Global Solar Energy Transition and End-of-Life PV Forecasts Through 2050 PDF Free Download
1 / 20
/20
100%
Global Solar Energy Transition and End-of-Life PV
Forecasts Through 2050
Xianlai Zeng
Tsinghua University https://orcid.org/0000-0001-5563-6098
Moisés Gómez
Tsinghua University
Vasilis Fthenakis
Columbia University
Mark Jacobson
Stanford University https://orcid.org/0000-0002-4315-4128
Jinhui Li
Tsinghua University https://orcid.org/0000-0001-7819-478X
Analysis
Keywords: Solar panels, Energy transition, Recycling, Critical metals, Circular economy
Posted Date: July 10th, 2025
DOI: https://doi.org/10.21203/rs.3.rs-7042450/v1
License: This work is licensed under a Creative Commons Attribution 4.0 International License.
Read Full License
Additional Declarations: There is NO Competing Interest.
1
Global Solar Energy Transition and End-of-Life PV Forecasts 1
Through 2050 2
Moisés Gómez1, Vasilis Fthenakis2, Mark Z. Jacobson3, Jinhui Li1, and Xianlai Zeng1*
3
4
1 School of Environment, Tsinghua University, Beijing 100084, China 5
2 Center for Life Cycle Analysis, Columbia University, New York, NY 10027, USA 6
3 Department of Civil and Environmental Engineering, Stanford University, Stanford, CA 94305-7
4020, USA 8
9
* Corresponding author: School of Environment, Tsinghua University, Beijing 100084, China. Email: 10
xlzeng@tsinghua.edu.cn
11
Abstract 12
Achieving net-zero emissions by 2050 relies on the rapid expansion of solar energy, projected to 13
dominate global electricity generation. However, the widespread deployment of solar panels introduces 14
a critical challenge: managing end-of-life panels as systems reach their 20–35-year lifespan. Existing 15
measurement of the obsolescence often rely on outdated data or limited regional scopes. Here, we 16
present a global forecast of obsolescent solar panels by integrating historical data, national energy 17
targets, and data-driven models to estimate installed capacity across 160 countries, covering over 95% 18
of the global population. Our projections indicate that meeting global solar targets will require 15.9–19
16.4 TW of installed capacity by 2050, generating 3.0-3.5 TW of obsolescent panels under a 25-year 20
lifespan and up to 5.3 TW under shorter lifetimes. With solar panels accumulation accelerating from 21
2024, urgent recycling strategies are needed to mitigate material losses and support a circular solar 22
economy. 23
Keywords: Solar panels, Energy transition, Recycling, Critical metals, Circular economy 24
Main 25
Solar photovoltaic (PV) has become a mature technology worldwide and has experienced exponential 26
growth over the past decade1. Since its first commercial application in 1958, the cost of solar PV has 27
plummeted by over three orders of magnitude, with a remarkable average annual decline of 15% 28
between 2010 and 20202. In addition to its low cost, solar PV has several significant advantages, 29
including safety, environmental friendliness, high efficiency, reliability, and ease of installation and 30
maintenance3. Due to these factors, global demand for solar PV has surged, with nameplate capacity 31
skyrocketing from 1.23 GW in 2000 to 1,411 GW in 2023, marking a growth rate that outpaces every 32
other renewable energy sector1. In 2023, solar PV contributed to a remarkable 75% of all new global 33
renewable electricity capacity additions and its rapid growth is expected to continue4. According to the 34
International Renewable Energy Agency (IRENA), in its 1.5°C scenario, solar energy is projected to 35
become the leading source of electricity generation by 2050, accounting for 37% of global electricity 36
supply5. Meanwhile, the International Energy Agency (IEA) projects that in its sustainable development 37
scenario, solar PV could provide approximately 62% of the world’s electricity and 26% of its total 38
energy supply by 20506. Recent reports from IRENA and the IEA highlight that to meet the Paris 39
Agreement's 1.5°C climate target and achieve net-zero emissions by 2050, global solar PV capacity 40
2
needs to reach 18.2 TW5 and 18.8 TW4, respectively. Earlier projections from these organizations had 41
estimated 14 TW (IRENA) and 9 TW (IEA) by 2050, underscoring the accelerating pace of solar 42
adoption. Jacobson et al., who modeled a transition of 149 countries (representing 99.75% of world 43
emissions) to 100% clean, renewable electricity and heat across all energy sectors upon electrification 44
of most all energy, estimated the need for ~22.8 TW of residential, commercial, and utility-scale PV 45
panels to provide about 57% of all-purpose end-use demand across the countries7. Nijsse et al. used 46
recent data-driven modeling to indicate that solar PV, driven by past policies that have spurred 47
technological advancements, is likely to become the dominant power source before mid-century, even 48
without new energy policy interventions8. 49
Despite the promising future of solar PV, the management of end-of-life (EoL) PV panels is an 50
emerging challenge. As PV deployment accelerates, the volume of decommissioned panels is projected 51
to increase exponentially, necessitating robust recycling strategies. The first major projection of EoL 52
solar panels, published by IRENA in 2016, brought critical attention to this emerging issue9. However, 53
their estimates were based on a global cumulative solar capacity of 4.5 TW by 2050, a figure that is 54
now significantly outdated. Subsequent studies have largely focused on specific regions or countries, 55
including Europe countries10-12, OECD nations13, China3,14, the United States12,15, South Korea16, and 56
Australia17. Estimates of EoL solar capacity have predominantly focused on developed countries, but 57
with the rapid global expansion of solar PV, these assessments are increasingly outdated. This 58
underscores the critical need for updated and comprehensive global analyses, encompassing both 59
developed and developing regions, that integrate recent advancements in solar deployment, evolving 60
energy targets, and diverse regional dynamics. Quantifying EoL PV at scale not only highlights the 61
magnitude of the challenge but also enhances the feasibility of organizing and implementing efficient 62
recycling systems15. By providing clear projections, stakeholders can better plan for resource recovery, 63
recycling infrastructure, and policy development, ensuring the sustainable integration of solar energy 64
into a circular economy framework9,14,15. 65
This study employs a data-driven approach to forecast global solar panel installations by analyzing 66
capacity growth from 2000 to 2023, incorporating government targets and net-zero milestones. Using 67
generalized logistic and Gompertz models, we project country-specific solar capacity through 2050 for 68
160 nations. We then estimate EoL solar panel volumes for lifespans of 20, 25, 30, and 35 years, 69
highlighting key challenges and opportunities. 70
Results 71
Current state of solar energy and PV modules production. The global installed solar energy capacity 72
has grown exponentially since 2000, with early growth driven primarily by European countries. 73
However, since 2016, Asia, led by China, Japan, and India, have overtaken Europe as the dominant 74
region in solar energy deployment (Fig. 1a). Currently, Asia leads with nearly 60% of global installed 75
solar capacity, followed by Europe (21.8%) and North America (12.4%); South America (3.1%), 76
Oceania (2.1%), and Africa (1.2%) account for smaller shares (Fig. 1b). This rapid adoption of solar 77
energy is largely attributed to a drastic decline in solar panels prices over time (Fig. 1c). Over the past 78
four decades, the price of solar panels has plummeted from approximately 106 USD ($336 in 2023 79
dollars) to 0.36 USD per watt. On average, for every doubling of installed capacity, the price of solar 80
panels decreased by approximately 20% (Fig. 1d). Some of the materials needed to produce solar panels, 81
e.g., aluminum (Al), cadmium (Cd), copper (Cu), gallium (Ga), indium (In), lead (Pb), selenium (Se), 82
silicon (Si), silver (Ag), tin (Sn), and tellurium (Te), are concentrated in a few countries, with China 83
leading production (Fig. 1e). In 2023, China alone accounted for 78% of global solar panel production, 84
followed by other Asian countries such as Vietnam, Malaysia, India, South Korea, and Japan, which 85
together contributed an additional 15%18. Given that solar panel production is heavily concentrated in 86
Asia, these countries also dominate the export market (Fig. 1f), while Europe plays a key intermediary 87
3
role. On the demand side, Europe is the largest importer of solar panels, accounting for 43% of global 88
imports, followed by North America (23%), Asia (20%), South America (8.7%), Oceania (2.7%), and 89
Africa (2.4%) (Fig. 1g). 90
91
92
93
94
95
2000 2010 2020
0
40
80
400
800
Cumulative installed solar capacity (GW)
Year
World
Africa
South Africa
Egypt
Morocco
Asia
China
Japan
India
Europe
Germany
Italy
Netherlands
North America
United States
Mexico
Canada
South America
Brazil
Chile
Argentina
Oceania
Australia
1980 1990 2000 2010 2020
0.1
1
10
100
Solar panel price (US$/W)
Year
59%
98%
12%
25%
13%
10%
42%
44%
12%
11%
23%
12%
19%
18%
79%
20%
66% 17%
39%
23%
10%
22%
42%
12%
25%
10%
100%
50%
Al
Cu
Te
Sn
Se
Si
In
Ga
Cd
Ag Pb
Metals Global
production
67% 78%
RoW
15%
0.6%
2.3%
4.2%
Global PV
production
Exports
1975
1980
1990
2000
2010
2020
10 100101102103104105106107
0.1
1
10
100
Solar panel cost (US$)
Cumulative installed solar PV capacity (MW)
Historical data
Wright's Law
95% Confident band
95% Prediction band
R2=0.9930
Wright's Law
3.1%
12.4%
2.6%
21.8% 58.9%
1.2%
Asia
Europe
North America
South America
Oceania
Africa
a
b
c
d
e
4
96
97
Fig. 1. Overview of solar energy trends: (a) global cumulative installed solar capacity from 2000 to 98
2023, categorized by continent, showing the three nations with the highest cumulative capacity in each 99
continent, (b) total global cumulative installed solar capacity for the period 2000–2023, (c) solar panel 100
price trends over time, (d) correlation between solar panel prices and cumulative installed capacity 101
(Wright’s law), (e) leading metal-producing countries for the solar panel industry, highlighting nations 102
with ≥10% share of the global market, alongside data on global solar production and panel flows, (f) 103
global solar panel export volumes in 2023, and (g) global solar panel import volumes in 2023. 104
105
Towards the 2050 Energy Transition with solar energy. Out of the 195 countries in the world, we 106
assessed 160 nations with populations exceeding 1 million. These countries represent both developed 107
and developing regions and account for over 95% of global electricity production and population. 108
Among them, 75 have set targets for installed solar capacity by 2050, with most aligning these targets 109
with long-term net-zero emission plans (Fig. 2a and Table 1). Additionally, 23 countries have 110
established medium-term solar capacity goals for 2030 or 2035, and the remaining 62 countries lack 111
official data on future installed solar capacity. Among all continents, European nations lead with nearly 112
88% having full or partial solar energy targets for their 2050 energy transition (Fig. 2b), followed by 113
Oceania (75%), Asia (66%), South America (54%), North America (46%), and Africa (31%). As 114
observed, this gap is most pronounced in African nations, where 70% (35 countries) have no reported 115
data, despite the continent's immense potential for solar energy, the highest in the world (Fig. 2d)19. By 116
2050, based on historical data, official targets and data-driven model, we project a global installed solar 117
capacity of 12,934 GW (Fig. 2c), and 15,488-16,410 GW including the replacement of retired solar 118
panels with a lifespan of 25 years. The results detailing the evolution of installed solar capacity for each 119
of the 160 countries analyzed can be found in the Supplementary Information (Figures S1 to S6). The 120
evolution of installed solar capacity by region (Fig. 2e) shows a clear trend towards dominance by Asian 121
countries, followed by North America, Europe, Africa, South America, and Oceania. To achieve these 122
2050 targets, substantial quantities of materials, including some classified as critical, will be required. 123
We analyzed 11 metals across four different solar cell technologies, examining five potential sub-124
technology scenarios (for more details, please see the Methods section). Meeting global solar capacity 125
targets will demand approximately 12.6 Gt of aluminum for frames, 134.5 Mt of aluminum for solar 126
cells, 3.2 Gt of silicon, 2.1 Gt of copper, and 168 Mt of other metals such as silver, cadmium, tellurium, 127
lead, tin, selenium, indium, and gallium (Fig. 2f-g). When replacement is factored in, material 128
requirements increase significantly, by approximately 23–27%, highlighting the substantial resource 129
demands associated with maintaining and expanding solar capacity over time (metal requirements for 130
both models, the five dominant PV sub-technology scenarios, and different solar panel lifespans are 131
detailed in Tables S6–S8 of the Supplementary Information). 132
133
United
China
India
Japan
Pakistan
Austria
Belgium
Germany Spain
France
UK
Greece
Italy
Netherlands
Poland
Portugal
USA
Australia
Brazil
Importers
Europe
North America
South America
Oceania
Africa
Asia
43%
20%
23%
2.4%
8.7% 2.7%
f
g
a
b
5
134
135
136
137
138
Africa Asia Europe North South Oceania World
0
50
100
150
Countries (number)
Goals by 2050
Partial goals
No data
No goals
America America
67%2%
12%
19%
49%
17%
6%
28%
70%
18%
5%
8%
33%
13%
53%54%
23%
23%
75%
25%
44%
13%
6%
36%
North America
South America
Europe
Asia
Oceania
Africa
2000
2010
2020
2030
2040
2050
10 TW
Scale
2 TW
5 TW
Ag Al Cd Cu Ga In Pb Se Sn Te Si Al
0.0
0.1
0.2
2
4
12
14
Cumulated metals required (Gt)
Metals
2DS (Base)
High c-Si
a-Si
CIGS
CdTe
(frames)
c
d
e
f
g
6
139
Fig. 2. Solar energy progression towards 2050 Energy Transition: (a) global mapping of national 140
targets for achieving the 2050 energy transition through solar energy by 2050, (b) analytics of the global 141
mapping of national targets, (c) installed solar energy forecasted by countries in 2050, (d) global 142
photovoltaic power potential by country, (e) evolution of installed solar capacity across continents, 143
featuring both historical data and future projections, (f) cumulative global metal requirements for solar 144
panel deployment by 2050, based on five scenarios representing different predominant sub-technologies 145
of solar panels, and (g) cumulative metal usage in installed solar panels from 2000 to 2050 in the 2DS 146
scenario. Metallic cubes, scaled by volume, illustrate the quantities of metals required. The Eiffel Tower, 147
Shanghai Tower, and Burj Khalifa (the tallest building in the world) are included as visual references, 148
with their representations scaled proportionally for comparison. Solar cells: crystalline silicon (c-Si), 149
amorphous silicon (a-Si), copper indium gallium selenide (CIGS), and cadmium telluride (CdTe). 150
PVOUT: Photovoltaic power output. Photovoltaic power potential from World Group Bank19. 151
152
Table 1. Solar energy goals, take-off year and growth of installed solar capacity in countries and 153
global regions. 154
Africa
Asia
Europe
North
America
South
America
Oceania
World
Goals by
2050
GM, GH,
KE, MU,
NG, RW,
TZ, TG,
UG
AM, BH,
BD, KH,
CN, GE,
IN, ID,
IR, IL,
JP, JO,
LB, MY,
NP, OM,
SG, KR,
LK, SY,
TW, TR,
VN
AT, BE,
BA, BG,
HR, CY,
CZ, DK,
FR, DE,
GR, HU,
IT, LV,
LT, LU,
MD, ME,
NO, PT,
RO, RS,
SK, SI,
CA, MX,
US, GT,
NI
AR, BO,
BR, CL,
CO, PE,
UY
AU, FJ,
NZ
75
7
ES, CH,
UA, GB
Partial
goals
DZ, CI,
LR, MZ,
ZA, TN
IQ, LA,
OM, PS,
PH, QA,
SA, TH
EE, FI,
IE, NL,
MK, PL,
SE
DO, PA
-
-
23
No goals
AO, BJ,
BW, BF,
BI, CM,
CF, TD,
DJ, CD,
EG, GQ,
ER, SZ,
ET, GA,
GN, GW,
LS, LY,
MG,
MW,
ML, MR,
MA, NA,
NE, CG,
SN, SL,
SO, SS,
SD, ZM,
ZW
AF, AZ,
KZ, KW,
MN,
MM, KP,
TJ, TL,
AE, UZ,
YE
AL, BY,
RU
CU, HT,
JM, PR,
TT, CR,
SV, HN
EC, PY,
VE
PG
62
No data
EH
BT, KG,
TM
IS, GL
-
GY, SR,
GF
NC
10
Logistic model
Take-off
year
2028
2026
2022
2025
2025
2025
2025
Average
growth (%)
13.2
13.9
9.1
10.8
13.5
11.0
11.9
Gompertz model
Take-off
year
2026
2025
2022
2025
2025
2025
2025
Average
growth (%)
14.2
14.1
9.0
10.8
13.3
10.8
12.0
Note: The take-off year is the year when installed solar capacity reaches or exceeds 10% of the projected final capacity by 155
2050. Bold countries indicate nations that have already surpassed their take-off year in both the generalized Logistic and 156
Gompertz models. Average growth is an average of their future growth (2023-2050). ISO codes for countries are given in 157
Table S16, in Supplementary Information. 158
159
160
Growth and evolution of installed solar capacity. The continental evolution of installed solar capacity, 161
considering the contributions from each country in both the generalized Logistic and Gompertz models, 162
indicates that Asia will lead this increase, reaching nearly 9.0-9.5 TW by 2050 (Fig. 3a and b). Europe 163
and North America are projected to follow, with approximately 2.0 and 1.9 TW, respectively. To a 164
lesser extent, Africa, South America, and Oceania are expected to contribute 0.65, 0.44, and 0.10 TW, 165
respectively. The national evolution of installed solar capacity of the 160 countries studied is found in 166
the Supplementary Information, Fig. S1-S6. While both models estimate similar installed capacity by 167
2050, their growth trajectories differ. The Logistic model follows an S-curve with rapid early growth, 168
while the Gompertz model shows slower initial growth, followed by a steep rise closer to 2050. This 169
8
contrast is reflected in the growth rates (Fig. 3c-d, Table 1): the Logistic model shows national growth 170
rates from 1.7% to nearly 40%, whereas the Gompertz model peaks between 2% and 59%, indicating 171
more acceleration in later years. The "take-off year", defined when installed capacity reaches 10% of 172
the projected 2050 level, serves as an indicator of national progress. Europe is the only continent to 173
have passed this milestone by 2022 (Fig. 3e). Other continents are expected to reach it between 2025 174
and 2028. A refined analysis using a 5-15% threshold (Fig. 3f-g) shows European countries already in 175
or beyond the take-off phase. North America, South America, and Oceania are reaching this in 2025; 176
Asia is expected in 2026; and Africa between 2026 and 2028, depending on policy action (Fig. S7). 177
Although a continent may be in the accelerating phase, intra-regional disparities exist. Overall, both 178
models signal that the global solar landscape is currently in its take-off phase, transitioning toward 179
exponential growth in installed solar capacity. 180
181
182
183
184
2000 2010 2020 2030 2040 2050
0
2
4
6
8
10
Cumulative installed solar capacity (GW)
Year
Africa
Asia
Europe
North America
South America
Oceania
Forecast
Historical data
Logistic model
2000 2010 2020 2030 2040 2050
0
2
4
6
8
10
Cumulative installed solar capacity (GW)
Year
Africa
Asia
Europe
North America
South America
Oceania
Historical data Forecast
Gompertz model
2000 2010 2020 2030 2040 2050
2000
2010
2020
2030
2040
2050
Africa
Asia
Europe
North America
South America
Oceania
World
Africa
Asia
Europe
Oceania
North America
South America
Continents (average)
Takeoff year (Gompertz model)
Takeoff year (Logistic model)
R2=0.9505
R2=0.7772
R2=0.8561
R2=0.9584
R2=0.9866
R2=0.9817
Accelerating stage
Initial stage
2020
2030
Logistic model
25%~75%
Range within 1.5IQR
Median Line
Mean
Data
15%
14%
13%
12%
11%
10%
9%
8%
7%
6%
5%
Africa
Asia
Europe
North America
South America
Oceania
World
Year
2020
2030
Africa
Asia
Europe
North America
South America
Oceania
World
Gompertz model
25%~75%
Range within 1.5IQR
Median Line
Mean
Data
15%
14%
13%
12%
11%
10%
9%
8%
7%
6%
5%
Year
a
b
d
c
e
f
g
9
Fig. 3. Forecasting future installed solar capacity: (a) projections using the Logistic model; (b) 185
average growth rate (2023-2050) based on the logistic model; (c) projections using the Gompertz model; 186
(d) average growth rate (2023-2050) based on the Gompertz model; (e) national, continental, and global 187
takeoff years for both models; and (f) take-off year corresponding to various percentages of total 188
cumulative solar capacity (5-15%) for the logistic model, and (g) for the Gompertz model. Take-off 189
year is defined as the point at which installed solar capacity reaches 10% of the total solar capacity 190
target set for 2050. 191
192
End-of-life (EoL) solar panels: Assuming a 25-year average lifespan (regular-loss scenario), the 193
global volume of EoL solar panels is projected to take off around 2024 (Fig. 4a), entering a phase of 194
exponential growth thereafter. This surge will be dominated by Asia, Europe, and North America, which 195
are expected to account for about 63.4%, 15.2%, and 15.0% of cumulative EoL solar panels by 2050, 196
respectively. South America (3.9%), Oceania (1.3%), and Africa (1.2%) will contribute smaller shares. 197
Altogether, EoL panels will generate approximately 193.7 Gt of material globally─ about 34 times the 198
mass of the Great Pyramid of Giza (Fig. 4b). China alone will lead with 937.6 GW of retired solar 199
nameplate capacity by 2050 (Fig. 4c). While the cumulative trend of EoL panels aligns with installed 200
capacity (Fig. 2c), their relative share presents a contrasting pattern (Fig. 4d). In countries such as 201
Australia, Brazil, Mexico, and the Netherlands, EoL panels will make up about 50-70% of total solar 202
capacity, indicating that even nations with lower overall deployment may face significant EoL material 203
streams. Regionally, South Africa will lead in Africa with 13.7 GW of retired capacity. In Asia, China 204
will be followed by India (474.6 GW) and Japan (228.3 GW). Australia will top Oceania with 43.4 GW. 205
In Europe, Germany (122.9 GW), Italy (48.6 GW), and France (47.9 GW) will generate the highest 206
volumes. The United States will dominate in North America (480.6 GW), while Brazil (79.7 GW) and 207
Chile (20.4 GW) will lead in South America. By 2050, EoL solar panels will constitute nearly 27% of 208
global installed solar nameplate capacity, amounting to about 3.5 TW. Detailed national projections 209
under both the Logistic and Gompertz models are available in Tables S10–S11 and Figures S9-S10 of 210
the Supplementary Information. 211
212
213
214
2000 2010 2020 2030 2040 2050
0
1
2
3
Cumulative EoL capacity (TW)
Year
Africa
Asia
Europe
North America
South America
Oceania
Takeoff year
63.4%
15.2%
3.9%
1.3%
1.2%
15%
1%
b
a
c
d
10
215
216
217
218
Fig. 4. Global distribution of End-of-Life (EoL) solar panels: (a) evolution of annual cumulative 219
EoL solar panel capacity by continent from 2000 to 2050, (b) cumulative weight of EoL solar panels 220
by 2050 by continents, (c) cumulative EoL solar panel capacity by countries in 2050, (d) percentage of 221
cumulative EoL solar panel capacity relative to the cumulative installed solar capacity by countries for 222
2050, and (e–j) cumulative EoL solar panel capacity by 2050, broken down by continent: (e) Africa, (f) 223
Asia, (g) Oceania, (h) Europe, (i) North America, and (j) South America. The Pyramid of Giza refers 224
to the Pyramid of Khufu. The size of the solar panels is proportional to their weight. 225
226
Varying the lifespan of solar panels significantly impacts the required installed solar capacity (Fig. 5). 227
In our early-loss scenario (PV modules’ lifespan of 20 years), the global cumulative decommissioned 228
solar PV capacity is projected to reach approximately 4,470-5.285 GW by 2050. This represents an 229
additional 35-41% of the global cumulative installed solar capacity by 2050, compared to 23-27% in 230
the regular-loss scenario (PV modules’ lifespan of 25 years), 12-13% in the new regular-loss scenario 231
(PV modules’ lifespan of 30 years), and 5.8-6% in the ideal-performance scenario (PV modules’ 232
lifespan of 35 years). In the early-loss scenario, embodied materials will total 249–295 Mt by 2050. In 233
comparison, the regular-loss, new regular-loss, and ideal-performance scenarios will account for 165-234
194 Mt, 83-94 Mt, and 40-43 Mt, respectively. Replacing retired solar panels is essential to meet the 235
energy transition targets set by nations. Without replacement, energy output would fall short of targets, 236
e
f
g
j
i
h
11
necessitating increased solar capacity. By 2050, the global forecast of 12.9 TW rises by 5.8-6% (13.6-237
13.7 TW) in the ideal-performance scenario, 12-13% (14.4-14.6 TW) in the new regular-loss scenario, 238
23-27% (15.9-16.4 TW) in the regular-loss scenario, and 35-41% (17.4-18.2 TW) in the early-loss 239
scenario. These figures are significant and should be carefully accounted for when forecasting the total 240
installed solar capacity. The materials embodied in EoL solar panels are substantial (Fig. 5g), amounting 241
to 1.0, 2.1, 4.3, and 6.6 Gt of critical and valuable metals for solar panels with lifespans of 35, 30, 25, 242
and 20 years, respectively. 243
244
245
246
247
2000 2010 2020 2030 2040 2050
0
1
2
3
4
5
Cumulative EoL solar capacity (TW)
Year
20y 25y 30y
Logistic model
World (20 years lifespan)
World (25 years lifespan)
World (30 years lifespan)
World (35 years lifespan)
Africa
Asia
Europe
North America
South America
Oceania
Takeoff year
35y
2000 2010 2020 2030 2040 2050
0
1
2
3
4
Cumulative EoL solar capacity (TW)
Year
World (20 years lifespan)
World (25 years lifespan)
World (30 years lifespan)
World (35 years lifespan)
Africa
Asia
Europe
North America
South America
Oceania
Takeoff year
20y 25y 30y
Gompertz model
35y
Ag Al Cd Ga In Pb Se Sn Te Cu Si Al (frames)
0
10
20
30
40
50
2500
5000
17.62%
11.67%
0.09%
0.25%
69.07%
0.25%
0.05%
1.64%
0.04%
0.01%
0.1%
0.74%
0.12%
Ga
In
Se
Te
Cd
Ag
Sn
Pb
Al
Cu
Si
Al (frames)
Cumulative metals in EoL solar panels (Mt)
Metals
Lifespan: 35y
Lifespan: 30y
Lifespan: 25y
Lifespan: 20y
a
b
c
d
e
f
g
35 years 30 years 25 years 20 years
0
5
10
15
20
Cumulative installed solar capacity (TW)
Targets
Replacement
12.9 TW
4.5 TW
12.9 TW 12.9 TW
3.0 TW
1.5 TW
35%
23%
12%
Gompertz model
12.9 TW
5.8%
0.72 TW
35 years 30 years 25 years 20 years
0
5
10
15
20
Cumulative installed solar capacity (TW)
Targets
Replacement
12.9 TW
5.3 TW
12.9 TW 12.9 TW
3.5 TW
1.7 TW
41%
27%
13%
Logistic model
12.9 TW
0.77 TW
6%
12
248
Fig. 5. End-of-Life (EoL) solar panels under varying lifespans (20, 25, 30, and 35 years) by 2050: 249
(a) cumulative solar capacity retirement for different lifespans (Logistic model), (b) total EoL solar 250
panel weight (Logistic model), (c) additional cumulative solar capacity required to replace retired panels 251
(Logistic model), (d) cumulative solar capacity retirement for different lifespans (Gompertz model), (e) 252
total EoL solar panel weight (Gompertz model), (f) additional cumulative solar capacity required to 253
replace retired panels (Gompertz model), (g) cumulative metals in EoL solar panels in the 2DS scenario, 254
(h) weight evolution of EoL PV modules across the four scenarios, and (i) comparison of fossil-fuel 255
waste with EoL PV modules through 2050. Fossil-fuel projections are based on data from the IEA Net 256
Zero by 2050 scenario. The size of the PV panels is proportional to their weight. The take-off year of 257
EoL PV panels was defined as the point at which cumulative EoL PV capacity reaches 1% of the 258
projected cumulative EoL capacity by 2050. 259
Discussion 260
End-of-life (EoL) solar panels were projected to have begun rapidly increasing around 2024 (ranging 261
from 2022 to 2028 in alternative scenarios), with exponential growth anticipated thereafter. The 262
geographic distribution of these discarded panels varies widely, with the largest concentrations expected 263
in Asia (Fig. 4). This presents both a significant challenge and an opportunity. Establishing effective 264
policy frameworks and value chains requires time, making it crucial for governments, industries, and 265
other stakeholders to proactively prepare for the approaching surge of end-of-life PV. The EU has led 266
the way with its regulatory framework for EoL PV. The Waste Electrical and Electronic Equipment 267
(WEEE) Directive mandates that producers supplying PV panels to the EU market finance collection 268
and recycling costs, enforcing an 85% recovery rate12. While the U.S. lacks a federal policy specifically 269
addressing EoL PV, several states, including California, Washington, North Carolina, and Minnesota, 270
have implemented programs to manage and recycle PV modules12,20. These initiatives highlight regional 271
efforts to bridge the policy gap, but a comprehensive national strategy remains absent. China, the 272
world’s largest PV market, has driven rapid solar adoption through policies such as the Photovoltaic 273
Poverty Alleviation Initiative. However, the country’s PV recycling industry is still in its infancy, with 274
a lack or minimal supporting policies or public awareness14,20. The National Solid Waste Law provides 275
a general framework for managing solid waste but does not explicitly address EoL PV installations3,20. 276
There is currently no clear mandate for handling EoL solar PV, leaving a significant regulatory gap in 277
the world’s largest solar-producing region. In Latin America, policies and initiatives for managing EoL 278
solar panels are at a nascent stage. Discussions are often integrated into broader waste management 279
frameworks, with a focus on e-waste. However, dedicated regulations for EoL solar PV have yet to be 280
developed. In Africa, while some countries have introduced waste management strategies, EoL solar 281
panel management is not yet a legislative priority. The rapidly growing global solar sector demands the 282
development of policies and infrastructure to address future challenges. Recycling companies 283
worldwide have already demonstrated the feasibility of recycling EoL solar panels21-24, achieving 284
recovery rates of up to 95%21. However, as the volume of EoL solar panels is set to surge soon, greater 285
2000 2010 2020 2030 2040 2050
0
10
20
Lifespan: 20y
Lifespan: 25y
Lifespan: 30y
Lifespan: 35y
Cumulative lifespan: 20y
Cumulative lifespan: 25y
Cumulative lifespan: 30y
Cumulative lifespan: 35y
Year
EoL PV (Mt)
0
100
200
300
Cumulative EoL PV (Mt)
2000 2010 2020 2030 2040 2050
0
10
20
Cumulative EoL PV
Cumulative fossil-fuel waste
Coal
Oil
Natural gas
EoL PV
Year
Fossil-fuel waste & EoL PV (Gt)
0.1
0.2
0.3
200
400
600
800
Cumulative fossil-fuel waste & EoL PV (Gt)
h
i
13
efforts are needed to develop innovative solutions for managing this emerging stream of secondary 286
resources. 287
The early replacement of solar panels and the shorter than expected lifespans of PV modules globally 288
is an emerging concern. The degradation of solar panels over time, losing efficiency at rates typically 289
ranging from 0.3 to 0.7 % annually25,26. In some cases, degradation rates can exceed these averages due 290
to poor manufacturing quality or exposure to harsh environmental conditions, resulting in early 291
replacements. Moreover, it has been shown that degradation tends to accelerate as solar panels age25. 292
Performance degradation, combined with technological advancements and declining costs, has led to 293
the concept of the economic replacement of solar panels, i.e. replacing PV systems before the end of 294
their functional lifespan to maximize returns. This approach could offer benefits beyond cost savings, 295
including improved efficiency and environmental advantages. A study has shown that replacing solar 296
panels with an initial lifetime of under 15 years can be economically favorable. Additionally, life-cycle 297
analyses indicate that most commercial PV technologies provide environmental benefits across most 298
impact categories, regardless of the replacement strategy, compared to the current electricity generation 299
mix26. Early replacement or shorter lifespan of PV modules, even if economic replacement offers 300
potential benefits, it also raises concerns about increased waste generation and resource consumption, 301
underscoring the need for circular economy strategies to manage retired panels responsibly. According 302
to the Global E-Waste Monitor 2024, 62 million metric tons (Mt) of e-waste were generated worldwide 303
in 2022, with projections suggesting this figure will rise to 82 Mt by 203027. The EoL solar PV 304
forecasted in this study for 2050 is estimated to be approximately 4–5 times greater than the global e-305
waste generated in 2022, underscoring the magnitude of this emerging challenge. Alarmingly, only 22.3% 306
of e-waste generated in 2022 was documented as properly collected and recycled, highlighting 307
significant gaps in global e-waste recycling infrastructure and measures27. The management of EoL 308
solar PV is becoming a significant environmental challenge, requiring coordinated global efforts in 309
research, policymaking, and technological innovation. Implementing effective recycling and material 310
recovery strategies, backed by robust legislation and international collaboration, is essential to minimize 311
environmental impacts while harnessing the economic potential of EoL PV materials3,12,20. Until a 312
comprehensive and regulated approach to EoL solar panel management is established, it remains crucial 313
to balance technological advancements in photovoltaics with sustainable practices, such as extending 314
module lifespans through improved materials, durable manufacturing, and enhanced recycling systems. 315
Comprehensive predictions of future EoL solar PV, like this study, are a crucial foundation for 316
developing effective PV recycling strategies. 317
Methods 318
Data collection. Data on solar energy capacity and targets for each country was collected from official 319
government reports and websites. Out of 195 countries, we assessed 160 nations with populations over 320
1 million. These countries represent both developed and developing regions, covering more than 95% 321
of global electricity production and population. Islands with populations under 1 million and small 322
countries (≤1000 km2) were excluded from consideration. Additionally, countries with no available 323
historical data, such as Iceland, Suriname, and Kyrgyzstan, among others (10 in total), were catalogued 324
with no data. Data for each country was collected individually and manually to minimize errors, 325
resulting in a comprehensive database of national solar energy targets. Table S1 in the Supplementary 326
Information summarizes this information, including relevant observations, references, and 327
governmental sources. For countries where specific data was unavailable, solar capacity for 2050 was 328
estimated using two methods: (i) for countries with partial goals (e.g., capacity targets for years other 329
than 2050, such as 2030 or 2035), the Logistic and Gompertz growth models were applied; (ii) for 330
countries with no available goals, we estimated the cumulative installed solar capacity in 2050 using an 331
exponential model with a conservative growth rate of 10%. These cases were predominant in countries 332
and regions with less supportive policies or lower economic growth. 333
14
Solar energy roadmaps. Historical solar capacity data (2000-2023) was sourced from the International 334
Renewable Energy Agency (IRENA)1, while projections for future installed capacity were estimated 335
using predictive models. Among the various mathematical models for analyzing the diffusion of new 336
technologies over time, we selected two models with distinct characteristics. The Logistic model, which 337
follows an S-curve trajectory, is widely used to study energy technologies, as it captures the typical 338
phases of growth, saturation, and maturity2,3,28. Solar capacity forecasting is often performed using a 339
standard logistic function. However, for more complex technologies or cases with multiple growth 340
phases, transitions, or distinct adoption limits, the generalized logistic function is typically more 341
appropriate. In this study, given the global scale and diverse national contexts, as well as the complex 342
factors influencing adoption rates (e.g., policy shifts, resource constraints), we opted for the generalized 343
Logistic function. To account for varying growth dynamics, we also considered the Gompertz model. 344
Unlike the symmetric curve of the generalized Logistic model, the Gompertz model incorporates 345
asymmetry, allowing for a slower, extended growth beyond the inflection point, characteristic that can 346
be particularly useful in cases where initial growth is influenced by policy measures28,29. Both the 347
generalized Logistic and Gompertz models, which capture distinct patterns of technological adoption 348
and market saturation, are described as follows: 349
350
351
where A1 and A2 represent the initial and final values (asymptotes) of the curve, respectively, x0 is the 352
inflection point, k is the growth constant, and A is the asymptote of the Gompertz model curve. 353
Models. The historical data-driven models employed in this study are anticipated to provide a more 354
accurate forecast when compared to traditional projections (more information in section 3 of the 355
Supplementary Information). While we acknowledge the relevance of other projection-based models, 356
which are useful for evaluating policy, deployment and cost scenarios aligned with climate targets, data-357
driven models offer distinct advantages28. Specifically, they are less constrained by assumptions about 358
the future and can often reflect more realistic growth trajectories. Organizations such as IRENA and the 359
IEA rely on projection-based models. IRENA adopts an aspirational approach, often illustrating what 360
should happen to achieve a renewable energy future30, while the IEA provides a broader energy outlook, 361
presenting multiple scenarios based on varying levels of policy expectations and economic conditions31. 362
In contrast, our work adopts a purely data-driven methodology. Projections have historically tended to 363
underestimate renewable energy deployment while overestimating costs, particularly for solar and wind 364
energy transitions2. Our primary objective is to empirically forecast global installed solar capacity on a 365
nation-by-nation basis to improve predictions of end-of-life (EoL) solar panels. Without accurately 366
estimating the evolution of installed solar capacity, reliable EoL forecasts cannot be made. For African 367
nations, where most countries lack specific renewable energy targets or goals, historical data is utilized. 368
An exponential growth assumption with a modest annual rate of 10% was applied. Even with an 369
increased growth rate of 15%, the impact on global installed solar capacity projections is minimal (≤5% 370
increase), largely due to the dominant influence of major players like China, the United States, and 371
Europe. 372
Photovoltaic (PV) market and sub-technologies. The PV module units were estimated using averages 373
of module power, sizes, and weight, for residential and power plants reported by the International 374
Technology Roadmap for Photovoltaics (ITRPV)32. The average solar panel size was estimated at 2 m2, 375
reflecting typical module dimensions used in residential and utility-scale applications, with an energy 376
generation rate of 180 W/m2. For weight, a 20 kg average was assumed for panels from 2010 to 2030, 377
increasing to 22 kg from 2031 onward (Table S2 in Supplementary Information). Only mature, 378
commercially established PV technologies were considered in this study, specifically crystalline silicon 379
15
(c-Si) from first-generation solar cells, and second-generation thin-film technologies such as amorphous 380
silicon (a-Si), copper indium gallium selenide (CIGS), and cadmium telluride (CdTe). Currently, 381
crystalline silicon is the dominant solar cell technology, representing over 85% of the global market 382
share and nearly 95% of new solar capacity added in 202032. However, the market shares of a-Si, CIGS, 383
and CdTe technologies are projected to increase gradually in the coming years33,34. Future PV market 384
projections from the World Bank Group (WBG) were used in this study to model different technological 385
scenarios toward 2050 (refer to Table S3 and Figure S8 in Supplementary Material). In total, five 386
dominant sub-technologies scenarios were studied (2DS, high c-Si, high a-Si, high CIGS, and high 387
CdTe scenarios)34. The annual evolution of each sub-technology was calculated using a linear trend. 388
Material requirements. Each of the PV technologies (c-Si, a-Si, CIGS and CdTe) requires distinct 389
materials for manufacturing, and their metal requirements by 2050 can vary significantly depending on 390
which technology becomes dominant. Historical data from 2000 to 2023 was used, assuming an initial 391
2020 market share of 90% for c-Si, 5% for CdTe, and 5% for CIGS32, reflecting global trends. The 392
future evolution of PV technologies was based on the WBG scenarios34. The annual material demand 393
(AMD) and total material demand by 2050 (TMD) in the different scenarios and technologies were 394
determined based on the bottom-up approach by: 395
396
397
398
where and are the sub-technologies installed capacity in year t, and 399 their material intensities for metal x in their respective sub-technologies. 400 is the accumulated total material demand for metals x and time t. 401
402
End-of-life (EoL) PV modules. On average, PV modules come with a product warranty of 403
approximately ten years. Manufacturers typically offer a performance warranty, ensuring no less than 404
98% of the labeled power output in the first year, followed by a maximum annual decline of 0.55%. 405
This means that after 25 years, the module is expected to perform at no less than 84.8% of its original 406
labeled power output35. Consequently, crystalline silicon (c-Si) PV panels have a nominal service life 407
or technical lifetime of 25 years, with some premium manufacturers now extending warranties up to 408
30-35 years. The lifespan of PV panels was set at 25 years, which is the average and widely accepted 409
lifespan for this technology3,15,17,35,36. To better capture the spectrum of performance and degradation, 410
we considered four distinct lifespan scenarios. The early-loss scenario (20 years) is characterized by 411
rapid failure due to factors such as poor quality, extreme weather, or manufacturing defects. The 412
regular-loss scenario (25 years) represents typical degradation patterns under normal operating 413
conditions, serving as a baseline for average performance. The new regular-loss scenario (30 years) 414
reflects degradation patterns observed in newer PV modules from recent manufacturers, which are not 415
yet representative of the general lifespan but show trends of newer technologies. Finally, the ideal-416
performance scenario (35 years) represents the best-case scenario, with minimal degradation and 417
optimal maintenance ensuring extended module life. Together, these scenarios encompass a range of 418
potential outcomes, from rapid decline to optimal longevity. 419
420
The EoL profile of PV modules can be modeled using various probability functions; however, the 421
Weibull distribution is widely regarded as the most suitable for this type of estimation9,14,37. The 422
probability density function for the Weibull distribution is expressed as follows37: 423
424
16
where being the initial year; and representing the shape and scale parameters, respectively. For 425
the early-loss scenario, the Weibull parameters used were 2.83 (shape, ) and 20 (scale, )14. In the 426
regular-loss scenario, the parameters were 3.5 () and 25 ()37, for the new regular-loss scenario, the 427
parameters were 5.38 () and 30.37 ()9,14, while for the ideal-performance scenario, the parameters 428
were 7 () and 35 (). In the case of the ideal-performance scenario, due to the absence of official 429
parameters, the values were estimated using the Weibull distribution relationship: 430
431
432
433
where is the mean lifespan for different scale parameters and is the gamma function. In the Weibull 434
distribution, the Gamma function helps compute the expected lifespan when given the scale parameter 435 and shape parameter . The term Γ (1 + 1/
) adjusts for how the distribution skews over time, 436
influencing the expected lifespan of solar panels. For a scale parameter of 35 years, the closest shape 437
parameter that results in a mean lifespan near 35 years is ~7, which is a moderate degradation pattern 438
observed in the industry. 439
440
The material content of PV modules at their EoL in each scenario was calculated based upon a bottom-441
up approach by multiplying the EoL solar capacity and their material content, using the AMD and TMD 442
equations. 443
Take-off year and Sensitivity analysis. We define the take-off year as the year when installed solar 444
capacity reaches or exceeds 10% of the projected final capacity by 2050. Generally, in the Gompertz 445
model, the take-off point occurs slightly earlier than in the logistic model28, typically when capacity is 446
between 5% and 15% of the asymptotic limit. To account for these differences, we conducted a 447
sensitivity analysis, varying the take-off year between 5% and 15% of the final capacity across different 448
countries and continents (Figure S7 in the Supplementary Material). For the take-off year of EoL PV 449
modules, we define it as the point at which cumulative EoL solar capacity reaches 1% of the projected 450
cumulative EoL capacity by 2050. This marks the onset of exponential growth in this EoL PV stream, 451
allowing us to highlight the starting point of this rapid increase. 452
Acknowledgements 453
This work was supported by the National Natural Science Foundation of China (Grant No. W2433173 454
and 92462301). 455
Author contributions 456
M.G. conceived the idea, collected the data, developed the methodology, conducted the analysis, and 457
wrote the manuscript–original draft and review & editing. M.G. and X.Z. developed the concept and 458
designed the project. X.Z. worked on the analysis and supervised the project. V.F., M. J., X.Z. and J.L. 459
worked on the discussion and manuscript writing–review & editing. 460
Competing interests 461
The authors declare no competing interests. 462
References 463
1 IRENA. Renewable energy statistics 2024. (Abu Dhabi, 2024). 464
2 Way, R., Ives, M. C., Mealy, P. & Farmer, J. D. Empirically grounded technology forecasts 465
and the energy transition. Joule 6, 2057-2082, doi:https://doi.org/10.1016/j.joule.2022.08.009
466
(2022). 467
17
3 Gómez, M. et al. Navigating the future: China’s photovoltaic roadmap challenges. Science 468
Bulletin 68, 2491-2494, doi:https://doi.org/10.1016/j.scib.2023.08.022 (2023). 469
4 International Energy Agency (IEA). Net Zero Roadmap A Global Pathway to Keep the 1.5 °C 470
Goal in Reach. 226 (2023). 471
5 IRENA. World Energy Transitions Outlook 2023: 1.5°C Pathway. (International Renewable 472
Energy Agency, Abu Dhabi, 2023). 473
6 International Energy Agency (IEA). Net Zero by 2050. A Roadmap for the Global Energy 474
Sector. 224 (Paris, 2021). 475
7 Jacobson, M. Z., Sambor, D. J., Fan, Y. F. & Mühlbauer, A. Effects of firebricks for industrial 476
process heat on the cost of matching all-sector energy demand with 100% wind–water–solar 477
supply in 149 countries. PNAS Nexus 3, pgae274, doi:10.1093/pnasnexus/pgae274 (2024). 478
8 Nijsse, F. J. M. M. et al. The momentum of the solar energy transition. Nature Communications 479
14, 6542, doi:https://doi.org/10.1038/s41467-023-41971-7 (2023).
480
9 IRENA. End-of-life management: Solar Photovoltaic Panels. (2016). 481
10 Czajkowski, A., Wajda, A., Poranek, N., Bhadoria, S. & Remiorz, L. Prediction of the Market 482
of End-of-Life Photovoltaic Panels in the Context of Common EU Management System. 483
Energies 16, doi:https://doi.org/10.3390/en16010284 (2023).
484
11 Bošnjaković, M., Galović, M., Kuprešak, J. & Bošnjaković, T. The End of Life of PV Systems: 485
Is Europe Ready for It? Sustainability 15, doi:https://doi.org/10.3390/su152316466 (2023). 486
12 Nain, P. & Anctil, A. End-of-life solar photovoltaic waste management: A comparison as per 487
European Union and United States regulatory approaches. Resources, Conservation & 488
Recycling Advances 21, 200212, doi:https://doi.org/10.1016/j.rcradv.2024.200212 (2024).
489
13 Mahmoudi, S., Huda, N. & Behnia, M. Critical assessment of renewable energy waste 490
generation in OECD countries: Decommissioned PV panels. Resources, Conservation and 491
Recycling 164, 105145, doi:https://doi.org/10.1016/j.resconrec.2020.105145 (2021).
492
14 Lin, M. et al. Response to the Upcoming Emerging Waste: Necessity and Feasibility Analysis 493
of Photovoltaic Waste Recovery in China. Environmental Science & Technology 56, 17396-494
17409, doi:https://doi.org/10.1021/acs.est.2c06956 (2022).
495
15 Domínguez, A. & Geyer, R. Photovoltaic waste assessment of major photovoltaic installations 496
in the United States of America. Renewable Energy 133, 1188-1200, 497
doi:https://doi.org/10.1016/j.renene.2018.08.063 (2019).
498
16 Kim, H. & Park, H. PV Waste Management at the Crossroads of Circular Economy and Energy 499
Transition: The Case of South Korea. Sustainability 10, 500
doi:https://doi.org/10.3390/su10103565 (2018).
501
17 Mahmoudi, S., Huda, N. & Behnia, M. Photovoltaic waste assessment: Forecasting and 502
screening of emerging waste in Australia. Resources, Conservation and Recycling 146, 192-503
205, doi:https://doi.org/10.1016/j.resconrec.2019.03.039 (2019).
504
18 The Observatory of Economic Complexity (OEC). Photosensitive devices; assembled 505
photovoltaic modules/panels, <https://oec.world/en/profile/hs/photosensitive-devices-506
assembled-photovoltaic-modulespanels> (2024).
507
19 Suri, M. B., Juraj; Rosina,Konstantin; Chrkavy,Daniel; Suriova,Nada; Cebecauer,Tomas; 508
Caltik,Marek; Erdelyi,Branislav. Global Photovoltaic Power Potential by Country (English). 509
( World Bank Group. , Energy Sector Management Assistance Program (ESMAP) Washington, 510
D.C. :, 2020). 511
20 Ali, A., Malik, S. A., Shafiullah, M., Malik, M. Z. & Zahir, M. H. Policies and regulations for 512
solar photovoltaic end-of-life waste management: Insights from China and the USA. 513
Chemosphere 340, 139840, doi:https://doi.org/10.1016/j.chemosphere.2023.139840 (2023).
514
21 SOLARCYCLE, <https://www.solarcycle.us/> (2025). 515
22 First Solar, <https://www.firstsolar.com/> (2025).
516
23 RESOLAR, <https://www.resolartech.com/en/> (2025). 517
24 PV CYCLE ORG, <https://pvcycle.org/> (2025). 518
25 Virtuani, A. et al. 35 years of photovoltaics: Analysis of the TISO-10-kW solar plant, lessons 519
learnt in safety and performance—Part 1. Progress in Photovoltaics: Research and 520
Applications 27, 328-339, doi:https://doi.org/10.1002/pip.3104 (2019).
521
18
26 Jean, J., Woodhouse, M. & Bulović, V. Accelerating Photovoltaic Market Entry with Module 522
Replacement. Joule 3, 2824-2841, doi:https://doi.org/10.1016/j.joule.2019.08.012 (2019). 523
27 Cornelis P. Baldé et al. The Global E-waste Monitor 2024. 148 (Geneva/Bonn, 2024). 524
28 Cherp, A., Vinichenko, V., Tosun, J., Gordon, J. A. & Jewell, J. National growth dynamics of 525
wind and solar power compared to the growth required for global climate targets. Nature 526
Energy 6, 742-754, doi:https://doi.org/10.1038/s41560-021-00863-0 (2021).
527
29 Davies, S. W. & Diaz-Rainey, I. The patterns of induced diffusion: Evidence from the 528
international diffusion of wind energy. Technological Forecasting and Social Change 78, 529
1227-1241, doi:https://doi.org/10.1016/j.techfore.2011.03.012 (2011).
530
30 IRENA. World Energy Transitions Outlook 2024: 1.5°C Pathway. (International Renewable 531
Energy Agency, Abu Dhabi, 2024). 532
31 International Energy Agency (IEA). World Energy Outlook 2024. 398 (2024). 533
32 Fischer, M., Woodhouse, M., Herritsch, S. & Trube, J. International Technology Roadmap for 534
Photovoltaic (ITRPV): Results 2020. 88 (2021). 535
33 European Commission et al. Raw materials demand for wind and solar PV technologies in the 536
transition towards a decarbonised energy system. Report No. 978-92-76-16225-4, (Publications 537
Office, 2020). 538
34 Kirsten Hund, D. L. P., Thao P. Fabregas, Tim Laing, John Drexhage. Minerals for Climate 539
Action: The Mineral Intensity of the Clean Energy Transition. (Climate Smart Mining Initiative 540
- The World Bank Group, 2020). 541
35 Ganesan, K. & Valderrama, C. Anticipatory life cycle analysis framework for sustainable 542
management of end-of-life crystalline silicon photovoltaic panels. Energy 245, 123207, 543
doi:https://doi.org/10.1016/j.energy.2022.123207 (2022).
544
36 Forti, V., Baldé, C. P., Kuehr, R. & Bel, G. The Global E-waste Monitor 2020: Quantities, 545
flows and the circular economy potential. p. 120 (United Nations University 546
(UNU)/UnitedNations Institute for Training and Research (UNITAR) – co-hosted SCYCLE 547
Programme, International Telecommunication Union (ITU) & International Solid Waste 548
Association (ISWA), Bonn, Geneva and Rotterdam, 2020). 549
37 Forti, V., Baldé, K. & Kuehr, R. E-waste Statistics: Guidelines on Classifications, Reporting 550
and Indicators, second edition. 71 (United Nations University, ViE – SCYCLE, Bonn, Germany, 551
2018). 552
553
