||Carbon materials play a fundamental role in the development of fusion reactors, for both the generation of electric power and the production of nuclear materials. It is possible to synthesise graphite and carbon materials from coal. Coal is available in large quantities and could be used for the production of high-purity carbon graphite. However, it contains large quantities of impurities that need to be removed prior to graphitisation/carbonisation. The impurity levels of certain elements in this graphite must be kept at very low levels. Boron, which absorbs neutrons strongly, should be below 500 ppb. Europium and gadolinium, which absorb neutrons and are activated to highly radioactive products, as is cobalt, should be as low as 50 ppb. Lithium transforms to tritium, which leads to the circulating helium becoming radioactive. Other elements, such as calcium, sodium, silicon, thorium and uranium, should not be ignored. The purpose of this study was to lower or remove completely the impurities and trace elements in coal that affect the quality of nuclear-grade graphite. The organic part of Tshikondeni coal was dissolved in a solvent, dimethylformamide (DMF), on addition of sodium hydroxide. The first stage of purification is centrifugation and filtration, which removes most of the impurities. The recovered organic material, known as ‘Refcoal’, may be converted to graphitisable coke. Some elements, significantly boron and cobalt, associate with the organic material in solution and are not sufficiently separated by centrifugation and filtration. Further purification was employed during each process step in the conversion of coal solution into graphite. Different methods of purification were employed in this study. They included chlorination, acid treatment and the ion-exchange or complexation method. Chlorine gas and hexachlorocyclohexane (benzene hexachloride) were used in the chlorination method. Acids such as hydrochloric, hydrofluoric and ascorbic were used in acid treatment. In the ion-exchange method, reagents such as methane, starch, potassium cyanide, ethylene-diaminetetraacetic acid, sodium fluoride, sodium sulphate, ice, glycerol and sodium nitrate were used. All the treated Refcoal was coked at 1 000º C. Pyrolysis was applied in other methods with the aim of volatilising elements that form volatile halides at higher temperatures. Analysis was done for elements such as calcium, cobalt, europium, gadolinium, lithium, sodium, silicon, thorium and uranium, and other elements in the periodic table. Inductively coupled plasma mass spectroscopy and inductively coupled plasma optical emission spectroscopy were used to analyse the concentrations of the trace elements in the coal (treated and untreated) and the coked Refcoal. In inductively coupled plasma mass spectroscopy, microwave digestion and fusion were applied as methods of preparation. However, the instrumentation gave different results for the same sample. The results showed that specific methods work for specific elements. The chlorination method and the acid-treatment method (especially using hydrofluoric acid and hydrochloric acid) gave better purification for most of the trace elements and other elements. Better purification was achieved with elements such as, boron, calcium, europium, gadolinium, lithium, sodium and silicon. All the treatments failed to lower uranium and thorium to the level required for nuclear-grade graphite. However, uranium has a low boron equivalent and does not pose serious problems with respect to nuclear usage. All the methods failed to remove cobalt and this remains a problem.