The objective of this research is to illustrate the need to consider the environmental burden and impact posed by the adoption of alternative, green process technologies. Specifically, it addresses the synthesis of catalysts using an alternative, environmentally benign solvent. Using the life cycle assessment methodology, a quantitative technique that assigns damage metrics to process emissions, the potential environmental damage is estimated. An extensive literature search validates the uniqueness and highlights the scientific merit of the research. Data gathered includes background information on supercritical processes, life cycle assessments of systems, aggregated in-house catalyst performance data, and a case study for catalyst application.

Catalyst synthesis via supercritical fluid deposition possesses several advantages over traditional techniques that can reduce production costs and environmental burden while maintaining high activity. The creation of catalysts using a “green” approach reduces harmful emissions but also drives the advancement of alternative fuel technology, an area known to make use of expensive catalytic materials. Additionally, this research provides a rationale for adoption of supercritical CO₂ at the commercial level for metal-oxide supported catalysts. This study demonstrates how improvements at the nanoscale propagate upwards through a product development system, sometimes with unpredictable outcomes. Furthermore, it illustrates the benefit of a holistic approach to chemical process design, an area that strives for a balance between economics, performance, and environmental sustainability.

CHAPTER I

INTRODUCTION

Within the engineering profession, there is an inherent moral obligation to preserve resources and implement policies and processes that benefit future generations and humanity. Presently, research in chemical engineering stresses the use of environmentally sustainable processes through “green chemistry” and “green engineering”.^1,2 Supercritical fluid deposition (SCFD) from CO₂, a catalyst preparation technique, is frequently touted as a green process, typically because of the relatively mild reaction conditions and the use of an environmentally benign solvent. These aspects should inherently leave a small carbon footprint. Despite these claims, there is a lack of evidence, qualitative or quantitative, evaluating what the environmental impact truly is. The figure on the following page highlights why it is advised to understand the process design fully before implementation.³

Figure 1: Contributions of design and development to environmental burden

Less than 10% of the overall environmental impact occurs during design and development, yet 70% of the impact is quantified and sourced. Time is allocated at the beginning of a product or process design to reduce and/or mitigate future outputs.

The aspects of the SCFD process mentioned above are not the only metrics that define a “green” process and are specifically focused on one stage of the product life cycle. Thus, they fail to take into account the quality of the catalyst/product being produced along with possible reuse and/or other end of life aspects. To build a case for this method’s viability as an alternative synthesis route, its overall environmental impact through a life cycle assessment (LCA) must be performed. It is important not to confuse the system life cycle with an environmental life cycle assessment (usually referred to simply as an LCA). Typical system life cycle stages along with their purpose and possible gate decisions at each stage are given in Table 1.

Table 1: System life cycle stages

Life Cycle Stages	Purpose	Decision Gates
Concept	Identify Stakeholder Needs
	Explore Concepts
	Propose feasible solutions	Decision Options:
Development	Refine system requirements	> Execute next stage
	Create solution description	> Continue this stage
	Build system	> Go to previous stage
	Verify and validate system	> Hold project activity
Production	Mass produce system	> Terminate project
	Inspect and test
Utilization	Operate system to satisfy users’ needs
Support	Provide sustained system capability
Retirement	Store, archive, or dispose the system

Table 1 is a simplified and truncated version of the system life cycle available in the NASA Systems Engineering Handbook.⁴ An environmental life cycle assessment is an element of the “Concept” stage, useful in alternatives analysis and trade studies. However, data generated from each system life cycle stage and lessons learned are gathered to enhance future life cycle assessments. For this study, only the environmental life cycle assessment is of interest although ramifications from the results will impact system life cycle decisions.

The development of life cycle assessment methods arose during the 1970s with the 1990s seeing a concerted effort by numerous organizations to develop a standardized LCA methodology.⁵ The definition of a life cycle assessment is given by Rebitzer et al:³

“Life cycle assessment (LCA) is a methodological framework for estimating and assessing the environmental impacts attributable to the life cycle of a product, such as climate change, stratospheric ozone depletion, tropospheric ozone (smog) creation, eutrophication, acidification, toxicological stress on human health and ecosystems, the depletion of resources, water use, land use, and noise—and others.”

The framework for the life cycle assessment and the system boundary is shown in Figure 2.

Figure 2: Life cycle assessment system boundary

Figure 2 is a segmented and a compartmentalized view of a systems/products life cycle. As mentioned earlier, there is significant overlap between the stages in a life cycle assessment and a system life cycle. A key difference is that an LCA investigates physical consequences that may occur from decisions made during the conceptual stage. This relates back to Figure 1, where the majority of environmental burden can be quantified and source during early system life cycle stages.

To produce any product, energy is the primary concern. This energy includes all processes required to fabricate the product, ranging from mining operations to recycling and/or reuse. From these processes, the outputs will always include unwanted products such as pollutants, emissions, and other wastes. These outputs can be linked through a series of assumptions and calculations to quantified potential damages, those being disability adjusted life years (DALY), potentially disappeared species x year, and depletion of mineral resources (USD2013). Each of these measurements are commonly accepted indicators for damage to human health, damage to ecosystem quality, and damage to mineral and fossil resource availability.⁶ Of these categories, does supercritical fluid deposition from CO₂ reduce these environmental burdens?

Thermal deactivation or agglomeration of platinum particles over a catalyst’s life is a major contributor to performance loss.⁷ As such, the ability of supercritical CO₂ to deposit highly dispersed, strongly anchored nanoparticles is significant. The link between processes at the nanoscale and that at the macro-scale and mega-scale is illustrated in the figure on the following page from Charpentier.⁸

Figure 3: Relationship between design at the nanoscale and the mega-scale⁸

Figure 2 shows the relationship between nanoscale properties and how their manipulation ultimately affects global issues. A life cycle assessment can provide an environmental-oriented rationale for the adoption of supercritical fluid deposition at an industrial level for the synthesis of supported metal-oxide nanosystems.

1.1 Research Statement

There are several studies that identify supercritical fluid deposition using carbon dioxide as a “green” method for chemical processes, including catalyst synthesis.^9–12 However, there is no quantification of how “green” the process is. Much of the carbon dioxide produced currently is an inevitable byproduct from the combustion of fossil fuels and other carbon containing compounds. Realizing this, the material balance of any process that uses CO₂ as the solvent will inherently have near zero net carbon output and have little contribution to the total CO₂ already in existence. However, the complexity of a systems life cycle necessitates a larger view. To understand the complete picture of the potential environmental impact posed by a supercritical process, a life cycle assessment is necessary.

1.2 Research Hypothesis

Hypothesis: Alumina-supported platinum/ceria catalysts for water-gas shift reaction prepared from a supercritical fluid deposition system contribute less potential environmental damage than those prepared by incipient wetness impregnation

Three damage metrics define potential environmental damage for this study:

Disability adjusted life years (DALY)
Potentially disappeared species x year
U.S.A. Dollars (2013)

These units are common for life cycle assessment studies.

1.3 Research Objectives

Several research objectives frame this study:

Quantify the environmental emissions of a supported metal oxide catalyst prepared by supercritical fluid deposition
Quantify the environmental emissions of a supported metal oxide catalyst prepared by incipient wetness impregnation
Determine the potential environmental damage of these catalysts over their entire product life cycle using three metrics
- Disability adjusted life years (DALY)
- Potentially disappeared species x year
- U.S.A. Dollars (2013)

Summarized, the objective of this research is to evaluate the potential environmental impact and burden that a precious metal-based catalyst prepared from supercritical fluid deposition presents to the environment. Regardless of which solvent and synthesis path employed, it is impossible to have zero impact on the environment. However, by making a quantitative assessment of SCFD’s impact as an alternative route for catalyst synthesis, it is possible to reduce harmful environmental outputs while maintaining high catalytic activity. It is also possible to apply the results of this work to other metal oxide nanosystems prepared using supercritical fluid deposition. This study serves as a starting point and/or source of data for other similar LCA studies.

1.4 Research Contributions

There is a scarce amount of published work involving supercritical processes and their environmental impact. An LCA study with a gate-to-gate scope was performed in 2012 that compares alternative methods of antioxidant extraction from rosemary to supercritical fluid extraction.¹³ The supercritical solvent in this study was a combination of CO₂ and ethanol. Another LCA study, performed in 2016, cradle-to-gate only, uses supercritical CO₂ to prepare barium strontium titanate nanoparticles. Both of these studies draw similar conclusions, that the major contributor of environmental burden from a supercritical process is the energy requirements. This research is the first to include the performance of a product prepared from supercritical fluid deposition in a life cycle assessment.

In a broader sense, a life cycle assessment pinpoints areas where refinement and improvements can reduce the environmental impact of catalyst synthesis processes. The comparisons of SCFD to traditional catalyst synthesis methods provides an understanding of environmental trade-offs involved with transitioning to a novel synthesis approach. This research lends support towards SCFD adoption on a larger, industrial scale.

This research contributes to the growing field of systems engineering at the nanoscale, and provides an example of the interdisciplinary nature of systems engineering. Studying nanoscale interactions can benefit top-level design while minimizing global environmental impacts. Through an in-depth study of a particular catalyst nanosystem, the conclusions and lessons learned can be applied to other supported metal oxide nanosystems prepared from supercritical CO₂. There has not been a life cycle assessment, a cradle-to-grave study, of supercritical fluid deposition as an alternative synthesis process for platinum group metal-based catalysts. Cradle-to-grave analyses embody systems engineering, with frameworks existing that can be applied to manage any system.¹⁴ This research tackles not only work at the nanoscale, but analyzes how optimization of material use and “green” technology benefits local, regional, and global environmental concerns.

CHAPTER II

LITERATURE REVIEW

Chemical engineering has become much more complex with “safety, health, environmental aspects, including non-polluting technologies, reduction of raw material and energy losses and product/by-product recyclability” all being factors in system development.¹⁵ As stated by Charpentier et al., “simple economics are no longer the defining factor in the production of a product”.⁸ This increase in complexity requires a systems thinking or holistic approach. Questions such as “How does implementation of a ‘green’ synthesis pathway affect product and overall system properties?” and maybe more importantly, “Are these changes necessary and/or beneficial?” are necessary for process and product development.

The Intergovernmental Panel on Climate Change reports that it is “extremely likely that more than half of the observed increase in global average surface temperature from 1951 to 2010 was caused by the anthropogenic increase in GHG concentrations and other anthropogenic forcings together”¹⁶. Therefore, even more so for those in the engineering profession, reducing the carbon footprint and harmful environmental emissions from chemical processes is an important and morally justifiable goal. Furthermore, current emission and pollution policies, along with increased public awareness of environmental ramifications, dictate environmentally conscientious products and processes. However, before implementation of a “green” or novel process, there must be a quantifiable and demonstrable rationale.

The goal of this literature review is to provide background information on a number of areas of this research, from a broad system-level viewpoint, down to molecular level, nanosystem behaviors. The rise of green chemistry in chemical engineering and catalysis is reviewed, along with processes typically used today for the production of metal-oxide catalysts. The history and implementation of life cycle assessment to systems and its relevance to chemical engineering is investigated. For the application of the catalyst of interest, a case study involving hydrogen generation for use in a typical fuel cell vehicle is studied. Specifically, the manufacturing and use phases of the catalyst are reviewed, as a priori knowledge predicts they are the most affected by a change in the manufacturing process. However, because of the summative nature of the LCA, all areas of the system life cycle are addressed.

2.1 Green Chemistry & Green Engineering

Green chemistry is defined as:¹

“The use of chemistry techniques and methodologies that reduce or eliminate the use or generation of feedstocks, products, by-products, solvents, reagents, etc., that are hazardous to human health or the environment.”

There are twelve principles of green chemistry or “design rules” for chemists proposed by Anastas in 1998,¹⁷ with some referring to them as a “reaction toolbox”.¹⁸

Table 2: Principles of green chemistry

Prevention	Use of Renewable Feedstocks
Atom Economy	Reduce Derivatives
Less Hazardous Chemical Synthesis	Catalysis
Designing Safer Chemicals	Design for Degradation
Safer Solvents and Auxiliaries	Real-Time Analysis for Pollution Prevention
Design for Energy Efficiency	Inherently Safer Chemistry for Accident Prevention

These principles are viewed as a framework/guideline for chemical engineers. They look to maximize energy and material efficiency while simultaneously minimizing material waste, mitigating process risks, reducing harmful environmental impact, and using sustainable resources. These principles were eventually generalized for all engineering disciplines, producing the twelve principles of green engineering.²

Table 3: Principles of green engineering

1. Designers need to strive to ensure that all material and energy inputs and outputs are as inherently nonhazardous as possible.

2. It is better to prevent waste than to treat or clean up waste after it is formed.

3. Separation and purification operations should be designed to minimize energy consumption and materials use.

4. Products, processes, and systems should be designed to maximize mass, energy, space, and time efficiency.

5. Products, processes, and systems should be “output pulled” rather than “input pushed” through the use of energy and materials.

6. Embedded entropy and complexity must be viewed as an investment when making design choices on recycle, reuse, or beneficial disposition.

7. Targeted durability, not immortality, should be a design goal.

8. Design for unnecessary capacity or capability (e.g., “one size fits all”) solutions should be considered a design flaw.

9. Material diversity in multicomponent products should be minimized to promote disassembly and value retention.

10. Design of products, processes, and systems must include integration and interconnectivity with available energy and materials flows.

11. Products, processes, and systems should be designed for performance in a commercial “afterlife”.

12. Material and energy inputs should be renewable rather than depleting.

Utilizing these guidelines, the application of a life cycle assessment studies their effect on a process, product’s or system’s environmental impact.

2.2 Life Cycle Assessment

Traditionally, economic criteria have been used as the basis for process decisions.¹⁹ However, rising interest in sustainable energy sources has led to increased investment in “green” technology.²⁰ The application of a life cycle assessment to these systems is beneficial in identifying key areas for optimization and in investigating the potential of alternative technologies. It is most useful in situations where both upstream and downstream inputs/outputs are known and all areas of the system life cycle are quantified. Life cycle assessment (LCA) is a methodology that assesses the potential environmental impact of a process or product from cradle to grave. The purpose of a life cycle assessment is to aid in achieving three specific goals. Hofstetter lists them as:⁶

To generate as many services as possible with the least amount of products
To produce these products with a minimal material flow
To reduce as much as possible environmental interventions per industrial throughput

An LCA can identify areas for refinement with regard to emissions or pollutants and highlight areas for improved energy and material use.

There are a number of LCA studies that focus on process and products.^21,22 Some compare the environmental impact of alternative technologies to traditional, such as cement manufacturing that employs carbon sequestration,²³ while others address the environmental feasibility of existing technologies, such as hydrogen production from electrolysis.²⁴ An example in line with “green” energy is a review by Peng et al. of solar photovoltaic (PV) systems and their environmental emissions.²⁵ Solar PV systems are examples of a technology that conjures the idea of being inherently and completely environmentally friendly, specifically in the “use” phase, but expansion of the system boundary demonstrates this to be somewhat misleading when accounting for material extraction, manufacturing, disposal, etc. The principles of green chemistry do not always reconcile with a life cycle assessment. Figure 4 highlights this conundrum and why the entire life cycle of the product needs consideration.

Figure 4: Does adherence to green design principles imply less environmental impact?²⁶

Figure 4, from a study by Tabone et al., shows the relationship between environmental burden and adherence to green design principles for petroleum/biological feedstocks for plastic production.²⁶ Although a product may adhere to the principles, it is not necessarily the best environmental choice when looked at from a larger perspective. The complex relationships between systems and the environment has shown multi-objective trade-off analysis to be necessary when performing life cycle assessments, especially when attempting to maximize the performance of a product while minimizing the environmental impact or the economic costs.¹⁹

Measuring a system’s environmental burden can be daunting. This sometimes makes execution of life cycle assessments a challenging and uneasy task. This is partly due to the sheer magnitude of variables involved in a product’s life cycle, the needed assumptions, and/or recognizing the environmental primary, secondary, and tertiary effects. For these reasons, it is common to focus LCA studies on specific phases of the life cycle, depending on the goals of the analysis.²⁷

2.2.1 Life Cycle Assessment Metrics

Life cycle assessment has met criticism during its development due to the amount of subjectivity in its methods, particularly in the impact assessment, which establishes a criterion for how to quantify emissions and abstract environmental concerns to environmental damages. Cumulative energy demand and associated greenhouse gas (GHG) emissions (typically coupled),^28–30 energy payback time,^31,32 and environmental, health, and safety (EHS)³³ are all possible metrics utilized in studies, many times in combination to paint a clear and accurate picture. Rather than reducing the environmental impact to a single measure, which can mask or hide certain environmental concerns, the effort has been made to add multidimensionality by the introduction of impact categories. Examples include climate change, acidification, eco-toxicity, eutrophication, etc. These environmental concerns, lacking a numerical value, are instead classified and seen as “midpoint” indicators or abstract concepts that produce a consequence. Because of their inherent qualitative and conceptual nature, the introduction of “endpoints” is necessary. These “endpoints” are damage categories, the potential ultimate outcome of the accumulation of environmental emissions and disturbances.

The life cycle impact assessment (LCIA) captures a quantitative understanding of the life cycle inventory phase of an LCA, providing the assumptions and calculations necessary to link “midpoints” and “endpoints”. Software provides practitioners of LCA the tools necessary to conduct independent studies of their processes and/or products, identifying environmental “hotspots” and reducing inefficiencies. SimaPro, developed by PRé Sustainability, is one of the leading LCA software packages, used in both industry and academia.^{22,23,27,28,34–39} SimaPro makes use of life cycle inventory databases to complete life cycle assessments, and relies on several user-specified weighting methods to quantify results. The ecoinvent database is the current leading life cycle inventory database and is bundled and sold with major LCA software suites. Its transparency and comprehensive data make it a popular choice used in industry, with large companies such as Procter & Gamble and Nestlé implementing it in their own internal life cycle assessments.^40,41

There are numerous methods that analyze the life cycle inventory and establish links to environmental burdens and quantified damages, with some taking a midpoint approach and others looking into endpoint categories.^37,42,43 A well-developed method known as ReCiPe, attempts to harmonize the relationship between these two valid approaches.⁴⁴ Based on numerous assumptions and calculations, ReCiPe takes abstract environmental concerns, such as climate change and ozone depletion, and relates them to three quantified damage categories. These groupings include damage to mineral and fossil resources, damage to ecosystem quality, and damage to human health. The figure on the following page shows the relationships that ReCiPe establishes between the outputs of a product’s life cycle phases and the potential environmental burden.⁴⁴

Figure 5: Relationship between product life cycle emissions and environment burdens in ReCiPe 2016⁴⁵

Figure 5 draws the relationships between emissions, land use, and resource use. In this impact assessment method, there are three final, numerical scores given: damage to human health (disability adjusted life years), damage to mineral and fossil resources (USD2013) and damage to ecosystem quality (potentially disappeared fraction of species x year). It is important to note the role of subjectivity in this impact assessment. The sheer number of assumptions introduces a relatively large degree of uncertainty, increasing more so as the 17 midpoint categories are merged into three endpoints. The methodology makes use of cultural theory to address this, using three view different perspectives with different assumptions: individualist, hierarchist, and egalitarian. Each perspective differs in time horizon considerations and toxicity assumptions. Further information regarding this impact method and how it draws these relationships is available in the ReCiPe methodology report and the Eco-indicator 99 methodology report.^44–46

2.2.2 Challenges in Life Cycle Assessment

Life cycle assessment is a useful and powerful tool; however, it is not without its faults. There are two major areas for improvement: data uncertainty and modeling uncertainty. A critique by Ayres from 1995 cites a shortfall in life cycle assessment being the lack of verifiable, transparent, and published data.⁴⁷ Furthermore, it addresses the absence of a mass balance in life cycle assessment. These concerns still exist; however, work has been undertaken to mitigate, mainly in the expansive and transparent ecoinvent database which allows tracing of material and energy flows.⁴¹ A more recent critique (2008) by Reap et al. cites similar problems in life cycle assessment, the major concerns being the definition of environmental boundaries (truncation), environment uniqueness (non-homogeneity), and the subjective nature of choosing midpoint and endpoint indicators.⁴⁸ There can be disagreement between different software conclusions as well, demonstrated in a study by Herrman et al.³⁴ The study found variance between two leading LCA software packages, and attributed the difference to how the software handled impact assessment methodologies. There is ongoing work to reconcile and harmonize LCA studies, with attempts so far including ISO standards to standardize the methodology.⁴⁹ Ultimately, the researcher, consumer, or stakeholder dictates what metric is most appropriate and important, while understanding the limits and assumptions of the LCA. It is essential for practitioners to clearly state the assumptions made for their analysis. Life cycle assessment is a decision-supporting tool, not the ultimate dictator for what is the correct choice.

2.2.3 Life Cycle Assessment of Catalyst Nanosystems

Nanosystems share the same definition of a system but differ from micro and macro scale systems due to the unique properties they exhibit and the complex interactions that occur at their interfaces. As technologies have advanced, manipulation of the atomic structure of a nanosystem has become possible, allowing adjustment and tailoring of properties to accomplish a specific task. From an environmental perspective, this has resulted in less material use and subsequently, less waste and emissions. This level of precision and control has led to the use of nanosystems in several areas including drug delivery systems,^50,51 solar cells,⁵² separation processes,⁵³ and advanced materials.⁵⁴ Additionally, this progress has led to an increased understanding of molecular interactions at their interfaces, leading some to designate multicomponent catalysts as nanosystems.^15,54–58 With this progress, nanoscale work has been stressed in catalyst synthesis;⁵⁹ however, relatively few works address catalyst and/or nanotechnology from an environmental perspective.

It is essential to recognize that nanoscale improvements will ultimately affect a system as a whole. With an increasingly environmentally aware public and the need for safe processes, consumers and the market have new product requirements. Consequently, the need exists to take a multidisciplinary and multiscale approach to these changes, with life cycle assessment an established method to manage these complex interactions.^8,15 Some researchers have taken note of this, applying life cycle assessment to catalytic research at the nanoscale. Platinum group metal (PGM) reduction has been investigated in automotive catalysts using life cycle assessment.⁶⁰ The study is more of a “what if” scenario, drawing conclusions from the hypothetical continued improvement of nanotechnology, but provides ample reasoning for why PGM reduction is beneficial, important, and a worthwhile investment. However, this study is constrained to raw material extraction. Most other LCA studies involving catalysis follow this format, using life cycle assessment to gauge what performance metrics and properties future catalysts should exhibit.⁶¹ A study by Holman et al. looks at using life cycle assessment to guide catalysis research.⁶¹ The study compares a cheaper but less efficient feedstock, propane, to the current and high-yield process for producing propylene. The conclusions are interesting in that they do not necessarily support implementing propane, but provide a theoretical “break even” point for engineers to strive towards. Dealing with costs and environmental impact is an intricate and complex balancing act. By achieving a yield of 60% through catalyst improvement, compared to a yield of 87% for propylene, propane becomes economically viable while also having a lower environmental impact. It should be noted that at lower yields (6%), propane is not economically viable but is environmentally favorable, an example of how the implementation of “greener” processes can be stifled by economic constraints.

Life cycle assessment is also used to address material selection for composites and in solvent selection when multiple criteria are to be satisfied.^33,62 The environmental impacts of nanosilver t-shirts have been investigated, with the study going so far as to consider the potential health effects of the nanomaterials themselves.⁶³ Carbon nanotubes prepared from chemical vapor deposition have also been studied to identify environmental “hot spots”.²² The study addresses the lack of LCA studies on nanotechnology. The study cites that only 7 out of 16,500 hits for “nanotechnology” contained any mention of life cycle assessment. Furthermore, of these studies, much uncertainty was present due to lack of data concerning nanomaterials. This agrees with the findings that most LCA studies involving catalysis and new technologies are anticipatory only, with some addressing only certain phases of the life cycle due to life cycle inventory uncertainties.

When performing assessments involving nanotechnology, data are unavailable at times, either due to the novelty of the technology or because of industry patents. Although the exploitation of nanotechnology can reduce material use significantly,⁶⁰ the environmental and human toxicological effects remain difficult to discern, leading to a higher degree of uncertainty in the impact assessment.^63–65 It is crucial to identify assumptions made to correctly interpret results. Furthermore, when comparing nanomaterials with traditionally produced products, some recommend stressing the toxicological aspects of the nanomaterial over other categories, with a much higher performance of the nanomaterial required for a “go” decision.⁶⁶ Much of the data for this research was available within European and North American databases utilized by SimaPro, with the remainder collected in-house and from literature..

Literature shows that a life cycle assessment of the production of metal oxide catalysts, specifically the Pt-CeO_x/Al₂O₃ nanosystem via supercritical fluid deposition, has not been performed. In fact, complete life cycle assessments of supercritical fluid processes in general are lacking; surprisingly, there is an LCA study that has been conducted by Tsang et al. involving the synthesis of barium strontium titanate nanoparticles from supercritical CO₂.³⁹ The study is limited to a cradle to gate window, not taking into account the use phase or any end of life considerations. Supercritical fluid deposition is a viable alternative production method because of its ability to produce a high quality product, a recurring characteristic of processes that make use of supercritical carbon dioxide.

2.3 Supercritical Fluids

Supercritical fluids (SCF) are fluids subjected to pressures and temperatures beyond their critical parameters. In the near supercritical region, the SCF displays properties characteristic of both a liquid and a gas.⁶⁷ They possess densities in the range of liquids yet retain the diffusivity of gases. The ability to tune properties of the SCF towards more liquid-like properties or gas-like properties by slight changes in either the pressure or temperature extends their practicality. These unique properties make supercritical fluids an attractive means for catalyst synthesis.

Carbon dioxide is the leading solvent for SCFD because it is environmentally benign, non-toxic, nonflammable, and has a relatively accessible critical point. Rates of deposition are higher due to an increase in mass transfer, liquid waste is minimal, and the substrate contains no solvent residue.⁶⁸ Characteristics of supercritical fluids that facilitate deposition processes include the ability to dissolve a wide range of organometallic precursors, high mass transfer rates, zero surface tension, and miscibility with reducing gases.⁶⁹

2.3.1 Supercritical Fluids in Industry

Currently, the main industrial use of supercritical fluids, specifically supercritical CO₂, is within the food industry and extraction processes in particular.^70–72 An example is the relatable and well-known decaffeination of coffee beans and tea.^73,74 The drive towards supercritical CO₂ as the solvent of choice for these processes has been mainly from public pressure and environmental regulations and the desire to use more “natural” solvents.⁷⁵ Of the few economic analyses available, the main cost of implementation of these processes is the initial capital investment; however, this is mitigated by high product recovery and reduced wastes.⁷⁶ Furthermore, for many green processes, it has been shown that the main contributor of environmental burden is the primary energy source for the process,¹³ although this can fluctuate dramatically if the energy source is a mix and/or renewable.⁷⁷ These examples demonstrate the intimate connection between energy consumption, process efficiency, and waste disposal along with the added complication of temporal/spatial dimensions. It is vital to approach the development of new technology and chemical processes from a holistic, or rather a “systems thinking” mindset.

2.3.2 Synthesis of Metal Oxide Catalysts from Supercritical CO₂

Metal oxide catalyst preparation traditionally involves exposing support material to a solution containing the dissolved metals where the dissolved metals then adsorb to the surface of the oxide.^78–86 The need for a liquid solvent and post-synthesis processing leads to an increase in waste generation and energy use. Furthermore, the deposition of metals can be highly inefficient resulting in increased material use, lower performance, and possible overburdening during the disposal phase. The disadvantages of these methods are established.⁸⁷ Supercritical fluid deposition offers advantages and benefits that make it a viable alternative, both from environmental and performance perspectives for catalyst synthesis.

Synthesis of metal oxide catalysts using supercritical fluids has innate advantages over conventional methods. Preparation techniques used today are based on wet impregnation which leaves behind impure solvent and a product that requires additional chemical treatment.⁸⁸ Furthermore, this method shows poor control of noble metal particle size leading to inefficient material use and non-optimized performance. Comparing to SCFD with CO₂ as a solvent, liquid waste is non-existent and after deposition, there is no leftover solvent residue on the product itself. This eliminates the need for additional purification processes that may negatively influence the product. Lack of surface tension results in a more even distribution of noble metals on the surface of the metal oxide, resulting in efficient material use.⁸⁹ Solvation of the noble metal precursors is also important, with a number of organometallics possessing low solubility in scCO₂.⁹⁰ Supercritical CO₂’s miscibility with reducing gases alleviates additional equipment use as well.⁹¹

The typical solvent of choice for deposition on metal oxides is supercritical carbon dioxide. It is the leading solvents for supercritical fluid deposition because it is environmentally benign, relatively inert, non-toxic, nonflammable, and widely available. The accessibility of the critical point of CO₂is important as well, especially in the context of energy requirements. The use of CO₂ as a solvent will greatly influence the manufacturing/synthesis phase of the life cycle. The characteristics of catalysts produced from supercritical CO₂are without question one of its strongest attributes, ultimately playing a role in the use phase of the catalyst.⁹²

Particle size, metal distribution and the concentration of the metal being deposited are key parameters that influence the activity of the catalyst.⁹³ SCFD of platinum on an alumina support using supercritical CO₂ has been shown to result in consistent and even deposition of platinum nanoparticles with average particle sizes ranging from 1.2 – 2.7 nm.⁹³ Deposition occurred over the course of 24 hours at 27.6 MPa and 80°C in a sealed, stainless steel vessel. The authors note how the concentration of the metal affects the particle size and distribution with higher concentrations leading to larger particle sizes from increased occurrences of coalescence. The interactions between the support and the metallic precursor also appeared to influence the size of the particles and their distribution. Use of silica supports resulted in larger platinum nanoparticles and an inconsistent metal distribution compared with alumina supports. This difference was attributed to the strength of the interactions between the organometallic precursors and the support. SCFD of platinum on alumina foam-supported SnO₂ was performed to generate a catalyst for the preferential oxidation (PROX) of carbon monoxide, another important reaction for CO minimization in a hydrogen production system.⁹⁴ Compared to a similar catalyst produced by wet deposition, the SCFD proved to be the superior method as the maximum particle diameter exhibited was no larger than 7 nm, while particles on the catalyst prepared by wet impregnation had an average diameter of 27 nm. Furthermore, because the reaction takes place at the interface between platinum and SnO₂, catalysts prepared using SCFD exhibited higher activity than those prepared via wet impregnation, even at temperatures approximately 75°C lower. These literature examples were chosen because of the support material; supercritical fluid deposition is well-known to deposit dispersed particles on a variety of substrates.^89,95

These studies demonstrate that supercritical fluids are capable of synthesizing a metal oxide supported catalyst with properties advantageous for reactions. Efficient use of materials will reduce energy costs for all phases of the life cycle. Strong adhesion of the expensive metals to the support will extend the life of the catalyst, reducing waste. Whether or not these benefits outweigh initial investment in the process itself are dependent on the catalytic application. As a case study for the life cycle assessment, and keeping in line with the general “green” theme of this research, the reaction of interest studied is the water-gas shift reaction, operating within a hypothetical automotive fuel cell system.

2.4 The Water-Gas Shift Reaction

Stringent emission controls and environmental standards, coupled with a general societal trend towards “green” energy and sustainability in the past 20 years, have led to a renewed interest in both environmentally friendly processes and fuels. Of the fuels, hydrogen is seen as a leading alternative to fossil-based fuels.^20,96,97 Due to safety concerns involving hydrogen storage and its low volumetric energy density, on-board or localized, small-scale hydrogen generation is viewed as a possible solution. A typical liquid fuel processing system for a fuel cell and its relationship with the synthesis phase of the life cycle is shown in Figure 6.⁹⁶

Figure 6: Standard fuel processing system coupled with an SCF catalyst synthesis process

The use of supercritical fluid deposition to produce water-gas shift catalysts, specifically combinations of platinum and/or ceria on alumina, has been previously performed in our laboratory.⁹² Although other catalysts formulations exist for this reaction, combinations of platinum and reducible metal oxides show the most promise and are more studied.⁹⁸ A pseudo performance baseline can therefore be established for SCFD prepared catalysts.

2.4.1 Current Lab-Scale Performance of Pt-CeO_x/Al₂O₃ WGSR Catalysts

Utilizing a holistic mindset for this research, the use phase of the product (catalyst) was thoroughly studied. The previous sections show that high implementation costs of supercritical technology can be alleviated if the output product displays superior performance.⁹⁹ Recent published work using data collected in-house produced the following table (Table 1), comparing the performance of our in-house catalyst prepared by SCFD (final row) to those found in literature and prepared by other methods.⁹²

Table 4: Comparison of WGSR catalysts

There is a large degree of variation in catalyst composition and feed gas composition in the table. The turn over frequencies are adjusted to reflect the variation in catalyst and feed gas composition, based off a proposed rate law reported in the literature and the platinum content. The SCFD catalyst exhibits the third highest overall performance (2.4 mmol CO/g*Pt*s) but also surpasses other preparations in its turn over frequency.¹⁰⁰ These results demonstrate that SCFD is competitive with other synthesis methods from a product performance perspective. From economic and environmental viewpoints, it also shows the efficiency of SCFD in regards to material use and a minimization of post-synthesis solvent waste.

2.5 Conclusions

The literature shows a definite void exists regarding a true establishment of the environmental benignity of supercritical fluid deposition. No life cycle assessment has been performed for a Pt-CeO_x/Al₂O₃ catalyst prepared using supercritical fluid deposition. Furthermore, there is scarce information regarding the environmental impacts of supercritical fluid deposition. The ability of SCFD to produce metal-oxide supported catalysts is recognized and catalysts prepared by SCFD exhibit high performance relative to other conventional methods. Additionally, supercritical carbon dioxide as a solvent is already industrially implemented, exhibiting high product recovery and little waste. These observations justify and support this research.

CHAPTER III

APPROACH & METHODOLOGY

Before beginning the life cycle assessment, the life cycle inventory must be completed. Accurate information is necessary and any assumptions made must be clearly stated and justified. Lack of data requiring assumptions and data uncertainty has led LCA studies to be deemed unreliable at times, but the life cycle assessment framework is generally accepted as a reliable decision making tool.^38,101 The life cycle inventory includes all inputs and outputs for materials, energy, and any emissions/waste. Because the results of a life cycle assessment can vary based on what system boundaries are chosen,¹⁰² a clear and reasonable delineation must be made between the environment and the system life cycle. Furthermore, to properly quantify the environmental impact, a functional unit must be defined as a basis for calculations and comparisons. The work in this chapter includes the documentation of all necessary starting/exiting materials and calculation of the energy requirements for the synthesis phase of the system life cycle. To properly evaluate energy mixes/sources, a designated region of operation is chosen. Additionally, because this is a cradle-to-grave study, an assessment of the product’s performance for a system of interest is presented. Example life cycle assessment studies performed with SimaPro and their associated output are provided as an example of the final research deliverable.

3.1 Life Cycle Assessment Phase Descriptions and Data Collection

The boundaries for the life cycle of a product are relatively self-explanatory. Raw material extraction, processing of the materials, manufacturing, the use of the material, and finally its disposal are the phases that a product will go through in its lifetime. For this research, the challenge is not finding data concerning the acquisition of materials, their processing, or their recycling. The energy costs, along with the environmental impact, are documented and quantified for material acquisition, processing, and disposal phases. However, due to the novelty of the synthesis method, in-house data are required to complete the life cycle inventory. Furthermore, use phase data requires a scenario of operation for comparison with other, traditionally prepared catalysts.

3.1.1 Raw Material Extraction & Processing

Data that concerns the extraction and processing of raw materials did not deviate largely from what is contained in the database. The main area examined were the materials called for by the synthesis process. Switching to a new synthesis method involves the use of a different solvent, requiring the use of a different platinum precursor. For the SCFD pathway, platinum acetylacetonate was used. The alternative method, incipient wetness impregnation (IWI), required a platinum salt. Both pathways use -alumina as the support. Processing of raw materials did not differ. The source for CO₂ was assumed to come from an already established process where carbon dioxide is an unwanted byproduct, generated from the combustion of carbonaceous fuels or from ammonia production. Carbon dioxide is essentially “free” in this regard as only 50% of industry generated CO₂ is recovered.¹⁰³ Manufacture of the catalyst at the CO₂ source will reduce emissions from rail or road transport as well and reduce the expensive transportation of CO₂. For this reason, no transportation processes included for CO₂. When comparing SCFD to other catalyst preparation methods, material extraction and processing will not impact the overall conclusions due to the similar composition of both catalysts. It is important to note that infrastructure and capital equipment are neglected for this study as it has been to shown to be a minor contribution to the total environmental burden over a system’s life cycle.¹⁰⁴

3.1.2 Manufacturing/Synthesis

For any life cycle assessment, a functional unit is defined to establish a criterion for comparison of systems. For this study, 1 kilogram of a catalyst composed of platinum and ceria supported on alumina is defined as the functional unit and labeled as Pt-CeO_x/Al₂O₃. The final elemental composition represents a normal range found in the literature. This composition does not necessarily reflect the optimal formulation but serves as the basis for the life cycle assessment. This functional unit was also chosen for ease of calculations in the life cycle assessment software.

There is a scarcity of LCA data available concerning supercritical fluid deposition. Much of the data required was collected within the lab. Figure 7 shows the reactor used for lab-scale synthesis from supercritical CO₂.

Figure 7: Lab-scale SCFD catalyst manufacturing vessel

The life cycle inventory for user-defined materials can be found in the appendices. The materials chosen for the synthesis are based on their solubility in carbon dioxide.⁹⁰ This is an important aspect of solvent selection and certain precursors are easier and/or cheaper to manufacture. The concentrations of the reactants were chosen for comparison with those in the literature and do not necessarily represent the optimal ratio for catalytic performance. Previous in-house work has demonstrated that the deposition occurs with a ≈ 70% efficiency; however, this could be from non-optimal reactor conditions and other synthesis factors. Therefore, it was assumed that the reaction went to 95% completion in a larger scale operation. Calculation of the energy requirements required to bring the solvent to supercritical conditions were performed using REFPROP with a pump efficiency of 65%.¹⁰⁵ It can be assumed from the low concentrations of the input materials that all the energy supplied to the reactor goes towards heating of the CO₂ solvent. No post-synthesis processes are included aside from drying of the pellets. The detailed experimental procedure for the preparation of the catalyst, both from SCFD and IWI, can be found in the appendices.⁹²

3.1.3 Use Phase Description

The switch to SCFD as an alternative synthesis route is not strictly for environmental reasons. There are advantages in the product that SCFD generates over those from traditional methods, specifically a high dispersion and homogeneity.¹⁰⁶ Literature values for sol-gel and incipient wetness impregnation of Pt on ceria and/or alumina range anywhere from 5% to 69%, highly dependent on metal loading and post deposition processes.^107–109 High dispersion of the metal on the metal oxide typically results in smaller average particle sizes, which increases the performance/conversion and provides optimal material usage. For the use phase, there are two performance metrics of interest: the reaction rate per mass platinum, and the time on stream, which is essentially a measure of the durability of the catalyst. For comparison and to ensure appreciable activity, catalyst performance is measured at 300°C. An appropriate amount of catalyst was loaded, approximately 0.3 grams. Realistically, this temperature is acceptable for a fuel cell system that makes use of hydrogen from a reformate stream or that relies on methanol for hydrogen generation. The economic and environmental viability of the fuel cell system itself is not within the scope of this research, but an operability scenario is required to properly compare the competing products and complete the life cycle assessment.

The performance of an automobile using a hydrogen fuel cell system should be approximately equal to that of current gasoline fueled automobiles. For simplicity, 250 miles as an average distance the vehicle should travel is reasonable, which equates to 4 kg of hydrogen.¹¹⁰ Calculations can also be based on the fuel cell power output, making a reasonable assumption of 50 kW (small car) for the fuel cell automotive system.¹¹¹ The basis for calculations in this study is 50 kW.

Fuel cell operation requires pure hydrogen, but safety concerns and storage pose limitations; however, onboard hydrogen generation is a solution. Due to current infrastructure, natural gas is considered one of the top sources of hydrogen, although gasoline could function as a hydrogen source too. The use of hydrocarbons also presents difficulties with carbon monoxide being a fuel cell poison. The typical CO allowance is specified at 10 – 50 ppm.^107,112 This requires a cleanup step in the process (water-gas shift reactor followed by preferential oxidation), with a typical CO concentration exiting the WGSR of 1% by volume.¹¹³ Subsequently, there are two key requirements derived:

The outlet from the water-gas shift reactor shall produce a minimum of 50 kW of power.
The outlet from the water-gas shift reactor shall contain a maximum of 1% carbon monoxide by volume.

From these specifications, the system boundary was drawn (segmented line) and is shown below. Molecules are listed to show material flow.

Figure 8: System boundary for catalyst performance testing

The efficiency of the fuel cell is assumed to be 50%, requiring a hydrogen flow rate of ≥ 0.84 grams/second. Calculations for the required methane feed rate are based on a paper by Ayabe et al.¹¹⁴ and can be found along with additional assumptions in Appendix A. From the requirements, the only concern is making sure that the carbon monoxide level is reduced to ≤ 1% by volume. An interesting question arose from the calculations: is it better to generate the necessary hydrogen strictly from the reforming step or reduce the performance of the reforming step, allowing the WGSR to make up the difference in hydrogen generation? It was concluded that the performance of the reformer and the preferential oxidation step is outside the scope of the scenario, and the environmental burden of the fuel cell system is limited strictly to the WGSR catalyst and no other associated emissions. The requirements outlined are therefore restricted to conversion and performance metrics. Furthermore, the feed to the water-gas shift reactor contains reducing gases which affect metal dispersion, requiring a “catalyst lifetime” or “robustness” be specified as well.

“Time on stream” is a measure of the catalyst’s degradation over time as coalescence of the metals reduces performance. Mitigation and elucidation of sintering mechanisms is of major interest in the catalysis community. It is difficult to extrapolate and predict the longevity of nanoparticles due to complex coalescence mechanisms. Work has been done in an attempt to understand the sintering kinetics of nanoscale platinum particles through the use of electron microscopy.¹¹⁵ Hansen et al. studied the sintering of palladium nanoparticles on silica, concluding that the majority of sintering occurs from atomic movement, rather than particle migration.⁷ Because SCFD produces extremely small particles, some of them atomically dispersed, addressing the performance over time is mandatory. Industrially and in real world applications, the problem is demonstrated as well. For example, a study by General Motors shows how the dispersion of the metal decreases through the life of a catalytic converter.⁶⁰

Figure 9: Loss of exposed metal over time in a commercial automotive catalyst⁶⁰

Within 40,000 miles, the dispersion has dropped around 90%. While it is true that emission standards are still being achieved, a large portion of the initial material costs could be saved if the metal was more resistant to thermal and kinetic forces. Raising the temperature of the reactor results in an increase in the rate of reaction, simultaneously increasing the likelihood that particle agglomeration will occur and reduce the dispersion and performance of the catalyst. The same reaction performance can be achieved at a lower temperature if the number of active sites is higher. For this reason, the reaction rate per mass of platinum is an excellent method for quantifying performance. Time on stream studies are fairly common in the literature for catalysts similar the ones in this study,^111,116 however, there have been no studies regarding stability on those produced from supercritical CO_2,let alone under the same gas composition. Validating the catalyst over the same range and time scale of the General Motors study is impractical for this work, and inappropriate because of the high operating temperature experienced by catalytic converters, normally >=650°C. The operability scenario involves a fuel cell vehicle operating under reformer conditions, the composition of the feed determined from the autothermal reforming of methane; therefore, the temperatures experienced are much milder. During vehicle use, there are kinetic considerations to account for in addition to thermal effects. A reasonable assumption can be made that the thermal forces will have a much larger impact on metal dispersion compared to any kinetic forces due to the structure of the catalyst. The temperature of the reactor will therefore play the largest role in the lifetime of the catalyst, influenced as well by the composition of the feed to the reactor. A study by Phatak et al. from 2007 varied the temperature from ≈ 250°C to ≈ 320°C to investigate the performance of Pt/Al₂O₃ catalysts operating under reformer conditions for fuel cell applications.¹⁰⁸ These operating temperatures are consistent with those found in Table 3. It was assumed that heat integration within the fuel cell system allows the entering feed to the water-gas shift reactor not to fluctuate substantially from the desired operating temperature of 300°C. It was important in this instance to look to the principles of green engineering for guidance. It is recommended not to design for “immortality”, and also not to design for unnecessary performance. Subsequently, testing at abnormally high temperatures and harsh conditions is unnecessary due to the small likelihood of their occurrence. Assessment of the stability of the catalyst is performed at a variety of temperatures > 600°C to ensure degradation and the time was recorded for each measurement. The longevity is based on the loss of catalytic activity over this time period, producing a rate of dispersion loss (agglomeration) per unit time. Through the use of the Arrhenius equation, an activation energy and pre-exponential factor are calculated to estimate the average catalyst life-time at 300°C.

3.1.4 Regeneration, Recycling, and Reuse

The expensive and rare materials, mainly platinum, that make up the catalyst are designated as essential for recovery. With environmental pressure and increasing PGM demand, particularly in developing nations, reclamation is essential to meet future needs. Because of the catalyst’s three-component composition, it is relatively simple to recycle and feed back into the manufacturing phase. This is a safe assumption to make as the PGM (platinum group metals) recycling industry is well-established.¹¹⁷ There is no additional investment needed as the industry is equipped to deal with the catalyst compositions; however, this also assumes that the material flow rate of aged/spent catalysts is large enough to meet the production demand of new catalysts.

Aside from the costs, recovery of platinum instead of mining new material will lower the environmental impact significantly. This is true for the majority of materials; it is nearly always more energy intensive to mine and extract virgin materials than it is to recycle them. The creation of new mines, depletion of current reserves, and the additional need for PGM/ore separation all contribute greatly to environmental burden. This goes for alumina as well, where as much as 16 MJ are needed to produce 1 kilogram.¹¹⁸

Resistance to deactivation data collected from “use” phase testing is essential for describing the best end of life path the catalyst should take. Unless the dispersed platinum can be stabilized on the surface, regeneration and reuse of the product is deemed unlikely due to agglomeration and loss of activity. However, it remains a possibility to use the catalyst for a different function in the future, perhaps in a different reaction. Although unlikely at this point in time, the assessment could still be performed if these options are available in the future, possibly through a detailed understanding of the surface chemistry that occurs during the synthesis phase.

3.2 Life Cycle Assessment of Supercritical Fluid Deposition

The in-house data is aggregated and used to complete the input for the life cycle assessment. SimaPro is used to complete the analysis. ReCiPe 2016 is used as the impact assessment method and the ecoinvent database contained within SimaPro provides inventory data that is not collected in-house or from literature. To gauge the environmental burden of SCFD in relation to other conventional synthesis pathways, sets of catalysts prepared from conventional methods are investigated. The diagram below shows the pathway for the processes being compared and their damage assessment categories.

Figure 10: System boundary of proposed study for competing catalyst preparation methods

Some processes are shared because there is no difference in the emissions. In the event that a database value is unavailable to obtain the necessary information to complete the life cycle inventory, the literature is consulted or experiment performed.

3.2.1 Handling of Uncertainties & Sensitivity Analysis

The uncertainty in a life cycle assessment is inherently a challenge. Some of the data within the ecoinvent database is an average supplied from a single source or industry expert. There occasionally is no uncertainty associated with the mean value. The ecoinvent database uses a mix of calculated and qualitative estimates to gauge uncertainty in data. The reasoning behind the calculations of uncertainties can be found in the ecoinvent methodology report.⁴¹ When using ReCiPe, the uncertainty in the results depends on the cultural perspective used: hierarchist, egalitarian, or individualist. Each perspective comes with different assumptions for characterization and relationships between the life cycle inventory and midpoint/endpoint indicators. For this research, due to the number of variables being manipulated, a single cultural perspective is implemented. The hierarchist cultural perspective is chosen because of its balance in weighting environmental/societal needs and its wide use in the scientific community.⁴⁴ For comparison of the damage metrics between the two production pathways, Monte Carlo simulations were performed with a stop criteria of standard error of mean being equal to .005 (99% confidence interval).

Implementation of the catalyst in a system requires looking at the possible range of operating conditions it will encounter. This could be anything from a change in the feed composition to an increase and/or decrease in the reactor temperature. These operating conditions influence the lifetime of the catalyst, and thusly, have an effect on the life cycle assessment. Feed conditions were kept at the calculated ratios available in the appendices.

Looking at the life cycle of the catalyst overall, several scenarios were investigated for sensitivity analysis, as the selection of a single process can drastically alter the conclusions of the assessment. The key areas or “hot spots” to focus on include:

Platinum content in the nanosystem
Recovery of energy in the supercritical fluid deposition system

Region of operation can affect results due to different material production and extraction methods. For example, crude oil from Canadian oil sands results in 5 times greater greenhouse gas emissions than obtaining crude oil from Saudi Arabia/Iraq, a scenario where LCA results can be substantially affected.²⁶ However, as environmental damages are a worldwide concern, average global production and associated source-region of materials are assumed when available. For user-defined materials, United States averages are used. Because this research focuses on reduction in platinum use, only platinum was assumed to be recycled with a rate of 95% recovery. Furthermore, the overall elemental composition of the catalyst and with the needed infrastructure already in place, the recycling is assumed to occur in a spent catalytic converter precious metal recovery facility.

3.2.2 Accelerated Life Cycle Testing

Strongly anchoring metals to the surface of the support is key in the design of precious metal water-gas shift catalysts. This robust surface bond enables particles to resist sintering, facilitating high catalytic activity and a longer operational lifetime. The temperature, feed composition, and dispersion of the metal all play a role in the performance of the catalyst, with complicated relationships.

Figure 11: Influence of feed composition, dispersion, and operating temperature on performance

Due to the practicality of running the catalyst for an extended time, shorter run times at higher than the operating temperature (300°C) are used in what is essentially highly accelerated life cycle testing (HALT). These results are used to estimate the stability of the catalyst at normal operating conditions, and to extrapolate and gauge the lifetime of the catalysts. Nucleation was assumed to be the dominating factor in product degradation.

3.3 Conclusions

3.3.1 Hypothesis Validation

Every process has an associated uncertainty with the emissions that it produces. The ecoinvent database includes uncertainties and uncertainty distributions for the majority of process data. SimaPro is a powerful software and can repeatedly sample from each process’ distribution to ensure proper conclusions are drawn from a life cycle assessment. To properly compare the two product pathways, Monte Carlo simulations are performed to validate the impact assessment. For comparison, these scenarios are applied to two separate catalysts. This results in a Pt-CeO_x/Al₂O₃ catalyst prepared by two production pathways:

Incipient wetness impregnation of both platinum and cerium precursors
Supercritical fluid deposition of platinum precursor only, IWI of cerium precursor

The hierarchist cultural perspective is used to define assumptions and normalization of the damages is performed for ease of comparison. The hypothesis was tested for two product life cycle phases:

Product development phase
Use phase

Validation is provided by a Monte Carlo simulation of the entire product life cycle. To say that supercritical fluid deposition is a “greener” alternative requires that the SCFD catalyst surpass the traditional method in all environmental damage categories, as a consensus on the importance of each category is undefined and left to interpretation.

3.3.2 Summary

Going through the literature, there are few studies that address products throughout their entire life cycle, let alone studies that model the situation in this research. The vast majority of the studies encountered investigate the production of a simple solvent or product, such as hydrogen. All hydrogen behaves the same, whereas all Pt-CeOx/Al₂O₃ catalysts have the potential to behave differently. Despite possessing the same composition, the properties at the surface, such as metal dispersion, particle size and geometry, and intermolecular forces, can drastically differ depending on the preparation method. In turn, these nanoscale features influence the performance of the catalyst is affected, causing differences in the environmental impact. This study addresses these differences, taking into account the overall product life cycle of a precious metal-based catalyst prepared by two competing methods, a unique attribute not found in other literature studies.

The research methodology includes a laboratory phase and a simulation phase. In order to properly address the environmental burden, a full characterization of both products is performed. This data includes an elemental analysis, particle size analysis, and performance data. The data collected are used as input for SimaPro to assess environmental burden. The work in this section includes the majority of the necessary life cycle inventory for the synthesis and use phase of the life cycle assessment, along with a scenario of operation. Calculations and further assumptions are attached in appendices. Figure 12 on the following page presents how the research is organized.

Figure 12: Research organization

This chapter establishes a framework and outline for future life cycle assessment work. In general, the research is an environmental analysis of the introduction of a new and novel manufacturing method for a catalyst product. The manufacturing method utilized is supercritical fluid deposition. However, the area of study is not limited to this phase only. Introduction of this new method requires different starting materials because of new process constraints. Also, the product will have arguably superior performance and different characteristics, requiring a study of the use phase as well. End of life concerns should not be a large factor due to existing precious metal recovery infrastructure. With literature data indicating that supercritical processes usually recoup energy losses through superior product performance, it is vital to perform the life cycle assessment for the entirety of the product or process life.

Upon completion of this study, a framework exists for the life cycle assessment of other catalyst compositions prepared from supercritical carbon dioxide. The materials chosen for this study are not the only components being investigated, and with the use of software, studies on multiple compositions is easily accomplishable if performance and inventory data are available. The interrelationship of the emissions from a products life cycle and the cumulative energy demand of its production make an economic analysis a future avenue of study as well.

The work presented in this chapter demonstrates the unique nature of this research. Understanding the chemistry that occurs at the nanoscale allows guided and thoughtful optimization. Once the process is understood, how it influences the life cycle of the product as a whole can be determined. Life cycle assessment provides a utility to see how the manipulation of materials at the nanoscale propagates to the macroscopic level, where ramifications as large as climate change can be studied. This research exemplifies the need for a systems thinking approach in rational catalyst and chemical process design.

CHAPTER IV

ANALYSIS & EXPERIMENT

This chapter is compartmentalized into three sections with the groupings aligning with the simulation layout provided in SimaPro. Raw material extraction, processing, and manufacturing constitute one segment, known in SimaPro as an “assembly”. From a systems development life cycle view, this constitutes every phase up to and including development. To identify the important areas for optimization and for general intrigue, platinum content and energy recovery in the supercritical product pathway are varied to assess their contribution to the overall environmental burden of the product development life cycle. The use phase and end of life phase (testing and disposition) constitute the remaining two. Each section contains the life cycle inventory and any information obtained from laboratory experiment for both production pathways. Furthermore, it includes the damage assessment of each pathway. The final section contains the overall results of the entire life cycle for each nanosystem production pathway. General trends in reduction of platinum content and energy recovery in the SCFD production pathway are also presented.

4.1 Raw Material Extraction, Processing, and Manufacturing

The boundary of analysis in a life cycle assessment must be properly drawn for any meaningful analysis to occur. Possibly due to the relatively new nature of life cycle assessment, there was inadequate life cycle data in the literature concerning industrial preparation of metal acetylacetonates and metal salts. Therefore, an estimate was made for both platinum precursors along with the cerium precursor. Guidelines proposed by the ecoinvent methodology were followed to estimate the life cycle inventory for materials not found in the econvent database.¹¹⁹ Many papers reference a 2015 environmental report from Gendorf industrial park,^39,119 a multi-commodity production facility in Bavaria; however, it was decided to use data based on United States production capabilities. Each reaction was assumed to operate at 95% efficiency. When energy use for the production of a chemical could not be found, an average energy consumption per kg of material was based on United States energy consumption data provided by Kim and Overcash.¹²⁰ The average calculated was in agreement with the Gendorf industrial park figure. Transportation distances were assumed to be the average United States mix of truck and rail. These values were obtained from the 2012 Commodity Flow Survey published by the Bureau of Transportation Statistics and U.S. Census Bureau.¹²¹ Rail transportation accounted for the initial transport and truck was used for final delivery with the average distance for the “Basic Chemical” commodity used for all materials. Average water consumption was based on a 2002 figure from The Dow Chemical Company and was similar to the Gendorf industrial park value as well.¹²² All materials that were not created for this study follow a global average in regards to production emissions and transportation values provided in the ecoinvent database. Production of the cerium salt was assumed to be from dissolution of bastnasite ore in hydrochloric acid. The source of the ore was a global average, with a typical concentration of 50 wt % cerium oxide. The carbon dioxide used was assumed to be released to the environment at the point of production of the catalyst. For ease of calculation, it was assumed that the excess water used for dissolution of the metal salt was equal to the water emitted from the process. During the manufacture of the catalyst from supercritical CO₂, it is important to understand the interaction at the interface between the platinum precursor and the alumina surface. In order to dissolve in carbon dioxide, platinum must be bound to another component, in this case, acetylacetone. Essentially, acetylacetone “masks” the chemical properties of platinum in order for CO₂ to dissolve the metal. Upon contact with the alumina surface, the platinum precursor reacts and deposits on the alumina surface. Preliminary experiments in the lab have found that acetylacetone leaves the surface upon depressurization of the reactor. *insert appendices* Acetylacetone, or 2,4-pentanedione, is known to be a toxic emission, although some studies suggest it is biodegradable.¹²³ Regardless, from experiment, it is proper to include acetylacetone as an emission during the supercritical CO₂ synthesis pathway. The table below contains all user-defined materials for this study and important assumptions.

Table 5: User-defined materials and important assumptions for this study

User-defined materials	Assumptions
Acetylacetone	2.84 grams of water per gram of product¹²²
Cerium Chloride	80% of water derived from river source

A Comparative Life Cycle Assessment of Precious Metal Catalysts: The Effect of Production Pathway

LIST OF TABLES

LIST OF FIGURES

ABSTRACT