*3.1. Define*

The aim of this stage is to identify opportunities for improvement that lead to e.g., an increase in benefits, reduced costs or losses, a mitigation of the environmental impact, etc. This requires pinpointing observed problems, framing them within the context of the corresponding processes, evaluating the costs and benefits of addressing them, and locating the most appropriate people to do it given the existing constraints on time and resources.

In this Six Sigma project, the focus was set on the purity (before separation) and volume of production (after purification) of one of the star products at the chemical plant where the continuous improvement program was implemented. This came as a result of an observed increase in the variability of this product's purity, due mostly to significantly lower values (compared to previous operations) being obtained once every four batches, approximately. This also meant a decrease in its average value of around 1%, starting on September of 2014, with an estimated monetary loss of more than 100,000 €/year with respect to previous years.

Figure 3a illustrates the 'Suppliers, Inputs, Process, Outputs, Customers' (SIPOC) diagram identifying the supplier (reaction 1) and inputs (one of the outputs from reaction 1) for the specific process under study (reaction 2), as well as its outputs (with primary focus on the composition of the so-called subproduct 2) and the customer (reaction 3). Figure 3b corresponds to a simplified process block diagram for reaction 2, in which four numbers have been included to indicate the points of the process where the process variables and critical to quality characteristics (CQC) relevant to this work are routinely measured.

**Figure 3.** (**a**) 'Suppliers, Inputs, Process, Outputs, Customers' (SIPOC) diagram; (**b**) simplified process block diagram for reaction 2.

Due to confidentiality reasons, the name of the reactants, products, and subproducts cannot be disclosed. However, it must be noted that both 'Reactants (1)' and 'Fresh Reactant (1)' come from reaction 1. Likewise, Premix 1 and Premix 2 are blends constituted by the same chemical species, these being freshly produced raw materials for Premix 1, and one of them (Rec. reactant (2)) having been recovered after the washing and separation following reaction 2. For each batch, each reactor was fed Premix 1 or Premix 2, but never a blend of both. A quick look at already available data revealed that the main concern that motivated this Six Sigma project was related to a loss of purity (in terms of the desired chemical species) of the Crude, (i.e., unrefined product) and therefore a lower amount of Product arriving to reaction 3 per batch of reaction 2.

The project team was constituted by the authors of this paper (Six Sigma Black Belts), championed by two high-profile members from the company and supported by three experts in the process from the technical team. Regarding the planning, the project was estimated to require over six months to be completed. Experimenting in a laboratory or at pilot-plant level was not recommended, since no proper scaling could be done. Additionally, altering or interrupting the production of this process was not allowed to any extent, and therefore experimenting on the plant itself was not an option either. Due to this, only historical data from past production could be used.
