Optimizing Physiochemical and Sensory Properties of Infrared-Hot Air Roasted Sunflower Kernels Using Response Surface Methodology

Roasting sunower kernels is a key process in production of nuts. In this study, the eect of roasting conditions, including hot air temperature (120–160°C), infrared (IR) power (400–600W) and roasting time (3–10min) on energy and specic energy consumption, color parameters (L∗, a∗, b∗, ΔE, BI, SI, WI, and h°), texture, moisture content, chemical properties (pH and total phenolic contents, peroxide value (PV), and sensory properties of sunower kernel were investigated. In addition, the best models for the responses were obtained, and the proper roasting conditions were determined using response surface methodology (RSM). A quadratic model was proposed for color change (L∗, ΔE, SI, andWI), moisture and total phenol contents, linear relation for a∗, b∗, h°, and 2FI for BI, texture, PV, and pH. Roasting at 425.7W IR power and 124.3°C for 3.7min was found to be convenient or proper roasting conditions.


Introduction
Sun ower (Helianthus annuus L.), an annual plant, is grown in temperate and subtropical climate [1].
e sun ower seeds are rich in unsaturated fatty acids, proteins, ber, and vitamins (such as B1, B5, and B6), phosphorus, copper, manganese, folate, iron, zinc, amino acids, and lowcarbohydrate diets.ey are consumed as snack products, either alone or in the blends of di erent nuts and other components [2].
Roasting is the key process in the production of valueadded nuts having better taste, aroma, and a crunchy texture and exhibit enhanced crispiness.One of the common treatment methods is dry roasting.In this process, the nuts are heated applying the conventional thermal treatment, such as air convection and pan or sand roasting at 250-300 °C for a short time [3][4][5][6].In dry roasting, some novel thermal treatments could be used applying irradiation of microwave, infrared (IR), and radiofrequency [7].In the conventional heating method, the outer surface of the seeds can be overroasted, whereas the core of them is not completely processed.is causes uneven roasting, burnt surface, and development of the undesirable aroma and avor.Roasting of grains and legumes leads to the gelatinization of starch and denaturation of proteins, too [8].On microwave roasting, the temperature on lower surface of the food, rather than its inner parts, causes over-or underroasted food [9].
IR heating has many advantages over the conventional method such as (1) production of the high-quality food due to consistent and uniform distribution of IR energy; (2) high energy e ciency as a result of direct radiation transferring from IR emitters to the surface of product, without requiring any heating medium; (3) short process time; (4) precise controlling the IR heating process; (5) compact design and difference in types of IR emitters; and (6) environmentfriendly energy [10].
Although IR heating is a new promising approach, it is limited because of its low penetrating power.Combination of two convection (hot air (HA)) and radiation heat transfer mechanisms is important to increase the thermal efficiency of dryers [11,12].
IR heating was applied lonely or in combination with hot air in roasting of several products like almonds [13], pistachios, coffee [14], tea [15,16], peanuts [17], pecans [18], hazelnuts [7,19], cereal grains [20], and sesame seeds [21,22].ere are only a few reports on the roasting sunflower kernels.Most of the previous studies on roasted sunflower kernels investigated the influence of various factors such as change of molecular species of triacylglycerols during microwave roasting [23], color, texture firmness, moisture, percentage of extracted oil, acidity, peroxide, sensory properties, and fatty acid composition sunflower kernels were roasted under two different conditions, i.e., by microwave and electrical oven [24].
e aim of this study was to investigate the effect of power, temperature, and time of combined IR-HA roasting method on some physicochemical properties of the sunflower kernels.e proper conditions for IR-HA roasting were also determined using response surface method (RSM).

Sample Preparation. Sunflower seeds (Slatki variety)
were obtained from the agricultural farm (KhanBebin, in Golestan Province).e seeds were manually cleaned and deshelled.e percent of kernel recovery was 39.3%.e seeds were sorted to obtain the uniform size and defect-free ones.en, they were kept at 4 °C until using in experiments.Before roasting process, sunflower kernels were soaked in 25% (w/w) salt solution for 30 min.en, the excess water of sieved seeds was removed using cloth.

Experimental Design.
e results were analyzed using RSM based on the performed pretests, the independent variables were IR power (A) (400-600 W), temperature of HA (B) (120-160 °C) having constant velocity (1 m/s), and roasting time (C) (4-10 min).e roasted samples were analyzed according to a central composite design (CCD).
e experimental design consisted of 20 treatments including 6 center points and 14 noncenter points, as shown in Table 1.

Combined IR-HA Roasting Process.
Roasting was performed in a combined IR-HA roaster equipped with a controller to adjust the air temperature and IR power of roasting.
e apparatus consisted of heating elements (4 electrical elements with power 750 watts), scale, pitot unit, a centrifugal fan, infrared radiator (Far infrared lamp, 1500 W), and a roasting chamber (Figure 1).A single layer of the seeds was spread on aluminum trays.Finally, the processed kernels were kept at room temperature to cool (25 ± 2 °C).Until performing the experiments, the samples were kept in polyethylene Zip-pack bag at room temperature (25 °C).

Specific Energy Consumption.
e energy consumption in each period of roasting was calculated using the following equation where E is the energy consumption (kW•h), P is the consumed power by the roaster system (W), and t presents the roasting time (min).Using (2), the specific energy consumption was calculated at different roasting conditions: e parameter hue angle (h °) expresses the color tone and is defined as red-purple: 0 °, yellow: 90 °, bluish green: 180 °, and blue: 270 ° [25].Total color difference (ΔE), browning index (BI), saturation index (SI), h °and whiteness index (WI) were calculated using Equations ( 3)-( 8) [26]: where L * , a * , and b * refers to the difference between the color parameters of the fresh sunflower kernel and the treated one.

Texture Analysis.
According to definition, texture hardness is the required force over compression test.e textural analysis of the sunflower kernels was performed using a TA.XT Plus Texture Analyzer (Texture Technologies Corp Surrey, UK TA-XT Plus, Stable, England).In texture profile analysis (TPA), 10 sunflower kernels were placed on the platform of the texture on the following settings as described by Soleimanieh et al. [24]: Probe P/36R used a load cell 250 N, a diameter of 32 mm, a speed of 5 mm/min, and a penetration index of the probe was 50% of the kernel thickness.Textural properties were derived from the force time curves.To study the textural properties of the samples, hardness (maximum peak of first compression) measured.

Moisture Content.
Moisture content of roasted kernels (10 g) was measured according to ISO665 [27] in oven drying method (103 ± 2 °C) until a constant weight was reached, and results are expressed as % db.
e extract was filtered through filter paper (Whatman no.40) and used for analysis of pH. e pH of the extract was measured using a pH meter (pH-lab, model 827, metrohm, Germany) with a glass electrode.

Total Phenolic Content (TPC).
e TPC was determined based on the Folin-Ciocalteu colorimetric method with slight modification as described by Bagheri et al. [29]. 1 g of powdered samples of sunflower kernel was mixed with 10 mL hexane, and the phenolic compounds were extracted by 12 mL methanol/water (60:40, v/v) for 30 min.e mixture was filtered through a filter paper, and the extracts (both phases) were separated by centrifugation (at 3500 rpm for 10 min).An aliquot (0.2 mL) of the methanolic phase was diluted using water to a total volume of 2.5 mL, followed by the addition of 0.25 mL Folin-Ciocalteu reagent.After 3 min, 0.5 mL sodium carbonate solution (35%, w/v) was added to the reaction mixture, which was finally mixed and diluted with distilled water to 5 ml.e absorbance of the solution was measured after 2 h against a blank sample (methanol) by a spectrophotometer (UV1601, Shimadzu, Kyoto, Japan) at a wavelength of 760 nm. e concentration of total phenolic compounds was determined by comparison with the absorbance of gallic acid at different concentration as standard.
2.6.Sensory Analysis.Sensory attributes of roasted kernels including appearance, hardness, taste, aroma, and total acceptance were assessed by using a five-point hedonic scale (5 point: the best, 1 point: the worst).A panel consisting of 10 members (all food experts of the Golestan Standard Office), with sensory evaluation experience, was trained in descriptive evaluation of sunflower kernels.Mean values were calculated and used to determine the sensory attributes of the roasted sunflower kernels.

Statistical Analysis and Optimization. Design-Expert
Software (Version 10.0.1.0,2016, Stat-Ease, Inc., Minneapolis, MN) was employed to find the mathematical model for any response and to present the relationship between the dependent and independent variables as three dimensions' Journal of Food Quality curves.e analysis was performed individually for all the responses.Also, optimization of the process conditions was simultaneously performed using numerical and graphical optimization tool of Design-Expert software.

Energy Consumption and Specific Energy Consumption.
e energy efficiency of the process and quality of roasted product are two key factors in food roasting.Figures 2(a e minimum and maximum values of specific energy consumption for roasting of sunflower kernels were 0.37 kW•h/kg in treatment 500 W, 140 °C, 2 min and 1.50 kW•h/kg in treatment 500 W, 140 °C, 10 min.Regression analysis indicated that the quadratic model (R 2 � 0.9915) sufficiently described the heat consumption (Equations ( 9) and ( 10)).Energy consumption increased with increasing power and time (p < 0.05), whereas no statistical difference in heat consumption was observed due to the temperature (p > 0.05).

E(kW•h)
Heat consumption of dried barley was compared under far infrared (FIR)-convection and convection alone.It was reported that the use of FIR radiation enhances the drying rate and the energy consumption was reduced considerably [31].In another research, it evaluated the energy consumption during IR-HA roasting of peanut kernels and reduced the energy consumption up to 31% in the HA method [17].Also, it was observed that in drying mushroom and pomegranate, using the combination of HA and IR decreased energy consumption relative to IR drying alone and increased it relative to HA drying [32,33].e use of IR method in drying persimmon sheet, drying time at three levels of 50, 60, and 70 degrees, respectively, decreased by 46.39%, 30.66%, and 22.38% [34].

Color Measurement.
Visual color is an important quality criterion for roasted foods and may be assessed in terms of quantitative parameters such as L * , a * , b * , color differences (ΔE * ), hue angle (h °), and saturation or chroma (SI) values.Color also is an important quality indicator in the case of roasted foods, and optimized color parameters may be used as quality control indicators during the roasting processes [25,26].
Figures 3(a) and 3(b) show the degree of lightness change (L * ) during roasting of sunflower kernels.e result showed that the highest value for L * was observed in raw kernels (66.55) (Table 2), and power, temperature, and time of thermal processing had significant effect (p < 0.05) on L * , but temperature and time had no interaction effect on it (p > 0.05).A quadratic model (R 2 � 0.9733) was applied between different parameters as As shown in Figures 4(a), 4(b) and 5, there was a linear relation (p < 0.05) between used power, time, and temperature and the changes in a * and b * : e value L * , a * , and b * of roasted kernels were 43.75-61.34,−0.56-5.61,13.44-20.8,respectively.a * value or the amount of redness of treatments increased by increasing the roasting variables (p < 0.05).Also, according to Figure 5, unlike the temperature of heating that had not significant effect (p > 0.05), the different amounts of power and time of roasting had significant effects (p < 0.05) on b * value or degree of yellowness of treatments.
e result of model fitting for ΔE compared with fresh samples (Table 2), SI, BI, h °, and WI of roasted kernels by RSM are given in Figures 6-10.ΔE, SI, and WI changes were described by a quadratic model and BI by two factor interaction (2FI) model and h °by a linear relation (Equations ( 13)- (17)).
e values ΔE, SI, BI, h °, and WI of roasted sunflower kernels were in the range of 6.38-23.20,13.93-21.20,28.67-51.93,72.17-89.84,and 42.29-57.95,respectively WI � +89.75183 − 0.27457A + 0.029266B + 2.40837C   and BI roasted peanut kernels by HA [29], but Sharma and Gujral [8] reported a decrease in ΔE during roasting of barley.Also, it was reported that increases in roasting temperature (90-190 °C) and time (5-65 min) caused a decrease in L * , b * , and yellow index except for a * value of ground state-roasted pistachio kernels with HA. e quadratic model sufficiently described the changes in color [35].ese results are in agreement with those found by Soleimanieh et al. [24] for oven and microwave-roasted sunflower seed, Wani et al. [36] for pan and microwave roasting arrowhead tubers, Gujral et al. [37] for sand and microwave roasting of oats, Uysal et al. [19] for roasted hazelnut by using microwave-infrared combination.e values of L * and h °color parameters coffee roasted using HA were evaluated.With increasing temperature, L * value and h °decreased.A second-order correlation was found between L * and roasting time [38].Change in color is due to formation of brown pigments with low and high molecular weight through nonenzymatic browning (NEB) reactions such as Millard reaction and caramelization and phospholipids degradation during roasting [26,36,39,40].

Texture Analysis.
Texture is one of the most important criteria in consumer acceptance of roasted sunflower kernels.In general, the roasting process is responsible for the crunchy texture of nuts resulting from structural and chemical changes such as the Millard reaction, decrease in moisture content, or lipid modifications reaction [41].e results of hardness of kernels during roasting process were shown in Figure 11.Regression analysis indicated that twofactor interaction (2FI) model (R 2 � 0.7256) was the best to explain the effects of the three independent parameters on texture as following: Hardness of texture(N•S) � +21874.49716− 28.70636A − 128.36302B − 974.46144C + 0.16857AB.(18) e result showed that used power, temperature, and time had significant effect (p < 0.05) on texture, but power and time or temperature and time had no interaction effect on it (p > 0.05).
e minimum and maximum values of hardness of texture were 256.9 NS in treatment 401 W, 158 °C, 3.6 min, and 1962 NS in treatment 401 W, 121 °C, 3.6 min.During the roasting, moisture content of kernels decreased and they became crispier and fragile that required force to break was less; therefore, hardness decreased.Similar results were reported for roasting roasted sunflower kernels by HA and microwave oven [24], peanut kernels by IR-HA roasting [17], roasted hazelnut by using microwave-IR [19], and roasted pistachio by HA oven [35,42].

Moisture Content.
e changes in the moisture content of kernels during roasting were presented in Figure 12.A quadratic model (R 2 � 0.7948) was applied to describe the changes of moisture during roasting process as following (Equation ( 19)).Power significantly (p < 0.001) affected the moisture removal during roasting.Effect of other roasting parameters (temperature and time) on moisture content was not significant (p > 0.05).e moisture contents of samples were in the range of 0.1% in treatment 500 W, 109 °C, 6 min, and 4.9% in treatment 334 W, 140 °C, 6 min: moisture content(%) � +12.80062 − 0.070874A Moisture content of IR-HA roasted peanut kernels decreased significantly by increasing IR power from 130 W to 200 W and led to decrease 22.4% in the moisture content [17].Also, by increasing and lower lamp powers of the microwave-IR combination oven for roasting of hazelnut, moisture content decreased [19].
ese results are in agreement with those found by Soleimanieh et al. [24] for roasted sunflower kernels and Nikzade et al. [42] and Shakerardekani et al. [35] for roasted pistachio.

pH.
In this study, regression analysis indicated that 2FI model (R 2 � 0.6141) was the best to explain the effects of the three independent parameters on pH as following (Equation ( 20)).Temperature and time had significant effect (p < 0.05) on pH (Figure 13).
e pH of sunflower kernels before roasting was 5.52.In general, the pH increased with an increase in temperature and time of roasting.Bagheri et al. [29] conversely reported that a decrease in pH of peanut kernels during roasting due to Millard reaction products and conversion of sugars to acidic compounds during roasting.e effect of pH has been evaluated in model studies finding a close relationship with the formation of volatile compounds; low pH (≤5) favors the formation of pyrazines and higher pH favors the methyl pyrazines and Strecker aldehydes [43].

Total Phenolic Content (TPC).
Roasting treatment influences TPC as well as antioxidant activities for nuts including cashew, hazelnut, and peanut [44].ANOVA showed that the model developed for the TPC was quadratic (R 2 � 0.7893) as in Equation ( 21) and Figure 14. e minimum and maximum values of TPC were 12.87 and 64.89 mg gallic acid equivalent (GAE)/g in treatment 401 W, 158 °C, 3.6 min and 665 W, 140 °C, 6 min, respectively.Interestingly, TPC increased compared with fresh samples (Table 2) by roasting.
ere are few works in the literature reporting data on the TPC of IR-roasted sunflower kernels.e TPC in roasted hazelnuts using IR and HA at two different time/temperature conditions ranged from 0.48 to 0.69 mg•GAE/g.e TPC increased during roasting compared with fresh samples (0.42 ± 0.01).e TPC in roasted samples by using HA was greater than the TPC roasted samples by using IR, probably because IR caused a higher heating in the hazelnut than HA and higher degradation of phenolic compounds [7].
Lin et al. [44] reported that the levels of total phenols substantially decreased in the initial roasting phase (5 min)  8 Journal of Food Quality and then increased gradually with roasting duration in roasted almond kernel into air convection oven.Other authors have observed the same behavior in other products [17,25,45].e reason for increasing the TPC upon roasting could be because of increasing in heat induced and extractable phenolic compounds after roasting to the formation of Millard products, complex physical and chemical reactions on phenolic, including leaching of water soluble phenolic, freeing phenolic from bond forms, degradation of polyphenols, and breakdown, and transformation of phenolic, such as formation of complex products from phenolic and proteins, and formation of Millard reaction products having antioxidative activity such as pyrroles and furans that react with Folin-Ciocalteu reagent [46,47].Lin et al. [44] indicated that Millard reaction products produced during thermal treatments may increase the antioxidant effects of processed foods.show primary oxidative rancidity.ANOVA showed that the model developed for the PV was linear (R 2 � 0.6220) as in Equation ( 22) and Figure 15.It can be concluded that roasting temperature had a more pronounced effect on sunflower kernels oxidation than power.Peroxide values up to 10 meq O 2 /kg oil in sunflower kernels can be considered safe according to ISIRI14615 [48].ese results were in agreement with Nikzade et al. [42], Yang et al. [13], Soleimanieh et al. [24], and Belviso et al. [7].    of Food Quality roasted kernel composition will vary with the processing conditions, which is reflected on the external color of the kernels (varying from light to dark brown), and development of typical taste and flavor [49].Sensory attributes for roasted kernel were evaluated using 5-hedonic scale and are presented in Figure 16.e sensory evaluation of roasted kernel with IR-HA method indicates that IR technology did not affect any undesired taste and flavor, texture, or appearance to the kernels that could be detected by panelists.

Determination of the Optimum Roasting Condition
Optimal roasting conditions were investigated based on the lowest energy consumption which is important in determining the optimal conditions.e best suggested condition by the software was 425.7 W of power and heating at 124.3 °C for 3.7 min (Figures 17(a

Conclusion
Roasting is one of the methods of improving color, flavor, and taste in seeds, and the conditions of roasting play an important role in the appearance of tissue and sensory response procedure in optimizing the IR-HA roasting process.In conclusion, this study showed that energy consumption increased with increasing power and time.TPC in most treatments increased compared with fresh samples during roasting.ese may provide inexpensive sources of natural antioxidants for use as functional food ingredients and nutraceuticals.A quadratic model was proposed for color change (L * , ΔE, SI, WI, moisture content, total phenol contents) and linear relation for (a * , b * , h °) and 2FI for BI, texture, pH, and PV.For sunflower kernels, roasting at 425.7 W IR power and 124.3 °C for 3.7 min was found to be proper roasting conditions.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.
) and 2(b) show the energy and specific energy consumption during roasting of sunflower kernels.

Figure 3 :
Figure 3: e effect of IR power and time (a) and IR power and temperature (b) of roasting on the L * value of sunflower seed kernels.

Figure 2 :
Figure 2: e effect of IR power and time of roasting on energy (a) and specific energy consumption (b) during IR-HA roasting of sunflower seed kernels.

CFigure 5 :Figure 4 :
Figure 5: e effect of IR power and time of roasting on color parameter (b * value) of sunflower seeds kernels.

∆Figure 6 :Figure 7 :
Figure 6: Response surfaces of ∆E value of sunflower seeds kernels during IR-hot air roasting in relation to IR power and temperature (a) and IR power and time (b).

Figure 8 :
Figure 8: Response surfaces of color parameters (BI) of sunflower seeds kernels during IR-hot air roasting in relation to IR power and temperature (a) and air temperature and time (b).

Figure 9 :
Figure 9: e effect of IR power and air temperature of roasting on hue angle of sunflower seed kernels.

AFigure 10 :Figure 11 :
Figure 10: e effect of IR power and air temperature (a) and IR power and time (b) of roasting on WI of sunflower seed kernels.

Figure 12 :
Figure 12: e effect of air temperature and IR power of roasting on moisture content of sunflower seeds kernels.

Figure 13 :
Figure 13: e effect of air temperature and time of roasting on pH of sunflower seeds kernels.

Figure 14 :
Figure 14: e effect of IR power and time of roasting on total phenol content of sunflower seeds kernels.

Figure 15 :
Figure 15: e effect of IR power and temperature of roasting on peroxide value of sunflower seeds kernels.

Figure 16 :
Figure 16: Effects of different roasting conditions on the sensory analysis of roasted sunflower kernels.

A
: p o w e r ( W ) B : te m p e r a tu r e (

Figure 17 :
Figure 17: e optimum values of the operating conditions of roasting of sunflower seed kernels in the IR-hot air roaster.

Table 2 :
Composition of raw sunflower seed kernels.
Mean value is the average of three replications.