An investigation into the availability and role of oxygen gas in

gold tailings dams of the Witwatersrand basin with reference

to their acid mine drainage potential

AC Nengovhela1*, B Yibas2 and JS Ogola1

1
Department of Mining and Environmental Geology, University of Venda, Private Bag X 5050, Thohoyandou 0950, South Africa
2
Council for Geoscience, Private Bag X112, Pretoria 0001, South Africa

Abstract

The oxygen content of tailings dams around the Witwatersrand Basin was quantitatively measured over a period of 2 months
using a multi-level gas sampling device (MLGS) in an attempt to understand the diffusion of oxygen in tailings dams as a
result of acid mine drainage. The measured oxygen showed that the diffusion of oxygen in some Witwatersrand tailings dams
is up to a depth of 4 m. In some instances the oxygen content in the layer 1m below the tailings surface decreased by a factor
of more than 97% compared to the atmospheric content. The findings show that the development of oxidation zones in the
tailings dams of Witwatersrand Basin which subsequently leads to acid mine drainage is limited by the amount of available
oxygen in the tailings materials.

Keywords: oxygen, oxidation zone, acid mine drainage

Introduction water infiltrating the impoundment contributes a negligibly

small amount to the flux of oxygen entering the impoundment
Over the past few years, there have been concerted efforts (Ritchie, 2003).
throughout the world to mitigate the impacts of mine waste Fine tailings are often deposited as slurry in a contain-
deposits on the environment. If sustainable development is to ing dam and allowed to settle. As diffusion of oxygen in air is
be realised, the impacts of mine waste cannot be overlooked. approximately 4 times greater than in water, ingress of oxygen
This paper is a product of a research project funded by the Water into tailings is relatively slow under water-saturated conditions.
Research Commission to Pulles Howard and De Lange (Yibas After closure of a tailings dam, or even during use, a decrease in
and Nengovhela, 2005). the water level may lead to the formation of an unsaturated zone,
The formation of acid mine drainage (AMD) from mine resi- thereby opening the way for oxygen diffusion in pore spaces, and
due deposits is one of the major problems faced by the mining hence faster oxidation processes (Lindvall and Eriksson, 2003).
industry. It occurs when sulphide minerals in rocks and/or tail- The shape of an oxygen profile is controlled by the rate and
ings are oxidised, usually because of exposure to moisture and manner in which oxygen diffuses or advects into the dam as
oxygen (Ritchie, 2003). This paper presents results of the deter- well as the rate at which it is consumed within the dam. The
mination of oxygen availability in tailings dams. The generally consumption of oxygen results in the development of oxygen
accepted acid mine drainage equation below, shows that oxygen profiles that evolve over time as the oxidation process progresses
availability is a prerequisite in AMD generation (Davis et al., (Ritchie, 1994). Several variables control the development of
1986). oxygen profiles and can be simplified as; the availability of oxy-
gen, the quantity of pyrite and other sulphides, the rate of pyrite
FeS2 + 7/2O2 + H2O → Fe2+ + 2SO42- + 2H+ oxidation, the age of the dam, the buffer capacity of the dam
rock, microbial activity, and the location of the profile on the
The tailings deposit environment is a complex system of coupled dam (Ritchie, 1994; Forstner and Salomons, 1988).
physical and biogeochemical processes which can be divided Oxygen can be transported into the tailings deposit by three
into primary processes, for example water infiltration, water- mechanisms, namely:
saturation, oxygen diffusion, sulphide oxidation and second- • Advective transport with water that contains oxygen
ary processes. Sulphides are consumed during oxidation, but it • Free or forced air convection through the top layer of the
is assumed that their consumption leads to a negligibly small tailings deposit
change in the bulk density of the material in the tailings. Oxy- • Diffusion in the gaseous and aqueous phases through pores
gen consumed in the oxidation of sulphides leads to a change in the tailings material (Werner, 2000).
in gas composition, with a consequent change in gas density.
It is further assumed that the quantity of gas dissolved in rain- The last mechanism is the dominant mechanism for transporting
atmospheric oxygen from the surface of the tailings to the depth
where the oxidation takes place (Jaynes et al., 1984; Nicholson et
* To whom all correspondence should be addressed.
 +27 76 926-8106; fax: 018 787 5972;
al., 1989; Pantelis and Ritchie, 1991; Yanful, 1993). Oxygen dif-
e-mail: colbert.nengovhela@goldfields.co.za fusion in the gaseous phase is often considered the main trans-
Received 24 August 2006; accepted in revised form 17 November 2006. port process for oxygen within the unsaturated zone. However,

Available on website http://www.wrc.org.za 271

ISSN 0378-4738 = Water SA Vol. 33 No. 2 April 2007
ISSN 1816-7950 = Water SA (on-line)
oxygen diffusion is sometimes conceptualised as a two-step
process, with ‘bulk diffusion’ in the pore gas as the first step and A
aqueous-phase diffusion near and/or within the sulphide min- C B
eral grains as the second step (Davis and Ritchie, 1986).
The oxygen diffusion mechanism in either gas or water E&F
phase may be treated as one-dimensional. The oxygen diffu-
sion rate in the unsaturated zone is high, relative to the saturated D
zone. Therefore, in the unsaturated zone, the sulphide oxidation
rate may be relatively rapid, since the potential availability of
both water and oxygen is large. Below the groundwater table,
oxygen diffusion is greatly reduced, and the oxygen content
will be limited to its solubility in water, which results in that
sulphide oxidation should occur slowly below the groundwater
table (Yanful, 1993).
Oxygen content inside a tailings dam varies depending on
the degree of water-saturation, temperature, the rate of water
infiltration and the rate of oxygen consumption. Elberling et al.
(1993) conducted field studies on uncovered unsaturated tail-
Figure 1
ings deposits in Canada. They reported that the pore gas oxygen Location map of all the study sites showing their spatial
content decreased from atmospheric contents (21 vol %) to less distribution across the Witwatersrand Basin (Gold Fields,
than 5 vol % within the upper 600 mm of the tailings dams. Driefontein Gold Mine, 2004)
In water-saturated conditions, the oxygen content is limited to
the solubility of oxygen in water, or 258 µM at 25oC (that is 21
vol %).

Materials and methods

Study area

Six sites, distributed across the Witwatersrand Basin were

selected for this study (Fig. 1). One in the Stilfontein area, two
in the Carletonville area on the border between Gauteng and
North West Provinces, one in the Central Rand area of the Wit-
watersrand Basin and the last two in the East Rand area of the
Witwatersrand Basin. The sampling was carried out between Legend
April and June 2005. The characteristics of the 6 dams are sum-
marised in Table 1.

TABLE 1
Summary of available background information for
the selected dams
Site Area Height Volume Date of
Name (ha) (m) (m3) decommission-
ing

Site A # 10.5 -22.5 # 1950s? Figure 2

Schematic diagram of the MLGS demonstrating the alternating
Site B 38 25 7 828 000 1976 layers of gravel, coarse and fine-grained sand used to promote
Site C 44.2 32 7 576 000 1975 horizontal pore-gas flow. The pipes are placed at 1 m intervals
Site D # 34 # 1973 up to the bottom of the hole, i.e. 9 m.

Site E # # 31 422 569 1997

Site F # # # 1999 structed using 9 x 6 mm diameter polyethylene sample tubes per
#Unavailable information. dam. This device has a built-in pipe, which allows for the gas
to be pumped from the various depths (Fig. 2). The inlets of the
Instrumentation sample tubes were placed between 200 mm thickness of 13mm-
diameter gravel stones, i.e. 100 mm below and 100 mm above
The instrument used for the measurement of oxygen is known the port to prevent particulate uptake and to prevent clogging of
as the Gas Alert Max, from BW Technologies. This instrument the sample tubes.
pumps O2 at a rate of 250 mℓ/min (minimum). Its detection range The MLGS was installed in holes made by a hand auger of
for oxygen is 0 to 30% in 0.1 increments and it uses a plug-in 80 mm diameter. The well screen at the bottom of the centre
electrochemical cell sensor type. The O2 measuring principle stock extended to a depth of 10 m with a 900 mm bentonite layer
used is the capillary-controlled content sensor that comes with between the 9 m and 10 m level to minimise vertical air flow
an automatic calibration system. At each dam, a multilevel gas from the bottom of the tailings dams.
sampling system (MLGS) adopted from Shaw (2004) was con- Fine-grained sand (0.2 to 0.3 mm) was placed in between

272 Available on website http://www.wrc.org.za

ISSN 0378-4738 = Water SA Vol. 33 No. 2 April 2007
ISSN 1816-7950 = Water SA (on-line)
 0
comes to an end at the 4 m level where there is a decrease by
84% from the atmospheric content. This shows that there is little
2
oxygen beyond the 3 m level. Comparing this site to Site A, its
oxygen content at the 4 m level is 48% less than that of Site A.
4
The 3rd site, i.e. C is the only site that has an oxygen con-
Depth (m)

6
tent of higher than 3% below the 3 m mark. It contains signifi-
cant oxygen at the 4m level of 3.78%, which is relatively high,
8 compared to all the other sites, which normally have levels of
between 0.6 and 1.2% at this level.
10 Site D has the lowest oxygen level of all the dams at 5.42%
within the first metre, which decreases to 1.62% at the 2 m mark
12 (Fig. 3). The drop from the atmospheric content of 21% to 5.42%
0 2 4 6 8 10 12 14 16 18
O2 (%) is by 74%. The decline from 1 to 2 m is by 70%. This point
shows that at Site D, the rate of oxygen depletion is approxi-
D E A B C F
mately 72% within the first 2 m.
Figure 3 Site E is the youngest site of all the dams (decommissioned
Oxygen content in the 6 tailings dams around the in 1997). Its oxygen content at the 1m level is only 7.38%, which
Witwatersrand Basin is significantly less than the oxygen content of the 3 m level of
Site C. The drop from the atmospheric content to the first metre
each sampling port for a width of 600 mm to minimise the verti- is by 65%. From the first metre to the second, the oxygen is
cal movement of pore gas. Coarse-grained sand (0.8 to 1.0 mm) depleted by 64%. At the 3 m level, from which relatively con-
was placed immediately above the 20 mm of gravel for a width stant contents of oxygen begin, the gradient is 68% from the 2 m
of 30 mm to promote horizontal flow of the pore-gas. At the level.
tailings surface the annular space in the drill hole was sealed Site F is located within 5 km of Site E. It is slightly older than
to 900 mm depth with bentonite. Due to the age of the dams, the latter and most of its surface is covered with vegetation. This
and the fact that they were being reclaimed, no water table was site contains the third lowest oxygen content with the first metre
encountered at any of the sites. One hole per dam was drilled for only having a content of 9.67%. The oxygen content decreases
the insertion of the MLGS which means that the data collected below 1% at the 3 m level and remains constant to the end of the
results from one profile of the dam. sampled depth then decreases up to the 8 m level before it stabi-
lises below this level.
Results and discussion From the data presented it can be observed that the oxygen
contents in the first 10 m of the 6 studied tailings dams varied
Oxidation of sulphide minerals in mine waste deposits is con- between 0.36 and 17.2% over the eight-week sampling period
trolled by the supply of oxygen. Diffusion in the air filled part from April to June 2005. Within each tailings dam, oxygen con-
of the pore system above the phreatic surface is usually con- tents remained relatively constant between 3 and 10 m, which
sidered to be the dominant transport mechanism for oxygen. indicates that there is little change in oxygen content below
Hence, the formation of oxidation zones in tailings deposits is 3 m, which explains the reported average oxidation zone of
strongly dependent on the oxygen flux by diffusion. Relevant 2.5 m. Previous studies (Yanful, 1993, Davis et al, 1986) of tail-
quantification of oxygen diffusion is therefore a key factor in ings dams have demonstrated that the consumption of pore-gas
the design and assessment of prediction and prevention of AMD O2 through in-situ reactions results in decreasing O2 contents
(Nengovhela, 2006). with depth when transport is controlled by the vertical diffusion
Site A shows a decreasing pattern of oxygen content and of O2 from the atmosphere.
depletion throughout the 9 m depth (Fig. 3). Differences in oxy-
gen contents are only found within the top 3m after which there Conclusions
is no variation until the end of the sampled depth. The average
oxygen content at the 1 m level is 10.76 %, which is a decrease From this study, it can be concluded that oxygen depletion in
by a factor of 48% from the atmospheric content (21%). The the studied Witwatersrand tailings dams occurs rapidly from the
abrupt decrease from the 1 m level to 2 m, i.e. from 10.76 to atmosphere to the first few metres of the studied tailings dams.
3.88% means that there is high oxygen consumption in the top The conclusion generated from the calculated average depletion
part of the dam. From 2 m to 3 m, the decrease is from 3.88% to gradient of all the dams is that the drop between the atmospheric
1.2%, which is a 69% decrease. This means the decrease from oxygen level of 21% and the level of 0.56% at a depth of 5 m
the atmospheric content of oxygen to the 3 m level is by 94.3%. is 97% as shown in Fig. 3. At a depth of 1 m where sulphide
The drop in oxygen gradient at this site is around 60%/m for the oxidation is active, the average oxygen content is 10.67% and
first 3 m. decreases by 88.5% at 4 m to 1.23%. The progressive decrease
Site B is slightly different from Site A in that its average in oxygen from 1 to 2 m is by 43.29% and at 3 m it decreases
oxygen content in the oxidation zone is almost twice that of Site again by a factor of 55.54% from the 2 m level. From 3 to 4 m,
A at 17.2%. This is the highest oxygen content in all the oxida- the gradient declines by 54.37%, which shows that there is a
tion zones of the sampled sites. Although the first metre contains significant difference between 3 m and 4 m levels. From 4 to 5m
the highest oxygen of all the sites by far, the content drops by there is a decrease of 54.47% in oxygen, which remains gener-
38% at the second metre level. In Site B, the decline from the ally constant until the end of the 9 m sampled depth. This means
atmospheric oxygen content to the first metre in the oxidation that below a depth of 4 m, there is no longer any variation in the
zone is by 18.09% which is the lowest of all the other dams. This amount of oxygen contained by the tailings dams. The overall
suggests that the dam’s material is highly diffusive at the top of oxygen depletion gradient within the first 5 m is an average of
the dam. The significantly high percentage of oxygen at this dam 52%. Most changes occur in the top part of the dam where the

Available on website http://www.wrc.org.za 273

ISSN 0378-4738 = Water SA Vol. 33 No. 2 April 2007
ISSN 1816-7950 = Water SA (on-line)
availability of oxygen and water is high. FORSTNER U and SALOMONS W (eds.) (1988) Environmental Man-
Oxygen in the studied tailings dams varies from 0.36% in agement of Solid Waste. Springer-Verlag, Berlin and Heidelberg,
the reducing zone to 17.2% in the oxidation zone. Considerable Germany.
GOLDFIELDS, DRIEFONTEIN GOLD MINE, GEOLOGY DEPART-
oxygen is only found in the top 4m of the Witwatersrand tail-
MENT (2004) Geology of Driefontein Gold Mine, Unpublished.
ings dams. There is no significant variation in oxygen content JAYNES DB, PIONKE HB and ROGOWSKI AS (1984) Acid mine
between 4 and 10 m. The flux of oxygen into the tailings is con- drainage from reclaimed coal strip mines, 2. Simulation results of
trolled by diffusion in the gaseous and aqueous phases through model. Water Resour. Res. 20 243-250.
pores in the tailings material. LINDVALL M and ERIKSSON N (2003) Investigation of weathering
Oxygen flux is strongly controlled by diffusion in the gase- properties of tailings sand from Boliden’s Aitik copper mine, Swe-
ous and aqueous phases through pores in the tailings material den – A summary of twelve years of investigations. Proc. ICARD
from the atmospheric content. This means that as long as the 2003, 6th Int. Conf. on Acid Rock Drainage. Cairns, Australia, July
14-17, 2003. 725-731.
tailings materials of the Witwatersrand remain sandy silt to silty
NENGOVHELA AC (2006) A Comparative Study of the Development
sand, oxidation zone development will be limited to the top 3 and Characterisation of Oxidation Zones of Witwatersrand Gold
m. This point shows that the tailings material that is prone to Mines’ Tailings Dams. Unpublished M.Sc. thesis, University of
oxidation and consequently acid mine drainage generation in Venda.
the Witwatersrand tailings dams is limited only to the first 4m. NICHOLSON RV GILLHAM RW CHERRY JA and REARDON EJ
Since some of the dams studied are older than 30 years, it can (1989) Reduction of acid generation in mine tailings through the use
be concluded that oxygen diffusion in tailings dams is limited to of moisture-retaining cover layers as oxygen barriers. Can. Geo-
the top 4m of a tailings dam. This is because, as was observed in tech. J. 26 1-8.
PANTELIS G and RITCHIE AIM (1991) Macroscopic transport mech-
Site E for instance, that despite having been decommissioned 10
anisms as a rate-limiting factor in dump leaching of pyretic ores.
years ago, it had a similar oxygen profile to a dam such as Site Appl. Math. Model. 15 136-143.
A that was decommissioned 50 years ago, for instance, Site A. RITCHIE AIM (1994) Sulfide oxidation mechanism – Control and
The age of a tailings dam does not appear to have a significant rates of oxygen transport. In: Jambor JL and Blowes DW (eds.) The
bearing on the severity of oxidation zone development and sub- Environmental Geochemistry of Sulfide Mine-Waste. Mineralogical
sequent acid mine drainage. Association of Canada, Short Course Handbook 22. 210-245.
RITCHIE AIM (2003) Oxidation and gas transport in piles of sulfidic
Acknowledgements material. In: Jambor JL, Blowes DW and Ritchie AIM (eds.)
Environmental Aspects of Mine Waste. Mineralogical Association
of Canada, Short Course Series, Volume 31. Vancouver British
This paper resulted from a research project funded by the Colombia. 77-94.
Water Research Commission of South Africa to Pulles Howard SHAW SA (2004) Determining the Long Term Persistence of Mercury
de Lange Inc (PHD) (now part of Golder and Associates). The Releases to the Environment from Cyanide-Rich Gold Mine Tail-
first author has also used the project for his M.Sc. study in the ings. Unpublished M.Sc. Thesis. The University of New Brunswick,
Department of Mining and Environmental Geology, University New Jersey, US.
of Venda. The authors also wish to acknowledge Professor Bao- WERNER K (2000) Soil-Cover Remediation of Mill Tailings Deposits:
jin Zhao (University of Fort Hare, Geology) for his contribution Effects on Oxygen Transport and Hydrological Conditions. Licenti-
ate Thesis. Division of Water Resources Engineering, Department
to the research.
of Civil and Environmental Engineering, Royal Institute of Technol-
ogy, Stockholm, Sweden. 111 pp.
References YANFUL EK (1993) Oxygen diffusion through soil covers on sulphidic
mine tailings. J. Geotech. Eng. 119 1207-1228.
DAVIS GB and RITCHIE AIM (1986) A model of oxidation in pyretic YIBAS B and NENGOVHELA C (2005) Oxygen profiles in gold tail-
mine wastes, 1, Equations and approximate solutions. Appl. Math. ings dams (Report to the WRC Deliverable 4): Development of oxi-
Model. 10 314-322. dation zones in tailings dams with specific reference to the Witwa-
ELBERLING B, NICHOLSON RV and DArren DV (1993) Field eva­ tersrand gold mine tailings dams. WRC Project No. K5/1554.
luation of sulphide oxidation rates. Nordic Hydrol. 24 (5) 323-338.

274 Available on website http://www.wrc.org.za

ISSN 0378-4738 = Water SA Vol. 33 No. 2 April 2007
ISSN 1816-7950 = Water SA (on-line)