FLOTATION OF CHRYSOCOLLA
a thesis submitted for the
degree of PhD in Engineering
in the
University of London
by
Anthony John. Wright BSc(Eng)
-partment of Mining and Mineral Technology
Royal School of Mines
London SW7
Janaary 1964 2.
1. SUMMARY. The existing information on the mineralogy and beneficiation of the hydrated copper silicate mineral, chrysocolla, has been critically analysed in conjunction with new data here determined on a typical specimen of the mineral obtained from the Kolwezi mine of Union Miniere du Haut Katanga. It is concluded that chrysocolla is a rigid xerogel with a crystallite size of 150 R. The most important physical property of the Kolwezi chrysocolla is its ability to sorb and desorb water: this supposes a porous structure, which has been confirmed.. The pores, which have an average diameter of 35-40 .g account for 40% (by volume) of the mineral. The relationship between the structure and solubility of chrysocolla, and its behaviour in xanthate and sodium sulphide - potassium xanthate flotation systems has been thoroughly investigated. It is concluded that flotation of chrysocolla is possible in either system, providing that large quantities of reagent are used and that the flotation conditions are quiescent. More turbulent flotation conditions reduce the floatability of chrysocolla by scouring the ' hydrophobic species from the external surface. In both quiescent and turbulent flotation conditions, the major portion of the hydrophobic species is formed either in the bulk solution or within the mineral pores, and not at the external mineral surface. The hydrophobic species are coprecipitated cuprous xanthate and dixanthogen. The effect of soluble sulphide is to reduce the subsequent reaction between potassium 3• xanthate and chrysocolla: no appreciable reaction occurs between precipitatedcupric sulphide and potassium xanthate.
Improved plant practice is considered possible providing the attrition of chrysocolla is minimised and reagent concentrations are maintained at high levels. 4.
2. INTRODUCTION.
Hydrous copper silicate minerals of the chrysocolla group are not industrially selectively recovered by flotation, despite the fact that various workers (e.g. 1 4 \ Arbiter , Bowdish and Chen`, Jaekel3 and Kovacs ) have shown flotation to be possible in laboratory-scale investigations.
It is particularly noted that many of the successful laboratory investigations are similar in some - even many - respects to the familiar sulphidation- xanthate process by which copper (and lead) carbonate minerals have been successfully beneficiated for a number of years on an industrial scale. Chrysocolla, however, is not apparently amenable to the industrial sulphidation- xanthate process. The mechanism of successful laboratory processes and the reasons behind their industrial inapplicability, have been largely ignored by many workers in this field. Typically (e.g. Dean-5-, Jal,kel3 and Ludt and Dewitt6), they have confined their researches to an assessment of floatability after some specific conditioning process. In the work described here, a more fundamental approach has been attempted. The aim of this research has been to define the basic physical, physico-chemical and chemical properties of a particular sample of chrysocolla, and show how these may preclude the satisfactory recovery of chrysocolla in present-day industrial flotation systems. Thus, it has not been the intention here to define the conditions for the successful industrial flotation of the Kolwezi chrysocolla. However, the work has suggested certain improvements which might be advantageously applied to the treatment of ores in which chrysocolla is a constituent. 5.
3. NOMENCLATURE. In 1950, Hey7 considered that the data available on hydrous copper silicates pointed to the existence of only three distinct varieties, that is dioptase, plancheite and chrysocolla. More recent work by Neumann et alb and Poljak and Gordillo', however, has revived the earlier contention that a further variety, shattuckite, is distinguishable. In order that the properties of chrysocolla may be more readily distinguished, the essential properties of these other hydrous copper silicates are now briefly discussed, and summarised in Table 1.
The physical and optical properties ofaoptase have 10 been well-described by Dana . It is a transparent, emerald-green mineral having hexagonal symmetry. X-ray powder diffraction data are reported by the ASTM powder 11 12 data file , Billiet and Neumann et a18, although the data of Neumann et alb are not in accord with those of the other two sources. The crystal structure of dioptase has been deduced by Belov et a113, Heidel4 and Moenke15, who conclude that the basic structural unit consists of six silicon-oxygen tetrahedra in a ring formation. Differential thermal analyses showing good agreement 16 have been reported by both Ivanova and Toussaint The principal feature of these analyses is an endothermal effect at approximately 550°C caused by the loss of structural water. The mineral is generally represented by the formula Cu6(Si6018). 6H20.
Although plancheite is not so well established as a distinct species, it is generally described as a pale- blue mineral having orthorhombic symmetry. Concordant x-ray powder data for various samples of the mineral are 6.
TABLE 1: PRINCIPAL PROPERTIES OF DIOPTASE AND PLANCHEITE.
MINERAL DIOPTASE PLANCHEITE
Colour Emerald-Green Pale-Blue
S.G. 3.3 - 3.4 3.3 - 3.8
Crystal System Hexagonal Orthorhombic
R.I. 1.64 - 1.71 1.64 - 1.72
2.11 3.32 d spacings - 2.60 3.50 7.24 4.40 DTA peaks °C 550 endo. 300 exo, 670 endo 7. 12 19 reported by Billiet , Guillemin and Pierrot18, Millman , 20 and cumin and Lasheva . The principal 'd' spacings are at 3.32, 3.50 and 4.40 R, although the work of Neumann et alb doesnot agree with these values (c.f. dioptase). Only Toussainti7 has recorded a differential thermal analysis for a plancheite sample: it shows a broad exothermal peak at 300°C, and a sharp endothermal peak at 670°C. Evidence that shattuckite is a distinct species is 21 put forward by Ford , Neumann et alb and Poljak and 21 Gordillo9. Ford describes shattuckite as being more dense than plancheite (e.g. 3.8 against 3.3), and having higher refractive indices (c.1.78 against c.1.66). Neumann et al8 and Poljak and Gordillo9 quote x-ray powder data which are different from those of dioptase, plancheite (and chrysocolla), although the two sources 12 22 are not in agreement. Although Dilliet , Sun and Toussaint17 claim the identity of shattuckite as a distinct species, their published data agree very well with that of plancheite.. Clearly, plancheite and shattuckite are closely related, if not identical, on the evidence presented here. The composition of these minerals, which for convenience will be termed plancheite, varies from 5Si02.6Cu0.1.5H20
to 14CuSiO3.6H2O. Hydrous copper silicates which do not conform to the above-mentioned properties (see Table 1) of dioptase and plancheite may be described under the group heading of chrysocolla. Different samples of this group, whilst manifesting various external forms and having different chemical compositions have been shown to exhibit very similar fundamental properties. Semmons23, in 1878, was 8. the first to suggest a chrysocolla group of minerals; but at that time - and for many years to follow - inadequate mineralogical techniques precluded the determination of any but the most superficial of properties, and his suggestion was somewhat impractical. Present-day methods of evaluating basic properties by x-ray and electron diffraction, differential thermal analysis, infra-red spectroscopy and thermogravimetric analysis enable the Semmons23 classification to be implemented. Included in the chrysocolla group of minerals are the previous designations asperolite 10‘ (Dana ),cornuite (Rogers2424, ), katangite (Billiet) and 10 pilarite (Dana ). 9.
4. REVIEW OF CERYSOCOLLA MINERALOGY a. Formation and occurrence.
Hydrous copper silicate minerals, including chrysocolla, are of epigenetic origin, and similar materials have been synthesised by Belov et a125 and Shcherbina and Ignatova26. Their methods involved the dehydration of gelatinous precipitates produced by reacting a solution of a copper salt with a solution of sodium metasilicate under alkaline conditions. Chrysocolla is frequently found either in massive forms, or as pseudomorphs encrusting, or interlayered with other copper minerals. Chrysocolla occurs in many oxidised Copper orebodies, including those in Chile, Congo, South-west USA and the Urals in the USSR. The layered appearance of marry samples of chrysocolla 27 known as colloform banding, is considered by Edwards , to have resulted from either of two processes. Continuous precipitation may have occurred from solutions which periodically changedin copper or silica concentration, or precipitation may have resulted from solution evaporation, followed later by a fresh influx of solution. b. Constitution.
Although Danal°, Winchell and Winche1128, Read29 and others represent chrysocolla by the formula CuSiO3.2H20, there are very few analyses in support of this. The majority of analyses show either an excess or a deficit of CuO, SiO 2 or H2O over the CuS103.2H20 formula, together with the frequent presence of other oxides especially A1203, Fe203, CaO or MgO. 31 23 After Foote and Bradley", Palmer and Semmons had noted the dependence of the.water content of 10.
chrysocolla upon relative humidity, some mineralogists modified the formula CuSiO O. 3.2H20 to CuSiO3'nH2 However, the relative humidity-water content relationship has not been closely studied, and the relative humidity to which reported chemical analyses apply, or at which other properties have been determined, is never stated. c. Optical, physical and chemical properties.
Because of the variability of chrysocolla, determinative mineralogical data for the mineral have tended to be generalised: specific properties are seldom associated with a specific chrysocolla of a particular composition. Some of the published data on optical and physical properties are summarised in Table 2. The optical properties especially, can vary in samples of apparently similar composition. It is reported that heating chrysocolla results in an evolution of water at temperatures near 100°C and, depending upon the atmospheric environment, a blackening (Ford21) or reddening (Yakhontova33) of the sample at higher temperatures. This is acknowledged to be caused by structural changes, yielding mixtures of cupric or cuprous oxide and silica. It is also reported that dilute acids decompose 21 chrysocolla without effervescence. Ford states that this occurs without gelatinisation of the resultant silica, but Butts34 that gelatinisation does occur. In neither case, however, is the term tgelatinisationt defined.
d. Structure. In 1916, Foote and Bradley30 proposed that chrysocolla should be classified as an amorphous 11.
TABLE 2: SOME EXAMPLES OF PUBLISHED OPTICAL AND PHYSICAL PROPERTIES OF CHRYSOCOLLA. i Source Crystal Refractive Hard- Formula 1 Colour S.G. of Data 1 Data Indices ness 10 N = 1.46 2.4 2-2.24 Dana CuSiO mountain- uniaxial o 2H 0 3. green N = 1.57 2 bluish- + E green crypto- sky-blue crystalline turquoise- blue
N = 1.46 c.2 c.2.4 Winchell CuSiO green, uniaxial o and 2H 0 3. blue + 2 2 N = 1.54 Winchell orthorhombic? E
biaxial N = 1.575, x - 1.585 orthorhombic? N = y 1.597 N z = 1.598 1.635 29 Read CuSiO bluish- amorphous - 2-4 2-2.2 2H 0 3. green 2 sky-blue turquoise- blue
Larsen CuO.Si0 green uniaxial mod. 2+ 2 :2' and nH 0 + birefri. 2 Berman32 N = 1.40+
N = 1.575 2+ 2.4+ Cu0.Si02' green biaxial x nH2O? - Ny = 1.598 N z = 1.597 12. hydrogel, because of its variable composition, its water sorption phenomenon and cryptocrystallinity indicated by the difficulty in determining optical properties. Suc. 1 a classification has been supported more recently by Toussaint17, who nonetheless acknowledges that a certain degree of crystallinity is apparent in many chrysocolla samples. On the other hand, Chukhr.)vand Anosov35'36 and Yakhontova33 have suggested that chrysocolla is more analogous to minerals of the montmorillonite group. (Montmorillonite has a layered structure which is ordered only in two dimensions: individual sheets in the structure are mutually parallel, but randomly tvisted relative to one another). The analogy is based principally upon chemical amalysis,Showing the existence of intermediate minerals between a chrysocolla of the standard constitution and copper-free montmorillenite; and secondarily, x-ray powder diffraction-patterns, the high variable water content, similar ion-exchange capacities and similar microscopic properties. In general, x-ray diffraction data for chrysocolla have not been interpreted quantitatively, probably because of the diffuseness of the patterns that are invariably obtained, Whilst the data of Table 3 indicate obvious variations between samples, there are, nevertheless some similarities - notably the occurrence of lines at approximately 1.5, 2.9, 3.3 and 4.4 X. Of these, only 1.5 and 2.9 X lines would seem to be characteristic; the other two lines are also found in plancheite and the 3.3 2 line, in dioptase, The application of differential thermal analysis to chrysocolla chiefly lies in determining the tenacity 13.
TABLE 3: X-RAY POWDER DIFFRACTION DATA ON SAMPLES OF CHRYSOCOLLA.
Source I d spacing range X of From 1.44 1.63 1.88 2.24 12.78 3.7 5.6 Data I 1 To 1.43 1.62 11.87 2.23 2.77 3.6 5.6 12 Billiet 1.491 1.64 2.49 X2.912 8.3 11 1 ASTM 1.48 2 2.46 .81 4.4 2.92 3.13
Kovacs 4 1.81 2.12 2.27 2.852 4.2 2.45 6.31 Neumann et al 1.36 1.49 2.27 3.3 4.5 2.571 3.7
Chukhrov and 1.31 1.48 2.54 2.86 Anosov 35 1.62 37 2 3 1 Sun 1.32 1.48 2.32 2.32 4.4 15.0 1.61 2.42 12.56
(Superscript figures, where shown, refer to the intensity of the most prominent lines) 14. of water to the remaining structure. The analyses of Table 4 show that in all cases an endothermic reaction occurs below approximately 250°C. Where this has been interpreted, it is ascribed to the loss of weakly-bound water (Chukhrovand Anosov35 confirmed this by thermogravimetry). No general agreement exists, however, in the interpretation of the other thermal reactions shown; they are variously ascribed to further water losses, phase changes in silica and crystallisation of cuprite or tenorite. 25 Only Belov of al have worked with synthetic chrysocolla, for which they have suggested a definite schistose structure. A differential thermal analysis of their material disproved the existence of an intimate Si02-Cu(OH)2 admixture, and showed a definite thermal effect at 640°C as a result of CuO crystallisation. In recent years, the determination of chemical groupings by infra-red spectroscopy has been applied successfully to various minerals, notably complex silicates. Again, the method is particularly useful for chrysocolla in determining the state of 17 constitutional water. Only Kovacs, Sun37 and. Toussaint appear to have applied the method to chrysocolla and 17 only Sun37 has recorded his spectra. Nether Toussaint nor Sun37 were able to detect OH groupings (as in montmorillonite), and Sun37 definitely refuted the earlier montomorillonite analogy. 15.
TABLE 4: REPORTED DIFFERENTIAL THERMAL ANALYSES OF CHRYSOCOLLA.
Source Temperature of Thermal Peaks-0C i of Data Endothermic Exothermic (Chukhrov and 38 80 300 McLaughlin Anosov do not, Kauffmannand 140 690 ; 910 in fact, assign Dilling3": 'endothermic' or Yakhontova33 220-250 700 ; 900 'exothermic' to 17 reactions listed Toussaint 155 670 ; 970 here. This Chukhrov and 120-140 ; 450-530; 1040 interpretation of Anosov35 690-700 their data has 16 170 ; 1030 Ivanova 340; 660 ; 910 been obtained 150 ; 1000 650; 850 from Chemical 110-220 660-700 ; 910 Kovacs4 Abstracts 44-7724b) 65-130 670 22 Sun 154 705 ; 920-980 37 Sun 150 ; 400-625 ; 950 700 ; 950
Sumin and 110 670-690 ; 300-330; Lasheva20 950 5. RESULTS AND DISCUSSION OF CHhicACTLRISATION OF CHRYSOCOLLA SAI4PLE. a. Preparation of sample for testwork.
The chrysocolla sample used for this investigation was obtained from the Kolwezi mine of Union Miniere du Haut Katanga, SA, Congo. As received, the material was a mixture of a turquoise-blue mineral (described by Union Miniere as chrysocolla, and hereafter referred to, and used, as such) of colloform appearance, occasionally banded with lighter blue minerals and malachite. These lighter blue minerals (ranging in colour from pale blue to white) were, however, more frequently encrusted upon the chrysocolla, together with a black resinous stain: both were shown by x-ray diffraction techniques to be similar materials to the chrysocolla, but they are not the subject of this thesis. Sulphides were not apparently present in the material, and the gangue minerals consisted of a grey, spongey, siliceous mineral of low specific gravity, and quartz, together with minor amounts of other rock-forming minerals.
An exploratory test showed that a visually homogeneous sample of chrysocolla was unobtainable using heavy liquid separation. Not only was a very poor separation made, but the liquid, bromoform, was sorbed into the structure, 30 a confirmation of the observations of Foote and Bradley with tetrabromoethane. The magnetic properties of chrysocolla and its encrustations proved too similar for effective separation on a Frantz tisodynamic, separator.
It was therefore necessary to upgrade the chrysocolla by hand, removing any contaminating minerals at the coarsest possible size, (about 1/4"). The selected mineral was stage-ground in a porcelain mortar to produce fractions 17. finer than 72 mesh BS; these were separated, dry, from any non-magnetic contaminant in the Frantz separator and then washed once in water, once in absolute alcohol, and a further twenty times in distilled water. (The first water wash served to fully hydrate the mineral, and prevent excessive subsequent sorption of alcohol). These washing operations were carried out in an ultrasonic device, the Soniclean (manufactured by Dawe Instruments Limited) which removed adhering fines, and almost certainly attrited the surface layer of the particles, since the cleaning solution rapidly assumed a turbid appearance.
The known dependence of the water content of chrysocolla upon relative humidity, suggested that the storage of samples should be at a known relative humidity. It was decided to use a relative humidity of nearly 100%, since the fully hydrated mineral state would be expected to be more stable than the fully desiccated condition, and was, in any case, more useful for the later work on flotation. This was achieved in a desiccator vessel, using a saturated potassium dichromate solution in place of the usual desiccant. Stokes and Robinson have measured the water activity of saturated K Cr 0 solutions 2 2 7 as 0.9800 at 25°C.
A specimen of the material received from Kolwezi is shown in Figure 1, b. Constitution.
In addition to copper, silica and water,semi- qualitative analyses using a quartz spectrograph and x-ray fluorescence spectroscopy showed the following to be present: Ca 0.4%, Mg, Al, Fe, Zn each <0.2%, The analysis for Ca, Mg, Al and Fe was performed on a Hilger E 742 quartz spectrograph, calibrated pgainst standards FIGURE 1: SPECIMTNS CF TflE. Y.()L—EZI C:HRYSOCOLLA
malacLito samile shown for comi7)arison 19. of similar composition. Zinc was not determinable by this means, so an ARL x-ray fluorescence machine was used on a 2g. sample of the chrysocolla. This latter analysis was performed by Mr. A.J. Smith of the Applied Geochemistry Department.
Copper and silica were determined on the chrysocolla by chemical analysis, using scaled-down macro-techniques owing to the small quantity of sorted sample available. The standard iodiometric method was used for copper, the liberated iodine titration being carried out with a N sodium thiosulphate solution, since only 10 mg. of sample were used. Silica was determined gravimetrically as Si02, and subsequently volatilised with hydroftioric acid in order to correct for co-precipitated oxides. The silica was determined on the same sample used for the copper analysis: a double evaporation procedure was employed to render the silica insoluble.
The water content of the sample was measured by thermogravimetry and by investigating the relationship between water content and relative humidity.
A Stanton Massflow thermobalance of lmg sensitivity was used for the thermogravimetric analysis of a 100mg. sample of the chrysocolla, During the analysis the furnace was open to the laboratory atmosphere. The heating rate for the range 20-1200°C was approximately 5°C/minute. The result of the analysis is shown in Figure 2.
A quantitative assessment of the dependence of water content on relative humidity was initially attempted as a batch method involving periodic transfer of a 200mg. sample from one fixed relative humidity environment to another, the humidities being fixed by saturated salt solutions beneath the sample. The procedure proved inaccurate and 20.
FIGIIQF 2: THFEC0GFIVF:IFTPIC ANALYSIS OF KOLWEZI CPRYSOCOLLA
wei:-.71-t 0 2 R5r,10 1..n0s102-1-67HO loss: %
10
0.P5C40.1.00-i02. 0.7H20 15
0.85cu0.1.00si02 20
0.21C110., SiO - 2
1
0 200 400 630 Roo 1.700 12'X o temperature: C 21.
involved considerable manipulation: it was abandoned in favour of the continuous gravimetric method of McBain and 41 Bakr , which was made 200 times more sensitive than the salt solution method. In the continuous method, water vapour was introduced to a previously evacuated sample suspended from a fused silica spring balance. Equilibrium water vapour pressure was measured directly with a mercury manometer. A diagram of the apparatus used is shown in Figure 3. Water sorption experiments were performed at a temperature of 20' + 1°C, achieved by siting the entire apparatus in a thermostatically-controlled, insulated room. They were initiated by loading the sample into the glass bucket, evacuating the apparatus and subsequently introducing water vapour in stages from an air-free water reservoir. When the sample had sorbed an equilibrium amount of water, the spring extension and the manometer levels were recorded with two travelling microscopes. After thus gradually hydrating the sample, it was then dehydrated in stages, the equilibrium weights and pressures again being recorded. The water sorption isotherm is shown in Figure 4, and an hysteresis loop is clearly apparent. Quite apart from the structural significance (which will be discussed in Section 5d) of the differing hydrations obtained on absorbing and desorbing water between relative humidity = 40% and relative humidity = 80%, this sorption hysteresis means that a particular relative humidity will not conclusively fix the hydration of the chrysocolla. It will also be necessary to know whether the sample hydrated or dehydrated to its present state and from which hydration it began. Not determined here, but nevertheless a contributory factor, will be the rate of hydration or dehydration, since this will not be an instantaneous process. In the apparatus used for determining the isotherm, 22.
FIG7IF 3: TH. S.-';2T._; BALANCF
vacuum taps
water reervoir
joints
fused :silica
spring b'ilance
mercury
manometer
glass bucket 23.
FIGURE 4: WATER SORPTION ISOTHERM OF .:0LWEZI. CITRYflCOLLA weight 0 1.67 moll loss: ••••-• of wa t it per mol SiO • • 1.45
5 1.22
0
1.0C,
.
A• I 10 /, x.77
• desorticn 0 absorption
A
15 0.3';
0 20 40 60 80 100
relF,tive humidity: 24. equilibrium hydration was not normally attained in less than half-an-hour.
When the sample was fully hydrated, the Kolwezi chrysocolla had the analysis shown in Table 5. Two results have been shown, the first calculated from chemical and spectrographic Cu, Ca and Si02 analyses H2O being deduced by difference. The second set has analysis, which is avoided the use of the chemical SiO2 the most likely to be in error, as a result of handling losses, the dispersion of silica, and the small weight of sample used: only the chemical Cu analysis, ane the deduced H2O content (from the thermogravimetry) have been used. The deduced H2O content has been obtained by assuming that the entire weight loss sustained on heating the chrysocolla to 700°C, was caused by the evaporation of free and combined water only.
Using a mean value of 14.5% for water loss at low temperatures (14% by thermogravimetry, 15% from water sorption measurements), an analysis has been calculated for the least hydrated state of this chrysocolla; it has been included in Table 5.
Water that is not removed on heating to 100°C is evolved between 280°C and 640°C, apparently irreversibly. By 600°C, only 13% of the 19% of original total H2O can be re-absorbed, which corresponds closely to the removable water value of 15%: this. may be coincidental, since radical structural changes which are clearly occurring in the mineral would be expected to influence the quantity of interstitial water. This high temperature water loss is accompanied by a colour change in the mineral, which remains the same on rehydration: by 500°C it had assumed a greenish-black appearance, and by 650°C it was black. This colour change is connected with the appearance of CuO 25.
TABLE 5: CONSTITUTION OF KOL\/EZI CHRYSOCOLLA.
H2O Constituent CuO CaO SiO2 Hydration State of Sample Min Max Min Max Min Max Min Max
Chem.+ Spectro. 42.4 0.7 0.6 33.9 10.6 24.1 Assays % 49.4 39.3 Based Upon Cu + Thermogray. 50.2 42.4 0.8 0.6 44.7 38.0 4.3 19.0 os Chem.+ Spectro. 0.95 0.95 0.02 0.02 1.00 1.00 0.90 2.43 Mole Rati Based Upon Cu + Thermograv; 0.85 0.85 0.02 0.02 1.00 1.00 0.32 11.67 26.
:_ines in x-ray diffraction patterns according to Kovacs4, and confirmed here. At 1070°C, a further 3% weight loss occurs. This corresponds to the loss of oxygen from the sample when CuO dissociates into Cu20 and 02.. If all the equivalent cupric oxide was reduced to cuprous oxide, there would be a 4% weight loss. BeeaUse the thermogravimetric sample was cooled in air, which resulted in re-oxidation (and a 3% weight gain), the presence of red cuprous oxide was not directly observed: o. Optical, physical and chemical properties. These have been determined on hand specimens, particulate material in the .-72 100 mesh BS size range, or polished thin sections. Colour: a blue-green, closely corresponding to the British Colour Council's4' colour standard BCC 119. Lustre: dull, subvitreous. Transparency: translucent. Feel: adheres to the tongue. Form: mamillated, colloform. Pseudomorphism: frequently encrusts a grey, spongey, siliceous material, and is frequently encrusted by malachite followed by a lighter blue mineral. Hardness: 3 to 4 on Moh's scale. Fracture: uneven to conchoidal. Specificgravity: Y- is will vary with the degree of hydration of the sample, and was determined here only on fully hydrated chrysocolla by three independent methods: A. Beckman air comparator pycnometer: in this device, the volume of a weighed sample (here 1,,8g) was obtained by air displacement. 27.
B. Heavy liquids: here, single particles were dropped gently into the surface of bromoform-carbon tetrachloride mixtures of various specific gravities, and the mixtures diluted until the particle would just float. However, the absorption of heavy liquids (mentioned in section 5a) occurred even on fully hydrated mineral, and rapidly increased its specific gravity.
C. Liquid pycnometer: this proved the most satisfactory and reproducible method. About 0.5g. of the chrysocolla was used with water in a 5 ml. pycnometer, From the specific gravity obtained by method C, the specific gravity of the mineral in equilibrium with dry air has been calculated from the quantity of sorbable water. It has been observed that only a negligible volume change occurs during dehydration, and it has been assumed that the specific gravity of the hydration water is 1.000. The specific gravity results are summarised in Table 6. Where the specific gravity of the minimum hydration state has been reported, this is a mean value, and includes the voids which must remain after dehydration occurs. From these data, it is possible to calculate the density of the solid matrix alone as follows:
If m , m are respectively the mass of the matrix and the m P pore water, amd v ,v are the corresponding volumes, then mp m + m m m m P m v + v - 2.42 (1); v + v - 2.04 (2); and v—2 = 1.00 (3) M P m P P From (2) = v = mm P 2.04 - vm (4) and from (1) and (3), m = 2.42 v + 1.42 v (5) m m P Substituting (4) into (5) and rearranging: 1.42 mm - vri rv = 2.42 + m vm 2.04 28.
TABLE 6: SPECIFIC GRAVITY OF KOLWEZI CHRYSOCOLLA.
Hydration State Method of Determination Min Max
Air Pycnometer - 2.23
Heavy Liquid - 2.55
Water Pyonometer 2.04* 2.42 * Calculated 29. from which -v2 = 3.33 (6) = specific gravity of matrix. m v _ The pore volume, v is calculated from (2) and (6) as follows: v v 2 m m .04 - 61%, and because - 100% v + v - 3.33 v + v v + v M Py m p m p then v + v - 39%. Refractive index: This is generally determined by locating the Becke line or using the shadow method with the particle immersed in a medium of known refractive index (e.g. Read29). Clove oil,oc-monobromonapthalene or methylene iodide are frequently used as mutually miscible media, but are very easily sorbed by the chrysocolla, - especially since the determination of refractive index can take a considerable time. However, in the absence of an alternative procedure, the method had to be used. In order to increase the speed of determination, and hence the precision, several measurements were made on different particles. The refractive index to sodium light of particles fully saturated with water was thus determined as 1.48 + 0.01 (25°C). Other optical properties: the chrysocolla shows no microscopic cleavage, is not pleochroic, is anisotropic, possesses no definite extinction and shows no interference figures. From such data, it must be concluded that the sample is not non-crystalline, (truly amorphous materials are isotropic) but has a crystallite size below the resolution of optical microscopes, viz: c.2 microns.
Chemical properties: the relevant chemical properties of the chrysocolla are presented and discussed in section 7, under the flotation behaviour of the mineral. Only two properties are considered here, the action of hydrogen 30.
peroxide, and of dilute acids on the mineral.
Hydrogen peroxide results in the chrysocolla blackening. The process is extremely rapid at acid pH's (although it slows considerably in the alkaline region), and is accompanied by the evolution of oxygen; since no such reaction was apparent with clioptase, plancheite or malachite, it may be useful for identification purposes.
Dilute hydrochloric acid solutions readily remove the blue colour from the chrysocolla without visibly altering the particle shape: the boundary between unreacted material and the colourless silica layer appears to retreat from the external surface evenly. The silica layer has far less strength than the original chrysocolla and peels off the unreacted mineral when the solution is agitated to any extent.
When a sample of chrysocolla, which is not fully hydrated, is immersed in solution (including alkaline solution), gas is displaced from the structure: the phenomenon is not apparent with fully-hydrated samples, and thus the gas evolved may be assumed to be air. The process may be repeated after drying. Care must be taken when differentiating between malachite and chrysocolla by the former's gas evolution with dilute acids. The volume of gas evolved with chrysocolla is comparable to the volume of the particle, whereas malachite liberates many times its volume of carbon dioxide when dissolved in acid.
d. Structure.
In addition to the water sorption measurements described in section 5b, four other techniques were employed in an attempt to elucidate the structure of the mineral. They were x-ray diffraction, infra-red absorption spectroscopy, differential thermal analysis and electron 31. microscopy.
X-ray powder diffraction photographs were taken with a 9 cm camera utilising copper radiation and a nickel filter. Finely-ground mineral was agglomerated onto a thin glass fibre with petroleum jelly, and irradiated for two hours. The resultant diffuse and faint pattern is reproduced in Figure 5, and the 'd' spacings recorded in Table 7. The infra-red absorption spectrum shown in Figure 6 was obtained by Miss J. Holroyd of the DSIR Warren Spring Laboratory on a modified S3 Grubb Parsons double-beam spectrophotometer. About lmg of finely-ground mineral was incorporated into a KC1 disc by the standard die-press technique. Although the chrysocolla was fully hydrated prior to incorporation in the KC1 disc, by the time the spectrum was recorded, some interstitial water was almost certainly lost. Since KC1 itself produces an absorption -1 peak at 3360 cm owing to the 0-H stretch vibration of sorbed water, the spectrum of chrysocolla dispersed in a Nujol mull was obtained. This spectrum also included the -1 3360 cm peak, which was thus shown to be a feature of the chrysocolla spectrum.
Figure 7 shows the differential thermal analysis result using a Netzsch apparatus. The analysis was performed at the DSIR Warren Spring Laboratory by Mr.L.D. Muller. Owing to the small bulk of the sample, it was necessary to dilute it with three times its weight of inert alumina; this thermally inert alumina was, however, run simultaneously as a compensating blank. The heating rate was 1000 per minute.
Electron microscopy of pulverised chrysocolla was performed by Mr. A.E.B. Presland of the Chemical •
Fi 5: TIn";)"CITIT 71.7719'-'11 OF KOL','TZI C117(r',OCOLL.A. 33.
TATiLE 7: X-RAY DIb.?RACTION PATTERN FOR KOLWEZI CHRYSOCOLLA.
---1 d Spacing : 1Rngstroms 1.31 1.47 1.61 2.60 2.87 4.4
Line Intensity Weak Strong Medium Medium Strong Strong
Line Width Broad Sharp Broad Tailing Broad Broad
hk0 (Dr.I.S.Kerr) 220 Basal Basal 110 Basal 100
34.
FIGURE 6: INFRARED SPFCTEJM OF KOLWEZI CHFYOCOLLA
t; absorption density: arbitrary scale
4 6 10 12 14 J wavelength: microns 5030 3000 2000 1000 z 4000 wave number: cm-1 35.
FT7,quE 7: :IFrEq'TTPI TIT7PM,',L ANALYSIS OF KDLWFZI CHFYSOCOLLA
(broken line io assumed base—line)
tem- perature differ- exoth9rmic reactions ence: arbitrar scale......
endothermic reactions J
200 400 600 Boo 1000 1200
temperature: °C 30.
Engineering Department with a Siemens Elmiskop 1 microscope at 40000 linear magnifications. Two samples were prepared for microscopy: for both samples chrysocolla was pulverised in water, in an agate tiortar for several hours and sized by sedimentation. Chrysocolla which was nominally finer than 311 (considerable flocculation occurred during sedimentation, and it would be expected that the nominally < 311 fraction would be finer) was, for one sample, washed with water in a high-speed centrifuge, and slowly dried over silica gel (to avoid rapid loss of water in the electron microscope); the other samDle was reacted with a 10-4 M Na2S.9H20 solution for 30 minutes at a pH of 10.0, and then water washed and dried as for the first sample. The micrographs obtained for the unreacted chrysocolla are difficult to interpret: the resolution and contrast obtained were inadequate for discerning any clear detail of the structure. However, the increased contrast obtained by sulphiding the chrysocolla does permit some structural detail to be seen especially on the thinner areas: the enlarged micrographs of suiphided chrysoc-411a are shown in Figure 8. Similar structural detail is just discernible on the micrographs of unreacted chrysocolla, but these micrographs are not suitable for reproduction.
The x-ray diffraction data conform quite closely with those of other workers, - especially with regard to diffuseness and relative prominence of the 1.5 and 2.9 . d spacings. The broad nature of all the discernible lines, except the 1.47 2 line, suggests a small crystallite size. 43 Using the Scherrer relationship (e.g. Azaroff and Buerger ) of L = >%r (where L is the crystallite size, the radiation wavelength, r the camera radius, and t the half-width of broadened lines - estimated with a millimetre scale owing to the relatively poor quality of the films obtained), it
8: ELF:OTT:01: OF P!- KOI;;EZI
plfAes show microgrEiphs of -2(Allpies successively tilted in 2 — 2i° itees •
0 1 cm = 500 A 38.
was estimated that the crystallite size lies between 100 and 150 R. An interpretation of the pattern in terms of a layer structure is possible, and lattice planes have been assigned to particular spacings in Table 7 p by Dr. I.S. Kerr of the Chemistry Department. The tailing-off of some reflections (notably the 2.60 R spacing) suggests disorder in the stacking of the layers.
The hysteresis loop of the chrysocolla water sorption isotherm (Figure 4) is characterised by a sloping absorption branch, and a steep desorption branch. De Boer44 has analysed such an isotherm and concluded that it is consistent with one of the following structuresC(i) a structure composed of aggregated polyhedra, with interstitial pores, (ii) a structure containing tubular or 'ink-bottle' capillaries with short necks and wide, sloping bodies; (iii) a structure containing tubular capillaries of varying cross-section. These structures are sketched in Figure 9. It can be inferred from the almost horizontal asymptote at relative humidity = 100% in Figure 4 that the pore size distribution is small, because there is no sudden filling of larger pores at high relative pressures, which would produce an almost vertical asymptote. This is suggestive of a structure similar to that shown in Figure 9a. The infra-red spectrum (Figure 6) has, as its principal feature, a strong broad absorption peak at 9.611 (1040 cm-1) which is found in the spectrum of non- crystalline or poorly crystallised silicates. In addition, the spectrum indicates the presence of two different types of constitutional water. Hydroxyl groups are indicated by the twin peaks near 311 (3340 cm 1). The 2.811 (3580 cm-1) peak is considered by Buswell and Dudenbostel45 to 39.
FIGURE 9: EXA7,1PLES OF RELEVANT PORE ST5 77:PF`7 p o5) a: interstitial pores arising from agglomeration of polyhedra.
b: tubular or 'ink bottle' capillaries with short necks and wide) sloping bodies.
c: t'ibular capillaries with wide parts of variolls widths. 4o. represent unbonded OH (such as in water), whilst they ascribe the 3.011 (3340 cm-1) peak to hydrogen-bonded OH. Sorbed water is also shown by the 6.11i (1640 cm-1) absorption peak according to Hunt'et a146. The infra-red spectrum of chrysocolla obtained by Sun37 does not show the twin hydroxyl peaks: this fact may arise from poor instrument resolution, since in other respects his spectrum is in close agreement with that obtained here. There are three distinct features of the DTA trace shown in Figure 7: two endothermic reactions at 80°C and 1025°C and an exothermic reaction at 630°C. Less clearly visible (owing to the non-linearity of the base- line) is a possible endothermic reaction near 400°C. The reaction shown by many previous differential thermal analyses at 850-1000°C was not apparent; in other respects, the differential thermal analysis performed here is in agreement with previous work with regard to temperature of reaction. Interpreting the data in conjunction with the thermogravimetric analysis of Figure 2, it is clear that the reaction at 80°C is a result of vaporisation of sorbed water; that the 1025°C endotherm is caused by the decomposition of cupric to cuprous oxide; and that the endotherm at 400°C is a result of the vaporisation of hydrogen-bonded OH as water. The 630°C exothermic reaction is most likely to be associated with decomposition of the copper silicate to form a mixture of CuO and Si02: it is at this point that the mineralassumes a black appearance. The peak areas of the differential thermal analysis trace approximately correspond to the heats of reaction involved. The latent heat of vaporisation of 1.35 moles of sorbed water is 13.38 Kcal(based on the normal latent 41.
heat at 10000). The heat of decomposition of 0.85 moles CuO to cuprous oxide and oxygen is 12.70 Kcal. These values are in the ratio of 1.00: 0.95, and the corresponding areas 1.00: 1.20. Accurate measurement of the areas is impossible owing to the non-linearity of the base-line. No data is available for the heat of decomposition of copper silicate or for the vaporisation c' 0.32 moles of non-se-bable water: the heats involved
appear to be some 3 Kcal/mole. (SiO2 basis) and 25 Kcal/ mole respectively.
Figure 8, showing the electron micrographs, is visible evidence for the porous structure of sulphided chrysocolla, the darker areas indicating the CuS deposits. ''his porous structure has been previously suggested by the evolution of air from unhydrated samples placed into water and the water sorption isotherms. Unsulphided r.!lrysocolla also showed a very similar structure; in ac latter case, darker areas, where these were discernible, were caused by the solid copper silicate matrix. Figure 8 shows the features are essentially independent of tilt, indicating that the pores resemble capillaries rather than surface configurations or regular interlamellar voids. It can be seen that the pore structure is irregular, although occasional bundles of nearly parallel pores can be ::_st.i.:nr7uished. The pores are seen to be of approximate uniform diameter (approximately 35 10, and variously orientated in the plane of the photographs, Their length is not readily determinable, although some of the bundles mentioned above have been seen to be 150 J1. in straight length.
The mean diameter of the pores can be arrived at somewhat differently. If the 0.32 mole of non-sorbable water in chrysocolla is assumed to be a hydration of the 42. silica framework, then it may be regarded, as an approximation, to be an adsorbed monolayer formation over the whole internal surface area (and the sorbable water as multilayers), then the internal surface area can be calculated as:
S = AWN /kJirgensons and Straumanis4 7 ) where S = surface area, A = area occupied by 1 molecule of adsorbate„ W = weight of adsorbate, N = Avogardro's number and M = molecular weight of adsorbate. Then S 155 m2/g. Knowing this, and the pore volume (from the earlier specific gravity calculations) it is possible to express the mean width of the pores (which are assumed 2 vp cylindrical) as T = (de Boer 44 \) . Where r is the mean radius, v the pore volume, and S the total surface area. P The 1. = 20 A, and a = 40 X. If the crystallites are taken as being cubes of 150 2 side, then the a on the basis of 39% voids would be approximately 30 2. e. Chrysocolla as a montmorillonite member. All the evidence obtained here either refutes, or does not specifically support the montmorillonite analogy for chrysocolla of Chukhrov and Anosov35'36, and Yakhontova33. This supports the recent contention of Sun37. Montmorillonite minerals all show a strong x-ray diffraction line between 9 and 16 X. Such a high reflection was not apparent in this chrysocolla sample (or most of previous workers' samples) even when a Guinier camera was used or when tetraethylammonium chloride was sorbed into the mineral. Furthermore, only the 1.47 and 4.4 .R lines of this 11 chrysocolla agree with lines reported by the ASTM index for montmorillonites. 43.
According to Chukhrov and Anosov's35 data it can be calculated that a hydrous copper silicate analogous to montmorillonite would have a composition equivalent to 0.76 CuO . 1.00 Si02 . 0.25 H20. This is not in good concordance with the analysis shown in Table 5. The most striking and unusual property of chrysocolla is its ability to absorb water. Two types of material show water sorption phenomena similar to that described here. First, certain minerals, such as montmorillonite and the zeolite, faujasite, absorb water in the crystal lattice. In the case of montmorillonite,water enters the crystal lattice between the silicate sheets and the lattice expands in one direction. The sorption of water by sodium montmorillonite takes place in three distinguishable 48% stages (Mering ). The first stage, from 0 to 90% relative humidity, corresponds to absorption of water vapour (up to 40% by weight) without visible swelling of the mineral, but the spacing between the silicate sheets increases from 9.6 R to 16.2 R. The second stage takes place between 90 and 99% relative humidity; up to 30% swelling of a mass of the mineral is observed and the silicate sheets are displaced to almost 20 R. The third stage occurs when the mineral is immersed in water; swelling increases until the mineral apparently occupies approximately 20 times its original volume. Calcium montmorillonite, however, is different; water is absorbed, as in the first stage above, but no visible swelling occurs, even when it is immersed in water. For several montmorillonites, of different" composition, the sorption increases rapidly with relative humidity above 95% (Mooney et al49)., Also, sodium and other alkali metal montinorillonites show an hysteresis of 50 sorption over the whole range of relative humidity (Brooks , 51N 49 Barrer and Reay ). However, Mooney et al report that 44. calcium montmorillonite shows hysteresis only at high relative humidit::_--;.
With faujasite,the water fills existing cavities in the lattice, thus no alteration of the lattice dimensions occurs. Also, in various cationic forms, faujasite absorbs water without hysteresis at any value of relative humidity 52\ (3arrer and Bratt ). The isotherms increase sharply at low relative humidity, become almost horizontal between 15% and 85% relative humidity and increase again above 85% relative humidity.
The second tyioe of material which absorbs water over a range of relative humidities is a xerogel, i.e. a dried hydrous precipitate such as silica gel. Substances of this type absorb 20-1,0% of water by weight between 0 and 99% ...relative humidity with only slight changes (2 to 3%) in vclue (van Bemmelen53) At low relative humidities, water is absorbed on external and internal surfaces and at relat. ive humidities, water is condensed in sub- 54\ microscot,ic pores in the particles (Kruyt ) . The pores are not df_rectly related to the crystal structure, as in the case of faujasite, but are of variable size and cl.c-1-=,_T:;utio7.1. Hysteresis is not observed below 54 ao-croxim,itely 30% relative humidity according to Kruyt n
In terms of porosity, the differences between the three structures are less marked than might be thought. .As mcntorillonites expand their lattice almost two-fold, when absorbing water, without visible swelling, this expansion mast be taken. up by relatively large spaces within a particle. Thus montmorillonites, zeolites and ::erogels all contain empty spaces when the loosely bound water has been removed. The first contains large and probably irregular spaces; the second contains small and regular spaces; the third contains irregular spaces of variable and intermediate size.
The sorption isotherms of this sample of chrysocolla shows hysteresis only at relative humidities greater than 30%, thus it resembles calcium montmorillonite and silica gel in this respect. k sample, previously dried, swelled by only 2 to 3% when immersed in water; thus it did not resemble sodium montinorillonite.
One observation, which assists in distinguishing between a Icalcium montmorillonitei structure and a xerogel structure, is the absorption of a non-polar liquid consisting of relatively large molecules. Degassed montmorillonites absorb such liquids (or their vapours) to a lesser extent than polar liquids such as water
(Barrer and KcLeod55) Xerogels (according to Kruyt54) and faujasite (according to Barrer56 ) absorb comparable volumes of non-polar and polar liquids. A sample of chrysocolla was degassed for several hours at approximately 7000 in a vacuum and the total uptake of benzene vapour, at a pressure almost equal to the vapour pressure of pure benzene, was determined in the same way as the water sorption. The volume of benzene sorbed was slightly greater than the volume of water sorbed, thus indicating a xerogel or zeolite structure.
The sorption properties of chrysocolla can be compared with those of opal or aged synthetic silica gels. Van Bemmelen 3 observed that the pores in synthetic silica gel increased in size over a period of a few years. Bachmann57 showed that, after ageing, synthetic silica gel resembled hydrophane (a form of opal) in its water sorption phenomena, i.e. appreciable sorption occurred only at relative humidity >80% and with no hysteresis. The conclusion is that the pores in these two materials are relatively large. If 46. chrysocolla does have a gel structure then it is apparent that the pores are prevented from increasing in size on standing.
In support of a xerogel model is the similarity in globular microstructure (Figure 9a) between chrysocolla and xerogels such as silica gel, and the indefinite composition of samples of chrysocolla which contain only Cu, Si3O and H in any significant proportion.
It is concluded therefore that the data obtained in this investigation do not clearly distinguish between the montmorillonite, zeolite and xerogel structures for chrysocolla. Some of the data are not inconsistent with any of the structures, but the remaining data are more consistent with a xerogel structure than either of the other two structures. This conclusion can only be tentative because there are no data for other minerals of established structure and similar composition to chrysocolla. There is no doubt that chrysocolla is porous since it absorbs large volumes of liquid, there is only doubt regarding the nature of the pores and the mechanism of absorption of water.
Infra-red spectra for montmorillonites, - such as is 4 presented by Hunt et al '6, are similar to the chrysocolla spectrum. The broad, diffuse nature of the silicate absorption peak in both spectra prevents any rigorous comparison although perhaps the peak is broad because the Si-0 structure is continuously varying throughout the mineral with both chrysocolla and the montmorillonite. A noticeable difference between the differential thermal analysis of this chrysocolla and montmorillonite is the high temperature endothermic reactions which occur only in various samples of the latter above 700°C owing to dehydration. 47.
Since the preparation of this thesis a contribution to the mineralogy of chrysocolla has been published:
MARTINEZ, E.: Chrysocolla studied by differential thermal gravimetric analysis and infra-red spectrophotometry: Trans. AIME: 1963; 226, p.424.
The data in this paper in no way alter the conclusions arrived at on the Kolwezi sample. 48,
6. REVIEW OF FLOTATION OF COPPER ORES. a. Processes.
Copper ores which contain copper sulphides, are almost universally processed, or partly processed, in mineral beneficiation operations by froth flotation, generally using collectors of the xanthate or dithiophosphate type under alkaline conditions. Native copper and cuprite are frequently concentrated together with sulphides in such operations, but other so-called 'oxide' copper minerals, hereafter more correctly referre6 to as secondary minerals (including malachite and chrysocolia, where these cccur),are generally lost to the tailings.
Malachite and azurite, basic copper carbonates, are, however, industrially recovered by two other processes of flotation. One is the direct flotation with crude carboxylic acids (such as palm kernel oil), and the other is a two-stage process in which the copper carbonate is first activated with a soluble sulphide, and the sulphided mineral subsequently collected as though it were a natural sulphide mineral, with a long-chained xanthate homologue. In some cases, the activation and collection stages of this process are carried out simultaneously, Both processes are normally only applied after copper sulphides have been first removed by direct xanthate flotation, for a number of practical and economic reasons. For example, natural sulphide minerals are easily depressed by the presence of even small quantities of hydrosulphide ion in solution.
Union Miniere du Haut Katangats Kolwezi mill is an example of malachite being recovered with some success by both the fatty acid and the sulphidation-xanthate technique (Anon58). Here, malachite is the principal secondary mineral, and is the principal loss of copper in both the sulphide and secondary mineral flotation circuits. Owing to the poorer recovery of chrysocolla, however, the relatively minor amount of this mineral in the ore contributes significantly to the copper lost in the tailings. A similar situation exists in the Northern Rhodesian operations of Nchanga Consolidated Copper Mines Limited. In this instance, O'Meara59 has shown that whilst 66% of the copper loss was caused by fine malachite, a further 12% was attributable to both fine and coarse chrysocolla.
Other minerals, including chrysocolla, which have poor amenability to flotation, are directly leached in sulphuric acid solutions. The copper is recovered from the pregnant leach liquor by either iron cementation or an electrolytic operation. Malachite flotation concentrates are also generally leached, since they are usually of too low a copper content for direct matte smelting operations. Examples of direct leaching operations, especially of chrysocolla, include the Chuquicamata property of the Chile Exploration Company, and the Inspiration mill in Arizona.
Ores which are neither floatable nor leachable (the latter as a result of insoluble complex copper minerals, or an acid-consuming gangue) may be treated by direct pyrometallurgy as in the segregation process. In this process the ore is heated to approximately 750°C with small quantities of a reducing agent and sodium chloride in a neutral atmosphere. Metallic copper is produced and recovered by grinding and flotation. b. General theory of flotation.
Flotation of a particle is possible when a finite angle of contact, (), is obtained between the surface of the particle and an air bubble in solution. Because the 50. work of adhesion between the particle and the bubble, W, is related to the contact angle by the modified Young (1 --cos 8), (where S is the surface cquation, W = S wa wa tension of the water-air interface), it follows that as 8 increases, the greater is the adhesive force between bubble and particle, and thus the more hydrophobic, and floatable, is the particle. In general, clean mineral particles are hydrophilic (i.e. 8 =0) owing to the atomic or molecular attraction of water dipoles by the mineral surface. With such minerals, surface hydrophobicity (i.e. 0>0) can only be achieved by forming; or adsorbing a new material at the particle-solution interface, such that the attraction of the new material at the surface for water molecules, is less than the attraction of water molecules for themselves. 6o Although Derjaguin and Shukakidse have calculated, for a naturally hydrophobic mineral, that mineral-air bubble attachment will just occur when: dC,I 2/A < 3 (d = thickness of ionic atmosphere, E = dielectric . permeability of water, j = zeta potential , A = difference in water-to-water, water-to-mineral Hamaker coefficients) rather more than just a finite contact angle is required in industrial flotation. The mineral-air bubble bond must be sufficiently strong to withstand the shearing forces present in practical flotation machines, and the new adsorbed species, where this is involved, must be fairly strongly adherent to the mineral. In this connexion it should be noted that the seemingly universal view, that flotation is the levitation of a hydrophobic particle to an air-water interface, where it can be stabilised in a froth column, may be invalid in industrial flotation involving large shearing forces, high collision rates and small contact periods. The possibility cannot be ruled out that, under such conditions, flotation is, in fact, the result of the random projection of all particles into the froth layer, where those particles which are hydro- phobic to the required degree can remain. This latter system is considered the only way in which minerals having a very low contact angle can be industrially floated, c. Theory of xanthate collection.
Collectors of the xanthate class are constituted as follows: •- + RU M S (R is an alkyl group, and M an alkali metal). These reagents are: soluble in water, where they exist in the ionised form shown: the xanthate ion is heteropolar, lathough the reaction of xanthate has been observed with secondary minerals, such as malachite, xanthate reactions have only been closely studied on sulphides. Two mechanisms have been proposed. 61 62 Cook and Nixon and Hagihara et a1 believe that a molecular species, derived from the hydrolysis of the xanthate ion, is adsorbed at sulphide surfaces which do 62 not have to be oxidised- Hagihara et al working with galena, PbS, and employing electron diffraction techniques, proposed that the adsorption bonding was between galena sulphur and the doubly-bonded sulphur of the xanthate molecular species, although adsorption occurred only near the lead sites on the surface
The second theory is better substantiated and involves the formation at the surface of heavy metal xanthate salts of very low solubility. Because metal sulphides are even more insoluble than the corresponding xanthates, it is now considered that an initial oxidation occurs at the surface, and that the surface hydrophobic compound is formed by an exchange of X (xanthate ion) for OH. With copper, two reactions occur depending upon the valence state of the copper, viz: 2+ 2 Cu + X + 4X--1. 2 CuXsolid 2 liquid + or 2 Cu + 2X 2 CuX solid cuprous xanthate and dixanthogen are formed with the cupric ion, and cuprous xanthate is formed with the cuprous ion. Both reaction products are hydrophobic, and their co-adsorption has been suggested by infra-red spectroscopy.
With sulphide surfaces, a tenacious film of solid cuprous xanthate (and possibly dixanthogen) is formed, and efficient industrial flotation is possible when only 15% of an equivalent mono-molecular layer films the external mineral surface. The film is not, however, regularly precipitated on the surface, but is markedly uneven, and forms multilayers in places.
From the scant data available for malachite, it is clear that a much faster reaction occurs to a greater extent. This may be caused by the relatively high solubility of the mineral compared with copper xanthates. The resultant thick, often visible film has poor tenacity, and it is not clear from the literature whether flotation would be possible at the low film coverages found successful with the sulphide minerals.
Many investigations have helped formulate and clarify the heavy metal xanthate theory; they include 63 67 Dewitt et al , Gaudin et al64'65'66, Leja et al , 68 69 70 71 Plaksin et al ' , Rao and Patel , Shorsher , Taggart J3.
72 73 74 et al ' and Wark . d. Theory of sulphidation-xanthate collection.
Despite the fact that sulphidation is a well- established plant procedure, especially for some secondary lead and copper minerals, little fundamental investigation has been undertaken on the underlying mechanism of the process, and the basic principles are not understood.
In view of the current theory for the interaction of natural sulphide minerals, and xanthates, a possible explanation might be considered to be the formation of a thin film of heavy metal sulphide at the surface of the secondary mineral followed by the usual xanthate reaction at this new sulphide surface. There is, however, some evidence against this. 71 For example, Rehbinder et al75 and Shorsher have measured the hydrophobicity of suiphided minerals that were not treated with xanthate. They were found to be considerably more hydrophobic than clean, natural sulphide minerals. On the other hand, only longer- chained xanthate homologues appear to be capable of collecting suiphided minerals in plant practice, whereas, short-•chained homologues are quite adequate for most natural sulphide minerals, It is possible that there is no real justification for not using short-chained xanthates on sulphided minerals. e. Previous chrysoco.11a investigations.
Investigational work on chrysocolla flotation has, in many cases, only involved measuring the flotation effects produced by various experimental conditions. The extent, location and rate of formation of any hydrophobic species formed, has been largely neglected. Unfortunately, insufficient data has generally been presented for any rigorous re-valuation of results to be made. Some previous work has involved the evaluation of experimental conditions under very quiescent conditions of flotation such as can be achieved in 'bubble pick-up' testing, or by using a simple Hallimond tube; other work has approached more practical flotation conditions, such as can be achieved in a standard 500g flotation machine. These two distinct methods of assessment are hereafter referred to as quiescent and turbulent flotation respectively.
Only Kovacs4 appears to have considered the possibility of direct xanthate flotation. He - found quiescent flotation possible, using a sample of chrysocolla from the Inspiration property, after heating the mineral to 500°C in the presence of water vapour for 40 minutes; flotation took place in 100 mgpl potassium amyl xanthate (hereafter KAmX) solution at pH 9.5. The author suggested that it was necessary to have some amorphous CuO present in chrysocolla in order to offer high energy adsorption sites for a xanthate reaction, since they were unable to obtain flotation without their initial steaming treatment. A more plausible explanation of their success would seem to lie in the reduction in internal surface area of xerogels that occurs on heating, especially in the presence of water vapour. This phenomenon has been investigated by a number of workers, including Adams and Voge76, working with silica-alumina catalysts. For a fixed xanthate conditioning time a greater external surface reaction will result from such an internal surface area reduction and the chrysocolla might be expected to be more hydrophobic. The sulphidation-xanthate collection problem has recei7-ed t17::- attention of a number of investigators, The first was Dean5 who claimed turbulent flotation with Na S Jr H 2 2S activation followed by XLmX. His p:2o'Jedure was improved by lowering the pH to k.0 and by using ammonium sulphate to control pH. Despite this success, it was 21 years before a more detailed study was made by Mitrofanov et al77'78 Kushnikova et a179 and Strigin 80 Kushnikova on sulphidation: their work, under quiescent conditions dealt largely with sulphidation kinetics, and how sulphidation rate could be increased by the use of either ammonium or aluminium sulphate. In addition, sulphidation of chrysocolla was shown to proceed less rapidly than that of a simiarly-sized malachite sample, and the film formed to be poorly adherent. !n examination of :anthate reactions on sulphided chrysocolla showed that these increased with pH.
The increased reaction rate observed when using ammonium salts in these investigations may be caused by the formation of the soluble cuprammonium comple:: which results in more copper being available in solution for reaction with the soluble sulphide. The advantages of aluminium sulphate may be attributable to the coagulative effect of the multivalent ions upon the sulphide film, which, because of the experimental technique, could appear as faster sulphidation. 1 Arbiter has suggested a sulphidation process involv:.ng a preliminary activation with copper sulphate solutions. He claims good recoveries on deslimed ores with very short conditioning times, and using sodium 2-mercaptobenzothiazole in a turbulent process. No pH is given for the process, and no explanation is proposed. Such an explanation lies, in all probability, in the 56. increased rate of sulphidation and collector reaction owing to increased copper being available in solution near the mineral for reaction, following the copper sulphate activation. 81 Bayula has described the use of sodium oleate in improving turbulent flotation of chrysocolla by sulphidation- xanthate collection techniques. He proposes that sodium oleate reduces the solubility of the surface-active substances and thereby prevents depression of chrysocolla by Na2S; but this explanation has little credibility because cupric oleate is considerably more soluble than either copper xanthates or copper sulphides.
A further example of the applicability of the sulphidation-xanthate collection process on a laboratory scale is to be found in Jaekelts3 work on Inspiration chrysocolla. Using turbulent flotation assessment, he experienced no difficulty in beneficiating synthetic ores containing only Inspiration chrysocolla .and quartz, or natural ores containing numerous minerals including malachite, azurite and chrysocolla. He sulphided and conditioned with collectors (dithiophosphate-type and the American Cyanamid Company's 404 and 425 reagents) at neutral to alkaline pH's. 2 Bowdish and Chen however, failed to find Jaekel's5 method applicable to another chrysocolla ore sample, and proceeded to study the range of sulphide concentrations, and pH's, which would produce a surface capable of xanthate collection. They appraised various experimental conditions by measuring the contact angles of captive air bubbles on submerged polished sections of chrysocolla, and concluded that an incipient leaching of copper was necessary for the formation of an adherent sulphide film. 57.
Optimum conditions for producing a surface for amyl xanthate collection were shown to be 10 minutes at pH 4 in a 400 mgpl Na2S.9H20 solution; these conditions could be improved (insofar as potential plant applicability was concerned) to pH 7 and 100 to 1200 mgpl Na2S.9H20 if the chrysocolla were treated with a titanium tetrachloride 2 solution. Bowdish and Chen were unable to explain the action of the TiC14. Because the xanthate-collection stage of Bowdish 2 and Chen's experiments took place in solutions essentially free of sulphide, hydrosulphide and hydrogen sulphide, it is unlikely that the low contact angles, found outside their range of optimum conditions, were caused by HS depression as they suggest. It is more plausible that low sodium sulphide concentrations, or high pH's resulted in 2+ too slow a reaction of HS and Cu to produce an adequate film in the conditioning time of their tests; and that high sodium sulphide concentrations, or low pH's, resulted in either a reaction which occurred substantially in the pores or in a non-adherent sulphide film which was. washed off prior to xanthate collection. From their results, it is difficult to re-assess the effect. It is possible that the mechanism of the TiC14 TiO2 colloid, formed by the hydrolysis of a TiC14 solution, is absorbed in the chrysocolla, resulting in more external may also have film being formed in a fixed time. The TiO2 a coagulative effect on the CuS film, resulting in a more adherent coating. A more rigorous explanation is not possible from the data. 2 Bowdish and Chen also reported on turbulent flotation testwork. Good recoveries and reasonable grades were obtainable with 3 lb/ton TiC14, 8 lb/ton Na2S.9H20, 58. 0.4 lb/ton KAmX and a conditioning pH of 3.5 to 5.0.
This account of previous sulphidation investigations 82 is concluded with the work of Bautista and Sollenberger . They found that a direct reaction between chrysocolla and sulphur took place over a period of 1-2 hours at 150°C - 2250C.
Some attention has also been given to the use of collectors other than xanthates, dithiophosphates or related compounds. Such collectors, generally complexing agents for copper, have been investigated by Dean5, 84 6, Gutsalyuk et a183, Livingood , Ludt and Dewitt Mitzmager 85 86 and Gailis and Tyurenkova ; varying degrees of success were claimed on natural ores.
Amongst those laboratory flotation processes for which success is claimed, it must be realised that the conditions in the turbulent procedures of Dean5 and Jaekel3 particularly, are very close to those encountered in plants where the copper ores (including those containing chrysocolla) are first sulphided and then floated with xanthate. In such operations chrysocolla is not, however, satisfactorily recovered.
Chrysocolla is not industrially floated in xanthate or soluble sulphide-xanthate systems because, clearly, the external surface of chrysocolla is insufficiently hydrophobic. Such a state of affairs may be brought about by a number of occurrences, including:
A: no hydrophobic species ever having adsorbed, or formed, at the external surface,
B: insufficient hydrophobic species having adsorbed, or formed, at theexternal surface. 59.
C: removal from the external surface of a hydrophobic species which had adsorbed, or formed, which, in turn, are a result of one or more of the following:
A': no reaction between chrysocolla and xanthate; or reaction yielding only a non-hydrophobic material; reaction taking place solely in the pores of the mineral; or a reaction taking place solely in the bulk solution surrounding the mineral.
B': insufficient time being allowed for reaction; insufficient xanthate available for reaction; a partial reaction in the pores of the mineral; a partial reaction in the bulk solution surrounding the mineral; another mineral or ion consuming xanthate at a faster rate; or the decomposition of xanthate.
C': scouring action on adsorbed or formed hydrophobic species; like electrical charges on the mineral and the hydrophobic species; or scouring action on mineral.
The role of soluble sulphide in this system, or other systems (see section 6d) is not clear from the previous work. It is unlikely, however, to affect directly the hydrophobic properties of the chrysocollain•axanthate system. 6o, 7, CHRYSOC.OLLA-XANTHATE, CHRYSOCOLLA-SOLUBLE SULPHIDE SYSTEMS. a. Investigational method. Many investigators of fundamental flotation problems have determined mineral wettability (by measuring contact angles) or the electrical characteristics of a mineral surface (by measuring electrokinetic potentials). Such phenomena are difficult to measure and may be related to real flotation systems in only a complex manner. It is here considered more important to relate the amount and disposition of the hydrophobic species to a direct, flotation method which can be interpreted in terms of actual flotation operations. Unfortunately, the ideal of turbulent flotation in a standard laboratory flotation machine is unsuitable for fundamental investigations, owing to the complexity of such a system; even working with sized, synthetic chrysocolla-quartz mixtures presents difficAlt problems of control, manipulation and analysis.
The method employed for proving, or disproving, the factors leading to poor chrysocolla flotation involved, in the main, two stages. The first stage involved the conditioning of a sized mineral sample under conditions of mild agitation; thus reducing thescouring action. In the second, the conditioned mineral was assessed for floatability in an apparatus giving quiescent flotation conditions. Additional, confirmatory, testwork was under- taken in more complex systems, which were more akin to industrial practice. The mechanism of the chrysocolla- xanthate, chrysocolla-soluble sulphide reactions was elucidated in various ways; these will be dealt with in detail as they arise.
Of the two methods available for assessing the amount of sorbed, or formed, hydrophobic species, only the chemical methods have been used here. Methods using 61.
radioactive isotopes are most suitable for very minor quantities of adsorbate, but nevertheless are complicated to operate, and interpret, and usually require lengthy calibration.
The chrysocolla-xanthate, chrysocolla-soluble sulphide systems can be adequately quantified by merely measuring the depletion of adsorbate from solution, and subsequently determining the disposition of any reaction products. The principal advantage of such a method is its simplicity in operation, and in the interpretation of results.
b. Preparation of samples for testwork-
The sample used for the investigation of flotation and related phenomena was identical to that prepared for the mineralogical investigation (see section 5a). Again, only the -72 +100 meshBSsize fraction was used. Although the mineral was initially prepared under water, in order to avoid contamination - it was soon apparent that such a procedure was unnecessary. Any contamination picked- up during the dry preparation procedure of section 5a was insufficient to produce any apparent hydrophobic effect as assessed by flotation.
A few tests were performed on an ore sample obtained from Union Miniere du Haut Katanga,.$A, in order to confirm some of the findings of the pure mineral system. This sample was not intended to be representative of that Company's milling operations but was sorted such that the chrysocolla was the major copper mineral and the gangue minerals fairly typical. The sample contained at least two visibly different types of chrysocolla (one of which was that described in section 5) as well as malachite. Prior to use, the sample was crushed, sized to pass a 62.
10 mesh BS screen, and stored in a closed container. c. Nature of systems.
Preliminary testwork showed that when chrysocolla was brought into contact with a solution of KEtX, a reaction occurred. The turquoise-blue mineral gradually assumed a green appearance, and sometimes turned bright yellow:in addition the solution changed from clear to noticeably turbid. The extent of the reaction, and its manifestations, depended very much upon the experimental conditions such as pH, xanthate concentration, degree of agitation and time. As the reaction proceeded, the flotation behaviour in quiescent systems improved.
The new yellow phase, which accounts for the colour in reacting chrysocolla, has been shown by infra-red spectroscopy to contain cuprous ethyl xanthate. The spectrum, obtained by Dr. E.S. Waight of the Chemistry Department, from a Nujol mull of reacted chrysocolla is shown in Figure 10. Tills, when compared with the chrysocolla spectrum of Figure 6 clearly indicates the 67 cuprous ethyl xanthate features described by Leja et al , despite the large background caused by thechrysocolla silicate absorption. (No data is available for the infra- red spectrum of cupric ethyl xanthate.) Unfortunately, it is not possible to detect ethyl dixanthogen against this background, although it must have been present either with the reacted chrysocolla, or dispersed into the reaction solution; for the reduction of cupric to cuprous ion necessitated its formation. However, when the reacted chrysocolla was agitated with iso-octane, ethyl dixanthogen was confirmed as a reaction product by ultraviolet spectroscopy. (Ethyl dixanthogen dissolved in iso-octane shows a definite, if broad absorption band at 2400 X). 63.
FIGURE 10: INFRARED ABSORPTION SPECTRUM OF CHRYSOCOLLA REACTED WITH POTASSIUM ETHYL XANTHATE
absorption density: arbitrary scale 8.
2 4 6 8 10 12 14
wavelength: microns //71 5000 3000 2000 1000 4000 -1 N wave number: cm 64.
Chrysocolla also produced a visible reaction with aqueous solutions of sodium sulphide or with dry hydrogen sulphide gas. Here again a distinguishable colour change was produced (ranging from green to black) and, in aqueous systems, a distinct turbidity of the solution was found. The black phase can only be cupric sulphide since there will be no tendency for a reduction of the cupric ion under these conditions.
When sulphided chrysocolla was contacted with a xanthate solution, no further visible colour change was apparent, but quiescent flotation was sometimes possible.
Factors affecting the xanthate and sulphide reactions were examined more closely. The processes of xanthate reaction and sulphidation are obviously analogous to a great extent, with a result that only one system need be studied in detail. The chrysocolla-xanthate reaction was chosen for closest study because it was experimentally easier to operate than was the sulphide system; the latter presented problems caused by volatilisation of hydrogen sulphide. d. Factors affecting chrysocolla-xanthate reactions.
(i) the reaction vessel. 87 Although Subba Rau concluded that adsorption experiments were more satisfactorily conducted by continuous flow methods, than by a batch procedure, it was here decided to employ the more convenient batch procedure.
Despite the difficulties of investigating the initial stages of adsorption processes, and the concentration change during reaction, the small Quantity of sorted mineral available for experiments in this investigation 65.
precluded the use of a flow method. In a flow method, excessively dilute solutions, or very small volumes, must be employed if accuracy is to be obtained. Further- more, flow methods involve a time lag between surface reaction and withdrawal of analytical sample; and experimental technique complicated by a pH change which tended to occur in the experiments.
The reaction vessel used is shown in Figure 11. It consisted of a flat-bottomed glass basin, 8 cm diameter, 5 cm height, sealed with a push-fit polyethylene top. A polyethylene stirrer, actuated by a 1/300 hp motor was used for agitation and entered the top of the cell via a mercury gas seal. A glass electrode-calomel electrode assembl;r for pH measurement also projected through the top; gas-tight seals were effected by a rubber sleeve on the glass electrode, and a simple push-fit in the polyethylene top for the calomel electrode. Three other apertures were provided in the cell top; one for adding acid or alkali with a micrometer syringe pipette, another for removing analytical samples or adding mineral; and a third for bubbling oxygen-free nitrogen into the cell, when this was required. When not in use during experiments the apertures were sealed with rubberstoppers.
(ii) experimental procedure.
Chrysocolla-xanthate reactions were conducted in solutions prepared from de-oxygenated distilled water (obtained by purging singly-distilled water with oxygen- free nitrogen for at least 30 minutes). Any changes in pH which tended to occur during filming were rectified by -1 the addition of 10 H H2SO4 or KOH solutions delivered from the micrometer syringe pipette.
The extent of reaction was ascertained by measuring 66.
FIGURE 11: THE REACTION VESSEL
polyethylene pH electrode stirrer assembly (calomel electrode shown)
JC mercury coal nitrogen inlet polyethylene top
glass basin
QQbOD the depletion of xanthate from the reaction solution: this necessitated removal, at intervals, of solution samples from the cell. Initially, 100 ml of adsorbate solution were introduced into the cell, and adjusted to the required pH level. When a stable pH had been obtained, a solution sample (1 - 10 ml, generally 2 ml, depending upon xanthate concentration) was pipetted fromthe cell, the stirrer started and 100 mg of fully-hydrated, but dry, chrysocolla sample introduced through a funnel into the system. The apparatus was then sealed until further solution samples were required in order to follow the progress of the reaction. The motor was temporarily halted during pipetting operations so as to prevent mineral particles from being drawn into the pipette.
The temperature during the experiments was not controlled, but invariably fell in the range 20 2°C. Nitrogen was used sparingly during the reaction in order to exclude air from the system; too fast a flow tended to float suitably hydrophobic particles, and would have, in any case, introduced unwanted oxygen and carbon dioxide into the system. (Oxygen-free nitrogen supplied by the British Oxygen Company does in fact contain 10 volumes per million 0,, and 5 volumes per million CO2). Agitation was always maintained at a level just sufficient to keep all the particles slowly rolling over the floor of the cell, without the formation of a vortex. Attrition was thereby kept at a minimum.
(iii) analytical procedure.
Xanthate was estimated directly on solutions, diluted to approximately 0.2 x 10-4 M, by a simple spectrophotometic technique. 68.
88 According to Hagihara , xanthates can be determined spectrophotometrically in the presence of their principal decomposition products and, if necessary, sodium sulphide. This is because of the distinguishing xanthate thiocarbonyl chromophore, C = S, corresponding to an ultraviolet absorption peak of 3500 In dixanthogen there are two thiocarbonyl chromophores and these do not give rise to a 3000 .R absorption peak. 1 cmihsed silica cells were used in a Unlearn SP 500 spectrophotometer, and the calibration 87 for the method is as described by Subba Rau .
The decomposition of xanthates in acid media is well known, and has received considerable attention in recent 90 years from Iwasaki and Cooke89, King and Dublon and Rao and Pate191,92. However, it was still necessary to determine decomposition under the present experimental conditions, if due allowance were to be made in calculating reaction progress. By running blank tests (i.e. no mineral present) in the reaction cell, it was shown (Figure 12) that xanthate decomposition was only significant at pH's below 5. The progress of the reaction is here described in terms of moles CuEtX formed per square metre of external surface. (This sample of chrysocolla has been estimated, on the basis of particle size and shape, to have an 2 external surface area of 200 cm /g). In interpreting such figures, three factors must be borne in mind: the proved porous nature of the mineral and its high total surface for each mole area; the formation of one-half mole EtX2 of CuEtX produced; the turbidity of reaction solutions suggesting some non-adherence of reaction products. 2 1 mole/m is equivalent to 0.02 mole/g -72 100 mesh BS chrysocolla. 69.
-4 FIGURE 12: DECOMPOSITION OF 10 M KEtX SOLUTIONS IN ACID CONDITI"NS
ETU 10-4 doncen- pH 6 tration: mblarity
0.9x10-4
0.8x10
0.7x10-4
pH 4 0.6x10-4
0 5 10 15 20 25 time: minutes 70.
(iv) effect of pH.
Despite precautions against carbon dioxide interference, the pH did not generally remain constant -during chrysocolla-xanthate reactions. In solutions more acid than pH 8, there was a tendency for the pH to rise, which was counteracted by the addition of sulphuric acid. At more alkaline pH's it was necessary to add alkali to the system, in order to counteract a tendency towards neutrality. A thorough investigation of these phenomena has revealed that, whilst the rise in pH in acid solutions is associated with the formation of CuEtX and EtX2: CuSiO (chrysocolla) + 2KEtX + (2-x)H 0 3 2 (Kx H2-x) SiO3 + CuEtX + 1/2 EtX2 + (2-x)KOH and the acid decomposition of xanthate. + (R.0CS2)- H CS2 + H.OH, the fall in pH in alkaline solution is not. In the latter case, no detectable pH change occurs in the experimental system when carbon dioxide from the air, or the nitrogen used for bubbling, is absolutely excluded.
The relationship between conditioning pH and the progress of the chrysocolla-xanthate reaction is illustrated in Figure 13. pH's below 4 were not investigated because of the rapid decomposition of xanthate, and because most of the reaction was seen to occur in the bulk solution and not at the mineral- solution interface. The reaction rate is seen to fall- off markedly as the pH rises from 4 to 7, and thereafter to decrease much more slowly, until at pH 12, there is apparently no reaction taking place.. 71.
FIGURE 13: EFFECT OF pH ON ETHYL XANTHATE REACTION
CuEtX 25 formed: m91es/4 m x10
152(31
10
5 after 30 minut es
after 5 0 minut ss
4 5 6 7 8 9 10 11 12
pH 72.
The external appearance of reacted chrysocolla after quiescent conditionhg, at various pH's is shown in Figure 14. When xanthate decomposition has been taken into account, the acid consumption during the reaction at acid pH's is shown in Table 8. Below pH 6, approximately twice as much acid is consumed as there is CuEtX formed.
Considerably greater quantities of CuEtX were formed in alkaline solutions when ammonium hydroxide (i.e. c. 10-4M at pH 10) is employed to obtain and control the desired pH: this is shown in Figure 15. This result is similar to the observations of Dean5 and Mitrofanov et al77'78 80 Kushnikova et al79 and Strigin and Kushnikova , using ammonium salts in sulphidation. 1-7) kinetics.
Using the same data employed for Figure 13, Figure 16 shows the rate of removal of ethyl xanthate from solution. Each curve corresponds closely to the equation Ft = a t1/2, where .'is a pH-dependent parameter. Values of Ware recorded in Table 9. The calcuated curves at pH's 4, 7 and 11 are shown as broken lines in Figure 16.
(vi) effect of concentration.
In order to evaluate the effect of xanthate concentration, it was necessary to choose conditions of pH and conditioning time such that sufficient reaction occurred to enable reasonably accurate measurements to be made. On the other hand, the concentration of xanthate in solution could not be allowed to fall, as a result of reaction, by too much. Suitable conditions were obtained by carrying out the reaction at pH 5.5 for 15 minutes, and departing from the standard experimental procedure by using a weight of chrysocolla proportional to the 7 •
rIG:JaE 14: IPPEIRANCF OF "X/1 7H/Ti-II(7,771, C=2COLLA IT 71.71fll.;: cAlls LI= C -7-37FS 74.
TABLE 8: CONSUMPTION OF SULPHURIC ACID DURING CuEtX FORMATION.
pH 4 5
2 4 CuEtX : moles/m x 10 21 9 4 a H SO : moles/m2 x 10' 8 2 4 32 17
CuEtX / H S 0.66 0.53 0.50 75.
FIGURE 15: EFFECT OF AMI4ONIA ON ETHYL XANTHATE REACTION AT 0 10
• CuEtX 10 formed: Iles/4 •with ammonia m x10
• 5
• with potassium
hydroxide 0
0 5 10 15 20 25 30 time: minutes 76.
FIGURE 16: KINETICS OF ETHYL XANTHATE REACTION
PH 4 CuEtX 20 ...- - ...- formed: .. mles/4 m x10
15 / / /
/ 10 /
/ v. 5 /
v.
5 .-11w-- 6 r 7 v• I 8 0 0 0 1
5 10 15 20 25 30
time: minutes 77-
1 TABLE 9: VALUES OF 'a' IN THE EXPRESSION r = at2.
pH 4 5 6 7 8 9 10 11 a = moles/m2/ I 1 minute x 104 4.01 1.48 0.78 0.45 0.40 0.32 0.22 10.07 78.
initial concentration of xanthate in solution. By these means, the solution concentration fell by only 20-25% of the initial xanthate concentration, and the highest possible degree of filming occurred, consistent with only a small degree of xanthate decomposition.
The results are plotted in Figure 17) and show a definite minimum (and thus an assumed maximum between 0 and 10-5M EtX, since the curve must pass through the origin), before indicating a steady increase in reaction rate with concentration. It is not possible to investigate the region between 0 and 10-5m Etrby the analytical method employed here, owing to the dilute xanthate solutions involved. (viii) comparison of ethyl and amyl xanthate.
In order to determine the effect of hydrocarbon chain length, five experiments with potassium amyl xanthate and chrysocolla were undertaken. (The amyl xanthate was purified in the same manner as the ethyl homologue) viz; ether precipitation from a filtered alcoholic solution). The comparison between identical tests with ethyl and amyl xanthate is shown in Table 10. In each instance, a greater reaction is seen to occur with amyl xanthate, within the period of the experiments. There is some indication that the initial rate of reaction is faster with amyl xanthate. e. Distribution of chrysocolla-xanthate reaction products.
It was not possible to show quantitatively how much CuEtX formed a film on chrysocolla, or how much was'held within the mineral pores. The scale on which the experiments had to be conducted here precltded the benzene leaching method of Gaudin and Schuhmann66, which,
79.
FIGURE 17: EFFECT CF XANTHATE CONCENTRATION ON ETHYL XANTHATE REACTIn
CuEtX 15 formed: m21es/4 m x10
10
55
0 1 2 3 4 5
ethyl xanthatR concentration: molarity x 10 80.
TABLE 10: COMPARISON OF ETHYL AND AMYL XANTHATE REACTIONS WITH CHRYSOCOLLA.
1 pH 1 8 8 8 10 11 J Reaction time-minutes 1 20 30 30 30 2 4 i Ethyl xanthate: CuEtX moles/m x 10 0.8 2.0 2.7 1.2 0 2 1! Amyl xanthate: CuAmX moles/M x 10 li 2.2 3.1 3.7 1.9 0.2
(Initial xanthate concentration was, in all cases, 10-4 M) 81.
in any case, was found by them not to be fully quantitative.
However, CuEtX can readily be estimated in the reaction by analysing for copper. The method consisted of evaporating 25 ml of the bulk solution to dryness and oxidising the residue to cupric nitrate with 5 ml of concentrated nitric acid, which was also evaporFted. The 2+ cupric nitrate was dissolved in water and Cu determined spectrophotometrically by the sodium diethyldithiocarbamate- carbon tetrachloride method at 4400 X. This method for the analysis of Cu2+ is summarised, and a calibration shown, in Figure 18. Although -72 + 100 mesh BS particles of chrysocolla were eliminated from the 25 ml of bulk solution by filtration through a 150 mesh BS nylon sieve cloth, any attrited chrysocolla that may have formed during the reaction could not be so collected. Such attrited material was assessed by a blank test in water, and the result deducted from xanthate experiment results: the data are shown in Table 11.
Ethyl dixanthogen can (as has been pointed out in section 7d) be determined in iso-octane, and can thus be determined in the bulk reaction solution, and at the mineral surface (EtX 2 within the mineral, cannot thus be determined because of the immiscibility of iso-octane and water, and the absence of any agitation within the pores). The method was as follows: separately, the reaction solution and the reacted chrysocolla were vigorously shaken with 10 ml iso-octane for two minutes in 250 ml volumetric flasks. After shaking, sufficient water was added to the flasks to bring the iso-octane phase into the neck. The iso-octane-dixanthogen solution was then sampled with a pipette for direct ultra-violet 82.
2+ FIGURE 18: COLORIMETRIC DETERMINATION OF Cu WITH SODIUM DIETHYLDITHIOCARBAMATE
optical0.5 density • 440mp • 1 cm cells 0.4 •
0.3
0.2
0.1
0
0 0.1 0.2 0.3 0.4 [Cu2+] in original solution: gpl
'(0.125x1 of original solution, diluted to a suitable volume, extracted into 20 ml CC1 at pH 9) 4 83,
TABLE 11: ESTIMATION OF CuEtX IN RE:XTION SOLUTION.
2+ pH Reaction Total Cu in solution Estimated Time moles/m2 x 104 CuEtX in Minutes Without With Reaction Solution Xanthate Xanthate
5.5 15 7.4 15.0 7.6
8.0 30 3.5 5.8 2.3
10.0 10 0.6 1.4 0.8 J4.
spectrophotometric determination. A calibration of the method is shown in Figure 19.
The results shown in Table 12 are based upon the stoichiometry of the reaction: 2+ Cu + 2EtX CuEtX + 1/2EtX2' Such stoichiometry has been proved by reacting a dilute copper sulphate solution with a 10-4H EtX solution at pH 5.5 for 15 minutes; 15.3 x 10-7 moles CuEtX were formed and 7.68 x 10-7 moles EtXz were detected.
Because some 40% of the stoichiometric quantity of EtX 2 could not be detected in the bulk solution, or at the external surface, it has been assumed that some was occluded within the pore structure. There was the possibility, of course, that the missing EtX2 had never been formed, owing to production of cupric ethyl xanthate; this seems, however, unlikely in view of the stoichiometry of the copper sulphate-ethyl xanthate reaction, Under the quiescent conditions of these distribution experiments, it is seen in Table 12 that the distribution of EtX (and possibly of CuEtX) is 2 independent of pH over the range explored. f. Flotation of xanthate-reacted chrysocolla.
(i) flotation cells
Various designs of milligram-scale batch flotation cells were given trials. Most designs were rejected because an appreciable portion of the mineral was trapped in places where contact with bubbles was impossible. Furthermore, when sufficient agitation was supplied to the mineral bed to reduce the amount of trapped mineral, the chrysocolla was severely attrited, or the air bubbles coalesced during aeration, 85.
FIGURE 19: ABSORPTIOMETRIO DETERMINATION OF EtX 2 IN ISO-OCTANE
optical 1.5 density • 224mil 1 cm cells 1.0
0.5
0
0 4 8 12 16
[EtX ] in iso-octane: gpl x 104 2 86.
TABLE 12: DISTRIBUTION OF CHRYSOCOLLA-ETHYL XANTHATE REACTION PRODUCTS.
Chrysocolla-Xanthate Reaction Product Distribution ( pH Reaction Products Solid-Liquid Bulk Solution Undetected time CuEtX EtX2* Interface:EtX2 EtX2 1 CuEtX EtX2* minutes moles/ moles/ moles / ai e moles/ m2 x104 m 2 x104 moles/ Imoleis/ m2 x104 A' m2 x104 % "m2 x1041 i° m2 xio4 , 1 1 5.5 15 7.6 3.80 0.22 6 2.14 56 14.05 t53 1.44 38 not 8.0 30 2.3 1.15 0.10 9 0.65 0.40 56 idetermined 35 10.0 10 0.8 0.40 0.02 4 0.70 55 [ 0.44 158 0.18 41
* Calculated.
(Distribution of Si0 .xH 0 has not been determined) 2 2 87.
Two designs eventually proved satisfactory - the conventional Hallimond tube and a 30 ml agitated froth flotation cell shown in Figure 20.
The Hallimond tube method of quiescent flotation was considered superior to classical bubble pick-up methods or vacuum flotation because a greater degree of hydrophobicity of particles was required for particle- bubble attachment; chrysocolla will 'pick-up' in bubble pick-up tests, when quite unsuitable for flotation in even a Hallimond tube. The Hallimond tube used had a particularly well-fitted sinter of only 1/8" diameter, and even porosity. Hold-up was eliminated by gently tapping the tube during flotation.
The agitated cell was aerated, via a windbox, through a recessed No.3 porosity sinter and agitation was effected with a magnetic follower. Turbulence was reduced in the froth layer by four small baffles. Despite manipulative difficulties when conducting a flotation test, the cell was useful for proving dynamic flotation in the earlier stages of the experimental programme.
(ii) experimental procedure.
Flotation, in the Hallimond tube, of xanthate conditioned chrysocolla followed this procedure. After conditioning the adsorbate solution was removed from the mineral by decanting through a 150 mesh BS nylon sieve cloth held in a glass funnel. The mineral was washed very gently on the cloth with water and flushed into the Hallimond tube. This flotation in a water medium prevented complications in the interpretation of results arising from conditioning continuing into the flotation period; it may also have removed dixanthogen 88„
fIGURE 20: 30m1 AGITATED FROTH FLOTATION CELL
square section froth removal lip SIDE VIEW
round section
fritted disc, porosity 3
wind—box
drawn actual size
FLAN 89.
from the surface of the mineral, since dixanthogen has been shown toemulsify readily in water.
Exploratory tests showed that, where floatable particles were present; flotation of the 100 mg samples was virtually complete within 30 seconds of the commencement of aeration (air rate = 1 ml/min). Thus, in order to exclude all kinetic effects the flotation period was extended to 2 minutes.
For a number of similar mineral particles conditioned for flotation under the same conditions, it would be expected that all the particles would be rendered floatable, or all the particles would remain non-floatable. With reacted chrysocolla in the Hallimond tube, this was not found to be so; there was, in fact, a gradation between 100% flotation, and 100% non-flotation. The phenomenon must be attributable to one, or a number of the following factors.
Al: that slight variations in contact angle between particles exist, and that these variations fall either side of the minimum contact angle necessary for flotation undernot-truly-static conditions of the Hallimond tube.
B': that the minimum contact angle necessary for flotation increases as flotation progresses, owing to changes in solution compositions,
Ct: that the hydrophobic film is removed from the surface of non-floated particles during flotation or conditioning.
D': that the rate of flotation is, in fact, slow.
Although these factors could mean that a completely hydrophobic sample of chrysocolla is not completely floated 90.
in the Hallimond tube, the gradation between flotation and non-flotation is clearly an asset in assessing various treatments. Flotation in the small froth cell followed the normal procedure of larger-scale batch flotation. A sized chrysocolla and quartz mixture was conditioned and floated in the same solution, and the froth removed continuously, during the flotation period, with a paddle. One 500g standard batch test was carried out in a Denver machine. The frother used in these flotation experiments was Dowfroth 250 (Dow Chemical Company, Midland, Michigan, USA) at a concentration of approximately 25 mgpl. (iii) flotation - total xanthate reaction relationship.
Flotation of samples conditioned at various pH's in the alkaline range have shown that flotation of both ethyl and amyl xanthate-treated chrysocolla is independent of pH, per se, but dependent only on the extent to which the reaction has progressed, (and thus, from section 7e on the quantity of CuEtX/EtX2 held at the external mineral surface). This can be seen from Figure 21, and from the fact that two samples of chrysocolla conditioned identically at pH 5.0 with 10-4M EtX under standard conditions were floated in their conditioning solution at different pH's under quiescent conditions. Flotation at pH 5.0 and at pH 11.0 were both virtually 100%.
In the Hallimond tube the maximum flotation obtainable with ethyl xanthate was 50%, when 3.7 x 10 4 2 moles CuEtX/m had been formed. Virtually 100% flotation was possible with KAmX when between 4.5 and 3.0 x 10-4 91.
FIGURE 21: CORRELATION BETWEEN XANTHATE REACTION AND QUIESCENT FLOATABILITY
recovery: 1
----9"8------an''.-- 08 apyl xanthate 80
09 0 LO
60
.9
40 --. T1-0 11 4.10 4.10 .11 • .9 oiD *9 8 .9 20
0
0 1 2 3 4 5 2 4 xanthate reaction: poles/m x10
(numbers at points refer to the pH of the xanthate reaction) 92.
2 moles CuAmX/m had been formed.
However, in the agitated froth cell, no flotation of chrysocolla wF,s obtainable with ethyl xanthate in the alkaline region despite adequate quantities of availableJlanthate and sufficient conditioning periods. But good flotation was possible under turbulent -4 2 conditions around pH 5. Here 2.9 x 10 moles CuEtX/m had formed at the end of the conditioning period, and 2 approximately 4.6 x 10-4 moles CuEtX/m at the end of the flotation period. These results are given in Figure 22.
This successful result at pH 5 is equivalent to a xanthate consumption of 4 - 7 lb/ton ore if the following assumptions are made: ore containing 2% Cu (as 2 chrysocolla); specific external surface of 2000 cm /g; no other ion or mineral consuming xanthate; all xanthate entirely reacted.
The method on a Union Miniere ore, assaying 6.0% Cu, was only partially successful in floating some chrysocolla at pH 5-6 using nearly 10 lb/ton of ethyl xanthate. Some of the concentrate was, however,miachite and the recovery was less than 30%. This experiment was conducted in a 500g Denver laboratory flotation machine. g. Factors affecting_ chrysocolla-sulphide reactions.
(i) the reaction vessel.
Chrysocolla-sulphide reactions were conducted in the apparatus described in section 7d (i) above.
(ii) experimental procedure.
Whilst the experimental procedure closely followed that described in section 7d (ii), two important 93.
FIGURE 22: TURBULENT FLOATABILITY OF CHRYSOCOLLA JITH ETHYL XANTHATE
recovery:100
80
60
40
20
2' 4 6 8 10 12
pH
t• 94.
difference should be noted. The equilibrium:
Na S + 2H20 --s• 2NaOH + H 2 2S means that a sodium sulphide solution will rapidly become a sodium hydroxide solution, unless especial care is taken to prevent loss of H 2S from above the solution surface. It was therefore necessary to ensure that all the seals on the reaction vessel were gas-tight, and secondly, that no purging nitrogen was used.
(iii) analytical procedure.
Sodium sulphide was estimated directly on solutions diluted to approximately 0.5 x 10-4M, by a simple spectrophotometric technique. This measured the intensity of the sulphydryl chromophore corresponding to an absorption peak of 2280 it was necessary to make solutions alkaline (>pH 9) to ensure virtually complete conversion to sulphydryl ion. The calibration for the method is given in Figure 23.
(iv) effect of pH.
The relationship between pH and the extent of sulphide reactions is illustrated in Figure 24. Because of difficulties in preventing the loss of H2S in the neutral and acid pH range, only the alkaline region was studied. The form of the curve shown is similar to that obtained for CuEtX formation (Figure 13), but the extent of the reaction in a particular time is approximately ten times as great.
(v) kinetics.
The rate of CuS formation is shown in Figure 25. As with CuEtX, each curve corresponds closely to 1/2 P = bt t , which is in accordance with the findings of Mitrofanov et al77'78, Kushnikova et a179 and Strigin 95.
FIGURE 23: ABSORPTIOMETRIC DETERMINATION CF Na2S.9H22
optical00 density 228mg 1 cm cells 0.4
0.3
0.2 •
0.1 •
0 0 0.2 0.4 0.6 0.8 1.0 LNa2S]: gpl 96.
FIGURE 24: EFFECT OF pH ON SODIUM SULPHIDE REACTION
CuS 25 formed: m2les/4 in x10
20
15
10
0
4 5 6 7 8 9 10 11 12 pH 97.
FIGURE 25: KINETICS OF SODIUM SULPHIDE REACTION
PH
8
CuS 20 miles/4 m x10
9 15
/ /
10 / /
/ 10 / ♦ 5 11 12
0
0 5 10 15 20 25 30
tine: minutes 98. 80 and Kushnikova . Values of the pH-dependent parameter 'b' are given in Table 13; and the calculated curves for pH 8 and 12 are sAown in Figure 25 as broken lines. (vi) effect of concentration. The work of Nitrafanov et al77'78, Kushnikova et 80 a179 and Strigin and Kushnikova , is also borne out to some extent by the data of Figure 26, which shows the amount of CuS formed at pH 10 after 30 minutes. r = c c l/d They described the form of the curve as t o where c and d are constants ka/1 . 0.25 - 0.65) and Co is the initial sulphide concentration. The curve obtained here gives a a value nearer to k.3 (shown in Figure 26 as a broken line). h. Distribution of chrysocolla-sulphide reaction products
In order to assess the degree of attrition resulting from turbulent agitation, small samples of chrysocolla were sulphided at pH 7.0, whilst being agitated with 10 times their weight of silver sand. In four similar tests (extending for differing periods of time), the sulphide solution was decanted, dried by evaporation, and the residue dissolved in 5 ml of cold 0.1 M HNO3. This solution was analysed for Cu2+polarographically - the conditions of the analysis, and a calibration for the method starting from Cu(NO3)25H20 are shown in Figure 27.
Figure 28 shows the relation between the total chrysocolla-sulphide reaction, and the copper sulphide formed polarographically in solution. It can be seen that under these circumstances of excess sodium sulphide -2 (50 ml 10 M Na S.911 0 were used with 10 mg chrysocolla) 2 2 even violent attrition does not remove all the sulphide 2+ film. Some of the Cu analysed could, of course, arise 99.
TAT-ME 13: VALUES OF lb' IN THE EXPRESSION P bt2.
pH 3 1 9 1 10 11 12 I .1. I I b - moles/m2//minute2 x 104 4.04 12.82 1.30 0.96 0.67 I 1 100. 59 CuS 40 --__ formed m2les/4 m x10 FIGURE 26: EFFECT CF SULPHIDE OONCENTRATIOk ON CDIUU EULPHIDE REAL TIC'U
35
•
30
25 1
/
/
20
/ /
• 15 1 /
/ 10
5
•
0
0 1 2 3 4 5 sulphide concentration: molarity x 10 101.
2+ FIGURE 27: POLAROGRAPHIC DETERMINATION OF Cu IN 0.1 M }1NO3
peak height M
15
10
5
0
0 0.01 0.02 0.03 0.04 0.05 2+1 Luu j: gpl 102. CuS 45 formed or a scoured m9les/4 m x10 40 FIGURE 28: .ISTRIBUTION OF CuS IN ITRBULENT CHRY\SCCOLLA—S I: IDE REACTI'
0
35
total formed
0 30
25 0
20
15
10
5 oun• .aspersed in solution
• 0
0 5 10 15 20 25 30
time: minutes 103.
from attrited chrysocolla.
i. Flotation of sulphide-reacted chrysocolla.
Sulphided chrys,3colla was shown to be very slightly floatable. However, its flotation response was not significantly increased by contacting the sulphided chrysocolla with potassium ethyl xanthate at alkaline pH's; when sulphidation was carried out to 2.0 x 10-4 2 moles/m , 2J subsequent ethyl xanthate reaction could be detected,
Much better results can be obtained with potassium amyl xanthate. In fact, it is shown in Table 14 that a 5-fold reduction in amyl xanthate necessary for good flotation can be made under quiescent conditons, over that which is required in the absence of sulphidation.
It has been established that simple pre-oxidation of the sulphided surface was not the cause of poor flotation with ethyl xanthate. Even lengthy periods of oxidation with 0 or H 0 produced no enhanced flotation 2 2 2 effect in aqueous solution. It was, however, possible to obtain 31% flotation in the Halliniond tube by drying the sulphided chrysocolla for 1 hour at 100°C, wetting the sample with water to re-effect full hydration and reacting with ethyl xanthate for 5 minutes at pH 9. When normal conditioning periods of 10-15 minutes were given, at neutral to alkaline pH turbulent flotation of the chrysocolla from the Union Miniere ore (total Cu% = 6.01%) was not apparent, even when amyl xanthate in large auantitiep IT-.r; used: shorter conditioning times, however, produced visible green concentrates at alkaline pH, These concentrates contained malachite, - despite hand-sorting prior to crushing the sample - but also some of the chrysocolla. 104.
TABLE 14: FLOTATION OF SULPHIDED CHRYSOCOLLA WITH AMYL XANTHATE.
Total Reaction Hallimond (pH of sulphidatiOn and tube subsequent xanthate CuS 1 CuAmX Recovery reaction = 9.0). moles/ moles/' m2x104 m2x104
0 1.00 35 0 2.50 95 0.5 1.00 43 1.0 0.50 95 2.5 0.25 47 2.5 0.50 99 1015.
Inssufficient ore sample was available for any real attempt at optimising flotation conditions for the sample, and it was found expedient to approach most of the variables on a 'trial-and-error' basis in a few qualitative experiments: the more interesting topic of critical conditioning period could thus be investigated more thoroughly.
The experimental conditions which proved suitable for investigating the conditioning period are summarised in Table 15. Grinding the ore under the conditions stated substantially liberated the chrysocolla (>95%); the product being 90% -100 mesh BS. The variation in recovery and grade of copper with conditioning period is shown in Figure 29. It is considered that these graphs of Figure 29 reflect what is happening to the chrysocolla: a mineralogical examinationsuggests that approximately the same quantity (about 2% of the ore by weight) of malachite is floated on each occasion, under these conditions. One of the more noticeable effects during the conditioning period, was the darkening of the pulp On adding the sodium sulphide-xanthate mixture. Flotation concentrates were 90% -150 mesh BS. j. Mechanism experiments.
(i) leaching.
Leaching, first mentioned in section 5c, will be discussed here in greater detail. It is important since it suggests a mechanism by which copper in chrysocolla may react with reagents such as acids, precipitants and ciamplexants.
Copper is only fully leached from the mineral at acid concentrations >c.1% H2SO4. When the acid concentration is less than this value, an equilibrium 106.
TABLE 15: EXPERIMENTAL CONDITIONS OF 500g FLOTATION TEST ON UNION MINIERE ORE.
Mill: 6" rod mill Ore charge: 500g
Pulp density: 50% solids ING L Milling time: 30 minutes MIL pH: 7.5 Speed of rotation: 72 rpm
Conditioner: Denver 500g laboratory machine Speed of rotor: 1000 rpm NG Pulp- density: 20% solids ONI
TI pH: 9.5 Reagents: Na2S.9H00: 5.0 lb/ton ) added simultaneously at the
ONDI ,. C KAmX : 0.2 lb/ton ) start of conditioning Dowfroth 250 : 0.2 lb/ton - immediately prior to flotation
Aeration: maximum possible under cell conditions outlined above 0 F-4 Concentrates: 0 - 1 minutes ) Ei 0 1 - 3 minutes ) analysed for copper separately 1 •i,ai 3 - 5 minutes 107.
FIGURE 29: EFFECT CF LENGTH OF CONDITIONING PERIOD ON FLOTATION OF KOLWEZI ORE
recovery: 25 grade
80 20
recovery 60 15
40 10 O grade
20 5
0 0
0 2 4 6 8 10
time: minutes 108. condition exists despite an excess of acid available, Here a colourless gel layer is not apparent, but the entire mineral is evenly coloured - frequently only slightly lighter than the original sample.
The rate of leaching was investigated in various nAvt solutions at two rates of agitation: a rapid in which the gel-layer, whore formed, was significantly removed; and a slower rate in which the gel layer remained significantly adherent. The quantity of copper leached was determined by the sodium diethyldithiocarbamate method an, 1 is shown in Figure 30.
The overall leaching process is the transfer of electrolytes to and from the exterior of the chrysocolla via micropores, filled with solution. Two stages are recognisable in the process: the transfer of electrolyte across a quiescent film of solution surrounding the particle; and the transfer of electrolyte through a layer of the particle. In each case transfer of the electrolytes (H2SO4 from the bulk solution into the particle, and CuSO4 from the interior of the particle into the bulk solution) will be effected by diffusion. The overall rate of leaching will be governed by the slowest stage which may be the diffusion of CuSO4 or H SO through the film or the particle, and is said to 2 4 be film or particle diffusion controlled respectively.
In general, the rate of leaching of chrysocolla was found to increase with decreasing pH, but the faster rate of extraction with turbulent agitation in which the gel layer was substantially removed (i.e. comparing the curves for 'quiescent : loo and 'turbulent : H2SO4 0.5% H So 1), indicates that the rate determining process 2 4 is the diffusion of ions through the porous solid. At 109.
FIGURE 30: ACID LEACHING OF KOLJEZI CHRYSOCOLLA AT DIFFERENT pHs A.712 AGITATIONS
Cu extraction turbul 5X, H2S 0 •
8
urbul n • .5/, H '04
0 60
quiescent; 1 H
40 e
20
quiescf24; /0 CH Quiescerlit; 1% HIBO
0 10 20 30 40 50 60
time: minutes 110. low pH's, therefore, particle diffusion is the rate determining step.
The leaching phenomenon is similar to ion exchange processes; copper being exchangeable for hydrogen. Both particle and film diffusion in ion exchange resins have been given the attention of Boyd et a193, who evaluated for spherical microporous particles that -3D C-t E = [1-exp (-5 2 )] 1/2 and E = 1-exp (1,5 x) for r t2 1 particle and film diffusion control respectively. In these expressions, E = e /e .(e eoo are respectively t oo t , the extraction at time t and 00); 1) is the particle 2 diffusion coefficient kl0I -7- 10-5 cm /sec for typical organic resins); t is the time of leaching and r is the particle radius (he.re 900; D is the film diffusion -5 2 coefficient (c. 10 cm /sec for typical organic resins); 2+ C is the total concentrations .of H+ and Cu in solution; S is the film thickness (c. 10-3 cm); and t": is the total concentration of H+ and Cu2+ in the particle,
.the. leaching experiments with chrysocolla, the data have been determined with insufficient precision to permit any strict comparison with either of Boyd et al's93 rate equations, For example, when the acid concentration is sufficient to leach all the copper,.and the agitation is turbulent, the data approximately fit both the particle and film-diffusion expression, as can be seen from Figure 31, If, however, the earlier contention of particle diffusion control is assumed for these strongly acid conditions, then it is calculated that the particle- -6 2 diffusion coefficient is equal to 1.3 x 10 em /sec. 6 2 ThiS compares with a typical value of 2 x 10 cm /sec for ion exchange resins.
Ion exchange and associated equilibria can readily FIT= 7,1: TESTING P1.P7ICLE nR FILM DIPUSIrrl r'n-7CL 1CIL LaC=G log(1-E2) 0 (pt- rticle diffusion) • or log(1-E) (film diffusion) x
-0.
;article Oilfusion
\ film diffusion .... \ -... .„. '--4( \ \
0 10 N. 20 30 40 time: :orates
(leaching conditions: turLulent; 5 H2SO4) 112.
be demonstrated under even weakly acid conditions. Three experiments were performed in the pH range 5.0 to 5.9 -1 using 10 M acetic acid-sodium acetate buffer solutions (100 ml) and chrysocolla (l00 or•5000) in the reaction vessel under mild agitation conditions for several hours. An experiment at pH 4.6 was conducted with the same buffer in a sealed vessel under static conditions for several weeks. In the agitated experiments, the rates of leaching, and the attachment of equilibrium were ascertained by frequent sampling of the solution and analysing for copper by the sodium diethyldithiocarbamate method. In the static experiment, equilibrium was not conclusively proved, although seven weeks of contact between the chrysocolla and the solution were allowed before an analysis of the solution for copper was made.
Ion exchange in alkaline solution is not so readily determined because of the low copper concentration in solutions. In order to obtain a measurable copper concentration in solutions, the complexant glycine, CH 2.NH2.CCOH was used at p11 11.0. In a system containing chrysocolla and glycine overall equilibrium is attained when three sets of individual equilibria are established. These are the 2+ equilibria between Cu and H1- in the mineral and the 2+ bulk solution; the equilibrium between Cu and copper- glycine complexes in solution; and the equilibrium between the three ionic forms of glycine and 11+ in solution. These equilibria are defined by the following equations:
[Cu2+] . XH 2 K - (1) 1-[11+]2 x Cu 113.
[cul,+ ] - (2) K2 = [Cu24][L-]
[CuL2] K - (3) 3 [CuL+][L-]
[HL][H+] K4 - (4) [H2L+]
[L] [114 ] K - (5) 5 [HL]
In these equations X1L, Xell are defined as follows:- The total number of equivalents of exchangeable cations, in a weight, w, of chrysocolla is n, and the number of equivalents of exchangeable hydrogen and copper is nH and nCu Thus n + n = n - H Cu (6) n and X_ H ; X nCu n Cu n - (7) Now if the number of gram atoms of hydrogen and copper in the mineral is initially qH and q Cu then: nH = cIH ; nCu = 2qCu (8) Substituting equations 6 and 8 into equation. 7-: 2q qH X = Cu H q + 2q XCu - q + 2q - (9) H Cu H Cu In equation 1 through 5, Kl_5, are the equilibrium constants, and L is equivalent to the ligand CH2.NH2.000. As the result of attaining a leaching equilibrium, x gram atoms of. Cu2+ are exchanged for 2x gram atoms of H+, and the new condition is described by 114. [Cu‘. 1(xeq)2 J k H / [14..+. ]2 (xe % - k10) will apply. Cu/ 2(q - x) eq H + 2x In equation 10, Xeq - Cu -(11) - c11-1 2c1Cu Cu qH + 2qCu Now the total copper concentration determined in solution [Cu]T by the analytical method is given by: [Cu]T = [Cu2+] + [CuL+] + [CuL2] - (12) Similarly, the concentration of glycine, in all forms is given by:
[L]T = [H2L+] + [HL] + [L-] + [CuL+] + 2[CuL2] - (13) Published data94 for K is 10-2.35 and for K is 10-9.78 4 5 so at pH 11.0 the ratio of [H2L+] : [HL] : [L] = 10-2735 : 10-9'78 : 1, and as the glycine was present in excess of [Cu]T (a factor of 20x), equation 13 can be approximated to:
[LIT = [L] (3.4) 94 -8.29 -7.61 Published data for K2 is 10 , and for K is 10 3 -8.29 so the ratio of [Cu2K ] : [CuL-1-] : [CuL2] = 10 -7.61 10 : 1. Thus equation 12 can be approximated to:
[Cu]T = [CuL2] (15). By combining equations 2 and 3, 14 and 15 a value 2+ , of [Cu ] was obtained for two values of XIS and XCu. If x V - (16), (where V is the volume in litres of solution in contact with a weight w of chrysocolla), then from equations 9, 11 and 16 eq 2[Cu]T V XH = X + - (17) H 2nCu + nH
115.
eq 2[Cti] V and X X T Cu Cu - (18 ) nCu nH Although ncu, nH can be estimated from the analysis of the chrysocolla sample, Xcu, XH cannot be measured with certainty. However, if it is assumed that the copper in chrysocolla is stoichiometrically associated with Si02, then since there is only 0.85 mole CuO to account for each mole of SiO X may be taken as 85%, and X as 2' Cu H 15%.
From equation 10, the equilibrium constant, K may 1 thus be calculated. The values of K 1 obtained at pH 11, together with those obtained at acid pHts are shown in Table 16. It is noted that the copper extracted from chrysocolla at pH 11 with gIycine is approximately 103 times that which reacts with potassium ethyl xanthate at the same pH, in the same period of time.
(ii) stability of CuEtX at alkaline pH.
In order to confirm the earlier findings of Du Rietz95, that solid copper hydroxide is unstable in the presence of 10-4M EtX- solutions (and therefore unlikely to be present in the chrysocolla-xanthate system investigated here), a series of experiments conducted in which Cu(OH), suspensions were reacted with 10 414 EtX solutions. Both reactants were previously adjusted to the same pH.
After 5 minutes reaction, the CuEtX (or remaining Cu(OH)2) suspension was pressure filtered through a plastic membrane; and a sample of the filtrate analysed spectrophotometrically for xanthate. The results are shown in Table 17, The work of Du Rietz95 is confirmed: below pH 11, Cu(OH) 2 (solid) is unstable in the presence
116.
[Cu21 TABLE 16: VALUES OF K 1 [11+]2 x AT VARIOUS pHTS AND XH'S. Cu
pH X K H 1 4.6 0.32 10*4*9
5.0 I 0.24 10+6.0
5.4 0,18 10+5.6
5.6 0.16 10+5.9
11.0 0,17 104-5.6
11.0 0.29 10+5.3 117.
TABLE 17: STABILITY OF CuEtX at ALKALINE pH.
pH 8 9 10 11
Cu(OH)2 --.), CuEtX + EtX2 -% 95 i 92 67 1 118.
of potassium ethyl xanthate kl0/ -4 M).\ Using the solubility product data of Geloso and Deschamps96 = 10-18.3.) (Cu(OH) = 10-24.2\ 2 and Du Rietz97 (CuEtX2 it was calculated that solid Cu(OH) 2 is stable only at pH's greater than 13. The rate of reaction, is, however, very slow when pH 11 is exceeded.
(iii) ethyl dixanthogen. In order to determine the role of ethyl dixanthogen in chrysocolla-xanthate systems, ethyl dixanthogen emulsions were contacted with chrysocolla. They have been shown not to react perceptibly with chrysocolla over the pH range 4 through 11: furthermore they did not induce flotation under quiescent conditions. Fifteen minutes conditioning time with 10-4M dixanthogen emulsions were used in these experiments. No visual, or detectable reaction occurrea between solutions 2+ containing Cu ions and ethyl dixanthogen emulsions. The dixanthogen was made by reacting iodine (dissolved in potassium iodide solution) with a concentrated potassium ethyl xanthate solution. The dispersed dixanthogen was concentrated and washed with water in a high-speed centrifuge.
(iv) hydrophobicity of co-precipitated CuEtX and EtX2.
Co-precipitated cuprous ethyl xanthate and ethyl dixanthogen produced in the experiments described in section 7j(ii) were found to be very hydrophobic, irrespective of the pH of formation: the precipitate was readily floated when air was bubbled through the solution. (v)surface charges
Some attempt has been made at assessing the 119.
electrical charges on the surface of chrysocolla, cuprous ethyl xanthate, ethyl dixanthogen droplets and cupric sulphide, Qualitative, rimple electrophoretic methods were used. The solutions for these experiments contained only H2SO4 or KOH, in order to obtain the desired pH, and some dissolved air.
From the migration of chrysocolla along a strip of filter paper immersed in a colloidal suspension of the mineral, it is apparent that below pH 6, chrysocolla has a definite positive charge, and above pH 10, a definite negative one Here, a 50 V potential was applied to the strip for 12 hours,
Tthyl dixantho,cen droplets were shown to have an iso-electrie point in the region pH 5-6, with a positive surface charge below pH 5, but the charge on cuprous ethyl xanthate could not be determined by these simple rraans, It muet be concluded that CuEtX has a very low surface charge, CuS has a negative charge at pH>41-. ty4 \ -/ e Sampler; of IraIachite have been shown to react with xantate and soluble sulphide about 1.7 times as rapidly as does chrysecollaa however, nearly the same quantity of xanthate reaction is required to achieve the same flotation effect with malachite and chrysocolla. These results are sl:mmarised in Table 18 and 19. A more significant difference between chrysocolla and malachite is in the ease ;.lith which they can be attrited, thus scouring the films from their surfaces, Samples of malachite and chrysocolla were reacted with ethyl xanthate under quiescent conditions, and then agitated in the reaction vessel with lOg of similarly-sized quartz for 120.
TABLE 18: RELATIVE REACTION RATES OF MALACHITE AND CHRYSOCOLLA.
Reaction Time - Minutes 5 1 10 20 30 (Reactions carried Malachite 0.71 1.2 1.9 2.5 out at pH 10.0 J 1;8H for 30 min. at Chrysocolla U ° X 0.5i 0.8 1.1 1.2 -4 10 M) Malachite ‹21 8-- 17.6 i 18.1 23.4 23:7
', rc)-4 GI 1 Chrysocolla C-.3 0 r., 7.9 10.0 12.7 159 121
TABLE 19: RELATIVE FLOATABILITY OF MALACHITE AND CHRYSOCOLLA.
1 Hallimond Tube Recovery % Total Reaction pH 2 4 CuAmX moles/im x10 Malachite Chrysocolla
0.5 5.5 20 14 1.3 5.5 70 56 1.3 9.0 41 31 2.5 5.5 98 97 2.5 9.0 98 95 122
various periods of time. Samples of the liquor (that 2+ passed a 150 mesh BS sieve cloth) were analysed for Cu : the results are presented in Figure 32. Chrysocolla is seen to scour much more readily than does malachite, by the greater slope of the chrysocolla curve in the scouring region. The scouring of copper exceeds, in the case of chrysocolla, the formation of CuAmX (i.e. 1.25 x ‘ 10-moles/m2} because of attrition of the mineral.
(vii) diffusivity of KEtX and NaHS. In order to estimate the relative diffusivity of KEtX and NaHS, the equivalent conductivity (since this is a measure of the current carried by diffusing ions) of potassium ethyl xanthate and sodium hydrosulphide -2 solutions were measured at 10 M. Water used for the solutions was vigorously outgassed with oxygen-free nitrogen, and a standard dipping-type conductivity cell with platinum black electrodes was employed with an alternating current conductance bridge in order to reduce polarisation errors. The values obtained for potassium ethyl xanthate and sodium hydrosulphide were -1 2 108 and 189 ohm .cm respectively: that is, the NaHS diffuses in aqueous solution 1.75 times as fast as the KEtX. r + 1 (viii)effect of LK 1 on xanthate reaction. When KC1 was added (up to 2M) to the solution in which the chrysocolla reaction occurred with xanthate, a noticeable correlation between [KC1] and extent of reaction in a certain period was apparent. Nearly 3 times the reaction occurred at pH 8 when [KC1] = 2M than when [KC1] = 0. -4 (ix.) stability of CuS in the presence of 10 M EtX- In the presence of 10-4M EtX-, CuS sols are stable 123.
FITURE 32: SOLDRiNG OF SURFACE FROM OHRYSOCOLLA AND NALACHITE 2+ scoured m9les/4 m x10
2.5
2.0
1.5
chrysccoll a •
1.0
0.5
malachite
0 filming scouring 0 5 10 15 +5 +10 +15
1.25 x10-4 moles CuAmX formed at time: minutes pH 9.0 for 15 minutes. Filmed chrysocolla and malachite then scoured by agitation with 100 times their weight of quartz in separate tests. 1214 in the absence of air; in air, however, they are slowly converted to CuEtX/EtX2 sols. 125.
8. DISCUSSION ON CHRYSOCOLLA FLOTATION SYSTEMS. a. Leaching of Cu2+
The leaching data described in section 7c suggest that chrysocolla is an ion exchanger of the weak acid, cationic type: exchangeable cupric ions being held in an insoluble, silicate matrix. The exchange of Cu2+ has been found to result in an even less crystalline material than is the original chrysocolla; in fact 2+ when the Cu is fully leached from the mineral a material resembling amorphous silica gel is produced, whilst partial leaching leads to an in-4rmediate degree of crystallinity. It seems, therefore, that the exchange 2+ involves rather more than a simple release of Cu for equivalent sorbed 114.: some structural changes are also occurring. The exact nature of such changes has not been elucidated.
The release of Cu2+ for other cations in solution is clearly important in accounting for the xanthate and sulphide reactions with the released Cu2+. This 2+ released Cu will, when the entire mineral particle is at equilibrium, be in the chrysocolla pores and in the bulk solution. Under non-equilibrium conditions the 2+ exchanged Cu will be concentrated in the mineral pores, and the external surface of the mineral.
As has been shown in section 7j(i), the equilibrium condition is represented by: [Cu2+] x2 C l r +1 2 LH J Cu A mean value of K 5.6 1 of 10 has been calculated from the conditions and results of the 6 leaching experiments (4 in the acid region, 2 in the alkaline). Individual
126.
values are shown in Table 16 and are seen to have a random scatter. Using the mean value of K1, the equilibrium [Cu2+] v pH diagram (Figure 33) has been constructed for various values of XH. Using this value for K1, it can be tested whether chrysocolla consists merely of cupric hydroxide with silica gel. If no cupric hydroxide r r 2 is present in the original mineral then LOU L OH-] < -18.3 10 - the solubility product constant for Cu(OH)2 (Du Rietz95\ 1. Since [0][0H] = 10 4.o -1 the -18.3 dissociation constant for water, then 10 [H+]2 10-28.0 < 109.7. No Cu(OH)2 is present, because it has been [Cu2+1 105.6 . 0.85 7.2. shown that 10 [H4 ]2 (0.15)2 For the entire mineral particle to attain equilibrium with a bulk solution requires a period of 2 days or so.. This period is greatly in excess of the time over which the xanthate and sulphide reactions were investigated; such reactions must therefore be considered either non- equilibrium, - or, and this is more plausible, - representing an equilibrium between a portion of the chrysocolla (an outer shell of the particle) and the bulk solution. In this latter case, a particle, diameter 9011, might be expected to be at equilibrium in only its outer 1011 shell after 30 minutes. b. Xanthate reactions
Because of the dependence of the exchange of H-4- for 0u2+ upon pH, the rate of formation, and deposition of CuEtX and EtX, can also be shown to be strictly a function of pH. The following argument applies strictly to the situation of the experiments undertaken here, viz; fully- hydrated chrysocolla being added to 10-4M EtX solutions 127.
FIGURE 33: EQUI_1_,:dUM LCu2+] v pH DIAGRAM FOR KOLWEZI CHRYSCCOLLA
log {Cu2+
0 ...... ---.0,(uH)2
.,, .„, -10 Et4_,E1,0„,.. -6 11.__ XH=0.001 Etr=L0-4M XH=0.01 AmX =10-cl XH=0.1 -2.0 XH=0.5 XH=0.9 XH.0.99
-30
-40 IP,-=10-4M
-50
4 5 6 7 8 9 10 11 12 PH (the probable effect of e, Na+ above pH 10.5 is shown for XH = 0.001 as a broken line) 128. maintained at the appropriate pH. With suitable minor modifications the theoretical picture now presented can be extended to situations w- ere only partially-hydrated chrysocolla is used; where xanthate is added to chrysocolla suspensions; or where the pH is uncontrolled. 2+ The reaction of xanthate with Cu in equilibrium with the outer layer of the mineral, and in the ,v,face pore 50.1ut- an at approximately p11 will produce, in itself, an insignificant quantity of CuEtX and EtX2. However, such precipitations will disturb the equilibrium condition in the surface pores and further exchange:Cof H for Cu will occur, with a result that the pH of the + surface pore solution will rise. The equilibrium LCu, ] will now be reduced, and further exchange of H for Cu (and subsequent xanthate reaction) will depend either on the diffusion of H+ into the pores from the bulk solution, or OH- into the bulk solution from the pores.
The situation at pH <4 with 10 4M EtX is shown in Figure 34a. Here the diffusion of acid will be more rapid than that. of xanthate since [H2SO4] > [KEtX] and the diffusion coefficient of H2SO4 is greater than that of KEtX. This will result in a high LCur 2+1j in the pores near the surface, where the pH will be only slightly higher than that of the bulk solution. Now CuEtX and r 2+ , EtX2 will be formed in any environment where the LCu j -16 2 -4 exceeds 10 ' M in a 10 M EtX solution, because Ksp for the cupric xanthate is 10-24.2 (Du Rietz97): even if the xanthate concentration is below 10-4M, owing to local depletion, then precipitation of CuEtX and EtX2 will nonetheless occur in the bulk solution, and at a rapid r 2+ 1 rate, since the LCu j at the external surface will be -1 as high as 10 M for XH = 0.15 (see Figure 33). 129.
FIGURE IONIC CONCENTRATIONS CF XA THATE—REACTING CHRYS0cOLLAS AT vARI0US..0
BULK SOLUTION PORE SOLUTION
BULK SOLUTION PORE SOLUTION
pH>7
10 4 Ltx OH
BULK SOLUTION PORE SOLUTION
Horizontal scale is the distance perpendicular to the mineral surface; vertical scale is ionic concentration. 130.
Figure 34b shows the ionic concentration profiles at 4 Above pH 7 in the bulk solution, there is no question + 2+ of H diffusing into the chrysocolla to leach more Cu from the pore walls, and precipitation will occur principally within the surface pores, since after the initial, insignificant reaction, there will be insufficient 2+ Cu at the external surface for precipitation to occur there. The reaction will be slow,primarily as a result 2+ of the low [Cu ] in the pores at alkaline pH's (figure 33). The fact that the xanthate reaction rate remains substantially constant over the pH range 7 to 9.5 - 10.0 is because the internal, pore pH. at the point of reaction is nearly constant over this range. For example, at pH 7.5, the alkali concentration in the bulk solution is 3 x 10-7M: when the reaction is at a steady state, the diffusion of xanthate to the point of reaction, and the diffusion of alkali from the point of exchange, - this latter, into the mineral, as well as towards the mineral surface, - must be equal. If the xanthate concentration -4 is 10 M, and the diffusion coefficient of alkali is 131. assumed twice that of xanthate, then because the concentration gradients for alkali and xanthate are proportional to 2(z/2 - 3 x 10-7) and 10-4 respectively, 4 they must be equal. Hence z 10 M (i.e. a pore pH of 10.0). At pH 8.5 the concentration gradients are /z/ 2k 2 - 3 x 10-6) and 10-4 for alkali and xanthate respectively, and z is still equivalent to a pH of some 10:0. AboVe pH 845 - 9.0, the internal pH rises above 10.0, and the reaction rate begins to fall again Furthermore, it can been seen from Figure 33, that if the xanthate concentration falls to 10 6M, no further precipitation of CuEtX and EtX can occur, because the 2 solubility product of CuEtX2 is not exceeded. No -4 reaction is possible between chrysocolla and 10 M EtX above an external solution pH of a. -3proximately 11.6. Concurrently with the pore reactions which take place above pH 7, it is possible that a certain amount of xanthate reaction will occur with Cu2+ originating from external surface sites of the solid. When the overall reaction has proceeded to any extent, however, diffusion of Cu2+ to the external surface, via the solid matrix, will be considerably slowed. The formation of CuEtX and EtX near the surface 2 of the mineral will slow down the diffusion rate between the pore reaction sites and the bulk solution when the pores are partially, or fully blocked with precipitate. The effect on reaction rate would be expected to be independent of pH, but critically dependent upon the extent of reaction in the pores: pore blockage is clearly a contributory factor for the slow reaction 132. apparent above pH 7. The enhanced reaction of chrysocolla and xanthate in the presence of ammonium hydroxide is the result of two effects. Primarily, the formation of the soluble cuprammonium complex ion will result in an increased total copper concentration in the pore solution, and as the cuprammonium ion Cu(NH ion is unstable in the 3) 2+ presence of EtX, more CuEtX and EtX 2 are formed, and at a faster rate. Secondarily, ammonium hydroxide will act as a buffer in the pores of chrysocolla, and prevent the pH rise that occurs at the point of reaction. Thus the overall effect is to produce precipitation conditions comparable to those found at more acid pills. The quantity of copper that can be extracted from chrysocolla in 30 minutes at pH 11.0 with glycine is approximately 103 x that which can be precipitated with xanthate, or ammonium hydroxide and xanthate. The rates of leaching and of the xanthate precipitation are determined by the rates at which 114- can diffuse to the 2+ reaction sites and Cu diffuse away, and would thus be expected to be comparable. But the formation of a pore blockage in the case of xanthate results in a slower diffusion rate. In addition, the stability of the copper-glycine complex (i.e. the low [Cu24] in equilibrium with it) will result in a greater extraction of Cu2+ from the outer layers of the mineral in this time. The glycine leaching rate must therefore be greatly in excess of the xanthate reaction. At pH 4.6 the quantity of copper that can be leached with a sodium acetate-acetic acid buffer is much more comparable to the CuEtX formation with xanthate. 133. -2 2+ 2 In the simple leaching of copper some 10 moles Cu /m is exchanged after 30 minutes and for the CuEtX case, -3 2+ 2 some 10 moles Cu /m are formed. The factor of 10 is well accounted for by the buffering action at the site of copper-hydrogen exchange and the absence of a scarcely-penetrable precipitate. The explanation of the effect of increasing [K+] on the CuEtX-EtX 2 reaction may lie in the exchange + within the mineral of Cu for K . However, it is not. unusual for reaction rate to change by a factor of 3 or more in the presence of excess salts. It would be expected that increasing the cationic concentration by any method would have resulted in a similar effect. The rise in pH observed when chrysocolla is reacted with xanthate in acid solution is consistent, in a qualitative sense, with the ion exchange mechanism proposed earlier, 2+ 2- _+ Cu .SiO 2KEtX + H SO 2h Si032 + CuEtX + 3 2 4 2 1 K SO (and CuEtX --0- CuEtX + —EtX ). 2 4 2 2 2 However, the ratio of consumed to CuEtX H2SO4 2 produced would be expected to be 1:1. The poor agreement of Table 8 with this theoretical result must be attributed to experimental error (probably in adding the small quantity of acid necessary to maintain the required pH), becEmse there is considerable independent evidence concerning the isoelectric point of silica precluding the adsorption of 114- by silica gel at pH's greater than approximately 3. No pH rise could be detected in alkaline solution because of the small amount of exchange, and because of difficulties irr absolutely 134. excluding CO2 from the cell during the reaction. The contention that the hydrophobic species produced in chrysocolla-xanthate reactions is based not only upon the determination of cuprous ethyl xanthate and, especially, dixanthogen: it can also be calculated that the alternative species, cupric ethyl xanthate may not be stable under the experimental conditions employed here. The species in equilibrium with a particular concentration of xanthate ion can be calculated, as follows, from the solubility products of cuprous and cupric ethyl xanthates, and the free energy change associated with the reaction. 2+ + 2Cu + 2EtX A 2Cu + EtX (1) 2 which can be expressed in terms of an e.m.f. The ratio a +/a 2+ in equilibrium with the ethyl xanthate ion Cu Cu concentration and dixanthogen is related to the e.m.f., e, associated with the redoxreaction (1) by the equation o 2.303 RT +)2.(a e = e log (aCu EtX2 ) (2) z F 10 2 (aCu2+)2 (aEtX-) o where e is the standard e.m.f., R is the gas constant, T the absolute temperature, z the number of electrons taking part in the reaction, F Faraday's constant, and 'a' activities. Using Du Reitz's97 value of -0.07v for the standard electrode potential for the reaction 2EtX EtX2 + 2e ( 3 ) and the value +0.17v for the standard electrode potential for the reaction + 2+ _ Cu Cu + e (4) 135. o then e is given by eo = +0.17 - (-0.07) = 0.24v (5) Ethyl dixanthogen is almost insoluble, so its activity may be taken as unity. Finally, when reaction (1) attains equilibrium the e.m.f. is zero. Substituting -4 these data, the value z = 2 and (aEtX -) = 10 into equation (2) Cu ) 2+ 2 E X - = 10 4 (6) aCu This is the condition for redox equilibrium; it is independent of any other equilibrium condition. The equilibrium with respect to cupric ethyl xanthate and cuprous ethyl xanthate precipitates is related to the solubility products as follows:- __ -24.2 S1 = (aCu 2+).(a_ PAX -) 2 = 10 (7)(Du Rietz97) and -19.3 98 32 = (aCu+)*(aEtX-) = 10 (8)(Kakovsky ) The critical condition, determining which precipitate is formed, is given by the combined conditions (7) and (8), i.e. when: (ac.-0 (9) (acu2+) si -4 which, for (aEtx-) = 10 is (aCu+) 10-4) (1o) (aCu2+) 8 (aEtx- When the ratio is less than this valuet solubility equilibrium corresponds to (7) being satisfied but not (8), i.e. cupric ethyl xanthate is precipitated. Thus, 136. in view of condition (6) overall equilibrium (i.e. redox equilibirum plus solubility equilibrium) is satisfied by a precipitate of cupric ethyl xanthate, and not cuprous ethyl xanthate, when the ethyl xanthate concentration is 10-4M. When the xanthate ion concentration is 2.5 x 10-5M the modified conditions represented by equations (6) and (10) are the same. This is then the critical xanthate concentration at which both precipitates are in equilibrium, and below this concentration the cuprous salt is the equilibrium product, However, chrysocolla-xanthate reactions may not occur under conditions of redox equilibrium, and at the reaction sites, the concentration of EtX has been shown in all instances to be well below the 10-4M concentration in the bulk solution. Furthermore, the data on which the transition xanthate concentration is based may be subject to large errors, which have not been assessed here. The effect of xanthate concentration at pH 5.5 on the extent of the chrysocolla-xanthate reaction has been shown in Figure 17. The pause in the general increase in reaction with xanthate concentration is considered to result from a changeover from a reaction occurring largely in solution, - to one occurring largely within the surface pores. The transition between these two sites of reaction has been shown to be near pH 7 at -4 [EtX ] = 10 M (section 8a): thus at pH 5.5 [EtX] = -4 10 M, the reaction would be expected to be in the bulk solution such that a poorly-adherent shell of precipitate surrounds the external mineral surface. Increasing the xanthate concentration at pH 5.5 will 137. result in precipitation occurring closer to the external mineral surface, or even inside the surface pores, - and the reaction rate will increL.se with xanthate concentration. The rate will not be directly proportional to concentration in fixed time experiments owing to the considerable reaction (and hence pore blockage) that occurs at this pH. -When the bulk solution xanthate concentration is reduced below approximately 10-4M, the reaction occurs more in the bulk solution, where the rate of diffusion is almost certainly faster than in the pores. The rate would be expected to rapidly increase with increasing xanthate concentration, - and then diminish to a slight extent as the reaction occurred more and more at the external surface, since this would tend to slow down diffusion. The data of Figure 17 show that, at pH 5.5, the reaction is entirely in solution below [EtX] = 0.25 x 4 10 M; and entirely in the pores above [EtX] = 2 x -4 10 H. The enhanced reaction apparent with amyl xanthate is primarily the result of the lower solubility product of the cupric amyl xanthate (Du Rietz97 CuEtX2, K sp = 10-24.2; CuAmX = 10-27.2 2'sp ). Although Figure 33 shows that virtually all the Cu2+ will be extracted from chrysocolla below pH 9.5 in [EtX-] = 10-4M or pH 11.0 in [AmX] = 10-4M, the fact that the xanthate concentration falls rapidly below 10-4M at the point of 2+ reaction, will mean that complete Cu extraction will not occur; a greater reaction will then occur with the amyl homologue, since this is precipitated at a lower [C1124]. More precipitates will thus be formed per unit thickness of outer mineral shell. Any reaction which occurs in the bulk solution at any pH and [XC] = 10-4m 138. will again favour the amyl xanthate because of the lower Ksp of CuAmX2. The observed distribution of chrysocolla-xanthate reaction products at various pests is in agreement with the previously-formulated hypothesis on product disposition. Products formed at the external surface of the mineral apparently possess little tenacity, and thus easily report in the bulk solution even when scouring action is mild. Pore reaction products are mechanically held and occluded; thus they will not be detected by the analytical method used here. It is not clear why the distribution of products should be approximately constant at various pH's. The low tenacity of the surface films may arise for reasons other than low mechanical strength of the film. The surface charge data of section 7j (v) suggest that chrysocolla has a negative charge when the solution also has its pH is greater than. 6, and that EtX2 isoelectric point in this region. If CuEtX also possesses a weak negative charge in this pH range, then a situation exists for the mutual dispersion of the three species quite apart from the dispersion caused by attrition. 0. Sulphide reactions. The process by which soluble sulphide reacts with chrysocolla is strictly analogous to the xanthate- chrysocolla reaction, - as will be any reaction, between chrysocolla and a precipitant for copper. Since only the alkaline pH range has been considered in the case of sulphrde reactions, it may be concluded 139. that only the mechanism of pore reactions (illustrated in Figure 34c) is possible, where the concentration of precipitating ion is of the order of 10-4M. According to the two dissociation constants obtained for hydrogen sulphide by Jellinek and Czerwinski99 , viz; K1= 10-7, -15 2 2 x 10 , it may be'calculated that at pH 9.0, 10-4 Na -4 4 2S.9H20 E l0 HS, and at 31-1 8, lo Na S.9H -4 2 20 m .91 x 10 1vi HS. The increased rate of sulphidation reactions over xanthate reactions is a result of a number of factors. Cupric sulphide has a very low solubility product -44.1\100 (10 and will thus be precipitated at very low [Cu2+]: in fact Figure 33 shows that all the copper will be leached from chrysocolla at all pH's in the presence of sodium sulphide. More precipitate will thus be formed per unit thickness of outer mineral shell. Furthermore, the molar volume of CuEtX and EtX2 may well be considerably greater than for CuS (which has a_ specific gravity of 4.6). It would be expected, therefore, that pore blockage would be less effective, and hence the rate of diffusion higher for the sulphidation process. Quite independent of pore blockage, it has been shown in section 7j(vii) that the NaHS diffuses at a rate 1.75 times that of the KEtX. :For comparable concentration gradients, this would lead to 1.75 times the reaction rate. The diffusing species HS which predominates in the alkaline region, reacts 2+ with Cu as follows: 2+ HS + Cu CuS + + Thus, one half of the H necessary for the exchange 2+ reaction which brings Cu in solution, is carried in 140 with the hydrosulphide ion. It follows that the amount of OH generated will only be one-half of that in a comparable xanthate case. This will lead to a further factor of 2 increase in the reaction rate of sulphide compared with that of xanthate. It is seen, therefore, that considerably more rapid reaction will occur with sulphide, than with xanthate. A quantitative factor of at least 1.75 x 2 = 3.50 can be seen from the foregoing argument: this, together with the qualitative effects of less pore + blockage and a greater ion exchange of H for Cu could readily account for the observed factor of 10. The inflections noticed with changes in xanthate concentration in Figure 17, are not apparent with sodium sulphide in Figure 26. At the pH of the sulphide experiment, the mechanism of Figure 34c applies, and over the entire concentration range explored, the reaction was occurring within the surface pores and to a small extent at the external surface. The rate will thus be dependent upon HS concentration,- though not directly proportional to it because of the considerable pore blockage resulting from the extensive reaction at this pH. The distribution of CuS which is similar to the distribution of CuEtX and EtX 2 in xanthate reactions, clearly is caused by the same processes. Again, the low tenacity may be caused, in fact, by the similarity of surface charge of CuS and chrysocolla at pH's> 6. The electron microscopy of sulphided chrysocolla has been previously described in section 5d. An examination of the plates in Figure 8 shows CuS lining 141 the pores: it appears darker than the background because it absorbs the electronsto a greater extent than the silica matrix. There is some indicat,Lon- that more CuS is present in. the outer layers pores than is present deeper in, - but this may be a confusion between. general particle opacity towards the particle centre. The encrusting layer shown in the electron micrographs must not be thought to be filmed CuS: it is, in fact, a layer of adsorbed, amorphous carbon which forms on all specimens in the electron. microscope, owing to traces of hydrocarbons in the atmosphere of the instrument. d. Flotation. quiescent flotation. Whore quiescent flotation occurs in xanthate- chrysocolla systems, it is the consequence of an adherent film of cuprous xanthate and dixanthogen. to the external surface of the chrysocolla.. The dixanthogon is apparently co-adsorbed with the CuEtX rather than at the chrysocolla surface. Flotation cannot be anticipated in cases where the hydrophobic species are formad in the bulk solution, scoured from the external surface, or occluded within the mineral pores. Under quiescent conditions, therefore, a fixed amount of CuEtX-EtX 2 reaction would be expected to produce no flotation at pHK4, good flotation at pH 4 - 6, and worse, but essentially constant, flotation at higher pH's, where the reaction follows the previously expounded pore mechanism. This confirms the observations of Figure 21. Even in quiescent flotation, approximately 100 times 142. the quantity of CuEtX and EtX (and hence, the 2 consumption of xanthate) are required for the flotation of chrysocolla compared with the flotation of natural sulphides. This is principally the result of the distribution of hydrophobic species between the mineral surface, the pores, and the bulk solution, which alone accounts for a factor of 10-20 times. Contributory factors in this high xanthate consumption are the far greater solubility of chrysocolla over the sulphides (even when the latter are partially oxidized), and the inherent hydrophilicity of the silica surface 2+ remaining after Cu extraction from chrysocolla which requires a greater coverage of hydrophobic species to offset this effect. The greater effectiveness of amyl xanthate compared with ethyl xanthate is considered to be for two reasons. Firstly, its lower solubility product will result in a more extensive precipitate in the surface shell, and thus contributP§: more to external surface hydrophobicity, even though the loss of CuAmX and AmX 2 to the bulk solution is probably very similar to the ethyl xanthate situation. Secondly, the amyl hydrocarbon chain is more than 40% longer (12.2 g and 8.4 R respectively, Klassen and Mokrousov101)than the ethyl chain: it is therefore effective over a greater area of external surface. The explanation of the inability of ethyl xanthate to enhance the floatability of sulphided chrysocolla at alkaline pH, even after extensive oxidation and in the absence of free sulphide, apparently lies in the site of the reaction. After 2+ sulphidation insufficient Cu lies either at the 1L13. external surface, or in the surface pores, and the only site for substantial reaction is deeper in the mineral. Because of pore blockage with CuS, this deep-seated xanthate reaction may be a slow process (undetectable in the experiments here) and, in any event, totally ineffective for flotation. It is inconceivable that the strong aqueous oxidants used in an attempt to 2+ produce Cu at the surface did not, in fact, oxidise 2+ the surface CuS: the Cu produced must have been insufficiently associated with the external surface and thus produced cuprous xanthate uselessly in the bulk solution. The effect of the heat treatment described in section 7i in successfully allowing an ethyl xanthate reaction, seems likely to have been in squeezing out CuS from the pores. This cupric sulphide is hydrophobic to a certain extent, but its partial removal from the pores may have allowed more 2+ Cu to be present in the surface oores for xanthate reaction, as well as blocking the pores to a lesser extent. The increased effectiveness of amyl xanthate on sulphided chrysocolla again may lie in the larger precipitation of CuAmX in the surface layers beneath the CuS, and the length of its hydrocarbon chain. (ii) turbulent flotation. In general, when conditioning and flotation was carried out under the more violent conditions of the 500g Denver flotation cell, good flotation of chrysocolla was not possible. Where the conditions of reagent concentration, external surface area, etc. were similar to those under which successful flotation 144.. was possible, the reason for failure is due to the removal from the external surface of more hydrophobic species thanwasnormally removed in quiescent conditions. Depending upon the degree of turbulence, this may simply be scoured CuEtX and EtX2 or additionally, the structurally-weakened copper- depleted silica of the surface layers. In the flotation of chrysocolla ores, additional factors are involved. If liberation of the chrysocolla values is effected by grinding in water at neutral pH's, much of the chrysocolla will be selectively slimed because it is relatively soft. When xanthate, or sodium sulphide, - or both, - is added to even an alkaline ground pulp, much reagent will be consumed in reacting with extremely fine chrysocolla particlesi of large external surface area per unit weight. With xanthate alone, the drop in effective 4 solution concentration from approximately 10 M (about 0.25 lb/ton in a Denver 500g laboratory machine) to substantially less than this, will result in the situation apparent at low pH and concentrations in Figure 17. Judging by the result of flotation tests, such a situation also apparently results at initial [EtX] of even as high as 3 x 10-3M. At higher pH's, it can be assumed that no such bulk solution reaction occurs. Where sodium sulphide is added, a further loss of reagent occurs both during conditioning and flotation (especially this latter) owing to the displacement of hydrogen sulphide gas from the system. This accounts for the very high soluble sulphide additions that must be made in order to achieve any reasonable flotation 145. effect with xanthate. Both the xanthate, and the sulphide-xanthate systems lose reagent, of course, to other ions that may be precipitated from solution. The optimum conditioning period apparent in Figure 29 is considered to be due to the following. Theoretically only when all the HS has disappeared 2+ from the pulp can any Cu - EtX reaction occur, - although a local depletion of HS in the mineral pores would induce a xanthate reaction. With the concentration of sulphide used, this apparently takes some 3 minutes. Further conditioning removes the hydrophobic species, and therefore reduces floatability, (iii) comparison with malachite. Because malachite reacts with xanthate and soluble sulphide faster than does chrysocolla, in a system in which both occurred, chrysocolla would be expected to form any hydrophobic coating slower. The better tenacity of surface films on malachite no doubt results from its good crystallinity, end lack of porous structure. Good crystallinity will provide a good anchorage at the surface for massive films, whilst the absence of porosity will mean that above a critical pH, -- for a certain precipitating ion concentration, - precipitation will occur at the surface: with chrysocolla, precipitation would then substantially occur within the pores. (iv) extension to other specimens of chysocolla. Chrysocolla of different Cu content and of different porosity would not be expected to follow exactly the data obtained here for the Kolwezi sample. Flotation, however, can be assumed less likely when the 146. copper content is low or when the diffusion of Cu?* from the mineral is slow because of narrow pores. On the other hand, flotation prospects would be expected to be better where the chrysocolla had a high copper content, where the copper was easily leached, or where the internal surface area was low. (v) improvement in chrysocolla flotation practice. Flotation of chrysocolla from ores with xanthates alone is unlikely to be an economic feasibility. Even xanthateflotation of sulphided chrysocolla is of doubtful practicality owing to the large amount of sodium sulphide required. However, where chrysocolla occurs in malachite ores, better recoveries will be achieved if sliming of chrysocolla during grinding, and the attrition of reacted chrysocolla is minimised. In particular, the latter criteria involves less turbulent flotation machines (e.g. air lift machines) operating at lower pulp densities with the minimum of conditioning time and the most rapid flotation rate. It is also necessary to ensure that sodium sulphide is not wasted by aeration or by allowing sulphidation to proceed in the absence of xanthate, - the formed CuS will only scour away. If these suggestions are coupled with rapid rate d such as is found at neutral or weakly acid pH's, optimum conditions for the practical flotation of chrysocolla with sulphidation-xanthate techniques will be more nearly approached. A more economical process for chrysocolla flotation 147. may lie in the use of nen-po1,7r reagents (such as fuel oil) in conjunction with xanthates. Exploratory testwork on quiescent flotation with a paraffin free of surface active agents (obtained by percolation through an activated alumina column), has shown that the flotation of chrysocolla is considerably enhanced by the use of a xanthate flotation following a paraffin activation. Only 5% of the amyl xanthate requirement is necessary for 100% Hallimond tube flotation when xanthate-reacted chrysocolla is subsequently conditioned in a 200 mgpl paraffin emulsion for 2 minutes at neutral pH's. 148. 9. ACKNOWLEDGEMENTS. The investigations described in this thesis were undertaken in the Bessemer Laboratory of the Royal School of Mines, London, SW 7, under the general direction of the Professor of Mineral Technology, Dr. M.G. Fleming. For their continued interest, encouragement and guidance, the author is very much indebted to many of the staff and postgraduate students of the Mineral Technology Department. In this connexion, his gratitude is particularly extended to Dr. A.P. Prosser, whose supervision of the project has proved invaluable. The financial assistance afforded by the Imperial College of Science and Technology, the Institution of Mining and Metallurgy, and the Bosworth Smith Trust Fund is also gratefully acknowledged. Ore and mineral samples were supplied without charge for the project by Union Miniere du Haut - Katanga, SA, Elisabethville, Congo. Assistance from other Imperial College departments, and from the DSIR Warren Spring Laboratory in obtaining certain data involving specialised determinative equipment and techniques has been acknowledged in the text. Such assistance has been very much appreciated. 149. 10. REFERENCES. (Abbreviations used here for the titles of journals, etc., are those authorised by the American Chemical Society, and published in Chemical Abstracts in 1961 and 1962). 1. ARBITER, N.: Air flotation of silica-bearing oxidised copper ore: US Pat. 2829770: 1958. 2. BOWDISF, F.W.; CHEN, T.P.: Studies in the flotation of chrysocolla; (1) Sulphidisation; (2) Activated sulphidisation: Trans. AIME.: 1963; 226, (1), p.21. 3. JAEKEL, J.A.: New guides to chrysocolla flotation: Mining World: 1959; 21 (8), p.44. 4. KOVACS, C.F.: Surface treatment of minerals with gases; chrysocolla and water vapour: Stanford Univ. Publ.; Dep. Miner, Engng. Prog. Rep.: 1961; 61, (2), p.35. 5. DEAN, R.S.: Flotation of oxidized copper minerals: US. Bur. Mines. Rep. Invest.: 1934; 3357. 6. LUDT, R.W.; DEWITT, C.C.: Flotation of copper silicates from silica: Mining Engng: 1949; 1, (2), p.49. 7. HEY, N.H.: An index of mineral species and varieties arranged chemically: British Museum Trustees, London: 1950. 8. NEUMANN, H.; SVERSDRUP, T.; SAEBO, P.C.: X-ray powder patterns for mineral identification. (3) Silicates: Avhandl, Norske Videnskaps- Akad. Oslo, (1)Mat. Naturv. Kl.: 1957; 6, p.1. 9. POLJAK, R.J.; GORDILLO, C.E.: Shattuckite from Del Alto de las Lecheras, Aicuna: Bol. Acad. Nac. Cienc. Cordoba, Rep. Arg.: 1957; 40, (la), p.97. 10. DANA, E.S.: The system of mineralogy of James Dwight Dana, 1837-1868; descriptive mineralogy: John Wiley and Sons, New York: 1909. 11. ANON.: X-ray powder data file: ASTM Stan., Philadelphia: 1960; 48-1. 150. 12. BILLIET, V.: Investigation on the connexion between chrysocolla, katangite, plancheite, bisbeite, shattuckite and dioptase: Verhandel. Koninkl. Vlaam. Acad. Uetenschap. Belg. Kl. Wetenschap.: 1942; 4, (1), p.l. 13. BELOV, N.V.; BUTOZOV, V.P.; GOLOVASTII7.0V, N.I.: Crystal structure of dioptase: Dokl. Akad. Nauk SSSR.: 1952; 87, p.953. 14. HEIDE, H.G.: The crystal structure of dioptase, Cu6(Si6018), 6H20: Naturwissenschaffen: 1954; 41, p.402. 15. MOENKE, H.: Infra-red absorption spectrophotometry and silicate investigation: Silikat Tech.: 1961; 12, p.323. 16. IVANOVA, V.P.: Thermal diagrams of minerals; Zap. Vses. Mineralog. Obshchestva.: 1961; 90, p.50. 17. TOUSSAINT, J.: A thermal study (DTA) of natural hydrated copper silicates: Ann. Soc. Geol. Belg., Bull.: 1957; 80, (B), p.287. 18. GUILLEMIN, C.; PIERROT, R.: New data on plancheite. The identity of plancheite and shattuckite: Bull. Soc. Franc. Mineral. Crist.: 1961: 84, p.276. 19. MILLMAN, A.P.: X-ray data of copper silicates: Private comlaunication: 1961. 20. SUMIN, N.G. ; LASHEVA, N.K.: New modifications of chrysocolla of the plancheite-type from Mednorudyansk in the Ural: Tr. Mineralog. Muzeya. Akad. Nauk SSSR: 1951; 3, p.106. 21. FORD, W.E.: A testbook of mineralogy by E.S. Dana: John Wiley and Sons, New York: 1945. 22. SUN, H-S.: Differential thermal analysis of shattuckite: Am. Mineralogist: 1961; 46, p.67. 23. SEMMONS, W.: Notes on some silicates of copper with remarks on the chrysocolla group: Mineral Mag.: 1878; 11, p.197. 151. 24. ROGERS, A.F.: A review of the amorphous minerals; J. Geol.: 1917; 25, p.515. 25. BELOV, N.Y.; MOLCHANOV, V.S.; N.E.: Synthesis anei structure of hydrosilicates containing simple and complex heavy metal cations: Tr. Pyatogo Soveshch. po Eksperim. i Tekhn. Mineralog. i Petrogr. Akad. Nauk SSSR, Inst. Khim. Silikatov: 1956; p.38, 26. SHCHERBINA, V.V.; IGNATOVA, L.I.: New data on the geochemistry of copper in the supergene mineralisation zone: Zap. Vses. Mineralog. Obshchestva: 1955; 84, p.353. 27. EDWARDS, A.B.: Textures of the ore minerals and their significance: Australian Institute of Mining and Metallurgy, Melbourne: 1960. 28. IiINCHELL, A.N.; WINCHELL, H.: Elements of optical mineralogy. (2) Description of minerals: John Uiley and Sons, New York: 1951. 29. READ, H.H.: Rutley's elements of mineralogy: Thomas hurby, London: 1960. 30. FOOTE, H.W.; BRADLEY, W.M.: On solid solution in minerals. (4) The composition of amorphous minerals as illustrated by chrysocolla: Am.J. Sci.: 1916; 36, p.180. 31. PALMER, C.H.: Chrysocolla, a remarkable case of hydration: Am. J. Sci.: 1903; 16, p.45. 32. LARSEN, E.S.; BERMAN, H.: Microscopic determination of non-opaque minerals: US. Gov. Printing Office, Washington: 1934. 33. YAKHONTOVA, L.K.: Copper-rich variety of chrysocolla: Vestnik. Moskov. Univ: 7, (6), Ser. Fiz. Mat. i Estestven. Nauk; 1952; 4, p.123. 34. BUTTS, A.: Copper: the science and technology of the metal, its alloys and compounds: Reinhold Publishing Corporation, New York: 1954. 35. CHUKHROV, F.V.; ANOSOV, F.Y.: The nature of chrysocolla: Zap. Vses. Mineralog. Obshchestva: 1950; 79, (2), p.127. 152. 36. CHUKHROV, F.V.; ANOSOV, F.V.: Medmontite, a copper-bearing montmorillonite mineral: Zap. Vses. Mineralog. Obshchestva: 1950; 79, (2), p.23. 37. SUN, M-S.: The nature of chrysocolla from Inspiration Mine, Arizona; Am. Mineralogist: 1963; 48, p.649. 38. MCLAUGHLIN, R.J.W.: The differential thermal analysis of clays: Mineralogical Society, London: 1957. 39. KAUFFMAN, A.J.; DILLING, E.D.; DTA curves of certain minerals: Econ. Geol.: 1950; 45, p.222. 40. STOKES, R.H.; ROBINSON, R.A.: Standard solutions for humidity control at 25°C: Ind. Eng. Chem.: 1949; 41, p.2013. 41. MCBAIN, J.W.; BAKR, A.M.: A new sorption balance: J. Am. Chem. Soc.: 1926; 48, p.690. 42. ANON: Dictionary of colour standards: British Colour Council, London: 1934. 43. AZAROFF, L.V.; BEURGER, M.J.: The powder method in X-ray crystallography: McGraw-Hill Book Company, New York: 1958. 44. DE BOER, J.H.: The shape of capillaries: Proc. Symp, Colston Res. Soc.: 1958; p.68. 45, BUSWELL, A.M.; DUDENBOSTEL, B.F.: Spectroscopic studies of base exchange materials: J. Am. Chem. Soc.: 1941; 63, p.2554. 46. HUNT, J.M. ; WISHERD, M.P. ; BONI-LA.1\1, L.C. : Infrared absorption spectra of minerals and other inorganic compounds: Anal. Chem.: 1950; 22, (12), p.1478. 47. JIRGENSONS, B.; STR\UMANIS, M.E.: A short textbook of colloid chemistry: Pergamon Press, C)x iord:- 1962. 48. MERING, J.: On the hydration of montmorillonite: Trans. Faraday Soc.: 1946; 42, (B), p.205. 153. 49. MOONEY, R.W.; KEENAN, A.G.; WOOD, L.A.: Adsorption of water vapour by montmorillonite. (2) Effect of exchangeable ions and lattice swelling: J. Am. Chem. Soc.: 1952; 74, p.1371. 50. BROOKS, C.S.: Free energies of immersion for clay minerals in water, ethanol, and n-heptane: J. Phys. Chem.: 1960; 64, p.532. 51. BARRER, R.M.; REAY, J.S.S.: Interlamellar sorption by montmorillonite: Proc. Intern. Congr. Surface Activity, 2nd, London: 1957; 2, p.79. 52. BARRER, P.M.; BRATT, G.C.: Non-stoichiometric hydrates (1) Sorption equilibria and kinetics of water loss for ion-exchanged near-faujasites: Phys. Chem. Solids: 1959; 12, p.130. 53. VAN BEMMELEN, J.M.: Absorption: T. Steinkopf, Dresden: 1910. 54. KRUYT, H.R.: Colloid Science, 2: Elsevier Publishing Company, Amsterdam: 1949, 55. BARRER, R.M.; MCLEOD, D.M.: Intercalation and sorption by montmorillonite: Trans. Faraday Soc.: 1954; 50, p.980. 56. BARRER, R.M.: Structure and properties of porous materials: Proc. Symp. Colston Res. Soc.: 1958; p.6. 57. BACHMANN, W.: Vapour pressure isotherms of substances with gel structure; 7 . Anorg. Allgem. Chem.: 1917; 100, p.l. 58. ANON: Kolwezi flotation plant treats three types of copper-cobalt bearing ores: World Mining: 1957; 10, (2), p.46. 59. O'MEARA, A.E.: A mineralogical approach to some Copperbelt metallurgical problems: Tech. Proc. Commonwealth Mining Met. Oongr., 7th, Kitwe; 1962; p.333. 60. DERJAGUIN, B.V.; SHUKAKIDSE, N.D.: . Dependence of the floatability of antimonite on the value of zeta potential; Trans. Inst. Mining Met.: 1961; 70, p.569. 154. 61. COOK, M.JL.; NIXON, J.C.: Theory of water repellent films on solids formed by adsorption from aqueous solutions of heteropolar compounds: J. Phys. and Colloid Chem.: 1950; 54, p.445. 62. HAGIHARA, H.; UCHIKOSH, H.; YANASHITA, H.: Adsorption and chemical reactions of sulphydric collectors as revealed by electron diffraction: Proc. Intern. Congr. Surface Activity, 2nd, London: 1957; 3, p.343. 63. DEWITT, C.C.; MAKENS, R.F.; HELZ, A.W.: The surface relations of the xanthates: J. Am. Chem. Soc.: 1935; 57, P.796. 64. GAUDIN, A.M.; PEELLER, G.S.: Surface area of flotation concentrates and the thickness of collector coatings: Trans. AIME: 1946; 169, p.248. 65. GAUDIN, A.M.; DEWEY, F.; DUNCAN, W.E.; JOHNSON, R.A.; TANGEL, 0.F.: Reaction of xanthates with sulphide minerals: Trans. AIME: 1934; 112, p.319. 66. GAUDIN, A.M.; SCHUHMANN, R.: The action of potassium n-amyl xanthate on chalcocite: J. Phys. Chem.: 1936; 40, p.257. 67. LEJA, J.; LITTLE, L.H.; POLING, G.W.: Xanthate adsorption studies using infrared spectroscopy. Oxidised and sulphidised copper substrates 2 Evaporated lead sulphide, galena and metallic lead substrates: Trans. Inst. Mining Met.: 1963; 72, p.407 68. PLAKSIN, I.N.; TYURNIKOVA, V.I.; TRETYAKOV, 0.V.; Disposition of xanthates at the surfaces of sulphide minerals: Izvest. Akad. Nauk SSSR, Otdel. Tekh. Nauk: 1958; 7, p.146. 69. PLAKSIN, I.N.; ANFIMOVA, E.A.: An investigation of several problems on the interaction of xanthate with the surfaces of sulphide minerals of copper and zinc under flotation conditions: Izvest. Akad. Nauk SSSR, Otdel. Tekh. Nauk: 1954; 5, p.95. 155. 70. RAO, S.R.; PATEL, C.C.: Adsorption of xanthates at chalcopyrite surfaces, and the effect of oxygen and carbon dioxide on the adsorbed xanthates: J. Sci. Ind. Res. (India): 1961; 20, (D), p.432. 71. SHORSHER, I.N.; Flotation of cerussite and malachite; Tr. Vses. Nauchn. Issled. i Proekt. Inst. Mekhan. Obrabotki Polezn. Iskop: 1957; 104, p.l. 72. TAGGART, A.F.; TAYLOR, T.C.; KNOLL, A.F.: Chemical reactions in flotation: Trans. AIME: 1930; La, p.217. 73. TAGGART, A.F.; DEL GIUDICE, G.R.M.; ZIEHL, 0.A.: The case for a chemical theory of flotation: Trans. AIME: 1934; 112, p.348. 74. WARK, I.A.: Principles of flotation: Australian Institute of Mining and Metallurgy, Melbourne; 1938. 75. REHBINDER, P.; LIPETZ, M.; RIMSKAYA, M.: Physical chemistry of wetting phenomena and flotation processes. (10) The dependence of wettability on the adsorption of flotation reagents (wetting isotherms at the interface mineral- water-air); Kolloidn. Zh.: 1934; 66, p.40. 76. ADAMS, C.R.; VOGE, H.H.: Ageing of Si02-L),03 cracking catalyst. (2) Electron microscope studies: J. Phys. Chem.: 1957; 61, p.722. 77. MITROFANOV, S.I.: Solution of some problems concerning the theory and practice of selective flotation in the USSR: Prog. Mineral Dressing, Trans. Intern. Mineral Dressing Congr., Stockholm; 1957; p.441. 78. MITROFANOV, S.I.; STRIGIN, I.A.; KUSIINIKOVA, V.G.; ROZHAVSTJI, G.S.; Kinetics or formation of sulphide films on the surface of oxidised minerals of heavy metals in flotation: Kolloidn. Zh.: 1955; 17, p.235. 156. 79. KUSHNIKOVA, V.G.; STRIGIN, I.A.; ROZHAVSKII, G.S.: Sulphidation reaction of oxidised minerals: Sb. Nauchn. Tr. Gos. Nauchn.-Issled. Inst. Tsvetn. Metal.: 1955;, 10, p.7. 80. STRIGIN, I.A.; KUSHNIKOVA, V.G.: Effect of ammonium sulphate and aluminium sulphate on the rate of sulphidation of malachite and chrysocolla, and on adsorption of collector on them: Sb. Nauchn. Tr. Gos. Nauchn.- Issled. Inst. Tsvetn. Metal: 1955; 10, p.30. 81. BAYULA, A.G.: Flotation of silicates and oxidised ores: USSR Pat. 75026: 1958. 82. BAUTISTA, R.G.; SOLLENBERGER, C.L.: Conversion of metallic oxide mineral surfaces to sulphides: Eng. Mining J.: 1962; 163, (11), p.81. 83. GUTSALYUK, T.G.; SOKOLOV, M.A.; KORABLINA, M.P.: Flotation of chrysocolla: Izv. Akad. Nauk Kaz. SSR, Ser. Met., Obegashch. i Ogneuporovt 1961; 1, p.3. 84. LIVINGOOD, M.D.: Flotation of copper silicates by selected akiyl-substituted polyhydroxynitroso- phenols: Univ. Microfilms: 1952. 85. MITZMAGER, A.; GAILIS, J.E.: Flotation of chrysocolla: Bull. Res. Council Israel: 1961; 10, A. 86, TYURENKOVA, G.N.: Flotation of oxidised carbonate and silicate copper ores: USSR Pat. 127962: 1960. 87. SUBBA RAU, M.G.: Studies in the adsorption of xanthate on galena: Ph.D. thesis, University of London: 1961. 88. HAGIHARA, H.: Study of the principles of flotation (3): Kobayashi Rigaku Kenkyushu Hokoku: 1954; 4, p.30. 89. IWASAKI, I.; COOKE, S.R.B.: The decomposition of xanthates in acid solution: J. Am Chem. Soc: 1958; 80, (1), p.285. 90. KING, C.V.; DUBLON, E.: Rate of decomposition of xanthic acid: J. Am. Chem. Soc.: 1932; 54, p.2177. 157. 91. RAO, S.R.; PATEL, C.C.: Kinetics of decomposition of xanthates in the presence of copper salt: J. Sci. Ind. Res. (India): 1961; 20, (D), p.299. 92. RAO, S.R.; PATEL, Kinetics of decomposition of xanthates: J. Mines, Metals Fuels: 1960; 7, p.9. 93. BOYD, G.E.; MYERS, L.S.; ADAMSON, A.W.: The exchange adsorption of ions from aqueous solution by organic zeolites. (2) Kinetics: J. Am. Chem. Soc.: 1947; 69, (2), p.2836. 94. ANON: Stability constants of metal-ion complexes, with solubility products of inorganic substances. (1) Organic ligands: The Chemical Society, London: 1957. 95. DU RIETZ, C.: Fatty acids in flotation: Prog. Mineral Dressing, Trans. Intern. Mineral Dressing Congr. Stockholm: 1957; p.417. 96. GELOSO, M.; DESCHAMPS, P.: Insoluble basic salts and the calculation of the solubility product constant: Compt. Rend.: 1947; 225, p.742. 97. DU RIETZ, C.: Chemical problems in flotation of sulphide ores: IVA: 1953; 24, p.257. 98. KAKOVSKY, I.A.: Physicochemical properties of certain organic flotation reagents and of their salts with ions of non-ferrous metals: Tr. Inst. Gorn. Dela, Akad. Nauk SSSR: 1955; 3, p.255. 99. JELLINEK, K; CZERWINSKI, J.: Dissolution of hydrogen sulphides sodium sulphide and sodium hydrogen sulphide in aqueous solutions: Z. Physik. Chem. (Leipzig): 1922; 102, p.438. 100. ANON: Handbook of chemistry and physics: Chemical Rubber Publishing Company, Cleveland: 1962. 101. KLASSEN, V.I.; HOKROUSOV, V.A.: An introduction to the theory of flotation: Butterworths and Company, London: 1963. 158. 11. CONTENTS. Page 1. SUMMARY • • • •• • •• • •• • •• • •• • 2 2 . INTRODUCTION 400 400 000 .00 P00 4 3 . NOMENCLATURE ... 060 4400 000 4.0 5 4. REVIEW OF CHRYSOCOLLA hINERALOGY a. Formation and occurrence ... 4.. 9 b. Constitution • • • • • • • • 9 c. Optical, physical and chemical properties . • • • • • • • 10 d. Structure . • • . * 4 • • • 10 5. RESULTS AND DISCUSSION OF CHARACThRISATION OF CHRYSOCOLLA SAMPLE a. Preparation of sample for testwork 16 b. Constitution ... ••• • • • •• • 17 c. Optical, physical and chemical properties • 26 d. Structure • •• ••• ••• . • P 30 e. Chrysocolla as a rnontrnorillonite member .., 004 • • • • • • 42 6. REVIEW OF FLOTATION OF COPPER ORES a. Processes 0•• ••• • • • • •• 48 b. General theory of flotation • • • 49 c. Theory of xanthate collection • •• 51 d. Theory of sulphidation-xanthate collection 040 64. 00. 53 e. Previous chrysocolla investigations 53 159. Page 7. CHRYSOCOLLA-XANTHATE, CHRYSOCOLLA-SOLUBLE SULPHIDE SYSTEMS a. Investigational method...... • • • 60 b. Preparation of samples for testwork . • • 61 c. Nature of the systems • • • • • • • • • 62 d. Factors affecting chrysocolla-xanthate reactions (i) the reaction vessel ... •• . 64 (ii) experimental procedure 004 000 65 (iii) analytical procedure 0.. ..e 67 (iv) effect of pH see ...... 70 (v) kinetics ...... 72 (vi) effect of concentration ... • • • 72 (vii) comparison of ethyl and amyl xanthate 000 040 . . 4 78 e. Distribution of chrysocolla-xanthate reaction products ...... 78 f. Flotation of xanthate-reacted chrysocolla (i) flotation cells 400 040 04. 84 (ii) experimental procedure ...... 87 (iii) flotation - total xanthate reaction relationship ... 90 g • Factors affecting chrysocolla-sulphide reactions (i) the reaction vessel • • • •• . 92 (ii) experimental procedure • • • •• • 92 (iii) analytical procedure • • • •• • 94 (iv) effect of pH 94 160. Page (v) kinetics • • • • • . •• • •• • 94 (vi) effect of concentration 000 98 h. Distribution of chrysocolla-sulphide reaction products 000 6O0 000 98 Flotation of sulphide-reacted chrysocolla . • • • • • • • • • • • 103 j. Mechanism experiments (i) leaching ...... 105 (ii) stability of CuEtX at alkaline pH 115 (iii) ethyl dixanthogen 000 600 118 (iv) hydrophobicity of co-precipitated CuEtX and EtX2 000 • 00 118 (v) surface charges ... .000 004 118 (vi) malachite • • • • • • • • • • • • 119 (vii) diffusivity of KEtX and NaHS 122 (viii) effect of [e] on xanthate reaction 6•0 ,6•0 0440 122 (ix) stability of CuS in the presence of 10-4M EtX- 06A 122 8. DISCUSSION ON CHRYSOCOLLA FLOTATION SYSTEMS a, Leaching of Cu2+ •• • .• • •• • •• • 125 b. Xanthate reactions ...... 000 126 c. Sulphide reactions 000 000 60* 138 d. Flotation (i) quiescent flotation 460 4104 141 (ii) turbulent flotation 0.4 ... 143 (iii) comparison with malachite 4041. 145 161. Page (iv) extension to other chrysocollas 145 (v) improvements in chrysocolla flotation practice ... • • • 146 9 . ACKNOWLEDGEMENTS • • • • • • • • • • • • • • • 148 10. REFERENCES 000 000 000 .o0 000 POO 149 11. CONTENTS . • • • • • • • • •• . •• • •. • 158 12. LIST OF FIGURES AND TABLES a. Figures •• • •.• ! • • • • i • • • 162 164 b . Tables ... • 4 • • • • • • • ••• ••• 162 12. LIST OF FIGURES AND TABLES Page a. Figures. 1. Specimens of the Kolwezi chrysocolla •• • 18 2. Thermogravimetric analysis of Kolwezi chrysocolla lee 400 ••• ••• • • • 20 3. The sorption balance • • • •• • • • * ... 22 4. Water sorption isotherm of Kolwezi chrysocolla 23 5. X-ray diffraction pattern of Kolwezi chrysocolla •• • • • • •• • ••• • • . 32 6. Infn›-red absorption spectrum of Kolwezi chrysocolla •• • ••• • • • • • • • •• 34 7. Differential thermal analysis of Kolwezi chrysocolla S.0 • • • • • • • • • 9 .•. 35 8. Electron micrographs of sulphided Kolwezi chrysocolla • • • • • • • • • • • • . • • 37 9. Examples of relevant pore structures • • • • • • 39 10. Infra-red absorption spectrum of chrysocolla reacted with potassium ethyl xanthates ... 63 11. The reaction vessel ... • • • • • • • • • • • • 66 4 12. Decomposition of 10 M KEtX solutions in acid conditions • . • • • • . • • • • • • . • 69 13. Effect of pH on ethyl xanthate reaction 0.* 71 14. Appearance of xanthate-reacted chrysocolla at various pH's after 30 minutes . • • • • • 73 15. Effect of aillonia on ethyl xanthate reaction at pH 10 ... • • • • • • • • •• • • • • 75 16. Kinetics of ethyl xanthate reaction • • • •• . 76 17. Effect of xanthate concentration on ethyl xanthate reaction . • • • • • • • • ... 79 163. Page 18. Colorimetric determination of Cu2+ with sodium diethyldithiocarbarnate • •• 82 19. Absorptiometric determination of EtX9 in iso-octane • • • • • . • • • • • • • .. 85 20. 30 ml agitated froth flotation cell ... 88 21. Correlation between xanthate reaction and quiescent floatability 4.. 000 91 22. Turbulent floatability of chrysocolla with ethyl xanthate ... • • . . . • • • • • • • 93 23. Absorptiometric determination of Na2S.9H20 000 95 24. Effect of pH on sodium sulphide reaction 400 96 25. Kinetics of sodium sulphide reaction ... ••• 97 26. Effect of sulphide concentration on sodium sulphide reaction *06 000 000 000 100 + 27. Polarographic determination of Cu'-' in 0.1M ,...... 060 000, • .. HNO3 101 28. Distribution of CuS in turbulent chrysocolla- sulphide reactions • • • • • • • • • • • . 102 29. Effect of length of conditioning period on flotation of Kolwezi ore OP. 900 30. Acid leaching of Kolwezi, cl--ysocolla at different pH's and agitation ••• 31. Testing particle or film diffusion control in acid leaching 000 000 • • • 32. Scouring of surface from chrysocolla and malachite • •• •• 0 • •• ••• 400 123 2+1 33. Equilibrium [Cue j v. pH diagram for Kolwezi chrysocolla .. • • OP • •• •• • 000 127 34. Ionic concentrations of xanthate reacting chrysocolla at various pH ... O4.0 006 129 164. Page b. Tables. 1. Principal distinguishing properties of dioptase and plancheite 00$ .00 0.0 6 2. Some examples of published optical and physical properties of chrysocolla 3. X-ray powder diffraction data on samples of chrysocolla • • • • • • • • • • • • ••• 13 4. Reported differential thermal analyses of chrysocolla A.* ••• 000 0.4 • • • 15 - 5. Constitution of Kolwezi chrysocolla • • • • • • 25 6. Specific gravity of Kolwezi chrysocolla 0.0 28 7. X-ray diffraction patternibr Kolwezi chrysocolla • • • ••• ••• • • • •• • 33 8. Consumption of sulphuric acid during CuEtX f ormation ... • • • • • • • • • . • • SO. 74 9. Values of 'a' in the expression r = at1/2 0.0 77 10. Comparison of ethyl and amyl xanthate reactions with chrysocolla • • • • • 80 11. Estimation of CuEtX in reaction solution •• • 83 12. Distribution of chrysocolla-ethyl xanthate reaction products .. 0.0 004 40. 86 l/2 13. Values of 'b' in the expression r = bt ... 99 14. Flotation of sulphided chrysocolla with amyl xanthate ... 000 Of. 0.0 000 VO. 104 15. Experimental conditions of 500g flotation test on Union Miniere ore .. 090 ... • 004 106 [cu24] . XH 16. Values of K = r +12 at various pH's 1 L HJ X ..Cu and XH's. . • • • • • • • • • ... 116 165. Page 17. Stability of CuEtX at alkaline pH ... *00 117 18. Relative reaction rates of malachite and chrysocolla with KEtX and Na2S.9H20 004 120 19. Relative floatability of malachite and chrysocolla • • • • • • • • • • . • • • • 121