Electronic Data Sheets
Electronic Data Sheets allow us to
predict the outcome of laboratory processes based on prior
data. For each tool, a custom-tuned simulator can be used to model
the data in the database. This also allows us to track equipment performance.
As with any such effort, it only works if people submit their data
to the database, and if such data is measured reliably and accurately.
It takes less than a minute to enter your data, and less than a
minute to produce a custom graph with automatic data regression.
Entering Data to the Database:
To enter data only takes about a minute. First go to the eData website:
http://www.lelandstanfordjunior.com/eData.html
Select the appropriate tool or process module from the "Enter Data" column.
Fill out the form, and press "Enter Data".
That's it. Less than one minute.
Outside of the lab, you will need a username and password to access
the database. Contact me (ericp@snf) for that, or get it from
another lab user.
Viewing Data in the Database:
To view the data in the database, visit the same
eData page,
and select "View Data in the Log", or navigate from the link on the Data Entry form.
The data is above- you are positioned at the end (most recent end) of
the database. You can click on any column Header/Footer to sort the database
by that column. For instance, to sort by "Process", click on the "Process"
footer, and the database will be sorted by process name. You can sort
by multiple criteria by sequentially selecting the columns. For instance,
if you want the database sorted first by "Process" and then by "Time", you
should click on the "Time" column FIRST, then click on the "Process" column
(reverse order). This will arrange the database by process, but within each
process name, the data will be sorted by time.
Instructions for Producing Graphs:
With just a few clicks, you can produce a custom graph with exactly the
data you want.
First select the x-axis and y-axis for your plot with the buttons.
You must select both an "x" and a "y". Next, you can optionally select
criteria to filter the data set that will be plotted. If left blank,
all values for that field will be graphed. If any value(s) are specified for a
field, those value(s) will be the limiting criteria for the plot:
- For the process, type the name of the process (program) that you wish to
limit the data to. For example, entering LONH378 here will only plot data for
the LONH378 program. Multiple values are allowed. If you wanted to plot values
for NITRIDE1 and for NITRIDE2, enter NITRIDE1 NITRIDE2 (separated by a space).
- For Thickness, Time and Rate fields, you specify first an operator, either
> or < to limit to data which is greater than, or less than a particular value.
Then enter that value (without a space). EG: >2000 would plot
everything greater than 2000A, and exclude everything less than 2000A. If you
want a range, you simply enter both the upper limit and the lower limit, separated by a
space, EG: >2000 <9000 would plot all data between 2000A and 9000A.
- For Metrology, suppose that you only wanted data that was measured by nanospec:
You would simply enter nanospec in the field for Metrology. But,
you probably want both the litho-nanospec and the nanospec, so you'd enter
nanospec litho-nanospec. The metrology field matches the word(s)
as a logical OR- That is, it will include that data point whether the field is nanospec
OR litho-nanospec. nanospec rudolph-ellipsometer litho-nanospec
would take any of these three.
- The Email and Comment fields work similar to each other, but differently than
the metrology field. A search is executed
on whatever you enter here. So for instance, if you enter maurice
in the email field, it will select maurice, as well as
maurice@stanford, and even SomePeopleCallMeMaurice.
You can use this feature to cull your own data, or staff runs. These fields do not,
however, add the words to the search as the metrology field does. It searches for
any occurence of the exact word that you enter.
As a final note, all of the fields are mutually exclusive:
In The example:
| Date |
Process |
Thickness |
Time |
Rate |
Metrology |
Email |
Comments |
|---|
|
LONH378 |
<2000 |
>30 <120 |
<100 |
nanospec |
maurice |
pp= |
|---|
the data would be limited to:
All runs using Process LONH378, also having a thickness <2000
A, also with a
recipe time >30minutes but <120 minutes, also
with a rate <100A/min, also
measured by nanospec, also run by maurice, also where
the user noted the partial pressure (pp=).
Again, leaving any field blank will include all data from that field.
Data Regression
A least-squares regesssion line will also be plotted through the data (on the data
plotted, not the whole data set). The equation for
this regression line will be displayed in the bottom right of the graph to help you predict the
outcome. If your x-axis is "Date", the regression is limited to just a least-squares average
of the data points in the range.
Custom calculator
If you select Thickness (y) versus Time (x):
For the data set you selected, based on your inclusion criteria, a custom calculator will
be produced.
This calculator is based on the Data Regression performed on your data set. You should
carefully observe the plot to make sure that the fit (green line) is reasonable, and
that you are not extrapolating too far off of the data set.
This is as close as you can get to a custom-tuned simulator for your process.
For Thermal Oxidation:
For thermal oxidation, Thickness (y) versus Time (x) data, the fit will be based on
a Deal-Grove model. The "out-of-the-box" deal grove model will be plotted in blue,
the Deal-Grove fit is plotted in green.
For the special case of dry oxidation, a Massoud model (out-of-the-box) will be
plotted as well, in purple.
Comments? Suggestions? Complaints?
Email me: ericp@snf.
Run the oxide thickness calculator to get my full address.