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