Effect Of Particle Size And Temperature On Rheological, Thermal, And Structural Properties Of Pumpkin Flour Dispersion 2014.pdf

Journal of Food Engineering 124 (2014) 43–53

Contents lists available at ScienceDirect

Journal of Food Engineering journal homepage: www.elsevier.com/locate/jfoodeng

Effect of particle size and temperature on rheological, thermal, and structural properties of pumpkin flour dispersion Jasim Ahmed a,⇑, Muhammad Al-Foudari a, Fatimah Al-Salman a, Abdulwahab S. Almusallam b a b

Food and Nutrition Program, Kuwait Institute for Scientific Research, P.O. Box 24885, Safat 13109, Kuwait Chemical Engineering Department, Kuwait University, Kuwait

a r t i c l e

i n f o

Article history: Received 24 March 2013 Received in revised form 25 September 2013 Accepted 26 September 2013 Available online 4 October 2013 Keywords: Particle size Mechanical strength Glass transition temperature Viscoelasticity Sediment volume fraction Complex viscosity

a b s t r a c t Controlling the rheological properties of dispersion has been of great interest in the food processing industry. Effects of particle size and temperature on oscillatory rheology of pumpkin flour dispersion were studied. Fresh pumpkin was freeze-dried, grinded and sieved through selected screens to obtain desired particle size fractions (74–841 lm). Most of the particles are spherical in shape. The glass transition temperature (Tg) and the melting temperature (Tm) of starch–lipid complex varied with particle size which is believed to be due to compositional variations. Rheological measurement of reconstituted particles as a function of temperature (10–90 °C) and concentration (4–10% w/w) indicated a solid-like behavior (G0 > G00 ). Sediment volume fraction (/) of isolated particle dispersions indicated a gradual decrease with decrease in particle size, which directly influences the mechanical strength and visco-elasticity of the dispersion. Particle size influenced the mechanical rigidity of pumpkin dispersion markedly whereas the temperature had the least effect. An unexpected increase in G0 of finest particle containing dispersion with temperature could be associated with gelatinization of starch and flocculation of particles with broken cell walls. Microscopic observation revealed the presence of a continuous network for the finest particle dispersion, as opposed to discontinuous one for other particle sizes. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction Pumpkins belong to the family Cucurbitaceae and the genus Cucurbita. Pumpkins and squash are available in different shapes and sizes with attractive orange colors around the world. It is a good source of carotenoids with the presence of relatively high contents of provitamin A carotenoids (principally b-carotene, acarotene, and sometimes b-cryptoxanthin) (Speek et al., 1988; González et al., 2001). Furthermore, pumpkins contain several biologically active components including polysaccharides, proteins and peptides, para-aminobenzoic acid, phenolic compounds and terpenoids and sterols (Kuhlmann et al., 1999). It is mostly considered to have active hypoglycaemic properties and it is reported that fruit pulp has anti-diabetic effects (Adams et al., 2011). Pumpkin flour powder (PFP) with or without sugar showed a significant increase in plasma insulin and reduction in blood glucose (Ju and Chang, 2001). Pumpkin mesocarp tissue has been used as a food matrix for iron supplement, and it is considered as a promising raw material for functional food product development (de Escalada Pla et al., 2009).

⇑ Corresponding author. Tel.: +965 24989789. E-mail addresses: [email protected], [email protected] (J. Ahmed). 0260-8774/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jfoodeng.2013.09.030

Pumpkin is used in various forms (e.g. puree, dry slice, powder) which are commonly used as an ingredient in pies, soups, sauces, stews, breads, instant noodle, and many other preparations as well as a natural coloring agent in pasta and flour mixes. Pumpkin shows a great diversity of texture in the cooked form, ranging from the smooth, pasty, dry, high-starch buttercup types to the stringy, watery, wet, low-starch types (Corrigan et al., 2001). It is difficult to understand whether texture attributes are inherited or attributed to constituents. Size reduction is an important unit operation where the ratio of surface area to volume of a food material is increased. Size reduction results in a mixture of particles, ranging a broad distribution starting from a larger size to a fine particle whereas sieving separates milled flours on the basis of particle size. The fullest description of a powder is given by its particle-size distribution (Snow et al., 1999). Because of the wide variation in the size and shapes of the particles and related properties in suspensions, it is really difficult to understand the contributing factors that affect the rheology. It is now accepted that the food powder properties are strongly dependent on the chemical composition and the surface properties of the particles (Cuq and Rondet, 2011). Separation of particles in uniform size range could provide uniform functional properties. Furthermore, the interaction of those known particle sizes with other ingredients could provide better understanding

44

J. Ahmed et al. / Journal of Food Engineering 124 (2014) 43–53

Nomenclature

List of abbreviations ASTM American Society for Testing and Materials ANOVA Analysis of variance DSC Differential scanning calorimetry LVR Linear visco-elastic range AOAC Official Methods of Analysis PF Pumpkin flour PFD Pumpkin flour dispersion PFP Pumpkin flour powder RSM Response surface methodology SEM Scanning electron microscope List of symbols x angular frequency, Hz P bulk density, kg/m3 g* complex viscosity, Pa s A constant G0 elastic modulus, Pa

of food structure and stability during food formulation, quality control and product development. Limited reports are available on the particle size dependency on rheological properties of food materials. Kerr et al. (2001) reported significant effects of particle size on the functionality of cowpea flour and reported textural problem with finer particles. Hayashi et al. (1976) obtained a good bread volume using fine fractions of hard red spring wheat flour, whereas coarse fractions were recommended for the cake applications. The viscoelastic behavior of suspensions has also been assessed by particle size distribution and shape as well as the volume fraction of particles (Nakajima and Harrell, 2001; Servais et al., 2002) and the particle–particle interactions (Shah et al., 2003). However, no attempt has been made to study the effect of specified particle range on the food rheology although it has tremendous effect on the food dispersion and even quality control of food suspensions especially soups and beverages. The specific particle size range and the volume fraction of swollen particles significantly influence the rheology of pumpkin flour dispersion (PFD). The objectives of this research work were to determine the effects of particle size, temperature, concentration, and their interactions (temperature–particle size) on rheological behavior of pumpkin flour (PF) particles dispersion. 2. Materials and methods 2.1. Sample preparation A single batch of mature fresh pumpkin (Cucurbita moschata) samples was purchased from the local market in the state of Kuwait during the winter season of 2012–2013. Samples were washed thoroughly, peeled and cut into small pieces with a sharp knife followed by manual separation of seeds from the pulp; finally pulps were macerated into puree and freeze dried. One set of fresh pureed sample was collected before freeze drying for rheological measurement. For the freeze drying operation, the samples were frozen in a freezer, and later transferred to the freeze-drier (GAMMA 2-16 LSC; Martin Christ GmbH, Osterode am Harz, Germany) for 38 h at a temperature between 47 °C and 50 °C, and a pressure of 0.7 Pa. Dried PF samples were grinded in a laboratory size grinder (Robot Coupe R5, France), and passed through a series of U.S. Standard sieve numbers 20, 30, 50, 100 200 and 230 mesh

n Tg HS S

frequency exponent glass transition temperature, °C height of the sediment, m particle size, lm g00 imaginary part of complex viscosity Xi and Xj independent variables b0, b1; bii; bij response surface coefficients Y measured response in RSM e random experimental error g0 real part of complex viscosity / sediment volume fraction Tm starch–lipid melting temperature, °C HT total height of the dispersed sample, m L tristimulus color value, lightness a tristimulus color value, redness b tristimulus color value, yellowness G00 viscous modulus, Pa

(Endecotts, London, UK), manually. The fractions obtained from those sieve analysis retained by the sieve were designated as 841 (20; +30), 595 (30; +50), 297 (50; +100), 149 (100; +200), and 74 (200; +230) lm. The ve sign represents pumpkin flour particles passed through the sieve and the retained particles are expressed through +ve sign. Fractionated samples were packed in amber glass bottles and stored at 5 °C till further use. 2.2. Physico-chemical properties The proximate compositions of the ground PF samples were analyzed according to AOAC methods (AOAC, 2002) for the determination of moisture, ash, and crude fat contents. Protein was calculated as nitrogen content (N)  5.3. Protein for each particle fraction was estimated by the CHNS analysis based on combustion method (AOAC, 2002). Total soluble solids and pH were measured by a refractrometer (Atago CM-780N-Plus; Bellevue, WA, U.S.A.) and pH-meter (Sension 3, Haach Co, Loveland, Columbia, USA), respectively. The loose bulk density was determined by weighing the mass of the dried powder sample which freely was poured in a 100 ml graduated cylinder and expressed as weight per unit volume (kg/m3) (ASTM D7481-09). The volume of bulk aggregate material includes the volume of the individual particles and the volume of the voids between the particles. 2.3. Determination of sediment volume fraction The volume fraction of the PFD of different particle sizes was measured using a simple centrifugation method as described by Hemar et al. (2011) for carrot cell wall particles dispersion with some modification. Simply, 1 g of flour was dispersed in 20 ml deionized water in a graduated centrifuge tube, mixed well in a vortex and kept for 6 h for hydration followed by centrifugation (Beckman GS-6R, USA) at constant centrifugation force (3000g) for 60 min. After centrifugation, the total height HT of the sample and the height of the sediment HS were measured and the effective volume fraction / occupied by the PF particles was expressed as:

/¼

HS  100 HT

ð1Þ

45

J. Ahmed et al. / Journal of Food Engineering 124 (2014) 43–53

Eq. (1) is valid only for closely packed particles without interstitial space and without damage during the packing. The volume fraction determination was performed at least in duplicate.

The enthalpy (DH) of the transition (associated with starch–lipid complex) was calculated from the area of the peak endotherm using the Universal Analysis Software (version 4.5A, TA Instruments, New Castle, DE, USA).

2.4. Tristimulus color measurement Visual color was measured using a Hunter colorimeter model ColorFlex (Hunter Associates Laboratory, Reston, VA) in terms of L (lightness), a (redness and greenness) and b (yellowness and blueness) as described earlier by Ahmed et al. (2002) for green leafy vegetables. The instrument (45°/0° geometry, 10° observer) was calibrated with a standard black and white tile followed by measurement of samples. A glass cell containing the PFD was placed above the light source and L, a and b values were recorded. Color measurements were taken in triplicates, and average values were taken for calculation. 2.5. Rheological measurement Oscillatory rheological measurements of fresh ground sample and PFD were carried out using a Discovery Hybrid Rheometer HR-3 (TA Instruments, New Castle, DE, USA). Samples were placed in a 1000lm gap between two stainless steel parallel plates (plate diameter 40 mm). The sample perimeter was covered with a thin layer of high-temperature-resistant silicone oil to prevent sample dehydration. The sample temperature was controlled by a peltier system and monitored by platinum resistance thermometer sensors (accuracy of ±0.1 °C) which are positioned at the upper and lower plates. The studied temperature range for oscillatory measurement was 10–90 °C. A sample concentration of 10% (w/w-10 g flour in 90 ml water) was used throughout the work for rheological study except for concentration effect where a range of 4–10% (w/w) was used. Small-amplitude oscillatory strain sweep experiments (0.001– 10%) were performed, and the elastic (G0 ) and viscous (G00 ) shear moduli, at a constant frequency of 0.1 Hz were monitored to determine the limit of the linear visco-elastic region (LVR). The LVR was carried out for the entire studied temperature range (data are not shown), and the measurement was carried out accordingly. Frequency sweep tests (0.01–10 Hz) were carried out in the linear regime, at constant strain (0.03) at selected temperatures. Following an initial equilibration of samples for 5 min at 10 °C, ramp heating was carried out at 5 °C/min to an endpoint of 95 °C (non-isothermal heating) at a frequency of 1 Hz. All rheological measurements were carried out in triplicate and rheological parameters were obtained directly from the manufacturer supplied computer software (TRIOS, TA Instruments, New Castle, DE, USA).

2.7. SEM and microscopic observation for average particle size and tissue structure The microstructure and particle dimension of PF particles were examined through a scanning electron microscope (SEM) (JEOL, JSM-5410LV, Tokyo, Japan). Each sample was coated with gold in a sputter coater (Structure Probe, West Chester, PA) before being scanned and photographed at 100 and 250 magnification. Particle size was measured by the software attached to the instrument which allows for detailed (average particle diameter and maximum length) measurements. About 50 particles were chosen randomly for the particle size measurement. A new generation DMRX Polarizing Microscope fitted with Image Processing Software (QWIN, Leica, Germany) was also used for particle size measurement and to study the heated PFD tissue structure. Microscopy of particles was carried out under crossed polarized Nicols using 5 and 10 objectives. 2.8. Experimental design and statistical analysis Since response surface methodology (RSM) designs have many advantages like adequate distribution of information across the experimental range (rotatability), good lack of fit detection, the fitted values are very close to the observed ones and they require the minimum number of treatment combinations (Kokkinidou and Peterson, 2013), and therefore, RSM has been selected to study the simultaneous effect of temperature and mesh size on rheological characteristics of PFD at a constant concentration of 10%. The experiments were based on a central composite rotatable design. Five levels of temperature (10, 30, 50, 70 and 90 °C) and particle size (74, 149, 297, 595 and 841 lm) were selected. The total number of experiments was 13 with five replications of the center point, as shown in Table 1. The independent variables temperature and mesh size were selected to optimize the responses g* and G0 at 1 Hz. A general second order model is mostly used in response surface methodology for predicting individual Y variables (Khuri and Mukhopadhyay, 2010). The model proposed for each response of Y is:

y ¼ b0 þ

k k X XX X bi xi þ bij xi xj þ bii x2ii þ e i¼1

i

<j

ð2Þ

i¼1

2.6. Differential scanning calorimetric (DSC) measurement A differential scanning calorimeter (DSC) (TA Q 2000, TA Instruments, New Castle, DE, USA) was employed to measure the thermal analysis for PF and dispersions. The DSC was calibrated with indium and sapphire for temperature and heat capacity calibration. The samples (10–12 mg) were run at a 10 °C/min heating/cooling ramp in heating–cooling cycles in a nitrogen atmosphere (flow rate 50 ml/ min). The samples were heated from 20 to 180 °C to detect thermal properties including, glass transition temperature (Tg), gelatinization of starch and starch–lipid complex in dry fraction and in dispersion. An empty pan was used as a reference. The DSC measurements were done in triplicate. Thermal transitions of PFD were measured for the peak gelatinization (Tm) and the melting temperature (Tm) of starch–lipid complex. Instrument software provides the onset temperature, end point temperature, and the change of heat flow of the glass transition region. The glass-transition midpoint value was calculated as the average of the onset and end point values and reported as the glass transition temperature.

where Y is the measured response associated with each factor level combination, b0 is the intercept, b1; bii; and bij are linear, quadratic and interaction coefficients of the model, Xi and Xj are the independent variables in coded values, and e is a random experimental error assumed to have a zero mean. The regression analysis, analysis of variance (ANOVA) of data was carried out using the Minitab Statistical Software (Version 16; Minitab Corp., USA) to fit second order polynomial equations for all response variables. Lack-of-fit tests were performed on the fitted models. For statistical tests, significance was defined at P 6 0.05. 3. Results and discussion 3.1. Proximate composition Freeze drying of pumpkin pulp into flour significantly reduced the moisture content to 6.7% (wet basis). The ash content and crude fiber of the flour were 5.6% and 11.7%, respectively in wet

46

J. Ahmed et al. / Journal of Food Engineering 124 (2014) 43–53

Table 1 Level of independent and response variables in experimental design. Treatment

Independent variables

1 2 3 4 5 6 7 8 9 10 11 12 13

Response variables

X1

X2

Temperature (°C)

Particle size (lm)

G0 (Pa)

0 0 1 1.41 1.41 0 0 0 1 0 1 1 0

1.41 0 1 0 0 0 0 1.41 1 0 1 1 0

50 50 70 90 10 50 50 50 30 50 30 70 50

841 297 149 297 297 297 297 74 149 297 595 595 297

47,458 16,695 10,093 15,599 19,249 16,700 16,767 96 9934 16,700 40,801 40,786 16,790

g* (Pa s) 7650 2672 1611 2500 3089 2673 2675 16 1501 2670 6568 6297 2677

basis. The lipid and protein content of the flour was 1.04% and 9.1%, respectively. These values are relatively different from reported values for PF (Aziah and Komathi, 2009; See et al., 2007). The difference is believed to be attributed to variations in geographical location, cultivar and method of sample preparation. The total soluble solids content and the pH of the pumpkin puree before freeze drying were 10.1 ± 0.2 °Brix and 6.31 ± 0.02, respectively. It is worth to mention that particle fractions did not show any significant difference in proximate compositions except nitrogen content which is discussed later on. 3.2. PF sieve analysis Particle mass (%) of ground PF after size distribution is shown in Fig. 1. It was observed that the maximum weight percent of particles (30%) were about 297 lm followed by 149 lm (29%), and 74 lm (25%). The grinding process produced minimum weight percent of coarser particles; about 1% and 15% of particles were in the sizes of 841 and 595 lm, respectively. Overall, ground PF had a broader distribution of particle sizes.

However, a sudden drop of bulk density to 562 kg/m3 was recorded for particle size of 74 lm. Similar decrease in bulk density with particle size has been reported earlier for sweet potato powder (Grabowski et al., 2006). The decrease has been explained by considering the stickiness of particles during dehydration and also from product agglomeration (Goula et al., 2004). Although particles may be small when measured individually, these agglomerates take up a larger volume and, thus, would contribute to a smaller bulk density. 3.4. Sediment volume fraction of dispersion The volume fraction (/) of suspended particles in aqueous dispersions of PF at 30 and 70 °C is presented in Fig. 2. Standard deviations are not indicated for the results where estimations were made using mean values of two runs. The volume fraction represents the particles occupancy after centrifugation and it mostly estimates the effective volume fraction of the particles. PFD containing particle sizes of 841 lm exhibited the highest / whereas the least value was obtained for powder particles of 74 lm. Generally, it is expected that the magnitude of / increases with fineness of the particle size. However, the opposite trend was observed for PF. The unusual trend could be attributed to cellular structure of PF and its water holding capacity. It is observed that more water was absorbed by porous particles compared to finer particles. This is further evidenced from the microscopic study as discussed later. The largest hydrated particle is expanded almost two times (i.e. 8 times volume increase based on spherical uniform particle) compared to 1.5 times (volume increase 3.4 times) for smaller one (Fig. 4). Secondly, the PF particles were disrupted during shearing and the cell walls separated from the endosperms especially for finer particles (evident from microscopic study-discussed later on) which is probably the cause for lower / value of 74 lm powder. Raising the temperature to 70 °C and heating isothermally for 30 min causes no significant change in the / value except for the sample of particle size 74 lm. It clearly indicates that the coarse granules already absorbed the maximum amount of water. Similar observation (constant granule sizes after heating) was made by Paterson et al. (2001) while working on corn starch from different sources.

3.3. Bulk density

3.5. PF particle size analysis by microscope and SEM

Particle size has a significant impact on bulk density. Generally, as particle size decreases the bulk density increases. Following this rule, the bulk density of PF powder increased linearly from 522 to 756 kg/m3 when the particle size decreased from 841 to 149 lm.

The particle sizes of the PF passed through sieves were observed through microscope and SEM, respectively. The average particle size, the largest and smallest particle sizes are reported in Table 2. The finest fraction seemed to have smaller particle size by almost

40

Volume fraction %

% Particle mass

30

20

10

0

80 30C

60

70C

40 20 0

841

841 595

595 297

Particle size

297 149

Particle size 74

Fig. 1. Mass distribution of freeze dried PF powder particles after sieving.

149 74

Fig. 2. Effect of particle size on sediment volume fraction of 5% (w/w) PFD.

J. Ahmed et al. / Journal of Food Engineering 124 (2014) 43–53

47

20 times lower than the coarsest fraction. SEM and microscopic images (Figs. 3 and 4) confirmed that the PF particles are irregular in shape, and most of them are ‘spherical’ shaped. A mixture of spherical, polyhedral and irregular shaped PF particles with less smooth granule surfaces have been reported (Saeleaw and Schleining, 2011). The larger particles were relatively more porous than finer particles. The hydrated particles indicated the water absorption capability of dried particles and the size of the particles increased thereafter. 3.6. Color of particle fractions The observed color values for PF with different particle sizes are shown in Table 3 and the values are comparable as reported by Kim et al. (2012) for orange color PF. Generally, the increase in the particle specific surface area by comminution caused better and more intense color values. As can be seen, decreasing the particle size caused an increase in the redness (+a value), yellowness (+b value) and lightness (L value) values of PF. The color values L and b decreased abnormally for 74 lm powder whereas +a value decreased for 149 lm sample. Mostly, the attractive orange/yellow color of PF is attributed by presence of lutein and b-carotene (Kim et al., 2012). The drop in color values for finer particles could be associated with removal of either lutein or b-carotene during size reduction or disruption of tissue structure during high shear and impact. 3.7. Thermal properties of particle fractions Pumpkin flour can exist in different solid-state forms like crystalline, amorphous or both. Many food powders and food ingredient mixes have components mainly in the amorphous state. Starch is semi-crystalline in nature and it shows both glass transition temperature (Tg) and the crystalline melting temperature (Tm). The glass transition temperature (onset, end and inflection) as influenced by particle size reduction is presented in Table 4. The coarser particles (841 lm) showed the minimum inflection Tg at about 7 °C, and the finest particles (74 lm) possessed the highest value of inflection Tg at about 13 °C although the change in Tg values were not systematic with particle size. The Tg is a very strong function of powder water content. As water is a plasticizer which enhances molecular mobility, increased water content will lead to a reduction in the glass transition temperature. However, the moisture content of the studied sample was almost alike (10.16 ± 0.12). The onset Tg represents the demarcation above which the molecules start to gain mobility or flow which will enable them to crystallize over time. The difference in observed Tg with particle size must be attributed to the constituents (starch and proteins) of the particles itself. However, it is really difficult to claim the actual contribution of each major constituent (starch or protein) to the particle here as the ground samples were not separated by air classification which is mostly used in cereal starch/protein separation. It has been reported that the protein after air classification was concentrated in fraction with fine

Table 2 Particle size analysis through SEM measurement of PF passed through selected sieves.

*

Sieve size (lm)

Avg. particle size (lm)

Largest particle size (lm)

Smallest particle size (lm)

841(20-mesh) 595 (30-mesh) 297 (50-mesh) 149 (100-mesh) 74 (200-mesh)

920 ± 38 414 ± 26 235 ± 21 122 ± 13 46 ± 06

1110 589 333 159 71

771 301 136 47 12

Values (mean ± SD, n = 15).

Fig. 3. Scanning electron microscopy of pumpkin flour particles: (a) passed through 841 lm and (b) passed through 149 lm.

particles (<15 lm) and starch was concentrated in fractions with coarser particle sizes (>15 lm) (Wu et al., 1994). Vasanthan and Bhatty (1995) studied scanning electron micrographs (SEM) of the coarse and fine barley fractions after an air classification and reported that most of the small starch granules were concentrated into finer fraction. Furthermore, coarser particles had lower protein and higher starch, and finer particles had higher protein and lower starch. Based on those above information, it could be assumed that both protein rich fraction and smaller starch granules contribute to the increase in Tg. The protein enrichment of fine particles is further supported by CHNS analysis data. The nitrogen (N) and carbon (C) contents showed a difference in the magnitude of each fraction of particles. For example, the N and C-contents of coarser particle (841 lm) fractions were 1.28 and 39.2 which changed to 1.42 and 38.1, respectively for fine particle (74 lm) fraction. DSC thermograms of PF with different particle sizes show an additional thermal transition peak in the temperature range of 144–158 °C. These melting peaks are different from starch gelatinization peak and it is believed that these endothermic peaks (Tm) are attributed to the melting of a complex between native lipids and starch/amylose (Bhatnagar and Hanna, 1994). Furthermore, it is reported in the literature that two types of complex formation (Type I and II) have been identified through DSC measurement; one at lower temperature range (96–100 °C) and another at higher temperature range (100–125 °C) (Biliaderis and Galloway, 1989; Biliaderis, 1992). The observed Tm values indicated that the

48

J. Ahmed et al. / Journal of Food Engineering 124 (2014) 43–53

157 °C to 144 °C (Table 3 and Fig. 5). The drop in Tm is noteworthy especially when the particle size reduced to 149 lm or less. The significant decrease in Tm for smaller particle fractions could be attributed to the presence of small starch granules which melted at relatively lower temperature than larger starch granules. The necessity for heat to dissociate the complex between amylose and its ligand suggests that there are certain forces that steady the helix conformation (Putseys et al., 2010). The only processing variable that had a significant effect on the formation of starch–lipid complexes was moisture. The Tm of the amylose–lipid complex decreased significantly (119–124 °C) with multiple peaks occurring when significant amount of water was added to the particles to produce a 10% dispersion. It has been reported in the literature that the semi-crystalline type IIa complexes melt around 115 °C (Biliaderis et al., 1986). In literature, there are many evidences where the presence of multiple peaks has been reported (Biliaderis and Seneviratne, 1990; Bulpin et al., 1982). This multiplicity of peaks could account for different kinds of lipids (Bulpin et al., 1982; Raphaedelis and Papavergou, 1991), which could form complexes with different structures (Raphaelides and Karkalas, 1988). 3.8. Rheology of PFD

Fig. 4. Microscopic view of PF particles: (a) particle passed through 74 lm screen and measured at 25 °C; (b) particle passed through 74 lm screen and measured after heat treatment; (c) particle passed through 841 lm screen and measured at 25 °C; (d) particle passed through 841 lm screen and measured after heat treatment; (e) heated 149 lm; and (f) heated particles network passed through 74 lm.

Table 3 Effect of particle size on tristimulus color values of pumpkin flour.

*

Particle size (lm)

L value

+a value

+b value

841 595 297 149 74

65.13 ± 2.10 65.18 ± 1.80 65. 70 ± 1.85 73.22 ± 2.31 68.43 ± 2.05

15.02 ± 0.73 15.24 ± 0.92 16.86 ± 1.10 15.88 ± 0.83 12.42 ± 0.65

28. 72 ± 1.43 31.24 ± 0.94 34.17 ± 1.10 37.43 ± 1.31 32.35 ± 1.00

Values (mean ± SD, n = 3).

Table 4 Glass transition temperature and melting of starch–lipid complex of PF as function of particle size.

*

Particle size

Tg (°C)

841 595 297 149 74

1.40 ± 0.32 6.80 ± 1.01 1.06 ± 0.23 4.04 ± 0.85 7.00 ± 0.80 12.12 ± 1.12 6.24 ± 0.72 12.09 ± 0.92 7.49 ± 0.64 12.97 ± 1.10

Onset

Inflection

Tm (°C)

DHm (J/g)

157.1 ± 1.54 150.5 ± 1.33 157.8 ± 0.96 144.0 ± 1.14 145.5 ± 1.23

137.6 ± 1.5 180.7 ± 1.1 129.9 ± 1.2 173.1 ± 0.8 133.6 ± 2.1

End 13.76 ± 1.02 9.96 ± 0.84 18.64 ± 1.22 16.24 ± 1.01 18.31 ± 1.42

Values (mean ± SD, n = 3).

complex comes under type II. Type II is more crystalline and is believed to have a lamellar-like organization of amylose complexes (Biliaderis, 1992; Biliaderis and Galloway, 1989). The Tm of the starch–lipid complex dropped with decrease in particle size from

The oscillatory measurements of all the samples were measured by performing frequency sweeps in the linear viscoelastic region. Fig. 6 illustrates representative rheological behavior of fresh pumpkin pulp (10% w/w dispersion) at 30 °C, showing elastic modulus (G0 ), viscous modulus (G0 0 ), and phase angle (°) as function of frequency. The dispersion shows a predominantly solid-like behavior, as indicated by the dominance of the G0 over the G0 0 . For example, the G0 value was almost six times higher than G00 at 1 Hz. Furthermore, G0 increases systematically within the studied frequency range (0.1–10 Hz). However, the G00 dropped at the similar condition except at higher frequency range. The G0 –x produces a smoother curve with better reproducibility than G00 , which is probably due to the more solid characteristic of the suspensions. Similar observation has been reported for tomato suspension (Bayod and Tornberg, 2011). The phase angle (°) decreased from 17 to 8.5 with the frequency indicating a gradual increase in solid-like property which levelled off later on. 3.9. Effect of concentration The mechanical rigidity of the heated PFD increased as the concentration was increased from 4% to 10% (w/w). It is worth mentioning that the mechanical strength was not detected below 4% concentration of PF since the concentration is too low for rheological measurement using parallel plate geometry. Therefore, 4% concentration can be considered as the minimum concentration to obtain a meaningful rheological structure/data. The G0 increased abruptly (almost one log cycle) when the concentration was increased further from 4% to 6%; above 6%, the value of G0 almost doubled. Such increment of the G0 of PFD as function of concentration can be explained by the sediment volume fraction (/) of dispersions which is found to be a direct function of concentration (data not shown). As the volume fraction increases gradually, the G0 gradually increases faster than the loss modulus which has been attributed to significant compression and interpenetration of the stabilizing polymeric layer. Similar observation has been reported by Luckham and Ukeje (1999) for polystyrene latex dispersion. The rigidity of PFD at lower concentration is controlled by the volume occupied by the swollen granules (termed as ‘dilute’ regime), and, to a lesser extent, by the soluble fraction. While concentration was increased significantly, powder granules cannot swell to their equilibrium volume due to limited availability of water, and then, the rheological characteristics contributed primarily by the particle

49

J. Ahmed et al. / Journal of Food Engineering 124 (2014) 43–53

––––––– 74 micrometer – – – – 841 micrometer ––– ––– 149 micrometer

-0.4

-0.6

-0.8

-1.0 100

143.51°C

156.18°C

144.27°C

120

140

160

180

Temperature (°C) Fig. 5. Effect of particle size on starch–lipid complex melting behavior.

20

100000

G''

15

δ

δ (o)

G', G'' (Pa)

G'

10000 10

1000 0.1

1

5 10

Frequency (Hz) Fig. 6. Typical rheograms for 10% PFD containing particles size of 841 lm at 30 °C.

rigidity of the swollen granules. Another explanation for increasing mechanical rigidity with concentration is based on the fact that particles replaced water and increasing amount of particles produced a low water activity environment which favored hydrophobic interactions between pectin chains, and therefore, increased the strength of the network (Genovese et al., 2010). The gel rigidity (G0 ) of 149 lm PFD as a function of concentration (C) at 70 °C can be well described by a quadratic polynomial relationship (Eq. (3)) with a coefficient of determination (R2) of 0.99. The equation is valid only at and above 4% (w/w) of PF concentration. Earlier a linear relationship between C and G0 has been reported for wheat and maize starch at concentration range of 6– 30% (Ring, 1985).

G0 ¼ 15:885C 2 þ 11:978C  239:38 where C  4%

g ¼ 0:0064S2 þ 15:638S  941:77 ðR2 ¼ 0:99Þ

ð4Þ

3.11. Effect of temperature Non-isothermal heating of 10% PF with 74 lm particle size at a frequency of 1 Hz is presented in Fig. 10. As can be seen from Fig. 9, 1000000

1000000

G' fresh nonuniform PS G' PF 841 micrometer CV PF nonuniform PS

G' PF nonuniform PS CV fresh nonuniform PS CV PF 841 micrometer 100000

ð3Þ

3.10. Effect of sieving and particle size Sieving of PF produces different particle sizes. Particle size distribution plays an important role in dispersion rheology and food formulation. A comparative rheological study between the reconstituted samples (10% concentration dispersion with uniform and non-uniform particle size) and the whole pulp (10 oBrix) with non-uniform particle size is shown in Fig. 7. It was observed that the fresh pulp possessed higher mechanical strength when compared to the systems reconstituted with particles with and without sieving. The higher G0 values of fresh sample have been attributed to a higher value of / than that of reconstituted sample and also to the retention of the original structure (anisometric cells and aggre-

100000

10000

η* (Pa.s)

Heat Flow (W/g)

-0.2

gates) without any processing effect. Comparing the G0 , the PF of uniform particle size distribution (841 lm) showed a relatively higher G0 than that of sample with non-uniform particles. Considering only hydrodynamic interactions, addition of equal amounts of water to the individual particle volumes produces different suspension viscosities. More water is required to fill the space between the particles in the uniform size distribution (occupies more space than widely distributed particles). Thus, the suspension of uniform size particles has less liquid between the particles compared to the suspension consisting of packed particles, and therefore, the suspension of uniform particles has higher viscosity. Further, it has been argued that increasing the polydispersity for a given solids volume fraction would decrease a suspension viscosity, which is also supported by the ‘Farris effect’ (Ferguson and Kemblowski, 1991). Such study infers that not only the particles mean diameter, but also the particle distribution is an important factor influencing the rheology of suspensions (Sato and Cunha, 2009). The effect of particle size on the visco-elastic behavior of 10% PFD at 70 °C is presented in Fig. 8. It is clearly evident that the complex viscosity (g*) decreased gradually as particle size reduced from 841 to 74 lm. Similar observation has been reported for jaboticaba pulp where increasing particle size from 64 to 550 mm promoted a rise in the viscosity from 0.06 to 0.40 Pa s (Sato and Cunha, 2009). Generally, it is expected that finer particle size produces more surface area and contributes to the higher mechanical strength. However, our results showed the opposite trend. It is assumed that the ground PF was less porous and would be unable to imbibe as much water as the coarse flour which is duly supported by decreasing trend of / with fineness of particles as discussed earlier. A blend of two distinct particle sizes (100 and 200 mesh) produces intermediate g* value which is higher than that of 200 mesh but lower than 100 mesh. This increase might be attributed to different interactions between pumpkin constituents with different particle size distribution. The complex viscosity (g of 10% PFD can be well described by a quadratic polynomial equation as a function of particle size (S) in the prescribed size limit (74– 841 lm) at 70 °C.

G' (Pa)

0.0

1000

10000 0.1

1

100 10

Frequency (Hz) Fig. 7. A comparison of rheological behavior between uniform and non-uniform particle size in 10% fresh and powder PFD (CV stands for complex viscosity and PS stands for particle size).

50

J. Ahmed et al. / Journal of Food Engineering 124 (2014) 43–53 100000

100

1000 Elastic modulus Complex viscosity

100

10

841 micrometer

595 micrometer

297 micrometer

140 micrometer

74 micrometer 1

0.1

10

100

η∗ (Pa.s)

1000

G' (Pa)

η* (Pa.s)

10000

Blend of 149 & 74 micrometer

1

10

Frequency (Hz) Fig. 8. Effect of particle size on complex viscosity of 10% PFD at 70 °C.

G0 ¼ Axn After linearization; ln G0 ¼ ln A þ n ln x

ð5Þ ð5aÞ

10

0

20

40

60

80

1 100

o

Temperature ( C) Fig. 9. Non-isothermal heating of 10% PFD passed through 74 lm sieve at a heating rate of 5 °C/min.

100000

100000 G' 30C G' 70C CV 30C CV 70C

G' 50C G' 90C CV 50C CV 90C

η* (Pa.s)

G' (Pa)

10000

10000

1000

1000 0.1

100 10

1

Frequency (Hz)

(a) 1000

1000

η* (Pa)

100

G' (Pa)

as heating proceeds, there is a sharp increase in G0 and g* associated to starch gelatinization especially above the starch gelatinization temperature of 73 °C (as detected from DSC measurement). However the developed gel was not a strong one. The gel strength is not improved even after non-isothermal heating followed by cooling to 25 °C (setback region) which is contrary to most of starch dispersions (data not shown). This could be the nature of pumpkin starch under heating–cooling condition and also the material inability to form a viscous paste or gel after cooking and cooling. The observed gelatinization temperature is very similar to the pasting temperature of unripe pumpkin (72 °C) however the value is relatively lower than that of ripe pumpkin (80 °C) (Roura et al., 2007). Further authors advocated that the differences with samples would indicate changes in samples chemical composition due to maturity that could produce different starch interactions with endogenous materials (sugars, lipids, pectin and so on) that would affect their behavior. Isothermal heat treatment of 10% for 149 and 74 lm particles containing PFD are illustrated in Fig. 10. A contradictory result was observed for those samples. It is evident from Fig. 10a that the temperature has insignificant effect on both G0 –x and g*–x curves of PFD. An increase in suspension rigidity was observed at 70 °C which was basically caused by swelling of starch granules and gelatinization. PFD made from 74 lm particles exhibited an unexpected behavior during isothermal heating (Fig. 10b). In fact, the mechanical rigidity of dispersion having 74 lm particles was about 200 times weaker than the suspension containing particle size of 149 lm at 30 °C. As it is mentioned in the earlier section, small starch granules are concentrated in fine particles fraction which is believed to lower the gel rigidity of dispersion. Furthermore, breakdown of starch and disruption of cellular structure during comminution could be possible other reasons for lowering the gel rigidity of fine powdered dispersion. Both G0 and g* increased as function of heating temperature; the increase was substantial at and above 70 °C which is mostly attributed by gelatinization of starch in the particle. Microscopic studies supported the hypothesis cited above; it was observed through microscope that most of the 74 lm particles were ruptured and cell walls were broken. Such hypothesis was also supported by the development of a continuous gel network (Fig. 4f) whereas other particles produced only discontinuous gel at similar condition (Fig. 4e). The frequency (x) dependence of gel rigidity (G0 ) can be described by the power-type relationship:

100

10

G' 30C G' 90C CV 70C

10 0.1

G' 50C CV 30C CV 90C

G' 70C CV 50C

1

1 10

Frequency (Hz)

(b) Fig. 10. Effect of temperature on oscillatory rheology of PF sample passed through (a) 149 lm and (b) 74 lm sieve (CV stands for complex viscosity).

51

J. Ahmed et al. / Journal of Food Engineering 124 (2014) 43–53

where A is a constant and n is the frequency exponent (slope) whose value lies between 0 and 1. The slope and intercept are calculated from the linear regression of ln x vs ln G0 . Table 5 provides magnitudes of slopes, intercepts and the goodness of fitting (R2) for 10% PFDs having different particle sizes at selected temperature. Experimental data were fitted well with R2 values being always higher than 0.95. The intercept, A, of a definite particle size did not show any trend as function of temperature. It is evident from the table that temperature had a minimal effect on the solid-like properties as assessed by slope, n, of PFD except for the dispersion having particle size of 74 lm where temperature played a significant role by changing the mechanical rigidity. The liquid-like property of 74 lm particles dispersion remained predominant up to 70 °C followed by a drop at 90 °C. The intercept (A) trends with particle size follow the aforementioned behavior for G0 and g*; they decrease from 841 lm to 74 lm. A blend (1–1 ratio) of two uniform particles (74 and 149 lm) exhibited predominant solid-like property (slope 0.05– 0.09) which was significantly different from fine particle (74 lm) dispersion.

50000

η'' (Pa.s)

40000

30000

20000

10000

0

0

841 micrometer

595 micrometer

297 micrometer

149 micrometer

5000

10000

15000

20000

η' (Pa.s)

(a) 800

3.12. Effect of particle size and temperature 74 micrometer

600

g ¼ 14S  558

ð6Þ

G0 ¼ 85S  1924

ð7Þ

η'' (Pa.s)

Although it is well established that the temperature greatly influences the rheological properties of food products, limited studies have been carried out on particle size effect on food rheology. When a system is filled with anisotropic particles, the type and degree of orientation of the particles can have a significant effect on the rheological behavior (Metzner, 1985). Yoo and Rao (1994) reported that there is a significant effect of particle size on the flow properties of tomato puree. Both magnitudes of consistency index and apparent viscosity of samples containing finer particles (60 mesh sieve) were higher than those of 40 mesh sieve sample. From a practical standpoint, it will be more informative to describe the effects of temperature and particle size on rheological characteristics by a combined model. In this study, g* and G0 at 1 Hz were related to temperature and particle size of 10% PFD by a multiple regression analysis. The effects of treatment variables as linear, quadratic, or interaction coefficients on response variables were obtained by ANOVA. The predictive models developed for g* and G0 of 10% PFD were considered adequate they possessed no significant lack-of-fit and had a satisfactory level of R2. The linear terms were significant for both g* and G0 (P < 0.05), whereas square and interaction terms were insignificant. The R2 values were greater than 0.96 for all the cases. The fitted equations were obtained considering only significant terms to predict g* and G0 as shown below:

400

200

0

0

75

150

225

300

η' (Pa.s)

(b) Fig. 11. Dispersion of selected pumpkin flour particles in aqueous solution at 90 °C.

The coefficient of linear term of particle size indicated a gradual increase of rheological parameters as influenced by particle size.

3.13. Pumpkin flour particle dispersion in solution Adequacy of PF particles dispersion in aqueous suspensions can be analyzed by the Cole–Cole plot of the rheological data which represents the relationship between real and imaginary parts of complex viscosity (Joshi et al., 2006; Ahmed et al., 2010a,b). A smooth, semicircular shape of the plotted curves represents good compatibility, that is, phase homogeneity in the solution, and any deviation from the shape indicates nonhomogenous dispersion or immiscibility. The g0 , the real part and g00 , the imaginary part of

where S is the PF particle size in lm. The above equations indicated that particle size has significant effect on both complex viscosity and elastic modulus, respectively.

Table 5 Slope of Eq. (5) for dispersion made by different PF particle size as function of temperature (R2 is reported after n value in the bracket). Particle size (lm)

841 595 297 149 74 Blend 74 & 149

10 °C

30 °C n

50 °C n

70 °C n

90 °C n

n

A (Pa s )

n

A (Pa s )

n

A (Pa s )

n

A (Pa s )

n

A (Pa sn)

0.12(0.98) 0.10(0.99) 0.08(0.99) 0.06(0.99) 0.17(0.99) 0.06(0.99)

52,029 45,073 19,136 10,297 56 978

0.11(0.98) 0.10(0.99) 0.07(0.98) 0.06(0.98) 0.24(0.95) 0.07(0.96)

45,960 38,482 17,073 9976 282 1126

0.12(0.96) 0.11(0.97) 0.09(0.97) 0.07(0.98) 0.30(0.95) 0.08(0.97)

46,005 37,971 16,374 9238 90 931

0.11(0.98) 0.10(0.99) 0.07(0.99) 0.06(0.99) 0.24(0.99) 0.09(0.99)

45,961 38,481 17,070 9986 285 1120

0.10(0.97) 0.09(0.99) 0.06(0.99) 0.04(0.98) 0.16(0.99) 0.05(0.96)

39,681 33,776 15,547 8988 670 1235

52

J. Ahmed et al. / Journal of Food Engineering 124 (2014) 43–53

Table 6 Regression equations describing PF properties as function of particle size (S). Property

2

q ¼ 0:0004S þ 0:157S þ 635

(10)

Volume fractiona

/ ¼ 0:0001S2 þ 0:160S þ 11

(11)

0.93

Complex viscositya

g ¼ 0:0002S2 þ 18:1S  2185

(12)

0.94

Glass transition

T g ¼ 0:00002S2  0:025S þ 16

(13)

0.78

Starch–lipid melting point

T m ¼ 0:00002S2 þ 0:035S þ 143

(14)

0.50

Bulk density

a

R2

Equation a

0.44

At 30 °C.

complex viscosity are calculated according to the following equations:

g0 ¼ g00 ¼

G00

x G0

x

ð8Þ

ð9Þ

Fig. 11 illustrates a Cole–Cole plot for PF particles in aqueous dispersion at 90 °C. It demonstrates a very smooth semicircular shape of the dispersion for all particle sizes except for particles with 74 lm. The curve deviates for dispersion containing 74 lm particles which indicates immiscibility. It is worth mentioning that the Cole– Cole plot is not an absolute technique to determine the miscibility of blends. Other supporting rheological data are required to strengthen the hypothesis. 4. Conclusion The functional properties of food powders are largely dependent of the surface composition, size and surface characteristics of the particles. Oscillatory rheology of PFD is strongly influenced by particle size of the powders and with the sediment volume fraction. The gel rigidity of the dispersion increased significantly above the peak gelatinization temperature. The glass transition temperature and starch–lipid complex melting temperature of PF were duly affected by particle size. The mechanical, thermal and optical properties of PF showed reverse trends when the particle size was reduced to 74 lm. Overall, a strong particle size dependency of PF properties are described by a second-order polynomial equations (Table 6). Microscopic study exhibited the disruption of cell walls above the critical particle size limit. Since the composition of the PF in the particles was affected by size reduction, and, therefore further investigations are needed to establish particle size and individual component (starch/protein/lipid) on structural property of PF gel. These rheological and thermal properties of PF provide important information for selecting particle size in food product development with desired texture. Acknowledgements The authors express their gratitude to Kuwait University for providing access to scanning electron microscopy equipment (General Facility grant number GE01/07). The authors are thankful to Dr. Vinod Kumar, Mr. Shaji Michael and Mr. Ejaj A. Bhatti, KISR for their help in microscopy, SEM and CNS measurements. References Adams, G.G., Imran, S., Wang, S., Mohammad, A., Kok, S., Gray, D.A., Channell, G.A., Morris, G.A., Harding, S.E., 2011. The hypoglycaemic effect of pumpkins as antidiabetic and functional medicines. Food Research International 44, 862–867. Ahmed, J., Kaur, A., Shivhare, U.S., 2002. Color degradation kinetics of spinach, mustard leaves and mixed puree. Journal of Food Science 67, 1088–1091. Ahmed, J., Varhney, S.K., Auras, R., 2010a. Rheological and thermal properties of polylactide/silicate nanocomposites films. Journal of Food Science 75, N17–N24.

Ahmed, J., Varhney, S.K., Auras, R., Hwang, S.W., 2010b. Thermal and rheological properties of L-polylactide/polyethylene glycol/silicate nanocomposites films. Journal of Food Science 75, N97–N108. AOAC, 2002. Official Methods of Analyses of Association of Analytical Chemist. 17th ed., Washington, DC. ASTM, 2009. American Society for Testing and Materials. Standard Test Methods for Determining Loose and Tapped Bulk Densities of Powders using a Graduated Cylinder. D7481-09, ASTM International, PA, USA. Aziah, N., Komathi, C.A., 2009. Physico chemical and functional properties of peeled and unpeeled PF. Journal of Food Science 74, 328–333. Bayod, E., Tornberg, E., 2011. Microstructure of highly concentrated tomato suspensions on homogenisation and subsequent shearing. Food Research International 44, 755–764. Bhatnagar, S., Hanna, M.A., 1994. Extrusion processing conditions for amylose lipid complexing. Cereal Chemistry 71, 587–593. Biliaderis, C.G., 1992. Characterization of starch networks by small strain dynamic rheometry. In: Alexander, R.J., Zobel, H.F. (Eds.), Developments in Carbohydrate Chemistry. American Association of Cereal Chemists, pp. 87–135. Biliaderis, C.G., Galloway, G., 1989. Crystallization behavior of amylose-V complexes: Structure–property relationships. Carbohydrate Research 189, 31–48. Biliaderis, C.G., Seneviratne, H.D., 1990. On the supermolecular structure and metastability of glycerol monostearate–amylase complex. Carbohydrate Polymers 13, 185–192. Biliaderis, C.G., Page, C.M., Maurice, T.J., 1986. On the multiple melting transitions of starch monoglyceride systems. Food Chemistry 22, 279–295. Bulpin, P.V., Welsh, E.J., Morris, E.R., 1982. Physical characterization of amylose– fatty acid complexes in starch granules and in solution. Starch 34, 335–341. Corrigan, V.K., Hurst, P.L., Potter, J.F., 2001. Winter squash (Cucurbita maxima) texture: sensory, chemical, and physical measures. New Zealand Journal of Crop and Horticultural Science 29, 111–124. Cuq, B., Rondet, E.AbecassisJ., 2011. Food powders engineering, between knowhow and science: Constraints, stakes and opportunities. Powder Technology 208, 244–251. de Escalada Pla, M., Campos, C., Gerschenson, L., Rojas, A., 2009. Pumpkin (Cucurbita moschata, Duchesne ex poiret) mesocarp tissue as a food matrix for supporting iron in a food product. Journal of Food Engineering 92 (4), 361–369. Ferguson, J., Kemblowski, Z., 1991. Applied Fluid Rheology. Elsevier. Genovese, D.B., Ye, A., Singh, H., 2010. High methoxyl pectin/apple particles composite gels: effect of particle size and particle concentration on mechanical properties and gel structure. Journal of Texture Studies 41, 171–189. González, E., Montenegro, M.A., Nazareno, M.A., 2001. Carotenoid composition and vitamin A value of an Argentinian squash (Cucurbita moschata). Archivos Latinoamericanos de Nutrición 51, 395–399. Goula, A.M., Adamopoulos, K.G., Kazakis, N.A., 2004. Influence of spray drying conditions on tomato powder properties. Drying Technology 22, 1129– 1151. Grabowski, J.A., Truong, V.-D., Daubert, C.R., 2006. Spray-drying of amylase hydrolyzed sweetpotato puree and physicochemical properties of powder. Journal of Food Science 71, E209–217. Hayashi, M., D’Appolonia, B.L., Shuey, W.C., 1976. Baking studies on the pin-milled and air-classified flour from four hard red spring wheat varieties. Cereal Chemistry 53, 525–540. Hemar, Y., Lebreton, S., Xu, M., Day, L., 2011. Small-deformation rheology investigation of rehydrated cell wall particles–xanthan mixtures. Food Hydrocolloids 25, 668–676. Joshi, M., Butola, B.S., Simon, G., Kukaleva, N., 2006. Rheological and viscoelastic behavior of HDPE/Octamethyl-POSS nanocomposite. Macromolecules 39, 1839–1849. Ju, L., Chang, D., 2001. Hypoglycemic effect of pumpkin powder. Journal of Harbin Medicine 21, 5–6. Kerr, W.L., Ward, C.D.W., McWatters, K.H., Resurreccion, A.V.A., 2001. Milling and particle size of cowpea flour and snack chip quality. Food Research International 34, 39–45. Khuri, A.I., Mukhopadhyay, S., 2010. Response surface methodology. WIREs Computational Statistics 2, 128–149. Kim, M.J., Shim, C.K., Park, J.H., Hong, S.J., Lee, M.H., Han, E.J., Kim, Y.K., Jee, H.J., 2012. Introduction to Colorimetric Analysis for Assessment of Carotenoid Pigments in Squash Germplasm. Korean Journal of Breeding Science 44, 433– 443. Kokkinidou, S., Peterson, D.G., 2013. Response surface methodology as optimization strategy for reduction of reactive carbonyl species in foods by means of phenolic chemistry. Food and Function 4, 1093–1104. Kuhlmann, H.K., Koetter, U., Theurer, C., 1999. Sterol contents in medicinal pumpkin (Cucurbita pepo convar. citrullinina var. styriaca). Acta Horticulturae 492, 175–178. Luckham, P.F., Ukeje, M.A., 1999. Effect of particle size distribution on the rheology of dispersed systems. Journal of colloid and interface science 220, 347–356. Metzner, A.B., 1985. Rheology of suspensions in polymeric liquids. Journal of Rheology 29, 739–776. Nakajima, N., Harrell, E.R., 2001. Rheology of PVC plastisol: particle size distribution and viscoelastic properties. Journal of Colloid Interface Science 1, 105–115. Paterson, J.L., Hardacre, A., Li, P., Rao, M.A., 2001. Rheology and granule size distributions of corn starch dispersions from two genotypes and grown in four regions. Food Hydrocolloids 15, 453–459.

J. Ahmed et al. / Journal of Food Engineering 124 (2014) 43–53 Putseys, J.A., Lamberts, L., Delcour, J.A., 2010. Amylose-inclusion complexes: formation, identity and physico-chemical properties. Journal of Cereal Science 51, 238–247. Raphaedelis, S.N., Papavergou, E.J., 1991. An investigation of the quantitative relationship of amylose–monoglyceride interactions. LWT-Food Science and Technology 24, 391–399. Raphaelides, S., Karkalas, J., 1988. Thermal dissociation of amylose–fatty acid complexes. Carbohydrate Research 172, 65–72. Ring, S.G., 1985. Some studies on starch gelation. Starke 3, 80–83. Roura, S.I., Del Valle, C.E., Aguero, L., Davidovich, L.A., 2007. Changes in apparent viscosity and vitamin C retention during thermal treatment of butternut squash (Cucurbita moschata Duch) pulp: effect of ripening stage. Journal of Food Quality 30, 538–551. Saeleaw, M., Schleining, G., 2011. Effect of frying parameters on crispiness and sound emission of cassava crackers. Journal of Food Engineering 103, 229–236. Sato, A.C.K., Cunha, R.L., 2009. Effect of particle size on rheological properties of jaboticaba pulp. Journal of Food Engineering 91, 566–570. See, E.F., Wan Nadiah, W.A., Noor Aziah, A.A., 2007. Physico-chemical and sensory evaluation of breads supplemented with pumpkin flour. ASEAN Food Journal 14, 123–130.

53

Servais, C., Jones, R., Roberts, I., 2002. The influence of particle size distribution on processing of food. Journal of Food Engineering 51, 201–208. Shah, S.A., Chen, Y.-L., Zukoski, C.F., Schweizer, K.S., 2003. Phase behavior and concentration fluctuations in suspensions of hard spheres and near ideal polymers. Journal of Chemical Physics 118, 3350. Snow, R.H., Allen, T., Ennis, B.G., Litster, J.D., 1999. Size reduction and size enlargement. In: Perry, R.H., Green, D.W. (Eds.), Perry’s Chemical Engineering Handbook. McGraw-Hill, New York, pp. 20–89. Speek, A.J., Speek-Saichua, S., Schreurs, W.H.P., 1988. Total carotenoid and bcarotene contents of Thai vegetables and the effect of processing. Food Chemistry 27, 245–257. Vasanthan, T., Bhatty, R.S., 1995. Starch purification after pin milling and air classification of waxy, normal, and high amylase barleys. Cereal Chemistry 72, 379–384. Wu, Y.V., Stringfellow, A.C., Inglett, G.E., 1994. Protein and b-glucan enriched fractions from high-protein, high b-glucan barleys by sieving and air classification. Cereal Chemistry 71, 220–223. Yoo, B., Rao, M.A., 1994. Effect of unimodal particle size and pulp content on rheological properties of tomato puree. Journal of Texture Studies 25, 421– 436.