Which Layer Absorbs the Light? Analyze and Optimize an a-Si Thin-Film Solar Cell

Total absorption is not useful absorption. In a thin-film solar cell, only light absorbed in the active a-Si layer can contribute to photoconversion; absorption in ITO is a parasitic loss.

This tutorial uses a simplified ITO/a-Si structure to identify which layer absorbs the light. You will use R/T/A and Layer Absorption to separate active-layer absorption from parasitic absorption, then improve average a-Si absorption over 400–700 nm with a thickness sweep and optimization.

Flexible amorphous-silicon thin-film solar cells can be integrated directly with roofing material (Source: Maury Markowitz / Wikimedia Commons · License: CC BY-SA 3.0)

Distinguish Active-Layer and Parasitic Absorption

Amorphous silicon (a-Si) is the active absorber in this model. ITO is the transparent front electrode; light absorbed by ITO is usually a parasitic loss and should not be treated as a-Si absorption.

Typical layers and incident-light direction in a thin-film solar cell (Source: Radiotrefoil and U.S. Department of Energy / Wikimedia Commons · License: Public Domain)

This tutorial simplifies the transparent conducting layer as ITO and the absorber as a-Si. The layer-absorption result then separates parasitic ITO loss from useful a-Si absorption.

The complex refractive index of an absorbing material is

$$ \tilde n=n+\mathrm{i}k, \qquad \alpha=\frac{4\pi k}{\lambda}. $$

Here, $\tilde n$ is the complex refractive index, $n$ is its real part, $k$ is the extinction coefficient, $\mathrm{i}$ is the imaginary unit, $\alpha$ is the absorption coefficient, and $\lambda$ is the vacuum wavelength. A larger $k$ makes light decay more rapidly inside the material. Interference also changes the electric-field distribution within a thin-film stack, so layer absorption cannot be inferred from thickness alone.

At every wavelength, a passive stack satisfies

$$ R+T+A=1, \qquad A=\sum_{j=1}^{L}A_j. $$

Here, $R$, $T$, and $A$ are power reflectance, transmittance, and total absorptance, respectively; $A_j$ is the fraction of incident power absorbed in layer $j$; and $L$ is the number of layers. Absorptance gives the total absorption, while Layer Absorption answers the more useful question: “Which layer absorbs the light?”

The same distinction applies to photodetectors, selective absorbers, and light-emitting devices. Absorption in the target layer is usually useful, whereas absorption in an electrode or package is often a loss.

Build a Simplified ITO/a-Si Structure

Build an air/ITO/a-Si/glass structure. Use the wavelength-dependent teaching data for ITO and a-Si from the case library, and set the bottom glass to a constant refractive index of 1.52.

Position	Material	Thickness	Index type
Top medium	Air	—	Constant, $n=1.00$, $k=0$
Layer 1	ITO	80 nm	File
Layer 2	a-Si	200 nm	File
Bottom medium	Glass	—	Constant, $n=1.52$, $k=0$

Baseline ITO/a-Si structure

This model is intended for learning optical absorption allocation. It omits the back electrode, doped layers, texture, and carrier transport, so its results are not solar-cell conversion efficiency.

Set the Optical Conditions

On the Optics page, set 400–900 nm with a 5 nm step, 0° incidence, and unpolarized light. Enable Reflectance, Transmittance, Absorptance, and Layer Absorption.

Optics settings for observing the absorption edge and layer-resolved absorption

The 400–900 nm range reveals the a-Si absorption edge; the design metric below is evaluated only from 400–700 nm. When no incident spectrum is enabled, a band average is the arithmetic mean of the sampled wavelengths, not a solar-spectrum-weighted efficiency.

Tip

Predict the result before running: absorption from 400–700 nm should occur mainly in a-Si, with little parasitic absorption in ITO. Increasing a-Si thickness should continue to raise active-layer absorption, but with diminishing returns.

Establish an Absorption Baseline

Run the calculation, verify R/T/A first, and then open Layer Absorption.

Layer-resolved absorption of the baseline structure

Condition	$R$	$T$	Total $A$	ITO absorption	a-Si absorption
400–700 nm average	14.217%	8.672%	77.110%	0.461%	76.650%
550 nm	5.269%	3.635%	91.096%	0	91.096%

Total absorption agrees with the sum of the two layer contributions, and the maximum energy-closure error across all wavelengths is approximately $1.1\times10^{-16}$. Energy conservation confirms that the result is complete, but only the layer-resolved result shows whether the absorption occurs in a-Si.

Sweep the a-Si Thickness

In Sweep, select a-Si Absorber → Thickness and sweep from 100 nm to 500 nm in 50 nm steps. Keep the rest of the structure and all optical settings unchanged.

a-Si thickness sweep

After selecting Sweep, open the Layer Absorption result and set Layer to a-Si Absorber. The nine curves correspond to a-Si thicknesses from 100 to 500 nm.

Active-layer absorption spectra for the nine a-Si thicknesses

a-Si thickness	Average $R$	Average $T$	Average a-Si absorption
100 nm	18.000%	18.594%	62.949%
200 nm	14.217%	8.672%	76.650%
300 nm	12.788%	4.635%	82.116%
400 nm	12.645%	2.551%	84.344%
500 nm	12.837%	1.433%	85.270%

A thicker a-Si layer increases active-layer absorption, but with diminishing returns

As a-Si thickness increases from 100 nm to 500 nm, average transmittance falls from 18.594% to 1.433% and active-layer absorption rises from 62.949% to 85.270%. Reflectance is not monotonic; it oscillates around 13% because interference redistributes the field inside the stack.

Optically, thicker a-Si absorbs more light. It also uses more material and may reduce carrier-collection efficiency. The thickness sweep reveals the optical trend, but it cannot determine the optimum thickness of a real cell by itself.

Optimize Active-Layer Absorption Directly

The objective is the average layer absorption of a-Si:

$$ \overline{A}{\mathrm{a\text{-}Si}} =\frac{1}{M}\sum{q=1}^{M}A_{\mathrm{a\text{-}Si}}(\lambda_q). $$

Here, $\overline{A}{\mathrm{a\text{-}Si}}$ is the average a-Si layer absorption; $A{\mathrm{a\text{-}Si}}(\lambda_q)$ is the a-Si absorption at wavelength $\lambda_q$; $q$ is the sample index; and $M=61$ is the number of samples from 400–700 nm at a 5 nm step.

In Optimizer, use the following settings:

Item	Setting
Objective	Maximize the average a-Si Layer Absorption
Band	400–700 nm, 5 nm step
Variable 1	ITO thickness, 40–140 nm, starting at 80 nm
Variable 2	a-Si thickness, 100–500 nm, starting at 200 nm
Algorithm	TRF, up to 160 evaluations
Global starts	Grid 5 × 5, retain 3 seeds

Defining active-layer absorption directly as the optimization objective

Grid seeds and TRF algorithm settings

Do not substitute total Absorptance for this objective. An increase in total absorption could come from ITO, where the absorbed light never reaches the a-Si layer.

After 72 objective evaluations and 34 iterations, the optimizer finds an ITO thickness of 62.372 nm and an a-Si thickness of 485.857 nm. Select Apply to Structure, then return to Structure and confirm that the thicknesses have been updated.

Active-layer absorption optimization report

Structure after applying the optimum solution

Validate the Optimum with a Forward Calculation

After applying the optimum thicknesses, run the calculation again. Do not treat the optimization report itself as the final result.

Layer-resolved absorption of the optimized structure

400–700 nm average metric	Baseline	Optimized	Change
$R$	14.217%	8.052%	−6.165 percentage points
$T$	8.672%	1.426%	−7.247 percentage points
Total $A$	77.110%	90.522%	+13.412 percentage points
ITO absorption	0.461%	0.473%	+0.012 percentage points
a-Si absorption	76.650%	90.049%	+13.399 percentage points

Average a-Si absorption improves by 13.399 percentage points, or approximately 17.48% relative, while parasitic ITO absorption is nearly unchanged. At 550 nm, a-Si absorption is 96.496%, and the final result still satisfies $R+T+A=1$.

The optimum a-Si thickness lies near the 500 nm upper bound, showing that this purely optical objective still favors a thicker absorber. An engineering design should add constraints for manufacturable thickness, material use, and electrical collection, and it should use weighting by a real solar spectrum.

Variation Exercise

In the current model, reset the a-Si thickness to 200 nm, then set the ITO thickness to 40, 80, and 120 nm in turn. Record the average ITO and a-Si layer absorption for each case, then determine whether any increase in total absorption actually reaches the active layer.

---

← Back to Tutorial Catalog · Next: Why Does the Optimized a-Si Cell Absorb More Light?

Going further

These tutorials cover the operating steps only. The physics behind them, the full parameter reference for each feature, how to read the results, and the algorithm validation all live on the documentation site:

• Transfer-Matrix Method — the physics and formulas behind these tutorials
• Feature Guide — complete reference for every screen and parameter
• Results — how to read each kind of output
• Open Dreapex TMM — build and simulate in the browser