Adsorption of Heavy Metals from Small-Scale Gold Processing in Baguio Mining District, Philippines

Small-scale mining in the Philippines lacks research and technologies for the application of an inexpensive and environmentally sound treatment of its wastewaters. Hence, laboratory-scale fixed-bed column tests were carried out to test and evaluate the sorbent adsorption performance of the Philippine Natural Bentonite (PNB) and the Philippine Natural Zeolite (PNZ) against the heavy metals present in the ball mill facilities within the Baguio Mining District. Results showed that removal percentages of PNB and PNZ in terms of lead (Pb) concentrations are 50.43% and 26.90%, respectively. Moreover, the dynamic uptake capacity and maximum adsorption capacity values of PNB showed greater values than PNZ and the sorption models, Bohart-Adams, Wolborska, Thomas, and Yoon-Nelson models, displayed good fit for the breakthrough curves. Furthermore, varying the flow rate and bed height manifested that an increase in flow rate yields a lesser saturation time while the rise in bed height increases uptake capacity, saturation time, and removal percentage of the system. Regarding the cost efficiencies of the sorbents, the study revealed that PNB is more cost-efficient than PNZ since only 1.13 Philippine pesos of PNB is required to remove a gram of Pb compared to PNZ, which costs 3.5 pesos. Therefore, this research concluded that PNB could be a better sorbent of Pb in terms of removal percentage and cost-efficiency. The study recommends further regeneration studies for the reusability of the sorbents after exhaustion and hopes that this research will be a springboard for further environmental studies. Philippine Natural Bentonite yields higher adsorption than Philippine Natural Zeolite. Increasing the inlet flow rate resulted in the decreased saturation time of the sorbents. Increasing the bed height results in increased uptake capacity and saturation time. Philippine Natural Bentonite is more cost-efficient adsorbent for lead than Philippine Natural Zeolite. Philippine Natural Bentonite yields higher adsorption than Philippine Natural Zeolite. Increasing the inlet flow rate resulted in the decreased saturation time of the sorbents. Increasing the bed height results in increased uptake capacity and saturation time. Philippine Natural Bentonite is more cost-efficient adsorbent for lead than Philippine Natural Zeolite.


Introduction
The Philippines is a significant gold producer which ranks 24th in gold production in 2018 (Rashotte 2019). The majority of the gold production is in the five well-known gold mining districts, namely the districts of Baguio and Paracale, which are located on the island of Luzon, while the remaining three: Masara, Surigao, and Masbate, are situated in the Mindanao islands. These mining locations comprise more than 60% of gold produced, all of which are smallscale mining operations (Simeon 2019). The Republic Act 7076, also known as the People's Small-Scale Mining Act of 1991, has allowed these small-scale mines (SSM) to be operated in areas declared by the law as "Minahang Bayan or Peoples Small-Scale Mining Area". Although lawfully recognized, a great majority of SSM operations have been running illegally for decades, wherein licenses are only temporary permits that are valid for less than a year. Most miners practice gold extraction processes using harmful chemicals such as mercury and cyanide due to their high gold recovery 1 3 71 Page 2 of 14 and low-cost characteristics resulting in a widespread and unregulated SSM operation that dumps untreated toxic wastewater directly into the environment (Mendoza 2014).
The non-compliance of SSM operators to environmental laws and regulations is one of the main reasons the approval rate of permits is very low and is the biggest obstacle in obtaining an Environmental Compliance Certificate (ECC) (Simeon 2019). In SSM, the waste disposal system is one of the major issues impeding the approval of the eleven applications for SSM permits in Tuba, Benguet. The study of Macasieb et al. (2014) detected traces of heavy metals such as arsenic (As), mercury (Hg), and lead (Pb) in the sediments of Bued River, which primarily covers the provinces of Benguet, Pangasinan, and La Union. Despite the consequences of improper wastewater disposal, which causes pollution to the Bued river and its tributaries, small-scale miners continue their gold processing operations. The need for intervention for this kind of undertaking is paramount for environmental protection.
Over the years, numerous treatment processes were developed to remove heavy metals from polluted waters. These include reverse osmosis, ultrafiltration, ion exchange, and electrodialysis; however, these technologies are costly, which most SSM industries cannot afford. One of the least expensive methods is adsorption, which depends on the sorbent material used (Acheampong et al. 2010), such as the naturally occurring clay minerals and zeolite. These minerals have small particle sizes and are mainly composed of silica, alumina, water, and weathered rocks. The adsorption capabilities of clay minerals and zeolite are owed to their complex porous structures and high specific surface area, allowing physical solid and chemical attractions with species dissolved in solutions such as heavy metals (Uddin 2017).
Despite the similarities, zeolites and clay minerals exhibit distinctive properties due to the difference in crystalline structures. The similarities and differences apply to many clay minerals and zeolites, but in particular, natural zeolites and natural bentonites are compared and contrasted for this discussion. The adsorption performance of bentonite is attributed to its predominant composition of clay minerals belonging to the smectite group, usually montmorillonite. Smectites are hydrous aluminosilicates with a 2:1 Si/ Al ratio structure of interlayers giving the bentonite a high specific surface area, porosity, water, compound retention capability, and cation-exchange capacity. In aqueous solutions, bentonite can exchange cations with those present in the solution (Belbachir and Makhoukhi 2017). Zeolite, on the other hand, is a group of aluminosilicate minerals. Its crystalline structure consists of tectosilicate interconnected tetrahedra of silicon, aluminum, and oxygen atoms. The bonding of the tetrahedra to common oxygen atoms prevents dense packing of the layers, leaving open pores or cavities in its framework. The open cavities in zeolite are often occupied by H 2 O molecules and cations that are commonly exchangeable (Ramesh and Reddy 2011). The isomorphous substitution between the Si 4+ and Al 3+ ions in bentonite and zeolite crystal lattices results in a negative net charge. This negative charge is neutralized by the exchangeable cations such as Na + , K + , and Ca 2+ . These cations can be exchanged with heavy metal ions during adsorption. The released ions from the exchanging of cations in the bentonite and zeolite adsorption are relatively safe, which is advantageous in choosing these materials in wastewater treatment for heavy metals (Olegario-Sanchez and Pelicano 2017).
The Philippines is known to bear mineralization of clay minerals zeolite and bentonite, which are found to adsorb heavy metals, as in the study of Gili et al. (2019). Zeolite and bentonite are locally found in regions of Pangasinan (Region I), and Albay (Bicol) (PSA 2013). The Philippine natural bentonite (PNB) is predominantly composed of mordenite, hectorite, and montmorillonite with kaolinite, feldspar, mica, and nontronite traces. It is chemically composed of 47.9% SiO 2 and 14.02% Al 2 O 3 (Olegario and Gili 2021) . On the other hand, the Philippine natural zeolite (PNZ), comprised mainly of mordenite, clinoptilolite, and chabazite, has a chemical composition of 73.20% silica and 10.72% alumina, with a calculated Si/Al ratio of 3.67, which is higher than the Si/Al ratio of PNB (Olegario-Sanchez and Pelicano 2017). According to literature, these adsorption materials, specifically the use of natural zeolite and bentonite, have not been fully employed in the Philippine Small-Scale Mining operations. Hence, this study aims to evaluate the adsorption capability, removal efficiency and cost-effectiveness of natural zeolite and bentonite as an alternative wastewater treatment method to remove heavy metals found in SSM wastewater so that the Minahang Bayan or the so-called small-scale mining areas will have a higher approval rate resulting from an improved wastewater treatment method.
In current adsorption studies, column experimental data were fitted to several mathematical models where design parameters may be determined. These parameters may be used in the initial stages of scaling up for industrial use of the sorption column. In similar studies, the most applicable models are (i) Bohart-Adams, (ii) Wolborska, (iii) Thomas, and (iv) Yoon-Nelson (Chittoo and Sutherland 2020;Khalfa et al. 2020). Bohart-Adams model was used to represent the portion of the breakthrough curve where C/C o is less than 50% (Calero et al. 2009). This model only describes the initial portion of the breakthrough curves. It is, however, used to determine model parameters such as the volumetric adsorption capacity and rate constants (Khalfa et al. 2020). The Wolborska model, on the other hand, is used to describe the lower concentration regions of the breakthrough curves (Chittoo and Sutherland 2020;Trgo et al. 2011). Thomas model ignores the effect of intraparticle mass transfer and external film resistances where the assumption states that the rate of adsorption is driven by surface reactions of the available binding sites and the sorbate (Han et al. 2009). As for the Yoon-Nelson model, it does not emphasize the importance of the characteristics of the adsorbate and adsorbent but rather the relationship of probabilities of adsorption of the sorbate molecules onto the sorbents (Aksu and Gönen 2004;Ayoob and Gupta 2007).

Study Area and Wastewater Sampling
Wastewater sampling was conducted in Barangay Loacan, Itogon, Benguet among the SSM processing facilities under Loacan, Itogon Pocket Miners Association (LIPMA) since the area is declared as a Minahang Bayan in 2019. However, the facilities still practice direct dumping of untreated wastewater into the environment. Wastewater samples were collected from five disposal points of ball mills in the LIPMA area. A total of 10 L wastewater discharge was collected from a full 24 h-operation, not including cleaning activities and downtime. The water samples were then characterized using X-ray Fluorescence (XRF) Vanta M Series to determine the heavy metals present in the wastewaters. According to Zhou et al. (2018), the XRF device is an effective alternative and portable.

Preparation of Sorbents
The PNB and PNZ materials were acquired from LITHOS Manufacturing, which was then ground with a ceramic mortar and pestle and sieved to 0.25-0.40 mm particles using stainless steel test sieves. No further treatment was applied to maintain the as-purchased properties of the sorbent materials. The samples of the PNB and PNZ were again analyzed for their chemical composition using the Vanta M Series XRF Analyzer in the Department of Chemical and Mining Engineering, Saint Louis University, Baguio City. The water samples and sorbents were analyzed in triplicate.

Experimental Set-up
Adsorption was studied by employing a laboratory-scale fixed-bed column setup for natural zeolite and natural bentonite, as shown in Fig. 1. A fixed-bed column setup is preferred because the effect of operational parameters such as the amount of sorbent and feed flow rate may be investigated, compared to batch sorption tests that only analyze a sorbent's adsorption capacity (Sazali et al. 2020). A standard borosilicate glass chromatography column of 11 mm inner diameter and 200 mm high was used to house the media sorbent. The bottom of the column includes a sintered glass disk to support the adsorbent bed and minimize the loss of adsorbent mass. A slurry of the sorbents was then added to the glass column to achieve uniform packing. The stationary phase (sorbent) was allowed to settle to a specific depth. Then, the mobile phase (solvent) was drained only up to the level of the stationary phase to avoid displacement of the packed sorbent when the feed water comes into the system. A peristaltic dosing pump with an adjustable flow rate was used to maintain the desired feed flow rate. The experiment was held at room temperature 22-23 °C.
The effects of the process parameters were studied by varying the sorbent's bed depth and the system's feed flow rate. The samples were collected at regular intervals from the outlet of the column. After that, the concentration of the samples was analyzed using the XRF Analyzer. The varied operational parameters per column setup are shown in Table 1. The range of parameters used for a glass column of 11 mm diameter and 200 mm height was adapted from studies by Adhikari et al. (2012) and Trgo et al. (2011).

Data Processing and Analysis
The performance of each column setup with varied sorbent type, amount, and feed flow rate was determined by plotting the breakthrough curves for Pb adsorption. The dimensionless concentration of C/C o is plotted against time , where C is the effluent concentration at any time t, and C o is the feed concentration. The total adsorbed species, q total in mg, is calculated as the area above the breakthrough curve times the feed flow rate as shown in Eq. (1) adapted from Karimi et al. (2012): where Q is the volumetric flow rate (mL min -1 ), (C o -C) is the adsorbed metal concentration represented by the equation of the breakthrough curve (mg/L), and t is the total time of flow (min).
(1) q total = Q 100 The dynamic uptake capacity of the columns at exhaustion point is calculated from Eq. (2) based on the study of Chittoo and Sutherland (2020): where q e is the mass of adsorbate adsorbed per mass of adsorbent at equilibrium (mg g −1 ) and m is the mass of adsorbent packed in the column, g.
The total removal percentage of metal was calculated from the following equation below: where m total (mg) is the total amount of metal entering the column at t total , which is calculated by the following equation (Karimi et al. 2012): The simulation of an adsorption column was achieved by fitting the experimental data into empirical sorption models and by calculating corresponding model parameters. The determined parameters are used to predict the adsorption performance given new operational conditions to facilitate the up-scale design of column systems.

Bohart-Adams Model
The Bohart-Adams model assumed that the adsorption rate is proportionate to the adsorbate initial concentration and the remaining capacity of the adsorbent (Bohart and Adams 1920). The Bohart-Adams model is expressed as: where k AB is the Bohart-Adams constant (L mg −1 min −1 ), Z is the bed height (cm), u is the linear flow rate (cm/min), and N o is the volumetric adsorption capacity (mg L −1 ) (Zhang et al. 2019) . The k AB and N o are empirical constants of the Bohart-Adams model, which values are characteristics of a specific column system. These constants are applied in scale-up studies wherein the desired column system is reproduced into a larger column.

Wolborska Model
The Wolborska model was a mathematical model that illustrates the lower concentration region of the breakthrough curve (Wolborska 1989). The model is established on the general equation of mass transfer accounting for the diffusion mechanism of the breakthrough curve (Trgo et al. 2011). It is expressed as: where β is the kinetic coefficient of external mass transfer, min −1 (Chittoo and Sutherland 2020), similar to the Bohart-Adams constant, β is a constant empirical characteristic of a specific Wolborska model of a particular column system. Also, it is applicable in scaling up laboratory-scale columns.

Thomas Model
The Thomas model was used to estimate the maximum adsorption capacity of a system based on the mass-transfer theory (Thomas 1948). It was established to represent adsorption without the consequences of intraparticle masstransfer resistance and external film resistance. The model assumes that the adsorption rate depends on the surface interaction between the adsorbate and the unused adsorbent capacity (Chittoo and Sutherland 2020). The model is expressed as: where q o is the maximum adsorption capacity (mg g -1 ), K TH is the Thomas rate constant, V is the volume of effluent (L), m is the adsorbent mass (g), and Q is the feed flow rate (L min -1 ) (Zhang et al. 2019) . Yoon and Nelson (1984) developed a simple empirical model which assumes that the rate of decrease of the likelihood of adsorption of adsorbate is proportionate to the probability of adsorption and breakthrough of an adsorbate and is expressed as:

Yoon-Nelson Model
where k YN is the Yoon-Nelson rate constant (min -1 ) and τ is the time at C/C o = 50% (Chittoo and Sutherland 2020).

PNB and PNZ Characterization
The elemental composition of the sorbents is summarized in Table 2 below. The results indicated that PNZ comprises 46.52% silica (SiO 2 ) and 9.00% alumina (Al 2 O 3 ), while PNB consists of 30.66% SiO 2 and 8.69% Al 2 O 3, assuming all Si detected are coming from SiO 2 and Al from Al 2 O 3 . Based on the study by Gili and Hila (2021), analytical methods for the characterization of the PNB and PNZ samples from LITHOS Manufacturing showed no other sources of Si and Al aside from SiO 2 and Al 2 O 3, respectively. Similarly, X-ray Diffractometry (XRD) was carried out by Gili and Hila (2021) and showed that PNZ is composed mainly of mordenite, clinoptilolite, and montmorillonite, while PNB is composed mainly of hectorite and montmorillonite. This study used XRF for the characterization of the sorbents which was reinforced and seen in the results of the previous work by Gili and Hila (2021). Based on the elemental analysis of the sorbents, the Si/Al ratio of PNB and PNZ are 3.53 and 5.17, respectively. The Si/Al ratio affects natural bentonite and zeolite adsorption performance. Higher adsorption is achieved at higher cation-exchange capacity (CEC), which is usually associated with a lower Si/Al ratio (Gili and Hila 2021). From this data, it is expected that PNB will yield higher adsorption performance than PNZ. This expectation on the adsorption performance of bentonite is from its predominant composition of clay minerals having Si/Al ratio of 2:1 which structure provides bentonite a high specific surface area, porosity, water, compound retention capability, and cationexchange capacity (Belbachir and Makhoukhi 2017). Similarly, the adsorption performance of zeolite is attributed to its structure which provides open pores or cavities in its framework. These cavities are often occupied by H 2 O molecules but can be exchanged in the presence of cations in aqueous solutions (Ramesh and Reddy 2011).

Wastewater Characterization
The heavy metal concentration of the wastewater composite is composed mainly of lead, cadmium, and iron, as shown in Table 3. Results showed that mercury (Hg) and arsenic (As) were below the limits of detection of the Vanta M Series XRF Analyzer. In contrast, a high amount of Pb, Cd, and Fe were detected, exceeding the Class D standard set by DAO 2016-08 (DENR 2016) which limits are at 0.2 ppm for Pb, 0.02 ppm for Cd, and 35 ppm for Fe. Hence, the wastewater quality of the samples indeed contained harmful heavy metals, which need to be treated.

Removal Efficiency
The performance of each column setup, with varied sorbent type, sorbent amount, and feed flow rate, were determined by plotting the breakthrough curves for Pb adsorption as shown in Fig. 2. The removal percentage of each setup is determined using Eq. (3) which q total and m total values are summarized in Table 4. Results showed that the mean removal percentage values between natural zeolite and natural bentonite are 26.90% and 50.43%, respectively, as shown in the statistical hypothesis testing for the difference of two means for independent groups in Table 5. The degrees of freedom are calculated using a 2-sample t test with unequal variances (Welch's t test). The test shows a significant difference with a p value of 0.02518 (p < 0.05) in the removal percentage between PNB and PNZ as sorbent for Pb. These results suggest that the mean removal percentage of PNB is greater than that of PNZ.  On the other hand, the total adsorbed species are presented in Table 4, in which the breakthrough curves do not follow a smooth curve. In terms of total adsorbed species, q total , the adsorption of each setup shows that PNB yields higher performance than PNZ. As determined in the Si/Al ratio calculation, PNB indeed showed a higher experimental adsorption performance based on the removal percentage owing to its lower Si/Al ratio, consequently a higher CEC, as compared to PNZ.
The average mass of Pb adsorbed per mass of adsorbent from q e values in Table 4 are 31.0 mg Pb/g PNB and 10.0 mg Pb/g PNZ. In batch sorption test, Olegario-Sanchez and Pelicano (2017) found that in static conditions, PNZ adsorbs 99.03% Pb ions from a simulated acid mine drainage solution. It is to be noted that this higher removal of Pb in static conditions is attributed to the single-ionic nature of the samples and does not reflect the actual adsorption of the sorbent on actual wastewater samples, which are multi-ionic in nature. The column adsorption of PNB, however, could not be compared to static conditions due to the lack of literature on PNB. It can, however, be compared to the natural bentonite studied by Sysa et al. (2019) which showed a Pb adsorption of 30.95 mg Pb/g natural bentonite. Although the adsorption capacities are similar for PNB under current conditions, and natural bentonite under static conditions, it is to be noted that the two sorbents were sourced in different localities and have different elemental compositions. The comparison of the calculated adsorption capacity in terms   Mousa et al. (2016) of the mass of Pb adsorbed per gram of adsorbent is summarized in Table 6. PNB and PNZ, generally have lower adsorption capacities compared to synthetic and modified adsorbents listed. PNB, however, has higher average adsorption capacity compared to the maximum adsorption capacity of silica modified calcium alginate-xanthan gum hybrid bead composite and activated carbon-calcium alginate composite. Both PNB and PNZ have higher average adsorption capacities compared to Biochar-alginate capsule Pb 2+ ion imprinted polymers. It is noteworthy that PNB and PNZ are natural adsorbents and were not modified to increase the surface area available for binding of Pb.

Application of Fixed Bed Adsorption Models
Linear models were initially obtained as a prerequisite of the non-linear models. Logarithms were used to linearize the experimental data to estimate characteristic parameters of the different adsorption models. The experimental data were fitted into the Bohart-Adams and Wolborska models for adsorption. Characteristic parameters were estimated by linear models ln(C/C o ) versus t in regions where the relative concentration C/C o < 0.75 at Q = 1 mL min −1 . The respective values of parameters K AB , N o, and ß a are shown in Table 7, wherein the parameters K AB and ß a increase with decreasing bed height while the N o values increase with increasing bed height. This shows that the maximum adsorption capacity described by N o is achieved at increased bed height which agrees with the results of the study by Aksu and Gönen (2004). Additionally, the overall system kinetics is initially driven by external mass transfer since the constants K AB and ß a increase with decreasing bed height. In general, N o values are greater in PNZ than in PNB, where at increased bed heights, PNZ shows higher adsorption capacity. In the Thomas model summarized in Table 8, parameters were estimated from the linear plot ln[(C o /C)-1] versus V eff at Q = 1 mL min −1 . The column data was fitted within the breakthrough curves' 0.1 < C/C o < 0.9 regions. Results show that K TH increases and q o decrease with the decreased mass of sorbent. This implies that the increase of the mass of sorbent which translates into the bed height increases the bed capacity described by q o (Zhang et al. 2019) .
The Yoon-Nelson parameters were estimated by plotting ln[C/(C o -C)] versus t. Table 9 shows constant rate values, K YN , and time at 50% concentration, τ. As the bed height increases, K YN decreases while τ increases. This shows that at higher bed height, breakthrough and saturation occur slower. This is attributed to the increase of available binding sites as the amount of adsorbent increases with bed height.
The fit of the non-linear models and the actual plot of C/C o versus time are shown in Fig. 3 Column systems, and its parameters are summarized in Table 10 (Q = 1 mL min −1 ). The Bohart-Adams and the Wolborska models follow the same upward curve for each system, while the Thomas model follows a downward curve. These three models have y-intercept values greater than zero. The Yoon-Nelson model closely follows the actual plot of C/C o versus time, including an estimated curve beginning from t = 0. The different adsorption models obtained provide valuable information for assessing the movement or the retention of metal ions from the liquid phase to the solid phase. Each column system may now be described by a specific adsorption model whose characteristic parameters be used in replicating the desired column behavior based on a chosen model. The information from the models is significant for the reproduction, scaling up, and improvement of a column system.

Effect of Varying Flow Rate
The results of varying the flow rate from 1 to 3 mL min −1 on both PNB and PNZ sorbents with a constant bed height of 10 mm are shown in Fig. 4. In this result, increasing the flow rate leads to a faster uptake (lesser saturation time) of both sorbents at 90% of the initial concentration. The plot of each sorbent in an increased flowrate is moving upward, resulting in a smaller area above the curve which translates into lower uptake capacity. This is explained by the accelerated movement of Pb in the system resulted in less sufficient residency time in the column. These results agree with the study by Chittoo and Sutherland (2020), where increased flow rates resulted in lessened contact time between the sorbent and sorbate.

Effect of Bed Height
Figure 5 below shows the effect of bed height relative to dynamic uptake capacity, saturation time, and removal percentage. The bed height of the columns was varied at 10 mm, 15 mm, and 20 mm for both PNB and PNZ sorbents with a constant flow rate of 1 mL min −1 . Results showed that increasing bed height yielded higher uptake capacity, longer saturation time at C/C o = 90%, and greater total removal percentage against Pb. These results coincide with the adsorption study by Zhang et al. (2019), where the increase in bed height resulted in a slower breakthrough that translates into higher saturation time. The gradual addition of sorbent in an increased bed height supports the idea that it takes longer to saturate the sorbent thoroughly. Hence, an increased amount of sorbent will have more adsorption sites available, resulting in capacity uptake improvement and a higher removal  Fig. 4 Breakthrough curves at a varied flowrate percentage of lead in the system. This result agrees with the study of Patel and Vashi (2011), where the increase in bed height shows an increase in surface area of adsorbent.
In the current condition, where the packing of particles was not varied, the increase in bed height, increases the available binding sites, thus, the accessible effective pores increased for both sorbents. However, the effect of packing on the effective porosity due to the increase in bed height was expected to be lower in PNZ than in PNB since Fitts (2013) states that the effective porosity is much smaller than the total porosity in crystalline rocks such as PNZ, while clays, such as PNB, relatively have higher porosity due to their platy clay mineral arrangement. As a result, PNB yields higher q e , saturation time, and removal percentage against lead, as shown in Fig. 5. Based on the results, the optimum bed height is limited to 20 mm, where q e , saturation time, and removal percentage are highest for the two sorbents. Any further increase in bed height cannot be extrapolated from three data points only.

Cost-Efficiencies
The prices of both sorbents per kilogram is 35 PHP. The cost efficiencies of each sorbent as cost per amount of Pb removed (PHP/g Pb) shown in Table 11. Results show that PNB can remove a gram of lead 2.37 PHP cheaper than PNZ. The costs per unit of both sorbents are the same; however, PNB removes a gram of lead for a much more affordable price than PNZ.

Conclusions and Recommendations
This study was carried out to evaluate and compare the utilization of PNZ and PNB sorbents in removing potentially toxic chemicals in the wastewater from small-scale mines. The wastewater from the ball mill facilities contained average heavy metal concentrations of 82.8 ppm, 19 ppm, and 432.4 ppm for Pb, Cd, and Fe, respectively. Column studies were conducted to evaluate the adsorption performance of the PNZ and PNB against Pb at varied flow rates and bed heights. Key findings of the study showed that PNB is a better adsorbent against heavy metals due to a lower Si/Al ratio than PNZ. This is confirmed by comparing the removal percentages of the two sorbents in terms of Pb removal. The removal percentage of PNB was 50.43% which is significantly greater than PNZ with 26.90%. The significant difference between PNB and PNZ was further confirmed by determining the dynamic uptake capacity and maximum adsorption capacity, wherein PNB yields higher values than PNZ. In applying empirical sorption models, the Yoon-Nelson model visually shows the closest estimation of actual values; however, the curves produced are not smooth. Nevertheless, the effect of flow rate for the two sorbents showed that an increase in flow rate results in a lesser saturation time at 90% initial concentration, which means that the faster the inlet flow rate is set, the faster it takes for the sorbent to be spent. Moreover, varying the bed height of the columns showed that the higher the bed height, the greater the uptake capacity, the longer the saturation time, and the larger the total removal percentage for both sorbents. Furthermore, in terms of the cost efficiencies of the sorbents, it revealed that PNB could remove a gram of lead 2.37PHP cheaper than PNZ; thus, PNB is more cost-efficient. To further investigate the adsorption performance of the PNB and PNZ sorbents, it is recommended to study the effect of sorbent particle sizes on the adsorption capacity. Since Cd and Fe were found to have concentrations above the DENR standards, it is recommended to study the adsorption efficiency and selectivity of PNZ and PNB on these elements. Additionally, the study only investigates the sorption properties of virgin PNB and PNZ sorbent. It is recommended to study the reusability of the sorbents after exhaustion. Regeneration studies to minimize the demand for virgin sorbents, reduce secondary pollution, and address the disposal problem.