Journal of Food Engineering 92 (2009) 261268

Contents lists available at ScienceDirect

Journal of Food Engineering

Oil migration in chocolate: A case of non-Fickian diffusion

Sabic Innovative Plastics Mt. Vernon, Inc., Mt. Vernon, IN 47602, United States Department of Food Science, Pennsylvania State University, 341 Food Science Building, University Park, PA 16802, United States The Solae Company, St. Louis, MO 63188, United States d Department of Chemical Engineering, United States
b c

a r t i c l e

i n f o

a b s t r a c t
Oil migration through lled chocolate products during storage periods is inevitable and is responsible for the loss of the original quality of the product. A model that predicts the extent of oil migration can help food engineers optimize the post-production period, thereby delivering an acceptable quality product to consumers. In this work, a predictive model is proposed based on an explicit formulation of the diffusion problem in terms of the molecular diffusivity and the internal microstructure of chocolate. This work overcomes the limitations of the previous models by including a moving boundary, which allows for swelling of the chocolate slab and replacing the effective diffusion coefcient with a more physically meaningful one, which includes a tortuosity term that varies with oil concentration. The model has been validated against experimental mass uptake curves. 2008 Elsevier Ltd. All rights reserved.

Article history: Received 28 April 2006 Received in revised form 9 October 2008 Accepted 2 November 2008 Available online 28 November 2008 Keywords: Chocolate Oil migration Phase behavior Tortuosity Model

1. Introduction The chocolate confectionery market is dominated by composite products such as pralines, lled chocolates or snack bars, where a chocolate or confectionery coating layer is in direct contact with a fat-based cream, biscuit, nut or nut paste. The marketable lifetime of these products is often limited by fat or oil migration, e.g. the llings may contain large amounts of highly mobile oils, such as peanut or hazelnut oil that migrates into the chocolate coating causing quality loss (see Fig. 1). Quality defects arising from oil migration include softening of the coating, hardening of the lling, deterioration in sensory quality and a greater tendency to fat bloom. For this reason, oil migration has been extensively studied (Aguilera et al., 2004; Khan and Rousseau, 2006; Ziegleder et al., 1996a,b; Choi et al., 2007). 1.1. Mechanism of oil migration Dark chocolate is a mixture of cocoa butter, chocolate liquor, and sugar. Milk chocolate contains milk solids as well. Commercial dark chocolate typically contains between 25% and 40% cocoa butter with the rest being variable amounts of sugar and solid cocoa powder. Chocolate is a dispersion of solid cocoa particles, sugar crystals, and milk powder (in the case of milk chocolate) in a con* Corresponding author. Tel.: +1 814 863 2960; fax: +1 814 863 6132. E-mail address: grz1@psu.edu (G.R. Ziegler). Deceased. 0260-8774/$ - see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jfoodeng.2008.11.003

tinuous phase of solid and liquid fat, whose proportions depend on temperature (Ziegler et al., 2004). Cocoa butter comprises a mixture of saturated and unsaturated fats; oleic acid (O), stearic acid (S), and palmitic acid (P) account for more than 95% of the fatty acids in cocoa butter. Cocoa butter contains high levels of the triacylglycerols (TAGs) POS, POP, and SOS that are crystalline at normal storage temperatures; some of the unsaturated TAGs in cocoa butter have low melting points making it partly liquid at room temperature. The triacylglycerols in nut-based llings, which often contain hazelnut or almond oils, such as triolein (OOO), LOO, LLO, POO, and SOO, where L stands for lauric acid, are predominately liquid at room temperature. Originally, the driving force for diffusion was assumed to be a difference in liquid fat content, but recently diffusion has been attributed to a gradient in triacylglycerol concentration within some domains of the product (Ghosh et al., 2002). In a series of experiments where a bar of chocolate was brought into contact with a nougat lling, Ziegleder et al. (1996a,b) showed that after sufcient time the main triacylglycerols of hazelnut oil had almost evenly distributed between chocolate and lling. On the other hand, little of the triacylglycerols of the cocoa butter had migrated into the lling. Molecular or Fickian diffusion is widely used by food engineers as a general model for mass transfer. In the literature, there are several attempts to model fat migration using simplied solutions to Ficks Second Law of diffusion (Ziegleder et al., 1996a; Choi et al., 2007). As the rst approximation, Ziegleder et al. (1996a) employed the short-time solution with a constant diffusion

262

J.R. Galdmez et al. / Journal of Food Engineering 92 (2009) 261268

Nomenclature C FP t Deff D0 jv HZ K L(t)  L t  t rSF!LF ^ Vi wFAT HZ x hazelnut concentration in lter paper [kg/m3] effective diffusion coefcient [m2/s] diffusion coefcient in the liquid phase of the cocoa butter [m2/s] diffusivity ux of hazelnut oil with respect to the volume average velocity [kg/m2 s] partition distribution constant slab thickness expressed as a function of time [m] dimensionless slab thickness Time [s] dimensionless time rate of crystal dissolution [kg/m3 s] specic volume of species i [m3/g] weight fraction of hazelnut oil in the cocoa butter phase position [m] k n time lag [hours] immobilization variable mass concentration of hazelnut oil in the total chocolate [kg/m3] mass concentration of solid crystals in the total chocolate [kg/m3] mass concentration of hazelnut oil in the liquid cocoa butter [kg/m3] mass concentration established instantaneously at the surface at time zero [kg/m3] dimensionless mass concentration global tortuosity factor volume fraction of solid fat in the cocoa butter phase volume fraction of non-fat solid in chocolate volume fraction of liquid fat in chocolate volume average velocity

qT HZ qT SF qLF HZ qs0
 q s /Fat SF /T NF /T LF

Greek letters aspect ratio

mv

Fig. 1. Schematic representation of hazelnut oil migration into chocolate.

coefcient. Effective diffusion coefcients were extracted from the experimental data of the mass of the migrated oil plotted against the square root of time. Ziegleder et al. (1996b) observed a linear correlation between the logarithm of the effective diffusion coefcients so obtained and the liquid fat content in the chocolate. Models based on the so-called Fickian diffusion with a constant diffusivity fail to accurately describe the migration of oil into chocolate (Choi et al., 2007), and because of this some have suggested that a mechanism other than diffusion, e.g. capillary pressure, is responsible for oil migration. However, it is equally likely that simplied models fail to accurately reproduce experimental data because simplifying assumptions are not fullled. In addition to constant diffusivity, these simplied solutions neglect swelling and interaction between the oil and the cocoa butter. Recent observations of oil migration through chocolate using magnetic resonance imaging have shown that the dominant mechanism is diffusion, and that if capillary pressure is involved it is a minor contributor (Deka et al., 2006). Shetty (2004) proposed a new model that modied the effective diffusivity depending on the fat phase behavior. The idea behind this model is simple: the migrating oil disturbs the equilibrium between the liquid and solid phase present in the fat phase of choc-

olate. Oil migration decreases the solid fat content and thereby increases diffusion. Shettys (2004) model used the same expression for the relationship between the effective diffusion coefcient and the liquid fat content obtained by Ziegleder et al. (1996b). The difference in the Shetty (2004) model was that the liquid fat content was calculated independently through an experimental phase diagram. On the basis that the exact mechanism of oil migration in chocolate remains poorly understood, the present paper proposes a new model, based on molecular diffusion, to try to help elucidate the mechanism underlying fat migration. The model will account for structural parameters such as fat crystal microstructure and tortuosity and allows for chocolate swelling in order to overcome some of the drawbacks present in the previous models. 1.2. Phase behavior Most chocolate confectionery products contain between 18% and 40% cocoa butter. Cocoa butter, as with most natural fats, contains at least two phases, i.e. liquid and solid, whose proportions vary with temperature. Cocoa butter can crystallize in six different polymorphic forms, referred to as IVI, with b forms V and VI being

J.R. Galdmez et al. / Journal of Food Engineering 92 (2009) 261268

263

the most stable. In good quality chocolate products, the formation of form V crystals is enhanced during manufacturing by a process called tempering. During tempering numerous small fat crystals are promoted. The solid fat content (SFC) depends not only on temperature but also on the presence of other fats. Fats like peanut or hazelnut oil, common components in lling products, have high levels of low melting point TAGs that can dissolve the crystals existing in cocoa butter. At each temperature and composition, there is a maximum amount of the solid phase that is soluble in the liquid phase. This equilibrium phase behavior is described by a phase diagram. Here, a simplication is necessary because of the complexity of the fat composition. Although numerous triacylglycerol species are present in cocoa butter and the migrating oil, the phase behavior is often assumed to be a pseudo-binary system. Shetty (2004) constructed an experimental phase diagram that related the amount of solid phase in cocoa butter hazelnut oil mixtures to composition and temperature. At a given temperature, the weight fraction of solid can be related to the weight fraction of hazelnut oil in cocoa butter using the following relationships (Shetty, 2004):

Implicitly, it is assumed that partial specic volumes of the components are independent of composition so there is no volume change on mixing, and combined with assumptions 6 and 13, this leads to a zero volume average velocity throughout the chocolate slab and elimination of the convective terms in the continuity equations. Given these simplifying assumptions, the continuity equations for the relevant components such us hazelnut oil and solid fat can be written as

oqT oj HZ HZ ot ox oqT SF rSF!LF ot

2 3

wFat 0:554 1:190 wFat 0:638 wFat 2 SF HZ HZ wFat 0:653 1:041 wFat 0:375 SF HZ wFat SF 0:694 1:067 wFat HZ 0:342 wFat 2 HZ wFat 2 HZ

at 26  C at 23  C at 20 C


In these equations, qT is the mass concentration of the species i in i v the total volume, ji is the diffusive ux of species i, and r SF!LF is the rate of transformation of the solid fat to liquid fat due to the presence of hazelnut oil. Since hazelnut oil diffuses only in the liquid phase of chocolate, i.e. the liquid cocoa butter, it is necessary to express the ux of this component in Eq. (2) as a function of the concentration gradient of hazelnut oil in the liquid fat. Previous models incorrectly set the driving force as the gradient of HZ in the total chocolate. The expressions for the HZ ux in the liquid and the relationship between different diffusive uxes relative to the volume average velocity are

1.3. Model development 1.3.1. Model assumptions It was assumed that: 1. Chocolate comprised four species: liquid fat (LF), solid fat (SF), non-fat solids (NF), hazelnut oil (HZ). 2. The diffusion process was considered to be isothermal and one-directional. 3. Diffusion takes place only in the liquid phase of the cocoa butter and the driving force for diffusion is a gradient in hazelnut oil concentration in that phase. 4. Solids and fat crystals are impermeable to hazelnut oil and no adsorption occurs on their surfaces. 5. Local equilibrium between the solid and liquid phases, i.e. dissolution was fast in comparison with diffusion. 6. Vapor pressure of hazelnut oil is negligible and bloom, or crystal formation on the chocolates surface was neglected (none was observed). 7. Counter-diffusion of cocoa butter into the lter paper was negligible. 8. Solid fat content is temperature dependent according to the experimental phase diagram for the solid fat content in equilibrium with a certain amount of hazelnut oil (Shetty, 2004). 9. Specic volumes are constant and independent of concentration of HZ. 10. Species continuity equations can be formulated in a volumeaveraged velocity coordinate frame. 11. Diffusion of fat crystals and other non-fat solids was negligible. 12. A moving boundary can be used to accommodate the change in volume of the sample (jump mass balances are used to describe the movement of the boundary). 13. Convection due to the difference in density of the liquid and solid fat is negligible. 14. A constant aspect ratio for crystals was appropriate to formulate the tortuosity factor. The diffusion process within the chocolate slab was formulated relative to the volume average velocity (Cussler, 1997, pp. 5254).

oqLF v jHZ Deff HZ ox X v ^ ji V i 0

4 5

where Deff is the effective diffusion coefcient based on information regarding the internal microstructure of the chocolate. Modeling of diffusion through heterogeneous materials is not as obvious as one could anticipate. The presence of particles, whether they are fat crystals or non-fat solids, forces the migrating species to follow a longer path. Factors such as the geometry of the dispersed phase (particle shape, size and size distribution) contribute to a considerable change in the net effect of these particles (Cussler, 1997). To describe the effective diffusion in these complex materials two models are extensively studied in the literature: the Maxwell model (Maxwell, 1873), developed for a dilute dispersion of spheres, and the model to describe diffusion in the presence of asymmetric akes (DeRocher et al., 2004; Moggridge et al., 2002). Whereas the Maxwell model does not have any parameter regarding the size of the spheres, the ake model depends strongly on the ake aspect ratio, a, i.e. the width to thickness ratio. In this work, a new model accounting for tortuosity is proposed, where the tortuosity depends on the volume fraction of both the fat and non-fat solids present in the chocolate.

Deff D0 s D0 fT1 /T fT2 /Fat NF SF

Here, D0 is the diffusivity of HZ in the liquid fat and s is the effective tortuosity. s is described as the product of the tortuosity that the diffusing species sees by the presence of the non-fat solids, fT1, and the tortuosity that the diffusing species experiences within the fat phase due to fat crystals, fT2. Expressions for these two terms are

fT1

1 /T NF for spheres 1 0:5/T NF a/Fat 2 SF 1 /Fat SF for flakes

7 8

1 fT2 1

We assumed that fat crystals are more likely to resemble akes than perfect spheres. Both leaf-like structures of classic form VI crystals and needle-like crystals have been observed in chocolate

264

J.R. Galdmez et al. / Journal of Food Engineering 92 (2009) 261268

products (Marangoni, 2005). Trial predictions using the Maxwell model (spheres), consistently reached equilibrium signicantly faster than the measured data. For this reason, the ake model was used. The diffusion of HZ into chocolate changes the solid fat content, decreasing the amount of solid crystals and, consequently, increasing the effective diffusion coefcient. For a coupled system like this, it is important to consider the relative rates of diffusion of the oil throughout the liquid phase versus dissolution of the solid phase. If the dissolution of crystals is relatively fast compared to diffusion, then the solid and liquid phase can reach equilibrium quickly at each position as the oil penetrates. In those cases diffusion is the rate-limiting mechanism. However, if diffusion is fast enough, hazelnut oil can diffuse into the chocolate without a signicant change in the solid fat content due to dissolution. In that case, the diffusion process will be slowed because the tortuous path will remain longer as the solid fat content has a lower rate of dissolution. Unfortunately, there are not many studies on dissolution of fat crystals during migration to discern which mechanism is dominant. Ziegler et al. (2004) and Shetty (2004) suggest that a change in the rate-limiting mechanism, from dissolution to diffusion, occurs somewhere around 20 C. Below this temperature, they observed that chocolate was mostly solid and little dissolution occurs. A big drop in solid fat content was observed at higher temperature, suggesting that the activation energy for dissolution is close to the thermal energy available at those temperatures, and therefore, the dissolution process is relatively fast. As the rst approximation, the model will assume that diffusion is the rate-limiting mechanism, and, therefore, Eq. (3) can be re-written as

where qT t; x 0 is the equilibrium surface concentration given HZ by Eq. (12) or (13), qs0 is the concentration established instantaneously at the surface at time zero, and k is the characteristic time for the surface resistance. The model includes a moving boundary that allows the chocolate slab to expand as a result of the diffusion of HZ oil. The movement of the boundary was obtained from a jump mass balance at the chocolate interface (Duda and Vrentas, 1969):

  dLt D0 s dqLF v HZ m  dt /LF ox xLt

15

These equations were transformed into dimensionless form making use of the following dimensionless variables for concentration, temperature, time, and space.

 HZ qLF

qLF qLF t!0 HZ HZ qLF t!1 qLF t!0 HZ HZ

16 17 18

 tD0 t L2 t0  Lt L Lt0

In addition, we include a new variable, n, to immobilize the boundary as explained elsewhere (Duda and Vrentas, 1969)

x Lt

19

The nal equations to solve in dimensionless form are

q q

T SF

T FAT SF wHZ

where the weight fraction of fat crystals is related to temperature and hazelnut weight fraction by means of the relationships shown in Eq. (1) (see Appendix). 1.3.1.1. Initial and boundary conditions. Initially, it was assumed that chocolate contains no hazelnut oil

   HZ  HZ oqT n d oqT 1 o L  HZ oqLF D0 s   2 on o t t L dt o L on    HZ d L D0 s dqLF    L  /LF on  dt

20 21

qT t 0 0 HZ

10

For the boundary conditions, assumption 6 effectively means that there is no ux of HZ on the slab surface in contact with air so

 dqLF  HZ  0 dx xLt

A nite difference scheme was used to solve this set of differential equations. For the special case where the diffusion coefcient depends on concentration, an implicit CrankNicolson method is used to convert the Eqs. (20) and (21) into a set of non-linear algebraic equations. This method improves the accuracy and convergence of the solution procedure. The resulting equations were solved using the DNEQNF subroutine included in the IMSL library (Visual Numerics, http://www.vni.com/). 2. Materials and methods Lindt Bittersweet Thins (World Wide Chocolate, www.worldwidechocolate.com) and hazelnut oil (pure unrened oil manufactured from roasted hazelnuts) from Huilerie J. Leblanc (Gourmet Country, www.gourmetcountry.com) were used in the model system to represent chocolate products with nut-based lling. Chocolate samples weighed 2.84 0.36 g and were 3.8 cm in length and width, and 1.4 mm in thickness, allowing diffusion to be considered uni-directional since the thickness was much less than the length and width. Two different amounts of hazelnut oil (0.57 and 1.14 g) were used in the experiments one equivalent to the amount of fat found in the chocolate sample (1.14 g) and one containing half the amount of fat found in the chocolate sample (0.57 g). An empty Petri dish (disposable Fisherbrand 100 15 mm, Fisher Scientic, Fair Lawn, NJ) was weighed on a lab scale (Mettler Toledo AB104-S Scale, Switzerland) and tared. The desired amount of hazelnut oil was pipetted into the tared Petri dish. Filter papers were cut to 3.3125 in. circles to t entirely in the bottom of the Petri dishes. One piece of lter papers was then placed on top of the hazelnut oil inside the Petri dish. The Petri dish was rotated by hand in order to allow for an even distribution of hazelnut oil

11

For the side in contact with the HZ oil, several alternative boundary conditions were considered. The simplest one is to keep the concentration of HZ constant at the chocolate surface:

qT t; x 0 qT t 1 HZ HZ

12

This boundary condition assumes that the concentration at the chocolate surface immediately rises to the equilibrium value when it is contacted with hazelnut oil in the lter paper. A more general form for the boundary condition is setting the concentration at the surface in equilibrium with the concentration of hazelnut oil left in the lter paper:

qT t; x 0 KC FP t HZ

13

In this equation, K is the partition coefcient and CFP is the hazelnut oil concentration in the lter paper at any given time. Both equations neglect any resistance at the surface between the lter paper and chocolate. If the surface resistance becomes important, Eqs. (12) and (13) can be combined with an exponential model developed by Long and Richman (1960):

T HZ t s0 T t; x 0 HZ

qs0

1 ek

14

J.R. Galdmez et al. / Journal of Food Engineering 92 (2009) 261268

265

across the lter paper. Filter papers were used to ensure the chocolate sample would have a uniform supply of oil at the chocolateoil interface, rather than having the chocolate oat in the oil. In this model system, the lter paper represented the non-oil portion of a nut-based lling. One chocolate sample was weighed on the same lab scale and placed in the Petri dish at the center of the lter paper. The dishes were covered and placed in an incubator (Imperial III Model 310 Lab Line Instruments, Inc Melrose Park, IL) at two different temperatures (20 and 23 C). The samples were removed from the Petri dishes at regular time intervals, weighed and measured to record any dimensional changes, then placed back in the incubator. In order to obtain the weight of the individual chocolate sample, a sheet of weighing paper was tared on the lab scale. Petri dishes containing the samples were removed from the incubator and the chocolate samples were weighed one at a time by removing the dishs lid and carefully lifting one corner of the chocolate from the lter paper using a spatula. Then, the sample was ipped onto a piece of weighing paper so that the side of the chocolate that had been in contact with the hazelnut oil was facing upward. The weighing paper with the chocolate was weighed on the lab scale and returned to the Petri dish with the side in contact with the HZ returned to its original position. This continued until the weight gain of the sample was negligible or the integrity of the sample was compromised and it was impossible to remove the sample without physical damage. Each concentration-temperature combination was run in duplicate. Nuclear magnetic resonance (NMR) (Bruker Optics Minispec Analyzer, Model mq20) was used to obtain the diffusion coefcients for cocoa butter (Cadbury Trebor-Allan, Inc., Toronto, Canada), hazelnut oil, and mixtures of hazelnut oil and cocoa butter using the diffusio application (Minispec Software v2.57Rev.09/ NT/XP, Bruker Optics, Inc.). Cocoa butter was held at 60 C in an incubator (VWR Model 1325F, Sheldon Manufacturing, Inc.) for two weeks to ensure any solids would settle out. Six mixtures of cocoa butter and hazelnut oil (0%, 20%, 40%, 60%, 80% and 100%) were created to total 1 cm of height in each NMR tube (diameter = 1 cm) for a total mass of 0.56 g/sample. The desired amounts of hazelnut oil and cocoa butter were added to the tube and mixed on the Maxi Mix II (Barnstead/Thermolyne Model M37615). Measurements were taken at six temperatures (35, 40, 45, 50, 55 and 60 C) using the NMR temperature control system (Bruker N2, Bruker Optics, Inc.). The diffusion coefcient was measured six times for each sample at each temperature using a gradient amplitude of 90% (180 G/cm). D0 had a weak temperature dependence in the experimental range.

ings, develop fat bloom faster because lling oils have to migrate a smaller distance to reach the outside surface. In the same way, the larger proportion of fat in llings relative to chocolate increases the concentration gradient of fat and speeds up the process. Experiments were designed to study this effect and that of temperature. These data are shown in Figs. 25, where the solid symbols represent the average value of both replicates and the bars are the difference between the average and the experimental values. It is shown that by decreasing the lling ratio, i.e. having less hazelnut oil relative to cocoa butter, leads to a slower diffusion process. The parameters used in the simulations for Figs. 25 are summarized in Table 1. The model accurately predicted the experimental trends. The experimental mass uptake curves (Fig. 2) show an initial time lag period where no important weight gain is observed. The same type of S-shape curves has been observed in systems where the surface resistance becomes important (Nielsen and Hansenm, 2005). k was estimated experimentally from the uptake curves as the point where a back extrapolation line, from the middle points of the curve, crosses the x-axis and qs0 was zero for all the simulations. By combining Eqs. (12) and (14) for the boundary condition, the model can predict both uptakes curves with only one parameter, the aspect ratio (if the time lag is not considered an adjustable parameter). A constant aspect ratio of 15 can t the experimental

1.00

average cumulative weight gain [g]

0.80

= 15

0.60
= 17

0.40

0.20

0.00

0.0

5.0

10.0

15.0

20.0

25.0

30.0

time0.5 [hours0.5]
Fig. 2. Comparison between experimental mass uptake curves at 20 C (initial load of hazelnut oil in lter paper: N 1.14 g, d 0.57 g) and model predictions (solid lines). Eqs. (12) and (14) were used for the boundary condition.

Though the model can predict both the spatial and temporal HZ concentration, in the current analyses the output of this model is the total mass uptake of HZ to compare with the experimental data available. With minimal modications, however, the program can produce concentration proles of HZ oil or calculate the average solid fat content in order to reproduce the experimental data obtained from other experimental techniques, such as magnetic resonance imaging (MRI). First, the model was validated with the experimental data taken in this work and the effect of the boundary conditions used was examined. Second, the model is used to understand the effect of other variables such as temperature. Lastly, the model was used to produce concentration proles to demonstrate its potential for comparison with other experimental techniques not used in this work. One way to delay migration and fat bloom in confectionery products is to control the lling to chocolate ratio. High lling ratio products, those with a thin layer of chocolate surrounding the ll-

average cumulative weight gain [g]

3. Results

0.8
= 20

0.6

0.4

= 22

0.2

0.0 0.0

5.0

10.0

15.0

20.0

25.0

30.0

time0.5 [hours0.5]
Fig. 3. Comparison between experimental mass uptake curves at 20 C (initial load of hazelnut oil in lter paper: N 1.14 g, d 0.57 g) and model predictions (solid lines). Eqs. (13) and (14) were used for the boundary condition.

266

J.R. Galdmez et al. / Journal of Food Engineering 92 (2009) 261268

0.80

0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 0.0

= 7

= 9

5.0

10.0

15.0

20.0

time0.5 [hours0.5]
Fig. 4. Comparison between experimental mass uptake curves at 23 C (initial load of hazelnut oil in lter paper: j 1.14 g, e 0.57 g) and model predictions (solid lines). Eqs. (12) and (14) were used for the boundary condition.

average cumulative weight gain [g]

0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 0 5 10 15

T=26C T=23C T=21C

0.5

[hours ]

0.5

Fig. 5. Hazelnut mass uptake curves at different temperatures. All curves were obtained using a constant concentration at the hazelnut-chocolate boundary (Eq. (12)) with an aspect ratio of a = 5.

Table 1 Parameters used for the simulations carried out in this work. Fig. 2 20 1.14 2 1010 10 0.2 15 0.57 2 1010 25 0.12 17 Fig. 3 20 1.14 2 1010 10 0.7 20 0.57 2 1010 25 0.7 22 Fig. 4 23 1.14 2 1010 1.2 0.2 7 0.57 2 1010 9 0.11 9

T [C] Load of HZ [g] D0 [m/ s2] k [hour] w K

average values for the 1.14 g of hazelnut oil experiment, but for the lowest hazelnut amount a slightly better t is achieved with 17. For comparison, mica, a well-known material for composite membranes, has an aspect ratio of 30 (DeRocher et al., 2004). These numbers are only empirical and semi-quantitative because no information regarding the structure and alignment of the fat crystals was available, and are the models only tted parameter. Nevertheless, the values obtained are reasonable ones and some information about the crystal shape may be inferred.

The partition distribution coefcient, K, was obtained experimentally from the weight gain at equilibrium of the two experiments shown in Fig. 2. Fig. 3 shows the results from the model using these values of K as the input in conjunction with Eqs. (13) and (14). In order to t the data using this construction, higher values for the aspect ratio, a, where needed (a = 20 for the 1.14 g of HZ oil experiment and a = 22 for the 0.57 g). Higher values for a were expected since the concentration at the surface and, therefore, the concentration gradient, was bigger than the case using Eq. (12) alone. With Eq. (12), the model uses the lowest possible concentration at the surface during the whole process. In contrast, the boundary condition represented by Eq. (13) has a higher initial concentration at the surface, which decreases as the concentration of hazelnut in the lter paper decreases because it has migrated into the chocolate. Only at the end will both cases show the same concentration at the boundary. Increasing the temperature accelerated the diffusion process as is shown in Fig. 4. Experimental data reach the equilibrium plateau at shorter times than experiments at lower temperature, Fig. 2. The model, using Eqs. (12) and (14), was used to t the data. Lower values for the aspect ratio parameter were needed, decreasing from 15 to 7 for the experiment with 1.14 g or HZ and from 17 to 9 for the other set of experiments. Temperature has a two-fold effect on diffusion. First, it decreases the amount of crystals and, second, the remaining crystals are smaller with a lower aspect ratio, i.e. they are less asymmetric. It is notable that the tted aspect ratio, a, decreases with an increase in the temperature as anticipated. Given that the model provided an accurate representation of the experimental data with reasonable values of the only tted parameter, a, it was used to simulate migration under other conditions. Temperature is a critical factor during storage. A poor selection of storage or distribution temperature may accelerate the diffusion process and shorten the life of the chocolate product. To study this effect, the model was run at three temperatures, 26, 23 and 21 C. Eq. (12) was used to model the boundary condition without mass transfer resistance. Fig. 5 shows the mass uptake of hazelnut oil at these temperatures, with no time lag and the same nal concentration in each run. The lower the temperature the longer it takes to reach the equilibrium concentration. This is the result of coupling the phase equilibrium behavior of the fat phase with the diffusion model. At lower temperatures, more crystals are present in the fat phase and, therefore, increase the tortuosity of the diffusion path. A signicant drop in solid fat content occurs between 23 and 26 C (Shetty, 2004), which results in a much more rapid diffusion of oil into the chocolate at this temperature (Fig. 5). The model was used to predict liquid phase proles inside the chocolate wafers for qualitative comparison to published MRI data which can provide spatially resolved concentration proles (Deka et al., 2006; Choi et al., 2007). Figs. 6 and 7 show the model predictions. In both cases, the simulation was carried out at 21 C with a 5 mm thick slab of chocolate containing 40% of cocoa butter. Fig. 7 was obtained using a xed concentration at the surface, i.e. using Eq. (12) alone. In this case, there is an instantaneous change in the concentration at the boundary and it is kept at this value during the whole process. If Eqs. (13) and (14) are used instead, a continuous concentration built up at the surface was observed (Fig. 7). For both cases a sharp gradient is established similar to the ones found in the literature. Experimentally, Deka et al. (2006) have shown an increase in the concentration at the boundary between the chocolate and the oil with time, reinforcing the idea of a significant surface resistance during migration of oils when lter paper is used to hold the hazelnut oil. Therefore, surface resistance should be taken into consideration for analyzing any diffusion experiment, unless this effect has been reduced by choosing an adequate experimental set up.

average cumulative weight gain [g]

J.R. Galdmez et al. / Journal of Food Engineering 92 (2009) 261268

267

0.4

Weight fraction of liquid phase

0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 0 0.1 0.2 0.3 0.4 0.5

time = 1 h time = 10 h time = 30 h time = 50 h time = 100 h

0.6

0.7

position [cm]
Fig. 6. Concentration proles for the liquid phase content (HZ oil and liquid cocoa butter) at different times. The model uses a x boundary condition (Eq. (12)) with an aspect ratio, a = 15.

0.6

weight fraction of liquid phase

time = 1h

0.5 0.4 0.3 0.2 0.1 0

time = 10h time = 30h time = 50h

experimental measurement of swelling was not attempted. An estimation of the nal thickness of the sample was made from digital images. The model overestimated swelling by 3040%, which can be attributed to the fact that swelling occurred in all three dimensions in the experimental sample, i.e. the sample increased in length and width as well as thickness. However, the model allowed swelling only in thickness. The model is based on the assumption of local equilibrium between solid fat and the migrating oil. This assumption is rather controversial because it implicitly assumes that diffusion is the rate-limiting mechanism, although it may not be true, especially at low temperatures. Nevertheless, it is an important starting point for future models when the necessary information regarding the rates of dissolution of crystals are available. This work addresses the importance of selecting an appropriate boundary condition because most of the experimental work shows a signicant surface resistance and reinforces the idea that temperature not only changes the amount of fat crystals in the chocolate, but also their size and shape. This study has demonstrated that a diffusion-based model can adequately capture the migration of liquid oil through chocolate. No other mass transfer mechanisms are required. Further improvements would include a model that also predicts changes within the lling phase and might include interactions of the oil with the dispersed particulate solids.

Appendix The model developed in this manuscript solves the continuity equation coupled with a thermodynamic equilibrium between the different fat phases as described by Shetty (2004). To incorporate the equilibrium (Eq. (1)) into the constitutive equation (Eq. (4)) the following relationship was used. The mass concentration can be expressed in terms of the weight mass fraction based on the total chocolate mass as follows:

Time

time = 100h

0.1

0.2

0.3

0.4

0.5

0.6

0.7

position [cm]
Fig. 7. Concentration proles for the liquid phase content (HZ oil and liquid cocoa butter) at different times. The model uses Eqs. (12) and (14) for describing surface resistant at the boundary and an aspect ratio, a = 15.

qi P
j

wTotal i ^ wTotal V j j

A1

4. Discussion and conclusion Controlling migration of liquid oil into chocolate with lled product is critical for maintaining an acceptable quality. Models that predict the migration of external fats can be extremely useful for food engineers for optimizing the manufacturing process and guaranteeing the best quality of the nal product for consumers. Published models are highly empirical, and usually introduce some lump parameters encompassing all possible forms of mass transfer. A more general form of oil migration models will be of great use in optimizing or developing new processes. This work proposes a new model that represents an important improvement over existing ones. It is conceived with a more realistic formulation of the problem, where in the driving force for migration is the concentration gradient of liquid oils in the liquid phase of the chocolate. Mass transfer is described by molecular diffusivity, and the presence of solid particles which affect the diffusion through a tortuosity term that varies with oil concentration and temperature. With independent measurements of the equilibrium between solid fat crystals in cocoa butter and hazelnut oil the proposed model is able to predict, within experimental error, the mass pickup of a chocolate wafer exposed to a source of liquid oil. As oil migration proceeded, manipulation of the chocolate wafers without causing physical damage became difcult. Therefore,

Here, j is any of the four species in the model as stated in assumption 1, wTotal is the weight mass fraction with respect to the total j ^ mass, and V j is the specic volume of each species. The phase equilibrium between the hazelnut oil and the chocolate fat only involves the fat fraction of the chocolate. For this reason, the equilibrium relationship shown in Eq. (1) is expressed in terms of the weight fraction based on the total amount of fat. To incorporate the equilibrium into Eq. (A1) the following expressions were derived to relate the weight mass fraction based on the fat and on the total chocolate mass.

wFat HZ

wTotal HZ 1 1 FATwTotal 1 HZ

A2

Here, FAT is the initial fraction of fat in the chocolate. Similar expressions could be derived to relate the weight fraction of the others species in terms of wTotal , so that Eq. (A1) has only one HZ dependence. Finally, the volume fractions used in Eq. (6) can be calculated as follows:

^ wTotal V i /Total P i i Total ^ wj Vj

where j LF; SF; HZ and NF

A3

/Fat i

^ wFat V i P i ^ wFat V j j
j

where j LF; SF and HZ

A4

268

J.R. Galdmez et al. / Journal of Food Engineering 92 (2009) 261268 Long, F.A., Richman, D., 1960. Concentration gradients for diffusion of vapors in glassy polymers and their relation to time dependent diffusion phenomena. Journal of the American Chemical Society 82, 513519. Marangoni, A.G., 2005. Fat Crystal Networks. Marcel Dekker, New York. Maxwell, J.C., 1873. A Treatise on Electricity and Magnetism, vol. 1. Clarendon Press, Oxford. p. 365. Moggridge, G.D., Lape, N.K., Yang, C., Cussler, E.L., 2002. Barrier lms using akes and reactive additives. Organic Coatings 1239, 110. Nielsen, T.B., Hansenm, C.M., 2005. Signicance of surface resistance in absorption by polymers. Industrial and Engineering Chemistry Research 44, 39593965. Shetty, A., 2004. Oil Migration in Fat-based Confectionery. M.S. Thesis. The Pennsylvania State University. Ziegleder, G., Moser, C., Geiger-Greguska, J., 1996a. Kinetics of fat migration in chocolate products Part 1: Principles and analytical aspects. Fett/Lipid 98, 196 199. Ziegleder, G., Moser, C., Geiger-Greguska, J., 1996b. Kinetics of fat migration in chocolate products Part 2: Inuence of storage temperature, diffusion coefcient, solid fat content. Fett/Lipid 98, 253256. Ziegler, G.R., Shetty, A., Anantheswaran, R.C., 2004. Nut oil migration through chocolate. Manufacturing Confectioner 84 (9), 118126.

References
Aguilera, J.M., Michel, M., Mayor, G., 2004. Fat migration in chocolate: diffusion or capillary ow in a particle solid? A hypothesis paper. Journal of Food Science 69, 167174. Choi, Y.J., McCarthy, K.L., McCarthy, M.J., Kim, M.H., 2007. Oil migration in chocolate. Applied Magnetic Resonance 32, 205220. Cussler, E.L., 1997. Diffusion: Mass Transfer in Fluid Systems, second ed. Cambridge, UK. pp. 174175. Deka, K., MacMillan, B., Ziegler, G.R., Marangoni, A.G., Newling, B., Balcom, B., 2006. Spatial mapping of solid and liquid lipid in confectionery products using a 1D centric SPRITE MRI technique. Food Research International 39, 365371. DeRocher, J.P., Gettelnger, B.T., Wang, J., Nuxoll, E.E., Cussler, E.L., 2004. Barrier membranes with different sizes of aligned akes. Journal of Membrane Science 257, 2130. Duda, J.L., Vrentas, J.S., 1969. Perturbation solutions of diffusion-controlled moving boundary problems. Chemical Engineering Science 24, 461470. Ghosh, V., Ziegler, G.R., Anantheswaran, R.C., 2002. Fat, moisture, and ethanol migration through chocolate and confectionery coatings. Critical Reviews in Food Science and Nutrition 42, 583626. Khan, R., Rousseau, D., 2006. Hazelnut oil migration in dark chocolate-kinetic, thermodynamic and structural considerations. European Journal of Lipid Science and Technology 108 (5), 434443.