2019 年 25 巻 3 号 p. 351-362
Palm date seeds can be a potential source to prepare a decaffeinated coffee-like brew and to produce oil. Hence, the objectives of this work are to model the changes in colour and hardness of date seeds during roasting process, as well as to study the effects of roasting conditions on the total specific grinding energy and the oil extraction yield of the seeds. General reaction models and general regression equation were applied to simulate the changes in colour, hardness, oil extraction yield, and total specific grinding energy during roasting process. The outcomes showed that the colour parameters (L*-value and b*-value) and the hardness of the roasted date seeds can be satisfactorily described by the first-order equation, while a*-value and ΔE were adequately defined by the zero-order model. A decrease in total specific grinding energy and an increase in oil yield were recorded as the roasting temperature and time were increased. Hence, this study concludes that the roasting of date seeds at 200 °C for 30 min generates the lowest grinding energy and a high amount of oil.
Date seeds represent a major waste of human consumption or from processing of palm date fruit. Due to the various palm date processes (date juice, date honey, date jam, date vinegar, and date pastes), approximately 846,000 tons of date seeds are produced annually (considering 10% of the total fruit weight) (FAO, 2016; Barreveld, 1993). This constitutes to major environmental issue and requires additional handling cost. Indeed, palm date seeds are considered as a source of minerals, dietary fibres, phenolic compounds, and antioxidants that can be used as nutritional and therapeutic components (Bouhlali et al., 2017; Al-Farsi and Lee, 2008). Bouhlali et al. (2017) determined the chemical composition of palm date seeds and found that these seeds contained eight minerals, including potassium (4153.3–22967.1 mg/kg), magnesium (827.6–615.3mg/kg), calcium (626.71–395.0 mg/kg), sodium (319.4 mg/kg–108.1 mg/kg), iron (70.3 mg/kg–27.7 mg/kg), zinc (8.8–14.7 mg/kg), manganese (5.5–11.0 mg/kg), and copper (4.8–8.3 mg/kg). Additionally, these seeds composed of 7.1–10.3% moisture, 5.0–6.3% protein, 9.9–13.5% fat, 46–51% acid detergent fibre, 65–69% neutral detergent fibre, and 1.0–1.8% ash (Hamada et al., 2002). It was reported that date seeds contained a considerable percentage of dietary fibre that varied between 70 and 83 g/100 g (Fikry, 2016; Al-Farsi et al., 2007). Furthermore, palm date seeds have high contents of total phenolic (3102–4430 mg of gallic acid equivalents/100 g) and antioxidant activity (580–929 µmol of Trolox equivalents/g) (Al-Farsi et al., 2007; Bouhlali et al., 2017). Apart from being used to treat some diseases, date seeds are commonly discarded or used as animal fodder. Date seeds also serve as an ingredient in food materials, such as bakery products (Platat et al., 2015), chocolate (Bouaziz et al., 2017), and drinks (Venkatachalam and Sengottian, 2016). Recently, date seed oil has been used for mayonnaise production (Basuny and Al-Marzooq, 2011). Apart from the food industry, date seed oil is used for cosmetic and pharmaceutical purposes (Devshony et al., 1992). In the Middle East countries, roasted palm date seeds have been used for production of non-caffeinated coffee (Rahman et al., 2007). Roasting is one of the most chief operations that gives food material the necessary changes to transform into value-added products. Indeed, some desirable or undesirable alterations in physical, chemical, and nutritional properties of the seeds can be resulted from the roasting process (Nizamlioglu and Nas, 2016). Conventional roasting is a common thermal process used by various food industries, including coffee, cocoa beans, nuts, and almonds (Nizamlioglu and Nas, 2016). Roasting process possesses a substantial number of merits for food manufacturing, such as promote the aroma and flavour of the beverages and improving the efficiency of post-operations. Roasting deactivates enzymes that can accelerate nutrient loss, terminates undesirable microorganisms and food contaminants, and extends the shelf-life of the product (Chung et al., 2013). Temperature and time dictate the degree of roasting, especially the interaction between the temperature and time should be sufficient for the required chemical changes to occur without burning the product or compromising the quality of the product (Mendes et al., 2001). Palm date seeds were roasted under varied conditions, such as temperature levels (125–220 °C) and roasting time of 30 min, in order to produce coffee-like powder and brew (El Sheikh et al., 2014; Venkatachalam and Sengottian, 2016). Quality and chemical properties of roasted foods are affected by roasting conditions (Bolek and Ozdemir, 2017b). Colour is considered the most important quality parameter of the foods and based on which the roasting process can be controlled (Driscoll and Madamba, 1994). Colour of the products is influenced by the roasting processes due to increment in brown pigments as the browning reaction progresses (Moss and Otten, 1989). Apart from colour attributes, hardness is another important quality attribute for roasting of palm date seeds. Roasting process makes seeds more crumble and brittle, which characterise the roasted products. Reaching a specific grade of hardness is essential to grind the seeds and produce powders (Kahyaoglu and Kaya, 2006). Grinding is an integral operation that has been widely used for size reduction across the food industry (Indira and Bhattacharya, 2006). For instance, grinding process consumes most of the total power while producing the powder form (Dziki, 2008). There are a variety of grinding tools, such as crushers, grinders, mills, disintegrators, cutters, shredders, mincers, and homogenisers (Kamdem and Hardy, 1995). The grinding energy consumption is an essential requirement to design a suitable grinder so as to obtain a certain particle size of palm date seed powder. Several empirical methods that measure energy have been reported and widely used for Asian-originated peanuts (Rozalli et al., 2015), soybean (Lee et al., 2013), wheat grains (Dziki, 2011), and wheat kernels (Dziki, 2008). Various studies have reported the effects of roasting conditions on the quality parameters of several products, such as P. terebinthus beans (Bolek and Ozdemir, 2017b), hazelnut (Marzocchi et al., 2017), peanuts (Shi et al., 2017; Devshony et al., 1992), Polish hazelnuts (Ciemniewska-Żytkiewicz et al., 2014), Arabica coffee (Wang and Lim, 2014), pistachio kernels (Shakerardekani et al., 2011), almond (Yang et al., 2010), coffee (Hernández et al., 2008), and sesame seeds (Kahyaoglu and Kaya, 2006). Nonetheless, to the best of the authors' knowledge, no information is available regarding the effects of roasting conditions on physical properties (colour and hardness), total specific grinding energy, and extracted oil yield from palm date seeds. The study objectives are to model the changes in colour and hardness of palm date seeds during roasting process and to assess the effects of roasting conditions on total specific grinding energy and oil extraction yield.
Raw materials Palm date fruits (Sukkari cultivar) were bought from a well-known Saudi market located at Riyadh, Saudi Arabia. The deseeding process was conducted manually to separate seeds from flesh. Next, the seeds were placed in hot water for an hour to remove flesh residue. After that, the seeds were arranged in thin layer in an environmentally controlled room at 25± 2 °C and 50 ±1% RH for 24 hr to remove excess water from the surface until no free surface water was observed, either visually or by touch. Five kg of palm date seeds were packed in sealed polyethylene bags (1 kg/bag) and stored in refrigerator at 5 °C until further use.
Characterization of raw palm date seeds Twenty seeds were randomly selected to determine their characteristics, including mass, diameter density, length, volume, surface area, sphericity, colour parameters, and hardness. A sensitive balance (Mettler-Toledo PG203-S, Toledo Comp, Switzerland) and a digital calliper (Mitutoyo, Absolute Digimatic, Japan) were used to measure mass and dimensions, respectively.
The mathematical formulae (Eqs. 1–7) used by Gastón et al. (2002) were applied to determine the characteristics of palm date seeds.
 |
 |
 |
 |
 |
 |
 |
Where l1 is length (the longest) (mm), l2 is width (the maximum diameter) (mm), l3 is thickness (the minimum diameter) (mm), lm refers to the mean diameter (mm), Vg is pycnometric seeds volume (mm3), Vgg is geometric volume (mm3) calculated from the measured axis by assuming an ellipsoidal shape, Agg the corresponding geometric surface area of this ellipsoid (mm2), U is eccentricity of the ellipsoid (unitless), Ag is the corrected surface area (mm2) obtained upon the differences of (Vg) and (Vgg), De is the equivalent spherical diameter (mm), corresponding to a sphere of the same volume as the seed, and fe is the sphericity factor, defined as the ratio of surface area of the equivalent sphere to the corrected surface area.
Roasting process Based on the method described by (Fikry et al., 2019), the roasting process was performed by using a natural convection oven (Memmert, UN, Germany) at three roasting temperatures (160, 180, and 200 °C) for 30 min to simulate the commercial roasting conditions of food products. The maximum roasting period of 30 min was chosen based on a preliminary study to avoid burning of the seeds and its effect on the seeds powder quality. Prior to the roasting process, the oven was operated for 1 hour to confirm that a steady-state condition was achieved. One hundred grams of palm date seeds were placed on aluminium foil (300 mm length × 250 mm width × 2 mm thickness) as a thin layer and placed on stainless steel oven trays (75 cm length × 50 cm width × 1 cm depth). Next, twenty samples were taken from the oven at every 2 min for a period of 30 min for colour determination. Among the selected samples, three and ten samples were randomly chosen to measure the moisture content and the hardness of the seeds, respectively. The samples were collected in less than 10 s to retain the steady-state condition during sampling. Fig. 1 illustrates the various samples of palm date seeds before and after the roasting process.
Photo of different roasted samples of palm date seeds at three temperatures (160, 180 and 200 °C).
Calculation of the thermal diffusivity (∝e) Internal temperatures of the seeds were measured by attaching a K-type thermocouple of 0.15 mm thickness to the data logger (Hoboware, USA) with ±0.5 °C accuracy. A seed was drilled to insert a thermocouple 4-mm below the surface (internal temperature). Temperature changes at the sample centre (Tc) was monitored every 1 min.
Thermal diffusivity was determined by using the general solution of Fourier's law of heat conduction (McCabe et al., 1993). Based on the first term of the general solution, the plot slope of ln (TR) versus (t) was applied to determine thermal diffusivity (∝e) from the linear form of Eq. 8.
 |
where TR, To, and Ts are temperature ratio, initial product temperature, and surrounding temperature (hot air), respectively.
Grinding process and its energy requirements The samples of palm date seeds (250 g) at various roasting temperatures and times were preliminary crushed using a cutting laboratory grinder (RT-CR30S, 2007, Taiwan) with 2 mm sieve, 3 HP motor (3-phase), and 450 main shaft RPM. The crushed samples (200 g) were ground by using a laboratory hammer mill (Perten, 120, Finland) equipped with 80-µm sieve, 750 W (1 phase) motor. The electrical current and voltage were measured over the grinding time with a digital AC clamp meter (KYORITSU, KEW SNAP, 2017, Thailand) and the collected data were used to determine electrical energy consumption during loading and unloading conditions. Eq. 9 and Eq.10 were applied to calculate the specific energy for both single and 3-phase grinders, respectively, using data obtained from clamp meter (Ngamnikom and Songsermpong, 2011; Rozalli et al., 2015). The total specific grinding energy (Et) was calculated as the sum of specific crushing energy (Esc) and specific grinding energy (Es) (Eq. 11).
 |
 |
 |
Where; I is the current (Ampere), V is the voltage (Volt), pf is the power factor (dimensionless), Mp is the mass of the prepared powder (kg), and t is the grinding time (sec).
Oil extraction process and oil yield calculation Oil extraction process was performed by adhering to that reported by Devshony et al. (1992) and Akbari et al. (2012). Ten grams of ground roasted date seeds were used to extract the oil by using Soxhlet apparatus. The oil was extracted from date seed samples using n-hexane solvent at 135 °C for 2 hr, and then, the vessel that contained the oil was dried in oven at 65 °C for 1 hr. The container was cooled in a desiccator and was weighed. The oil yield was determined using Eq. 12. The defatted samples were collected and preserved in plastic containers at 5 °C until further analysis. The experiments were repeated in triplicates and the average values ± standard deviation is used to present the outcomes.
 |
Colour measurement According to the procedure used by Mallek Ayadi et al. (2018), a portable colorimeter (CR 400, Minolta Co., Osaka, Japan) was used to measure the colour of palm date seeds using Lab system. The instrument was calibrated with a standard white calibration plate given along with the instrument and set to CIE Standard Illuminant C. Colour index is described by lightness (L* where L=0 for black; L=100 for white), red/green (+a* for red and -a* for green), and yellow/blue (+b* for yellow and -b* for blue). A white tile standard (L*=96.33; a*=+0.09; b*=+1.98) was used as reference (standard) colour. Twenty samples were randomly selected to assess the colour. The colour coordinates; L*, a*, and b*, were measured for each sample at different equidistant points on the seed surface. The average value of readings for each sample had been considered. The retrieved data are summarised as mean and standard deviation values. Additionally, total colour difference (ΔE) was used to express colour degradation/change during roasting process. The ΔE values were calculated using Eq. 13 (Palou et al., 1999; Akoy, 2014).
 |
Where, 
, 
, and 
are the colour coordinates for raw palm date seeds at zero time, while L*, a*, and b* refer to the colour coordinates for roasted palm date seeds at any time.
Hardness measurement The hardness of palm date seeds was measured by placing a single seed horizontally between two flat plates and force was applied on the thickness using Instron Universal Testing 5566 Machine (Canton MA, USA) at 10 mm/min speed. A 9.8 kN force was employed. The hardness values were derived from the force deformation curve in Newton (maximum beak of first compression) (Bourne, 1982). Ten measurements for each treatment were performed, and the results are reported as mean and standard deviation values.
Kinetics modelling of the colour and hardness changes The general reaction rate equation (Eq. 14) is usually utilised to model the quality indicators, such as colour and hardness (Peng et al., 2014). Table 1 presents the different derivatives of the general reaction rate equation using varied orders.
 |
| Reaction order | Model |
| Zero order (n = 0) | P = Po−kt |
| First order (n = 1) | P = Po exp (−kt) |
Where; P is quality parameter (color, hardness), t is roasting time (min), Po is the initial P and k is the reaction rate constant (m−1).
Calculation of the activation energy The Arrhenius-type equation (Eq. 15) was applied to calculate activation energy using the obtained reaction rate constant (k).
 |
Where k denotes the reaction rate constant (m−1), kref is the frequency factor (m−1), Tabs reflects absolute roasting temperature (K), and Rg means gas constant (8.314 kJ/mol. K).
Statistical data analysis Analysis of variance (two-way ANOVA) was applied to assess the effects of temperature and time on colour attributes, hardness, total specific grinding energy, and extracted oil yield of roasted palm date seeds. Zero- and first-order equations were applied to describe the experimental data via non-linear regression technique. The goodness of fit of the proposed models to the experimental data was examined by using two statistical indicators: coefficient of determination (Eq. 16), and sum of squared errors (SSE) (Eq. 17) using MATLAB R2010a software (MATLAB, 2010). The best model was selected based on the highest correlation coefficient (R2 ≥ 0.80) to signify a strong correlation and the lowest SSE (Prasad and Nath, 2002).
 |
 |
where, Xexp, Xpred and 
refer to experimental, predicted, and average moisture content values (% dry basis), respectively, whereas N represents the number of the experimental data.
Characterization of the raw palm date seeds The characteristics of raw palm date seeds (mass, length, diameter, equivalent spherical diameter, surface area, sphericity factor, colour attributes, and hardness) are summarised in Table 2.
| Characteristics | Mean ± SD |
|---|---|
| Mass (g) | 1.09±0.14 |
| Length (mm) | 18.07±0.73 |
| Diameter (mm) | 9.27±0.39 |
| Equivalent spherical diameter (mm) | 11.01±0.36 |
| surface area (mm2) | 416.89±25.9 |
| Sphericity factor | 0.92±0.01 |
|
0 36.62±0.75 |
|
1.60±0.62 |
|
30.16±3.00 |
| Hardness (kN) | 2.78±0.05 |
| Moisture content (% d.b.) | 2.26±0.11 |
Time–temperature profiles and thermal diffusivity The temperature of seeds against time serves as an indicator of roast-degree during the roasting process of palm date seeds. The typical time–temperature profile during roasting of palm date seeds at 160, 180, and 200 °C is portrayed in Fig. 2. Temperatures of seed centre were developed during the roasting process and it has been suggested that air temperature influences the temperatures of the seeds. Similar results were obtained by Hernández et al. (2007) for roasting coffee beans at different temperatures, ranging from 190 until 300 °C and Hii et al. (2017) for roasting of cocoa nibs at 120–140 °C. Thermal diffusivity (∝e) is a thermophysical property that defines speed of heat diffusion via conduction during changes of temperature. The higher the thermal diffusivity, the faster is the heat diffusion. Thermal conductivity can be determined via thermal diffusivity, specific heat capacity, and density. Fourier's law of heat conduction was applied to determine thermal diffusivity (Eq. 8). The values of thermal diffusivity for palm date seeds during the roasting process were 1.43x10−6, 1.52x10−6, and 2.74x10−6 m2/s for 160, 180, and 200 °C roasting temperatures, respectively. Hii et al. (2017) found that the thermal diffusivity of cocoa nibs during roasting process was 10−9 m2/s. This variance is attributed to the varying products and roasting temperatures. Fig. 3 illustrates the plot of thermal diffusivities versus roasting temperatures.
Time–temperature profiles, for the surrounding air, seed surface and seed centre, obtained during roasting of palm date seeds at the roasting temperatures (a) 160 °C, (b) 180 °C and (c) 200 °C.
Thermal diffusivity of the roasted palm date seeds versus roasting temperatures.
Analysis of colour and hardness changes and kinetics modelling Roasting processes are usually controlled by degree of colour formation due to the browning reaction progress that results from the increasing brown pigments (Saklar et al., 2001). Colour is an important indicator of roasting degree that has an impact on consumer preference (Maskan, 2001). Colour parameters (L*, a*, b*, and ΔE) were plotted against roasting time at varied temperatures, as displayed in Fig. 4.
Changes in colour coordinators (L* (A), a* (B), b* (C), ΔE (D)) of palm date seeds during roasting process at different temperatures for 30 min.
The L*-value represents the brightness or the darkness of a product. The L*-value is preferred for monitoring colour formation during roasting of product because it is analogous to the colour observation made by an operator in determining the degree of the roasted product (Moss and Otten, 1989; Wang and Lim, 2014). A two-way ANOVA test was performed to investigate the effects of roasting temperature and time on the colour attributes of palm date seeds during roasting. The results showed that the L*-values of palm date seeds decreased significantly as the roasting temperature and time were increased (see Fig. 4A). The decrease in L* value is attributed to the decreasing moisture content in the palm date seeds during roasting that slowed the Maillard reactions (Wang and Lim, 2014). It was reported that the phenolic polymerization and Maillard reaction are the chief reactions contributing to the formation of coloured compounds, called melanoidins, which give darker colour to roasted products (Bolek and Ozdemir, 2017a). As illustrated in Fig. 4A, the higher the temperature, the shorter is the time taken to reach the target roasting level. For instance, in order to reach dark roast (L* ≤ 25) based on the classification made by Pittia et al. (2007), the required roasting time had been approximately 30 and 17 min, when processed at 180 and 200 °C, respectively. The results are in agreement with the trend for whole hazelnut kernels during roasting using forced air pilot scale roaster at four temperatures (100, 120, 140, and 160 °C) for an hour (Özdemir and Devres, 2000), Arabica coffee beans roasted in a commercial fluidised bed roaster at four temperatures (220, 230, 240, and 250 °C) for 90 min (Wang and Lim, 2014), peanuts during the roasting process using a pilot plant dry roaster at temperatures that ranged from 149 until 204 °C for 100 min (Shi et al., 2017), and Polish hazelnuts roasted in a lab-scale ventilated oven at two temperatures (130 and 160 °C) for a maximum of 60 min (Marzocchi et al., 2017). Likewise, the b* value decreased as roasting temperature and time were increased (see Fig. 4C). This trend is similar to that observed for coffee during roasting process (Wang and Lim, 2014) and P. terebinthus (Bolek and Ozdemir, 2017a).
Fig. 4B portrays the changes in a*-value during roasting of palm date seeds. Generally, a significant increase was noted in a*-values of palm date seeds as the roasting time was increased. Similar trends were observed while roasting hazelnut (Özdemir and Devres, 2000) and P. terebinthus (Bolek and Ozdemir, 2017a). Therefore, it can be suggested that the non-enzymatic browning and pyrolysis reactions occurring during the roasting process of date seeds which enhance the development of brown pigments (Bolek and Ozdemir, 2017a), consequently gives the seeds a darker colour. This suggestion can be supported by the report of Pittia et al. (2007) who found that the brown colour increased with the increase of the roasting degree during the roasting of the coffee beans at three different degrees of browning (light, medium and dark). However, they stated that browning is, in turn, described by reduction in L*, a*, and b* attributes (Pittia et al., 2007).
Total colour difference (ΔE) is an important factor to determine colour change during roasting operations. Changes in ΔE at different roasting time and temperatures are displayed in Fig. 4D. Clearly, ΔE increased as the roasting temperature and time increased, signifying that the roasting process has a significant effect on colour change for palm date seeds. This observation is similar to that found by Yang et al. (2010), who iscovered that as the roasting temperature and time increased, ΔE of almond increased as well.
Apart from colour parameters, hardness is another aspect used to determine roasting level (Kahyaoglu and Kaya, 2006). Changes of hardness in palm date seeds during roasting process at varied temperatures are illustrated in Fig. 5. Statistical results indicate a significant decrease in hardness of palm date seeds as roasting temperature and time were increased. This finding proves gradually reduction in the strength of palm date seeds. Upon increment in roasting time and temperatures, the required force to break the seed appears to decrease. Similar observation was reported for sesame seeds (Kahyaoglu and Kaya, 2006), pistachio kernels (Shakerardekani et al., 2011), and P. terebinthus beans (Bolek and Ozdemir, 2017b). Modelling techniques are necessary to estimate the quality indicators during roasting processing or storage (Saguy and Karel, 1980). In order to evaluate the kinetic (zero- and first-order) models, two statistical criteria (SSE and R2) were determined using Levenberg-Marquard estimation method. The higher the R2 value, the higher the consistency between experimental and estimated data. Additionally, higher SSE indicates higher discrepancy between the real data and the model (Moore et al., 2013). Constants of different models and fitting criteria are presented in Table 3. In general, the first-order equation can adequately describe the colour parameters (L*-value, b*-value) and the hardness of the roasted date seeds. While zero-order model was the best to fit a*-value and ΔE data of the roasted date seeds during roasting process at different temperatures. As portrayed in Table 3, Average values of R2 of L*value, a*value, b*-value, ΔE and hardness were 0.962, 0.962, 0.901, 929 and 0.904, respectively. Although there are some high values of SSE in Table 3, R2 values are still in the acceptance area, judging from the high average value of R2 ≥ 0.80 (Lv et al., 2011). It was reported that the closer the value of R2 to the unity, the stronger correlation between the real and predicted data is (Moore et al., 2013). Notably, the order of chemical reaction is generally zero- or first-order for the food system (Özdemir and Devres, 2000). In fact, similar results were reported for roasted coffee (Wang and Lim, 2014).
Changes in hardness of palm date seeds during roasting process at different temperatures for 30 min.
| Model | Parameter(1) | L*-value | a*-value | b*-value | ΔE | Hardness (N) | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Roasting temperature (°C) | Roasting temperature (°C) | Roasting temperature (°C) | Roasting temperature (°C) | Roasting temperature (°C) | ||||||||||||
| 160 | 180 | 200 | 160 | 180 | 200 | 160 | 180 | 200 | 160 | 180 | 200 | 160 | 180 | 200 | ||
| Zero-order | Po | 36.8 | 35.5 | 34.3 | 1.9 | 1 | 1.2 | 31.1 | 29.3 | 27.2 | 0.4 | 2.6 | 3.7 | 2.9 | 2.4 | 1.7 |
| k | 0.229 | 0.366 | 0.523 | 0.121 | 0.314 | 0.345 | 0.323 | 0.402 | 0.365 | 0.387 | 0.581 | 0.711 | 0.032 | 0.063 | 0.062 | |
| R2 | 0.98 | 0.95 | 0.921 | 0.973 | 0.933 | 0.981 | 0.892 | 0.925 | 0.84 | 0.9423 | 0.919 | 0.926 | 0.958 | 0.869 | 0.555 | |
| SSE | 0.741 | 4.965 | 16.42 | 0.283 | 4.932 | 1.574 | 8.842 | 9.129 | 17.78 | 6.406 | 20.95 | 28.38 | 0.032 | 0.419 | 2.134 | |
| First-order | Po | 36.9 | 35.8 | 35.2 | 2.3 | 2 | 2.6 | 31.2 | 29.8 | 27.9 | 2.3 | 5.4 | 7 | 2.9 | 2.6 | 2.7 |
| k | 0.0068 | 0.0123 | 0.0206 | 0.0306 | 0.058 | 0.052 | 0.0121 | 0.0175 | 0.0178 | 0.0582 | 0.044 | 0.0432 | 0.013 | 0.0494 | 0.1656 | |
| R2 | 0.975 | 0.952 | 0.958 | 0.923 | 0.969 | 0.944 | 0.879 | 0.938 | 0.886 | 0.893 | 0.796 | 0.815 | 0.941 | 0.956 | 0.887 | |
| SSE | 0.941 | 4.727 | 8.842 | 0.806 | 2.258 | 4.73 | 9.93 | 7.558 | 12.69 | 11.93 | 52.46 | 70.88 | 0.04467 | 0.14 | 0.541 | |
Temperature dependency of colour attributes and hardness The impact of temperature on quality change reaction is usually determined via Arrhenius-type relationship (Eq. 15). Temperature dependence of the reactions rate of the colour attributes and hardness of the roasted palm date seeds using Arrhenius-type correlation is charted in Fig. 6. The dependency of the reactions rate of the colour attributes and hardness in relation to the roasting temperature was adequately described by the Arrhenius correlation. Table 3 presents the assumption of the best model, along with the Arrhenius-type relationship, to calculate activation energies for quality properties (colour and hardness) of palm date seeds during roasting process. As observed from Table 4, the activation energy for L*-value, a*-value, and hardness were 47.15, 45.32, and 108.34 kJ/mole, with R2 = 0.999, 0.836, and 1.00, respectively. These values are within the range of Ea values of colour changes in foods (41–125 kJ/mol), as reported by Saguy and Karel (1980). As for the b* coordinate and the total colour difference (ΔE), the activation energy values were 16.25 and 26.06 kJ/mole with R2 = 0.800 and 0.972, respectively. Notably, the activation energy values showed that L*-value, a*-value, and hardness are more sensitive to temperature variance during the roasting process, than b* and ΔE. Hence, it is inferred that the activation energy values (Ea) may change in accordance to heating system and product type (Kahyaoglu and Kaya, 2006).
Temperature dependence of the reaction rate constants using Arrhenius-type relationship.
| Attribute | Temperature (°C) | Kinetic model | Arrhenius equation | |||
|---|---|---|---|---|---|---|
| k (minȒ1) | R2 | Ea (kJ/mol) | ko (min−1) | R2 | ||
| L*-value | 160 | 0.007 | 0.975 | 47.15 | 8.11 | 0.999 |
| 180 | 0.012 | 0.952 | ||||
| 200 | 0.021 | 0.958 | ||||
| a*-value | 160 | 0.1205 | 0.973 | 45.32 | 10.6 | 0.836 |
| 180 | 0.314 | 0.933 | ||||
| 200 | 0.345 | 0.981 | ||||
| b*-value | 160 | 0.012 | 0.889 | 16.25 | 0.16 | 0.800 |
| 180 | 0.018 | 0.938 | ||||
| 200 | 0.018 | 0.893 | ||||
| ΔE-value | 160 | 0.3867 | 0.893 | 26.06 | 6.31 | 0.972 |
| 180 | 0.5813 | 0.796 | ||||
| 200 | 0.711 | 0.815 | ||||
| Hardness | 160 | 0.013 | 0.958 | 108.34 | 25.74 | 1.00 |
| 180 | 0.049 | 0.956 | ||||
| 200 | 0.166 | 0.887 | ||||
Effect of roasting conditions on the grinding energy consumption The analysis of regression showed that the total specific grinding energy is significantly influenced by roasting temperature and time. The second-degree polynomial equation (Eq. 18) was used to describe the experimental data (R2=0.986) with no lack of fit.
 |
Fig. 7 illustrates the total specific grinding energies of palm date seeds at varying roasting temperatures and time. Increment in roasting temperature and time led to a decrease in the total specific grinding energy. The total specific grinding energy reduced from 0.565 to 0.022 kW.hr/kg for raw to roasted seeds at 200 °C for 30 min. This result can be attributed to the fact that the seeds become more breakable and therefore, less energy is required for grinding as the moisture content and hardness of the seed are reduced. Analogous findings were reported for wheat grains and kernels (Dziki, 2011).
Contour plots for total specific grinding energies (kW.hr/kg) of palm date seeds at different roasting temperatures and times.
Effect of roasting conditions on the oil yield of palm date seeds Fig. 8 illustrates the contour lines of extracted oil yield, as affected by the roasting temperature and time. The regression analysis was performed to generate a polynomial equation (Eq. 19) that describes the oil yield data. The statistical outputs showed that the roasting temperature and the interaction between temperature and time significantly affected the oil yield, while there was no effect of roasting time on oil yield. Increment in roasting temperature caused a significant increase in oil yield percentage. In comparison, the oil yield of roasted palm date seeds increased by 75% with increment in roasting temperature. This result is attributable to the heating process that facilitates oil release by means of breaking the oil cell structure. Similar findings were reported for wild almond (Moghadas and Rezaei, 2017), locust bean (Akinoso and Raji, 2011), and orange seeds (Akinoso et al., 2011).
Oil yield (%) of palm date seeds at different roasting temperatures and times.
 |
Palm date seeds are a good source of nutritional components that can be incorporated in food applications and pharmaceutical industries. Oil can be extracted from roasted palm date seeds and the residuals of extraction process can be applied in preparing coffee-like brew. Roasting and grinding processes play crucial roles in several industries. Therefore, availability of information regarding the roasting conditions and their impact on the quality properties of palm date seeds is beneficial to manufacturers. In addition, prediction of quality attributes, grinding energy consumption, and oil yield as a function of roasting conditions would be useful for controlling these processes. In this study, changes in colour attributes and hardness of palm date seeds during roasting process via conventional oven were modelled. The effects of roasting conditions on grinding energy consumption and extracted oil yield had been studied. In short, the first-order equation can adequately describe L*-value, b*-value and hardness of the roasted date seeds, while zero-order model emerged as the best model to fit a*-value and ΔE data of the roasted date seeds during roasting process at various temperatures. The dependency of L*value and total colour difference on roasting temperature is adequately described by Arrhenius relationship. Increment in roasting temperature and time caused a decrease in the total specific grinding energy, while increasing the oil yield. As for powder and oil applications, the roasting of date seeds at 200 °C for 30 min can generate the lowest grinding energy and a high amount of oil.
Acknowledgment The authors acknowledge Universiti Putra Malaysia, IPS Grant with Vote Number: 9634800. The authors would like to thank the Research Centre in the faculty of food science and agriculture, King Saud University, Saudi Arabia for providing the palm date seeds to conduct this study. Also, many thanks to all members of Food Technology Research Institute, Agricultural research centre, Giza, Egypt as well as the authors acknowledge the efforts done by the members of the Supercritical Fluid Centre, UPM for their technical helping. The authors would like to take this opportunity to say warm thanks to Mr. Abdul Afiq from Halal Products Research Institute at UPM for his appreciated supporting.
