Separation of Mandelic Acid by a Reactive Extraction Method Using Tertiary Amine in Different Organic Diluents

Mandelic acid is a valuable chemical that is commonly used in the synthesis of various drugs, in antibacterial products, and as a skin care agent in cosmetics. As it is an important chemical, various methods are used to synthesize and extract this compound. However, the yields of the used processes is not significant. A dilute aqueous solution is obtained when using several production methods, such as a fermentation, etc. In this study, the reactive extraction of mandelic acid from aqueous solutions using tri-n-octylamine extractant at 298.15 K was investigated. Dimethyl phthalate (DMP), methyl isobutyl ketone (MIBK), 2-octanone, 1-octanol, n-pentane, octyl acetate, and toluene were used as diluents. The batch extraction results of the mandelic acid experiments were obtained for the development of a process design. Calculations of the loading factor (Z), distribution coefficient (D), and extraction efficiency (E%) were based on the experimental data. The highest separation yield was obtained as 98.13% for 0.458 mol.L−1 of tri-n-octylamine concentration in DMP. The overall extraction constants were analyzed for the complex of acid-amine by the Bizek approach, including K11, K12, and K23.


Introduction
Mandelic acid, also known as almond acid, is an aromatic alpha-hydroxycarboxylic acid derived from the hydrolysis of the extract of bitter almonds [1]. It is a raw material and synthetic intermediate used for the preparation of pharmaceutical compounds such as antibiotics, drugs, etc., in medicine [2]. In addition, mandelic acid is a beneficial compound for use in antibacterial, healthcare and skincare products in the cosmetic and chemical industries [3,4]. For many years, mandelic acid has been used for treating skin problems such as acne, sun damage, or photoaging in skincare products [5,6]. It is an influential agent for the treatment of wrinkles and irregular pigmentation, to the same extent as glycolic acid, which is commonly used in skincare [7]. Glycolic acid has a smaller molecular structure and penetrates the skin deeply [8]. However, mandelic acid provides a slow and uniform penetration of the skin and irritates it less than the glycolic acid [9]. Therefore, mandelic acid has become a preferred skin care agent in recent years, used preferentially over glycolic acid. The production of mandelic acid can be carried out using various methods, including its chemical synthesis from potassium cyanide/benzaldehyde with chloroform. In addition to chemical synthesis, it can also be produced via biotechnological processes, such as the conversion of benzoylformic acid into mandelic acid using the micro-organisms of micrococcus freudenreichii, micrococcus luteus, enterococcus faecalis, and enterococcus faecalis [10,11]. On the other hand, mandelic acid can be derived from the extraction of almonds with diluted hydrochloric acid [10]. By biotechnological processes and extraction methods, diluted aqueous solutions of mandelic acid are obtained. Therefore, the separation of mandelic acid from aqueous solutions and its conversion into commercial forms occur as significant processes.

Type of Extractant
Type of Carboxylic Acid Ref.

Results and Discussion
The reaction mechanism of mandelic acid (MA) with tri-n-octylamine (TOA) can be described by Equation (1): Herein, the m moles of undissociated mandelic acid react with n moles of tri-noctylamine at the external interface between the aqueous phase (aq) and the organic phase The loading factor (Z) is obtained by dividing the total amount of mandelic acid in the organic phase C ma,org by the total amount of tri-n-octylamine in the organic phase C TOA,org . This expression can be written as follows [28]: The distribution coefficients (D) of the mandelic acid extracted from the aqueous phase transitioning into the organic phase and the efficiency of extraction (E) can be calculated by Equations (4) and (5), respectively [29]: In Equation (5), C ma is the concentration of the mandelic acid in the aqueous phase after extraction and C ma0 is the initial concentration of mandelic acid in the aqueous phase. An efficiency of extraction of 100% means that all of the mandelic acid in the aqueous phase has been removed.
The results of the reactive extraction experiments are displayed in Table 2. Additionally, the conventional extraction results without tri-n-octylamine are presented in the same table. The initial concentration of mandelic acid was 0.74 mol.L −1 (10% w/w) in the aqueous phase, and the concentrations of tri-n-octylamine in the diluents ranged from 0.092 mol.L −1 and 0.458 mol.L −1 in both experiments. The results showed that the conventional extractions were realized with a low level of efficiency, and the distribution coefficients ranged from 0.01 to 0.78 without tri-n-octylamine in the n-pentane, octyl acetate, and toluene. It was observed that more than 70% of the mandelic acid could be separated with the other diluents, DMP, MIBK, 2-octanone, and 1-octanol. This can be attributed to the polarity of these diluents, which gives them a high extraction efficiency. However, the polarity alone is not sufficient to completely explain the solvating ability. In this study, the use of an alcohol, such as 1-octanol, which has a high hydrogen binding capacity, led to high distribution coefficients [33,61].
To achieve higher yields, the reactive extraction experiments were also performed, and the results indicated that the extraction efficiency and the dispersion coefficients increased with an increasing tri-n-octylamine concentration in all the diluents. The high extraction efficiency was found to be 98.13% using DMP with 0.458 mol.L −1 of tri-n-octylamine concentration. It was seen that the E% values ranged between 30.67 and 98.13 with increasing concentrations of tri-n-octylamine from 0.092 to 0.458 mol.L −1 . Figure 1 shows the plot of the E% values for all the diluents employed. The distribution coefficients, which ranged from 0.44 to 52.5 with increasing of trin-octylamine concentrations, were calculated using Equation (4). The tri-n-octylamine concentration vs. distribution coefficient for each diluent is presented in Figure 2.  Figure 3 displays the concentration of tri-n-octylamine for each diluent against the loading factor. The loading factors had high values, ranging from 1.53 to 6.99, and it can be stated that the system was an overloading one. In this case, the complex of [(MA) m .(TOA) n ] was formed with more than one mandelic acid per tri-n-octylamine [61]. The increase in the basic amine compound concentration could lead to a decrease in the polarity and dissolution. As can be seen from Figure 3, the values of Z decreased with the increasing tri-n-octylamine concentration. In the literature, the decrease in the loading factors has been described as an indication of the reaction mechanism, in which the complexes contain more than one amine per complex [33,61].
The Bizek approach is commonly used for predicting complex formations in reactive extractions [62]. By this approach, the stoichiometry of the complexes can be formed as (MA).(TOA), (MA).(TOA) 2 and (MA) 2 .(TOA) 3 . K 11 , K 12 , and K 23 , which were the overall extraction constants, were analyzed using the following equations, respectively:  Table 3 shows the overall extraction constants calculated using Equations (6)-(8). K 11 was calculated for all the solvents. According to the Bizek approach, we calculated K 23 for the non-protonating diluents DMP, MIBK, 2-octanone, n-pentane, octyl acetate, and toluene, and only K 12 for 1-octanol.
SCHOTT TitroLine ® Easy M2 was used to analyze the concentration of mandelic acid in the aqueous phase with sodium hydroxide (relative uncertainty: ±1%). Each analysis was performed three times to minimize the errors. Experimental data were used to calculate the loading factor (Z), distribution coefficient (D), and extraction efficiency (E%).

Conclusions
The extraction of mandelic acid from aqueous phases by tri-n-octylamine in seven diluents at 298.15 K was investigated. As a result, the extraction efficiency of mandelic acid with tri-n-octylamine was found to be high, especially in the case of polar diluents such as DMP, 2-octanone, and MIBK. The maximum extractability of mandelic acid was 98.13% with DMP (0.458 mol.L −1 concentration of tri-n-octylamine). The maximum extraction efficiencies for diluents used with maximum concentration of tri-n-octylamine were identified as DMP > 2-octanone > MIBK > toluene = octyl acetate > 1-octanol > n-pentane. Using Bizek approach, the overall extraction constants of K 11 , K 12 , and K 23 were calculated as 4.81-114.63 mol.L −1 , 123.02-620.05 L 2 .mol −2 , and 9.37-8286.93 L 4 .mol −4 , respectively. These results show that the reactive extraction method is an effective method for the separation of mandelic acid, and tri-n-octylamine is a compatible reagent for this separation process.

Acknowledgments:
The authors wish to thank the University of Oradea, Oradea, Romania, for the financial support in publishing this paper.

Conflicts of Interest:
The authors declare no conflict of interest.

C ma
Molar concentration of acid in the aqueous phase (mol.L −1 ) C ma0 Initial molar concentration of acid in the aqueous phase (mol.L −1 ) C ma,org Molar concentration of acid in the organic phase (mol.L −1 ) C TOA,org Molar concentration of amine in the organic phase (mol.L −1 ) D Distribution coefficient DMP Dimethyl phthalate E Extraction efficiency MA Mandelic acid MIBK Methyl isobutyl ketone TBA Tri-n-butylamine TBP Tri-n-butyl phosphate TOA Tri-n-octylamine TOPO Tri-n-octyl phosphine oxide TPA Tri-n-propylamine Z Loading factor