Novel 2D and 3D Imaging of Internal Aerated Structure of Ultrasonically Treated Foams and Cakes Using X-Ray Tomography and X-Ray Microtomography

Accepted Manuscript

Novel 2D and 3D Imaging of Internal Aerated Structure of Ultrasonically Treated Foams and Cakes using X-ray Tomography and X-ray Microtomography

M.C. Tan, N.L. Chin, Y.A. Yusof, J. Abdullah

PII: S0260-8774(16)30084-X DOI: 10.1016/j.jfoodeng.2016.03.008 Reference: JFOE 8506

To appear in: Journal of Food Engineering

Received Date: 20 August 2015 Revised Date: 5 March 2016 Accepted Date: 20 March 2016

Please cite this article as: Tan, M.C., Chin, N.L., Yusof, Y.A., Abdullah, J., Novel 2D and 3D Imaging of Internal Aerated Structure of Ultrasonically Treated Foams and Cakes using X-ray Tomography and X- ray Microtomography, Journal of Food Engineering (2016), doi: 10.1016/j.jfoodeng.2016.03.008.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. ACCEPTED MANUSCRIPT

1 Novel 2D and 3D Imaging of Internal Aerated Structure of Ultrasonically Treated Foams 2 and Cakes using X-ray Tomography and X-ray Microtomography 3

4 1Tan, M.C., 1Chin, N.L.*, 1Yusof, Y.A., and 2Abdullah, J. 5 6 1Department of Process and Food Engineering, Faculty of Engineering, Universiti Putra 7 Malaysia, 43400 UPM Serdang, Selangor, Malaysia. 8 2Plant Assessment Technology Group, Malaysian Nuclear Agency (Nuclear Malaysia), Bangi, 9 43000 Kajang, Selangor, Malaysia. 10 11 *Corresponding author: Tel.: +603 89466353; fax: +603 89464440 12 Email address: [email protected] 13 14 15 16 17 18 19 20 21 22 23 MANUSCRIPT 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 ACCEPTED 40 41 42 43 44 45

ACCEPTED MANUSCRIPT

46 Abstract

47 The aerated structure of ultrasound treated foams and its resulted cake structures were examined 48 using X-ray tomography and X-ray microtomography, leading to highly contrasted two- 49 dimensional (2D) and three-dimensional (3D) images. Through these imaging techniques and 50 software approaches, the effect of ultrasound treatment on the bubble size distribution was 51 distinguished clearly. Microbubbles in foam which were in the size range of 0 to 0.00125 mm 3 52 and in cakes which were in the range of 0 to 1 mm 2 increased by 48% and 29% respectively after 53 ultrasonic treatment at a frequency of 20 kHz.

54 55 Keywords

56 X-ray microtomography; X-ray tomography; 2D image; 3D image; Aerated foam; Cake

57 58 1. Introduction

59 Imaging is one of the direct methods to visualize and analyze bubbles or cells matrix in food 60 products. Traditionally, bubble sizes in cake batter have been examined by taking photographs of 61 aerated batter spread out on a thin film using micr MANUSCRIPToscope-linked camera (Niranjan and Silva, 62 2008; Sahi and Alava, 2003) or tinted cross sections of cake samples at 5 mm thickness with 63 black oil-based ink delineating pore walls using a digital camera (Barrett and Ross, 1990; Kocer 64 et al., 2007; Lange et al., 1994; Smolarz et al., 1989). The photographs were analyzed using 65 imaging software. The three-dimensional structure was then drawn based on the assumed 66 symmetrical two-dimensional images. This imaging technique is a destructive method and it also 67 arises problems in generating a sharp image due to inaccurate measurement of the walls and 68 curvature effects (Niranjan and Sahu, 2009; Niranjan and Silva, 2007). As such, bubble size 69 distribution is difficult to be measured and analyzed (Niranjan and Silva, 2007), especially with 70 the traditional imaging method. The poorly understood aerated foods’ structures, however, did 71 not stop the evolutionaryACCEPTED food manufacturers from seeking to exploit the versatility of bubbles as 72 a food ingredient (Campbell and Mougeot, 1999). Air is an important ingredient in aerated foods 73 such as beverages, baked, dairy, egg and confectionery products to enhance the appearance, 74 texture, and digestibility of the food products (Campbell and Mougeot, 1999). For bakery 75 products, air attribution has a significant influence on the quality of the final product because the

ACCEPTED MANUSCRIPT

76 cells in the final products are originated as bubbles in the batter during the mixing stage, and that 77 integrity of air cells would determine the volume of the final products (Edoura-Gaena et al., 78 2007). 79 80 The rapid developments in technology have improved those traditional imaging methods into 81 more effective and accurate ones. Today, information about microstructure of food products and 82 ingredients can be obtained using various imaging techniques such as bright-field, polarizing and 83 fluorescence light microscopy, confocal scanning laser microscopy and electron microscopy. A 84 relatively new technique, X-ray tomography and more advanced X-ray microtomography can 85 probe the microstructure of samples non-invasively up to a few millimeters and down to few 86 micrometers of resolution, the advantages also include observations under environmental 87 conditions without sample-disturbing preparations which occur in the fluorescence light 88 microscopy and electron microscopy techniques (van Dalen et al., 2003). The X-ray tomography 89 and X-ray microtomography techniques have emerged as important and useful non-invasive 90 imaging tools to measure the internal microstructure of cellular food products. X-ray 91 microtomography has been attempted for use to generate the 3D images of porous rice kernels 92 and whipped cream (van Dalen et al., 2003), aeratedMANUSCRIPT chocolate, mousse, marshmallow and 93 muffins (Lim and Barigou, 2004), bread crumb (Falcone et al., 2005; Moussawi et al., 2014), 94 extruded rye bran (Alam et al., 2013), dough (Bellido et al., 2006), biscuit, breadstick, emulsion 95 and coffee (Laverse et al., 2012), biopolymer foams (Trater et al., 2005), ice crystal formation in 96 strawberry (Nurzahida et al., 2010), banana slices (Léonard et al., 2008), apple tissue (Mendoza 97 et al., 2007), mango during ripening (Cantre et al., 2014) and processed meat (Frisullo et al., 98 2009) while X-ray tomography has been used to generate images of extruded starch (Babin et al., 99 2007), bread crumbs (Lassoued et al., 2007), cornflakes (Chaunier et al., 2007) and fruits (Rogge 100 et al., 2013). 101 102 The objective ofACCEPTED this study is to investigate the effects of ultrasound treatment in protein 103 suspensions for making foams to be used in the baking of cake. The highly aerated foam and 104 baked cake structure were then assessed via imaging using X-ray microtomography and X-ray 105 tomography techniques. The collected X-ray microtomography and X-ray tomography images

ACCEPTED MANUSCRIPT

106 were then reconstructed into 2D and 3D images for a complete evaluation of the internal aerated 107 structures. 108 109 2. Materials and methods

110 2.1. Materials

111 The cake making process uses protein suspensions from protein powder, i.e. whey protein 112 concentrate (Textrion PROGEL 800, DMV International, BA Veghel, Netherlands) because it is 113 a protein with the highest nutritional value (García-Garibay et al., 2008) and contains all of the 114 essential amino acids in higher concentration compared to vegetable proteins such as soy, corn 115 and wheat gluten (Recio et al., 2008). The typical composition of the whey protein concentrate is 116 80% protein, 6.5% lactose, 6.3% fat, 4.4% ash and 4.9% moisture with 6.5 pH. 117 118 2.2. Cake preparation from ultrasonically treated foam

119 The aqueous suspensions of whey protein powder at 20% (w/w) concentration were prepared by 120 dispersing 50 g of dry matter into 200 g of distilled water in a 500 mL beaker and stirrer using a 121 mechanical stirrer (RW20 DZM.n S2, IKA Works MANUSCRIPT (Asia) Sendirian Berhad, Malaysia) at 355 122 rpm for 20 min until homogenous suspensions were obtained. The solution was sonicated with 123 20 kHz - 400 W high intensity ultrasound probe (Digital Sonifier Model 450, Branson 124 Ultrasonics Corporation, Danbury, Connecticut, USA) at 60% amplitude for 25 min. The control 125 sample has no ultrasound treatment and is known as an untreated sample. 250 g of the treated or 126 untreated whey protein suspension was whipped into foam at room temperature in a mixer 127 (5K5SSS, Kitchen Aid Inc., St. Joseph, Michigan, USA) at 330 rpm for 15 min. 128 129 The whipped foam was used for cake making following the Angel food cake recipe with an 130 adapted formulation given in Table 1 (Yamazaki and Lord, 1971). Sugar was added into the 131 foam during the extendedACCEPTED mixing of time 4 min. Mixing continued for another 4 min at a reduced 132 speed of 160 rpm when pre-mixed salt with flour was added to obtain cake batter. The speed of 133 mixing the cake batter was increased to 330 rpm during the final 15 s. 450 g of cake batter was 134 poured into the cake tin with dimensions showed in Fig. 1 and baked at 170°C with oven heat 135 levels setting of 20% (top): 30% (side): 10% (bottom) for 40 min in an electronic baking oven

ACCEPTED MANUSCRIPT

136 (ST-02, Salva Industrial, Spain). The baked cakes were inverted on a wire rack immediately and 137 cooled for 5 min before removing them from the tins. The cakes were left in ambient for 1 hour 138 for further cooling (Gómez et al., 2008; Tan et al., 2011) before cake imaging analyses. 139 140 2.3. Preparation of X-ray microtomography sample holder

141 Fig. 2a is the schematic diagram of the sample holder designed specifically for the loading of the 142 foam sample with minimal disturbance of its aerated structure. It was made with Perspex 143 (polymethyl metacrylate) and placed inside the chamber of the X-ray microtomography system. 144 The entire sample holder is separated into three parts, namely the cap, body and stand to ease 145 sample loading and for cleaning purposes. The cap and stand are sealed hermetically using O- 146 rings to tight up the closure. For sample loading, the body with 11 mm inner diameter and 1 mm 147 wall thickness was gently pressed into the aerated foam, then plugged on the stand and closed 148 with the cap as showed in Fig. 2b. 149 150 2.4. X-ray microtomography and X-ray tomography techniques 151 Fig. 3 shows the Skyscan 1172 desktop microtomograp MANUSCRIPThy system used for scanning images of 152 foam and cake samples. The X-ray microtomography system includes a chamber where it is kept 153 closed by a lead shield when the experiment is running to prevent X-ray beams escaping from 154 the chamber (Nurzahida et al., 2010). The foam sample loaded in the sample holder was fixed 155 securely to the right position in the chamber. For cake imaging, a small cake sample was cut into 156 a cube in the dimension of 1.5 cm x 1.5 cm x 1.5 cm and placed on the 1.5 cm diameter 157 specimen holder platform. The lead shield was closed and the scanning operation started after the 158 desired settings were set. The X-ray beams from the X-ray tube produced by its X-ray source 159 penetrated into the sample. The sample was rotated on its vertical axis by a rotational stage over 160 180° with a step size of 1° to provide different sampling directions. While the dataset is 161 massively under-sampled,ACCEPTED there are still insights about the structure which can be gained. The 162 images data of line integrals were collected from many angles by a charge-coupled device (CCD) 163 camera and conveyed to the computer for construction into a two dimensional image, known as 164 the tomography process. The settings for foam and cake samples were same, i.e. with the X-ray 165 tube set at a power of 10 W, 80 kV of voltage and 124 µA of current source. The total acquisition

ACCEPTED MANUSCRIPT

166 of images and read-out time was about 11 min for one projection. The image pixel size was set at 167 17.2 µm resolution and the camera pixel size used was 1280 x 1024 pixels with no filter. A set of 168 flat cross sections of 2D images was obtained after tomographical reconstruction by using the 169 NRecon software of the Skyscan system. 170 171 The concept of X-ray tomography is similar to the X-ray microtomography and it was used for 172 larger cake samples (Tan et al. 2014). The whole cake sample was placed on an adjustable 173 rotational stage within the X-ray source and detector with a 54 cm distance between the X-ray 174 source and cake sample and 58 cm distance between the cake sample and the detector. The cake 175 images were taken every 0.5° around a 360° path. The X-ray tube was set with 100 kV of voltage 176 and 2 mA of current source. The detector size was 230 x 1310 pixels, corresponding to a field of 177 view of 61 x 347 mm. Each cake was scanned at the top, center and bottom layers with 1.5 cm 178 from the center to the top or bottom. 179 180 2.5. Reconstruction of 2D and 3D images analysis of foam and cake samples 181 The collected images from the X-ray microtomographyMANUSCRIPT were reconstructed into 2D images by 182 NRecon software of Skyscan system and 3D images by using VGStudio MAX 2.2 software. The 183 region desired for volume reconstruct can be set by keying in the range of top and bottom slices 184 or by dragging the lines on the image stack displayed at the start page. A middle slice of the 185 image stack was chosen as a previewing slice together with the data displayed in a histogram. It 186 is important to choose a slice which goes through the dense parts of the object to avoid 187 truncation of higher values in the final image. The dynamic range of the displayed histogram in 188 the output page in NRecon software was set as 0.01 to 0.08 to obtain a contrast image. 189 190 VGStudio MAX is the high-end software from the Volume Graphics range of products for the 191 visualization and analysis of computer tomography data and it has become a virtual standard for 192 industrial voxel dataACCEPTED analysis (Anonymous, 2013). The 2D image stack reconstructed by NRecon 193 software was imported through the directory in VGStudio MAX 2.2 software and mapped to 194 unsigned 8 bit in ramp mode with X, Y and Z in resolution of 17.2 µm. The region of interest to 195 reconstruct the 3D image was tuned from the top, right and front projections in the preview 196 window. The contrast in the image reconstructed was then adjusted by tuning in the opacity

ACCEPTED MANUSCRIPT

197 manipulation area, and the image was set to display in rendering setting of volume renderer 198 (Scatter HQ). The 2D and 3D images were constructed in a similar way for both the foam and 199 cake. For whole cake imaging, the X-ray tomography images of the cake at the top, center and 200 bottom layers were constructed as 2D images using the Backprojection software (Tan et al., 201 2014). 202 203 An analysis of bubbles size distribution of foam was performed on the constructed 3D images 204 from x-ray microtomography using voids analysis with default algorithm and defect analysis tool 205 by the VGStudio MAX 2.2 software. The probability threshold was set at 0.5. The number of 206 bubbles at three different size ranges, 0 to 0.0025 mm 3, 0.00251 to 0.005 mm 3 and 0.0051 to 0.01 207 mm 3 was averaged from three randomly selected areas of 155 x 155 x 155 voxels in each foam 208 sample. An analysis of bubbles size distribution in the cake was performed on the constructed 2D 209 images from x-ray tomography using ImageJ software. For each layer, the number of bubbles at 210 different size ranges, from 0 to 1 mm 2, consecutively ranged at every 1 mm 2 until 10 mm 2 and 211 above was analyzed at each quarter at 200 x 200 pixels. The average was taken from three layers 212 of cake. A statistical comparison of analysis of variance and Bonferroni correction post-hoc test 213 on the number of bubbles at every size range due to MANUSCRIPT ultrasound effect was performed. 214 215 3. Results and discussion

216 3.1. Aerated foam structure

217 The X-ray microtomography technique was used to examine the changes in the structure of 218 bubbles in the ultrasonically treated foam. A layer of reconstructed 2D image is presented in Fig. 219 4a, while Fig. 4b shows the constructed 3D images. It is obvious that the bubbles in the foam 220 formed by treated whey protein suspension in Figs. 4-a2 and 4-b2 are smaller in size and more 221 evenly distributed than the untreated whey protein foam in Figs. 4-a1 and 4-b1. 222 ACCEPTED 223 An analysis of the 3D images of foam using the defect detection analysis shows that the overall 224 total number of bubbles in the size range of 0 to 0.01 mm 3 has increased 45% averagely from 225 1296 to 1883 after treatment with P < 0.0005 (Fig. 5a). Zooming into that range, bubbles in the 226 size range of 0 to 0.00125 mm 3 was found to be the biggest population, which occupied 89% and

ACCEPTED MANUSCRIPT

227 91% of the total bubbles respectively before and after treatment (Fig. 5b). The number of this 228 biggest population of bubbles has increased 48% after treatment with P < 0.0005, while the 229 number of bigger bubbles in size ranges of 0.001251 to 0.0025 mm 3, 0.00251 to 0.005 mm 3 and 230 0.0051 to 0.01 mm 3 has increased by 17%, 23% and 35% respectively, with P < 0.5. This shows 231 that ultrasound has increased the number of microbubbles in whipped whey protein foam, where 232 Bonferroni correction post-hoc test also indicated that the number of bubbles in size range 0 to 233 0.01 mm 3 after treatment were significantly higher than control. The cavitation effect of 234 ultrasonication which causes implosion of the bigger bubbles in whey protein suspension 235 (Suslick, 1988; Soria and Villamiel, 2010) to implode giving a higher number of microbubbles. 236 237 3.2. Aerated cake structure

238 The images of the top, center and bottom layers of an uncut cake scanned using X-ray 239 tomography are shown in Fig. 6 while Fig. 7 are 2D and 3D images scanned using X-ray 240 microtomography from a smaller cube sample. From both figures, which show the whole cake 241 and a section of a cake, it is clear that the bubbles in the control cake are not as evenly 242 distributed as the treated. The air cells seem to coalesceMANUSCRIPT and form bigger air cells. The structure 243 of the control cake is also coarser with ring defec ts throughout the cake layers observed by its 244 brighter color in Fig. 6a. This suggests that ultrasonic treatment on protein suspension has helped 245 to minimize and distribute the bubbles more evenly. Fig. 8 shows the analysis of bubbles in 246 terms of cumulative and frequency distribution at various size ranges. The overall total number 247 of bubbles in the size range of 0 to 10 mm 2 has increased 32% from 1641 to 2196 after treatment 248 with P < 0.005 (Fig. 8a). The bubbles in the size range of 0 to 1 mm 2 are the largest in population, 249 which occupied 65% and 63% of the total bubbles, respectively before and after treatment (Fig. 250 8b). The average number of this group of bubbles has increased 29% after treatment with P < 251 0.05, while the average number of bubbles in other size ranges of 1.001 to 2 mm 2, 2.001 to 3 252 mm 2 and 3.001 to 4 mm 2 has increased 69% (P < 0.005), 51% (P < 0.05) and 48% (P < 0.05) 253 respectively. TheACCEPTED number of bubbles with area larger than 10 mm 2 has decreased 49% after 254 treatment with P < 0.1. The Bonferroni correction post-hoc test indicated that ultrasound 255 treatment significantly increased the number of bubbles in size ranges of 0 to 10 mm 2, 0 to 1 256 mm 2 and 1.001 to 2 mm 2 when compared with untreated sample. 257

ACCEPTED MANUSCRIPT

258 4. Conclusion

259 X-ray microtomography can visualize and differentiate the effects of ultrasound treated whey 260 protein foams and cakes made from it by its aerated structure in terms of bubbles attributes, 261 while X-ray tomography visualized the whole cake structure non-destructively. With the defect 262 detection analysis of VGStudio MAX 2.2 and ImageJ software, both tomography techniques 263 have quantitatively indicated that ultrasound treatment has increased the number of microbubbles 264 in foam and cake systems with more even distributions. 265 266 Acknowledgements

267 This work was supported by the Ministry of Education (MOE), Malaysia, through the 268 Fundamental Research Grant Scheme (FRGS/2/2013/TK05/UPM/02/5). The authors would like 269 to thank Alex Corporation (M) Sendirian Berhad, Selangor (Malaysia) for providing VGStudio 270 MAX 2.2 software for this research. 271 272 References 273 Alam, S.A., Järvinen, J., Kirjoranta, S., Jouppila,MANUSCRIPT K., Poutanen, K., & Sozer, N. (2013). 274 Influence of particle size reduction on structural and mechanical properties of extruded rye 275 bran. Food and Bioprocess Technology , 1-13. 276 Anonymous. (2013). Volume Graphics. http://www.volumegraphics.com/en/products/vgstudio- 277 max.html. Accessed 23 October 2013. 278 Babin, P., Della Valle, G., Dendievel, R., Lourdin, D., & Salvo, L. (2007). X-ray tomography 279 study of the cellular structure of extruded starches and its relations with expansion 280 phenomenon and foam mechanical properties. Carbohydrate Polymers , 68(2), 329-340. 281 Barrett, A.H., & Ross, E.W. (1990). Correlation of extrudate infusibility with bulk properties 282 using image analysis. Journal of Food Science , 55(5), 1378-1379. 283 Bellido, G.G., Scanlon, M.G., Page, J.H., & Hallgrimsson, B. (2006). The bubble size 284 distribution in wheat flour dough. Food Research International , 39(10), 1058-1066. 285 Campbell, G.M., & Mougeot, E. (1999). Creation and characterisation of aerated food products. 286 Trends in Food Science & Technology , 10(9), 283-296. 287 Cantre, D., Herremans,ACCEPTED E., Verboven, P., Ampofo-Asiama, J., & Nicolai, B.M. (2014). 288 Characterization of the 3-D microstructure of mango ( Mangifera indica L. cv. Carabao) 289 during ripening using X-ray computed microtomography. Innovative Food Science and 290 Emerging Technologies . http://dx.doi.org/10.1016/j.ifset.2013.12.008 291 Chaunier, L., Della Valle, G., & Lourdin, D. (2007). Relationships between texture, mechanical 292 properties and structure of cornflakes. Food Research International , 40(4), 493-503.

ACCEPTED MANUSCRIPT

293 Edoura-Gaena, R.B., Allais, I., Trystram, G., & Gros, J.B. (2007). Influence of aeration 294 conditions on physical and sensory properties of aerated cake batter and biscuits. Journal of 295 Food Engineering , 79(3), 1020-1032. 296 Falcone, P.M., Baiano, A., Zanini, F., Mancini, L., Tromba, G., Dreossi, D., & Nobile, M.A.D. 297 (2005). Three-dimensional quantitative analysis of bread crumb by X-ray microtomography. 298 Journal of Food Science , 70(4), E265-E272. 299 Frisullo, P., Laverse, J., Marino, R., & Nobile, M.D. (2009). X-ray computed tomography to 300 study processed meat microstructure. Journal of Food Engineering , 94(3), 283-289. 301 García-Garibay, M., Jiménez-Guzmán, J., & Hernández-Sánchez, H. (2008). Whey proteins: 302 Bioengineering and health. In G.F. Gutierrez-Lopez, G.V. Barbosa-Canovas, J. Welti-Chanes, 303 & E. Parada-Arias (Eds.), Food engineering: Integrated approaches (pp. 415-430). New 304 York, USA: Springer Science + Business Media, LCC. 305 Gómez, M., Oliete, B., Rosell, C.M., Pando, V., & Fernández, E. (2008). Studies on cake quality 306 made of wheat-chickpea flour blends. LWT – Food Science and Technology , 41(9), 1701- 307 1709. 308 Kocer, D., Hicsasmaz, Z., Bayindirli, A., & Katnas, S. (2007). Bubble and pore formation of the 309 high-ratio cake formulation with polydextrose as a sugar- and fat-replacer. Journal of Food 310 Engineering , 78(3), 953-964. 311 Lange, D.A., Jennings, H.M., & Shah, S.P. (1994). Image analysis techniques for 312 characterization of pore structure of cement-based materials. Cement and Concrete Research , 313 24(5), 841-853. 314 Lassoued, N., Babin, P., Della Valle, G., Devaux, MMANUSCRIPT.F., & Réguerre, A.L. (2007). Granulometry 315 of bread crumb grain: contributions of 2D and 3D image analysis at different scale. Food 316 Research International , 40(8), 1087-1097. 317 Laverse, J. Frisullo, P., Conte, A., & Nobile, M.A. (2012). X-ray microtomography for food 318 quality analysis. Food Industrial Processes-Methods and Equipment , 339-362. 319 Léonard, A., Blacher, S., Nimmol, C., & Devahastin, S. (2008). Effect of far-infrared radiation 320 assisted drying on microstructure of banana slices: an illustrative use of X-ray 321 microtomography in microstructural evaluation of a food product. Journal of Food 322 Engineering , 85(1), 154-162. 323 Lim, K.S., & Barigou, M. (2004). X-ray micro-computed tomography of cellular food products. 324 Food Research International , 37(10), 1001-1012. 325 Mendoza, F., Verboven, P., Mebatsion, H.K., Kerckhofs, G., Wevers, M., & Nicolai, B. (2007). 326 Three-dimensional pore space quantification of apple tissue using X-ray computed 327 microtomography. Planta , 226(3), 559-570. 328 Moussawi, A., Xu,ACCEPTED J., Nouri, H., Guessasma, S., & Lubineau, G. (2014). Volume digital image 329 correlation to assess displacement field in compression loaded bread crumb under X-ray 330 microtomography. Innovative Food Science and Emerging Technologies . 331 http://dx.doi.org/10.1016/j.ifset.2014.02.002

ACCEPTED MANUSCRIPT

332 Niranjan, K., & Silva, S.F.J. (2007). Bubble-containing foods. In J.M Aguilera, & P.J. Lillford 333 (Eds.), Food materials science: Principle and practice (pp. 281-303). New York, USA: 334 Springer Science + Business Media, LCC. 335 Niranjan, K., & Silva, S.F.J. (2008). Bubbles in foods: Creating structure out of thin air. In G.F. 336 Gutierrez-Lopez, G.V. Barbosa-Canovas, J. Welti-Chanes, & E. Parada-Arias (Eds.), Food 337 engineering: Integrated approaches (pp. 183- 192). New York, USA: Springer Science + 338 Business Media, LCC. 339 Niranjan, K., & Sahu, J.K. (2009). Gas-liquid mixing. In P. J. Cullen (Ed.), Food mixing: 340 Principle and applications (pp. 230-252). USA: Wiley-Blackwell. 341 Nurzahida, M.Z., Mimi Sakinah, A.M., & Abdullah, J. (2010). Influence of thickness on ice 342 crystal formation in strawberry during freeze-drying. Journal of Applied Sciences , 10(21), 343 2741-2744. 344 Recio, I., Ramos, M., & Pilosof, A.M.R. (2008). Engineered food/ protein structure and bioactive 345 proteins and peptides from whey. In G.F. Gutierrez-Lopez, G.V. Barbosa-Canovas, J. Welti- 346 Chanes, & E. Parada-Arias (Eds.), Food engineering: Integrated approaches (pp. 399-414). 347 New York, USA: Spinger Science + Business Media, LCC. 348 Rogge, S., Beyene, S.D., Herremans, E., Hertog, M.L., Defraeye, T., Verboven, P., & Nicolai, 349 B.M. (2013). A geometrical model generator for quasi-axisymmetric biological products. 350 Food and Bioprocess Technology , 7(6), 1783-1792. 351 Sahi, S.S., & Alava, J.M. (2003). Functionality of emulsifiers in sponge cake production. 352 Journal of the Science of Food and Agriculture , 83(14), 1419-1429. 353 Smolarz, A, Van Hecke, E., & Bouvier, J.M. (1989).MANUSCRIPT Computerized image analysis and texture 354 of extruded biscuits. Journal of Texture Studies , 20(2), 223-234. 355 Soria, A.C., & Villamiel, M. (2010). Effect of ultrasound on the technological properties and 356 bioactivity of food: a review. Trends in Food Science & Technology , 21(7), 323-331. 357 Suslick, K. (1988). Ultrasound: Its chemical, physical and biological effects . New York: VCH 358 Publishers. 359 Tan, M.C., Chin, N.L., & Yusof, Y.A. (2011). Power ultrasound aided batter mixing for sponge 360 cake batter. Journal of Food Engineering , 104(3), 430-437. 361 Tan, M.C., Chin, N.L., Yusof, Y.A., Taip, F.S., & Abdullah, J. (2014). Improvement of eggless 362 cake structure using ultrasonically treated whey protein. Food and Bioprocess Technology , 1- 363 10. 364 Trater, A.M., Alavi, S., & Rizvi, S.S.H. (2005). Use of non-invasive X-ray microtomography for 365 characterizing microstructure of extruded biopolymer foams. Food Research International , 366 38(6), 709-719. 367 van Dalen, G., Blonk,ACCEPTED H., van Aalst, H., & Hendriks, C.L. (2003). 3-D imaging of foods using 368 X-ray microtomograpghy. GIT Imaging and Microscopy , 3, 18-21. 369 Yamazaki, W.T., & Lord, D.D. (1971). Soft wheat flour products. In Y. Pomeranz (Ed.), Wheat 370 chemistry technology (pp. 743-776). USA: American Association of Cereal Chemists, 371 Incorporated. 372

ACCEPTED MANUSCRIPT

Figure Captions

Fig. 1. Schematic diagram of hollow baking tin with diameters given at the top, center and bottom sections.

Fig. 2. (a) Schematic diagram of X-ray microtomography sample holder and (b) Sample holder with foam filled-in.

Fig. 3. Schematic diagram of X-ray microtomography system.

Fig. 4. (a) 2D and (b) 3D microtomography images of foam whipped with (1) untreated and (2) treated whey protein suspension.

Fig. 5. (a) Cumulative distribution and (b) frequency distribution of bubbles at various size ranges in ultrasonically treated and untreated aerated whey protein foams.

Fig. 6. 2D tomography images of cake baked with (a) untreated and (b) treated whey protein suspension at the (1) top, (2) center, and (3) bottom layers.

Fig. 7. (a) 2D and (b) 3D microtomography images of cake baked with (1) untreated and (2) treated whey protein suspension.

Fig. 8. (a) Cumulative distribution and (b) frequency distribution of bubbles at various size ranges in cakes formulated by whey protein foam with and without ultrasound. MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT

Table 1 Angel food cake formulations. Ingredients Mass of cake loading (g) Flour 100 Sugar 250 Salt 3 Protein suspension 250 Total Mass 603

MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT

Fig. 1. Schematic diagram of hollow baking tin with diameters given at the top, center and bottom sections.

MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT

Cap

b Cap

O-ring

Body

Stand

Stand Body MANUSCRIPT

Fig. 2. (a) Schematic diagram of X-ray microtomography sample holder and (b) Sample holder with foam filled-in.

ACCEPTED ACCEPTED MANUSCRIPT

Chamber X-ray CCD camera

Sample holder Computer X-ray tube X-ray beam

Rotation stage

Controller

Fig. 3. Schematic diagram of X-ray microtomography system.

MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT

a1 a2

b1 b2

MANUSCRIPT

Fig. 4. (a) 2D and (b) 3D microtomography images of foam whipped with (1) untreated and (2) treated whey protein suspension.

ACCEPTED ACCEPTED MANUSCRIPT

120 (a) 100

40 Control Treatment 20 Cumulative Cumulative distribution(%)

0 0 0.002 0.004 0.006 0.008 0.01 Bubbles volume (mm 3)

2000 (b) 1800 1600 1400 MANUSCRIPT 1200 Control 1000 Treatment

Frequency 800 600 400 200 0 0 0.00125 0.00125 0.0025 0.0025 0.005 0.005 0.01 0.01 Bubbles Volume (mm 3)

Fig. 5. (a) CumulativeACCEPTED distribution and (b) frequency distribution of bubbles at various size ranges in ultrasonically treated and untreated aerated whey protein foams. ACCEPTED MANUSCRIPT

a1 a2 a3

b1 b2b2 b3 b3

Fig. 6. 2D tomography images of cake baked with (a) untreated and (b) treated whey protein suspension at the (1) top, (2) center, and (3) bottom MANUSCRIPTlayers.

ACCEPTED ACCEPTED MANUSCRIPT

a1 a2

b1 b2

MANUSCRIPT

Fig. 7. (a) 2D and (b) 3D microtomography images of cake baked with (1) untreated and (2) treated whey protein suspension.

ACCEPTED ACCEPTED MANUSCRIPT

120 (a)

100

40 Control Treatment 20 Cumulative Cumulative distribution(%)

0 0 1 2 3 4 5 6 7 8 910 2 Bubbles area (mm )

1600 (b) 1400 1200 MANUSCRIPT 1000 Control Treatment 800

600 Frequency 400

200

0 0 0 1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 910 10 >10 Bubbles area (mm 2)

Fig. 8. (a) Cumulative distribution and (b) frequency distribution of bubbles at various size ranges in cakes formulatedACCEPTED by whey protein foam with and without ultrasound. ACCEPTED MANUSCRIPT

Highlights

• Nondestructive X-ray tomography and X-ray microtomography imaging techniques provided constructed 2D and 3D images of highly aerated foams and baked cakes. • The ultrasound treatment was helpful in increasing the number of microbubbles with more even distribution.

MANUSCRIPT

ACCEPTED