Eco-efficiency assessment of technological innovations in high-grade concrete recycling

the technologies for high value-added concrete recycling. critical environmental and economic evaluation of such technological innovations is important to inform making, been a of this field. Hence the present study aimed to assess the efficiency of the technical innovations in high- grade concrete recycling, using an improved eco-efficiency analytical approach by integrating life cycle assessment (LCA) and life cycle costing (LCC). Four systems of high-grade concrete recycling were analyzed for comparison: (i) business-as-usual (BAU) stationary wet processing; (ii) stationary advanced dry recovery (ADR); (iii) mobile ADR; (iv) mobile ADR and Heating Air Classification (A&H). An overarching framework was pro- posed for LCA/LCC-type eco-efficiency assessment conforming to ISO standards. The study found that technological routes that recycle on-site and produce high-value secondary products are most advantageous. Accordingly, policy recommendations are proposed to support the technological innovations of CDW management.


Introduction
Construction and demolition waste (CDW) is widely acknowledged as one of the most important sources of waste (Koutamanis et al., 2018). This is especially true for Europe, where the stock of buildings and infrastructure was built during World War II and renewal including https://doi.org/10.1016/j.resconrec.2019.06.023 demolition of such stocks is now a main activity for the building and construction sector. Eurostat estimated an annual CDW generation of 970 million tons in the European Union (EU)-27 (Vegas et al., 2015). The CDW has been identified by the European Commission (EC) (2001) as a priority stream because of the large amounts that are generated and the high potential for re-use and recycling embodied in these materials.
For this reason, the Waste Framework Directive (EC, 2008) requires member states to take any necessary measures to prepare for material recovery, by 2020, of at least 70% (by weight) of CDW. The current recycling percentages of CDW per European country vary between less than 5% in Montenegro and more than 90% in countries including Belgium, Portugal, and the Netherlands (Eurostat, 2018). The vast majority of CDW is down-cycled, for instance in road foundation, or even landfilled in some European countries. For example, in 2003, the Spanish construction sector only recycled 10.3% of CDW, while 25.6% was deposited in inert waste landfills, and 64.1% was dumped in the absence of controls in debris sites, pits or watercourses (Rodr and Alegre, 2007). In 2012, Switzerland recycled 51%, landfilled 26%, incinerated 8% (combustible fraction such as wood), and re-used 15% onsite (Hincapié et al., 2015). In Europe, the average composition of CDW shows that up to 85% of the waste is stony waste (Gálvez-Martos et al., 2018) such as End-of-life (EOL) concrete. An alternative market of recycled aggregates derived from EOL concrete was already established in Europe, where the EOL concrete was re-used for road base material (Anastasiou et al., 2014). Experts foresee that landfill of EOL concrete can be reduced to 0% and that the use of recycled concrete aggregates in road construction can significantly contribute to reaching the 70% target for CDW recycling in the EU (Bio Intelligence Service, 2011).
The Netherlands has achieved 100% recycling of EOL concrete and has a more advanced concrete recycling and CDW management system than China (Zhang et al., 2018;Huang et al., 2018), Australia (Tam et al., 2010), Canada (Yeheyis et al., 2013) and other European member states (Eurostat, 2018). The most common practice for concrete recycling in the Netherlands is simply crushing and subsequent use as a base in road construction, which is considered a low-grade or low value-added route. Currently, the most commonly applied method for high-grade recovery of concrete is the wet process, which produces clean aggregate for concrete by washing the coarse aggregate, leaving the fine fraction (sieved sands) for road base filling and generating sludge, which needs be treated. A downside of the wet process is that it requires a large washing plant, which is expensive. Therefore, more than 90% of the waste concrete in the Netherlands is still processed low-grade for use in road base materials.
In the coming years, a continuous increase of the amount of CDW and EOL concrete is expected in Europe because of the large number of constructions built in the 1950s which are coming to the end of their life. At the same time, options for low-graded reuse will become more limited, since road construction will stabilize (Bio Intelligence Service, 2011). So, higher value-added solutions are needed for the EOL concrete that cannot be absorbed in road construction.
In 2011, UNEP (2011) advocated "greening the waste sector", referring to a shift from less preferred waste treatment and disposal methods, such as landfilling, towards options that contribute to the highest reduction of the use of primary resources. The growth of the waste market, increasing resource scarcity and the feasibility of new technologies create opportunities for high value-added recovery options, also in the case of the EOL concrete. Technical progress and green technical innovation are necessary not only to improve the productivity of industries, but also to enhance the environmental benefits of re-use, recovery, and recycling (Song and Wang, 2018). Governments are imposing more stringent regulations, while other parties, including suppliers, consumers, and banks, are formulating requirements for ecoproducts and green technology (Klostermann and Tukker, 1998). Moreover, new products need to be prepared for upcoming challenges concerning lower carbon footprints, resource depletion and shortages and also concerning cost-effectiveness in a competitive market place . Over the last few years, novel technologies have been developed that aim to guarantee high-quality recycled raw materials for manufacturing new construction products, thereby closing the concrete loops.
In Europe, the European Commission (EC) funded an innovation project called C2CA (Concrete to Cement and Aggregate, www.c2ca. eu), which aims to develop a cost-effective approach for recycling highvolume EOL concrete streams into prime-grade aggregates and cementitious fines (Lotfi et al., 2014). The C2CA project proposes an innovative solution called Advanced drying recovery (ADR). It constitutes a dry alternative to the existing wet process, which significantly reduces the processing cost for high-grade recovery of the coarse fraction of EOL concrete. However, the initial plan to use the fine product of ADR as a feed-in kiln for cement production was not optimal due to the required long-distance transportation of fines.
In the C2CA project, the equipment for the ADR process was a semimobile facility that could not yet be used for in-situ EOL concrete processing. The challenge to make the ADR technology transportable for in-situ use was taken up by a follow-up project called HISER (Holistic Innovative Solutions for an Efficient Recycling and Recovery of Valuable Raw Materials from Complex Construction and Demolition Waste, www.hiserproject.eu). In this project, a mobile ADR set was developed that can be transported by one truck and assembled in one day.
Although the mobility of the ADR set has been improved, the fine fraction (0∼4 mm) materials generated during the high-grade concrete recycling are still not valorized, being left on site or used as filling material for road base or land leveling. This issue was taken care of by the EC VEEP project (Cost-Effective Recycling of CDW in High Added Value Energy Efficient Prefabricated Concrete Components for Massive Retrofitting of our Built Environment, www.veep-project.eu). In the VEEP project, the ADR system was combined with a Heating-Air Classification System (HAS) to refine the fine fraction of the output of the ADR process for the production of high value-added products -clean secondary sand and cementitious fine materials.
The environmental benefits and economic consequences of different recycling routes are commonly assessed via eco-efficiency evaluation that combines Life cycle assessment (LCA) and Life cycle costing (LCC). Although the concept of eco-efficiency itself is not new or complex, a better specification is desirable to assess the co-benefits of technological innovations. A series of innovations in high-grade concrete recycling offers a good study case to investigate how technological development would influence the efficiency changes in CDW management. Using field data collected from the C2CA, HISER and VEEP projects, this study presents an eco-efficiency assessment, from a practical perspective, to understand whether each step of the innovation generates environmental benefits and if so, at what financial cost. Is it possible to achieve an environmental-economic win-win situation in high-grade concrete recycling? Would the innovations trigger any potential burden-shifts (environmental and economic)? The findings of such an investigation are expected to shed light on the technological development of future concrete recycling and on the feasibility of a circular economy in the construction sector. Moreover, from a theoretical perspective, this case study on concrete recycling proposed a framework for LCA/LCC-type eco-efficiency assessment.

Literature review of eco-efficiency analysis
The concept of eco-efficiency was designed to guide the ecological and economic efficiency improvement in a production system within a company, by measuring the environmental impact caused per monetary unit earned. Eco-efficiency can be mathematically expressed as shown in Eq. (1) (Keffer et al., 1999).

Eco efficiency
Value added Environmental impact (1) ESCAP (2009) defines eco-efficiency as a key element for promoting fundamental changes in the way societies produce and consume resources, and thus for measuring progress in green growth. It is commonly accepted that eco-efficiency was first mentioned by Sturm and Schaltegger in 1989: "the aim of environmentally sound management is increased eco-efficiency by reducing the environmental impact while increasing the value of an enterprise" (Bohne et al., 2008). Later, it was popularized by the World Business Council for Sustainable Development (WBCSD) for the business sector in the course of the United Nations Conference on Environment and Development (UNCED) in 1992. Eco-efficiency was the first developed academically in 1990 and prominently promoted by WBSCD in 2000 (Kicherer et al., 2007). Since then, eco-efficiency has been variously defined and analytically implemented, and in most cases, eco-efficiency is taken to mean the ecological optimization of overall systems while not disregarding economic factors (Saling et al., 2002). The "eco-efficiency assessment" is a concept rather than a specific appraisal tool. Eco-efficiency analysis can be deployed by using data envelopment analysis (DEA) as the efficiency measurement vehicle (Korhonen and Luptacik, 2004). However, DEA is more likely to explore efficiency issues at meso-and macro-level (Mardani et al., 2017;Chen and Jia, 2017;Tajbakhsh and Hassini, 2015;Atici and Podinovski, 2015;Gerdessen and Pascucci, 2013), whereas the environmental and economic impacts of technological innovations on concrete recycling are essentially product-level issues.
In 2012, eco-efficiency assessment was standardized in ISO 14045 (2012) as a quantitative management tool which enables the study of environmental impacts of a product system along with its product system value for a stakeholder from a life cycle perspective. In this manner, the eco-efficiency assessment which examines the life cycle of a certain product is more adaptable to product-oriented issues. The framework of eco-efficiency assessment, which is based on LCA standards, was outlined in 6 steps in ISO 14045 (2012), and in this framework, LCA is employed for "environmental assessment" conforming to ISO 14040 (2006) and14044 (2006). ISO 14045 (2012) defines three ways to present a value system: functional value, monetary value, and other values (e.g. aesthetic, brand, cultural and historical). However, it does not specify the tool for the economic value assessment. Based on Eq. 1, Bohne et al. (2008) argued that "value added" cannot be used in a recycling-system context in the same way as at the firm level, because profits which stakeholders seek to make along the way do not necessarily increase the value of the material but arise from their performance of services, and "cost" is used to denote all economic transaction. As LCC is a methodology for the systematic economic evaluation of life cycle costs (ISO, 2017), we reckon that the financial analysis for this study via an LCC assessment would be an appropriate approach for making decisions on the cost-effectiveness of a product. We reviewed some typical LCA/LCC-type eco-efficiency studies and listed their methodological choices in Table 1. Table 1 shows that eco-efficiency assessment has been applied to multiple domains: waste management, energy, construction, and daily necessities. However, the assessment method is far from standardized yet. First, ISO 14045 (2012) did not specify the method for the product value system assessment, and a guideline on LCC and LCA under an overarching eco-efficiency framework is lacking. Second, in LCC cost structures were broken down in different ways and life-cycle costs were randomly expressed in different cost forms. Third, even though sensitivity and uncertainty analysis are mandatory in ISO 14045 (2012), they are not common practice yet either on LCC and LCA separately or on the eco-efficiency index as a whole.
To fill the knowledge gap, the present study proposes a protocol 1) to embed LCA and LCC inside a joint eco-efficiency framework under ISO standards; 2) to add an additional "economic impact assessment" step to multi-dimensionally break down the cost structure and classify cost stressors; and 3) to present a solution for the quantification of sensitivity and uncertainty in an LCA/LCC-type eco-efficiency assessment.

Table 1
Literature related to LCA/LCC-type eco-efficiency analysis.

Framework for integrating LCC and LCA for eco-efficiency
According to the ISO 14040 (2006) and14044 (2006), an LCA is organized in four steps: 1) goal and scope definition; 2) life cycle inventory (LCI) analysis; 3) life cycle impact assessment; 4) life cycle interpretation. We tried to apply the environmental LCC conforming to the guidebooks published by the Society of Environmental Toxicology and Chemistry (SETAC): Environmental Life Cycle Cost (Hunkeler et al., 2008) and Environmental Life Cycle Cost：a Code of Practice (Swarr et al., 2011), in which LCC is classified into three types: conventional LCC, environmental LCC, and societal LCC. In this eco-efficiency study, the cost indicator is supposed to relate the environmental indicator which is based on LCA; therefore the cost indicator was calculated according to the environmental LCC methodology (Swarr et al., 2011), in which the LCC is constructed in three steps: 1) goal and scope definition; 2) LCI analysis; 3) life cycle interpretation. According to the SETAC guide, LCC need not include the step of "impact assessment" as it is already clear that a lower cost is better. However, the types of cost and the time factor of the cost are also important when evaluating the economic impacts of technological innovations. We argue that not only the sum of the life cycle costs but also the breakdown of the cost structure needs to be investigated in the LCC analysis. Therefore, in this study, an "impact assessment" step is added to the LCC analysis, which consists of a definition of cost categories and cost impact category selection. In the first step, "cost category definition", cost breakdown structures were applied to present the cost distribution and to identify cost stressors. The second step, "cost impact category selection", introduced issues such as whether to employ a discount rate over time, and how the proposed life cycle cost will facilitate decision making. Fig. 1 gives an overview of the updated integrated framework for the LCA/LCC type eco-efficiency analysis. The "economic impact assessment" step for the LCC analysis is depicted with a dashed rectangle in Fig. 1, in analogy to the environmental impact assessment in LCA.

Goal
The goal of this study was to assess and compare the eco-efficiencies of four high-grade concrete recycling alternatives enabled by the technological innovations of ADR and HAS. The presently available high-grade recovery method -the wet process -serves as a reference to illustrate the potential changes led by the innovations. The geographic scope of the study is the Netherlands, where the field data of the case study were collected. The temporal scope of the study is recent years (2015-2019).

Description of the innovative technologies 3.2.2.1. Wet process.
In 2010, when the C2CA project started, about 2% of the EOL concrete in the Netherlands was processed for high-grade applications, such as recovered clean aggregates for concrete. The commonly applied method is the wet process. Within the C2CA project, the wet process data were collected from a wet treatment plant located in Utrecht, which represents the BAU high-grade concrete recycling method. In the wet process, the pre-crushed concrete rubble (0∼0.5 mm) is transported by a truck to a stationary wet process treatment plant with a productivity of 150 ton/h. There the EOL concrete is broken down to 22 mm, and sieved into recycled coarse aggregate (RCA) above 4 mm and sieved sand (SS) below 4 mm. Then the coarse fraction (4∼22 mm) of the aggregates enters a long water bed for washing. After crushing and washing, the high-grade 4∼22 mm RCA is sold for concrete manufacturing, substituting natural coarse aggregate (NCA). The washing residues are pumped to a thickener for sedimentation, and sludge is generated and sent to a landfilling site. The 0∼4 mm SS is a mixture of dirt, sand and hydrated cement, which  (Auer et al., 2017) Glass container Eutrophication potential (EP),-Photochemical ozone creation potential (POCP),-Global warming potential (GWP) and-Acidification potential (AP) Zhang, et al. Resources, Conservation & Recycling 149 (2019) [649][650][651][652][653][654][655][656][657][658][659][660][661][662][663] prevents its high-grade application, e.g. clean sand for new concrete manufacturing. Consequently, SS is seen as a residue in the production of the 4∼22 mm RCA. Due to its chemical inertness, SS is often piled up in-situ. However, if a nearby construction project needs to balance earthworks, SS could be given away free of charge or sold at a very low price. Since the application of SS is uncertain, in present study the environmental and economic impact of SS is cut-off. The mass balance of the investigated wet process is presented in Fig. 2.

Advanced dry recovery (ADR).
When the C2CA project started, the ADR technology had already been successfully applied for the recovery of incineration bottom ash. In the C2CA project, the technology was used to recover the high-grade concrete aggregate.
The original version of the ADR system was already much smaller than the wet processing plant; however, in the C2CA project, dismantling and assembling the ADR system took a week, which meant that in practice it was used as a stationary recycling plant. In the HISER project, the mobility of the ADR equipment was improved, and in the VEEP project, a truly mobile ADR set was developed, which can be transported by one truck and assembled and dismantled on site within one day. In the case studies carried out in the C2CA, HISER and VEEP projects, the ADR process is combined with pre-crushing. In an ADR system, the EOL concrete of about 0.5 m is crushed to 22 mm and sieved to a fraction above 12 mm as a final product and below 12 mm as ADR feed. The 12∼22 mm fraction is about 20% of the crusher output, which is quite clean and was used as clean coarse aggregate for concrete. About 80% of the crusher output is in the 0∼12 mm fraction and is fed into the ADR set. The ADR breaks up the feed material and classifies it into 4∼12 mm RCA, which is used as highgrade concrete aggregate, and 0∼4 mm SS, which contains pollutants and for which no suitable high-value applications are found yet in the C2CA and HISER projects, hence it is usually stacked on site or left for land leveling or road foundation due to its inertness. As the mass balance of the ADR system ( Fig. 3) shows, the ADR set transforms 68% of its feed material into high-grade coarse aggregate and generates 32% of 0∼4 mm SS, for which suitable applications have to be found. Otherwise, the more concrete is recycled with the ADR system, the more 0-4 mm fines will require disposal. Thus, the impact of 0∼4 mm SS is cut-off in the ADR process, as it is in wet process.

Heating air classification system (HAS).
The VEEP project took up the challenge to valorize the fine products of ADR. In the exploration, the Heating Air Classification System (HAS) was  C. Zhang, et al. Resources, Conservation & Recycling 149 (2019) 649-663 proposed for treating the 0∼4 mm SS residue of ADR. The HAS is capable of separating cementitious powder from the sandy part in the fine particle fraction. The HAS uses simultaneous heating, grinding and separation in a fluidized bed, which can remove most of the 0∼0.125 mm recycled ultrafine particles (RUP, 6.4%) from the 0.125∼4 mm recycled fine aggregate (RFA, 25.6%). The RUP can be used to reduce the cement consumption in concrete manufacturing, and the 0.125∼4 mm RFA can be used to substitute natural sand in concrete manufacturing. The mass balance of HAS is presented in Fig. 4.

Recycling systems used in the comparative evaluation
Offered by the series of innovations in high-grade concrete recycling, four systems representing the potential alternatives for the treatment of EOL concrete in the Netherlands are assessed to capture changes in eco-efficiency: 1) BAU WP (wet process) system; 2) ADR-S (stationary) system; 3) ADR-M (mobile) system; 4) A&H (ADR & HAS) system. Details of each system are presented below.

BAU WP (wet process) system.
In the wet process system, EOL concrete is transported to the wet processing plant. Through the wet recycling process, 4∼22 mm RCA for concrete manufacturing is produced together with the below 4 mm SS.

ADR-S (stationary) system.
In the ADR-S system, crushed concrete rubble is transported by truck to the plant with a stationary ADR set. 4∼12 mm and 12∼22 mm RCA for concrete manufacturing is produced together with the below 4 mm SS.

ADR-M (mobile) system.
In the ADR-M system, a mobile ADR set is transported and reassembled for in-situ treatment. 4∼12 mm and 12∼22 mm RCA for concrete manufacturing is produced together with the below 4 mm SS.

Functional unit
Comparability of assessment is particularly critical when different systems are being evaluated (ISO, 2006a). Since the wet process, the ADR and HAS system deliver different products, each product system was expanded to ensure the comparability. Since the residue 0∼4 mm SS is cut-off due to its uncertain position as a good or waste, the basket of functions for the comparison of the expanded product systems are: 1) EOL concrete treatment, 2) coarse aggregate for concrete production, 3) fine aggregate for concrete production, 4) cementitious material for concrete production. Based on the mass balance of the combined ADR and HAS system the functional unit for the comparative study is defined as following (see The reference flows of each system are presented in Table 2.

Life cycle assessment (LCA)
3.3.1. Environment inventory analysis 3.3.1.1. System boundary and unit processes. Inventory analysis is the phase that defines the product system, including system boundaries, flow diagram with unit processes, data collection and allocation for multifunctionality (Guinée et al., 2001). Since the Netherlands is one of the major European countries involved in C2CA, HISER and VEEP projects for technological systems development. This study takes the Netherlands as the geographical reference area. Since selective demolition and sorting is a common practice in the Netherlands, very few contaminations are contained in the EOL concrete waste. To simplify modeling, unnecessary process like residue disposal is omitted in this study. It is assumed that the target EOL concrete for analysis does not contain any contamination. After selective demolition and sorting, EOL concrete generated at the construction site in the Netherlands will be crushed into 0∼0.5 m size and then sorted on site, and cost and impacts of this procedure will not be considered.
The life cycle considered in the study comprises three phases: I) Transport; II) Recycling; and III) virgin material production. The first phase considers the transportation of the EOL concrete for treatment. It varies from different technology systems. For the off-site ones, it includes the transportation of the EOL concrete from the demolition site to the recycling plant. For the in-situ recycling pathways, it refers to the transportation of the processing equipment. The recycling phase is about processing EOL concrete into diverse secondary products, which can be used as raw materials for concrete manufacturing, so save virgin materials, accordingly. In order to guarantee the compatibility across different technology systems, virgin material production processes are added in several systems, which are grouped in the phase of virgin material production. It is assumed that the transport costs for the secondary products are the same as for virgin materials to their next destination. Based on the defined 3 phases, the flows diagrams for 4 systems are depicted in Fig. 5. As experiments have shown that the use of the secondary raw materials (0∼0.125 mm RUP and 0.125∼4 mm RFA) produced by HAS can reduce comparable amounts of the virgin cement and virgin sand in concrete production (Technalia, 2018). It is modeled as that the generation of HAS fine products 0.125∼4 mm RFA and 0∼0.125 mm RUP will lead to the avoided production of the virgin sand and cement.
3.3.1.2. Data collection. As indicated, process-based LCA was used for the environmental impact assessment. We used the software OpenLCA  C. Zhang, et al. Resources, Conservation & Recycling 149 (2019)  1.7.4 to perform the LCA analysis with the Ecoinvent 3.4 database in combination with foreground processes, which are listed in Table A1 in the supporting information. The background processes that were linked to the foregrounds are listed in Table A2 in the supporting information. Data for the BAU WP system were obtained from an industrial wet treatment plant located in Utrecht, the Netherlands (within the C2CA project). Data for ADR was collected from the semi-mobile installation in the C2CA project and from the ADR demonstration in the HISER project. The mobile HAS data is gathered from the recycling lab of the TUD, the Netherlands. Data of energy use and emissions were compared to those of relevant diesel-engine equipment for verification. For the technical systems which do not generate certain secondary products as specified in the functional unit in Table 2, the production of their natural counterpart materials (e.g. gravel, virgin cement, virgin sands) were modeled by using Ecoinvent datasets.
3.3.1.3. Multifunctionality. When a process delivers more than one function, we encounter a 'multifunctionality' problem. ISO 14040 (2006) recommends avoiding allocation by either dividing the process or expanding the system boundary. According to the data obtained, multifunctional processes cannot be divided into discrete sub-processes, thus the system boundary was expanded by using a basket-type functional unit. In S1 BAU WP especially, recycling of 100 tons of EOL concrete through WP will generate 52.9 tons of RCA but less than the amount of 68 tons in the functional unit. Thus, an additional 15.1 tons of NCA is produced in S1 BAU WP. Besides, in S1 BAU WP, S2 ADR-S, and S3 ADR-M, 25.6 tons of virgin sand and 6.4 tons of cement are produced.

Environment impact assessment
The impact assessment phase in an LCA includes characterization of the result based on an impact category selected, followed by an optional normalization and weighting process (Guinée et al., 2001). ISO 14044 requires a deliberate assessment of all relevant impact categories for an LCA study; therefore, it is not allowed to leave out impact categories that have a significant impact. Besides, evaluation of a range of novel technologies indicates the need for a broader environmental perspective. Joint Research Centre of the European Commission (EC-JRC) recommended a comprehensive ILCD life cycle impact assessment method. The impact categories in the ILCD method (ILCD 2011,

Fig. 5.
Flow diagrams for the four systems: (i). S1 BAU WP (wet process) system; (ii). S2 ADR-S (stationary) system; (iii). S3 ADR-M (mobile) system; and (iv). S4 A&H (ADR & HAS) system, those flows which cross through the system boundary are functional flows, those lower-case letters at the end of the reference flow arrows refer to the sub-functional unit in Table 2.
C. Zhang, et al. Resources, Conservation & Recycling 149 (2019) Table 3. Normalization and weighting are optional steps of LCA according to ISO 14040/14044 to rank the impacts of a system. However, decisionmaking becomes easier when the impacts are normalized, as this compares the contribution of a particular service with the overall environmental problems under eco-efficiency consideration (Kicherer et al., 2007). Normalization was based on "JRC EU 27, 2010, total [year]", which stands for impact in 2010 of the 27 European Union countries.
After normalization, the next step is to combine the normalized values via a weighting scheme. ISO 14045 (2012) regulates that weighting shall not be used in a comparative eco-efficiency analysis intended to be disclosed to public. However, in order to present a solution to the sensitivity and uncertainty analysis of the final eco-efficiency results, this case study tried to weight the environmental indicators in a relatively objective way. In fact, in an eco-efficiency context, it may be found that one recycling system is better than another for some impact categories but poorer for others. In that case, it is difficult to figure out whether the total environmental performance was improved or deteriorated. Thus, a weighting method is indispensable to aggregate all impact category indicators into one sole environmental score, making it possible to calculate an eco-efficiency ratio. There is no scientific basis for weighting LCA results, since weighting requires value choices (ISO, 2006a). However, the expert opinions about impact category weights are sensitive to either subjective biases in elicitation situations or in local characteristics (Seppälä et al., 2005), which may consequently result in a wide range of uncertainty. To render the results universally compatible and applicable for all EU member states, this study applied an equal weight (0.066) recommended by EC-JRC (2016).

Life cycle costing (LCC)
3.4.1. Economic inventory analysis 3.4.1.1. Data collection. LCC analysis shares the same system boundary as that of LCA. All costs are expressed in the currency of the Netherlands: Euro (€). It is also a problem that some economic values keep fluctuating over time, such as the price of aggregate, which shifts with market supply and demand. We therefore used historically observed data from different sources and then adjusted those data according to confirmation with relevant actors. To perform the LCC study, Microsoft office 2016 Excel was used to investigate the main contributions of costs, connected with a parametric cost database. The cost data were validated by comparing them to the Ecoinvent 3.4 cost database to avoid noticeable deviation. Details and sources of the price data are presented in Table 4. Table 2 Reference flows in each system under a coherent functional unit. Note: Secondary products are marked in [bracket]; primary products are underlined. EOL denotes end-of-life, RCA means Recycled Coarse Aggregate, NCA means Natural coarse aggregate, RCA means Recycled coarse aggregate, RFA means Recycled fine aggregate and RUP represents Recycled ultrafine particle. was the same as that of LCA, and system expansion was used.

Economic impact assessment
LCC quantifies costs to operate the same technological systems that were evaluated in LCA, while SETAC suggested not to have an impact assessment step for LCC (Swarr et al., 2011). Moreover, the life cycle costs of a product is a number expressed in monetary units; thus normalization and weighting are not performed either (Swarr et al., 2011). However, for different product systems, we were faced with different cost categories, while costs and benefits could also be incurred at different moments in time. An economic impact assessment was performed in this section to better align the economic information with the environmental ones generated by LCA. We propose two stages in the economic impact assessment: 1) cost category definition, which answers the question how the cost will be structured in LCC analysis; and 2) cost impact category, which answers questions on how the time factor will be considered and how the final cost value will be expressed.

Cost category definition.
Given the diversity of LCC equations, the selection of LCC equations can play a central role in how LCC results are interpreted (Miah et al., 2017). The life cycle cost can always be broken down according to the life cycle phases, such as in the concrete recycling case, as shown in Eq. (2), where C I , C II , and C III represent the costs of Phase I Transport, Phase II Recycling and Phase III Production of virgin material, respectively.

Life cycle cost = C I + C II + C III
(2) On the other hand, the costs can also be categorized into different types of cost, such as transport cost (TC), equipment cost (EC), personnel cost (PC), utility cost (UC), waste treatment cost (WC), virgin material cost (VC). Thus, the life cycle cost can be estimated as in Eq. (3).

Life cycle cost = TC + EC + PC + UC + WC + VC
If life cycle cost is estimated via Eq.
(2), it will be clear how the cost is attributed to each phase; via the Eq. (3), we would know the share for each category of cost. Thus, in this study we could deploy cost structure breakdown using these two forms of cost category. In principle, further differentiation of costs and benefits is possible, i.e. which actors over the life cycle are confronted with costs and benefits. Since in this LCC analysis there is only one stakeholder (the recycling company), adding actors as a third dimension was not considered here.

Cost impact category selection.
In the cost impact category selection stage, two main issues were addressed: 1) will the incurring moment of the costs and benefits in time be considered? 2) how will the final cost value be expressed? If costs and benefits are spread out over a long time span, a conscious decision is needed on whether one wants to discount costs and benefits that occur in the future, and which discount rate is applied, which leads to a dynamic-type LCC model; on the other hand, if costs and benefits occur in a very short time span, discounting does not need to be considered, which results in a static-type LCC model (Hunkeler et al., 2008). In this study, all unit processes take place in a short period; therefore, we add costs and benefits without considering any discounting over time. In fact, as mentioned by in the SETAC LCC book (Hunkeler et al., 2008), environmental LCC usually is a steadystate method. Discounting of the final result of an environmental LCC Table 4 Cost data in three life cycle phases and their sources.

Life cycle phase Explanation
Phase I Transport Transport cost (TC): costs related to the transport of raw and ancillary materials, EoL concrete, products, and equipment. Waste transport: the transport cost is 0.1 €/t⋅km (including the cost of fuel and personnel costs of the staff) 1 . Personnel cost (PC): costs related to wages and salaries. Wages and salaries in the construction sector are set as 35.9 €/man-hour 5 . Especially personnel cost for wet processing plant is 0.65 €/t 4 . Utility cost (UC): costs related to utilities (e.g. electricity, diesel, water). Diesel price is 0.73 €/L 6 ; electricity price is 0.06 €/kWh 7 ; water (for dust control) price is 0.16 €/L 1 ; tap water (for wet process) is 0.003 €/L 1 . Lubricating oil for machine is omitted from this study. Waste treatment cost (WC): costs related to sludge treatment (wet process methods only). Sludge treatment is 25 €/ton (including transport) 4 .

Phase III Production of virgin material
Virgin material cost (VC): costs related to the procurement of primary raw materials which cannot be produced through wet process and ADR. NCA price is 10.2 €/ton 1 ; sand price is 12 €/ton 1 ; cement price is 75 €/ton 1 .
Notes: 1 data from Strukton BV without documental support; 2 data from investigation in TUD; 3 data from HISER project unpublished report "Final Report of Integrated environmental and economic assessment for the HISER case studies" in 2018; 4 data from C2CA project unpublished report "Life cycle costing of concrete recycling: comparison between a conventional and the C2CA technology" in 2016; 5 EUROSTAT, Labour cost levels by NACE Rev. 2 activity (the Netherlands, 2018), via http://ec.europa.eu/eurostat/web/products-datasets/-/lc_lci_lev; 6 data from Ecoinvent 3.4 cost database; 7data from Eurostat "Electricity prices for nonhousehold consumers -bi-annual data" via https://ec.europa.eu/eurostat/web/products-datasets/-/nrg_pc_205;  Zhang, et al. Resources, Conservation & Recycling 149 (2019) 649-663 will be specified in detail in our further studies. For the question of how to express the cost values, 9 approaches can be considered (see Table 5). Some other approaches were mentioned by Miah et al. (2017), such as "Net LCC" by Menikpura et al. (2016), "Total Annualized Equivalent Cost" by Pretel et al. (2016), and "Resale Value" by Minne and Crittenden (2015). These are conceptually overlapping with those in Table 5. Furthermore, the "Present Worth" method developed by Afrane and Ntiamoah (2012) includes monetization of externalities, which does not fit in this eco-efficiency approach, where the environmental dimension is a separate one covered by LCA. For those systems that contain costs over longer time spans, discounting can play a role in considering the time value, and cost can be expressed in forms from C 1 to C 8 . Since the costing system is defined as a static-state type, two possible LCC impact categories C 8 and C 9 were selected. Firstly, the exact cost of each technology was investigated without considering the time span; thus all costs in this LCC analysis are presented in static-state cost (C 9 ). Then, to make each technology more comparable in the form of an eco-efficiency indicator, life cycle cost of each system was normalized based on the baseline reference in relative value (C 8 ).

Integrated impact assessment: integration of LCA and LCC for ecoefficiency indicators
In this phase the environmental and economic results were elaborated by contribution analysis for identification of dominating factors, then the form of how the eco-efficiency indicator will be expressed was selected, and sensitivity and uncertainty analysis were conducted to evaluate the robustness. Firstly, should the eco-efficiency be expressed graphically or numerically? It is clear from Table 1 that there are two methods to express eco-efficiency: via numeric value and a two-dimensional diagram, and the last one is the most frequently used method. Providing a unified numeric value is convenient for decision making. However, it does not give easy insight into the relative scale and importance, and into the trade-offs between environmental and cost aspects. To overcome this drawback, the economic and environmental aspects can be plotted in a more visible and evident manner in a two-dimensional diagram. Therefore, a two-dimensional diagram was used to visualize the eco-efficiency results.
Secondly, there is an issue on whether the LCA and LCC results in the eco-efficiency graphs should be expressed in an absolute-value way (Huppes and Ishikawa, 2005) or relative-value way (Woon and Lo, 2016). In this study, evaluation of the eco-efficiency of technological innovations in high-grade concrete recycling would lead to different scores. The eco-efficiency of the existing recovery technology wet process was used as the reference basis. In this context, we believe LCA/ LCC in the percentage form would better reflect the improvement of an innovative system compared to the BAU system, thus a modified ecoefficiency indicator was adopted, presenting LCA/LCA results in a relative-value method. The eco-efficiency was interpreted through a twodimensional graph in Fig. 6.
The cost saved is presented through a relative LCC index in Eq. (4) as the Y axis of the graph; the relative LCA index is expressed through a relative LCA index in Eq. (5) as the X axis of the graph. Zone I represents full eco-efficiency (lower environmental impact and cost); Zone II (higher environmental impact, lower cost) and III (lower environmental impact, higher cost) indicate half eco-efficiency; Zone IV depicts non-eco-efficiency (higher environmental impact and cost). Therefore, if the location of a recycling system is closer to upper-right it represents a higher rate of eco-efficiency.
where LCC NOV , life cycle economic score of novel treatment; LCC BAU , life cycle costs of BAU treatment; LCA NOV , life cycle environmental score of novel treatment; LCA BAU , life cycle environmental score of BAU treatment. The S1 BAU WP is set as the origin point. Table 6 presents the indicator results calculated with OpenLCA, 15 impact categories for each system. The normalized results of 15 impact categories for each system are presented in Fig. 7. Finally, by applying weight, the aggregated impact for each system is shown in Fig. 8. Fig. 7 shows that from impact category indicator 1 to 10, the values of the environmental impact of systems from S1 to S4 presents an ascending trend; on the contrast, from indicator 11 to15 (in the rectangle), S4 A&H has the highest environmental impact (resulting from diesel consumption). Thus, the selection of impact category method will probably affect the environmental performance superiority of S4 A&H. All 15 impact indicators are considered in this study, however, uncertainty on the choice of impact categories cannot be modeled due to the limitation of the software. Fig. 8 shows that generally technological development is associated with a clear descending trend in the weighted environmental impact. Firstly, transportability is essential for the comparative advantages of an EOL concrete waste recycling system. Transport accounts for around 25% of the life cycle environmental impact in stationary recycling methods (S1 BAU WP and S2 ADR-S). After optimization of the transportability of the recycling equipment (S3 ADR-S and S4 A&H), less than 1% of the life cycle environmental impact is contributed by transport. Another factor contributing to the comparative advantages of the HAS system is the high-value recovery of secondary raw materials. For the first three systems, S1 BAU, S2 DAR-S, and S3 ADR-M, the impact of virgin material production contributes 69%, 72% and 95% to their life cycle impact, respectively. Even though the HAS technology shows a surging increase of environmental impact in the recycling phase, from the calculated results we can see that its advantages can certainly be realized since the virgin cement and sand consumption can, in fact, be reduced by using the recovered RUP and RFA. Compared to the wet process (S1 BAU WP) and the ADR system (S2 and S3), HAS technology (S4 A&H) can reduce the total environmental impact by 31%∼54%. C. Zhang, et al. Resources, Conservation & Recycling 149 (2019) Fig. 9 show a similar trend as the LCA results. From an economic perspective, four systems show a descending cost trend, and S4 A&H is the most cost-efficient pathway. In general, cost savings are mainly realized by a reduction in transport and production of higher value-added materials. Compared to the stationary recycling methods (S1 BAU and S2 ADR-S), on-site recycling systems (S3 ADR-S and S4 A&H) can reduce the share of transport in life cycle costs from 33%∼44% to 1%. Furthermore, the life cycle costs of S3 ADR-S are slightly higher (9%) than that of S4 A&H, although they both can be considered as economically feasible methods for concrete recycling. However, there is a clear trade-off between virgin material cost (in S3 ADR-S) and personnel cost (in S4 A&H).

Eco-efficiency index
Based on the modified eco-efficiency Eq. (5) and (6), the life cycle cost and life cycle environmental impact are translated into the relative life cycle cost and relative life cycle environment impact, respectively. Then those relative values are located in the eco-efficiency graph as shown in Fig. 10. Graphically, all comparative systems are located in Zone I, and the S4 A&H is the best choice for concrete recycling from an eco-efficient perspective, as it can noticeably reduce both life cycle environmental and economic burdens by about 55%. The S1 BAU WP turns out to be the costliest and most environmentally unfriendly pathway, and S2 ADR-S only slightly improved the eco-efficiency by around 20%.

Sensitivity analysis
For the assessments to be useful in the actual decision-making processes, knowledge of the uncertainty and sensitivity of the data is of great significance. In the assessment, LCA and LCC were used in an ecoefficiency assessment to estimate the environmental impact and economic value. The sensitivity and uncertainty in the calculation may result from inventory data, allocation options, characterization factors, and weighting factors. According to the Handbook on life cycle   C. Zhang, et al. Resources, Conservation & Recycling 149 (2019) [649][650][651][652][653][654][655][656][657][658][659][660][661][662][663] assessment (Guinée et al., 2001), sensitivity and uncertainty analysis cannot be made obligatory due to the limited functionality of LCA software, but it is recommended to implement such an analysis at least partially. Since the latest version OpenLCA 1.7.4 is unavailable for full sensitivity and uncertainty analysis, the present study only considered the sensitivity and uncertainty of critical cost data and unit process data from environmental and economic inventories. Based on the economic and environmental results highlighted, we list the most relevant 15 factors with respect to variations as shown in Table 7. The sensitivity analysis was conducted to identify the factors that are the most sensitive to economic and environmental performance. By decreasing 10% of each factor, their sensitivity is shown in Table 8. A positive value of sensitivity is presented in red and a negative value of sensitivity is presented in green. The darker its color, the more sensitive the factor will be. For stationary recycling systems, S1 BAU WP and S2 ADR-S are the most sensitive to transport-relative factors (price and travel distance), followed by cement price; in contrast, mobile recycling systems S3 ADR-M, S4 A&H are insensitive to transport. S3 ADR-S is the most sensitive to cement and sand price, while S4 A&H is noticeably sensitive to HAS productivity, which, however, will not be improved currently. S4 A&H is also sensitive to personnel cost.

Uncertainty analysis
The factors which were evaluated in the sensitivity analysis were selected for an uncertainty analysis. Their values of the range were determined by consulting with relevant actors, as shown in Table 9. Since according to the HAS developer, the productivity and unit diesel usage of HAS will remain steady in the near future, therefore their uncertainty were not considered. A single standard error range of ± 5% for the LCI data was chosen in this study, which is an accepted approach to the uncertainty of LCI data (Huijbregts et al., 2003). Thus, a market price fluctuation range ± 5% for LCC uncertainty factors (from u 1 to u 9 ) and for environmental inventory data (u 14 ) was selected. Apart from that, truck travel distance and the amount of EOL concrete generation at demolition site have a larger range of uncertainty, more than 50% of fluctuating rate was given to those factors u 10 to u 13 as shown in Table 9.
Taking into account the uncertainty of those data, the final economic and environmental performance with uncertainty ranges of each scenario is shown in Fig. 11. The stationary recycling systems S1 BAU and S2 ADR-S have a wider range of uncertainty mainly because of the fluctuation of truck travel distance.

Policy implications
The results indicate that different technological innovations have different potentials to improve eco-efficiency. Technological innovations are responsible for improving the product quality and reducing the recycling cost, while policies are responsible for fostering a functional market for the recycled concrete product to evolve. The results do not intend to be precise quantifications, but rather to demonstrate the potential contributions of those EOL concrete technological strategies toward sustainable growth. Following the eco-efficiency assessment, if policymakers want to support the eco-efficient growth of concrete recycling networks and technologies, then they should impose relevant policies at least for the following perspectives: (1) Avoiding transport of waste. In this study, the on-site recycling routes, S3 ADR-M and S4 A&H, demonstrated an obvious advantage in transport distance reduction. Mobile recycling solutions are assumed to be a good answer to solve EOL logistic issues, from an economic and an environmental point of view. Therefore, on the one hand, further efforts are required to optimize the mobility of recycling facilities. Plant/equipment could be designed for modular construction and efficient dismantling, transportability as well as assembling, and recycling companies would have the opportunity to share one processing facility among several production sites instead of transporting a massive amount of EOL concrete to a recycling site. Policies, on the other hand, could constrain the waste transport via tax tools or by increasing road toll.
(2) Enacting regulations and standards for secondary raw material. Standards for building materials are based on virgin materials and are not always useful for secondary materials. Local governments of EU member states need to consider introducing a quality certification mechanism and maximum percentage use of recycled material, especially for emerging products such as RUP from HAS.   Productivity of HAS C. Zhang, et al. Resources, Conservation & Recycling 149 (2019) [649][650][651][652][653][654][655][656][657][658][659][660][661][662][663] (3) Enhancing publicity and promotions of technological innovations. Even though the ADR and HAS pathways were shown to generate more financial and environmental gains than the traditional wet process, if virgin material will still be massively consumed, the practical application of more eco-efficient high-grade recycling methods is hindered by misconceptions and information asymmetry between the construction or recycling companies and technology developers. Governments are responsible for helping those developers to improve the visibility and publicity of those innovations.

Conclusion
EOL concrete is the predominant constituent in CDW with a high potential for re-use and recycling. In EU countries, EOL concrete is usually down-cycled for road base or even used in landfills. It is important to shift from less preferred EOL concrete treatment and disposal way towards methods maximizing resource efficiency. In Europe, novel technologies have been developed aiming to guarantee high-quality recycled secondary raw material from EOL concrete for use in the manufacturing of new concrete products, thereby closing the concrete loops. Eco-efficiency assessment provides a useful tool for steering decisions towards sustainable resource management, considering economic and environmental aspects at the same time. This paper presents a comparative eco-efficiency analysis methodology for assessing the environmental and economic performance of technological innovations ADR and HAS for EOL concrete recovery by comparing them to the BAU method wet process. This study proposes a framework protocol for LCA/LCC-type eco-efficiency assessment. Besides, an "economic impact assessment" step is proposed for LCC to specify cost breakdown structure, types of cost expressed, and cost stressors, in analogy with the "environmental impact assessment" step in LCA. Next, this case study presents a solution for conducting sensitivity and uncertainty analysis in an eco-efficiency assessment.
The study showed that the most advantageous technological routes are recycling on-site and producing high-value secondary products. The higher eco-efficiency performance system S3 ADR-M and S4 A&H reduced the life cycle environmental impact to a large extent and minimized the life cycle cost by ensuring transportability of the recycling facility. However, for the fine fraction of HAS, the recovered product (0∼0.125 mm RUP and 0.125∼4 mm RFA) cannot replace cement and sand 100%, but it can reduce the use of cement and sand in the production of concrete. Calculation of the achievable reduction of cement and sand led to a modeling choice in favor of HAS. Besides, S4 A&H has the worst performance on some impact categories indicators such as photochemical ozone formation, acidification, etc., which, however, are compensated by other indicators under an eco-efficiency context, thus somehow concealing the energy-intensive personality of HAS. With respect to policy implications, relative policy recommendations are as follows: avoiding the transport of waste; enacting regulations and standards for secondary raw material; enhancing the publicity and promotions of technological innovations.
This study has several limitations. First, the cost data is largely based on a Dutch context, and higher availability and lower cost of primary material in some other EU member states will challenge the competitiveness and market share of secondary material. Second, this study used lab-scale data of HAS; the performance of HAS in a more developed stage (i.e. on a pilot scale and industrial scale) will be discussed in further research. Third, we excluded some factors, such as the exact distribution of the recycling plants, transportation cost of the products and virgin material to the next destination, the variation of some recycling technologies, and the uncertainty of impact category indicators selection, which may have influenced the results. Finally, this Table 8 Sensitivity analysis results (each factor decreased by 10%). Other environmental inventory data in LCA: ± 5% Fig. 11. The uncertainty of eco-efficiency for four systems.
C. Zhang, et al. Resources, Conservation & Recycling 149 (2019) 649-663 study demonstrates a preliminary concept of an "economic impact assessment" step for LCC with a case study on eco-efficiency assessment; a more comprehensive and systematic illustration will be presented in the near future.