Removal of dissolved organic nitrogen amino acid from aqueous solutions using activated carbon based on date pits

Disinfection, an integral part of water treatment, is carried out daily on large volumes of water worldwide, treating both drinking and, increasingly, recycled waters for use in agriculture, etc (Vo et al. 2014; Feng et al. 2019). Chemical disinfection may result in the unintended production of disinfectant by-products (DBPs) due to reactions between disinfectants and natural organic matter present in the water. DBPs can be associated with adverse human health effects, including bladder and rectal cancers, and unfavorable pregnancy outcomes (Li & Mitch 2018; Zhang et al. 2019). Amino acids (AAs), which are used in various industrial fields, are important precursors for halogenated nitrogenous disinfection by-products (N-DBPs) (Shim et al. 2008; Bond et al. 2012). These compounds are widely detected in soil, and surface- and ground-water (Thurman 2012). AAs are also precursors for carbonaceous disinfection by-products (C-DBPs), including halo-acetic acids (HAAs) and trihalomethanes (THMs) (Fang et al. 2010; Sharma et al. 2014).

Chlorinating water containing tyrosine, tryptophan, phenylalanine, asparagine, aspartic acid, histidine and glutamine is reported to lead to the formation of N-DBPs such as haloacetonitriles (HANs) and haloacetamides (HAMs) (Chu et al. 2012). HANs and HAMs, commonly detected in chlorinated drinking water, are of growing concern due to their higher cyto- and geno-toxicities in mammalian cell assays than the currently regulated C-DBPs (Muellner et al. 2007; Plewa et al. 2008). As a result, the occurrence of AAs, including tyrosine, in water could lead to the production of unregulated disinfection by-products, which may have adverse effects on human health. Thus it is necessary to remove tyrosine from water, and find rapid techniques and effective adsorbents for the purpose.

An effective strategy for controlling DBP formation is to remove their precursors before chlorination. Several effective amino acid removal techniques have been applied in water treatment including nanofiltration over polyelectrolyte membranes (Hong & Bruening 2006), electrodialysis (Readi et al. 2013), combined techniques (flotation complexation extraction) (Bi et al. 2008), manganese oxide oxidation (Casbeer et al. 2013) and adsorption using ionic liquids (Jonckheere et al. 2017). Among these, adsorption by activated carbon (AC) has become one of the most reliable methods used (Clark et al. 2012). The ability to use ACs as amino acid adsorbents is due to their versatility, which results from their high specific surface, porous structure, high adsorption capacity, and surface chemistry (Kopecka et al. 2014; Laksaci et al. 2017). Fruit stones – pits – are appropriate raw materials for AC production, as they can be modified by physical and/or chemical treatments to enhance adsorption. Date pits, as an abundant by-product of dates, bring the issue of sustainable use to the preparation of low-cost ACs (Belhamdi et al. 2016).

As far as can be ascertained, l-tyrosine removal using date pit-based ACs for water treatment is not yet reported in the literature. This study focuses on the use of granular activated carbons (GACs) to adsorb free amino acids during water chlorination. The main objective was to evaluate the adsorption capacities of ACs prepared from date pits by chemical activation with KOH and ZnCl₂ ACK and ACZ respectively, for l-tyrosine removal under various conditions – that is, initial concentration, contact time, solution pH and temperature.

The free amino acid (l-tyrosine) used came from Sigma Aldrich (Figure 1). It has been identified in natural waters at higher concentrations than many other free amino acids (Fang et al. 2010).

Two activated carbons were used as low-cost adsorbents to remove l-tyrosine from aqueous solution. Both are microporous and designed for water treatment based on the results of previous studies, including adsorption of low molecular weight natural organic matter (l-phenylalanine amino acid), and taste and odor compounds (Belhamdi et al. 2016). The activated carbons were synthetized from date pits using KOH and ZnCl₂ as activating agents. The raw material was first washed repeatedly with distilled water, and then dried at 120 °C. The samples were crushed and sieved into uniform size ranging from 0.5 to 1 mm. The dried precursor was impregnated with a chemical activating agent in solid form (1 g ZnCl₂: 1 g date pits) and (9 mmol KOH: 1 g date pits). These mixtures were heated in a tubular furnace at 600 and 800 °C, respectively, for 1 hour under nitrogen flow, a heating rate of 5 °C·min⁻¹. After cooling at room temperature, the activated products were washed with HCl (0.1 M) and hot distilled water until the wash-water test with AgNO₃ was negative, and then dried at 120 °C. The resulting activated carbons are denoted ACZ (ZnCl₂) and ACK (KOH).

The textural properties of ACK and ACZ are given in Table 1 and Figure 2.

A standard adsorbent dose (50 mg) was added in 100 mL flasks containing 50 mL of l-tyrosine solution (200 mg L⁻¹) at different pH values. The solution pH was adjusted by adding NaOH or HCl (0.1 M) to achieve pH values between 2 and 10, and the reaction mixture agitated on a water bath shaker at 293 K for 360 min, a time interval pre-determined as sufficient for adsorption equilibrium to be reached. The l-tyrosine uptake, (mg·g⁻¹), was calculated using Equation (1).

The l-tyrosine aqueous solution adsorption isotherms were measured using a static method. 50 mg of each AC was added to 50 mL of l-tyrosine solution at different initial concentrations from 20 to 240 mg·L⁻¹. The flasks were covered and stirred at 150 rpm for 360 min at 293 K, before the solution was filtered and analyzed using a UV–visible spectrophotometer (JASCO V-630) at 275 nm. The amount adsorbed (mg·g⁻¹) was calculated using Equation (1).

Adsorption isotherms are used to describe adsorbate/adsorbent interactions. The parameters obtained from different models provide information on the adsorption mechanisms, surface properties and adsorbent affinities. There are several equations for analyzing experimental adsorption equilibrium data and, in this study, the Langmuir, Freundlich and Redlich–Peterson (R–P) isotherm models were used.

The change in the amount of l-tyrosine adsorbed onto ACK and ACZ as a function of contact time is illustrated in Figure 3. Adsorption efficiency increases rapidly at first, then more moderately. The increasing trend stopped when equilibrium was reached, within 240 minutes for each AC. Three l-tyrosine concentrations (60, 100 and 200 mg·L⁻¹) were also used to investigate the effect of initial concentration on its adsorption. The adsorption capacity of both ACK and ACZ increases with increasing initial l-tyrosine concentration (Figure 3), reaching experimental values of 154.68 and 68.76 mg·g⁻¹, respectively, at 200 mg·L⁻¹. ACK has the higher l-tyrosine adsorption capacity due to its relatively larger microporous surface area and high microporous volume (Belhamdi et al. 2019). The enhanced adsorption capacity of both ACs at higher initial l-tyrosine concentrations can be explained by an increase in the driving force, due to the concentration gradient with increased initial l-tyrosine concentration, overcoming the mass transfer resistance of l-tyrosine molecules between the aqueous phase and solid surface.

The controlling mechanism of l-tyrosine adsorption by ACK and ACZ was studied by fitting the experimental data with the pseudo- first and second order, and the intra-particle diffusion kinetic models – see Table 2. Although the R² values of the pseudo-first order kinetic model are all acceptable (>0.96), the adsorption capacities estimated from experimental data for l-tyrosine do not correlate with that model, which is invalid for testing the l-tyrosine/ACs system. An evaluation of both the R² coefficients and the estimated (mg·g⁻¹) values indicates that l-tyrosine adsorption onto ACK and ACZ is described better by the pseudo-second order model.

Weber and Morris’ proposed intra-particle diffusion model (Figure 4) was applied to identify whether intra-particle or film diffusion is the rate-limiting step. As the intra-particle diffusion plot does not pass through the origin, with values of R² > 0.954 (Table 2), the model suggests that the l-tyrosine adsorption mechanism onto ACK and ACZ is complex and governed by more than one mechanism, and that intra-particle diffusion was not the only rate-limiting stage (Alves et al. 2013b). It is assumed that external mass transfer is significant only in the first stage of adsorption.

The l-tyrosine adsorption by ACK and ACZ is affected strongly by the solution's initial pH (Figure 5). The pH determines species of l-tyrosine in the solution as well as affecting the ACs’ surface properties. The pH at the point of zero charge (pH_PZC) was respectively 6.18 and 6.80 for ACZ and ACK (Belhamdi et al. 2016) – that is, their surfaces have a negative charge when the pH exceeds the pH_PZC. Cationic (⁺NH₃–R–COOH), zwitterionic (⁺NH₃–R–COO⁻), and anionic (NH₂–R–COO⁻) compounds are formed according to the isoelectric point of l-tyrosine (Tran et al. 2018). At pH 2 and 4, the repulsive electrostatic interactions between ACs (strong positive charge) and l-tyrosine protonated amino groups (⁺NH₃-R-COOH) in an acidic medium reduced the adsorption capacity. In contrast, adsorption was enhanced significantly when solution pH increased from 4.0 to 6.0. A possible explanation, based on l-tyrosine's hydrophobicity, is proposed in Figure 6. The higher hydrophobicity of l-tyrosine molecules at pH 6 means that hydrophobic interactions must occur to minimize contact with water (Goscianska et al. 2014) – the l-tyrosine molecules stand with their hydrophobic elements together to displace water molecules. The carboxylic groups of the l-tyrosine molecules interact strongly with positively charged sites on the ACK and ACZ surface (pH > pH_PZC) by electrostatic attraction.

Adsorption capacity fell significantly in basic solutions at pH 8 and 10. The dominant species is (NH₂-R-COO^–, pH > PI 5.66), the surface charge of the two ACs is negative, and the repulsive electrostatic interactions are unfavorable to l-tyrosine adsorption.

The adsorption isotherms, represented as the amount of l-tyrosine adsorbed onto ACK and ACZ at equilibrium (⁠⁠, mg·g⁻¹) versus the amount of amino acid remaining in solution (⁠⁠, mg·L⁻¹), are shown in Figure 7. The l-tyrosine isotherm is L-shaped (Langmuir) according to the classification by Giles et al. (Giles et al. 1960), which means that there is limited competition between the solvent and l-tyrosine molecules to occupy ACK and ACZ surface sites. The initial part of the L-curve suggests low interaction between amino acid molecules and AC adsorption sites at low concentrations, while, for high initial concentrations, adsorption occurs more readily. This may be due to l-tyrosine amino acid's non-polar groups getting closer to each other until they are within their van der Waal radii at high concentrations, leading to dense molecular packing on ACs’ active sites (Goscianska et al. 2014). The experimental equilibrium data are described by the Langmuir, Freundlich and R-P models (see section 2.2.2. Equilibrium adsorption isotherm).

The parameters deduced from different isotherm models give information on the adsorption mechanisms, surface properties and l-tyrosine affinity. The model parameters obtained from linear plots are given in Table 3. On the basis of R², the equilibrium data for l-tyrosine adsorption onto ACK and ACZ are fitted well by the Langmuir and R-P models. In contrast, the Freundlich isotherm does not provide a good fit – R² 0.90 and 0.97. It is thought, therefore, that the surfaces of ACK and ACZ display homogeneous adsorption site distribution and monolayer adsorption of l-tyrosine (Konggidinata et al. 2017). The maximum monolayer adsorption capacities were found to be 178.57 and 102.04 mg·g⁻¹ on ACK and ACZ, respectively at 293 K and pH 6. The heterogeneity factor, n, of the Freundlich model is in the range 1–10, indicating favorable adsorption conditions. Non-linear fitting of the isotherm models was also studied – see Figure 8(a) and 8(b). Both the Langmuir and R-P isotherms can generate satisfactory fits to the l-tyrosine experimental data, unlike the Freundlich model. The calculated R_L values versus the initial l-tyrosine amino acid concentration for the two ACs (Figure 8(c)) are between 0 and 1, indicating that the adsorption process is favorable. This confirms that the Langmuir isotherm is favorable for l-tyrosine adsorption onto ACK and ACZ at 293 K.

The effect of temperature on l-tyrosine adsorption is reported in Figure 9. The quantity adsorbed decreases with increasing temperature – that is, low temperatures favor l-tyrosine removal. This is likely to be due to the tendency of l-tyrosine molecules to form hydrophobic bonds in aqueous solution, hindering their hydrophobic interactions with the AC surface (Alves et al. 2013b).

The standard free energy (ΔG°), standard enthalpy (ΔH°), and standard entropy (ΔS°), evaluated from Equations (9) and (10), are shown in Table 4 and Figure 10. Negative ΔG° indicates that l-tyrosine removal from water is spontaneous for both ACs. Negative ΔH° confirms that adsorption was exothermic and physical in nature for both ACs. Positive ΔS° indicates that the randomness of the l-tyrosine molecules when in solution increases during their adsorption onto the ACs (Belhamdi et al. 2019).

The efficiency of ACK and ACZ was also compared with results reported from some other studies. Table 5 shows that the adsorption capacity of ACK and ACZ exceeds that of many adsorbents, suggesting that the two ACs have potential as adsorbents for the removal of dissolved organic nitrogen amino acid in water treatment, especially l-tyrosine.

In this study the dynamic adsorption of l-tyrosine onto date pit-based ACs was investigated. Removal of this typical amino acid from aqueous solution was controlled by factors affecting adsorption capacity such as contact time, initial l-tyrosine concentration, pH and temperature. Adsorption capacity decreased with temperature and increased with increasing initial l-tyrosine concentration. On the basis of the R² values, and the difference between the experimental and calculated values of ⁠, the pseudo second-order kinetic model provides a better fit than the pseudo-first-order model for l-tyrosine adsorption by ACK and ACZ. Weber and Morris’ proposed intra-particle diffusion model suggested that adsorption in this case is governed by more than one mechanism and that intra-particle diffusion is not the only rate limiting step.

The adsorption system was evaluated using the Langmuir, Freundlich and R–P models. The highest correlation coefficient (R²) obtained indicated that the adsorption process is best described by the Langmuir and R-P models. The two adsorbents – ACK and ACZ – have monolayer adsorption capacities as 178.57 and 102.04 mg·g⁻¹, respectively under optimized conditions – temperature 293 K, pH 6 and mixing time 360 minutes. The thermodynamic data show that l-tyrosine adsorption onto ACK and ACZ is feasible and is exothermic.

The authors wish to express their thanks for financial support from the Faculty of Chemistry (USTHB, Algiers), and for the Directorate General of Scientific Research and Technological Development (DGSRTD, Algiers).

All relevant data are included in the paper or its Supplementary Information.