Application of environmental and economic metrics to guide the development of biocatalytic processes

2.1.1 AE

AE, also known as atom economy, measures how much of the starting material ends in the desired product [27] [Eq. (1)]. Hence, this metric assesses the reaction stoichiometry as it does not consider the Y_reaction. The driver behind this metric is the selection of a synthetic route or step targeted at product synthesis with a high mass yield and low waste [27, 28]. This principle ensures that process chemists design a reaction where as many as possible of the atoms of the substrate are included in the structure of the final product. Hence, reactions with lower AE are usually labelled as less environmentally friendly reactions [26]. AE is an easy to use metric, based on the reaction stoichiometry and mechanism [29]. However, it does not consider the by-products produced, or co-substrates used and it is based only on the reaction chemistry, not taking into account the overall process:

2.1.2 CME

CME, also known as carbon efficiency, is a green chemistry metric developed by GlaxoSmithKline (GSK) to measure the sustainability of their processes within the framework of the 12 Principles of Green Chemistry [23]. CME is defined as the percentage of carbon in the reagents that remains in the final product [Eq. (2)] and thus, the preferred reaction systems corresponds to CME=100% and a reaction with a lower CME lead to a less desired system, as there is a greater amount of carbon that will eventually end in a waste stream or as emissions. CME takes into account the stoichiometry and the amount of carbon in the reagents that is incorporated into the final product [23].

However, as noted for the previous metric (CME), it does not account for the waste generated during the process:

3.1.1 Reaction mass efficiency

Reaction mass efficiency (RME) (also known as mass efficiency) is a metric developed by GSK. This metric takes into account the Y_reaction and the actual molar quantities of reagents, as well as the atom economy [23]. RME can be calculated by computing the quotient of the mass of the product by the mass of all the reagents in the process [Eqs. (4) and (5)] [23]. However, this metric does not account for the type of waste generated, solvent use, work-up (e.g., quenching) or energy use [26, 35]:

3.1.2 E-factor

The E-factor analyzes the amount of waste formed in the synthesis of chemical compounds [36], quantifying it as the amount of waste created/kg of product produced. Despite the simple definition of E-factor, this metric can be made as complex and thorough as is required. In a simplified viewpoint, this metric can be defined as the inverse of the RME as, at this stage, assumptions on solvent and other factors can be made. By contrast, a more thorough analysis can be performed when including solvent and catalyst recycling. Therefore, this green chemistry metric has been recognized as a valuable measure to provide information about the environmental performance and waste footprint of a given synthetic route or process [37]. In the E-factor, the waste is defined as everything leaving the process boundaries, except the desired product. Therefore, the E-factor takes into account the Y_reaction, including solvent and reagent losses and process aids [38]. The E-factor includes not only the reaction chemistry and the process options related with the reaction, but can also include all the steps in a chemical synthesis, upstream or downstream of the reaction step [Eq. (6)]:

Under stoichiometric conditions (i.e., no excess), assuming that solvents and catalysts are recycled, the E-factor and RME are related according to Eq. 7, and in such situations, the use of both metrics is unnecessary:

Water is generally excluded from the E-factor, as this could lead to excessive E-factors for some processes (such as bioprocesses), making a meaningful comparison of the results difficult [37]. Since water is usually benign, the solvent of choice for green chemistry is frequently water, whereas organic solvents are more often used in conventional chemical synthesis [39]. There is a historical perception that water by itself does not have a significant environmental impact. However, one must remember that many chemical processes require highly purified water and there are life cycle impacts related to the water purification step. In addition, in many chemical synthetic routes, a mixed aqueous-organic waste stream is generated and therefore additional recovery units are required to further separate the phases prior to the wastewater treatment plant. Furthermore, nowadays, assessing the water footprint of a given process is also a measure of sustainability, since in many parts of the world competition for water is becoming more of a concern and seems certain to become a greater issue in the future [40]. Therefore, two types of E-factor should be considered: one excluding water [Eq. (6)] and one including water [Eq. (8)]:

However, this green chemistry metric does not consider what type of by-product or waste is generated. The E-factor can also be used for multi-step reactions, although the result only provides a benchmark guide for different sectors and markets of the chemical industry [41].

In the study presented by Henderson et al. [13], where a biocatalytic route was compared with the conventional chemical synthesis, the biocatalytic route led to a higher E-factor, whereas when using a simple life cycle assessment (LCA) model (FLASC tool [42]) this synthetic route shows better results with respect to energy consumption and greenhouse gas emissions [13]. However, in this case, there is a high water contribution to the E-factor of the biocatalytic route, whereas for the E-factor of the conventional chemical route, only the mass of unrecovered solvents are accounted for, leaving the solvent recovery processing steps out of the environmental evaluation. Hence, a drawback of the E-factor as an accurate metric is the need to define system boundaries before calculations. Furthermore, this simplified metric only accounts for one environmental impact (i.e., waste generation), leaving aside other relevant outputs and thus, the exclusive use of this metric can lead to biased results.

3.1.3 PMI

PMI is the metric chosen by the American Chemical Society Green Chemistry Institute’s Pharmaceutical Roundtable as a high-level metric to evaluate the extent to which a given manufacturing process is sustainable [43]. PMI is defined as the total mass of materials used to produce a specified mass of product [43], Eq. (9). When calculating PMI the starting point is the commonly available materials [43]. Hence, the metric accounts for all the steps in the chemical synthesis including the catalyst production and all reagents (including water) that are used directly in the process [44]. Additionally, PMI also includes the downstream process (DSP) steps required for isolation and purification of the final product at the required quality [44]. However, PMI does not include specific concerns regarding the environment, health and safety of the raw materials used or the waste produced. However, such concerns are normally addressed during route selection:

3.1.4 SI and WI

From a careful assessment of many of the synthetic routes to chemical products, solvents have been found to be one of the biggest mass contributors [45], as the use of solvents can lead to higher yields and thus, an improved environmental performance. This is particularly the case in the fine chemicals and pharmaceutical sectors, where the solvents typically contribute 80% to 90% of the mass intensity of a process [46]. For these sectors, it was observed that solvents contribute to about 75% of the total energy usage (used in solvent work-up and recycling), around 70% to volatile organic compounds (VOCs) emissions (and thus, to the photochemical creation potential) and about 50% for the global warming potential (GWP) [18, 19, 47]. Furthermore, the solvent production process also contributes greatly for the solvents total environmental impact [48, 49]. Hence, SI was defined as a metric in order to address the limitations raised when applying the E-factor, by analyzing and quantifying the amount of all solvents used in a given synthetic process [Eq. (10)]:

There are several solvent selection guides available in the scientific literature from GSK [50, 51], Pfizer, AstraZeneca and the American Chemical Society (ACS); these have a great focus on safety, health and environmental issues. A systematic methodology for solvent selection from an operational and environmental perspective (also integrating partially solvent selection guides) has been developed by Gani and co-authors [52, 53]. When choosing a solvent for the development of a process, it is important to take into account the environmental impact of the chosen solvent, and the potential safety and health risks associated with handling and operating with the given solvent [52]. Solvents with issues should therefore be avoided. Furthermore, other issues related with the ease of separation and recovery of the products and enzyme compatibility are often only regarded as a secondary aspect by the abovementioned solvent guidelines. Hence, solvent selection should include a score for options to recycle the solvent. For instance, for water-miscible solvents, where the solvent recovery is commonly done by distillation and evaporation, the solvents should be ranked according with the differences on the vapor pressure, whereas for immiscible solvents, the operation unit mostly used is liquid-liquid extraction and thus, the solubility differences should be the basis for the ranking. Moreover, an adequate solvent for biocatalytic processes should not cause catalyst inactivation or degradation. For water-immiscible systems (two liquid-phase systems), the solvent should be as immiscible as possible in water, in order to reduce biocatalyst contact with the potentially toxic solvent. Nevertheless, for these systems, the liquid-phase interfaces have toxic effects on the biocatalyst, since solvents with high partition coefficients (log P>4) are argued to have less toxic effects on biocatalysts [54, 55]. For water-miscible solvents, the solvent effect on the biocatalyst is dependent on its concentration and thus, the biocatalyst should be developed to tolerate these conditions. Finally, another constraint when selecting the organic solvent is the solvent reactivity, as it should not interact with the substrates nor products by reacting or degrading, or likewise the biocatalyst (e.g., avoid using organic solvents with the same functional group as the substrate).

A particular version of SI is WI, where the focus is on analyzing the amount of water used throughout the whole process [Eq. (11)]. WI is of special interest in biocatalysis and fermentation processes, since one of the most attractive features of bioprocesses in organic synthesis is the possibility of operating with an environmentally compatible solvent (water) [39, 49] often representing more than 50% of the resources mass [19]. However, water has become a scarce and overexploited natural resource. In addition, these processes usually lead to a large amount of wastewater that needs to be properly cleaned and this is usually very energy intensive [56]:

Nevertheless, the type of solvent used (or the quality of the water used) is not specified and a methodology for measuring the relative greenness of a given solvent is still required. Furthermore, the environmental issues related with the solvent recovery are not addressed by this metric.

3.1.5 Other metrics for environmental evaluation

The Green Chemistry community has generally agreed that a suitable metric should be clearly defined, simple, measurable and objective rather than subjective [57]. Despite the weaknesses of the abovementioned metrics, these fulfill (at least, partially) the requisites for suitable metrics. There are other metrics that can be applied at the early development stage (such as environmental quotient, EQ [58], EMY [59] and the C-factor [60]). However, the correct application of these metrics requires more information that is often not readily available at this development stage.

The EQ is an attempt to overcome the flaws of the E-factor, by multiplying the E-factor by a quotient that measures the type of waste [Q, Eq. (12)]. This quotient has an arbitrary value between 1 and 1000 (where 1 is the optimal waste type), depending on whether the type of waste produced is innocuous, waste toxicity and ease of recycling [58]. However, the quotient is set on a rather arbitrary and subjective ground:

The EMY is defined as the percentage of desired product relative to the mass of all non-benign materials used in its synthesis [Eq. (13)] [59]. Unlike the previously described metrics, this metric highlights the ecological effect of the reagent(s) and reaction additive(s). However, the EMY lacks a precise definition of non-benign reagents, currently defined as “those by-products, reagents or solvents that have no environmental risk associated with them” [59]. Nevertheless, this definition cannot specify or quantify the environmental risks in an objective way. Furthermore, the EMY only highlights one potential environmental impact (toxicity), disregarding others, such as the GWP or waste generated. Hence, the use of the EMY is hindered by the availability of information regarding human toxicity and ecotoxicity. For example, when assessing a biofuel production process, this metric would give a favorable score to the process, since typically the reagents are renewable feedstocks and would be considered benign reagents. This is clearly erroneous:

The C-factor quantifies the amount of CO₂ produced/kg of product formed [60] [Eq. (14)]. The innovation in this green chemistry metric is the life cycle thinking, since the C-factor includes all the CO₂ produced from the raw material production, preparation, conversion and purification of the chemical. Nevertheless, use of this metric has some potential shortcomings and for this reason it was excluded from the suggested metrics listed in Figure 2. Briefly, while the production of chemical products (in particular fine chemicals and pharmaceuticals) involves many environmental issues, such as emission of VOCs, toxicity or nutrient enrichment, the C-factor focuses solely on one environmental concern (GWP) [61]. Thus, the exclusive use of the C-factor for environmental assessment leads to a risk of these other issues being neglected. Furthermore, the C-factor as defined, only accounts for the emission of CO₂, as part of the GWP. This may be misleading, especially when using renewable resources where emissions of methane (CH₄) and nitrous oxide (N₂ O) have an equal (if not greater) potential contribution to global warming [62]. Furthermore, for fine chemicals and pharmaceuticals the calculated C-factor for the reaction performance often leads to biased results, as the large downstream impact of the process is not considered. While for bulk chemicals, the C-factor is often small [61], but it disregards the contribution of the raw materials production phase, which is often more energy intensive and thus with a higher C-factor than the production phase itself [63]. Finally, the inventory of CO₂ emissions itself is a time consuming task to ensure it is sufficiently comprehensive.

The energy loss index was introduced on a decision framework for process design developed by Sugiyama and co-workers [21]. This metric is an indicator to estimate the energy-related efforts associated to the process, using only the available reaction information forecasting the energy and utilities that will be required in a large scale. This index has a value of 0 (low) to 1 (high estimated energy demand) according with: reaction medium (aqueous or organic solvent), [P], difference of boiling points, inherent waste and reaction energy [21]. However, this metric only gives a qualitative score of the process energy demand and it was thought to be applied exclusively for a process requiring distillation processes on their recovery and purification processes, whereas the impact of the reaction adjuvants (such as co-solvents and catalysts), operating conditions and other unit of operations are not accounted for in the process energy demand.

3.2.1 Y_reaction

Efficient conversion of the raw material is an essential requirement for process success, since high Y_reaction (% mol_Product/mol_Substrate) simplifies the downstream separation and leads to a more cost-effective process, lowering the environmental and economic contribution of the raw material to the final process performance. Ultimately, the yield (and the associated cost differential of the substrate and product) determines the available opportunity for processing. For example, a low yield implies very little freedom for costs in the rest of the process. Since raw material costs are often in the range of 35%–90% of the total processing cost [18, 66], dependent on the industrial sector, there are different threshold values for the suggested Y_reaction for each market (Table 2). When aiming for the production of a low value (bulk) chemical, there is a small gap between the cost of the raw material and the profitable product cost (>80% of the total processing cost [66]). Therefore, the production costs are often dominated by the cost of the raw materials and very high Y_reaction are required. A higher Y_reaction also has a positive effect on the environmental performance, leading to higher RME (by virtue of better raw material use), lower PMI (due to the lower mass contribution of the raw materials) and a lower E-factor (since there is a smaller amount of raw materials in the waste stream).

Furthermore, achieving a high conversion of a fossil-based raw material (as is the case for many fine chemicals and pharmaceuticals [34]) leads to lower cumulative energy requirements [and consequently lower global warming, acidification potential (AP), human and eco-toxicity impacts]. For bio-based raw materials, a high Y_reaction might lead to lower nutrient enrichment potential (NEP) (due to reduced use of fertilizers during the growing of crops) and other land use aspects [34]. Due to the high selectivity of enzymes, when highly functionalized molecules or those with several chiral centers are aimed at, the biocatalytic route might appear to be a suitable alternative to avoid protection and de-protection steps (with an overall reduction of the process steps) and thus, the theoretical yield can be greatly improved and the amount of waste minimized [34]. Nevertheless, at industrially relevant conditions, high yields (i.e., high concentrations, presence of solvents) might be difficult to achieve, not only due to catalyst inhibition and/or stability (substrate or product at concentrations above their inhibitory level and/or presence of organic solvents), but also many of the industrially relevant biocatalytic reactions are not thermodynamically favorable (e.g., chiral amine synthesis using ω-transaminase). For these reactions, the allowable cost contribution of the raw material determines the efforts required for displacing the equilibrium (such as use of an excess of co-substrate or in situ product removal). However, for low value chemicals, the reaction thermodynamics might influence the process viability, since these types of chemicals have a smaller allowable cost for downstream processing (for co-substrate recovery and recycle) and often cannot afford the implementation of expensive process technologies to shift the equilibrium (e.g., membrane technology or resins for in situ product removal).

For instance, the asymmetric synthesis of amines can be performed with high Y_reaction (>90%) using a vanadium catalyst and tert-butanesulfinamide as the amine donor [67] and thus, the Y_reaction of the ω-transaminase-catalyzed reactions should at least match the overall Y_reaction of the chemical route. However, under the mild operating conditions at which the biocatalytic reaction takes place, the synthesis of the amine from a pro-chiral ketone is very often hindered by the unfavorable thermodynamic equilibrium. Hence, by putting in place first a target to the Y_reaction, the process can be designed in such a way that process technologies are put in place and the equilibrium can be shifted towards the amine production, as demonstrated (for instance) by Truppo and co-workers [68]. Among the strategies to shift the equilibrium, the use of an excess of one of the reagents is often used to improve the Y_reaction. However, the use of this strategy leads to higher waste produced in the reaction (if the excess of substrate is not recovered) and might lead to additional downstream processing steps to recover or separate the unreacted substrate from the product.

3.2.2 Y_biocatalyst

Even enzymes for which the optimization of the expression system is not fully realized can be applied in industrially relevant processes [5], as the biocatalyst cost in itself does not mean much. The relevant question to address is how much the cost of the biocatalyst (including fermentation and biocatalyst formulation) contributes to the final processing cost (when compared with the competing synthetic processes). Furthermore, some of the environmental considerations of the biocatalyst production are also covered by the Y_biocatalyst (Table 3), since the fermentation step can have a large impact on land use footprint, NEP (due to growing of the carbon source crop) and AP (due to energy demand) [69]. Moreover, a process that has a low Y_biocatalyst (i.e., a higher biocatalyst use) implies higher PMI (due to a higher mass contribution of the biocatalyst), a higher E-factor (since the amount of biocatalyst in the waste stream is higher) and higher WI (related with the higher fermentation broth volumes). Furthermore, for the production of the catalyst itself, the media preparation and the immobilization procedure are the two main contributors for the environmental impacts [42, 69]. Typically, for optimized processes, the contribution of the enzyme to the overall environmental impact is small compared to the environmental impact of the raw material and production process. Nevertheless, at early development stages, often the biocatalyst activity and half-life time have not been optimized to match the large-scale conditions and thus, using this process metric tells the biologist the targets for the biocatalyst improvement and the large-scale conditions (i.e., concentrations, presence of inhibitors, etc.). Therefore, processes where the biocatalyst contribution is significant lead also to a less sustainable process (Table 3). It is therefore necessary to set a threshold value for the efficiency of the biocatalyst (Y_biocatalyst, g_Product·g_biocatalyst^-1).

3.2.3 Product concentration

Recovery of the final product from the reaction phase is a critical step and it is often overlooked in laboratory assessments of biocatalytic processes. Clearly, the extent (and consequently the allowable cost) of the DSP is dependent on the product’s subsequent use.

Despite being beneficial for biocatalyst activity and stability, operating at low [P] (g_Product·L_reactor^-1) compromises the DSP, due to the high volume of water (or organic solvent) that it is necessary to remove (often by evaporation), prior to effective processing. Hence, operating at low [P] increases the PMI (due to an increase in the mass contribution of water and solvents [19]), the E-factor (since the amount of water or solvents in the waste stream increases), the process energy requirements [49] (and consequently the energy costs and the emissions related with energy production), the SI (leading to VOCs emissions), the process water footprint (or WI) and volumetric capacity of the DSP units of operation (with subsequent increase of the capital costs of the DSP), as also reported by Ravelli et al. [49] for a photocatalytic reaction. Hence, unless the product is removed during the course of the reaction by in situ product removal, there is a trade-off between the biocatalyst activity and stability (and consequently the Y_biocatalyst) and the final [P], that determines the process performance (Table 3).

3.2.4 STY

When developing a new process, the business drivers are two-fold: economic (CAPEX and OPEX) and environmental (often greenhouse gas emissions) and thus, energy requirement reductions are often targeted [70]. High STY, or volumetric productivity, g_Product (L_reactor·h)^-1, is required to lower the capital costs (CAPEX) and energy requirements during the reaction (lowering utility costs, labor related costs, emissions and environmental impacts related with energy production, such as the C-factor, AP and GWP). Indeed, a significant share of the greenhouse gas emissions in bioprocesses originates from the energy use and the type of energy input [71]. Tufvesson and co-workers [71] reported that changing the input energy from fossil (hard coal and natural gas) to a less carbon intensive (such as bioenergy) led to a reduction of about 20% of the GWP [71].

This process metric assesses the speed at which the reaction occurs, for a given equipment occupancy time and required production rate. When performing cost evaluation, there is often a trade-off between the STY and Y_biocatalyst, since operating at higher biocatalyst loading increases the STY (reducing the time that the reaction takes to reach completion and consequently CAPEX, utilities and labor related costs) at the expense of increasing the biocatalyst cost contribution, unless recycling and reuse is possible. Hence, the STY and [P] are indicators of the cost and impact of the process energy required, representing the upstream (i.e., reaction section) and the downstream (i.e., product recovery and purification section) costs. Furthermore, together, the Y_biocatalyst and STY determine the requirements (targets) for biocatalyst specific activity and stability.