ll

Perspective
Solar photovoltaics is ready
to power a sustainable future
Marta Victoria,1,2,13,* Nancy Haegel,3 Ian Marius Peters,4 Ron Sinton,5 Arnulf Jäger-Waldau,6
Carlos del Cañizo,7 Christian Breyer,8 Matthew Stocks,9 Andrew Blakers,9 Izumi Kaizuka,10
Keiichi Komoto,11 and Arno Smets12

SUMMARY Context & scale

Thanks to fast learning and sustained growth, solar photovoltaics (PV) Limiting global temperature
is today a highly cost-competitive technology, ready to contribute sub- increase to 1.5 C requires a rapid
stantially to CO2 emissions mitigation. However, many scenarios as- and profound transformation of
sessing global decarbonization pathways, either based on integrated our energy system. Solar
assessment models or partial-equilibrium models, fail to identify the photovoltaics (PV) is a mature
key role that this technology could play, including far lower future PV technology ready to contribute to
capacity than that projected by the PV community. In this perspective, this challenge. Throughout the
we review the factors that lie behind the historical cost reductions of last decade, a higher capacity of
solar PV and identify innovations in the pipeline that could contribute solar PV was installed globally
to maintaining a high learning rate. We also aim at opening a construc- than any other power-generation
tive discussion among PV experts, modelers, and policymakers technology and cumulative
regarding how to improve the representation of this technology in capacity at the end of 2019
the models and how to ensure that manufacturing and installation of accounted for more than 600 GW.
solar PV- can ramp up on time, which will be crucial to remain in a However, many future low-carbon
decarbonization path compatible with the Paris Agreement. energy scenarios have failed to
identify the potential of this
technology.
INTRODUCTION
Our ability to reduce greenhouse gas emissions by 2030 will determine whether we In this perspective, we present
remain on a path compatible with the Paris Agreement or whether limiting temperature arguments for anticipating that
increase to 1.5 C above the preindustrial level is beyond our reach.1 Solar photovoltaics PVs could become our majority
(PV) is now a mature technology, which is ready to deploy at the multi-terawatt scale and global energy source and argue
contribute to emission reduction in the short term. Global electricity generation from so- for an improved representation of
lar PV is an order of magnitude lower than conventional technologies (it accounted for this technology in the models.
2.8% at the end of 20192), but it shows a very steep progression, see Figure 1. Several New innovations, at both the solar
factors lie behind the plummeting cost and fast ramp up of this technology. One partic- cell and system levels, could
ularly interesting factor is the fact that PV is modular, also referred to as ‘‘granular.’’3 Iden- contribute to keeping the high
tical solar panels of hundreds of watts are combined, by the dozens in rooftop installa- learning rate shown in the past.
tions, or by the millions in utility-scale power plants. The small unit size and low unit Neither materials nor land use will
investment have enabled a much faster scaling through replication compared with other prevent PV expansion. The
technologies,3 whereas several markets were addressed in parallel based on high modu- integration of strategies, both
larity and access to solar resources in all countries. Today, solar provides the cheapest existing and under development,
electricity in many parts of the world,4,5 and the milestone of attaining a cumulative ca- could enable solar PV to
pacity higher than 1 TW is expected before 2023.6,7 Despite the technological achieve- contribute not only to
ments, many of the analyses involving future energy scenarios have failed to identify the decarbonization of the power grid
key role that solar PV could play in 2030 and in 2050 when net-zero emissions are but also other sectors through
required.8–12 direct or indirect electrification.

Integrated assessment models (IAMs), which include a global representation of

energy, economy, land, and climate, are typically used both to investigate transition

Joule 5, 1041–1056, May 19, 2021 ª 2021 Elsevier Inc. 1041

 ll
Perspective

We identify the following

challenges for a sustained scaling
up of solar PV in the next decade:
ensuring adequate regulatory
frameworks that reduce soft costs,
reducing capital expenditure via
industrial innovations, untapping
the demand for PV by enabling
electrification of other energy
sectors assisted by proper tax
schemes, and strengthening
research on improving efficiency
and reliability of PV systems.

Figure 1. Historical expansion of electricity generation technologies

Original plot with data from BP, 2009. 2

paths under stringent carbon budgets and to determine cost-optimal future sce-
narios.8–12 Some IAMs include lower solar PV contributions to sustainable scenarios
than those predicted by other analyses (Figure 2A.)12,13 For instance, the 890 IAM
scenarios included in the 5th Assessment Report (AR) of the Intergovernmental Panel
on Climate Change (IPCC)8 considered, on average, a global solar PV electricity gen-
eration of 4.9 PWh/year in 2050. For the 311 IAM scenarios covered in the IPCC
1.5 C special report,9 the average is 12.5 PWh/year. Conversely, researchers in
1Department of Mechanical and Production
the PV community estimate that in 2050 solar electricity could contribute 41 to 96
Engineering, Aarhus University, 8000 Aarhus,
PWh/year.6,14 The main reasons that cause the outputs to downplay the role of solar Denmark
PV in IAM results are cost assumptions and constraints that limit the maximum 2iCLIMATE Interdisciplinary Centre for Climate
renewable penetration in the grid, together with low levels of direct and indirect Change, Aarhus University, 8000 Aarhus,
Denmark
electrification of other sectors. Those reasons will be examined in detail in the
3NREL, Golden, CO 80401, USA
following sections.
4Forschungszentrum Jülich, 52425 Jülich,
Germany
Although the underestimation of the solar PV potential in IAMs was initially assessed 5Sinton Instruments, Boulder, CO 80301, USA
by Creutzig et al.,12 Breyer et al,13 and Fraunhofer ISE,16 expanding this discussion 6European Commission - Joint Research Centre
seems relevant because most of the limitations persist in some IAMs, and the evolu- (JRC), 2749 Ispra, Italy
tion of PV technology in recent years has significantly reduced its costs. Furthermore, 7Instituto de Energı́a Solar, Universidad
Politécnica de Madrid, 28040 Madrid, Spain
the outcomes of IAMs constitute the results included in the IPCC ARs8,17 and influ-
8LUTUniversity, Yliopistonkatu 34, 53850
ence the narratives on the energy transition.18 As an example, the role of solar PV
Lappeenranta, Finland
and wind is particularly understated in IAMs when technologies with uncertain devel- 9Research School of Electrical, Energy, and
opment prospects, such as bioenergy with carbon capture and storage (BECCS), are Materials Engineering Australian National
considered. Reinforced by discount-rate effects, such technologies emerge in IAM University, Canberra 2601, Australia
10RTS Corporation, Tokyo, Japan
results showing contributions to primary energy similar or higher than wind or solar
PV by 205010,19,20 (Figure S1). There are also doubts as to whether the biomass for
11Mizuho Information & Research Institute,
Tokyo, Japan
the assumed BECCS volumes would be sustainably available and whether a solar- 12Photovoltaic Materials and Devices Laboratory,
PV-based CO2 direct air carbon capture and storage (DACCS) route would not be Delft University of Technology, 2628 Delft, the
more attractive.21,22 Some IAMs have recently made efforts to improve the represen- Netherlands
tation of wind and solar PV,12,20,23,24 but there are indications that they could be 13Lead contact
insufficient in many cases to capture the rapid evolution and critical role of sector *Correspondence: [email protected]
coupling. Furthermore, the 7-year cycle for every new AR and the rapid evolution https://doi.org/10.1016/j.joule.2021.03.005

1042 Joule 5, 1041–1056, May 19, 2021

 ll
Perspective

Figure 2. Historical evolution and projected tendencies for solar photovoltaics

(A) Global PV electricity in 2050. Violin plots represent the distribution of results included in IPCC 5 th AR 8 and 1.5  C SR. 9
(B) PV module price versus cumulative capacity.
(C) Cost assumptions for PV plants. The red star (circle) marks the average (minimum) cost in 2019. 15
(D) Historical efficiency for multi and monocrystalline solar cells and efficiency targets included in the industry roadmap.
(E) Age distribution for existing power plants.
(F) Ratio of silver consumption to PV manufactured capacity. Data sources are listed in the supplemental experimental procedures.
See also Figure S1.

of the PV technology in recent years have widened the gap between the PV mitiga-
tion potential estimated by the mainstream IAMs (and included in the ARs) and that
based on the technology’s current achievements and trajectory.

Similarly, partial-equilibrium models using annual resolution also tend to result in

low solar contributions. For example, the PRIMES model, which supports the ana-
lyses of policy options for the European Commission,25,26 estimates that solar PV
supplies less than 20% of electricity demand in all the cost-optimal 1.5 C and 2 C
scenarios for 2050. This is well below the cost-optimal solar penetration obtained
by other energy models including hourly resolution (Figure 3; Table S1). Cost as-
sumptions and lack of proper modeling of the grid-integration challenges are again
the key factors behind the underestimation of solar contribution.

In this perspective, we review assumptions and limitations that have prevented a ma-
jor role of solar PV for some decarbonization scenarios and argue that most of them
are too pessimistic in the light of current technology achievements. First, we discuss

Joule 5, 1041–1056, May 19, 2021 1043

 ll
Perspective

Figure 3. Solar and wind contribution to electricity supply in 2050 obtained with different models
The circles indicate models for Europe, and the squares indicate global models. Gray squares mark
the IAM scenarios included in the IPCC 5 th AR. 8 Red dots correspond to cost-optimal scenarios
obtained with energy models that include sector coupling, uninterrupted hourly resolution, and
network modeling. 27–29 The black line marks the range obtained by Creutzig et al. using an
integrated assessment model with up-to-date PV cost. 12 Black dots indicate results obtained with
the partial-equilibrium models PRIMES 26,30 and TIMES-VTT, 31 as well as results in Jacobson et al. 32
Scenarios, for Europe and globally, predicted by ENTSOE, 33 BNEF, 4 and Shell Sky 34 are also shown.
The scenarios in Victoria et al., 27 Bogdanov et al., 28 and Ram et al. 35 include the cost assumptions
for solar PVs from Vartiainen et al., 36 as shown in Figure 2C.
See also Figures S2–S4.

cost and efficiency evolution, then materials and land limitation, and finally grid and
energy system integration. The final section gathers the main conclusions.

Solar PV is cheap and still have room for learning and reducing costs
The rapid technological evolution of solar PV has made future-cost assumptions
obsolete in most IAM models. Krey et al.37 reviewed fifteen different global and na-
tional IAMs and found that all but one assume capital costs for solar PV plants close
to or higher than 1V/W in 2050, whereas the average cost in 2019 was 0.995 V/W
and the minimum was 0.618 V/W.15,38 The partial-equilibrium model PRIMES also
assumes higher future costs than those predicted by the industry26,30 (see
Figure 2C).

Solar PV modules have maintained a learning rate of 23% since 1976, i.e., their cost
reduces by 23% every time the capacity doubles.39 The main drivers for solar cost
reductions include technological improvements, such as efficiency increase40,41
and those described in Note S1, and high-level mechanisms,41 including economies
of scale, automation, and standardization in manufacturing42,43 (Note S1), and the
existence of a large body of knowledge on silicon materials and devices created
by international research that has been gradually implemented by the indus-
try.42,44,45 The dominant high-level mechanism was different in the various phases
of the learning curve.41 Given that the learning rate is based on module prices, it
also includes the elimination of big parts of the margins in PV manufacturing due
to strong competition between suppliers.45,46 Furthermore, the learning rate

1044 Joule 5, 1041–1056, May 19, 2021

 ll
Perspective

estimated with data from 2007 onward is even higher at 40% (Figure 2B). This recent
learning acceleration is associated with the standardization of the material supply
chain for silicon modules, once the silicon shortage was overcome, and the stan-
dardization of manufacturing tools being made in China, which has allowed higher
throughputs and reduced capital expenditure on equipment.42 Additional cost re-
ductions in the balance of system (BoS) components,36 reliability improvements,
and lifetime extension further decreased the levelized cost of electricity (LCOE)
(Note S2). Access to better financing conditions has also reduced solar LCOE.47 It
is noteworthy that all the previous mechanisms were catalyzed by policies that stim-
ulated market growth through feed-in tariffs or renewable portfolio standards.41
Looking ahead, those key drivers, together with further reductions in soft costs,48
are expected to keep lowering the cost.

The roadmaps for PV technologies39,49 describe diverse strategies to increase effi-

ciency that we briefly summarize here. Silicon solar cells comprise more than 95%
of PV capacity installed in 2019. Passive emitter rear contact (PERC) technology50
has been quickly adopted by the industry (65% of market share in 2019) reaching
an average module efficiency of 20.3%, which is predicted to rise to 22.5% by
2030.39 The record efficiency for silicon solar cells is 26.7% and was attained by using
amorphous silicon-crystalline silicon heterojunction (HIT) and interdigitated back
contacts (IBCs).50,51 The latter increases efficiency by eliminating the shadow of
the front contact. Average HIT module efficiency is at 21% and the industry roadmap
predicts attaining 24% in 2030.39 Tunnel-oxide passivated contact (TOPCon) could
also provide an additional percentage increase in efficiency in the short term.49 The
quick learning of perovskite solar cells could enable an additional route for efficiency
increase once the stability is improved.52,53 When combined with silicon in a double-
junction solar cell, 29.1% efficiency has been reported,51 enabling a commercializa-
tion pathway toward 30% efficient perovskite/c-Si tandem modules.53 Finally, multi-
junction solar cells based on III-V semiconductors are the PV devices attaining the
highest efficiency. The record six-junction solar cell achieves 47.1% efficiency at
143 suns by converting different parts of the spectrum into electricity.51,54 Multijunc-
tion solar cells are used in space applications and can be combined with concen-
trating systems to generate electricity on the ground if significant cost reduction is
achieved for such systems.55

Industrial innovations in the pipeline include pushing toward larger cells and mod-
ules, cutting them in half or more to reduce electrical losses,42 and using multi-
wire to reduce resistance and shadow losses, while at the same time reducing the
use of silver.56 Moreover, existing developments outside of standard PV modules
show that there is additional room for innovation. Single-axis tracking, which ac-
counts for 30% of the market today,39 increases the energy yield by boosting elec-
tricity generation in the early morning and late evening when it matches electricity
demand.57 Bifacial solar cells,58 which convert irradiance reaching both sides of
the panel into electricity, account today for 20% of the market and are expected
to gain market share.39 Combined with single-axis tracking, they can achieve very
competitive costs in most parts of the world.59 On top of module cost, BoS and
soft costs impact the total system cost (Note S2). BoS comprises wiring, mounting,
and other area-dependent components, which will also benefit from an increase in
module efficiency.36,60 BoS also includes the inverter required to convert DC power
generated by the solar panels into AC power and the grid connection. Inverters
traditionally dimensioned with a DC/AC ratio of ~1.2 are shifting toward higher ra-
tios, i.e., clipping part of the DC power in hours with high irradiance is worthwhile

Joule 5, 1041–1056, May 19, 2021 1045

 ll
Perspective

given that the utilization of the inverter and grid-connection capacity is increased.
The cost-optimal DC/AC ratio can be as high as 1.7 for certain locations.61,62

As the module and BoS costs shrink, decreasing soft costs also becomes crucial. Soft
costs, which include financing, planning, permitting, inspection, and interconnec-
tion (PPI), as well as land acquisition, installation labor, and profit, show a wide range
among different regions15,38,48 indicating that there is room for cost reductions as
solar PV diffuses locally. The perception of solar PV as a mature technology and auc-
tions, which have become the preferred method for governments to select new solar
capacity worldwide,38 decrease the risk and the financial costs, which in turn reduces
the cost of solar electricity.36 Improvements in maintenance and lifetime will also
lower the cost. The efficiency of crystalline silicon modules degrades by ~0.5%/
year.63 This might enable lifetimes longer than the 25 years traditionally assumed.
Because the average age of solar plants is still low (Figure 2E), research in degrada-
tion and maintenance strategies for solar modules and BoS materials64 will be key to
extending their lifetimes. Under low investment costs, operation and maintenance
(O&M) become increasingly important and can account for 25% of the life cycle costs
in solar power plants.65 The existence of a high learning rate for O&M costs, esti-
mated at 18% in Germany,65 means that PV plants hold cost-reduction potential
even after commissioning. Application of AI to diagnose operational status and fail-
ure of PV systems is expected to reduce O&M costs. The poor representation of pol-
icy-driven demand has also been identified as a potential limitation in IAMs.12
Agent-based modeling has been proposed66,67 to study PV diffusion and consumer
choices, overcoming some of the limitations of IAMs and partial-equilibrium models.

We conclude this section by providing some suggestions. The previous experience

and the innovations in the pipeline showed that solar PV is a dynamic, fast-evolving
technology, and this should be reflected in the models. When modeling future en-
ergy scenarios, up-to-date cost assumptions must be included. The PV industry
keeps up-to-date records on the learning rate and cost evolution of the technol-
ogy.15,38,39 We recommend including an endogenous learning curve that captures
the learning effects and conducting sensitivity analyses to investigate the impact
that the assumed learning and growth rates can have on the results. For instance,
Creutzig et al.12 found that implementing this strategy in REMIND, a specific IAM,
resulted in solar PV covering 30%–50% of global electricity demand in 2050
(compared with 5%–17% share in previous results68).

Materials and land-use limitation

The availability of raw materials is not a real issue that limits the growth of PV
manufacturing. Materials availability could be a constraint for thin-film technol-
ogy69,70 but not for silicon solar cells, which account for 95% of the market today.71
Silicon is one of the most abundant materials on Earth. In 2019, the PV industry
consumed 10% of global silver demand72 which is used to manufacture the cell
contacts. Thanks to the increase in efficiency and the use of thinner contact fingers,
the use of silver per watt has significantly reduced in the last years (Figure 2F), and
copper or aluminum could be used as a replacement if necessary.39,43 The noncell
materials in PV (glass, plastic, aluminum, concrete, and steel) are not expected to
represent a limit either.70 Fortunately, incorrect assumptions on material constraints
are not common among IAMs, but there are some exceptions, e.g., an outdated
silver consumption of 133 mg/W, which is more than five times higher than current
value, is assumed in the analysis73 associated with one IAM. The PV industry is
already on the way to a circular economy. As a standard practice today, PV modules
run through existing glass- or metal-recycling lines where the aluminum frame,

1046 Joule 5, 1041–1056, May 19, 2021

 ll
Perspective

external copper wire, and glass are recovered, which represents a mass-recovery
rate that is higher than 75%.74 Processes whose technical potential enable recovery
rates of 94% for silver and 97% for silicon have been described.74,75 The interest to
industrialize those processes will increase as the large capacity installed in very
recent years reaches its end of life (Figure 2E).

Supplying the current global electricity consumption with solar PV would imply
covering 0.3% of the land area of the world (Note S3). PV modularity enables resi-
dential and commercial installations, as well as utility-scale power plants. In the
former, solar panels can be installed on rooftops and facades potentially supplying
more than 25% of electricity demand in most of the USA states76 and Europe.77
Utility-scale plants can be installed in dry areas that are not suitable for agriculture,
although their ecological and social effects must be considered.78 Furthermore,
other strategies have recently emerged. In floating PV, solar panels are placed on
water bodies, such as calm equatorial seas or hydro reservoirs,79,80 enabling the res-
ervoirs to act as virtual batteries.81 Agrivoltaics, i.e., combining agriculture land and
PV, can offer additional benefits.82 In dry areas, soil humidity beneath the panel in-
creases, reducing the need for irrigation.83 In windy regions, vertical solar panels
could act as artificial trees reducing the wind and increasing crop yields. Sparse
strips of solar panels installed on agriculture land can serve as biodiversity reser-
voirs.84 Other dual-use of infrastructures such as irrigation channels or sound barriers
have also been proposed.85 Of course, the forecasted PV development, land cover,
and cost projections are not based on these PV configurations but on classic densely
packed utility-scale power plants. Nevertheless, these embryonic applications show
that there is still room for innovation at the system level. In summary, although avail-
able land can limit solar PV at local levels, it will not be a limitation at a larger scale,
and therefore, we recommend that models include accurate and up-to-date con-
straints based on materials and land availability. The increase in efficiency and the
strategies to improve annual energy yield described above will help to address
the land-limitation issues in specific locations.

Grid and energy system integration of high shares of solar electricity

One of the main factors that hinders the role of solar PV in decarbonization pathways
envisaged by IAMs, or partial-equilibrium models such as PRIMES (Figures 3 and S2–
S4; Table S1), is related to the fact that such models use annual resolution. To
circumvent the limitation of not modeling the supply and demand in every hour,
additional constraints are imposed, which might be too pessimistic. Some models
include integration costs, which are added on top of technology costs, and which
can reach 23 $/MWh for solar and 37$/MWh for wind in low-penetration scenarios
(<20% of demand)68 and increase to above 100$/MWh for higher penetrations.86
Assuming high integration costs, despite the existence of alternative, cheaper
balancing strategies, hampers the role of solar PV in cost-optimized results. In other
cases, an exogenous constraint limits the maximum share of wind and solar power in
electricity generation, e.g., 30% limitation is used in some models.12,68 Pietzcker
et al.24 have assessed different strategies to improve the representation of integra-
tion costs in IAM. The average electricity supplied by wind and solar PV increased
from a model-averaged of 38% to 62%, but the new, simplified representations of
integration cost might still be over constraining the results. Some IAMs have started
to conduct harmonization exercises with hourly resolved models of the power
sector87,88 and investigated high electrification pathways,89 which could provide
useful insights. The trade-offs of reducing time resolution have been analyzed by
the energy modeling community90–93 and insights could be incorporated to IAMs.

Joule 5, 1041–1056, May 19, 2021 1047

 ll
Perspective

Table 1. Challenges and strategies for high solar PV penetration in the grid at different timescales
Timescale Interannual Seasonal Hours Minutes-sub seconds
Challenge solar generation interannual significant seasonal difference in balancing the strong daily lower inertia in the power system is
variability is low. PV generation only for high- generation pattern. a challenge, but it does not limit
climate change impact on solar latitude locations.a the maximum wind and solar
generation expected to be not significant for most of the world penetration in the grid.
low.101,102 population living in the Sun Belt.
Strategy robust design and operation of - balance with wind, which - electric batteries - grid-forming inverters
energy systems considering shows opposite seasonal - pumped hydro storage - synchronous condensers
interannual variability dependence103 - demand-side management - batteries providing synthetic
- reinforce interconnections - additional flexibility brought inertia
among neighboring countries by sector coupling - pumped hydro storage
- sector coupling provides - EVs with demand-side man- - PV and wind can contribute to
higher demand flexibility and agement and power-to-grid upward regulation (if oper-
lower storage costs in other capabilities ated at reduced capacity) and
sectors - high temporal correlation of downward regulation (if
- long-term storage (e.g., cooling demand and solar curtailed)
power-to-X) generation
- interconnections over large
areas to smooth out local
weather
a
This is the only challenge showing high complexity. The others show low complexity.

The share of global electricity supplied by solar PV in 2050 average to 10.6%/19.6%

in IAM scenarios included in the IPCC 5thAR8 / SR1.59 (Figures 3, S3, and S4).
Although some scenarios include higher shares, historical growth rates, combined
with the issues discussed above, suggest that significant underestimation might still
be a concern. Another key factor is that most of those IAM scenarios include a limited
direct or indirect electrification of other sectors. The contribution of solar electricity
to primary energy in 2050 averages to 3.1%/6.8% in the IPCC 5thAR8/SR1.59 (Figures
S2 and S4). Conversely, several analyses based on sector-coupled energy modeling
approaches27–29,94–96 found that deep electrification of other sectors (directly or
indirectly by, e.g., synthetic fuels) is a cost-effective strategy to enable timely
decarbonization.

Next, we first review the existing challenges and strategies to balance PV generation
at different timescales. We then discuss how the proper modeling of all the existing
strategies for balancing enables the emergence of solar PV as one of the main
energy sources in cost-optimal decarbonized scenarios.

Matching variable renewable generation and demand entails challenges at different

timescales. We focus here on those particularly relevant for solar PV (Table 1), but we
keep in mind that onshore and offshore wind generation, and its associated vari-
ability, will also be a significant contributor in future power systems. We start by dis-
cussing seasonal variation. For high-latitude locations, solar generation in winter and
in summer show large differences. Fortunately, wind generation typically shows the
opposite seasonal trend.97 Furthermore, most of the world’s population lives in the
Sun Belt close to the equator where the solar resource is abundant and seasonal vari-
ation is low (see Figure 4). Tropical locations, despite repeating weather patterns
such as monsoon, show low seasonal variation in solar resources.98 Moving now to
the hourly balancing, the strong diurnal solar-generation pattern produces an excess
of generation in the middle of the day and requires ramping up balancing technol-
ogies as solar generation vanishes after sunset. Power systems are already prepared
to deal with hourly variation in electricity demand, so the initial solar deployment
(demand being covered by solar <20%) can be easily integrated into the system
without further modifications. For higher penetrations, short-term storage with

1048 Joule 5, 1041–1056, May 19, 2021

 ll
Perspective

Figure 4. Relative seasonal variation in solar resource-

For every location, relative variation is calculated as the 80 percentile of daily insolation on the
horizontal surface in the sunniest month minus the 80 percentile of daily insolation in the least sunny
month, divided by the average daily insolation. Original plot with satellite data from NASA. 104 The
depicted seasonal variation in solar resource is an upper limit for the variation that can be expected
in solar power generation because of local variation in the utilization of direct and diffuse radiation.

high-efficiency, i.e., electric batteries, pumped hydro storage (PHS),99 and demand-
side management contribute to energy arbitrage to ease the intraday balancing of
solar PV-.100

Finally, it is well understood that solar does not contribute mechanical inertia to the
grid stability in the minute-subsecond timescale. Several strategies exist to deal with
the reduced system inertia caused by increasing variable renewable penetration.
First, grid-forming inverters can regulate the system voltages and frequency through
local, decentralized control.105 Second, the use of synchronous condensers to pro-
vide reactive power and inertia has been successfully implemented in Denmark and
Australia, and conventional power plants can be retrofitted into synchronous con-
densers.103 Solar can contribute to upward regulation (when operating at reduced
capacity) and to downward regulation by curtailing. Batteries can also be operated
to contribute to ensuring grid stability.106–108 Pumped hydro can provide both me-
chanical inertia and rapid response (in tenths of seconds). It is interesting to note that
South Australia recently operated for an hour with 100% PV electricity,109 and
already in 2015, Denmark’s power system was operated without dispatching primary
central power stations for several consecutive days in which wind supplied most of
the electricity demand.103 Frew et al.110 showed that, with appropriate changes to
grid operation, 55% PV penetration could be achieved in the US power system in
2050, while ensuring resource adequacy, addressing net-load variability, and
providing sufficient operating reserves. In summary, the existing field experience
and literature105,111 do not indicate a maximum wind and solar penetration above
which the system’s operation is not feasible in the sub hourly timescale.

Besides the power system, solar PV can significantly contribute to decarbonizing

other sectors while benefiting from the additional flexibility provided by sector
coupling. Rooftop PV enables charging electric vehicles (EVs) reducing the need
for distribution-grid reinforcement. The short-term storage provided by EV batteries
synergizes with PV, easing the daily balancing of solar generation,112 in particular
when demand-side management and vehicle-to-grid capabilities are assumed for

Joule 5, 1041–1056, May 19, 2021 1049

 ll
Perspective

EVs.94,100 EVs can also be considered a granular technology3; their short renovation
times enable mutual benefits from a fast PV deployment in the next decade. PV can
also contribute to decarbonizing the residential and services sector by successfully
competing with gas for low-temperature space heating and water heating with the
assistance of electric heat pumps. Solar can also supply refrigeration demand, which
is highly correlated in time with solar generation113,114 and is expected to increase
due to climate change. As solar becomes cheaper, it begins to compete with gas
and coal for industrial heating through the use of high-efficiency electric furnaces.
This can be coupled with low-cost on-site thermal storage in water, hot rocks, molten
salts, molten silicon, and other materials to take advantage of low-cost daytime solar
electricity.

Moreover, low-cost solar electricity could be used to produce electrolytic hydrogen

that substitutes hydrogen from steam-methane reforming. This could reduce emis-
sions associated with ammonia production and fertilizers in the agricultural sector
and cover the entire chemical industry with related green methanol as the new
bulk chemical.115 Electrolytic hydrogen could also curb emissions in steel
manufacturing by shifting to direct-reduced iron. Moreover, hydrogen can also be
used as fuel for long-distance road transport and shipping116,117 or used to supply
high-temperature heat in the industry. In the Sabatier reaction, hydrogen can be
combined with direct-air-captured CO2118 to produce synthetic methane, ensuring
system operation in the critical weeks with low renewable generation and high heat-
ing demand.94 It can also be combined with additional CO2 in the Fischer-Tropsch
reaction to produce synthetic hydrocarbons for aviation or to be used as feedstock
in the chemicals industry. The flexibility provided by electrolyzers can not only stabi-
lize the grid119 but might also enable a most effective balancing of the entire energy
system as indicated in the case of Europe.94,100,35

Before concluding this section, we should highlight that solar PV emerges as one
of the main, if not the main, energy sources in cost-optimal future decarbonized
scenarios when using models that include proper representation of all the afore-
mentioned balancing strategies. i.e., when the models (1) use uninterrupted hour-
ly modeling for a whole year, (2) use time series that represent solar generation at
a country level by aggregating data with high spatial resolution (~40 km2),120 (3)
allow the exchange of electricity among countries, (4) and include detailed
coupling of the power system with other sectors such as heating, transport, or in-
dustry. For instance, Victoria et al.27 and Child et al.29 agree in estimating the
share of solar electricity by ~45% in Europe in 2050. Energy models with high
spatial resolutions (dozens of nodes per country) have also predicted a prominent
role of solar PV.121–124 Although the large seasonality in heating demand in
Europe, opposed to solar seasonal variation, limits the solar PV penetration in
that region, it is important to acknowledge that most of the global population,
expected population growth, and energy demand growth are located at lower
latitudes where the solar resource and energy demand show low seasonal
variations (Figure 4). Bogdanov et al.28 and Ram et al.125 found that the optimal
solar share increases to 68% for a global analysis not only for the power sector
but also the entire energy system (Table S1).

Finally, historical field experience indicates that solar PV penetration is increasing

rapidly. In 2010, no large power system existed in which solar PV supplied more
than 3% of the annual demand. In 2019, solar PV supplied 9% of electricity demand
in Germany and 19% in California (Figure 5). Existing plans contemplate penetration
higher than 20% in several power systems by 2030.

1050 Joule 5, 1041–1056, May 19, 2021

 ll
Perspective

Figure 5. Percentage of electricity demand covered by solar PV in different markets worldwide

The circles’ sizes are proportional to the annual electricity demand in 2019 in every region. Table S2
gathers references for historical values and planned expansion.

Conclusions
Solar PV is ready to become one of our main energy sources based on the argu-
ments provided in this perspective: (1) learning and cost reductions are expected
to continue, (2) neither materials nor land use will prevent PV expansion, and (3)
existing integration strategies and those under development will allow large
penetration of solar PV not only in the power grid but in the entire energy system.
Some IAMs and partial-equilibrium models used to investigate alternative transi-
tion paths do not include proper representation of those strategies and use
outdated technology costs. Consequently, they miss the large mitigation poten-
tial of solar PV and describe the 1.5 C-compatible scenarios that rely on a tech-
nology mix that will most likely be more expensive and over-reliant on far less
mature technologies. A critical assessment of the results provided by those
models needs to pay attention to cost assumptions, time resolution, and whether
state-of-the-art modeling of sector coupling and renewable balancing strategies
is used.

Failing to identify the prominent role that solar PV will play in a future climate-neutral
energy system weakens the communication of an important message: PV technol-
ogy is ready to ramp up fast and contribute to mitigating emissions by 2030, which
will be key to remain on a path compatible with the Paris Agreement.1 Installation
times are shorter for solar PV than for other conventional technologies and PV modu-
larity is again a benefit for ramping up. For developed countries, rooftop PV systems
owned by citizens and small companies not only increase awareness, but they can

Joule 5, 1041–1056, May 19, 2021 1051

 ll
Perspective

raise additional investments for the energy transition. Proper policy interventions
and business models can ensure that rooftop PV also diffuses among low- and mod-
erate-income households.126 For less developed countries, solar PV could be used in
solar home systems or microgrids to provide electricity to the 860 million people
who still live without it.127,128 The modularity of the technology allows consumers
to gradually climb the ladder of electricity access by extending their consumption
as their income increases, from tier 1 (electricity for task lighting and phone
charging) to tier 5 (use of high-power appliances).129

Throughout the last decade, global solar electricity generation maintained a 50%
annual growth, rising from ~12 TWh in 2008 to ~880 TWh in 2018. Keeping a 50%
annual growth for 9 additional years would mean producing ~34,000 TWh (more
than the global electricity demand in 2019, which accounted for ~27,000TWh2).
This highlights the large potential for solar PV expansion. Adequate policies are
needed to ensure that high installation rates are maintained in the short term and
that the potential environmental, social, and economic benefits of solar PV are effec-
tively materialized. We identify the following challenges for sustained scaling up of
solar PV in the next decade: ensuring adequate regulatory frameworks that reduce
soft costs, reducing capital expenditure via industrial innovations,130 untapping
the demand for PV by enabling the electrification of other energy sectors assisted
by proper tax schemes, and strengthening research on improving efficiency and reli-
ability of PV systems. Certainly, developing PV is not the only action needed to limit
temperature increase, but future scenarios must properly capture the mitigation po-
tential of this technology.

EXPERIMENTAL PROCEDURES
Resource availability
Lead contact
Further information related to the data and code should be directed to the Lead
Contact, Marta Victoria ([email protected]).

Materials availability
No materials were used in this study.

Data and code availability

The data and code to produce the figures in the main paper and Supplemental
Materials can be retrieved from the repository https://github.com/martavp/pv-is-
ready.

SUPPLEMENTAL INFORMATION
Supplemental Information can be found online at https://doi.org/10.1016/j.joule.
2021.03.005.

ACKNOWLEDGMENTS
This article resulted from input associated with the ‘‘100% renewable energies’’
session at the 47th IEEE PVSC Conference, June 2020. Arnulf Jäger-Waldau works
at the European Commission - Joint Research Centre (JRC), Ispra, Italy. The views
expressed are purely those of the author and may not in any circumstances be
regarded as stating an official position of the European Commission. This work
was authored in part by the National Renewable Energy Laboratory, operated by
Alliance for Sustainable Energy, LLC, for the U.S. Department of Energy (DOE) under
contract number DE-AC36-08GO28308. The views expressed in the article do not

1052 Joule 5, 1041–1056, May 19, 2021

 ll
Perspective

necessarily represent the views of the DOE or the U.S. Government. The U.S. gov-
ernment retains and the publisher, by accepting the article for publication, acknowl-
edges that the U.S. government retains a nonexclusive, paid-up, irrevocable, world-
wide license to publish or reproduce the published form of this work, or allow others
to do so, for U.S. government purposes. This work was supported by the Bavarian
State Government (project ‘‘PV-Tera – Reliable and cost efficient photovoltaic power
generation on the terawatt scale,’’ no. 44-6521a/20/5). M.V. is partially funded by
the RE-INVEST project, which is supported by the Innovation Fund Denmark under
grant number 6154-00022B, and the HYPERFARM project funded by the European
Commission-H2020 under grant agreement number 101000828. C.d.C. acknowl-
edges the partial support of the Spanish Ministerio de Ciencia, Innovación y Univer-
sidades under the project MOBISOL ENE2017-89561-C4-2-R and the Comunidad
de Madrid under the project Madrid-PV2 (S2018/EMT-4308).

The responsibility for the contents lies solely with the authors.

AUTHOR CONTRIBUTIONS
Conceptualization, M.V., N.H., I.M.P., R.S., A.J.-W., C.d.C., C.B., M.S., and A.B;
project administration, M.V.; visualization, M.V. and I.M.P.; writing–original draft,
M.V.; writing–review & editing, M.V., N.H., I.M.P., R.S., A.J., C.d.C., C.B., M.S.,
A.B., I.K., K.K., and A.S.

DECLARATION OF INTERESTS
The authors declare no competing interests..

REFERENCES
1. Figueres, C., Schellnhuber, H.J., Whiteman, 9. IPCC (2019). Global warming of 1.5 C, 15. IRENA (2020). Renewable power generation
G., Rockström, J., Hobley, A., and Rahmstorf, intergovernmental panel on climate change costs in 2019. https://www.iea.org/reports/
S. (2017). Three years to safeguard our (IPCC). https://www.ipcc.ch/sr15/ Data renewables-2020/solar-pv#abstract.
climate. Nature 546, 593–595. esemble with scenarios results. https://data.
ene.iiasa.ac.at/iamc-1.5c-explorer/. 16. Philipps, S., Kost, C., and Schlegl, T. (2014).
2. B Statistical review of world energy (2020). BP Up-to-date levelised cost of electricity in
2020. https://www.bp.com/en/global/ 10. Rogelj, J., Popp, A., Calvin, K.V., Luderer, G., photovoltaics. background from Fraunhofer
corporate/energy-economics/statistical- Emmerling, J., Gernaat, D., Fujimori, S., ISE relating to IPCC WGIII 5th assessment
review-of-world-energy.html. Strefler, J., Hasegawa, T., Marangoni, G., et al. report, final draft, September 2014. http://
(2018). Scenarios towards limiting global www.go100re.net/wp-content/uploads/2014/
3. Wilson, C., Grubler, A., Bento, N., Healey, S., mean temperature increase below 1.5  C. 10/2014_10_24_Statement_LCOE_PV_
De Stercke, S.D., and Zimm, C. (2020). Nat. Clim. Change 8, 325–332. Fraunhofer_ISE.pdf.
Granular technologies to accelerate
decarbonization. Science 368, 36–39. 11. van Vuuren, D.P., Stehfest, E., Gernaat, 17. IPCC (2021). IPCC sixth assessment report.
D.E.H.J., van den Berg, M., Bijl, D.L., de Boer, https://www.ipcc.ch/assessment-report/ar6/.
4. BNEF (2020). New energy outlook 2020. H.S., Daioglou, V., Doelman, J.C.,
https://about.bnef.com/new-energy- Edelenbosch, O.Y., Harmsen, M., et al. (2018). 18. McLaren, D., and Markusson, N. (2020). The
outlook/. Alternative pathways to the 1.5  C target co-evolution of technological promises,
reduce the need for negative emission modeling, policies and climate change
5. IEA (2020). World energy outlook. https:// technologies. Nat. Clim. Change 8, 391–397. targets. Nat. Clim. Chang. 10, 392–397.
www.iea.org/reports/world-energy-outlook- 12. Creutzig, F., Agoston, P., Goldschmidt, J.C., 19. Gambhir, A. (2019). Planning a low-carbon
2020. Luderer, G., Nemet, G., and Pietzcker, R.C. energy transition: what can and can’t the
(2017). The underestimated potential of solar models tell us? Joule 3, 1795–1798.
6. Haegel, N.M., Atwater, H., Barnes, T., Breyer, energy to mitigate climate change. Nat.
C., Burrell, A., Chiang, Y.M., De Wolf, S.D., Energy 2, nenergy2017140. 20. Gambhir, A., Butnar, I., Li, P.-H., Smith, P., and
Dimmler, B., Feldman, D., Glunz, S., et al. Strachan, N. (2019). A review of criticisms of
(2019). Terawatt-scale photovoltaics: 13. Breyer, C., Bogdanov, D., Gulagi, A., integrated assessment models and proposed
transform global energy. Science 364, Aghahosseini, A., Barbosa, L.S.N.S., approaches to address these, through the
836–838. Koskinen, O., Barasa, M., Caldera, U., lens of BECCS. Energies 12, 1747.
Afanasyeva, S., Child, M., et al. (2017). On the
7. International Energy Agency (IEA) (2020). role of solar photovoltaics in global energy 21. Creutzig, F., Breyer, C., Hilaire, J., Minx, J.,
Renewables 2020. https://www.iea.org/ transition scenarios. Prog. Photovolt.: Res. Peters, G.P., and Socolow, R. (2019). The
reports/renewables-2020/solar-pv#abstract. Appl. 25, 727–745. mutual dependence of negative emission
technologies and energy systems. Energy
8. IPCC (2020). IPCC Fifth Assessment Report, 14. Kurtz, S.R., Leilaeioun, A.M., King, R.R., Environ. Sci. 12, 1805–1817.
2014, Data ensemble with scenarios results: Peters, I.M., Heben, M.J., Metzger, W.K., and
https://www.ipcc.ch/assessment-report/ar5/ Haegel, N.M. (2020). Revisiting the terawatt 22. Breyer, C., Fasihi, M., Bajamundi, C., and
https://tntcat.iiasa.ac.at/AR5DB/. challenge. MRS Bulletin 45, 159–164. Creutzig, F. (2019). Direct air capture of CO2:

Joule 5, 1041–1056, May 19, 2021 1053

 ll
Perspective

a key technology for ambitious climate TYNDP_2020_Scenario_Report_entsog- challenge in U.S. small-scale solar PV system
change mitigation. Joule 3, 2053–2057. entso-e.pdf. pricing. Energy Policy 134, 110956.

23. Luderer, G., Pietzcker, R.C., Carrara, S., de 34. Shell. (2020). Sky Scenarios. https://www.shell. 49. Wilson, G.M., Al-Jassim, M., Metzger, W.K.,
Boer, H.S., Fujimori, S., Johnson, N., Mima, S., com/energy-and-innovation/the-energy- Glunz, S.W., Verlinden, P., Xiong, G.,
and Arent, D. (2017). Assessment of wind and future/scenarios/shell-scenario-sky.html. Mansfield, L.M., Stanbery, B.J., Zhu, K., Yan,
solar power in global low-carbon energy Y., et al. (2020). The 2020 photovoltaic
scenarios: an introduction. Energy Econ 64, 35. Solar Power Europe (2020). 100% renewable technologies roadmap. J. Phys. D: Appl. Phys.
542–551. Europe. https://www.solarpowereurope.org/ 53, 493001.
100-renewable-europe/.
24. Pietzcker, R.C., Ueckerdt, F., Carrara, S., de 50. Blakers, A.W., Wang, A., Milne, A.M., Zhao, J.,
Boer, H.S., Després, J., Fujimori, S., Johnson, 36. Vartiainen, E., Masson, G., Breyer, C., Moser, and Green, M.A. (1989). 22.8% efficient silicon
N., Kitous, A., Scholz, Y., Sullivan, P., and D., and Román Medina, E.R. (2020). Impact of solar cell. Appl. Phys. Lett. 55, 1363–1365.
Luderer, G. (2017). System integration of wind weighted average cost of capital, capital
expenditure, and other parameters on future 51. Green, M.A., Dunlop, E.D., Hohl-Ebinger, J.,
and solar power in integrated assessment
utility-scale PV levelised cost of electricity. Yoshita, M., Kopidakis, N., and Hao, X. (2020).
models: a cross-model evaluation of new
Prog. Photovolt.: Res. Appl. 28, 439–453. Solar cell efficiency tables, (version 56). Prog
approaches. Energy Econ 64, 583–599.
photovolt: Res Appl. 28, 629–638.
25. European Commission (2018). In-depth 37. Krey, V., Guo, F., Kolp, P., Zhou, W., Schaeffer,
R., Awasthy, A., Bertram, C., de Boer, H.-S., 52. Jung, E.H., Jeon, N.J., Park, E.Y., Moon, C.S.,
analysis in support on the COM(2018) 773: a Shin, T.J., Yang, T.Y., Noh, J.H., and Seo, J.
clean planet for all - a european strategic Fragkos, P., Fujimori, S., et al. (2019). Looking
under the hood: a comparison of techno- (2019). Efficient, stable and scalable
long-term vision for a prosperous, modern, perovskite solar cells using poly(3-
competitive and climate neutral economy. economic assumptions across national and
global integrated assessment models. Energy hexylthiophene). Nature 567, 511–515.
https://knowledge4policy.ec.europa.eu/
publication/depth-analysis-support- 172, 1254–1267.
53. Sahli, F., Werner, J., Kamino, B.A., Bräuninger,
com2018-773-clean-planet-all-european- M., Monnard, R., Paviet-Salomon, B., Barraud,
strategic-long-term-vision_en. 38. Masson, G., and Kaizuka, I. (2020). Trends in
photovoltaic applications 2019 (IEA PVPS). L., Ding, L., Diaz Leon, J.J., Sacchetto, D.,
https://iea-pvps.org/wp-content/uploads/ et al. (2018). Fully textured monolithic
26. Impact Assessment Accompanying the EU perovskite/silicon tandem solar cells with
Commission Communication Stepping up 2020/02/5319-iea-pvps-report-2019-08-lr.pdf.
25.2% power conversion efficiency. Nat.
Europe’s 2030 Climate Ambition. (2020). Mater. 17, 820–826.
39. International Technology Roadmap for
https://ec.europa.eu/clima/sites/clima/files/
Photovoltaic (ITRPV) 11th Ed. 2019 results.
eu-climate-action/docs/impact_en.pdf. 54. Geisz, J.F., France, R.M., Schulte, K.L., Steiner,
https://itrpv.vdma.org/documents/
M.A., Norman, A.G., Guthrey, H.L., Young,
27. Victoria, M., Zhu, K., Brown, T., Andresen, 27094228/29066965/ITRPV%302020.pdf/
M.R., Song, T., and Moriarty, T. (2020). Six-
G.B., and Greiner, M. (2020). Early ba3da187-3186-83de-784e-6e3b10d96f3f.
junction III–V solar cells with 47.1% conversion
decarbonisation of the European energy efficiency under 143 Suns concentration. Nat.
40. Nemet, G.F. (2006). Beyond the learning
system pays off. Nat. Commun. 11, 6223. Energy 5, 326–335.
curve: factors influencing cost reductions in
28. Bogdanov, D., Farfan, J., Sadovskaia, K., photovoltaics. Energy Policy 34, 3218–3232. 55. Wiesenfarth, M., Philipps, S.P., and Bett, A.W.
Aghahosseini, A., Child, M., Gulagi, A., (2017). Current status of concentrator
41. Kavlak, G., McNerney, J., and Trancik, J.E.
Oyewo, A.S., Barbosa, L.de S.N.S., and photovoltaic (CPV) technology. https://www.
(2018). Evaluating the causes of cost reduction
Breyer, C. (2019). Radical transformation nrel.gov/docs/fy16osti/65130.pdf.
in photovoltaic modules. Energy Policy 123,
pathway towards sustainable electricity via
700–710. 56. Ballif, C., Hessler-Wyser, A., Baumgartner, Y.,
evolutionary steps. Nat. Commun. 10, 1077.
Cattin, J., Cattaneo, B., Strahm, B., Papet, P.,
42. Chen, Y., Altermatt, P.P., Chen, D., Zhang, X.,
29. Child, M., Kemfert, C., Bogdanov, D., and Ufheil, J., Gragert, M., Grischke, R., et al.
Xu, G., Yang, Y., Wang, Y., Feng, Z., Shen, H.,
Breyer, C. (2019). Flexible electricity (2014). SmartWire solar cell interconnection
and Verlinden, P.J. (2018). From laboratory to
generation, grid exchange and storage for technology (29th European Photovoltaic Solar
production: learning models of efficiency and
the transition to a 100% renewable energy Energy Conference and Exhibition), pp. 2555–
manufacturing cost of industrial crystalline
system in Europe. Renew. Energy 139, 80–101. 2561.
silicon and thin-film photovoltaic
technologies. IEEE J. Photovoltaics 8, 1531– 57. Afanasyeva, S., Bogdanov, D., and Breyer, C.
30. European Commission (2018). In-depth
1538. (2018). Relevance of PV with single-axis
analysis in support of the Commission
Communication COM (2018) 773 A clean tracking for energy scenarios. Sol. Energy 173,
43. Verlinden, P.J. (2020). Future challenges for 173–191.
planet for all. A European long-term strategic photovoltaic manufacturing at the terawatt
vision for a prosperous, modern, competitive level. J. Renew. Sustain. Energy 12, 053505.
and climate neutral economy. https:// 58. Luque, A., Eguren, J., and Del Alamo, J.
knowledge4policy.ec.europa.eu/publication/ (1978). Internal quantum efficiency of back
44. Altermatt, P.P., Yang, Y., Chen, Y., Zhang, X., illuminated n+ pp+ solar cells. Rev. Phys.
depth-analysis-support-com2018-773-clean- Chen, D., Xu, G., and Feng, Z. (2020).
planet-all-european-strategic-long-term- Appl. (Paris) 13, 629–632.
Requirements of the Paris climate agreement
vision_en. for the coming 10 years on investments, 59. Rodrı́guez-Gallegos, C.D., Liu, H., Gandhi, O.,
technical roadmap, and expansion of PV Singh, J.P., Krishnamurthy, V., Kumar, A.,
31. Pursiheimo, E., Holttinen, H., and Koljonen, T. manufacturing (37th European Photovoltaic Stein, J.S., Wang, S., Li, L., Reindl, T., and
(2019). Inter-sectoral effects of high Solar Energy Conference and Exhibition), Peters, I.M. (2020). Global techno-economic
renewable energy share in global energy pp. 1999–2004. performance of bifacial and tracking
system. Renew. Energy 136, 1119–1129.
photovoltaic systems. Joule 4, 1514–1541.
45. Green, M.A. (2019). How did solar cells get so
32. Jacobson, M.Z., Delucchi, M.A., Bauer, Z.A.F., cheap? Joule 3, 631–633. 60. Peters, I.M., Rodriguez Gallegos, C.D., Sofia,
Goodman, S.C., Chapman, W.E., Cameron, S.E., and Buonassisi, T. (2019). The value of
M.A., Bozonnat, C., Chobadi, L., Clonts, H.A., 46. Chase, J. (2019). Solar Power Finance Without efficiency in photovoltaics. Joule 3, 2732–
Enevoldsen, P., et al. (2017). 100% Clean and the Jargon (World Scientific). 2747.
renewable wind, water, and sunlight all-sector
energy roadmaps for 139 countries of the 47. Egli, F., Steffen, B., and Schmidt, T.S. (2018). A 61. Oversizing White Paper, SMA. https://www.
world. Joule 1, 108–121. dynamic analysis of financing conditions for sma.de/fileadmin/content/global/specials/
renewable energy technologies. Nat. Energy documents/oversizing/Whitepaper_
33. ENTSOE (2019). Ten years network 3, 1084–1092. Oversizing_EN_180530_01.pdf.
development plan (TYNDP) 2020, scenario
report. https://www.entsos-tyndp2020- 48. O’Shaughnessy, E., Nemet, G.F., Pless, J., and 62. Fluence. (2018). How to optimize your inverter
scenarios.eu/wp-content/uploads/2019/10/ Margolis, R. (2019). Addressing the soft cost loading ratio for solar + energy storage

1054 Joule 5, 1041–1056, May 19, 2021

 ll
Perspective

projects. https://blog.fluenceenergy.com/ European Union. Renew. Sustain. Energy Rev. cheapest: increasing electrification of global
solar-energy-storage-how-to-optimize-your- 114, 109309. energy use on the road to climate neutral
inverter-loading-ratio. energy systems.
78. Grodsky, S.M., and Hernandez, R.R. (2020).
63. Jordan, D.C., Kurtz, S.R., VanSant, K., and Reduced ecosystem services of desert plants 90. Pfenninger, S. (2017). Dealing with multiple
Newmiller, J. (2016). Compendium of from ground-mounted solar energy decades of hourly wind and PV time series in
photovoltaic degradation rates. Prog. development. Nat. Sustainability 3, 1036– energy models: a comparison of methods to
Photovolt.: Res. Appl. 24, 978–989. 1043. reduce time resolution and the planning
implications of inter-annual variability. Appl.
64. DuraMAT (Durable Module Materials) 79. Liu, H., Krishna, V., Lun Leung, J.L., Reindl, T., Energy 197, 1–13.
Consortium https://www.duramat.org/. and Zhao, L. (2018). Field experience and
performance analysis of floating PV 91. Kotzur, L., Markewitz, P., Robinius, M., and
65. Steffen, B., Beuse, M., Tautorat, P., and technologies in the tropics. Prog. Photovolt.: Stolten, D. (2018). Impact of different time
Schmidt, T.S. (2020). Experience curves for Res. Appl. 26, 957–967. series aggregation methods on optimal
operations and maintenance costs of energy system design. Renew. Energy 117,
renewable energy technologies. Joule 4, 80. Lee, N., Grunwald, U., Rosenlieb, E., Mirletz, 474–487.
359–375. H., Aznar, A., Spencer, R., and Cox, S. (2020).
Hybrid floating solar photovoltaics- 92. Frew, B.A., and Jacobson, M.Z. (2016).
66. Rai, V., and Henry, A.D. (2016). Agent-based hydropower systems: benefits and global Temporal and spatial tradeoffs in power
modeling of consumer energy choices. Nat. assessment of technical potential. Renew. system modeling with assumptions about
Clim. Change 6, 556–562. Energy 162, 1415–1427. storage: an application of the POWER model.
Energy 117, 198–213.
67. Lamperti, F., Bosetti, V., Roventini, A., and 81. Farfan, J., and Breyer, C. (2018). Combining
Tavoni, M. (2019). The public costs of climate- floating solar photovoltaic power plants and 93. Kannan, R., Turton, H., and Panos, E. (2015).
induced financial instability. Nat. Clim. hydropower reservoirs: a virtual battery of Methodological significance of temporal
Chang. 9, 829–833. great global potential. Energy Procedia 155, granularity in energy-economic models—
403–411. insights from the MARKAL/TIMES framework.
68. Luderer, G., Krey, V., Calvin, K., Merrick, J., In Informing Energy and Climate Policies
82. Schindele, S., Trommsdorff, M., Schlaak, A., Using Energy Systems Models (Springer
Mima, S., Pietzcker, R., Van Vliet, J., and
Obergfell, T., Bopp, G., Reise, C., Braun, C., International Publishing), pp. 185–200.
Wada, K. (2014). The role of renewable energy
Weselek, A., Bauerle, A., Högy, P., et al.
in climate stabilization: results from the
(2020). Implementation of agrophotovoltaics: 94. Brown, T., Schlachtberger, D., Kies, A.,
EMF27 scenarios. Clim. Change 123, 427–441.
techno-economic analysis of the price- Schramm, S., and Greiner, M. (2018).
69. Kavlak, G., McNerney, J., Jaffe, R.L., and performance ratio and its policy implications. Synergies of sector coupling and transmission
Trancik, J.E. (2015). Metal production Appl. Energy 265, 114737. reinforcement in a cost-optimised, highly
requirements for rapid photovoltaics renewable European energy system. Energy
83. Barron-Gafford, G.A., Pavao-Zuckerman, 160, 720–739.
deployment. Energy Environ. Sci. 8, 1651–
M.A., Minor, R.L., Sutter, L.F., Barnett-
1659.
Moreno, I., Blackett, D.T., Thompson, M., 95. Kwok, G., Farbes, J., and Jones, R. (2020).
Dimond, K., Gerlak, A.K., Nabhan, G.P., and Low-carbon transition strategies for the
70. Jean, J., Brown, P.R., Jaffe, R.L., Buonassisi, T.,
Macknick, J.E. (2019). Agrivoltaics provide southeast united states (Evolved Energy
and Bulovic, V. (2015). Pathways for solar
mutual benefits across the food–energy– Research/Sustainable Development
photovoltaics. Energy Environ. Sci. 8, 1200–
water nexus in drylands. Nat. Sustainability 2, Solutions Network). https://www.unsdsn.org/
1219.
848–855. low-carbon-transition-for-the-southeast-
71. Fraunhofer Institute for Solar Energy Systems, united-states.
84. Schneider, J., Pannicke, N., Haufe, N., Biger,
ISE (2019). Photovoltaics report. https://www.
J., Volz, B., Schmeichel, A., Zwosta, N., and 96. Bogdanov, D., Gulagi, A., Fasihi, M., and
ise.fraunhofer.de/content/dam/ise/de/
Mattiza, M. (2020). Vertical bifacial solar Breyer, C. (2020). Full energy sector transition
documents/publications/studies/
module systems on wild flower strips towards 100% renewable energy supply:
Photovoltaics-Report.pdf.
combining energy and crops production with integrating power, heat, transport and
72. (2020) World Silver Survey. preservation of biodiversity (37th European industry sectors including desalination. Appl.
Photovoltaic Solar Energy Conference and Energy, 116273.
73. Valero, A., Valero, A., Calvo, G., and Ortego, Exhibition).
97. Heide, D., Greiner, M., von Bremen, L., and
A. (2018). Material bottlenecks in the future
85. Jäger-Waldau, A. (2020). The untapped area Hoffmann, C. (2011). Reduced storage and
development of green technologies. Renew.
potential for photovoltaic power in the balancing needs in a fully renewable
Sustain. Energy Rev. 93, 178–200.
European Union. Clean. Technol. 2, 440–446. European power system with excess wind and
74. Heath, G.A., Silverman, T.J., Kempe, M., solar power generation. Renew. Energy 36,
86. Pietzcker, R.C., Stetter, D., Manger, S., and 2515–2523.
Deceglie, M., Ravikumar, D., Remo, T., Cui, H., Luderer, G. (2014). Using the sun to
Sinha, P., Libby, C., Shaw, S., et al. (2020). decarbonize the power sector: the economic 98. World Bank (2017). Solar resource and
Research and development priorities for potential of photovoltaics and concentrating photovoltaic potential of indonesia, Energy
silicon photovoltaic module recycling to solar power. Appl. Energy 135, 704–720. Sector Management Assistance Program
support a circular economy. Nat. Energy 5, (World Bank Group). http://documents.
502–510. 87. Eurek, K., Vimmerstedt, L., Lamers, P., Cohen, worldbank.org/curated/en/
S., Podkaminer, K., Lyer, G., Binsted, M., Wise, 729411496240730378/Solar-resource-and-
75. Latunussa, C.E.L., Ardente, F., Blengini, G.A., M., and Graham, N. (2020). Challenges for photovoltaic-potential-of-Indonesia.
and Mancini, L. (2016). Life cycle Assessment modeling high-penetration renewable
of an innovative recycling process for energy and energy efficiency technologies in 99. Stocks, M., Stocks, R., Lu, B., Cheng, C., and
crystalline silicon photovoltaic panels. Sol. integrated assessment models (Presented at Blakers, A. (2021). Global atlas of closed-loop
Energy Mater. Sol. Cells 156, 101–111. the 13th IAMC). pumped hydro energy storage. Joule 5,
270–284.
76. Gagnon, P., Margolis, R., Melius, J., Philipps, 88. IAMC (2020). Analyses presented at the 13th
C., and Elmore, R. (2016). Rooftop solar integrated assessment modeling consortium 100. Victoria, M., Zhu, K., Brown, T., Andresen,
photovoltaic technical potential in the United (IAMC) annual meeting. https://www. G.B., and Greiner, M. (2019). The role of
States: a detailed assessment (NREL). https:// iamconsortium.org/event/thirteenth-annual- storage technologies throughout the
www.nrel.gov/docs/fy16osti/65298.pdf. meeting-of-the-iamc-2020/. decarbonisation of the sector-coupled
European energy system. Energy Convers.
77. Bódis, K., Kougias, I., Jäger-Waldau, A., 89. Luderer, G., Madeddu, S., Ueckerdt, F., Manag. 201, 111977.
Taylor, N., and Szabó, S. (2019). A high- Pietzcker, R.C., Levesque, A., Pehl, M., Rottoli,
resolution geospatial assessment of the M., Dirnaichner, A., and Kriegler, E. (2020). 101. Jerez, S., Tobin, I., Vautard, R., Montávez, J.P.,
rooftop solar photovoltaic potential in the When the most valuable becomes the López-Romero, J.M., Thais, F., Bartok, B.,

Joule 5, 1041–1056, May 19, 2021 1055

 ll
Perspective

Christensen, O.B., Colette, A., Déqué, M., comprehensive review of the feasibility of International Conference on the European
et al. (2015). The impact of climate change on 100% renewable-electricity systems. Renew. Energy Market (EEM), pp. 1–7.
photovoltaic power generation in Europe. Sustain. Energy Rev. 92, 834–847.
Nat. Commun. 6, 10014. 122. Sasse, J.P., and Trutnevyte, E. (2020). Regional
112. Kaufmann, R.K., Newberry, D., Xin, C., and impacts of electricity system transition in
102. Schlott, M., Kies, A., Brown, T., Schramm, S., Gopal, S. (2021). Feedbacks among electric Central Europe until 2035. Nat. Commun. 11,
and Greiner, M. (2018). The impact of climate vehicle adoption, charging, and the cost and 4972.
change on a cost-optimal highly renewable installation of rooftop solar photovoltaics.
European electricity network. Appl. Energy Nat. Energy 6, 143–149. 123. Tröndle, T., Lilliestam, J., Marelli, S., and
230, 1645–1659. Pfenninger, S. (2020). Trade-offs between
113. Zhu, K., Victoria, M., Andresen, G.B., and
103. Orths, A., and Eriksen, P.B. (2016). The future Greiner, M. (2020). Impact of climatic, geographic scale, cost, and infrastructure
has come: the 100% RES driven power system technical and economic uncertainties on the requirements for fully renewable electricity in
is reality. Revue de l Electricité et de l optimal design of a coupled fossil-free Europe. Joule 4, 1929–1948.
Electronique 5, 84–86. electricity, heating and cooling system in
Europe. Appl. Energy 262, 114500. 124. Lombardi, F., Pickering, B., Colombo, E., and
104. Doelling, D.R., Loeb, N.G., Keyes, D.F., Pfenninger, S. (2020). Policy decision support
Nordeen, M.L., Morstad, D., Nguyen, C., 114. Laine, H.S., Salpakari, J., Looney, E.E., Savin, for renewables deployment through spatially
Wielicki, B.A., Young, D.F., and Sun, M. (2013). H., Peters, I.M., and Buonassisi, T. (2019). explicit practically optimal alternatives. Joule
Geostationary enhanced temporal Meeting global cooling demand with 4, 2185–2207.
interpolation for Ceres flux products. photovoltaics during the 21st century. Energy
J. Atmos. Oceanic Technol. 30, 1072–1090. Environ. Sci.. 12, 2706–2716. 125. Ram, M., Bogdanov, D., Aghahosseini, A.,
Gulagi, A., Oyewo, A.S., Child, M., Caldera,
105. Kroposki, B., Johnson, B., Zhang, Y., 115. Kätelhön, A., Meys, R., Deutz, S., Suh, S., and U., Sadovskaia, K., Farfan, J., Barbosa,
Gevorgian, V., Denholm, P., Hodge, B.-M., Bardow, A. (2019). Climate change mitigation L.S.N.S., et al. (2019). Global energy system
and Hannegan, B. (2017). Achieving a 100% potential of carbon capture and utilization in based on 100% renewable energy – power,
renewable grid: operating electric power the chemical industry. Proc Natl Acad Sci USA heat, transport and desalination sectors
systems with extremely high levels of variable 116, 11187–11194. (Study by Lappeenranta university of
renewable energy. IEEE Power and Energy technology and energy watch group). http://
Mag 15, 61–73. 116. Horvath, S., Fasihi, M., and Breyer, C. (2018).
energywatchgroup.org/wp-content/uploads/
Techno-economic analysis of a decarbonized
EWG_LUT_100RE_All_Sectors_Global_
106. Garralaga Rojas, E.G., Sadri, H., and Krueger, shipping sector: technology suggestions for a
Report_2019.pdf.
W. (2020). Case study of MW-sized power fleet in 2030 and 2040. Energy Convers.
generation at St. Eustatius island combining Manag. 164, 230–241.
126. O’Shaughnessy, E., Barbose, G., Wiser, R.,
photovoltaics, battery storage, and gensets.
117. Kakoulaki, G., Kougias, I., Taylor, N., Dolci, F., Forrester, S., and Darghouth, N. (2021). The
Prog. Photovolt.: Res. Appl. 28, 562–568.
Moya, J., and Jäger-Waldau, A. (2020). Green impact of policies and business models on
107. AURECON. (2018). Hornsdale power reserve, hydrogen in Europe – a regional assessment: income equity in rooftop solar adoption. Nat.
year 1 technical and market impact case substituting existing production with Energy 6, 84–91.
study. https://www.aurecongroup.com/ electrolysis powered by renewables. Energy
markets/energy/hornsdale-power-reserve- Convers. Manag. 228, 113649. 127. IEA Sustainable Development Goal 7. Data
impact-study. and Projections. https://www.iea.org/reports/
118. Fasihi, M., Efimova, O., and Breyer, C. (2019). sdg7-data-and-projections.
108. Oyewo, A.S., Farfan, J., Peltoniemi, P., and Techno-economic assessment of CO2 direct
Breyer, C. (2018). Repercussion of large scale air capture plants. J. Cleaner Prod. 224, 128. Szabó, S., Pinedo Pascua, I., Puig, D., Moner-
hydro dam deployment: the case of Congo 957–980. Girona, M., Negre, M., Huld, T., Mulugetta, Y.,
grand Inga hydro project. Energies 11, 972. Kougias, I., Szabó, L., and Kammen, D. (2021).
119. Breyer, C., Tsupari, E., Tikka, V., and Vainikka, Mapping of affordability levels for
109. Parkinson, G. (2020). Solar meets 100 per cent P. (2015). Power-to-gas as an emerging photovoltaic-based electricity generation in
of South Australia demand for first time. profitable business Through creating an the solar belt of sub-Saharan Africa, East Asia
Renew Economy. https://reneweconomy. integrated value chain. Energy Procedia 73, and South Asia. Sci. Rep. 11, 3226.
com.au/solar-meets-100-per-cent-of-south- 182–189.
australia-demand-for-first-time-78279/.
120. Victoria, M., and Andresen, G.B. (2019). Using 129. Bhatia, M., and Niki, A. (2015). Beyond
110. Frew, B., Cole, W., Denholm, P., Frazier, A.W., validated reanalysis data to investigate the Connections. Energy Access Redefined
Vincent, N., and Margolis, R. (2019). Sunny impact of the PV system configurations at high (World Bank)).
with a chance of curtailment: operating the us penetration levels in European countries.
grid with very high levels of solar Prog. Photovolt.: Res. Appl. 27, 576–592. 130. Needleman, D.B., Poindexter, J.R., Kurchin,
photovoltaics. iScience 21, 436–447. R.C., Marius Peters, I.M., Wilson, G., and
121. Hörsch, J., and Brown, T. (2017). The role of Buonassisi, T. (2016). Economically
111. Brown, T.W., Bischof-Niemz, T., Blok, K., spatial scale in joint optimisations of sustainable scaling of photovoltaics to meet
Breyer, C., Lund, H., and Mathiesen, B.V. generation and transmission for European climate targets. Energy Environ. Sci. 9, 2122–
(2018). Response to ‘burden of proof: a highly renewable scenarios. In 14th 2129.

1056 Joule 5, 1041–1056, May 19, 2021