- Email: [email protected]
Visualization of oil migration in chocolate using scanning electron microscopy–energy dispersive X-ray spectroscopy
Food Structure 8 (2016) 8–15
Contents lists available at ScienceDirect
Food Structure journal homepage: www.elsevier.com/locate/foostr
Visualization of oil migration in chocolate using scanning electron microscopy–energy dispersive X-ray spectroscopy Hironori Hondoh* , Kenta Yamasaki, Miharu Ikutake, Satoru Ueno Graduate School of Biosphere Science, Hiroshima University, 1-4-4, Kagamiyama, Higashi-Hiroshima, Hiroshima 739-8528, Japan
A R T I C L E I N F O
Article history: Received 10 June 2015 Received in revised form 10 March 2016 Accepted 1 April 2016 Keywords: Chocolate Oil migration Energy dispersive X-ray spectroscopy SEM Silicone oil Chocolate structure
A B S T R A C T
We have developed a novel method to observe oil migration in chocolate using scanning electron microscopy and energy dispersive X-ray spectroscopy (EDX). If silicone oil is used as the mobile phase in migration, the amount of migrated silicone oil can be evaluated using the EDX signal. Compound chocolate placed on the silicone-oiled cotton resulted in the appearance of liquid oil droplets on the surface of the chocolate within 10 days at 24 C. Fat blooming was also induced by these oil droplets at 2.5 months. The amount of silicone oil that migrated into the chocolate was evaluated based on the weight change of the chocolate and the EDX signal from the silicone oil. Quantitative analysis indicated that two independent migration mechanisms, diffusion and capillary force, induced migration. In addition, the EDX signal indicated superfast migration from a spot that was distinct from the area of diffusion or capillary force migration. The EDX signals from phosphorus or magnesium atoms were helpful in identifying cacao particles in chocolate since these elements were specific to the particles in chocolate. Homogeneous distribution of silicone oil was observed around the particles in chocolate, thus, these particles might not help to produce a specific migration pathway in chocolate. ã 2016 Elsevier Ltd. All rights reserved.
1. Introduction Fat bloom, caused by the formation of needle-shaped fat crystals, is a significant problem in the confectionery industry. It is a greyish haze which forms on the surface of chocolate during storage and is promoted by storage near the melting point, temperature fluctuation, and oil migration (Lonchampt & Hartel, 2004; Rousseau, 2006). Customers dislike the visual appearance and textural attributes of bloomed chocolate, and this deterioration results in decreased consumer confidence. When chocolate is stored at a temperature above the melting point of form V crystals, melt-mediated transformation leads to the development of form VI crystals which grow to cause fat blooming. Temperature fluctuation causes liquid oil to pump out from the inside of the chocolate to the surface, even at a low temperature (Sonwai & Rousseau, 2010). The pumped liquid oil dissolves the surface of the chocolate and promotes recrystallization for blooming. These two issues are promoted by temperature change, however, fat bloom induced by oil migration occurs independently of temperature. If chocolate
Abbreviations: TAG, triacylglycerol; EDX, energy dispersive X-ray spectroscopy; SEM, scanning electron microscopy. * Corresponding author. E-mail address: [email protected] (H. Hondoh). http://dx.doi.org/10.1016/j.foostr.2016.04.001 2213-3291/ ã 2016 Elsevier Ltd. All rights reserved.
comes into contact with other ingredients containing liquid oil, the liquid oils in chocolate and the other ingredients migrate and mix together. Thus, migration promotes the softening of filled chocolate, the hardening of fillings, and fat bloom. Therefore, oil migration is a serious problem in praline chocolates, chocolatecovered cookies and chocolates with nuts. Oil migration is thought to occur by diffusion or capillary flow in chocolate to the surface (Ziegler, 2009). The driving force for diffusion is assumed to be the difference in chemical components: the triacylglycerol (TAG) concentration gradient in domains of confectionery chocolate products. If the liquid fraction of oil in a filling has a different TAG composition than cocoa butter, oil migration occurs. This diffusion has been the preferred explanation for oil migration in chocolate products (Aguilera, Michel, & Mayer, 2004; Altimiras, Pyle, & Bouchon, 2006; Ghosh, Ziegler, & Anantheswaran, 2002; Miquel, Carli, Couzens, Wille, & Hall, 2001). The possibility of capillary flow has also been studied: pores exist inside chocolate because of volume shrinkage during crystallization and the volume of this vacancy has been calculated to be 1% of chocolate (Loisel, Lecq, Ponchel, Keller, & Ollivon, 1997). Mass transportation occurs through these pores via capillary force. The contribution from these two mechanisms is difficult to distinguish because they exhibit similar time dependence (Aguilera et al., 2004). In addition, another migration mechanism,
H. Hondoh et al. / Food Structure 8 (2016) 8–15
pressure-driven convection flow, has been proposed (Altimiras et al., 2006; Dahlenborg, Millqvist-Fureby, Bergenståhl, & Kalnin, 2011; Dahlenborg, Millqvist-Fureby, Brandner, & Bergenståhl, 2012; Dahlenborg, Millqvist-Fureby, & Bergenståhl, 2015a). Numerous approaches have been utilized in the quantitative investigation of oil migration. Quantification of fat composition using chromatography is a common technique in the analysis of oil migration (Talbot, 1996; Ziegleder, Moser, & Geiwe-Greguska, 1996a; Ziegleder, Moser & Geiwe-Greguska, 1996b), however, spatial resolution is low because the samples must be divided for this analysis. Magnetic resonance imaging has been used to observe the inner structure of chocolate and to monitor oil migration in chocolate (Duce, Carpenter, & Hall, 1990; Guiheneuf, Couzens, Wille, & Hall, 1997). Magnetic resonance imaging is a nondestructive quantitative analytical method and thus is advantageous for oil migration investigation. Using this method, kinetic analysis of oil migration has made much progress, and has been used to determine the diffusion coefficient in model confectionery products (Guiheneuf et al., 1997; Lee, McCarthy, & McCarthy, 2010; McCarthy and McCarthy, 2008; Miquel et al., 2001; Rumsey & McCarthy, 2012). A flat-bed scanner combined with dye stain is a convenient technique for observing oil migration (Marty, Baker, Dibildox-Alvarado, Rodrigues, & Marangoni, 2005; Marty, Baker, & Marangoni, 2009). These techniques provide quantitative information on oil migration kinetics. A combination of scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDX) is a novel approach in the investigation of migration and blooming in chocolate. EDX is an analytical technique to identify elemental composition of a sample by measuring energy of characteristic X-ray from a specimen. Kinta and Hatta (2005, 2007) were the first to report on SEM–EDX in chocolate analysis (Kinta & Hatta, 2005, 2007). Using brominated vegetable oil as a die in SEM–EDX, Dahlenborg, Millqvist-Fureby, and Bergenståhl (2015b) found that the size of particles in chocolate affects migration velocity, and significantly increased the developing rate of fat bloom. They found that smaller particles of cocoa powder and sugar resulted in increased migration. A theoretical approach calculated that the velocity of capillary force migration was considerably faster than the migration rate obtained in the experimental results (Altimiras et al., 2006). In addition, Reinke et al. (2015) used small angle X-ray scattering to report that oil migration through pores occurred within seconds. These findings may support the significance of capillary force in oil migration. However, the oil migration pathway has not been observed directly, and details of oil migration remain obscure. The aim of this study was to demonstrate a novel approach to quantitative analysis of oil migration using SEM–EDX. Silicone oil was used as the mobile phase in this study and its distribution was visualized using EDX. Silicone oils have advantageous physical properties, such as high thermal and oxidation stability, and low viscosity and surface tension. The surface tension of silicone oils is lower than that of other natural oils, so that they spread easily over an interface, and are widely used as deformers and release agents in the industrial field. In the present study, the significant property of silicone oils was low solubility in cocoa butter: silicone oil and cocoa butter are immiscible therefore the diffusion of silicone oil
9
into chocolate is negligible. Using silicone oil as the mobile phase of oil migration in chocolate, we investigated the significance of capillary force and diffusion in oil migration. 2. Materials and methods 2.1. Materials Commercial compound milk chocolate bars (Meiji Co., Ltd., Japan) and canola oil (The Nisshin Oillio Group, Ltd., Japan) were purchased at a local market. The chocolate contained sugar, cocoa powder, vegetable oil, whole milk powder, cocoa butter, and soy lecithin, in descending order of content. Silicone oil, Element14* PDMS 10-JC was obtained from Momentive Performance Materials Inc. (Japan). The physical properties of the silicone and canola oils are presented in Table 1. 2.2. Sample preparation For the migration experiments, a chocolate bar was broken into pieces (32 26 mm2) using a sharp knife. A piece of chocolate was placed on a cotton pad (also 32 26 mm2) impregnated with silicone oil or canola oil in a polystyrene case. The volume of the oils was 600 ml, the weight of the canola oil was 552 mg, and that of the silicone oil was 562 mg. The samples were incubated at 24 C in an air incubator. To confirm the migration ability of silicone oil in the relatively high liquid oil environment of chocolate, the chocolate samples were stored at 24 C. Stored chocolates were used for further experiments. 2.3. Weight change and hardness measurements Weight change during storage was measured. The cotton pad was removed from the chocolate, and the weight of the chocolate was measured without wiping oil from the contact face. The measured samples were returned to the cotton pad in an incubator and used for further weight measurements. Hardness was measured using a penetrometer (EX-210E, Elex Science Co., Ltd., Japan) with a penetration rod that had a base diameter of 64.91 mm, a cone height of 29.08 mm, and a cone weight of 102.51 g. A penetration test was performed on the oil contacting face of the chocolate samples. These measurements were repeated at least five times. 2.4. SEM and EDX observations The central part of the stored chocolate was cut out in a block shape (3 3 4 mm3) using a sharp knife. When the tip of the knife was inserted into the chocolate, the chocolate was loosened and a piece of chocolate was released. The fragment was carefully trimmed to become a block shape. The surface of the chocolate block was coated with Au using a sputter coater (E-1010, Hitachi Instruments, Japan), and the cross-section was observed using SEM (S-2380, Hitachi Instruments, Japan). The acceleration voltage was 15 kV. To avoid artifacts in the SEM image, the area of chocolate where the knife had made contactwas not observed. Element
Table 1 Physical properties of silicone and canola oils.
Element14 PDMS 10-JC Canola oil
Dynamic viscosity (mm2/s)
Density (g/cm3)
Surface tension (mN/m)
Pour point ( C)
10 (25 C) 50.91 (37.8 C)
0.935 (25 C) 0.906 0.920 (25 C)
20.1 (25 C) 31.7 (20 C)
Below 60 Below 18
10
H. Hondoh et al. / Food Structure 8 (2016) 8–15
mapping was performed using EDX (E-MAX-7000S, Horiba, Japan) in the same field of view as the SEM. These measurements were carried out at room temperature. The EDX images were analyzed using Image-J (Rasband, 1997) to obtain the integrated intensity of EDX signal. Curve fitting was performed using Igor Pro (WaveMetrics, USA). The fitting was evaluated by x2 value. 2.5. Statistical analysis The results of the penetration test were statistically tested using two-way ANOVA and the Tukey-Kramer test. Significant differences (p < 0.05) were evaluated using Microsoft Excel. The experiments were repeated 5 times. 3. Results and discussion 3.1. Changes in weight, hardness and appearance We measured changes in the weight of compound chocolates with oil migration throughout the storage time to determine whether silicone and canola oils migrated into the chocolates. Fig. 1 indicates weight changes in chocolates stored with silicone and canola oils at 24 C. The weight of both chocolate samples clearly increased during the storage time, confirming that silicone and canola oils migrated into the chocolates. In addition, the weightincreasing rate of the silicone oil-chocolate was almost twice that of the canola oil-chocolate. After 20 days of storage, 38% of the applied silicone oil had migrated into the chocolate, compared with 22% of the canola oil, if counter diffusion was ignored. This result implied that silicone oil migration is faster than canola oil migration. The lower viscosity of silicone oil could be the reason for faster migration, or counter diffusion was significant in the canola oil-chocolate. The results of the penetration test are presented in Table 2. There was no significant difference in the hardness of the bare compound chocolate compared with the silicone oilchocolate after storage for 20 days at 24 C. In contrast, the penetration depth of the canola oil-chocolate was deeper from day 1 compared with the reference sample, and the depth was increased until day 10. Cocoa butter crystals are soluble in vegetable oils, such as canola oil, and chocolate softens by dissolving in vegetable oils. In contrast, the low solubility of cocoa butter in silicone oil did not affect the hardness of the chocolate. Fig. 2 depicts appearance changes in the chocolate samples during storage. The appearance of the reference compound chocolate did not change during 2.5 months of storage (Fig. 2(a, b)). After 9 days storage of the silicone oil-chocolate, liquid droplets appeared on the surface (Fig. 2 (c)) and the area of liquid increased until the chocolate became soaking wet within 4 weeks
Fig. 1. Weight gain of chocolate with silicone oil (squares) and canola oil (diamonds).
Table 2 Penetration depth (mean SD in 1/10 mm) in chocolate samples. Mobile phase
Day 1
Day 4
Day 10
Day 20
Reference Silicone oil Canola oil
7.3 0.7Aa 11.1 1.7Bab 14 2Cb
7.9 1.4Aa 11.2 1.6Ba 18 6CDb
7.8 1.9Aa 11.1 1.1Ba 22 4Db
7 3Aa 10.4 1.6Ba 18 4CDb
ab Same superscript indicates no significant difference (p > 0.05) in a column. ABCD Same superscript indicates no significant difference (p > 0.05) in a row.
(Fig. 2 (d)). At the end of the 2.5 months storage period, the surface of the silicone oil-chocolate bloomed and whitened (Fig. 2 (e)). In contrast, liquid droplets did not appear on the surface of the canola oil-chocolate for the entire storage period. The chocolate with canola oil bloomed after 4 weeks, and the bloom spread the whole surface after 2.5 months in storage (Fig. 2 (g, h)). An aliquot of the liquid droplets on the surface of the silicone oil-chocolate was taken and cooled to 5 C, and then heated until crystals were disappeared on a temperature-controllable stage (THMS500, Linkam Scientific Instruments, UK) under polarized optical microscopy. The liquid crystallized at 8 C while cooling, and the crystals melted and disappeared at 11 C while heating (Fig. 3). The pour point of silicone oil is below 60 C, so silicone oil does not crystallize at this temperature, therefore, we concluded that the liquid which had collected on the surface of the compound chocolate with silicone oil was a low melting point fraction of cocoa butter or vegetable oil from the chocolate. Because the solubility of cocoa butter and vegetable oil in silicone oil is quite low, silicone oil seldom diffuses in the liquid fraction of cocoa butter. Consequently, the liquid oil fraction in the chocolate is pumped out from the chocolate by migrating silicone oil. This is clear evidence that the liquid fraction of cocoa butter is pumped out via silicone oil migration. 3.2. SEM and EDX images SEM and EDX images of the chocolate cross-section are presented in Fig. 4. The edge of the cut end of chocolate sample was flattened where the blade touched it, however, most of the section had a rough surface indicating that the chocolate was cleaved and the microstructure of the chocolate was maintained. Spherical voids (air bubbles) were preserved, however, cocoa butter crystals, sugar and cacao particles could not be distinguished in the obtained SEM images. Comparison of the SEM and EDX images of cross-sections indicated that EDX intensity clearly depended on the topography of the chocolate. Thus, the absolute EDX intensity from the specimen in this analysis was not directly linked to the concentration of the specimen. In general, EDX intensity from the specimen is affected by the surface topography of the samples because the detection rate of the emitted X-ray to the detector depends on the surface topography. Thus, EDX intensity from the carbon on the EDX image depended on the undulation on the chocolate surface (Fig. 4). Identification of the polymorph of cocoa butter from EDX intensity was quite difficult in our experiments even though Kintra & Hatta (2005) distinguished form VI from V with the concentration of carbon and oxygen measured by EDX. The EDX signal from silicon in the reference chocolate was as low as the background noise level. The SEM image of the cross-section of the chocolate with silicone oil was similar to that of the chocolate without silicone oil. In contrast, the EDX images clearly illustrated the distribution of silicon in the chocolate. The signal from silicon spread from the silicone oil contact face of the chocolate. The area exhibiting a silicon signal became wider as the storage period increased. This result indicates that the signal was coming from the silicone oil migrating into the chocolate.
H. Hondoh et al. / Food Structure 8 (2016) 8–15
11
Fig. 2. Appearance change on the surface of the chocolates. (a), (b) reference chocolate without liquid oil. Fat bloom with (c), (d), (e) silicone oil and (f), (g), (h) canola oil. (a), (c) and (f) the ninth day, (d), (g) the 28th day, and (b), (e), and (h) after 2.5 month of storage.
The integrated intensity on the cross-section of the silicone oil contact face is depicted in Fig. 4. To analyze the migration mechanism, we fitted the silicon profiles for different storage periods using a variety of equations. After the trials, we obtained a well-fitted equation (smallest x2) of the sum of double exponentials I = A1exp( x/t1) + A2exp( x/t2)
(1)
Here I is EDX intensity, x is distance from the chocolate-silicone oil interface. The physical meaning of exponential terms is not clear but the two independent exponential terms indicate that there exist two different migration modes. A1,A2, t1 and t2 are the coefficients for the fitting. A1 and A2 are the relative maximum intensity of each exponential term, and t1 and t2 are the positions at which the intensity decreases to 1/e of the maximum intensity. The coefficient t corresponds to the migration rate. The higher t value
means silicone oil can reach further indicating faster migration. The relative maximum A shows the relative amount of migrated oil at the interface. The total amount of migrated silicone oil can be calculated by the integration of each term of Eq. (1). The silicon profiles for different storage periods were well-fitted using the sum of the double exponentials. Table 3 summarizes the fitting. The obtained parameters did not show significant variations, but the time evolution between them was unclear. These results may imply that not enough silicone oil was applied for steady-state migration, since 38% of the applied silicone oil had already migrated at day 9. The variation in the internal structures of the samples, which was induced during poor storage at the market, might also be the reason for the variation in the parameters. For all periods of the experiment, the equation contained double exponential terms, implying that there are slow and fast
Fig. 3. Melting behavior of crystals in liquid oil fraction obtained from the surface of the chocolate with silicone oil migration. These images were taken under polarized light microscopy.
12
H. Hondoh et al. / Food Structure 8 (2016) 8–15
Fig. 4. (a), (d), (g) SEM and (b), (e), (h) EDX images of reference sample and silicone oil migration in the chocolate. (c), (f), (i) Integrated intensity of silicon signals and the fitted lines. The left side of the images depicts the silicone-contacting side. The green dots represent silicone, and red dots represent carbon in (b), (e) and (h). (a), (b), (c) The reference chocolate without silicone oil. (d), (e), (f) The first day and (g), (h), (i) the ninth day of migration. The integrated intensity of silicon signals is indicated in red, and the fitted lines in blue in (c), (f) and (i).
migrations. Previous studies suggested that liquid oil migrates into chocolate via diffusion and capillary force. However, in our experiments, silicone oil and the low melting fraction of cocoa butter seldom diffused with each other, because the solubility of silicone oil in vegetable oil is low, and they exhibited liquid–liquid separation. So, the two migration modes suggested from the fitted equation did not imply diffusion but capillary force migration. One possibility of these two terms in the equation is that silicone oil migrated through gaps in cocoa butter crystals, which were fully or partially filled by the liquid fraction of cocoa butter or vegetable oil. A previous study indicated that compound chocolate contained several percent liquid fraction at room temperature and 1–4% cavities (Loisel et al., 1997). These cavities in compound chocolate Table 3 Summary of the fitting parameters using Eq. (1) for silicon profile. Day
A1
t1 (mm)
A2
t2 (mm)
x2
1 2 3 4 7 9
0.758 0.015 0.62 0.04 0.36 0.02 0.480 0.017 0.594 0.019 0.610 0.014
35.4 1.4 19 2 38 5 31 2 29.6 1.9 58 3
0.141 0.009 0.222 0.013 0.513 0.019 0.448 0.011 0.272 0.011 0.303 0.008
440 40 440 40 280 11 320 10 380 20 710 40
0.207 0.855 0.287 0.194 0.274 0.249
should be partially or fully filled with the liquid cocoa butter fraction or vegetable oil. If the gap was fully filled with liquid oil, silicone oil migrated into chocolate via capillary force while extruding the liquid oil in the gap. This migration should be slow because the silicone oil had to push all the liquid oil in the gap. In contrast, if the gap was partially filled with liquid oil, the silicone oil migrated faster because migration into a void capillary was much easier than migration into a filled capillary. The slow and fast migration fractions in our experiments were calculated by integrating the equation using the parameters and the amount of migrated silicone oil (Table 4). The integrated amount of migrated silicone oil indicated that silicone oil tended to migrate in the fast migration mode. Therefore, the fast migration with capillary force through partially filled capillary should be more Table 4 Calculated weight ratio for fast and slow migrations. Day
1 3 9
Weight gain (mg)
67 106 140
Fast migration
Slow migration
weight (mg)
ratio (%)
weight (mg)
ratio (%)
47 97 120
70 92 86
20 9 20
30 8 14
H. Hondoh et al. / Food Structure 8 (2016) 8–15
significant in silicone oil migration. If miscible oil (e.g., nut oil) was applied instead of silicone oil, it diffused into the liquid oil fraction in compound chocolate. Once the foreign oil reached unfilled parts in the gap, it migrated rapidly via capillary force. Therefore, we conclude that foreign oil migrates between cocoa butter crystals via diffusion if the gap is filled with a liquid oil fraction, and via a combination of diffusion and capillary force if the gap is a partially filled gap. Fig. 5 presents an EDX image on the first day of migration. A high density area of silicone oil was observed on the silicone oilcontacting side and, as mentioned previously, the density decreased with distance from the interface. In addition, a clear spot of silicone oil was found at a distance from the interface and was clearly isolated from the silicon at the interface. This isolated silicon spot could be assumed as a result of fastest migration, superfast migration. The diffusion or diffusion/capillary force mixed mechanisms were suggested from the analysis of the
13
distribution of silicon at the interface, while a different mechanism was suggested for this distant spot. The migration rate should be fastest if the gap is completely free of liquid oil (Altimiras et al., 2006). Therefore, it is suggested that this spot of silicone oil is correspond to a superfast migration through the air gap between cocoa butter crystals via a single capillary force mechanism. The migration rate with this mechanism was faster than 1.8 mm/day from Fig. 5. This superfast migration was rarely observed in our experiments. These results indicate that compound chocolate contains gaps between cocoa butter crystals, and most of these gaps are partially filled with a liquid fraction of cocoa butter or vegetable oil. 3.3. Identification of cacao particles Cocoa butter and sugar consist of carbon and oxygen atoms, so they could not be distinguished in the EDX images. It is possible to
Fig. 5. (a) SEM, and (b) EDX images of superfast migration of silicone oil in the chocolate, and (c) a magnified SEM image of the framed area in (a) and (b) combined with EDX. The left side of the images depicts the silicone-contacting side. Red denotes carbon and green denotes silicone in the EDX image (b) and (c). These images were taken on the first day of migration.
14
H. Hondoh et al. / Food Structure 8 (2016) 8–15
Fig. 6. Effect of particles on migration. (a) SEM image overlapped with phosphorous distribution and (b) EDX signal from silicon. Red represents EDX signals from phosphorous, and green represents those from silicon. Solid circles indicate the cacao particles. (c) Integrated intensity of EDX signals from phosphorous (red) and silicon (green) along the square in (a) and (b).
identify cacao particles in SEM images of the EDX results because cacao particles contain phosphorus, magnesium and potassium (Radak et al., 2013). We used overlapping images of SEM and EDX: Fig. 6 illustrates the distribution of phosphorus in compound chocolate, and indicates a localized distribution in the chocolate. This localized phosphorous is suggested to be from cocoa particles, since no other ingredients in the chocolate contain detectable amounts of phosphorous. The EDX signal from silicon in the cacao particle was not significantly stronger than that from other parts of the chocolate, suggesting that the silicone oil hardly penetrated or became concentrated in cacao particles, and the surfaces of these particles could not form an advanced pathway for oil migration. The effects of particles were investigated in previous studies: Altimiras et al. (2006) and Dahlenborg et al. (2015a) demonstrated that smaller particles increased the migration rate. In contrast, Choi, McCarthy, and McCarthy (2005) and McCarthy and McCarthy (2008) reported a positive correlation between particle size and oil migration rate: migration rate increased with larger particle size. Larger particles had a much narrower interface than smaller particles when the total particle volume was the same. Thus, the migration rate will be higher for larger particles if the interface forms a migration pathway. The interface between the particles and cocoa butter crystals was not advantageous for silicone oil migration in this study, suggesting our results would be consistent with the observations of Altimiras et al. (2006) and Dahlenborg et al. (2015b). Dispersed small particles may affect the density of the cavity in chocolate, which is related to oil migration.
4. Conclusion Oil migration in chocolate was visualized using SEM–EDX with silicone oil as the mobile phase. Migrating silicone oil pushed the liquid oil fraction in the compound chocolate towards the surface, and the liquid oil droplets induced fat bloom. The distribution of migrating silicone oil was analyzed quantitatively, based on weight change and EDX intensity. The obtained distribution of silicone oil demonstrated two possible mechanisms of migration: diffusion and/or capillary force. Superfast migration was observed for the first time in the present study using SEM–EDX. This is evidence of pure capillary force-induced migration. Since silicone oil did not concentrate around the solid particles, such as sugar and cacao particles, the surface of the particle might not work as a migration pathway. The application of silicone oil was an effective, simple approach in the investigation of oil migration. We intend to apply this method in further research on oil migration in chocolates, biscuits, and breads. Oil migration in chocolate is significant issue because it produces deterioration of chocolate with fat bloom. We developed a novel approach for the investigation of oil migration in foods such as chocolate. A quantitative analysis of oil migration was carried out with this method. The results in this work provide the knowledge for deeper understanding of oil migration. Acknowledgement This work was supported by JSPS Grant-in-Aid for Young Scientists (B) grant number 15K18698.
H. Hondoh et al. / Food Structure 8 (2016) 8–15
References Aguilera, J. M., Michel, M., & Mayer, G. (2004). Fat migration in chocolate: diffusion or capillary flow in a particulate solid?—A hypothesis paper. Journal of Food Science, 69, R167–R174. Altimiras, P., Pyle, L., & Bouchon, P. (2006). Structure-fat migration relationship during storage of cocoa butter model bars: bloom development and possible mechanisms. Journal of Food Engineering, 80, 600–610. Choi, Y. J., McCarthy, K. L., & McCarthy, M. J. (2005). Oil migration in a chocolate confectionery system evaluated by magnetic resonance imaging. Journal of Food Science, 70, E312–E317. Dahlenborg, H., Millqvist-Fureby, A., & Bergenståhl, B. (2015b). Effect of microstructure on oil migration and fat bloom development in model pralines. Food Structure, 5, 51–65. Dahlenborg, H., Millqvist-Fureby, A., & Bergenståhl, B. (2015a). Effect of particle size in chocolate shell on oil migration and fat bloom development. Journal of Food Engineering, 146, 172–181. Dahlenborg, H., Millqvist-Fureby, A., Bergenståhl, B., & Kalnin, D. J. E. (2011). Investigation of chocolate surfaces using profilometry and low vacuum scanning electron microscopy. Journal of the American Oil Chemists’ Society, 88, 773–783. Dahlenborg, H., Millqvist-Fureby, A., Brandner, B. D., & Bergenståhl, B. (2012). Study of the porous structure of white chocolate by confocal Raman microscopy. European Journal of Lipid Science and Technology, 114, 919–926. Duce, S. L., Carpenter, T. A., & Hall, L. D. (1990). Nuclear magnetic resonance imaging of chocolate confectionery and the spatial detection of polymorphic states of cocoa butter in chocolate. Lebensmittel-Wissenschaft Technology, 23, 545–549. Ghosh, V., Ziegler, G. R., & Anantheswaran, R. C. (2002). Fat, moisture, and ethanol migration through chocolates and confectionary coating. Critical Reviews in Food Science and Nutrition, 42, 583–626. Guiheneuf, T. M., Couzens, P. J., Wille, H.-J., & Hall, L. D. (1997). Visualisation of liquid triacylglyceril migration in chocolate by magnetic resonance imaging. The Journal of the Science of Food and Agriculture, 73, 265–273. Kinta, Y., & Hatta, T. (2005). Composition and structure of fat bloom in untempered vhocolate. Journal of Food Science, 70, S450–S452. Kinta, Y., & Hatta, T. (2007). Composition, structure, and color of fat bloom due to the partial liquefaction of fat in dark chocolate. Journal of the American Oil Chemists’ Society, 84, 107–115. Lee, W. L., McCarthy, M. J., & McCarthy, K. L. (2010). Oil migration in 2-component confectionery system. Journal of Food Science, 75, E83–E89. Loisel, C., Lecq, G., Ponchel, G., Keller, G., & Ollivon, M. (1997). Fat bloom and chocolate structure studied by mercury porosimetry. Journal of Food Science, 62, 781–788. Lonchampt, P., & Hartel, R. W. (2004). Fat bloom in chocolate and compound coatings. European Journal of Lipid Science and Technology, 106, 241–274.
15
Marty, S., Baker, K., Dibildox-Alvarado, E., Rodrigues, J. N., & Marangoni, A. G. (2005). Monitoring and quantifying of oil migration in cocoa butter using a flatbed scanner and fluorescence light microscopy. Food Research International, 38, 1189–1197. Marty, S., Baker, K., & Marangoni, A. G. (2009). Optimization of scanner imaging technique to accurately study oil migration kinetics. Food Research International, 42, 368–373. McCarthy, K. L., & McCarthy, M. J. (2008). Oil migration in chocolate–peanut butter paste confectionery as a function of chocolate formation. Journal of Food Science, 73, E266–E273. Miquel, M. E., Carli, S., Couzens, P. J., Wille, H.-J., & Hall, L. D. (2001). Kinetics of the migration of lipids in composite chocolate measured by magnetic resonance imaging. Food Research International, 34, 773–781. Radak, Z., Silye, G., Bartha, C., Jakus, J., Banyai, E. S., Atalay, M., et al. (2013). The effects of cocoa supplementation, caloric restriction, and regular exercise, on oxidative stress markers of brain and memory in the rat model. Food and Chemical Toxicology, 61, 36–41. Rasband, W.S. (1997–2016). ImageJ. Bethesda, Maryland, USA: U.S. National Institutes of Health. http://imagej.nih.gov/ij/. Reinke, S. K., Roth, S. V., Santoro, G., Vieira, J., Heinrich, S., & Palzer, S. (2015). Tracking structural changes in lipid-based multicomponents food materials due to oil migration by microfocus small-angle X-ray scattering. Applied Materials & Interface, 7, 9929–9936. Rousseau, D. (2006). On the porous mesostructure of milk chocolate viewed with atomic force microscopy. Lebensmittel-Wissenschaft Technology, 39, 852–860. Rumsey, T. R., & McCarthy, K. L. (2012). Modeling oil migration in two-layer chocolate-almond confectionery products. Journal of Food Engineering, 111, 149– 155. Sonwai, S., & Rousseau, D. (2010). Controlling fat bloom formation in chocolate— impact of milk fat on microstructure and fat phase crystallization. Food Chemistry, 119, 286–297. Talbot, G. (1996). The washer test: a method for monitoring fat migration. The Manufacturing Confectioner, 76, 87–90. Ziegleder, G., Moser, C., & Geier-Greguska, J. (1996a). Kinetics of fat migration within chocolate products, part I: principles and analytics: fett/lipid. European Journal of Lipid Science and Technology, 98, 196–199. Ziegleder, G., Moser, C., & Geier-Greguska, J. (1996b). Kinetics of fat migration within chocolate products. part II: influence of storage temperature, diffusion coefficient, solid fat content: fett/lipid. European Journal of Lipid Science and Technology, 98, 253–256. Ziegler, G. (2009). Product design and shelf-life issues: oil migration and fat bloom. In G. Talbot (Ed.), Science and technology of enrobed and filled chocolate, confectionery and bakery productsUK: Woodhead Publishing.
Contents lists available at ScienceDirect
Food Structure journal homepage: www.elsevier.com/locate/foostr
Visualization of oil migration in chocolate using scanning electron microscopy–energy dispersive X-ray spectroscopy Hironori Hondoh* , Kenta Yamasaki, Miharu Ikutake, Satoru Ueno Graduate School of Biosphere Science, Hiroshima University, 1-4-4, Kagamiyama, Higashi-Hiroshima, Hiroshima 739-8528, Japan
A R T I C L E I N F O
Article history: Received 10 June 2015 Received in revised form 10 March 2016 Accepted 1 April 2016 Keywords: Chocolate Oil migration Energy dispersive X-ray spectroscopy SEM Silicone oil Chocolate structure
A B S T R A C T
We have developed a novel method to observe oil migration in chocolate using scanning electron microscopy and energy dispersive X-ray spectroscopy (EDX). If silicone oil is used as the mobile phase in migration, the amount of migrated silicone oil can be evaluated using the EDX signal. Compound chocolate placed on the silicone-oiled cotton resulted in the appearance of liquid oil droplets on the surface of the chocolate within 10 days at 24 C. Fat blooming was also induced by these oil droplets at 2.5 months. The amount of silicone oil that migrated into the chocolate was evaluated based on the weight change of the chocolate and the EDX signal from the silicone oil. Quantitative analysis indicated that two independent migration mechanisms, diffusion and capillary force, induced migration. In addition, the EDX signal indicated superfast migration from a spot that was distinct from the area of diffusion or capillary force migration. The EDX signals from phosphorus or magnesium atoms were helpful in identifying cacao particles in chocolate since these elements were specific to the particles in chocolate. Homogeneous distribution of silicone oil was observed around the particles in chocolate, thus, these particles might not help to produce a specific migration pathway in chocolate. ã 2016 Elsevier Ltd. All rights reserved.
1. Introduction Fat bloom, caused by the formation of needle-shaped fat crystals, is a significant problem in the confectionery industry. It is a greyish haze which forms on the surface of chocolate during storage and is promoted by storage near the melting point, temperature fluctuation, and oil migration (Lonchampt & Hartel, 2004; Rousseau, 2006). Customers dislike the visual appearance and textural attributes of bloomed chocolate, and this deterioration results in decreased consumer confidence. When chocolate is stored at a temperature above the melting point of form V crystals, melt-mediated transformation leads to the development of form VI crystals which grow to cause fat blooming. Temperature fluctuation causes liquid oil to pump out from the inside of the chocolate to the surface, even at a low temperature (Sonwai & Rousseau, 2010). The pumped liquid oil dissolves the surface of the chocolate and promotes recrystallization for blooming. These two issues are promoted by temperature change, however, fat bloom induced by oil migration occurs independently of temperature. If chocolate
Abbreviations: TAG, triacylglycerol; EDX, energy dispersive X-ray spectroscopy; SEM, scanning electron microscopy. * Corresponding author. E-mail address: [email protected] (H. Hondoh). http://dx.doi.org/10.1016/j.foostr.2016.04.001 2213-3291/ ã 2016 Elsevier Ltd. All rights reserved.
comes into contact with other ingredients containing liquid oil, the liquid oils in chocolate and the other ingredients migrate and mix together. Thus, migration promotes the softening of filled chocolate, the hardening of fillings, and fat bloom. Therefore, oil migration is a serious problem in praline chocolates, chocolatecovered cookies and chocolates with nuts. Oil migration is thought to occur by diffusion or capillary flow in chocolate to the surface (Ziegler, 2009). The driving force for diffusion is assumed to be the difference in chemical components: the triacylglycerol (TAG) concentration gradient in domains of confectionery chocolate products. If the liquid fraction of oil in a filling has a different TAG composition than cocoa butter, oil migration occurs. This diffusion has been the preferred explanation for oil migration in chocolate products (Aguilera, Michel, & Mayer, 2004; Altimiras, Pyle, & Bouchon, 2006; Ghosh, Ziegler, & Anantheswaran, 2002; Miquel, Carli, Couzens, Wille, & Hall, 2001). The possibility of capillary flow has also been studied: pores exist inside chocolate because of volume shrinkage during crystallization and the volume of this vacancy has been calculated to be 1% of chocolate (Loisel, Lecq, Ponchel, Keller, & Ollivon, 1997). Mass transportation occurs through these pores via capillary force. The contribution from these two mechanisms is difficult to distinguish because they exhibit similar time dependence (Aguilera et al., 2004). In addition, another migration mechanism,
H. Hondoh et al. / Food Structure 8 (2016) 8–15
pressure-driven convection flow, has been proposed (Altimiras et al., 2006; Dahlenborg, Millqvist-Fureby, Bergenståhl, & Kalnin, 2011; Dahlenborg, Millqvist-Fureby, Brandner, & Bergenståhl, 2012; Dahlenborg, Millqvist-Fureby, & Bergenståhl, 2015a). Numerous approaches have been utilized in the quantitative investigation of oil migration. Quantification of fat composition using chromatography is a common technique in the analysis of oil migration (Talbot, 1996; Ziegleder, Moser, & Geiwe-Greguska, 1996a; Ziegleder, Moser & Geiwe-Greguska, 1996b), however, spatial resolution is low because the samples must be divided for this analysis. Magnetic resonance imaging has been used to observe the inner structure of chocolate and to monitor oil migration in chocolate (Duce, Carpenter, & Hall, 1990; Guiheneuf, Couzens, Wille, & Hall, 1997). Magnetic resonance imaging is a nondestructive quantitative analytical method and thus is advantageous for oil migration investigation. Using this method, kinetic analysis of oil migration has made much progress, and has been used to determine the diffusion coefficient in model confectionery products (Guiheneuf et al., 1997; Lee, McCarthy, & McCarthy, 2010; McCarthy and McCarthy, 2008; Miquel et al., 2001; Rumsey & McCarthy, 2012). A flat-bed scanner combined with dye stain is a convenient technique for observing oil migration (Marty, Baker, Dibildox-Alvarado, Rodrigues, & Marangoni, 2005; Marty, Baker, & Marangoni, 2009). These techniques provide quantitative information on oil migration kinetics. A combination of scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDX) is a novel approach in the investigation of migration and blooming in chocolate. EDX is an analytical technique to identify elemental composition of a sample by measuring energy of characteristic X-ray from a specimen. Kinta and Hatta (2005, 2007) were the first to report on SEM–EDX in chocolate analysis (Kinta & Hatta, 2005, 2007). Using brominated vegetable oil as a die in SEM–EDX, Dahlenborg, Millqvist-Fureby, and Bergenståhl (2015b) found that the size of particles in chocolate affects migration velocity, and significantly increased the developing rate of fat bloom. They found that smaller particles of cocoa powder and sugar resulted in increased migration. A theoretical approach calculated that the velocity of capillary force migration was considerably faster than the migration rate obtained in the experimental results (Altimiras et al., 2006). In addition, Reinke et al. (2015) used small angle X-ray scattering to report that oil migration through pores occurred within seconds. These findings may support the significance of capillary force in oil migration. However, the oil migration pathway has not been observed directly, and details of oil migration remain obscure. The aim of this study was to demonstrate a novel approach to quantitative analysis of oil migration using SEM–EDX. Silicone oil was used as the mobile phase in this study and its distribution was visualized using EDX. Silicone oils have advantageous physical properties, such as high thermal and oxidation stability, and low viscosity and surface tension. The surface tension of silicone oils is lower than that of other natural oils, so that they spread easily over an interface, and are widely used as deformers and release agents in the industrial field. In the present study, the significant property of silicone oils was low solubility in cocoa butter: silicone oil and cocoa butter are immiscible therefore the diffusion of silicone oil
9
into chocolate is negligible. Using silicone oil as the mobile phase of oil migration in chocolate, we investigated the significance of capillary force and diffusion in oil migration. 2. Materials and methods 2.1. Materials Commercial compound milk chocolate bars (Meiji Co., Ltd., Japan) and canola oil (The Nisshin Oillio Group, Ltd., Japan) were purchased at a local market. The chocolate contained sugar, cocoa powder, vegetable oil, whole milk powder, cocoa butter, and soy lecithin, in descending order of content. Silicone oil, Element14* PDMS 10-JC was obtained from Momentive Performance Materials Inc. (Japan). The physical properties of the silicone and canola oils are presented in Table 1. 2.2. Sample preparation For the migration experiments, a chocolate bar was broken into pieces (32 26 mm2) using a sharp knife. A piece of chocolate was placed on a cotton pad (also 32 26 mm2) impregnated with silicone oil or canola oil in a polystyrene case. The volume of the oils was 600 ml, the weight of the canola oil was 552 mg, and that of the silicone oil was 562 mg. The samples were incubated at 24 C in an air incubator. To confirm the migration ability of silicone oil in the relatively high liquid oil environment of chocolate, the chocolate samples were stored at 24 C. Stored chocolates were used for further experiments. 2.3. Weight change and hardness measurements Weight change during storage was measured. The cotton pad was removed from the chocolate, and the weight of the chocolate was measured without wiping oil from the contact face. The measured samples were returned to the cotton pad in an incubator and used for further weight measurements. Hardness was measured using a penetrometer (EX-210E, Elex Science Co., Ltd., Japan) with a penetration rod that had a base diameter of 64.91 mm, a cone height of 29.08 mm, and a cone weight of 102.51 g. A penetration test was performed on the oil contacting face of the chocolate samples. These measurements were repeated at least five times. 2.4. SEM and EDX observations The central part of the stored chocolate was cut out in a block shape (3 3 4 mm3) using a sharp knife. When the tip of the knife was inserted into the chocolate, the chocolate was loosened and a piece of chocolate was released. The fragment was carefully trimmed to become a block shape. The surface of the chocolate block was coated with Au using a sputter coater (E-1010, Hitachi Instruments, Japan), and the cross-section was observed using SEM (S-2380, Hitachi Instruments, Japan). The acceleration voltage was 15 kV. To avoid artifacts in the SEM image, the area of chocolate where the knife had made contactwas not observed. Element
Table 1 Physical properties of silicone and canola oils.
Element14 PDMS 10-JC Canola oil
Dynamic viscosity (mm2/s)
Density (g/cm3)
Surface tension (mN/m)
Pour point ( C)
10 (25 C) 50.91 (37.8 C)
0.935 (25 C) 0.906 0.920 (25 C)
20.1 (25 C) 31.7 (20 C)
Below 60 Below 18
10
H. Hondoh et al. / Food Structure 8 (2016) 8–15
mapping was performed using EDX (E-MAX-7000S, Horiba, Japan) in the same field of view as the SEM. These measurements were carried out at room temperature. The EDX images were analyzed using Image-J (Rasband, 1997) to obtain the integrated intensity of EDX signal. Curve fitting was performed using Igor Pro (WaveMetrics, USA). The fitting was evaluated by x2 value. 2.5. Statistical analysis The results of the penetration test were statistically tested using two-way ANOVA and the Tukey-Kramer test. Significant differences (p < 0.05) were evaluated using Microsoft Excel. The experiments were repeated 5 times. 3. Results and discussion 3.1. Changes in weight, hardness and appearance We measured changes in the weight of compound chocolates with oil migration throughout the storage time to determine whether silicone and canola oils migrated into the chocolates. Fig. 1 indicates weight changes in chocolates stored with silicone and canola oils at 24 C. The weight of both chocolate samples clearly increased during the storage time, confirming that silicone and canola oils migrated into the chocolates. In addition, the weightincreasing rate of the silicone oil-chocolate was almost twice that of the canola oil-chocolate. After 20 days of storage, 38% of the applied silicone oil had migrated into the chocolate, compared with 22% of the canola oil, if counter diffusion was ignored. This result implied that silicone oil migration is faster than canola oil migration. The lower viscosity of silicone oil could be the reason for faster migration, or counter diffusion was significant in the canola oil-chocolate. The results of the penetration test are presented in Table 2. There was no significant difference in the hardness of the bare compound chocolate compared with the silicone oilchocolate after storage for 20 days at 24 C. In contrast, the penetration depth of the canola oil-chocolate was deeper from day 1 compared with the reference sample, and the depth was increased until day 10. Cocoa butter crystals are soluble in vegetable oils, such as canola oil, and chocolate softens by dissolving in vegetable oils. In contrast, the low solubility of cocoa butter in silicone oil did not affect the hardness of the chocolate. Fig. 2 depicts appearance changes in the chocolate samples during storage. The appearance of the reference compound chocolate did not change during 2.5 months of storage (Fig. 2(a, b)). After 9 days storage of the silicone oil-chocolate, liquid droplets appeared on the surface (Fig. 2 (c)) and the area of liquid increased until the chocolate became soaking wet within 4 weeks
Fig. 1. Weight gain of chocolate with silicone oil (squares) and canola oil (diamonds).
Table 2 Penetration depth (mean SD in 1/10 mm) in chocolate samples. Mobile phase
Day 1
Day 4
Day 10
Day 20
Reference Silicone oil Canola oil
7.3 0.7Aa 11.1 1.7Bab 14 2Cb
7.9 1.4Aa 11.2 1.6Ba 18 6CDb
7.8 1.9Aa 11.1 1.1Ba 22 4Db
7 3Aa 10.4 1.6Ba 18 4CDb
ab Same superscript indicates no significant difference (p > 0.05) in a column. ABCD Same superscript indicates no significant difference (p > 0.05) in a row.
(Fig. 2 (d)). At the end of the 2.5 months storage period, the surface of the silicone oil-chocolate bloomed and whitened (Fig. 2 (e)). In contrast, liquid droplets did not appear on the surface of the canola oil-chocolate for the entire storage period. The chocolate with canola oil bloomed after 4 weeks, and the bloom spread the whole surface after 2.5 months in storage (Fig. 2 (g, h)). An aliquot of the liquid droplets on the surface of the silicone oil-chocolate was taken and cooled to 5 C, and then heated until crystals were disappeared on a temperature-controllable stage (THMS500, Linkam Scientific Instruments, UK) under polarized optical microscopy. The liquid crystallized at 8 C while cooling, and the crystals melted and disappeared at 11 C while heating (Fig. 3). The pour point of silicone oil is below 60 C, so silicone oil does not crystallize at this temperature, therefore, we concluded that the liquid which had collected on the surface of the compound chocolate with silicone oil was a low melting point fraction of cocoa butter or vegetable oil from the chocolate. Because the solubility of cocoa butter and vegetable oil in silicone oil is quite low, silicone oil seldom diffuses in the liquid fraction of cocoa butter. Consequently, the liquid oil fraction in the chocolate is pumped out from the chocolate by migrating silicone oil. This is clear evidence that the liquid fraction of cocoa butter is pumped out via silicone oil migration. 3.2. SEM and EDX images SEM and EDX images of the chocolate cross-section are presented in Fig. 4. The edge of the cut end of chocolate sample was flattened where the blade touched it, however, most of the section had a rough surface indicating that the chocolate was cleaved and the microstructure of the chocolate was maintained. Spherical voids (air bubbles) were preserved, however, cocoa butter crystals, sugar and cacao particles could not be distinguished in the obtained SEM images. Comparison of the SEM and EDX images of cross-sections indicated that EDX intensity clearly depended on the topography of the chocolate. Thus, the absolute EDX intensity from the specimen in this analysis was not directly linked to the concentration of the specimen. In general, EDX intensity from the specimen is affected by the surface topography of the samples because the detection rate of the emitted X-ray to the detector depends on the surface topography. Thus, EDX intensity from the carbon on the EDX image depended on the undulation on the chocolate surface (Fig. 4). Identification of the polymorph of cocoa butter from EDX intensity was quite difficult in our experiments even though Kintra & Hatta (2005) distinguished form VI from V with the concentration of carbon and oxygen measured by EDX. The EDX signal from silicon in the reference chocolate was as low as the background noise level. The SEM image of the cross-section of the chocolate with silicone oil was similar to that of the chocolate without silicone oil. In contrast, the EDX images clearly illustrated the distribution of silicon in the chocolate. The signal from silicon spread from the silicone oil contact face of the chocolate. The area exhibiting a silicon signal became wider as the storage period increased. This result indicates that the signal was coming from the silicone oil migrating into the chocolate.
H. Hondoh et al. / Food Structure 8 (2016) 8–15
11
Fig. 2. Appearance change on the surface of the chocolates. (a), (b) reference chocolate without liquid oil. Fat bloom with (c), (d), (e) silicone oil and (f), (g), (h) canola oil. (a), (c) and (f) the ninth day, (d), (g) the 28th day, and (b), (e), and (h) after 2.5 month of storage.
The integrated intensity on the cross-section of the silicone oil contact face is depicted in Fig. 4. To analyze the migration mechanism, we fitted the silicon profiles for different storage periods using a variety of equations. After the trials, we obtained a well-fitted equation (smallest x2) of the sum of double exponentials I = A1exp( x/t1) + A2exp( x/t2)
(1)
Here I is EDX intensity, x is distance from the chocolate-silicone oil interface. The physical meaning of exponential terms is not clear but the two independent exponential terms indicate that there exist two different migration modes. A1,A2, t1 and t2 are the coefficients for the fitting. A1 and A2 are the relative maximum intensity of each exponential term, and t1 and t2 are the positions at which the intensity decreases to 1/e of the maximum intensity. The coefficient t corresponds to the migration rate. The higher t value
means silicone oil can reach further indicating faster migration. The relative maximum A shows the relative amount of migrated oil at the interface. The total amount of migrated silicone oil can be calculated by the integration of each term of Eq. (1). The silicon profiles for different storage periods were well-fitted using the sum of the double exponentials. Table 3 summarizes the fitting. The obtained parameters did not show significant variations, but the time evolution between them was unclear. These results may imply that not enough silicone oil was applied for steady-state migration, since 38% of the applied silicone oil had already migrated at day 9. The variation in the internal structures of the samples, which was induced during poor storage at the market, might also be the reason for the variation in the parameters. For all periods of the experiment, the equation contained double exponential terms, implying that there are slow and fast
Fig. 3. Melting behavior of crystals in liquid oil fraction obtained from the surface of the chocolate with silicone oil migration. These images were taken under polarized light microscopy.
12
H. Hondoh et al. / Food Structure 8 (2016) 8–15
Fig. 4. (a), (d), (g) SEM and (b), (e), (h) EDX images of reference sample and silicone oil migration in the chocolate. (c), (f), (i) Integrated intensity of silicon signals and the fitted lines. The left side of the images depicts the silicone-contacting side. The green dots represent silicone, and red dots represent carbon in (b), (e) and (h). (a), (b), (c) The reference chocolate without silicone oil. (d), (e), (f) The first day and (g), (h), (i) the ninth day of migration. The integrated intensity of silicon signals is indicated in red, and the fitted lines in blue in (c), (f) and (i).
migrations. Previous studies suggested that liquid oil migrates into chocolate via diffusion and capillary force. However, in our experiments, silicone oil and the low melting fraction of cocoa butter seldom diffused with each other, because the solubility of silicone oil in vegetable oil is low, and they exhibited liquid–liquid separation. So, the two migration modes suggested from the fitted equation did not imply diffusion but capillary force migration. One possibility of these two terms in the equation is that silicone oil migrated through gaps in cocoa butter crystals, which were fully or partially filled by the liquid fraction of cocoa butter or vegetable oil. A previous study indicated that compound chocolate contained several percent liquid fraction at room temperature and 1–4% cavities (Loisel et al., 1997). These cavities in compound chocolate Table 3 Summary of the fitting parameters using Eq. (1) for silicon profile. Day
A1
t1 (mm)
A2
t2 (mm)
x2
1 2 3 4 7 9
0.758 0.015 0.62 0.04 0.36 0.02 0.480 0.017 0.594 0.019 0.610 0.014
35.4 1.4 19 2 38 5 31 2 29.6 1.9 58 3
0.141 0.009 0.222 0.013 0.513 0.019 0.448 0.011 0.272 0.011 0.303 0.008
440 40 440 40 280 11 320 10 380 20 710 40
0.207 0.855 0.287 0.194 0.274 0.249
should be partially or fully filled with the liquid cocoa butter fraction or vegetable oil. If the gap was fully filled with liquid oil, silicone oil migrated into chocolate via capillary force while extruding the liquid oil in the gap. This migration should be slow because the silicone oil had to push all the liquid oil in the gap. In contrast, if the gap was partially filled with liquid oil, the silicone oil migrated faster because migration into a void capillary was much easier than migration into a filled capillary. The slow and fast migration fractions in our experiments were calculated by integrating the equation using the parameters and the amount of migrated silicone oil (Table 4). The integrated amount of migrated silicone oil indicated that silicone oil tended to migrate in the fast migration mode. Therefore, the fast migration with capillary force through partially filled capillary should be more Table 4 Calculated weight ratio for fast and slow migrations. Day
1 3 9
Weight gain (mg)
67 106 140
Fast migration
Slow migration
weight (mg)
ratio (%)
weight (mg)
ratio (%)
47 97 120
70 92 86
20 9 20
30 8 14
H. Hondoh et al. / Food Structure 8 (2016) 8–15
significant in silicone oil migration. If miscible oil (e.g., nut oil) was applied instead of silicone oil, it diffused into the liquid oil fraction in compound chocolate. Once the foreign oil reached unfilled parts in the gap, it migrated rapidly via capillary force. Therefore, we conclude that foreign oil migrates between cocoa butter crystals via diffusion if the gap is filled with a liquid oil fraction, and via a combination of diffusion and capillary force if the gap is a partially filled gap. Fig. 5 presents an EDX image on the first day of migration. A high density area of silicone oil was observed on the silicone oilcontacting side and, as mentioned previously, the density decreased with distance from the interface. In addition, a clear spot of silicone oil was found at a distance from the interface and was clearly isolated from the silicon at the interface. This isolated silicon spot could be assumed as a result of fastest migration, superfast migration. The diffusion or diffusion/capillary force mixed mechanisms were suggested from the analysis of the
13
distribution of silicon at the interface, while a different mechanism was suggested for this distant spot. The migration rate should be fastest if the gap is completely free of liquid oil (Altimiras et al., 2006). Therefore, it is suggested that this spot of silicone oil is correspond to a superfast migration through the air gap between cocoa butter crystals via a single capillary force mechanism. The migration rate with this mechanism was faster than 1.8 mm/day from Fig. 5. This superfast migration was rarely observed in our experiments. These results indicate that compound chocolate contains gaps between cocoa butter crystals, and most of these gaps are partially filled with a liquid fraction of cocoa butter or vegetable oil. 3.3. Identification of cacao particles Cocoa butter and sugar consist of carbon and oxygen atoms, so they could not be distinguished in the EDX images. It is possible to
Fig. 5. (a) SEM, and (b) EDX images of superfast migration of silicone oil in the chocolate, and (c) a magnified SEM image of the framed area in (a) and (b) combined with EDX. The left side of the images depicts the silicone-contacting side. Red denotes carbon and green denotes silicone in the EDX image (b) and (c). These images were taken on the first day of migration.
14
H. Hondoh et al. / Food Structure 8 (2016) 8–15
Fig. 6. Effect of particles on migration. (a) SEM image overlapped with phosphorous distribution and (b) EDX signal from silicon. Red represents EDX signals from phosphorous, and green represents those from silicon. Solid circles indicate the cacao particles. (c) Integrated intensity of EDX signals from phosphorous (red) and silicon (green) along the square in (a) and (b).
identify cacao particles in SEM images of the EDX results because cacao particles contain phosphorus, magnesium and potassium (Radak et al., 2013). We used overlapping images of SEM and EDX: Fig. 6 illustrates the distribution of phosphorus in compound chocolate, and indicates a localized distribution in the chocolate. This localized phosphorous is suggested to be from cocoa particles, since no other ingredients in the chocolate contain detectable amounts of phosphorous. The EDX signal from silicon in the cacao particle was not significantly stronger than that from other parts of the chocolate, suggesting that the silicone oil hardly penetrated or became concentrated in cacao particles, and the surfaces of these particles could not form an advanced pathway for oil migration. The effects of particles were investigated in previous studies: Altimiras et al. (2006) and Dahlenborg et al. (2015a) demonstrated that smaller particles increased the migration rate. In contrast, Choi, McCarthy, and McCarthy (2005) and McCarthy and McCarthy (2008) reported a positive correlation between particle size and oil migration rate: migration rate increased with larger particle size. Larger particles had a much narrower interface than smaller particles when the total particle volume was the same. Thus, the migration rate will be higher for larger particles if the interface forms a migration pathway. The interface between the particles and cocoa butter crystals was not advantageous for silicone oil migration in this study, suggesting our results would be consistent with the observations of Altimiras et al. (2006) and Dahlenborg et al. (2015b). Dispersed small particles may affect the density of the cavity in chocolate, which is related to oil migration.
4. Conclusion Oil migration in chocolate was visualized using SEM–EDX with silicone oil as the mobile phase. Migrating silicone oil pushed the liquid oil fraction in the compound chocolate towards the surface, and the liquid oil droplets induced fat bloom. The distribution of migrating silicone oil was analyzed quantitatively, based on weight change and EDX intensity. The obtained distribution of silicone oil demonstrated two possible mechanisms of migration: diffusion and/or capillary force. Superfast migration was observed for the first time in the present study using SEM–EDX. This is evidence of pure capillary force-induced migration. Since silicone oil did not concentrate around the solid particles, such as sugar and cacao particles, the surface of the particle might not work as a migration pathway. The application of silicone oil was an effective, simple approach in the investigation of oil migration. We intend to apply this method in further research on oil migration in chocolates, biscuits, and breads. Oil migration in chocolate is significant issue because it produces deterioration of chocolate with fat bloom. We developed a novel approach for the investigation of oil migration in foods such as chocolate. A quantitative analysis of oil migration was carried out with this method. The results in this work provide the knowledge for deeper understanding of oil migration. Acknowledgement This work was supported by JSPS Grant-in-Aid for Young Scientists (B) grant number 15K18698.
H. Hondoh et al. / Food Structure 8 (2016) 8–15
References Aguilera, J. M., Michel, M., & Mayer, G. (2004). Fat migration in chocolate: diffusion or capillary flow in a particulate solid?—A hypothesis paper. Journal of Food Science, 69, R167–R174. Altimiras, P., Pyle, L., & Bouchon, P. (2006). Structure-fat migration relationship during storage of cocoa butter model bars: bloom development and possible mechanisms. Journal of Food Engineering, 80, 600–610. Choi, Y. J., McCarthy, K. L., & McCarthy, M. J. (2005). Oil migration in a chocolate confectionery system evaluated by magnetic resonance imaging. Journal of Food Science, 70, E312–E317. Dahlenborg, H., Millqvist-Fureby, A., & Bergenståhl, B. (2015b). Effect of microstructure on oil migration and fat bloom development in model pralines. Food Structure, 5, 51–65. Dahlenborg, H., Millqvist-Fureby, A., & Bergenståhl, B. (2015a). Effect of particle size in chocolate shell on oil migration and fat bloom development. Journal of Food Engineering, 146, 172–181. Dahlenborg, H., Millqvist-Fureby, A., Bergenståhl, B., & Kalnin, D. J. E. (2011). Investigation of chocolate surfaces using profilometry and low vacuum scanning electron microscopy. Journal of the American Oil Chemists’ Society, 88, 773–783. Dahlenborg, H., Millqvist-Fureby, A., Brandner, B. D., & Bergenståhl, B. (2012). Study of the porous structure of white chocolate by confocal Raman microscopy. European Journal of Lipid Science and Technology, 114, 919–926. Duce, S. L., Carpenter, T. A., & Hall, L. D. (1990). Nuclear magnetic resonance imaging of chocolate confectionery and the spatial detection of polymorphic states of cocoa butter in chocolate. Lebensmittel-Wissenschaft Technology, 23, 545–549. Ghosh, V., Ziegler, G. R., & Anantheswaran, R. C. (2002). Fat, moisture, and ethanol migration through chocolates and confectionary coating. Critical Reviews in Food Science and Nutrition, 42, 583–626. Guiheneuf, T. M., Couzens, P. J., Wille, H.-J., & Hall, L. D. (1997). Visualisation of liquid triacylglyceril migration in chocolate by magnetic resonance imaging. The Journal of the Science of Food and Agriculture, 73, 265–273. Kinta, Y., & Hatta, T. (2005). Composition and structure of fat bloom in untempered vhocolate. Journal of Food Science, 70, S450–S452. Kinta, Y., & Hatta, T. (2007). Composition, structure, and color of fat bloom due to the partial liquefaction of fat in dark chocolate. Journal of the American Oil Chemists’ Society, 84, 107–115. Lee, W. L., McCarthy, M. J., & McCarthy, K. L. (2010). Oil migration in 2-component confectionery system. Journal of Food Science, 75, E83–E89. Loisel, C., Lecq, G., Ponchel, G., Keller, G., & Ollivon, M. (1997). Fat bloom and chocolate structure studied by mercury porosimetry. Journal of Food Science, 62, 781–788. Lonchampt, P., & Hartel, R. W. (2004). Fat bloom in chocolate and compound coatings. European Journal of Lipid Science and Technology, 106, 241–274.
15
Marty, S., Baker, K., Dibildox-Alvarado, E., Rodrigues, J. N., & Marangoni, A. G. (2005). Monitoring and quantifying of oil migration in cocoa butter using a flatbed scanner and fluorescence light microscopy. Food Research International, 38, 1189–1197. Marty, S., Baker, K., & Marangoni, A. G. (2009). Optimization of scanner imaging technique to accurately study oil migration kinetics. Food Research International, 42, 368–373. McCarthy, K. L., & McCarthy, M. J. (2008). Oil migration in chocolate–peanut butter paste confectionery as a function of chocolate formation. Journal of Food Science, 73, E266–E273. Miquel, M. E., Carli, S., Couzens, P. J., Wille, H.-J., & Hall, L. D. (2001). Kinetics of the migration of lipids in composite chocolate measured by magnetic resonance imaging. Food Research International, 34, 773–781. Radak, Z., Silye, G., Bartha, C., Jakus, J., Banyai, E. S., Atalay, M., et al. (2013). The effects of cocoa supplementation, caloric restriction, and regular exercise, on oxidative stress markers of brain and memory in the rat model. Food and Chemical Toxicology, 61, 36–41. Rasband, W.S. (1997–2016). ImageJ. Bethesda, Maryland, USA: U.S. National Institutes of Health. http://imagej.nih.gov/ij/. Reinke, S. K., Roth, S. V., Santoro, G., Vieira, J., Heinrich, S., & Palzer, S. (2015). Tracking structural changes in lipid-based multicomponents food materials due to oil migration by microfocus small-angle X-ray scattering. Applied Materials & Interface, 7, 9929–9936. Rousseau, D. (2006). On the porous mesostructure of milk chocolate viewed with atomic force microscopy. Lebensmittel-Wissenschaft Technology, 39, 852–860. Rumsey, T. R., & McCarthy, K. L. (2012). Modeling oil migration in two-layer chocolate-almond confectionery products. Journal of Food Engineering, 111, 149– 155. Sonwai, S., & Rousseau, D. (2010). Controlling fat bloom formation in chocolate— impact of milk fat on microstructure and fat phase crystallization. Food Chemistry, 119, 286–297. Talbot, G. (1996). The washer test: a method for monitoring fat migration. The Manufacturing Confectioner, 76, 87–90. Ziegleder, G., Moser, C., & Geier-Greguska, J. (1996a). Kinetics of fat migration within chocolate products, part I: principles and analytics: fett/lipid. European Journal of Lipid Science and Technology, 98, 196–199. Ziegleder, G., Moser, C., & Geier-Greguska, J. (1996b). Kinetics of fat migration within chocolate products. part II: influence of storage temperature, diffusion coefficient, solid fat content: fett/lipid. European Journal of Lipid Science and Technology, 98, 253–256. Ziegler, G. (2009). Product design and shelf-life issues: oil migration and fat bloom. In G. Talbot (Ed.), Science and technology of enrobed and filled chocolate, confectionery and bakery productsUK: Woodhead Publishing.