**Eco-Energetical Life Cycle Assessment of Materials and Components of Photovoltaic Power Plant**

#### **Izabela Piasecka 1, Patrycja Bałdowska-Witos 1,\*, Katarzyna Piotrowska <sup>2</sup> and Andrzej Tomporowski 1,\***


Received: 12 February 2020; Accepted: 12 March 2020; Published: 16 March 2020

**Abstract:** During the conversion of solar radiation into electricity, photovoltaic installations do not emit harmful compounds into the environment. However, the stage of production and post-use management of their elements requires large amounts of energy and materials. Therefore, this publication was intended to conduct an eco-energy life cycle analysis of photovoltaic power plant materials and components based on the LCA method. The subject of the study was a 1 MW photovoltaic power plant, located in Poland. Eco-indicator 99, CED and IPCC were used as calculation procedures. Among the analyzed elements of the power plant, the highest level of negative impact on the environment was characterized by the life cycle of photovoltaic panels stored at the landfill after exploitation (the highest demand for energy, materials and CO2 emissions). Among the materials of the power plant distinguished by the highest harmful effect on health and the quality of the environment stands out: silver, nickel, copper, PA6, lead and cadmium. The use of recycling processes would reduce the negative impact on the environment in the context of the entire life cycle, for most materials and elements. Based on the results obtained, guidelines were proposed for the pro-environmental post-use management of materials and elements of photovoltaic power plants.

**Keywords:** CED; Eco-indicator 99; IPCC; LCA; photovoltaics panels; recycling; landfill

#### **1. Introduction**

Climate change has occurred many times in the history of the planet Earth. For the first time, however, the climate is changing faster than before. The natural greenhouse effect, necessary for life, has been intensified in modern times as a result of human activity, and the thermal balance has been significantly shaken. For the protection of the climate, the energy sector is of strategic importance as the largest consumer of energy raw materials and emitter of pollution. A sustainable energy policy should ensure that the social needs of current and future generations are best met by maintaining a balance between energy security, competitiveness of the economy and environmental protection, including the climate [1–4].

Modern civilization has become almost completely dependent on energy. Economic and social analyses indicate that the civilization changes taking place are deepening this relationship (Table 1). Energy in all forms will play an increasingly important role not only in the economic but also in the social sphere [5–7].


**Table 1.** Energy consumption over the centuries per person per day [7].

With the current state of technological development, energy is most often obtained by processing energy raw materials, such as coal, natural gas or oil, and to a lesser extent from renewable energy sources (e.g., solar radiation, water, wind, biomass). Energy resources are not evenly distributed everywhere. Some countries do not have them at all or the resources they have at their disposal do not fully meet their energy needs. For this reason, they are forced to obtain the necessary raw materials from regions where they occur in excess. The problem with energy resources is further complicated by the fact that, in the opinion of numerous experts, their resources are limited and are running out [8–10].

One of the reasons that leads to rapid environmental degradation is excessive consumption of energy obtained from conventional sources. Pollution caused by burning fossil fuels is associated with the production of a very large amount of harmful compounds, which include SO2, NOx, CO2, CO, as well as ashes and waste heat. The effects of air pollution by conventional fuel power plants include: human and animal diseases, destruction of vegetation, destruction of building structures (including historic buildings), metal corrosion and increased machine wear, etc. [11–13].

The need to increase the share of renewable energy in the energy balance of each country results from the obligation to reduce CO2 and other greenhouse gas emissions as a result of the growing greenhouse effect, the need to replace depleting fossil fuel resources with other energy sources, and the desirability of reducing dependence on energy suppliers from other countries [14,15].

Solar installations are becoming increasingly popular around the world. The advantage of photovoltaic cells is undoubtedly that their long-term, trouble-free operation allows for a significant reduction of harmful emissions. However, their production is very energy-intensive, which entails the emission of combustion products. Due to the presence of heavy metals in PV panels, their future recycling may also become a problem. The advantage of this branch of energy is the ubiquity of the sun's rays, environmental friendliness and inexhaustibility. However, disadvantages include daily and annual cyclicity, radiation dispersion and significant costs of the equipment used [16–19].

In the global literature, many analyses can be found, mainly regarding the evaluation of solar panels with particular emphasis on the conditions of the production process. Kumar et al. (2018) indicated that the effect of shape of abrasive and silicon crystal is relevant to the yield and the life cycle of the solar cells over 20+year lifetime. In addition, the production process and the materials and raw materials used are very important for the quality of the solar farm. A detailed analysis of quality and durability was carried out by Kumar et al. (2017), proving that the impact of diamond wire wear impact on surface morphology, roughness and properties of silicon wafer subsurface. Despite numerous publications describing the stages of the production process, no studies have been found on the impact of selected system elements on the condition and development of the natural environment. Therefore, the following hypothesis is worth considering: which of the analyzed elements of the solar plant show the highest level of negative impact on the environment?

The pro-ecological attitude adopted by the authors is aimed at reducing gas emissions due to the operation of PV cells, which must correspond with environmentally friendly technology for producing photovoltaic cells. To this end, the authors made a detailed LCA (Life Cycle Assessment) analysis showing stages throughout the life cycle of the system that have a negative impact on the environment.

Increasing care for nature leads to the development and use of increasingly complex methods that give control of, and the ability to counteract, the human impact on the environment. Therefore, many new ways of assessing the impact of processes, products and industries on the environment have been created. One of them is the method of analysis and assessment in the context of the entire life cycle of products, i.e., their impact from the acquisition of raw materials to development. The Life Cycle Assessment (LCA) method covers the environmental impact of production, operation and post-use management and is in accordance with the principle of sustainable development [20–22].

Each source of energy, even classified as renewable, has a certain impact on the environment. Photovoltaics are widely regarded as a "green", environmentally friendly energy source. During photovoltaic installation exploitation, solar radiation is converted into electricity. This process does not cause emissions of harmful substances into the environment, unlike analogous ones, when using conventional resources (e.g., emissions of CO2, SO2, NOx, dust, etc., as a result of burning coal). The fact that the production and post-use development of plastics and components of photovoltaic power plant is usually overlooked is the need for large material expenditures, for example related to the extraction of raw materials for the production of plant components or chemicals necessary for recycling processes. In addition, the accompanying processes (for example production of PV cells using the Czochralski method) are extremely energy-consuming. During the entire lifecycle of a photovoltaic installation, many compounds and chemicals are emitted that can have a negative impact on the environment, and large amounts of energy are required (especially at the production stage). In view of the above, main target of this study is an ecological and energetical life cycle assessment of materials and components of photovoltaic power plant.

#### **2. Materials and Methods**

#### *2.1. Object and Plan of Analysis*

The object of this study is a photovoltaic power plant with a capacity of 1 MW, situated in the northern Poland, which produces from 950 to 1100 MWh of electricity per year. As a reference for the purpose of further analyses, it was assumed that the system produced 1,000 MWh per year. The basic elements of the parsed photovoltaic power plant are: supporting structures, photovoltaic panels, cables and straight connectors for electrical installations, container station along with the static inverters (including DC switchgear, DC/AC inverters, AC/LV switchgear, LV/MV transformer, MV switchgear, control and surveillance system, the system of measurement of energy generated).

An LCA study (in accordance with ISO 14000) consists of four stages: determination of goal and scope, life cycle inventory (LCI), life cycle impact assessment (LCIA) and interpretation (Figure 1) [23–25].

**Figure 1.** Life Cycle Assessment (LCA) framework.

In accordance with the mentioned ISO standards, the analysis plan consisted of four basic steps. In the first of them, the purpose of the study and its scope were determined, which are described in detail in Section 2.2. The basis for their formulation was the collection of the largest possible amount of data on the studied object. Key data for the analysis were provided by the owner of a 1 MW

photovoltaic power plant located in Poland. In addition, detailed information on the production and operation of photovoltaic panels, inverter stations, cables and cabling accessories was obtained from their manufacturers. We also managed to obtain data from a post-use PV panel management company. Details on the second step of the analysis are provided in Section 2.3. The third step involved performing a comprehensive ecological and energetical analysis of the life cycle of the photovoltaic power station under study. SimaPro 8.4 software (PRé Sustainability, LE Amersfoort, Netherlands) was used for this purpose. The basic calculation procedure was the Eco-indicator 99 method, which allows the assessment of the impact of the photovoltaic power plant's life cycle on the environment, including human health, ecosystem quality, and resources. To determine the energy demand and CO2 emissions at each stage of the life cycle, in addition the CED and the IPCC methods were used. The IPPC method was used for CO2 emission quantitative assessment to indicate which of the material stages of the life cycle involves the highest level of greenhouse gas emissions (CO2), the reduction of which is one of the key aims of the European Union countries. The CED method was used to identify the share of a photovoltaic power plant life stages in the energy demand from different sources. In this way, a comprehensive analysis of a photovoltaic power plant cycle, including important areas of sustainable development, that is, CO2 emission (IPCC method), energy consumption (CED method), human health, ecosystem quality and resources (Eco-indicator 99 method), was provided.

The characteristics of this stage of the analysis are described in Section 2.4 and the results are presented in Sections 3.1–3.5. The last, fourth step was the interpretation of the results of the analysis, which are presented in Sections 2.5 and 4 ("Conclusions").

#### *2.2. Determination of Goal and Scope*

This work analyzes several single products connected in a one system—a photovoltaic power station. The analysis, which was conducted as part of this publication, aimed at the numerical determination of the value of the environmental impact related to the life cycle of a 1 MW PV power station. The purpose of the analysis is, most of all, to describe the existing reality (retrospective LCA), but also to model future changes and determine recommendations aimed at developing more pro-environmental solutions (prospective LCA). The procedure will constitute a classic process LCA, the purpose of which will be to determine the extent of the negative environmental impact of the life cycle of the analyzed object [26–28]. For this purpose, selected elements of the solar installation were analyzed: photovoltaic panels, supporting structures, inverter station, electrical installations. The environmental assessment included 11 impact categories: Carcinogens, Resp. Organics, Resp. Inorganics, Climate change, Radiation, Ozone layer, Ecotoxicity, Acidification / eutrophication, Land use, Minerals, Fossil fuels. The research results are divided into four phases, described as: production, exploitation, landfill and recycling. Among the eleven categories available, categories with the highest level of significance were selected for which detailed emissions of compounds into the environment were presented.

Most of the processes performed under the analyzed stages of the life cycle of the photovoltaic power station (production, exploitation, post-use management) take place in Europe. Therefore, the study scope was referenced to the European conditions. The territory of Poland was taken as the geographical area, while the time horizon taken was 20 years (average operation time of photovoltaic systems). Electric energy production was assumed as a function of the photovoltaic power plant. A functional unit was defined as a production of 1000 MWh of electric power by the relevant system in a year. The analysis did not cover the stages of transport, sales, technical tests, and storage. The main reason for this was the lack of appropriate data and large differences in the effects of transport depending on the power plant location.

#### *2.3. Life Cycle Inventory (LCI)*

To collect data, special sheets were prepared. Each sheet was assigned to a specific unit process, with a division into process inputs, process performance, and process outputs (Figure 2). Process inputs

included main materials, auxiliary materials, and water; process performance involved duration and media consumption; process outputs included main product, waste, and emissions. Data concerning processes and materials less significant from the point of view of environmental impact were obtained from databases included in the SimaPro 8.4 software(PRé Sustainability, LE Amersfoort, Netherlands). Due to confidentiality agreements with companies manufacturing photovoltaic power station elements, any detailed information regarding the design of the analyzed objects and process data are not subject to disclosure in this publication [29,30].

**Figure 2.** The material life cycle of photovoltaic power plants.

After the data were assigned to the unit processes, they were validated through bilateral energy and mass balance. Models were constructed systematically and filled with data. The input value was equilibrated by the output value. This operation made data aggregation and quantification per functional unit and reference flows possible. By totaling environmental interventions of the same type (inputs of material, energy, waste, emissions, etc.) for all the unit processes, input–output matrices were obtained that were referenced to the reference flows. The next step was to adapt them to a format compatible with the SimaPro 8.4 software. This allowed us to enter data into a calculator and proceed with another stage of the analysis. Information supplied by the manufacturer allowed a precise determination of the values of the materials and energy used in the photovoltaic power station life cycle [31–33].

The relevant power station was equipped with photovoltaic panel support structures made of galvanized steel (mainly due to economic benefits and numerous technical advantages). The support structure in a dual system was placed directly in the ground. Properly selected photovoltaic panels constitute a key element of the entire photovoltaic plant. The construction of the relevant power plant needed a system of 4170 polycrystalline photovoltaic modules with a capacity of 240 W demonstrating a performance up to 17.7%. The manufacturer's declared performance amounts to 91.2% of rated capacity for the first 10 years and 80.7% for a period of another 15 years. Each single module is composed of 60 photovoltaic cells. The panels are made of glass, aluminum, silica, EVA (ethylene-vinyl acetate copolymer), PVB (polyvinyl butyral), cadmium, lead, copper, nickel, selenium, and silver. The modules are connected in series and are inclined at an angle of 35◦ in the southern direction (this is due to the fact that the angle of the PV panels tilt relative to the horizon depends on the latitude (ϕ). The location of the analyzed object is 54◦N. The angle of the sun above the horizon is H = 90 ◦-ϕ; therefore, for the location considered it will be 36◦. The highest efficiency is obtained with a perpendicular angle of incidence of sunlight on the panel surface. In Poland, for year-round operation, the optimal angle for placing PV panels is about 30-40◦ and pointing southwards). To function properly, each photovoltaic system should be equipped with proper wires and cabling accessories. The relevant plant, located on the ground, employed wires with pre-terminated ends (galvanized copper conductor, internal insulation and external sheath made of cross-linked polyolefin), control cables and wires, different types of connectors and splitters (galvanized copper contacts), cable glands (body: PVDF—polyvinylidene fluoride and polyamide PA6; seal: silicone or neoprene) and protection hoses (modified polyamide PA12). In the relevant solution of photovoltaic power station, a central inverter substation was used. Megawatt substation constitutes a comprehensive solution dedicated for solar power plants with a high installed capacity. It contains the electrical equipment necessary to connect photovoltaic power stations to a medium voltage electric power network. The station accommodates two central inverters, optimized transformer, MV switchgear, DC connections for PV modules, and a monitoring system. Made of steel, the insulated container is placed on a concrete base. The entire megawatt substation weighs nearly 20 tons, while the volume of the container amounts to almost 50 m<sup>3</sup> (manufacturer's data).

#### *2.4. Life Cycle Impact Assessment (LCIA)*

Life cycle impact assessment was performed with the use of the SimaPro 8.4 calculation software. Cut-off level amounted to 0.1%. LCIA results are presented in Section 3 of this article.

#### 2.4.1. Eco-indicator 99 method

Eco-indicator 99 method was chosen as the base calculation procedure. Eco-indicator 99 belongs to a group methods for modeling the environmental impact of environmental endpoint mechanism. The process of characterization is done for the eleven categories of impact, coming within three larger groups referred to as impact areas or categories of damages. There are the following areas of impact: human health, ecosystem quality, and resources. The results of the impact area indicators are further analyzed through normalization, grouping and weighting into the final Ecolabel. Eco-indicator 99 method offers 11 impact categories with a wide spectrum of analysis areas (Figure 3). The first damage category (carcinogens, resp. organics, resp. inorganics, climate change, radiation, ozone layer) is expressed in DALY (Disability Adjusted Life Years)—the number of years spent in the disease or lost. Assessment of the impact on the environment was performed with the use of a scale from 0 to 1, where 0 stands for a lack of impact on the human health and 1 stands for death. The damage to ecosystem quality is expressed in terms of the percentage of species that have disappeared in a certain area due to the environmental load. Ecotoxicity covers the percentage of all species present in the environment living under toxic stress (PAF—Potentially Affected Fraction). Regarding acidification/eutrophication and land use, the damage to a specific target species (vascular plants) in natural areas is modeled (PDF—Potentially Disappeared Fraction). The damage category covering resource extraction gives a value expressed in MJ surplus energy to indicate the quality of the remaining mineral and fossil resources. The final goal of the grouping and weighting analysis was to obtain environmental factors expressed in environmental points (Pt), constituting aggregated units enabling comparisons of

eco-balance sheets. A thousand environmental points are equal to the impact on one's environment, the average European within a year. The value of 1 Pt (eco-point) is representative for one thousandth of the yearly environmental load of one average European inhabitant. It is calculated by dividing the total environmental load in Europe by the number of inhabitants and multiplying it with 1000. Due to the lack of express premises for exclusions, all the impact categories functioning within the Eco-indicator 99 (Eco-indicator 99 (H) V2.06/Europe EI 99 H/A) were subject to analysis [34–37].

**Figure 3.** Structure of LCA impact category groupings, Eco-indicator 99 method.

Due to the fact that the purpose of the research was to carry out ecological and energy life cycle analysis of photovoltaic power plant materials and components, in addition to the Eco-indicator 99 method, it was decided to additionally use two other methods—CED and IPCC. As part of Eco-indicator 99, two impact categories ("minerals" and "fossil fuels") refer to energy analyzes (value expressed in MJ surplus energy), but due to the need for expand the scope of their research with an additional assessment of cumulative energy demand in each phase lifecycle. Therefore, the CED method was used, as described in Section 2.4.3. Nowadays, great attention is also paid to the issue of excessive CO2 emissions, which is why it was decided to use another method, IPCC, which makes it possible to assess the impact of greenhouse gases on the increase of the greenhouse effect and obtain results for each stage of the life cycle expressed in kg CO2 eq. As a result of this procedure, the results obtained under the category "climate change" in the Eco-indicator 99 method are more detailed. A more detailed description of the IPCC method is provided in Section 2.4.2.

#### 2.4.2. IPCC Method

The IPCC (Intergovernmental Panel on Climate Change, Global Warming Potential) method has made it possible to perform a quantitative assessment of the impact of particular greenhouse gasses (GHG) for the greenhouse effect, with respect to CO2. The carbon dioxide indicator in order to assess the impact on the greenhouse effect is equal to 1 (GHG = 1). This research was conducted in accordance with the IPCC standard: IPCC 2007 GWP 100a V1.01 (Intergovernmental Panel on Climate Change, Global Warming Potential, time horizon: 100 years) [38–41].

#### 2.4.3. CED Method

The CED (Cumulative Energy Demand) method allows the determination of the cumulative energy demand. The impact indicators are divided into seven impact categories: two non-renewable (nuclear power, fossil fuels) and five renewable (biomass, water, solar, wind and geothermal energy). The research was conducted in accordance with the CED standard: Cumulative Energy Demand V1.05 [42–44].

#### *2.5. Interpretation*

During the analysis, its completeness was checked against the positive result. All the important information and data necessary for interpretation were complete and obtainable directly from the manufacturer, the recycling company, and from the databases of the SimaPro software (PRé Sustainability, LE Amersfoort, Netherlands). Conformity was checked during the analysis. Assumptions, methods, analysis depth, specificity and precision of data for both systems are compliant with the previously assumed goal and scope of analysis. Detailed interpretation of the results obtained is presented in Sections 3 and 4.

#### **3. Results and Discussion**

#### *3.1. Eco-Indicator 99*

Table 2 presents the results of characterizing the environmental consequences occurring in the life cycle of selected components of the 1 megawatt photovoltaic power plant. Impacts are presented in the 11 categories of impact characteristic of the Eco-indicator 99 method. Two impact models were distinguished: the first one was the life cycle, including landfill disposal as a form of post-disposal management, while the second one was recycling. For all adverse effects in the area of human health, all tested groups showed the highest negative impact on the category of inorganic compounds causing respiratory diseases (e.g., life cycle of photovoltaic panels with landfill: 0.22 DALY; inverter life cycle with storage: 0.14 DALY). The largest quantity is formed at the stage of production of materials and elements, and the maximum share is characterized by sulfur dioxide and nitrogen oxides. These compounds are poisonous to humans and animals and have a harmful effect on plants. Sulfur dioxide is a by-product of burning fossil fuels, which, for example, contributes to atmospheric pollution (smog). In turn, nitrogen oxide is a compound with high biological activity and easily penetrates biological membranes. It is also created, among other things, as a result of burning fossil fuels and industrial processes that can cause smog. For categories affecting environmental deterioration, the ecotoxic compounds category was the most important, for which the maximum level of harmful impact was recorded for photovoltaic panels deposited in landfill: 137,741 PAF·m2/a. Ecotoxic compounds are substances which, due to their origin, chemical, biological or other properties, constitute or may pose a direct or delayed threat to humans, animals and plants. In the life cycle of a photovoltaic power plant, the largest amount arises from the storage of materials and components at the landfill. A particular threat is copper ion emissions, which in addition to reducing the quality of the environment, can contribute to the formation of diseases of the nervous and digestive systems in humans (for example: mental disorders or liver damage). In the area of processes affecting the depletion of raw material resources, the highest level of harmful impact was recorded in the raw materials category for the life cycle of electrical installations with post-consumer use in the form of landfill (322,646 MJ), and in the fossil fuels category for the life cycle of photovoltaic panels placed in landfill after use (400,584 MJ). The largest amount of fossil fuels is consumed during the plastics and materials plant production phase. The most processes with the highest energy demand include, for e.g., the production of PV cells. Burning of conventional fuels is associated with many hazardous emissions to the atmosphere,

water and soil, which are the causes of, for examplem, diseases, increasing the greenhouse effect, ozone layer depletion or increased smog and acid rain. In most of the categories considered, there is a positive impact of the use of recycling processes, to reduce the harmful impact of particular groups of components of the analyzed photovoltaic power plant.


**Table 2.** Results of characterization of environmental consequences, occurring in the life cycle of selected groups of 1 megawatt photovoltaic power plant.

The results of grouping and weighting the environmental after-effects of the existence of selected groups of 1 MW photovoltaic power plants are summarized in Table 3. For the life cycle of photovoltaic panels deposited in landfill, the highest level of harmful impact was recorded in terms of: fossil fuel extraction (9534 Pt), inorganic compounds causing respiratory diseases (5729 Pt), mining of minerals (1621 Pt) and carcinogenic compounds (1822 Pt). Silicon cell production processes are associated with a huge demand for energy, which in the case of the analyzed power station is about 5 million MJ. This energy is most often obtained from non-renewable sources, which causes many negative impacts in relation to human and animal health, and a reduction in the quality of the environment. Another problem is the excessive exploitation of raw material deposits, including silver, used in the electrical contacts of the cells, whose extraction causes the most negative consequences in comparison to other substances and chemical compounds used in the production of PV cells. For the life cycle of supporting structures, including post-use disposal, the categories were: ecotoxic compounds (832 Pt), carcinogenic compounds (827 Pt), and fossil fuel extraction (606 Pt). Supporting structures were made mainly of galvanized steel. The galvanizing process of steel poses a threat to health and the environment. In thermal processes, smoke and zinc vapors (especially particles smaller than 1 μm) can, for example, get into the respiratory system, causing many diseases. In the life cycle of the inverter station including landfill disposal, the highest levels of negative impact were reported in terms of: inorganic compounds causing respiratory disease (3725 Pt), and fossil fuel extraction (1887 Pt). In the cycle of existence of an electrical installation, which after ending its life cycle will be deposited in landfill, the most negative influence on the environment was attributed to mineral extraction (7679 Pt) and inorganic compounds causing respiratory diseases (1327 Pt). An important element found in both inverter station and electrical installation are electrical cables and wires. The main raw material for their production is copper, the extraction and processing of which is associated with very high energy inputs, obtained from conventional sources. As a consequence, many harmful compounds are emitted into the environment, and the resources of raw materials and fuels are depleted. The application of recycling would reduce the harmful impact on the environment of all analyzed groups of photovoltaic elements.


**Table 3.** Results of grouping and weighting environmental consequences, occurring in the cycle of existence of selected groups of 1 megawatt photovoltaic power plant [unit: Pt].

The highest total level of harmful impact on the environment is the life cycle of photovoltaic panels ending in landfill storage (22,029 Pt), but in this case, the use of recycling processes significantly reduces their negative impact on the environment. The key reason for this is the abovementioned very high energy demand in the production of PV cells. Reuse of cells recovered in the recycling process is associated with large savings in both energy and materials (e.g., elimination of significant material losses arising during cutting silicon rollers). The lowest total damaging effect was found in the supporting structures (1089 and 3197 Pt) (Figure 4).

**Figure 4.** Results of grouping and weighting of environmental impacts occurring during the material life cycle of selected groups of 1 megawatt photovoltaic power plant for all categories of influence [unit: Pt].

The highest level of harmful impact on the environment of 1 MW photovoltaic power plant was observed at the manufacturing stage. The largest category of harmful effects was characterized by the categories of fossil fuel extraction and minerals, while the smallest – compounds causing the increase of the ozone hole. The largest share in the negative impact on the environment was characterized by the production processes of PV panels, which are part of the photovoltaic power plant with the highest demand for energy and materials. The lowest level of negative impact was the exploitation stage (32 Pt total). By comparing the received forms of post-use management, the most negative influence on the environment is the landfill. Carcinogenic compounds and ecotoxic compounds can be classified as the most potent adverse effects. The use of recycling processes would reduce the impact of the life cycle in

most of the impact categories analyzed. The fossil fuel mining processes and emissions of inorganic compounds affecting respiratory diseases would be most positively affected (Figure 5).

**Figure 5.** Grouping and weighting results of environmental impacts occurring in the material life cycle stages of a 1 megawatt photovoltaic power plant [unit: Pt].

Post-use disposal of photovoltaic power plant components could be the most important source of oncogenic factors contributing to the development of cancer by mutation of genetic material. The highest level of harmful emissions was recorded for cadmium ions (2547 Pt). Cadmium is an element that easily concentrates in air, water and soil and quickly moves in the soil–plant–human trophic chain. Due to easy absorption and bioaccumulation in living organisms and toxic influence, it is one of the most serious threats to the natural environment and man. For this reason, it is important to minimize the storage of materials and components of solar power plants in landfills. The production phase also has a significant impact on the carcinogenic emissions—538 Pt total—and the highest share of arsenic: 363 Pt. The use of recycling would reduce the harmful impact of carcinogens by a total of 824 Pt, mainly in the range of arsenic (-514 Pt) (Table 4).


**Table 4.** Results of grouping and weighting environmental after-effects for carcinogenic compounds present in the 1 megawatt photovoltaic power plant [unit: Pt].

The production stage is distinguished by the highest level of negative impact in organic compounds causing respiratory diseases, altogether 11 Pt, especially in the emissions of non-methane volatile organic compounds (10 Pt). NMVOC is a group of organic compounds that occur as by-products in many industrial processes and are a source of environmental pollution, including, for example: acetone (paints, protective covers, sealants), aliphatic hydrocarbons (paints, glues, sealants, combustion processes), aromatic hydrocarbons (paints, glues, combustion processes), polycyclic aromatic hydrocarbons (paints, polymeric materials, incomplete combustion processes, e.g., car exhaust gas) or chlorine-containing

compounds (varnishes, solvents). During storage of the power plant components in landfill, biogenic methane may be the greatest risk: 1 Pt. Recycling processes would reduce the damaging impact of the total lifetime by 17 Pt, including non-methane volatile organic compounds by 16 Pt (Table 5).


**Table 5.** Results of grouping and weighting of environmental after-effects for organic compounds causing respiratory diseases, occurring in the 1 megawatt photovoltaic power plant [unit: Pt].

The highest level of harmful impact on the environment of respiratory organisms caused by respiratory diseases occurring during the 1 MW photovoltaic power plant cycle was recorded for the production stage–10,837 Pt in total. The highest share was in sulfur dioxide (3194 Pt) and nitrogen oxide (3091 Pt). Recycling would allow a total reduction of harmful effects of 7423 Pt, mainly in the sulfur oxide (-3274 Pt) and particulates in total (-2521 Pt). A key negative health role is played by atmospheric aerosols or particulate matter (PM). They are drops or solid particles of natural or anthropogenic origin (impurities). PM 2.5 (particle size 2.5 μm or smaller) is the most harmful because prolonged exposure to it results in a reduction in life expectancy, while short-term exposure to high concentrations causes an increase in deaths from respiratory and circulatory diseases and increases the risk of emergencies that require hospitalization (for example: worsening of asthma, decreased lung function), because dust enters the blood directly through the lungs. Atmospheric aerosols also contribute to smog (Table 6).

**Table 6.** Results of grouping and weighting of environmental after-effects for inorganic compounds causing respiratory diseases, occurring in the 1 megawatt photovoltaic power plant [unit: Pt].


The highest level of emissions of substances causing climate change is the production of materials, and components, which results in a total of 2049 Pt, mainly composed of carbon dioxide (891 Pt). Over the past two centuries there has been a marked acceleration of climate change. The basic factor shaping the speed of these changes is the emission of carbon dioxide and other greenhouse gases, which arise, for example: during the exploitation of fossil fuels, used in the life cycle of a photovoltaic power plant most often for obtaining electricity for various processes, including the energy most

intensive—PV cell production. Landfill disposal can result in total negative emissions of 292 Pt, primarily biogenic methane (201 Pt). Recycling, as a form of post-use management, would allow the reduction of dangerous emissions by a total of 2685 Pt, mainly in carbon dioxide (-1947 Pt) (Table 7).


**Table 7.** Results of grouping and weighting of environmental after-effects for compounds causing climate change, occurring in the 1 megawatt photovoltaic power plant [unit: Pt].

Radioactive compounds are characterized by possessing nuclear nuclei with radioactive decay, most commonly associated with alpha particle emission, beta particles, and gamma radiation. The highest number of such elements in the 1 MW photovoltaic power plant cycle was noted for the production phase (total 28 Pt). It is associated with the processes of extracting mineral resources and fossil fuels, during which there are not only emissions of dust and gases containing many harmful substances (e.g., sulfur and nitrogen oxides, chlorine, fluorine, heavy metals), but also radioactive elements such as uranium, thorium, and potassium, and their breakdown products, for example: radium and radon. In this case, these are mainly radon isotopes—222Ra (21 Pt)—and carbon—14C (7 Pt) (Table 8).

**Table 8.** Results of grouping and weighting of environmental after-effects for radioactive compounds, occurring in the 1 megawatt photovoltaic power plant [unit: Pt].


The ozone hole is a phenomenon of a decrease in the concentration of ozone (O3) in the stratosphere, resulting in a decrease in the level of absorption of ultraviolet radiation reaching the Earth from the Sun. It is, therefore, a threat to living organisms. During the processes of producing materials, and components of the plant under investigation, harmful substances causing the ozone hole to increase in total 2 Pt are formed, mainly bromotrifluoromethane (1,5 Pt). Halon 1301 may have a toxic effect on the central nervous system and other bodily functions (Table 9).


**Table 9.** Results of grouping and weighting of environmental after-effects for ozone-increasing compounds, occurring in the 1 megawatt photovoltaic power plant [unit: Pt].

Comparing all phases of the life cycle, particularly high levels of harmful impact of waste disposal in the form of waste landfills are visible in the category of ecotoxic compounds (a total of 1503 Pt). The most significant level of negative emissions was copper ions (1200 Pt). The high level of emissions of ecotoxic substances is also characterized by the production phase (total 837 Pt), which consists mainly of the harmful effects of nickel (280 Pt) and zinc (263 Pt). Recycling could significantly reduce emissions by a total of -397 Pt, primarily in terms of minimizing the negative impact of nickel (-212 Pt). Nickel is used in the manufacture of many materials and components of a photovoltaic power plant, ranging from steel elements to resistors. The main source of nickel in the environment is the combustion of conventional fuels (especially coal and oil), as well as steel production and electroplating processes. The absorption of nickel into the body is primarily through the respiratory system. Nickel tends to accumulate in the lungs. With wind and rain, it gets into soil and groundwater. It is also one of the components of smog (Table 10).


**Table 10.** Results of grouping and weighting of environmental after-effects for ecotoxic compounds, occurring in the 1 megawatt photovoltaic power plant [unit: Pt].

Acidification of the environment is a phenomenon of progressive decrease in the pH value of its individual components. It can be caused by anthropopressure, for example: by emissions of air pollutants (e.g., SO2, NOx, NH3) as a result of combustion of conventional fuels. Eutrophication, on the other hand, consists of enriching the environment with biophilic elements, mainly phosphorus, which causes an excessive increase in their trophic (biological productivity). Materials and elements of a 1 megawatt photovoltaic power plant can be distinguished primarily by nitrogen oxide (596 Pt) and sulfur dioxide (182 Pt). The total negative impact of the production stage is 975 Pt. Recycling processes would minimize harmful emissions by a total of 502 Pt, including 314 Pt for nitrogen oxide and 187 for sulfur oxide (Table 11).

The highest level of negative impact of land use category was characterized by the production stage (total of 1551 Pt), including primarily the processes related to the transformation to mineral extraction site (338 Pt). A significantly lower level of adverse impact is the potential for landfill disposal—a total of 35 Pt. Extraction of mineral resources is an area of economic activity with one of the highest harmful impacts on human health and the quality of the environment. It is associated with environmental destruction, especially serious in the case of open pit mines. It is characterized by the consumption of huge amounts of water, often causing shortages, and at the same time results in the contamination of surface and groundwater, as well as a reduction in their level by up to several meters (Table 12).



**Table 12.** Results of grouping and weighting of after-effects of environmental land use processes, occurring in the 1 megawatt photovoltaic power plant [unit: Pt].


Economic development entails an increase in demand for various types of natural resources, resulting in the depletion of non-renewable resources. Although their deposits are limited, their exploitation continues to grow. The stage of production of 1 megawatt photovoltaic power plant, characterized by the highest value of harmful influence in the mineral mining sector, is the production phase (total 9453 Pt), mainly in the field of tin mining (6610 Pt) and copper (2230 Pt). Due to its physical and chemical properties, tin is very important for industry. Its use in the metallurgical industry is the largest. In addition, this element is used for solders alloys. Tin is also used to coat other metals, e.g., steel, with a thin anti-corrosive layer. Recycling would minimize the pervasive effects analyzed, a total of 1889 Pt, mainly in the field of bauxite mining (-1835 Pt) (Table 13).

The highest level of harmful impacts in the category of fossil fuel extraction processes is characterized by a production phase (total of 12,159 Pt), which consists primarily of processes related to extraction of natural gas (5509 Pt) and crude oil (2366 Pt). This is connected to the high energy demand of production processes, in particular PV cells. Recycling as a form of post-disposal management would reduce several adverse effects by a total of 10,095 Pt, mainly in the oil-related processes (-7794 Pt). However, landfill disposal would result in an increase in the unfavorable impact on the life cycle of the tested power plant by 137 Pt (Table 14).


**Table 13.** Results of grouping and weighting of after-effects of environmental processes related to mineral extraction, occurring in the 1 megawatt photovoltaic power plant [unit: Pt].

**Table 14.** Results of grouping and weighting of after-effects of environmental processes related to the extraction of fossil fuels, occurring in the 1 megawatt photovoltaic power plant [unit: Pt].


#### *3.2. IPCC*

The analyzed groups of photovoltaic power plant components were also subjected to an IPCC analysis to determine greenhouse gas emissions in kilograms CO2 equivalents. The results are shown in Figure 6. It follows from that that the largest amount of greenhouse gases is generated in the life cycle of photovoltaic panels that end in landfill storage (269,099 kg CO2 eq). However, if they are recycled, there is the possibility of significantly reducing the volume of the emissions in question. The lowest greenhouse gas emissions were recorded for post-consumption in recycling cycles (support structures: 3880 Pt, inverter 10,017 Pt, electrical installation: 16,091 Pt). The sources of CO2 emissions throughout the life cycle include, above all, the burning of fossil fuels to obtain electricity. Its highest demand was observed during the production of photovoltaic cells.

**Figure 6.** Results of the characterization of environmental consequences for cumulative greenhouse gas (GHG) emissions occurring in the cycle of existence of selected groups of 1 megawatt photovoltaic power plant [unit: kg CO2 eq].

#### *3.3. CED*

The last factor considered was the energy consumption of the life cycle of individual groups of photovoltaic power plants, which was estimated using the CED method. Sustainable development of technical facilities, in addition to having the lowest demand for materials and the least harmful environmental impact of the life cycle, also includes the maximum reduction of energy consumption within its individual stages (which also translates into improving the quality of the environment, especially considering the fact that the main source of energy are conventional fuels). The largest amount of energy is needed to produce photovoltaic panels (e.g., to produce silicon with proper purity), a life cycle that covers the production, operation and management of panels in the form of landfill, consumes nearly 5 million MJ. Similarly, the case of an inverter station represents over 800 thousand MJ, and for supporting structures, a further more than 1 million MJ, and with regard to the electrical installation, almost 250 thousand MJ. The use of recycling processes reduces the energy consumption of all groups of components of the power plant in question. Reuse of photovoltaic cells recovered in recycling processes would minimize energy and material consumption (Figure 7).

**Figure 7.** Results of the characterization of environmental consequences, in relation to cumulative energy demand (in MWh), occurring in the cycle of existence of selected groups of 1 megawatt photovoltaic power plant [unit: MJ].

#### *3.4. Recapitulation*

Figure 8 shows the results of grouping and weighting the environmental consequences of 1000 kg of selected plastics and materials that are part of the photovoltaic farm components. The highest level of harmful influence on the environment is distinguished by silver (20,512 Pt/1 Mg), nickel (3842 Pt/1 Mg), copper (2363 Pt/1 Mg), PA 6 (656 Pt/1 Mg), lead (638 Pt/1 Mg) and cadmium (586 Pt/1 Mg). The most widely used in photovoltaic panels are: silver, nickel, lead, cadmium, EVA, selenium, silicon, and aluminum; in supporting structures: nickel and steel; in inverter stations: copper, PA6, PVDF, rubber, and steel; and in electrical installations: copper, PA6, PVDF, and rubber.

**Figure 8.** Results of grouping and weighting the environmental consequences of 1,000 kg of selected plastics and materials included in photovoltaic elements for all categories of influence [unit: Pt].

#### *3.5. Other Energy Sources*

Using the databases available in the SimaPro software and calculations made previously for a 1 MW photovoltaic power plant, a comparison was made of the environmental impact of the processes of obtaining electricity from photovoltaics with selected, commonly used conventional energy sources and with the structure of the mixed energy characteristic of Poland, mainly based on hard and brown coal (approximately 80%).

The analyzed photovoltaic power station, during its 20-year life cycle, is able to produce about 20,000 MWh of electricity, which value was taken as a reference value. The degree of impact on the surroundings resulting from the combustion of an amount of hard coal, brown coal, and heating oil necessary to obtain the same amount of electricity was analyzed. In addition, an analogous analysis was carried out when 20,000 MWh of energy was obtained in Poland.

Table 15 summarizes the results of the characterization of the environmental after-effects arising from the generation of 20,000 MWh of electricity from the selected energy sources. The highest level of harmful impact of the analyzed energy sources is visible in the area of emissions of compounds having a negative impact on the quality of the environment (category: ecotoxic compounds, compounds causing acidification/eutrophication, land use) and in processes related to the depletion of raw material resources (categories: mineral extraction, extraction of fossil fuels). This is characteristic of energy extraction processes, especially from conventional sources.


**Table 15.** Results of the characterization of environmental consequences arising as a result of the generation of 20,000 MWh of electricity from selected energy sources.

The results of grouping and weighting environmental consequences arising as a result of generating 20,000 MWh of electricity from selected energy sources are presented in Table 16. The highest value of harmful impact on the environment was noted in the following categories: inorganic compounds causing respiratory diseases (from 3437 to 357,703 Pt), compounds that cause climate change (from 2351 to 163,556 Pt), and the extraction of fossil fuels (from 2198 to 246,591 Pt).

**Table 16.** Results of grouping and weighting of environmental consequences arising as a result of generating 20,000 MWh of electricity from selected energy sources [unit: Pt].


The highest total level of harmful impact on the environment was from the production of 20,000 MWh of electricity was determined to be from hard coal (709,17 Pt) and brown coal (633,445 Pt). A high degree of negative impact on the environment was also noted for the Polish energy mix (683,220 Pt), due to the fact that it is mainly based on the burning of brown and hard coal. Obtaining energy from solar radiation possessed the lowest level of adverse impact on the environment, with this type of installation exerting an impact from 13,997 (recycling) to 42,492 Pt (storage) throughout its entire life cycle, depending on the form of post-use management. Despite some expenditure of energy and materials in the production and post-use management phase, the use of a renewable energy source, i.e., photovoltaics, causes the least negative environmental consequences compared to conventional energy sources (Figure 9).

**Figure 9.** Results of grouping and weighting environmental consequences arising as a result of generating 20,000 MWh of electricity from selected energy sources [unit: Pt].

Additionally, the amount of greenhouse gas emissions resulting from the production of 20,000 MWh of electricity was analyzed from the same energy sources using the IPCC method. The highest level of GHG emissions was found when obtaining energy from natural gas (29,989 Mg CO2 eq), hard coal (22,872 Mg CO2 eq), and brown coal (22,241 Mg CO2 eq). In the case of photovoltaic power plants, the emission level was the lowest and amounted to about 400 Mg CO2 eq. The obtained results confirm that photovoltaic power plants can be a source of energy enabling the reduction of greenhouse gas emissions, and hence are one of the ways of reducing the greenhouse effect (Figure 10).

**Figure 10.** The results of the characterization of the environmental after-effects in relation to cumulated greenhouse gas (GHG) emissions arising from the production of 20,000 MWh of electricity from selected energy sources [unit: Pt].

As a result of these considerations it was found that the total highest level of harmful impact on the environment is the life cycle of post-consumer photovoltaic panels when stored in landfill (22,029 Pt). The largest amount of greenhouse gases generated during the post-consumer life cycle of photovoltaic panels is from storage in landfill (269,099 kg CO2 eq). The greatest amount of energy absorbed during the life cycle of photovoltaic panels is from the use of this form of storage—nearly 5 million MJ. The use of recycling processes reduces the energy consumption of all groups of components of the power

plant in question. Silver, nickel, copper, PA 6, lead, and cadmium are among the materials with the most harmful influence on the environment.

An additional element of the analysis was the comparison of the environmental impact of the processes of obtaining electricity from photovoltaics with selected, most commonly used conventional energy sources and with the structure of the mixed energy characteristic of Poland, which is mainly based on hard and brown coal (approximately 80%). The highest level of harmful impact of the analyzed energy sources is visible in the area of compound emissions: ecotoxic (2,671,933 PAF·m2/a brown coal), compounds causing acidification/eutrophication (335,746 PAF·m2/a hard coal), land use (192,997 PAF·m2/a hard coal) and within processes related to the depletion of raw material resources includes: mineral extraction (398,914 MJ PV power plant (landfill)), and extraction of fossil fuels (10,360,976 MJ heating oil).

#### **4. Conclusions**

Renewable energy sources, including solar energy, possess many positive environmental aspects in terms of local, regional and national and, most importantly, global aspects. The beneficial effect of the use of alternative energy sources can be considered in three areas: continuous sustainable development, the environment protection and natural resources, and the timeless and endless nature of the raw materials used [45,46].

Energy security is understood as providing the security of energy supplied to recipients in a particular time and place, and it is one of the priority actions in economic policy. Scattered generation ensures even distribution of heat sources and energy (derived from a wide variety of energy sources, including renewable sources of energy) and have been the subject of considerable interest. They are considered to be important not only for increasing energy security, but also for the reduction of greenhouse gas emissions, particularly carbon dioxide, from burning fossil fuels. Each energy type impacts the quality of the environment, but not all to an identical extent [47–49]. That is why processes related to solar cell production have become so important in environmental assessment, determined by assessing the surface properties, including roughness, thickness and machining [50,51].

In the opinion of society, the use of only conventional energy sources (for example coal or oil), is a threat to human health and the quality of the environment. However, the depletion of non-renewable resources concerns not only traditional ways of obtaining energy, but also alternative ones. As a consequence, we are constantly striving to minimize negative environmental impacts, e.g., by reducing CO2 emissions [52]. Bearing in mind the future prospects of sustainable development, directions for the environmental assessment of wind farms, small hydropower plants and solar farms have become attractive. This approach makes it possible to strive to meet the growing demand for electricity without burning fossil fuels [53,54]. Thus, this work proved that at the production stage of the elements for a photovoltaic power plant, there is also a demand for raw materials and energy. The same applies to processes related to post-use management. Assuming that the analyzed power station produces annual energy equal to about 1000 MWh, it must work for up to about 2 years to produce an amount of energy equal to the demand for it throughout its entire life cycle (total energy demand for the life cycle with post-use management in the form of storage on the dump is about 2024 MWh—Figure 7). For this reason, it was considered justified to conduct research aimed at ecological and energetical assessment of the life cycle of materials and components of a photovoltaic power plant, and the hypothesis adopted in the work was confirmed.

Based on the results of the research performed and an evaluation of the material stages of the life cycle of the analyzed PV power plant, in terms of pro-environmental, post-production use of materials, materials and elements of photovoltaic power plants, the following suggestions are proposed:



**Author Contributions:** Conceptualization, I.P. and P.B.-W.; methodology, I.P. and P.B.-W.; software, I.P.; validation, I.P., P.B.-W. and K.P.; formal analysis, I.P, P.B.-W, K.P. and A.T..; investigation, I.P. and P.B.-W.; resources, I.P.; data curation, I.P.; writing—original draft preparation, I.P.; writing—review and editing, I.P, P.B.-W, K.P. and A.T.; visualization, I.P.; supervision, I.P, P.B.-W, K.P. and A.T.; project administration, P.B.-W. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Influence of Waste Management on the Environmental Footprint of Electricity Produced by Photovoltaic Systems**

#### **Sina Herceg \*, Sebastián Pinto Bautista and Karl-Anders Weiß**

Department of Material Analysis, Fraunhofer Institute for Solar Energy Systems ISE, Heidenhofstr 2, 79110 Freiburg, Germany; sebastian.pinto.bautista@ise.fraunhofer.de (S.P.B.); karl-anders.weiss@ise.fraunhofer.de (K.-A.W.) **\*** Correspondence: sina.herceg@ise.fraunhofer.de

Received: 30 March 2020; Accepted: 19 April 2020; Published: 1 May 2020

**Abstract:** PV waste management will gain relevance proportionally to the amounts of waste that are expected to arise with the phasing-out of old installations in the upcoming years and decades. The Life Cycle Assessment (LCA) methodology is used here to analyze the environmental performance of photovoltaic systems and the waste management methods that have been developed recently. Several LCA studies have already been performed for PV technologies, but in most cases these do not include the end of life stage, thus there is still uncertainty about the impacts of recycling on the environmental footprint of PV electricity. The present study offers a more detailed analysis of different end-of-life approaches for the main photovoltaic technologies that are found on the market. The results from the analysis demonstrate that recycling has the potential to improve the environmental profile of PV electricity but at the same time there is room for further improvements in developing dedicated recycling technologies.

**Keywords:** photovoltaics; waste management; life cycle assessment

#### **1. Introduction**

#### *1.1. Motivation*

Energy transition towards cleaner electricity grids has led to a large deployment of photovoltaic (PV) installations on a global scale. However, until now only small flows of PV waste have had to be handled. This has restrained the consolidation of a specialized waste processing industry due to a lack of profitability. As a consequence, little is known regarding the performance of waste management schemes and the potential ecological benefits that could come along with them. Appropriate waste management is required to mitigate potential ecological impacts of the waste material. If not handled in a proper manner, critical and valuable resources might be lost and toxic materials such as lead or cadmium contained within the modules could leak into the soil and groundwater, representing a threat for biodiversity and human health. While the risk of leakage is low under neutral atmospheric conditions, it rises when the module is exposed to acidic environments as those produced by rainfall in certain locations. Furthermore, PV systems represent a potential source of valuable materials. However, the small flow of End-of-Life (EoL) modules has restrained the profitability of recycling, since dedicated recycling processes demand high volumes of input material to make this activity economically viable [1–3]. The study of waste management will gain increasing relevance in the long term when dealing with big amounts of photovoltaic waste that are already emerging. It is nevertheless difficult to accurately predict these flows of waste since the real lifetime of PV modules still remains uncertain. Most installations have not yet met their expected lifespans, which have been indicated to

be around 25–30 years by most manufacturers. Little is known about the modules' performance after this timespan, but a couple of studies indicate that an economically viable operation of PV plants is realistic even after 30–35 years [4–6].

Several environmental assessment studies have proved the environmental benefits of PV technologies, but in most cases the end-of-life stage has been neglected. Likewise, a couple of approaches to processing waste material have been proposed, but their impact on the environmental profile of PV electricity remains uncertain. This study intends to reduce this gap through a comprehensive analysis of different waste management approaches based on recycling. The main objective is to quantify the impacts of EoL management approaches and assess their contribution to the overall environmental footprint of electricity produced by different PV technologies with a special focus on Germany.

#### *1.2. LCA of Waste Management for PV Systems*

The most recent Life Cycle Inventories (LCIs) have been published by Wambach et al. [7]. For crystalline Silicon (c-Si) modules, the inventories were built alongside plants designed for the processing of laminated glass, metals or electronic and electric waste. For the specific case of Cadmium-Telluride (CdTe) modules, the inventories included were based on the recycling activities of the company First Solar in Germany. The study concluded that the benefits of recycling were higher than the impacts caused by processing the waste material in all the categories under study with an exception for CdTe in the category 'human toxicity (cancer effects)'. Another study [8] found a reduction of about 4–11% of the modules' environmental burdens and highlighted the influence of transports and electricity use in the burdens of the recycling process. The authors also presented a preliminary study on a new approach for EoL management of thin film modules, which was mostly based on mechanical processes eliminating the need for thermal treatments and reducing the use of chemical agents [9]. The technical feasibility of a new c-Si module recycling technique was studied by Duflou et al. [10] by means of a selective mechanical delamination, performing a comparative LCA with two other existing methods. Others have illustrated and analyzed an innovative process that enables the recovery of other valuable materials present in the modules such as silicon, silver and other metals that cannot be retrieved by traditional means [11,12]. Those studies applied the LCA methodology to a pilot process developed within the project called 'Full Recovery End of Life Photovoltaic' (FRELP), which is composed of a series of mechanical, thermal and chemical treatments. A more detailed analysis on the environmental impact categories according to the different processes involved was given recently [13]. This study shows net benefits from recycling in all the impact categories of analysis, as well as the identification of transportation and thermal treatments as the main burdens within the entire recycling process. A specific analysis focused on waste backsheets in order to evaluate the effects of fluorinated layers on the EoL processing found that fluorine-free backsheets performed better than fluorinated ones for both incineration and pyrolysis; thus, the use of fluoropolymers should be avoided [14]. Other studies evaluating existing waste management techniques and new alternative methods for recycling PV panels have been performed, but without offering the disclosure of inventories or a deep analysis of the ecological footprint of such techniques [15–18].

#### **2. Materials and Methods**

The evaluation of the environmental profile of a product or service is the characterization of its ecological footprint; a description of the interactions with its environment, and thus the associated impacts. Several tools have been developed to perform such an evaluation. One of these is the so-called Life Cycle Assessment (LCA), a systematic methodology which determines indicators related to impact categories that quantify the environmental burden of each phase in the lifecycle of the product. Within this paper, a total of six waste management approaches proposed by industry and research programs are evaluated under this methodology. Following the recommendations from [19,20] the impact assessment method for the analysis is the ILCD-2011 and the following impact categories were

chosen: (1) Climate change; (2) Human toxicity (non-carcinogenic); (3) Human toxicity (carcinogenic); (4) Particulate matter; (5) Acidification; (6) Freshwater ecotoxicity; (7) Resource depletion.

#### *2.1. Product Systems Description*

The system under study comprises the generation of AC electricity produced with a slanted-roof photovoltaic plant installed in Germany and manufactured under the approximated market and technological conditions of 2010–2013, which can be seen in Table 1. This specific period was chosen due to the great increase in deployment of PV installations at the time, also describing the gross composition of the waste to be processed in the future:



As proposed by [22], the following lifetimes and parameters will be used in this study:

• Life expectancy



Figure 1 shows the system boundaries defined for the assessment of electricity production, which include raw materials and energy supply in processing at the initial manufacturing plant. The EoL stage includes burdens from transportation, waste processing and final disposal of the modules. In addition, environmental credits gained from energy and material recovery are given. However, the burdens from the post-recycling refining processes of these materials needed before reintroducing them in the production lines have been left out. The manufacturing of the balance-of-system components (BOS), e.g., mounting structure and the electric installation required for PV systems, is included in the analysis; however, recycling of these has not been considered, since they are treated separately from the PV modules, and thus no credits have been accounted for.

The functional unit (FU) is 1 kWh of electricity produced with a PV system installed in Germany, which allows consistent comparability of the different technologies under study. This FU shall describe the purpose of the system studied and will serve as a reference measure to which the impacts in the different categories will be expressed. The reference flow refers to the size of the system that quantifies the functional unit, which in this case is the 3 kWp plant.

**Figure 1.** System boundaries of the current study.

#### *2.2. Characterization of Modules and Input Waste*

For practical purposes, the mass composition of c-Si input waste was taken from [11]. However, in real life, recycling plants will receive a wide variety of modules, each one with a specific environmental profile. This is because the technological development has diversified the market making it difficult to define a standard composition. The mass fractions assumed for a c-Si module can be found in Table 2.


**Table 2.** Mass composition of c-Si module input waste.

#### *2.3. Description of the EoL Approaches*

The waste management approaches under study are either state-of-the art basic recycling as described, basic recycling with additional benefits from backsheet recovery, or more dedicated recycling strategies as proposed by other research programs (Table 3). These approaches have been subject to research and testing for a long time and are thus likely to define the future scenario. Their main differences consist in the recovered material, the involved processes and the final disposal mechanism that the unrecovered materials undergo.

Basic material recovery refers to those approaches limited to recovering only bulk materials such as aluminum and glass by simple processes in order to comply with the European legal requirements of mass recovery. These approaches leave behind other valuable materials such as wafers and silver due to the added complexity of these processes. Currently, this type of approach is performed mostly in laminated glass, metal and electronic waste recycling plants where PV modules can also be processed. Figure 2 presents a diagram flow chart of one of the recycling approaches studied as well as how the boundary conditions have been defined.

**Table 3.** Modelled approaches and respective features.

**Figure 2.** Exemplary process flow chart of a recycling approach (appr. n◦1).

Approaches n◦1 and n◦2 correspond to the most common methods performed nowadays as described by Wambach et al. [7].

Approaches n◦3 and n◦4 are more specified processes described by Aryan et al. [14] which seek to determine the impacts of (A) fluorine-free and (B) fluorinated backsheets. These approaches consist essentially of the following processes: The initial preprocessing in which the frame, junction box and cables are removed. The cables are used for copper extraction while the plastic cover is sent to an incineration plant. The next stage comprises a series of mechanical steps such as shredding, crushing, sieving and separation of the different fractions. The recovered materials are glass and aluminum, while the unrecovered fraction is sent to a landfill and in some cases is partially incinerated.

In addition to basic bulk recovery, the approaches described as 'dedicated recovery' will be based on the work from Latunussa et al. [11] and correspond to approaches n◦5 and n◦6, presented in Table 3. The first step of these approaches consists in the dismantling of the modules; this includes removal of frame and wires, which later on receive the same subsequent treatments as described before. The following step is the glass separation by means of heat and mechanical detachment. The recovered glass is later refined, and the clean fraction is reused in other industrial processes, while the contaminated fraction is sent to a landfill. The remaining 'sandwich laminate' is cut into small pieces and sent to an incineration plant. Alternatively, depending on the backsheet composition, pyrolysis could be performed (appr. n◦6). The outputs of this process are heat and fly ashes which are sent to a landfill and bottom ashes rich in silicon and other metals that are returned to the recycling plant for further processing. The final processing stage consists in a series of mechanical and chemical treatments such as sieving, leaching, filtration and electrolysis which aim to separate silicon, silver and copper scrap, as well as the remaining fractions of aluminum connectors. These usable fractions are

sent to third party processors for further refinement while the sludge and liquid waste also produced within these processes are sent to a landfill.

#### *2.4. Sensitivity Analysis*

The initial analysis was performed assuming that recycling was done within the same time frame as the manufacturing of the modules (2010–2013, due to data availability). To determine how changing industrial parameters would affect the environmental performance of recycling technologies, a sensitivity analysis based on technology forecasts and the author's understanding is performed until the year 2040, in order to evaluate the impact of recycling modules produced over the last decade:

• Scenario n◦1: Changing electricity mix with higher penetration of renewables. Figure 3 shows the variation in the energy mix as expected in the year 2040 based on actual conditions (Figure 3a) [24] and predictions (Figure 3b) [25].

**Figure 3.** (**a**) Electricity mix for the 2013 baseline conditions [24], (**b**) Electricity mix for 2040 according to predictions from [25].


#### **3. Results**

#### *3.1. Electricity Production*

Figures 4 and 5 show the impact of the different EoL approaches analyzed relative to the impacts of the production of 1 kWh of electricity. For mono-Si and multi-Si the same recycling approaches with the same impacts per kg of waste have been used. Due to different conversion efficiencies, different amounts of waste have to be recycled per kWh of electricity produced. The results are displayed in the negative axis as this describes reductions of the original profile.

**Figure 4.** Potential contribution of waste management to the environmental profile of electricity production with mono-Si modules (for numerical values see Appendix A, Table A1).

**Figure 5.** Potential contribution of waste management to the environmental profile of electricity production with multi-Si modules (for numerical values see Appendix A, Table A2).

• Mono-Si modules

Figure 4 shows that the recycling of mono-Si modules has a high capacity of reducing the impacts within the impact category *'human toxicity\_(cancer e*ff*ects)'* for every approach, with potential reductions ranging from −6.6% to −7.9% of the total. *'Resource depletion'* could be reduced by about 12% when dedicated recovery is considered, but it is negligible for any other approach. The lowest effect can be seen for *'freshwater ecotoxicity'*, ranging from 0% to −1.1%. It can be seen that for the basic recycling approaches (n◦1 to n◦4) the environmental benefits are comparable in all the categories. Approaches n◦5 and n◦6 (dedicated recycling) also perform in a similar way in all impact categories.

#### • Multi-Si modules

The relative contribution of Multi-Si recycling in every impact category, with an exception of *'resource depletion'*, entails greater potential than the ones from Mono-Si recycling, as presented in Figure 5. For *'climate change'* and *'particulate matter'*, potential reductions of about 3% arise from approaches n◦1 to n◦4, and of about 6–7% from approaches n◦5 or n◦6. With potential reductions of up to 1.2%, effects from any approach on *'freshwater toxicity'* are negligible. Within 'resource depletion', impacts from approaches n◦1 to n◦4 are negligible while those from approaches n◦5 and n◦6 are 11% respectively.

#### *3.2. Sensitivity Analysis*

The results of the sensitivity analysis displayed in Figure 6 are presented in the following way: The 0% line defines the respective baseline approach; values on the positive side represent improvements with regard to the baseline while values on the negative side reflect detriments of performance or lesser benefits obtained.

The values displayed here are proportions of the results presented in Figures 4 and 5. Due to the drastic effects observed in the sensitivity analysis for the impact category *'freshwater ecotoxicity'*, the category has been plotted in an independent graph (Figure 6f). Given that the order of magnitude differs drastically from the other impact categories, a separate display has been chosen. In the baseline impact assessment for approaches n◦1 to n◦4B, net benefits in this category have been close to zero (Figures 4 and 5). Therefore, small parameter variations as disposed by the sensitivity analysis will easily translate into a large percentage of difference from the baseline. For example, in approach n◦4B (Figure 6f) a decline of 1336% of the 0.007% environmental benefits for multi-Si modules (Figure 4) translate into absolute numbers of only 0.1% additional burdens with respect to PV electricity production without EoL management.

From scenario n◦1, it can be seen that a higher penetration of renewables lowers the benefits obtained from incineration (appr. n◦2) where the recovered energy is assumedly substituting generation in a cleaner grid. This is, however, subject to whether the substituted energy is effectively electricity or just heat. For all other approaches, mostly net environmental benefits can be observed, except for the category *'resource depletion'*, since the input electricity required to run the processes will come from sources with high criticality in this category, e.g., increased silver demand for a greater share of PV. A changing recycling efficiency as analyzed in scenario n◦2 entails improvements of between 0.2% and 5.3% for most categories, whereas *'human toxicity (non-cancer)'* benefits the most. A variation in primary material content in PV modules (scenario n◦3) leads to fewer environmental benefits from recycling in every category, due to there being less primary material content to be replaced. For the special case of *'freshwater ecotoxicity*'; however (Figure 6f), the horizontal −100% line indicates that the original benefits from recycling face a reduction of 100%. In other words, benefits went down to zero. Values above this line mean that recycling will still lead to environmental benefits, fewer than in the baseline (Figures 4 and 5). Values below this, as is the case in this category, mean that recycling is now leading to additional burdens on the environment. Reducing transport distances as assumed in scenario n◦4 has on average the greatest influence, especially in the category *'resource depletion'*, with improvements up to 20%. What can be seen from scenario n◦5 is that when combining all the above assumptions, recycling has smaller relative environmental benefits as it would have under current conditions (Figures 4 and 5). Still, recycling would lead to overall environmental benefits of up to 17.2% for 'human toxicity (non-cancer)', with an exception for *'freshwater toxicity',* where the burdens overlay the benefits.

**Figure 6.** Results from the different scenarios considered in the sensitivity analysis: (**a**) high penetration of renewables, (**b**) increased processing efficiency, (**c**) lower primary material content, (**d**) Optimized collection network, (**e**) combined conditions in 2040, and (**f**) specific plot for *'freshwater ecotoxicity'.*

#### *3.3. Greenhouse Gas Emissions*

Potentially avoided greenhouse gas emissions are calculated for each approach as a measure of the global warming potential. For practical purposes, these are displayed as kg CO2-eq per kilogram of recycled waste, since these units make the calculation of cumulated greenhouse gas over a time period more comprehensible. It can be seen in Figure 7 that the potential to avoid CO2 emissions is highest for approaches n◦5 and n◦6, almost twice as high as the CO2 savings from the basic recycling approaches.

**Figure 7.** Potentially avoided greenhouse gas production of each approach.

#### *3.4. Energy Payback Time*

Table 4 shows the cumulative energy demand (CED) and Energy Payback Time (EPBT) of a 3 kWp plant without the EoL phase as a reference to evaluate the contribution of EoL management. The table further shows the potential reduction in CED and EPBT that arises from the energy savings obtained through the different recycling scenarios. For mono-Si systems the EPBT reduction potential lies between 2.4% and 4.7%. In the case of multi-Si, reductions of about 3.6% up to 7.1% are plausible. This equals an improvement of about one to three months for the EPBT.


**Table 4.** Energy metrics with and without effects from EoL management.

#### **4. Discussion**

The analysis performed within this study shows that dedicated material recovery has a significant influence on reducing the global warming potential of PV electricity. It can be seen that implementing dedicated recovery approaches could contribute, to a great extent, to the reduction of greenhouse gas emissions inherent in module production. The high savings of CO2 emissions come from the reduced need for primary materials which are now being substituted by those recovered through recycling. Table 5 shows the potentially avoided greenhouse gas emissions in tons of CO2-eq when combining the results obtained from the analysis with the average amounts of cumulative PV waste as predicted by IRENA [28] until the year 2040 and 2050. Minimum and maximum values are given in consideration of the approaches with the lowest and highest potentials:


**Table 5.** Potentially avoided greenhouse gas emissions.

By 2040, the avoided greenhouse gas emissions from recycled PV material in Germany could add up to two to four million tons, which equals around 10% of the possibly avoided average global emissions. By 2050, the same analysis results in possible emissions avoided of around three to seven million tons in Germany, accounting for 6% of the global greenhouse gas emissions that could be avoided by PV recycling. These estimations represent an ideal scenario under the assumption that all the PV waste produced until then will be recycled using any of the studied approaches. To put this in perspective, the average global CO2 emissions in 2014 were about 5 tons per capita [29].

The results show that current module recycling has moderate to low influence on the EPBT of PV systems. The maximum potential EPBT reduction is found for multi-Si and is of about three months. It is worth mentioning again that the CED used to calculate the EPBT takes into account the energy demand of the whole plant while the energy savings obtained through recycling were calculated only for the treatment of the modules. Additional energy savings and thus higher improvements in the EPBT will be obtained if the whole system undergoes recycling.

This study was conducted for the specific conditions in Germany, which differ greatly from those considered in other studies. The same applies for the boundary conditions, as is the case for some transportation steps and final disposal processes, which were often not taken into account. Additionally, different initial module compositions were used in all of the studies analyzed. For example, the study from [12], where dedicated recovery was initially studied, assumes a module mass of 22 kg, which is 37.5% heavier than the model studied here. These facts make direct comparability impossible. Clarity and transparency when describing the system boundaries become of high relevance for the correct understanding of these differences.

Cost-effectiveness of investments in recycling infrastructure will be heavily influenced by the waste composition and volumes of material to process. Accurate estimation of these parameters is therefore necessary when assessing technical and economic viability for a proper settlement of a recycling industry. Miscalculations, in a business where profits are already low, will most likely lead to economic loss. This can be exemplified with the silver content present in the modules. Reasonably, and due to scarcity and the environmental burden of its mining, research efforts are already aiming at reducing the amount of metal used. However, if the silver content reaches a certain minimum, dedicated recycling could lose economic appeal and, if no other recovery technique has been developed, silver will probably, at a slow rate, end up in landfills.

From the sensitivity analysis it can be seen that the benefits of energy recovery are lower when the production energy comes from renewable sources. Scenario n◦1 shows that energy recovery (as in approach n◦2) does not have the same impacts in a system based on a cleaner grid as when substituting electricity coming from a more fossil-based grid (as in approach n◦1). Additionally, it can be seen that the primary material content is the parameter with the highest influence on the environmental benefits from waste management. It was demonstrated that reducing the primary material content in the modules (scenario n◦3) lowers the potential benefits of recycling to a great extent. In other words, the more recycled material is used, the lower the impacts are from module recycling, as can especially be observed in the impact category 'Freshwater ecotoxicity'. With the assumed primary material content of the sensitivity analysis, this effect overcomes the estimated increased benefits obtained from recycling (as observed in scenario 5); however, it is also expected that the impacts of module manufacturing will lower considerably under the same consideration thus improving the overall footprint of electricity.

PV waste management will gain relevance when the impacts of electricity generation become smaller. PV recycling benefits will be more significant for low-impact optimized PV electricity production, for which the benefits of recycling will be found in a similar order of magnitude as the quantified environmental impacts of electricity generation. This can be achieved with better and more efficient industrial manufacturing processes both energy- and material wise.

#### **5. Conclusions**

Within this contribution, six different approaches of recycling of PV modules have been compared with respect to their impact on the environmental footprint of electricity production from a standard PV system. As has been shown, dedicated recycling as presented in approach n◦5 and n◦6 carries the best potential to improve the environmental footprint of PV electricity production. Future analysis should take into account the benefits of recycling of the whole system, including inverter and other BOS components, to further highlight the potential of environmental benefits through recycling. The overall balance shows that the benefits overlay the burdens in all the approaches and for all the impact categories studied, improving the environmental profile of PV and lowering its EPBT. The specific treatment of the backsheet layer, as seen in approaches n◦3 and n◦4, does not have a major impact on the environmental profile of recycling the modules under study due to their low mass fraction. It should, however, be taken into consideration that certain materials contained, like the fluorine in the backsheet, can become problematic when taking into account the forecasted increase in waste material flow. Efforts in research should focus on reducing the unrecovered fraction that is being landfilled containing polymers, silicon and metals. The ITRPV technology roadmap [30] already contemplates the content reduction for some of these as a countermeasure. While the processing of PV waste until 2040, as studied here, will be characterized by high silver and silicon content, the roadmap estimates significant material content reductions that could make recycling activities unfeasible from the economic point of view driving more material into landfills. Research is thus required into that subject.

The development of an appropriate recycling network will not only bring benefits in environmental aspects but will also have a great impact on the economics and financial balances of the logistic schemes. Such an assessment must also take into account a policy and legislative framework so that the outcomes meet realistic conditions. Additionally, it is very likely that trade of waste between neighboring countries will arise, given the small amount of installed capacity of certain regions which might be more conveniently managed by imports and exports, adding complexity to the analysis.

There is still plenty of room for improvement in PV waste management. Even when the results from the current approaches seem to entail moderate contributions to the ecological footprint of electricity, further development of these and new other techniques will turn recycling into a relevant tool to make photovoltaics an exemplary technology for the energy transition. Additionally, recycling should still be studied as a whole, including the processing of the BOS components as well as the environmental credits that come with it. This will most likely lead to a greater reduction of the electricity production's footprint.

**Author Contributions:** Conceptualization, S.H. and S.P.B.; methodology, S.H. and S.P.B.; software, S.P.B.; validation, S.H.; formal analysis, S.P.B.; investigation, S.H. and S.P.B.; resources, S.H. and S.P.B.; data curation, S.H.; writing—original draft preparation, S.H. and S.P.B.; writing—review and editing, K.-A.W.; visualization, S.P.B.; supervision, S.H.; project administration, K.-A.W.; funding acquisition, K.-A.W. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Appendix A**

**Table A1.** Potential contribution of waste management to the environmental profile of electricity production with mono-Si modules.


**Table A2.** Potential contribution of waste management to the environmental profile of electricity production with multi-Si modules.


#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

## **Environmental Assessment of a Coal Power Plant with Carbon Dioxide Capture System Based on the Activated Carbon Adsorption Process: A Case Study of the Czech Republic**

#### **Kristína Zakuciová 1,2,\*, Jiˇrí Štefanica 1, Ana Carvalho <sup>3</sup> and Vladimír Koˇcí 2,\***


Received: 3 April 2020; Accepted: 19 April 2020; Published: 4 May 2020

**Abstract:** The Czech Republic is introducing new technological concepts for mitigation of greenhouse gases (GHG) in coal-based energy industries. One such technology, in power plants, is post combustion CO2 capture from flue gases by activated carbon adsorption. A life cycle assessment (LCA) was used as the assessment tool to determine the environmental impacts of the chosen technology. This article focuses on a comparative LCA case study on the technology of temperature-swing adsorption of CO2 from power plant flue gases, designed for the conditions of the Czech Republic. The LCA study compares the following two alternatives: (1) a reference power unit and (2) a reference power unit with CO2 adsorption. The most significant changes are observed in the categories of climate change potential, terrestrial acidification, and particulate matter formation. The adsorption process shows rather low environmental impacts, however, the extended LCA approach shows an increase in energy demands for the process and fossil depletion as a result of coal-based national energy mix. The feasibility of the study is completed by the preliminary economical calculation of the payback period for a commercial carbon capture unit.

**Keywords:** carbon dioxide capture; activated carbon; environmental impacts; life cycle assessment

#### **1. Introduction**

In the Czech Republic, around 52.4% of the total gross electricity production (87.6 TWh) is generated from coal, which is approximately 41% of the energy mix [1]. The Czech industry emitted around 120.5 million tons CO2, with the largest proportion of 98 million tons in 2017 [2]. The major problem of reducing CO2 emissions and its sequestration lies with the implementation of carbon capture and storage (CCS) systems. It is commonly perceived that the implementation of CCS decreases local CO2 emissions and, if applied globally, supports mitigation efforts concerning the anthropogenic contribution to climate change. However, CCS technologies can be related to more complex, unexpected, or non-obvious environmental impacts. Thus, there is a common need for a holistic environmental approach that assesses and evaluates new industrial techniques and applications, which can significantly influence the environment. Results and conclusions based on such a holistic approach support the decision making of scientists, environmentalists, and governments concerning the implementation of new techniques and allow the environmental analysis in a wider and more detailed context. Life cycle assessment (LCA) is a tool for the assessment of technologies such as CCS, from both an environmental and sustainable point of view. A LCA uses different database methods

and offers several approaches for the optimization of the processes and subsequent calculation of the environmental suitability of the chosen technology.

Considering several scientific references available (summarized in Table 1), there is a need for systematic environmental studies and reports that are built on local and site-specific operational data of the pilot CO2 capture plants. Most of the available studies deal with averaged environmental data from LCA databases and use hypothetical and mathematical models to describe a specific CO2 capture system. Current LCA studies focus on the comparison of several sorbents. Most studies specifically target post-combustion capture using monoethanolamine (MEA) sorbent. Comparative LCA analyses of MEA and potassium carbonate solvents have revealed that potassium carbonate solvents contribute to lower environmental impacts [3]. Compared to MEA, this process shows a reduction of emissions and energy cost savings. In [4], Manuilova compared power units with MEA sorbents and without CO2 capture and corroborated the general conclusions of other research works which reported a decrease in SO2 and particulate matter and an increase in NOx emissions due to MEA emissions. Additionally, increased levels of smog, water consumption, and water toxicity were also calculated. Koornneef [5] stated that SO2 and solid particle emissions decreased due to CCS implementation, but NOx, NH3, and volatile organic compound emissions increased due to the utilization of amine and ammonia-based absorbents for CO2 capture. Petrakopoulou and Tsatsaronis [6] evaluated the environmental impacts of electricity generated by natural gas and coal power plants. The facts published by others [5,7,8] also revealed that CCS required additional energy consumption, leading to a decrease in power plant efficiency and a greater potential of fossil fuel consumption. One recently studied technology using CaCO3 as the solvent, was the CaO looping system. A comparative LCA study by Clarens et al. [9] showed that CaO looping decreased CO2 emissions by 73 percent. A suitable alternative to the absorption processes based on amines can be adsorption separation of CO2. Under certain circumstances, adsorption can exceed CO2 absorption if the corrosive absorption medium is replaced by a solid sorbent, the absorption media treatment is removed, and the operational costs are decreased due to lower energy consumption during the regeneration step as compared with the regeneration of liquid solvents.

Regarding the activated carbon (AC) adsorption process, our literature survey revealed a study published by Hjaila et al. [10] about LCA of the production of AC from olive waste cakes. This study highlighted the most significant impacts of acidification and eutrophication due to the use of H3PO4 and the electricity demand for the AC production process. The software used in this study was SimaPro 7.3 based on the Ecoinvent database and the assessment method was CML 2 Baseline 2000. Another recent study associated with LCA and AC was conducted by Arena et al. [11]. This study evaluated the LCA of the production of AC from coconut shells. The life cycle inventory was based on the Ecoinvent 3.0 database, using CML-2001 as the LCA method and the software GaBi 6.0. The results demonstrated that global warming potential, acidification, and human toxicity represent the most significant environmental impacts. The environmental burdens are mainly associated with the production of electrical energy (based on hard coal) used in the production process of AC.

Currently, in the Czech Republic, CCS is in the stage of technical drafts and optimization of the systems, as well as economic assessment of the optimized solutions. These studies can be subsequently supported by the evaluation of environmental gains and impacts using the LCA approach. The environmental impacts have already been assessed for two technical solutions in the Czech Republic, i.e., a power plant with ammonia scrubbing of CO2 [12] and a power plant with high-temperature carbonate looping [13]. These processes face various operational issues that could be omitted by using low-temperature solid sorbents, such as zeolites or activated carbon.

**Table 1.** Summary of life cycle assessment (LCA) studies on carbon capture and storage (CCS) technologies, including methods and results. PC, pulverized coal power plant; NGCC, natural gas combined cycle power plant; IGCC, integrated gasification combined cycle cycle power plant; SC-PC, subcritical pulverized coal; SPC-PC, supercritical pulverized coal; GWP, global warming potential; POP, photochemical oxidation potential; AP, acidifying potential; EP, eutrophication potential; PM10, PM-10 equivalents; ADP, abiotic depletion; ODP, ozone layer depletion; HTP, human toxicity; FAETP, freshwater aquatic ecotoxicity; MAETP, marine aquatic ecotoxicity; TET, terrestrial ecotoxicity.


This study aims to analyze an integrated system of Czech brown coal power unit with CO2 capture based on an AC system, drafted and optimized as part of a national-scale project [14]. The main goal of this study is to identify key environmental impacts and the economic feasibility of the unit with integrated capture of CO2. The study intends to use operational data from pilot testing of the CO2 capture method based on adsorption to evaluate environmental impacts on the national conditions. In order to perform a holistic and systematic evaluation, the LCA study was performed in different levels of decision-making processes such as characterization, normalization, Pareto analysis, and input-output analysis. This study was completed by the economical evaluation of such CCU (carbon capture and utilization) unit. In summary, the main contributions of the paper are:


• The identification of areas within a carbon capture technological process that can be improved to enhance environmental and economic performance.

The structure of this paper is comprised of the definition of a life cycle approach according to related international standards; the definition of the case study, i.e., description of Czech power unit and adsorption technology; implementation of the LCA methodology for the case study; characterization and interpretation of the environmental results; and finally, an economic evaluation of the case study.

#### **2. Methods**

#### *2.1. Environmental Assessment: The Life Cycle Assessment Method*

LCA is a tool to evaluate the environmental impacts of products and processes, such as the production of electricity. The LCA method is certified and defined by international standards ISO 14040 [15] as a cradle-to-grave analysis which facilitates a comparison of technological processes regarding their environmental characteristics. This includes all phases of a product´s lifetime. According to the international standards, LCA consists of four steps, i.e., goal and scope definition, inventory analysis, impact assessment, and interpretation which are described as follows:

**Step 1** Goal and scope definition

The depth of the analysis is determined by the goal of LCA. This study aimed to create a model and analyze the potential environmental impacts of CO2 adsorption on activated carbon connected to a 250 MW brown coal power unit. Therefore, two scenarios were considered:


For the comparison of the LCA results, a compatible functional unit must be defined for each scenario. The functional unit for both scenarios was defined as the power capacity (250 MW) of the power unit. System boundaries included the operational part of the power plant and activated carbon production, emission treatment, CO2 capture process, and waste generation. System boundaries excluded CO2 compression, transport, and final storage due to limited information about CO2 storage in the Czech conditions. Moreover, the environmental assessment included the operational part of the power plant rather than the life cycle of the CO2. Therefore, the approach used was considered to be "cradle-to-gate."

**Step 2** Life cycle inventory (LCI)

LCI starts with data collection and model construction, in compliance with the goal and scope definition, followed by the collection of input-output data and calculation of the resource depletion and emission release during the production process. Operational data for the case study was collected from the pilot project report [14] with detailed technical requirements and descriptions.

**Step 3** Life cycle impact assessment (LCIA)

LCIA can be divided into three steps, i.e., characterization, classification, and normalization. For the characterization and classification steps, the impact potentials were calculated. Normalization is an optional step of LCIA. Normalization uses a common reference impact to express results after the characterization and supports the comparison between alternative scenarios by using reference numerical scores. Normalization also gives a basis for comparing different types of impact categories [16]. The additional approach of Pareto analysis was chosen for defining the most significant impact categories. The selected method for the LCA analyses of the study was the ReCiPe 1.08 method. This method combines the problem-oriented approach of the CML method and the damage-oriented approach of EcoIndicator 99. The ReCiPe method is characterized by the following 18 midpoint indicators: ozone depletion (OD), human toxicity (HT), ionizing radiation (IR), photochemical oxidant formation (POF), particulate matter formation (PMF), terrestrial acidification (TA), climate change (CC), terrestrial ecotoxicity (TET), agricultural land occupation (ALO), urban land occupation (ULO),

natural land transformation (NLT), marine ecotoxicity (MET), marine eutrophication (ME), fresh water eutrophication (FE), fresh water ecotoxicity (FET), fossil depletion (FD), metal depletion (MD), and water depletion(WD); in addition, there are three endpoint indicators, i.e., human health, ecosystems, and resource surplus costs [17,18].

#### **Step 4** Characterization and interpretation

This step is based on the results of the LCIA phase. The results are defined as a potential environmental impact. The environmental impact is calculated using characterization methods that associate the scale of a pollutant emission to selected characterization factors. The interpretation of the results includes an identification of significant issues, evaluation of completeness, and sensitivity of results. The interpretation phase also includes key conclusions and recommendations [19]. Normalized results are further assessed by the statistical method of Pareto´s rule (80/20 rule), which states that 20 percent from all impact categories cause 80 percent of the total environmental impacts [20,21].

In our case study LCA, for the Pareto analysis, we chose values after normalization for each impact category.

In summary, the full LCA analysis was performed for both scenarios applying the ReCiPe method. Then, the characterization and normalization (according to ReCiPe 1.08, midpoint normalization of the Europe region) steps were performed to interpret the environmental impacts of the chosen scenarios. Additional Pareto analyses with more detailed input-output analyses of the specific processes were performed to identify the most significant processes which influence relevant environmental impacts.

#### *2.2. Economical Assessment and Economical Inventory of the Carbon Capture Unit (CCU)*

The economical evaluation of the CCU can have a significant impact on the actual feasibility of the project and contributes to the sustainability assessment of the technology. The calculation predicts the cost for the construction of a newly built commercial CCU and payback period. The economical inventory of the required construction materials was estimated based on previous projects made for the Czech market by the experts of the biggest Czech energetic research group (ÚJV Rež, a. s.). ˇ The evaluation was part of a national scale project [14] for the CCU application and connection to the 250 MW power unit.

#### **3. Case Study Definition: Reduction of CO2 Emissions in the Czech Republic**

To understand and define the technological boundaries for comparing the systems, it is important to describe both scenarios from a technical point of view. Scenario 1 defines the basic systems of the reference power plant. Scenario 2 is described by the reference power unit with the adsorption process of CO2 capture systems.

#### *3.1. Scenario 1: Reference Power Plant*

The first scenario considers the concept of the reference case scenario of the real Czech power plant. Mass and energy flows for further LCA analysis are related to the power plant's operation, which consists of three independent power blocks, each with an installed power of 250 MW. Each power block includes a dry bottom boiler, a turbine and its auxiliaries, a generator, a fly-ash separator, a cooling tower, a transformer, and a desulphurization unit. Coal feeding, water management (pipeline, pump, and chemical treatment), stack, auxiliaries for fly-ash handling, and desulphurization are shared systems. Since 1996, several equipment modernizations have been added in the power plant, such as a desulphurization system based on wet-limestone scrubbing. The gypsum, a product of desulphurization, is deposited into an adjusted mining dump site. Moreover, a hydraulic ash removal system has been replaced by deposition of a mixture of ash, gypsum, and wastewater into an adjusted mining dump site. The modernization includes the research and development of suitable CO2 sorbents for specific conditions. One of the most viable and commercially affordable options seems to be capture by activated carbon [22].

#### *3.2. Scenario 2: Activated Carbon Adsorption for Reference Power Plant*

The second scenario considers the same reference power plant with the connected CO2 adsorption unit. Mass and energy flows for the next LCA analysis include the operational phase of the power plant and the adsorption unit [23]. For the Czech power plant, the pilot adsorption facility was designed by the UJV Rez group and it was a pilot project [14] for the adsorption of operational flue gases of the power unit. The adsorption unit is based on a rotative adsorber (Figure 1). The main advantages are the continuous operation of adsorption, easy regulation of the adsorption process, and being a viable source of activated coal. The pilot facility for CO2 separation from flue gases by the adsorption was designed for continuous operation in the conditions of real flue gases from the lignite power unit. The concept is based on the rotational adsorption part, where the main functional part is the fixed adsorption bed connected to the motor driven rotor. The adsorption wheel rotates at a predetermined velocity around its own axis and the speed determines the time of the whole adsorption–desorption cycle. The CO2 separation follows the desulphurization process, where flue gases are purified and cooled by NaOH, and then the oxides SOx and NOx are removed. Cooled and purified flue gases pass through ventilators and through heating to the rotational adsorber for CO2 adsorption. Purified flue gases without CO2 are led out of the separation technology to the chimney or cooling tower. In the section of desorption, adsorbed CO2 is thermally displaced from the carbon sorbent by circulating gas heated by external steam. CO2 is continually transported for potential compression with 95% purity. The next step is the cooling of the sorbent to the requested temperature for adsorption. The whole process operates continuously by the rotation of the wheel. The pilot case rotative adsorber is pictured in Figure 1 [24]. The primary source of the activated carbon in Czech conditions is assumed to be hard coal from the CSM mine site (Karvina region) with an annual mining of six million tons of hard coal [ ˇ 2]. The activation of the hard coal is based on two main steps, i.e., carbonization of the raw material in the absence of oxygen and activation of the carbonized product with water vapor. The heat supply necessary for the activation is obtained by combustion of gases produced during activation.

**Figure 1.** Case rotative adsorber.

#### *3.3. LCA Study: System Boundaries Definition*

As stated previously, this study aims to identify the environmental impacts of a power plant with a carbon capture system integration using the designed adsorption method and comparing those impacts with a reference power plant without an adsorption system. The study also focuses on assigning the environmental impacts to the designed CO2 capture technology itself. The system boundary for Scenario 1 (Figure 2) includes a brown coal production chain from the mining process, transport of fuel to the power unit with power production, combustion, and flue gas treatment processes. Scenario 1 also includes waste and gas emissions production (residuals of flue gas after treatment as nitrogen oxides, carbon dioxide, and sulphur dioxide, released through a stack into the air). Scenario 2(Figure 3) includes a brown coal production chain, power unit operation with power and waste production, adsorption process with all relevant inputs, such as activated carbon production and production of 40% NaOH, and finally, output flows from the adsorption process (captured CO2, gas emissions and waste products). Although the CCS chain also includes transport and storage of captured CO2, our specific study does not include any operational data for transport and storage of CO2, as there is no specific solution of CO2 transport and storage in the Czech Republic and the distance from an emission source to a storage site with adequate storage capacity and lifetime is unique for every case.

**Figure 2.** System boundaries for Scenario 1.

**Figure 3.** System boundaries for Scenario 2 (differences from Scenario 1 are marked in red).

#### *3.4. Life Cycle Inventory*

The input data for the power unit was based on the real power plant operational parameters. Its characteristics are given in Table 2. The heat and mass balances for CO2 capture technology (Table 3) were evaluated in relation to the power plant characteristics [14] and the results obtained from the pilot testing of the adsorption method.

**Table 2.** Characteristics of power unit without CO2 capture.



**Table 3.** Input data for the CO2 adsorption process.

In addition to the process of adsorption, the process of activation and carbonization of hard coal is also required to be included in the model. The input data was calculated for the initial batch of 7.6 t of hard coal and 76 t of tar (then activated and transformed into activated carbon). The energy consumption for the hard coal activation was calculated as 1133 MJ. Before the introduction of the flue gas from coal combustion into the CO2 adsorption stage, it must be secondary treated to decrease the amount of acid gases. To do so, flue gas after coal combustion is washed with a NaOH solution. The composition of the reacted products (output flows) after the reaction with NaOH is described in Table 4.

**Table 4.** Composition of products after the reaction between flue gas compounds and NaOH.


The emission gases released into the air after the CO2 adsorption stage, the amount of wastewater, and of captured CO2 are depicted in Table 5.


**Table 5.** Output data from the CO2 adsorption process.

Moreover, other conditions and assumptions in the LCA model were taken into consideration with respect to the technical concept of CO2 adsorption technology and its energy and mass balances as follows:


#### *3.5. Economical Calculation*

For the calculation of the CO2 capture unit construction and connection to the average power plant, the following parameters were assumed:


The cost estimation of the construction of CCU is calculated according to the price of the required appliances. Then, the capital expenditure (CapEXP) is multiplied by the coefficient 1.7 (conservative estimate), which refers to the assumption that the construction is built as a fully new technology, and therefore some unexpected expenses could arise.

The operational costs (OpCost) for each item are difficult to estimate. Therefore, 5% of the capital expenditures were used as the operational cost value per year.

Incomes (In) are calculated from the cost of saved CO2 allowance for each ton of CO2 and the current market price of CO2. The authors used a more pessimistic scenario according to the EU commission reference for CO2 allowance of 25 Euro/t CO2 and the actual market price which is estimated to be 120 Euro/t CO2. Then, the payback period (PBT, simple, not discounted) for the commercial CO2 capture unit is calculated as following:

$$\text{In} = \text{CO}\_2 \cdot \text{captured} \circ (\text{CostCO}\_2 \cdot \text{allowance} + \text{CostCO}\_2 \cdot \text{market}) \tag{1}$$

$$\text{PBT} = \frac{\text{CapEXP}}{(\text{In} - \text{OpCost})} \tag{2}$$

#### **4. Results**

#### *4.1. Life Cycle Impact Assessment*

Steps 3 and 4 of LCIA are involved in the Results section. First, each scenario is analyzed separately, and then compared to each other. The results of both considered scenarios are represented in Table 6. Table 6 summarizes the environmental impact categories into the following three groups of results: values in category units, normalization results without any units, and the relative contribution of each impact category to the sum of all categories. The relative contributions are computed from the normalization values.

According to the results, the relative contribution to the sum of impacts in Scenario 1 shows the highest contribution of 46.81% by fossil depletion and in 29.27% by climate change potential. In addition, almost 10% of the environmental impact contribution refers to the terrestrial acidification potential. For Scenario 1, the highest contribution of 77.41% is shown by fossil depletion. All other potential environmental impacts refer to much smaller contributions. For the comparison of both scenarios among the environmental impact categories, the normalization level of the decision-making process was considered. Further Pareto analysis describes the difference between values among the scenarios.


#### **Table 6.** LCA results for both scenarios.

#### *4.2. Pareto Analysis of the Scenarios and Processes*

The Pareto analysis defines 20% of the potential environmental impact categories that contribute to 80% of the total impact. The environmental impacts were divided into the following two groups of flows: (1) input flows which use, consume, or transform primary soil, land, or resources (agriculture land occupation, natural land transformation, and fossil depletion) and (2) output flows which are considered to be emissions from the considered processes. In the case of the input flows, for both scenarios, all the mentioned environmental categories have equal values and the highest among them has fossil depletion.

Figures 4 and 5 illustrate the most significant environmental categories for output flows by Pareto graphics. On the one hand, Scenario 1 identifies fossil depletion (FD), climate change (CC) potential, and terrestrial acidification (TA) as the highest contributors. On the other hand, Scenario 2 shows that the terrestrial acidification has a higher impact value than CC. For the fossil depletion category, the brown coal mining is shown as having the highest contribution. Therefore, the climate change category is affected mainly by the combustion of brown coal and thermal energy production for the power unit operation.

For the comparison of both scenarios the following graphics (Figures 6–9) represent differences in the significant (CC, TA, PMF, and POF) environmental categories. The graphs show lower impacts in each category for Scenario 2. The most significant difference is seen in CC where the values for Scenario 2 are almost two-thirds lower than those in Scenario 1.

**Figure 5.** Pareto graph for scenario 2.

**Figure 6.** Comparison of scenarios-climate change.

**Figure 8.** Comparison of scenarios- particular matter formation.

**Figure 9.** Comparison of scenarios -photochemical oxidant formation.

For Scenario 2, Table 7 shows that the category of fossil depletion is affected mainly by brown coal production and mining. Moreover, terrestrial acidification potential is mainly affected by the CO2 adsorption process. Brown coal combustion and, consequently, production of thermal energy for the adsorption process are also contributors to the acidification potential. Transport has a minor role for both scenarios


**Table 7.** Causes of significant environmental impacts in Scenario 2.

Environmental Impact Assessment of Activated Carbon Production

The question of the activated carbon production is crucial for the whole environmental impact analysis. Therefore, further detailed input-output analyses of the process of the activated carbon production is required. Table 8 shows the relative contribution of each category for the activated carbon production for the CO2 adsorption process. The highest contribution refers to the categories of climate change potential and fossil depletion.

**Table 8.** Environmental assessment of activated carbon production.


*4.3. Economical Evaluation of the Payback Period of the Pilot CO2 Capture Unit*

Total capital expenditures according to the required components price are listed in Table 9.


**Table 9.** Total costs of CCU construction.

Results in Table 10 for the economic feasibility and payback time period were calculated according to Formulas (1) and (2).


**Table 10.** Payback period (simple, not discounted) for CCU.

#### **5. Discussion**

According to the characterization values in Table 6, the most obvious difference between the two scenarios is the climate change category (Figure 6). This result corresponds to the decrease of CO2 levels by the adsorption process (in Scenario 2), modelled at 75% CO2 capture from flue gases. The gains in terrestrial acidification potential in Scenario 1 have higher values (227 kg SO2 eq.), which are mainly influenced by SO2 emissions (119 kg) released into the air after the flue gas treatment. In Scenario 2, this amount of SO2 in flue gases is the input into the adsorption process, thus, the values for the acidification category are lower. In addition, Scenario 1 refers to the higher values in the PMF and POF categories. These categories are influenced by fuel combustion emissions.

Studies by Kantová [25] and Vassilev [26] showed that ash from the brown coal combustion consisted of a high volume of particulate matter (93.08% from brown coal) which was volatile and persistent in the atmosphere. Therefore, brown coal combustion and the quality of brown coal are significant factors that influence the level of potential environmental harm. According to Kantová, the key parameters for controlled emissions in the process of combustion are low ash content, low moisture levels, and a constant size of volatile particles. In the context of the Czech integrated register of pollutants and emissions restrictions [27], there are no further chemical descriptions of particulate matter (PM), and thus it is complicated to get parameters of produced PMs that directly affect the environmental category of particulate matter formation. Therefore, the chemical analysis of PM produced from Czech brown coal would be an interesting subject for further research.

The Pareto analysis was chosen to distinguish which processes have the greatest impact, particularly for output flows (emissions) in both scenarios which are harder to decide upon. The input flows, which consider the change of the land and resource depletion, show the highest contribution of fossil depletion potential in both scenarios. In both cases, fossil depletion is related to the processes of brown coal mining and production which require 214 t/h of primary brown coal for the actual operation of the studied power unit. However, the result values for fossil depletion in Scenario 1 does not show any significant difference as compared with Scenario 2, although there is a slightly higher demand for raw material extraction, such as hard coal for activated carbon production. This demand refers to a need for fresh carbon, in the amount of 23 kg/h, which in the evaluation of the whole life cycle means only a small resource demand. The Pareto analysis for output flows (Figures 4 and 5) considers fossil depletion, climate change, and terrestrial acidification as the most significant contributors to the overall environmental impacts. The only difference lays in the degree of priority for each scenario. For both scenarios, fossil depletion has equal significance, as the values of characterization are the same. In Scenario 1, the impact contribution in second place is climate change and, in Scenario 2 it is terrestrial acidification. Climate change in Scenario 1 is mainly caused by the combustion process of the coal and the production of thermal energy for the power unit. For Scenario 2, Table 7 shows that the processes influencing the category of terrestrial acidification are brown coal mining and SO2 emissions (released into the air after the adsorption process in the amount of 82.3 kg SO2 eq). The impact of transportation, in both scenarios, is insignificant.

It is also important to mention the category of water consumption. In the relative contribution to environmental impacts, this category does not show any significance. However, in the characterization phase, the consumption of 6400 m<sup>3</sup> of water shows that the energy industry plays a role in water management. This is a reminder that the process must also aim to mitigate excessive water consumption, especially in times of global warming and drought danger.

The next step was to assess the production of the activated carbon. The literature review for the LCA of the activated carbon adsorption process showed that environmental studies have focused on the type of activated carbon production. The case study is considering the production of activated carbon from hard coal as a primary source. Therefore, resource depletion, as a direct connection to the category of fossil depletion potential, is affected by hard coal mining. Tar, as an input flow for the activated carbon production, also contributes to resource depletion. On the site of emissions, CO2 emissions from the combustion of natural gas (1133 MJ for activation and carbonization of 7.6 t

of activated carbon) cause major environmental consequences. Moreover, among all environmental impact categories, climate change has the highest relative contribution (almost 99%). The impact relates to the combustion of natural gas (Table 8).

The results clearly demonstrate that the power unit with the connection of adsorption process leads to decreased environmental impacts, specifically in the categories of climate change, terrestrial acidification, particulate matter formation, and photochemical oxidant formation. The problem is seen in a primary source, i.e., coal extraction, which, in both scenarios, shows relatively high and equal values. The Czech national energy mix is based on brown coal power plants, and therefore the raw materials extraction and resource depletion creates a significant environmental burden. The extraction of hard coal for activated carbon production also contributes to this fact. The case study counts with just an input of 23 kg fresh activated carbon, but the basic batch of hard coal is rather bigger and counts with 7.6 t, which is an additional amount of raw materials extracted from the ground. The production of the activated carbon could be optimized using different resources such as biomass, which could contribute to reduced consumption of the fossil source.

Finally, the economical consideration (Tables 9 and 10) of the newly built CCU shows a payback period of almost six years (relatively fast for such a small CCU). It must be considered that the market price for a ton of CO2 could be lower, due to lower purity of the CO2 product. If the market price was one-third lower (40 Euro), incomes would change to 78,000 Euros and the payback period would increase to almost 18 years. We conclude that, even if the process of CO2 capture was highly effective, the purity of the final product has a significant role in the economy of the whole process. To make the project feasible, there is a technical requirement to solve the purification process of CO2, leading to an increase of total capital expenditures and of the payback period. Moreover, purified CO2 as a final product could be more attractive for sectors such as agriculture or the food processing industry, and therefore contribute to the national circular economy.

#### **6. Conclusions**

Carbon dioxide capture by activated carbon adsorption seems to be a promising technology from an environmental perspective. The LCA assessment highlights the main environmental impacts that can arise during the life cycle of the technology.

The robust LCA analyses which included characterization, normalization, and Pareto analyses with input-output analyses are approaches that create a precise model to reflect specific conditions of the technology. The LCA helps to identify the key processes that can be improved with respect to their environmental performance. In addition, the economic calculation completes the sustainability assessment of the newly built technology and gives the perspective of the final product (CO2) utilization. It must be stressed that the designed emissions of the capture method are site specific and reflect the local conditions, for example, the type of power plant, fuel type, natural sources (for capture media production), etc. The presented adsorption method was designed for the purpose of CO2 capture from subcritical, coal-fired power plants in the Czech Republic. The sum of the environmental impacts with carbon capture is generally lower than the power unit itself. Nevertheless, this study shows that the Czech energy mix (in both scenarios) leads to high levels of CO2, SO2 emissions, and solid particulates. As the Czech national energy mix is primarily from brown coal, the depletion of fossils by a primary energy source still remains the main environmental problem, but monitoring of the coal quality, as well as testing the chemical composition of particulate matter, could contribute towards lower potential environmental impacts.

Nevertheless, further research that focuses on various sources (as biomass) for activated carbon production should be considered. Moreover, CCU could become part of the Czech circular economy, if the purification processes and measures of the CO2 product are wisely chosen and adapted.

Worldwide pressure for low carbon economy transition is forcing the national energy systems to find viable solutions to mitigate the levels of greenhouse gases (GHG). The Czech Republic is slowly shifting towards increased integration of renewable energy systems. However, the coal industry is still

the prevailing sector, where a sudden shift could be drastic for the national economy and coverage of the power demand. Therefore, the current systems with optimized CCU could be the solution that would help to overcome the transition process. The implementation of all available tools and knowledge to reach this goal is required and would assist choosing and creation of a reasonable and wise strategy for the sustainable development of the country.

**Author Contributions:** Conceptualization, K.Z. and A.C.; methodology, K.Z. and V.K.; formal analysis, K.Z.; resources, K.Z.; data curation, K.Z. and J.Š.; writing—original draft preparation, K.Z.; writing—review and editing, K.Z.; supervision, V.K. and A.C.; All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** This work was supported by the Technology Agency of the Czech Republic, project number TH03020169 and project number TA02020205 and by institutional support from the University of Chemistry and Technology Prague.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **A Prospective Net Energy and Environmental Life-Cycle Assessment of the UK Electricity Grid**

**Marco Raugei 1,2,\*, Mashael Kamran 1,2 and Allan Hutchinson 1,2**


Received: 30 March 2020; Accepted: 27 April 2020; Published: 2 May 2020

**Abstract:** National Grid, the UK's largest utility company, has produced a number of energy transition scenarios, among which "2 degrees" is the most aggressive in terms of decarbonization. This paper presents the results of a combined prospective net energy and environmental life cycle assessment of the UK electricity grid, based on such a scenario. The main findings are that the strategy is effective at drastically reducing greenhouse gas emissions (albeit to a reduced degree with respect to the projected share of "zero carbon" generation taken at face value), but it entails a trade-off in terms of depletion of metal resources. The grid's potential toxicity impacts are also expected to remain substantially undiminished with respect to the present. Overall, the analysis indicates that the "2 degrees" scenario is environmentally sound and that it even leads to a modest increase in the net energy delivered to society by the grid (after accounting for the energy investments required to deploy all technologies).

**Keywords:** LCA; EROI; net energy; energy scenario; energy transition; electricity; grid mix; storage; decarbonization

#### **1. Introduction**

Energy is an essential commodity for productivity, economic growth and well-being. However, ensuring the delivery of the energy that societies need while preventing major environmental disruption is quickly becoming one of the major challenges of this century. The worldwide increase in energy demand and continuous burning of fossil fuels for energy production since industrialization has led to a rapid increase in carbon dioxide (CO2) concentration in the atmosphere from 280 ppm (parts per million) to over 400 ppm [1].

Historically, the UK was one of the first countries to experience massive industrialization across all sectors. Before industrialization, most of the UK energy demand was fairly renewable, met by water and wind mills and by burning woody biomass, mostly for agricultural and heating purposes [2]. Subsequently, with the onset of industrialization, the UK population and its appetite for energy grew rapidly. The UK economy shifted from reliance on agriculture to manufacturing, and most manufacturing sectors were based on coal consumption. This propelled UK economic growth in the 18th and 19th century [3] but, as a consequence, it also led to a rapid increase in CO2 and other greenhouse gas (GHG) emissions to the atmosphere. The burning of coal and peat for manufacturing and heating purposes created a visual smog in major UK cities, too. This led to the first 1956 and 1968 Clean Air Act to reduce the emissions caused by coal [3]. In 2008, the UK legislated the Climate Change Act to reduce its contribution to global warming by cutting down GHG emissions by 80% by 2050, relative to its 1990 level [4].

While the UK was not the only country with legislation already in place aimed at reducing GHG emissions, the Paris Agreement held in December 2015 was the first global agreement to combat climate change signed by 195 countries. It aims to keep the global temperature below 2 ◦C above pre-industrialization level, while encouraging further efforts to limit global warming to below 1.5 ◦C to reduce the risk of irreversible climate consequences [5]. To accomplish this target, all parties of the United Nation Framework Convention for Climate Change (UNFCCC) understood the necessary action to limit the CO2 and other GHG concentration in the atmosphere by decarbonizing their current energy system. In 2019, the UK Committee on Climate Change (CCC) recommended a new national target to fulfil the commitment made when signing the Paris Agreement. This new ambitious target aims to bring all GHG emissions to net zero by 2050 [6].

The current UK energy system comprises of three main energy carriers: oil-derived fuels, natural gas and electricity. Electricity is known to be one of the most desirable energy carriers for a low-carbon future due to its flexibility to operate a wide range of services such as lightning, heating, and transportation, with its transportability over long distances with comparatively little loss, and the efficiency with which it can be converted into useful work at the point of use [7]. Electricity can be generated from a range of primary sources, and in order to meet the climate target, there has recently been an increased deployment of various low-carbon technologies in the UK electricity grid (primarily wind and solar photovoltaics), with an on-going shift away from the combustion of fossil fuels in thermal power plants, and a concomitant increase in the electrification of the heating and transportation sectors. Such trends are expected to be sustained for the coming decades, with a positive overall effect in terms of decarbonization; however, the increased reliance on electricity as the energy carrier of choice for all sectors, paired with the shift to more and more variable renewable energy (VRE) generation, has also led to growing concerns about the future requirement for large-scale energy storage (including lithium-ion batteries) and the possible associated adverse consequences.

Over the past decade, a growing body of literature has analyzed the feasibility and environmental consequences of various energy transition pathways characterized by an increased degree of electrification across all sectors, coupled with a larger reliance on renewable technologies to generate the required electricity. Among such studies that appeared relatively early, a few attracted significant attention by non-specialist media, such as Fthenakis et al.'s seminal "Solar Grand Plan" [8,9], while others seemed to polarize scientific opinion and generated strong controversy [10–14]. More recently, a string of similarly-framed studies respectively focusing on Europe [15], the Americas [16] or the whole world [17,18] have seemingly pointed to a degree of high-level agreement on the general desirability of such transition pathways, albeit with a number of caveats.

Crucially, although key electricity generation technologies such as wind, solar photovoltaics and nuclear are often considered to be "zero carbon" at the point of consumption, the same is not true for their manufacturing, nor are they necessarily 100% environmentally sound in all respects. Life cycle assessment (LCA) is the preferred methodology in order to address all these points while taking into consideration all the life-cycle stages of the various energy supply chains (resource extraction, processing, and delivery, and power plant manufacturing, operation and decommissioning). For instance, a comprehensive LCA of a range of International Energy Agency (IEA) world electricity supply scenarios indicated that, in broad terms, those scenarios that rely more heavily on renewable technologies do indeed tend to stabilize or even reduce global pollution, but entail larger (and sometimes potentially critical) demand for key materials, noticeably among which is copper [19]. Another recent study, whose scope also extended to the whole world, found that those decarbonization strategies relying heavily on wind and solar technologies are comparatively more effective in reducing human health impacts than those relying on carbon sequestration, while the use of bioenergy in the mix raises concerns in terms of land use and associated ecosystem damage [20]. Pehl et al. [21] carried out a thorough life-cycle comparison of all the direct and indirect GHG emissions from various renewable and non-renewable low-carbon technologies, also including seldom-considered factors such as indirect carbon emissions from land use change. Their study produced ranges of results for each technology that reflect the sometimes-large variability that is not only due to technological aspects, but also, critically, system siting (from insolation level for solar photovoltaics, to the rather less obvious methane emissions from biomass degrading in poorly-chosen hydro reservoir sites).

Coming to more UK-centric LCA studies, Stamford and Azapagic [22] compared a range of possible grid mix scenarios up to 2070, and concluded that a low-carbon mix with nuclear and renewables provides the best overall environmental performance, albeit with some increased impacts in terms of terrestrial eco-toxicity, and material resource depletion (elements). Raugei et al. [23] investigated a number of stakeholder-informed photovoltaic-heavy grid mix options, and found that, despite the comparatively low insolation levels that are characteristic of the UK, no such scenarios would be detrimental to the grid performance from a wide range of technical and environmental metrics.

There also remains a question about whether the future electricity grid can continue providing the same level of "net" energy for all societal needs as before, to support continued economic growth and prosperity. The concept of net energy was first described by ecologist Howard T. Odum in 1973 [24] and then reprised by his former student Charles A.S. Hall, who went on to coin the popular term "energy return on investment" (EROI) [25]. Net energy is intended to represent the "true value" of the energy that is delivered to society. In other words, for a society to prosper, the energy delivered to the society must at least be higher than the energy invested in the chain of processes required to convert primary energy resources into useful energy carriers. Carbajales-Dale et al. [26] discuss the importance of carrying out net energy analysis (NEA) to measure the productivity of the energy system and support a sustainable energy transition. Moving towards a low carbon future would require a different re-investment of energy for the electricity grid due to the deployment of new energy generation and storage technologies, and NEA is the way to identify the ensuing changes in the final net energy delivered to society by the electricity grid.

To this date, however, literature is divided as to this important question, with some studies questioning the capability of future renewable-heavy grid mixes to deliver sufficient net energy (e.g., for the case of Australia [27]), and others instead concluding that the net energy set-backs could be minor (e.g., for New York state [28]) to non-existent (e.g., for Chile [29]). Ultimately, the devil is, once again, in the details, and the answer is likely to depend on specific conditions such as the exact grid mix composition, location, degree of gird-level storage, and the demand profile.

This paper presents the results of a joint environmental LCA and NEA of one of the leading energy transition scenarios for the UK, specifically to understand the changes that may be expected in environmental impacts as well as net energy delivery when moving from the current UK grid mix to a future one featuring larger amounts of VRE and associated energy storage (the latter including dedicated stationary solutions as well as mobile—also referred to as "vehicle-to-grid"—schemes).

#### **2. Materials**

#### *2.1. Current and Future (Projected) UK Grid Mix*

National Grid Electricity System Operator, the UK's largest utility company, produced a study called "Future Energy Scenarios (FES)" [30] analyzing the energy demand and supply of the future electricity and gas networks in Great Britain based on policy support, customer engagement, technological development, economic growth and energy efficiency (National Grid does not provide electricity to Northern Ireland, which instead is served by three interconnectors with the Republic of Ireland. Consequently, while the FES are based on targets set at the national (UK) level, all the data contained therein are actually for Great Britain (GB) only, i.e., excluding Northern Ireland. For the sake of clarity and simplicity, however, in this paper the distinction between GB and UK is dropped in favour of a unified reference to "UK" throughout). Four potential energy pathways to achieve various degrees of decarbonization for the UK are described in the FES report, respectively named: "steady progression", "consumer evolution", "community renewables" and "2 degrees". Out of these, "community renewables" and "2 degrees" are the more aggressive in terms of decarbonization. Both target 80% reduction in GHG emissions by 2050 compared to 1990, based on the 2008 Climate Change Act, and the main difference between the two is that "2 degrees" assumes more "large-scale", centralized electricity generation, while "community renewables" relies more on smaller, decentralized

installations. This paper focuses on the "2 degrees" scenario, due to it being the one for which a greater level of detail is provided in the FES report.

The FES "2 degrees" scenario projects electricity generation capacities for all the grid mix technologies and the associated annual electricity generation for the next 30 years. The overall electricity generation is estimated to increase by 35% in 2035 and 72% in 2050, relative to 2019, in response to the increasing demand partly due to the electrification of transport and heating. Figure 1 illustrates the UK grid mix composition in terms of total electricity generated in the year 2019, and Figures 2 and 3 the corresponding projected grid mix compositions in the years 2035 and 2050.

**Figure 1.** Current (2019) UK grid mix composition, in terms of total electricity generated. NGCC = natural gas combined cycles; PWR = pressure water reactors.

**Figure 2.** Expected UK grid mix composition in 2035, in terms of total electricity generated (under "2 degrees" scenario assumptions). NGCC = natural gas combined cycles; NGCC + CCS = natural gas combined cycles plus carbon capture and storage; PWR = pressure water reactors; SMR = small modular reactors.

**Figure 3.** Expected UK grid mix composition in 2050, in terms of total electricity generated (under "2 degrees" scenario assumptions). NGCC = natural gas combined cycles; NGCC + CCS = natural gas combined cycles plus carbon capture and storage; PWR = pressure water reactors; SMR = small modular reactors.

The 2019 grid mix includes a small share of coal-fired electricity, which is expected to be completely phased out within the next few years. Gas-fired generation, currently accounting for 25% of the grid mix, is expected to move towards a peaking plant role in the coming years to provide flexibility in the electricity grid by supplying electricity when the demand is high or when renewable generation is low. While most conventional combined-cycle gas-fired generation is also expected to be phased out (−92% by 2050), a non-negligible share of gas-fired generation is expected to re-emerge from 2030 onwards but equipped with carbon capture and storage. Nuclear generation in 2019 provides 22% of the total electricity generation, but over half of the existing power plants are expected to reach their designated end of life and close down before 2035. However, small modular reactors are then projected to take over, eventually providing around 16 GW of generation capacity, and 13% of total electricity generation, by 2050.

By 2035, 99% of electricity generation is expected to come from "zero carbon" technologies at the point of use (i.e., nuclear, biogas, biomass, waste, hydro, tidal, wind and solar)—cf. blue line in Figure 4. Out of this, a large proportion will be met by renewable technologies (cf. green line in Figure 4), and especially by offshore wind (accounting for approximately 44% of total generation in 2035 and 2050). In fact, most renewable technologies (wind, solar, tidal, and biogas) are projected to increase significantly in output over the next decades, with the sole exceptions of hydro, which is already essentially maxed out in the UK, and biomass, which is expected to drop significantly by 2035. It should be mentioned that in the "2 degrees" scenario tidal generation is expected to account for 76% of total marine generation, with the remaining marine output to be provided by wave technologies; however, due to the lack of data on the latter, in this study, the simplifying assumption was made to consider all marine generation to be tidal.

More specifically, the yellow line in Figure 4 shows that approximately 68% of total generation in 2035 and 2050 is expected to be provided by VRE technologies (i.e., wind, solar and tidal), which, as the phrase implies, are inherently intermittent. As the cumulative share of these technologies increases over the next decades, the grid will thus require increased levels of energy storage to provide the required flexibility to match the energy demand profile.

Storage technologies will therefore increasingly be deployed alongside VRE, to provide such flexibility, as shown by the red line in Figure 4.

**Figure 4.** Expected trends (under "2 degrees" scenario assumptions) for, respectively: % of total generation that is considered "zero-carbon" at point of use (i.e., nuclear, biogas, biomass, waste, hydro, tidal, wind and solar); % of total generation that is considered "renewable" (same as above but excluding nuclear); % of total generation that is classified as "Variable Renewable Energy" (VRE; i.e., wind, solar and tidal); % of VRE generation that is sent to energy storage (as opposed to used directly).

National Grid's "2 degrees" scenario considers the deployment of five energy storage technologies: pumped hydro storage (PHS)—in large part relying on existing dammed hydro installations located overseas and tapped via interconnectors, compressed air energy storage (CAES), dedicated lithium-ion batteries (LIB), vehicle-to-grid (V2G) storage provided by the electric vehicle (EV) fleet, and liquid air energy storage (LAES). The first two technologies are generally capable of providing longer-duration storage than the latter three. Specifically, LAES is a fairly new technology that is only projected to ultimately provide 4% of the total storage capacity; due to the very limited information available on its supply chain, in this study the associated storage capacity was instead lumped together with that provided by LIB. The resulting specific assumptions on storage technologies for the future of the UK grid mix are summarized in Table 1.



<sup>1</sup> Own assumption.

Electricity transmission was also included within the boundary of this assessment, albeit limited to the high voltage (HV) network. This was deemed an acceptable simplification, since the vast majority of the electricity generation plants comprising the grid mix at present and in the "2 degrees" scenario are large centralized units which inject HV electricity into the grid (the only exception being

rooftop-mounted PV systems). The current HV transmission lines were modelled using the UK-specific life-cycle inventory (LCI) information provided in the Ecoinvent database [33], and then scaled for 2035 and 2050 by simple linear extrapolation based on the projected total gross electricity generation. In terms of energy balance, transmission losses were conservatively set at 5% [34].

Finally, electricity interconnectors with the continent are expected to provide net electricity imports during the demand peaks in winter in the early 2020s. Such imports are, however, expected to decrease significantly in the following years, and to become almost negligible by 2035, eventually leading to very slight net exports by 2050. The choice was therefore made to limit the scope of this analysis to the electricity generated and delivered domestically within the UK, disregarding all electricity exchanged via the interconnectors (with the sole exception of a relatively small share of the future VRE that is assumed to be sent to be stored in PHS facilities located on the continent).

#### *2.2. Electricity Generation and Storage Technologies*

#### 2.2.1. Coal

Currently, there are six coal-fired power stations in UK; all are expected to shut down by 2025, except for Drax Power Station, for which there are plans to convert its units to biomass- and gas-fired generation in the near future [35]. The Ecoinvent LCI database "GB" (Great Britain) hard coal electricity production model was adopted for the life cycle inventory of coal-fired electricity generation in the UK.

#### 2.2.2. Natural Gas Combined Cycles

UK natural gas combined cycle (NGCC) inventory was based on the corresponding Ecoinvent GB database model. The model output was reduced by 19% to account for the difference in assumed vs. reported natural gas energy input per kWh of electricity output at the power plant stage, as well as the actual average heating value of the UK gas feedstock [36].

#### 2.2.3. Natural Gas Combined Cycles Plus Carbon Capture and Storage (Future Technology)

At present, there is no detailed inventory information available for natural gas combined cycles with carbon capture and storage (NGCC-CCS). Literature data on the carbon capture process were adopted for 90% CO2 capture, and incorporated in the adjusted model for conventional NGCC plants. The electricity output of the plant was reduced by 15% to account for the CO2 capture and compression process (in addition to the initial output adjustment described in Section 2.2.2) [37,38]. The life cycle inventory for the transportation of CO2 captured and stored is excluded from the study due to the uncertainty involved in the location of the plant and possible storage size.

Appendix A—Tables A1 and A2 provide the construction and operational inventory quantities and the adjusted emissions per kWh of electricity generated.

#### 2.2.4. Biomass

The biomass used for electricity generation in the UK mainly consists of wood pellets from North America forestry, domestic wood chips from UK forests, and a small percentage of domestic residues [39,40]. According to the digest of UK energy statistics (DUKES) 2019 report, based on the 2018 renewable flow-chart, 45% of the biomass share was domestically sourced and 55% was imported [36]. The GB heat and power co-generation model from the Ecoinvent database was selected to represent biomass electricity generation. However, this model does not include mixed inputs of woody biomass feedstocks, and it was therefore modified to account for such. Specifically, the Ecoinvent "RER" (regional European) wood chip model was used to represent the share of wood chips used (since there is not one available specifically for the UK), and two additional processes were added to the model to account for the wood pellet imports. The first process accounts for wood pellet production, and the second one for the transportation of the wood pellets from North America by freight ship. The transport

distance was assumed to be 5300 km [41] and the mass of the dry pellets was multiplied by a factor of 1.2 to account for the average moisture content in the transported pellets [42].

#### 2.2.5. Biogas

Biogas is a mixture of gases (mainly methane and carbon dioxide, but with multiple trace gases such as ammonia and hydrogen sulphide) produced by the anaerobic degradation ("digestion") process of organic waste from landfills and, to a lesser extent, agricultural sludge. The biogas is fed into a combined heat and power plant station (CHP), where it is combusted to generate electricity and heat [43]. The Ecoinvent GB biogas-fired heat and power co-generation model was selected to assess biogas electricity generation, and all energy and environmental impacts of the CHP multi-output process were allocated on an energy content basis.

#### 2.2.6. Waste

Electricity generation from waste incineration is also a co-product of a multi-output process. According to ISO recommendations [44], in this case system expansion was adopted in preference to allocation, since waste incineration with energy recovery represents an almost textbook example in which: (i) a primary function is clearly identified (i.e., getting rid of the waste), and (ii) a comparable alternative process exists which delivers only one of the two outputs (i.e., an incinerator without energy recovery). Consequently, the energy recovery process and associated electricity generation was calculated to have negligible emissions assigned to it, since the only additional up-front inputs required vs. the incinerator without energy recovery are those for the boiler and turbine system, while the use-phase emissions at the stack are virtually the same.

#### 2.2.7. Nuclear

There are 15 operating nuclear reactors in UK, 14 of which are advanced gas-cooled reactors (AGR) which are expected to shut down before 2035, and one is a large pressurized water reactor (PWR) which was initially also expected to shut down in 2035, but whose operation may be extended for 20 more years [45]. Out of the currently operating nuclear reactor technologies in the UK, life cycle information in Ecoinvent was only available for PWRs, and therefore the latter was used to model the life cycle inventory associated to nuclear electricity generation

#### 2.2.8. Nuclear (Small Modular Reactors—Future Technology)

Small Modular Reactors (SMR) are factory-built nuclear reactors of less than 300 MWe installed power, inspired by the current large nuclear power plants [46]. They are also known as integral PWR since their main components, such as the stream generator, reactor and pressurizer, are all located in one vessel. SMRs offer the opportunity to add nuclear generating capacity with a smaller capital cost and thus reduce construction risks. They can be categorized in two groups: (1) Generation III water-cooled SMR based on existing large nuclear plants but on a smaller scale, and (2) Generation IV SMRs based on the use of novel fuels and coolants, which can provide other services such as heat for industrial processes [47]. Generation IV small modular reactors are not expected to achieve commercial maturity until 2030 onwards [48], while Generation III SMR are considered to be more mature technologies as they are based on the existing large nuclear plants concept. According to the world nuclear association there are currently two potential SMR projects, one with NuScale and other with Rolls Royce both based on light-water pressurized SMR designs [45].

At present there is no existing LCI database model for SMR technologies, therefore, in this study the Ecoinvent model for PWRs was adopted as a basis for the life cycle inventory of light-water pressurised SMR. The latter are the scaled-down version of large PWR which utilize the same working concepts, but instead of having pumps and coolant loops for directing the flow of water, they utilize natural circulation to direct the cool water to the reactor core after going through the steam generator to turn the turbine to generate electricity [49]. The main components of both systems are considered to

be the same, and both are expected to have the same lifetime of 60 years. The Ecoinvent model was adjusted to account for the reduction in efficiency and improvement of the capacity factor (CF) for SMRs compared to large PWRs, which lead to an overall reduced output (−15% in relative terms) [48] (The Capacity Factor is defined as the ratio of the actual power generated by a system, averaged over its service lifetime, to its nominal installed power).

#### 2.2.9. Hydroelectric

Hydropower electricity generation in UK consist of 24% run-of-river and 76% reservoir [36]. The Ecoinvent model for GB run-of-river hydroelectricity model was adopted, and the Ecoinvent "DE" (Germany) model for hydro-reservoirs was used as a proxy, as the database does not contain a corresponding GB model.

#### 2.2.10. Marine Tidal (Future Technology)

Tidal energy is generated through the rise and fall of tides, due to the interaction of gravitational pull of moon and to a lesser extend the sun on the ocean and the rotation of the Earth [50]. There are three types of tidal technologies: lagoon, barrage and stream turbines. National Grid's FES 2019 scenarios assume that the target tidal capacities will be met primarily using tidal lagoons, and secondarily stream turbines. However, there is mounting uncertainty on the future development of tidal technologies, with on the one hand, the recent cancellation of one large tidal lagoon project [51], and on the other hand, new upcoming developments and installed projections on stream turbine [52]. In this study, the assumption was therefore made that the electricity generated by tidal will be harnessed by tidal stream turbines.

There is no model in the Ecoinvent database for this technology. Therefore, all life-cycle inventory and technical information for use in this study was sourced from published scientific literature [53] on an OpenHydro tidal stream turbine. The inventory information includes energy inputs for the installation, manufacturing and maintenance of the system and the material inputs for the construction of the device, power cabling and foundation. The system as described was expected to have a lifetime of 20 years and was rated at 2 MW. The average CF for the stream turbine tidal plant was taken as 5.5% from the DUKES 2019 report [36], and all energy and material inputs were duly scaled to 1 kWh of electricity generated over the life-time of the system.

The resulting foreground inventory information is provided in Appendix A-Table A3.

#### 2.2.11. On-Shore Wind

The Ecoinvent electricity production model for GB 1–3 MW onshore wind turbines was used to represent the total onshore wind electricity generation in UK. The model assumes a 20-year lifetime for all moving components and a 40-year lifetime for all the stationary components of the wind installation [33], and a 26% CF.

#### 2.2.12. Off-Shore Wind

The Ecoinvent electricity production model for GB 1–3 MW offshore wind turbine was used to represent the total offshore wind electricity generation in UK. The model assumes the same lifetimes as for onshore wind turbines, and a 30% CF.

#### 2.2.13. Solar Photovoltaic

National Grid's "2 degree" scenario considers distribution-connected and micro-connected solar capacity and accordingly in this study the assumption was made that most of the solar photovoltaic (PV) generation will come from roof-top mounted systems; additionally, the latest Fraunhofer Institute for Solar Energy report [54] confirms that multi-crystalline silicon (mc-Si) continues to be the leading PV technology by far in terms of global annual production. In order to limit the complexity of the model, a single Ecoinvent process (GB roof-top mounted mc-Si PV) was therefore adopted as the basis for the assessment of solar PV electricity generation in UK. However, since PV systems are still on a continuously and rapidly improving trend, the model was adjusted to reflect the current and expected future mc-Si module efficiencies, respectively reported at 17% in 2019 [54], and estimated at 20% in 2035 and 25% in 2050 [55]. This information was used to adjust the area of solar panels required to produce 1 kWp of installed power in the model.

An average insolation of 1000 kWh/(m2·yr) was then assumed [56], which, combined with a performance ratio (PR) of 80% [57], led to a calculated CF of 9.1%. Finally, the expected lifetime of the PV modules was kept at 30 years until 2035 [57], and then increased to 35 years for 2050 [55].

#### 2.2.14. Pumped Hydro Storage (PHS)

Pumped hydro storage (PHS) uses electricity to pump water into the high-elevation reservoirs during high generation and low demand, and then releasing the water to generate electricity at peak demand. Since PHS for the UK is projected to utilize pre-existing hydro reservoir systems (mainly located overseas and accessed via the interconnectors), which were built for the primary function of generating hydroelectricity, and since the electricity used for pumping the water uphill would otherwise have to be curtailed, the life-cycle impacts associated with PHS were taken to be zero, thus avoiding any double-counting.

#### 2.2.15. Compressed Air Energy Storage (CAES)

Compressed Air Energy Storage (CAES) systems store excess electricity by compressing air to high pressure in underground reservoirs such as pre-existing salt mines. The stored air is then heated and expanded to drive a turbine to generate electricity when the electricity demand is high and the generation is low [32]. Currently there are two CAES systems operating worldwide, one is in Huntorf, Germany since 1969 and the other is in McIntosh, United States since 1991 [58]. Both work by burning natural gas to provide heat for the expansion of air to drive the turbine generator.

However, the FES 2019 "2 degrees" scenario expects CAES systems to be deployed from 2030 onwards, and therefore in this study the assumption was made that by that time the UK's CAES installed capacity will be of the more advanced adiabatic type (A-CAES). This type of CAES works by retaining and storing the heat generated during the compression of the air using a thermal energy storage (TES) system, and then reusing the stored heat for the expansion process instead of burning natural gas. There has been a lot of on-going research on A-CAES over the last decade, including the planned EU-based research Project "ADELE" [59], the Storelectric project planning to build large-scale A-CAES in Holland [60], and a recently completed demonstration project by Hydrostor in Toronto Island, Canada [61].

Life cycle inventory (LCI) information on A-CAES is not available in the Ecoinvent database. The technical data were therefore adopted from available literature; specifically, the maximum number of storage cycles was taken to be 10,000 [62], and the cycle efficiency of the plant was taken to be 67% [32]. It was also assumed that the compressed air will be stored in pre-existing underground caverns, with a cumulative energy storage capacity of 6 GWh deployed by 2035 (cf. Table 1). The information on material and energy inputs for plant construction, compression unit, heat expander and thermal energy storage system was adopted from published literature [63] and rescaled linearly in terms of storage capacity.

The inventory information for the input quantities per kWh of electricity storage capacity is provided in Appendix A-Table A4.

#### 2.2.16. Lithium-Ion Battery Storage (LIB)

Dedicated grid-level lithium-ion battery (LIB) storage was modelled on the basis of the Ecoinvent model for lithium manganese oxide (LMO) technology. LMO is among the most mature options for LIBs, and although it lags behind some of the other cathode formulations in terms of energy density [64,65], this was deemed relatively unimportant for dedicated stationary applications, and counterbalanced by its comparatively long cycle life, its overall stability, and its reliance on abundant and eco-friendly materials [64]. The maximum number of charge-discharge cycles was assumed to be 7000 [66].

#### 2.2.17. Vehicle-to-Grid Storage (V2G)

For each year of analysis, vehicle-to-grid (V2G) storage schemes rely on the Li-ion batteries already installed in the existing electric vehicle (EV) fleet, when connected to the network of charging points, to provide short-duration storage (i.e., frequency response services) for grid support. In order to avoid incurring in double-counting, energy storage made available through V2G is therefore regarded to have zero impacts assigned to it, since the primary function of vehicle batteries is to provide electricity storage for transportation.

#### **3. Methods**

#### *3.1. Life Cycle Assessment (LCA)*

From a methodological perspective, the work presented here was conducted in strict adherence to the current International Organization for Standardization norms on LCA [44,67].

The functional unit (FU) of this study was set as 1 kWh of electricity delivered by the UK grid as a whole, including energy storage and transmission.

The main data source used for the life cycle inventory (LCI) analysis was the reputable and widely-used Ecoinvent version 3.5 database [33], complemented where appropriate and required by a range of other literature sources as described in detail in Section 2.2.

As concerns life cycle impact assessment (LCIA), the choice was made to focus on a set of key impact categories, individually discussed in Sections 3.1.1–3.1.5, and all evaluated at "mid-point" using the widely-adopted and well-regarded CML method [68]. Normalization and weighting were not conducted because, while potentially facilitating the interpretation of the results by a less technical audience, they inevitably remain the most arbitrary steps in any LCA, and the choice of the weighting factors is to a large extent political, with very little if any scientific relevance. In fact, because of this, according to ISO, normalization and weighting are always optional steps and are discouraged for any "comparative assertion intended to be disclosed to the public" [44].

Finally, from a practical point, the whole analysis was carried out using the latest release of the dedicated LCA software package GaBi [69].

#### 3.1.1. Global Warming Potential (GWP)

Global Warming Potential is calculated applying IPCC-derived characterization factors to gaseous emissions, on the basis of their respective equivalent warming potential relative to carbon dioxide (CO2), with a time horizon of 100 years.

Two alternative accounting rules are possible with regard to biogenic carbon emissions, i.e., those that arise from the combustion of biomass (wood chips and pellets, and biogas derived from the anaerobic degradation of organic matter). The argument for excluding them from the calculation of the grid's GWP is that the same amount of carbon (on a molar basis) was previously absorbed from the atmosphere during the biomass growth phase (including the trees used for the wood chips and pellets, as well as those used for paper and cardboard and all the agricultural and food crops that are eventually biodegraded to produce biogas), thereby effectively "closing the loop" and resulting in net zero carbon emissions. Of course, this calculation only applies to C emitted as CO2 (i.e., under complete combustion conditions), otherwise each mole of C that is absorbed as CO2 and later emitted as CH4 would result in a net contribution to GWP, as quantified by Equation (1):

Net GWP per mole of biogenic C = MMCH4 (CFCH4-1) [g CO2-eq]

where:

$$\text{MM}\_{\text{CH4}} = \text{molar mass of CH4}\_4 \text{ [g/mol]}$$

$$\text{CF}\_{\text{CH4}} = \text{GWP} \text{ characterized factor for CH4 [g CO}\_2\text{-eq/g CH4}]$$

(1)

Even so, the argument for excluding biogenic C from the accounting only holds fully if it can be proven that the totality of such biomass was in fact grown sustainably (e.g., from well-managed short-rotation forestry that leads to net zero standing biomass change over time). In reality, this condition can rarely be expected to be met completely. Specifically, wood chips coming from domestic forestry residues may often be closer to being net zero C than wood pellets imported from overseas and paper and cardboard coming from a wide range of international sources. As a result, the real-world net carbon emissions of biogenic feedstocks will always be higher than zero, with considerable uncertainty on the exact figures, depending on often hard-to-ascertain factors such as feedstock origin and supply chain practices.

To reflect such uncertainty, in this work the choice was made to calculate and report GWP under both assumptions, i.e., respectively including and excluding biogenic carbon emissions, and to discuss the results in the light of the considerations made above.

#### 3.1.2. Acidification Potential (GWP)

Acidification Potential is calculated applying stoichiometry-derived characterization factors to airborne emissions, on the basis of their respective acidification potential relative to sulphur dioxide (SO2).

#### 3.1.3. Human Toxicity Potential (HTP)

Human Toxicity Potential is calculated applying characterization factors to all emissions (to air, water and soil), on the basis of their respective toxicity potential relative to 1,4-dichlorobenzene (1,4-DB).

It is noteworthy that HTP results are inevitably affected by a larger margin of uncertainty than those for all other impact categories, due to the intrinsic methodological difficulty of comparing and combining the individual toxicity potentials of a wide and diverse range of organic and inorganic emissions into a single indicator. The uncertainty is especially large in the case of metal emissions [70].

#### 3.1.4. Abiotic Depletion Potential (ADP)

Abiotic Depletion Potential is calculated applying characterization factors to all non-living LCI inputs from the geosphere, expressing their respective scarcity relative to the element antimony (Sb), based on estimates of ultimate reserves and current extraction rates. In order to avoid partially duplicating the information provided by the nr-CED indicator (cf. Section 3.1.5), and potentially obfuscating the impact arising from the depletion of non-energy resources, this indicator is calculated excluding the contributions of all energy inputs (such as fossil fuels and uranium).

It should be noted that abiotic depletion is an impact category that is still frequently the object of methodological discussion, and alternative approaches exist to the quantification of the associated impact, often reflecting differences in problem definition [71–73]. Additionally, ADP's specific dependence on the estimate of ultimate reserves makes it susceptible to obsolescence, especially when using this indicator to assess a depletion-related impact taking place several decades into the future [74].

#### 3.1.5. Non-Renewable Cumulative Energy Demand (nr-CED)

Non-renewable Cumulative Energy Demand is calculated by applying characterization factors to all LCI inputs, based on the total non-renewable primary energy directly and indirectly harvested from the environment for their provision, and expressed in terms of Joules of crude oil equivalent [75].

#### *3.2. Net Energy Analysis (NEA)*

Net Energy Analysis (NEA) provides a different and complementary viewpoint to LCA [23,26,28,29,76], and it is carried out here using the same underlying inventory analysis, thereby providing an internally coherent platform for the calculation and discussion of the results. Specifically, while LCA draws the boundary of the system under analysis so as to account for all natural resources used as inputs (and all emissions to the environment as outputs), NEA is only concerned with those energy inputs that are already available as energy carriers within the technosphere, and which are deliberately "invested" in the system (Inv) for the purpose of harvesting more primary energy (PE) from nature and delivering (the term "returning" is also often used) a usable energy carrier (EC) to society. This is illustrated in Figure 5 with a simplified diagram of a generalized energy supply chain.

**Figure 5.** Simplified diagram of a generalized energy supply chain, illustrating the different boundaries set by life cycle assessment (LCA) and net energy analysis (NEA). PE = Primary Energy; Inv = energy "investment"; EC = Energy Carrier (the energy "return"); S = energy dissipated to the environment.

For the purpose of consistency, all energy "investments" (Inv) to the energy system are accounted for in terms of the total (i.e., renewable plus non-renewable) primary energy directly and indirectly harvested from the environment for their provision (expressed in MJ of oil equivalents). This corresponds to including in the analysis the whole supply chain for "Inv", like in LCA.

The energy "returned" by the same system in the form of a usable energy carrier may then alternatively be accounted for either on the basis of the actual energy in the carrier (EC), or on the basis of its equivalent primary energy (ECPE-eq, measured in MJ of oil equivalents).

When EC is electricity, its equivalent primary energy may be calculated according to an LCA substitution logic, i.e., by calculating how much primary energy is directly and indirectly harvested (in total) from the environment to produce one unit of electricity using the current grid mix. Expressing the electricity "return" in terms of equivalent primary energy thus makes the NEA of any of the individual electricity technologies (e.g., natural gas combined cycles, wind, PV, etc.) inextricably linked to the specific grid mix into which it is embedded. However, doing so also has the following advantages:

(i) It enables the calculation of an "energy return ratio", often referred to in literature as Energy Return on Investment (EROI) [25,77], which features consistent units of primary energy equivalents at both the numerator and denominator. This simplifies the interpretation, since otherwise, "if the numerator and denominator are not measured by the same rule, one loses the intuitively appealing interpretation that EROI > 1 is the absolute minimum requirement a resource must meet in order to constitute a net energy source" [78]. In order to clarify when EROI is calculated using units of equivalent primary energy at the numerator, the notation EROIPE-eq is used in this work—cf. Equation (2).

$$\text{EROL}\_{\text{PE-eq}} = \text{EC}\_{\text{PE-eq}} / \text{Inv} \tag{2}$$

(ii) It enables the methodologically consistent calculation of the Net-to-Gross (NTG) energy return ratio, defined as per Equation (3). NTG thus provides a clear and easy-to-interpret indication of how much of the "gross" energy delivered to society by the UK grid is a "net" energy output that remains available for all societal uses, vs. how much needs to be re-invested to keep the grid itself operational.

$$\text{NTG} = (\text{EC}\_{\text{PE-eq}} - \text{Inv}) / \text{EC}\_{\text{PE-eq}} \tag{3}$$

By combining Equations (2) and (3) and simply re-arranging the terms, NTG may also be expressed as in Equation (4), which leads to what has often been referred to as the "Net energy cliff" [28,79–81] (Figure 6).

$$\text{NTG} = 1 - 1/\text{EROI}\_{\text{PE-eq}} \tag{4}$$

**Figure 6.** "Net energy cliff" diagram displaying the relation between Energy Return on Investment (EROIPE-eq) and Net-to-Gross ratio (NTG).

The interpretation of the "Net energy cliff" is that once the EROIPE-eq of an energy system starts dropping below approximately 10, the share of its gross energy output that remains available for other societal uses (after subtracting the primary energy required to sustain the operation of the energy system itself) starts being reduced significantly for each additional reduction in EROIPE-eq; in other words, for EROIPE-eq < 10, NTG quickly "falls off a cliff". Conversely, for all values of EROIPE-eq > 10, NTG remains safely > 0.9, which means that beyond such "threshold" there is less and less significant competitive advantage to a system characterized by increasingly larger EROIPE-eq.

#### **4. Results and Discussion**

#### *4.1. Life Cycle Impact Assessment Results*

#### 4.1.1. Global Warming Potential

Figures 7 and 8 both illustrate the expected life-cycle GHG emissions associated to the future UK grid mix in 2035 and 2050, expressed per unit of electricity delivered, and relative to the emissions in 2019. The difference between the two figures is that the former includes the contribution of biogenic carbon, while the latter excludes it.

**Figure 7.** Global warming potential (including biogenic carbon) per unit of electricity delivered by the UK grid mix (including energy storage and transmission), expressed as relative to the 2019 value (100% = 269 g CO2-eq/kWh).

**Figure 8.** Global warming potential (excluding biogenic carbon) per unit of electricity delivered by the UK grid mix (including energy storage and transmission), expressed as relative to the 2019 value (100% = 170 g CO2-eq/kWh).

As expected, the two technologies that are most significantly affected by the different assumptions on how to account for biogenic C are biomass- and biogas-fired electricity generation. Additionally, it is noteworthy that even when excluding biogenic C emissions (Figure 8), the life-cycle GHG emissions caused by biogas-fired electricity are expected to still amount to 9% of the grid mix total in 2035, despite such technology delivering only 5.7% of the total electricity generation (cf. Figure 2). This points to biogas being almost twice as carbon intensive as the grid mix as a whole at that point in time, and definitely nowhere near as deserving of the "zero carbon" designation as the leading renewable technology for the UK, i.e., wind, which is responsible for just 5% of total GHG emissions while

generating 58% of the total electricity. On the one hand, this might lead to question of whether retaining, and if fact increasing, biogas-fired electricity generation in the future grid mix even makes sense at all. On the other hand, though, one must also consider what the alternative would be, i.e., what would happen to the biogas that is produced by anaerobic degradation of organic matter in landfills and in sewage sludge if it were not used as a feedstock for electricity generation. Clearly, releasing it directly to the atmosphere would not be an option, as biogas is rich in methane and would cause even more global warming. Therefore, the focus really shifts from the energy sector to the waste management sector, and points to a need to reduce the reliance on all landfills (municipal, industrial and agricultural) to the maximum extent possible, and instead incentivize the reuse, recycling and thermovalorization of waste flows (as applicable, and broadly in that order of merit), as well as composting (the latter providing the added benefit of returning valuable nutrients to the soil).

Unsurprisingly, the planned decommissioning of coal- and gas-fired power plants is shown to have the largest effect on the decarbonization of the grid mix. It is also noteworthy that the contribution of energy storage technologies (A-CAES and LIB) to the total carbon budget is absolutely negligible, even out to 2050 when 10% of the total yearly VRE generation is assumed to be routed into storage. This is a reassuring result that dispels any potential concerns about the negative effect that the requirement for energy storage might have on the planned decarbonization of the grid.

Considering the total grid mix results as a whole, from 2019 to 2035, GHG emissions may be expected to drop by 60% (i.e., from 263 to 106 g CO2-eq/kWh, when including biogenic C emissions), or even as much as 73% (i.e., from 170 to 46 g CO2-eq/kWh, when excluding biogenic C emissions). Further, but less significant, emission reductions are then expected when extending the analysis to 2050. As discussed in Section 3.1, the real total net GHG emissions lie somewhere in between those reported respectively in Figures 7 and 8, but are likely to be somewhat closer to the latter. Be that as it may, it is worth noting that in neither case does the reduction in the total life-cycle grid mix emissions approach 99%, as one might naively assume when considering the share of electricity output that is generated by technologies that are considered "zero carbon at point of use" (cf. blue line in Figure 4). This should not be interpreted as a damning result per se, but rather as a stark reminder of the importance of always duly including all life cycle stages in the analysis.

#### 4.1.2. Acidification Potential

When looking at the results for acidic emissions (Figure 9), two things are readily apparent: (i) the total reductions in potential impact that may be expected for 2035 and even 2050 are much less significant than in the case of global warming potential (respectively, only −8% and −28% with respect to 2019); and (ii) biogas-fired electricity is by far the worst offender in the mix (>80% of total acidic emissions). The latter result is striking, but it is broadly confirmed by other independent studies [82,83], and it may be understood if one considers that the anaerobic degradation of biomass produces significant levels of ammonia and hydrogen sulfide alongside methane, and that those two gases are readily oxidized to, respectively, NO2 (then hydrated to nitric acid) and SO2 (then hydrated to sulphuric acid).

While these AP results may look somewhat discouraging, it should also be recognized that there is potentially ample scope for improvement if the biogas is "upgraded" to biomethane by subjecting it to water scrubbing and membrane separation prior to feeding it to the thermal power plant, which would essentially rid it of most ammonia and hydrogen sulfide and render it almost indistinguishable from natural gas [84]. While not yet common practice in the UK, such pre-treatment of the biogas feedstock may become more widespread in the future and contribute to curbing the acidic emissions of biogas-fired electricity, and hence of the whole grid mix, by 2035 or 2050.

**Figure 9.** Acidification potential per unit of electricity delivered by the UK grid mix (including energy storage and transmission), expressed as relative to the 2019 value (100% = 2.1 g SO2-eq/kWh).

#### 4.1.3. Human Toxicity Potential

When moving to consider the life-cycle environmental impacts of the grid mix in terms of human toxicity (Figure 10), a reversal of the trend is observed for the first time, whereby the total impact increases going from 2019 to 2035 (+10%), only to then decrease again slightly in 2050 (+8% relative to 2019). Because of the unavoidably large uncertainty associated with the quantification of HTP (cf. Section 3.1.3), such relatively small changes are probably at the limit of what ought to be considered statistically significant, and another, possibly more scientifically valid way of looking at the results is that the overall human toxicity potential of the grid mix is expected to remain broadly stable for the next three decades.

**Figure 10.** Human toxicity potential per unit of electricity delivered by the UK grid mix (including energy storage and transmission), expressed as relative to the 2019 value (100% = 83 g 1,4-DB-eq/kWh).

What is interesting, nonetheless, is that the relative contributions of the various technologies to the total impact are significantly different than for GWP or AP, and for the first time even those technologies that are conventionally regarded as the "greenest" (i.e., wind and solar PV) end up being responsible for sizeable shares of the total impact. These results are due to a combination of these technologies' comparatively large demand for heavy metals (mainly Cu, Al and Ni) per functional unit (also corroborated by previous independent studies [19,85]), and the toxic emissions associated to the respective metal supply chains (mainly at the mining and beneficiation stages) [33,86]. For the same reasons, electricity transmission lines and LIBs are also non-negligible contributors to this impact category. The HTP of nuclear electricity is likewise significant in the mix, almost entirely due to the emissions arising from the uranium supply chain.

#### 4.1.4. Abiotic Depletion Potential (Elements)

The abiotic resource depletion results illustrated in Figure 11 ("elements", i.e., excluding energy resources such as fossil fuels and uranium) are even more striking, in that they point to a significant increase of the total grid mix impact: initially +76% in 2035 and then down to +55% in 2050 (both values relative to 2019). Even more so than for HTP, these results are mainly driven by wind and PV's increased demand for metals (mainly Cu) per unit of electricity delivered, and LIB and transmission lines once again play a non-negligible role.

**Figure 11.** Abiotic depletion potential (elements) per unit of electricity delivered by the UK grid mix (including energy storage and transmission), expressed as relative to the 2019 value (100% <sup>=</sup> 3.6·10−<sup>4</sup> <sup>g</sup> Sb-eq/kWh).

#### 4.1.5. Non-Renewable Cumulative Energy Demand

The analysis of the grid's overall demand for non-renewable primary energy (Figure 12) is straightforward to interpret, with those thermal technologies relying on non-renewable energy feedstocks (i.e., coal, natural gas and uranium) being responsible for the vast majority of the impact. The continued reliance on nuclear power as a significant contributor to the future mix (15.8% in 2035 and up then to 19.5% in 2050—cf. Figures 2 and 3) also poses a major constraint on the achievable improvement at grid level (only −24% in 2035 and −40% in 2050), which contrasts with the more significant gains in terms of decarbonization (cf. Figures 7 and 8).

**Figure 12.** Non-renewable Cumulative Energy Demand (nr-CED) per unit of electricity delivered by the UK grid mix (including energy storage and transmission), expressed as relative to the 2019 value (100% = 6.0 MJ oil-eq/kWh).

#### *4.2. Net Energy Analysis Results*

Figure 13 illustrates the total primary energy that must be invested to operate the various energy supply chains that "feed" the UK grid, per unit of final electricity delivered by the grid itself. It is important to recall that, as explained in Section 3.2, these figures exclude the primary energy resources that are directly harvested from nature and converted into electricity (e.g., the coal itself that is extracted, pulverized and delivered to the coal-fired power plants; the woody biomass is harvested, compressed into pellets and delivered to the biomass-fired power plants; the solar energy is captured by the PV panels; and so on and so forth). What these results show is that, over time, the UK grid mix will require less and less commercial energy to be diverted from other societal uses and (re)invested to support its operation, per unit of electricity delivered.

In terms of the break-down of the total energy investment by technology, the results for the three years considered in the analysis broadly reflect the changing grid mix composition, with a few notable observations to be made, as follows.

(i) The energy investment for biomass-fired electricity is disproportionally large (over 30% of the total in 2019, vs. a share of only 10% of electricity generated). The planned reduced reliance on biomass as an energy resource for electricity generation in the future (3% of the grid mix in 2035 and 2% in 2050—cf. Figures 2 and 3) therefore seems justified from a net energy perspective.

(ii) At the opposite end of the scale, the expected future energy investment for wind electricity (approximately 30% of the total in 2035 and 2050) appears to be well justified in view of the associated energy returns (the share of total electricity generated by wind approaches 60% in the same years—cf. Figures 2 and 3).

(iii) The energy investment required for energy storage technologies (A-CAES and LIB) remains very small, even when such technologies are relied upon to store 10% of the VRE generated. This is a reassuring result that indicates that, from an energy balance perspective, there likely will not be any disruptive consequences when moving to a future grid mix which will rely heavily on intermittent renewable generators and will therefore require energy storage.

**Figure 13.** Total energy investment per unit of electricity delivered by the UK grid mix (results include the effects of energy storage and transmission), expressed as relative to the 2019 value (100% = 0.65 MJ oil-eq/kWh).

Figure 14 then illustrates the same overall results, but in terms of the Net-to-Gross primary energy return ratio. NTG is shown to creep steadily upwards on the "net energy cliff", from 2019 to 2050, once again indicating that the planned energy transition according to National Grid's "2 degrees" scenario is sound from a net energy perspective. This is a significant result, which should allay some of the often-voiced concerns about potentially reduced future availability of net energy.

**Figure 14.** Evolution of the Net-to-Gross (NTG) primary energy return ratio of the UK grid mix (results include the effects of energy storage and transmission).

Having said all this, it is also important to underline that the present analysis was carried out using an "integrative" approach [87], which is characteristic of LCA and which "discounts" (i.e., virtually spreads out) all inputs and emissions associated to each individual energy system

over its respective full life cycle. In other words, each individual energy supply chain is treated as a "black box" from a temporal perspective, irrespective of exactly when, during its life cycle, a particular (energy) input is supplied or a specific emission occurs. Instead, in reality, most of the material and energy inputs required for renewable technologies such as wind and PV take place up front during their manufacturing phase, while the associated energy "returns" are reaped over the course of approximately three decades (the typical lifetime of these systems). When these technologies are deployed quickly and on a massive scale, this type of temporal mismatch could, in some instances, result in a temporary reduction in the initial year-by-year availability of net energy, as discussed elsewhere in literature [88]. While this fact needs to be duly taken into account in terms of energy policy planning, it is fair to argue that, in light of the overall positive longer-term NTG trend identified here, it should not be seen as reason to dismiss the fundamental soundness of the whole long-term strategy, but instead as a further call for the application of a "sower's strategy" (whereby today's energy "seeds" are planted to reap the energy "crops" of tomorrow) [89].

#### **5. Conclusions**

This thorough life-cycle analysis of National Grid's "2 degrees" future energy scenario has produced a wide range of quantitative results which, when analyzed together, allow for a comprehensive and balanced appraisal of its strengths and weaknesses. To summarize, the following considerations may be made, which capture the key findings and provide policy guidance.

Firstly, these life-cycle carbon budget calculations have confirmed that the "2 degrees" strategy would be very effective at decarbonizing UK electricity, albeit not to the radical extent that might have been inferred by just taking at face value the share of electricity that is projected to be generated by "zero carbon" technologies (nuclear, wind, tidal, hydro, biogas, biomass and gas with CCS).

Secondly, the analysis has shown that biogas and, to a lesser extent, biomass are not especially benign energy resources for electricity production, despite their intuitively reassuring "bio" designation. Both are negatively affected by low energy return on investment, and biomass-fired electricity is responsible for very large acidic emissions. However, while reducing the reliance on biomass would be relatively easy to achieve by just curbing imports of wood pellets from North America, biogas-fired electricity may prove difficult to phase out, since biogas is produced in landfills and agricultural sludge, and using it to produce electricity may be the lesser evil in terms of global warming. Cleaning up biogas electricity generation by either upgrading the biogas feedstock to biomethane or applying drastic scrubbing at the power plant stack output would however still be recommendable.

Thirdly, the results have indicated that moving to the large-scale penetration of variable renewable energy (VRE) generation entails some trade-offs in terms of the depletion of metal resources and potential associated life-cycle toxicity impacts. This is partly due to wind and PV's demand for technology-specific metals and semi-metals, but also in large part simply determined by their less spatially concentrated nature, which calls for more copper transmission lines per unit of electricity delivered.

Fourthly, the planned continued (and even increased, after 2030) reliance on nuclear generators puts a hard cap on the achievable gains in terms of decreased dependence on non-renewable (and imported) energy resources. While the ultimate availability of nuclear ore reserves may not pose a major issue for a long time still, this fact does have immediate negative consequences in terms of the UK's energy sovereignty (since no such ores are available domestically).

Importantly, the deployment of energy storage technologies has shown not to cause any major set-backs in terms of any of the impact categories (with the possible partial exception of human toxicity and abiotic depletion). This is a welcome and reassuring result in and of itself, and it may also be interpreted to indicate that perhaps an even more aggressive roll-out of wind and PV might be attempted (the intermittency of which could be mitigated with even more energy storage), which would lead to a reduced demand for new small modular nuclear reactors.

However, a margin of uncertainty remains on the possible residual requirement for some curtailment of VRE, despite the substantial energy storage capacity that is assumed to be deployed in the "2 degrees" scenario. Such uncertainty could potentially be reduced by employing a high-temporal-resolution modelling approach whereby the hourly electricity supply and demand profiles are extrapolated on the basis of historical datasets.

Additionally, further research is already under way, whereby the interlinkages between the electricity and transport sectors will be explicitly modelled dynamically, with specific focus on the twin roles played by lithium-ion batteries (LIBs). On the one hand, LIBs are expected to be used as on-board storage in electric vehicles (EVs), which may be privately-owned or used for transport-as-a-service (TaaS), and in both cases potentially provide vehicle-to-grid (V2G) storage capacity. On the other hand, LIBs will also be used for dedicated stationary grid-level storage, in which case they may be produced using a combination of virgin and recycled materials, or even re-purposed after their first use in EVs (i.e., a second life application). The ensuing coupled mass-flow model will enable a more detailed and robust assessment of the energy and environmental impacts of energy storage on the future electricity grid mix, by improving on the current blanket assumptions about the future deployment of LIB and V2G storage.

Finally, this analysis has produced significant, and to some extent, possibly even ground-breaking results by dispelling the often-voiced concerns that a massive transition to renewable energy technologies must necessarily entail a worrisome reduction in the net energy that is made available to society. In fact, it was shown that the evolution of the UK grid under "2 degree" scenario conditions would even result in a modest increase of its net-to-gross energy return ratio.

All in all, this study's results should be taken as a sobering reminder that there is more to "greening the grid" than meets the eye, and that forecasting the complex effects of any energy policy strategy calls for a holistic life-cycle approach.

**Author Contributions:** Conceptualization, M.R.; methodology, M.R.; software, M.R. and M.K.; validation, M.R., M.K. and A.H.; formal analysis, M.R.; investigation, M.R. and M.K.; resources, M.R. and M.K.; data curation, M.K.; writing—original draft preparation, M.R. and M.K.; writing—review and editing, M.R. and A.H.; visualization, M.R.; supervision, M.R. and A.H.; project administration, A.H.; funding acquisition, A.H. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the Faraday Institution [grant number FIR005].

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Appendix A**

This appendix contains foreground life-cycle inventory tables for those electricity generation and storage technologies for which no pre-existing data were available in the Ecoinvent database, and which had to be modelled from scratch on the basis of literature information.

**Table A1.** Foreground inventory of life-cycle inputs for carbon capture and storage (CCS) technology, complementing natural gas combined cycle (NGCC) power plants. All values are per kWh of electricity generated.


<sup>1</sup> Accounted for by deduction from plant output.


**Table A2.** Foreground inventory of use-phase emissions per kWh of electricity generated by the NGCC + CCS system.

**Table A3.** Foreground inventory of life-cycle inputs for stream turbine tidal electricity generation. All values are per kWh of electricity generated.


**Table A4.** Foreground inventory of life-cycle inputs for adiabatic compressed air energy storage (A-CAES) technology, including plant construction, compressors, thermal energy storage (TES) and heat expanders. All values are per MWh of electricity storage capacity.


#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
