*4.1. PV Panels LCA*

To operate with LCA results performed for different system boundaries and functional units, we used data only from the sources where the results were given in physical units (not in dimensionless points). For this reason, the analysis did not include the results of studies where environmental impact assessments were presented in normalized values (e.g., [16,68–71]). Additionally, because recycling technologies are rapidly progressing, we selected for comparison only the results of the decade. Table 2 presents the results of a comparative analysis of the photovoltaic plant's life cycle.

The results of LCA in [41] were calculated according to the ReCiPe endpoint method with GaBi software. The functional unit was 1 kg of silicon-based PV waste modules. The study considered two types of PV modules (multi- and monocrystalline silicon modules) and several end-of-life treatment scenarios: landfill, incineration, and thermal, chemical, and mechanical recycling. In the case of recycling, the system boundaries included manufacturing, installation, operation, recycling, and reuse of recycled materials in a new cycle of manufacturing. One of the notable features of the study was a separate analysis of the transportation phase. In this case, the impacts do not depend on the recycling technology but the waste collection system and distance from it.

The environmental effects of transporting used modules to the place of recycling or disposal were taken into account in two ways: (1) assuming a transportation distance of 50 km; (2) assuming a transportation distance of 100 km (for the case of recycling only). Modeling results showed that transportation for 50 km increases impacts on human health up to 2.1–2.3 <sup>×</sup> 10−<sup>4</sup> DALY in case of landfill or incineration, and up to 1–1.2 <sup>×</sup> 10−<sup>4</sup> in case of recycling. Transportation for 100 km for recycling increases the impact on human health up to 1.6–1.7 <sup>×</sup> 10−<sup>4</sup> DALY, which is comparable to the landfill or incineration indicators on site (without transportation). Similar results were obtained in two other categories of environmental impact. Transportation for 50 km increased ecosystem effects to 5.8–6.2 <sup>×</sup> 10−<sup>11</sup> species·year in case of landfill or incineration and 3–3.5 <sup>×</sup> 10−<sup>11</sup> species·year in case of recycling. Transportation per 100 km for processing increased the adverse effects up to 4.5–5.5 <sup>×</sup> 10−<sup>11</sup> species·year, which offset all the positive effects of recycling compared to landfill or incineration.

Thus, the results of this analysis demonstrate that the recycling plant should be at most 80 km away from a PV plant. Otherwise, landfill and incineration scenarios have lower impacts when considering human health and ecosystems. This is a critical practical conclusion for the eco-design of the energy system regarding the territorial location of PV-plants. Solar plants must be placed compactly in a relatively small radius from the processing plant, or they should be some kind of mobile recycling plant. In addition, it is essential to consider not only the distance but also the method of transportation.

In [67], an LCIA is made for the recycling of c-Si and CdTe PV modules using both the cut-off and end-of-life approaches. Data were collected from several recycling companies in central Europe. The functional unit was 1 kg of used framed c-Si and unframed CdTe PV modules, but the final results were presented for 3 kW modules, mounted on a slanted roof. The life cycle impact assessment competed with the ILCD Midpoint 2011 method, but only six of the most relevant impact categories were taken into consideration.







For the case of c-Si PV modules, the weighted average of multi- and monocrystalline Si PV modules was considered. It was estimated that the used modules would be transported by truck over a total distance of 500 km. The technology of recycling is mechanical. The desirable outputs of the recycling process are the bulk materials of glass cullets, aluminum scraps, and copper scraps. For the case of CdTe PV modules, the average transport distance from the place of installation to the recycling plant was considered as 678 km (data of First Solar's recycling facility located in Germany). The recycling process includes shredding and milling the used CdTe PV modules in the first step, then removing and dissolving the semiconductor film as the second step. The LCIA results for both types of panels are presented in Table 2. As one can see, CdTe PV modules are superior in environmental performance to c-Si PV modules in all categories of environmental impact. This conclusion can also be used in the eco-design of energy systems.

Note that even though both LCAs of [41,67] were performed throughout the entire life cycle, they still do not provide a complete picture of all the adverse effects of a PV plant. They do not take into account the productivity of a PV plant at the operation stage. More efficient PV plants (both by capacity factor and by energy conversion coefficient) can produce more useful products (electricity) for their life cycle and thereby reduce the need for the production and installation of additional PV capacity. Therefore, it is more appropriate to use 1 kWh of generated electricity as a functional unit for the LCAs of all energy facilities, including solar panels. Recalculation of the results of these two studies into other functional units (1 kWh of electricity), unfortunately, is impossible, since the location of the installation of solar panels is a source of uncertainty in this study. In locations with a high level of solar radiation, the total adverse environmental effects over the entire life cycle can be lower due to the greater volume of produced useful products (electricity).

Latunussa and co-authors [72] apply the LCA methodology to a pilot process of recycling of crystalline-silicon (c-Si) PV panels on the Italian "SASIL S.p.A" company. The functional unit was 1000 kg of PV waste panels, including internal cables. The system boundaries of the LCA included "gate-to-gate" recycling processes, starting from the delivery of the waste to the recycling plant and ending with the sorting of the different recyclable material fractions and the disposal of residues. The transportation of PV waste to the recycling plant was considered, while the decommissioning of the PV plant was not. Transportation was considered under the assumption that the distance from the PV plant to the nearest collection point of electronic waste is no more than 100 km, and the distance between the collection point and the recycling site is 400 km.

The novel process of recycling has a sequence of physical (mechanical and thermal) treatments, followed by acid leaching and electrolysis. The amounts of energy produced by the incineration process (for example, the incineration of the sandwich layer and plastics from cables) are considered as coproducts and their positive impact calculated as a credit (avoided environmental impact). By contrast, the environmental credits derived for potentially substituted primary materials are not included.

The modeling and calculation were implemented with SimaPro software version 8.0. The ILCD midpoint method was used for the life cycle impact assessment (Table 2). The "mineral, fossil, and renewable resource depletion" impact category is replaced with "abiotic depletion, fossil" and "cumulative energy demand (CED)" in order to distinguish the contributions of energy sources from those of nonenergy materials. The "land use" and "water resource depletion" impact categories were not taken into consideration due to their high uncertainty.

The results of this study do not have direct applications for the eco-design of the power system. However, they can be used to calculate the entire life cycle of c-Si PV modules for any method of installing panels and for any location.

Ardente [11] extended the results obtained in [72] by introducing options for the recycling process that depend on the material of the back-sheet. This study also introduced several additional impact categories. The results indicated that there is little potential to reduce the environmental impact of the recycling process in some categories due to the variation of materials for the back-sheet and recycling technologies for the auxiliary parts of the PV panel.

Held and Ilg [73] considered both individual stages and the entire life cycle of thin-film CdTe modules. They used two different functional units: 1 m<sup>2</sup> of the module and 1 kWh of energy produced. An interesting feature of the study is the comparison of impacts of the module with and without the balance of systems (BoS). It was revealed that the relative contribution of the BoS on the total impact of the PV power plant is around 35% to 45%. Estimates of environmental impacts for 1 kWh of electricity as a functional unit are made under three different assumptions about the amount of solar radiation in the location of the solar power plant. The disadvantage of the study is the lack of accounting for the effects of transporting new modules to the installation site and used modules to the recycling site. As shown above, transportation can have a significant influence on all environmental impact assessments. In addition, in this paper, the lifetime of the PV modules is assumed to be 30 years, which is an overestimation.

The authors of [74] presented a rather unusual approach to life cycle analysis, which the authors defined as "grave-to-cradle". They suggested that recycled silicon should replace a virgin material in the production of PV panels and considered the process in terms of industrial symbiosis. Recycling is carried out using thermal and chemical methods.

Corcelli and coauthors [75] investigated two c-Si PV panel recycling scenarios: one with a high level of material recovery and another with a low level. Environmental impacts were calculated in two versions: including credits and excluding them. The disadvantage of this study is the lack of accounting for transportation.

Comparing the environmental impacts of recycling with the impacts of all previous stages of the life cycle presented in EcoInvent (Tables 3 and 4), one can conclude that any technology for recycling PV waste is more ecologically friendly than landfilling. However, this is true only if the recycling plant is located in the same region where the panels are manufactured and used. In this case, the transportation of heavy modules over long distances was not required. This is consistent with what has been found in [46,47] for economical and in [76] for environmental parameters of recycling. By comparing the results from Tables 3 and 4, we can conclude that CdTe panels are preferable over silicon panels for the full life cycle (with EoL stage) in most categories of environmental impacts. These results go beyond previous reports [77], showing that current techniques used in the recycling of PVs produce higher impacts in the case of c-Si than in the case of CdTe. A further novel finding is the following: despite the fact that current technologies for recycling of PVs can be significantly improved from an environmental point of view [78], the contribution of these improvements to the negative impacts throughout the life cycle is insignificant. Therefore, it is advisable to focus on further improvement in economic parameters, in particular, on achieving the economic feasibility of recycling small volumes of PV waste.

#### *4.2. LCA of Wind Turbines or Their Components*

Aggregating the LCA results for wind turbines, as in the previous case, we also did not consider the results presented in dimensionless units (e.g., [68,79,80]), or received more than ten years ago (e.g., [81]). For a more detailed understanding of the development possibilities of recycling, we also separately examined the studies analyzing the LCA of the composite materials of blades (Table 5).

Garrett and Rønde [82] performed an LCA for a 50-MW wind park. The functional unit was 1 kWh of electricity produced. GaBi DfX software and primary data from Vestas were used. The analysis included all stages for the manufacturing and transportation of raw materials, turbine and wind plant components, as well as maintenance and end-of-life disposal. The study used data on primary fuel consumption collected by Vestas for truck and sea vessel transportation of turbine components. The transportation distance corresponded to the one that is part of Vestas' supply chain. Turbine recyclability (in percent turbine mass) was estimated to be between 81% and 85% (mainly metals), depending on the class of the turbine. In modeling, an avoided impact approach was used.


**Table 3.** Environmental impacts of di fferent life cycle stages of c-Si PV panels (for 1 kWh 1).

1—The conversion to the functional unit of 1 kWh was carried out on the assumption that the capacity factor is 14%, the weight of the 200 Wp module is 16.8 kg, and the weight of 1 m2 of the module is 13.2 kg. 2—P, manufacturing (production) of PV panel with upstream activities; O, operation; R, recycling.



3—Recalculation into a functional unit of 1 kWh was carried out by assuming that the mass of 1 m2 of the module is 14.63 kg (according to First Solar), module e fficiency 10.5%, solarirradiation1200kWh/m2/year.







Bonou [83] used the ILCD method for the assessment of the environmental impact of two types of wind power plants (onshore and offshore) from the extraction of raw materials to the EoL. The primary data were collected from four representative European power plants with state-of-art technology provided by Siemens Wind Power. The functional unit was 1 kWh. The EoL of the power plants consisted of the management of construction and demolition wastes. The recycling of turbine blades and foundation was included. The recycling process for composite blades was mechanical shredding and incineration in cement production. The recycling process for the cement foundation was crushing with the positive output of crushed gravel. The results in physical units were obtained only for the impact category "climate change", and they indicated the preference of land-based wind parks.

Poujol [84] studied a floating 24-MW offshore wind farm's life cycle from "cradle-to-grave". The multicriteria approach was used for LCIA. It is based on the combined performance of ILCD, CED, and ReCiPe 2016 MidPoint method (H). An additional parameter of "water use" was estimated according to the AWARE method [85]. The functional unit was 1 kWh of electricity. The recycling technology was not specified. Presumably, this is the usual recycling of metals and landfill for composite blades. The contribution of the EoL stage to all the adverse effects was quite large due to the need to use diesel-powered marine vehicles for decommissioning a wind farm located 16 km from the coast.

In a report of Vestas [86], the environmental impacts associated with the production of electricity from a 50-MW onshore wind plant were studied using a cradle-to-grave LCA. The functional unit was 1 kWh of electricity. At the EoL stage, it was assumed that metals are recycled (85–87% of the total turbine mass), and composite blades are incinerated by 50% and landfilled by another 50%. Blades and foundation treatment did not bring avoided environmental impacts. An important distinguishing feature of this work is the high certainty of data on all processes. It included the process of transporting equipment to the installation site and the recycling site and the process of connecting to the grid. The estimated transportation of the turbine components ranged from 50 km for the base to 2200 km for the blades; the transportation distance to the recycling site was assumed to be 200 km.

Al-Behadili and El-Osta [87] calculated the emissions of a 1.65-MW wind turbine over its entire life cycle (including EoL). The functional unit was 1 kWh of electricity. An important feature was a fact that emissions related only to energy consumption were taken into account. The initial data were obtained from literature and taken in an averaged form. EoL treatment technologies were not specified, but from the context of the paper, one can conclude that it was recycling of the metal parts of the turbine and the landfilling of the blades. Transportation distance was not specified.

Chipindula [88] performed a classic LCA for several types of turbines: medium ground-level power (1, 2, and 2.5 MW), powerful offshore in shallow water (2 and 2.5 MW), and offshore in deep waters (2.5 and 5 MW). All intermediate and final calculated data were averaged over the class of turbines. The functional unit was 1 kWh of electricity. EoL treatment involved the recycling of metal components of the turbine in the range from 55% (for aluminum) to 90% (for iron, steel, and copper) and 100% disposal of composite parts and a concrete base. Transportation of turbine components to the installation site was taken into account, and the maximum distance was assumed to be 10,000 km. Transportation to the place of recycling was not taken into account.

The results of the study show that although offshore wind parks have a higher capacity factor (45–47% compared to 35% for onshore wind turbines), in most categories, they produce more significant adverse impacts. This is precisely due to the contribution of the EoL stage, in which concrete foundations are left in the ground.

Alsaleh and Sattler [57] performed a life cycle inventory and assessment of the various phases of the 2-MW Gamesa onshore wind turbine life cycle in terms of TRACI impact categories with SimaPro Software version 8.3.2. The study considered RES recycling (OLR) with 98%, 90%, and 50% recycling rates for metals, plastic, and electronic components, respectively. Fiberglass (blade material) and lubricants were considered 100% landfilled. The most notable result of the study is the conclusion regarding the environmental friendliness of transporting new turbines to the installation site: the impact of truck transport for the distance 656 km is, in most cases, comparable or even greater than that of transoceanic ship transport, which was 8325 km. The effect of transporting the used parts of the turbines to the place of recycling or landfill was not taken into account.

Guezuraga [89] used GEMIS software for modeling the entire life cycle of two types of wind turbines: a 2-MW turbine with a gearbox and a 1.8-MW turbine without a gearbox. The paper assumed that the 2-MW turbine was transported for a distance of 2700 km, and the 1.8-MW turbine was transported for a distance of 1100 km by truck. The transportation distance to the place of recycling or landfill was not indicated. Only energy-related impact categories were considered. Recycling of stainless steel, cast iron, and copper was considered, whereas epoxy, plastic, and fiberglass were incinerated. The concrete foundation was 100% landfilled. Comparing the results for the two turbines, the authors concluded that the impact of the 1.8-MW turbine without gearbox was a little less.

The authors of [90] compared energy demand for the entire life cycle of 2-MW onshore wind turbines with a tall (76.16 m) tower. The study compared the ecological impact of traditional steel and innovative lattice towers. A lattice tower needs around 35% less steel, and it has an almost 33% lighter foundation, which gives a significant advantage for transportation and construction. GEMIS software was used for modeling total energy demand. The transportation for 240 km by truck and 1020 km by ship to the location of installation was considered. EoL treatment included recycling of metals with 5–10% loss, the 100% incineration of epoxy, fiberglass, and plastic, and 100% landfill of the concrete foundation. The distance to the place of processing, incineration, or landfill was not indicated. The study concluded that the turbine with a lattice tower was 32% less impactful on the environment in terms of CO2 emissions.

Of the studies modeling LCAs of composite materials, only a few can be distinguished. For example, in [91], recycling by pyrolysis of carbon fiber reinforced polymers (CFRPs) was investigated with an LCA according to the ReCIPe Midpoint (H) 1.10 method. The functional unit was 1 kg of CFRP waste. It was estimated that recycling can avoid −0.08 kg CO2 eq. due to the recovery of methanol and ethyl acetate.

In [92], a limited version of LCA was carried out for several possible options for handling CFRP waste. Only CO2 emissions were taken into account. The models in this study were based on hypothetical CFRP treatment routes because exact facility locations were not identified. "Gate-to-grave" models have been developed for CFRP waste treatment by landfilling and incineration, beginning at the point of waste collection and including waste processing (disassembly, shredding), transport, and waste treatment (landfill, incineration). For recycling, a "gate-to-gate" approach is taken, which includes the production of composite materials from recycled carbon fibers (r-CF) and the use and/or disposal of other recyclate materials. Maintenance and facility construction is also not included in the LCI boundaries.

The results demonstrated that landfilling produced minor GHG emissions (24 kg CO2eq./t CFRP) due to the inert nature of CFRP waste; incineration resulted in the greatest net GHG emissions (2011 kg CO2eq./t CFRP) from the combustion process. Energy outputs from incineration are assumed to displace the electricity and natural gas-fired heat generation, which gave a credit (−1041 kg CO2eq./t CFRP). Mechanical recycling with landfilling of the coarse recyclate fraction and displacement of GF production resulted in a net global warming potential reduction of 378 kg CO2eq./t CFRP.

The results of a comparative analysis of environmental assessments of the life cycle of wind turbines are summarized in Table 6. The following conclusion can be drawn: increasing the capacity of wind turbines by increasing its height and the span of the blades from an environmental point of view is unreasonable since transporting bulky components of the turbine to the installation and recycling site or landfill contributes too much to the overall ecology footprint. For the same reason, it is inappropriate to build offshore wind farms far in deep waters. Contrary to the findings of [93], we did not find solid evidence that the environmental impacts of onshore wind farms are higher. This situation can only change if environmentally friendly modes of transport (for example, electric trains) are used. Overall, these findings are in accordance with findings reported by Wang et al. [93]. The technologies for recycling the blades need further development. So far, it can be stated with a degree of certainty that incinerating the blades of wind turbines is an environmentally unacceptable alternative to their

disposal. The development of technologies to reduce the weight of the tower for a wind turbine, for example, by improving its design, has significant prospects from the environmental point of view.


**Table 6.** Environmental impacts of different life cycle stages of wind plants (for 1 kWh).

4—Minimum and maximum values are given in single units of measurement from sources [57,82–84,86–89] of Table 5. Data from [88] were not taken into account since they are 3–4 orders of magnitude higher than data from other sources, which most likely indicates significant differences in the choice of primary data. The exception was the data on the land occupation category since they were not found in other sources, and it was impossible to compare them with any other results. 5—Data from EcoInvent for Vestas 2 MW wind turbine (global).

A major limitation of our study is the uncertainty induced by parameters of technological, spatial, and temporal nature in LCA models of wind and solar plants. For example, recent research [94,95] suggests that there is great variability in results within sets of wind turbines with similar nominal power output. Nevertheless, we can still state several important common principles that can be applied to the eco-design on energy systems based on RES. This is particularly important when investigating new possibilities for the development of renewables in countries where the issues of eco-design have not been properly addressed yet.
