Page 430 - A Comprehensive Guide to Solar Energy Systems

440 A COmPREhEnSIVE GuIdE TO SOLAR EnERGy SySTEmS

creased due to the energy intensive ion implantation process. module 1 and 3 require
the same amount of feedstock, but the umG-Si feedstock and seeded dS reduce the PE
requirement for this step by almost a factor of three (although this comes at an efficiency
penalty).
The breakdown for the aforementioned recent LCA results on monocrystalline silicon
modules into poly-Si, ingot growth and wafering, cell processing, and encapsulation is
−2
approximately 500, 1260, 390, and 600 mJ p m (or 18, 46, 14, and 22%) respectively [15].
Comparing this to the breakdown shown in Fig. 22.3B we see that in module 1a modest
reductions have been achieved in feedstock, crystallization and wafering (owing to thin-
ner wafers), and encapsulation. The cell processing energy requirement has not changed
much, most likely because low-temperature processing has offset increases due to the in-
creased overall process complexity. As mentioned before, poly-Si consumption is drasti-
cally reduced in module 1b, and module 2 uses low-energy feedstock. Because module 2 is
based on quasi-mono wafers, the energy requirement here is reduced further (the corre-
sponding energy requirement for ingot growth and wafering in 2011 in a multi-Si module
−2
was ∼670 mJ p m [16]).
The confidence intervals are quite large: [1570, 2640], [1140, 1930] and [1070, 1630]
−2
mJ p m for modules 1, 2, and 3 respectively (Fig. 22.2A). This is mainly due to the tem-
poral uncertainty: About half (or even 63% for module 1a) of the PE requirement comes
from poly-Si feedstock, ingot growth and wafering, and these production steps have a
large potential for reduction in energy intensity by scaling up. We therefore expect the
actual CEd to decrease towards the lower bound rather than increase.
We also calculated the EPBTs for these modules (Fig. 22.4A). The mean payback time of
module 1 is just over 0.6 yr, while 2 and 3 have an EPBT of a little under 0.5 yr. Adding 0.2 yr
for BOS brings those EPBT to 0.7–0.8 yr. As shown in Table 22.1 the most recent module
EPBT values are 1.6 for mono-Si and 1.2 for multi-Si [7], so a reduction in the EPBT of over
a factor of two may be feasible. Correspondingly, the EROI of produced electricity for a
lifetime of 30 yr would be about 40.

22.7.2 Balance of System

Although the BOS lies outside of the scope of this investigation, it is a crucial part of any
PV energy system and we will therefore briefly discuss its impacts on our results. One can
distinguish between two common types of grid-connected PV systems: building-integrat-
ed systems (e.g., rooftop panels) and ground-based systems (e.g., large scale centralized
power plants). In Table 22.3, we show the EPBT and EROI for both types of system, calcu-
lated with data from Fthenakis et al. [2]. The rooftop system consists of a 2.5 kW inverter
and a Schletter mounting system. The PV plant uses average mounting structure data from
market surveys and inverter data based on a 4.6 mW p plant in Tucson, uSA. According to
the IEA LCA guidelines a PR of 0.80 is used for a utility PV plant [6].
To reduce costs associated with the BOS the EuPVTP has identified some research pri-
orities [16]. For the EPBT the most relevant goals are an increase in the inverter lifetime (of