Minerals in foods are essential for the maintenance of good health. It is therefore paradoxical that a criteria widely used for defining the quality of white flour is a low ash value.
The milling industry and its customers are not motivated by the desire to undermine the health of those who consume their products. So, why should flours with higher mineral content — as revealed by a high ash value — be penalized in the marketplace?
Many nutritionally important substances, including vitamins and minerals, reside in the outer parts of the wheat grain (see table below). But it is in the endosperm where the functional components (starch and gluten-forming proteins) reside — functional, that is, in the context of baking.
There is no doubt that these functional components are valuable, as only they can confer the light, aerated texture that consumers so enjoy in baked products. But do they have to be separated to quite the degree inflicted by the requirement for a low ash value? The purpose of this article is to examine which non-endosperm components of the grain can be tolerated and how the unacceptable components should be measured, and perhaps controlled, if the milling industry chose not to rely on ash values.
To examine this possibility further, it is necessary to look at wheat bran in some detail. Wheat bran is usually referred to as a single commodity but it is actually a complex laminate consisting of several layers, most of which are complete but some of which are fragmentary. To complicate the issue further, the composition and structure of each layer is different.
From a technological point of view, there are other differences as well because the inclusion of some wheat bran layers in a flour causes more reduction in quality of baked products than do others. There are several reasons for this but most are related to color and structure. Dark particles disfigure the flour and its products, and the shape and texture of particles determines how they accommodate to the delicate structure of, for example, an aerated crumb.
The layer that causes most problems is the outer epidermis of the pericarp, sometimes known as "beeswing." This layer is colored and consequently adversely affects the whiteness of the product, but its shape is also undesirable in baked products because it includes needle-like hairs that in the whole grain form the brush at one end (see Figure 1). These hairs have been shown to puncture the gas cells in aerated doughs and baked goods (see Figure 2), reducing their capacity to retain gas.
At the other extreme is the aleurone layer, which contains many minerals as well as vitamins and nutritionally well-balanced protein. This layer has a color that is similar to the endosperm and, when present in a flour, integrates well into the gas cell walls (see Figure 3), causing no leaks.
Ironically, it is the aleurone that is most sensitively detected and discriminated against by the ash test.
In large pieces of bran the various layers remain together, but in the fine particles that manage to pass through flour sieves to enter the flour streams there is greater likelihood of the layers being separated. Beeswing is particularly likely to become separated because it is only loosely attached to the adjacent tissues.
Because the composition of the layers varies, the sensitivity with which each is detected by a chemical method also varies. The mineral contents of the layers are indicated by their ash values (see Figure 4 on next page). The effects of these differences (see Figure 5) shows how much of each would be present in a flour with an ash value of 0.5% if it were the only bran layer present and the endosperm ash was 0.3%. The harmless aleurone with its desirable nutritional attributes would be given very low toleration, while the beeswing, with its needle-like trichomes, could be tolerated with relative impunity.
By varying milling practice, it is possible to include in a white flour different amounts of the respective bran layers, while maintaining the same ash value.
EFFECT OF DEBRANNING. An illustration of an extreme case of such variation is provided by comparing flours from conventional milling with those from the same grist milled by a system including an initial debranning stage.
Flows including debranning allow the outer colored layers of bran to be removed without the entire aleurone layer being abraded off. This is quite different from the conventional roller milling flow, where the entire aleurone layer, and some outer endosperm, enter the bran stream attached to the dark outer layers.
Analyses of all the machine flours from both millings were carried out and ash curves were plotted for both (see Figure 6a on page 62). The curves were almost identical. However, cumulative curves were also plotted for the amount of aleurone layer and the amount of dark bran specks (see Figures 6b and 6c).
It is clear that the minerals recorded in the ash curves of the conventionally milled flours originated in the dark bran layers while minerals in the flours milled from pearled wheats originated in the aleurone layer. Because of the high concentration of minerals in the aleurone layer and the similarity in ash value of the two flour types, the amount of non-endosperm material in the pearled flour must have been much less than in the conventionally milled flour.
The beneficial effects of minerals in the diet may be partly attributed to claims that the aleurone layer has functional food properties. Because of this claim, a major baking company in Australia has marketed a type of bread in which aleurone tissue is intentionally included in the flour. The exercise has proved two points: first, that aleurone can be included at high levels while maintaining good bread texture; and second, that a high ash value would indicate successful flour production rather than inferior processing.
Even without invoking shortcomings in nutritional potential resulting from maintaining ash as a criterion, it is possible to demonstrate its inadequacy as a measure of the proportion of endosperm in white flour. Although minerals are concentrated in the grain's outer layers, they are also present in a lower concentration in the endosperm itself. This in itself does not make ash a bad measurement, but the fact that the concentration can and does vary considerably undermines it as an acceptable criterion of purity.
Scientists at TNO Nutrition and Food Research, Zeist, The Netherlands, demonstrated the extent of variation that occurs. They harvested a range of wheat varieties, grown on a number of sites in three consecutive years, and found that flours milled to the same extraction rate from the 300-plus samples produced flours with widely differing ash values.
The TNO scientists examined the ash values of the parent wheats and found a high correlation between flour and wheat values. As ash purports to be a measure of quality, a high correlation with good baking parameters would be expected. However, the relationship was not significant.
ALTERNATIVE METHODS. If ash is not acceptable as a criterion of purity and quality, do any alternative methods perform better? The Dutch experiment, in addition to measuring ash, also measured the amount of colored bran using image analysis. A good correlation was demonstrated between this measure and quality parameters (see table on page 61).
There are, of course, other means of measuring non-endosperm particles in flour. In some countries, the Kent Jones and Martin Color Grader provides the standard while the Tristimulus system is used elsewhere. However, both methods suffer in a similar way to ash in that they respond to other factors as well as to the presence of non-endosperm grain components.
This was demonstrated in an experiment in which bran powder was added at levels between 0 and 5% to five patent flours milled from different wheat varieties. The results (see Figure 7 on page 63) reveal that at each level of addition a wide range of values was recorded for all methods except image analysis.
The extent of the range was not the result of poor reproducibility, because the order of the samples was consistent throughout. Rather, it resulted from a response to factors other than added bran powder.
In the case of color measurements, the response was to pigmentation in the endosperm. In the case of ash, variation in mineral content of the endosperm was the reason. With Tristimulus, particle size was responsible. The fact that the ranges overlapped indicates that the levels of addition could not be adequately distinguished by most methods. Only in the case of image analysis were the ranges for each addition level discreet.
In spite of image analysis having been applied in the cereals industry only relatively recently, several useful applications have arisen, including analyzing the size of cells in breadcrumb and identifying non-complying grains in seed or other samples where species purity and freedom from disease are important.
Image analysis exploits the possibilities that arise through digitizing a picture generated by a CCD (charge coupled device) or other television camera, a digital still camera or a scanner such as a flat-bed type. The digitized image consists of a symmetrical, two-dimensional array of uniformly sized rectangular pixels or picture cells so that when seen at high resolution it has the appearance of having a squared grid placed over it.
Each pixel can be interrogated electronically to identify its color or, in a monochrome image, its shade of grey. By relating information from pixels in the same neighborhood, the size of objects of the same color or shade can be distinguished from the background and their sizes determined as a proportion of the image area.
Applied to flours, the technique can measure the number of individual specks that are darker than a threshold grey level and can express the total proportionate area of the image that is occupied by darker pixels. The darker particles are usually of bran, so a value for the percentage of dark bran in the sample can be derived.
Sometimes black particles are present as well as bran. By selecting two threshold values, the two unwanted components can be quantified separately. This is particularly relevant to durum semolina because "blackpoint" problems commonly arise.
In the early days of image analysis, the data processing computer filled a room. Now, a standard desktop personal computer is more than adequate for making the analysis and driving a control system that responds to the results of a continuous monitoring system.
Systems can be applied in the lab or on-line in a mill. The hardware for both applications can be simple, reliable and durable. Testing in both applications is rapid, non-destructive and simple.
Some of the shortcomings of ash have been known for many years, but it has been retained in the milling industry because no sufficiently attractive alternative method has been available. The situation has now changed and a viable alternative that can be applied both on and off line is available in image analysis-based systems.
To assure the fairest possible trading in which flour quality is clearly defined, it is time for the milling industry throughout the world to review its practices with regard to ash.
Tony Evers lives in the United Kingdom and has been a consultant for the cereals industry for the past three years after retiring as head of the Cereals and Milling Department at Campden and Chorleywood Food Research Association (U.K.) Ltd. Finlay McRitchie is a professor in the Department of Grain Science and Industry at Kansas State University, Manhattan, Kansas, U.S. The authors are grateful to Campden & Chorleywood; TNO Nutrition and Food Research; BRI Australia Ltd.; Journal of Cereal Science; and Branscan for permission to include their results and illustrations.