Economic and environmental evaluation of the EM-Ferro plastic sorting technology

............................................................................................................................. iii Table of

iv the plastic waste should be sorted into the different plastic types it is composed of.
Traditional plastic sorting methods used optical or manual sortation which was not only costly but also susceptible to high nonconformance rates. Newer technologies have been developed that have higher output and are more economically justifiable.
This thesis performs and economical and environmental comparison between five plastic sorting methods that are currently in use and one emerging technology which uses electromagnetic (EM) waves and ferrofluid to sort plastics. Economic input-output life cycle assessment (EIO-LCA) is the method used to collect the data and perform the analyses. Two economic measures are used to evaluate the methods from an economical point of view. The objective is to study how the new method compares to the existing methods both economically and environmentally using a case study in Toledo, Ohio. The data related to cost, energy requirements, and carbon emissions were collected through contacting local vendors. Also, sensitivity analyses were performed using different Recycling rate of MSW generated has increased from 9.6% to 34.5% in the same period of time. According to EPA, the MSW generated in the US is composed of 9 major To be able to be utilized, the outgoing plastic pellets must have a certain degree of consistency and purity. Therefore, plastic sorting technologies have enjoyed a lot of attention in the past two decades in order to yield more precise sortation technologies which lead to optimal purity and consistency.

Objective
The objective of this study is to primarily evaluate the EM-Ferro plastic The research carries out economic and environmental analyses using the estimates collected from vendors and provides rankings in three categories which might be utilized in public or private applications.

Current technologies for plastic sortation
There are currently five plastic sorting technologies in use by the recycling industry. These methods use different technologies to sort plastics. These five methods are known as electrostatic separation, the sink/swim differential method, surfactant based separation, near-infrared scanning, and ultrasound scanning [24]. This section will provide a brief overview of how these methods work. In addition, we will also introduce the new EM-Ferro plastic sorting method which uses electromagnetic waves and ferrofluid to sort plastic particles.

Electrostatic Separation
Electrostatic separation was first developed in the 1990's as an example of an electrically based solution to sorting plastic particles. In this method, plastic particles are first statically charged by friction and then introduced to an electrostatic field with two oppositely charged electrodes. Different plastic types tend to take different electrostatic charges as shown in table 1.1. The particles are then sorted in different bins based on their charges. There are six components in a triboelectric separator: a feeder system, a blower, a cyclone shaped tunnel for the triboelectric friction to occur, assorted containers for collecting sorted plastic bins, and two vertical-plate electrodes along with their accompanying DC power supply [4]. The system is installed inside a temperature and humidity control chamber whose duty is to keep temperature and humidity constant.
Plastic particles are fed to the tribo-cyclone using an air current provided by a blower. The air flow transports the particles into the tribo-cyclone and frictionally charges them by rubbing them against the inner lining of the cyclone. After enough time being statically charged in the cyclone, the air flow feed is switched off and the particles fall into the electrostatic field generated by the two oppositely charged electrodes. The particles are deflected towards the two electrodes based on their charge and then collected into the assigned bins. The particles with insufficient charge are collected in the central bins.

Sink/Swim Differential Method
The first studies on plastic flotation were carried out in 1970's. however, it has been receiving more attention in the past few years [5].
The main problem in this method is that different plastic types do not have a substantial difference in density. To address issue, two major solutions have been proposed. The first solution is to use liquids with pressure-variant densities. Pressurevariant liquids are liquids whose density changes relative to the container pressure. Being able to change the density of the liquid through changing the pressure of the container allows for a more accurate separation. This is usually done in a pressurized container and by setting the liquid density to specific values between the densities of plastic particles, higher precisions are attained. The second solution is applying this method in a multiplestage form where different stages are applied with multiple known-density liquids which results in the separation of plastic particles with closer densities [6].

Surfactant Based Separation
This method use a similar mechanism to the sink/swim method. The difference is in this method the plastic particles are first treated with surfactants or wetting agents and then are mixed with a liquid of known surface tension ( / ). After the system reaches a stable condition, air in introduced to a system through a pump. This method makes use of the concept of critical surface tension of wetting for a solid (γc). γc is defined as "the surface tension of the liquid at which the solid surface exhibits a hydrophobic to hydrophilic transition". The value of γc is different for different types of plastic. Air bubbles are interested to adhere to plastic particles with lower values of γc which causes them to float to the surface of the liquid whereas plastic particles with higher values of γc are reluctant to adhere to the air bubbles which causes them to sink to the bottom [6,7].
The method offers three advantages. The first advantage is that it does not require any particularly advanced technology. Secondly, the chemicals that are used as reagents are often used in chemical processing are not environmentally unfriendly. The third benefit is that this method resolves the problem of the sink/swim's method inability to separate certain types of plastic with overlapping densities such as PVC and PET.

Near-Infrared Scanning
Near-Infrared (NIR) scanning systems are used widely in recycling industries.
NIR separation systems are macro-sorting systems i.e. they work with bulk sizes of plastics such as bottles, containers, etc. The vision machine of the system can identify almost every type of plastic resins [9].
In NIR plastic sortation, plastic parts are moved one by one on a conveyor belt where the NIR detection systems uses high-speed cameras and infrared light to specify the resin color and density spectrometry of the parts. The mechanism makes use of the fact that each plastic type return certain wavelengths after being exposed to infrared light.
By analyzing the returning wavelengths, the system identifies the plastic type. After identification, the parts are optically marked and later removed from the system using ejectors such as high pressure air jets. The main benefit of this method is that surface contaminations of the parts do not hinder the system performance.

Ultrasound Scanning
One of the most recent technologies utilized in plastic recycling is ultrasound scanning. Similar to NIR systems, ultrasound systems also identify the plastic types through exposing the plastic samples to ultra-sonic waves in water. After identification, the plastics are removed from the line into their respective collection bins. This is typically carried out by a mechanical arm. Ultrasound scanning offers two advantages compared with other plastic sorting methods. Unlike other optical technologies, the ultrasound scanning system can identify plastic densities in non-clear liquids such as ferrofluid. It also builds a 3D-image of the parts which is a unique feature to this approach.

EM -Ferro Technology
This method is recently developed by a team in MIME department at the University of Toledo. EM-Ferro method uses electromagnetic waves to separate plastic particles suspended in a ferrofluid. Ferrofluids are colloidal liquids in which tiny particles of iron, magnetite or cobalt are suspended. The carrier liquids are usually oils, water, acids, or organic solvents. The interesting property of ferrofluids is that, when exposed to electromagnetic waves, they form temporary domains and thus, the viscosity of the fluid changes. The ability to change the ferrofluid's density through changing the intensity of the electromagnetic waves allows for complete control over the process.
By altering the density of the ferrofluid to specific values between the densities of different types of plastic, one can determine which plastic particles rise to the surface and which sink to the bottom. While the electromagnetic field is active, the sorted plastic particles are collected. As soon as the electromagnetic field is switched off, the liquid returns to its normal situation and this allows for further separations.
In this method, an iron core electromagnet is used to generate the electromagnet field and the process is done inside an antistatic acrylic vessel. Being new, this method needs more experiments and modifications to reach its optimal performance and throughput.

Life Cycle Assessment (LCA)
1960s were probably the first years in which the western scientific community really gained vision into the fact that the Earth's resources such as oil, coal, minerals, etc.
are not infinite. The increasing rate of resource extraction as well as the pollution and detrimental environmental effects caused by its consumption gave rise to the question of can we be more environmental friendly? It was in the 1960s when the importance of environmental considerations in manufacturing systems was realized and issues such as resource and energy efficiency, pollution control, and solid waste were brought to public attention. As a response to this concern, in late 1960s and early 1970s, scientists in Europe and USA started developing a method for assessing all of the environmental impacts associated with a product or service during its entire life cycle [10].

LCA Definition
According to ISO, LCA is defined as: "Compilation and evaluation of the inputs, outputs and the potential environmental impacts of a product system throughout its life cycle" (ISO 14040, 1997). SETAC has defined LCA as: "A process to evaluate the environmental burdens associated with a product, process, or activity by identifying and quantifying energy and materials used and wastes released to the environment; to assess the impacts of those energy and material uses and releases to the environment; and to identify and evaluate opportunities to affect environment improvements. The assessment includes the entire life cycle of the product, process or activity, encompassing extracting and processing raw materials; manufacturing; distribution and transportation, use, re-use, maintenance; recycling and final disposal" LCA is a tool through which we can assess all of the environmental impacts of a product, process or service throughout its life cycle. LCA help us do this assessment through: 1. Building an inventory of all the energy and materials input and environmental consequences.
2. Assessing the potential environmental impact of all of these inputs and consequences.
3. Interpretation of the results to help the decision making process [13] The precise identification of inputs, outputs, and what is considered as life cycle is an important part of LCA. We should make sure to include all of the data related to relevant inputs, life cycle, and outputs in our LCA study [10,13]

Goal and Scope Definition
The goal and scope definition phase is of great importance in every LCA study. In this phase, the reason as well as the depth, context, and audience of the LCA study is defined and elaborated [10]. The scope of the study has to be well defined to ensure that the breadth and depth of the analysis are compatible with the stated goal, and sufficient to address it. There are several relevant concepts that need to be identified in goal and scope definition phase which are as followed:  System boundaries (technical, geographic and time)  Functional unit (or reference function)  Rules and assumptions  Type of impact assessment and valuation  The groups to be addressed by the study  Peer (Expert) review

System Boundaries
In the first phase of the LCA, the boundaries of the system that is going to be studied should be clearly defined; boundaries such as:  The boundary between the system under study and the environment  The boundaries between the system under study and other systems and  The geographic and time boundaries. Infrastructure, technologies available, and existing ecosystems are some of the factors that affect the geographic boundaries of a system. Also, time horizon is affected by pollutants' life span and other things such as technologies involved. [14] As shown in figure 2.2 the outline of the box denotes the 'system boundaries' and includes the 'foreground system' (what is studied) and its surroundings called 'the background system'. The background system is the source of all inputs to the system and the sink for all outputs from the system (Beukering et al., 1998). In the case of a waste management system the foreground would be the processes within waste management (collecting, sorting and disposal). The background system would include processes such as grid electricity production and raw materials. The boundaries exclude the use of packaging before it became waste. The system boundaries need to reflect the inputs and outputs that will be included in the inventory. These will be determined by the environmental issues a study wishes to address.

Fig 2-2 System Boundaries for a typical LCA study Functional Unit
Functional unit is the measurement unit that is used in an LCA study so that a specific system or product can be compared to another system or product based on their environmental impacts and energy efficiency. It has to b clearly defined, measurable, and relevant to input and output data. For example, the functional unit for a paint system can be defined as the unit surface protected for 10 years or the functional unit for a refrigerator can be "a refrigerator" or "a refrigerator year" depending on the perspective of the LCA practitioner.

Data quality requirement
These requirements are defined to enable the goals and scope of the LCA study to be met. The data quality requirement should address: time, geography and technology coverage; precision, completeness and representativeness of the data, as well as consistency and reproducibility of the methods used throughout the LCA. Data sources should also be acknowledged with their representativeness and/or uncertainty (ISO,

Inventory Analysis
Inventory analysis is the most developed as well as the most resource intensive part of LCA [10,13]. In the inventory analysis phase, all of the inflows and outflows generated by the system or product under study are identified and the data about each of these flows is collected per functional unit. Relevant data can be collected from several resources such as manufacturers' databases, government databases, previous LCAs, surveys and audits, test results, etc. There are several software packages and databases that can be accessed for data collection [13]. The important issue in data collection is that the data should be current and precise to avoid data gaps because data collected from one or several manufacturers or data collected from one specific country or region might not be a good representation of data required for another LCA study. So the validity and quality of data is of great significance.

Fig 2-3 Flow Diagram of Life Cycle Inventory according to SETAC
The results from this phase are presented in an "Inventory Table". Inventory table depicts all of the inputs and outputs per functional unit [9]. It generally includes inputs such as raw materials, energy, and transportation, outputs such as air emissions, water waste, and solid waste and other factors such as land use [10,13].
The total life cycle inventory is defined as: -Direct burdens: arise from the operations being studied (foreground system), and could be directly affected by the decisions based on the study.
-Plus indirect burdens: arising in the supply chains of materials and energy provided to the activity being study (background system) and are not directly affected by the outcome of the study.
-Minus avoided burdens: associated with economic activities displaced by material and/or energy recovered.

Life Cycle Impact Assessment (LCIA)
Life cycle impact assessment (LCIA) is the phase in which the potential impact of each of the aforementioned inflows and outflows is calculated. LCIA can serve as the basis for the comparison between different production systems based on their overall environmental impact. LCIA associates each input and output with a particular environmental issue (e. g. global warming), and evaluate the significance of potential (not actual) environmental impacts. The level of detail, choice of impacts evaluated and methodologies used, depend on the goal and scope of the study. There are no generally accepted methodologies for consistently and accurately associating inventory data with specific potential environmental impacts.
The need for impact assessment depends on the purpose and results of the study.
In a comparative study, it may be that one alternative is better than all other alternatives on all environmental burdens in the inventory, and thus the conclusion is easy to reach.
However, more often than not, one scenario will do better on some environmental burdens but worse on others, thus it might be desirable to attach some degree of importance to the environmental burdens. According to ISO, LCIA consists of three mandatory and three optional steps. The three mandatory stages are as follows [14]: -Selecting and Defining the Impact Categories: in this step, a list of environmental impact categories associated with the system is created. Impact categories can be divided into input related categories and output related categories. Table 2.1 indicates a list of categories adopted by SETAC-Europe. -Characterization: in this step, the results obtained from the last phase are converted into a common unit of measurement within each category for the sake of simpler aggregation.
Through characterization, one can assign an impact indicator to each category so we can better compare different systems. The formulation used to calculate an impact indicator is as follows: Inventory Data × Characterization Factor = Impact Indicator Characterization factors are constant numbers that are defined for a specific amount (e.g. one pound) of the given item. For example, Methane has a Global Warming Potential (GWP) of 21 per pound so if 10 pounds of Methane is released per functional unit, the total GWP for Methane equals 10*21=210. We can find a uniform unit for all of the greenhouse gases using the same approach.
The three optional steps are defined as follows: -Normalization: normalization is converting the potential impacts obtained from LCIA to a unit which provides us with the possibility of comparing different systems together.
For example, we can construct the ratio of the product's GWP (Global Warming Potential) to the national or regional GWP and then compare the products.
-Grouping: it is the organizing of the impact factors based on geographical realm, company's priorities, etc.
-Weighting: is the ranking of potential impacts based on their importance. The problem with weighting is that it is subject to subjectivity because the ranking system used depends on several factors such as stakeholders' values, geographical conditions, the scale that is used, etc. For example, the importance of the "noise" factor can be much different in an urban location compared with an out-of-town location.

Interpretation
Interpretation is the final phase of a LCA study. In this phase, three major steps are performed [14]: LCA is a tool designed for studying and comparing different products, services or processes based on their overall environmental impact throughout their life cycle. The ultimate goal of an LCA study is to help decision-makers select the better system from an environmental point of view. It is important to note that LCA cannot identify the best overall system by itself rather it should be used in simultaneity with other decision making tools because like any other tool, LCA has its own limitations. One of the most important limitations LCA has is that it does not take factors such as cost, performance or social and political factors into consideration e.g. a system with a lower overall environmental impact could be much costlier to set up. Also, one should be aware that the results of LCA studies can differ greatly based on the geographical region in which they are performed. For example, a LCA study done in the US might not be of so much help to practitioners in the Middle East. Therefore, it is essential to use LCA in the right context.

LCA in Waste Management
An Finnveden (1995,1999) discusses five important issues that should be considered in an LCA application to waste management systems [15,17]. He argues that in order to be compatible with LCA definition, the system boundaries should be defined in a way that all products are "identical" in all of the systems. If so, one can exclude the stages that are common between all systems from the analysis. They also state that multi-input allocation is another issue that should be paid attention to. They argue that in performing life cycle inventory for different municipal waste management, the allocation process should be clear as to which emission or energy inputs and outputs are related to which material. Building more on this point, J.R. Barton (1996) introduces a "dual classification approach" for the LCI phase of an LCA for mixed waste management systems in which the input/output data is divided into two categories: waste -independent and wastedependent [16]. The paper argues that the potential impacts associated with different types of materials present in the mixed municipal waste should be identified and properly allocated to the different materials. Speaking more to this point, CLIFT et al. (2000) also discuss the problem of allocation in LCA applications to Integrated Waste Management (IWS) systems. They argue that there should be a clear distinction between "wasterelated" emissions which depend on the waste composition and "process-related" emissions which are produced directly by the waste management operations [18].
The issue of allocation in applying LCA to integrated waste management systems has been the subject of much debate. To avoid the complications arising from studying mixed waste streams, scholars have been trying to apply LCA to material-specific treatment options where possible i.e. they study different waste treatment systems for one specific material type. The emergence of material-specific recovery methods has enabled researchers to avoid the allocation problem and have a clear-cut inventory analysis. Molgaard (1995) used LCA to compare six alternatives for recycling plastics.
Material recycling with separation based on vision or chemical analysis, material recycling with separation based on selective dissolution, material recycling without separation, pyrolysis, incineration with heat recovery, and landfilling were the six methods studied in Molgaard's work [19]. Molgaard used the concept of "Eco-profile" in his study which is an approach based on LCA that provides the ability to rank the alternatives based on certain impact categories. The impact categories used in this study were environmental effects (gaseous emission) and resource consumption (crude oil, natural gas, and pit coal).
Several other LCA studies have been carried out for plastic recycling systems.
Two case studies for plastic packaging wastes carried out by Arena et al. showed that recycling scenarios offer considerable environmental benefits compared to non-recycling scenarios [25,28]. A similar research has been done by Wollny et al. using a case study in Germnay [29]. Li Shen et al. [20] performed an LCA study on four different PET bottles to fiber recycling systems (mechanical and semi-mechanical). The study shows that all of the studied recycling systems offer substantial non-renewable energy and global warming potential (GWP) savings.
Another flaw associated with LCA that has been discussed extensively is its inability to incorporate economic and social impacts into its analyses. Assies argues that Simple economic evaluation that includes costs and benefits of different scenarios has been carried out in several works [26,27,29]. These works consider land acquisition costs, transportation costs, equipment costs, etc. as negative costs whilst considering revenues generated by selling the recycled materials and savings in energy and raw material as benefits.
Craighill and Powell (1996) utilize a new method named "Lifecycle Evaluation" in their study [21].

Region
The region chosen for this study is Toledo, Ohio. Toledo, Ohio is a city located in Northwest Ohio with a population of 284,012 and expands over an area of 217.87 km 2 .
Waste collection data for the region was collected from the Lucas County Solid Waste Management District (LCWMD). There is currently 20 drop off sites operating within the city of Toledo. The LCWMD facility collect recyclable materials such as plastic, paper, cardboard, etc. from the drop off sites and deliver them to recycling facilities after sortation.
LCWMD collects 8,700 metric tons of solid waste annually. Mixed plastic waste composes 90% of the total solid waste collection which is equal to 7,850 metric tons.

Life Cycle Analysis
The six plastic sorting technologies are compared using Economic Input-Output Life Cycle Assessment (EIO-LCA). The Economic Input-Output Life Cycle Assessment (EIO-LCA) method uses the materials and energy resources required as inputs, and the environmental emissions resulting as outputs with the associated costs.

Scope
The system boundary for this study encompasses the installation/manufacturing processes and the transportation processes to/from the facility. This allows us to focus on economic, energy, and environmental aspects of each system and disregard the communal stages between the systems. Figure 3.1 displays the stages considered in this study.

Functional Unit
In LCA studies, the functional unit serves as a reference for comparison between multiple systems. Functional unit should clearly define what is being studied and must be selected in a way that all of the inputs and outputs can be assigned with regards to it.
The functional unit chosen for this study is the total amount of plastic waste generated in the region which is equal to 7,850 metric tons of comingled plastic waste.

Life cycle inventory
Energy requirements and carbon footprint are the two main impact categories selected for this study. Carbon footprint data include carbon emissions incurred from installing and operating the manufacturing systems as well as the carbon emissions generated by transporting the waste from drop-off sites to the sorting facility for a 10 year period.
Energy data include the equivalent amounts of energy required to set up and operate the manufacturing systems in addition to the energy required for the transportation processes in each of the manufacturing systems over a 10 year period. For better comparison, payback periods are also calculated for each scenario.
Payback period is the period of time in which the initial investment is compensated for through the aggregation of revenues. Economically, shorter payback periods are preferable since it means that the system regains its initial costs in a shorter amount of time. The formulae for calculating the payback period in the case of even cash flows are as follows:

= ℎ
Where: In this study, we have considered one year as the period used to calculate the payback period. For example, in the first scenario, we have an initial investment cost, annual benefit, and annual cost of $3,100,000, $2,355,000, and $1,800,000 respectively.
Considering one year as the period, we have:

Sensitivity Analysis
Sensitivity analysis is a technique used to study how changes in different inputs of a system affects the outcome of that system. It can be used to study the relationship between different variables in a model. In finance, sensitivity analysis is widely used to assess the economic attractiveness of different financial opportunities. It is often used to study the different scenarios that could occur if the actual outcomes or financial requirements of a project is different compared to predictions. It is vital to conduct sensitivity analyses beforehand for different predictions to make sure that the project remains in company's acceptable financial definitions even if the actual costs or revenues are different than the predictions.
In this study, it is assumed that a potential investor is planning to use the data to invest in one of the six scenarios. Therefore, it is necessary to provide a more comprehensive overview of how changes in costs or revenues can change the attractiveness of a scenario so all of the risks can be accounted for before making a decision. In order to reach this comprehensive overview of the six scenarios, different sensitivity analyses were performed to study the range in which each scenario stays financially attractive if the actual parameters at implementation are different from the estimates made through the data gathered from the vendors.
Initial investment cost, annual operating cost and the annual revenue from the system were considered as the main changing variables and the effects of changing then were studied using economic attractiveness measures (payback period and internal rate of return) as the target variables.
In order to be able to better compare the systems, the least financially attractive scenario (scenario 1) was used as the base of comparison. The payback period of scenarios one (5.64 years) was considered to be the Maximum Acceptable Payback Period (MAPP). It was assumed that a higher payback period was not acceptable for a potential investor. for each scenario were collected through contacting local vendors and getting their estimates. It is assumed that the waste is delivered once daily to the sorting facility from each of the 20 community drop-off sites whose distance from the sorting facility ranges from 5km to 20 km.

Calculation and Assumptions
The emission data is reported in million metric tons of carbon dioxide equivalent (MTCO2E). MTCO2E is a measure that aggregates various greenhouse gas emissions into one single measure and is widely used in LCA studies as a measure to compare the global warming potential (GWP) of different systems.

Scenario 1: Electrostatic Separation
The data were collected by contacting two local vendors and getting their estimates. The average initial cost reported by the vendors was $3.1 million which is the cost of purchasing the required buildings and equipment for a processing plant with a capacity of 10000 metric tons per year. Additionally, it was estimated that 1.48 GJ of energy would be required to set up the system while 288.01 MTCO2E would be generated during the process in terms of carbon emissions. The average annual cost of operating the system was estimated to be $1.8 million which includes the salaries and benefits for two administrative staff and ten workers as well as the system material,  It should be noted that the energy use data for transportation is excluded because it is a communal stage for all of the scenarios. Figure 4.1 compares the internal rate of returns of the six scenarios. As discussed before, IRR is a measure that is used to evaluate and compare the financial attractiveness of projects. Obviously, scenario 4 and 6 are the most attractive ones from an economic standpoint whereas scenario 1 and scenario 5 have the weakest financial performance which is partly due to the larger initial investments that are required to install the systems.

Economic Analysis
Scenario 4 also has the least annual operational cost while the largest annual operational cost belongs to scenario 3 which is due to the operational (surfactants) requirements.
Ranking of the scenarios in terms of financial attractiveness from highest to lowest is Scenario 4 = Scenario 6 > Scenario 2 > Scenario 3 > Scenario 5 > Scenario 1.

Sensitivity Analysis
Sensitivity analysis is a technique used to study possible future changes in different variables and predicting the effects of these changes on the target variable. In this study, to better understand the effects of potential changes on each scenario's financial attractiveness, three types of sensitivity analysis were performed for each scenario using different parameters of Table 4.1 as variables. The data from Table 4.1 was used as the base amount for each of the variables and the changes were studied relative to these amounts.
In the first two groups of sensitivity analyses, "Initial Investment Cost" and "Annual Operating Cost" were used as the two changing variables. A range of -20% to +20% was considered for each variable with 5% intervals. These changes could realistically reflect changes in costs associated with the initial or annual cost such as energy, material, transportation, etc. Given the nine different amounts for each of the variables, the corresponding Payback Period (PP) and Internal Rate of Return (IRR) for each possible combination was calculated in the first and second group of sensitivity analyses, respectively.
For the third group of sensitivity analyses, "Annual Amount of Plastic Collected" and "Annual Operating Cost" were considered the changing variables, Again, a range of -20% to +20% was considered for each variable with 5% intervals and the relative PP was calculated for each of the combinations. The change in the amount of plastic collected each year could be a result of potentially different waste collection systems, higher or lower budgets for waste management across the region, or even a change in the population of the region. It is important to note that for this study, the selling price for one metric ton of sorted plastic was assumed to be constant. A change in the amount of plastic collected affects the annual revenue from the system which in turn changes the cash flow of the scenario each year which leads to higher or lower payback periods.
For the sake of comparison, for each of the scenarios, the payback period calculated from the data obtained from the vendors (given in Table 4.1) for scenario 1 was assumed as the Maximum Acceptable Payback Period (MAPP) which is equal to 5.64 years. MAPP is a measure defined by a company internally and is considered the highest payback period that is acceptable for the company in a potential future investment. Obviously, projects with lower payback periods than the MAPP are more attractive while higher payback periods mean less attractiveness from an economic point of view. In the following sensitivity analysis tables, the payback periods that are more than their respective MAPP are highlighted with red as "not-attractive".

Discussion
Sensitivity Analyses provide decision-makers with valuable information about how changes in different variables can affect the attractiveness of a potential investment opportunity and how to deal with the uncertainty.
The results of the sensitivity analyses performed were shown in the previous sections. It is important to note that the red-highlighted section of the above tables shows unattractive scenarios while considering a MAPP of 5.64 years i.e. they are economically viable but are not attractive for the purpose of this study. The darker red sections of the results represent scenarios that are economically not-viable meaning that the initial investment cost will never be recouped through aggregation of revenues.
The first group of sensitivity analyses provide information about the sensitivity of each of the scenarios to a decrease/increase in initial or annual operating costs. The general trend of the analysis is in line with our expectation which was that lower costs would lead to shorter payback periods and more attractive projects; however, the results of the analysis shows that, in the range studied, all scenarios show limitations on how much the annual operating cost could be increased while keeping the scenario economically attractive. An increase of 10 percent or more in scenarios 2 and 5 and an increase of 15 percent or more in scenarios 2, 4, and 6 would render the systems economically unattractive. Overall, scenario 4 shows the least sensitivity to changes in annual operating costs while scenario 5 is the most sensitive.
Considering the initial cost of investment, the results show that most scenarios show positive performances when the initial cost of investments increases as long as it is accompanied by a reduction of at least 5% in the annual operating cost. Therefore, it can be recommended that while given the option, for all scenarios, it is more feasible to increase the initial investment cost to avoid increasing the operating cost e.g. buying newer, more energy-efficient equipment even if they are more expensive.
The second group of sensitivity analyses considers the same variables as the first group but study the effects of the changes on the IRR of each scenario. The results are congruent with our prediction and the results of the first group of analyses that is less cost leads to higher payback period and higher rate of returns.
The third group of analyses incorporates the annual amount of plastic collected as one of the changing variables. The changes in the amount of plastic collected could be the result of different recycling policies and budgets or different plastic collection methods.
The results of the analyses show that all of the scenarios are very sensitive to the amount of plastic collected annually. A decrease of 5 percent or more in the amount of plastic collected without decreasing the annual operating cost of the system results in scenarios with payback periods that are higher than the MAPP and therefore are financially unattractive. Decision-makers and investors should watch the regulations potentially affecting the collection rate very closely so they can be prepared for such a fall. Overall, scenario 4 is the least sensitive to the amount of plastic collected annually whereas scenario 5 shows the most sensitivity.

Conclusions
This study compared six emerging post-consumer plastic sortation technologies in the three categories of cost, energy, and CO2 emissions via a case study. The data obtained were for 7,850 metric tons of plastic that is collected annually in the city of Toledo from 20 drop off sites. The method used in this study is Economic Input-Output LCA (EIO-LCA) method which provided a comprehensive account for all of the related initial and startup cost, energy, and CO2 emissions data for each of the six scenarios for a 10 year period lifetime. In order to compare the scenarios from an economic point of view, IRRs and payback periods were calculated for all of the scenarios. The economic analysis showed that all of the scenarios have positive economic performances with IRRs ranging from 12% to 19%. Over a 10 year period, scenario 4 (infrared) and scenario 6 (EM-Ferro) are the most preferred scenarios from an economic standpoint with IRRs of 19% (Table 4.1).
The energy data showed that scenario 2 (differential sink/swim) uses the least energy over a 10 year lifetime. With regards to carbon emission, all pf the scenarios have similar performances with total emissions over 10 year time ranging from 344.44 MTCO2EE to 389.2 MTCO2EE. From a mere environmental point of view, scenario 2 has the best environmental performance.
The results also show that scenario 6 (EM-Ferro) offers substantial cost and environmental benefits on a larger scale and it can compete with other technologies that are currently in use. The standings of this method in the three categories show that with further improvements in performance and output, it could be widely applicable in the recycling industry.
Sensitivity analyses were performed to study how the economic attractiveness of each scenarios would change in case the actual costs or revenues were different compared to the estimated in Table 4.1. Using initial investment cost, annual operating cost and metric tons of plastic collected annually as the changing variables, the effects of changing these variables were studied on the payback period and IRR of each scenario and the financial attractiveness of the scenarios were compared using the payback period as a measure while considering a MAPP of 5.64 years.
The results of the sensitivity analyses showed that all of the scenarios are more sensitive to annual operating costs compared to the initial cost of investment so it is generally preferable to increase the initial investment cost rather than the annual operating cost if different options exist. Also, all of the scenarios were extremely sensitive to the amount of plastic collected annually which the main component of their revenue streams is so it is advisable for the investors/decision makers to watch out for policies or regulations that could change the plastic collection rate.
The results of the analysis show that all of studied scenarios are economically viable.
They also provide rankings for three categories; however, further research needs to be carried out to provide a more comprehensive analysis. It is important to notice that the EM-Ferro method is still under development and improvement to reach its optimum success rate. Having known the success rates, the boundaries of the study can be broadened to encompass the avoided burdens of each scenario and cost, energy, and CO2 emission related inputs and outputs corresponding to the disposal of unsorted plastics for each scenario.