Exergy references for chemical elements have been discussed in specialized literature for a long time, but so far these data are not included in standard handbooks. One reason is that the development of the reference system took place over many years in a rather small group of scientists; their ideas are still dominant today. Some assumptions and procedures are quite unusual from a thermodynamic point of view. Examples are discussed and suggestions are given to improve the situation. Very low concentrations of elements in the atmosphere and oceans and the abundance of elements in the earth＇s crust (Clarke number) are no longer used in determining reference states for chemical elements. Thermodynamic theory shows that generally liquid and gas mixtures are not useful as reference states. A generalized air-like gas is used to get reference states for the first few elements. For all other elements well defined data of pure compound from thermodynamic tables are used to develop the reference system in successive approximations. Linear optimization is used in all steps to minimize the sum of exergy values of the elements, It is shown that the allocation of reference materials to elements can be done in a logic procedure not requiring assumptions. It is also shown that this allocation procedure is unnecessary once the set of exergy values for elements is established. The values obtained are reported for 90 elements. The use of a generalized composition of air is the only aspect that reminds to the original set-up of the reference system. Otherwise the procedure is strict thermodynamics. The procedure might be used to develop a standard list of exergy values of chemical elements.