Different routes of bismuth mineral transformation during pertechnetate and perrhenate uptake for subsurface remediation

Abstract

We investigated basic bismuth subnitrate for removal of radioactive technetium-99 as pertechnetate (⁹⁹TcO₄⁻) from contaminated groundwater. This material removed 93% of the initial concentration of ⁹⁹TcO₄⁻ within a week via formation of pH-dependent mineral phases that were identified here, but not reported previously. Perrhenate (ReO₄⁻) removal was also studied because it is a widely used non-radiological analogue for ⁹⁹TcO₄⁻, considering their similar physicochemical properties. We found that removal of ReO₄⁻ was not identical to removal of ⁹⁹TcO₄⁻ and led to formation of an additional transitional phase. This demonstrates that perrhenate and pertechnetate have different kinetics of contaminant removal as a result of variations in mineral transformation.

---

Water impact

Basic bismuth subnitrate was studied to capture a contaminant ⁹⁹TcO₄⁻ and its non-radioactive analogue ReO₄⁻ in a groundwater simulant. Removal of ⁹⁹TcO₄⁻ without its release back within a two-month period occurred during bismuth mineral transformation. Application of the non-toxic bismuth minerals can be one of the efficient options for subsurface remediation but requires more in-depth investigations and insights.

Introduction

Bismuth(III) (Bi³⁺, referred to as Bi) materials are non-toxic and versatile, and comprise a group of remediation materials with numerous advantages, e.g., environmentally friendly, high affinity for many contaminants, availability, cost effective, and easy to handle.^1,2 In environmental research, Bi-based materials, mainly oxyhalides BiOX (X = Cl, Br, I), are predominantly studied for their photocatalytic properties for water disinfection, groundwater remediation, air purification, carbon dioxide reduction, energy production, and for electrochemical sensing of pollutants.^2–6 However, there are synthesis challenges with BiOX-based materials, and lack of in-depth understanding of structure–property relationships.⁷

Here, we investigated Bi-based materials for contaminant removal via mechanisms other than photocatalysis. Past nuclear waste disposal practices at the U.S. Department of Energy's Hanford Site (WA, USA) have driven the need for groundwater remediation. We have conducted a series of studies on sequestration of contaminants existing as anions in the subsurface, i.e. Cr(VI) as CrO₄²⁻, U as carbonate complexes of UO₂²⁺, ¹²⁹I as IO₃⁻, and ⁹⁹TcO₄⁻ (further TcO₄⁻), from contaminated groundwater with Bi-based materials.^8–11 In these studies, we synthesized a disordered Bi oxyhydroxide (BOH), which was effective for uptake of all the abovementioned contaminants except pertechnetate, TcO₄⁻. We found that pertechnetate can be removed with another Bi material, commercially available Bi subnitrate (BSN).^9,10 The removal was pH-dependent and further studies were needed to elaborate removal mechanism and Bi mineral transformation of BSN in a groundwater simulant. Specifically, we aim to investigate which mineral phase is responsible for TcO₄⁻ uptake during BSN transformation and kinetics of TcO₄⁻ removal.

Bi aqueous chemistry is complex, with hydrolysis leading to olation and formation of polynuclear hydroxo species, and oxolation leading to oxo-bridged species through hydrolytic polymerization of olated cations. These oxo-hydroxo polynuclear species form Bi subnitrate during crystallization of dilute Bi³⁺ solutions in HNO₃.¹² If concentrated HNO₃ is used, the pentahydrate Bi(NO₃)₃·5H₂O is crystallized from Bi³⁺ solutions, and the basic salt BiO(NO₃) precipitates upon dilution.¹³ A mixture of the basic compounds BiO(NO₃), Bi(NO₃)(OH)₂, Bi₂O₂(OH)(NO₃), and BiOOH, as well as a variety of Bi polynuclear species, were also reported in literature as a result of hydrolysis of Bi³⁺ aqueous solutions.^14–16 Bi hexanuclear clusters are defined as [Bi₆O_x(OH)_8−x]^(10−x)+, i.e. [Bi₆O₅(OH)₃](NO₃)₅·3H₂O and [Bi₆O₄(OH)₄](NO₃)₆·H₂O,¹⁷ – both described first by Lazarini (1978 and 1979).^15,18 They are represented by the two main complex ions [Bi₆O₅(OH)₃]⁵⁺ and [Bi₆O₄(OH)₄]⁶⁺ charge balanced by NO₃⁻ and OH⁻ ligands through Bi–ONO₂ bonds or OH–ONO₂ hydrogen bonds,^14,19 and their structural isomers, e.g., [Bi₆O₄(OH)₄(NO₃)₅(H₂O)](NO₃), [Bi₆O₄(OH)₄(NO₃)₆(H₂O)₂]·H₂O.²⁰ According to Christensen and Lebech (2012),²¹ the overall composition of a basic Bi nitrate can be described as a mixture of [Bi₆O₄(OH)₄]⁶⁺ and [Bi₆O₅(OH)₃]⁵⁺ with a general formula [Bi₆O₄(OH)₄]_0.5+x[Bi₆O₅(OH)₃]_0.5−x(NO₃)_5.5+x (ref. 21) or [Bi₆O_4.5(OH)_3.5]₂(NO₃)₁₁.²² In this mixture, the equilibrium between the clusters is pH-dependent:²¹

[Bi₆O₄(OH)₄]⁶⁺ + H₂O ⇋ [Bi₆O₅(OH)₃]⁵⁺ + H₃O⁺ (1)

where [Bi₆O₄(OH)₄]⁶⁺ would be preferably formed at lower pH.

For BSN, Liu et al. (2007)²³ solved the structure that had been previously reported by Lazarini (1978),¹⁵ describing it as a dumbbell-like cluster consisting of two bridged [Bi₆O₅(OH)₃]⁵⁺ polynuclear species: Bi₁₂O₁₀(OH)₆(NO₃)₁₀·6H₂O. Contaminant uptake by this Bi cluster in the BSN starting material was the foci of our work.

In our studies, we used groundwater simulants containing chloride (Cl⁻), sulfate (SO₄²⁻), bicarbonate/carbonate (HCO₃⁻/CO₃²⁻), silicate (SiO₃²⁻), and nitrate (NO₃⁻),^9–11 to determine the effect of these common groundwater anions on the removal of contaminants of interest. We showed that the starting Bi polynuclear species transformed to different Bi mineral phases, and that the co-mingled effect of different anions in solution changed the Bi transformation paths. More detailed discussion is given in Escobedo et al. (in prep.).¹⁰

The layered crystalline structures of these Bi minerals exhibit electronic and spatial flexibility, due to the lone-pair electronic effect.²⁴ This flexibility allows them to sequester various contaminants as negatively charged species. The layered structure is composed of Bi–O sheets ([Bi₂O₂]²⁺ in a fluorite-related structure) and two large groups, the Aurivillius and Sillén, cover numerous Bi minerals of various composition – e.g., BiPb₂OCl, Bi₂WO₆, BiBa₂Nb₂O₉, Bi₄Ti₃O₁₂etc., − with alternating arrangement of anions of different nature between sheets of [Bi₂O₂]²⁺.^1,25

Bismutite Bi₂O₂(CO₃) and bismoclite BiOCl – layered Sillén related and Sillén structures, respectively – are two of the most common bismuth minerals found in the environment under ambient conditions. Bismutite is more thermodynamically stable (bismutite ΔG_f = −916.2 kJ mol⁻¹; bismoclite ΔG_f = −322 kJ mol⁻¹) and will dominate except under acidic conditions and the presence of chloride (eqn (2)).²⁶

Bi₂O₂CO_3(s) + 2H⁺_(aq) + 2Cl⁻_(aq) = 2BiOCl_(s) + CO_2(g) + H₂O (2)

In this series of experiments with pertechnetate (TcO₄⁻), ⁹⁹Tc radioisotope, and its common non-radioactive analogue perrhenate (ReO₄⁻), we used Hanford Site simulated groundwater (HSGW) representative of the Hanford Site Central Plateau. We also used ReO₄⁻ as a non-radioactive analogue to TcO₄⁻ as they have similar physicochemical properties: low hydration energies (−251 kJ mol⁻¹ for TcO₄⁻ and −330 kJ mol⁻¹ for ReO₄⁻) and charge densities on the O atoms (low enthalpic contribution to the binding process); the same charge volumes 0.048 Å⁻³ for both anions; similar Re–O (1.72–1.73 Å) and Tc–O distances (1.71 Å);²⁷ similar ionic radius 252 pm for TcO₄⁻ and 260 pm for ReO₄⁻ explaining their homologous chemical properties.^28,29 Many studies in the literature use Re as a Tc surrogate, e.g., ReO₄⁻ sequestration by iron minerals, clays, polymers, metal–organic frameworks, etc.^30–34 Hence, a comparative analysis of TcO₄⁻ and ReO₄⁻ behavior with respect to uptake by Bi-based materials is of interest to the wider scientific community.

Materials and methods

Caution! ⁹⁹Tc is a β⁻ emitter; energy 0.29 MeV (ref. 35) and should be handled in a suitably controlled facility. The experimental work with radioactive samples was conducted in a nuclear research facility by trained personnel in a fume hood designated for radiological contamination control.

Materials

Bismuth(III) subnitrate, BSN, and bismuth(III) oxychloride, BiOCl, were procured from Sigma Aldrich. A cluster compound Bi₆O₄(OH)₄(NO₃)₆ was synthesized following the procedure described in Henry et al. (2006).¹⁹ Briefly, 0.5 g Bi₂O₃ was dissolved in 2 mL concentrated HNO₃ and added dropwise to a 50 : 50 water/ethanol solution (200 mL) with stirring. The precipitate was filtered, washed with deionized water (DI water), air-dried at room temperature, and characterized by PXRD. Groundwater simulant, here HSGW (Hanford simulated groundwater), was prepared using chemical reagents of analytical grade: silicic acid, H₂SiO₃·xH₂O – Sigma Aldrich, potassium chloride, KCl – Fisher Chemical, magnesium carbonate, MgCO₃ – Sigma Aldrich, sodium chloride, NaCl – Fisher Chemical, calcium sulfate dihydrate, CaSO₄·2H₂O – Acros Organics, calcium carbonate, CaCO₃ – Baker. The in-house solution of ammonium pertechnetate, NH₄TcO₄, was prepared at the Radiochemical Processing Laboratory (Pacific Northwest National Laboratory), and sodium perrhenate, NaReO₄, was acquired from Sigma Aldrich. Double deionized water of at least 18 MΩ cm was used to prepare all solutions.

Batch kinetics experiments

The preparation procedure for HSGW was adopted from Truex et al., 2017 (i.e. Table 9).³⁶ Final pH of the HSGW after filtering was 8.1 ± 0.1. The initial TcO₄⁻ and ReO₄⁻ stock solutions were prepared in DI water to have concentrations 40 mM KTcO₄ and 0.1 M NaReO₄. The same liquid to solid ratio was used in all batch experiments, i.e. 100 mL of liquid and 0.1 g of the solid bismuth material. The preparation of the samples was as follows: 100 mL of HSGW were added to a 250 mL Nalgene bottle, spiked with either TcO₄⁻ or ReO₄⁻ stock solution to have a target concentration 0.17 mM of each contaminant. After that, 0.1 g of bismuth material was mixed into the solution, samples were placed on an orbital shaking table at 200 rpm (MaxQ 2000, A class, J-Kem Scientific), and time was recorded as the starting point of the experimental series. Samples were taken every day for 3 days after the start of the experiment, followed by weekly sampling. Sacrificial samples were terminated at different times to examine changes to the Bi-based materials after exposure to the contaminants for up to almost two months. The control series had the same liquid to solid ratio of HSGW to BSN, excluding the contaminants; and the same concentrations of TcO₄⁻ and ReO₄⁻, excluding BSN. Aliquots from the TcO₄⁻ series (50 or 100 μL) were filtered into 10 mL of liquid scintillation cocktail (Ultima Gold LLT, Perkin Elmer), and analyzed using a liquid scintillation counter (Tri-Carb 3100TR, PerkinElmer). The data were processed with correction for the background. Aliquots from the ReO₄⁻ series (1 mL) were collected for inductively coupled plasma mass spectrometry (ICP-MS). Both TcO₄⁻ and ReO₄⁻ aliquots were filtered using 0.45 μm pore size syringe filters (PTFE membrane, 25 mm, EZFlow). pH data were collected systematically for the ReO₄⁻ and control series.

All experiments were conducted under aerobic conditions, at ambient temperature and pressure.

Powder X-ray Diffraction (PXRD)

The PXRD analysis with BSN starting material was conducted upon the termination of the experiments, at 59 days, when the samples were centrifuged, aqueous phase was decanted and the solid phase air dried. Additional samples were prepared and terminated for PXRD analysis at 24–28 and 45–49 days after the start of experiment. The PXRD patterns were collected using Cu Kα radiation (λ = 1.5418 Å) employing a Rigaku Miniflex II XRD unit. The Miniflex was operated at 30 kV and 15 mA and patterns were collected between 3 and 90° 2θ, scanning at 2° min⁻¹ at 0.02° 2θ intervals. Identification of material phases was carried out using JADE® XRD pattern processing software (Materials Data Incorporated, CA) and reference patterns from the International Centre for Diffraction Data (ICDD) powder diffraction database.

Results and discussion

In our earlier study we showed that the starting BSN material is the Bi₁₂O₁₀(OH)₆(NO₃)₁₀·6H₂O compound described by Lazarini (1978)¹⁵ and Liu et al. (2007),²³ comprised of bridged hexanuclear [Bi₆O₅(OH)₃]⁵⁺ clusters (Fig. S1, ESI†).⁹ It is helpful to distinguish between the dominant structural motifs in Bi compounds, and so we use the prefix “clus-” to denote compounds based on clusters, “lay-” for compounds based on layered [Bi₂O₂]²⁺ sheets, “dis-” for disordered compounds which give very broad PXRD peaks, and “unk-” for compounds where the structure is unknown. Hence, BSN is denoted as clus-Bi₁₂O₁₀(OH)₆(NO₃)₁₀·6H₂O.

The clusters are charge balanced by anions, with the NO₃⁻ in the starting material, released and replaced by OH⁻, Cl⁻, SO₄²⁻ and CO₃²⁻, depending on composition of aqueous media.¹⁰ In deionized water, there is decrease in pH (Fig. 1), as clusters undergo NO₃⁻ ligand exchange with OH⁻ (eqn (3)). The pH decrease is less significant in aqueous media containing HCO₃⁻/CO₃²⁻ ions that participate in ligand exchange and buffer pH. Evidence of release of NO₃⁻ from BSN starting material in DI water with consequent pH drop was shown in our previous study.¹⁰

[Bi₆O₄(OH)₄](NO₃)₆ + (6 − x)H₂O = [Bi₆O₄(OH)₄](NO₃)_x(OH)_(6−x) + (6 − x)H⁺ + (6 − x)NO₃ (3)

BSN starting material, clus-Bi₁₂O₁₀(OH)₆(NO₃)₁₀·6H₂O composed of [Bi₆O₅(OH)₃]⁵⁺ polycations. One day of BSN contact with DI water resulted in formation of clus-Bi₆O₄(OH)₄(NO₃)₆ structure,¹⁰ similar to what was described in Henry et al. (2006).¹⁹ This provides evidence of transformation of [Bi₆O₅(OH)₃]⁵⁺ to [Bi₆O₄(OH)₄]⁶⁺ polycations in aqueous solution at low pH (eqn (1)). The transformation paths differed in groundwater simulants of different composition – the role of anions in mineral transformation of BSN material with time was evaluated in our earlier study.¹⁰

Fig. 1 Change of pH with time in the 0.17 mM and 2.0 mM ReO₄⁻ series with 0.1 g BSN starting material in 100 mL HSGW simulant. Control series: HSGW – groundwater simulant only; Re-HSGW – 0.17 mM Re in the simulant; HSGW-BSN – 100 mL groundwater simulant only; Re-HSGW – 0.17 mM Re in the simulant; HSGW-BSN – 100 mL groundwater simulant and 0.1 g BSN starting material. Error bars within the symbols are ± 1 st. dev. of triplicate measurements.

Fig. 1 shows changes of pH in the samples with BSN and 0.17 and 2.0 mM of ReO₄⁻ as compared to ReO₄⁻ in the groundwater simulant without BSN and to the groundwater simulant itself, where pH did not change by the end of the experiment. The decrease in pH caused by mineral transformation of BSN in HSGW simulant without ReO₄⁻ is shown for the control sample HSGW-BSN.

There were similar changes in pH trend, plateauing around 4.4, in the 0.17 mM ReO₄⁻ sample and HSGW-BSN control (BSN in the groundwater simulant without perrhenate) while pH decreased slightly below 4.0 in the 2.0 mM ReO₄⁻ sample.

This trend in pH for the 0.17 mM ReO₄⁻ samples agrees with pH measurements in our previous study for 0.17 mM TcO₄⁻ samples with BSN in HSGW at the same solution-to-solid ratio,¹⁰ where pH dropped to 6.7 after 1 day. After 60 days of the contact time, pH was 3.8, no intermediate pH measurements were performed. In that study, pH of the HSGW-BSN blank was 6.5 after 1 day and 3.5 after 60 days, which implies slightly higher buffering capacity of the batch of HSGW simulant prepared for these experiments, resulting in pH of 4.3 at the end of the experiment. The HSGW is saturated with respect to CaCO₃, and the extent of saturation (thereby the buffering capacity) will vary depending on temperature, exposure to atmosphere etc. The overall similar trend between the studies shows that the pH data for the 0.17 mM ReO₄⁻ series is representative of the 0.17 mM TcO₄⁻ series.

NO₃⁻ anions in the BSN starting material undergo rapid ligand exchange with OH⁻, with pH decreasing below 6 after only three hours from the start of the experiment and plateauing after 168 hours (7 days) from the start of the experiment. Addition of 0.17 mM ReO₄⁻ to the HSGW simulant without BSN had no impact on the pH time profile. Consequently, we would expect contaminants to be removed by 7 days, or 168 hours, if the main mechanism for their removal is related to changes in BSN material caused by anion exchange. The removal kinetics for TcO₄⁻ (Fig. 2a) showed that 93% of the initial 0.17 mM TcO₄⁻ was removed from HSGW after 168 hours from the start of the experiment, but only 78% of the ReO₄⁻ was removed over the same time period (Fig. 2b). It took up to 672 hours for ReO₄⁻ removal to be below ICP-MS detection limit (0.027 mM of Re). The rate constant for TcO₄⁻ removal within the first 72 hours (1.2 ± 0.2) × 10⁻⁴ min⁻¹ was slightly higher than the rate constant for ReO₄⁻ removal (0.7 ± 0.2) × 10⁻⁴ min⁻¹ with an initial concentration of 0.17 mM and over the same time period (Fig. 2b).

Fig. 2 Kinetics of TcO₄⁻ and ReO₄⁻ removal with BSN starting material (0.1 g) in HSGW simulant (100 mL). A. Removal of 0.17 mM TcO₄⁻. Control series: 0.17 mM TcO₄⁻ in HSGW simulant. B. Removal of 0.17 and 2.0 mM ReO₄⁻. Control series: 0.17 mM ReO₄⁻ in HSGW simulant (only select time points were analyzed). Insets: First order linear models for 72 hours (4320 min). ND – not detected (ICP-MS EQL for Re was 0.027 mM).

The rate constant for ReO₄⁻ removal from the 2.0 mM sample was (1.1 ± 0.3) × 10⁻⁴ min⁻¹ for the first 72 hours. Concentrations of Re in solution at 72 hours were 0.13 mM and 1.3 mM for 0.17 mM and 2.0 mM ReO₄⁻ samples respectively. Later, after one week, it reached equilibrium, removing only, on average, 67% from the initial 2.0 mM Re concentration. Lower uptake of ReO₄⁻ in the 2.0 mM sample at the later time can be related to the different initial Bi to Re molar ratio, which in the 0.17 mM ReO₄⁻ sample was around 20.6, while in 2.0 mM ReO₄⁻ sample only 1.7. The equilibrium pH of this sample (3.91 ± 0.06) was lower compared to the 0.17 mM ReO₄⁻ sample (4.43 ± 0.08) revealing a greater extent of NO₃⁻ exchange for OH⁻ in the Bi material during transformation.

The differences in the TcO₄⁻ and ReO₄⁻ removal kinetics with BSN imply different removal mechanisms, and we employed PXRD analysis to investigate the nature of the BSN mineral transformation in these experiments.

A summary of the PXRD patterns for TcO₄⁻ or ReO₄⁻ in contact with BSN material is given in Fig. 3, including the PXRD pattern for the blank, i.e. BSN in HSGW.

Fig. 3 Bi minerals after contact of the Bi subnitrate, or BSN, starting material with 0.17 mM TcO₄⁻ or 0.17 mM ReO₄⁻ in HSGW simulant for different number of days. Blank is BSN in HSGW simulant. Bi rhenium oxide Bi₃ReO₈ indicated as “o”; daubreeite related structure lay-BiO(OH,Cl) indicated as “d”; and the unknown phase unk-Bi(NO₃)_x(OH)_yO_z indicated as “x”. Numbers show peak positions of unidentified phases. Only the most clear and prominent peaks are marked. Disordered phase dis-BiO_w(OH)_x(NO₃)_y is evident as broad peaks in all patterns, except 2.0 mM Re 49 days.

With both 0.17 mM ReO₄⁻ and TcO₄⁻, the layered daubreeite related structure, lay-BiO(OH,Cl), and a disordered structure, with a general chemical formulae dis-BiO_w(OH)_x(NO₃)_y defined in our previous study,⁹ were present in all patterns. The disordered structure transformed into the unknown, i.e. not resolvable by PXRD analysis, unk-Bi(NO₃)_x(OH)_yO_z phase by 59 days (growing peaks in the region of 27–31° 2θ). Formation of this structure is consistent with our previous study, where unk-Bi(NO₃)_x(OH)_yO_z and lay-BiO(OH,Cl) formed after 60 days in the presence of TcO₄⁻.¹⁰ The blank sample was less crystalline, composed of lay-BiO(OH,Cl) and dis-BiO_w(OH)_x(NO₃)_y structures. This sample is less crystalline compared to the blank sample in the previous study exposed to HSGW for 60 days.¹⁰ The blank samples were prepared identically for these two studies, but with different batches of the HSGW variable in carbonate saturation and buffering capacity. Differences in crystallinity and amount of the disordered phase in the blanks show that phase transformation of BSN starting material is sensitive to the presence of carbonate in a groundwater simulant. The disordered structure was described in more detail in Pearce et al.,⁹ where the broad diffraction pattern was like that for a metastable δ-polymorph of bismite (Bi₂O₃). The presence of 0.17 mM TcO₄⁻ and ReO₄⁻ in the HSGW enhanced the crystallinity of the solid phase, compared to the blank further enhanced in the 2.0 mM ReO₄⁻ sample (Fig. 3).

The experiments were conducted under aerobic conditions and no color change was observed. We assume that TcO₄⁻, as well as ReO₄⁻, is removed by non-specific adsorption and ion exchange with bismuth subnitrate and preserves its chemical form as pertechnetate, or perrhenate, anion.⁹ Despite expected similarities in TcO₄⁻ and ReO₄⁻ uptake by BSN, a different product, Bi rhenium oxide (Bi₃ReO₈), was formed in the presence of ReO₄⁻. The relative amount of Bi₃ReO₈ to lay-BiO(OH,Cl) decreased with time in 0.17 mM ReO₄⁻ series and is more prominent in the 28 day Bi–Re sample than in the 49 day Bi–Re sample, disappearing in the 59 day Bi–Re sample. We hypothesize that this is due to the prolonged dissolution of Bi₃ReO₈. This hypothesis is supported by the Pourbaix diagram showing that ReO₄⁻, not Bi₃ReO₈(s), is the predominant Re species at low pH (Fig. S2 (ref. 37) ESI†). The PXRD pattern for the 2.0 mM Bi–Re sample showed a more crystalline structure comprised of two clearly resolvable mineral phases, lay-BiO(OH,Cl) and Bi₃ReO₈ by 49 days (Fig. 3). Compared to 0.17 mM ReO₄⁻ sample at 49 days, the relative amount of Bi₃ReO₈ to lay-BiO(OH,Cl) was larger in the 2.0 mM ReO₄⁻ sample, suggesting that a larger fraction of the initial BSN material participated in formation of Bi₃ReO₈ in the 2.0 mM ReO₄⁻ sample. This is similar to the formation of lay-BiO(OH,Cl), which is controlled by the amount of Cl⁻ in HSGW.

The presence of Bi₃ReO₈ in 0.17 mM Re sample and absence of Bi₃TcO₈ in 0.17 mM Tc sample is demonstrated in Fig. S3 ESI.† The formation of different products as a result of TcO₄⁻ and ReO₄⁻ uptake by the BSN starting material was an unexpected outcome of this study, as it was anticipated that these two oxyanions would behave in the same way, given their similar physical and chemical properties. However, it is known from the literature that ReO₄⁻ undergoes changes in coordination more readily than TcO₄⁻, and it has a lower hydration energy (−330 kJ mol⁻¹ for ReO₄⁻ and −251 kJ mol⁻¹ for TcO₄⁻); with both only weakly binding to positively charged binding sites (ligands with strong specific coulombic interactions).²⁷ The difference in the oxidizing potentials of Re^VII/Re^IVvs. Tc^VII/Tc^IV is −0.51 vs. −0.74, meaning that Tc^VII is more prone to reduction than Re^VII at the same conditions; and it was demonstrated in the study with steel coupons that TcO₄⁻ was predominantly reduced to Tc^IV and bound to the steel (iron) corrosion products, while ReO₄⁻ remained oxidized and did not bind to the corrosion products.³⁸ It is also described that for Bi–Re–O and Bi–Tc–O systems, Bi₃ReO₈ and Bi₂Tc₂O₇ are the predominant phases (studied under non-oxidizing conditions), respectively, where Re has oxidations state VII and Tc has oxidations state IV; Bi₃TcO₈ (Tc^VII) synthesis was only possible from the Bi₂Tc₂O₇ (Tc^IV) with Bi₂O₃ in O₂ upon heating.²⁹ Formation of Bi₂Tc₂O₇ (the precursor to Bi₃TcO₈) is not expected in this study as it would require reducing conditions (Tc^VII to Tc^IV).

Formation of the Bi₃ReO₈ phase resulted in slightly slower removal of ReO₄⁻ by BSN compared to the removal of TcO₄⁻ by BSN without formation of a Tc-bearing Bi phase, under the same experimental conditions (starting concentration of 0.17 mM). Formation of Bi₃ReO₈ may also explain the plateau at around 70% ReO₄⁻ removed for the higher starting concentration of 2.0 mM (Fig. 2b), as compared to almost complete removal with a starting concentration of 0.17 mM ReO₄⁻. PXRD shows that formation of Bi₃ReO₈ is the initial ReO₄⁻ removal mechanism. At a starting ReO₄⁻ concentration of 0.17 mM, the initial molar ratio of Bi to Re was 20.6, which is sufficient to removal all of the ReO₄⁻ from solution by formation of the Bi₃ReO₈ phase. At a starting ReO₄⁻ concentration of 2.0 mM, the molar ratio of Bi to Re was at most 1.7 in the system, which is less than the ratio of 3 required to form the Bi₃ReO₈ phase. Therefore, there was insufficient Bi in the system to remove all the ReO₄⁻ from solution by formation of Bi₃ReO₈, and in this case 42% of ReO₄⁻ should have remained in solution if no uptake was caused by lay-BiO(OH,Cl). The fact that there was approximately 30% of the ReO₄⁻ suggests some surface interaction together with incorporation into the Bi solid phase.

After 59 days in the presence of 0.17 mM TcO₄⁻, the dominant Bi mineral phases were unk-Bi(NO₃)_x(OH)_yO_z and the daubreeite-related structure lay-BiO(OH,Cl). We hypothesize that the unk-Bi(NO₃)_x(OH)_yO_z phase was responsible for Tc removal from the HSGW simulant. To test this hypothesis, bismoclite (lay-BiOCl or BiOCl; Fig. S4 ESI†), which is isostructural with the daubreeite-related structure lay-BiO(OH,Cl), was used as the starting Bi phase for TcO₄⁻ removal (Fig. S5a and b†). No TcO₄⁻ removal from HSGW was observed up to 59 days (Tc average fraction of 1.00 ± 0.01) and pH plateaued at 6.00 ± 0.05 after 24 days likely indicating completion of the phase transformation in this time. The PXRD patterns revealed that both the daubreeite-related structure lay-BiO(OH,Cl) and bismutite lay-Bi₂O₂(CO₃) were present after 24 and 59 days, with the amount of bismutite, qualitative estimate, increasing over time (Fig. 5a). Thus, interlayer exchange of Cl⁻ ions in bismoclite for OH⁻ and CO₃²⁻ ions in HSGW resulted in bismuth mineral transformation (eqn (2)), but the bismutite and daubreeite-related phases formed did not uptake TcO₄⁻.

A final set of experiments was conducted to compare TcO₄⁻ removal by the commercial BSN with TcO₄⁻ removal by a synthesized clus-Bi₆O₄(OH)₄(NO₃)₆ (Fig. S6 ESI†). It was more effective for TcO₄⁻ removal in the short-term, with 73% removed after 1 day (9.1 ± 0.1) × 10⁻⁴ min⁻¹ (Fig. 4a), compared to 25% removed by BSN over the same time period. However, clus-Bi₆O₄(OH)₄(NO₃)₆ took longer to reach near quantitative (99%) TcO₄⁻ removal (59 days compared to 14 days for BSN, Fig. 2a), implying a slow process of mineral incorporation of TcO₄⁻ (unidentified mineral phase). With clus-Bi₆O₄(OH)₄(NO₃)₆, pH decreased more rapidly to 3.68 ± 0.08 after 34 hours (Fig. 4b), compared to 4.93 ± 0.01 after 24 hours for BSN. PXRD patterns revealed that both clus-Bi₆O₄(OH)₄(NO₃)₆ and BSN formed unk-Bi(NO₃)_x(OH)_yO_z and lay-BiO(OH,Cl) after 59 days (Fig. 5b). However, the peaks corresponding to the unk-Bi(NO₃)_x(OH)_yO_z phase are clearly visible in the PXRD pattern after 1 day with clus-Bi₆O₄(OH)₄(NO₃)₆, in both test and blank samples, but are not visible until after 59 days with BSN. This suggests that the readily available [Bi₆O₄(OH)₄]⁶⁺ polycations in clus-Bi₆O₄(OH)₄(NO₃)₆ rapidly transformed into the unk-Bi(NO₃)_x(OH)_yO_z phase, whereas BSN required an additional step to convert the [Bi₆O₅(OH)₃]⁵⁺ polycations in its structure to [Bi₆O₄(OH)₄]⁶⁺, followed by transformation to unk-Bi(NO₃)_x(OH)_yO_z.

Fig. 4 A. Removal of 0.17 mM TcO₄⁻ with clus-Bi₆O₄(OH)₄(NO₃)₆ starting material (0.1 g) in HSGW simulant (100 mL). Control series: 0.17 mM TcO₄⁻ in HSGW simulant. Inset: First order linear model for up to 24 hours (1440 min); first order constant k = (9.1 ± 0.1) × 10⁻⁴ min⁻¹. B. PH measurements for control series: HSGW only and HSGW with clus-Bi₆O₄(OH)₄(NO₃)₆.

Fig. 5 Bi minerals after contact with 0.17 mM TcO₄⁻ in HSGW simulant for different number of days. Blanks are the corresponding Bi materials in HSGW simulant. A. Bismoclite lay-BiOCl starting material. Daubreeite related structure lay-BiO(OH,Cl) indicated as “d”; bismutite lay-Bi₂O₂(CO₃) indicated as “b”. B. clus-Bi₆O₄(OH)₄(NO₃)₆ starting material. Daubreeite related structure lay-BiO(OH,Cl) indicated as “d”; and the unknown phase unk-Bi(NO₃)_x(OH)_yO_z indicated as “x”. Numbers show peak positions of unidentified phases.

Formation of unk-Bi(NO₃)_x(OH)_yO_z also occurs in the absence of TcO₄⁻ through ligand exchange of BSN (clus-Bi₁₂O₁₀(OH)₆(NO₃)₁₀·6H₂O) and synthesized clus-Bi₆O₄(OH)₄(NO₃)₆, which are present as [Bi₆O₄(OH)₄]⁶⁺ polycations in acidic solution (eqn (1) and (3)). However, transformation of both clus-Bi₆O₄(OH)₄(NO₃)₆ and BSN in the presence of TcO₄⁻ results in the formation of an additional unidentified phase (with peaks at approximately 25.8, 32.4, and 34° 2θ), which has also been observed in our previous studies.¹⁰ We hypothesize that mineral transformation into both this unidentified phase, and the unk-Bi(NO₃)_x(OH)_yO_z phase, is involved in the uptake of TcO₄⁻ (Fig. 3 and 5b), and further studies are needed to resolve the unknown Bi mineral phases. Peaks absent in the blanks, but present in the experiments with TcO₄⁻, regardless of the starting Bi material (17 and 21.5° 2θ, Fig. 3 and 5a and b) also require additional evaluation.

Overall, removal of TcO₄⁻ and ReO₄⁻ from HSGW by the in situ transformation of clus-Bi₁₂O₁₀(OH)₆(NO₃)₁₀·6H₂O (BSN) and clus-Bi₆O₄(OH)₄(NO₃)₆ starting materials is dependent on ligand exchange of Bi polynuclear oxo-hydroxo species, or clusters, removing OH⁻ from solution and decreasing pH. Several research questions remain for future studies. For example, identification of the structure and formation mechanism of unk-Bi(NO₃)_x(OH)_yO_z; evaluation of long-term transformation of the unk-Bi(NO₃)_x(OH)_yO_z phase, possibly, into bismutite at higher pH and bismoclite as more thermodynamically stable products of Bi mineral transformation (eqn (2)), which may affect long term performance for TcO₄⁻ removal from groundwater in the subsurface.

Conclusions

Basic Bi subnitrate, BSN, starting material showed effective TcO₄⁻ removal from HSGW reaching 93% within 168 hours (7 days). Experiments were conducted up to almost two months and no release of removed TcO₄⁻ back to the aqueous phase was observed. BSN transformation by this time resulted in the “unknown” unk-Bi(NO₃)_x(OH)_yO_z phase and the mineral phase with the unidentified peaks, consistent with our previous studies,¹⁰ which were hypothesized to be responsible for TcO₄⁻ uptake. Formation of another mineral phase, daubreeite-related structure lay-BiO(OH,Cl), did not participate in TcO₄⁻ removal.

The same starting material showed different removal kinetics for ReO₄⁻ from HSGW, the rate constants for 0.17 mM ReO₄⁻ removal vs. 0.17 mM TcO₄⁻ removal within the first 72 hours were (0.7 ± 0.2) × 10⁻⁴ min⁻¹ and (1.2 ± 0.2) × 10⁻⁴ min⁻¹, respectively. Along with unk-Bi(NO₃)_x(OH)_yO_z and lay-BiO(OH,Cl) phases formation of the intermediate Bi₃ReO₈ mineral phase was indicated by PXRD. Analogous mineral phase was not observed in the pertechnetate series of experiments, despite similarities in physical and chemical properties of these two species. Therefore, we conclude that BSN starting material is a promising efficient material for subsurface remediation, but more research is needed to investigate its long-term mineral transformation, characterization and investigation of the pH dependent formation of unk-Bi(NO₃)_x(OH)_yO_z and another unidentified phase, mechanisms of pertechnetate removal, as well as relevance to the conditions of the subsurface environment. Moreover, care should be taken when its non-radioactive analogue is being used for these and other kinds of studies.

Data availability

The data that support findings of this study are available from the corresponding author, D. Boglaienko, upon request.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

Pacific Northwest National Laboratory (PNNL) is operated by Battelle Memorial Institute for the U.S. Department of Energy (DOE) under Contract DE-AC05-76RL01830. We acknowledge the Deep Vadose Zone, Applied Field Research Initiative at PNNL. Funding for this work was provided by the DOE Richland Operations Office.

Notes and references

1. T. G. Levitskaia, N. P. Qafoku, M. E. Bowden, R. M. Asmussen, E. C. Buck, V. L. Freedman and C. I. Pearce, A Review of Bismuth (III)-Based Materials for Remediation of Contaminated Sites, ACS Earth Space Chem., 2022, 6, 883–908.
2. F. Franceschini, P. Jagdale, M. Bartoli and A. Tagliaferro, Perspectives on the use of bismuth based materials for sensing and removal of water pollutants, Curr. Opin. Environ. Sci. Health., 2022, 100345.
3. S. Nagendran, P. Sivasubramanian, J. H. Chang, S. Y. Shen, A. K. Nayak and M. Kumar, Fundamentals of environmental remediation techniques, in Bismuth-Based Materials for Environmental Remediation, IOP Publishing, 2022.
4. H. Chawla, A. Chandra, P. P. Ingole and S. Garg, Recent advancements in enhancement of photocatalytic activity using bismuth-based metal oxides Bi2MO6 (M = W, Mo, Cr) for environmental remediation and clean energy production, J. Ind. Eng. Chem., 2021, 95, 1–15.
5. Y. Liu, B. Yang, H. He, S. Yang, X. Duan and S. Wang, Bismuth-based complex oxides for photocatalytic applications in environmental remediation and water splitting: A review, Sci. Total Environ., 2022, 804, 150215.
6. J. Z. Hassan, A. Raza, U. Qumar and G. Li, Recent advances in engineering strategies of Bi-based photocatalysts for environmental remediation, Sustainable Mater. Technol., 2022, 00478.
7. Z. Wang, M. Chen, D. Huang, G. Zeng, P. Xu, C. Zhou, C. Lai, H. Wang, M. Cheng and W. Wang, Multiply structural optimized strategies for bismuth oxyhalide photocatalysis and their environmental application, Chem. Eng. J., 2019, 374, 1025–1045.
8. A. R. Lawter, T. G. Levitskaia, O. Qafoku, M. E. Bowden, F. C. Colon and N. P. Qafoku, Simultaneous immobilization of aqueous co-contaminants using a bismuth layered material, J. Environ. Radioact., 2021, 237, 106711.
9. C. I. Pearce, D. Boglaienko, A. R. Lawter, E. A. Cordova, K. J. Cantrell, M. E. Bowden, N. Lahiri, O. Qafoku, S. T. Mergelsberg, C. T. Resch, F. C. Colon, V. A. Garayburu-Caruso, N. D'Annunzio, M. Balasubramanian, S. A. Saslow, N. P. Qafoku, V. L. Freedman and T. G. Levitskaia, Mechanisms of interaction between bismuth-based materials and contaminants for subsurface remediation, in prep.
10. N. M. Escobedo, S. A. Saslow, E. C. Cordova, A. R. Lawter, M. E. Bowden, O. Qafoku, C. T. Resch, N. Lahiri, N. D'Annunzio, D. Boglaienko, T. G. Levitskaia, C. I. Pearce, V. L. Freedman and R. D. Mackley, Part I: Structural transformation of bismuth-based materials in dynamic aqueous environments and implications for subsurface contaminant remediation, in prep.
11. A. R. Lawter, N. P. Qafoku, E. C. Cordova, M. E. Bowden, O. Qafoku, F. C. Colon, N. D'Annunzio, D. Boglaienko, N. M. Escobedo, T. G. Levitskaia, C. I. Pearce and V. L. Freedman, Part II: Sediment interactions with layered bismuth materials and implications for subsurface contaminant remediation, in prep.
12. D. T. Richens, The chemistry of aqua ions: synthesis, structure, and reactivity: a tour through the periodic table of the elements, Wiley, New York, 1997.
13. N. N. Greenwood and A. Earnshaw, Chemistry of the Elements, Elsevier, 2nd edn, 2005.
14. A. N. Christensen, M. A. Chevallier, J. Skibsted and B. B. Iversen, Synthesis and characterization of basic bismuth (III) nitrates, J. Chem. Soc., Dalton Trans., 2000, 265–270.
15. F. Lazarini, The crystal structure of a bismuth basic nitrate, [Bi₆O₅(OH)₃](NO₃)₅·3 H₂O, Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem., 1978, 34, 3169–3173.
16. B. S. Brčić, D. Kolar, F. Lazarini and M. Malešič, Oxydation of bismuth with atmospheric oxygen in the presence of diluted nitric acid, Monatsh. Chem., 1973, 104, 365–375.
17. N. Henry, M. Evain, P. Deniard, S. Jobic, F. Abraham and O. Mentré, [Bi2O2] 2+ layers in Bi2O2 (OH)(NO3): synthesis and structure determination, Z. Naturforsch. B, 2005, 60, 322–327.
18. F. Lazarini, Bismuth Basic Nitrate [Bi₆(H₂O)(NO₃)O₄(OH)₄](NO₃)₅, Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem., 1979, 35, 448–450.
19. N. Henry, O. Mentré, F. Abraham, E. J. MacLean and P. Roussel, Polycationic disorder in [Bi6O4(OH)4](NO3) 6: structure determination using synchrotron radiation and microcrystal X-ray diffraction, J. Solid State Chem., 2006, 179, 3087–3094.
20. L. Miersch, T. Ruffer, M. Schlesinger, H. Lang and M. Mehring, Hydrolysis studies on bismuth nitrate: synthesis and crystallization of four novel polynuclear basic bismuth nitrates, Inorg. Chem., 2012, 51, 9376–9384.
21. A. N. Christensen and B. Lebech, Investigation of the crystal structure of a basic bismuth (iii) nitrate with the composition [Bi6O4 (OH) 4] 0.54 (1)[Bi6O5 (OH) 3] 0.46 (1)(NO3) 5.54 (1), Dalton Trans., 2012, 41, 1971–1980.
22. N. Henry, M. Evain, P. Deniard, S. Jobic, O. Mentré and F. Abraham, [Bi6O4. 5 (OH) 3.5] 2 (NO3) 11: a new anhydrous bismuth basic nitrate. Synthesis and structure determination from twinned crystals, J. Solid State Chem., 2003, 176, 127–136.
23. B. Liu, W. W. Zhou, Z. Q. Zhou and X. Y. Zhang, Hydrolysis to the first dumbbell-like high-nuclearity bismuth-oxo cluster [Bi12 (μ3-OH) 4 (μ2-OH) 2 (μ3-O) 8 (μ4-O) 2 (NO3) 6] 4+: Synthesis, structure and spectroscopic characterizations, Inorg. Chem. Commun., 2007, 10, 1145–1148.
24. D. M. Driscoll, R. C. Shiery, N. D'Annunzio, D. Boglaienko, M. Balasubramanian, T. G. Levitskaia, C. I. Pearce, N. Govind, D. C. Cantu and J. L. Fulton, Water Defect Stabilizes the Bi3+ Lone-Pair Electronic State Leading to an Unusual Aqueous Hydration Structure, Inorg. Chem., 2022, 61, 14987–14996.
25. S. Sun and X. Yin, Progress and perspectives on aurivillius-type layered ferroelectric oxides in binary Bi4Ti3O12-BiFeO3 system for multifunctional applications, Crystals, 2020, 11, 23.
26. M. E. Clissold, Aspects of the supergene geochemistry of copper, nickel and bismuth, 2007.
27. E. A. Katayev, G. V. Kolesnikov and J. L. Sessler, Molecular recognition of pertechnetate and perrhenate, Chem. Soc. Rev., 2009, 38, 1572–1586.
28. C. Eiroa-Lledo, L. Lecrivain, T. G. Parker, D. E. Wall and N. A. Wall, Comparison of ReO4− and TcO4− in solvent extraction systems, Radiochim. Acta, 2020, 108, 443–449.
29. E. E. Rodriguez, F. Poineau, A. Llobet, K. Czerwinski, R. Seshadri and A. K. Cheetham, Preparation and crystal structures of bismuth technetates: a new metal oxide system, Inorg. Chem., 2008, 47, 6281–6288.
30. B. A. Lenell and Y. Arai, Perrhenate sorption kinetics in zerovalent iron in high pH and nitrate media, J. Hazard. Mater., 2017, 321, 335–343.
31. B. Wakoff and K. L. Nagy, Perrhenate uptake by iron and aluminum oxyhydroxides: An analogue for pertechnetate incorporation in Hanford waste tank sludges, Environ. Sci. Technol., 2004, 38, 1765–1771.
32. E. M. Pierce, W. W. Lukens, J. P. Fitts, C. M. Jantzen and G. Tang, Experimental determination of the speciation, partitioning, and release of perrhenate as a chemical surrogate for pertechnetate from a sodalite-bearing multiphase ceramic waste form, Appl. Geochem., 2014, 42, 47–59.
33. B. H. Hamilton, T. A. Wagler, M. P. Espe and C. J. Ziegler, Sequestering perrhenate with a borate-based coordination polymer: A model for pertechnetate separation, Inorg. Chem., 2005, 44, 4891–4893.
34. Q. H. Hu, Y. G. Wang, X. Gao, Y. Z. Shi, S. Lin, R. P. Liang and J. D. Qiu, Halogen microregulation in metal-organic frameworks for enhanced adsorption performance of ReO4-/TcO4, J. Hazard. Mater., 2023, 130744.
35. K. Schwochau, Technetium, Wiley-VCH, 2000.
36. M. J. Truex, J. E. Szecsody, N. Qafoku, C. E. Strickland, J. J. Moran, B. D. Lee, M. Snyder, A. R. Lawter, C. T. Resch and B. N. Gartman, Contaminant Attenuation and Transport Characterization of 200-DV-1 Operable Unit Sediment Samples, Pacific Northwest National Laboratory, Richland, WA (United States), 2017.
37. A. M. Patel, J. K. Nørskov, K. A. Persson and J. H. Montoya, Efficient Pourbaix diagrams of many-element compounds, Phys. Chem. Chem. Phys., 2019, 21, 25323–25327.
38. S. M. Heald, K. M. Krupka and C. F. Brown, Incorporation of pertechnetate and perrhenate into corroded steel surfaces studied by X-ray absorption fine structure spectroscopy, Radiochim. Acta, 2012, 100, 243–253.

---

Footnote

† Electronic supplementary information (ESI) available: PXRD patterns for BSN starting material, BSN after 0.17 mM Tc and Re uptake, pristine BiOCl, and clus-Bi₆O₄(OH)₄(NO₃)₆ synthesized material; removal kinetics and pH profile for 0.17 mM Tc uptake with BiOCl; and a Pourbaix diagram. See DOI: https://doi.org/10.1039/d4ew00496e

---