Recovery of electronic wastes as fillers for electromagnetic shielding in building components: An LCA study

The present study reports the development of sandwich panels for building walls having electromagnetic interference (EMI) shielding abilities. Conductive polymer composites (CPCs) have started being employed as EMI shielding materials. In this paper we propose the use of a conductive polymer composite flat sheet made of high-density polyethylene (HDPE) recovered from municipal solid wastes (MSW) used as polymeric matrix, “doped” with dispersed metal fillers recycled from e-wastes. Test results proved that the recycled metal fillers enhance the electrical conductivity and enable EMI shielding. Different sandwich panels were discussed in the context of building applications, using identical HDPE/metal-filler EMI sheets, but different thermal insulation material (polystyrene and glass wool). The life cycle assessment (LCA) methodology was applied to evaluate the environmental impact generated during the following steps: a) recycling of thermoplastic materials fromMSW; b) recovering of metallic components from waste PCB; c) re-use of the recovered components into sandwich panels with electromagnetic shielding properties for buildings. The goal of the LCA was to perform a comparative analysis of the composite sandwich structures manufactured to be used as EMI shielding in buildings applications in order to assist the materials selection and eco-design. By means of the LCA results it was possible to manufacture a building component with good EMI shielding properties and reduced environmental impacts. © 2020 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).


Introduction
Over the last 50 years, the increase in population and industrialization has increased the demand for electrical and electronic equipment (EEE) such as cell phones, printers, televisions, and personal computers with the increasing of societal living standards (Siddique and Siddique, 2019). But when it comes to the disposal of such electronic devices, many concerns arise due to the content of some toxic metals which have severe human health and environmental effects if they are not disposed of carefully (Kyereet al., 2899;Rehman et al., 2020). Therefore, such electronic wastes (ewaste or WEEE) require a proper management both for environmental protection and economic savings, considering that most metals are valuable materials if they are recovered and reused. Developed countries differ from developing countries in respect to e-waste as in developing countries threats from inadequate handling are greater, formal recycling systems are lacking and recycling legislations are either weak or absent (Heeks et al., 2015). Many studies exist in this topic as reported in a comprehensive literature review conducted by Ikhlayel in 2018 (Ikhlayel, 2018). Here the author recommends the application of a systematic integrated e-waste management (IEWM) for developing countries as a necessary alternative to the existing approach. The European Directive on WEEE (European Union, 2012) aims at reducing the disposal of wastes and promotes efficient use of resources by reuse, recycling and other forms of recovery of such waste.
However, the integration of the materials, components and structures does not make for ease of recovery of the metals, polymers, ceramics, etc.

Novelty of the work
The study we present hereafter fits very well into this contest as we propose a possible recovery solution for metal components from e-waste and their reuse in the production of an innovative composite material where the use of recovered metals brings both benefits of enhancing the final properties and reducing the environmental burdens. We studied and tested a possible process to recover metals from printed circuit boards (PCB) and reuse them in the production of conductive metal-thermoplastic panels that are proved to act as ElectroMagnetic Interference (EMI) shielding materials (Chung, 2002). ElectroMagnetic Interference (EMI) shielding materials prevent penetration or leaking of electromagnetic radiation from enclosures (Guan et al., 2006). The rapid development of the electrical industry has led to EMI becoming a serious problem in modern society. EMI not only causes operational malfunction of electronic instruments but can be harmful for humans under certain circumstances. Many diseases such as leukaemia, miscarriages, breast cancer, are correlated to continuous exposure to ElectroMagnetic (EM) fields. Effective shielding is of primary importance to protect the indoor and outdoor environments from unwanted electromagnetic waves. It is particularly needed for the buildings containing power transformers, mobile communications and other electronic facilities (Li et al., 2008). Research studies on new materials solutions with good electromagnetic interference shielding effectiveness are very current Yang et al., 2020;Song et al., 2020).
Recently, conductive polymer composites (CPCs) have started being employed as EMI shielding materials owing to their low cost, strong resistance to corrosion, lightweight and good processability in comparison with conventional metals. The electrical conductivity (s) of polymers can be enhanced by the addition of conductive fillers developing CPCs with EMI shielding properties. The maximum fraction of conductive fillers was reported to be 10 wt%; greater concentrations can lead to poor dispersion of particles and increase the viscosity of the polymer (Lin et al., 2003;Sankarana et al., 2018).
In the present study we propose the use of HDPE recovered from MSW as polymer matrix to produce a conductive polymer composite flat sheet by addition of dispersed powder rich in metals, recovered from waste PCB boards. Recovering and reusing HDPE from MSW is a realistic practice due to the good resistance of this thermoplastic material. For example, studies on the addition of asphaltenes in HDPE have shown that introduction of up to 4% of filler to the polyethylene matrix does not significantly change the physical-mechanical properties (Borisova et al., 2019).

Why LCA
LCA methodology was applied to evaluate the environmental impact of the materials under study. We selected the LCA method as it is recognized as the most effective tool in line with the recent Circular Economy (CE) (Ellen MacArthur Foundatio, 2016) principles, referring to the challenges of resource scarcity, environmental impact or economic development, to promote a transition from a "Cradle to Grave" approach which means from materials extraction, manufacture use and waste production to a "Cradle to Cradle" approach in a "closed loop", where the wastes produced become itself nutrient for the next cycle.
Circular Economy Facilitates UN Sustainable Development Goals. The practice of recycling waste materials and reusing them in new products encounters the principles of circular economy towards the SDG goals (Stahel, 2016).
As this paper covers different topics it was necessary to arrange them into a clear structure, as follows: section 2 describes the LCA methodology and the application to the present case study; section 3 presents the experimental program including testing results and data collection for the life cycle inventory analysis; section 4 provides the life cycle impact assessment results and discussion and section 5 reports the conclusions.

LCA methodology
LCA was applied according to the ISO standards 14040e44:2006(O 14040. Environmental, 2006O 14044. Environmental, 2006) that define the following phases: 1) Goal and scope definition; 2) Life Cycle Inventory (LCI) analysis; 3) Life Cycle Impact Assessment (LCIA); 4) Life Cycle Interpretation. In particular, all data collected and used for the LCI were loaded into SimaPro v.8.3 ( e-Product Ecology Consu, 2012) accessing the Ecoinvent database v.2.2. The ISO standard allows the use of impact category indicators that distinguished into two levels: the "mid-point" level and the "endpoint" level. In general, indicators that are chosen close to the inventory result have a lower uncertainty, as only a small part of the environmental mechanism needs to be modeled, while indicators near endpoint level can have significant uncertainties. However, indicators at endpoint level are much easier to understand and interpret by decision makers than indicators at midpoint. In this study we used the CML2000 mid-point level indicator and the Recipe H/A endpoint level indicator. It is important to point out that the LCA provides a model that is a simplification of a complex reality. The present application can be considered a good approach for of eco-design as it gives information on materials selection and process optimization to improve existing products and the development of new products. In this way, an LCA study can be performed to assess the environmental impact e.g.: of a companyinternal product improvement, product development or technical innovations.

Scope & goal definition
In this phase it is needed to identify the objectives of the study and the system boundaries in order to avoid omitting relevant parts of the system to be investigated.
The goal of the present LCA was to perform a comparative analysis of the composite sandwich structures manufactured to be used as EMI shielding in buildings applications in order to assist the materials selection and eco-design.
Moreover, at this stage of the assessment, the Functional Unit (FU) is required to be chosen.

Functional unit and system boundaries
Cradle-to-gate analyses for two different EMI shielding systems were carried out. The functional unit is defined as the material assembly used for the production of one sandwich with dimensions 700 mm square by 2.5 mm thick with EMI shielding properties. According to the ISO 14040e44, an inventory analysis was carried out for each system under investigation to quantify the environmentally significant inputs and outputs by means of mass and energy balance.
As shown in Fig. 1, the system boundaries include the following processes: -Recovery of HDPE from MSW; -Recovery of metallic components from PCB waste; -Production of the metal-modified composite through extrusion and injection moulding; -Assembly of the sandwich shielding panels for building applications.

Life cycle inventory analysis
Primary data, literature data and Ecoinvent v2.2 database were used as following: -HDPE from MSW: primary data were collected from the company All Green AS; -Metal-powders from PCB: inventory data were taken from the Ecoinvent database and at laboratory scale; -Metal-HDPE flat sheet production: primary data were collected at laboratory scale; -EMI shielding sandwich panels: inventory data were collected at laboratory scale. Two different panels were produced.

Metallic components recovered from wasted
PCB usually consist of glass fibres in an epoxy resin matrix with various microscale metal features, posing a significant challenge to separate them during recycling. The thermoset polymers used in PCB do not melt even with the application of extreme heat. Further, brominated flame retardants (BFR) in resins, make combustion difficult and such disposal can generate toxic gases. Therefore, PCB pose a great recycling issue (Hino et al., 2009) so some form of reuse is preferable for discarded personal computers, etc.
Nevertheless, the metal content is very high (about 70% weight, as reported in www.wrap.org) and this makes PCB wastes a rich metal source. Furthermore, after the RoHS directive in 2006 (Directive on the restrict, 2006) the use of lead (Pb) in solder has decreased. If cadmium and lead contents exceed the limits allowed under legislations, the PCB can be classified as hazardous waste as they can potentially cause significant contamination of soil, groundwater and surface water by heavy metals (lead and cadmium) if the boards and components are not correctly disposed of (Bizzo et al., 2014;Ron and Penev, 1995;Xiang et al., 2007;Lee et al., 2004). Locking the non-volatile/non-leachable wastes in (clearly labelled) building products is preferable to disposable in moist landfill where the toxic material could escape to the ecosystem.
The present study is considering a lead-free PCB sample with high content of iron oxide. The metallic parts of PCB were manually separated from the plastic component and were ground using a SPEX mill, 8000M series. The milling time was 4 h, using a rotation speed of 875 cycles/min. Metal-plastic powders at a size ranging between 100 and 200 mm were obtained. Fig. 2 shows the different processing steps.  For the Life cycle Inventory Analysis, the following data were collected.
-Milling: the electricity used for grinding was calculated.
-Metal-particles obtained: (2.7 kg of metallic powder and 0.38 kg of other waste).
The metal distribution of the recovered powder was analysed using a Scanning Electron Microscope (SEM) at magnitude of 5000Â coupled with an X-ray diffraction spectrum and reported in Fig. 3. An elemental analysis was also carried out by means of an Xray fluorescence spectrometer (WDXRF S8 Tiger Bruker) to check the metallic distribution in the PCB powder. Results are reported in Table 1.

Polymer recovered from MSW
The recycling process of HDPE (performed by SC All Green SRL (Romania), is schematically illustrated in Fig. 4. The electricity consumption data, required for the inventory analysis, is in Table 2.
For the LCA, the functional unit (f.u.) chosen was the mass of mixed plastics from MSW processed during 1 batch (350 kg) for the duration of 1 h. All the required inventory data was collected according to the chosen f. u. and listed in Table 3.

Metal filler modified HDPE
The recycled HDPE obtained from MSW was used as the matrix for flat polymeric sheets 'doped' with 10 w/o of metal micro-fillers recovered from the PCB wastes. The chosen filler fraction was adequate to render HDPE into a CPC. Fig. 5 illustrates the final metal-particle modified HDPE panel obtained through injection moulding, after extrusion.
The conductivity, dielectric features and the percolation phase are demonstrated by broadband dielectric analysis by means of Novocontrol BDS 1200 broadband dielectric spectrometer. Samples in the form of circular pills (diameter 25 mm, thickness between 2 and 4 mm) were monitored in the frequency range between 1 Hz and 3 GHz, at temperatures between 25 C and 120 C. Fig. 6 reports the dielectric properties for the recycled HDPE with 10 w/o of metal micro-fillers. The dielectric permittivity increases with the increase of the temperature due to the presence of the metal particle content as for pure HDPE this behaviour does not occur.
The dielectric analysis will the subject of a future paper.
3.3.1. Life cycle inventory for the metal-particle modified HDPE panel production Primary data were collected at laboratory scale. The functional unit was one 700 mm square by 2.5 mm thick panel with materials data reported in Table 4. 3.4. The suggested EMI shields 3.4.1. EMI shielding effectiveness Shielding effectiveness (SE) measures the capability for the material to attenuate the intensity of EM waves (Lakshmi et al., 2009). The EMI shielding mechanism depends on the absorption of the EM energy, the material surface reflection and the multiple internal reflections of the EM radiation. For adequate shielding behaviour, the shielding material must have electrical conductivity in the range 10 À 4 to 10 1 S/m (Gupta and Choudhary, 2011;Kulkarni et al., 2017).
In this work, two sandwich material assemblies were produced and tested: Shield type 1, illustrated in Fig. 7(aed), consists of the metal-particle modified HDPE sheet (thickness 2.5 mm), glass wool (thickness 95 mm) as thermal insulating material, adhesive and an aluminium foil (thickness: 20 mm) to enhance the EMI shielding ability. Shield type 2 consists of the same metal-particle modified HDPE sheet (thickness 2.5 mm), a polystyrene sheet (thickness 50 mm) as thermal insulating material, adhesive and aluminium foil (thickness: 20 mm) to enhance the EMI shielding ability.
In all cases, the metallic particles provide the EMI shielding ability due to the increased electrical conductivity of the studied sandwich components.
The EMI shielding effectiveness of the developed sandwich panels were studied by measuring signal attenuation (decibel, dB) in an anechoic chamber employing an antenna Horn 3115, at a frequency of 10 GHz, according to the standard EN 50147e1.
The optimum attenuation value (À48.3 dB) for the Shield type 1, was obtained when the HDPE surface is oriented towards emission. For the Shield type 2, the strongest attenuation value (À42.7 dB) was recorded with the HDPE surface oriented towards reception. The apparatus used is shown in Fig. 8. In both cases, there is additional attenuation within the panel to complement reflection from the aluminium foil.  Decontamination container with screw 5.5 7 Friction Device 7.5 8 Washing tank (I) 3 9 Washing tank (II) 3 10 Dewatering by centrifugation 15 11 Drying device (I) 3 12 Drying device (II) 3 Total electricity consumption 79.7 Table 3 Life Cycle Inventory data of the process to recover HDPE from MSW at the plant.

Item Quantity Assumptions
Waste plastic mix 350 kg (250 kg of HDPE) All the plastic is recovered. Only recovery of HDPE is considered with 100% allocation Electricity consumption 79,7 kW Calculated for the processing duration of 1 h Water consumption 12,045 m 3 The water is destined to wastewater treatment after use Non-recovered mat 5 kg Waste production Transportation <40 km The waste collection from the municipal area The obtained SE values are within the range of EMI shielding performance reported in the literature. Singh et al., 2014 (Singh andKulkarni, 2014) reported EMI shielding performance of 0.1, 0.5, 1.0, 5 and 10 wt% transition metal oxide (TMO) filled PVA composites prepared by the solvent casting method. Measurements were conducted in the 4e12 GHz (C band<8 GHz < X band) frequency regimes with minimum reflectivities reported in Table 5.

Life cycle inventory of EMI shielding systems
The Inventory data of the two studied EMI shields are presented in Table 6.

Results
The first LCA run carried was the calculation of the impact during the recycling of mixed plastics derived from MSW. In the network of Fig. 9 a schematic output depicts the quantities of materials and electricity required to obtain 1 kg of recycled HDPE. The numbers, expressed as percentage (%), are related to the GWP (Global Warming Potential) impact for each step. The box marked with green lines indicates the avoided impact due to the reuse of HDPE, meaning that most of the impact is compensated by the recovery and re-use of HDPE. Fig. 10 is a flowchart containing both impact types, generated and avoided, during the recycling process. The created impacts (values above zero, in the upper part of the flowchart) are mainly associated to the electricity consumption during the recycling process (light-blue bars) The avoided impacts (values below zero, in the bottom part of the flowchart) are associated to the materials recovery (HDPE).
In order to calculate the reduction of the impacts due to the recycling process of HDPE a comparative assessment shown in Fig. 11 compares recycled and virgin HDPE (obtained from the Ecoinvent database). Single point damages were reported according to the ReciPe endpoint method, where the strongest damage reduction is associated to the resources section: the use of recycled HDPE eliminates the use of virgin raw material.
The network of Fig. 12 is a schematic representation of the impact associated to the production of one metal-modified HDPE sheet obtained by using recycled HDPE and metallic powder. GWP contributions, for each step involved, are expressed in percentage (%). The avoided impacts (green boxes) are evaluated for the    Fig. 13 is the network of the processes involved in the production of the EMI shield type 1 showing the GWP contributions evaluated through CML 2000. The major contributions are associated to the metal-HDPE panel (32.4%) and to glass wool met (62%). Fig. 14 is the network of the processes involved in the production of the EMI shield type 2. The major contributions are associated to the metal-HDPE panel (47.7%), the polystyrene foam (32.2%) and steel (9.51%). Fig. 15 reports a comparative analysis of EMI shields type 1 and type 2 by means of the CML2000 characterization method. All impact categories result higher for type 1. The quantitative evaluation of GWP reported in Table 7 shows that 6,43 kg CO2 eq are released in the case of type 2 while 9,48 kg CO2 eq are involved for type 1. The major contribution to this result was the use of glass wool mat (5,88 kg CO2) in type 1. Glass wool mat was used in type 1 in order to provide thermal insulation properties to the structure whilst in type 2 polystyrene was used which contributed for 2.07 kg CO 2 eq. However, type 1 performed better in EMI shielding, due to the presence of aluminium foil in the sandwich structure (Table 8).

Uncertainty
LCA results deal with uncertainty. Quantity uncertainty is present for any data used in the inventory and the impact assessment. In this study characterization factors (CFs) from LCIA methods that are currently provided without uncertainty ranges, are reported. Uncertainty is also induced by the methodological choices e.g.: related to the goal and scope of the study, such as the definition of the functional unit, cut-off rule, allocation rule. In this study no allocation was required and the cut-off rule of 1% was applied. Finally, in this comparative LCA the compared systems use the same background processes and therefore a common propagation of uncertainty from the background is expected.

Discussion
According to the LCA results, it seems that the EMI shielding sandwich under study needs to be optimised in order to perform adequately as EMI shields with the least environmental impact. One way to achieve this could be the reduction of the amount of   glass wool in Shield type 1. A hypothetical evaluation is reported in Fig. 16 where a new Sandwich panel (type 3) containing half amount of glass wool is analysed. As can be seen, all kinds of impact are significantly eliminated. EMI shielding properties remain unaffected by the reduction of glass wool, however thermal insulation may be influenced. Table 8 taken from literature, reports interesting data for several materials used as insulators for building. Rock wool (assumed to  behave similarly to glass wool) has a low thermal conductivity and very low primary energy demand and GWP compared to expanded polystyrene and polyurethane. This means that glass wool in Shield 1 was a good choice. However, the glass wool generated a significant impact because of its increased thickness. The insulating material thickness is a critical parameter to provide thermal resistance in buildings. Table 9 reports the thermal resistance evaluated for the Shield containing polystyrene and the Shield containing glass wool. As the two insulating materials possess similar thermal conductivity, they are expected to provide similar thermal resistance when the thickness is similar. Since in Shield type 1 the thickness of glass wool is approximately double the thickness of polystyrene of Shield type 2, a glass wool panel of half the thickness, i.e. 0.5 m, was tested. This new system (type 3) exhibited weaker thermal resistance than type 2, but the environmental impact was much lower as shown by Fig. 16.
The choice of the functional unit is crucial for LCA evaluation as the results may change considerably. In this paper, on one hand, Shield type 2 and 3 employ the same functional unit as they are both EMI shields with similar thermal resistance values i.e. ranging from 12.4 to 13.5 m 2 k/W. Thereby, the corresponding LCA outputs are comparable. On the other hand, for Shield type 1, the functional unit is defined as an EMI shield with thermal resistance equal to 22.5 m 2 k/W. In such case, LCA outcomes of Shield 1 are not comparable with the other two Shield cases. Obviously, excluding the thermal properties from the functional unit definition and referring solely to the EMI properties, all three Shield types will be comparable in terms of LCA.
One disadvantage of using glass wool instead of polystyrene might be the weight as the density of glass wool is higher and for some applications this can be a hurdle. Another solution of environmental-friendly insulating material could be cork instead of glass wool, but in this case, the total cost will increase. It is noteworthy that an earlier LCA proved that cork has very low environmental impact when employed as natural insulating material for building applications (La Rosa et al., 2014). Also, cork performs well in terms of fire resistance (Di Modica et al., 2015) but it is a very expensive material and the cost is obviously a factor that cannot be disregarded.
The case study reported in this paper can be considered an addition to the existing literature in this field. A recent review on LCA application in WEEE management from Ismail and Hanafiah, 2019) (Ismail and Hanafiah, 2019) reports, as consideration from the authors, that different studies evaluated different WEEE management strategies with different objectives along with different research subjects, even for similar research scope. For an example, under the research scope that evaluated final disposal strategies for WEEE (i.e., evaluation of recycling, recovery and disposal strategies for WEEE), the management strategies discussed were unique from one study to another, and this make it difficult to compare, even if similar research subject was selected for comparison (Ismail and Hanafiah, 2019). This is still a limit of the LCA method.

Conclusions
This study investigated the development of electromagnetic shielding and thermal insulating components for building applications, employing recovered polymers and metal fillers from municipal solid wastes and printed circuit boards, respectively. Two different components were studied in terms of shielding effectiveness (primarily) and thermal insulation (secondarily), built into a sandwich assembly: Sandwich type 1 and Sandwich type 2. The two component types were evaluated in terms of electromagnetic shielding effectiveness and thermal insulating ability whereas the involved materials and processes were evaluated using life cycle assessment.
The sandwich components feature a polymeric sheet made of recycled HDPE, sourced from MSW and modified (doped) with

Table 9
Thermal resistance for the insulating materials used in the structure of EMI shields. 10w/o metal particles recovered from lead-free waste printed circuit boards. The metal particles are fundamental to the shielding effectiveness of the sandwich components. Sandwich type 1, containing glass wool, performed better in terms of electromagnetic attenuation and was the preferred choice. However, the LCA comparison revealed a high environmental impact due to the high thickness of the glass wool layer. Since the EMI shielding is not influenced by the thickness of the glass wool, a third system (type 3) fabricated with half the thickness of glass wool was evaluated with LCA. The thermal conductivity for Sandwich type 3 was similar to that of Sandwich type 2, however the overall environmental impacts were significantly reduced.
This study remarks the importance of the choice of the correct functional unit in LCA applications as the results may change considerably.
The paper we present here does not deal with methodological innovation, as we use the standard LCA methodology applied to a case study. Nevertheless, this case study can be considered an example of eco-design useful in the phase of material selection for the manufacture of new products. Data collection for the inventory analysis was included in the experimental section (section 3) in order to remind that the environmental assessment of a product should be used as a characterization method together with the other characterization methods (e.g.: physic-chemical, mechanical, thermal, electrical etc.) to allow a full understanding of the materials under study.
Finally, this paper represents an interesting contribution to the growing literature on WEEE recyclable and reuse processes as increasing the variations of LCA studies provides increasing support to quantify and assess environmental impact of various product systems within the complex WEEE management.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.