Trends in Scientific Literature on Energy Return Ratio of Renewable Energy Sources for Supporting Policymakers

: The scarcity of fossil fuels and their environmental impact as greenhouse gas (GHG) emissions, have prompted governments around the world to both develop research and foster the use of renewable energy sources (RES), such as biomass, wind, and solar. Therefore, although these efforts represent potential solutions for fossil fuel shortages and GHG emission reduction, some doubts have emerged recently regarding their energy efficiency. Indeed, it is very useful to assess their energy gain, which means quantifying and comparing the amount of energy consumed to produce alternative fuels. In this context, the aim of this paper is to analyze the trend of the academic literature of studies concerning the indices of the energy return ratio (ERR), such as energy return on energy invested (EROEI), considering biomass, wind and solar energy. This could be useful for institutions and to public organizations in order to redefine their political vision for realizing sustainable socio-economic systems in line with the transition from fossil fuels to renewable energies. Results showed that biomass seems to be more expensive and less efficient than the equivalent fossil-based energy, whereas solar photovoltaic (PV) and wind energy have reached mature and advanced levels of technology.


Introduction
Energy plays a fundamental role in systems, as it allows for the operation and, consequently, the transfer of information among its components. In an economic system, it allows for human development. In a society, the availability and quality of various types of energy and material resources drive growth, the settlement of people, and the eventual decline experienced by each civilization (Hall et al. 2014). For 200 years, fossil fuels have represented dominant sources of energy, powering human societies and contributing to prosperity across fields, such as economics, agriculture, transportation, and the sciences. According to British Petroleum (2019), in 2018, the primary world energy consumption was equal to 13.9 Gtoe, and fossil fuels, such as oil, natural gas, and coal, contributed 84.7% of the total. This was followed by hydroelectricity (6.8%), nuclear energy (4.4%), and other renewables such as wind, geothermal, solar, biomass, and waste (4.1%).
Although renewable energy sources (RES) satisfy a small percentage of global primary energy demand, their importance has been increasing in most countries due to unquestionable environmental and social benefits (lower greenhouse gas (GHG) emissions that result in mitigating climate change) and the possibility to be managed at local and participative level (Lyu and Shi 2018;Capellán-Pérez et al. 2019). Numerous strategies have been adopted to mitigate GHG emissions with an administrative science approach in the form of green information technology and green information systems (Imasiku et al. 2019). Moreover, governments are addressing the effects of climate change and managing the associated transition to a larger share of renewable energy-based generation (Chapman et al. 2016).
Therefore, the transition from fossil fuels to renewable energies is a crucial condition for realizing sustainable socio-economic systems even if, currently, the performance of RES technologies can be, in some cases, worse than those of fossil fuels (Murphy and Hall 2010;Capellán-Pérez et al. 2019). Actually, scientists and economists have already proved that, in energy production processes, the quantity of energy delivered (energy output) by renewable sources is often less than or is equal to energy consumed in capturing and delivering it to customers (energy input) (Heinberg and Mander 2009). Hence, the technologies to produce renewable energies are not sufficiently mature enough to replace fossil fuels in terms of energy gain. This means that policymakers should implement policies and strategies to support, financially, the improvement of their energy efficiency for the socio-economic and technical consequences of their large-scale replacement in the energy system.
In this context, it could be very useful to provide information to institutions and/or public organizations about the current state of the art of RES energy efficiency to help them to redefine their political vision (Clack et al. 2017) for realizing sustainable socio-economic systems, in line with the transition from fossil fuels to renewable energies.
Consequently, the aim of this paper is to offer a critical analysis of the academic literature of studies concerning the EE indices of RES in the last 20 years, in order to present the state of the art in this field and identify recent research and energy efficiency trends.
The authors point out that the paper does not represent a systematic review due to the encountered methodological limits, mainly related to the heterogeneity of the contents of the bibliographic sources.
Specifically, the authors will focus on the energy return ration (ERR) and some its typologies, such as the net energy ratio (NER) or renewability factor (RF), the net energy gain (NEG) or net energy value (NEV), and the energy returned on investment (EROEI) applied to biomass, solar and wind energy. They will not consider hydroelectric power production, since it is a mature technology and the possibility of improving further its energy efficiency is limited (Fagnart et al. 2020). The reason of this choice is due to the fact that these indices are able to assess the amount of useful energy that remains after energy consumed to drill, pipe, refine, or build infrastructure (including solar panels, wind turbines, dams, nuclear reactors, and drilling rigs) and that subtracted from the total quantity of energy produced from a certain source.
In this framework, an economy that can obtain energy with a higher value of NER, NEG or EROEI can distribute more labor and capital among different activities. Therefore, such an economy has a greater potential for development and growth than another (Fagnart and Germain 2016).

Energy Return Ratio and Relative Indices
To assess the energy efficiency of RES, several methodologies and relative indices can be used. Some scholars prefer to adopt, for example, energy analysis (EA). This method, introduced by H.T. Odum in 1983, evaluates the integrated relationship between the economic system and its environment. In other words, it considers "total direct and indirect available energy of one kind used to make a product or service, measured in solar emjoules (sej)" . The EA calculates some indices, such as: i) the emergy yield ratio (EYR), which measures the ability of a process to exploit local resources by using economic resources from outside; ii) the environmental loading ratio (ELR), which quantifies the environmental stress caused by the production system; iii) and the Energy Sustainability Index (ESI), which measures the potential contribution of a resource or process to the economy per unit of environmental loading (Buonocore et al. 2015).
Although this analysis is important to assess the efficiency of an energy process, the authors focused expressively on net energy analysis, as stressed in the introduction section, for calculating the different typologies of energy return ratio. Generally, this choice allows us to obtain more strictly economic evaluations than those of energy analysis.
NEA is a wide set of methods used to determine the efficiency of a process in capturing and converting energy. For this reason, this type of analysis is specifically applicable to energy producing industries. NEA's goal is to calculate whether the energy output supplied by a production system, surpasses the energy input required to build, operate and maintain the infrastructure (Palmer and Floyd 2017). Moreover, it is necessary to take into account the quantity of energy to extract the energy resources and convert them, etc. (Evans 1982). In fact, the cumulative quantity of commercial energy (i.e., renewable, nuclear or fossil), invested to extract, produce and manufacture a product and to transport it to its point of use (embodied energy), can be converted only in a portion of useful energy (or exergy). This is because a certain amount of energy will be unavoidably lost as entropy effects and technological inefficiency (or anergy) of the process (Liao et al. 2015).
NEA is often useful for calculating the ERR of an energy source, which compares the amount of energy consumed in the production process and the amount of energy suitable for societal use. This assessment is similar to the embodied exergy analysis, which consists of calculating how much useful energy is lost and embodied in a product; additionally, it quantifies the emission and waste amount, caused by energy depreciation, allocated either to producers or to consumers (Liao et al. 2015). For instance, while a photovoltaic (PV) cell has a well-defined conversion efficiency (e.g., MJ electricity/MJ solar radiation), the energy efficiency of its entire supply chain and electricity distribution needs to be calculated using an ERR (Brandt and Dale 2011).
Over the years, several ERR typologies have been proposed, such as NER, NEG, and EROEI, which assess both the quality of an energy source and the sustainability of processes to harness it.
According to Spath and Mann (2000), NER can be defined as: where NER is the net energy ratio of a defined energy source (such as electricity, fuels, etc.); Eg is the energy gained from an energy source (output) (i.e., electric energy delivered by the utility grid) and expressed in Joules or kWh; Eff is the energy consumed within the energy production system (input) and expressed in Joules or kWh. NER is dimensionless, and to assess energy source efficiency, it should reveal a value greater than one. NEG refers to the difference between the energy consumed to produce an energy source and the quantity of energy gained from that production (Rahimi et al. 2018). NEG is equal to: NEG = energy content output -net energy input (2) where NEG is the net energy gain (Joules or kWh); energy content output is the total energy (gross) gained and corresponds to the term Eg in Equation (1) (expressed in Joules or kWh); net energy input is the total (gross) energy consumed and corresponds to the term Eff in Equation (1) (expressed in Joules or kWh). Clearly, a NEG is achieved when the value is positive. EROEI is the ratio of the amount of energy gathered from a source and the energy consumed directly or indirectly by the process (Fagnart and Germain 2016) and is equal to: EROEI= (total energy output) ⁄ (total energy input) where EROEI is the energy return on energy investment; total energy output is the energy left over after an energy source is produced (expressed in Joules or kcal); total energy input is all energy required to produce an energy source (expressed in Joules or kcal). As the numerator and denominator are generally expressed in the same units, the index is dimensionless. To measure the energy efficiency of a fuel, EROEI should be a ratio of at least up to 5:1 for non-renewable energy sources and 3:1 for renewable sources (Murphy et al. 2011). When comparing Equations (1) and (2), EROEI and NER seem to be similar. This is true, but the lack of a standardization method prevents a comparison between these indices (Brandt and Dale 2011). In fact, Fagnart and Germain (2016) consider these indices different, as EROEI assesses only the input and output energy of the main process, while NER also considers intermediary consumption energy flows. Consequently, the latter is more detailed than the former. Moreover, Hall et al. (2014) recognize the following four typologies of EROEIs depending on system boundary: 1) standard (EROIST); 2) point of use (EROIPOU); 3) extended (EROIEXT); 4) societal (EROISOC). The first approach considers all energy outputs (direct and indirect) and all energy inputs without considering labor and financial services. Moreover, if EROIST is applied to a fuel, the assessment of energy output includes all energy consumption from production to the firm gate. For instance, EROIST takes into account, for the production phase, the energy consumption for oil extraction, but the agricultural phase for biodiesel.
Conversely, EROIPOU considers a system boundary more expanded than the previous one because it involves energy inputs associated with refining and transporting a fuel. Thus, for biofuel production, all phases of transformation in fuels are evaluated; for instance, with biodiesel, agricultural, oil extraction and trans-esterification ones are evaluated, while for ethanol, agricultural phase, sugar extraction, and distillation ones are evaluated.
In EROIEXT, the system boundary is further expanded to include the energy required to obtain and transport a fuel to the point of use (i.e., oil or diesel pump).
Finally, EROISOC differs from the others, as it considers the numerator the sum of the energy content of all fuels delivered to a city, region, or nation and the denominator the total energy consumed to get those fuels.
To calculate the ERR, in the beginning, both direct and indirect (embodied) energy requirements for producing an energy source were considered. This method is derived from engineering process modeling and follows a bottom-up approach, where all energy inputs of the main production phase are considered (Lenzen and Munksgaard 2001). Subsequently, this procedure was standardized by life cycle assessment (LCA), a methodology regulated by ISO 14040 (ISO 2006a) and ISO 14044 (ISO 2006b) that analyses environmental impacts of a product from resource extraction (cradle) to processing to disposal (grave). When an LCA study focuses only on energy aspects, it is often called an energy life cycle assessment (ELCA), which consists of an energy balance accounting for energy used in the production process and compares it to the amount of energy contained in the resulting energy sources (Pradhan et al. 2008). For instance, to assess the EER of the electricity produced in a PV plant, it must be considered as the unit of energy obtained (e.g., 1 kWh) and the total energy consumed over the lifetime of the device (e.g., 20 years). The latter refers to energy necessary for: constructing and decommissioning the plant, extracting and transporting silicon and other materials, producing the cell, transportation of materials, etc.
In the Figure 1, the relationship and hierarchical structure of the abovementioned indices are reported.

Literature Review
It was very difficult to find some general reviews on the ERR linked to renewable energy resources, such as biomass, solar and wind energy. This was due to the very extensive field of the considered study and, above all, because scholars preferred mostly to carry out a systematic review or meta-analysis on a specific RES, comparing, where possible, the same energy efficiency index. Thus, according to this premise, the authors aimed to analyze mainly the direction taken by the research into the energy efficiency of RES, highlighting both the prior state of the art and recent trends in the academic literature. Nevertheless, in the following section, some (mostly representative) examples of reviews have been reported. Only a few of them, as mentioned, focus properly on the aim of this study; the other ones, though equally important, focus only on the energy efficiency of specific RES.
For instance, Murphy et al. (2016) analyzed the inconsistencies common within the NEA literature, and in the assessment of energy return ratio and its typologies. Indeed, though it is important to have a broad measure of energy profitability, "comparisons, among indices, are often misleading, particularly if the units being compared are different". Consequently, the authors proposed, as a possible solution, the application of the LCA methodology within the NEA community.
Walmsley et al. (2018) "collected energy assessment data for the most common electricity generation methods and evaluates five Energy Ratios", i.e., EROI (standard and external), Energy Payback Time (EPT), Primary Energy Factor (PEF), and Resource Utilization Factor (RUF). Thus, the authors compared the values of both fossil fuels and renewables sources. They showed that EROIs correspond to short Energy Payback Times and vice versa and that "Energy Ratio performance levels for renewable energy generation sources -hydro, wind, geothermal and solarheavily rely on the quality of the primary natural resource available". The authors recommended EROI (external) and Resource Utilization Factor as the most useful metrics for inclusion in full sustainability assessment.
Capellán-Pérez et al. (2019) applied a new methodology to "assess the energy and material investments required in the next decades to achieve the transition from fossil fuels to renewable energy sources in the electricity sector". They considered a number of novel methodological contributions in relation to the state of the art of energy systems and EROI analysis, for resolving some of the existing differences in the literature. The indices, identified in the literature, showed levels well below the thresholds needed to sustain industrial complex societies, making questionable the transition towards a totally green electricity sector.
Fabre (2019) tried to "estimate future EROIs in a decarbonized electricity system by examining a broad range of scenarios". These scenarios were given by the several estimations of EROIs made in literature, derived from diverse data sets and methodologies, of which none was authoritative. The author arrived at the same conclusion as Capellán-Pérez et al. (2019)-renewable electricity is not as energetically efficient as previously thought. However, the choice to undertake a renewable transition-or not-has to take into account not only EROI indices but also the negative externalities and the scarcity of fossil fuels.
Moving towards reviews on EE indices applied to a specific RES, many of them are related to energy from biomass, followed by solar-PV and, lastly, by wind.
Regarding the latter, an initial study was carried out by Lenzen and Munksgaard (2002). They reported results in terms of energy intensity, stressing that the existing life cycle assessments of energy and CO2 intensity for wind turbines showed considerable variations due to "discrepancies in the energy contents of materials and the analyses' methodology and scope". To minimize this issue, they proposed to use a standardized methodology, and input-output-based hybrid techniques. Kubiszewski et al. (2010) published a meta-analysis of EROIs based on a survey of wind turbine literature, ranging in date from 1977 to 2007 and adding almost 50 more studies to the publication of Lenzen and Munksgaard (2002). They evaluated only operational studies with a power rating for turbines less than 1 MegaWatt (MW). They concluded that even if "wind energy could yield significant economic and social benefits relative to other power generation systems" … "technical progress and a high EROI are not sufficient conditions for the continued rapid expansion of wind energy. A number of social, economic, environmental and regulatory issues need resolution". Price and Kendall (2012) stressed, two years after, that the conclusion of Kubiszewski et al. (2010) lacked sufficient data or used unreliable data, highlighting a lack of transparency in reporting modeling inputs and outputs. Thus, they stated that "it is necessary to establish clear methods for assessing transparency and to develop conventions for LCA reporting that promote future comparisons". Thus, they identified guidelines for evaluating reporting transparency, that predict the necessities of meta-analyses of LCA applied to renewable energy technologies.
Regarding reviews on solar PV, Hernandez and Kenny (2012) published the first important paper in 2012. The authors assessed the methods for calculating and measuring the energy performance of Domestic Solar Water Heating, considering the existing studies on life cycle, embodied energy and 'net energy' performance. They revealed: "that while there is a real potential for life cycle energy savings through domestic solar water heating installations, devising mechanisms to ensure proper design, installation and operation of systems is essential for this technology". Bhandari et al. (2015) carried out, instead, a systematic review and a meta-analysis of the embedded energy, energy payback time (EPBT), and EROI metrics of PV technologies, published in 2000-2013. They stated that "Across different types of PV, the variation in embedded energy was greater than the variation in efficiency and performance ratio suggesting that the relative ranking of the EPBT of different PV technology today and in the future depends primarily on their embedded energy and not their efficiency". Good (2016) presented an evaluation of the published results relative to life cycle assessments of different hybrid photovoltaic-thermal (PV/T or PVT) concepts and installations, aiming to find a common ground. The author stated that "payback time for both energy and greenhouse gas emissions of the PV/T systems are much shorter than their expected lifetime. However, due to the use of different methods and unclear data sources, it is difficult to make any wide-ranging conclusions about the environmental impact of PV/T modules".
Al-Waeli et al. (2017) reviewed various aspects (such as performance parameters and efficiencies) of PV/T systems through the existing literature, aiming to highlight the key points in this field and show the different techniques used for such systems. They did also a critical review of many findings of previously conducted research. They concluded: "the PV/T air heater system is promising for future preheating air applications. Furthermore, the PV side views require more attention in technical and cost terms. However, more research is essential to reduce the cost and, improve the effectiveness and technical design of such systems".
Koppelaar (2017) analyzed twenty-nine studies on mono-and polysilicon solar-photovoltaic (PV) modules, providing an Energy Payback Time (EPT) or NER value and examining the influence of data age, system boundaries, and technological configurations. Considering the different problems in comparing several studies, the author recommended: "in future studies to use recent data reflecting up-to-date technological standards and include the collection year of any used datasets. And to strictly follow existing ISO14040, ISO14044, and IEA-PVPS T12 standards, especially by transparent reporting of: solar module specifications, energy inputs for individual facilities and non-module components, technology assumptions, and electric/thermal conversions". Palmer and Floyd (2017) reviewed the differing approaches of NEA as relevant tools for assessing the role of PV in future energy systems and considering the EROI and energy payback time (EPBT). They discovered that there were many apparent divergences between studies due the different goal definitions. Thus, it is important to consider EROI or EPBT "with specific reference to the details of the particular study context, and the research questions that it seeks to address. NEA findings in a particular context cannot definitively support general statements about EROI or EPBT of PV electricity in all contexts".
The last part of this section focused on reviews about energy obtained from biomass. A first paper was published by Pradhan et al. (2008) and investigated models used to calculate the net energy ratio (NER) of biodiesel production. The aim was to indicate the reasons for the opposing results deriving from several studies done on this supply chain. The authors discovered that "the most significant factors in altering the results were the proportions of energy allocated between biodiesel and its co-products. The lack of consistency in defining system boundaries has apparently led to very different results". Consequently, they proposed a unified model for biodiesel energy analysis to answer the renewability question because, often, the same definitions of NER used among the models were different. Milazzo et al. (2013) reviewed the performance and prospects of rape biodiesel production on a global basis using some 40 life cycle assessments, identifying best (agricultural) practice and laggards. In their concluding remarks, the authors stressed that: "Opportunities to improve rape biodiesel life cycle energy efficiency and environmental impact by implementing new technologies in agriculture as well as in industrial processing have been identified for various Brassica oil crop cultivars in relevant production areas. The consequences of large-scale renewable energy action plans have been considered. Improvements are needed for rape biodiesel to stay in business". Rocha et al. (2014) assessed and compared, by meta-analysis with previous published studies, the main environmental life cycle impacts and energy balance (NER) of ethanol, from sugarcane, and biodiesel from soybean and palm oil. The authors concluded: "the choices of co-products allocation method, transport distance and inventory database of the country have significant influence on the results of the life cycle environmental performance of biofuels". Arodudu et al. (2017a) studied tools, methodologies and frameworks that can be adopted individually or together with each other for bridging the methodological gaps in application to agro-bioenergy systems. They argued that using "energy return on energy invested-EROEI, human appropriation of net primary production-HANPP, net greenhouse gas or carbon balance-NCB, water footprint individually and in combination with each other will also lead to further derivation of indicators suitable for assessing relevant environmental, social and economic impact categories". Carneiro et al. (2017) aimed to identify global tendencies based on a statistic and critical interpretation of previously published study results, reviewing 61 recent papers addressing the environmental evaluation of microalgae biofuels. They considered, among other elements, Energy Ratio (ER) allowing a broad estimation of the net energy efficiency of biofuels. The authors stated: "in terms of energy efficiency, algae biofuels cannot compete with other biofuels or fossil fuels. They present very low performances, even demanding more energy for its production then the energy they can deliver. Energetic and GWP performances can be improved if production pathways are carefully chosen and optimized". Ketzer et al. (2018) developed a comparative meta-analysis of 16 recent LCA studies on algal bioenergy production pathways, calculating the energy return on investment (EROI) for each study. Of course, also in this review, large variations for the EROI were underlined. This depended "on the assumptions made on biomass productivity, lipid content, required energy, and the output of the system, the energetic performance was assessed". Ribeiro et al. (2018) presented new information regarding the conventional torrefaction process, as well as analyzing the literature and studies of the innovative techniques that have been in development for the improvement of torrefied biomass qualities. They argued that "the publications regarding this topic have been strongly increasing, which suggests a strong academic and industrial interest for this subject in the last few years". They stressed the importance of the torrefaction parameters on the final properties of torrefied biomass, as well as the importance of reactor scales for the definition of ideal protocols.
Considering this literature Considering this literature review, it is evident that, in each paper published in recent years, emerge discrepancies resulting in difficulties to compare them even if come from the same field of research. Most of the reviews stressed the importance of developing guidelines, including, for instance, a common methodology such as LCA to overcome this issue and to measure more properly the energy efficiency of RES.

Methodology
In this paper, a qualitative analysis has been used to develop an in-depth and comprehensive summary of the academic literature on the ERR indices related to energy from wind, solar and biomass. Data came from the scientific literature generated through a combination of: a) database searches (cross-discipline platform of Elsevier and Clarivate) as of March 2020, and b) screening references of studies retrieved under a). The aim was to select studies highly referenced to identify, underline, and assess all data related to the abovementioned subject. For an initial selection, the keywords used for the research were: i) EROEI or EROI; ii) NEG or NEV); iii) and NER. For each, the term "renewable" was added. Publications were chosen if they included, as a minimum, the calculation of one of these indices. The authors decided to not limit the search to a specific timespan.
The total number of energy efficiency scientific papers was 162. Six were related to wind, 31 concerned solar, and 125 regarded biomass energy. The first paper was published in 1998, and the most recent studies were published in 2020, covering more than a 20-year interval. Since 2009, the quantity of studies has significantly increased with an average of 15 papers per year, but there was a considerable reduction in the last years (Figures 2 and 3).    Figure 3 shows that the most calculated index is the EROEI (78), followedby NER (45), others (21), and NEG (18), because it provides powerful insights into the real balance sheets that drive our "petroleum economy" over the transition to lower carbon energy sources (Hall and Klitgaard 2012;Rye and Jackson 2018).
According to Brockway et al. (2019), numerous aspects have affected the increasing attention that has been paid to the study of EROEI, for example: i) the decrease in EROEI ratios of fossil fuels due to the depletion of finite resources, the leading energy source up to at least 2050; ii) the fact that the EROEI ratios for RES are often controversial, since they can change according to the adopted methodology, and are commonly lower than fossil fuels. Thus, it must be considered that renewable-led energy transition has unfavorable socio-economic impacts.
Capellán-Pérez et al. (2019) underlined that many studies have been carried out to estimate the EROEI of individual RES technologies (as the authors have already shown in Figures 2 and 3) and important differences exist depending on the technology, system design and location, and the field is plagued with methodological differences related to the functional units (e.g., an MJ of heat energy versus an MJ of grid electricity) or the boundaries of the analysis (i.e., mine mouth vs. end use or energy technology vs. energy system) (Price and Kendall 2012;Weißbach et al. 2013;De Castro et al. 2014;Raugei et al. 2015;Ferroni and Hopkirik 2016;Murphy et al. 2016, Hall 2017De Castro and Capellán-Pérez 2018).

Results and Discussion
The authors decided to limit the discussion to only EROEI/EROI, according to the abovementioned premise, choosing among those studies ones that were more similar and thus more easily comparable. Thus, a final number of 30 main studies were taken into account (Table 1).
The analysis showed that, in general, all renewable energy sources present a good level of energy efficiency, even though it was impossible to compare all values calculated in the various studies. Indeed, as already stressed in Section 3, there are different energy conversion technologies, raw materials, and system boundaries that, without a standardization method, do not permit accurate assessment.  Table 1 showed that wind energy presents the highest EROEIs when considering models and conceptual installations, although results were slightly lower when considering operational installations. For instance, in New Zealand, EROEIs were about 34 with peaks of about 58 over a lifespan of 20 years, whereas offshore installations had values equal to 16.7-17.7 depending on the presence of a substation (Kubiszewski et al. 2010;Walmsley et al. 2017;Huang et al. 2017;Dupont et al. 2018).
The EROEIs of PV farms have been revisited and computed in Switzerland, with values ranging from seven to 10 according to the different system boundaries considered (Raugei et al. 2017). PV installations showed EROEIs ranging widely according to the used technology (semitransparent building-integrated PV, thin film technologies of crystalline silicon, etc.) and locations (insulations) from five5 to 34 (Poh Khai and Mithraratne 2014;Bhandari et al. 2015;Pickard 2017;Zhou and Carbajales-Dale 2018; Lui and Van den Bergh 2020). Goel et al. (2020) focused on two different PV systems (standalone and grid-connected systems) showing an EROEI equal to 5.42 and 6.62, respectively. However, even if the EROEI value increases, it is unthinkable to power the European electrical system using renewable energies. In fact, Dupont and Jeanmart (2019), who have assessed the potential electricity production from wind and solar power in the European Union, affirm that, considering the current EROEI values, it is impossible to supply the electricity required in Europe using only these technologies because it is necessary to increase energy storage systems. This implies the decrease in EROI value and, consequently, a reduction in electricity efficiency.
With regards to biomass, the variability of the results were greater than wind and solar energy due to the high number of available conversion technologies (thermochemical, biological, and chemical-physical) and of several types of feedstock (dedicated crops, animal and vegetal residual biomass, marine biomass) for obtaining biofuels and bioenergy. This is also the reason why there are many more studies published in this research field.
Consequently, it was very difficult to compare all the different EROEI results. Even if they considered the same technologies and feedstock, they also varied in some way. Thus, the authors decided to divide in three main categories the EROEI scores: below one; from one to five; over five.
In the first ranking, there is the production of ethanol, from any kind of grain (oat, wheat, barley, corn and sweet sorghum) in papers published in recent years, as well as bio-crude oil and synthetic natural gas (SNG) from micro-algae (Beal et al. 2012;Brandenberger et al. 2013;Uusitalo et al. 2017;Wang and Cheng 2018). In the second, some efficiency values often slightly exceeded the threshold of the unit, especially from biodiesel from rapeseed and ricinum (Restuccia et al. 2013;Firrissa et al. 2014;Van Duren et al. 2015;Amouri et al. 2017;Uusitalo et al. 2017) and ethanol from corn stover, corncorb, cassava (Manihot esculenta C.), corn (high intensity agriculture) and straw (Arodudu et al. 2017b;Murphy et al. 2016;Uusitalo et al. 2017;Wang and Cheng 2018). However, the EROEI value also depends on the technology used to produce a biofuel. For instance, a study on ethanol from spring triticale cultivated with traditional systems reduced tillage technology and nitrogen fertilization, showing EROEI values equal to 1.138, 1.089 and 1.114 respectively (Kikocla et al. 2020). Moreover, the typologies of considered systems, as well as the various obtained energies, could lead to different results. For instance, Yeo et al. (2019) proposed Smart Food Waste Recycling Bin (S-FRB) technology to produce three different typologies of energy (i.e., biofuel, electricity and heat, biogas) from food waste and wood chips in four different scenarios. The study found that producing biogas or, simultaneously, electricity, heat and biogas, obtained high values of EROEI, i.e., six and eight, respectively. On the contrary, biofuel and electricity/heat had an energy efficiency index equal to two. It is more favorable the production of biogas from grass, wood, willow, roadside vegetation, and microalgae (Meyer et al. 2014;Meyer et al. 2015;Uusitalo et al. 2017;Barbera et al. 2018). In the last category, the score fluctuates significantly from 5.9 for ethanol from corn to 22.4 for biogas from the same crop, both with the application of a low intensity agriculture system, characterized by a small quantity of energy inputs like fertilizers, fossil fuels, and pesticides (Arodudu et al. 2017a). Interesting results are also shown by the performance of the use of sugarcane (Saccharum officinarum L.) and stems, together with organic waste, equal to eight (Preston and Rodríguez 2009), and combustion in the heating of industrial hemp, equal to 12.6 (Kolarikova et al. 2014). Finally, the work of Zupko (2019) differs from the aforementioned categories because it assessed the EROI of renewable gasoline and diesel produced from woody biomass residues using integrated hydropyrolysis and hydroconversion technology. The study showed an energy index equal to 4.19 and 4.31 per kilogram of diesel and gasoline, respectively.
Though not reported in Table 1, some interesting considerations could be made regarding the NER index. In the case of biomass, for instance, this varies from 1.09 to 1.19 for the production of ethanol from grains (Dong et al. 2008), while it reaches higher levels, from 3.14 to 9.3, if it is made from sugarcane or cassava (Dias De Oliveira et al. 2005;Macedo et al. 2008;Hanif et al. 2017). The most efficient process is that of the gasification of willow with an NER equal to 13.3 (Heller et al. 2008), while diesel from palm oil (Elaeis guineensis J.) has the same range of ethanol production (3.58) (Pleanjai and Gheewala 2009). These results are almost similar to those of the EROEI values calculated for the same biofuels.
In this context, some final considerations can be made: there is no evident evolution of EROEI increment over time according to the papers collected in this review. It is likely that this is due to the fact that the year of publication does not always correspond with the conducted study. Obviously, only the advancement of technology can improve the level of EE of RES and not an increase in biomass yield. Some technologies, like anaerobic digestion and combustion, are already mature and so the final EROEI score generally presents higher values.
Finally, the EE values can increase if the calculation takes into account the energy content of byproducts of the supply chain. For instance, in the case of biodiesel from microalgae or soybean, the inclusion of the calorific value of glycerine and flour resulting from the extraction of oil increased the energy indices (Rana et al. 2013;Piastrellini et al. 2017).

Conclusion
Energy production from RES may be more widespread and less expensive in a scenario with high energy efficiency and could support the transition towards decarbonizing the global energy system. Therefore, policymakers should consider any visions of a rapid, reliable, and low-cost transition to whole energy systems completely based on these alternative eco-friendly sources.
Our analysis showed that the trend is characterized by many studies relating to biomass because it is the renewable source with a wide range of raw materials and conversion technologies. It is also highlighted that the references have decreased in the last two years, after a peak in 2017 and an average increase of 15 papers per year. The reduction is probably due to the lack of industrial applications of advanced technologies.
The analysis showed that all RES generally have an acceptable level of energy efficiency, in terms of ERR, even with peculiar differences among sources. In particular, solar and wind energy, although they display possible room for improvement, have reached mature and advanced levels of technology, unlike biomass. Biomass energy can be used to produce almost all energy carriers, and its efficiency depends on the large variety of biomass feedstock, different energy extraction and conversion methods, and the manner in which the biomass is produced and collected. Thus, it can be considered as a versatile RES that is able to bridge the energy supply and demand gap in different energy applications such as electricity, transport, and buildings (Tareen et al. 2018). However, according to the analyzed research, bioenergy and biofuels seem to be much more expensive and less efficient than equivalent fossil-based energy. This is due to their lower density, higher moisture content, and hydrophilic nature, which causes their heating value to decrease, making it difficult to use biomass for large-scale productions (Ribeiro et al. 2018). Consequently, it is important to overcome these issues in improving the energy efficiency in order to allow energy from biomass to play an important role in future energy models.
In this direction, Moriarty and Honnery (2019) suggested, for example, the introduction of a green EROIfor measuring the global net energy, produced in an ecologically sustainable manner. Many scholars, indeed, believe that it is necessary to consider, along with EROEI value, sustainable indices, such as CF, EROC and ESA.
Policymakers should take into consideration all of this evidence for redefining their policies and strategies to foster the large-scale replacement of fossil fuels with RES in the energy system.