Optimization of grape juice deacidification using mixture of adsorbents: A case study of Pekmez

Abstract Grape syrup (Pekmez or Dooshab) is one of the nutritious products developed through grape processing. One of the main challenges in the industrial manufacture of this product is the utilization of traditional pekmez earth for tartaric acid adsorption. The objectives of this study were to investigate the effect of calcium carbonate, nano‐silica, alumina, and activated carbon as adsorbents and also contact time in grape juice deacidification, and to determine the effects of these adsorbents on the physicochemical properties of grape juice by using the Box–Behnken statistical design. By applying different amounts of these adsorbents in grape juice, the magnitude of acidity decrement and the physicochemical properties such as acidity, pH, transmittance, the amount of reducing sugars, formalin index, and adsorption efficiency were investigated. Data analysis showed that different mixtures of adsorbents at different concentrations had significant effects on acidity and pH of the samples but no effects on the level of reducing sugars and formalin index were observed (p > .05). According to the results, the adsorption capacity with the highest calcium carbonate content (0.7 g/100 ml) was about 88%; the maximum acidity decrements of up to 92% were achieved using the treatments containing calcium carbonate, nano‐silica, and activated carbon, while alumina failed to affect the acidity of the samples. Optimum conditions were obtained in 1.27, 0.21, 0.7, and 0.07 g/100 ml alumina, nano‐silica, calcium carbonate, and activated carbon, respectively, resulting pH 4.3 and acidity 0.37% in grape syrup.

is valuable in situations demanding urgent energy, and it could play an important role in the diet of different age groups, especially children and athletes (Akbulut & Ozcan, 2009;Heshmati, Ghadimi, Ranjbar, & Khaneghah, 2019;Heshmati, Ghadimi, Ranjbar, & Mousavi Khaneghah, 2020). Grapes and grape products contain a wide range of polyphenolic constituents that have been reported to show anticancer and anti-inflammatory effects in vitro, as well as the ability to block cellular events that lead to atherosclerosis and coronary heart disease. The compounds presumed to provide these positive health effects are mainly flavonols, procyanidins, anthocyanins, and phenolic acids (Capanoglu, de Vos, Hall, Boyacioglu, & Beekwilder, 2013). Grape syrup, as one of the most popular traditional food products in eastern culture (Heshmati et al., 2019), is a concentrated and shelf life-extended form of grape juice formed by boiling without the addition of sugar or other food additives after the acids are removed (Kaya & Belibagli, 2002).
Flow diagram of a typical grape syrup processing operation is presented in the Scheme 1. During the processing of grape syrup, a calcareous soil called "grape syrup earth," containing approximately 90% calcium carbonate is added to grape juice. The grape syrup earth lowers the acidity caused by naturally existing tartaric and malic acids by precipitating them as calcium tartrate and calcium malate. Grape juice is concentrated usually in open vessels and rarely under vacuum to obtain 65-68° Brix; this product is called liquid grape syrup (Kaya & Belibagli, 2002). It is also known by the names of "Pekmez" and "Petimezi" in Turkey and Greece, respectively. The most important problem in grape syrup production is grape syrup earth addition, as an acid adsorbent, into the juice. Due to the presence of heavy metals in soil composition containing different calcium purities, this can be very harmful to the human body and cause kidney or liver disease, etc (Jarup, 2003). Also, soil usage in grape syrup production causes other problems, such as high consumption of soil, high waste production, high costs, limitation of daily production, and low performance of the production line. In recent years, the usage of nanoparticles and chemical adsorbents has been studied for the adsorption studies.
Calcium carbonate (CaCO 3 ) with a pH of 9 is a highly adsorbent material used in various industries, especially for the neutralization of juices. It combines with tartaric acid and malic acid to form calcium tartrate crystals and calcium malate (Kaiser, 2012). Silica (SiO 2 ) has adsorbent hydroxyl groups and a large atomic structure.
It is one of the most important adsorbents for the separation of materials in chemical technologies and chromatography. Factors influential on the ability of silica to adsorb include: pH, temperature, and ionic strength; silica is apparently more efficient for the adsorption of cations relative to anions (Papirer, 2000). Alumina (Al 2 O 3 ) is highly resistant to various chemical environments and can withstand extreme stress and maintain its properties until up to 1,500°C (Uslu & Inci, 2009). The utilization of activated carbon is a widely used technique for reducing organic pollutants from water and is most used in the food industry for bleaching and removing heavy metals or phenolic compounds (Farouk Mohamed, 2011).
In recent years, many similar researches about tartaric acid adsorption have been done. A study by Kontogiannopoulos, Patsios, and Karabelas (2016) showed that tartaric acid recovery from winery lees using Lewatit ® MonoPlus S108 as a strongly acidic cation exchange resin at room temperature is possible. In that study, the pH, water dosage, and cation exchange resin dosage were the three variables that were investigated (Kontogiannopoulos et al., 2016).
Another study was conducted by Shayanfar and Bodbodak (2014) on the effect of different physicochemical detartration methods on red grape juice quality. In that study, the application of cold treatment, tartar cream (potassium bitartrate), carboxymethyl cellulose (CMC), and mannoprotein were investigated. Then, all samples were stored at 5°C to stimulate tartrate crystal precipitation and were investigated for changes in color indices, sediment levels, acidity, pH, and taste. Among these methods, the addition of mannoprotein and CMC proved to have the ability to stabilize tartrate crystals in red grape juice (Shayanfar & Bodbodak, 2014). The adsorption equilibria of L-(+)-tartaric acid onto alumina from wastewaters of wineries were studied by Uslu and Inci (2009). In that study, the adsorption experiments were carried out at three different temperatures (298, 310, and 325 K). The adsorption of L-(+)-tartaric acid was found to be dependent on the acid concentration and the amount of alumina.
The maximum percentage removal of L-(+)-tartaric acid was 22% using alumina at 298 K.
The aim of this research was to find the best mixture of nano-silica, calcium carbonate, alumina, and activated carbon as acid adsorbents for grape syrup production. In this research, the pH, titratable acidity and deacidification rate, transmittance, turbidity, Brix, reducing sugars content, polyphenol content, and formalin index of grape juice samples introduced with mixtures of adsorbents were investigated and the proper mixture of adsorbents was obtained through statistical optimization.

S C H E M E 1
Production diagram of grape syrup from fresh grapes (Kaya & Belibagli, 2002).

| Materials
Grapes (Vitis vinifera "Trebbiano") were obtained from a local grape farm (Urmia, Iran) as the raw materials. Nano-silica was prepared by Pioneer Iranian Nano Materials Company. Calcium carbonate, alumina, activated carbon, and other laboratory compounds and solutions were obtained from Merck.

| Statistical design and analysis
The present study was conducted in two steps, through two experimental designs. At the first stage, a mixture experimental design was applied to investigate the effect of four components (nanosilica: 0%-20%; calcium carbonate: 0%-100%; alumina: 0%-100%; and activated carbon: 0%-10%), with total amount of 0.7 g/100 ml, on physicochemical properties of grape juice which was carried out across 24 treatments in four blocks. According to the results of the first step, the effects of two variables, namely calcium carbonate and activated carbon, were highly significant. Due to the high uptake of the acid by these two adsorbents and thus their great impact, the second step was performed using a Box-Behnken experimental design to investigate the effect of three factors, nano-silica amount, alumina amount, and contact time ( Table 1). The calcium carbonate and activated carbon amount were constant in all treatments (0.7 and 0.07 g/100 ml, respectively). The number of treatments was 16; these treatments were tested, and the data analyzed and optimized against a control sample (grape juice). Analysis of variance (ANOVA) at 95% confidence level (p < .05) was performed, and Design-Expert 7 software was used for analyzing data, optimizing, and plotting curves. Data were fitted to second-order polynomial Equation 1 for each dependent Y variable, through a stepwise multiple regression analysis.

| Preparation of grape juice
Firstly, all grapes were washed and clustered. The grapes were then pressed with a juicer and filtered. The volume of each sample was 400 ml to begin the test. In the next step, calcium carbonate, nano-silica, alumina, and activated carbon adsorbents were added into the samples according to the experimental design ( Table 1).
The amounts of calcium carbonate and activated carbon were kept constant at 0.7 and 0.07 g/100 ml, respectively. According to the experimental design, contact times ranging from 30 to 240 min were applied. After data collection, optimization tests were performed. The pH and acidity of the optimum treatment, the stock solution of tartaric acid (0.6% m/w), and the control sample were investigated. Also, precipitates of chemical adsorbents inside the samples were separated by filter paper and subjected to structural testing by FT-IR.

| pH
In order to determine the pH of the samples, the electrode of the AZ 86502 pH meter was put in contact with the samples before the pH values were recorded (Bruijn, Venegas, Martinez, & Borques, 2003).

| Acidity
This test was performed using the potentiometric method. In brief, 50 ml of distilled water was mixed with 20 g of grape juice. Then, 0.1 N sodium hydroxide solution was added drop-wise to bring the pH of the juice to 8.1; the volume of sodium hydroxide consumed was recorded, and the acidity of the samples was obtained by the following equation: where A is the total acidity in terms of tartaric acid in grams per 100 ml; V is the volume of sodium hydroxide consumed (0.1 normal per ml), and m is the sample weight in grams (Hashim, Zubairi, Wan Mustapha, & Masket, 2018).

| Deacidification rate
The rate of acid uptake by sorbents from grape juice samples was calculated using the following equation: where C i is the initial acidity, C f is the final acidity, and A is the acidity uptake percentage (Ghasempour, Alizadeh-Khaledabad, Vardast, & Bari, 2016).

| Turbidity
The Hach 2100AN Laboratory Turbidimeter was used to determine the turbidity of the samples after filtering (Bruijn et al., 2003).

| Transmittance
In order to determine the transparency of the samples after filtering, a US-based UV2100 spectrophotometer was employed at a wavelength of 625 nm. Distilled water was used as the blank (Bruijn et al., 2003).

| Reducing sugars
The content of reducing sugars was estimated using the dinitrosalicyclic acid (DNS) method. The DNS method is a colorimetric technique that consists of a redox reaction between 3,5-dinitrosalicyclic acid and the reducing sugars present in a sample. The reducing power of these sugars comes from their carbonyl group (C=O), which can be oxidized to the carboxyl group by mild oxidizing agents, while the DNS (yellow) is reduced to 3-amino-5-nitrosalicylic acid (redbrown) that can be quantified by spectrophotometry at 540 nm. The intensity of the color is proportional to the concentration of sugars.

| Polyphenol content
In order to measure the total polyphenol content, the Folin-Ciocalteu reagent method was used. In this experiment, the absorbance of the samples was measured at 765 nm. The phenolic compounds of the samples were calculated using the standard gallic acid curve. The amount of total phenolic compounds was expressed in milligrams of gallic acid equivalents per gram of dry sample. In this test, 200 µl of Folin reagent (10 times diluted), 1 ml of saturated sodium carbonate, and 1 ml of grape juice were mixed and stirred for all treatments before being incubated in the dark for half an hour. The absorbance of the samples was then evaluated. It should be noted that the blank sample employed contained the same amount of Folin reagent and sodium carbonate but 1 ml of distilled water instead of grape juice (Liu et al., 2019).

| Formalin index
The first step of the procedure employed to evaluate formalin index included the addition of 25 ml of juice to a beaker. While stirring was performed, the juice was titrated drop-wise to pH = 8.1 with sodium hydroxide. Then, 10 ml of neutral formaldehyde solution was added before the mixture was stirred for 1 min. Next, it was titrated with 0.1 N sodium hydroxide to pH = 8.1. The volume of sodium hydroxide consumed was recorded, and the formalin index was obtained using the following equation (Garcia, Barros, Fidalgo, & Ilharco, 2007).
where F denotes the formalin index, V is the volume of sodium hydroxide consumed (0.1 normal per ml), N is the normality of sodium hydroxide, and V 0 is the sample volume in ml.

| Brix
This test was performed using the JK-ARM refractometer. The device was first set to zero with distilled water. Then, droplets of grape juice (20°C) were placed on a refractometer prism graded according to sucrose. The results were expressed in grams per gram of the samples (Liu et al., 2019).

| Optimization
After analysis of the physicochemical properties, optimization was performed on the 16 samples related to the Box-Behnken experimental design (Table 1) using the response surface statistical method with the Design-Expert 7 software. This optimization was done with the goal of achieving minimal acidity and turbidity rate but maximal acid adsorption rate, sample transmittance, reducing sugars content, and phenolic compounds content. After optimization, pH, acidity, and FT-IR tests were performed on the optimum sample.
Also, the stock solution of tartaric acid with acid concentration of 0.6 g/100 ml was prepared based on the total acidity of grape juice, and the adsorbents were added to the stock solution at optimum amounts. The pH, acidity, and FT-IR tests were also performed on the filtrate of adsorbents separated from the stock solution as well those separated from the grape juice samples. (

| FT-IR spectroscopy
FT-IR spectroscopy of the chemical adsorbents was performed before and after contact with grape juice (Spectrum Two, Perkin Elmer). The chemical adsorbents were separated and filtered using filter paper before being placed in an oven set at 70°C. After drying the adsorbents, the optimum treatment of grape juice, the adsorbents of the stock solution, and the chemical adsorbents before contact were analyzed with FT-IR. FT-IR spectroscopy was run at a wavenumber ranging from 400 to 4,000 cm −1 with a resolution of 0.5 cm −1 (Ghasempour, Alizadeh-Khaledabad, Vardast, & Bari, 2018).

| pH
According to the primary studies, the pH of the blank sample of grape juice was 3.84 before contact with any chemical adsorbent, improving to 6.51 after contact with calcium carbonate, nano-silica, and activated carbon adsorbents; here, calcium carbonate was the most effective adsorbent. The findings of the second part of study showed that adding nano-silica and alumina to samples across different contact times (30-240 min) had no significant effect on the F I G U R E 1 Contour plot curve of interaction effect of nano-silica and alumina contents on acidity (a) and acidity adsorption (b) pH of the examined grape juice. The pH remained at 4.5 in all treatments, which may be related to the presence of a constant amount of calcium carbonate within all treatments or to the existence of buffer substances such as proteins in the grape juice. Our findings are in accordance with those of the study of Kaiser (2012) in that the increase in pH was in line with the increase of calcium carbonate adsorbent content.

| Acidity and acid adsorption
In total, the acidity of the grapes was related to approximately 0.3% of grape juice content, mostly being due to tartaric acid (about 90%), with small amounts of malic and citric acid also being present.
In the primary experiments, this amount decreased to 0.03% in the presence of a mixture of calcium carbonate, nano-silica, alumina, and activated carbon adsorbents at levels of 0.74, 0.14, 0.02, and 0.1 g per 100 ml, respectively. In other words, the acidity was reduced by 10%. Figure 1a demonstrates the synergistic influence of nano-silica and alumina on grape juice acidity. The findings prove that the grape juice acidity was reduced by increasing the contents of both nano-silica and alumina. In fact, by increasing the nanosilica level from 0.2 to 0.3 g per 100 ml, the acidity was gradually reduced by this adsorbent. The maximum amount of acid reduction was observed by adding 0.2 g of nano-silica and 1.5 g of alumina; the interaction effect of nano-silica and alumina on grape juice acidity reduction was highly significant (p < .05). However, the contact time had no significant effect on acidity (p > .05).
The amount of acid adsorbed by the adsorbents was measured through the acidity index of the samples, which was obtained using Equation 3; the total acid (mostly comprised of tartaric acid) adsorbed using nano-silica and alumina adsorbents was thus determined. Figure 1b shows the effect of nano-silica and alumina on grape juice acidity adsorption. The results show that the grape juice acidity adsorption increased with rises in both nano-silica and alumina contents. It should be noticed that the slope related to alumina shows considerable differences compared to the nano-silica adsorbent in acidity reduction. In addition, by increasing the nano-silica from 0.2 to 0.3 g per 100 ml, acid adsorption was gradually increased by this adsorbent. In a research conducted by Uslu and Inci (2009)

| Turbidity
Turbidity is an indicator that shows the amount of light transmission through fluids and liquids. This indicator expresses the turbidity level of food samples in nephelometric turbidity units (NTU). The reason for turbidity is the existence of some suspended particles or insoluble solids such as food colloids, proteins, and phenols. Figure 2 demonstrates the interaction effect of nano-silica and alumina contents on grape juice turbidity. The results show that the grape juice turbidity considerably decreased with increasing nano-silica and decreasing alumina contents. Therefore, the effect of nano-silica and alumina on the turbidity factor of grape juice samples was highly significant (p < .05), but contact time had no significant effect (p > .05).
According to the findings of Aljohani et al. (2018), it seems that the reason for the reduction in grape juice turbidity with increased nanosilica content is the adsorption of proteins and phenolic compounds by silica (Aljohani et al., 2018).

| Reducing sugars
The obtained results clarified that the effects of alumina adsorbent and contact time variables as well as their interaction on the content of reducing sugars were significant in grape juice samples (p < .05).
The results also showed that with a simultaneous rise in the time

| Total phenolic compounds (TPC)
Nowadays, it is known that phenolic compounds have protective effects on the antioxidative capacity of biological systems due to reducing hydrogen peroxidation, activating antioxidant enzymes, reducing α tocopherol radicals, and inhibiting oxidases. Besides, phenolics are suitable antimicrobial, antiviral, and anticancer compounds. Several studies have reported many polyphenols found in grapes such as flavan-3-ols, quercetin, and anthocyanins (Pasini, Riciputi, Fiorini, & Caboni, 2019). However, removal of phenolics and proteins during grape juice production is essential, otherwise they cause the hazing of grape products, which is maintained by their interactions (Kulcan, Öziyci, Tetik, & Karhan, 2015). In this experiment, the changes occurring in the TPC amount of the samples was investigated. The findings reveal that the minimum amount of

| Formalin index
Formalin index is known as the indicator of amino acid presence in fruit juice products. Figure 3a shows that the effect of the two vari-

| Optimization
Considering all results, the optimization was done in two steps.
Firstly, the primary test was performed based on the desired objectives. Furthermore, based on the acidity of nontreated grape juice, a solution of tartaric acid was obtained at the same amount of acidity (0.6 g per 100 ml). In the second step, by selecting optimum amounts of adsorbents, the effect of the optimum treatment on pH, acidity, and FT-IR spectroscopy was studied on grape juice and tartaric acid solution. The aim of this procedure was to determine the optimum mixture of applied adsorbents in order to decrease tartaric acid adsorption and allow a minimal level of adsorption of nutritional compounds such as phenolic compounds. Table 2 shows the conditions of the optimized treatment.

| FT-IR spectroscopy
FT-IR spectroscopy is applied as an appropriate approach to determine the functional groups and or chemical bonds of a material.

| pH and acidity
The pH of the stock solution and grape juice samples was evaluated before and after contact with the optimum mixture of TA B L E 2 Optimized treatment conditions with the goal of achieving minimal acidity and turbidity rate but maximal acid adsorption rate, sample transmittance, reducing sugars content, and phenolic compounds content in grape juice deacidification with adsorbents adsorbents. The pH of the stock solution was initially 2.5, which increased to 3.39 after contact with optimal adsorbents. Also, the pH of grape juice was initially 3.2, which increased to 4.6 after contact with optimal adsorbents. According to the obtained results, the amount of tartaric acid in the stock solution was 0.675 g per 100 ml, which decreased to 0.154 g per 100 ml after contact with optimal adsorbents. Additionally, the total acidity of the control grape juice sample attributable to tartaric acid was 0.615 g per 100 ml, which decreased to 0.176 g per 100 ml after contact with optimal adsorbents.

| CON CLUS ION
The objective of this research was to separate tartaric acid from grape juice by usage of a novel adsorbent mixture of nano-silica (nanoparticles of SiO 2 ), calcium carbonate (CaCO 3 ), alumina (Al 2 O 3 ), and activated carbon in order to produce grape syrup, which was characterized during contact time. The applied adsorbents were studied in order to obtain low-cost and safe adsorbents with high efficacy for the replacement of traditional methods with a glance of mass industrial production. The calcium carbonate (0.7%) and activated carbon (0.07%) contents were kept constant across all treatments, while nano-silica and alumina adsorbents were varied in content and examined in the contact time of 30-240 min. The evaluated parameters were pH, total acidity, reducing sugars, transmittance, total phenol compounds, turbidity, formalin index, and Brix. Also, after optimization of the pH value, the acidity, and FT-IR spectra were determined for the optimal adsorbents after contact with grape juice and tartaric acid solutions. According to the results of this study, the effects of nano-silica and alumina on Brix and pH were not significant (p > .05). In addition, an increase in the amount of nano-silica adsorbent had a significant effect on the increase of transmittance and decrease of turbidity in the samples in contact with chemical adsorbents. The results show that an optimal adsorption rate of 60% was achieved for the adsorption of total grape juice acidity through 240 min of contact with 0.21 g/100 ml of nano-silica, 1.27 g/100 ml of alumina, 0.7 g/100 ml of calcium carbonate, and 0.07 g/100 ml of activated carbon. Hence, the usage of this optimal adsorbent mixture could be introduced in the processing of industrial grape syrup products in order to enhance the efficacy of clarification.

CO N FLI C T O F I NTE R E S T
The authors declare that they do not have any conflict of interest.

E TH I C A L S TATEM ENT
This study does not involve any human or animal testing.

I N FO R M E D CO N S E NT
Written informed consent was obtained from all study participants.