DETERMINATION 0F CARBONYL
COMPCUNDS DEVELOPED DURENG STORAGE
OF FREEZE-DRIED BEEF

Thesis {or Hm Degree of M. 5.
MlCfiISfin‘ STATE UE‘HVERSH’Y
Toshio I-L Saga
1965

114E515

 

LIBRARY

Michigan State 4
University

 

.‘. rll ' Jul

ABSTRACT

DETERMINATION OF CARBONYL COMPOUNDS DEVELOPED
DURING STORAGE OF FREEZE-DRIED BEEF

by Toshio H. Saga

This study has been undertaken to investigate
whether the determination of carbonyl compounds can measure
progressive oxidative changes in stored freeze-dried meats
in relation to: degree of change with time, role of mono-
carbonyl compounds, and the effectiveness of certain anti-
oxidants and their method of application.

For the carbonyl analysis, the Henick method and the
Keith and Day method were employed with some modifications.
The major modifications involved attempts to extract and
measure the carbonyls of the protein bound lipids in freeze—
dried meat. Both procedures involved colorimetric determi—
nation of carbonyls based on the formation of a colored
quinoidal ion of 2,4-dinitrophenylhydrazone derivatives in a
basic solution, and calculation by substituting the spectro-
photometric data for the quinoidal ion into simultaneous
equations. Thus, the major classes of carbonyls were dis-
tinguished through calculation by means of the equations
from the data for a single solution without actual fraction-

ation into the separate components.

Toshio H. Saga

The accumulation of carbonyl compounds in the stored
freeze-dried beef was observed to occur to variable degrees.
However, the increase of carbonyl compounds could not be
definitely correlated to other factors such as the effective—
ness of the antioxidants and their method of application.

Due to the complexity of the autoxidation process,
it appeared that no simple analytical method could be
directly applied to the evaluation of oxidative change in

freeze-dried meat with complete satisfaction.

DETERMINATION OF CARBONYL COMPOUNDS DEVELOPED

DURING STORAGE OF FREEZE-DRIED BEEF

BY

Toshio H. Saga

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Food Science

1965

ACKNOWLEDGMENTS

Sincere gratitude is extended to Dr. L. R. Dugan, Jr.
for his unfailing help and guidance throughout the course of
this study and the preparation of this thesis.

Acknowledgment is given to Dr. L. E. Dawson, Pro-
fessor of Food Science, and Dr. E. J. Benne, Professor of
Biochemistry, for their critical reading of this manuscript.

The author greatly appreciates the encouragement of
Dr. B. S. Schweigert, Chairman of the Food Science Depart-
ment, and Mr. K. Hokketsu, President of The Kyokuyo Hogei
Co., Ltd.

It is with the utmost gratitude that this thesis is
dedicated to my parents and members of my family for their

encouragement and support.

ii

TABLE OF CONTENTS

IMRODUCTION o O 0 O O O 0 0 O O 0 O O O 0 O 0 O 6 O 0

REVIEW OF LITERATURE . . o

‘0 O O O O O O O 0 O O O O 0

Process of Oxidative Rancidity of Lipids
Relationship of Rancidity to Carbonyl Value
Flavor Thresholds of Monocarbonyl Compounds
Analytical Methods for Carbonyl Content
Stability of Freeze-Dried Meats

EXPERIMENTAL METHODS AND PROCEDURE . . . o o o o . . .

Preparation of Reagents

Extraction of Carbonyls from Freeze-Dried
Meats

Determination of Crude Fat Content

Determination of Moisture

Procedure of Cooking and Freeze-Drying

Experiment I. Determination of Carbonyl Compounds
by a Modification of Henick's
Method

Determination of Two Classes of Carbonyls

Experiment II. Changes of Carbonyl Content in
Freeze-Dried Raw and Cooked Beef
during Storage at Room Temperature

Sample Preparation for Experiment II

Experiment III. Determination of the Classes of
Free Monocarbonyl Compounds in
Oxidizing Lipids of Freeze—Dried
Beef using Benzene—Ethanol
Extracts

Determination of Three Classes of Monocarbonyls

Preparation of a Model System with Corn Oil
Incorporated in Freeze—Dried Egg White

iii

U)

KO\]\IU'|UJ

12
12
14
15

15
15

l6

l7

l8

18

20

22

Experiment IV.

Determination of the Classes of
Free Monocarbonyl Compounds in
Oxidizing Lipids of Freeze-Dried
Beef using an Acid-Benzene Mixture
as Extracting Means

Extraction of Carbonyls with an Acid-Benzene

Mixture

RESULTS AND DISCUSSION . . . . . . . . o . o . . . . .

CONCLUSIONS

BIBLIOGRAPHY

APPENDICES

0

O

O

0

O O O O O O O O O 0 O O O O O O 0

O C O O O 0 O 0 O O O O O O O O 0

0 . O C O O O O O 0 O O O O O O 0

iv

Page

23

23
25
37
4O

46

LIST OF TABLES

Table Page
1. Standardization of Henick method with
hexanal O O O O O O 0 O O O O 0 0 O O 9 ° 0 27
2. Carbonyls of freeze-dried cooked beef by the
modified Henick method . . . . . . . . . . . 27
3. Changes of carbonyl content in freeze—dried

beef during storage in aluminum foil at
room temperature (23 i 2 C.) . . . . . . . . 28

4. Changes of free monocarbonyls in oxidizing
freeze—dried raw beef stored at room
temperature . . . . . . . . . . . . . . . . 32

5. Monocarbonyls in crude fats extracted from
freeze-dried raw beef by an acid-benzene
mixture and by benzene from a water slurry . 35

LI ST OF APPENDICES

Appendix Page

I Scheme for Formation of Hydroperoxides and
Dismutation to Monocarbonyl Compounds . . . 47

II Reaction of DNP-Hydrazine with a Carbonyl
Compound and the ChromOphoric Quinoidal
Ion DevelOped in the Presence of Base . . . 48

III Hydroperoxides and Aldehydes Derived from
Fatty Acids by the Scheme in Appendix I . . 49

IV Common and Chemical Designations of
Antioxidants . . . . . . . . . . . . . . . . 51

V Derivation of Simultaneous Equations by
Matrix Inversion from Average Molar Ab-
sorptivities of DNP-Hydrazone Derivatives
of Monocarbonyl Compounds . . . . . . . . . 52

vi

INTRODUCT ION

Rancidity, which can be defined as any off—odor or
flavor developed in lipid or fatty portions of food, degrades
the quality of food during storage. In this study, this
term is confined to oxidative rancidity by atmospheric oxygen.

The history of objective tests for rancidity begins
many years ago with development of the Kreis test (35). Al-
though numerous workers have attempted to make quantitative
evaluations of rancidity very few of them have proved to be
satisfactory. Among the many methods attempted, the carbonyl
test appears to have promise for this purpose (12). Carbonyl
compounds, especially volatile monocarbonyls, have been
directly related to flavor deterioration of oxidizing
lipids (5, 19, 33, 43).

In this study the carbonyl content was determined in
several samples of freeze-dried raw and cooked beef. Freeze-
drying is considered one of the most promising methods for
dehydrating and preserving food (7, 22). Freeze-drying,
while it protects against microbial spoilage because of the
low moisture content, does not prevent the development of
oxidative rancidity in the lipid components.

In the first two experimental sections, Henicks'

procedure (24) with some modification was employed to

determine carbonyls in freeze-dried meat. The method of
Henick differentiates saturated carbonyl from (xlg-unsatu-
rated carbonyl by means of substituting absorbancy values at
430 and 460 m/(into equations. In the last two sections, a
modification of the method developed by Keith and Day (33)
was applied to freeze-dried beef and to a model system with
corn oil on freeze-dried egg white. The Keith and Day
method makes it possible to distinguish and account for
three major classes of free monocarbonyl compounds using
simultaneous equations with spectrophotometric data obtained
on derivatives of carbonyls. Both procedures (24, 33) in-
volve a colorimetoric method for carbonyl compounds based on
the formation of a colored quinoidal ion of 2,4—dinitro-
phenylhydrazones* of the carbonyl compounds in a solution of
base (36).

The aim of this study has been to learn whether the
application of the methods for carbonyl compounds can measure
the progressive oxidative changes in stored freeze-dried
meats in relation to: degree of change with time, role of
monocarbonyl compounds, and the effectiveness of certain

antioxidants and their method of application.

 

*
2,4-Dinitrophenyl is hereafter abbreviated DNP.

REVIEW OF LITERATURE

Process of Oxidative Rancidity
of Lipids

Among lipid constituents of foods, those containing
polyunsaturated fatty acids are most susceptible to autoxi-
dation or spontaneous oxidation by atmOSpheric oxygen (23,
28). The generally accepted theory for lipid autoxidation
is that the oxidation of the unsaturated fatty acid is initi-
ated by the attack of molecular oxygen at a double bond to
form a diradical. This diradical activates d—methylene
groups of other unsaturated systems, and oxygen attack at
these sites leads to the formation of allylic hydroperoxides
(3, 16).

It is commonly known that the formation of the hydro-
peroxides develops through a chain reaction that is con-
sidered to occur in three stages, namely initiation, propa-
gation and termination, involving free radicals. The energy
necessary to form the very first radical may come from heat,
light, or ionizing radiation.

A characteristic feature of lipid oxidation is the
ease with which it can be influenced by factors other than

the usual ones of concentration of reactant and temperature.

The kind and quantity of unsaturated fatty acid in
the lipids affect the rate and course of development of
rancidity. Phosphatides which usually contain a greater
amount of polyunsaturated fatty acids than the associated
neutral lipids are more readily subjected to oxidation (37,
46, 62, 63). An anomaly results from the fact that phospha-
tides such as lecithin may also act as an antioxidant or
synergist (45). This phenomenon may result if the phospha-
tides react first with any pro—oxygenic substance present
and effect the removal of this substance from the oxidizing
lipid system (18). Some other so-called synergists such as
ascorbic acid may have both functions, that is, they some-
times act as an antioxidant and sometimes as a pro-oxidant,
according to various conditions such as pH, concentration of
synergist, and so on (41).

Factors which inhibit lipid oxidation are low temper-
ature (59), the presence of antioxidant and synergist (8, 34,
50), the exclusion of oxygen, light, and ionizing radiation
(4, 13, 60), low moisture levels (17, 30, 59), and exclusion
of trace metals and salts. High temperature (59), exposure
to oxygen, light, and ionizing radiation (4, 13, 60), the
presence of enzymes and catalysts such as lipoxidase (49),
hematin compounds (58), and metals, etc., are all considered
to play a role in accelerating the oxidation of lipids.

The hydroperoxides are the primary products of lipid

autoxidation. The so-called secondary degradation products

are formed by the dismutation of hydroperoxides (2, 3).
Carbonyl compounds are typical secondary products which are
usually unstable and may decompose or react further (32).

Badings (3) has published a review of the principles
of the autoxidation process in lipids. Appendix III shows
monocarbonyls or aldehydes resulting from cleavage of mono-
meric hydroperoxides formed during autoxidation of different
unsaturated fatty acids. Appendix I illustrates the scheme
of formation of aldehydes from linoleic acid by a concept
first postulated by Bell et a1. (3). Aldehydes formed from
other unsaturated fatty acids by this scheme are those in
Appendix III. These hypothetical components have been
proved by many researchers (2, 15, 20, 44, 47) to exist in
mildly oxidized esters of unsaturated fatty acids or natural
fats and oil. Gaddis _t__l. (20) have shown that appreciable
amounts of heptanal can be isolated from autoxidized lamb
and beef fat. They also found heptanal and decanal in oxi-
dizing butterfat, although these could not be theoretically
predicted from the four unsaturated fatty acids in Appendix
III (40).

Relationship of Rancidity to
Carbonyl Value

 

 

Although organoleptic evaluation is the ultimate
standard of rancidity of foods, it does not lend itself to
quantitative measurement or reproducibility. Determination

of the degree of rancidity of stored food may indicate the

stability or keeping quality of the food. The most commonly
used objective tests to determine the degree of rancidity
are: TBA test (7, 9, 10, 54, 64), carbonyl test (5, 24, 33,
36, 42, 48, 53), and peroxide value (1). Among these the
carbonyl test appears to be one of the most promising for
this purpose, since odors or flavors associated with typical
oxidative rancidity are mainly due to carbonyl type compounds
(12). The greater part of the carbonyls determinable by the
conventional methods (24, 36) are non-volatile carbonyls (5,
19, 38). It has become apparent that many of the carbonylic
substances which accumulate as secondary oxidation products
are non—volatile and that, like the hydrOperoxides, these
make little or no major contribution to off-odor or flavor,
although they are probably precursors of volatile odorous
substances and may well have deleterious effects on stability
(5, 12, 19).

McKerrigan (43) reported that volatile carbonyl shows
a better correlation with flavor than either the peroxide or
total carbonyl value. Volatile carbonyl compounds are
usually collected by vacuum or steam distillation (5, 19).

Much attention has recently been directed toward the
separation, estimation, and identification of the volatile
carbonylic degradation products from oxidizing lipids (21,

33, 44, 53, 55, 61).

Flavor Thresholds of Mono-
carbonyl Compounds

 

 

Flavor thresholds for a series of aliphatic alde—

hydes have been determined by Lea gt 1. (39). It was found

that the C8 - C12 9

detectable in water at dilutions as great as one part in 10

2-enal were
8

saturated aldehydes and the C
or 109. This level is similar to the value reported by
Patton_§tia1. (47).

Analytical Methods for Carbonyl
'Content

 

A widely used analytical method for carbonyl content
is the colorimetric DNP—hydrazine procedure of Henick _t _1.
(24). This procedure is a modification of the method intro-
duced by Lappin and Clark (36). The major modifications are
the substitution of trichloroacetic acid for hydrochloric
acid and the use of benzene as a solvent to make this method
universally applicable to fats. The reaction of DNP-
hydrazine with a carbonyl compound, and the wine-red chromo-
phoric quinoidal ion developed in the presence of a base are
shown in Appendix II (36). Since this color is unstable and
the absorbancy decreases with time at a rate of 0.5% to 0.7%
per minute, the time element is very important in determi-
nations using this color reaction (31, 48).

Henick__t.al. (24) determined saturated and cfiala'

unsaturated aldehydes by absorbancy measurement at 430 and

460 my4. The values which were obtained were much higher

than those determined by Pool and Klose (48). Data presented
by Gaddis 33 a1. (19) indicates that decomposition of
carbonyl precursors occurs during the DNP-hydrazine reaction
when using Henick's method (24).

Pool and Klose (48) applied a chromatographic mixture
of carbonyl compounds to a 15% hydrated alumina column im-
pregnated with DNP—hydrazine, and found that this column re—
tained the excess reagent and the derivatives of poly-
carbonyl compounds. Monocarbonyl derivatives were eluted
and determined colorimetrically.

While the procedures of Pool et a1. (48) and Henick
_§,§l. (24) possessed certain desirable features, it was
observed that neither gave an adequate picture of carbonyl
distribution (33). Neither procedure distinguished between
the alkanals, a1k—2-enals and alk-2,4—dienals which have
been found in oxidizing lipids (20, 33).

In 1963, Keith and Day (33) introduced a modification
of the Pool and Klose method. Simultaneous equations were
derived to distinguish between and account for the three
major classes of monocarbonyls (Appendix V). The Keith and
Day method (31) also involved formation of DNP-hydrazones on
an alumina reaction column and determination of quantity by
colorimetry. The values obtained by the procedure have been
referred to as free carbonyl (19). It was found that the

quantitative composition of the DNP—hydrazone derivatives ob-

tained from milk fat by the alumina reaction column were

identical to the volatile carbonyl fraction (33). Moreover,
there was less than 5% difference in the volatile mono-
carbonyl content of milk fat and the quantity measured by
the modified Pool and Klose procedure (33, 48). This pro—
cedure did not measure the carbonyls nor those aldehydes

present but bound through an enol ether linkage (33, 51).

Stability of Freeze-Dried Meats

 

Among the dehydration methods applicable to foods,
freeze—drying has proved to be highly effective for the
maintenance of desirable palatability (7, 22). Some un-
desirable changes such as decreased tenderness and loss of
texture are frequently encountered in foods which have been
freeze—dried.

An important application of freeze—drying appears to
be for meat products, but, unfortunately, they are among the
more difficult foods to freeze-dry successfully (6). Al—
though the primary characteristics of freeze-dried meat,
especially for military and emergency use, lies in its
storage stability, there are deteriorative changes which
occur during extended storage under ordinary conditions.
Deterioration of dehydrated foods due to microbial growth is
not common because of the low moisture content. Oxidation
is involved in most of the undesirable effects which develop
during storage.

Tappel t l. (56, 57) have studied deteriorative

changes in freeze—dried meats. The most labile components

10

are lipids, particularly phosphatides with their high content
of polyunsaturated fatty acids, and protein bound lipids,
which readily undergo the oxidative changes and consequent
flavor deterioration (29, 37, 62, 63). Tappel__§_al. (56)
have reported that oxidation of non—ether—soluble conjugated
lipids could account for approximately 50% of the total
oxygen absorption, and that the most unusual and most de-
teriorative oxidation reaction in freeze-dried beef appears
to involve the oxidation of the protein fraction.

Another reaction limiting the storage life of freeze-
dried meat is nonenzymatic browning. Browning reaction of the
Maillard type between a carbonyl group and the free amino-
groups of protein or amino—acids have long been recognized
as potent causes of change in the color and flavor of foods.
The several types of browning, the many different reactants
and reactions involved in each type, and the large number of
controlling variables make the pathway of browning reactions
an intricate subject (27).

Lea (37) has suggested that several kinds of phospho-
1ipids containing free amino-groups, could undergo the same
kind of reactions with sugar and other aldehydes in freeze—
dried meat as do the amino—group of amino acids and proteins.

Henrickson gt a1. (25, 26) have shown that the
storage life of dehydrated pork can be extended by reducing
the glucose content. Browning has been reported by Tappel

gt 1. (57) to be responsible not only for deteriorative

11

changes but also for the desirable flavor change of pre-
cooked freeze—dried beef during storage.

Flosdorf (17) has reported that moisture content
should be reduced to below 2 percent, and preferably to 0.5
percent for best keeping quality while Goldblith gt _1. (22)

have pointed out that storage stability is greater with raw

materials of higher quality.

EXPERIMENTAL METHODS AND PROCEDURE
Preparation of Reagents

Carbonyl-Free Benzene. The following two methods were used
to purify benzene.

Method I. (For use in Experiments I and II)

Two liters of benzene were refluxed for 1 hour with
10 g. of DNP-hydrazine and 2 g. of trichloroacetic acid, and
then distilled through a Vigreux column.

Method II. (For use in Experiments III and IV)

This method is a modification of the method by
Schwartz §t_al. (52). One g. of DNP-hydrazine was dissolved

by grinding in a mortar with 12 ml. of 85% H3PO and 8 ml. of

4
H20 were added to the clear yellow solution. The precipitated
DNP-hydrazine was redissolved by further grinding. Twenty g.
of Celite (Johns-Manville, Hyflo Super-Cel) were then ground
with the solution until a homogeneous damp preparation was
obtained. The bright yellow impregnated Celite was then
transferred to a chromatographic tube of I. D. 28 mm. which
had a stopcock at the outlet and which contained about 50 m1.
of carbonyl-free n—hexane. A faster flow rate was obtained

by using Celite 545 instead of Hyflo Super-Cel in some of

the preparations. The height of the impregnated Celite

l2

13

column was 12.5 cm. The carbonyl—free hexane in the column
was drained from the column. The benzene to be purified was
stirred with excess DNP-hydrazine, and the benzene saturated
with DNP-hydrazine was added to the column. After discarding
the first 300 ml. of effluent, the remainder of the effluent
was collected at a flow rate of 4 liters per 24 hours. The
yellow effluent was stirred with adsorptive magnesia (Sea
Sorb 43, Fisher Scientific Co., Pittsburgh 19, Pa.) to remove
residual DNP-hydrazine, and the mixture was filtered. The
filtrate was distilled at atmospheric pressure to obtain

carbonyl—free benzene.

Carbonyl-Free Absolute Ethanol. 14 g. of aluminum granules

 

and 18 g. of KOH were added to 2 liters of ethanol. The
mixture was refluxed for 1 hour. 0n distillation, the first

100 m1. of distillate was discarded, and distillation was

stopped before the last 100 ml. had distilled.

DNP—Hydrazine Solution. DNP-hydrazine, twice recrystallized

 

from carbonyl—free methanol, was dissolved in amounts of 0.5
g. in 1 liter of carbonyl-free benzene to form a 0.05%

solution (W/V).

Trichloroacetic Acid Solution. Forty-three g. trichloracetic
acid were dissolved in carbonyl-free benzene and made to 1

liter to form a 4.3% solution (W/V).

Potassium Hydroxide Solution. A 4% solution (W/V) was made

by dissolving 4 g. of KOH with stirring in 100 ml. of

l4

absolute carbonyl-free ethanol. This solution was prepared
fresh and centrifuged for 10 min. immediately before use at

a speed of about 2,000 r.p.m.

Hydrated Alumina. Activated alumina, F-20 grade, (Aluminum

 

Co. of America, East St. Louis, Ill.) was modified by mixing
with 15% fully hydrated material prepared by exposing the
alumina in a thin layer to water vapor in a vacuum desiccator.
The mixed alumina was allowed to stand and equilibrate in a
closed container for several hours before use.
Extraction of Carbonyls from
Freeze-Dried Meats

A sample of freeze—dried meat weighing 5 - 7 g. was
weighed into a homogenizer flask (Virtis '45). Then 20 ml.
of carbonyl-free ethanol were added and the mixture was al-
lowed to stand for l min. while the ethanol penetrated the
meat. Forty ml. of carbonyl-free benzene was then added and
the mixture was blended by a Virtis '45 homogenizer at a
speed-setting of 50 for 2.5 min. while cooling the flask
with water at a temperature of approx. 50C. The resulting
slurry was filtered through a Buchner funnel and the residue
in the funnel was washed with approx. 40 ml. of a mixture of
benzene and ethanol (2 to 1; v/v). Depending on the concen-
tration of carbonyls, 3 — 5 m1. of the filtrate were pipetted

out for analysis of carbonyls.

15

Determination of Crude Fat Content

Method I. A known volume of a portion of the fil-

 

trate which had been used for carbonyl analysis, was trans-
ferred into a tared flask with a standard tapered joint. The
solvent was removed carefully under vacuum using a Rinco
flask evaporator. The flask containing the crude fat was
held in a vacuum desiccator overnight, then the gross weight
was determined on an analytical balance to calculate the
crude fat content.

Method II. The fats, soluble in hydrocarbon solvents,
were extracted from dry meat samples using petroleum ether

(Boiling range; 300C — 600C) in a Soxhlet apparatus.

Determination of Moisture

 

Approximately 5 g. of ground meat sample were
weighed into a tared weighing bottle. The weighing bottle
and its contents were kept in a drying oven at a temperature
of 1000C at atmospheric pressure for 24 hours. After being
cooled in a vacuum desiccator, the weighing bottle was re—
weighed and the loss in weight was assumed to be due to

moisture loss.

Procedure of Cooking and Freeze-Drying

 

Beef (Commercial grade, TOp round) was obtained from
the Michigan State University Food Store. It was closely

trimmed of depot fat and connecting tissue. The trimmed

l6

meat was sliced into cubes of approximately 2 cm on edge and
divided into two lots. One lot was freeze-dried without
cooking, and the other was freeze-dried after being cooked in
a Radarange Micro-Wave Oven, Model 1161 (Raytheon Manufactur—
ing Co., Newton, Mass.) in a single layer for 5 min.

Both cooked and raw meat samples were freeze—dried in
a Stokes Freeze—Dryer, Model 2003-F2 (F. J. Stokes Machine
Co., Philadelphia, Pa.). The frozen cubed meat samples were
laid on a tray in a single layer, and placed in the freeze-
dryer. During the last half period of freeze—drying, the
gauge controlling the heating plate was set to control the
temperature at approximately 420C. The vacuum in the drying
chamber was observed to fall to a minimum pressure of ap-

proximately 0.1 mm. of Hg.

Experiment I.

 

Determination of Carbonyl Compounds by a Modification of

Henick's Method (24)

Preliminary tests were conducted to determine the
applicability of the Henick method for carbonyls in freeze-
dried meats. The following two meat samples were used.
Sample 1: Cooked freeze-dried beef stored in an Open glass
jar at room temperature (23 i 20C.) for approximately 12
months. Sample 2: Cooked freeze-dried beef stored in an
Open glass jar at room temperature for approximately 5

months. (These were not from the same source.)

17

The procedure described below details the modified

Henick method used (23).
Determination of Two Classes of Carbonyls

Three m1. of trichloroacetic acid solution (4.3%)
and 5 ml. of DNP-hydrazine solution (0.05%) were placed in a
50 ml. volumetric flask and then 5 ml. of the benzene-ethanol
solution containing the carbonyls to be analyzed was added.
The flask was loosely stoppered and heated in a water bath
at 600C. for 30 min., and then cooled to room temperature.
To develop the color, 10 ml. of KOH solution was added,
carbonyl-free absolute ethanol was added to make a total
volume of 50 ml. and all components mixed thoroughly. After
exactly 10 min., the absorbancy values of the wine-red
colored solution were read at 430 and 460 mxi in a Beckman
DU spectrophotometer against a blank prepared in the same
. manner by substituting 5 m1. of the carbonyl—free benzene—
ethanol mixture (2 : 1, V/v).

The absorbancy values at 430 and 460 mft were substi—
tuted into the following equations to calculate the saturated
and the oCfl -unsaturated carbonyls.

Saturated

5.803 A - 4.412 A

430 460
Unsaturated = -3.366 A430 + 4.338 A460
where A430 indicates absorbancy at 430 In)»

A460 indicates absorbancy at 460 m}&.

Unit isji moles of carbonyl per 50 m1. soln.

18

Experiment II.

 

Changes of Carbonyl Content in Freeze—Dried Raw and Cooked

Beef during Storage at Room Temperature

 

Ten different samples of a single lot of freeze-
dried beef were prepared, and stored, wrapped with aluminum
foil, at room temperature (23 i 20C.) for 12 weeks. Controls
and antioxidant treated samples were prepared as noted below.
During the storage, saturated and CXUB-—unsaturated carbonyls
were determined by a modification of the method by Henick

t l. (24).

——

Sample Preparation for Experiment II.

 

Sample 1; The cooked meat was dipped into an aqueous
solution of 0.01% propyl gallate (P.G.) at approximately 170
C. for 15 min. After being drained on a screen, the P.G.-
treated meat was freeze-dried to obtain a final moisture of
less than 2% for about 24 hours.

Sampl§_2; Sample 2 was prepared in the same manner
as Sample 1 except that trihydroxybutyrophenone (T.H.B.P.)
was used instead of P.G.

Sample 3; The cooked meat was freeze-dried for 8
hours. The partially dehydrated meat was removed from the
freeze-dryer, and dipped into an aqueous solution of 0.01%
P.G. at approximately 170C. for 15 min. After being drained,
the meat was again placed in the freeze-dryer for an ad-
ditional 16 hours until the final moisture content was less

than 2%.

19

Sample 4; Procedure was the same as for Sample 3 ex-
cept that T.H.B.P. was used instead of P.G.

Sample 5; Fifty g. of the freeze-dried cooked beef
was placed in a 2.59 liter desiccator together with 7.5 mg.
of butylatedhydroxyanisole (B.H.A.). The desiccator was
evacuated by a water aspirator for 5 min., and sealed
hermetically at approximately 20 mm. Hg at room temperature.
In order to make the solid B.H.A. vaporize and penetrate the
freeze-dried meat sample, the desiccator was placed in an
oven at 600C. for 20 hours.

Sample 6; Procedure was the same as for Sample 5
except that butylatedhydroxytoluene (B.H.T.) was used in—
stead of B.H.A.

Sample 7; Control sample comparing to the cooked
antioxidant-treated samples (Samples 1 - 6).

Sample 8; Freeze-dried raw beef was treated with
B.H.A. exactly as the cooked freeze-dried beef in Sample 5.

Sample 9; Freeze—dried raw beef was treated with
B H.T. exactly as the freeze-dried cooked beef in Sample 6.

Sample 10; Control sample to compare with the raw

 

antioxidant-treated samples (Samples 8 — 9).

2O

Experiment III.

Determination of the Classes of Free Monocarbonyl Compounds

 

in Oxidizing Lipids of Freeze—Dried Beef using Benzene—

 

Ethanol Extracts

 

Different treatments of freeze-dried beef, a model
system with corn oil incorporated in freeze—dried egg white,
and a pure chemical two component system consisting of
hexanal and cinnamaldehyde were used. The preparation of
the freeze—dried beef samples was similar to that in Experi-
ment II except for the following two points: (a) In the
treatment of B.H.A. or B.H.T. vapor, a vacuum pump was used
to evacuate the desiccator for 2 min. The desiccator was
sealed hermetically at a pressure of less than 1 mm. Hg at
room temperature. (b) Aluminum foil was not used for
wrapping the sample. Six g. samples were weighed accurately
into 100 ml. beakers. All samples were weighed at the same
time, and allowed to stand in the beakers covered with watch
glasses at room temperature until analysis.

The following modified procedure of the method by
Keith and Day (33) was used in this portion of the study:

Determination of Three Classes of
Free Monocarbonyls

 

A chromatographic tube of 15 mm I.D. x 30 cm. was
plugged at the lower constricted end with sintered glass and/
or glass wool, and the outlet was clamped. Carbonyl-free

benzene was added to a level of 5 cm. and the 15% hydrated

21

alumina was added to a depth of 3 cm. The tube was tapped
lightly to remove air bubbles and benzene was allowed to
drain to the level of alumina. Ten g. of DNP-hydrazine
reagent were added and sufficient 15% hydrated alumina was
immediately added to increase the depth of the column by 1
cm. After the reagent had passed onto the column, an ad-
ditional 10 ml. of carbonyl—free benzene was added and the
total depth of the column was made to 8 cm. by further ad-
dition of 15% hydrated alumina. Finally, the column was
washed with 5 ml. of carbonyl—free benzene prior to being
used. Three to five ml. of the benzene-ethanol solution
containing less than l/umole of carbonyls was added to the
column. The solution was completely washed into the column
with small aliquots of carbonyl-free benzene. A total of
100 ml. of carbonyl-free benzene was then percolated through
the column at a flow rate of 5 ml. per min. The eluate was
collected in a 250 ml. standard taper round bottom flask.
The solvent was carefully removed from the flask under re-
duced pressure, and the following were added in the sequence
listed: 5 ml. of carbonyl-free benzene, 10 ml. of 4%
ethanolic KOH, 35 ml. of carbonyl-free absolute ethanol.

The flask was stoppered and the contents were thoroughly
mixed by inverting and shaking. The absorbancy values of the
above solution were read at 430, 460, and 480 mwt. com-
mencing exactly 10 min. after addition of the ethanolic KOH.
Blank determinations were run at the same time on 100 ml. of

the purified benzene and the color pigment in sample. The

22

blank of color pigment was determined on a portion of the
sample which had been passed over an alumina column free of
DNP-hydrazine. After accounting for the absorbancy of the
benzene blank and the pigment of the sample, the quantity of
each of three monocarbonyl classes was calculated by means

of the following simultaneous equations: (See Appendix V for
derivation of constants).

Alkanals = 7.158 A - 11.142 A + 6.496 A

430 460 480
Alk-2-enals :—5°477 A430 + 15.374 A460 - 11.090 A480
Alk—2,4-dienals = 1.514 A430 - 6.634 A460 + 6.423 A480

where A430 indicates absorbancy at 430 m}4., etc.
Unit is jlmoles of carbonyl per 50 ml. solution.
Preparation of a Model System with Corn Oil
incorporated in Freeze-Dried Egg White
Raw egg white was beaten, and corn oil was added to
the whipped egg white to a level of 5% of raw egg white by
weight. The mixture was whipped until firm, and then held
at —4OOC. for 1 hour. The frozen sample was transferred to
the freeze—dryer, where it was dried for 11 hours to obtain
a final moisture content of 2% with the heating plate temper—
ature of approximately 450C. under a vacuum of approximately
0.1 mm. Hg.
The finished sample was transferred into a brown
glass bottle with a large head space and stored at room

temperature (23 i 20C.) until analysis.

23

Experiment IV.

 

Determination of the Classes of Free Monocarbonyl Compounds
in Oxidizing Lipids of Freeze-Dried Beef using an Acid~

Benzene Mixture as Extracting Means

The freeze-dried beef samples were prepared in the
same way as in Experiment III. In extraction of carbonyls,
pure benzene and an acid—benzene mixture were used. The
method using the pure benzene was essentially that of Keith
and Day (33). The method using the acid—benzene mixture was
as follows:

Extraction of Carbonyls with an
Acid-Benzene Mixture

Seven 9. of sample were placed in a 100 ml.
homogenizer flask (Virtis '45) and then 30 ml. of 15% H2804
(5.64 N) and 60 ml. of carbonyl-free benzene were added. The
mixture was homogenized with a dial setting of 45 on the
Virtis "45 homogenizer for 2.5 min. The slurry was trans-
ferred to a centrifuge bottle with the aid of a small amount
of carbonyl-free benzene, and centrifuged at a speed of 2,000
r.p.m. for 10 min. The supernatant was transferred to a
250 ml. graduated cylinder. Forty ml. of carbonyl-free
benzene were added to the centrifuge bottle containing the
lower phase, this was thoroughly mixed with a spatula. The
mixture was again centrifuged in the same manner. The
supernatant was combined with the previously obtained

solution, and the total volume of the extract was measured.

24

A 3 — 5 ml. aliquot containing less than 1 /ymole of
carbonyls was used for analysis following the procedure
described in the previous section. For a control sample,
the 15% H2SO4 was replaced by distilled water, and the re—

maining procedure was exactly as previously described.

RESULTS AND DISCUSSION

The experiments in this study were conducted on the
premise that since carbonyl compounds have been correlated
with rancidity in some systems, the degree of rancidity of
freeze-dried beef could be evaluated by determining its
carbonyl content, eSpecially the monocarbonyls (5, 19, 33,
43). In assessing the results of this investigation, a
number of general observations have been made as well as
those pertinent to each experiment.

The change in moisture level in the freeze-dried
beef probably has some effect on carbonyl formation. When
the freeze—dried beef was exposed to the atmosphere during
storage, the level of moisture fluctuated and increased from
1.15% to as much as 10.55% according to the humidity of the
air. The procedure of weighing all samples at one time,
which was mentioned in Experiment III, eliminated the weight
errors due to moisture change. In freeze—drying the meat
samples, low moisture levels were obtained with less diffi-
culty in the precooked sample than in the uncooked.

In measuring the absorbancy values of quinoidal ion,
it was necessary, in order to obtain dependable and re-
producible results, to make all absorbancy readings at the

same time interval after adding ethanolic KOH. This was

25

26

required because the unstable colored quinoidal ion faded with
time at a rate of 0.5% to 0.7% per min. (31, 48). A 60

second period was maintained between the readings of the ab-
sorbancy values for each sample.

In Experiment III, it was observed that an appreciable
amount of carbonyl compounds was released from a rubber tube
attached to the lower end of the chromatographic column. A
rubber tube, 5 mm. I. D. x 5 cm., liberated an average of
0.145 fimoles of carbonyls by passing 100 ml. of benzene
through the tubing at a flow rate of 5 ml. per min. There—
fore, in Experiment IV a Teflon stopcock replaced the clamped
rubber tube thus eliminating the source of extraneous
carbonyl compounds.

In the simultaneous equations (Appendix V) used in
Experiments III & IV, the coefficients for the absorbancy
values were slightly smaller, by about 0.1%, than those re—
ported in the paper by Keith and Day (33). This small
difference did not appear to be grossly meaningful in evalu—
ating results but it seemed to show that the constants used
by Keith and Day were not precisely determined.

Initial studies were made using a modification of
the method of Henick _p.al. (24). These began with evalu—
ation of known substances. An adequate recovery of carbonyl
at a range of less than 250 x 10-6 molar concentration of
carbonyls in a one—component system using this method is
shown in Table 1. When applied to freeze-dried beef, as

shown by the data in Table 2, no definite correlation was

27

made between the period of storage and the amount of carbonyls
determined in two different samples of freeze—dried beef.

This could have been due to a number of factors which include
difference in composition, difference in storage conditions
over the period of storage, and difference in condition of

the meat prior to preparation and freeze-drying.

Table l. Standardization of Henick method with hexanal.

 

 

 

 

Molarity of Hexanal Percentage of
Added Found Recovery
41 x 10‘6 44 x 10'6 107.2
51 x 10’6 48 x 10'6 94.1
82 x 10'6 86 x 10'6 104.9
102 x 10‘6 110 x 10’6 107.8
204 x 10"6 228 x 10‘6 111.8

 

 

 

 

 

Table 2. Carbonyls of freeze-dried cooked beef by the
modified Henick method.

 

 

 

S l Stor e Saturated* Unsaturated* Total*
amp es ag Carbonyls Carbonyls Carbonyls
1 12 months 50.2 69.6 119.8
2 5 months 60.4 85.0 145.4

 

 

 

 

 

 

 

*Unit: [imoles of carbonyls in 1 g. of sample on dry
basis.

The data in Table 3 obtained by the modified Henick

method demonstrates the short term erratic changes of

28

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 3. Changes of carbonyl content in freeze—dried beef
during storage in aluminum foil at room temperature
(23 1 20 c.).
* Periods of Storage
** Carbonyl
Samples Compounds 0 1 2 4 12
week week weeks weeks weeks
Saturated 4.56 5.35 7.00 3.08 4.64
l Unsaturated 0.63 1.16 1.73 0.27 0.93
Total 5.19 6.51 8.73 3.35 5.57
Saturated 5.20 4.72 5.45 6.12 6.75
2 Unsaturated 0.17 0.67 0.99 —l.35 0.68
Total 5.37 5.39 6.44 4.77 7.43
Saturated 4.46 5.42 3.48 3.57 5.95
3 Unsaturated 0.73 0.95 1.28 0.00 1.96
Total 5.19 6.37 4.76 3.57 7.91
Saturated 5.08 5.34 6.05 4.58 7.80
4 Unsaturated 0.32 0.83 0.72 0.18 0.12
Total 5.40 6.17 6.77 4.76 7.92
Saturated 1.03 2.89 5.35 5.55 7.80
5 Unsaturated 2.42 0.44 0.53 0.27 0.67
Total 3.45 3.33 5.88 5.82 8.47
Saturated 3.44 4.44 5.18 5.64 7.84
6 Unsaturated -0.11 -0.04 0.37 0.16 0.36
Total 3.33 4.40 5.55 5.80 8.20
Saturated 6.79 6.32 5.47 3.11 5.96
7 Unsaturated 0.06 0.96 1.79 0.15 1.04
Total 6.85 7.28 7.26 3.26 7.00
Saturated 2.04 4.45 4.74 5.50 7.97
8 Unsaturated 1.50 —0.22 0.51 0.22 1.23
Total 3.54 4.23 5.25 5.72 9.20
Saturated 3.90 5.34 4.62 5.49 5.71
9 Unsaturated -0.01 -0.18 0.26 0.19 0.32
Total 3.89 5.16 4.88 5.68 6.03
Saturated 4.05 5.27 5.13 5.01 7.22
10 Unsaturated -0.23 0.02 0.21 0.10 0.31
Total 3.82 5.29 5.34 5.11 7.53
*Unit isflmoles of carbonyls in l g. of meat sample

on dry basis.

 

**Sample numbers refer to the sample preparations
described in the section on Experiment II.

 

29

carbonyl content in freeze-dried beef both raw and cooked.
Antioxidant treated and untreated samples appeared to behave
similarly. The effect of the antioxidants in suppressing
oxidation in freeze-dried beef could not be detected by this
analytical method. When analysis of variance (11) was
applied to the experimental data for the total carbonyls in
Table 3, there was no significant difference between any two
of the different treatments even at the 5% level as judged
by the total carbonyl content. Values of the carbonyl
content in all samples went up by the end of 12 week storage.
These are denoted by a larger calculated F—value 8.15 than
the tabular value 3.89 at 1% level (Degrees of freedom: 4
for numerator: 36 for denominator).

In Table 2, the crude fat content by Method I in
Samples 1 and 2, were 30.8% and 20.3%, on a dry basis, re-
spectively. The crude fat content of the fresh beef sample
used in Experiment II was 17.5% on a dry basis. The lower
fat content of the beef used in Experiment II appeared to be
manifested by the lower levels of carbonyl content in the
data for freeze-dried product as shown in Table 3 than for
those shown in Table 2. The average value of the crude fat
content in the raw beef was 15.35% for that used in Experi—
ment III, and 18.8% for that used in Experiment IV. It ap—
peared that the level of carbonyl content in a given sample
was influenced to some degree by the crude fat content of

the sample.

30

The formation of DNP—hydrazone derivatives in the
benzene—ethanol solution containing trichloroacetic acid and
the DNP-hydrazine reagent was carried out by refluxing at a
temperature of 600C. During the reaction, decomposition of
carbonyl precursors probably occured, and the carbonyl values
obtained by this method were much higher than those determined
by using the alumina reaction column. This decomposition of
precursors was probably partially responsible for the higher
levels of the total carbonyls found in Experiment II than
those in Experiment III.

While the procedure of Henick_gp._l. (24) possessed
the desirable features of simplicity and convenience, this
procedure was found not to give an adequate picture of
carbonyl distribution in the oxidizing freeze-dried beef.
Since it was reported that alkanals, alk-2-enals, and a1k—
2,4-dienals were the major classes of monocarbonyls in oxidiz-
ing lipids by several researchers (15, 20, 33), a modifi-
cation of the method of Keith and Day, which afforded the
separation of these classes of carbonyls, was employed in
Experiments III and IV.

After investigation with several kinds of samples in
the portion of the study devoted to Experiment III, it was
found that the solvent system used to liberate the protein
bound lipids had an unfortunate eluting effect on the alumina
reaction column impregnated with the DNP-hydrazine reagent.
The ethanol washed down undesirable substances such as the

unreacted DNP—hydrazine reagent and the DNP-hydrazone

31

derivatives of polycarbonyls from the column. The elution
of the unreacted DNP—hydrazine reagent was recognized by the
fast migration of a yellow band in the column, and by the
higher absorbancy values compared to those of a control in
which the pure benzene used. The elution of the DNP-
hydrazone derivatives of polycarbonyls from the column was
also demonstrated by the higher absorbancy values.

The model system in Experiment III was set up to
examine whether it might be possible to eliminate some of
the unknown factors which influence the determination of
carbonyls. Due to the unfortunate eluting effect of the
solvent system used, the results obtained by the model system
with corn oil incorporated in freeze-dried egg white were
difficult to correlate with other observations.

The simultaneous equations used in this study
(Appendix V) were valid only for the three component system
consisting of alkanals, alk-2-ena1s, and alk-2,4-dienals
formed under the same conditions as those for which the
equations had been developed. The eluate obtained by the
use of the benzene-ethanol solvent system was found to con-
tain more than three components. It was also found that the
simultaneous equations did not distinguish between the three
major classes of monocarbonyls with significant accuracy.
However, the total monocarbonyl content still showed a steady
increasing tendency.

Typical results obtained in Experiment III are shown

by the data for freeze-dried raw beef in Table 4. These

32

showed a maximum in carbonyl content after a 9 day storage
period. Although many carbonyl compounds were formed and
measured, the monocarbonyl content still appeared to be low
level, and most of the monocarbonyls measured were the alka-
nals or saturated aldehydes. The non—detectable values in
Table 4 indicated that the difference in absorbancy values
of the blank and the sample was not significant. In order
for a maximum value to be develOped and followed by lower
values, the carbonyl formed must have decreased either

through volatilization, which is improbable for the higher
have been reacted

molecular weight entities, or they must

with other components of the system, or they were possibly

oxidized further to products which were non-carbonyl in

 

 

 

 

 

 

 

 

 

 

nature.
Table 4. Changes of free monocarbonyls in oxidizing freeze~
dried raw beef stored at room temperature
(23 4 200.)
* * . * Total
Storage Alkanals Alk-Z—enals Alk—2.4_d1enals Carbonyls*
0 day ___*t -__*k --1** ___**
1 day ___+;+; ___tlrir _____*iz‘ ___$:‘k
2 days 0 012 «0.067 —0.021 —0.076
7 days 0.174 0.310 —0.114 0.370
9 days 1.570 -0.320 0.160 1.410
13 days 0.700 -0.044 0.205 0.861
*Unit is lkmoles in 1 g. of sample meat on dry basis.

if
Dash marks refer to non-detectable values.

33

After discovering the eluting effect of ethanol on
the products developed on the alumina reaction column, the
original Keith and Day method (33) using only benzene as
extracting solvent, was again applied to a series of freeze—
dried beef samples, but the free monocarbonyl contents re-
mained practically nil for a 20 day storage period at room
temperature (23 i 20C.) in beakers. This negative result
did not indicate that the free monocarbonyl compounds were
not formed in the oxidizing samples, but probably meant that
the free monocarbonyls extractable by neutral solvent did
not accumulate as stable compounds. It could indicate also
that many of the monocarbonyl compounds measured in other
experiments were not present as such in the samples but were
liberated by the conditions of extraction and derivative
formation from oxidized precursors.

The lipid fractions primarily involved in the oxi-
dative reaction are phospholipids and protein bound lipids,
rather than triglycerides which are generally termed neutral
lipids (63). It is often found that the phospholipids from
a particular tissue are appreciably more unsaturated than
triglycerides from the same source (29, 37, 62). Protein
bound lipids which are referred to as proteolipids cannot be
extracted readily with the usual fat solvents. Phospholipids
which have higher polarity than neutral lipids, and which
are more likely to be bound to protein, are practically in-
soluble in a nonpolar solvent such as benzene. Protein

bound lipids appear to share a role with proteins in

34

maintaining the integrity of the cell. Bonding by water
molecules seems to play an important part in the union be-
tween the lipid and the protein. Therefore, polar solvents
such as methanol or ethanol, which are dehydrating agents,
are required for the liberation and extraction of tissue
lipids, and even with the use of a polar solvent some of the
tenaciously bound lipids appear not to be completely
liberated (37, 46, 63).

An acid—benzene mixture was employed as liberating
and extracting solvent system in Experiment IV. As shown in
Table 5, the carbonyls from the acid-benzene extraction were
higher than those from the water benzene extraction, in which
the freeze-dried, uncooked beef stored at a temperature of
ca. -l2OC. for a month was used as sample. The extraction
by the acid—benzene mixture yielded an extra 21% crude fat
compared to that by the water-benzene mixture. The color of
the extract obtained by the acid—benzene mixture was more in-
tense pinkish-yellow than that by the benzene from the water
slurry. The acid treatment thus appeared to liberate the
protein bound lipids. However, 15% acid probably caused
enough deterioration of oxidized precursors in the meat
samples to form additional carbonyls. In order to obtain a
more precise evaluation of this acid-benzene extraction

method, further study is needed.

35

Table 5. Monocarbonyls in crude fats extracted from freeze-
dried raw beef by an acid-benzene mixture and by
benzene from a water slurry.

 

 

 

 

* * . * Total
Solvents Alkanals Alk—Z-enals Alk-2,4-dienals *
Carbonyls
Acid-
Benzene 0.434 —0.003 0.018 0.449
Water_
Benzene 0.083 -0.025 0.007 0.065

 

 

 

 

 

 

*
Unit is /{moles in l g. of meat sample on dry basis.

The accumulation of carbonyl compounds in the stored
freeze—dried beef has been observed to a certain extent as
shown in Tables 3 and 4. The increase of the carbonyl com—
pounds with storage time, as yet, can not be correlated to
other factors such as the effectiveness of the antioxidants
and their methods of application. Due to the complexity of
the autoxidation process, it seems that no simple analytical
method can be directly applied to the evaluation of change in
the freeze-dried meat with complete satisfaction.

As an example of what could be considered in a
further study, the procedure used in Experiment III could be
modified as follows (14): Extract the freeze-dried meat
sample with a benzene—ethanol mixture. Let the extract con—
taining essentially total lipids and all kinds of carbonyls
react in the alumina column impregnated with DNP—hydrazine
reagent by percolating the benzene—ethanol extract through

the column. Remove the solvent from the eluate containing

36

all kinds of DNP-hydrazone derivatives and the unreacted
DNP—hydrazine reagent. Dissolve the dry material in carbonyl
free benzene and percolate the benzene solution over an
alumina column without any DNP-hydrazine reagent in order to
separate the DNP—hydrazone derivatives of monocarbonyls from
those of polycarbonyls and the unreacted DNP-hydrazine
reagent. The eluate containing the DNP—hydrazone deriva-
tives of monocarbonyls would be used to measure absorbancy
at 430, 460 and 480 m/(. The three absorbancy values would
be substituted into the simultaneous equations with appropri-
ate constants to calculate the three major classes of
monocarbonyls.

This modified procedure still includes the treatment
with a dehydrating agent such as ethanol which ruptures the
lipid-protein linkage. It suggests a means whereby the ab-
sorbancy due to polycarbonyls and the unreacted DNP—hydrazine
reagent could be eliminated and permit measuring the three

classes of monocarbonyls.

CONCLUSIONS

The carbonyl values of stored freeze—dried beef as
determined by the method of Henick_§pl_1. (24) were much
higher than those obtained by that of Keith and Day (33).
While the former method determined the monocarbonyls as
well as the polycarbonyls, the latter measured only the
monocarbonyls. However, the different values obtained
by use of the two methods indicated that decomposition
of carbonyl precursors probably occured with the Henick
method giving a higher yield of measurable carbonyl.
Although many carbonyls were found and measured in the
oxidizing lipids in freeze-dried beef, only a few per-
cent of the total carbonyls appeared to be the free
monocarbonyls. Most of the free monocarbonyl compounds
seemed to be the alkanals or the saturated aldehydes.
Carbonyl compounds were observed to accumulate in the
stored freeze-dried beef. This increase with time could
not be correlated with other factors such as the effective-
ness of the antioxidants used in this study and their
methods of application.

The effectiveness of the antioxidants applied to the
freeze-dried beef could not be detected from the analyses

of carbonyl content by the methods used in this study.

37

38

An analysis of variance denoted no significant difference
even at the 5% level between any two of the ten samples
treated in the various ways.

The two methods employed in this study, namely the modi—
fied Henick method and the modified method of Keith and
Day, possessed certain desirable features. The main
feature were the simplicity and convenience by which
major classes of carbonyls were distinguished through
calculation by simultaneous equations from data for a
single sample without actual fractionation into the
separate components.

The presence of ethanol cannot be tolerated in the
solvent used for eluting DNP-hydrazones from a column
since it has the undesirable feature of eluting more than
the monocarbonyl derivatives from the alumina reaction
column impregnated with the DNP-hydrazine reagent. The
ethanol eluted the DNP—hydrazone derivatives of poly-
carbonyls and the unreacted DNP-hydrazine reagent in
addition to the DNP-hydrazone derivatives of monocarbonyls.
The absorption data obtained from the solution containing
more than the three major classes of monocarbonyls, that
is, alkanals, alkn2-enals, and alk-2,4-dienals, could

not be applied to the simultaneous equations developed
under the three component system in order to calculate
prOperly the three major classes with significant

accuracy.

39

Further research was indicated for studies of the acid-
benzene extraction method on the long storage series,
and of the further modified Keith and Day method which

included extraction by a mixture of ethanol—benzene.

BIBLIOGRAPHY

American Oil Chemists' Society. Official and Tentative
Methods of Analysis. 1946 ed., 1958 rev.
Tentative method Cd 8-53, Chicago 1, I11.

Badings, H. T. 1959. Isolation and identification of
carbonyl compounds formed by autoxidation of
ammonium linoleate. J. Am. Oil Chemists' Soc.
pp, 648.

Badings, H. T. 1960. Principles of autoxidation
processes in lipids with special regard to the
development of autoxidation off-flavors. The
Netherlands Milk and Dairy J. 14, 215.

Batzer, O. F., M. Sribney, D. M. Doty, and B. S.
Schweigert. 1957. Production of carbonyl com-
pounds during irradiation of meat and meat fats.
J. Agr. Food Chem. p, 700.

Berry, N. W., and A. A. McKerrigan. 1958. Carbonyl
compounds as a criterion of flavor deterioration
in edible fats. J. Sci. Food Agr. 9, 693.

Connell, J. J. 1962. The effects of freeze—drying on
the proteins of flesh foods. Freeze-Drying of
Foods. pp. 50. National Academy of Sci.,
Washington, D. C.

 

Corliss, G. A. 1963. The influence of cooking and
storage on oxidative changes in freeze—dried
beef as evaluated by the 2—thiobarbituric acid
test. M. S. thesis, Michigan State Univ.

Cowan, J. C., P. M. Cooney, and C. D. Evans. 1962.
Citric acid: Inactivating agent for metals or
acidic synergist in edible fats? J. Am. Oil
Chemists' Soc. 32, 6.

Dahle, L. K., E. G. Hill, and R. T. Holman. 1962. The
thiobarbituric acid reaction and the autoxidations
of polyunsaturated fatty acid methyl esters.

Arch. Biochem. BiOphys. 28, 253.

40

10.

11.

12.

13.

14.

15.

l6.

17.

18.

19.

20.

41

deKoning, A. J., and M. H. Silk. 1963. The 2-
thiobarbituric acid reagent for determination of
oxidative rancidity in fish oils. J. Am. Oil
Chemists' Soc. 49, 165.

Dixon, W. J., and F. J. Massey, Jr. 1957. Introduction
39 Statistical Analysis. pp. 139. McGraw—Hill
Book Co., New York, N. Y.

 

Dugan, L. R., Jr. 1955. Stability and rancidity. J.
Am. Oil Chemists' Soc. 32y 605.

Dugan, L. R., Jr. 1957. High energy irradiation and
fats. Am. Meat Institute Foundation Bull. No.
35. Chicago 37, Ill.

Dugan, L. R., Jr. 1964. Personal communication. Dept.
of Food Sci., Michigan State Univ.

Ellis, R., A. M. Gaddis, and G. T. Currie. 1961.
Carbonyls in oxidizing fat. IV. The role of
various fatty acid components in carbonyl gener—
ation. J. Food Sci. gp, 131.

Farmer, E. H., and D: A. Sutton. 1943. The course of
autoxidation reactions in polyisoprene and allied
compounds. IV. The isolation and constitution
of photochemically-formed methyl oleate peroxide.
J. Chem. Soc. 119.

Flosdorf, E. W. 1945. Drying by sublimation. Food
Ind. 11, 22.

Frankel, E. N., P. M. Cooney, H. A. Moser, J. C. Cowan,
and C. D. Evans. 1959. Effects of antioxidants
and metal inactivators in tocopherol-free soy-
bean oil. Fette, Seifen, Anstrichmittel p1, 1036.

Gaddis, A. M., R. Ellis, and G. T. Currie. 1960.
Carbonyls in oxidizing fat. III. The distribu—
tion of volatile and non-volatile carbonyls.
Food Res._g§, 495.

Gaddis, A. M., R. Ellis, and G. T. Currie. 1961.
Carbonyls in oxidizing fat. V. The composition
of neutral volatile monocarbonyl compounds from
oxidized oleate, linoleate, linolenate esters
and fats. J. Am. Oil Chemists' Soc. 38, 371.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

42

Gaddis, A. M., R. Ellis, and G. T. Currie. 1964.
Carbonyls in oxidizing fat. VI. The Girard T
reagent in the isolation and determination of
microamounts of n—aliphatic aldehydes and 2—
alkanones. J. Food Sci. 29, 6.

Goldblith, S. A., M. Karel, and G. Lusk. 1963. The
role of food science and technology in the
freeze dehydration of foods. Food Technol. 11,
139.

Gunstone, F. D. 1958. .Ap Introduction £p_the Chemistpy
pf Fats and Fatty Acids. John Wiley & Sons,
Inc., New York, N. Y.

 

Henick, A. S., M. F. Benca, and J. H. Mitchell, Jr.
1954. Estimating carbonyl compounds in rancid
fats and foods. J. Am. Oil Chemists' Soc. 31,
88.

Henrickson, R. L., D. E. Brady, C. W. Gehrke, and R. F.
Brooks. 1955. Dehydrated pork studies. Re-
moval of glucose by yeast fermentation. Food
Technol. 2, 290.

Henrickson, R. L., D. E. Brady, C. W. Gehrke, and R. F.
Brooks. 1956. Dehydrated pork studies: Re-
moval of glucose by glucose—oxidase enzyme.

Ibid. 19, l.

Hodge, J. D. 1953. Dehydrated foods. Chemistry of
browning reaction in model systems. J. Agr.
Food Chem. 1, 928.

Holman, R. T. 1954. Progress Lg the Chemistry 9f Fats
and other Lipids. vol. 2. PP. 51. R. T. Holman,
W. O. Lundberg, and T. Malkin, Eds. Academic
Press Inc., Publishers, New York, N. Y.

 

 

Hornstein, I., P. E. Crowe, and M. J. Heimberg. 1961.
Fatty acid composition of meat tissue lipids.
J. Food Sci. gp, 581.

Huang, I. Y., and H. N. Draudt. 1964. Effect of
moisture on the accumulation of carbonyl-amine
browning intermediates in freeze—dried peaches
during storage. J. Food. Technol. 18, 124.

Jones, L. A., J. C. Holmes, and R. B. Seligman. 1956.
Spectrophotometric studies of some 2,4—dinitro-
phenylhydrazones. Anal. Chem. 28, 191.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

43

Keeney, M. 1962. Symposium pp Foods: Lipids and
their Oxidation. pp. 83. H. W. Schultz, E. A.
Day, and R. O. Sinnhuber, Eds. The Avi Publish-
ing Co., Inc., Westport, Conn.

 

 

Keith, R. W., and E. A. Day. 1963. Determination of
the classes of free monocarbonyl compounds in
oxidizing fats and oils. J. Am. Oil Chemists'
Soc. 49, 121.

Kraybill, H. R., L. R. Dugan, Jr., B. W. Beadle, F. C.
Vibrans, V. N. Swartz, and H. Rezabek. 1949.
Butylated hydroxyanisole as an antioxidant for
animal fats. Ibid. gp. 449.

Kreis, H. 1902. Mittheilungen aus der analytischen
Praxis. Ueber neue Farbenreactionen fetter Oele°
Chemiker—Zeitung gp, 897.

Lappin, G. R., and L. C. Clark. 1951. Colorimetric
method for determination of traces of carbonyl
compounds. Anal. Chem. 23, 541.

Lea, C. H. 1957. Deteriorative reactions involving
phospholipids and lipoproteins. J. Sci. Food
Agr. 8, l.

Lea, C. H., and P. A. T. Swoboda. 1958. A new procedure
for the determination of volatile carbonylic
substances in autoxidizing edible fats. Chem.

& Ind. (London) 974.

Lea, C. H., and P. A. T. Swoboda. 1958. The flavor of
aliphatic aldehydes. Ibid. 1289.

Lillard, D. A., and E. A. Day. 1964. Degradation of
monocarbonyls from autoxidizing lipids. J. Am.
Oil Chemists' Soc. 41, 549.

Marcuse, R. 1956. On the function of ascorbic acid in
fat autoxidation. Experiments with salted
herring. Fette, Seifen, Anstrichmittel 58, 1063.

Mashio, F., and Y. Kimura. 1960. Spectrophotometric
determination of malonaldehydes in aqueous solu—
tion. Nippon Kagaku Zasshi (Tokyo) 18, 434.

McKerrigan, A. A. 1957. Flavor deterioration in edible
fats containing antioxidants. Chem. & Ind.
(London) 724.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

44

Mookherjee, B. D., and S. 8. Chang. 1963. Characteri-
zation of the carbonyl compounds in reverted soy-
bean oil. J. Am. Oil Chemists' Soc. 49, 232.

Olcott, H. S., and J. Van der Veen. 1963. Role of
individual phospholipids as antioxidants. J.
Food Sci. 2§J 313.

Ostrander, J., and L. R. Dugan, Jr. 1962. Some differ-
ences in composition of covering fat, inter-
muscular fat, and intramuscular fat of meat
animals. J. Am. Oil Chemists' Soc. 39, 178.

Patton, S., I. J. Barnes, and L. E. Evans. 1959.
n-Deca—2,4-dienal. Its origins from linoleate
and flavor significance in fats. Ibid. 39, 280.

Pool, M. F., and A. A. Klose. 1951. Estimation of
monocarbonyl compounds in rancid foods. Ibid.
28, 215.

Privett, O. S., C. Nickell, W. O. Lundberg, and D. D.
Boyer. 1955. Products of lipoxidase catalyzed
oxidation of sodium linoleate. Ibid. 32, 505.

Ramsey, M. B., and B. M. Watts. 1963. The antioxidant
effects of sodium tripolyphosphate and vegetable
extracts on cooked meat and fish. Food Technol.
11, 102.

Schogt, J. C. M., P. H. Begemann, and J. Koster. 1960.
Nonphosphatide aldehydogenic lipids in milk fat,
beef tallow, and ox heart. J. Lipid Res._l, 446.

Schwartz, D. P., and O. W. Parks. 1961. Preparation
of carbonyl-free solvents. Anal. Chem. 33, 1396.

Schwartz, D. P., O. W. Parks and M. Keeney. 1962.
Separation of 2,4-dinitrophenylhydrazone deriva-
tives of aliphatic monocarbonyls into classes on
magnesia. Ibid. 34, 669.

Sinnhuber, R. 0., and T. C. Yu. 1958. 2—Thiobarbituric
acid method for the measurement of rancidity in
fishery products. II. The quantitative determin-
ation of malonaldehyde. Food Technol. ng 9.

Soukup, R. J., R. J. Scarpellino, and E. Danielczik.
1964. Gas chromatographic separation of 2,4—
dinitrophenylhydrazone derivatives of carbonyl
compounds. Anal. Chem. 3Q, 2255.

56.

57.

58.

59.

60.

61.

62.

63.

64.

45

Tappel, A. L. 1956. Freeze-dried meat. II. The
mechanism of oxidative deterioration of freeze-
dried beef. Food Res. 21, 195.

Tappel, A. L., R. Martin, and E. Plocher. 1957. Freeze-
dried meat. V. Preparation, properties, and
storage stability of precooked freeze-dried
meats, poultry, and seafoods. Food Technol. 11!
599.

Tarladgis, B. G. 1961.. An hypothesis for the mechanism
of the heme catalyzed lipid oxidation in animal
tissues. J. Am. Oil Chemists' Soc._3§, 479.

Thomson, J. S., J. B. Fox, Jr., and W. A. Landmann.
1962. The effect of water and temperature on
the deterioration of freeze-dried beef during
storage. Food Technol. 1p, 131.

Witting, L. A., and B. S. Schweigert. 1958. A
Characterization of volatile carbonyl compounds
isolated from meat fat subjected to gamma
radiation. J. Am. Oil Chemists' Soc. 35, 413.

Wyatt, C. J., and E. A. Day. 1963. Autoxidation of
fish oils. II. Changes in the carbonyl distri—
bution of autoxidizing salmon oil. J. Food Sci.
28, 305.

Younathan, M. T., and B. M. Watts. 1959. Relationship
of meat pigments to lipid oxidation. Food Res.
24, 728.

Younathan, M. T., and B. M. Watts. 1960. Oxidation of
tissue lipids in cooked pork. Ibid. 25, 538.

Zipser, M. W., and B. M. Watts. 1962. A modified 2-
thiobarbituric acid (TBA) method for the determin—
ation of malonaldehyde in cured meats. Food
Technol. 16, 102.

APPENDICES

APPENDIX I

Scheme for Formation of HydrOperoxides and
Dismutation to Monocarbonyl Compounds

CH3- (CH2) 4- CH=CH- CH
Rl- CH=CH- CH— CH: CH- R2

R1- CH= CH- CH= CH- CH- R2 R1

#

2-CH=CH- (CH2) 7-COOH

- CH- CH: CH- CH: CH- R2

 

 

 

 

Rl- CH=CH- CH ( OOH) -CH=CH- R2
‘f ‘V
Rl- CH=CH- CH=CH- CH ( OOH) -R2 R1- CH ( OOH) -CH=CH- CH=CH- R2
V 1
CH3- (CH2) 4- CH=CH- CH0 CH3- (CH2) 4-CHO

CH -(CH.

3 2) 4- (CH=CH) 2-CHO

47

APPENDIX II

Reaction of DNP-Hydrazine with a Carbonyl
Compound and the Chromophoric
Quinoidal Ion Developed in
the Presence of Base

 

 

 

 

 

N0 N0
R. H l 2 R1 ’ 2
‘\\ \\. ——‘ ‘HZO \\ "“ 0
C=O + , N-N4 -NO C=N—N -N
R H H R H ‘
2 2
carbonyl DNP-hydrazine DNP-hydrazone derivative
R1 0"
base \\\ +///
7* C=N—-N= =N
/// \\\
R2 0H

Chromophoric quinoidal ion of
DNP—hydrazone derivative

48

 

H L I Il/III,

m
125 N Q U ImH . NH . m G thHwQOHUALIwa
m Uflom Oflcwaocflq

HMI U I hxouwmon a I
¢H.ma.mq U S ma

 

I I04 I Ima
HMCGHU .VaN Q U NH.®< U

m INH.OH.WN>MOHOQOHU%£IG

Hmcml Q I U .<.
N |

6 NH.@

AusmxonwmoupwnIHH
Hal 0 ( ..
I . hxoummou w I
Ha.m.flw p n ma

6.0a oawaocflq

 

Hmcwl AN IHHU
m .. 4u>xoumm0HU>£Im w
Hmchm Q IOHO m < ma
I hxoummou m I I I
Hmumo OH Am 6 n m m.mm o
m Im.flN>XOHmQOHU%£IOH

Hml O Uflom UHOHO
Imnflwmx0Hmmoupmnlaa

 

 

 

mmpxnwpafl mmpfixoummoupmm mpflom mpumm

 

 

 

H xflpcwmm< ca menom map >3
mpfiofi muummnsfiflm.pw>flpmn mwphnwpafl pcm mwpflxouwmoup>m

HHH XHQmem<

 

6H

.m m I I
H cm 0000 oa.s.6.mnw “0
«a

macaw I . I
a . 0 m.m.m.uw 6
ma
MCWHIH I . I
H . u s.4.mxu. o
I , INA
H6636 6 .m AM o

IHH
Hmcmflfl m.Nnm U
. OH
MCMH I ,VI
H .0 fi‘N , U

0
Im
HMCOIN Q U

HmIoo

I¢H.HH.m.oAflwmxoummouphnlm

I¢H.HH.m.mAcwhxonwmoup%£Ih
Iva.aa.m.mAflwmxoummoupmnIm
I¢H.HH.h.m.uwwxonomonp%nlm
I6H.HH.m.m”HumxoummonpmnIoa
I¢H.ma.m.m mexoummoupthHH
I6H.oa.w.m.dwmxonmmoupmnImH
Iva.aa.m.m.flwmxoummoupmzlma

I . wxouwmou m I
mH.HH.m.m.m. 6 n ma

 

I Iom
4H.Ha.m.m.flw o

pflo< 0flcopflzomn4

 

OH
MCGHH I 4% I
H - g N- § V \ N 1.1. U

m
MC m H I 1IN I
H a U m ‘ N .(\ U

h
META. I IN I

o
HMCOIM 1% I U

 

 

I hxouwmou % I
ma.ma.oa Am 6 n m

I . hxoummou a I .
ma.ma.m,m 6 a As
I . xx Homou % I

mH.mH.m,w o 6 6 NH

I . . , hxoummou w I
ma.aa.m.flm 6 6 ma

 

I ImH
ma.ma.mnm. o

 

50

APPENDIX IV

Common and Chemical Designations of

 

 

Antioxidants
Name and Abbreviation Chemical Names
PrOpyl gallate
P.G. 3,4,5—trihydroxypropylbenzoate

 

Trihydroxybutyrophenone

 

T. H. B. H. 2,4,5—trihydroxybuterphenone
Butylatedhydroxyanisole a mixture of the isomers
B. H. A. 2-tert—butyl-4-methoxyphenol

3-tert-butyl-4—methoxyphenol

 

 

Butylatedhydroxytoluene J
B. H. T. 2,6-di-tert-butyl-p—cresol

 

 

51

 

APPENDIX V

Derivation of Simultaneous Equations by Matrix
Inversion from Average Molar Absorptivities
of DNP-Hydrazone Derivatives of
Monocarbonyl Compounds
The average molar absorptivities of alkanals, alk-2—
enals, and alk-2,4-dienals for the three wave lengths corres-

ponding closely to the absorption maxima of the DNP-hydrazone

derivatives, are given as follows:

430 m/‘° 460 m/I. 480 m};.
Alkanals = 20,930 15,290 10,860
Alk-2-enals = 23,670 30,050 25,460
Alk—2,4-dienals = 19,700 36,420 40,760

The Lambert-Beer law is mathematically expressed in
the form:
A = aCl + bC2 + ...
_ 100

where A is the absorbancy (log——;E) of the sample at
the given wave length, a, b, etc. are the absorbancys
for each of the components in unit concentration, and
C1’ C2, etc. are the concentration for each component.

Hence, in a three component system with the three

wave lengths, that is, 430, 460 and 480 mlk;

52

53

A430 = a430C1 + b430C2 + d430C3
A460 = a460C1 + b460C2 + d460C3 ..... .......... (l)
A480 = a480C1 + b480C2 + d480C3/

where Ai's are the absorbancys of the sample con—
sisting of three components at the wave length of

i mfx, ai's are the average molar absorptivities of
alkanals at the wave length of i mxfi, and similarly
bi's and di's are the average molar absorptivities

of alk—2-enals and alk—2,4-dienals, respectively.

The given average molar adsorptivities are substituted

(1).

into equations

A430 = 20,930 C1 + 23,670 C2 + 19,700 C3
A460 = 15,290 C1 + 30,050 C2 + 36,420 C3 . ........ (1')
A480 = 10,860 C1 + 25,460 C2 + 40,760 C3

/

The above simultaneous equations are inverted by

Cramer's rule.

 

 

 

 

 

 

A430 23,670 19,700 A430 23,670 19,700
A460 30,050 36,420 A460 30,050 36,420
A480 25,460 40,760 A480 25,460 40,760
20,930 23,670 19,700 2,078,735,680,000
15,290 30,050 36,420

10,860 25,460 40,760

2,975I848 A430

 

+ 2,700,764 A

60 480

 

20,787,356,800

54

Unit is converted from molar concentration to fimoles

per 50 ml. by multiplying by 5 x 104.

 

 

 

50,000Cl = 7.158 A430 — 11.142 A460 + 6.496 A480 ....... (2)
Similarly, for C2 and C3,

20,930 A430. 19,700

15,290 A460 36,420

10,860 A480 40,760
50,000C2 = x 50,000

2,078,735,680,000
= - 5.477 A430 + 15.374 A460 — 11.090 A480 ...... (3)

20,930 23,670 A430

15,290 30,050 A460

 

 

10,860 25,460 A480

50,000C3 = x 50,000
2,078,735,680,000

 

= 1.514 A - 6.634 A + 6.423 A

430 460 480

Therefore, from (2), (3) and (4) the simultaneous

.... ..... (4)

equations, which give the three major classes of monocarbonyls

in terms of jumoles per 50 m1., are as follows:
J/ Alkanals = 7.158 A430 — 11.142 A460 + 6.496 A480
L Alk-2—enals = —5.477 A430 + 15.374 A460 - 11-090 A480
Alk-2 4-dienals = 1.514 A430 — 6.634 A460 + 6.423 A480

”1.... «I
.| . 4..
”In \"

 

 

 

 

. w.._.. w

 

MICHIGAN STA‘IE UNIVERSITY LIBRARIES

III IIIIIIIIII'III

3174 6864

3 1293