Waste Management 84 (2019) 91–101

Contents lists available at ScienceDirect

Waste Management
journal homepage: www.elsevier.com/locate/wasman

End-of-life of silicon PV panels: A sustainable materials recovery process

Valeria Fiandra ⇑, Lucio Sannino, Concetta Andreozzi, Giorgio Graditi
ENEA, Italian National Agency for New Technologies, Energy and Sustainable Economic Development – Research Centre ENEA Portici, Naples, Italy

a r t i c l e i n f o a b s t r a c t

Article history: In this paper, the management of end-of-life PV modules based on an advanced eco-sustainable process
Received 14 March 2018 has been presented and discussed. The thermal removal of the polymeric compounds contained in c-Si PV
Revised 14 November 2018 modules has been investigated to separate and recover Si, Ag, Cu, Al and glass. A two-step thermal pro-
Accepted 19 November 2018
cess has been employed. In the first step, the rear polymeric layer has been removed without emissions of
dangerous fluorinated substances. In the second step, the remaining polymers have been completely
removed with low volatile organic compounds (VOCs) emissions. The polymers degradation has been
Keywords:
studied at combustion equivalent ratios U varying from 0.5 to 2 and at 500 °C.
Photovoltaic module
Recycling
The materials recovery has been evaluated from an environmental point of view and optimized by con-
Polymer removal sidering the energy cost, through the identification of the best operating conditions, in terms of temper-
Thermal treatment ature, time, atmosphere and gas flow. One hour of heat treatment and a slightly oxidizing atmosphere
Energy consumption have been enabled to separate and recover the different materials of the module. The elemental compo-
sitions of the PV sample and the residue condensed organic products have been determined. The gaseous
degradation products have been characterized by gas chromatographic analysis (GC).
Ó 2018 Elsevier Ltd. All rights reserved.

1. Introduction germanium) and conventional materials (e.g., aluminium and

glass) and important environmental issues deriving from the
1.1. PV waste management release of hazardous substances (e.g., lead and cadmium)
(D’Adamo et al., 2017). Instead, a correct end-of-life management
E-waste is under global attention due to its potential environ- of the PV modules supports the development of circular economies
mental pollution and human health risk besides its content of valu- assuring the availability of secondary materials by a cost efficient
able resources (e.g., metals, plastics). In order to limit the negative recovery of available resources (Malandrino et al., 2017). Nowa-
impact of the relentless growth in PV waste volume and to enforce days, the affirmation of technologies for the green energy produc-
solar module recycling, the European Union (EU) has recently tion cannot be separated from measures and assessments of the
included the PV waste into the new Waste of Electrical and Elec- economic, environmental and social impacts associated with their
tronic Equipment (WEEE) directive (Shin et al., 2017). In detail, entire life cycle.
the EU imposes legal obligations on PV manufacturers, distributors Shiue and Lin (2012) has discussed the environmental advan-
and sellers to ensure take-back and recycling of their waste prod- tages and financial benefits of the business companies implement-
ucts. Despite regulatory efforts, WEEE continues to be badly man- ing the PV waste recycling programs. He has stated that the
aged and the proper recycling rate is extremely low worldwide. adoption and implementation of an optimal recycling strategy
Illegal disposal routes in developing countries and environmental not only reduce environmental pollution but also increase the
pollution issues are constantly reported (Bakhiyi et al., 2018). financial profits for the solar energy industry (Shiue and Lin, 2012).
Exploring the issues of the main technologies for the materials Huang et al. (2017) has estimated the costs of a recycling pro-
recovery from PV waste, intertwined gaps, issues and complexities cess of the valuable materials deriving from silicon PV modules.
are often revealed. The total revenue of a typical 60-cell Si module can generate

An incorrect management of the PV waste involves the loss of $16–17. He has stated that this revenue is more than enough to
scarce and precious metals (e.g., silver, copper, gallium, indium, cover the recycling cost and keep up a profitable recycling business
without any government support (Huang et al., 2017).
However, the lack of tested, implemented and validated tech-
⇑ Corresponding author. nologies on a large scale for the disposal of the various types of
E-mail address: [email protected] (V. Fiandra). PV modules at end-of-life and the recovery of valuable materials

https://doi.org/10.1016/j.wasman.2018.11.035
0956-053X/Ó 2018 Elsevier Ltd. All rights reserved.
92 V. Fiandra et al. / Waste Management 84 (2019) 91–101

constitutes a deep limit to the realization of a sustainable manage- removed subsequently from the cells surface by chemical etching
ment of the photovoltaic supply chain. for crystalline silicon regeneration (Wambach, 2003). However,
The current PV waste treatment technologies should be signifi- this approach presents some drawbacks. Its low automation degree
cantly improved with the aim to reduce/avoid the production of and the necessity to manage the significant environmental impact
toxic substances downstream of the material recovery process. make this process much expensive.
Furthermore, new regulations concerning the hazardous by-
products deriving from recycling processes should be also intro- 1.3. State of art
duced (Zeng et al., 2017).
The scientific knowledge gaps concern the extent of the con- The removal of the encapsulating polymeric materials is usually
tamination resulting from waste of the recovery processes. In par- the first step for the materials recovery from end-of-life PV mod-
ticular, a complete characterization of the quantity and quality of ules. The most commonly used encapsulating polymer is EVA
the by-products of the recovery processes and an assessment of and represents the major problem for the separation and subse-
their potential environmental damage are lacking. quent recovery of the different materials (Tao and Yu, 2015;
The goal of this work is to highlight some of the problems Klugmann-Radziemska and Ostrowski, 2010; Dias et al., 2017).
occurring in the conventional PV waste management practice by Different methods have been proposed to remove EVA. Chemi-
also suggesting possible solutions. The development of an cal approaches are based on dissolution by acids or solvents (Doi
advanced material recovery process for industrial application, is et al., 2001; Kang et al., 2012; Kim and Lee, 2012). Physical meth-
proposed to improve the energy balance and environmental per- ods are based on decomposition by pyrolysis (Frisson et al., 2000;
formance of the entire life cycle of a silicon PV module. Dias et al., 2016b; Dias et al., 2017). It is unlikely that the chemical
treatment becomes a viable industrial process due to large
1.2. End-of-life of PV modules amounts of needed chemicals. The chemical waste disposal and
the toxic gases treatment increase the material recovery complex-
In 2016, the PV market continued its global expansion with a ity and the energy involved as well as the financial cost signifi-
50% growth, bringing the market to at least 75 GW. The total cantly (Sica et al., 2018; Kim and Lee, 2012). The thermal
installed capacity at the end of 2016 globally amounted to at least approach seems to be more favourable than the chemical one even
303 GW (Jäger-Waldau, 2017; Ludin et al., 2018). With the growing if it is not without problems from both economic and environmen-
number of PV installations, the proper end-of life management of tal points of view (Wang and Xu, 2014; Strachala et al., 2017).
modules is becoming increasingly urgent since offers a sustainable There is no unique method for conducting the separation and
solution in terms of resource availability, economic feasibility and recovery treatment since it depends on the type of module to be
environmental risks (Sener and Fthenakis, 2014; Vellini et al., processed (Tao and Yu, 2015; Duflou et al., 2018). According to dif-
2017). Furthermore, the PV waste contains numerous precious ferent PV module’s production technologies, the thermal process
and scarce metals such as: Cu, Ag, Si, Se, In, Ga, Te (Bigum et al., control depends heavily on the type of design, the size of laminate
2012; Klugmann-Radziemska and Ostrowski, 2010; Rocchetti and and the module dimensions. These factors influence the efficiency
Beolchini, 2015). Recovering these metals would increase the eco- of the process and the type and quantity of gaseous emissions dur-
nomic and environmental sustainability of PV panels recycling ing the thermal treatment (Bombach et al., 2006).
(De-wen et al., 2004; Xu et al., 2018). This work proposes an advanced thermal treatment to reduce
Most of the PV waste derive from silicon based PV modules at the problems of environmental sustainability and costs of the con-
the end of their life. These waste represent the most urgent prob- ventional thermal processes carried out on silicon based PV mod-
lem which needs the proper management in the current context of ules at end-of-life. It investigates the thermal treatment
the e-waste. The main issues are the lack of a proper material operating conditions, in terms of temperature, gas atmosphere
recovery technology. and duration, and the thermal treatment by-products in the oper-
Furthermore, if the PV waste is not disposed correctly, haz- ating conditions chosen.
ardous metals (Pb, Cr, Cd, Ni) or toxic and carcinogens substances It originates from the attempt to respond to the main critical
risk to be released in the environment (Widmer et al., 2005; issues emerging from the analysis of the literature on this topic.
Marwede et al., 2013). The procedure presented from Klugmann-Radziemska et al.
The silicon based PV modules are the most common type of (2010) to recover the silicon wafers and to use them in the produc-
solar cell manufactured in the world. These modules consist of a tion of new solar cells is mainly a two-step process based on cells
glass layer, aluminum frame, encapsulating layer of transparent separation and cells refinement. The separation stage is a thermal
EVA (ethylene-vinyl acetate) copolymer, photovoltaic cells, junc- delamination used to remove the EVA resin and separate the other
tion box, TedlarÒ protective foil as back-sheet. The solar cell is layers, the refinement step is performed by a chemical method.
composed of Si wafer, Ag electrode on the front side and Al elec- However, the authors optimized and examined only the environ-
trode on the rear side. The solar cells are connected with lead mental aspects of the chemical step, whereas factors involved in
coated copper wires. the thermal operation (e.g. temperature, emissions, etc.) were not
The valuable elements contained in silicon panels are silicon taken into account (Klugmann-Radziemska et al., 2010).
(Si), aluminum (Al), silver (Ag), copper (Cu) and glass. At present, Wang et al. (2012) have proposed a two-step heating process,
the Al frames may be separated mechanically from the solar pan- followed by a chemical treatment, to recover the glass plate with-
els, the glass may be recovered by thermal or mechanical treat- out breaking and recycle silicon and Cu. In the first step, TedlarÒ is
ments, while Si, Cu, Ag by thermal and chemical methods (Jung separated from the PV module back surface heating at 330 °C for
et al., 2016; Dias et al., 2016a; Zhang and Xu, 2016). 30 min. Then the second heating step is carried out to burn EVA
The only one currently industrialized process to recover PV and residual TedlarÒ at 400 °C for 120 min (Wang et al., 2012). At
materials was developed by Deutsche Solar company (Solar 330 °C EVA is not decomposed and therefore during the thermal
World). It provides for the preliminary manual dismantling of sili- treatment at 400 °C, the glass plate will not breaking due to the
con modules and subsequently two main steps, a thermal and a deforming of EVA and TedlarÒ. However, other authors have stated
chemical one. First the polymeric laminate is burned off at 600 °C that thermal degradation of TedlarÒ film already starts at 204 °C
to facilitate the materials separation and recover silicon cells, then (Drobny, 2000; Teng, 2012). Therefore, at 330 °C TedlarÒ decom-
the metallization, antireflective coating and p/n-junction are poses to form dangerous fluorinated by-products requiring special
 V. Fiandra et al. / Waste Management 84 (2019) 91–101 93

abatement systems (Hagiwara et al., 1977; Farneth et al., 1993). In this way, the process is addressed by taking in consideration
Thus, in the process proposed by Wang et al. (2012), the environ- the yield of the materials recovery, the treatment efficiency and
mental issues are neglected. Moreover, at 400 °C, the EVA decom- the environmental pollutions emerging from it. The obtained
position is not complete (Girardin et al., 2015). Therefore, results can act as practical tips to aid the industrial practice of
recovered Silicon and Cu are contaminated with organic residues. the materials recovery process from the PV modules.
Dias et al. (2016b) has investigated the silicon PV module’s compo- Another original aspect dealt with the manuscript is on the
nents separation. He has studied the degradation of the polymeric thermal treatment selectively performed on the PV waste delami-
fraction by pyrolysis to detach the layers bonded by the adhesive nated from backsheet of TedlarÒ. Other processes reported in the
material. His study proved that at 500 °C, all polymeric fractions literature performed the thermal treatment for EVA degradation
present in the PV module are degraded. However, further studies using the whole mass of PV panel. The proposed approach is based
are recommended to evaluate the ideal dwell time for the pyroly- on the backsheet preliminary removal, with the consequently
sis. Furthermore, the characterization of the recovered materials reduction of harmful gaseous emissions and of the process costs
suggests that new studies should be conducted in order to improve concerning the abatement systems.
the separation methods (Dias et al., 2016b). In a following study,
the removal of the polymers, EVA, TedlarÒ and poly-ethylene 2. Materials and methods
terephthalate (PET), were performed by pyrolysis on PV modules
previously milled. It was demonstrated that the pyrolysis at The separation of the valuable materials deriving from PV pan-
500 °C, in 1 h can remove >99% of the polymeric matter. Tempera- els at the end of their life was achieved by means of a thermal
tures above 500 °C continue to degrade the organic matter, but the treatment carried out with an experimental apparatus developed
mass loss rate is 6.25 times lower. However, the optimal pyrolysis in the laboratory for batch tests on samples of about 110 g of
conditions have not been identified for PV panels treated as such weight. Samples taken from a typical PV panel with polycrystalline
(Dias et al., 2017). silicon wafers were subjected to thermal tests. Elemental and gas
The pyrolysis behaviour of EVA under the partial oxidizing chromatographic analysis were performed to determine the degra-
atmosphere was also investigated (De-wen et al., 2004). Pyrolysis dation products and evaluate the environmental sustainability of
reactions can be accelerated by enriching the atmosphere of oxy- the treatment.
gen but in this case, an exothermal peak occurs at about 450 °C.
The thermal stress associated with the rapid increase in tempera- 2.1. Case study: waste from silicon based PV panels
ture, causes the Si wafer damage. The authors have stated that this
can be avoided by performing the EVA thermal degradation with In the present paper, a PV module with polycrystalline-silicon
an oxygen content lower than 10 vol%. Furthermore, an abundant wafers made by Pragma was investigated. The panels under study
amount of coke is formed. Pyrolysis gas amount and composition, were part of a PV system called Delphos (Demonstrative Electrical
condensate and residual coke are influenced by the O2 content. Photovoltaic System) designed and realized by ENEA (Italian
However, De-wen’s experimental determinations concern pure National Agency for New Technologies, Energy and Sustainable
EVA samples and not PV panels. Further investigations are there- Economic Development). This plant was installed in Manfredonia,
fore necessary on more representative samples of PV waste. in the South of Italy, and was operating since August 1986 for
All the processes for the PV materials separation and recovery around 15 years. The plant had a total installed capacity of
are currently very expensive, are not completely efficient in terms 600 kWp. The Table 1 shows the main features of the PV plant
of recovery yield and have a significant environmental impact. and used modules.
Many aspects of the thermal treatment to degrade the polymeric The structure of the module includes the TedlarÒ as backside
matter, contained inside the PV panel structure, have to be foil, the glass on the upper side and an aluminium frame. The sili-
improved. con cells are embedded front and back in EVA. The modules con-
In this work, some aspects of the thermal process for recovery of tain cells with a typical backside metallization of aluminium.
silicon, glass and metals from silicon PV panels were investigated. Before experimentation, the modules were inspected and vari-
A sustainable treatment for PV panels at end-of-life has been pro- ous types of defects were found: glass breakage, damaged cells,
posed in order to reduce environmental impact and energy costs of laminate defects and electrical insulation defects.
the conventional recycling process. In Fig. 1, the PV module under investigation and a typical sam-
The information in literatures about the EVA degradation as a ple withdrawn from the module are shown.
result of a combustion and the necessary data for the furnace
design, such as required gas volume, influence of the heating
Table 1
regimes, end temperature, time of dwell at the end temperature Technical data of the c-Si PV module.
and oxidizing atmosphere, are not complete. Therefore, this paper
aims to contribute to the understanding of the degradation beha- PV module’s technical data

viour of EVA under partially oxidizing atmosphere. Item Description

The main goal of the work is to identify the optimal operating Year of construction 1986
conditions of the thermal treatment. For this purpose, the fractions Installed power in KWp 300
obtained from thermal treatment have been characterized in terms Total Silicon area 3917 m2
Covered ground 3880 m2
of carbon content. The determination of the carbon content in the
Total number of modules 5760
three classes of combustion products: gas, condensate and solid Module mass 16.8 Kg
residue, allows to carry out a carbon material balance. The distri- Module size 130  68 cm
bution of carbon in the three phases allows to verify the evolution Frame Aluminium alloy
and efficiency of the thermal treatment in regard to the chosen Front glass Tempering glass 3.2 mm
EVA sheet thickness 0.5 mm
operating conditions. Wafer thickness 200 mm ± 20 mm
A further and not secondary target of the work is to propose a Number of cells per module 72
new advanced eco-sustainable treatment that can be achieved Cells material Polycrystalline silicon
through the elimination of the highly polluting substances deriving Cells size 10  10 cm
Year of shutdown 2001
from the degradation of TedlarÒ.
94 V. Fiandra et al. / Waste Management 84 (2019) 91–101

Fig. 1. c-Si PV module at end-of-life: front (a) and back surface (b). Sample taken from the PV module and used in the experimentation: front (c) and back surface (d).

The samples were taken from the PV module after manual dis- pling bag was filled using the V three-way valve. It is equipped
mantling of external Al frame. Each sample was obtained by cut- with a tap with built-in porous septum that allows filling, empty-
ting a piece of about 10  10 cm by using a diamond blade for ing and picking of the gaseous mixture by means of a chromato-
glass incision and then for panel cutting. Each sample weighed graphic syringe. Once the bag is filled, the gaseous stream is fed
110 g approximately. to a hood. During the heat treatment, the little sampling bag col-
The specification of the relative mass fractions for the compo- lects some gas samples for the GC analysis in continuous.
nent of a typical standard c-Si PV module, and therefore of the
module under test, has detailed in Table 2 (Sander et al. 2007). 2.3. Pre-treatment of delamination

As well-known in literature, the main products generated by

2.2. Experimental apparatus
combustion of TedlarÒ are the following: HF, CO, CO2, volatile
hydrocarbons, fluorinated organic compounds, aromatic hydrocar-
The thermal treatment experiments were performed in a labo-
bons, particulate matter (Baker and Kasprzak, 1993; Garcìa et al.,
ratory apparatus as shown in Fig. 2. A high temperature Lenton
2007; Yamada et al., 2005; Stapler et al., 1968). These substances
tubular furnace was used as source of heat. The reactor for thermal
are harmful to human health and the environment, accordingly
decomposition of EVA was made of a quartz tube with an inner
their elimination from the emissions associated with the thermal
diameter of 117 mm and length of 900 mm. The tubular reactor
treatment is particularly beneficial. To this end, the PV waste sam-
has a constant temperature zone of 370 mm in the furnace. The
ples were placed in the quartz sample holder, inserted into the
flanges at the ends of the quartz tube have an inlet for gas and
tubular furnace and brought to the temperature of 200 °C, with a
an outlet for the decomposition products contained in the exhaust
heating rate of 450 °C/h, in the absence of gas flow. Samples were
gas. Both the flanges are cooled by water network. A pro-
kept at 200 °C for 30 min and then cooled to room temperature.
grammable temperature controller provided the process tempera-
This gentle heating allows the rear polymeric layer of TedlarÒ
ture control. A type ‘‘K” thermocouple affixed near the quartz
and EVA to be easily removed from the starting PV sample by sub-
sample holder boat inside the tube was used to precisely monitor
sequent manual peeling off.
the temperature. The tubular reactor was connected with two
Drechsel bottles through flexible tubes to condense the high-
2.4. Thermal treatment
boiling products. The pressure in the tube was monitored by a
pressure gauge.
The delaminated samples were weighed and placed in the mid-
The experiments of thermal degradation were carried out at the
dle of the tubular reactor in a sample holder boat.
temperature of 500 °C, in atmosphere that differs in the concentra-
Two flowmeters allow to manage the gas flow rates to be sup-
tion of oxygen and nitrogen. During the normal operation of the
plied to the furnace and to produce nitrogen/oxygen mixtures in
plant, the gas stream is sent to the S collection bag, to collect up
different ratios. The supply gas, air, was flowed inside the furnace
to 80 Lt of gas at room temperature. The non-condensable volatiles
and the programmed temperature was started.
were collected in the gas sampling bag. By using the valve of the
The gas was supplied at a flow rate of 24 L/h, then the reactor
sampling bag, gaseous samples were extracted during the treat-
was heated up to the process temperature (500 °C) at the heating
ment time to check the combustion performance by qualitative
rate of 450 °C/h, and finally held at this temperature for 1 h. The
and quantitative analysis of the produced substances. The S sam-
performance of the treatment was followed in time by monitoring
the carbonaceous degradation products, solid, liquid and gaseous
Table 2 deriving from the combustion of the sample, by gas-
Main constituents amounts of a typical stan- chromatographic, gravimetric and elemental analysis. The matter
dard c-Si PV module (215Wp).
balance on the carbon has allowed to fix the required treatment
Materials Amount (wt %) time in 1 h.
Glass 74.16 The degradation products were dragged by the gas flow from
Aluminium frame 10.30 reactor to the collection traps by cooling in the two Dewar vessels
Silicon wafer 3.48 with a water-ice bath. The cooling produces the separation of high
Encapsulating (EVA) 6.53 boiling substances, such as polyoleines or other partial combustion
Backsheet (TedlarÒ) 3.60
Adhesives 1.16
products of high molecular weight, driven by the gaseous stream.
Cu 0.57 During the heat treatment, the degradation gas products with a
Ag 0.01 high boiling point had condensed in the Drechsel bottles and in
Sn 0.12 the flexible tubes. The weight of the condensed products was
Pb 0.07
obtained by weighting the Drechsel bottles and the flexible tubes
 V. Fiandra et al. / Waste Management 84 (2019) 91–101 95

Fig. 2. Experimental setup for thermal treatment and sampling. V: 3-way valve, F: Tubular furnace, FI: Flow indicator, T: Trap, FC: Flow controller, S: Sampling bag, P:
Pressure indicator, TC: temperature controller.

before and after the experiments. At the end of the heat treatment, Fisher Scientific Trace 1300 with a TCD and a FID was used. Each
the collection bags of the emitted gas during combustion was detector has a limited quantitative working range, which remains
detached and the gas was analyzed. under control when the peak height is measured in standard units
like volt or ampere. Thus the peak height was taken as alarm value
for the measurement error.
2.5. Analysis For TCD, the minimum detection limit (MDL) is less than 400 pg
tridecane/mL with He carrier or less than 20 pg tridecane/s with a
The elemental composition of the organic matrix in the PV sam- total flow through the cell of 3 mL/min. For FID, MDL is less than
ples waste subjected to thermal treatment were analyzed with ele- 1.4 pg C/s. The guaranteed percentage values of relative standard
mental analyzer Elementar Inductar EL Cube, equipped with a deviation (% RSD) are 2% for TCD and 1% for FID.
solid-state high-frequency induction furnace for both CS and A capillary packed column TG Bond Q
ONH analysis. The high temperature CS/ONH analyzer allowed to (15 m  0.53 mm  20 mm) was used to analyze CO2. The temper-
analyze all the five elements. C, S, O were analyzed with a wide- ature ramp choose in the used chromatographic method was been
range infrared detector. For the precise determination of H an elec- from 50 °C to 200 °C at a rate of 50 °C/minute. The selected detec-
tronic hydrogen sensor was used. Depending on the selected func- tor for this chromatographic analysis was been the TCD with a
tionality, the analyzed sample was either combusted in Oxygen at temperature of 210 °C. The split flow used was been 68 mL/min.
2000 °C (for CS analysis) or fused in Helium or Argon at 3000 °C A CP-Carboplot P7 column (28 m  0.53 mm  25 mm) was used
(for ONH analysis). Subsequently the resulting gases were ana- to analyze CO. The used chromatographic method was been an
lyzed. The samples were weighed (
0.3 g) into graphite sample isothermal analysis at a temperature of 30 °C. The selected detector
containers (CS) or ceramic sample containers (ONH). The measure- for this chromatographic analysis was been the TCD with a temper-
ment accuracy (RSD %) for each element is: C = 0.3, S = 0.2, H = 0.1, ature of 80 °C. The used split flow was been 20 mL/min.
O = 0.5, N = 0.2. The separation of the hydrocarbons C1, C2, C3 and C4 was car-
In Table 3, the elemental compositions of the PV sample waste, ried out with a TG-Bond Q column (15 m  0.53 mm  20 lm). The
of the EVA layer and polymeric backsheet respectively are shown. used temperature ramp was been from 50 °C to 200 °C at a rate of
All samples were analyzed 3 times. The CHOS average content and 10 °C/minute. The selected detector for this chromatographic anal-
its absolute standard deviation are shown in Table 3. ysis was been the FID with a temperature of 250 °C. The used split
The gas products were totally collected in the gas sampling bag. flow was been 9 mL/min. The calibration of the species relative
Gas-chromatography was used to analyse the gas and determine profiles, in mole fraction, was performed by using a standard gas-
the mole fraction of the major stable species. The species identifi- eous mixture of hydrocarbons in helium, prepared and supplied by
cation and their relative concentrations were obtained by coupling Rivoira Pharma. The mixture volume composition was been:
GC with a classical thermal conductivity detector and a flame ion- dimethyl ether (109 ppm), acetylene (579 ppm), isobutene
ization detector (GC/TCD/FID). A gas-chromatograph Thermo (886 ppm), n-butane (8073 ppm), n-propane (0.263%), methane
(0.405%), hydrogen (0.42%), propene (0.532%), n-ethane (0.662%),
Table 3
ethane (0.658%), carbon monoxide (1.08%), carbon dioxide
CS/OH Elemental analysis on: PV waste sample object of study, EVA contained in the (3.11%), helium (to 100%).
structure and polymeric backsheet. The fluorine value was calculated by difference: The analysis of acetic acid was performed using acetic acid ana-
100%  (%C + %H + %O + %S). *Mean value (wt %) and absolute deviation (a.d.). lytical standard for GC, supplied by Sigma Aldrich.
Element * PV sample EVA Polymeric
(wt % ± a.d.) waste backsheet
C 15.83 ± 0.02 81.59 ± 0.08 75.72 ± 0.09
3. Results and discussion
H 2.41 ± 0.04 13.94 ± 0.05 13.20 ± 0.02
O 0.62 ± 0.09 4.47 ± 0.07 3.34 ± 0.06 In order to reduce the harmful emissions resulting from the
S 0.08 ± 0.007 n.d. n.d. combustion of TedlarÒ, representative samples of a c-Si PV panel
F – 7.74
at end-of-life, described in Table 1, were subjected to a preliminary
96 V. Fiandra et al. / Waste Management 84 (2019) 91–101

delamination of the backsheet. In this way, the rear layer, consist- combustion, in O2 depleted conditions. In fact, its greater or less
ing predominantly of the TedlarÒ fluorinated polymer, is elimi- concentration is also a function of the oxygen concentration in
nated before the thermal treatment which allows the separation the atmosphere inside the furnace. The different equivalence ratio
of the recoverable materials. The total removal of the backsheet, conditions were created by varying the mass of sample, in order to
which remains attached to the metal layer only on the edges of feed only air to the furnace. The air flow was kept constant at each
the sample, as shown in Fig. 3c), was carried out by manually peel- U ratio in order to ensure identical conditions in the furnace in all
ing off. Fig. 3 shows a sample submitted to delamination at 200 °C, the experiments.
the separation of the polymeric backsheet from the remaining The choice of the treatment temperature (500 °C) was made on
metal layer and some recovered rear metal strips. the basis of literature analysis. Published thermogravimetric stud-
The removal of the rear polymeric layer is a particularly advan- ies have shown that EVA is completely degraded at 500 °C (Miller
tageous preliminary step from an environmental point of view in et al., 2013). The goal of the second step of the recovery treatment
the separation and recovery process of valuable materials. In fact, is the complete degradation of the remaining EVA. It remains after
it allows the reduction of the total gas emissions which would be the preliminary step to delaminate the panel backsheet. Therefore,
produced by treating the entire panel. Moreover, harmful emis- it was considered more economically convenient to work at a
sions to human health and the environment of hydrofluoric acid lower temperature (500 °C) than temperature used in the conven-
and fluorinated organic compounds, are eliminated. During the tional thermal treatment (600 °C).
conventional thermal treatment at 600 °C, for the separation of A typical thermally treated sample has the appearance shown in
the recoverable PV materials, fluorinated organic compounds and Fig. 4, it is devoid of the organic matrix and separated into the con-
hydrogen fluoride are developed. Then, the decomposition gaseous stituent components.
products were sent to a gas-fueled afterburner, installed down- The separation of the different materials was carried out by
stream of the furnace. It is used to efficiently burn all the gaseous sieves. Silicon, glass and metal strips were recovered as shown in
organic by-products and transform they in carbon dioxide and Fig. 5.
water. However, the afterburners for the destruction and removal Table 4 shows the amount of the recovered materials after the
of the by-produced halogenated hydrocarbons are not effective. thermal treatment at 500 °C.
Therefore, these substances resulting from the partial degradation
of TedlarÒ, require appropriate management and solutions to meet
the environmental regulations in force. The proposed backsheet
delamination pretreatment can solve this problem, by considerably
reducing the environmental impact.
After the delamination of the backsheet, the samples were sub-
jected to a further thermal treatment to completely thermo-
degrade the residual organic matrix, predominantly consisting of
EVA, which holds the glass layer together with that of Silicon/
metal. The thermal treatment was performed using a stoichiomet-
ric amount of oxygen compared to the contained carbon in the
organic matrix. To calculate the fuel/air stoichiometric ratio, it
was assumed that all the remaining polymer was composed only
of EVA and that EVA decomposes, producing only water and CO2.
The combustion was studied at equivalent ratio U varying from
0.5 to 2.0, at 500 °C.
The combustion equivalent ratio U is defined as follows:

U ¼ ðfuel=airÞactual =ðfuel=airÞstoichiometric
Depending on the fuel/air ratio, the polymeric matrix degrada-
tion can lead to the higher formation of volatile hydrocarbon spe-
cies. Under conditions of incomplete oxidation in fact, organic
compounds can react to form VOCs, some of which are the basic
constituents for soot formation. The equivalent ratio U also influ- Fig. 4. Appearance of a PV panel sample previously delaminated and then subjected
ences the formation of CO in the gaseous products resulting from to thermal treatment at 500 °C for 1 h.

Fig. 3. PV sample as is before backsheet delamination (a), sample at the end of delamination treatment at T = 200 °C (b), recovered metal strips (c).
 V. Fiandra et al. / Waste Management 84 (2019) 91–101 97

Fig. 5. Recovered materials after thermal treatment of a c-Si PV panel sample: Silicon (a), glass (b) and metal strips (c).

Table 4 It is interesting to note in Fig. 6a) that the Carbon in the pro-
Recovered materials amount (weight %) after the two duced CO2 tends to have a low and almost constant value for
steps of thermal treatment. The Copper amount is
U > 1. Instead, for the values of U < 1, the Carbon in CO2 has higher
related to the recovered strips in the first step of the
backsheet delamination.
values than those obtained at U  1, downstream of the thermal
treatment at 500 °C. The higher Carbon values in CO2 mean greater
Recovered materials Amount (wt %)
combustion efficiency. At U  1, the low yield of CO2 means that
Glass 67.00 the polymer is only partially decomposed and is lost partly as vola-
Silicon 7.70 tile organic carbon and prevalently as viscous residue.
Metals 1.35
Aluminium Frame 12.13
On the other hand, for values of U < 1, the organic matter is
Copper 0.51 preferentially oxidised to CO2 rather than CO as shown in Fig. 6b)
Backsheet (Tedlar Ò / EVA) 11.31 by the amount of carbon in the produced CO. The amount of CO
produced is however, less than that of CO2 in every case evaluated.
The analysis of the elemental carbon content in the recovered For 1.5 < U < 2 the slight decrease in CO is probably due to the
materials was performed to evaluate the thermal treatment effi- increased production of VOC and particulate, which is instead more
ciency. Table 5 shows the carbon content in the recovered materi- moderate in oxidative conditions.
als after the thermal treatment at 500 °C. The analysis has shown The Carbon in the volatile organic compounds (VOCs), repre-
very low carbon concentrations in all the recovered materials. This sented in Fig. 6c), show a trend of always rising values in the whole
indicates the complete degradation of the polymer and justifies the range of U investigated. However, at the processing temperature of
choice of the used temperature and time of the process. 500 °C, the lower production of VOCs is observed for low values of
U.
3.1. Degradation gaseous products of the polymeric matrix contained
in the PV panel 3.2. Composition determination of the degradation products and mass
balance on carbon
The thermal degradation products of the organic compounds
contained in the c-Si-PV samples treated in this work, were classi- The gaseous compounds and the low volatile condensate are the
fied into three groups: gas products, condensed liquid products products in greater quantity formed by the thermal degradation of
and solid residues. The decomposition gaseous products are CO2, the organic matter contained in the PV panel structure.
CO and volatile organic compounds (VOCs) consisting mainly of To perform the carbon material balance, the elemental analysis
C1, C2, C3, C4 hydrocarbons. Acetic acid was also found among of the viscous condensate, formed after thermal treatment at
the products deriving from the EVA thermal decomposition. It 500 °C, was carried out. The carbon content in the solid residue
was in the gaseous state, with white fumes having typical pungent was calculated for difference between the initial value and the
smell. The GC-FID analysis has revealed its presence in the sampled carbon contained in CO2, CO, VOCs and viscous condensate. The
gaseous mixture. elemental composition of the condensed liquid products, col-
The partial degradation products consist of high-boiling hydro- lected after the thermal degradation at 500 °C, is shown in
carbons, collected by condensation, which represent a viscous liq- Table 6.
uid residue. The solid residues constitute the particulate and the The condensed products were in the form of wax at the room
ashes. temperature whereas they were in the form of liquid at tempera-
Fig. 6 shows the trends of the Carbon contained in the emitted ture higher than 40 °C. Probably the hydrocarbons making up these
main gases (CO2, CO and VOCs) during the thermal treatment at products are of paraffinic, olefinic and aromatic type. These prod-
500 °C, vs combustion equivalent ratio. ucts could be used as fuels for combustion systems in the industry
or could be burned to fulfil the heat requirement in the thermal
Table 5 process of material recovery.
Elemental analysis on recovered materials. Carbon The particulate was recovered and quantified by gravimetric
content (wt %) and absolute deviation (a.d.).
analysis through the use of a filter that was weighed before and
Recovered material Carbon (wt% ± a.d.) after the treatment.
Glass 0.00035 ± 0.00005 The solid residue was composed of ashes and particulate. Its
Silicon 0.00971 ± 0.00160 quantity, the average on three tests of thermal treatment at
Copper 0.00739 ± 0.00038 500 °C, was calculated with respect to the delaminated sample
Metals 0.00271 ± 0.00021
from backsheet and is shown in Table 7.
98 V. Fiandra et al. / Waste Management 84 (2019) 91–101

Fig. 6. Mass ratio between Carbon and polymer (g/g) vs U. Carbon contained in the produced gaseous species: (a) CO2, (b) CO and (c) VOCs, after thermal treatment at 500 °C.

Table 6 can easily be recovered and used for energy production in industry.
Elemental composition of the condensed degradation products after thermal treat- Probably, the drastic reduction of CO2 that is observed at
ment at 500 °C and U
1. *Mean value in weight % and absolute deviation a.d.
0.5 < U < 1 is related to the greater amount of viscous organic
Elemental analysis (wt % ± a.d.)* Liquid organic matter products that are formed as one moves away from oxidative condi-
C 84.25 ± 0.05 tions, which subtract carbon to the gaseous phase.
H 12.49 ± 0.03 Similarly to VOCs, the Carbon in particulate increases as U
O 3.26 ± 0.08 increases, as shown in Fig. 7b) and it is minimal in oxidizing con-
S 0.01 ± 0.008 ditions (U < 1). Probably, the carbon content in the particulate is
also due to heavy hydrocarbons dragged by the air flow. Fig. 8
shows the mass balance on the carbon and the average products
Table 7 distribution deriving from the thermal treatment of the PV sample
Quantitative composition (mean value
waste. The residue consists of particulate matter and ashes. The
in weight %) of the solid residue, after
thermal treatment at 500 °C and U
1. few residue constitute a favourable quantity for the next surface
*Delaminated sample. cleaning treatments of the recovered Silicon and glass after the
combustion.
Solid residue wt % (g/gsample*)
Particulate 0.15
Ash 0.14
3.3. Qualitative composition of gaseous emissions

It is known that the thermal degradation of EVA occurs in two

stages (Allen et al., 2000). The first stage, between 300 and
Fig. 7 shows the trends of the Carbon contained in the viscous 350 °C, corresponds to the elimination of acetic acid from the poly-
organic residue and in the particulate vs equivalent combustion mer chain leaving a highly unsaturated residue or polyene. The
ratio, during the heat treatment at 500 °C. second stage, carried out under a temperature running in the range
In Fig. 7a) an important representative aspect of the thermal from 400 to 500 °C, corresponds to the decomposition of the poly-
treatment at 500 °C is shown: the high amount of the Carbon in olefin chain resulting from the previous stage.
the viscous organic residue. Its value is between 0.4 and 0.5 g/g The main gaseous products CO2, CO and VOCs, constituted from
of polymer in the whole range of U investigated. Probably, these C1-Cn hydrocarbons, are the sum of the secondary degradation
condensed products could be eliminated or however greatly productions of the acetic acid and the products deriving from the
reduced, performing the treatment at higher temperatures. On second stage of the chain degradation at higher temperature
the other hand, in the used process conditions in this work, they (Gasnot et al., 2005, Hull et al., 2002).

Fig. 7. Mass ratio between Carbon and polymer (g/g) vs U. Carbon contained in the produced different residues: (a) viscous organic residue and (b) particulate (Carbon in
particulate is calculated by difference: [Cin  (CCO2 + CCO + CVOCS + Ccondensate)], obtained after thermal treatment at 500 °C.
 V. Fiandra et al. / Waste Management 84 (2019) 91–101 99

Fig. 8. Mass balance on Carbon after thermal treatment of c-Si PV panels a 500 °C and U
1. The carbon content in the solid residue was determined by difference:
[Cin  (Cgas + Ccondensate)].

According to literature, the distribution of degradation gaseous Operating in slightly oxidizing conditions, the volatile organic
products (vol %) from the PV waste thermal treatment, at 500 °C species decrease considerably and are oxidised to CO2.
and U
1, is given in Table 8. Under inert conditions, the polyene structure formed by the
The ratio by weight of the main volatile compounds deriving deacetylation, undergoes chain scission reactions producing com-
from the organic matrix combustion vs the treated delaminated pounds not reaching the full degradation.
PV panel are reported in Table 9. Instead, in an oxidative environment, the polyene probably
As can be seen from Table 8, Ethene (C2) is the volatile organic forms carbonaceous sub-products, which are oxidised into CO2 at
compound generated in larger quantities, followed by Propene 500 °C, so decreasing the VOCs content in the gas emissions.
(C3), Butane (C4) and Methane (C1). The lesser VOC is Isobutane Though the oxidizing atmosphere can accelerate the degrada-
(C4). tion reaction, using an inert gas atmosphere, such as nitrogen,
The gaseous products distribution deriving from the heat treat- the chemical oxidation of the metals is prevented, their surface
ment conducted in a slightly oxidizing atmosphere, differs consid- results clean, almost residue free (De-wen et al., 2004).
erably from that obtained in the case of thermal treatments On the other hand, in comparison with oxidizing conditions, in
conducted under pyrolytic conditions. In fact, in the absence of the temperature range from 450 °C to 600 °C, under inert atmo-
oxygen, the main products deriving from the polymeric matter sphere, many liquid and gaseous organic products of partial degra-
decomposition consist of C1-C5 hydrocarbons, predominantly C3, dation are formed (Choi and Kim, 2017; Girardin et al., 2015). To
C4 and C5 and small amounts of CO2, CO and H2 (Choi and Kim, reduce their formation it is necessary to work at a higher
2017, Marcilla et al., 2005, Zhang and Xu, 2016). temperature.
Although the thermal treatment under nitrogen, at the same
degradation rate, in most cases must be conducted at higher tem-
perature than in oxidizing atmosphere, the end pyrolysis tempera-
ture, however, is much lower than under oxidizing atmosphere due
Table 8
Composition of the degradation gas- to exothermal reactions occurring (García and Font, 2004). The
eous products of the PV panels exothermal peak occurring in oxidizing atmosphere may indicate
waste at 500 °C and U
1. burning reactions of the partial degradation products (Delfosse
Gas Product Volume % et al., 1989).
Therefore, to decrease the exothermal effect and the end ther-
Carbon Oxide 32.46
Carbon Dioxide 41.86 mal treatment temperature, pyrolysis under weak oxidizing atmo-
Methane 5.63 sphere is highly recommended.
Ethene 14.69 In this study, the optimal conditions for thermal treatment have
Ethane 0.69 been determined and correspond to a temperature of 500 °C and a
Propene 3.84
slightly oxidizing atmosphere. In particular, the oxygen content
1-Butene 0.23
n-Butane 0.27 has to be slightly higher than the necessary stoichiometric value,
i-butane 0.33 so that all the carbon contained in the PV panel structure is com-
pletely degraded.

Table 9 4. Conclusions
Type and amount of gaseous organic products deriving from the combustion of the
organic matrix contained in the PV modules.
In this work, an energy-environmental sustainable process to
Gas product Gas product/delaminated PV panel (g/g) separate the different materials deriving from silicon-PV panels is
Carbon Oxide 0.0611 proposed and developed. This process allows to achieve the effi-
Carbon Dioxide 0.1238 cient management of PV waste and the recovery of valuable mate-
Methane 0.0061 rials, by minimizing dangerous gaseous emissions, while also
Ethene 0.0277
Ethane 0.0014
assuring, a good trade-off between technical and environmental
Propene 0.0109 targets. In detail, an advanced thermal treatment aimed to elimi-
1-Butene 0.0009 nate the polymeric matter embedded in the PV panel structure
n-Butane 0.0011 has been developed by assuring a good performance from ener-
i-butane 0.0013
getic and environmental point of view. The main innovative aspect
100 V. Fiandra et al. / Waste Management 84 (2019) 91–101

of the proposed treatment lies in the total elimination of the emis- Dias, P., Javimczik, S., Benevit, M., Veit, H., Bernardes, A.M., 2016a. Recycling WEEE:
extraction and concentration of silver from waste crystalline silicon
sions of hydrofluoric acid and fluorinated organic compounds
photovoltaic modules. Waste Manage. 57, 220–225. https://doi.org/10.1016/j.
deriving from the TedlarÒ degradation, which are instead present wasman.2016.03.016.
in the conventional heat treatment. Dias, P.R., Benevit, M.G., Veit, H.M., 2016b. Photovoltaic solar panels of crystalline
The thermal process consists of two steps. The first one allows silicon: characterization and separation. Waste Manage. Res. 34 (3), 235–245.
https://doi.org/10.1177/0734242X15622812.
to completely removes the rear polymeric layer, mainly consisting Doi, T., Tsuda, I., Unagida, H., Murata, A., Sakuta, K., Kurokawa, K., 2001.
of TedlarÒ. Accordingly, the gaseous exhausts deriving from the Experimental study on PV module recycling with organic solvent method. Sol.
TedlarÒ burning, which contain fluorinated compounds and Energy Mater. Sol. Cells 67, 397–403. https://doi.org/10.1016/S0927-0248(00)
00308-1.
hydrofluoric acid, are not produced. The second step allows to sep- Drobny, J.G., 2000. Technology of Fluoropolymers. Taylor & Francis group.
arate the materials and consists of a thermal degradation process Duflou, J.R., Peeters, J.R., Altamirano, D., Bracquene, E., Dewulf, W., 2018.
of the remaining organic matter inside the PV structure. It is car- Demanufacturing photovoltaic panels: comparison of end-of-life treatment
strategies for improved resource recovery. CIRP Annals – Manuf. Technol. 67,
ried out under a slightly oxidizing atmosphere and presents a max- 29–32. https://doi.org/10.1016/j.cirp.2018.04.053.
imization of CO2 yield compared to the produced VOCs, when it is Farneth, W.E., Aronson, M.T., Uschold, R.E., 1993. Degradation mechanisms of poly
performed at a temperature of 500 °C, in air atmosphere, with an (vinyl fluoride) films. Macromolecules 26, 4765–4769. https://doi.org/
10.1021/ma00070a006.
equivalent combustion ratio U < 1. Frisson, L., Lieten, K., Bruton, T., DeClercq, K., Szlufcik, J., De Moor, H., et al., 2000.
An important innovation introduced with the proposed treat- Recent improvements in industrial PV module recycling. In: Proceedings of the
ment, is the reduction of global emissions deriving from the poly- 16th European Photovoltaic Solar Energy Conference and Exhibition, Glasgow,
United Kingdom, May 2000, pp. 3–7.
meric materials degradation, obtained with the introduction of the
García, A.N., Font, R., 2004. Thermogravimetric kinetic model of the pyrolysis and
TedlarÒ preliminary delamination step. Moreover, it is possible to combustion of an ethylene-vinyl acetate copolymer refuse. Fuel 83, 1165–1173.
reduce the process costs as compared to the conventional thermal https://doi.org/10.1016/j.fuel.2003.10.029.
treatment, through the operating conditions of the second thermal Garcìa, A.N., Viciano, N., Font, R., 2007. Products obtained in the fuel-rich
combustion of PTFE at high temperature. J. Anal. Appl. Pyrol. 80, 85–91.
step, the reduced time and treatment temperature. https://doi.org/10.1016/j.jaap.2007.01.004.
In the industrial scale-up, the heat recovery from the fumes of Gasnot, L., Decottignies, V., Pauwels, J.F., 2005. Kinetics modelling of ethyl acetate
the second step treatment could be carried out to sustain the oxidation in flame conditions. Fuel 84, 505–518. https://doi.org/10.1016/
j.fuel.2004.10.003.
energy costs of the first step. Girardin, B., Fontaine, G., Duquesne, S., Försth, M., Bourbigot, S., 2015.
Additional studies and tests are in progress to evaluate the per- Characterization of thermo-physical properties of EVA/ATH: application to
formance of the proposed thermal treatment under different tem- gasification experiments and pyrolysis modeling. Materials 8, 7837–7863.
https://doi.org/10.3390/ma8115428.
peratures and atmospheres conditions. The main goal of this Hagiwara, M., Ellinghorst, G., Hummel, D., 1977. Thermal decomposition of y-Ray
activity is to further increase the economic and environmental sus- irradiated poly (viny1 fluoride). Makromol. Chem. 178, 2901–2912.
tainability of the process, thereby fostering its industrialization. Huang, W.H., Shin, W.J., Wang, L., Sun, W.C., Tao, M., 2017. Strategy and technology
to recycle wafer-silicon solar modules. Sol. Energy 144. https://doi.org/10.1016/
j.solener.2017.01.001.
Acknowledgements Hull, T.R., Quinn, R.E., Areri, I.G., Purser, D.A., 2002. Combustion toxicity of fire
retarded EVA. Polym. Degrad. Stab. 77, 235–242. https://doi.org/10.1016/
S0141-3910(02)00039-3.
This work was supported by the National Project FERGE Jäger-Waldau, A., 2017. Snapshot of photovoltaics-March 2017. Sustainability
‘‘Devices, technics and enabling technologies for Renewable Energy (Switzerland) 9, 1–9. https://doi.org/10.3390/su9050783.
Jung, B., Park, J., Seo, D., Park, N., 2016. Sustainable system for raw-metal recovery
Sources towards Green Economy”, funded by the Italian Ministry of
from crystalline silicon solar panels: from noble-metal extraction to lead
Education, Universities and Research (MIUR). removal. ACS Sustain. Chem. Eng. 4, 4079–4083. https://doi.org/10.1021/
acssuschemeng.6b00894.
Kang, S., Yoo, S., Lee, J., Boo, B., Ryu, H., 2012. Experimental investigations for
References recycling of silicon and glass from waste photovoltaic modules. Renew. Energy
47, 152–159. https://doi.org/10.1016/j.renene.2012.04.030.
Allen, N.S., Edge, M., Rodriguez, M., Liauw, C.M., Fontan, E., 2000. Aspects of the Kim, Y., Lee, J., 2012. Dissolution of ethylene vinyl acetate in crystalline silicon PV
thermal oxidation of ethylene vinyl acetate copolymer. Polym. Degrad. Stab. 68, modules using ultrasonic irradiation and organic solvent. Sol. Energy Mater. Sol.
363–371. https://doi.org/10.1016/S0141-3910(00)00020-3. Cells 98, 317–322. https://doi.org/10.1016/j.solmat.2011.11.022.
Baker, B.B., Kasprzak, D.J., 1993. Thermal degradation of commercial fluoropolymers Klugmann-Radziemska, E., Ostrowski, P., 2010. Chemical treatment of crystalline
in air. Polym. Degrad. Stab. 42, 181–188. https://doi.org/10.1016/0141-3910 silicon solar cells as a method of recovering pure silicon from photovoltaic
(93)90111-U. modules. Renew. Energy 35, 1751–1759. https://doi.org/10.1016/j.
Bakhiyi, B., Gravel, S., Ceballos, D., Flynn, M.A., Zayed, J., 2018. Has the question of e- renene.2009.11.031.
waste opened a Pandora’s box? An overview of unpredictable issues and Klugmann-Radziemska, E., Ostrowski, P., Drabczyk, K., Panek, P., Szkodo, M., 2010.
challenges. Environ. Int. 110, 173–192. https://doi.org/10.1016/j.envint.2017. Experimental validation of crystalline silicon solar cells recycling by thermal
10.021. and chemical methods. Sol. Energy Mater. Sol. Cells 94, 2275–2282. https://doi.
Bigum, M., Brogaard, L., Christensen, T.H., 2012. Metal recovery from high-grade org/10.1016/j.solmat.2010.07.025.
WEEE: a life cycle assessment. J. Hazard. Mater. 207–208, 8–14. https://doi.org/ Ludin, N.A., Mustafa, N.I., Hanafiah, M.M., Ibrahim, M.A., Teridi, M.A., Sepeai, S.,
10.1016/j.jhazmat.2011.10.001. Zaharim, A., Sopian, K., 2018. Prospects of life cycle assessment of renewable
Bombach, E., Röver, I., Müller, A., Wambach, K., Kopecek, R., Wefringhaus, E., 2006. energy from solar photovoltaic technologies: a review. Renew. Sustain. Energy
Technical experience during thermal and chemical recycling of a 23 year old PV Rev. 96, 11–28. https://doi.org/10.1016/j.rser.2018.07.048.
generator formely installed on Pelllworm island. 21st European Photovoltaic Malandrino, O., Sica, D., Testa, M., Supino, S., 2017. Policies and measures for
Solar Energy Conference Sept 4–8, 2048–2053. sustainable management of solar panel end-of-life in Italy. Sustainability
Choi, S.S., Kim, E., 2017. Analysis of pyrolysis products of ethylene-vinyl acetate (Switzerland) 9, 1–15. https://doi.org/10.3390/su9040481.
coploymer (EVA) using pre-deacetylation. J. Anal. Appl. Pyrol. 127, 1–7. https:// Marcilla, A., Gómez, A., Menargues, S., 2005. TGA/FTIR study of the catalytic
doi.org/10.1016/j.jaap.2017.09.015. pyrolysis of ethylene-vinyl acetate copolymers in the presence of MCM-41.
D’Adamo, I., Miliacca, M., Rosa, P., 2017. Economic feasibility for recycling of waste Polym. Degrad. Stab. 89, 145–152. https://doi.org/10.1016/j.
crystalline silicon photovoltaic modules. Int. J. Photoenergy 2017, 1–7. https:// polymdegradstab.2005.01.011.
doi.org/10.1155/2017/4184676. Marwede, M., Berger, W., Schlummer, M., Mäurer, A., Reller, A., 2013. Recycling
De-wen, Z., Born, M., Wambach, K., 2004. Pyrolysis of EVA and its application in paths for thin- fi lm chalcogenide photovoltaic waste e Current feasible
recycling of photovoltaic modules. J. Environ. Sci. 16, 889–893. processes. Renew. Energy 55, 220–229. https://doi.org/10.1016/j.
Delfosse, L., Baillet, C., Brault, A., Brault, D., 1989. Combustion of ethylene-vinyl renene.2012.12.038.
acetate copolymer filled with aluminium and magnesium hydroxides. Polym. Miller, D.C., Muller, M.T., Kempe, M.D., Araki, K., Kennedy, C.E., Kurtz, S.R., 2013.
Degrad. Stab. 23, 337–347. https://doi.org/10.1016/0141-3910(89)90056-6. Durability of polymeric encapsulation materials for concentrating photovoltaic
Dias, P., Javimczik, S., Benevit, M., Veit, H., 2017. Recycling WEEE: polymer systems. Prog. Photovolt.: Res. Appl. 21, 631–651. https://doi.org/10.1002/
characterization and pyrolysis study for waste of crystalline silicon photovoltaic pip.1241.
modules. Waste Manage. 60, 716–722. https://doi.org/10.1016/j.wasman.2016. Rocchetti, L., Beolchini, F., 2015. Recovery of valuable materials from end-of-life
08.036. thin-film photovoltaic panels: environmental impact assessment of different
 V. Fiandra et al. / Waste Management 84 (2019) 91–101 101

management options. J. Cleaner Prod. 89, 59–64. https://doi.org/10.1016/j. Vellini, M., Gambini, M., Prattella, V., 2017. Environmental impacts of PV technology
jclepro.2014.11.009. throughout the life cycle: importance of the end-of-life management for Si-
Sander, K., Schilling, S., Reinschmidt, J., Wambach, K., Schlenker, S., Muller, A., panels and CdTe-panels. Energy 138, 1099–1111. https://doi.org/10.1016/j.
Springer, J., 2007. Study on the Development of a Take Back and Recovery energy.2017.07.031.
System for Photovoltaic Products. Okopol GmbH, Hamburg, Germany. Wambach, K.S.E., 2003. Method for Separating the Components of a Laminated
Sener, C., Fthenakis, V., 2014. Energy policy and financing options to achieve solar Glazing. EP0893250.
energy grid penetration targets : accounting for external costs. Renew. Sustain. Wang, R., Xu, Z., 2014. Recycling of non-metallic fractions from waste electrical and
Energy Rev. 32, 854–868. https://doi.org/10.1016/j.rser.2014.01.030. electronic equipment (WEEE): a review. Waste Manage. 34, 1455–1469. https://
Shin, J., Park, J., Park, N., 2017. A method to recycle silicon wafer from end-of-life doi.org/10.1016/j.wasman.2014.03.004.
photovoltaic module and solar panels by using recycled silicon wafers. Sol. Wang, T.Y., Hsiao, J.C., Du, C.H., 2012. Recycling of materials from silicon base solar
Energy Mater. Sol. Cells 162, 1–6. https://doi.org/10.1016/j.solmat.2016.12.038. cell module. Conf. Rec. IEEE Photovolt. Spec. Conf. 2355–2358. https://doi.org/
Shiue, Y.C., Lin, C.Y., 2012. Applying analytic network process to evaluate the 10.1109/PVSC.2012.6318071.
optimal recycling strategy in upstream of solar energy industry. Energy Build. Widmer, R., Oswald-Krapf, H., Sinha-Khetriwal, D., Schnellmann, M., Böni, H., 2005.
54, 266–277. https://doi.org/10.1016/j.enbuild.2012.07.032. Global perspectives on e-waste. Environ. Impact Assess. Rev. 25, 436–458.
Sica, D., Malandrino, O., Supino, S., Testa, M., Lucchetti, M.C., 2018. Management of https://doi.org/10.1016/j.eiar.2005.04.001.
end-of-life photovoltaic panels as a step towards a circular economy. Renew. Xu, Y., Li, J., Tan, Q., Peters, A.L., Yang, C., 2018. Global status of recycling waste solar
Sustain. Energy Rev. 82, 2934–2945. https://doi.org/10.1016/j.rser.2017.10.039. panels: a review. Waste Manage. 75, 450–458. https://doi.org/10.1016/j.
Stapler, J., Barnes, W., Yelland, W., 1968. Thermal Degradation of polyvinylidene wasman.2018.01.036.
fluoride and polvinyl fluoride by oven pyrolysis. Technical Report 69-7-C14, Yamada, T., Taylor, P.H., Buck, R.C., Kaiser, M.A., Giraud, R.J., 2005. Thermal
Natick Laboratories, Natick, MA, 1968. degradation of fluorotelomer treated articles and related materials.
Strachala, D., Hylsky, J., Vanek, J., Fafilek, G., Jandova, K., 2017. Methods for recycling Chemosphere 61, 974–984. https://doi.org/10.1016/j.chemosphere.2005.03.
photovoltaic modules and their impact on enviroment and raw material 025.
extraction. Acta Montanistica Slovaca 22, 257–269. Zeng, X., Yang, C., Chiang, J.F., Li, J., 2017. Science of the Total Environment
Tao, J., Yu, S., 2015. Review on feasible recycling pathways and technologies of solar Innovating e-waste management : from macroscopic to microscopic scales. Sci.
photovoltaic modules. Sol. Energy Mater. Sol. Cells 141, 108–124. https://doi. Total Environ. 575, 1–5. https://doi.org/10.1016/j.scitotenv.2016.09.078.
org/10.1016/j.solmat.2015.05.005. Zhang, L., Xu, Z., 2016. Separating and recycling plastic, glass, and gallium from
Teng, H., 2012. Overview of the development of the fluoropolymer industry. Appl. waste solar cell modules by nitrogen pyrolysis and vacuum decomposition.
Sci. 2, 496–512. https://doi.org/10.3390/app2020496. Environ. Sci. Technol. 50, 9242–9250. https://doi.org/10.1021/acs.est.6b01253.