Send the link below via email or IMCopy
Present to your audienceStart remote presentation
- Invited audience members will follow you as you navigate and present
- People invited to a presentation do not need a Prezi account
- This link expires 10 minutes after you close the presentation
- A maximum of 30 users can follow your presentation
- Learn more about this feature in our knowledge base article
Do you really want to delete this prezi?
Neither you, nor the coeditors you shared it with will be able to recover it again.
Make your likes visible on Facebook?
Connect your Facebook account to Prezi and let your likes appear on your timeline.
You can change this under Settings & Account at any time.
Spectroscopic Ellipsometry for the casual user
Transcript of Spectroscopic Ellipsometry for the casual user
(for the casual user) Spectroscopic -
Multiple energies /wavelengths Ellipsometry -
Exploits changes in polarization states of light (most generally an ellipse) The multiplicity of representations can make life confusing:
eV, nm, (1/cm), etc.
or Conversion isn't hard. The analysis software doesn't help with this confusion, even if you're sticking to J.A. Woollam's products: WVASE32 CompleteEASE -Requires a dongle
-Newer versions capable of converting .dat to .SE and .SE to .dat ! (If your tool has this for acquisition, talk to them about an update if you cannot do this!)
-Multi sample analysis is possible but somewhat painful
-Older program, but well supported for some time. Data files
SAVE EVERYTHING! .SE or .iSE (real-time)
.mod, new format
Snapshots .dat (text files)
.mod, old format
Environments -No dongle!
-Cannot currently convert data types
-Multi-sample analysis is easier/faster than WVASE
-"Neither complete, nor easy" The analysis capabilities of both programs are similar, however: Okay, we've wandered a bit into the details, and this is supposed to be for the casual user.
What were you trying to do again? I tend to group efforts made using SE into two types:
'Research' SE (sometimes hard)
Building NEW optical and/or structural models for unknown materials (substrates and/or films)
'Routine' SE (not too hard, but easy to get lost)
Measuring thickness/roughness of common or transparent materials on well known substrates (Native or Thermal SiO2, SiNx, Al2O3, on c-Si or SiO2)
Measuring thickness/roughness of new materials using models developed using Research SE Whether you're performing SE for research or routine measurement:
1. Acquire Data (Simple, but not foolproof)
2. Analyze Data (Harder) Data Acquisition: Important questions to ALWAYS ask yourself:
Is my SE system calibrated?
Is my sample well-aligned?
Is my substrate transparent
(over some or all of my spectral range)? Sanity checks are often quick and/or painless and may save you from yourself. Do a calibration (system check) before collecting data. Data obtained with a poorly calibrated instrument is notoriously hard to model. Alignment of samples prior to measurement is KEY. Don't rely on the single alignment often imposed prior to measurement. That should be a final check that your prior alignments were good, NOT THE ONLY ALIGNMENT YOU PERFORM! Data Acquisition: Important questions to SOMETIMES ask yourself:
Is my experimental
Is my sample stable? Elaborate experimental setups are often required for specialized study, however a sloppy setup may lead to headaches later on. Your system should be optically stable if you want to get useful data. Aging and oxidation are not trivial phenomena. Generally, these processes are fast (happen long before you measure) or slow (happen slowly relative to measurement) relative to your measurement time. Otherwise you should be taking real-time data. If so, what am I doing regarding backside reflections? Okay okay, say we have good data. THEN WHAT?! Your fit of any data consists of:
a) the physical model representing your sample stack (layers and thicknesses thereof)
b) the optical properties of the components in your model:
c) the specific parameters of the above which are allowed to vary Modeling and Fitting SE Data -Reference values
-Direct Inversion -Surface Roughness
-Substrate Alpha = 4 (pi) k / lambda Optical Models Generally, you're trying to assess the thicknesses and roughness of films. Thus 'bulk' film thickness and the thickness of the Surface roughness effective medium approximation (EMA) will often vary. If you have any dielectrics or underlayers, you SHOULD MEASURE AND MODEL THOSE separately FIRST! Most of the time, your experimental film will fall into one of the following categories: Transparent over most/all of your spectral range Absorbing over most/all of your spectral range Absorbing and Transparent within your spectral range Stoichiometric oxides, insulators, wide -gap materials Metals, Metal Oxides, other conductors Semiconductors with observable Eg, polymers, small molecule films, TCOs, etc. Transparent Materials k(E) = 0 implies e2(E) = 0
for all measured Energies, E. This means that you only need to modify the real part (n,e1) -Thickness fringes!
-Your sensitivity to surface roughness may not be so good Absorbing Materials Determining thickness for a heavily absorbing film may be difficult, especially if it is thick enough to absorb all incident light.
You can determine the optical properties and surface roughness of a thick metal film by treating it as semi-infinite, or exaggerating its thickness. Both: Use Tauc-Lorentz or Cody-Lorentz* to model asymmetric absorption behavior
In the infrared, try gaussians to account for vibration modes Most 'interesting' materials are partially absorbing and partially transparent.... Other Considerations:
-Your models should be as simple as you can make them but no simpler. Sometimes extra EMA layers are needed for complex samples, but more fit parameters may not always lead to a better result.
-Visual fit is important. If it looks bad, it probably is bad.
-ALWAYS check the error on your fit parameters. If the error on a parameter is not small relative to that parameter, YOU SHOULD BE WORRIED! This means that your parameter is coupled to something else or otherwise not changing the fit in a useful way.
-Understand the limitations of what you're doing: You don't get 'the answer' you get an answer whose accuracy and importance is relative to the quality of the measurement and the data analysis.