Pvi20 Front Cover indd

Figure 3. Schematic illustration of optical losses at the front of a silicon wafer solar cell.
Metal shading
27%
Shunt resistance
2%
Front-surface
11%
Non-perfect
IQE
40%
Forward bias
currrent at MPP
17%
Front-surface escape
2%
Figure 4. Pie chart of the relative current losses at the maximum power point for a 
standard industrial p-type Al-BSF solar cell. The total current losses amounted to 
12.7mA/cm
2
 in this specific case.

---

Ph o to v o l t a i c s I n te r n at i o n a l
51
Cell 
Processing
Absorption in the front-side silicon 
nitride passivation/anti-reflective coating
A current loss mechanism that is typically 
grouped under the ‘non-perfect IQE’ 
category is the parasitic photon absorption 
in the front SiN
x
AR coating. Measuring 
the reflectance of the solar cell surface 
is not sufficient for an optimization of 
AR coatings to be performed if there is 
non-negligible absorption in the coating 
material. This is particularly important 
for AR coatings that are optimized for 
solar cells within PV modules. If account 
is taken of the optical properties of the 
encapsulation material and the glass cover 
sheet, the desirable refractive index for the 
cell’s AR coating is actually higher than the 
refractive index for optimum reflectance 
in air. A higher refractive index is typically 
conne c te d to a higher absor ption 
coefficient and, consequently, causes 
higher parasitic absorption losses. 
Unfortunately there is no direct way 
to measure the absorption in an AR 
coating once it is deposited onto a solar 
cell. Furthermore, the absorption losses 
occur in the short-wavelength range (< 
500nm) and many of the materials used 
also absorb light in this region. Glass, 
ethylene-vinyl acetate (EVA) and silicon 
– either because of their large thicknesses 
or high absorption coefficients – absorb 
considerable fractions of the light. 
Moreover, these fractions are much 
larger than the fraction of light absorbed 
by the AR coating. This also makes an 
indirect extraction of the absorption by 
the AR coating from measurements very 
difficult.
The absorption by the AR coating can, 
however, be calculated very accurately 
if the optical material parameters – 
refractive index and absorption coefficient 
– are well known. For this purpose, a 
computer-based simulation method has 
been developed which allows the light 
absorption by the AR coating of a textured 
silicon wafer to be calculated. The method 
is based on work published elsewhere 
[4] and spe ctroscopic ellipsometr y 
measurements of the optical parameters. 
Rather than minimizing the reflectance 
from the solar cell surface, the method is 
used to maximize the fraction of light 
transmitted into the silicon. This can 
lead to significant differences in the 
assessment of the coatings. The example 
given in Fig. 5 shows the investigation of 
two AR coatings, with refractive indices of
n 
= 1.9 and 
n
= 2.7, on a silicon wafer with 
pyramidal texture and encapsulated with 
a material with refractive index 
n
= 1.5 
(corresponding to glass). The material with 
refractive index 
n
= 1.9 has a very small 
absorption coefficient, while the material 
with 
n
= 2.7 is considerably absorbent. Fig. 
5 shows the weighted average reflectance 
( WAR) and the weighte d aver age 
transmission (WAT). These two quantities 
represent the fraction of solar photons 
reflected or transmitted respectively, and 
are calculated via the expressions:
(1)
where 
φ
(
λ
) is the solar photon flux 
and 
R
(
λ
) and 
T
(
λ
) are the reflectance and 
transmission.
From the reflectance, which is the 
quantity that can be measured directly, the 
material with 
n
= 2.7 appears to be more 
favourable. However, when looking at the 
transmission, it becomes clear that the 
material with 
n
= 1.9 is the better choice. 
This is because the material with 
n
= 2.7 
shows a very strong absorption.
The example presented shows clearly 
that parasitic absorption needs to be 
considered in the optimization of AR 
coatings. As a general rule, if there is a 
choice between two AR materials, it is 
better to opt for the material with the 
lower absorption instead of the one with 
the more favourable refractive index. In 
the actual cell investigated in this paper, 
a coating material similar to the one with 
n
= 1.9 was used. The contribution of AR 
coating losses to the non-perfect IQE 
is negligible for this material. However, 
the AR coating absorption will have a 
noticeable contribution in the case where 
higher index materials are required (for 
example if the cell is encapsulated or a 
stack is used).
“
If there is a choice between 
two AR materials, it is better 
to opt for the material with the 
lower absorption instead of the 
one with the more favourable 
refractive index.
”