Remediation of PO₄³⁻ in Water Using Biodegradable Materials Embedded with Lanthanum Oxide Nanoparticles

Abstract

Eutrophication, a process in which algae grow inordinately, adversely affects aqueous fauna. Phosphorous at levels above 0.1 mg/L is adequate to cause eutrophication. In this study, we aimed to reduce the amount of PO₄³⁻ in water using biodegradable and ecofriendly sorbents. Lanthanum oxide nanoparticles were doped in agar and cellulose sponge to produce two new sorbents, agar–La and sponge–La, respectively. Both sorbents showed high efficacy in remediating up to 10 mg/L PO₄³⁻ in water. Sponge–La was found to be more proficient in terms of adsorption than agar–La because it required just 1 h to achieve 80% adsorption when the initial concentration of PO₄³⁻ was 10 mg/L. Sponge–La was effective at pH levels ranging from 4 to 8, with a removal rate of 80–100%. Although agar–La displayed a slow sorption process, it presented a high adsorption capacity (156 mg/g); moreover, the cake-shaped agar–La could be easily manufactured and separated from an aqueous matrix or any water-based solutions. These two sorbents could effectively remove high concentrations of PO₄³⁻, and their preparation requires a simple step. Agar–La was easier to manufacture, whereas the adsorption process using sponge–La was more rapid. In addition, both sorbents can be easily separated from the matrix after sorption.

1. Introduction

Eutrophication, during which algae grow inordinately, is a global problem that threatens the survival of aqueous fauna due to hypoxia and toxins [1,2]. Some coastal areas, such as the Bohai Sea, China, are vulnerable to eutrophication owing to their relatively enclosed terrain [3]. Despite efforts for decades to manage nutrient loads in the Chesapeake Bay watershed, the water quality of the estuary area has not improved [4]. P is one of the primary factors causing eutrophication; P at levels above 0.1 mg/L is adequate for the occurrence of eutrophication [5]. The level of P has been higher than that of N in the Bohai Sea area over the last 40 years [6]. P is predominantly present as PO₄³⁻ in water because of its high solubility [7]. A large amount of PO₄³⁻ in water is chemically precipitated with Fe, Ca, and Al compounds [8,9]; however, the precipitated P is difficult to recycle. Liu and his coworkers applied an electrical field to intensify the migration of PO₄³⁻ to an Al/C electrode, which is derived from a metal–organic framework (MOF) [10]; however, the installation of this device is expensive.

The adsorption process has been used to effectively remove PO₄³⁻ from water. The ability of natural minerals, such as dolomite and hydroxyapatite, to adsorb PO₄³⁻ has been investigated [11]. Using structural modulation during calcination, CaFe-layered double hydroxides were synthesized to adsorb PO₄³⁻ [12]. To date, the most common strategy has been the use of a composite comprising a carrier with a high surface area, such as meso-silica [13], or with high encapsulation ability, such as MOFs [14,15], and with another compound with a high affinity to PO₄³⁻. A magnetic MOF, Fe₃O₄@MIL-101 (Cr), which is feasible for recycling, has been synthesized to adsorb arsenate and PO₄³⁻ without interference [16]. Fe/Al(NO₃⁻) MOF shows a high PO₄³⁻-adsorption capacity but the introduction of N into the matrix aggravates eutrophication [17]. Sulfonic acid introduced into a Cu-based MOF shows high efficacy for PO₄³⁻ adsorption [18]. However, the complexity of the ligand increases the difficulty of synthesis. Zr(IV) has been embedded in biochar activated with chitosan–soya-bean husk for PO₄³⁻ adsorption [19]. As biochar is highly dispersive, a gel-like sorbent made of zirconium oxide and polystyrene offsets this limitation [20]. Moreover, some rare earth elements other than Zr that hardly coordinate with ligands, such as La and Ce, have a high affinity for PO₄³⁻, which is supported by the hard–soft acid–base theory; therefore, these metals have been intensively studied [21,22]. A Ce-doped MOF, UiO-66-NH₂, captured PO₄³⁻ intensively, and the input of amine-introduced N in the solution [23]. Min et al. [24] modified a UiO-66 MOF using La, which showed a phenomenal PO₄³⁻-adsorption capacity. However, the collection of La after each cycle of adsorption is not convenient. Although La-loaded biochar [25] and bentonite [26] are effective in removing PO₄³⁻, it is necessary to find a suitable carrier that is intact and unbreakable in harsh environments. Ribet et al. [27] imparted magnetite iron oxide (Fe₃O₄) to a cellulose sponge, which is cheap and biodegradable, to adsorb PO₄³⁻ proficiently. An La-loaded magnetic hydrogel has been reported to achieve 105.72 ± 5.99 mg P/g adsorption capacity [28]. The membrane was designed to modify the La–Zn(4,4-dipy)(OAc)2/bacterial cellulose composite with tannic acid [29]. Al(OH)₃ nanocrystals have been grown in an Artemia Cyst shell to achieve a high PO₄³⁻-adsorption capacity [30].

In this study, we aim to reduce the amount of PO₄³⁻ in water using biodegradable and ecofriendly sorbents. We postulate that lanthanum oxide nanoparticles (LONP) combined with cellulose sponge (sponge–La) can adsorb PO₄³⁻ with high efficacy. Additionally, we designed an LONP-containing sorbent using agar as the carrier (agar–La). Agar is an ideal carrier because it is cheap and biodegradable, and its shape can be easily manipulated. The preparation of these two sorbents requires a simple step: assembling the two components, the LONP and the carrier. In addition, it is easy to separate the sorbent from the matrix after sorption; no complex filtration techniques are required. We aim to evaluate the detailed practical aspects of each sorbent and justify its application.

2. Materials and Methods

2.1. Reagents and Apparatus

Lanthanum (III) nitrate, sodium hydroxide, and potassium dihydrogen PO₄³⁻ were purchased from Tianjin Jinke Fine Chemicals Co., Ltd. (Tianjin, China). Agar was procured from Qingdao Hope Bio-Technology Co., Ltd. (Qingdao, China). Cellulose sponge was obtained from Scotch-Brite (3M Co., Shanghai, China).

A horizontal shaker (SHY-2 water-bath shaker, Dadi Auto-Instrumentation Co., Baoding, China) was applied for the adsorption experiment.

2.2. Instruments

Instruments used to characterize the materials were as follows: X-ray photoelectron spectroscopy (XPS) system (ESCALAB 250Xi; Thermo Fisher Scientific, Waltham, MA, USA), Fourier transform infrared spectroscopy (FTIR) system (TENSOR 27; Bruker, Germany), transmission electron microscopy (TEM) system (JEM 2100F; JEOL, Tokyo, Japan), scanning electron microscopy associated with energy dispersive spectroscopy (EDS) system (Apreo 2C; Thermo Fisher Scientific, San Jose, CA, USA), and X-ray diffraction (XRD) analysis system (D8 Discover; Bruker AXS, Karlsruhe, Germany). MI-100S Water Analysis Device (Massinno Co., Ltd., Tianjin, China) was used to determine the concentration of total P. Alternatively, a microplate reader (JC1086-A; Qingdao Juchuang Environmental Protection Co., Ltd., Qingdao, China) was used to determine P concentration.

2.3. Synthesis of LONPs

First, LONPs were synthesized using Fang’s method [31]. In 40 mL of deionized water, 6 mmol La(NO₃)₃·6H₂O and 18 mmol NaOH were dissolved and stirred for 10 min. The white precipitate produced was transferred to an 80 mL autoclave tube and heated for 12 h at 120 °C. The product was then washed with deionized water and centrifuged five times at 100,000× g for 10 min each time. The supernatant was decanted, and the precipitate (LONP) was freeze-dried overnight.

2.4. Preparation of Agar–La

To prepare the agar–La membrane, 390 mg of agar and 50 mg of La were added to 10 mL of deionized water and boiled. The boiled solution was immediately poured into a ring mold (diameter = 3.8 cm) on a glass plate. The hot fluid was cooled for 10 min and solidified as a malleable membrane (Supplementary Figure S1).

2.5. Preparation of Sponge–La

To prepare sponge–La, 0.10 g of LONP was added to 100 mL of deionized water and stirred at 35 °C. The cellulose sponge was immersed in the emulsion for 2 h. Thereafter, the cellulose sponge was dried at 60 °C in a box furnace for 1 h.

2.6. Analysis of Sponge–La

To examine the actual effect of the sponge sorbent in sewage, we added pure sponge and sponge–La to the sewer from a residential kitchen. Several parameters, including total P (TP), total N (TN), and NH₄⁺-N, reflecting the water quality were measured using a water analysis device (MI-100S; Massinno Co., Ltd., Tianjin, China).

2.7. Analysis of PO₄³⁻

PO₄³⁻ was analyzed using the ammonium molybdate spectrophotometric method [32]. This is a standardized method regulated by the Ministry of Ecology and Environment in China. The water quality testing device that we used was equipped with reagent kits for the analysis.

2.8. Adsorption and Desorption Study

For adsorption study, 10 mL phosphate-spiked solution and 0.1 g sorbent were added to a small glass vial with pH ~6, which was then placed into a horizontal shaker to agitate at 30 °C.

For desorption study, the phosphate-contained sorbent was added to 10-mL 0.1 M NaOH in the glass vial, which was then placed into the horizontal shaker to agitate at 25 °C.

Formula used are:

Adsorption capacity:

$$\mathit{c}=\frac{\mathit{V}\left({\mathit{c}}_{\mathit{i}\mathit{n}\mathit{i}}-{\mathit{c}}_{\mathit{e}\mathit{q}\mathit{u}}\right)}{\mathit{m}}$$

where V (L) is the volume of spiked solution; c_ini (mg/L) and c_equ (mg/L) are the initial and equilibrium concentration of phosphate, respectively; m (g) is the mass of sorbent.

Adsorption rate:

$$\mathit{r}=\frac{{\mathit{c}}_{\mathit{i}\mathit{n}\mathit{i}}-{\mathit{c}}_{\mathit{e}\mathit{q}\mathit{u}}}{{\mathit{c}}_{\mathit{i}\mathit{n}\mathit{i}}}\times \mathbf{100}\%$$

3. Results and Discussion

3.1. Characterization

The XRD spectra of the agar–La membrane displayed sharper peaks in the crystalline structural features than the control (pure agar) because of La doping (Supplementary Figure S2a). In addition to the peaks of pure sponge, a sharper peak was observed for sponge–La owing to the crystalline form of La (Supplementary Figure S2b). When the sponge–La sorbent was added to actual sewage, the overall spectra shifted to the left, implying that the sponge–La could adsorb more fouling agents than pure sponge (Supplementary Figure S2c).

Because of the addition of LONP, sleeker lines were observed in the TEM image of agar–La (Supplementary Figure S3a,d) and sponge–La (Supplementary Figure S3f) than those of agar (Supplementary Figure S3b,e) and pure sponge (Supplementary Figure S3c). The rough surface of pure agar created spots for the binding of LONP. Moreover, the tendency to form a lamellar structure after doping LONP in the agar (Figure 1) was confirmed by EDS mapping, which showed that the La weight percent of agar–La was higher (Figure 1d) than that of pure agar (Figure 1c). The pure agar had a rugged structure, whereas agar–La displayed crystalline lines attributed to La, which was confirmed by EDS mapping, where the weight percent of La was 22.1% and that of P was 1.8% because of the adsorbed PO₄³⁻ (Figure 1e).

The honeycombed multiporous structure of sponge (Figure 2a,b) and sponge–La (Figure 2c,d) are observed. Meanwhile, the loaded La are distinctive for sponge–La, which can be confirmed from EDS analysis (Figure 2f), which presents 1.67 wt% La.

Two agar–La samples, pH6-A and pH6-B (replicates under pH 6) showed adsorption at pH 6. After the sorption process, pure agar, agar–La, and the two agar–La samples were characterized using FTIR (Supplementary Figure S4a). The intensity of the broad peak representing hydrogen bonds associated with hydroxyl groups at approximately 3450 cm⁻¹ gradually weakened. This weakening occurred the fastest in agar, followed by agar–La, pH6-A, and pH6-B, because of the adsorbed La and P.

The carboxyl groups (at approximately 1650 cm⁻¹) evolved in a similar pattern, indicating that the adsorbed PO₄³⁻ chemically disrupted the original structure of agar. The peak at 1410 cm⁻¹, belonging to agar, gradually split into two peaks at 1450 and 1376 cm⁻¹ after the adsorption of P on agar–La. As a result, the peak at 1450 cm⁻¹ appeared only for pH6-A and pH6-B.

To compare the difference between FTIR spectra of sponge and sponge–La (Supplementary Figure S4b), the moderately weak peak under ~1400 cm⁻¹ (carboxylic acid group) is flattened for sponge–La, indicating the chemical modification between sponge and La³⁺.

The XPS spectra of agar–La sorbents at initial PO₄³⁻ concentrations 5 and 10 mg/L (Supplementary Figure S5) demonstrated the difference between the following two adsorption scenarios: the different initial concentrations (5 and 10 mg/L) of PO₄³⁻ and the new bond formed at 830.1 eV when the initial concentration was 10 mg/L. The results imply that in the matrix containing a higher concentration of PO₄³⁻, La³⁺ had a higher probability of forming a chemical bond with PO₄³⁻.

3.2. Adsorption Study

Figure 3 illustrates the adsorption performance of the membranes when the initial concentration of total P is different. The actual initial concentration of total P was measured at 10 h. on the first day. After the measurement, all tubes were placed in a horizontal shaker and incubated at 30 °C. The next day, each sample was collected at 9 h (23 h after) to determine the P concentration, and the results are denoted as A on the x-axis. The samples were tested again at 15 h (29 h after) on the same day, and the results are denoted as B on the x-axis. On day 3, the samples were tested at 9 h (35 h after), and the results are denoted as C on the same axis. To analyze the adsorption capacity, Figure 3b shows that as initial concentration of phosphate elevates to 20 mg/L, the saturation of adsorption starts to surface. When it comes to 50 mg/L, the maximum adsorption capacity is stabilized to 150 mg/g.

For Columns 1 and 2, denoting 5 mg/L initial concentration on the second day of adsorption, the maximum adsorption capacity, approximately 80% removal rate, was achieved. Subsequently, the adsorption started to recede. However, when the initial concentration started to increase on day 3, the removal rate was higher than that on the previous day. The highest removal rate, approximately 90%, was observed at an initial concentration of 10 mg/L. As shown in Supplementary Figure S5, a new bond occurred in the La 3D spectra at this initial concentration, and it might be related to the chemical bonding between La³⁺ and PO₄³⁻. The highest removal rate declined to 10% at an initial concentration of 50 mg/L.

To determine the best-fit isotherm curve, all adsorption scenarios for agar–La were plotted (Supplementary Figure S6a). The maximum adsorption capacity was approximately 150 mg/g, which was observed when the equilibrium concentration of PO₄³⁻ was higher than 15 mg/L. The Langmuir curve (R² = 0.82) fit the adsorption more adaptably than the Freundlich curve (R² = 0.81). However, the curves were also similarly overlapped, indicating the co-existence of monolayer and multilayer adsorption.

Cellulose sponges completed the sorption process faster than agar membranes (Figure 4). The initial concentration of PO₄³⁻ used was 10 mg/L. For all solutions, the PO₄³⁻ concentration was tested 1 h after adsorption. The PO₄³⁻ removal rate was 30% when only the pure sponge was used, and the sponge–La significantly improved this to above 80%. The maximum removal rate was achieved in both control (40%) and sponge–La groups (100%) 3 h after adsorption. This was favorable, and it indicates the scope of field studies for PO₄³⁻ removal in the future. After 4 h, the PO₄³⁻ removal rate of sponge–La remained 100% but that of the sponge started to recede thereafter, as indicated by its poor PO₄³⁻ content (Supplementary Figure S7). The adsorption capacity quickly approached to the maximum, after only 1 h of the adsorption process (Figure 4b). Subsequently, in the isotherm study, the adsorption capacity levels up at 20 mg/L; of which the maximum value approaches to 160 mg/g. The simulation shows that Langmuir is the best-fit (Supplementary Figure S6b), indicating the monolayer chemisorption process. Besides, the adsorption rate is high at low concentration but declines when it is above 20 mg/L (Figure 4c). Additionally, pseudo-first- and second-order models were attempted for simulation. Pseudo-first-order (R² = 0.92) indicates the rapid adsorption process for sponge–La, while pseudo-second-order (R² = 0.996) is attributed to chemisorption process (Figure 4d). Overall, sponge–La proved to be more effective and efficient than pure sponges in removing excess PO₄³⁻ from water.

When agar–La was used, pH had a profound effect on adsorption. Samples of different pH levels were prepared by adding 0.1 M HCl (aq.) and 0.1 M NaOH dropwise to 10 mg/L PO₄³⁻ solution. The samples were collected after 16 h of adsorption, and pH was measured to determine data fluctuations (Figure 5a). The pH showed an increasing trend, and the actual pH was skewed from the initial pH. When the initial pH was 4, it shifted to approximately 6. When the initial pH was 8, it shifted to a slightly lower pH and when the initial pH was 10, it transitioned to 9. This fluctuation can be attributed to the acidic factor of the reagent used (KH₂PO₄). When the removal rate was the highest, above 60%, the optimal pH was 6. After continuing for another 24 h, the removal rate reached approximately 90% (Figure 5b). The overall curve was bell-shaped, indicating poor sorption at either very low or very high pH. Moreover, the dark blue color of the water indicates a high concentration of PO₄³⁻ (Supplementary Figure S8).

The analyses showed promising results, particularly for total N, which decreased to 0 mg/L after the application of sponge–La (Figure 6). The adsorption method was identical to that used in PO₄³⁻ adsorption. Although the sorbent was designed to ameliorate substantial amounts of total P, it was found to be effective for several other substances related to eutrophication. Nitrate input can indirectly affect PO₄³⁻ leaching to groundwater-fed wetlands because nitrate can directly lead to the mobilization of sulfate and immobilization of ferrous iron [33]. The relationship between N and P becomes the driving force for the selective adsorption of N rather than P when sponge–La is used in domestic sewage. Further research is required to ensure the selectivity and efficacy of P and N removal in an aqueous environment.

As for the desorption process (Figure 7a), after 3 h, majority phosphate is effectively released, when the original concentration is 10 mg/L, which brings potential to recycle phosphate after highly phosphate-polluted water is remediated using sponge–La. Furthermore, this sponge sorbent can maintain its ability after 4, 5 cycles (Figure 7b).

As shown in

Table 1. Comparison of other sorbents.

Type	Maximum Adsorption Capacity (mg/g)	Advantages	Disadvantages
La-MOF [24]	348.43	High adsorption capacity	Difficult to separate from solutions
Sulfonic acid-derived Cu-MOF [18]		High affinity to perchlorate and PO43−	Difficult to synthesize
Ce-MOF [23]	211.86	High adsorption capacity	Introduction of N
Fe/Al(NO3−) MOF [17]	130	Selectivity to PO43−	Introduction of N
La-modified bentonite [28]		Applicable to real aqua, and inexpensive	Difficult to separate from solutions
La biochar [25]	46.37	Applicable to real aqua	Low adsorption capacity, and difficult to separate from solutions
Agar–La (this study)	156	High adsorption capacity, easily separable from solutions, and simple synthesis
Sponge–La (this study)	160	Rapid sorption process, applicable to actual sewers, which contain N

, our approach shows high potential to alleviate eutrophication compared to other sorbents, as agar–La has a high adsorption capacity, involves a simple synthesis method, and can separate from an aqueous matrix or water-based solutions. In addition, sponge–La can rapidly complete PO₄³⁻ adsorption.

3.3. PO₄³⁻ Analysis

During the ammonium molybdate method, the reaction between orthophosphate and ammonium molybdate occurs to form phosphorous-molybdenum hybrid acid, which is immediately reduced by ascorbic acid to produce a blue chelate. The darker the blue color, the higher the concentration of PO₄³⁻ in the water. This was observed when the sponge sorbent was used to remove 10 mg/L PO₄³⁻ in water. Supplementary Figure S9 represents the color distinction for samples with sponge–La (the first three replicates) and pure sponges (the last three replicates). A clear solution indicates that the majority of the PO₄³⁻ was absorbed by sponge–La.

However, this method required large quantities of reagents to complete the reaction, as each sample should be at least 5 mL, to which 0.2 mL of ammonium molybdate and 0.4 mL of ascorbic acid had to be added. Consequently, large quantities of reagents were wasted, which impeded the routine testing of PO₄³⁻. Herein, we propose a rapid testing method that requires only a small number of samples and the reagents. In a 96-well transparent plate, 100 µL of sample was added, to which 10 µL of ammonium molybdate and 20 µL of ascorbic acid were added. Horizontally, the plate was gently shaken a few times, and the reaction was allowed to continue for 20 min. The plate was then placed in a microplate reader to measure the absorbance at 630 nm. The higher the absorbance, the higher the concentration of PO₄³⁻. After 3 h of absorption, the absorbance was several times higher for the pure sponge than for sponge–La (Figure 8a). The absorbance for sponge–La was 0.1 throughout, which inversely correlated to the removal rate tested—the lower the absorbance, the higher the removal rate. The relationship was examined in the pH study as well (Figure 8b). The lowest absorbance appeared at an initial pH of 4 (actual pH was 6), which corresponded to the highest removal rate. The highest absorbance was observed at extremely low and high pH levels.

To further confirm the effectiveness of the proposed adsorbents and determine any interference of other ions present in water, common competing ions for PO₄³⁻ in water, such as F⁻ and Na⁺, should be considered in future studies.

4. Conclusions

We designed two biodegradable and ecofriendly sorbents, agar–La and sponge–La, and demonstrated through removal experiments that they have promising applications in aqueous environments containing high PO₄³⁻ levels. Agar–La is cheaper and has a more modifiable shape than sponge–La; however, it requires a longer time to achieve adequate removal of PO₄³⁻ ions. In contrast, even though the raw material (sponge) for Sponge–La is expensive to procure, this sorbent can complete the sorption process rapidly. Both sorbents can acquire the maximum adsorption capacity around 150~160 mg/g. Furthermore, agar–La is stable in a wide pH range from 4–8, within which it maintains a high adsorption rate around 80~90%. Thus, an informed decision should be made when applying the two sorbents.