The Changes in Food Coating Characteristics During Coating a Powder Mixture And

The Changes in Food Coating Characteristics during Coating a Powder Mixture and

Salting Potato Chips Nonelectrostatically and Electrostatically

Dissertation

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of

Philosophy in the Graduate School of The Ohio State University

Teerarat Likitwattanasade, M.S.

Graduate Program in Food Science and Technology

The Ohio State University

2015

Dissertation committee:

Sheryl A. Barringer, Advisor

Luis Rodriguez-Saona

Valente Alvarez

V.M. (Bala) Balasubramaniam

Teerarat Likitwattanasade

2015

Abstract

Changes in coating characteristics, when powder mixtures and extra fine NaCl were coated nonelectrostatically and electrostatically, were determined in this study.

Mixtures contained powders of two different compositions or two different sizes of one composition, of common food ingredients such as salts, sugar, starch, proteins, and maltodextrin were coated at 0 and -25 kV on aluminum targets. The separation and dustiness after coating the mixtures were determined. Extra fine NaCl was also coated individually at 0-95 kV onto potato chips by different coating systems to investigate which system provides better transfer efficiency, more even distribution and lower amount of dust. When the mixtures were coated, separation occurred in most mixtures. In the mixtures containing the same size powders but different composition, differences in individual transfer efficiency of powders was the greatest cause of separation. Moreover, interactions between the powders during coating generally decreased the separation. In the mixtures of the powders containing the same composition but different size powders, the differences in individual transfer efficiency was also a cause of separation. Interactions between the powders generally increased the separation. In mixtures with salt either mixed with other powders or salt itself of a different size, the magnitude of separation was higher than other mixtures.

Electrostatic coating decreased separation in the mixtures of non-salt powders when the same size but different composition powders were mixed. Being in a mixture generally had no effect on dustiness during either nonelectrostatic or electrostatic coating. For potato chip salting, there was no significant difference in the percent of ii

NaCl on the chips when nonelectrostatic or electrostatic salting was performed by spray gun or scarf plate. During nonelectrostatic salting, the spray gun produced more dust than the scarf plate. However, when electrostatics (≥50 kV) was applied, the amount of dust produced by the spray gun decreased up to 88%. NaCl distribution across the bed of chips was more even when using the spray gun than the scarf plate.

With the spray gun the evenness of the distribution decreased at high voltage (≥75 kV).

iii

Dedication

Dedicated to the student at The Ohio State University

Acknowledgements

I would like to express my sincere gratitude to my advisor, Professor Sheryl

A. Barringer for her understanding, wisdom, and patience throughout my Ph.D. program – especially for all precious opportunities she gave me. Without her guidance this dissertation would not have been possible. I thank to my committee, Professor

Valente Alvarez, Professor Luis Rodriguez-Saona, Professor V.M. (Bala)

Balasubramaniam, for their suggestions and kindness. I also thank friends and staff at

Department of Food Science and Technology, The Ohio State University for your academic support, assistance and friendship. I humbly extend my thanks to all persons who assist me throughout my degree. I am deeply indebted to Mrs. Sherry Walter who always is with me when I have problem.

I would like to thank Ministry of Science and Technology, Thailand for financial support. Finally, I would like to acknowledge with gratitude, the support and love of my family and Thai friends.

Vita

2006…………………...B.S. Food Science and Technology, Kasetsart University,

Thailand

2009…………………...M.S. Food Science, Kasetsart University, Bangkok, Thailand

2010 to present ……….Graduate Student, Department of Food Science and

Technology, The Ohio State University

Publications

Likitwattanasade T, Hongsprabhas P. Effect of storage proteins on pasting properties and microstructure of Thai rice. Food Res Int. 43(5): 1402-1409., 2010.

Likitwattanasade T, Barringer SA. Separation of powder mixtures during electrostatic and nonelectrostatic coating. J Food Process Eng. 36: 731-738., 2013.

Fields of Study

Major Field: Food Science and Technology

Table of Contents

Abstract ...... ii

Dedication ...... iv

Acknowledgements ...... v

Vita ...... vi

List of Tables ...... xi

List of Figures ...... xii

CHAPTER 1 LITERATURE REVIEW ...... 1

1.1 Electrostatic Coating ...... 1

1.2 Effect of Physical Properties on Nonelectrostatic and Electrostatic Coating ...... 4

1.2.1 Particle size ...... 4

1.2.2 Flowability ...... 6

1.2.3 Particle resistivity ...... 8

1.2.4 Chargeability ...... 10

1.2.5 Density ...... 11

1.3 Effect of Particle Size on Mixture during Nonelectrostatic and Electrostatic coating .. 13

1.4 Effect of Powder Composition on Mixtures during Nonelectrostatic and Electrostatic

coating ...... 15

1.5 Dustiness during Nonelectrostatic and Electrostatic coating ...... 16

vii

CHAPTER 2 THE SEPARATION OF MIXTURES DURING

NONELECTROSTATIC AND ELECTROSTATIC COATING ...... 18

2.1 Abstract...... 18

2.2 Introduction ...... 19

2.3 Materials and Methods ...... 22

2.3.1. Powder samples ...... 22

2.3.2. Coating conditions ...... 22

2.3.3 Powder property determinations ...... 23

2.3.4 Determination of deposited powder ...... 24

2.3.5 Targeting loss and adhesion loss determination ...... 25

2.3.6 Statistical analysis ...... 27

2.4 Results and Discussion ...... 27

2.4.1 Separation in the mixtures ...... 27

2.4.2 Nonelectrostatic coating of the mixtures with NaCl ...... 30

2.4.3 Nonelectrostatic coating of the mixtures with starch, protein and sugar ...... 36

2.4.4 Electrostatic versus Nonelectrostatic Coating ...... 36

2.4.5 Electrostatic coating of the mixtures with NaCl ...... 38

2.4.6 Electrostatic coating of the mixtures with starch, protein and sugar ...... 38

2.5 Conclusion ...... 38

2.6 Practical Applications ...... 39

viii

CHAPTER 3 THE INFLUENCE OF PARTICLE SIZE ON SEPARATION AND

DUSTINESS IN POWDER MIXTURES DURING NONELECTROSTATIC AND

ELECTROSTATIC COATING ...... 40

3.1 Abstract...... 40

3.2 Introduction ...... 41

3.3 Materials and Methods ...... 44

3.3.1 Powder samples ...... 44

3.3.2 Coating process ...... 45

3.3.3 Determination of the Amount of Powder in a Mixture after Coating ...... 46

3.3.4 Separation...... 46

3.3.5 Composition versus Particle size ...... 47

3.3.6 Dustiness...... 47

3.3.7 Statistical analysis ...... 48

3.4. Results and Discussion ...... 48

3.4.1 Separation...... 48

3.4.2 Composition vs. Particle size ...... 55

3.4.3 Dustiness...... 59

3.5 Conclusion ...... 61

3.6 Practical Applications ...... 61

CHAPTER 4 ELECTROSTATIC SALTING OF POTATO CHIPS ...... 62

4.1 Abstract...... 62

4.2 Introduction ...... 63

4.3 Materials and Methods ...... 66

4.3.1 Potato chip salting ...... 66

4.3.2 NaCl determination ...... 67

4.3.3 Dust determination ...... 68

4.3.4 Statistical analysis ...... 68

4.4 Results and Discussion ...... 68

4.4.1. NaCl distribution ...... 68

4.4.2. Average % NaCl on the chip ...... 71

4.4.3. Dustiness...... 73

4.5 Conclusions ...... 75

4.6 Practical Applications ...... 76

References ...... 77

Appendix A: Standard curves ...... 86

List of Tables

Table 2.1 Number of locations where separation occurred in the mixture, and were predicted to occur based on coating of the individual powders...... 28

Table 2.2 Physical properties of NaCl , starch, protein and sugar ...... 29

Table 3.1 Physical properties of NaCl, KCl, sucrose, starch, maltodextrin, whey protein, casein and soy protein...... 54

Table 3.2 Transfer efficiency of fine NaCl, sucrose, starch, and soy protein when coating as a mixture with powders of different composition or different particle siz58

List of Figures

Figure 1.1 An illustration of corona charging and triboelectric charging ...... 3

Figure 2.1 Targeting setup reference ...... 23

Figure 2.2 Targeting loss, adhesion loss and transfer efficiency ...... 26

Figure 2.3 Targeting loss (TL) and adhesion loss (AL) of powders when coated individually nonelectrostatically (NE) and electrostatically (E) ...... 32

Figure 2.4 Nonelectrostatic coating: targeting loss (TL) and adhesion loss (AL) of the powder coated individually and in a mixture...... 34

Figure 2.5 Difference in losses in each mixture with two powders of similar size. .... 35

Figure 2.6 Electrostatic coating: targeting loss (TL) and adhesion loss (AL) of the powder coated individually and in a mixture...... 37

Figure 3.1 Coating setup ...... 46

Figure 3.2 Proportion of coarse and fine powder on the target after nonelectrostatic and electrostatic coating ...... 52

Figure 3.3 Transfer efficiency of NaCl, KCl, sucrose, starch, maltodextrin, whey protein, casein and soy protein ...... 53

Figure 3.4 Dust collected during nonelectrostatic coating and electrostatic coating of the mixtures containing different size powders...... 60

xii

Figure 4.2 NaCl ring on the drum wall during nonelectrostatic salting of the potato chips by using the scarf plate ...... 71

Figure 4.3 Percent NaCl on potato chips after salting the potato chips with 1.1% NaCl by using the spray gun and the scarf plate systems ...... 73

Figure 4.4 The dust amount produced during salting by the spray gun and the scarf plate ...... 75

Figure A 1 Standard curve of powdered sugar solution ...... 86

Figure A 2 Standard curve of soy protein solution ...... 86

Figure A 3 Standard curve of extra fine salt solution ...... 87

xiii

CHAPTER 1

LITERATURE REVIEW

1.1 Electrostatic Coating

Electrostatic powder coating has been widely used in the automotive and painting industry for several decades before it was introduced to the food industry.

Due to superior coating performance; higher transfer efficiency, more even powder distribution, better adhesion, and also reducing dust production in processing line when compared to traditional coating processes, electrostatic powder coating applications have been studied by numerous researchers. The quality of tomato dices calcified by electrostatic powder coating and traditional liquid dipping were compared. A firmer tomato with no water waste was a result when using electrostatic coating (Rao and Barringer 2005). The shelf life of shredded cheese, coated with antimycotic agent, can be improved by 15-30% using electrostatic coating (Elayedath and Barringer 2002). Improvement in color and microbial inhibition of meat (bologna, salami, pepperoni and ground turkey patties) when electrostatically coated with sodium erythorbate and GDL has been reported (Barringer and others 2005).

Electrostatic coating can also be used to coat French fries with sugar powder before frying and results in a decrease in processing time because the fries achieve the desired color faster than by traditional methods (Amefia and others 2006). Moreover

some of those works and others reported great reduction of powder waste and dust amount when using electrostatic coating (Rao and Barringer 2005, Elayedath and

Barringer 2002, Barringer, Abu-Ali and Chung 2005, Beihl and Barringer 2004).

The concept of electrostatic powder coating was originally in the US in the

1950s. The technique is widely used across several industries, but the core concept is all the same. Firstly, powder particles which are blown by the air from a reservoir are charged. Two common methods to charge powder particles are 1.corona charging which the particles are charged by passing them through the electric field and 2. triboelectric charging in which charged particles are produced by the collision between the particles themselves found between particle and the pipe wall resulting in charge exchanging (Figure 1.1).

Powder

Air Un-charged particles Charged particle Free ions

Powder

Air Un-charged particles Charged particle

Figure 1.1 An illustration of corona charging and triboelectric charging (adapted from

Khan and others 2012)

The amount of charge on the particles surface is dependent on the ability of the powder to pick up the charge which is related to the electrical property of powder particles such as electrical resistivity. A group of charged particles is formed after particle charging. Due to similarity in charge, the particles repel each other and create an even cloud of the power particles across the target surface. Subsequently, the charged particles are deposited on ground target by gravitational force which naturally occurs in traditional coating (nonelectrostatic coating) and Coulomb force which

accelerate the charged particles toward the target. As the distance decreases the attractive Coulomb force increases. Coulomb’s law equation is shown below.

Where F represents the electrical force, Q1 represents the quality of charge on particle 1, Q2 represents the quality of charge on the particle 2 and d2 represents the distance of separation between the two particles. The symbol k is a proportionality constant known as the coulomb’s law constant.

1.2 Effect of Physical Properties on Nonelectrostatic and Electrostatic Coating

Electrostatics has been proved by several researches to be positively influential to coating performance. However, different powders respond to the electrostatics differently resulting in difference in coating profile. This indicated the impact of powder property on the coating. There are many studies proved that powder properties play an important on the coating performance in both nonelectrostatic coating and electrostatic coating (Biehl and Barringer 2003; 2004, Mayr and

Barringer 2006, Halim and Barringer 2007, Ratanatriwong and Barringer 2007,

Yousuf and Barringer 2007).

1.2.1 Particle size

The size of the powder is the factor that most affects the efficiency of both electrostatic and nonelectrostatic coating; the direction of the impact, however, is different. In nonelectrostatic coating, smaller particles show lower transfer efficiency

than larger particles (Ricks and others 2002, Miller and Barringer 2002, Yousef and

Barringer 2007, Ratanatriwong and Barringer 2007). The small particles tend to remain in the air that passes out of the coating system with the air flow, while large particles with high mass that are dominated by gravitational force during coating fall onto the target resulting in less coating loss (Ricks and others 2002, Mayr and

Barringer 2006, Xu and Barringer 2008). For electrostatic coating, the transfer efficiency of most powders decreased and then increased with an increase of particle size (Ratanatriwong and Barringer 2007). As the size of the particle increases, charge to mass ratio on the powder decreases resulting in lower transfer efficiency when using electrostatic coating. Transfer efficiency decreases with particle size increase until it reaches a certain particle size where the electrostatic force is overcome by gravitational force. Thus, small particles are recommended for electrostatic coating to achieve better coating efficiency (Ratanatriwong and Barringer 2007).

Adhesion is one of the coating characteristics that are influenced by particle size. Particle size affects both electrostatic and nonelectrostatic adhesion. As the particle size increases, both nonelectrostatic and electrostatic adhesion significantly decreases. In nonelectrostatic coating, smaller powders adhere to the target more strongly than larger powders (Halim and Barringer 2007). This is because the greater surface area of the small powders leads to stronger van der Waals forces and capillary force per unit mass (Bowling 1988). In cases where the particle size is smaller than the pore size on the surface of the target, better adhesion is caused by the occurrence of mechanical interlocking (Bowling 1988). Salt with smaller size showed better adherence on potato chips than larger salt (Buck and Barringer 2007). With electrostatic coating, small particles have a higher charge to mass ratio compared to

large particles. Since electrostatic adhesion forces increase with increasing charge to mass ratio (Stozel and others 1994), small particles attain better adhesion then large particles (Buck and Barringer 2007). Additionally, large particles with lower charge easily fall off of the target after coating resulting in lower adhesion (Sumonsiri and

Barringer 2011).

Besides transfer efficiency and adhesion, particle size also affects the percent side coverage. The percent side coverage increased when the size of the powder decreased (378µm - 22µm), in both nonelectrostatic and electrostatic coating. The higher gravitational force on large particles causes less powder to deposit on the sides of the target. Moreover, large particles were more likely to fall off the target even if they are deposited. On the other hand, small particles which are more affected by aerodynamic force, have higher charge to mass ratio when compared to larger powder, so there is more attraction between the powders and target when using electrostatic coating (Sumonsiri and Barringer 2010).

1.2.2 Flowability

Flowability is the ability of a powder to flow. It has been reported in many papers for their effects on coating efficiency in both nonelectrostatic and electrostatic coating (Barringer 2007, Ratanatriwong and Barringer 2007, Amefia and others 2006,

Sumawi and Barringer 2005, Biehl and Barringer 2003, Ricks and Others 2002).

Flowability is influenced by other properties of the powder such as particle size, moisture content (Fitzpatrick and others 2004, Teunou and other 1999) and also can depend on the chemical composition of the powders (Dhanalakshmi and other 2011).

Powders with small size are more cohesive due to their high surface to volume ratio

producing low flowability powders. In contrast, larger powders typically have higher flowablity (Ratanatriwong and Barringer 2007, Biehl and Barringer 2003).

Transfer efficiency has been reported to depend on the flowability. Transfer efficiency increases with increasing flowability of powder in both nonelectrostatic and electrostatic coating (Ratanatriwong and Barringer 2003; 2007, Biehl and Barringer

2003; 2004, Ricks and others 2002,). There are many reasons explaining why free flowing powders exhibit better transfer efficiency. Free flowing powders are better capable to flow through the system and be well dispersed (Mazumder and other

1997). The powders also maintain a uniform curtain over the entire length of the high voltage wire placed across the coating chamber. Charging occurs efficiently (Biehl and Barringer 2003). In contrast, particles with high cohesiveness are likely to form clumps, have low flowability and cause variation in mass flow rate and transfer efficiency. Additionally, the powder clumps are more likely to fall off the targets and conveyor belt during coating due to gravitational force resulting in lower transfer efficiency (Ricks and others 2002, Biehl and Barringer 2004). Besides, cohesive powders are prone to choke or form bridges within the outlet they are flowing from and adhere to the transporting tube wall, decreasing the amount of input powder and resulting in lower transfer efficiency (Ratanatriwong and Barringer 2007, Biehl and

Barringer 2003; 2004).

Flowability also significantly affects the dustiness. Free flowing powder produces more dust than cohesive powder. However, using electrostatic coating with free flowing powder, not only improves transfer efficiency, it also reduces the amount of dust. Percent side covered is also affected by flowability (Sumonsiri and Barringer

2010). Percent side covered increases with decreasing flowability. This is opposite to 7

the effect of flowability on the transfer efficiency which is increased with increasing flowability. This is due to cohesive powder adhering well not only with powder themselves but also with the target causing less falling off of the target which produces higher percent side coverage.

There are a number of methods used for determining the flowability of the powder including angle of repose, Hausner ratio, cohesion and flow index. Angle of repose is a quick and reproducible method measuring the flowability of powders

(Geldart and others 2006). It represents the angle which is formed from the highest peak before collapse of a pile of powder. According to this method, the powder can be classified into four groups based on their angles. Powder with angles of repose below

30° indicate good flowability, 30° - 45° some cohesiveness, 45° - 55° true cohesiveness and more than 55° sluggish or very high cohesiveness and very limited flowability (Carr 1965). Hausner ratio is another method commonly used to determine powder flowability. It is the ratio between untapped bulk density divided by the tapped bulk density of powders. Powders with Hausner ratio greater than 1.4 are cohesive and do not flow well. Hausner ratio of 1.25 to 1.4 indicates fairly free flowing powders and powders with Hausner ratios of 1 to 1.25 are free flowing (de

Jong and others 1999).

1.2.3 Particle resistivity

Powder resistivity plays an important role in the performance of electrostatic coating. Powder can be grouped into three ranges based on their resistivity (Bailey

1998). Powder with resistivity below 1010Ωm is a conductive powder. The powders in this range are charged effectively but their charges are quickly lost when they contact the target, resulting in them detaching from the target (Bailey 1998). Thus, 8

this type of powder frequently provides poor adhesion. Salt and maltodextrin are conductive powders which do not show significant electrostatic adhesion (Halim and

Barringer 2007). Others examples of food powders in this group are tapioca starch,

NaCl, sugar, soy protein isolate and wheat protein (Sumonsiri and Barringer 2010).

The second group is powder in the resistivity range between 1013 - 1010Ωm.

The performance of coating of the powder in this range is very difficult to predict.

Generally, powders in this range have a few seconds charge decay and poor adhesion.

However, some powders show good adhesion depending on particle composition, surface chemistry and other factors (Bailey 1998). In this range, corn starch and sour cream powder do not show significant electrostatic adhesion while cellulose, whey, soy flour, nonfat dry milk and cocoa powder do (Halim and Barringer 2007). Percent side coverage, another coating characteristic, on fresh and dry bread is significantly higher than those of the powders with low resistivity. Other food powders in this group are corn starch, modified starch, cocoa powder and cheese powder (Sumonsiri and Barringer 2010).

Particles with resistivity higher than 1013Ωm are insulating. Powders having a resistivity in this range produce good adhesion and have a self-limiting layer thickness because of slow charge decay (Bailey 1998). In other words, after high resistivity powders are charged, they held those charges longer than powder with lower resistivity. Higher amount of charge on the powders depositing on the target results in better adhesion of the high resistivity powder. Thus, high resistivity powder is recommended for good adhesion properties. Other workers have reported that powders with a high resistivity (cocoa powder with a high resistivity of 1.15x1013) had stronger electrostatic adhesion than powder with a low resistivity (starch with a 9

medium resistivity 2.56x1010 and NaCl with a low resistivity 7.31x105, respectively)

(Huang and Barringer 2011). However, only a few food powders have resistivity in this range.

1.2.4 Chargeability

Chargeability is the ability of a powder to become charged. The higher the chargeability, the easier the powder can pick up the charge. Chargeability was reported to have a significant effect on transfer efficiency in both nonelectrostatic and electrostatic coating. Using nonelectrostatic coating, as the charge on the powder increases, the amount of powder landing on the target decreases (Ricks and others

2002). Carrying the higher charge may have caused the particles to repel each other resulting in powder remaining in the air and also have inadequate charge to seek the grounded target. Note that powders have charge even during nonelectrostatic coating which develops from tribocharging while being ground, shipped or conveyed in a coating system.

However, in one study there was no relationship between improvement of coating efficiency and chargeability when using electrostatic coating (Ricks and others 2002). This result contrasted with those of Biehl and Barringer (2003) which determined that the chargeability was the most important variable for electrostatic coating. They noted that larger charge caused more repulsion and the powder to spread out farther, so more uniform deposition and greater transfer efficiency are produced. The possible reason for this difference might be the insensitivity of measurement which cannot differentiate between strongly charged powders, because with electrostatic coating, most of the powders were charged close to 100% (Ricks and others 2002). 10

The effects of relative humidity on the resistivity of particles have been reported by many researchers (Ong and others 1975, Sharma and others 2001, Halim and Barringer 2007, Xu and Barringer 2008). Powder resistivity decreases with increasing relative humidity. As relative humidity increases, the amount of water absorbed onto the powder surface increases. Since water has lower resistivity than most food powders, it thereby decreases the powder resistivity. Cocoa powder’s resistivity decreases with increasing relative humidity (Halim and Barringer 2007).

With increasing relative humidity and decreasing resistivity, the electrostatic adhesion decreases because particles lose charges quickly, as previously discussed in powder with low resistivity. Soy protein, pork gelatin and whey protein shows the greatest decrease in resistivity at 80% relative humidity compared to those with 60, 40 and

20% relative humidity (Xu and Barringer 2008).

1.2.5 Density

The effect of particle density on coating efficiency has been reported by a few researchers (Ricks and other 2002, Biehl and Barringer 2003; 2004, Yousuf and

Barringer 2007). In a conveyor system, as the density of food powders (1.20 - 2.20 g/cm3) increases, transfer efficiency increases for both nonelectrostatic and electrostatic coating (Yousuf and Barringer 2007). This is because particles with higher density have larger mass and inertia than particle with lower density of the same size. Therefore, high density particles have a high propensity to deposit onto the target more than low density particles which are more susceptible to remain in the air as dust and potentially pass out of the coating system, producing lower transfer efficiency (Yousuf and Barringer 2007). Such finding is in agreement with Ricks and others (2002) who have shown that an increase in particle density increases

improvement in coating transfer efficiency during electrostatic coating. However, particle density is not found to be the factor influenced the transfer efficiency during nonelectrostatic coating in their experiment.

The results of an increase in improvement of coating transfer efficiency with an increase in particle density during electrostatic coating are contrasted with those of

Biehl and Barringer (2004) who found that a greater improvement is shown in less dense particles with electrostatic coating. The reason is that coating electrostatically with low density particles, electrostatic forces may dominate over gravitational forces resulting in a stronger Coulombic force attraction between charged particles and the target in the low density particles when compared with the high density particles.

Particle density has a marked impact on only nonelectrostatic powder transfer efficiency when coating in a tumble drum (Biehl and Barringer 2003). As the density of powders in the range 1.28 - 2.21 g/cm3 increased transfer efficiency increased. The reason why the denser powders showed higher transfer efficiency was not clear.

However, the reason why the particle density was not a significant factor for electrostatic coating might be due to the domination of gravitational force over electrostatic force (Biehl and Barringer 2003).

Besides transfer efficiency, particle density also affects coating evenness.

Coating evenness increases with a decrease in particle density in both nonelectrostatic and electrostatic coating (Yousuf and Barringer 2007). Low density particles have lower mass than high density particle comparing with the same size. Having low mass, low density particles are more easily carried by air and well dispersed on the target than dense particles. Greater improvement in evenness due to using electrostatic

is exhibited when low density particles are used (1.4-15 g/cm3) when compared with higher density particles (NaCl at 2.2 g/cm3).

1.3 Effect of Particle Size on Mixture during Nonelectrostatic and Electrostatic coating

The changes in proportion of powder mixture after coating are mainly due to losses and uneven distribution. Somboonvechakarn and Barringer (2009) identified two types of losses; targeting loss and adhesion loss. Targeting loss is the loss of powder during the targeting step, referring to the powders which do not deposit on the target are either lost with the air or deposit on non-target area. The other loss is adhesion loss which occurs after powders deposit onto the target. Weakness in adhesion or in interaction between powder and target causes the deposited powder slip off the target resulting in adhesion loss. All three factors; targeting loss, adhesion loss and uneven distribution have been studied and it was found that they are influenced by the individual powder characteristics and interaction between powders in the mixture.

Considering targeting loss when using nonelectrostatic coating, particle size was found to have a huge impact on powder separation during coating. Basically, fine particles have a higher tendency to be carried away by the air resulting in lower amounts of fine particles landing on the target than coarse particles. However when the coarse and fine particles were mixed, midair collision between coarse particles and fine particles caused changes in particle trajectories, decreased the velocity of coarse particles and increased the velocity of fine particles. Consequently, more

coarse particles and fewer fine particles land on the targets during mixture coating when compared with the individual coating. For example, the mixture consisted of pregelatinized starch with particle size 64µm and 191µm. After nonelectrostatic coating, targeting loss of 64µm starch increased and targeting loss of 191µm starch decreased compared to when they were coated individually. However, a NaCl mixture

(44µm and 256µm) showed a different result. The coarse NaCl had a significant increase in targeting loss while fine NaCl had the same targeting loss when individual and mixture coating were compared. The fact that a high percent (70%) of the fine powder was already lost when coated individually may explain why there was no difference in targeting loss compared to coating in a mixture. However, the reason caused high targeting loss in coarse NaCl is still unclear.

Unlike nonelectrostatic coating, the effect of collisions and scattering between the particles was lessened when electrostatic coating was used. There was no significant difference in targeting loss when compared between in mixture coating and in individual coating of both fine and coarse powders. This was due mainly to the electrostatic force which leads charged particles sought the ground and reduce the effects of collision (Miller and Barringer 2002).

Adhesion loss did not have a significant effect on the proportion on the target for a NaCl and starch mixture during nonelectrostatic coating. However, during electrostatic coating, adhesion loss plays a major role on the proportion change for both powders. Because of better holding of the charge and strong electrostatic adhesion, fine particles showed a decrease or zero value in adhesion loss in the mixture when compared to coarse particles which the adhesion loss increased because

the fine particles with high adhesion created a charged layer on the target and repelled the incoming coarse particles.

Using nonelectrostatic coating, the large particles are greatly affected by gravitational force and small particles are more likely to be carried away by the air means the position on the target close to the nozzle contained a larger proportion of the coarse particles while fine particles were more evenly distributed. However, by using electrostatic coating, the charge on the particles during electrostatic coating caused both fine and coarse particles to land closer to the nozzle. This significantly reduced the targeting loss of fine particles. However, the uneven distribution of the powder was still problematic.

1.4 Effect of Powder Composition on Mixtures during Nonelectrostatic and

Electrostatic coating

Besides the particle size, composition of the powder mixture also played a role in the proportion change during coating (Somboonvechakarn and Barringer 2011).

Three different effects; targeting loss, adhesion loss and uneven distribution are also responsible for the change in proportion. In a mixture of fine NaCl with particle size

44µm and starch with particle size 64µm, during targeting more NaCl particles were lost in the air than starch causing a change in proportion. For a coarse mixture, the effect of density was predominant. Coarse starch (191µm) with lower density was more likely to be carried away by the air than coarse NaCl (197µm) resulting in a change in proportion. Nevertheless, using electrostatic coating can reduce the

targeting loss because the charged powder particles prefer to seek the ground rather than staying in the air.

Although the changes in the proportion after coating of a mixture containing similar size powders has been explained by the difference in their size and density, other properties such as flowability and resistivity might also influence the mixture separation. So studying of the effect of the others properties such as flowability and resistivity are interested.

1.5 Dustiness during Nonelectrostatic and Electrostatic coating

The food powders generally are small particles and easily create dust during coating process (Anon 1992). Dust is undesired causing powder lost, powder waste, further cleaning cost and respiratory hazard to the workers. The physical properties have been studied and found their effects on the coating dustiness. During nonelectrostatic coating, powders with smaller size generate more dust than larger ones (Ratanatriwong and others 2003, Biehl and Barringer 2004, Sumawi and

Barringer 2005). Small particles (less in mass) have less effect by gravitational force than large particles (higher in mass), resulting in suspending in the air rather than fall onto the targets (Mazumder and others 1997). Free flowing powders produce more dust than cohesive powders (Biehl and Barringer 2004). The free flowing powders are much easier to driven by air to the coating system than the cohesive powders. Also, the cohesive powder is likely to form clumps, increasing their total mass (Ricks and others 2002). In some case, more cohesive powders show an increase in dustiness

(Ratanatriwong and others 2003). This is a result of dominating of effect of particle

size over particle flowability. Some cohesive powders are very fine and they have a tendency to suspend in the air as mentioned earlier. Powder resistivity is also reported having an effect on dustiness even in nonelectrostatic coating (Ricks and others 2002,

Biehl and Barringer 2004). Charges generating during handling (tribocharging) may cause the powder repelling each other during coating and creating more dust. Density does not have significantly effect on dustiness (Biehl and Barringer 2004).

Electrostatic coating has been proved that it can efficiently reduce dust amount during powder coating (Ricks and others 2002, Sumawi and Barringer 2005,

Ratanatriwong and others 2009). When the electrical charges are applied, the powders become charged and seek out the target resulting in a decrease in dust amount.

Although, small powder have a tendency to remain airborne, the greater absolute reduction in dustiness by using electrostatic coating is found greater than large powders. This can be explained by the charge to mass ratio; the small particles can achieve a greater charge to mass than large particles (Mazumder and others 1997).

Free flowing and highly charging powders is reported having a greater dust reduction

(Mazumder and others 1997, Biehl and Barringer 2004). However, Ratanatriwong and others in 2003 reported cohesive powder has a greater reduction in dust and powder chargeability has no effect on dust reduction. Comparing among the types of coating, dust reduction is greater found in pneumatic conveyor system than tumble drum system (Biehl and Barringer 2004). Further study in electrostatic coating showed positive and negative corona has no different effect on dustiness during electrostatic coating (Sumawi and Barringer 2005).

CHAPTER 2

THE SEPARATION OF MIXTURES DURING NONELECTROSTATIC AND

ELECTROSTATIC COATING

2.1 Abstract

Powder separation causes uneven flavor and color on coated food products.

Understanding the basis behind separation is needed to decrease separation. NaCl, starch, protein, sugar and mixtures of pairs of those powders were coated nonelectrostatically and electrostatically. Separation occurred in most mixtures.

Individual targeting loss and adhesion loss caused separation while interactions between powders decreased both of these losses and separation during mixture coating. The difference in individual targeting loss was the greatest cause of separation. During nonelectrostatic coating, when NaCl was one of the powders in the mixture, there were a greater number of locations both where separation actually occurred and where it was predicted to occur. During electrostatic coating, the individual targeting loss of all powders during electrostatic coating was lower than in nonelectrostatic coating and the difference in individual targeting loss was also lower.

Electrostatic coating generally decreased separation in the mixtures without NaCl.

2.2 Introduction

Powder coating is used to apply food powder ingredients such as seasonings to create flavor and color variety in food products. Powders are typically dispersed from the powder hopper aerodynamically or gravitationally, and deposit onto the food surface. During the coating process, some powders are lost with the air as dust, and some powders that are deposited on the food, fall off the food surface because of poor adhesion. These losses cause problems in coating. To ensure that an adequate amount is on the food products, an excess amount of powders is used and consequently causes a higher processing cost. The lost powders may remain suspended in the air and cause respiratory distress to the operators after long periods of continuous breathing. Waste powders also require a higher labor cost to clean (Clark 1995). Besides the losses, uneven distribution of powders on the coated foods causes uneven flavor and color on the coated products, which is undesirable in the food industry (Clark 1995).

Electrostatic powder coating was widely used in the automotive and painting industry for several decades before it was introduced into the food industry. Several studies have proven that electrostatic coating offers superior coating performance over convectional coating. After powders are dispensed into the coating chamber where a corona zone is produced, the powders are charged and seek the nearest ground, which is the food products, by electrostatic force described in Coulomb's law (Hughes 1997,

Bailey 1998). The combination of gravitational force and electrostatic force causes powders to deposit on the food products, resulting in a reduction of dust and less overuse. Less labor is needed to clean the processing line because of less powder waste. Uneven distribution of flavor and color may also be solved by electrostatic coating. Because powders, after passing through the corona zone, are similar in 19

charge, they repel each other and are evenly dispersed across the surface of the food products.

Physical properties have a significant impact on the coating performance.

Coarse particles coat more efficiently than fine particles in nonelectrostatic coating

(Ricks and others 2002, Biehl and Barringer 2004, Yousuf and Barringer 2007).

While for electrostatic coating, fine particles coat more efficiently than coarse particles (Ricks and others 2002, Biehl and Barringer 2003, Ratanatriwong and others

2003), until the size increases to the point where gravitational force overcomes electrostatic force, and coarse particles show better coating efficiency than fine particles (Mayr and Barringer 2006, Ratanatriwong and Barringer 2007). Powders with high ability to flow produce better coating efficiency than cohesive powder for nonelectrostatic coating (Ricks and others 2002, Ratanatriwong and others 2003,

Biehl and Barringer 2004). For electrostatic coating, some studies found cohesive powders coat more effectively than free-flowing powders (Ricks and others 2002,

Ratanatriwong and others 2003), while one study found the opposite (Biehl and

Barringer 2004). The greater the chargeability, the better the transfer efficiency when using electrostatic coating (Miller and Barringer 2002, Ricks and others 2002,

Ratanatriwong and others 2003). Increasing the particle density increases improvement in both nonelectrostatic and electrostatic coating transfer efficiency

(Ricks and others 2002, Biehl and Barringer 2003).

The other important issue in food powder coating is the separation of powders in a mixture during coating. Powders used in coating typically consist of a mixture of powders with different physical properties (Seighman 2001). After coating, the ratio of powders coated on the food changes from the original ratio. This is undesirable 20

because uneven appearance and distribution of flavors occur on the products. A difference in particle size of the powders in the mixture causes separation during mixture coating (Ye and Domnick 2003, Wang and others 2005, Somboonvechakarn and Barringer 2009). When powders of different sizes were coated nonelectrostatic coating, greater targeting loss occurred in fine powders than in coarse powders both in individual and mixture coating, causing the powders to separate (Somboonvechakarn and Barringer 2009). Using electrostatic coating increased the difference in adhesion loss of powders in the mixture.

When mixtures of similar size, but of different densities and compositions were nonelectrostatically coated, the fine mixture showed separation while little separation occurred in the coarse mixture (Somboonvechakarn and Barringer 2011).

In the fine mixture, the separation is because the high-density powders have significantly higher individual and mixture adhesion loss than the low-density powders. During electrostatic coating, the high-density powders have significantly greater individual adhesion loss than low-density powders in both the fine and coarse powder mixtures, resulting in separation.

Despite several studies on the effect of physical properties affecting coating efficiency, there is limited information on how the physical properties of powders, and use of electrostatic coating, impact mixture separation. Thus, the objective of this study was to determine the major causes of separation during mixture coating, when mixtures of powder with similar size (36–48μm) were coated nonelectrostatically and electrostatically. The amount of powder coated on the target, targeting loss and adhesion loss of individual powders and mixtures were determined.

2.3 Materials and Methods

2.3.1. Powder samples

Food powders tested in this study included potato starch (AVEBE American

Inc., Princeton, NJ, U.S.A.), NaCl (325 Extra Fine Salt, Morton International, Inc.,

Chicago, IL, U.S.A.), soy protein (ADM, Decatur, IL, U.S.A.) and powdered sugar

(Dixie Crystal, Sugar Land, TX, U.S.A.). Mean diameters of the powders were measured using the Malvern Mastersizer (X standard bench, Malvern Instrument Ltd.,

Worcestershire, U.K.). The volume mean diameter D[4,3] of each powder was measured three times and the average diameter were reported. Because water activity of the powder has been reported to affect the coating efficiency, all powders were equilibrated for 7 days at 20–25 ˚C and stored over saturated magnesium chloride solution (32.8% relative humidity) in sealed desiccators until used.

2.3.2. Coating conditions

An electrostatic powder-coating machine (Terronics Development Corp.,

Elwood, IN, U.S.A.) was used to coat the food powders on nine 15 × 10 cm aluminum sheets (Figure 2.1). No voltage was applied for nonelectrostatic coating and −25 kV was used for electrostatic coating. An air compressor (Hitachi EC79, Hitachi Koki

Co., Ltd., Tokyo, Japan) was used to supply airflow to drive the powder through the coating chamber. All experiments were carried out at an air velocity of 3.3 m/s, 30–

35% relative humidity and at 20–25 ˚C.

Figure 2.1 Targeting setup reference

2.3.3 Powder property determinations

Angle of repose was used to measure the flowability of the powders. It was determined by the modified fixed base method using a 2.81-cm diameter and 1.22-cm high Petri-dish. Powders were sifted with a powder sifter through a funnel with its tip

11 cm from the top edge of the Petri-dish. The maximum peak height of the powders prior to collapse was measured by a caliper and recorded. The arctangent of the peak height over the radius gives the angle of repose.

Resistivity of the powders was measured using a powder resistivity test cell.

The powder test cell was filled with 5 cm3 of the powder. Air was dispelled by tapping the cell for 5 s. Voltage (125 V) was applied to the cell via a high-voltage supply unit. The current value was read from the electrometer when there was no

change over a 15-s period. Resistivity was calculated using (K × I) / V, where K represents the cell constant (0.014); I represents electrical current; (A) and V represents the voltage applied (V).

Bulk density was determined by sifting the powder (120 ± 10 g) into a 250-mL graduated cylinder. The powders in the cylinder were weighted; their weights and volumes were recorded. Bulk density was calculated by dividing powder weight (g) by powder volume (cm3).

2.3.4 Determination of deposited powder

Both individual powders and powder mixtures were coated in order to determine interactions between the two powders during mixture coating. To calculate the targeting loss and adhesion loss, 20.000 ± 0.002 g of individual powder or a mixture of the two powders in a 1:1 ratio were coated nonelectrostatically and electrostatically on unoiled and oiled targets. For the oiled target, approximately 1 g vegetable oil (Kroger Pure Vegetable Oil, Cincinnati, OH, U.S.A.) was applied on each aluminum target. After coating, the mass of the individual or mixture of powder on each location was measured and was used to calculate the percent targeting loss and adhesion loss. For the powder mixture, the targets were weighed and the deposited powder was rinsed off the target using deionized water. The weight of the solution was recorded in order to calculate the concentration of each powder in the mixture. The solution was filtered through filter paper (No.5 filter paper Whatman,

GE Healthcare UK Ltd., Buckinghamshire, U.K.) to remove the water-insoluble materials: starch, oil and insoluble protein which interfere with the NaCl reading, ultraviolet (UV) absorbance and the brix reading.

The protein content of all mixture containing protein was determined using

UV absorption. The absorbance of the mixture solutions were measured at 280 nm with a UV-visible spectrophotometer (UV2450, Shimadzu, Kyoto, Japan). The amount of starch, NaCl or sugar in the mixture with protein was calculated by subtracting the protein content from the total mixed powder content after coating. The

NaCl concentration of the mixture between NaCl and starch or sugar was determined using a salt analyzer (Newport M-10 Digital Salt Analyzer, Newport, Santa Ana, CA,

U.S.A.). The amount of starch or sugar in the mixture with NaCl was calculated by subtracting the NaCl content from the total mixed powder content after coating. Sugar concentration in the mixture of sugar and starch was measured using a digital refractometer (Mark II, ABB, Reichert, NY, U.S.A.). The amount of starch was calculated by subtracting the sugar content from the total mixed powder content after coating.

2.3.5 Targeting loss and adhesion loss determination

Targeting loss and adhesion loss of the mixture were calculated. Targeting loss is a loss of powder to the environment occurring during the targeting step (Figure

2.2). Targeting loss occurs because powder is lost with the air as dust resulting in less powder deposition onto the targets compared to powder put into the system. It also includes the powder that missed the target. Targeting loss was calculated by the total mass of the powder fed into the system minus the powder deposited on the oiled target. Adhesion loss occurs after powders are already deposited on the targets. Not all deposited powder stays on the target, some powders roll off of the targets due to their poor adhesion. Adhesion loss was calculated by total loss of powder on unoiled targets minus the targeting loss.

The targeting loss was calculated as:

( )

The adhesion loss was calculated as:

( ) -

Oiled target Unoiled target

Deposited powder Powder coated on the target

Adhesion loss Powder slide off the target

Figure 2.2 Targeting loss, adhesion loss and transfer efficiency

2.3.6 Statistical analysis

For statistical analysis, one-way ANOVA with Tukey method for means was performed. Independent two tailed, two-sample T-test with unequal variance was performed to determine significant differences between the proportions of powders on the targets, and also to compare the means for the actual proportion and the predicted proportion of the powder on the targets. A p-value of 0.05 indicates a significant difference between the two groups.

2.4 Results and Discussion

2.4.1 Separation in the mixtures

Mixtures of two powders in equal percentage were coated onto nine targets.

The amount of each powder deposited on the target surfaces was measured to determine the occurrence of separation. On each location, separation is said to occur when the deposited amount of each of the two powders is significantly different.

During both nonelectrostatic and electrostatic coating, most mixtures showed separation in at least half of the locations (Table 2.1). Only the mixture of protein and sugar showed little to no separation. Particle size has been reported to have a large effect on mixture separation during powder coating (Somboonvechakarn and

Barringer 2009). In the current study, all of the powders had the same particle size to eliminate this effect (Table 2.2). Thus, differences in other properties are responsible for separation of the mixtures.

Table 2.1 Number of locations (out of nine) where separation occurred in the mixture, and were predicted to occur based on coating of

the individual powders. (Separation is defined as a significant difference between the amounts of each powder on a location.)

Nonelectrostatic coating Electrostatic coating

Oiled target Unoiled target Oiled target Unoiled target Mixture (Targeting loss) (Targeting loss (Targeting loss) (Targeting loss

+ Adhesion loss) + Adhesion loss)

Mixture Individual Mixture Individual Mixture Individual Mixture Individual

NaCl /Starch 6 7 5 9 8 5 6 9 28

NaCl/Protein 4 9 9 8 5 4 9 6

NaCl/Sugar 6 7 4 6 7 8 6 8

Starch/Protein 7 5 6 8 4 4 2 6

Starch/Sugar 1 4 6 8 1 6 8 8

Protein/Sugar 2 5 0 8 0 6 0 6

Separation is defined as a significant difference between the amounts of each powder on a location.

Table 2.2 Physical properties of NaCl , starch, protein and sugar

Properties NaCl Starch Protein Sugar

Particle size (μm) 45.91a 40.71bc 40.89b 38.31c

Flowability (degree) 41.09b 40.48b 62.41a 63.91a

Bulk density (g/cm3) 2.04a 2.15a 1.68c 1.87b

Resistivity (×105 Ωm) 9d 9,000c 89,000b 150,000a

aSamples in the same row with different letters are significantly different.

Separation during coating of the mixture is caused by the difference in individual coating performance of each powder and/or interactions among powders during coating. Differences in coating performance of each individual powder caused different amounts of each powder to be deposited in each location, causing separation.

In addition, during coating of a mixture, collisions among the powders may cause their coating performance to be different from when they are coated individually.

Collisions among powders with different properties may change the trajectory of the powder, causing the powder to be scattered beyond the target and producing, less powder on the target. Conversely, the collision may change the trajectory of the powder so that more powder is deposited on the target.

To determine if the differences between the coating profiles of individual powders caused separation in coating of the mixture, each powder was coated

individually and the amounts on the targets were compared. A significant difference in the amount between two powders when they are coated individually predicts that separation will occur during coating of the mixture. A difference in predicted amount was found in at least four out of the nine locations in all pairs of powders based on the individual coating (Table 2.1). This indicates that the difference in individual coating characteristics of powders was a major cause of separation in the mixtures. However, the interactions between the powders during coating also affect separation.

Comparing the amount predicted to be on a location based on individual coating, to the actual amount when the mixture is coated, indicates how much interaction occurs.

Interactions decreased separation in most mixtures.

During coating of most food items, both targeting loss and adhesion loss occurs because most foods have oil or water on the surface where the powders can adhere to. Targeting loss is a loss of powder during application, including powders lost with the air as dust and powders deposited beyond the target. Adhesion loss describes the loss of powder that deposits on to the target, but falls off because of poor adhesion between the powder and the target. In order to separate these two types of losses to be able to understand where loss is occurring, oil was applied on the surface of one set of targets to eliminate adhesion loss. Thus, the oiled targets measure only the targeting loss while unoiled targets measures both targeting loss and adhesion loss. Overall, differences in mixture targeting loss were the biggest cause of separation in most mixtures.

2.4.2 Nonelectrostatic coating of the mixtures with NaCl

When NaCl was one of the powders in the mixture, there were a greater number of locations both where separation actually occurred and where it was 30

predicted to occur, than with the other powder mixtures (Table 2.1). NaCl had much higher individual targeting loss (72%) than protein, sugar and starch, whose targeting loss values were similar to each other (44–50%; Figure 2.3). The difference in targeting loss between NaCl and the other powders predicted that most locations would have separation (Table 2.1). The high targeting loss of NaCl was also a major cause of separation in mixtures of 44 μm NaCl and 64 μm starch in a previous study

(Somboonvechakarn and Barringer 2011). They attributed the higher targeting loss of

NaCl to differences in particle size between their powders. However, there were minimal size differences in this study (Table 2.2), but still major differences in transfer efficiency. NaCl was actually the larger particle, which should have decreased its targeting loss. Thus, other physical properties, such as differences in density or resistivity, must be responsible for the higher targeting loss of NaCl.

Figure 2.3 Targeting loss (TL) and adhesion loss (AL) of powders when coated individually nonelectrostatically (NE) and electrostatically (E) (aSamples with different letters are significantly different between the powders in the same treatment.)

Interaction among powders in the mixture decreased the separation

(Table 2.1). In the mixture, the targeting loss of NaCl did not change, while the targeting loss of the powders it was paired with increased compared with when they were coated individually (Figure 2.4). During coating, NaCl, which has a high intrinsic density (2.20 g/cm3), may collide with and scatter away the powders it is paired with, which have lower intrinsic density (starch 1.51 g/cm3, protein 1.31 g/cm3, sugar 1.63 g/cm3, Ricks and others 2002, Setyo and Barringer 2007). These collisions greatly decreased the difference in losses between NaCl and the other powders, thus

these interactions created less separation than was predicted based on individual coating (Figure 2.5).

Based on differences in individual adhesion loss, adhesion loss was predicted to have a significant effect on separation (Figure 2.5). However, interactions during coating decreased the difference in adhesion loss. Thus, the effect of adhesion loss on the separation was small and much lower than the effect of targeting loss (Table 2.1).

Adhesion loss of NaCl was very small; 1% when it was coated individually and 0% when it was coated as a mixture (Figure 2.4). Thus, only the change in adhesion loss of the powders that NaCl was mixed with, affected the separation.

Other properties such as flowability and bulk density showed no effect on the separation. They have been shown to be important to other aspects of powder coating.

For instance, the more free-flowing and the denser the powder, the greater is the transfer efficiency (Ricks and others. 2002, Biehl and Barringer 2003; 2004,

Ratanatriwong and others 2003). However, NaCl and starch had similar flowability and density, but showed separation in five out of nine locations (Tables 2.1 and 2.2).

It may be that the much higher intrinsic density and much lower resistivity of salt were more important than the other properties.

Figure 2.4 Nonelectrostatic coating: targeting loss (TL) and adhesion loss (AL) of the powder coated individually and in a

mixture. (Samples with a triangle on top are mixtures significantly different from individual coating.)

Figure 2.5 Difference in losses in each mixture with two powders of similar size. (A solid circle on top shows significant

difference between the losses.)

2.4.3 Nonelectrostatic coating of the mixtures with starch, protein and sugar

Most mixtures with starch, protein and sugar showed separation, but it was less than for the mixtures with NaCl (Table 2.1). The greatest factor causing separation was again the mixture targeting loss (Figure 2.5). Individual adhesion loss predicted an increase in separation; however, the interaction during coating decreased the difference between the adhesion losses of powders in the mixtures, resulting in less impact of adhesion loss on the separation than targeting loss. Protein and sugar had no separation. Interestingly, they had the same flowability. It may be that when

NaCl is not one of the powders, flowability becomes an important factor in separation.

2.4.4 Electrostatic versus Nonelectrostatic Coating

The individual targeting loss of all powders during electrostatic coating was lower than in nonelectrostatic coating (Figure 2.3) and the difference in individual targeting loss was also lower (Figure 2.5). The decrease in both individual targeting loss and difference in individual targeting loss when electrostatic coating is used has also been reported by others (Somboonvechakarn and Barringer 2011). The decrease in targeting loss during electrostatic coating is due to the fact that charged powders did not remain suspended in the air, but rather sought the nearest target by the force described in Coulomb's law (Bailey 1998).

Generally, adhesion loss of NaCl increased and the other powders decreased when they were coated electrostatically compared with nonelectrostatically

(Figure 2.3). When charged powder lands on the target, it induces the target to create an image charge that is equal and opposite in polarity to the charged particle. This image charge attracts the charged powder with electrostatic force, resulting in a 36

decrease in adhesion loss (Hughes 1997). In the case of NaCl, the large decrease of targeting loss when electrostatics was used, producing more NaCl landing on the target, may have produced the higher adhesion loss.

Individual adhesion loss predicted an increase in the separation in electrostatic coating, but the interactions decreased the adhesion loss for many powders (Table 2.1,

Figure 2.6). Thus, there was no significant difference between the adhesion losses of the powders and their adhesion loss did not affect separation (Figure 2.5).

Figure 2.6 Electrostatic coating: targeting loss (TL) and adhesion loss (AL) of the powder coated individually and in a mixture. (Samples with a triangle on top are mixtures significantly different from the individual coating.)

2.4.5 Electrostatic coating of the mixtures with NaCl

Targeting loss, although lower than in nonelectrostatic coating, was still the greatest source of loss during electrostatic coating when NaCl was present. Using electrostatic coating, the mixtures with NaCl had increased separation when compared with nonelectrostatic coating and the biggest cause of separation was the mixture targeting loss (Table 2.1, Figure 2.5).

2.4.6 Electrostatic coating of the mixtures with starch, protein and sugar

For mixtures of starch, protein and sugar, electrostatic coating generally decreased the separation when compared with nonelectrostatic coating (Table 2.1).

Both individual targeting loss and adhesion loss decreased with electrostatic coating.

The greatest factor causing separation was the difference in adhesion loss in the mixture, especially when one of the powders was protein.

2.5 Conclusion

Even when powder mixtures have the same particle size, separation was found in most mixtures. Individual targeting loss and adhesion loss caused separation while interactions between powders decreased both these losses and separation during mixture coating. The difference in individual targeting loss was the greatest cause of separation. Separation is decreased by minimizing differences in individual targeting loss and by mixing powders together. During nonelectrostatic coating, when NaCl was one of the powders in the mixture, there was a large number of locations where separation occurred. During electrostatic coating, the individual targeting loss of all powders was lower than in nonelectrostatic coating and the difference in individual 38

targeting loss was also lower. Electrostatic coating generally decreased separation in the mixtures without NaCl.

2.6 Practical Applications

Using mixtures with similar size powders has been shown to reduce separation during powder coating. However, even if the powders are the same size, separation still occurs because different powders have different properties. Increasing the amount of salt in a mixture may decrease separation because salt has greater targeting loss than other powders. Minimizing differences in individual targeting loss decreases separation, as does mixing powders together. Electrostatic coating can be used to decrease losses and separation in mixtures without salt.

CHAPTER 3

THE INFLUENCE OF PARTICLE SIZE ON SEPARATION AND DUSTINESS IN

POWDER MIXTURES DURING NONELECTROSTATIC AND

ELECTROSTATIC COATING

3.1 Abstract

Many foods, especially snack foods, have powdered seasonings in order to enhance their flavor and increase product variety. A seasoning is usually a mixture of powder ingredients with different size. The effect of particle size of powder on coating characteristics has been studied, however it has been limited to the coating of a single powder. Separation of the powders after coating a mixture, which can negatively affect flavor and color of final product, can occur. Thus the objective of this study was to determine the effect of particle size on the separation and dustiness of powders in a mixture. Common food powders and their mixtures, consisting of two powders with the same composition but different in particle size: fine (51-95µm) and coarse (244-401µm) NaCl, KCl, sucrose, rice starch, maltodextrin, whey protein, casein and soy protein, were coated on a target at 0 and -25 kV at 30-35% relative humidity. Over half of the mixtures showed separation due to a difference in particle size during either nonelectrostatic or electrostatic coating. Separation was caused by the difference in individual transfer efficiency of the powders in the mixture in four

cases. Coarse powders typically had higher transfer efficiency than fine powders when coating individually. However, interactions during coating were found to have more influence on the separation than the difference in individual transfer efficiency.

Interaction generally caused a further increase in transfer efficiency of coarse powders and a further decrease in transfer efficiency of fine powders. When comparing whether composition or differences in size are more important, the results for nonelectrostatic and electrostatic coating were opposite, and the results for NaCl were opposite the results for the less dense powders. Being in a mixture did not change the amount of dust formed.

3.2 Introduction

Many foods, especially snack foods, are coated with powdered seasonings in order to enhance their flavor and increase product variety. A seasoning is usually a mixture of at least two powder ingredients because of the unique function of each ingredient. The most commonly used ingredients are salt, filler such as maltodextrin and corn flour, dairy powders such as cheese powders or sour cream powders, dehydrated vegetable powders, spices, compounded flavors, flavor enhancers, sweeteners, acids, color, processing aids such as vegetable oil or silicon dioxide, and antioxidants (Seighman 2001). In spite of the high frequency of use, research related to coating characteristics with powder mixtures is still limited.

Separation of powder in a mixture is undesirable. Separation can negatively affect flavor and appearance of the final product. While the composition of the powders contributes to the separation of mixtures (Likitwattanasade and Barringer 41

2013), particle size has been reported to cause separation in a mixture as well (Ye and

Domnick 2003, Somboonvechakarn and Barringer 2009). Differences in particle trajectory cause the separation in the mixture containing powders with different size.

Coarse powders fall faster than fine powders because of gravity force and miss the target, leading to separation (Ye and Domnick 2003). A greater transfer efficiency for coarse powders than fine powders due to gravity force was also found by others

(Ricks and Barringer 2002, Mayr and Barringer 2006, Xu and Barringer 2008). Not only differences in targeting loss cause separation, but also the differences in adhesion loss cause separation, particularly when the mixture is coated electrostatically

(Somboonvechakarn and Barringer 2009). However, some studies showed that there was no separation when a mixture with different size powders was coated (Yousuf and Barringer 2007).

Most powders used for coating are very small and cause dust during coating.

Dust is undesirable because it causes powder waste, overuse of powders, additional time and cost for cleaning, and respiratory distress to the employees. Several physical properties affect dustiness (Ricks and Barringer 2002, Ratanatriwong and others

2003). With increasing particle size and density, dustiness increases (Ricks and

Barringer 2002, Ratanatriwong and others 2003). As the powder becomes more cohesive, dustiness decreases (Ratanatriwong and others 2003). Although there are no studies on the effect of food powder mixtures on dustiness, electrostatic coating is generally known to be efficient in reducing dust during powder coating (Ricks and

Barringer 2002, Ratanatriwong and others 2003, Biehl and Barringer 2004).

Electrostatic powder coating has been used to improve the coating performance.

It has been shown to reduce powder loss, produce more even distribution and reduce 42

the amount of dust (Moore 1968). During electrostatic coating, powders pass through the corona discharge area, where the intense voltage gradient ionizes the air around the negative electrode, and are charged by negative ions. The charged powders are accelerated by electrostatic force according to Coulomb's law to deposit onto the nearest grounded target (Bailey 1998). Gravitational force, which depends on the powder mass, and electrostatic force cause powders to deposit on food products, resulting in an increase in transfer efficiency (Miller and Barringer 2002, Yousuf and

Barringer 2007, Biehl and Barringer 2004) and a reduction of dust (Rick and

Barringer 2002, Ratanatriwong and others 2003, Biehl and Barringer 2004). Even though electrostatic coating is more efficient than nonelectrostatic coating in individual coating, it can increase separation when mixtures are coated. In a mixture of coarse and fine powders, differences in adhesion loss increase the separation. Fine powders with a higher charge adhere well on the target and increase the adhesion loss of coarse powders by creating a charged layer and repelling the incoming coarse powders (Somboonvechakarn and Barringer, 2009). However, in a mixture with similar size powders, using electrostatic coating decreased separation by decreasing targeting loss and adhesion loss of powders in the mixture (Somboonvechakarn and

Barringer 2011, Likitwattanasade and Barringer 2013).

The objective of this study was to determine whether separation occurs when a mixture containing different sizes of powder was coated nonelectrostatically and electrostatically. The causes of the separation: differences in individual transfer efficiency and the interaction of powder during coating were determined. In addition, the amount of dust generated was measured during coating powders individually and coating as a mixture.

3.3 Materials and Methods

3.3.1 Powder samples

The mixtures used in this study consisted of two powders with the same composition but different in particle size: fine and coarse. The fine powders were

NaCl (325 Extra fine salt, Morton International Inc., Chicago, IL, U.S.A.), KCl

(Fisher Scientific, Fair Lawn, NJ U.S.A.), powdered sugar (Dixie Crystal, Sugar

Land, TX U.S.A), rice starch (National starch and chemical Ltd., Thailand)

Maltodextrin (Maltrin M700, GPC, Muscatine, IA U.S.A.), whey protein isolate

(Proliant, Iso-Chill 9000, Ames, IA, U.S.A.), casein (acid casein 720, Fonterra,

Rosemont, IL, U.S.A.), and soy protein (toasted soy grits, ADM, Decatur, IL, U.S.A.).

Among these eight powders, KCl, rice starch, maltodextrin, whey protein, casein and soy protein were ground using an ultra-centrifugal mill (ZM100, Retsch GmbH &

Co.KG, Clifton, NJ, U.S.A.) three times at 18,000 rpm to reduce particle size. The coarse powders were NaCl (Alberger flake salt, Cargill Inc., Minneapolis, MN,

U.S.A.), KCl (Sigma-Aldrich, St. Louis, MO, U.S.A.), granulated sugar (Domino,

Yonkers, NY, U.S.A.), rice starch, maltodextrin, whey protein isolate (Proliant, Iso-

Chill 9010, Ames, IA, U.S.A.), casein and soy protein. The mean diameters of the powders were measured using the Malvern Mastersizer (X standard bench, Malvern

Instrument Ltd., Worcestershire, U.K.). The volume mean diameter D[4,3] of each powder was reported. Because the water activity of the powder has been shown to affect the coating efficiency, the powders were equilibrated over saturated magnesium chloride solution (32.8% relative humidity) in sealed desiccators until used. The mixtures contained 20g of the fine powder mixed with 20g of the coarse powder.

Forty grams of each powder were used for individual coating. 44

3.3.2 Coating process

An electrostatic powder-coating machine (Terronics Development Corp.,

Elwood, IN, U.S.A.) was used to coat the food powders on an aluminum tray

(Figure 3.1). A previous study found that targeting loss is the main factor in powder loss during coating but there was little effect of adhesion loss (Likitwattanasade and

Barringer 2013). Thus this study focused only on the targeting loss. Adhesion loss was prevented by using a tray with a 2 cm height edge as a target. No voltage was applied for nonelectrostatic coating and −25 kV was used for electrostatic coating. An air compressor (Hitachi EC79, Hitachi Koki Co., Ltd., Tokyo, Japan) was used to supply airflow to drive the powder through the coating chamber. All experiments were carried out at 30–35% relative humidity and 20–25°C. Either individual powders or mixtures (40 g) were blown through the coating chamber at an air velocity of

3.3 m/s. A vibratory feeder (Syntron Magnetic Feeder, Model FTO-C, FMC

Corporation, Homer City, PA, U.S.A.) was used to control the consistency of the powder feeding.

Figure 3.1 Coating setup

3.3.3 Determination of the Amount of Powder in a Mixture after Coating

In order to identify how much fine and coarse powders were in the deposited mixture, the cutoff particle size (particle size used to differentiate fine powder from coarse powder) was determined using the malvern mastersizer. The fine and coarse powders were individually measured to determine their particle size distribution. The cut off particle size was selected to produce the least overlap or error of percent fine and coarse powders when coated individually. To ensure that the selected cutoff also works for a mixture, the cutoff was tested with mixtures of the known concentration.

The error produced by using the cut off was between 0-10%.

3.3.4 Separation

Separation was defined as a significant change in percentage between fine and coarse powders from the percentage entering the coating chamber, which was 46

50:50%. The amount of fine and coarse powders on the target when they were coated individually was used to calculate the predicted percentage between fine and coarse powders. The separation would be a result of the individual transfer efficiency when the ratio between the percentage of coarse and fine powders is different from the original ratio. Additionally, the separation would be a result of the interaction between fine and coarse powders when the ratio of percentage of powders after coating a mixture is different from the predicted one calculated from individual transfer efficiency.

3.3.5 Composition vs Particle size

Transfer efficiency of fine powders from previous study (Likitwattanasade and

Barringer 2013) was compared to the transfer efficiency of fine powders from this study to determine the effect of composition versus particle size on transfer efficiency. The mixtures used in previous study were the mixtures of two fine powders, which were different in composition but similar in size. Any changes occurred during coating these mixtures were assumedly caused by the difference in composition while any changes occurred during this study were assumedly caused by the difference in particle size.

3.3.6 Dustiness

During coating, dust was collected by a dust collector consisting of a cassette, cellulose support pad, 5.0 µm polyvinyl chloride filter (SKC Inc., Eighty Four, PA,

U.S.A.) and air sampling pump (Gillian Model HFS 513A, Sensidyne, Clearwater,

FL, U.S.A.). The pumps were operated at 4 liters per minute. Dust was collected from two different locations at the outlet of the coating chamber for the entire coating time

(3.30 min) (Figure 3.1). The reported amount of the dust for each replicate was a 47

summation of the dust amounts from these two locations.

3.3.7 Statistical analysis

Independent two tailed T-test with unequal variance and one-way analysis of variance ANOVA with post-hoc analysis using Tukey’s test were performed. A p- value of 0.05 was used to indicate significantly different results.

3.4. Results and Discussion

3.4.1 Separation

Separation is defined as a change in the proportion of powder ingredients on the final product compared to the proportion entering the coating chamber. Any powder separation can negatively affect the flavor profile and the appearance of the final product. In this study, the worst separation was with NaCl, at 79:21 (coarse: fine). The other mixtures were between 67:33 and 45:55. Separation was found in nine out of sixteen mixtures during either nonelectrostatic or electrostatic coating

(Figure 3.2). In those nine mixtures, there was always more coarse powder than fine powder on the target. Others have also reported that a difference in particle size causes separation in a mixture (Ye and Domnick 2003).

There are two main sources of separation: differences in individual transfer efficiency and interaction between the powders in a mixture. Four out of the nine mixtures that showed separation were predicted to show separation based on differences in individual transfer efficiency (Figure 3.2). Transfer efficiency of most fine powders was different from coarse powders when they were coated individually

(Figure 3.3). For many powders, coarse powders had higher transfer efficiency than

the corresponding fine powders during both nonelectrostatic and electrostatic coating.

The coarse and fine powders used in this study were the same in composition, thus the larger the powders, the greater the mass. Coarse powders with higher mass are more affected by gravitational force than fine powders, resulting in a higher amount of coarse powders depositing on the target. In addition, the lighter fine powders were more likely to be blown out as dust by the air used to disperse the powder mixture into the coating chamber. A higher transfer efficiency of coarse powder than fine powders has been reported by others (Ricks and others 2002, Biehl and Barringer

2004, Yousuf and Barringer 2007). In a previous study, the difference in individual transfer efficiency was the biggest cause of separation in mixtures of NaCl and starch

(Somboonvechakarn and Barringer 2009).

The second source of separation is the interaction between fine and coarse powders during coating. The interaction caused separation in seven out of the nine mixtures that showed separation (Figure 3.2). When the interaction caused a separation, it was generally because the interaction further decreased the transfer efficiency of fine powders while further increasing the transfer efficiency of coarse powders, compared to the individual transfer efficiency (Figure 3.3). A decrease in transfer efficiency of fine powders and an increase in transfer efficiency of coarse powders in starch mixture were also observed by others (Somboonvechakarn and

Barringer 2009). For some powders, such as NaCl, KCl and soy protein, the changes in transfer efficiency was significant, while for other powders the trend was the same but was not statistically significant. The change in transfer efficiency could be due to midair collisions between powders during coating. According to momentum conservation, when a small object collides with a large object, the velocity of the

small object increases and the velocity of the large object decreases. In a mixture, fine powders collide with coarse powders in a mixture during coating. These fine powders are scattered at a higher velocity and thus missed the target, resulting in a decrease in the deposition of the fine powders during coating. Opposite to the fine powders, the velocity of coarse powders decreased after collision, resulting in an increase in the deposition of the coarse powders on the target. The increase in transfer efficacy of coarse powders and decrease in transfer efficiency of fine powders was significant in

NaCl and KCl mixtures, possibly due to a difference in density of the powders. The density of NaCl and KCl is higher than the other powders (Table 3.1). The momentum of high density powders is higher than low density powders when collision occurs, thus the scattering between the powders will be greater in a mixture of high density powders than low density powders.

However, Somboonvechakarn and Barringer (2009) found interaction decreased the separation in the mixture of NaCl during nonelectrostatic coating.

Based on momentum conservation, expected result when coating a mixture of different size powders is a higher targeting loss of fine powders and a greater transfer efficiency of coarse powders, which consequently results in an increase of separation.

In their study, transfer efficiency of fine NaCl in the mixture did not change from individual coating, while the transfer efficiency of coarse NaCl decreased. Fine NaCl used in their study (44μm) was smaller than fine NaCl used in our study (58μm). Thus no change in transfer efficiency of fine NaCl in their study may be due to the fact that

70% of fine NaCl was already lost during individual coating, so there was a little loss that could occur during coating the mixture. For the coarse NaCl, the reason why the transfer efficiency decreased after coating NaCl was reported to be unknown. It

should be pointed out that interaction is going to increase or decrease the separation, is also depending on the particular size of powders. Interaction in the same mixture such as a mixture of NaCl could either increase or decrease the separation.

Besides the particle size, the amount of fed powders between these two studies was different. Forty gram of powders was used in our study while three gram of powders was used in their study. Frequency of collision during coating may be higher in our study than their study and possibly caused a difference in deposition of powders.

100% Fine/Nonelectrostatic 90% Coarse/Nonelectrostatic 80% Fine/Electrostatic 70% Coarse/Electrostatic 60%

50%

40%

30%

20% 52

10%

Proportion of coarse and fine powders on the target the on powders fine and ofcoarse Proportion NaCl KCl Sucrose Starch Maltodextrin Whey protein Casein Soy protein

Figure 3.2 Proportion of coarse and fine powder on the target after nonelectrostatic and electrostatic coating ( Significant

difference from 50:50%, Separation predicted by significant difference in individual transfer efficiency, separation was

significantly better than predicted, separation was significantly worse than predicted. (Predicted separation was calculated

based on the transfer efficiency of the powders when they were coated individually.)

NaCl a 100 KCl a a ab 100

ab a a NE/ Individual coating/ Fine

b 80 80 b NE/ Individual coating/ Coarse c c c NE/ Mixture coating/ Fine 60 60 cd d NE/ Mixture coating/ Coarse c 40 40 E/ Individual coating / Fine d E/ Individual coating/ Coarse 20 20

Transfer efficiency (%) efficiency Transfer E/ Mixture coating/ Fine Transfer efficiency (%) efficiency Transfer E/ Mixture coating/ Coarse 0 0

Sucrose Starch Maltodextrin 100 100 100

a a

ab a abc bcd ab a bc 80 80 ab c 80 abc bcd ab cd d c bcd cd 60 d 60 e e 60

40 40 d

40 d

Transfer efficiency (%) efficiency Transfer Transfer efficiency (%) efficiency Transfer Transfer efficiency (%) efficiency Transfer 20 20 20

0 0 0

Whey protein Casein Soy protein 100 100 100 a

a a a a 80 a 80 a ab 80 a b ab bc c bc 60 b b 60 c 60 cd cde e b de e 40 40 40

c Transfer efficiency (%) efficiency Transfer 20 (%) efficiency Transfer 20 Transfer efficiency (%) 20

0 0 0

Figure 3.3 Transfer efficiency of NaCl, KCl, sucrose, starch, maltodextrin, whey protein, casein and soy protein (aSamples of the same powder with different letters are significantly different. NE = Nonelectrostatic coating, E = Electrostatic coating)

Table 3.1 Physical properties of NaCl, KCl, sucrose, starch, maltodextrin, whey protein, casein and soy protein

Particle size Flowability Bulk Density Resistivity Property (µm) (Degree) (g/cm3) (x109Ωm)

Powder Fine Coarse Fine Coarse Fine Coarse Fine Coarse

NaCl 58.2fg 278c 55.1d 48.6fg 1.08b 1.17a 0.0000655c 0.0140c

KCl 59.9fg 366b 65.8b 48.0fg 1.15a 1.07b 2.07c 3.81c

Sucrose 56.1fg 357b 70.3a 45.9gh 0.744e 0.955c 24.1c 614a

Starch 62.0fg 244d 53.0de 46.8gh 0.659f 0.551h 1.90c 1.77c 54 fg b c de h j c c Maltodextrin 61.9 357 58.6 52.4 0.584 0.183 5.29 2.53

Whey protein 51.8g 250d 55.0d 53.0de 0.590gh 0.429i 90.9b 11.3c

Casein 93.8e 391a 50.2ef 46.7gh 0.801d 0.631fg 4.10c 1.80c

Soy protein 68.0f 400a 61.4c 44.4h 0.740e 0.794d 2.28c 2.59c

a Samples in the same category of physical property with different letters are significantly different.

3.4.2 Composition versus Particle size

While particle size caused separation in mixtures containing powders with different size, separation is also found in mixtures of powders with the same size but different composition (Likitwattanasade and Barringer 2013). Composition and particle size both have a significant effect on the transfer efficiency of powders in the mixtures. The effects of particle size and composition on transfer efficiency for nonelectrostatic and electrostatic coating were opposite, and the results for NaCl were opposite the results for the less dense powders (Table 3.2). In nonelectrostatic coating with fine sucrose, starch or protein, the fine powders had higher transfer efficiency when mixed with the coarse powder, and lower transfer efficiency when mixed with powders of a different composition. The lowest transfer efficiency occurred when the powder was mixed with NaCl. In contrast, NaCl had lower transfer efficiency when mixed with the coarse powder and higher transfer efficiency when mixed with different composition, opposite to the results for the less dense powders.

In electrostatic coating, the opposite occurred and for fine sucrose, starch and protein, the fine powders had lower transfer efficiency when mixed with the coarse powder, and higher transfer efficiency when mixed with powders of different composition. However, NaCl had lower transfer efficiency when mixed with powders of different composition and higher transfer efficiency when mixed with the coarse powders. These both particle size and composition had an impact on the transfer efficiency of powders in the mixture. To achieve the highest transfer efficiency, both the particle size of powders in the mixture and the selection of the powders to be mixed, must be designed.

Transfer efficiency is influenced by the properties of powder such as particle

size, flowability, density and resistivity (Biehl and Barringer 2003; 2004, Mayr and

Barringer 2006, Halim and Barringer 2007, Ratanatriwong and Barringer 2007,

Yousuf and Barringer 2007). During nonelectrostatic coating, as particle size increases, nonelectrostatic coating transfer efficiency increases (Ricks and others

2002, Miller and Barringer 2002, Yousef and Barringer 2007, Ratanatriwong and

Barringer 2007). In this study, individual transfer efficiency of most coarse powders was higher than fine powders, as expected (Figure 3.3). Only KCl and soy protein powders showed no difference in transfer efficiency between fine and coarse powders.

Flowability is one of the properties that have been reported to effect on transfer efficiency. Flowability is partially influenced by particle size (Fitzpatrick and others

2004). Small powders are more cohesive due to their high surface to volume ratio, causing low flowability in powders (Ratanatriwong and Barringer 2007, Biehl and

Barringer 2003). In this study, the flowability of powders which was determined using

Angle of repose method showed that most fine powders had a lower flowability

(higher in a degree of repose) than coarse powders (lower in a degree of repose)

(Table 3.1). When compared between fine and coarse of the same composition powders, the nonelectrostatic transfer efficiency of fine powders generally was lower than coarse powders either during individual coating and coating as a mixture (Figure

3.3). This may be a result of particle size as discussed previously and/or the result of flowability. One study reported that among particle size, flowability and powder charge, particle size is the most influential factor on the transfer efficiency (Ricks and others 2002).

The effect of density has also been reported by a few researchers (Ricks and others 2002, Biehl and Barringer 2003; 2004, Yousuf and Barringer 2007). Higher

density powders have a higher mass than powers with lower density of the same size.

According to the force of gravity, high density powders with a higher mass are more likely to deposit onto the target than low density powders, resulting in a higher transfer efficiency. However, the powders used in the mixture in this study were different in size, the bulk density may not be a related factor used to explain the transfer efficiency of the powders.

During electrostatic coating, resistivity of powder was reported to play a role on transfer efficiency (Bailey 1998, Halim and Barringer 2007). Powders with higher resistivity (resistivity > 1013Ωm) produce higher transfer efficiency than powders with lower resistivity (Bailey 1998). The low resistivity powders are charged better, but the charges are lost quickly. In this study, there was no powder that the resistivity was higher than 1013Ωm (Table 3.3). The resistivity of most powders was 1010 - 1011Ωm.

Powders in this resistivity range were reported to be difficult to predict their transfer efficiency because the transfer efficiency depends on many factors such as composition and surface chemistry (Bailey 1998). When compared the resistivity between fine and coarse powders, most powders showed no significant difference between the resistivity (Table 3.1). This indicated that the resistivity would not have any different effects on fine and coarse powders during electrostatic coating.

However, the magnitude of an increase in the transfer efficiency of fine powders was higher than coarse powders either during individual coating and coating as a mixture

(Figure 3.3). This was expected because charge to mass ratio of powder increases when the particle size decreases (Ratanatriwong and Barringer 2007).

Table 3.2 Transfer efficiency of fine sucrose, starch, soy protein and NaCl when coating as a mixture with

powders of different composition or different particle size

Transfer efficiency (%) when powder mixed with:

Different size powder Different composition powder*

Coarse Fine NaCl Fine sucrose Fine starch Fine soy protein

Nonelectrostatic coating

Fine sucrose 62.75a 36.48c - 43.56bc 53.62ab

Fine starch 52.35ab 46.10c 55.33a - 49.95bc

Fine soy protein 42.16b 22.88d 61.75a 31.54c - 58 c b b a Fine NaCl 25.14 - 31.91 33.95 37.29

Electrostatic coating

Fine sucrose 55.05b 80.79a - 87.50a 78.73a

Fine starch 67.82b 89.48a 91.00a - 83.23a

Fine soy protein 41.86b 53.38b 70.60a 70.69a -

Fine NaCl 86.51a - 46.88b 58.62b 98.88a

aSamples in the same row (same powder and coating treatment) with different letters are significantly different.

* Data was taken from Likitwattanasade and Barringer 2013.

3.4.3 Dustiness

When fine or coarse powders were coated individually, the dust produced by fine powders was higher than coarse powders, as expected (Figure 3.4). Using electrostatic coating significantly decreased the amount of dust, particularly for the fine powders. Fine powders can achieve a greater charge to mass than coarse particles, so they were more influenced by the electrostatic force (Mazumder and others 1997). The potential of using electrostatic coating to achieve dust reduction during food powder coating was also reported by others (Ricks and others 2002,

Sumawi and Barringer 2005, Ratanatriwong and others 2009).

When the mixtures were coated, there was little change in the amount of dust when compared to the average of fine and coarse powders during either nonelectrostatic or electrostatic coating. Twelve out of the sixteen mixtures had no significant difference between the amount of dust produced during coating the mixture and the average of individual fine and coarse powders. Among the ones that were different, the decrease in dust for the whey protein mixture and the increase in dust for the soy protein mixture were expected based on the transfer efficiency. Dust particles are usually smaller than 75µm (World Health Organization, 1999), so only fine powder is responsible for dustiness. In a mixture, the transfer efficiency increased for fine whey protein powders and decreased for fine soy protein powders, meaning there was less and more powder lost as dust respectively.

a 120 a a Fine powder/ Nonelectrostatic coating

Coarse powder/ Nonelectrostatic coating 100 a Mixture/ Nonelectrostatic coating

Fine powder/ Electrostatic coating 80 b Coarse powder/ Electrostatic coating

b Mixture/ Electrostatic coating 60

60 b

a b Amount of (mg) dust Amount c 40 a ab bc bc a c b cd b c c 20 a d bc d c b b c c c bc d bc b d c c d c c c c e e c c c 0 NaCl KCl Sucrose Starch Maltodextrin Whey protein Casein Soy protein

Figure 3.4 Dust collected during nonelectrostatic coating and electrostatic coating of the mixtures containing different size

powders. (aSamples of the same powder with different letters are significantly different. Significant difference from the

amount of dust predicted when the powders were coated individually)

3.5 Conclusion

In this paper, eight mixtures of common food powders, which were the same in composition but different size, were coated nonelectrostatically and electrostatically.

Differences in particle size of the powder mixture caused separation, however it was dependent on the exact powder. Differences in individual transfer efficiency and interaction between powders during coating were the causes of separation. In general, interaction increased the separation. Dustiness was generally not affected by being in a mixture of different size. Using similar size powders, optimizing the coating parameters and adjusting the proportion of original powders in the mixture may be needed to achieve the desired taste and appearance when coating a mixture.

3.6 Practical Applications

A mixture containing different size powders has a propensity to separate. The separation was caused by the difference in individual transfer efficiency of powders in a mixture and/or the interaction between the powders. To reduce the difference in individual transfer efficiency and effect of interaction, powders with similar size should be used. However, the effect of compositional differences also needs to be considered. Coating parameters such as air velocity could be optimized to decrease the differences in individual transfer efficiency and interactions. Additionally, adjusting the proportion of original powders in the mixture may be needed to achieve the desired taste.

CHAPTER 4

ELECTROSTATIC SALTING OF POTATO CHIPS

4.1 Abstract

Consumers, food companies, and public health agencies all express continuing interest in reducing sodium content in foods efficiency of salting of potato chips was recently investigated. Since only part of NaCl on a potato chip is sensed by human tongue receptors, smaller NaCl particles can be added at lower concentration to decrease total NaCl content per chip while maintaining taste perception quality. In general, smaller particle size salt (<100µm) dissolves faster and produces a higher salt sensation versus larger particle size salt. However, smaller NaCl creates more dust during the salting process. Electrostatic coating has been shown to produce significant dust reduction. Thus the objective of this study was to determine the effect of using electrostatic coating by different methods on the dustiness, evenness, and transfer efficiency during salting of potato chips with extra fine NaCl. Potato chips were salted non-electrostatically (0 kV) and electrostatically (25-95 kV) with using three delivery different delivery methods: an auger system, a spray gun and a scarf plate inside a tumble drum at lab scale. The dust was collected during the salting process and salted chips were sampled at 12 different locations inside the drum. There was no significant difference in the percent of NaCl on the chips when nonelectrostatic or electrostatic 62

salting by spray gun or scarf plate was used. During nonelectrostatic salting, the spray gun produced much more dust than the scarf plate. However, when electrostatics (≥50 kV) was applied, the amount of dust produced by the spray gun decreased up to 88%.

NaCl distribution across the bed of chips was more even when using the spray gun than the scarf plate, indicating the need for air assist system for scarf plate to spread the salt. With the spray gun the evenness of the distribution decreased at high voltage

(≥75 kV). Electrostatics significantly reduces dusting irrespective of the application system.

4.2 Introduction

Salt (NaCl) is a common constituent used in food production and processing impacting many final product attributes including but not limited to: stability

(physical and microbiological), and palatability (Hui 2007, DeSimone and others

2013). In European and North American populations it is estimated that as much as

70% of dietary sodium exposure originates from food manufacturing sources (Brown and others 2009). In some populations, excessive sodium intake has been associated with an increase in blood pressure, increase risk of stroke, heart attack, gastric cancer, calcium-containing kidney stones and osteoporosis (Strazzullo and others 2009,

Cappuccio and others 2000). However the reproducibility and consistency of these findings and related sodium exposure across populations has been variable leading to continued debate about the most appropriate public health recommendations around dietary sodium intake (O’Donnell and others 2013). Current dietary guidelines for

Americans recommend that adults consume no more than 2,300 mg of sodium per day

and that certain individuals/groups further limit daily amounts to no more than 1,500 mg per day (USDA/HHS Dietary guidelines for Americans 2010). However, less than

1% of Americans achieve these recommendations with most consuming about 1.5-2- fold current recommended daily intake, according to the Centers for Disease Control and Prevention.

Reducing sodium utilization in manufacturing processes is a goal of many food companies, in spite of the continued debate regarding appropriate dietary sodium exposure (DeSimone and others 2013). While it is estimated that contribution of potato chips to overall population sodium exposure is lower than many other foods, research has been done attempting to reduce the amount of sodium in potato chips

(Drewnowski and Rehm 2013). Only about 20-50% of the salt on a chip actually dissolves on the tongue before the chip is chewed and swallowed, and the remaining salt is swallowed without contributing to the taste (McKay 2010). In addition, the particle size of salt (NaCl) has a significant impact on the delivery rate of sodium into the saliva, the maximum concentration of sodium in the saliva, the maximum perceived saltiness, and the saltiness onset time (Rama and others 2013). Finer salt particles (<100 µm) have a faster delivery of sodium per unit sodium, greater concentration of sodium in the saliva, and greater saltiness sensory perception, compared to coarse salt. Thus, fine salt can be added at a lower concentration to decrease total NaCl while achieving the same salty taste. However, fine NaCl creates more dust during the salting process.

Dustiness is undesirable because it causes powder waste, overuse of powders, and reduction in production efficiency, additional dust cleaning time and cost, equipment corrosion, and respiratory distress to the employees. Electrostatic coating 64

is known to be efficient in reducing dust during powder coating (Ricks and others

2002, Ratanatriwong and others 2003, Biehl and Barringer 2004). During electrostatic coating, powders pass through the corona discharge area, where the intense voltage gradient ionizes the air around the high negative potential electrode, and are charged by negative ions. The charged powders are accelerated by electrostatic force and deposit onto the nearest grounded target (Bailey 1998). Since the powder seek the target rather than being suspended in the air, the amount of dust in the air decreases dramatically in air. Electrostatic coating also provides significantly better transfer efficiency (Biehl and Barringer 2003, Miller and Barringer 2006, Yousuf and

Barringer 2007) and adhesion over non-electrostatic coating (Halim and Barringer

2007, Sumonsiri and Barringer 2011, Huang and Barringer 2012).

To achieve top coating performance, factors affecting the coating efficiency including coating equipment (Biehl and Barringer 2004), coating conditions (Halim and Barringer 2006, Xu and Barringer 2008), charging systems (Mayr and Barringer

2006, Sumawi and Barringer 2007), powder properties (Ricks and others 2002, Biehl and Barringer 2003, Ratanatriwong and Barringer 2007) and target properties

(Sumonsiri and Barringer 2010; 2011) have been studied. There are several different coating systems used in the industry with the auger, spray gun, and scarf plate being commonly used. The choice system depends on the powder flow characteristics. In the auger system, powders are conveyed by rotating auger through a pipe with holes where it is discharged over target product. In the spray system, powders are pumped from a feed hopper by compressed air, conveyed through a powder feed hose to a spray gun and propelled toward targets. In the scarf plate system, powders are dispensed from a vibrating scarf plate to targets by gravity force. Different equipment

will produce different coating profiles and amounts of dust. The objectives of this study are to compare three different salting systems and determine if electrostatic coating can be used to reduce dust, improve transfer efficiency, and salt distribution during salting of potato chips with extra fine salt (<20µm).

4.3 Materials and Methods

4.3.1 Potato chip salting

Unsalted chips (2.27 kg, Lay’s, Plano, TX, U.S.A.) were warmed in a heated cabinet (43°C, a common temperature when chips are salted) and were loaded into a tumble drum (Master Series TM Tumble Drum, Spray Dynamics, Ltd., St. Clair, MO,

U.S.A.). The drum was 2.13 m in drum cylinder length and 0.76 m diameter. The processing area was decreased to minimize the potato chip load by having a baffle ring in the middle of the drum, thus the processing area was 1.07 m × 0.76 m diameter. NaCl (Microsized® 95 Extra Fine Salt (<20µm), Cargill, Minneapolis, MN,

U.S.A.) was introduced to the drum by three different systems, auger system (Spray

Dynamics, Ltd., St. Clair, MO, U.S.A.), pneumatic spray gun (Sure Coat Manual

Powder Spray Gun, Part 302123D, Nordson Corporation, Amherst, OH, U.S.A.) at 0,

25, 50, 75 or 95 kV and scarf plate (Model SAS 1 DV, Oxford, England, U.K.) at 0,

35, 50, 75 or 85 kV. For the gun, the flow rate air pressure was set at 1.7x104 Pa (0.25 psi) and the atomizing air pressure was set at 1.7x104 Pa (0.25 psi). The auger system did not convey the fine salt so it was not studied further.

A vibratory feeder (Syntron Magnetic Feeder, Model FTO-C, FMC

Corporation, Homer City, PA, U.S.A.) was used to control the consistency of NaCl feeding. NaCl (25 g) was placed in the magnetic feeder placed in front of the drum.

Vibration of the feeder gradually fed the NaCl into the scarf plate or into a funnel attached to the feeding hole on the spray gun. The chips were salted in the drum tumbling at 8.2 rpm for 3 min 30 s. Chips from 12 locations in the drum were sampled to determine the salt content (Figure 4.1). The drum wall was cleaned by vacuum and wiped with cleaning paper between runs. The chip was coated under 45-55% relative humidity at 20-25°C. Runs were completed in duplicate for the spray gun treatment and triplicate for the scarf plate treatments.

4.3.2 NaCl determination

Target salt content of chips was 1.1% on weight basis. Amount of NaCl on the salted chips was determined using Mohr titration (Nielsen 2009). Chips (50 g) were crushed and transferred to a 500 mL Pyrex bottle. Hot deionized distilled water (70-

80°C, 100 mL) was added to the bottle to dissolve NaCl on the chips. The bottle was shaken for 30 s, allowed to rest for 1 min and shaken again for 30 s before being cooled down to room temperature. The salt solution was separated from the chips by filtering it through a Buchner funnel. The salt solution (50 mL) was transferred to a

250 mL volumetric flask with 1 mL of 5% K2CrO4 (indicator) and titrated with 0.075

M AgNO3 until first perceptible pale red-brown color due to precipitate of Ag2CrO4 appeared. Percent salt on the chips is the average of three replicates carried out for each treatment.

4.3.3 Dust determination

The dustiest location of each salting method was determined using an aerosol monitor (Casella, MicroDust, Aerosal monitoring system kit, Bedford, U.K.). The dustiest locations, which were at the back of the drum for the spray gun treatment and at the front for scarf plate treatment, were used to monitor dust produced during coating. The averages of dust amount (mg/m3) during 3 min 30 s were reported.

4.3.4 Statistical analysis

Independent two tailed T-test with unequal variance and one-way analysis of variance ANOVA with post-hoc analysis using Tukey’s test were performed.

Statistical significance was assumed at p= 0.05.

4.4 Results and Discussion

4.4.1. NaCl distribution

Distribution or evenness was determined by comparing the percent NaCl on the chips at 12 different locations (Figure 4.1). When using the spray gun with nonelectrostatic and low voltage electrostatic coating (25 and 50 kV), there was no significant difference in the percent NaCl among the 12 locations. At 75 and 95 kV, the NaCl deposited closer to the spray gun.

Electrostatic coating has previously been shown to cause more powder to be deposited closer to the nozzle (Yousuf and Barringer 2007, Somboonvechakarn and

Barringer 2009). At high voltage, the NaCl has a greater charge to mass ratio than at low voltage. At higher charge to mass ratio, there is greater acceleration by

electrostatic force to deposit closer to the nozzle. Similarly, the relative standard deviation (RSD) of the percent NaCl among the 12 locations was significantly increased at 75 and 95 kV compared to 0 and 25 kV, indicating a decrease in the evenness.

When using the scarf plate, differences in the percent NaCl among the 12 locations were found at all voltages (Figure 4.1). There was no significant difference in the NaCl distribution (RSD) when salting the chips non-electrostatically and electrostatically by the scarf plate. Similar to the spray gun, chips with a higher percent NaCl were found in the front row, except it occurred at all voltages. RSD of the percent NaCl when using the scarf plate was much higher (85-93%) than in the spray gun (13-34%). It was also noticed that a white ring of NaCl was found at the front of the drum because gravity caused most of the NaCl to fall directly underneath the feeder in nonelectrostatic coating with scarf plate (Figure 4.2), which decreased with the use of electrostatics. When the electrostatics was applied, the NaCl particles repelled each other in the air and deposited further into the drum.

Further, the air assist system in spray gun produced a more even distribution than salting with the scarf plate without air assist. It should be pointed out that a better distribution in scarf plate can be achieved with the air assist system to spread the salt evenly, which will be complemented by movement of chips from front to back in a continuous commercial salt application system.

Figure 4.1 Percent NaCl on potato chips in 12 different location of the tumble drum after salting the potato chips with 1.1% NaCl by using the spray gun and the scarf plate systems at a range of voltages. (aNaCl contents with different letters within each voltage and coating method are significantly different. ARSD values with different letters within the same coating method are significantly different.)

Figure 4.2 NaCl ring on the drum wall during nonelectrostatic salting of the potato chips by using the scarf plate.

4.4.2. Average % NaCl on the chip

In the potato chip industry, the salt content of the potato chips varies widely and in this study, 1.1% NaCl was added into the drum to determine how much NaCl deposited on the potato chips and how much NaCl was lost. Coating nonelectrostatically, the NaCl on the chips was 0.69% and 0.68% when using the spray gun and the scarf plate respectively, thus they had equal transfer efficiency (Figure

4.3). With electrostatic salting (25-95 kV), the spray gun produced 0.59-0.75% NaCl on the chips while the scarf plate produced 0.66-0.81% NaCl on the chips. Using either the spray gun or the scarf plate, there was no significant difference in the percent of NaCl on the chips when nonelectrostatic salting or electrostatic salting was used. Additionally, no significant difference in the percent NaCl was observed between using the spray gun or the scarf plate at all voltage levels. However, standard deviation of the percent NaCl on the chips when using the scarf plate was much 71

higher than the spray gun at all voltage levels. Thus, there was a more uneven distribution of NaCl across the chip bed when salting with the scarf plate due to lack of suitable mechanism to spread the salt.

It was expected that the percent NaCl on the chips would have increased with electrostatics. Some studies found an improvement with electrostatics, however a different electrostatic charging system was used (Biehl and Barringer 2003, Miller and Barringer 2006, Yousuf and Barringer 2007) or the coating occurred on a conveyor belt instead of tumble drum system (Ricks and others 2002, Ratanatriwong and Barringer 2003, Sumawi and Barringer 2005). One study that used the same spray gun system found an increase in transfer efficiency by up to 27% when coating with electrostatics at 95 kV (Mayr and Barringer 2006). However, the coating was carried out in a stationary coating system on graham crackers at 35% RH. Increasing surface oil content increases adhesion (Buck and Barringer 2007). When using electrostatic coating, no increase in adhesion force was found in samples with oily surfaces, such as pork rinds and potato chips (Halim and Barringer 2007). In addition, charging efficiency decreases as humidity increases resulting in a decrease in transfer efficiency (Xu and Barringer 2008). Thus, the higher surface oil content and humidity, or difference in coating systems, may have caused the lack of improvement in transfer efficiency in this study.

Figure 4.3 Percent NaCl on potato chips after salting the potato chips with 1.1%

NaCl by using the spray gun and the scarf plate systems. (aSamples with different letters are significantly different.)

4.4.3. Dustiness

Dust produced during the coating process is undesirable. During nonelectrostatic salting, the spray gun system produced more dust than the scarf plate system (Figure 4.4). Typically, there are three forces directing powders during powder coating: pneumatic force from the air supply, gravitational force and electrostatic force. Owing to the strength of pneumatic force from the spray gun, air turbulence was created within the drum and caused more dust from the spray gun than the scarf plate. Dust reduction has previously been found to be greater when using electrostatics with a pneumatic coating system rather than a non-pneumatic coating system (Biehl and Barringer 2004).

Many studies have shown the potential of using electrostatic coating to achieve dust reduction during food powder coating (Ricks and others 2002, Sumawi and Barringer 2005, Ratanatriwong and others 2009). In this study, the spray gun system showed a significant reduction in dust when using electrostatic coating, at 50 kV or higher (Figure 4.4). Dust was reduced by the mechanism of electrostatic force.

When the NaCl left the gun and passed through the corona discharge area where the surrounding air was ionized by corona, the NaCl became charged. The charged NaCl particles were accelerated by electrostatic force, according to Coulomb’s law (Bailey

1998), to deposit onto the nearest object, which was the chips in this situation. Using electrostatic salting, only the spray gun showed a decrease in dustiness while no dust reduction was found when using the scarf plate.

The laboratory scarf plate generated such low levels of dust that the effect of electrostatics on dusting was difficult to observe. Larger pilot and commercial scale scarf plates are known to generate much higher levels of dust. Yet the NaCl distributions in the spray gun demonstrated the effectiveness in combining a pneumatic assist with electrostatics for greater homogeneity. Thus, a scarf plate system with electrostatics and pneumatic assist should produce more controlled NaCl distributions while reducing dusting.

250 a Spray gun a Scarf plate

200

150

100

Dust (mg/m3) Dust b 50 bc bc c c c 0 0 20 40 60 bc 80 bc 100 Voltage (kV)

Figure 4.4 The dust amount produced during salting by the spray gun and the scarf plate (aSamples with different letters are significantly different.)

4.5 Conclusions

Electrostatic salting (≥50 kV) reduced the dust amount during salting of potato chips and spreading the salt with pneumatic air assist increased evenness of application. Using electrostatic salting did not improve the percent NaCl on the chips for either system. Electrostatics made salt distribution less even for the spray gun, but had no effect on evenness for the traditional scarf plate system. Unlike the first two systems, the auger was not able to move the 16µm particles of the extra fine salt.

Thus it could not be considered as an effective system even by changing screws for different particle sizes when spray gun and scarf plates handle much wider ranges of particles.

Scarf plates and spray guns are both suitable for salting applications. Yet a combined system of a scarf plate with electrostatics and pneumatic assist should yield 75

the best combination of dust suppression, manufacturing robustness, and homogeneity in salt distributions. Deployment of such a system is a business decision, but the authors believe the technical viability has been demonstrated.

4.6 Practical Applications

Three salting systems commonly used to salt potato chips were compared: auger system, spray gun, and scarf plate. The auger system is not suitable for fine salt as it compacted and discharged the salt in a completely inconsistent fashion. Salting with a spray gun created more dust, but distributed more evenly across the salting area than scarf plate. Electrostatic salting can be used to minimize dust when salting the potato chips with the spray gun, and improve even distribution when salting with the scarf plate.

References

AOAC Official analytical methods. Ultraviolet 280 nm Absorption Method. AOAC

International, Gaithersburg, MD, U.S.A., 1984.

Amefia AE, Abu-Ali JM, Barringer SA. Improved functionality of food additives with

electrostatic coating. Innov Food Sci Emerg Technol 7(3): 176-181., 2006.

Bailey AG. The science and technology of electrostatic powder spraying, transport

and coating. J Electrostat 45: 85-120., 1998.

Biehl HL, Barringer SA. Physical properties important to electrostatic and

nonelectrostatic powder transfer efficiency in a tumble drum. J Food Sci

68(8): 2512–2515., 2003.

Biehl HL, Barringer SA. Comparison of the effect of powder properties on coating

transfer efficiency and dustiness in two nonelectrostatic and electrostatic

systems. Innov Food Sci Emerg Technol 5(2): 191-198., 2004.

Bowling AR. A theoretical review of particle adhesion. In: Mittal KL, editor. Particles

on surfaces. Vol.1: detection, adhesion and removal. New York: Plenum

Press. p 129-142., 1998.

Brown IJ, Tzoulaki I, Candeias V, Elliott P. Salt intakes around the world:

implications for public health. Int. J. Epidemiol 38: 791–813., 2009.

Buck VE, Barringer SA. Factors dominating adhesion of NaCl onto potato chips. J

Food Sci 72(8): E435-E441., 2007.

Cappuccio FP, Kalaitzidis R, Duneclift S, Eastwood JB. Unraveling the links between

calcium excretion, salt intake, hypertension, kidney stones and bone

metabolism. J. Nephrol 13(3): 167-177., 2000.

Carr RL. Classifying flow properties of solids. Chem. Eng 1: 69-72., 1965.

Clark G. Processing: electrostatic coating technology for savory snack. Food Tech

Europe 2(3): 90-96., 1996.

De Jong JAH, Hoffmann AC, Finkers HJ. Properly determine powder flowability to

maximize plant output. Chem Eng Prog 95(4): 25-34., 1999.

Desimone, JA, Beauchamp GK, Drewnowski A, Johnson GH. Sodium in the food

supply: challenges and opportunities. Nutr Rev 71(1): 52-59., 2013.

Drewnowski A, Rehm CD. Sodium Intakes of US Children and Adults from Foods

and Beverages by Location of Origin and by Specific Food Source. Nutrients

5(6): 1840-1855., 2013.

Fitzpatrick JJ, Barringer SA, Iqbal T. Flow property measurement of food powders

and sensitivity of Jenike’s hopper design methodogy to the measured values. J

Food Eng 61: 399-405., 2004.

Geldart D, Abdullah EC, Hassanpour A, Nwoke LC, Wouters I. Characterization of

powder flowability using measurement angle of repose. China Particuology

4(3-4): 104-107., 2006.

Goula AM, Adamopoulos KG. Spray drying of tomato pulp in dehumidified air: II.

The effect on powder properties. J. Food Eng 66(1): 35-42., 2005.

Halim F, Barringer SA. Electrostatic adhesion in food. J Electrostat 65(7): 168-173.,

2007.

Huang Y, Barringer SA. Adhesion of food powders with nonelectrostatic and

electrostatic coating. J Food Proc Eng 35(2): 264-277., 2011.

Hughes JF. Electroparticle charging: industrial and healthcare applications. Taunton,

Somerset, England: Research Studies Press, Ltd. 170p., 1997.

Hui YH. Handbook of Food Products Manufacturing. Hoboken, New Jersey: John

Wiley & Sons, Inc., 2007

Likitwattanasade T, Barringer SA. Separation of powder mixtures during electrostatic

and nonelectrostatic coating. J Food Process Eng 36(6): 731-738., 2013.

Lui F, Cao Z, Wang H, Liao X. Changes of tomato powder qualities during storage.

Powder Tech 2004(1): 159-166., 2010.

Mayr MB, Barringer SA. Corona compared with triboelectric charging for

electrostatic powder coating. J Food Sci 71(4):171–177., 2006.

Mazumder MK, Wankum DL, Sims RA, Mountain JR, Chen H, Pettit P, Chaser T.

Influence of powder properties on the performance of electrostatic coating

process. J Electrostat 40/41: 369-374., 1997.

McKay B. PepsiCo develops ‘Designer salt’ to chip away at sodium intake. The wall

street journal., 2010.

2283791566> Accessed on: 17 April 2015.

Miller MJ and Barringer SA. Effect of sodium chloride particle size and shape on

nonelectrostatic and electrostatic coating popcorn. J Food Sci 67(1):198–201.,

2002.

Moore AD. Electrostatics. New York: Doubleday & Company Inc. p 127-131., 1968.

O’Donnell MJ, Mente A, Smyth A, Yusuf S. Salt intake and cardiovascular disease:

why are the data inconsistent? Eur. Heart J 34: 1034-1040., 2013.

Peleg M. Physical characteristics of food powders. In: Peleg M, Bagley EB, editors.

Physical properties of foods. Westport CT: AVI Publishing Co. P 293-324.,

1983.

Rama R, Chiu N, Da Silva MC., Hewson L, Hort J, Fisk ID. Impact of salt crystal size

on in-mouth delivery of sodium and saltiness perception from snack foods. J.

Texture Stud 44(5): 338-345., 2013.

Ratanatriwong P, Barringer SA, Delwiche J. Sensory preference, coating evenness,

dustiness and transfer efficiency of electrostatically coated potato chips. J

Food Sci 68(4): 1542-1547., 2003.

Ratanatriwong P, Barringer SA. Particle size, cohesiveness and charging effect on

electrostatic and nonelectrostatic powder coating. J Electrostat 65(7): 704-

708., 2007.

Ratanatriwong P, Suwansri S, Barringer SA, Tanasukarn P. Effect of electrostatic

coating on consumer acceptance and process efficiency of seasoned coated

snacks. J. Food Ag-Ind 2(01): 51-64., 2009.

Ricks NP, Barringer SA, Fitzpatrick JJ. Food powder characteristics important to

nonelectrostatic and electrostatic powder coating and dustiness. J Food Sci

67(6): 2256–2263., 2002.

Seighman J. Snack food seasonings. In: Lusas EW, Rooney RW, editor. Snack foods

processing. Lancaster PA: Technomic Publishing Co Inc. p 495–527., 2001.

Somboonvechakarn C, Barringer SA. Effects of particle size on mixtures during

nonelectrostatic and electrostatic powder coating. J Food Process Eng 34(4):

1086-1100., 2009.

Somboonvechakarn C, Barringer SA. Effect of particle density and composition on

mixtures during nonelectrostatic and electrostatic powder coating. J Food

Process Eng 35(2): 236-249., 2011.

Strazzullo P, D’Elia L, Kandala N-B, Cappuccio FP. Salt intake, stroke and

cardiovascular disease: meta-analysis of prospective studies. BMJ 339:

b4567., 2009.

Sumawi H, Barringer SA. Positive vs. negative electrostatic coating using food

powder. J Electrostat 63(6-10): 815-821., 2005. 83

Sumonsiri N, Barringer SA. Effect of powder and target properties on wrap around effect during coating. J Food Sci 75(8): E537–E543., 2010.

Sumonsiri N, Barringer SA. Effect of sodium chloride and food target properties on

nonelectrostatic and electrostatic coating. J. Electrostat 69(6): 578-586., 2011.

U.S. Department of Agriculture, and U.S. Department of Health and Human Services.

Dietary guidelines for Americans (7th edition). Washington, DC, U.S.A.

Government Printing Office., 2010.

Wang F, Martinuzzi R, Zhu J. Experimental study of particle trajectory in

electrostatics powder coating process. Powder Technol 150(12): 20-29., 2005.

World Health Organization. Prevention and control exchange (PACE). Hazard

Prevention and Control in the Work Environment: Airborne Dust.

WHO/SDE/OEH/99.14. Geneva, Switzerland. p 1., 1999. Available at: <

http://www.who.int/occupational_health/publications/en/oehairbornedust.pdf

>. Accessed on: 17 April 2015.

Ye Q, Domnick J. On the simulation of space charge in electrostatic powder coating

with a corona spray gun. Powder Techno135-136(10): 250-260., 2003.

Yousuf S, Barringer SA. Modeling nonelectrostatic and electrostatic powder coating.

J Food Eng 83(4): 550-561., 2007.

Xu Y, Barringer SA. Factors effect of relative humidity on coating efficiency in

nonelectrostatic coating. J Food Sci 73(6): E297-E330., 2008.

Appendix A: Standard curves

Figure A 1 Standard curve of powdered sugar solution

1.4 y = 2.3203x - 0.0273 1.2 R² = 0.9981

1.0

0.8

0.6

Absorbance 0.4

0.2

0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 Concentration (% w/w)

Figure A 2 Standard curve of soy protein solution 86

14 y = 1.0589x - 0.1602 R² = 0.9992 12 10 8 6 4

2 Reading salt concentration (%) concentration salt Reading 0 0 2 4 6 8 10 12 14 16 Actual salt concentration (%)

Figure A 3 Standard curve of extra fine salt solution