Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Surface Recombination- Solar Cell Electrical Simulation
Posted Apr 26, 2015, 2:50 p.m. EDT Semiconductor Devices 3 Replies
Please login with a confirmed email address before reporting spam
Hello everyone,
My question is more of a conceptual one rather than a technical problem. I'm quite new to solar cells and i'm trying to add a surface recombination model to my 2D thin film(10um) solar cell simulation. However, i'm not sure how to implement an accurate model which has an effect that varies with the distance from the surface.
There are mainly two formulas on the internet, one of which uses SRH recombination model with (1/S) instead of recombination lifetime, which results in a recombination rate 1/m^2*s rather than 1/m^3s, which i'm not sure how to apply to a 2D structure, and the one with 'surface lifetime= (W/s)+Dn(W/pi)^2' which is only applicable to very thick cells as far as i understand.
I couldn't find a most general formula that is applicable to every thickness besides these two.
I would appreciate any help from someone who knows how to implement this.
Thanks in advance
My question is more of a conceptual one rather than a technical problem. I'm quite new to solar cells and i'm trying to add a surface recombination model to my 2D thin film(10um) solar cell simulation. However, i'm not sure how to implement an accurate model which has an effect that varies with the distance from the surface.
There are mainly two formulas on the internet, one of which uses SRH recombination model with (1/S) instead of recombination lifetime, which results in a recombination rate 1/m^2*s rather than 1/m^3s, which i'm not sure how to apply to a 2D structure, and the one with 'surface lifetime= (W/s)+Dn(W/pi)^2' which is only applicable to very thick cells as far as i understand.
I couldn't find a most general formula that is applicable to every thickness besides these two.
I would appreciate any help from someone who knows how to implement this.
Thanks in advance
3 Replies Last Post May 15, 2015, 12:08 a.m. EDT
Please login with a confirmed email address before reporting spam
Posted:
10 years ago
Surface recombination is typically modeled as a velocity, the rate of recombination the product of a surface concentration. And the recombination velocity, the units workin out correctly in 1, 2, or 3 dimensions.
The only problem with this would be if you were using a quantum electrostatic chore too. Like the density gradient method (Ancona, 1987) which results in surface carrier concentrations of zero or near zero depending on whether wacefunction penetration into adjacent insulators is modeled. But for macroscopic solar cells you likely are. It doing this sort of thing. It's more common with mosfet channels.
It should be relatively easy to implement a surface concentration based model. A disadvantage of using a microscopic proximity based model is you could become very sensitive to meshing in the near interface region. So I'd use the surface model as long as there is no density gradient method.
The SRH model in bulk is:
R - G = (n p - ni^2) / (n taup + p tauh)
So an analagous model for surfaces would be:
R - G = (n p - ni^2) / (n / vp + p / vn)
So if n >> p, then this becomes:
R - G ~= p vp
and if p >> n it becomes:
R - G ~= n vn
The only problem with this would be if you were using a quantum electrostatic chore too. Like the density gradient method (Ancona, 1987) which results in surface carrier concentrations of zero or near zero depending on whether wacefunction penetration into adjacent insulators is modeled. But for macroscopic solar cells you likely are. It doing this sort of thing. It's more common with mosfet channels.
It should be relatively easy to implement a surface concentration based model. A disadvantage of using a microscopic proximity based model is you could become very sensitive to meshing in the near interface region. So I'd use the surface model as long as there is no density gradient method.
The SRH model in bulk is:
R - G = (n p - ni^2) / (n taup + p tauh)
So an analagous model for surfaces would be:
R - G = (n p - ni^2) / (n / vp + p / vn)
So if n >> p, then this becomes:
R - G ~= p vp
and if p >> n it becomes:
R - G ~= n vn
Please login with a confirmed email address before reporting spam
Posted:
10 years ago
Thanks for the fast response Daniel.
However, this is where my problem begins. How exactly do i select this recombination's region? I'm working at 2D and 'User Defined Recombination' allows me to choose a domain, not a boundary. So i cannot choose the 'surface' for this purpose.
By the way, to approximate the surface recombination effect before, I've inserted a very thin layer (10nm) of the same material with a highly degraded lifetime right under the surface. (For example 0.02us to represent the effect of s=100cm/s, as opposed to 10us bulk lifetime). I've gotten results that make sense, but i'm not sure about the accuracy of this method.
I'm quite new to COMSOL, forgive me if all this is too trivial by the way.
Deniz
However, this is where my problem begins. How exactly do i select this recombination's region? I'm working at 2D and 'User Defined Recombination' allows me to choose a domain, not a boundary. So i cannot choose the 'surface' for this purpose.
By the way, to approximate the surface recombination effect before, I've inserted a very thin layer (10nm) of the same material with a highly degraded lifetime right under the surface. (For example 0.02us to represent the effect of s=100cm/s, as opposed to 10us bulk lifetime). I've gotten results that make sense, but i'm not sure about the accuracy of this method.
I'm quite new to COMSOL, forgive me if all this is too trivial by the way.
Deniz
Please login with a confirmed email address before reporting spam
Posted:
10 years ago
Hi, Deniz. I have met the same problem like yours.
I think a more accurate method for handling surface recombination effect is to add the corresponding
boundary conditions directly into the semiconductor interface.
for the electron, the controlling equation is Dn*(dn/dt)=Sn*(n-n0);
for the hole, the controlling equation is Dp*(dp/dt)=-Sp*(p-p0);
I think both of these equations should be added in the boundaries as the flux boundary conditions.
Unfortunately, in the semiconductor interface, COMSOL does not provide the user defined boundary conditions, which means that we can only use the default boundary conditions in COMSOL, like metal contact, insulator...
I think a more accurate method for handling surface recombination effect is to add the corresponding
boundary conditions directly into the semiconductor interface.
for the electron, the controlling equation is Dn*(dn/dt)=Sn*(n-n0);
for the hole, the controlling equation is Dp*(dp/dt)=-Sp*(p-p0);
I think both of these equations should be added in the boundaries as the flux boundary conditions.
Unfortunately, in the semiconductor interface, COMSOL does not provide the user defined boundary conditions, which means that we can only use the default boundary conditions in COMSOL, like metal contact, insulator...