key: cord-0253420-j8mmjsp2 authors: Raoufi, Kamyar; Manoharan, Sriram; Etheridge, Tom; Paul, Brian K.; Haapala, Karl R. title: Cost and Environmental Impact Assessment of Stainless Steel Microreactor Plates using Binder Jetting and Metal Injection Molding Processes date: 2020-12-31 journal: Procedia Manufacturing DOI: 10.1016/j.promfg.2020.05.052 sha: cc337714f8e1188f12caa0adf97cb806af1e4d9f doc_id: 253420 cord_uid: j8mmjsp2 Abstract Modular chemical process intensification provides a platform for down-scaling chemical process equipment and reducing the number of chemical process steps through increased surface area-to-volume ratios and increased heat and mass transfer rates. In turn, these improvements reduce capital investment and operating costs, facility size, feedstock consumption, and emissions from chemical processing. However, conventional microreactor fabrication has employed inefficient material removal and bonding processes to create devices formed of layers of metal shim stock. This research investigates the relative costs and environmental impacts of metal additive manufacturing and powder metallurgy processes for making microreactor components for a range of market sizes. Binder jetting additive technology and metal injection molding were studied for producing two plates for a microscale chemical reactor used in dimethyl ether production. The manufacturing process design method was applied to quantify the cost of goods sold, and life cycle assessment was applied to model environmental impacts. The manufacturing process design analysis showed that metal injection molding would have better cost performance than binder jetting for a greenfield facility due to lower capital tooling cost and shorter cycle time in making the green part. However, the life cycle assessment indicated, at lower annual production volumes, metal injection molding would have higher cumulative energy demand, global warming potential, and other impacts due to the mold plates and solvent use. While this work reports on model development and a single use case, it motivates a focus on validating analysis results for a range of part sizes and geometries, as well as alternative production routes. This future work would provide design and manufacturing decision makers with richer information regarding process capabilities, production costs, and product environmental impacts for a range of product complexities. Process intensification in chemical production facilities can lower capital investments due to size reductions achieved through the high surface-to-volume ratios of microscale unit operations and reactors [1] . These devices often utilize superalloys and rely on the use of microchannels to obtain high heat and mass transfer rates [2] [3] [4] [5] [6] [7] . Subtractive manufacturing, which sequentially removes material, is often not suitable for producing such devices due to the high cost of machining high aspect ratio features and other complex geometries [8, 9] . In place of subtractive manufacturing, conventional powder Process intensification in chemical production facilities can lower capital investments due to size reductions achieved through the high surface-to-volume ratios of microscale unit operations and reactors [1] . These devices often utilize superalloys and rely on the use of microchannels to obtain high heat and mass transfer rates [2] [3] [4] [5] [6] [7] . Subtractive manufacturing, which sequentially removes material, is often not suitable for producing such devices due to the high cost of machining high aspect ratio features and other complex geometries [8, 9] . In place of subtractive manufacturing, conventional powder metallurgy processes such as metal injection molding (MIM) are capable of producing microstructural geometries at high production volumes [10, 11] . MIM starts with injection molding of a metal powder-binder feedstock to create a green part. This is followed by debinding to remove the binder, and sintering to densify the part into a strong metal product [12] . Additive manufacturing has the potential to provide a reliable manufacturing process flow for producing microchannel-based chemical production devices. Additive manufacturing processes build up parts layer by layer [13] , enabling fabrication of parts with complex geometries, while using fewer tools and equipment [14] . Additive manufacturing Process intensification in chemical production facilities can lower capital investments due to size reductions achieved through the high surface-to-volume ratios of microscale unit operations and reactors [1] . These devices often utilize superalloys and rely on the use of microchannels to obtain high heat and mass transfer rates [2] [3] [4] [5] [6] [7] . Subtractive manufacturing, which sequentially removes material, is often not suitable for producing such devices due to the high cost of machining high aspect ratio features and other complex geometries [8, 9] . In place of subtractive manufacturing, conventional powder metallurgy processes such as metal injection molding (MIM) are capable of producing microstructural geometries at high production volumes [10, 11] . MIM starts with injection molding of a metal powder-binder feedstock to create a green part. This is followed by debinding to remove the binder, and sintering to densify the part into a strong metal product [12] . Additive manufacturing has the potential to provide a reliable manufacturing process flow for producing microchannel-based chemical production devices. Additive manufacturing processes build up parts layer by layer [13] , enabling fabrication of parts with complex geometries, while using fewer tools and equipment [14] . Additive manufacturing a l i d a t i n g a n a l y s i s r e s u l t s f o r a r a n g e o f p a r t s i z e s a n d g e o m e t r i e s , a s w e l l a s a l t e r n a t i v e p r o d u c t i o n r o u t e s . T h i s f u t u r e w o r k w o u l d p r o v i d e d e s i g n a n d m a n u f a c t u r i n g d e c i s i o n m a k e r s w i t h r i c h e r i n f o r m a t i o n r e g a r d i n g p r o c e s s c a p a b i l i t i e s , p r o d u c t i o n c o s t s , a n d p r o d u c t e n v i r o n m e n t a l i m p a c t s f o r a r a n g e o f p r o d u c t c o m p l e x i t i e s . 48th SME North American Manufacturing Research Conference, NAMRC 48 (Cancelled due to processes can reduce material waste and enable rapid product customization, leading to improved environmental, social, and economic performance [15, 16] . Powder bed fusion (PBF), binder jetting (BJ), directed energy deposition (DED), and sheet lamination processes have been demonstrated for producing metal components in an additive manner [17] . PBF and BJ, which create parts by fusing metal powders, are perhaps the most promising for producing microscale products, such as metal microreactor plates. PBF utilizes a laser or electron beam as a heating source to melt metal powders layer by layer, forming the desired product as the output of the process [18] . The product may require postprocessing, e.g., machining, to obtain the desired dimensions and tolerances of precision parts. Rather than directly fusing the metal powder, BJ uses a binding agent to bond the powders together layer by layer, creating a green part. The green part then undergoes debinding and sintering; the resulting sintered part may require additional post-processing to obtain the final specifications. The BJ process does not require a heat source, reducing cycle time and improving cost-effectiveness relative to PBF for high production volumes for many applications [19] . While the technical advantages and disadvantages of MIM and BJ, which are often cost-effective for higher production volumes, have been explored in prior research [12, 20] , few studies have reported their relative sustainability performance. The objective of the work presented herein is to investigate cost and environmental performance of these two processes for the production of microscale chemical reactor components for a range of production volumes. Manufacturing process design (MPD) and life cycle assessment (LCA) methods are utilized to quantify costs and environmental impacts, respectively. The remainder of the paper is organized into four sections. The MPD method, which defines process requirements, process flow diagrams, and cost models is presented in Section 2. The main steps of the LCA method, which define the study goal and scope, process inventory approach, and impact assessment methods, are presented in Section 3. The analysis results are then presented in Section 4, and conclusions and directions for future research are discussed in Section 5. The MPD method [21] investigates the physics and the chemistry of the capital equipment, shaping tools, and work holding tools for the unit manufacturing processes (UMPs) required for making a product in order to generate supporting process flows and related costs in a bottom-up manner. Product development starts with defining product functional requirements, which dictate the specifications for materials and part geometries. In turn, material and geometry specifications provide inputs used to define the UMPs, or the smallest elementary manufacturing activity[ies] required for a specific taxonomological transformation and composed of machines, devices, or equipment [22] , required for component production. The product investigated here is a microscale chemical reactor ( Fig. 1) used to convert syngas into dimethyl ether (DME). The DME microreactor plates are made of 316L stainless steel due to the elevated operating temperatures (250-280 °C) and pressures (7-20 bar). Table 1 summarizes the specifications of the three microreactor plates. Since the top plate can be produced using traditional subtractive methods (e.g., laser cutting), its production is not evaluated as a part of this study. Process flow diagrams, developed based on the process requirements, present the sequence of manufacturing process steps required for making the middle and bottom microreactor plates. The process flow diagrams for the BJ and MIM processes were developed based on inputs from technical literature and industry experts (Fig. 2) . The process flow for MIM involves feedstock preparation and three major process steps [12] : injection molding, debinding, and sintering. Injection molding starts with heating the feedstock (binder and metal powder mixture) and injecting it into the mold cavity. Next, once packed into the mold cavity, the feedstock is allowed to cool. Finally, the solidified part is ejected from the mold and the machine is readied for the next process cycle. Following injection molding, the next major step is primary debinding to initiate primary binder removal. Removing the primary binder forms a network of interconnected pores, which are then used by the secondary binder at higher temperatures to ensure binder evaporation without the generation of cracks or loss of integrity within the part. The most common debinding methods use solvent or water-based debinding, or catalytic debinding [23] . Solvent debinding using an organic solvent (Tergo, a transdichloroethlylene-based agent) is selected here based on the binding agent used in the feedstock. The primary debinding step is followed by thermal debinding, to remove the secondary binders, and then sintering, to densify the material and increase the strength of the metal product. The process flow for making the middle and bottom plates of the microreactor using BJ additive technology involves four major steps: printing, debinding, depowdering, and sintering. Printing starts with loading the build box with metal powder. Next, the part is printed using a repetitive cycle, which starts with spreading a layer of powder, followed by depositing the binder, and completed by curing the layer using a heat source (e.g., lamp). This cycle repeats until all layers of the part are printed. The build box is removed at the end of the printing step. In the debinding step, the binder is thermally removed from the part in an oven. The unbonded metal powder is then manually removed during depowdering, which is time-and labor-intensive due to the care that must be taken to avoid breakage of the green part. Finally, the part is sintered to densify the material and reach the desired strength. Production cost models are required to compare the resulting cost of goods sold (COGS) for the two manufacturing process flows. A bottom-up cost modeling approach developed in prior work is applied here [24] [25] [26] . Total COGS includes seven cost elements: tool, facility, labor, maintenance, raw materials, consumables, and utilities costs, which must be calculated for each UMP in the process flow. Common assumptions between the two processes are shown in Table 2 . It was also assumed that one laborer performs manual depowdering in the BJ process, as well as injection molding in the MIM process. Development of the cost models requires a process step analysis, which includes capability and capacity analysis. Capability analysis considers selection of a machine tool capable of implementing the process step, while capacity analysis includes calculating the cycle time for each process step and considering the correlation with the overall annual production required [26] . The capability and capacity analyses, along with the cost data from vendors for each process step and raw material, enable calculation of the unit costs for each of the seven cost elements. The total unit cost (COGS for the two plates considered) can then be calculated based on the intended production volume (market size). Ultimately, market size and process utilization are key factors in determining COGS. Conducting an LCA study includes four steps: defining the goal and scope of the study, conducting an inventory analysis, conducting an environmental impact assessment, and interpreting results [27] . The goal of this study is to compare the environmental impacts of the BJ and MIM processes for the production of the middle and bottom plates of the microreactor presented above. Since both processes are assumed to produce equivalent parts, the functional unit is defined as one middle and one bottom plate. The scope of the study is cradle-to-gate (Fig. 3) , considering impact of material and part production. Since the parts are assumed to be functionally equivalent, the remaining process steps as well as product use performance and end of life strategy would have the same impacts. The required materials, consumables, and utilities for each UMP under both production scenarios were obtained from the MPD analysis. Thus, life cycle inventory (LCI) data for the processes in the system boundary were captured from literature and by consulting with design and manufacturing practitioners, as well as using the ecoinvent 3 database. Detailed materials and energy inputs for the two process flows are reported in Tables 3 and 4 . SimaPro 9 [28] , an industry-standard LCA software, was used to capture LCI data and conduct the impact assessments. Impact assessments used the cumulative energy demand (CED), IPCC 2013 V1.03 [29] , and ReCiPe 2016 [30] methods. CED captures indirect and direct energy use over the life cycle. The IPCC 2013 method applied assumes a 100-year time horizon to estimate global warming potential. Fig. 3 . DME microreactor plate production using MIM (left) and BJ (right) process flows (activities in gray boxes are outside of the system boundary) The ReCiPe 2016 method applies weightings to composite midpoint and endpoint indicators [31] . Here human health, ecosystems, and resource availability endpoint indicators quantify environmental impacts under a hierarchist weighting. With the parameters for the studies established, the MPD and LCA analyses were completed, as presented in Sections 4.1 and Section 4.2, respectively. Figure 4 illustrates the reduction of COGS with increasing annual production for the middle and bottom reactor plates using the MIM and BJ processes. It can be seen that MIM results in lower COGs than BJ for all production volumes. At 1,000 reactors per year, the unit cost for MIM ($294) is 17% lower than for BJ ($356). As production volume increases to 10,000 reactors per year, COGS for both processes decreases with the same relative cost savings, while at 100,000 reactors per year, COGS for MIM is 22% less than that for BJ. Fig. 4 . Effect of annual production volume on cost of goods sold for the production of DME microreactor plates using the MIM and BJ processes The production cost breakout by cost element for production volumes of 1,000 and 100,000 reactors per year are presented in Figs. 5 and 6, respectively. In both cases, materials become primary cost drivers as production volume increases, while capital cost effects fall dramatically. The capital tooling cost for MIM is 13% and 6% lower than for BJ at 1,000 (Fig. 5) and 100,000 (Fig. 6) reactors per year, respectively, mainly for two reasons. First, the injection molding machine is lower cost than the binder jet printer. Also, MIM has shorter cycle time, which improves equipment utilization relative to additive manufacturing (both tools are only capable of producing two plates per cycle), requiring fewer machines to meet the same demand. A second injection molding machine is needed for an annual production volume of 100,000 reactors, while a second printer is needed to produce 5,000 reactors per year. For annual production volumes of 10,000, 50,000, and 100,000 reactors, three, 14, and 28 printers are required by the BJ process, respectively. As production volume increases, injection molding keeps its better cost performance, resulting in an 80% cost savings over the printing process step for BJ at 100,000 reactors per year. The longer cycle time to make the green part in the BJ process is partly due to the curing step. However, this curing step pays off in sintering, since it reduces sintering cycle time relative to sintering after injection molding. Moreover, the sintering process step in the MIM process flow includes thermal debinding, which increases the sintering cycle time relative to sintering during the BJ process. A second sintering furnace is required at 10,000 reactors per year for the MIM process, while a second furnace is needed for the BJ process at an annual production volume of 20,000 reactors. The MIM process requires sequentially more sintering furnaces than the BJ process as production volumes increase. Thus, MIM has lower green part costs, but higher sintering cost (Figs. 7 and 8) . It should be noted that the injection molding mold and debinding solvent are consumables (solvent is recycled and changed annually). Costs of MIM consumables are higher than those in the BJ process for an annual production volume of 1,000 reactors. However, as production volume increases, mold and solvent costs are amortized over more products, consequently reducing the consumables costs for MIM. At the lower production volume, labor cost for BJ ($78) is nearly double that of MIM ($40), mainly due to manual depowdering. In fact, though capital equipment cost is low for depowdering, it is a primary cost driver at higher production volumes due to associated labor cost (labor is more than four times the cost for BJ than MIM at 100,000 reactors/year). Figure 9 presents the impact assessment results of both process flows for an annual production volume of 1,000 (top) and 100,000 (bottom) reactors for the five selected indicators. The trends in the impact assessment results for the two process flows are similar under the two different production volumes. Therefore, for simplicity and due to space limitations, only GWP is reported for the full range of volumes considered. Fig. 9 . LCA analysis endpoints for the production of DME microreactor plates using the MIM and BJ processes Figure 10 presents the effect of annual production volume on GWP for the middle and bottom reactor plates under both production routes. For an annual production volume of 1,000 reactors, the MIM process flow has a slightly higher impact compared to the BJ process flow. As the annual production volume increases to 10,000 reactors, the environmental impact of MIM (15 kg CO2 eq./unit) decreases by 32% per pair of reactor plates, while for the BJ process GWP decreases by 10%. Fig. 10 . Effect of annual production volume on GWP per reactor for the production of DME microreactor plates using the MIM and BJ processes The reduction for MIM is larger since the environmental impacts of the consumables are amortized over more products. For an annual production volume of 100,000 reactors, the environmental impacts of the BJ process flow do not change significantly compared to impacts at 10,000 reactors, while the impacts of MIM decrease by 7%, mainly due to amortizing the impacts of the consumables (mold) over more products. In fact, at 1,000 reactors/year, the impacts of injection molding is 70% higher compared to printing (Fig. 11 ), but impacts of injection molding are dramatically reduced (90% of printing) at 100,000 reactors/year (Fig. 12) . It should be noted, while printing remains a cost driver in the BJ flow, related impacts reduce by 34% as production volume increases from low to high. While seeing reductions of 34% and 10% for MIM and BJ, respectively, from low to high production volumes, sintering remains an impact driver for both process flows. The objective of the study presented herein is to analyze and to compare the relative cost and environmental performance of the MIM with BJ processes for making a microscale chemical reactor for dimethyl ether production. The results of the cost assessment showed that binder jetting has higher cost per reactor for different annual production volumes compared to metal injection molding. Using the production cost curves by cost elements and process steps, it was found that the main cost driver was capital tooling cost. Specifically, metal injection molding has lower capital tooling cost, coupled with a shorter cycle time for making the green part. Equipment utilization in the injection molding step exceeds the capacity of a single tool at a higher production volume (100,000) than the printing equipment in binder jetting (20, 000) . The other key cost driver in the binder jetting process was labor cost associated with the depowdering step. Overall, the labor cost in the binder jetting process is double that for the metal injection molding process at an annual production volume of 1,000 reactors. Although labor cost decreases as the annual production volume increases, labor cost in the binder jetting process remains higher than in the metal injection molding process. This cost differential indicates an opportunity for improving the economic performance of the binder jetting process by automating the current manual depowdering step. Life cycle assessment results indicate the metal injection molding process has higher environmental impacts than the binder jetting process for an annual production volume of 1,000 reactors. Impacts are mainly due to consumables, i.e., the mold in the injection molding step and the solvent in the debinding step. As production volume increases, however, the environmental impacts per part reduce significantly for the metal injection molding process due to amortizing the associated impacts of the mold and solvent across more products. However, this effect of amortization is not as apparent for the binder jetting process since the main environmental drivers are raw material and the utilities. To build upon this study, a wider range of products and processes should be investigated to identify appropriate operational windows on a basis of sustainability metrics. To analyze and compare alternatives, products explored should be producible by all selected processes. For example, to investigate the potential advantages of a broader set of metal additive manufacturing processes over MIM, which is limited by mold, more complex geometries (product shapes and sizes) should be investigated. This would provide richer information for design and manufacturing decision makers about process capabilities, production costs, and product environmental impacts for various product complexities. The benefits of additive manufacturing, such as improving supply chain flexibility, eliminating work-in-process, and enabling just-in-time manufacturing, which consequently reduce supply chain environmental impacts have been reported previously [16] . However, investigating the sustainability performance of the additive manufacturing supply chain was out of the scope of the research presented herein and can be explored in future work. Future Production Concepts in the Chemical Industry: Modular -Small-Scale -Continuous Microchemical Engineering in Practice Micro Energy and Chemical Systems (MECS) and Multiscale Fabrication. Micromanufacturing and Nanotechnology Development of an adhesive-bonded counterflow microchannel heat exchanger Improving the Energy Efficiency of Adsorption Chillers by Intensifying Thermal Management Systems in Sorbent Beds Conversion of Biomass Syngas to DME Using a Microchannel Reactor Highly active and stable Rh/MgOAl2O3 catalysts for methane steam reforming The mechanics of machining at the microscale: assessment of the current state of the science Micromachining Techniques for Fabrication of Micro and Nano Structures. BoD -Books on Demand Microstructural comparison and mechanical properties of stainless steel 316L fabricated by selective laser melting and metal injection moulding processes Metal powder injection molding (MIM): key trends and markets Handbook of Metal Injection Molding. 1 edition Terminology for Additive Manufacturing Technologies (Designation: F2792-12a) Environmental Performance Evaluation of Direct Metal Laser Sintering through Exergy Analysis Defining Near-Term to Long-Term Research Opportunities to Advance Metrics, Models, and Methods for Smart and Sustainable Manufacturing Environmental Performance Evaluation of a Fast Mask Image Projection Stereolithography Process Through Time and Energy Modeling Metal Additive Manufacturing: A Review Strengthening of 304L Stainless Steel by Addition of Yttrium Oxide and Grain Refinement during Selective Laser Melting Binder jetting additive manufacturing of metals: a literature review A Review on Binder Jet Additive Manufacturing of 316L Stainless Steel A Pedagogical Framework for Manufacturing Process Design Terminology to Support Manufacturing Process Characterization and Assessment for Sustainable Production Debinding and sintering of metal injection molding (MIM) components An Economic and Environmental Assessment Model for Microchannel Device Manufacturing: Part 1 -Methodology An Economic and Environmental Assessment Model for Microchannel Device Manufacturing: Part 2 -Application Comparing the Economics of Metal Additive Manufacturing Processes for Micro-scale Plate Reactors in the Chemical Process Industry Environmental Management -Life Cycle Assessment -Principles and Framework Direct Global Warming Potentials -AR4 WGI Chapter 2: Changes in Atmospheric Constituents and in Radiative Forcing PRé Consultants The Eco-indicator 99: A damage oriented method for Life Cycle Impact Assessment This material is based upon a research seed fund project supported by Oregon State University and HP, Inc. The authors gratefully acknowledge the assistance of Ms. Mackenzie Smith (HP Inc.) in determining binder jetting process data inputs.