Selenite adsorption and reduction via Iron(II) impregnated activated carbon produced from the Phosphoric acid activation of landfill waste wood

doi:10.21203/rs.3.rs-2507587/v1

Download PDF

Research Article

Selenite adsorption and reduction via Iron(II) impregnated activated carbon produced from the Phosphoric acid activation of landfill waste wood

https://doi.org/10.21203/rs.3.rs-2507587/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 10 Oct, 2023

Read the published version in Archives of Environmental Contamination and Toxicology →

You are reading this latest preprint version

Chemical activation of waste materials to form activated carbon, (AC) is complicated by the large amounts of chemical activating agents required and wastewater produced. To address these problems, we have developed an optimized process for producing AC by phosphoric acid activation of construction waste. Waste wood from construction sites was ground and treated with an optimized phosphoric acid digestion and activation that resulted in high surface areas (> 2000 m²/g) and a greater recovery of phosphoric acid. Subsequently the phosphoric acid activated carbon (PAC), was functionalized with iron salts and evaluated for its efficacy on the adsorption of selenite and selenate. Total phosphoric acid recovery was 96.7% for waste wood activated with 25% phosphoric acid at a 1:1 ratio, which is a substantially higher phosphoric acid recovery than previous literature findings for phosphoric acid activation. Post activation impregnation of iron salts resulted in iron(II) species adsorbed to the PAC surface. The iron(II) chloride impregnated AC removed up to 12.45 ± 0.025 mg selenium per g Iron-PAC. Competitive ions such as sulfate and nitrate had little effect on selenium adsorption, however, phosphate concentration did negatively impact the selenium uptake at high phosphate levels. At 250 ppm, approximately 75 of adsorption capacity of both the selenate and the selenite solutions was lost, although selenium was still preferentially adsorbed. Peak adsorption occurred between a pH of 4 and 11, with a complete loss of adsorption at a pH of 13.

Activated carbon

activation

phosphoric acid

selenium

selenate

selenite

iron doping

Annually there are approximately 2.5 million tons of waste wood generated in Canada from industrial, commercial, land clearing, and demolition. Waste wood accounts for almost 50% of all demolition materials deposited in landfill annually.(Environment and Climate Change Canada Public Inquiries Centre) This makes it a prime target as an affordable source for generating new value-added materials such as activated carbons. Activated carbon (AC) is a versatile adsorbent due to its high porosity, surface area and surface chemistry. It is commonly produced by gasification, of agricultural, forestry, and industrial waste materials, but there has been increasing interest in production through chemical activation due to the lower activation temperatures, higher yields, and increased porosity. The phosphoric acid activation of cellulose-based materials such as wood, succeeds in all these areas, with significantly lower activation temperatures (in the 400–500 ^oC range as opposed to the 800–900 ^oC employed for steam, CO₂ and other chemical activations) and much higher yields (35–45 % as opposed to the 10–15 % of KOH). The phosphoric acid work as a templating agent whre both its concentration and mass ratio effect the porosity that forms. The H₃PO₄ produces micro and mesoporous activated carbons in the range of 0.5–15 nm which can lead to faster kinetics than its more microporous competitors.(Sevilla and Mokaya 2014) As the acid would be the largest continuous cost of this process it is important to be able to recover and reuse it in subsequent activations. It was for this reason that the lower temperature of 400 ^oC was maintained in all activations. The recovery and the reuse of the recovered acid was monitored to ensure that the process would remain economically feasible.

It is not enough to create a high surface area activated carbon it is also necessary to demonstrate its efficacy. For this reason, the adsorption of selenite and selenate was explored. Selenium contamination in river and ground water is an issue in several sites throughout North America.(He et al. 2018) Selenium is a vital micronutrient for humans and other mammals, but it becomes toxic at higher concentrations in the body. The World Health Organization recommends a 10 \({\mu }\text{g}\) level of selenium in drinking water with a maximum of 40 \({\mu }\text{g}\) per day(J.K. Fawell et al. 2011). It plays an essential dietary role as it is required for the formation of the enzyme glutathione peroxidase which is involved in fat metabolism(N. N. Greenwood 1997). Selenium is commonly found as selenite (Se(IV) SeO₃ ^− 2) and selenate (Se(VI) SeO₄ ^− 2) in natural water. Se(IV) is more toxic than Se(VI), but both have been found to bioaccumulate(Long et al. 2014)(Olegario et al. 2010).Whether selenium takes the form of selenite or selenate is strongly governed by water chemistry. Both the pH and oxidizing conditions have strong effects. Selenium (VI) is present in oxidizing conditions, generally as HSeO₃⁻, while Se(IV) is more present in reducing conditions generally as SeO₄²⁻ (Mirbagheri et al. 2008). Higher levels of selenium are generally found in ground water than in surface waters due to contact with minerals(Santos et al. 2015). Strategies for the removal of selenite and selenate, include weakly basic poly-amine ion exchange columns and precipitation followed by adsorption on an adsorbent material(Nishimura et al. 2007).

The principal mechanism of adsorption for unfunctionalized activated carbon is via Van der Waals forces. This makes it an excellent adsorbent for larger non-polar molecules, but less adsorptive of anionic species. As selenate and selenite are both anions, the predominant mechanism for their adsorption onto activated carbon surfaces has been found to be through surface functionalities and defects (Herrero Latorre et al. 2013)(Gonzalez et al. 2010a). Because of this, iron salts have been a particular focus due to their abundance and ability to complex with both selenite and selenate (Yamani et al. 2016). Work with solutions of nanoscale zero-valent iron has shown intraparticle reduction of Se(IV). During the reaction, iron is oxidized and diffuses outward, as selenium and oxygen are attracted inward and form a thin layer of reduced zero valent selenium(Ling et al. 2015). This reduction appears to be crucial to prevent the desorption of selenium(Gezer et al. 2011)(van der Hoek1 et al.). Finding economical solutions for remediating industrial sites is still needed. Even though solutions currently exist to solve some of the many current issues, these methods are often considered too expensive to be commercially feasible (bioreactors, iron nano-particles). (Zhang et al. 2008)(Xue et al. 2018)With these considerations in mind, in this study the effectiveness of post activation iron salt additions as a source of functionalization were thoroughly evaluated on the powdered activated carbon (PAC) by adsorbing iron salts (FeCl₂, and C₆H₅FeO₇) onto the PAC surface. The impact of this functionalization on adsorption of selenium species was examined from a range of aqueous contexts.

2.1 Raw Materials

Waste wood from construction sites in Ontario, was chipped and ground until it passed through mesh 18 screen (less than 1.0 mm). Waste wood was found to have approximately 1% atomic nitrogen, 70% carbon, and 29% oxygen with some trace amounts of calcium and silicon. The primary surface modification experiments done with iron salts examined as sources for impregnation include: iron(II)chloride (Fe(II)Cl₂), iron(III)citrate (C₆H₅FeO₇), ethylenediaminetetraacetic acid iron(III) sodium salt (C₁₀H₁₂N₂NaFeO₈), and iron(III)pyrophosphate (Fe₄(P₂O₇)₃).

2.2 Preparation of Activated Carbon using H₃PO₄

Wood was digested by 25%, 50% and 85% H₃PO₄ at a 1:1, 1.5:1, and 2:1 dry mass ratios for 20 hours at 35 \(\text{℃}\) in a convection oven. The digested waste wood was then placed in crucibles in a muffle furnace for 30 minutes at 400\(\text{℃}\) under nitrogen at 5 L/min. Samples were then ground and washed with 10 mL 0.1 M HCl to 1 g unwashed PAC at 80 \(\text{℃}\) for 1 hr with agitation, followed by 10 mL 0.1 M NaOH to 1 g unwashed PAC and finally 10 mL DI water to 1 g unwashed PAC at 80 \(\text{℃}\). Samples were then dried at 110 \(\text{℃}\)overnight.

2.3 Re-concentration of H₃PO₄

Recovery and recycling of the primary chemical input, phosphoric acid, is important for both cost reduction and reduction of environmental impact. Phosphoric acid was recovered from aqueous washes of the activated carbon following the thermal activation step. These wash solutions were combined and boiled to remove excess water. Recycled phosphoric acid and recycled phosphoric acid with addition of fresh phosphoric acid were used to activate new waste wood feedstock.

2.4 Impregnation with iron(II)chloride and iron(III)citrate

Surface modification with iron(II) chloride and iron(III)citrate was achieved by first base treating PAC with 0.1 M NaOH for 1 hour (10 mL: 1g PAC), followed by washing with DI to a constant pH and dried. The base treated PAC was then placed in 50 mL iron salt solutions of 0.05 M, 0.10 M, 0.15 M, and 0.25 M concentrations.

2.5 Adsorption of Selenite to Iron impregnated PAC

The adsorption of the target analyte anions, selenite and selenate, were investigated by placing 100 mg of iron impregnated PAC in 100 mL solutions of 25 ppm selenate and selenite. The kinetics of the reaction were examined by testing the removal of selenium over time. 100 mg of Iron-PAC was placed in 100 mL solutions of 25 ppm selenite and 100 mL solution of selenate. Solutions were filtered, and selenium concentration was measured after 10 minutes, 30 minutes, 60 minutes, 120 minutes and 1200 minutes using microwave plasma atomic emission spectrometry (MP-AES).

To evaluate the effect of competitive anions on selenium uptake, 100 mg Iron-PAC was added to solutions of 25 ppm selenite with concentrations of 10 ppm, 25 ppm, 50 ppm and 250 ppm nitrate, phosphate, and sulfate. All solutions were then vacuum filtered, and the filtrate was examined for selenium content using microwave plasma emission spectrometry(Gonzalez et al. 2010b).

To investigate the effect of pH on selenium adsorption, solutions of 100 ppm selenite with pH values between 1 and 14 were made by adjusting pH with NaOH and HCl. 100 mg of Iron-PAC was then placed in these solutions and left for 1 hr. with no agitation. Solutions were then vacuum filtered, and the filtrate was analyzed for selenium content using MP-AES (Herrero Latorre et al. 2013)(López-Antón et al. 2007).

2.6 Characterization Techniques

Analysis by phosphorous-31 nuclear magnetic resonance spectroscopy (³¹P NMR) and ion chromatography (IC) was carried out on the washing filtrate. Etidronic acid was used as the ³¹P-NMR calibration standard.

Specific surface area and pore size analysis was measured using a Micromeritics Tristar II plus with Tristar II Plus v3.02 software. The samples were analyzed using N₂ adsorption at 77 K with 104 points monitoring adsorption and desorption between 0.0065 p/p⁰ and 0.995 p/p⁰ and 52 points desorption between 0.995 p/p⁰ and 0.104 p/p⁰. Some samples were additionally investigated using CO₂ adsorption at 273 K. All specific surface areas are reported using Brunauer-Emmet-Teller (BET) specific surface area analysis with pore size distributions developed using DFT with slit geometry modeling 2D-NLDFT with N₂ carbon finite pores(Jagiello and Olivier 2009). Microporosity was measured using T-plot with points between 0.35 and 0.5 nm using the Harkins and Jura thickness curve. Total pore volume was taken from the single point adsorption from the isotherm.

The X-ray photoelectron spectroscopy (XPS) analyses were carried out with a Thermofisher Scientific K(alpha) spectrometer using a monochromatic Al K(alpha) source (15mA, 15kV). The instrument work function was calibrated using sputter cleaned Au, Ag and Cu to determine the absolute linearity of the binding energy scale, with 83.96 eV for the Au 4f7/2 line for metallic gold, 368.21 eV for Ag 3d5/2, and 932.62 eV for the Cu 2p3/2 line of metallic copper. Survey scan analyses were carried out with a 200 eV pass energy and 1 eV/step. The x-ray spot size is a 2:1 ellipse with a 400 \({\mu }\text{m}\) major axis. High resolution analyses were carried out with an analysis area of 300 x 700 \({\mu }\text{m}\) and a pass energy of 20 eV. Fitting was done using CASA XPS (version 2.31) with the spectra being corrected to the main line of the carbon 1s spectrum at 284.85 eV.

3.1 Optimization of H₃PO₄ activation

Waste wood feedstock unlike coal and other feedstocks is not already carbonized and so the wettability and permeation of activating agents becomes a key factor in controlling pore sizes and surface areas of the produced activated carbons. Table 1 demonstrates the effects of 25, 50 and 85% concentrations of H₃PO₄ acid to wood with ratios of 1:1, 1.5:1, and 2:1 on both the surface area and pore size distribution. The 25% concentration at a ratio of 1:1 showed the best results with an average surface area of 2236 m²/g with the 1:1 50% concentration a little lower at 2106 m²/g. The higher surface area of the 25% 1:1 ratio, is most likely a result of the lower viscosity allowing better mixing and permeation of the wood during the digestion stage. At the 85% concentration, mixing was problematic due to the viscosity of the acid resulting in poor permeation. This was especially impactful for the 1:1 ratio which used the smallest volume and resulted in the lowest surface areas for the 85% concentration. Higher ratios of acid mixed better and a linear increase in porosity from 1:1 to 2:1 might be expected but is not seen. This could be a product of the increased acid content resulting in the formation of larger polyphosphates that, permeate more poorly and form larger, less-stable pores which when washed out; ending in pore collapse and a subsequent loss of surface area. This is supported by Fig. 1, which shows the pore size distribution for 1:1, 1.5:1 and 2:1. At a concentration of 1.5:1 there is a widening of the pores between 2 and 6 nm relative to other ratios. When the ratio is increased to 2:1, porosity between 2–6 nm drops, indicating either poorer permeability of the acid, or more likely a loss of pore volume due to these larger pores collapsing when the pyrophosphate is removed. The pore size distribution of 85% H₃PO₄ showed the same trend. The mesoporosity decrease seen in Table 1 also supports the collapse of larger pores with a decrease across all concentrations when the acid to wood ratio was increased from 1.5:1 to 2:1. This lower permeation and pore collapse is also validated by the phosphorous content seen in Table 2 which shows lower phosphorous contents for higher concentrations and ratios in the final product.

Table 1

Surface areas, pore volumes and yields of activated carbons produced with varying concentrations and ratios of acid to wood
Sample	Surface area (m²/g)	Total pore volume (cm³/g)	Micropore volume cm³/g	Mesopore volume cm³/g	% yield
1:1 25%	2236\(\pm 213\)	1.27 \(\pm\)0.13	0.23 \(\pm\)0.02	1.04 \(\pm\) 0.13	47.67 \(\pm\)1.46
1.5:1 25%	1859 \(\pm\)50	1.07 \(\pm\)0.03	0.21 \(\pm\)0.02	0.86 \(\pm\) 0.05	48.95 \(\pm\)0.35
2:1 25%	1361 \(\pm\)104	0.79 \(\pm\)0.10	0.15 \(\pm\)0.02	0.64 \(\pm\)0.12	47.45 \(\pm\)0.70
1:1 50%	2106 \(\pm\)42	1.08 \(\pm\)0.01	0.34 \(\pm\)0.04	0.74 \(\pm\) 0.08	47.00 \(\pm\) 0.82
1.5:1 50%	1989 \(\pm\)104	1.02 \(\pm\)0.10	0.15 \(\pm\)0.02	0.87 \(\pm\) 0.12	52.50 \(\pm\)7.77
2:1 50%	1523 \(\pm\)86	0.89 \(\pm\)0.08	0.16 \(\pm\)0.02	0.73 \(\pm\) 0.10	52.00 \(\pm\)4.20
1:1 85%	1734 \(\pm\)93	0.96 \(\pm\)0.09	0.18 \(\pm\) 0.03	0.78 \(\pm\) 0.15	44.90 \(\pm\)1.20
1.5:1 85%	1859 \(\pm\)50	1.07 \(\pm\)0.03	0.21 \(\pm\) 0.02	0.86 \(\pm\) 0.05	45.00 \(\pm\)1.00
2:1 85%	1763 \(\pm\)119	1.01 \(\pm\)0.12	0.18 \(\pm\) 0.04	0.83 \(\pm\) 0.17	46.30 \(\pm\)0.25

Table 2

Atomic % of phosphoric acid activated carbons produced under different dry mass ratios and concentrations
Sample	Atomic % C	Atomic % O	Atomic % P	Atomic % Na
25% H₃PO₄ 1 to 1	83.41	13.15	2.61	0.98
25% H₃PO₄ 1.5 to 1	84.50	12.68	1.91	0.90
25% H₃PO₄ 2 to 1	88.66	9.98	1.35	0.12
50% H₃PO₄ 1 to 1	79.85	16.26	2.86	1.03
50% H₃PO₄ 1.5 to 1	85.08	12.50	2.33	0.09
50% H₃PO₄ 2 to 1	88.17	10.13	1.62	0.07
85% H₃PO₄ 1 to 1	81.77	15.18	2.33	0.72
85% H₃PO₄ 1.5 to 1	82.30	14.39	2.40	0.91
85% H₃PO₄ 2 to 1	87.06	10.93	1.38	0.62

3.2 Recovery and recycling of H₃PO₄

To promote the economic viability and reduced environmental impact of the activated carbon produced using the methods developed in this research, it was necessary to show that recovering and reusing the H₃PO₄ was achievable with no detrimental effects on surface areas. XPS of the crude, unwashed PAC contained 13.3 atomic percent phosphorous, with the fitting of the high resolution phosphorous 2p peak at 133 eV showing that the majority was phosphoric acid. This is indicative of the phosphoric acid input remaining within the sample during activation. Phosphate is highly soluble and the subsequent XPS of the washed product show the phosphorous content falls to less than 2.5 atomic % (Table 2).

The high-resolution scan in Fig. 2 shows the peak splitting of the 3/2 and ½ phosphorous 2p peak. The fitting shows that 58% of the phosphorous is bound to the activated carbon as PO₄³⁻ and P₂O₇⁴⁻ (133.2eV). These peaks have overlapping binding energies and are therefore indistinguishable. The remaining phosphorous appears in the form of H₃PO₃ and PO₃⁻ at 134.2 eV.(Puziy et al. 2008) The results are in line with other work that shows the C-O-P bond is more stable and forms rather than the C-P-O bond(Lupul et al. 2014). To promote increased recovery additional washing could be implemented to remove the remaining phosphate species although much of it may be trapped.

³¹P NMR was used to identify the speciation and recovery of the phosphoric acid species washed out of the activated carbon product. The 0.1 M HCl washing filtrate showed an NMR resonance at 0.074 ppm in the ³¹P spectrum that was characteristic of phosphoric acid demonstrating that this is the only significant phosphorous species(Silverstein 1963). The ³¹P NMR of the NaOH washings, showed peaks characteristic of phosphoric acid at 0.4 ppm, and the pyrophosphates, diphosphoric acid at -10.6 ppm and triphosphoric acid at -21.1 ppm and − 22.6 ppm (Long et al. 2014). Pyrolyzed phosphoric acid could have been present in the acidic wash, and may have decomposed quickly prior to the ³¹P NMR experiment. The formation of pyrophosphate species by phosphoric acid under thermal treatment is well known and occurs by a simple dehydration process (Harry Marsh 2006). The presence of pyrophosphate in the basic environment of the 0.1 M NaOH (pH 13) washing suggests that the acidic environment of the 0.1 M HCl wash is regenerating phosphoric acid from polymerized pyrophosphate chains. The ³¹P NMR of the DI washings showed similar peaks to those found in the NaOH wash with one new smaller peak at 3.4 ppm most likely associated with short alkyl chain phosphates(Puziy et al. 2008).

Ion chromatography was used to evaluate the concentration of phosphate within the washings. The average amount of phosphoric acid recovered in the 0.1 M HCl wash, from AC that was treated with 85% phosphoric acid with a 1:1 (dry weight) impregnation ratio of activating agent to feedstock, was 63.9\(\pm\)1.6%. In the 0.1 M sodium hydroxide (NaOH) wash, 6.3\(\pm\)0.3% was recovered, and in the DI water wash a negligible amount below 0.5% was recovered.

The recovery of phosphoric acid from activation using 25% phosphoric acid at a dry mass 1:1 impregnation ratio, was much higher with 87.5\(\pm\)0.6% recovered in the 0.1 M HCl wash, 9.2\(\pm\)1.5% recovered in the following NaOH wash, and a negligible amount below 0.5% in the DI wash. The difference in recovery is most likely the failure of the more viscous, highly concentrated acid to permeate the wood, in which case the acid has been left behind on glassware and lost before the activation has taken place. This is supported by the percent of phosphorous remaining in the activated carbon which is comparable between the 1:1, 25 %and 85 %pointing to a failure to permeate and not excess phosphorous remaining trapped. The much lower surface areas and pore volumes of the 85% phosphoric acid activation also substantiates the failure to permeate due to an increase in the concentration.

3.2.1 Production of AC Using Recycled H₃PO₄

To evaluate the effect of using recycled H₃PO₄ recovered from the 0.1 M HCl wash for further activation of waste wood, two sets of experiments were carried out. In the first H₃PO₄ was recovered from the wash and re-concentrated. This re-concentrated H₃PO₄ was then used to activate a fresh sample of waste wood feedstock.

Figure 3: The percent of phosphorous recovered from each individual activation and the percent of total phosphorous recovered from the initial activation, over successive activations using recovered phosphoric acid from the 0.1M HCl hot wash. Sample filtrates were analyzed using microwave atomic plasma emission spectroscopy (MPAES) to determine phosphorous content.

Figure 3 shows the percent of phosphorous recovered from each individual activation and the percent of total phosphorous recovered from the initial activation over successive activations using recovered phosphoric acid from the 0.1M HCl wash. The average phosphorous recovery was 80% per activation resulting in a continuous loss with each successive activation. Figure 4 shows that successive activations without additional top ups to the recovered acid decreased in surface area with each subsequent activation whereby the 6th activation produced had only 60% of the 1st activations surface area. The downward trend in surface area is expected as less and less acid is available to digest and then crosslink within the waste wood during activation.

In the second experiment H₃PO₄ was recovered, re-concentrated and topped up with fresh acid to account for losses in the previous activation cycle. In Fig. 7, the surface areas that result from this top-up process remains steady. With this top-up in place, the acid could be used repeatedly with no loss in porosity of the PAC. This could result in significant cost savings and substantial reduction in chemical inputs for the process.

Surface Modification with iron salts

Post-activation functionalization using iron salt solutions (Iron(II)chloride, Iron(III)citrate, iron(III)ethylenediaminetetraacetic acid) were investigated. For selenite and selenate removal it was found that PAC loaded with Fe(II) was more effective than Fe(III). Because of this post-activation functionalization with Fe(II)Cl was focused on for further study.

PAC was first treated with 0.1 M NaOH to increase the oxygen groups available for bonding. This resulted in a decrease of 313 \(\pm\) 52 m²g⁻¹ in the observed surface area. This indicates that some etching and pore collapse may occur(Gao et al. 2013), although some sodium and hydroxide ions though highly soluble, are most likely still being adsorbed to the AC surface filling or blocking smaller pores and increasing the overall density of the PAC. Impregnation of the iron species was dramatically improved after treating the PAC surface with NaOH, going from uptake of less than 2% of the iron available in solution to up to 95%. The distribution of the speciation of the iron(II) adsorption (Table 3) can be seen in the XPS of the iron 2p peak (Grosvenor et al. 2004) as a mixture of both goethite ((FeO)OH), hematite (Fe₂O₃) and magnetite (Fe₃O₄). The inclusion of Fe(II) species such as magnetite is important for any surface reactions that require reduction, where the iron would act as the reducing agent.

Table 3

Distribution of Fe species adsorbed to PAC from fitting the Fe 2p peak
Species	Percent of Fe Species
Fe metal	1
Fe₂O₃	16
Fe₃O₄²⁺	14
Fe₃O₄³⁺	30
FeOOH	39

4.1 Selenium adsorption on Iron Loaded PAC

In order to target the most common selenium oxyanions, selenate and selenite, PAC impregnated with iron salts was investigated. To analyze the removal efficiency of the Iron-PAC, standard solutions of selenite (SeO₃²⁻) and selenate (SeO4²⁻) were prepared. Impregnation with iron citrate resulted in negligible adsorption whereas impregnation with iron(II)chloride resulted in removal of both selenate and selenite. Iron(II)chloride was therefore used in all subsequent selenium experiments. The ineffectiveness of iron(III) species points to the necessity for iron(II) to be present on the PAC surface for the removal of selenium to take place. This points to a reduction mechanism most likely being part of a chemical adsorption of selenite and selenate and not simply a ligand exchange or physical adsorption.

4.1.1 Selenite and Selenate Removal

Figure 5 shows the results of two time-trials. The first shows that when Iron-PAC was in excess, all of the Se(IV) was removed to within the limit of detection of the MP-AES instrument which is 1 ppm for Se species. The second shows that 90% of the Se(IV) was removed after 20 hrs. with a maximum loading of 4.19 \(\pm\) 0.024 mg Se removed per gram Iron-PAC. Both these results are in good agreement with literature results showing the majority of adsorption occurring within the first 10 minutes(Ling et al. 2015)(Seyed Dorraji et al. 2017).

In order to increase selenium uptake, PAC samples were subsequently impregnated with iron under nitrogen to prevent oxidation of the iron(II) to iron(III). This increased the adsorption of selenium to 12.45 \(\pm\) 0.025 mg Se per g Fe-PAC. This also points to the importance of iron(II) in the removal of selenite and selenate from solution.

The removal efficiency for Se(IV) fell between published results of other studies using other substrates. Dorraji et al demonstrated that nanosized magnetite impregnated on wet spun chitosan adsorbed 1.34 mg Se(IV)/ per g while Dobrowolski showed that magnetic iron/manganese oxide nanomaterials adsorbed 6.573 mg Se(IV) per g and 0.769 mg Se(VI) per g(López-Antón et al. 2007) (Kwon et al. 2015)(Kim et al. 2017). Results have been as high as 120.1 mg Se(IV) /g and 83.8 mg Se(VI) /g adsorbed on polyamine modified graphene oxide nanocomposites(Lu et al. 2017).

The efficiency for Se(VI) was lower at 1.01 \(\pm\) 0.024 mg removed per gram. This is consistent with previously published research showing that selenate adsorption is typically lower than that of selenite on most composite surfaces(Kwon et al. 2015). This is most likely due to the mechanism of adsorption. Figures 6 through 9 show possible mechanisms for the formation of iron metal complexes with selenium oxyanions including a monodentate ligand, a bidentate mononuclear ligand, and a monodentate binuclear ligand (bridging ligand).

The formation of these metal ligand complexes with selenite and selenate is most likely irreversible (Peak 2006) (Sabur and Al-Abadleh 2015). Since iron(III) impregnation showed no observed adsorption of selenite and selenate this implies that there is more than a simple ligand exchange occurring. Previous research shows the reduction of selenate and selenite by zero valent iron and by magnetite or iron(II) containing molecules is central to adsorption(Zhang et al. 2008). This implies that after the formation of a metal ligand complex there is a reduction of the selenite and selenate and an oxidation of iron. Selenium combines readily with most metals and many non-metals to form selenides. After reduction to Se²⁻ the selenium is bound to the iron complex, or if reduced to elemental selenium, Se(0), it is no longer soluble and therefore has little mobility within the solution and can be retained within the PAC porous network.

XPS can be used to examine the reduction of selenite to elemental selenium and selenide. The most significant challenge with using XPS to examine the interaction of selenium species with iron loaded PAC, is that the most prominent peak for selenium, is the 3d5 peak (57 eV), which overlaps with the iron 3p peak (55–56 eV). Although the iron 3p peak does not have the strongest iron signal it can still provide problems due to its much larger atomic % within the samples in question.

A more creative approach is required to understand the selenium adsorption on these iron functionalized samples. This can be done by examining some of the selenium peaks with weaker signals in order to generate the complete picture of the reduction of selenium by iron on the activated carbon surface.

The weak signals belonging to the Auger L3M45M45 peak (184 eV) and the 3s peak (232 eV) of selenium were identified. The deconvolution of the iron 3p peak after selenium adsorption showed an increased signal at 58 eV. This can be attributed to the Se 3d peak and shows the presence of both elemental selenium and what is most likely selenium dioxide (Fig. 10).

Continuing with this analysis of the selenium and iron XPS signals, a high-resolution scan of the Se 3p peak was completed. Prior to selenium adsorption the Fe-PAC surface showed no signal in this region as it contains no selenium. Figure 11 clearly shows two peaks at 161.5 eV and 165 eV which can be attributed to elemental selenium and SeO₂ respectively (Alexander V. Naumkin 2022). This supports a selenium reduction mechanism being important for the adsorption of the selenite and selenate anions.

4.1.2 Effect of Competitive Ions

Within tailings ponds and other aqueous environments there is typically a wide range of species that can compete with selenium ions for adsorption on the activated carbon surface. Depending on the source of the tailings ponds, the content within it can differ significantly. Because of their prevalence in oil sands process-affected water, sulfate (SO₄²⁻), phosphate (PO₄³⁻), and nitrate (NO₃⁻) were chosen to be examined for their effect on selenate and selenite uptake on the Iron-PAC surface. Oil sands process-affected water (OSPW) has been recorded as containing as high as 536\(\pm\)90 ppm SO₄²⁻ and as low as 85\(\pm\)6 ppm. Content of phosphate and nitrate is much lower in OSPW with a high of 0.07\(\pm\)0.01 ppm in the former and 0.10\(\pm\)0.03 ppm in the latter.

Phosphate showed the largest effect on selenium species adsorption as seen in Fig. 12 with concentrations at 250 ppm losing ~75% of adsorption capacity in both the selenate and the selenite solutions. The molar concentration of phosphate was 13 times higher than that of the selenate and selenite and yet both selenium species still had significant adsorption to the Iron-PAC surface (2.6 x 10⁻³ mol/L PO₄³⁻ : 2.0 x 10⁻⁴ mol/L Se) indicating preferential binding of the selenium species. The drop off in adsorption, seen with the competitive phosphate ions, is most likely a result of a build-up of negative charge on the surface as phosphate adsorbs. This would form an electrostatic charge pushing away the selenium anions and preventing their adsorption. For species with lower charges such as sulfate and nitrate this would not be observed at the same concentrations. Figures 13 and 14 show that SO₄²⁻, and NO₃⁻, had little effect as competitive anions with the selenate and selenite. This suggests that selenate and selenite preferentially bind to the surface and is consistent with literature results that show low binding affinities for sulfate and nitrate with iron(II)(Dobrowolski and Otto 2013) (Gonzalez et al. 2010a).

4.1.3 Effect of pH on Selenium Adsorption

Tailings ponds within the Alberta oil sands have pH’s that vary from 7.77 to 8.63 (Mahaffey and Dubé 2017). Water purification of other contaminated aquatic systems may require activated carbon to be functionally active in a variety of pH values. It is therefore, important to characterize the active pH threshold of the Iron-PAC. Selenate can be found as \(\text{S}\text{e}{\text{O}}_{4}^{-2}\) at all pH’s above 2 and predominantly as \(\text{H}\text{S}\text{e}{\text{O}}_{4}^{-1}\)below a pH of 1. Selenite can be found as \({\text{H}}_{2}\text{S}\text{e}{\text{O}}_{3}\)at a pH of less than 2, as \(\text{H}\text{S}\text{e}{\text{O}}_{3}^{-1}\)at a pH between 4 and 7 and finally as \(\text{S}\text{e}{\text{O}}_{3}^{-2}\) at a pH greater then 9. The pH dependent behaviour is controlled by several competing factors; the migration of selenite to the surface of the Iron-PAC, the deprotonation of complexed selenite and finally surface complexation.

Peak adsorption of Se (IV) by Iron-PAC, was seen to occur between a pH of 4 and 10. Between a pH of 3 and 8.5 the selenite exists predominantly as the protonated form HSeO₃ ^− 1.(Kim et al. 2017b)(Kim et al. 2017b)(Kim et al. 2017b)(Kim et al. 2017b)(Kim et al. 2017b) Selenium adsorption to the Iron-PAC was less pH dependent than coagulants such as iron or aluminum alone, which show sharper declines in adsorption at pH values lower than 4 and higher than 9.(Hu et al. 2015)(Hu et al. 2015)(Hu et al. 2015)(Hu et al. 2015)(Hu et al. 2015) This may be a result of a difference in the pH gradient within the bulk solution and the pores where the redox transitions take place.(López-Antón et al. 2007c)(López-Antón et al. 2007c)(López-Antón et al. 2007c)(López-Antón et al. 2007c)(López-Antón et al. 2007c) There was a complete loss of Se(IV) adsorption at pH 13 most likely due to electrostatic repulsion between the anionic selenite species and the now negatively charged Iron-PAC surface. This could potentially be used to desorb selenium for recovery.

4.2 Conclusion

In this work we have clearly demonstrated that high surface activated carbons could be produced through the phosphoric acid activation of wood and in this case specifically, waste wood from construction sites (Muñoz et al. 2003). More uniquely, we have shown that up to 97% of the phosphoric acid could be recovered, with 87.5% readily available, already regenerated in the hydrochloric acid wash. This was higher than the 50–80% previously reported in literature and was achieved by using lower concentrations of phosphoric acid (Lim et al. 2015). Recovering and reusing the phosphoric acid, was found to produce high surface area activated carbons, provided that the acid was replenished to account for losses.

The Iron-PAC was shown to be very effective for the adsorption of selenate and selenite with peak adsorption over a wide range of pH from 4 to 10. The maximum adsorption for selenite observed was 12.45 \(\pm\) 0.03 mg Se per g Fe-PAC with high selectivity over competitive anions SO₄²⁻, NO₃⁻, and PO₄³⁻ although higher concentrations of PO₄³⁻ did result in a reduction in adsorption of both selenite and selenate. While there are expensive adsorbents that provide higher removal efficiencies of selenite and selenate (Crane and Scott 2012), the method presented here is cost effective and makes use of repurposed waste wood with recycled phosphoric acid activating agent to reduce both costs and negative environmental impacts of manufacturing the adsorbent.

Declarations

Conflict of interest

The authors declare that we have no conflicts of interest. The manuscript is approved by all authors and has not been submitted to more than one journal for simultaneous consideration. The manuscript described has not been published before. The submitted work has not received any financial support from any third party, and there is no financial relationship with any entities. All of the financial organizations associated with this work have been disclosed. There is no patent, whether planned, pending or issued, relevant to the submitted work.

Alexander V, Naumkin AK-VSWG, CJP (2022) The NIST X-Ray Photoelectron Spectroscopy Database Number 20. In: National Institute of Standards and Technology
Crane RA, Scott TB (2012) Nanoscale zero-valent iron: Future prospects for an emerging water treatment technology. J Hazard Mater 211–212:112–125. https://doi.org/10.1016/J.JHAZMAT.2011.11.073
Dobrowolski R, Otto M (2013) Preparation and evaluation of Fe-loaded activated carbon for enrichment of selenium for analytical and environmental purposes. Chemosphere 90:683–690. https://doi.org/10.1016/J.CHEMOSPHERE.2012.09.049
Environment and Climate Change Canada Public Inquiries Centre National Waste Characterization Report :The composition of Canadian residual municipal solid waste
Gao Y, Yue Q, Gao B et al (2013) Preparation of high surface area-activated carbon from lignin of papermaking black liquor by KOH activation for Ni(II) adsorption. Chem Eng J 217. https://doi.org/10.1016/j.cej.2012.09.038
Gezer N, Gü M, Osman Aydın A (2011) Adsorption of Selenite and Selenate Ions onto Thiourea-Formaldehyde Resin. J Appl Polym Sci 122:1134–1141. https://doi.org/10.1002/app.34246
Gonzalez CM, Hernandez J, Parsons JG, Gardea-Torresdey JL (2010a) A study of the removal of selenite and selenate from aqueous solutions using a magnetic iron/manganese oxide nanomaterial and ICP-MS. Microchem J 96:324–329. https://doi.org/10.1016/J.MICROC.2010.05.005
Gonzalez CM, Hernandez J, Parsons JG, Gardea-Torresdey JL (2010b) A study of the removal of selenite and selenate from aqueous solutions using a magnetic iron/manganese oxide nanomaterial and ICP-MS. Microchem J 96:324–329. https://doi.org/10.1016/j.microc.2010.05.005
Grosvenor AP, Kobe BA, Biesinger MC, Mcintyre NS (2004) Investigation of multiplet splitting of Fe 2p XPS spectra and bonding in iron compounds. Surf INTERFACE Anal Surf Interface Anal 36:1564–1574. https://doi.org/10.1002/sia.1984
Harry Marsh FR-R (2006) Activated Carbon. Elsevier Science & Technology Books
He Y, Xiang Y, Zhou Y et al (2018) Selenium contamination, consequences and remediation techniques in water and soils: A review. Environ Res 164:288–301
Herrero Latorre C, Barciela García J, García Martín S, Peña Crecente RM (2013) Solid phase extraction for the speciation and preconcentration of inorganic selenium in water samples: A review. Anal Chim Acta 804:37–49
Hu C, Chen Q, Chen G et al (2015) Removal of Se(IV) and Se(VI) from drinking water by coagulation. Sep Purif Technol 142:65–70. https://doi.org/10.1016/J.SEPPUR.2014.12.028
Jagiello J, Olivier JP (2009) A simple two-dimensional NLDFT model of gas adsorption in finite carbon pores. application to pore structure analysis. J Phys Chem C 113:19382–19385. https://doi.org/10.1021/JP9082147/. ASSET/IMAGES/LARGE/JP-2009-082147_0004.JPEG
Fawell JK, Lund U, Mintz B (2011) Selenium in Drinking-water Background document for development of WHO Guidelines for Drinking-water Quality
Kim YJ, Yuan K, Ellis BR, Becker U (2017) Redox reactions of selenium as catalyzed by magnetite: Lessons learned from using electrochemistry and spectroscopic methods. Geochim Cosmochim Acta 199:304–323. https://doi.org/10.1016/J.GCA.2016.10.039
Kwon JH, Wilson LD, Sammynaiken R (2015) Sorptive uptake of selenium with magnetite and its supported materials onto activated carbon. J Colloid Interface Sci 457:388–397. https://doi.org/10.1016/J.JCIS.2015.07.013
Lim WC, Srinivasakannan C, al Shoaibi A (2015) Cleaner production of porous carbon from palm shells through recovery and reuse of phosphoric acid. J Clean Prod 102:501–511. https://doi.org/10.1016/J.JCLEPRO.2015.04.100
Ling L, Pan B, Zhang W (2015) Removal of selenium from water with nanoscale zero-valent iron: Mechanisms of intraparticle reduction of Se(IV). Water Res 71:274–281. https://doi.org/10.1016/j.watres.2015.01.002
Long C, Qi D, Wei T et al (2014) Nitrogen-doped carbon networks for high energy density supercapacitors derived from polyaniline coated bacterial cellulose. Adv Funct Mater 24:3953–3961. https://doi.org/10.1002/adfm.201304269
López-Antón MA, Díaz-Somoano M, Fierro JLG, Martínez-Tarazona MR (2007) Retention of arsenic and selenium compounds present in coal combustion and gasification flue gases using activated carbons. Fuel Processing Technology 88:799–805. https://doi.org/10.1016/J.FUPROC.2007.03.005
Lu Z, Yu J, Zeng H, Liu Q (2017) Polyamine-modified magnetic graphene oxide nanocomposite for enhanced selenium removal. Sep Purif Technol 183:249–257. https://doi.org/10.1016/J.SEPPUR.2017.04.010
Lupul I, Yperman J, Carleer R (2014) Tailoring of porous texture of hemp stem-based activated carbon produced by phosphoric acid activation in steam atmosphere. https://doi.org/10.1007/s10934-014-9894-4. Gra_ Zyna Gryglewicz •
Mahaffey A, Dubé M (2017) Review of the composition and toxicity of oil sands process-affected water. Environ Reviews 25:97–114. https://doi.org/10.1139/ER-2015-0060. /SUPPL_FILE/ER-2015-0060SUPPL.PDF
Mirbagheri SA, Tanji KK, Rajaee T (2008) Selenium transport and transformation modelling in soil columns and ground water contamination prediction. Hydrol Process 22:2475–2483. https://doi.org/10.1002/hyp.6842
Muñoz Y, Muñoz M, Ren´ R et al (2003) Phosphoric and boric acid activation of pine sawdust. J Chem Technol Biotechnol J Chem Technol Biotechnol 78:1252–1258. https://doi.org/10.1002/jctb.923
Greenwood NN (1997) AE Chemistry of the Elements, 2nd edn
Nishimura T, Hashimoto H, Nakayama M (2007) Removal of selenium(VI) from aqueous solution with polyamine-type weakly basic ion exchange resin. Sep Sci Technol 42:3155–3167. https://doi.org/10.1080/01496390701513107
Olegario JT, Yee N, Miller M et al (2010) Reduction of Se(VI) to Se(-II) by zerovalent iron nanoparticle suspensions. J Nanopart Res 12:2057–2068. https://doi.org/10.1007/s11051-009-9764-1
Peak D (2006) Adsorption mechanisms of selenium oxyanions at the aluminum oxide/water interface. J Colloid Interface Sci 303:337–345. https://doi.org/10.1016/J.JCIS.2006.08.014
Puziy AM, Poddubnaya OI, Socha RP et al (2008) XPS and NMR studies of phosphoric acid activated carbons. Carbon N Y 46:2113–2123. https://doi.org/10.1016/j.carbon.2008.09.010
Sabur MA, Al-Abadleh HA (2015) Surface interactions of monomethylarsonic acid with hematite nanoparticles studied using ATR-FTIR: Adsorption and desorption kinetics. Can J Chem 93:1297–1304. https://doi.org/10.1139/CJC-2015-0350/SUPPL_FILE/CJC-2015-0350SUPPL.PDF
Santos S, Ungureanu G, Boaventura R, Botelho C (2015) Selenium contaminated waters: An overview of analytical methods, treatment options and recent advances in sorption methods. Sci Total Environ 521–522:246–260
Sevilla M, Mokaya R (2014) Energy storage applications of activated carbons: Supercapacitors and hydrogen storage. Energy and Environmental Science. Royal Society of Chemistry, pp 1250–1280
Seyed Dorraji MS, Amani-Ghadim AR, Hanifehpour Y et al (2017) Performance of chitosan based nanocomposite hollow fibers in the removal of selenium(IV) from water. Chem Eng Res Des 117:309–317. https://doi.org/10.1016/J.CHERD.2016.10.043
Silverstein RM (1963) Spectrometric Identification of Organic Compounds, 8th edn. Wiley, New York
van der Hoek1 EE, van Elteren2 JT, Comans ~ RNJ DETERMINATION OF As,Sb AND Se SPECIATION IN FLY-ASH LEACHATES
Xue J, Huang C, Zhang Y et al (2018) Bioreactors for oil sands process-affected water (OSPW) treatment: A critical review. Sci Total Environ 627:916–933. https://doi.org/10.1016/J.SCITOTENV.2018.01.292
Yamani JS, Lounsbury AW, Zimmerman JB (2016) Towards a selective adsorbent for arsenate and selenite in the presence of phosphate: Assessment of adsorption efficiency, mechanism, and binary separation factors of the chitosan-copper complex. Water Res 88:889–896. https://doi.org/10.1016/j.watres.2015.11.017
Zhang N, Lin LS, Gang D (2008) Adsorptive selenite removal from water using iron-coated GAC adsorbents. Water Res 42:3809–3816. https://doi.org/10.1016/J.WATRES.2008.07.025

Download PDF

Journal Publication

published 10 Oct, 2023

Read the published version in Archives of Environmental Contamination and Toxicology →

Reviewers invited by journal
05 May, 2023
Reviewers agreed at journal
08 Feb, 2023
Editor assigned by journal
27 Jan, 2023
First submitted to journal
26 Jan, 2023

You are reading this latest preprint version

Selenite adsorption and reduction via Iron(II) impregnated activated carbon produced from the Phosphoric acid activation of landfill waste wood

Status:

Journal Publication

Version 1

Abstract

Figures

1. Introduction

2. Materials And Methods

3. Results And Discussion

4. Analyte Adsorption

Declarations

Declarations

References

Status:

Journal Publication

Version 1