsaturated fatty acids, which are susceptible to lipid oxidation and 
hydrolysis. However, suitable processing methods and storage conditions can 
enhance the stability o f  lipid in soyflour.
The objective o f  the project was to investigate the effects o f  some processing 
methods and storage conditions on the stability o f  lipid in soyflour and to define 
the parameters for the prediction o f  the shelf life soyflour.
Some whole commercial soybeans were milled into raw soyflour. Some o f  the 
whole beans were cooked for one hour, and milled into cooked-dried soyflour. 
The final portion was roasted in an open pan and then milled into roasted 
soyflour. The raw, cooked-dried and roasted flours produced were stored at 
temperatures o f  5, 16, 30, 42, 68 and 80°C, and at water activities o f  0.15, 0.23,
0.45, 0.68 and 0.75 for 12 weeks. Indices o f  lipid oxidation (peroxide value and 
ihiobarbituric acid number) and an index o f  lipid hydrolysis (free fatty acids) 
were determined at time intervals o f  0, 2, 6, 10 and 12 weeks. Sensory 
evaluation was performed on the samples stored at 5 and 30°C at storage times 
o f  4, 8,10 and 12 weeks.
Results showed that the rates o f  lipid oxidation and hydrolysis were higher in 
raw soy flour than in the heat-processed flours at the same storage temperature 
and time and at (he same water activity. The rale o f  lipid hydrolysis was found
i i i
to be minimal at 5°C and maximal at 3-0 °C in both raw and heat-processed 
soy II ours. The rate o f  lipid oxidation was also found lo be minimal at water 
activity corresponding to the average monolayer value o f  both raw and heat- 
processed soydours. The flavour of  raw and heat-processed Hours began to 
change significantly after 6 and 12 weeks of  storage respectively; and this 
occurred at peroxide value o f  4.21 meq/kg and thiobarbituric acid number o f  
9.76 mg/kg.
Hcat-treatment o f  soybeans prior to processing into Hours yield Hours with 
lower rates o f  chemical reactions, which result in minimal lipid oxidation and 
hydrolysis. Storage o f  soy Hour under cold condition or maintaining (he moisture 
content o f  soyllour at or close to the monolayer value o f  soy Hour results in 
lower rates o f  chemical reactions, which increases the stability o f  lipids in 
soyHour. The shelf life of  soyllour can be predicted when the peroxide value 
and the TBA Number o f  soyHour are known. It is possible to extend the shelf 
life o f  soyHour by heat-processing the beans prior to milling into Hour and 
storing the flour under cold conditions.
IV
ACKNOWLEDGEMENT
I wish lo thank the Lord Almighty for the sound mind and strength given me 
towards the successful completion of this work. I owe thanks to Professor G.S. 
Aycrnor who supervised the execution o f  this work thoroughly and provided 
a lot o f  scientific advice.
I remain particularly grateful to my parents Mr. And Mrs. Bonsi and my 
husband Mr. Paul Omari for their prayers and moral support.
I am indebted to Dr. P.N.T. Johnson of the Food Research Institute for his 
academic and technical support. I continue to be indebted to all who helped in 
various ways during the execution o f  this work. 1 am particularly grateful to 
Emmanuel Afoakwa for editing this work.
Finally I wish to thank the Bank o f Ghana for awarding me a Fellowship, which 
contributed immensely to the completion o f  this work.
May God bless you All.
TABLE OF  CONTENT
PAGE
DECLARATION...............................................................................................................i
DEDICATION.................................................  ii
ABSTRACT..................................................................................................................... ui
ACKNOWLEDGEMENT.............................................................................................v
LIST OF FIGURES...................................................................................................... xii
LIST OF TABLES...................................................................................................... xvi.
INTRODUCTION
1.1 History o f  soybeans :................................................................ 1
1.2 Uses o f  soybeans.................................................................................................. 1
1.3 Nutritional contribution o f  soybeans...............................................................2
1.4 Some problems associated with soybean and soybean products.............. 3
1.5 Main objective......................................................................................................5
1.5.1 Specific objectives.............................................................................5
LITERATURE REVIEW
2.1 Soybean.................................................................................................................... 7
2.1.1 Advantages of soybean................................................................... 7
2.1.2 Disadvantages o f  soybean..............................................................8
2.2 Components of  interest in soybean.................................................................. 8
2.2.1 Protein................................................................................................. 8
vi
2.2.2 Fat...................................................................................................: -9
2.2.3 Carbohydrates.................................................................................... 9
2.2.4 Vitamins and minerals.......................... ; ...................   10
2.2.5 Enzymes.............................................................................................. 11
2.2.6 Phytales............................................................................................... II
2.2.7 Protease inhibitors........................................................................... 11
2.2.8 Hemagglutinins................................................................................ 12
2.2.9 Miscellaneous components........................................................... 12
2.3 Chemical composition o f  soybean oil......................................................... 12
2.3.1 Component fatty acids..................................................................13
2.3.2 Glycerides.........................................................................................14
2.3.3 Phospholipids.................................................................................. 14
2.3.4 Unsaponifiables................................................................................14
2.4 Problems associated with soy and soy products...................................16
2.4.1 Oxidation of  lipids in soybean and soybean products 17
2.4.1.1 Factors which affect lipid oxidation........................ 17
2.4.1.2 Types o f  lipid oxidation................................................23
2.4.1.3 Products o f  lipid oxidation............................................. 31
2.4.1.3.1 Reaction o f  oxidized
lipids with proteins..............................................32
2 .4 .1.4 Effects o f  lipid oxidation on the nutritive
value of' food......................................................................36
vi i
2.4.1.5 ni'lbct o f  lipid oxidation on the
sensory value of  food................................................................ 37
2.4.1.6 Effect o f  lipid oxidation on health............................ 39
2.4.1.7 Laid down guidelines for avoiding rancidity 43
2.4.1.8 Indices o f  lipid oxidation and rancidity...................43
2.4.1.8.1 Other indirect indices.................................. 45
2.5 Soybean processing and products............................................................ 46
2.5.1 Soybean flour............................................................ 46
2.5.1.1 Full-fat flour..................................................46.
2.5.1.2 Low-fat flour..................................................47
2.5.1.3 Deffated flour.................................................. 48
2.5.2 Protein concentrates and isolates.................................. 50
2.5.3 Other soybean products................................................... 51
2.5.4 Effect o f  processing on the
nutritional value o f  soybean........................................51
2.5.5 Effect o f  processing on the flavour
of soybean........................................................................ 52
2.5.6 Utilization o f  soybean flour
in other dry foods............................................................52
MATERIALS AND METHODS
3.1 Materials...................................................................................................................56
3.2 Methodology............................................................................................................ 56
3.2.1. Preparation o f  samples........................................................ 56
3.2.2 Chemical analyses............................................................. 57
3.2.2.1 Total lipids.............................................................57
3.2.2.2 Crude fat................................................................ 59
3.2.2.3 Free latty acids.....................................................59
3.2.2.4 Peroxide value.....................................................59
3.2.2.5 Thiobarbituric acid number................................60
3.2.2.6 Moisture content.................................................. 61
3.2.3 Chemical analyses o f  stored samples..............................................61
3.2.3.1 Investigation o f  the effect o f  storage 
temperature and time on lipid stability
in the various stored soyflours.......................61
3.2.3.2 Investigation of  the effect o f  relative 
humidity and time on lipid stability in the 
various stored soyflours.......................................63
3.2.3.2.1 Determination o f  moisture
sorption isotherms................................................63
3.2.3.2.2 Determination o f  the effect o f  water activity 
on lipid stability in soyllour............................. 65
3.2.4 Sensory evaluation.............................................................................67
ix
3.25 Statistical analyses 67
RESULTS AND DISCUSSION
4.1 General observation orsoyllour samples............................................68
4.2 Particle size analyses................................................................................68
4.3 Chemical analyses o f  freshly prepared soyflour samples...............69
4.4 Chemical analyses of  stored soyflour samples.................................. 72
4.4.1 Effect o f  storage temperature and time on lipid stability .. ..72
4.4.1.1 Effect o f  storage temperature and time on
free fatty acids in soyflour........................................... 74
4.4.1.2 Effect o f  storage temperature and time on the P.V. and 
TBA no. in stored soyflour..............................................77
4.5 Effect o f  water activity and storage time on
lipid stability in stored soyflour......................................................................... 93
4.51 Moisture sorption isotherms and B.E.T. monolayer values........ 93
4.5.2 Effect o f  water activity and time on the P.V. and
TBA no. o f  stored soyflours..............................................................98
4.5.3 Effect o f  water activity and storage lime on
FFA in stored soyflours...................................................................110
4.5.4 Effect o f  lipid oxidation and hydrolysis on
total lipids and erode fat conlcnl in soyllour .......................115
4.6 Sensory evaluation o f  soyllour samples.....................................................119
CONCLUSION AND RECOMMENDATIONS
5.1 Conclusion........................................................................  124
5.2 Recommendations................................................................................................124
References...............................................................................................................126
Appendices............................................................................................................. 136
xi
LIST OF FIGURES
Page
Fig 3.1 Flow charts for the processing o f  roasted soyflour and
cooked-dried soyflour...............................................  58
Fig 4.1 Particle size analysis o f  soyflours..............................................................70
Fig 4.2 Free fatty acids in raw, cooked-dried and roasted soyflours
stored at different temperatures for twelve w eek s ......................................... 75
Fig 4.3 3-D response surface plots for FFA in raw soyflour
as a function of  temperature and lime................................................................ 78
Fig 4.4 3-D response surface plots for FFA in cooked-dried soyflour
as a function o f  temperature and lime................................................................ 79
Fig 4.5 3-D response surfacc plots for FFA in roasted soyflour
as a function o f  temperature and time................................................................ 80
Fig 4.6 Effect o f  storage temperature and time on the P.V.
o f  raw soyflour..........................................................................................................82
Fig 4.7 Effect o f  storage temperature and time on the TBA no.
o f  raw soyflour..........................................................................................................82
Fig 4.8 Effect o f  storage temperature and time on the P.V.
o f  cooked-dried soyllour........................................................................................ 83
Fig 4.9 Effect o f  storage temperature and time on the I ]i A no.
o f  cooked-dried soyllour.........................................................................................83
Fig 4.10 Effect o f  storage temperature and time on the P.V.
o f  roasted soy flour.................................................................................................84
Fig 4.11 Effect o f  storage temperature and time on the TBA no.
o f  roasted soyflour.................................................................................................84
Fig 4,12 3-D response surface plots for P.V. in raw soyflour
as a function o f  temperature and time............................................................... 87
Fig 4.13 3-D response surface plots for P.V. in cooked-dried soyflour
as a function o f  temperature and time............................................................... 88
Fig 4.14 3-D response surface plots for P V. in roasted soyllour
as a function o f  temperature and time............................................................... 89
Fig 4.15 3-D response surface plots for TBA no. in raw soyflour
as a function o f  temperature and time............................................................... 90
Fig 4.16 3-D response surface plots for TBA no. in cooked-dried soyflour
as a function o f  temperature and time............................................................... 91
Fig 4.17 3-D response surface plots for TBA no. in roasted soyllour
as a function o f  temperature and time............................................................... 92
Fig 4.33 B.E.T plot for raw soyflour at 30°C.......................................................94
Fig 4.34 B.E.T. plot far cooked-dried soyllour at 30°C.................................... 95
x i i i
Fig 4.35 B.O.T. plot for roasted soyflour at 30°C .............................................96
Fig 4.18 Moisture sorption isotherms for raw, cooked and
roasted soyflours at 30°C..............................................................................97
Fig 4.19 P.V. o f  raw, cooked-dried and roasted soyflour:- stored at
various water activities at 30°C for a period of  twelve weeks..99
Fig 4.20 TBA no. o f  raw, cooked-dried and roasted soyflours stored at
various water activities at 30°C for a period o f  twelve weeks..100
Fig 4.21 3-D response surface plot for P.V. in raw soyflour as
a function o f  water activity and time............................................................104
Fig 4.22 3-D response surface plot for P.V. in cooked-dried soyflour as
a function o f  water activity and time...........................................................105
Fig 4.23 3-D response surface plot for P.V. in roasted soy 11 our as
a function of  water activity and time...........................................................106
Fig 4.24 3-D response surface plot for TBA no. in raw soy Hour as
a function o f  water activity and time...........................................................107
Fig 4.25 3-D response surface plot for TBA no. in cooked-dried soyflour as
a function of  water activity and time...........................................................108
Fig 4.26 3-D response surface plot for TBA no. in roasted soyflour as
a function of  water activity and time...........................................................109
Fig 4.27 Effect o f  water activity and storage time on the FFA
x i v
in raw soyflour at 30°C..................................................................................111
Fig 4.28 Effect o f  water activity and storage time on the FFA
in cookcd-dried soyflour at 30°C.................................................................1 12
Fig 4.29 Effect o f  water activity and storage time on the FFA
in roasted soyflour at 30°C............................................................................. 113
Fig 4.30 3-D response surface plot for raw soyflour as a
function o f  water activity and time.................................................................116
Fig 4.31 3-D response surface plot for cooked-dried soyflour as a
function of  water activity and time................................................................ 117
Fig 4.32 3-D response surface plot for roasted soyflour as a
function o f  water activity and time................................................................ 118
x v
Page
Table 2.2 Typical composition of soybean and the various
soybean products (%) DMB.......................................................................11
Table 2.3 Percentage ranges o f  component fatty acids o f  soybean oils ..............14
Table 2.4 Component fatty acids o f  phospholipids o f  soybean oils....................16
Table 2.5 Component fatly acids o f  phospholipids and glycerides
o f  soybean oils................................................................................................. 17
Table 2.6 Tocopherol levels in soybean......................................................  22
Table 2.7 Hydroperoxides and aldehydes, which may be
formed in autoxidation o f  some unsaturated fattj' acids..................... 28
Table 2.8 Reactive groups taking part in the interaction
of oxidized lipids with proteins................................................................. 37.
Table 2.9 Flavour threshold values for oxidation products in
soybean.........................................................................................................40
LIST OF  TABLES
Table 2.10 Flavour descriptions used for crude, processed
and reverted soybean oil..............................................
Tabic 2.11 Proximate Analyses of  Commercial Soybean
41.
Table 2.12 Proportions of  phospholipid fractions in soyllouras a
percentage o f  total lipids.......................................................................... 51
Table 2.13 Percentage composition o f  fatty acids in isolated fractions o f  bound
lipids present in defatted soyflour......................................................... 52
Table 2.14 Effect o f  heat treatment on protein efficiency ratio (P.E.R.),
biological value (B.V.), and digestibility of  soybeans
and soybean products................................................................................... 55
Table 2.15 Effect o f  steaming on flavour o f  soyflour............................................56
Table 2.16 Protein and fat contents and protein value of
different maize - soy preparations.............................................................. 58
fable 3.1 Design matrix and variable combination in experimental set up .........62
Table 3.2. Variables and their levels used in the central composite
rotatable design for k=2....................................................................62
Table 3.3 Water activity of  selected Saturated salt solutions at 30°C.................. 65
Table 3.4 Variables and their levels used in the central composite
rotatable design for k=2...................................................................................66
Table 4.1 Moisture content, crude fat content, free fatty acids hydrolytic and 
oxidative products in freshly prepared raw and processed 
soyflours............................................................................................................. 71
Table 4.2 Effect o f  storage temperature and time on the moisture
x v i i
con lent o f  packaged roasted soyflour........................................................ 73
Table 4.3 Effect o f  storage temperature and time on the moisture
content o f  packaged raw soy llour................................................... 73
Table 4.4 Further ANOVA for variation in the order filled
(showing only the P-values) for FFA content in soj'flours
stored at various temperatures over a twelve-week period..................... 76
Table 4.5 Further Anova for variation in the order fitted
(showing only the P-values) for Peroxide value in soyflours
stored at various temperatures over a twelve-week period........................86
Table 4.6 Further Anova for variation in the order fitted
(showing only the P-values) for Peroxide value in soyflours
stored at various temperatures over a twelve-week period.................... 86
Table 4.7 B.E.T. monolayer values for different types
o f  soyflour at 30°C ............................................................................................93
Table 4.8 Further Anova for variation in the order fitted
(showing only the P-values) for peroxide value in soyflours
stored at various water activities (aw) over a twelve-week period 102
Table 4.9 Further Anova for variation in the order fitted
(showing only the P-values) for TBA no. in soyflours 
stored at various waler activities (aw) over a twelve-week period... 102 
Table 4.10 Further Anova for variation in the order filled
(showing only the P-values) for FFA in soyflours
stored at various water activities (aw)
over a twelve-week period.........................................................................114
Table 4.11 Total lipids in various forms o f  freshly prepared soyflour
as well as those stored for twelve (12) weeks at 30CC .......................115
Table 4.12 Crude fat content in various forms of freshly prepared
soyflours as well as that stored for twelve (12) weeks at 30°C ........119.
Table 4.13 Sensory perceptions of  flavour changes in various
types of  soyflour stored at 5°C ............................................................121
Table 4.13 Sensory perceptions o f  flavour changes in various
types o f  soyflour stored at 30°C............................................................122
x i x
CHAPTER ONE 
INTRODUCTION
1.1 History of soybean
Soybean (Glycine max) is a piolcin/oil legume that originated from the South- 
Eastern Asian countries. It grows best in warm climates and matures within five 
months'(Scott and Aldrich, 1970). Soybean also known as the “miracle crop” or 
the “golden bean” is one of  the oldest o f  all food plants, and a modern-day 
success with many uses (Thio, 1975). It has been used for human food in China 
since long before the birth of  Christ (Scott and Aldrich, 1970). More recently, 
soybean has become the most important source o f  edible oil in the western world. 
In the Far East where animal protein remains a luxury, soybeans have been a 
prominent part o f  the diet for centuries (Scott and Aldrich. 1970). Soybean was 
first introduced in Ghana in 1909 for cultivation as both cash and food crop 
(Mercer-Quarshic and Nsowah, 1975).
1.2 Uses o f soybean
Soybeans are used for the preparation of  animal feed, human food and for some 
industrial applications. Well over 90% o f  soybean meal was used in livestock 
feeding, however, the proportion used in human food has continued to increase 
since 1929 when the first soyflour and grits were produced commercially (Scott 
and Aldrich, 1970). Soybean is also an important source o f  edible oil in the 
western world (Perrier, 1975).
I
In the Orient, soybeans arc traditionally used as food in two ways, that is, 
production o f  unfermcnted foods such as tofu (soymilk curd) and production o f  
fermented foods such as miso and lempeh (Perrier, 1975). There are three distinct 
categories o f  soy protein products and these include soyflour and grits, protein 
concentrates and protein isolates (Ferrier, 1975). These products can be used as 
ingredients in bakery products, baby foods, meat products and confections.
In recent times, in Ghana, the use o f  soybean in human diet is being promoted 
since soybean has been observed to reduce the incidence o f  protein deficiency. It 
is for instance used to fortify low protein foods such as cassava products. Soybean 
flour is used a lot in the preparation of  infant and weaning foods, such as Cere lac 
maize-soya  produced by Nestle7 Ghana Limited.
1.3 Nutritional contribution of soybean
Soybean has been recognised as the best and cheapest source o f  food energy in 
terms o f  calories per unit cost o f  production provided it is consumed directly 
(Hammond and Call, 1970). Soy protein for human food has been shown to be o f  
the highest nutritional value o f  all the well-known plant proteins (Robinson. 
1971). It is particularly high in lysine, an essential amino acid, which is usually 
limiting in cereal-based diets, and would thus complement other plant proteins 
(Lin el a!., 1975). Properly processed soyflour generally has protein efficiency 
ratio (P.E.R) of  2.2-2.3 compared with 2.5 for a cascin standard (Williams, 1970).
Soybean contains 36-44% proteins, 30-35% carbohydrate?, 20- 25% (at, vitamins 
and minerals such as calcium and phosphorus (Waggle d  nl., 1981).
1.4 Some problems associated with soy and soy products
The greatest obstacle to the general use o f  soybean as a source o f  food for human 
include:
i. the bitter, beany taste,
ii. objectionable odour,
iii. poor keeping qualities,
iv. presence o f  anti-nutritional factors such as trypsin inhibitors,
v. high tendency to undergo rancidity and
vi. presence o f  hard-and-inlimate seed coat, which leads to prolonged cooking 
time (Ayernor, 1993).
The unsaturated fatty acids in soybean lipids are highly susceptible to oxidation 
(Hamilton and Berger, 1995^. Rackis cl a!. (1979) reported that even defatted soy 
(lour contains residual fat and a considerable amount o f  bound lipids, all o f  which 
arc susceptible to oxidation. Lipid oxidation which occurs via autoxidation, 
thermal oxidation, enzymatic oxidation and photo-oxidation leads to problems 
such as flavour deterioration and destruction o f  some nutrients including
unsaturated fatty acids and vitamin E (Hamilton and Berger, 1995). Kamel and
Kakuda (1995) reported that oxidation products o f  polyunsaturated laity acids 
could be carcinogenic or co-carcinogenic.
It has also been reported that soybean lipids are susceptible to hydrolysis as a 
result o f  the presence o f  considerable level o f  lipases. The long-chain fatty acids 
produced from lipid hydrolysis are responsible for the development o f  soapy 
flavours in soy products (Satouchi and Matsushita, 1976).
Certain factors such as amount o f  oxygen present, degree o f  unsaturation o f  lipids, 
presence o f  antioxidants, presence of  pro-oxidants, nature o f  packaging material, 
moisture content, light exposure and temperature of  processing and storage have 
been shown to affect the rate o f  lipid oxidation (deMan, 1990). One successful 
way o f  improving lipid stability in soy products is to adequately control some of 
these factors (deMan, 1990).
This project focused on studying the effect o f  processing, storage temperature and 
relative humidity on lipid stability in soyflour and how these factors could be 
regulated to improve lipid stability and hence retard the development o f  flavour 
changes in soyflour.
1.5 Main objective
The main objectives of  the study were to investigate (he effects o f  some 
processing methods and storage conditions on lipid stability in soyflour and to 
define the parameters for the prediction o f  the shelflife o f  soyflour .
1.5.1 Specific objectives
The specific objectives o f  the study were as follows:
1. to process soybean into various soyllour products using different processing 
methods.
2. to analyse the various soyflour samples obtained in (1) above, for indices o f  
lipid stability including:
(a) Total lipids
(b) Fat content
(c) Free fatty acids
(d) Peroxide value
(e) Thiobarbituric acid (T.B.A.) Number
3. to study (he effect of:
(a) storage temperature and time on lipid stability, and crude fat and total 
lipids contents in (he various soyflours.
(b) relative humidity and time on lipid stability under storage in the various 
soyllours by determining the peroxide value and the TBA number as 
indiccs for lipid oxidation, and free fatty acids as an index for lipid 
hydrolysis.
5
4. lo establish when flavour changes become detectable during storage o f  the 
soyflours by sensory analysis.
5. to establish the most suitable processing methods, favourable storage 
conditions, and shelf  life for soy (lour.
6
CHAPTER TWO  
LITERATURE REVIEW
2.1 SOYBEAN
Soybean is a protein/oil legume that originated from the South-Eastern Asian 
countries. It grows best in warm climates and matures within five months (Scott 
and Aldrich, 1970). Both traditional and new soybean products are marketed 
extensively in Bangladesh, India and other countries in the Far East (Markley, 
1951). Soybean has proved to be a versatile and highly successful food product in 
meeting food shortages in Europe and Asia during and since World War II 
(Markley, 1951). Soybean products are used as valuable ingredients in Occidental 
diets (Markley, 1951). Soybean has been shown to be o f  the highest nutritional 
value o f  all the well-known plant proteins (Robinson, 1971).
2.1.1 Advantages of soybean
a. It has higher amino acid content and better nutritive value than most vegetable 
proteins (Robinson, 1971). When supplemented with methionine it can be 
equivalent to good quality animal protein (Robinson, 1971).
b. It is relatively easy to store and transport (Scott and Aldrich, 1970).
c. Technology for processing into food is well developed (Robinson, 1971).
d. It produces good quality oil in addition to proteins (Robinson, 1971).
7
e. Il is leguminous and fixes atmospheric nitrogen, thus requiring little or no 
nitrogen fertilizer (Scott and Aldrich, 1970).
f. Its indigestible carbohydrates and possibly its hull prove useful as a source 
o f  dietary fibre (Wolf, 1970).
2.1.2 Disadvantages of soybean
a. Its products have distinctive beany flavour and odour, which are 
objectionable in some food applications (Scott and Aldrich, 1970).
b. Proteins from soybean do not have the same desirable functional properties in 
food systems as do animal protein such as casein (Wolf, 1970).
c. It contains antinutritional factors, which must be inactivated before used for 
food (Wolf  1970).
2.2 Components of interest in soybean
2.2.1 Protein
Soybean contains 36-44% protein on dry matter basis (Wolf, 1970). The chief 
protein in soybean is a globulin, glycinin, which is very similar to that o f  cow ’s 
milk (Wolf, 1970). Soybean proteins have high biological value and contain all 
the essential amino acids with exceptionally high levels o f  lysine but low levels o f  
sulphur - containing amino acids: methionine and cysteine, and tryptophan (Wolf. 
1970). The high level o f  lysine in soybean enables it to complement cereal-based 
diets, which arc usually low in lysine. Properly processed soyllour and grits 
generally have a protein efficiency ratio of  2.2 -  2.3 compared with 2.5 for a
8
casein standard (Williams, 1970). It has been shown that soy proteins can be used 
to supplement and/or extend meat and fish proteins and to supplement single or 
mixed vegetable-based protein diets (Bressani, 1981). Soybean protein-fortilied- 
foods have been shown to be useful in relieving malnutrition among specified 
segments o f  the population such as infants and weaned children (Williams, 1970).
2.2.2 Fat
Soybean has a high fat content between 20 and 25% depending on the soil type, 
environmental factors and variety (Markley, 1951). The fat is made up o f  15% 
palmitic acid and 80% unsaturated fatty acids, that is, 70% oleic, 24% linoleic and 
6% linolenic acids (Wolf, 1970). The fat contains no free fatty acids and consists 
most entirely o f  natural triglycerides (Markley, 1951).
2.2.3 Carbohydrates
Soybean contains 30-35% carbohydrates. The carbohydrates can be subdivided 
into soluble and insoluble fractions. The soluble carbohydrates include hexose, 
sucrose and oligosaccharides such as raffinose, stachyose and verbacose while the 
insoluble fraction include hcmicellulose, celluloses, lignin, pectin, and other 
complex carbohydrates (Waggle el al., 1981). The oligosaccharides verbacose, 
stachyose and raffinose are the major contributors to soybean flatulence (Rackis 
el al., 1979). These Components escape digestion and arc fermented by intestinal 
microflora to produce excess amounts o f  intestinal gases.
<)
2.2.4 Vitamins and Minerals
Green soybean has carotene content ranging from 2 - 7u/g while that oi the 
mature bean is about 0 - 8u/g (MarkIcy. 1951). The immature soybean is richer in 
riboflavin, niacin and vitamin C than the mature bean, and pyridoxin and thiamine 
concentrations increase with maturation (Wolf, 1970). Soybean contains 0.12% -
0.21% tocopherols (Markley, 1951). It has fairly high amount o f  calcium but in 
relation to animal requirements this is not adequate for the young animal 
(Markley, 1951). It has adequate amount of  potassium, phosphorus and 
magnesium. The most notable vitamin deficiency in soy protein is the low level 
o f  vitamin B12, which is virtually absent in all vegetable proteins (Markley, 
1951).
Table 2.2 Typical composition of soybean and various soybean  
products (%) [DMB],
Protein * Fat Carbohydrates Ash Crude fibre Moisture
Whole soybean 42 20 30.5 5.0 5.5
Full-fat soyllour 40 20 30.5 5.0 5.5
Defatted soyflour 54 1.0 35.5 6.0 3.5 6.0
Soy  p r o t e i n  concentrate 70 1.0 21.5 5.0 3.5 6.7
Soy p r o t e in  isolates 92 0.5 2.2 4.5 0.3 4.7
* N X  6.25
Source: Waggle el al. (1981)
10
2.2.5 Enzymes
Enzymes in soybean include cc and |3 amylases, lipases, lipoperoxidases, 
lipoxygenases, urease and proleinases (Whitaker, 1972). The lipases, 
lipoxygenase and lipoperoxida.se catalyse the oxidation o f  lipids and may result in 
flavour changes in improperly processed soy oils and so}' products (Rackis et al., 
1979). Lipoxygenase is the most important enzyme involved in enzymatic 
peroxidation o f  soybean lipids (Satouchi and Matsushita, 1976).
2.2.6 Phytatcs
Defatted soybean meal contains about 1.5% phytic acid, which is associated 
mainly with the germ o f  the intact seed (Erdman et al., 1981). Processing of  
soybean can produce protein phytic mineral complexes that may reduce the 
bioavailability o f  minerals (Erdman et a l ,  1981). Soaking o f  bean reduces phytic 
acid content through enzymatic hydrolysis by the action o f  the enzyme phytase, 
which is an indigenous constituent (Rackis et al., 1979).
2.2.7 Protcase inhibitors
Soybean contains several different factors that inhibit the action o f  the digestive 
enzymes, trypsin and chymotrypsin, thereby decreasing protein digestion. 
Heating o f  soy meal improves the nutritional value as a result o f  reduced protease 
inhibitor activity (Kakade el a l., 1974).
2.2.8 Hemagglutinins
.Soy hemagglutinins are glyco-protcins with molecular weight of  approximately
110,000 and contain about 5% carbohydrates (Anderson el al., 1979). They are 
responsible for agglutination o f  red blood cells (Lin cl al., 1975). Soy 
hemagglutins do not appear to be a problem in properly processed soyflour and 
other soy products used for human food (Rackis el a l, 1979).
2.2.9 M iscellaneous components
These components include saponins. possibly sterols and triterpene alcohols, and 
an antithyrotoxic factor (Rackis el al., 1979). Other biologically active substances 
are soybean allergenic compounds and lysinoalanine, which is not a constituent o f  
raw soybean but are often found in processed soy protein products (Rackis el al.,
1979). Less o f  the population is allergic to soy-milk than cow,,; , :,k or milk from 
other legumes (Anderson el al., 1979). Lysinoalanine is formed during alkaline 
treatment o f  soy protein (Struthers el al., 1979). Addition o f  either 
mercaploethanol or cystine to an alkaline solution or proper temperature control 
reduces the formation o f  lysinoalanine (Struthers el al.. 1979). Presence of  
lysinoalanine in food reduces the bioavailability o f  cystine, cysteine and lysine 
(Struthers el al., 1979).
2.3 Chemical composition of soybean oil
Soybean oil may be defined as the composition of  lipid materials cxtractable from 
ground or flaked soybeans, with organic solvents such as hexane, chloroform or
petroleum ether, Lipid can be defined as any substance that furnishes fatty acids 
and/or related compounds on hydrolysis.
2.3.1 Component fatty acids
Soybean oil contains 14.6% Saturated acids consisting solely o f  palmitic acid and 
80% unsaturated fatty acid (Kamel and Kakuda, 1994). The unsaturated acids 
were found to comprise 56% oleic, 19% linoleic, and 4.8% Iinolenie acid 
(Markley 1951).
Table 2.3 Percentage Ranges of Component Fatty Acids o f Soybean Oils
Saturated Acids (AV. 15%) Unsaturated Acids (AV. 85%)
Acid Percent Acid Percent
Laurie 0.2 Lauroleic -
Myristic 0.4 Myristolcic 0.64
Palmitic 11.0 Palmitoleic 1.60
Stearic 4.0 Oleic 25
Arachidic 0.9 Linoleic 51
Behenic tracc Linolenic 9
Arachidonic Trace
Source: Kamel and Kakuda (1995).
2.3.2 Glyccrkles
Soybean oil is composed predominantly of  a mixture of glycerol esters or 
triglycerides. In soybean oil, the five principal fatty acids, which form the various 
types o f  triglycerides are palmitic, stearic, oleic, linoleic and Underlie acids 
(Markley, 1951). The component glycerides o f  soybean oil were found to 
comprise 58% diunsaturated monosaturated triglycerides and 42% triunsaturated 
triglycerides (Markley, 1951).
2.3.3 Phospholipids
Phospholipids or phosphatides are lipids containing phosphorus and in many 
instances nitrogen. Soybean phospholipids include lecithin (Phosphatidylcholine), cephalin 
(phosphatidyl ethanolamine) and sphingomyelins (Markley, 1951). Soybean 
phospholipids contain about 35% lecithin and 65% cephalin (deMan, 1990). The 
fatty acid composition o f  these phospholipids is usually different from that o f  the 
oil in which they are present. The acyl groups are usually more unsaturated than 
those o f  the triglyerides (deMan, 1990). The cephalins are insoluble or slightly 
soluble in ethanol while the lecithins are relatively soluble in this medium.
2.3.4 Unsaponifinblcs
The unsaponifiable fraction o f  soybean oil, although representing only a small 
proportion o f  total lipids, comprises a diversity o f  components, which include fat- 
soluble pigments and vitamins, anti-and pro-oxidants, and other compounds of  
unknown function (deMan, 1990). The fat-soluble pigments include carolenoid
14
pigments and chlorophyll. The major sterol present in soybean oil is stigmasterol 
(deMan, 1990). Also present are sterol glycerides and other components such as 
tocophcrols, waxes, aldehydes, ketones, alcohols, hydrocarbons, tocoquinones 
and dicarbonyl compounds (Marklcv. 1951).
Tabic 2.4 Component fatty acids of the phosphatidcs of soybean oil
Fatty Acid Phosphatides, % (wt)
Alcohol Insoluble Alcohol Soluble
Myristic
Palmitic 11.7 17.3
Stearic 4.0
Arachidic 1.4
Ilexadecenoic 8.6 5.5
Oleic 5.5 19.0
Linolcic 63.3 53.0
Linolenic 3.7
Unsaturated C2,, 5.5 1.5
Source: Zadernowski el al. (1983)
Table 2.5 Component fatty acids of the phosphatidcs
and glycerides o f soybean oil (%)
Fatty acid Lecithin Glycerides
Palmitic 15.77 6.8-14.8
Stearic 6.30 2.4-5.5
Arachidic 0.00 0.3-0.9
Hexadecenoic
Oleic 12.98 25.9-33.7
Linolcic 62.90 50.7-58.8
Linolenic 2.00 2.1-2.6
Unsaturated C 20-22
Source: Zadernowski cl al. (1983)
2.4 Problems associated with soybean and soybean products
The major problems associated with soybean and soybean products include the 
bitter beany taste, objectionable odour and colour, poor keeping qualities, 
presence o f  anti-nutritional factors such as trypsin inhibitors, high tendency to 
undergo rancidity, and presence o f  hard-and-intimate seed coat which leads to 
prolonged cooking time (Ayernor, 1993). Most o f  these problems can be 
eliminated during processing of  soybean. For instance, the beany taste and 
objectionable odour can be removed by heat treatment to inactivate lipoxygenase, 
which is believed to be responsible for the objectionable flavour (Ferrier, 1975). 
Dehulling the beans also reduces the cooking time.
16
2.4.1 Oxidation o f lipids in soybean and soybean products
Soybean is known lo contain 20 - 25% fat which is made up o f  15% saturated and 
80% unsaturated acids (Markley. 1951). Soybean phospholipids, lecithin and 
cephalin, also contain high proportion o f  unsaturated fatty acids (deMan, 1990). 
These unsaturated acids are highly susceptible lo oxidation (Hamilton and Berger, 
1995). Oxidation usually takes place at the carbon atom next to the double bond 
(deMan, 1990).
2.4.1.1 Factors which affect lipid oxidation
(i) Amount of oxygen present
The higher the amount o f  oxygen present in the food sample the higher the rate of  
lipid oxidation (deMan 1990). Hoffman, (1995) calculated the solubility o f  
oxygen in oil to be 22ml/kg at ambient temperature, that is, 1 mmol/mg. If the 
oxygen reacts specifically with linolenic acids, 2 meq/kg o f  hydroproxide should 
be formed, that is, peroxide value o f  2. The production o f  3-Cis-hexenal from 
these hydroperoxides at the level o f  0.1 mg/kg would result in the oil being 
considered rancid (Hoffman, 1995). The rate o f  oxidation is independent o f  
oxygen concentration at high oxygen partial pressure while it is proportional to 
oxygen concentration at low oxygen partial pressure (Labu/.a, 1975).
(ii) Degree o f unsaturation of the Lipid
The unsaturated bonds present in all fats and oils represent active centers, which, 
among other things, may react with oxygen (deMan, 1990). Although even
17
saturated acids may be oxidized, the rate of oxidation greatly depends on the 
degree o f  unsturation. In the series of  18 carbon atom laity acids 18:0, 18:1, 18:2, 
18.3, the relative rate o f  oxidation has been reported in the ratio o f  
1:100:1200:2500 (deMan, 1990). In general, the greater the degree o f  unsaturation 
the more susceptible are the fatty acids to attack by oxygen. The reaction of  
unsaturated compounds proceeds by the abstraction o f  hydrogen from the cc 
carbon, and the resulting free radical is stabilised by resonance (deMan, 1990). 
Free fatty acids are much more susceptible to oxidation than those bound to 
alcohols (Zadernovvski el al., 1983).
(iii) Prcscncc of pro-oxidants
Pro-oxidants are substances that facilitate lipid oxidation. They include metals 
such as copper and some organic compounds such as heme - containing molecules 
and enzymes such as lipoxygenase (deMan, 1990). Any metal with two valence 
states can be active in initiating autoxidation. In general metals in their higher 
valence stale tend to be more effective as free radical initiators (Hamilton and 
Berger, 1995). When the metal has been converted to its lower valence state it 
can then react with oxygen to form complexes, which may provide singlet oxygen 
(Hamilton and Berger, 1995). Transition metals aid the decomposition reactions 
o f  peroxides by an oxidation mechanism and a reductive mechanism in which the 
metal goes from higher valence state to lower valence state in the former and from 
lower valence slate lo higher valence state in the latter (Hamilton and Berger, 
1995):
is
ROOM + M("+ l ) --------------► RDj + H' + M"h
ROOM + M1" --------------► RO- + O il ‘
M = metal, ROOM = Fatty acid
The two most prevalent metals in food processing, storage and transport are iron 
and copper. Metal catalysts may enter foods via the water or spices used in food 
preparation (Taylor 1984).
(iv) Presence o f  antioxidants
Antioxidants are substances that inhibit lipid oxidation. They could be naturally 
present in food or added to food. Naturally occurring antioxidants in soybeans 
include tocopherols, genistein, diadzein and glycitein (Hamilton and Berger, 
1995). The antioxidants interfere with lipid oxidation by interacting with the 
peroxy free radical and the resultant antioxidant free radical does not initiate or 
propagate further oxidation (dcMan, 1990).
The synthetic antioxidants, which are added to food to prevent lipid oxidation are 
often phenolic compounds such as bulylated hydroxy anisole (BHA) and 
butylatedhydroxytoluene (BUT). Only phenolic compounds that can easily 
produce quinones are active as antioxidants (Pokorny. 1971). At higher 
concentrations antioxidants may have a pro-oxidant effect and one o f  the 
reactions may be as follows:
l<>
A’ + RH  ► AH +  R-
Anlioxidant- Fatty acid Anti- Fat-free
free radical. Oxidant radical’
Carotenoids have been shown to have some stabilizing effect on soybean fat 
under certain conditions (Markley, 1951). The resistance o f  soybean products to 
oxidation can be somewhat increased by the addition o f  acidic compounds which 
apparently act synergistically with the naturally occurring antioxidants in soybean 
(Pokomy el al., 1983). Examples are citric acid, phosphoric, ascorbic and tartaric 
acids. These function by chelating pro-oxidant metals in the product and to some 
extent by inhibiting peroxide decomposition and by regeneration or sparing o f  
primary antioxidants (Hamilton and Berger, 1995).
Tabic 2.6 Tocopherol levels in soybean
Tocopherol Type Level (mg/kg)
oc 116
p 34
y 737
d 275
Source: Hamilton and Berger (1995).
20
A' -I- RH  ► AH +■ R'
Antioxidant- Fntty acid Anti- Fat-free
free radical. Oxidanf radical
Carotcnoids have been shown to have some stabilizing effect on soybean fat 
under certain conditions (Markley, 1951). The resistance o f  soybean products to 
oxidation can be somewhat increased by the addition o f  acidic compounds which 
apparently act synergistically with the naturally occurring antioxidants in soybean 
(Pokorny at al., 1983). Examples are citric acid, phosphoric, ascorbic and tartaric 
acids. These function by chelating pro-oxidant metals in (he product and to some 
extent by inhibiting peroxide decomposition and by regeneration or sparing of  
primary antioxidants (Hamilton and Berger, 1995).
Table 2.6 TocophcroI levels in soybean
Tocopherol Type Level (mg/kg)
OC 116
p 34
y 737
d 275
Source: Hamilton and Berger (1995).
20
(v) Nature of packaging material
Any packaging material, which is permeable to oxygen and other factors such as 
proxidants favours lipid oxygen (dcMan, 1990). To prevent or reduce lipid 
oxidation fat-containing products need to be stored in airtight packaging 
materials. The interaction of  the food product with catalytic metals, which may 
be components o f  the packaging material should be minimised.
(vi) Light exposure
Light has the ability to increase the rate o f  lipid oxidation. Autoxidation is 
increased rapidly in the presence o f  light (deMan, 1990). Photooxidation involves 
reaction o f  an alkene with oxygen in the presene o f  light and a suitable sensitiser 
(Hamilton and Berger, 1995). Photosenstizers include riboflavin, erythrosine and 
methylene blue. Exposure o f  foods in transparent containers in supermarkets 
leads to a great deal o f  oxidation (Hamilton and Berger, 1995).
(vii) Temperature
The rate o f  oxidation o f  lipids approximately doubles for every 10° rise in 
temperature. As a general rule, it can be argued that storage temperature should 
be 5 - 10°C above slip melting point o f  the material (Hamilton and Berger,
1995). At higher temperatures (100 - I40°C), formic acid is produced from 
aldehyde decomposition and can be used to indicate the end of the induction 
period (Hamilton and Berger. 1995).
21
Martz el al. (1975) observed that as moisture content o f  cercal products was 
lowered, the products became rancid much sooner. Sal win (1959) proposed that at 
the B.E.T. monolayer, the water formed a protective barrier preventing the oxygen 
from reaching the underlying unsaturated fats. Ilalton and Fischer (1987) also 
proposed that the water retarded the diffusion o f  oxygen to the sites o f  the 
unsaturated double bonds. The B.E.T. monolayer value in defined as the moisture 
content at which each polar and ionic groups has a water molecule bound to it, to 
form the start o f  a liquidlike phase. Generally, at water activity below the 
monolayer value lipid oxidation rate decreases with increasing water activity 
(Labuza, 1975). The rate reaches a minimum around the monolayer value and 
increases with a further increase in the water activity (Labuza and Karel, 1980). 
The “antioxidant effect” o f  water at low water activity has been attributed to the 
bonding o f  hydroperoxides and hydration o f  metal catalysts whereas the “pro­
oxidant effect” o f  water at higher water activity is due to the increased mobility o f  
reactants (ITeidelbaugh and Karel, 1970).
(ix) Other factors which affcct lipid oxidation
In most foodstuffs, lipids are dispersed in a system containing various non-lipid 
materials such as proteins, sugars or minerals, mostly in presence o f  water 
(Pokorny el al., 1983). Oxidation o f  the lipid fraction is affected by the non- 
lipidic components and interactions occur between oxidised lipids and various 
non-lipidic substances (Pokorny et al., 1983). Among non-1 ipidic substances
(vii) Moisture content
22
several mineral components, especially derivatives o f  transition heavy metals are 
known as pro-oxidants, and various polyphenolic substances as antioxidants. 
Substituted and polyvalent acids behave as synergists o f  the antioxidants as they 
bind heavy metals into inactive complexes. Polysaccharides often protect lipids' 
slerically from oxidation (Pokornv el al., 1983).
(x) Effect of protein on the oxidation of lipids
In a mixture o f  lipids and proteins the reaction course depends on the water 
content. In dry systems the oxidation o f  lipids proceeds slowly during the 
induction period but becomes very rapid in the subsequent stages (Hamilton and 
Berger, 1995). Low peroxide value in mixtures of  protein and lipids are caused by 
rapid destruction o f  hydroperoxides in contact with protein solution (Pokornv el 
al., 1983). The hydroperoxidation of  lipids is probably suppressed by their 
interaction with amine groups o f  protein. When amine groups o f  protein 
molecules are blocked with lormaldelyde, hydroperoxides accumulate on storage 
(Pokorny el al., 1983).
2.4.1.2 Types o f lipid oxidation
Lipid oxidation occurs via autoxidation, photooxidation, enzymatic oxidation and 
thermal oxidation (deMan. 1990). Autoxidation is the most important route to 
oxidation o f  lipids (deMan, 1990).
23
The unsaiurnted bonds present in all fats and oils represent active centers, which 
among other things, may react with oxygen. This reaction leads to the formation 
o f  primary, secondary, and tertiary oxidation products which may make the fat or 
fat-containing foods unsuitable for consumption (deMan, 1990). The process o f  
autoxidation and the resulting deterioration in flavour o f  fats and fatty foods are 
often .described by the term rancidity (deMan, 1990). Autoxidation reaction can 
be divided into three parts namely, initiation, propagation and termination. It is a 
free radical chain reaction.
Initiation Step: it involves tire formation o f  fat-free radical (FFR) when loosely 
held hydrogen atoms are lost from the fatty acid groups, 
traces o f  Cu, o r o lhcr m etals
(i) Autoxidation
Propagation Step: FFR combines with oxygen to form peroxy free radical 
(PFR). The PFR removes a hydrogen atom from unsaturated lipid to form another 
FFR and unstable hydroperoxide.
1. RFI IT
heal, light
Free Fatty Acid (FFA) FFR
2. R  + 0 2
(FFR)
p*- R ()i 
(PER)
3. RO z '+RH >  R- + RDOII 
FFR hydroperoxide
24
The FFR reacts with oxygen to form more hydroperoxides as propagation 
proceeds. The unstable hydroperoxides breakdown into a variety o f  compounds, 
which include hydrocarbons (such as ethane and ethcne from linolenic and other n 
- 3 polyene acids and pentane form linoleic and other n 6 polyene acids), 
aldelydes, ketones, esters, lactones, alcohols and ethers all o f  which may be 
saturated or unsaturated (deMan, 1990). These compounds constitute the 
secondary oxidation products and are responsible for flavour deterioration.
Termination Step
This occurs when all the oxygen in the system is used up, or when two FFR 
reacts, or when an antioxidant radical reacts with the FFR.
4. R' +  R'  R2
Role o f antioxidant in termination
The antioxidants can function in two ways:
(a) It can donate a hydrogen atom to the FFR to form the fatty acid molecule and 
(hereby terminating the oxidation in its first step (initiation step).
R‘ + AH  ► RI1+ A'
FFR antioxidant FFA
Autoxidation is thus delayed until all the AH are used up.
(b) It can donate a hydrogen atom to PFR to from a hydroperoxide.
RO-2 + AH   ROOM -I- A-
(Instead of RH)
25
A’ is more stable than R because of  the electron resonance structure in the 
aromatic ring o f  A . The chain reaction in the step (3) is thus terminated. 
Antioxidants cannot reverse the oxidation o f  lipids therefore the)' must be added as 
early as possible in the manufacturing process or to the finished product (deMan,
1990). The aldehydes formed are further oxidized to form fatty acids that are 
considered as tertiary oxidation products.
Table 2.7 Hydroperoxides and aldehydes, which may be formed in 
autoxidation of some unsaturated fatty acids.
Fatty acid Methylene
group
involved
Isomeric hydroperoxides formed 
from the structures contributing to 
intermediate free radical resonance 
hybrid
Aldehydes formed 
by decompsition of  
the hydroperoxides
Oleic 11
8
11- hydroperoxy-9-ene
9-hydroperoxy-10-ene 
8-hydroperoxy-9-ene
10-hydroperoxy-8-ene
Octanal
2-deccnal
2-undecenal
nonanal
Linoleic 11 13-hydroperoxy-9,l 1-diene 
1 l-hydroperoxy-9,12-diene 
9-hydroperoxy-10,12-diene
I icxanal
2-octanal
2,4-decadienal
Linolenic 14
11
16-hydroperoxy-9,12,14-trienc 
14 -hvdroperoxy-9.12,15-triene
12-hydroperoxy-9,13,15-tricne
13-hydroperoxy-9,11,15-trienc 
11 -hydroperoxy-9,12,15-triene 
9-hydroperoxy-10,12,15-tricne
Propanal
2-pentenal
2.4-heptadienal
3-hexenaI
2.5-octadienal 
2.4,7-decatrienal
Source: Keeney (1962).
26
Photo-oxidation involves reaction o f  an alkene with oxygen in the presence of 
light and a suitable sensitizer. It involves the participation o f  a singlet oxygen 
(Hamilton and Berger, 1995). The extent o f  the changes in this pathway covers 
the transfer o f  energy from light to a photosensitizer, which then helps to form 
singlet oxygen. Photosensitizers such as erythrosine or methylene blue (the 
dyestuffs), flavin, porphyrin (the pigments) and anthracene, rubrene, the 
polycyclic (aromatic hydrocarbons) absorb the visible or near ultraviolet light to 
be converted to the excited state (Hamilton and Berger, 1995). The singlet, 
excited oxygen is converted to excited triplet state by inter-system crossing. 
Triplet activated sensitizer can then react with triplet oxygen to form singlet 
oxygen, which reacts with substrate to give hydroperoxides, which are 
.characteristic o f  this mechanism (Hamilton and Berger, 1995).
Photo-oxidation is much quicker than autoxidation and the difference in reactivity 
between oleate, linoleate, and linoleate (1:1.3:2.3) is close to the number o f  
double bonds in these esters (deMan, 1990). The hydroperoxides produced in this 
way differ from those resulting from autoxidation. It has been suggested that 
autoxidation o f  natural lipids may be initiated by photo-oxidation due to pigments 
remaining in the lipids after processing (Gunstone and Norris, 1983).
(ii) Photo-Oxidation
27
Some enzymes such as lipases, lipoxygenase and lipoperoxidase catalyze the 
oxidation o f  lipids. The most important agent for enzymatic peroxidation is the 
enzyme lipoxygenase (Salouchi and Matsushita, 1976). Chief sources o f  
lipoxygenase (lipoxidases) arc soybean, cereal grains and oil seeds, peas and 
beans. Soybean lipoxygenase are localized in mitochondria, plastids, chloroplasts 
and the cytosol o f  cells (Markley, 1951). When the raw tissue is broken the 
enzyme and substrate (oil) are liberated, and, provided some moisture is present a 
bitter, beany taste develops very rapidly (Baker and Mustakas (1972).
Lipoxygenase generally accelerates the addition o f  oxygen to the double bond o f  
carotcnoids and unsaturated fats to form peroxides, the pH for optimum activity 
being about 6 (Markley, 1951). The natural substrate o f  the enzyme is linoleic 
acid, but other acids such as linolenic and arachidonic acids are also oxidized 
(Gunstone and Norris, 1983). Lipoxygenase (Linoleate: Oxidoreductase) is highly 
specific and will attack the cis-cis-1, 4 -pentadiene group contained in the fatty 
acids linoleic, linolenic and arachidonic (deMan, 1990). The exact mechanism o f  
the reaction is still in doubt, but initially a hydrogen atom is abstracted from the 
o>-8 methylene group to produce a free radical. The free radical isomerizes, 
causing conjugation o f  the double bond and isomerization o f  the trans- 
configuration. The free radical then reacts to form the to-6 hydroperoxide. The 
peroxide formation by lipoxygenase is interrupted by the common lipid
(iii) Enzymatic Oxidation
28
antioxidants. The antioxidants are thought to react with the free radicals and thus 
interrupt the oxidation (Gunstonc and Norris, 1983).
Lipase is a lipolytic enzyme present in soybeans (Whitaker, 1972). Lipases are 
capable o f  hydrolyzing lipids to fatty acids and glycerol. The production o f  long- 
chain fatty acids by the action o f  lipases leads to the development o f  soapy 
flavour in foods. The fatty acids so produced are more susceptible to oxidation 
(Satouchi and Matsushita, 1976). Although lipase inhibitor has been reported in 
soybeans (Satouchi and Matsushita, 1976) there is no inhibitor available and 
acceptable for food use.
(iv) Thermal Oxidation
This term refers to the oxidation o f  lipids at high temperature. During food 
preparation involving the application o f  heat oxidative reactions are greatly 
accelerated (Frankel, 1984). Prolonged exposure o f  lipids at elevated temperature 
in the presence o f  air and moisture leads to the formation o f  various oxidation 
products, including polymeric compounds (Chow, and Gupta, 1994). The rate o f  
hydroperoxide formation and decomposition is markedly increased during thermal 
oxidation. A major pathway o f  thermal oxidation appears lo involve a homolytic 
cleavage o f  the 0 - 0  bond of  the hydroperoxides and formation o f  alkoxyl and 
hydroxyl radicals (Frankel, 1984). Alkoxyl radicals formed may undergo C-C 
bond scission lo produce an alkyl radical and a vinyl radical (Frankel. 1984). On 
the other hand, the alkyl radical can react with a hydrogen radical, hydroxyl
29
radical or molecular oxygen to generate hydrocarbon, alcohols, and 
hydroperoxides, respectively. Also the vinyl radical may react with hydroxyl 
radical, hydrogen radical or molecular oxygen to form aldehydes and olefins 
(Nawar, 1984). Thus, hydroperoxides can be broken down into non-volatile oxy - 
and cyclic acids and volatile products such as saturated and unsaturated 
aldehydes, ketones, hydrocarbon, alcohols, acids and esters (Frankel, 1984).
During thermal oxidation, intra - and intermolecular reactions o f  alkoxyl, alkyl and 
peroxyl radicals may lead to the formation o f  dimers, trimers and large
molecular weight polymers with C-O-C and C-O-O-C cross links (Nawar,
1984). Dimers are a major component o f  non-volatile products formed in
oxidized and heated lipids (Nawar, 1984). By combining two radicals to
form a non-radical dimer, the free radical chain reaction is thus terminated.
At high temperature, hydroperoxides begin to decompose spontaneously, and the 
radical concentration becomes relatively high allowing radical - radical interaction 
to proceed faster (Nawar, 1984).
Heating conditions (time, temperature and aeration) and antioxidant levels in 
lipid-containing foods can modulate the degree o f  thermal oxidation and type o f  
products generated (Hsieh and Kinsella, 1989).
30
2.4.1.3 Products o f lipid oxidation
The products o f  lipid oxidation arc divided into primary, secondary and tertiary 
oxidation products.
Primary oxidation products: These are (lie hydroperoxides formed. The. 
hydroperoxides are odourless hcnce do not conlribuic to flavour themselves 
(deMan, 1990). They are also unstable hence decompose into other compounds. 
The transition metals aid the decomposition reaction of  peroxides by an oxidative 
and reduction mechanism in which the metal goes from higher valence state to 
lower valence state in the former and from lower valence state to a higher valence 
state in the latter (Hamilton and Berger 1995).
ROOM + M(n+1)+ ----------► R 0 2 + H+ + M n+
ROOM + M n+  ► RO' + OIT + M(n+I)l'
If a reducing agent such as ascorbic acid is present with Fe2+, the Fe31 produced in 
the decomposition o f  the hydroperoxide will be reduced and the smallest quantity 
o f  metal ion catalyze very large quantities o f  lipid breakdown.
The decomposition of  the hydroperoxide is also increased markedly during 
thermal oxidation (Chow and Gupta, 1994).
Secondary oxidation products: These are formed from the decomposition 
reaction o f  the hydroperoxides. Different levels o f  secondary oxidation products
31
are formed depending on the condition o f  heating (Chow and Gupta, 1994). 
Kanazawa el al. (1985), have shown that the secondary product fraction o f  
peroxidized methyl linoleatc consisted o f  about 35% polymers, 25%> 
endoperoxide-rich components and 40% low molecular weight compounds. 
Approximately 16% o f  the total low molecular weight fraction obtained from 
peroxidized methyl linoleate is identified as 8 hydroxyl methyl octanoate, 41% 
as 4 - formyl-9-decenoate (Oarada et al., 1986).
Secondary oxidation products include a variety o f  compounds including 
hydrocarbons (such as ethane and ethene), aldehydes, ketones, esters, lactones, 
alcohols and ethers, all o f  which may be saturated or unsaturated (deMan, 1990). 
Oganoleptic changes in fats and fatly foods are generally related to the formation 
o f  secondary oxidation products, especially aldehydic compounds. The aldehydes 
are strong flavour compounds and have very low flavour thresholds.
Tertiary oxidation products: These are formed from the oxidation o f  the 
secondary products, especially the carbonyls.
2.4.1.3.1 Reaction of oxidised lipids with proteins
The most important lipid oxidation products, which react with proteins are 
hydroperoxides and aldehydes produced by their rearrangements cleavage. 
Malondialdehyde (MDA) produced by cleavage of  dihydroperoxide or 
hydroperoxy alkenals arising from polyunsaturated lipids react with thiol, amine.
32
and phenolic groups o f  protein. The reaction o f  a hydroperoxide molecule with 2- 
amino acids, either free or bound in protein, proceeds by a mechanism analogous 
lo the strccker degradation;
ROOM + R ‘CH(NH2) C O O I I  ► ROM + R rCI 10 + Ni l., + C 0 2
'The products of  Strccker degradation are very reactive so that molecular brown 
pigments are the most often detected end products (Pokoi ny el al., 1983).
Other very reactive functional groups of  compounds present in oxidised lipids are 
various aldehydes, usually a mixture o f  alkanals, 2 alkenals, and 2, 4 
alkadienals. When these substances interact with protein or free amino acids a 
Schiff base is formed by the addition o f  amines to aldehydes and by a subsequent 
dehydration reaction;
RCITO + H2N-R 1  ► RCHOH   * RCH = N -R '+H 20  
l lH -R 1
The Schiff base readily reacts with another molecule of  aldehydes;
RCI I2 CIT = N -R1 + R "  Cl 1 0  ► RCCII = N-R' + H20
Cl I R "
The reaction can proceed with the formation of  higher oligomers (Pokorny el al, 
1983). The reaction is very rapid at room temperature or cold storage 
temperatures.
.13
The autoxidation o f  aldehydes is an extremely rapid reaction, more rapid than that
o f  polyunsaturated fatty acid esters. The oxidation results in the formation of
peroxy acids, which are often very stable.
RCIIO + 0 2   ^  RC=0
OOH 
Peroxy acid
They are decomposed by reaction with another molecule o f  aldehyde;
RC =  O + RCI-IO ------------► 2 RC=0
Aoi-I OH
The resulting fatty acid belongs to the main reaction products in the system 
oxidised at room temperature (Pokorny et al, 1983).
The aldolization and oxidation reactions o f  aldehydes give rise to various other 
volatile products (Jirousova el al., 1975). The hydroperoxides reacts with sulphur 
containing amino acids such as cysteine, which is oxidised to lysine.
2R - SH + R 'OOH  ► R-S-S-R + R 'OH + 1I20
Cystine is further oxidized into thiosulphinate;
R-S-S-R +R 'OOH   ► R-S-S-R + R 'OII
I!
o
Methionine is attacked by lipid peroxides and oxidized into methionine
su lphox ide :
34
r s c h 3 + r 'o o i i >  RCHi + R 'OH
II
0
The oxidation may proceed further with the formation of  methionine sulphoile at 
least under extreme conditions.
Table 2.8 Reactive groups taking part in the interaction of oxidized lipids 
with proteins.
Amino group bound 
in protein
Functional group 
present
Reacting product o f  
lipid oxidation
Resulting functional 
group
Cysteine Thiol Hydroperoxide
Aldehyde
Disulphide
Thioacetal
Cystine Disulphide Hydroperoxide Thiosulphinal
Methionine Sulphide Hydroperoxide Sulphoxide
Lysine Amine Hydroperoxide
Aldehyde
Hydroketone
Schiff base 
Schiff base 
Schiff base
Tryptophan Indole Hydroperoxide Various products
Tyrosine Phenol Aldehyde Methylene bridge
Serine,Threonine I Iydroxyl Epoxide
Carboxyl
Hydroxy ether 
Ester
Aspartic,Glutamic 
acids
Carboxyl Epoxide
Hydroxyl
Hydroperoxide
I lydroxy ester 
Ester + products 
Ester
Source: Pokornv ct al. (1983)
35
2.4.1.4 Effcct o f lipid oxidation on (lie nutritive value of food
Essential fatty acids and tocophcrols present in lipids and lipid-containing foods 
are readily oxidizable during processing, storage and usage, especially at high 
temperatures (Hamilton and Berger, 1995). If extensively oxidized lipids are used 
as the sole source o f  dietary lipids and vitamin E, deficiency o f  essential fatty 
acids, vitamin E or both may result (Kamel and Kakuda, 1995). Using 15% 
autoxidized oil as the sole source o f  dietary lipid, Kamel and Kakuda (1995), 
observed a marked growth inhibition in growing rats. However when fresh 
cottonseed oil was added to the diet, normal growth resumed. These findings 
suggest a deficiency o f  essential fatly acids, vitamin E or both in oxidized lipids 
used in the diet.
The interaction products o f  oxidising lipids with proteins have lower nutritional 
value for the following reasons:
(a) Destruction o f  essential factors such as essential amino acid in bound 
proteins particularly lysine, tryptophan and methionine, or essential fatty 
acids in the lipid fraction, oxylabile vitamins and various non-essential 
componcnl;; which arc iinpuitunt far miliilinii.
(b) Decrease o f  digestibility due lo incomplete and slower enzymic hydrolysis o f  
bound lipids and protein.
(c) In vivo formation o f  precursors o f  various diseases (Pokorny el al., 1983).
Reactions of  oxidized lipids with proteins proceed in vivo very slowly producing 
brown insoluble deposits called ceroid (age pigment). The ceroid formation is
36
enhanced if diet contains excess polyunsaturated lipids while deficient in 
locopherols (Horowitz and I lartroft, 1971), and may be inhibited by 
supplementation with locopherols.
It has been shown that methyl linoleate has a depressing effect on carotene 
utilization, which can be overcome when an excess o f  pro-vitamin A is given. In 
spite o f  the insolubility o f  the water- soluble vitamins (B and C) in fats, some o f  
them may be destroyed during lipid oxidation. Biotin and vitamin C have been 
found to be destroyed during lipid oxidation (Pokornv et al., 1983).
2.4.1.5 Effcct o f lipid oxidation on sensory value of food
Since the human palate is very discriminating, it can detect odoriferous molecules 
at very low levels. It is believed that when as few as 1 in 1000 carbon-carbon 
double bonds in a fatty food reacts with oxygen, it is already too late (Hamilton 
and Berger, 1995). The carbonyls, especially aldehydes produced from the 
breakdown o f  hydroperoxides are responsible for flavour changes in soy products 
and other foods. The process of  oxidation and the resulting deterioration in 
flavour o f  fats and fatty foods are often described by the term rancidity. Labuza 
(1975) defined rancidity as (he development o f  off-flavour, which makes a food 
unacceptable for the consumer. . Usually rancidity refers to oxidative 
deterioration, but in the field of  dairy science, rancidity refers to hydrolytic 
changes resulting from enzyme activity (deMan, I9<)()). Rancidification o f  
soybean lipids occurs in two forms. One form involves the rapid development of
37
off-11 avows commonly referred to as reversion and the other form involves the 
slow development of  strong odours and flavours characlcristic o f  lipid oxidation 
(deMan, 1990). Flavour reversion is a particular type of  oxidized flavour that 
develops at comparatively low levels o f  oxidation. The off-flavours may develop 
in oils that have a peroxide value o f  as little as 1 or 2. Other oils may not become 
rancid until the peroxide value reaches 100. LinoJenic acid is generally 
recognized as the determining factor o f  flavour reversion (deMan, 1990).
The interactions o f  oxidized lipids with proteins result in the following 
organoleptic changes:
(a) Changes o f  the colour due to browning reactions.
(b) Changes of  the flavour because o f  binding of off-flavour compounds into 
neutral or less active substances or because o f  the formation o f  new flavour- 
active compounds.
(c) Changes o f  the texture caused by denaluration o f  proteins and by crosslinking 
o f  polypeptide chains (Pokorny et al., 1983).
The activity o f  lipases and lipoxygenases are responsible for soapy and rancid 
flavours in foods. The former is due to the presence o f  long chain fatty acids and 
the latter is due to the oxidation of  unsaturated fatty acids (Satouchi and 
Matsushita, 1976). Lipases hydrolyze fats into fatty acids and glycerol.
38
Tabic 2.9 Flavour threshold values for oxidation products in 
soybean
Substance Flavour description Threshold (ppm)
Oct-1 -ene-3-onc Metallic 0.001
2-Pentyl furan Liquorice 2.000
Oet-l-ene-3-ol Mushroom 0.007
Pent-l-cne-3-one Metallic 0.001
Source: Hamilton and Berger, (1995)
Table 2.10 Flavour descriptions used for crude, processed and reverted 
soybean oil.
Slate Flavour
Crude Grassy, beany
Freshly processed Sweet, pleasant, nutty
Reverted Grassy, beany, buttery, melony 
Tallowy, painty, fishy
Source: Hoffman (1962).
2.4.1.6 Effect of lipid oxidation on health
The biological and toxicological properties of oxidized lipids have been 
extensively investigated (Alexander, 1986). There is general agreement that 
undesirable or harmful materials are formed during storage and usage o f  lipids. 
But, there is a considerable disagreement regarding the levels o f  such materials 
formed and the nature of  harmful effects caused by oxidized lipids (Alexander. 
1986). This can be partly attributed to the very large variety o f  oxidation products
that may be formed according to the degree o f lipid oxidation and the adequacy o f 
experimental diets employed by various investigations (Chow and Gupta, 1994).
Oxidation products o f  polyunsaturated fatty acids may also be carcinogenic or co- 
carcinogenic (Kamel and Kakuda. 1995). The ability o f certain anti-oxidants to 
protect against experimentally induced carcinogenesis (Wallenberg, 1972) had led 
to fhe suggestion that oxidation products o f  lipids may play a role in 
carcinogenesis and mutagenesis by damaging genetic materials.
Oxidation o f  lipids results in deficiencies o f  essential fatty acids and vitamin E. 
Animals fed on oxidized lipids containing low proteins had elevated serum 
glutamate -oxaloacetate transaminase and glutamate - pyruvate transaminase 
values, suggesting injury o f the heart or liver (Huang et a!.., 1988). Insufficiency 
o f  dietary protein has also been . shown to aggravate the enhanced lipid 
peroxidation and decreased activities o f antioxidant enzymes in rats fed on high 
polyunsaturated fats (Huang et al., 1988).
Pure fatty acid hydroperoxides are very toxic to experimental animals when 
administered intravenously (i.v.) but not orally (Findlay et al., 1970). The same 
authors have shown that the 24-h lethal dose o f a high purity preparation o f 
methyl linoleatc hydroperoxides in adult male rats was about 0.07mmol/100g 
body weight. The major effect o f the linoleatc hydroperoxide was on the lungs.
40
Secondary oxidation products such as polymeric materials o f high molecular 
weight arc not easily absorbed (Kanazawa el al., 1985) and are generally less 
toxic than monomeric or dimeric compounds. Consumption o f  large amounts of 
polymeric fatty acids may result in diarrhoea. Cyclic monomers, when fed at high 
levels to rats, have been shown to cause fatty livers (Poling el al., 1970). Levels 
o f  0.20 0.15% o f cyclic monomers have produced fatty livers (Poling et al.,
1970).
Many secondary oxidation products o f fatty acids have been shown to be more 
toxic to experimental animals. This is partly due to the fact that low 
molecular weight products have shorter carbon chain lengths and are more 
easily absorbed into the intestinal wall than lipid hydroperoxides or their 
polymeric materials (Chow and Gupta, 1994). Malondialdehyde (MDA) is a 
three-carbon dialdchyde and is toxic to experimental animals.
The LD 50 levels o f  MDA in rats are 632 mg/kg for its enolic sodium salt and 527 
mg/kg for is acetal form (Frankel, 1984), and both are more toxic than 
formaldehyde or glyoxal. MDA can react with mitochondrial membranes and 
disrupt red blood cell membranes (Frankel, 1984). MDA may be carcinogenic or 
co-carcinogenic under certain conditions. Mutagenicity and cytotoxicity o f  MDA 
have been demonstrated in mammalian lymphoma cells (Begin, 1987). Also, 
MDA may play a role in regulating tumour metastasis, host immune mechanism 
and the proliferation and differentiation o f  tumour cells (Frankel, 1984).
41
However the significance o f MDA to human risk o f cancer remains to be 
established (Begin. 1987).
Another group o f  toxic compounds formed in secondary oxidation is 4- 
hydroxyalkenal. An example is 4-hydroxynonenal, which has a lethal dose o f  68 
mg/kg in  mice (Frankel, 1984). When 13-18 mg/kg o f  4-hydroxynonenal or 4- 
hydroxyhexanal is injected intravenously as a phospholipid emulsion, severe liver 
damage is produced in rats (Segall et al., 1985) similar lo that seen after CCU 
administration.
Components o f oxidised lipids may also accelerate the turnover o f  vitamin E and 
increase the susceptibility o f the red cells to hemolytic stress (Chow and Gupta,
1994). There may also be a reaction o f peroxidizing lipids intracellularly with cell 
proteins leading to the formation o f ceroids. Ceroid is deposited mainly in the 
brain and neuronal tissue but also in other organs such as the liver and uterus 
(Begin, 1987). Ceroid may also be deposited into aorta walls.
Generally oxidised lipids cause appetite and growth depression, diarrhoea, tissue 
enlargement, interference with reproduction, and even death in some cases 
(Poling el al., 1970).
42
2.4.1.7 Laid down guidelines for avoiding rancidity
i. Maximal retention o f natural antioxidants.
ii. Use as low a temperature as possible for processing and storage.
iii. Where high temperatures are unavoidable, reduce the time o f exposure to a
minimum.
iv. Reduce access o f  air (i.e., O2).
v. M inimise the interaction o f  the food with catalytic metals.
vi. Maintain good practice in terms o f  stock rotation and cleanliness.
vii. If  these good practices are found to be insufficient use permitted chelating 
agents and antioxidants in the minimum amount needed (Hamilton and 
Berger, 1995).
2.4.1.8 Indices of lipid oxidation and rancidity
i. Lipid stability
Stability o f  lipids refers to resistance to the development o f  any o ff flavours, 
whether resulting from hydrolysis, reversion or oxidation (Hamilton and Berger, 
1995). Methods such as the Swift stability test or the active oxygen method 
(AOM) can be used to measure stability o f  lipid. AOM predicts (he susceptibility 
o f  lipid to oxidation and rancidity. In this test, air is continuously passed through 
the sample at a specified temperature and the length o f time required for peroxide 
value to rise to a level indicative of'rancidity  is noted (Hamilton and Berger,
1995). For soybean oil (he peroxide value at which the oil becomes rancid is 
between 20-40 mcq/kg (Zadernowski et al., 1983). The time is expressed in
43
hours, AOM. In general the higher the AOM the longer the shelf life o f  the 
sample.
ii. Peroxide value
The peroxide value (P.V) is a measure o f  active oxygen in lOOOg o f fat or oil and 
is expressed as millimoles or milliequivalents o f  peroxide (Ronald and Ronald, 
1991). Since peroxides are the intermediate products formed in the autoxidation 
o f  oils, the test is used as measure o f  stability or for following course o f  the 
development o f  rancidity o f oils. The test is also used for evaluating the 
effectiveness o f  antioxidant compounds on the keeping quality o f fats and fat 
products. Peroxide value is an indication o f  the extent o f rancidity; the lower the 
peroxide value, the less the oxidation that has occurred. Peroxide value is 
determined by measuring iodine liberated from potassium iodide by a given 
quantity o f  fat under prescribed conditions.
iii. Tliiobarbituric acid (TBA) test
Products o f  lipid oxidation are apparently responsible for the colour reaction with 
TBA reagent (example, malonic dialdehyde and methyl oleate hydroperoxide both 
give colour reaction). The test is more sensitive and responsive at earlier stages 
o f  autoxidation. There .is increase in the red pigment formation as oxidative 
rancidity advances. Zadernowski et al. (1983) reported that soybean oil becomes 
rancid at TBA Number between 15 and 20 meq/kg.
44
2.4.I.S.1 Other indirect indices
a. Free fatly acid and acid value
I he concentration o f  free fatty acids in soybean oii is influenced by factors 
including agronomic and hydrolysis during processing and storage (Markley, 
1951). Some free fatty acids could also result from the oxidation o f  secondary 
oxidation products such as the aldehydes (deMan, 1990). The free fatty acid is 
expressed in terms o f  percentage oleic acid (Ronald and Ronald, 1991). High 
fatty acid content indicates high instability o f  the soybean lipids. The term acid 
value is generally used to express the quantity o f  free fatty acid present in a food. 
It is defined as the number o f  milligrams o f potassium hydroxide required to 
neutralise the free fatty acids in gram o f a lipid sample (Ronald and Ronald,
1991). The acid value is determined by dissolving a known weight o f  oil or fat in 
hot neutral alcohol and titrating with 0.25ml alkali to a phenolphthalein end point 
(Ronald and Ronald, 1991). The acid value o f  solvent extracted soybean oil is 
between 0.5 and 1.92 (Zadernowski et al., 1983).
b. Iodine value (I.V.)
The iodine value is defined as the number o f  grams o f iodine absorbed by lOOg o f 
an oil or fat (Ronald and Ronald, 1991). It is an indication o f  the amount o f 
unsaturation o f  a fat or oil but provides no information concerning (he specific 
types or arrangement o f  the unsaturated bonds in the fatly acid components o f  the 
glyceride molecules. The iodine value can be used as a measure o f  the stability o f 
a lipid. The higher the iodine value (he more susceptible the lipid is lo oxidation
'15
hence the lower the stability. The iodine value o f solvent-extracted soybean oil is 
between 134.1 and 135.8 (Zadernowski el al.. 1983).
2.5.0 Soybean processing anti products
Soybeans were processed first in 1911, in the United States for oil and meals 
(Markley, 1951). Since 1941, soybean flour (including grils) has been the principal 
soybean product (Maikley, 1951). Technology for processing soybean is well developed.
2.5.1 Soybean flour
There are three main types o f  soybean flour. These are full-fat flour, low-fat flour 
and defatted flour (Ferrier, 1975), which are revised below.
2.5.1.1 Full-fat flour
Full-fat flour contains all the iat originally present in the soybean seed. It is 
prepared by blanching or steaming the beans to debitter them and to inactivate 
lipid-oxidising enzymes such as lipases and lipoxygenase. The inactivation o f 
these enzymes completely prevents the formation o f any bitter, beany, or painty 
flavour (Ferrier. 1975). Blanching simultaneously destroys trypsin inhibitors, 
hemaglutinins and other known toxic factors present in the law  beans. The length 
o f  time required for these components to be destroyed decrease with increased 
moisture content o f  the whole bean.
•ifi
Lipoxygenase is inactivated in rehydrated soybeans by boiling for less than five 
m inutes (Ferrier, 1975). Boiling is also essential to produce an acceptable texture. 
Soaking and boiling also remove about one-third o f  (he oligosaccharides in 
soybeans, some o f which are responsible for the production o f  intestinal gas or 
Halits. Only a small amount o f  protein (1%) is lost during soaking and blanching 
(Ferrier, 1975).
After the soybeans have been blanched or steamed, the hulls, mainly cellulose and 
polysaccharides are removed and the dehulled beans are dried and then ground 
into full fat flour. The flour contains about 40% protein, 20% fat, 30% 
carbohydrate, 5% ash and 5.5% crude fibre (Waggle et a l 1981).
The flour may be used as an ingredient in different bakery products to improve 
crumb body and resilience, as well as colour and toasting characteristics as a 
result o f  its sugar content (Waggle et al., 1981). They are also used in cereal and 
infant foods.
2.5.1.2 Low-fat flour
This type o f  soybean flour has greater part o f  its fat removed by a continuous 
mechanical pressing melhod. It is also used as an ingredient in bakery products 
and has many other uses as well (Ferrier, 1975).
47
2.5.1.3 Defatted flour
This type o f flour ordinarily contains less than 1% fat. The protein content ranges 
from 53 55% (Waggle cl t/L. 1981). It is prepared by cracking and dehulling the 
beans and tempering the cracked meats to about 11% moisture. The tempered 
meats are then passed through smooth rolls to form thin Hakes that are extracted 
with hexane to remove the oil. The defatted Hakes are then milled into flour and 
grits. These products are used in the preparation o f beverages, crackers, cereal 
foods and infant foods. They are also used as meat extenders (Wolf, 1990).
Table 2.11 Proximate Analyses of Commercial Soybean Flour and Grits (%)
Flour or 
Grit Type
Moisture Protein Fat Carbohydrale Ash Fibre
Full-Fat 5 41.5 21 25.2 5.2 2.1
Low-Fat 5.5 46.0 6.5 33.5 5.5 3.0
Defatted 5 53.0 0.9 32.3 6.0 2.9
Source: Meyer (1970)
Fat extraction constitutes one o f the essential stages o f the technological process 
o f  soyflour production. After the extraction flours contain small amount o f  free 
lipids and lipids bound with proteins and other hydrophilic components. 
Percentage bound fat in defatted soyflour is 2.2 ± 0.5. The significant quantity o f 
bound lipids in defatted flour suggests that stability problems would be common
'18
during storage. Polar lipid such as the predominant phospholipids, phosphatidyl 
choline would create an immediate flavour problem (Zadernowski el a l., 1983).
Tabic 2.12 Proportions of phospholipid fractions in soyflour as a 
percentage of total lipids
Phospholipid Proportion (%)
Unknown 15.9
Phosphatidylethanolamine 23.1
Phosphatidyl glycerol 0.9
Phosphalidic acid 4.2
Phosphatidyl choline 32.7
Phosphatidyl serine 0.7
Phosphatidylinositol 20.4
Lyso Phosphatidylhanolamide 0.2
Source: Zadernowski el al. (1983).
The ratio between neutral and polar lipids in soyflour is 1:0.9 (Zadernowski el al., 
1983). The stability o f soyflour varies directly with the fat content. The 
development o f a rancid off-odour always precedes the appearance o f  a rancid 
taste in the dried flour. The solvent-extracted flour having a fat content o f  less 
than 1.5% never attains appreciable concentrations of peroxides.
Table 2.13 Percentage composition of fatty aciils in isolated fractions of 
bound lipids present in defatted soyflour
Fatty acids Neutral lipid Polar lipid Glycolipid
16 : 0 29.3 30.9 28.5
18 : 0 6.9 7.6 5.7
18 : 1 8.5 10.8 8.3
1 8 : 2 48.0 46.3 50.8
18 : 3 7.3 4.3 5.6
Source: Zadernowski et al., (1983).
2.5.2 Protein concentrates and isolates
Protein concentrates are defatted soy flake or soyflour, which has been upgraded 
in protein content by further fractionation to remove about one-half o f  the 
carbohydrates and some o f the minor constituents (Markley, 1951). Soy protein 
concentrate has protein content between 60 and 70%, 1% fat, 21% carbohydrates, 
5% ash and 3.5% crude fibre (Waggle el al., 1981). U is mainly used as an 
ingredient in bakery industry and as meat additive.
Protein isolates are obtained from soyflour by an isolation process based on the 
differences in solubility o f proteins as pH is varied. They are prepared by 
extracting alkaline solution at pH 7 - 8  and centrifuging out the insoluble 
polysaccharide residue. The clarified extract is then acidified to pH 4.5 thereby 
precipitating the major proteins as a white curd. The protein curd is separated
50
front the soluble fractions (vvhey) by centrifuging. The curd is washed and 
slurried in water and spray-dried in the isoelectric condition to give isoelcctric 
protein isolate. The isolates have protein content between 90 and 97%,, 0.5% fat, 
2.5% carbohydrates, 4.5% ash and 0.3% crude fibre (Waggle cl a l., 1981). These 
products are used mainly in simulated meat products, and in the bakery industry.
2.5.3 Other soy products
These include soymilk that can be used to make soy yoghurt and ice cream. Some 
traditional soy products include miso, lempeh, nalo , soy sauce and sufo  all o f 
which are consumed extensively in the South Eastern Asian countries (Ihekoronye 
and Ngoddy, 1985).
2.5.4 Effcct o f processing on the nutritional value of soybean
Nutritive value o f  soybean may be improved by heat treatment or by processing 
into soybean products. The degree o f nutritional improvement by heating 
depends upon, among other things, the temperature, duration o f  heating, and 
whether moist or dry heat is used (Hamilton and Berger, 1995).
Trypsin inhibitors and hemagglutinins are thermolabile hence the beneficial effect 
o f  heat treatment on the nutritive value is related to the inactivation o f  both 
components (Zadernowski cl al, 1983). Lipoxygenase, which is responsible for the bitter, 
beany flavour as well as catalyzing lipid oxidation is very sensitive to heat and is 
destroyed at 82°C in 15 min (Baker and Mustakas, 1972).
51
Excessive heating may destroy certain amino acids, such as lysine and particularly 
cystine, which are sensitive to heat and the loss by excessive heating may be more 
than 50% (Smith and Circle, 1972). Other amino acids such as arginine, 
tryptophan, histidine and serine may also be partially destroyed (Smith and Circle, 
1972).
2.5.5 Effect o f processing on flavour of soybean
Raw matured soybeans have a bitter, astringent taste and when eaten in this form 
will generate gas (flatulence) and cause diarrhoea (Ferrier, 1975). Raw meal is 
usually described as tasting beany, bitter and green but the green taste disappears 
after steaming for 3 minutes at atmospheric pressure (Thio, 1975). The beany, 
bitter flavour as well as the nutty, sweet, and toasted flavours vary in intensity 
with continued streaming (Thio, 1975).
2.5.6 Utilization of soybean flour in other dry foods
Soybean can be used in the fortification o f  cereal products. Cereals and legumes 
generally are individually not nutritionally adequate. For optimum utilisation, all 
the essential amino acids must be present in the right proportions. Soybeans are 
good sources o f  lysine but are deficient in sulphur-containing amino acids (Wolf. 
1990). On the other hand, cereals such as maize, is generally deficient is lysine 
but have adequate amounts o f  sulphur-containing amino acids. These 
characteristics make soybean and other legumes natural complements to cereal- 
based diets. Studies have shown that such mixes in appropriate proportions will 
increase the quality o f  the diet above that o f any single ingredient (Bressani and
52
l l ia z , 1983). In general a maximum supplementary effect is observed when about 
50% o f  the legumes protein is replaced by cereal prolein (Bressani and Eliaz, 
1983).
Tabic 2.14Effect of steaming on flavour of soyflour
Steaming
(Minutes)
Flavour
(Score00)
Flavour Description
0 1.5 Beany, Bitter, Green
3 4.5 Beany, Bitter, Nutty, Toasted Sweet
10 6.0 Beany, Nutty, Bitter, Toasted Sweet
20 6.3 Beany, Nutty, Bitter, Toasted Sweet
40 6.1 Beany, Nutty, Bitter, Toasted Sweet
(a) 1 = strong 10 = Bland
Source: Rackis et al. (1979).
Cereal and soybean can play an important role in the diet o f the infant. In 
developing countries, most infants show satisfactory growth for the first six 
months o f life when breast milk solely meets the nutritional needs. However with 
the onset o f  weaning, malnutrition usually sets in when protein and other nutrient 
requirements o f infants are much higher than provided b) the weaning foods used 
(Orraca-Tetteh, 1972). In Ghana, porridges and gruel made from maize, millet 
and sorghum are popular foods used during the weaning period. The net prolein 
utilisation (N.P.IJ) is 44.5% for maize porridge and 50% for sorghum gruel 
(Orraca-Tetteh, 1972), compared with breast milk with N.P.IJ o f  about 100%.
53
Table 2.15 Effcct of heat treatment on protein efficiency ratio (P.E.R.), 
biological value (13. V.), and digestibility of soybeans and soybean products.
Products Treatment (a) P.E.R (b) B.V. Digestibility
Raw, immature None
Autoclaved
1,1
2.0
49 88
Raw vine- None 0.5 69 85
ripened autoclaved 1.5
Raw, mature None
Dry heated
0.7
0.7
58 82
Steamed 1.3 92
autoclaved 1.3 64 90
Raw, None 1.4
germinated autoclaved 1.9
Meal, solvent None 52 74
extracted autoclaved 69 89
Soymilk none 2.0 79 91
Tofu none 1.8 68 96
Tempeh Steamed(2hrs) 
Deep fat 
fried(7min)
2.2
0.6
-..... .. . ..............
a. P.E.R: Gain in body weight divided by weight o f  protein consumed 
(FAO/WHO, 1965).
b. B.V: The proportion o f  absorbed nitrogen retained in the body for 
maintenance and/or growth (FAO/WHO, 1965).
Source: Thio, 1975.
The protein quality and content o f  these traditional weaning foods can be 
improved by the addition o f high-protein soybean products. The best ratio o f 
soybean: Maize is 28:72 (Bressani et al., 1974). This mixture gives a P.E.R o f 
2.54 whereas 100% maize results in a P.E.R o f 0.69 compared with a casein 
standard o f 2.87 (Bressani el a!., 1974). In Ghana, infant foods like Brownilac,
54
Cerelac and JVeaniniix have been formulated from mixture of cereals and
proteins.
Tabic 2.16 l’rotciu and fat contents and protein value of different maize - soy 
preparations
Mixture (%) Content (%) P.E.R
Soy bean Maize Protein Fat
0 100 9.9 4.5 0.69
21 79 16.9 8.9 2.08
28 72 17.9 10.9 2.54
38 62 18.1 11.3 2.37
100 0 40.0 25.6 2.03
casein 2.87
Source: Bressani and EIiaz(1983).
Soybean flour has been used to fortify cassava flour for the baking industry 
(Grace, 1977). Research done at the International Institute o f Tropical 
Agriculture (IITA) has led to the production o f  nutritious bread and other baked 
products such as cakes and biscuits using cassava and soybean flour and other 
relevant ingredients (Kordylas, 1991).
In Ghana, a non-governmental organisation (NGO), the Global farm ers' Wives 
Association has produced soy-fortified gciri and cassava flour on commercial 
scale (Industrial and Technology Fair, Ghana, 1999).
55
CHAPTER THREE  
MATERIALS AND METHODS 
3.1 MATERIALS
The main raw material used for the project was commercial soybeans grown in the 
northern savanna zone o f  Ghana, bought from the Madina  market, Accra.
3.2 METHODOLOGY
The commercial soybeans were processed into flour using three different 
processing methods. Each o f  the samples prepared was stored and analyzed 
chemically and organoleptically over a period o f  twelve (12) weeks.
3.2.1 Preparation of samples
1. Raw soyflour
Sorted whole beans were milled into flour using a hammer mill (Christy and 
Norris Laboratory [Chelmsford, England] Mill Size 8). The Raw soyllour was 
packaged in high-density polythene bags obtained from the Poly-Products, 
Ghana Ltd., Accra and stored in the cold room for analyses.
2. Cooked-dried soyflour
The whole beans were washed thoroughly after sorting. The beans were 
transferred into a saucepan containing water and boiled for one (1) hour on a hot 
plale. The boiled beans together with the hulls were pul in an air-oven set at
56
60°C to diy over night. The dried beans were milled into flour using the hammer 
mill (Christy and Norris Laboratory [Chelmsford, England] Mill Size 8). The 
cooked-dried soyflour was packaged in high-density polythene bags 
obtained from the Poly-Products, Ghana Ltd.-Accra and stored for analyses.
3. Roasted soyflour
Whole beans were first cleaned and roasted in a hot open pan for thirty (30) 
minutes. To get rid o f  burnt and broken pieces o f  hulls the roasted beans were 
sorted. The whole beans were then milled into flour using the hammer mill 
(Christy and Norris Laboratory [Chelmsford, England] Mill Size 8). The 
roasted soyflour was packaged in high-density polythene bags obtained from the 
Poly-Products, Ghana Ltd.-Accra and stored for analyses
3.2.2. Chemical analyses
3.2.2.1 Total Lipids
The total lipids were determined using the method according to Bligh and Dyer (1981) 
in A.O.A.C. (1990). In this method, lOg o f sample were accurately weighed into a 
200ml-homogenizing flask. Water, chloroform and methanol were added in the volumes 
10ml, 20ml and 40ml respectively. The mixture was homogenized for one ( I)  minute 
after which 20ml chloroform was added and the mixture homogenized again for 30 
seconds. This was followed by the addition o f  20ml water and homogenizing for 30 
seconds. The homogenate was transferred into glass centrifuge tubes and centrifuged at 
2000 r. p.m. for 20 minutes using the Den fey BS400 centril'ugc. The aqueous layer was
57
Flow chart for the 
Preparation of roasted 
Soyflour____________
Soybeans
Clean
nP
Roast
(Roasted soybean) 
sU
Winnow
4^
Mill
sU
(Roasted soybean flour) 
sU 
Package 
sU
(Packaged roasted soybean Hour)
Fig 3.1 Flow charts for the processing 
soybean flour.
Flow chart for the 
preparation ofcooked- 
dricd ; ovflour
Soybeans
Clean
nU
Cook
(Cooked soybeans)
4 '
Oven-dry
4/
(Cooked-dried soybeans)
>1/
Mill
4^
(Cooked-dried soybean flour)
4/
Package
(Packaged Cooked-dried soybean flour) 
o f  roasted soybean flour and cooked-dried
58
removed by suction, The lipids in 20ml o f  chloroform was then determ ined after 
evaporation in a dried, weighed flask, initially on a steam bath and finally in an 
air-oven set at 105°C for 30 minutes. To get the total lipids the weight o f  the 
lipid was then multiplied by 2.
3.2.2.2 Crude Fat
The crude fat content o f  the samples was determined using the Soxhlets 
extraction method (A.O.A.C. 1990)
3.2.2.3 Free Fatty Acids
Free fatty acids were determined using the A.O.A.C. (1980) method 28.029. In 
this method chloroform extract o f  the sample was used for the analysis. The 
extract was prepared using the method o f Bligh and Dyer (1981) in A.O.A.C. 
(1990).
3.2.2.4 Peroxide Value (P.V.)
The peroxide value o f  the sample was determined using the method according to 
Ronald and Ronald (1991). In this method Ig o f  chloroform extract o f  the 
sample was weighed into a clean diy boiling tube followed by the addition o f  lg  
powdered potassium iodide and 20ml o f  solvent mixture (2 vol. glacial acetic 
acid = 1 vol. Chloroform). The boiling tube was then placed in boiling water so 
that it boiled within 30 seconds and then allowed to boil vigorously for 30 
seconds. The content was quickly poured into a flask containing 20ml o f  5% KB
59
solution. The boiling tube was washed out twice with 25ml water and the 
solution was titrated with 0.002M sodium thiosulphate solution using starch 
indicator. A blank was performed at the same time. The volume o f sodium 
thiosulphate utilised in the titration was measured and this was used to calculate 
the peroxide value in mil (equivalents per kilogram o f sample.
3.2.2.5 Thiobarbituric Acid (T.B.A) Number
The T.B.A. Number was determined using the method according to Ronald and 
Ronald (1991). In this method lOg o f soyflour and 50ml water were placed in a 
homogenizing flask and homogenized for 2 minutes. The homogenate was 
poured into a distillation flask and the homogenizing ilask was washed with 
47.5ml water. 2.5ml o f  4M HC1 was added and the pH was adjusted to 1.5. This 
was followed by the addition o f  a few glass beads. The solution was heated by 
means o f  an electric mantle and 50ml o f  distillate was collected 10 minutes from 
the time o f  boiling. 5ml o f  the distillate was pipetted into a glass-stoppered tube 
and TBA reagent (0.288g thiobarbituric acid/lOOml o f 90% glacial acetic acid) 
was added. The tube was stoppered, shaken and heated in boiling water for 35 
minutes. A blank was prepared similarly using 5ml waler with 15ml reagent. 
The boiling tubes were cooled in water for 10 minutes and the absorbance was 
measured against the blank using a spectrophotometer at 538nm. The TBA 
Number (mg/kg sample) was given by 7.8 x Absorbance.
60
The moisture content o f  the samples was determined using an air-oven 
(A.O.A.C., 1990)
3.2.3 Chemical analyses of stored soyflours
The samples were stored in environment with different temperatures and relative 
humidity as shown in Tables 3.2 and 3.3 for a period o f twelve (12) weeks. The 
following analyses were performed on the samples at various storage periods 
over the twelve- week period:
1. F.F.A. (3.2.2.3)
2. P.V (3.2.2.4).
3. T.B.A. Number (3.2.2.5)
4. Moisture content (3.2.2.6)
Total lipids and crude fat [(3.2.2.1 and 3.2.2.2) were determined only at times 0 
and 12 weeks],
3.2.3.1 Investigation of the effect of storage temperature and time on lipid 
stability in soyflour.
Experimental design
The factors considered (temperature and time) were designed according to the 
Central Composite Rotatable design for K=2 as in Table 3.1.
Five levels o f  each o f  the variables were established as shown in Table 3.2.
3.2.2.6 Moisture Content
61
Table 3.1 Design matrix and variable combination in experimental set up.
Coded X, -1 -1 1 1 0 0 0 1.414 -1 .414 0 0 0 0 0
var iab le s -! 1 -1 1 0 0 0 0 0 1.414 -1 .414 0 0 0
1 2 3 4 5 6 7 8 9 1(1 11 12 13 14
Table 3.2. Variables and their levels used in the central composite 
rotatable design for k=2._______________ ________ ________ ____________
Variable Code -1.414 -1 0 + 1 + 1.414
Temperature
(°C)
X, 5.0 15.9 42.2 68.5 80.0
Time
(weeks)
x 2 90 2 6 10 12
The packaged samples were put in the environment with the following 
temperatures:
(a) 5°C (cold room),
(b)l 5.9°C(incubator),
(c) 30°C (laboratoiy temperature),
(d) 42.2°C (incubator),
(e) 68.5°C (incubator) and
(f) 80°C (incubator) for twelve weeks.
The three samples (raw, roasted and cooked-dried soyflours) were analysed 
chemically for indices o f lipid oxidation and hydrolysis (that is, P.V., TBA
62
Number, FFA) at storage times o f 0, 2, 6, 10 and 12 weeks. Moisture content 
was also determined along side the lipid stability indices (see scction 3.2.2).
3.2.3.2 Investigation of the effect of relative humidity ami time on lipid 
stability in the various stored soyflours.
A moisture sorption isotherm was constructed and the monolayer value 
calculated for all the samples. The minimum relative humidity/water activity 
used for storage o f  the samples was obtained using the monolayer value as the 
reference point.
3.2.3.2.I Determination moisture sorption isotherm
All the samples used for the experiment were o f  the same o f particle sizes. The 
uniform particle sizes were obtained using the modified form o f  the method o f 
Ken-Jones el al. (1967). In this method 250g o f each o f the flours was placed in 
the uppermost sieve o f  the five sieves placed one on top o f  other having sieve 
aperture o f  2mm. 1mm, 0.5mm, 0.25mm and 0.063mm respectively. The 
amplitude o f vibration used was 50Hz and the mode o f  vibration was 
continuous. The vibration was monitored for five (5) minutes. At the end o f the 
vibration, the weight o f  the flour in each sieve plus the adhering particles on the 
underside o f  each sieve was determined. The weights were expressed as a 
percentage o f  the total weight. The flours in sieves with apertures o f  I mm, 
0.5mm and 0.25mm were pooled together and used for the sorption isotherm
63
determination as well as for the storage studies. The result o f  the particle size 
analysis is shown in Fig. 4.1.
The method used for the determination o f  moisture sorption isotherm is as 
follows:
Atmospheres o f  different relative humidities were established using saturated 
solutions o f  different salts as described by Speiss and W olf (1987). The 
saturated salt solutions and the corresponding relative humidities are shown in 
Table 3.1.The set-up used for exposing the samples to the various relative 
humidities is based on the principle o f  proximity equilibrium cell (Lang el al., 
1981). Each saturated salt solution was held in small sorption containers as 
described by Lang el al. (1981). This consisted o f a small cylindrical glass ja r 
(90mm long and 75mm diameter) with a fitted lid. Each sorption container was 
filled to about a third o f  its volume with saturated salt solution. About 2g o f  the 
flour was weighed into polypropylene weighing boats, 44 x 44mm, in size, and 
incubated under concentrated sulphuric acid for 24 hours so as to dry the 
samples to approximately zero moisture content. The samples were then 
suspended over the saturated salt solution, with a wire siring attached to the lid 
o f  the sorption container. The sorption containers with the food samples were 
placed inside an incubator set at 30°C. Triplicate determinations were made at 
each relative humidity and the weights o f  samples recorded. Samples became 
equilibrated between eight and fourteen days and the moisture content o f  each
64
sample was determ ined using the air-oven method (A.O.A.C., 1990).
Tablc3.3 Water activity of selcctcd saturated  
salt solutions at 30°C.
Salt Water 
activity (aw)
Lithium chloride 0.11
Glycerine 0.15
Potassium acetate 0.23
Potassium carbonate 0.45
Sodium bromide 0.58
Potassium iodide 0.68
Sodium chloride 0.75
Potassium chloride 0.84
Source:,Speiss,and, Wolf,(1987).
3.2.3.2.2 Determination of the effect o f water activity (relative humidity) 
and time on lipid stability in stored soyflour.
Experimental design
The factors considered (water activity and storage time) were designed 
according to the Central Composite Rotatable Design for k=2 as shown in Table 
3.1. Five levels o f  each o f the variables were established as shown in Table 3.4. 
The minimum water activity used for storage o f  samples was obtained using the 
water activity corresponding to the B.E.T. monolayer value as the reference 
point. The maximum water activity used for storage o f  samples was chosen
65
based on the fact that it is the average normal relative humidity in the coastal 
belt o f  Ghana.
Table 3.4 Variables and their levels used in the central composite 
rotatable design for k=2. ______  __
Variable Code -1.414 -1 0 + 1 +1.414
Water activity X3 0.15 0.23 0.45 0.68 0.75
Time (weeks) X, 0 2 6 10 12
Experimental procedure
Saturated solutions o f  salts corresponding to the water activities chosen were 
prepared using Table 3.3 as reference. Each o f  the solutions was poured into a 
desiccator and wire gauze was placed inside the desiccator about 8cm above the 
surface o f  the solution. The packaged soyflour samples were placed on the wire 
gauze and the desiccator covered with an airtight lid. The whole set-up was put 
in an incubator set at 30°C. Samples o f  each o f  the soyf lours were taken at time 
intervals o f  0, 2, 6, 10 and 12 weeks for the following chemical analysis:
(a) Moisture content
(b) Peroxide value
(c) Thiobarbituric acid number
(d) Free fatty acids
(e) Total lipids and crude fat content were determined only at times 0 and 12 
weeks o f  storage (see section 3.2.2)
66
3.2.4 Sensory evaluation
The samples used for the sensory evaluation were stored at cold room 
temperature (5°C) and ordinary room temperature (30°C). A panel o f  fifteen 
(15) judges selected randomly from the Nutrition and Food Science Department 
o f  the University o f  Ghana were used for the test. The soy flour samples in 
storage were analysed organoleptically over a period o f  twelve weeks. The first 
two sensory evaluations were performed after four and eight weeks o f  storage 
respectively. The subsequent analyses were performed after ten and twelve 
weeks o f  storage. The questionnaire used for the sensory test is shown on 
Appendix 1.
3.2.5 Statistical analyses
The data obtained from the experiment were subjected to multiple regression 
analysis with dependent variables P.V., TBA Number and FFA. Three 
dimensional response surface plots were generated from the multiple regression 
models obtained using a stepwise multiple regression technique. The response 
surface plots were generated to illustrate the simultaneous triple effects o f 
variables. The regression models were developed from Statgraphics software 
(Statgraphics, STSC Inc. version 4.2 U.S.A.).
67
CHAPTER FOUR
RESULTS AND DISCUSSION
4.1 General observation of soyflour samples.
The raw, cooked-dried and roasted soyflours were different in terms o f  
appearance (colour) and flavour. The raw soyflour had a crcam colour; the 
cooked-dried had a light brown colour while the roasted flour had a brown colour. 
The brown colour o f  the processed flour might be di'e lo browning reaction 
between the amino groups o f  proteins and the carboxyl group o f  carbohydrates 
(sugars) present in the beans (deMan, 1990). The roasted beans were subjected to 
prolonged dry heat than the cooked -d ried  hence the difference in shades o f  the 
brown colour. All the three Hours had slightly different flavours. The raw  flour 
had a beany, bitter flavour. The roasted flour however did not have a detectable 
beany flavour but had a strong flavour characteristic of roasted bean or cereal 
flour. The absence o f  beany flavour in the processed flour, that is, roasted and 
cooked-dried, might be due lo heat inactivation o f  the enzyme lipoxygenase 
which is believed to be responsible for the development o f the bitter, beany o ff 
flavour (Baker and Muslakas, 1972).
4.2 Particle size analysis o f soyflour
The results for the particle size analysis are shown on Fig.4.1. For all the three 
samples, that is, raw, cooked-dried and roasted soyflours, a high proportion o f  the
68
flour was retained in the sieves with apertures between 0.25min and lmm , with 
the highest proportion at 0.5mm. Apart front the cooked-dried Hour, both raw and 
roasted flour had more o f  the sample retained in the sieve with aperture 2mm 
(largest) than in the sieve w ith aperture 0.063mm (smallest). Among the three 
samples raw flour had the highest amount o f samples retained in the largest sieve 
probably because o f  the hard seed coat, which might have been difficult to grind. 
The roasted flour had fewer larger particles than the raw flour probably because 
the roasting might have made the seed coat and the bean drier and more fragile 
hence grinding was easily effected.
4.3 Chemical analyses o f freshly prepared soyflour samples 
The results o f  the chemical analyses performed on the freshly prepared raw, 
roasted and cooked-dried soyflours are shown in Table 4.1. The raw soyflour 
showed a higher level o f  moisture than the processed flours and this could be 
attributed to the fact that the two processing methods (roasting and cooking- 
drying) involved heat drying, which might have led to some dehydration. The raw 
soyflour also appeared to have higher values o f  free fatty acids (FFA), peroxide 
value (P.V.) and thiobarbituric acid number (TBA no.) than the processed flours. 
This might be an indication that lipid-hydrolysing enzyme, lipase, and lipid- 
oxidizing enzyme, lipoxygenase, might have been inactivated through heat 
processing hence the mode o f  lipid oxidation might be mainly autoxidation and 
thermal oxidation (Hamilton and Berger, 1995).
69
W
ei
gh
t 
of 
sa
m
pl
e 
re
ta
in
ed
 
(%
)
40
0.063 0.25 0.5 1 2
Sieve aperture (mm)
Fig.4.1 Particle size analysis o f  soy flours
H  Raw " Cooked 13 Roasted]
70
T able 4.1 M oisture content, crude fat content, free fatty acids, hydrolytic and 
oxidative products in freshly prepared raw and processed soyflours.
Soyflour Moisture 
content 
(% DMB)
Crude fat 
(%)
Total lipids
(%)
Free fatty 
acids (%)
Peroxide
value
(meq/kg)
Thiobarbituri 
acid number 
(mg/kg)
Raw 9.78±0.52 20.1610.16 26.29+0.09 1.76+0.10 2.56+0.21 12,50+1.03
Roasted 5.51±0.81 20.56+0.09 26.56+0.11 1.13+0.06 1.42+0.18 7.25+0.67
Cooked-
dried
4.74+0.46 20.48+0.11 26.82 +0.15 1.39+0.08 1.72+0.12 6.84+0.52
Note: Hydrolytic products include free fatty acids while oxidative products 
include hydroperoxides and thiobarbituric acid reactive substances.
In the case o f  the raw flour the mode o f  lipid oxidation might be mainly 
enzymatic and autoxidation (Hamilton and Berger, 1995). The lipoxygenase 
enzyme acts on the unsaturated acids as soon as the raw bean tissue is raptured 
and the enzyme and the substrate are released (Satouchi and Matsushita, 1976). 
Baker and Mustakas (1972) showed that when the raw soybean tissue is broken 
the enzyme and the substrate are liberated and provided some moisture is present 
a bitter, beany taste develops very rapidly. Lipoxygenase have been shown to 
generally accelerate the oxidation o f  lipids (Markley, 1951).
71
The higher moisture content and the absence o f heat treatment might have been 
the cause o f  the high FFA content in the raw soyflour. Lipases are capable o f  
hydrolysing lipids in the presence o f moisture to fatty acids and glycerol (deMan, 
1990). The lipases are however inactivated at temperatures above 50°C 
(Whitaker, 1972). It has also been reported that the fatly acids produced from the 
hydrolysis are even more susceptible to oxidation (Satouchi and Matsushita, 
1976). This observation explains further why the raw  soyflour had more oxidative 
products (hydroperoxides and TBA reactive substances or TBARS).
4.4 Chemical analyses o f stored soyflour samples
4.4.1 Effect of storage temperature and time on lipid stability
Labuza (1975) reported that the rate o f  lipid oxidation in diy foods depends 
greatly on the moisture content. Whitaker (1972) also reported that water is an 
important requirement for lipid hydrolysis. With reference to these reports the 
moisture content in all the samples were determ ined along side the oxidative and 
hydrolytic properties. Generally the moisture contents o f  all the three soyflours 
(raw, roasted and cooked-dried) were found to decrease with increase in 
temperature and time. At 30°C (room temperature) however the moisture content 
was observed to increase slightly, thal is, from 9.78 to 9.98 for raw flour, 5.51 to 
5.71 for roasted flour and 4.74 to 5.09 for cooked-dried flour, at the end o f  the 
twelve weeks storage period. The slight increase in moisture content might be due 
to the absorption o f  moisture from the environment, which had a relative 
humidity o f approximately 75% (ambient). The slight decrease in moisture
72
content o f  all the samples at 5°C (cold room) might be as a result o f  loss o f 
moisture resulting from the process o f  cooling (Badings, 1970). Tables 4.2 and
4.3 show the variation o f  moisture with temperature and time in the raw and 
roasted soyflour samples.
Table 4.2 Effect of storage temperature and time on the moisture content of 
packaged roasted soyflour_________________________________________________
Temperature
(°C)
Moisture content o f  Roasted soyflour (%)
WeekO Week2 Week6 Week 10 Week 12
5 5.51 5.49 5.42 5.11 5.00
15.9 5.51 5.52 5.51 5.55 5.55
3 0 5.51 5.56 5.62 5.67 5.71
42.2 5.51 5.35 5.22 4.86 4.47
68.5 5.51 5.02 4.52 421 4.01
80 5.51 4.83 4.22 3.52 2.79
Table 4.3 Effect o f storage temperature and time on the moisture content of 
packaged raw soyflour_____________________________________________________
Temperature
(°C)
Moisture content o f raw  soyflour (%)
WeekO Week2 Week6 Week 10 Week 12
5 9.78 9.78 9.64 9.60 9.51
15.9 9.78 9.77 9.80 9.82 9:82
30 9.78 9.78 9.86 9.93 9.98
42.2 9.78 9.55 8.57 8.00 7.21
68.5 9.78 9.54 9.03 7.34 6.29
80 9.78 9.02 8.54 7.03 5.56
7.1
4.4.1.1 Effect o f storage temperature and time cn Free Fatty Acids of 
soyflau rs
The FFA for all the three samples were observed to have generally increased with 
time at all the selected temperatures. The increase in FFA was higher in raw 
soyflours at all the selected temperatures than in the processed flours. This could 
be attributed to the inactivation o f  lipases during heat processing o f  the soybeans. 
At 5°C however, the increase in FFA was very minimal, that is, from 1.39 to 1.61 
for cooked-dried, 1.13 to 1.35 for roasted and 1.76 to 2.13 for raw  soyflour by the 
end o f  twelve weeks o f  storage.
The Free fatty acids in all the three soyflours increased as temperature was raised 
from 5°C, to reach a maximum value (that is, 2.42% in roasted flour, 2.54% in 
cooked-dried and 4.02% in raw soyflour) at 30°C and thereafter decreased as 
temperature was increase from 42.2°C to 80°C. These observations could be 
explained based on the observation by Whitaker (1972) that lipases have their 
optimum temperature around 30°C and about 50°C denaturation begins. At about 
85°C the lipases are completely denatured (Whitaker, 1972). The decrease in the 
FFA content beyond 30°C might also be due to the accumulation o f  a lot o f  FFA 
resulting from the high rate o f  hydrolysis at 30°C, which might have acted as a 
competitive inhibitor o f  the enzyme (Whitaker, 1972). Fig.4.2 shows the 
relationship between the FFA and temperature for all the three soyflours after 
twelve weeks o f  storage.
74
Fr
ee
 
fat
ty 
ac
id
s 
(%
)
5.5 
5 -
4.5 - 
4 -
3.5 
3 -
2.5 - 
2
1.5 - 
1
0.5 -
0 -I 1----------- 1— -------- 1----------- [—--------1----------- 1----------- 1----------- 1----------- 1
0 10 20 30 40 50 60 70 80 90
Temperature (°C)
Fig 4.2 Free fatty acids in raw, cooked-dried and roasted soyflours stored at 
different temperatures for twelve (12) week?
A multiple regression analysis gave a model with an R2 e f  91%, 97% and 95% for 
raw, cooked-dried and roasted soyflours, respectively. None o f  the models had a 
significant “lack o f  fit” (Appendix 2,3,4) implying that each model is sufficient 
for predicting the FFA in soyHour at any given storage temperature and storage 
time.
In all the three soyflours storage time had both linear and quadratic effects on the 
FFA content (Table 4.4). The linear effect o f  storage temperature on FFA content 
was significant only in the cooked-dried soyflour whereas the quadratic effect was 
significant in all the three soyflours. There was however a non-significant effect 
o f  the interaction between the two factors (storage temperature and time) on the 
FFA content in all the three soyflours.
Table 4.4 Further ANOVA for variation in the order fitted (showing only the 
P-values) for FFA content in soyflours stored at various temperatures over a 
twelve-week period_________________________________________________________
Source o f  variation P-values
Raw Cooked Roasted
Time 0.0001* 0.0001* 0.0001*
Temp 0.5380 0.0257* 0.1719
Time2 0.0114* 0.0004* 0.0001*
Temp2 0.0107* 0.0340* 0.0056*
Time *Temp 0.7516 0.1519 0.8466
* = Significant P-values
76
3-D response surface plots, for FFA content in soyflour es  a function of storage 
temperature and time were generated from the following regression equations and 
are shown in Figs. 4.3-4.5.
4.4.1.2 Effect of storage temperature and time on Peroxide value and the 
TBA number of stored soyflours
The peroxide values and the TBA numbers o f  the soyflours were observed to 
increase generally with time and temperature. These can be seen in Figs. 4.6- 
4.11. It was observed that, as the peroxide value increased there was a 
corresponding increase in the TBA number.
In all the three soyflours there was no significant change in the P.V. at storage 
temperatures o f  5°C and 15.9°C after four weeks o f storage. However beyond 
four weeks the P.V. o f  the raw soyflour began to increase gradually and by the 
eighth week the increase was quite pronounced. At 68.5°C and 80°C storage 
temperatures, the P.V. o f  the raw soyflour increased at an accelerated rate 
compared to the TBA number, reaching a maximum at 8 weeks and thereafter 
declined.
An inverse relationship was observed between the P.V. and the TBA Number 
after eight weeks o f  storage at 68.5°C and 80°C in r?w soyflour, indicating 
progression o f  oxidation from a primary to a secondary stale.
77
Regression equation for Fig 4.3
Z = 0.95233 - 0.0235It + 0.045IT  + 0.02069212-  0.000467T2 - 
0.000476Tt 
Z = FFA
T = Storage temperature 
t = storage time.
R2 =91%
10
°  Temperature (°Q
12
Time (weeks)
Fig. 4.3 The 3-D response surface for the free fatty acids (Z) in raw soy 
flour as a function of temperature (T) and time (t).
i s
T = storage temperature
t =  storage time.
r 2 =  9 7%
Regression equation for Fig 4.4
Z = 1.377721 -  0.0453751 + 0.005544T + 0.008994t2 -  0.000073T2 + 0.000546TI
Z = FFA
Time (weeks)
Fig 4.4 The 3-D response surface for the free fatty acids (Z) in cooked- 
dried soyflour as a function of temperature (T) and time (t)
~ f9
T = storage temperature
Regression equation fo rF ig  4.5
Z = 1.23069 -  0.0821711 + 0.015625T + 0.014325t2 -  0 .000154T2 -  0.000095Tt
Z = FFA
t = storage time.
R2 = 95%
i ^ ce
Vo®$ e e
80
F '\9
o.O  e^ s9 ° ' oPra^ e
A. S ^ e  , n O ° ^ e
^ ° ° °
By the end o f the twelfth week storage period the processed (lour had not shown 
any decline in the P.V. indicating that a longer storage time was required for the 
P.V. to reach the maximum level. This also meant that the rate o f  production o f 
the thiobarbituric reactive substances (TBARS) was still low since the rate had 
been found to increase sharply as soon as the P.V. began to decline.
In general the raw soyflour showed a much higher increase in both the P.V. and 
the TBA Number than the processed soyflours. These results suggest that both 
moist and diy heat processing o f  soybeans have the ability to increase lipid 
stability o f  soy products. The difference in the rate o f  lipid oxidation in raw  and 
processed soyflour could be attributed to the fact that the heat treatment given to 
the processed flours might have inactivated the lipid oxidising enzyme, 
lipoxygenase (Baker and Mustakas, 1972). The oxidation o f  lipids in the 
processed samples might therefore be due to autoxidation, thermal oxidation and 
photo-oxidation whereas in the raw flour there was an additional mode o f  
oxidation, that is, enzymatic (Hamilton and Berger, 1995). The decline in the P.V. 
after reaching a maximum value could be due to a decrease in the level o f  
substrate (oxygen or unsalurated fatty acids) as peroxidation progressed (deMan 
1990). The fact that the samples were packaged in high-clensity polythene might 
have reduced the transfer o f  oxygen into the package hence the oxygen inside 
could be used up almost completely (Hamilton and Berger. 1995).
Pe
ro
xi
de
 
va
lu
e 
(m
eq
/k
g)
T ime (weeks)
Fig 4.6 Effect o f  storage temperature and time on the peroxide value 
o f raw soy flour.
-5°C  —B— I6eC — 42°C —H— 68°C - ® - 8 0 ’C
27
24
/•—sto 21
1  18
ja 15
o
Z 12
< 9
P 6
3
0
0 2 4 6 8 10
Time (weeks)
Fig 4.7 Effect o f  storage temperature and time on the 
o f  raw soyflour.
TBA Number
TB
A 
N
o.
 (
m
g/
kg
) 
Pe
ro
xid
e 
va
lue
 (
m
eq
/k
g)
Time (weeks)
Fig 4.8 Effect of storage temperature and time on Ihe peroxide value  
of roasted soyflour.
—#— 5°C —*8— 16°C — 42°C -X -68°C  — B0°c]
Fig 4.9 Effect o f  storage temperature and time on the TBA
Number o f roasted soyflour.
5°C 16°C —a—42°C — 68°C ~ ^ - 8 0 ° c ]
83
Time (weeks)
Fig 4.10 Effect o f  storage temperature and time on the peroxide 
value o f cooked-dried soyflour.
■5°C 16°C -42°C ■68°C — 80°C
2 4 6 8 10 12 14
Time (weeks)
Fig 4.11 Effect o f  storage temperature and time oil the TBA 
Number o f  cooked-dried soyflour.
j—» - 5 0C —m—16°C — 42°C H * -68 °C  - o - 8 0 ° C
84
The general increase in the P.V. and the TBA Number as temperature increased 
confirms the general rule that the oxidation o f lipids approximately doubles for 
every 10° rise in temperature (Hamilton and Berger, 1995).
A multiple regression analysis gave a model with an R2 of 96%, 98% and 92% for 
P.V. in raw, cooked-dried and roasted soyflours respectively. The regression 
models for TBA Number in soyflour gave R2 o f 99%, 98% and 97% for raw, 
cooked-dried and roasted soyflours respectively. None o f  the models had a 
significant “ lack o f  fit” (Appendices 5,6, 7,8,9,and 10).
In all the three soyflours the linear terms o f  temperature and time had significant 
effects on the P.V. The quadratic term o f  temperature however affected only the 
P.V. in the roasted flour significantly. The quadratic term o f time also affected 
only the P.V. in the cooked-dried soyflour significantly. The interaction between 
temperature and time had significant effect only on the raw  and cooked-dried 
soyflours (Tables 4.5 and 4.6). The TBA Numbers o f  all the three soyflours were 
affected significantly by the linear terms o f  temperature and time and also by the 
interaction between the two factors. The quadratic terms o f temperature and time 
affected significantly the TBA Number o f  raw and cooked-dried soyflours only.
3-D response surface plots for P.V. and TBA Number o f  soyflour as a function o f 
temperature and time were generated from the regression equations and are shown 
in Figs 4.12-4.17.
85
Tabic 4.5 Further Anova for variation in the order fitted (showing only the 
P-values) for Peroxide value in soyflours stored at various temperatures over 
a twelve-week p e r i o d . _________________________________________________
Source o f variation P-values
Raw Cooked Roasted
Time 0.0001* 0.0001* 0.0001*
Temp 0.0001* 0.0001* 0.0004*
(Time)2 0.7848 0.0150* 0.4548
(Temp)2 0.4582 0.8491 0.0253*
Time *Temp 0.0337* 0.0001* 0.1195
* = Significant P-value
Table 4.6 Further Anova for variation in the order fitted (showing only the 
P-values) for TBA Number in soyflours stored a t various temperatures over 
a twelve-week period._______________________________________________________
Source o f  variation P-values
Raw Cooked Roasted
Time 0.0001* 0.0001* 0.0001*
Temp 0.0001* 0.0001* 0.0001*
(Time)2 0.0001* 0.0001* 0.1532
(Temp)2 0.0258* 0.0012* 0.0562
Time *Temp 0.0001* 0.0001* 0.0082*
»=  Significant P-value
86
T = storage temperature
Regression equation for Fiiz 4.12
Z -  0.399762 + 0.534072t + 0.015502T -  0.006632t2 + 0.000464T2 + 0.014082Tt
Z = P.V.
T = storage time
R2 = 96%.
10
o  Temperature (°C)
12
Time (weeks)
Fig 4.12 The 3-D response surface for the peroxide value (Z) in raw 
soyflour as a function of temperature (T) and time (t).
$ 7
T = storage temperature
Regression equation for Fig 4.13
Z = 1.787188 - 0.0715071 -  0.003508T -  0.006632t2 + 0.00001T2 + 0.005366Tt
Z = P.V.
t = storage time
R2 = 98%
Time (weeks)
Fig 4.13 The 3-D response surface for the peroxide value (Z) in cooked- 
dried soyflour as a function of temperature (T) and time (t)i
Regression equation for Fig 4.14
Z = 1.75484 + 0.024894t -  0 .048513T + 0.00839 lt2  +0.00074T2 + 0.004132Tt 
Z =  P.V.
T  =  storage temperature
t = storage time
R2 = 97%
^  Temperature ( ° C )
10 12
Time (weeks)
Fig 4.14 The 3-D response surface for the peroxide value (Z) in roasted 
soyflour Vas a function of temperature (T) and time (t)
Regression equation for Fig 4.15
Z = 12.922889 -  0.2326341 -  0.025856T + 0.030753t2 + 0.000382T2 +
0.007883Tt
Z= TBA no.
T = temperature 
t = storage time).
R2 = 99%
a
o
G
<
Time (weeks)
Fig 4.15 The 3-D response surface for the TBA no. (Z) in raw soyflour: as 
a function of temperature (T) and time (t).
<>0
Regression equation for Fin 4.16
Z = 8,007328 -  0.208362t -  0 .033737T+ 0.017071t2 +0.000304T2 +0.005794Tt 
Z  =TBA no.
T = storage temperature
t = storage time
R2 = 98%
Time (weeks)
Fig 4.16 The 3-D response surface for the I'BA no. (Z) in cooked-dried 
soyflour as a function of temperature (T) and time (t).
Regression equation for Fig 4.17
Z = 6.932735 + 0.021054t -  0.01765T + 0.00674It2 - +0.000256T2 +  0.003372Tt 
Z  = TBA no.
T  = storage temperature
t = storage time
R2 = 94%
TB
A 
no
. 
(m
g/
kg
)
Time (weeks)
Fig 4.17 The 3-D response surface for the TBA no. (Z) in roasted
soyflour as a function of temperature (T) ,and time ({), .
9X
4.5 Effect of water activity (a„) and time on lipid stability in stored soyflour.
4.5.1 Moisture sorption isotherms and IJ.E.T. monolayer values 
According to Labuza (1975), moisture content elose to the monolayer coverage o f 
food has a protective action against' lipid oxidation and results in minimum lipid 
oxidation rate. Based on these findings, the BET monolayer values o f  all the 
soyflour samples were determined from the BET plots shown in Figs 4.33-4.35, 
which were plotted using the following BET equation:
A/(1-A)M = 1/M()C + (C-1/M0C)A, where A = water activity; M  =  equilibrium 
moisture content; M0 = monolayer value; C = a constant related to heat o f 
adsoiption.
Moisture sorption isotherms for all the three soyflours are shown on Fig. 4.18.
The calculated B.E.T. monolayer values and the corresponding aw are shown on 
Table 4.7. The corresponding a„s were obtained from the sorption isotherms 
(Fig.4.18).
Table 4.7 B.E.T. monolayer values for different types of soyflour at 30°C.
Soyflour B.E.T. monolayer value Corresponding aw
Raw 2.87 0.24
Roasted 2.96 0.23
Cooked-dried 2.93 0.22
a/
(l
-a
)m
W ater activity 
F ig 4.33 B.E.T plot for raw  soyflour at 30°C
Water activity 
Fig4.34 B.E.T. plot for cooked-dried soyflour at 30°C.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Water activity
Fig4.35 B.E.T plot for roasted soyflour at 30°C
Eq
ul
ib
riu
m
 
m
oi
stu
re
 
co
nt
en
t 
(g
/lO
Og
 
so
lid
s)
W ater activity
Fig 4.18 Moisture sorption isotherms for raw, cooked-dried and roasted
soyflours at 30°C
Raw —re— Cooked —<*>— Roasted
3 1
1 he monolayer water is the amount o f water in a system, which is unavailable as 
a solvent (Labuza, 1975). As indicated in Table 4.18, there was no significant 
difference in the monolayer values as well as the corresponding water activities in 
all the three soyflours.
4.5.2 Effect of water activity and time on peroxide value and TBA Number of  
stored soyflours
In all the three soyflours the P.V. and the TBA Number generally increased with 
time at all the selected water activities, that is, 0.15, 0.23 (average BET 
monolayer value), 0.45, 0.68 and 0.75. The P.V. was high at 0.15 aw and then 
decreased to a minimum value at 0.23 avv (that is, about 1.94 meq/kg in roasted 
soyflour, 2.21 meq/kg in cooked-dried soyflour and 3.89 meq/kg in raw soyflour) 
and thereafter increased gradually as aw was increased (Fig.4.19).
A similar trend was observed for the TBA Number (Fig 4.20). The minimum 
value o f P.V. and that o f  TBA Number obtained at 0.23 a,v (the approximate aw 
corresponding to the monolayer value o f  all the three soyllours could be due to 
the water present in each sample, that is, 2.87% in raw, 2.96% in roasted and 
2.93% in cooked-dried soyflours. This water might have formed a protective 
barrier, preventing the oxygen from reaching the underlying unsaturated fatty 
acids (Salwin, 1959). I-Ialton and Fischer (1967) also proposed that the monolayer 
water retarded the diffusion o f oxygen to the sites o f  the unsaturated double 
bonds.
98
Pe
ro
xi
de
 
va
lu
e 
(m
eq
/k
g)
13 - 
12 
11 -
10  -  
9 
8 
7 
6
5 H 
4 
3 
2 
1 
0
14
0 0.1 0.2 0.3 0.4 0.5
Water activity
0.6 0.7 O.f
Fig 4.19 Peroxide values of raw, cooked-dried and roasted soyflours 
stored at various water activities, at 30°C, for a period o f twelve (12)
weeks.
-Raw ■Cooked —a—Roasted
3 5
TB
A 
N
o.
 (
m
g/
kg
)
24 
22 
20 H 
18 
16 - 
14 - 
12  -  
10 -  
8 
6 
4
2 H 
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Water activity
F ig  4.20 TBA Nos. o f different types of soyflour stored at various water 
activities, at 30°C, over a period of twelve (12)weeks.
-Raw ■Cooked Roasted
loO
Al a„. below the monolayer value the P.V. and the TBA Number decreased with 
increasing aw and this could be explained from the fact that water was acting as an 
antioxidant at those water activities. This is attributed to water hydrogen- bonding 
to hydroperoxides and hydration o f  metal catalysts hence reducing the peroxides 
concentration and overall lipid oxidation rate (Heidelbaugh and Karel, 1970). The 
increase in the rate o f  oxidation at aw above the monolayer value could be 
attributed to the increased water content, which reduces the viscosity o f  the 
solution hence promoting mobility o f  reactants (Heidelbaugh and Karel, 1970). 
Labuza (1975) also proposed that the increase in the rate o f  lipid oxidation as aw 
increases might be due to dissolution o f precipitated catalysts and swelling o f 
solid matrices hence exposing new catalytic surfaces.
A multiple regression analysis for P.V in soyflour gave an R2 o f 94%, 92% and 
95% for raw, cooked-dried and roasted soyflours respectively. None o f  the 
models showed a significant “ lack o f  fit” (Appendices 14, 15, and 16) implying 
that all the models are sufficient for predicting the P.V. in soyflour at any given 
value o f  water activity and time.
In all the soyflour samples both the linear terms o f  watar activity and time, and 
the interaction between time and water activity affected the P.V. significantly. 
Storage time however had no significant second power effect on the P.V. in all 
the three soyflours (Table 4.8).
101
A 3-D response surface plot showing the P.V. as a function o f water activity and 
time were generated from the regression equations and are shown in Figs 4.21 — 
4.23.
Table 4.8 Further Anova for variation in the order fitted (showing only the 
P-values) for peroxide value in soyflours stored at various water activities 
(a„) over a twelve-week period. .____________________________ ___
Source o f  variation P-values
Raw Cooked Roasted
Time 0.0001* 0.0001* 0.0015*
avv 0.0004* 0.0001* 0.0001*
(Time)2 0.5142 0.6211 0.2993
(aw)2 0.1518 0.0422* 0.3168
Time * aw 0.0019* 0.0016* 0.0107*
* =  Significant P-value
Table 4.9 Further Anova for variation in the order fitted (showing only the 
P-values) for TBA Number in soyflours stored at various water activities (aw) 
over a twelve-week period.__________________________________________________
Source o f  variation P-values
Raw Cooked Roasted
Time 0.0001* 0.0001* 0.0001*
aw 0.0006* 0.0001* 0.0001*
(Time)2
0.5240 0.0339* 0.0275*
(aw) 0.2344 0.0012* 0.2106
Time * aw 0.0047* 0.0001* 0.0016*
. = Significant P-value
10 'J
A multiple regression analysis lor TBA Number in soyflour produced models 
with R2 o f 92%, 98% and 97% for raw, cooked-dried and roasted soyflours 
respectively. All the models had no significant “ lack if fit” (Appendixl7, 18, 
and 19) hence the models could predict the TBA Number in soyflour for any given 
water activity and storage time.
The TBA Number in all the three soyflours were affected significantly by the 
linear terms o f  water activity and storage time as well as the interaction between 
the two factors (Table 4.9). Storage time had a second power effect on only the 
heat processed soyflours.
3-D response surface plots showing the TBA Number as a function o f  water 
activity and time were generated from the regression equations and are shown in 
Figures 4.24-4.26.
ID3
Regression equation for Fig 4.21
Z = 0.972049 -  0.002047t + 7.825284w -  0.007642t2 -  9.836994W2 + 
1.3338256wt 
Z = P.V.
w = water activity 
t = storage time 
R2 =92%
bO
£(L)
g,
<D ■ 3
i •<si>
•a
Xo
v-.<u
P-,
Time (weeks)
F i g  4 .21  The 3-D response surface for the peroxide value (Z) in r a w
soyfloui as  a function of water activity (w) and time (t)-1
10'f
Regression equation for Fig 4.22
Z = 1.33564 -  0.1213t + 1.266546w + 0.006453t2 -  2.499652W2 + 0.552486wt 
Z = P.V.
w = water activity 
t =  storage time
bO
cr0J>
E,
<D
”cd>
1)
o1-.<L>Oh
Time (weeks)
F ig  4 .2 2  The 3-D response surface for the peroxide value (Z) in c o o k e d -
d r ie d  soyflour5 as a function of water activity (w) and time (t).
105
Regression equation for Fig 4.23
Z = 1.38135 -  0 .1122t + 1.256866w + 0.006293t2 -2.487654W2 +0.540447wt 
Z = P.V.
w = water activity 
t= storage time 
R2 = 95%
o o A — - /  
2  4 «  s  , 0
Time (weeks)
Water activity
12
F i g  4.23 The 3-D response surface for the peroxide value (Z) in r o a s t e d
soyflour as a function of water activity (w) and time (\)
Regression equation for Fig 4.24
Z = 12.168641 -0 .1253271 + 3.8 5 3 9 9 8w -0 .011933t2 -  7.480869w2 + 
1.581264wt 
Z = TBA no. 
w = water activity 
t = storage time 
R2 = 92%
1  a,
oc
<
ea
Time (weeks)
Fig 4.24 The 3-D response surface for the TBA no. (Z) in raw  soyflour as
a function of water activity (w) and time (t) '
\Q -f-
Water activity
Regression equation for Fig 4.25
Z = 8.125469 -  0.1959711 -  4.025325 w + 0.010167t2 +4.217257W2 + 0.715521 wt 
Z -  TBA no. 
w = water activity 
t = storage time).
R2 =98%
Time (weeks)
F i g  4 .2 5  The 3-D response surface for the TBA no. (Z) in c o o k e d - d r ie d
soyflour as a function of water activity (w) and time (t)*
108
Regression equation for Fig 4.26
Z= 7.346747 -  0 .161642t -  2.648091 w + 0 .014505t2 +2.6804W2 +  0.643557wt 
Z = TBA no. 
w = water activity 
t =  storage time).
R2= 97%
tan
bh
E
o'c
<
P
Time (vv
Fig 4 .26  The 3-D re sp o n se  su rface  for the TBA no. (Z) in ro a s te d
soyflour as a function of walei activity (w) and time (I).
109
4.5.3 Effect o f water activity and time on free fatty acids in stored soyflours
The free fatty acid content o f all the three soyflours increased with storage time 
and aw as shown on Figures 4.27-4.29. These findings confirm the report by Acker 
and Wiese (1972) that lipid hydrolysis by the enzyme lipase increases with 
increase in aw from the dry state.
Among the three soyflours raw soyflour recorded the highest rate o f  lipid 
hydrolysis and this might probably be due the fact that the sample was not given 
any heat treatment hence the lipase was not inactivated. It was also observed in all 
the three samples that the rate o f  FFA production increased at aw 0.45, 0.68 and aw
0.75 after about eight (8) weeks o f storage. This might be explained from the fact 
that at these aws (he samples might have absorbed more moisture hence growth o f 
moulds and other microorganisms might have been encouraged. Some o f  these 
microorganisms might have produced lipases, and in the presence o f  adequate 
moisture rate o f  lipid hydrolysis might be increased (Whitaker, 1972).
HO
Fr
ee
 
fat
ty 
ac
id
s 
(%
)
T ime (weeks)
Fig 4.27 Effect o f  water activity and storage time on the free fatty acids in raw
soyflour at 30°C.
—*—aw=0.15 -® -aw =0 .23  —a— aw=0.45 aw=0.68 —'9— aw=0.75
Fr
ee
 
fat
ty 
ac
id
s 
(%
)
Time (weeks)
Fig 4.28 EfFect o f water activity and storage time on the free fa lty  acids in cooked-
dried  soyflour at 30°C.
—♦— aw=0.15 —s*— aw=0.23 — * r —  aw=0.45 aw=0.68 —e—aw=0.75
112.
Fr
ee
 
fat
ty 
ac
id
s 
(%
)
Time (weeks)
Fig 4.29 Effect o f water activity and storage time on the free fatty acids in 
roasted soyflour at 30°C.
—♦— aw=0.15 — aw=0.23 — aw=0.45 aw=0.68 — a\v=0.75
A multiple regression analysis for FFA in soyflour produced regression models 
with II2 o f  99%, 99% and 98% for raw, cooked-dried and roasted soyflours 
respectively. None o f  the models had a significant “lack o f  fit” (A ppend ix ll, 12, 
and 13).
The linear terms o f  the two factors, water activity and time, had significant effect 
on the FFA in all the three soyflours (Table 4.10). Interaction between the two 
factors also affected the FFA significantly. The quadratic term o f  time however 
affected only the FFA in raw and cooked-dried soyflours significantly. Water 
activity also had a second power effect on the free fatty acids in cooked-dried and 
roasted soyflours.
3-D response surface plots showing the FFA in soyflour as a function o f  water 
activity and storage time, generated from the regression equations are shown in 
Figs 4.30 -  4.32.
Table 4.10 Further Anova for variation in the order fitted (showing only the 
P-values) for FFA content in soyflours stored at various water activities (aw) 
over a twelve-week period.__________________________________________________
Source o f  variation P-values
Raw Cooked Roasted
Time 0.0001* 0.0001* 0.0001*
aw 0.0001* 0.0001* 0.0001*
(Time)2 0.0001* 0.0001* 0.0609
(aw / 0.8391 0.0027* 0.0303*
Time *aw 0.0001* 0.0001* 0.0001*
»= Significant P-value
M4-
4.5.4 Effect of lipid oxidation and hydrolysis on tolal lipids and crude fat 
content in soyflour
I lie total lipids in all the three soyflours were found to decrease at the end o f  the 
storage period as shown on Tabic 4.3. Raw soyflour had the highest total lipid 
loss o f 6.50% followed by cooked-dried and roasted soyflours with losses o f  
1.64% and 1.58% respectively. Like the total lipids the crude fat content o f  the 
soyflours decreased with storage time. At the end o f  fhe twelve-week storage 
period raw soyflour lost 8.58% o f the crude fat whilst cooked-dried and roasted 
flours lost 2.78% and 2.38% respectively.
Table 4.11 Total lipids in various forms o f  freshly prepared soyflour as well as 
those stored for twelve (12) weeks at 30°C._______________________ ________ ____
Soyflour type Moisturecontent (%) Total lipids (%) Storage loss
Fresh Stored Fresh Stored Total lipids 
loss (%)
Raw 9.78 9.98 26.29 + 0.09 24.58 + 0.21 6.50
Roasted 5.51 5.71 26.56 + 0.11 26.14 + 0.15 1.58
Cooked-dried 4.94 5.02 26.82 + 0.15 26.36 + 0.24 1.64
M i'
Regression equation for Fig 4.30
Z = 1.835393 -  0.1832031 -0.143718w + 0.032932t2 + 0.247847W2 + 0,572356wt 
Z = FFA
w = water activity 
t = storage time).
R2 =99%
cN
£m
Water activity
Time (weeks)
Fig 4.30 The 3-D response surface (or the free fatty acids (Z) in raw
-soy tour as a function of water activity (w) and time (|)
life
Regression equation for Fig 4,31
Z = 2.24105 -  0.2568991 -  2.838473w + 0.016738t2+ 1.984741 w2 + 0.53058wt 
Z = FFA
w = water activity 
t = storage time
R2 =99%
N<ro-'
&
Time (weeks)
F ig  4 .3 1  The 3-D response surface for the free fatty acids (Z) in cooked-
d ried  soyflour as a function o f water activity (w) and time (t) <
U7
Regression equation for Fiu 4.32
Z = 1.889946 -0 .1717171 -3 .404782w  + 0.00698712 +2.920305w2 + 0.658826vvt 
Z = FFA
w = water activity 
t = storage time 
R2 =98%
e
<Uhu_
Time (weeks)
Fig 4.32 The 3-D response surface for the free fatty acids (Z) in roasted
soyflour as a function of water activity (w) and time (t).
ii(5
Table 4.12 Crude fat content in various forms o f freshly prepared soyflours as 
well as that stored for twelve (12) weeks at 30°C.
Soyflour type Moisture content (%) Crude fat (%) Storage loss
Fresh Stored Fresh Stored Crude fat 
loss (%)
Raw 9.78 9.98 20.16 ±0 .09 18.43 ±0.11 8.58
Roasted 5.51 5.71 20.56 ±0.11 20.07 ±0 .15 2.38
Cooked-dried 4.94 5.02 20.48 ±0 .15 19.91 ±0 .12 2.78
4.5 Sensory evaluation of soyflour samples
Sensory evaluation was performed on the samples at four (4) different times 
during storage. A panel, made up o f  sixteen (16) judges, was used and each was 
given a fresh and a stored sample o f  each o f  the three soyflour types to compare 
in terms o f flavour. The results are shown oil Tables 4.13 and 4.14. Whenever at 
least 50% o f panellists detected a difference in the flavour o f  the samples in each 
pair the results was considered significant.
0 . 9
According to the results shown on table 4.14, only 31.25% o f the panellists were 
able to detect changes in the flavour o f  the raw soyflour stored at 30°C. All the 
sixteen panellists could not detect any change in the flavour o f  all the other 
samples. By the end o f  the eighth week however, all the panellists could detect a 
difference between the fresh raw soyflour and the stored one at 30°C.
At the end o f  the twelve weeks storage period, 50% and 43.75% o f  the panellists 
detected a change in the flavours o f  cooked-dried and roasted soyflours, stored at 
30°C, respectively. This is an indication that, flavour changes in processed 
soyflour become detectable after storage for three months. This detection 
occurred at peroxide value and TBA Number o f  4.21 meq/kg and 9.76 mg/kg 
respectively in cooked-dried soyflour and 4.01 meq/kg and 9.57 mg/kg 
respectively in roasted soyflour.
Zademowski et al. (1983) reported that refined soybean oils in good condition 
have TBA Number in the range 0.02 to 0.08 mg/kg whereas crude oil or badly 
stored oils have TBA Number in the range 15 to 20 mg/kg. He also reported that 
the P.V. o f  fresh soybean oil is well below 10 meq/kg and that rancidity begins 
beyond 20 meq/kg. The disparity in this report and my findings could be 
explained as follows. The oil in the soybean flour is in the presence o f  non- lipidic 
substances such as proteins and polysaccharides. It was reported by Pokorny el al. 
(1983) that in the presence o f  protein there is rapid destruction o f hydroperoxides 
as some amino acids react with the peroxides. As a result the peroxide value 
measured in the soyflour appears to be lower than in the extracted oil. Pokorny et. 
al. (1983) also reported (hat the secondary oxidation products especially
I ',7.0
malondialdehyde react with proteins to produce compounds including some 
volatile compounds which do contribute to flavour changes. As a result the T BA 
lest does not measure all the TBA reactive substances produced from secondary 
oxidation hence the TBA Number measured appears 1o be lower than in the 
extracted oil. However since some o f the volatile products produced from the 
reaction o f  the TBA reactive substances with proteins affect flavour, changes can 
thus occur in flavour o f soyflour at a lower TBA Number than it will in the 
extracted oil.
Table 4.13 Sensory perceptions o f flavour changes in various types o f  soyflour 
stored at 5°C.
Length o f
storage
(weeks)
Judges’ Response (%)
Raw Cooked-dried Roasted
Y N Y N Y N
4 0 100 0 100 0 100
8 12.5 87.5 0 100 0 100
10 25 75 12.25 87.5 6.2.5 93.75
12 43.75 56.25 18.75 81.25 1875 81.25
Y = percentage o f  judges who could detect a difference in the flavours o f  freshly 
prepared and stored soyflours.
12.1
N = percentage o f  judges who could not detect any difference in the flavours o f  
freshly prepared and stored soyflours.
Table 4.14 Sensory perceptions o f flavour changes in \arious types o f  soy flour 
stored at 30°C.
Length o f 
storage (weeks)
Judges’ Response (%)
Raw Cooked-dried Roasted
Y N Y N Y N
4 31.25 .68.75 0 100 0 100
8 100 (s) 0 25 75 18.75 81.25
10 31.25 68.75 25 75
12 r 50 (s) 50 43.75 56.25
Y = percentage o f  judges who could detect a difference in the flavours o f  freshly 
prepared and stored soyflours.
N = percentage o f  judges who could not detect any difference in the flavours o f 
freshly prepared and stored soyflours.
(s) =Significant difference
I T i l
Flavour changes in all the samples stored at 5°C occurred very slowly. The raw 
flour however showed the highest rate o f  flavour change. As shown on Table 
4.13, 43.75% o f the panellists detected flavour change in (lie raw soyllour stored 
at 5°C after twelve weeks while only 18.75% o f the panellists detected flavour 
change in both the cooked-dried and the roasted Hours. These results indicate that 
raw soyflour has the tendency to undergo flavour changes faster than processed 
soyflours. It also shows that both moist (cooked-drying) and dry (roasting) heat 
processing do not have any significant difference in terms o f  slowing down the 
rate o f  o ff  flavour development in soyflour. Dry heat processing however 
appeared to have the most pronounced effect.
I Z Z
CHAPTER  FIVE 
CONCLUSION  AND RECOMMENDATION
5.1 CONCLUSION
1. Heat-treatment o f  soybeans prior to processing into flours yield flours with 
lower rates o f  chemical reactions, which result in minimal lipid oxidation 
and hydrolysis.
2. Cold storage o f  soyflour results in lower rates o f  chemical reactions, which 
cause increase in the stability o f  lipids in soyflour.
3. M aintaining the moisture content o f  soyflour at or close to the monolayer 
value o f  soyflour leads to lower rates o f  chemical reactions, which result in 
increased rate o f  lipid stability.
4. The shelf life o f  soyflour can be predicted when the peroxide value and the 
TBA Number o f  soyflour are known.
5. It is possible to extend the shelf life o f  soyflour by
1. Heat-processing the beans prior to milling into flour
2. Storing the soyflour under cold condition.
5.2 Recommendations
1. This studies has revealed that full-fat soyflour is liable to oxidative activities 
under tropical conditions therefore it is recommended that studies be 
conducted for the development o f  appropriate low-cost effective packaging 
for soyflour at ambient storage conditions.
U 4
2. This study lias shown that soybean is used in the fortification o f  some cassava 
products and maize flour in Ghana. It is therefore recommended that more 
studies be conducted on soybean-fortified foods so as to determine the effect 
o f  other ingredients on lipid stability in soyflour.
REFERENCES
Acker L.N. and Wiese R. (1972). Fd Techno/. 25, 1252. In: W ater relations o f 
food edited by R.B. Duckworth (1974), New York Academic press.
Alexander J.C. (1986). Heated and oxidized fats. In: Technological advances 
in improved and alternative sources o f  lipids edited by Kamel and 
Kakuda (1994). Blackie Academic Press and Professional, London.
Anderson, R.L., Rackis, J.J., and Tallent, W.H., (1970). In: soy protein and  
human nutrition edited by R.M. Lauk (1975). Academic Press, New 
York.
A.O.A.C. (1990). Association o f Official Analytical chemists, Official Methods 
o f  analysis. Arlington, V.A.
Ayernor, G.S. (1993). Utilization o f  legumes. Crop management guide 47. 
C.R.I., Ghana.
Baker, E.C. and Mustakas, G-C. (1972). Heat inactivation o f  trypsin inhibitor, 
lipoxigenase and urease in soybean; effect o f  acid and base additive. In: 
Expanding the use o f  soybeans. Proceedings o f  a conference fo r  Asia  
and Oceania. Thailand, Feb. 1976 edited by R.M. Goodman.
Badings, ITT. (1970). Cold storage defects in butter and their relation to the 
autoxidation o f unsaturated fatty acids. Milk dairy J. 24: 147-256. In: J. 
Fdsc 1996 vol. 61.40.3 pp.486-487.
Begin, M.E. (1987), Effects o f  polyunsaturated fatly acids and their oxidation 
products on cell survival. In: Technological advances in improved and
alternative sources o f  lipids edited by Kamel, B.S and 
Kakuda,Y.( 1994).BIackie academic and professional ,London.
Bressani, R. (1981). Soy protein products. In: Evaluation o f  proteins fo r  humans 
edited by B.J.F. Hudson (1983). Avi. Publishing co. Westport 
Connecticut.
Bressani, R., Murillo, B., and Elias L.G. (1974). Whole soybean as a means o f  
increasing protein and calories in maize based diets. In: Expanding the 
use o f  soybeans. Proceedings o f  a conference fo r  Asia and Oceania, 
Thailand Feb (1976) edited by R.M. Goodman.
Bressani, R. and Elias, L.G. (1983). Guidelines for the development o f
processed and packaged weaning foods. In: Effective post-harvest 
management o f  legumes and cereals in Ghana. Proceedings o f workshop, 
University o f  Ghana, Jan. 1989, edited by S. Sefa-Dedeh and R. Orraca- 
Tetteh.
Chow, C.K. and Gupta, M.K. (1994). Treatment, oxidation and health aspects o f  
fats and oils. In: technological advances in improved and alternative 
sources o f  lipids edited by Kamel, B.S and Kakuda,Y.(1994). Blackie 
academic and professional, London.
deMan, J.M .(1990). Principles o f  fo o d  chemistry. AVI Publishing Co. New 
York.
Erdman, J.W. and Forber, R.M. (1981), Soy protein foods In: Development in 
plantproteins-2, ppl 10-112, edited by B.J.F. Hudson (1983), Applied 
science Publishers, London.
Perrier, B.W. (1975). Heat processing o f  soybeans. In: Proceedings o f  a
conference for Asia and Oceania, Thailand Feb (1976) edited by R.M. 
Goodman.
Findlay, M.G., Draper, 11.11. and Bcrgan, J.G. (1970). Metabolism o f  1-4 C-
methyl linoleate hydroperoxides in the rabbit. In: Technological 
advances in improved and alternative sources o f  lipids edited by Kamel, 
B.S and Kakuda,Y.(1994). Blackie academic and professional, London.
Frankel, E.N. (1984). Lipid oxidation: mechanisms, products and biological 
significance. In: Technological advances in improved and alternative 
sources o f  lipids edited by Kamel, B.S and Kakuda, Y.(1994).Blackie 
academic and professional,London.
Grace, M.R. (1977). Cassava processing. (FAO plant production and protection 
series). FAO o f  the UN, Rome.
Gunstone, F.D. and Norris, F.A. (1983). Lipids in foods: Chemistry,
Biochemistry and Technology. Pp. 78-156. Pergamon Press, New York.
Halton, P. and Fisher, E.A. (1967). Cereal chem.14, 267. In: Water relations o f  
fo o d  edited by R.B. Duckworth (1974), N ew  York Academic press.
Hamilton, R.J. and Berger, K.G. (1995). Developments in oils and  fa ts. Lipids 
and oxygen: Is rancidity avoidable in practice? Blackie academic and 
professional, London.
Hammonds, T.M. and Call, D.L. (1970). Utilization o f  soy ingredients marketed  
in the U.S. fo o d  industry-part /.The current market for protein 
ingredients. A.E. Res. 320. Ithaca, N.V: Cornell
1.18
Heidelbaugh, N. and Karel, M. (1970). Effects o f  water binding agents on
oxidation o f methyl linoleate. In: Water activity; theory and application  
to fo o d  edited by L.B. Rockland and L. R. Beuchat. N.Y, Marcel Dekker.
Hoffman, R.S. (1962). Solubility o f  oxygen in oils. Developments in oils and
fats. In: Lipids and oxygen: Is rancidity avoidable in p racticel Edited by 
R.J. Hamilton and K.G. Berger (1975). Blackie academic and 
professional, London.
Horowitz, J. and Hartroft, W.S (1971). J. Nutr; 101.959. In: Fat sc ience -pa rt B  
edited by J. Hollo. Proceedings o f  the 16th ISF congress. Budapest, 
Hungary, Oct. 1983,
Hsieh, R.J. and Kinsella, J.E. (1989). Oxidation o f  polyunsaturated fatty acids: 
mechanisms, products and inhibition with emphasis on fish. Adv. Food 
Nutr. Res. 33,233-241. In: Technological advances in improved and  
alternative sources o f  lipids edited by Kamel, B.S and 
Kakuda,Y.(1994).Blackie academic and professional,London.
Huang, C.J., Cheung, N.S. and Lu, V.R. (1988). Effect o f  deteriorated frying oil 
and dietaiy protein level on liver microsomal enzymes in rat. In: 
Technological advances in improved and alternative sources o f  lipids 
edited by Kamel, B.S and Kakuda,Y.(1994).Blackie academic and 
professional,London.
Ihekoronye, A.I. and Ngoddy, P.O. (1985). Integrated Food science and  
technology fo r  the tropics. Macmillan Publishers, London.
Jirousova, J. and Davidek, J. (.1975). Health effects o f  lipid oxidation. In: Fat
I3L-9
science p a r t B  edited by J. Hollo. Proceeding; o f  the 16th ISF congress. 
Budapest* Hungary, Oct. 1983,
Kakade, M X ., Simons, N.R., Liener, I.E. and Lambert, J.W. (1972). J. Agric. 
Food Chem, In: Development in plant proteins-2, pp) 10-112, edited by 
B.J.F. Hudson (1983), Applied science Publishers, London.
Kamel, B.S. and Kakuda, Y. (1994). Technological advances in improved and  
alternative sources o f  lipids. Blackie Academic and Professional, 
London.
Kanazawa, K., Kanazawa, E. and Natake, M. (1985). Uptake o f  secondary
autoxidation products o f  lionoleic acid by rat. In: Technological advances 
in improved and alternative sources o f  lipids edited by Kamel, B.S and 
Kakuda,Y.(1994).Blackie academic and professional,London.
Keeney, B.J. (1962). Products o f  lipid oxidation. In: Principles o f  food
chemistry edited by R.S. deMan (1990), AVI Publishing Co. New York.
Ken- Jones S. and Amos A.J. (1967). Modern cereal chemistry. Pp 612, Avi 
Publishing co. Westport Connecticut.
Kordylas, J.M. (1991). Processing and preservation o f  tropical and subtropical 
foods. Macmillan education Ltd. London.
LabuzaT.P. (1975). Oxidative changes in food at low and intermediate moisture 
level. In: Water activity; theory and application to food  edited by L.B. 
Rockland and L. R. Beuchat. N. Y. Marcel Dekkcr.
Lang, K.W., McCune, T.D. and Steinberg, M.P. (1981) Proximity Equilibration 
Cell for rapid determination o f  isotherm. Food sc. 46:936-938.
Lin, Humber, E.S., Sosulski, F.W. and Cord, J.W. (1975). Soybean as a
food source. In: Development in p lan t proteins-2, p p l  10-112, edited by 
B.J.F. Hudson (1983), Applied science Publishers, London.
Markley, K.S. (1951). Soybeans and soybean products-vol.2. Inter Sciencc 
Publishers Inc. New York.
Martz, S., McWilliams, R. Larsen, J.H., McMullen, M. and Layman, B. (1965). 
Fd Technol. 9, 276. In; Water relations o f  fo o d  edited by R.B. 
Duckworth (1974), New York Academic press.
Meyer, E.W. (1970). Soybean flours and grits. Proc. SOS 170. 3rd Intern. Congr. 
FoodSc. Technol. Inst. Food Tech. Chicago. In: Encyclopaedia o f  fo o d  
Technol. Edited by A.H. Johnson and M.S. Peterson (1974). AVI 
Publishers Westport Connecticut.
Mercer-Quarshie and Nsowah (1975). Soybean consumption in Ghana. Food  
Research Institute Newsletter, vol. 15, No. 1, April 1988.
Nawar, W.F. (1984). Chemical changes in lipids produced by thermal
processing. J. Chem. Educ. 61, 299-302. In: Technological advances in 
improved and alternative sources o f  lipids edited by Kamel, B.S and 
Kakuda,Y.(1994). Blackie academic and professional, London.
Oarada, M., Miyazawa, T. and Kaneda, T. (1986). Distribution o f  C-14 after oral 
adminitration o f  U-C-14 leballed methyllinoleate hydroperoxides and 
their secondary products in rats. Lipids 21,150-154. In: Technological 
advances in improved and alternative sources o f  lipids edited by Kamel, 
B.S and Knkuda,Y,( 1994), Blackie academic and professional, London.
Orraca-Tetteh, R. (1972). Evaluation o f the protein value of traditional weaning 
ioods and children’s diets in Ghana. Ghana J. Sc. 12:73. In: Effective  
post-harvest management o f  legumes and cereals in Ghana pp. 103. 
Proceedings o f  workshop, University o f  Ghana, Jan. 1989, edited by S. 
Sefa-Dedeh and R. Orraca-Tetteh.
Pokorny, J. Novotna-Moravkia, Alexova, H. and Davidek J. (1983). Oxidation 
o f  polyunsaturated lipids in presence o f  non-lipidic components. In: Fat 
science -part. B  edited by J. Hollo. Proceedings o f the 16th ISF congress. 
Budapest, Hungary, Oct. 1983,
Poling, C.E, Warner, W.D., Mone, P.E. and Rice, E.E. (1970). The nutritional 
value o f  fats after use in commercial deep fat frying. J. Nutr. 72, 109- 
120. In: Technological advances in improved and alternative sources o f  
lipids edited by Kamel, B.S and Kakuda,Y.(1994). Blackie academic and 
professional, London.
Rackis, J.J., Sessa, D.J. and Honig, D.H. (1979). JAOCS 56, 262. In: Fat. 
science -P a rt A edited by J. Hollo. Proceedings o f  the 16th ISF congress. 
Budapest, Hungary, Oct. 1983,
Robinson, R.F. (1971). Plant proteins: their advantages as extenders in meat 
products. Fd technol. April 1995, pp l23 .
Ronald S.K. and Ronald S. (1991). Pearson's Composition and  Analysis o f  
foods. 9lh edition. Longman Scientific and Technical Press, U.K.
Salwin, H. (1959). Defining minimumn moisture contents for dehydrated foods. 
I'ood Chetn. 18:71. In: Water activity; theory and application to food
edited by L.B. Rockland andL . R. Beuchat. N.V. Marcel Dekker.
Satouchi, K. and Matsushita, S. (1976). Purification and properties o f  a lipase- 
inhibiting protein from soybean cotyledons. Agiicol. Biol. Chem, 
40,889-897. In: Enzymes in fo o d  processing  edifed by G.A. Tucker and 
L.F.J. Woods (1991). Thompson Press (India) Ltd, New Delhi.
Scott, W.O. and Aldrich, S.R. (1970). Modern soybean products. Pp 165-167, 
Avi Publishers Westport Connecticut.
Segall, H.J., Wilson, D.W., Dallas, J.L. and Haddon, W F. (1985). Trans-4-
hydroxy-2-hexenal: a reactive metabolite from macrocyclic pyrrolizidine 
alkaloid senecionine. Sc. 229, 472-475. In: Technological advances in 
improved and alternative sources o f  lipids edited by Kamel, B.S and 
Kakuda,Y.(l 994). Blackie academic and professional, London.
Smith, A.K. and Circle, S.J. (1972). Functional properties o f  commercial edible 
soybean products. Avi. Publishers. Westport Connecticut. In: Expanding  
the use o f  soybeans. Proceedings o f  a conference fo r  Asia  and Oceania, 
Thailand Feb (1976) edited by R.M. Goodman.
Speiss, W.E.L. and Wolf, W. (1987). Critical evaluation o f  methods to
determine moisture sorption isotherms. In: Water activity; theory and  
application to fo o d  edited by L.B. Rockland and L.R. Beuchat. N.Y. 
Marcel Dekker.
Struthers, B.J., Dahlgren, R.R. Hopkins, D.T. and Raymond, M.L. (1979). Soy 
protein and Human nutrition. In: Developments in proteins-2  edited by
B.J.F. Hudson (1983). Applied science Publisher;, New York.
Taylor, A.J. (1987). Effects o f water quality on lipid oxidation. Fd sc. And
Technol. 1'odoy, 1, 158-159. In: Technological advances in improved  
and alternative sources o f  lipids edited by Kamel, B.S and 
Kakuda, Y.( 1994).
Blackie academic and professional, London.
Thio, G.L. (1975). Small scale and home processing o f  soybeans with
application and recipes. Royal tropical inst., Amsterdam. In: Expanding  
the use o f  soybeans. Proceedings o f  a conference fo r  Asia and  Oceania, 
Thailand Feb (1976) edited by R.M. Goodman.
Waggle, D.H., Decker, C.D. and Kolar, C.W. (1981). Soy Protein Products.
J.AOCS. In: Developments inproteins-2  edited by B.J.F. Hudson (1983) 
pp. 103.Applied science Publishers, New York.
Wattenberg, L.W. (1972). Inhibition o f  carcinogenic and toxic effects o f
polycyclic hydrocarbons by phenolic antioxidants ethoxy-quin. J. Natl, 
cancer inst. 48,1425-1430. In: Technological advances in improved and  
alternative sources o f  lipids edited by Kamel, B.S and Kakuda,Y.(1994). 
Blackie academic and professional, London.
Whitaker, JR . (1972). Principles o f  enzymology fo r  the fo o d  sciences. Pp. 483- 
499. Marcel Dekker Inc. New York.
Williams, L.D. (1970). Soy proteins. In: Symposium on vegetable proteins. The 
British Manufacturing Industries Res. Association, London.
Wolf, J.W. (1970). Soybean proteins: their functional, physical and chemical 
properties. In: Food Technol. April 1985, pp l20 .
Zadernovvski, R., Kozlowska, H. and Sosulski, F.W. (9183). Composition o f
bound lipids in flours obtained from oleaginous seeds. In: Fat science  -  
par/ A edited by J. Hollo. Proceedings o f  the 16h ISF congress.
Budapest, Hungary, Oct. 1983,
APPENDICES
APPENDIX  1: NUTRITION AND FOOD SCIENCE DEPARTMENT 
UNIVERSI TY OF GHANA, LEGON 
MPIIIL PROJECT.
SENSORY EVALUATION OF SOYFLOUR SAMPLES
Date............................... Panelist number............................
You are provided with three (3) pairs o f soyflour samples. Please taste and smell 
each pair (one at a time) and record whether the samples in each pair are different 
in flavour or not.
Note: where difference exists between a pair o f samples indicate it by writing ‘Y \  
where no difference is detected indicate it by writing ‘N ’ at the appropriate place.
Sample Pair Flavour
1   ..............
2   --------
3 .............  ...........
Comment....................................................................................................................................
136
APPENDIX2: Analysis o f  variance for the full regression for FFA in raw
soyflour as a function o f temperature and time
Source
of
variation
Sum of 
squares
D.F Mean square F-ralio P-value
Model 7.7162 5 1.5433 16.9767 0.0004
Error 0.7272 8 0.0909
Total 
(co rr)
8.4435 13
Lack o f 
(It
0.4282 3 0.1427 2.3
Pure
error
0.2990 5 0.0598
APPENDIX3: Analysis o f variance for the full regression for FFA in cooked-
dried soyflour as a function o f temperature and time
Source
o f
variation
Sum of 
squares
D.F Mean square F-ratio P-value
Model 1.2666 5 0.2533 48.0931 0.0001
Error 0.0421 8 0.0053
Total
(corr)
1.3087 13
Lack of 
lit
0.0293 3 0.0098 3.78
Pure
error
0.0129 5 0.0026
137
APPENDIX4: Analysis o f variance for the full regression for FFA in roasted  
Soyflour as a function o f  temperature and time
Source
o f
variation
Sum o f 
squares
D.F Mean square F-ratio P-value
Model 1.6231 5 0.3246 33.5679 0.0001
Error 0.0774 8 0.0097
Total
(corr)
1.7004 13
Lack o f 
fit
0.06 3 0.02 5.78
Pure
error
0.0173 5 0.0034
APPENDIX5: Analysis o f  variance for the full regression for P.V in raw
soyflour as a function o f  temperature and time
Source
of
variation
Sum o f 
squares
D.F Mean square F-ratio P-value
Model 267.741 5 53.5482 39.9118 0.0001
Error 10.7333 8 1.3416
Total
(corr)
278.474 13
Lack o f 
fit
6.8333 3 2.2778 2.9
Pure
error
3.9 5 0.78
138
APPENDIX6: Analysis of  variance for the full regression lor P.V in cooked-
dried soyllour as a function o f temperature and time
Source
of
variation
Sum of 
squares
D.F Mean square F-ratio P-value
Model 16.3158 5 3.2632 144.1.97 0.0001
Error 0.1810 •8 0.0226
Total
(corr)
16.4969 13
Lack o f 
fit
0.1732 3 0.05776 37
Pure
error
0.0078 5 0.0016
APPENDIX7: Analysis o f variance for the full regression for P.V in roasted
soyflour as a function o f temperature and time
Source
o f
variation
Sum of 
squares
D.F Mean square F-ratio P-value
Model 23.5544 5 4.7109 18.9013 0.0003
Error 1.9939 8 0.2492
Total
(corr)
25.5482 13
Lack o f
fit
1.4439 3 0.4813 4.8
Pure
error
0.5 5 0.1
n o
APPENDIX8: Analysis o f  variance for the full regression for I BA in raw
soyflour as a function o f  temperature and time
Source
o f
variation
Sum of 
squares
D.F Mean square F-ratio P-value
Model 51.3228 5 10.2646 235.805 0.0001
Error 0.3082 8 0.0435
Total
(corr)
51.6711 13
Lack o f 
fit
0.3438 3 0.1146 127.33
Pure
error
0.0045 5 0.0009
APPENDIX9: Analysis o f  variance for the full regression for TBA in cooked-
dricd soyflour as a function o f  temperature and time
Source
o f
variation
Sum o f 
squares
D.F Mean square F-ratio P-value
Model 14.2961 5 2.8592 280.274 0.0001
Error 0.08161 8 0.01020
Total
(corr)
14.3777 13
Lack o f 
fit
0.0691 3 0.0230 9.208
Pure
error
0.0125 5 0.0025
APPENDIX 10: Analysis of variance for the full regression for TBA  in roasted
soyllour as a function o f  temperature and time
Source
of
variation
Sum of' 
squares
D.F Mean square F-ratio P-value
Model 12.2189 5 2.4438 59.0055 0.0001
Error 0.3310 8 0.0414
Total
(corr)
12.5499 13
Lack o f 
fit
0.2471 3 0.0824 4.9
Pure
error
0.084 5 0.0168
APPENDIX1I: Analysis o f  variance for the full regression for FFA in raw  
soyflour as a function o f water activity and time
Source
of
variation
Sum of 
squares
D.F Mean square F-ratio P-value
Model 38.4550 5 7.6910 355.870 0.0001
Error 0.1729 8 0.0216
Total
(corr)
38.6279 13
Lack of 
lit
1.629 3 0.0543 27
Pure
error
0.0099 5 0.0019
APPENDIXI2: Analysis of variance lor the lull regression for FFA in cooked-
dried soyflour as a function ol water activity and time
Source
o f
variation
Sum o f 
squares
D.F Mean square F-ratio P-value
Model 7.9576 5 1.5915 463.569 0.0001
Error 0.0275 8 0.0034
Total
(corr)
7.9851 13
Lack o f 
lit
0.0211 „■> 0.0070 5.48
Pure
error
0.0064 5 0.0013
APPENDIX13: Analysis o f  variance for the full regression for FFA in roasted
soyflour as a function o f  water activity and time
Source
o f
variation
Sum of 
squares
D.F Mean square F-ratio P-value
Model 1 1.5539 5 2.3108 117.529 0.0001
Error 0.1573 8 0.0197
Total
(corr)
11.7112 13
Lack o f 
lit
0.1365 3 0.0455 10.9351
Pure
error
0.0208 5 0.0042
APPENDIX14: Analysis o f  variance for the full regression for P.V in raw
soyflour as a function o f water activity and time
Source
o f
variation
Sum of 
squares
D.F Misitt square F-ratio P-value
Model 61.4134 5 12.2827 46.4994 0.0001
Error 2.1132 8 0.2641
Total
(corr)
63.5265 13
Lack o f 
fit
1.5032 3 0.5011 4.1
Pure
error
0.61 5 0.122
APPENDIX15: Analysis o f  variance for the full regression for P.V in cooked-
dried soyflour as a function o f water activity and time
Source
o f
variation
Sum o f 
squares
D.F Mean square F-ratio P-value
Model 3.2732 5 0.6546 33.9167 0.0001
Error 0.1544 8 0.0193
Total
(corr)
3.4276 13
Lack of 
lit
0.1271 3 0.0424 7.8
Pure
error
0.0273 5 0.0055
143
APPENDIX16: Analysis o f  variance for the full regression lor P.V in roasted
soyflour as a function o f water activity and time
Source
o f
variation
Sum o f 
squares
D.F Mean square F-ratio P-value
Model 8.9545 5 1.7909 20.7366 0.0002
Error 0.6909 8 0.0864
Total
(corr)
9.6454 13
Lack o f 
fit
0.5316 3 0.1772 5.56
Pure
error
0.1593 5 0.0318
APPENDIX17: Analysis o f variance for the full regression for TBA in raw
soyflour as a function o f water activity and time
Source
o f
variation
Sum o f 
squares
D.F Mean square F-ratio P-value
Model 52.8308 5 10.5662 19.6279 0.0003
Error 4.3066 8 0.5383
Total
(corr)
57.1374 13
Lack o f 
fit
2.6224 3 0.8741 2.59
Pure
error
1.6842 5 0.3368
APPENDIX18: Analysis o f  variance for the full regression for TBA in cooked-
dried soyflour as a function o fw ater activity and time
Source
o f
variation
Sum of 
squares
D.F Mean square F-ratio P-value
Model 17.0330 5 3.4066 112.259 0.0003
Error 0.2427 8 0.0303
Total
(corr)
17.2758 13
Lack o f 
fit
0.2374 3 0.0791 73.26
Pure
error
0.0054 5 0.0011
APPENDIX19: Analysis o f variance for the full regression for TBA in roasted
soyllour as a function o f water activity and time
Source
o f
variation
Sum o f 
squares
D.F Mean square F-ratio P-value
Model 19.5235 5 3.9047 63.264 0.0003
Error 0.4938 8 0.0617
Total
(corr)
20.0173 13
Lack o f 
fit
0.4303 J 0.1434 11.3
Pure
error
0.0635 5 0.0127
_
145