Method for the Routine Quantitative Gas Chromatographic Analysis of Major Free Fatty Acids in Butter and Cream1

Method for the Routine Quantitative Gas Chromatographic Analysis of Major Free Fatty Acids in Butter and Cream 1 A. H. WOO and R. C. L I N D S A Y Department of Food Science University of Wisconsin Madison 53706 ABSTRACT

A rapid quantitative method was developed for routine analysis of the major, even carbon-numbered free fatty acids in butter and cream. Free fatty acids were isolated directly from intact samples by a modified silicic acidpotassium hydroxide arrestant column and were separated by gas chromatography with a 1.8 m × 2 mm inner diameter glass column packed with 10% neopentyl glycol adipate on 80/100 Chromosorb W. Purified, formic acid-saturated carrier gas was required for minimal peak tailing and extended column life. The accuracy and reproducibility of the method was established through quantitative recovery studies of free fatty acid mixtures, free fatty acids added to butter, and replicate analysis of butter and cream samples. INTRODUCTION

Sporadic occurrences of lipolytic rancidity or lipase flavors in both fresh and frozen butter continue to create quality assurance problems and attendant economic losses. While distinctly rancid cream and butter are identified easily by the usual routine indexing methods o f flavor scoring and acid degree values (ADV), these methods have proven to be of little value in detecting and predicting lipase flavors developed during commercial handling of butter. An inability to detect nearly excessive free fatty acids ( F F A ) by sensory analysis and an invalidation of ADV indexing by large, normal variations in individual F F A prevent reliable detection and prediction of lipase flavors in butter. These problems could be overcome if

Received October 12, 1979. 1 Research supported by the College of Agricultural and Life Sciences, University of Wisconsin, and by a grant from Madison Dairy Produce Co., Madison. 1980 J Dairy Sci 63:1058--1064

suitable quantitative data for individual F F A could be obtained for cream and butter on a routine basis. Earlier reports on the gas chromatographic (GC) quantification of F F A in dairy products have included isolation of fat (1) and F F A esterification (1, 9) in the analysis. These procedures frequently lead to distorted F F A profiles that are different from intact samples (8) and are not adapted easily to routine analyses. More recently, the availability of stable polar stationary phases has led to the GC of F F A without derivatization (2, 3, 7, 12), but these reports have not dealt with the establishment of reliable, routine methods. Similarly, Gray (5) described a m e t h o d for the GLC analysis of free Ca to C1s:3 fatty acids in dairy products, but the reliability and suitability of the method for routine applications was not established. This paper presents a GC method for accurately quantifying individual major F F A in butter and cream and culminates extensive efforts to develop a stable GC system for analysis of free acids from intact products. MATERIALS AND METHODS

The overall procedure column, clean-up step for from intact butter or cream followed by GC separation of individual FFA.

involves an openisolation of F F A samples, and this is and quantification

Sample Clean-up Procedure

This technique (14) involves a modification of the combined methods o f Harper (6) and McCarthy and Duthie (13). Materials required include glass chromatographic columns (32 cm by 2.5 cm id) fitted with Teflon stopcocks, acid-washed glass wool (Supelco), Celite Analytical Filter Aid (Johns-Mansville), and anhydrous sodium sulfate (ACS, Fisher). Silicic acid (powder, J. T. Baker) was prepared for use by removing fines by the method o f Harper (6).

1058

FREE FATTY ACIDS IN BUTTER AND CREAM Isopropanolic-KOH was prepared by dissolving 25 g of KOH (85%) in 800 ml of isopropanol (ACS) as described by Keeney (11). Solvents employed were ethyl ether (AR, Mallinkrodt), petroleum ether (b.p. 37 to 57 C, Fisher), acetonitrile (Fisher Certified), acetone (AR, Mallinkrodt), and formic acid (97%+, Aldrich). All solvent systems were stored over excess anhydrous sodium sulfate after preparation. Reference C4, C7, C10, Cl2, C14, and e l 6 F F A were obtained from Aldrich Co.; C6 and Cs from Eastman Co.; and C18, C l s : I , C1s:2, and C1s:3 from Supelco, Inc. The internal standard F F A solution was prepared by dissolving .12% heptanoic acid (C7, Aldrich) in 10% acetonitrile in ethyl ether. The procedure also required 1 N H2SO4, a waterbath (35 C), a rotary vacuum evaporator (Rotavapor R, Buchi), and sample concentrate tubes (glass, 114 mm by 18 mm id with polyethylene stoppers; Laboratory Research Co., Los Angeles, CA). Preparation of the Column Body

Four grams of silicic acid and 1 g of Celite Analytical Filter Aid were mixed with 30 ml of ethyl ether and 10 ml of isopropanolic-KOH, and the mixture was allowed to stand for 5 rain. The slurry then was transferred with ca. 30 ml of ethyl ether to a chromatographic column which was fitted with a small plug of acid-washed glass wool. The column was drained and washed with an additional 100 ml of ethyl ether. During column preparation, nitrogen pressure (1400 kg/m 2) was applied to move solvents through the column.

1059

ethyl ether were held above the body of the column, and the cap material was added slowly through the solvent layer. The column then was washed successively with 100 ml of 20% petroleum ether in ethyl ether and 50 ml of 10% acetonitrile in ethyl ether. Up to this point eluates, including those containing neutral lipids, were discarded, except in studies to determine the extent of fat hydrolysis on the column. Free fatty acids were eluted from the column body with 50 ml of 2% formic acid in ethyl ether followed by 100 ml of .5% formic acid/10% petroleum ether in ethyl ether. Nitrogen pressure (1400 kg/m 2 ) also was used to move solvents through the column during isolation. Final FFA Sample Concentration

The eluate containing F F A was vacuum evaporated with a Rotavapor under water aspiration to ca. 10 ml at 30 to 35 C. The concentrate then was transferred with 5 ml of ethyl ether to a sample concentrate tube. A slow stream of nitrogen was used to distill excess solvent until about 3 ml of cloudy solution was obtained. A solution of ethyl ether/acetone (80:20) was added by drop until an absolutely clear solution was obtained. This usually required 1.2 and 1.6 ml of solvent for butter and cream samples, respectively, Samples then were capped and warmed to 35 C in waterbath before sampling for injection i n t o the GC. The sample preparation and F F A isolation procedure required 2.5 h for a series of replicate columns.

Preparation of the Cap Material

Ten grams of butter were warmed to 25 C, mixed to a salvelike consistency, and then were mixed thoroughly with 7 g of silicic acid, 15 g of sodium sulfate, 3 g 'of Celite, and 1 ml of internal standard solution. The cap material then was acidified by mixing with .2 ml of 1 N sulfuric acid. For analysis of cream, the cap consisted of 15 g of 35% cream, 7 g of silicic acid, 40 g of anhydrous sodium sulfate, 3 g of Celite, 1 ml of internal standard solution, and i ml of 5.5 N sulfuric acid. Isolation of the FFA Fraction

Forty milliliters of 20% petroleum ether in

Gas Chromatography Conditions

The F F A were separated with a 1.8 m long by 2 mm id glass column packed with 10% neopentyl glycol adipate (NPGA; HI-EFF-3AP, Anspec Co.) on 80 to 100 mesh Chromosorb W, AW-DMCS support (Anspec Co.) using a Varian-Aerograph Model 1740-10 gas chromatograph. The column was conditioned .for 12 h at 220 C before use, and it was programmed from 100 to 215 C at 6 C/min during each analysis. The gas chromatograph was equipped with a flame ionization detector (FID) which was supplied with flows of 30 ml of H 2 , 2 4 0 ml o f air, and 30 ml of purified N2 carrier gas per min. Oxygen and water were removed from the Journal of Dairy Science Vol. 63, No. 7, 1980

1060

w o o AND LINDSAY

carrier gas by an inline heated gas purifier (Supelco). Carrier gas was then saturated with formic acid immediately prior to column entry. Formic acid was replenished daily by injecting 1.5 ml into an inline reservoir (15). Injector and detector temperatures were maintained at 225 C. The gas chromatograph was usually operated with a range setting of 10 "1° and x 8 attenuation, and chromatograms were obtained with a Varian Model A-25 recorder operated at a range of 1 mv full-scale response and 76 cm/h. Gas Chromatographic Analysis of the F FA Concentrate

One to five microlitersof sample were injected on-column with a Hamilton 10 bd syringe. The entire GC run required 50 min for the analysis to C4 to e l 8 FFA. Chromatographic peaks were identified by retention time, and peak areas were measured by triangulation. Individual FFA were quantified according to the equation [1]: ppm of FFA = (response factor)(peak area of FFA)(wt of C7 in/~g) (peak area of C7)(wt of sample in g) Response factors required in the above equation were determined with a solution prepared by dissolving approximately .01 to .05 g of each of the C4 to C18:3 F F A and .01 g of C7 standard FFA in 75 ml of acetone/ chloroform (1:1). This solution then was made up to 100 ml and an aliquot was injected into the GC. Response factors for individual F F A were calculated according to the equation [2] : Response factor of F F A = (wt of FFA)(peak area of C7) (wt of C7 )(peak area of FFA) Response factors are in Table 1. The NPGA GC column does not separate effectively Cas F F A (stearic, oleic, linoleic, and linolenic acids). Consequently, a pooled response factor is reported for them. Acid degree values (ADV) were determined by the method of Deeth and Fitz-Gerald (4) and involved titration of an isolated fat sample (9 ml for melted butter and 3.5 ml for cream) with standard base (.02 N methanolic KOH). Journal of Dairy Science Vol. 63, No. 7, 1980

TABLE 1. Flame ionization detector response factors for even-numbered free fatty acids. Fatty acid

Response factor a SD

C4:o C6 :o

1.16 1.10

.04 .05

C8 :o Cxo:o Cx2:o

1.02 1.05 1.10

.04 .03 .02

C,4:0 C16 :o

.94 1.00

.03 .08

C1~ Congeners

1.09

.07

aFire determinations.

RESULTS AND DISCUSSION Sample Clean-up Procedure

The column body functions as an arrestant for FFA as potassium salts when they are swept from the cap material by organic solvents. The isopropanolic-KOH solution in each column contained ca. 25 mg KOH/ml, which is one-half the amount of the original preparation (11). This reduction in base minimized fat hydrolysis on the column and eliminated extremes in variability in the degree of hydrolysis for the original preparation (14). The cap material uniformly distributes the sample in an acidic system. This readily allows selective elution of fats and FFA from other components and permits use of an intact butter or cream sample rather than a fat isolate. Acidification of the cap material below pH 1.5 was necessary to mobilize the FFA in the solvent systems. Addition of dry cap material slowly through a solvent layer prevented the inclusion of excessive amounts of air bubbles into the column when slurry additions were employed. Solvents containing 20% petroleum ether in ethyl ether and 10% acetonitrile in ethyl ether sequentially eluted nonpolar and polar lipids, respectively. Similarly, the longer chain FFA were moved onto the basic portion of the column body by the less polar solvent system while the short chain F F A were mobilized to the arrestant KOH-section of the column by the

FREE FATTY ACIDS IN BUTTER AND CREAM more polar solvent system. Other nonfatty components which were insoluble in the solvent systems were retained in the cap. Salts of FFA retained in the KOH arrestant were eluted after they were converted to the solvent-soluble acid form when the formic acid in the final etuting solvents acidified the column body. Gas Chromatographic Analysis

A typical gas chromatogram of the FFA from butter is in Figure 1. Small peaks were minor FFA in milk fat (10), and these were not quantified. Small amounts of C18:3 were in some butter samples and were measured where possible. Cream had the same general FFA pattern as butter, but peaks were generally smaller than those for butter. For routine GC analysis of major even carbon-numbered FFA, the volume of FFA concentrate after solvent removal but before addition of ethyl ether/acetone (80:20) should be about 3 ml. Further concentration accentuated any phase separations or precipitations and caused sizable impurity peaks that obscured the C4 and C6 peaks. The addition of a few drops ethyl ether/acetone (80:20) and warming to 35 C insured a clear, single phase FFA solution for injection into the GC. This portion of the preparation was critical for reliable FFA analysis. Samples can be held for short periods, but the concentration of C18 unsatured FFA decreased when final F F A samples were stored at 25 C for more than 24 h. Consequently, all F F A concentrates should be analyzed on the day of preparation. The entire procedure for a replicate series of FFA analyses can be completed in 5 h. Cochrane (3) evaluated polar stationary phases for separation of free acids and found that NPGA was one of the most suitable phases available. In the present study, NPGA was unstable when water was in the system. Inclusion of anhydrous sodium sulfate was necessary to reduce the water content of FFA isolates. It was also absolutely necessary to remove traces of water and oxygen from the carrier gas with an efficient, heated gas purifier. However, with suitable precautions, a column life of 2.5 mo can be expected, but after prolonged usage higher molecular weight fatty acid peaks begin to tail. It is then necessary to repack a column to allow accurate peak measurements. Generally,

1061

zu

C16

C4

C7 Cs

C14

C18"2

C10

TIME {min)

Figure 1. Gas chromatogram of free fatty acids (FFA) isolated from butter (sensitivity: range 10-10; attn. X8. C7 F F A added as an internal standard).

the use of a glass column, on-column injection, and formic acid saturated carrier gas were adequate to prevent adsorption and distinct peak tailing. The NPGA stationary phase gives a moderate column-bleed signal during some temperature regimes, but this did not interfere with quantification of the F F A concentrations in butter and cream. Several brands of acetonitrile other than the one reported gave impurity peaks which interfered with measurement of the butyric acid peak. Therefore, solvent systems should be checked for purity when lots are changed. Analysis of Synthetic FFA Mixtures

Synthetic F F A mixtures (C4 to C18: 3) were prepared in acetone/chloroform (1:1) for analysis and were carried through the entire procedure described for butter. Recovery data were obtained for three concentrations of FFA (low, intermediate, and high). Data in Table 2 show that nearly 100% recovery of FFA was achieved for each of the concentrations. The greatest deviations in recovery occurred when peak sizes were small and errors in measurements were magnified. The three levels of FFA were approximately equivalent to the F F A concentrations in 10 g of nonrancid, intermediately rancid, and very rancid butters, respectively. These data demonstrated that in the absence of fat the arrestant column was capable of quantitative recovery of F F A over a wide range of molecular weight and concentration. Reproducibility of the Method

Reproducibility of the method was established for butter analyses, and the data in Table 3 show that the method is reproducible. Results Journal of Dairy Science Vol. 63, No. 7, 1980

1062

WOO AND LINDSAY

TABLE 2. Recovery of synthetic free fatty acid mixtures. Low concentration Fatty acid

~tg Added

Intermediate concentration

Recovery (%)a

~g Added

SD

Recovery (%)a X

High concentration t*g Added

SD

Recovery (%)a X

SD

C4: o

68

79

3

340

98

3

1700

101

1

C~: o

32

137

5

160

119

7

800

112

4

Cs: o

30

103

2

150

124

7

750

111

3

Clo:o

148

104

4

740

106

7

3700

94

2

C~:0

298

108

2

1490

103

5

7450

109

7

C14:0

656

110

7

3280

105

5

16400

105

1

C16:0

1778

107

1

8890

99

2

31115

104

6

C~s Congeners

3138

92

6

15690

91

10

78450

96

5

aThree determinations.

o n t h e silicic acid-KOH a r r e s t a n t p o r t i o n o f t h e c o l u m n . This appears t o o c c u r b e f o r e m a x i m u m c a p a c i t y of t h e c o l u m n b o d y is r e a c h e d a n d indicates t h a t milk fat physically interferes w i t h t r a p p i n g o f t h e relatively n o n p o l a r C18 F F A . However, t h e e x t e n t of r e c o v e r y o f t h e C18 F F A is relatively c o n s t a n t a n d allows q u a n t i f i c a t i o n o f these acids t h r o u g h use o f a n a p p r o p r i ate factor.

were similar for s y n t h e t i c F F A m i x t u r e s a n d c r e a m samples. Recovery of F F A Added to Butter

To verify f u r t h e r t h e validity o f t h e m e t h o d a n d t o d e t e r m i n e the i n f l u e n c e o f fat o n F F A recovery, s y n t h e t i c F F A m i x t u r e s (Ca to C 1 8 : 3 ) were a d d e d to b u t t e r samples b e f o r e analysis. All of t h e F F A were r e c o v e r e d n e a r l y q u a n t i t a t i v e l y w i t h the e x c e p t i o n o f t h e C18 F F A (Table 4). Since t h e C18 F F A are n o n p o l a r , it is likely t h a t t h e y coelute w i t h the n e u t r a l fat in the first s o l v e n t fraction, w h i c h creates an e n v i r o n m e n t less c o n d u c i v e for h o l d i n g t h e m

Extent of Fat Hydrolysis on the Silicie Acid-KOH Column

T h e n a t u r e o f t h e m o d i f i e d silicic acid-KOH a r r e s t a n t c o l u m n suggests t h a t s o m e h y d r o l y s i s

TABLE 3. Reproducibility of the method for replicate analyses of free fatty acids in a butter sample.

FFA concentration (gg/g) Fatty acid

1

2

C4: o

40

3

4

5

X

42

38

41

42

SD 41

1 1

C6: o

16

18

15

16

19

17

Cs: o

12

10

11

13

14

12

1

C1o:0

117

103

119

131

122

118

9

C12:o

213

214

214

238

218

219

9

C14:o

600

554

575

604

566

580

19

C~6:o

1854

1555

1583

1644

1602

1648 107

Cla Congeners

2560

2020

1806

2407

2101

2179 271

Journal of Dairy Science Vol. 63, No. 7, 1980

FREE FATTY ACIDS IN BUTTER AND CREAM

1063

TABLE 4. Recovery of free fatty acids added to butter. #g/lO g butter Fatty acid

Measured endogenous

Amount added

SD

.~

6

346

Measured total

Calculated total

800

SD 6

X 713

Recoverya SD

~

15

(%) 89

C, :o

454

C6: o

140

7

194

334

7

339

9

101

C,: o

81

7

204

285

7

305

8

107

C:o:o

935

24

805

1740

24

1924

33

111

C12:o

2377

11

1500

3877

11

4494

25

116

C1,: o

6558

189

3285

9843

189

9980

418

101

C16 :o

18508

248

10760

29268

248

27827

1129

95

C~ s Congeners

27691

342

14855

42546

342

35020

572

82

aRecovery (%) =

(~g measured) X 100. (t~g present) + (~g added)

of triglycerides still could be induced w h e n fats are in the basic section of the c o l u m n b o d y . To d e t e r m i n e the e x t e n t of hydrolysis, fat f r o m b u t t e r samples was analyzed repeatedly. Eight grams of milk fat simulated the fat c o n t e n t of 10 g o f butter, and new internal standard was added to each of the c o l u m n eluates before it was added to the n e x t column. The first c o l u m n passage trapped most of the F F A in the sample (except for C18). Thus, F F A isolated f r o m

TABLE 5. Correction factors for milk fat hydrolysis occurring during isolation of free fatty acids on silicic acid-KOH columns.

Fatty acid

Column hydrolyzed FFA (#g/g fat) .X

SD

C4:0

9

C6:0

9

2 1

Ca: 0

3

1

C10:0

6

2

C12:0

10

2

C14:0

22

1

C16:0 Cls Congeners

42 31

3 4

subsequent sequential columns could be attribu t e d to fat hydrolysis. The correction factors in Table 5 were developed for b u t t e r and can be applied to prevent o v e r e s t i m a t i o n of F F A for any sample size. Longer chain fatty acids are m o r e a b u n d a n t in milk fat than short-chain fatty acids, and the figures for c o l u m n hydrolysis reflect these concentrations. The data r e p o r t e d herein were obtained with a m o d i f i e d silicic acid-KOH column (14). This c o l u m n is substantially less h y d r o l y t i c than the version described in the earlier literature (13) considered by some workers (9) to give negligible hydrolysis. The standard deviations for each set of fatty acid c o l u m n hydrolysis data along with standard deviations in the sample analyses roughly define the lower limits o f d e t e c t i o n o f individual fatty acids by the m e t h o d . Generally, the lower accurate estimation limit for the C4 to C12 acids is a b o u t 2 p p m on a fat basis. Longer chain acids (Cx4 to C18) can be quantified d o w n to a b o u t 30 p p m on a fat basis, because the c o l u m n trapping efficiency for C18 increases w h e n c o n c e n t r a t i o n is decreased greatly. Free Fatty Acid Composition of Butter and Cream Samples

Analytical data for e x a m p l e b u t t e r and cream samples are in Table 6 along with correJournal of Dairy Science Vol. 63, No. 7, 1980

1064

WOO AND LINDSAY

TABLE 6. Free fatty acid (FFA) composition o f butter and cream samples. FFA concentration a Sample

C4:0

C6:0

C=: o

C10:0

Cx2:o

Acid

C~4:0

C~6:0

C1 s congeners Total

degree value

Flavor quality

Bland, no criticism Bland, no criticism Full flavor, no criticism Full flavor, no criticism Distinctly rancid Distinctly rancid Strongly rancid Soapy rancid Bland, no criticism

(ppm) Butter A

5

1

11

60

119

293

933

1571

2993

1.53

Butter B

2

0

2

26

43

93

347

833

1346

.81

Butter C

26

3

4

32

60

114

381

847

1467

.79

Butter D

29

12

16

73

141

282

900

1438

2891

1.72

Butter E

71

23

26

78

97

172

565

1494

2526

1.22

Butter F

87

41

59

152

148

280

884

2481

4132

2.11

Butter G

118

46

57

148

135

244

870

2373

3991

1.88

Butter H

58

14

11

122

311

878

2447

3446

7287

4.39

Fresh cream (35% fat)

13

11

9

20

29

60

264

515

921

1.08

aData corrected for fat hydrolysis.

sponding ADV and flavor descriptions. Distinct patterns or profiles of FFA can be noted for each of the samples. The FFA profiles appear t o r e l a t e well t o t h e f l a v o r a s s e s s m e n t s . T h e s e r e l a t i o n s h i p s are u n d e r i n v e s t i g a t i o n in a comprehensive continuing study of rancid f l a v o r d e v e l o p m e n t in b u t t e r a n d o t h e r d a i r y products.

8 9

10

REFERENCES

1 Bills, D. D., L. L. Khatri, and E. A. Day. 1963. Method for the determination of the free fatty acids of milk fat. J. Dairy Sci. 46:1342. 2 Byar, B., and G. Jordan. 1964. An efficient packed column for free fatty acids. J. Gas Chromatogr. 2:304. 3 Cochrane, G. C. 1973. The analysis of free fatty acids by gas chromatography. Proc. Soc. Analyt. Chem. 10:212. 4 Deeth, H. C., and C. H. Fitz-Gerald. 1976. Lipolysis in dairy products: A review. Australian J. Dairy Tech. 31:53. 5 Gray, I. K. 1975. The quantitative analysis of free fatty acids in dairy products. New Zealand J. Dairy Sci. Tech. 10:158. 6 Harper, W. J. 1953. Direct chromatographic determination o f acetic, propionic and butyric acids in cheese. J. Dairy Sci. 36:808. 7 Hrivnak, J., and V. Palo. 1967. Separation o f

Journal of Dairy Science Vol. 63, No. 7, 1980

11

12

13

14

15

C~-C t s free fatty acids in dairy products by dual column PTGC. J. Gas Chromatogr. 5:325. Humbert, E. S., and R. C. Lindsay. 1969. Comparison o f methods to determine the free fatty acid content o f butter. J. Dairy Sci. 52:1862. lyer, M., T. Richardson, C. H. Amundson, and A. Boudreau. 1967. Improved technique for analysis of free fatty acids in butteroil and provolone cheese. J. Dairy Sci. 50:285. Jensen, R. G., J. G. Quinn, D. L. Carpenter, a n d J . Sampugna. 1967. Gas-liquid chromatographic analysis o f milk fatty acids: A review. J. Dairy Sci. 50:119. Keeney, M. 1956. A survey o f United States butterfat constants. II. Butyric acid. J. Assoc. Offic. Agric. Chem. 39:212. Mayberry, W. R., and G. J. Prochazka. 1965. Polyethylene glycol 600: An efficient liquid phase for alcohols and free fatty acids. J. Gas Chromatogr. 3:232. McCarthy, R. D., and A. H. Duthie. 1962. A rapid quantitative method for the separation of free fatty acids from other lipids. J. Lipid Res. 3:117. Woo, A. H., and R. C. Lindsay. 1980. An improved silicic acid-KOH arrestant column for the quantitative isolation of free fatty acids. J. Am. Oil Chem. Soc. (Submitted). Woo, A. H., and R. C. Lindsay. 1980. Simple apparatus for saturating gas chromatography carrier gas with formic acid for free fatty acid analysis. J. Chromatogr. Sci. (In press).