608 Analyst, September, 1974, Vol. 99, pp. 608-611 A Potentiometric Titration Method for the Rapid Determination of Salt in Meat Products BY M. KAPEL AND J. C. FRY (Procter Department of Food and Leather Science, Lceds University, Leads, LS2 9 J T ) h method is described for the rapid determination of salt in meat pro- ducts by titration of the macerated sample with silver nitrate. The acidic and oxidising conditions specified permit the use of an ion-selective electrode with a silver sulphide membrane in the potentiometric detection of the end-point. The method is slightly more precise and very much more rapid than existing routine and reference methods. MOST existing methods for the determination of salt in meat products depend upon either an aqueous extraction or an ashing pre-treatment with the subsequent titrimetric deterrninatioii of chloride, typically by the Volhard or Mohr m e t h ~ d .l - ~ Alternatively, the meat can be digested with nitric acid in the presence of silver nitrate.* Such pre-treatments are slow, and the methods based on their use involve several further time-consuming operations. In addi- tion, ashing must be carried out with great care and in the presence of sodium carbonate in order to minimise the risk of loss of hydrogen chloride, while extraction procedures are not applicable to certain meat products. This latter fault arises because it is essential, in the Volhard titration, that chloride be precipitated as silver chloride and that the precipitate should be either filtered off or effectively removed from the system by a surface coating of a substance such as nitrobenzene.The colloidal precipitates of silver chloride that are produced by some meat products (e.g., luncheon meat) cannot be removed in this way and accurate titration is therefore impossible. The following method overcomes the above difficulties and is generally applicable to meat products, being accurate and faster than existing methods. It depends upon the direct titration of chloride in a sample macerate under conditions that permit the use of an ion-selective electrode to detect the end-point. In this context certain amino-acids and their residues in proteins seriously interfere with solid-state electrodes that include silver salt membranes. In particular, histidine and cysteine are known to form silver complexes,5~6 which have been shown to interfere with the solubility product equilibria and hence the potentials of such electrodes.The mode of action of these amino-acids will be reported elsewhere, but the conditions required for the elimination of such interference are embodied in the present method. These conditions are, firstly, a low pH to ensure that protonation of the imidazole group in histidine takes place, and, secondly, the presence of an oxidising agent to convert cysteine into cystine. Advantage has been taken of the necessity to provide a low pH in order to effect control over the ionic strength of the macerated sample by using a strong acid at a high concentration. The ionic strength and, consequently, activity coefficients and liquid-junction potentials, are thus maintained virtually constant.A mercury - mercury(I) sulphate reference electrode was employed in preference to the more widely used calomel instrument, as the latter would cause difficulties with the diffusion of chloride into the sample solution as well as the precipitation of silver chloride at the liquid junction. EXPERIMENTAL APPARATUS- All readings were made by means of an ion-selective electrode with a silver sulpliide membrane (Philips 1s 650-S) and a mercury - mercury(1) sulphate reference electrode, with saturated potassium sulphate solution as the outflowing electrolyte (Activion Glass Ltd.) . These electrodes were connected to a pH meter (Corning - EEL Digital 110) used in the millivolts mode. 0 SAC and the authors.An MSE homogeniser was used in order to macerate samples.KAPEL AND FRY 609 REAGENTS- All of the inorganic reagents used were of analytical-reagent grade. Background sohtion-This was a 0.01 M solution of potassium dichromate in 1 M nitric Titrant-A 0.171 M solution of silver nitrate in 1 M nitric acid. Emulsijier-Teepol-L detergent. SAMPLE PREPARATION- In accordance with British Standard methods,l bulk meat samples, normally greater than 200 g, were rendered homogeneous by at least two treatments with a meat mincer that had plate holes less than 4 mm in diameter. They were then thoroughly mixed and stored in air-tight containers. PREPARATION OF MACERATED SAMPLES- which were added about 50 ml of background solution and about 0.2 ml of Teepol-I1. meat was then macerated on the homogeniser for not less than 15 s.DETERMINATION OF CHLORIDE- The macerated sample was carefully rinsed into a 250-ml beaker with approximately 50 ml of background solution. While the liquid was being stirred by means of a magnetic rotor, the electrodes were introduced into the sample, which was then titrated to a potential of +30 mV with the silver nitrate solution titrant (Note 1). The salt content was obtained by substitution of the values into the following equation- acid. About 10 g of the prepared meat sample were accurately weighed into a vortex flask, to The Volume of titrant used (ml) Mass of sample taken (g) Sodium chloride, per cent. mlm = After each determination, the electrodes were rinsed in distilled water and the silver sulphide membrane was dried on a tissue by use of very firm pressure and a polishing motion (Note 2).COMPARISON WITH EXISTING METHODS- The above procedure was compared with two reference methods for a range of meat products. The reference techniques were : firstly, a modification of the British Standard method for the “Determination of Chloride in Meat and Meat Products,”l in which the British Standard method was followed, except that deproteinated extracts were first filtered and then made up to volume, the reverse order being specified in the standard; and secondly, preliminary ashing for 3 hours at 450 “C under an ample layer of sodium carbonate, this process being followed by dissolution in a slight excess of nitric acid, filtration and a Volhard titration, as in the British Standard method.From every bulk sample six sub-samples were taken and analysed for chloride by all three methods. RESULTS In Table I NA indicates that analyses by the modified British Standard method were attempted but that colloidal silver chloride precipitates were produced, so that no results were TABLE I SALT CONTENTS OF VARIOUS MEAT PRODUCTS KESULTING FROM THE USE OF DIFFERENT ANALYTICAL METHODS Bulk sample Canned sausage Chopped ham with pork Canned ham Canned meat balls Luncheon meat Corned beef Salami Salt, mean per cent. m/ni on “as received” basis 7- A \ Modified BS Ashing Present method method method 1.36 1-37 1-36 NA 1.43 1.45 3.38 3.38 3.41 NA 0.99 1.00 NA 1-95 1.93 2.02 1.95 1.97 6-34 6.03 5.92 Best estimate of standard deviation 7-- \ Modified RS Ashing Present method method method 0.009 22 0.0198 0.007 00 NA 0.0156 0.006 24 0.02 15 0.0227 0-0124 NA 0-006 06 0.004 89 NA 0-008 11 0.0137 0.007 14 0.0211 0.009 06 0.115 0-0905 0-0537610 KAPEL AND FRY: A POTENTIOMETRIC TITRATION METHOD [Analyst, vO1.99 obtained. minations. error relative to the mean result indicated under “Present method” in Table I, are shown. ,4s already explained, each of the values in the table is based upon six deter- In Table I1 the expected limits of error for the proposed method, recorded as percentage TABLE I1 . ERRORS IN PROPOSEI) METHOD Relative error limits, per cent. Bulk sample Canned sausage Chopped ham with pork Canned ham Canned meat balls Luncheon meat Corned beef Salami for 95 per cent. population (1.9ti x s.d.) $I 1.01 3: 0.843 A0.713 f 0.958 f 1.39 *0.901 f 1.78 for 99 per cent.population (2-58 x s.d.) f 1.33 f l . 1 1 *Om938 5 1-26 f 1.83 f 1-19 4 2-34 Correlation coefficients, calculated from the mean results in Table I, were : modified British Standard method with proposed method, r = 0.999 37 ; ashing method with proposed method, The results indicate a close correlation between the proposed method and the two refer- ence procedures. In a majority of the experiments the proposed technique proved to be more precise than the reference methods and, for most materials, it can be expected to give results to within about &l per cent. relative error. = 0.999 86. RAPIDITY OF THE METHODS- The British Standard method took about 34 hours per single sub-sample, excluding the bulk sample preparation time; the ashing procedure required rather more time, while the proposed method furnished single results in 10 to 12 minutes with a routine throughput of about eight samples per hour.Although virtually the whole duration of the proposed method comprises operator time, the actual working time involved is at most only one third of that required by either of the reference methods. The proposed method can also be carried out with an auto-titrator, a device that would both reduce the total operator time and increase the throughput. NOTES- The exact end-point potential depends upon the electrodes used and is best determined by replicate titrations of standard sodium chloride solution with standard silver nitrate solution under conditions identical with those under which subsequent determinations are to be made. The quoted value of f 3 0 mV was obtained by means of this method and can be expected to be correct only for the electrodes specified.The proposed method is unsuitable for use with silver-metal electrodes, which are attacked by the background solution. The silver sulphide membrane of the ion-selective electrode used is also dissolved slowly, and the indicated drying procedure is accordingly necessary in order to prevent the formation of a porous surface layer, with consequent deterioration in the speed of response of the electrode. 1. 2. C o N c L u s I o N The proposed method yields results which show excellent correlation with those obtained by use of two reference methods over a range of typical samples.The results obtained with the proposed method are usually of greater precision than those furnished by the reference methods. The proposed method also makes available a large saving in operator time, being rapid in execution. We gratefully acknowledge the generous financial assistance for this work provided by J. Sainsbury Ltd., to whom one of us (J.C.F.) is also indebted for a maintenance grant. Our thanks are also due to Professor A. G. Ward for his interest and encouragement.September, 19741 FOR THE RAPID DETERMINATION OF SALT IN MEAT PRODUCTS 61 1 REFERENCES 1. “Methods of Test for Meat and Meat Products, Determination of Chloride Content,” British 2. “Official Methods of Analysis of the Association of Official Analytical Chemists,” Eleventh Edition, 3. Pearson, D., “The Chemical Analysis of Foods,” Sixth Edition, J. and A. Churchill Ltd., London, 4. “Official Methods of Analysis of the Association of Official Analytical Chemists,” Eleventh Edition, 5 . Greenstein, J. P., and Winitz, M., “Chemistry of the Amino Acids,” Volume 3, John Wiley and 6. -,- , op. cit., p. 1973. Standard 4401 : Part 6 : 1970. The Association of Official Analytical Chemists, Washington, D.C., 1970, p. 398. 1970, p. 378. The Association of Official Analytical Chemists, Washington, D.C., 1970, p. 392. Sons, New York, 1961, p. 1885. Received February 25th, 1974 Accepted March 2292d, 1974