LéPED OXEDATEON 0F MILK FAT 'N FREEZE-DRlED CELLULOSE MODEL SYSTEMS 0F VARYING WATER ACTIVITiES ‘.: ‘.~ . 923:.’-Q {‘5 .'- 3%.3333 5U; Luv au&.3u v. “I :3. {mat-7min? Q".»3":’tT-‘ Fiatri.ff::Q-'*'v ith‘v.;IOJ‘ .‘i U .i.01_ u“:~‘-‘:‘“ I AQ'AT' ? F“? Mar”: 4 «f» Ul’lisd Hulddh it“=u $679 in vi 3 w.) m #. U w Y. Rm n M Am B...‘ I Lm L fl/ifl/I E W]: W 293 mg m /7/ ABSTRACT LIPID OXIDATION OF MILK FAT IN FREEZE-DRIED CELLULOSE MODEL SYSTEMS OF VARYING WATER ACTIVITIES BY Sang Choul Han The organoleptic deterioration of milk fat components in dehydrated food systems has been a serious problem from the standpoint of consumer acceptance. The effects of several different water activities on autoxidation of milk fat in a freeze—dried microcrystalline cellulose model system were investigated using a peroxide test, a 2-thiobarbituric acid test and flavor panel evalua- tion. The results of autoxidation of milk fat were compared with the calculated BET monolayer value. The correlation of the Z-thiobarbituric acid test to the other tests are discussed. Three different water activities, aw 0.12, aw 0.47 and aw 0.75 of freeze-dried milk fat - microcrystalline cellu— lose model systems were prepared for these studies. The calculated BET monolayer value of the model system was aW 0.23 and 0.37% of moisture content, dry basis. Sang Choul Han In comparison of results of TBA to peroxide value and sen- sory evaluation during 2 months storage periods at 20‘: 0.5°C, the TBA test showed a poor correlation with the peroxide test and the sensory evaluation; the peroxide test proved to be the most applicable test for the autoxidation studies of the freeze-dried model systems for long term storage stability. The effects of varying water activities on the autoxi- dation studies demonstrated that the model system aw 0.47 (above BET monolayer value) demonstrated a protective effect, while the model system, aw 0.12 and aw 0.75 (below BET monolayer value and far above BET value) showed a pro— oxidant effects. Possible reasons are suggested for the poor correlation of the TBA test results with those of the peroxide test and the sensory evaluation. LIPID OXIDATION OF MILK FAT IN FREEZE-DRIED CELLULOSE MODEL SYSTEMS OF VARYING WATER ACTIVITIES BY Sang Choul Han A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Food Science 1976 ACKNOWLEDGMENTS The author wishes to express his appreciation to his major professor, Dr. C. M. Stine for his generous assistance in all phases of this study. His conscientious guidance and constant availability for discussion were largely responsi- ble for the success of this work. Sincere appreciation is also extended to Dr. R. C. Nicoholas and Dr. H. A. Lillevik for serving on the author's guidance committee. The author also wishes to thank his parents, Mr. and Mrs. H. S. Han, his mother-in—law, S. H. Cho, and his wife, Aikyung, for their sacrifices, understanding and constant inspiration for higher education. ii TABLE OF CONTENTS Page LIST OF TABLESOOOOOOOOOOOOOOO...OOOOOOOOOOOOOOOOOOOOO Vi LIST OF FIGURESOOOOOOOOOOOO0.000000000000000000000000 Viii INTRODUCTIONOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 1 LITERATURE REVIEWOOOOOOOOO...OOOOOOOOOOOOOOOOOOOOOOOO 4 Water Activity and Lipid Oxidation................ 4 Definition of water Activity................... 4 Moisture Adsorption Isotherm................... 5 Theoretical Description of Isotherms........... 6 The Storage Stability of Food Products and water ActiVitYOOOOOOOOOOOOOOOOOOCOOOOOOOOOOO 6 Hypotheses Explaining the Protective Effect of waterOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 9 Oxidation Mechanism of Milk Lipids................ 10 Theoretical Aspects of Autoxidations........... 10 Dismutation of Hydroperoxides and Off-flavor PrOduCtionOOOOOOO...OOOOOOOOOOOOOOOOOOOOOOO. 11 The Composition of Milk Lipid and Its PropertieSOOOOOOOOOCOOOOOOOOOOOOOOOOOOOOOOOO 12 Autoxidation of Saturated Fatty Acid of Milk FatOCOOOOOOOOOOOOOOOOOCOOOOOOOOOOOOOOOOOOOOO l3 Autoxidation of Oleic Acid..................... 13 Oxidative Deterioration of Dairy Products...... 16 Chemical Tests for Lipid Oxidation and Correlation to sensory EvaluationOOOOOOOOOOOOOOOOOOIOOOOOOO 16 2-Thiobarbituric Acid Test for Measurement of OXidative RanCiditYOOOOOOOOOOOOOOOCOOOOOCOOO 16 The PrinCipal TBA ReaCtantSooooooo0000000000coo 17 iii TABLE OF CONTENTS—~continued TBA Reactive Material and Water Extraction..... The Correlation Between TBA-Reactive Materials and Rancid Odor of Food Products............ EXPERIMENTALOOOOOOOOC0.0.0.000...OOOOOIOOOOOOOOOOOOOO separation Of Milk FatOOOOOOIOOOOOOOOOOOOOOOOOOOOO Preparation of Freeze-dried Model Systems......... Adjusting Water Activity by Moisture Equilibration Determination of Water Activity................... Calculation of Water Activity..................... Chemical Test for Lipid Oxidation................. PerOXide TeStOOOO00.00.0000...OOOOOOOOOOOOOOOOO 2-Thi0barbituric ACid TeStOOOOOOOOOOOOOOOOOOOOO Chromatographic Separation of Pink TBA Pigment. RESULTS AND DISCUSSION.0.000.000.00000000000000000000 Preparation of Model Systems...................... Moisture Equilibration of the Milk Fat Model SYStemSooooooooooooo0.0000000000.000.000.000... Adsorption Moisture Isotherm Curve of the Model systemSOOOOO0....0.0.0.0...OOOOOOOOOOOOIIOOOOOO The Change in Moisture Content of the Milk Fat Model Systems During Autoxidation Periods...... sensory EvaluationOOOOOOOOOIOOOOOOOOOO00.0.0000... Effects of Varying Water Activities on Autoxida- tion of Milk Fat in Freeze-dried Model Systems. The Effect of water Activity on the Rate of Milk Fat OXidationOOOO0.0.0.0....OOOOOOOOOOOOOOOOOIO Z—Thiobarbituric Acid Test in Freeze—dried Milk Fat OXidationOOO...0.....OOCOOOOOOOOOOCOCCOOOQO Effects of Fatty Acid Component of Purified Milk Fat on TBA TeStOOIOOOOOOOOOOOOOOOOOOOOOOOOOOOO. iv Page 18 19 21 21 21 22 22 26 26 26 27 28 29 29 30 35 40 40 42 52 55 56 TABLE OF CONTENTS--continued Page The Theoretical Reasons for Poor Correlation Between the TBA and Peroxide Test or Sensory EvaluationOOOO00.00....OOOOOOCIOOOIOOOOOOOOOOOO 60 TBA Color Production and Decomposed Products of the OXidized FatOOOOOOOOOOOOOOO0.00.00.00.00 60 The Effects of Chemical Reactivity of Unsatur- ated Fatty Acid in Milk Fat and TBA Reagent. 64 The Effects of Physical Structure of Freeze- dried Model Systems on the Accumulation of TBA Reactive Materials...................... 65 SUWARY MD CONCLUS ION O O O O C O O O O O O O I O O O O O O O O O O O O O O O O O O 6 8 BIBLIWMPHYOOCOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 72 LIST OF TABLES TABLE 8. 9. 10. ll. 12. 13. The Composition of Milk Lipid................... The Fatty Acid Composition of Milk Fat.......... Comparison of Unsaturated Fatty Acid in Mole Percent of Milk Fat (Total Unsaturated Fatty ACid components)O0......OOOOOOOOOOOOOOOOOOOOCOOO Relative Humidities of Saturated Salt Solutions at Given Temperature............................ Saturated Salt Solutions and Equilibrium Mois— ture contentOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOIOO Change of Moisture Content of the Model System During Attainment of Equilibrium Moisture....... Change in water Activity of the Freeze-dried Model System at Varying Moisture Contents....... The Value for BET Monomolecular Layer Plot...... Flavor Scores of Freeze—dried Model Systems..... TBA Values and Peroxide Values of Butter Fat Stored for Approximately 2 Years at -20, 0, 72 and 100 0F...0.0.0....00....OOOOOOOOOOOOOOOOOCI. Peroxide Value of Milk Fat Model System Runs 1, 2 and 3 During Autoxidation Periods............. The Effects of Water Activity on Autoxidation Rate of Milk Fat in Freeze—dried Cellulose Model Systems at 20 :_0.5°C (MEQ O/KG FAT per day).... TBA Color Production at 532 nm of Autoxidized Milk Fat in Freeze-dried Cellulose Model Systems at Various Ranges of water Activity............. vi Page 12 14 15 23 31 32 35 38 43 47 48 53 58 LIST OF TABLES—-continued TABLE Page 14. Number and Type of Isomeric Peroxide Radicals Formed in Early Stage of Autoxidation of Methylene Interrupted Polyene................... 61 15. Absorbance of Products of TBA Reaction with Different Types of Aldehydes per Mole of AldehYdeOOOOOOO0.0.0.0000...00......0.000.000... 63 vii LIST OF FIGURES FIGURE Page 1. Moisture isotherm curve of food products....... 7 2. Simplified Makower-Meyer/Taylor apparatus...... 24 3. Moisture equilibrium curve of a milk fat- cellulose model system at 20 :_0.7°C (Run 1)... 33 4. Moisture equilibrium curve of a milk fat- cellulose model system at 20 :_0.7°C (Runs 2 and 3)0.0COOOOOOOOCOOO00......OIOOOOOOOOCOOOOOO 34 5. Moisture adsorption isotherm of the freeze- dried model system at 20 : 0.2°C............... 36 6. The BET monolayer plot of the freeze—dried model system at 20 :_0.2°C..................... 39 7-1. Sensory evaluation of freeze-dried model sys— tems during storage at 20° C and aw of 0.12, 0.47 and0.75..COO...0.0000000000000000000ZOOOO 44 7-2. Sensory evaluation of freeze-dried model sys- tems during storage at 20 :_°C and aw of 0.12, 0.47 and 0.75.00.00.000000000000000000000000000 45 7-3. Sensory evaluation of freeze—dried model sys— tems during storage at 20°C and aw of 0.12, 0.47 and 0.750.000.0000...OOOOOOOOOOOOOOOOOOOOO 46 8. Lipid oxidation in freeze-dried model systems in varying water activities at 20 :_0.5°C (Run 1)000......0..0.0...OOOOOOOOOOOOOOOOOOCOOO 49 9. Lipid oxidation in freeze-dried model systems in varying water activities at 20 :_0.5°C (Run 2)...C...CCOOOOOOOOOOOOOOOOOOO0.000.000... 50 viii LIST OF FIGURES--continued FIGURE 10. 11. 12. 13. 14. Lipid oxidation in freeze-dried model systems in varying water activities at 20 :_0.5°C (Run 3).....................‘................... The effects of various water activities on milk fat oxidation in the model systems at 20;: OOSOCOOOOO...00.0.00...OOOOOOOOOOOOOOOOOOOOOOOO TBA color production of oxidized milk fat in cellulose model systems at various water aCtiVitieSOOOOOO00.0.00...OOOOOOOOOOOOOOOOOOOOO Relationship between peroxide value and TBA value in the oxidation of poly—unsaturated acids. (Source: Holman et al., l962)......... The accumulation of TBA reactive materials and effects of water activity in freeze—dried milk fat mOdel SYStemSooooooooooooooooooo00000000000 ix Page 51 54 59 62 66 IN TRODUCT ION The milk of animals was used for human food before the dawn of history. Milk has long been referred to as one of the most nearly perfect foods. In the last half of this century the dairy industry has made remarkable progress in the technology of fluid and dehydrated foods. In recent years, research related to dry by-products has intensified, partially as a result of declining per-capita consumption of fluid milk in developed nations. The milk lipids are recognized as a major factor in determining economic value as well as consumer acceptability of most dairy products. Many dehydrated dairy products have been produced: nonfat dry milk, dry cream, dry whole milk, ice cream.mix, dehydrated butter, dehydrated cheese and coffee whiteners. These products offer great promise since their versatility, novelty and convenience of use have inherent consumer appeal. However, their mediocre storage stability has been a major obstacle to wider acceptance. Initially, in the preparation of dehydrated foods, the objective was to produce a product as dry as possible. But for most food, it was found that for optimum stability a small amount of water must be left in the products. The presence of a small amount of water content appears to be an extremely important factor in preventing or inhibit- ing fat oxidation. Much research on lipid oxidation in various freeze— dried model systems has been done recently. Most of these studies have been based upon the method of oxygen uptake which uses rather purified unsaturated fatty acid in metal catalyzed conditions. However, the method of oxygen uptake provides good kinetic results in oxidation studies of lipids, but in complex food systems such as dairy products which contain large amounts of a low peroxide fatty acid constituent, it is not very applicable for the practical routine examinations for a long term storage stability. This is because the errors caused by the empirical oxygen uptake method are often greater than those detectable by chemical or sensory evaluation of oxidation of milk fat. For practical purposes, the peroxide test, the 2—thio— barbituric acid test and sensory evaluation have been widely used in studies of organoleptic deterioration of high moisture content food systems. But in the case of dehyé drated food systems which contained low peroxide fatty acid, the correlation of the TBA test to other tests such as the peroxide test and sensory evaluation have not been fully investigated. The purpose of this research was to determine the effects of water activity on milk fat oxidation in a model system and also to find the correlation of TBA to other chemical and sensory tests. LITERATURE REVIEW water Activity and Lipid Oxidation One of the major problems in research on dehydrated food is the prevention of deleterious change in the products during storage. Removing water from a food by freeze dehy- dration results in a product that has a porous, sponge-like matrix which permits ready access of oxygen to the compon- ents of the food, thereby facilitating oxidation changes (Maloney et al., 1966). Historically, in the preparation of dehydrated food, the objective was to produce a product as dry as possible. But for most foods (the exception being those high in sugar), it was found that a small amount of water must be left in the product. Without a residue of water, good stability generally cannot be attained. Moreover, water activity appears to be an extremely important factor in preventing or inhibiting lipid oxidation (Rockland, 1957; Salwin, 1963). ‘ Definition of Water Activity The availability of water for spore germination, micro— bial growth and chemical deterioration of food is closely related to its relative vapor pressure (Scott, 1957), com- monly designated as water activity. Water activity (aw) is defined as the ratio of the vapor pressure (P) of water in the food to the vapor pressure of water (PS) at the same temperature (Brockman, 1970). It can also be approximated as the mole fraction of water; that is, the moles of water divided by the sum of the moles of water and the moles of solute. where P # water vapor pressure exerted by the food PS= vapor pressure of pure water at temperature T T = equilibrium temperature of systems Nw= moles of water Ns= sum of mole fraction of all soluble constituents. Moisture Adsorption Isotherm Moisture adsorption isotherms have been used to calcu— late the Brunauer, Emmet and Teller (BET) monomolecular layer value, and they have been used to predict the optimum moisture content of numerous food (Rockland, 1969; Salwin, 1963). The adsorption isotherm of a food is best described as a plot of the amount of water adsorbed as a function of the relative humidity or activity of the vapor space surrounding the material all at a constant temperature. Many investi- gators (Taylor, 1961; Stitt, 1958; Rockland, 1960) have described the procedures for obtaining water vapor isotherms for food. One general method is that the dehydrated food material is placed in a vacuum desiccator containing a specific saturated salt solution known to provide a definite equilib- rium relative humidity necessary for determination of the isotherms. The other method involves measurement of the vapor pressure of water in equilibrium with a food at given mois— ture content and temperature, using a sensitive manometric system or electric—hygrometer. Theoretical Description of Isotherms According to the theory of Brunauer et al. (1938), water is bound (strongly adsorbed) in a monomolecular layer in the region of the first slope of a moisture adsorption isotherm, up to point A from Figure 1. Above point A, in the linear section, bi- or multimolecular adsorption occurs. From about B on, water is condensed in capillaries with increasing water activity (Acker, 1969). The Storage Stability of Food Products and7water Activity, The effect of water and water activity on lipid oxida- tion in foods has been studied in model systems by several MOISTURE, PER CENT OF SOLIDS Figure l. Moisture l I 1 L IO I5 20 25 RELATIVE HUMIDITY isotherm.curve of food products. investigators (Heidelbaugh et al., 1970; Labuza et al., 1966; Chou et al., 1973; Heidelbaugh et al., 1971). In general it was found that at very low water activities, lipid oxidation was relatively rapid. Addition of water up to a critical level reduced the rate of lipid oxidation. Salwin (1959, 1963) conjectured that the monolayer moisture acts as a film, protecting the food from attack by oxygen. Exceptions to this concept appear to be mainly food of high carbohydrate content. Maloney and Karel (1966) reported that water had an inhibitory effect on the oxidation of a freeze-dried model system consisting of micro-crystalline cellulose and methyl linoleate, varying with.water activity up to 0.5, and then at intermediate moisture levels, presumably aw 0.5 to aw 0.75, the lipid oxidation of model systems becomes acceler— ate (Labuza et al., 1969, 1970; Heidelbaugh et al., 1971). A series of related studies on the intermediate moisture range were carried out in an attempt to determine the rate of lipid oxidation. Labuza and Chou (1973) observed that in‘a non-swellable cellulose system at low metal content, the rate of oxidation increased as the aw increased into the range of moisture that describes intermediate moisture foods. Labuza et a1. (1970), concluded that with high water activity, where the water is not bound tightly, the moisture content of the system has a strong effect on control of oxidation rate. At an extremely high trace metal content, 500 ppm, the oxidation rate pattern was completely reversed with the rate of oxidation being faster at the lower water content and lower water activity. Hypotheses Explaining the Protective Effect ofiwater Several hypotheses have been suggested to explain the protective effect of water in retarding lipid oxidation. The most important are: 1. That water has a protective effect due to retarda— tion of oxygen diffusion (Halton and Fisher, 1937). That water lowers the effectiveness of metal catalv ysts such as copper and iron (Uri, 1956). That water is attached to sites on the surface thereby excluding oxygen from these sites (Salwin, 1959). That water promotes non—enzymatic browning and browning can result in the formation of antioxidant compounds (Lea, 1958). That water forms hydrogen bonds with hydroperoxide and retards hydroperoxide decomposition (Maloney, 1966). 10 Oxidation Mechanism of Milk Lipids Theoretical Aspects of Autoxidations The autoxidative deterioration of lipid containing foods has been a serious problem in the food industry be- cause it leads to a decrease in organoleptical quality and nutritional value. A good understanding of the principal mechanism will be of value in examining the studies of the correlation between the peroxide test and the 2-thiobarbi— turic acid test in the autoxidized milk fat. Badings (1960) and Holman et a1. (1962), have pre- sented a good review of the theoretical aspects of autoxida— tion processes in lipids. The free radical chain reaction is a well—established explanation for the majority of autoxidation mechanisms. Peroxides are usually the primary products and the major general reactions are: l. Initiation: formation of free radicals 2. Propagation: the free radical chain reactions +R— + O >ROO- ROO- + RH )- ROOH-I-R- I 3. Termination: formation of non-reactive products with R-, ROO-, RO—, etc. 11 The rate of autoxidation is dependent on the energy required for the rupture of the a-methylene bond. The inter- mediately formed a-methylene radical is stabilized by re sonance o * —CH-CH=CH- <————+ -CH=CH-C:I- The two structures contributing to the resonance hybrid give rise to the formation of isomeric hydroperoxides. The re- action can be accelerated by pro-oxidant factors such as metal, free radicals, UV light, elevated temperature and moisture (which presumably increases mobility of pro- oxidants). On the other hand, autoxidation is inhibited by compounds which react with free radicals to form non-radical products. The autoxidation pattern is often complicated by numer- ous free radical reactions such as: a. Formation of epoxides b. Polymerization c. Formation of poly-peroxides d. Formation of cyclic peroxides e. Formation of secondary autoxidation products by dismutation of the hydroperoxides. Dismutation of HydrOperoxides and Off-flavor Production The hydroperoxides formed in the autoxidation of un- saturated fatty acid are usually not stable and will 12 decompose. The formation of these dismutation products is a serious problem in food products which contain lipids because of the formation of very objectionable off-flavor products such as short carbon chain aldehydes, ketones,; alcohols and other volatile carbonyl compounds. The Composition of Milk Lipid and Its Properties Milk fat consists chiefly of triglycerides of fatty acid. The composition of cow's milk is presented in Table 1. The unsaturated acids are responsible for the * Table l. The Composition of Milk Lipid Class of lipid Range of occurrence % Triglycerides of fatty 97-98 acid Di—glycerides of fatty 0.25-0.48 acid Mono-glycerides 0.016-0.038 Phospholipids 0.2-l.0 *Source: Byron H. webb and Arnold H. Johnson, 1965. development of oxidized flavor in the milk fat because of the lability of the carbons alpha to the unsaturated bonds. 13 The fatty acid composition of cow's milk fat is pre- sented in Table 2. As Tables 2 and 3 indicate, the major unsaturated fatty acid components are carbon chain 18 and the number of un— saturated double bonds are principally monoene and diene (oleic and linoleic acid). Autoxidation of Saturated Fatty Acid of Milk Fat From a practical standpoint, the breakdown of saturated fatty acids through oxidation can be ignored. Brodnitz (1968) observed that at 100°C the rate of saturated fat oxidation was about 100 times slower than for unsaturated linoleate. Autoxidation of Oleic Acid Oleic acid, which is a major unsaturated fatty acid of milk fat, has two alpha—methylene groups and these are the point of attack in the free radical chain reaction. Farmer et a1. (1943); Ross et a1. (1949); and Privett et a1. (1953) have proved that four hydroperoxides are formed in oleic acid oxidation. Badings (1960) quoting from Horikx and Schogot (1959) observed that saturated aldehydes with a chain length of C5 to C10 and the 2-enals with a chain length of C 6 to C11 are formed in the initial phase of the autoxidation of the fatty acid. 14 Table 2. The Fatty Acid Composition of Milk Fat Number of Percentage Fatty acid carbon atoms of total fat Butyric 4 3.7 Caproic 6 2.0 Caprylic 8 1.3 Capric 10 2.7 Lauric 12 4.0 Myristic 14 7.9 Palmitic 16 23.8 Stearic 18 10.7 Arachidonic* 20 0.5* Oleic* 18 38.5* Linoleic* 18 4.7* Unsaturated* 20 to 22 0.4* * Major unsaturated fatty acid. Source: Hilditch and Thompson, 1936. 15 Table 3. Comparison of Unsaturated Fatty Acid in Mole Percent of Milk Fat (Total Unsaturated Fatty Acid Components) Hilditch Smith Hilditch Length of Jack and and and and carbon Henderson Longenecker Dastur Paul chain 1945 1938 1938 1940 Iodine value 31.9 37.5 36.6 46.9 Clo 0.5 0.4 0.3 0.2 C12 0.3 0.9 0.3 0.2 C14 1.3 1.7 1.0 0.9 'k , * * * C16 3.6 3.7 3.0 2.8 C18 16.4* 24.8* 30.5* 31.4* C20 1.4 0.2 0.6 0.5 C 18:2 1.7* 2.9* 1.0* 4.9* *Major unsaturated fatty acid components. Source: Henderson, 1970. 16 Oxidative Deterioration of Dairprroducts It is generally accepted that in the oxidation of the unsaturated fatty acid of milk, triglycerides and phos- pholipids are always involved (Greenbank, 1953; Lea, 1953). El-Negoumy et a1. (1962, 1968), and Lea (1957) also con- cluded that off-flavor production is closely connected with phOSpholipid deterioration in butter, since phospholipid- free butter fat, on oxidation, gives rise only to “oily" and ultimately, "painty" flavors. Swanson and Sommer (1940) isolated the crude phospho— 1ipid fraction from milk which had developed a copper— induced oxidized flavor and showed that it had decreased in iodine value by a large amount, while the triglyceride fat had hardly changed. The copper induced "carboard—like“ flavor of whole milk is largely due to a series of C to 4 C11 2—unsaturated and 2:4 di—unsaturated carbonyl compounds, particularly the C7 to C9 aldehydes present at concentra— tions as low as 1 part in 109 of milk. These compounds are typical products of the autoxidation of a polyvunsatura— ted fatty acid (Lea, 1953). Chemical Tests fo£_Lipid Oxidation and Correlation to Sensory Evaluation 2—Thiobarbituric Acid Test for Measure- ment of Oxidative Rancidity Since Kohn and Liversedge (1944) reported that a red color resulted when tissue had been incubated aerobically 17 with 2-thiobarbituric acid (TBA), the production of malon- aldehyde has been widely used as a measurement of the oxidation of unsaturated fatty acid in food (dairy products: Patton et al., 1951; Biggs et al., 1953; Sidwell et al., 1954; Jennings et al., 1955; King, 1962; fish: Sinnhuber, 1958; pork: Younathan et al., 1960; and bakery products: Caldwell, 1955). The Principal TBA Reactants The mechanism and chemical structure of TBA—reactive materials are not very well understood even at the present time, though numerous investigators have postulated possible TBA—reactive materials. Bernheim and Bernheim and Wilber (1947); for example, suggested that TBA pigment from oxi- dized lipids involves the reaction of a three carbon come pound with the reagent. Patton and Kurtz (1951) observed that malonaldehyde, a three carbon compound, gives the characteristic absorbance at 532 nm in the test. Sinnhuber et a1. (1958) indicated that the principal TBA—reactive material is malonaldehyde and Patton and Kurtz (1955) re- ported that freshly prepared or unsaturated aldehyde did not give the TBA test, but when they were autoxidized or heated with Cu++ ions a positive reaction resulted. Sinnhuber et a1. (1958) demonstrated that crystalline TBA derivatives could be isolated from rancid salmon oil, 18 sulfadiazine and malonaldehyde. The molecular configuration of the pigment is believed to be a condensation product of two molecules of TBA with one molecule of malonaldehyde. Holman et a1. (1962) observed that an a,B—unsaturated peroxide radical is a precursor of malonaldehyde and has a relation to the TBA color production. Finally, Marcuse and Johansson (1971) reported that TBA color production is quan- titatively related to the amount of dismutation products of hydroperoxide such as conjugated triene, tetraene, pentaene and hexaene. TBA Reactive Material and water Extraction The true chemical structure and the origin of the total TBA-reactive materials are still unknown. However, Patton and Kurtz (1955) reported that the TBA—reactive material from.mi1k fat was water soluble and of low molecular weight. Tarladgis et a1. (1964) indicated that free malonalde- hyde is produced as a result of the oxidative breakdown of the unsaturated fatty acid and can be extracted with water. King (1962) demonstrated a sensitive TBA test of an intact milk sample and suggested that the TBA reaction should not be carried out in the presence of the oxidized or oxidizing lipids since further oxidation may occur during the reac— tion with TBA. The pigments derived probably do not entire— 1y reflect the organoleptic property of the intact sample. 19 The Correlation Between TBA—Reactive Materials and Rancid Odor of Food Products The correlation between TBA values and rancid odor of food products has not been fully investigated. Dunkley (1951) observed, however, that the color production in the 2-thiobarbituric acid test and the organoleptic flavor score of a whole milk sample having a oxidized flavor are closely correlated. Lillard and Day (1961) examined the relation- ship between the off—flavor of milk fat oxidized in diffused daylight to different levels and several criteria of chemi- cal change, including peroxide value and TBA value. They reported that TBA and peroxide test correlate well with oxi— dized flavor in this particular system. King (1962) also reported that the TBA test closely reflects the organoleptic conditions of the intact samples in the studies of oxidized flavor in the model systems containing fat globule material of cow's milk and ascorbic acid. Holman et a1. (1954) also demonstrated that the TBA color production and peroxide value have an essentially linear relation below a peroxide value of 1000 in oxidized poly-unsaturated fatty acids which contain conjugated diene, triene, tetraene, hexaene and pentaene. However, other investigators have reported a poor cor— relation between TBA test and off-flavor production. For example, Holland (1971) observed that in freeze-dried meat 20 the TBA values were low and could not be correlated with the appearance of off—odor or the absorption of oxygen. Tarladgis et a1. (1960) and Corliss et a1. (1963) also ob- served that the TBA value reached a peak at an early stage of oxidation and later fell. The relationship between TBA reactive materials and rancid odor in dehydrated food systems needs further inves— tigation with respect to the possible decomposition products of lipid oxidation. EXPERIMENTAL Separation of Milk Fat Three different lots of fresh, raw, winter creams were obtained from the Michigan State University dairy plant, pasteurized at 77°C for 10 minutes, cooled, and held at refrigerator temperature for subsequent churning. The milk fat was separated by conventional churning in a 5 liter Erlenmeyer flask at 12°C-16°C. After the cream was churned, the butter was recovered and carefully melted in a 50°C water bath. The melted milk fat was then washed with 50°C distilled water until a fairly clean aqueous solution was obtained. The washed milk fat was then dried under vacuum and filtered through filter paper in a 50°C incubator. Preparation of Freeze-dried Model Systems The freeze-dried model system, which contained micro- crystalline cellulose and milk fat, was prepared as follows. Components of Model Systems: 1. One part milk fat and six parts of microcrystalline cellulose, dry basis W/W. 21 22 2. The melted milk fat was homogenized with 10 parts of distilled water with 2,500 psi at 45°C by using a laboratory homogenizer. The small portion of unhomogenized milk fat was separated with a separaw tory funnel, re—homogenized and combined with the major portion. 3. The homogenized milk fat was mixed with 6 parts of micro-crystalline cellulose. 4. The thoroughly mixed model system was freeze—dried at 100 microns absolute pressure. Adjusting water Activity bthoisture ‘EquilibratiOn The sifted freeze-dried model system was:mixed uni“ formly in a nitrogen inflated vinyl bag. An accurately weighed amount of the freeze-dried system was humidified in vacuum desiccators which contained a specific saturated salt solution for desired water activity (see Table 4). Determination of water Activity An apparatus capable of measuring aw by either the Makower—Meyers (1943) or the Taylor (1961) methods was used to determine the water activity of the freeze—dried model systems (see Figure 2). In the present study a 23 Table 4. Relative Humidities of Saturated Salt Solutions at Given Temperature Saturated salt Relative Temperature °C, solution humidity, % and reference Lithium chloride 11 20 (l) (LiCl) 12 25 (4) 12 22.5 (1) Potassium thio— 47 25 (3) cyanate (KSCN) Sodium chloride 75.8 20 (4) (NaCl) 75 25 (1,2,4) 75.1 22.5 (3) References: 1. Weston et al., 1954 2. Rockland, 1957 3. Handbook of Chemistry and Physics, 55th ed. 4. wexler and Hasegawa, 1954 24 .msumusmmm Hoahms\uomoztum3oxoz toenaamsflm .N shaman . $65822 . 1.. 85% 2-2 .258... 1 3:28 £6 IIIIII o 3 ill! m Em and“ m/_ m as 2. +532 2 Wk ml. m o>_m> 20: SE: $65822 25 modified Taylor method was used. The sample tube was filled quickly with 3 g. of freeze-dried sample and then st0ppered with a greased ground-glass cap. After the sample tube was frozen in a mixture of dry ice and acetone for 30 minutes, the tube was quickly installed in the Taylor apparatus and held an additional 10 minutes in the dry ice—acetone mixture. The initial reading of the difference of oil "B", manometric fluid, levels between the right and left legs was made in the following ways: 1. The valve A_was set so that the vacuum pump would exhaust air from both legs of the manometer for five minutes after the pump had approached maximum vacuum as indicated by pump sound. 2. The valve A_was then set to pump the air only from the left leg of the manometer. The initial pressure difference in the system was read after reaching constant oil levels. The freezing solution was then replaced by water at ambient temperature and the equilibrated food sample was maintained at a given temperature under the constant vacuum pressure so that a maximum pressure change could be read. The total pressure change by the equilibrium moisture con— tent of food sample was calculated by subtracting initial pressure difference of both legs from the total pressure difference after reaching equilibrium moisture content at given temperature. 26 Possible errors which might be caused by trapped air in the sample were minimized by refreezing the sample tube 10 minutes and subtracting the difference of the pressure change caused by the refreezing from the total pressure difference by the food sample. Calculation of Water Activity In Taylor's apparatus the water activity can be calcu— lated as follows: where P(oil B) = Vapor pressure of pure water at temperature T., in terms of mm of oil B. Ah = pressure difference of manometer at equilibrium relative humidity after thawing with water at ambient temperature (at T.) Ah = pressure difference of manometer after re-freezing the sample. Chemical Test for Lipid Oxidation Peroxide Test Peroxide values were determined by a minor modifica— tion of the method used by Stine et a1. (l954)——fat extrac— tion, which was modified by the butanol—salicylate 27 de—emulsification method of Pont (1955), proceeded in the following manner: Ten 9. of the freeze—dried model system was reconsti— tuted with 25 m1. of distilled water and 15 ml. of the de-emulsification reagent were added. The mixture was then transferred quantitatively to a 9 g. 50% cream test bottle. After the reagent was thoroughly mixed, the bottle was placed in a 70°C water bath for 10 minutes and then centri- fuged for 3 minutes in a heated Babcock centrifuge. Distilled water at 70°C was added to bring the fat into the neck of the Babcock bottle and the bottle was again centri- fuged for 3 minutes. The peroxide determination of a weighed amount of milk fat was accomplished by the method described (Stine et al., 1954). Z-Thiobarbituric Acid Test The 2-thiobarbituric acid test of King (1962) was modi- fied to accommodate a freeze-dried model system based on micro-crystalline cellulose and milk fat by using water extraction of TBA—reactants, increasing reaction time with the TBA reagent and the chromatographic separation of pink TBA pigment. Five 9. of the freezeedried model system was reconsti— tuted with 20 m1 of distilled water and shaken vigorously for 10 seconds in a glass stoppered 50 ml Erlenmeyer flask 28 and tempered for 10 minutes at 45°C. One ml of trichloro- acetic acid solution (1 g. TCA/l ml) was added. The solu— tion was tempered for 5 minutes at the same temperature and was shaken vigorously. The sample was then filtered through No. 42 Whatman ashless filter paper. Four m1 of filtrate and 2 ml of TBA reagent (0.05 M) were reacted in the 50 ml screw capped glass tube for exactly 35 minutes at 100°C. The interference of yellow pigment caused by high heat treatment was elimi- nated by passing the reactants through a chromatographic column. Chromatographic Separation of Pink TBA Pfgment The TBA reacted solution was poured into a chromato— graphic column and forced through under 5 psi nitrogen pres— sure. The preparation of chromatographic column with cellulose (Whatman standard grade) used the procedure of Caldwell et al., 1955. Each column was washed with 2 m1 of the washed reaction mixture from the test tube, then with two 2-m1 portions of distilled water. A 10 ml volumetric flask was placed under each column, and absorbed red pigment eluted with aqueous pyridine (water:pyridine, 5:1), exactly 10 ml being collected. Optical density was read at 532 nm against distilled water using a Beckman Model DU-2 spectrophotometer. RESULTS AND DISCUSSION Lack of sensitivity of chemical methods in analyzing lipid oxidation of low-peroxide fat has been an obstacle to the study of autoxidation of purified milk fat. In recent years most studies of lipid oxidation in dehydrated food systems have been based upon the various methods of oxygen uptake methods. But for practical purposes, other methods, have been required. There is also scant information regard— ing the effects of water activity on off-flavor development; correlating of the 2—thiobarbituric acid test with other tests such as peroxide test and sensory evaluation. The purpose of this investigation was to determine the effects of water activity on milk fat oxidation and the correlation of two chemical tests, the TBA test and the colorimetric peroxide test to sensory evaluation of freeze—dried milk fat model systems. Preparation of Model Systems Three different lots of fresh, raw winter creams were pasteurized at 77°C for 10 minutes to inactivate enzymes and effectively inhibit microbial growth, which might alter the result of autoxidation studies. 29 30 The churned milk fat was melted carefully and washed with distilled water at 50°C to prevent possible coagulation of protein residue. For the further separation of water soluble components, the washed milk fat was concentrated under vacuum at 50°C, and the clear lipid layer was decanted. Any insoluble matter was filtered off at the same temperature. In the preparation of freeze—dried milk fat model sys— tems utilizing micro—crystalline cellulose (6 parts) as the carrier, the melted milk fat (1 part) was homogenized in distilled water (10 parts) to attain homogeneity. The freeze—dried systems were sifted with a stainless steel screen and stored at —30°C under nitrogen. Moisture Equilibration of the Milk Fat Model Systems Table 6 and Figures 3 and 4, respectively, present the moisture change and moisture equilibration curves during the humidification periods at 20 :_0.5°C under reduced pres- sure. The equilibrium moisture data were obtained by exposing the freeze-dried milk fat model systems to the required equilibrium moisture content (RH) at 20 :_0.5°C in vacuum desiccators. The equilibration temperatures and vacuum were selected in order to obtain optimum determina— tion conditions of water activities by the Taylor method 31 (manometric) and to prevent further lipid oxidation during humidification periods. The saturated salt solutions which were used in adjusting water activity are presented in Table 5 . Table 5. Saturated Salt Solutions and Equilibrium Moisture Content RH by Saturated Relative hmmidity reference salt solution by Taylor method * at 25°C ** LiCl 12.1 12.0 KSCN 46.8 47.0 NaCl 74.7 75.8 *Measurement water activity at 20.5°C. **Data from Rockland, 1957; Wexler and Hasegawa, 1954; in Strolle et al., 1965. The moisture content of the freeze—dried model systems during moisture equilibration was periodically determined on duplicate samples by the method described in A.O.A.C. (1965). At the final stage of the moisture equilibration curve, the water activities of the model systems were sys- tematically determined until constant values were obtained. 32 Table 6. Change of Moisture Content of the Model System During Attainment of Equilibrium Moisture Run 1 * % moisture in sample, wet basis Time over satd over satd over satd (hrs) LiCl KSCN NaCl 24 0.95 3.45 5.97 48 1.78 4.65 7.15 72 2.26 5.30 7.51 144 2.50 5.40 7.70 (aw 0.12) (aw 0.447) (aw 0.685) 168 2.51 5.45 7.75 (aw 0.12) (aw 0.462) (aw 0.745) 192 2.50 5.45 7.75 (aw 0.12) (aw 0.465) (aw 0.745) Runs 2 and 3 Time (days) 3 1.82 3.05 5.84 6 2.27 4.32 6.78 9 2.35 5.1 7.30 13 2.45 5.15 7.55 (aw 0.67) 20 2.50 5.40 7.75 (aw 0.12) (aw 0.447) (aw 0.742) 24 2.50 5.40 7.74 (aw 0.12) (aw 0.467) (aw 0.746) * Average of duplicate determination. 33 NS mg 3: o2 .AH csmv ooh.o.H on no Bowman H0008 mmoHsHHmOIumm MHHE a mo w>nso Esflunaawsow musumwoz 3.50: V ooEmE 202.10.229.22: 8 us mu .m madman I . _. _ . 4 A a peg: - Mien. .‘ o O oN IO.\\\ 0 0 o \ I o6 III\\\\\\\\\\\\\‘~ 4 4 d . 4 \\\\\¥1 1 o.o I 1 Il\ aIIIIlla _ _ _ _ _ ow aldwes “6001/ fiJazeM‘mseq 19M '1N31N00 slamsnow 34 .3. one m mssmv 01.6 H cm as Bowman Hmpoa mmoHsHHmouumm v3.2: a no 9550 gfiuowawsom onsumfloz is 52.3 assesses: cm am fl 3 Na 0 0 cm A I l q _ fl _ a .1I4.|I.fl I I \ III- all II_ I _ . a _ _ r .v musmflm :3 OJ C) .:>:o< $.23 .m onsmwm 36 we so no me o M _ L T _ _ I F _ J H I. n I m 0 nu W. low M . I G \ 1 ,M o m o\ I: .s v.0 1 w. m o m 13 w m m0 I. .1 AU m . rA .J o m mw _ o b l _ _ _ _ 1 N 37 The selection of water activities for the lipid oxida- tion studies of milk fat in freeze-dried model systems was based on an arbitrary figure representing each of these three regions: aw 0.12, aw 0.47 and aw 0.75. Several theories and hypotheses have been offered to explain the moisture adsorption isotherms by Langmuir (1918), Brunauer et a1. (1938), Henderson (1952) and Rockland (1957). The BET isotherm by Brunauer et a1. (1938) has been most widely used. The limitation of the BET theory is confined to its ability to explain the observed facts in the range of water activity between 0.1 to 0.4. The BET monolayer value has frequently been used to predict the optimum mois- ture content of a food in relation to shelf life. The BET monomolecular layer value of the milk fat model systems was calculated from the moisture adsorption isotherm data by the following procedure: BET equation: a = l (C — l) a (l - aT'm ml m: Grams of water per 100 g. of dry matter corre' sponding to a relative vapor pressure C: Constant related to the heat of adsorption C = exp (OS/RT) where 38 OS = the site of interaction energy of adsorp- tion of water molecules to the solid surface at temperature T R = universal gas constant T = absolute temperature ml: Grams of water equivalent to monolayer absorbed on 100 g. of dried solid. The data a(100)/(l - a)m and relative humidity are presented in Table 8 and Figure 6. Table 8. The Value for BET Monomolecular Layer Plot water activity Moisture content a(100) (a) Dry basis (m) 771 - a)m 0.12 2.51 5.43 0.208 3.53 7.44 0.28 4.23 9.19 0.465 5.60 15.52 0.618 6.95 23.20 0.745 8.43 34.65 From the straight line and intercept of the line, the monolayer value was calculated. The BET monolayer value was water activity of 0.23 and at the moisture content of 0.37 g. of water per 100 g. dry solid. 39 .u.~.o H om um ewumwm Hmoofi owwuolmnmmnu may no uon HmwmaosOE 9mm one .m musmfim as x as? as; 1 Ca 2,5: 122%. 8 E 8 cm 8 cm 8 2 o e _ _ a r _ 4|. _ o\ o\ o\ A \\ I s \ \ \ \\\\ I a 0. HS 0 m.~ u A H V ummonDsH H ~sa.o u I flww o macaw _ b om :IO EIITIVA UJ(I?"I) I001)? 40 The Change in Moisture Content of the Milk Fat Model Systems During Autoxidation Periods The model systems, which were adjusted to water activ— ity figures aw 0.12, aw 0.47 and aw 0.75 were placed in containers with air tight covers in 10 9. portions and stored over the saturated salt solutions in desiccators to provide the same relative humidity conditions as the sample and also to prevent additional moisture change. During two months of storage, the changes of moisture content were monitored by the A.O.A.C. method (1965) at the initial stage, after 1 month, after 2 months by the random selection of 3 samples of each water activity through dupli- cate determination. The average moisture change of the model systems during the storage periods was negligible. Sensory_Eva1uation Sensory evaluation of the freezeedried model systems was performed periodically during two months of storage at 20 :_0.5°C. The scoring system, adapted from a similar method used by Dunkley (1951), is as follows: a. Excellent (no detectable oxidized . flavor)OOOCOOOOOOOOOO00.0.0.0...0.. 4O pOints b. Good (detectable oxidized flavor... 37-40 points c. Fair (slightly oxidized f1avor..... 34-37 points d. Poor (distinctly oxidized flavor... 25-34 points 41 The selection of panel members are based upon their ability to differentiate three standard samples: fresh, a two month old sample at room temperature, a mixture of 50% fresh, and a 50% two month old sample. In the preparation of sample, 5 g of duplicate por- tions of the freeze-dried sample in glass bottles with air- tight plastic lids were distributed to the panel members in random order during mid-morning or mid-afternoon. The three criteria based on different degrees of oxidized sample were used to obtain consistent results. The sensory evaluation of tasting was not applicable in the freeze-dried model systems by confusing of character— istic cellulose flavor against the mild oxidized flavor. However, in sensory evaluation, some difficulties and confusing of flavor judgments were encountered at the ini— tial stage of the autoxidation by a small amount of chemical change responsible for oxidized flavor defects. The total pattern of off-flavor production at two month storage periods was able to differentiate among the model systems, aW 0.12, aw 0.47 and aw 0.75. The best storage stability was obtained at a water activity of 0.47, while the samples stored aw 0.12 showed slightly increased off-flavor production and those held at aw 0.75 produced a strongly detectable off—flavor after a two month storage period. 42 The off—flavor production shows some correlation with the peroxide test, but not with TBA test. The results of off—flavor evaluation in freeze-dried model systems con- tained various amounts of moisture content is presented in Table 9 and Figure 7 (see the following four pages). Effects of Varying Water Activities on Autoxidation of Milk Fat in FreezeJdried Model Systems Lipid oxidation of milk fat in freeze-dried model sys— tems based on micro—crystalline cellulose was determined by the peroxide test, TBA test and off-flavor evaluation dur- ing a two month storage period. The major difficulties encountered in this investigation included the resistant nature of purified milk fat to autoxidation and the narrow range of chemical change which occurred during storage. This may be due to the composition of the milk fat which is high in oleic and low in polyunsaturated fatty acids. A good example of this resistant nature over a longer period was reported by Sidwell et a1. (1954). His data for the storage stability of butter fat are presented in Table 10 (see page 47). The peroxide test, TBA test and sensory evaluation were used to monitor the degree of lipid oxidation in this study; the most applicable chemical test was the peroxide test. The results of the TBA test showed a poor correlation to both the peroxide test and the sensory evaluation. In the 43 Table 9. Flavor Scores of Freeze—dried Model Systems Number Storage Average Score* of Run days a 0.12 a 0.47 a 0.75 w w w 0 38.2 38.7 37.3 13 38.3 38.2 34.7 Run 1 20 37.4 36.7 34.8 30 36.0 36.9 32.8 43 35.1 36.4 31.7 0 38.0 37.8 38.3 Run 2 14 36.8 37.3 36.1 30 36.5 37.1 32.0 48 35.2 36.5 31.2 0 37.5 37.4 37.7 8 37.4 36.9 35.7 Run 3 21 36.5 37.3 34.0 35 35.5 36.7 31.6 46 35.0 36.4 31.3 *The average score of duplicate evaluation from the results of 6 panel members. 44 .36 new $6 .35 mo 3... 98 0.3 p... 3303 mceuso mfimumam Hmoofi omwuotoummum mo sowpmsamtwm muomsmm .th 95mm; m>m .IIII. I mud n O 0 NH .0 26 (l4 Hm mm mm R 38035 HO/W'H .35 PB 25 .36 «o a... can u. H om um omnuoum msHHso mewummm Hmoofi oowuolmumoum mo sowussam>m muonsmm .Nth Gasman m>m mucosow .mlh musmflm m>wuom nouns msamum> sH newsman Hoooa omwuotonmonm cw sowumoaxo owned .a magmas m>wuom “mums mswmuob CH newsman Hmooa_oofinotmuomnm cw sowumpwxo owqu .oH gunmen m> mo muowmmm one :5: £25: 2,2221 53:2 32; .HH wusmom wd Nd ed md ed md N .o To _ l s I 4‘ . I _ ._ I was, 53.82: Em _ I m SE wuoo your? msowum> us newsman Hmoofi mmoHSHHmo me you Mafia omueoflxo mo coauosooum Hoaoo «m9 .NH gunmen m>wo pumpssum.H A>zv mosam> sows * .H0 #0 museums ”mousom 100 ~.0 H.H I00 H.0 0.0 100 0.0 0.0 Hmcmenmomnoauv.~ 100 H.0 0.0 100 H.0 0.0 Inc 0.0 0.0 Hmcmenmomecosw.~ I~0 0.0 m.m 100 ~.0 0.0 100 0.0 «.0 Hmcoeemomauv.~ Imo 0.0 0.~ 100 0.0 0.H INV 0.0 0.0 Hmcmeomcozne.m 100 0.0 0.0 100 0.0 0.0 100 0.0 0.~ Hmcmeemxmmuv.m mamcmeemxam Ame ~.0 0.H Ame H.~ m.0 Haemomonm “mo 0.0 0.0 “m0 H.0 0.H 100 ~.H «.0 Hmamaozum I00 0.0 «.0 Imo 0.0 0.0 100 0.0 H.m HmcmuoOIm 100 0.0 «.0 100 0.0 0.0 x00 0.H m.H Hmcoummmum 100 0.0 H.0 100 0.0 0.0 100 0.H 0.0 Hmcmxmmum mamamxaa Ame 0.0 0.0 100 0.0 0.~ 100 0.0 0.0 awesome Ame 0.0 0.0 Ame 0.0 0.0 I00 0.H 0.0 Hmcmummm 5 0.0 0.0 E 0.0 0.0 80 0.0 NJ. 18980 100 0.0 0.0 Imc H.0 H.H Ame 0.H 0.0 Hmcmucmm measures a 0m >2 0 am >2 0 am >2 moanmeam ommeus H .3: 00m 000 .msIm .sa.mw« 000 .ue N .sa 000 meanweam «0 0002 000 amusemoas mo momma uswuwmmwo sues coeuomom «me no monsoonm mo wocmoHOmn< .ma manna 64 poly-unsaturated fatty acid compositions in food systems such as purified milk fat because the ratio of the fraction of B,Y-unsaturated, polyene conjugated components derived from autoxidized milk fat to the total off-flavor component will be much less than that of higher unsaturated fatty acid. The Effects of Chemical Reactivity of Unsaturated Fatty Acid'in Milk Fat and TBA Reagent Kenstone et a1. (1955) observed that oxidized linolen- ate (C 18:3) produced 60 to 100 times as much TBA color as oxidized linoleate (C 18:2); whereas, oleate (C 18:1) pro— duced no color when all were measured at the same level of autoxidation indicated by peroxide value. From the standpoint of autoxidation and chemical reactivity of poly-unsaturated fatty acid with TBA reagent, it may be true that the early stage of milk fat oxidation is predominantely governed by the trace amount of poly- unsaturated fatty acid components. The reasons for the higher oxidation rates at the initial period of the TBA test in this investigation may be due to the oxidative behavior of trace poly-unsaturated fatty acids. 65 The Effects of Physical Structure of Freeze—dried Modelgystems on the Accumulation of TBA Reactive Materials In the comparison of the TBA color reactive materials of the milk fat oxidation (Runs 2 and 3, see Figure 14 on the following page), the possible total accumulation of TBA color reactants during the autoxidation period can be repre- sented as a total integrated area of the autoxidation curves in the plots of TBA absorbance at 532 nm vs. time of storage period. The integrated areas at the early stages of oxida- tion increased as did the order aw 0.75, aw 0.12 and aw 0.47, which represent an inversion of the oxidative pattern of the oxidative curves. From these results, it can be postulated that the TBA color reactants could be further oxidized to produce a less TBA reactive, volatile, very unstable and water soluble short carbon chain compounds probably satur— ated aldehyde, ketone, alcohols in freeze-dried model systems. In the nature of porous and sponge like matrix of de- hydrated food systems, the holding capacity of the water soluble, volatile, and unstable TBA reactive materials will be affected in the different ways from high moisture content food or liquid food systems. This could be a part of the explanation as to why many researchers have failed to demon— strate a correlation of the TBA test to other chemical and 66 2 u .mfiwumhm .3on you head: omwucnoummum sun hauntflnuom noun? mo muoommm can mamwumuma @3393.» «me no cowumasfisoom can. m>