Page 412 - A Comprehensive Guide to Solar Energy Systems

Chapter 21 • (EROI) and (EPBT) for PVs 421

quality correction and harmonizing for large battery systems. Koppelaar also examined
the age of systems being analyzed compared with publication dates and found that some
authors were using systems that were 6 years, in an extreme case 18 years old. This had a
serious impact on values for both EpBT and EROI of the pv modules and systems studied.
Five popular types of pv modules often analyzed are monocrystalline (mono­Si), mul­
ticrystalline, or polycrystalline (multi­Si or poly­Si), amorphous silicon (a­Si), CdTe thin­
film (CdTe), and CIS thin­film (CIS). Kumar et al. [20] studied these five types of modules
in varying environments and systems. As can be expected, they found that EpBT varied
greatly from case to case and also that Si volume and methods of production were large
factors in determining CED, which along with efficiency were the largest factors for EpBT.
Overall they found mono­Si to have the highest lifecycle energy demand and a large range
in EpBT due to estimations of silicon purification and crystallization processing. poly­Si
had a much lower EpBT and the thin­film systems (a­Si, CdTe, and CIS) had the lowest
EpBTs. Among the thin­films, CIS consumed the most primary energy and the a­Si had the
longest EpBT due to lower conversion efficiency. CdTe held the lowest EpBT across sys­
tems analyzed. Bhandari et al. [38] found similar results when they examined several LCAs
for EpBT and CED. After harmonizing the variables for output, they calculated mean EpBT
varying from 1.0 to 4.1 years. CdTe modules ranked lowest, followed by copper indium
gallium diselenide, a­Si, poly­Si, and mono­Si with the highest EpBT. They also found that
across different types of pv, variation in CED was greater than that in efficiency and perfor­
mance ratio, concluding that CED, not efficiency, has greater influence on EpBT.
If a low CED is the important factor for a low EpBT, we can assume that new manufac­
turing technologies and application methods, such as advanced production processes and
reducing raw material consumption, especially silicon, would show a reduction in EpBTs
over time. Wong et al. [39] studied the difference in mono­Si versus poly­Si pv process­
ing and found that the former had a larger CED due to an additional Czochralski process,
which has a significant energy requirement. Although this results in greater conversion
efficiency for mono­Si, it is insufficient to lower the EpBT below that of poly­Si. Wong et al.
describe another approach to manufacturing a combination of mono­Si and poly­Si into
a hybrid c­Si, which eliminates the need for the extra Czochralski process and produces
a pv module with greater efficiencies than poly­Si but lower energy costs than mono­Si.
They also note that wafer thickness has been greatly reduced without losing efficiency by
the use of reflective back­coatings, which increase chances of photon capture.
In the few studies that have applied harmonization to NEA, EpBT, and EROI analyses,
a reduction in CED across the industry was found. Gorig and Breyer [40] calculated and
compared the CED of different modules and systems over time using LCAs. They employed
financial learning curve concepts to determine the energy demand for major pv systems.
They weighted both module and system energy demand according to their share in the
pv market at the time and found that the CED for all modules and systems, for which
there was adequate marketing data, had decreased over time. The modules and systems
with higher amounts of Si decreased at faster rates. They showed that energy consumption
in pv manufacturing followed the log­linear learning curve and BOS, such as aluminum