7 Crystalline Silicon Solar Cells: Heterojunction Cells
177
Achieving these three characteristics at the same time is not at all easy—in practice
it means making compromises. If, however, the c-Si(n)/a-Si:H(p) interface is placed
on the back side, a completely new situation arises, with interesting advantages. We
are now collecting the electrons at the front contact. The electrons are the majority
charge carriers in the n-type wafers we generally employ for HJT cells—therefore
the electrons can also utilise the lateral conductivity of the bulk to reach the metal
contacts. This allows us to reduce the thickness of the TCO on the front side by
approximately 20% and to optimize the front TCO for its optical function, e.g. no
compromises have to be made with respect to the anti-reflection properties. On the
back side, however, reduced transmission of the TCO layer plays a minor role. This
is because less light reaches the cell at the back. On the back, the TCO layer can
be optimized for conducting and can be thicker than on the front side.
15 A thicker
TCO layer on the back increases light trapping for long-wave light. In the case of
bifacial cells, however, the TCO layer on the back cannot be as thick as otherwise—
the albedo effect (see Chap. 10) must be taken into account in case of bifacial cells
when determining the thickness.
Figure 7.10 illustrates the difference between an HJT cell with the pn-junction on
the front (left) and an HJT cell with the pn-junction on the back (right).
If we compare both designs then we see that the TCO layer can be thinner in the
case where the junction is on the back side.
There is another reason to move the pn-junction to the back. Indium-Tin-Oxide
(ITO) is an n-type semiconductor. If the ITO layer lies on top of the amorphous p-type
layer, which is doped with boron, a diffusion of boron atoms into the ITO layer can
occur. As a result, the boron-doped p-type amorphous layer becomes “depleted”.
16
This leads to a band bending and to performance losses [11]. To prevent this, the
thickness of the a-Si:H(p)-layer must be relatively high, with all the disadvantages
for the front side we have described above. This is an additional reason for moving
the pn-junction to the back side.
Because the TCO layer is conductive, there is no PID (Potential Induced
Degradation) effect here (see Chap. 10). The sodium ions Na
++ from the glass or
the aluminum ions Al
++ from the aluminum frame are neutralized by the TCO layer
and cannot influence the space charge zone.
Figure 7.11 shows a cross-section of an HJT cell with the pn-junction on the back
side.
(c) Illustration of the High Passivation in HJT Cells
Unilluminated and Undoped Silicon Crystal
In an unilluminated and undoped silicon crystal, the equilibrium relationship follows the law of mass action [12]. The product of electron and hole densities is at
15 The sheet resistivity of the TCO layer close to the amorphous layer is ~100 /sq. Close to the
glass of the module the sheet resistivity is much lower ~40 /sq. Thus, with the thickness of the
TCO layer, the resistivity can be optimized.
16 “depleted” means here: it loses a part of the holes which are otherwise present in a p-type
amorphous layer.
177
Achieving these three characteristics at the same time is not at all easy—in practice
it means making compromises. If, however, the c-Si(n)/a-Si:H(p) interface is placed
on the back side, a completely new situation arises, with interesting advantages. We
are now collecting the electrons at the front contact. The electrons are the majority
charge carriers in the n-type wafers we generally employ for HJT cells—therefore
the electrons can also utilise the lateral conductivity of the bulk to reach the metal
contacts. This allows us to reduce the thickness of the TCO on the front side by
approximately 20% and to optimize the front TCO for its optical function, e.g. no
compromises have to be made with respect to the anti-reflection properties. On the
back side, however, reduced transmission of the TCO layer plays a minor role. This
is because less light reaches the cell at the back. On the back, the TCO layer can
be optimized for conducting and can be thicker than on the front side.
15 A thicker
TCO layer on the back increases light trapping for long-wave light. In the case of
bifacial cells, however, the TCO layer on the back cannot be as thick as otherwise—
the albedo effect (see Chap. 10) must be taken into account in case of bifacial cells
when determining the thickness.
Figure 7.10 illustrates the difference between an HJT cell with the pn-junction on
the front (left) and an HJT cell with the pn-junction on the back (right).
If we compare both designs then we see that the TCO layer can be thinner in the
case where the junction is on the back side.
There is another reason to move the pn-junction to the back. Indium-Tin-Oxide
(ITO) is an n-type semiconductor. If the ITO layer lies on top of the amorphous p-type
layer, which is doped with boron, a diffusion of boron atoms into the ITO layer can
occur. As a result, the boron-doped p-type amorphous layer becomes “depleted”.
16
This leads to a band bending and to performance losses [11]. To prevent this, the
thickness of the a-Si:H(p)-layer must be relatively high, with all the disadvantages
for the front side we have described above. This is an additional reason for moving
the pn-junction to the back side.
Because the TCO layer is conductive, there is no PID (Potential Induced
Degradation) effect here (see Chap. 10). The sodium ions Na
++ from the glass or
the aluminum ions Al
++ from the aluminum frame are neutralized by the TCO layer
and cannot influence the space charge zone.
Figure 7.11 shows a cross-section of an HJT cell with the pn-junction on the back
side.
(c) Illustration of the High Passivation in HJT Cells
Unilluminated and Undoped Silicon Crystal
In an unilluminated and undoped silicon crystal, the equilibrium relationship follows the law of mass action [12]. The product of electron and hole densities is at
15 The sheet resistivity of the TCO layer close to the amorphous layer is ~100 /sq. Close to the
glass of the module the sheet resistivity is much lower ~40 /sq. Thus, with the thickness of the
TCO layer, the resistivity can be optimized.
16 “depleted” means here: it loses a part of the holes which are otherwise present in a p-type
amorphous layer.