Quantifying of the Best Model for Prediction of Greenhouse Gas Emission, Quality, and Thermal Property Values during Drying Using RSM (Case Study: Carrot)

Abstract

The aim of this study is to use the response surface methodology (RSM) to mathematically model the response parameters and emission of greenhouse gases (GHG) and optimize the drying variables for a carrot dried with the microwave method using various pretreatments. To this end, the influence of the drying parameters (independent), such as microwave power and slice thickness dried by two pretreatments of ultrasonication at 30 °C for 10 min and blanching at 70 $℃$ for 2 min, was explored on the dependent (response) parameters including the thermal properties (drying time, effective moisture diffusion coefficient (D_eff), specific energy consumption, energy efficiency, quality features (color changes and shrinkage), and GHG emission (including CO₂ and NO_x). It should be mentioned that the emission of GHG was determined based on the energy consumption of various types of power plants such as the gas turbine steam power turbine, and combined cycle turbines using various fuels such as natural gas, heavy oil, and gas oil. The results indicated that the ultrasonication and blanching pretreatments can decrement the drying time (linearly), energy consumption (linearly or quadratically), shrinkage(quadratically), and color changes(quadratically) and enhance the D_eff (linearly) and energy efficiency (linearly or quadratically) in all samples with R² > 0.86. Moreover, the shortest drying time (42 min), lowest SEC (9.51 MJ/kg), and GHG emission ((4279.74 g CO₂ in the combined cycle turbines plant, and 18.16 g NO_X in the gas turbine plant) with natural gas for both plants) were recorded for the samples pretreated with blanching while the lowest color changes (13.69) and shrinkage (21.29) were observed in the ultrasonicated samples. Based on the optimization results, a microwave power of 300 W and steam power turbine of 2 mm were the best variables with a desirability of about 80% which resulted in the highest-quality products at the lowest GHG emission.

1. Introduction

A carrot is one of the vegetables belonging to the group of apiaceae with the scientific name of Daucus carrot L. [1]. The root of this plant contains various minerals such as iron, copper, magnesium, and phosphorous, as well as vitamins such as group B vitamins, and C [2]. A carrot is a rich source of beta-carotene, i.e., a precursor of vitamin A, such that the daily use of 100 g of this product can supply 17% of the total recommended daily intake of vitamin A [3]. These benefits, along with the desirable taste of this vegetable, have resulted in the application of carrots in various foods [4].

As most of the vegetables are seasonal (i.e., available only at a specific time of the year), their abundance in the harvesting seasons and thus their low price may result in their spoilage [5]. In this regard, the use of processing and preservation can prevent this spoilage, making them available at the other times of the year at profitable prices [6]. Drying is one of these methods in which the moisture content (MC) of the sample is decreased to a constant level [7]. One of the most common drying techniques in food processing is microwave drying. The superiority of this method over the others lies in the uniform drying of the products which improves their quality [8]. With the increasing demand for pre-cooked foods, the development of the microwave drying method is of crucial importance for agricultural products.

Various methods have been proposed for preserving the quality of dried agricultural products, among them being pretreatments with edible coatings [9], osmotic dewatering [10], microwaving [11], ultrasonication [12], blanching [13], ethanol and citric acid [14], and electric pulsing [15] before and after drying. The results have shown that the application of different pretreatments can improve the thermal properties while preserving the product quality.

As the drying process accounts for 12–20% of the energy use of the processing industry, optimization of this process in terms of energy use is of crucial significance [16]. Diverse methods have been developed for optimization, among which is the response surface method (RSM). In the RSM, the relationship between the response and independent variables can be determined through mathematical modeling [17]. Numerous studies have optimized and modeled the impact of the drying variables (such as air speed and temperature, microwave power (MP), and slice thickness (ST)) on the response parameters including mechanical and thermal qualities as well as the quality of the agricultural products such as cumin seeds [18], sweet potato slices [8], and Cornus mas [10]. In all the mentioned studies, the researchers have reported the best drying variables and pretreatments to achieve the best parameters; however, no similar study was found for carrot slices.

Iran is one of the richest countries in terms of energy sources including fossil fuels. Natural gas (NG), heavy oil (HO), and gas oil (GO) are among the main sources of energy applied for electricity generation. High energy consumption of agricultural product dryers will result in greenhouse gas (GHG) emissions such as CO₂ and NO_x which will result in serious environmental consequences. Such a high GHG emission can be one of the major sources of global warming. Global estimates have suggested that 80% of the raw energy in the world is supplied from fossil fuels. In 2013, 76 million barrels of oil were consumed, which is expected to grow to 123 million barrels by 2025 [11]. Therefore, any decline in energy consumption will lead to lowering GHG emission. The emission of GHG arisen from overconsumption of fossil fuels has led to environmental consequences such as ozone layer depletion, global warming, climate change, flooding and acidic rainfalls, and the rising level of sea water, which will have a serious negative impact on human lives. As the energy use of dryers is based on fossil or electricity fuels, the consumption of energy, and hence, GHG emission of the drying process, should be decremented. So far, several researchers have addressed the effect of the drying methods combined with various pretreatments on GHG emission from drying agricultural products [19,20,21,22].

Previous reports have shown that no study has addressed the optimization and modeling of the effects of the drying variables and pretreatments on the thermal and quality properties, as well as the GHG emission of the drying process of carrot slices. In this regard, the present research is highly necessary. The aim of this study is, thus, the establishment of the best mathematical model for the optimization of quality (color and shrinkage) and thermal (drying time (D_t) properties, effective moisture diffusivity (D_eff), specific energy consumption (SEC), energy efficiency (E_ef), and GHG emission (including NO_x and CO₂), which were determined for six different plants (including steam power-natural gas (SP-NG), steam power-heavy oil (SP-HO), gas turbine-natural gas (GT-NG), gas turbine-gas oil (GT-GO), combined cycle turbines-natural gas (CC-NG), and combined cycle turbines-natural gas (CC-GO)). The drying variables (independent) included MP (including 100, 200, and 300 W) and ST (including 2, 4, and 6 mm). Two pretreatments of ultrasonication and blanching at 30 °C and 70 °C for 10 and 2 min were also considered, respectively.

2. Materials and Methods

2.1. Carrot Sample

Fresh medium-sized carrots (length: 0.17 m and diameter: 0.03 m) were bought from a local market in Parsabad (Ardabil, Iran). The carrots were transferred to the agricultural science and natural resource faculty of Moghan. The samples were then washed, and their MC was measured using an oven (Memmert company, UFB500, Schwabach, Germany) at 105 °C for 24 h [23] and the MC was evaluated based on the standards [24]. The experiments were carried out in triplicates to ensure the validity of the results. Then, Equation (1) was employed to calculate the moisture content (MC) of the carrot which was 11.95 ± 0.55 (d.b.) [14].

$$M{C}_{d.b.}=\frac{M_i-M_f}{M_f}$$

(1)

In Equation (1), MC_d.b. shows the initial MC while M_f and M_i are the final and initial weight of the sample in kg, respectively.

2.2. Ultrasonication and Blanching Pretreatments

The samples were classified into three groups before drying: (I) controls (with no pretreatment), (II) ultrasonicated samples, and (III) blanched samples. Ultrasonication was carried out under laboratory conditions using an ultrasonic bath (Parsonic company, 750S model, Iran, Tehran) at 30 °C and 28 kHZ frequency with ratio of 1:4 for sample-to-water [15]. For blanching, the samples were put in hot water at constant temperature of 70 °C for 2 min [14]. After the pretreatment, the slices were withdrawn from the water and their surface-adsorbed water was cleaned by a towel tissue. All the samples were then transferred to the microwave dryer until reaching a constant weight.

2.3. Microwave Dryer

The samples were transferred into a domestic microwave oven (Sharp R-861SLM) for drying. All the experiments were carried out for three different ST (2, 4, and 6 mm) and MP (100, 200, and 300 W) in triplicate. The tray with the samples was put on a balance scale to measure the weight of the slices at regular 2-min intervals using a digital balance (AND company, GF-6000 model, Tokyo, Japan). About 0.01 kg of the carrot slices were put in the microwave chamber. The process of drying continued from the initial MC of 11.95 ± 0.55 until the final MC of 0.11 ± 0.1 (d.b.) [25]. The time taken by this process is called D_t.

2.4. Effective Moisture Diffusivity

To measure the D_eff, the diagram of the drying kinetics and the Crank equation based on the second Fick’s law were applied. To this end, the MC was plotted versus the time of the drying process. Then, the moisture ratio was determined, and its logarithmic diagram was drawn versus time and fitted with a straight line. The second Fick’s law deals with the mass diffusivity during the drying process’s descending phase, and it may be solved for a variety of geometries by employing the right boundary conditions. Equation (2) can be used to calculate the Fick equation for a thin layer:

$$MR=\frac{M_t-M_e}{M_b-M_e}=\frac{8}{{\pi }^2}{\sum }_{n=1}^{\infty }\frac{1}{(2n+1)}exp\left(\frac{-D_{eff}{(2n+1)}^2{\pi }^{2}t}{4L^2}\right)$$

(2)

In which L denotes the sample thickness (m); n shows the number of the drying terms (1, 2, 3, 4, …). M_t, M_e, and M_b represent the sample moisture at time of t, equilibrium, and initial MC (d.b.), respectively. MR shows the moisture ratio, while t stands for the D_t. D_eff is also expressed in m²/s.

If both sides of the equation are considered a natural logarithm, Equation (3) is obtained which is a straight line [26]. Putting the slope of this line equal with the slope of the fitted line of the natural logarithm, the value of D_eff can be determined (Equation (4)) [27].

$$ln(MR)=ln(\frac{8}{{\pi }^2})-ln\left(\frac{-D_{eff}{\pi }^{2}t}{4L^2}\right)$$

(3)

$$Slope=-D_{eff}\frac{{\pi }^2}{4L^2}$$

(4)

2.5. Specific Energy Consumption

SEC shows the ratio of the total energy use to the water lost during the drying process. Based on the experiments, the total energy use was obtained by measuring the energy consumption of the microwave, ultrasonic wave generator, and blanching heat according to the reported equation (Taghinezhad et al., 2020 [11]). The SEC was obtained by Equation (5) [28].

$$SEC=\frac{TEC}{M_w}$$

(5)

In Equation (5), SEC shows the specific energy consumption (MJ/kg⁻¹ water), while TEC denotes the total energy use (MJ). M_w also represents the value of water leaving the sample during the process of drying (kg).

The E_ef (${\eta }_{M_w}$) of the microwave dryer for drying carrot slices can be obtained by Equation (6) [29].

$${\eta }_{M_w}=\frac{M_w·{\lambda }_w}{P·t}$$

(6)

where P (W), λ_w and t are the microwave power (W), the latent heat for water vaporization (2257 kJ/kg), and D_t (min), respectively.

2.6. Greenhouse Gas Emission

Consumption of the dryer energy is based on the fossil or electricity fuels (NG, HO, and GO). Therefore, various methods have to be explored to reduce energy use, and hence the emission of GHG. Regarding the climate of Iran, electricity is mainly produced in the PP using fossil fuels. Regarding the fact that Iran is one of the countries with the highest NG and HO resources, it is also one of the biggest producers of GHG. As all the PP of Iran are GT-GO, GT-NG, SP-HO, SP-NG, and CC-NG, the mean GHG emission (NO_x and CO₂) was determined for various types of plants with different fuels for the production of 1 kWh energy as depicted in Figure 1 [30].

2.7. Color

Color is one of the major features of agricultural products. Color and its variations are directly related to the chemical and physical characteristics of the sample [31]. Portable colorimeter (HP company–200 model, Guangdong, China) was employed to obtain the index of the sample color. For each test, 10 carrots were evaluated, and their mean values were considered and compared with the fresh samples. Parameters of L* (varying from 0–100), a* (varying from −60 to 60), and b* (altering from −60 to 60), respectively, indicate the white to black, green to red, and blue to yellow colors [32]. The total color variations of the carrot slices can be calculated by Equation (7) [25].

$$\Delta E=\sqrt{(L_1^{*}-L_0^{*}{)}^2+(a_1^{*}-a_0^{*}{)}^2+(b_1^{*}-b_0^{*}{)}^2}$$

(7)

In which $a_1^{*}$, $b_1^{*}$, and $L_1^{*}$ show the color values of the dried carrot using a microwave dryer with various pretreatments under different experimental conditions. Variables $a_0^{*}$, $b_0^{*}$, and $L_0^{*}$ also represent the initial color of the slices.

2.8. Shrinkage

For determination of carrot shrinkage, first the initial volume of the carrot sample before and after the drying process with toluene were measured using a pycnometer. Then the shrinkage was determined from the ratio of the difference between the initial and dried samples to the initial volume as expressed in Equation (8) [4].

$$S_v=\frac{V_i-V}{V_i}\times 100$$

(8)

In which V_i and V, respectively, denote the sample volume before and after drying (cm³). S_v also represents the shrinkage (%).

2.9. RSM

In this research, RSM was used in the form of Historical data design as implemented in Design Expert 10 software to analyze the data and optimize the drying process of carrots with a microwave dryer under different pretreatments. The independent variables (input) of the Design Expert software in the microwave dryer included the MP (at three levels of 100, 200, and 300 W) and the ST (2, 4, and 6 mm). The experiments were performed in the form of factorial design on a completely random basis in 3 repetitions. The dependent variables are D_t, D_eff, SEC, E_ef, GHG emissions (CO₂, NO_x), color, and shrinkage. They were used to model the behavior of two-line operating systems and investigate their interaction effects [10].

$$y_k={\beta }_0+{\sum }_{i=1}^k{\beta }_i{x}_i+{\sum }_{i=1}^k{\beta }_{ii}{x}_i^2+{\sum }_i{\sum }_{j=2}^k{\beta }_{ji}{x}_j{x}_i$$

(9)

where y_k is the response variable, x_i shows the independent variable value, x_j is the actual value of each parameter, β₀ and β_i denote the intercept and the coefficients of the variables, respectively, while β_ji is the coefficients of the interaction effects of the variables. Drying variables (independent) are shown in

Table 1. The coded levels of the independent variables studied and their levels.

Independent Variables	Coded Variables	Levels
		−1	0	+1
Thickness (mm)	A	2	4	6
Microwave power (W)	B	100	200	300

with their coded levels.

Three important steps in the process optimization with RSM include a statistical design of experiments, a prediction of coefficients in a mathematical model, and an estimation of response parameters and a study of model accuracy in the experimental designs [18]. Optimum operating conditions were achieved using optimization with RSM. For this purpose, optimization objectives, response levels, and independent variables were first set. Then, the best answers were obtained using the desirability function technique. In this study, achieving the shortest D_t, lowest SEC, GHG emission, color, and shrinkage, and the highest D_eff and E_ef were considered as the objectives of experiments in statistical analyses for optimization.

3. Results and Discussion

3.1. Drying Time (D_t)

The results of the variance analysis and modeling of different drying parameters to predict the response variables are listed in

Table 2. Coefficients of the models fitted to the regression equation of response variables (A: Thickness and B: microwave power).

Pretreatment	Parameters	No	Equation	Model	R2	Adj R2	Pred R2	CV (%)	p-Value
Control	Time	10	+135.77 + 12.88A − 0.37B	Linear	0.9603	0.957	0.9483	7.15	<0.0001
	Deff	11	−3.68 × 10−10 − 1.80 × 10−10A − 1.52 × 10−13B + 9.34 × 10−10AB	2FI	0.9818	0.9795	0.9751	9.36	<0.0001
	SEC	12	17.01 + 2.06A + 0.005B − 0.0001B2	Quadratic	0.9340	0.9254	0.9089	6.14	<0.0001
	Eef *	13	+9.97 – 0.47A − 0.004AB	2FI	0.9046	0.8921	0.8714	8.72	<0.0001
	Color	14	+28.29 − 6.73A + 0.01B − 0.94A2	Quadratic	0.8810	0.8655	0.8326	4.75	<0.0001
	Shrinkage	15	+30.59 − 4.36A − 0.02B − 0.62A2	Quadratic	0.8602	0.8420	0.8076	3.67	<0.0001
US **	Time	16	+124.17 + 12.97A − 0.35B	Linear	0.9572	0.9516	0.9415	5.82	<0.0001
	Deff	17	−2.00 × 10−11 − 1.76A × 10−10 − 4.30 × 10−12B + 9.81 × 10−13 AB + 1.06 × 10−14B2	Quadratic	0.9844	0.9815	0.9762	8.82	<0.0001
	SEC	18	+19.20 − 2.01A + 0.03B	Linear	0.9361	0.9308	0.99178	6.61	<0.0001
	Eef	19	+15.65 − 1.78A + 7.44B − 6.27AB + 0.19A2 + 0001B2	Quadratic	0.9406	0.9265	0.9007	7.17	<0.0001
	Color	20	+16.44 − 3.69A + 0.02B + 0.43A2	Quadratic	0.8826	0.8672	0.8373	5.21	<0.0001
	Shrinkage	21	+26.45 − 4.08A + 0.02B + 0.55A2	Quadratic	0.8785	0.8627	0.8328	3.46	<0.0001
BL ***	Time	22	+117.148 + 11.58A − 0.33B	Linear	0.9699	0.9674	0.9616	6.49	<0.0001
	Deff	23	−9.10 × 10−11 − 1.73A × 10−10 − 5.61 × 10−12B + 1.02 × 10−12 AB + 1.42 × 10−14B2	Quadratic	0.9838	0.9809	0.9749	8.95	<0.0001
	SEC	24	+14.74 − 2.05A + 0.03B	Linear	0.9298	0.9239	0.9113	7.17	<0.0001
	Eef	25	+14.43 − 1.43A + 0.01B − 0.008AB + 0.18A2 + 0001B2	Quadratic	0.9432	0.9297	0.9063	7.02	<0.0001
	Color	26	+18.38 − 3.00A + 0.01B + 0.002AB + 0.47A2	Quadratic	0.9130	0.8972	0.867	5.15	<0.0001
	Shrinkage	27	+24.80 − 2.86A + 0.02B + 0.44A2	Quadratic	0.8973	0.8839	0.8572	3.56	<0.0001

A: Thicknesses (mm), B: Microwave power (W), R²: determination coefficient and CV: Coefficient of variation. * E_ef = energy efficiency; ** US = ultrasonication; *** BL = blanching. Notice: Lack of fit for all conditions is not significant.

along with the statistical coefficients. Accordingly, the effect of MP and ST in different pretreatments is significant on D_t (p < 0.0001), while the interaction effect of temperature × moisture on D_t was not significant for any of the pretreatments. Equations (10), (16), and (22) in Table 2 show the model of the changes in carrot D_t in relation to the independent parameters of the microwave dryer for untreated samples, ultrasonicated, and blanched samples, respectively.

Changes in the D_t of sample slices for the control sample are depicted in Figure 2 for (a) controls, (b) ultrasonicated, and (c) blanched samples. The longest D_t (170 min) was observed for control samples with 6 mm thickness dried at the 100 W power, while the shortest D_t (42 min) was achieved at the MP of 300 W and 2 mm thickness in the samples pre-treated with blanching. According to Figure 2, the D_t has decreased linearly and significantly (p < 0.0001) by reducing the ST and increasing the MP. The microwave-induced power raises the temperature within the plant due to the collision of the ions placed in the electric field. The placement of water molecules along the field direction leads to friction, and hence, to heat production in the product. Subsequently, thermal energy is transferred to all different parts of the carrot through convection and conduction; therefore, the heat diffusion mechanism at higher MP_s will shorten the D_t [33]. Similar findings were published by other researchers for different products such as kiwi [34], dragon fruit [35], and mushrooms [36]. Moreover, thicker slices of carrot samples dried slower than thinner ones for all experimental treatments, indicating that thinner materials approached equilibrium MC in a shorter time. Compared to the thicker slices, the moisture can be easily removed from the surface of the thinner products [34]. A study on drying yam slices under a hot air dryer stated that the D_t decreases by reducing the ST [26].

Comparison of the D_t of ultrasound pretreated carrot slices with control samples shows that ultrasound pretreatment can accelerate the drying process. Ultrasound waves can open the micropores of carrot cells (due to the cavitation phenomenon), facilitating the moisture diffusion and significantly shortening the D_t. These results were confirmed for different products [37,38,39]. Abbaspour et al. [40] revealed a significant difference in the decrease in the D_t of ultrasound-treated terebinth dried by hot air and infrared dryers compared to untreated samples at all temperatures and powers, which is in agreement with the results of this study. According to Figure 2, blanching-pretreated samples had the shortest D_t. The reduction of D_t by blanching is probably attributable to the great damage of the cell structure and the reduction of the resistance of the cell wall and cell membrane against the movement of water [38,41]. Rybak et al. [12] reported that the D_t of blanched bell peppers was approximately 72% shorter than the samples pretreated with ultrasound and electric pulses.

3.2. Effective Moisture Diffusivity

Based on Table 2, the independent effect of the MP and ST, as well as the interaction effects of ST × MP, were significant on the D_eff for different pretreatments. The quadratic effect of the MP on the D_eff of blanched and ultrasonicated samples was significant (p < 0.0001), while it was not significant in the controls. Equations (11), (17), and (23) in Table 2 show the model of the D_eff variations in relation with the independent parameters. According to the model coefficients, the influence of the MP was higher than the effect of the ST on the D_eff value of the pretreated samples. Figure 3 also reveals the interaction effects of the MP and ST on the D_eff value of (a) untreated, (b) ultrasonicated, and (c) blanched carrots. The D_eff value showed a significant (p < 0.0001) rise by enhancing the MP and ST. In general, a higher MP leads to higher temperatures and, hence, lower water viscosity and higher water activities. This phenomenon can result in easier diffusion of water within the product [33]. The same observations were reported in drying lemon [29] and tomato [6] slices. The impact of the ST on moisture diffusion can be discussed by the surface hardening of the samples. Due to the higher rate of moisture evaporation during the drying process, the surface of the thin slices is generally hardened faster than the thicker ones, which will inhibit the diffusion of moisture from the center to the surface of the samples [42]. The results of the current research are well in line with the reports on aloe vera gel [26], apple slices [43], and potatoes [44]. Based on the results, the use of ultrasonication and blanching pretreatments can quadratically increase the value of the D_eff compared to the control samples. The influence of the blanching was more profound on the increase in the D_eff as compared with ultrasonication. These results are consistent with the drying kinetic data, which imply that the stronger the mobility of internal water molecules, the shorter the D_t [45]. The formation of more pores and microchannels due to the ultrasound-induced cavitation and more enzyme degradation and tissue loosening due to blanching led to higher permeability [14]. The results of this research are in agreement with the researchers from other reports for cranberry snacks [46] and pumpkin products [47] that were pre-treated with ultrasound and blanching. Furthermore, when ultrasound waves pass through an environment, they cause successive expansion and contractions, leading to the cavitation phenomenon [39,48]. As a result, ultrasonic waves increased the D_eff due to the small channels created inside the food which facilitated the moisture removal [49].

3.3. SEC

Table 2 lists the results of the influence of different parameters (MP and ST) in different pretreatments on the SEC of the microwave drying process. As can be seen, the MP and the ST have a significant effect (p < 0.0001) on the SEC value of carrot slices, while the interaction effect of the MP and the ST had no significant effect on the SEC value. Figure 4 also depicts the variations of the SEC for (a) control, (b) ultrasonicated, and (c) blanched samples. The value of the SEC decreases linearly or quadratically by increasing the MP and decreasing the ST. The use of the MP can accelerate the drying due to changes in the cell structure and faster moisture diffusion. Such a reduction in the D_t can decrement the SEC. In other words, the thermal gradient and the acceleration of moisture extraction from the samples are improved by increasing the MP [50]. Additionally, in thicker samples, the transfer of product moisture to the surface for final evaporation is associated with a slower movement, hence, the D_t is prolonged, leading to a higher SEC [51]. Similar findings were shown for drying potato [8], banana [52], and kiwi [34] slices.

According to Figure 4, the application of a pretreatment reduced energy consumption as compared to the untreated samples. It is worth mentioning that sonication (damaging carrot tissue due to creation of microscopic channels and cavitation) before drying is effective in reducing the D_t and the SEC [37]. Additionally, the use of the blanching pretreatment had a greater effect on the SEC reduction compared to the ultrasound pretreatment. This phenomenon may be assigned to the pressure difference between the surrounding environment and the sample tissue during the blanching stage. The blanching process leads to the formation of a porous structure, resulting in a better and faster moisture transfer during the microwave drying [53]. The highest SEC value (28.06 MJ/kg) was achieved at the MP of 100 W for control samples with the thickness of 6 mm, while the lowest SEC value (9.51 MJ/kg) was observed at a power of 300 W in the blanching-pretreated samples with the thickness of 2 mm. Similar findings were reported on the reduction of the SEC by using blanching and ultrasound pretreatment for different products, such as mushroom slices [15], parsley leaves [37], and blackberry [54].

3.4. Energy Efficiency (E_ef)

As suggested by Table 2, the influence of the drying parameters (MP and ST) and the interaction effect of MP–ST were significant (p < 0.00001) on the E_ef of the samples pretreated by blanching and ultrasound. Figure 5 also represents the interaction effect of MP–ST on the E_ef of (a) controls, (b) ultrasonicated, and (c) blanched carrots. As observed, the E_ef rose linearly or quadratically by increasing the MP and ST. The highest E_ef (22.39%) was achieved at the power of 300 W in the ultrasonicated samples with 2 mm thickness; the lowest E_ef (7.78%) was also related to the power of 100 W and 6 mm thickness in the control samples. The simultaneous use of higher MP and thinner samples led to a faster removal of the surface moisture. A rise in the MC gradient also increased the microcracks within the samples. These cracks caused a volume change, further accelerating the removal of moisture and declining the D_t, which in turn enhanced the E_ef [6]. The increment in the E_ef by raising the MP and reducing the ST was reported before [8,55]. On the other hand, a comparison of the various pretreatments showed that the highest E_ef was achieved in the blanching pretreatments while the untreated samples exhibited the lowest E_ef. Pretreatments improved the E_ef as they increased the destruction and accelerated the removal of moisture. These results coincided with the previous reports [18,28].

3.5. GHG

Table 3. Fitting the effect of different levels of microwave power and sample thickness on independent parameters in a microwave dryer without a pretreatment.

Pretreatment	Power Plant—Type of Fuel	Equation for CO2	Equation for NOx	p-Value
Control	SP-NG	14,074.16 + 2625.24A − 2363.16B − 682.08B2	65.5 + 12.22A − 11B − 3.17B2	<0.0001
	SP-HO	22,682.41 + 4230.94A − 3805.56B − 1099.27B2	55.77+ 10.4A − 9.36B − 2.70B2	<0.0001
	GT-NG	17,305.02 + 3227.9A − 2905.65B − 838.66B2	42.27 + 7.88A − 7.1B − 2.05B2	<0.0001
	GT-GO	23,191.38 + 4325.88A − 3894.02B − 1123.94B2	128.13 + 23.9A − 21.51B − 6.21B2	<0.0001
	CC-NG	9958.13 + 1857.48A − 1672.05B − 482.61B2	65.28 + 12.18A − 10.96B − 3.16B2	<0.0001
	CC-GO	13,764.35 + 2567.46A − 2311.14B − 667.07B2	83.65 + 15.60A − 14.05B − 4.05B2	<0.0001
US *	SP-NG	12,304.77 + 2563.27A − 2518.20B	56.27 + 11.93A − 11.72B	<0.0001
	SP-HO	19,830.81 + 4131.06A − 4058.43B	48.75 + 10.16A − 9.98B	<0.0001
	GT-NG	15,129.45 + 3151.7A − 3096.28B	+36.95 + 7.70A − 7.56B	<0.0001
	GT-GO	20,275.79 + 4223.76A − 4149.49B	112.02 + 23.34A − 22.93B	<0.0001
	CC-NG	8706.21 + 1813.64A − 1781.75B	57.07 + 11.89A − 11.68B	<0.0001
	CC-TGO	12,033.91 + 2506.85A − 2462.77B	73.13 + 15.23A − 14.97B	<0.0001
BL **	SP-NG	10,762.27 + 2613.9A − 1921.75B	50.09 + 12.17A − 8.94B	<0.0001
	SP-HO	17,344.85 + 4212.66A − 3097.16B	42.64 + 10.36A − 7.61B	<0.0001
	GT-NG	13,232.85 + 3213.95A − 2362.91B	32.32 + 7.85A − 5.77B	<0.0001
	GT-GO	17,734.05 + 4307.19A − 3166.66B	97.98 + 23.80A − 17.50B	<0.0001
	CC-NG	7614.81 + 1849.46A − 1359.73B	49.92 + 12.12A − 8.91B	<0.0001
	CC-TGO	10,525.36 + 255,636A − 1879.45B	63.96 + 15.54A − 11.42B	<0.0001

A: Thicknesses (mm), B: Microwave power (W), R²: determination coefficient and CV: coefficient of variation. * US = ultrasonication; ** BL = blanching.

lists the results of GHG modeling including CO₂ and NO_x for various MP and ST, as well as ultrasonication and blanching pretreatments using electricity from plants operating on diverse fuels. The R² values were above 0.96. According to Table 3, the influence of the independent parameters was significant (p < 0.0001) on the emission of CO₂ and NO_x for all plants and their corresponding fuels. The second power of the independent parameters was only significant (p < 0.0001) on the emission of CO₂ and NO_x due to drying the control samples. Figure 6 and Figure 7, respectively, represent the interaction influences of MP–ST on the response variables of CO₂ emission from a GT-GO plant and NO_x emission of a CC-NG plant obtained based on the response surface diagrams for untreated and ultrasonicated and blanched samples. The highest CO₂ and NO_x emissions (29,414.5 and 162.5 g, respectively) were observed in the untreated samples with a thickness of 6 mm at the MP of 100 W dried by the electricity supplied from a GT-GO plant. However, the lowest CO₂ (4279.7 g) was from a CC-NG plant in the drying of the blanched samples with the thickness of 2 mm at the MP of 300 W, while the lowest NOx emission (18.16 g) was related to the blanched samples with a thickness of 2 mm which were dried at the MP of 300 W of which electricity supply was a GT-NG plant. Regarding the obtained results, a rise in the MP and a decline in the ST decreased the GHG emission. Such a descending trend in the GHG emission can be assigned to the reduction in the SEC during drying at a high MP and thin thicknesses [56]. These findings are in line with the previous reports on the reduction of the GHG emission by increasing the MP for drying green pea [57], aloe vera [58], and Pistacia Atlantica [20].

The results suggested a significant reduction in GHG production upon a pretreatment of the samples. The reason could be the prevention from the formation of the hard layer upon a pretreatment, which can help in the fast removal of moisture, hence, declining the drying duration and the energy required for the drying process. Upon a decrement in the energy consumption, the amount of the required fuel and hence the GHG emission due to the combustion of the fossil fuels in the plants will be decremented [59]. In line with this research, the results of other studies on drying blackberries [60] and goldenberries [21] also indicated a decline in GHG emission upon a pretreatment of the samples.

3.6. Color Changes

One of the prominent parameters in the quality of the dried agricultural products is color. Based on Table 2, the influence of the independent parameters of the MP and ST, as well as the quadratic effect of the thickness, were significant (p < 0.0001) on the color changes of the carrot slices. For blanched samples, the interaction effect of the MP–ST was also significant (p < 0.0001) on the color changes of the samples. Figure 8 presents the interaction effects of the MP and ST on the color changes during the process of the carrot sample drying. Accordingly, the color variation range of the untreated samples was 19.29–28.68 which altered to 13.69–21.22 and 15.59–25.94 in the samples pretreated by ultrasound waves and blanching, respectively. The lowest color change (13.69) was observed in the ultrasonicated samples with a thickness of 4 mm dried at the MP of 100 W, whereas the controls with thickness of 6 mm dried at the MP of 300 W experienced the highest color changes. At constant thickness, an increase in the MP increased the difference between the color of the dried and fresh samples in all pretreatments. This observation can be assigned to oxidation, non-enzymic browning reactions, and changes in the surface structure of the carrot samples during the drying process [18]. In drying red bell pepper [61] and okra pods [62], researchers found similar trends in which the increasing in the MP enhanced the color difference. As shown in Figure 8, the color change declined by increasing the ST from 2 to 4 mm in the untreated and sonicated samples, whereas an increase in the ST from 4 to 6 mm led to an ascending trend in the color change. In the case of the blanching pretreatment, a rise in the ST from 2 to 6 mm increased the color changes which was in line with the findings of other researchers who dried pumpkin [7] and potato [63]. For all the samples, big color changes were observed in the thickness of 6 mm as these samples required a longer time for drying. Therefore, they were exposed to oxidation and thermal degradation for a longer period of time, which enhanced the color changes in the final product [7].

After the ultrasonication and blanching pretreatments, the range of the color changes was maintained at an acceptable level. In other words, the application of the pretreatments better preserved the color features of the samples, suggesting that these pretreatments managed to prevent the browning of the carrot slices. High-temperature blanching also deactivates the enzymes such as polyphenol oxidase and peroxidase, hence decrementing the enzymic browning reactions during the drying process [32,64]. On the other hand, ultrasonication decreased the oxygen content of the carrot samples, hence inhibiting the browning reaction [47]. This is why the least color changes occurred in the samples pretreated by ultrasound waves. The results of this research on the effect of the ultrasound waves and blanching on a better color preservation are in line with previous studies such as studies on drying apples in a hot-air dryer (Bai et al., 2023 [32]), pumpkin samples dried in an IR dryer [47], and cranberry snacks dried in a microwave-vacuum dryer [65].

3.7. Shrinkage

As listed in Table 2, the influences of the MP and ST, as well as the quadratic effect of sample thickness, were significant (p < 0.0001) on the shrinkage of the dried samples. The interaction effect of the MP–ST was not significant. Based on Figure 9, the highest shrinkage (33.39%) was recorded in the untreated samples with 6 mm thickness dried at the MP of 300 W, while the ultrasonicated samples with a thickness of 4 mm dried at the MP of 100 W showed the lowest shrinkage (21.29%); as expected, a rise in the MP increased the shrinkage due to the contraction stress at higher powers, which led to the cellular collapse (deformation of the product and its cracking), and these results are in line with previous works [11,66]. An increase in the ST from 2 to 4 mm decremented the shrinkage followed by an increment in the shrinkage by further thickness enhancement from 4 to 6 mm. The highest shrinkage was generally observed in the thickest samples (6 mm) for all treatments. Drying will result in deformation and volume reduction, hence the shrinkage of the product. Heating and water removal from the matrix of the food products will lead to cellular stress, decrementing the dimensions of the product and contraction of its capillary structure. At higher thicknesses, these changes are more profound [67]. These findings coincided with the reports by other researchers on drying apple [68] and Persian shallot [69] of different thicknesses.

According to Figure 9, the use of an ultrasonic pretreatment and blanching led to a lower shrinkage compared to the untreated samples. The carrot slices pre-treated with ultrasound waves showed the lowest shrinkage at all MP_s and ST followed by the blanched samples and controls, respectively. Tissue hydration and damage during pretreatment increased the MC of the blanched samples which led to more shrinkage in the blanched samples than in sonicated samples [70].

3.8. Optimization

A vital step in food process engineering to achieve an effective performance in different operations and high-quality food products is optimization. Selecting the proper drying parameters to achieve the desired response surfaces is the key challenge in the optimization process using RSM and its desirability function [61]. In this study, the MP and the ST were selected. The results of optimal drying conditions for carrots are listed in

Table 4. Results of optimization using a desirability method.

Pretreatment	Parameters	Optimal Point	Power Plant—Type of Fuel	Optimal Point for CO2	Optimal Point for NOx	Desirability
Control	Time	48.96	SP-NG	8630.0	40.16	0.776
	Deff	5.087 × 10−10	SP-HO	13,908.6	34.19
	SEC	13.22	GT-NG	10,611.2	25.91
	Eef	16.92	GT-GO	14,220.6	78.56
	Color	24.54	CC-NG	6106.1	40.02
	Shrinkage	31.48	CC-TGO	8440.1	51.29
Ultrasound	Time	44.46	SP-NG	7218.9	33.59	0.807
	Deff	5.90 × 10−10	SP-HO	11,634.3	28.60
	SEC	11.35	GT-NG	8876.1	21.67
	EE	19.42	GT-GO	11,895.4	65.71
	Color	19.36	CC-NG	5107.7	33.48
	Shrinkage	26.95	CC-TGO	7060.0	42.90
Blanching	Time	39.31	SP-NG	6225.5	28.97	0.82
	Deff	6.50 × 10−10	SP-HO	10,033.3	24.66
	SEC	9.79	GT-NG	7654.6	18.69
	Eef	21.98	GT-GO	10,258.4	56.67
	Color	19.08	CC-NG	4404.8	28.87
	Shrinkage	29.13	CC-TGO	6088.5	37

. The desirability obtained at optimal conditions for the investigated variables and responses were 0.776, 0.807, and 0.82 for untreated, ultrasonicated, and blanched samples, respectively. According to Table 4, the optimal conditions for carrot slices dried by the microwave method are as follows.

For samples without pretreatment, the MP (300 W) and the ST (2 mm) resulted in the following response values: D_t (48.96 min), SEC (13.22 MJ/kg)), E_ef (16.92%), D_eff (5.087 × 10⁻¹⁰), CO₂ (14,220.6 g, in the GT-GO power plant), NO_x (40.02 g, in the CC-NG power plant), color (24.54), and shrinkage (31.48%).

In the case of samples pretreated with ultrasound waves, the MP of 300 W and the ST of 2 mm led to the following response values: D_t (44.46) min, SEC (11.35 MJ/kg)), E_ef (19.42%), D_eff (5.90 × 10⁻¹⁰), CO₂ (11,895.4 g, in the GT-GO power plant), NO_x (33.48 g; in the CC-NG power plant), color (19.36), and shrinkage (26.95%).

For the samples with blanching pretreatment, the MP of 300 W and the ST of 2 mm resulted in the following response values: D_t (39.31 min), SEC (9.79 MJ/kg), D_eff (6.5 × 10⁻¹⁰), CO₂ (10,258.4 g; in a GT-GO plat), NO_x (28.87 g; in a CC-NG plant), color (19.08), and shrinkage (29.13%).

4. Conclusions

In this research, microwave drying was utilized to dry carrot slices of different thicknesses. The RSM was employed to explore the effects of the MP (100, 200, and 300 W) and the ST (2, 4, and 6 mm) on D_t, D_eff, SEC, E_ef, GHG (CO₂ and NO_x), and qualitative characteristics (color and shrinkage) of carrot slices considering the ultrasonication and blanching pretreatments. The results showed a reduction in the D_t (linearly), SEC (linearly or quadratically), and GHG (linearly and quadratically) by increasing the MP and decreasing the ST, while causing an ascending trend in the E_ef (linearly or quadratically) and the D_eff (linearly). A statistical and graphical analysis indicated a significant effect of drying parameters (MP and ST) on the response parameters of drying carrot slices in all pretreatments (p < 0.0001). The optimization results showed that the MP of 300 W and the ST of 2 mm are the best independent variables for drying carrot slices. Different pretreatments reduced the GHG emission, while the lowest CO₂ (4279.7 g in a CC power plant) and NO_x (18.16 g in a GT power plant) emissions with NG were obtained for the blanched samples with a ST of 2 mm dried at a MP of 300 W. Regarding the serious global warming crisis, knowing that the increasing population results in an ever-increasing demand for agricultural products, and considering the high energy consumption in the drying process of agricultural products, the application of different pretreatments seems to be a practical and important approach to reduce the SEC and GHG emission, hence reducing climate change.