ISSN:0003-2654    

DOI:10.1039/AN95075FX009

出版商:RSC    

年代:1950

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN95075BX011

出版商:RSC    

年代:1950

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN95075FP017

出版商:RSC    

年代:1950

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN95075BP021

出版商:RSC    

年代:1950

数据来源: RSC

摘要:

MARCH, 1950 Vol. 75, No. 888 The Estimation of Tomato Solids in Tomato Products by a Method Involving the Determination of Lycopene by Absorption Spectroscopy BY F. G. STOCK SYNOPSIS-A method is proposed for the estjmation of tomato solids in tomato products by the determination of lycopene by absorption spectroscopy. The history of work done on the tomato pigment is traced. The theoretical aspects of the determination of lycopene in the presence of carotenes are discussed in the light of experiments carried out by the author. A method giving consistent and reproducible results has been evolved for the extraction of lycopene from tomato products. A suggestion is advanced that the origin of the extracted pigment, whether from a ripe or green tomato, is indicated by the shape of the absorption curve obtained.A series of five graphs is used as an illustration. The lycopene content of (a) English tomatoes in various stages of ripeness, and (b) commercially available concentrated purkes are given. The method is applied to sauces and ketchups at present available. THE Food Standards (Tomato Ketchup) Order, S.R. and 0. No. 1817, 1949,* prescribes a standard for tomato ketchup, catsup, sauce and relish. A tomato solids content of not less than 6 per cent. by weight is specified, and the Order comes into operation for sales by retail on October lst, 1950. The characteristic red pigment of the tomato was first investigated in 1876 by Millardet,l who obtained it in a crystalline state, and also observed the crystals in the flesh of the ripe fruit. He established that absorption in the visible portion of the spectrum of a carbon disulphide solution of the pigment was characterised by two bands in the green region and one in the blue.Later investigators2~3~4~5 found that carotene was present in tomatoes, *See Analyst, 1950, 75, 112. 1 1 i118 STOCK: ESTIMATION OF TOMATO SOLIDS I N TOMATO PRODUCTS BY A [VOl. 75 and considered it to be identical with the pigment obtained by Millardet. The tomato pigment was again studied in 1903 by Schunck,6 who found that the red colouring matter was clearly distinguishable from carotene in appearance, crystal form, solubility and absorption spectrum. The first chemical investigation of lycopene was undertaken by Montanari7 in 1904. His results led him to consider that the tomato pigment was a condensation product of two molecules of carotene, having the formula C,,H,,.In 1910, Willstatter and Escher8p9 isolated from Italian tomato conserve a relatively large quantity of the pure crystalline red pigment, together with a smaller quantity of the yellow pigment, carotene. They found lycopene to be an unsaturated hydrocarbon and to possess the same composition and molecular weight as carotene, namely C4oH5,. It was less soluble than carotene in ether, carbon disulphide, light petroleum and alcohol, and oxidised and bleached more readily. Xantho- phylls and xanthophyll esters were also reported to be present. As a result of the work of Karrer and others,10,11~12,13~14,16 the following structural formula has now been assignedu to the pigment- CH, CH3 CH3 CH, I I I (CH,) 2C : CH-CH,CH,*(C : CHCH : CH) ,*C : CHCH : CHCH : (CCH : CHCH), : dCH&H,CH : C(CH$ , Willstatter and Escher, as well as Montanari, carried out their investigations on the pigment isolated from Italian-grown varieties.Matlack and Sandole proved the pigments contained in Italian-grown tomatoes and in American red and purple tomatoes to be identical. Lubimenko17 considered lycopene to be formed by the oxidation of chlorophyll, but Euler et aZ.18 found that green tomatoes kept at 20" to 21" C. ripened normally in a few days, developing the usual red lycopene colouring. At 30" C. they developed only a yellow colour, and tests showed the presence of carotene, xanthophyll and an unidentified flavone dye, but no lycopene.At 37" C. they turned yellow and spoiled rapidly. Neither light nor its absence had any effect on the rate of colouring. It is believed that lycopene is formed by an enzyme action which is inhibited by the higher temperature. Kuhn and Grundmann16 stated that the chlorophyll content of the green fruit is far too small to account for the formation of lycopene from it. Smithla showed that protection from intense light favours lycopene formation, whereas light is conducive to the maximum production of carotene. Since lycopene is present in mature fruits which have been grown in the complete absence of light and which have therefore never contained chlorophyll, lycopene cannot be an oxidation product of chlorophyll as has been assumed by some investigators. The oil-soluble colour of the tomato has been suggested as a basis for an approximate estimation of tomato solids in tomato products.20 The natural colour of the tomato was extracted with light petroleum, and its visual density determined by means of a Lovibond tintometer fitted with the Rothamsted device.Three analytical techniques are of use in the study of a problem of this nature, namely- (a) The separation of carotenoids by the partition technique between immiscible solvents as described in detail in Thorpe's Dictionary, volume 11.21 (b) Chromatographic adsorption. (c) Spectroscopic criteria using absorption maxima in different solvents. These afford much assistance in the task of characterising carotenoids. EXPERIMENTAL A chromatographic study of the pigment extracted from ripe tomatoes by light petroleum Three zones (i) At the bottom of the column, a rather indistinct, yellowish, diffuse zone, viz., (ii) A well-defined reddish zone, viz., lycopene, in the middle.(iii) At the head of the column, a pinkish ill-defined zone, viz., xanthophylls. A rough idea of the respective proportions present could be obtained from the depths of the various zones. Very approximately, the pigment appeared to be composed of about 90 per cent. of lycopene, the remainder being carotenes with only a very small amount of xanthophylls. This assumption is borne out by the work of Kuhn and Grundmann on the const it u t ion of 1 ycopene. For work of b.p. 80" to 100" C., was made in this laboratory, using an alumina column. separated- a- and p-carotenes.The spectral absorption curve of the extracted pigment was next examined.March, 19601 METHOD INVOLVING THE DETERMINATION OF LYCOPENE 119 in the visible region, a light petroleum (b.p. 80" to 100" C.) solution was employed, and a chloroform solution for the investigation of absorption in the ultra-violet. uch general absorption was observed in the ultra-violet, and it was found necessary to extract the chloro- form solution with 90 per cent. methanol, after which satisfactory absorption curves were b t I 4 cu go.9 V - $0.8 0 v c 0-7 0 $0.6 X - - - - * - 0 4 0.3 0.2 01 00. Fig. 2. - - - - \ ' 1 1 I - ' ' 1 ' 1 ' 1 1 1 ' 1 ' 1 Absorption curve in ultra-violet spectrum for methanol-washed chloroform solution 0.1 -120 [Vol. 75 obtained with very little irrelevant absorption.The spectral absorption curve of the pigment dissolved in light petroleum (b.p. 80" to l0OOC.) was next studied (Fig. 1). Solutions in chloroform and in carbon disulphide were also examined. The following maxima were recorded- STOCK: ESTIMATION OF TOMATO SOLIDS IN TOMATO PRODUCTS BY A Observed Recorded Solution in maxima, mp. maxima, mp.* Light petroleum . . .. .. 605 475 450} cf. { 606 474 444 Chloroform . . .. .. .. 610 480 455 513 480 451 * Morton,Be chart V, p. 60. This identified the pigment as being chiefly lycopene. As further evidence, the ultra-violet curve of a chloroform solution, extracted with 90 per cent. methanol to remove substances giving irrelevant absorption, was determined (Fig. 2). This showed a well-defined maximum at 305 mp., which is in agreement with recorded data.Furthermore, the extinction coefficient at 305 mp. was one-third of that at 475 mp., which again is in accordance with published results.= The shape of the lycopene curve is extremely characteristic. The absorption curves obtained in different solvents have the same shape, but there is a "shift" in the wavelengths of the maximum absorption peaks of the curves. If we compare the carbon disulphide and light petroleum curves, we find that the 505mp. maximum in light petroleum is displaced to 545mp. in carbon disulphide, i e . , towards the red end of the spectrum. The solution does, in fact, appear quite red in carbon disulphide, whereas in light petroleum it is yellowish- red. The effect of carotene on the shape of the lycopene curve was studied using a light petroleum extract of carrots as a source of carotenes.The absorption curve of this solution is shown in Fig. 3. The point of interest is that the absorption at 505mp. is practically Carbon disulphide . . .. .. 5.45 605 - 548 507 - f I \ k lycopene \ ' " L L 1 I r ' ' ' " a ' 1 ' ' A 600 590 580 570 560 550 540 530 520 510 500 490 480 470 460 450 440 430 420 & Wavelength my. Fig. 3. Comparison of a mixed a- and ,%carotene curve and a lycopene curve (light petroleum solutions) negligible, and the first absorption maximum is reached at 485 mp., thus enabling lycopene to be estimated spectroscopically in the presence of carotenes. As confirmation, equal volumes of two solutions of approximately equal concentration, as indicated by the extinction coefficients, were mixed, and the mixed solution evaporated to the same volume.The shape of the absorption curve for this solution is shown in Fig. 4. The actual maximum at 505 mp. was practically unaltered, but the shape of the curve is vastly different. If we assume that the lycopene/carotene ratio in the pigment from ripe tomatoes approximates to 9/1, then Fig. 4 is the curve of a mixture of lycopene and carotenesin the ratio of approximately 9/11.March, 19M)J METHOD INVOLVING THE DETERMINATIOX OF LYCOPEKE 121 0.3 m r n I L 430 60 go Wavelength my. Fig. 4. Effect on the lycopene cume of the addition of ct- and /3-carotenes (light petroleum solutiOn) Fig. 5 . 0.0 6do 5b sio s;o sio s ; ~ s40 s;o sio 5io A 4;o do 4;o 4 ~ 0 4;0 4;0 i;o 4;o 4to Wavelength mp.The absorption curve of the pigment extracted from a green tomato (light petroleum solution)122 STOCK: ESTIMATION OF TOMATO SOLIDS IN TOMATO PRODUCTS BY A [VOl. 76 This confirms the findings of Mills,= that, by selecting suitable wavelengths, accurate quantitative spectral analysis of carotenoid mixtures can be made with an error of less than 2-5 per cent. Another question arising here is the possibility of determining by the shape of the curve whether a given pigment originates from a green or a ripe tomato. This would be possible if the relative proportions of lycopene and carotene altered upon ripening. Kuhn and Grundmann in their work on the constitution of lycopene gave the following figures for the lycopene, carotene and xanthophyll content of the tomato pigment.Green, Half ripe, Fully ripe, mg. per 100 g. mg. per 100 g. mg. per 100 g. Lycopene * . .. . . 0.11 0.84 7.85 Carotene . . .. .. 0.16 0-43 0.73 Xanthophyll . . .. . . 0.02 0.03 0-06 Xanthophyll esters . . . . 0.00 0.02 0.10 This indicates that by inspection of the shape of the curve an inference regarding the origin of the pigment may be drawn. The ratio Emax 505mp./Ema, 475mp. is approximately 8/10 in a normal ripe tomato pigment, and the use of this ratio is now suggested as a criterion. The curve obtained from the pigment of a green tomato is shown in Fig. 5, and its shape is obviously very different from that obtained for the ripe fruit. In order to obtain results on a statistical basis for the lycopene content of tomatoes and tomato products, a method was evolved which gave consistent and reproducible results.Extraction with light petroleum of an aqueous solution gave rise to troublesome emulsions, and it is also doubtful whether a complete extraction of the lycopene is obtained in this way. The use of anhydrous sodium sulphate, followed by hot extraction in a Bolton extractor, was also unsatisfactory. PROPOSED METHOD The method finally decided upon was to dry the material, well mixed with sand, in a vacuum desiccator, and then, after powdering, to transfer to a Bolton extractor. Acetone was used as solvent, light petroleum not being very successful, owing to a form of chromato- graphic separation taking place on the sand column. Further, the penetration of the column did not seem as effective with light petroleum as with acetone.PROCEDURE Weigh accurately about 5 g . of the purke, sauce or ketchup into a stainless steel or porcelain dish. Incorporate carefully, and thoroughly, sufficient sand to make a powdery mass. Allow to dry in an efficient vacuum desiccator over calcium chloride for at least 16 hours. Transfer to a mortar and powder thoroughly. Pack tightly into the liner of the Bolton extractor. Extract with acetone for 3 hours, using a 100-ml. Quickfit pyrex flask, keeping it away from bright sunlight. Evaporate the acetone until nearly dry, then blow the last drops off in a current of air, with the flask removed from the water-bath. Add light petroleum (b.p. 80" to 100" C.) and heat to boiling on the water-bath, cool, and make up to a convenient volume, usually 150 ml.Determine the absorption at 505 mp. in a l-cm. cell, using a constant-deviation spectro- meter in conjunction with a Hilger photometer, a ribbon filament lamp being employed as light source. The full absorption curve may be plotted if required, special note being made of the Em,,. 475mp. Normally, however, only one determination a t 505mp. is required, and this is made in duplicate. If the solution is left in the dark overnight, the E alters very little, but it is advisable t o make the determination without delay. A light petroleum (b.p. 80" to 100" C.) solution of the pigment was refluxed for 3 hours and gave the same reading before and after the refluxing. So for all practical purposes, there is very little destruction of the lycopene under the conditions of the experiment.It was found in a series of experiments that approximately 70 to 80 per cent. of the lycopene was extracted in a few minutes, and 85 to 90 per cent. in about half an hour. To obtain a 95 to 98 per cent. extraction, however, 3 hours were required. It was possible to obtain still more complete extraction by prolonging the time beyond 3 hours, but the further amount extracted, even after prolonged boiling, was proportionately very small. The method outlined above will extract at least 95 per cent. of the lycopene present, and since all theMarch, 19501 METHOD INVOLVING THE DETERMINATION OF LYCOPESE 123 determinations are made under the same conditions, the figures obtained are comparable.The method will give consistent results on the same sample. The question of complete extraction is closely connected with the distribution of lycopene in the tomato. The bulk is easily extracted, but a proportion appears to be held in the fruit in some way. Information on this question is difficult to find, but it seems probable that, with a more complete extraction, the figures obtained by Darbishire on commercial purkes, using a Lovibond tintometer, would have shown a smaller variation than the 40 per cent. he reported.20 Extraction for 3 hours gives the acetone solutions a cloudy appearance, but if the experimental details regarding the evaporation of the acetone are closely followed, the resulting light petroleum solution is quite clear. Light petroleum (b.p.80" to 100" C.) was used as final solvent because of its comparatively high bding-point, so that there is less risk of concentration due to evaporation when transferring the solution to the l-cm. cell of the photometer. Furthermore, the 505 mp. maximum is a very easy one to determine. In calculation it is assumed that the E: 2;. 505 mp. for lycopene in light yetroleurn is approximately 2000, and the Eft:! is calculated for the product under examination. In the case of tomato puries, the total solids, and salt, if any, were also determined, and hence the true tomato solids content. The lycopene content was expressed as (a) micrograms of lycopene per gram of product, or (b) micrograms of lycopene per gram of tomato solids (on the purkes only). RESULTS ,4 number of determinations were made on tomatoes in various stages of ripeness, after preparing purkes by mashing and passing through a sieve to obtain a purke free from skins and seeds.I t will be observed that the lycopene The results obtained are given in Table I. TABLE I THE LI-COPENE CONTENT OF P U R ~ E S PREPARED FROM EKGLISH-GROWN TOMATOES I N VARIOUS STAGES OF KIPENESS Description of tomatoes Green . . .. .. Red-green . . .. Ripe, red-green .. Red, firm (very small) Red, firm . . .. Red, firm . . .. Ripe, red, firm . . Very ripe, some rotten Total solids .. 5.4 . I 4-8 .. 5.4 . . 10.8 . . 7.0 . . 5-2 .. 4.4 .. 5.2 Lycopene +per g. , assuming &&. 505 mp. = 2000, PLg. 3.2 16-0 3.3 16.5 6.6 28.0 12-8 64-0 15.8 79-0 12.5 62-5 12.9 64.5 22.0 110 E:? 505 mp. Lycopene per g.of tomato solids, CLg. 296 344 519 693 1128 1201 1466 2115 content gradually increases, with a very considerable variation between green and over-ripe tomatoes. It is worth noting that, for an average ripe, red, firm tomato, the concentration of lycopene approximates to 1500pg. per g. of tomato solids. With respect to the actual commercial products available to the sauce manufacturers, it should be remembered that nearly all, if not all, concentrated tomato purkes used in this country are imported, and consist of a product manufactured from ripe tomatoes. Unripe tomatoes are used for tomato chutney rather than for tomato sauce or ketchup. In a bulletin issued by the National Canners' Association Research Laboratory of the U.S.A.,25 reference is made to the exclusion of greenish tomatoes in the preparation of purke, because of the development of an objection- able colour.Not only do green tomatoes, or tomatoes with green portions, not have the desired amount of red colouring matter, but the yellow and greenish particles mask and dull the red colour present. Also the amount of pectin naturally occurring in tomatoes is greatest at the time of complete ripeness; this is important, because the pectin present in the finished product contributes to its viscosity. The avoidance of over-ripe tomatoes also is in the interests of the manu- facturer, because the absence of mould is of great importance. The colour of tomato products is an important index of their quality, and greatly influences their commercial value. The manufacturers of purees have a financial interest in avoiding the use of green, partly green and over-ripe tomatoes, and it is highly improbable that they will do anything likely to decrease the commercial value of their products. Thus we may be reasonably assured of a Good pulp can be made only from thoroughly ripe, sound tomatoes.124 STOCK: ESTIMATION OF TOMATO SOLIDS IN TOMATO PRODUCTS BY A [VOl.75 fairly uniform concentrated purke being commercially available to sauce manufacturers] and this is borne out by Table 11. TABLE I1 THE LYCOPENE CONTENT OF MANUFACTURERS’ COXCENTRATED PUREES AND ALSO OF TWO SAMPLES Origin (u) MANUFACTURERS’ PURBES Italian “Rago” . . . . .. Italian “Valnure” ,. .. Italian “Helvia” . . . . .. I talian “Suprema’ ’ . . .. Italian “Soriso” .. . . .. Italian “Catalano” . . .. Portugese “Toiro” . . .. Hungarian “Gschwindt” .. Hungarian “Helins” . . . . Hungarian “Globus” . . . . Hungarian “Golden Pheasant” . . French “U.D.C.” I . .. French “Gourmet” . . . . French “Rolli” . . . . . . Canadian “Smith” . . . . South African “Barclay Vale” . . Australian “O.T. Ltd.” . . . . Total solids, % 33.9 26.4 26.7 33.1 32.5 28.3 37.9 28.0 25.8 27.2 27.6 30-7 27.5 28.1 25-9 27.4 23-8 (b) CANNED UNCONCENTRATED PURBES Australian .. . . . . 11.0 New Zealand . . . . . . 11.2 OF UNCONCENTRATED PURBE NaCII, Yo 2.3 1.9 1.6 3-1 0.7 1.5 2.4 0.2 0.2 0.2 0.2 2.3 0.3 4.5 0.3 0-4 1.9 0.7 0.9 Tomato Lycopene solids, E:%?’ 605 mp. per g . , % PE5 31.6 24.5 25.1 30-0 31.8 26.8 35-5 27-8 25-6 27.0 27.4 28.4 27.2 23-6 25-6 24.0 21-9 108 87 106 91 107 87 110 85 95 82 87 70 92 69 07 77 75 540 435 530 455 535 435 550 425 475 410 435 350 460 345 485 385 375 Average 10-3 31 155 10.3 33 165 Lycopene per g.of tomato solids, 1709 1775 1677 1517 1683 1624 1550 1529 1856 1518 1588 1652 1691 1462 1505 1604 1712 1620 1505 1601 The variation in lycopene content is surprisingly small, and this is all the more remarkable when we consider that these purkes represent random samples from the produce of seven different countries, covering practically the whole world. It would seem a reasonable assumption that 1 g. of tomato solids contains a minimum of 1500 pg. of lycopene, and it is proposed that this figure should be used as a means of estimating the tomato-solids content of an unknown sauce or ketchup.In the manufacture of such products, colouring matter may occasionally be used to maintain a standard colour. In this case, however, a water-soluble dye is used, which is not extracted by the proposed method. In any event, the shape of the absorption curve of lycopene is so characteristic that an analyst, if ever confronted with an oil-soluble dye, would be able to detect its presence from the shape of the curve. It is very unlikely, however, that this would be encountered. A number of sauces and ketchups were analysed, with the results given in Table 111. The samples giving low figures were further investigated, and in every case the shape of the absorption curve was the normal curve for lycopene. The E values at 475mp. bore normal relationships to the E values a t 505mp., so that the use of green tomatoes was excluded.To test the accuracy of the method, a number of samples of ketchup of known tomato- solids content were obtained. A series of five samples of ketchup, obtained at different stages in the manufacture of a single batch of tomato ketchup, was analysed, and the lycopene content was found to be practically constant, showing that there is very little loss during the manufacture. Furthermore, about a dozen different purkes were included in this batch of over a ton of ketchup, and the tomato-solids content of each one was known, so that the theoretical cQmposition of the batch could be calculated. The calculated tomato solids approximated to 12.5 per cent., and the actual percentage found by lycopene determination averaged 11.5 per cent. over the five samples.The manufacturers themselves from time to time had analysed their product by Morpeth’s method,26 and this gave figures ranging between 11 and 12 per cent. A ketchup containing exactly 33 per cent. of tomato purke was obtained from one manufacturer, together with a sample of the purke from which it was made. The Eto2:5 505mp. of the purke was 94 and that of the ketchup 29. The solids content of theMarch, 19501 METHOD INVOLVING THE DETERMINATION OF LYCOPENE 125 purke was 27.0 per cent. after allowance for a small amount of salt. The calculated tomato- solids content of the ketchup was therefore 9.0 per cent., and the amount found by lycopene determination was 9.7 per cent. TABLE I11 ,~LPPLICATION OF THE METHOD TO THE ESTIMATION OF THE TOMATO-SOLIDS CONTENT OF A NUMBER OF TOMATO SAUCES AND KETCHUPS AVAILABLE FOR RETAIL SALE (a) TOMATO SAUCES Sample .. .. .. .. 1 E:E% 505 mp. . . .. . . 14 Lycopene, pg. per g. . . . . 70 Tomato solids, assuming 1500 pg. lycopene = 1 g. of tomato solids . . .. .. . . 4.7 Sample . . .. .. .. 10 * . . . 14 lm. 605 mp. . . Lycopene, pg. per g. . . .. 70 Tomato solids, assuming 1500 pg. lycopene = 1 g. of tomato solids . . .. .. . . 4.7 Elo” % (b) TOMATO KETCHUPS Sample . . .. .. .. 18 E:%% 605 mp. . . .. . . 26 Lycopene, pg. per g. . . .. 130 Tomato solids, assuming 1500 pg. lycopene zz 1 g. of tomato solids . . .. .. . . 8.7 Sample . . .. .. . . 27 .. .. 18 Elcm. 505 mp. . . Tomato solids, assuming 1500 pg. lycopene E 1 g. of tomato solids .. .. .. . . 6.0 loo a‘& Lycopene, pg. per g. . . . . 90 2 27 135 9.0 11 20 100 6.6 19 26 130 8.7 28 26 130 8.7 3 21 105 7.0 12 34 170 11.0 20 43 215 14.3 29 30 100 6.7 4 17 85 5.7 13 29 5-8 1.9 21 30 150 10.0 30 31 155 10.3 5 2.1 10-5 0.7 14 21 105 7.0 22 24 120 8.0 31 32 160 10.7 6 24 120 8.0 15 24 120 8.0 23 36 180 1290 32 55 275 18-3 7 0-5 2.6 0.2 16 17 87 6.2 24 12 60 4.0 33 22 110 7.3 8 15 75 5.0 17 11 2.1 0.7 25 31 155 10.3 34 25 125 8.3 9 18 90 6.0 $8 19 75 5.0 From a series of six samples of tomato ketchup supplied by another manufacturer the following results were obtained. Two samples made from Hungarian purke to contain 15.3 per cent. of tomato solids were found to give E:Eh 505 mp. of 40 and 39, corresponding to a tomato-solids content of 13.3 and 13.0 per cent.respectively. Two samples made from French puree to contain 15.0 per cent. of tomato solids were found to give El$:% 505 mp. of 48 and 47, corresponding to a tomato-solids content of 16.0 and 15.7 per cent. respectively. A further two samples made from French pur6e to contain 13.7 per cent. of tomato solids both gave E:::! 505 mp. of 42, corresponding to 14.0 per cent. of tomato solids. One important point remains to be stressed, namely the desirability of defining “tomato solids.” In the experiments on the lycopene content of tomatoes and concentrated purbes, approximately 5 g. were taken for the solids determination, a little water was added, and the mixture was dried to an even film on the water-bath for approximately 1 hour, followed by drying at 100” C.in the oven for 1 hour. Various methods are given in the literature for tomato-solids determination,” q., vacuum desiccation, refractometer methods, etc. It would be in the interests of all concerned if the Food Standards (Tomato Ketchup) Order, S.R. and 0. No. 1817, 1949, defined “tomato solids,” and prescribed a standard method for determination. SUMMARY AND CONCLUSIONS It has been demonstrated that the lycopene content of the tomato and tomato products can be ascertained by a spectroscopic method using the spectral absorption at 505 mp. as a criterion. Furthermore, the effect of carotene on the absorption maximum has been shown to be negligible at 605 mp., even though the shape of the remainder of the absorption curve *See report of A.M.C.Tomato Products Sub-Committee, Analyst, 2941, 66, 319.126 STOCK: ESTIMATION OF TOMATO SOLIDS IN TOMATO PRODUCTS alters as the percentage of carotene increases. It is suggested that the relationship between Em,,. 505 mp. and Em,,. 475 mp. can be used as a criterion of the degree of ripeness of the tomatoes used. A method of extracting and determining the lycopene which gives consistent results has been developed. A suggested factor is given, based on a number of determinations of the lycopene content of English-grown tomatoes in various stages of ripening, and also of commercially available tomato purkes from which sauces and ketchups are normally manu- factured. The factor converts micrograms of lycopene to grams of tomato solids. A table is given of the results obtained in the application of the method to a number of ketchups and sauces at present available.The method does not claim extreme accuracy, but it gives a definite indication of tomato-solids content, and is undoubtably a good “sorting out” test, which may be used in conjunction with other methods of analysis. The analytical procedure involved is very simple, and the determination rapid. [Vol. 75 The author is indebted to the Birmingham Public Health Committee for facilities for carrying out this work, and also wishes to record his thanks to Mr. H. H. Bagnall, the Birmingham City Analyst, for his very helpful encouragement. The writer is very grateful to Mr. D. J. Munns, the Chief Chemist at H.P. Sauce Ltd., Birmingham, for supplying the samples of commercial purke.REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. Lubimenko, V., Rev. Gen. Botan., 1914, 25, 476. 18. 19. 20. 21. 22. Morton, R. A., Vitamins, Adam Hilger, London, 1942. 23. 24. 25. 26. Millardet, A., J. Bot. Jahresber., 1876, 4, 783. Arnaud, A., Compt. Rend. Acad., 1886, 102, 1119. Passerini, N., Staz. Sper. Agrar. Ital., 1890, 18, 545. Kohl, F. G., Untersuchungen ubey des Cavotins und seine physiologischefi Bedeutung it& der Pjianze, Ehring, C., Dissertation Munich, 1896. Schunck, C. A., Proc. Roy. SOC. Lond., 1903, 72, 165. Montanari, C., Staz. Sper. Agrar. Ital., 1904, 37, 909. Escher, H. H., Zur Kenntnis des Carotins und des Lycopiiis, Proxnotionsarbeit, Zurich, 1909. Willstatter, R., and Escher, H. H., 2. physiol. Chem., 1910 64, 47. Karrer, P., and Widmer, R., HeZv. Chim. Acta, 1928, 11, 751. Karrer, P., and Bachmann, W. E., Ibid., 1929, 12, 285. Karrer, P., Helfenstein, A., and Wehrli, H., Ibid., 1930, 13, 88. Karrer, P., Helfenstein, A., Wehrli, H., and Wettstein, A,, Ibid., 1930, 13, 1084. Karrer, P., Helfenstein, A., Pieper, B., and Wettstein, A., Ibid., 1931, 14, 435. Kuhn, R., and Grundmann, C., Bey. Chem. Ges., 1932, 65, 898. Matlack, M. B., and Sando, C. E., J . Biol. Chem., 1934, 104, 407. von Euler, H., Karrer, P., von Krauss, E., and Walker, D., Helv. Chinz. Acta, 1931, 14, 164. Smith, A., Cornell Univ. Agr. Expt. Sta. Mem., 1936, 187, 3. Darbishire, 0. B., Analyst, 1948, 73, 457. Thorpe’s Dictionary of Applied Chemistry, 4th Edition, Vol. 11, 1949, p. 399. Brode, W. R., Chemical Spectroscopy, Wiley, New York, 1946, p, 253. Mills, E. S., Plant Physiol., 1934, 9, 193. Bigelow, W. D., Smith, H. R., and Greenleaf, C. A., “Tomato Products,’’ Bulletin No. 27-L, Morpeth, J. C., Analyst, 1948, 73, 449. Leiflsic, 1902, p. 41. 194 1, Nat. Canners’ Assoc. Research Laboratory, Washington. CITY ANALYST’S LABORATORY BIRMINGHAM

ISSN:0003-2654    

DOI:10.1039/AN950750117b

出版商:RSC    

年代:1950

数据来源: RSC

摘要:

March, 19501 KENT- JONES AND MARTIN 127 A Photo-Electric Method of Determining the Colour of Flour as Affected by Grade, by Measurements of Reflecting Power BY D. W. KENT-JONES AND W. MARTIN (Read at the meeting of the Society on Wednesday, October 5th, 1949) SYNOPSIS-A piece of apparatus that has been designed for assessing flour colour as an indication of flour grade is described. Its readings are independent of the effect of natural or artificial bleaching. It consists in a source of light rays that are directed by a lens system on to standard white surfaces of magnesium oxide, the reflected light from which is received by two screened photo-electric cells with filters having their main transmission in the 630 mp. band. After balancing the optical system by means of mechanically operated shutters, one of the standard surfaces is replaced by a similar surface of flour paste and the decrease in the amount of reflected light determined by reducing the light that reaches the other standard surface by means of a cam-operated shutter until the system is again balanced.The duller the flour, the greater the “cut-off” required. The apparatus is easy to manipulate and a numerical evaluation of the flour colour can be obtained in five minutes. THE problem of evaluating flour colour, which is influenced by the length of extraction and is therefore a measure of the grade of the flour, is of commercial importance, but the difficulties involved in making useful and sufficiently exact measurements are considerable. In the main. the factors controlling the colour of flour are- (1) The grade of the flour, which depends upon the extent of the contamination of the ground-up endosperm by fragments of the outer coverings of the grain, i.e., b r a and associated substances.(2) The degrees of yellowness due to the amount of the natural yellow colouring matter present and the extent to which this has been removed by natural and artificial bleaching, or by both. (3) The granularity; the more granular the flour the darker and duller it appears. Dullness from this cause will not persist in the crumb of the resulting loaf. (4) The presence of dirt, smut (I’iZZetia tritici) and other extraneous matter. In pre-war days, when white flour was the normal product commercially made, minor differences in colour often affected the price by several shillings per sack.Even to-day, when 85 per cent. extraction has to be made, there is still considerable interest and competition in obtaining the best possible colour in flour. Despite considerable work on the subject, there is as yet no simple and satisfactory method available for measuring and recording numerically the colour of flour as an index of grade. What is known as the Pekar test, which involves the visual comparison of compressed slabs of flour, has remained the stand-by of millers and bakers for the assessment of the comparative brightness of flour samples. This test is open to much criticism, as is indicated by Kent-Jones, Amos and Martin.l It is true that the literature abounds with methods which have been suggested for the measurement of flour colour but, as is pointed out by Kent-Jones, Amos and Martin,l these have not been found sufficiently rapid and reliable to gain general acceptance. A useful distinction between the yellowness due to natural yellow pigments and dullness principally due to bran powder contamination of the endosperm was made by Kent-Jones and Herde2 This convenient division of flour colour has been generally accepted and has been adopted in this investigation.The degree of yellowness of flour can be easily determined by extracting the unoxidised carotene, under standardised conditions, and measuring the intensity of the yellowness acquired by the solvent. Many such methods have been Some criticism of the early method of Kent-Jones and Herd2 was made by Visser’t Hooft and128 KENT- JONES AND MARTIN : A PHOTO-ELECTRIC METHOD OF de Leeuw,ll but this is not applicable to the modified procedure of Kent-Jones and Amos.f2 The dullness factor, which Kent- Jones and Herd2 called the “grade colour” as opposed to the creaminess factor, is much more difficult to measure and has received much less attention.Kent-Jones and Herd2 devised a method of measuring this factor by extracting the bran pigments, but certain weaknesses in this method were pointed out by Markley and Bailey.13 It has been our experience that this grade colour-or “brightness” as millers and bakers call it, and which is so noticeable in the crumb of the loaf-is correlated with the extent to which a smooth surface of flour in paste form reflects light.The method of measuring this aspect of flour colour which is described in this paper relies upon this correlation. Preliminary work revealed that the influence of differences in granularity could be overcome by the use of a flour paste and that the effect of differences in degree of bleach could be largely eliminated by the employment of light of a prescribed wavelength. In this paper we shall describe the apparatus which was eventually evolved and the technique advised for the accurate comparison and numerical recording of grade colour by the measurement of the reflecting power of the surfaces of flour pastes. We, and those associated with us, have studied this problem for many years, and, although not unaware of the difficulties involved and the almost impossible task of devising a method to which no objection could be raised, we have found that the procedure described in this paper is of considerable practical use.lyol. 75 PRINCIPLE OF THE METHOD- The principle of the method is to utilise a balanced circuit containing two photo-electric cells to measure the amount of light of a particular wavelength which, under the conditions of the test, is reflected from the surface of a paste prepared from the flour. The amount of light reflected is not measured in absolute units but is recorded as a proportion of the amount of light of the same wavelength which is reflected from a standard white surface. As explained previously, the determination is performed upon a paste of the flour in order to obviate the differences which would otherwise arise between flours which were similar in grade and colour but which differed in granularity.Whereas previous methods have suffered from the defect of insensitivity, the apparatus and technique described in this paper enables the operator, uninfluenced by personal judgment, to assess flour colour quite as accurately as can the experienced and skilled miller and baker and moreover to express the result numerically. The data recorded in this paper reveal that different operators obtain results in close agreement and, as a simple method of standardisation is available, all instruments should give, within the normal experimental error, the same results. APPARATUS- The apparatus (Fig. 1) consists of a source of light, L, which is a short filament 36-watt lamp (12 volts) fed from the normal A.C.mains supply via a transformer. This lamp projects light via the lens systems, X and Y , which are designed to give parallel light, on to two standard surfaces, S, and S,, which consist of glass cells, approximately 5 cm. square and 1 cm. thick, containing heavy magnesium oxide. The reflected light from the standard surfaces is picked up by the photo-cells, C and D. Immediately in front of each photo-cell is a filter, W (Wratten No. 58), having its main transmission in the 630 mp. waveband. It has been established that the reflected light from the surface of a flour paste transmitted by this filter is not influenced to any appreciable extent by the degree of bleach, whether natural or artificial. Thus the amount of light transmitted by the filter is substantially dependent purely upon the grade of flour, k., upon the amount and nature of the bran powder present.This green filter was selected after considerable experimental work with an earlier form of colorimeter, namely, that of Bolton and Williams,l* from which the principle of comparing the light reflected by the surface of a sample under test against that reflected by a standard white was adopted. The photo-cells are connected in the usual way to a galvanometer with a tapping key in circuit so that the null point can be easily obtained. In front of lens X is a shutter B (Figs. 1 and 2) operated by the “set zero” control, K, by means of which a fixed portion of the beam of light passing to cell S, can be cut off when K is raised.It is found necessary to have this arrangement because of the marked difference in reflecting power of the standard surface and the surfaces presented by the flour pastes. If this cut off were not used, the dial G would have to be rotated considerably before the null point was reached, even with a flour of high grade, and hence an appreciable portion of the scale on dial G would be wasted. By suitable adjustment of the amount of light cut offMarch, 19501 DETERMINING THE GRADE COLOUR OF FLOUR 129 by the “set zero” control, K, which can be adjusted by means of screw F (Fig. 2), the range of the instrument can be altered to suit the materials being tested. There is a second shutter, E, in front of the lens X, which is operated by a screw ter- minating in a knob, J, on the panel and’this serves to effect the initial balance of the current after the “set zero” knob has been raised.In front of the lens Y is a cut-off screen, A, operated by the calibrated dial G, through a linear camshaft. The total movement of the cam is very small, being approximately 0-16 inch for a complete revolution of the dial. The setting of the instrument adopted by the authors at the present time will cover all grades of flour from the brightest patent flour in the mill (C flour) to flours as dull as those C 0 P L Fig. 1 given by flour of 90 per cent. extraction. With this setting, there should be a fixed reading on the dial when a Kodak gelatin filter of 80 per cent. transmission is placed in front of the standard surface S, (heavy magnesium oxide), and the instrument operated as if testing flour paste.If a reading of 80 is not obtained, the screw adjustment F must be rotated (this alters the extent of the movement of the shutter B operated by the “set zero” control K), until the instrument, upon being rebalanced as usual for zero, does give a reading of SO when the SO per cent. transmission filter is in position in front of surface S,. The differences obtained in dial readings when different samples of Kodak filters of 80 per cent. transmission are used are very small, so that all instruments can be standardised to yield substantially the same results provided that the magnesium oxide is of the same quality and does not change. The range covered by this instrument is naturally intimately bound up with the cam movement and the original cam fitted to the instrument had only half the movement of the one at present in use.With this very fine cam movement.the instrument was, of course, even more sensitive than it is at the moment, but the range of brightness which could be measured was naturally much narrower; in our opinion the present cam gives a more useful form of instrument. Bolton and Williams14 pointed out that it was necessary to remove infra-red light in order to obtain a good result with this type of instrument. They accomplished this by using a cell of dilute copper sulphate solution. We have not used a copper sulphate cell in our instrument on account of practical disadvantages, such as the need for fairly frequent replacement of the solution and the difficulty of maintaining the inner surfaces of the cell in a clean condition.Instead, we have incorporated in the light path of our instrument In practice, with flour, we have selected the dial reading of 80.130 KENT-JONES AND MARTIN A PHOTO-ELECTRIC METHOD OF [vol. 75 a special glass (Chance’s No. ON19), labelled in Fig. 1, H, which has the power of absorbing a high proportion of the infra-red rays. The extreme sensitivity of the instrument is shown by the fact that mere reversal of the faces of the cell may give a reading difference as great as 10” on the dial-the dial being calibrated 0” to 360”. It is, therefore, most important to make sure in any series of tests that the same cell face is always used. We have found that normal mains fluctuations do not call for the use of a stabilising transformer but, if the supply were rather abnormal, a transformer of this type might be necessary.J K SET ZERO Fig. 2 METHOD Switch on the lamp L and raise the “set zero” control K, thereby cutting down by a pre-determined amount the light falling on the standard surface S, and thus on the photo-cell C. Set the dial G at zero and balance the photo-cells by adjusting the zero control J (Fig. 2) until, on using the tapping key (not shown in diagram as it is part of the electrical circuit), there is no movement on the galvanometer, which indicates that both cells are receiving the same amount of light. Remove the standard cell S, and replace by a similar cell into which has been poured a flour paste prepared by mixing together, until smooth and homo- geneous, 30g.of flour and 50ml. of distilled water. Depress the “set zero” control K, which then allows the full amount of light to fall on the sample cell, and rotate the calibrated dial G from its zero setting, cutting down the light from S, until a null point is obtained on the galvanometer, indicating that both cells are again receiving the same amount of light. Record the degrees through which the scale G has been rotated. GENERAL REMARKS ON THE METHOD- The figures obtained from the calibrated dial are purely empirical, but they do serve to give a numerical representation of the grade of flour; the higher the dial reading, the lower the grade of the flour, Le., the duller the colour. The results can be very easily duplicated and, as bleaching has practically no effect at the wavelength of light employed, flours can be re-checked, if necessary, after a lapse of time, such as several weeks, although in the interim they will have undergone natural bleaching.The reproducibility of results with a given instrument is quite satisfactory, as can be seen from the data of the tables which follow. The concordance of the results provided by different models of this instrument will, of course, be influenced by the accuracy with whichMarch 19501 DETERMINING THE GRADE COLOUR OF FLOUR 131 the cam and its associated shutter is manufactured. Naturally we have no experience of the “spread” of the data which would be provided by different models of this instrument, but we do not contemplate any serious difficulty in obtaining reasonably close duplication with different models.This is essentially a matter for the manufacturers and depends upon the uniformity of material and construction. With accuracy in making the cam and with standard magnesium oxide to the specification indicated, there is no reason why all instruments checked by the Kodak transmission filter should not give reasonably similar results. I t is, however, important to bear in mind two matters. We have, in fact, worked with other matched cells and these have given results almost identical with the cells originally used. Secondly, there may be errors introduced from the nature cf the filaments in various lamps as the light thrown on the surface being examined is not com- pletely uniform.This is an important matter and lamps must be selected so as to have the filament in the same relative position, which can be conveniently judged by the operator. An arrangement for pre-focussing the lamp assists in overcoming the trouble when lamp replacement becomes necessary. We have not, so far, encountered any serious difficulties due to differences in power of reflection of the heavy magnesium oxide, provided the material has approximately the same granularity. We used B.D.H. heavy magnesium oxide, all of which passes a No. 25 flour silk (approximately 197 meshes per linear inch or 38,809 per square inch). It has been our practice to test a flour paste within a few minutes of its being mixed and, when a series of flours is to be tested, we find it convenient to weigh out all the samples at one sitting and then to mix each of them with water when its turn for testing arrives.It is necessary to maintain absolute cleanliness in the sample cell surfaces because very small differences in reflecting power are being recorded. DISCUSSION OF RESULTS AND SUMMARY- The degree of reproducibility attainable is revealed in Table I, which gives the dial readings furnished by a number of flours at different times. The dial can be read to half a degree. I t should be noted that the readings in this and subsequent tables are those read directly from the dial and not the corresponding figures of the empirical scale recommended in the paper by Kent-Jones, Amos and Martin.l TABLE I REPRODUCIBILITY OF RESULTS Dial readings obtained on separate flour pastes Firstly, the variation which might arise with different photo-electric cells.Sample First day Second day One week later Long patent flour .. . . 48, 48, 48.5 49, 50, 50.5 50.5, 50, 61 Lower grade flour .. . . 211.5, 210.5, 212 212.5, 210, 210-5 208, 207, 206 Little effect is noticed if the amount of water used in making the paste is varied. Thus, if instead of 50 ml. of distilled water, 60 ml. is used for the 30 g. of flour, there is noappreciable change in the dial reading, as is shown in Table 11. TABLE I1 EFFECT OF VARYING QUANTITY OF WATER USED IN MAKING FLOUR Quantity of water added to 30 g. flour PASTE Sample Flour 1 2 3 r 50 ml. 203 48 121 3 60 ml. 202 45 120 TABLE I11 REPRODUCIBILITY BY DIFFERENT OPERATORS Sample Operator I Operator I1 Operator I11 4 37 38 39 5 255-5 254 257 6 70 69 71132 small, as is shown by the figures in Table 111.in the result if the test is carried out within half an hour of the initial mixing. the results obtained on the same paste tested after intervals of 30 minutes. KENT-JONES AND MARTIN : A PHOTO-ELECTRIC METHOD OF [Vol. 75 The differences in readings obtained on a flour by different operators is remarkably As indicated earlier, we use the paste as soon as it is made, but there is no marked change Table IV gives TABLE IV EFFECT OF TIME ON THE READING Readings A r 7 Sample As made 30 minutes later 1 hour later A 50.5 51.5 57 B 71 74 81 C 155 158 I66 Table V gives results indicating in a broad way that the readings of grade colours furnished by this apparatus are substantially independent of the degree of bleach.This matter is, however, discussed in more detail by Kent-Jones, Amos and Martin.1 TABLE V EFFECT OF BLEACh Sample Reading A. Unbleached .. .. .. 50.5 A. Bleached . . .. .. .. 48 B. Unbleached .. .. .. 70 B. Bleached . . * . .. .. 72 C. Unbleached .. .. .. 155 C. Bleached . . .. .. .. 153 The nature of the skins of the grain or of the branny particles present in the flour does have some influence on the colour of the resulting flour. Thus, branny particles from white wheats have rather less deleterious effect with respect to colour than those of red wheats. The difference in reading obtained, when finely ground white and red bran is added at the rate of approximately 6 per cent. to a sample of flour, is of the order of 15".Also bran powder, as opposed to the same weight of bran in larger form, gives a poorer colour to the flour as assessed by commercial judgment. It is, therefore, of interest to check the results obtained under these conditions, although it should be pointed out that normally the differences in commercial flours in these respects are not great. Our broad experience is that when one flour is brighter than another, solely because of commercial differences in the bran size, there is a difference in dial reading of less than 10". It is claimed that, by means of this instrument, it is possible to express, on a reliable numerical scale, the commercial evaluation of flour colour and grade, uninfluenced by the effect of bleach. The method is rapid (5 minutes per sample) and gives easily reproducible results. We wish to record our thanks to Mr. F. Widdis of Messrs. H. Tinsley & Co., Ltd., for his interest in this work and his great assistance in the final design of the apparatus, to Dr. A. J. Amos for his advice and assistance in preparing this paper, and to Mr. R. Donaldson of the National Physical Laboratory for helpful criticism of the instrument. 1. 2. 3. 4. 5. 6. 7. 8. REFERENCES Kent-Jones, D. W., Amos, A. J., and Martin, W., Analyst, 1950, 75, 133. Kent-Jones, D. W., and Herd, C. W., Ibid., 1927, 52, 443. Markley, M. C., and Bailey, C., Cereal Chem., 1935, 12, 33. Binnington, D. S., Hutchinson, W. S., and Ferrari, C. G., I b i d . , 1941, 18, 10. Binnington, D. S., Sibbitt, L. D., and Geddes, W. F., Ibid., 1938, 15, 119. Ferrari, C. G., and Bailey, C. H., Ibid.. 1929, 6, 218. , , Ibid., 1929, 6, 347. , , Ibid., 1929, 6, 457. -- --March, 19501 DETERMINING THE GRADE COLOUR OF FLOUR 133 9. 10. 11. 12. Kent-Jones, D. W., and Amos, A. J . , Modern Cereal Chemistry, Northern Publ., Co., Ltd., 13. 14. LONDON, W.5 Ferrari, C. G., Croze, A. B., and Bailey, C. H., Ibid., 1932, 9, 491. Ferrari, C. G., and Croze, A. B., Ibid., 1934, 9, 505. Visser’t Hooft, F., and De Leeuw, F. J . G., Ibid., 1927, 5, 351. Markley, M. C., and Bailey, C. H., CereaE Chem., 1935, 12, 40. Bolton, E, R., and Williams, K. A., Analyst, 1935, 60, 447. Liverpool, 1947. 88, MADELEY ROAD

ISSN:0003-2654    

DOI:10.1039/AN9507500127

出版商:RSC    

年代:1950

数据来源: RSC

摘要:

March, 19501 DETERMINING THE GRADE COLOUR OF FLOUR 133 Experiments in the Photo-Electric Recording of Flour Grade by Measurements of Reflecting Power BY D. W. KENT-JONES, A. J. AMOS AND W. MARTIN (Read at the meeting of the Society on Wednesday, October 5th, 1949) SYNoPsrs-The reliability and usefulness of the instrument described at the same meeting by Kent-Jones and Martin for the determination of the grade colour of flour has been investigated. It is recommended that the grade colour of a flour should be expressed as the dial reading in degrees divided by 10 and corrected to the nearest 0.5 unit. Statistical analysis of the data of reproducibility shows that if the grade colour figure for a given flour is assessed by different operators on different days the results can be expected to agree within 0.5 unit.With flours of ash content below 1 per cent., a range that embraces flours from high-grade patent to 90 per cent. extraction flour, the effect of bleach is within the experimental error of the method. The dial readings of the instrument are linearly related to the logarithms of the corresponding ash contents and the correlation coefficient between dial reading and ash content is 0.959. Some of the aerating ingredients used for self-raising flour influence the dial reading, but it is shown that this influence can be annulled by using a buffer solution in place of water for making the paste. With this modification the instrument gives reliable results with self-raising flour. THE quality of flour is determined by two factors, namely, the suitability of the flour for its intended purpose, eg., bread-making or cake-making, and the colour of the flour.This latter aspect of quality is of such importance that in pre-war days it regulated to a great extent the price of flour. The “colour” of flour is, however, a complex characteristic that is largely governed by two independent factors; these are, first, the brightness or dullness, which is related to the extent to which the endosperm has become contaminated with powdered wheat skins and associated substances during the milling process and, secondly, the degree of creaminess which is dependent upon the amount of carotene in the original wheat and the extent to which this has undergone natural or artificial bleaching. The former aspect of flour colour, i.e., the dullness or brightness, is referred to as the “grade” colour because it is correlated with the degree of extraction, that is, the grade of the flour.Since the two factors which together mainly determine the colour of flour are quite independent, any degree of brightness, or dullness, can be associated with any tint between the full creaminess of an unbleached flour and the extreme whiteness of a highly bleached one. The possible variations in the over-all colour of flour, therefore, cover a very extensive range from a very bright and very white flour on the one hand to a very dull and fully creamy flour on the other. The great commercial importance which attaches to flour colour has led to an evaluation of this characteristic becoming a very frequently applied test in mill and bakery.The earliest method of assessing flour colour, and one which is still the standby of millers and bakers, was a visual test-the Pekar test-in which a comparison is made between the surface of com- pressed slabs of flour before and after they have been dipped in water. The immersion of the slabs in water tends to magnify the colour differences through the action of enzymes,134 KENT-JONES, AMOS AND MARTIN: EXPERIMENTS IN THE [Vol. 75 probably oxidative in nature, which exist in the flour. The visual appearance of such a compressed slab of flour depends, however, not only upon the over-all colour of that flour, but also upon other factors such as the degree of compression to which the slab has been subjected, the moisture content of the flour, the conditions of the immersion of the slab in water and the length of time which has elapsed since the immersion.Although this test does undoubtedly serve a useful purpose in the flour industry, it is only a comparative test, which does not lend itself to numerical expression and which relies upon personal judgment. Attempts have been made to devise methods of evaluating the over-all colour of flour which permit the i-esults to be recorded on a numerical basis. Jagol described a method in which the colour exhibited by the surface of a compressed slab of flour was matched visually against a combination of red and yellow glasses in a Lovibond tintometer. Baker, Parker and Freese2 decided that the over-all colour of flour embraced four colour factors, namely, yellow, red, black and white and that a satisfactory match could be obtained by visual comparison with Maxwell discs rotated at high speed.These methods, however, have never gained popularity in the cereal field and no method of measuring the over-all colour of flour on a numerical basis has become generally accepted. In the milling industry, of course, interest is taken not only in the over-all colour of flour, but in each of the two main independent colour factors, since one, the grade colour, reflects the efficiency with which the milling operation has been performed, and the other, the creami- ness, is a measure of the degree of bleach. It is only natural, therefore, that attention should have been given to the independent determination of each of these factors.The determination of the degree of creaminess is a relatively simple matter involving only the extraction of the carotene with a suitable solvent and the measurement of the intensity of colour in the resulting solution. I t is not surprising, therefore, that numerous methods of performing this test have been recorded. Investigations into this determination have been reported by Kent- Jones and Herd3; Markley and Bailey4; Binnington, Hutchinson and Ferrari6; Binnington, Sibbett and Geddese ; Ferrari and Bailey’ p 8 p9 ; Simpsodo ; Visser’t Hooft and De Leeuwll ; Kent- Jones and Herd12; Kent-Jones and Amos.ls The determination of the “grade” colour of flour is a more complex problem, but it also has received attention in the past.Kent-Jones and Herd3 reported a method involving the extraction of the pigments of the admixed bran powder, i.e., the powdered wheat skins and the measurement of the colour of the resulting solution against a standard solution. This method furnished reliable and reproducible results with flours of 75 per cent. extraction or less, but did not prove to be so satisfactory when applied to long extraction flour such as the present 85 per cent. National flour. The grade colour or brightness of flour is, however, related to the extent to which a smooth surface of the flour reflects light-as is evident from the long-used Pekar test-and Kent- Jones and Martin14 have recently u tilised this relation- ship as the basis of a new procedure for measuring grade colour. They have devised and described an instrument which is capable of measuring fine differences in the proportion of incident light reflected by a flour surface and, by employing a flour paste and light of a prescribed wavelength, have made the measurement independent of variations in the granularity of flour and in the degree of bleach.The technique of Kent-Jones and Martin14 is an advance upon previous procedures for assessing grade colour in two directions. First, the method is identical in principle with the procedure employed throughout the flour industry for judging flour colour in that it depends upon the reflecting power of a flour surface and, secondly, relying as it does upon the reactions of photo-electric cells, it is independent of personal judgment. A numerical evaluation of the grade of flour as provided by the instrument of Kent-Jones and Martin,14 supplemented by a numerical measure of the degree of bleach obtained, for example, by the method of Kent- Jones and Amos,ls should provide a complete picture of the over-all colour of the flour and, moreover, reveal the extent to which each of these two main colour factors contributes to the whole.The purpose of this paper is to present the data which we have accumulated during our experience with the method of Kent-Jones and Martin, and to illustrate the value which the method can have for those in the various sections of the cereal industry who are called upon to make frequent assessments of flour colour. Although we have restricted our present review to wheat flours, the instrument can be used to advantage for measuring the colour of many other cereal products and undoubtedly has possibilities in other fields.March, 19501 PHOTO-ELECTRIC RECORDING OF FLOUR GRADE 136 METHOD OF EXPRESSING RESULTS- The circular scale carried by the Kent- Jones and Martin instrument and from which the final readings are taken is calibrated in degrees.With the instrument adjusted so as to give a reading of SO" when balance is effected after an 80 per cent. transmission screen has been inserted in front of the standard cell, as advocated by Kent-Jones and Martin, pre-war patent flours give readings of about 15" to 40°, while flours of 85 per cent. extraction give readings of about 100" to 140". There is much to be said for the use in the cereal industry of a less widely spread scale, particularly as many millers in this country are accustomed to the grade colour scale of Kent-Jones and Herd,3 which ranged from 4.5 to 6.0 for patent flour to 8.5 to 10.0 for basic grades. A scale in this region for the grade colour, which is easily understood by the miller, has the further advantage that it falls in line with the widely used bleach figure scale of Kent-Jones and Herd3 which runs from about 2 to 10.With these considerations in mind we decided to narrow the scale and we have done this by accepting as the "grade colour figure" the figure obtained by dividing the dial reading in degrees by 10 and correcting to the nearest 0.5. REPRODUCIBILITY OF DIAL READINGS- In view of the possibility of this method being adopted by chemists in both the milling and the baking industries, that is by representatives of buyers and sellers of flour, it is most important that the results it gives should show good reproducibility.It is not sufficient that duplicate readings on a given flour paste should show good agreement; the spread of results obtained upon a flour by different operators on different days must be small. The reliability of the method from this angle was therefore determined by statistical analysis of appropriate experimental data. The design of this experiment involved the determination of the dial readings of five flours by two operators on each of 4 days. The operators were unqualified assistants accustomed to performing routine tests. The experimental results appear in Table I.TABLE I DIAL READINGS OF FLOURS DETERMINED BY TWO OPERATORS ON DIFFERENT DAYS Dial readings Flour No. A I 3 1 2 Operator A 1st day . . .. .. 104 122 2nd $7 . . . . .. 108 120 3rd 39 . . .. .. 101 121 4th >> . . * . . . 96 118 Operator B 1st day . . .. . . 100 123.5 2nd 93 . . .. . . 98 119.5 3rd 99 . . .. .. 111 132 4th 33 . . .. .. 102 126 Statistical analvsis of these- data shows that the ratio 3 4 5 146 159 34 150 161 33 147 158 35 143 157 32 153 160 - 150 158 31.5 148 161 32 146 158 32.5 of "between-operators" variance for the "between-dad' variance to the error varianck is 2.01, and the corresponding figure is 2.13; for these ratios to attain significance at a probability level of 0.05 t h e i would need to exceed 4-84 and 3-59 respectively.The standard deviation is 3.2". These data mean that if the grade colour figure for a given flour is assessed by different operators on different days, the results (expressed as dial readings divided by 10) may be confidently expected to agree within h0.5 unit. INFLUENCE OF DEGREE OF BLEACH ON DIAL READINGS- If the instrument of Kent-Jones and Martin does in fact do what it purports to do, that is measure the grade colour or brightness of a flour, then it is essential that the dial reading given by a flour should be uninfluenced by any bleach, natural or artificial that has been conferred upon the flour. Kent-Jones and Martin state that they have rendered the dial reading substantially independent of degree of bleach by the insertion of a filter in the light path.In their paper they quote three tests in support of their contention but state that further tests may be desirable. Since the maximum difference between the dial readings before and after treatment obtained in the tests quoted by Kent-Jones and Martin was136 KENT-JONES, AMOS AND MARTIN: EXPERIMENTS IN THE [Vol. 75 less than the standard deviation established by our experiments, we extended the study to a wider range of flours. A wider discrepancy than that recorded by Kent-Jones and Martin between the readings given by a flour in the unbleached and in the bleached state would not be perturbing, provided it was still within the experimental error of the method. In these experiments flours of various ash contents, obtained from several mills, were tested on the instrument in the unbleached state and also after the flours had been bleached by the addition of one of the bleaching agents in use in the milling industry.The bleaching agents were applied in the proportions in which they are commonly used in the industry. The data obtained in these experiments are given in Table 11. TABLE I1 EFFECT OF BLEACHING REAGENTS ON THE DIAL READING Colour figures (dial reading divided by 10) 1 7- Dial readings Sample Ash, Bleached Unbleached Bleached Unbleached % Benzoyl peroxide, 1/16 oz. per 280 1b.of flour K1 K2 K3 K4 K5 B1 B2 s 1 s 2 s 3 s1 s 2 s 3 s 1 s 2 s 3 0.49 0.61 0.67 0 75 0.88 0.43 0.99 0.42 0.46 1-15 0.42 0.46 1-15 0.42 0.46 1-15 54 77 86 93 116 42 126.5 34 48 124 53.5 72.5 84 93 114.5 31.5 124.5 28 40 119.5 Nitrogen trichloride, 7 g.per 280 Ib. of flour 34 29.5 48 43 124 119 Chlorine, 1 oz. per 280 Ib. of flour 34 31 48 41.5 124 119.5 5.5 7.5 8 5 9 5 11.5 4 0 12-5 3.5 5-0 12.5 3.5 5.0 12.5 3.5 5-0 12.6 5.5 7.0 8 5 9.5 11.5 3.0 12-5 3.0 4.0 12.0 3.0 4-5 12.0 3.0 4.0 12.0 As we had anticipated, differences appreciably greater than the 2.5” quoted by Kent- Jones and Martin were obtained in this extended series of tests of unbleached and bleached samples. The encouraging feature of the tests is, however, that the effect of bleach upon flours with ash contents below 1 per cent.-a range which embraces flours from high grade patent flours to 90 per cent. extraction flours-is within the experimental error of the method, i.e., 3~0.5 units. CORRELATION BETWEEN DIAL READING AND ASH CONTENT- We have explained in the introduction to this paper that the “grade” colour of flour, which Kent-Jones and Martin claim to be measurable by their instrument, is related to the extent to which the endosperm of the grain has become contaminated with powdered wheat skins and associated substances during the milling process.Provided a flour contains no extraneous mineral matter, the proportion of powdered wheat skins it contains is correlated with its ash content, because the ash content of pure endosperm is 0.3 per cent. or less, whereas the ash content of pure wheat skin is in the region of 8 to 10 per cent. It follows, therefore, that if this instrument is to provide a reliable measure of the “grade” colour of flours, there must be a correlation between dial reading and natural ash content.Our next step in this investigation, therefore, was to determine the dial readings and the ash contents of various mill stocks from several mills. As the instrument was designed for use with commercial flours the stocks used in this experiment were restricted to those with ash contents not substantially greater than 1 per cent. The data obtained in this series of tests are given in Table 111.March, 19501 PHOTO-ELECTRIC RECORDING OF FLOUR GRADE TABLE I11 DIAL READINGS AND ASH CONTENTS OF A SERIES GF MILL STOCKS Sample A flour . . B 99 .. c 3) .. D 79 .. B, 79 . . A flour . . B 99 .. c '9 .. D 31 .. A flour .. B 99 .. c '7 .. D 39 .. B, 39 . . A flour . . B 9' .. c 79 .. D 99 .. €3, 9) . . .. .. .. .. .. .. .. .. .... .. .. .. .. .. .. .. .. .. .. .. * . .. .. .. Ash, % Mill 1 . . 0.52 . . 0.54 . . 0.46 . . 0.80 . . 1.11 Mill 2 . . 0.43 . . 0.42 . . 0.43 . . 0.87 Mill 3 . . 0.53 . . 0.42 . . 0-47 . . 1.08 . . 1.06 Mill 4 . . 0.44 . . 0.44 . . 0.51 . . 0-68 . . 1.08 Mill 5 . . 0.37 . . 0.43 . . 0.43 . . 0-66 . . 1.10 . . 0-79 Dial reading 61 59.5 49 99 121 40.5 36 47 108 54 38.5 50.5 122 110 55 56 62.5 98 130 39 39.5 39.5 96 149.5 99 Sample A flour . . .. B 79 .. * . c 9' .. .. D 73 .. .. 1st Bk. flour . . A flour .. .. B '9 .. .. c 99 .. . . A flour . . . . c 11 .. * . D 99 .. . . B, 9' . . . . B.M.R. .. .. A flour .. .. B 99 .. .. c 99 .. .. D n .. .. A, B, C, D flours A flour . . .. B 99 .. .. c 9) .. .. D 3) .. . . 1st Bk. flour . . 2nd Bk. flour . . B, flour . . .. Ash, % Mill 6 .. 047 .. 0.44 . . 0.43 . . 0.74 . . 0-60 Mill 7 . . 0.44 . . 0.42 . . 0-47 Mill 8 . . 0.48 . . 0.44 . . 0-66 . . 0.50 . . 0-77 Mill 9 . . 0.46 . . 0.62 . . 0.47 . . 0-58 . . 0.61 Mill 10 . . 0-45 . . 0.43 . . 0.49 . . 0.61 . . 0.59 . . 0.59 . . 0.59 137 Dial reading 48 43.5 40-5 93.5 82 36.5 37.5 51 56 39.5 80.5 57 96 47.5 51.5 48 68 54 39 39 47 66 80 75.6 70.5 St at istical data reveals that for these ex~eriments the correlation analysis of these coefficient between dial reading and ash content is 0.959. This very' high degree of correlation leaves no doubt that the colour figure of a normal uncontaminated flour furnished by this instrument is a reliable index of the grade of that flour. When the dial readings of Table I11 were plotted against the corresponding ash contents, the scatter diagram thus produced showed that the relationship between the two parameters was not linear.Re-examination of the data established that a linear relationship did exist, however, between dial reading and the logarithm of the ash content. The regression of log ash on dial reading was therefore calculated and was found to be- It must be emphasised that this equation connecting dial reading with log ash applies only to the instrument we used and at its present setting. Each setting and each instrument, therefore, will have its own regression equation. In Fig. 1 the regression line corresponding to this equation has been drawn and on the same diagram the dial readings of Table I11 have been plotted against the logarithms of the corresponding ash contents.Although the degree of correlation between dial reading and ash content is so high, as can be seen from the scatter diagram of Fig. 1, it is possible that two flours of significantly different ash contents may furnish the same dial reading and, conversely, two flours of identical ash contents may yield significantly different dial readings. Examples from Table I11 are A flour from Mill 5 and A flour from Mill 10, flours which give the same dial reading of 39 but have ash contents of 0.39 and 0.45 per cent. respectively; D flour from Mill 5 and B.M.R. flour from Mill 8, each of which gives a dial reading of 96 but which have respectively ash contents of 0.66 and 0.77 per cent.; B flour from Mill 4 and A flour from Mill 7, each of which has an ash content of 0.44 per cent.and yet which give Log ash = 0.0044 x dial reading - 0.5454.138 KENT-JONES, AMOS AND MARTIN: EXPERIMENTS IN THE [Vol. 75 dial readings of 56 and 36.5 respectively; and D flour from Mill 5 and D flour from Mill 8 which are identical in ash content of 0.66 per cent., but which give respectively dial readings of 96 and 80.5. Occasional apparent discrepancies of this nature are not surprising and are, in fact, to be expected. The ash contents of different types of wheat show an appreciable spread and the ash content of the pure endosperm of one wheat may be different from that of another. Furthermore, the various skins on a wheat kernel vary among themselves in both ash content and colour, and hence the relationship between the dial reading and the ash content of an individual mill stock can be influenced by the origin of the stock in question.0.100 0,050 0 - 1.950 T.900 T-850 - 1.800 - 1.750 ‘ - 1.700 ’i. 650 . ’T.550 T500 30 40 5 0 6 0 70 80 90 100 110 120 130 140 150 Fig. 1 Although these facts have not been unrealised in the past, there has been a tendency to ignore them and to accept ash content as a definite measure of flour colour. Our experience, however, has proved that ash content is not always a fully reliable index of the brightness of the flour as deterrnined by its ability to reflect light, which is what is meant by “grade” colour. If an instrument is to be a reliable means of measuring the “grade” colour of flour, the readings it gives will bear a high degree of correlation with the corresponding ash contents, but since the ash content is not always a sure index of the colour, occasional deviations from the relationship should occur.This instrument conforms to these expectations and it is our opinion that providing, as it does, a direct measure of reflecting power, it is a more reliable means of evaluating “grade” colour The position is, in our opinion, as follows.March, 19501 PHOTO-ELECTRIC RECORDING OF FLOUR GRADE 139 than is the indirect estimate from the ash content, and hence, where a discrepancy between the two methods occurs, it is the instrument that provides the truer index of the colour. As has been mentioned, the miller in pre-war days paid considerable attention to the ash content of his flour because this figure was accepted as an index of the grade colour.To-day, however, the miller is without this guidance because by Government decree all bread flour contains a proportion of added calcium carbonate (Creta praeparata) and the reported ash content of a flour is not, therefore, a measure of the natural ash content of the flwr. Furthermore, it is not possible to apply a standard correction for the effect of the addition of the statutory proportion of calcium carbonate because it is impossible to effect completely uniform distribution of this addendum. This new method for the measurement of grade colour should, therefore, prove exceptionally valuable at the present time in that it will enable the miller to obtain a numerical assessment of the success of his milling operations despite the influence of added mineral matter upon the ash determination.In Table IV are given the grade colour figures, Le., dial readings divided by 10, of a series of commercial samples of 85 per cent. National flour received recently at our laboratories for analysis and it will be seen that between the worst and 'the best of these, for samples covering the extreme range likely to be encountered under normal conditions, there is a spread of 4 units of grade colour. Included also in this table are the ranges of colour figures to be expected from pre-war white flours and from 90 per cent. extraction flours of the type made for a short period during the war. TABLE IV TYPICAL GRADE FIGURES FOR COMMERCIAL FLOURS Sample Pre-war patent flours (ash contents 0.32 to 0.40 per cent.) Pre-war straight-run flours (72 per cent.extraction) (ash contents 0-44 to 0-50 per cent.) . . . . .. .. National flour (85 per cent. extraction) . . .. .. n * n >¶ .. .. .. n n n n .. .. .. 91 * n n .. n I> n n .. .. .. .. .. I3 n n n n n n n .. .. .. .. .. * . * 99 9s n n n n n .. .. .. .. .. .. n n n n .. .. * . n n n n n n 19 n n n n n n 99 n n .. .. .. .. .. .. .. .. .. .. .. .. 90 per cent. extraction flour . . .. .. .. Grade colour figure 1-4.0 4.5-6.5 9.5 10.5 11.0 9.0 10.5 13.0 10.0 9.5 11.0 13.5 10.0 10.0 12.0 9.5 13*0-16*0 DETERMINATION OF GRADE COLOUR OF SELF-RAISING FLOURS- The evaluation of the grade colour of self-raising flour has always been a problem. The inclusion of over 7 lb. of mineral aerating ingredients in 280 lb.of flour and the possibility of minor variations in the distribution of the ingredients renders the ash content of a self- raising flour valueless as an index of colour. Separation of the added mineral ingredients by high-speed centrifuging prior to the determination of ash content has been suggested,ls but the method has not been widely adopted. The present colour method, therefore, seemed to offer a simple solution to the problem, provided any interference caused by the added chemicals could be overcome. A preIiminary experiment (Table V) revealed that the addition of acid calcium phosphate and sodium bicarbonate to a flour in the proportions used for self-raising purposes significantly diminished the dial reading given by a flour by the normal technique.The addition of normal proportions of acid sodium pyrophosphate and sodium bicarbonate did not, however, seriously alter the dial reading. In view of the interference caused by the presence of acid calcium phosphate, which is the commonly used acid ingredient of self-raising flours, it was decided t o investigate the effect of using a buffer solution in place of water in the preparation of the flour paste. Experi- ments showed that if the paste were made with a citrate buffer solution, the reading of a plain140 KENT- JONES, DIAL READINGS Natural ash content, yo . . AMOS AND MARTIN: EXPERIMENTS I N THE [Vol. 76 TABLE V OF FLOURS CONTAINING MINERAL ADDITIONS PASTE MADE WITH WATER Flour A B C D A f \ .. .. .. . . 0.41 0.56 0.78 1.20 Dial reading with water (normal procedure) .. .. 39 68 105 169 Dial reading witth water after addition of 341b. NaCO, and 4+ lb. A.C.P. per 280 lb. (normal self-raising flour) 25 48 87 151 Dial reading with water after addition of 34lb. NaCO, and 64 lb. proprietary pyrophosphate mixture per 280 lb. (normal self-raising flour) . . .. .. 39 62 105 168 flour was not altered and, moreover, the conversion of the flour to a self-raising flour by the addition thereto of sodium bicarbonate and acid calcium phosphate or acid sodium pyro- phosphate did not markedly affect the dial reading. Furthermore, when the buffer solution was used to make the paste, a flour containing a marked excess of acid calcium phosphate, acid sodium pyrophosphate or sodium bicarbonate, gave a dial reading not vastly different from that of the untreated flour.A series of results from these experiments is given in Table VI. TABLE VI PASTE MADE WITH BUFFER SOLUTION DIAL READINGS OF FLOURS CONTAINING MINERAL ADDITIONS Flour Natural ash content, % . . .. .. . . .. Dial reading with water (normal procedure) . . . . Dial reading with buffer . . .. .. * . .. Dial reading with buffer after addition of 34 lb. NaHCO, and 44 lb. A.C.P. per 280 lb. (normal self-raising flour) Dial reading with buffer after addition of 3) lb. NaHCO, and 64 lb. proprietary pyrophosphate mixture per Dial reading with buffer after addition of 4+1b. A.C.P. Dial reading with buffer after addition of 64 lb. pro- prietary pyrophosphate mixture per 280 lb. (excess of Did reading with buffer after addition of 34 lb.NaHCO, 280 lb. (normal self-raising flour) . . .. .. per 280 lb. (excess of acid) . . .. .. . . acid) .. .. . . .. . . .. per 280 lb. (excess of alkali) . . .. .. .. r A 0.4 I 39 41 40 39 41 39 35 B C D‘ 0.56 0.78 1.20 68 105 169 65 103 172 67 106 164 62 105 168 70 I 173 174 - - A stock buffer solution is prepared by dissolving 110 g. of disodium hydrogen phosphate and 77 g. of citric acid in water and making up to 1 litre. When required for testing self-raising flour this solution is diluted fivefold. These tests show, therefore, that the method can be applied to self-raising flours equally as well as to plain flours provided the flour is made into a paste not with water but with a citrate buffer solution of specified strength. By this technique it is possible to obtain a reliable numerical assessment of the colour of a self-raising flour.This instrument, therefore, provides a long-felt want in that it enables the grade colour of flour to be expressed on a numerical basis, even when that flour contains mineral additions which invalidate the use of the ash test as a measure of colour. SUMMARY The instrument of Kent-Jones and Martin enables the grade coldur to be expressed numerically and the results it gives have good reproducibility in the hands of different operators (irrespective of the day upon which the test is made). The “grade colour figure” obtained by dividing the dial reading by 10 can be determined within f0-5 unit irrespective of the degree of bleach on the flour. The dial readings obtained upon this instrument are linearly related to the logarithms of the corresponding ash contents and there is a very high degree of correlation between dial reading and ash content.March, 19501 PHOTO-ELECTRIC RECORDING OF FLOUR GRADE 141 The grade colour of self-raising flours can be determined by this instrument by employing a buffer solution in the preparation of the paste.in the 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. It is with pleasure that the authors place on record their appreciation of the ready willingness of Dr. E. C. Wood to guide them in the theoretical and practical issues involved statistical aspects of this paper. REFERENCES Jago, W., and Jago, W. C., The Technology of Breadmaking, Simpkin, Marshall, Hamilton, Kent Baker, J . C., Parker, H. K., and Preese, F.B., Cereal Chem., 1933, 10, 437. Kent-Jones, D. W., and Herd, C. W., Analyst, 1927, 52, 443. Markley, M. C., and Bailey, C., Cereal Chem., 1935, 12, 33. Binnington, D. S., Hutchinson, W. S., and Ferrari, C. G., Ibid., 1941, 18, 10. Binnington, D. S., Sibbitt, L. D., and Geddes, W. F., Ibid., 1938, 15, 119. Ferrari, C. G., and Bailey, C. H., Ibid., 1929, 6, 218. t , Ibid., 1929, 6, 347. -- , Ibid., 1929, 6, 457. Sirnison, A. G., Ibid., 1935, 12, 569. Visser’t Hooft, F., and De Leeuw, F. J. G., Ibid., 1928, 5, 351. Kent-Jones, D. W., and Herd, C. W., Ibid., 1929, 6, 33. Kent-Jones, D. W., and Amos, A. J., Modern Cereal Chemistry, 4th Edition, Northern Publ. Co., Ltd., Liverpool, 1947, p. 473. Kent-Jones, D. W., and Martin, W., Analyst, 1950, 75, 127. Gustafson, C.B., Cereal Chem., 1931, 8, 475. & Co., Ltd., London, 1911, p. 709. -- 88, MADELEY ROAD LONDON, W.5 DISCUSSION MR. A. L. BACHARACH called attention to the difference between “dial readings” for bleached and unbleached flour. Although the maximum difference between pairs of readings was 10, and although this represented only twice the expected difference from the mean (5 units), the effect of bleaching might well be significant because all the differences produced by i t were in the same direction. DR. J. H. HAMENCE asked whether there was any means of checking the instrument against a standard. MR. D. M. FREELAND pointed out that the correlation of colour units with ash content demonstrated by the authors could apply only these days to “pure” mill streams.Most users of flour found ash figures were increased because of the presence of 14 oz. of prepared chalk per 280 lb. of flour and the ash - colour relation was thereby disjointed. Would not some approximate ranges of colour units to the usual observa- tions of the Pekar test be more useful to the flour buyer? MR. E. SEAL said that, as the grade colour was really that of the flour-water paste and not the flour itself, it was not exactly a measure of flour colour. A soft flour would appear much brighter in the dry state than a strong granular flour of equal grade. Had Dr. Amos observed these actual differences between weak and strong flours of equal grade-figure as measured on this instrument? He also asked whether the use of the buffer with plain flours would extend the period of time during which a measurement could be accurately made on the paste.DR. C. R. JONES endorsed the claim that there was need for a method of measuring brightness of flours, and considered the results given most encouraging, subject to the important question already raised as to reproducibility of results with different instruments. He suggested the term “flour-water suspensions” be used in place of “flour pastes” to avoid confusion with gelatinised preparations. DR. A. GREEN said that he considered that the reference by the authors to the presence of carotene in wheat flour was not justified. Both Zechmeister and von Euler had failed to find it (see also Bacharach, Analyst, 1941, 66, 36). The degree of bleaching of the test flours was not in general that which is used in commercial practice.Would the application to the more branny stocks of heavier treatments than those described by the authors still yield practical identity in light reflection for the same mill product when unbleached and when bleached ? In view of the fact that nitrogen trichloride and chlorine dioxide have different effects on the pigments of bran, had the authors tested the performance of their instrument on mill products containing finely divided bran, before and after these products had been treated with chlorine dioxide ? It is possible that light of a different wavelength might be needed in this case. DR. K. A. WILLIAMS mentioned that he used a very similar method in determining the colour of oils and fats. In that case he had preferred to determine the ratio of the transmissions a t two wavelengths rather than a figure a t only one wavelength.Attention had recently been paid to this method in America. DR. J. STRAUB called attention to the fact that greyness of flour is measured by the apparatus much more accurately than it could be assessed visually. He felt that such precision would prove its usefulness in research in preventing greyness of flour. When trying out new techniques of milling for the removal of142 KENT- JONES, AMOS AND MARTIN [Vol. 75 greyness, the slightest invisible change in greyness would show whether progress was being made. The laboratory was t o be congratulated on its results. DR. N. L. KENT asked whether a high correlation between dial reading and ash content was obtained if a series of flours taken right down the reduction of a mill, from A to K, L or M, Le., including both high grade and low grade flours, was used.He felt that the method which had been described for measuring the colour of flour in a positive way would be welcomed by millers, and more particularly by the miller abroad. For the first time i t would be possible for him to compare the colour he obtains in a simple and reliable way with that achieved by millers in this country. Therefore the speaker asked for an assurance that no local factors would upset the figures he might obtain, thus preventing a true comparison with the figures obtained in this country. He asked whether the character of the water would affect the result, and whether the temperature of the water would alter the readings.DR. AMOS said in reply that the fact that the slight effect of bleach upon the dial readings had always been in one direction might have some significance, but the purpose of the quoted figures was t o show that the magnitude of this effect upon the readings was within the experimental error of the test. One of these, which was mentioned in the paper, was to insert an 80 per cent. transmission screen in one light path and to adjust the instrument so that a given reading was obtained upon the dial. The other method was t o adjust the instrument so that the readings of two or three flours of known ash contents were those required by the regression equation for the instrument in question. The ideal method would probably be calibration against a standard white surface, but this would necessitate a standard which was absolutely permanent, and it had not been possible to meet this essential criterion.Mr. Freeland had perhaps misunderstood the position) The authors had demonstrated that their dial readings were closely correlated with the true ash contents of flours as proof that their readings were a measure of grade colour. The fact that this correlation would not be apparent in National flours because of the presence of chalk was beside the point; the instrument was designed to measure grade colour, and this it did whether chalk was present or absent. Although the reading was made on a flour paste, it was nevertheless a true measure of the grade colour of the flour, as had been shown by the high degree of correlation between the readings and ash contents. Certainly a soft flour appeared brighter than a strong flour of equal grade because of the influence of granu- larity, and i t was for this very reason that the authors had eliminated the influence of granularity by the use of a paste.The use of the buffer with plain flours had not been tried, but there seemed no point in investigating the possibility of extending, by this means, the time that a paste could be kept before the reading was taken, since with a water paste no change in the reading occurred in 30 minutes. The rates of bleaching treatment that were employed were average values for the commercial bleach- treatment of straight-run flours, and they had not significantly affected the readings when applied t o flours ranging from patent grade to a grade equivalent to 90 per cent.extraction. It was true that in commercial practice certain very low-grade stocks would receive a much higher dosage, but this point had not arisen because in this paper the authors had restricted their investigations to stocks with ash contents not significantly greater than 1 per cent. The use of chlorine dioxide was not included in the bleaching experi- ments because it was not in commercial use in this country, but in view of the fact that it was likely to be so used at some time in the future, tests with this product were in hand. The correlation coefficient between dial reading and ash content quoted in the paper applied to a series of flours ranging from high patent grade to those with ash contents slightly higher than 1 per cent. Tests with low-grade stocks having ash contents of 3 per cent. and over had shown a very good degree of correlation. The nature of the water used to make the paste would not be likely to influence the results seriously; the authors had tried distilled water and tap water and had found no difference in their results. Tests made during the summer had not been affected by the temperature, but the high temperature experienced in some countries overseas might lead to a slight difference in the readings; this could be tested and due allowance made. MR. F. T. HOLDEN said that he was a miller from British East Africa. The instrument could be standardised by two methods. THE CALCULATION OF THE BOTANICAL COMPOSITION OF WHEAT FLOURS AND OFFALS FROM THE CHEMICAL ANALYSIS AT the meeting at which the above two papers were read, a lecture with this title was delivered by J. Straub, Chem. Ing., of the Central Institute for Nutrition Research, Utrecht. The following is a summary of Mr. Straub’s lecture, which has been published in full in Rec. Trav. Chim. Pays-Bas, 1950, 69, 141. A method was described for calculating the botanical composition, in terms of the percentage of pericarp, aleurone and endosperm in wheat flours and offals, from the amounts of crude fibre, phosphorus and starch in the milled product; these three constituentsMarch, 19501 EL-SOKKARY AND HASSAN : EGYPTIAN MILK 143 being characteristic respectively of the pericarp, the aleurone and the endosperm. Hence, from a knowledge of the amounts of fibre, phosphorus and starch that characterise the different botanical tissues it is possible to calculate the relative proportions of each that might be present in a sample, without having to separate them in a pure state. As a further consequence of the work it was possible to calculate the protein and fat content of the botanical fractions. It was also shown that the endosperm consists of two distinct parts, an inner and an outer, of widely differing starch and protein contents, and of which the proportions could be determined in milled products.

ISSN:0003-2654    

DOI:10.1039/AN9507500133

出版商:RSC    

年代:1950

数据来源: RSC

摘要:

March, 19501 EL-SOKKARY AND HASSAN : EGYPTIAN MILK 143 Lactose and Chloride Contents of Egyptian Cows’ and Buff aloes’ Milks BY A. M. EL-SOKKARY AND H. A. HASSAN SyNoPsIs-Previous studies of Egyptian milks have not included the direct chemical estimation of lactose and chloride contents. An investigation has therefore been carried out, as an extension of previous work, into the lactose and chloride contents of the milks of Egyptian cows and buffaloes, 100 of each animal being taken at random. Buffaloes’ milk was found to have a significantly higher lactose c6ntent and lower chloride content than cows’ milk. Mean values, ranges of variation and frequency distributions of values have been calculated for both milks. The chloride - lactose relation, calculated by the Mathieu and Fen6 formula, was found to be approximately the same for both milks, although the value was lower than had originally been supposed.OWING to the lack of adequate data on the subject, there is still considerable scope for an investigation of the milks of Egyptian cows and buffaloes. In a previous paper,l the authors have given some results for both milks, but lactose and chloride contents were not given, since they were not determined. Lactose is generally obtained by difference methods in routine milk analysis, and thus there are very few available data based on its actual chemical determination, as compared with the data concerning the other constituents of cows’ milk. This inadequacy is even more noticeable in the case of buffaloes’ milk, owing to the small amount of work that has been carried out on it.It was decided, therefore, to carry out an investigation on both milks to complete the previous study. The complementary relationship between lactose and chloride in milk has been recognised for over 50 years, and Mathieu and Ferrk’s formula2 is still looked upon as a valuable means of assessing the part played by lactose in the variation of non-fatty solids in milk. The chloride content of milk is taken as a criterion of the performance of the animals’ glands in milk secretion, since a low chloride content denotes a high level of lactose secretion, and thus indicates a good condition of the secreting tissue. Recent investigation^,^^^ however, although still indicating a close negative correlation between the lactose and chloride contents of milk, show wide variation in Mathieu and Ferrk’s value or the “simplified molecular constant.” The problem, therefore, needed further investigation, particularly with respect to the Egyptian milk animals, in which field this paper is the first to deal with the subject.EXPERIMENTAL This study was carried out on individual samples taken from the well-mixed milk from the morning milking of 100 cows and the same number of buffaloes kept in Cairo and on neighbouring farms. Samples were taken at random, but animals with clinical cases of disease were avoided. Lactose and chloride contents were determined, and their relationship calculated according to the Mathieu and Ferrk formula2: Lactose, per cent. + 19.6 x chloride, per cent. = constant (7).METHODS OF ANALYSIS-Lactose was determined by the gravimetrk method of A.O.A.C. for milk: and chloride by the wet digestion method developed by Davies.9144 EL-SOKKARY AND HASSAN: LACTOSE AND CHLORIDE CONTENTS [Vol. 75 RESULTS Table I shows the means, S.D., S.E.M. and significance of differences in lactose content, The significance of differences chloride content and chloride - lactose number, for both milks. is calculated by the t-test.’ TABLE I: THE SIGNIFICANCE OF DIFFERENCES BETWEEN cows’ AND BUFFALOES’ MILKS Lactose, % Chloride, yo Chloride - lactose number 7- 7- EzZ-za Cows’ Buff aloes’ Cows’ Buff aloes’ Mean .. . . 4-71 4-87 0.0765 0.0649 6.20 6.15 Range of variation 3-75-5-39 3.91-5.46 0.0403-0-1492 0-0338-0.1 190 5.73-6-89 6.83-6-58 S.D. ... . 0.33 0.27 0.0209 0.0149 0.20 0.18 S.E.M. . . . . 0.03 0.03 0.002 1 0.0015 0.02 0.02 Difference between means, C-B . . - 0.16 + 0.0116 + 0.05 t .. .. .. 3.7324 4.5366 1.8746 Probability, P . . < 0.01 < 0.01 0.1-0*05 - Significance* .. + + The frequencies with which the threevalues occur were calculated, and are given in * Plus sign (+) = significant; minus sign (-) = insignificant. Tables 11. I11 and IV. TABLE I1 FREQUENCY DISTRIBUTION OF PERCENTAGE OF LACTOSE 3-40 3.80 4.20 4.60 5.00 5-40 Intervals . . to to to to to to 3-80 4.20 4.60 5.00 5.40 5.80 Cows’milk .. .. 2 5 32 39 22 - Buffaloes’ milk . . - 1 13 52 33 1 TABLE I11 FREQUENCY DISTRIBUTION OF PERCENTAGE OF CHLORIDE 0.0300 0.0500 0.0700 0.0900 0.1100 0.1300 0.0500 0.0700 0.0900 0.1100 0.1300 0.1500 Intervals .. to to to to to to Cows’ milk . . .. 7 35 38 14 3 3 Buffaloes’ milk . . 12 51 33 3 * 1 - TABLE IV FREQUENCY DISTRIBUTION OF CHLORIDE - LACTOSE NUMBER 5.6 5.8 6.0 6.2 6.4 6.6 6.8 Intervals . . to to to to to to to 5.8 6.0 6.2 6.4 6-6 6.8 7.0 Cows’ milk .. 2 8 41 36 9 1 3 Buffaloes’ milk . . - 16 54 20 8 2 - DISCUSSION OF RESULTS LACTOSE CONTENT- Total 100 100 Total 100 100 Total 100 100 It can be seen from Table I that the lactose content of cows’ and buffaloes’ milks ranged from 3-75 to 6.39 per cent. and 3.91 to 5-46 per cent. respectively; the mean values being 4.71 per cent., S.E.M. 0.03, and 4.87 per cent., S.E.M. 0-03. It is thus observed that buffaloes’ milk has a higher lactose content than that of cows; the difference averaging 0.16 per cent.This difference, though little, is consistent, and, as calculated by the t-test, has a P of less than 0.01, which indicates its high significance. This was also proved by Anantakrishnan ei! al.? who found that Indian buffaloes’ milk contained more lactose than did cows’ milk. Table I1 shows that the highest frequency distribution of lactose lies in the class 4-60 to 5.00 per cent. in both milks, but for that of cows this highest frequency represents 39 per cent. of the samples, while it represents 52 per cent. of the samples of buffaloes’ milk. It is also noticeable that for cows’ milk, 39 per cent. of the samples are below this class, whileMarch, 19501 OF EGYPTIAN cows’ AND BUFFALOES’ MILKS 145 for buffaloes’ milk, the corresponding percentage is only 14, and 34 per cent.of the samples are above the 4.60 to 5.00 per cent. class. This distribution clearly shows that the value tends to be higher for buffaloes’ than for cows’ milk. Results given by Ghosb and Datta-Roy8 show that lactose in buffaloes’ milk ranged from 4.0 to 5.3 per cent., and that of cows’ milk ranged from 3.8 to 5.3 per cent. These results are in close agreement with those of the present investigation, and with most of the available data for European and other breeds of cows. CHLORIDE CONTENT- I t is well known that the chloride contents of individual samples of cows’ milk show wide variation. The mean values are 0.0765 per cent., S.E.M. 0.0021, and 0.0649 per cent., S.E.M. 0.0015, for cows’ and buffaloes’ milks respectively. In the former, the value varied from 0.0403 to 0.1492 per cent., while in the latter it varied from 0.0338 to 0.1190 per cent., thus indicating that, on the average, cows’ milk is 0.0116 per cent.higher in its chloride content than that of buffaloes. The difference is statistically significant, as shown in Table I. This observation, together with the higher content of lactose in buffaloes’ milk and its higher value of casein nitrogen as a percentage of the total nitrogen - casein number (El- Sokkary and Hassanl), indicates that buffaloes have a more efficient secreting tissue than cows have. It can also be seen, by comparing the present results with those given by D a ~ i e s , ~ that Egyptian cows’ milk tends to contain less chloride than does the milk of other breeds of Table I11 shows that the highest percentage distributions of the chloride value for both cows’ and buffaloes’ milks lie in the classes 0.07 to 0.09 per cent.and 0.05 to 0-07 per cent. respectively, representing 35 and 51 per cent. of the samples. For cows’ milk, 58 per cent. of the samples had chloride values above the highest frequency distribution, while for buffaloes’ milk, only 37 per cent. were above i t ; hence the significance of the difference between the two milks is explained. This is demonstrated by the magnitude of the variations in Table I. cows. CHLORIDE AND LACTOSE RELATIONSHIP- Owing to the fact that chloride and lactose concentrations in milk account for approxi- mately 75 per cent. of its osmotic pressure, many attempts have been made to establish the correlation between them on a complementary basis.Among these attempts are those of Kopatschekl* and Koestler,ll but their ratios, however, are not so commonly used as that given by the formula suggested by Mathieu and known as the “simplified molecular constant.” It is stated that if the value is below 7, the milk sample may be suspected of containing added water. Since these workers published their results many investigations have been carried out on the subject, and these showed the value to vary over a remarkably wide range. Richmond et aLl2 state that the value is said to lie, for most milks, between 7.4 and 7.9, but they quoted a report of a range of 6.92 to 8.28. Recently, some investigators have given results which show wider variations. Mathieu,13 working on the milk of cows in Alpine regions, found that the ratio, in most samples, was higher than 8.0.Anantakrishnan et aL3 found that, with Indian cows’ and buffaloes’ milks, the lactose - chloride relationship did not hold. Moreover, working out the ratio from results given by Roadhouse and Henderson* on milk from American breeds of cows, the authors found that, for one set of results, the ratio varied from 6.19 to 7.34, corresponding to lactose and chloride contents of 4.74 and 0.074 per cent. respectively for the lower ratio, and 4.16 and 0.162 per cent. respectively for the higher ratio. In another set, a ratio as low as 5.99 is given, corresponding to 5.05 per cent. lactose and 0.048 per cent. chloride. These latter results agree to a large extent with those of the present study, where the mean value for the ratio in cows’ milk is 6.20, S.E.M.0-02, ranging from 5.73 to 6-89, and that of buffaloes’ milk is 6.15, S.E.M. 0.02, ranging from 5.83 to 6.58. Difference between both means is shown to be insignificant (see Table I). Table IV gives the frequencies with which values are distributed among the various classes. The highest percentage distributions for both milks lie in the same class, namely 6.0 to 6.2, and represent 41 per cent. of the cows’ milk samples and 54 per cent. of those of buff aloes’ milk.146 HALES THE INFRA-RED SPECTROSCOPIC ESTIMATION OF [Vol. 75 SUMMARY- Buffaloes’ milk was found to have significantly higher lactose and lower chloride contents than cows’ milk. The following are the respective means for cows’ and buffaloes’ milks: Lactose: 4-71 per cent., S.E.M.0.03, and 4.87 per cent., S.E.M. 0.03. Chloride: 0-0765 per cent., S.E.M. 0.0021, and 0.0649 per cent., S.E.M. 0.0015. Using the formula suggested by Mathieu and Ferr6 the lactose - chloride relation was calculated, and was found to be approximately the same for both milks, although the value was lower than had originally been supposed. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. El-Sokkary, A. M., and Hassan, H. A., J . Dairy Res., 1949, 16, 217. Mathieu, L., and F e d , L., J . SOC. Chem. Ind., 1914, 33, 214. Anantakrishnan, C. P., Dastur, N. N., and Kothavala, 2. R., Indian J . Vet. Sci., 1944, 13, 297. Roadhouse, C. L., and Henderson, J. L., Calif. Agric. Expt. Sta. Bull., 1935, 395. Oficial and Tentative Methods of Analysis of the A.O.A.C., 6th Edition, 1945. Davies, W. L., Analyst, 1932, 57, 99. Fisher, R. A., Statistical Methods for Research Workers, Oliver & Boyd, Edinburgh, 1936. Ghosb, N. K., and Datta-Roy, B. K., Indian Med. Gazette, 1941, 76, 279. Davies, W. L., J . Dairy Res., 1938, 9, 327. Kopatschek, F., quoted by Davies, W. L., Chemistry of MiEk, Chapman & Hall, 1936, p. 49. Koestler, G. A,, Mitt. A u s dem Geb. Lebensmie. Untersuch. u. Hyg., 1920, 11, 154. Richmond, H. D. et al., Dairy Chemistry, Chaxles Griffin & Co., Ltd., London, 1942, p. 130. Mathieu, M. G., Lait, 1942, 22, 219-220, 317-323. DAIRY DEPARTMENT, FACULTY OF AGRICULTURE FOUAD I UNIVERSITY, CAIRO July, 1949

ISSN:0003-2654    

DOI:10.1039/AN9507500143

出版商:RSC    

年代:1950

数据来源: RSC

摘要:

146 HALES THE INFRA-RED SPECTROSCOPIC ESTIMATION OF [Vol. 75 The Infra-Red Spectroscopic Estimation of 4-Methyl-2 : 6-Ditertiary Butyl Phenol in Mixtures Containing 2-Methyl and 3-Methyl-4 : 6-Ditertiary Butyl Phenols BY J. L. HALES SYNOPSIS-Industrial requirements set the problem of finding a rapid, accurate method for the estimation of 4-methyl-:! : 6-ditertiary butyl phenol in crude and refined products. It was also desirable to obtain a check on the type and amount of impurity in the products. A description is given of an infra-red spectroscopic technique whereby this may be obtained. The limitations inherent in the method because of the presence of certain impurities are indicated. Experimental values are given for estimations on several synthetic mixtures and on a crude product, and the magnitude of the errors introduced by impurities is indicated. INTRoDucTIoN-Weinrichl has shown that although m- and 9-cresol form mixtures which are difficult to separate, the two corresponding ditertiary butyl derivatives have a marked difference in volatility.Should, however, a cresol fraction containing o-cresol in addition to the m- and 9- derivatives be ditertiary butylated, the product will contain, in addition to the 4-methyl-2 : 6-ditertiary butyl phenol (b.p. 191" C./lOO mm.) and 3-methyl-4 : 6- ditertiary butyl phenol (b.p. 211" C./100 mm.), a proportion of 2-methyl-4 : 6-ditertiary butyl phenol (b.p. 194"C./100mm.) which will render more difficult the separation by distillation of the first-mentioned component in a reasonably pure state. The 4-methyl-2 : 6-ditertiary butyl phenol has technological applications as a softener and anti-oxidant, and the present purpose was to work out a rapid infra-red spectroscopic method of analysing mixtures in which it is the preponderant component, but which contain the other two isomerides, and in addition small quantities of monobutyl derivatives of the three cresols.March, 19501 4-METHYL-2 : 6-DITERTIARY BUTYL PHENOL 147 A method is described which gives an accurate value for the quantity of the desired component, and also a check on the amounts of the other two components present.A correction may be applied to allow for errors introduced due to the presence of monobutyl cresols in the crude mixture. EXPERIMENTAL Samples of the following were available- (a) The three individual dibutyl cresols.(b) A crude mixture of all three dibutyl cresols. (c) A fraction representative of the monobutyl cresol content of (b). The spectra of these samples and of their solutions in cyclohexane were obtained on a Hilger D209 double-beam infra-red spectrometer,2 using a rocksalt prism (see Fig. 1).* The region from 700 to 800 cm.-l was found to be the most useful for analytical purposes, since it includes key bands suitable for the estimation of each of the three components. Furthermore, rocksalt has a high dispersion in this region. Since all the aromatic ring frequencies lie in this range, the analysis is liable to interference from aromatic impurities, such as unchanged or monobutylated cresols. cycZoHexane, which is fairly transparent in this region, was chosen as the solvent for this analysis.Standard solutions of various concentrations of the individual components were made up and their spectra were measured. Sample cell thickness was 1-00 mm. for all analytical measurements. Wide slits (1.20 mm., corresponding to spectral slits of 6 cm.-l) were used with the spectro- meter, so as to give a high signal-noise ratio for the thermopile, and to avoid resolving the key band at 775 cm.-l (Fig. 1) into two bands at 771 and 776 cm.-l, which would complicate the analysis. Optical density at the three key frequencies for each of the three components was plotted against concentration (Fig. 2). Scattered radiation was allowed for, using a mica shutter and a l-mm. thick cell filled with cyclohexane, so as to approximate closely to the experimental conditions.A number of synthetic mixtures were prepared, their optical density a t the key frequencies measured, and from the curves illustrated in Fig. 2 the analytical figures were worked out by the method of successive approximation^.^ The results are given in Table I. TABLE I SYNTHETIC MIXTURES This is referred to later in the paper. Ditertiary butyl phenols Found .. Found (corrected) Taken . . .. Found .. Found (corrected) Taken . . .. Found .. Found (corrected) Taken . . .. Found .. Found (corrected) Taken . . .. .. .. . . .. . . . . . . . . . . .. .. .. .. .. .. .. .. . . .. .. ,. .. .. .. r 3 4-methyl-2 : 6-, 2-methyl-4 : 6-, 3-methyl-4 : 6-, g./100 ml. g./100 ml. g./100 ml. .. 1-81 2.48 2.42 ..1-67 1.78 1.81 . . 1.72 1.68 1.69 .. 1.75 . . 1-70 . . 1.72 . . 1.55 .. 1.51 . . 1.52 .. 2.10 . . 2.07 . . 2.05 1-30 0.75 0.7 1 0.99 0.5 1 0.44 0.83 0.26 0.23 1.08 0-73 0-64 0.87 0.60 0.55 0.53 0.27 0-22 Corrected figures take into account the overlapping of key bands. It should be noted (Fig. 2) that the key band at 775 cm.-l, for the 4-methyl-2 : 6- ditertiary butyl phenol has an extinction coefficient considerably higher than those of the bands at the key frequencies for the other two components. This enhances the accuracy and reliability of the estimation of this component, particularly as it is the major component * The spectrometer that was used incorporated a D.C. amplifying system, and slight drifts in the latter were the main factor limiting the consistency of the quantitative results.148 100 80 - 60- 40..20 - O -i HALES : THE INFRA-RED SPECTROSCOPIC ESTIMATION OF ? A I I I 1 I I Wol. 75 100 80- 60 - 40- $ 20- U B Wave numbers (cm.-l) Fig. 1. Absorption Spectra. The arrows indicate analytical bands. A : 2-Methyl-4 : 6-ditertiary butyl phenol. C: 4-Methyl-2 : 6-ditertiary butyl phenol. B: 3-Methyl-4 : 6-ditertiary butyl phenol. D. Monobutyl cresol fraction. Concentration (g./lOO ml.) y = 755 crn.-l -0 I Concentration (g./lOO rnl.) Concentration (g./100 mt.) Fig. 2. Analytical curves for ditertiary butyl phenols. A: 2-Dfethyl-4 : 6 ditertiary butyl phenol. B: 3-Methyl-4: 6-ditertiary butyl phenol. C: 4-Methyl-2 : 6-ditertiary butyl phenol.March, 19501 4-METHYL-2 : 6-DITERTIARY BUTYL PHENOL 149 in the present case.In Table I the uncorrected figures illustrate the error introduced when no allowance is made for the overlapping of the key bands, and it is seen that the effect is smallest in the case of the band at 775 cm.-l. These considerations are substantiated by the relative magnitude of the errors in the figures given in the table. MONOBUTYL CRESOL FRACTION-This fraction had several absorption bands near 775 cm.-I (Fig. 1) and the presence of monobutyl cresols in the crude product to be analysed will thus reduce the accuracy of estimation of the dibutyl cresols. A correction for errors due to the presence of monobutyl cresols may be applied using an “internal reference” method. The monobutyl cresol fraction had a fairly intense band at 1081 cm.-l which does not coincide with any bands in the dibutyl cresols, and this can be used to detect and estimate the effect of small quantities of monobutyl cresols in the crude mixture.However, a l-mm. layer of cyczohexane absorbs strongly at 1081 cm.-1, so that any comparison of optical densities must be made on the undiluted components. Although this may introduce deviations from Beer’s law, the correction under consideration need not be known to a very high accuracy. Hence the optical density of the band at 1081 cm.+ was compared with the densities at the three key frequencies for the monobutyl cresol fraction, and from the density of the 1081 cm.-l band in the crude dibutyl cresol mixture, the respective corrections for density at the three key frequencies could be estimated.Their magnitude is shown by the figures in Table 11, which gives the results of two analyses on the crude mixture. TABLE I1 ANALYSIS OF CRUDE DIBUTYL CRESOL MIXTURE Ditertiary butyl phenols r 4-methyl-2 : 6-, 2-methyl-4 : 6-, 3-methyl-4 : 6- g./100 ml. g./lOO ml. g./100 ml. A .I Found .. .. .. . . 1.28 0.40 0.18 Corrected for presence of monobutyl cresols . . .. .. .. 1-24 0.26 0.01 Found .. .. .. .. 1-26 Corrected for presence of monobutyl cresols . . .. .. . . 1.22 0-41 0.27 0.19 0.02 Total crude material was 2 g. per 100 ml. in each case. Thanks are due to the Esso European Laboratories for providing the specimens of pure and crude materials required for this investigation. The work has been carried out as part of the research programme of the Chemical Research Laboratory, and this paper is published with the approval of the Director of the Laboratory. REFERENCES 1. 2. 3. Weinrich, W., Ind. Eng. Chem., 1943, 35, 264. Hales, J. L., J . Sci. Inst., 1949, 26, 359. Fry, D. L., Nusbaum, R. E., and Randall, H. M., J . A$pZied Phys., 1946, 17, 150. D.S.I.R. CHEMICAL RESEARCH LABORATORY TEDDINGTON, MIDDLESEX First submitted, July, 1949 Amended, January. 1960

ISSN:0003-2654    

DOI:10.1039/AN9507500146

出版商:RSC    

年代:1950

数据来源: RSC

摘要:

150 STOCKDALE : THE ESTIMATION OF IRON BY DICHROMATE Pol. 75 The Estimation of Iron by Dichromate BY D. STOCKDALE SYNOPSIS-The relative merits of diphenylamine, diphenylbenzidine, barium diphenylamine sulphonate, ferrous phenanthroline and potassium ferricyanide as indicators in the estimation of iron by dichromate have been examined. The source of the iron was either ferrous ammonium sulphate or an ore, and the titrations were made in both the presence and in the absence of ortho- phosphoric acid. All these indicators proved efficient when the conditions were suitable, Barium diphenylamine sulphonate was found to be the best, because its colour change is marked and because the end-point that it gives is the least affected by the conditions of the titration. NUMEROUS indicators, both internal and external, and a wide range of conditions have been recommended for this estimation.Because of certain discrepancies in the literature and occasional failures in the laboratory, it seemed desirable to make a systematic examination of the position. The indicators used were as follows- Diphenylamine (DA) . . . . Diphenylbenzidine (DB) . . . . Barium diphenylamine sulphonate Ferrous phenanthroline (FeP) . . Potassium ferricyanide . . . . Dilute solutions, used externally. ELECTROCHEMISTRY Since the pioneering work of J. H. Hildebrandl the electrochemistry of the oxidation of iron by dichromate and of the indicators has been much explored, notably by I. M. Kolthoff2J94 and G. F. Smith6 and their collaborators. Only the principal points need be recapitulated here.When a bright platinum wire is placed in a solution of ferrous and ferric ions, the half-cell potential at 25" C. is given by 0.04 ml. of a 1 per cent. solution in concentrated 0-04 ml. of a 1 per cent. solution in concentrated sulphuric acid. sixlphuric acid. (BaDS) . . .. .. . . 1.0 ml. of a 0.2 per cent. solution in water. 0.05 ml. of an aqueous solution 0-025 M with respect to both ferrous sulphate and 1 : 10 phenanthroline monohydrate. [Fe" ] [Fe"'] E = Eo - 0.059 loglo--* E,,, the standard electrode potential at infinite dilution, is -0.771 volt.6 However, as the ferric ion is very liable to form complex ions, the formal electrode potential in a solution of significant concentration, such as would be obtained by dissolving ferrous and ferric sulphates in the ratio of their equivalent weights in N sulphuric acid, is substantially less than Eo.It is about -0.69 volt when the ferrous and ferric salts are each about 0.05 N in either N sulphuric or N hydrochloric acid. In general, the lower the pH of the solution, the lower this potential, and it can be depressed further by the addition of some compound, such as orthophosphoric acid, which forms complexes with the ferric ion even more readily than do the common strong acids. The electrode potential for the reduction of the dichromate ion, CrzO," + 14H' + 6e -+ 2Cr"' + 7H,O, is similarly given by 0.059 [Cr"']2 E = E o - - 6 loglo[Cr,O,"] [H]14 * The standard electrode potential of this reaction is -1.36 volts, but more usually the activity coefficients of the ions are far from unity, and the potential obtained by adding 50ml.ofMarch, 19501 STOCKDALE: THE ESTIMATION OF IRON BY DICHROMATE 151 0.1 N dichromate to 25 ml. of 0.1 N ferrous solution is about -1.10 volts when the final concentration of strong acid is formally normal. This potential is much influenced by the pH, as the equation suggests, and will be increased by increasing the hydrogen ion concentra- tion of the solution in which the reduction is taking place. These points are illustrated by curves which have been obtained during the present investigation (Fig. 1). These curves show the results obtained when 25 ml. of 0.1 N ferrous salt in a solution of acid of the nature and concentration indicated were titrated with 50 ml. of 0.1 N potassium dichromate containing such an excess of the acid that the formal con- centration of acid remained constant at that shown for each curve until the equivalence-point .was reached, The volume at the equivalence-point was 50ml., except with phosphoric acid present, when it was 60 ml., since 10 ml.of 50 per cent. by volume of orthophosphoric acid had been added initially. The horizontal lines near the curves mark the potentials of first colour change for the diphenylamine group of indicators and the formal electrode potential of ferrous phenanthroline in N sulphuric acid. 10 3Q Milfllitrer of 0.1 N K,Cr,O, Fig. 1. Millilitres of potassium dichromate plotted against 25 ml. of 0.1 N solution of Fe". Curve (a) 0.1 N hydrochloric acid; curve (b) N sulphuric acid ; curve (c) N sulphuric + orthophosphoric acids The rate of change of potential per unit volume of dichromate was found to be greatest a t -0.92 volt in N acid in the absence of phosphoric acid, and this was accepted as the equivalence-point potential.Fig. 1 shows that the first colour change of DA and DB should occur substantially before, and that of BaDS slightly before this point. It is to be expected that the maximum intensity of colour would be reached at potentials about 0.1 volt more than those of the first colour change. If this were so, the end-points of these indicators assessed on maximum intensity, not on first colour change, would coincide substantially with the equivalence-point. These potentials were found to be -0.86 volt for DA, -0.91 volt for BaDS and -0.94 volt for DB, the first two results being in accordance with expectation.The result for DB, which should be the same as that for DA because DB is the first oxidation product of diphenylamine, is high, presumably because DB is so insoluble (0.06 mg. per litre in water at 25" C.2). Most of this indicator is precipitated when a drop of it is added to the ferrous solution. When that remaining in the solution has been oxidised, more will slowly enter into solution and be oxidised in turn. The maximum colour intensity, therefore, can be obtained only slowly, and an estimation of the oxidation potential made during a titration carried out a t a practicable rate will necessarily be high.152 STOCKDALE: THE ESTIMATION OF IRON BY DICHROMATE [Vol. 75 Fig. 1 shows that FeP is an unsuitable indicator when in a normal solution of a strong acid.However, as Smith and Richter5 have shown, its oxidation potential is decreased by increasing the concentration of acid. The oxidation potential of the dichromate system is increased by the same means. It should therefore be possible to lower the line representing the potential of FeP relatively until it cuts the vertical part of the titration curve (Fig. l ) , and under this condition the indicator would prove efficient. In an experiment using 2 N sulphuric acid, FeP was fully red at -0.91 volt. The addition of 0.05 ml. of 0.1 N dichromate at 25 ml. raised the potential to -1.05 volts and caused an immediate and noticeable change in the intensity of the colour. The red tinge faded out completely in about 30 seconds.In 10 N sulphuric acid the indicator was fully red at -0.87 volt, and a further addition of 1 drop of dichromate raised the potential to -1.00 volt, causing the indicator to change colour rapidly and completely. It would seem that 2 N is approximately the minimum concentration of sulphuric acid in which it is possible to make this titration. A final con- centration oi '2.5 N was later successfully adopted as standard. Results at the lower concentrations of hydrochloric acid were similar. With 1.5 N acid a titration was just possible, provided it was made slowly. With 2.5 N acid, 1 drop changed the potential from -0.84 volt (red) to -11.06 volts (colourless), and the oxidation of the indicator took place in about 10 seconds.In 6 N solution, on the other hand, it was almost impossible to obtain a result, because the greater part of the indicator was precipitated on the electrode tubes and on the sides of the beaker, and because the yellow colour of the ferric chloride ion made the final colour change difficult to see. In these experiments, drops of the solutions were withdrawn from time to time and tested with potassium ferricyanide. It was found that in a normal solution of a strong acid the critical potential for this indicator was near -0.85 volt, a potential very slightly lower than that required Tor DA and definitely lower than those needed for the other indicators. It is also lower than the value of -0.92 volt taken as the equivalence-point potential. These results suggest that potassium ferricyanide ceased to give a blue colour when about i part per thousand of the bivalent ion remained unoxidised.This corresponds to a concentration of approximately 3 mg. per litre. This conclusion is confirmed by later work (see Tables I and 11). There is, therefore, an appreciable indicator error with potassium ferricyanide, and it can be used in exact estimations only when a solution of dichromate is used to connect a known with an unknown quantity of iron, the two ferrous solutions being of approximately the same concentration. EXPERIMENTS WITH FERROUS AMMONIUM SULPHATE Portions from 1.2 to 1.4 g. of the salt were weighed into conical flasks, dissolved in 100 ml. of approximately N sulphuric acid, and titrated with 0.1 N potassium dichromate (exact, assuming the salt to be pure).The final volume was 150ml. and the final concentration of sulphuric acid was approximately 0-65 N. In a parallel set of experiments, 10 ml. of 1 : 1 by volume orthophosphoric acid was added in each titration, the final volume again being 150ml., and the weight of sulphuric acid being the same as before. The above conditions apply for all the indicators except FeP, when 100 ml. of 4 N sulphuric acid was used to give, a final concentration of 2.5 iV with respect to this acid in a final volume of 150 ml. The results, each the arithmetical mean of four titrations, are given in Table I. TABLE I Without H,PO,, With H,PO,, rnl. ml. 25.50 f 0-01 Diphenylamine . . ,. .. .. 25-49 rt 0.015 Diphenylbenzidine .. .. .. 25.48 * 0.02 Barium diphenylamine sulphonate .. 25.49 f 0.01 25.50 rt 0.01 No result possible Ferrous phenanthroline . . .. .. 25-48 f 0.01 - Potassium ferricyanide . . .. .. 25.44 f 0.01 - Table I gives the volume in ml. of 0.1 N potassium dichromate found to be equivalent to 1 g. of ferrous ammonium sulphate. The ferrous ammonium sulphate used was referred by potassium permanganate to pure sodium oxalate. Its equivalent weight was found to be 392.9 (for pure salt, 392.1). One gram of the salt should be oxidised by 25.45 ml. of 0.1 N potassium dichromate. It seems reasonable to assume that the sample of potassium dichromate was reasonably pure and that the end-points as shown by the various indicatorsMarch, 19501 STOCKDALE: THE ESTIMATION OF IRON BY DICHROMATE 163 (Table I) are all close to the true equivalence-point.The standard deviations are given in the table; they mean only that if the groups of experiments were repeated under the same conditions the probability that the arithmetical means would fall within the range indicated is about 60 per cent. It was hoped that a deviation would serve as a measure of the value of the indicator, but, surprisingly, this was far from being the case. All the deviations are small, and the spread of results in any group can be entirely accounted for by the errors inherent in volumetric analysis with the type of apparatus used. The numerical results show that all the indicators worked perfectly; personal impressions of their behaviour were quite different. Some gave clear colours, with sharp changes; other gave muddy colours with the intensity of the colour changing noticeably with time after the addition of the oxidant, in such a way that considerable personal judgment seermd to be required in making an assessment.I t seemed probable that such judgment could not be free from considerable error. It would appear, however, that satisfactory results can be obtained in volumetric analysis, even when using what seems to be an indifferent indicator and with the observer in doubt about his gauging of the end-point. NOTES ON INDICATORS Diphenylamine is first oxidised irreversibly to DB, which is then oxidised reversibly to diphenylbenzidine violet. The behaviour of BaDS is similar. The end-points given in Table I are those of maximum intensity. I t is probably advisable to assess them without reference to a blank, partly because of the difficulty of obtaining equal concentrations of indicator when such small volumes of indicator solution are used, and partly because the diphenylbenzidine violet is itself further oxidised by a small excess of dichromate to com- pounds of less intense colour. This further oxidation under the experimental conditions was slow, usually taking not less than an hour for the removal of the violet colour, and it is unlikely that it interfered with the assessment of the point of maximum intensity, but the colour of a spent and oxidised solution is often not permanent enough to be reliable as a standard in a subsequent titration.The results given in Table I show that when an end-point is obtainable reasonably quickly with one of these three indicators it makes no difference to the numerical result which indicator is used, or whether orthophosphoric acid is present or not.DIPHENYLAMINE-This indicator behaved somewhat erratically, particularly in the earlier stages of its oxidation in the absence of phosphoric acid. In two of the titrations the violet colour appeared early and deepened over a range of about 0.6 ml. in a titration of some 30 ml. Usually a dirty muddiness appeared in the solution, often as much as 1 ml. before the end-point. Despite these variations the end-point was consistent. In the presence of phosphoric acid the solution remained clear, the colours were bright, and the range was approximately 0.04 ml. In some cases, however, the colour at the end-point was blue, not violet. Titrations without phosphoric ' acid are perhaps to be preferred, because of the warning given of the approach to the end-point.DIPHENYLBENZIDINE-The colours, in the absence of phosphoric acid, were much brighter than with diphenylamine. Blue, not violet, predominated, and the blue colour changed to violet when the solution was allowed to stand after the titration. The range was approximately 0.7 ml. The titration must be carried out slowly with this indicator, because the colour deepens only slowly after the addition of the dichromate. This is perhaps due to its marked insolubility. In all cases a precipitate was produced when the indicator was added initially. Conversely, it also reacted slowly when ferrous iron was added to a solution of potassium dichromate.When phosphoric acid was present there was a further slowing of the oxidation of the indicator to such an extent that to obtain consistent end-points was impracticable. BARIUM DIPHENYLAMINE SULPHONATE-without phosphoric acid the approach to the end-point was marked by the appearance of a slight muddiness in the solution. The full rose-violet colour was developed after its first appearance by about 0.05 ml. of dichromate, and the range from first appearance of muddiness to maximum colour was about 0.1 ml. With phosphoric acid the colours were clear and the full colour was developed by about 0.03 ml. of dichromate. The indicator proved pleasant to use by either method, the element of doubt in judging the end-point being almost entirely absent.FERROUS PHENANTHROLINE-The oxidation of this indicator, and indeed of the other oxidation - reduction indicators discus~ed,~ appears to be complex, with ferrous iron playing In two others, the range was only about 0.25 ml.154 STOCKDALE: THE ESTIMATION OF IRON BY DICHROMATE [Vol. 75 some part in the mechanism. It is possible to prepare a solution of FeP containing an appreciable excess of dichromate, the indicator being obviously in the reduced form, and to discharge the red colour by the addition of a small quantity of a solution containing ferrous iron. It follows that FeP is a more satisfactory indicator when the ferrous solution is run into the dichromate. This will usually entail the use of a standard iron solution in a titration, the stages being the oxidation of the iron for estimation with excess dichromate, followed by titration of this excess with standard iron. With DA and BaDS this process would seem to be unnecessary, because so little trouble arises in the direct titration, and, indeed, the direct titration with FeP is quite practicable provided a generous concentration of sulphuric acid is present.EXPERIMENTS WITH IRON ORE These experiments were made to find out whether the elements likely to be present in a mineral or the compounds introduced by its reduction with stannous chloride interfered in any.way with the indicators. The material used was Iron Ore A supplied by British Chemical Standards, and contained 58.20 per cent. of iron, together with very small quantities of sulphur, phosphorus, arsenic, copper and titanium, in addition to lime, magnesia, alumina and silica.About log. of the mixed and dried ore were digested with concentrated hydrochloric acid, potassium chlorate being added as a saturated solution from time to time in the later stages of the attack, a total of about 1 g. of the solid being used. The excess of hydrochloric acid was removed by evaporation, and the contents of the dish were baked a t 110" C. for 1 hour. The ferric chloride was extracted with 1 : 1 hydrochloric acid, and the residue was fused in a platinum crucible with 3 6 g . of sodium carbonate. After treatment with hydrochloric acid, the residue was evaporated to dryness and baked again at 110" C. for 1 hour. This second extraction in hydrochloric acid was filtered into the main solutiop, which was diluted to 1 litre.For each titration 25 ml. of this solution was used. The iron was reduced with approximately 0.4 M stannous chloride in hydrochloric acid, and the small excess of this reagent was oxidised by 0.25 g. of mercuric chloride added as a solution. The final volume was about 120 ml., and was approximately normal with respect to hydrochloric acid, except that twice this quantity of acid was used for FeP. When phosphoric acid was present the volume added was 10 ml. of 1 : 1 acid in a final volume of 120ml. Titrations were made with the sample of potassium dichromate used previously in 0.1 N solution. The results are given in Table 11. TABLE I1 DICHROMATE REQUIRED FOR 2 5 ~ ~ . OF SOLUTION OF IRON ORE Without H,PO,, With H,PO,.Diphenylamine . . .. .. .. 26.67 rt 0.005 26-61 f 0-005 Diphenylbenzidine . . .. .. . . 26.66 3z 0-005 No result possible Barium diphenylamine sulphonate .. 26-66 f 0.01 26.65 f 0.015 26.69 rt 0.015 ml. ml. - Ferrous phenanthroline . . .. .. Potassium ferricyanide . . .. .. 26-57 rt 0.03 - COMMENTS ON INDICATORS The various substances present in the more complex solution made little difference to the behaviour of the indicators. The diphenylamine group always gave a violet colour at the end-point, whereas a blue was sometimes obtained with pure ferrous ammonium sulphate. The colours faded more rapidly in the presence of a small excess of the oxidising agent. DIPHENYLAMINE-PhOSphOriC acid reduced the indicator range from about 0-2 to about 0.03 ml.It also improved the quality of the colour change, which was, however, moderately good in its absence. The reduction by the phosphoric acid of the volume of dichromate equivalent to the ore remains unexplained. The effect was checked by titrating a volume of the solution to just short of the end-point. It was then divided, and phosphoric acid was added to one half. One drop of dichromate caused this portion to change colour, but 2 drops were required to develop the full colour in the other. DIPHENYLBENZIDINE-AS before, the full violet colour was slow to develop, and therefore the final additions of the dichromate had to be made slowly. Phosphoric acid again delayedMarch, 19501 STOCKDALE: THE ESTIMATION OF IRON BY DICHROMATE 155 the oxidation to such an extent that the use of this indicator in conjunction with this acid was not possible.BARIUM DIPHEXYLAMINE SULPHONATE-The end-point in the absence of phosphoric acid appeared to be somewhat indefinite and the colours obtained were again muddy. The indicator range was about 0-25ml. In all titrations made in the presence of phosphoric acid the colours were clear and the change at the end-point was marked. The indicator range was only about 0-03 ml. in this case. FERROUS PHENANTHROLINE-The colour of the FeCL" ion masked the red colour of the indicator to such an extent that it was almost impossible to obtain an end-point. Even when phosphoric acid was present the colour change was not particularly obvious, and it was found advisable to compare the colour of the solution under titration with the colour of a similar solution containing a small quantity of ferrous iron. The end-point was taken a t the first sign of lightening.POTASSIUM FERRICYANIDE-The end-point with this particular sample of ore proved somewhat troublesome. It will be noticed that the standard deviation for this indicator, though still small, is larger than those for the others. PERCENTAGE OF IRON I N THE ORE With 25.49 ml. of potassium dichromate equivalent to 1 g. of ferrous ammonium sulphate of equivalent weight 392-9 (Table I), and 26-66 ml. of the dichromate equivalent to 25 ml. of the solution of the ore at 10.244 g. per litre (Table II), the percentage of iron in the ore is 58.05. The average value returned by the fourteen analysts co-operating with British Chemical Standards is 58.20 & 0.02.The organiser's own value is 58.09 per cent. Better agreement is not to be expected, in view of the length of the standardisation chain-sodium oxalate, potassium permanganate, ferrous ammonium sulphate, potassium dichromate, ore. Interference by the foreign substances present during the titration of the solution of the haematite can, therefore, only have been small. CONCLUSION s Of the indicators examined, barium diphenylamine sulphonate is the best for use in the estimation of iron with dichromate. Kolthoff and Sandell7 have previously arrived at the same conclusion. The indicator gives a satisfactory result when the solution is at least 0.5 N with respect to hydrochloric or sulphuric acid. If orthophosphoric acid is present the colours are clear and the range of colour change is small. In the absence of this acid, titrations must be continued until the violet colour of the indicator is fully developed. Without phosphoric acid the colours may be somewhat muddy, but the colour change takes place over a rather larger range of dichromate, so giving some warning of the approach of the end- point. Both methods give satisfactory results, and which is selected seems largely to be a matter of personal choice. REFERENCES 1. Hildebrand, J . H., J. Amer. Chem. Soc., 1913, 35, 847. 2.. Kolthoff, I. M., and Sarver, L. A., Ibid., 1930, 52, 4179. 3. Sarver, L. A., and Kolthoff, I. M., Ibid., 1931, 53, 2902. 4. Hume, D. N., and Kolthoff, I. M., Ibid., 1943, 65, 1895. 6. Smith, G. F., and Richter, F. P., Ind. Eng. Chem., Anal. Ed., 1944, 16, 580. 6. Latimer, W. M., Oxidation Potentials, Prentice-Hall, New York, 1938. 7. Kolthoff, I. M., and Sandell, E. B., Textbook of Quantitative Inorganic AnaZysis, Macmillan, New York, 1946. THE UNIVERSITY CHEMICAL LABORATORY CAMBRIDGE June, 1949

ISSN:0003-2654    

DOI:10.1039/AN9507500150

出版商:RSC    

年代:1950

数据来源: RSC