Uncorrected Proof

1 © IWA Publishing 2017 Water Science & Technology | in press | 2017

The economic pre-treatment of coal mine drainage water

with caustic and ozone
B. H. Boyden, L. Nador, S. Addleman and L. Jeston

ABSTRACT
B. H. Boyden (corresponding author)
Coal mine drainage waters are low in pH with varying amounts of iron and manganese and are
L. Nador
generally brackish. The Austar Coal Mine in NSW, Australia sought alternatives to their current lime CNF & Associates,
Level 2, 393 George Street,
dosing as the pre-treatment before the downstream reverse osmosis plant. Undesirable operating Sydney,
NSW 2000,
aspects of the current system include manganese and gypsum scaling/fouling, the need for anti- Australia
E-mail: bhboyden@yahoo.com
scalants and reduced water recovery. Thirteen processes for acid mine drainage were initially
S. Addleman
considered. The preferred process of caustic and ozone for Mn(II) oxidation was pilot tested at up to
Ozone Technologies Group Inc.,
0.74 kL/hr at the mine site. Under proper conditions and no aeration, about 81 per cent of the Fe 253 Portman Lane, Suite 107,
Bridgeville,
could be removed (initially at 156 mg/L) as green rust. Supplemental aeration followed first-order PA 15017

kinetics and allowed 99.9 per cent Fe(II) oxidation and removal but only with a hydraulic residence L. Jeston
Austar Coal Mine,
time of about 47 minutes. The addition of supplemental Cu catalyst improved Fe removal. Ozone Middle Road,
Paxton,
applied after caustic was effective in stoichiometrically oxidising recalcitrant Mn(II) and any remaining
NSW 2325, Australia
Fe(II). Control of the ozonation was achieved using the oxidation reduction potential during oxidation and
Austar Coal Mine,
of the Mn(II) species. The use of caustic, followed by ozone proved economically comparable to the Locked Bag 806,
Cessnock,
current lime pre-treatment. NSW 2325,
Key words | acid mine drainage, caustic, iron and manganese oxidation, lime, ozone, pilot scale Australia

INTRODUCTION

Australia provides around 30 per cent of the world coal
trade and in 2011 was the world’s largest exporter of metal-
lurgical coal and second largest exporter of thermal coal.
One of the major production expenses of mining in general
is the management of potentially acid generating wastes
during production and post mine closure. The Australian
mining industry spends an estimated $US100 million per
year managing potentially acid generating wastes from oper-
ating mine sites (Ciccarelli et al. ) and over 15 years the
costs can amount to about $A900 million (1997 dollars)
Harries () for the whole Australia industry. Wilson
() estimated the global cost for environmental liabilities
of acid generating wastes at $US100 billion.
Acid mine drainage (AMD) occurs when sulphide-con-
taining ores are exposed to oxygen, water and certain
types of acidophilic bacteria. The ferric or Fe(III) leaching
of pyrite to form ferrous iron or Fe(II) is shown in Equation
(1) and the bacterial oxidation of ferrous iron to ferric in
Equation (2) (Rawlings et al. )
Fe(II)S2 þ 14Fe(III) þ 8HOH
¼ 15Fe(II) þ 2SO2 þ
4 þ 16H (1)
O2 (dissolved) þ 4Fe(II) þ 4Hþ ¼ 4Fe(III) þ 2HOH (2)
The acidic pH of the drainage subsequently dissolves
other constituents within its geological environment. Moha-
jane et al. () noted that the main components of acid
mine water are free acid, Fe(II), Fe(III), and concentrations
of other metals present such as Al(III), Mn(II), Ca(II), and
Mg(II), in addition to the sulphate from bacterial sulphide
oxidation collectively termed Acidithiobacillus ferrooxi-
dans. These bacteria can thrive in low pHs of 1 to 2 and
are used commercially for the bioleaching of metal ores
and the cleaning of pyrite from coal.

doi: 10.2166/wst.2017.263
 Uncorrected Proof
2 B. H. Boyden et al. | Treatment of coal mine drainage with caustic and ozone Water Science & Technology | in press | 2017

The traditional method for treating mine drainage has
been its collection into dams, lagoons and ponds and the
addition of alkaline chemicals, usually lime because of its
lower cost. Lime converts carbonic acid and bicarbonate
alkalinities into carbonate alkalinity as per Equations (3)
and (4), respectively. Roughly, 1.3 kg of calcium carbonate
hardness is created for every kg of 93 per cent hydrated
lime added.
Ca(OH)2 þ H2 CO3 ¼ CaCO3 þ 2H2 O
Fe(III) forms solute complexes with OH and can pre-
cipitate as iron oxyhydroxides (e.g. FeOOH), with the pH
decreasing during oxidation. Mohajane et al. () with a
synthetic coal mine water showed that Fe(II) oxidation
had first-order dependence on the Fe(II) concentration,
oxygen pressure, amount of recycled sludge, limestone
dosage and the mixing rate as shown in Equation (6).
@[Fe(II)]
¼ k[Fe(II)]0:90 [PO2 ]0:6 [CaCO3 ]0:3 [SS]0:03 [SR]0:03 (6)
(3) @t

where Fe(II) is the concentration of ferrous iron, PO2 is the

Ca(OH)2 þ Ca(HCO3 )2 ¼ 2CaCO3 þ 2H2 O
The higher the amount of total carbonate in a system,
the higher the buffering capacity against acid and alkali.
Heavy metals will precipitate under the resulting alkaline
conditions according to their solubility product chemistry.
Mn(II) and Fe(II) species, however, are better precipitated
after oxidation, with Mn(II) being the most recalcitrant to
natural oxidation until a pH  9.5 as shown in Table 1.
The kinetics of Equation (2) for ferrous iron can be
expressed empirically as Equation (5) (Stumm & Lee ).
d[Fe(II)]
¼ k[Fe(II)][O2 ](OH )2 ,
dt
where k ¼ rate constant L3 mol3 and equal to 1.1 ± 0.5 at
T ¼ 20.5 C with total alkalinity between 0.009 and 0.039
W
(4)
oxygen pressure above the mine drainage; CaCO3 is the
limestone concentration, SS is suspended gypsum and
ferric hydroxide sludge, and SR is the shaking or mixing
rate. However, the exponents over the SS and SR are so
small as to be neglected.
Oxidation of Mn(II) by natural dissolved oxygen is not
practical until the pH is at least 9.5 (Table 1 and Pourbaix
diagram in Figure 3). A lime system can only achieve this
pH level with substantial lime dosage and sludge production
as shown in the bench scale testing in Table 3. The presence
of sulphate in mine drainage in combination with lime
addition also means downstream gypsum scaling in addition
to the large amount of waste solids. These disadvantages for
(5)
a lime pre-treatment before a reverse osmosis (RO) system
were sufficient to consider alternative approaches.
This study explored alternatives to lime pre-treatment of

AMD specifically for the Austar Coal Mine (Austar) of Yan-

as mg CaCO3/L; t is time in minutes; [O2] is the dissolved
coal located in the Newcastle Coalfields of New South
oxygen concentration in mg/L; and [OH] is concentration
Wales, Australia. The mine produces about 2 million
of hydroxyl anions. The time for complete oxidation of fer-
tonnes per annum of coking coal. The mine utilises deep
rous iron is of the order of minutes in an aerated solution
underground coal mine longwall top coal caving technol-
when pH is above 7.0. Stumm & Lee () showed a 100-
ogy, used extensively in China (Humphries et al. ) to
fold increase in the Fe(II) oxidation rate for single pH unit
maximise thick seam coal recovery. The study sought to
increase as per the quadratic hydroxide functionality of
reduce lime sludge quantities and reduce gypsum and
Equation (5).
manganese scaling prior to the RO units. Initial jar testing
and bench studies were conducted, followed by pilot testing
Table 1 | Indicative times for the oxidation of Fe(II) and Mn(II) at various pHs
of the preferred option.

Temperature Time for 90 per cent Time for 90 per cent

W
pH ( C) Fe(2þ) Oxidation Mn(2þ) Oxidation
MATERIALS AND METHODS
6.0 21 100 hours Slowest
7.0 21 1 hour Slower
Current mine drainage treatment system
7.0 5 10 hours Slower still
8.0 21 <2 minutes Slowa The Austar mine drainage originates from two shaft pumps
9.5 21 instantaneous ∼30 minutesb that direct flow into a series of impoundment dams, the
a
Hem (1963) measured only 2 per cent oxidation of Mn(II) after 120 hours. first of which is 65 ML (65,000 m3). The overflow from the
first
b
Hem (1963) measured 95 per cent oxidation after 30 minutes. dam is subsequently pumped into a second 15 ML
 Uncorrected Proof
3 B. H. Boyden et al. | Treatment of coal mine drainage with caustic and ozone Water Science & Technology | in press | 2017

(15,000 m3) dam, where it is joined by the flow from a third
shaft pump. These two dams in total oxidise about two-
thirds of the Fe(II) and about one-third of the Mn(II)
before lime pre-treatment. The typical water quality that is
pumped for pre-treatment with lime is shown in Table 2.
After addition of up to 350 mg/L of hydrated lime in pre-
treatment, the drainage flows to a third in-series dam, a
6 ML (6,000 m3) precipitation dam, the design of which
acts as a plug-flow reactor. The effluent from the
precipitation dam shows removal of essentially all of the
iron (down to 1 ± 1 mg/L) and manganese down to
8.4 ± 1.2 mg/L. The flow from the precipitation dam flows
into a forth in-series dam, a 140 ML (140,000 m3) process
dam before being dosed with anti-scalant and pumped to
the RO treatment train. The RO product water is used for
site reuse and environmental discharge. The brine is put
back into the mine.
The disadvantages of the current system include: (i) a
restricted treatment capacity if operations need to increase,
(ii) piping and equipment are plagued by gypsum scaling,
(iii) significant difficult-to-remove waste solids, (iv) mem-
brane life reduction and (v) low water recovery.
Consideration of pre-treatment process options
The management of acid drainage has continually evolved
over the years. In addition to augmentation and optimis-
ation of the current lime system, three broad classes of
process alternatives were considered in this study: (1) Use
of oxidants, (2) Use of biological oxidation and reduction,
and (3) Ion exchange for a total of 13 separate options.
Augmentation of current lime system
It is well known that limestone or hydrated lime treatment
are the most cost-effective treatments for neutralisation of
acid mine water, and partial sulphate removal to levels of
2,000 mg/L can be achieved. Greater lime use (pH >11)
can reduce sulphate further to 1,200 mg/L (Geldenhuys
et al. ). This could potentially reduce scaling and cor-
rosion problems but would acerbate the waste solids
problem. Preventing gypsum scaling in the current effluent
management system has generally not been cost-effective
and the anti-scaling agents used are expensive and usually
not recommended for lime treatment plants.
Use of oxidants
If lime were to be replaced, an oxidising agent would still be
necessary to reduce the Mn(II) concentration. Zogo et al.
() reported that Okpara dam water with 30–50 mg/L of
iron and 1.5–4.5 mg/L of Mn at a pH of 5.6 to 6.7 was success-
fully pre-treated with 2.5 mg/L potassium permanganate
(KMnO4), for 99 per cent elimination of iron and about 72
per cent of manganese at a pH of 6.5. Sodium hypochlorite
(NaOCl) is a site chemical and another oxidant possibility
but a constant use would incur a background demand penalty
of 50 mg NaOCl/L due up to 7 mg/L ammonia. Ozone (O3) is
the strongest commonly used oxidant and can be made on site
less expensively than KMnO4. Ozone also reacts with ammo-
nia but the reaction is slow at pHs < 9 (Holgné & Bader
). Tewalt et al. () used a USGS patented process
(described by Sato & Robbins ) at the Little Toby Creek

Table 2 | The mean mine drainage water quality before and after lime pre-treatment

mg/L or as designated mg/L

Parameter Before After Parameter Before After

pH, units 4.4 ± 1.9 7.5 ± 0.2 Na 4,242 ± 924 3,916 ± 463
TDS 16,522 ± 2,300 14,863 ± 1,038 Cl 774 ± 80 645 ± 40
ΣHardness, mg CaCO3/L 3,256 ± 424 3,610 SO4 10,511 ± 1,769 10,074 ± 736
Alkalinity, mg CaCO3/L 10 þ 16 66 ± 19 ΣFe 76 ± 33 0.9 ± 1.0
Ca 360 ± 73 529 ± 24 ΣMn 13.9 ± 5.1 8.4 ± 1.2
Mg 561 ± 80 458 ± 48 Al 7±9 0.2 ± 0.2
TOC 2.7 ± 2.2 1.5 ± 0.4 B 0.5 ± 0.1 0.5 ± 0.04
COD 22 ± 28 0±0 Rctv Si 33 ± 25 7 ± 0.6
Turbidity, NTU 22 ± 7 4 Sr 2.1 ± 0.3 2.0 ± 0.3
Acidity, mg CaCO3/L 183 ± 223 NA NH3-N 6±1 6 ± 0.03

Nomenclature: COD, chemical oxygen demand; NA, not available; Rctv Si, reactive silica; SO4, sulphate; TDS, total dissolved solids; TOC, total organic carbon; Σ reads ‘total’ in front of a
parameter; values are means and most include standard deviations.
 Uncorrected Proof
4 B. H. Boyden et al. | Treatment of coal mine drainage with caustic and ozone
Treatment Plant in Pennsylvania, a limestone-based process
for AMD, for removal of iron at a pH of 6.0. The use of
ozone at an oxidation reduction potential (ORP) of 450 to
600 mV reduced Mn(II) from 11.6 mg/L to 0.2 mg/L after
the existing limestone treatment. It was noted that the settle-
ment of particulate Mn(IV) proved difficult, which is well
known in potable water treatment with manganese and ozone.
Biological treatment
The current dams do oxidise some iron and manganese, likely
by Acidithiobacillus ferrooxidans due to the prevalent low pH.
Meruane & Vargas () concluded that oxidation by this
bacteria of Fe(II) predominates over its chemical oxidation
at pHs below 5.0. If one considers more passive treatment sys-
tems for acidic mine waters, the Western Australian proposed
guidelines (Degens ) considered a number of options for
‘end of drain treatment’ including anoxic and oxic limestone
drains, composting wetlands, etc. At over 300 mine sites in
the bituminous coal region of the eastern United States,
AMD is being treated biologically in constructed wetlands
(Hammack & Hedin ). Biological neutralisation of
AMD in man-made wetlands has also been described by Skou-
sen et al. (). More active biological treatment is highly
dependent on pH and redox. Conditions are significantly
different for optimal Fe(II) removal as opposed to Mn(II)
removal and two separate process units can be required.
Anaerobic treatment with an external carbon source was
used by Hammack & Hedin () with a synthetic acid mine
water with sulphate of about 400 mg/L and Fe(II) of 100 mg/L.
Removal of 80 per cent of the iron, 50 per cent of the manga-
nese and 55 per cent of the sulphate with an influent pH
between 2.5 to 3.5 was achieved with a final effluent pH of
between 6.4 and 7.2. Case studies for sulphate removal at
the Landau Colliery in South Africa reported by Lorax
Environmental () concluded that biological sulphate
removal was amongst the better methods trialled. Modelling
of the Austar RO train with the Dow ROSA version 9.1 soft-
ware showed that reducing sulphate by 50 per cent could
potentially improve water recovery by 10 per cent.
Ion exchange
West et al. () describe an ion exchange based technology
to reduce sulphate levels in mine drainage, involving firstly a
strong acid cation resin unit to selectively remove Ca(II) and
Mg(II), followed by weak base anion exchange unit to selec-
tively remove the sulphate. In both stages, gypsum was
separated and filtered.
Water Science & Technology | in press | 2017
The literature review and the practicalities of using
specific technologies on site narrowed the options to those
that were further explored with jar and bench scale testing.
Bench scale studies
Jar testing was conducted to better understand the current
use of lime and the possible use of caustic as its replacement.
Caustic is already used on site to increase the pH of the lime
pre-treatment. The use of ozone was tested in bench scale
studies and carried further to the pilot trials.
Pilot study equipment
An ozone and chemical dosing pilot plant was designed and
housed in a 20-foot container as shown in Figure 1.
A side stream of the feed normally directed to the lime pre-
treatment system was used to feed the 1 m3/hr pilot plant.
Ozone was generated on site by a plant sourced from Plastic
Technics, Inc., Racine, WI, USA from site enriched 90 per
cent oxygen stream. The ozone gas analyser was supplied by
OzoneLab Instruments of Alberta, Canada. The unit was
added in-line between the iron oxidation reactor (IOR) and
the manganese oxidation reactor (MOR) with ORP control.
The pilot plant was fitted with an Allen-Bradley Compact
Logix L23 monolithic PLC to control the process with cascade
PID controllers. The PLC was connected with a mini SCADA
to provide historical trending and data storage for the test runs.
The ORP was the main control parameter with the SCADA.
Chemicals and analyses
Fifty percent (50%) caustic for the pilot plant was procured
from a local supplier. Iron, manganese and total organic
carbon (TOC) concentrations were determined colourime-
trically with a Hach Model DR/2700 Spectrophotometer
and respective Hach test kits. Low range total iron of 0.02
to 3 mg Fe/L was determined via the Hach FerroVer
Method, number 8008. High range total iron of 10 to
1,000 mg Fe/L was titrated using the Hach TitraVer Titra-
tion Method, number 8214. Manganese concentration was
determined using Hach Method 8034 for the range of 0.1
to 20.0 mg Mn/L. TOC was determined by Hach Method
10129 for the range 0.3 to 20.0 mg/L. Chemical oxygen
demand (COD) was determined using Hach Method 8000
in the low range from 0 to 150 mg/L.
Coal mine drainage from the second in-series dam (see
Table 3 ‘Before’ for water quality) was blended with caustic
in the mix tank (MT) as shown in Figure 1 to bring the pH to

5 B. H. Boyden et al. | Treatment of coal mine drainage with caustic and ozone
Figure 1 | Process flow diagram of pilot plant with instrumentation.
Table 3 | Results of bench scale testing of oxidants for Fe(II) and Mn(II) removal
Potassium Permanganate
Dose rates of ∼118 mg/L (dry basis) were
required for 100 per cent stoichiometric
dosing requirements of Fe(II) and Mn(II).
Larger doses turned solution pink
Iron removal of 99.9 per cent was achievable
(with 40 per cent stoichiometric dosing and
an 8 minute reaction time)
Mn(II) removal of 99.8 per cent was achieved
with 145 per cent stoichiometric dosing
and 16 minute reaction time. However,
94.9 per cent removal was still achieved
with 74 per cent stoichiometric dosing and
8 minute reaction time
a designated set point. The contents of the MT flowed into
IOR, which could be optionally aerated, and its contents sub-
sequently optionally ozonated and pumped into the MOR.
RESULTS AND DISCUSSION
Bench scale testing
The current lime pre-treatment system doses to a pH of
around 8.3. Jar testing using the Austar mine drainage (see
Table 2) found that:
Uncorrected Proof
Water Science & Technology | in press | 2017
Sodium Hypochlorite Ozone
NaOCl dose rates of ca. 127 mg/L (as Residual Mn(II) and Fe(II) were easily
Cl2) were required for breakpoint (BP) oxidised
þ100 per cent stoichiometric dosing
requirements
Iron removal of 99.9 per cent was Fe(II) was stoichiometrically oxidised
achievable with BP þ50 per cent
stoichiometric dosing and 10 minute
reaction time.
Mn(II) removal of only 32 per cent was Mn(II) removal of 99 per cent plus was
achievable with BP þ97 per cent achieved but dosages of over 500 per cent
stoichiometric dosing and 20 minute stoichiometric dosing were indicated,
reaction time. attributed later to over oxidation to Mn(VII)
• Lime dosing: no calcium and little sulphate removal
occurred at pHs up to 10.5. Approximately 23 per cent
of the sulphate could be removed with a reaction pH of
11.5, but lime dose rates of 1,280–1,480 mg/L (dry
basis) would be required, with subsequent sludge volumes
that occupied about 28 per cent of the 1 L testing jar. This
compares with 5 per cent solids occupation of the testing
jar at a pH of 10.5.
• NaOH dosing: dose rates of 200 to 218 mg/L (dry basis)
were required to raise the pH to neutral, with additional
dosages of 6–20 mg/L (dry basis) to adjust the pH back
up to neutral following the oxidation reactions.
 Uncorrected Proof
6 B. H. Boyden et al. | Treatment of coal mine drainage with caustic and ozone
The results of testing the alternative oxidants potassium
permanganate (KMnO4), sodium hypochlorite (NaOCl) and
ozone (O3) on this mine drainage are shown in Table 3. The
results are the averages from six mine drainage samples
taken over three consecutive days. All samples for oxidation
testing were initially raised by NaOH to a pH of near neutral
before addition of the oxidant. As per the kinetics, the pH
decreased as the oxidation proceeded. Reaction times were
controlled by quenching the oxidation with sodium thiosul-
phate, the exception being the ozone testing. NaOCl
oxidation was carried to breakpoint due to the background
presence of 6 ± 1 mg/L of ammonia (Table 2).
Fe(II) was the most easily oxidised by all the oxidants,
including air. Both the KMnO4 and NaOCl achieved 99.9
per cent oxidation in 10 minutes of reaction time with less
than the stoichiometric requirement of 0.94 mg KMnO4/
mgFe(II) and 0.62 mg Cl2/mg Fe(II), respectively. Ozone
oxidation of Fe(II) closely followed the stoichiometric
requirement of 0.43 mg O3/mg Fe(II). Mn(II) oxidation
was as expected more challenging, requiring 145 per cent
of stoichiometric dosing of 1.92 mg KMnO4/mg Mn(II)
and a 16 min reaction time to oxidise 32 per cent of the resi-
dent Mn(II). NaOCl required 97 per cent of stoichiometric
dosing of 1.3 mgCl2/mg Mn(II) but only achieved 32 per
cent Mn(II) oxidation in 20 minutes.
Pilot testing
Pilot testing utilised a side stream from the pre-treatment
plant. The liquid temperature at the time ranged from 16
W
to 18 C and the water quality as shown in Table 2.
Fe(II) oxidation
The intent of the IOR in the pilot plant process was to
remove remaining Fe(II) through pH adjustment and dis-
solved air oxidation. When adjusting the pH of the Austar
mine water to near neutral without air, we noticed the for-
mation of the bluish-green precipitant later identified as
‘green rust’ as noted by Run No. 1 in Table 4. Stauffer &
Lovell () also noticed this phenomena when injecting
alkali in coal mine drainage. The precipitant settled easily
and upon further investigation was found to exist potentially
as several different species (Morgan & Lahav ), depend-
ing on the composition of the water and the dissolved
oxygen. The pilot feed had the presence of Fe(II), Fe(III),
sulphate and chloride so the compositions of the precipitant
were postulated to be:
Water Science & Technology | in press | 2017
• Fe(II)3Fe(III) (OH)8]þ [Cl·H20] (because of the high
chloride levels) and/or
• Fe(II)4Fe(III)2(OH)12SO4· ∼ 8H2O (because of the high
sulphate levels).
These compounds will slowly oxidise (Morgan &
Lahav ) into a variety of stable iron oxide end products
such as hematite, magnetite, goethite and lepidocrocite that
settle well as opposed to the voluminous Fe(OH)3 that
does not. This oxidation via a colour change from green
to orange was observed in the pilot trials. One would
expect that the addition of lime causes the same reaction
that is ultimately masked by the white lime. Because lime
also raises the pH initially to over 8, part of the green
rust would almost instantaneously oxidise to Fe3O4·nH20
(the orange colour) but only to the extent of the dissolved
oxygen content of the water. Fe(II) concentrations of 100þ
mg/L would be rate limited by the dissolved oxygen
(Equation (2)) as was also noted by Hustwit et al. ().
When the air was turned on in the IOR, as shown in
Runs No. II and III in Table 4, the green rust turned
into the orange FeOOH species. Much of the Fe(II)
could be removed without aeration with this mine drainage
but lower levels of Fe(II) were achieved with aeration as
noted by the theoretical residual ozone demand figures.
Run IV in Table 5 showed the best Fe(II) removal using
a 47 minute HRT in the MT plus the IOR, with aeration
in the IOR and no ozone. Run No. VI in Table 5 showed
that a residence time of 20 minutes or less could be used
to get Fe(II) down to about 1 mg/L but doubling the resi-
dence time to over 40 minutes (Run IV) ensured
complete Fe(II) oxidation.
Mn(II) oxidation
Table 5 includes runs in the pilot plant specifically for
Mn(II) oxidation. Runs V through VII all utilised ozone
before the MOR and aeration in the IOR. It is well known
that ozone thermodynamically favours the oxidtion of
Fe(II) over that of Mn(II) and would therefore oxidise any
Fe(II) residual first. It is also well known from water treat-
ment that over-ozonation of manganese oxidises Mn(II) to
Mn(VII) or permanganate, which can turn the water pink-
ish. This was at times purposely observed in the pilot
trials. Often this is used for visual control of ozonation in
manganese containing potable water source waters. In
these pilot runs ozone reduced the measured Mn(II) levels
to the 1.2 to 2.4 mg/L level. Likely, the levels were lower
 Uncorrected Proof
7 B. H. Boyden et al. | Treatment of coal mine drainage with caustic and ozone Water Science & Technology | in press | 2017

Table 4 | Dosing of caustic and iron removal

Run Raw MT Set MT IOR MOR Max. Theo. Residual O3 Demand

No. Parameter Feed Pt. Exit Exit Exit (mg/L) Conditions/Comments

I pH: 5.47 8.04 7.27 7.15 NR *Q ¼ 522 L/hr; 31 min. HRT before MOR
ΣFe, mg/L: 156 27 ± 3 NR 12.9 þ Mn(II) demand *No air in IOR; Fe removal by green rust
*No O3 in MOR
II pH: 5.47 8.05 7.37 6.70 NR *Q ¼ 522 L/hr; 31 min. HRT before MOR
ΣFe, mg/L: 156 13.0 NR 5.6 þ Mn(II) demand *Diffused air in IOR; Fe removal by
FeOOH
*No O3 in MOR
III pH: 4.61 8.01 7.74 6.9 6.5 5.9 þ Mn(II) demand *Q ¼ 794 L/hr; 20 min. HRT before MOR
ΣFe, mg/L: 166 20.0 11.6 13.8 *Diffused air in IOR; Fe removal by
FeOOH
*No O3 in MOR
*ORP 268 mV MOR

Nomenclature: refer to Figure 1; MT, mixing tank; IOR, iron oxidation reactor; MOR, manganese oxidation reactor; Q is flow rate through the pilot plant; HRT, hydraulic retention time; NR, not
recorded; Σ reads ‘total’ in front of a parameter; values are single analyses and means with standard deviations.

Table 5 | Ozone oxidation of residual iron and manganese post IOR

MT
Run Set MT Max. Theo. Residual O3
No. Parameter Raw Feed Pt. Exit IOR Exit MOR Exit Demand (mg/L) Conditions/Comments

IV pH: 4.27 8.30 8.17 6.52 6.33 – *Q ¼ 340 L/hr; 47 min HRT before O3
ΣFe, mg/L: 156 1.13 <1 <1 0.4 *Diffused air in IOR; Fe removal by
ΣMn, mg/L: 12.8 NR 12.9 ± 0.1 10.3 11.4 FeOOH
11.8 *No ozone

V pH: 4.27 7.89 7.3 6.66 NR – *Q ¼ 740 L/hr; 22 min HRT before O3
ΣFe, mg/L: 173 22 26 2.2 0.9 *Diffused air in IOR; Fe removal by
ΣMn, mg/L: 13.0 NR 12.9 ± 0.1 2.4 2.1 FeOOH
3.0 *21.3 mgO3/L; ORP 338 mV post
MOR

VI pH: 4.27 7.89 7.3 6.25 6.15 – *Q ¼ 740 L/hr; 22 min HRT before O3
ΣFe, mg/L: 173 22 5 1.1 0.5 *Diffused air in IOR; Fe removal by
ΣMn, mg/L: 12.9 NR NR 1.7 1.5 FeOOH
2.0 *21.3 mgO3/L; ORP 338 mV post
MOR
*20 mg/L Cu(II) added as catalyst
VII pH: 4.25 8.17 7.48 6.04 5.99 – *Q ¼ 480 L/hr; 32 min HRT before O3
ΣFe, mg/L: 186 ± 3 NR 35 2.54 1.1 *Diffused air in IOR; Fe removal by
ΣMn, mg/L: 10.8 NR NR 1.6 ± 0.4 1.8 FeOOH
2.9 *28.9 mgO3/L; ORP: 378 mV post
MOR

Nomenclature: refer to Figure 1 and Table 4; ORP, oxidation reduction potential.

but inadequate filtering of the samples on site could not
remove all the particulate MnO2. Other work on US mine
drainage (Addleman ) with a similar design reduced
Mn(II) concentrations to <0.1 mg/L with ozone and
Tewalt et al. () achieved a Mn(II) level with ozone at
Little Toby Creek of 0.2 mg/L.
Addition of a catalyst
It has been reported that the presence of some chemical
species will catalyse the oxidation Fe(II) with dissolved
oxygen like the presence of copper (Cher & Davidson ).
This could in theory allow the use of a lower pH and less
 Uncorrected Proof
8 B. H. Boyden et al. | Treatment of coal mine drainage with caustic and ozone
caustic usage. Although not optimised, in two separate tests,
Runs V and VI, run as closely as possible in terms of pH
and residence times, the addition of 20 mg Cu(II)/L (added
in excess, 0.02 mg/L is often enough) in Run VI seemed to
enhance Fe(II) oxidation. This was shown by the iron concen-
trations and pHs in the IOR: 26 mg Fe(II)/L at pH 6.66
without Cu(II) and 5 mg Fe(II)/L at pH 6.25 with Cu(II).
Oxidation depresses the pH so the overall oxidation rate is
self-limiting. The pilot studies often started with a pH of
near 8 in the MT after caustic dosage, which would fall
between a pH of 6 to 7 in the IOR when the aeration was
on. The use of a catalyst can help to overcome these limit-
ations. Figure 2 shows that the Fe(II) oxidation and removal
with this mine drainage also followed first-order kinetics
shown in Equation (7), where the first-order rate constant
was calculated to be 0.0977 min.1 with a R2 of 0.76 for six
points.
[Fe(II)]f
¼ exp (  0:0977HRT, min:)
[Fe(II)]i
Figure 2 is a graph of the points from Tables 4 and 5 and
also contains a plot of data extracted from Morgan & Lahav
() for comparison. The points were pulled from a plot of a
W
batch test done at 25 C with a Fe(II) initial concentration of
100 mg/L at a constant pH of 7.0, hence the high R2 value.
Our pH varied between the MT set point of near 8.0 to as
Figure 2 | Oxidation of Fe(II) against residence time in the MT þ IOR ([Fe]f represents the final iron concentration; [Fe]i the initial iron concentration; the ratio of pHf /pHi represents the
fraction of the final pH over the initial pH; curve fits are first-order least square fits).
Water Science & Technology | in press | 2017
low as 6.25 out of the IOR. The two first-order rate constants
were, however, similar. The Morgan & Lahav () first-
order rate constant at pH 6.3 (not shown) was calculated to
be 0.0312 min.1 as opposed to 0.1151 min.1 at pH 7.0.
The effect of the addition of Cu(II) in Run VI (the square
point in Figure 2 that was not included in the curve fit) puts
the point significantly below our curve fit or an increased
rate of oxidation. The parallel Run V, without the Cu(II) cat-
alyst and everything else being approximately the same
(included in the curve fit) and exists vertically above the
Run VI point and definitely above the curve fit. Figure 2
also includes the pH decreasing (the top curve) as the
Fe(II) oxidation proceeded.
Control
The cost-effective operation of any ozonation process
depends on four factors: cost of production, required ozone
(7)
dose, transfer efficiency from gas to liquid phases and the
need for ozone destruction. Soluble Mn(II) can be oxidised
by ozone to form manganese dioxide MnO2, a fine particulate
that is removed in drinking water plants with a sand filter.
This process stoichiometrically consumes 0.88 mg of ozone
per mg of Mn(II). Over oxidation of manganese will form
pink soluble permanganate MnO4. Permanganate will nor-
mally return to manganese dioxide MnO2 over time (20–30
 Uncorrected Proof
9 B. H. Boyden et al. | Treatment of coal mine drainage with caustic and ozone
minutes) but over oxidation is a waste of ozone and costly,
whilst under dosing has ramifications downstream.
The ozone control philosophy most widely practiced in
the water industry is based on a constant ozone dose. The
required ozone dose is determined from laboratory batch
or pilot scale tests on the water to be ozonated. A dissolved
ozone analyser measures residual ozone concentration at a
selected sampling point after ozonation. The ozone reading
on the analyzer is compared to a pre-determined set-point
and the ozone dose is adjusted. Over ozonation for Mn(II)
control can be seen by a colour change in the water. How-
ever, a more robust control scheme was sought for the use
of ozone on this mine drainage. After reviewing a number
of options, it was decided to employ the electrochemical
properties of manganese in water.
The most useful tool to visualise these properties is the
Pourbaix diagram (named after Marcel Pourbaix 1904–
1998). The Pourbaix diagram (also known as an Eh-pH dia-
gram) presents as a phase diagram that shows equilibrium
conditions as straight lines along which activities are equal
for species on either side of the line, as calculated from
the Nernst equation. Vertical lines separate species that
are in acid-base equilibrium, non-vertical lines separate
species related by redox equilibria and dashed lines enclose
the practical region of stability of the water solvent to oxi-
dation or reduction. The Pourbaix diagram for manganese
and water is shown in Figure 3.
Figure 3 | Modified Pourbaix diagram for manganese in water (shaded area shows a
control region that can be moved as desired).
Water Science & Technology | in press | 2017
The ORP when kept within a defined ORP range was
found to be an effective control for ozone dosage for
Mn(II) oxidation in this mine drainage as it was for Tewalt
et al. (). The target control region EV range shown
was curve fitted and programmed into the PLC. The utility
of this control approach proved stable and was demon-
strated in the pilot runs when the EV trending was
monitored against the Mn(II) concentration.
Analysis of Figure 3 indicates that the species that we
wish to have in predominance for removal by subsequent fil-
tration is insoluble MnO2. The application of ozone
followed sequentially after the IOR in the pilot plant. The
ozone oxidation further depressed the pH from between 6
and 7 out of the IOR to around 6 after the MOR. A simple
spreadsheet curve fitting exercise was used to derive the
target EV as shown Equation (8).
Target EV ¼ 1:475  0:088pH[V] (8)
Plotting this relationship on the Pourbaix diagram in
Figure 3 shows that the maximum and minimum EV excursion
in order to remain within the MnO2 region is ±0.15 [V]. Con-
sequently, the full relationship can be presented Equation (9).
Target EV ¼ 1:475  0:088pH ± 0:15[V] (9)
This relationship was programmed into the Pilot Plant’s
control PLC, along with alarms and status indicators to alert
the operator that the conditions were right for taking of
samples. Run VII showed almost full removal of manganese
from the effluent. The SCADA system was used to prepare a
data dump for that period as shown in Figure 4. Two salient
points arise from Figure 3: firstly, the range of acceptable EV
values is relatively large, and secondly, the measured EV is
very stable once it is in the acceptable range. Each of these indi-
cates that the control scheme is very stable and easily
controlled. This control philosophy can be further developed
into a final plant control system by using the operator status
alert signals to control a pair of automated control valves:
one to direct the flow to a recirculation line, and the other to
direct the flow to discharge.
Cost of treatment alternatives
Table 6 gives a cost comparison of various management alterna-
tives for AMD. The table is split into those processes that:
• Promote Chemical Precipitation: (i) use of lime and lime-
stone (mostly for pH control and Fe(II) oxidation), (ii) the
 Uncorrected Proof
10 B. H. Boyden et al. | Treatment of coal mine drainage with caustic and ozone
Figure 4 | Test Results from the Pilot Plant SCADA for Run VII.
use of barium salts (used for sulphate removal by precipi-
tation of BaSO4), (iii) the SAVIN process (lime þ
Al(OH3) þ CO2 for sulphate removal) and iv) the CESR
process (‘cost effective sulphate removal’ by step wise pH
increases with lime to remove sulphate, nitrate and metals);
• Pre-treatments Only: (i) use of oxidants, (ii) use of mem-
brane treatments of microfiltration (MF) and
ultrafiltration (UF), (iii) use of caustic only, (iv) use of
ozone only and (v) use of caustic and ozone;
• Membranes and Ion Exchange Total Treatments: (i)
reverse osmosis (RO), (ii) SPARROW process (addition
of seed crystals of gypsum to feed water of RO), (iii)
EDR (electrical dialysis reversal for total dissolved
solids (TDS) removal, where polarity of electrodes is
periodically reversed) and (iv) GYP-CIX (ion exchange
for removal of sulphate) and
• Biological Total Treatments: (i) utilisation of sulphate
reducing bacteria (SRB), (ii) designed wetlands (free
water surface and subsurface flow, (iii) Alkalinity Produ-
cing Systems (SRB systems that reduce sulphate to
sulphides and then to elemental sulphur to precipitate
metal sulphides and produce bicarbonate) and iv) Per-
meable Reactive Barrier (creation of a reactive zone in
an aquifer, perpendicular to groundwater flow).
Addleman () treated mine water drainage in the U.S.A.
with caustic with an Opex of 0.2 USD/kL and with
only ozone (mine water pH at 6.3) for an Opex of
0.07 USD/kL. This study sequentially used caustic (mine
Water Science & Technology | in press | 2017
water with pH as low as 2.5), followed by ozonation for a com-
bined cost of 0.16 USD/kL, but this process was not
optimised. The removal of sulphate was shown by modelling
to be beneficial to increase RO water recovery and this could
be accomplished by SRB in an anaerobic treatment for an
Opex of 0.27 to 0.30 USD/kL, with an ion exchange system
like GYP-CIX for an Opex of 0.3 to 0.69 USD/kL or with
barium salts for an Opex of around 0.39 USD/kL. Wetlands
are one of the least expensive of the options with an Opex
0.09 USD/kL but may not meet final effluent objectives and
depends heavily on land availability. Other elements of
Table 6 include Capex and Opex cost comparisons of full
treatment systems for mine water drainage for comparison
purposes.
CONCLUSIONS
In the pilot plant trials conducted, caustic was shown to easily
replace the current lime system and would greatly reduce the
production of waste solids. Under proper conditions for this
mine drainage, about 81 per cent of the Fe could be removed
(initially at 156 mg/L) as green rust without supplemental
aeration with a hydraulic residence time of about 30 minutes.
Supplemental aeration followed first-order oxidation and
allowed 99.9 per cent Fe(II) oxidation and removal (also
initially at 156 mg/L) but only with a hydraulic residence
time of about 47 minutes. A hydraulic residence time of 20
to 22 minutes oxidised with supplemental aeration removed
 Uncorrected Proof
11 B. H. Boyden et al. | Treatment of coal mine drainage with caustic and ozone Water Science & Technology | in press | 2017

Table 6 | List and costing of various management options for AMD

Capex (Million
Approach USD/MLD) Opex(USD/kL) Advantages Disadvantages Comment

Chemical Precipitations
Limestone þ Lime
Lorax Environmental
USD 1 to 2
(2002) (1)
0.10(1)
(Nov 2002)
• Cheap, has wide • Sludge production/ Better suited as pre-
treatment
use removal
() (1)
Bowell () (2)
0.23(2) 0.61(2) • SO4 down to
1.2 g/L
• Metal removal
Ba salts 0.48(1) (1990) 0.36(1) (1990) • SO4 down to • Expensive chem. Capex incl. for
Lorax Environmental
() (1)
(25 MLD plant)
0.45(2)
0.39(2) 200 mg/L • Limited metal recycle of salts
(rotary kiln & S(0)
removal
Bowell () (2) • Sludge production)
SAVMIN Process 0.31(1) (2001) 0.17(1) (2001) • SO4 to 70 mg/L • Sludge production Whole treatment
Lorax Environmental
() (1)
(30 MLD plant)
0.32(2)
0.22(2) • Metal removal • Cheap lime at (lime þ Al(OH3) þ
CO2 for pH)
USD39/MT
Bowell () (2)
CESR Process Not given(1) 0.79(1) (2001), • SO4 down to • Sludge volume • Whole treatment
Lorax Environmental 0.39(2) Not given (2) 200 mg/L • Similar to
() (1)
Bowell () (2)
• Metal removal SAVMIN Process

Pre-treatments Only
RPSEA ()
Oxidant
0.0026
0.27–0.53
0.01
0.26–0.53
• Oxidation: No pre- • Oxidation: sludge • Oxidation
treatment from settling requires post
MF/UF • MF: 85 per cent þ • MF: pre-treatment, settling
recovery CIP, addnl. filters
Caustic and O3 Not given(1) 0.20(1) • NaOH: Effective (1) • NaOH: sludge & • 25 mg/L of Mn
Addleman () (1) 0.11(1) 0.07(1) • O3: Effective, less chem. transport (II) & Fe(II) at
NaOH only: (pH to 9.5) Not calc. (2) 0.12(2) sludge and O2 sat. • O3: MnO2 difficult pH 6.3; both to
O3 only: (pH at 6.3) 0.09(2) 0.04(2) • Made on site to settle (1)(2) <2 mg/L (1)
This work (6.5 MLD)(2)
NaOH Process Step
• Fe gone by NaOH • Processes were • Caustic taken at
(2) combined but not 700 USD/DMT,
O3 Process Step • O3 for Mn(II) optimised (2) feed pH 8 (2)
oxidation only (2)
Membranes and Ion Exchange Total Treatments
RO, Lorax 0.44–0.53(1) 0.88(1)(2) • Low TDS • Scaling • Highly dependent
Environmental () (1);
Schoeman & Steyn ()
(2001, 80 MLD)
0.82(2), 1.24(4)
(2001)
1.3 to 3.8(3)
• SO4 to 113 mg/L • Short Membrane upon pre-
(NaOCl, FeCl3, Life treatment(s) and
(2); Energy-Water Nexus
() (3); RPSEA ()
(RO þ Sparrow) 0.93(4)
(RO þ lime)
settling, media filter • Fe fouling sulphate
before RO)
(4)
SPARRO Process, 0.52(1) (1992) 0.22(1) (1992) • Low TDS • As above • Phase1 adds
Lorax Environmental (4 MLD plant) 0.30(2) • SO4 down to • Fouling without NaOH þ
() (1); 0.85(2) 150 mg/L metal pre-treatment KMnO4þ
Schoeman & Steyn ()
(2)
Pilot testing(2) • Better RO • Salt rejection went polymer before
performance with down 90 per cent to RO
P1 50 per cent • Lime added
EDR, Lorax 0.56–0.67(1) 0.48(1)(2) • Low TDS • Scaling problems • ED with polarity
Environmental () (1);
Schoeman & Steyn ()
(2001)
0.69(2)
(2001)
0.95(3)
• SO4 down to • Short membrane reversal
250 mg/L life
(2); RPSEA () (3) • Tolerates higher
TSS

(continued)
 Uncorrected Proof
12 B. H. Boyden et al. | Treatment of coal mine drainage with caustic and ozone Water Science & Technology | in press | 2017

Table 6 | continued

Capex (Million
Approach USD/MLD) Opex(USD/kL) Advantages Disadvantages Comment

GYP-CIX (2 stage IX), 0.33–0.37(1), 0.69(1) • Low TDS • Sludges • Treat raw
Ibid. (1)(2)(3) 0.39(2), 0.59(2) • SO4 down to • Resin needs regen. • Ca(OH)2 &
0.5–0.69(3) 0.3–0.45(3) <240 mg/L • System complex H2SO4 regen.
Biological Total Treatments
Anaerobic Bioreactor 0.24(1) (1992) 0.27(1) (1992) • SO4 to 200 mg/L • Needs Carbon • SO4 ! (HS)! S
Lorax Environmental 0.32(2) 0.30(2) • Metal Rem./ • Long start up (0) (reduction to
() (1) Recover • Toxic sensitive elemental S)
Bowell () (2) • Low Sludge
• Commercial
variants
Designed Wetlands Depends on Not given but • Passive treatment • Fe(III) or NO3 • Surface or sub-
Ibid. (1)(2) land cost(1) low(1) • Metal Removal inhibits SO4 red. surface flow
0.11(2) 0.09(2) • Large land area
Alkalinity Prod. Depends on Not given but • Metal Removal • Little SO4 reduction Fe should be
Systems
Ibid. (1)(2)
land cost(1)
0.15(2)
low(1)
0.11(2)
• Passive • Gypsum sludge removed in pre- or
post-treatment
(Includes anoxic
limestone drains and
others)
Permeable Reactive
Barrier (reactive zone in
USD 65 k(1)
(2002)
USD 30 k per
annum(1)(2)
• SO4 down to • Long term Construct.
Perpendicular to
850 mg/L unknown
an aquifer) USD 80 k(2) (2002) • Passive system • Site and EPA ground water flow
Ibid. (1)(2) • Used in N. America approval issues

Nomenclature: refer to Table 3; MT, metric tonne; DMT, dry metric tonne; kL, kilolitre or 1 m3; Opex, annual operational and maintenance costs; Capex, initial capital costs.

85 to 92 per cent of the Fe. In parallel trials, each with 22 min-
utes of hydraulic residence time, providing a Cu(II) improved
the Fe removal from 85 per cent (initially at 173 mg/L) to 97
per cent. A lower pH set point could be theoretically be used
with a Cu(II) catalyst, thus affecting Opex savings with the
use of more expensive caustic. The Cu(II) would sub-
sequently be removed in the RO units. Ozone applied after
caustic was effective in stoichiometrically oxidising recalci-
trant Mn(II) and any remaining Fe(II). Control of the
ozonation process was achieved by monitoring the ORP to
keep it above 450 mV. The presence of TOC had little to no
effect with regards to ozone background demand. The Opex
cost of using hydrated lime for pre-treatment ranged between
0.10 to 0.61 USD/kL, the larger figure for significant gypsum
removal. The use of caustic, followed by ozone proved favour-
able, ranging from 0.16 to 0.27 USD/kL.
REFERENCES
Addleman, S.  Personal Communication on Pilot Studies
Conducted at a Coal Company in Somerset County,
Pennsylvania.
Bowell, R. J.  A Review of Sulphate Removal Options for Mine
Waters. pp. 1–23. Available from: http://citeseerx.ist.psu.edu/
viewdoc/download;jsessionid=2D7FFD9F884037B947B71
E706EB6A26B?doi=10.1.1.631.7837&rep=rep1&type=pdf
(accessed June 2016).
Cher, M. & Davidson, N.  The kinetics of the oxidation of
ferrous iron in phosphoric acid solution. J. Am. Chem. Soc.
77, 793–798.
Ciccarelli, J. M., Weber, P. A., Stewart, W. S., Li, J., Schumann, R.,
Miller, S. D. & Smart, R. St. C.  Estimation of long-term
silicate neutralisation of acid rock drainage. In: 8th
International Conference on Acid Rock Drainage (ICARD)
and Securing the Future: Mining, Metals and the
Environment in a Sustainable Society, Proceedings of a
Meeting Held 22–26 June 2009, Skelleftea, Sweden.
Degens, B.  Proposed Guidelines for Treating Acidic Drain
Water in the Avon Catchment, Western Australia: Adapting
Acid Mine Drainage Treatment Systems for Saline Acidic
Drains. Report of Government of Western Australia,
Department of Water, Salinity and Land Use Impacts series,
No. SLUI 54.
Energy-Water Nexus: Information on the Quantity, Quality, and
Management of Water Produced during Oil and Gas
Production . Report of the United States Government
Accountability Office – Report to the Ranking Member,
Committee on Science, Space, and Technology, House of
 Uncorrected Proof
13 B. H. Boyden et al. | Treatment of coal mine drainage with caustic and ozone
Representatives. GAO-12-156. Available from: http://www.
gao.gov/assets/590/587522.pdf (accessed June 2016).
Geldenhuys, A. J., Maree, J. P., de Beer, M. & Hlabela, P.  An
integrated limestone/lime process for partial sulphate
removal. Journal of the South African Institute of Mining and
Metallurgy.
Hammack, R. W. & Hedin, R. S.  Microbial sulphate
reduction for the treatment of acid mine drainage: a
laboratory study. Report from Pittsburgh Research Center, U.
S. Bureau of Mines, U.S. Department of the Interior,
Pittsburgh, PA. Also in Proceedings of the ninth annual West
Virginia surface mine drainage task force symposium.
Harries, J. R.  Acid Mine Drainage in Australia: Its Extent and
Potential Future Liability, Supervising Scientist Report 125,
Supervising Scientist, Canberra. Available from: http://www.
environment.gov.au/science/supervising-scientist/
publications/ssr/acid-mine-drainage-australia-its-extent-and-
potential-future-liability (accessed April 2017).
Hem, J. D.  Chemistry of Manganese in Natural Water, USGS
Paper 1667-A, pg A16. Available from: https://pubs.usgs.gov/
wsp/1667a/report.pdf (accessed April 2017).
Holgné, J. & Bader, H.  Ozonation of water: kinetics of
oxidation of ammonia by ozone and hydroxyl radicals.
Environ. Sci. Technol. 12 (1), 79–84.
Humphries, P., Poulsen, B. & Ren, T.  Top Coal Caving
Application Assessment in Australia, Report CSIRO-ACARP
Project C13018, P2006/308. Available from: http://www.
acarp.com.au/abstracts.aspx?repId=C13018 (accessed April
2017).
Hustwit, C. C., Ackman, T. E. & Erickson, P. M.  Role of Oxygen
Transfer in Acid Mine Drainage Treatment. U.S. Bureau of
Mines Report of Investigation # 9405, Pittsburgh, PA, 18 pp.
Lorax Environmental  Treatment of Sulphate in Mine Effluents,
Report by the International Network for Acid Prevention
(INAP). http://www.inap.com.au/public_downloads/
Research_Projects/Treatment_of_Sulphate_in_Mine_
Effluents_-_Lorax_Report.pdf (accessed April 2017).
Meruane, G. & Vargas, T.  Bacterial oxidation of ferrous iron
by Acidithiobacillus ferrooxidans in pH range 2.5–7.
Hydrometallurgy 71 (1–2), 149–158.
Mohajane, G. B., Maree, J. P. & Panichev, N.  Treatment of
iron(II)-rich acid mine water with limestone and oxygen.
Water Sci. Technol. 70 (2), 209–217.
Morgan, B. & Lahav, O.  The effect of pH on the kinetics of
spontaneous Fe(II) oxidation by O2 in aqueous solution – basic
principles and a simple heuristic description. Chemosphere 68,
2080–2084.
First received 21 December 2016; accepted in revised form 19 April 2017. Available online 5 May 2017
Water Science & Technology | in press | 2017
Rawlings, D. E., Tributsch, H. & Hansford, G. S.  Reasons
why ‘Leptospirillum’-like species rather than Thiobacillus
ferrooxidans are the dominant iron-oxidizing bacteria in
many commercial processes for the biooxidation of pyrite
and related ores. Microbiology 145, 5–13.
RPSEA  An Integrated Framework for Treatment and
Management of Produced Water: Technical Assessment of
Produced Water Treatment Technologies, 1st ed. Report by
Colorado School of Mines, Golden CO, Project 07122-12.
Available from: http://aqwatec.mines.edu/produced_water/
treat/docs/Tech_Assessment_PW_Treatment_Tech.pdf
(accessed June 2016).
Sato, M. & Robbins, E. I.  USGS U.S. patent 6,485,696.
Schoeman, J. J. & Steyn, A.  Investigations into alternative
water treatment technologies for the treatment of
underground mine water discharged by Grootelvei
Proprietary Ltd into the Blesbokspruit, South Africa.
Desalination 133, 13–30.
Skousen, J., Sexstone, A., Garbutt, K. & Sencindiver, J. 
Wetlands for treating acid mine drainage. Greenlands 22 (4),
31–39. Available from: http://anr.ext.wvu.edu/r/download/
44884 (accessed June 2016).
Stauffer, T. E. & Lovell, H. L.  The Oxygenation of Iron I1
Solutions Relationships to Coal Mine Drainage Treatment.
Available from: https://web.anl.gov/PCS/acsfuel/preprint%
20archive/Files/13_2_MINNEAPOLIS_04-69_0088.pdf
(accessed April 2017).
Stumm, W. & Lee, G. F.  Oxygenation of ferrous iron.
Industrial and Engineering Chemistry 53, 143–146.
Tewalt, S. J., Motoaki, S., Dulong, F. T., Neuzil, S. G., Kolker, A. &
Dennen, K. O.  Use of ozone to remediate manganese
from coal mine drainage waters. In: Presented at the 2005
National Meeting of the American Society of Mining and
Reclamation, Breckinridge, CO, June 19–23, Published by
ASMR, 3134 Montavesta Rd., Lexington, KY 40502.
West, A., Kratochvil, D. & Fatula, P.  Sulfate Removal from Acid
Mine Drainage for Potential Water Re-use. In: Presented at
72nd Annual International Water Conference, Orlando,
Florida.
Wilson, G. W.  Why are we still struggling with acid rock
drainage? Geotechnical News June, 51–56.
Zogo, D., Bawa, L. M., Soclo, H. H. & Atchekpe, D.  Influence
of pre-oxidation with potassium permanganate on the
efficiency of iron and manganese removal from surface water
by coagulation-flocculation using aluminium sulphate: case
of the Okpara dam in the Republic of Benin. J. Environ.
Chem. Ecotoxicol 3, 1–8.