The percentage of consumable to the overall PMI is less than 1% and was demonstrated as consumables in the breakdown shown in Figure 5.1. As the use of plant-based production routes are still considered an emerging technology, literature surrounding PMI values for plant bio-processing practices remain limited. Thus, for interpretation of this result, we compare our PMI value to the value found in Budzinski et al., which details their PMI for producing monoclonal antibodies in the biopharmaceutical field. The authors describe using a process where an average input of 7,700 kg is required to obtain 1 kg of product, resulting in a PMI of 7,700 kg inputs/kg mAb. This simple comparison assures us our calculations are within an acceptable range and our processes is not operating at similar conditions as one designed for clinical and pharmaceutical applications. Another interpretation is the plant based bio-manufacturing facility detailed here is material and process efficient in producing high concentration of resveratrol at 100 MT. While the PMI value for our process is only an order of magnitude less than a mAb production facility, this may be attributed to the addition of the CIP to the PMI calculation and the limited cleaning procedures initialized within our simulation. Here, the CIP process was simulated to follow the steps needed for an industrial chemical application, grow bags garden thus limiting the cleaning procedure to the following order: water rinse, caustic rinse, water rinse, acid rinse, and water rinse.

Certain bio-pharmaceutical CIP procedures are much more rigorous, including a larger variety of components than listed here. The PMI attributed to the CIP procedure was 163 kg/kg resveratrol, about 30% of the total PMI calculated for the entire facility. Notably, the same authors mention that the water input accounts for a large percentage of the PMI, 90% of the total mass used within their process. This claim is in alignment with our results shown below.After performing an assessment on the mass intensity of the process, further analysis was conducted using the EH&S method to evaluate how the chemicals implemented within the simulation effected the environment. Here, the materials being utilized were categorized into impact groups, where individual environmental indices can be calculated2 . The sum of the individual component EI’s gives an overall EI value for the process. This value may be used to assess the facility’s effect on the environment. Notably, components with high EI values can also identified here as they represent hazardous materials which can be replaced to provide a safer working environment. To perform the EH&S method described by Biwer and Heinzle, the MSDS sheets for all components in the input and output streams in the simulation were assessed. Using the information contained in the MSDS for each chemical, ABC classifications for the various impact groups were given to each chemical. These classifications were assigned environmental factor values of 1 , 0.3 , and 0 . These EF’s could be averaged to give a single EF for each input and output, where high values indicate potentially dangerous chemicals and lower factors indicate safechemicals.

To keep track of the amount of each material used, we calculated mass indices . The calculations for MI’s were performed by dividing the total mass of each chemical in or out of the process per batch by the total amount of product produced per batch. Lastly, the EI’s are calculated by multiplying each EF by its respective MI for all components. This analysis yields a series of bar graphs where the most impactful components make up the largest portions of either the input or output bars. Results of this analysis are shown in Figure 5.2a and Figure 5.2b. As discussed, the MI’s presented in Figure 5.2a are rescaled based on each components EF yielding Figure 5.2b highlighting any dangerous components. In Figure 5.2a we see that the proposed process generates input and output MIs of 480 and 487 kg/kg resveratrol, respectively. Both MIs are dominated by water which is absent from the overall EI’s presented in Figure 5.2b, because water is nontoxic and generates an EF value of 0. The EI’s for the inputs and outputs are 26 and 31.6, respectively. As shown in Figure 5.2a, the MIs are dominated by ethanol and plant biomass. The consistency of ethanol dominating in both breakdowns alludes to reducing the usage of ethanol as one method for operating a more environmentally friendly process.Here, a techno-economic analysis was performed on a base case production facility modeled for an annual production target of 100 MT of 98% pure resveratrol using Japanese knotweed rhizomes as the source. The economic results for a bio-manufacturing facility operating under similar conditions to the base case model are estimated to be as follows: CAPEX of $44.7 million, OPEX of $15.0 million per year, and a COGS of $150/kg resveratrol. As described, the model was built using certain assumptions. Notably, the largest assumption was the limiting the concentration of resveratrol present in Japanese knotweed to 0.5 mg/g FW.

As demonstrated by the sensitivity analysis which altered the concentration of resveratrol present in the processed rhizomes, the design of the facility and economic costs are altered significantly when the concentration is varied. When a concentration of 1.5 mg/g FW was used within the model, the CAPEX for such a facility dropped 38% from $45 to $28 million. To control the yield grown domestically, the allocation of resources such as additional R&D personnel is required to research and generate data on the factors which largely contribute to the variation between plants. Factors may include treating the soil with fertilizers to replenish any depleted nutrients within the soil, controlling moisture content via irrigation, and the testing of different field sites in search for optimal environmental conditions and low seasonal variation. These are all additional costs to be considered within the annual operating costs but are deemed insignificant, a couple of orders of magnitude less compared to the total cost required. At this time, containment costs were ignored during the design of the model. Due to the limited information available on containment methods for Japanese knotweed grown in plantations, strategy for physical containment of vegetably propagated plants from the USDA are suggested as an alternative solution until further research or information on this topic presents itself. One improvement might be to include the cost for discing 100 ft of land surrounding the specified cropland and the cost of herbicides required to prevent any rhizomes from propagating outwards. The use of invasive species for manufacturing is a controversial method which may require multiple regulatory oversight from local, state, and federal regulatory agencies. Another improvement would be initiating conversations with these regulatory bodies and aligning production of Japanese knotweed with any regulations outlined. Here, the analysis of the upstream portion of the facility was determined using information provided by UC Davis Center for Agriculture and Resource Economics. Further improvement might be to model the upstream portion using SuperPro and incorporate the production process with the purification steps, so a single simulation file can reflect the entire bio-manufacturing process as well as total CAPEX, OPEX and COGS. Downstream operations available for Japanese knotweed processing are vast. The method deployed to model the base case bio-manufacturing facility simulation is acknowledged to only be one strategy used for resveratrol production. Certain techniques and processing equipment can be added or removed to best serve the incorporation of future technology and improved manufacturing practices. Future experimental work might investigate alternative processing methods and compare both the economics and processing capabilities to explore optimal conditions and a more standardized process. An example might be to investigate the use of supercritical CO2 rather than ultrasonic technology for extraction of phenylpropanoids from plants. During the downstream processing, grow bag for tomato multiple bio-processing parameters were initialized using data provided in publicly available literature/patent. Sensitivity and scenario analysis were performed to assess the effects few conditions had towards the economics of the bio-manufacturing facility.Certain parameters such as resveratrol content, cost of ethanol, cost of enzymes, enzymatic conversion, extraction efficiency, and resin exchange frequency were altered.

Each parameter was varied using reasonable increments aligned with what can be expected. For example, the cost of ethanol and enzymes were varied within a certain ranged and assessed for their effect towards the annual operating costs. The increase of resveratrol concentration demonstrated a decrease in all economic factors measured. A decrease in CAPEX by $13 million and COGS by $48 was observed during a single incremental concentration change from 0.5mg/g to 1.0 mg/g. The cost of ethanol was varied between $1.00 – $3.00 a gallon to align with volatile costs for the commodity currently being faced within the market. As expected, the annual operating cost is expected to heavily move towards the direction of the cost of ethanol since the solvent remains the largest bottleneck in the process. The cost of enzymes was investigated as enzymes are also a commodity which can be subjected to change. Nonetheless, the low amount of enzymes used within the process allowed total costs for enzymes to remain relatively low compared to the annual operating cost, about 1% of the OPEX. Enzymatic conversion values for polydatin to resveratrol is widely detailed in literature for laboratory scale experiments but is limited for large scale processing. A sensitivity analysis was performed varying different possible conversion values and analyzing its effect on capital required for operation. It was determined that a reduction in equipment size was possible when conversions increased thus resulting in lower CAPEX and COGS values compared to the base case. The extraction efficiency of resveratrol from Japanese knotweed in an ultrasonic assisted extractor is a bioprocessing parameter which requires additional research to be conducted in an effort to model with more accuracy. Modeling the process after other plant sources demonstrated a downward trend and decrease in the CAPEX and COGS when extraction efficiency was optimized. Resins were used within the adsorption vessel used to bind and capture resveratrol arriving from the filtrate. The frequency these resins were replaced and exchanged with new resins was varied between every 25 cycles to 150 cycles in a sensitivity analysis to determine how much it impacted the COGS. Replacing the resin every 25 batches lead to an increase in COGS of about $4 while replacing the resin every 150 cycles only decreases the COGS by $0.43 . The implementation of the ethanol recycling stream was explored as one solution to combatting the high estimated cost of ethanol required for processing. Additional processing equipment needed for recycling resulted in a higher CAPEX but decreased the COGS . The bio-manufacturing facility simulation was modeled to produce annual production amounts ranging from 25 MT to 200 MT of resveratrol. As anticipated, the capital required for equipment and operating the facility steadily increased as the production amount increased but the COGS decreased exponentially. The CAPEX for the facilities producing 25 and 200 MT were $26 million and $71 million, respectively. The COGS for the 25 MT and 200 MT cases were $311/kg and $124/kg, respectively. Further sensitivity and scenario analysis can be explored to examine other variables within the process. The cost of Japanese knotweed rhizomes was limited to $0.19/kg for each scenario analysis as calculated for the base case. As the mass of Japanese knotweed rhizomes required for processing varies across different analysis, the production amount is also expected to vary. Calculating the cost of rhizomes for each sensitivity analysis may reflect a more realistic value for each case and improve the overall economics estimates. Another improvement which can be made is using other parametric uncertainty analysis such as using crystal ball or monte carlo rather than using a define set of values to test1 . Two methods were performed on the model to demonstrate the environment, health, and safety of the simulated facility. The first method which examined the process mass intensity determine the facility had a PMI value of 529 kg inputs/kg, well under the 7000 kg inputs/kg value shown for a bio-pharmaceutical facility producing mAbs2 . When environmental and mass indices were calculated using the methods described by Biwer and Heinzle, the EI’s for the inputs and outputs are 26 and 31.6, respectively and the input and output MIs of 480 and 487 kg/kg, respectively3 . These calculated values allow this stimulated bio-manufacturing facility to be classified as operating a safe and environmentally-friendly process.