QUANTITATIVE ANALYSIS OF GOLD IN LOW-GRADE TAILINGS FROM 

DIFFERENT MATRICES, COUPLED WITH A STUDY INTO THE ASSOCIATED 

UNCERTAINTIES 

 

 

 

 

 

 

 

 

 

 

 

 

Kedibone Nicholine Mashale 

Supervisor: Prof Luke Chimuka 

Co-supervisors: Dr James Tshilongo 

                   : Dr James Sehata 

 

 

A thesis submitted to the Faculty of Science, University of the Witwatersrand, in fulfilment 

of the requirements for the degree of Doctor of Philosophy. 

 

 

November 2023 

 

 

 



i 
 

DECLARATION 

I declare that this research is my own, unaided work. It is being submitted for the Degree of 

Doctor of Philosophy at the University of the Witwatersrand, Johannesburg. It has never been 

submitted before for any degree or examination at any other university. This research was 

conducted at Mintek, Analytical Chemistry Division. 

 

 

 

 

 

--------------------------- --------------------------------- 

Kedibone Nicholine Mashale 

 

 

This 23th day of November 2023 in Johannesburg 

 

 

 

Supervisor Name: Prof Luke Chimuka 

Co-supervisors: Dr. James Tshilongo and Dr. James Sehata 

 

 

 

 

 



ii 
 

PUBLICATIONS AND CONFERENCES 

1. Statistical evaluation of the uncertainties in the characterization of South African low-grade 

gold mine tailings 

Kedibone Mashale, Bambesiwe May, James Sehata, James Tshilongo and Luke Chimuka. 

Article published under the South African Journal of Chemistry (2023); 77(1), 42-47. 

https://doi.org/10.17159/0379-4350/2023/v77a07 

2. Analysis of low-grade gold from mine tailings following gold deportment analysis, size 

fractionation, and acid digestion by ICP‒MS and the associated uncertainties 

Kedibone Mashale, Napo Ntsasa, James Sehata, James Tshilongo and Luke Chimuka. 

(Manuscript) 

3. The effectiveness of Fe and Sn in the gravimetric quantification of gold from low-grade 

tailings, supported by evaluation of measurement uncertainties 

Kedibone Mashale, Napo Ntsasa, James Sehata, James Tshilongo and Luke Chimuka. 

(Manuscript) 

Other work 

1. Detection and separation of silver ion from industrial wastewaters using fluorescent D-

glucose carbon nanosheet and quaternary silver indium zinc sulphide quantum dots 

Bambesiwe M. May, Olayemi J. Fakayode, Kedibone N. Mashale, Mokae F. Bambo, Ajay 

K. Mishra, Edward N Nxumalo 

This article was published in the Journal of Water Processing Engineering (2022); 49, 102944 

https://doi.org/10.1016/j.jwpe.2022.102944 

 

Conferences 

K.N. Mashale, J Sehata, J Tshilongo and L. Chimuka, Gravimetric quantification of gold in 

low-grade mine tailings, Test and Measurement, 24-26 October 2022, Persequor Park, Pretoria. 

(Oral). https://www.nla.org.za/?page_id=4191 

 

K.N. Mashale, J Sehata, J Tshilongo and L Chimuka, Gravimetric quantification of gold in 

low-grade mine tailings, 44th SACI National Convention. 08-13 January 2023, Stellenbosch, 

South Africa. (Oral). https://www.saci.co.za/SACI2023 

 

  

https://doi.org/10.17159/0379-4350/2023/v77a07
https://doi.org/10.1016/j.jwpe.2022.102944
https://www.nla.org.za/?page_id=4191
https://www.saci.co.za/SACI2023


iii 
 

ABSTRACT 

Gold is one of the precious group elements that is used for various purposes, such as jewellery, 

auto catalysts and as a form of investment. Various countries have gold reserves, with South 

Africa being the leading gold producer between 1980 and 2007. However, as of 2022, it is 

ranked as the eighth largest producer of gold, contributing 3% to the global contribution. The 

majority of gold is mainly mined from the Witwatersrand Basin in Johannesburg. It is well 

known that mining has been ongoing for decades, which means that a significant amount of 

land has been mined across the country. During gold mining, a large proportion of the ore 

material from which the gold is extracted is waste, together with the chemicals that were used, 

and this waste is termed mine tailings. This implies that based on the years that gold mining 

has occurred for and the depth of mining, a significant amount of the tailings have been 

deposited into free land around the mines, some of which are close to communities. 

The tailings consist of traces of gold that were left due to inefficient extraction processes and 

other components, such as base metals. The disadvantage of this is that due to the other 

chemical composition of these tailings, they have the potential to be dangerous to the 

environment. Some tailings contain minerals such as jarosite (KFe2(SO4)2(OH)6) that cause 

acid mine drainage, while heavy metals such as lead, mercury, arsenic and chromium can leach 

into surface and ground waters, causing pollution. Furthermore, they pose a danger if the dams 

that they are stored in collapse, which was recently witnessed in South Africa. Because of these 

factors, there have been various advances made towards the beneficiation of tailings, such as 

utilizing them to make glass or bricks for construction. A major advancement was the 

reprocessing of these mine tailings to recover or extract the remaining gold, which benefits 

both the environment and the mining houses. Therefore, in a move to support this initiative, 

scientists have taken to the laboratory to develop new or optimize existing methods for the 

extraction and quantification of gold, which is expected to be of a low grade over time. 

Various methods can be used for the quantification of gold, including the conventional fire 

assay, wet and dry chlorination and acid digestion. Most of these are suitable for medium- to 

high-grade gold ores but are known to experience challenges in regard to low-grade ores. The 

aim of this research was therefore to find the optimum method for the quantification of gold 

from mine tailings emanating from the Ventersdorp Contact Reef (VCR) and Barberton 

Greenstone Belt (GBS). Subsequent to chemical analysis, the samples were characterized for 

mineralogy using X-ray diffraction (XRD) and Brunauer‒Emmett‒Teller (BET) surface area 



iv 
 

analysis. The mineralogical characterization of the material showed high quartz content (68-

78%), muscovite (11.4-13.8%), chlorite (6.9-7.6%), pyrite, dolomite and calcite as minor 

minerals. The BET analysis reported a surface area of 2.3 m2.g-1 for the feed sample. The 

tailings were treated with acid digestion employing reverse aqua regia, which was both in the 

presence and absence of hydrofluoric acid (HF). This process was preceded by a gold 

deposition study to determine the association of gold with minerals such as quartz and pyrite. 

In the fire assay, the variables varied were sample mass (50, 100 and 120 g), the ratio of litharge 

to carbon (PbO:C, 8, 10 and 12) and the enhancement of the lead collection by adding tin (Sn) 

and iron (Fe). These methods were developed for the tailings sample (0.1 g.t-1), which was a 

residue of a 0.32 g.t-1 sample from a cyanide leaching process. Furthermore, it was investigated 

whether the lead started with at the beginning of the fusion was recovered, as its optimum 

recovery would indicate that the gold is being efficiently recovered. In an attempt to curb issues 

relating to sample homogeneity due to sample size, the use of commercial sodium hypochlorite 

(NaOCl) for the chlorination of gold in the ore sample (0.32 g.t-1), in which the concentration 

of NaOCl, pH and leaching time were varied, was investigated. To assess the performance of 

the acid digestion and fire assay methods, measurement uncertainty was evaluated using the 

bottom-up approach, which focused on the major contributors. 

The recoveries obtained from the methods were 50-79% for aqua regia, 81-83% for reverse 

aqua regia, 63-83% for aqua regia with HF and 81-111% for reverse aqua regia with HF, which 

fit the criteria set. Reverse aqua regia offered the highest recoveries, irrespective of the addition 

of HF, which implied that the process did not suffer from any gold encapsulation by minerals 

such as quartz. This is further supported by the gold depot test, in which 78% of the gold was 

free milled. This method yielded limits of detection (LOD) of 0.14-0.334 µg.kg-1, which was 

impressive because there was no use of harsh acids and the sample mass of 1 g did not result 

in any homogeneity problems. In the assessment of the homogeneity, a homogeneity indicator 

of 8.9% was obtained, which indicated that the sample, at 1 g, was satisfactorily homogeneous. 

From the fire assay, a sample mass of 100 g, flux ratio of 10 and addition of 10 mg. L-1 Sn, 

obtained the highest recoveries greater than 80%. It was also proven that there is a correlation 

between the lead recovery after fusion and the recoveries of gold. Although a large mass of 

300 g can be used for wet chlorination, this method encounters various issues. The results of 

these tests showed that at a pH of 4, 1.5 M NaOCl and 2 h extraction time, recoveries as high 

as 97% can be obtained. Furthermore, analysis of the initial sample and the residues indicated 

a change in the mineral composition, with minerals such as kaolinite appearing in the 



v 
 

diffractograms due to increased crystallization, while the surface area of the residues increased 

from 3.3 to 5.2 m2.g-1, showing that the oxidation process had occurred. 

To statistically support the methods, evaluation of measurement uncertainty was carried out in 

which the bottom-up approach, which included quantifying the overall method uncertainty by 

assessing all the parameters involved, was used. The measurement uncertainty evaluations 

showed that the fire assay with a gravimetric finish is the best method, as it has values of ± 

0.00432 g.t-1, while that of acid digestion with inductively coupled plasma‒mass spectrometry 

(ICP‒MS) finish and fire assay with inductively coupled plasma optical emission spectroscopy 

(ICP‒OES) finish had measurement uncertainties of 0.1 and 0.09 g.t-1, respectively. This brings 

the expressions of the measurand as follows: 

- Fire assay by gravimetry: 0.0904 ± 0.00432 g.t-1 

- Fire assay by ICP‒OES: 0.26 ± 0.1 g.t-1 

- Acid digestion by ICP‒MS: 0.258 ± 0.09 g.t-1 

This shows the importance of quantifying uncertainties in the methods for gold quantification, 

especially those that have major drawbacks. It was therefore concluded that although all three 

methods gave satisfactory results, a fire assay with the addition of Sn and a gravimetric finish 

would be best suited to quantify gold in these low-grade tailings. 

The advantages of the methods presented in this study were the use of commercial bleach 

(NaOCl), the use of a small sample mass (1 g), the use of a simple acid combination of HCl 

and HNO3 and the coupling of these to an uncertainty study. In terms of chlorination, most 

studies have used it on high-grade ores of at least 10 g.t-1, obtaining recoveries of less than 

80%, while in this study, the feed sample was 0.32 g.t-1 and offered as much recoveries. To 

curb the use of instrumental analysis, the gravimetric method using Sn and Fe showed good 

recoveries, which again, is an enormous achievement. In general, the superiority that these 

methods have over other studies is that they worked satisfactorily on low-grade samples, 

whereas in the literature, they were applied on high-grade ores. 

 

 

 

 

  



vi 
 

DEDICATION 

 

 

 

 

To my mother; Elizabeth Mbiza, late father; Ishmael Mashale and grandmother, Ellah 

Mbuyane. 

 

 

 

To my daughter, Mokgethwa Duduzile. 



vii 
 

ACKNOWLEDGEMENTS 

I would like to acknowledge my supervisors, Prof Luke Chimuka, Dr James Sehata and Dr 

James Tshilongo, for their guidance in completing this Ph.D. 

To Dr James Tshilongo, Ms. Nompumelelo Leshabane, and Mr. Napo Ntsasa, thank you for 

all the words of encouragement that came my way. Thank you for motivating me to be the best 

that I can and for always being available. Your efforts are very much appreciated. 

Thank you to Ms. Hlengiwe Mnculwane, Dr. Odwa Mapazi, and Dr. Bambesiwe May for the 

ongoing support and for the outings that always relieved my stress. 

A massive appreciation to my friends, Mr Basil Munjanja, Ms Portia Makhubela, Ms Amanda 

Mahlangu, Ms Leah Malapane, and Ms Patricia Mawela, for always encouraging me to push 

through. 

NRF and Mintek are acknowledged for funding my doctoral studies. 

Thank you to the ACD team at Mintek for assisting me when I needed help. Especially the fire 

assay team, thank you so much, and to everyone who helped in any way. 

I would also like to acknowledge Dr Sabine Verryn of XRD Consulting, who gave me 

discounts on my XRD analysis. 

Last, I offer massive appreciation to my family for their support, Elizabeth Mbiza, Diemetse 

Mashale, Clement Nake, Luyanda Mkhonto, Kamohelo Mashale, Accolade Sehlabela, and my 

precious daughter, Mokgethwa Duduzile. 

  



viii 
 

CONTENTS 

DECLARATION ........................................................................................................................ i 

ABSTRACT ............................................................................................................................. iii 

ACKNOWLEDGEMENTS ..................................................................................................... vii 

LIST OF FIGURES ............................................................................................................... xiii 

LIST OF TABLES ................................................................................................................... xv 

LIST OF SYMBOLS ............................................................................................................. xvii 

ABBREVIATIONS ............................................................................................................. xviii 

1 INTRODUCTION .............................................................................................................. 1 

1.1 Background ................................................................................................................. 1 

1.2 Gold tailings economics in South Africa .................................................................... 4 

1.3 Problem statement ....................................................................................................... 7 

1.4 Aim .............................................................................................................................. 7 

1.5 Objectives .................................................................................................................... 7 

1.6 Justification of study ................................................................................................... 8 

1.7 Novelty of the study .................................................................................................... 8 

1.8 Structure of thesis ........................................................................................................ 9 

1.9 References ...................................................................................................................... 10 

2 Literature review ............................................................................................................... 12 

2.1 Literature review on gold .......................................................................................... 12 

2.1.1 Gold.................................................................................................................... 12 

2.2 Literature review on tailings ..................................................................................... 13 

2.2.1 Type of tailings .................................................................................................. 15 

2.2.2 Storage of tailings .............................................................................................. 16 

2.3 Sampling methods ..................................................................................................... 16 

2.4 Sample size ................................................................................................................ 17 

2.5 Characterization techniques ...................................................................................... 18 



ix 
 

2.5.1 Mineralogical composition ................................................................................ 19 

2.5.2 Elemental composition....................................................................................... 19 

2.6 Determination of gold in high- and low-grade samples, including mine tailings ..... 20 

2.6.1 Fire assay ................................................................................................................. 20 

2.6.2 Chlorination ............................................................................................................. 24 

2.6.3 Acid dissolution/digestion ....................................................................................... 25 

2.6.4 Solvent extraction .................................................................................................... 26 

2.6.5 Adsorption ............................................................................................................... 26 

2.7 Quantitative determination of gold ................................................................................ 29 

2.7.1 Instrumental analysis ............................................................................................... 29 

2.7.2 Gravimetric analysis ................................................................................................ 29 

2.7 References ................................................................................................................. 32 

3 STATISTICAL ANALYSIS ............................................................................................ 40 

3.1 Basic statistical evaluations ....................................................................................... 40 

3.2 Paired t test and Analysis of Variance (ANOVA) .................................................... 40 

3.2.1 Paired t test ......................................................................................................... 40 

3.2.2 ANOVA ............................................................................................................. 41 

3.3 Measurement uncertainty .......................................................................................... 42 

3.3.1 Approaches to determining measurement uncertainty ....................................... 42 

Monte Carlo Method (MCM) ........................................................................................... 44 

3.4 Quantification of measurement uncertainty .............................................................. 45 

3.4.1 Bottom-up approach........................................................................................... 45 

3.4.2 Top-Down approach .......................................................................................... 47 

3.5 Evaluation of measurement uncertainty literature .................................................... 49 

3.6 References ................................................................................................................. 51 

4 Analysis of gold from tailings AFTER gold deportment, size fractionation and acid 

digestion by ICP‒MS and the associated uncertainties ........................................................... 53 



x 
 

ABSTRACT ............................................................................................................................. 53 

4.1 Introduction ............................................................................................................... 54 

4.2 Methods and methods................................................................................................ 55 

4.2.1 Materials ............................................................................................................ 55 

4.2.2 Particle size distribution, homogeneity, and purity analysis .............................. 56 

4.2.3 Mineralogical analysis ....................................................................................... 56 

4.2.4 Elemental composition....................................................................................... 57 

4.2.5 Determination of the head grade ........................................................................ 57 

4.2.6 Gold depot study ................................................................................................ 57 

4.2.7 Acid digestion .................................................................................................... 58 

4.2.8 Instrumental analysis ......................................................................................... 58 

4.2.9 Statistical evaluations......................................................................................... 60 

4.3 Results and discussion ............................................................................................... 61 

4.3.1 Instrumental analysis ......................................................................................... 61 

4.3.2 Mineralogical composition ................................................................................ 62 

4.3.3 Purity analysis and homogeneity ....................................................................... 65 

4.3.4 Gold deportment ................................................................................................ 68 

4.3.5 Size-based fire assay .......................................................................................... 69 

4.3.6 Acid digestion .................................................................................................... 70 

4.3.7 Statistical evaluations......................................................................................... 74 

4.4 Conclusion ................................................................................................................. 77 

4.5 References ................................................................................................................. 79 

5 The effectiveness of iron and tin in the gravimetric quantification of gold from low-grade 

tailings ...................................................................................................................................... 82 

Abstract ................................................................................................................................ 82 

5.1 Introduction ............................................................................................................... 83 

5.2 Materials and methods .............................................................................................. 85 



xi 
 

5.2.1 Materials ............................................................................................................ 85 

5.2.2 Sample information ............................................................................................ 85 

5.2.3 Cyanide leaching for higher grades ................................................................... 86 

5.2.4 Fire assay ........................................................................................................... 87 

5.2.5 Instrumental analysis ......................................................................................... 89 

5.3 Results and Discussion .............................................................................................. 90 

5.3.1 Instrumental analysis ......................................................................................... 90 

5.3.2 Cyanide leaching ................................................................................................ 90 

5.3.3 Effect of sample mass and flux .......................................................................... 93 

5.3.4 Effect of co-collectors ........................................................................................ 95 

5.3.5 Recovery of lead ................................................................................................ 97 

5.4 Statistical evaluations. ............................................................................................... 99 

5.4.1 ANOVA ............................................................................................................. 99 

5.4.2 Measurement Uncertainty ................................................................................ 100 

5.5 Conclusion ............................................................................................................... 101 

5.6 References ............................................................................................................... 103 

6 Low-grade gold extraction in tailings using A chlorination approach ........................... 106 

Abstract .................................................................................................................................. 106 

6.1 Introduction ............................................................................................................. 107 

6.2 Materials and methods ............................................................................................ 108 

6.2.1 Materials .......................................................................................................... 108 

6.2.2 Characterization ............................................................................................... 109 

6.2.3 Chlorination ..................................................................................................... 109 

6.2.4 Time-integrated chlorination ........................................................................... 110 

6.3 Results and discussion ............................................................................................. 111 

6.3.1 Sample composition. ........................................................................................ 111 

6.3.2 Change in the mineralogy ................................................................................ 111 



xii 
 

6.3.3 Effect of time on recovery ............................................................................... 112 

6.3.4 Effect of pH on recovery.................................................................................. 114 

6.3.5 Surface area and morphology variation ........................................................... 115 

6.4 Conclusion ............................................................................................................... 118 

6.5 References ............................................................................................................... 120 

7 Conclusion and recommendations .................................................................................. 122 

8 Appendix: first page of publications .............................................................................. 124 

 

 

  



xiii 
 

LIST OF FIGURES 

Figure 1.1 Uses of gold .............................................................................................................. 1 

Figure 1.2 Satellite image of gold dumps in South Africa. ....................................................... 3 

Figure 1.3 Tailings dam in West Rand in 2005 ......................................................................... 5 

Figure 1.4 Tailings dam in West Rand in 2016 ......................................................................... 6 

Figure 1.5 Graphical representation of the amount of gold expected to be produced per year at 

the Elikhulu and BRTP treatment facilities.. ............................................................................. 6 

Figure 2.1 Map of the Witwatersrand basin............................................................................. 12 

Figure 2.2 Map of the Ventersdorp Contact Reef. ................................................................... 13 

Figure 2.3 Schematic diagram of the fire assay process and its various finishes .................... 21 

Figure 2.4 Diagram of cupellation during the fire assay ......................................................... 22 

Figure 2.5 Phase diagram of the Na2O-SiO2 (mass-%) system ............................................... 23 

Figure 2.6 Process of dry chlorination ..................................................................................... 25 

Figure 2.7 Process of wet chlorination. The reaction vessel consists of the sample slurry and 

NaOCl and HCl ........................................................................................................................ 25 

Figure 3.1 GUM framework operation. ................................................................................... 43 

Figure 3.2 Propagation of distributions by the Monte Carlo method ...................................... 45 

Figure 4.1 External calibration curves for Au on the ICP‒OES (a) and ICP‒MS (b). ............ 62 

Figure 4.2 Mineralogical data of the fractions screened by XRD (n=3). ................................ 64 

Figure 4.3 Diffractograms of the screened fractions. .............................................................. 64 

Figure 4.4 Gold recoveries of the screened fractions (n=3). ................................................... 70 

Figure 4.5 Gold recoveries obtained from the four experiments (n=3). .................................. 72 

Figure 5.1 Pouring of the fused sample into iron molds. ......................................................... 88 

Figure 5.2 External calibration curve for gold on the AAS at a wavelength of 242.80 nm. ... 90 

Figure 5.3 Recovery of gold in the residue and filtrate from cyanide leaching (n=3). ............ 92 

Figure 5.4 The tailings materials used for the study. ............................................................... 93 

Figure 5.5 Effect of the mass on the recovery of gold from the tailings (n=3) ....................... 95 

Figure 5.6 Recovery of gold in the VCR matrix after the addition of Sn and Fe during fusion 

(n=3). ........................................................................................................................................ 96 

Figure 5.7 Recovery of gold in the GSB matrix after the addition of Sn and Fe during fusion 

(n=3). ........................................................................................................................................ 96 

Figure 5.8 Comparison of lead recovery against gold recovery in CRMs and random samples 

after fusion (n=3). .................................................................................................................... 98 



xiv 
 

Figure 5.9 Comparison of lead recovery against gold in the tailings sample (n=3). ............... 99 

Figure 6.1 Comparison of the mineralogical composition of the head sample and the 

chlorinated residues for GSB. ................................................................................................ 112 

Figure 6.2 Comparison of the mineralogical composition of the head sample and the 

chlorinated residues for VCR................................................................................................. 112 

Figure 6.3 Relationship between extraction time and gold recovery at pH 12 (a) and pH 4 (b).

................................................................................................................................................ 113 

Figure 6.4 Relationship between the concentration of the oxidant and gold recovery at pH 12 

(a) and pH (4). ........................................................................................................................ 114 

Figure 6.5 Surface area plot comparing the head sample and the chlorinated sample at different 

concentrations. ....................................................................................................................... 116 

Figure 6.6 SEM images of the samples before chlorination. ................................................. 117 

Figure 6.7 SEM images of samples after chlorination. .......................................................... 118 

  



xv 
 

LIST OF TABLES 

Table 1.1 Distribution of mine dumps in the Witwatersrand Basin .......................................... 4 

Table 1.2 Various gold tailings treatment plants currently running in South Africa. ................ 5 

Table 2.1 South African tailings characterization from the literature. .................................... 14 

Table 2.2 Types of tailings emanating from different mineral processing techniques. ........... 15 

Table 2.3 Methods commonly employed in the decomposition of gold in various ore bodies.

.................................................................................................................................................. 27 

Table 2.4 Various instrumental analysis techniques for the quantification of gold in 

geological matrices. ................................................................................................................. 30 

Table 3.1 Divisors for various distributions of data for Type B measurement uncertainty. ... 44 

Table 3.2 Example of a measurement uncertainty budget table. ............................................. 47 

Table 3.3 Evaluations of measurement uncertainty from different matrices in different fields.

.................................................................................................................................................. 50 

Table 4.1 ICP‒MS operating conditions .................................................................................. 59 

Table 4.2 Operating conditions for the ICP‒OES. .................................................................. 60 

Table 4.3 Criteria set for the calibration of the ICP‒OES and ICP‒MS. ................................ 60 

Table 4.4 Calibration errors for ICP‒OES and ICP‒MS ......................................................... 62 

Table 4.5 Mineralogical information of the original tailings sample and the certified reference 

material in %wt. ....................................................................................................................... 63 

Table 4.6 Elemental composition (%) of the screened fractions by ICP‒OES (n=3).............. 65 

Table 4.7 Purity analysis of the flux for alkaline fusion by ICP‒OES (n=5) in %. ................. 66 

Table 4.8 Purity analysis of the flux for the fire assay by ICP‒OES (n=5) in %. ................... 66 

Table 4.9 Homogeneity indicators for base metals and gold at uncertainties of 5 and 25%. .. 67 

Table 4.10 Homogeneity evaluation employing ANOVA. ..................................................... 68 

Table 4.11 Distribution of gold in various minerals by gold deposition. ................................ 69 

Table 4.12 Gold recovery results for the certified reference materials determined by ICP‒MS 

(n=3). ........................................................................................................................................ 72 

Table 4.13 Comparison of proposed methods to published methods for the quantification of 

gold from geological samples, including tailings. ................................................................... 74 

Table 4.14 Quantification of the uncertainties from various parameters from acid digestion 

and fire assay with ICP‒MS finish. ......................................................................................... 77 

Table 5.1 Experimental parameters for cyanide leaching. ....................................................... 87 

Table 5.2 Components of the three fluxes employed. ............................................................. 88 



xvi 
 

Table 5.3 Criteria for calibration of AAS for gold analysis. ................................................... 89 

Table 5.4 Calibration information showing the calculated concentrations and calibration 

errors. ....................................................................................................................................... 90 

Table 5.5 Mass balance from the cyanide leaching process. ................................................... 92 

Table 5.6 Guide on sample mass depending on the grade. ...................................................... 94 

Table 5.7 Recovery of Au from the tailings sample and the CRM, AMIS 610. ...................... 96 

Table 5.8 ANOVA at 95% confidence interval, as obtained from Microsoft Excel. ............ 100 

Table 5.9 ANOVA with the decision rule. ............................................................................ 100 

Table 6.1 Experimental parameters for the chlorination processes. ...................................... 110 

Table 6.2 Concentrations of the main ions in chlorination. ................................................... 110 

Table 6.3 Chemical compositions and content of minerals. .................................................. 111 

  



xvii 
 

LIST OF SYMBOLS 

mg.L-1 
Milligram per litre 

ng.L-1 Nanogram per litre 

µg.L-1 Microgram per litre 

kg Kilogram 

g.L-1 Gram per litre 

kg.t-1 Kilogram per litre 

g.t-1 Gram per tonne 

cm Centimetre 

km Kilometre 

g Gram 

mg Milligram 

°C Degrees Celsius 

% Percent 

µg.kg-1        Microgram per kilogram 

mg.kg-1       Milligram per kilogram 

m2.g-1           Square metres per gram 

  



xviii 
 

ABBREVIATIONS 

AAS 
Atomic absorption spectroscopy 

ANOVA Analysis of variance 

AR Aqua regia 

BET Brunauer‒Emmett‒Teller 

BIC Bushveld igneous complex 

BTRP Barberton tailings retreatment plant 

CIL Carbon in leach 

CRM Certified reference material 

CV Correlation of variance 

DIBK Diisobutyl ketone 

DLLME Dispersive liquid‒liquid microextraction 

EDXRF Energy dispersive X-ray fluorescence 

FA Fire assay 

FACT Fast automated curve fitting 

GSB Greenstone belt 

GUM Guide to uncertainty measurement 

HF Hydroflouric acid 

HRCS-GFAAS High resolution continuum source GF-AAS 

ICP‒MS Induced couple plasma mass spectrometry 

ICP‒OES 

Induced couple plasma optical emission 

spectroscopy 

INAA Instrumental neuron activated analysis 

ISO International standardization organization 

KED Kinetic energy discrimination 

LA-ICP‒MS Lase ablation coupled to ICP‒MS 

LOD Limit of detection 

LOQ Limit of quantification 

MCM Monte carlo method 

MIBK Methyl isobutyl ketone 

MU Measurement uncertainty 

Pb-FA Lead fire assay 

PGE Precious group elements 



xix 
 

PGM Platinum group metals 

PSD Particle size distribution 

PT Proficiency testing 

PTFE Polytetrafluoroethylene 

RAR Reverse aqua regia 

RF Radio frequency 

RMS Root mean square 

RSD Relative standard deviation 

SEM Scanning electron microscopy 

SVDV Synchronous vertical dual view 

TDS Total dissolved solids 

UNC Uncertainty 

VCR Ventersdorp contact reef 

XRD X-ray diffraction 

  

 

 



1 
 

1 INTRODUCTION 

1.1 Background 

Gold is one of the most mined minerals in the world and is often referred to as a precious metal 

(Neingo & Tholana, 2016). It is mostly used for jewellery, investments, to make coins and in 

dentistry. In Figure 1.1, as of 2023, 50% of gold is utilized in the jewellery industry (Hughes 

et al., 2021; Carpinteyro et al., 2021). 

 

Figure 1.1 Uses of gold (Natural-resources, 2023). 

 

In South Africa, the mining of gold started in the 19th century; thus, a significant amount of land has 

been mined, with South Africa being the 8th largest producer of gold (Minerals Council, 2023). This 

has subsequently led to the depletion of the valuable minerals at the level currently being mined and 

would have to mine deeper into the Earth’s surface, which can pose various challenges. South Africa 

currently has the deepest gold mine, namely, the Mponeng gold mine, which is located along the 

Ventersdorp contact reef in Carletonville (Scheepers & Malan, 2022). The challenges faced by the 

mining industry include technology and power to reach lower levels, safety for personnel and the 

costs that will be incurred, plus the generation of increased mine waste(Neingo & Tholana, 2016) 

This has led to interest in the recovery of precious metals from mine tailings, which are the residues 

of the ore that the valuables were extracted from and would often consist of the chemicals used in the 

processes (Kossoff et al., 2014). 

 

Jewellary Investments Central banks Technology



2 
 

With the significant exploitation of gold deposits in South Africa, an increase in the number of tailings 

produced is inevitable. In Figure 1.2, the distribution of the gold tailings dumps in Gauteng is shown, 

which is quite significant. A major impact of these tailings on the environment is that they occupy a 

significant amount of land that could be used for other purposes. This is supported by a report in 

which it was discovered that in the Johannesburg area, more than 321 km2 of land is covered by toxic 

and radioactive tailings (DRDGOLD, 2016). In addition to occupying land, tailings introduce sulfides 

and pyrite minerals into the soil, which could leach into the ground and surface water (Dold, 2014). 

Previous studies by Rösner and Van Schalkwyk (2000) on quantifying the impacts of tailings have 

found that tailings consist of harmful substances such as cyanides, with the concentrations of copper, 

chromium, cobalt, nickel, and zinc exceeding the threshold concentration for soils. 

A further adverse effect of tailings is their ability to cause acid mine drainage, which is a problem 

that South Africa is trying to solve. Recently, there have been cases in which dams storing tailings 

collapsed and led to tailings being introduced into nearby communities. This was witnessed in 

Jaarsfontein in the Free State, South Africa, which led to fatalities and destroying houses (Torres-

Cruz and O’Donovan, 2023)



3 
 

 

Figure 1.2 Satellite image of gold dumps in South Africa. 

 



4 
 

Therefore, the possibility of mining deeper into the Earth and the adverse effects of tailings on the 

environment have led to various studies into the viability of these tailings. Between 2010 and 2015, 

various projects were launched that were aimed at the reclamation of valuable minerals from tailings. 

Studies by Ahmari and Zhang (2012), Li et al., (2017), Zhu et al., (2017), and Luo et al., (2020) have 

examined the possible use of tailings in brickmaking for construction purposes and glass 

manufacturing. 

1.2 Gold tailings economics in South Africa 

Due to the amount of mining carried out in South Africa, the amount of tailings produced is 

significant. In Johannesburg, there are approximately 446 mine tailing dumps that cover a total area 

of 18000 hectares, with the East Rand having the majority of the dumps (Table 1.1). 

Table 1.1 Distribution of mine dumps in the Witwatersrand Basin (Mabaso, 2023). 

Goldfield Number of dumps Total area (ha) 

East rand 154 4,893.4 

Central rand 127 2,588.5 

West rand 69 1,195.1 

Carletonville 18 1,710.6 

Venterskroon 15 1.941.0 

Klerkdorp 17 4,902.0 

Welkom 46 1,365.1 

Total 446 18,495.17 

 

Due to the various effects of mine tailings on the environment, processing plants for tailings have 

been developed by different mining houses to reduce their ecological footprints. Pan African 

Resources have developed the Elikhulu and Barberton Tailings Retreatment Plant (BTRP), which 

processes approximately 1.2 Mt historic tailings per month and recovers about 55 000 ounces of gold, 

as shown in Table 1.2. Another tailings processing plant that has been developed is the Ergo plant by 

DRDGOLD, which processes 1.7 Mt of tailings and recovers 80 000 ounces of gold. A substantial 

amount of tailings have already been processed. Figures 1.3 and 1.4 show the tailings dumps before 

2016 and after 2016, in which the tailings have been cleared that emanated from a mine in the West 

Rand. Reprocessing of mine tailings benefits mining houses in terms of finances, the reduced 

resources needed and their environmental management. 



5 
 

Table 1.2 Various gold tailings treatment plants currently running in South Africa. 

Treatment facility Tailings 

processed per 

month (ton) 

Gold 

recovered 

(ounce) 

Grade of 

gold (g/t) 

References 

Barberton Tailings 

Retreatment Plant 

100 000 20 000 0.7 (Pan African 

Resources, 2020) 

Elikhulu 1 200 000 60 000 0.1 (Pan African 

Resources, 2020) 

DRDGOLD 1 700 000 80 000 0.3 (DRDGOLD, 

2016) 

Mine Waste 

Solutions 

- 100 000 0.25 (Harmony, 2016) 

 

 

Figure 1.3 Tailings dam in West Rand in 2005 (Source: Google Earth). 

 

 



6 
 

 

Figure 1.4 Tailings dam in West Rand in 2016 (Source: Google Earth). 

 

As an example of the financial benefit from reprocessing mine tailings, the Elikhulu project, which 

is estimated to cost approximately R2 billion, will process 179 million tonnes that were accumulated 

during Evander's history of gold mining and produce up to 50 000 oz per year from a reserve of 1.7 

million oz mine (Pan African Resources, 2020). The BTRP produced 28,590 oz of gold at a 

remarkably low all-in-sustaining cost of $332 per oz (Pan African Resources, 2020). Additionally, 

gold production increased by 18% and led to self-sufficiency in a year and a half. Furthermore, the 

reprocessing of tailings from various plants is forecasted to continue until 2032, as shown in Figure 

1.5. 

 

Figure 1.5 Graphical representation of the amount of gold expected to be produced per year at 

the Elikhulu and BRTP treatment facilities. LOM represents the life of mine (Pan African 

Resources, 2020). 

0

10000

20000

30000

40000

50000

60000

70000

80000

2
0
1

8

2
0
1

9

2
0
2

0

2
0
2

1

2
0
2

2

2
0
2

3

2
0
2

4

2
0
2

5

2
0
2

6

2
0
2

7

2
0
2

8

2
0
2

9

2
0
3

0

2
0
3

1

2
0
3

2

O
u
n
ce

s 
p

ro
d

u
ct

io
n
 p

er
 y

ea
r

Elikhulu LOM

production profile

BTRP LOM ProfileActual Forecast



7 
 

Based on the information obtained on mine tailings and industry research, analytical techniques are 

necessary for the quantification of gold with the lowest measurement uncertainties. Additionally, few 

to no studies describe the outcomes of their method development without the corresponding 

uncertainties. In the analytical quantification of gold, the methods involved or used have associated 

errors that can be quantified and calculated as to how much they contribute to the final measurement 

value. This can lead to the identification of steps or methods that need improvement. This can also 

benefit the mining industry in reprocessing tailings for the reduction of the ecological footprint and 

to recover any lost gold. Therefore, this study looked into the quantification of gold from low-grade 

tailings using various methods together with the evaluation of the statistical measurement 

uncertainties. 

1.3 Problem statement 

The continued mining of gold in South Africa contributes significantly to the economy of the 

country. However, at the same time, it is leading to the increased production of mine waste or 

tailings, which tend to affect the environment in various ways. For example, they occupy a large 

amount of land, and toxic minerals can end up in nearby communities via airborne or water 

transportation. This has then sparked a shift from mining companies to re-mine the tailings in a 

bid to decrease the land they cover and to cut operation costs, safety concerns and their impact on 

nearby communities. 

Subsequently, the challenge that was then encountered was for methods to extract and quantify the 

gold in an accurate manner, as it is in low amounts (low grade). Although methods such as fire 

assays are often problematic for low-grade samples, they can offer high errors. Therefore, this 

study looked into optimizing various methods and evaluating the uncertainties of these methods. 

This is important because it stands to lead to a high amount of tailings being recovered, with an 

accurate method, thereby holistically benefitting different sectors. 

1.4  Aim 

- The aim of the study was to develop an environmentally friendly, cost-effective and low 

measurement uncertainty-associated method for the quantification of gold from low-grade 

tailings. 

1.5 Objectives 

a. To characterize the mine tailings samples by X-ray diffraction and alkaline fusion to 

obtain the mineralogy. 



8 
 

b. The purity of the reagents was assessed by analysing the preparation blanks of the 

various methods by ICP‒MS and ICP‒OES to determine the purity of the reagents. 

c. To determine the optimum method for analysis of gold, the following methods were 

used: 

 Fire assay 

 Acid dissolution with solvent extraction 

 Wet chlorination using sodium hypochlorite (NaOCl) and hydrochloric acid 

(HCl) 

d. To quantitatively determine the gold in the tailings by employing gravimetric and 

instrumental determination using ICP‒MS and ICP‒OES; 

 Gravimetric determination 

 Instrumental determination using ICP‒MS and ICP‒OES 

e.  To evaluate the associated uncertainties in the investigated methods (fire assay and 

acid digestion. 

1.6 Justification of study 

With the depth of mining increasing (going deeper into the Earth), it is becoming impractical to mine 

further. Furthermore, large amounts of tailings are being produced in the process, which is not 

beneficial for the environment. The impacts or effects are reported to be the introduction of harmful 

compounds or substances into the environment, the cause of acid mine drainage, and the occupation 

of land that could be used for other purposes (Dold, 2014). 

Therefore, it is important to shift the focus to tailings, as they might be of economic value. Having a 

practical method of determining gold and other valuable metals from tailings can lead to the optimum 

recovery of the minerals that were lost during the extraction and concentration process due to 

ineffective methods at that time. This can be a positive step into the reclamation of land occupied by 

tailings in South Africa, as there will be greater benefits. 

In the analysis of mineral-containing ores, the contribution to the measurement uncertainties comes 

from all the steps involved, from decomposition to analysis. Usually, the handling and analysis of 

low-concentration samples often lead to high measurement uncertainties, depending on the tolerance. 

In the case of low-grade tailings, being able to determine the associated measurement uncertainties 

from different matrices can offer an improved understanding of the errors involved and the credibility 

or accuracy of the methods developed. 

1.7 Novelty of the study 



9 
 

The variety of gold quantification methods focuses on medium- to high-grade ores, which allows for 

better recoveries. In this study, the focus was on developing methods for low-grade gold in tailings, 

as they will contribute significantly to the existing body of knowledge. This included investigating 

the use of small sample weights, less-invasive acid combinations and assessing the associated 

measurement uncertainties. The incorporation of the measurement uncertainty will provide guidance 

as to which method performs satisfactorily, which is important in the analytical chemistry world. 

1.8 Structure of thesis 

The thesis is structured as outlined. 

Chapter 1: Introduction 

The backdrop of the current issue and the research design are provided in Chapter 1. 

 

Chapter 2: Literature Review 

A review covering the occurrence of gold, mine tailings, and techniques for estimating the amount of 

gold in geological samples. 

 

Chapter 3: Statistical analysis 

This consists of a short review of measurement uncertainty and its evaluation. 

 

Chapter 4: Acid digestion (Manuscript) 

The methodology and strategy employed for the acid digestion studies as well as the measurement 

uncertainties were examined in this chapter. 

 

Chapter 5: Fire assay (Manuscript) 

In this chapter, the fire assay method by gravimetric analysis will be the focus, and the measurement 

uncertainties will be addressed. 

 

Chapter 6: Chlorination (Manuscript) 

Al look into the chlorination of the gold tailings sample. 

 

Chapter 7: Conclusion and future prospects 

A conclusion based on the three experimental chapters and future prospects. 

 



10 
 

1.9 References 

Ahmari, S. and Zhang, L. (2012). Production of eco-friendly bricks from copper mine tailings 

through geopolymerization. Construction and Building Materials, 29, 323–331. 

https://doi.org/10.1016/j.conbuildmat.2011.10.048 

Dold, B. (2014). Evolution of acid mine drainage formation in sulphidic mine tailings. 

Minerals, 4(3), 621–641. https://doi.org/10.3390/min4030621 

DRDGOLD. (2016). Sustainability. https://www.drdgold.com/sustainability/tailings-

management. Accessed: 16 September 2021. 

Harmony. (2016). Tailings management. 

https://www.harmony.co.za/sustainability/environment/tailings-management/. Accessed: 16 

September 2021 

Hughes, A.E., Haque, N., Northey, S. A and Giddey, S. (2021). Platinum group metals: A 

review of resources, production and usage with a focus on catalysts. Resources, 10(9), 1–40. 

https://doi.org/10.3390/resources10090093 

Kossoff, D., Dubbin, W. E., Alfredsson, M., Edwards, S. J., Macklin, M. G and Hudson-

Edwards, K. A. (2014). Mine tailings dams: Characteristics, failure, environmental impacts, 

and remediation. Applied Geochemistry, 51, 229–245. 

https://doi.org/10.1016/j.apgeochem.2014.09.010 

Li, C., Wen, Q., Hong, M., Liang, Z., Zhuang, Z and Yu, Y. (2017). Heavy metals leaching in 

bricks made from lead and zinc mine tailings with varied chemical components. Construction 

and Building Materials, 134, 443–451. https://doi.org/10.1016/j.conbuildmat.2016.12.076 

Luo, L., Li, K., Weng, F., Liu, C and Yang, S. (2020). Preparation, characteristics and 

mechanisms of the composite sintered bricks produced from shale, sewage sludge, coal gangue 

powder and iron ore tailings. Construction and Building Materials, 232, 117250. 

https://doi.org/10.1016/j.conbuildmat.2019.117250 

 

Mabaso, S. M. (2023). Legacy Gold Mine Sites & Dumps in the Witwatersrand: 

Challenges and Required Action. Natural Resources, 14(05), 65–77. 

https://doi.org/10.4236/nr.2023.145005Mia Breytenbach. (2016). Gold tailings retreatment an 

attractive proposition in current environment. Mining Weekly. 



11 
 

https://www.miningweekly.com/print-version/south-africa-has-absolute-potential-for-vibrant-

sustainable-tailings-projects-industry-players-2016-10-14. Accessed: 20 May 2023 

Mineral Council Siuth Africa (2023). Gold. https://www.mineralscouncil.org.za/sa-

mining/gold. Accessed: 15 November 2023. 

Natural-resources. (2023). Platinum facts. https://natural-resources.canada.ca/our-natural-

resources/minerals-mining/minerals-metals-facts/platinum-facts/20520. Accessed: 20 May 

2023 

Neingo, P. N and Tholana, T. (2016). Trends in productivity in the South African gold mining 

industry. Journal of the Southern African Institute of Mining and Metallurgy, 116(3), 283–290. 

https://doi.org/10.17159/2411-9717/2016/v116n3a10 

Pan African Resources. (2020). Barberton Tailings Retreatment Plant. 

https://www.panafricanresources.com/btrp/. Accessed: 16 September 2021 

Rösner, T and  Van Schalkwyk, A. (2000). The environmental impact of gold mine tailings 

footprints in the Johannesburg region, South Africa. Bulletin of Engineering Geology and the 

Environment, 59(2), 137–148. https://doi.org/10.1007/s100640000037 

Scheepers, L., & Malan, D. F. (2022). A case study of geotechnical conditions affecting 

mining-induced seismicity in a deep tabular mine. Journal of the Southern African Institute of 

Mining and Metallurgy, 122(3), 115–124. https://doi.org/10.17159/2411-9717/1511/2022 

Torres-Cruz, L. A and O’Donovan, C. (2023). Public remotely sensed data raise concerns about 

history of failed Jagersfontein dam. Scientific Reports, 13(1), 4953. 

https://doi.org/10.1038/s41598-023-31633-5 

Zhu, M., Wang, H., Liu, L. L., Ji, R and Wang, X. (2017). Preparation and characterization of 

permeable bricks from gangue and tailings. Construction and Building Materials, 148, 484–

491. https://doi.org/10.1016/j.conbuildmat.2017.05.096 

 

  



12 
 

2 LITERATURE REVIEW 

2.1  Literature review on gold 

2.1.1 Gold 

In South Africa, gold is mainly mined from the Witwatersrand Basin (Wits basin), which is 

situated in Johannesburg, Gauteng, the Ventersdorp Reef (VCR), which is situated alongside 

the Wits basin, and the Greenstone belt (GSB), which is situated in Barberton, Mpumalanga. 

Witwatersrand basin 

The Witwatersrand basin (Figure 2.1) is the largest gold field in the world, from which over 52 

000 tonnes of gold were sourced. It is estimated that approximately 30 000 tonnes of gold 

remain in the basin. Seven significant conglomerate reefs, including Evander, East Rand, 

Central Rand, West Rand, West Wits, Klerksdorp, and Welkom, make up the basin. According 

to reports, the Central Rand Super group holds the majority of the gold (Tucker et al., 2016). 

 

 

Figure 2.1 Map of the Witwatersrand basin (da Costa et al., 2020). 

 

Ventersdorp Contact Reef (VCR) 

This reef is the basal conglomerate of the overlying Ventersdorp formation at the base of the 

Ventersdorp Supergroup and is identified as part of the Witwatersrand basin. The VCR (Figure 



13 
 

2.2) is one of the major gold-bearing regions in the Witwatersrand, with an estimated gold 

grade of 5-12 g.t-1 (Birch, 2022). 

 

Figure 2.2 Map of the Ventersdorp Contact Reef (Manzi et al., 2014). 

 

2.2  Literature review on tailings 

Tailings are defined as the material that remains from the ore after the extraction of various 

minerals such gold and often consists of the chemicals used during the extraction processes 

(Hudson, 2001). In most cases, valuable minerals exist in the earth’s crust in small 

concentrations, resulting in many tailings being produced to obtain a low quantity of minerals. 

Previous studies by Lottermoser (2007) reported that the ratio of gold to gangue material is 

10:1. With the grade of minerals such as gold decreasing, it is expected that the ratio will 

increase in favour of the gangue. In Australia, the grade of copper decreased by 1% from 10% 

between 1885 and 2005, whereas in Canada, the nickel grade dropped by 1.5% from 5% in the 

same years (Giurco et al., 2009). For gold, it is reported that the amount of tailings to 

concentrate is higher than that of base metals because they already exist in parts per million 

amounts in nature, with the grade expected to fall to 2 g.t-1 by the year 2050 (Giurco et al., 

2009). 



14 
 

The composition of tailings is dependent on the mineralogical makeup of the ore body, the 

processing chemicals, and the efficiency of extraction and weathering (Kossoff et al., 2014). 

The most abundant minerals that are typically found in tailings are quartz and pyrite, along 

with base metals such as aluminium, calcium, magnesium, manganese, sodium, phosphorus, 

titanium and sulfur (Table 2.1) (Kossoff et al., 2014). Most studies have looked into 

quantifying toxic elements such as arsenic, copper, lead and zinc from tailings (Hudson-

Edwards et al. 2003). In the mineralogical characterization of tailings from AngloGold Ashanti 

and base metal extraction by Malatse and Ndlovu (2015), it was found that SiO2 is the major 

mineral in the gangue fraction and minerals such as calcite and dolomite in the minority. 

Secondary minerals also form from tailings, and the composition of these minerals depends on 

the mineralogy of the source and environmental conditions such as pH and climate. Examples 

of these are geothite, scondite, etc. (Kossoff et al., 2014). 

Table 2.1 South African tailings characterization from the literature. 

Mine or Matrix Minerals Elements Reference 

Louise Moore mine Dolomite, Mica, 

Amphibole, Serpentine, 

Smectite, Quartz 

As (277), Au (0.37), Cd 

(0.4), Cr (359), Fe (40,47), 

Ni (224)* 

(Singo & Kramers, 2021) 

Gold mine in 

Krugersdorp 

 Zn>Ni>Co>As>Pb>Cd (Omotola Fashola et al., 

2020) 

East Rand basin Quartz, Magnetite, 

Marcasite, Kyanite, 

Gupeite, 

Magnesioferrite 

Si (45.98), Al (2.35), K 

(1.29), Fe (0.99), Ti (0.42), 

Mg (0.27), Ca (0.16)# 

(Okereafor et al., 2020) 

AngloGold Ashanti SiO2 (77.7), Al2O3 

(10.2), Fe2O3 (4.51), 

CaO (1.93), Ti (0.47)# 

 (Malatse & Ndlovu, 2015) 

Witwatersrand 

basin 

Quartz (58-82), Mica 

(3-10), Chlorite (3-10)# 

 (Nengovhela et al., 2006) 

West rand basin SiO2 (36.6), Al2O3 

(7.25), Fe2O3 (4.27), 

CaO (0.68), TiO2 

(0.27)# 

 (Gcasamba et al., 2019) 



15 
 

Ventersdorp 

Contact Reef 

Quartz, muscovite, 

chalmosite, gypsum, 

pyrite, talc 

As (0.088), Si (34), Fe (3.4), 

Al (3.4), Mg (0.97), Ti 

(0.23)# 

(Mashale et al., 2023) 

*Unit: µg.g-1, #Unit: % 

2.2.1 Type of tailings 

As mentioned in Section 2.2, tailings are generated as a result of extraction processes applied 

on the ore, which includes cyanidation, flotation and gravity concentrating. In Table 2.2, the 

different types of tailings and the processes from which they are generated are shown. 

Table 2.2 Types of tailings emanating from different mineral processing techniques. 

Type of tailings Formation Reference 

Gravity concentration tailings 

Gravity concentration: the process by 

which particles of different sizes, 

shapes and densities are separated 

from each other by the force of gravity. 

Fine gold is usually too small to be 

recovered by installed equipment. 

(Marsden 

and House, 

2006) Flaky gold which presents a large 

surface area in one plane and 

carried away with the lighter 

particles. 

Hydrophobic gold that adheres to 

the water‒air interface, giving the 

gold particle an artificially low 

density. 

Unliberated gold which is 

associated with oxides, silicates or 

sulfide gangue. 

Cyanidation tailings 

Cyanidation: a hydrometallurgical 

technique for extracting or leaching 

gold from ores by converting the gold 

to a water soluble complex. 

Free or exposed gold that is too 

coarse to dissolve in the initial 

process. 

(Marsden 

and House, 

2006) 

Gold locked in silicates or oxide 

gangue 

Gold encapsulated within 

unreactive and nonporous sulfide 

grains such as pyrite. 



16 
 

Gold coated with iron-oxides or 

hydroxides which are formed by 

dissolution or precipitation 

reactions. 

Flotation tailings 

Flotation: Separates concentrate ores 

by changing their surface 

Sulfide gold tailings coated in 

nonfloatable material such as iron 

oxides and silicates or if 

associated with nonhydrophobic 

materials. 

(Marsden 

and House, 

2006) 

 

2.2.2 Storage of tailings 

Various ways exist in which tailings can be stored, which is extremely important. These include 

storing as slurries or as dried material in storage dams (Kossoff et al., 2014), with the former 

being the most used. Most of the impoundments are constructed near water sources to prevent 

dust from emerging from the tailings surface and possible acid mine drainage (Kossoff et al., 

2014). In some cases, due to poor construction, the dams fail, leading to major environmental 

effects. The most adverse impacts are that the failures lead to the discharge of toxic elements 

into river systems, which affects water and sediment quality, impacting the ecosystem. In South 

Africa, gold tailings dams failed in 1994 due to a breach following heavy rains, and 600 000 

m3 of tailings were released into the community of Merriespruit (Fourie et al., 2001). This led 

to the death of 17 people due to drowning and suffocation. In another failure in Spain in 1998, 

1.3 x 106 m3 of Ag, Cu, Pb and Zn tailings were released into a river, and the entire fish and 

shellfish population was killed (Hudson-Edwards et al., 2003). In a more recent incident, a 

tailings dam storing mine waste from a diamond mine collapsed in September 2022 in the Free 

State, South Africa. This led to one death and several injuries with houses destroyed (Torres-

Cruz & O’Donovan, 2023). These are examples of fatalities that are a result of tailings dam 

failures, which shows how dangerous they are. However, once they dry, tailings still pose a 

threat with dust to surrounding communities and AMD leaching that has toxic metals. 

2.3 Sampling methods 

In the handling of a small sample from a larger original sample, it is important to have a sample 

that is a representation of the bulk sample, which will influence the homogeneity. Depending 

on the method and application, there are various methods for splitting the bulk sample. Splitting 



17 
 

the bulk sample also allows for the assessment of the sample homogeneity across the fractions 

before proceeding with other methods. 

a. Riffle splitting 

This involves the splitting of the sample using chutes that move in alternative 

directions, resulting in two randomly divided equal-sized fractions (Petersen et al., 

2004). This process can be repeated until the preferred size is obtained. The riffle 

splitting method is reported to have a relative standard deviation (RSD) of 1% (Petersen 

et al., 2004). 

b. Rotary splitting 

In this method, the sample is fed into a feed hopper and split into rotating bins at a 

uniform rate using a vibratory feeder (Nenuwa et al., 2018). These bins can be set to 

collect the sample more than once until the preferred sample size is obtained. The 

percent RSD for the rotary splitting method is 0.1% (Nenuwa et al., 2018). 

c. Cone and quarter 

This method is optimum for large samples, whereby the sample is shovelled into a 

uniform conical shape pile, then flattened and divided into four sections using two 

boards (Taggart, 2009). As expected, this method has a high percent RSD of 7% 

(Taggart, 2009). 

d. Grab sampling 

In grab sampling, the homogenized sample is placed on a rolling mat, and random 

portions are taken using a spatula or a shovel, depending on the sample size (Petersen 

et al., 2004). This method is reported to have a relative standard deviation of 5% 

(Petersen et al., 2004). Although faster and cheaper, the grab sampling splitting method 

introduces a significant amount of errors. 

e. Slurry splitting 

In slurry splitting, the homogenized sample is mixed with water to form a slurry that is 

put through a rotary splitter at a slower rate than the dry samples. 

2.4 Sample size 

When dealing with geological samples, the amount or mass of the sample being sampled for 

analysis is crucial, as it can severely affect the subsequent analysis. According to the study of 

Stanley (2007), the sampling error contributes more to the component errors in the 

determination of gold from various ores than the actual analysis (Stanley, 2007). This is mainly 



18 
 

because gold is inhomogenously distributed in the ore and the gold particles usually form 

nuggets, which are when the gold particles form clumps (Stanley, 2007). An example of a 

nugget effect is usually seen when the sample is analysed in duplicate and the results are not 

comparable. However, such problems are rarely encountered in the analysis of base metals 

because the ore grains containing them are more abundant than those containing gold (Stanley, 

2007). 

Various ways exist to try and ensure that, first, the sample is homogenous and that the right 

amount of the sample is taken, which will guarantee satisfactory homogeneity with the lowest 

measurement uncertainties. One method used to reduce or avoid large sampling errors is to 

have a sufficient number of replicates, but this often leads to biases in the interpretation of the 

data, especially when the sample is unknown (Mao et al., 2013). The other method includes 

screening the sample for particle size, either using the old sieve analysis method or by using 

instrumental analysis such as the Malvern Mastersizer. These tests allow for the establishment 

of the particle sizes, and if the particle size is not satisfactory, the sample is subjected to milling 

to make it finer. In the literature, it is reported that for the analysis of gold, the finer the material 

is, the greater the chances of eliminating the nugget effect and obtaining improved 

homogeneity (Mao et al., 2013). 

Furthermore, the homogeneity of the sample after carrying out certain chemical analyses can 

be quantified using the homogeneity indicator (Equation 2.1). This homogeneity indicator is 

calculated using the equation below, where RSD is the rsd calculated from the repetitive 

measurements and m is the mass that was weighed. A value of less than or equal to 10 indicates 

that the sample is of satisfactory homogeneity (Mao et al., 2013). 

                                       𝐻𝐸 = (𝑅𝑆𝐷) × √𝑚                                                   Equation 2.1 

 

Upon determining the relative homogeneity, Pauwell’s equation (Equation 2.2) can be used to 

establish the minimum sample mass, which is the smallest accurate fraction of the bulk that 

may be regarded as representative at a certain measurement uncertainty (Mao et al., 2013). 

                                               𝑀 = (
𝐾2

′×%𝑅𝑆𝐷)

𝑈𝑁𝐶
) × 𝑚                                             Equation 2.2 

where K’2 is a factor for tolerance limits for a 2-sided normal distribution and UNC is the 

needed uncertainty in that specific chemical analysis. 

2.5 Characterization techniques 



19 
 

2.5.1 Mineralogical composition 

Mineralogical information was obtained using XRD. When equipped with the X’Pert 

Highscore plus software and PAN-ICSD, this technique can give the dominant minerals 

together with the quantitative results. 

2.5.2 Elemental composition 

Determination of the elemental composition of the samples is important in the characterization 

stage, as it provides information on what the sample contains. This is mostly important in the 

selection of the subsequent methods downstream of characterization. In this case, the common 

methods employed to decompose geological samples are alkaline fusion and acid digestion, 

which are analysed by ICP‒MS, ICP‒OES, atomic absorption spectroscopy (AAS) and pressed 

pellets or fused beads, which are analysed by X-ray fluorescence (XRF). 

2.5.2.1 Sample preparation 

Alkali fusion 

This method includes the fusing of the sample on an open flame in a crucible using flux 

reagents such as sodium carbonate (Na2CO3), sodium peroxide (Na2O2), lithium metaborate 

(LiBO2), lithium borate (LiB4O7), potassium borate (K2B4O7), potassium hydroxide (KOH) 

and sodium hydroxide (NaOH) (Liu et al., 2019). Although all the reagents serve their purpose, 

Na2CO3 and Na2O2 are the most preferred, as they can readily decompose refractory minerals 

such as zircon. The alkali fusion method offers advantages such as time efficiency, as it is 

usually short, works for refractory minerals and has high decomposition efficiencies, as it is at 

high temperatures (Enzweiler & Potts, 1995). Due to the high flux-to-sample ratio, it often 

leads to high blanks with high total dissolved solids (TDS), which becomes a problem during 

instrumental analysis. However, this is often overcome by dilution of the melt in high volumes. 

For example, a 0.2 g sample plus flux is usually dissolved into a 200 mL volumetric flask. 

Furthermore, gel-like insoluble silicates can form in the process, which introduces difficulties 

during analysis (Liu et al., 2019). 

Acid digestion 

In this method, the samples are digested by acid attack from the use of various acids, such as 

HCl, nitric acid (HNO3), perchloric acid (HClO4) and hydrofluoric acid (HF) (Liu et al., 2019). 

These can be through an open vessel or a closed vessel, which offer different advantages. These 

acids can be used in their individual form or a mixed form. For example, HCl, HNO3, HClO4 

and HF are often used as a combination, and HF has the added advantage of being able to 



20 
 

decompose silicates (Wang et al., 2016). This decomposition allows for decreased spectral 

interference and the liberation of minerals that were enclosed in the silicates (Balaram & 

Subramanyam, 2022). However, other minerals, such as zircon, might be partially digested, as 

the maximum temperature is determined by the boiling point of the acid combination, in which 

case, zircon might need high temperatures (Balaram & Subramanyam, 2022). Furthermore, this 

method can be performed in close vessels, such as microwave-assisted digestion, in which the 

sample and the acids are added into a microwave vessel and digested for a certain time. This 

has an advantage of short digestion times and a minimum amount of reagents used (Wang et 

al., 2016). However, this can also be limiting, as some minerals require longer digestion times. 

2.6 Determination of gold in high- and low-grade samples, including mine tailings 

In most cases, the valuable metals are deposited in the ore, which needs to be decomposed and 

concentrated to separate the minerals from it and quantify thereafter. With high-grade ores, the 

conventional method to use is the fire assay (FA), which has worked for decades. However, 

other alternatives, such as dry or wet chlorination, thermochemistry and dissolution, have been 

applied successfully in high-grade ores and a few low-grade ores (Table 2.3). Alternative 

methods were employed to minimize the errors that FA introduces and to reduce the time 

needed for FA. However, with tailings, which are expected to be very low grade, there is limited 

research on whether FA is still the optimum method and whether it can be determined 

gravimetrically. There is also limited research on whether the characteristics of the ore change 

significantly after being chemically processed, which subsequently influences the selection of 

the decomposition and analysis methods. 

2.6.1 Fire assay 

Fire assay is a decomposition (and concentrating) method that has been used for decades. It 

involves fusing the sample at temperatures of at least 1100 °C with a flux that contains a 

collecting medium such as lead oxide (Figures 2.3) (Rao and Reddi, 2000). The fused mixture 

is then poured into iron molds to allow for cooling and phase separation, upon which the gold-

containing button is separated from the slag in a process called deslagging. The button is 

subsequently put on a cupel and heated at 1000 °C (Figure 2.4), in which the lead and other 

base metals are removed from the gold as oxides. The prill that is obtained after cupellation is 

then subjected to further chemical analysis, either instrumental or gravimetric analysis. Lead 

FA (Pb-FA) is usually used for the collection of gold, platinum and palladium (Rao and Reddi, 

2000). FA is the most preferred decomposition method because it can handle large samples and 



21 
 

allows for complete separation of the gold from the gangue material (Rao and Reddi, 2000). In 

a study by Rabatho et al. (2010), before subjecting low-grade tailings to FA, gold was 

concentrated using magnetic separation, and grades of 0.7 and 0.2 g.t-1 were reported for gold 

and palladium (Pd). Subsequently, the decomposed samples are concentrated using methods 

such as adsorption and precipitation to gravimetrically or instrumentally determine gold using 

ICP‒OES/MS and AAS, among others (Wang et al., 2016). Many studies prefer instrumental 

analysis, as it is much easier and multiple elements can be determined, with low detection 

limits. There is limited information on the gravimetric determination of gold from low-grade 

tailings. 

Since it can handle large samples, it can be assumed that FA can be applied for low-grade ores. 

However, it also requires an extremely experienced assayer, as errors in the fluxes and masses 

added can lead to incomplete collection of metals. Common problems encountered with FA 

include the prolonged time of dissolving the button, which can lead to low recoveries in the 

case of incomplete dissolution (Rao and Reddi, 2000). However, various studies have looked 

into using tellurium and stannous chloride to precipitate out the gold, which leads to improved 

collection efficiencies (Jackson et al., 1990, Morcelli et al., 2004). 

 

Figure 2.3 Schematic diagram of the fire assay process and its various finishes (Balaram & 

Subramanyam, 2022). 

 



22 
 

 

Figure 2.4 Diagram of cupellation during the fire assay (Mcintosh, 2004). 

 

Components of fire assay flux 

In the fire assay, the flux is the main component of the process, as it contains reagents that 

ensure that the process occurs efficiently. The flux assists by leading to the formation of two 

phases: a liquid borosilicate slag that contains everything except for the noble metals and a 

liquid lead phase that contains the noble metals. 

Litharge (PbO) is the source of the metallic lead that is responsible for collecting gold, which 

is reduced by carbon from maize or flour (Equations 2.3 to 2.5). This process of reduction 

occurs in the presence of oxygen, which enters the furnace through an opening, and the 

reduction follows the following reactions: 

                                      2𝐶(𝑠) + 𝑂2(𝑔) → 2𝐶𝑂(𝑔)                                             Equation 2.3 

                                        𝐶𝑂(𝑔) + 0.5𝑂2(𝑔) → 𝐶𝑂2(𝑔)                                            Equation 2.4 

                                     𝑃𝑏𝑂(𝑠) + 𝐶(𝑠) → 𝑃𝑏(𝑙) + 𝐶𝑂(𝑔)                                          Equation 2.5 

In addition to being a collector, PbO can destroy any sulfides in the sample and functions as an 

oxidant, allowing gold encapsulated by sulfides such as pyrite to be liberated, as in Equations 

2.6 and 2.7 (Mcintosh, 2006). Furthermore, the litharge reacts with the slag to form lead 

silicates, which ensure the homogeneity and fluidity of the slag; therefore, the amount of lead 

started with is not 100% recovered. 

                                                     
                                     3𝑃𝑏𝑂(𝑠) + 𝐹𝑒𝑆(𝑠) → 3𝑃𝑏(𝑙) + 𝐹𝑒𝑂(𝑔) + 𝑆𝑂2(𝑔)                       Equation 2.6 

                                              𝑃𝑏𝑂(𝑠) + 𝑆𝑖𝑂2(𝑠) → 𝑃𝑏𝑆𝑖𝑂3(𝑔)                                          Equation 2.7 

Most importantly, the targeted analytes must be soluble in the lead collection phase to be 

quantitatively collected. 



23 
 

The other constituents of the flux that are common; 

- Na2CO3 is a basic compound that reacts with silica-based minerals to form a fusible 

sodium silicate. It also acts as a desulfurizing and oxidizing agent. Sodium carbonate 

also decreases the melting points of the minerals in the samples so that fusion can occur 

(Mcintosh, 2006). For example, quartz melts at 1700 °C by itself, but when reacted with 

sodium carbonate, the melting point decreases to 790 °C (Figure 2.5). 

 

 

Figure 2.5 Phase diagram of the Na2O-SiO2 (mass-%) system (Vadász et al., 2019). 

 

- Quartz (SiO2): SiO2 is a strong oxidizing agent that forms silicates, which are the basis 

for the formation of slag. It can also protect crucibles from the corrosive action of 

litharge. 

- Borax glass or fused borax (Na2B4O7): a strong oxidizing agent that lowers the 

temperature of fusion so that the slag can form and it fluxes metal oxides well (Equation 

2.8). It also increases the fluidity of slags, thereby making pouring easier. 

                                               𝐵2𝑂3(𝑠) + 𝐶𝑎𝑂(𝑠) → 𝐶𝑎𝑂. 𝐵2𝑂3(𝑙)                            Equation 2.8 

 



24 
 

- Calcium fluoride (CaF2): increases fluidity and helps with difficult pours. 

- Reducing agent (maize meal or flour): source of carbon that reduces the PbO to Pb to 

enable it to collect the noble metals. 

It has been reported by Rao and Reddi (2000) that the flux constituents have to be in certain 

amounts concerning each other for maximum collection of the targeted metals. However, the 

ratio of the litharge to reducing agent (PbO:C) is of utmost importance, as the reaction that 

occurs, which leads to the production of lead rain, is crucial in the collection of the targeted 

analytes by lead. 

2.6.2 Chlorination 

Chlorination of ores is a metallurgical process that involves treating ores with sources of 

chloride ions, which produces chlorinated metal complexes that are often soluble in most 

solvents (Atasoy, 2002). Chlorinated metal complexes usually have different properties from 

their precursor, such as high volatility, low melting point, water solubility and ease of 

reduction, which leads to the ultimate aim of separating them from gangue or other metals. 

Chlorination is the preferred leaching method because it is environmentally safer than 

cyanidation (Baghalha, 2007). 

 

Chlorination can occur in a dry environment (Figure 2.6), whereby chlorine gas is used as the 

source, or in a wet environment (Figure 2.7), whereby HCl, sodium chloride (NaCl), or 

ammonium chloride (NH4Cl) are used as the source of the chloride ions (Atasoy, 2002). 

Chlorination offers the advantage of being applied to low-grade ores and allows for large 

sample masses of up to 250 grams (Rao and Reddi, 2000), resulting in detection limits ranging 

from 0.06 ng.g-1 to 0.25 ng.g-1 for platinum. In Baghalha (2007), calcium hypochlorite 

(Ca(OCl)2) vs sodium hypochlorite (NaOCl) were used for a gold sample of 1-2 g.t-1, and the 

maximum gold recovered was 58% after 48 h of Ca(OCl)2, while with NaOCl, the leaching 

time was 24 h under acidic conditions. However, in Fu et al. (2017), where NaOCl was used, 

good recoveries of at least 60% were obtained under alkaline conditions. In a study by Matsau 

and Koch (2003) that looked into a method that excludes FA, it was found that chlorination 

under specific conditions was able to determine the targeted gold from an ore. 

 



25 
 

 

Figure 2.6 Process of dry chlorination (Dosmukhamedov et al., 2022). 

 

Figure 2.7 Process of wet chlorination. The reaction vessel consists of the sample slurry and 

NaOCl and HCl (Baghalha, 2007). 

 2.6.3 Acid dissolution/digestion 

This is based on the fact that gold can be digested in acids such as HF, HCl, HClO4, bromide 

(Br2) and hydrogen peroxide (H2O2) under specific conditions. However, the extraction of gold 



26 
 

from geological samples using acids is often not satisfactory, as minerals such as braggite and 

copper do not dissolve in HCl and HNO3, matrices of chromite are also resistant to acid attack, 

and sulfur-rich samples have fine-sized particles, which hinders high recoveries. A previous 

study by Matsau and Koch (2003) reported a roasting process that was added before acid 

dissolution to help with the mentioned problems, and high extractions were obtained. 

2.6.4 Solvent extraction 

Solvent extraction of gold from solutions as a method of preconcentration has gained popularity 

because it is easy to carry out and offers high extraction efficiencies (Nguyen et al., 2016). In 

Nguyen et al. (2016), reagents such as tri-butyl phosphate and quaternary ammonium salts such 

as Aliquat-336 in diisobutyl ketone (DIBK) were used for the selective extraction of Au(II) 

from an HCl solution. In another study by Radzyminska-lenarcik et al. (2021), gold was 

concentrated and extracted using a new amine derivative, β-diketone, while a greener solvent, 

diethyl carbonate, was evaluated in Raiguel et al. (2020). Both studies reported extraction 

efficiencies of over 80%. Other extractants reported to offer high recoveries are N-n-

octylaniline, amilorie monohydrochloride, etc. 

2.6.5 Adsorption 

Ion exchange is often used where solvent extraction is lacking, whereby a reversible exchange 

of ions takes place between a solid and liquid phase of ion exchangers during two sequential 

processes, namely, adsorption and elution. According to Izatt et al. (2012) and Izatt et al. 

(2015), Au, Pd, platinum (Pt), iridium (Ir), and rhodium (Rh) were sequentially separated using 

Superlig resins with varying affinities, whereas in Komendova (2020), preconcentration of gold 

was achieved using the Strata SDB-L sorbent. However, it offers the benefits of concentrating 

selected elements to measurable levels and can also remove sample matrices that often 

contribute to interferences during instrumental analysis. The drawbacks of this technique are 

that repetitions are often necessary to achieve complete removal of base metals. 



27 
 

Table 2.3 Methods commonly employed in the decomposition of gold in various ore bodies. 

Method Key reagents Sample 

mass, g 

Analysis method Reference 

Lead FA Pb/Na2CO3,Na2B4O7,SiO2,flour 10-30 ICP‒MS (Hall & Pelchat, 1994) 

Lead FA Pb/Na2CO3,Na2B4O7,SiO2,flour 50 FS-LA-ICP‒MS (Bédard & Barnes, 2002) 

Lead FA Pb/Na2CO3,Na2B4O7,SiO2,flour 10-20 HRCS-GFAAS (Ni et al., 2019) 

Lead FA Pb/Na2CO3,Na2B4O7,SiO2,flour 15 ICP‒MS (R. Juvonen et al., 2002) 

Precipitation Hydroxylamine hydrochloride 0.2 Gravimetry (Singh, 2012) 

Alkaline fusion with ion exchange Na2O2 0.5 GF-AAS (Enzweiler & Potts, 1995) 

Alkaline fusion with Se-Te 

coprecipitation 

Na2O2/NaKCO3/KOH 1-20 ICP‒MS (Amassé, 1998) 

Alkaline fusion with Te 

precipitation 

Na2O2 2 ICP‒MS (Jin & Zhu, 2000) 

Acid digestion with Te 

coprecipitation 

HF+ HCl+HNO3 2-5 GF-AAS (Gupta, 1989) 

Acid digestion with Te 

precipitation 

HNO3+HF+HClO4+HCl 0.3-1.3 INAA (Elson & Chatt, 1983) 

Acid digestion with Hg 

precipitation 

HCl+HNO3 0.5 GF-AAS (Niskavaara & Kontas, 1990) 

Acid digestion with solvent 

extraction 

HCl+HNO3 

MIBK 

0.1-2.0 GF-AAS (Terashima, 1988) 

Acid digestion with solvent 

extraction 

HNO3+HF+HCl 

DLLME 

0.02 GF-AAS (Fazelirad et al., 2014) 

Acid digestion with solvent 

extraction 

HCl+HNO3 10 GF-AAS (Monteiro et al., 2003) 

Acid digestion with solvent 

extraction 

HBr +Br2 5-10 GF-AAS (Chattopadhyay & Sahoo, 1992) 

Acid digestion HCl+HNO3 10 GF-AAS (Liu et al., 2013) 

Acid digestion HNO3+HF+HCl+AR 

DIBK-loaded CG71 resin 

4 ICP‒MS (Pitcairn et al., 2006) 

Acid digestion (MW) HCl+HNO3 

Single granular carbon 

0.2 GF-AAS (Hassan et al., 2011) 

Acid digestion (MW) HCl+HNO3 

Modified carbon nanotubes 

0.5 SS-HR-CS-GF-AAS (Dobrowolski et al., 2017) 



28 
 

Acid digestion HCl+HNO3 

Magnetic nanoparticles 

5-10 FI-column-GF-AAS (Ye et al., 2014) 

#Based on determining the purity of the gold 

 



29 
 

2.7 Quantitative determination of gold 

Following the procedures above, the gold can be quantitatively determined using techniques 

such as instrumental and gravimetric analysis. 

2.7.1 Instrumental analysis 

For decades, gold has been and still is determined mainly on instruments such as ICP‒OES or 

ICP‒MS, AAS, LA-ICPMS. The common advantage of instrumental analysis is that multiple 

elements can be determined simultaneously; hence, most studies use this technique (Jarvis et 

al., 1995; Juvonen et al., 2004; Morcelli et al., 2004). Instruments such as ICP‒OES and ICP‒

MS have detection limits as low as 10 µg. L-1 (Komendova, 2020). For F-AAS, the detection 

limits range from 3 to 500 µg. L-1 (Komendova, 2020). Disadvantages of instrumental analysis 

include the costs associated with performing the analysis and the expertise of the analyst, which 

is important in the correct interpretation of the results. Table 2.4 shows the various instrumental 

techniques and their limits of detection. 

 

2.7.2 Gravimetric analysis 

Gravimetric analysis, whereby it is determined by weight, is an easy and cost-effective 

procedure. However, it can be challenging in regard to the determination of gold that is in 

association with platinum group metals (PGMs). Studies such as Vasekina et al. (2014) and 

Singh (2012) have quantified gold using gravimetry through the use of hydroxylamine 

hydrochloride. In Singh (2012), gold was gravimetrically quantified from a gold metal by 

reduction in a zero-valent state by hydrolamine hydrochloride, and it was found to be a rugged 

and precise method, offering low uncertainties. The biggest challenge with gravimetric analysis 

is the increase in the work that must be carried out. Another disadvantage of this method is that 

it cannot be carried out on low-grade samples, such as tailings. However, gravimetric analysis 

or determination offers direct measurement traceability to the physical units. 

 

These two quantification techniques have been used in various studies for the longest times, 

thus leaving few gaps. However, a gap that was identified was that they were rarely associated 

with measurement uncertainty. Incorporating the measurement uncertainty with these 

techniques strengthens their capabilities and provides a bigger picture into the accuracy and 

limitations of the results they produce 



30 
 

 

Table 2.4 Various instrumental analysis techniques for the quantification of gold in geological matrices. 

Technique Advantages Disadvantages Typical LOD References 

ICP‒OES Fast analysis, simultaneous 

detection of analytes. 

Suffers from a great deal 

of interferences and 

method development can 

be time-consuming 

0.01-0.1 ng.L-1*, 

10 µg.L-1* 

 

(Juvonen et al., 2004; Komendova, 2020) 

 

ICP‒MS Low detection limits, wide linear 

range 

Intensive spectral 

interferences for 

geological samples 

0.053 ng.g-1*, 

0.90 ng.g-1*, 

0.71 µg.L-1*, 

0.0068 ng.L-1#, 

0.007 µg.L-1! 

 

(Cadar et al., 2021; Jackson et al., 1990; Juvonen et al., 2004; 

Oguri et al., 1999; Wang et al., 2016) 

 

GF-AAS Require low sample volume, 

complex organic matrices can be 

eliminated proper to atomization of 

sample 

Expensive, low sample 

throughput and requires 

highly skilled operators 

0.3 ng.L-1#, 

0.5 µg.L-1@, 

 0.1 µg.L-1# 

 

(Balaram et al., 2012; Medved’ et al., 2004; Niskavaara & Kontas, 

1990; Todand et al., 1995) 

 

LA-ICP‒MS Minimal sample preparation, access 

to isotopic information 

Ions detected following 

m/z separation are not 

typical of the original 

sample's makeup. 

0.007 µg.L-1! (Limbeck et al., 2015) 

INAA It is not affected by problems 

arising from digestion due to 

difficult matrices. 

Can analyse a limited 

number of analytes. 

2 µg.L-1*, 

0.7 µg.L-1# 

 

(Asif et al., 1992; Elson & Chatt, 1983) 

 



31 
 

*= fire assay, #=digestion,!= fusion, @=ion exchange resins 



32 
 

2.7  References 

Amassé, J. (1998). Determination of platinum-group elements and gold in geological matrices 

by Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) after separation with selenium 

and tellurium carriers. Geostandards Newsletter, 22(1), 93–102. 

https://doi.org/10.1111/j.1751-908X.1998.tb00548.x 

 

Asif, M., Parry, S. J and Malik, H. (1992). Instrumental neutron activation analysis of a nickel 

sulfide fire assay button to determine the platinum group elements and gold. The Analyst, 

117(8), 1351–1353. https://doi.org/10.1039/an9921701351 

 

Baghalha, M. (2007). Leaching of an oxide gold ore with chloride/hypochlorite solutions. 

International Journal of Mineral Processing, 82, 178–186. 

htps://doi.org/10.1016/j.minpro.2006.09.001 

Balaram, V., Mathur, R., Satyanarayanan, M., Sawant, S. S., Roy, P., Subramanyam, K. S. V., 

Kamala, C. T., Anjaiah, K. V., Ramesh, S. L and Dasaram, B. (2012). A Rapid Method for the 

Determination of Gold in Rocks, Ores and Other Geological Materials by F-AAS and GF-AAS 

After Separation and Preconcentration by DIBK Extraction for Prospecting Studies. Mapan - 

Journal of Metrology Society of India, 27(2), 87–95. https://doi.org/10.1007/s12647-012-0012-

2 

Balaram, V and Subramanyam, K. S. V. (2022). Sample preparation for geochemical analysis: 

Strategies and significance. Advances in Sample Preparation, 1, 100010. 

https://doi.org/10.1016/j.sampre.2022.100010 

Bédard, L. P and Barnes, S. J. (2002). A comparison of the capacity of FA-ICP‒MS and FA-

INAA to determine platinum-group elements and gold in geological samples. Journal of 

Radioanalytical and Nuclear Chemistry, 254(2), 319–329. 

https://doi.org/10.1023/A:1021632118200 

Birch, C. C. (2022). Optimizing cut-off grade considering grade estimation uncertainty - A case 

study of Witwatersrand gold-producing areas. Journal of the Southern African Institute of 

Mining and Metallurgy, 122(7), 337–346. https://doi.org/10.17159/2411-9717/1403/2022 

Cadar, O., Mocan, T., Roman, C and Senila, M. (2021). Analytical performance and validation 

of a reliable method based on graphite furnace atomic absorption spectrometry for the 

determination of gold nanoparticles in biological tissues. Nanomaterials, 11(12). 



33 
 

https://doi.org/10.3390/nano11123370 

 

Chattopadhyay, P and Sahoo, B. N. (1992). Modified decomposition procedure for the 

determination of gold in geological samples by atomic absorption spectrometry. The Analyst, 

117(9), 1481–1484. https://doi.org/10.1039/an9921701481 

 

da Costa, G., Hofmann, A and Agangi, A. (2020). A revised classification scheme of pyrite in 

the Witwatersrand Basin and application to placer gold deposits. Earth-Science Reviews, 

201(July 2019), 103064. https://doi.org/10.1016/j.earscirev.2019.103064 

 

Dobrowolski, R., Mróz, A., Dąbrowska, M and Olszański, P. (2017). Solid sampling high-

resolution continuum source graphite furnace atomic absorption spectrometry for gold 

determination in geological samples after preconcentration onto carbon nanotubes. 

Spectrochimica Acta - Part B Atomic Spectroscopy, 132, 13–18. 

https://doi.org/10.1016/j.sab.2017.03.011 

 

Dosmukhamedov, N., Kaplan, V., Zholdasbay, E., Argyn, A., Kuldeyev, E., Koishina, G and 

Tazhiev, Y. (2022). Chlorination Treatment for Gold Extraction from Refractory Gold-

Copper-Arsenic-Bearing Concentrates. 1–14. 

Elson, C. M and Chatt, A. (1983). Determination of gold in silicate rocks and ores by 

coprecipitation with tellurium and neutron activation-γ-spectrometry. Analytica Chimica Acta, 

155, 305–310. https://doi.org/10.1016/S0003-2670(00)85610-X 

Enzweiler, J and Potts, P. J. (1995). The separation of platinum, palladium and gold 

fromsilicate rocks by the anion exchange separation of chloro complexes after a sodium 

peroxide fusion: an investigation of low recoveries. Talanta, 42(10), 1411–1418. 

https://doi.org/10.1016/0039-9140(95)01577-X 

Fazelirad, H., Taher, M. A and Nasiri-Majd, M. (2014). GFAAS determination of gold with 

ionic liquid, ion pair based and ultrasound-assisted dispersive liquid‒liquid microextraction. 

Journal of Analytical Atomic Spectrometry, 29(12), 2343–2348. 

https://doi.org/10.1039/c4ja00272e 

Fourie, A. B., Blight, G. E and Papageorgiou, G. (2001). Static liquefaction as a possible 

explanation for the Merriespruit tailings dam failure. Canadian Geotechnical Journal, 38(4), 

707–719. https://doi.org/10.1139/t00-112 



34 
 

Gcasamba, S. P., Ramasenya, K., Ekolu, S and Vadapalli, V. R. K. (2019). A laboratory 

investigation on the performance of South African acid producing gold mine tailings and its 

possible use in mine reclamation. Journal of Environmental Science and Health - Part A 

Toxic/Hazardous Substances and Environmental Engineering, 54(13), 1293–1301. 

https://doi.org/10.1080/10934529.2019.1642694 

Gupta, J. G. S. (1989). Determination of trace and ultra-trace amounts of noble metals in 

geological and related materials by graphite-furnace atomic-absorption spectrometry after 

separation by ion-exchange or co-precipitation with tellurium. Talanta, 36(6), 651–656. 

https://doi.org/10.1016/0039-9140(89)80257-7 

Hall, G. E. M and Pelchat, J. C. (1994). Analysis of geological materials for gold, platinum and 

palladium at low ppb levels by fire assay-ICP mass spectrometry. Chemical Geology, 115(1–

2), 61–72. https://doi.org/10.1016/0009-2541(94)90145-7 

Hassan, J., Shamsipur, M and Karbasi, M. H. (2011). Single granular activated carbon 

microextraction and graphite furnace atomic absorption spectrometry determination for trace 

amount of gold in aqueous and geological samples. Microchemical Journal, 99(1), 93–96. 

https://doi.org/10.1016/j.microc.2011.04.003 

Jackson, S. E., Fryer, B. J., Gosse, W., Healey, D. C., Longerich, H. P and Strong, D. F. (1990). 

Determination of the precious metals in geological materials by inductively coupled plasma‒

mass spectrometry (ICP‒MS) with nickel sulfide fire-assay collection and tellurium 

coprecipitation. Chemical Geology, 83(1–2), 119–132. https://doi.org/10.1016/0009-

2541(90)90144-V 

Jin, X and Zhu, H. (2000). Determination of platinum group elements and gold in geological 

samples with ICP‒MS using a sodium peroxide fusion and tellurium coprecipitation. Journal 

of Analytical Atomic Spectrometry, 15(6), 747–751. https://doi.org/10.1039/b000470g 

Juvonen, M. R., Bartha, A., Lakomaa, T. M., Soikkeli, L. A., Bertalan, E., Kallio, E. I and 

Ballók, M. (2004). Comparison of recoveries by lead fire assay and nickel sulfide fire assay in 

the determination of gold, platinum, palladium and rhenium in sulfide ore samples. 

Geostandards and Geoanalytical Research, 28(1), 123–130. https://doi.org/10.1111/j.1751-

908X.2004.tb01048.x 

Juvonen, R., Lakomaa, T and Soikkeli, L. (2002). Determination of gold and the platinum 

group elements in geological samples by ICP‒MS after nickel sulfide fire assay: Difficulties 



35 
 

encountered with different types of geological samples. Talanta, 58(3), 595–603. 

https://doi.org/10.1016/S0039-9140(02)00330-2 

Komendova, R. (2020). Recent advances in the preconcentration and determination of platinum 

group metals in environmental and biological samples. TrAC - Trends in Analytical Chemistry, 

122, 115708. https://doi.org/10.1016/j.trac.2019.115708 

Kossoff, D., Dubbin, W. E., Alfredsson, M., Edwards, S. J., Macklin, M. G and Hudson-

Edwards, K. A. (2014). Mine tailings dams: Characteristics, failure, environmental impacts, 

and remediation. Applied Geochemistry, 51, 229–245. 

https://doi.org/10.1016/j.apgeochem.2014.09.010 

Latypov, R., Chistyakova, S and Mukherjee, R. (2017). A novel hypothesis for origin of 

massive chromitites in the Bushveld igneous Complex. Journal of Petrology, 58(10), 1899–

1940. https://doi.org/10.1093/petrology/egx077 

Limbeck, A., Galler, P., Bonta, M., Bauer, G., Nischkauer, W and Vanhaecke, F. (2015). 

Recent advances in quantitative LA-ICP‒MS analysis : challenges and solutions in the life 

sciences and environmental chemistry. Analytical and Bioanalytical Chemistry, 407, 6593–

6617. https://doi.org/10.1007/s00216-015-8858-0 

Liu, X. I., Wen, T., Sun, W., Yao, W., Wang, T and Wu, J. (2013). Determination of Au and 

Pt in geological samples by graphite furnace atomic absorption spectrometry with concentrate 

and extraction by foam plastics and thiourea. Rock and Mineral Analysis, 32(4), 576–580. 

Liu, Y. H., Wan, B and Xue, D. S. (2019). Sample digestion and combined preconcentration 

methods for the determination of ultralow gold levels in rocks. Mo