Optimizing Mixing during the Sponge Cake Manufacturing Process

Optimizing Mixing during the Sponge Cake Manufacturing Process Julia Rodríguez-García, 1 Sarabjit S. Sahi, 2 and Isabel Hernando 1,3 ABSTRACT Sponge cakes have traditionally been manufactured using multistage mixing methods to enhance potential foam formation by the eggs. Today, use of all-in (single-stage) mixing methods is superseding multistage methods for large-scale batter preparation to reduce costs and production time. In this study, multistage and all-in mixing procedures and three final high-speed mixing times (3, 5, and 15 min) for sponge cake production were tested to optimize a mixing method for pilotscale research. Mixing for 3 min produced batters with higher relative density values than did longer mixing times. These batters generated well-aerated cakes with high volume and low hardness. In contrast, after 5 and 15 min of high-speed mixing, batters with lower relative density and higher viscosity values were produced. Although higher bubble incorporation and retention were observed, longer mixing times produced better developed gluten networks, which stiffened the batters and inhibited bubble expansion during mixing. As a result, these batters did not expand properly and produced cakes with low volume, dense crumb, and high hardness values. Results for all-in mixing were similar to those for the multistage mixing procedure in terms of the physical properties of batters and cakes (i.e., relative density, elastic moduli, volume, total cell area, hardness, etc.). These results suggest the all-in mixing procedure with a final high-speed mixing time of 3 min is an appropriate mixing method for pilot-scale sponge cake production. The advantages of this method are reduced energy costs and production time. Cake quality depends on several factors, including selection of ingredients, knowledge of their function, and use of a balanced formula that incorporates precise measurement of ingredients and optimal mixing and baking procedures (6). Optimal mixing procedures differ based on the production scenario, i.e., laboratory study, pilot-scale experiment, or full-scale production. The major process parameters include batch volume, equipment, and mixing method. There are three main goals of cake batter mixing: 1) combine all ingredients into a smooth, uniform batter; 2) form and incorporate air cells in the batter; and 3) develop proper texture in the finished product (7). Complete dispersion of ingredients is a fundamental requirement for producing a good quality cake, and the presence and dispersion of air bubbles also are essential because these bubbles act as the nuclei for cake expansion (6). Multistage mixing methods are based on the separation of specific ingredients to prevent the formation of gluten and to enhance potential foam formation by the eggs (4). Mixing methods used for foam-style cakes, such as sponge cakes, depend on the occlusion of air and stable foam formation provided by the eggs in conjunction with other ingredients, such as sugar and acid. To achieve maximum batter volume, the egg whites and yolks are beaten separately (6). First, the egg whites 1 Research Group of Food Microstructure and Chemistry, Department of Food Technology, Universitat Politècnica de València, Camino de Vera s/n 46022, Valencia, Spain. 2 Campden BRI, Chipping Campden, Glos GL55 6LD, U.K. 3 Corresponding author. E-mail: mihernan@tal.upv.es; Tel: +34 963877000 ext. 78230. http://dx.doi.org/10.1094/cfw-59-6-0287 2014 AACC International, Inc. are whipped to form a foam, and then, sugar is added to stabilize the egg-white foam and form a meringue. Next, an acid is added to the egg-white foam to lower the ph and further stabilize the egg whites (5). Finally, the egg yolks are incorporated; the presence of proteins and lecithin in egg yolks provides unusual extensibility, which is conducive to foam formation (6). Whipped cake batters prepared using wire whips require production in smaller batches (6). In large-scale cake manufacturing, ingredient incorporation and mixing procedures must be optimized to reduce costs and production time. The all-in (single-stage) mixing method has become increasingly common, and in a significant number of cases, this method now supersedes multistage methods of batter preparation (4). In many modern bakeries, all-in methods are used in production scenarios where batter preparation is completed in one stage, particularly when using high-speed mixers (3). Emulsifiers containing mono- and diglycerides are important in all-in mixing methods because they help to trap air in the mixture and promote a finer texture in the final product (7). During cake batter mixing, the movement of the mixing tool pushes the material aside, creating a void behind the trailing edge. As the batter flows into the void, small pockets of gas (air) are entrained. The ongoing movement of the mixing tool through the batter continues to trap air, decreasing the density of the batter. As a result, the length of the mixing time profoundly affects the density of the cake batter: the longer the mixing time, the lower the batter density will be until the minimum density is reached (4). Low batter density is associated with good air incorporation, as well as good bubble retention during mixing. However, air and leavening gas retention during baking are also affected by batter viscosity (2).Therefore, the ability to occlude air and the bulk properties of the batter are important for air bubble retention and creation of a stable sponge structure in the oven that will not collapse after baking (9). This paper addresses aspects of cake production related to the mixing procedure used; two different mixing methods (multistage and all-in) with three final high-speed mixing times (3, 5, and 15 min) were studied. The objective of the experiment was to confirm whether the characteristics of the cakes produced significantly differed depending on mixing procedure and time. Moreover, the study helped determine whether replacing the multistage mixing method with an all-in procedure is possible for cakes formulated without an emulsifier. MATERIALS AND METHODS Ingredients The ingredients used in the cake batter preparations were based on a traditional Spanish formulation (10) that contains a leavening system of citric acid and sodium bicarbonate. Ingredients used included the following (percentages based on flour): 100% plain white flour (13.9% moisture, 9.7% protein) (Golden Dawn, ADM Milling Ltd.), 27% pasteurized liquid egg yolks and 54% egg whites (Framptons Ltd.), 100% white granulated sugar (British Sugar plc), 50% skim long-life milk (Tesco PLC), CEREAL FOODS WORLD / 287

46% sunflower oil (Olympic Oils Ltd.), 4% sodium bicarbonate (Brunner Mond), 3% citric acid (VWR International Ltd.), and 1.5% salt. Batter and Cake Preparation The ingredients were weighed and allowed to reach 20 C. Two different mixing procedures were tested: multistage and all-in. In addition, the final portion of the mixing process was performed at the highest mixer speed for three different lengths of time: 3, 5, and 15 min. The multistage mixing process was performed according to a method previously reported by Baixauli et al. (1), with slight modifications. The egg whites were whipped in a mixer (N50, Hobart Manufacturing Company Ltd.) for 2 min at 255 rpm (speed 3), and the sugar was added and mixed for 30 sec at 255 rpm. The egg yolks, citric acid, and half of the milk were added and mixed at 60 rpm (speed 1) for 1 min, after which the wheat flour was added and mixed at 60 rpm for 1 min. Next, the sodium bicarbonate and salt were added and mixed at 60 rpm for 2 min. Finally, the oil and remaining milk were added and mixed at 255 rpm; at this speed three different mixing times were tested (3, 5 and 15 min). The all-in mixing procedure was performed according to the Campden BRI method (9), with some modifications. The liquid eggs and milk were placed in a mixer (N50, Hobart Manufacturing Company Ltd.). The dry ingredients were sieved and added to the liquids, and the oil was placed on top. The mixing proceeded using a wire whisk at 60 rpm (speed 1) for 30 sec, followed by 1 min at 124 rpm (speed 2), and varying intervals at An ad appeared here in the print version of the journal. 255 rpm (speed 3); at this speed three different mixing times were tested (3, 5, and 15 min). Two replicates of the same formulation were prepared on different days, resulting in two sets of cakes for each formulation. Batter samples were collected for measurement of relative density and rheology. Batters were scaled at 300 g in paper cases, placed in 400 g bread tins (145 mm 75 mm at the base), and baked in a reel oven (Frederick Bone & Co. Ltd.) for 45 min at 180 C. The baked products were cooled at room temperature and analyzed within 24 hr for weight, volume, crumb cellular structure, and texture. The results represent an average of six cakes baked using two different batter preparation methods and three final high-speed mixing times. Batter Properties Relative Density. The relative density of the cake batters was measured with a calibrated density cup of known volume. After mixing, the cup was filled to the brim with batter and weighed. The same procedure was performed using water. Relative density was determined gravimetrically by dividing the weight of the known volume of batter by the weight of an equal volume of water. Relative density was measured first during the mixing procedure (once per minute) to study its evolution and then, after the mixing procedure (after 3, 5, and 15 min at the highest speed [255 rpm]). Measurements were performed in duplicate. Rheology. A rheological study of the cake batters was performed using a rheometer (Rheometrics ARES model, TA Instruments Ltd.). Measurements were taken using a 50 mm diameter parallel plate immediately after mixing. The strain was selected by performing a strain sweep at 1 Hz, and a strain of 0.1, which corresponded to the region in which the batter displayed linear behavior, was chosen. Dynamic oscillatory frequency tests were performed in duplicate with a frequency sweep from 0.1 to 20 Hz. Viscoelastic functions were monitored (Rheometric Scientific software, version 6.4.3.), including elastic (G ) and viscous (G ) moduli and complex viscosity ( *). Measurements were performed in triplicate. Cake Properties Weight Loss during Baking. Weight loss (WL) during baking was calculated as WL (%) = (W batter W cake /W batter ) 100, where W denotes weight (g) (11). Measurements were performed in triplicate. Volume and Crumb Cell Structure. Cake volume was measured using a volume analyzer (BVM-L 370, Tex-Vol Instruments AB, Perten Instruments), and the measurements were analyzed and stored in a database (Bread Calcu, version 7.2.4d_ contin). Measurements were performed in triplicate. The cakes were cut into four vertical slices (15 mm thick) using a slicing machine (Gebr. Graef GmbH & Co. KG). Two slices were scanned in a C-cell (CCFRA Technology Ltd.) following the standard method for collecting images. The scanned images were analyzed using image processing software (ImageJ, National Institutes of Health). Each image was split into color channels, contrast was enhanced, and pixels were converted to either black or white units. Total cell area within the crumb (%), cell area (mm 2 ), and cell circularity were calculated. Three cakes produced from each formulation were used for the measurements. Because each batter formulation was baked in duplicate, producing two sets of cakes, data were obtained by measuring the cells in 12 different images for each formulation. 288 / NOVEMBER DECEMBER 2014, VOL. 59, NO. 6

Texture. Texture profile analysis was performed using a texture analyzer (TA-TXT. plus, Stable Micro Systems Ltd.) and software (Texture Exponent 32, version 4.0.8.0, Stable Micro Systems Ltd.). Measurements were taken for three cakes from each batch, and each formulation was replicated for a second batch of batter. Three cakes from each formulation were selected, and four slices (15 mm thick) were cut from the central region using a slicing machine (Gebr. Graef GmbH & Co. KG). The texture profile was analyzed at 5 mm sec 1 using a strain that was 25% of the original height and a 1 sec interval between the two compression cycles. A trigger force of 5 g was selected. The double compression test was performed with a disc-shaped steel probe (45 mm diameter). The parameters measured by the curves were hardness, springiness, and cohesiveness. Statistical Analysis A categorical multifactorial experimental design with two factors (mixing procedure and final mixing time at highest speed) was used for statistical analysis. An analysis of variance (ANOVA) was performed on the data using statistical software (Statgraphics Centurion XVI, version 16.1.11, StatPoint Technologies, Inc.). A least significant difference (LSD) Fisher s test was used to evaluate the differences between the mean values (P < 0.05). No significant interactions (P > 0.05) were observed between mixing procedure and time when relative density of cake batters was measured. Significant differences (P < 0.05) in relative density were observed for different mixing times (Fig. 2). As mixing time increased, relative density significantly (P < 0.05) decreased. Minimum relative density of cake batters was reached after 12 min of mixing (Fig. 1). During the mixing process, the amount of entrained air eventually equals the amount of air released; this equilibrium coincides with minimum batter density and is unique for each formulation and type of mixer (4). When mixing continued beyond 15 min, the multistage mixed batters showed an increase in relative density, while the all-in mixed batters showed a plateau. Air was disentrained with continued mixing, which allowed some gas bubbles and any carbon dioxide present to escape with the air. Cauvain and Young (4) showed that after 10 min of mixing, the combined loss of carbon dioxide and air disentrainment results in an increase in relative density. Moreover, if the bubble-stabilizing mechanism in a batter begins to deteriorate during continued mixing, air disentrainment and relative density will increase. In this study, the all-in mixing procedure was more efficient at reducing the relative density of batters from the inception of mixing (Fig. 1), and it was possible to achieve lower relative RESULTS AND DISCUSSION Batter Properties Relative Density. Changes in the relative density of cake batters during mixing are shown in Figure 1. Relative density increased during the first 2 min of multistage mixing and first 4 min of all-in mixing because at those times the ingredients were not yet fully mixed. Fig. 1. Changes in relative density during cake batter mixing. Fig. 2. Mean plot, with 95.0% LSD intervals, of relative density of cake batters at three mixing times. Fig. 3. Mean and interactions plots, with 95.0% LSD intervals. A, Mean values for elastic moduli (G ) for each mixing time; B, interactions between mixing time and procedure for viscous moduli (G ); and C, interactions between mixing time and procedure for complex viscosity ( *). CEREAL FOODS WORLD / 289

density in the batters in less mixing time. When longer mixing times were used, however, the multistage method produced lower relative density values. Despite these variations, when relative density values for the three mixing times studied were evaluated, no significant (P > 0.05) differences between the procedures were observed (data not shown). Rheology. No significant interactions (P > 0.05) were found between mixing procedures and times when elastic modulus (G ) was evaluated. Significant differences (P < 0.05) in G were observed for mixing time (Fig. 3A). Significant interactions (P < 0.05) between mixing procedures and time were observed for viscous moduli (G ) and complex viscosity ( *), as shown in Figure 3B and C, respectively. Final high-speed mixing for 3 and 5 min produced batters with similar (P > 0.05) G values. When batters underwent 15 min of high-speed mixing, G, G, and * were significantly higher (P < 0.05) than for the shorter mixing times for both procedures due to greater development of gluten networks with the longer mixing time, which generated a stiffer batter. Loewe (8) also observed that during batter mixing at ambient or refrigerated temperatures, viscosity increased due to development of gluten. Mixing procedure significantly affected (P < 0.05) G and * values, although the effects varied for different mixing times. When multistage mixing was used, no significant differences (P > 0.05) between G and * at 3 or 5 min of high-speed mixing were observed. However, when the all-in mixing procedure was used, G and * were significantly higher (P < 0.05) after 3 min of mixing than after 5 min. When the longest mixing time (15 min) was used, the all-in method produced batters with lower G and * values than did multistage mixing. Cake Properties Weight Loss during Baking. No significant interactions (P > 0.05) were observed between mixing procedure and time when weight loss during baking was evaluated. Therefore, all of the values for each mixing procedure were combined into one mean value. Significant differences (P < 0.05) between the two mixing procedures were observed for weight loss during baking (Fig. 4). Cakes produced from all-in mixed batters showed sig- Fig. 4. Mean plot, with 95.0% LSD intervals, of weight loss during baking for mixing procedures. Fig. 5. Mean plots, with 95.0% LSD intervals, of mean values for volume (A) and total cell area (B) for each mixing time. Fig. 6. Mean and interaction plots, with 95.0% LSD intervals. A and B, mean values for cell area for each mixing time and procedure, respectively; and C, interactions between mixing time and procedure for cell circularity. 290 / NOVEMBER DECEMBER 2014, VOL. 59, NO. 6

nificantly (P < 0.05) less weight loss than those produced from multistage mixed batters. The method used to incorporate ingredients for mixing significantly affected the dispersion, dilution, and hydration of ingredients. Because the weight lost from cakes reflects moisture loss (11), the results suggest all-in mixing may enhance the dilution and hydration of ingredients, thereby alleviating moisture loss during baking. Volume and Crumb Cell Structure. Mixing procedure did not interact significantly (P > 0.05) with mixing time for cake volume, total cell area, and cell area. Significant differences were observed for volume and total cell area (Fig. 5A and B, respectively) and cell area (Fig. 6A and B). Significant interactions (P < 0.05) between factors were observed for cell circularity (Fig. 6C). Mixing batter for 3 min at high speed produced cakes with significantly (P < 0.05) higher volumes and total cell areas. When mixing time was increased to 5 and 15 min, cake volume and total cell area (Fig. 5A and B, respectively) and cell area (Fig. 6A) decreased significantly (P < 0.05). Although batters mixed for 15 min had lower relative densities and higher complex viscosities, cakes with lower volumes were produced. Moreover, these cakes had crumb structures characterized by significantly (P < 0.05) lower total cell areas (Fig. 5B) and smaller cell areas and higher circularities (Fig. 6A and C, respectively) due to higher batter viscosity, which restricts cake expansion during baking, prevents proper bubble expansion during heating, and generates inadequately developed cell crumb structure. No significant differences (P > 0.05) were observed for volume and total cell area when different mixing procedures were used. Multistage batter mixing resulted in cakes with more uniform crumb structures, as characterized by small cell areas and high cell circularities (Fig. 6B and C, respectively). Texture. No significant interactions (P > 0.05) were observed between mixing procedure and time when cake hardness, springiness, and cohesiveness were evaluated. Significant differences (P < 0.05) were observed in hardness and springiness for mixing time (Fig. 7A and B, respectively) and in cohesiveness for mixing time and procedure (Fig. 8A and B, respectively). When the shortest batter mixing time (3 min) was used, the cakes produced exhibited the lowest (P < 0.05) hardness values. When batter mixing time increased, cakes with significantly (P < 0.05) higher hardness values were produced (Fig. 7A). The structure of a cake depends on air bubble retention, which allows controlled expansion and maintains volume after cooling. Cakes produced from batters mixed at high speed for 3 min had the best volume, crumb structure, and hardness qualities. No significant differences (P > 0.05) were observed in hardness when different mixing procedures were used. The cakes produced from batters mixed at high speed for 3 and 5 min had significantly (P < 0.05) higher springiness values than cakes produced from batters mixed for 15 min (Fig. 7B). Decreased springiness was associated with a decrease in volume and total percentage of cells, which generated a denser crumb structure. No differences were found for springiness values related to mixing procedure. Cakes produced from batters mixed for 3 min at high speed had the highest cohesiveness values (P < 0.05). When mixing time increased (5 and 15 min), cakes with lower cohesiveness were produced (Fig. 8A). The all-in mixing procedure produced cakes with significantly (P < 0.05) higher cohesiveness values than did the multistage mixing procedure (Fig. 8B). Cohesiveness is related to the energy required for the second compression and provides information related to the density and energy required to chew a food (10). High cohesiveness values indicate more energy was required for the second compression. CONCLUSIONS Based on the results of our study, a final high-speed mixing time of 3 min is suitable for producing high-quality cakes. These Fig. 7. Mean plots, with 95.0% LSD intervals, of mean values for hardness (A) and springiness (B) for each mixing time. Fig. 8. Mean plots, with 95.0% LSD intervals, of mean values for cohesiveness for each mixing time (A) and procedure (B). CEREAL FOODS WORLD / 291

cakes had the highest volume and greatest total cell area, as well as the lowest hardness values. When mixing time was increased, higher bubble occlusion occurred, but inadequate expansion was observed due to the higher viscosity of these batters. In general, there were no significant differences between the all-in and multistage mixing procedures for most of the parameters, including relative density, volume, total cell area, and hardness. Therefore, an all-in mixing procedure used in place of a multistage mixing procedure to improve the pilot-scale cake manufacturing process (i.e., time and costs) should provide satisfactory results regarding cake quality, even when the formulation does not incorporate emulsifiers. Of the cake batter mixing methods and times tested, the most suitable mixing procedure at a pilot scale was the all-in mixing procedure with a final 3 min mixing time. This mixing method produced the highest quality cakes and optimized energy costs and processing time. Because these results may have important implications for commercial production applications, further studies should be conducted to analyze the repercussions of using this mixing procedure in large-scale production. Acknowledgments We thank the Spanish Ministry of Science and Innovation (Project AGL2009-12785-C02-02) for their financial support and the Conselleria de Educación of Valencia Government for financing the contract and providing a supplementary grant for a stay of research for Julia Rodríguez- García. We also thank Campden BRI for access to its materials, methods, and laboratories, particularly Gary Tucker for his support and Paul Catteral for his guidance during the research and discussions on the results. References 1. Baixauli, R., Sanz, T., Salvador, A., and Fiszman, S. M. Muffins with resistant starch: Baking performance in relation to the rheological properties of the batter. J. Cereal Sci. 47:3, 2008. 2. Bath, D. E., Shelke, K., and Hoseney, R. C. Fat replacers in highratio layer cakes. Cereal Foods World 37:7, 1992. 3. Bennion, E. B., and Bamford, G. S. T. Baking fats. In: The Technology of Cake Making. A. J. Bent, ed. Blackie Academic and Professional, London, 1997. 4. Cauvain, S. P., and Young, L. S. Interactions between formulation and process methodologies. Page 120 in: Baked Products: Science, Technology and Practice. S. P. Cauvain and L. S. Young, eds. Blackwell Publishing, Ames, IA, 2006. 5. Conforti, F. D. Fundamentals of cakes: Ingredients and production. Page 307 in: Handbook of Food Products Manufacturing. Y. H. Hui, ed. John Wiley & Sons, Hoboken, NJ, 2006. 6. Conforti, F. D. Cake manufacture. Page 393 in: Bakery Products: Science and Technology. Y. H. Hui, ed. Blackwell Publishing, Ames, IA, 2006. 7. Lai, H. M., and Lin, T. C. Bakery products: Science and technology. Page 3 in: Bakery Products: Science and Technology. Y. H. Hui, ed. Blackwell Publishing, Ames, IA, 2006. 8. Loewe, R. Role of ingredients in batter systems. Cereal Foods World 38:9, 1993. 9. Sahi, S. S., and Alava, J. M. Functionality of emulsifiers in sponge cake production. J. Sci. Food Agric. 83:14, 2003. 10. Sanz, T., Salvador, A., Baixauli, R., and Fiszman, S. M. Evaluation of four types of resistant starch in muffins. II. Effects in texture, colour and consumer response. Eur. Food Res. Technol. 229:2, 2009. 11. Sumnu, G., Sahin, S., and Sevimli, M. Microwave, infrared and infrared-microwave combination baking of cakes. J. Food Eng. 71:2, 2005. An ad appeared here in the print version of the journal. 292 / NOVEMBER DECEMBER 2014, VOL. 59, NO. 6