FROM ULTRA-HIGH-TEMPERATURE PROCESSED MILK A THESIS FOOD SCIENCE FRANK TAKACS, JR. MASTER OF SCIENCE. ^loiame^ty

Transcription

1

2 SOLID PHASE EXTRACTION OF STALE FLAVOR MPONENTS FROM ULTRA-HIGH-TEMPERATURE PROCESSED MILK by FRANK TAKACS, JR. B.B.A., Ranke Cllege, 1977 A THESIS Submitted in partial fulfillment f the requirements fr the degree MASTER OF SCIENCE FOOD SCIENCE Department f Animal Sienes and Industry KANSAS STATE UNIVERSITY Manhattan, Kansas 1989 Apprved by: ^loiame^ty Majr Prfessr

3 Lb Knlrt AllEDfi 3D471A TABLE OF NTENTS FN W. PAGE INTRODUCTION 1 C, 2 REVIEW OF LITERATURE 4 I. DEFINITION 4 II. HISTORY 4 III. NSUMPTION 5 IV. UHT PROCESSING SYSTEMS 7 V. EFFECTS OF UHT PROCESSING AND STORAGE 8 A. Nutritin 8 B. Mirbilgial 12 C. Clr 13 D. Sedimentatin 13 E. Flavr 14 VI. OFF-FLAVORS IN UHT MILK 15 A. Cked Flavr 15 B. Stale Flavr 16 C. Islatin 18 VII. FACTORS AFFECTING FLAVOR OF UHT MILK 19 A. Raw Milk Quality L9 B. Strage Time and Temperature l l) C. Disslved Oxygen 21 i

4 MATERIALS AND METHODS 23 I. EVALUATION OF SOLID PHASES 23 II. ORGANOLEPTIC ANALYSIS 24 A. Intrdutin 24 B. Training Sessin 25 C. Evaluatin f Samples 25 III. ANALYSIS OF FREE FATTY ACIDS 27 A. Intrdutin 27 B. Extratin Predure 28 C. GC Analysis 30 D. Quantifiatin 32 IV. ANALYSIS OF NEUTRAL VOLATILE MPOUNDS 33 A. Intrdutin 33 B. Extratin Predure 35 C. GC and GC-MS Analysis 37 D. Quantifiatin 41 V. STALE FLAVOR REPLICATION 42 A. Intrdutin 42 B. Repliatin Predure 42 C. Evaluatin f Samples 44 RESULTS AND DISCUSSION 45 I. EVALUATION OF SOLID PHASES 45 ii

5 II. FLAVOR ANALYSIS 46 III. ANALYSIS OF FREE FATTY ACIDS 50 IV. ANALYSIS OF NEUTRAL VOLATILE MPOUNDS 77 V. STALE FLAVOR REPLICATION ANALYSIS 93 NCLUSIONS 98 REFERENCES 99 in

6 LIST OF TABLES TABLE PAGE 1. Typial vitamin lsses by heat treatment Lsses f sme water-sluble vitamins during strage f UHT milk in the dark Results frm the evaluatin f slid phases Means f rganlepti flavr sres Free fatty aid nentratins in varius milk samples Changes in UHT milk free fatty aid nentratins during strage Cmpunds identified in 7 m. stale UHT milk Cmpunds identified in fresh UHT sample Means f sensry flavr sres fr stale flavr repliatin 93 IV

7 LIST OF FIGURES FIGURES PAGE 1. Types f UHT sterilizing presses 7 2. Example f the frmat used by taste panel Summary f milk samples analyzed fr free fatty aids Samples analyzed fr vlatile mpunds Summary f rganlepti results - ked arma Summary f rganlepti results - ked flavr Summary f rganlepti results - stale arma Summary f rganlepti results - stale flavr Cmparisn f butyri aid (C 4 ) nentratins (ppm) in varius milk samples Cmparisn f apri aid (C 6 ) nentratins (ppm) in varius milk samples Cmparisn f apryli aid (C 8 ) nentratins (ppm) in varius milk samples Cmparisn f apryli aid (C 8 ) nentratins (ppm) in fresh, stale, and stale filtered UHT milk Cmparisn f apri aid (C 10 ) nentratins (ppm) in varius milk samples Cmparisn f apri aid (C 10 ) nentratins (ppm) in fresh, stale, and stale filtered UHT milk Cmparisn f Laurie aid (C 12 ) nentratins v

8 (ppm) in varius milk samples Cmparisn f lauri aid (C 12 ) nentratins (ppm) in fresh, stale, and stale filtered UHT milk Cmparisn f myristi aid (C 14 ) nentratins (ppm) in varius milk samples Cmparisn f palmiti aid (C 16 ) nentratins (ppm) in varius milk samples Cmparisn f steari aid (C 18 ) nentratins (ppm) in varius milk samples Cmparisn f lei aid (C 18:1 ) nentratins (ppm) in varius milk samples Changes in butyri aid (C 4 ) nentratins (ppm) ver a 90 day perid Changes in apri aid (C 6 ) nentratins (ppm) ver a 90 day perid Changes in apryli aid (C 8 ) nentratins (ppm) ver a 90 day perid Changes in apri aid (C 10 ) nentratins (ppm) ver a 90 day perid Changes in lauri aid (C 12 ) nentratins (ppm) ver a 90 day perid Changes in myristi aid (C 14 ) nentratins (ppm) ver a 90 day perid Changes in palmiti aid (C 16 ) nentratins (ppm) ver a 90 day perid Changes in steari aid (C 18 ) nentratins (ppm) ver a 90 day perid Changes in lei aid (C 18:1 ) nentratins vi

9 (ppm) ver a 90 day perid GC hrmatgram f 7 m. ld stale UHT milk GC hrmatgram f 7 m. ld stale UHT milk filtered thrugh C 18 Sep-Pak GC hrmatgram f 20 m. ld stale UHT milk GC hrmatgram f 20 m. ld stale UHT milk filtered thrugh C 18 Sep-Pak GC hrmatgram f fresh pasteurized milk GC hrmatgram f raw milk GC hrmatgram f 90% raw/10% pasteurized milk GC hrmatgram f free fatty aid standards GC hrmatgram f fresh UHT milk Chrmatgram fr 7 m. stale UHT sample Chrmatgram fr 7 m. stale UHT filtered thrugh Sep-Paks Chrmatgram fr methylene hlride elutin f Sep-Paks 7 m. stale UHT sample Chrmatgram fr ethyl-ether elutin f Sep-Paks fr 7 m. stale UHT sample Chrmatgram f 20 m. stale UHT sample Chrmatgram f 20 m. stale UHT filtered thrugh Sep-Paks Chrmatgram f methylene hlride elutin f Sep-Paks f 20 m. stale UHT sample Chrmatgram fr ethyl-ether elutin f Sep-Paks f 20 m. stale UHT sample 89 vii

10 47. Chrmatgram f fresh UHT sample Chrmatgram f mpunds eluted frm Sep-Pak with methylene hlride Chrmatgram fr 3 m. stale UHT sample with Hl added Summary f rganlepti results - ked arma Summary f rganlepti results - ked flavr Summary f rganlepti results - stale arma Summary f rganlepti results - stale flavr 97 vin

11 ACKNOWLEDGEMENTS I wuld like t sinerely thank Dr. Ike J. Jen, my majr prfessr, fr the utstanding supprt he has prvided me thrughut my studies. His inspiratinal leadership, prfessinal knwledge, and persnal mentrship has enabled me t amplish my aademi gals and bjetives. Withut his guidane, it wuld have been impssible t mplete this investigatin. Additinally, I wuld like t thank the members f my supervisry mmittee, Dr. William G. Ikins and Dr. Leniel H. Harbers fr their help in reviewing this manusript. I am als grateful t Dr. William G. Ikins fr the valuable advie and assistane he has prvided me. Appreiatin is extended t the many faulty, staff and graduate students wh have helped me alng the way. A speial thanks t Liz Raddatz wh heerfully typed this manusript while sarifiing her free time. Finally, I want t regnize my wife, Carla, fr her unfailing supprt thrughut all my numerus endeavrs. IX

12 INTRODUCTION Ultra-high-temperature (UHT) pressed milk an be desribed as milk that has been pressed t extremely high temperatures fr shrt perids f time and aseptially pakaged. The bjetive f this methd f pressing is t prdue a mmerially sterile prdut while minimizing the heat-indued hemial reatins that are respnsible fr flavr defets. Althugh there is n ne definitin fr UHT milk, it is generally aepted that the pressing temperature is frm 135 C-150 C and held fr a few sends. Inreasing the heat treatment will prvide an even greater sterilizatin effet n bateria with a derease in baterial spilage, but is limited by the unaeptable hemial hanges whih are brught abut by the heat pressing. These hemial reatins an ause adverse effets n the lr, flavr and nutritinal attributes f UHT pressed milk. UHT milk pressing time and temperature parameters, therefre, prvide a gd balane between a maximum sterilizatin f bateria and minimum hemial reatin rate. Shelf-stability at ambient strage temperatures is UHT pressed milk's greatest advantage ver fresh pasteurized milk. It maintains its high nutrient value withut a requirement fr refrigeratin. Renner and Berlage-Weinig (1983) summarized that UHT milk has lr, flavr, and nutritive value equal t pasteurized milk and superir t nventinally sterilized milk. Beause f these fatrs, UHT milk uld prvide a gd slutin fr the preservatin f the large

13 milk surpluses in the United States and uld als be used t explit new markets where nutritin, strage, r distributin prblems are mmnplae. These new markets may inlude underdevelped untries, the military, and amping enthusiasts (Jelen, 1982). Replaing r substituting UHT milk fr fresh pasteurized milk by nsumers wuld be an even mre fundamental use f this tehnlgy. The ptential st and energy savings when mpared t fresh pasteurized milk uld represent a signifiant amunt. A redutin in st and energy uld be realized at the whlesale, retail, and nsumer level. The need t purhase refrigeratin equipment suh as mpressrs, insulated warehuses, display ases, refrigerated truks and even hme refrigeratin uld be redued r eliminated. Additinally, the redutin r eliminatin f this equipment wuld result in a signifiant derease in energy sts as well. Finally, distributin sts uld be lwered by reduing the frequeny f replenishment by inreasing stkage levels. This uld redue persnnel, equipment and energy sts fr the manufaturer and retailer by allwing fr mre effiient prdutin sheduling while minimizing dwn time. Althugh UHT pressed milk has been widely aepted by nsumers in Eurpe, it has nt gained the same ppularity in the United States. This lak f aeptane an be partially redited t st and lak f serius marketing. Hwever, the flavr f UHT milk is prbably the mst limiting fatr. UHT milk is ritiized as having an ff-flavr (ked r stale) when mpared t fresh

14 pasteurized milk as a result f the varius hemial reatins that have taken plae. Amerian nsumers, hwever, may be verly sensitive t the flavr f UHT milk beause they have been nditined t equate the taste f fresh pasteurized milk as the standard fr what fresh milk shuld taste like. This nditining has been reinfred by the availability f fresh milk supplies, refrigeratin equipment, and the advaned distributin systems present in this untry. Sine the ff-flavr f UHT pressed milk serves as a serius impediment twards nsumer aeptane and marketing suess, the study in this area is bth relevant and benefiial. The purpse f this study is t determine the ause f stale flavr develpment in UHT pressed milk. Speifi bjetives f ur researh are t islate and identify the hemial mpunds respnsible fr stale flavr develpment in UHT pressed milk utilizing C 18 Sep-Pak artridges as a tl fr entrapping r remving these mpunds.

15 REVIEW OF LITERATURE I. Definitins There are many definitins available fr ultra-high-temperature (UHT) milk. Burtn (1988) desribed the UHT press as ne in whih the prdut is heated t a temperature f C in ntinuus flw using a heat exhanger. The Internatinal Dairy Federatin defines UHT milk as having been subjeted t a ntinuus flw heating press at a high temperature fr a shrt time. The PHS/FDA Grade A milk Ordinane and Cde requires that the prduts be heated t nt less than C (280 F) with at least a 2 send hlding time (Wainess, 1982). Althugh n tw definitins may be exatly alike, it is generally aepted that UHT milk has been heat-treated t extremely high temperatures fr shrt perids f time and aseptially pakaged. The time-temperature standards are usually in the range f C fr 2 t 8 sends (Hsu, 1970). II. Histry The develpment f UHT pressed milk is extensively reviewed (Burtn, 1988; Westhff, 1978). Jnas Nielsen had pineered the first rerded UHT pressing plant by 1913 and later develped an asepti anning system. In the late

16 1940's, plants started using higher pressing temperatures and shrter hlding times t give prduts better baterilgial quality whih resulted in less hange f lr, flavr, and nutritive value f the milk. UHT milk was prdued and distributed as early as 1940 in the United States and marketed in Switzerland in 1953 Burtn (1988). "Fresh-tasting," lng shelf-life milk was prdued by sterilizing milk in bttles in the 1940's and in ans in the 1950's (Hsu, 1970). In 1951, a retangular plyethylene-ated paper bnd ntainer, with an aluminum fil laminate barrier was intrdued in Switzerland by Tetra Pak, Sweden (Mehta, 1980). The asepti systems, first by Tetra Pak alng with the develpment f a wide variety f heat exhangers has led t the present state f asepti pressing wrldwide (Burtn, 1988). III. Cnsumptin UHT pressed milk has been available fr ver 30 years in Eurpe. This pressing methd with asepti filling f milk and milk prduts is nw pratied in mre than 60 untries (Burtn, 1988). In Germany, apprximately 40% f the milk nsumptin in 1983 was in the frm f UHT milk (Ann, 1983). In 1986, hwever, Italy led the way with 56.9%; Prtugal 50.9%; Spain 39.8%; Germany 38.6%; and Switzerland 32.3% (Burtn, 1988). In the United States, nsumptin

17 f UHT pressed milk is less than 1% f the ttal milk sales. This is a result f pr nsumer aeptability brught abut by the ff-flavrs assiated with UHT milk as well as ther fatrs already disussed. The largest user in the United States is the military whih nsumed UHT pressed milk fr pratially all f their training exerises. The fllwing table mpares UHT milk nsumptin (%) in 1986 t 1975 fr Western Eurpe. Cuntry Austria Belgium Denmark Finland Frane Germany Greee Italy Netherlands Nrway Prtugal Spain Sweden Switzerland United Kingdm Sure: (Burtn, 1988) The abve data shws large inreases in nsumptin fr sme untries and little r n inrease fr thers. This may be due t different nsumer preferenes, st nsideratins, and effetive marketing strategies amng the varius untries.

18 IV. UHT Pressing Systems The tw primary methds used fr the manufature f UHT pressed milk tday are diret and indiret systems. Figure 1 belw prvides a summary f the systems. - Plate - Cnentri Tube Indiret heating Tubular Sraped Surfae I Shell and Tube Steam r ht water - Injetin, steam-int-milk Diret Heating with Steam - Infusin, milk-int-steam Figure 1. Types f UHT Sterilizing Presses (Burtn, 1988). In diret heating systems, the prdut is mixed diretly with steam under pressure, s that the steam is ndensed and its latent heat f vaprizatin is

19 transferred t heat the prdut very rapidly (Mehta, 1980; Burtn, 1988). Steam an either be injeted int the milk r milk an be sprayed int an atmsphere f steam (Infusin). Cndensatin f the steam int the milk auses dilutin f the prdut whih is mpensated using a "flash-dwn" vauum treatment whih remves the added water and ls the milk at the same time (Rerkrai, 1986). In the indiret heating system, milk is heated thrugh a physial heatnduting barrier (usually f stainless steel) between the prdut and the heating medium f steam r pressurized ht water. Examples inlude tubular, plate and sraped surfae indiret presses (Mehta, 1980). V. Effets f UHT Pressing and Strage A. Nutritin The prperties f UHT-pressed milk are influened by hanges that result frm the heat treatment and thse urring after the heat treatment during strage. Overall, there is little hange in the nutritinal quality f UHT pressed milk when mpared t fresh pasteurized milk. The nutritive value f prtein in severely heated milk may be impaired, beause f a drp in the availability f lysine brught abut by the Maillard reatin. Hwever, there is little r n adverse effet n the nutritinal quality f prteins during UHT heat treatment (Burtn, 1988; Renner and Berlage-Weinig 8

20 1983; Frd & Thmpsn, 1981). During strage, the Maillard reatin leads t a valent plymerizatin f the aseins, and t an inreased resistane t prtelysis. It is nt likely that these hanges have a signifiant effet n nutritinal and rganlepti qualities. Physial r hemial hanges in milk fat d nt appear t have nutritinal nsequenes frm UHT pressing. There is sme lss f fatty aids in the milk triglyerides, linlei aid, 33%; linleni aid, 13%; and arahidni aid, 7% (Pl & Grt, 1960) and up t a 30% lss f individual free fatty aids (Waymbe & Lindsey, 1969). During strage, the nentratin f free fatty aids inreases in UHT milk; hwever, their inrease appears t be dependent n the strage temperature and type f pressing. Free fatty aids are prdued mre rapidly at high strage temperatures than at lw; with higher-fat milks; and fr diret than indiret pressing (Shmidt & Renner, 1978a). This is prbably aused by the survival f resistant lipases f milk r frm thse prdued by psyhrtrphi rganisms. The free fatty aids eventually lead t ff-flavrs in the prdut when their nentratin exeeds the threshld f detetability (Burtn, 1988). The lss f vitamins during pressing is dependent n the severity f the heat treatment and expsure t light r xygen befre r after pressing. Typial vitamin lsses due t varius heat treatments are given in Table 1.

21 Table 1. Typial vitamin lsses by heat treatment (derived frm Frd and Thmpsn, 1981). Vitamin Lss (%) Past. Steril UHT Thiamin <10 Ribflavin NS Nitini Aid NS Vitamin B 6 <10 Vitamin B 12 <10 Panttheni Aid NS Bitin NS Fli Aid <10 Asrbi Aid 20 Vitamin A NS Vitamin D NS Vitamin E B-artene NS NS NS NS NS NS <90 10 NS NS NS NS NS NS NS NS NS NS NS NS NS = Nt signifiant During strage and in the absene f light, the fat sluble vitamins A, D, and E are stable at rm temperature at least fr 3 mnths (Frd et al., 1969). Sme water sluble vitamins suffer sme lss during strage in the dark. The data in Table 2 shws the lsses f sme f the water-sluble vitamins. 10

22 Table 2. the dark. Lsses f sme water-sluble vitamins during strage f UHT milk in Vitamin Lss (%) B Thiamin NS 10 NS Ribflavin NS Nitini Aid NS 20 - Vitamin B Vitamin B Panttheni aid NS 30 - Bitin NS 20 - NS = Nt signifiant A = Frd et al. (1969). 3 mnths at C B = Grner & Uherva (1980). 6 weeks at C C = Thmas et al. (1975). 9 weeks at 23 C The greatest lss f vitamins during strage f UHT milk is asrbi aid and fli aid. Lsses f these vitamins are interlked, and are dependent n the availability f xygen (Frd et al., 1969; Thmas et al., 1975). If the 2 ntent is 1 mg/liter r less, and the UHT is stred in the dark in ntainers whih are impermeable t xygen, bth asrbi and fli aid are stable and may shw a lss f nly abut 20% ver a 3-mnth perid (Frd et al., 1969; Thmas et al., 1975). There is n lss f minerals as a result f UHT pressing and strage (Henry & Tuthill, 1960; Pelet & Dnath, 1974). The mineral ntent f UHT 11

23 pressed milk is just as high as fr fresh pasteurized milk. B. Mirbilgial UHT pressing is intended t destry all the mirrganisms present in bth vegetative and spre frming states r at least make them inapable f grwth in the prdut, s that a lng keeping quality is btained withut refrigerated strage (Burtn, 1988). This des nt mean that the prdut is sterile in an abslute sense beause it may ntain sme mirrganisms, hwever, the remaining viable mirrganisms are inhibited under the nrmal strage nditins. The term "mmerially sterile milk" has been frequently used t desribe UHT milk (Lembke, 1972). Mirrganisms are destryed by heat when the mirbial prteins agulate and enzymes required fr their metablism are inativated. Althugh mst mirrganisms and their spres are destryed by the UHT press, a few may be resistant t the UHT heat treatment. Spres f the bligate thermphili sil baterium Baillus stearthermphilus and spres f mesphili bailli and Clstridia may survive UHT treatment. Atwal et al. (1974) examined different types f bulk milk frm a mmerial plant and fund that the mst mmn resistant spre strains were B. subtilis and B. stearthermphilus. 12

24 C. Clr UHT pressing inreases the refletane f milk due t the denaturatin f serum prteins and their aggregatin with asein resulting in a whiter prdut (Williams et al., 1955; Burtn, 1955; Hrak & Kessler, 1981; Kessler, 1981). Milk bemes brwner with the inreased severity f heating as a nsequene f the Maillard reatin between the lysine f the milk prteins and latse (Adrian, 1974). The Maillard reatin lwers the degree f refletane and therefre lightness aused by an inrease in the green and yellw mpnent in the milk (Bsset et al., 1979). During strage, lightness further dereases with a redutin in the green mpnent and an inrease in the yellw mpnent (Zadw, 1970; Ismail and El Deeb, 1973). Only a small hange in lr an be attributed as a result f UHT pressing. D. Sedimentatin and Age Gelatin Sedimentatin is aused by the denaturatin f the milk prteins r preipitatin f the salts in milk resulting frm the intensive heat used in UHT pressing. This sediment settles t the bttm f the artn with sme f the sediment returning t slutin during strage (Burtn, 1988). The degree f sedimentatin inreases with the intensity f heat treatment (Thme et al., 1964). UHT pressed milk shws a greater tendeny t thiken and agulate during strage than pasteurized milk. The ause f age gelatin is nt fully 13

25 understd althugh a lt f literature is available n the subjet (Harwalker, 1981). Hstettler and Imhf 1965 suggests that a asein-serum prtein mplex is invlved while Muir (1984) nsiders the milk fat t be invlved. Other wrkers attribute it t prtelyti enzymes that riginated frm the grwth f psyhrtrphi vegetative rganisms in the ld-stred raw milk that survived the UHT heat treatment (Burtn 1988). The enzymes prdued by a pseudmnad an breakdwn K-asein t para-k-asein during strage in a similar way as rennet. This destabilizes the asein mielles and leads t gel frmatin. E. Flavr UHT pressed milk exhibits varius flavr hanges. When fresh, there is a strng heated r ked flavr with a sulfurus dr (Shipe et al., 1978). After a few days, the ked flavr disappears t leave a harateristi flavr mparable t fresh pasteurized milk with a slight ked residue. On strage, a stale flavr develps and inreases in intensity ver time (Burtn, 1988). Even when the flavr f UHT milk is at its best, it is judged as inferir t that f pasteurized milk (Eer-kins-, 1985). Cnsumer judgments, hwever, are made by peple using raw r pasteurized milk as their standard and may therefre be biased r prgrammed when hsing a respnse. Finally, twards the end f the UHT shelf-life and just prir t age gelatin, a bitterness may be detetable. 14

26 VI. OFF-FLAVOR IN UHT MILK Ashtn et al. (1965) ategrized the flavr hanges in UHT milk int five perids. Immediately after pressing, there is a strng biled abbage flavr. After tw r three days, there is a strng ked flavr. Frm five t twelve days, a reamy taste develps whih resembles the flavr f pasteurized milk. Frm twelve t eighteen days, a flat and halky taste develps. After nineteen days, a stale flavr develps. A. Cked Flavr The flavr f UHT pressed milk is relatively pr after prdutin, but imprves within the first few days f strage (Ashtn, 1965). The initial flavr is referred t as "ked" r "sulfurus" whih dereases in intensity during strage giving way t a residual ked flavr (Shipe et al, 1978; Heath, 1983). The initial sulfurus flavr arises frm the frmatin f free sulfhydryl grups released during the denaturatin f the serum prteins, mainly /3-latglbulin f milk (Huttn & Pattn, 1952; Burtn, 1988). This gives it a strng hydrgen sulfide smell desribed befre as "biled abbage." The free sulfhydryl grups are xidized by disslved xygen during the first few days f strage resulting in a derease in intensity f the sulfurus smell and ked flavr (Clarke, 1967). The rate f disappearane depends upn the disslved xygen ntent (Thmas et al., 1975) as 15

27 well as strage temperature (Patrik & Swaisgd, 1976). The xidatin is slwer in diret pressed UHT milk than indiret UHT pressed milk whih has higher levels f disslved xygen available t xidize the free sulfhydryl grups (Jrdan 1968). In indiret-uht pressed milk with an initial high r medium disslved xygen ntent, the sulfhydryl grups dereased rapidly (Thmas et al., 1975). Patrik and Swaisgd (1976) fund that the -SH grups dereased due t xidatin mre rapidly at rm temperature than at refrigeratin temperatures. There are several ways t redue the ked flavr in UHT milk arding t the literature (Badings et al., 1978; Renner and Berlage-Weinig, 1983; Swaisgd, 1980; Ferretti et al., 1974). B. Stale Flavr The flavr f UHT milk begins t deterirate detetably after tw t three weeks strage at rm temperature (Burtn, 1955; Thmas et al., 1975; Shmidt & Renner 1978a). Jen (1976) summarized the fllwing terms t desribe the stale flavr in UHT milk: "xidative ranidity r ardbardy" (Ashtn, 1965), "ranid r xidized" (Zadw and Birtwistle, 1973), and "stale" (Kirk et al., 1968; Thmas et al., 1975). Other researhers have desribed stale flavr as "aged" (Muk et al., 1963), "ld rubber" (Patel et al., 1962), "ereal-stale" (Bassette, 1958), "ardbard" and "gluey" (Henry and Tthill, 1960). Hansen and Swartzel (1982) and Shipe et al. (1978) used the term "lak f freshness" instead f stale. 16

28 Badings and Neeter (1980) islated and identified at least 400 vlatile mpunds frm pasteurized milk and UHT pressed milk by a mbinatin f gas hrmatgraphy and mass-spetrmetry. These mpunds represent vary different lasses f hemial mpunds arding t varius researhers. Carbnyl mpunds (Harper & Huber, 1956), dimethyl sulfide (Bassette et al., 1966), alkanls, fatty aids, latnes, esters, sulfur mpunds, nitrgen mpunds, aliphati and armati hydrarbns (Badings and Neeter, 1980). Jen (1976) identified 8 methyl ketnes (C 3, 5, 7 _ n, 13 ), 6 n-alkanals (C )» 4 aliphati alhls (Q3-6)> 2,3-butanedine, 2-furfurals, benzaldehyde, tluene, ethylbenzene, and methyldisulfides. Sanln et al. (1968) identified C3.5, 7 _n, 13 n-methylketnes, C 8, i> 12 delta-latnes, benzaldehyde, furfural phenylaetaldehyde, vanillin, t-l-en- 3-l, n-heptanal, 2-butxy-ethanl, maltl, aetphenne, benzthiazle and diaetyl. Rerkrai (1986) reprted that aid degree valve (ADV), disslved xygen and titratable aidity rrelated lsely with the stale flavr in UHT milk. Inreases in nentratins f aetaldehyde, prpanal, 2-hexanal, 2-pentanne, 2-hexanne and 2-heptanne paralleled the stale flavr (Rerkrai, 1986). Many psyhrtrphi lipases and prteases are heat stable and are able t hydrlyze lipids ausing liplyzed flavr during strage (Adams and Brawley, 1981; Driesen, 1983; Christen and Wang, 1985; Mttar, 1981). Als, heat stable lipases frm pseudmnas bateria that survived the high temperature may ntribute t liplyzed flavr (Adams and Brawley, 1981; Driessen and Stadhuders, 1974). Bth 17

29 heat resistant prteases and lipases an appear in milk as the result f baterial grwth and lead t ff-flavr develpment during strage (Burtn, 1988). Arding t Shmidt and Renner (1978a), free fatty aids are prdued in strage as a result f the survival f resistant lipases frm either the natural lipases f milk r thse prdued by psyhrtrphi rganisms leads t ff-flavrs. Althugh many mpunds have been identified in UHT milk t date, whih f the many mpunds respnsible fr stale flavr develpment is nt yet knwn. C. Islatin Islatin f the stale flavr mpunds is extremely diffiult beause f the large number f mpunds that have been identified in milk. Many f these mpunds have diverse funtinal grups, lw sensry threshlds, and that ften, flavr mpnents in minute nentratins ntribute mre twards flavr than thse in higher nentratins. 18

30 VII. Fatrs Affeting Flavr f UHT Milk A. Raw Milk Quality Arding t Rerkrai (1986), in rder t prdue UHT milk whih is lse t pasteurized milk in physial, hemial, and rganlepti harateristis, three majr fatrs must be taken int aunt. They are the quality f the raw milk, the quality f the UHT milk immediately after pressing, and the quality f stred UHT milk. Flavr defets due t liplysis an be minimized by reduing the ativities f the serum lipases in raw milk r lipases frm baterial rigin that survive the UHT press. Baterial lipase levels an be restrited by aviding ntaminatin f the raw milk with pseudmnads and lwering the strage temperature befre heat treatment (Burtn, 1988; Zadw, 1980; Ksari et al., 1981). B. Strage time and temperature Chemial reatins and stale flavr develpment in UHT pressed milk inrease when stred at higher temperatures (Mgensen and Pulsen, 1980; Kirk et al., 1968). It has been nfirmed by many researhers that refrigeratin minimizes the hanges whih ur in UHT milk during strage (Bassette and Jen, 1983; Aki et al., 1977; Grner et al., 1977; Hansen and Swartzel, 1982). Zadw (1984) fund that the heat treatment resulted in a rapid inrease in the rate f 19

31 Maillard reatin with the rate ntinuing in strage depending n strage nditins. Burtn (1988) believes that the Maillard reatin urring due t prtein-latse interatin during strage ntributes t the deline in the aeptability f UHT milk. Mehta and Bassette (1980) bserved that stale flavr develped mre rapidly when stred at a higher temperature than a lwer temperature. They als reprted that inreased prpanal, n-pentanal, and n-hexanal paralleled the stale flavr intensity (Mehta and Bassette, 1978). Bassette and Jen (1983) fund inreases in nentratins f vlatile mpunds primarily aliphati aldehydes (aetaldehyde, n-pentanal and n-hexanal) at rm temperature with little hange during refrigeratin temperatures. Similar results were reprted by Wadswrth (1984) and Jen et al. (1978). Rerkrai (1986) reprted that stale flavr f UHT milk develped sner at rm temperature than at refrigerated temperature and that nentratins f aetaldehyde, prpanal, n-hexanal, 2- pentanne, 2-hexanne and 2-heptanne inreased mre at rm temperature than refrigeratin temperature. During strage f UHT milk at rm temperature fr 2-3 mnths, free fatty aids are prdued frm the triglyerides whih lead t detetable flavr hanges (Shmidt & Renner, 1978b). The prdutin f free fatty aids are mst likely aused by the atin f lipases whih have survived the UHT press (Burtn, 1988). This enzymati ativity uld be either frm the natural lipase in milk r frm the liplyti and prtelyti ativity frm the grwth f psyhrtrphi bateria 20

32 in the raw milk befre pressing that survived the UHT pressing. Rerkrai (1986) measured milk fat liplysis f UHT milk using Aid Degree Value (ADV) and fund that hanges in ADV paralleled very lsely t stale flavr develpment. Lipid xidatin may nt be a majr fatr in the staling f UHT pressed milk. Rerkrai (1986) and Wadswrth and Bassette (1985) fund that lipid xidatin values as measured by Thibarbituri Aid (TBA) methd did nt rrelate with stale flavr develpment. C. Disslved Oxygen Thmas et al. (1975) studied the effet f disslved xygen ntent in indiret UHT milk and fund that UHT milk with initial higher xygen ntent resulted in the disappearane f the ked flavr faster than milk with a lwer disslved xygen ntent. They als fund that, althugh the ked flavr disappeared faster, stale and xidized flavrs develped sner. This was nfirmed by wrk dne by Wadswrth and Bassette (1985). UHT pressed milk with lwer initial disslved xygen ntent was less aeptable due t a ked flavr persisting in the milk a lnger time (Rerkrai, 1986). The speed f disappearane f the ked flavr depends n the initial disslved xygen ntent. The reasn behind this is that the disslved xygen xidizes the free -SH grups released during the denaturatin f /?-latglbulin. These free -SH grups are respnsible fr the ked r hydrgen sulfide flavr. Arding t Thmas et al., (1975) free -SH 21

33 grups disappeared in tw days if the disslved xygen ntent is 9 mg/liter, while it tk 3 weeks if the disslved xygen level is 1 mg/liter t xidize 65% f the available free -SH grups. 22

34 MATERIALS AND METHODS Grade A raw, pasteurized, and ultra-high-temperature (UHT) milk samples were btained fr analysis frm the fllwing sures. All f the raw and fresh pasteurized milk samples were prvided by the Kansas State University dairy bar and maintained at a refrigerated temperature. The UHT milk samples in 8 z/236 ml artns were prvided by Real Fresh, In., Vasilia, CA 93277, Plant N and stred at rm temperature. All hemials and standards used in the experiments were analytial grades. Sep-Pak filters used fr the islatin f rgani mpunds frm milk were purhased frm Waters Assiates, Milfrd, MA Other materials and methds will be desribed in the fllwing setins. I. Evaluatin f Slid Phases As a preliminary study, varius slid phase filters frm Waters Assiates, Maple Street, Milfrd, MA 01757, were tested t determine their effet in remving the stale flavr in UHT milk. The fllwing filters were tested: Sep-Pak C 18 Cartridge Sep-Pak Silia Cartridge Sep-Pak Flrisil Cartridge Sep-Pak Alumina A Cartridge 23

35 Sep-Pak Alumina N Cartridge Sep-Pak Alumina B Cartridge Sep-Pak NH 2 Cartridge Sep-Pak CN Cartridge Sep-Pak Dil Cartridge Sep-Pak filters were munted n a 5 ml glass syringe and 2 ml f 95% ethyl alhl was passed thrugh t ativate the filter fllwed by 10 ml f distilled water. Stale UHT milk (17 ml) was slwly passed thrugh the filter at an apprximate flw rate f 3 ml/min. The milk that was passed thrugh the filter was evaluated fr stale arma and flavr by an infrmal but 3-membered experiened taste panel. Results were expressed as either psitive r negative fr the presene f stale flavr and used as a sreening tl. II. Organlepti Analysis A. Intrdutin Based n results f the preliminary study with varius slid phase filters, fresh whle pasteurized milk, UHT stale milk, and UHT stale milk that were eluted thrugh a C18 Sep-Pak and were evaluated by a trained sensry panel fr ked arma intensity, stale arma intensity, ked flavr intensity, and stale flavr intensity. 24

36 B. Training Sessin Referene samples f fresh, stale and ked flavr milk were used in training a 6 membered taste panel t familiarize them with these flavrs. During the training sessins, panelists disussed flavr intensities and harateristis f the samples and agreed upn defets and sres. C. Organlepti Evaluatin f Samples Samples f fresh whle pasteurized, UHT stale, and UHT stale milk filtered thrugh a C18 Sep-Pak were intrdued t a 6 membered taste panel. The samples were evaluated fr ked arma intensity, stale arma intensity, ked flavr intensity and stale flavr intensity. Evaluatins were made n three separate days with stale and ked referene samples prvided eah time. Panelists were asked t judge arma by smelling and flavr by tasting the samples. Cked r stale armas and flavrs were determined r sred using a 5-pint intensity sale (1 = nne t 5 = extremely intense). Cmments were enuraged t desribe ff-armas r flavrs. Fig. 2 prvides an example f the frmat used by the taste panel t sre the UHT milk. 25

37 Cked intensity Briefly Briefly intensity Briefly UHT Milk Taste Panel Sample I.D. arma intensity 1 Stale arma 1 desribe stale arma desribe muthfeel Cked flavr intensity 1 Stale flavr 1 stale desribe flavr Intensity sale 1 nne 2 slight 3 mderate 4 definite 5 extremely Figure 2. Example f the frmat used by the taste panel t sre UHT milk. 26

38 III. Analysis f Free Fatty Aids A. Intrdutin Dupliate milk samples were prepared and analyzed fr the presene and quantity f free fatty aids using a gas hrmatgraphi methd develped by Deeth et al. (1983). All the experiments were repeated twie. A desriptin f these samples is as fllws: Identifiatin Press Date Fresh UHT (Cde N ) 12/12/88 Stale UHT - 7 mnths (Cde N ) 6/8/88 Stale UHT - 20 mnths (Cde N ) 5/23/87 Fresh whle pasteurized 1/15/89 Raw milk 1/15/89 90% Raw/10% Fresh pasteurized 1/15/89 In additin t the desriptin prvided, the fresh UHT milk was examined at ne, tw and three mnth intervals. The tw stale samples were analyzed in bth a filtered and unfiltered frm using a C 18 Sep-Pak filter (filtering predure disussed in Part I, Materials and Methds). The milk ntaining 90% raw milk and 10% fresh whle pasteurized was refrigerated fr eight hurs prir t examinatin. 27

39 Fig. 3 Summary f milk samples analyzed fr free fatty aids I 1 mnth -Fresh UHT -- 2 mnths 12/12/88 I 3 mnths I 7 m. ld unflltered -Stale UHT 6/8/88 «filtered 20 m. ld unflltered Free Fatty Aid Analysis -Stale UHT 5/23/87 I Altered Fresh Whle Past- 90% Raw 10% Fresh Whle Past. Raw Milk- * Press Date 28

40 B. Extratin Predure A gas hrmatgraphi methd fr determinatin f free fatty aids in milk develped by Deeth. et al. (1983) was utilized fr all the samples disussed. Alumina, Neutral, (Brkman Ativity, mesh) (38.4 g) was deativated by adding 1.6 ml distilled water in a 50 ml glass beaker and stirred vigrusly with a glass rd fr apprximately 5 minutes. The alumina was vered with para-film and set aside at least tw hurs befre use. A 10 ml milk sample was added t a 50 ml srew-apped glass tube ntaining a mixture f 20.7 ml ie ld diethyl ether, 3.3 ml HC1, and 1 ml f a trideani aid (C^) standard slutin. The C 13 standard slutin was prepared by weighing 0.10 g C 13 aid using an analytial balane and disslving it int a 100 ml vlumetri flask ntaining diethyl ether. Tw ml f this slutin was withdrawn using a 2 ml pipet and added t a 10 ml vlumetri flask ntaining diethyl ether. One ml f this slutin was added t the milk sample. The milk and diethyl ether - HC1 slutin was agitated gently fr 10 min with a mehanial agitatr and then entrifuged (2,000 rpm) fr 10 min at 0 C using a Bekman Mdel J-21B Centrifuge. After entrifugatin, 15 ml f the ether layer was lleted using a pasteur pipet and transfered t a 50 ml srew-apped glass tube ntaining 15 ml hexane and 1 g sdium sulfate. This was mehanially agitated fr 10 min. A small glass hrmatgraphy lumn (10 mm i.d.) was prepared by paking apprximately 0.5 g f glass wl tightly int the bttm f the tube with a glass rd and adding 1 g f deativated alumina int the tube. The 29

41 ether - hexane slutin was then slwly passed thrugh the lumn at a rate f apprximately 3 ml/min and lleted. The ether-hexane slutin lleted was passed thrugh the lumn a send time. A 1:1 slutin f hexane and diethyl ether was prepared and 5 ml f this slutin was passed thrugh the lumn. Anther 5 ml f the hexane-ether slutin was passed thrugh the lumn and the eluate was disarded. The alumina in the lumn ntaining the absrbed free fatty aids was dried by applying a vauum t the bttm f the lumn utlet fr 5 min. The dry alumina was transferred int a small srew-apped glass tube and 1 ml f a slutin ntaining is-prpyl ether with 6% frmi aid was added. The tube was mehanially agitated fr 10 min and entrifuged at 2000 g fr 5 min, again using a Bekman Mdel J-21B Centrifuge. The isprpyl ether layer was transferred int a small glass vial and 4 /xl was injeted int the gas hrmatgraph. C. GC Analysis Free fatty aids were determined by analyzing the prepared samples using a Hewlett-Pakard 5880A Series gas hrmatgraph equipped with a flameinizatin detetr and Hewlett-Pakard 5880A Series GC Terminal (Level Fur Data System). The fatty aids were separated n a Nukl fused silia apillary lumn with plyglyl liquid phase (15 m x 0.53 mm i.d.) made by Supel, In., Bellefnte, PA. The perating nditins were prgrammed as fllws: 30

42 Oven Temperature Prfile: Initial value 100 C Initial time = 1.0 min Level 1 Prgram rate = C/min Final value = 220 C Final time = min Pst value = 220 C Pst time 3.00 min Other perating nditins: Helium arrier flw rate = 27 ml/min Hydrgen flw rate = 28 ml/min Oxygen flw rate = 430 ml/min Chart speed = 0.5 m/min Attenuatin = 2t x 1 Peak width =.08 Threshld = -1 % Offset = 10 31

43 D. Identifiatin and Quantifiatin The free fatty aids in the milk samples were identified by injeting knwn analytial standards (C 4, C 6, Q, C 10, C 12, C 14, C 16, C 18, C 18 :l) int the gas hrmatgraph and mparing the peak retentin times fr the standards t the peak retentin times fr the milk samples. The nentratins f free fatty aids in the milk samples were determined by taking a rati f the peak area f the free fatty aid in the milk and that f the C 13 fatty aid standard added t the milk during sample preparatin. Beause f the large number f samples and trials as well as nsidering the variability f the data, the areas were adjusted t reflet an average area fr the C 13 standard. This was neessary t allw fr the mparisn f the different hrmatgrams. The fllwing equatins depit the adjustment f peak areas fr eah hrmatgram. (1) Average Area f C 13 Standard Area f C 13 Standard = Crretin Fatr (2) Crretin Fatr X Area f Free = Adjusted Area f Fatty Aid Free Fatty Aid Fur ml f eah sample was injeted int the gas hrmatgraph. A hrmatgram alng with an Area % Cmpensated Analysis Reprt was btained and utilized fr the identifiatin and quantifiatin f the free fatty aids in the samples. 32

44 IV. Analysis f Neutral Vlatile Cmpunds A. Intrdutin Dupliate samples f UHT fresh milk, stale milk, and stale milk filtered thrugh a C 18 Sep-Pak were steam distilled and the distillates analyzed fr vlatile mpunds. The C 18 Sep-Pak filters used t filter the stale milk samples were eluted with rgani slvents t flush ut the trapped mpunds and then the eluants were analyzed. The samples were analyzed by gas hrmatgraphy and mass spetrmetry (GC-MS) t determine the presene and identifiatin f vlatile mpunds. The experiments were repeated three times and a mparisn was made amng the different samples and the differenes in results were evaluated. Fig. 4 depits a flw diagram f the experiments. 33

45 -Fresh UHT --Unfiltered UHT -Filtered UHT Stale UHT-- 7 mnths -Filter Elutin Slvent 1 -Sep-Pak Filtered UHT -Filter Elutin Slvent 2 -Unfiltered UHT -Filtered UHT Stale UHT 20 mnths -Filter Elutin Slvent 1 -Sep-Pak Filtered UHT --Filter Elutin Slvent 2 Figure 4. Samples Analyzed fr Vlatile Cmpunds 34

46 B. Extratin Predures The UHT milk samples were prepared by adding fifty milliliters f milk and 1 ml f a defamer-internal standard slutin int a Kemmerer-Hallet type distillatin flask. The filtered milk samples, hwever, were first passed thrugh a C 18 Sep-Pak and then plaed int the distillatin flask. The samples were steam distilled in a mir-kjeldahl distillatin unit with ie water irulating in the ndenser. The defamer-internal standard slutin was prepared by adding ne gram f defamer (10% ative siline ingredient frm Christian Hansen's Labratry, In., Milwaukee, WI) int a 100 ml steam distillatin flask ntaining frty-five ml f duble-distilled water, whih had been treated with ptassium permanganate. The slutin was gently biled fr 20 min t strip ff and remve the vlatile mpunds present in the defamer. The slutin was led t rm temperature and made t a vlume f 100 ml with duble-distilled permanganate water. This slutin ntained 1000 ppm defamer. Seventy-five ml f the 1000 ppm defamer slutin and 1 ml f a 1-butanl internal standard slutin were mixed and made t a final vlume f 100 ml by adding duble distilled permanganate water. The 1-butanl internal standard slutin was made by weighing g f 1-butanl and disslving int 45 ml duble distilled permanganate water. This was made t a vlume f 100 ml and 1 ml f this standard slutin was added t the defamer making a ttal slutin f 750 ppm defamer and 37.5 ppm 1-butanl. 35

47 Five milliliters f distillate were lleted in apprximately five minutes using a 15 ml graduated nial test tube plaed in an ie water bath. The distillate was rapidly remved and pured int a separatry funnel ntaining 20 ml f redistilled ld ethyl ether maintained in the refrigeratr. This predure was repeated a ttal f twenty times, lleting apprximately 100 ml f distillate in the separatry funnel with ld ethyl ether fr eah individual sample. The distillate and ether slutin were mixed gently by rtating the separatry funnel bak and frth (180 angles) fr apprximately five minutes. The slutin was allwed t stand fr ten minutes in the refrigeratr and then the water phase (bttm layer) was separated frm the ether phase (tp layer) by draining the water frm the bttm f the funnel. The ethyl ether phase (10-15 ml) was then plaed in a 25 ml glass beaker ntaining 1.0 g f anhydrus sdium sulfate t remve exess water. The ethyl ether was pured int a 15 ml graduated nial test tube and nentrated dwn t 1 ml, 0.3 ml, and finally, 0.1 ml nentratins by evaprating the ether with a gentle nitrgen stream. The nentrated samples were transferred int a 1 ml graduated serum vial and 1 mirliter was injeted int a gas hrmatgraph and mass spetrmeter. A ttal f 60 Sep-Paks were used during the preparatin f the Sep-Pak filtered UHT milk samples. These Sep-Pak artridges were eluted with rgani slvents and the eluates were analyzed fr the presene f mpunds that were trapped in the filter. The 60 Sep-Pak artridges were eah eluted with 5 nil methylene hlride (Slvent N. 1) fllwed by 5 ml ethyl ether (Slvent N. 2) and 36

48 lleted in tw separate beakers. The slvents with extrated mpunds frm the Sep-Pak filters were nentrated dwn t 1 ml, 0.3 ml, and 0.1 ml, respetively, and 1 mirliter samples were injeted int the gas hrmatgraph and mass spetrmeter. C. GC and MS Analysis Neutral vlatile mpunds were determined by analyzing the prepared samples n a Hewlett-Pakard Mdel 5880A Gas Chrmatgraph equipped with a flame inizatin detetr as well as a Hewlett-Pakard Mdel 5970B Mass Seletive Detetr mbined with a Hewlett-Pakard Mdel 5890A Gas Chrmatgraph. A 5% phenyl methyl siline rsslinked fused silia apillary lumn (50 m x 0.2 mm i.d. x 0.33 /Lim film thikness) manufatured by Hewlett- Pakard was utilized fr the separatin f the vlatile mpunds in bth instruments. The perating nditins fr the GC were prgrammed as fllws: Oven temperature = 40 C Injetin prt temperature = 230 C Detetr temperature = 230 C Oven temperature prfile = Initial value = 40 C Initial time = 1.00 min 37

49 Level 1 Prgram rate 5.00 C/min Final value 100 C Final time 0.10 min Level 2 Prgram rate C/min Final value 180 C Final time 0.10 min Level 3 Prgram rate C/min Final value 210 C Final time Pst value 220 C Pst time 3.00 Helium arrier flw rate 0.7 ml/min Hydrgen flw rate Oxygen flw rate Make-up gas Chart speed 28 ml/min 430 ml/min 20 ml/min 0.5 m/min Attenuatin 2t x Peak width

50 Threshld -1 % ffset 10 The perating nditins fr the GC-MS were prgrammed as fllws: Aquisitin Parameter Slvent delay 7.00 min. Start time Lw mass High mass San threshld a/d samples Sans per send Ttal in 2 Ttal in sale sale

51 Temperature Prgram and Heated Znes Run time: Equilibretin time: Purge time ff: min min min Level Initial Temp Initial Time Rate C /Min Final Temp Final Time Ttal Time Atual Set Pint Limit Oven Inj Prt Transfer line Run Table Valves Splitless Off 0.75 min Grup min Mass Spe On 7.00 min Stp Run min 40

52 D. Identifiatin and Quantifiatin The neutral vlatile mpunds in the milk samples were tentatively identified by analyzing the individual spetrums frm the Ttal In Chrmatgram (TIC) and mparing the spetrums with thse listed in the autmated Natinal Bureau f Standards (NBS) library. These were later nfirmed by injeting analytial standards f aldehydes, ketnes, and free fatty aids frm Supel, In., Bellefnte, PA int the GC and GC-MS. Identifiatin was nsidered as nfirmed when the retentin time f the unknwn mpunds mathed the retentin time f the analytial standards. The nentratin f the neutral vlatile mpunds in the milk samples was determined by mparing the peak area f mpunds in the sample t the peak area f the 1-butanl standard added t the milk during sample preparatin. One mirliter f eah sample was injeted int the GC and GC-MS. A hrmatgram alng with an Area % Cmputed Analysis Reprt was btained and utilized fr the identifiatin and quantifiatin f the neutral vlatile mpunds in the samples. 41

53 V. Stale Flavr Repliatin A. Intrdutin Cmparisns were made amng the GC and GC-MS hemial prfiles f the UHT fresh, UHT stale, and UHT stale milk filtered thrugh a C18 Sep-Pak. Cmpund nentratins were determined fr eah sample by using a referene mpund as an internal standard. The differenes in mpund nentratin between the fresh UHT and the stale UHT milk were determined. Cmpunds refleting signifiant differenes in nentratins were seletively added diretly t fresh UHT milk in an attempt t repliate stale flavr develpment. Samples were evaluated fr stale flavr develpment by a trained taste panel. B. Repliatin Predure GC and GC-MS hemial prfiles were used t determine mpund identifiatin and nentratins fr UHT fresh and UHT stale milk. Differenes in mpund nentratins between the UHT fresh and UHT stale milk were determined in the fllwing manner: UHT Stale Milk UHT Fresh Milk Cnentratin - Cnentratin = Differene (p.p.m./p.p.b) (p.p.m./p.p.b) (p.p.m./p.p.b) 42

54 Cmpunds refleting signifiant differenes in nentratins were added diretly t fresh UHT milk t investigate their effet n stale flavr develpment. The fllwing samples were prepared: Identifiatin Cmpund added Cnentratin *ppm/ppb Sample 1 tani aid (C 18) *34.0 Sample 2 deani aid (C 10 ) *80.0 Sample 3 tani aid (C 8 ) *34.0 deani aid (Ci) *80.0 Sample 4 n-hexanal heptanne 60.0 n-heptanal nnanne 50.0 n-nnanal butanne deanne undeanne trideanne 1.0 Sample 5 tani aid (Q) *34.0 deani aid (Ci) *80.0 n-hexanal heptanne 60.0 n-heptanal nnanne 50.0 n-nnanal butanne deanne undeanne trideanne

55 C. Organlepti Analysis UHT fresh milk samples with added nentratins f seleted mpunds were evaluated by a trained six member taste panel. Stale and fresh UHT milk samples were prvided as referenes fr their arma and flavr harateristis. The taste panel evaluated the five samples in three separate sessins ver a three day perid. Stale and fresh UHT samples were prvided as a referene during eah sessin. Panelists were asked t judge arma by smell and flavr by tasting the samples. Stale armas and flavrs were sred using a 5-pint intensity sale (1 = nne t 5 = extremely intense). Cmments were requested fr stale armas and flavrs. The frmat used t sre the samples is shwn in Figure 2. 44

56 RESULTS AND DISCUSSION I. Evaluatin f Slid Phases Results frm the infrmal taste panel are prvided in Table 3. Table 3. Results frm the evaluatin f slid phases. Type Slid Phase Result Sep-Pak Alumina A Cartridge Sep-Pak Flrisil Cartridge Sep-Pak Silia Cartridge Sep-Pak C 18 Cartridge + Sep-Pak Alumina N Cartridge Sep-Pak Alumina B Cartridge Sep-Pak NH 2 Cartridge Sep-Pak CN Cartridge Sep-Pak Dil Cartridge (+) = stale flavr remved (-) = stale flavr nt remved The data indiates that when passing stale UHT pressed milk thrugh C 18 Sep-Pak artridges, the stale flavr was remved. The ther filters that were tested shwed n derease in stale flavr intensity. This finding suggests that the stale flavr mpnents in UHT pressed milk were entrapped in the C 18 Sep-Pak 45

57 artridges. By utilizing this tehnique, it was then pssible t analyze the differenes in hemial mpsitin between UHT stale and UHT stale filtered milk. Als, the mpunds entrapped in the C 18 Sep-Pak filter, thught t be respnsible fr stale flavr develpment, uld als be subjeted t analysis. The C 18 Sep-Pak artridge therefre prvides a signifiant step frward tward the suessful islatin and subsequent identifiatin f the stale flavr mpnents in UHT pressed milk. II. Organlepti Analysis Results frm the rganlepti analysis f fresh pasteurized milk, stale UHT milk, and stale UHT filtered thrugh a C 18 Sep-Pak are shwn in Table 4. Table 4. Means f rganlepti flavr sres. Sample Defet Pasteurized Whle UHT UHT Sep-Pak Trt Sres 1 Cked Arma Intensity Stale Arma Intensity Cked Flavr Intensity Stale Flavr Intensity Sres: 1 = nne t 5 = extremely intense 46