GEOCHEMISTRY AND ISOTOPIC CHEMISTRY OF ACID ROCK DRAINAGE

GEOCHEMISTRY AND ISOTOPIC CHEMISTRY OF ACID ROCK DRAINAGE AND THE EVALUATION OF PYRITE OXIDATION RATE AT MINE DOYON, QUÉBEC, CANADA Ondra Sracek 1, René Lefebvre 2 1 Palacký University, Olomouc, Czech Republic 2 Université Laval, Québec, Canada

Mine Doyon pile - description(1) • Mine Doyon is a gold mine, located north of Montreal, in Abitibi region, • Average height of south pile is 30 -35 m, rocks are sericite schists with pyrite content up to 7. 0 wt % (locally 9. 0 wt %), initial calcite content about 2. 0 wt % has been generally depleted, • AMD appeared 2 years after construction of pile and reached steady-state in 1988; typical p. H values are about 2, 0 and concentrations of sulfate in pore water are > 200 g/L; under such low p. H conditions, behavior of Al and Mg can be considered conservative in long term.

Mine Doyon pile – description (2) • Behavior of sulfate in the pile is not conservative due to the precipitation of gypsum (Ca. SO 4*2 H 2 O) and K-jarosite (KFe 3(SO 4)2(OH)6), • Precipitation of these minerals occurs within metamorphic foliation, with resulting decomposition of rocks and increased access of oxidants to unoxidized pyrite; this process increases greatly pyrite oxidation rate;

Mine Doyon pile: sampling and mesurement sites

Conceptual model of processes in Mine Doyon pile (1) Oxygen diffusion Waste rock Baserock (2) Production and conduction of heat (3) Infiltration of water (4) Oxygen convection

Two types of oxygen transport are evident in temperature and O 2 profiles • Convection oxygen transport close to slope (Site 6) • Diffusion oxygen transport in the core (Site 7)

Profiles of dissolved species concentrations

SI values for K-jarosite and Si concentrations

Isotopes D and 18 O a) Dry period (Summer 1995) b) Recharge period (Spring 1996)

Determination pyrite oxidation rate (POR) There are 3 principal possibilities in waste rock piles: (1) Interpretation of temperature and oxygen profiles (TOP) (2) Mass balance for sulfate/pyrite (SMB/PMB) (3) Oxygen consumption method in laboratory (OCM) Experiments in different scales generally produce different results (Malmström et al. , 2000).

(1) POR from T and O 2 profiles (TOP) • When O 2 is dominated by diffusion, O 2 concentration profile is more or less exponential, • Analytical solution of Crank can be applied to determine transport parameters, • In Mine Doyon pile, O 2 profile at Site 7 indicates O 2 transport by diffusion, • At this site analytical solution works for O 2 profile, but not for T profile - there probably is lateral heat transport, • Calculated POR at Site 7 is 4. 4 x 10 -10 mol(O 2)·kg-1·s-1,

A comparison between calculated and measured profiles of T and O 2 – Site 7

(2) POR from mass balance for sulfate/pyrite (SMB/PMB) • Oxidation of pyrite for p. H < 3. 0 is described as: Fe. S 2 + 3. 5 O 2 + H 2 O Fe 2+ + 2 SO 42 - + 2 H+ (oxidation of 1 mol of pyrite produces 2 moles of sulfate), • When behavior of sulfate is conservative, the amount of sulfate flushed out of the pile can be converted to oxidized pyrite, e. g. sulfate mass balance (SMB), • As an alternative, pyrite mass balance (PMB) can be used, assuming known initial pyrite content and pyrite content at time zero; calculated POR is an average value for whole oxidation period.

Role of secondary minerals in mass balance calculations • At Mine Doyon, precipitation of gypsum, Ca. SO 4. 2 H 2 O, and jarosite, KFe 3(SO 4)2(OH)6, in the pile affects mass balance for sulfate (SMB) and values of POR are underestimated, • For this reason, pyrite mass balance (PMB) was a better option.

(3) POR from oxygen consumption method (OCM) • Sample of a broken waste rock is placed into a closed chamber, sprinkled with water and changes in headspace oxygen concentration are recorded, • A graph of voltage as a function of time is plotted and straight line in the graph is used to calculate POR in mol(O 2)·kg-1·s-1, • Scale of the test is small compared to field and POR values are generally high (oxygen transport limitations are not present);

Mesurement of oxygen consumption Rock Sample

Voltage vs. time plot • Fresh material from Mine Doyon, size of particles 4. 5 cm:

Comparison of various methods of POR [mol(O 2)·kg-1·s-1] determination Site/method Temperature and O 2 profiles (TOP) Pyrite mass balance (PMB)& Oxygen consumption method (OCM) Site 6 (slope) 3. 9 x 10 -9 2. 2 x 10 -9 n. a. 1. 8 x 10 -9 n. a. Site 7 (core) 4. 4 x 10 -10 2. 0 x 10 -9 n. a. Pile average 1. 5 x 10 -9 2. 0 x 10 -9 6. 6 x 10 -8** 3. 4 x 10 -9* Site TBT (core) - based on 5 maximum values of pyrite content in deep unoxidized zone, n. a. -not available, **- fresh material, 5 runs, *- weathered slope material, 4 runs &

CONCLUSIONS • In Mine Doyon waste rock pile oxygen convection plays an important role, temperature and oxygen concentration profiles indicate convection close to slope and diffusion in the pile core, • Concentrations of contaminants in pore water (sulfate, iron, aluminum etc. ) are extremely high close to slope an in dry period and decrease in recharge, high concentrations are linked to internal evaporation of pore water in the pile due to high temperature (up to 67º C), • Internal evaporation is also confirmed by enriched values of isotopes D and 18 O, • Principal secondary minerals are gypsum and jarosite, their precipitation within metamorphic foliation breaks rocks and facilitates the access of oxidants to the surface of unoxidized pyrite,

CONCLUSIONS (continuation) • POR values obtained by different methods are comparable, • PMB method with a minimum precision can still distinguish between POR values close to slope (Site 6) from central zone (Sites 7 and TBT), • When oxidized material from the slope is used, results obtained by OCM are consistent with TOP results from Site 6 close to slope, • This means that when transport of O 2 is not a limiting factor, values of POR obtained from small scale experiments may give representative results; this is different from sites like Aitik in Sweden with oxygen transport dominated by diffusion, where POR values depend very much on the scale of experiments.