Consequential cradle-to-gate carbon footprint of water treatment chemicals using simple and complex marginal technologies for electricity supply

Chemicals produced via chlor-alkali electrolysis are widely used throughout the water industry worldwide, with treatment chemicals often the second largest source of environmental impacts from potable water production after electricity use. Population-driven increases in the future demand for potable water will require concomitant increases in the production of water treatment chemicals, with the associated environmental impacts of chemicals production primarily arising from the additional demand for electricity. Due to the dominance of electricity in the environmental performance of chlor-alkali chemicals, assessment of the future environmental impacts of potable water production is largely dependent on proper identification of the marginal source of electricity. In this paper, we present a consequential cradle-to-gate carbon footprint (cCF) for the most widely used chlor-alkali-produced disinfectant (sodium hypochlorite (13 % w/w)) and coagulant (ferric chloride (42 % w/w)) in Australia, with special emphasis placed upon the identification of future marginal electricity supply and the substitution of hydrogen gas and sodium hydroxide during production. While this analysis is presented in an Australian context, commonalities in potable water and chlor-alkali chemical production processes internationally give the findings a broader relevance. Consequential models for sodium hypochlorite (13 % w/w) and ferric chloride (42 % w/w) production were developed, and the identification of the marginal source of electricity was modelled using a simple marginal technology approach via operationalisation of the Weidema framework and a complex marginal technology using a partial equilibrium model. For the simple marginal technology, the levelised cost of electricity was used to select the most competitive energy generation technologies and those most relevant for the Australian market. For the complex marginal technology, the energy sector model was used to simulate the most likely electricity supply mix. Details of the different paths taken in the substitution of hydrogen gas and sodium hydroxide are also presented. To allow for proper incorporation of uncertainties arising from these key factors in the cCF, several scenarios were developed covering fuel and carbon prices for identifying the marginal supply mix of electricity, as well as the likely production routes for sodium carbonate in the context of sodium hydroxide substitution. cCF results of sodium hypochlorite (13 % w/w) and ferric chloride (42 % w/w) are presented using simple and complex marginal technologies, and the implications of choosing one marginal technology over the other in the context of water treatment chemicals are presented. For the simple marginal technology approach, the global warming potential (GWP) per megagram of chemical varied from 68 to 429 kg CO2-eq for sodium hypochlorite (13 % w/w) and 59-1,020 kg CO2-eq for ferric chloride (42 % w/w). For the complex marginal technology approach, the GWP per megagram of chemical varied from 266 to 332 kg CO2-eq for sodium hypochlorite (13 % w/w) and 214-629 kg CO2-eq for ferric chloride (42 % w/w). Insights are given in relation to the impact of the price of fossil fuels, the carbon price, and the different substitution routes. The use of a partial equilibrium model (PEM) has enabled a better understanding of the variability of the results in this study. For example, the use of PEM for the identification of the complex marginal source of electricity shows that, for the case of Australia, any benefit from a carbon price is lost with high prices of natural gas due to the incentive to use cheaper fuels such as black and brown coal. Likewise, the use of explorative scenarios was decisive to manage the inherent uncertainty of the parameters included in the model. In relation to substitution, the case of ferric chloride (42 % w/w) indicated that using only one substitution route was not enough to fully understand the potential continuum of cCF results. The simple marginal approach, where an exclusive marginal source of electricity or substitution route is considered, presents significant risks for the modelling accuracy of the cCF as shown here for sodium hypochlorite (13 % w/w) and ferric chloride (42 % w/w), therefore, it is not recommended.