The following article has been condensed and adapted from “Recent advances in application of modified starches for breadmaking,” by Miyazaki, M. et al., first published in Trends in Food Science & Technology, November, 2006, pp. 591 - 599. The original article reviews a number of studies on this subject and references over 50 papers. See end of this article for more information.—Ed.

Starch, a main component of cereal grains and roots, has increased functionality for food use after chemical, physical or enzymatic modifications are made. (See chart “Classification of Modified Starches.”) For example, amylolytic enzymes hydrolyze parts of starch molecules into lower molecular weight ingredients such as maltodextrin or dextrin. Physical modification may involve pre-gelatinization and heat-treatment of the starch. Pre-gelatinized starches are ‘‘convenience products.’’ They are pre-cooked and dried by ingredient manufacturers and reconstituted in water by the end user to give a viscous paste. Heat-treatment processes include heat-moisture and annealing treatments, both of which cause a physical modification of starch without gelatinization, damage to granular integrity or loss of birefringence. Heat-moisture treated starches show improved consistency and viscosity stability in food with a pH of less than 4.5, whereas annealing-treated starches show enhanced viscosity profiles and are considered to be native starches and, therefore, can be labeled as ‘‘food starch.’’ Many kinds of chemically modified starches have been developed with broad application not only in food, but also in the paper and textile industries.

In breadmaking, both wheat protein (especially gluten), and starch play an important role in the formation of the texture of the dough and bread. Starch, about 75-85% of flour, dilutes the gluten, interweaves with gluten and absorbs water from the gluten by gelatinization; thus, it provides a bread structure permeable to gas, so that the bread does not collapse while cooling. As temperature increases and starch gelatinizes, starch competes with other components for the water available in the system. Thus, starch ‘‘sets’’ the structure of baked product systems. Studies reveal that modified starches, which were developed to suppress undesirable properties of native starches, also affected the properties of dough and quality of bread.

Characteristics of Hydroxypropylated, Acetylated and/or Cross-linked Starches

Native starch often does not have the functional properties required for food processing, such as thickening and stabilization. Therefore, starches are modified to help prevent undesirable changes in product texture and appearance such as those caused by retrogradation or breakdown of starch during processing and storage.

Starch contains abundant hydroxyl groups. These hydroxyls are potentially able to react with chemicals having reactivity with alcoholic hydroxyls. Commercially modified starches include hydroxypropylated and/or cross-linked, acetylated and/or cross-linked starches.

Cross-linked starches are widely used as thickeners in foods, particularly where a high and stable viscosity is needed. Cross-linking minimizes granule rupture, loss of viscosity and formation of a stringy paste during cooking. Cross-linking is performed by treating granular starch with multifunctional reagents capable of forming ether or ester linkages with hydroxyl groups in starch. When the specific reagent contains two or more moieties capable of reacting with hydroxyl groups, the two different hydroxyls may react, resulting in cross-linking between hydroxyls on the same molecule or on different molecules. Cross-linking reinforces the hydrogen bonds in the granule with chemical bonds.

Cross-linking a starch molecule improves its stability during cooking, especially under shear and acidic conditions. However, cross-linking also reduces paste clarity and stability against cold storage. Therefore, further modifications such as hydroxypropylation and acetylation have been used to decrease these undesirable characteristics.

Hydroxypropylated starch, formed by reaction of starch with propylene oxide, is commonly used in the food industry. This modification improves the shelflife, freeze/thaw stability, cold-storage stability, clarity and texture properties of the starch paste. In the formation of hydroxypropylated starch, propylene oxide is believed to be substituted primarily at the HO-2-hydroxyl groups in the starch anhydroglucose unit. (See illustration “Hydroxypropylation of Starch.”)

Hydroxypropyl groups are hydrophilic in nature. When they are introduced into starch granules, they weaken the internal bond structure holding the granule together and also prevent water separating from the starch paste through syneresis. This reduction in bond strength also impacts the pasting temperature (the point at which the granules become swollen) of the starch. That is, the higher the level of hydroxypropyl substitution, the lower the pasting temperature. Since the hydroxypropyl groups prevent retrogradation, a more desirable fluid paste with improved clarity results.

When hydroxypropylated starch is then also cross-linked, the resulting paste has increased viscosity stability and a short texture. Swollen but intact starch granules are usually desired in most food starch applications in order to maintain rheological properties. However, for each application, there is an optimum level and balance between hydroxypropyl substitution and cross-linking.

As for starch modified through acetylation, the FDA has approved the use of those with a low degree of substitution (DS) of 0.01-0.2. They are used to improve the binding, thickening, stability and texturing of foods. The introduction of acetyl groups greatly reduces the tendency of the starch solutions to deteriorate in clarity and texture and to synerese when held at low temperatures. The substituted groups stabilize starch by interrupting the linearity of amylose and segments of amylopectin branches. They sterically interfere with intermolecular alignment. Therefore, acetylated starch typically has physicochemical characteristics such as low gelatinization temperature, high solubility, good cooking and storage stabilities.

Modified Starches and Dough Rheological Properties

Wheat protein possesses the unique and distinctive property of forming gluten when wetted and mixed with water. When water is added and mixed with wheat flour, water-soluble—along with water-insoluble proteins—hydrate and form gluten in which starch and other components are embedded. Gluten imparts physical properties that differ from those of dough made from other cereal grains. The gluten is the “skeleton” of wheat flour dough and plays an important role in gas retention. Substituting part of the wheat flour with starches results in a weaker skeleton. Thus, vital gluten is needed as a dough improver. One study recommended adding vital gluten at about 8% of the weight of starch substitution. This is almost equal to the percentage of insoluble wheat protein, glutenin and gliadin combined in wheat flour. Other research shows that if modified corn or tapioca starches are used to replace wheat flour, they should be at less than 20%.

As the amount of starch substituted for wheat flour increases, it is to be expected that the dough and bread will show characteristics of the substituted starches. Both starch characteristics and types of modification also affect the amount of water absorbed by dough.

For example, several studies report that dough made from waxy wheat flour (or partially of waxy wheat flour) absorbs more water than dough made from non-waxy wheat flour. It therefore follows, as one study shows, that dough containing 5-10% of cross-linked, waxy cornstarch (CWCS) and vital gluten increases water absorption by 0.2-1.2%, when compared to 100% wheat flour. In contrast, substitution with non-waxy starch derivatives usually decreases dough water absorption. Other studies show that dough containing 5-10% of cross-linked, non-waxy corn starch (CNCS) or 20% of several chemically modified tapioca starches (along with vital gluten) decreases water absorption by 1.2-2.2% and 1.5-4.0%, respectively, as compared with 100% wheat flour use.

Substituting 5-10% of the wheat flour with cross-linked cornstarch—but without adding vital gluten—decreases dough water absorption. However, when the substitution is with hydroxypropylated starch, there is a slight increase in water absorption of dough made from the same amount of native starch, or wheat flour alone. And, when potato starch that was both hydroxypropylated and cross-linked replaced 10-30% of wheat flour, more water was absorbed, as compared to when the same amount of native potato starch was substituted. Likewise, when 20-30% of wheat flour was replaced with hydroxypropylated tapioca or wheat starch, water absorption increased, compared to when native starch replaced wheat flour or when wheat flour alone was used. Other chemically modified starches, like acetylated starches, generally decrease wheat flour dough water absorption when substituted at less than 20%.

Starch, present in the native state in the dough, absorbs up to about 46% water during dough preparation. However, the cross-linking causes granules to become compact; less water is absorbed. In contrast, when the hydrophilic hydroxypropyled starch is substituted for wheat flour, the dough absorbs more water.

Other studies show that when 20% of the wheat flour in a dough (with added vital wheat gluten) is replaced by modified tapioca starches, the time it takes for water absorption into wheat flour (shown as arrival time on a Farinogram chart) and development time of dough (which appears as peak time) is shorter. However, the mixing tolerance is the same or only slightly decreased.

Studies also show that replacing wheat flour with cross-linked corn starch (CLCS) and vital gluten improves dough texture by making it stronger and more extensible during proofing. The effects were more pronounced when the CLCS had a higher degree of swelling. Generally, bread volumes of substituted and unsubstituted wheat flour were similar. However, bread substituted with a CLCS with a greater degree of swelling exhibited more volume than bread made of 100% wheat flour, or bread using a corn starch with a lower degree of swelling. Bottom line, these results indicate that the level of cross-linkage of starches affects the cohesive, gummy texture of the dough during molding, proofing and baking. However, the role of native and modified starches in the dough is still ambiguous.

Modified Starches and Dough Gelatinization Properties

Starches from different sources differ in granule shape, amylose and amylopectin content, gelatinization temperature and so on. Modified starches have essentially similar properties with their native starches.

Differential Scanning Calorimetry (DSC) is an effective instrument for evaluation of gelatinization properties of dough. It has been found that the gelatinization peak temperature decreases in dough substituted with 5-15% of CLCS and vital gluten, although gelatinization endothermic enthalpy does not change. Dough substituted with 20% of tapioca starch and vital gluten increases both gelatinization peak temperature and endothermic enthalpy, when compared with 100% wheat flour dough. Dough containing 20% of acetylated (DS 0.03-0.04) or cross-linked (with phosphorus substituted at 0.015-0.02%) tapioca starch slightly decreases gelatinization peak temperature, as compared with the same amount of native tapioca starch, but shows higher gelatinization temperature than wheat flour. However, dough containing 20% of hydroxypropylated tapioca starch (DS 0.09-0.11) significantly decreases gelatinization peak temperature and its endothermic enthalpy, showing the same gelatinization properties as the wheat flour dough in another study. This suggests that the hydroxypropylated tapioca starch with DS 0.09-0.11 does not compete with wheat flour for the available water in the system.

Brabender amylographs and RVA (rapid visco analyzer) have been used to evaluate viscosity changes in wheat flour or starch-substituted wheat flour during heating and cooling. Wheat flour substituted with corn starch and CLCS—especially waxy corn starch—shows higher gelatinization viscosity than wheat flour alone. Wheat flour substituted with native, hydroxypropylated or acetylated tapioca starches also shows higher gelatinization viscosity and breakdown than wheat flour alone. However, phosphorylated, cross-linked tapioca starch significantly decreases gelatinization viscosity and breakdown. What this means is that native, hydroxypropylated and acetylated tapioca starches swell and collapse easily during heating, while phosphorylated, cross-linked tapioca starch is hard to swell and disperse—as compared with wheat starch or native starch.

Other studies with cross-linked corn starches of low and high degrees of swelling indicate that the chemically bonded cross-link may provide sufficient granule integrity to keep the swollen granules intact and minimize or prevent loss in viscosity. Thus, the cross-linkage by phosphorous has a significant effect on the pasting properties.

Texture and Staling of Bread

Bread crumbs prepared from flour that has been substituted with native, hydroxypropylated and acetylated tapioca starches have a tacky texture, while bread with phosphorylated, cross-linked tapioca starch has a dry feel. On amylographs, flour substituted with the first types of modified starches show higher peak viscosity and larger breakdown than 100% wheat flour, while flour substituted with the second type of starch show just the opposite. These results show that bakers can develop uniquely textured breads using modified starches.

Bread staling has been extensively studied. One significant factor is the rearrangement of starch fractions. A number of studies have looked at the impact of substituting wheat starch or wheat flour with modified starches. They indicate that some cross-linked waxy starches possess anti-staling ability, although substitution with cross-linked, non-waxy starches for wheat flour generally speeds bread staling. Other studies with hydroxypropylated or acetylated tapioca starch indicate that the firming of bread with various chemically modified tapioca starches is associated with amylopectin retrogradation.

Gelatinization and retrogradation characteristics are considered to be quite important for bread texture and quality. Among various modified starches, hydroxypropylated starches are the most effective to retard staling. DSC data show that dough containing 20% of hydroxypropylated tapioca starch and vital gluten has the same gelatinization temperature and endothermic enthalpy as wheat flour dough. This would be a key factor in determining if hydroxypropylated tapioca starch is suitable for breadmaking. As ether or ester modification to starches lowers gelatinization temperature and changes endothermic enthalpy, these modifications are thought to be good strategies to apply to various non-wheat starches (where they may slow staling). Although a large amount of substitution with cross-linked, non-waxy starches for wheat flour speeds bread staling, cross-linking is an essential method to control stable viscosities of starch gel. Researchers report that cross-linked, hydroxypropylated potato starch shows softer crumb than that containing native potato starch. Therefore, treatment with both hydroxypropylation and cross-linking, or acetylation and cross-linking, might be more suitable for breadmaking.

On a final note, it is, of course, important to consider ingredient costs, which may vary greatly among countries. For example, in Japan, due to low tariffs, modified starches tend to be competitively priced against wheat flours, thus making the replacement of flour more financially feasible.

Megumi Miyazaki, Pham Van Hung, Tomoko Maeda and Naofumi Morita, Contributing Editors

The original article, with some 50 references, has been condensed and adapted by Claudia D. O’Donnell, chief editor. It is reprinted from Trends in Food Science and Technology, Vol. 17, Issue 11, Miyazaki, et al., “Recent Advances in Application of Modified Starch,” pp. 591-599, copyright 2006, with permission from Elsevier.