KOH Etch Rate Calculator
This calculator determines the etch rate of several major silicon crystal planes and thermally grown silicon dioxide (SiO2). KOH as a silicon etchant has been known in the semiconductor industry almost since the industry's inception. 
The model for KOH etching of silicon has evolved over the years and this calculator coallesces the best parts of different data and models into one simulation.
Today, anisotropic alkaline etching of silicon is perhaps one of the most important steps in the manufacture of silicon MEMS devices, used in both bulk and surface micromachining processes.
This calculator computes etch rates of silicon <100>, <110> and <111> crystal planes and SiO2 versus solution temperature and solution percent weight, and incorporates the effects of substrate doping and the addition of isopropyl alcohol to the solution.
Is the Calculator Exactly Correct?
How Accurate is it?
This calculator is really a simulation, based on experimental data from many sources. There is a large amount of data
available for <100> undoped Si, and for SiO2. For other crystal orientations, there is less. For doped silicon,
the data is sparse. For different crystal planes and doped Si with the addition of IPA, there is almost none. In fact,
at higher %wt for doped Si with IPA, this simulation is a complete extrapolation of the data. So results may vary.
The mechanisms of etching are not universally agreed upon in the literature. 
Has the calculator been tested?
Yes, but not for all possible cases. The most relaible data is for KOH etch rates of silicon <100> and <110>
and for thermally grown SiO2. For Silicon <111>, the measurement of etch rate is ellusive. The etch rates are small, and may change with time. It has also been reported that the etch rate may depend on the specific geometry of the sample being etched, including the open area, nearby actively etching <100> planes, and surface orientation.
Also, Boron etch stop models are not as thoroughly tested. For one, it is often not known what the exact B concnetration is in a particular sample. The B concentration may not be constant in depth. And, like in the <111> case, the rates tend to be low, and therefore more difficult to measure accurately.
How do you blend data from many sources into one coherent model?
This is the art of science, statistics and modeling. Data is taken from many sources, analyzed statistically to see what data is consistent with the whole set and which data might be errant. Then, using a system of equations with physical
relevance (not a random curve fit), a best-fit is made through available and statistically selected data using regression and enhanced visualization tools.
Do you improve the models as more data becomes availbale?
Yes! In fact, you can help that effort by sharing your data, and it will be tested and incorporated where possible.
It is important to include all experimental detail, such as how you measured (particularly, doping concentration).
The following data would be particularly interesting:
Go to the calculator main page:
and enter data according to:
- Any <111> data-- it is vital to explain how it was measured (eg: undercut method, <111> wafers, etc)
- Any doped Si data -- again, measurement details important. Especially interesting would be B-Si data over %wt.
- Any IPA solution data.
- Doped Si <110> and <111> data, with and without IPA
- Percent Weight of KOH in solution:
- General: This is the amount of KOH in the solution used to etch the sample.
- Purpose: Etch rates are strong functions of solution chemistry.
- Options: 10% to 55%
- Units: %
- Default: This defaults to a commonly used 30% solution.
- limitations: 10 to 55 Percent Weight KOH
The MEMS/semiconductor industry has become accustomed to percent units, usually percent by weight.
The definition, used here and most places, is %Weight = Weight of KOH / (Weight of KOH + Weight of Water).
Stronger solution does not always imply faster etching. Etch rate usually
peaks somewhere in the intermediate concentrations.
- General: This is the Temperature, in degrees Celcius, at which the etching is carried out.
- Purpose: The etching process is highly sensitive to temperature.
- Options: 20°C-100°C.
- Units: degrees Celcius
- Default: set to 80°C, as a commonly used temperature.
- specifics/details: Higher temperature will cause faster etching. However, temperature can be used to
cause the relative etch rate between two entities to change (IE: the etch selectivity). For instance, going to lower temperatures lowers the etch rate of all surfaces, but it affects the etch rate of Si <100> less than it affects the etch rate of SiO2, and therefore the selectivity of Si <100> to SiO2 improves by lowering the temperature, though the same amount of Si <100> etched will take longer.
- Select the Model Options:
- Saturated IPA:
- General: Etch rates and selectivities can be altered with the addition of Isopropanol (Isopropyl Alcohol).
- Purpose: If IPA is added to the KOH, check this box to indicate a saturated solution.
- Options: IPA (box checked) or no IPA (box unchecked)
- Units: (none: saturation)
- Default: no IPA (box unchecked)
- limitations: Only a saturated solution or no IPA is allowed.
- specifics/details: Only saturated solutions are considered, because KOH solution saturates with the addition of very small amounts of IPA and any condition with IPA below saturation is very difficult to maintain over time. IPA rapidly evaporates from the solution, making precise control difficult, even with a condensor apparatus.
Saturation is easy to see with the naked eye, as a stratified layer of different optical index floating on top of the KOH solution. As long as this layer is present, it will maintain the solution in saturation. IPA can be periodically added as it evaporates to maintain a layer. Note that this is assumed to be "100%" IPA, not the more dilute variety that is sold at the pharmacy.
- Active Doping Concentration:
- General: This is the concentration of active dopants in the silicon at the region where the etching will occur.
- Purpose: There is a strong dependence on p-type doping (Eg: Boron).  The retardation in etching is related to the electrical carriers (electrons and holes) in silicon, not the actual dopant atoms directly. This effect is not significant below about 1e18 /cm3.
- Options: Value up to about 10% of silicon atomic density.
- Units: concentration per cubic centimeter. (eg: 1e20 /cm3)
- Default: Defaults to 1e14 /cm3, which is below the level of having any impact on etch rate. This default is set when doping is selected, only to illustrate the entry format (ie: exponential notation).
- limitations: Only accurate for the <100> etch rate. Little data is available to confirm the impact on other crystal axes.
- specifics/details: Specify the active carrier concentration, not the chemical doping level to get the correct answer. If the concentration that you specify is too low, then the calculator will decide not to use doping models. You must select the "Substrate has heavy p doping" option for this model to be invoked. It is up to you to account for solid solubility, clustering, or other effects which might limit the dopant electrical activation.
- Click Calculate Etch Rates, and the following OUTPUTS will appear:
- Etch Rates for three Crystal Orientations and SiO2:
- General: This is the orientation of the silicon crystal surface (or SiO2) on which you are etching.
- Purpose: KOH attacks some crystal planes significantly faster than others. 
- Possibilities: <100>, <111>, <110> and SiO2.
- Units: (standard miller indices for crystal orientation)
- Default: Etch rates for all three major axes are returned by the calculator along with SiO2
- limitations: <100>, <111>, <110>, SiO2.
- specifics/details: The most common silicon wafer has a <100> surface orientation, though the other major orientations are available for special purposes. The calculator returns the etch rate of all three axes, though the <111> etch rate may depend on many factors not taken into account. 
The angles between the various common planes are:
Complements of angles also possible, not shown
|(110)||90°||90°||35.26° or 90°|
|(111)||54.7°||35.26° or 90°||70.53°|
- Various comments will appear to notify you of the quality of the data. For instance, if you have requested
etch rates for conditions that are not tested, or based on extrapolated data, these comments will so indicate.
- Comments may also alert you to known issues with the solution you selected, eg: rough surface may result. 
- Solution Mixing Instructions:
Many times, things don't come out right because solutions were improperly mixed. Much of the data available is corrupted by improperly mixed solutions, which if known can be fixed by adjusting the reported %wt with a corrected value. In an effort to reverse this trend, the calculator provides mixing instructions for the solution you've requested:
- Instructions are provided for mixing from a 45% stock solution. This is a commonly sold liquid concentration found in the laboratory.
Obviously, you are limited to 45% or less solution concentration if you use this stock solution.
- Instructions are provided for mixing from powdered KOH, which may be anhydrous (completely dry, without adsorbed water) or may be in various states of hydration. The container should specify the state of hydration (anyhydrous/monohydrate/dihydrate .../hexahydrate) or specify hydration as "KOH * 2H2O" which would be the same as "potassium hydroxide di-hydrate". In either case, the calculator allows you to specify the state of hydration for proper mixing instructions. Just put the number
in the box provided: 0=anhydrous, 1=monohydrate, 2=dihydrate, etc. The calculator will handle the proper accounting of the hydrated powder.
What's not accounted for in this calculator?
- Commercial KOH, even semiconductor-grade, is known to have significant impurities (eg: NaOH), which this calculator cannot predict
Under what conditions might the calculator have significant error?
© 2013-2014 Maria Pace. All Rights Reserved.
Send comments or suggestions to Maria Pace mariaepscience@ (run the calculator to get my full email).
- Many factors can alter experiments & measurements, for instance diffusion effects in certain feature geometries.
- As always, Garbage-in/Garbage-out: If your %wt is not known or calculated improperly , or temperature is poorly calibrated or non-uniform, experiment will not agree with this calculator. If your silicon is not pure, or has significant crystal damage, rates will vary.
- If your KOH powder has been unsealed (factory seal broken) for any length of time, the true state of hydration is likely unknown. Just like in cooking, Use fresh ingredients.
- Short-time etches, or frequently interrupted etches may not allow etch rate to stabilize.
Referencing this Calculator in a Publication:
The MLA7 format for citing a website is:
Last name, First name. "Article Title." Website Title. Publisher of Website, Day Month Year article was published. Web. Day Month Year article was accessed. <URL>.
Therefore, a recommended format for referencing this calculator is:
Pace, Maria. "KOH Silicon Etch Calculator." N.p., June 2013. Web. 08 Apr. 2014. <http://lelandstanfordjunior.com/KOH.html>.
You should change the accessed date to the date you accessed it
for your paper, and for the date of the page, it is recommended that you look at the "code base" at the bottom
of the results page to reference the date (so that any changes in the code can be tracked):
Calculator code build: M1312
For example, would be best to reference the page date as Dec 2013
Calculator code build: M1312 I1401
would be referenced at Jan 2014 since this is the most recent date of code modification for the particular sections of
code that your simulation used. Although non-standard from a library reference point of view, you
could just reference the code build.