Flavor Analysis - American Chemical Society

Chapter 5

Downloaded by STANFORD UNIV GREEN LIBR on September 28, 2012 | http://pubs.acs.org Publication Date: September 10, 1998 | doi: 10.1021/bk-1998-0705.ch005

Relationship Between Aroma Compounds' Partitioning Constants and Release During Microwave Heating Deborah D. Roberts and Philippe Pollien Nestle Research Center, Vers Chez les Blanc, 1000 Lausanne 26, Switzerland

A method was designed and tested to quantitate aroma compounds eluting from food heated in the microwave. It was readily able to measure aroma compounds present at 4 mg/kg in the meal. The trap and condenser glassware were designed to accommodate the high air and water flow rates, in effect using the distillant nature of microwave heating. SPME was an effective method to quantitate aroma compounds from the trapped fractions. When analyzing frozen spaghetti, aroma compounds were found to elute at different times during heating. The initial release starting time correlated (R = 0.98) with the compounds' air-water partition coefficients. Thus, predictions for which types of compounds will be more readily lost can be made and aroma formulations for microwaveable food can be modified accordingly. 2

The use of a microwave to reheat foods is highly popular, especially in the U.S. The convenience and quickness of preparation have instigated a range of food products for microwave heating. The mechanisms for heating food in the microwave are different from the conventional oven and can cause differences in the flavor of the prepared food. Due to the differences in the surface moisture and temperature, conventionally cooked foods can form a crust at the surface which reaches high temperatures and acts as a barrier to water and aroma compound evaporation. In microwave food, this crust is usually not formed because of the 100°C maximal temperature and high surface moisture. Thus, there is significant water and aroma migration to the surface and into the air. These losses can cause microwave foods to have less desirable flavor. A recent ACS Symposium Series book (7) covers the challenges of flavoring microwave food. In this chapter, a method of aroma trapping is demonstrated that gives information about aroma losses and their relationship to aroma compound properties.

©1998 American Chemical Society In Flavor Analysis; Mussinan, C., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1998.

61

62


Materials and Methods Sample Preparation. Buitoni brand spaghetti (1700 g) broken up into 7 cm pieces was immersed in 11.8 L of boiling water and cooked for 8 minutes. After draining and rinsing with cold water for 2 min, the final weight was 3992 g. The aroma mixture contained (a-l): dimethyltrisulfide, ethyl-2methylbutyrate, 2E-nonenal, l-octen-3-ol, diacetyl, and 2,3-diethyl-5methylpyrazine at 3 g/kg dissolved in medium chain triglycerides (MCT), containing approximately 60 % caprylic and 40 % capric acid. The aroma compounds were obtained from commercial sources.

a

d

b

e

e

f

The aroma mixture (0.285 g) was dispersed on top of 200 g of spaghetti in a plastic container and sealed. This gave a final concentration of 4.275 mg/kg in the spaghetti. Samples were frozen at - 25 °C until microwave oven analysis. Eight samples were analysed whose frozen storage times were 2, 3, 5, 11, 13, 14, 16, and 17 days. Device to Analyze Aroma Released in Microwave. A Panasonic Genius N N 5852 microwave oven with turntable, 23 L volume, 1420 W required power, and 100-900 W output power was used. The actual power output at full power was found to be 711 watts, determined by heating 1 kg water (28 °C) for 1 minute, and measuring the increase in temperature. The microwave oven was modified by placing a hollow cylinder in the center of the top cover whose dimensions prevent leakage of microwave energy. Specially designed glass vessels, as shown in Figure 1, allowed the trapping of aroma released in the microwave. A l l glassware was silanized with Sylon CT (5% dimethyldichlorosilane in toluene, Supelco) to avoid adsorption of aroma compounds. The food vessel (19 cm in diameter by 5 cm high) was glued to the center of the turntable to allow reproducible placement of the food. A Teflon washer with Viton o-rings, which connected the food vessel to glass tubing, allowed the food vessel, but not the glass tubing, to turn during microwave heating. The glass tubing was connected to a three-way valve that allowed the air stream to be directed to either the liquid nitrogen trap, or the condenser. The liquid nitrogen trap was placed in this configuration, instead of after the condenser, to avoid a partial

In Flavor Analysis; Mussinan, C., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1998.

63


trapping of the initial aromas on the condenser. The glass trap was placed in a Dewar of liquid nitrogen and its exit was connected to a deflated plastic bag which prevented oxygen in laboratory air from condensing in the trap. The liquid nitrogen trap and the condenser incorporated serpentine designs which maximized surface area and time for contact of the air flow with the cold surface. The condenser was kept at -5 °C and the flasks, in ice, were rotated to allow multiple fraction collections. Microwave Analysis Procedure. Immediately before analysis, liquid nitrogen was poured into the liquid nitrogen trap and alcohol coolant was circulated in the condenser. The entire system was filled with nitrogen gas to prevent oxygen condensation in the liquid nitrogen trap. The sample was placed in the food vessel, sealed, and heated for a total of 255 sec. During the first 165 sec, when the temperature was less than 100 °C, the air stream was collected in the liquid nitrogen trap. Then, water began to elute and the air stream was diverted into the condenser. Two consecutive samples were collected (165 - 225 sec, 225 - 255 sec). The liquid nitrogen trap was rinsed with 5 mL of water to recover the trapped aroma. A l l fractions, which were aroma solutions in water, were weighed. The amounts of aroma compounds in the solutions were determined by SPME. The experiment was repeated eight times in its entirety with separate spaghetti samples. Solid Phase MicroExtraction (SPME) - GC/FID Quantification. Six external standard curves for quantitation were produced using standard solutions in water, one for each molecule of interest. The most concentrated solution was prepared by dissolving the aroma compounds in water with continued agitation. This solution was diluted to produce five different concentration levels which included the concentration range of the samples. These ranges were (in mg/L) l-octen-3-ol (0 16), diacetyl ( 0 - 28), ethyl-2-methyl butyrate (0 - 13), 2,3-diethyl-5methylpyrazine (0 - 19), and 2E-nonenal (0 - 17). As most of the curves were reproducibly slightly curvilinear, a quadratic function was used for regression with R values of at least 0.999. The standard samples and the microwave elution samples (all flavor solutions in water) were analyzed in the same manner using the following procedure. 3 mL were placed in a 4 mL septum-closed vial and stirred with a stir bar (800 rpm). A carbowax/divinylbenzene fiber (Supelco) was immersed for 10 minutes. A trial with 10 % sodium chloride addition was also performed. The fiber was then desorbed for 5 minutes at 200 °C in the G C injection port containing a 0.75 mm ID liner. During the 5 minutes, the volatiles volatilized almost immediately but the fiber remained for conditioning. Good G C resolution was obtained due to the narrow liner and quick desorption. The GC/FID contained a D B W A X column: 30 m, 0.32 mm, 0.25 \im film thickness, 10 PSI; and was programmed to begin at 50 °C for 3 minutes, heat to 140 °C at 8 °C/min and then to 220°Cat25°C/min. Temperature of Meal. The heating profile of the spaghetti dish was obtained by placing fiberoptic probes at different places in the spaghetti and measuring the


64 temperature during microwave heating, using the same microwave as in the release analysis. Eight measurements were taken from different places in the meal. Determination of K . A method that does not require the use of internal or external standards was used to determine the air-water partition coefficients (Ka ) of the compounds studied (2). One or two mL of sample were equilibrated with 312 mL of headspace at 30 °C in an enclosed stainless steel sampling cell. After a 30 min. equilibration period, the contents of the headspace were pushed onto a Tenax trap, which was subsequently thermally desorbed using an A T D 400 (Perkin Elmer, Beaconsfield, U.K.). Figure 2 shows how the partition coefficient is calculated. a w


W

Results and Discussion Method Development. In order to study aroma release kinetics during microwave heating, a method was developed based on the distillant properties (3). After the attainment of maximal temperature, usually close to 100 °C, a large amount of water evaporates from high moisture food. This leads to dehydrated foods with a high amount of surface moisture. It also causes problems if trying to trap the aroma compounds released. In the case of frozen spaghetti, 15 mL of water evaporated per minute during the last few minutes of heating. This would be too much for a thermally desorbed adsorbent trap. The method developed, therefore, used a series of cold traps especially designed with a large surface area. Liquid nitrogen was used as the coolant for trapping during the first few minutes. When water began to distill, the eluting air-stream was diverted into a condenser in which water and flavor compounds were condensed together. Both traps were very effective in trapping the compounds as traps connected in-series recovered very little. After the experiment, trap 1 was removed from liquid nitrogen and rinsed with water to recover the aroma compounds. Unfortunately, it also trapped oxygen from the air which quickly volatilized as it returned to ambient temperature. This could have resulted in some losses of aroma. The SPME method was found to be more sensitive than direct G C water injection for compound quantitation. The use of solvents was avoided, allowing analysis of early eluting compounds. The solutions had to be in the concentration range of aroma compound solubility. One of the key advantages of this method is that it allowed a measurement of released aroma compounds, rather than analysis of the food before and after microwave heating. Another method, developed for popcorn, offered the same advantage (4) by using purging gas to strip water and volatiles from food during microwave heating with retention in a cold trap. This trap was then heated to transfer the volatiles to an adsorbent trap, followed by trap heating and G C analysis. The method described here did not require a purging gas and it offered an alternate method (SPME) for quantitation of trapped aroma compounds. Loss amounts during storage and release. The 8 replicates were sampled with differing storage times, up to 17 days. For most of the aroma compounds, the same results were obtained thoughout this storage period. However, the two compounds with the highest Ka showed a weak relationship between storage time and amount W


65

Hollow Cylinder 1


Teflon Washer Oven Cavity

200 g Spaghetti Aroma AddedtoFood and Frozen I 6 compounds, 4 ppm

Quantitation by Solid Phase Microextraction and GC-FID

Figure 1. Apparatus and method used for aroma release analysis during microwave heating (Adapted from ref. 3).

slope 2 m L

900000

a =• slope 1 m L

600000 •^...-•""lml

O 300000

(l-«)V, ,V. 2 i q

( 2

4

6

8

10

a

^liq2 " V

l i q l

i q

) V gas

12

Concentration [mg/kg] Figure 2. Example data (ethyl-2-methylbutyrate) and calculation for Kaw


66


released (Figure 3). This was seen for the first trap, where highest amounts of these compounds were collected. The loss during storage for some but not all compounds shows that frozen food can undergo a selective volatilization of compounds, resulting in a product with a different aroma than the freshly prepared counterpart. In addition, the aroma compounds showed large losses during heating. Respectively for compounds a-f: 27, 27, 33,46,41, and 44 %. For a and b, these values were probably larger due to losses from the liquid nitrogen trap. These amounts are similar to what was reported from frozen pancakes (10-56%) (5). Table I. Relationships between the values of K , coefficient) and the starting release time

(air-water partition

w

Temperature of meal at T

T Extrapolated Release starting time, (from Figure 4) 62 16 2,3-diethyl-5-methylpyrazine 5.0 x 10 4 61 diacetyl 1.1 x 10" 53 3.1 x 10" 1 l-octen-3-ol 2 49 2E-nonenal 7.0 x 10 20 ethyl-2-methylbutyrate 1 1.5 x 10" the first to be dimethyltrisulfide 1.9 x 10 1 released (concentration air/concentration water) at equilibrium Compounds

Kaw RSD

Kaw

0

0

95 94 78 71 19

-4

3

3

-3

2

-2

Relationship between estimated release starting time (t ) and K . Figure 4 shows that the aroma compounds were released at different times during microwave heating. The linear relationship between release amount and time of heating indicates that the compounds exhibited a constant release over heating. The point at 225 s of heating was low for ethyl-2-methylbutyrate probably due to losses of this compound in the condenser before the water began to elute. The x-intercept for the release corresponds to the point when the particular aroma compounds began to elute. These points were estimated and are found in Table I. The estimations are more precise for the later eluting compounds than for the earlier eluting compounds. The aroma compounds exhibited different release kinetics, as seen by these estimated times of release commencement, and the differences followed the measured air-water partition coefficients of the compounds. Dimethyltrisulfide and ethyl-2-methylbutyrate were the first two compounds to be released and are also those with the highest air-water partition coefficients. These compounds started to be released early in the heating of the frozen spaghetti, when the average temperature in the meal was still below room temperature. 2E-nonenal was the next compound to elute and also exhibits the next lowest Ka . l-Octen-3-ol followed in time and also had the next lowest Ka . And lastly, diacetyl and 2,3-diethyl-5methylpyrazine eluted primarily at the end of heating and had the lowest Ka . These compounds, which have low volatility in water, did not begin their elution until the meal reached its maximum temperature of close to 100 °C. Thus, for a meal that has low fat content, the K was found to be an excellent approximation of 0

a w

W

W

W

a w


67

Dimethyltrisulfide

Ethyl-2-methylbutyrate


| 200 ua

200 ug

I 10

15

15

20

20

Storage Time (days) Figure 3. Relationship between release during microwave heating (0-165 s) and storage time of spiked frozen spaghetti.

65

130

195

Time of Heating (s) Figure 4. Graph showing the total cumulative release percent for six aroma compounds (a-f) and their relationship to the temperature of spaghetti. The line begins at 165 s because the first trap was from 0 - 165 s.


68 2


the microwave elution kinetics. A high correlation was found (R = 0.98) between the air-water partition coefficient and time of elution commencement (Figure 5) for this low-fat food. Previously, compound parameters such as the air-product partition coefficient (6) or the Henry's law constant were used (7) to predict losses during microwave heating. This work shows that the partitioning parameters, in addition, predict the kinetics of release.

0

25

50

t

0

75

(sec)

Figure 5. Relationship between the extrapolated release starting time (to) and the air-water partition coefficient for compounds (a-f).

Literature Cited 1. Thermally Generated Flavors: Maillard, Microwave, and Extrusion Processes; Parliment, T.H., Morello, M.J., McGorrin, R.J., Eds.; American Chemical Society: Washington, D.C., 1994; 2. Chaintreau, A.; Grade, A.; Munoz-Box, R.Anal.Chem. 1995, 67, 3300-3304. 3. Roberts, D.D.; Pollien, P. J.Agric.Food Chem. 1997, 45, 4388-4392. 4. Risch, S.J.; Keikkila, K.; Williams, R.J. U.S.Patent 5,177,995. Analysis of Migration of a Volatile Substance During Heating with Microwave Energy. 1993. 5. L i , H.C.; Risch, S.J.; Reineccius, G.A. In Thermally Generated Flavours: Maillard, Microwave, and Extrusion Processes; Parliment, T.H., Morello, M.J., McGorrin, R.J., Eds.; American Chemical Society: Washington, D.C., 1994; pp 466-475. 6. de Roos, K.B.; Graf, E. J. Agric. Food Chem. 1995, 43, 2204-2211. 7. Lindstrom, T.R.; Parliment, T.H. In Thermally Generated Flavours: Maillard, Microwave, and Extrusion Processes; Parliment, T.H., Morello, M.J., McGorrin, R.J., Eds.; American Chemical Society: Washington, D.C., 1994; pp 405-413.


Flavor Analysis - American Chemical Society

Recommend Documents