How the RVA can contribute to understanding extruded foods and feeds

B Elliott, J.M.C. Dang and M.L. Bason
Newport Scientific Pty. Ltd., Warriewood NSW 2102. Australia.

INTRODUCTION

The extrusion process offers fast and continuous production of value added products such as ready to eat breakfast cereals, snack foods, pet foods and aqua feeds. However, control of the extrusion process to provide consistent product quality has proved difficult. Changes in ingredient quality, water addition rate and process conditions and parameters such as barrel temperature, screw speed, motor power, die pressure and wear cause variations in final product quality.

Extrusion processing of cereal based products involves the breakdown of starch and its interaction with the other components in the formulation. Starch, and its degree of transformation by thermal and mechanical inputs, plays a dominant role in determining final product flavour, texture, expansion, binding, digestibility, calorific value, resilience and shelf life. Therefore, measuring starch transformation is a reasonable approach to assess final product quality. Since starch transformation persists well beyond loss of birefringence and other measures of gelatinisation, Paton and Spratt (1981) suggested that ‘degree of cook’ be used to describe the degree of starch transformation.

Off-line pasting viscometry has been shown to adequately measure degree of cook in starch (Harper 1994). The method involves re-cooking the product and monitoring viscosity changes. This has been used to quantify the cold-swelling 'cooked' component, the 'raw' component that pastes during the test, and the overall viscosity that indicates degree of starch dextrinisation. Ryu et al (1993) suggested that the Rapid Visco Analyser (RVA) could be used for at-line control as it provides a relatively rapid measure of starch degradation.

RVA METHODS TO ASSESS EXTRUDATES

Sample preparation

Direct expanded products and low moisture extrudates (<10% moisture) are generally stable for up to 1 hour. Sample preparation involves only grinding and sieving. Product is stored (Whalen et al 1996). Above this moisture (above starch glass temperature), polymeric doughs are snap-frozen with dry ice and kept frozen throughout preparation to preserve the structure of the product by locking it in the solid phase. The milled sample is sieved and the fine portion kept frozen for testing (Paton and Spratt 1991).

All products need to be reduced in size for RVA analysis. Blade mills are suitable, particularly those that can hold a small amount of liquid nitrogen or dry ice to maintain the frozen state and fracturability of the sample and to prevent thermal degradation. Samples need to be sieved to obtain a uniform distribution of material for RVA analysis. The screen chosen should not affect sample composition (separate out fibre or other components) and generally 50-75% of material should pass the screen.

Rapid Visco Analyser

The RVA provides an index of how cooked a sample is by re-cooking under relatively low shear in excess water and measuring the pasting viscosity throughout the test.

Typically 3.00-6.00g (dry weight basis) of sample is added to distilled water to give 28.0g total mass, and analysed using the RVA extrusion method pasting profile (Table 1). To aid dispersion, samples may be shaken vigorously for 5-10 seconds in the RVA canister prior to starting the test. Hygroscopic samples may be first wetted with ethyl or isopropyl alcohol to prevent lumping, and analysed using a modified RVA extrusion method pasting profile (Table 1).

Table 1. RVA Extrusion Methods

Extrusion method (no alcohol)

 

Extrusion method (with alcohol)

Time

Type

Value

 

Time

Type

Value

00:00:00

Temp

25°C

 

00:00:00

Temp

25°C

00:00:00

Speed

960 rpm

 

00:00:00

Speed

960 rpm

00:00:10

Speed

160 rpm

 

00:00:10

Speed

160 rpm

00:02:00

Temp

25°C

 

00:02:00

Temp

25°C

00:07:00

Temp

95°C

 

00:07:00

Temp

90°C

00:10:00

Temp

95°C

 

00:10:00

Temp

90°C

00:15:00

Temp

25°C

 

00:15:00

Temp

25°C

Idle Temperature: 25 ± 1°C

End of Test: 20 min.

Time Between Readings: 4 sec.

 

Idle Temperature: 25 ± 1°C

End of Test: 20 min.

Time Between Readings: 4 sec.

Extrusion

Generic cereal-based formulae based on rice flour, corn pollard no. 2 and freshly milled wheat meal were extruded a commercial scale twin-screw co-rotating extruder (Clextral BC-72) fitted with three segments. Two 8 mm diam. dies were used for the doughs and six 2 mm dies for direct expanded product, and the feed rate was set to 300 kg/h. Extruder conditions were varied to give temperatures of 130 - 200°C, screw speeds of 100 - 200 rpm, and water additions of 12 - 29%. (Whalen et al 1996).

RVA RESULTS AND DISCUSSION

Effect of cooking starch

The overall effect of cooking shows pasting curves of a raw corn formulation, a dough from the same product indicating the effect of moderate cooking, and a direct expanded, highly cooked product (Figure 1). With increased levels of cooking, the raw peak and overall viscosities were reduced, but the cold-swelling viscosity increased. These results are consistent with previous observations (Paton and Spratt 1981).

Cold viscosity is low in raw starch samples. In cooked samples, however, a starch is effectively pregelatinised and will swell in cold water indicating the presence of cold water-soluble products such as damaged granules, granule fragments and perhaps leached polymers. Cold viscosity first rises as starch becomes soluble, then falls as it is progressively dextrinised to lower molecular weight (Mason and Hoseney 1986, van Lengerich 1989). For this reason it is difficult to determine degree of cook from cold viscosity.

A ‘hot swelling’ peak will appear in the hot viscosity region of the pasting curve as heating in the RVA causes swelling of remaining ‘hot swelling’ starch granules (Figure 1, dough curve). Similarly, the setback viscosity will be higher in less degraded starches. Hot and setback viscosities always decline with increasing degree of cook, and are lower than the pasting curve of the raw mix. Samples that are highly cooked in extrusion will exhibit no raw peak at all and only minor increases in viscosity on cooling.

Figure 1. RVA pasting curves of raw corn-based formula, an extruded dough and a direct expanded product (Whalen et al 1996).

Effect of screw speed

Increasing the barrel screw speed, or adding reverse elements to the screw, imparts a higher shear rate on the melt. This effect is reflected in RVA pasting patterns (Figure 2(a)). Higher shear causes a reduction in peak and especially setback viscosity, but a higher cold viscosity (Figure 2), consistent with a higher degree of cook. High shear can cause a larger cell structure and thinner walls in extrudates, increased starch-lipid complexing, and reduced digestibility. Many products are potentially over-sheared.

Effect of temperature

Thermal inputs also affect degree of cook in extrudates, and this is similarly reflected in RVA pasting curves (Figure 2 (b)). Barrel heating, rather than energy through shear, will achieve cooking with reduced dextrinisation. Less sheared materials typically have higher final viscosities in the RVA.

Effect of water

Water acts as a plasticiser for starch, lowering its gel temperature, and also acts as a lubricant thus reducing the amount of shear applied to the starch. Relatively high water additions are used to produce doughs used to make half products, whilst low water additions are used for directly expanded products.

Decreasing the added water progressively increases the shear, causing a progressive change in paste viscosity in the RVA curves from the raw to cooked pattern (Figure 2 (c)). The peak viscosity, reflecting the amount of remaining raw material, reduces, along with the setback viscosity, showing increased starch degradation and the cold viscosity increases.

Figure 2. Effect of extruder settings on RVA pasting curves of corn based extrudates (a) screw speed, (b) barrel temperature and (c) added water. (Whalen et al 1996).

APPLICATIONS

Off the shelf products

RVA analysis can readily be applied to off-the-shelf products. Such information could be used to improve products by providing 'good quality' target pasting curves (Figure 3). Product A shows a curve that suggests a traditional corn flake made by batch cooking of flake grits. Product B indicates an extrusion cooked product, with more large polymers responsible for high cold viscosity. Product C suggests an extrusion cooked dough with high shear resulting in a combination of dextrins and less cooked corn (Whalen 1999).


Figure 3. Comparison of cornflakes made by different processes. (Whalen 1999).

Aged products

Starch retrogradation causes the migration of moisture from starch granules into the interstitial spaces, forming an insoluble starch structure. This is observed as a decrease in the cold viscosity area of the RVA curve (Figure 4).

Figure 4. Comparison of fresh and aged cat food.

Formulation effects

Four different fish feed formulations are evident in the RVA curve (Figure 5). The effect of added emulsifier in some formulations is clearly seen by a rapid rise in the RVA viscosity at approximately 14 minutes, presumably due to the formation of an inclusion complex with the starch.


Figure 5. Comparison of four different fish feed formulations

CONCLUSIONS

The RVA can be used to evaluate all phases of extruded product during production, including raw material, dough, intermediates, dried half-products and final products. Pasting, mechanical and thermal pathways of starch degradation can be observed in the RVA curve, thus it is possible to use RVA paste rheology to assess the degree of cook of an extrudate.

The ability to directly measure product conversion in cook processes in this way is attractive because it has inherent advantages in both routine production and research work.

Defining the target RVA parameters of a desirable product could be used to improve products by providing 'good quality' target pasting curves. By quantifying the interaction between the extrusion process conditions and the transformation of the product for a given cooker, the basis for manipulation of conditions to efficiently achieve or maintain a product quality target, whether this be a pre-existing product being translated to a new cooker during scale-up or equipment change, or for developing new products, is provided.

The RVA also offers the possibility to assess off-the-shelf and competitive products.

REFERENCES

Harper, J. M. (1994) Extrusion Processing of Starch. R. Alexander, J. and Zobel, H. F. (eds.) Developments in Carbohydrate Chemistry, Amer. Assoc. Cereal Chem., St Paul MN. 37-64.

Lengerich, B. van (1989). Influence of extrusion processing on in-line rheological behavior, structure, and function of wheat starch. Faridi, H. and Faubion, J. M. (eds.). Dough Rheology and Baked Product Texture, Van Nostrand Reinbold, NY. 421-471.

Mason, W. R. and Hoseney, R. C. (1986). Factors affecting the viscosity of extrusion-cooked wheat starch. Cereal Chem. 63(5):436.

Paton, D. and Spratt, W. A. (1981) Simulated approach to the estimation of degree of cooking of an extruded cereal product. Cereal Chem. 58:216-220.

Ryu, G. H., Neumann, P. E. and Walker, C. E. (1993) Pasting of wheat flour extrudates containing conventional baking ingredients. J. Food Sci. 58(3):567-573.

Whalen, P. J. (1999). Measuring process effects in read-to-eat breakfast cereals. Cereal Foods World 44(6):407-412.

Whalen, P. J., Bason, M. L. and Booth, R. I. (1996). Measuring degree of cook in extruded foods using the Rapid Visco Analyser, Proc. 45th Aust. Cereal Chem. Conf., Y.A. Williams and C.W. Wrigley (eds.), Royal Aust. Chem. Inst. 289-293.