Zirconium fumarate metal-organic framework: a selective adsorbent for fluoride from industrial wastewater

Fluoride is one of the reactive non-metallic elements. It combines directly or indirectly with many other elements and generates various fluoride compounds. Most fluoride compounds are stable and generally dissolve in nature. So fluoride remains in the surrounding environment (Loganathan et al. 2013; Halla et al. 2015). Fluoride originates from various sources, such as daily-use products (e.g., dental products, dietary supplements, and beverages), geogenic sources, and industrial production. The fluoride concentration in industrial wastewater is the highest among the different sources of origin, and the pollution level is also high (100–10,000 mg/L) (Wan et al. 2021).

The iron and steel industry, aluminium electrolysis, lead, aluminium, zinc, and copper smelting, rare earth element separation, uranium enrichment, photovoltaic industry, lithium-ion battery industry, cement industry, and thermal power generation are all fluoride-releasing industries (Wan et al. 2021). In the iron and steel industry, fluoride is present in iron smelting, waste gas, wastewater, and slag generated during ore mining, beneficiation, sintering, and smelting, which are essential for steel manufacturing by using a blast furnace. The high-temperature process in steelmaking releases potassium fluoride, sodium fluoride, and other fluoride compounds (Wan et al. 2021). Similarly, metallurgical industries, e.g., copper, zinc, aluminium smelting, etc., generate fluoride in the wastewater. High fluoride-containing wastewater (10,000 mg/L) in copper smelting is challenging to treat (Castillo et al. 2020). In the zinc smelting process, fluoride comes from zinc concentrate electrolytes (1,000 mg/L) (Menad et al. 2003). A tonne of aluminium requires 30 kg of fluoride to be produced (Shan & Guo 2013). Photovoltaic solar panels generate 1,000–3,500 mg/L fluoride-containing wastewater (Drouiche et al. 2007). In the battery industry, hydrofluoric acid contributes to fluoride in wastewater (500 mg/L) (Drouiche et al. 2012). Other industrial processes resulting in fluoride wastewater include uranium enrichment, phosphate fertilizer production, and ceramic, and cement factories (Dehghani et al. 2016).

Physical processes reduce the concentration, but not sufficiently to bring it down to the discharge standard (15 mg/L) (Schedule-VI, Central Pollution Control Board, India, The Environmental Protection Rules, 1986). Treatment of such high fluoride concentrations in industrial wastewater by physical (e.g., adsorption, membrane process) or chemical processes is also difficult (Wu et al. 2017). A literature review shows that various techniques have been reported to treat industrial wastewater with high fluoride concentrations. It can be divided into two categories: (i) chemical processes (such as chemical precipitation and coagulation/electro-coagulation) and (ii) physical processes (such as reverse osmosis (RO), ion-exchange (IX), electrodialysis, and adsorption) (Mohapatra et al. 2009). Fluoride-containing wastewater from the solar cell manufacturing process (Joint National Egyptian Chinese Renewable Laboratory, Egypt) was treated by chemical precipitation with quick lime and potassium hydroxide (Diwani et al. 2022). Hydrated lime was used for the separation of fluoride in wastewater from the fertilizer industry (Zueva et al. 2020). However, chemical precipitation is slow, and it is not efficient to reduce the fluoride concentration up to the discharge limit. Coagulation and chemical precipitation of fluoride produced in the microelectronics industry for wastewater treatment have been reported (Chang & Liu 2007). Similar high fluoride removal was reported in the electro-coagulation process with aluminium electrodes for treating steel industry wastewater (Mobarake Steel Complex, Esfahan, Iran) (Khatibikamal et al. 2010). In the electro-coagulation method, energy consumption and equipment costs are relatively high. Also, the selectivity of electrodes gets reduced frequently. Chemical precipitation and coagulation are the traditional methods of fluoride removal from waste streams (Khatibikamal et al. 2010). However, during precipitation and chemical coagulation processes, a large amount of chemicals is added, resulting in a high amount of fluoride precipitate sludge daily (Wan et al. 2021). Further, membrane technology can reduce fluoride in a wide pH range (Waghmare & Arfin 2015). The hybrid lime and RO system used for wastewater treatment at the aluminium fluoride manufacturing plant was reported to be efficient (Ezzeddine et al. 2015). However, treating high-strength and high-salinity wastewater is difficult, as membranes are frequently fouled (Singh et al. 2013). The coagulation-ultrafiltration process was studied for fluoride removal from wastewater generated in Shanghai, China's semiconductor manufacturing industry (Qiu et al. 2022). However, this process requires a large amount of coagulant. Ion-exchange technology reported up to 90–95% fluoride removal (Meenakshi 2006). However, other ions such as phosphate, sulphate, and bicarbonates in industrial wastewater lower its efficiency. Thus, chemical precipitation, coagulation, membrane processing, and ion-exchange resin used to reduce the fluoride content in the effluent are applicable only for low fluoride concentrations, as they are expensive (Singh et al. 2013). Also, in some cases, it has a low fluoride uptake capacity (Wan et al. 2021).

Among all the processes mentioned above, the adsorption process is considered one of the most promising methods for fluoride uptake because of its high uptake capacity, easy operating conditions, low operational cost, absence of harmful by-products, and low energy consumption (Patel et al. 2016; Dhanasekaran et al. 2017). In a recent study, calcium and zirconium modified acid-activated alumina (CAZ) reported 92% fluoride uptake (Kumari et al. 2020). Similarly, sulphuric acid-activated modified alumina resulted in 94% fluoride uptake from aluminium industry wastewater (National Aluminium Company Limited, Orissa, India) (Kumari et al. 2019). Drinking water treatment residues (polyaluminium chloride used as a coagulant) based on adsorptive material used for the treatment of industrial wastewater (smelting factory, Gyeonggi-do, South Korea) reported 90% fluoride uptake (Jung et al. 2016). The integrated process of calcium hydroxide precipitation and adsorption on hydroxyapatite was also reported to treat fluoride-containing wastewater from the aluminium finishing industry (Xanthi, Greece) (Melidis 2015). Alumina-modified citric acid showed fluoride uptake of 99% from phosphogypsum slag yard mine leachates (Li et al. 2022). These adsorbents show high uptake capacity, but no study on the adsorption–desorption process has been reported.

Metal-organic frameworks (MOFs) have received a lot of attention in recent years for their high fluoride adsorption due to their controllable pore structure, high stability, large surface area, and good physical and chemical properties (Ma & Zhou 2010; Li et al. 2011; DeCoste & Peterson 2014). Their uniform surface structure, comparatively high void ratio, and outstanding chemical and thermal stability make them different from other inorganic porous materials (Eddaoudi et al. 2002; Furukawa et al. 2013). As a super fluoride adsorbent, aluminium fumarate (AlFu) MOF was reported; it has a large specific surface area (1,156 m²/g) and higher uptake capacity (600 mg/g at 25 °C) than previously reported and commercially available water-stable MOFs (Basolite^®-MOFs, Merck) (Karmakar et al. 2016). Reuse of AlFu MOF has been reported (Kumari et al. 2022). Similarly, MIL-96(Al) MOF showed high selectivity towards fluoride ions and high pH stability (3–10), with an adsorption capacity of 32 mg/g at 25 °C (Zhang et al. 2014). MOF-801, a zirconium-based MOF, also demonstrated a high rate of fluoride adsorption capacity (Tan et al. 2020). Hydroxyapatite (HAp)-based layered lanthanum-benzene tricarboxylic acid-based MOFs (La-BTC MOFs) showed promising results for fluoride removal from water (Jeyaseelan et al. 2022). However, these MOFs as adsorbents were reported to have low initial fluoride concentrations (2–15 mg F/L) and were formulated with synthetic water.

Zirconium fumarate (ZrFu) MOF has been reported to be thermally, chemically, and water stable, having a large surface area and high fluoride adsorption capacity (Mukherjee et al. 2016). It was prepared from ZrCl₄ and fumaric acid as Zr₆O₄ (–OH)₄ (fumarate)₆ XH₂O. In addition, the presence of Zr-bound-OH groups in the structure improves fluoride adsorption by the ion-exchange process. It has been found effective for fluoride removal in drinking water. But no study has been done to see if fluoride can be selectively separated from industrial wastewater at its source and then reclaimed for use by desorption. This has motivated us to prepare materials suitable for removing fluoride from industrial wastewater. The efficacy of ZrFu MOF as an adsorbent for selective fluoride recovery from industrial wastewater has been prepared and studied. The wastewater was collected from the clarifier tank of the coke plant and blast furnace at the Tata Steel Plant, Jamshedpur, India. Every tonne of steel produced requires 25 m³ of water, and the rate of effluent generation is 5 m³/h (Mukherjee et al. 2016). The wastewater is treated by primary, secondary (biological process), and tertiary (advanced oxidation process) treatment processes, respectively. Fluoride, chloride, nitrate, phosphate, total solids, total dissolved solids, and salinity are all present in high concentrations in the treated water (Mukherjee et al. 2016). In this study, the effects of contact time, pH variation, adsorbent dose, initial concentration, adsorption kinetics, and isotherms were studied with synthetic water to understand the adsorption mechanism and factors affecting the adsorption process. The MOF was then studied for industrial wastewater under optimized conditions. A simultaneous adsorption–desorption study was also carried out for the re-use of the material.

Zirconium oxychloride octahydrate (98%) and fumaric acid (99.5%) were purchased from Loba Chemie Pvt. Ltd, India. N,N-Dimethyl formamide (DMF) (99%) and sodium fluoride (98%) were purchased from Merk Life Science Pvt. Ltd, India. Formic acid (85%) was purchased from Nice Chemical Pvt. Ltd, India. Ethanol (99.9%) was purchased from Changshu Hongsheng Fine Chemical Co., Ltd. The chemicals were used as-received.

Zirconyl chloride octahydrate (1.6 g) and fumaric acid (0.58 g) were mixed with a mixed solvent to make ZrFu MOF. A mixed solvent was prepared by adding DMF (20 mL) and formic acid (7 mL). The mixture was filled in an autoclave (KLB Instruments Co. Pvt. Ltd, India), and it was heated at 130 °C for 6 h in a hot air oven (i-therm, Mumbai, India). The white-coloured precipitate was separated by centrifugation at 8,000 rpm for 20 min (REMI R-24, India). The MOF was washed at least three times with DMF followed by ethanol. Washed ZrFu MOF was dried at 105 °C for 1 day in a hot air oven and then mortar pestle into fine powder. The MOF was then dried at room temperature for 1 h in a vacuum oven (Stericox, Delhi, India) and stored for future use.

After optimizing the operating conditions with synthetic water (150 mg/L), the adsorption study of ZrFu MOF for industrial wastewater was done. The experiments were carried out in 500 mL plastic containers containing 100 mL of synthetic fluoride solutions with compositions ranging from 10 to 150 mg/L (Mukherjee et al. 2016), an adsorbent dose of 1–5 g/L, and an adsorption time of 10–60 min. The tests were carried out in an orbital shaker set to a constant speed of 100 rpm (T = 35 °C and pH = 6). Wastewater was collected (January 2021) in a 10 L Jerry can from the clarifier of the coke plant (CP) and blast furnace (BF) of TATA Steel, Jamshedpur, India. The samples were centrifuged to remove the suspended solids, and the supernatant was used for the experiments. The wastewater was tested for chemical oxygen demand (COD), phosphate, chloride, nitrate, total hardness, and total solids using standard methods for water and wastewater analysis (American Public Health Association (APHA) et al. 1998). The pH was measured using a pH metre (LAB INDIA, Mumbai). The total solid was calculated using the gravimetric method. COD, chloride, and total hardness of the samples were estimated by the titration method. The nitrate and phosphate concentrations of the samples were determined by a UV–Vis spectrophotometer (Systronics 119, Mumbai). The fluoride desorption study was performed by adding 100 mL of DI water to the exhausted MOF. The mixture was shaken for 1 h at 100 rpm and 35 °C in a shake flask. The concentration of fluoride in the water was determined after the reaction. The regenerated MOF was again used in a fresh batch of synthetic and industrial wastewater. This process continued for several cycles for synthetic and industrial wastewater, respectively.

The structural characterization of the ZrFu MOF was conducted by an X-ray diffraction technique (XRD) using an X-ray diffractometer (Rigaku, Japan, Smart Lab 9Kw). Brunauer–Emmett–Teller (BET) analysis was performed at 77 K using a surface area analyser (Quantachrome Instruments, USA) to estimate the surface area, pore diameter, and pore volume of ZrFu MOF. The surface morphology of ZrFu MOF was analysed by a Field emission scanning microscope (FESEM; SM-6390LV, Jeol, Japan). The functional groups of the samples before and adsorption studies were analysed using Fourier transform infrared (FTIR) spectroscopy (IR-Prestige 21, Shimadzu Corporation, Japan, IR-Prestige 21).

ZrFu MOF was chosen among some of the water-stable MOFs (MIL-96, AlFu, ZIF-8, ZIF-67, and CuBTC) because it showed complete fluoride uptake capacity under the given operating conditions (MOF dose of 1 g/L, C_o = 10 mg F/L, pH = 6, and T = 35 °C) (Supplementary Figure S1). The experiments were conducted in a shake flask. The study shows that the fluoride uptake capacity of the MOFs varies from one another. Under similar operating conditions, the ZrFu MOF showed 100% fluoride uptake compared to other studied MOFs. The variation in the MOFs adsorption could be due to differences in (i) electrostatic adsorption, (ii) complexation mechanisms (Zhu et al. 2018; Jeyaseelan et al. 2020), and (iii) surface area/volume ratio. Increased fluoride uptake capacity was observed in ZrFu MOF due to Zr⁺⁴ compared to Al⁺³, Zn⁺², Co⁺², and Cu⁺². The positively charged higher valence ion (Zr) present in the synthesized MOF attracts more negatively charged fluoride ions by electrostatic attraction and complexation (Zhu et al. 2018). The functional groups (e.g., carboxylate group, hydroxyl group) present on the MOF structure provide active sites for the adsorption of fluoride (Ke et al. 2018). Thus, ZrFu MOF was selected for further studies.

The high selectivity of ZrFu MOF for fluoride ions was due to the comparatively higher electronegativity and electron-withdrawing property of fluorine compared to halogen elements and oxygen. This results in the replacement of hydroxyl ions by fluoride ions on ZrFu MOF. In addition, the octahedral structure of ZrFu MOF improves the accessibility of the exchange site. It increases diffusion in the entire structure of the MOF, which improves the uptake capacity of the MOF (Tan et al. 2020).

A similar study with blast furnace and coke plant wastewater having an initial concentration of 170 and 130 mg/L, respectively, and different contents (2, 3, 4, 6, 8, and 10 g/L) of ZrFu MOF was conducted. More than 90% of fluoride uptake for coke plant wastewater was obtained with 3 g ZrFu/L. Complete uptake was obtained with 10 g ZrFu/L (Figure 3(b)). In a similar study with blast furnace wastewater, 4 g ZrFu/L achieved more than 90% fluoride uptake and 10 g/L achieved complete uptake (Figure 3(c)). Both of these reactions were carried out for 60 min. While optimizing the reaction time, almost complete fluoride uptake was obtained within 5 min of the reaction, with 10 g ZrFu/L for the wastewater studied. This shows that the reaction with ZrFu MOF was rapid, indicating the presence of many active sites, which help quick fluoride uptake from wastewater. The wastewater characteristics of the coke plant and blast furnace show that the concentration of fluoride competing ions (Cl⁻, ⁠, ⁠, ⁠) was high for comparable active sites (Table 2). The concentration of coexisting anions before and after the adsorption with ZrFu MOF in coke plant and blast furnace wastewater was studied. Interestingly, after the adsorption process, only phosphate concentration was reduced in coke plant and blast furnace wastewater. This indicates that ZrFu MOF selectively adsorbs fluoride from the solution, and phosphate competes with fluoride for the same active site. Similar observations were reported earlier (Ke et al. 2018).

This study reported fluoride recovery from industrial wastewater by adsorption using ZrFu MOF. The morphology, porous structure, and fluoride uptake capacity were studied systematically. Different process parameters showed that the fluoride uptake capacity improved with an increase in MOF addition in the batch reactor and reaction time. Fluoride adsorption followed the Langmuir adsorption isotherm and the pseudo-second-order rate law, according to kinetic studies. ZrFu MOF absorbed fluoride completely from synthetic water, coke plant wastewater, and blast furnace wastewater, with uptake capacities of 49.66, 13, and 17 mg/g, respectively. Because of the high surface area of ZrFu MOF (538 m²/g), it quickly absorbed fluoride from wastewater and reached adsorption equilibrium. When the solution pH was less than pH_ZPC, protonated hydroxyl groups encouraged fluoride attraction to ZrFu MOF. The MOF regeneration studies showed that ZrFu MOF reaches a steady fluoride uptake capacity after repeated adsorption–desorption cycles. This is C7 (70%) for synthetic water, C4 (62%) for coke plants, and C2 (65%) for blast furnace wastewater. This variation was due to (i) a different initial fluoride concentration and (ii) with or without interfering ions at the same adsorption sites in the solution. Thus, fluoride may be recovered on-site through the adsorption column by an adsorption–regeneration cycle operation.

R.K. acknowledges the Ministry of Education (GoI), for the doctoral fellowship.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.