Medium-Temperature Pressure Leaching of Copper Concentrates – Part II

By Marsden, J O Wilmot, J C; Hazen, N

Abstract Part I in this series of papers considered the chemistry of medium-temperature pressure leaching of copper sulfide concentrates and reviewed the metallurgical testing and initial process development for this technology. This initial work resulted in the development of a process flowsheet utilizing superfine grinding (to a P^sub 98^ of 12 to 15 [mu]m and P^sub 80^ of 6 to 7 [mu]m), pressure leaching at 160[degrees] to 170[degrees]C, followed by conventional solution extraction (SX) and electrowinning (EW) for the recovery of copper. However, this process required a significant amount of concentrated acid to be fed to the pressure leach vessel to meet the reaction requirements, representing a cost of between $0.03 and 0.05/lb copper (2006 basis) produced by the process. It was immediately recognized that the development of a process that did not require such a large acid addition would be advantageous. This paper, which is Part II in the series, describes the development of a process that incorporates medium-temperature pressure leaching and direct electrowinning (DEW) in a novel and innovative process flowsheet. Phelps Dodge (a subsidiary of Freeport- McMoRan Copper & Gold Inc.) refers to this as the "MT-DEW-SX" process. Part III in this series covers the large-scale demonstration of the MT-DEW-SX process at Bagdad, Arizona; and Part IV addresses the development and design of a full-scale commercial application at Morenci, Arizona.

Key words: Copper, Pressure leaching

Introduction and background

Part I in this series of papers considered the chemistry of medium-temperature sulfate-based pressure leaching of copper sulfide concentrates in detail and reviewed the initial process development work that was conducted in an effort to develop this technology for potential commercial application. This work was conducted on the heels of the successful development of a commercial flowsheet for applying high-temperature pressure leaching to chalcopyrite copper concentrates, which was developed under the terms of aTechnology Development Agreement between Phelps Dodge (a subsidiary of Freeport- McMoRan Copper & Gold Inc.) and Placer Dome Inc. The chronology of this development work has been reported elsewhere (Marsden et al., 2003; Marsden and Brewer, 2003).

As discussed in Part I, the initial work resulted in the development of a flowsheet utilizing the following steps (see Fig. 1):

* superfine grinding to a P9g of approximately 12 to 15 [mu]m and a Pg0 of approximately 6 to 7 [mu]m;

* pressure leaching at 160[degrees] to 170[degrees]C, 690 to 1,380 kPa (100 to 200 psi) oxygen overpressure, 400 to 500 kg/t acid addition, 10 kg/t calcium lignosulfonate (CLS) addition and 90 minutes retention time;

* solid-liquid separation;

* SX for recovery of copper from leach solution;

* EW of copper; and

* neutralization and disposal of the solid residue.

The acid contained in the pressure-leach discharge could be used beneficially for heap, stockpile and/or agitated leaching processes. Alternatively, excess acid must be neutralized, which represented a cost to the process.

Figure 1 – Simplified MT-DEW-SX process flowsheet.

It should be noted that, throughout this paper, the terms "solution extraction" and "SX" are used to refer to the process that is commonly referred to as "solvent extraction" in the industry.

The process required a significant amount of concentrated acid to be fed to the pressure leach vessel to meet the reaction requirements, representing a cost of $0.03 to 0.05/lb copper produced by the process, depending on the cost of acid ($40 to $70/ t) and the exact amount added (400 to 500 kg/t). It was immediately recognized that the development of a process that did not require such a large, concentrated sulfuric acid addition would be advantageous.

Phelps Dodge and Hazen Research staff identified several flowsheet configurations that provided the necessary supply of acid to the pressure leaching process. The preferred flowsheet considered direct electrowinning (DEW) of copper from strong pregnant leach solution (PLS) produced by pressure leaching, recycling lean electrolyte from DEW back to the pressure leaching step to supply the required acid and treatment of weak PLS by SX (referred to as the "MT-DEW-SX" process). This circuit (Fig. 1) represented a simple and elegant method to meet the acid requirements of leaching, provided that acceptable copper extractions could be achieved with a significantly higher copper tenor in the leaching step (due to copper content of the recycled electrolyte), provided the iron concentration in the pressure leach discharge could be controlled within acceptable limits, and provided that the copper could be maintained in solution (rather than crystallization as copper sulfate) in the copper-rich solution feeding DEW.

Other flowsheets were pursued and developed in some detail by Phelps Dodge and Hazen Research in 2000 and 2001 (Marsden and Wilmot, 2007), but these are not considered further here.

By early 2003, the development of the Cerro Verde (Peru) primary sulfide resource was being advanced through a preliminary feasibility study that considered large-volume grinding and flotation facilities to produce copper and molybdenum sulfide concentrates. At that time, Cerro Verde was processing secondary sulfide ore (predominantly chalcocite copper mineralization) through crushing, agglomeration, heap leaching, SX and EW This process required acid for the heap-leaching step, both in agglomeration and as make-up to raffinate and EW electrolyte. The potential benefits of an effective flowsheet to process the primary sulfide concentrates at the site and generate byproduct sulfuric acid that could be used in heap leaching were recognized.

The high-temperature pressure-leaching process, developed and implemented at Bagdad (Arizona) in a commercial demonstration scale since March 2003, generated more acid than could be used beneficially by the heap-leaching operations at Cerro Verde. However, medium-temperature pressure leaching coupled with DEW offered the chance to match the acid generated by pressure leaching with the requirements of heap leaching at the site, while also minimizing oxygen consumption and eliminating the need for concentrated acid addition to the pressure leaching step.

Also, during 2003, other potential applications of mediumtemperature pressure leaching and DEW process were identified, including the Western Copper deposit in the Morenci district, Arizona, and the El Abra primary sulfide resource in northern Chile.

The following sections of this paper cover the development of the medium-temperature, DEW and SX process flowsheet (MT-DEW-SX). The process chemistry for medium-temperature pressure leaching has been reviewed previously (Marsden et al., 2007) and is not discussed further here.

MT-DEW-SX pilot plant description

During August and September 2003, a series of continuous pressure leaching pilot plant campaigns were conducted at Hazen Research to develop and demonstrate the MT-DEWSX flowsheet (Fig. 1) applied to Cerro Verde primary sulfide concentrate. A representative sample of Cerro Verde primary sulfide concentrate was obtained from a grinding and flotation pilot plant run on large-diameter core samples from Cerro Verde. During the initial pilot plant work, Bagdad concentrate was used to demonstrate the flowsheet and material balance prior to consuming the limited amount of Cerro Verde concentrate that was available for testing. Mineralogical and chemical analyses for the various concentrates are provided in Tables 1 and 2.

The pilot plant process consisted of superfine grinding of the concentrate toatarget size of 98% passing 15 urn, followed by pressure leaching at 160[degrees]C for 90 minutes with continuous oxygen sparging. The pressure leach discharge – containing copper, iron and acid in solution and a mixture of hematite, elemental sulfur and hydrated iron sulfate species in the solid residue – was subjected to solid-liquid separation. The copperrich solution was further clarified (filtered) and fed directly to electrowinning (DEW) for copper recovery. Lean electrolyte was recycled back to the pressure-leaching step as a source of acid. Additional cooling water for the pressure-leaching step was provided by recycling pressure- leach discharge solution. (It should be noted that discharge slurry can also be recycled to the pressure leaching feed or into one or more compartments – this may be desirable in certain applications.) A portion of the lean electrolyte was bled out of the DEW circuit and processed through SX and conventional EW for copper recovery. The MT-DEW-SX pilot plant flowsheet is shown in Fig. 2.

Three pilot-plant campaigns were conducted on the MTDEW-SX flowsheet, processing a total of 834 kg of concentrate containing 251 kg copper to produce 160 kg of copper cathode (the balance of the copper was held up in inventory). Two concentrates were used for the campaigns. To demonstrate the flowsheet and to demonstrate operability, the first campaign was conducted using Bagdad concentrate (of which there was essentially an unlimited supply available) and was run at a constant acid addition rate. Calculation of target acid additions took into account not only the free acid, but also the acid that was formed by hydrolysis of the iron in the recycled electrolyte. This was necessary because, as the acid concentration increases in the pressure leach discharge solution, so too does the iron concentration, which lowers the current efficiency in DEW and increases the iron concentration in the recycled solution. Thus, controlling the pressure-leach discharge acid concentration was critical to effective operation of the MT-DEW-S W process. The pilot plant demonstrated that the acid and iron concentrations could be controlled effectively. Once the ability to control the flowsheet had been demonstrated, a switch was made to use Cerro Verde primary sulfide concentrate generated by pilot plant flotation of a bulk ore sample. Superfine grinding. Concentrate feed for each campaign was subjected to superfine grinding in a Netzsch LME4 unit, which is a continuous, pressurized mill with a volume of 2.7 L and that is fitted with a horizontal shaft and stirring discs. The mill was operated at 1,750 rpm and filled with a charge of 3.4 kg of 1.2- by 2.4-mm silica "Colorado river" sand (Oglebay Norton Industrial Sands, Colorado). Grinding was conducted in batches using a slurry density of 50% solids. Generally, slurry was fed through the mill twice to achieve the desired particle size. The gross mill operating power was 3.5 kW, and the slurry temperature exiting the mill varied between 55[degrees] and 75[degrees]C. Freshly ground slurry was prepared for each campaign, and the slurry was stored under water in plastic 208-L (55-gal) drums to minimize oxidation prior to processing. Effective monitoring of the condition of the mill internals (i.e., wear parts) and regular maintenance throughout the campaigns were required to maintain the desired particle size.

A laser diffraction particle size analyzer, the Malvern Mastersizer (Malvern Instruments), was used to determine the size distribution following grinding. A significant amount of time and effort was devoted to understanding particle size distributions and the different methods of measurement.

Table 1 -Chemical analyses of finely ground concentrate used during MT-DEW-SX pilot plant runs.

Table 2-Mineralogical analysis of copper concentrate materials used for pilot plant work.

Operation of the Netzsch mill was complicated by a tendency for the mill to plug, causing numerous shut downs. Unplugging the mill typically required flushing it with relatively large quantities of water to free up the feed lines and mill internals. However, during the pilot plant operation, one run was made using Cerro Verde concentrate that had been passed through the Netzsch mill only once, i.e., a single pass, to investigate the effect of particle size. To conserve concentrate, the single-pass material remaining after this run was ground for a second time and was used as feed for the next run. However, calcium lignosulfonate (CLS) had already been added to the single-pass material. When regrinding in the mill, this CLS- dosed material operated much more smoothly, and operators reported that the slurry viscosity appeared to be much lower than the viscosity of material without CLS added, with lesser tendency to plug the mill. In addition, this batch yielded a finer product top size (P98) than the previous double-pass grind batches, indicating a beneficial effect of the reduced viscosity on grinding efficiency. Once this was discovered, the Cerro Verde concentrate was ground with CLS added to the feed for the remaining runs. It was expected that this practice would prove to be beneficial for a commercial application and will be investigated at Morenci.

Figure 2 – Hazen pilot plant MT-DEW-SX flowsheet.

Pressure leaching. For each campaign, ground slurry was fed continuously into a pressure leach vessel. The vessel was a horizontal, titanium cylinder divided into four compartments by three vertical dividers, with a total volume of approximately 30 L. Each compartment was equipped with a mechanical agitator fitted with two 97-mm-diameter, 45[degrees] pitched-blade impellers. For the majority of the runs, an agitator speed of 600 rpm was used, but this was increased to 800 rpm and then 950 rpm towards the end of the final run. This is discussed in more detail in the "Analysis of results" section below. The vessel was fed from a slurry feed tank using one of two variable-speed, 12-stage progressive cavity Moyno pumps (one standby). The CLS and/or other dispersant were added into the feed tank, as required, prior to feeding the slurry in the pressure leach vessel. The slurry feed entered the vessel below the slurry level in the first compartment and then overflowed to each successive compartment. High-pressure steam was available to be injected into the first compartment for heating during the initial startup of the system.

Material and energy balance models were prepared by Aker- Kvaemerforacommercial-scaleMT-DEW-SXflowsheet (for the Cerro Verde application) in advance of the pilot plant campaigns, and these were used as the basis for the three MTDEW-SX campaigns. These balances were used to calculate the amount of pressure-leach vessel discharge solution to be recycled back to the vessel as cooling solution, as well as the amount of lean electrolyte from DEW to be recycled. This was important because, in a commercial application, these streams would provide the entire acid addition requirement for the pressure- leaching reactions.

The feed slurry density was carefully controlled. To compensate for the pilot plant operation without a true flash letdown system, the slurry density was maintained at a solids concentration approximately 10% higher than that determined by the material balance for the commercial-scale application. This was done to compensate for the lack of evaporation in the pilot plant operation, to ensure that the copper sulfate concentration in solution would closely approximate the expected DEW concentration in a commercial operation.

The acid-to-concentrate ratio was used as a critical variable for controlling the pressure leach vessel. Because the slurry density used in the pilot plant was higher than would be used in a commercial-scale process, additional fresh acid was added to maintain the target acid-to-concentrate ratio. To achieve this, a small amount of concentrated sulfuric acid was injected into the vessel under pressure through a sidewall port in Compartment 1 using a variable-speed, air-actuated piston diaphragm pump.

The primary cooling solution (recycled DEW electrolyte) was injected into the discharge of the slurry feed pump using a variable- speed, electrically actuated piston diaphragm pump. Oxygen was injected into the slurry feed line immediately before entering the first compartment (i.e., a cooler location), to avoid scaling and reduce plugging in the oxygen tubing. Oxygen was also injected into Compartments 2 and 3. It should be noted that the calculated oxygen requirement for Compartment 4 was combined with the Compartment 3 supply to allow for effective metering of such small gas flows. Potential measurements in Compartment 4 indicated that the solution remained highly oxidizing, even though no oxygen was sparged directly into the compartment. Excess oxygen was discharged from the pressure leach vessel through a pressure control valve to a set of water-cooled condensers. Condensate from this stream was collected and measured for mass-balance calculations.

Cooling liquor for Compartments 2 and 3 comprised recycled pressure-leach discharge solution. This solution was injected through the same tubing used for oxygen sparging, which humidified the oxygen prior to entering the vessel. A variable-speed, air- actuated piston diaphragm pump was used for injecting cooling liquor.

Slurry was discharged from the vessel through a two-vessel pressure letdown system. A programmable logic controller (PLC)- based timer control system was used to control the pressure letdown equipment. This system allowed the pressureleach vessel exhaust gas to be continuously monitored while discharging slurry from the vessel. The off gas was cooled to remove moisture, and then the oxygen content of the gas was measured using an oxygen analyzer. Once the partially cooled slurry was discharged from the second letdown vessel, it was discharged into a catch basin and then pumped to a surge tank for weighing and sampling. Periodically, individual compartment samples were withdrawn from the vessel into water- cooled bombs through short lengths of tubing. The tubing was located in each compartment so that the samples were withdrawn from well- mixed zones to ensure that representative samples were obtained. These samples provided information about the kinetics of the leaching reactions.

Other peripheral equipment associated with the pressure-leach vessel operation was linked to a PLC computer to monitor and control their operations – for example, feed pump and vessel pressures; feed slurry and vessel compartment temperatures; and feed, cooling liquor and acid-addition flow rates.

Pressure-leaching product treatment and handling. Following pressure leaching, the objectives of the subsequent product treatment steps were as follows:

* to recover a low-iron, solids-free, rich electrolyte solution suitable to feed DEW;

* to recover sufficient solids-free solution at the required copper concentration to recycle back to the pressure leach vessel (Compartments 2 and 3) as cooling solution; and

* to wash the pressure leach discharge filter cake (solid residue) thoroughly to remove all of the soluble copper to accurately determine copper extraction.

The pressure-leaching product treatment steps were typically conducted on batches collected over 90-minute operating periods. The liquid fraction of the discharge slurry was supersaturated in copper sulfate at temperatures below 40[degrees]C. Because of this, equipment and procedures for product recovery were used to keep the solution temperature above 40[degrees]C, and preferably above 50[degrees]C, to prevent crystallization of copper sulfate. Also, these procedures were established to minimize contact between the residue and liquor to help minimize iron dissolution or redissolution. The pressure-leach discharge slurry was first flocculated with Hyperfloc NF 301 (a high molecular weight, nonionic flocculant manufactured by Hychem Inc.) After settling, a portion of the pressure-leach vessel discharge solution was decanted and used as rich electrolyte for DEW. The DEW feed solution was filtered through a 1-um cartridge and stored hot until it was fed into the DEW circuit. Additional decant solution was recycled back to the vessel as cooling liquor (into Compartments 2 and 3). The settled solids and remaining solution were remixed and transferred to a heated tabletop vacuum-filtration system, where the solids were separated from the solution. Because of the additional aging that occurred during the filtration operation, the filtrate in the vacuum receiver was higher in iron and was not used as feed to DEW. Rather, this filtrate was combined with the solution to be recycled back to the vessel as cooling liquor.

The filter cake was subjected to multiple displacement washes to completely remove the remaining soluble copper. The initial wash liquors, containing some copper, iron and acid, were used for cooling liquor makeup. The more dilute wash solution was collected for subsequent disposal. The washed filter-cake residue was sampled and analyzed for mass-balance calculations.

Occasionally, the pressure-leach discharge slurry would cool too much, and copper sulfate would crystallize out of solution and blind the filtration systems. A repulping procedure was developed to redissolve the copper sulfate crystals and remove the soluble copper from the filter cake.

Direct electrowinning. Copper cathode was produced from rich electrolyte in an 11.2-kg Cu/day pilot EW cell containing two cathodes of full commercial length and one third of the commercial width. The EW pilot plant configuration is shown in Fig. 3, and the cell configuration and design details are shown in Fig. 4.

Copper starter sheets, produced at the Bagdad operation (Arizona), were cut to the desired dimensions at Hazen Research. They were then washed with 100-g/L sulfuric acid to prepare the surface, water-rinsed and inserted into the cell. A proprietary flow- through, metal-mesh, anode design, developed by the Phelps Dodge Process Technology group (a subsidiary of Freeport-McMoRan Copper & Gold Inc.), was used in the cell. Anode insulators were mounted along the edges of each anode and were positioned in the upper third and lower quarter of both faces of the center anode and on one face of the end anodes. The insulators were fabricated from half-sphere polyethylene stock.

Electrical connections between the cell header bars and the power supply (rectifier) were secured with bolts. During the first two campaigns, the header bar connections were the only electrical connections attached to the electrodes. During the third campaign, additional jumper connections were made across the anodes to help ensure a balanced current distribution. These jumpers were bolted to the anodes at the opposite end of the anode bus bar.

The DEW circuit was equipped with a gravity-overflow commercial electrolyte system that included two tanks in series. Both tanks were insulated, equipped with baffles and a variable-speed mixer and had a working volume of approximately 15 L. The bulk of the flow to these tanks was supplied by a recycled lean electrolyte stream that was recovered from the lean electrolyte overflow surge tank of the EW cell. The lean electrolyte recycle was delivered via a variable- speed, magnetically coupled centrifugal pump through a line fitted with a rotameter and pinch valve, allowing the flow rate to be monitored and controlled.

Figure 3 – Pilot EW process flow diagram.

Rich electrolyte was stored in a 25-L-capacity tank fitted with an electric immersion heater. The rich electrolyte solution was metered into the first commercial electrolyte tank using a speed- controlled peristaltic pump. The rich electrolyte was fed through a 1-u.m cartridge filter to minimize entrained solids. A cathode- leveling (or smoothing) reagent (PD4201, manufactured by Chemstar) was added to the first commercial electrolyte tank at a dose of 334- g/t copper plated, while operating at a current density of 308 A/ m2. A mist-control reagent (FC-1100 Fluorad, manufactured by 3M) was added into the second commercial electrolyte tank at a dose of 0.084 L/t (10 gal/10^sup 6^ lb) copper plated. In the case of both of these reagent additions, the reagents were diluted with water to allow the reagent addition to be controlled at the desired level, and the diluted flow rates were monitored.

The temperature of the electrolyte was controlled via a solution heat exchanger and a 6-kW electrically powered immersion heater. This system compensated for radiant heat losses through the EW system and enabled the electrolyte to be reheated to a temperature that maintained the target EW cell temperature. The solution heat exchanger discharged into a commercial electrolyte distribution manifold. The transfer lines and distribution manifold were insulated to minimize heat loss. The commercial electrolyte flow was split into two streams, allowing the electrolyte to be added into the EW cell through two lances. Each lance line was equipped with a rotameter and a control valve so that the flow to each stream could be controlled and balanced. The discharge from each rotameter fed a 316 stainless-steel injection lance that entered just above the bottom of the cell. The lances were fabricated with small orifices (drilled or laser-slotted) with correspondingly small cross- sectional areas, pointing either directly up or at an angle into the cell. A variety of different orifice configurations were tried.

The lean electrolyte from the cell overflowed into a 15-L surge tank. A significant portion of this solution was recycled back to the commercial electrolyte circuit; the volume of lean electrolyte recycle was about ten times that of the incoming rich electrolyte. The balance of the lean electrolyte was either stored in a discharge tank or recirculated back to the pressure leach vessel as one of the constituents of Compartment 1 cooling liquor.

MT-DEW-SX pilot plant results

Pressure leaching. The first campaign ran for 116 hours using Bagdad concentrate as the feed material. The average size of the superfine ground product was 98% passing 21 ujti and 80% passing 5 urn. The pressure-leach vessel operated at 160[degrees]C with an average retention time of 87 minutes, at an oxygen over-pressure of 1,380 kPa (200 psi) and a total pressure of 1,920 kPa (278 psi). For the initial 40 hours of the run, the acid addition was 378 kg/t feed, but this was increased to a target of 450-kg/t feed for the remainder of the campaign. The initial copper extraction was 96.5%,but this increased to 97.5% after the increase in effective acid addition. As expected, iron increased from 1.6 g/L during the first 40 hours of the run to more than 3 g/L by the end of the campaign as a result of the increase in acid concentration in the pressure leach discharge. The copper, acid and iron concentrations in the pressure leach discharge solution during the final 36 hours of the run were 111, 25.3 and 3.11 g/L, respectively. The average conversion of sulfide sulfur to elemental sulfur was 67%.

The second campaign was split into a 104-hour run with Bagdad concentrate and a 55-hour run with Cerro Verde concentrate. The Bagdad results are summarized in Table 3. Bagdad concentrate yielded a slightly lower copper extraction (96.1 %) under conditions similar to those of the first campaign. However, the P9g had increased to 25 urn as a result of wear to the Netzsch mill. This was thought to be the major reason for the lower copper extraction. Copper, acid and iron concentrations for this portion of the run were 112, 19.8 and 1.97 g/L, respectively. Conversion of sulfide sulfur to elemental sulfur averaged 70%.

Figure 4 – Pilot EW cell details.

For the portion of the second campaign operated with Cerro Verde concentrate, a significantly lower copper extraction of 93.6% was obtained under similar conditions to those applied for the Bagdad material. However, due to the finer size of the as-received concentrate, only a single pass through the Netzsch mill was required to yield a product close to the desired particle size. The single pass generated a product with 98% passing about 22 [mu]m and 80% passing about 7.5 (im. Copper, acid and iron concentrations for this portion of the run were 107,22.6 and 2.89 g/L, respectively. Conversion of sulfide sulfur to elemental sulfur averaged 69%.

Following the second campaign, a series of batch tests were run on the Cerro Verde concentrate to investigate the cause of the lower copper extraction. A number of potential causes for the lower extraction were investigated. It was concluded that using an excessive amount of flotation reagents in the pilot plant treatment of the bulk ore sample from Cerro Verde had resulted in excessive residual reagent in the concentrate. It was thought that this could be causing chalcopyrite to float in the pressure leach slurry. A batch test was run using concentrate that had been washed with acetone to remove residual flotation reagents. This yielded a much lower residue copper concentration, corresponding to a significantly higher copper extraction (i.e., 96% to 97%), supporting this theory. This was further substantiated by the examination of the pressure leach vessel at the end of the second campaign. Inspection of Compartment 4 showed the walls of the vessel in the vapor space were coated with unreacted sulfide particles. This coating was not observed when the Bagdad concentrate was processed. Table 3 – Summary data (averages) of Campaign 1 and Campaign 2 runs using Bagdad concentrate.

Several options were considered to address this issue in the pilot plant. Washing the concentrate with acetone was deemed impractical and unacceptable. hot water washing was tested, but was found to be far less effective than acetone washing. Another option was to depress sulfide mineral flotation in the pressure leach vessel by adding a suitable reagent. Batch tests were conducted on Cerro Verde concentrate using various addition amounts of CLS and quebracho tannin. These showed that 10 kg/t CLS and 5 kg/t tannin reduced the residue copper concentration down to 0.6% to 0.7%, similar to the results on Bagdad material at a CLS dosage rate of 10 kg/t. Other methods considered included increasing the agitator speed to try to better submerge floating sulfide minerals. This was only partially successful. After completing all this work, it was determined that the use of both CLS and tannin provided the best route to complete pilot plant runs on the Cerro Verde concentrate.

Table 4 – Summary data for Campaign 5 and 6 runs with Cerro Verde concentrate.

For the third and final campaign, the Cerro Verde concentrate was subjected to a double pass through the Netzsch mill, resulting in product with 98% passing about 15 u,m and 80% passing about 5 urn. Under similar conditions to the first two campaigns,copperextractionsaveraged95.9%,almost reaching the levels achieved with Bagdad concentrate. Copper, acid and iron concentrations were 114,23.2 and 3.19 g/L, respectively, as shown in Table 4. Conversion of sulfide sulfur to elemental sulfur averaged 65%.

During portions of the third campaign, the oxygen overpressure was reduced from 1,380 to 690 kPa (200 to 130 psi). The results indicated a slight decrease in copper extraction (about 0.5%) from this change. There was no significant change in sulfur content in the residue, or in acid and iron concentrations in the filtrate solution.

The results indicated that copper extractions were still increasing at the end of the pressure-leaching period (see Fig. 5). Extending the leach time from approximately 104 minutes to 120 minutes was expected to yield between 0.5% and 1.0% additional copper extraction; however, no work was done in the pilot plant to confirm this.

One important difference between the original mediumtemperature (MT) process (Marsden et al., 2007) and the MT-DEW-SX flowsheet is that the sulfate concentration in the solution phase in the pressure leach vessel is much higher, due to recirculating pressure-leach discharge solution and lean electrolyte (from DEW) as cooling solution. Table 5 compares the results of the MT process pilot plant runs on Candelaria concentrate with the MT-DEW-SX pilotplantresults for Bagdad and Cerro Verde. Pilot plant experience indicated that Candelaria and Bagdad concentrate behave similarly at medium- temperature conditions. Both the amount of sulfide sulfur conversion to elemental sulfur and the amount of sulfate in the residue are similar for the MT and MT-DEW-SX processes, indicating that the increased sulfate concentration in pressure leaching did not significantly change the sulfide mineral oxidation reactions under these conditions.

Table 6 – Particle size distribution comparison for pressure- leach feed and discharge residue.

The pilot plant equipment did not allow for the direct measurement of oxygen consumption; however, based on the sulfur- species analysis in the feed/residue samples and given an understanding of the process chemistry, the theoretical oxygen consumption could be calculated. For the Bagdad material, an estimated 69.4% of the sulfide sulfur was converted to elemental sulfur and 28.6% to sulfate. The remaining 2.0% was unreacted sulfides. For Cerro Verde concentrate, these values were estimated at 64.4%, 33.6% and 2.0%, respectively. Based on the feed mineralogy, these values yielded very similar oxygen consumptions of approximately 360 kg 02/t concentrate for both materials.

Table 6 shows that the pressure-leaching residue was significantly coarser than the feed concentrate. This was not unexpected because about 30% of the residue was elemental sulfur that froze on discharge from the pressure leach vessel.

Direct electrowinning. For the three pilot plant campaigns, DEW was conducted using conditions similar to those applied in commercial copper electrowinning tank houses, i.e., approximately 33 to 38 g/L Cu, 133 to 140 g/L H2S04,2.4 to 3.6 g/L Fe, 50[degrees]C electrolyte temperature, 2.5 to 3.0 L/min/m2 cathode electrolyte specific flow rate and 290 to 310 A/m2 current density. Copper starter sheets were used as the cathodes; two cathodes and three anodes were used. The cathode area was 0.36 m wide by 0.99 m long, and the spacing between anode and cathode was 0.039 m. A cell voltage of 1.70 to 1.75 V was applied as a result of the special anode employed. The special anode utilized was proprietary Phelps Dodge technology, developed in parallel with concentrate leaching technology. The application of these anodes is not limited to use in DEW. The cell voltage is affected by the copper and acid concentrations in the electrolyte, as illustrated in Fig. 6.

Current efficiencies of 88% to 90% were typically achieved during the runs. While these current efficiency levels were deemed to be acceptable, several opportunities to improve the current efficiency above these levels were identified during the pilot campaigns, but these are not discussed further here. Total EW energy consumptions varied between 1.57 and 1.68 kWh/kg copper deposited. It should be noted that the equivalent energy consumption in an EW system where conventional lead (Pb-Sn-Ca alloy) anodes are used was measured at 1.79 kWh/kg copper in other pilot testing at Hazen Research. The concentration of iron in electrolyte had a direct effect on current efficiency (and energy consumption), as illustrated in Figs. 7 and 8. It can be seen from these figures that, for every 1 g/L increase in iron concentration, the current efficiency decreased by 1.99%, and the energy consumption increased by 0.083 kWh/kg copper. The optimal iron concentration in the electrolyte was a compromise between optimizing copper extraction during pressure leaching (which is favored by higher acid concentrations with correspondingly high iron concentration in the resulting solution) and maximizing current efficiency (which is favored by lower acid concentration). Economic evaluation of the pilot plant results indicated that the optimum iron concentration in the electrolyte would be at, or close to, 3 g/ L.

Figure 5 – Copper extraction as a function of retention time, from compartment samples.

Figure 6 – Calculated cell voltage as a function of acid concentration (assuming the same overpotentials at 300 and 308 A/ m^sup 2^; T = 50[degrees]C; specific flow rate = 3 L/min/m^sup 2^ cathode).

Figure 7 – Effect of iron concentration on cathode current efficiency. Range of operating conditions: lean electrolyte = 30 to 42 g/L Cu, 133 to 145 g/L H^sub 2^SO^sub 4^; T = 50[degrees]C; 250 to 308 A/m^sup 2^; 2 to 3 L/min/m^sup 2^ cathode; copper starter sheet cathode, proprietary PD anode.

Figure 8 – Effect of iron concentration on energy consumption. Range of operating conditions: lean electrolyte = 30 to 42 g/L Cu, 133 to 145 g/L H^sub 2^SO^sub 4^; T = 50[degrees]C; 250 to 308 A/ m^sup 2^; 2 to 3 L/min/m^sup 2^ cathode; copper starter sheet cathode, proprietary anode).

Table 7 – Calculated and actual electrolyte impurity analyses for the MT-DEW-SX Process using Bagdad concentrate.

Table 8 – Calculated and actual final electrolyte impurity analyses for the MT-DEW-SX process using Cerro Verde concentrate.

The bleed from EW was split, with a portion returning to the pressure leach vessel as cooling solution. In a commercial operation using the MT-DEW-SX process, the remaining portion of the bleed is mixed with stockpile/heap-leaching solution and is processed through conventional SX/EW. This is considered in more detail in Parts IH and IV of this series of papers.

The key issue for the DEW pilot plant operation was copper cathode quality, which could be affected by the presence of impurities in the electrolyte and, more importantly, by the ultimate concentrations of these impurities at steady state. Any impurities dissolved during pressure leaching that did not precipitate prior to solid-liquid separation reported to the rich electrolyte. Obviously, impurity concentrations also vary depending on the chemistry of the concentrate feed. As a result, impurities in the pressure-leach discharge solution, the composition of both rich and lean electrolyte, were tracked carefully throughout the MT-DEW-SX pilot plant campaigns.

Based on initial results and a material balance for the proposed MT-DEW-SX flowsheet, the expected steady-state electrolyte impurity levels for the more important elements that would affect the bulk solution properties (i.e., aluminum, magnesium, sodium and zinc) were estimated. To reach steadystate concentrations for these elements in the electrolyte as quickly as possible, the solution was spiked to these levels at the start of the pilot plant DEW run. The calculated and actual steady-state electrolyte impurity concentrates for the first and second campaigns using Bagdad material are summarized in Table 7.

A similar approach was taken for the Cerro Verde material (Table 8), but in this case the list of elements was expanded to include nickel, cobalt and cadmium, in addition to aluminum, magnesium, sodium and zinc. This step-wise approach to developing an electrolyte that was representative of what could be expected for a commercial-scale operation for Cerro Verde was considered to be effective. Table 9 – Impurity analyses during Campaign 5 using Bagdad concentrate.

Table 10 – Impurity analyses during Campaign 6 using Cerro Verde concentrate.

Periodically during the pilot plant runs, samples of concentrate, pressure leach discharge solid residue, pressure leach discharge solution, and electrolyte were taken and analyzed for a suite of impurities. The results for the Bagdad and Cerro Verde pilot plant runs are summarized in Tables 9 and 10, respectively. The change in concentration of every element could be tracked from one sampling period to the next. This change in solution concentration was then checked against changes in the concentration of that element in the cathode for a given campaign. Any element for which both the solution and cathode concentrations were increasing during a campaign run was considered to be a potential problem for cathode quality. By the end of the third campaign, which used Cerro Verde concentrate, the cathode produced by DEW was meeting LME Grade A specifications. However, two elements remained a concern. Firstly, the zinc concentration in solution was increasing, and it was not clear that the zinc concentration in cathode had flattened out. secondly, the silicon concentration was increasing in both the solution and the cathode. In the first case, zinc was not considered to be a major issue because the electrolyte bleed to SX would remove zinc from the electrolyte continuously and keep the concentrations within reasonable levels. In the second case, while there is no LME specification for silicon in cathode, there was a concern that silicon could affect cathode quality adversely at high levels and there was a possibility that colloidal and/or other forms of silica could adversely affect SX operation.

For the three campaigns, seven cathode harvests were completed: two from the first campaign and two from the second campaign using Bagdad concentrate and one from the second campaign and two from the third campaign using Cerro Verde concentrate. Each harvest consisted of two cathodes. As discussed above, this provided the opportunity to track cathode quality during the various campaigns and to correlate this with electrolyte chemistry. The harvested cathodes were washed with fresh water, air-dried and weighed. Samples were cut from each cathode using a Saber saw and punch, as shown in Fig. 9.

Figure 9 – Cathode sample locations.

The samples were submitted for chemical analysis and metallographic work. An example of the chemical analysis results for cathode produced in the third campaign using Cerro Verde concentrate is provided in Table 11.

Table 11 – Example of typical chemical analysis of pilot plant cathodes while treating Cerro Verde concentrate.

In general, the chemical results met LME Grade A; however, some outliers for lead and sulfur were observed. Silicon was found to be highly variable, both between cathodes and by location within a single cathode, with concentrations varying from <1 to 25 g/t. Silicon was thought to be present as fine, particulate inclusions of precipitated hydrated silica. Elevated silica in the cathode was thought to be caused by elevated concentrations of soluble silica coupled with higher cathode porosity. This was confirmed to some extent by the metallography work.

X-ray diffraction (XRD) analyses showed that the cathode crystal structure had a preferred (220) orientation on both the blank (i.e., against the starter sheet) and non-blank sides. By contrast, the starter sheet had a strong preferred (220) orientation on the non- blank side. It is interesting to note that a commercial cathode from the EW tank house at Bagdad, Arizona, showed random orientation on one side and the preferred orientation (220) on the other side, similar to the starter sheets, rather than the preferred (220) orientation on both sides that was observed for the MT-DEW-SX pilot cathodes.

All cathodes produced in the three pilot plant campaigns had some areas where nodules had grown, but were otherwise smooth. Also, the cathodes showed some roping along the edges, which was not unexpected because of the higher current density at the edges. A reflection of the anode insulators was also visible on the final deposit, as expected. In general, the nodules formed in similar locations on all cathodes and grew to be as large as 3 mm in diameter and 5 mm tall after 2 to 3 days of plating. Because of the extremely localized nature of the nodules and the ability to manage the degree of nodulation by controlling electrolyte flow and distribution in the cell, this was not considered to be a problem for a commercial operation.

Pressure-leaching residues. Pressure-leaching residues from the MT-DEW-SX pilot plant were essentially the same as residues from previous medium-temperature pressure leaching pilot runs (Marsden et al., 2007). The residue consisted predominantly of elemental sulfur and hematite, with minor unreacted chalcopyrite and traces of pyrite and covellite. The elemental sulfur varied in texture and size from dominantly spherical prills, ranging from 10 to 20 [mu]m for the fine prills to more irregularly shaped coalesced agglomerates, fragments and coarse prills from 15 to 200 [mu]m, and occasionally up to 500 [mu]m.

The spherical prills were typically rimmed and enclosed by relatively compact hematite, whereas the coalesced agglomerates generally lacked any hematite rimming. The hematite occurred as finely dispersed pigment; as compact, solid subspherical to spherical particles about 5 urn in size; as rims 2 to 5 um thick around the spherical sulfur prills; and as hollow spheres with cell walls 2 to 5 urn thick. Unreacted chalcopyrite occurred preferentially encapsulated in or attached to the coalesced sulfur agglomerates, the coarse prills and the fragments, which all lacked the hematite rimming. A subordinate to minor amount of chalcopyrite was evidently liberated. The encapsulated chalcopyrite tended to occur in the outer zones of the sulfur agglomerates or attached along the periphery of such agglomerates and was less commonly disseminated. Chalcopyrite particle sizes were typically 5 to 15 [mu]m, but occasionally 30 u.m and larger. The edges of the chalcopyrite particles appeared to be corroded. In all of the residues, the constituents described occurred as aggregations where all the phases were intimately intermixed, with the exception of the coarse sulfur prills and fragments, which tended to occur as isolated particles.

Summary and conclusions

Three pilot plant campaigns were used to further develop a medium- temperature pressure-leaching process that incorporated direct electrowinning (DEW) to recover the majority of the dissolved copper values. Chalcopyrite concentrates from the Bagdad, Arizona, operation were used for the first campaign and for a portion of the second campaign. Chalcopyrite-dominant concentrates produced by grinding and flotation of a bulk sample obtained from the primary sulfide zone at Cerro Verde, Peru, were used in the balance of the pilot work. The results indicated that a copper extraction of approximately 97.5% could be achieved from the Bagdad material.

Lower copper extractions were observed initially for the Cerro Verde material, but this was found to be due to a large excess of flotation reagents (frother and collector) that were present in the concentrate. After removing a significant portion of the excess reagents, recoveries increased to 96% to 97%. It is expected that extractions could be improved to the levels achieved with Bagdad material in a commercial operation. Further opportunities for increasing the copper extraction exist in the form of particle size and acid concentration optimization. The pressure-leaching step generated solution containing approximately 115 g/L Cu, 3 g/L Fe and 24 g/L sulfuric acid, which was suitable for direct feed to EW. The pressure-leaching step was acid-autogenous, with all of the acid required supplied in the form of recycled electrolyte and recycled PLS.

Continuous pilot testing of the DEW step in conjunction with pressure leaching indicated that LME Grade Acathode could be produced from both the Bagdad and Cerro Verde concentrates. Reasonable current efficiencies (88% to 90%) were achieved with iron concentrations of 2.5 to 3.6 g/L, resulting in overall energy consumption of about 1.62 kWh/kg copper.

The next step in the development of a commercial MT-DEW-SX process was a large-scale demonstration of the technology. This is considered in Part III of this series of papers.

Acknowledgments

The authors thank Freeport-McMoRan for permission to publish this paper. Many Freeport-McMoRan and heritage Phelps Dodge staff members contributed to the concentrate leaching developments described in this paper. In particular, the efforts of the late Bob Brewer and Jodi Robertson are recognized. We would also like to thank David Baughman, Christel Bemelmans, Dennis Gertenbach, K.C. Oberg and Roland Schmidt of Hazen Research (Golden, Colorado) and Phil Thompson of Dawson Metallurgical Laboratories (Salt Lake City, Utah) for their assistance with metallurgical and process development work.

Paper number MMP-07-028. Original manuscript submitted online July 2007 and accepted for publication August 2007. Discussion of this peer-reviewed and approved paper is invited and must be submitted to SME Publications Dept. prior to May 31, 2008. Copyright 2007, Society for Mining, Metallurgy and Exploration Inc.

References

Marsden, J.O., Brewer, R.E., and Hazen, N., 2003, "Copper concentrate leaching developments by Phelps Dodge Corporation," Hydrometallurgy 2003: Proceedings of the 5th International Symposium in Honor of Professor Ian Ritchie, Volume 2, Electrometallurgy and Environmental Hydrometallurgy, C.A. Young, A.M. Alfantazi, C.G. Anderson, D.B. Dreisinger, B. Harris and A. James, eds.. The Metallurgical Society of AIME, pp. 1429-1446.

Marsden, J.O., and Brewer, R.E., 2003, "Hydrometallurgical processing of copper concentrates by Phelps Dodge at the Bagdad Mine in Arizona," ALTA 2003 Copper-8 Technical Proceedings, ALTA Metallurgical Sen/ices, Castlemaine, VIC, Australia. Marsden, J.O., Wilmot, J.C., and Hazen, N., 2007, "Medium-temperature pressure leaching of copper concentrates – Part I: Chemistry and initial process development," Minerals & Metallurgical Processing, Vol. 24, No. 4, November, pp. 193-204, SME, Littleton, Colorado.

Marsden, J.O., and Wilmot, J.C., 2007, "Sulfate-based process flowsheet options for hydrometallurgical treatment of copper sulfide concentrates." Proceedings of Copper/Cobre 2007 Conference, Riveras P. et al., eds., CIM, Montreal, Quebec, Canada.

J.O. Marsden and J.C. Wilmot

Senior vice president, technology and product development, and manager, concentrate leaching, respectively,

Freeport-McMoRan Copper & Gold Inc., Phoenix, Arizona

N. Hazen

President, Hazen Research Inc., Golden, Colorado

Copyright Society for Mining, Metallurgy, and Exploration, Inc. Nov 2007

(c) 2007 Minerals & Metallurgical Processing. Provided by ProQuest Information and Learning. All rights Reserved.