Analysis of activated carbon, as used in the carbon-in-pulp process, for gold and eight other constituents Gerda E.E. Balaes, Kenneth Dixon, Gerda M. Russell, and George J. Wall National Institute for Metallurgy, Randburg

Methods involving atomic-absorption spectrophotometry (AAS), X-ray fluorescence (XRF) spectrometry, and the use of a direct-reading spectrometer - optical emission spectrometry using inductively coupled plasma (OES-ICP), are considered for the determination of nine constituents (silver, gold, copper, cobalt, nickel, iron, zinc, calcium, and silicon) that are adsorbed onto activated carbon during the carbon-in-pulp (CIP) process. Analyses of three reference samples are reported, and the statistical significance of the mean values are evaluated in relation to the relative standard deviations of the method. Limits of determination and times of analysis are compared, and it is concluded that OES-ICP and XRF offer the best means for the multi-element analysis. However, if the analysis of gold alone is required, the times of analysis and results for all three methods are comparable. S Air J Chem., 1982 3 5 , 4 — 8 Metodes is ondersoek vir die bepaling van nege elemente (silwer, goud. koper, kobalt, nikkel, yster, sink, kalsium, en silikon) wat tydens die koolstof-in-pulpproses (KIP) op geaktiveerde koolstof geadsorbeer word. Die metodes behels atoomabsorpsiespektrofotometrie (AAS), X-straalfluoressensiespektrometrie (XSF), en direkaflesende optieseemissiespektrometrie wat van induktief-gekoppelde plasma (OES-IKP) gebruik maak. Drie verwysingsmonsters is ontleed en die statistiese betekenis van die gemiddelde waardes is met betrekking tot die relatiewe standaardafwykings van die metodegeevalueer. Die bepalingsgrense en ontledingstye word vergelyk en die gevolgtrekking word gemaak dat OES-IKP en XSF die beste metodes vir die ontleding van die nege elemente is Die ontledingstye en resultate van al drie metodes vergelyk goed met mekaar as daar net 'n goudontleding verlang word. S -Air.

Tydskr

Chem

, 1982, 35, 4—8

G.E.E. Balaes, K. Dixon,* G.M. Russell, and GJ. Wall National Institute for Metallurgy, Private Bag X3015, Randburg 2125, Republic of South Africa *To whom correspondence should be addressed Received 25 May 1981; revised 26 August 1981

The development of the carbon-in-pulp (CIP) process for the recovery of residual amounts of gold from cyanidation pulps at the National Institute for Metallurgy (NIM), has received considerable attention for several years. Because of the need to achieve a greater understanding of the process and to determine the possible effects of toxicity in relation to the adsorption of gold, there is also interest in analysing for other elements adsorbed on carbon. These include iron, copper, silver, nickel, silicon, calcium, cobalt, and zinc. It was expected that iron, copper, and silver would be present in concentrations up to several hundred p.p.m., nickel in concentrations up to several thousand p.p.m., silicon and calcium in concentrations between 0,1 and 2%, and gold in concentrations of up to 10 000 p.p.m. The probable concentrations of cobalt and zinc could not be predicted. Some analyses of activated carbon were reported by Davidson, 1 without details of the method used. Kirmura 2 adsorbed trace elements onto activated carbon in the presence of xanthate, and recovered at least some of them by first digesting the carbon with concentrated nitric acid and then evaporating to dryness, after which the trace constituents were taken into solution with nitric acid (150 g/1). Tests showed that nickel, silver, and copper could be recovered quantitatively in this way, and measured by atomic-absorption spectrophotometry (AAS). Extraction with aqua regia enabled the gold to be recovered. The ideal approach would be direct analysis of the carbon, but this is not possible at present owing to certain problems. For example, the carbon could be pelletized and the emission spectra excited by spark excitation. Such a procedure has been tried, but lacks sensitivity. The preparation of synthetic standards is difficult, and the degree of excitation is dependent on whether the constituents are added to the carbon as chlorides or cyanides. Excitation of the solids by flame atomization is not generally applicable and, whereas solid material is the normal medium for analysis by X-ray fluorescence spectrometry (XRFS), initial tests showed that, when discs were prepared from carbon, mixed with alumina as the diluent and boric acid as the binder, the resulting pellets were not of sufficient strength. The high capacity of the carbon for absorbing moisture led to the disintegration of the pellets after less than a week, and would have necessitated the continuous preparation of standards. In view of these considerations, it was decided that the

S.Afr. . Chem., 1982, 35 (1) Table 1 Instrumental parameters for atomic-absorption spectrophotometry Diluting medium

Wavelength/ nm

Slit width/ nm

Flame type

Background correction

409b HC1 HNO, 10% HCI and 5% HNO, and 1% La

C2H2—air CJHI-air N2O2-C2H2 C2H2—air

Fe Ni

10% HNOJ 10% HNO, and 1% K N 0 3 10% HNO } and 1% K N 0 3

324,8 248,3 232,0

0,5 0,5 0,5 0,2 0,5

Yes No

10% HNOj and 1% KNO,

328,1 242.8 422,7 240,7

Si Zn

10% HNO, and 1% KNO, 10% H N 0 3 and 1% K N 0 3

251,6

Element Ag Au Ca Co Cu

213,9

carbon should be ashed and the ash dissolved by fusion. The resulting solution could then be analysed by AAS or by optical emission spectrometry (OES) with excitation by inductively coupled plasma (ICP), which would simplify calibrations. Analysis by XRFS could be carried out on pellets prepared from the ash. Experimental Samples, as received, were ground to pass a 200-mesh sieve, during which process they readily absorbed moisture. The usual range of moisture content was 2—10%, although it has been reported that up to 25% moisture can be absorbed. Ground samples were therefore dried prior to analysis, through overnight heating at 100 °C, after which they were cooled and transferred to a desiccator before aliquots were weighed out. Measurement by AAS3 Dried carbon (1 g) was ashed in a porcelain crucible for 3 h at 450 °C, transferred to a zirconium crucible, and fused with sodium peroxide (I g). The fused sample was leached with 6M-hydrochlonc acid (10 ml) and concentrated nitric acid (2 ml), and the volume was adjusted to 50 ml. The operating parameters are given in Table 1. Calibration was effected by standards matched with respect only to the acid medium. Measurement by OES with excitation by ICP* Dried carbon (1 g) was ashed in a zirconium crucible for 2 h at 450 °C. The resulting ash was fused with sodium peroxide (2 g) and extracted with aqueous 50% nitric acid (v/v), after which nitric and hydrochloric acids were added. Before the solution was made up to volume (200 ml), scandium was added to serve as an internal standard, the final scandium concentration being 50 p.p.m. The solution was then excited in an ICP torch, and the line intensities were measured on a direct-reading spectrometer. The operating parameters and spectral lines used for measurement are given in Tables 2 and 3. Calibration was, as for AAS, effected by the use of synthetic standards matched with respect only to sodium and the acid medium. Measurement by XRFS' The ash from carbon (3 g). prepared by heating on a bed of alumina (2 g) in a silica dish (10 cm diam.) for 30 min at 450 °C, was transferred to an agate Siebtechnik bowl together with boric acid (2 g) and additional alumina to give a total mass of 6,5 g. The mixture was ground and mixed for 120 s, and then pelletized for 2 min at a pressure of 66 Χ 10 3 kPa. On irradiation of the pellet by Λ'-rays from a tube with a chromium target, the emitted fluorescence was measured, together with the background, at suitable 2Θ angles. Count rates were corrected for background directly after the application of a background factor. This factor converts the background value at a given 2Θ angle to that under the peak measured, to give a net peak intensity