HINTS & TIPS

Q: Can ATLAS handle photogeneration from non-normally incident light? What syntax is used to determine the correct photogeneration rate?

A: The models with Luminous allow simulation of photogeneration within ATLAS. Arbitrary light sources are available within Luminous using the BEAM statement. The program uses geometric ray tracing to determine the path of a light beam including refraction and reflection of non-normally incident light. It then applies models for the absorption of the light to determine the photogeneration rate.

To use Luminous the use should first define a light source using the BEAM statement and then choose the light intensity using a SOLVE statement. The BEAM statement can be explained by a simple example. The following syntax generates the ray trace in Figure 1.

BEAM NUM=1 X.ORIGIN=3.0 Y.ORIGIN=-5.0 ANGLE=60.0 \
MIN.W=-1 MAX.W=1 WAVEL=0.6 REFLECT=2

The parameter NUM sets the beam number. Luminous can applications up to 10 independent beams. X.ORIGIN and Y.ORIGIN define the initial starting point for the light beam. This must be outside the device coordinates and for non-normal beams it is important to keep this point well away from the device. The ANGLE parameter determines the direction of the light beam relative to the x-axis. ANGLE=90 gives normal incidence from the top of the device. The light is defined as coming from a line perpendicular to the direction set by ANGLE and passing through (X.ORIGIN,Y.ORIGIN). MAX.W and MIN.W set a window along this line through which the light passes. The default for these parameters is +/- infinity.

The wavelength of the light beam is defined in microns using the WAVEL parameter. It is also possible to specify spectral sources of light by replacing WAVEL with POWER.FILE={filename}, where the {file} is a UNIX text file defining the spectrum in terms of an XY list of wavelength vs intensity.

Luminous automatically chooses the number of rays needed to resolve the geometry. In the case of Figure 1 only one initial ray is needed. The number of rays used by Luminous is purely a function of the geometry and is not related to optical intensity, photogeneration rate or MIN.W and MAX.W. The REFLECT parameter is used to limit the number of reflections the light beam is allowed. In ATLAS Version 3.0.0.R an alternative parameter MIN.POWER is used to set a relative intensity below which no more rays are traced.

The use has a choice in Luminous as to whether the rays should reflect from the sidewalls and bottom of the device structure. If the simulation is of a partial wafer (such as is typical for CCD simulation) the light should not reflect. This is the default shown in Figure 1. If the simulation is of a complete device (such as is typical for solar cells) the light should reflect. The parameter BACK.REFL is used to enable back and side reflection. The result of adding this to the previous syntax is shown in Figure 2. In this figure the limit set by REFLECT=2 is also clear as each ray reflects only two times.

The ray trace is only done once a SOLVE statement is used to turn on the light beam. For example the syntax SOLVE B1=0.5 sets the power of beam #1 to 0.5W/cm^2. DC and transient ramps of light intensity can also be performed.

The refraction and reflection are determined by the real portion of the refractive index for each material. The imaginary portion on the refractive index controls the absorption of the light. Wavelength dependent defaults exist in ATLAS for common materials, but can be defined by the user as follows:

MATERIAL MATERIAL=Silicon REAL.INDEX={value} IMAG.INDEX={value}

From the ray trace information and the imaginary refractive indices the photogeneration rate is calculated at each node in the structure. An integration method is used to ensure the charge generated is independent of the mesh density. The photogeneration rate from the ray trace in Figure 2 is shown in Figure 3. The electrons and holes generated are included in electrical simulations in ATLAS to determine collection efficiency, spectral response and other device parameters.

 

Figure 1 Figure 2 Figure 3