The Amanzi concept for acid mine drainage
R E Robinson
Turning environmental threats into sustainable development projects
AMANZI is the African word for water, and is also the code name given to a
project under the auspices of the Rand Water Board (RWB) to develop a process to
handle the Acid Mine Drainage (AMD) problem.
AMD is the international name for a polluted effluent from base metal, coal
and gold mines. It contains sulphuric acid and toxic metals formed by bacterial
oxidation of mineral sulphides using oxygen dissolved in the underground water
or rain and surface water in contact with waste dumps. The South African gold
mines are prolific producers of AMD and even if a mine has ceased operating, the
formation of AMD continues underground and on surface. On the Witwatersrand gold
mining area alone, it has been estimated that approximately 500 megalitres per
day have to be pumped from underground and discharged into surface water courses
to prevent seepage back into the mine. The underground water is in vast
interconnected reservoirs and the hydrology is complex. If a mine ceases
operations, the AMD might flow into an adjacent operating mine or can rise to
dangerous levels may overflow into the shallower fresh water reservoirs that
feed the springs and rivers in the Vaal River catchment basin.
Currently AMD is treated by adding lime to neutralize the acid and
precipitate the heavy metals as hydroxides. These can be flocculated forming a
High Density Sludge which once settled can give a relatively clear overflow,
which is discharged to the rivers (the HDS process). This is far from being
satisfactory; firstly because the overflow contains dissolved solids (tds),
mainly calcium sulphate, which is well above the specifications for domestic
water. Secondly, the sludge, which often contains radioactive solids, has to be
disposed of in "safe" areas thus representing a long-term
environmental hazard.
Some of the mines on their own account have established pilot plants to treat
AMD more satisfactorily (e.g. the GYP-CIX process at the Western Areas Gold
Mine). The working group of the AMANZI team selected a reverse osmosis (RO)
process of French design, for a prolonged pilot plant test at the Randfontein
Mine. The tests were completed several years ago and pronounced as being
technically successful, but the AMANZI project has never been implemented on a
commercial scale. Full details are not available in the public domain but it is
believed that the cost of treatment was, like the best of the other processes,
of the order of R5 per cubic meter (m3), excluding capital amortization. This
cost, (in total R 2,5 million/day) would bankrupt the few remaining operating
mines. The RWB and the various government departments involved disclaimed any
responsibility for such expenditure. Thus the existing practice continued in
spite of the environmental mishaps and threats and the high hidden cost of fresh
water needed to create a 'bleed' from the recirculated water in the Vaal
catchment area with its high total dissolved solids content.
The complacency is now shattered by an investigation reported by South
African scientist Dr Peter Wade in the March
issue of Science in Africa. They found
high levels of radioactivity in the sediments of the Mooi River, which is fed
from the Wonderfontein Spruit, a fresh water spring close to the Randfontein
mine with known uranium (and thus the daughter product, radium) contaminated
AMD. Action must be taken and the challenge is to find the best options to
prevent even more serious pollution.
Two exciting developments are highlighted which indicate that this crisis may
be turned into a remarkable opportunity.
The first relates to the GYP-CIX process, which uses the well-known cation
and anion exchange resins to absorb from the AMD, cations (e.g.Ca++) and anions
(SO4--) by exchanging them for hydrogen and hydroxide ions respectively. This is
a well-known method to de-ionize solutions. When the resins are fully loaded
with the pollutants, they have to be regenerated with an acid and an alkali
respectively. Conventionally, sulphuric acid and lime are used because of their
low cost, which nevertheless represented 85% of the total operating costs of the
GYP-CIX process. In a recently published paper it was proposed that nitric or
phosphoric acid be used for the cation resin regeneration and ammonia or
potassium carbonate or hydroxide for the anion resin. Although the cost of these
reagents is much higher than the conventional ones, the regenerant effluents can
be reacted with each other to form the well-known fertilizers, ammonium and
potassium nitrates and phosphates. After removing the precipitated calcium
sulphate a range of concentrated fertilizer solutions are obtained with the same
intrinsic value as the industrial products which are invariably made by direct
reaction between these same acids and alkalis.
In effect the regeneration of the resins is achieved at close to zero cost,
and with no significant changes to the basic engineering features of the proven
process. The total operating costs reduce to < R1.0 / m3. The modified
process, known as the FERRIX Process offers other advantages such as reduced
capital costs and a method of incorporating radioactive sludge into insoluble
blocks that can be deposited underground in abandoned areas of the mine. The
product water is ideally suited for agricultural use.
The second development of immense significance was the establishment of an
agricultural activity at Western Areas Gold Mine using partially treated AMD to
which fertilizers had been added to grow high value fruit and olive crops using
sophisticated control technology with drip irrigation. This activity is still
operating successfully on waste land surrounding the mine. Phenomenal value
outputs are achieved of the order of > R100 000 per ha, equivalent to an
added value potential to the water used of >R20 per m3. Can these remarkable
results be achieved on a large scale? Let us compare them with some of the
latest concepts in Australia.
Richard Pratt, the entrepreneurial head of a multibillion rand industrial
empire, spurred by a devastating drought in Queensland has proposed that water
be stored in the huge aquifers in Northern Australia during flood times and then
pumped to the agricultural areas in times of drought. It would be an enormous
undertaking, which will cost many billions; but exciting because tens of
millions of dollars have been spent by his research staff in examining the
economic feasibility. This boils down to the viability of paying about $A5 for
irrigation water. His scientists have shown that conventional spray irrigation
methods are only about 15% efficient in terms of the proportion of water
actually used in promoting plant growth. By using advanced computer controlled
drip irrigation the efficiency can be increased five times. And the cost of
pumping water over long distances can be justified in terms of the added value
potential for crops. Pratt's figures correspond well with what has been achieved
at Western Areas Gold Mine.
There is a wealth of research still to be done before such concepts can be
seriously considered, but the full significance of what might be achieved has
not been fully appreciated in South Africa.
Some simple conceptual calculations will illustrate the possibilities.
The 500 megalitres/day of treated AMD pumped from the Witwatersrand gold
mines could provide an agricultural turnover of the order of >R2 billion per
annum. Of this, probably 50% would be employment related to give, say, a
spectrum of at least 100 000 jobs allowing a factor of 2 for peripheral
activities. Land area needed for such farms would be ca 20 000 to 50 000 ha,
which in magnitude is equivalent to a single piece of land <30 km in
diameter, and way below any "land grab" requirement! The FERRIX
process is not particularly sensitive to scale factors so that many localities
could be selected for a plant and associated agricultural activity.
There are similar quantities of AMD from the coal mining areas around Witbank
and Middelburg and from the mines of the Carltonville, Klerksdorp and the Free
State Gold mining areas. The versatility and very low cost of the modified ion
exchange process allows considerable scope to pump the appropriately treated
water to the most suitable locations. Water stored during flood periods can be
stored in the vast dolomitic caverns of Gauteng and NW Province and treated by a
modified FERRIX plant. Excess fertilizer could be distributed to surrounding
conventional farming activities. Nor does the land have to be prime arable
farmland. It can be mountain terraces (for, say, vineyards and olives). It can
be bushveld or even semi-desert areas.
Capital expenditure would be of the order of R1 to R10 billion depending on
the character of the land, the infrastructure and the marketing facilities
needed. The water treatment plant would be only a fraction of this capital.
This article can in the space available deal only with the basic concept.
There is a vast scope for a national multidisciplinary evaluation and research
effort and evolutionary development. There is great antagonism towards the use
of water for irrigation and this is understandable if this is using outdated
techniques.
It can only be hoped that scientific entrepreneurship will prevail to take
further a uniquely large opportunity to generate real sustainable development.
|
