Adsorption Study of Uremic Toxins (Urea, Creatinine, and Uric Acid) Using Modified Clinoptilolite

Abstract

The development of materials for uremic toxin removal is under continuous research. In this work, a natural zeolite (clinoptilolite) was modified using tartaric acid through two different methods: conventional reflux heating and ultrasound energy. The resulting materials were used as an adsorbent material for the removal of uremic toxins such as urea, creatinine, and uric acid. In the uremic toxin removal study, it was observed that the material modified using ultrasound for 100 min had the highest removal values (74.49%, 40.31%, and 51.50% for urea, creatinine, and uric acid, respectively), while unmodified zeolite removed 30.57%, 18.07%, and 22.84% of the same toxins. The best results for conventional heating modification were 67.08%, 31.97%, and 32.39%, respectively. Therefore, acid group incorporation considerably improved the adsorption properties of the clinoptilolite. Regarding adsorption kinetics, it was found that the pseudo-second-order model better described the behavior of all the modified materials. Equilibrium adsorption data were adjusted to the Langmuir and Freundlich models. The Freundlich model (multilayer adsorption) described urea adsorption, while the Langmuir model (monolayer adsorption) described creatinine and uric acid.

1. Introduction

Chronic kidney failure (CKD) is a major cause of death around the world. CKD is a disorder caused by partial or total damage to kidney function, resulting in the inability to excrete toxic metabolic products such as urea, creatinine, and uric acid, among others [1].

Dialysis is a common process used to extract these toxins from organisms. It is important to highlight that due to the increase in patients with CKD in recent years, researchers have seen the need to investigate, innovate, and explore new techniques, methods, and materials to increase the adsorption of uremic toxins during the dialysis process as well as reduce the toxin concentrations in wastewater [2].

In recent years, membranes, nanoporous materials, and polymeric composites have attracted attention and have been important in toxin removal through dialysis processes. The current research focuses on modifying such materials to increase the adsorption efficiency of uremic toxins [3].

From a large number of materials, different types of zeolites, such as kaolinite [4], cancrinites [5], and phillipsite [6], have been used for the removal of uremic toxins.

An example of this is the research carried out by Lu et al. [7], who fabricated and characterized polyacrylonitrile (PAN), nanofiber membranes, and PAN–zeolite membranes using different zeolite concentrations (5–35 wt%). These authors investigated the creatinine adsorption in HSZ-series zeolite powders: 500-KOA (L), 720-KOA (Farrierite), 840-NHA (ZSM-5), and 940-HOA (Beta). They found that 940-HOA zeolite performed better with an adsorption of around 2.2 mg/g for a 5.7 mg/L creatinine solution (50 mmol/L) and 9.05 mg/g for a 22.70 mg/L solution (200 mmol/L). When the PAN–zeolite membrane was used, an increase in the adsorption was observed (25.42 mg/g for a 70.7 mg/L (625 mmol/L) solution). Also, they reported that 0.025 g of 940-HOA zeolite can adsorb 91% of the creatinine from a 200 mmol solution in five minutes.

Another work, carried out by Koubaissy et al. [8], studied the adsorption capacities of dealuminated zeolites (Faujasite, HFAU, and Beta, HBEA) for the uremic toxins urea and creatinine at a solution volume/adsorbent weight of 1.33 at room temperature. They found that HFAU was more efficient in the adsorption of the toxins. An increase in the adsorbent dose was found to decrease the adsorption capacity (q_e). For creatine, the q_e = 71 mg/g for 15 mg of adsorbent while q_e = 50 mg/g for 30 mg of zeolite for a 340 mg/L solution (3 mmol). The authors explained that at low adsorbent doses, all types of sites are completely exposed, and the adsorption on the surface is saturated faster, resulting in a higher q_e value. For urea, q_e = 218 mg/g for the same solution concentration. This larger adsorption capacity is attributed to the larger size of creatinine, which means it cannot easily access the pores of the adsorbents.

Lu et al. [9] analyzed the influence of the size and shape of zeolites, used as filler in PAN membranes or alone as powders, on the adsorption capacity for uremic toxins. They tested spheric microparticles and spheric and rod-like nanoparticles. The adsorption capacities reported were 1.41, 1.62, and 1.85 mg/g, respectively, for 5.7 mg/L creatinine solutions, and they improved to 10.51, 10.82, and 10.44 mg/g for the 70.7 mg/L solution. When the micro-zeolites were incorporated inside the membranes, a significant effect of the size and shape of the zeolite on creatinine adsorption was observed. The microparticles performed better because the polymer fiber did not block the surface of the particles.

Some other materials were tested by Fu et al. [10], who studied the adsorptive removal of p-cresol and creatinine using polyethersulfone (MMM) mixed-matrix membranes, incorporating the adsorbents activated carbon (AC), zeolite ZSM-5, and graphene oxide (GO). They compared the adsorption capacity using the different prepared membranes. The authors found that for a 150 mg/L creatinine solution, the adsorption was 133.3 mg/g for the AC membrane, 87.7 mg/g for GO-MMM, and 3.9 μg/g for ZSM5-MMM, while for the adsorbents alone, the q_e values were 270.3, 263.2, and 6.8 mg/g for AC, GO, and ZSM-5, respectively. They attributed the reduction in adsorption ability of ZSM-5 to its small particle size (2 μm), which causes the ZSM-5 particles to be wholly embedded in the membrane matrix.

De Pascale et al. [11] developed cellulose acetate MMMs to eliminate uremic toxins from aqueous solutions. Various percentages of different sorbent particles (activated carbon, zeolite ZSM-5, and clinoptilolite ZUF) were tested. An HCl dealumination treatment activated the zeolites. They observed that the absorbents had different adsorption affinities for each toxin; activated carbon was the most suitable adsorbent for urea and creatinine, while ZSM-5 had the highest adsorption capacity for uric acid. For the batch adsorption test in mixed-matrix membranes, adding at least 20% of the sorbent particles improved the removal capacity for urea and creatinine. Still, no effect was noticed for uric acid. For the removal of urea and creatinine, an increase of 19% was obtained when ZUF was added. There was an increase of 16% when ZSM-5 was added and 28% in the case of AC.

Hsiao et al. [12] also developed cellulose acetate membranes using adsorbents activated carbon (AC), zeolite ZSM-5, and graphene oxide (GO) to remove p-cresol and creatinine. These authors found that the AC-MMM had better results, around 74% of p-cresol (initially 50 mg/L) and 7% of creatinine (initially 150 mg/L) within 4 h. For the zeolite, the removal was less than 5% of creatine and 27% of p-cresol.

Among the most recent techniques in the modification of materials is ultrasound energy. Different nanoparticles and micrometric-sized particles have been modified by ultrasound for adsorbents, whether for toxic or contaminating compounds [13]. The ultrasound technique is highly effective due to the cavitation phenomenon [14] and because it can activate the surface of the particles or generate vacancies [15] that promote a better interaction with other molecules and help the adsorption processes of different chemical compounds.

Over the years, countless advances have been made in the materials and techniques applied to remove uremic toxins; chemical modification has been demonstrated to increase the adsorption percentage. Also, ultrasound radiation, an innovative method used in the modification of materials, has allowed improvements in chemical modification.

However, it is necessary to continue looking for economical options that increase the removal of toxins. To test low-cost materials, a natural zeolite (clinoptilolite) was modified. It is an abundant mineral that can be easily extracted and is considered a renewable and economical resource because of these characteristics since other adsorbents require an expensive synthesis process.

The zeolite was chemically modified by conventional heating and ultrasonic radiation, an emerging activation energy that helps a chemical reaction occur in less time, allowing a homogeneous dispersion of particles. Tartaric acid is used as a modifier for its application in the adsorption of uremic toxins (creatinine, urea, and uric acid), raising the possibility of using this designed material to produce a new filter medium in the healthcare sector.

The novelty is that this is the first study that compares two methods, conventional chemical modification and ultrasonic radiation, to introduce carboxylic groups from a modifier (tartaric acid) to the surface of a natural zeolite for its study as an adsorbent material for uremic toxins. Currently, nothing similar has been reported.

2. Materials and Methods

Natural zeolite was used, provided by the Mexican company Zeomex S. A. (San Luis Potosí, México), and was composed of 95% by weight of clinoptilolite. The oxides present in this natural zeolite were Al₂O₃ (7.91% by weight), SiO₂ (71.70% by weight), CaO (1.80% by weight), MgO (0.47% by weight), K₂O (0.78% by weight), and Fe₂O₃ (1.51% by weight), and had an average particle size of 2–4 µm. The tartaric acid was purchased from Analytyka^® (Monterrey, Nuevo León, México), the creatinine (anhydrous ≥ 98%) and uric acid (≥99%, crystalline) were provided by Sigma-Aldrich^® (St. Louis, MO, USA), and the urea (analytical grade) was supplied by Fagalab (Sinaloa, México). All materials were used as received. Distilled water was used in the experiment.

Chemical modification by conventional heating was performed according to the methodology reported by Cabello-Alvarado et al. [16]. Two different systems were tested (zeolites/acid mass ratio): 0.05:1 for Clinop/ATarC1 and 0.01:1 for Clinop/ATarC2. The modification was carried out in a flask where an aqueous acid solution (acid/water mass ratio 1:10) and unmodified clinoptilolite were added. Reflux heating was performed under gentle stirring for 24 h. After the modification time, the samples were filtered and washed with distilled water until a neutral pH was reached. The samples were dried in a conventional oven for 24 h at 40 °C.

Chemical modification by ultrasound treatment was performed in an Erlenmeyer flask by dispersing 1 g of zeolite in a solution of 10 g of tartaric acid in 100 mL of distilled water. For ultrasound treatment, a 25 mm diameter ultrasonic probe was coupled to a homemade ultrasonic generator with an output power of 750 W, a wave amplitude of 50%, and a variable frequency of 15–50 kHz. For safety reasons, all experiments were performed in a soundproof room. Three different ultrasound times were applied (25, 500, and 100 min). All treatments were performed at room temperature, and at the end of the reaction time, the samples were washed several times with distilled water to remove unreacted acid, filtered, and dried at 80 °C for 24 h.

Table 1. Identification of samples according to modification characteristics.

Sample Name	Modification Type	Modification Time	Modification Group Relation (Zeolite: Acid)
Clinop/ATarC1	Conventional heating	24 h	0.05:1
Clinop/ATarC2	Conventional heating	24 h	0.01:1
Clinop/ATar25US	Ultrasound	25 min	0.1:1
Clinop/ATar50US	Ultrasound	50 min	0.1:1
Clinop/ATar100US	Ultrasound	100 min	0.1:1

shows the identification of the samples obtained.

Fourier-transform infrared spectroscopy, ATR-FTIR, and thermogravimetric analyses (TGA) were performed to verify the incorporation of carboxyl functional groups from tartaric acid in clinoptilolite. ATR-FITR analysis was carried out in a Nicolet Magna 550 spectrometer (Thermo Scientific, Waltham, MA, USA), with 10 scanners and a resolution of 16 cm⁻¹, in the range of 400 to 4000 cm⁻¹. TGA was performed using a Dupont 951 Instrument (TA Instruments, New Castle, DE, USA) from 25–700 °C with a heating rate of 10 °C/min in a nitrogen atmosphere at a 50 mL/min flow.

For the removal analysis of urea, creatinine, and uric acid in clinoptilolite, non-modified and modified with the tartaric acid, a UV–vis spectrometer Thermo Scientific Genesys model 10-S was employed (Waltham, MA, USA). The following procedure was used. A total of 50 mL of an aqueous solution of urea, creatinine, or uric acid at an initial concentration of 160 mg/mL and 0.1 g of clinoptilolite were added in a beaker glass under agitation at 100 rpm at room temperature for 4 h (which is a similar duration of hemodialysis treatment). An aliquot was taken at different contact times: 0,15, 30, 45,60, 90, 120, 150, 180, 210, and 240 min, filtered, and read in the UV–vis spectrophotometer, obtaining the concentration values from a calibration curve. For equilibrium adsorption (24 h for contact time), toxin dissolutions with 160, 140, 120, 100, 80, 60, 40, and 20 mg/L were used.

Different concentrations (20, 40, 60, 80, 100, 120, 140, and 160 mg/L) were used to create the calibration curves for each compound. The concentration of uremic toxins in the solution was determined using the Beer–Lambert law; the maximum wavelength (λ_max) for urea, creatinine, and uric acid are 200, 220, and 293 nm, respectively.

The removal percentage was calculated according to the Equation (1):

$${\displaystyle \frac{C_{i0}-C_f}{C_i}}\times 100$$

(1)

where C_i and C_f are the initial and final concentrations, respectively.

The adsorption capacity in equilibrium, q_e, of the tested materials was calculated with Equation (2)

$$q_e={\displaystyle \frac{\left(C_0-C_{eq}\right)V}{m}}$$

(2)

where C_eq and V are the concentration at the equilibrium and volume of toxin solution (L), respectively, and m is the mass in mg of the absorbent.

The Langmuir and Freundlich models were used to describe the adsorption equilibrium data. The absorption at equilibrium data was fitted to evaluate both models, and the correlation coefficient (R²) was calculated using the trendline command in Microsoft Excel^® (2021 18.0).

The Langmuir isotherm was calculated by the Equation (3):

$${\displaystyle \frac{C_{eq}}{q_{eq}}}={\displaystyle \frac{C_{eq}}{q_{max}}}+{\displaystyle \frac{1}{K_L{q}_{max}}}$$

(3)

where q_max is the maximum adsorption capacity, and K_L is the constant obtained from the graph of C_eq/q_e against C_eq.

The Freundlich isotherm was calculated by the Equation (4):

$$ln{q}_e=ln{K}_F+\left({\displaystyle \frac{1}{n}}\right)ln{C}_{eq}$$

(4)

where K_F and 1/n are the Freundlich constants related to the adsorption capacity and n is the heterogeneity factor calculated by linearly plotting ln·q_e against ln·C_eq.

Also, the removal data for all the adsorbents used were fitted to the pseudo-first- and pseudo-second-order kinetics models; the correlation coefficient (R²) was calculated using the trendline command in Microsoft Excel^® (2021 18.0).

3. Results and Discussion

3.1. Zeolite Modification

Figure 1 shows the FT-IR spectra of the unmodified clinoptilolite modified with tartaric acid by conventional heating. A signal at 1100 cm⁻¹ was observed in all the samples, which is typical for clinoptilolite. For the modified samples, a signal at 3407 cm⁻¹ was observed, corresponding to the O–H bounding, and at 2500 cm⁻¹, a broadband characteristic of the COOH functional group. Both modified samples exhibited signals at 1370 cm⁻¹, corresponding to a tension band C=O (stretching vibration of COOH bounding from the carboxylic acid), and at 1410 cm⁻¹, corresponding to the COOH group stretching, a combination of OH flexion and the antisymmetric O–C–C stretching. Also, there was a signal around 1285 cm⁻¹, corresponding to a C–O tension, resulting from a stretching–bending of OH groups, which is a characteristic signal for COOH groups. Finally, a signal near 900 cm⁻¹ for both modified samples corresponds to a C–O–H flexion.

Furthermore, the FT-IR spectra of the modified clinoptilolite Figure 1 show the presence of characteristic bands of tartaric acid functional groups. The characteristic peak of clinoptilolite is broadened due to the interaction between the Si(Al)–O aluminosilicate network of the zeolite and the carboxylic groups of tartaric acid.

Figure 2 shows the FT-IR spectra of unmodified and modified clinoptilolite by ultrasound, Clinop/Atar25US, Clinop/Atar50US, and Clinop/Atar100US. Figure 2 presents one band in 1100 cm⁻¹, a result of the structural units of the aluminosilicate lattice Si (Al)–O of the zeolite. All ultrasound samples show a wide band at 3420 cm⁻¹, attributed to the O–H bond, and the Clinop/Atar50US and Clinop/Atar100US samples present a small band around 2160 cm⁻¹, corresponding to the COOH bond, which indicates that there was a chemical modification when the functional groups were found on the surface of the clinoptilolite.

The thermograms of the samples of the unmodified and modified zeolite by the conventional method are shown in Figure 3. Although functional groups corresponding to tartaric acid are observed in the FT-IR, it is not possible to detect the corresponding weight losses in the TGA. This is because the zeolite was not homogeneously modified (there are modified and unmodified materials) since mechanical stirring does not promote a good dispersion or mixing and provides slow chemical kinetics.

The thermograms for the samples modified by the ultrasound technique are shown in Figure 4. A weight loss corresponding to humidity in the samples is observed for all the samples around 100 °C. Even though the percentage weight loss curve is not easily noticed, a change at 600 °C in the derivative curve is observed, which can be attributed to the oxygen loss from the clinoptilolite. For the modified samples, a weight loss mass in the temperature interval of 295 to 580 °C can be attributed to the tartaric acid decomposition [17], confirming the modification of the zeolite with the acid groups.

Table 2. Thermogravimetric analysis results for samples modified by ultrasound.

Label	Water (wt%)	The Initial Temperature of Acid Group Decomposition (°C)	The Final Temperature of Acid Group Decomposition (°C)	Tartaric Acid Incorporation (wt%)	Total Weight Loss (%)
Clinoptilolite	3.76	It does not have	It does not have	It does not have	10.24
Clinop/ATar25US	1.06	298	578	12.79	15.41
Clinop/ATar50US	1.34	298	580	15.94	19.39
Clinop/ATar100US	1.29	299	580	18.42	21.63

summarizes the results of the analysis for ultrasound modification. As can be noticed, the sample Clinop/ATar100US had a higher loss in weight, indicating a significant incorporation of the acid functional groups.

3.2. Textural Properties

The results of the textural properties of pristine clinoptilolite and modified clinoptilolite by conventional heating and ultrasound treatment, including the surface area and pore volume, are summarized in

Table 3. Textural parameters of clinoptilolite and modified clinoptilolite.

Material	BET Surface Area m2/g−1	Pore Volume cm3/g−1
Clinoptilolite	50.83	0.19
Clinop/ATarC1	62.31	0.21
Clinop/ATarC2	63.58	0.20
Clinop/ATar25US	83.91	0.27
Clinop/ATar50US	85.30	0.30
Clinop/ATar100US	121.46	0.32

. The surface area and pore volume increased after modification with conventional heating and ultrasound treatment with tartaric acid. These results are in agreement with those of Akyalcin et al. They modify the zeolite with 0.5 M HNO₃ at 80 °C for 2 h [18].

3.3. Adsorption Results

A comparison of the adsorption results for urea and creatinine, using as adsorbent unmodified and modified samples by ultrasound and conventional heating, are presented in Figure 5. It was observed that Clinop/ATar100US had the greatest removal value for urea, 74.5%. Clinop/ATar50US had an adsorption value of 71%, less than Clinop/ATar100US. These results suggest that after 50 min of ultrasound modification, there is no significant difference in acid group incorporation at the clinoptilolite surface.

For the modified samples by conventional heating, Clinop/ATarC2 had a urea removal percentage of 67.10–7.40% lower than Clinop/ATar100US—while Clinop/AC1 removed 62.15%, which was 4.95% less than Clinop/ATarC2. This demonstrates that the modification of clinoptilolite with tartaric acid amount influences the number of functional groups incorporated. A lower zeolite quantity implies less competition for the functionalization of clinoptilolite sites. This allows the free acidic groups dispersed in an aqueous modification solution to adhere to the clinoptilolite surface. However, as can be observed in Figure 3, the removal percentage is similar; hence, increasing the zeolite amount permits the use of the acidic groups available in a better way. Unmodified clinoptilolite reached 30.60% for urea removal; therefore, the surface modification of zeolite improves urea adsorption. On the other hand, the sample Clinop/ATar25US presents removal percentages of 40% and 17% for urea and creatinine, respectively. This may be because the degree of modification is lower than in other materials; therefore, the percentages of removal of uremic toxins are lower.

To consider a material as a good adsorbent, the urea reduction must be at least 60% [12]. In this research, the removal percentages reached are 74.50, 71, 67.10, and 61.15 for the materials Clinop/ATar100US, Clinop/ATar50US, Clinop/ATarC1, and Clinop/ATarC2, respectively, indicating that these materials are a good candidate for urea removal.

Concerning creatinine, the Clinop/ATar100US sample had the best removal value, 40.30%. Clinop/ATarC2 had a better performance for conventional heating modification, removing 31.97% of the creatinine. As it occurred for urea, there is not a big difference, 0.7%, in the removal results between Clinop/ATar100US and Clinop/ATar50US, confirming that extending the modification time beyond 50 min does not significantly influence the removal.

The adsorption kinetics using unmodified clinoptilolite as an adsorbent shows a better fit to the first-order model (R² = 0.997), indicating that urea adsorption is carried out by physical adsorption. On the other hand, when modification is performed, whether by conventional heating or ultrasound, the adsorption data fits better with the second-order model, indicating chemical adsorption when there are acid groups. Physisorption allows the toxin to maintain its chemical nature because only van der Waals forces are responsible for binding the urea to the zeolite surface, explaining why the percentage of adsorption is lower in comparison with the results obtained when the modified zeolite is used as an adsorbent. Instead, when chemisorption occurs, the bonds formed are stronger, and the desorption of urea is avoided during agitation.

Table 4. Parameters for the pseudo-first- and second-order models adjusted, and correlation coefficients are calculated for the kinetics adsorption of urea.

Material	Pseudo-First-Order			Pseudo-Second-Order
	qe (mg/g)	k1 (1/min)	R2	qe (mg/g)	k2 (g·mg−1·min−1)	R2
Clinoptilolite	47.9844	0.0111	0.9978	66.2252	0.00017	0.9936
Clinop/ATarC1	69.3426	0.0106	0.8873	114.9425	0.00024	0.9908
Clinop/ATarC2	89.7016	0.0122	0.9675	133.3333	0.00013	0.9880
Clinop/ATar25US	38.3707	0.0159	0.9195	72.9927	0.00064	0.9965
Clinop/ATar50US	96.6941	0.0233	0.9726	133.3333	0.00018	0.9794
Clinop/ATar100US	105.1962	0.0246	0.9713	136.9863	0.00021	0.9867

and

Table 5. Parameters for the pseudo-first- and second-order models adjusted, and correlation coefficients calculated for the kinetics adsorption of creatinine.

Material	Pseudo-First-Order			Pseudo-Second-Order
	qe (mg/g)	k1 (1/min)	R2	qe (mg/g)	k2 (g·mg−1 min−1)	R2
Clinoptilolite	28.2228	0.0081	0.9801	38.0228	0.00031	0.995
Clinop/ATarC1	24.5132	0.0154	0.9206	48.7805	0.00120	0.9985
Clinop/ATarC2	25.6094	0.0147	0.8993	52.9101	0.00113	0.9978
Clinop/ATar25US	25.4624	0.0101	0.9592	37.1747	0.00057	0.9941
Clinop/ATar50US	46.3020	0.0182	0.9369	74.0741	0.00031	0.9705
Clinop/ATar100US	44.0961	0.0147	0.9192	75.1880	0.00032	0.9751

summarize the results for the fitted adsorption data to the first- and second-order models for urea and creatinine, respectively. As it can be observed, the value for the adsorption velocity constant for urea, k₂ (g mg⁻¹min⁻¹), is 1.3 × 10⁻⁴, 1.8 × 10⁻⁴ y 2.1 × 10⁻⁴ for Clinop/ATarC2, Clinop/ATar50US, and Clinop/ATar100US, representing a higher velocity in adsorption for Clinop/ATar100US—reaching the equilibrium faster in comparison with the other materials; moreover, the adsorption capacity in equilibrium, q_e, is higher, indicating a greater capacity for adsorption for urea. Although Clinop/ATarC2 and Clinop/ATar50US have the same theoretical adsorption ability at equilibrium (133.33 mg g⁻¹), it is observed that the material modified by ultrasound has a higher velocity of adsorption, indicating a higher effectiveness in adsorption. Considering the adsorption results, it can be concluded that Clinop/ATar100US is the best adsorbent among all the materials tested.

The adsorption process for creatinine is described by the pseudo-second-order kinetic model for all the materials used as adsorbents, including the non-modified zeolite. The higher adsorption velocity was achieved for the samples modified by conventional heating, Clinop/ATarC1 and Clinop/ATarC2, k₂ = 1.20 × 10⁻³ and 1.13 × 10⁻³ g mg⁻¹ min⁻¹, respectively, indicating that equilibrium was reached faster. Clinop/ATar100US shows the highest removal percentage; however, it is one of the slowest to reach equilibrium, almost as it occurs for unmodified clinoptilolite. Nevertheless, even when Clinop/ATar100US delays in reaching equilibrium, it is the best material for creatinine adsorption. Moreover, in practice, it is necessary for the toxins to be in contact with the removal material only for four hours of dialysis time, which Clinop/ATar100US demonstrated to have one of the best removal percentages for creatinine. Comparing the two best removal materials, Clinop/ATar50US versus Clinop/ATar100US, the last one reaches the equilibrium adsorption faster. Considering the kinetics and equilibrium results, the best material for removing creatinine is Clinop/ATar100US.

Adsorption for uric acid was performed using just the modified samples that presented the highest removal percentages for urea and creatinine, Clinop/ATar100US, Clinop/ATar50US, and Clinop/ATarC1, as well as unmodified clinoptilolite. Figure 6 shows the removal results in acid uric using unmodified and modified clinoptilolite as an adsorbent. It can be noticed that Clinop/ATar100US exhibits the best removal capacity, obtaining 51.50% of adsorption, which is 6.65% higher than Clinop/ATar50US. The sample modified by conventional heating removed 32.39%.

Table 6. Parameters of the pseudo-first- and second-order models adjustments and correlation coefficients calculated for the kinetics adsorption for uric acid.

Material	Pseudo-First-Order			Pseudo-Second-Order
	qe (mg/g)	k1 (1/min)	R2	qe (mg/g)	k2 (g·mg−1 min−1)	R2
Clinoptilolite	29.0202	0.0101	0.9734	41.6667	0.00062	0.9989
Clinop/ATarC2	42.0146	0.0104	0.9817	59.1716	0.00041	0.9987
Clinop/ATar50US	62.5893	0.0177	0.9864	80.0000	0.00038	0.9994
Clinop/ATar100US	62.5893	0.0177	0.9864	90.0901	0.00044	0.9995

summarizes the kinetic parameters obtained for the lineal adjustment to the experimental data for the acid uric adsorption using the pseudo-first- and second-order models. Notice that, for all the samples, a better adjustment is obtained with the second-order model, indicating chemical adsorption. The highest value k₂ = 4.4 × 10⁻⁴ (g·mg⁻¹min⁻¹) was calculated for the unmodified clinoptilolite, showing that the equilibrium in adsorption is reached faster. The modified material Clinop/ATar100US presents the highest value in removal, and the equilibrium is reached faster. As can be seen, the removal percentage for uric acid is lower than for urea. It can be attributed to the molecular size of uric acid, as it occurred for creatinine, which is bigger than that for urea, causing it not to anchor properly and desorbs easily during agitation, explaining the lower removal for uric acid.

The kinetic models allow us to obtain the general behavior of adsorption in a solid material. However, to widen the information about the mechanisms of adsorption that occur during the urea removal, the homogeneous solid diffusion model (HSDM) will be used, which describes mass transfer in an amorphous and homogeneous sphere [19].

Figure 7 shows the fit of the experimental data to the homogeneous solid diffusion model for urea adsorption. It can be observed that the regression lines deviated from the origin of the coordinate system, which can be attributed to the difference in the mass transfer velocity in the initial and final adsorption steps. The double linear region indicates that many stages conditionate the kinetics of the adsorption process, which must be analyzed individually. The interpretation of the literature for these regions is confusing. Dogan and Alkan [20] attribute the initial stage of the curve to the diffusion in the macropores, while Kumar and Porkodib [21] explain that such a part informs about the diffusion process in the boundary layer. Therefore, it is considered that stage I, consistent with minor time, corresponds to the adsorption in the zeolite external surface, including macro and mesopores, until this surface reaches saturation. Stage II corresponds to micropore diffusion; this is the real internal diffusion. From this moment, the urea begins to penetrate the least accessible pores, increasing the resistance to mass transport and decreasing the diffusion velocity. Therefore, k_dI > k_dII means that as the path available to the diffusion is reduced because of the obstruction of some pores, the magnitude of the diffusion parameters is reduced [20].

Table 7. Results for the experimental kinetic data fitting with the homogeneous solid diffusion model for urea, creatinine, and uric acid adsorption.

Urea
Material	Region I							Region II
	kd, I (mg·g−1 h−0.5)		c			R2		kd, II (mg·g−1 h−0.5)			c		R2
Clinoptilolite	35.396		−8.1384			0.9842		17.918			12.957		0.9779
Clinop/ATarC1	85.627		−12.677			0.9728		10.414			76.922		0.9171
Clinop/ATarC2	88.678		−19.176			0.9865		22.232			61.713		0.9178
Clinop/Atar25US	52.399		1.9553			0.9266		8.3644			50.658		0.8304
Clinop/ATar50US	113.17		−28.357			0.9757		7.2905			95.658		0.9962
Clinop/ATar100US	113.17		−22.907			0.9757		7.2905			101.21		0.9952
Creatinine
Material	Region I							Region II
	kd,I (mg·g−1 h−0.5)		c			R2		kd,II (mg·g−1 h−0.5)			c		R2
Clinoptilolite	21.460		−5.3115			0.9889		9.6452			8.9725		0.9871
Clinop/ATarC1	42.525		−1.179			0.9910		4.4851			36.593		0.9695
Clinop/ATarC2	48.671		2.9398			0.9792		4.7471			40.033		0.9309
Clinop/Atar25US	30.620		−7.6442			0.9969		8.0131			15.296		0.9772
Clinop/ATar50US	81.793		28.926			0.9932		6.5680			48.138		0.9881
Clinop/ATar100US	81.465		28.404			0.9862		7.2905			95.764		0.9962
Uric acid
Material	Region I				Region III					Region II
	kd,I (mg·g−1 h−0.5)	c		R2	kd,II (mg·g−1 h−0.5)		c		R2	kd,II (mg·g−1 h−0.5)		c		R2
Clinoptilolite	25.517	0.1311		0.9997	16.33		8.19		0.9975	7.6666		20.867		0.9601
Clinop/ATarC2	35.039	0.2048		0.9993	24.079		8.8998		0.9977	11.697		27.765		0.9668
Clinop/ATar50US	51.775	0.9686		0.9995	32.062		18.032		0.9997	11.607		47.335		0.9668
Clinop/ATar100US	65.551	0.2116		0.9	34.147		28.567		0.9685	8.5208		63.886		0.9819

summarizes the results of the experimental data that fit the homogeneous solid diffusion model for urea adsorption. Following them, region II is the stage that limits the process since parameter c has higher values; however, as can be observed in Figure 6, the trend lines do not pass at the origin of the coordinate system, indicating that intraparticle diffusion is not the limiting stage and that the adsorption on the zeolite surface controls the adsorption kinetics. However, if the external diffusion limits the process, the adsorption equilibrium would be achieved in a short time (<10 min). Nevertheless, more than 240 min are required for this to happen. Therefore, these results suggest that the mechanism for the urea removal in the materials tested here is a complex process and that intraparticle diffusion and diffusion in the boundary contribute to the limiting stage.

The Clinop/ATar50US and Clinop/ATar100US samples present a higher velocity for adsorption in stage I (kdI), which can be the reason for a higher percentage of removal in a minor time. For stage II, the Clinop/ATarC2 sample presents a higher velocity for internal diffusion, explaining the increasing behavior in the adsorption curves.

As it occurs for urea, region II is the limiting stage (higher values for parameter c), while for uric acid, three regions are observed, where the third region corresponds to the limiting stage; however, again, the trend lines do not cross at the origin of the coordinates system in none of the cases. Therefore, the acid uric and creatinine removal mechanism is complex, and all stages limit the adsorption process. Also, Clinop/ATar50US and Clinop/ATar100US have a higher adsorption velocity (kd) in stage I, explaining the minor removal times.

Notice that the urea adsorption using clinoptilolite without modification as an adsorbent has a better fit with the Langmuir model, indicating a monolayer formation (see

Table 8. Parameters obtained by the experimental data fit the Langmuir and Freundlich models.

Urea
Material	Langmuir			Freundlich
	qmax (mg/g)	KL (L/mg)	R2	1/n	KF (L/g)	R2
Clinoptilolite	58.8235	0.0670	0.9942	0.3818	9.43192	0.9375
Clinop/ATarC1	192.3077	0.0239	0.9560	0.6757	7.62781	0.9971
Clinop/ATarC2	208.3333	0.0261	0.9061	0.6713	8.91046	0.9970
Clinop/ATar25US	90.9091	0.0358	0.9642	0.4607	8.95365	0.9952
Clinop/ATar50US	166.6667	0.0404	0.9040	0.5821	11.95639	0.9899
Clinop/ATar100US	156.2500	0.0727	0.9393	0.5026	18.78019	0.9975
Creatinine
Material	Langmuir			Freundlich
	qmax (mg/g)	KL (L/mg)	R2	1/n	KF (L/g)	R2
Clinoptilolite	35.2113	0.0919	0.9937	5.5617	13.58626	0.9789
Clinop/ATarC1	50.2513	0.0737	0.9665	5.5432	18.40772	0.9423
Clinop/ATarC2	53.7634	0.1177	0.9802	7.6394	25.98364	0.9572
Clinop/ATar25US	36.7647	0.0697	0.9780	5.4318	13.29842	0.9526
Clinop/ATar50US	68.4932	0.0794	0.9638	5.9952	26.21803	0.8880
Clinop/ATar100US	70.9220	0.0698	0.9394	5.9277	26.19389	0.8737
Uric acid
Material	Langmuir			Freundlich
	qmax (mg/g)	KL (L/mg)	R2	1/n	KF (L/g)	R2
Clinoptilolite	40.8163	0.0911	0.9841	5.0684	14.62177	0.9732
Clinop/ATarC2	58.1395	0.0975	0.9845	3.9370	16.72245	0.9747
Clinop/ATar50US	74.6269	0.1110	0.9823	3.6955	20.62529	0.9797
Clinop/ATar100US	74.6269	0.1110	0.9823	3.6955	20.62529	0.9797

). This behavior increases the number of molecules at the zeolite surface until a limit value is reached, corresponding to a monolayer covering. Monolayer formation is a characteristic of a chemisorption process, which is supported by the kinetics results (second-order kinetic model).

For modified clinoptilolite, by ultrasound treatment and conventional heating, a better data fit was obtained with the Freundlich model (see Table 8), suggesting multilayer formation and a heterogenous surface in the clinoptilolite. These results contradict the results obtained by the kinetic models because, to obtain a multilayer formation, a physical adsorption process is necessary, which was not predicted by the kinetics analysis. Linares et al. [6] researched the adsorption process of urea using philipsite zeolite as an adsorbent and found a similar behavior. They explained that a monolayer (corresponding to chemical adsorption) is formed at the beginning. Still, when the surface is saturated, multilayer adsorption occurs because new layers of solute are physically bonded and formed over the initial chemical layer produced. The adsorption mechanism can be established by the hydrogen bonds from the urea molecules and the oxygens coming from the crystalline network of the zeolite.

The parameter 1/n, which measures the surface heterogeneity, indicates that a minor value to one is characteristic of the systems that predominate multilayer formation (Freundlich model). Therefore, it is possible for the urea multilayer formation over the clinoptilolite because the values for such parameters are minor for all the samples.

For creatinine and uric acid adsorption, notice that for all the samples, it was obtained better with the Langmuir model, indicating a monolayer formation and a homogeneous surface for the zeolite. Monolayer formation is a characteristic process for chemisorption, which coincides with the results obtained by the kinetic analysis. This is when the amount of the toxin molecules increases until a limit value is reached, forming a monolayer. This behavior has also been reported for other materials [21,22,23,24,25]. However, the value for parameter n suggests that multilayer formation is possible. Nevertheless, a monolayer formation predominates because a better adjustment of the experimental data is obtained with the Langmuir model.

Figure 8 shows the adsorption capacity in function of the concentration of the toxin solutions that reached equilibrium. As can be observed, the samples Clinop/ATar100US and Clinop/ATar50US have a higher adsorption capacity for all the cases. Moreover, all the curves have a sigmoidal behavior (s shape), classified as type 2, according to Giles et al. [26]. Linares et al. [6] explained that this behavior is due to cooperative adsorption; as the solute concentration increases, the adsorption in the solid surface (adsorbent) increases. This indicates that the previously adsorbed molecules attract the other molecules of solute, and as time elapses, the solute layer formed becomes more stable. This can be confirmed with the values of q_e obtained because a higher concentration of toxin results in a significant capacity of adsorption for all the materials tested in this research. In conclusion, a multilayer formation is possible, starting as a monolayer formed by chemisorption, indicating that the adsorption energy for the active sites is homogeneous, and with the subsequent addition of solute molecules by physical adsorption (attracted by themselves).

Even though the removal results obtained in this work are minor in comparison with those obtained with other materials reported in the literature (

Table 9. Comparison of adsorption results for urea, creatinine, and uric acid.

Material	Uremic Toxins (Adsorption, %)			Adsorption Isotherm			Reference
	Urea	Creatinine	Uric Acid	Urea	Creatinine	Uric Acid
ZnFe2O4	84	77	21	Langmuir	Langmuir	Freundlich	[27]
Zn0.5 Mg0.5 Fe2O4	85	77	20	Langmuir	Langmuir	Freundlich	[27]
ZnO/Activated carbon			90			Langmuir	[23]
MCB/Nylon 6			78–82			Freundlich	[28]
NB20		86					[29]
Kaolinite	34.8			Freundlich			[4]
Clinoptilolite	30.57	18.07	22.84	Langmuir	Langmuir	Langmuir	This work
Clinop/ATarC2	67.88	31.97	32.39	Freundlich	Langmuir	Langmuir	This work
Clinop/ATar100US	74.49	40.31	51.50	Freundlich	Langmuir	Langmuir	This work

), the modified zeolite has the advantage of being antimicrobial (characteristic given by the modification), economical and biodegradable, which allows it to compete with the other materials.

Figure 9 shows the scheme of the modification of clinoptilolite by two methods: conventional heating and ultrasound. Based on the results of FT-IR and TGA, functional groups such as carbonyl and hydroxyl are found on the surface. Weight loss attributed to organic compounds was also detected on the surface. Possibly, there is a link with clinoptilolite through hydrogen bonding electrostatic interactions between the carboxylic acids of tartaric acid and Si⁺, K⁺, Al³⁺, and Ca²⁺.

In the same figure, the adsorption scheme is observed. The adsorption of uremic toxins on the modified clinoptilolite can occur through physisorption, chemisorption, or both. This process can originate from the hydrogen bonding between the hydroxyl group from the acid modification and the group amino of the uremic toxins. Also, hydrogen bonding can occur with an increasing Si/Al molar ratio, which is achieved by dealumination originating from the acid treatment. Moreover, an electrostatic interaction can occur between the amino group and the Si⁺, K⁺, and Ca²⁺ ions in the zeolite, while physisorption occurs in the material pores.

4. Conclusions

Clinoptilolite (natural zeolite) was modified using tartaric acid by ultrasound treatment or conventional heating. It was found that for conventional heating, the relation g_zeolite/g_acid influences the modification results; as the relation increases, the removal percentage increases even in 5.08% for urea and 2.78% for creatinine. For the ultrasound technique, it was observed that a higher modification time increases the adsorption capacity of toxins. Modification by the US technique is more effective than conventional heating because the highest removal percentages were obtained for the removal of urea, creatinine, and uric acid (74.49%, 40.31%, and 51.50%, respectively) in solutions with an initial concentration of toxin of 160 mg/L. In comparison, the removal values for unmodified clinoptilolite were 30.57%, 18.07%, and 22.84% for the same toxins. These results demonstrate that incorporating acid groups into the surface of the zeolite increases the adsorption properties of the clinoptilolite. The experimental kinetic data fitting showed that the pseudo-second-order model better describes the adsorption process for all the toxins; therefore, chemisorption occurs. Regarding the equilibrium adsorption data adjustment, a better fit for urea was found with the Freundlich model, while for creatinine and uric acid, the Langmuir model describes the process, indicating a monolayer formation in the zeolite surface. These new materials are an alternative to add to a polymeric membrane for hemodialysis treatment.