Hydrodynamic Decontamination of Groundwater and Soils Using ZVI

Abstract

Polluted aquifers can be decontaminated using either ZVI (zero valent iron) permeable reactive barriers (PRB) or injected ZVI. The placement of ZVI within the aquifer may take several decades to remediate the contaminant plume. Remediation is further complicated by ZVI acting as an adsorbent to remove some pollutants, while for other pollutants, it acts as a remediation catalyst. This study investigates an alternative aquifer decontamination approach to PRB construction or n-Fe⁰ injection. The alternative approach reconstructs the potentiometric surface of the aquifer containing the contaminant. This reconstruction confines the contaminant plume to a stationary, doughnut shaped hydrodynamic mound. Contaminated water from the mound is abstracted, decontaminated, and then reinjected, until all the water confined within the mound is decontaminated. At this point, the decontaminated mound is allowed to dissipate into the surrounding aquifer. This approach is evaluated for potential use in treating the following: (i) immiscible liquid plumes; (ii) miscible contaminant and ionic solute plumes; (iii) naturally contaminated aquifers and soils; and (iv) contaminated or salinized soils. The results indicate that this approach, when compared with the PRB or injection approach, may accelerate the decontamination, while reducing the overall amount of ZVI required.

1. Introduction

Over the last three decades, permeable reactive barriers (PRBs) containing zero valent iron (Fe⁰, ZVI) [1,2,3,4,5,6,7] and injected n-Fe⁰ [8,9,10] have been used to decontaminate (confined and unconfined) aquifers and soils. This technology is addressed by >1000 patents and patent applications, together with >10,000 academic publications (e.g., references [1,2,3,4,5,6,7,8,9,10]). A PRB places a reactive permeable wall or barrier containing Fe⁰ into an aquifer, in order to intercept the flowing, contaminated water plume. The injected n-Fe⁰ approach either creates a PRB by n-Fe⁰ injection, or is associated with the direct injection of n-Fe⁰ into the contaminated plume or soil [11,12,13,14].

The initial Fe⁰ PRB patent (US5266213) focused on the treatment of organo-halides. Since then, the commercial process has been adapted to include decontamination of a variety of organic pollutants (e.g., KR100476115B1; KR20060027634A; US5611936; US5616253; US575389); and a variety of inorganic pollutants (e.g., As (US6132623; US6387276); Cd (CN105502671B); Pb (CN105502671B)). The initial patents (e.g., US5266213) constructed the PRB as a vertical barrier (0.3 to 1.5 m wide) that extended from the ground surface to the base of the contaminated aquifer. The PRB was filled with Fe⁰ particles. These barriers may be straight or funneled. In addition, they may include sections that are non-reactive.

Some more recent patents have created a PRB using boreholes to target deeper aquifers, unconsolidated aquifers, and soils. Five groups of borehole approaches have been adopted. The first approach (e.g., US5975798) constructs a Fe⁰ slurry, which is then combined with a pressurizing gas for delivery into the subsurface (by injection). The aquifer can be fracked before or during this process. The second approach (e.g., US8210773; US8366350; US85960351; US9061333; US9943893) places one or more horizontal wells within the aquifer. These are then filled with Fe⁰ to create a PRB. A third approach (e.g., US6664298; US7037946; US7008964; US7271199; US7582682; US8062442; US8163972) directly injects a n-Fe⁰–oil–water-emulsion-slurry into the aquifer or into the contaminant plume to create a barrier, or zone of remediation. A fourth approach (e.g., US5641020; US5733067; US7179381) hydraulically fractures the contaminated aquifer and uses the Fe⁰ as a reactive proppant to effect remediation. A fifth approach (e.g., US 5857810) injects a colloidal suspension of Fe⁰ into the aquifer or soil to create a PRB. This list of variants is not exclusive. The decontamination rates may be enhanced by using the PRB as an electrode, or by constructing it to include or use electrodes to facilitate electrolytic decontamination (e.g., US5868941; CN1899717B; US8968550B2; US10500618B2). A number of different patent specifications (e.g., US5833388; US5975800; US6207114; US9884771) address the hydrodynamic design of a PRB in order to focus the fluid flow within the aquifer towards a PRB, through the use of nonreactive barriers. The selection of an installed PRB, or n-Fe⁰ injection, to address a contaminant problem is a commercial decision. This decision will consider the aquifer properties, the location, and the contaminant present in the water.

1.1. Fe⁰ Particle Size

Three sizes of Fe⁰ are used for water remediation. They are as follows: n-Fe⁰ (particle size = 1–1000 nm); m-Fe⁰ (particle size = 1000 nm–1 mm); and Fe⁰ (particle size = 1 mm–200 mm). From a practical perspective, only n-Fe⁰ (particle size 1 = 100 nm) or m-Fe⁰ is suitable for injection into an aquifer as part of a water-ZVI slurry. Fe⁰ (0.5 to 5 mm) is suitable as a reactive proppant in some fracked boreholes. Fe⁰ and m-Fe⁰ (and n-Fe⁰ precipitated onto inert supports in the micron-millimeter particle size) are used in PRBs. The underlying chemical assumption for the PRB decontamination approach is that the remediation reaction occurs on the particle surface. If there are x remediation sites m⁻² on the surface of a particle of Fe⁰, then it follows that, by decreasing the particle size, or by increasing the degree of irregularity on the particle surface, it is possible to increase the surface area (a_s, m²) per unit weight (g) of Fe⁰. This creates a natural assumption that particle reactivity increases with increasing particle surface area (decreasing particle size).

Oxidation of the particles’ surface by water converts the Fe⁰ to Fe_xO_yH_z (rust). The rust crust, comprising Fe_xO_yH_z species, tends to be porous (Figure 1). It has a lower density than Fe⁰. This density can approach 1 g cm⁻³ in more hydrated and porous forms. The surface area, associated the Fe_xO_yH_z species, is typically within the range 30 to 800 m² g⁻¹. If Fe⁰ reactivity is solely a function of a_s, then it would be reasonable to expect the Fe⁰ to be less reactive than the Fe_xO_yH_z. While it is true that n-Fe⁰ is more reactive than Fe⁰, it is not always true that Fe_xO_yH_z is more reactive than n-/m-/Fe⁰. This is because different redox (Eh, pH) conditions are present on the Fe⁰ surface, within the Fe_xO_yH_z and at the Fe_xO_yH_z–water interface, including hydration shells (Figure 1).

In an aqueous environment, at any specific temperature and pressure, the redox conditions define the reaction quotient for each pollutant and the range of product outcomes that can occur (i.e., the environment is Faradaic or Nernstian). A further complication is that some pollutants form precipitants, which adhere to the Fe⁰ particle surface. When the precipitant is a result of a Fe–pollutant reaction, this will reduce the number of active sites available for pollutant removal, and will reduce the reactivity.

1.2. Fe⁰ Catalysis

Fe⁰ catalyzes the redox reduction of H₂O: (H₂O = OH⁻ + H⁺; Log(OH⁻) = −14 − pH). This results in the water pH increasing (relative to the water body) at the Fe⁰–water interface (within its hydration shell). Chemical diffusion between the Fe⁰ hydration shell and the water body results in the pH of the water body increasing with time. However, the pH of the water body is always less than the pH of the Fe⁰ hydration shell. The Fe_xO_yH_z corrosion crust (Figure 1) tends to remove the produced OH⁻ ions. This results in a reduction in pH. Therefore, the pH of the hydration shell that is associated with the Fe_xO_yH_z corrosion crust, will be less than the pH associated with the Fe⁰ hydration shell at the Fe⁰: Fe_xO_yH_z interface (Figure 1).

The Eh (V) of the Fe⁰ hydration shell varies with pH. In acidic water, the Fe⁰ corrosion reaction at the Fe⁰:water interface within the hydration shell is as follows: Fe⁰ = Fe²⁺ + 2e⁻ (Eh, V = −0.44 + 0.0295Log(Fe²⁺)). In alkaline water, the reaction within the Fe⁰ hydration shell is as follows: Fe⁰ + 2H₂O = Fe(OH)₂ + 2H⁺ + 2e⁻ (Eh, V = −0.047 – 0.0591 pH). At the Eh within the Fe⁰ hydration shell, a proportion of the H⁺ will form H₂ (H₂ = 2H⁺ + 2e⁻; Eh = −0.0591 pH–0.0295 Log(pH₂)). From this brief discussion, it is evident that the principal controls in the ionic reaction quotient for a dissolved contaminant are the Eh and pH. Obviously, some ionic species will interact directly with Fe⁰ to form a chemical precipitate or an adsorbed species. Contaminant removal by a direct reaction with Fe⁰, or by adsorption onto Fe⁰, will normally result in a maximum of 20–300 mg of contaminant removed g⁻¹ Fe⁰. By way of contrast, redox remediation has been demonstrated to remove >1000 g of contaminant g⁻¹ Fe⁰. These concepts are addressed further in the Appendix A, Appendix B, Appendix C, Appendix D, Appendix E, Appendix F, Appendix G, Appendix H and Appendix I.

This study focuses on the redox removal of contaminants. This requires a modification of the aquifer’s pH and Eh to achieve the required remediation level. pH = −Log[H₃O⁺] = −Log[H⁺], where a hydronium ion molar concentration of 1 × 10⁻⁸ provides a pH of −Log [1 × 10⁻⁸] = 8. Eh is the potential (volts, V) of the water determined using a standard hydrogen electrode. The Standard EMF (electromotive force) ΔE^o of an ideal ionic reaction (e.g., Fe⁰ = Fe²⁺ + 2e⁻) is determined at a pH of 0. The Eh measurement adjusts ΔE^o for pH, temperature, the molar reaction quotients, and for the competing reactions in the water. Eh can be measured directly, using a standard hydrogen electrode (SHE). Eh is most commonly measured indirectly, using an ORP (oxidation reduction potential) meter, which has been calibrated to the standard hydrogen electrode. The linear calibration between the ORP and the Eh is normally undertaken at pH 4 and 7, using quinhydrone reference solutions.

1.3. Contaminant Plumes

Contaminant plumes, within an aquifer or soil, fall into three basic groups: (i) historical anthropogenic contamination, which needs to be removed; (ii) ongoing pollution, associated with continuing anthropogenic activity. PRB’s which are associated with continuing anthropogenic activity, are designed to have a long life, of perhaps 20–50 years; and (iii) natural contamination of the aquifer with heavy metals, (including As, Se), NaCl or another contaminant.

This study provides an alternative solution (to a PRB, or aquifer n-Fe⁰ injection) for the following: (i) the decontamination of contaminant plumes, resulting from historical anthropogenic activity; (ii) the periodic decontamination of contaminant plumes, resulting from continuing anthropogenic activity; (iii) the decontamination of water, containing naturally acquired pollutants (e.g., As, Se, NaCl, etc.).

1.4. Permeability Reduction

The PRB remediation approach permanently reduces the aquifer porosity and permeability [4,5,15,16,17]. This reduction occurs within the treatment areas (PRB, injected zone and the adjacent sediments). They permanently leave Fe_xO_yH_z polymers and their related minerals within the water [16,17]. They also permanently increase the pH (and decrease the Eh) of the water within the vicinity of the Fe⁰ and polymers [18]. These polymers act as adsorption sinks for many cations and anions within the water [19,20,21,22,23,24,25]. These remediation approaches are costly [17] and may take decades to completely, or partially, decontaminate the aquifer.

ZVI can reduce the permeability of a PRB (and its flow rates) by >2 orders of magnitude within 2 months of installation. Consequently, a PRB with an initial permeability of 10⁻⁸ m³ m⁻² s⁻¹ Pa⁻¹, may, after 2 months, have a permeability of 10⁻¹⁰ m³ m⁻² s⁻¹ Pa⁻¹. A PRB with a design expectation of treating x m³ d⁻¹, would, after 60 days, have a reduced flow rate through the PRB of 0.01 × x m³ d⁻¹. This would increase the time required to treat the contaminant plume by a factor of 100. This is not a new ZVI operational problem. In the commercial surface-based reactors, the problem is normally addressed through the use of a revolving drum or moving bed reactors (e.g., US439588; US439589; US439590; US443737; US458887; US458946; US462537; US513536; US513686; US559816; US559817; US563811), slurry/fluidized bed reactors (e.g., US2086753; US6533499), high porosity fixed bed reactors (e.g., US5803174; US6787034), or diffusion reactors (e.g., GB2520775A). In PRBs, the permeability issue is sometimes addressed by the dilution of the ZVI with inert material (e.g., US7347647; US20210213498A1), through the use of emulsifiers (e.g., US6664298; US7037946; US7008964; US7271199; US7582682; US8062442; US8163972; USRE40448E1), or through the use of ZVI support systems (e.g., US6787034; US7670082; US8262318). These permeability reduction concepts are addressed further in the Appendix A, Appendix B, Appendix C, Appendix D, Appendix E, Appendix F, Appendix G, Appendix H and Appendix I.

1.5. Modelling

It is possible to model the expected hydrological and chemical performance of a PRB prior to installation [26,27,28,29]. However, issues such as Fe⁰ corrosion (Figure 1), PRB by-pass by flowing water within the aquifer, reactivity loss with time, and both PRB and aquifer permeability loss over time, can lead to the conclusion that, for some applications and locations, the installation of a PRB is a commercially inefficient use of the Fe⁰. In most, but not all, ZVI reactors and ZVI remediation applications, the ZVI is reusable. Therefore, a single charge of ZVI may be able to process >100 volumes of water. A typical PRB may contain 500–1000 t of Fe⁰, and may process 50 to 1000 m³ of water over its life time, Without Fe⁰ recycle, in a surface-based facility, 1000 t of Fe⁰ may be able to process 33,000 m³. With Fe⁰ recycle, the 1000 t of Fe⁰ charge may be able to process >3,000,000 m³. Therefore, a commercial argument can be made, to the effect that a more efficient use of the ZVI would be to undertake the aquifer remediation in a surface-based reactor. This is the approach adopted in this study.

1.5.1. Contaminant Plume Availability

Globally, there are over 1 million sites that have been contaminated with toxic metals, chlorinated organics, nitro-organics, hydrocarbons, petrochemicals, agricultural chemicals, industrial chemicals, mining waste water, associated water produced with hydrocarbon extraction, flowback water associated with oil/gas extraction from shales, and foul (sewage) waste [26]. This is in addition to sites where the aquifer is naturally polluted with As, Se, heavy metals, nitrates and chlorides.

1.5.2. ZVI Operation

ZVI, when placed in water, corrodes [30]. The corrosion products are a mixture of ions and polymers (Fe⁰:Fe(a,b,c) [31,32,33,34] (Figure 1). Abrasion of the iron particles can release the corrosion products into the water as n-Fe(a,b,c) polymer colloids. A typical released polymer is spherical, has a 7% hydration shell, and may have a diameter of between 6 and 11 nm [35]. The polymers are degraded in acid to release water (US10913665B2), and are preferentially formed in alkaline water [36].

The acid treatment of ZVI, prior to use, has the following effects: (i) it reduces the water body Eh, which is associated with a specific batch of ZVI, relative to the control water body Eh, using the unacidified ZVI charge (US10913665B2); (ii) it increases the effectiveness of the ZVI when measured as the cumulative H₂ (moles) released g⁻¹ Fe⁰ (US10913665B2); and (iii) it maintains a similar water body pH to the control water body pH, using the unacidified ZVI charge. (US10913665B2). These observations, allow the relative efficiency of ZVI acidification to be assessed using a pH-specific Eh parameter (PSE,V), which is defined as (US10913665B2):

PSE, V = Eh/pH,

(1)

The PSE value determined for acidified ZVI is smaller than the PSE value determined for unacidified ZVI, and can be negative (US10913665B2). The rate of contaminant removal and the total amount of contaminant removed by a ZVI charge increases with a decreasing PSE (US10913665B2).

1.5.3. ZVI Definition

The term ZVI is used in this study to refer to Fe⁰ and its associated corrosion products and polymers (Figure 1). The ZVI interacts with pollutants (cations and anions) within the water by reaction, adsorption and catalysis.

The adsorption and reaction models for the ZVI decontamination of water consider that ZVI becomes ineffective once the available reaction or adsorption sites are removed. Adsorption, reaction and corrosion (where the corrosion removes active sites) progressively reduce the effectiveness of ZVI with time [37,38]. Consequently, the amount of ZVI used in a PRB or an n-Fe⁰ injection is frequently substantially in excess of the optimal charge required. PRB and n-Fe⁰ injection provide a tried, proven, and relatively low technical risk approach, which can be used to remediate an aquifer.

1.5.4. Significance of Iron Bacteria

Some biological experiments have demonstrated that, if the PRB or the injected n-Fe⁰ is contaminated with Shewanella putrefacients, then the aging of the Fe⁰ (Figure 1) can be partially reversed; this allows an in situ reactivation of the Fe⁰ to occur [39,40]. All PRBs and zones of Fe⁰ injection contain iron bacteria. These bacteria use the Fe²⁺ ions present within the pore spaces of the PRB to form an external shell of Fe(OH)₃ or FeOOH (Figure 2). Once the bacteria dies, these shells (exoskeletons) form high permeability tubes within the corrosion crust. In these PRBs, the normal abiogenic decline in the particulate ZVI permeability with time may be partially offset by the development of residual, high permeability iron bacteria exoskeletons (Figure 2). These may extend through the corrosion crusts of the Fe⁰ (Figure 1).

1.6. Aquifer Contaminants

Aquifer contaminants fall into five broad categories:

• Group 1: anthropogenic, immiscible fluid plumes within a soil or aquifer (Figure A1a), e.g., oils [41,42,43,44,45,46,47,48,49,50];
• Group 2: anthropogenic, miscible and soluble contaminant fluid plumes (Figure A1b), e.g., organo-halides, organo-nitrates, explosive residues, pesticides, herbicides, dyes, etc. [23,49,50,51,52,53,54];
• Group 3: seawater incursion of saline water into an aquifer (Figure A1c) [55,56,57]. This can be associated with anthropogenic aquifer depletion in coastal areas [56,57];
• Group 4: dispersed naturally sourced contaminants (Figure A1c), e.g., As [58,59,60], Se [61,62,63,64], NaCl, etc., within a confined aquifer, unconfined aquifer, or a soil;
• Group 5: soil contaminated by salinization associated with agricultural activity or natural processes, soils contaminated by chemicals and left by chemical spills, and chemicals in vertically migrating groundwater mounds in areas of anthropogenic activity [65,66,67].

1.7. Reconstruction of the Aquifer Potentiometric Surface

Groups 1, 2 and 3 plumes, and their plume boundaries, migrate with time within the aquifer. Their migration rates and the direction of migration are controlled by the potentiometric surface within the aquifer, and by the aquifer permeability and recharge/discharge point sources that affect the aquifer.

The potentiometric surface within an aquifer is relatively easy to reconstruct [68,69,70,71]. Fluid will flow from a point of high (natural or artificial) potential (e.g., high pressure, or head), to a point of low (natural, or artificial) potential (e.g., low pressure or head) [68].

If the fluids within an aquifer are flowing in a direction x, the insertion of a ring of recharge points (injection wells, infiltration devices) around the plume can be used to create a mound of stationary fluid within the aquifer [68] (Figure A1). If the center of the mound contains an abstraction well and the abstracted water is returned to the injection wells, then the ring of recharge points will surround a stationary hydrodynamic mound, or plume, of fluid. This stationary plume of fluid will have little, or no, effective interaction with the surrounding water body [68] (Figure A1).

While the detailed hydrology of the stationary plume is complex and site specific, the general process design of the stationary plume is relatively simple. Leakage into, and out of, a plume, is a function of abstraction and injection rates. If the abstracted water is passed through a reactor containing ZVI before being reinjected, then it will be possible to decontaminate the constructed stationary plume in a controlled manner. This can be achieved within a definable and predictable time frame, using a defined amount of ZVI.

1.7.1. Group 1 and 2 Plumes

If the original plume was a Group 1 plume (e.g., gasoline) or a Group 2 plume (e.g., trichloroethylene, dissolved heavy metal salts, dissolved radionuclides, dissolved organic pollutants, etc.), then following decontamination, the stationary plume can be allowed to dissipate into the wider aquifer environment. This dissipation requires the abstraction and injection operations to cease. Site remediation will require the plugging of the injection and abstraction wells, coupled with a removal of their well heads.

1.7.2. Seawater Incursion

Group 3 seawater incursion is associated with an aquifer that has both freshwater recharge and seawater recharge at different points along its length. The potential associated with seawater recharge is controlled by the sea level, and will cyclically vary during the day. The magnitude of the daily cycle variation will vary during a year. The magnitude of this variation is a function of the local tidal range. In most inland freshwater aquifers, the initial water table elevation is above that of sea level. Anthropogenic aquifer depletion reconstructs the potentiometric surface within the aquifer to a negative slope from the sea towards the land within the depleted area (i.e., the net flow is from the sea towards the land). This results in seawater incursion into the aquifer. The reconstruction of the potentiometric surface in this area, through the creation of stationary plumes, may allow the incursion to be stabilized or reversed, and allow the salinized section of the aquifer to be desalinated.

1.7.3. Group 4 Aquifers

The water contained within Group 4 aquifers, containing NaCl, As, Se, etc., can be decontaminated through the creation of hydrodynamic stationary plumes. Once the plume has been decontaminated, it can be used to provide a source of decontaminated water for anthropogenic use.

1.7.4. Group 5 Contaminated Soils

The soils, located below areas of anthropogenic activity, may become contaminated with the organic chemicals, herbicides, inorganic chemicals, sewage, pesticides, metals, nitrates, oils, or salts, that are associated with the infiltration of water, develop at mineral pans, or are concentrated at locations of permeability change within the soil. These Group 5 contaminated soils are located above the regional confined, unconfined, or perched aquifers. A process of infiltration and abstraction can be used to create hydrodynamic groundwater mounds within these soils. These hydrodynamic groundwater mounds can be used to decontaminate the soils.

1.8. Study Purpose

This study is a scoping study that is designed to determine whether the Group 1 to Group 5 hydrodynamic stationary plumes and groundwater mound routes may provide a viable technical alternative for aquifer decontamination through the use of a PRB or injected n-Fe⁰. The evaluated analysis process flow is summarized in Figure 3.

2. Materials and Methods

This scoping study integrates hydrodynamic modelling, chemical process modelling and published experimental ZVI chemical remediation data to assess the viability of using a hydrodynamic aquifer reconstruction (Figure 3).

Two hydrodynamic approaches are considered, and are as follows:

• Firstly, the formation of stationary hydrodynamic plumes (SHP) in both confined and unconfined aquifers. The issues associated with the advection–diffusion migration of pollutants [72] across the hydrodynamic boundaries of the SHP, due to chemical potential gradients, is not addressed in this study; the detailed modelling of the aquifer recharge using wells [73] associated with the periphery of the SHP is outside the scope of this study;
• Secondly, the formation of perched groundwater mounds (PGM) in soils located above an unconfined aquifer or an aquitard [74,75,76]. These stationary plumes (PGM) are designed to decontaminate part of the groundwater contained within an unconfined aquifer or soil. These groundwater mounds are created using a freshwater source. The PGM is used to dissolve salts (contaminant ions) within the soil, or an unconfined aquifer. These ions are then abstracted with the water, from the PGM, to allow for ion recovery using a surface located ZVI reactor, and for the appropriate disposal of the recovered pollutants.

The Appendix A, Appendix B, Appendix C, Appendix D, Appendix E, Appendix F, Appendix G, Appendix H and Appendix I are used to present the underlying hydrological/hydrodynamic/chemical principles, process assumptions, methods, and illustrative examples allow remediation to occur. They are documented as follows:

• Appendix A—Process assumptions
  • Stationary Hydrodynamic Plume Assumptions
  • Example Stationary Hydrodynamic Plume
  • Perched Groundwater Mound Assumptions
• Appendix B—Reaction Process Assumptions
  • Confined or Unconfined Aquifer Treatment
  • Stationary Groundwater Mound Treatment
• Appendix C—ZVI Formation
  • Formation via an Fe₃O₄ Intermediary
  • n-Fe⁰ Formation using Borohydrides
  • n-Fe⁰ Formation using Dithionite or Hydrosulfite Ions (S₂O₄²⁻)
  • n-Fe⁰ Formation using Hydrazine
  • n-Fe⁰ Formation using Organic Reductants
  • Regeneration of n-Fe⁰
• Appendix D—Contaminant Removal Kinetic Models
  • Adsorbent Models
• Appendix E—Contaminant Processing Assumptions
  • Redox Remediation
• Appendix F—Reactor Assumptions;
• Appendix G—Creation of a Stationary Hydrodynamic Plume
• Appendix H—Formation of a Groundwater Mound
  • Chemical Potential
  • Standard Infiltration Groundwater Mound
  • Vertically Static, Laterally Expanding Groundwater Mound
  • Reconstructing a Perched Groundwater Mound to Remediate the Soil
  • Modelling the Remediation of soil using a Groundwater Mound
• Appendix I—Arsenic Removal
  • Unsupported n-ZVI
  • Active Carbon (AC)-Supported n-ZVI (C⁰@n/m-Fe⁰)
  • Biochar (BC)-Supported n-ZVI (C⁰@n/m-Fe⁰)

The appendices contain Figure A1, Figure A2, Figure A3, Figure A4, Figure A5, Figure A6, Figure A7, Figure A8, Figure A9, Figure A10, Figure A11, Figure A12, Figure A13, Figure A14, Figure A15, Figure A16, Figure A17, Figure A18, Figure A19, Figure A20, Figure A21, Figure A22, Figure A23, Figure A24, Figure A25, Figure A26, Figure A27, Figure A28, Figure A29 and Figure A30,

Table A1. Parameters associated with the hydrodynamic stationary plume.

Item	Parameter	Value
1	Aquifer Thickness, m	5
2	Aquifer Porosity	30%
3	Aquifer Permeability, m3 m−2 s−1 Pa−1
3a	Vertical (z) direction	5 × 10−6
3b	Horizontal (x) direction	5 × 10−6
3c	Horizontal (y) direction	5 × 10−6
4	Stationary plume radius, m	150
5	Gross Rock volume within the plume, m3	353,429
6	Net to Gross Ratio for the aquifer	1
7	Saturations
7a	Irreducible Water Saturation
7aa	Tixier Model	30.19%
7ab	Timur Model	3.66%
7ac	Coates Dumanoir Model	28.70%
7ad	Assumed for Modelling	30.00%
7b	Reducible Water Saturation	70.00%
8	Total Mobile Water volume, m3	74,220
9	Ground area covered by the stationary plume, ha	7.07

and Equations (A1)–(A73).

2.1. Data Sources

This study utilizes measured kinetic data and stoichiometric information from the academic literature in order to establish what the constraints on remediating an aquifer, using a stationary plume or groundwater mound, may be.

2.2. Hydrological Modelling

For the purposes of this study, all modelling is performed at a relatively high level. It is assumed that the basic flux equation applies, where the flow rate, J, at any point located at a 3D location, x, y, z (where x, y and z, are spatial reference points), is defined for flow in a x direction, J_x, as follows:

J_x, m³ m⁻² s⁻¹ = k_x P_x,

(2)

where k_x = permeability in a x direction, m³ m⁻² s⁻¹ P⁻¹; P_x = pressure, Pa, potential, or head, expressed in pressure of the potentiometric surface, head (m, ft), or vertical length, relative to a reference elevation. Solving the basic flux equation for simultaneous flow in the x, y and z directions is normally undertaken using a variant of the Richards equation. In this study, the hydrodynamic stationary plume modelling assumes homogeneity, where k_x = k_y = k_z. The hydrodynamic perched groundwater plume modelling assumes k_x = k_y > k_z. The x, y, z subscripts indicate the flow directions.

The 3D flow rates between each cell, where the stationary plume or groundwater mound is subdivided into n × 1 m³ cells, is assumed to be constant with time, when P remains constant.

The hydrodynamic stationary plume structure requires that: J entering the plume (J₁) = J exiting the plume (J₂). When (J₁) > (J₂), the outer margins of the plume expand into the surrounding aquifer. The amount of expansion is controlled by the lateral dissipation of the potential associated with inflow. When (J₁) < (J₂), the plume may receive fluids from the surrounding aquifer.

In the field, assumptions and analyses have to be made about aquifer heterogeneity, the Dietz shape factors [77,78], and the associated Euler constants [78], which are associated with both the abstraction and injection points. The modelling associated with both the abstraction wells [77,78,79] and injection wells can be very complex [80,81,82,83,84,85].

In this study, it is assumed that the location of the injection and abstraction points have considered these items, and that (J₁) = (J₂).

2.3. Aquifer Remediation Modelling

For the purposes of this study, it is assumed that chemical potential differences will ensure that all the soluble ions held in the irreducible water saturation (S_wir) will equilibrate with the concentration of the same ions in the mobile water saturation (S_w). This assumption allows the surface remediation of S_w, followed by reinjection, in order to effectively remove the pollutant from the S_wir. The remediation flowchart adopted in this study is summarized in Figure 4.

3. Modelling Remediation

Figure 3 and Figure 4 indicate that a key element in the decontamination process is the ZVI reactor. There are a large number of diverse views as to how ZVI remediates pollutants [7,86,87,88,89,90]. ZVI is a smart material and can operate using several routes simultaneously [91].

The behavior of ZVI is unusual. Its sensitivity to changes in Eh, pH, temperature, pressure, light, and chemical compounds, allow it to be self-sensing and to undertake a multitude of functions simultaneously in a controlled manner [91]. In this study, ZVI is assumed to be able to simultaneously undertake a number of catalytic, hydrogenation, redox, and auto-activation reactions.

3.1. Fe-Reaction, Adsorption and Redox Interpretations

Two assumptions permeate through the ZVI and PRB literature:

• Assumption 1: ZVI reacts directly with a contaminant ion to remove it. Contaminant removal is by reduction, while the Fe⁰ is oxidized. The evidence for this interpretation is based on the observation that a soluble contaminant has been reduced to allow its precipitation. At the same time, the Fe⁰ has been oxidized to form an oxyhydroxide, e.g., [7]. There is no doubt that the two reaction start and end products occur. For example [7]:

Fe⁰ + CrO₄²⁻ + 8H₃O⁺ = Fe³⁺ + Cr³⁺ + 12H₂O,

(3)

(1 − x)Fe³⁺ + xCr³⁺ + 2H₂O = Fe_(1−x)Cr_xOOH + 3H⁺,

(4)

This model results in 1 g of Fe⁰ removing < 1 g of Cr⁰, and demonstrates that the reactions in Equations (3) and (4) are dependent on each other. The recovered Cr is retained in the PRB. The redox model adopted in this study assumes that the oxidation of Fe⁰ to Fe³⁺, and the reduction of Cr⁶⁺ to Cr³⁺, are independent reactions (e.g., US9909221B2; US10913665B2; US11370677B2). This model assumes that the actual redox remediation and precipitation reaction is as follows [92]:

CrO₄²⁻_(aq) + 5H⁺ + 3e⁻ = C_rOOH_(s) + 2H₂O,

(5)

The driving force for this redox reaction is the availability of H⁺ and e⁻ ions. These are produced as a by-product of the Fe⁰-catalyzed decomposition of water [91]. The equilibrium reaction quotient (Q) [92,93] for this precipitation reaction is favored by decreasing Eh in the water, where at 298 K and 0.1 MPa, the following is determined [92]:

Log(Q) = Log(CrO₄²⁻) = (Eh − (1.386 − 0.0985pH))/0.0197,

(6)

This Nernstian relationship demonstrates that decreasing Eh (V), while maintaining a constant pH, will result in the precipitation of C_rOOH_(s) [92]. This analysis indicates that the redox Eh and pH conditions, suitable for n-C_rOOH precipitation [92], could occur in the Fe⁰ hydration shell, in the Fe_xO_yH_z hydration shells, in the water adjacent to the corroded Fe⁰ (as entrained n-C_rOOH particles (Figure 1)), or on the surface of Fe_xO_yH_z or Fe⁰ particles. If the Cr removal reaction is redox [92], as suggested in reference [92] and patents US9909221B2, US10913665B2, and US11370677B2, then 1 g of Fe⁰ will be able to remove >> 1 g of Cr⁰.

This redox model [92] demonstrates that the oxidation of Fe⁰ is not a necessary pre-requisite for the reduction of CrO₄²⁻_(aq), as required by Approach 1 [7]. Either of these approaches, or an alternative approach, may be the correct explanation for the observed production of CrOOH and Cr(OH)₃ precipitates.

2.

Assumption 2: ZVI reacts directly with the contaminant and chemically adsorbs the contaminant onto its surface [94,95,96,97,98,99]. This approach assumes that the Fe⁰ becomes functionalized, e.g.,

2Fe⁰ + H₂O = Fe⁰:H⁺ + Fe⁰:OH⁻,

(7)

This allows CrO₄²⁻_(aq) to be adsorbed, as follows:

2Fe⁰:H⁺ + CrO₄²⁻_(aq) + 2OH⁻ = 2Fe⁰:H⁺:CrO₄²⁻ + 2OH⁻ = 2FeOOH + CrOOH + H⁺ + 3e⁻,

(8)

The redox approach assumes (Equation (6)) that the rate constant increases for CrOOH production as Log(Q) decreases. A Log(Q) value of >0 requires an Eh of >0.5 at pH = 9. The Fe⁰ surface at pH 9 can react as follows:

Fe⁰ = Fe²⁺ + 2e⁻,

(9)

where

Log(Q) = Log(Fe²⁺) = (Eh − (−0.44))/0.0295,

(10)

At pH = 9, the Eh of the Fe⁰ will be −0.44 V, and Log(CrO₄²⁻) at this surface will be −46.47. Since the rate constant, k, for the forward reaction increases, as Log(Q) decreases, it is probable that the precipitation of CrOOH on the Fe⁰ surface would be virtually instantaneous. In this instance, the reaction would not be an adsorption reaction. Instead, it would be a redox surface precipitation reaction. This interpretation is consistent with reference [92] and patents US9909221B2, US10913665B2, and US11370677B2.

There are numerous experimental examples where the redox conditions in the water body do not allow the precipitation of the ion. However, the soluble ion concentration has been demonstrated to decline with time in the water body. In these examples (e.g., reference [99]), analyses of the iron surface demonstrate either a chemical adsorption or a physical adsorption of the removed ion on the iron surface, or adsorption/concentration within its hydration shell. The redox model [92] interprets this adsorption as indicating that Log(Q) was >0 at the measured locations within the water body. It assumes that the observed concentration at the Fe⁰ surface indicates a lower value of Log(Q) (i.e., Log (Q) <0) on the Fe⁰ surface. This results in a chemical potential difference between the Fe⁰ surface and the water body. This potential difference accounts (under a redox model) for the observed concentration of the ions, with their apparent (or interpreted) adsorption.

The approach taken in this study assumes that all the ZVI remediation reactions that are considered can be interpreted as redox reactions [92]. It is assumed that Log(Q) and the rate constant, k, for the appropriate remediation reaction is a function of the Eh and pH of the water, Fe_xO_yH_z and Fe⁰ (Figure 1). This is demonstrated in Figure A13 and Figure A14. As noted above, this redox remediation view is not universally held, and alternative interpretations of the stoichiometry may be preferred by others.

3.2. Redox Changes

The placement of ZVI in water changes the Eh and pH of the water [100,101]. Continuous flow trials (Appendix E, Figure A13 and Figure A14) demonstrate that the Eh and pH shifts can be maintained through the processing of 1500–25,000 m³ t⁻¹ of ZVI. These shifts place the redox regime of the product water within the zone of n-Fe(a,b,c) polymer formation (Figure A10). This will result (for the example of feed water) in Log(Q) changing sufficiently to result in the precipitation (Figure A11 and Figure A12) of all or part of the ion in the following locations:

(i)

Mode A: On the ZVI surface

(ii)

Mode B: In the hydration shell of the ZVI, where the Eh and pH conditions of the shell are intermediate between those in the water body and on the ZVI surface;

(iii)

Mode C: In the main water body.

The reaction quotient (Log(Q)) at each of these three locations is directly related to the PSE (Equation (1)).

Mode A: Contaminant precipitation on the Fe⁰ surface is dependent on the continuing access of the water (from the water body) to the appropriate redox conditions provided by Fe⁰. This water will travel through the Fe(a,b,c) polymer crust (Figure 1) from the water body by diffusion, and to a lesser extent by viscous flow. Ion removal may cease when the (i) site availability decreases, or when (ii) the formation of oxyhydroxides on the Fe⁰ has resulted in an increase in Eh and/or change in pH, relative to those present on the surface of fresh Fe⁰. This type of precipitation may mimic ion removal, using a conventional adsorption reaction. It is most likely to occur when the redox conditions in the water do not favor rapid precipitation. It is commonly associated with a high chemical potential gradient between Log(Q) on the Fe⁰, and Log(Q) in the water body;

Mode B: This will result in contaminant precipitation, both on the Fe⁰ surface and as entrained nano-particles within the hydration shell. It will also result in precipitation as nano-particles, within the porosity, associated with the Fe⁰. Ion removal may be described using a pseudo-zero to second order reaction.

Mode C: This will result in contaminant precipitation, as entrained nano-particles and aggregated colloids within the water body. It will also result in precipitation on surfaces within the reactor (e.g., conduit walls, reactor walls). Ion removal may be described using a pseudo-first to second order reaction.

3.2.1. pH Changes

It is obvious (Figure A1c) that, if the aquifer pH is P₁, and the plume contains x m^3,, that the abstraction of y m³ with a transformation of pH to P₂ in a reactor, will, following reinjection, result in a revised simplified aquifer pH (R_P) of the following form:

R_P = ((P₁ × (x − y)) + (P₂ × (y))/x,

(11)

Since P₁ = −Log₁₀[H₃O⁺], a pH of 6 signifies that there are 10⁻⁶ moles of H₃O⁺ L⁻¹. If P₁ = a pH of 10, this signifies that there are 10⁻¹⁰ moles of H₃O⁺ L⁻¹. If y = 1 and x = 10, based on pH, R_p = 6.4 (Equation (11)).

This revised aquifer pH may partially reverse the remediation when the remediated contaminant products are retained in the injected/infiltrated water. This is because, while Log(Q) at pH = 6 is positive, and Log(Q) at pH = 10 is negative, Log (Q) in the water will increase (as the pH approaches 6.0) and may turn from negative to positive, thereby favoring the reverse reaction, as follows:

k_obs = k_forward − k_reverse,

(12)

k_obs = observed rate constant; k_forward = forward rate constant; k_reverse = reverse rate constant.

A similar situation occurs with Eh. However, over time, the pH, Eh, and PSE (Equation (1)) of the hydrodynamic stationary plume or hydrodynamic perched groundwater plume will stabilize.

3.2.2. ZVI Characteristics

ZVI has a number of characteristics:

(i)

Its zeta potential in water switches from a positive value to a negative value as the pH increases [100]. The pH (7.5 to 8.5) where this switch occurs (the iso electric point (IEP)) can decrease as the P_w increases. P_w = weight of Fe⁰ in the water, g L⁻¹.

(ii)

For a specific Fe⁰ particle size and P_w, the product water pH increases with time to an equilibrium level [100]. For a specific reaction time, pH increases as P_w increases [100]. Consequently, in a continuous flow reactor containing a fixed volume of Fe⁰, the product water pH decreases as the water flow rate increases. i.e., pH is a function of the space velocity, SV:

SV = F_R/Pv,

(13)

F_R = water flow rate m³ Unit time⁻¹; P_v = volume (m³) or weight of Fe⁰.

(iii)

For a specific Fe⁰ particle size and P_w, the product water Eh decreases with time to an equilibrium level [100]. For a specific reaction time, Eh decreases as P_w increases [100]. Consequently, in a continuous flow reactor containing a fixed volume of Fe⁰, the product water Eh increases as the water flow rate increases. i.e., Eh is a function of the space velocity, SV. This, in turn, establishes that the PSE (Equation (1)) is also a function of SV.

(iv)

pH increases and Eh decreases for a specific value of P_w, as a_s (the particle surface area, m² g⁻¹) increases, and the particle size decreases. It is not unreasonable to expect that, for a particle size of <50 nm, a_s of >20 m² g⁻¹, and a P_w of >0.2 g L⁻¹, the water Eh to reduce by <−0.8 V and the pH to increase by >3 units after approximately 15 min in a batch reactor [100].

3.2.3. Impact of Space Velocity in a Continuous Flow Reactor

Most experimental data is based on a batch operation, where the Fe⁰ is discarded after a single use. The economic viability of the proposed remediation application requires each batch of Fe⁰ to be reused, to process multiple volumes of water.

The experimental results (e.g., Figure A10, Figure A11, Figure A12, Figure A13 and Figure A14) indicate that, in a fixed bed reactor with a single charge of Fe⁰, the continuous flow processing rate (P_R) can exceed 25,000 m³ t⁻¹ of Fe⁰. These experimental results establish that the amount of Fe⁰ required to increase pH and decrease Eh by a constant amount can be assessed, as follows:

P_R, m³ unit weight⁻¹ = SV × t_p,

(14)

where t_p = cumulative processing time. The magnitude of the pH increase and Eh decrease for a specific a_s and specific F_R, increases as P_v (in the reactor) increases. During a prolonged processing time, the Fe⁰ surface will gradually corrode to create a surface covered in Fe(a,b,c) polymers (Figure 1). After the polymer:Fe⁰ ratio exceeds a critical value, the observed PSE will start to increase. When the PSE exceeds a critical value, the Fe⁰ in the reactor will need to be either replaced or regenerated.

3.2.4. Use of Fe⁰ to Produce Entrained n-Fe(a,b,c) Polymers

Some remediation reactors use the Fe⁰ to restructure and repurpose the Feⁿ⁺ ions contained within the water body, in order to produce entrained n-Fe(a,b,c) polymers (which are retained within the water body). These entrained polymers are used to remove pollutants by reaction and adsorption [101]. For these reactions, P_R can exceed 3,000,000 m³ t⁻¹ of Fe⁰ [101].

Other remediation reactors use supported Fe⁰:Fe(a,b,c) [102] or SiO₂@Fe(a,b,c) polymers [103] to catalyse the formation of entrained n-Fe(a,b,c) polymers. These polymers are used to remove pollutants (e.g., Na⁺ and Cl⁻ ions) by reaction and adsorption. For these reactions, P_R can exceed 3,000,000 m³ t⁻¹ of Fe⁰ [103].

3.3. Remediation Rate Constant

All remediation reactions operate at different rates [92,93]. A specific reaction will have a rate constant, which is a function of water composition, pollutant concentration, temperature, pressure, reactor type and turbulence within the reaction environment [92,93,104,105].

The remediation rate constant [93] for a specific a_s, specific F_R, and specific P_v, may decline with time, as the initially available reaction sites on the Fe⁰ become blocked (or poisoned) with precipitated remediation (and oxyhydroxide) products [106].

3.4. Remediation of Immiscible Liquids (Oils)

Immiscible liquids include halogen organics and oils [107,108,109,110,111,112]. As the pollutant plume migrates through the aquifer, it will create a porosity containing mobile water (S_w), irreducible (immobile) water (S_wir), mobile oil (S_o) and irreducible (immobile) oil (S_oir). The ratio of S_o:S_oir will decrease towards the plume margins.

Encasing the pollutant plume within a hydrodynamic stationary mound (Figure A1a), will result in a situation where the mobile oil migrates towards the abstraction well, leaving a residual porosity containing S_w, S_wir and S_oir. The recovered oil, or immiscible fluid, is removed in the reactor. S_oir may account for 20–70% of the oil or immiscible liquid present in the aquifer. A substantial part of the residual irreducible oil, or immiscible liquid, may be recoverable from the aquifer through the injection/infiltration of surfactants or polymers, which are designed to reduce their viscosities, etc. The use of surfactants to recover S_oir is well established in the hydrocarbon industry (US7581594B2; US8146666B2; US20150233223A1), and has been used in the subsurface to help remove contaminant plumes [108,109,110,111,112]. It has been demonstrated that lowering the salinity of the aquifer pore-waters can reduce S_oir and increase S_o (US8550163B2). The ZVI reactor (Figure 3 and Figure 4), may, when remediating an immiscible oil, be used to both degrade the oil [101] and desalinate the water [102,103]; this is prior to the injection/infiltration of the product water into the hydrodynamic stationary plume.

The surface recovery of the water + immiscible liquid will allow the separation and recovery of all or part of the liquid, prior to the residual water being transferred to a ZVI reactor (Figure 3 and Figure 4). A variety of ZVI remediation approaches have been used to remove residual hydrocarbons and their associated miscible liquids and solutes from water. These are documented extensively elsewhere, e.g., references [113,114,115,116,117,118,119].

3.5. Removal of Ionic Salts (Anions and Cations)

Redox ion removal from a hydrodynamic stationary plume or a hydrodynamic perched groundwater mound requires the removal of the precipitate from the product water before reinjection and the precipitation of an ion; this is as either M⁰ MO_x, MOOH, or M(OH)_x within the reactor (Figure 3 and Figure 4).

Using this approach, a suitably designed and operated ZVI redox reactor would expect to recover the following [91]: Ag, Al, Am, Au, As, Ba, Be, Bi, Ca, Cd, Ce, Co, Cr, Cu, Dy, Er, Eu, Fe, Ga, Gd, Ge, Hf, Hg, Ho, In, Ir, In, La, Lu, Mg, Mn, Mo, Nb, Nd, Ni, Np, Os, Pb, Pd, Pm, Po, Pr, Pu, Re, Rh, Ru, S, Sb, Sc, Se, Sm, Sn, Sr, Ta, Tb, Tc, Te, Th, Ti, Tm, U, V, W, Y, Yb, Zn, and Zr.

Monovalent metals (e.g., K, Li, Na) and common anions H_2−xCO₃^x−, H_2−xSO₄^x−, Cl⁻, Br⁻, F⁻, NO_x⁻, H_2−xPO₄^x− can also be removed in a suitably designed and operated ZVI redox reactor [100,101,102,119,120].

The basic principles of ion removal (within a stationary plume or a perched groundwater mound) are the same in each example. In this study, the removal of As (Appendix I) is used to demonstrate the redox removal of a multi-valent ion from a hydrodynamic stationary plume, or perched groundwater mound.

The removal of monovalent ions (e.g., Na⁺, Cl⁻) is more complex, where removal involves catalytic removal by Fe(a,b,c) polymers [120,121] or n-Fe⁰:Fe(a,b,c) polymers [121]. The chemical products associated with ion removal, using the polymers or Fe⁰, have not been defined. However, in saline water, n-Fe⁰ and n-Fe⁰:Fe(a,b,c) polymers reorientate to initially form hydrogen and water-filled spheres [121]. These then grow with time, and sequester (by desorption) ions adsorbed from the water body into their hollow cores (e.g., Na⁺, Cl⁻, Ca²⁺, Mg²⁺, HCO₃⁻, SO₄²⁻) [121]. A typical example is provided in Figure 5. The mechanism associated with this process is described elsewhere [102,122].

n-Fe(b) polymers, when first placed in water, form a mixture of agglomerated networks and colloids (Figure 6a–c) comprised of fluid-filled n-Fe(a,b,c) nano/micron-sized spheres [121]. These spheres sequester ions into their hollow cores (e.g., Na⁺, Cl⁻, Ca²⁺, Mg²⁺, HCO₃⁻, SO₄²⁻). These sequestered minerals (e.g., NaCl) crystallize within the hollow spheres as the polymers are dried (Figure 6d).

4. Applications

The hydrodynamic groundwater plume (or mound), can be used to decontaminate the aquifer or soil. The process duration, as demonstrated in the Appendix A, Appendix B, Appendix C, Appendix D, Appendix E, Appendix F, Appendix G, Appendix H and Appendix I, is a function of the size of the contaminated area and the pollutant considered.

Natural contaminants, such as NaCl or As, can be removed from the hydrodynamic groundwater plume (or mound). The plume can then be used as an anthropogenic source of water. Abstraction for anthropogenic use only occurs when the water composition falls between the pre-set limits (e.g., salinity (NaCl) between 0.5 and 2.5 g L⁻¹ for irrigation applications).

The general remediation process followed for all contaminants is:

• Identify the contaminant plume and define its extent (Figure A1);
• Establish the nature of the aquifer, in terms of permeability, porosity and fluid flow characteristics for both the water and the contaminant;
• Define a target area, and gross rock (aquifer) volume, for the construction of the stationary hydrodynamic mound (Figure A1 and Figure A17), using a mixture of infiltration and abstraction wells;
• Insert the required infiltration and abstraction wells (Figure A17). Start abstracting water from the central part of the mound, and reinjecting/infiltrating water around the periphery (Figure 3). The potentiometric surface across the area will reorientate, to create a stationary hydrodynamic mound, e.g., Figure 7. The hydrology of this change is summarized in Appendix A and Appendix I.
• An idealized mound structure is shown in Figure 7. The idealized mound has no water leakage from the aquifer, into the hydrodynamic mound. In an idealized environment the contaminant concentration will reduce to zero, within the mound, over a short time span (Figure A2a). In reality, there will always be some fluid leakage, from the wider aquifer into the mound. This leakage, will control the amount of remediation, which is possible (e.g., Figure A2b), at a specific aquifer location, when the contaminant, is a natural pollutant (e.g., As). When the contaminant is an anthropogenic pollutant, fluid leakage from the wider aquifer, will not affect the ability of the process to remove the pollutant. The expected pollutant removal profile, in this situation, is shown in Figure A2a.
• The choice of ZVI (Appendix C) used in the reactor (Figure A6), and the reactor type selected will define the reactor size required, the amount of ZVI required, the Eh and pH of the product water (Figure A13 and Figure A14), and the amount of water, which can be processed by, a single batch of ZVI. For example, Appendix I examines (e.g., Figure A29) the removal of As, using three different forms of n-Fe⁰. The same level of removal, could be achieved, with each form of n-Fe⁰. However, the amount of n-Fe⁰ varies, by a factor of 10, between the different types of n-Fe⁰. The example, n-Fe⁰ formulations in Appendix I, indicate that to achieve the idealized outcome, illustrated in Figure A2a, the reactor, processing 40 m³ h⁻¹, would require 400 kg n-Fe⁰ h⁻¹, or 40 kg n-Fe⁰:AC h⁻¹; or 80 kg n-Fe⁰:BC h⁻¹. Using 400 kg n-Fe⁰:AC h⁻¹, would significantly reduce the remediation time (e.g., Figure A2c). Therefore, the process economics and remediation duration, is impacted by the type Fe⁰ selected, and the amount used. These details are site specific.
• The redox modification within the reactor (Figure A13 and Figure A14) will result in the Eh of the stationary mound decreasing with time, while its pH will in-crease with time.

4.1. Example Aquifer Desalination

As an example, a traveling line reactor train (internal volume 3.76 L, 0.0423 m O.D.; water flow rate 0.4 L m⁻¹ (24 L h⁻¹)), processing saline water (pH = 9.06; Eh = 0.215 V; PSE = 0.0237 V; Cl⁻ = 7.89 g L⁻¹; Na⁺ = 6.03 g L⁻¹), produced a product water over 10 h with an average composition of pH = 12.58, Eh = −0.53 V, PSE = −0.042 V, Cl⁻ = 1.39 g L⁻¹, and Na⁺ = 0.97 g L⁻¹, using a sol-gel Fe(a,b,c) polymer constructed from 0.4 g L⁻¹ of FeSO₄ + 0.13 cm³ L⁻¹ of formic acid + 0.06 g L⁻¹ of tartaric acid + 0.38 g L⁻¹ of CaO + 0.16 g L⁻¹ of K₂CO₃; + 0.27 g L⁻¹ of CaCO₃ + 0.18 g L⁻¹ of MnO₂ + 0.36 g L⁻¹ of ZnO. This reactor, achieved an 82.3% Cl⁻ removal and an 83.9% Na⁺ ion removal, over a reaction period of 9.4 min. Increasing the reactor O.D. to 0.2 m would allow 40 m³ h⁻¹ (960 m³ d⁻¹) to be processed using 67 parallel reactor trains.

If the stationary plume held 100,000 m³, then abstraction of 40 m³ h⁻¹ (960 m³ h⁻¹), followed by a reinjection of 40 m³ h⁻¹, would result in a gradual decline in the salinity of the hydrodynamic stationary plume (Figure 8a). This profile form is common to all perched groundwater mounds and hydrodynamic stationary plumes during the initial desalination or decontamination phase.

Once the plume, or mound, has achieved the desired ion concentration, it can be allowed to dissipate into the aquifer or soil. Alternatively, it can be used as a source of decontaminated water for anthropogenic use. This will, in the case of a stationary plume, result in water from the aquifer replacing the abstracted water. This will then lead to a rise in the salinity (or contaminant concentration) of the stationary plume [121]. Once the contaminant concentration exceeds a critical level, abstraction for anthropogenic use ceases until the contaminant concentration reaches a new lower base level [121]. An example of an application where the stationary plume salinity oscillates between a higher value (2.5 g L⁻¹) and a lower value (1 g L⁻¹) during periods of abstraction is provided in Figure 8b.

4.2. Groundwater Mound Decontamination

In many areas, the soil permeability characteristics are not suitable to allow the formation of a perched groundwater mound, and the groundwater mound will descend until it intersects with either an aquitard or an unconfined aquifer. Where the groundwater mound intersects with an unconfined aquifer, it may dissipate into the aquifer [123]; this is after it has initially formed a lens or mound of decontaminated water, resting in/on the upper surface of the unconfined waterbody [123,124]. Continual recharge may make the freshwater lens a permanent or semi-permanent feature that can be exploited for anthropogenic use [123]. The morphology and longevity of the decontaminated lenses is a function of the water density contrast between the decontaminated lens and the surrounding decontaminated aquifer, and the lens recharge rate and the flow rate within the aquifer [125]. These decontaminated (e.g., desalinated) lenses can be used to provide sources of fresh water for anthropogenic use [125]. While previous studies [123,124,125] use natural recharge to create the freshwater lenses, this study has demonstrated that a hydrodynamic descending perched groundwater mound, or a hydrodynamic stationary plume, could be used to produce freshwater lenses, freshwater groundwater mounds, or freshwater stationary plumes in saline confined or unconfined aquifers and soils. This approach may provide a larger and more sustainable access to freshwater within areas where natural freshwater lenses are associated with saline unconfined aquifers.

5. Conclusions

The decontamination of an aquifer or soil that is either naturally contaminated (e.g., saline water) or contaminated as a result of anthropogenic activity, can be undertaken using a PRB or an injected n-Fe⁰. This study has examined whether it is conceptually possible to decontaminate these soils or aquifers using either hydrodynamic stationary plumes or hydrodynamic perched groundwater mounds. The contaminated water is abstracted from the stationary plume or the perched groundwater mound. It is then decontaminated using a reactor containing ZVI, before being reinjected into the stationary plume or the perched groundwater mound.

Integrated hydrological, hydrodynamic and ZVI redox/catalytic/adsorbent kinetic modelling suggests the following:

• Immiscible liquids, which are lighter than water, can be abstracted from the aquifer (or soil) using both a water flood and a surfactant flood approach. The liquids are either recovered or are processed in a surface reactor containing ZVI before the residual water is reinjected.
• The miscible liquids and dissolved ionic salts that are contained in contaminant plumes can be abstracted from an aquifer, by converting a part of the aquifer into a stationary plume. This results in both the sub-surface dilution of the contaminant and the surface-based ZVI remediation of the contaminant.
• The miscible liquids and dissolved ionic salts that are contained in naturally contaminated aquifers can be abstracted from an aquifer by converting a part of the aquifer to a stationary plume. This results in surface-based ZVI remediation of the contaminant, followed by the reinjection of the decontaminated water in order to produce a diluted contaminant level. Following a prolonged period of operation, the process will produce a decontaminated aquifer within the stationary plume.
• The miscible liquids and dissolved ionic salts that are contained in naturally (and anthropogenically) contaminated soils can be abstracted from a soil by converting a part of the soil to a perched groundwater mound. This results in the surface-based ZVI remediation of the contaminant, followed by the reinjection of the decontaminated water, which produces a diluted contaminant level. Following a prolonged period of operation, this process produces a decontaminated soil within the groundwater mound.

Experiments have established that the effective processing of the contaminated water in surface-based reactors could be operated with a redox reaction; this would involve increasing the pH by 1 to 4 units and decreasing the Eh by 0.1 to 0.5 V over a sustained period. The magnitude of the pH increase and the Eh decrease increases with the decreasing space velocity, decreasing particle size, and increasing particle surface area. These experiments suggest that it may be possible to use a single ZVI charge to process >25,000 m³ t⁻¹ of ZVI.

Hydrodynamic models established the technical viability of creating a hydrodynamic stationary plume. Integrating these models with surface-based ZVI water treatment kinetic data, indicates that the stationary plumes can be used to decontaminate reservoirs areas, zones, or volumes, that contain contaminant plumes or natural contaminants.

Infiltration experiments have established that it is possible to sequester large volumes of water in perched groundwater mounds, and that it is possible to abstract water from these mounds. Integrating this experimental data with surface-based ZVI water treatment kinetic data indicates that these mounds could be used to decontaminate soils, and could act as repositories or reservoirs of decontaminated (or desalinated) water for anthropogenic use.