Desalination of Water Using ZVI (Fe⁰)

Abstract

Batch treatment of water (0.2 to 240 L) using Fe⁰ (44,000–77,000 nm) in a diffusion environment operated (at −8 to 25 °C) using: (a) no external energy; (b) pressurized (<0.1 MPa) air; (c) pressurized (<0.1 MPa) acidic gas (CO₂); (d) pressurized (<0.1 MPa) anoxic gas (N₂); (e) pressurized (<0.1 MPa) anoxic, acidic, reducing gas (H₂ + CO + CO₂ + CH₄ + N₂), reduces the salinity of water. Desalination costs increase with increasing NaCl removal. The cost of reducing water salinity from: (i) 2.65 to 1.55 g·L⁻¹ (over 1–24 h) is $0.002–$0.026 m⁻³; (ii) 38.6 to 0.55 g·L⁻¹ (over 210 days) is $67.6–$187.2 m⁻³. Desalination is accompanied by the removal, from the water, of one or more of: nitrate, chloride, fluoride, sulphate, phosphate, As, B, Ba, Ca, Cd, Co, Cu, Fe, Mg, Mn, Na, Ni, P, S, Si, Sr, Zn. The rate of desalination is enhanced by increasing temperatures and increasing HCO₃⁻/CO₃²⁻ concentrations. The rate of desalination decreases with increasing SO₄²⁻ removal under acidic, or pH neutral, operating conditions.

1. Introduction

Water treatment using zero valent metals (ZVM) in a reactor (or water body) [1,2,3,4,5,6], by aquifer injection/infiltration [7,8,9,10,11,12,13,14], or by placement in a permeable reactive barrier (PRB) located within an aquifer [15,16,17], results in an increase in water pH [5,6,7,18] and a change in water Eh [5,6,7,18]. This is accompanied by a change in water electrical conductivity (EC) [5,6,7,18] which is associated with: (i) the removal of cations and anions from the water by the ZVM (water treatment) [5,6,7,18]; (ii) dissolution of part of the ZVM into the water body [5,19]. Assessment of how the results of this study impact on the removal of microbiota, cations and anions from the water by the ZVM (water treatment) is addressed in Section 6 and Appendix A. ZVM includes Fe⁰, Cu⁰, and Al⁰. In all the trials used in this study, the dominant ZVM is zero valent iron, ZVI (Fe⁰). The abbreviation ZVI is used for Zero Valent Iron (Fe⁰).

Analyses of the salinity of water in aquifers passing through PRB’s and in reactors containing ZVM have established that both Na⁺ and Cl⁻ can be removed from water by ZVM [6,15,20,21,22,23,24,25], or mixtures of ZVM + ion exchange material such as aluminium silicates (e.g., Ca-montmorillonite) [5]. The ZVM is considered to form anodic and cathodic sites in water [18,22,23,24]. The Cl⁻ ions are interpreted as attaching to the anodic sites [22,23]. The iron corrosion rate increases with water velocity [26]. Electrical conductivity (EC) measurements can be used to assess both water salinity and the release of iron corrosion products into the water [20,27].

This study evaluates the partial desalination of water using: (i) 44,000–77,000 nm Fe⁰ powders which are derived from carbon steel; and (ii) a control/reference air stable 50 nm n-Fe⁰ powder (PS7) which is impregnated with polyvinylpyrrolidone (PVP) and coated with tetraethylorthosilicate (TEOS) [28].

Na:Fe:C electrochemical capacitor studies have demonstrated that a high capacitance is associated with: (i) structures where a Fe^III corrosion product is templated onto Fe⁰ [29]; (ii) and structures containing a Fe^III corrosion product and carbon [29].

1.1. Potential Agricultural Application for Partial Desalination by ZVI

Agricultural water usage for irrigation, livestock and cleaning represents about 70% of global water usage [30] (i.e., about 2600 billion·m³·a⁻¹ [31]). This is projected to rise to between 2800 and 3900 billion·m³·a⁻¹ by 2050 [31]. 260–1000 billion·m³·a⁻¹ of global arable irrigation water is adversely affected by salinity [32,33,34,35]. Any reduction in irrigation water salinity, or livestock feed water salinity, can be expected to result in an increase in crop, or livestock, yield (Appendix B, Figure B1 and Figure B2).

A series of background Figures (Figure B1a–p) have been provided (Appendix B) to show the UN FAO (Food and Agricultural Organization) estimates of the relationship between crop yield and water salinity. This is to demonstrate:

(1)

The impact of a decrease in water salinity on the crop yields;

(2)

The impact of a small decrease in water salinity on the number of crops which could be grown on an agricultural holding.

For example, a reduction in water salinity from:

(1)

8 to 4 g·L⁻¹ has the potential (Appendix B, Figure B1a) to increase the possible crop yield (t·ha⁻¹) associated with wheat by 300%.

(2)

4 to 2 g·L⁻¹ has the potential (Appendix B, Figure B1i) to increase the possible crop yield (t·ha⁻¹) associated with potato by >200%.

(3)

3 to 1 g·L⁻¹ has the potential (Appendix B, Figure B1g) to increase the possible crop yield (t·ha⁻¹) associated with soft fruit such as blackberry, raspberry or strawberry by >800%.

These changes in crop yield with reduced salinity indicate that if the salinity of saline irrigation water can be reduced for a reasonable cost using a solution that utilises existing reservoirs, impoundments and tanks, then it may be possible to:

(i)

Significantly reduce the proportion of global agricultural land which is adversely affected by salinity;

(ii)

Increase crop yields on individual agricultural holdings (Appendix B);

(iii)

Increase the range of crops that can be grown commercially on a specific agricultural holding (Appendix B).

Irrigation water, supplied as desalinated water produced by large (>100,000 m³·d⁻¹) reverse osmosis (RO), or multistage flash distillation (MSFD) desalination plants, has a delivery cost of US$0.9–3 m⁻³ [36,37,38].

An agricultural holding may require an irrigation rate of 1000 m³·ha⁻¹·a⁻¹, but may not have access to the finance, or energy, required to operate a suitably sized RO, or MSFD, desalination plant. ZVM desalination has the potential to: (i) reduce the irrigation cost; (ii) utilise existing tanks, ponds and impoundments on the agricultural holding, thereby removing a requirement for new capital investment; (iii) undertake desalination without requiring an energy source (electricity or heat); and (iv) provide an economically viable partial desalination solution for agricultural holdings utilizing 10 to more than 100,000 m³·a⁻¹ of irrigation water.

1.2. Background

1.2.1. Historical ZVI Desalination Experiments

There have been five experimental studies which have evaluated desalination associated with ZVI and ZVM (Fe⁰ + Al⁰ + Cu⁰). They are:

(1)

ZVM (Fe⁰ + Al⁰ + Cu⁰)-Ca-montmorillonite combination and ZVI (Fe⁰)-Ca-montmorillonite. This study [5] demonstrated (Temperature, T = 12–25 °C; Brunauer-Emmett-Teller (BET) surface area, a_s = 0.00289–0.01732 m²·g⁻¹; ZVI (Fe⁰) concentration, P_w = 90 g·L⁻¹) (in an open, unstirred, static flow, batch diffusion reactor operated at ambient temperatures) declines in salinity of 25%–50% over 60 days from an initial salinity of about 1 g·L⁻¹.

(2)

ZVM (Fe⁰ and Fe⁰ + Al⁰ + Cu⁰). This study [6] demonstrated (T = 8–20 °C; a_s = 0.00289–0.01732 m²·g⁻¹; P_w = 90 g·L⁻¹) in an open, unstirred, static flow, batch diffusion reactor operated at ambient temperatures with an initial salinity of about 1 g·L⁻¹, no effective decline in salinity over 60 days.

(3)

ZVI (Fe⁰). This study [20] established (T = 15.3 °C; a_s = 77.26 m²·g⁻¹; P_w = 3.33 g·L⁻¹) in an open, unstirred, static flow, batch diffusion reactor over 48 h a decline in Cl⁻ concentration from 1.52913 to 1.19831 g·L⁻¹. This was associated with an increase in electrical conductivity (EC) from 3.9 to 4.51 mS·cm⁻¹, and an increase in pH from 6.47 to 9.59.

(4)

ZVI (Fe⁰). This study [21] established (T = 20–22 °C; a_s = 77.26 m²·g⁻¹; P_w = 8 g·L⁻¹) in an open, continuously stirred, batch diffusion reactor over a 24 h period a statistical relationship between Cl⁻ concentration in the feed water [C_F] and product water [C_R] over a 24 h period, where C_R, mg·L⁻¹ = 1.1 C_F^0.98 [R² = 0.99] [21]. The absorbed Cl⁻, mg·L⁻¹ (C_L) = C_F − C_R.

(5)

ZVI (n-Fe⁰). This study [25] established over a 3 h period using highly reduced n-Fe⁰, (structured with a Fe⁰ core and an outer carbon coating (P_w = 1.25 g·L⁻¹; 0.1 gNO₃⁻·L⁻¹), in a fluidised, batch diffusion reactor saturated with argon), a linear relationship between Cl⁻ removal [C_L] (g Cl⁻·g⁻¹ Fe⁰) and NaCl concentration g·L⁻¹. C_L = Adsorbed Cl⁻, gCl⁻·g⁻¹ n-Fe⁰; after 3 h C_L = 0.055 g Feed NaCl·L⁻¹. At a salinity of 20 gNaCl·L⁻¹, after 3 h [C_L] = 1.1 gCl⁻·g⁻¹ Fe⁰. This implies a NaCl removal of 1.77 gNaCl·g⁻¹ Fe⁰.

These studies establish five important points:

(1)

The rate of desalination increases with increasing ZVM particle surface area. This indicates that the desalination is either an adsorption process [20,21,25], or a reaction which involves ZVM catalysis on the particle surface [18];

(2)

The rate of desalination increases with increasing fluid flow rates through the Fe⁰ [6,21,25]. This indicates [26] that desalination is associated with a corrosion reaction on the Fe⁰ particle surface;

(3)

The rate of desalination increases with increasing water salinity [21,25]. This indicates that the reaction rate is a function of electrolyte [NaCl] strength.

(4)

The rate of desalination can be significantly increased by combining ZVM with ion exchange material (e.g., aluminium silicates such as Ca-montmorillonite) [5].

(5)

Desalination associated with Fe⁰ can be accompanied by an increase in EC [20]. This indicates that the EC reduction associated with the removal of NaCl [27] has been more than offset by the increase in EC associated with the release of ions (Feⁿ⁺) [19] into the water [27].

1.2.2. ZVI Composition

Most low cost commercial iron powders (<0.1 mm diameter) are constructed from electrolytic iron, or milled iron (44,000–77,000 nm), or iron carbonyl (1000–10,000 nm spheres).

The iron powders will typically contain <0.4 wt % Mn; <0.35 wt % O; <0.03 wt % S; <0.1 wt % Si; <0.04 wt % C; <0.03 wt % P, e.g., [39,40,41,42].

The carbonyl iron will typically contain 0.01–2 wt % C; 0.01–2.5 wt % N, and 0.15–0.5 wt % O [43].

Milled iron produced from carbon steel will have a composition which varies with the grade. For example Q235C/U12358 grade (Chinese Standard GB/T 700-2006 [44]) can contain 0.17 wt % C; 0.35 wt % Si; 1.4 wt % Mn; 0.04 wt % P; 0.04 wt % S; 0.3 wt % Cr; 0.3 wt % Ni; 0.3 wt % Mo; 0.3 wt % Cu, e.g., [45,46].

1.2.3. ZVI Cost

The price of Fe⁰ powders is a function of particle size, quantity purchased, purity, packaging, stabilization, supplier, and prevailing commodity prices. Indicative FOB (free on board) prices on www.alibaba.com for volume purchases (>10 t) were:

(i)

1–100 nm n-Fe0 powders = US$1,000–>$500,000 t⁻¹ (e.g., [47,48,49,50,51]).

(ii)

1000–10,000 nm carbonyl iron powders = US$1,000–$150,000 t⁻¹ (e.g., [43,52,53,54,55].

(iii)

44,000–77,000 nm iron powders = US$150–$3,500 t⁻¹ (e.g., [56,57,58,59,60]).

The price of small quantities (<5 kg) of chemical grade Fe powders varies with particle size and can fall in the range $50 to greater than$10,000/kg for some powders with a particle size of 1–100 nm (e.g., [47,48,61]).

The term FOB is defined, for international trade, by the Incoterms^® 2010 [62,63,64]. The Incoterms^® 2010 define FOB as meaning that the seller delivers the goods on board the vessel nominated by the buyer at the named port of shipment, or to a named destination. The risk of loss, or of damage, to the goods passes to the buyer when the goods arrive on board the vessel, or arrive at the nominated destination. The buyer bears all costs from that moment onwards. The Incoterms® 2010 define a number of other alternative international contract structures [62,63,64], which can be applied to the pricing and sale of ZVM powders.

1.2.4. Potential Cost of Partial Desalination Using ZVI

The cost of partial water desalination is a function of: (i) the cost of the ZVI ($·t⁻¹); (ii) the ZVI loading, P_w, g·L⁻¹ (kg·m⁻³); (iii) length of time required to achieve the required level of desalination; and (iv) the number of times a specific batch of ZVI can be reused.

This study establishes (for a salinity reduction of <4 gNaCl·L⁻¹·d⁻¹), that it is possible to reuse a batch of ZVI at least 18 times. This allows achievement of an effective partial desalination cost for irrigation water of <$0.1 m⁻³.

1.3. Study Structure

This study includes the results from 137 separate desalination trials and 144 control trials. Trial data, detailed methodology descriptions and background information has been placed in the Appendices in order to improve text readability. The Appendices comprise:

(a)

Appendix A: Microbiota, cations, and anions removed by ZVM.

(b)

Appendix B: Impact of changing water salinity of arable crop yields and changing feed water salinity on livestock yields;

(c)

Appendix C: Trial results: feed and product water cations and anion analyses, gas flow rates, Eh, pH, EC, salinity vs. time, UV-visible spectroscopy analyses of entrained particles.

(d)

Appendix D: ZVM compositional and pre-treatment details; Control Reactor Trials;

(e)

Appendix E: Interpretation of salinity from electrical conductivity and UV-visible absorbance data.

(f)

Appendix F: Fe, Al, Cu corrosion in saline water; Relationship between EC (salinity) and reaction kinetics;

(g)

Appendix G: Corrosion species involved in desalination;

(h)

Appendix H: Identification of radicals removed during desalination;

Reference to a specific figure or table in an Appendix is prefixed by the Appendix letter, e.g., Figure C1 is located in Appendix C.

1.3.1. ZVM TP and ZVM TPA

This study examines if it is possible to:

• Treat the iron powders (and ZVM combinations) prior to use to form air stable, attrition resistant porous pellets which can partially desalinate water over a 30–250 day trial period. This treated ZVM is termed ZVM TP in this study; TP = treatment product;
• Use a combination of untreated Fe⁰ and potassium aluminium silicate powder (K-feldspar) to force partial desalination to occur with a 1–24 h period. This combination of untreated Fe⁰ + K-feldspar is termed ZVM TPA in this study; TPA = treatment product, Type A.

The desalination results associated with:

• The ZVM TP are provided in Section 4 and Section 6, Appendix C (

  Table C1. Measured Anion Concentrations in the Feed and Product Water.

  Water	Cl			N(NO3)	S(SO4)	P(PO4)		F	N(NO2)	Product Water (% Feed Water)
  	mg/L			mg/L	mg/L	mg/L		mg/L	mg/L
  Fresh Spring Water
  Base Water	11.67			11.28	4.16	<0.10		0.024	0.04	-
  PS4
  Feed Water	6004.17			11.28	4.16	<0.10		0.024	0.04	-
  Product Water	1434.2			0.12	3.12	<0.10		0.03	2.48	55.00%
  PS5
  Feed Water	4568.84			11.28	4.16	<0.10		0.024	0.04	-
  Product Water	1613.63			0.73	2.32	<0.10		<0.020	3.17	55.00%
  PS11
  Feed Water	2464.2			1.38	1.7	0.229		<0.020	2.49	-
  Product Water	1554.33			1.53	2.78	<0.10		0.046	2.7	92.80%
  PS14
  Feed Water	1914.05			10.81	4.46		<0.10	0.024	3	-
  Product Water	1822.42			0.13	4.84		<0.10	0.024	3.22	78.30%
  ST3b
  Feed Water		6519		8.41	4.5		<0.010	0.023	2.2	-
  Product Water		2060		0.16	1.32		<0.010	0.017	1.31	75.00%
  ST3f
  Feed Water		6519		8.41	4.5		<0.010	0.023	2.2	-
  Product Water		2066		3.69	2.48		<0.010	0.025	2.08	75.00%
  ST6d
  Feed Water			5457.6	11.28	4.16		<0.10	0.024	0.04	-
  Product Water			5269.3	3.33	4.27		<0.10	<0.020	2.34	82.60%
  ST8e
  Feed Water			3983.85	10.04	4.72		<0.10	<0.020	0.29	-
  Product Water			3631.88	0.09	4.74		<0.10	0.034	0.02	94.20%
  MT1b
  Feed Water			4769	11.03	4.86		<0.010	0.015	1.06	-
  Product Water			3937	4.61	4.58		<0.010	0.067	2.21	82.60%
  MT2b
  Feed Water			4037.39	10.5	4.51		<0.10	<0.020	<0.020	-
  Product Water			3700.83	0.53	4.49		<0.10	<0.020	0.08	79.20%
  MT3d
  Feed Water		2092.1		10.81	4.55		0.436	<0.020	2.4	-
  Product Water		1982.02		9.77	7.35		<0.10	0.079	2.43	82.60%
  MT4d
  Feed Water		2464.2		1.38	1.7		0.229	<0.020	2.49	-
  Product Water		1873.3		1.39	5.5		<0.10	0.041	2.42	91.30%
  PS15
  Feed Water		16,243.98		27.55	10.15		<0.10	0.06	0.09	-
  Product Water		10,076.22		3.94	6.04		<0.10	0.13	2.47	66.80%
  PS16
  Feed Water		23,162.04		26.03	9.59		<0.10	0.06	0.09	-
  Product Water		348.03		0.64	1.52		0.15	0.03	0.20	26.50%

  ,

  Table C2. Measured Cation Concentrations in the Feed and Product Water.

  Water		K		Ca	Mg	Na	Al		Fe	Mn
  		mg/L		mg/L	mg/L	mg/L	µg/L		µg/L	µg/L
  Fresh Spring Water
  Base Water		1.69		32.91	10.54	6.32	<150.0		<30.0	1.70
  PS4
  Feed Water		1.69		32.91	10.54	3995.83	<150.0		<30.0	1.70
  Product Water		15.71		10.79	16.45	899.63	<150.0		47.6	2.10
  PS5
  Feed Water		1.69		32.91	10.54	3039.15	<150.0		<30.0	1.70
  Product Water		10.11		20.52	7.58	1009.59	173.4		<30.0	5.50
  PS11
  Feed Water		1.80		9.27	2.13	1522.07	<150.0		<30.0	2.80
  Product Water		2.37		3.64	44.29	1107.61	<150.0		<30.0	2.10
  PS14
  Feed Water		2.67		31.13	9.6	1245.32	<150.0		<30.0	2.80
  Product Water		4.92		6.94	3.58	1197.14	<150.0		<30.0	1.80
  ST3b
  Feed Water		3.49		26.65	7.65	4527	<150.0		<30.0	8.00
  Product Water		5.48		3.26	2.49	1309	<150.0		<30.0	13.00
  ST3f
  Feed Water		3.49		26.65	7.65	4527	<150.0		<30.0	8.00
  Product Water		7.76		17.63	7.38	1293	<150.0		<30.0	5.90
  ST6d
  Feed Water		1.69		32.91	10.54	3630.53	<150.0		<30.0	32.90
  Product Water		5.53		32.48	9.58	3275.18	<150.0		<30.0	69.90
  ST8e
  Feed Water		5.89		31.26	9.5	2390.85	<150.0		34.7	14.20
  Product Water		6.97		14.36	4.5	2175.29	<150.0		<30.0	3.30
  MT1b
  Feed Water	2.30		31.16		9.54	3052		<150.0	42.1	2.90
  Product Water	156.80		6.08		49.23	2521		377	<30.0	2.40
  MT2b
  Feed Water	4.18		30.17		9.09	2434.31		<150.0	<30.0	4.30
  Product Water	23.05		12.27		3.85	2198.69		2085.2	<30.0	2.90
  MT3d
  Feed Water	2.80		30.55		9.42	1370.7		<150.0	<30.0	5.10
  Product Water	7.04		33.65		0.41	1319.14		10,720	<30.0	1.40
  MT4d
  Feed Water	1.80		9.27		2.13	1522.07		<150.0	<30.0	2.80
  Product Water	3.24		2.84		35.75	1273.08		<150.0	<30.0	2.00
  PS15
  Feed Water	4087.42		80.39		25.75	8406.66		<150.0	<30.0	4.15
  Product Water	3266.23		25.51		6.23	4197.07		<150.0	<30.0	23.90
  PS16
  Feed Water	3.89		75.94		24.33	15,411.17		<150.0	<30.0	3.92
  Product Water	5.24		7.46		1.82	223.21		288.8	<30.0	4.00

  ,

  Table C3. Measured Cation Concentrations in the Feed and Product Water.

  Water	P	S	B	Ba	Cd	Co	Cr
  	mg/L	mg/L	µg/L	µg/L	µg/L	µg/L	µg/L
  Fresh Spring Water
  Base Water	<0.005	4.31	29.40	135.60	<0.2	<0.2	<2.0
  PS4
  Feed Water	<0.005	4.31	29.40	135.60	<0.2	<0.2	<2.0
  Product Water	0.01	3.3	12.00	15.10	<0.2	1.30	<2.0
  PS5
  Feed Water	<0.005	4.31	29.40	135.60	<0.2	<0.2	<2.0
  Product Water	0.01	2.55	<10	<15	<0.2	1.50	<2.0
  PS11
  Feed Water	0.25	1.89	<10	16.40	<0.2	1.80	<2.0
  Product Water	0.04	3.15	25.80	<15	<0.2	1.40	<2.0
  PS14
  Feed Water	0.03	4.51	22.20	117.10	<0.2	1.50	<2.0
  Product Water	0.01	4.8	<10	<15	<0.2	1.50	<2.0
  ST3b
  Feed Water	0.07	4.59	14.90	103.10	0.9	3.30	<2.0
  Product Water	0.01	1.56	59.40	<15.0	1.2	1.20	<2.0
  ST3f
  Feed Water	0.07	4.59	14.90	103.10	0.9	3.30	<2.0
  Product Water	0.02	2.63	17.50	<15.0	1.0	1.60	<2.0
  ST6d
  Feed Water	<0.005	4.31	29.40	135.60	<0.2	<0.2	<2.0
  Product Water	0.01	4.31	<10	68.50	<0.2	3.90	<2.0
  ST8e
  Feed Water	0.01	4.8	28.90	118.70	<0.2	2.20	<2.0
  Product Water	0.01	4.9	34.60	<15	<0.2	1.10	<2.0
  MT1b
  Feed Water	0.04	4.79	20.80	102.80	0.4	1.30	<2.0
  Product Water	0.01	4.79	70.40	<15.0	0.4	1.60	<2.0
  MT2b
  Feed Water	0.02	4.66	15.80	115.40	<0.2	2.40	<2.0
  Product Water	0.01	4.68	11.30	<15	<0.2	2.20	<2.0
  MT3d
  Feed Water	0.47	4.66	36.00	118.90	0.3	0.80	<2.0
  Product Water	0.03	7.55	22.30	<15	<0.2	1.80	<2.0
  MT4d
  Feed Water	0.25	1.89	<10	16.40	<0.2	1.80	<2.0
  Product Water	0.02	6.02	97.30	<15	<0.2	2.40	<2.0
  PS15
  Feed Water	<0.005	10.52	71.82	331.20	<0.2	<0.2	<2.0
  Product Water	0.01	6.12	38.10	55.20	10.2	3.60	<2.0
  PS16
  Feed Water	<0.005	9.94	67.80	312.90	<0.2	<0.2	<2.0
  Product Water	<0.005	3.74	109.60	<15	<0.2	0.20	<2.0

  and

  Table C4. Measured Cation Concentrations in the Feed and Product Water.

  Water	Cu	Ni	Pb	Si	Sr	Zn	As
  	µg/L	µg/L	µg/L	mg/L	µg/L	µg/L	µg/L
  Fresh Spring Water
  Base Water	77.7	<3	<10	5.21	144.9	37.4
  PS4
  Feed Water	77.7	<3	<10	5.21	144.9	37.4
  Product Water	<20	<3	<10	0.05	76	<4
  PS5
  Feed Water	77.7	<3	<10	5.21	144.9	37.4
  Product Water	54.3	<3	<10	0.09	54.2	33.8
  PS11
  Feed Water	<20	<3	<10	1.43	28.5	137.5
  Product Water	<20	<3	<10	0.9	4.3	14.1
  PS14
  Feed Water	122.1	<3	<10	4.77	136.1	125.6
  Product Water	<20	<3	<10	0.04	54	<4
  ST3b
  Feed Water	45.1	3.7	<10	3.99	112.4	337.6	2.3
  Product Water	<20.0	<3.0	<10	0.29	15.6	<4.00	5.4
  ST3f
  Feed Water	45.1	3.7	<10	3.99	112.4	337.6	2.3
  Product Water	<20.0	<3.0	<10	0.62	103.4	<4.00	5.3
  ST6d
  Feed Water	77.7	<3	<10	5.21	144.9	37.4	-
  Product Water	736.5	<3	<10	0.41	134.6	68.2	-
  ST8e
  Feed Water	41	<3	<10	4.89	138	286.9	-
  Product Water	<20	<3	<10	0.28	32.5	29	-
  MT1b
  Feed Water	52.9	<3.0	<10	4.92	139.9	170	4.8
  Product Water	<20.0	<3.0	<10	0.37	6.7	17.6	2.5
  MT2b
  Feed Water	93.1	<3	<10	4.67	133.3	101.1	-
  Product Water	<20	<3	<10	0.06	10.2	5.1	-
  MT3d
  Feed Water	74.9	<3	<10	4.83	134.3	342	-
  Product Water	<20	<3	<10	0.12	209.1	<4	-
  MT4d
  Feed Water	<20	<3	<10	1.43	28.5	137.5	-
  Product Water	<20	<3	<10	0.44	0.8	7.7	-
  PS15
  Feed Water	189.80	<3	<10	12.73	353.9	91.36	-
  Product Water	31.80	<3	<10	0.22	86.1	205.10	-
  PS16
  Feed Water	179.30	<3	<10	12.03	334.4	86.30	-
  Product Water	29.00	<3	<10	0.35	30.6	828.50	-

  ; Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16, Figure C17 and Figure C18), Appendix D (Figure D1, Figure D2, Figure D3, Figure D4, Figure D5 and Figure D6), Appendix F.
• The ZVM TPA are provided in Section 5, Appendix C (

  Table C5. ZVM TPA Example (Trial E143): Feed Water Composition.

  Trial	Water Charge	Feed Water	Eh, mV	pH	EC, mS·cm−1
  	L	NaCl, g·L−1
  E143a	5.4	5.74	0.2	5.66	11.15
  E143b	5.4	4.44	0.193	6.61	8.63
  E143c	5.4	4.81	0.199	6.11	9.35
  E143d	5.4	5.37	0.214	5.59	10.43
  E143e	5.4	6.48	0.22	5.78	12.59
  E143f	5.4	1.85	0.166	5.65	3.59
  E143g	5.4	2.59	0.188	6.46	5.04
  E143h	5.4	2.63	0.176	6.46	5.04
  E143i	5.4	2.96	0.185	6.18	5.75
  E143j	5.4	3.33	0.17	5.18	6.63
  E143k	5.4	3.35	0.219	6.19	6.63

  ,

  Table C6. ZVM TPA Example (Trial E143): Product Water Composition.

  Trial	Eh, mV	pH	EC, mS·cm−1	Storage	Stored Water Analysis	pH	EC, mS·cm−1
  				Weeks	Eh, mV
  E143a	0.214	6.80	0.725	6	0.232	6.82	0.814
  E143b	0.226	6.19	3.29	4	0.214	7.00	3.62
  E143c	0.248	5.22	4.10	4	0.227	6.86	4.25
  E143d	0.258	5.31	5.73	4	0.243	6.51	5.80
  E143e	0.224	5.54	9.35	4	0.238	6.58	9.57
  E143f	0.185	6.68	6.71	4	0.220	6.96	7.01
  E143g	0.197	6.76	4.70	4	0.215	7.13	4.75
  E143h	0.201	6.90	6.27	4	0.219	7.05	6.28
  E143i	0.165	5.12	9.51	4	0.261	6.13	9.52
  E143j	0.212	5.31	6.15	4	0.271	6.17	6.19
  E143k	0.207	6.04	6.56	4	0.261	6.34	6.47

  ,

  Table C7. ZVM TPA Example (Trial E143): Product Water Volume Recovered, Trial Duration, Temperature, Gas Flows.

  Trial	Water Recovered	Water Consumed	Duration	Air Flow	CO2 Flow	Air Flow Rate	CO2 Flow Rate	Temperature
  	L	L	h	h	h	L·h−1	L·h−1
  E143a	4.60	0.80	9.00	9.00	0.00	13.2	3.3	5–12 °C
  E143b	4.60	0.80	8.33	5.05	3.28	13.2	3.3	5–10 °C
  E143c	4.70	0.70	4.37	2.56	1.81	13.2	3.3	5–10 °C
  E143d	4.80	0.60	4.02	1.76	2.26	13.2	3.3	5–10 °C
  E143e	4.70	0.70	3.87	1.71	2.16	13.2	3.3	5–10 °C
  E143f	4.65	0.75	2.70	2.70	0.00	13.2	3.3	5–10 °C
  E143g	4.50	0.90	3.08	3.08	0.00	13.2	3.3	5–10 °C
  E143h	3.90	1.50	14.30	14.30	0.00	13.2	3.3	5–10 °C
  E143i	4.00	1.40	8.90	4.50	4.40	13.2	3.3	5–10 °C
  E143j	3.90	1.50	7.50	2.20	5.30	13.2	3.3	5–10 °C
  E143k	3.90	1.50	3.58	2.58	1.00	13.2	3.3	5–10 °C

  ,

  Table C8. ZVM TPA Example (Trial E143): Salinity Changes calculated from changes in EC and changes in water volume.

  Trial	Product Water	Apparent NaCl Removed	Actual NaCl Removed	Cumulative NaCl Removed	Cumulative Water Processed	Cumulative
  	g·L−1	g·L−1	g·L−1	g·L−1	L	NaCl in Feed g·L−1
  E143a	0.064	5.68	5.686	5.69	5.4	5.74
  E143b	1.514	2.93	3.154	8.84	10.8	10.19
  E143c	1.933	2.88	3.132	11.97	16.2	15.00
  E143d	2.786	2.58	2.894	14.87	21.6	20.37
  E143e	4.686	1.80	2.403	17.27	27.0	26.85
  E143f	3.491	−1.64	−1.154	16.12	32.4	28.70
  E143g	2.363	0.23	0.623	16.74	37.8	31.30
  E143h	3.248	−0.62	0.284	17.02	43.2	33.93
  E143i	4.948	−1.98	−0.702	16.32	48.6	36.89
  E143j	3.028	0.31	1.146	17.47	54.0	40.22
  E143k	3.268	0.08	0.992	18.46	59.4	43.57

  ,

  Table C9. ZVM TPA Example (Trial E144): Feed Water Composition.

  Trial	Water Charge, L	Feed Water	Eh, mV	pH	EC, mS·cm−1
  		NaCl, g·L−1
  E144a	5.4	4.07	0.203	6.50	7.91
  E144b	5.4	4.07	0.192	6.35	7.94
  E144c	5.4	1.00	0.181	6.31	2.16
  E144d	5.4	1.10	0.191	6.16	2.34
  E144e	5.4	2.05	0.218	6.37	4.04
  E144f	5.4	5.83	0.192	6.34	11.25

  ,

  Table C10. ZVM TPA Example (Trial E144): Product Water Composition.

  Trial	Eh, mV	pH	EC, mS·cm−1	Storage Weeks	Stored Water Analysis	pH	EC, mS·cm−1
  					Eh, mV
  E144a	0.199	5.64	7.03	3	0.282	6.51	5.8
  E144b	0.222	6.64	7.32	3	0.282	6.57	6.6
  E144c	0.154	5.23	3.92	-	-	-	-
  E144d	0.169	5.38	2.74	-	-	-	-
  E144e	0.153	5.30	4.08	-	-	-	-
  E144f	0.080	5.56	8.58	-	-	-	-

  ,

  Table C11. ZVM TPA Example (Trial E144): Product Water Volume Recovered, Trial Duration, Temperature, Gas Flows.

  Trial	Water Recovered	Water Consumed	Duration	Air Flow	CO2 Flow	Air Flow Rate	CO2 Flow Rate	Temperature
  	L	L	h	h	h	L·h−1	L·h−1	°C
  E144a	4.00	1.40	5.1	3.8	1.3	13.2	3.3	5–12
  E144b	4.00	1.40	104.3	70.5	33.8	13.2	3.3	8–15
  E144c	4.50	0.90	690.0	0.0	690.0	13.2	3.3	11–15
  E144d	4.15	1.25	291.0	0.0	291.0	13.2	3.3	12–16
  E144e	4.00	1.40	537.0	0.0	537.0	13.2	3.3	16–19
  E144f	3.50	1.90	1295.0	0.0	1295.0	13.2	3.3	13–19

  ,

  Table C12. ZVM TPA Example (Trial E144): Salinity changes calculated from changes in EC, spectrometry, and changes in water volume.

  Trial	Product Water Salinity Based on EC	Product Water Salinity Based on Spectrometry	Apparent NaCl Removed	Actual NaCl Removed	Cumulative NaCl Removed	Cumulative Water Processed	Cumulative
  	g·L−1	g·L−1	g·L−1	g·L−1	g·L−1		NaCl in feed g·L−1
  E144a	3.55	-	0.52	1.44	1.44	5.4	4.07
  E144b	3.69	-	0.38	1.34	2.78	10.8	8.15
  E144c	1.90	0.53	0.47	0.56	3.34	16.2	9.15
  E144d	1.27	1.04	0.06	0.30	3.64	21.6	10.25
  E144e	2.03	1.30	0.75	1.09	4.73	27.0	12.30
  E144f	4.35	2.77	3.06	4.04	8.77	32.4	18.13

  and

  Table C13. ZVM TPA Example (Trial E146): Feed Water and Product Water Composition. The NaCl in the feed water is dissolved halite (Cheshire). The product water contains a settled layer of precipitant including clays and Fe hydroxides/peroxides derived from the halite. The apparent increase in salinity in Trial E146l is due to the rate of nano-particle formation exceeding the rate of desalination.

  Trial	Water Charge, L	Feed Water	Eh, mV	pH	EC, mS·cm−1		Product Water	pH	EC, mS·cm−1	NaCl, g·L−1	Desalination
  		NaCl, g·L−1					Eh, mV				NaCl, g·L−1
  E146a	240	1.38	0.188	6.83	2.88		0.152	8.01	4.24	1.04	0.34
  E146b	240	2.73	0.164	6.67	5.4		0.151	8.31	5.35	2.09	0.64
  E146c	240	3.76	0.202	6.92	7.34		0.172	8.29	10.16	0.64	3.12
  E146d	240	6.27	0.156	6.99	12.2		0.178	8.27	18.19	4.68	1.59
  E146e	240	2.92	0.196	6.96	5.76		0.156	8.10	6.51	1.96	0.96
  E146f	240	2.24	0.162	6.90	4.47		0.142	8.19	5.29	1.39	0.85
  E146g	240	2.45	0.164	6.89	4.94		0.145	8.15	5.73	0.71	1.74
  E146h	240	3.73	0.164	6.97	7.39		0.157	8.37	8.62	1.87	1.86
  E146i	240	2.25	0.164	7.08	4.56	0.134		8.57	4.84	0.62	1.63
  E146j	240	2.75	0.176	6.19	5.48	0.155		7.57	5.57	1.85	0.90
  E146k	240	2.07	0.176	6.34	4.25	0.174		7.50	4.59	0.69	1.38
  E146l	240	1.39	0.181	6.24	2.9	0.173		7.65	3.69	2.73	−1.34
  E146m	240	2.17	0.181	6.33	4.38	0.174		7.50	4.59	0.61	1.56
  E146n	240	1.00	0.181	6.31	2.26	0.117		6.86	2.86	0.5	0.50
  E146o	240	1.10	0.191	6.16	2.34	0.134		7.10	3.14	0.76	0.34
  E146p	240	2.05	0.218	6.37	4.04	0.129		7.05	4.77	1.65	0.40
  E146q	240	5.83	0.192	6.35	11.25	0.099		7.40	12.27	2.15	3.68

  ; Figure C19, Figure C20, Figure C21, Figure C22, Figure C23, Figure C24, Figure C25, Figure C26, Figure C27, Figure C28, Figure C29, Figure C30, Figure C31, Figure C32, Figure C33, Figure C34, Figure C35, Figure C36 and Figure C37), Appendix H.

1.3.2. Partial Desalination Information Provided in This Study

• Control Trials and Reference Data (Section 2, Table C1, Table C2, Table C3 and Table C4, Figure D1)

  • Natural spring water used to construct the synthetic saline water used in the trials;
  • Initial Control Data Set: Variation in Eh, pH, EC of untreated ZVM (Fe⁰, Al⁰, Fe⁰ + Al⁰, Fe⁰ + Cu⁰, Fe⁰ + Al⁰ + Cu⁰) in fresh water and saline water;
  • Trial Control Data Set: Variation in Eh, pH, EC of treated particulate ZVM TP (P1) in fresh water (P1c) and saline water (P1);
  • Saline feed water used in the trials, at the trial onset and trail conclusion;
• ZVM TP (Section 4 and Section 6, Table C1, Table C2, Table C3 and Table C4; Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16, Figure C17 and Figure C18,

  Table D1. ZVM TP Manufacturing Processes and Trials; Cu sheathed pellets are 15 mm OD. MDPE sheathed pellets are 20 mm and 25 mm OD. A (g) indicates that the trial group was manufactured in a gaseous environment. All other trial groups were manufactured in an aqueous environment. Trial Group 9 (PS14) was constructed from fine steel wool.

  Trial Group	Manufacturing Process Type	ZVM TP Usage	Trials	Trial Results
  1	B	Particles	PS1–PS4, PS16	Table 1, Table C1, Table C2, Table C3 and Table C4; Figure C1
  2	A	Particles	PS5, PS15	Table 1, Table C1, Table C2, Table C3 and Table C4; Figure C2
  3	C	Particles	PS8	Figure C2
  4	C	Particles	PS9	Figure C2
  5	C	Particles	PS10	Figure C2
  6	D	Particles	PS11	Table 1, Table C1, Table C2, Table C3 and Table C4; Figure C3
  7	D	Particles	PS12	Figure C3
  8	D	Particles	PS13	Figure C3
  9	A (g)	Cu sheathed pellets	PS14	Table 1, Table C1, Table C2, Table C3 and Table C4; Figure C3
  10	D	Container	CSD1a–CSD1d	Table 5
  11	B	Cu sheathed pellets	AS4–AS6, ST1a–ST5j, ST6a–ST6d	Table 1, Table C1, Table C2, Table C3 and Table C4; Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9 and Figure C10
  12	A	Cu sheathed pellets	AS1–AS3	Figure C11
  13	C	Cu sheathed pellets	ST8a	Figure C12
  14	C	MDPE sheathed pellets	ST8b–ST8e	Table 1, Table C1, Table C2, Table C3 and Table C4; Figure C12
  15	C	MDPE sheathed pellets	MT1a–MT1d	Table 1, Table C1, Table C2, Table C3 and Table C4; Figure C13
  16	C	MDPE sheathed pellets	MT2a–MT2d	Table 1, Table C1, Table C2, Table C3 and Table C4; Figure C14
  17	D	MDPE sheathed pellets	MT3a–MT3d	Table 1, Table C1, Table C2, Table C3 and Table C4; Figure C15
  18	D	MDPE sheathed pellets	MT4a–MT4d	Table 1, Table C1, Table C2, Table C3 and Table C4; Figure C16
  19	B	Reused Cu sheathed pellets	ST7a (originally used as ST2d); ST7b (originally used as ST2g); ST7c (originally used as ST3d); ST7d (originally used as ST2f); ST7e (originally used as ST3b); ST7f (originally used as ST3f);	Figure C17

  and

  Table D2. ZVM molar compositional ratios used to manufacture ZVM TP. Data in Table 1 from PS7 refer to 50 nm Fe⁰ (Nanofer Star), a_s = 20 m²·g⁻¹ [28] (Manufacturing Process Type C; 1 Mole n-Fe⁰ + 2.14 Moles Al⁰ + 0.85 Moles Cu⁰). The Nanofer Star was obtained from Nano Iron s.r.o., Stetanikova 116, Rajhrad, 664 61 Czech Republic.

  Trial Group	Fe	Al	Cu	MnO2	ZnO	CaCO3	MgCO3	K2CO3	C	CaO
  1	1	0.3950	0.040	0.006	-	-	-	-	-	-
  2	1	1.4200	0.162	-	-	-	-	-	-	-
  3	1	0.1100	-	-	0.110	-	-	-	-	-
  4	1	0.1580	0.006	-	-	0.0900	0.060	0.017	0.16	-
  5	1	0.5200	0.019	-	-	-	-	-	-	-
  6	1	-	-	-	0.056	0.2086	0.152	-	-	-
  7	1	0.8400	0.090	-	0.068	0.0640	0.090	-	-	-
  8	1	0.9700	0.040	-	-	0.0530	0.090	-	-	0.07
  9	1	-	-	-	-	-	-	-	-	-
  10	1	0.3200	-	-	-	-	-	-	-	-
  11,19	1	0.3950	0.040	0.006	-	-	-	-	-	-
  12	1	1.4200	0.162	-	-	-	-	-	-	-
  13	1	0.1100	-	-	0.110	-	-	-	-	-
  14	1	0.1100	-	-	0.110	-	-	-	-	-
  15	1	0.1580	0.006	-	-	0.0900	0.060	0.017	0.16	-
  16	1	0.5200	0.019	-	-	-	-	-	-	-
  17	1	0.8100	0.080	0.100	0.070	0.1500	0.100	-	0.06	-
  18	1	0.8200	0.040	-	-	0.1100	0.080	-	0.14	-

  , Figure D1, Figure D2, Figure D3, Figure D4, Figure D5 and Figure D6)

  • Impact of ZVM pre-treatment on desalination rates;
  • Variation in EC (salinity) with time;
  • Variation in Eh and pH with time;
  • Relationship between desalination and ZVM TP loading (g·L⁻¹);
  • Relationship between desalination, particle size and capacitance;
  • Economics of partial desalination using ZVM TP.
• ZVM TPA (Section 5, Table C5, Table C6, Table C7, Table C8, Table C9, Table C10, Table C11, Table C12 and Table C13, Figure C19, Figure C20, Figure C21, Figure C22, Figure C23, Figure C24, Figure C25, Figure C26, Figure C27, Figure C28, Figure C29, Figure C30, Figure C31, Figure C32, Figure C33, Figure C34, Figure C35, Figure C36 and Figure C37,

  Table D3. ZVM molar compositional ratios used to manufacture ZVM TPA. K-Feldspar (11.3% K₂O + 3.2% NaO + 18.5% Al₂O + 65.8% SiO₂); Manufacturing Process Type E = dry mixing in air. Al = 5 mm punchings.

  Trial Group	Manufacturing Process Type	ZVM TP Usage	Trials	Trial Results	Cartridge Size, L	Fe, g	K-Feldspar, g	Al, g
  20	E	Cartridge	E143a–E143k	Table C5, Table C6, Table C7 and Table C8	0.4	135	97	-
  21	E	Cartridge	E144a–E144f	Table C9, Table C10, Table C11 and Table C12	0.4	51	90	38
  22	E	Cartridge	E145	Figure C18	2.0	449	1000	-
  23	E	Cartridge	E146a–E146p	Table C13, Figure C19, Figure C20, Figure C21, Figure C22, Figure C23, Figure C24, Figure C25, Figure C26, Figure C27, Figure C28, Figure C29, Figure C30, Figure C31, Figure C32, Figure C33, Figure C34, Figure C35, Figure C36 and Figure C37	0.4	124	276	-

  )

  • Relationship between desalination rate, ZVM TPA reuse, and time;
  • Economics of partial desalination.
• Mechanism of desalination (Section 4 and Section 5, Appendix F, Appendix G, Appendix H)

  • Removal of NaCl by incorporation into Fe corrosion products;
  • Removal of NaCl by inclusion in a hydration shell;
• Implications for the co-removal of cations and anions contained in the water (Section 6, Table C1, Table C2, Table C3 and Table C4, Figure C18)

  • Water containing other cations, anions, and microbiota;
  • Water enriched in sulphates;
  • Water enriched in carbonates.

2. Materials, Methods and Equipment

Tabulated and graphical experimental/trial data used in this study are provided in Appendix C (Table C1, Table C2, Table C3, Table C4, Table C5, Table C6, Table C7, Table C8, Table C9, Table C10, Table C11, Table C12 and Table C13; Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16, Figure C17, Figure C18, Figure C19, Figure C20, Figure C21, Figure C22, Figure C23, Figure C24, Figure C25, Figure C26, Figure C27, Figure C28, Figure C29, Figure C30, Figure C31, Figure C32, Figure C33, Figure C34, Figure C35, Figure C36 and Figure C37). All particles used to construct the ZVM TP and ZVM TPA were powdered (44,000–77,000 nm particle size). Details of the ZVM TP and ZVM TPA compositions are provided in Appendix D (Table D1, Table D2 and Table D3).

2.1. Feed Water

The natural untreated spring water used in the trials was extracted from a well located in fractured andesites and acid (rhyolitic) pyroclastics (Devonian, Old Red Sandstone Volcanic Series, Ochil Hills, Scotland). This water has previously been used in ZVM studies [5,6]. The variation in EC, Eh, and pH with time of this water source is documented elsewhere [6]. The cation and anion composition of the natural spring water water is provided in Table C1, Table C2, Table C3 and Table C4. The term freshwater in this study is used to refer to this natural spring water.

2.2. Equipment

Eh, pH, EC, and temperature were measured using equipment manufactured or branded by Hanna Instruments Ltd. (Leighton Buzzard, Bedfordshire LU7 4AD, UK), Extech Instruments (Nashua, NH, USA), HM-Digital Inc. (Culver City, CA, USA) and Oakton Instruments (Vernon Hills, IL, USA). EC, Eh and pH calibration standards used were manufactured or branded by Hanna Instruments, HM-Digital, Milwaukee Instruments Inc. (Rocky Mount, NC, USA). Eh is calibrated to the standard hydrogen electrode (SHE) (Appendix D2.1). Current, voltage, resistance and capacitance of the ZVM TP were measured using equipment manufactured or branded by Philex Electronic (UK) Ltd. (Bedford, UK). Cation and anion analyses were contracted to Forest Research (Farnham, Surrey, UK), the commercial laboratories of the UK Forestry Commission. Anions were determined using Dionex Ion Chromatography (Thermo Fischer Scientific Inc. (Waltham, MA, USA)). Cations were determined using a Thermo Icap 6500 Spectrometer.

Real time salinity changes were determined using EC (Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16 and Figure C17 and Figure D1, Figure D2, Figure D3, Figure D4, Figure D5 and Figure D6) and UV-visible spectroscopy (Figure C19, Figure C20, Figure C21, Figure C22, Figure C23, Figure C24, Figure C25, Figure C26, Figure C27, Figure C28, Figure C29, Figure C30, Figure C31, Figure C32, Figure C33, Figure C34, Figure C35 and Figure C36) analysed using a YXUV-5200 spectrometer supplied by Shanghai Selon Scientific Instruments, Shanghai University National Science Park, Shanghai, China. The gas compositions (CH₄, CO, CO₂, N₂, H₂) entering and leaving the water from the 3.5 L, 5.4 L and 8 L capacity reactors were monitored using a SRI 8610C TCD GC (manufactured by SRI Instruments Inc. (Torrance, CA, USA)) with a silica gel column and He carrier gas. GC = gas chromatograph, TCD = Thermal Conductivity Detector). Gases and gas calibration standards used in this study were purchased from BOC/Linde, (Guildford, Surrey, UK).

2.3. Control and Reference Trials

2.3.1. Reference Trials: Feedwater Analysed at the Start and End of Each Trial

The trials used saline water which was constructed by either dissolving NaCl into natural spring water (control trials and ZVM TP trials), or by dissolving halite into natural spring water (ZVM TPA trials). Cation and anion compositions are provided in Table C1, Table C2, Table C3 and Table C4 for the saline feed water constructed using NaCl. The amount (wt) of NaCl/halite dissolved into the water was recorded.

The EC, Eh, pH and temperature of the saline water was determined prior to each trail. Some of the saline feed water used in each trial was stored as a control. At the conclusion of the trial the EC, pH and temperature of the saline feed water was redetermined. In each instance the redetermined pH reading was within 0.1 units of the original reading and the redetermined EC reading was within 2% of the original reading. The redetermined results are within the margins of error of the analytical tools used. The reference trials establish that any differences in pH and EC between the original feed water and the product water containing ZVM, or ZVM TP, or ZVM TPA, arise from the presence of the ZVM, ZVM TP and ZVM TPA.

2.3.2. Initial Control Trials Using Untreated ZVM

A series of batch reactors were constructed where each reactor contained ZVM powders (Fe⁰, Al⁰, Fe⁰ + Al⁰, Fe⁰ + Cu⁰ and Fe⁰ + Al⁰ + Cu⁰) + water. Each reactor was not sealed, was open to the air and had an air-water contact. The water was not stirred and the principal interaction between the water and ZVM was by diffusion. These control trials are documented in [6]. Two of the trials (containing Fe⁰ and Fe⁰ + Al⁰ + Cu⁰) were undertaken in both saline water and freshwater [6] to identify the differences which arise with time in Eh, pH and EC. The EC declines recorded over 60 days in the freshwater and saline water were similar [6]. Therefore it is reasonable to conclude that no noticeable desalination had occurred during the trials with untreated ZVM.

In order to allow comparison with the earlier trials documented in references [5,6], the untreated Fe⁰, Al⁰, and Cu⁰ powders used (and illustrated [5]) in references [5,6] were used to construct the ZVM TP (containing treated ZVM) and ZVM TPA (containing a mixture of untreated ZVM and ion exchange material). The ZVM TPA operating procedure effectively treats the ZVM during desalination.

2.3.3. Control Trial Using ZVM TP

A control trial was established where two trials were run simultaneously (PS1 (Figure C1) and PS1C (Figure D1)) under the same temperature and pressure conditions. In each trial the ZVM TP powder was placed in an unsealed, unstirred, reactor containing a batch of water, and an air-water contact. Reactor PS1 contained saline water. Reactor PS1C contained fresh water. Both Reactors displayed a rapid increase in pH with time (Figure D1), and an initial decrease in Eh followed by a rise in Eh (Figure D1). PS1 displayed a general decrease in EC with time due to desalination (Figure D1), while PS1C displayed a general increase in EC with time due to the release of Feⁿ⁺ (aq) ions (Figure D1). Water consumption in PS1C was higher than in PS1. 35% of the feed water was consumed in PS1C within 42 days.

Cross plots of Eh vs. pH, Eh vs. EC, and pH vs. EC (Figure D1) indicate that PS1C had a consistently higher pH than PS1 and operated with an Eh over a wider range. The pH in PS1 reduced as the salinity reduced towards the pH of the initial feed water. The pH in PS1C remained at elevated levels. The bulk of the salinity reduction in PS1 occurred at Eh > 0 mV. The Eh in PS1 reduced below 0 mV when the EC fell below 3 mS·cm⁻¹ (Figure D1).

The performance (Eh, pH, EC) of PS1C was similar to that observed for untreated ZVM [6].

These control trials demonstrated that pre-treatment of the ZVM (when compared with the performance of untreated ZVM [6]) can result in a major decrease in EC (salinity) when the treated ZVM TP is placed in saline water (Figure C1 and Figure D1).

2.4. ZVM TP Trials

2.4.1. Temperatures and Pressures Used in the ZVM TP Trials and ZVM TP Control Trial

The reactors were all placed in an external environment. The temperatures and pressures used in the ZVM TP trials (Appendix C) and ZVM TP control trials (Figure D1) are provided in Figure 1.

Figure 1. Diabatic operating conditions. (a) Temperature; (b) Air Pressure. Data Source: Strathallan weather station (www.weatheronline.com). All trials are referenced to Day 1, where Day 1 is 17 July 2012.

Figure 1. Diabatic operating conditions. (a) Temperature; (b) Air Pressure. Data Source: Strathallan weather station (www.weatheronline.com). All trials are referenced to Day 1, where Day 1 is 17 July 2012.

2.4.2. Manufacture of ZVM TP

ZVM TP particles and pellets were manufactured by placing the raw ZVM in saline water which was saturated with either:

• Type A: [16.79% CH₄ + 16.88% H₂ + 11.97% CO + 8.33% CO₂ + 46.03% N₂], Trials: AS1–AS3, PS5; Manufacturing time = 17 days; Manufacturing time for Trial PS15 = 35 Days;
• Type B: [16.79% CH₄ + 16.88% H₂ + 11.97% CO + 8.33% CO₂ + 46.03% N₂ for 110 days] followed by [80% N₂ + 20% CO₂ for 135 days] followed by [100% N₂ for 46 days]; Trials: ST1–ST7, AS4–AS6, PS1–PS4, PS1C; Manufacturing time = 291 days; Manufacturing time for Trial PS16 = 65 Days;
• Type C: [N₂] Trials: PS8–PS10, ST8, MT1, MT2; Manufacturing time = 34 days;
• Type D: [air] Trials: PS11, PS12, PS13, MT3, MT4; Manufacturing time = 42 days.

During the manufacturing process the ZVM was maintained at a temperature which varied within the range −8 to 25 °C, and at a pressure which was between atmospheric pressure and 0.1 MPa.

2.4.3. ZVM Composition Used to Manufacture ZVM TP (Table D1 and Table D2)

The molar ZVM compositions entered into the manufacturing reactor to construct the ZVM TP are referenced to Fe⁰. The recovered ZVM TP was used as particulate matter (Figure 2) or was placed in the water as either:

(i)

Copper sheathed pellets (Figure 2),

(ii)

MDPE sheathed pellets,

(iii)

A cartridge.

Figure 2. Macroporous ZVM TP (a) End view of a partially oxidized Cu sheathed ST1a–ST5j series ZVM TP pellet (15 mm OD; OD = Outer diameter). Porosity produced by degassing during manufacture; (b) Particulate ZVM TP (PS1–PS4, PS1C, PS15, PS16), field of view = 2 cm.

Figure 2. Macroporous ZVM TP (a) End view of a partially oxidized Cu sheathed ST1a–ST5j series ZVM TP pellet (15 mm OD; OD = Outer diameter). Porosity produced by degassing during manufacture; (b) Particulate ZVM TP (PS1–PS4, PS1C, PS15, PS16), field of view = 2 cm.

2.4.4. Reactors: ZVM TP Tests at Ambient Temperatures

The reactors used for testing the ZVM TP pellets and particles were containers with capacities of 0.3, 2.3 and 10 L. Each container had an air-water contact and was operated at atmospheric pressure and temperature (Figure 1). The ZVM TP was placed in the reactors and settled at their base. The water in the reactors was not stirred or agitated during the desalination. The upper surface of the reactor was not sealed. This allowed fresh air to continually interact with the water body.

2.4.5. Reactors: ZVM TP Tests Using Pressured Reactors where a Gas Is Used to Maintain the Pressure

Sealed reactors (3.5 and 8 L) containing particulate ZVM TP held in a cartridge, were used for ZVM TP trials CSD1 (8 L reactor), PS15, and PS16. Trials PS15 and PS16 used a reactor with a capacity of 3.5 L. Both reactors had a 1.5 m water column above the gas distributor with >0.5 m gas located above the gas-water contact. The ZVM TP cartridges were located below the gas distributors.

The 8.0 L reactor contained a heat exchanger which allowed the water temperature to be increased into the range 30 to 70 °C.

2.5. ZVM TPA Trials

2.5.1. Cartridge and Particle Manufacture: ZVM TPA (Table D3)

The ZVM TPA was constructed by mixing untreated ZVM with ion exchange material (Figure 3) prior to placement in cartridges. This series of tests determined if this ZVM combination could partially desalinate water over a 1 h to 24 h period, and whether a specific batch of ZVM TP could be used multiple times without regeneration.

Figure 3. ZVM TPA powder used in the initial commercial trials (E145, E146, Table C5, Table C6, Table C7, Table C8, Table C9, Table C10, Table C11, Table C12 and Table D3), prior to placement in the cartridges. Field of view = 2 cm.

Figure 3. ZVM TPA powder used in the initial commercial trials (E145, E146, Table C5, Table C6, Table C7, Table C8, Table C9, Table C10, Table C11, Table C12 and Table D3), prior to placement in the cartridges. Field of view = 2 cm.

2.5.2. Reactors: Used for Testing ZVM TPA

The reactors used for testing the ZVM TPA cartridges had capacities of 5.4, 8.0, 114 and 240 L. Each reactor contained a ZVM TPA cartridge and a gas (80% N₂ + 20% CO₂, or air) was bubbled through the reactor during operation. The 5.4 and 8.0 L reactors were structured to allow no interaction between air and the gas-water contact in the reactor. The 114 and 240 L reactors were not sealed. They were specifically structured to allow some interaction between atmospheric air and the water body.

3. Interpretation of Salinity

The methodology used to interpret the real time salinity in the trials is detailed in Appendix E, (Figure E1, Figure E2, Figure E3 and Figure E4,

Table E1. Equations used to calculate salinity. Calibrated using Cheshire Halite. Valid Range = 0 to 22 gNaCl·L⁻¹. Applicability: two component system only, water + dissolved halite. x_raw = raw absorbance value at the measured wavelength of the partially desalinated water. x_initial = the raw absorbance value at the measured wavelength of the feed water. x = the normalized absorbance value at the measured wavelength of the partially desalinated water. This is calculated as x = [x_raw (at a specific wavelength)] − [x_raw (at a wavelength of 900 nm)]. R = the initial salinity of the feed water, g·L⁻¹. Absorbance is referenced to the freshwater (0 g NaCL L^-1) control sample.

Wavelength, nm	Normalized Absorbance	Raw Absorbance
	Salinity, g·L−1	Salinity, g·L−1
200	Salinity, g·L−1 = 766.28x2 − 3.2388x + 1.8149	Salinity, g·L−1 = [xraw/xinitial]·R
205	Salinity, g·L−1 = 636.34x2 − 10.996x + 1.8025	Salinity, g·L−1 = [xraw/xinitial]·R
210	Salinity, g·L−1 = 533.61x2 + 6.1084x + 1.8424	Salinity, g·L−1 = [xraw/xinitial]·R
215	Salinity, g·L−1 = 784.61x2 + 3.7002x + 1.8436	Salinity, g·L−1 = [xraw/xinitial]·R
220	Salinity, g·L−1 = 866.19x2 + 32.111x + 1.8318	Salinity, g·L−1 = [xraw/xinitial]·R
225	Salinity, g·L−1 = 450.66x2 + 33.898x + 1.5123	Salinity, g·L−1 = [xraw/xinitial]·R
230	Salinity, g·L−1 = 118.7x2 + 27.487x + 1.3347	Salinity, g·L−1 = [xraw/xinitial]·R
235	Salinity, g·L−1 = 7.8675x2 + 52.237x + 0.9155	Salinity, g·L−1 = [xraw/xinitial]·R
240	Salinity, g·L−1 = 11.138x2 + 49.401x + 0.957	Salinity, g·L−1 = [xraw/xinitial]·R
245	Salinity, g·L−1 = 38.641x2 + 42.294x + 0.948	Salinity, g·L−1 = [xraw/xinitial]·R
250	Salinity, g·L−1 = 18.317x2 + 47.257x + 1.0025	Salinity, g·L−1 = [xraw/xinitial]·R
255	Salinity, g·L−1 = 37.841x2 + 43.73x + 1.0558	Salinity, g·L−1 = [xraw/xinitial]·R
260	Salinity, g·L−1 = 10.321x2 + 52.13x + 1.0098	Salinity, g·L−1 = [xraw/xinitial]·R
265	Salinity, g·L−1 = 20.714x2 + 49.407x + 0.9941	Salinity, g·L−1 = [xraw/xinitial]·R
270	Salinity, g·L−1 = 27.368x2 + 48.753x + 1.0832	Salinity, g·L−1 = [xraw/xinitial]·R
275	Salinity, g·L−1 = 27.115x2 + 47.752x + 1.1226	Salinity, g·L−1 = [xraw/xinitial]·R
280	Salinity, g·L−1 = 32.059x2 + 52.604x + 0.9819	Salinity, g·L−1 = [xraw/xinitial]·R
285	Salinity, g·L−1 = 27.529x2 + 55.578x + 0.9332	Salinity, g·L−1 = [xraw/xinitial]·R
290	Salinity, g·L−1 = 19.31x2 + 59.508x + 0.8993	Salinity, g·L−1 = [xraw/xinitial]·R
295	Salinity, g·L−1 = 29.723x2 + 62.585x + 0.8414	Salinity, g·L−1 = [xraw/xinitial]·R
300	Salinity, g·L−1 = 17.616x2 + 65.145x + 0.8971	Salinity, g·L−1 = [xraw/xinitial]·R
350	Salinity, g·L−1 = 56.721x2 + 83.129x + 0.9896	Salinity, g·L−1 = [xraw/xinitial]·R
400	Salinity, g·L−1 = 116.87x2 + 77.832x + 1.0311	Salinity, g·L−1 = [xraw/xinitial]·R
500	Salinity, g·L−1 = 302.83x2 + 101.92x + 1.2602	Salinity, g·L−1 = [xraw/xinitial]·R
600	Salinity, g·L−1 = 1385x2 + 158.9x + 1.3455	Salinity, g·L−1 = [xraw/xinitial]·R
700	Salinity, g·L−1 = 7993.4x2 + 187.23x + 1.6814	Salinity, g·L−1 = [xraw/xinitial]·R
800	Salinity, g·L−1 = 22,332x2 + 100.29x + 1.6396	Salinity, g·L−1 = [xraw/xinitial]·R
900	-	Salinity, g·L−1 = [xraw/xinitial]·R

).Two methods are used. They are determination of salinity using EC and determination of salinity using absorbance determined by UV-visible spectroscopy (Figure E1). The basic equations used to calculate salinity are:

(a) Salinity at time t = n calculated using EC:

EC_t=n = [A] + [B] + [C] − [D] − [E]

(1)

where [A] = EC due to salinity at time t = 0; [B] = EC due to other components in the water at time t = 0; [C] = EC added to the water by the ZVM, ZVM TP, ZVM TPA between time, t = 0 and time, t = n; [D] = EC due to salinity which has been removed from the water between time, t = 0 and time, t = n; [E] = EC due to other components in the feed water which has been removed from the water between time, t = 0 and time, t = n.

EC is related to salinity (Appendix E, Figure E3) using a regression equation of the form:

Salinity = a[EC] + b

(2)

a and b are constants where b is a function of [B]. a is feed water specific and may vary with temperature.

(b) Salinity at time t = n calculated using Absorbance at wavelength x nm

Absorbance_t=n = [A]_a + [B]_a+ [C]_a − [D]_a − [E]_a

(3)

where [A]_a = Absorbance due to salinity at time t = 0; [B]_a = Absorbance due to other components in the water at time t = 0; [C]_a = Absorbance added to the water by the ZVM, ZVM TP, ZVM TPA between time, t = 0 and time, t = n; [D]_a = Absorbance due to salinity which has been removed from the water between time, t = 0 and time, t = n; [E]_a = Absorbance due to other components in the feed water which has been removed from the water between time, t = 0 and time, t = n;

Absorbance is related to salinity (Appendix E, Table E1) using a regression equation of the form:

Salinity = a[absorbance] + b, or Salinity = a[absorbance]ⁿ…… e[absorbance] + f

(4)

where a, e and f are constants, n is a polynomial number. The regression equation (and applicable equation type, linear, polynomial, etc.) which is applicable will vary with water composition. Different relationships will apply at different wavelengths. Measurement accuracy is enhanced by calculating salinity at multiple wavelengths [65] and then using the average calculated salinity as the salinity of the water [65]. In this study, each salinity value calculated from absorbance is the average of the salinity indicated by 27 different wavelengths (Table E1, Figure E1 and Figure E2). The absorbance attributable to ([A]_a − [D]_a) is 0 when the salinity is zero.

Absorbance due to nano particles in the water was analysed (e.g. Figure E4) in accordance with the UV-Visible spectroscopy standard best practice methodology [66].

4. Results: ZVM TP

Placement of ZVM TP in a static water body at ambient temperatures, utilising no external energy, established a general decline in salinity with time (Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16 and Figure C17). The process produces product water and used ZVM TP. The magnitude of the salinity declines (documented in Table C1, Table C2, Table C3 and Table C4, Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16 and Figure C17 and Figure D1, Figure D2, Figure D3, Figure D4, Figure D5 and Figure D6) indicates that ZVM TP pellets or powders have a potential application in the desalination of irrigation water (Appendix B). Salinity declines of up to 8 gNaCl·L⁻¹ are recorded in Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16 and Figure C17. Example analyses are summarised in Table 1.

The economic effectiveness of ZVM TP as a potential desalination agent for irrigation water is a function of:

(1)

The weight of NaCl removed/unit weight ZVM TP (Q_e);

(2)

The ZVM TP loading in the water body (weight of ZVM TP/unit volume of water, P_w);

(3)

The time, t, taken to achieve the required level of desalination;

(4)

The amount of water consumed during desalination;

(5)

The number of times the ZVM TP can be reused;

(6)

The residual value of the ZVM TP;

(7)

The cost of the ZVM TP.

Q_e and P_w and the inter-relationship between Q_e, P_w, pH, Eh, surface charge, and ZVM TP capacitance are used to assess the economic effectiveness of different pre-treatment methods. The relationship between surface charge/capacitance and EC decline rates is used to provide a measure of quality control for the ZVM TP (i.e., Q_e, P_w and the length of time required to achieve a targeted level of desalination).

The significance of Eh, pH, EC, P_w, Q_e, temperature, pre-treatment, EC decline rates, and surface charge/capacitance, is discussed here in the context of established Fe corrosion theory in order to demonstrate the mechanism which ZVM TP uses to facilitate desalination and to facilitate the manufacture of more effective ZVM TP desalination pellets and particles.

Table 1. Changes in water salinity associated with ZVM TP. Gross Q_e is calculated from the gross salinity reduction after consideration of water reduction. EC and Salinity Data: Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16 and Figure C17; Table C1, Table C2, Table C3 and Table C4. Colour coding reflects the manufacturing process, which is defined in Table D1: Yellow = Type A; Blue = Type B; Tan = Type C; Green = Type D.

Trial	Feed Water	Feed Water Na + K + Cl	Product Water	Product Water Na + K + Cl	Number of Days	Reduction in Water Volume	Salinity Reduction	EC Reduction	Qe	Gross Salinity Reduction	Gross Qe	ZVM TP
	EC mS·cm−1	mg·L−1	mS·cm−1	mg·L−1		%	mg·L−1	mS·cm−1	mg g−1	mg·L−1	mg g−1	g·L−1
MT1b	14.65	7,823.30	12.53	6,614.80	100	17%	1,208.50	2.12	28.1	2,359.48	54.87	43
MT2b	12.66	6,475.88	10.93	5,922.57	98	21%	553.31	1.73	4.56	1,785.20	14.72	121.3
MT3d	7.23	3,465.60	6.46	3,308.2	98	17%	157.40	0.77	2.5	733.03	11.65	62.9
MT4d	7.84	3988.07	6.57	3,149.62	57	9%	838.45	1.27	9.31	1,112.47	12.35	90.1
ST3b	18.84	11,049.49	6.73	3,374.48	126	25%	7,675.01	12.11	307	8,518.63	340.75	25
ST3f	18.84	11,049.49	6.66	3,052.11	126	25%	7,997.38	12.18	139.08	8,760.41	152.35	57.5
ST6d	16.59	9,089.82	15.74	8,550.01	120	17%	539.81	0.85	105.85	2,027.51	397.55	5.1
ST8e	12.35	6,380.24	10.95	5,814.14	98	6%	566.10	1.4	15.51	903.32	24.75	36.5
PS4	19.71	10,001.69	5.05	2,195.88	150	45%	7,805.81	14.66	141.92	8,793.96	159.89	55
PS5	15.18	7,609.68	5.32	2,633.33	150	45%	4,976.35	9.86	165.88	6,161.35	205.38	30
PS7	14.05	7,391.68	4.7	1,941.37	198	45%	5450.30	9.35	2180.12	6,323.93	2529.57	2.5
PS11	7.84	3,988.07	5.43	2,664.32	57	7%	1,323.75	2.41	49.95	1,515.58	57.19	26.5
PS14	6.41	3,162.04	5.66	3,024.48	57	22%	137.56	0.75	13.49	793.87	77.83	10.2
PS15	57.49	23,707.28	31.12	17,284.81	79	33%	6,422.47	26.37	47.93	12,212.88	91.14	134
PS16	78.18	38,577.1	1.09	576.48	210	74%	38,000.62	77.09	182.7	38,424.33	184.73	208

Table 1. Changes in water salinity associated with ZVM TP. Gross Q_e is calculated from the gross salinity reduction after consideration of water reduction. EC and Salinity Data: Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16 and Figure C17; Table C1, Table C2, Table C3 and Table C4. Colour coding reflects the manufacturing process, which is defined in Table D1: Yellow = Type A; Blue = Type B; Tan = Type C; Green = Type D.

Trial	Feed Water	Feed Water Na + K + Cl	Product Water	Product Water Na + K + Cl	Number of Days	Reduction in Water Volume	Salinity Reduction	EC Reduction	Qe	Gross Salinity Reduction	Gross Qe	ZVM TP
	EC mS·cm−1	mg·L−1	mS·cm−1	mg·L−1		%	mg·L−1	mS·cm−1	mg g−1	mg·L−1	mg g−1	g·L−1
MT1b	14.65	7,823.30	12.53	6,614.80	100	17%	1,208.50	2.12	28.1	2,359.48	54.87	43
MT2b	12.66	6,475.88	10.93	5,922.57	98	21%	553.31	1.73	4.56	1,785.20	14.72	121.3
MT3d	7.23	3,465.60	6.46	3,308.2	98	17%	157.40	0.77	2.5	733.03	11.65	62.9
MT4d	7.84	3988.07	6.57	3,149.62	57	9%	838.45	1.27	9.31	1,112.47	12.35	90.1
ST3b	18.84	11,049.49	6.73	3,374.48	126	25%	7,675.01	12.11	307	8,518.63	340.75	25
ST3f	18.84	11,049.49	6.66	3,052.11	126	25%	7,997.38	12.18	139.08	8,760.41	152.35	57.5
ST6d	16.59	9,089.82	15.74	8,550.01	120	17%	539.81	0.85	105.85	2,027.51	397.55	5.1
ST8e	12.35	6,380.24	10.95	5,814.14	98	6%	566.10	1.4	15.51	903.32	24.75	36.5
PS4	19.71	10,001.69	5.05	2,195.88	150	45%	7,805.81	14.66	141.92	8,793.96	159.89	55
PS5	15.18	7,609.68	5.32	2,633.33	150	45%	4,976.35	9.86	165.88	6,161.35	205.38	30
PS7	14.05	7,391.68	4.7	1,941.37	198	45%	5450.30	9.35	2180.12	6,323.93	2529.57	2.5
PS11	7.84	3,988.07	5.43	2,664.32	57	7%	1,323.75	2.41	49.95	1,515.58	57.19	26.5
PS14	6.41	3,162.04	5.66	3,024.48	57	22%	137.56	0.75	13.49	793.87	77.83	10.2
PS15	57.49	23,707.28	31.12	17,284.81	79	33%	6,422.47	26.37	47.93	12,212.88	91.14	134
PS16	78.18	38,577.1	1.09	576.48	210	74%	38,000.62	77.09	182.7	38,424.33	184.73	208

The NaCl is removed by the ZVM TP and is not present as a separate precipitate within the water. This demonstrates that the NaCl interacts with, and is bound into, a reaction product associated with the corrosion of ZVM TP in water.

The NaCl is removed by one or more of:

(1)

Incorporation into a Fe corrosion product. Green Rust One (Cl⁻), GR1, incorporates Cl⁻ ions in its crystal structure. β-FeOOH requires the presence of Cl⁻ ions to form. All other Fe corrosion products do not incorporate Cl⁻ ions within their crystal structure. In this instance Q_e < 70 mg NaCl removed g⁻¹ ZVM [67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100].

(2)

Incorporation into the hydration shell of a Fe corrosion product (e.g., green rust, FeOOH) [96]. A hydrated amorphous FeOOH corrosion product containing 46 wt % H₂O will contain 5 hydration shells. Radicals containing Na⁺ and Cl⁻ can substitute for H₂O in the hydration lattice [96]. This may potentially allow Q_e to exceed 3 gNaCl removed g⁻¹ ZVM [96].

(3)

Concentration of NaCl in the pore waters within the ZVM TP/ZVM TPA corrosion product bed, when the ZVM interacts with the overlying water body by diffusion. This may potentially allow Q_e to exceed 20 gNaCl·removed·g⁻¹ ZVM. This method of NaCl removal is investigated in the ZVM TP trials CSD1 and ZVM TPA trials E143, and E144 (Section 5).

The method of NaCl removal is considered in the context of established iron corrosion theory (Section 4.2).

4.1. Pre-Treatment

The relative effectiveness of a treatment process can be evaluated using a standardized measure of salinity reduction, Q_e, (mgNaCl·removed·g⁻¹ ZVM). The relative effectiveness of different pre-treatments (for similar sized ZVM particles) is demonstrated by the data in Table 1 and reference [6] to be:

[H₂ + CH₄ + CO + CO₂ + N₂] > [N₂] > [air] > no pretreatment

(5)

The pre-treatment is designed to increase the amount of NaCl incorporated into the hydration shell of a Fe corrosion product during desalination.

Example Calculation of the Relative Efficiency of a Pre-Treatment: ST1a, MT2b

The rate of decline in salinity is a measure of the size of the facilities required to process a specific volume of water over a specific time period. For example, a reduction in the length of time taken to achieve a reduction of x g·L⁻¹ from 100 days to 20 days, allows:

• The size of the desalination tanks (water bodies) required to be reduced by 80% (e.g., from 1000 to 200 m³);
• The land take required for the desalination tanks to be reduced from 100–500 m² to 50–150 m².

The EC declines associated with ZVM TP can be analysed in accordance with the methodology in Section F3, where [a_t+1] is a normalized measure of the relative efficiency of the desalination process. [a_t+1] increases with increasing desalination efficiency. This analysis, summarised in Table 2 and Figure 4 for example calculations using ST1a and MT2b, establishes:

• The expected value of [a_t+1] if the ZVM is used at a specific P_w, without pre-treatment;
• The expected value of [a_t+1] following pre-treatment without considering water losses;
• The expected value of [a_t+1] following pre-treatment after consideration of water losses.

This analysis demonstrates the benefit of pre-treatment, and allows the effectiveness of different types of pre-treatment to be evaluated. In this example, a Type B pre-treatment is more effective than a Type C pre-treatment.

Table 2. Assessed Impact of Pre-Treatment on the desalination rate using [a_t+1] is an adjustable variable. Data: Figure 4a–d. Reference data set without pre-treatment: [6,21]: a_s = 0.017 m²·g⁻¹ [6]; Methodology: Appendix F3.

Trial	Expected without Pre-Treatment	With Pre-Treatment	Adjusted for Water Consumption	Improvement Due to Pre-Treatment	Improvement Due to Pre-Treatment
		Observed		Observed	Adjusted for Water Consumption
ST1a	0.00020	0.12000	0.15000	60,000%	75,000%
MT2b	0.00134	0.02000	0.10000	1498%	7491%

Table 2. Assessed Impact of Pre-Treatment on the desalination rate using [a_t+1] is an adjustable variable. Data: Figure 4a–d. Reference data set without pre-treatment: [6,21]: a_s = 0.017 m²·g⁻¹ [6]; Methodology: Appendix F3.

Trial	Expected without Pre-Treatment	With Pre-Treatment	Adjusted for Water Consumption	Improvement Due to Pre-Treatment	Improvement Due to Pre-Treatment
		Observed		Observed	Adjusted for Water Consumption
ST1a	0.00020	0.12000	0.15000	60,000%	75,000%
MT2b	0.00134	0.02000	0.10000	1498%	7491%

The data required to calculate [a_t+1] is:

• The surface area, a_s, and P_w of the ZVI used in the reference desalination data set using untreated ZVI (Section 1.2.1 and Section 2.3.2 [21]);
• The a_s and P_w of the ZVI used in the control desalination data using untreated ZVI (Section 2.3.2 [6]);
• Q_e and P_w for the individual trial (Table 1, Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15 and Figure C16).

The desalination is associated with changes in Eh (Figure C5 and Figure C14). Changes in Eh reflect changes in the rate of hydrolysis, where decreases in Eh represent increased hydrolysis, and increases in Eh represent reduced hydrolysis [67]. They indicate that NaCl removal is associated with the OH radical and related species (e.g., HO₂⁻) [67]. This is confirmed by the [a_t+1] analysis in Table 2 where hydrolysis improves the effective NaCl removal efficiency of a Type B pre-treatment by 25%, and a Type C pre-treatment by 500%.

Figure 4. Example EC changes vs. time. (a) ST1a; (b) MT2b; (c) ST1a, adjusted for water losses; (d) MT2b, adjusted for water losses. Spikes where the EC rises represent periods when the water was partially frozen. Spikes where the EC reduces to zero represent periods when the entire water body was frozen. Data: Figure C5 and Figure C14. Methodology: Appendix F3.

Figure 4. Example EC changes vs. time. (a) ST1a; (b) MT2b; (c) ST1a, adjusted for water losses; (d) MT2b, adjusted for water losses. Spikes where the EC rises represent periods when the water was partially frozen. Spikes where the EC reduces to zero represent periods when the entire water body was frozen. Data: Figure C5 and Figure C14. Methodology: Appendix F3.

PS14 (Figure F1) used a Type A gaseous pre-treatment (Table D1). A [a + t] analysis (Section F3), normalising the particle size in PS14, to the particle size used in ST1a, and setting a P_w of 20 g·L⁻¹, indicates that PS14 would be expected to provide a Q_e of 346 mgNaCl·g⁻¹ ZVM TP. Similar values of Q_e are recorded for ST3b (Table 1) which was manufactured using a Type B pre-treatment. The Type A pre-treatment undertaken in a gaseous environment can be as effective as a Type B pre-treatment undertaken in an aqueous environment.

4.2. Iron Corrosion in Saline Water

The removal of NaCl is associated with ZVM TP corrosion. The rate of corrosion, the nature of the ZVM TP corrosion products, and the nature of the Eh and pH environment created by the ZVM TP controls the rate of desalination [18,19,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100].

Two roles have been proposed for Cl⁻ during iron corrosion, they are:

• A role as a catalyser [68]: Fe⁰ + nCl⁻ = Feⁿ⁺nCl⁻ + ne⁻ and Feⁿ⁺nCl⁻ = Feⁿ⁺ + nCl⁻ when the Fe⁰ is in direct contact with the saline water. This results in the release of Feⁿ⁺ ions into the water [19,68].
• A role as phase distributor [68]. As the surface rust (FeOOH) species grows the (–OH) groups on the surface of the rust change to (–OH₂)⁺ groups. This both attracts Cl⁻ ions and allows Cl⁻ ions to migrate through the rust to the metal surface [68].

Appendix F1, Appendix F2 and Appendix G consider the principal corrosion products associated with ZVM TP and ZVI in the Eh and pH environment created by the ZVM TP (Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15 and Figure C16).

4.2.1. Pourbaix Relationships

Eh and pH define the stable (equilibrium) phases of iron which will be present in the pore waters and ZVM TP body [19]. They also define the corrosion sequence that occurs [19]. Analysis of aqueous ion (and ion adduct) equilibrium using temperature, pressure, pH and Eh is termed a Pourbaix analysis [19]. For any aqueous reaction involving one or more of ions, ion adducts (precipitates) and gaseous species, a Pourbaix analysis defines [19] the equilibrium constant, K, the Reaction Quotient, Q, the Gibbs Free Energy, G, the heat of formation, H, and the enthalpy, S, associated with the reaction (Appendix F2). Therefore, if two or more reaction products are theoretically possible a Pourbaix analysis will identify which species is the equilibrium product [19]. A Pourbaix analysis will allow the equilibrium ion concentrations and gas partial pressures to be determined for a specific Eh and pH [19].

Iron, as it corrodes, adjusts the Eh and pH of the water with time [5,6,18]. The initial change is an increase in pH followed by a subsequent decrease in pH [5,6,18]. The Eh either initially drops, or remains stable, before rising and then falling [5,6,18]. These changes are associated with the progressive formation in fresh water of [19]:

Fe⁰ → Fe(OH)₂ → FeOOH

(6)

The Eh and pH at any moment in time defines [19] the equibrium constants for Fe(OH)₂ and FeOOH. During periods when Eh and pH are constantly changing (e.g., Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16 and Figure C17 and Figure C19, Figure C20, Figure C21, Figure C22, Figure C23, Figure C24, Figure C25, Figure C26, Figure C27, Figure C28, Figure C29, Figure C30, Figure C31, Figure C32, Figure C33, Figure C34, Figure C35 and Figure C36) the water is in disequilibria [19]. Once the Eh and pH values achieve a stable level (at a constant temperature and pressure (e.g., Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16 and Figure C17 and Figure C19, Figure C20, Figure C21, Figure C22, Figure C23, Figure C24, Figure C25, Figure C26, Figure C27, Figure C28, Figure C29, Figure C30, Figure C31, Figure C32, Figure C33, Figure C34, Figure C35 and Figure C36) the water is in a chemical equilibrium [19].

Eh, pH relationships do not define reaction rates (though the rate of change and direction of change is a reflection of underlying reaction rates) [19]. They do define the equilibrium product species and where two, or more, species can occur, the equilibrium molar ratios between the products [19].

A byproduct of iron corrosion is the removal and consumption of water [19]. Water removal was observed in the trials associated with both ZVM TP and ZVM TPA (Appendix C). The principle reactions are [19]:

Fe⁰ → Feⁿ⁺ + ne⁻

(7)

H₂O = H⁺ + OH⁻

(8)

The rate of water ionisation is higher than the rate of iron ionisation [19] and the amount (wt) of water removed can exceed the weight of Fe placed in the reaction environment (e.g., Table 1, Appendix C).

4.2.2. Fe⁰ Corrosion Series

The established redox (Eh, pH) boundaries for the stable phases of iron [19,69,70,71,72,73,74,75] in saline water are shown in Figure 5a. The stable equilibrium species are Fe⁰, Fe²⁺ ions, Fe(OH)₂, Green Rust One (GR1, GR:Cl), and FeOOH. During corrosion (oxidation), the sequential reaction series in saline water which is free of sulphates and bicarbonates/carbonates is [19,69,70,71,72,73,74,75]:

Fe⁰ → Fe(OH)₂ → GR1(Cl⁻) → FeOOH

(9)

The GR1 either [43,44,45,46,47,48,49]:

• transforms directly to β-FeOOH (akaganeite), or,
• adopts the reaction series, GR1 → α-FeOOH (goethite) → β-FeOOH, or,
• adopts the reaction series GR1 → γ-FeOOH (lepidocrocite) → α-FeOOH → β-FeOOH

In the presence of sulphate the corrosion reaction series is:

Fe⁰ → Fe(OH)₂ → GR1(Cl⁻) → GR1(SO₃²⁻) → GR2(SO₄²⁻) → FeOOH (γ-FeOOH and/or α-FeOOH)

(10)

In the presence of carbonate/bicarbonate the corrosion reaction series is:

Fe⁰ → Fe(OH)₂ → GR1(Cl⁻) → GR1(CO₃²⁻) → FeOOH (γ-FeOOH and/or α-FeOOH)

(11)

The corrosion of Fe⁰ in the presence of sulphate, carbonate, and bicarbonate is addressed in Section 6.

The trials analysed in Section 4 and Section 5 were operated in saline water containing <11 mgSO₄²⁻·L⁻¹ (Table C1) and <100 mgHCO₃⁻·L⁻¹.

The corrosion of Fe⁰ to FeOOH produces adsorbed hydrogen (Fe + 2H₂O = FeOOH + 3H⁺ + 3e⁻) [19]. The adsorbed hydrogen can interact with Feⁿ⁺ ions at the FeOOH–water boundary to initiate the templated growth of GR1 growing from a FeOOH corroded ZVM particle (Figure F1) [69,70,71,72,73,74,75]:

FeOOH → GR1(Cl⁻) → FeOOH

(12)

Excess H₂ (gas) formed as 2H⁺ + 2e⁻ = H₂ [19], can form small trapped accumulations of hydrogen [5,18] within the pore space of the ZVM TP (e.g., Figure F1). This hydrogen interacts with the FeOOH corrosion products to produce Fe₃O₄ (e.g., Figure F1) [19,83]:

3FeOOH + 0.5H₂ → Fe₃O₄ + 2H₂O

(13)

The corrosion environment associated with both iron corrosion and desalination is complex and dynamic. The trials documented in Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15 and Figure C16 used water which contained trace quantities of sulphates (Table C1) and bicarbonates/carbonates.

Figure 5. Eh vs. pH for the ST1 to ST5 series of ZVM TP trials (Figure C5, Figure C6, Figure C7, Figure C8 and Figure C9). (a) Pourbaix diagram identifying the dominant Fe species phase as a function of Eh and pH [19,69,70,71,72,73,74,75]. GR = green rust. 2.8 = 2.8 mgCl⁻·L⁻¹ for R = 8; (b) ST1; (c) ST2; (d) ST3; (e) ST4; (f) ST5. Increasing the concentration of Cl⁻ (salinity) increases the pH required to achieve R = 8, and increases the pH range where akaganeite is the dominant FeOOH corrosion species. Data taken for the time period between the date the ZVM TP was added to the water to the date the partial freezing event identified in Figure C5, Figure C6, Figure C7, Figure C8 and Figure C9. This is to demonstrate the Eh vs. pH redox regime during the principal desalination phase.

Figure 5. Eh vs. pH for the ST1 to ST5 series of ZVM TP trials (Figure C5, Figure C6, Figure C7, Figure C8 and Figure C9). (a) Pourbaix diagram identifying the dominant Fe species phase as a function of Eh and pH [19,69,70,71,72,73,74,75]. GR = green rust. 2.8 = 2.8 mgCl⁻·L⁻¹ for R = 8; (b) ST1; (c) ST2; (d) ST3; (e) ST4; (f) ST5. Increasing the concentration of Cl⁻ (salinity) increases the pH required to achieve R = 8, and increases the pH range where akaganeite is the dominant FeOOH corrosion species. Data taken for the time period between the date the ZVM TP was added to the water to the date the partial freezing event identified in Figure C5, Figure C6, Figure C7, Figure C8 and Figure C9. This is to demonstrate the Eh vs. pH redox regime during the principal desalination phase.

4.2.3. Transformation from α-FeOOH → β-FeOOH

The transformation from α-FeOOH → β-FeOOH is a function of the pH and Cl⁻ concentration in the water [69,70,71]. The molar ratio, R, log (Cl⁻/OH⁻) in the feed water has been proven to be an accurate predictor of the dominant (equilibrium) precipitated FeOOH species [69,70,71]. These studies have established that when:

1

R ≤ 1.16 the only equilibrium FeOOH species which will be present is α-FeOOH [69,70,71].

2

1.16 < R < 2.25, then the observed equilibrium FeOOH will contain α-FeOOH + β-FeOOH [69,70].

3

2.25 < R < 8, the stable equilibrium FeOOH species is β-FeOOH [69,70,71]

4

R ≥ 8, the only equilibrium FeOOH species which will be present is β-FeOOH [71].

The transformation from α-FeOOH → β-FeOOH is not instantaneous, therefore analyses of FeOOH species prior to equilibrium being achieved may show the presence of α-FeOOH.

At a pH of 10, R = 8 corresponds to a Cl⁻ concentration of 2.8 mgCl⁻·L⁻¹ (4.5 mgNaCl·L⁻¹) (Figure 5a). In all the examples (ZVM TP and ZVM TPA) considered in this study (Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16, Figure C17, Figure C19, Figure C20, Figure C21, Figure C22, Figure C23, Figure C24, Figure C25, Figure C26, Figure C27, Figure C28, Figure C29, Figure C30, Figure C31, Figure C32, Figure C33, Figure C34, Figure C35, Figure C36 and Figure D1, Figure D2, Figure D3, Figure D4, Figure D5 and Figure D6), R is greater than 8, when the water pH < 13.

4.2.4. Dominant Corrosion Species Associated with ZVM TP

The Eh vs. pH cross plots (Figure 5) from 50 trials (ST1a to ST5j) indicate that the water chemistry throughout the trials was in the FeOOH precipitation zone [19]. The dominant Fe corrosion product surrounding the Fe⁰ core will be FeOOH [19]. R > 8, therefore any entrained n-Fe particles will be β-FeOOH [19,69,70,71]. Similar Eh, pH observations were made in the trials documented in Figure C1, Figure C2, Figure C3, Figure C4, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16, Figure C17, Figure C19, Figure C20, Figure C21, Figure C22, Figure C23, Figure C24, Figure C25, Figure C26, Figure C27, Figure C28, Figure C29, Figure C30, Figure C31, Figure C32, Figure C33, Figure C34, Figure C35 and Figure C36.

The corrosion product reaction sequences result in a series of concentric halos of different corrosion products surrounding the Fe⁰ particle core [68]. Example halos associated with ZVM TP are illustrated in Figure F1. In this example the initial particle corrosion results in a structure where an Fe⁰ core is surrounded by Fe(OH)₂ and the Fe(OH)₂ is then surrounded by β-FeOOH (e.g., Figure F1). The β-FeOOH acts as a template for the growth of Schiller sheets of GR1. These are transformed at their margins to β-FeOOH (e.g., Figure F1) [76,77,87,88]. Figure F1 demonstrates the presence of, and growth of, spherulitic, amorphous, entrained particles of β-FeOOH. These particles display the characteristic colour of β-FeOOH (Figure F1). The margins of the Schiller sheets have a purple colour (Figure F1). This colour is a characteristic of Na incorporation in the lattice (see Appendix H).

This corrosion structuring in saline water is demonstrated for ZVM TP (PS14) with a Q_e of 77.83 mg·L⁻¹ (Table 1) in Figure F1. This demonstrates that desalination is part of the normal corrosion process.

4.2.5. NaCl Removal in the Hydrated Layers Surrounding ZVM TP

The Pourbaix analysis [19,69,70,71,72,73,74,75] has established that desalination is associated with green rust and FeOOH corrosion species. Figure F1 demonstrates incorporation of Cl⁻ ions into entrained amorphous β-FeOOH and incorporation of Na⁺ ions into Fe^III species associated with growing Schiller sheets within a ZVM TP body. Incorporation into these species is unable to account for the high levels of NaCl removal (e.g., PS14 (Table 1, Figure F1) Q_e = 77.83 mg·g⁻¹) observed (Table 1, Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15 and Figure C16). The removed NaCl is located principally in the hydration shells associated with ZVM TP and its corrosion products [87,88,96].

The hydrated shells surrounding FeOOH can remove NaCl by adsorption [78,79,80] or by concentration in the pore water within the ZVM body, surrounding the ZVM particles (e.g., Figure D1).

Analysis of ZVM TP will be expected (e.g., Figure F1) to contain a mixture of Fe⁰, Fe^II and Fe^III species. The entrained Fe^III species will be expected to contain a mixture of γ-FeOOH, α-FeOOH, β-FeOOH.

4.2.6. Identification of FeOOH or Other Fe Species Involved in NaCl Removal

Fe corrosion species have been identified from solid material using a number of different techniques including: Raman spectroscopy; X-Ray diffraction (XRD), Differential scanning calorimetry, FTIR, UV-Visible spectroscopy, fluorospectroscopy, vibrating sample magnetometry, electron paramagentic resonance, optical analyses, Mossbauer, Synchron X-ray powder analyses etc. Entrained Fe corrosion products in aqueous solutions are commonly identified using UV-Visible-NIR Spectroscopy. The identification of specific species in aqueous solution is addressed in Appendix H.

Figure F1 clearly illustrates the problem of using a tool such as XRD to analyse the corrosion products. This problem arises as a number of different species are present which represent different stages in the corrosion process and different products in each of the different micro-environments contained in the ZVM TP.

The use of Fe derived from carbon steel in this study results in the characteristic XRD pattern for β-FeOOH changing in the presence of carbon from having dominant peaks (CPS) at 2-Theta of 11.90° (110), 26.77° (310), 33.94° (400), and minor peaks at 16.92° (200), 35.20° (211), 39.31° (301) [80,81,82], to a dominant peak at 26.65° and a minor peak at 35.31° [80]. The other peaks are either absent or within the noise range of the spectra [80]. The numbers in brackets refer to crystal faces. The FeOOH peaks are labeled and indexed to a tetragonal FeOOH phase (JCPDS File No. 34-1266) [81,82]. Highly hydrated amorphous green rust, γ-FeOOH, α-FeOOH, β-FeOOH have either no XRD peaks, or very subdued peaks. The presence of other species such as carbon, further complicates the interpretation of the traces.

XRD analyses, and similar analyses, are suited to analyses associated with incorporation of Na and Cl in the crystal lattices. They are not suitable for the analysis of ions contained within the hydration shells of the crystallites. In these circumstances, the most reliable identification tool for the entrained FeOOH species is UV-Visible-NIR spectrography, which is used in this study.

The concentration of these entrained n-Fe corrosion species in the product water body associated with ZVM TP was less than 0.05 mg·L⁻¹ in the product water. This is demonstrated in Table C2.

The reference standard UV-Visible-NIR spectra for GR1 (Cl⁻), GR2 (SO₄²⁻), hematite, γ-FeOOH, α-FeOOH, β-FeOOH (including hydrated polyionic β-FeOOH) are provided in Table 3. Hydrated polyionic β-FeOOH demonstrates a dominant absorption peaks at 225/8 nm. Minor peaks may be present at 350 and 500 nm, but are commonly absent.

Table 3. UV-Visible Spectral Bands Associated with GR1, GR2 and associated iron corrosion species. Data: [90,92,101,102,103,104,105,106,107].

Species	Hydrated Ligand Field Transition Fe3+	Ligand Field Transition Fe3+	Ligand Field Transition Fe3+	Ligand Field Transition Fe3+	Excitation to an Fe-Fe Pair	Fe(II)-Fe(III) Intervalence and Fe(III) Absorption	Fe(II)-Fe(III) Intervalence	Excitation to an Fe-Fe Pair	Excitation to an Fe-Fe Pair	Excitation to an Fe-Fe Pair
Hematite	-	300	370	430	485	-	-	555	-	900
Ferrate	-	-	-	450		-	-	550	800	-
Akaganeite (hydrated)	225	-	350	-	500	-	-	-	-	-
Akaganeite (low hydration)	-	-	-	-	500	-	-	-	-	-
Goethite	-	300	-	450	-	-	-	590	760	910
Lepidocrocite	-	300	350	420	485	-	-	-	-	-
GR2–SO4	-	-	320	410	-	550	690	-	-	-
GR1–Cl	-	-	350	485	-	550	690	-	-	-

Table 3. UV-Visible Spectral Bands Associated with GR1, GR2 and associated iron corrosion species. Data: [90,92,101,102,103,104,105,106,107].

Species	Hydrated Ligand Field Transition Fe3+	Ligand Field Transition Fe3+	Ligand Field Transition Fe3+	Ligand Field Transition Fe3+	Excitation to an Fe-Fe Pair	Fe(II)-Fe(III) Intervalence and Fe(III) Absorption	Fe(II)-Fe(III) Intervalence	Excitation to an Fe-Fe Pair	Excitation to an Fe-Fe Pair	Excitation to an Fe-Fe Pair
Hematite	-	300	370	430	485	-	-	555	-	900
Ferrate	-	-	-	450		-	-	550	800	-
Akaganeite (hydrated)	225	-	350	-	500	-	-	-	-	-
Akaganeite (low hydration)	-	-	-	-	500	-	-	-	-	-
Goethite	-	300	-	450	-	-	-	590	760	910
Lepidocrocite	-	300	350	420	485	-	-	-	-	-
GR2–SO4	-	-	320	410	-	550	690	-	-	-
GR1–Cl	-	-	350	485	-	550	690	-	-	-

The major difference in spectral behaviour between the different FeOOH species (see Section 5, and Table 3 allows UV-Visible-NIR spectra to be used (Figure C37, Appendix H) to provide a first level identification of the dominant entrained Fe^III corrosion species within the water. The spectra (Figure C37) identified hydrated polyionic β-FeOOH as the dominant entrained species. Both α-FeOOH and ferrate were identified as minor entrained Fe corrosion species (Figure C37).

FeOOH species have a high capacitance, and are widely used as adsorbant and catalytic material [80].

These observations identify two areas of investigation which may assist in increasing Q_e. They are an analysis of pH and an analysis of the parameters associated with capacitance (e.g., voltage, current, surface charge).

4.2.7. Significance of pH Change

Desalination is associated with an initial increase in pH, followed by a gradual decline in pH to an equilibrium or constant level (Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16 and Figure C17). The decrease in pH results in:

• An increase in surface protonation (H⁺, (–OH₂)⁺), with an increase in Cl⁻ ions adsorbed on the FeOOH surface [84].
• The competition between Cl⁻ ions and OH⁻ ions for the protonated sites decreasing due to both a decreased relative availability of OH⁻ and increased surface protonation resulting in an increase in the number of available adsorption sites [84].
• The replacement of blocking OH⁻ ions and H₂O at the entrances to tunnels within the β-FeOOH by Cl⁻ ions [84].

The change in Cl⁻ resulting from incorporation in the Fe corrosion products, as the pH changes, can be assessed from the equation: μ mol·Cl⁻·m⁻² FeOOH = 0.0375pH² − 0.75pH + 3.7125 [84].

Adsorption of ions from water by FeOOH increases as the FeOOH surface charge increases [84].

4.2.8. Significance of Water Consumption

The water consumption rates increased as:

• Q_e decreased when a Type A or B pre-treatment method had been used (Figure 6a).
• Q_eincreased when a Type C or D pre-treatment method had been used (Figure 6a).

Figure 6. Water consumption vs. Q_e. (a) Type A + B Pre-treatment; (b) Type C + D Pre-treatment. Data: Table 1.

Figure 6. Water consumption vs. Q_e. (a) Type A + B Pre-treatment; (b) Type C + D Pre-treatment. Data: Table 1.

This demonstrates that the Type C pre-treatment, or D pre-treatment methods create a ZVM TP product which increasingly removes NaCl as the surface charge increases, while a Type A pre-treatment or B pre-treatment method creates a ZVM TP product which replaces water consumption with NaCl consumption in the Fe corrosion products. These observations demonstrate that the removed NaCl is concentrated in the hydration shells of the Fe corrosion products.

4.2.9. Significance of Surface Charge

The surface charge on FeOOH increases with increasing NaCl concentration and decreasing pH [58,59]. Adsorption on the FeOOH proton active site FeOH^0.5− takes the form [85]

• Protonation: FeOH^0.5− + H⁺ = FeOH₂^0.5+;
• Cl⁻ capture: FeOH₂^0.5+ + Cl⁻ = [(FeOH₂^0.5+)(Cl⁻)];
• Na⁺ capture: FeOH^0.5− + Na⁺ = [(FeOH^0.5−)(Na⁺)].

The highest concentrations of the Cl⁻ ions associated with the β-FeOOH crystallites, are held within 0.25 and 0.4 nm of an –OH₂⁺ site [85,96], while the highest concentrations of the Na⁺ ions are held with 0.05 and 0.1 nm of the –OH site [85,96]. Both Cl⁻ and Na⁺ ions are also held in the hydrated shell around the crystallite [85,96].

The relative ratio, R_A, of active FeOH₂^0.5+: FeOH^0.5− sites, is demonstrated by the cation (Na⁺ + K⁺) and anion (Cl⁻) removal analyses (Table C1 and Table C2). R_A approximates 1.0 for most desalination examples (Figure 7) where R_A = Moles [Na⁺ + K⁺] Removed/Moles Cl⁻ Removed.

Figure 7. Relative removal ratios associated with Cl⁻ and Na⁺ ions in saline water. Data: Table C1 and Table C2.

Figure 7. Relative removal ratios associated with Cl⁻ and Na⁺ ions in saline water. Data: Table C1 and Table C2.

4.2.10. Recovery of Adsorbed NaCl from ZVM TP

The adsorbed [Na⁺ + K⁺ + Cl⁻] can be recovered from the hydration shells by:

• Increasing pH and reducing the surface charge. This releases adsorbed products associated with FeOOH [86].
• Extensive washing of β-FeOOH. This removes incorporated Cl⁻ ions and Cl⁻ ions held in the hydration shells [87,88].
• Actively discharging the capacitance charge held in the FeOOH. This removes incorporated Na⁺ ions (and adsorbed Na⁺ ions held in the hydration shell) [89].

These observations demonstrate that it is possible to:

• recover the Na⁺ and Cl⁻ ions separately from the used ZVM TP;
• recover the charge held in the ZVM TP following desalination;
• regenerate the ZVM TP for reuse following desalination.

Regeneration can involve reduction to Fe⁰, or changing the dominant Fe corrosion species within the recovered ZVM TP to facilitate corrosion via GR1 to FeOOH, or change the density and structure of active sites on the ZVM TP, or changes to the capacitance structure of the ZVM TP to facilitate removal of NaCl.

4.3. Removal of NaCl from Water in the Hydration Shells

NaCl removal is related to: pH, Q_e, water consumption, surface charge and capacitance. NaCl removal in hydration shells can be analysed using an adsorption model (e.g., Langmuir, Freudlich, and Redlich–Peterson) [78,79,80]:

Langmuir: Q_e = [Q_maxK_LC_e]/[1 + KLCe]
Freudlich: Q_e = K_FC_e^1/n
Redlich-Peterson: Q_e = [AC_e]/[1 + BC_e^g]
Q_e (mg/g) = [(C₀ − C_e)V]/m = V[(C₀ − C_e)/m] K_R = 1/[1 + K_LC₀]
Removal Efficiency (%) = 100 − 100C_e/C₀

(14)

where Q_e = the quantity of NaCl absorbed per unit weight of solid absorbent at equilibrium; Q_max = the maximum quantity of NaCl which could be absorbed per unit weight of solid absorbent; K_L = the Langmuir adsorption equilibrium constant, L·mg⁻¹; K_F = the amount adsorbed at unit concentration (i.e., 1 mg·L⁻¹); n = adsorption intensity, 1/n < 1.0; C₀ = The initial concentration of the solute (NaCl) in the water, mg·L⁻¹; C_e = The equilibrium concentration of the solute (NaCl) in the water, mg·L⁻¹; V = volume of the adsorbate, L; m = mass of the adsorbent, g·L⁻¹; A, B and g (where 0 < g < 1) are constants. K_R = dimensionless separation (or equilibrium) factor. K_R indicates the shape of the isotherm, where K_R > 1, is unfavorable; K_R = 1, is linear; 0 < K_R < 1, is favorable; K_R = 0, is irreversible [79]. Adsorption on β-FeOOH can decrease with increasing temperature [78], though K_R commonly decreases with increasing temperature [79].

An adsorption isotherm results in a general decline in the rate of adsorption (as absorbent sites [a_a] are utilized) until the adsorbent is saturated. At that point adsorption ceases or proceeds at a reduced rate. This pattern of EC decline is observed in Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16 and Figure C17.

4.3.1. Characteristics of NaCl Removal by Adsorption in the Hydration Shells

Pre-treatment Type B trials, ST1a to ST5j (Figure C5, Figure C6, Figure C7, Figure C8 and Figure C9), established that, Q_e:

• is not influenced (Figure 8a) by the initial NaCl concentration;
• is influenced by the concentration of ZVM TP in the water (Figure 8b);
• increases with decreasing pH (Figure 8c,d);
• is not influenced by changes in Eh (Figure 8g,h).

The EC, Eh and pH in each example (Figure C5, Figure C6, Figure C7, Figure C8 and Figure C9) were stable during the last 50 days of each trial. This demonstrated [19] that a stable chemical equilibrium position had been reached in the reaction environment. The final Eh and pH measurements made in each trial were representative of these equilibrium conditions. The term equilibrium is used in Figure 8 to refer to this Eh, or pH, in the final equilibrium phase of operation (See Section F2 for further analysis). The term maximum pH refers to the maximum pH recorded in Figure C5, Figure C6, Figure C7, Figure C8 and Figure C9. The term minimum Eh refers to the minimum Eh recorded in Figure C5, Figure C6, Figure C7, Figure C8 and Figure C9.

In a conventional adsorbent model, increasing the amount of adsorbent will increase the amount of NaCl removed. Figure 8b establishes that this assumption is not applicable to desalination as:

• Q_e decreases with increasing P_w;
• pH increases as P_w increases (Figure 8e,f);
• Q_e decreases with increasing pH (Figure 8c,d);
• Q_e is not affected by changing Eh (Figure 8g,h).

These observations demonstrate that the surface charge associated with the hydration shells controls the effective desalination rate and total amount of desalination that occurs.

Figure 8. Observed Relationships: between desalination (Q_e) and: (a) Desalination (Q_e) vs. C₀; (b) Desalination (Q_e) vs. ZVM TP concentration; (c) Desalination (Q_e) vs. Equilibrium pH; (d) Desalination (Q_e) vs. Maximum pH; (e) ZVM TP concentration vs. Equilibrium pH; (f) ZVM TP concentration vs. Maximum pH; (g) Desalination (Q_e) vs. Equilibrium Eh; (h) Desalination (Q_e) vs. Minimum Eh. Data: ST1a to ST5j Figure C5, Figure C6, Figure C7, Figure C8 and Figure C9. n = 50 samples; All trials operated under identical conditions. Water losses are not considered in calculating (Q_e) (see Appendix C for further details of water losses).

Figure 8. Observed Relationships: between desalination (Q_e) and: (a) Desalination (Q_e) vs. C₀; (b) Desalination (Q_e) vs. ZVM TP concentration; (c) Desalination (Q_e) vs. Equilibrium pH; (d) Desalination (Q_e) vs. Maximum pH; (e) ZVM TP concentration vs. Equilibrium pH; (f) ZVM TP concentration vs. Maximum pH; (g) Desalination (Q_e) vs. Equilibrium Eh; (h) Desalination (Q_e) vs. Minimum Eh. Data: ST1a to ST5j Figure C5, Figure C6, Figure C7, Figure C8 and Figure C9. n = 50 samples; All trials operated under identical conditions. Water losses are not considered in calculating (Q_e) (see Appendix C for further details of water losses).

4.3.2. Impact of Water Consumption

The analysis in Figure 8 does not consider water consumption during desalination. Integrating (into Figure 8) the salinity data in Table 1 (Table C1 and Table C2), demonstrates that Q_e reaches a maxima at a ZVM TP concentration [P_w] of between 15 and 20 g·L⁻¹ (Figure 9a).

The relationship between Q_e calculated both before and after water loss consideration (Figure 9b) provides an estimate of:

• the adjusted value of Q_e as a function of ZVM TP concentration (Figure 9c) and
• the total amount of NaCl removed as a function of ZVM TP concentration (Figure 9d).

This analysis demonstrates that the total amount of NaCl removed can approximate 10 g·L⁻¹ for ZVM TP concentrations above 40 g·L⁻¹ (Figure 9d).

Figure 9. (a) Q_e (without adjustment for water consumption) vs. ZVM TP concentration; (b) Measured relationship (Table 1) between Q_e (without adjustment for water consumption) vs. Q_e (with adjustment for water consumption); (c) Q_e (with adjustment for water consumption) vs. ZVM TP concentration; (d) Total NaCl removed, g·L⁻¹ vs. ZVM TP concentration; (e) Q_e (with adjustment for water consumption) vs. Fe⁰ BET surface area, based on a ZVM TP concentration of 10 g·L⁻¹ (for ST1-ST5, PS14) and 2.5 g·L⁻¹ for PS7; (f) Total NaCl removed, g·L⁻¹ vs. Fe⁰ surface area for a nominal ZVM TP concentration of 40 g·L⁻¹. Ion analyses samples are: ST3b, ST3f, ST6d, PS4, PS5, PS7 (Table 1): Data: Appendix C. Surface Area Data: ST1–ST5 [6]; PS14 [6]; PS7 [28].

Figure 9. (a) Q_e (without adjustment for water consumption) vs. ZVM TP concentration; (b) Measured relationship (Table 1) between Q_e (without adjustment for water consumption) vs. Q_e (with adjustment for water consumption); (c) Q_e (with adjustment for water consumption) vs. ZVM TP concentration; (d) Total NaCl removed, g·L⁻¹ vs. ZVM TP concentration; (e) Q_e (with adjustment for water consumption) vs. Fe⁰ BET surface area, based on a ZVM TP concentration of 10 g·L⁻¹ (for ST1-ST5, PS14) and 2.5 g·L⁻¹ for PS7; (f) Total NaCl removed, g·L⁻¹ vs. Fe⁰ surface area for a nominal ZVM TP concentration of 40 g·L⁻¹. Ion analyses samples are: ST3b, ST3f, ST6d, PS4, PS5, PS7 (Table 1): Data: Appendix C. Surface Area Data: ST1–ST5 [6]; PS14 [6]; PS7 [28].

4.3.3. Impact of ZVM Particle Size

The [a_t+1] analysis established that NaCl removal is a function of a_s and P_w (Appendix F3). The measured relationship between Q_e and a_s (Figure 9e,f), demonstrates that Q_e increases with a_s.

This relationship confirms that it is possible for 1 g ZVM to remove >1 g NaCl [25]. This observation is consistent with a model of accretionary, polyionic, akaganeite/FeOOH nano rod formation where the bulk of the adsorbed NaCl is electrostatically bound in the hydration shells and the ionic layer surrounding each FeOOH, e.g., [22,23,25,78,79,90,91,92,93,94,95,96,97,98,99,100,108].

4.4. Desalination Associated with NaCl Concentration in the Pore Waters within the ZVM TP Bed

Molecular modeling demonstrates [96] that when the water contains 0.05–0.1 M NaCl·L⁻¹, the Cl⁻ concentrations associated with the ionic layer at the akaganeite crystal end (001) face are 7.34 M Cl⁻·L⁻¹ [96]. The Cl⁻ concentrations in the hydrated shell adjacent to the longitudinal (100), (110) crystal faces are 1.47 M Cl⁻·L⁻¹ [96].

This situation can only arise if the Fe(H₂O)₆³⁺ species create chemical gradients in the water which physically attract NaCl to the Fe(H₂O)₆³⁺ species to enable akaganeite formation and growth [96].

The molecular modeling [96] for akaganeite formation creates an osmotic gradient between the hydrated shells surrounding the akaganeite and the surrounding water body [108]. Osmotic theory would expect the Cl⁻ ions to migrate from the akaganeite water shell (and ionic layer) into the surrounding water body [108,109].

4.4.1. Osmotic Pressure

The osmotic pressure created between fresh water and the longitudinal water shell of the akaganeite nano-crystals is 3.6 MPa, where osmotic pressure [O_p] is calculated as [O_p] (MPa) = [c]RT [82,83], where [c] = salt concentration in the water (M·L⁻¹), R = gas constant, T = temperature, K. At 25 °C, RT = 2.480 kJ·M⁻¹ [108,109], and [O_p] (MPa) = 2.480 [c].

Experiments which have examined the molar relationship between OH:Fe in water, in conjunction with the molar relationship between Cl:Fe in water, have established [110] that:

• Cl⁻ is removed in Fe^III polymers (general form [Fe_m(H₂O)_6m−n(OH)_n]^3m−n) [110].
• No Cl⁻ is observed in the first co-ordination shell of Fe in the polymers when OH:Fe > 2 [110].
• Increasing the Fe³⁺:Cl⁻ ratio in the water results in water molecules in the first hydration shell being exchanged for Cl⁻ [110]. The resultant polymers and ferric hydroxide gels (e.g., orange spheroids in Figure F1) possess an akaganeite, or goethite, structure [110].

These observations indicate that the critical control on the rate of desalination is the rate of discharge of Feⁿ⁺ ions into the pores within the ZVM mass and the ability of those pores to access saline water from the overlying water body. This process creates an effective osmotic gradient between the water body and the pore water surrounding the ZVM, which results in a net migration of NaCl into the pore waters surrounding the ZVM.

This mechanism predicts that in a diffusion environment [the salinity of the hydration shells] > [the salinity of hydrated ZVM TP + associated pore waters] > [Salinity of the principal water body].

4.4.2. Role of Hydration Shells

The β-FeOOH polymorph is an unstable structure, which is stabilized by Cl⁻ [111]. When Cl⁻ is absent, or drops below a critical level in the pore waters within the ZVM TP, another FeOOH species (e.g., goethite) will form [111].

The Al⁰ in the ZVM TP corrodes in water to produce Al(–OH₂)ⁿ⁺ sites and is able to remove both Na⁺ and Cl⁻ ions as: [(AlOH₂^0.5+)(Cl⁻)], [(Al₃O^0.5−)(Na⁺)], and [(Al₃OH^0.5+)(Cl⁻)] [112]. Both Fe and Al adopt similar and complementary surface sites [111,112,113].

The precipitated FeOOH species form a repeating sequence of atoms and radicals [113]:

• terminal surface (no hydration, e.g., FeOH^0.5−): (OH)–(OH)–Fe–O–O–Fe–R (where R = a repeat of the stoichiometric atomic layer sequence, or tethering surface. R can included hydrated layers);
• interface terminal surface (no hydration, e.g., FeOH₂^0.5+): (H₂O)–(H₂O)–(OH₂)–(OH)–Fe–O–O–Fe–R
• double hydrated terminal surface: (H₂O)–(H₂O)–(OH)–(OH)–Fe–O–O–Fe–R
• double hydrated interface terminal surface: (H₂O)–(H₂O)–(OH₂)–(OH)–Fe–O–O–Fe–R

The Na⁺ ions are removed within the (H₂O)–(H₂O)–(OH)–(OH)–Fe–O–O–Fe–R sequence. The Cl⁻ ions are removed within the (H₂O)–(H₂O)–(OH₂)–(OH)–Fe–O–O–Fe–R sequence.

The example illustrates two hydration shells. The actual number of hydration shells can exceed 2, i.e.,

(H₂O)–……..–(H₂O)–(OH)–(OH)–Fe–O–O–Fe–R and (H₂O)–……–(H₂O)–(OH₂)–(OH)–Fe–O–O–Fe–R

(15)

The hydration shells can account for >41% of the weight of FeOOH [114], and increase the effective surface area of the FeOOH [114]. 5 hydration shells approximate to 46 wt % of FeOOH·mH₂O.

This general phase structuring allows NaCl removal to be undertaken at the terminal sites and within the corrosion zone. It potentially allows any hydrated FeOOH species to be structured to form an effective desalination agent.

FeO_xH_y nano-particles rapidly aggregate in saline water [115]. The aggregated colloidal particle size increases with increasing water salinity [115].

4.5. Role of Surface Charge and Capacitance in Assessing Desalination Efficiency

This section considers:

• the observed surface charge and capacitance associated with desalination;
• the capacitance characteristics of ZVM TP prior to use for desalination;
• the capacitance characteristics of the ZVM TP following desalination.

The trial results in Appendix C indicate that an understanding of surface charge and capacitance may allow elucidation of the ZVM TP characteristics required to maximize Q_e and reduce the time required to achieve a specific level of Q_e.

4.5.1. Observations Made during Desalination

The increase in Q_e which is associated with a decrease in pH (Figure 8c,d) can be linked to surface charge and proton uptake associated with FeOOH, i.e.,

• the proton uptake increases with both increasing salinity and decreasing pH; (e.g., at 0.1 M (5.844 g·L⁻¹) NaCl proton uptake ([H⁺]_ads (μM NaCl)⁻¹·m⁻² = −0.7006pH + 4.3174; at 0.01 M (0.5844 g·L⁻¹) NaCl proton uptake ([H⁺]_ads (μM NaCl)⁻¹·m⁻² = −0.7006pH + 3.8174 [116]), M = moles NaCl·L⁻¹; m⁻² refers to the BET surface area;
• the surface charge (C·m⁻²) increases (at 0.1 M (5.844 g·L⁻¹) NaCl) with decreasing pH (e.g., Surface charge, C·m⁻² = −0.0009pH² − 0.033pH + 0.2957 [116]).

Q_e increases as a function of the magnitude of the proton uptake shift and surface charge shift (Figure 10a,b). The Q_e increase is also a function of pH (Figure 10c–j). A statistical correlation is present between Q_e and the change in surface charge (i.e., the difference between the surface charge at the maximum pH and the equilibrium pH) (Figure 10g).

Decreases in salinity removal with time (e.g., Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16 and Figure C17) can be interpreted as being related to a decrease in the surface charge and a decrease in the availability of H⁺ sites.

4.5.2. Impact of Changes in Surface Charge on Q_e during Desalination

Q_e is maximized under conditions where the change in surface charge and change in proton adsorption (associated with the change in redox environment between the maximum pH and equilibrium pH) is minimized (Figure 10e–g).

The surface charge is proportional to the capacitance [117]. The capacitance increases with BET surface area of the FeOOH [91], e.g., Inner Layer Capacitance, C, μF cm⁻² = −1.629a_s + 208.955; a_s = −0.61939C +128.27; C = Surface Charge/(D₁ – D₂); D₁ = potential at the 0 plane. D₂ = potential at the beta plane [117]. Surface Charge = C(D₁ − D₂) [117].

The surface area was calculated for the case where (D₁ − D₂) = 1 (Figure 10k). This analysis demonstrates that the surface area increases as Q_e increases (Figure 10k), and the effectve surface charge decreases as surface area increases (Figure 10a).

The formation of FeOOH species such as goethite and ledidocrocite (which have a lower capacity for Cl⁻ adsorption) may result in different surface charge relationships and lower NaCl adsorption rates [116,118].

Figure 10. ZVM TP: Surface charge and Q_e during desalination. (a) Q_e (without adjustment for water consumption) vs. surface charge; (b) Measured relationship Q_e (without adjustment for water consumption) vs. proton adsorption; (c) Surface charge vs. proton adsorption; (d) Maximum and equilibrium Eh vs. proton adsorption; (e) Q_e (without adjustment for water consumption) vs. change in surface charge; (f) Measured relationship Q_e (without adjustment for water consumption) vs. change in proton adsorption; (g) Expected Q_e (without adjustment for water consumption) vs. surface charge; (h) Equilibrium salinity vs. change in surface; (i) Change in salinity vs. change in surface charge; (j) Desalination cycle shown by water pH vs. surface charge; (k) Q_e vs. calculated FeOOH surface area, when (D₁ − D₂) = 1. Data: ST1a–ST5j.

Figure 10. ZVM TP: Surface charge and Q_e during desalination. (a) Q_e (without adjustment for water consumption) vs. surface charge; (b) Measured relationship Q_e (without adjustment for water consumption) vs. proton adsorption; (c) Surface charge vs. proton adsorption; (d) Maximum and equilibrium Eh vs. proton adsorption; (e) Q_e (without adjustment for water consumption) vs. change in surface charge; (f) Measured relationship Q_e (without adjustment for water consumption) vs. change in proton adsorption; (g) Expected Q_e (without adjustment for water consumption) vs. surface charge; (h) Equilibrium salinity vs. change in surface; (i) Change in salinity vs. change in surface charge; (j) Desalination cycle shown by water pH vs. surface charge; (k) Q_e vs. calculated FeOOH surface area, when (D₁ − D₂) = 1. Data: ST1a–ST5j.

4.5.3. Measured Capacitance of the ZVM TP Prior to Usage for Desalination

Figure 10 demonstrates that Q_e is a function of surface charge and capacitance. Two randomly selected Cu⁰ sheathed pellets (from the ZVM TP used in the analysis series ST1 to ST5 (which demonstrate desalination in Figure C5, Figure C6, Figure C7, Figure C8 and Figure C9)) were converted into dry cells by placing a steel electrode in the centre of the ZVM and a second steel electrode on the Cu⁰ shell of the pellet. The cell was placed on non conductive material (PVC) and attached to a multi-meter. The size of the current and voltage varied between pellets and with time.

• Cell 1: 7.5 cm × 1.5 cm OD; total weight = 56 g ZVM product + Cu⁰ sheath. The two ends of the dry cell were exposed to air (Figure 11).
• Cell 2: 5 cm × 1.5 cm OD containing 11 g ZVM product + 5 g NaCl + 2 g H₂O. The two ends of the dry cell were sealed with PVC (Figure 11).

These cells established that fresh ZVM TP pellets (manufactured by a Type B pre-treatment) placed in the saline water (Appendix C) were electrically active (Figure 11). The dry cells also established that the treated ZVM displayed significantly higher levels of electrical activity in the presence of NaCl + H₂O (Figure 11).

Figure 11. Measured currents and voltages as a function of time, associated with ZVM TP (ST1 to ST5 series) prior to usage for desalination. (a) Voltage vs. Time; (b) Current vs. Time. (c) Current vs. Voltage.

Figure 11. Measured currents and voltages as a function of time, associated with ZVM TP (ST1 to ST5 series) prior to usage for desalination. (a) Voltage vs. Time; (b) Current vs. Time. (c) Current vs. Voltage.

4.5.4. Relationship between Capacitance and Desalination

The electrical analysis (Figure 11) demonstrates that ZVM TP can act as a charge receiver and storage unit (i.e., a capacitor or battery). In carbon steels (and carbon steel powders) the carbon content is typically within the range 0.15%–1.3% [119]. When the carbon content is <0.9% the corrosion sequence is Fe⁰ → Fe^II → Fe^III [119]. When the carbon content is >0.9% the corrosion sequence is Fe⁰ → Fe^III [119]. The Fe^III species act as an anodic layer and carbon acts as a cathode.

The ZVM TP (formed from carbon steel) during desalination is effectively structured as a Fe⁰:Fe₂O₃:C capacitor, where the saline water acts as an electrolyte. Fe capacitor studies place Fe^III species within their generic group (Fe₂O₃ [19]).

Under this model (e.g., [120]) NaCl removal at anodic sites (Fe₂O₃) will be associated with capacitor charging (e.g., formation of Fe³⁺ ion adduct species) while capacitor discharge will be associated with releases of NaCl into the water [120]. These gross changes are reflected in the EC measurements where a general EC decline is associated with oscillations between higher and lower EC values (Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16 and Figure C17).

4.5.5. Capacitance

The capacitance can be calculated [121,122] as:

C_s = i/(m(ΔE/Δt)

(16)

where C_s = specific capacitance, F·g⁻¹, i = average current (mA·s⁻¹) applied, m = mass of ZVM, ΔE/Δt = average slope of the discharge curve after iR drop where i and R are current and resistance.

The Energy Density, E, (Wh·kg⁻¹) is [121]:

E = 0.5C_s (ΔV)²

(17)

where (ΔV) (V) = the potential range. The power density, P (kW·kg⁻¹) is [121]:

P = E/t_d

(18)

t_d = time (s) to discharge.

The dry cells in Figure 11 analyze two situations:

• Cell 1 represents the typical capacitance associated with discharged ZVM TP prior to placement in saline water.
• Cell 2 demonstrates that in saline water (following, or during desalination), the presence of NaCl around the ZVM TP has the effect of charging the cell and increasing its specific capacitance. These features are quantified in Table 4.

Table 4. Specific capacitance, energy density and power density associated with ZVM TP pellets prior to use in desalination. These pellets were used in ST1a to ST5j (Appendix C), Cell 1 and Cell 2. x = hours since start of measurement.

Variable		Cell 1	Cell 2
Current	mA	0.08 exp(−0.24x)	4.33 exp(−0.0119x)
R2		0.8463	0.408
Voltage	mV	947.23 exp(−0.0216x)	781.47 exp(−0.0028x)
R2		0.858	0.494
i	mA	0.016	2.860
m	g	56.000	11.000
(ΔE/Δt)		0.01531	0.00081
Cs	F·g−1	0.018	321.208
ΔV	V	0.199	0.734
E	Wh·kg−1	0.00036	86.411
td	s	12,469	4452
P	kW·kg−1	0.00000003	0.01941

Table 4. Specific capacitance, energy density and power density associated with ZVM TP pellets prior to use in desalination. These pellets were used in ST1a to ST5j (Appendix C), Cell 1 and Cell 2. x = hours since start of measurement.

Variable		Cell 1	Cell 2
Current	mA	0.08 exp(−0.24x)	4.33 exp(−0.0119x)
R2		0.8463	0.408
Voltage	mV	947.23 exp(−0.0216x)	781.47 exp(−0.0028x)
R2		0.858	0.494
i	mA	0.016	2.860
m	g	56.000	11.000
(ΔE/Δt)		0.01531	0.00081
Cs	F·g−1	0.018	321.208
ΔV	V	0.199	0.734
E	Wh·kg−1	0.00036	86.411
td	s	12,469	4452
P	kW·kg−1	0.00000003	0.01941

Table 4 demonstrates that the net impact of NaCl is to increase C_s, E and P. The energy density (E) of Cell 2 is comparable with the energy density recorded from Fe₂O₃:C composite capacitors (Figure 12).

Figure 12. Energy densities and specific capacitance recorded by Fe₂O₃:C composite capacitors. Reference Data Sources: [121,122,123,124,125,126,127,128,129,130,131,132,133,134,135,136,137,138,139,140,141,142]. ZVM Cell 2 = unused ZVM TP prior to use for desalination.

Figure 12. Energy densities and specific capacitance recorded by Fe₂O₃:C composite capacitors. Reference Data Sources: [121,122,123,124,125,126,127,128,129,130,131,132,133,134,135,136,137,138,139,140,141,142]. ZVM Cell 2 = unused ZVM TP prior to use for desalination.

4.5.6. Voltage/Current Changes Associated with the Cessation of Desalination

The gradual removal of Na⁺ and Cl⁻ from the water will result in the gradual reduction in both the specific capacitance and the columbic efficiency with time. These factors when combined with the trapping of Na⁺ within the ZVM, will result in a major fading of the desalination rate after either a critical number of charge:discharge cycles (Eh:pH oscillations) has occurred [143], or the salinity (electrolyte concentration) drops below a specific level. This situation was demonstrated in all the trial examples when the desalination period was extended to Day 250 (Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16 and Figure C17).

The effective floor on desalination (or cessation of desalination) indicates that a minimum salinity (or electrolyte concentration) is required in order to drive the discharge reaction. Similar observations are associated with Na–Fe₂O₃:C battery analyses [143,144].

The current and voltage associated with discharge from the fresh pellets declined with time from a Level, L₁ to a Level, L₂ (Figure 13) In the presence of NaCl the initial current and voltage increased from a Level L₁ to a Level L₃ (Figure 13).

Figure 13. Relationship between voltage and current associated with used ZVM TP pellets following desalination (ST1a to ST5j), L₄, and the unused ZVM TP prior to desalination, Cell 1 (L₁, L₂) and Cell 2 (L₃). The pellet series are labeled S1 to S5 to distinguish them from the analyses (ST1a to ST5j) in Appendix C, as the measurements were made 2 years after termination of the desalination trials.

Figure 13. Relationship between voltage and current associated with used ZVM TP pellets following desalination (ST1a to ST5j), L₄, and the unused ZVM TP prior to desalination, Cell 1 (L₁, L₂) and Cell 2 (L₃). The pellet series are labeled S1 to S5 to distinguish them from the analyses (ST1a to ST5j) in Appendix C, as the measurements were made 2 years after termination of the desalination trials.

The voltage and current produced by the pellets associated with ST1a to ST5j were measured two years after the conclusion of the desalination trials in order to investigate the impact of desalination on capacitance.

The measured current clustered at a Level L₄ (Figure 13). The measured voltage was between L₂ and L₁. The ZVM pellets following conclusion of the desalination trials were in a discharged state where the discharged voltage is intermediate between L₁ and L₂. This demonstrates [143,144] that the effective cessation of desalination (Figure 13) is associated with a partially discharged capacitor. Appendix H addresses the significance of this observation for Na ion removal.

This interpretation (supported by electrical measurements) demonstrates that the initial model of NaCl removal by adsorption [20,21,25] is over-simplistic, and that focussing on the capacitance of the ZVM TP during manufacture will allow Q_e to be increased and allow the time required to desalinate water to a specific salinity to be reduced.

4.5.7. ZVM TP Pellet Capacitance Following Desalination

The capacitance of each ZVM TP pellet used in trials ST1a to ST5j (Figure C5, Figure C6, Figure C7, Figure C8 and Figure C9) was measured following the cessation of desalination and the amount of charge stored in the pellets was calculated as:

Energy Stored, E_s, (Joules) = 0.5[Measured Capacitance (F) × (Voltage (V))²] Potential Power Generated (W) = E_s/Dissipation time, s

(19)

This analysis (Figure 14) establishes that the recovered (discharged) pellets have a stored charge in the range 0.00001 to 0.3 J. It was observed that if the pellets were subsequently moistened, the capacitance (F) increased by a factor of 100–1000. The dry pellets had a measured capacitance in the range 0.1–20 × 10⁻⁶ F. The pellets when in operation in a water body (at the conclusion of the desalination trials) had a capacitance (F) of between 10 and 20,000 × 10⁻⁶ F and had an effective stored charge within the range 0.00001 to 0.3 kJ (E = 0.0006 to 16.7 Wh·kg⁻¹).

Figure 14. Relationship between measured voltage and stored charge associated with ZVM pellets.

Figure 14. Relationship between measured voltage and stored charge associated with ZVM pellets.

This observation is consistent with a desalination model where the capacitance comes from charge storage at the ZVM TP–electrolyte (water) interface [143] and the specific capacitance and current density reduces with time. This is discussed further in Appendix F2.2 and Appendix H and is illustrated in Figure F2, Figure F3 and Figure H1.

4.5.8. Desalination Model Associated with ZVM TP

In Fe⁰:Na–Fe₂O₃:C capacitors [143,144] structural dislocations resulting from Na⁺ release reduce the permeability (and efficiency) of the capacitor with time. The Fe₂O₃ section of the Fe⁰:Na–Fe₂O₃:C ZVM unit can be schematically viewed [89] as a composite slab (Figure 15) composed of different Fe_nO₂ layers (AB, CA, BC).

Figure 15. (a) Schematic representation of the structural changes which occur in the anodic Fe₂O₃ schiller sheets (Figure F1) during desalination of water. C_s values are taken from Table 4 and Figure 15b; (b) Stored energy contained in the ZVM TP during desalination and following desalination.

Figure 15. (a) Schematic representation of the structural changes which occur in the anodic Fe₂O₃ schiller sheets (Figure F1) during desalination of water. C_s values are taken from Table 4 and Figure 15b; (b) Stored energy contained in the ZVM TP during desalination and following desalination.

The first stage in the desalination process is the formation of prismatic structures (e.g., colloidal nano-particles, or Schiller sheets) containing tetrahedral sites occupied by Fe³⁺. FeOOH grows as Schiller sheets, when templated on a substrate [87,88].

The initial Fe₂O₃ corrosion product adopts a cuboid structure (P3) dominated by prismatic sites of the form (AB, BC, AC) (Figure 15) during the initial part of the charge cycle [89]. The Na⁺ ions are accommodated at the octahedral sites located between the Fe_nO₂ layers (AB, CA, AB) [89]. This O₃ structure develops during cell charging (desalination) [89]. The prismatic sites are energetically stabilized by the Na_xO ions [89]. The glide vectors of the Fe_nO₂ layers during this transition can create a variety of staking faults in the crystal structures particularly during Na⁺ extraction (reversals of the desalination process) [89] caused by temperature, Eh and pH oscillations [5,6,18]. During Na⁺ extraction vacant face sharing tetrahedral sites are formed between the Fe_nO₂ layers (Figure 15) [89]. These sites are filled by Fe³⁺ (and other cations contained in the water) [89]. The Fe₂O₃–water interface is characterized by an outer layer of adsorbed H⁺ ions [145]. This positively charged surface attracts the negatively charged Cl⁻ ions (which are present as Na⁺:Cl⁻ dissolved ion adducts [145].

These observations demonstrate the type of Fe corrosion structure required in the ZVM TP to maximise NaCl removal.

4.6. Statistical Analysis of the Change in Salinity with Time

The analyses in C1 to C17 (collated in Figure 16) indicate that passive diabatic desalination at ambient temperatures is gradual with a progressive increase in the amount of NaCl removed with time. A proportion of trials showed an initial increase in EC due to the release of ions from the ZVM TP into the water. All of the trials established some level of desalination after 120–150 days. The highest levels of desalination were associated with the most saline water (i.e., the rate of desalination increases with increasing water salinity) (Figure 16).

Figure 16. Statistical Trials: Relationship between EC change and the time interval which has lapsed following placement of the ZVM TP in the water. (a) Observed EC declines within 4 h; (b) ZVM TP concentrations associated with declines observed in a 4 h period; (c) Observed EC changes within 7 days; (d) Observed EC changes within 30 days; (e) Observed EC changes within 60 days; (f) Observed EC changes within 90 days; (g) Observed EC changes within 120 days; (h) Observed EC changes within 150 days; (i) Observed EC changes within 180 days; All the trials were undertaken simultaneously. Temperature: Figure 1.

Figure 16. Statistical Trials: Relationship between EC change and the time interval which has lapsed following placement of the ZVM TP in the water. (a) Observed EC declines within 4 h; (b) ZVM TP concentrations associated with declines observed in a 4 h period; (c) Observed EC changes within 7 days; (d) Observed EC changes within 30 days; (e) Observed EC changes within 60 days; (f) Observed EC changes within 90 days; (g) Observed EC changes within 120 days; (h) Observed EC changes within 150 days; (i) Observed EC changes within 180 days; All the trials were undertaken simultaneously. Temperature: Figure 1.

4.7. Cost of Desalination Using ZVM TP

Figure 16 and Figure C1, Figure C2, Figure C3 , Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16 and Figure C17 have established that placement of ZVM TP in water will result in a decline in water EC and salinity. The magnitude of the salinity decline is sufficient to have a substantial impact on crop yields (Figure B1) if the water is used for irrigation, or livestock yields if the water is used as livestock feed water (Figure B2).

4.7.1. Economics of ZVM TP without Reuse or Regeneration

The cost of Fe⁰ powders varies with time, grade, and location. The desalination cost is:

Desalination Cost, $·m⁻³ = (((W_s/Q_e)(P_Fe))/R_u) − (P_wR_v)

(20)

where W_s = water salinity drop required, kg·m⁻³; P_FE = ZVM cost, $·t⁻¹; R_u = number of times the ZVM TP is reused; R_v = Residual value, $·t⁻¹, of the ZVM TP. The desalination cost reduces as W_s reduces, P_Fe reduces, Q_e increases, R_u increases, P_w decreases, and R_v increases.

For example, if the Fe⁰ powders are sourced for $400–$900 t⁻¹, then based on Q_e = 300 kg·t⁻¹, 1 t Fe⁰ will be expected to remove 300 kg NaCl. The cost of desalination ($·m⁻³) is a function of the salinity reduction required (Figure 17) and the number of times a batch of ZVM TP can be reused.

Figure 17. Desalination cost using ZVM TP, assuming no residual value for the ZVM TP and no reuse of the ZVM TP. P_w of 20 g·L⁻¹, and Q_e = 300 mg·g⁻¹.

Figure 17. Desalination cost using ZVM TP, assuming no residual value for the ZVM TP and no reuse of the ZVM TP. P_w of 20 g·L⁻¹, and Q_e = 300 mg·g⁻¹.

4.7.2. Reuse of the ZVM TP

The residual ZVM TP can be reused as a water treatment agent, or reprocessed for reuse as a desalination agent, or can be reused as an iron ore feedstock, or can be reused as a gas desulphurization agent (e.g., in an iron sponge), or reused in the construction of semi-conductors. The used ZVM TP can be regenerated as Fe⁰ (by reduction) or restructured (e.g., by NaCl removal) to allow reuse for desalination, or another purpose. An example life cycle for a batch of ZVM TP is illustrated in Figure 18.

Figure 18. Example Life Cycle of a batch of ZVM TP which is used to desalinate water.

Figure 18. Example Life Cycle of a batch of ZVM TP which is used to desalinate water.

This residual value associated with reuse is location and application specific and will reduce the desalination cost further.

4.7.3 Reuse of ZVM TP as an Iron Sponge

The iron sponge gas desulphurization process has been in commercial use for >100 years [146]. It is based on the reaction [147,148]:

In gas: Fe_xO_y (s) + yH₂S (g) = Fe_xS_y (s) + yH₂O
In water: H₂S (g) + H₂O = H₂S (aq) = HS⁻ + H⁺
2Fe³⁺ (e.g., Fe₂O₃) + HS⁻ = 2Fe²⁺ (e.g., Fe₃O₄) + H⁺ + S (s)

(21)

1 t Fe (contained in ZVM TP) may remove 0.3 to 0.83 t S contained in a gas (or water) before regeneration. The sulphur is recovered by floatation from water [148]. Regeneration of the iron sponge by blowing air through the iron sulphide product [146,148], or by blowing air through water containing the 2Fe²⁺ (e.g., Fe₃O₄) [148]:

2Fe₂S₃ (s) + 3O₂ (air) = 2Fe₂O₃ (s) + 6S + heat
2Fe₃O₄ (s) + 0.5O₂ (air) = 3Fe₂O₃ (s)

(22)

Treatment of the nFe₂O₃ (s) + mS mixture with water under H₂ or N₂ at a pressure of 10.1 MPa (and elevated temperature) will result in the Fe₂O₃ catalysing the reaction [149]:

nS(s) + 2H₂O = SO₂ + (n − 1)H₂S + (3 − n)H₂

(23)

This cyclic process allows the ZVM TP to be reused for desulphurization, desalination and water treatment. A potential life cycle for a batch of ZVM is shown in Figure 18. The full life cycle allows the costs (depending on location and market conditions) associated with the ZVM and pre-treatment to be amortized over the various reuse/reapplication stages (Figure 18). Equation (21) demonstrates that any HS⁻ ions in the water will be reformulated during desalination to release sulphur and H⁺ ions. This may allow both desulphuristion and desalination to be undertaken simultaneously.

The H₂S removal operating costs for a sour gas producer using Fe based technologies vary with gas composition, competition between suppliers, field, and geographical location. An indicative operating cost is in the range $10,000–30,000 t⁻¹ sulphur removed. This indicates that if the ZVM TP (following desalination) can be sold for reuse as a sulphur removal agent for $400–$950 t⁻¹, the effective cost of desalinating the water will reduce to $0 m⁻³. The gross sales price is $550–$1,350 t⁻¹ sulphur removed.

4.7.4. Reducing the Capital Costs Associated with Desalination

Q_e and P_w define the amount of ZVM TP required to achieve a specific level of desalination. The time taken to achieve the required level of desalination is defined by the EC decline rate. The size of desalination facilities required is therefore a function of the amount of desalinated water required per unit time and the time taken to process a batch of water to a specific salinity level.

Trials ST1a to ST5j established that the salinity declined to a stable or equilibrium level and then abruptly ceased (Figure C5, Figure C6, Figure C7, Figure C8 and Figure C9). A number of freezing or partial freezing events occurred during the trials. The general decline in EC between the onset of the trial and the first freezing event was examined. In order to distinguish discussion of part of the trial, from the full trial data set, the following notation is used; (i) ST is replaced by S, i.e., ST1 becomes S1; (ii) The individual trial within a salinity grouping is labelled T1 to T10, i.e., ST1a becomes S1:T1. The individual plots of EC vs. time for each of these trials are provided as Figure D2, Figure D3, Figure D4, Figure D5 and Figure D6.

The decline in EC with time in each example can be modelled with a polynomial equation of the form:

EC, mS·cm⁻¹ = (Ax² − Bx + C)/F

(24)

where x = day number. In this study the first day of the trial is Day 59. A, B and C are constants. C represents the salinity (EC) at Day = 0. C can be replaced with salinity of the water at the onset of desalination. In this instance x can be replaced with the number of days since the onset of desalination. F = a factor required to the change units in the equation to mS·cm⁻¹ (e.g., Figure E3) The observed polynomial regression equations for each trial are provided in Figure D2, Figure D3, Figure D4, Figure D5 and Figure D6.

The same data set (Figure D2, Figure D3, Figure D4, Figure D5 and Figure D6) could be interpreted using a regression power, or exponential function. The general exponential equation is:

EC, mS·cm⁻¹ = (C·e^(−Dx))/F

(25)

where D = constant. The observed values of D ranged between 0.0042 and 0.0109 (Figure 19). R² (for the exponential equation) was typically >0.91, but ranged between 0.604 and 0.995 (Figure D2, Figure D3, Figure D4, Figure D5 and Figure D6). Constant D reduces as the concentration of ZVM TP, or ZVM TP + Cu⁰ copper sheathing increases (Figure 19). The observed values of D and R² are provided in Figure D2, Figure D3, Figure D4, Figure D5 and Figure D6 for Trials ST1a–ST5j).

Figure 19. Redox desalination constants (D) associated with an exponential regression correlation. (a) Constant D vs. ZVM concentration; (b) Constant D vs. weight of Cu⁰ pellet sheath; (c) Constant D vs. weight of ZVM + Cu⁰ pellet sheath; (d) EC vs. Length of desalination period (days), for the observed range of values of Constant D, and Constant C (initial salinity) = 17 mS·cm⁻¹ at t = 0 days.

Figure 19. Redox desalination constants (D) associated with an exponential regression correlation. (a) Constant D vs. ZVM concentration; (b) Constant D vs. weight of Cu⁰ pellet sheath; (c) Constant D vs. weight of ZVM + Cu⁰ pellet sheath; (d) EC vs. Length of desalination period (days), for the observed range of values of Constant D, and Constant C (initial salinity) = 17 mS·cm⁻¹ at t = 0 days.

The unit cost of desalination ($·m⁻³) decreases as P_w decreases. The size of the facilities required to process a specific volume of water each year (and the length of time required to process a batch of water) decreases as Constant D increases.

Q_e and P_w control the amount of NaCl that can be removed by a specific batch of ZVM and control operating costs. Constant D controls the size of the facilities required (capital costs), and associated land take required, to deliver a specific volume of water per unit time.

Trials PS15, and PS16 (see Section 6) demonstrate that Constant D can be increased to >0.04 by operating the desalination process in a pressured reactor (operated at <0.1 MPa) where air is replaced by CO₂ and N₂. For example, processing of 1000 m³ in an open tank or reservoir with Constant D = 0.07 (Figure 19) will take 148 days to reduce the salinity from 10 to 3.68 g·L⁻¹. Operating the same ZVM TP (PS16) in a pressured reactor increases Constant D (Section 6) to 0.462. This will allow the same volume of water to be processed to the same salinity level (3.68 g·L⁻¹) in 22 days. This would allow delivery of 1000 m³ of partially desalinated water over 148 days using a pressured batch reactor with a treatment capacity of 152 m³.

4.7.5. Improving the Economics of Partial Desalination by Increasing the Operating Temperature

The desalination can be defined in terms of a rate constant (k_observed) [18] which is linked to both the amount of ZVM and surface area of the ZVM in the water, where [R_ad] = ([a_s]·[a_a]). [R_ad] represents the number of available reaction or adsorption sites on the ZVM TP. It follows that:

k_observed = k_actual × [P_w] × [R_ad]

(26)

At a constant [P_w] the reduction of NaCl to products must therefore adopt the general form [150]:

Rate = k × [NaCl]^m

(27)

where m = the reaction order, k = rate constant. The [Rate] according to the Rate law relates the rate of reaction to the concentration of reactants (and catalyst) raised to various powers [150]. It therefore follows that if the concentration of NaCl in the water is changed from a concentration [C₁] to a concentration [C₂] the [Rate] of desalination must change to [150]:

Rate = ([C₂]/[C₁])^m × k × [NaCl]^m

(28)

The change in ([C₂]/[C₁])^m accounts for all or part of the decline in the rate of desalination with time associated with different starting salinity concentrations.

The [Rate] can be calculated [90] as:

Rate = −Δ[NaCl]/Δt = ([C₂]/[C₁])^m × k × [NaCl]^m

(29)

The Arrhenius equation requires that [124]:

k = Ae^[−E/RT]

(30)

where A = frequency factor, E = activation energy, kJ·M⁻¹, R = gas constant, T = temperature, K. It follows that the desalination rate will vary with temperature, and E, where different ZVM combinations will have different values of E.

The desalination reaction will progress to equilibrium, where the equilibrium constant (K_c) is [18,150]

K_c = [Products]^a/[Reactants]^b
ΔG⁰ = −RT·(lnK_c)
ΔE⁰ = −RT/nF·(lnK_c)

(31)

where ΔE⁰ = standard electropotential; n = number of electrons transferred; F = Faraday constant; ΔG⁰ = Standard Gibbs Free Energy. A change in state (e.g., T, Eh and pH) changes ΔG, ΔE and K_c.

4.7.6. Relationship between EC (Salinity) Reduction and Temperature

Equations (26)–(31) demonstrate that the desalination rate is a function of temperature. Therefore operation of a reactor at elevated temperatures may increase the rate of desalination.

A trial (CSD series (Table D1, Table D2 and Table 5)) was established using a reactor containing 8 L saline water and a 0.4 L cartridge containing ZVM TP constructed from 224 g Fe⁰ + 35 g Al⁰ (Table D1 and Table D2) The partial pressure in the water was maintained within the range 0–0.1 MPa. The same ZVM TP batch was reused in each of Trials CSD1a to CSD1d. The operating parameters were:

• Trial CSD1a: The trial duration was 15 h during which time the water temperature was raised from 5 to 50 °C. Air was bubbled through the water at a rate of 55 mL·min⁻¹.
• Trial CSD1b: The trial duration was 124 h during which time the water temperature was raised from 5 to 44 °C. Air was bubbled through the water at a rate of 64 mL·min⁻¹.
• Trial CSD1c: The trial duration was 30 h during which time the water temperature was raised from 7 to 39 °C. Air was bubbled through the water at a rate of 60 mL·min⁻¹.
• Trial CSD1d: The trial duration was 11 h during which time the water temperature was raised from 7 to 36 °C. Air was bubbled through the water at a rate of 22 mL·min⁻¹ for 4.7 h followed by 15 mL·min⁻¹ 80% N₂ + 20% CO₂.

The gases were not bubbled through the ZVM TP. The trial results are summarized in Table 5. This trial established that increasing temperature:

• increases the rate of desalination;
• allows the ZVM TP to be reused to sequentially partially desalinate batches of saline water;
• allows the salinity of the product water to reduce below 0.2 gNaCl·L⁻¹.

Table 5. Feed and product water compositions: CSD Trials.

Trial	Feed Water		EC, mS·cm−1	Salinity, g·L−1	Product Water		EC, mS·cm−1	Salinity, g·L−1	NaCl Removed, g·L−1
	Eh, V	pH			Eh, V	pH
CSD1a	0.277	6.150	4.540	2.500	0.167	6.500	0.544	0.194	2.306
CSD1b	0.188	5.920	4.540	2.500	0.130	6.900	0.556	0.200	2.300
CSD1c	0.088	6.700	4.760	2.625	0.219	6.020	1.830	0.888	1.737
CSD1d	0.174	5.970	5.450	3.000	0.233	5.440	3.670	1.882	1.118

Table 5. Feed and product water compositions: CSD Trials.

Trial	Feed Water		EC, mS·cm−1	Salinity, g·L−1	Product Water		EC, mS·cm−1	Salinity, g·L−1	NaCl Removed, g·L−1
	Eh, V	pH			Eh, V	pH
CSD1a	0.277	6.150	4.540	2.500	0.167	6.500	0.544	0.194	2.306
CSD1b	0.188	5.920	4.540	2.500	0.130	6.900	0.556	0.200	2.300
CSD1c	0.088	6.700	4.760	2.625	0.219	6.020	1.830	0.888	1.737
CSD1d	0.174	5.970	5.450	3.000	0.233	5.440	3.670	1.882	1.118

The cost of reducing the average feed water salinity from 2.65 gNaCl·L⁻¹ to an average salinity of 0.791 g·L⁻¹ (CSD1a–CSD1d) (based on a Fe⁰ price of $400 t⁻¹) is $2.8 m⁻³ ($2.28–$6.3 m⁻³ at a Fe⁰ price of $325–$900 t⁻¹).

The product water from Examples CSD1a and CSD1b (Table 5) meet the EU, WHO and Indian definition of potable water. These results indicate that ZVM TP can be used to create potable water at temperatures of <50 °C over a period of <24 h. The cost of providing the potable water in CSD1a + CSD1b (based on a Fe⁰ price of $400 t⁻¹) is $5.6 m⁻³ ($4.55–$12.6 m⁻³ at a Fe⁰ price of $325–$900 t⁻¹).

Potable water definitions vary with country. Examples include:

• EU [151]: Salinity = 0.45 gNaCl·L⁻¹, EC < 2.5 mS·cm⁻¹, 6.5 < pH < 9.5;
• WHO [152,153]: Salinity ≤ 0.6 gCl⁻·L⁻¹ + less than 0.2 gNa⁺·L⁻¹, pH = no restriction [146];
• India [154]: Salinity = 0.25 gCl⁻·L⁻¹ (may be locally extended to 1 g·L⁻¹), 6.5 < pH < 8.5 (may be locally extended to pH 9.2).

Trials CSD1a and CSD1b established that a reduction in temperature from 50 to 44 °C increased the length of time required to produce potable water (<0.2 gNaCl·L⁻¹) (feed water salinity = 2.5 g·L⁻¹) from 15 to 124 h. Trial CSD1 demonstrates that operating ZVM TP (CSD1a) at a temperature in the range 50–75 °C will produce potable water within 15 h for a cost of <$2.8 m⁻³ from a small (e.g., <10 m³) or large scale (e.g., >10,000 m³) heated water tank.

Following completion of trial CSD1d, the residual water (0.4 L) in the manifold and ZVM cartridge was recovered (on recovery: Eh = 0.097 V; pH = 5.84; EC = 10.650 mS·cm⁻¹; after 6 weeks storage: Eh = 0.166 V; pH = 7.41; EC = 17.12 mS·cm⁻¹). These observations demonstrate the presence of a halocline between the main water body (EC = 3.67 mS·cm⁻¹) and the manifold and ZVM container (EC = 10.650 mS·cm⁻¹). They also demonstrate that:

• The ZVM TP modifies the salinity of the water adjacent to the ZVM TP by removing NaCl from the overlying water body.
• The removal of NaCl can occur when the water body is static (e.g., ST1a–ST5j series of trials) or when the bulk of the water body is in constant movement due to gas bubbling and the water column connecting the ZVM TP to the moving water body is static.

5. Results: ZVM TPA

The ZVM TP analyses established that the rate of desalination could be enhanced to allow partial desalination in less than 24 h if:

• The main water body was kept in a state of constant movement by an acid oxygenating (N₂:CO₂), or oxygenating (N₂:O₂) gas which was bubbled through the water;
• The ZVM TP was held in a container which was separated from the main water body by a static water column.

Earlier studies [21,25] had established that high rates of desalination occurred in a continuously stirred water column [21], or a water column which was subject to argon gas bubbling [25]. In both the earlier studies the ZVM was fluidised by the water movement in the reactor.

High rates of desalination were also observed when ZVM was mixed with ion exchange material (Ca-montmorillonite) [5] in a static water environment.

The ZVM TPA trials were designed to create a reaction environment where the main water body was kept in a state of constant movement by an acid oxygenating (N₂:CO₂), or oxygenating (N₂:O₂) gas which was bubbled through the water and the ZVM TPA was held in a container which was separated from the main water body by a static water column.

5.1. Desalination Using ZVI + Ion Exchange Material (ZVM TPA)

Experiments using mixtures of ZVM and Ca-montmorillonite (ion exchange material) [5,6] have established that the presence of the ion exchange material limits the maximum pH (attributable to ZVM) and creates a low variation between the maximum pH and the equilibrium pH. This can be reinforced by the addition of Al⁰ to the ZVM-ion exchange medium mixture [5]. The ZVM TP analysis has established that this will result in a low surface charge and a high Q_e.

Layered aluminium silicates (e.g., feldspars, feldspar derivatives (clays), and zeolites) contain charged sites (when placed in water). These sites (and associated hydration layers) allow them to attract [Fe(H₂O)₆]³⁺ ions and act as nuclei for Fe species growth.

Switching from ZVM TP to ZVM TPA allows the [ZVI-ion exchange material] combination to be used to facilitate (and accelerate) NaCl removal at ambient temperatures.

5.2. Oxygenation and CO₂

5.2.1. Oxygenated Water

In an oxygenated environment radicals which include NaO⁻, ClO⁻ (Figure C37, Appendix H) develop:

Na⁺ + 0.5O₂ + nH₂O + 2e⁻ = [(NaO)(H₂O)_n]•¹⁻
Na⁺ + 0.5O₂²⁻ + nH₂O + e⁻ = [(NaO)(H₂O)_n]•¹⁻
Na⁺ + 0.5O₂⁻ + nH₂O + 1.5e⁻ = [(NaO)(H₂O)_n]•¹⁻
Cl⁻ + (1 + n)H₂O = [(ClO)(H₂O)_n]•¹⁻ + 2H⁺ + e⁻
Cl⁻ + Na⁺ + 0.5O₂ + (1 + n + m)H₂O = [(NaO)(H₂O)_n]•¹⁻ + [(ClO)(H₂O)_m]•¹⁻ + 2H⁺

(32)

5.2.2. Acid Oxygenated Water

In the presence of CO₂ the radicals which include HCO₃⁻, CO₃²⁻, NaO⁻ and ClO⁻ (Figure C37, Appendix H) develop:

Na⁺ + CO₂ + (2 + n)H₂O = [(CO₃)(NaO)(H₂O)_n]•¹⁻ + 4H⁺ + 2e⁻
Na⁺ + CO₂ + (2 + n)H₂O = [(HCO₃)(NaO)(H₂O)_n]•¹⁻ + 3H⁺ + 2e⁻
Cl⁻ + CO₂ + (2 + n)H₂O = [(CO₃)(ClO)(H₂O)_n]•¹⁻ + 4H⁺ + 4e⁻
Cl⁻ + CO₂ + (2 + n)H₂O = [(HCO₃)(ClO)(H₂O)_n]•¹⁻ + 3H⁺ + 4e⁻
Cl⁻ + Na⁺ + 2CO₂ + (4 + n + m)H₂O = [(CO₃)(NaO)(H₂O)_n]•¹⁻ + [(CO₃)(ClO)(H₂O)_m]•¹⁻ + 8H⁺ + 6e⁻
Cl⁻ + Na⁺ + 2CO₂ + (4 + n + m)H₂O = [(HCO₃)(NaO)(H₂O)_n]•¹⁻+ [(HCO₃)(ClO)(H₂O)_m]•¹⁻+6H⁺ + 6e⁻

(33)

5.2.3. Reaction Quotient Associated with Desalination in Oxygenated or Acid Oxygenated Water

The reaction quotient, Q, is defined [19,124] as:

Q = ([NaO⁻] × [ClO⁻] × [H⁺]²)/([Na⁺] × [Cl⁻] × [P_O2]^0.5)
Q = ([NaO⁻] × [ClO⁻] × [CO₃²⁻]² × [H⁺]⁸)/([Na⁺] × [Cl⁻] × [P_CO2]²)

(34)

5.2.4. Secondary Reaction Associated with Oxygenation by Air or Oxygen

Oxygenation of water creates the secondary reactions:

0.5O₂ + H₂O + 2e⁻ = 2OH⁻
2OH⁻ = H₂O₂ + 2e⁻
OH⁻ + H⁺ = H₂O

(35)

These result in a gradual increase in pH with oxidation to an equilibrium level (Figure C18, Figure C19, Figure C20, Figure C21, Figure C22, Figure C23, Figure C24, Figure C25, Figure C26, Figure C27, Figure C28, Figure C29, Figure C30, Figure C31, Figure C32, Figure C33, Figure C34, Figure C35 and Figure C36).

5.2.5. Adverse Impact of CO₂ Demonstrated by ZVM TP

ZVM TP Trial CSD1c (Table 5) established that increasing the amount of H⁺ in the product water decreased the amount of desalination. This was confirmed by trial CSD1d (Table 5) where CO₂ is used to supply a source of H⁺ [19]:

CO₂ + H₂O = HCO₃⁻ + H⁺
CO₂ + H₂O = CO₃²⁻ + 2H⁺

(36)

This adverse effect on desalination was not observed when the trials used ZVM TPA (Figure C33, Figure C34, Figure C35 and Figure C36).

5.3. Primary Desalination Mechanism Associated with ZVM TPA

Equation (32), Figure C37, and Appendix H demonstrate that the primary desalination mechanism involves oxidation of both Na⁺ and Cl⁻ and the NaCl is removed as [(NaO)(H₂O)]•¹⁻, [(ClO)(H₂O)]•¹⁻, [(ClO–ClO)(H₂O)]•, [(ClO-ClO)(H₃O)]•¹⁺ radicals. These radicals are removed by electrostatic attraction to the ZVI [68] and ion exchange material. This allows Q_e > 1.0 [25]. It also allows a single charge of ZVM TPA to be reused multiple times at ambient temperatures.

This desalination mechanism was demonstrated using saline water (constructed using Cheshire halite (NaCl, rock salt)) and ZVM TPA, constructed from ZVM + K-feldspar (held in cartridges) using four series of trials. The four series of trials are: E143, E144, E145 and E146 (Table C5, Table C6, Table C7, Table C8, Table C9, Table C10, Table C11, Table C12 and Table C13, Table D3; Figure C18, Figure C19, Figure C20, Figure C21, Figure C22, Figure C23, Figure C24, Figure C25, Figure C26, Figure C27, Figure C28, Figure C29, Figure C30, Figure C31, Figure C32, Figure C33, Figure C34, Figure C35 and Figure C36).

Trials E143 and E144 were operated in a reactor containing 5.4 L water, with a basal ZVM TPA cartridge attached to the reactor via a manifold.

Trial E145 used a 114 L reactor containing 12 g·L⁻¹ ZVM TPA. The test results from E145 indicated that a larger reactor (Trial E146) utilizing 240 L and a lower ZVM TPA concentration (1.67 g·L⁻¹ ZVM TPA (Table C3)) would be more effective. The ZVM TPA in Trial E145 and E146 was held in a cartridge. In Trial E146, the ZVM TPA cartridge was separated from the main water column by a static water body.

5.4. Initial Trial: E143

The ZVM TPA charge (P_w = 25 gFe⁰·L⁻¹) was reused 11 times in the E143 series of trials (Table C5, Table C6, Table C7 and Table C8, Table D3). Details of the trial feed water parameters, product water parameters, salinity change, water consumption, gas charge, gas composition, trial duration, temperature, and product water stability are provided in Table C5, Table C6, Table C7 and Table C8. The treatment:

• Achieved Q_e(apparent) = 0.738 gNaCl·g⁻¹ Fe⁰; Q_{e(after adjustment for water consumption)} = 1.735 gNaCl·g⁻¹ Fe⁰;
• Processed 2.4 L water·g⁻¹ Fe⁰;
• Processed 59.4 L feed water (3.96 gNaCl·L⁻¹ average salinity) to produce 48.28 L product water (2.28 gNaCl·L⁻¹).
• Demonstrated a partial desalination cost of $0.9 m⁻³ feed water (for a Fe⁰ cost of $400 t⁻¹).

Specific observations are:

• Product water pH > feed water pH when the CO₂ content of the feed gas is 0%, otherwise Product water pH < feed water pH.
• NaCl removal decreased with increasing usage of the ZVM TPA.
• NaCl removal is associated with relatively high water consumption.
• The NaCl removal reaction is partially reversible. Some trials (Table C8) demonstrated increases in product water salinity when one, or more, of the water Eh, pH and/or temperature changed;
• The product water composition is stable and initial pH values in the range 5–6, are changed after 6 weeks to pH values in the range 6–7 and remained stable at that level thereafter.

5.4.1. Munsell Analysis

Following completion of the Trial E143k the manifold and ZVM TPA cartridge were drained separately and analyzed.

• 0.3 L (light orange colored; Munsell Parameters: hue = 5YR; chroma = 12; value = 7) water was recovered from the manifold (on recovery: Eh = 0.222 V; pH = 5.31; EC = 44.75 mS·cm⁻¹; after 4 weeks storage: Eh = 0.275 V; pH = 6.47; EC = 50.17 mS·cm⁻¹).
• 0.1 L (deep orange colored; Munsell Parameters: hue = 5YR; chroma = 8; value = 4) water was recovered from the cartridge (on recovery: Eh = 0.263 V; pH = 5.15; EC = 425.4 mS·cm⁻¹ (3.92 M NaCl·L⁻¹); after 4 weeks storage: Eh = 0.225 V; pH = 6.48; EC = 350.37 mS·cm⁻¹).

These observations confirm that NaCl is concentrated in the static water beside the ZVI.

The Munsell parameters confirm that the nano-particles are akaganeite [155]. Akaganeite nano-particles darken in color as the particle size decreases [155], i.e., the Munsell value decreases [155]. The YR hue indicates that the n-Fe(OH) crystallites adopt a needle, or rod like, morphology [155]. A 5YR hue is consistent with the hydrated akaganeite (FeOOH) crystallites containing >0.15 M Na⁺·M⁻¹ Fe⁰ [155].

5.4.2. Halocline Analysis

The average concentration of NaCl (g·L⁻¹) in the cartridge approaches the eutectic concentration of NaCl at 23.3 wt % NaCl (37.6 wt % dehydrate (NaCl:2H₂O)) [156] and may exceed the critical dehydrate concentration of 37.6 wt % [156]. The ZVM cartridge and manifold acted as a collector of NaCl. The high concentrations of NaCl indicate that direct NaCl precipitation can occur within the ZVM bed once the concentration of NaCl exceeds the eutectic concentration [156].

These observations demonstrate that the water body in the reactor can become structured during operation with separate haloclines forming between the main water body and the manifold, and the manifold/cartridge and the ZVM. Each of the three water zones identified has a separate density, with density increasing as salinity increases.

The major variable in each trial was the amount of, and type of dissolved gases (N₂, O₂, CO₂), entered into the water. The permeability of the water (k_w) to a migration location (within the water) away from the gas source decreases exponentially with increasing distance (D_w) [157]. The differences in oxygenation may also result in the development of a thermocline, where the temperature of the oxygenated layer is greater than the temperature of the anoxic layer [157]. Haloclines can be manufactured by placing less saline water on top of more saline water [158].

The observations in Table C4, Table C5, Table C6, Table C7 and Table C8 together with the development of chemoclines is consistent with a model that dimeric/ion adducts, or radicals, of the structure [(NaO)(H₂O)_n]•¹⁻, [(Na)(H₂O)_n]•¹⁺, [(Cl)(H₂O)_n]•¹⁻, [(ClO)(H₂O)_n]•¹⁻, [(ClO-OCl)(H₂O)_n]• are formed in the main water body and migrate (and are concentrated) by electrostatic attraction towards the ZVM TPA.

5.5. Long Duration Trial: E144

This trial sought to determine if increasing the treatment time using ZVM TPA (P_w = 9.4 gFe⁰·L⁻¹), resulted in increased NaCl removal. The trial results (Table C9, Table C10, Table C11 and Table C12) established that:

• Product water salinity based on EC overestimated the actual water salinity (calculated from UV-visible spectrometry) due to the formation of nano-particles in the water;
• Desalination using ZVM TPA has an optimum treatment time within the range 1 to 24 h, though the amount of NaCl removed increases with time.

The trial demonstrated:

• ZVM TPA can be reused
• 32.4 L of feed water (3.02 gNaCl·L⁻¹) produced 24.15 L of product water (1.56 gNaCl·L⁻¹)
• An achieved Q_e(apparent) of 0.932 gNaCl·g⁻¹ Fe⁰; Q_{e(after adjustment for water consumption)} = 1.25 gNaCl·g⁻¹ Fe⁰;
• Processimg of 0.63 L water·g⁻¹ Fe⁰;

5.6. Initial Upscaled Trial: E145

A trial (Figure C19) in a 114 L reactor containing 1.449 kg ZVM TPA held in 5 cartridges established an increase in pH to an equilibrium level, which was associated with a decrease in Eh. EC was found to increase to an equilibrium level. The UV-visible spectrometry established a salinity decline which was associated with nano particle formation (Figure C19).

This initial trial indicated (Figure C19) that the oxygenation level used would have been more effective in a larger water body containing less ZVM TPA.

5.7. Upscaled Trial: E146

Following completion of the trial E145, a cartridge containing 400 g ZVM TPA was removed from the reactor used in Trial E145. This cartridge was reused by being placed in a 240 L reactor (P_w = 0.52 gFe⁰·L⁻¹). This trial series, termed Trial E146, (Table C13, Figure C20, Figure C21, Figure C22, Figure C23, Figure C24, Figure C25, Figure C26, Figure C27, Figure C28, Figure C29, Figure C30, Figure C31, Figure C32, Figure C33, Figure C34, Figure C35 and Figure C36) establishes that the cartridge can be used to treat 18 batches of water (average feed water salinity = 2.66 g·L⁻¹; average product water salinity = 1.53 g·L⁻¹). The desalination cost demonstrated was $0.011 m³ (for a Fe⁰ cost of $400 t⁻¹). The 18th usage of the cartridge removed 3.6 g·L⁻¹ (Table C13). The trial established that the reduced ZVM TPA concentration reduces the total nano-particle discharge into the water.

All the trials were undertaken using an air feed gas. Some of the trials were operated with a parallel trial using the same feed water (in a 5.4 L reactor), the Trial E144 ZVM TPA cartridge, and a feed gas containing 20% CO₂ + 80% N₂ (Figure C33, Figure C34, Figure C35 and Figure C36). These trials demonstrated lower rates of desalination than the air fed trials. The lower rates of desalination were associated with lower pH values and higher Eh values (Figure C33, Figure C34, Figure C35 and Figure C36).

In each example, using an air feed (Figure C20, Figure C21, Figure C22, Figure C23, Figure C24, Figure C25, Figure C26, Figure C27, Figure C28, Figure C29, Figure C30, Figure C31, Figure C32, Figure C33, Figure C34, Figure C35 and Figure C36), the product water pH is 1 to 1.5 units higher than the feed water, while the product water Eh is lower than the feed water Eh. The product water EC is higher than the feed water EC due to the formation of nano-particles. UV-visible spectrometry was used to assess the associated decline in salinity (Figure C20, Figure C21, Figure C22, Figure C23, Figure C24, Figure C25, Figure C26, Figure C27, Figure C28, Figure C29, Figure C30, Figure C31, Figure C32, Figure C33, Figure C34, Figure C35 and Figure C36).

A cross plot of feed water salinity vs. product water salinity (Figure 20), indicates that NaCl removal is a function of feed water salinity, where the total amount of NaCl removed increases with increasing feed water salinity. Removal rates are within the range 0.2–1 gNaCl·L⁻¹·h⁻¹.

Figure 20. ZVM TPA example for the Trial Series E146a to E146q (Table C13). (a) Relationship between feed water salinity and product water salinity; (b) Relationship between feed water salinity and the rate of salinity removal.

Figure 20. ZVM TPA example for the Trial Series E146a to E146q (Table C13). (a) Relationship between feed water salinity and product water salinity; (b) Relationship between feed water salinity and the rate of salinity removal.

Trials E145 and E146 confirmed the earlier observation [20] that placement of Fe⁰ powders in a circulating, oxygenated, water body will result in both an increase in EC and a decrease in salinity.

5.8. Cost of ZVM TPA Desalination

The economics of partial desalination using ZVM TPA is a function of both P_w and the number of times a specific charge can be reused. The number of times a specific charge can be reused will be a function of the feed water composition. Figure C19, Figure C20, Figure C21, Figure C22, Figure C23, Figure C24, Figure C25, Figure C26, Figure C27, Figure C28, Figure C29, Figure C30, Figure C31, Figure C32, Figure C33, Figure C34, Figure C35 and Figure C36 (Table C13) demonstrate that a single charge of 0.4 kg can be reused 18 times to partially desalinate >4.0 m³ (i.e., >10,500 m³·t⁻¹ (ZVM + ion exchange material)). The ZVM TPA cost will vary with location, volume and commodity prices.

K-feldspar prices (free on board (FOB)) range from $45–$300 t⁻¹ [159,160,161]. The iron powder prices range from $325–$900 t⁻¹ FOB. The ZVM TPA (Trial E146) material costs fall within the range $132–$1020 t⁻¹ FOB, i.e., the partial desalination cost range demonstrated by Trial E146 (Table C13) is between $0.009 and $0.026 m⁻³.

Trial E146 (Table C13) demonstrated that a single batch of 1 t ZVM TPA could be used (with reuse) to partially desalinate 10,500 m³ water. Additional costs include transport, loading/unloading, ZVM TPA disposal, water tanks, air compression and monitoring costs.

The ZVM TPA is designed for multiple reuse. The ZVM TP analyses established that a single use of a ZVM TP charge could result in high levels of desalination over a long time period. ZVM TPA Trials E144 and E145 (Figure 19, Figure 20 and Figure 21, Table C13) establish that a more appropriate application for ZVM TPA is as a multiple use desalination agent.

The decreases in salinity, e.g., 40%–55% (Figure 21), indicate that the ZVM TPA can be used to partially desalinate irrigation water, or livestock feed water, within a 24 h period.

Figure 21. Partial desalination of water using ZVM TPA: E146, 0.4 kg ZVM TPA; 240 L water. (a) Regression relationship between feed water salinity and product water salinity; (b) Regression relationship between the proportion of the salinity removed and the feed water salinity; (c) Feed and product water salinities vs. number of times a batch of ZVM TPA is reused; (d) Amount of NaCl removed vs. number of times a batch of ZVM TPA is reused.

Figure 21. Partial desalination of water using ZVM TPA: E146, 0.4 kg ZVM TPA; 240 L water. (a) Regression relationship between feed water salinity and product water salinity; (b) Regression relationship between the proportion of the salinity removed and the feed water salinity; (c) Feed and product water salinities vs. number of times a batch of ZVM TPA is reused; (d) Amount of NaCl removed vs. number of times a batch of ZVM TPA is reused.

The minimum cost of partially desalinated water used for irrigation on the basis of an irrigation requirement of 1000 m³ water ha⁻¹·a⁻¹ is $12–$98 m³·ha⁻¹·a⁻¹. This compares with about $3000 m³·ha⁻¹·a⁻¹ if the desalinated water is produced by a large scale (>100,000 m³·d⁻¹) desalination plant (reverse osmosis, or multi-stage flash distillation) at a sale price of $3 m⁻³ [36,37,38].

6. Cations and Anions (Pollutants) Removed during Desalination by ZVM TP and ZVM TPA

ZVM (Fe⁰, Cu⁰, Al⁰) (and ZVM TP, ZVM TPA) will remove a variety of pollutants and microbial organisms from water (Appendix A). Removal is achieved by either direct precipitation [18,19], or by incorporation into ZVM corrosion products [18], or by reduction of a pollutant to a lower oxidation state (e.g., many organic anions) [18], or by oxidation of a pollutant to a higher oxidation state [18]. A detailed discussion of the mechanisms involved in these processes is beyond the scope of this study, but is addressed elsewhere [18,19,162]. These pollutants are not removed by the ZVM, but are removed by the effect the ZVM has on the surrounding water. These effects are the increase in availability of: (i) e⁻ to facilitate/catalyze electron shuttle reactions [18,19,162]; (ii) H⁺_(ads) to increase the number of potential anion removal sites within the water; (iii) OH⁻ to increase the number of potential cation removal sites within the water; (iv) adjustment of water Eh and pH to change the stable equilibrium form of a specific pollutant species.

Appendix A summarises the microbiota, cations and anions that can be removed by Eh and pH changes in the water which result from the presence of ZVM [18,19,162].

This study has demonstrated that ZVM TP and ZVM TPA create the same Eh and pH changes in water as ZVM [5,6]. Therefore it is reasonable to assume that similar pollutant removal mechanisms will apply to all three groups of material.

6.1. Cations and Anions (Pollutants) Removed during Desalination Using ZVM TP

Placement of ZVM TP in water results in the removal of a variety of cations and anions (Table C1, Table C2, Table C3 and Table C4). The proportion removed is a function of their original concentration, the composition of the water, treatment time, temperature and the ZVM TP composition.

The cations whose concentrations were reduced include: As, B, Ba, Ca, Co, Cu, Mg, Mn, Na, Ni, P, S, Si, Sr (Table C2, Table C3 and Table C4). The anions whose concentrations were reduced include: Cl, F, NO₃, NO₂, PO₄, SO₄ (Table C1). The concentrations of the cations and anions (apart from Na, Cl) in the water samples were low.

This ion removal data (Table C1, Table C2, Table C3 and Table C4) is cross plotted against NaCl removal in Figure C18. With the exception of sulphur (and sulphate) whose removal decreases with increasing NaCl removal, and manganese whose removal increases with increasing NaCl removal, there appears to be no direct correlation between cation/anion removal and NaCl removal. In all other cases, high levels of ion removal are associated with both low and high rates of NaCl removal.

The inter-relationship between desalination and the removal of other cations and anions in the water has not been specifically analysed in this study. The presence of high concentrations of other cations and anions (e.g., 1–400 g·L⁻¹) may adversely impact on the rate of desalination.

6.2. Impact of Sulphates: ZVM TP and ZVM TPA

Sulphates can be present in chloride rich groundwater containing Fe, Na and other cations. In the presence of ZVI, sulphate interacts with the green rust product to preferentially displace Cl from the ion lattices (at low pH). The sulphate rich green rust product is termed Green Rust 2 (GR2) [163]. In the presence of sulphates the dominant FeOOH species is goethite [96,163]. The initial molar ratio, R, of SO₄²⁻/OH⁻, controls the formation of GR2 [163]. GR2 has only been observed when R > 0.5 [163]. In an environment containing both Cl⁻ and SO₄²⁻ ions, the overall reaction sequence is an initial precipitation of Fe(OH)₂. This then transforms to GR1. The Cl⁻ ions are then replaced by SO₄²⁻ ions to form GR2 and can be subsequently oxidised to form lepidocrocite (γ-FeOOH) [164].

The boundary between GR1 and GR2 is defined by the relative concentration of Cl⁻ and SO₄²⁻ ions where Eh_GR2 = −0.507 + 0.4137 log [Cl⁻] − 0.2364 log [SO₄²⁻] [164]. The initial transition from GR1 to GR2 is made by the replacement of Cl⁻ by SO₃²⁻ [165]. This maintains the GR1 structure [165]. The SO₃²⁻ ions are replaced by SO₄²⁻ ions as the GR2 structure forms [165].

This results in the effective desalination rate decreasing as the effective sulphate removal rate increases. This situation is demonstrated by the ZVM TP analyses (Figure C18) and indicates that the desalination process may be more effective in either desulphurized water or sulphurous water where the initial molar ratio R < 0.5 [163] and Eh < Eh_GR2.

The transformation of GR2 to lepidocrocite is controlled by the relative concentration of SO₄²⁻, where Eh = 0.55 + 0.0148 log [SO₄²⁻] − 0.089 pH [166]. The relative position of this redox boundary is shown in Figure 22. The UV-visible spectral peaks [101] associated with each Green Rust (GR) species (GR1, GR2) are well defined (Table 3). These spectra allow the dominant entrained Fe corrosion species in the water to be identified (Table 3).

Figure 22. Pourbaix diagram showing the Eh, pH regions where the dominant stable precipitate is Fe(OH)₂, GR1, and GR2 in water containing both Cl⁻ and SO₄²⁻ ions.

Figure 22. Pourbaix diagram showing the Eh, pH regions where the dominant stable precipitate is Fe(OH)₂, GR1, and GR2 in water containing both Cl⁻ and SO₄²⁻ ions.

The FeOOH chemistry is complex. In the absence of Cl⁻ ions, lepidocrocite forms [166]. In the absence of SO₄²⁻ ions, akaganeite forms, e.g., [167]. Lepidocrocite is replaced by goethite as the SO₄²⁻:Cl⁻ ratio increases at the transition from GR2 to FeOOH [167]. Akaganeite which forms at the GR1:FeOOH boundary is replaced by goethite as the SO₄²⁻:Cl⁻ ratio increases [167].

The presence of high concentrations of SO₄²⁻ ions will reverse any Cl⁻ desalination (by incorporation) which is associated with GR1 or akaganeite. Na⁺ ions are incorporated with green rusts and the transformation from GR1 to GR2 is unlikely to expel the Na⁺ ions [168]. Both Na⁺ and SO₄²⁻ ions are released as the GR2 transforms to goethite [168].

6.3. Impact of Sulphides: ZVM TP and ZVM TPA

In the presence of water, Fe₂O₃ (spp.) catalyse the transformation of HS⁻ ions to elemental sulphur (Equation (21)). HS⁻ is stable under reaction conditions [19] which have a significantly lower Eh than those observed in this study (Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16 and Figure C17). The lower Eh (without additional oxidation) is expected to adversely impact on the rate of formation of green rust and FeOOH species. HS⁻ (without oxidation) can result in the formation of Fe_1+xS, Fe_1−xS, Fe₃S₄, Fe_3+xS₄, FeS₂, The main iron sulphide precipitation zone is within the green rust/Fe(OH)₂ precipitation zone, and can result in reassignment of some of the ZVM from desalination to sulphide removal (Figure 23).

Figure 23. Pourbaix diagram showing the Eh, pH regions where the dominant stable precipitate is FeS₂, Fe_1+xS. Data: [169,170].

Figure 23. Pourbaix diagram showing the Eh, pH regions where the dominant stable precipitate is FeS₂, Fe_1+xS. Data: [169,170].

6.4. Impact of Bicarbonates: ZVM TP and ZVM TPA

Bicarbonates (HCO₃⁻) are a common constituent of most groundwater. In volcanic areas (such as those used to provide the base water in this study), the dissolved organic content is low (e.g., 1–18 mg·L⁻¹) [171]. Natural bicarbonate concentrations in this water are in the range 10–150 mg·L⁻¹.

6.4.1. Trial E146: ZVM TPA

The ZVM TPA Trials, E146, established that increasing the HCO₃⁻ content of the water (with associated acidification) either had (i) no effect on the rate of desalination, e.g., Figure C33; or (ii) reduced the amount of NaCl removed by 30%–50% (Figure C34, Figure C35 and Figure C36).

6.4.2. Trials PS15 and PS16: ZVM TP

Two ZVM TP trials (PS15, PS16) were undertaken, using a reactor containing 3.5 L water (Table C1, Table C2, Table C3 and Table C4). The reaction environment in Trial PS15 was saturated with [16.79% CH₄ + 16.88% H₂ + 11.97% CO + 8.33% CO₂ + 46.03% N₂ for 79 days]. The reaction environment in Trial PS16 was saturated with [16.79% CH₄ + 16.88% H₂ + 11.97% CO + 8.33% CO₂ + 46.03% N₂ for 29 days] followed by [80% N₂ + 20% CO₂ for 135 days] followed by [100% N₂ for 46 days].

The associated Eh, pH, and EC for the feed and product waters establish a general rise in pH (Table 6), indicating that the dominant carbonate species is a mixture of HCO₃⁻ and CO₃²⁻ [19]. Both examples PS15 and PS16 demonstrated a substantial removal of NaCl (Table 6, Table C1, Table C2, Table C3 and Table C4).

Trial PS15 removed 9.51 gCl⁻·L⁻¹ + 1.29 gK⁺·L⁻¹ + 6.23 gNa⁺·L⁻¹ over 79 days. The normalized molar removal ratio was: 1 Cl⁻: 1.26 Na⁺:0.54 K⁺. The normalised removal ratio is calculated as (K or Na Removed/K or Na in Feed)/(Cl Removed/Cl in Feed). Trial PS 16 removed 23.07 gCl⁻·L⁻¹ + 15.35 gNa⁺·L⁻¹ over 121 days. The normalized molar removal ratio was: 1 Cl⁻:1 Na⁺.

Trial PS16 reused the ZVM TP used in Trial PS15 and increased the amount of ZVM TP present (Table 6). These trials demonstrated:

• ZVM TP can be reused if it is held in an oxygen free environment.
• In an oxygen free environment, ZVM TP can increase the desalination rate [Constant D] to a value in the range 0.0128–0.0462. This compares with a [Constant D] of 0.004–0.011 using ZVM TP from the same batch in an oxygenated environment.
• Desalination using ZVM TP can occur when the water contains bicarbonate or carbonate ions.

Table 6. PS15, PS16. Particulate ZVM TP (Trial Group 1, Table D1 and Table D2); (i) PS15: P_w = 134 g·L⁻¹; Constant D = 0.0128; (ii) PS16: P_w = 208 g·L⁻¹. Constant D = 0.0462; Operating temperature = −8 to 25 °C (atmospheric temperature); Pressure = 0.1 MPa.

Water	Feed and Recovered Product Waters	pH	EC, mS·cm−1	Storage Period, Years	Product Water Following Storage	pH	EC, mS·cm−1	HCO3−, g·L−1
	Eh, V				Eh, V
PS15
Feed	0.207	6.14	57.49	-	-	-	-	-
Product	0.094	9.49	31.12	1.11	0.220	8.9	35.70	>10
PS16
Feed	0.207	6.14	78.16	-	-	-	-	-
Product	−0.111	10.26	1.09	0.58	0.109	9.18	1.25	1.14

Table 6. PS15, PS16. Particulate ZVM TP (Trial Group 1, Table D1 and Table D2); (i) PS15: P_w = 134 g·L⁻¹; Constant D = 0.0128; (ii) PS16: P_w = 208 g·L⁻¹. Constant D = 0.0462; Operating temperature = −8 to 25 °C (atmospheric temperature); Pressure = 0.1 MPa.

Water	Feed and Recovered Product Waters	pH	EC, mS·cm−1	Storage Period, Years	Product Water Following Storage	pH	EC, mS·cm−1	HCO3−, g·L−1
	Eh, V				Eh, V
PS15
Feed	0.207	6.14	57.49	-	-	-	-	-
Product	0.094	9.49	31.12	1.11	0.220	8.9	35.70	>10
PS16
Feed	0.207	6.14	78.16	-	-	-	-	-
Product	−0.111	10.26	1.09	0.58	0.109	9.18	1.25	1.14

6.4.3. Desalination in an Alkaline Bicarbonate Saturated Anoxic Environment: ZVM TP

In this anoxic environment the ZVM TP (PS15, PS16) will corrode to have a Fe₃O₄ interface between the Fe⁰ and the Fe^IIFe^III corrosion layers [172]. High rates of desalination occur (Table 6). The Fe⁰ corrodes directly to form a GR1 structure containing HCO₃⁻ [73,172]. In the presence of sulphate, the GR1 transforms to GR2 [172,173]. The redox boundary between GR1 (CO₃²⁻) and GR2 (SO₄²⁻) approximates to the redox boundary between GR1 (Cl⁻) and GR2 (SO₄²⁻) (Figure 18). When sulphate is absent, the dominant green rust is green rust carbonate containing some Cl⁻ ions [173]. GR1 (Cl⁻) is absent, even when the Cl⁻:HCO₃⁻ ratio is >40 [73]. The presence of carbonate-bicarbonate increases the rate of cathodic and anodic reactions and accelerates the rate of corrosion [174]. High corrosion rates are maintained by the interaction of Fe²⁺ species with HCO₃⁻ and the reduction of protons supplied by HCO₃⁻ ions [174]. A rapid decrease in corrosion rates is observed when the pH is decreased to move the reaction environment away from the CO₃²⁻:HCO₃⁻ buffer zone [174]. When the water contains no Cl⁻ ions the GR1 (CO₃²⁻) will oxidize to the FeOOH species, goethite [73].

6.4.4. Impact of CaCO₃ or MgCO₃ in the Water: ZVM TP and ZVM TPA

The presence of CaCO₃ or MgCO₃ in the water (i.e., a Ca²⁺ − HCO₃⁻, or Mg²⁺ − HCO₃⁻ groundwater type) will result in the formation of GR1 (CO₃²⁻) [172]. Trials PS15 and PS16 demonstrate that ZVM is able to desalinate a mixed ground water type containing [HCO₃⁻ + Na⁺ + K⁺ + Cl⁻] or [HCO₃⁻ + Na⁺ + Cl⁻].

6.4.5. Impact of Chlorinated Organic Compounds in the Water: ZVM TP and ZVM TPA

TCE (trichloroethylene) removal experiments in water containing Ca²⁺ − HCO₃⁻, or Mg²⁺ − HCO₃⁻, or Na⁺ − Cl⁻ have established that passivation of the ZVI surface by GR1 (CO₃²⁻) reduces the ability of ZVI to dechlorinate hydrocarbons [175]. This may indicate that the dechlorination process involves a GR1 (Cl⁻) intermediate. In water containing high concentrations of NaCl, hydrocarbon dechlorination rates are inhibited due to competition for available sites by Cl⁻ ions [175].

6.4.6. Bicarbonate/Carbonate Buffers: ZVM TP and ZVM TPA

The carbonate environment contains an acidic (H₂CO₃:HCO₃⁻ or CO₂:HCO₃⁻) buffer, and an alkaline (HCO₃⁻:CO₃²⁻ or CO₂:CO₃²⁻) buffer [19].

The pH in the proximity of the alkaline buffer is a function of the ratio of log (CO₃²⁻/HCO₃⁻) or log (CO₃²⁻/pCO₂)

Examples PS15 and PS16 (ZVM TP) were operated in the vicinity of the alkaline buffer, where: Observed pH = 10.34 + log (CO₃²⁻/HCO₃⁻) [19], or in the presence of CO₂, where: Observed pH = (18.14 + log (CO₃²⁻/pCO₂))/2 [19].

The examples in Trials CSD1 and E146 (ZVM TP and ZVM TPA) were operated in the vicinity of the acidic buffer, where: Observed pH = 6.38 + log (HCO₃⁻/H₂CO₃) [19], or in the presence of CO₂, where: Observed pH = (7.81 + log (HCO₃⁻/pCO₂) [19].

The example MT1d (Figure C13) investigated the effect of acidified carbonate (formic acid + calcium carbonate) on desalination. This example demonstrated that first placing the water in the H₂CO₃ regime and then using CaCO₃ to increase the pH to move the water across the H₂CO₃:HCO₃⁻ buffer and into the HCO₃⁻ buffer resulted in a progressive increase in pH. As the pH increased above 7 and the Eh increased to >0.1 V, the EC demonstrated a rapid drop of >5 mS·cm⁻¹ as dissolved species, Feⁿ⁺, Ca²⁺, Na⁺, Cl⁻ were removed.

6.4.7. Acid Bicarbonate Buffer: ZVM TP and ZVM TPA

Example analyses at the acid bicarbonate buffer demonstrated (Trials CSD1d, E143, E144, E146) a very low HCO₃⁻/pCO₂ ratio which was accompanied by partial desalination of the water, e.g.,

Trial CSD1d demonstrated:

• Table 5: HCO₃⁻/pCO₂ ratio = 0.004266:1; HCO₃⁻/H₂CO₃ ratio = 0.1022; Feed water = 3 g·L⁻¹; Product water = 1.88 g·L⁻¹. Desalination time required = 16 h;

Trial E143 demonstrated (Table C5, Table C6, Table C7 and Table C8):

• E143b: HCO₃⁻/pCO₂ ratio = 0.1023:1; HCO₃⁻/H₂CO₃ ratio = 2.74; Feed water = 4.44 g·L⁻¹; Product water = 1.67 g·L⁻¹. Desalination time required = 16.5 h;
• E143c: HCO₃⁻/pCO₂ ratio = 0.0109:1; HCO₃⁻/H₂CO₃ ratio = 0.299; Feed water = 4.81 g·L⁻¹; Product water = 2.11 g·L⁻¹. Desalination time required = 15 h;
• E143d: HCO₃⁻/pCO₂ ratio = 0.0135:1; HCO₃⁻/H₂CO₃ ratio = 0.343; Feed water = 5.37 g·L⁻¹; Product water = 2.99 g·L⁻¹. Desalination time required = 15.5 h;
• E143e: HCO₃⁻/pCO₂ ratio = 0.0229:1; HCO₃⁻/H₂CO₃ ratio = 0.594; Feed water = 6.48 g·L⁻¹; Product water = 4.95 g·L⁻¹. Desalination time required = 15.5 h;

Trial E146 demonstrated:

• 13th Reuse (Figure C33): HCO₃⁻/pCO₂ ratio = 0.00195:1; HCO₃⁻/H₂CO₃ ratio = 0.063; Feed water = 1 g·L⁻¹; Product water = 0.35 g·L⁻¹. Desalination time required ≤ 3 h;
• 14th Reuse (Figure C34): HCO₃⁻/pCO₂ ratio = 0.00219:1; HCO₃⁻/H₂CO₃ ratio = 0.072; Feed water = 1.1 g·L⁻¹; Product water = 0.7 g·L⁻¹. Desalination time required ≤ 4 h;
• 15th Reuse (Figure C35): HCO₃⁻/pCO₂ ratio = 0.00275:1; Feed water = 2.05 g·L⁻¹; Product water = 1.6 g·L⁻¹. Desalination time required ≤ 9 h;
• 16th Reuse (Figure C36): HCO₃⁻/pCO₂ ratio = 0.00245:1; Feed water = 5.83 g·L⁻¹; Product water = 1.8 g·L⁻¹. Desalination time required ≤ 17 h;

Trial E143i (Table C5, Table C6, Table C7 and Table C8) established that if the ZVM TP had previously been operated in an environment containing a high HCO₃⁻/pCO₂ ratio (0.524:1) which is then reduced to (0.0087:1), this change is associated with a discharge of previously removed NaCl into the product water.E144 established (Table C9, Table C10, Table C11 and Table C12) that desalination ocurred at the acid bicarbonate buffer.

In each example the amount of NaCl removed was in the range 0.2–5 g·L⁻¹. The E146 control trials (Figure C33, Figure C34, Figure C35 and Figure C36) where the trial was operated using 80% N₂ + 20% CO₂, and repeated using air (78% N₂ + 21% O₂) demonstrated similar levels of desalination in each example. This indicates that the type of oxygenating gas does not markedly affect the rate of desalination over the pH range 5–8.

6.4.8. Alkali Bicarbonate Buffer: ZVM TP

Two trials (PS15, PS16) were operated at the alkali bicarbonate buffer (Table 6). Trial PS15 was operated with a (CO₃²⁻/pCO₂) ratio of 6.92:1 and Trial PS16 was operated with a (CO₃²⁻/pCO₂) ratio of 239.88:1. Trial PS16 demonstrates that high NaCl removal rates (Table 6) can be associated with a high (CO₃²⁻/pCO₂). This example (Trial PS16) demonstrated a net reduction in salinity from 38.6 to 0.55 gNaCl·L⁻¹ product water over 210 days (Table 1, Table C1 and Table C2) which was associated with a ZVM desalination cost in the range $67.6–$187.2 m⁻³ (based on a ZVM cost of $325–$900 t⁻¹) + gas cost + reactor costs. Trial PS16 demonstrated a gross reduction in salinity (after consideration of water consumption during desalination) from 38.6 to 0.15 gNaCl·L⁻¹ of feed water (Table 1). Trial PS16 demonstrates 8.07 gSO₄²⁻·L⁻¹ removal (Table C1) and indicates that desalination of water with a high g SO₄²⁻·L⁻¹ content is possible if the water has a (CO₃²⁻/pCO₂) ratio of 239.88:1, or (CO₃²⁻/HCO₃⁻) ratio of 0.88:1 (i.e., the reactor is operated to ensure that GR2 is not an intermediate reaction product).

6.5. Desalination of Water Containing Sulphates and Bicarbonates: ZVM TP and ZVM TPA

GR2 (SO₄²⁻) can template on an FeOOH corrosion surface [176]. Its formation requires [176] the presence of an excess of OH⁻ ions in the water, where Fe(OH)₂ sheets grow parallel to the FeOOH or Fe⁰ substrate, trapping anions within the accreting layer (e.g., Figure F1). The water composition can be defined in terms of a ratio R, where R = [OH⁻]/[Feⁿ⁺], and in terms of the ratio of Fe³⁺/Feⁿ⁺ ions (Figure 20) [176]. The initial accreting product (Fe(OH)₂) will, under stable anoxic conditions, as the [Fe³⁺/Feⁿ⁺] ratio increases start to accrete green rust (Figure F1) [176]. As the [Fe³⁺/Feⁿ⁺] ratio continues to increase these green rusts will either form FeOOH species (Figure F1), or will be reduced (by adsorbed H⁺ and released H₂ (g)) to form magnetite (Fe₃O₄) (e.g., Figure F1) [176]. Entrained precipitates formed under these conditions will be dominantly FeOOH sp. (e.g., Figure F1) [176]. This sequence of progressive oxidation is summarized in Figure 24.

The initial FeOOH species can be lepidocrocite which is transformed into goethite and then subsequently akaganeite. Goethite and lepidocrocite are the dominant species when Cl⁻ ions are absent or a relatively high level of SO₄²⁻ ions are present [176].

While the exact ratio of R required for each phase may be water composition dependent, the general reaction scheme remains the same (Figure 24), where the end points associated with green rust formation are either (Fe₃O₄) resulting from reduction, or FeOOH formed from continued oxidation [176]. The presence of the initial (Fe(OH)₂), green rust intermediate and the FeOOH and (Fe₃O₄) end points is illustrated in Figure F1.

The trial results have demonstrated that desalination will progress if the water contains Cl⁻, or [Cl⁻ + HCO₃⁻, CO₃²⁻, CO₂]. Sulphates will expel both Cl⁻ and CO₃²⁻ ions from the green rust structure and will change the final oxidized product to goethite or lepidocrocite. This change is associated with an inverse relationship where the rate of desalination decreases as the sulphate concentration of the water increases (Figure C18).

Figure 24. Relationship between Ratio R, and Fe³⁺ content surrounding ZVI and the nature of the corrosion products growing in the waters and accreting onto the ZVI. Data: [176].

Figure 24. Relationship between Ratio R, and Fe³⁺ content surrounding ZVI and the nature of the corrosion products growing in the waters and accreting onto the ZVI. Data: [176].

This water can be desalinated if an absorbent is used to first remove the sulphur. The residual salinated (desulphurized) water can then be desalinated using ZVM TP. Alternatively, Trial PS16 demonstrates that if the desalination process is operated in a pressured, reducing acid environment, both SO₄²⁻ and NaCl removal will occur, provided the reactor is operated with a pH around the alkaline bicarbonate buffer.

7. Desalination Economics

The trials have demonstrated desalination costs. For example water with a salinity of:

(1)

2.65 gNaCl·L⁻¹ can be desalinated to:

(a)

ZVM TPA—multiple use:

1.55 g·L⁻¹ (over 1–24 h, Trial E146) and 6 gNaCl·L⁻¹ reduced (Trial E146) to 3.35 gNaCl·L⁻¹) for a cost of less than $0.009–$0.026 m⁻³. This cost is based on the demonstrated 18 times reuse of the same ZVM TPA charge. The number of times a ZVM TPA charge can be reused needs to be demonstrated in field operation, but may be in the range 20–100 times. This would allow the desalination cost to fall within the range $0.002–$0.024 m⁻³.

(b)

ZVM TP—multiple use:

0.791 g·L⁻¹ (over 11–124 h, Trials CSD1a–CSD1d) for a cost of $2.28–$6.3 m⁻³. This cost is further reduced if the used ZVM TP is sold for sulphur removal, or is reprocessed for reuse. The number of times a ZVM TP charge can be reused needs to be demonstrated in field operation, but may be in the range 5–20 times. This allows the desalination cost to fall within the range $0.4–$6.1 m⁻³. ZVM TP can be reused for sulphur removal from natural gas. This reuse application can result in a residual ZVM TP value which is greater than the cost of ZVM TP purchase and use as a desalination agent.

(c)

ZVM TP—single use: 0.2 g·L⁻¹ (over <15 h, Trial CSD1) for a cost of $4.55–$12.6 m⁻³; Reuse of the ZVM TP for sulphur removal can reduce the cost of producing potable water to <$0.0 m⁻³, depending on the resale value realised for the used ZVM TP.

(2)

ZVM TP—single use: 11.05 gNaCl·L⁻¹ can be desalinated to 3.34 g·L⁻¹ (over 126 days, Trial ST3b) for a cost of $5.6–$22.48 m⁻³. Reuse of the ZVM TP for sulphur removal can potentially reduce the cost of producing partially desalinated water to <$0.0 m⁻³, depending on the resale value realised for the used ZVM TP.

(3)

ZVM TP—single use: 38.6 gNaCl·L⁻¹ can be desalinated to 0.55 g·L⁻¹ (over 210 days, Trial PS16) for a cost of $67.6–$187.2 m⁻³. Reuse of the ZVM TP for sulphur removal can potentially reduce the cost of producing partially desalinated water to <$3.0 m⁻³, depending on the resale value realised for the used ZVM TP.

These illustrative examples of costs demonstrated by the trials identify a number of potential market applications for the ZVM TP and ZVM TPA. They include:

• Partial desalination of irrigation water for arable crops;
• Partial desalination of livestock feed water;
• Partial desalination of water for municipal activities and grey water applications;
• Emergency water for people and livestock in areas affected by natural or anthropogenic disasters;
• Desalination of saline water in impoundments (including flowback water from shale gas wells and reject brine from RO desalination plants);

The economics of desalination is impacted by: (i) the number of times a specific batch of ZVM TP/ZVM TPA can be reused; and (ii) the residual (resale or reuse) value of the ZVM TP/ZVM TPA.

ZVM TPA can potentially be used for the rapid partial desalination of large volumes of water in most areas (for agriculture, municipal applications and greywater applications) and allows partially desalinated water to be delivered for a cost which is in the range $0.002–$0.026 m⁻³.

ZVM TP can fulfil a similar role for a similar or lower price in areas where there is high value added reuse market for the used ZVM TP as a gas desulphurization agent (e.g., onshore oil and gas fields in the USA). The ZVM TP can have a role as a single use agent, or a multiple use agent.

ZVM TP can fulfil a role as a single use agent to produce desalinated or partially desalinated water from a wide range of saline water (2–39 g·L⁻¹). It is cost competitive with technology such as reverse osmosis (RO) in areas which are remote from infrastructure (e.g., disaster relief areas), or where the water is heavily fouled with other components, or where the water salinity is outside the applicable range for an RO plant.

7.1. Irrigation Market

UN Water estimates that globally, there are more than 16,000 desalination plants producing 70 million·m³·d⁻¹ (25.5 Bm³·a⁻¹) [177]. These plants consume 75.2 TWh (0.5% of global electricity consumption [177]). The basic cost of desalination (excluding extraction, pre-treatment, post desalination treatment, storage, distribution and wastewater disposal) is estimated for large scale (e.g., >100,000 m³·d⁻¹) plants at $0.5–2.5 m³ for seawater desalination, and $0.2–2 m³ for brackish water desalination [178,179,180,181,182]. Small scale (<100 m³·d⁻¹) desalination costs are estimated at $1 to more than 12 m³ [183].

An indication of the full cycle costs associated with large scale desalination is provided by the volume supply costs to industrial users in countries which rely on desalinated water for the bulk of their water supply (e.g., United Arab Emirates (UAE)). In the UAE, the tariff [184] for the supply of more than 20,001 imperial gallons (277.8 m³) month⁻¹ is (April 2015) 4.6 fils/imperial gallon ($3.45 m³ (based on an exchange rate of 1 Dirham = US$0.27)). This indicates that actual cost of delivered water is in the range 3 to 7 times the basic desalination cost associated with the desalination plant. The supply cost of desalinated water to an agricultural holding may therefore fall within the range $3.5–$90 m³.

Desalinated water is currently used for agriculture in Spain (308,000 m³·d⁻¹), Kuwait (130,000 m³·d⁻¹), Italy (986 m³·d⁻¹), and Bahrain (248 m³·d⁻¹) [185]. Small quantities are also used in Qatar, and the USA [159]. Other countries such as Australia, Chile, and China are evaluating the feasibility of using water from desalination plants for agriculture [185].

In Australia, the sustainable potable groundwater resurce is about 10.2 Gm³·a⁻¹, and the sustainable saline groundwater resource is about 15.5 Gm³·a⁻¹ [185]. Current Australian agricultural water usage is 7.55 Gm³·a⁻¹ [185]. This saline water resource is likely to be subeconomic for agricultural usage if the cost of extraction, desalination and delivery is >US$0.96 [185]. The desalination costs established by Trial E146, indicate that the cost of partially desalinating the sustainable saline groundwater resource of 15.5 Gm³·a⁻¹ using ZVM TPA will be in the range $0.009–$0.026 m⁻³ ($140–$400 million·a⁻¹ + extraction, tank, and distribution costs). This compares with the cost of water at the Melbourne desalination plant of $2.18 m⁻³ + distribution and storage costs [185].

It is estimated that four crops account for >40% of global agricultural water usage (rice = 21.3%; wheat = 12.4%; maize = 8.6%; soybeans = 4.6%) [181]. Other major consumers of irrigated water include cotton and grazing pasture [185]. Existing usage of desalinated water for agriculture has focussed on high value added crops which can sustain a relatively high water supply cost. High value crops include grape, fruit trees, nut trees, berry fruits, and vegetables (e.g., tomatoes, pepper, etc.) have lower irrigation requirements [185].

Ratios Used to Assess Irrigation Water Quality

Standard guidelines for water suitability for irrigation focus on a number of parameters including EC, TDS (total dissolved solids), SAR (Na⁺/((Ca²⁺ + Mg²⁺)/2)^0.5), sodium percentage (100(Na + K))/(Ca + Mg + K + Na), Wilcox plot (sodium percentage vs. EC), Kelly’s ratio (Na/(Ca + Mg)), magnesium adsorption ratio ((100 Mg)/(Ca + Mg)), Doneen’s permeability index (100(Na + ((HCO₃)^0.5))/(Ca + Mg + Na)), chloroalkaline index [CAI-I = (Cl − (Na + K))/Cl; CAI-II = (Cl − (Na + K))/(SO₄²⁻ + CO₃²⁻ + NO₃⁻), bicarbonate hazard index (mg·L⁻¹ HCO₃⁻/6100), residual sodium carbonate index ((CO₃²⁻ + HCO₃⁻) − (Ca²⁺ + Mg²⁺)) [27,186,187]. Water which has been partially desalinated by ZVM will show declines in sodium percentage and chlorine percentage. Other ions, Mg, Ca, bicarbonate/carbonate, sulphate, nitrate are also removed by the desalination process (Table C1, Table C2, Table C3 and Table C4), therefore analyses of product water suitability for irrigation based solely on ratio analysis of Na, K, Mg, Ca, Cl, HCO₃/CO₃, SO₄, NO₃ may not be meaningful. Desalination can remove Mg, Ca, Cl, HCO₃ at the same rate, or a faster rate, than K, Na and Cl.

7.2. Cost of Irrigation Water

7.2.1. ZVM TP

The ZVM TP analyses (Table 1, Figure C1, Figure C2, Figure C3, Figure C4, Figure C5, Figure C6, Figure C7, Figure C8, Figure C9, Figure C10, Figure C11, Figure C12, Figure C13, Figure C14, Figure C15, Figure C16 and Figure C17) demonstrate a base desalination cost (based on a single usage at 20 kg·m⁻³) of $6.5–$18 m⁻³ (at a ZVM TP cost of $325–$900 t⁻¹; desalination cost = P_w × Price/number of times the ZVM TP is reused). Trial CSD1 establishes that the ZVM TP can be reused at least four times without reprocessing or regeneration. This reduces the effective desalination cost to $1.6–$4.5 m⁻³.

Example ST3b (Table 1) demonstrates a reduction in salinity from 11.05 gNaCl·L⁻¹ to 3.34 g·L⁻¹ (over 126 days) for a cost of $5.6–$22.48 m⁻³. This change would increase the yields of barley, cotton, wheat, date palm, sugar beet, wheat grass, and Bermuda grass by between 300% and 800%, and would allow the agricultural holding to diversify to produce one or more of soybean, safflower, sorghum, rice, sesbania, flax, beans, figs, olives, cantaloupe, corn, lettuce, pepper and a wide variety of grasses for pasture (Appendix B, Figure B1).

Widespread usage of ZVM TP as an agricultural desalination agent requires the effective ZVM TP cost (following regeneration and reprocessing, or resale as a desulphurization agent (or for another application)) to reduce to <$58 t⁻¹ or P_w to reduce to <4.5 g·L⁻¹.

7.2.2. ZVM TPA

The ZVM TPA analyses (Trial E146) demonstrated a partial desalination cost of $0.009–$0.022 m⁻³. This was associated with a regression relationship (Figure 20) which links feed water salinity to product water salinity. This regression equation indicates that feed water with a salinity of 6 g·L⁻¹, will produce a product water with a salinity of 3.35 g·L⁻¹. This change in water salinity would allow a crop such as rice which cannot be grown in water with a salinity of >6 g·L⁻¹ to be grown with a yield decrement of 40% [188] (Figure B1).

7.3. Potential Use of ZVM TPA to Boost Rice Yields

Rice yields and production costs vary with time, location and variety, but generally fall within the range: yield = 6–24 t·ha⁻¹; wholesale price = $300–400 t⁻¹; production cost = $300–$400 ha⁻¹. Irrigation requirements for rice can fall within the range 5000–10,000 m³·ha⁻¹·a⁻¹ [189]. The cost of supplying partially desalinated irrigation water for rice paddy (based on Trial E146) is estimated to be $45–$220 ha⁻¹·a⁻¹ (i.e., a total rice production cost of $350–$420 ha⁻¹). These costings indicate (Table 7) that the ZVM TPA partial desalination technology could be used to provide water for rice cultivation in areas which contain brackish saline water. Coupling partial desalination with high yield rice varieities may allow the expected rice yield to fall in the range 10–15 t·ha⁻¹.

Field scale pilot trials are required to establish that the initial trials of ZVM TP and ZVM TPA technology can be upscaled to provide a low cost partial desalination approach for agricultural applications. The site specific water, including agricultural chemicals, fertilisers and microbiota, may interact with ZVM TP and ZVM TPA to accelerate, or retard, the desalination process.

Table 7. Potential impact of partial desalination on the economics of rice cultivation. Assumptions: ZVM TPA (Trial E146); Desalination determination: Figure 20.

Item	No Desalination		Partial Desalination
Irrigation, m3·ha−1	5000	10,000	5000	10,000
Feed Water salinity, g·L−1	6	6	6	6
Irrigation Water Salinity, g·L−1	6	6	3.35	3.35
Rice Yield, t·ha−1	-	-	-	-
Low	6	6	6	6
High	24	24	24	24
Yield Decrement	100%	100%	40%	40%
Actual rice yield, t·ha−1	-	-	-	-
Low	0	0	3.6	3.6
High	0	0	14.4	14.4
Production Cost, $·ha−1	-	-	-	-
Low	300	300	345	390
High	400	400	500	620
Rice price, $·t−1	-	-	-	-
Low	300	300	300	300
High	400	400	400	400
Net Sales Revenue, $·ha−1	-	-	-	-
Low	−400	−400	580	460
High	−300	−300	5415	5370

Table 7. Potential impact of partial desalination on the economics of rice cultivation. Assumptions: ZVM TPA (Trial E146); Desalination determination: Figure 20.

Item	No Desalination		Partial Desalination
Irrigation, m3·ha−1	5000	10,000	5000	10,000
Feed Water salinity, g·L−1	6	6	6	6
Irrigation Water Salinity, g·L−1	6	6	3.35	3.35
Rice Yield, t·ha−1	-	-	-	-
Low	6	6	6	6
High	24	24	24	24
Yield Decrement	100%	100%	40%	40%
Actual rice yield, t·ha−1	-	-	-	-
Low	0	0	3.6	3.6
High	0	0	14.4	14.4
Production Cost, $·ha−1	-	-	-	-
Low	300	300	345	390
High	400	400	500	620
Rice price, $·t−1	-	-	-	-
Low	300	300	300	300
High	400	400	400	400
Net Sales Revenue, $·ha−1	-	-	-	-
Low	−400	−400	580	460
High	−300	−300	5415	5370

7.4. Economic Implications of Corrosion Scale

ZVI desalination has been associated with a high surface area [20,21,25,190], indicating that the contact area between the Fe⁰ and water surface is important [15,18,191,192]. Experiments using Fe + organic chemicals have demonstrated [5,6,193,194,195,196,197,198,199,200] that increases in water salinity are associated with in an increase in the rate of degradation of the organic chemicals and (CO_x, H_xC_yO_z). This is associated with the construction of simple organic compounds (C_xH_yO_z, alkanes and alkenes), where carbon can act as both an intermediary and reactant and Fe acts as a catalyst [5,6,193,194,195,196,197,198,199,200]. Trials PS15 and PS16 demonstrate that the presence of CO, CO₂ and CH₄ can be associated with increases in the rate of desalination.

This study has demonstrated that the ZVM TP is electrically active and contains both cathodic and anodic centres. Under these circumstances the potential cathodic half reactions [150] in saline water are:

Na⁺ (aq) + e⁻ = Na ΔE⁰ = −2.71 V

(37)

2H₂O (l) + 2e⁻ = H₂ (g) + 2OH⁻   ΔE⁰ = −0.83 V

(38)

Half reaction (Equation (38)) is favoured over half reaction (Equation (37)) [150]. This is confirmed by Figure D1 which demonstrates production of H₂ (g) + 2OH⁻ during desalination. Half reaction (Equation (38)) can result in a film of NaOH coating [201] the ZVI and a corrosion scale (Fe(OH)₂, GR1, FeOOH) coating the ZVI [18,68,69,70,71,72,73,74,75,76,77,87,88,93,94,96,110,111,112,167,172,173,176,191,202,203,204].

Na⁺ + OH⁻ = NaCl

(39)

Fe⁰ + nOH⁻ = Fe(OH)_n
Fe(OH)₂ = Fe(OOH) + H⁺ + e⁻

(40)

The associated anodic half reactions are [124]:

2Cl⁻ (aq) = Cl₂ (g) + 2e⁻   ΔE⁰ = −1.36 V

(41)

2H₂O (l) = O₂ (g) + 4H⁺+ 4e⁻   ΔE⁰ = −1.23 V

(42)

H₂O is oxidised in preference [150] to Cl⁻ at the anodic sites. However, as Cl⁻ concentrations increase, overcharging (development of FeOOH [18]) at the anodic sites (corrosion scale) results in a switch from half reaction Equation (42) to half reaction Equation (41) [150]. Any Cl₂ formed is removed, as demonstrated in Figure C37, by an anodic reaction as ClO⁻, e.g.,

Cl₂ (aq) + 2OH⁻ = 2ClO⁻ + 2H⁺ + 2e⁻

(43)

This model [150] results (as demonstrated in Section 4) in the amount of NaCl removed (and corrosion scale formed on the ZVI particles) being directly proportional to the inherent charge generated by ZVM TP/ZVM TPA, via the Nernst relationships [150], where [150]:

Charge, coulombs = current (amperes) × time (seconds)

(44)

The charge associated with desalination is [150]:

Charge = (1/Molecular weight (MW) Cl⁻) × Faraday Constant (F) × Weight (W) of Cl⁻ Removed

(45)

For reactions involving other species associated with corrosion (e.g., formation of Fe(OH)₂, GR1, FeOOH) the charge is a function of e⁻ [150]:

Charge = (1/MW(species)) × F/e_n × W (species) × e_n/Moles Species Weight of
Species formed or removed = (1/MW(species) × F/e_n × e_n/Moles Species)/Charge
e_n = Moles e⁻ associated with removal of 1 mole of the contaminant species

(46)

These relationships demonstrate that the rate of corrosion (and GR1, GR2, FeOOH scale formation around the n-Fe⁰ particles) will increase with increased contaminant concentration (e.g., NaCl) and with increased oxidation of the water. This has been demonstrated experimentally [205,206] and is used in this study (ZVM TPA, Section 5) to combine accelerated iron corrosion by oxidation with accelerated desalination. The charge associated with the development of the corrosion scale can be linked directly to the effective pseudo-specific capacitance (pSC) associated with the corrosion reactions as [207]:

pSC, F·g⁻¹ = charge/(P_w(Eh_t=0 − Eh_t=n))

(47)

These relationships allow changes in the reaction quotient, Q, [19,150], (by altering the composition (and concentrations) of dissolved gases in the water) to be used to accelerate desalination and corrosion product (or scale) formation.

7.5. Applicability to Specific Sites

This study has evaluated desalination in batch treatment sizes ranging from 0.2 L to 240 L. Commercial applications will require larger plant sizes and may require the consideration of more complex water chemistries. Scale up performance into the 10–1000 m³ reactor/tank size range and application of the process using complex saline water (e.g., seawater, flowback water) has not been trialled at this time. The desalination rates at a larger scale and with more complex feed waters may be higher than, or lower than, the rates recorded in this study. This study has not optimized the rate of desalination, but has demonstrated a number of different routes which can be trialled to optimise the desalination rate. Facilities, offsites, and land requirements associated with commercializing this technology will be location, size and application specific [208,209].

8. Conclusions

This study has demonstrated that partial desalination of water (combined with water treatment (cation and anion removal)) using pellets, particles and cartridges, has the potential to be a green, sustainable, economically viable commercial treatment option for some saline water bodies. The operating limits evaluated in saline water include:

• pH range = 2–12,
• Electrical conductivity range of the feed water = 2–78 mS·cm⁻¹,
• Salinity range of the feed water = 1–39 gNaCl·L⁻¹,
• Carbonate range in the feed water = 10 mg·L⁻¹ to >10 g·L⁻¹; sulphate ≤ 10.2 mg·L⁻¹.
• Temperature range considered = −8 to 50 °C.

This treatment option can either be operated using:

• No external energy, or
• Pressurized (<0.1 MPa) air, or
• Pressurized (<0.1 MPa) acidic gas (CO₂), or
• Pressurized (<0.1 MPa) anoxic gas (N₂), or
• Pressurized (<0.1 MPa) anoxic, acidic, reducing gas (H₂ + CO + CO₂ + CH₄ + N₂).

The desalination rates observed:

• Increase with increasing temperature;
• Increase with decreasing particle size (and increased particle BET surface area);
• Are optimised using P_w = 0.3–20 gFe⁰·L⁻¹;
• Are optimised by maintaining the ZVM TP/TPA in a static water body, while allowing the overlying water body to be circulating and pressurised using a reducing gas, or an oxidising gas.

The cost of reducing water salinity was demonstrated to fall in the range $0.002–$190 m⁻³.

The technology has applications for the desalination of irrigation water (example cost: $0.002–$0.026 m⁻³), desalination of livestock feed water (example cost: $0.002–$0.026 m⁻³) and the provision of emergency water in areas following the destruction of infrastructure (example cost: $4–$190 m⁻³).