The removal of toxic metals from liquid effluents by ion exchange resins. Part VI: Manganese(II)/H⁺/Lewatit K2621

Francisco J. Alguacil^*

Centro Nacional de Investigaciones Metalúrgicas (CENIM, CSIC), Avda. Gregorio del Amo 8, 28040 Madrid, España

*Corresponding author:  fjalgua@cenim.csic.es

1. INTRODUCTIONTOP

Manganese is a natural occurring element, that as many others metallic elements is essential to human health. In fact, manganese is a normal component of human tissues and fluids, but as can be expected, manganese overexposure can be toxic.

Manganese(II) intake by humans can be done via air, water or food, and health effects of inorganic manganese on humans include: death, systemic, neurologic and developmental (US ATSD, 2012). In the case of waters, and particularly drinking waters, the intake limit established by the World Health Organization is of 400 μg·L⁻¹ in a daily basis, whereas concentrations about 0.05 mg·L⁻¹ is normally acceptable for consumption, though the presence of the element in waters can be somewhat objectionable if causes water discoloration, black staining and bitter metallic taste (WHO, 2011; USEPA, 2017). Manganese is non-classified as human carcinogenic (group D). A number of disclosures and future developments on manganese toxicity related to humans can be found elsewhere (O´Neal and Zheng, 2015).

Having manganese(II) these hazardous characteristics, recently, several authors reported about the removal of this ion from aqueous solutions using different types of methodologies (Coskun et al., 2016; Li et al., 2016; Zue et al., 2016; Alexander et al., 2017; Kebabi et al., 2017; Vries et al., 2017).

Next in the series of articles published by the author about the use of ion exchange technology in the removal of toxic metals (Alguacil et al., 2002; Alguacil, 2002; Alguacil, 2003; Alguacil, 2017a; Alguacil, 2017b), the present work reports about the use of the cationic exchange resin Lewatit K2621 on the removal of manganese(II) from aqueous solutions. The experimental results were fitted to models in order to a better understanding of the exchange process.

2. EXPERIMENTALTOP

Lewatit K2621 (Fluka) has the characteristics shown in Table 1. Other chemical used in the investigation were of AR grade, whereas the experimental procedure used here was the same as described in previous works of this series. The multiwalled carbon nanotubes (MWCNTs) presented the next characteristics: carbon content > 98%, outer diameter: 10nm±1nm, inner diameter: 4.5 nm ± 0.5 nm, length: 3-6 μm) (Alguacil et al., 2016), whereas the carboxylic functionalized multiwalled carbon nanotubes (MWCNTs-ox) presented >8% carboxylic acid functionalized, 9.5 nm average diameter and 1.5 μm length.

All the loading experiments were carried out in a glass reactor using 200 mL of the aqueous solution, 1200 rpm and 20 °C, except when these variables were investigated.

Metals in solution were analysed by AAS spectrophotometry, whereas metals load in the resin was calculated by the mass balance.

3. RESULTS AND DISCUSSIONTOP

3.1. Manganese uptake onto the resinTOP

The equilibrium which represented the loading of Manganese(II) onto the resin Lewatit K2621 can be represented by a cationic exchange between protons of the sulfonic group of the resin and Mn²⁺ from the aqueous solution, thus:

where R is the non-reactive part of the resin, and r and aq represent to the resin and the aqueous solution, respectively.

To investigate the effect of the stirring speed on metal loading onto the resin, aqueous solutions containing 0.01 g·L⁻¹ manganese(II) at pH 5 were put into contact with 0.5 g·L⁻¹ of the resin at 20 °C and various stirring speeds, ranging from 250 to 1400 rpm. The results of this set of experiments being shown in Fig. 1, it can be seen (Fig.1 top) that the percentage of metal loaded onto the resin increased as the stirring speed of the system is increased, however in the 1000-1400 rpm range, and specially in the 1200-1400 rpm, this difference is, if any, negligible. Thus, in the 1200-1400 rpm a maximum in metal loading is achieved and a minimum in the thickness of the aqueous boundary layer is reached. Fig. 1 bottom represented the variation of the fractional approach to equilibrium (F) versus time. F values were calculated accordingly with the next expression:

where [Mn]_t,r and [Mn]_e,r represented the manganese concentrations in the resin at an elapsed time t and at equilibrium, respectively. When the fractional approach to equilibrium is considered, it can be seen that there is an initial influence of the stirring speed on the F value, but after 1 h of reaction, in all cases, near or more than 50% of the metal loading is achieved, being practically the F value independent of the stirring speed after 2 h or 3 h of reaction (near 80% of 90% of the total metal loading, respectively), this means that at these times, the rate of metal uptake is independent of the stirring speed, being this independence also evident for stirring speeds greater than 1000 rpm and shorter contact times. Experimental data at 1200 rpm were used to investigate the loading kinetics followed by this system. Thus, data were fit to the pseudo-first or pseudo-second kinetics order models, the results of this fitting showed that the system responded better to the pseudo-first kinetic order model (r² = 0.998):

where [Mn]_e,r and [Mn]_t,r are the manganese concentrations in the resin at equilibrium and at elapsed time, respectively, k is the rate constant and t the elapsed time. From the numerical data the value of k was estimated as 0.014 min⁻¹.

Metal loading onto the resin was also investigated as a function of the temperature. In this case, the resin dosage was of 0.25 g·L⁻¹, whereas the aqueous solution contained 0.01 g·L⁻¹ Mn(II) at pH 5. From results showed in Table 2, it can be seen that with the increase of the temperature the metal uptake in the resin increases. In this table, the values of the distribution coefficient D were calculated accordingly with:

where [Mn(II)]_e,r and [Mn(II)]_e,s are the manganese concentrations in the resin and in the solution at equilibrium, respectively. From the experimental values, ΔH° was calculated as 102.1 kJ·mol⁻¹, being the exchange reaction endothermic. The values of ΔG° and ΔS° calculated for the present system are −6.5 kJ·mol⁻¹ (spontaneous system) and 374.2 J·K⁻¹·mol⁻¹ (increase of the process randomness), respectively.

Various models were used to investigate the influence of the temperature on the mechanism in which manganese is loaded onto the resin (López Díaz-Pavon et al., 2014). From this fitting, it is showed that at 20 °C, the metal loading onto the resin is better explained by the aqueous diffusion, whereas at 60 °C the data best fit to the moving boundary model (Table 3).

The variation of the metal loaded onto the resin when the pH of the solution changes was also considered, and for these experiments solutions containing 0.01 g·L⁻¹ manganese(II) at pH values of 5 or 1, resin concentrations of 0.5 g·L⁻¹ were used. Fig. 2 showed the variation in the Manganese(II) concentration in the aqueous solution versus time at these pH values. The pH of the solution clearly influences the removal of Manganese(II) from the solution, increasing the remaining Mn(II) concentration in the solution as the pH of the solution becomes more acidic. Metal uptakes at equilibrium are 14.4 and 19.9 mg·g⁻¹ at pH values of 1 and 5, respectively, for Manganese(II) concentration in the solution of 2.8 and 0.04 mg·L⁻¹.

The ionic strength of the aqueous solution is another variable which potentially can affect to the exchange reaction and thus the Manganese(II) loads onto the resin. These experiments were carried out with resin doses of 0.25 g·L⁻¹ and aqueous solutions containing 0.01 g·L⁻¹ Mn(II) and various concentrations of LiCl, as the source for the ionic strength variation. Table 4 summarized the results from these experiments. It can be seen how clearly the increase of the ionic strength influences the metal uptake, specially when compared with solutions in which no LiCl are present, thus, with ionic strength almost zero a maximum in metal uptake (34.0 mg.g⁻¹) is reached.

Resin concentrations in the 0.05-0.5 g·L⁻¹ range were used to investigate the influence of the resin dosage on metal uptake, when aqueous solutions containing 0.01 g·L⁻¹ Mn(II) at pH 5 were used as the source for Mn²⁺ ions. The results of this investigation are summarized in Table 5, as can be expected, the decrease in the resin concentration resulted in an increase of the metal uptake onto the resin, but at the same time, the residual Mn²⁺ concentration in the solution increases. The experimental data better responded to the Freundlich model (r² = 0.901):

where [Mn]_e,r and [Mn]_e,s are the manganese concentrations at equilibrium in the resin and in the aqueous solution, respectively, K_F is a Freundlich constant and 1/n is a parameter. For the present system ln K_F = 3.6 and 1/n = 0.199.

The selectivity of the resin respect to Manganese(II) when other metals are present in the aqueous solution was studied using binary solutions of Manganese(II) and a given metal in equimolar concentrations. Table 6 summarized these results in the form of the experimental separation factors, β_Mn/metal:

where D represented the distribution coefficient for the metals, calculated as in Eq. (4). From these results it can be deduced that this resin presented a bad selectivity with respect to manganese(II), since this element only is exchanged favourably in the presence of zinc(II) and indium(III), β>1, whereas the others metals experimented in this work are exchanged in preference to manganese(II) (β<1). The differences in selectivity may be attributed, among others, to:

i) The easiness (affinity) in which the cation is exchanged with the proton from the resin, resulting in greater metal uptake.

ii) The different rate of the metal loading onto the resin.

Also the exchange capacities of the resin with respect to other smart metal-adsorbents or exchangers, as functionalized and non-functionalized multiwalled carbon nanotubes are examined. The results of this investigation are shown in Table 7, it can be seen that the results obtained within the resin are far better than those obtained with the two types of multiwalled carbon nanotubes.

3.2. Metal elution from the loaded resinTOP

Hydrochloric acid and sulfuric acid were used as eluents for the present system. The elution step in this system can be related to an exchange of the manganese ions loaded onto the resin with the correspondent concentration of protons released from the acids, that is a shifting of Eq. (1) to the left.

The use of both acids results in an acceptable yield in the recovery of manganese, 85.0 % and 92.3% for sulfuric and hydrochloric acid, respectively. Reaching the equilibrium, also in both cases, within 30 min of contact between the Mn(II)-loaded resin and the respective elution solution (Fig. 3). The conditions for the above experiments were: 0.1 g resin loaded with 1.6 mg Mn(II), aqueous solution: the corresponding acid, liquid/resin ratio: 125/1 mL·g⁻¹, temperature: 20 °C, time: 15, 30 and 60 min. The effect of varying the solution/resin ration was also investigated in the case of H₂SO₄ in order to evaluate if this variation has any effect on the percentage of Manganese(II) elution. The results from these tests are shown in Table 8, and it can be seen that this variation has not any influence in the recovery of the metal from the loaded resin. In any case it can be observed that near a ten-fold increase in the concentration of Manganese(II) in the solution, with respect to the initial one, can be achieved using the proper liquid volume/resin weight ratio.

4. CONCLUSIONSTOP

ACKNOWLEDGEMENTTOP

To the CSIC (Spain) for support. To Mr. J.M. Medina for assistance in part of the experimental work.

REFERENCESTOP