Predicting Baking Performance through Evaluation of Short-crust Dough

Rana Cheaib
Masters Thesis project at the Department of Food Technology, Faculty of Engineering, Lund University, SwedenAarhusKarlshamn AB (AAK) in Malmö, Sweden

Abstract

The three major components of short-crust dough are flour, sugar, and fat. Since high fat contents have been shown to have a major effect on the development of the gluten network, studying how these ingredients could affect the texture of the dough and the baked product became interesting. In addition, there are no existing methods on short-crust dough that allow predicting the characteristics of the baked product based on those of the dough. Therefore, the task was to study whether developing such a method is possible. In this paper, the amount of ingredients was varied and textural analyses were run on dough and baked samples. The results were analyzed using chemometrics and statistical tools and different graphs were plotted to visualize the relations between the variables and parameters. The analyses showed the margarine and egg amounts have a significant effect on the texture of the dough in terms of resilience, gumminess, and hardness. In addition, the hardness of the baked product has been shown to be positively correlated with the same parameters, which in turn can be controlled by the addition of fat and/or eggs according to requirements. Thus, prediction of baking performance based on dough characteristics is shown to be possible.

Keywords: short crust dough, rheology, textural characteristics, texture, shortening, margarine, texture analyzer, baking

Introduction

In daily life, at least at home, no more than basic knowledge is required to be able to produce bread and other bakery products. However, in order to improve such production, especially in industrial processing and applications, proper scientific understanding of the components and their interactions is a priority. Taking this to another level, it is valuable to be able to predict baking performance and the characteristics of the finished product based on those of the dough. This will help improve production in terms of quality and quantity, since time and materials will be saved. While methods for predicting bread-baking performance through dough evaluation already exist, no such methods have been developed for short crust dough. Therefore AAK suggested a study on the possibility of developing such a method.

Flour, sugar, and fat are the three major components of short crust cookies (Zucco, et al., 2011). Fat could be added as butter, shortening, or margarine. The high content of fat increases incorporation of air in the dough especially when subjected to a creaming stage. In addition, fat acts as a lubricant and competes with the aqueous phase and limits the gluten formation (Maache-Rezzoug, et al., 1998, Slade & Levine, 1994, Wade, 1990). However, during baking fat melts and together with sugar they increase the mobility of the dough and hence result in larger dough spread (Pareyta, et al., 2009). The dough components thus influence the dough rheology, dough making and handling, and baking quality that influence the rheology characteristics (Pedersen, et al., 2004). According to Bourne (1990) methods such as compression, bending-snapping, and puncture principle have been used to study rheological and textural properties of dough and baked products.

Materials and Methods

Materials

Wheat flour produced by Nord Mills (moisture content less than 15.5%; Kärnvetemjöl (ARTNR 140233)) granulated sugar, egg powder, and cold drinkable tap water. Four types of fat based on S100 were used in the experiments especially produced at AAK for this study: (1) shortening, (2) shortening with 2% emulsifier, (3) margarine, and (4) margarine with 2% emulsifier.

Dough preparation

Mixing: sugar, fat, and egg powder were mixed at low speed (60 rpm) in a mixer (Hobart N50 5-Quart Mixer) for 30 sec., scraped down, water was added, ingredients creamed for additional 30 sec., and the content was scraped down. The speed was increased to 124 rpm and creaming was allowed for 4 min. The cream was scraped down, flour was added, and the content was mixed at 60 rpm for 90 sec., the content was then scraped down, speed was increased to 60 rpm, and the content was mixed for 3 min. and 30 sec. Note that when margarine was used instead of shortening, in the first step the content was mixed for 60 sec. instead of 30, and no water was added since margarine consists of 20% water.

Sheeting: two rulers of the same height were placed on each side of a baking paper, and the sheeting was performed by allowing the edges of a rolling pin to roll on the rulers so that the dough would obtain a standardized height of 6 mm.

Sampling: a mold with a diameter of 42 mm was used to cut the samples. The excess dough surrounding the samples was carefully removed by lifting or pushing away. Using a knife, the baking paper was cut around the sample and a thin pie lifter was used to lift the samples with the baking paper and transfer onto a small petri dish. Thus sample deformation would be avoided and a standardized method would be obtained. The samples were then divided into three groups (baking, compression, and holding) composing five samples in each group.

Baking: the samples in the baking group were weighed and using a caliper the diameter was measured. A baking tray covered with baking paper was prepared and samples were lifted and tilted over (upside-down). The small piece of baking paper was gently removed to avoid deformation. The samples were baking for 20 min. at 180°C, and stored for at least two weeks before analyses.

Color and textural measurements

Color measurements: the color was measured using a color measuring instrument (CR-400 Chroma Meter, Konica Minolta).

Textural measurements: to determine the textural characteristics of the samples TexVol Texture Analyzer TVT-300XP was used. Three different methods were used, method (1) and (2) were used on dough and (3) on the baked product.

1. Double compression: double compression cycle with compression 4 mm over plate and 50 sec. pause between cycles. Probe P-Cy75A was used and the pre-test, test, and post-test speed was set to 0.2 mm/s.

2. Hold until time: a single cycle with compression 4 mm over plate and holding time 62 sec. Probe P-Cy75A was used and the pre-test, test, and post-test speed was set to 1.0 mm/s.

3. Three point bend: a single cycle that breaks the sample with a compression of 10%. Probe P-BP70A and rig R-TPBR were attached and the pre-test, test, and post-test speed was set to 1.0 mm/s.

Statistics and chemometrics

Experimental design: in order to detect the effect of the different ingredients as well as the interactions between them, a central composite design known as CCD was used. It consists of a complete factorial design combined with replicates in the centrum (Brereton, 2003). The variables were varied at 3 levels (-1 0 +1), which allow neglecting triplicates (Miller & Miller, 2010).

Statistical analysis: to analyze data obtained from the different measurements, different tools were used:
-       PCA: consists of multivariate projections of the observations onto a two-dimensional plane. This enables visualization of the structure of the investigated data set and reveals the relationships between variables and observations as well as the relationships within the variables themselves (Brereton, 2003, Eriksson, et al., 2001).

-       PLS2: a projection method as PCA that handles complex models and strongly correlated responses or parameters (Brereton, 2003, Eriksson, et al., 2001). This method was used to observe relations between variables and parameters as well as within parameters themselves.

-      ANOVA: a tool used to study the significance of the differences between two or more vectors.

Results and Discussion

Effect of ingredients on characteristics of dough and baked product

PLS2 was used to analyze the effect of ingredients on the parameters. Figure 1 shows that the amount margarine is negatively related to resilience (DCResilience), gumminess (DCGumminess), and hardness (DCForceA) obtained by double compression on dough in addition to the hardness of the final product (TPBForceA). This is due to the angle between their vectors being much larger than 90° and the considerable length of their vectors which are almost the same, giving a high scalar product (Brereton, 2003). This means that higher amounts of margarine result in decreasing the values of these parameters. Note that resilience is how much energy an object can absorb without causing permanent deformations (Campbell, 2008). Margarine seems also to be positively correlated with the adhesiveness, springiness, and stringiness of the dough, which means that higher margarine values result in higher elasticity and stickiness of the dough. This observation has also been stated by Lai and Lin (2006).

On the other hand, Figure 1 shows that shortening has no relation to DCGumminess and DCForceA as the angle between their vectors is 90°, yet, has a negative relation to DCResilience. What also can be seen in Figure 1 is that the amount of eggs is positively related to DCResilience, DCGumminess, DCForceA, and TPBForceA, but negatively correlated with DCAdhesiveness, DCSpringiness, and DCStringiness, which clearly is the opposite of margarine. Regarding brightness of the cookies (TPBBrightness), both eggs and margarine showed to be strongly related, however in an opposite manner, which is very logical since a high egg amount gives darker cookies due to the protein content in eggs reducing sugars which enhances Maillard reactions (Maillard, 1912).

These relations were studied further by calculating the standard error of estimating the different parameters by the ingredient values, and calculating the confidence interval for parameter estimation with 95% significance level. The results showed that shortening and margarine values could be used to predict all parameters since the zero was not included in the calculated confidence interval. Regarding egg powder, results showed that almost all parameters could be estimated except DCCohesiveness and DCChewiness. By this, it could be concluded that all the observations are true on 95% significance level.


Figure 1: Relation between input and output parameters

Relations between dough characteristics and those of the baked products

Figure 1 revealed that there seems to be a positive relation between TPBForceA, DCForceA, DCGumminess, DCResilience, DCChewiness, but negative relations to DCAdhesiveness, DCSpringiness, and DCStringiness. Therefore we used PCA to see if this observation could be supported. Initially, a cross validation was made (Figures 2 and 3) to analyze how many principal components should be considered to obtain a proper explanation of the data. The scree-plot (Figure 2) shows that two components are needed to describe the data, however, in the PRESS-plot (Figure 3) the slope increases between components 1 and 2 and then decreases to three components. Consequently, it was concluded that three components would be the best choice to describe the data, however since it is not possible to visualize the relation between the parameters in a 3D plot, the different principal components were plotted against each other in Figures 4 and 5.


Figure 2: Scree-plot illustrating the number of components needed to describe the data


Figure 3: PRESS-plot illustrating the number of components needed to describe the data


Figure 4: Loadingplot PC2 against PC1 with 59% degree of determination


Figure 5: Loadingplot PC3 against PC1 with 59% degree of determination

When plotting PC2 against PC1 interesting results were obtained (Figure 4). This graphs shows that there are strong positive correlations between TPBForceA to DCResilience, DCGumminess, and DCForceA, as well as negative relations to DCAdhessiveness, DCSpringiness, and DCStringiness. However, since the explanation degree is 60%, PC3 against PC1 was considered to see if this pattern is still obvious.

In Figure 5, the relations above become clearer. Looking at the right side of the graph, the correlated parameters are ordered on a straight line over TPBForceA. In addition, the correlations seen in Figures 1 and 4 could still be seen in Figure 5 which is a good sign.

Conclusions

The amount of margarine added to short-crust dough has a negative effect on the hardness of the finished product, the gumminess, and the resilience of the dough. Meaning that increasing the amount of margarine would increase the ability of the cookie to deform upon application of force. However, this could be overcome by the addition of egg powder in appropriate amounts, since the egg powder has been shown to be negatively correlated with the parameters above. In addition, the hardness of the baked product is positively linked to the resilience, gumminess, and hardness of the dough. Since these results were obtained by three methods (PLS2, significance test, and PCA) and observed in three plots (Figures 1, 4, and 5), it is possible to say that measuring the resilience, gumminess, or hardness of the dough would allow the prediction of the characteristics of the baked cookie. Conclusively, in order to control the dough properties by balancing the amount of fat and eggs to the needs, it is possible to control the hardness of the baked product. 

Acknowledgments

This work is done within the framework of a Masters thesis in Food Technology at the University of Lund, Sweden. Special thanks to Malin Sjöö, Ann-Charlotte Eliasson, Jörgen Andersson, and Jeanette Purhagen for the help they provided during this work.

References

Bourne, M. C. (1990). Basic Principles of Food Texture Measurement. In H. Faridi and J. M. Faubion (Eds.), Dough Rheology and Baked Product Texture. New York: Van Nostrand Reinhold.

Brereton, R. G. (2003). Chemometrics: Data Analysis for the Laboratory and Chemical Plant. Chichester: John Wiley & Sons.

Campbell, C. F. (2008). Elements of Metallurgy and Engineering Alloys. United States of America: ASM International.

Eriksson, L., Johansson, E., Kettaneh-Wold, N., & Wold, S. (2001). Multi- and Megavariate Data Analysis- Principles and Applications. Umeå: Umetrics AB.

Lai, H. M., & Lin, T. C. (2006). Bakery Products: Science and Technology (1st ed.). Ames: Blackwell Publishing.

Maache-Rezzoug, Z., Bouvier, J. M., Allaf, K., & Patras, C. (1998). Effect of Principal Ingredients on Rheological Behaviour of Biscuit Dough and on Quality of Biscuits. Journal of Food Engineering, 35(1),(23-42).

Maillard, L. (1912). Action of Amino Acids on Sugars. Formation of Melanoidins in a Methodical Way. Compt Rend, 154,(66-68).

Miller, J. N., & Miller, J. C. (2010). Statistics and Chemometrics for Analytical Chemistry. Edinburgh Gate Pearson Education Limited.

Pareyta, B., Talhaouia, F., Kerckhofsb, G., Brijsa, K., Goesaerta, H., Weversb, M., & Delcoura, J. A. (2009). The Role of Sugar and Fat in Sugar-Snap Cookies: Structural and Textural Properties. Journal of Food Engineering, 90(3),(400-408).

Pedersen, L., Kaack, K., Bergsoe, M. N., & Adler-Nissen, J. (2004). Rheological Properties of Biscuit Dough from Different Cultivars, and Realtionship to Baking Characteristics. Journal of Cereal Science, 39,(37-46).

Slade, L., & Levine, H. (1994). Structure-Function Relationships of Cookie and Cracker Ingredients. In H. Faridi (Ed.), The Science of Cookie and Cracker Production, (23-141). New York: Champan and hall.

Wade, P. (1990). Biscuits, Cookies, and Crackers. The Principles of the Craft-Vol 1. Food Chemistry, 37(1),(77-78).

Zucco, F., Borsuk, Y., & Arntfield, S. D. (2011). Physical and Nutritional Evaluation of Wheat Cookies Supplemented with Pulse Flours of Different Particle Sizes. LWT- Food Science and Technology, 44(10),(2070-2076).