The Advantages of O_2

Adding pure oxygen to a conventional cyanide leach circuit has several advantages. It reduces the consumption of cyanide, lime and lead nitrate; it reduces the requirement for compressed air; and it improves recovery through more stable, consistent leaching, and through increased leach kinetics in a circuit which has insufficient residence time. Since oxygen is a much cheaper reagent, these advantages lead to considerable cost savings. They also have important implications for new gold mills since the use of oxygen is less capital-intensive.

In terms of improved recovery, oxygen benefits existing mills that experience periodic vicissitudes in recovery due to changes in incoming ore. A fluctuating day-to-day leach profile will flatten out, allowing for a net increase in recoveries.

Good examples are the Doyon and Terrains Auriferes mills near Val d’Or, Que., where gold recoveries have increased while reagent consumptions (cyanide, lime, and lead nitrate) have declined because of the introduction of pure oxygen. Implementing and operating the oxygen system, as documented here, is simple and straightforward. But more research is required to demonstrate that powerful, high-efficiency mixers increase dissolved oxygen levels complementary with oxygen. Lac Minerals and Air Products of Brampton, Ont., are considering verification of this assumption.

Air Products approached LAC in 1987 to initiate trials with pure oxygen in the leach circuit at several of lac’s facilities in Canada. Lac also uses oxygen at the Macassa mill in Kirkland Lake, Ont. (see N.M.M., May, 1990). In addition to these mills, oxygen is used for test purposes at the Holt McDermott mine, near Kirkland Lake, and the Camflo mine, near Val d’Or.

There are two requirements for oxygen in the cyanide leach mill. First, oxygen is required in the familiar gold dissolution reaction. 4 AU(s) + 8 NaCN(aq) + 0_2(aq) + 2H_2(aq) 4NaAu(CN)_2(aq) , 4Na 0H(aq)

Of the reagents on the left side of the equation, the availability of gold is fixed by the ore in question. Obviously, water is available in excess, leaving cyanide and oxygen as the limiting constituents in the process. Most mills operate with an excess of cyanide, which is closely monitored. Dissolved oxygen (do) levels, however, are not normally watched closely and find their own level based on the temperature, the availability of compressed air and the ability of the ore to consume oxygen.

Typical DO levels are in the range of three to eight parts per million (p.p.m.); however, levels less than one p.p.m. are not uncommon for refractory ores. With pure oxygen, DO levels as high as 30 p.p.m. can be achieved. (Figure 1). This greatly accelerates the kinetics of the leach reaction and may also shift the reaction further toward completion, thus increasing gold dissolution slightly.

The second oxygen consumer in the process is the so-called “refractory” minerals found in many ores. These materials are soluble in the leach solution, and they compete with the gold dissolution reaction for oxygen and cyanide. (Figure 2). In severe cases, a pre-aeration stage to oxidize these minerals is added before cyanidation. Lead nitrate and litharge have also been used to passivate soluble sulphides in pre-aeration and in leaching.

In the case of refractory ores, it is possible to reduce cyanide consumption by operating with higher DO levels. In other words, soluble minerals can be preferentially oxidized with oxygen rather than cyanide. In a pre-aeration circuit, DO levels always jump noticably at the first cyanidation tank. This is because the soluble minerals that depress DO levels in pre-aeration are now being consumed by cyanide, thus allowing DO levels to increase.

For slurry oxygenation techniques, there are three methods available. They are:

*sparging;

*side stream pumping; and

*total stream oxygenation.

Each offers advantages with regard to cost, installation and operation.

Bubble diffusion is a method of injecting oxygen through a porous medium or diffuser from which bubbles are dispersed into the slurry. Oxygen utilization efficiency for this type of system is determined by the size of the bubble produced and the hydrostatic head above the point of discharge. The advantage of bubble diffusion is its simplicity and low capital cost.

Side Stream Pumping

Since the oxygen saturation levels can be increased only with reduced temperature or increased pressure, it is desirable to inject the oxygen into a region of high pressure such as a pump discharge. With side stream oxygenation, a side stream of the slurry is pressurized to several atmospheres to enable high levels of DO content to be obtained. This super-saturated side stream is then re-injected into the agitator. When the slurry stream hits the low pressure of the agitator, oxygen comes out of solution in very small, microscopic bubbles (with a high surface-area-to-volume ratio) which then rise to provide oxygen to the rest of the slurry. With side stream pumping (ssp), oxygen utilization efficiency will be affected by the discharge depth of the agitator, the volume of the side stream relative to the main slurry stream, and the retention time provided for oxygen dissolution. A disadvantage of SSP is its higher capital cost for pumps and side stream piping. Additional costs are also incurred in operating and maintaining the pumps. This method is unsuitable in a carbon-in-leach circuit because of the obvious carbon attrition problem. Total Stream Oxygenation

A third method of oxygen injection is total stream polishing where oxygen is fed into an existing pipeline. This location is usually the feed line from a pump to the first of a series of agitators.

This method offers similar advantages of SSP in terms of high oxygen utilization efficiencies. Also, by utilizing existing piping and equipment, the installation cost is low. As with the other methods of oxygen injection, the hydrostatic head at the point of discharge affects utilization efficiency.

Total stream oxygenation is also very successful in increasing DO levels in the grinding circuit. Mill water is oxygenated prior to its addition to a ball or rod mill, thus the mill becomes a more effective dissolution vessel (assuming grinding is done in cyanide). DO levels at a grind circuit discharge are as high as 25 p.p.m., compared to a more typical 1-to-7 p.p.m. without oxygen. In addition, the oxygen requirement of this application is minimal. One widely expressed concern regarding the use of oxygen in the grinding circuit is that consumption of grinding media may increase. While this method is now in limited operation, no data are yet available regarding increased corrosion in the grinding mill.

This method should also work quite well in heap leach applications, allowing for faster leaching of a given heap. Addition of only a small flow of oxygen into the discharge of a leach solution pump would provide a great deal more useful oxygen to the heap.

Oxygen consumptions vary considerably, depending upon mill size, type of agitator, and ore characteristics. Since oxygen solubilities are reduced considerably at higher temperatures, oxygen consumption is generally greater in the summer months. In terms of oxygen storage and supply, typical volumes are such that a cryogenic oxygen storage vessel is required on site, periodically filled from an oxygen tanker. While liquid oxygen is not particularly dangerous material, it must be treated with respect and any personnel involved with it should be thoroughly informed of its potential hazards.

Lac Minerals’ Terrains Auriferes division began adding pure oxygen to its gold leach circuit in September, 1987. This mill, treating Bousquet ore at a rate of 1,650 tons per calendar day (1,500 tonnes per day), is a straight cyanidation mill using the Merrill-Crowe process for gold precipitation. The Bousquet ore characteristics require lead nitrate to inhibit cyanide solution fouling by sulphides. The initial objectives were to reduce or eliminate use of lead nitrate and continue to improve overall plant recovery. In O
ctober, 1988, oxygen was implemented at the Doyon mill, a 3,450 ton-per-day (3,135 tonne-per-day) operation employing leach/carbon-in-leach/ leach/carbon-in-pulp.

High consumption of lead nitrate is noted at the Bousquet and Doyon operations. At Doyon, consumptions as high as 0.6 lb. per ton have been experienced. The following are some initial problems that we were facing prior to lead nitrate addition:

*During flowsheet elaboration of Doyon ore, maximum attainable recovery was 85% and roasting was even considered. Finally straight cyanidation (leach/filter/Merrill-Crowe) was selected with 72 hours leach time and a 95% minus 200 mesh grind objective. Milling in these conditions gave a recovery fluctuating from 88% to 92.75% with associated solid tails ranging from 0.005 to 0.02 oz. per ton (0.17 to 0.69 grams per tonne). In general, pushing more tonnage resulted in higher tails, and on occasion the tails were exceptionally high.

During these periods, we encountered the following problems:

*Cyanide concentration (titration) in the circuit was dropping to trace levels in the thickener underflow and primary leach tanks. Despite massive additions of cyanide in the tanks or thickeners, we were unable to regain control of the circuit. The throughput was then reduced and reagent dosage increased.

*Because it was impossible to predict when these periods would arrive, a daily leach profile was done on the circuit. We observed when we were losing cyanide concentration that a strange metallurgical phenomenon that we call “reprecipitation” occurred. In fact, for example, gold concentrations in solid residues in the thickener underflow were higher than the cyclone overflow feeding the thickeners. Gold dissolved in the grind circuit was precipitating in the thickeners. Sometimes the gold in solids in the primary agitator (#1) was higher than the thickener underflow which was feeding it. Sampling and assaying confirmed the occurrence of this metallurgical phenomenon. We also noted, at the assay laboratory, occurrence of a blue precipitate in the mill solution acidification process for assaying which showed the occurrence of dissolved iron.

*Between the primary and secondary leach, we had an intermediate filtering stage. During the problem periods, we often observed that the residues of the secondary #1 leach tank were considerably lower than the last primary agitator. In fact, high leach kinetics was occurring on the filters. We consequently arrived at the following observations during these periods:

A premature dissolved gold precipitation phenomenon was occurring onto the solids which, under specific chemical conditions, was reversing the normal leach process.

The drum filters were becoming more efficient dissolution units than the agitators.

Cyanide consumption of up to 3.5 lb. per ton and lime to 10 lb. per ton were observed. We saw solid tail residues up to 0.04 oz. per ton (1.38 grams per tonne), in which most of the gold was in fractions finer than 325 mesh. Solution losses at the last filtering stages, because of leaching on filters, was as high as 0.007 oz. per ton (0.24 grams per tonne). We were always able to re-establish normal conditions but at the cost of reduced throughputs and higher reagent dosages. An investigation into these occurrences revealed the following:

*Sequential pit mining demonstrated that problematic milling periods were associated with two specific mining blocks which proved, at later dates, to have more geological alterations and higher contents of chalcocite, tellurides such as calaverite, secondary iron oxides, and some metallic iron, copper and sulphur.

*Adequate lead nitrate dosage to the circuit curtailed considerably the negative effects of contaminants and helped us in reaching a much more stable operation. It also lead to revised operating criteria such as:

*grinding fineness: to 78% minus 200 mesh from 95% minus 200.

*residence time: to 50 hours from 72 hours, therefore an increase in throughput.

*recovery stability (no more evidence of marked ups and downs) and an overall gain of about 1% in recovery.

However, to achieve this, we had to develop a control strategy for lead nitrate addition. Reducing power was not precise enough. The Prussian Blue test proved to be the answer. Lead nitrate addition points were also critical in order to get the desired result in the process as soon as possible but also to avoid interfering with filter press operation. A high lead nitrate ratio in the grind may alter the efficiency of the Merrill-Crowe circuit and, in the worst case, reverse the precipitation reaction. With what we have learned to date, we try to maintain lead levels in solution below two p.p.m. It was also demonstrated that during milling of these zones, without proper lead nitrate control, dissolved oxygen in the circuit fell as low as trace levels. With lead nitrate addition, oxygen contents are higher but not to where we normally want them.

We believe lead nitrate helps the process for the following reasons:

*it precipitates sulphides as a lead sulphide precipitate;

*it acts as a strong oxidizing agent which influences leach kinetics; and

*it cleans the surface of “tarnished gold” because of sulphide deposition.

Despite the positive impact of lead nitrate, there are also some drawbacks, as follows:

the potential health hazard of working with certain quantities of lead;

lead fumes in precipitate refining;

susceptible world markets;

metallurgical concerns (Lead additions can be misjudged, therefore affecting leach characteristics); and

possible gold loss at the filter press operation. Oxygen Addition

We all know Elsner’s gold dissolution equation. Oxygen is the key element to dissolution. When we started to use pure oxygen on bench tests, we first observed that dissolution kinetics was increased considerably and that lime and cyanide consumptions were reduced. These results convinced us that we should experiment with oxygen addition on a full-scale operation.

Oxygen additions were tried at various points in the plant, taking advantage of the high pressure (and resultant high oxygen solubility) of the available slurry lines.

At Doyon, oxygen was added to the #1 tank in the secondary leach circuit. Air spargers to this tank were shut off and, during winter months, oxygen levels well above 20 p.p.m. were maintained (slurry temperature: 4/7ch/C). Oxygen levels remained above 10 p.p.m. in tank #2 through #5 compared with two to four p.p.m. prior to oxygen addition with high levels of lead. This altered the leach profile of the secondary circuit considerably. Before oxygen addition, solid residues in #1 secondary leach tank would go as high as 0.1 oz. per ton (3.43 grams per tonne) while now, with oxygen, they vary between 0.045 and 0.0095 oz. per ton (betwen 1.5 and 0.32 grams per tonne). Before, oxygen addition leaching in the case of higher-grade ore was occurring in the carbon-in-pulp (cip) circuit while, with oxygen, this phenomenon is reduced considerably. In fact, most of the time, leach is now completed in #4 secondary leach tank (Figure 3).

It is also interesting to mention that at Les Terrains Auriferes mill, oxygen addition has been tried in the grinding circuit; cyclone overflow showed, despite transfer boxes and cycloning action, higher dissolved oxygen level. With the current understanding we feel that the best addition points would be the following:

*Oxygenation of mill solution at the pump discharge to supply the grinding mills.

*Touch up oxygen in first leach tank of primary and secondary circuits and possible replacement of compressed air to other leach tanks depending on process economics.

*In mill design, we can avoid air in the CIP circuit to reduce carbon attrition. However, we always observe some slight leaching in these tanks because of residual DO and also the fact that the solution loading in gold is decreasing with CIP contact time. This gold depletion seems to enhance further leaching. Oxygen is then an ideal element. We can increase DO levels without increasing carbon attrition because
oxygen volume requirements will be relatively low. Therefore, the storage capacity of the tanks will be improved.

Since implementing oxygen, we have been able to reduce lead nitrate consumption by 55%. We tried cutting it out completely but experienced some problems. Our work continues; however, it is impossible for us at this stage to confirm whether we will be able to eliminate it completely. Because of a more stable leach curve, we then tried cutting cyanide addition. We have cut cyanide so far by about 25%. Overall cost trend is then favorable by about $0.65 per ton, the difference between the cost of oxygen versus savings for cyanide and lead nitrate.

Another advantage is the fact that once the leach profile is well established and is less susceptible to head grade variation, operators will find it easier to evaluate the impact on the process of other variables or adjustments for process optimization such as grind fineness influence on tails, throughput or residence time, etc.

So far, the only drawback is in the efficiency of oxygen utilizaiton during summer months. It is a basic law of chemistry that the higher the temperature of a solution, the higher the dissolution kinetics will be. However, the higher the temperature, the lower the attainable dissolved oxygen level. To maintain 15 p.p.m. in a 35/7ch/C solution requires five times the oxygen consumption required in a 4/7ch/C solution. Experience will dictate requirements. In our case at Doyon, we aimed for DO levels of about 10 p.p.m. last summer, which required two or three times as much oxygen as did winter operation.

In terms of overall recovery, we increased our recoveries by about 1% using oxygen. Oxygen consumption is about two tons (1.8 tonnes) per day. Recently, more oxygen addition points were added to the circuit at Doyon. An oxygen line was added to the mill solution feed to the ball mills, however it has not yet been used. A side stream pump arrangement was added to the number one primary agitator, and while this is in operation no significant data have yet been collected.