High Adsorption of Hazardous Cr(VI) from Water Using a Biofilter Composed of Native Pseudomonas koreensis on Alginate Beads

Most conventional methods to remove heavy metals from water are efficient for high concentrations, but they are expensive, produce secondary pollution, and cannot remove low concentrations. This paper proposes a biological system to remove Cr(VI) from aqueous solutions; the biofilter is composed of a native Pseudomonas koreensis immobilized in calcium alginate beads. Lab experiments were conducted in batch reactors, considering different operating conditions: Cr(VI) concentration, temperature, pH, and time. At 30 °C and a pH of 6.6, the immobilized bacteria achieved their optimal adsorption capacity. In the chromium adsorption system, saturation was reached at 30 h with a qmax = 625 mg g−1. By adjusting the experimental data to the Langmuir and Freundlich models, it is suggested that P. koreensis forms a biofilm with a homogeneous surface where Cr(VI) is adsorbed and that the bacteria also incorporates the metal in its metabolism, leading to a multilayer adsorption. On the other hand, using Fourier transform infrared spectroscopy, it was inferred that the functional groups involved in the adsorption process were O-H and C=O, which are a part of the P. koreensis cell wall.


Introduction
The presence of heavy metals in drinking water is an important issue regarding human health. The properties (toxicity, persistence, accumulation, and immutability) of these pollutants cause several human diseases [1]. Chromium stands out, since it is recognized as very toxic [2]; some of the reported risks to humans by chromium include carcinogenic potential, apoptotic cell death, and altered gene expression [3][4][5]. Moreover, it has been reported that at an accumulation of 0.1 mg by g body weight, chromium can cause death [6]. Chromium could be released into the environment by some industrial processes (metallurgy, agrochemical, and petrochemical industries), and other anthropogenic activities [4,7]. In aqueous solutions, this metal could be in two ionic forms Cr(VI) and Cr(III). It is accepted that Cr(VI) is 300 times more toxic than Cr(III) due to its ability to diffuse through biological membranes [8]. For these reasons, different organizations have developed regulations about chromium in water. The maximal permissible level for chromium in water is 1 and 0.5 mg L −1 for agriculture and public uses, respectively. In soils, the maximum levels are 510 and 280 mg kg −1 for industrial and agriculture/commercial/residential uses, respectively [9]. Concentrations (µg L −1 ) of chromium in rivers and lakes around the world were determined as 388.77 (Africa), 383.93 (Asia), 13.61 (Europe), 5.42 (North America), and 903.78 (South America) [7]. In Mexico, Cr(VI) is regulated in the Official Mexican Standard NOM-147-SEMARNAT/SSA1-2004, and it is considered hazardous waste by the standard NOM-052- SEMARNAT-2005. Two decades ago, the Federal Attorney for Environmental Protection (PROFEPA, report 1995(PROFEPA, report -2000 reported 61 sites contaminated by heavy metals in Mexico. More recently, Cr(VI) has been found in several regions in the country [9][10][11], as MEB, SEM, EDS, FTIR, were employed to explain the remotion of Cr(VI) by immobilized P. koreensis on alginate beads.

Biofilter Preparation
Alginate beads of different sizes were synthesized by coprecipitation, following the method reported elsewhere [30]. An alginate solution (15 g L −1 ) was placed in a separatory funnel and dropped into a calcium chloride solution (15 g L −1 ) under constant magnetic stirring. The obtained beads were washed three times with distilled water.
For the preparation of different sizes of alginate beads, the separatory funnel was conditioned by placing a micropipette tip of different capacity at the outlet, namely: blue tip (100 to 1000 µL), yellow tip (20 to 200 µL), and white tip (2 to 20 µL).
The quantification of the synthesized beads was obtained by manual count, considering three samples of 1 g.
To select the best size of alginate beads, significant differences were determined by an ANOVA at 95% confidence and 10% error. Since it is expected to obtain a high quantity of beads in each set, it is required to select a representative number of them to be measured and weighed as required by the statistical analysis. This number is computed using Equation (1).
where n is the number of beads that must be measured to obtain a representative sample, N is the number of beads per set, p is the proportion of beads that have the characteristics of bacteria immobilization, and q is the proportion of beads which does not have this characteristic (q = 1 − p), K is the confidence level, and E is the desired sampling error. Once the sample size was computed, the diameter (d) of the selected beads was measured with a digital Vernier. To determine the available surface (S A ) to immobilize bacteria, the beads were considered as a sphere, and their surface was computed (S A = πd 2 ); after that, the surface per gram was obtained, considering the number of beads per gram. From this information, the best kind of beads was determined.

Bacteria Immobilization
Regarding the biofilm formation, 50 mL of agar (17 g of Cetrimide L −1 agar) previously inoculated with P. koreensis (24 h) were mixed with 5 g of calcium alginate beads. The mixture was incubated at 30 • C with orbital shaking at 150 rpm for 24 h. The biofilm formation was confirmed by scanning electron microscopy (SEM): the sample was prepared with a fixation with 2% formaldehyde and 70% ethanol for 18 h, subsequently dehydrated in ascending ethanol gradients: 70, 80, and 95%.
To find the best biosorption conditions, a statistical design of experiments was proposed considering the following variables: temperature (20 and 30 • C), amount of adsorbent (0.021, 0.024, and 0.027 g), and pH (adjusted to 6.6 and unadjusted). The adsorption tests were carried out in batch reactors stirred at 100 rpm, with a reaction volume of 30 mL and an initial metal concentration of 10 mg L −1 ; all tests were performed in triplicate.

Kinetics Study
The study was carried out using different Cr(VI) concentrations in the range of 10 and 120 mg L −1 . For the preparation of the aqueous Cr(VI) stock solution, a specific amount of anhydrous K 2 Cr 2 O 7 (2.828 g) was dissolved in distilled water; the volume of the solution was made up to 1 L to obtain a concentration of 1000 mg L −1 . Solutions of different Cr(VI) concentrations, i.e., 10,20,30,50, and 100 mg L −1 , were prepared by dilution from the stock solution. After that, a concentration of 0.36 g L −1 of adsorbent was placed with 100 mL of Cr(VI) solutions at different initial concentrations; the mixture was kept in orbital shaking considering 100 rpm for 36 h at 30 • C. All experiments were carried out by duplicate.
The amount of Cr(VI) adsorbed by the immobilized bacteria was calculated using the following equation: where q (mg metal mg adsorbent −1 ) represents the uptake of the metal by the biofilter, V is the reaction volume (L), C i is the initial concentration of Cr (mg L −1 ), C f is the final concentration of the metal (mg L −1 ), and m is the mass of the bioadsorbent on a dry basis (g).
The obtained data from the experiments were used to build the adsorption isotherms, which are analyzed considering the models of Langmuir and Freundlich, which are presented in Equations (3) and (4), respectively.
where C e is the adsorbate concentration in the solution at equilibrium, q max is the maximum adsorption capacity of the bioadsorbent, q e is the amount of adsorbate adsorbed per unit weight of adsorbent at equilibrium, K L is the Langmuir equilibrium constant, K F is the Freundlich constant, and 1/n is defined as the adsorption intensity according to with the next criteria: for (1/n) = 1, the relative adsorption (adsorption partition) is a linear function of the adsorbate concentration; when 0.7 < (1/n) < 1, the relative adsorption decreases as the concentration increases, which implies saturation of adsorption sites; (1/n) < 0.7 describes highly curved isotherms; finally, cooperative adsorption is represented by (1/n) > 1. Regarding the Langmuir model, the separation factor R L , also known as the equilibrium parameter, is computed with the following equation: The separation factor can be used as a guide to determine the adsorption strong or adsorbent/adsorbate interaction according to the following criteria: R L > 1 implies nonfavorable conditions for adsorption, R L = 1 corresponds to a linear adsorption function, R L = 0 represents non-reversible adsorption and 0 < R L < 1 implies favorable conditions for adsorption [31].
To ease the computation of the parameters q max , K L , a reciprocal linearization of Equation (3) is recommended; for the computation of the parameters 1/n and K F , a logarithmic linearization of Equation (4) is suggested. The linearized equations are (6) and (7), respectively. 1 log To evaluate the rate of the adsorption, pseudo-first and pseudo-second order kinetic models were applied [19,32,33]. The linear form of those models is represented by Equations (8) and (9), respectively.
where q e , q t , K 1 , and K 2 are the equilibrium adsorption capacity, adsorption capacity at time t, first-order constant, and second-order constant, respectively. For the pseudo-first-order equation, the plot of left side versus time provides a linear graph from a linear regression; from the slope of this graph, the constant K 1 is computed. Regarding the pseudo-secondorder model, the intercept and slope of the line obtained from the linear regression of t/q t versus time are used to obtain K 2 .

Determination of Dry Weight of Bacteria Attached to Alginate Beads
The amount of adsorbent (P. koreensis) in the system was determined with the following procedure: 50 g (by wet weight) of the biofilter (P. koreensis adhered to the alginate beads) were taken and washed three times; the biofilm was removed by adding 200 mL of Tween 80 (0.1%) to the beads and stirring for 3 h in a magnetic stirrer [34]. Subsequently, the moisture on the supernatant obtained from the stirring stage was removed by drying the material in an oven at 110 • C until a constant weight was achieved.

Alginate Beads
The characteristics of the beads obtained using the different micropipette tips are listed in Table 1. The size lends noticeable differences to the beads, especially for the amount and the available surface to carry the microorganisms. This inference is corroborated by the ANOVA (Table 2). Then, since F > FCV, significant differences exist between the three synthesized beads' sizes. The relationship between the beads' size and the surface per gram was considered to determine the best bead size: the fewer the beads, the larger the available surface to immobilize P. koreensis. Then, beads with S = 12.56 mm 2 (2 mm diameter) were allowed to obtain the largest contact surface for the biofilter, and it was selected for successive experiments.

Biosorption System Implementation
The adhesion of the bacteria to the alginate beads was verified by observation under a scanning electron microscope (SEM at 5000×), complemented by an energy-dispersive X-ray spectroscopy (EDS) analysis (XL-30 ESEM, Philips; North Billerica, MA, USA). First, virgin alginate beads were analyzed; the results are shown in Figure 1.
It can be appreciated there is a coarse surface in the micrograph of Figure 1, which corresponds to virgin alginate beads. The EDS analysis showed that Ca is the main component of the beads, and no other elements are present since the beads were not in contact with either the bacteria or other material; this agrees with the report of Ohemeng-Boahen et al. [35]. It can be appreciated there is a coarse surface in the micrograph of Figure 1, which corresponds to virgin alginate beads. The EDS analysis showed that Ca is the main component of the beads, and no other elements are present since the beads were not in contact with either the bacteria or other material; this agrees with the report of Ohemeng-Boahen et al. [35].
After that, representative samples of beads taken from the immobilization system were analyzed at 8, 12, and 24 h. The respective micrographs are presented in Figure 2, where the adhesion of P. koreensis to the support is demonstrated. After that, representative samples of beads taken from the immobilization system were analyzed at 8, 12, and 24 h. The respective micrographs are presented in Figure 2, where the adhesion of P. koreensis to the support is demonstrated.
At 8 h (Figure 2a), the micrograph is already different from the obtained for the virgin bead. It is possible to observe the presence of bacteria colonies randomly dispersed in space, forming a biofilm over the surface of alginate beads. The micrograph analysis also reveals the presence of cell aggregates, gaps, and embedded biomass in the alginate matrix, approximately 1 µm in length, which is consistent with the characteristics P. koreensis.
The micrograph of the beads sampled at 12, and 24 h (Figure 2c,e) shows slight changes compared with that obtained for the beads sampled at 8 h. The bacillary forms of P. koreensis are clearly remarked on the surface of the alginate beads. Then, it is deduced that 8 h is enough time for biofilm formation.
The presence of P. koreensis on the surface of the beads was confirmed by EDS analysis. The characteristic elements of the microorganisms' prokaryotic wall (P, N, K) were identified in the spectrums shown in Figure 2b,d,f, which are similar to those found for other Pseudomonas species [36,37]. None of those above characteristics was observed in the virgin beads.

Cr(VI) Removal Tests for Optimization
The data obtained from the batch tests described in Section 2.2 were processed by an ANOVA at 95% confidence to determine the best operating conditions. Table 3 presents the values of the corresponding analysis, considering the associated variability of time and dose of bio-adsorbent and their interactions.  At 8 h (Figure 2a), the micrograph is already different from the obtained for the virgin bead. It is possible to observe the presence of bacteria colonies randomly dispersed in space, forming a biofilm over the surface of alginate beads. The micrograph analysis also reveals the presence of cell aggregates, gaps, and embedded biomass in the alginate matrix, approximately 1 µm in length, which is consistent with the characteristics P. koreensis. It should be noted that significant differences were obtained in the triple interaction pH, temperature, and adsorbent concentration. This is because the effect of a factor is defined as the change in the response produced by a difference in the factor level; in this case, almost all the terms were statistically significant (p < 0.05), except the influence of temperature and the interaction of the temperature versus the concentration of the adsorbent.
However, even if this analysis allows determining that the experiments show statistically significant differences, it is not possible to obtain information regarding which treatment is different from the others. Then, a Fisher test was performed (Table 4). This test allows direct comparisons between different treatments of individual groups, through a less significant difference test [38]. Thus, from the results of the Fisher test, it is concluded that Treatment 1 (30 • C, pH 6.6, amount of adsorbent 0.027 g) was the best one. Therefore, these conditions were selected for the subsequent experiments. It is important to remark that the selected conditions agree with the optimal growth conditions of Pseudomonas [39]. On the other hand, the quantity of microorganisms in a biosorption system is directly proportional to its efficiency; this situation confirms that the best treatment has the largest number of microorganisms.
Although the results of Fisher's test indicate that Treatment 1 corresponds to the best adsorption conditions, other treatments allow reaching similar final chromium concentrations. Therefore, it is inferred that P. koreensis present biological activity at 20 and 30 • C, and the two pH tested, thus proving to be a robust microorganism. This represents an advantage for continuous reactors and scaling up the adsorption system.

Biosorption Effect on the Biofilter Surface
Samples of the biofilter were taken before adsorption (0 h) and after Cr adsorption (48 h). The samples were analyzed by SEM and EDS to determine morphological and chemical changes that occurred on the biosorbent surface due to the metal adsorption process, and to visualize the interaction between the adsorbed contaminants and the P. koreensis/alginate beads system. The corresponding results are presented in Figure 3.
At 0 h, the biofilm on the surface is still clearly observed (Figure 3a) as for the biofilter preparation stage (Section 3.2); in addition, the EDS analysis (Figure 3b) allows corroboration of the presence of P. koreensis on the alginate beads through the signals of P, N, and K, which conform to the cell wall.
On the other side, at 48 h of experimentation, the corresponding micrograph (Figure 3c) shows that P. koreensis are still on the alginate beads after the bioabsorption process. However, from the EDS analysis, it is not possible to demonstrate the presence of the characteristic elements of the bacterial cell wall (Figure 3d) as for the previous analysis. This may be explained since, although the bacteria are still present in the biofilter, they are highly dispersed. On the other side, at 48 h of experimentation, the corresponding micrograph ( Figure  3c) shows that P. koreensis are still on the alginate beads after the bioabsorption process. However, from the EDS analysis, it is not possible to demonstrate the presence of the characteristic elements of the bacterial cell wall (Figure 3d) as for the previous analysis. This may be explained since, although the bacteria are still present in the biofilter, they are highly dispersed.
For this reason, it was necessary to resort to another analytical technique to better understand how the process of adsorption of contaminants was carried out on the P. koreensis/alginate system. Fourier transform infrared (FTIR) analysis was selected for this goal for two reasons: it is possible to identify functional groups, and the preparation of the sample to be analyzed allows including a larger number of bacteria, so then the microorganisms can be analyzed with greater precision. The FTIR spectrum of samples taken before (0 h) and after (32 and 48 h) the adsorption is presented in Figures 4 and 5, respectively. For this reason, it was necessary to resort to another analytical technique to better understand how the process of adsorption of contaminants was carried out on the P. koreensis/alginate system. Fourier transform infrared (FTIR) analysis was selected for this goal for two reasons: it is possible to identify functional groups, and the preparation of the sample to be analyzed allows including a larger number of bacteria, so then the microorganisms can be analyzed with greater precision. The FTIR spectrum of samples taken before (0 h) and after (32 and 48 h) the adsorption is presented in Figures 4 and 5, respectively.
The FTIR spectrum before the adsorption (Figure 4) exhibits several bands corresponding to the characteristic functional groups of the bacteria biochemical compounds, specifically peptidoglycans, proteins, and lipopolysaccharides [40]. The signal centered at 3293 cm −1 is associated with the vibration of hydroxyl and amino groups (O-H and N-H stretching) corresponding to peptidoglycans [41]. Likewise, a characteristic band of C-H is detected at 2929 cm −1 ; the signal of the carboxylic group C=O is observed around 1626 cm −1 , and the symmetrical stretching of C-O-C at 1038 cm −1 . Finally, the functional groups of nitrogen of the amino groups can be observed at 1425 cm −1 , which corresponds to the C-N bond, as well as the vibration outside the plane of the group N-H at 929 cm −1 [42]. The FTIR spectrum seems like the one obtained for Arthrobacter viscosus [41] and for Pseudomonas species such as P. alcaliphila [37] and for P. stutzeri [43]. The FTIR spectrum before the adsorption (Figure 4) exhibits several bands corresponding to the characteristic functional groups of the bacteria biochemical compounds, specifically peptidoglycans, proteins, and lipopolysaccharides [40]. The signal centered at 3293 cm −1 is associated with the vibration of hydroxyl and amino groups (O-H and N-H stretching) corresponding to peptidoglycans [41]. Likewise, a characteristic band of C-H is detected at 2929 cm −1 ; the signal of the carboxylic group C=O is observed around 1626 cm −1 , and the symmetrical stretching of C-O-C at 1038 cm −1 . Finally, the functional groups of nitrogen of the amino groups can be observed at 1425 cm −1 , which corresponds to the C-N bond, as well as the vibration outside the plane of the group N-H at 929 cm −1 [42]. The FTIR spectrum seems like the one obtained for Arthrobacter viscosus [41] and for Pseudomonas species such as P. alcaliphila [37] and for P. stutzeri [43].
The FTIR spectrum of the biofilter samples taken after the adsorption ( Figure 5) showed some changes from the spectrum of the sample taken before the adsorption. First, for the spectrum of the sample taken at 32 h, an intensity decrease of the band centered at 3293 cm −1 (assigned an O-H stretching vibration) was observed (Figure 5a). Meanwhile, the bands related to C-H bonds (~2900 cm −1 ), showed no remarkable changes. On the other hand, a remarkable change in the carbonyl group C=O band was observed; the intensity of this band is drastically reduced in the sample taken at 32 h (Figure 5a) until it almost disappears in the sample taken at 48 h (Figure 5b). Moreover, a shift in C-O-C symmetric stretch from 1038 to 1090 cm −1 was observed. Finally, a relative increase in the intensity of the bands corresponding to the N-H bond was observed at 32 and 48 h after adsorption. These remarks allow us to deduce that the functional groups O-H, and C=O are the main ones responsible for the adsorption of Cr(VI) in this system. This behavior is similar to that reported for A. viscosus in a Cr(VI) adsorption batch reactor [41]; the authors associated the removal of the metal to the carboxyl, phosphoric, or amine groups. In addition, the FTIR spectrum after adsorption presented some differences in comparison with those observed by El-Naggar et al. [37] for P. alcaliphila removing Cr(VI); due to the modifications in the bands of the spectra, the authors associate the Cr(VI) adsorption to groups such as phenolic, carbonyl ester, acetyl, carboxylate, alkanes, and carbonyl. These differences could be due to the specific characteristics and metabolism of the Pseudomonas genus.

Kinetic Adsorption Experiments
Here, it is important to mention that in preliminary assays it was observed that the pristine alginate beads have a considerable contribution in the chromium removal (about 20%); however, the P. koreensis presence in the support increased the removal until it reached nearly 100% in a synergic process. Then, considering the results of the tests for optimization (Section 3.3), a new experimental stage was implemented: a series of adsorption kinetic assays were performed to identify interactions in the adsorbent/adsorbate system. The operating conditions considered were: pH 6.6, stirring 100 rpm, temperature 30 °C, initial concentrations of Cr(VI) of 0, 10, 25, 40, 55, 70, 85, 100, and 120 ppm, amount of adsorbent (bacteria dry weight) remained constant as 0.36 g L −1 , and reaction volume (Cr(VI) solution) was 100 mL. The experiments were performed by duplicate, and samples for monitoring were taken at 0, 4, 8, 12, 24, 28, 30, 32, 34, and 36 h. The adsorption system behavior is shown in Figure 6. The FTIR spectrum of the biofilter samples taken after the adsorption ( Figure 5) showed some changes from the spectrum of the sample taken before the adsorption. First, for the spectrum of the sample taken at 32 h, an intensity decrease of the band centered at 3293 cm −1 (assigned an O-H stretching vibration) was observed (Figure 5a). Meanwhile, the bands related to C-H bonds (~2900 cm −1 ), showed no remarkable changes. On the other hand, a remarkable change in the carbonyl group C=O band was observed; the intensity of this band is drastically reduced in the sample taken at 32 h (Figure 5a) until it almost disappears in the sample taken at 48 h (Figure 5b). Moreover, a shift in C-O-C symmetric stretch from 1038 to 1090 cm −1 was observed. Finally, a relative increase in the intensity of the bands corresponding to the N-H bond was observed at 32 and 48 h after adsorption. These remarks allow us to deduce that the functional groups O-H, and C=O are the main ones responsible for the adsorption of Cr(VI) in this system. This behavior is similar to that reported for A. viscosus in a Cr(VI) adsorption batch reactor [41]; the authors associated the removal of the metal to the carboxyl, phosphoric, or amine groups. In addition, the FTIR spectrum after adsorption presented some differences in comparison with those observed by El-Naggar et al. [37] for P. alcaliphila removing Cr(VI); due to the modifications in the bands of the spectra, the authors associate the Cr(VI) adsorption to groups such as phenolic, carbonyl ester, acetyl, carboxylate, alkanes, and carbonyl. These differences could be due to the specific characteristics and metabolism of the Pseudomonas genus.

Kinetic Adsorption Experiments
Here, it is important to mention that in preliminary assays it was observed that the pristine alginate beads have a considerable contribution in the chromium removal (about 20%); however, the P. koreensis presence in the support increased the removal until it reached nearly 100% in a synergic process. Then, considering the results of the tests for optimization (Section 3.3), a new experimental stage was implemented: a series of adsorption kinetic assays were performed to identify interactions in the adsorbent/adsorbate system. The operating conditions considered were: pH 6.6, stirring 100 rpm, temperature 30 • C, initial concentrations of Cr(VI) of 0, 10, 25, 40, 55, 70, 85, 100, and 120 ppm, amount of adsorbent (bacteria dry weight) remained constant as 0.36 g L −1 , and reaction volume (Cr(VI) solution) was 100 mL. The experiments were performed by duplicate, and samples for monitoring were taken at 0, 4,8,12,24,28,30,32,34, and 36 h. The adsorption system behavior is shown in Figure 6.

Kinetic Adsorption Experiments
Here, it is important to mention that in preliminary assays it was observed that the pristine alginate beads have a considerable contribution in the chromium removal (about 20%); however, the P. koreensis presence in the support increased the removal until it reached nearly 100% in a synergic process. Then, considering the results of the tests for optimization (Section 3.3), a new experimental stage was implemented: a series of adsorption kinetic assays were performed to identify interactions in the adsorbent/adsorbate system. The operating conditions considered were: pH 6.6, stirring 100 rpm, temperature 30 °C, initial concentrations of Cr(VI) of 0, 10, 25, 40, 55, 70, 85, 100, and 120 ppm, amount of adsorbent (bacteria dry weight) remained constant as 0.36 g L −1 , and reaction volume (Cr(VI) solution) was 100 mL. The experiments were performed by duplicate, and samples for monitoring were taken at 0, 4,8,12,24,28,30,32,34, and 36 h. The adsorption system behavior is shown in Figure 6.  The adsorption is progressively developed following a similar trend with all initial concentrations assayed; at 30 h, a quasi-equilibrium point is reached in all cases. After that time, lower concentrations remain in the equilibrium, but higher concentrations still present increasing adsorption. This implies that the immobilized bacteria can remove even larger concentrations of Cr. Since the experimentation time was restricted to 36 h, and knowing that the adsorption achieved at higher concentrations could be larger, the quasi-equilibrium point at 30 h was taken for further analysis.
It is pertinent to indicate that none of the experiments demonstrated the formation of a precipitate at the bottom of the reactor, which indicates that during the chromium adsorption process, there were no reactions for the formation of compounds not assimilable by the biofilter.
Further analysis of the adsorption isotherms using the Langmuir and Freundlich models was performed to determine the strength and type of interaction between the biofilter and the contaminant. Even if these models were proposed to describe the adsorption of gases on a solid, they are used for other systems, since it is accepted that they provide valid information for more complex systems such as biological ones, as reported in several studies (Table 5). Then, for the isotherm analysis, it was considered that the biofilter saturation was reached at 30 h. Figure 7 includes the linearized isotherms; Table 6 lists the corresponding parameters obtained from Equations (5) and (6) and the linear equation representing the isotherm.   The linearization of the adsorption isotherms data in the Langmuir model coordinates is shown in Figure 7a; a determination coefficient R 2 = 0.981 was obtained, indicating a good fitting of the adsorption systems by this model.
Concerning the adsorption rate, Table 7 lists the correlation coefficients and constants for low, middle, and high initial concentrations.

Discussion
The coarse morphology observed on virgin alginate beads agrees with those reported in other works [49]; this characteristic seems adequate for a bacteria immobilization as reported for other materials used as microorganism carriers [37,50,51].
The biofilm formation on the surface of alginate beads is similar to that reported by Mangwani et al. [52], who observed P. mendocina on the glass; also, El-Naggar et al. [37], achieved the immobilization of P. alcaliphila in sodium alginate beads. In their studies, using SEM, the authors verified the presence of a biofilm composed of bacillary aggregates [37,52].
The maximal adsorption capacity obtained from experiments (qmax-ex) is 168 mg g −1 ; this value is similar and even better than the results obtained by other authors in several heavy metals biosorption systems [14,19,22,26]. In this sense, the P. koreensis/alginate beads biofilter tested in the present study is a promising option for the decontamination The linearization of the adsorption isotherms data in the Langmuir model coordinates is shown in Figure 7a; a determination coefficient R 2 = 0.981 was obtained, indicating a good fitting of the adsorption systems by this model.
Concerning the adsorption rate, Table 7 lists the correlation coefficients and constants for low, middle, and high initial concentrations.

Discussion
The coarse morphology observed on virgin alginate beads agrees with those reported in other works [49]; this characteristic seems adequate for a bacteria immobilization as reported for other materials used as microorganism carriers [37,50,51].
The biofilm formation on the surface of alginate beads is similar to that reported by Mangwani et al. [52], who observed P. mendocina on the glass; also, El-Naggar et al. [37], achieved the immobilization of P. alcaliphila in sodium alginate beads. In their studies, using SEM, the authors verified the presence of a biofilm composed of bacillary aggregates [37,52].
The maximal adsorption capacity obtained from experiments (q max-ex ) is 168 mg g −1 ; this value is similar and even better than the results obtained by other authors in several heavy metals biosorption systems [14,19,22,26]. In this sense, the P. koreensis/alginate beads biofilter tested in the present study is a promising option for the decontamination of water containing Cr(VI). In addition, from the kinetic study and using the Langmuir adsorption isotherm, it was determined that the theoretical q max is 625 mg g −1 ; the values is high in comparison with other works. In this context, further studies could be developed to prove this theoretical q max by experiments.
From the application of the Langmuir model to the P. koreensis kinetics, a separation factor (R L = 0.6304) was obtained that satisfies the condition 0 < R L < 1, which corresponds to a favorable adsorption process. That means the P. koreensis/alginate beads system presents adsorption characteristics to attract and to grab the metal [13,31]. Moreover, the R L value is in the middle of the interval |0 1|, which is interpreted as the limits for reversibility since R L = 0 corresponds to irreversible adsorption and R L = 1 to reversible adsorption [13,33]; then, it is considered that the strength of the adsorption by P. koreensis is enough to remove Cr(VI) and it is also convenient for desorption, which is useful for biofilter regeneration.
Regarding the Freundlich model, the behavior of the Cr(VI) adsorption indicates a directly proportional relationship between the adsorption capacity of the adsorbent material (P. koreensis/alginate system) and the chromium concentration at the equilibrium state. A quite acceptable correlation was obtained since the coefficient of determination is R 2 = 0.978. According to the fundamentals of the Freundlich model [53], the behavior of the Cr(VI) adsorption indicates a directly proportional relationship between the adsorption capacity of the adsorbent material (P. koreensis/alginate beads system) and the chromium concentration at the equilibrium state. Moreover, the parameter 1/n < 1 implies the adsorbent material is able for normal adsorption, which corresponds to the separation factor obtained from the Langmuir model. In addition, since 1/n = 0.9345 matches the condition, 0.7 < (1/n) < 1, it is inferred that the relative adsorption of Cr(VI) by the P. koreensis biofilter decreases as the concentration of the metal increases, which implies that saturation of adsorption sites is occurring, as it was deduced for another biological adsorption system [54].
In addition, the determination coefficient for both models is similar (R 2 = 0.978 for Freundlich and R 2 = 0.981 for Langmuir), which implies that the two approaches describe the adsorption process. A similar situation (adjustment to both models Freundlich and Langmuir) has been observed in other biosorption systems [41,48,54]. For Cr(VI) removal, some authors proposed that the behavior of adsorption in microorganism/support systems is as a monolayer with heterogeneous distribution of active sites [32], and including intracellular bioaccumulation [41]. This could suggest that the P. koreensis/alginated beads system forms a biofilm with a homogeneous distribution of active sites in the surface where Cr(VI) is adsorbed. After that, P. koreensis also assimilates the metal inside its cellular structure; then, the adsorption system follows a multilayer adsorption. For this reason, the capacity for chromium removal is high. This highlights the potential of P. koreensis to remove heavy metals, and that it is suitable for further investigations with other operating conditions, higher adsorption volumes, and reactor configurations.
Moreover, the experimental data better fit the pseudo-second-order model, which corresponds with the reported in other sorption studies [19]. Therefore, it can be deduced that the adsorption depends on the active sites available on the P. koreensis/alginate beads system, which acts as a biosorbent [19,33]. In addition, this situation suggests that adsorption of Cr(VI) on the biofilter occurs, which corresponds to the obtained from the adsorption isotherms [32].
On the other hand, FTIR analysis suggests a bonding between the hydroxyl groups of the biofilter surface and the chromium when the signal at~3300 cm −1 after the adsorption process was reduced in intensity. Moreover, since the signals observed at~2920 cm −1 did not suffer notorious changes before and after adsorption, the reduction of Cr(VI) to Cr(III) [55] was discarded.

Conclusions
The biofilter formed by P. koreensis/alginate beads is able to remove up to 96% of Cr(VI) in an aqueous solution at 30 • C and pH 6.6. A theoretical maximal adsorption capacity of q max = 625 mg g −1 was determined. From the isotherm analysis, it was observed that both models, Langmuir and Freundlich, fit the experimental data. This suggests that the adsorption is performed in homogeneous distribution of active sites and multilayer adsorption, since P. koreensis also assimilate chromium in their cellular structure, according to the separation parameter. In addition, the functional groups O-H and C=O, which are part of the prokaryotic wall of bacteria, were identified as responsible for the adsorption of Cr(VI).

Perspectives
From this research, it is possible to mention that P. koreensis could be used as a basis to develop systems for water purification. However, complementary works are required to evaluate the P. koreensis performance to remove heavy metals in presence of other elements such as anions, cations and multimetals, as well as in real water environments. Moreover, the recycling of the proposed biofilter should be considered for further studies. It is required to determine whether alginate beads are able to keep the mechanical resistance for several cycles; also, the system P. koreensis/alginate beads should be evaluated economically for scaled-up application.
On the other hand, a deeper analysis has to be conducted to solve the questions related to the nature of the reactions occurring during the adsorption process. In fact, this kind of study could help to understand the mechanisms better and eventually predict a specific behavior with similar systems or parameter change.
However, once the biofilter is exhausted and it is impossible to regenerate, a disposal strategy is required. Underground enclosure is an easy method to dispose of the materials; but other options should be explored. In this context, circular economy approaches can be an alternative: the adsorbed metal could be recovered through thermal treatment of the biofilter; after that, the recovered metal could be used in its current applications, obtaining economic benefits.