The role of phosphate in applied biotechnology in ...

1998 Kalin, M., M.P. Smith and A. Fyson, "The Role of Phosphate in Applied Biotechnology in Mine Waste Management: Reduction in AMD from Pyritic Waste Rock" Proceedings of the International Symposium, the Metallurgical Society of the Canadian Institute for Mining, ‘Waste Processing and Recycling’, Calgary, Alberta, August 16-19, pp. 15-29.

15

The role of phosphate in applied biotechnology in mine waste management: Reduction in AMD from pyritic waste rock M. Kalin, M.P. Smith and A. Fyson Boojum Research Limited 468 Queen Street East Toronto, Ontario M5A 1T7 Tel: (416) 861-1086

ABSTRACT

This work examined sedimentary phosphate rock additions to pyritic waste rock in order to reduce the amount of acid generated. This research was initiated when phosphate (2.5%) was identified in precipitate scrapings from rock surfaces which originated from a section of a copper leach dump of the Gibraltar Mine in British Columbia, Canada, where acid generation and copper recovery had ceased. These findings provided the rationale for experimentation with pyritic waste rock from a zinc/copper mine, where Natural Phosphate Rock (NPR, 115 kg·fl waste rock) was added to 60 L of waste rock in 70 L drums and exposed to weathering for three (3) winters. The chemistry of the effluent from the drums was monitored over a period of 989 days. In the presence of NPR, the cumulative acidity discharged was 74% to 90% less than that discharged from drums without NPR, with the exception of one drum. Mixing NPR throughout the drums appeared more effective in reducing the acidity ~eleased than placement of the NPR throughout only the upper half of the drum. Better effluent quality appeared to emerge from drums with waste rock which was 'fresh' (not exposed to winter conditions), in comparison to waste rock which had weathered for more than 4 years. The results suggest that sedimentary NPR integrated into pyritic waste rock may present an avenue to greatly reduce the acidity generated from waste rock piles. The application rates of NPR that achieve the reductions have not been defined, but appear to be related to the surface area of pyritic material exposed to weathering.

Waste Processing and Recycling III Edited by S.R. Roa, L.M. Amaratunga, G.G. Richards and P.O. Kondos The Metallurgical Society of elM, 1998


16

WASTE PROCESSING AND RECYCLING III

INTRODUCTION Waste rock dumps are significant sources of acid and metals which are generated by the oxidation of iron sulphides (most commonly present as pyrite (FeS 2) and as pyrrhotite (Fe1_xS»Oxygen and water supply are the factors controlling the rate of oxidation, and oxygen supply is the reaction-limiting component Rapid pyrite oxidation occurs at low pH due to direct oxidation of pyrite by dissolved ferric iron, which is increasingly soluble below a pH level of 3.5. At low pH, bacterial activity accelerates the oxidation of ferrous to ferric iron by a factor greater than 106 (1). The generation of AMD can be viewed as a corrosive process, taking place on the pyrite surface. Phosphate is an agent used to curtail corrosion and Fe(III) oxides chemisorb phosphates on their surfaces. The main sinks of phosphate in nature are the iron rich sediments oflakes and oceans. AMD environments are dominated by the effects of oxidation of reduced iron leading to low pH, where iron phosphate is stable and precipitates. Phosphate can be expected to be key in reducing both corrosion of the pyrite surface and precipitation of iron. Ifprecipitates could be induced and coatings could be formed on the surface of the rocks, the oxidation or weathering of pyrite would be reduced and, therefore, acid generation resulting from the hydrological transport of the oxidation products from the rock surfaces would be curtailed or inhibited. It is general practice to neutralize the acid generated from the weathering process, creating metal-laden hydroxide sludges which require disposal. Neutralizing agents such as limestone do not prevent the pyrite from oxidizing, but treat the resulting acid mine drainage (AMD) or acid rock drainage (ARD). Reducing the access of oxygen to pyritic surfaces slows down the rate of the weathering process and reduces the environmental impact Waste management practices integrate sub-aerial deposition of tailings in order to reduce the acid generation rate. Meek (2) determined in pilot-scale experiments that apatite (Cas(P0 4»)OH) mixed with tailings was more cost effective than either liming or using PVC liners to reduce oxygen and water penetration into the tailings. In their laboratory work on the application of natural phosphate to acid-generating coal wastes, Hart et al. (3) and Hart and Stiller (4) report that NPR serves as an effective amelioration agent. For waste rock piles, submergence of the waste rock in mined out open pits is costly. Moving a ton of waste rock is estimated at $0.50, depending on conditions. In addition, it is frequently impossible to re-submerge the waste rock in the pits, since the ore-to-waste ratio is greater than 1, with a typical swell factor of 1.3 for broken rock. In waste rock piles, chimneys or oxidation pockets often develop where, due to the exo-thermic reaction of pyrite oxidation, temperatures increase significantly, and steam emerges from the pile. Such chimney or pocket development suggests that reducing acid generation by inducing coating with precipitates has to be focussed on those areas of the waste rock pile where the hydrological pathways have formed, generating the ARD toe seepages. The coating- or precipitate-inducing product has to be a substance which is easily weathered, such that it can be transported with atmospheric precipitation (rain) to those locations in the pile where rain drives the acid generation process. The NPR used in the experiments is a low phosphate waste product from a sedimentary phosphate deposit in North Carolina. The kinetics, thermodynamics and chemical equilibria of phosphate-iron interactions in natural environments are complex, as is evident from perusing several chapters of Stumm and Morgan (5). It followSthat the chemical reactions which may lead to precipitate formation at the pyrite surface are likely to be equally complex. Discussions of the potential reactions which lead to phosphate precipitation are speculative and empirical in nature, and make the assumption that iron-



17

phosphate reactions are dominant, although many other factors playa presently-undefmed role. Spotts & Dollhopf (6) describe potential reactions of phosphate in acid generating material, and suggest that phosphate may create the chemical conditions on sulphide mineral surfaces which prevent further oxidation, rather than form a reactive compound which precipitates the products of sulphide oxidation. According to Evangelou (1,7), phosphate reacts with dissol ved ferrous and ferric iron to form vivianite (FelP04)2.8H20) and strengite (FeP0 4.2H20), respectively. Through this precipitation, very low dissolved iron concentrations are maintained, thereby minimizing the direct oxidation of pyrite by ferric iron. He further postulates that direct adsorption of phosphate molecules onto pyrite surface iron atoms takes place, which eliminates electron transfer between pyrite and oxidizers. This would represent a further mechanism which would lead to inhibition of pyrite oxidation. The formation of iron phosphate and iron hydroxide precipitates over the pyrite surfaces, and the associated reduction in the rate of oxygen diffusion to the pyrite surface, has been reported by Ziemkiewicz (8) as physical isolation of the pyrite surface. Such a physical isolation of the surfaces corroborates the description by Scott (9) of secondary mineral formation in the dump leach pile at the Gibraltar mine, suggesting a link between oxidation and the mineral formation process. The experiments with NPR have been described previously by Kalin et al. (10), giving the effluent characteristics of the drums with the additions ofNPR to the waste rock for a period of 709 days. Effluent from the drums was collected for a further 280 days and the experiment was terminated after 989 days. The waste rock drums were disassembled, and the pyrite surfaces on the rocks were quantified along with the remaining NPR. The rock sizes were quantified for each drum. The waste rock from the experiment is presently stored indoors to further analyse the ongoing processes, and to test the reversibility of the inhibition. This paper summarizes the effluent characteristics in relation to some of the quantified parameters of the rocks in the drums after exposure to weathering for 989 days.

MATERIALS AND METHODS

Sampling and Experimental Set-Up The waste rock samples used in the experiment originated from a base metal mine located in northern Quebec, where the principal hypogene minerals are pyrite, sphalerite and chalcopyrite. With the assistance of the mine geologist, five rock types were selected for the study which reflected the operational classification of the waste rock of the site. Type A was described as fresh (defined as broken from the pit and not having weathered for one winter), low pyrite waste rock which formed the majority of the waste rock pile proper. Type B was fresh, high pyrite waste rock which was segregated and buried in the tailings pond, following the principle of sub-aerial deposition. Type C was low pyrite waste rock which was collected from the waste rock pile and had weathered for four or more years prior to sampling. Type D consisted primarily of dacite tuff, classified as nonacid generating material. It was generally used as fill material for construction in the waste management area. Type E waste rock was low grade ore containing sphalerite «Zn,Fe)S), chalcopyrite (CuFeS 2) as well as high pyrite, which was stockpiled since mine deVelopment and had weathered for 4 or more years. The size of rocks collected were generally in the range of O.Olm to 0.25m in diameter. Si}{ty litres of waste rock were set up in 70 L drums outdoors in August, 1992. The waste rock in each drum was held 0.1 m above the drum bottom by a styrofoam block in order to facilitate drainage (a schematic diagram of the set-up is presented in Kalin et al. (10». Three drums were set


18


up for each of the following rock types. A single drwn containing 70% non-reactive rock was mixed with 30% low pyrite, to simulate the operational reality of the classification of waste rock in a mining operation. Segregation of waste rock is prone to error, in that the occasional truck ends up putting a load of rock in a pile when it should have been placed in another pile. This 30% blend was described by the on-site geologist as being representative of both waste rock excavated from the pit and unsegregated waste rock piles. Natural phosphate rock (NPR) was obtained from TexasGulf (Raleigh, North Carolina). Code 48 NPR is the coarse by-product (0.001 m to 0.01 m diameter), with its main constituents being calcium phosphate (36%) and calcium carbonate (48%). NPR (Table I) is a product remaining after the finer rock is screened out for use as fertilizer plant (11).

Table I - Composition of TexasGulf Code 48 Natural Phosphate Rock

Element P S Ca Mg Na Al

Assay % 7.1 0.8 31.3 0.3 0.5 0.3

Probable Compound Ca)(P04h CaS04 CaCO) MgCO) Na2CO) AI(OH»)

% 35.8 3.4 47.9 1.0 1.2 0.7

Element K Fe Si Sr B F

Assay % 0.1 0.5 0.01 0.2 0.03 1.8

Probable Compound KOH Fe(OH») Si02 srCO) B2O) CaF 2

% 0.2 1.0 0.03 0.3 0.2 3.7

Three series of tests were conducted. Drwns A-I, B-1, C-1, D-l and E-l (series 1) did not recei ve applications of natural phosphate rock and, therefore, served as contro Is. In drums A -2, B-2, C-2 and E-2 (series 2), a fibre glass screen (0.001 m mesh size) was placed horizontally over the styrofoam block in the drum prior to the placement of waste rock. NPR added to the surface could trickle throughout the drwn, referred to as the mixed application. After half the waste rock was placed in drwns A-3, B-3, C-3 and E-3 (series 3), a fibreglass screen was positioned over this rock before the second half of the drwn was filled with waste rock. NPR in this series was added to the surface of the drwn, but could only distribute half-way down the drwn, simulating placement of the NPR in a waste rock pile on successive lifts being built. Following the placement of waste rock in the drwns of series 2 and 3, 8.2 kg of NPR, equivalent to a layer of about 0.05 m, was spread over the rock surface. This application rate is equivalent to approximately 115 kg·r l of waste rock, and was chosen because the shipping and handling costs of the NPR in relation to perpetual treatment of ARD were considered cost effective for scale up. Economic criteria were the prime considerations in choosing an application rate, and it is not related to factors affecting acid reduction. Those have to be based on the results of the ongoing experiments, such as pyrite surface distribution, hydrological pathways in the waste rock pile and others.



19

Experiment Monitoring A water sampling port was installed, extending to the bottom of the drum, and an overflow drain was positioned at the styrofoam-waste rock interface to remove excess water and to prevent flooding of the waste rock. Monitoring commenced on August 25, 1992 (day 0) when set-up of the experiment was completed. The experiment was monitored for 989 days following set-up. Water samples (0.25 L) were collected monthly from the sampling ports, exc~pt when the drums were frozen. Effluent water samples were collected on 27 occasions between September 22, 1992 and May 11, 1995. Acidity (NaOH additions to pH 8.3) and alkalinity (H2S04 additions to pH 4.5) determinations were carried out with a Brinkman Metrohm SM 702 Titrino autotitrator. Dissolved metal concentrations in filtered (0.45 J.lm) and acidified samples were determined by ICAP (Inductively Coupled Argon Plasma spectroscopy). Atmospheric precipitation and temperature (maximum and minimum) were monitored on-site for calibration to the meteorological data available from Environment Canada for the Toronto (Queens Park) station, located 2.7 km north west of the experiment set-up.

Disassembly of Drums The rocks were removed from each drum, defining five layers, and the presence of NPR grains on the waste rock in each layer were recorded, the rock characterized for appearance (surface coatings), and the diameter of the rocks were measured. Approximations of rock surface areas were made by wrapping various areas of graph paper around rocks and selecting the best fit. Visual estimates of exposed pyrite content were made using a hand lens (12). The NPR remaining in the drums was recovered, dried and weighed to determine NPR consumption.

RESULTS AND DISCUSSION Acidities, in mg'L- 1 CaC03 equivalents, were determined for 27 sets of water samples collected from the waste rock drums between August 25, 1992 and May 11, 1995. The rock type which produced the highest acidity, in the control series (A-l to E-l, no NPR added), was type C (1,052 mg'L-I, number (n)=27), the low pyrite, weathered waste rock. The lowest average acidities were produced by type E (163 mg'L- 1, n=27), the high pyrite, low grade ore. Types A and B produced intermediate acidities (426 and 443 mg'L- 1, respectively). Type D, dacite tuff classified as non-reactive with the operational component oflow pyrite additions, producing 810 mg' L -I acidity on average (Table II). These effluents indicate that the characteristics of the effluent are not proportional to the amount of waste rock and the classification, i.e., high or low pyrite, as all the drums had a similar weight of volume of rock exposed to weathering under identical conditions. Thus the application rate of NPR is unlikely to be a function of the tonnes of rock which require treatment, given the effluent characteristics of the waste rock drums. Lower acidities were measured in the effluent from the drums treated with NPR. In those drums where NPR mixed throughout the waste rock column (series 2), the average acidities did not exceed 79 mg'L- 1 (Table II). Higher average acidities were measured in those drums where NPR was added only to the upper half of the rock column (series 3; up to 250 mg'L- 1 acidity on average).


20

WASTE PROCESSING AND RECYCLING III Table II - Waste Rock Drum Effluent Acidities (mg.L- I ), August 25, 1992 to May 11, 1995

Rock Type

Unit

Control

8.2 kg Mixed NPR

8.2 kg LayerNPR

A

low pyrite fresh, weathered < 1 y

Average Minimum Maximum

426 54 1788

57 3 1102

20 7 59

B

high pyrite fresh, weathered < 1 Y


443 36 1514

18 4 47*

48 13 145

C

low pyrite weathered > 4 Y


1052 157 4411

79 14 321

210 46 874

D

70% Dacite Tuff 30% Type C


810 66 2871*

Not Set Up

Not Set Up

high pyrite weathered> 4 y


163 43 464

13 4 45

250 48 819

E

Number of samples is 27. * denotes 26 samples. Cumulative acidity generated by each waste rock drum was calculated using available meteorological data. Annual precipitation data collected on site between August 25, 1993 and August 24, 1994 (745 mm), and between August 25, 1994 and November 28, 1994 (167 mm), compared very closely (±3%) with Environment Canada atmospheric precipitation for the same periods (726 mm and 173 mm, respectively). The meteorological data were used to estimate the volume of water which entered the waste rock drums for the period of the experiment. Cumulative acidity generated by the drums was estimated by multiplying the volume of water collected by a drum over a period by the acidity in samples collected at the end of the period. The seasonal pattern of acidity release is given in Figure 1 for drums E-l (control) and E-2 (NPR mixed) for the duration of the experiment as an example. The seasonal pattern was generally similar for all of the rock types, with the greatest acidity release during the wanner periods of each year. For the onset of the experiment (August, 25 to December, 1992), acidity released was greater than'in the subsequent six winter and spring months (January to June, 1993). Between July and September, 1993, acid generation was rapid, while it slowed over the following winter between October, 1993 and June, 1994 (see Kalin et al. (10) for cumulative acidity data for all rock types). Acidity release again increased over the 1994 summer, then slowed over the 1994-1995 winter months. These seasonal patterns show that water continued to flush out oxidation products in relation to the precipitation, and the process of inhibiting or reducing the forming of oxidation products is not instantanoous, as would be expected from pure neutralisation of the effluent. It is a reduction process which corresponds to the oxidation process.


WASTE PROCESSING AND RECYCLING £II

1.2

I Aug 25-Dec Jan-Mar Apr--lun 1992 1993

Jul- SeF Oct-Dec Jan-Mar Apr--lun 1994

Jul-Sep

a: 0.8 Cl

--"".~ Cl

'0

'0

~

0.6

~

~

r

/

I!

I

I

II

'1f