- Email: [email protected]
A novel technique for in situ measurements of stress development within a drying film
Journal of Food Engineering 92 (2009) 383–388
Contents lists available at ScienceDirect
Journal of Food Engineering journal homepage: www.elsevier.com/locate/jfoodeng
A novel technique for in situ measurements of stress development within a drying film Jianshe Chen a,*, Rammile Ettelaie a, Haiyan Yang b, Lin Yao a a b
School of Food Science and Nutrition, University of Leeds, Leeds, West Yorkshire LS2 9JT, UK College of Food Science, Xinjiang Agricultural University, 42 Nanchang Road, Urumqi, Xinjiang 830052, China
a r t i c l e
i n f o
Article history: Received 20 October 2008 Received in revised form 11 December 2008 Accepted 14 December 2008 Available online 25 December 2008 Keywords: Dehydration Film drying Internal stress Caseinate Waxy maize starch
a b s t r a c t A novel technique capable of monitoring the stress evolving process within a drying film has been developed in this work. The device uses a thin cantilever beam made of stainless steel to hold a thin layer of fluid material. The (bending) force exerted on the beam as a result of the lateral stress developed within the film is measured via a pendant-balance transmitting mechanism. The reliability of the device was tested with two different biopolymer fluids (18 wt% sodium caseinate and 30 wt% waxy maize starch) dried at different temperatures and relative humidity (RH). The results revealed three stages of stress development for these biopolymer films: an initial delay, a sharp stress increase, followed by a steady stress plateau. Further analysis of experimental data showed that stress development and moisture loss can be normalised to form a master curve. However, it appeared that the magnitude of stress increase has no direct link with the rate of moisture loss, indicating possibly different mechanisms of the stress increase and the loss of moisture. Results also showed that the total stress of a dried starch film was much higher than that of a dried caseinate film. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction Drying or dehydration is an operation commonly used in food industry to remove water (or the solvent) from a solid or slurry matrix for the purposes of either a decreased water activity for a longer shelf life, or a reduced weight for easy transportation, or simply for texture modification. The removal of water molecules involves simultaneous heat and mass transfer, regulating respectively the amount and rate of thermal energy supply and the amount and rate of solvent migration. The kinetics of drying or the rate of mass transfer is determined mainly by two sets of controlling parameters: the environmental conditions (e.g. temperature, relative humidity, air flow speed, air flow direction) and the material properties (e.g. surface properties, surface area-volume ratio, microstructure, and composition). The mechanisms and principles underpinning the drying process have been well studied and described in various textbooks (Barbosa-Cásanovas, 1996; Toledo, 1999). Even though controlling a drying process (the drying time and/ or the rate) can be achieved by using a particular drying device operated at specific drying conditions, some undesirable effects of drying (such as material deformation, shrinking, surface cracking, loss of glossiness) are rather difficult to control. These issues
* Corresponding author. Tel.: +44 113 3432748; fax: +44 0 1133432982. E-mail address: [email protected] (J. Chen). 0260-8774/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jfoodeng.2008.12.009
are often great concern to food manufacturers because of their detrimental effects on the visual appearance of a food product and consumers’ preference. The internal stress is believed to be the main cause of the above identified side effects. During a drying process, the moisture loss at the external surface will lead to moisture migration from the interior to the exterior surface and develop a moisture content gradient. Particles/molecules within the drier plane will have an intensified tendency of displacement and therefore a lateral stress develops. In order to release this stress, a material will have to either shrink from the outside or simply to crack (mostly at the surface). For a thin film, bending or curling is often an effective way for stress release. The exact rate and mechanisms for such stress development and release are not yet clear, but is generally believed to be material-dependent. De Gennes (2002) studied the drying of polymer films and indicated that a ‘‘crust” glassy surface was the direct result of solvent evaporation. He further argued that the crust is under a mechanical tension and could lead to some cracks. Thill and Spalla (2003) and Lee and Routh (2004) investigated the cracking of drying colloidal films and indicated more specifically that lateral capillary pressure, induced by packed particle fronts travelling horizontally across film, was responsible for the failure (crack) in these films. The occurrence of surface crack depends of course on the balance on the magnitude of the capillary force and the mechanical strength of the material itself. Using both numerical analysis and NMR technique for moisture diffusion, Augier et al. (2002) showed that the highest risk of surface cracking was at the beginning of a drying process,
384
J. Chen et al. / Journal of Food Engineering 92 (2009) 383–388
when the yield stress of the material was relatively low compared to the internal stress caused by moisture loss. In order to understand the evolving process of the internal stress (including the creation and the relaxation) within a drying material, various efforts have been made in developing a reliable technique to monitor and quantify such a process. One major progress was an in situ stress measurement apparatus developed a decade ago (Payne et al., 1997). They used a metal beam as the substrate coated with a thin layer of drying material on its top side. Since the beam was clamped firmly only on one end, it tended to bend slowly upwards as the lateral stress developed. The displacement or the deflection of the beam was monitored using a laser beam reflection technique and was used to quantify the magnitude of the internal stress of the drying film. This cantilever deflection principle was further modified with an improved accuracy and more sophisticated functionalities (Martinez and Lewis, 2002; Wedin et al., 2004). The improved device was not only able to measure in situ the internal stress but also the rate of moisture loss (or the drying kinetics). However, one potential problem of this set up is that the internal stress relaxes continuously as a result of film bending. This continuous relaxation makes it questionable whether the technique is capable to determine the maximum stress of a drying film and to observe surface cracking. This work aims to develop an alternative method to monitor the lateral stress development and to investigate effects of temperature and humidity on drying and stress development. The technique is still based on the cantilever principle, but the beam is restricted for free bending by a pendant acting on the free end. This pendant also functions as a force transmitting mechanism so that the bending force acting on the beam can be sensed and measured. The advantage of this design is that it allows only a minimal film deformation so that the maximum stress can be measured. In this paper, we present results from our initial investigation to demonstrate the feasibility of this novel method. 2. The apparatus The operation principles of self-developed stress apparatus for film drying is illustrated in Fig. 1. The whole device was housed inside a sealed transparent wooden-framed Perspex chamber with dimensions of 710 mm 445 mm 705 mm (length width height). The chamber was secured on the top of a solid bench to avoid any vibration and external disturbance. Inside the chamber,
a thick cast-iron plate topped with a layer of polished stainless still was used as the base, also for the benefit of anti-vibration stability. A stainless metal beam was used as the substrate for film coating and drying. The beam was clamped at one end on to a brass block, which sits on a lubricated sliding trough of the base so that it can be pushed and pulled smoothly to its designated position. The beam was made from a Starrett feeler gauge and has a thickness of 400 lm (Starrett Feeler Stock, Cat No. 667M, The L.S Starrett Company Ltd., Jedburgh, Scotland). The beam has a suspend length of 80 mm and a width of 12.6 mm. Near the free end of the beam, a small indent was made so that the tip of the pendant can rest on it with a precise position. This makes the effective length for film coating 73 mm, from the centre of indent to the clamping wall. In order to enhance the grip or adhesion of coating fluid, the surface of the beam was slightly and uniformly treated with a fine sand paper. A texture analyser (Stable Micro Systems, Surrey, UK) was used to determine the mechanical strength of the beam by measuring the bending force and the deflection displacement of the beam (shown in Fig. 2). The elastic modulus of the beam, E, can then be calculated using the following equation:
E ¼ RM=I;
ð1Þ
3
I ¼ bh =12
ð2Þ
where R is the radius of the bended beam, M is the moment of bending, I is the second moment of cross-section area, b and h are the width and the thickness of the beam. The radius of the bended beam can be estimated as an approximation using the following equation: 2
0
R ¼ l =2d
ð3Þ
where l is the effective length of the beam and d is the deflection distance of the beam. Taking the fact that the beam has an effective length of 73 mm, a width of 12.6 mm and a thickness of 0.4 mm, one can estimate that the beam should have an elastic modulus of 289.0 GPa, larger than the modulus for the one used in previous work (190 GPa) (Wedin et al., 2004). Beneath the indent, a small disc was attached and dipped into an oil reservoir to form a dashpot. This arrangement was found to be extremely useful in restraining beam vibration for a stable reading. The pendant has its own weight of 97.6423 g. The length of the pendant hook was adjustable so that it can be firmly hooked to the balance but only in a gentle touch with the beam. The net force/weight acting on the beam can be read from the difference of balance reading before and after touching the beam. Normally a net weight of around 1 to 1.5 g was applied to the beam, which makes sure that the tip of pendant is in firm touch with the beam but that the initial bending to the beam is also negligible.
0.25
Force (N)
0.2
y = 0.0991x R2 = 0.9957
0.15 0.1 0.05 0 -0.05 0
0.5
1
1.5
2
Deflection (mm) Fig. 1. A schematic illustration of the apparatus capable of measuring the stress development of a drying film.
Fig. 2. The deflection curve of the substrate beam used for film drying.
385
J. Chen et al. / Journal of Food Engineering 92 (2009) 383–388
Two biopolymer materials have been used in this investigation: sodium caseinate (Veghel, Netherlands) and Dexylose I231 (Roquette, France). The sodium caseinate powder was a spraydried, food-grade product with a minimum of 91% dry protein (based on Kjeldahl method using a Nitrogen coefficient of 6.38), a maximum moisture content of 5.0% (Gravimetric method at 105 °C), 0.04 wt% calcium, and less than 4.0% fat and ash (using Soxhlet extraction). Dexylose I231 was a phosphated waxy maize starch with a degree of substitution (DS) > 0.07 and, as a result, is cold water soluble. This starch contains 100% amylopectin and no amylose. According to the specification certificate provided by the manufacturer, the sample was in granular form and had a moisture content of 11.2% w/w (gravimetric method at 105 °C) and a particle size distribution of 99% <1000 lm and 65% >200 lm. The 18% w/w sodium caseinate solutions were prepared by adding 18 g of sodium caseinate powder into 82 g of mixture containing 81 g distilled water and 1 g of preservatives (1:1 of ActicdeR IPS 15 and Rocima 607 Microbicide). The mixture was gently stirred by a magnetic stirrer at a slow speed under a constant temperature of about 40 °C (which is the optimum temperature for solubility of sodium caseinate in water) for at least 6 h. The 30% w/w dexylose mixtures were prepared by adding 30 g of dexylose powder into a solution consisting 69 g distilled water and 1 g of the same preservatives. The dexylose samples were agitated moderately into a consistent texture for 2.5 h under room temperature. All the samples were used within a couple of days and disposed after.
4. Results and discussion 4.1. Stress development as a function of drying time Fig. 3 shows the results from four repeated experiments of film drying for18 wt% sodium caseinate fluid (3a) and their average (3b). The drying was conducted at a controlled temperature of 25 oC and within a small variation of relative humidity (between 30% and 35% RH). The force acting on the drying film was recorded as a function of time. A positive force means that the beam tends to bend upwards, indicating lateral stress acting on the drying film. Even though the four repeats showed certain degree of variation, the overall trend remained the same and the reproducibility was satisfactory with an acceptable standard deviation (3b) (<10%).
a Force Increase (g)
3. Materials
The main reason of using these two biopolymers was because of their wide usage in food applications as structure build agents and because of their consistency and flowability at such particular concentrations. Sodium caseinate is a water soluble milk protein widely used as an important functional ingredient for food and other applications because of its excellent structure-building or emulsifying capabilities and its physicochemical properties have been thoroughly investigated in literature (Wong et al., 1996; Fox and Brodkorb, 2008). At a concentration of 18 wt%, the fluid has an apparent viscosity close to 50 Pas (Chan et al., 2007). Similar viscosity profile was also observed for 30% waxy maize starch. Both samples were highly viscous and consistent while still flowable (Chan et al., 2007). This is important in making sure that the fluid can be easily and evenly coated to the beam surface but is still firm enough to hold its shape relatively well against the gravity, with little dripping or spreading during the experiment.
3 2.5 2 1.5 1 0.5 0 -0.5 0
10
20
30
40
50
60
40
50
60
Time (Min)
b
3.0 2.5
Average force (g)
The small gate on the left side wall was used for film coating. Initially the beam was pushed to the left end of the trough and a sliding gate with an opening of 0.4 mm height and 10.6 mm width rests on the beam. An appropriate amount of testing material (already thermally equilibrated to the experimental temperature) was carefully transferred to the front of the sliding gate and, by slowly pulling the brass block inwards, a thin coating film (0.4 mm 10.6 mm 73 mm) can be created on the top of the beam, leaving 1 mm edge space on both sides of coating film. Once the coated beam was in its position, the pendant was carefully put into its place to make a firm connection of the beam to the balance. The drying process is then monitored by recording the data for every minute from the balance. It should be noted that, in order to maintain reproducibility, a two minute delay was given to each test after the completion of film coating to the start (time zero) of force recording. The chamber was fitted with two sets of heating elements. The temperature inside the chamber can then be controlled by the thermal probe in the centre of the chamber. A small and slow rotating fan was used to create a very gentle air flow to help maintain thermal and humidity homogeneity of air without causing any disturbance to the pendant. The closed chamber has not yet been equipped with a humidity control device. However, a separate thermo-hydrometer was placed inside the chamber to record the relative humidity as well as the experimental temperature. This device was also used to monitor the kinetics of moisture loss of the film. A film was coated on a metal beam using the same technique. The beam was attached to a brass cube for easy handling. After coating, the beam was quickly relocated into the balance and the total weight of the metal beam and the film was recorded at 30 s intervals after the chamber was sealed. Same as stress development tests, moisture loss tests were also conducted under a controlled temperature and a monitored relative humidity. All experiments were repeated for at least three times for reproducibility.
2.0 1.5 1.0 0.5
TM
0.0 -0.5
0
10
20
30
Time (Min) Fig. 3. Reproducibility tests of the drying stress for 18 wt% sodium caseinate films. The force (stress) increase of the drying film was recorded as a function of time. (a) Repeats of four experiments conducted at 25 ± 1 oC and relative humidity of 30 ± 2% and (b) an average of the four experiments.
J. Chen et al. / Journal of Food Engineering 92 (2009) 383–388
system may reach a frozen state. At this stage, the amount (and the rate) of moisture loss becomes very limited and, as a result of this, there should be very limited stress increase, but the film becomes very rigid and solid enough to support the stress (Lei et al., 2002). The relaxation for such a film approaches infinite long and the stress will remain unless a cracking or physical bending to release the stress. This observation may contradict to the conclusion made by Augier et al. (2002), who claimed that surface cracking occurs most likely at the beginning of a drying process because of the low yield stress of the material. However, with no detectable lateral stress, it is hard to believe that cracking could occur at the first phase of drying. It is more likely that material deformation (shrinking or cracking) to occur during the phase II of drying, where the lateral stress becomes significantly large but the yield stress of the material remains low. 4.2. Effects of drying temperature and relative humidity It is well-known that temperature and relative humidity of the drying air are the two main controlling factors for dehydration. Higher temperature and lower relative humidity are often favoured for a faster moisture loss. However, it is rather not clear how these two factors influence the stress development with the drying system. For this purpose, the stress development of concentrated sodium caseinate and waxy maize starch systems has also been examined as affected by the temperature and relative humidity. The recorded stress developments are shown in Fig. 4 for the
Force Increase (g)
a
100% 4.5 80% 3.5 60%
2.5
40%
1.5
20%
0.5
Remaining Mass (%)
0%
-0.5 0
10
20
30
40
50
60
Time (Min)
b
7
100% 80%
5 60% 3 40% 1
20%
-1
Remaining Mass (%)
Contributions to the experimental error could be from a number of factors, but we recommend that the control of relative humidity could be the main contributing source. The difference of the relatively humidity at the film surface and of the drying air is the driving force of mass transfer and determines the rate of moisture loss. Unfortunately, there was no internal control of relative humidity in the current experimental set up. The relative humidity of each experiment was determined by the atmospheric condition of the lab. Even though 5% difference in relative humidity was probably the best we could wait for, its contribution to experimental error should not be ignored. Further improvement plan is currently undergoing to add a regulated air circulation system to the chamber which will give a properly controlled temperature and relative humidity. The thickness and evenness of coated film is another important factor influencing the accuracy and reproducibility of experiments. Currently the film coating is done manually by pulling the cantilever beam through the coating gate. Numerous coating practices have to be conducted to make sure that the pulling is conducted at an even and constant speed when the beam is slid through the coating gate. The force curves in Fig. 3 suggest that the stress development for a drying caseinate film can be divided roughly into three phases. At the beginning of the drying (phase I), the dehydration appears to lead no detectable stress to the film. This phase lasts for about 12 min after drying starts. However, it should be noted that after the completion of film coating to the start of recording, there was a delay of 2 min. This delay was necessary for the transfer of the beam from the coating gate to its drying position and for the set up of pendant. This was also necessary to make sure that every drying test has the same starting point. In phase II, the lateral stress starts to show a relentless increase with time, a sharp and almost a linear increase over the next 15 min or so. This could be the stage of most drastic changes to the film in terms of its material properties. Huge internal stress increase within a short period of time may cause various changes to a material, such as internal flow, yielding, bending, and of course possibly surface cracking. At phase III, the stress starts to level off and remains at a high but constant value. This phase tends to last for many hours or even overnight. There were one or two occasions that the coating film partly peeled off from the substrate and gave no detected force (stress). Fig. 3b gives the average of four experiments. We can see that the bending force of the drying film acting on the beam reaches its maximum of around 2.5 g (or 24.5 mN) and levels off after about half hour drying. In analysing the three phases stress development, one should note that the internal stress development is always accompanied by a stress relaxation process which tends to release stress by displacing particles and molecules within the network. Therefore, the recorded stress should be the combined effect of the two counteracting processes. While the rate of stress increase depends on the magnitude of moisture content gradient or the amount of moisture loss, the rate of stress relaxation is determined in principle by the microstructural nature of the material. Internal stress develops only if the rate of stress increase is faster than that of stress release. If the rate of stress relaxation exceeds the rate of stress increase, no internal stress becomes observable. From Fig. 3, one can say that the stress relaxation at phase I is very fast. This could be partly because that the material at this stage is still flowable and highly flexible for fast stress dissipation and partly because that the moisture content gradient is still small due to fast moisture migration from the interior to the surface. It was believed that, at this stage, the fluid coating is liquid enough that any shrinkage stress is rapidly relieved by viscous flow (Lei et al., 2002). However, the two processes changes its places in the second phase, where the moisture content gradient is well established but the stress relaxation becomes limited. In phase III, a constant stress suggests that the
Force Increase (g)
386
0% 0
10
20
30
40
50
60
Time (Min) Fig. 4. The force increase and the remaining mass as a function of drying time at different drying conditions. (a) Drying of 18 wt% sodium caseinate (s: force increase at 25 oC and 30% RH; d: force increase at 30 oC and 21% RH; D: remaining mass at 25 oC and 30% RH; N remaining mass at 30 oC and 21% RH) and (b) drying of 30 wt% waxy maize starch (s: force increase at 25 oC and 33% RH; d: force increase at 25 oC and 24% RH; D: retaining mass at 25 oC and 33% RH; N retaining mass at 25 oC and 24% RH).
387
J. Chen et al. / Journal of Food Engineering 92 (2009) 383–388
4.3. Normalisation of stress and moisture loss A time factor has been used to normalise the stress and moisture loss of the films dehydrated at different atmospheric conditions. For example, for the pair of drying curves shown in Fig. 4a for 18% caseinate films, both stress content curves were normalised as a fraction of their maximum stress values and the time scales were shifted by the normalisation time factor, Kn. In this case, a time factor of 0.84 was used (Fig. 5a) and we can see that both stress curves are effectively merged into one single master curve. Interestingly, this same time factor was also applicable for the normalisation of the remaining mass curves (Fig. 5a). Same normalisation approach has also been found applicable for waxy maize starch films dehydrated at different relative humidity (Fig. 4b). It was found that a time factor of 0.93 was applicable for the normalisation of both stress development and the remaining mass for the starch system (Fig. 5b).
Normalised Fraction
a
1.2 1 0.8 0.6 0.4 0.2 0 0
10
20
30
40
50
60
50
60
Normalised Time (Min)
b Normalised Fraction
caseinate sample dried at two different temperatures (4a) and for the starch sample dried two different relative humidities (4b). The temperature change was regulated by the thermal control system, but the relative humidity change was achieved by conducting the experiment at different weather conditions. Overall the stress development showed very similar pattern for the two biopolymer films dried at different conditions. However, the results clearly showed to our expectation that a higher drying temperature and a lower relative humidity led to an early onset of the stress increase and a quicker reach of the maximum stress. For example, the onset of stress increase for the caseinate film was shortened from 12.3 min at 25 oC to less than 10 min at 30 oC, while the time to reach the maximum was also decreased from around 27.5 min at 25 oC to around 24 min at 30 oC (Fig. 4a). Similar pattern was also clearly observable for the starch film when the relative humidity was decreased from 33% to 24% (Fig. 4b). The main reason for the quick onset of the stress increase at a higher temperature or a low relative humidity is probably because of the fast moisture loss under these conditions, which leads to an early set of the moisture gradient. Another significant impact of the temperature and humidity of drying air was seen on the magnitude of the maximum stress. It was found that an increase of five degree in drying temperature almost doubled the internal stress of the film. The final force recorded for the caseinate film was around 2.5 g when it was dried at 25 oC, but this was increased to higher than 4.6 g at the drying temperature of 30 oC. This is probably due to the fact that a fast moisture loss (as a result of higher drying temperature and/or drier air) leads to a quicker set up of moisture content gradient, which gives a faster stress increase and also at the same time a much slowed relaxation process (or an increase relaxation time) of the film. However, the change of 9% relative humidity seemed to have a much smaller effect than that of 5 oC. The maximum stress for the starch film increased from around 4.9 g at 33% relative humidity to around 6.2 g at 24% relative humidity, an increase of around 26%. Fig. 4 also gives the percentage of remaining mass as a function of drying time. These tests were conducted separately under the same experimental conditions (see details in apparatus section) and the percentage of remaining mass was calculated as the ratio of sample weight recorded at time t against its initial weight. It can be seen clearly that a higher temperature leads to a faster loss of moisture (Fig. 4a); while a decrease of relative humidity shows a very limited effect on the rate of moisture loss (Fig. 4b). However, it is worth to note that, for both varied temperature and varied relative humidity, the remaining mass reaches to its final steady value almost at the same time as the stress reaches its maximum plateau for both biopolymer systems.
1.2 1 0.8 0.6 0.4 0.2 0 0
10
20
30
40
Normalised Time (Min) Fig. 5. Normalised stress development curve and remaining mass curves from Fig. 4a a normalisation factor of 0.84 was used for the normalisation of sodium caseinate films dried at two different temperatures as shown in Fig. 4a (s: force increase at 25 oC and 30% RH; d: force increase at 30 oC and 21% RH; D: remaining mass at 25 oC and 30% RH; N remaining mass at 30 oC and 21% RH) and (b) a normalisation factor of 0.93 was used for the normalisation of waxy maize starch films dried at two different relative humidity as shown in Fig. 4b (s:force increase at 25 oC and 33% RH; d: force increase at 25 oC and 24% RH; D: remaining mass at 25 oC and 33% RH; N remaining mass at 25 oC and 24% RH).
From the normalisation analysis, it is clear that the stress increase and the amount of moisture loss within a drying film are coherently relevant if not directly linked. A single time factor for each sample dried at two different conditions suggests that the drying of the film undergoes through a very same pathway and is probably controlled by the same mechanisms. It is also worth to indicate that the magnitude of the normalisation factor may reflect different impacts of drying conditions. A value further away from a unit may suggest a greater impact of the drying condition. This means that 5 oC temperature variation (Fig. 5a) may have a greater influence on drying dynamics than 10% relative humidity change (Fig. 5b). The former has a time factor of 0.84, while the latter has a time factor of 0.93. 4.4. Relationship to the rate of film dehydration A dehydration process is often divided into different stages based on the rate of moisture loss: constant rate, first falling rate, second falling rate, etc. (Toledo, 1999). Fig. 6 shows the rate of dehydration for the 18% sodium caseinate film, where the rate of moisture loss was calculated as the amount of moisture loss per unit of time and per unit amount of dry matter. We can see that the rate of moisture loss in this case can be roughly categorized into three different stages: first falling rate, second falling rate, and finally levelling off with a minimal drying rate. Factors and mechanisms controlling the rate of moisture loss have been well studied in the literature (Srikiatden and Roberts, 2007) and are not main concern of this work. Our concern was whether the stress development within a drying film has any correlation with the rate
J. Chen et al. / Journal of Food Engineering 92 (2009) 383–388
3.0
0.18
2.5
0.15
2.0
0.12
1.5
0.09
1.0
0.06
0.5
0.03
0.0
0
-0.5
Drying rate
Force increase (g)
388
-0.03 0
10
20
30
40
50
60
Time (Min) Fig. 6. The stress development (D) and the rate of dehydration (s) as a function of drying time (18 wt% sodium caseinate films dried at 25 ± 1 oC and 30 ± 2% RH). The results were the average of four experiments.
of moisture loss. From the results shown in Fig. 6, one can see that the three phases of stress development and the three stages of dehydration are in rather different time scales. There was no stress increase when the moisture loss was in the highest rate at the beginning of drying. The most significant stress increase occurred at the later stage of first falling rate and early stage of second falling rate. Therefore, one could conclude that, even though the stress development within a drying film is the direct result of moisture loss, its evolving process has no direct correlation with the rate of moisture loss. This may be because that the rate of moisture loss is more or less an instantaneous property of a material and has little dependence on its previous rate, but the stress increase within a film is an integrated effect and could depend highly on the whole drying history. 5. Conclusions A novel technique has been developed to monitor the stress development of a fluid film during its drying process. This technique directly measures the bending force exerting on the substrate beam and is easy to operate. Two biopolymer fluids have
been tested using this methodology: 18 wt% sodium caseinate and 30 wt% waxy maize starch, dried at two different temperatures (25 and 30 oC) and two different relative humidities (24% and 33%). It was found that the stress development of these systems showed very similar pattern: an initial delay, a sharp stress increase, followed by a stress plateau. The normalisation of stress increase and amount of moisture loss revealed same time factor for two processes conducted at different atmospheric conditions. It was also observed that the stress development of a biopolymer film had no direct relevance to the rate of moisture loss. Acknowledgement Authors wish to acknowledge partial financial support from Food Processing Faraday (UK) for this Project. References Augier, F., Coumans, W.J., Hugget, A., Kaasschieter, E.F., 2002. On the risk of cracking in clay drying. Chemical Engineering Journal 86, 133–138. Barbosa-Cásanovas, G.V., 1996. Dehydration of Foods. Chapman & Hall, New York. Chan, P.S.-K., Chen, J., Ettelaie, R., Law, Z., Alevisopoulos, S., Day, E., Smith, S., 2007. Shear and extensional rheology of casein and starch mixtures: effects of concentration and temperature. Food Hydrocolloids 21, 716–725. De Gennes, P.G., 2002. Solvent evaporation of spin cast films: crust effects. European Physical Journal E 7, 31–34. Fox, P.F., Brodkorb, A., 2008. The casein micelle: historical aspects, current concepts and significance. International Dairy Journal 18, 677–684. Lee, W.P., Routh, A.F., 2004. Why do drying films crack? Langmuir 20, 9885–9888. Lei, H., Francis, L.F., Gerberich, W.W., Scriven, L.E., 2002. Stress development in drying coatings after solidation. AICHE Journal 48, 437–451. Martinez, C.J., Lewis, J.F., 2002. Shape evolution and stress development during latex-silica film formation. Langmuir 18, 4689–4698. Payne, J.A., McCormick, A.V., Francis, L.F., 1997. In situ stress measurement apparatus for liquid applied coatings. Review of Scientific Instruments 68, 4564–4568. Srikiatden, J., Roberts, J.S., 2007. Moisture transfer in solid food materials: a review of mechanisms, models, and measurements. International Journal of Food Properties 10, 739–777. Thill, S., Spalla, O., 2003. Aggregation due to capillary forces during drying of particle submonolayers. Colloids and Surfaces A 217, 143–151. Toledo, R.T., 1999. Fundamentals of Food Process Engineering, second ed. An Aspen Publication, Gaithersburg, Maryland. Wedin, P., Martinez, C.J., Lewis, J.A., Daicic, J., Berström, L., 2004. Stress development during drying of calcium carbonate suspensions containing carboxymethylcellulose and latex particles. Journal of Colloid and Interface Science 272, 1–9. Wong, D.W.S., Camirand, W.M., Pavlath, A.E., 1996. Structures and functionalities of milk proteins. Critical Reviews in Food Science and Nutrition 36, 807–844.
Contents lists available at ScienceDirect
Journal of Food Engineering journal homepage: www.elsevier.com/locate/jfoodeng
A novel technique for in situ measurements of stress development within a drying film Jianshe Chen a,*, Rammile Ettelaie a, Haiyan Yang b, Lin Yao a a b
School of Food Science and Nutrition, University of Leeds, Leeds, West Yorkshire LS2 9JT, UK College of Food Science, Xinjiang Agricultural University, 42 Nanchang Road, Urumqi, Xinjiang 830052, China
a r t i c l e
i n f o
Article history: Received 20 October 2008 Received in revised form 11 December 2008 Accepted 14 December 2008 Available online 25 December 2008 Keywords: Dehydration Film drying Internal stress Caseinate Waxy maize starch
a b s t r a c t A novel technique capable of monitoring the stress evolving process within a drying film has been developed in this work. The device uses a thin cantilever beam made of stainless steel to hold a thin layer of fluid material. The (bending) force exerted on the beam as a result of the lateral stress developed within the film is measured via a pendant-balance transmitting mechanism. The reliability of the device was tested with two different biopolymer fluids (18 wt% sodium caseinate and 30 wt% waxy maize starch) dried at different temperatures and relative humidity (RH). The results revealed three stages of stress development for these biopolymer films: an initial delay, a sharp stress increase, followed by a steady stress plateau. Further analysis of experimental data showed that stress development and moisture loss can be normalised to form a master curve. However, it appeared that the magnitude of stress increase has no direct link with the rate of moisture loss, indicating possibly different mechanisms of the stress increase and the loss of moisture. Results also showed that the total stress of a dried starch film was much higher than that of a dried caseinate film. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction Drying or dehydration is an operation commonly used in food industry to remove water (or the solvent) from a solid or slurry matrix for the purposes of either a decreased water activity for a longer shelf life, or a reduced weight for easy transportation, or simply for texture modification. The removal of water molecules involves simultaneous heat and mass transfer, regulating respectively the amount and rate of thermal energy supply and the amount and rate of solvent migration. The kinetics of drying or the rate of mass transfer is determined mainly by two sets of controlling parameters: the environmental conditions (e.g. temperature, relative humidity, air flow speed, air flow direction) and the material properties (e.g. surface properties, surface area-volume ratio, microstructure, and composition). The mechanisms and principles underpinning the drying process have been well studied and described in various textbooks (Barbosa-Cásanovas, 1996; Toledo, 1999). Even though controlling a drying process (the drying time and/ or the rate) can be achieved by using a particular drying device operated at specific drying conditions, some undesirable effects of drying (such as material deformation, shrinking, surface cracking, loss of glossiness) are rather difficult to control. These issues
* Corresponding author. Tel.: +44 113 3432748; fax: +44 0 1133432982. E-mail address: [email protected] (J. Chen). 0260-8774/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jfoodeng.2008.12.009
are often great concern to food manufacturers because of their detrimental effects on the visual appearance of a food product and consumers’ preference. The internal stress is believed to be the main cause of the above identified side effects. During a drying process, the moisture loss at the external surface will lead to moisture migration from the interior to the exterior surface and develop a moisture content gradient. Particles/molecules within the drier plane will have an intensified tendency of displacement and therefore a lateral stress develops. In order to release this stress, a material will have to either shrink from the outside or simply to crack (mostly at the surface). For a thin film, bending or curling is often an effective way for stress release. The exact rate and mechanisms for such stress development and release are not yet clear, but is generally believed to be material-dependent. De Gennes (2002) studied the drying of polymer films and indicated that a ‘‘crust” glassy surface was the direct result of solvent evaporation. He further argued that the crust is under a mechanical tension and could lead to some cracks. Thill and Spalla (2003) and Lee and Routh (2004) investigated the cracking of drying colloidal films and indicated more specifically that lateral capillary pressure, induced by packed particle fronts travelling horizontally across film, was responsible for the failure (crack) in these films. The occurrence of surface crack depends of course on the balance on the magnitude of the capillary force and the mechanical strength of the material itself. Using both numerical analysis and NMR technique for moisture diffusion, Augier et al. (2002) showed that the highest risk of surface cracking was at the beginning of a drying process,
384
J. Chen et al. / Journal of Food Engineering 92 (2009) 383–388
when the yield stress of the material was relatively low compared to the internal stress caused by moisture loss. In order to understand the evolving process of the internal stress (including the creation and the relaxation) within a drying material, various efforts have been made in developing a reliable technique to monitor and quantify such a process. One major progress was an in situ stress measurement apparatus developed a decade ago (Payne et al., 1997). They used a metal beam as the substrate coated with a thin layer of drying material on its top side. Since the beam was clamped firmly only on one end, it tended to bend slowly upwards as the lateral stress developed. The displacement or the deflection of the beam was monitored using a laser beam reflection technique and was used to quantify the magnitude of the internal stress of the drying film. This cantilever deflection principle was further modified with an improved accuracy and more sophisticated functionalities (Martinez and Lewis, 2002; Wedin et al., 2004). The improved device was not only able to measure in situ the internal stress but also the rate of moisture loss (or the drying kinetics). However, one potential problem of this set up is that the internal stress relaxes continuously as a result of film bending. This continuous relaxation makes it questionable whether the technique is capable to determine the maximum stress of a drying film and to observe surface cracking. This work aims to develop an alternative method to monitor the lateral stress development and to investigate effects of temperature and humidity on drying and stress development. The technique is still based on the cantilever principle, but the beam is restricted for free bending by a pendant acting on the free end. This pendant also functions as a force transmitting mechanism so that the bending force acting on the beam can be sensed and measured. The advantage of this design is that it allows only a minimal film deformation so that the maximum stress can be measured. In this paper, we present results from our initial investigation to demonstrate the feasibility of this novel method. 2. The apparatus The operation principles of self-developed stress apparatus for film drying is illustrated in Fig. 1. The whole device was housed inside a sealed transparent wooden-framed Perspex chamber with dimensions of 710 mm 445 mm 705 mm (length width height). The chamber was secured on the top of a solid bench to avoid any vibration and external disturbance. Inside the chamber,
a thick cast-iron plate topped with a layer of polished stainless still was used as the base, also for the benefit of anti-vibration stability. A stainless metal beam was used as the substrate for film coating and drying. The beam was clamped at one end on to a brass block, which sits on a lubricated sliding trough of the base so that it can be pushed and pulled smoothly to its designated position. The beam was made from a Starrett feeler gauge and has a thickness of 400 lm (Starrett Feeler Stock, Cat No. 667M, The L.S Starrett Company Ltd., Jedburgh, Scotland). The beam has a suspend length of 80 mm and a width of 12.6 mm. Near the free end of the beam, a small indent was made so that the tip of the pendant can rest on it with a precise position. This makes the effective length for film coating 73 mm, from the centre of indent to the clamping wall. In order to enhance the grip or adhesion of coating fluid, the surface of the beam was slightly and uniformly treated with a fine sand paper. A texture analyser (Stable Micro Systems, Surrey, UK) was used to determine the mechanical strength of the beam by measuring the bending force and the deflection displacement of the beam (shown in Fig. 2). The elastic modulus of the beam, E, can then be calculated using the following equation:
E ¼ RM=I;
ð1Þ
3
I ¼ bh =12
ð2Þ
where R is the radius of the bended beam, M is the moment of bending, I is the second moment of cross-section area, b and h are the width and the thickness of the beam. The radius of the bended beam can be estimated as an approximation using the following equation: 2
0
R ¼ l =2d
ð3Þ
where l is the effective length of the beam and d is the deflection distance of the beam. Taking the fact that the beam has an effective length of 73 mm, a width of 12.6 mm and a thickness of 0.4 mm, one can estimate that the beam should have an elastic modulus of 289.0 GPa, larger than the modulus for the one used in previous work (190 GPa) (Wedin et al., 2004). Beneath the indent, a small disc was attached and dipped into an oil reservoir to form a dashpot. This arrangement was found to be extremely useful in restraining beam vibration for a stable reading. The pendant has its own weight of 97.6423 g. The length of the pendant hook was adjustable so that it can be firmly hooked to the balance but only in a gentle touch with the beam. The net force/weight acting on the beam can be read from the difference of balance reading before and after touching the beam. Normally a net weight of around 1 to 1.5 g was applied to the beam, which makes sure that the tip of pendant is in firm touch with the beam but that the initial bending to the beam is also negligible.
0.25
Force (N)
0.2
y = 0.0991x R2 = 0.9957
0.15 0.1 0.05 0 -0.05 0
0.5
1
1.5
2
Deflection (mm) Fig. 1. A schematic illustration of the apparatus capable of measuring the stress development of a drying film.
Fig. 2. The deflection curve of the substrate beam used for film drying.
385
J. Chen et al. / Journal of Food Engineering 92 (2009) 383–388
Two biopolymer materials have been used in this investigation: sodium caseinate (Veghel, Netherlands) and Dexylose I231 (Roquette, France). The sodium caseinate powder was a spraydried, food-grade product with a minimum of 91% dry protein (based on Kjeldahl method using a Nitrogen coefficient of 6.38), a maximum moisture content of 5.0% (Gravimetric method at 105 °C), 0.04 wt% calcium, and less than 4.0% fat and ash (using Soxhlet extraction). Dexylose I231 was a phosphated waxy maize starch with a degree of substitution (DS) > 0.07 and, as a result, is cold water soluble. This starch contains 100% amylopectin and no amylose. According to the specification certificate provided by the manufacturer, the sample was in granular form and had a moisture content of 11.2% w/w (gravimetric method at 105 °C) and a particle size distribution of 99% <1000 lm and 65% >200 lm. The 18% w/w sodium caseinate solutions were prepared by adding 18 g of sodium caseinate powder into 82 g of mixture containing 81 g distilled water and 1 g of preservatives (1:1 of ActicdeR IPS 15 and Rocima 607 Microbicide). The mixture was gently stirred by a magnetic stirrer at a slow speed under a constant temperature of about 40 °C (which is the optimum temperature for solubility of sodium caseinate in water) for at least 6 h. The 30% w/w dexylose mixtures were prepared by adding 30 g of dexylose powder into a solution consisting 69 g distilled water and 1 g of the same preservatives. The dexylose samples were agitated moderately into a consistent texture for 2.5 h under room temperature. All the samples were used within a couple of days and disposed after.
4. Results and discussion 4.1. Stress development as a function of drying time Fig. 3 shows the results from four repeated experiments of film drying for18 wt% sodium caseinate fluid (3a) and their average (3b). The drying was conducted at a controlled temperature of 25 oC and within a small variation of relative humidity (between 30% and 35% RH). The force acting on the drying film was recorded as a function of time. A positive force means that the beam tends to bend upwards, indicating lateral stress acting on the drying film. Even though the four repeats showed certain degree of variation, the overall trend remained the same and the reproducibility was satisfactory with an acceptable standard deviation (3b) (<10%).
a Force Increase (g)
3. Materials
The main reason of using these two biopolymers was because of their wide usage in food applications as structure build agents and because of their consistency and flowability at such particular concentrations. Sodium caseinate is a water soluble milk protein widely used as an important functional ingredient for food and other applications because of its excellent structure-building or emulsifying capabilities and its physicochemical properties have been thoroughly investigated in literature (Wong et al., 1996; Fox and Brodkorb, 2008). At a concentration of 18 wt%, the fluid has an apparent viscosity close to 50 Pas (Chan et al., 2007). Similar viscosity profile was also observed for 30% waxy maize starch. Both samples were highly viscous and consistent while still flowable (Chan et al., 2007). This is important in making sure that the fluid can be easily and evenly coated to the beam surface but is still firm enough to hold its shape relatively well against the gravity, with little dripping or spreading during the experiment.
3 2.5 2 1.5 1 0.5 0 -0.5 0
10
20
30
40
50
60
40
50
60
Time (Min)
b
3.0 2.5
Average force (g)
The small gate on the left side wall was used for film coating. Initially the beam was pushed to the left end of the trough and a sliding gate with an opening of 0.4 mm height and 10.6 mm width rests on the beam. An appropriate amount of testing material (already thermally equilibrated to the experimental temperature) was carefully transferred to the front of the sliding gate and, by slowly pulling the brass block inwards, a thin coating film (0.4 mm 10.6 mm 73 mm) can be created on the top of the beam, leaving 1 mm edge space on both sides of coating film. Once the coated beam was in its position, the pendant was carefully put into its place to make a firm connection of the beam to the balance. The drying process is then monitored by recording the data for every minute from the balance. It should be noted that, in order to maintain reproducibility, a two minute delay was given to each test after the completion of film coating to the start (time zero) of force recording. The chamber was fitted with two sets of heating elements. The temperature inside the chamber can then be controlled by the thermal probe in the centre of the chamber. A small and slow rotating fan was used to create a very gentle air flow to help maintain thermal and humidity homogeneity of air without causing any disturbance to the pendant. The closed chamber has not yet been equipped with a humidity control device. However, a separate thermo-hydrometer was placed inside the chamber to record the relative humidity as well as the experimental temperature. This device was also used to monitor the kinetics of moisture loss of the film. A film was coated on a metal beam using the same technique. The beam was attached to a brass cube for easy handling. After coating, the beam was quickly relocated into the balance and the total weight of the metal beam and the film was recorded at 30 s intervals after the chamber was sealed. Same as stress development tests, moisture loss tests were also conducted under a controlled temperature and a monitored relative humidity. All experiments were repeated for at least three times for reproducibility.
2.0 1.5 1.0 0.5
TM
0.0 -0.5
0
10
20
30
Time (Min) Fig. 3. Reproducibility tests of the drying stress for 18 wt% sodium caseinate films. The force (stress) increase of the drying film was recorded as a function of time. (a) Repeats of four experiments conducted at 25 ± 1 oC and relative humidity of 30 ± 2% and (b) an average of the four experiments.
J. Chen et al. / Journal of Food Engineering 92 (2009) 383–388
system may reach a frozen state. At this stage, the amount (and the rate) of moisture loss becomes very limited and, as a result of this, there should be very limited stress increase, but the film becomes very rigid and solid enough to support the stress (Lei et al., 2002). The relaxation for such a film approaches infinite long and the stress will remain unless a cracking or physical bending to release the stress. This observation may contradict to the conclusion made by Augier et al. (2002), who claimed that surface cracking occurs most likely at the beginning of a drying process because of the low yield stress of the material. However, with no detectable lateral stress, it is hard to believe that cracking could occur at the first phase of drying. It is more likely that material deformation (shrinking or cracking) to occur during the phase II of drying, where the lateral stress becomes significantly large but the yield stress of the material remains low. 4.2. Effects of drying temperature and relative humidity It is well-known that temperature and relative humidity of the drying air are the two main controlling factors for dehydration. Higher temperature and lower relative humidity are often favoured for a faster moisture loss. However, it is rather not clear how these two factors influence the stress development with the drying system. For this purpose, the stress development of concentrated sodium caseinate and waxy maize starch systems has also been examined as affected by the temperature and relative humidity. The recorded stress developments are shown in Fig. 4 for the
Force Increase (g)
a
100% 4.5 80% 3.5 60%
2.5
40%
1.5
20%
0.5
Remaining Mass (%)
0%
-0.5 0
10
20
30
40
50
60
Time (Min)
b
7
100% 80%
5 60% 3 40% 1
20%
-1
Remaining Mass (%)
Contributions to the experimental error could be from a number of factors, but we recommend that the control of relative humidity could be the main contributing source. The difference of the relatively humidity at the film surface and of the drying air is the driving force of mass transfer and determines the rate of moisture loss. Unfortunately, there was no internal control of relative humidity in the current experimental set up. The relative humidity of each experiment was determined by the atmospheric condition of the lab. Even though 5% difference in relative humidity was probably the best we could wait for, its contribution to experimental error should not be ignored. Further improvement plan is currently undergoing to add a regulated air circulation system to the chamber which will give a properly controlled temperature and relative humidity. The thickness and evenness of coated film is another important factor influencing the accuracy and reproducibility of experiments. Currently the film coating is done manually by pulling the cantilever beam through the coating gate. Numerous coating practices have to be conducted to make sure that the pulling is conducted at an even and constant speed when the beam is slid through the coating gate. The force curves in Fig. 3 suggest that the stress development for a drying caseinate film can be divided roughly into three phases. At the beginning of the drying (phase I), the dehydration appears to lead no detectable stress to the film. This phase lasts for about 12 min after drying starts. However, it should be noted that after the completion of film coating to the start of recording, there was a delay of 2 min. This delay was necessary for the transfer of the beam from the coating gate to its drying position and for the set up of pendant. This was also necessary to make sure that every drying test has the same starting point. In phase II, the lateral stress starts to show a relentless increase with time, a sharp and almost a linear increase over the next 15 min or so. This could be the stage of most drastic changes to the film in terms of its material properties. Huge internal stress increase within a short period of time may cause various changes to a material, such as internal flow, yielding, bending, and of course possibly surface cracking. At phase III, the stress starts to level off and remains at a high but constant value. This phase tends to last for many hours or even overnight. There were one or two occasions that the coating film partly peeled off from the substrate and gave no detected force (stress). Fig. 3b gives the average of four experiments. We can see that the bending force of the drying film acting on the beam reaches its maximum of around 2.5 g (or 24.5 mN) and levels off after about half hour drying. In analysing the three phases stress development, one should note that the internal stress development is always accompanied by a stress relaxation process which tends to release stress by displacing particles and molecules within the network. Therefore, the recorded stress should be the combined effect of the two counteracting processes. While the rate of stress increase depends on the magnitude of moisture content gradient or the amount of moisture loss, the rate of stress relaxation is determined in principle by the microstructural nature of the material. Internal stress develops only if the rate of stress increase is faster than that of stress release. If the rate of stress relaxation exceeds the rate of stress increase, no internal stress becomes observable. From Fig. 3, one can say that the stress relaxation at phase I is very fast. This could be partly because that the material at this stage is still flowable and highly flexible for fast stress dissipation and partly because that the moisture content gradient is still small due to fast moisture migration from the interior to the surface. It was believed that, at this stage, the fluid coating is liquid enough that any shrinkage stress is rapidly relieved by viscous flow (Lei et al., 2002). However, the two processes changes its places in the second phase, where the moisture content gradient is well established but the stress relaxation becomes limited. In phase III, a constant stress suggests that the
Force Increase (g)
386
0% 0
10
20
30
40
50
60
Time (Min) Fig. 4. The force increase and the remaining mass as a function of drying time at different drying conditions. (a) Drying of 18 wt% sodium caseinate (s: force increase at 25 oC and 30% RH; d: force increase at 30 oC and 21% RH; D: remaining mass at 25 oC and 30% RH; N remaining mass at 30 oC and 21% RH) and (b) drying of 30 wt% waxy maize starch (s: force increase at 25 oC and 33% RH; d: force increase at 25 oC and 24% RH; D: retaining mass at 25 oC and 33% RH; N retaining mass at 25 oC and 24% RH).
387
J. Chen et al. / Journal of Food Engineering 92 (2009) 383–388
4.3. Normalisation of stress and moisture loss A time factor has been used to normalise the stress and moisture loss of the films dehydrated at different atmospheric conditions. For example, for the pair of drying curves shown in Fig. 4a for 18% caseinate films, both stress content curves were normalised as a fraction of their maximum stress values and the time scales were shifted by the normalisation time factor, Kn. In this case, a time factor of 0.84 was used (Fig. 5a) and we can see that both stress curves are effectively merged into one single master curve. Interestingly, this same time factor was also applicable for the normalisation of the remaining mass curves (Fig. 5a). Same normalisation approach has also been found applicable for waxy maize starch films dehydrated at different relative humidity (Fig. 4b). It was found that a time factor of 0.93 was applicable for the normalisation of both stress development and the remaining mass for the starch system (Fig. 5b).
Normalised Fraction
a
1.2 1 0.8 0.6 0.4 0.2 0 0
10
20
30
40
50
60
50
60
Normalised Time (Min)
b Normalised Fraction
caseinate sample dried at two different temperatures (4a) and for the starch sample dried two different relative humidities (4b). The temperature change was regulated by the thermal control system, but the relative humidity change was achieved by conducting the experiment at different weather conditions. Overall the stress development showed very similar pattern for the two biopolymer films dried at different conditions. However, the results clearly showed to our expectation that a higher drying temperature and a lower relative humidity led to an early onset of the stress increase and a quicker reach of the maximum stress. For example, the onset of stress increase for the caseinate film was shortened from 12.3 min at 25 oC to less than 10 min at 30 oC, while the time to reach the maximum was also decreased from around 27.5 min at 25 oC to around 24 min at 30 oC (Fig. 4a). Similar pattern was also clearly observable for the starch film when the relative humidity was decreased from 33% to 24% (Fig. 4b). The main reason for the quick onset of the stress increase at a higher temperature or a low relative humidity is probably because of the fast moisture loss under these conditions, which leads to an early set of the moisture gradient. Another significant impact of the temperature and humidity of drying air was seen on the magnitude of the maximum stress. It was found that an increase of five degree in drying temperature almost doubled the internal stress of the film. The final force recorded for the caseinate film was around 2.5 g when it was dried at 25 oC, but this was increased to higher than 4.6 g at the drying temperature of 30 oC. This is probably due to the fact that a fast moisture loss (as a result of higher drying temperature and/or drier air) leads to a quicker set up of moisture content gradient, which gives a faster stress increase and also at the same time a much slowed relaxation process (or an increase relaxation time) of the film. However, the change of 9% relative humidity seemed to have a much smaller effect than that of 5 oC. The maximum stress for the starch film increased from around 4.9 g at 33% relative humidity to around 6.2 g at 24% relative humidity, an increase of around 26%. Fig. 4 also gives the percentage of remaining mass as a function of drying time. These tests were conducted separately under the same experimental conditions (see details in apparatus section) and the percentage of remaining mass was calculated as the ratio of sample weight recorded at time t against its initial weight. It can be seen clearly that a higher temperature leads to a faster loss of moisture (Fig. 4a); while a decrease of relative humidity shows a very limited effect on the rate of moisture loss (Fig. 4b). However, it is worth to note that, for both varied temperature and varied relative humidity, the remaining mass reaches to its final steady value almost at the same time as the stress reaches its maximum plateau for both biopolymer systems.
1.2 1 0.8 0.6 0.4 0.2 0 0
10
20
30
40
Normalised Time (Min) Fig. 5. Normalised stress development curve and remaining mass curves from Fig. 4a a normalisation factor of 0.84 was used for the normalisation of sodium caseinate films dried at two different temperatures as shown in Fig. 4a (s: force increase at 25 oC and 30% RH; d: force increase at 30 oC and 21% RH; D: remaining mass at 25 oC and 30% RH; N remaining mass at 30 oC and 21% RH) and (b) a normalisation factor of 0.93 was used for the normalisation of waxy maize starch films dried at two different relative humidity as shown in Fig. 4b (s:force increase at 25 oC and 33% RH; d: force increase at 25 oC and 24% RH; D: remaining mass at 25 oC and 33% RH; N remaining mass at 25 oC and 24% RH).
From the normalisation analysis, it is clear that the stress increase and the amount of moisture loss within a drying film are coherently relevant if not directly linked. A single time factor for each sample dried at two different conditions suggests that the drying of the film undergoes through a very same pathway and is probably controlled by the same mechanisms. It is also worth to indicate that the magnitude of the normalisation factor may reflect different impacts of drying conditions. A value further away from a unit may suggest a greater impact of the drying condition. This means that 5 oC temperature variation (Fig. 5a) may have a greater influence on drying dynamics than 10% relative humidity change (Fig. 5b). The former has a time factor of 0.84, while the latter has a time factor of 0.93. 4.4. Relationship to the rate of film dehydration A dehydration process is often divided into different stages based on the rate of moisture loss: constant rate, first falling rate, second falling rate, etc. (Toledo, 1999). Fig. 6 shows the rate of dehydration for the 18% sodium caseinate film, where the rate of moisture loss was calculated as the amount of moisture loss per unit of time and per unit amount of dry matter. We can see that the rate of moisture loss in this case can be roughly categorized into three different stages: first falling rate, second falling rate, and finally levelling off with a minimal drying rate. Factors and mechanisms controlling the rate of moisture loss have been well studied in the literature (Srikiatden and Roberts, 2007) and are not main concern of this work. Our concern was whether the stress development within a drying film has any correlation with the rate
J. Chen et al. / Journal of Food Engineering 92 (2009) 383–388
3.0
0.18
2.5
0.15
2.0
0.12
1.5
0.09
1.0
0.06
0.5
0.03
0.0
0
-0.5
Drying rate
Force increase (g)
388
-0.03 0
10
20
30
40
50
60
Time (Min) Fig. 6. The stress development (D) and the rate of dehydration (s) as a function of drying time (18 wt% sodium caseinate films dried at 25 ± 1 oC and 30 ± 2% RH). The results were the average of four experiments.
of moisture loss. From the results shown in Fig. 6, one can see that the three phases of stress development and the three stages of dehydration are in rather different time scales. There was no stress increase when the moisture loss was in the highest rate at the beginning of drying. The most significant stress increase occurred at the later stage of first falling rate and early stage of second falling rate. Therefore, one could conclude that, even though the stress development within a drying film is the direct result of moisture loss, its evolving process has no direct correlation with the rate of moisture loss. This may be because that the rate of moisture loss is more or less an instantaneous property of a material and has little dependence on its previous rate, but the stress increase within a film is an integrated effect and could depend highly on the whole drying history. 5. Conclusions A novel technique has been developed to monitor the stress development of a fluid film during its drying process. This technique directly measures the bending force exerting on the substrate beam and is easy to operate. Two biopolymer fluids have
been tested using this methodology: 18 wt% sodium caseinate and 30 wt% waxy maize starch, dried at two different temperatures (25 and 30 oC) and two different relative humidities (24% and 33%). It was found that the stress development of these systems showed very similar pattern: an initial delay, a sharp stress increase, followed by a stress plateau. The normalisation of stress increase and amount of moisture loss revealed same time factor for two processes conducted at different atmospheric conditions. It was also observed that the stress development of a biopolymer film had no direct relevance to the rate of moisture loss. Acknowledgement Authors wish to acknowledge partial financial support from Food Processing Faraday (UK) for this Project. References Augier, F., Coumans, W.J., Hugget, A., Kaasschieter, E.F., 2002. On the risk of cracking in clay drying. Chemical Engineering Journal 86, 133–138. Barbosa-Cásanovas, G.V., 1996. Dehydration of Foods. Chapman & Hall, New York. Chan, P.S.-K., Chen, J., Ettelaie, R., Law, Z., Alevisopoulos, S., Day, E., Smith, S., 2007. Shear and extensional rheology of casein and starch mixtures: effects of concentration and temperature. Food Hydrocolloids 21, 716–725. De Gennes, P.G., 2002. Solvent evaporation of spin cast films: crust effects. European Physical Journal E 7, 31–34. Fox, P.F., Brodkorb, A., 2008. The casein micelle: historical aspects, current concepts and significance. International Dairy Journal 18, 677–684. Lee, W.P., Routh, A.F., 2004. Why do drying films crack? Langmuir 20, 9885–9888. Lei, H., Francis, L.F., Gerberich, W.W., Scriven, L.E., 2002. Stress development in drying coatings after solidation. AICHE Journal 48, 437–451. Martinez, C.J., Lewis, J.F., 2002. Shape evolution and stress development during latex-silica film formation. Langmuir 18, 4689–4698. Payne, J.A., McCormick, A.V., Francis, L.F., 1997. In situ stress measurement apparatus for liquid applied coatings. Review of Scientific Instruments 68, 4564–4568. Srikiatden, J., Roberts, J.S., 2007. Moisture transfer in solid food materials: a review of mechanisms, models, and measurements. International Journal of Food Properties 10, 739–777. Thill, S., Spalla, O., 2003. Aggregation due to capillary forces during drying of particle submonolayers. Colloids and Surfaces A 217, 143–151. Toledo, R.T., 1999. Fundamentals of Food Process Engineering, second ed. An Aspen Publication, Gaithersburg, Maryland. Wedin, P., Martinez, C.J., Lewis, J.A., Daicic, J., Berström, L., 2004. Stress development during drying of calcium carbonate suspensions containing carboxymethylcellulose and latex particles. Journal of Colloid and Interface Science 272, 1–9. Wong, D.W.S., Camirand, W.M., Pavlath, A.E., 1996. Structures and functionalities of milk proteins. Critical Reviews in Food Science and Nutrition 36, 807–844.