See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/248933800

Review of agglomeration practice and fundamentals in heap leaching

Article  in  Mineral Processing and Extractive Metallurgy Review · July 2005

DOI: 10.1080/08827500590944009

CITATIONS READS
42 1,449

1 author:

Sylvie C. Bouffard
BHP
44 PUBLICATIONS   351 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Jansen Potash Project View project

Heap biooxidation View project

All content following this page was uploaded by Sylvie C. Bouffard on 28 July 2016.

The user has requested enhancement of the downloaded file.

 Review of Agglomeration Practice and Fundamentals
in Heap Leaching

Sylvie C. Bouffard, Ph.D.

Submitted for Review and Publication in the Journal

Minerals Processing and Extractive Metallurgy Review

December 1, 2004

Affiliation:
Research Metallurgist
Placer Dome Research Centre
323 Alexander St.
Vancouver, B.C., Canada
V6A 1C4
Tel: 604-682-1730 ext. 239
Fax: 604-682-1767
E-mail: [email protected]

1
1. ABSTRACT

This paper reviews agglomeration practices for precious metal and copper heap leaching. Both

industries prefer drum to conveyor agglomeration, particularly for clayey ore or ore having a high fines

content. Precious metal heap leaching operations opt for cement in a dosage from 2.5 to 10 kg cement/t

of ore (5-20 lb/ton) added to a cyanide solution. Copper ores are agglomerated with water and up to 40

kg sulfuric acid/t of ore (80 lb/ton) without binder. The agglomerate physical characteristics, with the

exception of their strength, can be measured precisely and automatically. The impact of agglomeration

on the in-situ physical characteristics of the heap, other than the observable ponding and slumping, is not

well understood. The most substantial benefits of agglomeration include up to 90% metal recovery from

poorly-permeable ores, shorter leach cycles, extra metal recovery from already-leached tailings, and

better environmental heap closure.

2. KEYWORDS

Agglomeration, gold, copper, heap leaching, practice, fundamentals

3. INTRODUCTION AND OUTLINE

The Institute for Briquetting & Agglomeration defines agglomeration as “ … the consolidation of

solid particles into larger shapes by means of agitation alone (i.e. without application of mechanical

pressure in molds, between rolls or through dies). The work of T.C. Scrutton in 1905 is one of the earliest

references to agglomeration. Scrutton agglomerated ore by rolling it down a chute inclined at 60° and

stacked the agglomerates in a vat leach (very large vessel filled with ore and completely saturated with

leaching solution). The initial work on iron ore pelletization began a few years later in 1911 and quickly

expanded to include other materials, including manganese, fluorspars, and phosphate, for better furnace

operation. Advances in agglomeration for heap leaching occurred twenty-five years later, in 1937

precisely, with Shepard’s work on the agglomeration of gold tailings with lime and calcium carbonate.

There are numerous publications, proceedings, and conferences about the matured practice,

fundamentals, and benefits of agglomeration in the ferrous mining industry. There are relatively few, very

recent references on agglomeration for heap leaching authored by a couple of pioneers such as Gene

McLelland and Paul Chamberlin. This review paper summarizes the fundamental and practical advances

2
of agglomeration for heap leaching. This paper discusses, in order (1) the preliminary ore characterization

that determines the agglomeration requirements, (2) the most common agglomeration equipment

employed, (3) the nature, dosage, and chemistry of the solution and binders added, (4) the methods to

evaluate the agglomerate quality, (5) the impact of stacking and irrigation methods, and (6) the benefits of

agglomeration. The review of more than 100 publications produced two compilation tables (Table 1 and

Table 2) about the past and current practice of agglomeration in copper and precious metal heap leaching

sectors. The two tables contain data about (1) the type and size of agglomeration equipment employed,

(2) the nature and amount of solution added, (3) the nature and dosage of binders added, (4) the curing,

stacking, and irrigation practices, and (5) the measured impact of agglomeration on copper, gold, and

silver recovery. These tables are referred to throughout this paper. The paper utilizes SI units, with

imperial (e.g. lb/ton for dosage) or American (e.g. GPM/ft2 for irrigation rate) units listed in brackets. In

this paper, the unit “kg/t of substance” means “kg of substance per tonne of ore”, unless otherwise

indicated.

The most significant advances in agglomeration for heap leaching occurred in the late 1970s to

early 1980s at the U.S. Bureau of Mines in Reno, Nevada (Potter, 1983). This organization had

previously developed in the 1960s cyanide heap leaching processes for recovering precious metals from

low-grade ores. Cyanide heap leaching is the preferred process to treat low-grade gold and silver ores

that could not be economically ground and cyanide-leached in stirred tanks. Heap leaching consists of

stacking run-of-mine or crushed ore to form a pile several feet high, then irrigating the crest (top) of the

heap with a solution that percolates, by gravity, through the pile to solubilize the metal value, and finally

separating the soluble metals from the pregnant solution (called pregnant because of its high metal value,

in contrast to the barren solution – low in metal value – applied on top of the heap) discharging at the

bottom (toe) of the pile. In 1981, there were approximately 78 active precious metal heap leach

operations worldwide, of which 34 were based in the United States (Kappes, 2002). In 1981, 22 of the 34

American precious metal heap leach operations were located in Nevada.

Some commercial heaps commissioned prior to the U.S. Bureau of Mines agglomeration studies

were suffering from low solution permeability, which resulted in lower gold and silver recovery and longer

3
leaching cycle. The wide range of particle sizes in the heap, the segregation of these sizes during

stacking and irrigation, and the ore mineralogy were the ultimate culprits of low permeabilities.

The U.S. Bureau of Mines determined the optimum practical requirements (moisture level,

binder type, binder dosage, curing time, agglomeration equipment, residence time) for agglomeration of

precious metal ores and tailings. Original methods (e.g. drainage rate from a flooded column), albeit

atypical of heap leaching environments, were designed to compare quantitatively various agglomeration

conditions. A decade later, in 1979, the work culminated in the first pilot-scale heap leach test on

crushed and agglomerated ore in Eastern Nevada. The first commercial agglomeration heap leach began

operation in 1980. According to Gomes (1983), there were 36 commercial operations in the Western

United States in 1983 that agglomerated ore.

The 2002 statistics compiled by Kappes revealed that 12 gold operations heap-leached run-of-

mine ore, 7 heap-leached run-of-mine and crushed ore, and 24 heap-leached crush ore. 32 operations

reported an average heap height of 8.9 m, an average irrigation time of 70 days during the leach, and an

average solution application of 1.3 t/t of ore. 33% of 24 operations that heap leached crushed ore only did

not agglomerate; 66% agglomerated crushed ore prior to stacking. The survey did not indicate the

proportion of run-of-mine and run-of-mine/crushed operations that agglomerated.

Advances in agglomeration in precious metal heap leaching paralleled those in copper heap

leaching. Chilean copper production boomed from 20,000 tpa in the 1980s to 1,000,000 tpa in the 1990s

(Scheffel, 2002). This phenomenal expansion of the heap leach, solvent extraction, and electrowinning

procress is attributed to the development of the “thin layer leaching” concept. Sociedad Minera Pudahuel

commercialized this concept for copper oxide ores in 1980, and then for copper sulfide ores as “bacterial

thin layer leaching” in 1985. The ore, crushed to less than 12-16 mm (½”), is agglomerated typically with

concentrated sulfuric acid, stacked in heaps 6-8 m (20-26 ft) tall, and cured for a couple of days prior to

irrigation with raffinate or intermediate pregnant leach solution. The majority of copper operations

agglomerate in one form or another by mixing the ore with concentrated sulfuric acid and water (Table 1).

Pelletization, granulation, and agglomeration are common terms describing a size enlargement

process. The term agglomeration is most often employed for ores containing both fine and coarse

particles whereby fine particles coat larger particles. Lipiec and Bautista (1998) define these

4
agglomerates as rim agglomerates. The Geobiotics agglomerates (Kohr, 1998), whereby a sulfide

concentrate is coated onto an inert, coarse substrate, also belong to this category. The agglomeration of

ore containing only fine particles of less than typically 74 µm (200#), such as tailings or iron ore, is

referred to as pelletization. As shown in Figure 1, the pellet structure is more homogeneous than the

agglomerate structure.

Figure 1 Simplistic representation of a pellet (left) and an agglomerate (right).

The size enlargement chapter in Perry’s Chemical Engineering Handbook offers a

comprehensive outlook on kinetics, equipment, and modeling of agglomeration processes. Certain topics

directly applicable to agglomeration processes for heap leaching were extracted and are summarized

below.

The binding process occurs in typically four consecutive or parallel stages:

• Wetting – coating of particles with a liquid film, of great importance for tailings because of the

comparable size between the water droplet and the solid material

• Growth – characterized by three phases: nucleation, followed by coalescence of nuclei to form an

agglomerate, and layering of fines onto the nuclei or agglomerate

• Consolidation – bed agitation intensity and compaction pressure applied to an agglomerate that

reduce its porosity

• Breakage – classified as shatter, fragmentation, wear, and abrasion of agglomerates

A population balance model can describe mathematically the rates of nucleation, layering,

coalescence, and breakage. Such equations consider the changes in the number of agglomerates and in

their mass. For instance, the layering process does not change the number of agglomerates, but

increases their mass. The coalescence process (two agglomerates forming one) has the opposite effect.

To this author’s knowledge, there were no available references on the modeling of crushed ore, except for

some glimpses into the distribution of particles making up agglomerates of various sizes (Bouffard, 2003).

There are four types of bonding mechanisms:

5
• Solid bridges – created by the crystallization of dissolved substances, the hardening of bonding

agents such as glue and resins, or the chemical reactions between wetting fluid, added binder, and

ore. Cement curing for precious metal ore agglomeration is an example of the latter type. Glue and

resins are expensive and may interfere with the heap leach chemistry.

• Mobile liquid binding – created by the surface tension and capillary water suction which exist in three

forms:

o Pendular state corresponding to discrete lens-shaped rings at the points of contact of particles

o Funicular state corresponding to a network of liquid interspersed with air and particles

o Capillary state corresponding to the complete saturation of pores between particles.

• Intermolecular and electrostatic forces – created by the short-range forces that attract very fine

particles (< 1 µm) under agitation

• Mechanical interlocking – created amongst fibrous particles under agitation or compression

Of the four types of bonding mechanisms, solid and liquid bridges are the most common in the

agglomeration of crushed ore and tailings. Solid bridges are likely to better survive wetting in the more

saturated heap than liquid binding forces.

6
 Table 1 Agglomeration practice in the copper heap leaching industry

Operation Agglomeration conditions Leaching

Sulfuric Water Final Stacking Copper
Producti Initial particle Agglomeration Curing Stacking Irrigation Leaching
Operation acid (W)/raffinate moisture height 2 recovery
on (tpd) size (mm) equipment (d) equipment (L/m /h) time (d) 1
(kg/t) (R) (kg/t) (%) (m) (%)
6.1
Mine-for-leach,
2 drums, 14 ft Mobile and (18”X18”
Morenci, Phelps
85,000 P80 12.5 wide, 40 ft 2.1-3.1 W/R 7 0 6.85 radial drip
Dodge (O’Brien et
long stacker emitter
al.,2003)
grid)
Mantoverde
(Trincado et al., 23,000 P100 15 drum 6.8 10-12
2003)
Nifty Copper
Operation Radial 6-10 (31”
68 P80 17 15 W 5-10 9
(Readett et al., stacker center)
2003)
Drum, 8.8 ft
(Holle, 1996) 8,400 P80 12 wide, 26 ft 6-10 7-80 W 7-8
long, 6°
Quebrada Blanca P100 9 2 drums, 9.7 ft 6-9 on/off
(Canello and
13,200 P80 6 wide, 29 ft 5-7 65 W 9-10 5 6-6.5 (0.4 m × > 300 d 80%
Schnell (1995);
long, 5° 0.8 m)
Schnell, 1997)
El Abra (Jenkins 3 drums, 14 ft
Mobile
and Canello, 8,250 P100 19 wide, 43 ft 16-20 80-100 W/R 2 6-10 15 45 85%
stacker
1997) long, 6°
Zaldívar, Placer
Mobile 8 (drip
Dome Inc. 50,000 P80 12 conveyor 12 W 8 15-30 9 300 80%
stacker emitter)
(Bouffard, 2004)
Doña Inés de
P100 9 28-35 80-100 W 1 5 10-15 60 68-85%
Collahuasi S.A.
Lince P100 9 35 50 W 4 24 30 80%
Cerro Colorado P100 12.5 7 60 W 2 6 12 210 80%
Lo Aguirre P93 6 15-30 60-70 W 1 3-4 6-12 180-360 63-68%
El Salvador,
owned by Codelco P97 12.5 11-20 55-60 W 1-2 5 10 240-360 75%
(Chile) (Pino,

7
 1995)
Ivan Mine P100 12.5 20-50 150 W 20-25 30 (oxide) 84% (oxide)
Sociedad Punta
del Cobre S.A. P100 12.5 25 65 W 1 20 30-120 80% (oxide)
Planta Biocobre
Radomiro Tomic,
P100 51 5-8 40 W 8 10 45 78% (oxide)
Codelco (Chile)
Empresa Mineral
Mantos Blancos P100 10 26 90 W 1 5 15 50 83% (oxide)
S.A. Manto Verde
Tesoro Project,
Anaconda Chile P100 25 20 100 W 1 3-5 15 30-60 76% (oxide)
S.A.
La Cascada,
owned by
P100 9.5 32 0W <1 1.9 7-8 75% (oxide)
Sociedad Minera
La Cascada Ltd.
Minera Disputada
P100 8 60 W 1 3 36 14 65% (oxide)
de Las Condes,
Compania Minera
P90 9.5 45 130 W 1 1.8 20 75% (oxide)
de Tocopilla
1
The designation (oxide) following the percent copper recovery indicates an operation processing copper oxide ores only.

Table 2 Agglomeration practice in the precious metal heap leaching industry

Operation Agglomeration equipment Agglomeration conditions Leaching

Moisture Gold
Drum (# of drums,
Maximum after recovery
Producti Reverse diameter in m, length Cement Other Curing Stacking Irrigation Leaching
Operation particle Stockpile Belt Solution agglome 2 (%,
on (tpd) belt in m, angle in °, speed (kg/t) (kg/t) (h) height (m) (L/m /h) time (d)
size (mm) ration unless
in rpm)
(%) noted)
Arizona 1 kg
gold heap 240 9.5 1,0.9,2.1,5,17 5 NaCN/t 10 48 2.1 81 5 90%
leaching solution
Colorado 5 kg
Fly ash
gold vat 1500 12.5 1,2.6,9.8,8,10 1.5 NaCN/t 13 8 vat 3 90%
– 3.5
leaching solution
90%
Arizona
Lime – (before
silver heap 180-200 12.5 × 3.5
Water 10-12 72 32 18.3 7
37% in
leaching
90 d)
N. Nevada 9-13 20-30 60%
gold heap
2500 15.3 × 3.5-5 Water
(zones
48-72 3.7 11.0
(before 50 gain

8
 leaching in the d)
heap
retain
30%)
E. Nevada
5,160- NaCN
gold heap 19 1,2.7,6.4,N/A,N/A 2-5 8 4.6 9.8 20-80 70-90%
10,800 solution
leaching
Central
Nevada
tailings Lime –
P65 200# 1,2.6,6.7,4,10.5 5 Water 12-14 4.9 7.3 24 76%
operation 25
(McClellan
d, 1986)
Southeast
ern
California
25 leach
silver 24,000 P90 200# 2,3.0,9.1,N/A,N/A 17.5 Water 12-15 4.0 24.4 73%
40 rinse
operation
(McClellan
d, 1986)
Alligator
Ridge
(DeMull
and 0.6 kg
Lime –
Womakc, 19 × 1.5-5
NaCN/t 5-11 3.7 14.6 30-40 70%
1984; solution
Strachan
and van
Zyl, 1987)
Cerro Rico
by
1 kg
Compania
400 P50 100# N/A,2.0,7.0,N/A,N/A 10 NaCN/t 15-17 6 17.8
Minera del
solution
Sur S.A.,
Bolivia
Anaconda’
s Darwin
Silver Mill
Lime - 5
(Milligan 3,960 P87 100# × 17.5 15.7 96 1.4 35 25-72%
and
Engelhardt
, 1983)
Kennecott
Barneys
Canyon
Mining Co
(LeHoux,
1997)
Haile Gold NaCN
1,600 49 N.A,1.8,7.9,7,9 4-7.5 12 48 4.3 30-45 90%
Mines Inc. solution

9
 (Phifer,
1988)
Masbate,
Philippines NaCN
1,200 12.5 N/A,2.7,8.8,N/A,N/A Lime – 9 9-12 24 3.4 17.1 20 58%
(Pizzaro et solution
al., 1986)
Houston
Internation
al Minerals 0.25 kg
Borealis 1,400 37 × 5 NaCN/t 4.6 14.6 82-84%
Project solution
(Outzen,
1983)
Gooseberr 85%
y Mine gold
P85 200# N.A.,3.0,9.8,12.5,N/A 7.5-11 17.6 9.1 7.3 60-70
(Butwell, 75%
1990) silver
Brewer
Nalco
Gold NaCN 90 leach
(Pautler et × 4 9760 –
solution
10-12 72 12.2
15 rinse
7% gain
0.12
al., 1990)
Cripple
Creek, CO NaCN
1,500 N/A,2.4,9.8,N/A,N/A 5 13 8 2.7 (vat)
(Gomes, solution
1983)
Tombston
e, AZ
(Gomes,
4,800 × 3.2 7 90% Ag
1983)
Maggie
Creek, NV Fly ash
(Gomes,
2,000 × 3.5
– 1.5
13
1983)

10
4. ORE CHARACTERIZATION

Several publications have stressed the essential need to characterize the physical, chemical,

and mineralogical properties of the ore to be heap leached. With regards to the physical characteristics,

the methods include: particle size distribution, silt vs. clay content, Atterberg limits of plasticity and solidity,

weathering and swelling characteristics, triaxial strength and internal friction angle, and permeability.

The particle size distribution (or ore gradation) obtained by wet screening constitutes the starting

point for further testing. Most consultants (Heinen et al., 1979; McClelland, 1986; Garcia and Jorgensen,

1997; Kinard and Schweizer, 1987) concur that the proportion of fines smaller than 50-75 µm (200-270#)

determines the need for agglomeration. Garcia and Jorgensen (1997) recommended agglomerating ores

containing more than 5% of –74 µm (200#) fines. McClelland (1986) and Garcia and Jorgensen (1997)

recommended agglomerating with binder if –74 µm (200#) fines account for more than 10-15%.

The Atterberg limits are classified as liquid and plastic limits. The Atterberg liquid limit

defines the moisture content at which the material changes from a liquid to a plastic state. The Atterberg

plastic limit defines the moisture content at which the material changes from a plastic to semisolid state. If

the liquid limit is greater than 20 and the plastic limit is greater than 10, the ore appears to contain high

clays. The liquid and plastic limit can be measured using the ASTM D4318 standard.

The triaxial strength and internal friction angle, measured using the ASTM D4767 standard,

helps to determine whether the heap will have sufficient static and seismic strength. A friction angle of 30-

37% indicates sufficient stability. Lower friction angles could be evidence of poor stability.

ASTM standards D4546 and D5890 measure the swelling index of soils and clay mineral

component of geosynthetic clay liners. A common procedure for measuring the swelling index consists of

placing 0.1 kg of material in a cylinder equipped with a porous plate and submerged in water. The height

difference after 24 h of soaking is a direct measure of the swelling.

The greater the permeability of a bed of rock or soil, the greater is the ability of the bed to carry

larger flows of solution without flooding and to drain rapidly. The solution application rates in heap

leaching are so small that the fluid can be assumed to obey Darcy’s law:

11
 k ' ρg
u = − K∇P = − ∇P Eq. 1
µ

where u: vector flow velocity (m/s)

P: hydrostatic head (m)

g: gravitational acceleration (9.81 m/s2)

K: saturated permeability of the bed (m/s)

k ’: intrinsic permeability (m2)

ρ: fluid density (kg/m3)

µ: liquid viscosity (kg/m/s)

There are numerous ASTM standards (D5084, D2434, D5856, D5093, and D6391) to measure

the permeability of a bed of rocks, i.e. when all pores in the bed are filled with solution. The constant and

falling head permeameters are popular methods.

On a scale from excellent to worse permeability, clean gravel has the largest permeability

ranging typically between 10 to 100 cm/s. Gravel drains very easily. Clean sand ranks second to gravel,

with a typical permeability of 1 cm/s. Very fine sand has a much lower permeability of 10-5 cm/s and

drains poorly. Organic and inorganic silts, mixture of sands, silts, and clays have a very low permeability

of 10-6 cm/s. Lastly, impervious soils made up of clays have the worst permeability ranging from 10-7 to

10-9 cm/s. These soils drain very slowly, if at all, due to the swelling of the clays.

This paragraph presents three examples of permeability measurements on unleached and

leached ore. The permeability of the unleached agglomerated high-clayey ore decreased rapidly from 0.2

cm/s to 10-4 cm/s with an applied normal stress of 25 psi (Garcia and Jorgensen, 1997). This load

corresponds to a heap height of 8 m (26 ft). With increasing loads of up to 200 psi, the permeability was

reduced further to only 10-7 cm/s. Uhrie et al. (2003) found that the permeability decreased with

increasing clay content. The permeability of an ore containing low clay remained above 10-1 cm/s under

loads corresponding to an equivalent heap height of 183 m (600 ft). In comparison to the guidelines

referred to in the previous paragraphs, the high-clayey ore tested had a very high permeability of 0.003

cm/s under load. Kinard and Schweizer (1987) measured the permeability of leached clayey

agglomerates, originally 0.3 to 1.0 cm in diameter, after cyanide heap leaching. The agglomerates were

12
recovered from the 9-m (29.5-ft) tall heap with a backhoe, after measuring the heap density at different

depths. The permeability ranged from 10-4 to 4×10-7 cm/s and was inversely proportional to the density.

The bulk density was surprisingly low, ranging from 1.19 to 1.43 t/m3 (74-89 lb/ft3), but not correlated with

heap depth.

A decision to agglomerate should be based on the proper evaluation of the physical, chemical,

and mineralogical characteristics of the ore. Agglomeration alone may suffice to increase solution

permeability. In other instances, agglomeration of ores of high fines content, high clay content, or brittle in

nature may still not provide adequate solution and air permeability. A viable process consists of screening

out (desliming) fines and heap leaching the remaining coarse particles. Trent Parker and Harmel Dawson

of Dawson Metallurgical Laboratories were awarded in November 1979 the first patent related to

screening (U.S. patent 4,173,519). It is worth noting, though, that the remaining coarse particles may still

decrepitate in contact with the leaching solution. This is more problematic for copper or zinc sulfide ores,

which contain a larger percentage of metal value, compared to gold ores. Phifer (1988) noted a significant

improvement in the gold recovery from 30% to 80% after screening out –74 µm (200#) fines, which

accounted for 30-40% of a run-of-mine ore. A 3-m (10-ft) tall heap stacked with non-screened run-of-mine

ore had previously slumped to 2.2 m (7 ft). Fines had migrated due to heavy rainfall. A hybrid tank/heap

flowsheet should increase the overall recovery because of the larger recovery achievable from fines

leached in tanks, and the larger recovery obtained from coarse particles in the heap. In addition, it is

possible to envision crushing the remaining coarse particles, without generating too many more fines, to

further increase the recovery. An alternative to screening consists of screening the ore into fine and

coarse fractions and regulating the rate of each fraction fed to the agglomerator.

5. EQUIPMENT

Estimates proposed by Rose et al. (1990) and Kappes (2002) indicate that agglomeration and

stacking account for 6 to 10% of the total capital costs and 10-21% of the total operating costs (Table 3).

The cost of cement alone is about $US 1.00/t of the $1.15/t, i.e. 90% of the agglomeration and stacking

operating costs.

13
Table 3 Influence of the scale of the heap leaching operation on the capital and operating costs of the

agglomeration/stacking unit relative to the total cost. All figures in US dollars.

Production 400 tpd 1000 tpd 3000 tpd 15,000 tpd 30,000 tpd
(Rose et (Rose et (Kappes, (Kappes, (Kappes,
al., 1990) al., 1990) 2002) 2002) 2002)
Agglomeration/stacking $460,000 N/A $1,000,000 $3,500,000 N/A
Capital

Total $4,600,000 N/A $14,000,000 $53,600,000 N/A

Percentage 10% N/A 7% 6.5% N/A
Agglomeration/stacking N/A $1.15/t $1.20/t $1.10/t $1.10/t
Operating

Total N/A $7.20/t $11.50/t $8.30/t $5.20/t

Percentage N/A 16% 10% 13% 21%

According to Kappes (2002) recent survey of heap leach design and practice in the precious

metal industry, of the 24 of the 43 responding mining companies that crushed ore before heap leaching, 8

did not agglomerate, 5, including Barney’s Canyon and La Quinua operation at Yanacocha, turned to belt

agglomeration, and 11 utilized rotating drums.

This section discusses belt conveyor and drum agglomeration - the two principal industrial-scale

agglomeration equipment employed for heap leaching. Some references to rotating disc agglomerator

and pug mills are included.

5.1 BELT AGGLOMERATION

Conveyor belts are well suited to agglomerate ore containing typically less than 15% of –104 µm

(150#) fines. Belt conveyors can be operated in three ways.

When all belt conveyors are inclined at about the same angle (about 15°) and are moving in the

same direction, agglomeration occurs when particles touch each other at the transfer point between belts

or when they bounce on the belt upon landing (Figure 2). The belt moves typically at a rate of 1.25-1.50

m/s (250-300 ft/min) (Chamberlin, 1986). Dispersion bars hanging at the discharge of a belt improves the

mixing of the ore in its free fall of 1.2 to 1.8 m (4-6 ft). The number of required transfer points increases

with increasing fine content. The number of transfer points depends also on the nature of the solution

14
addition. LeHoux (1997) found that the addition of a slurry rather than water required 3-4 times more drop

points.

The solution can be sprayed at the transfer points or along the belts. Too little solution results in

excessive dusting at the transfer points. Too large solution addition results in spillage at the transfer

points, damages to the belt, more frequent shutdown for cleanup, solution running down the belt, as well

as compaction, rather than rolling, of the agglomerates upon landing on the pile. Startup and shutdown

may also lead to solution running onto an empty belt and washing off the pile. Because some ores cannot

absorb the prescribed amount of moisture all at once at the transfer points, LeHoux (1997) recommends

staging the points of solution addition at the tail of the conveyor trains to avoid buildup of slimes

underneath the equipment and solution running. In his experience, adding a polymer or a filtration aid

reduced splashing and allowed for large solution additions.

Figure 2 Belt conveyor agglomeration. Reproduced from Chamberlin (1986).

When the ore falls from a low-angle conveyor moving relatively slowly onto a high-angle (35-55°)

conveyor moving rapidly in the opposite direction of the low-angle conveyor (Figure 3), agglomeration

occurs at the transfer point and primarily on the high-angle belt due to the opposite forces in action: the

forward momentum of the belt attempting to move agglomerates to the top, counterbalanced by the

gravity forcing the agglomerates to roll down to the bottom of the high-angle belt. Too high of a belt angle

would cause the agglomerates to slide down the belt rather than roll. McClelland et al. (1986) rated the

quality of agglomerates produced from tailings with a reverse belt as inferior to the quality obtained from a

drum agglomerator, even though both equipment had the same residence time of only 10-15 s.

15
Figure 3 Reverse, high-angle belt conveyor agglomeration. Reproduced from Chamberlin (1986).

When the ore falls from a low-angle belt onto a vibrating conveyor, particle agglomeration occurs

at the transfer point and when agglomerates slightly bounce on the vibrating conveyor and hit each other.

5.2 DRUM AGGLOMERATION

Drum agglomeration consists of injecting ore into a cylindrical, inclined drum that rotates to

impart rolling, cascading, and tumbling (Figure 4). Particles are set in motion by the balance between

gravitational and centrifugal forces. Commercial drums average 6.4 m (21 ft) in length and 2.1 (6.8 ft) in

diameter. The drum will rarely exceed 15 m (50 ft) in length. Another drum is installed in parallel to

handle larger throughput. The solution is pumped through nozzles or perforated pipes located

preferentially along the first 2/3 of the drum length. The drum may be rubber-lined to prevent corrosion

and equipped with loose chains or rubber strips to prevent ore sticking. Partially closing the front entrance

of the drum with berms prevents ore spillage. There were no available references on the number and

size of baffles.

Drum agglomeration is well suited for ores containing high clays or a large fine content.

Chamberlin (1986) prefers drum agglomerator to belt conveyor when a binder must be added. In rare

applications, a drum agglomerator may be equipped with screens on the discharge end to separate the

oversize from the undersize. Recycle ratios between 2:1 and 5:1 are common in iron-ore pelletization and

fertilizer granulation circuits. Recycling affects the moisture content, already susceptible to feed moisture

variations, and the agglomerate size distribution.

16
Figure 4 Drum agglomeration equipment. Reproduced from Chamberlin (1986).

There are three steps involved in the calculation of the drum throughput, starting with the

calculation of the drum critical and normal rotation speed, then the drum residence time, and lastly the

drum throughput.

The critical speed, i.e. the speed at which a single particle is held stationary in the drum due to

centripetal forces, is given by Eq. 2:

g sin θ 42.3
C= ≈ Eq. 2
2π 2 D D

where C: critical speed (rpm)

D: drum diameter (m)

g: gravitational acceleration (m/min2)

θ: set to 80 and 90° for drum agglomerator

The normal rotation speed, N, is a fraction, α, of the critical speed, C. The fraction, α, is equal to

30% to 50%. Slow rotation allows the agglomerate to roll rather than cascade.

An empirical equation (Eq. 3), derived from rotary dryer by the U.S. Bureau of Mines, calculates the

residence time as a function of the drum length, the drum diameter, the rotation speed, the incline of the

drum, and the angle of repose of agglomerates:

1.77 θ L
t= Eq. 3
SDN
where t: retention time (min)

θ: angle of repose of agglomerates (about 45°)

17
 L: drum length (m)

S: drum incline ranging from 1 to 12.5° (typically 5-7°) from the horizontal

D: drum diameter (m)

N: normal rotation speed (rpm)

According to Eq. 3, longer drum, drum of larger diameter, drum inclined at a lower angle, and slow

rotation increase the residence time. Assuming the material to move in a piston flow through the drum,

the throughput can be calculated using Eq. 4:

πD 2 L ρf
Q = 1440 Eq. 4
4 t
where f: solid volume holdup (m3 drum filled with solids/m3 total drum volume), typically

10-20%

ρ: density of the solid volume holdup (t/m3)

Q: agglomerate throughput (tpd)

Combining Eqs. 2 and 3 into Eq. 4 yields:

πD 2 L ρf πD 2 L ρf
Q = 1440 = 1440 = 4039 D 5 2 ρ f S α
4 ⎛ 1.77 45 L ⎞ 4 ⎛ ⎞
⎜ ⎟ ⎜ ⎟
⎜ SDN ⎟ ⎜ 1.77 45 L ⎟ Eq. 5
⎝ ⎠
⎜ ⎛ α ⋅ 42.3 ⎞ ⎟
⎜ SD⎜ ⎟⎟
⎜ D ⎠ ⎟⎠
⎝ ⎝

Assuming ρ = 1300 t/m3, f = 0.15, S = 5°, and α = 0.40, Eq. 5 simplifies to:

Q = 1571D 5 2 Eq. 6

Predictions from Eqs. 2, 3, and 6 were compared to laboratory and industrial data from the heap

leaching, fertilizer, and iron ore industries.

Figure 5 compares the predicted and actual normal rotation speeds of drum agglomerators. The

lower and upper curves correspond to 30% and 50%, respectively, of the critical speed, C, as calculated

from Eq. 2. There is very good agreement between the various sets of data and the predicted rotation

speed. Half of the data points lie on the invisible 40% α-line. This value seems reasonable for future

calculations.

18
 30% of critical speed
45
(Eq. 2)
40
50% of critical speed
Normal rotation speed, N (rpm)
35 (Eq. 2)

30
Iron ore (Capes, 1980)
25

20 Heap leaching data

(Table 2)
15

10 Pilot-scale drum
(McClelland et al.,
5 1986)
Fertilizer (Capes,
0
1980)
0 1 2 3 4
Drum diameter (m)

Figure 5 Comparison of the predicted and actual normal rotation speeds of laboratory and industrial

drum agglomerators. The lower and upper curves define the boundaries of the predicted

normal rotation speed, N.

Eq. 3 was used to compare the predicted and actual residence times of the drums listed in Tables

1 and 2. The drum diameter and length data was obtained directly from these tables. If unavailable, the

rotation speed was estimated from the reliable Eq. 2, assuming α equal to 40%. Eq. 3 assumed θ and S

equal to 45° and 5°, respectively. The predicted residence times ranged from 14 to 53 s. These agree

well with the actual residence times in Tables 1 and 2. The actual residence times may be too short for

proper agglomeration. Chamberlin (1986) suggested a residence time of at least 60 s for coarse ore and

240 s for fines. For better mixing and longer residence time, Ranchers Exploration and Development

Corporation equipped a drum with a dam located 1/3 of the way in.

Eq. 6 was used to predict the throughput of all agglomerators shown in Figure 5, and all

agglomerators in Tables 1 and 2 for which the diameter was known. Figure 6 shows a poor agreement

between the predictions of Eq. 6 and the actual throughputs. The discrepancies may be attributed to the

empiricism of Eq. 3 and the assumption of piston flow behaviour in the drum. If the flow behaviour

19
resembled more well-mixed conditions, the residence time would be shorter, and thus the agglomerate

throughput would be greater.

Attempts were made to derive a simple empirical equation that predicts the throughput as a

function of the drum diameter. Such an equation exists for disc agglomerators, as will be shown in Eq. 10.

The proposed equation takes the form:

Q = kD m Eq. 7

Parameters k and m were evaluated by minimizing the sum of the squared difference between the

predictions of Eq. 7 and the actual data shown in Figure 6. Parameters k and m were found to be 23.2

and 4.7, respectively, for values of drum diameter, D, ranging from 0 to 4.5 m, and for ore throughput, Q,

expressed in tpd. This equation provides reasonable estimates of the ore throughput for drums up to 2.5

m in diameter (Figure 6). The wide data scatter above 2.5 m in diameter was not well predicted by this

model.

80000
Fertilizer and iron ore
70000 Heap leaching

60000 Eq. 6: Q = 1571*D^2.5

Eq. 7: Q = 23.2*D^4.7
Ore throughput (tpd)

50000

40000

30000

20000

10000

0
0 1 2 3 4 5
Drum diameter (m)

Figure 6 Comparison of the predictions of Eqs. 6 and 7 and actual ore throughputs of drum

agglomerators. Variables Q and D have units of tpd and metre, respectively.

20
 Another empirical equation is proposed to predict the drum ore throughput as a function of the

drum diameter, length, and rotation speed:

Q = kD m Ln N p Eq. 8

To estimate the four parameters, only data sets with known diameter, length, and rotation speeds

were considered in the least square minimization calculation. There were 6 sets from the fertilizer

agglomeration tests, 3 sets from the iron ore agglomeration tests, and 4 sets from the heap leaching

agglomeration tests. The parameters k, m, n, p took values of 823, 0.34, 1.08, and –0.88, respectively for

D expressed in metre, L in metre, N in rpm, and Q in tpd. Inserting these values in Eq. 8 yielded a very

good fit of the 13 actual throughput values across a wide range of drum diameters (Figure 7). This

equation also predicted very well the throughput of 4 incomplete sets from heap leaching agglomeration

tests. However, the predictions were unreliable for throughputs above 5000 tpd.

45000

40000

35000

30000 Fertilizer and

Ore throughput (tpd)

iron ore
25000 Heap leaching
20000
Eq. 8 prediction
15000

10000

5000

0
0 1 2 3 4 5
Drum diameter (m)

Figure 7 Comparison of the prediction of Eq. 8 and actual ore throughputs of drum agglomerators.

Since the length of the drum is typically 3 times longer than the drum diameter (Figure 8), Eq. 8

can be modified to include the length to diameter ratio. The revised equation is:

D 0.33 L1.08 D 1.41

Q = 823 = 2695 Eq. 9
N 0.88 N 0.88

21
 14
y = 2.98x - 0.30
12
R2 = 0.74
10

Drum length (m)

8

0
0 1 2 3 4 5
Drum diameter (m)

Figure 8 Influence of the drum diameter on the drum length. Data obtained from Tables 1 and 2.

More data on the rotation speed and residence time should be acquired to improve the fit of Eq.

9 for throughputs above 5000 tpd.

5.3 DISC AGGLOMERATION

A disc agglomerator (Figure 9), also known as disc granulator or pan granulator, consists of a

rotating, tilted disc or pan with a rim. Disc agglomerators produce pellets of uniform size. There is little to

no solid recycling. Solids and solution are continuously added to the disc. A coating of the feed material

builds up on the disc and the thickness of this layer is controlled by scrapers or a plow, which may

oscillates mechanically. Feed rate variations affect the rolling action. Solution is applied by a series of

spray nozzles distributed across the face of the bed. A coarse spray onto fines favors the formation of

nuclei for growth. Solid feed and spray nozzle locations influence significantly the agglomeration

performance and pellet structure. Pellets collect in the eye of the pan, located above the bottom rim of

finer pellets.

22
Figure 9 Disc agglomeration equipment. Reproduced from Chamberlin (1986).

The normal rotation speed and the throughput can be calculated using similar equations as for

the drum agglomerator. The critical rotation speed, C, is also given by Eq. 2 where θ is replaced by a

value between 45° to 65°. To calculate the normal rotation speed, the fraction, α, takes a value between

50 to 75%. Since a disc can range in size from 0.3 m (1 ft) to 10 m (33 ft), the rotation speed, N, can vary

from when 41 to 6 rpm, respectively, when θ = 50° and α = 60%. The disc angle and the fraction of the

critical speed are the two parameters to maintain constant in scale-up. If the rotation speed is too low,

sliding occurs. If the rotation speed is too high, particles are thrown off the disc or openings develop in the

disc, allowing spray blowthrough and uneven buildup on the disc bottom.

The residence time varies typically from 60 to 120 s. It can be increased by lowering the solid

feed rate or the disc angle, or by increasing the rotation speed or the bed depth. The solid flow pattern

lies between well-mixed and piston behavior.

The thickness of the pan walls is typically 10-20% of the diameter. The throughput is

proportional to the square of the pan diameter:

Q = kD 2 Eq. 10

where k takes a value between 0.5 and 1.2, Q has units of tph, and D has units of m (Capes, 1980). Eq.

10 applies for throughputs ranging from 5 to 1700 tpd.

Disc agglomerators are used in the iron ore, agricultural, and chemical industries, but not for heap

leaching.

23
5.4 OTHER AGGLOMERATION EQUIPMENT

Ores containing little clays and fairly coarse can also be agglomerated by far simpler means,

such as dumping water and possibly a solution of low cyanide concentration into the haul trucks, or

spraying moisture (water or cyanide solution) on the sides of the heap throughout the construction. As the

ore is rolling down the slopes, fines will stick to coarse particles. This is referred to as stockpile

agglomeration (Figure 10). Even the dozer moving the ore up or down the slopes can help particles

agglomerate.

Tibbals (1987) was a fervent proponent of pug mills, which are devices equipped with a

horizontal trough in which slowly rotates a central shaft, to which is attached mixing blades, bars, rods, or

paddles. In his opinion, the power consumption of pug mills is less than the drum agglomerator, the

capital cost is 25-50% of the costs of the drum, and the mixing action is comparable to the drum. Pug

mills were used at Haywood-Santiago and Florida Canyon operations. Pug mills and disc agglomerators

are very rarely used, if at all, in heap leaching, as confirmed by Kappes recent survey (2002).

Figure 10 Stockpile agglomeration. Reproduced from Chamberlin (1986).

6. MOISTURE REQUIREMENTS

In the mining industry, agglomeration and pelletization utilize a liquid, such as a cyanide solution

in the gold industry or a sulfuric acid solution in the copper industry, to bind particles together by a liquid

film. Without the addition of a binder, the surface tension and van der Waals forces that hold the particles

24
together are weak. Particle adhesion created by surface tension may fail when agglomerates retain more

moisture under irrigation. In fact, over wet ore looses its ability to agglomerate. Curing is also non-

essential when agglomerating with solution alone, unless using sulfuric acid, thought to react with gangue

minerals (e.g kaolinite) to render them amorphous and to inhibit silica dissolution (Cruz et al., 1980, Farias

et al., 1995). In fact, agglomerates inoculated with a microbial inoculum should remain moist to avoid cell

dessication.

Agglomerating gold ores with a cyanide solution does not affect the chemistry of the

agglomeration, as opposed to the chemical bridges created by the hydration reaction between dicalcium

silicate, tricalcium silicate, and water when using cement. Chamberlin (1980) reported, however, that a

cyanide solution improved particle bonding compared to water alone, but did not have any effect on gold

extraction. The uniform presence of cyanide in the heap before irrigation may shorten the time required

for cyanide transport to the gold surfaces. Worstell (1987) believes this phenomenon to be the rate-

limiting step in cyanide heap leaching, amongst oxygen gas/liquid transfer, oxygen and cyanide

adsorption on the gold surface, electrochemical dissolution of gold, desorption of the gold-cyanide

complex, and gold transport to the bottom of the pile. The benefits of adding cyanide during

agglomeration would translate in an overall faster recovery. O’Brien (1982) demonstrated that the

agglomeration of two gold ores with Portland cement and a cyanide solution reduced the leach cycle by 4-

10 days compared to the Portland cement and water combination. The tests were performed in columns

only, not in a heap.

Evidence suggests that the use of strong cyanide solution in agglomeration increases the

cyanide consumption. The use of a strong cyanide solution can also be a safety hazard for toxic gas

emanation at too low solution pH’s. A dosage of 1 kg NaCN/t of ore (2 lb/ton) is fairly typical for

agglomeration. At least half of the former and present gold heap leach operations that agglomerate(d)

use(d) a cyanide solution rather than water alone (Table 2). There is no specific guideline for the

appropriate moisture addition. The following equation has been proposed to estimate the moisture

content of agglomerate:

25
 1
ϖ= for size > 30 µm
ρ Eq. 11
1 + 2.17 s
ρl

where ω is the moisture content in kg liquid/kg dry ore, ρs is the solid true density, and ρl is the density of

the liquid having similar properties as water (Perry’s Chemical Engineering Handbook). For a solid

density of 2.8 g/cm3 and a water density of 1.0 g/cm3, the moisture content is approximated to be 14 wt%.

This equation does not account for the initial moisture content of the ore, its mineralogy, its particle size

distribution, and the use of binders. All affect the agglomerate final moisture content. Tailings, which

contain a greater proportion of fines, and thus more surface area for wetting, require typically twice as

much moisture as crushed ore (15-30 wt% vs 5-15 wt%). If the material to be agglomerated were too wet,

it should be dried, as excessive moisture does not produce individual agglomerates, but rather clumps of

agglomerates that do not roll down the slopes of the heap. A starting material that is too wet also limits

the quantity of soluble reagents that can be mixed with the ore. To further support the negative impact on

the agglomerate quality of adding too much moisture, Zárate and Guzmán (1987) found that the quality of

cement/lime-based pellets was optimum at a moisture of 17 wt%, but declined at higher water dosages.

All three methods employed in their work to measure the agglomerate quality (dip, flooding, compaction)

recommended the same optimum moisture content. Heinen et al. (1979) observed the same

phenomenon with agglomerated tailings.

Copper heap leaching operations agglomerate with water and sulfuric acid. Figure 11 illustrates

the water and concentrated sulfuric additions for 14 copper heap leaching operations. On average, 15-25

kg sulfuric acid/t of ore is added to 60-100 kg water/t of ore (120-200 lb/ton). The final moisture content

ranges between 7.5 to 12.5 wt%, a typical moisture for agglomerated crushed ore. There appears to be a

proportional relationship between water and sulfuric acid. The high sulfuric acid and water additions on

the right-hand side of Figure 11 could be associated to finely crushed ore having a larger acid demand.

The low sulfuric acid and water addition on the left-hand side of Figure 11 could be related to the

coarseness of the particles, to the higher than usual initial moisture content of the feed material, possibly

due to wet screening, or to the low acid consumption by gangue minerals.

26
 Holle (1996) found that larger amounts of acid added during agglomeration increased the copper

recovery by 30%. Phelps Dodge’s Morenci Mine-for-Leach operation recovered also 15% more copper

with the addition of 5 kg/t of sulfuric acid (10 lb/ton) (previously 0 kg/t) to agglomeration. The increase in

copper recovery was proportional to the acid dosage (Uhrie et al., 2003).

160

140

120
Water addition (kg/t)

100

80

60

40

20

0
0 10 20 30 40 50
Sulfuric acid addition (kg/t)

Figure 11 Survey of the sulfuric acid and water additions of 14 copper heap leaching operations.

Each data point represents one (1) plant.

Sulfuric acid was also used to agglomerate radioactive uranium-bearing tailings from Ranchers

Exploration and Development Corporation in a 3.7 m (12 ft) wide by 10.1 m (33 ft) long drum rotating at 7-

16 rpm (Scheffel, 1982). 544,000 tonnes of agglomerates prepared with 45-68 kg/t (90-136 lb/ton) of

sulfuric acid were stacked. Some success had been obtained using Polyox WSR 301 and sulfuric acid,

but its addition was considered unnecessary.

Others have employed a hypochlorite solution to oxidize sulfidic refractory ore (Perez et al.,

1990; Ahmadiantehrani et al., 1991). Perez et al. (1990) patented the use of a sodium or calcium

hypochlorite solution in the amount of 2.5-22.5 kg/t (5-45 lb/ton) of chlorine and 2.5-10 kg/t (5-20 lb/ton) of

cement or gypsum to destroy, modify, or passivate sulfidic minerals. Agglomerates cured for 1 to 3 days

before rinsing with water and cyanide leaching. Previous tests by Ahmadiantehrani et al. (1991) had

consumed up to 80 kg/t (160 lb/ton) of hypochlorite (equivalent to 55 kg/t or 110 lb/ton of chlorine) due to

27
the oxidation of gangue minerals, but recovered 80% of the gold content. This consumption was reduced

to 9.5 kg/t (19 lb/ton) of chlorine by agglomerating with cement and by operating at low temperatures (3-

15°C). This did not affect the gold recovery.

The method of solution addition is not critical for crushed ore since there are several large

particles that can act as a nucleus for fine adhesion. Tibbals (1987) suggested that the amount of energy

input for agglomeration of crushed ore was more important than the method of solution addition. When

the ore particle size is smaller than the droplet size, as for tailings, these authors (Tibbals, 1987; Eisele

and Pool, 1987) agree that the moisture should be added as droplets rather than as an atomizing spray.

Fines adhere to the droplet to create a nucleus. Atomizing spray do not form nuclei; the wet fines do not

agglomerate.

7. BINDER REQUIREMENTS

The weak forces of adhesion between water and ore particles can be strengthened with binders.

A binder can be a liquid or solid that forms a bridge, film, or matrix, or that causes a chemical reaction.

The proportion of fines and clays in the ore determines the need and dosage of binder. For instance,

Chamberlin (1986) recommended adding cement for gold ores containing more than 10% of material

smaller than 75 µm (200#). Dry binders should be mixed with the ore preferably in the crushing circuit, at

the entrance of the drum agglomerator, or added on the first conveyor. Liquid and viscous binders should

be mixed with the agglomerating solution for better distribution.

The pH of the heap leach operation determines the nature of the binder selected. There are

numerous binders for alkaline pH’s, including lime, cement, silicate, pozzolan, and polymer. 14 of 19 past

and present cyanide heap leaching operations agglomerated with cement (Table 2). Very few copper

heap leaching operations add any binder to the sulfuric acid/water combination, possibly because of the

limited selection of acid-tolerant and microbial-resistant binders (polymers and gypsum). The next

sections describe the chemistry, dosage, and strength-producing mechanisms of several binders.

7.1 LIME, POZZOLAN, AND SILICATE

To minimize cyanide losses as HCN at pH’s less than 9, lime (Ca(OH)2) is added to gold and

silver ores in amounts from 1.5 to 25 kg/t (3-50 lb/ton) to provide alkalinity. Lime was found to be a less

28
effective binder than Portland cement. Lastra and Chase (1984) recommended lime agglomeration for

ores containing no clays, either on belts for coarse particles or on belts with vibrating chutes for particles

smaller than 12 mm (½”). Lime has long been known to strengthen clays through three mechanisms: (1)

formation of carbonates, (2) destructuring the clay minerals by raising the pH, and (3) by ion exchange, in

which natural monovalent cations adsorbed onto the clay minerals are exchanged by divalent cations.

The latter mechanism reduces the diffuse double layer thickness, thereby minimizing clay swelling.

Litz (1993) and an article in the November 1992 issue of the Engineering & Mining Journal (New

agglomeration aids help leaching, pp. 24-25) reported on the development of Leach-It, a modified lime

patented (U.S. Patent 5,116,417) by Chemstar Lime Co., Phoenix, AZ. This product crystallizes into

cementitious torbemorite and ettringite minerals, producing a dentritic structure.

Walker and Oliphant (1992) also developed and patented a mixture comprised of 10-80%

calcerous component, such as quick lime, 5-50% siliceous-calcerous component, such as fly ash, and 10-

80% sulfated component, such as gypsum. The mixture should be added in the amount of up to 2 wt% to

precious metal ores for cyanide heap leaching.

Pozzolan, such as cement kiln dust, fly ash, granulated blast furnace slag, and other metallic

slags, is a fine siliceous or siliceous and aluminous material that can be used as a binder. When mixed

with lime and water at ambient temperature, the glassy, fine particulates produce cementitious

compounds such as calcium silicate hydrate gel and calcium aluminosilicate. Pozzolan sets slowly in

comparison to cement.

Solid or liquid silicates are produced by the fusion of typically 1.5 to 3 parts sand with 1 part

sodium or potassium carbonate. They are employed as agglomeration binders, particularly if the

agglomerates are to be dried. The sand to carbonate ratio is determined by the binding mechanism, by

the set-up time, and by the material agglomerated. Starch, glycerin, molasses, dextrin, and lime may be

added as additives to increase the agglomerate strength.

Silicates produce different types of binding mechanisms, such as hydration, precipitation,

polymerization, and surface charge modification. Polymerization occurs rapidly when the pH of liquid

silicate drops below 10.7. Sodium silicates react almost instantly with multivalent metal cations to form

29
the corresponding insoluble metal silicates. The metal-robbing property makes silicates undesirable in

heap leaching where the dissolved metals must remain soluble.

A crushed ore containing uranium was agglomerated in a drum with 4.1-4.5 kg/t (8.2-9 lb/ton) of

sodium silicate, 11-12.6 kg/t (22-25 lb/ton) of concentrated sulfuric acid, and 66-96 kg/t (132-192 lb/ton) of

water to produce agglomerates 10-30 mm (3/8”-1”) in diameter (Videau and Roche, 1990). The 1,000-

tonne demonstration-scale heap containing no silicates ponded and channeled. The heap agglomerated

with silicates could be irrigated at 20 L/m2/h (0.008 GPM/ft2). Even though the heap containing silicates

eventually blew out in certain zones, this heap leached uniformly.

Lime, pozzolan or silicates are not as popular as Portland cement for precious metal heap

leaching.

7.2 CEMENT

Cement is the preferred binder for precious metal ore agglomeration. Cement is comprised of

50-70% tricalcium silicate (3CaO•SiO2), 15-30% dicalcium silicate (2CaO•SiO2), 5-10% tricalcium

aluminate (3CaO•Al2O3), 5-15% tetracalcium aluminoferrite (4CaO•Al2O3•Fe2O3), and various hydrated

forms of gypsum (CaO•SO3•H2O) (Kosmatka et al., 2002). There are five types of cement (10, 20, 30, 40,

and 50) under the Canadian Standards Association Standard A5 (CSA), and eight types of cement (I, IA,

II, IIA, III, IIIA, IV, V) under the ASTM C150 standard. CSA A5 types 10 to 50 are, respectively, essentially

the same as ASTM C150 cements Types I to V.

The two calcium silicates hydrate to form calcium hydroxide (CaO•H2O) and calcium silicate

hydrate (3CaO•2SiO2•8H2O). Tricalcium aluminate participates in three reactions. It reacts with gypsum

to produce ettringite (6CaO•Al2O3•3SO3•32H2O), with ettringite to produce calcium monosulphoaluminate

(4CaO•Al2O3•3SO3•12H2O), or with calcium hydroxide to produce tetracalcium aluminate hydrate

(4CaO•Al2O3•13H2O).

Tetracalcium aluminoferrite reacts with water and calcium hydroxide to produce calcium

aluminoferrite hydrate (6CaO•Al2O3•Fe2O3•12H2O). These reactions are summarized in Table 4. The

30
strength of hydrated cement is due primarily to calcium silicate hydrate. Tricalcium silicate and aluminate

hydrate and harden rapidly. The addition of silicates accelerates the set of cement.

Table 4 Cement compound hydration reactions

2 (3CaO•SiO2) + 11 H2O Æ 3CaO•2SiO2•8H2O + 3 (CaO.H2O)

2 (2CaO•SiO2) + 9 H2O Æ 3CaO•2SiO2•8H2O + 1 (CaO.H2O)

3CaO•Al2O3 + 3 (CaO•SO3•2H2O) + 26 H2O Æ 6CaO•Al2O3•3SO3•32H2O

2(3CaO•Al2O3) + 6CaO•Al2O3•3SO3•32H2O + 4 H2O Æ 3 (4CaO•Al2O3•3SO3•12H2O)

3CaO•Al2O3 + CaO.H2O + 12 H2O Æ 4CaO•Al2O3•13H2O

4CaO•Al2O3•Fe2O3 + 2 (CaO.H2O) + 10 H2O Æ 6CaO•Al2O3•Fe2O3•12H2O

Cement types I to V are not resistant to acids or highly corrosive substances. Calcium, sodium,

and magnesium sulfates may attack calcium aluminate hydrates and calcium hydroxide to form ettringite,

gypsum, and brucite (magnesium hydroxide). Cement with a low percentage of tricalcium aluminate is

more resistant to sulfate.

Cement accounts for 7-15% of a concrete mixture, while it typically represents less than 1% in

agglomeration of ore and tailings. Electron microscopy showed calcium to be uniformly distributed

throughout agglomerates of clayey particles (Heinen et al., 1979). The nature of the bridging holding the

clayey particles was unidentifiable.

Keuhey and Coughlin (1983) reported a rare instance of lower gold extraction from cement-

based agglomerates. 8% less gold was recovered after 32 days from a column containing agglomerates

prepared with 10 kg/t (20 lb/ton) of cement type II. Cement has been found to be most effective in high

siliceous ores (crushed rocks) and less effective in ores having high clay content.

Eisele and Pool (1987) recommend a dosage of 7.5 kg/t (15 lb/ton) of cement and 7.5 kg/t (15

lb/ton) of lime for tailings. They measured drainage rates in excess of 20,000 L/m2/h (compared to

31
typically 6-12 L/m2/h for heap leaching) from five columns loaded with agglomerates prepared with these

dosages of cement and lime. They also recommended this equal dosage of cement and lime for crushed

ores. Based on three independent tests to assess the agglomerate quality, Zárate and Guzmán (1987)

recommended instead twice as much of each binder, i.e. 8-12 kg/t (16-24 lb/ton) of cement and lime. On

the other hand, McClelland (1986) and Heinen et al. (1979) recommended 2.8-5.5 kg/t (5-11 lb/ton) of

cement, without or with lime at 3.9 kg/t (7.7 lb/ton).

Lime and cement dosages are reduced to 2.5 to 5 kg/t (5-10 lb/ton) of cement alone (only in few

instances with lime) for crushed ore in former and current industrial cyanide heap leaching operations

(Figure 12). Tailings require roughly twice as much cement as crushed ores, anywhere from 5 to 17 kg/t

(10-35 lb/ton) (Figure 12).

40
35
Cement dosage (lb/ton)

30

25
20
15
10

0
0.01 0.1 1 10 100
Average maximum particle size (mm)

Figure 12 Comparison of the cement requirements for the agglomeration of precious metal tailings

(left-hand side of graph) and crushed ore (right-hand side of graph).

7.3 GYPSUM

Sulfide heaps are irrigated with an acidic sulfate solution. Adding cement or lime to sulfide ores

results in the precipitation of gypsum and jarosite (Ca1/2Fe3(SO4)2(OH)6). Lastra and Chase (1984) saw

promise in gypsum and gypsum-derived binders. These precipitates could potentially passivate the

mineral surfaces and yield solution of too high pH’s.

32
7.4 POLYMERS

Nalco Chemical Company and Betz Dearborn Inc. have developed several polymers for ore

agglomeration under acidic or alkaline conditions. These polymers are either anionic, cationic, or both in

nature. Table 5 to Table 7, summarizing the chemicals patented by each company, show that

acrylamide is the basis for most polymers. The combination of anionic 70/30 mole percent of

polyacrylamide/poly(acrylamide/sodium acrylate) and cationic 90/10 mole percent of

poly(diallyldimethylammonium chloride) to poly(acrylamide/diallyldimethylammonium chloride) in a dosage

ranging from 0.1 to 2 kg/t (0.2-4 lb/ton) is a novel binder developed by Nalco Chemical Company.

Table 5 Cationic polymer binders.

Company Inventor(s) Patent Agglomerating agent Suited for

Block polymer containing a block of
Betz Dearborn Inc. Polizzotti et al. 5,668,219 Copper ores
an ammonium cation or acrylamide
Polymers of acrylamide and diallyl Copper or gold
Betz Dearborn Inc. Polizzotti et al. 5,512,636
dimethyl ammonium chloride ores

Table 6 Anionic polymer binders.

Company Inventor(s) Patent Agglomerating agent Suited for

5,077,021
5,077,022
Polymers of acrylamide and acrylic
Betz Dearborn Inc. Polizzotti et al. 5,112,582 Gold ores
acid
5,186,915
5,211,920
Nalco Chemical 4,898,611 Polymers of acrylamide and acrylic
Gross Gold ores
Company 5,100,631 acid
4,342,653
Nalco Chemical Polymers of acrylamide with
Gross et al. 4,786,318 Copper ores
Company carboxylate or sulfonate groups
4,875,935
Polymers of acrylamide with 2-
Allied Colloids Limited MacDonald 4,587,108 acrylamido-2-methyl propane Uranium ores
sulphonic acid

Table 7 Other polymer binders.

Company Inventor(s) Patent Agglomerating agent Suited for

Polypropylene glycol and alkylphenol
Betz Dearborn Inc. Cifuentes WO99/63123 ethoxylate in a paraffin oil solvent Gold ores
(known as PEG 400 MOT)
Nalco Chemical Cationic and anionic polymers added Copper and
Kerr WO99/20803
Company sequentially gold ores

33
 Nalco Chemical Company has also developed polymeric binders that would also not inhibit the

microorganisms present, and that would not be consumed by the microorganisms (Kerr, 1999). Of the

following binders (9704, CX-2131, CX-2134, CX-2185, CX-2194, 98DF108, and 97DF125) at doses

ranging from 0.2 to 0.8 kg/t (0.4-1.6 lb/ton), products 9704, 98DF108, and 97DF125 yielded the smallest

slump between 12-19% after 30 min of irrigation in columns loaded with a 27% clayey ore. The control

without binder slumped by 31%. The performance of these binders in the field was comparable to the

slump in columns. The columns did not slump further after leaching. Other indicators of the performance

of these binders included (1) the rate of ferrous oxidation as an indirect method of binder compatibility with

microbial species, (2) the turbidity of the column discharge solution, and (3) the solution flow rate after

compression (Kerr, 1999). Binder 97DF125 was superior based on these criteria.

Nalco binders were utilized by the Toquepala/Caujone Copper Mine in Chile and Asarco’s Ray

Mines Division in Arizona to belt-agglomerate copper ores. Lately, WMC Resources’ Nifty Copper

Operation has utilized two other products commercialized by Nalco in a demonstration-scale heap. A very

large column containing ore agglomerated with the same Nalco binders had achieved 90% copper

recovery in 55 days (Efthymiou et al., 1998). 30,000 tonnes of copper ore were agglomerated with 1 kg/t

(2 lb/ton) of Nalco’s Extract-Ore® 9560 and 5,000 tonnes of ore were agglomerated with 1 kg/t (2 lb/ton)

of a less expensive Nalco polymeric binder. Extract-Ore® is a medium molecular weight latex copolymer

of moderate anionic charge. The polymer consists of water-swollen micro gels, approximately 1 µm in

diameter. When diluted in water, the micro gels instantly solubilize, exposing the functional groups along

the polymeric backbone to the charged sites of the ore particles. 73.5% of the copper was recovered in

303 days. Although the large column performed better than the heap, the addition of Nalco’s Extract-

Ore® 9560 yielded a net improvement to Nifty’s heap operation. After 303 days, the heap contained

unleached and partially-leached ore.

Newmont Gold Mines in Carlin, NV and Philex Gold’s Sibutad Project in the Philippines also

employed Nalco binders and cement to agglomerate gold ores. Brewer Gold found that the column gold

recovery was superior with the use of 125 g/t (0.25 lb/ton) of Nalco 9760 and 2.5 kg/t (5 lb/ton) of cement

compared to an ore agglomerated with 7.5 kg/t (15 lb/ton) of cement (Pautler et al., 1990). A column

containing ore agglomerated with 125 g/t (0.25 lb/ton) of Nalco 9760 drained more slowly than a column

34
containing ore agglomerated with cement. In the field, the dual mixture of cement and Nalco 9760

performed better than cement alone. Two 10.7-m (35-ft) tall heaps were stacked side-by-side and

irrigated at 12.5 L/m2/h (0.005 GPM/ft2). One half was agglomerated with cement and Nalco 9760; the

other half was agglomerated with cement. Half of the solution applied on the heap without Nalco 9760

discharged from the heap with Nalco. This suggests severe channeling.

The Betz Dearborn Company developed binders HL 9120 and HL 9121 for agglomeration of

gold ores (Polizzotti et al., 1997, Polizzotti, 1993). Binder HL 9120 (US patents 5,077,021, 5,077,022,

5,186,915, 5,211,920) is an anionic 70/30 to 90/10 acrylamide/acrylic acid polymer of medium molecular

weight (1-10 million) added in a dose of 50 g/t (0.1 lb/ton) to 2.5 kg/t (5 lb/ton) of lime. Binder HL 9121

(US patent 5,472,675) is a cross-linked borated polyvinyl alcohol added in a dose of 50 g/t (0.1 lb/ton) to 3

kg/t (6 lb/ton) of lime and 3 kg/t (6 lb/ton) of cement. The influence of the binder on the gold recovery and

drainage rate was measured in columns. The binder did not affect the gold recovery. Binder HL 9121

improved the gold extraction by 5% under simulated heavy rainfall conditions (Polizzotti et al., 1997). The

column agglomerated with the binder drained 5 times faster than the column agglomerated with cement

(Polizzotti, 1993).

7.5 OTHERS BINDERS

The Research Centre for the Mining and Metallurgical Industry of Cuba has developed a

proprietary binder termed Additive 1 to agglomerate clayey copper ores (Serrano, 2003). Serrano (2003)

claims that Additive 1 is low cost, resistant to acid solutions, and forms porous pellets with good

mechanical resistance. The clayey ore, mixed with 25-50 kg/t (50-100 lb/ton) of Additive 1 and 20 kg/t (40

lb/ton) of sulfuric acid, was cured for 72 hours in columns. A column agglomerated with 25 kg/t (50 lb/ton)

of Additive 1 slumped by 34%. A column containing clayey ore mixed with coarse ore, and agglomerated

with 50 kg/t (100 lb/ton) of Additive 1, and 20 kg/t (40 lb/ton) of sulfuric acid slumped by 15%. However, a

column containing the same material without coarse ore slumped by 28%. Although the dosage of

Additive 1 is large and its benefits were undermined by the coarse ore addition, Additive 1 did not affect

the copper recovery or the acid consumption.

35
 Because of its sticky nature, molasses seems an appealing binder, acting as glue to

agglomerate particles. The forces of adhesion are relatively weak. What’s more, under irrigation, these

forces weaken due to the molasses dissolution. However, if it is combined with hydrated lime, an

exothermic reaction occurs between sucrose and lime that forms calcium sucrate, a rigid and stable

material that bonds firmly to particles. Because the reaction is fast, the reagents should be mixed

immediately before agglomeration. The molasses/lime combination has been used for many years for

briquetting, pelletization, and other applications involving coal fines, metal ores, fly ash, limestone, and

steel mill waste. There is no mention of its use in the mining industry.

The U.S. Bureau of Mines has also tested a high molecular weight polyethylene oxide (PEO, or

also known as Polvox). This chemical is an effective flocculant for colloidal silica and clays in basic

solution. Heinen et al. (1979) observed an improved drainage rate with the addition of 0.05 kg/t (0.1

lb/ton) of PEO mixed with lime.

8. CURING

Curing refers to hydration reactions between calcium silicates and water that form calcium

hydroxide and calcium silicate hydrate. The latter is by far the most important cementitious component in

concrete. The chemical composition of calcium silicate hydrate varies, but it contains typically 3 parts lime

(CaO) to 2 parts silicate. The calcium silicate hydrate forms dense bonds between particles.

Cement does not cure by drying. Heinen et al. (1979) observed that agglomerates cured and

dried broke down upon wetting. If the relative humidity of the mixture drops below 80%, cement stops

gaining strength. Hydration resumes after resaturation; the strength increases again. Approximately 40

kg (88 lb) of water per 100 kg (220 lb) of cement is necessary for curing. If an ore is agglomerated with 5

kg/t (10 lb/ton) of cement and has a moisture content of, say, 8 wt% before irrigation, there should be 4

times more water available than the required amount for proper curing.

The strength continues to increase provided that unhydrated cement is still present, that the

concrete temperature remains favorable, and that sufficient space is available for hydration products to

form. Although 28 days of curing is the standard in the concrete industry, 8 to 24 hours sufficed in

previous agglomeration studies of crushed ore (Chamberlin, 1986; McClelland, 1986; Eisele and Pool,

1987; Zárate and Guzmán, 1987). 72 hours was preferable for tailings (Eisele and Pool, 1987).

36
 Herkenhoff (1987) recommends letting cement-based agglomerates cure in a separate pile for

72 hours. Forming a separate pile does not seem necessary for two reasons. Firstly, if the agglomerates

can withstand being stacked in a separate pile and then moved to the heap, then they should as easily

withstand the impact of stacking the heap. Secondly, given that the installation of solution header lines

and emitters on top of the heap takes longer than 3 days, agglomerates should have more than enough

time to cure before the solution is applied.

Like the hydration of cement, there exist other binders (molasses and lime, glue, and resins) that

form solid bridges between ore particles. Other substances, like cyanide and ferrous ions present in the

agglomerates, may be oxidized before the start of irrigation. Reactions of the binder or of the

agglomeration solution may occur during the period between agglomeration and irrigation. Their reaction

products may affect the leach chemistry during irrigation.

9. AGGLOMERATE QUALITY

The agglomerate quality can be defined in terms of:

• the size distribution of the agglomerate, and particularly the uniformity of the agglomerate size

• the agglomerate moisture content immediately after agglomeration and during irrigation

• the agglomerate strength, and

• the agglomerate internal porosity

The iron ore, pharmaceutical, and fertilizer industries have developed quantitative and less

subjective techniques for measuring the agglomerate quality. The reader is referred to the review paper

of Pietsch (1985) for details on these methods. In this author’s opinion, the unavailability of quantitative

techniques lies in the inconsistencies of the feed material, in the specificities of the agglomeration

conditions dictated by the ore type, and in the general acceptance that the slow, difficultly-controllable,

and ever larger heap leaching processes overwrite the imperfections of the agglomerates produced.

Quantitative techniques employed in the heap leaching industry include the use of sieves to

measure the size distribution of moist and dry agglomerates. Nevertheless, very few records were found

in the literature about the size distribution of agglomerates.

Moist agglomerates can be screened onto a standard vibrating Gilson screen, which, if operated

for too long, can produce better agglomerates by the bouncing action imparted by the vibrations. Care

37
should be exercised not to blind the screen to achieve rapid separation. Snap freezing agglomerates with

liquid nitrogen prior to sizing may alleviate these inconvenients (Hall, 1986).

Although more tedious, dry agglomerates can be screened by gently manually rolling

agglomerates over screens. These techniques are very practical in the laboratory and in the field during

periodic sampling. For more frequent and on-line assessment of quality, some companies have

developed image analysis software that calculates the agglomerate size distribution from a reference

length and a digital picture taken over the conveyor belt carrying the agglomerates. Let us not forget that

what is an acceptable agglomerate size distribution for a given ore may not be for another. Nevertheless,

as a very rough guideline, Chamberlin (1986) suggested that no particle smaller than 104 µm (150#)

should remain unattached.

Lipiec and Bautista (1998) emphasized that agglomeration should eliminate free fines and

produce uniformly-sized agglomerates. A disc agglomerator is well suited to produce uniformly-sized

agglomerates from tailings. However, the heap leaching industry prefers drums to discs for agglomerating

crushed ore. Therefore, to produce more uniformly-sized agglomerates from crushed ore without the use

of a disc agglomerator, mining operations may choose one or more of the following options:

• Crush run-of-mine ore to eliminate boulders

• Screen ore to eliminate fines

• More transfer points with belt conveyors

• Longer drum operated at lower speed and set up at a lower angle

• Tighter control of the moisture content of the agglomerates

• Recycle the undersize agglomerates

• Add coarse particles to shift the apparent particle size distribution, as tested by Serrano (2003) on

clay-containing copper ores

The agglomerate moisture content can be determined by drying a representative sample.

Although very accurate, this method requires at least a few hours until dryness, a period of time. In a few

hours, the moisture content of the ore fed continuously to the agglomerator may have changed from the

sample collected. Chamberlin (1986) estimated the moisture content of the agglomerates as 1-3% less

than the moisture content of a dewatered filter cake. Both methods (drying and filter cake) lack the

38
immediateness sought after of a continuous operation. Observing that no free moisture glistens on the

surface of the agglomerate is also too subjective of a method.

One of the greatest contributions to the field of agglomeration since the early work of the U.S.

Bureau of Mines was the recent use of electrical conductivity at Phelps Dodge Cerro Verde operation

(Fernández, 2003). The electrical conductivity increases exponentially, with the largest signal detected

when a liquid film forms around the agglomerates. Agglomerates prepared with three types of ores (no

clay, medium clay, high clay content) all registered 150 mA when the proportion of agglomerates smaller

than 4.8 mm (4#) remained constant or was nil. Fernández (2003) related this conductivity reading to the

optimum moisture content of 4% (no clay), 6.5% (medium clay), and 10% (high clay).

In addition to measuring the conductivity, each type of agglomerates was prepared at different

moisture content and submitted to compaction tests. At the optimum moisture content, the slump was

10% for the no-clay ore and 25% for the high-clay ore. The extent of compaction for each ore type also

increased with increasing moisture content up until the optimum moisture, and then remained fairly

constant. Fernández (2003) did not describe the setup employed nor mentioned the initial height of the

bed prior to irrigation. This data could help explain why non-optimum agglomerates produced a more

permeable bed of lesser bulk density than optimum agglomerates. Expressing the slump as a percentage

can be misleading if the bed does not contain the same amount of material and if the initial heights are

different between tests.

Subjective tests for measuring the agglomerate strength include squeezing agglomerates into

someone’s hands and looking for clumping from good agglomerates. Others suggest that a clump of

good agglomerates should fall apart if poked. Others look for the rolling of agglomerates on the ground

after they have been thrown up into the air. Herkenhoff (1987) proposed to tumble dry agglomerates and

to measure the proportion of abraded fines. Others measure the height of the drop that leads to complete

agglomerate disintegration. The latter two techniques offer some quantitative basis for comparison, but

can be applied only to pellets that contain particles of uniform size. No technique obtains a direct

measure of the strength of agglomerates made up of particles of various sizes, such as rim agglomerates.

Other tests quantify the disintegration of agglomerates upon contact with water. For instance,

Chamberlin (1986) suggested that good agglomerates submerged in water should not disintegrate for

39
many hours. Milligan and Engelhardt (1983) measured the amount of fines produced when dipping pellets

in water ten times. The pellets must be previously cured for 6 hours at 90°C and cooled before dipping.

Such dip tests do not simulate the unsaturated conditions prevailing in a heap and apply primarily to

pellets because of their homogeneous structure. Dip tests are more qualitative for agglomerates.

Rather than dipping pellets, Chamberlin (1986) placed them in a burette and covered the top

with glass wool. He then applied water at increasing flow rates and measured the fine content in the

discharge solution. This technique does not measure the disintegration of the pellets throughout the

burette. The U.S. Bureau of Mines (Heinen et al., 1979; McClelland et al., 1985; Eisele and Pool, 1987)

rated the agglomerate strength using a more severe method that consisted of flooding a column of

agglomerates and measuring the rate of drainage. A column containing good agglomerates drained

rapidly.

Additional quantitative parameters can be obtained about the agglomerate quality by loading and

irrigating a column. For instance, one can measure the pore space in a column before and after irrigation.

The pore space, εb, is related to the bulk density of the column, ρb, and of the agglomerates, ρa, through

the following relationship:

ρb
εb = 1− Eq. 12
ρa

where εb : pore space between agglomerates (does not include pores within the agglomerates)

ρb : bulk density of a bed (mass of dry ore agglomerated/total volume of the bed)

ρa : bulk density of an agglomerate (mass of dry ore agglomerated/volume of the agglomerates)

To overcome the challenge of measuring the bulk density of an agglomerate, the pore space in

the above equation can be approximated as:

ρb
εb ≈ 1 − Eq. 13
ρs

where ρs is the ore true density (mass of dry ore/volume of dry ore).

This equation combines the pore space in and between agglomerates in a single variable. It

becomes straightforward to calculate the pore space by measuring the mass of dry ore loaded in a

40
column, the height of the bed after slumping, the ore true density, and the volume of water necessary to

flood the column.

Scaling down this concept at the agglomerate scale, one defines the agglomerate pore space,

εa, by the following relationship:

ρs
εa = 1− Eq. 14
ρa

where εa is the pore space in the agglomerates, and ρa is the bulk density of an agglomerate (mass of dry

ore agglomerated/volume of the agglomerate).

Videau and Roche (1990) found that a column containing large and wet agglomerates and

another column containing smaller and drier agglomerates slumped by different extent, but ultimately had

the same final bulk density. The slump is thus a poor indicator of the agglomerate quality and does not in

fact inform about the final pore space in the heap. An even better parameter than the pore space value is

the pore space value diminished by the proportion of pores filled with solution. The air-filled pore space is

critical for mutiphase reactions. The corrected pore space still does not define the size and inter-

connectivity of pores, two important aspects for heap aeration.

10. STACKING AND IRRIGATION

The stacking equipment, the heap height, and the irrigation equipment are as important as the

agglomeration itself to maintain adequate permeability.

With regards to the stacking equipment, 15 operations stack with conveyors and 9 stack with

trucks (Kappes, 2002). 100% of precious metal heap leaching operations that agglomerated in drums

stacked with conveyors (Kappes, 2002). Likely 4 of 5 operations that agglomerated onto belt conveyors

also stacked with conveyors. The fifth operation that belt-agglomerated, and likely 100% of operations

that did not agglomerate, stacked with haul trucks.

Truck stacking is suitable for run-of-mine ores. Trucks compress the surface of the newly-

stacked lift if stacked from the top down, or the bottom lift if stacked from the bottom up. At the Alligator

Ridge Mine, haul truck traffic was ultimately responsible for the low permeability of an agglomerated lift.

This resulted in lateral flow of solution that broke out through the slopes of the heap and around the

access ramps (DeMull and Womakc, 1984). Limiting the traffic to a central road did not improve the

41
permeability. The ground pressure and vibration from the dozer was believed to cause as severe

compaction as haul trucks. The leach performance improved significantly when the haul trucks dumped

ore at the toe of the new lift, leaving it to the dozer to push up against the slopes.

There are two types of conveyor stacking systems: 1) mobile conveyor unit (or grasshopper)

combined with radial stacker, and 2) spreader conveyor employed primarily for dynamic pads of constant

width and height. Because spreader conveyors travel across the entire width of the pile, there is less

segregation across the length of the pile. Radial stackers tend to create discontinuities at the intersection

between ridges and fingers (Scheffel, 2002).

According to published data on 17 former and current copper heap leaching operations, the

maximum and average heap height reach 10 and 5.5 m (32.5 and 17.9 ft), respectively (Figure 13). Static

heap designs, such as practiced as Monto Verde and Escondida, stack lifts of 5-10 m (16.5-33 ft), but

install liners and drainage pipes between lifts (Scheffel, 2002).

According to a 1987 survey of North American precious metal heap leaching operations, 22% of

operations stacked lifts of 0.9 to 1.8 m (3-6 ft) tall, and 30% from 2.1 to 3.7 m (7-12 ft) (Worstell, 1987).

These statistics agree well with those gathered by this author from publications dating back to the 1980s

and presented in Figure 13. Kappes’ more recent survey (2002) revealed that the average heap height of

32 gold heap leaching operations has more than doubled to 8.9 m (29 ft). In 2002, the highest heap had

10 lifts that rose to 120 m (394 ft) (Kappes, 2002).

100%

90%
Proportion of heaps of given height (%)

80%

70%

60% >7m
5-7m
50%
3- 5 m
40% <3m
30%

20%

10%

0%
Cu Au

42
Figure 13 Survey of the heap height of 17 copper heap leaching operations and 11 gold heap

leaching operations (Former and current operations included).

Figure 14 shows the proportional relationship between the heap height and its bulk density, with

the exception of one data point. The taller the heap, the greater is the bulk density. The probability of

retaining more solution and of reducing the air-filled pore space increases with increasing bulk densities.

1.9

1.8
Bulk density (t/m3)

1.7

1.6

1.5

1.4

1.3

1.2
0 2 4 6 8 10
Stack height (m)

Figure 14 Influence of the heap height on the heap bulk density. Data obtained from Miller (2003).

The mode of solution application and the irrigation rate may also affect the permeability of the

heap. There are three common methods of irrigation:

• drip emitters manufactured originally for agricultural purposes

• wobbler sprinkler, such as Senninger™ wobblers

• reciprocating sprinkers, often referred to as Rainbird™ sprinklers

Drip emitters consist of perforated plastic tubes or soft pipes laid parallel to each other on top of

the heap. Some tubes and pipes consist only of holes evenly spaced along the length. Others contain a

small labyrinth inside the tube or pipe that evens out the pressure, thus producing more uniform flow over

long distances. Drops emitted from these systems have the gentlest and the most local impact on the

heap surface. In addition, drip emitters reduce evaporation, especially if buried. At Phelps Dodge’s

Morenci Mine-for-Leach heap irrigated with drip emitters, the evaporation was measured to be 12%

43
(O’Brien et al., 2003). The main disadvantage of drip emitters is hole plugging, thus requiring the use of

antiscalants and in-line filters.

Wobblers are vibrating and rotating devices mounted on the header line producing multiple,

small jets of coarse water droplets across a radius of 3 m (10 ft). Droplets hit the entire heap surface,

producing more damages to the agglomerates than drip emitters. The better surface coverage with

wobblers than drip emitters yields a greater extraction of the ore at the top of the heap.

Rainbird™ sprinklers emit a single water stream 5 to 8 m (16-26 ft) long that rotates 360°. They

are ideal for irrigating slopes, but have high evaporative losses. Rainbird™ sprinklers are not suitable for

agglomerated heaps. The strong jet disintegrates agglomerates, unless the agglomerated heap surface is

covered by coarse particles to dampen the impact, as practiced by the Cerro Rico operation of Compania

Mineral del Sur S.A. Bolivia.

In the late 1980s, Rainbird™ sprinklers were the most popular irrigation method of precious

metal heap leaching operations (41%) (Worstell, 1987). 9% used Bagdad wigglers, 9% used Senninger™

wobblers, 14% drilled holes in the header lines, and 5% actually built walls on top of the heap for the

solution to pond on the surface. Fifteen years later, the popularity of drip emitters has grown markedly at

the expense of Rainbird™ sprinklers. According to Kappes 2002 survey of 37 operations reporting, 35%

used drip emitters, 14% – all based in a tropical climate with heavy rainfall – used only wobblers, and 51%

used both drip emitters and wobblers. No operation reporting used Rainbird™ sprinklers. One third of

operations using drip emitters buried the tube.

With regards to the irrigation rate, precious metal heap operations that crush ore irrigate, on

average, at larger flows than run-of-mine heap operations (11 vs 8.3 L/m2/h) (0.0049 vs 0.0037 GPM/ft2)

(Kappes, 2002). Table 8 compares the variability of the irrigation rates amongst run-of-mine and crushed

ore heap leaching operations.

44
Table 8 Comparison of irrigation rates of precious metal heap leaching operations of run-of-mine and

crushed ore (Kappes, 2002).

Irrigation rate Run-of-mine ore Crushed ore

Number of operations reported 17 19
2 2
< 8 L/m /h (< 0.0036 GPM/ft ) N/A 21%
2 2
8-10 L/m /h (0.0036-0.0045 GPM/ft ) 88% 63%
2 2
> 10 L/m /h (> 0.0045 GPM/ft ) 12% 16%

11. BENEFITS OF AGGLOMERATION

The benefits of agglomeration can be classified into three categories: heap physical structure,

leach chemistry, and environmental impact.

Agglomeration improves the heap physical structure by minimizing or avoiding ponding, slope

failure, or solution channeling. Channeling occurs in zones containing coarse particles. Because

agglomeration reduces the spread of the material size distribution, it minimizes segregation whereby

coarse particles roll down to the toe of the heap, leaving smaller particles at the top. Segregation is

nonetheless ideal for upward gas flow, but unsuitable for parallel downward solution flow due to increased

chances of ponding. The following example is an exception to this acknowledged benefit of

agglomeration. Despite the fact that a very clayey ore (60%) containing kaolinite and montmorillonite had

produced uniformly-sized agglomerates of 0.3 to 1 cm in diameter using 5 kg/t (10 lb/ton) of lime and 4

kg/t (8 lb/ton) of cement, trenches digged out in a 9 m (30 ft) tall heap showed coarse, well-graded, and

fine gradations after leaching (Kinard and Schweizer, 1987). The segregation was attributed to the radial

stacker. Agglomeration does not prevent the expansion of swelling clays upon contact with water, but will

avoid the formation of zones impermeable to flow by distributing the clay particles more evenly into the

heap.

Agglomeration is also thought to lessen fines migration – a phenomenon apparently observed by

many but poorly quantified. In this author’s opinion, low solution rates applied to cyanide and copper

sulfide heaps carry enough momentum to transport fines at the most 0.3 m (1 ft) below the surface, but

not to the bottom of the heap. The gradation observed in heaps that some have attributed to fine

45
migration could simply be the result of segregation caused by changes in the ore properties and the

method of stacking.

The cross-section of a leached heap at the Alligator Ridge mine confirmed that surface fines

don’t migrate too far, only 8-10 cm (3-4”) below the surface (Strachan and van Zyl, 1987). The gradation

was uniform everywhere else (Strachan and van Zyl, 1987). Rainfall and sprinklers are usually to blame

for damages (agglomerate disintegration, ore decrepitation) caused to the heap surface by the impact of

water droplets. The impact of a stalled emitter on the heap surface should not be ignored. A stalled

wobbler still sprays solution, while a sprinkler emits a single stream that hits in a single spot.

Agglomeration helps create a more porous heap with better air and solution distribution. Two

indicators of greater porosity are a lower bulk density and a lesser extent of slumping. Miller (2003)

showed that the bulk density of a heap containing non-agglomerated clayey ore increased from 1.15-1.30

t/m3 (71-81 lb/ft3) at the surface to 2.0-2.1 t/m3 (125-130 lb/ft3) 4 m (13 ft) below. This was equivalent to a

porosity of 50-55% at the surface to less than 30% 1 m (3.3 ft) below, approaching only 15-20% 6 m (20

ft) below the surface. At Candelaria, the bulk density decreased from 1.59 to 1.49 t/m3 (99 to 93 lb/ft3)

after agglomerating an ore containing as little as 2.5% of –147 µm (100#) fines (Chamberlin, 1980). The

bulk density of a heap agglomerated with both 125 g/t (0.25 lb/ton) of Nalco 9760 and 2.5 kg/t (5 lb/ton) of

cement was 3% lower than a heap agglomerated with cement alone (Pautler et al., 1990). There are no

general guidelines for the optimum heap bulk density after agglomeration. The bulk density can be as low

as 0.88 t/m3 (55 lb/ft3) at the Gooseberry Mine (Butwell, 1990) and as high as 1.67 t/m3 (104 lb/ft3) at the

Masbate operation (Pizzaro et al., 1986). Agglomeration minimizes but not does eliminate slumping. At

Little Bald Mountain, cement agglomeration reduced slumping from 24 to 8% (Tibbals, 1987).

Agglomerated copper sulfide heaps 4 to 8 m (13-26 ft) tall still slumped rapidly by about 12% (James and

Lancaster, 1998).

If neither agglomeration or desliming improve the heap permeability, ripping the heap surface or

remining the entire heap (ore turnover by backhoe) are methods commonly employed for underperforming

heaps. Blasting the leach pile has been suggested.

Remining has been pioneered by Girilambone Copper Company and is performed in several

Chilean operations. Remining may increase copper recovery by 2 to 10% after the same leach cycle.

46
 Scheffel (2002) recommends ripping the heap surface two to four times in criss-cross direction

and to rip the lower lift before stacking the next lift. Uhrie and Koons (2000) have shown that truck traffic

areas should be ripped to 2.4 m (7.9 ft) deep before irrigation. This supports the recommendations of the

Alligator Ridge Mine to use 3 m (10 ft) deep shanks (DeMull and Womakc, 1984). The severe

consolidation of some heaps at the Alligator Ridge Mine even stalled a dozer equipped with shanks. AT

this mine, ripping temporarily increased the gold extraction rate by establishing new flow paths through the

impermeable zones. However, only the zones actually disturbed by the dozer leached gold thoroughly,

leaving higher-grade ore below. The mine finally chose to also rip immediately after stacking, rather than

only after consolidation.

Ripping the surface of a run-of-mine heap, initially 3 m (10 ft) tall and only 2.1 (7 ft) tall after

irrigation and heavy rainfall, was not successful (Phifer, 1988). The –74 µm (200#) fines, which accounted

for 30-40% of the run-of-mine ore, had already migrated to the bottom of the heap. Ripping the surface

was thus disturbing a couple of feet of the remaining coarse particles.

It is ultimately the proportion of connected pores filled with air that determines the efficiency of

oxygen-based heap leaching systems, whether cyanide, thiosulfate, thiourea, or sulfide minerals.

Therefore, by producing a material of more constant size with agglomeration, there should be fewer

contact points between the wetted surfaces, which, in turn, would reduce the stagnant moisture held up

between agglomerates and increase the gas/liquid surfaces. At Little Bald Mountain, cement

agglomeration did in fact reduce the moisture content of the heap (Tibbals, 1987). In addition, the initial

presence of moisture everywhere in the heap may contribute to the even spreading of the solution during

wetting.

A more porous heap could sustain larger irrigation rates, which may decrease the leach cycle of

certain heap leaching applications. Agglomeration may also lead to a more structurally stable heap,

capable of bearing greater loads. An increase of the heap height directly translates into increased metal

production. If the heap can support more than its own weight, one may opt for multiple lift stacking rather

than dynamic (on/off) pads. Besides the criterion of structural stability, other factors, such as stackers and

heat control in sulfide heaps, determine the heap height.

47
 Used in conjunction with a binder, agglomeration can also increase the overall metal extraction

of a heap stacked with material of smaller size than could have been stacked without agglomeration.

From the perspective of the leach chemistry, the greatest benefits of agglomeration are to

reduce the travel time of reagents and increase the initial recovery rate. These benefits arise because of

the faster contact between the mineral grains and the reagents introduced by the agglomeration solution.

The Alligator Ridge Mine has observed a faster initial recovery (DeMull and Womakc, 1984). Compared

to non-agglomerated heaps, agglomerating gold ores with cyanide also reduced the overall cyanide

consumption. However, adding more than 50 g NaCN/t of ore did not improve the gold extraction and

increased the overall cyanide consumption (DeMull and Womakc, 1984). However, in copper heap

leaching, the better and faster contact of the reagents with the mineral surfaces also lead to undesirable

chemical reactions between gangue minerals and sulfuric acid. To avoid such reactions, the rest period

between stacking and irrigation should be minimized.

According to the sulfide bioheap model developed by the University of British Columbia, sulfide

heap bioleaching environments should also benefit significantly from agglomerating the ore with the

leaching solution (Bouffard, 2003). Model simulations suggest that mixing an inoculum of mesophilic

microorganisms (cell viability between 15 and 45°C) with the ore could eliminate the downward microbial

colonization wave that would otherwise advance through the heap at a slower rate than the barren

solution. With microbial preinoculation of the ore, the simultaneous growth of microorganisms everywhere

throughout the heap could increase the leaching rate. According to the model, the more rapid colonization

of the heap could occur in spite of the relatively few initial number of microorganisms added, a condition

imposed by the appropriate moisture content of the ore (typically 5-15%) and the number of

microorganisms in the inoculum (at the most 109 cells/mL). Pre-inoculating the ore with a variety of

temperature-sensitive microorganisms (mesophiles, moderate thermophiles, extreme thermophiles) would

further accelerate the oxidation of certain sulfide minerals, such as pyrite and sphalerite. In copper heap

leaching, the effects on the microbial viability of simultaneously mixing the inoculum with the 5-45 kg

concentrated sulfuric acid/t of ore are not well understood.

Agglomeration was a technical and economical breakthrough technology for heap leaching of

clayey ore and ore containing high fines content. Figure 15 and Figure 16 illustrate the most important

48
performance indicator of the success of agglomeration. Up to 80% of the metal value of tailings could be

extracted in 20 to 70 days. Gold or silver extraction from crushed ore was as high as 90% in sometimes

as little as 10 days. Agglomeration at an Arizonian silver heap leach operation yielded incredible results.

The extraction increased from 37% to 90%, while the leaching time dropped from 90 to only 7 days.

Significant improvements were also obtained at a North Nevada gold heap leaching operation where the

gold extraction increased by 60%, and the leaching time was reduced by half from 50 to 20-30 days. The

Alligator Ridge Mine in Northeastern Nevada has well documented the implementation of agglomeration

of crushed ore at its site (DeMull and Womakc, 1984; Strachan and van Zyl, 1987). Ten pads were tested

for cement dosage, lime dosage, and stacking method. The combination of agglomeration at 50 g NaCN/t

of solution and 1.5-5 kg lime/t of ore and a different stacking method (pushing the agglomerates up the

slopes) increased the gold recovery to 70% and reduced the leach time from 60-90 d to 30-40 d. The

heaps contained a reasonable amount of moisture (5-11%), no saturated zone, and leached uniformly.

100

90

80

70
Recovery (%)

60

50
40

30
20

10

0
0.01 0.10 1.00 10.00 100.00
Approximate maximum particle size (mm)

Figure 15 Influence of agglomeration of tailings (on the left of the X-axis) and crushed ore (on the

right of the X-axis) on the precious metal recovery.

49
 70

60

50

Leach time (d)

40

30

20

10

0
0.01 0.10 1.00 10.00 100.00
Approximate maximum particle size (mm)

Figure 16 Influence of agglomeration of tailings (on the left of the X-axis) and crushed ore (on the

right of the X-axis) on the leach time.

From an environmental perspective, the better solution distribution in an agglomerated heap

should also increase the recovery of the remainder of the soluble metal value during rinsing. It may also

reduce the duration of the rinse cycle or the volume of wash water applied. Producing pellets that will

remain intact long after leaching and rinsing may also reduce dust emissions.

12. CONCLUSIONS

This review paper reported on advances in agglomeration in the heap leaching industry. To

achieve the greatest benefits, the agglomeration process depends on the proper characterization of the

material and on the appropriate selection and design of the agglomeration equipment. Such decisions

affect, in turn, the design and operating conditions of the heap. However, as demonstrated at the Alligator

Ridge Mine, the proper combination of crushing, agglomeration, stacking, and irrigation guarantees the

success of heap leaching operations.

Agglomeration was a breakthrough technology for heap leach producers faced with ore of high

fines or clay content. Achieving up to 80-90% precious metal recovery from this material at first thought to

be heap-unleachable attests of the success of agglomeration.

The benefits of increased recovery and shorter leach cycles must be weighed, though, against

the 5-10% extra capital costs incurred with agglomeration. With regards to the agglomeration operating

50
costs, labor and energy costs evaluated at $US 0.10 to 0.30 per tonne of ore pale in comparison to the

cost of the binder alone at roughly $US 1.00 per tonne of ore. The total agglomeration operating costs

account for 10-20% of the total.

All operations reviewed that practiced agglomeration used a binder. Precious metal heap

leaching operations prefer cement in an amount from 2.5 to 10 kg cement/t of ore (5-20 lb/ton) added to a

cyanide solution containing typically less than 300 ppm NaCN. Copper heap leaching operations

agglomerate with typically 20 kg sulfuric acid/t of ore (80 lb/ton) and 80 kg water/t of ore.

To reduce the binder costs, a combination of two or more binders could be envisaged. The

performance of polymeric binders should also become more predictable, as they are currently one of two

binders suitable for copper ores. A greater selection of inexpensive binders, tolerant of low pH’s and

resistant to microbial attacks, should be developed for the copper heap leaching industry.

Most heap leaching operations, including some of the largest heap leach producers, have

chosen mobile or spreader conveyors for stacking and drums for agglomeration. The residence time in

the drum and the amount of moisture added determine the agglomerate size and size distribution, but not

necessarily the agglomerate strength. A residence time of less than 60 seconds in industrial drums does

not live up to recommendations from laboratory trials. An empirical equation that includes the drum

diameter, drum length, and rotation speed was proposed to predict drum throughputs of less than 5000

tonnes per day.

A non-recognized disadvantage of using drum agglomerators is the production of large

agglomerates or not sufficiently porous agglomerates. The influence of the pore length and tortuosity

inside an agglomerate on the diffusion rate has been examined up until now from a theoretical standpoint

only (Bouffard, 2003). Future trials should attempt to define the optimum agglomerate structure.

Although not representative of the unsaturated conditions in a heap, laboratory dipping, flooding,

or compaction tests can help compare the performance of the binders on a quantitative basis and optimize

the binder requirements. In this author’s opinion, none of these methods, including even the full height

columns or silos of several square feet in surface area, can accurately predict the expected performance

of agglomerates in an industrial-scale setting. Test pads are undeniably necessary for scaling-up

production. On-line methods for controlling the moisture and size of agglomerates are now available

51
industrially. Nevertheless, the criteria of size and moisture content, as well as an increased in metal

recovery, do not directly relate to the true benefits of agglomeration, i.e. increased porosity and uniform

flows. What are needed are methods for measuring the strength of agglomerates. Such methods exist

for agglomerates made up of fines, i.e. particles of the same size, but none is available for agglomerates

comprised of particles of very different sizes. What are also needed are methods for visualizing the

movement and possible disintegration of agglomerates in a heap, and for measuring the moisture

retention, air pore space, and pore connectivity in the heap. Existing methods for measuring the moisture

content, such as electrical resistivity tomography, are not properly calibrated in the field, and current

methods for measuring pore space lack scalability. Hence, other than the financial benefits of

agglomeration on the bottom line, opportunities still exist to quantify the real physical impact of

agglomeration to ultimately make this process more controllable, predictable, and robust.

13. ACKNOWLEDGEMENTS

The author expresses sincere thanks to Mr. Ryan Baker and the staff of the Placer Dome Inc.

Library for collecting and referencing papers. The author is grateful to Dr. Robert H. Cuttriss, Dr. Ralph P.

Hackl, and Dr. J. Pierre Constantineau for helpful suggestions in the preparation of the manuscript.

14. REFERENCES

Ahmadiantehrani, M., Hendrix, J.L., and Nelson, J.H., 1991, Hypochlorite pretreatment of a low grade
carbonaceous gold ore, Part II: Effects of agglomeration, temperature and bed height, EPD Congress
1992, Ed. J.P. Hager, The Minerals, Metals & Materials Society, pp. 63-79.

Bouffard, S.C., 2004, Personal information acquired during visit.

Bouffard, S.C., 2003, Understanding the heap biooxidation of sulfidic refractory gold ores, Ph.D. Thesis,
The University of British Columbia, 351 pp.

Butwell, J.W., 1990, Heap leaching of fine agglomerated tailings at Gooseberry Mine, Nevada, Advances
in Gold and Silver Processing, Chapter 1, pp. 3-13.

Capes, C.E., 1980, Particle size enlargement, Elsevier Scientific Publishing Company.

Chamberlin, P.D., 1980, Heap leaching and pilot testing of gold and silver ores, Paper given at the
Precious-Metals Symposium, Sparks, NV, November 17-19, 1980, Nevada Bureau of Mines, pp. 77-83.

Chamberlin, P.D., 1986, Agglomeration: cheap insurance for good recovery when heap leaching gold and
silver ores, Mining Engineering, 38(12), pp. 1105-1109.

52
Cifuentes, R., 1999, Heap leach agglomeration/percolation extraction aids for enhanced gold and silver
recovery, WO 99/63123.

Cruz, A., Lastra, M. and Menacho, J., 1980, Sulphuric acid leaching of copper silicate ores, Leaching and
Recovering Copper from As-Mined Materials, pp. 61-70.

DeMull, T.J., and Womakc, R.A., 1984, Heap leaching practice at Alligator Ridge, Chapter 2, Au & Ag
Heap and Dump Leaching Practice, Proceedings from the 1983 SME fall Meeting, Salt Lake City, Utah,
October 19-21, 1983, (Ed. J.B. Hiskey), Sponsored by Solution Mining Committee, Mining and Exploration
Division, Society of Mining Engineers of AIME, pp. 9-21.

Efthymiou, M., Henderson, L., and Pyper, R., 1998, Practical Aspects of the Application of Extract-Ore®
9560 for Heap Leach Agglomeration at Nifty Copper Operation, The world’s knowledge, 15 pp.

Eisele, J.A. and Pool, D.L., 1987, Agglomeration heap leaching of precious metals, CIM Bulletin, 80(902),
pp. 31-34.

Farias, L.L, Reghezza, A.I., Cruz, A.R., Menacho, J.L., and Zivkovic, Y.D, 1995, Acid leaching of copper
ores, Copper Hydrometallurgy Short Course, Copper 1995, Eds. D. Dreisinger and J.D. Vasquez, pp. 13-
17.

Fernández, G.V., 2003, The use of electrical conductivity in agglomeration and leaching, Copper 2003 –
Cobre 2003, Volume VI – Hydrometallurgy of Copper (Book 1), Eds. P.A. Riveros et al., Santiago, Chile,
pp. 161-175.

Garcia, A.J. and Jorgensen, M.K., 1997, Agglomeration and heap leach testing requirements for high clay
ores, Randol Gold Forum ’97, Conference and Exhibition, May 18-21, 1997, Monterey, California,
Publisher Randol International Ltd., pp.143-146.

Gomes, J., 1983, Heap leaching of gold ores, Fifth Annual Conference on Alaskan Placer Mining,
Fairbanks, AK, USA, pp. 80-83.

Gross, A.E., 1990, Polymeric ore agglomeration aids, U.S. Patent 4,898,611.

Hall, J. S., 1986, Sizing of particulate-water-air agglomerates using liquid nitrogen. Chemical Engineering
Science, 41(1), pp. 187-188.

Heinen, H.J., McClelland, G.E. and Lindstrom, R.E., 1979, Enhancing percolation rates in heap leaching
of gold-silver ores, Report of Investigation 8388, Bureau of Mines, Reno, NV, 20 pp.

Herkenhoff, E.C., 1987, Heap leaching: agglomerate or deslime?, Engineering Mining Journal, 188(6), pp.
32-39.

Holle, H., 1996, Process innovations at Girilambone Copper Company’s heap leach/SX/EW operation,
Randol at Vancouver ’96, 5th Global Mining Opportunities and 2nd Annual Copper Hydromet Roundtable,
International Conference and Exhibition, November 12-15, 1996, Vancouver, British Columbia, pp. 289-
299.

James, B. and Lancaster, T., 1998, Physical parameters of heap leaching at Girilambone Copper
Company, Randol at Vancouver ’98, Global Mining Opportunities, Copper Hydromet Roundtable ’98,
International Conference and Exhibition, Vancouver, British Columbia, November 17-20, 1998, pp. 57-61.

Jenkins, J. and Canello, R., 1997, Description of the El Abra plant, Randol at Vancouver ’97, Global
Mining Opportunities and Copper Hydromet Roundtable ‘97, International Conference and Exhibition,
November 4-7, 1997, Vancouver, British Columbia, pp. 183-187.

53
Kappes, D.W., 2002, Precious metal heap leach design and practice, Proceedings of the Mineral
Processing Plant Design, Practice, and Control, Littleton, CO, Society for Mining, Metallurgy, and
Exploration, pp.1606-1630.

Keuhey, K.Y. and Coughlin, W.E., 1983, Getty Mining Company’s approach to heap leaching at the
Mercur Mine, Chapter 6, Proceedings of Au & Ag Heap and Dump Leaching Practice at the 1983 SME
Fall Meeting, Salt Lake City, UT, Editor J.B. Hiskey, October 19-21, 1983, pp. 51-57.

Kinard, D.T. and Schweizer, A.A., 1987, Engineering properties of agglomerated ore in a heap leach pile,
Chapter 10, Geotechnical Aspects of Heap Leach Design, Ed. Dirk van Zyl, pp. 55-64.

Kohr, W.J., 1998, Method of biotreatment for solid materials in a nonstirred surface bioreactor, US patent
5,766,930.

Kosmatka, S.H., Kerkhoff, B., Panarese, W.C., MacLeod, N.F., and McGrath, R.J., 2002, Design and
Control of Concrete Mixtures, Engineering Bulleting 101, 7th Ed., Cement Association of Canada, 356 pp.

Lancaster, T. and Walsh, D., 1997, The development of the aeration of copper sulphide ore at
Girilambone, Proceedings of the IBS-Biomine ’98, Glenside, SA: AMF 1997, pp. M5.4.1-M5.4.10.

Lastra, M.R. and Chase, C.K., 1984, Permeability, solution delivery, and solution recovery: Critical factors
in dump and heap leaching of gold, Mining Engineering, 36(11): pp. 1537-1539.

LeHoux, P.L., 1997, Agglomeration practice at Kennecott Barneys Canyon Mining Co, Global Exploitation
of Heap Leachable Gold Deposits, Ed. D.M. Hausen et al., The Minerals, Metals & Materials Society, pp.
243-249.

Lipiec, A.I. and Bautista, R.G., 1998, Considerations in the extraction of metal from heap leaches, Trans.
Indian Inst. Met., 51(1): pp. 7-16.

Litz, J.E., 1993, Hydrometallurgy: A review of 1992 activities, Mining Engineering, June 1993, p. 609.

MacDonald, A.R.F., 1986, Flocculation of acid leach slurries, U.S. Patent 4,587,108.

McClelland, G.E., 1986, Agglomerated and unagglomerated heap leaching behavior is compared in
production heaps, Mining Engineering, 38(7), pp. 500-503.

McClelland, G.E., 1986, Heap leaching and agglomeration: heap leaching practice in the United States,
Proceedings of the International Conference on Gold, Volume 2: Extractive Metallurgy of Gold,
Johannesburg, SAIMM, pp. 57-69.

McClelland, G.E., Pool, D.L, and Eisele, J.A., 1983, Agglomeration-heap leaching operations in the
precious metals industry, U.S. Bureau of Mines, Reno, NV, IC 8945, 16 pp.

McClelland, G.E., Pool, D.L, Hunt, A.H., and Eisele, J.A., 1985, Agglomeration and heap leaching of finely
ground precious-metal-bearing tailings, Information circular 9034, Bureau of Mines, 11 pp.

Miller, G., 2003, Ore geotechnical effects on copper heap leach kinetics, Hydrometallurgy 2003 – Fifth
International Conference in Honour of Professor Ian Ritchie, Volume 1: Leaching and Solution
Purification, Edited by C.A. Young et al., The Minerals, Metals & Materials Society, pp. 329-342.

Milligan, D.A. and Engelhardt, P.R., 1983, Agglomerated heap leaching at Anaconda’s Darwin silver
recovery project, Chapter 4, Proceedings of the Au & Ag heap and dump leaching practice with panel
discussion ,1983 SME Fall Meeting, Salt Lake City, UT, October 19-21, 1983, Editor J.B. Hiskey,

54
Sponsored by Solution Mining Committee, Mining and Exploration Division, Society of Mining Engineers of
AIME, pp. 29-39.

Outzen, R.L., 1983, Heap leaching at the Borealis gold project, Chapter 4, Practical Hydromet ’83, pp. 19-
22.

O’Brien, M.F., Griffin, J.B., Cuthbertson, W.F., Lamanna, J.R., and Wilton, L.E., 2003, Operational
changes at Phelps Dodge’s Morenci Mine, Volume VI – Hydrometallurgy of Copper (Book 1), Copper
2003 – Cobre 2003, Santiago, Chile, Eds. P.A. Riveros et al., pp. 39-52.

O’Brien, R.T., 1982, Agglomeration pre-treatment in the heap leaching of gold and silver, Seminar
Carbon-In-Pulp Technology in the Extraction of Gold, Australasian Institute of Mining and Metallurgy,
Symposia Series, pp. 297-311.

Pautler, J.B., Gross, A.E. and Strominger, M.G., 1990, New polymeric agglomeration aid improves heap
leach efficiency at Brewer Gold, Chapter 2, Advances in Gold and Silver Processing, Eds. Maurice C
Fuerstenau and James L. Hendrix, pp. 15-21.

Perez, J.W., Barrois, M.J., McCord, T.H., and O’Neil, G.R., 1990, U.S. Patent 4,961,777,
Pretreatment/agglomeration as a vehicle for refractory gold ore treatment.

Perry’s Chemical Engineering Handbook, Size enlargement equipment and practice, 7th ed., Eds. R.H.
Perry, D.W. Green, and J.O. Maloney, McGraw-Hill, pp. 20.57-20.89.

Phifer, S.E., 1988, Agglomerating gold ores at the Haile gold mine, Mining Engineering, June 1988, pp.
447-450.

Pietsch, W., 1985, Agglomeration, Fortschritte der Verfahrenstechnik. Abt. B, Mechanische

Verfahrenstechnik (Section B, Mechanical process technology). Düsseldorf [Germany], VDI-Verlag, vol.
23, pp.125-139.

Pizzaro, R.S., Pato, V.V. and Ricafort, L.R., 1986, Gold heap leaching operations, Masbate, Philippines,
In Gold 100, Proceedings of the International Conference on Gold, Volume 2: Extractive Metallurgy of
Gold, Johannesburg, South Africa, SAIMM, pp. 71-81.

Polizzotti, D.M., Rice, D.A., and Rak, R., 1999, Polymeric agglomerating agents for the gold mining
industry, Polymers in Mineral Processing Industry, Ed. J.S. Laskowski, pp. 455-462.

Polizzotti, D.M., 1993, Agglomerating agents for clay containing ores, U.S. Patent 5,211,920.

Polizzotti, D.M., Liao, W.P. and Roe, D.C., 1997, Cationic block polymer agglomeration agents for mineral
bearing ores, U.S. Patent 5,668,219.

Polizzotti, D.M., Liao, W.P. and Roe, D.C., 1996, Cationic graft polymer agglomeration agents for mineral
bearing ores, U.S. Patent 5,512,636.

Potter, G.M., 1983, Some factors in the design of heap leaching operations, Nevada Bureau of Mines &
Geology, Report 36, Papers given at the Precious metals Symposium, Sparks, NV, November 17-19,
1980, (Eds. V.E. Kral et al.), pp. 69-76.

Readett, D., Sylwestrzak, L, Franzmann, P.D., Plumb, J.J., Robertson, W.R., Gibson, J.A.E., Watling, H.,
2003, The life cycle of a chalcocite heap bioleach system, Hydrometallurgy 2003 – Fifth International
Conference in Honor of Professor Ian Ritchie, Volume 1: Leaching and Solution Purification, Eds. C.A.
Young et al., TMS (The Minerals, Metals & Materials Society), pp. 365-374.

55
 Rose, W. R., Pyper, R., Eguivar, J. and Mirabal, C., 1990, Heap leaching with fine grinding and
agglomeration at Potosi – Bolivia, Chapter 3, Proceedings of the Symposium at GOLDTech 4, Ed.s
Maurice C. Fuerstenau and James L. Hendrix, pp. 23-32.

Scheffel, R.E., 2002, Copper heap leach design and practice, Proceedings of the Mineral Processing
Plant Design, Practice, and Control, Littleton, CO, Society for Mining, Metallurgy, and Exploration, pp.
1571-1605.

Scheffel, R.E., 1982, Retreatment and stabilization of the Naturita tailing pile using heap leaching
techniques, Chapter 27, Interfacing Technologies in Solution Mining: Proceedings of the 2nd SME-SPE
International Solution Mining Symposium, AIME, New York, pp. 311-321.

Schnell, H.A., 1997, Leach practice at Quebrada Blanca, Randol at Vancouver ’97, Global Mining
Opportunities and Copper Hydromet Roundtable ‘97, International Conference and Exhibition, November
4-7, 1997, Vancouver, British Columbia, pp. 319-321.

Serrano, E.M., 2003, New agglomeration techniques for the acid percolation leaching of clay-bearing
copper ores, Volume VI – Hydrometallurgy of Copper (Book 1), Copper 2003 – Cobre 2003, Santiago,
Chile, Eds. P.A. Riveros et al., pp. 189-201.

Strachan, Cl and van Zyl, D., 1987, Feasibility assessment for increasing heap thicknesses at the Alligator
Ridge Mine, Chapter 11, Geotechnical Aspects of Heap Leach Design, Ed. Dirk van Zyl, pp. 65-76.

Tibbals, R.L., 1987, Agglomeration practice in the treatment of precious metal ores, Proceedings of the
International Symposium on Gold Metallurgy, Winnipeg, Canada, August 23-26, 1987. New York:
Pergamon, 1987. pp. 77-86.

Trincado, L., Troncoso, U., Vargas, C., Bunger, C. and Zárate, G., 2003, Process improvements at
Mantoverde heapleach-SX-EW plant, Hydrometallurgy 2003 – Fifth International Conference in Honor of
Professor Ian Ritchie, Volume 1: Leaching and Solution Purification, Eds. C.A. Young, A.M. Alfantazi,
C.G. Anderson, D.B. Dreisinger, B. Harris, and A. James, TMS (The Minerals, Metals & Materials
Society), pp. 343-349.

Uhrie, J. L. and Koons, G. J., 2001. The evaluation of deeply ripping truck-dumped copper leach
stockpiles. Mining Engineering. 53 (12). December 2001, pp. 54-56.

Uhrie, J.L., Wilton, L.E., Rood, E.A., Parker, D.B., Griffin, J.B., and Lamanna, J.R., 2003, The
metallurgical development of the Morenci MFL project, Volume VI – Hydrometallurgy of Copper (Book 1),
Copper 2003 – Cobre 2003, Santiago, Chile, Eds. P.A. Riveros et al., pp. 29-37.

Videau, G. and Roche, M., 1990, Developpement industriel d’un nouveau procede d’agglomeration de
mineral uranifere par boulettage avant traitement par lixiviation en tas (voie acide): application au mineral
tres argileux de Nord-Aquitaine, Industrie Minérale Mines et Carrières. Les Techniques, Mars-Avril 1990
(2-3). pp. 135-142.

Walker Jr., D.D. and Oliphant, J., 1992, Composition and method for agglomerating ore, U.S. Patent
5,116,417.

Worstell, J.H., 1987, Guide to successful precious metal heap leaching, Paper presented at the African
Mining Conference, Institute of Mining and Metallurgy, pp. 425-436.

Zárate, G.E. and Guzmán, J.C., 1987, Gold tailings processing by heap leaching, Chapter 24, Small
Mines Development in Precious Metals. Littleton, Colorado: Society of Mining Engineer, pp. 151-156.

56

View publication stats