|
||||||||||||||||||||||||
| Home > Resources > Application Notes > IR-Optical Material | ||||||||||||||||||||||||
|
|
Application Notes:
|
|||||||||||||||||||||||
IR-VASE Analysis of IR-Optical Materials |
||
|
||
MgF2-Al2O3 Film on SiliconThis IR-VASE study determined the thickness and mid-IR optical functions, (λ) and (λ), for thin films of ZnS, MgF2-Al2O3, GeO2 and Al2O3. The samples consisted of various single layers on standard silicon wafers.All of these films are substantially transparent between 2µm – 10µm wavelengths (1000 – 5000 cm-1), as indicated by the full range coverage of the interference spectra (Figure 1). We exploit this transparent region to determine film thickness and refractive index. We created a two-term Sellmeier (pole) dispersion model in the GenOsc layer to account for dispersion on both ends of the 2µm – 10µm spectral region. ![]() Figure 1. Model, experimental data and fits for MgF2-Al2O3 combined layer.
Mid-IR vibrational absorptions of amorphous materials are generally best fit with Gaussian
oscillators. Each Gaussian represents the summation of a large population of narrow molecular resonances with a broader normal distribution of center energies created by slightly different bond lengths and bond angles. The fit to a portion of the ε2 spectrum for the MgF2-Al2O3 point-by-point fit is shown in Figure 2. Both ε2 and ε1 reference spectra are fit using procedures described in the addendum of the WVASE32 software manual. ![]() Figure 2. Dielectric function (ε2) after fitting a portion of MgF2-Al2O3 point-by-point reference spectrum. Red Curve is reference spectra and black is model. Blue curves are the individual Gaussian Oscillators. |
||
All the parameters of the GenOsc layer, as well as layer thickness, are then fit to the ellipsometric data. The resulting fits are shown in Figure 1, with n(λ) and k(λ) shown in Figures 3a and 3b. Beyond the determination of n and k at specific wavelengths, reproducible, accurate data from the IR-VASE can produce additional information about a film. In Figure 3b, subtle details in n(λ) and k(λ) often yield useful information about the chemical make-up of the films. Vibrational absorptions in the 2 to 10µm region can indicate the presence of contaminants that are incorporated in the film either during or after deposition. For example, the MgF2 layer appears to have incorporated some water, as indicated by the O-H stretch absorption at ~3µm combined with what appears to be the O-H-O scissor vibration at ~ 6µm. |
||
Thickness Non-uniformity
The ability to quantitatively measure percent depolarization is an important diagnostic tool for analysis of non-ideal thin films, and is a quantity that is readily measured by the IR-VASE. In this example, “%Depolarization” gave quantitative information regarding the effects of non-uniformity in the film thickness (see Figure 4). The depolarization effects of such non-uniformities are amplified by thicker films. |
![]() Figure 3. a) The optical functions n(λ) and k(λ) for MgF2-Al2O3 film, for 2 ≤ λ ≤ 33µm. b) Close-up between 2 and 11µm showing subtle O-H asborption bands. |
|
Because the IR-VASE yields accurate Ψ-∆ data, one can be confident that these misfits are real and not measurement artifacts. In this case, misfits indicated that optical properties of at least one of the films changed when deposited in the multilayer stack rather than as individual layers on silicon. Despite the fact that the fits are not perfect in some regions, this model still provided a very good match to IR-Reflectance data. |
![]() Figure 4. Percent Depolarization data for GeO2 layer. |
|
![]() Figure 5. Model, experimental data and fits for the multilayer MgF2-ZnS/Al2O3 on Si. |
||