Hello to all,

Some time ago, I asked a question on "alignment to crystal planes".
I have compiled the answers to the request.
There are 2 suggestions emerging:
A) ALIGN TO WAFER FLAT        (simple)
B) PRE-ETCH and ALIGN TO BEST ETCHED SHAPE    (more precise)



here are the replies I received:


A) ALIGN TO WAFER FLAT

1-  [email protected] (Satish Prabhakaran)
[...]"I work with 100 type silicon wafers and have been doing anisotropic
etching on them. My patterns are v-grooves and hence have to be aligned parallel
to the 110 plane of silicon. I found it easy by generating a mask which had a
long straight line  parallel to my patterns. This line , when transferred onto
the wafer (you could design it such that it is transferred to the edge of the
wafer) can be easily aligned with the primary flat of 100 silicon (the 110
plane).  The patterns on the wafer are all now aligned parallel to the 110
plane.

This is a very simple approach and works fine with a little care while aligning.
You could use a magnifier during this step. I guess, a similar approach would
work for other crystallographic plane. Hope this is useful.

Good luck.

Satish Prabhakaran
Graduate Student
Thayer School of Engineering
Dartmouth College
Hanover, NH - 03755
USA
"

2-  Marc DeVincentis 
[...]"I only know of one way, and it has suited my purposes fine so far.  Your
wafers should come with a tolerance specified on the primary flat, like {110}
+/- 5%. Just align your photomask to the flat.  I'm not doing anything too
precise, so the tolerance I get from the wafer supply has been good enough.
But, maybe you could specify how close you need, like {110} +/- 1%.  Then you
are limited in precision only by 1) your alignment to the flat, and 2) the
tolerance delivered by the wafer supply.  The first error source is probably
pretty small, given the accuracy of even modest photolith setups.  So let the
wafer supplier do the work for you.  They should have an x-ray crystallographic
projection setup to monitor quality control anyway.  Make them put it to good
use.  If you need to work with small pieces of silicon, just dice parallel to
the flat, and align to the dice cut.

Hope this helps,
Marc
"

5-  Ninad Shinde 
"Do you think aligning the wafer flat with the flat on the mask aligner will not
do a good enough job ? You could align those initially when you load the wafer.
Another way would be to have some sort of reference line or tape on parallel to
the edges of the photomask. It should'nt be a problem to align the photomask to
the edges of the mask holder. You could then align the wafer flat with the
reference line on the mask. I will be doing something of this sort in the next
month or so with my pressure senors and could let you know how it worked out.

-ninad
"

B) PRE-ETCH and ALIGN TO BEST ETCHED SHAPE

3-  Briand Danick 
[...]"on fait une mini attaque KOH (TMAH) avec masque d'oxyde au debut du
process, peut se faire aussi avant l'attaque finale, pour reveler l'orientation
des plans cristallins. Le masque est fait de rectangles a differents angles qui
permettent de determiner l'orientation du plan d'arret d'attaque (111).
"[...]
Here is an english translation:
We have done a small KOH (TMAH) etch with an oxide mask at the begining of the
process (it can also be done right before the final etch), to reveal the the
orientation of the crystal planes.  The mask is made of rectangles at different
angles to help in determining the orientation of the {111} etchstop planes.

suggested reference:
"J. MicroMecha. Microeng. 8 (1998) 327-329
"


4-  "Schubert, W Kent" 
"See the article by M. Mattias and Y. Backlund in J. Micromech. Microeng., Vol
6, 279(1996).  Their method requires an extra patterning step and a short
anisotropic etch (about 15 minutes).  We have tried several methods and this is
the best we've found.

W. Kent Schubert
Microsensors R&D Department
PO Box 5800
MS1425
Sandia National Laboratories
Albuquerque, NM 87185-1425

phone:  505-844-5935
"

I believe the reference is: M. Vangbo and Y. Backlund in J. Micromech.
Microeng., Vol 6, 279(1996)


6-  "Wilson, Alan" 
"If you want a really accurate alignment to the crystal planes you could do a
test alignment to form a trench  structure. Over etch the trench in KOH or TMAH
and the etch will show you the exact orientation of the crystal planes. The
longer the trench the more accurately you will be able to align to it.

Dr Alan Wilson
Defence Science and Technology Organisation
Australia
E-mail: [email protected]
Tel: +61 3 9626 7508
Fax: +61 3 9626 7087
"

7-  [email protected] (Mattias Vangbo)
"I wrote an article in JMM in -96 that might be of help. We did +/-0.05° -
limited by the pattern rather than the method. However, the technique presented
requires a pre-etch to determine the direction before the patterning and etching
of your actual pattern. The pre-etch pattern used is available in GDSII and CIF
but I think you should check what it is all about before I send the file.
Please, refere to 'Precise mask alignment to the crystallographic orientation of
silicon wafers using wet anisotropic etching' Mattias Vangbo and Ylva Bäcklund
J. Micromech. Microeng., 6 (1996), p. 279-284.

(It was also adopted for InP:' High precision crystallographic alignment of
(100) InP'
Mattias Vangbo, Åse Richard, Mikael Karlsson and Klas Hjort J. Electrochem. Soc.
Letters, 2 (1999), p.407-408.)

Regards,
Mattias Vangbo
"


8- Mirza Andy 
[...]"Have you tried patterning a circlular mask (e.g. with silicon nitride) on
your wafer and then performing a KOH etch to undercut the circular mask to
reveal the <111> crystal planes which could them be used for alignment ?

Regards,
Andy Mirza
Technology Manager

EV Group-Technology, Tel: (602) 437 9492 x 112, Fax: (602) 437 9435
E-mail: [email protected] , Web
www.EVGroup.com
"

9- Ken Westra 
[...]"  There are a number of ways you can align a pattern to a crystal flat.
You can simply align the major flat to a line on the mask. We were using contact
mode mask aligners, so we could get the line and the flat aligned closely (0.1
to 0.2 degrees is easily possible with a bit of experience). This coupled with a
wafer with good alignment between the flat and the crystal planes (+/- 0.5
degrees iseasy to get for prime wafers), will give you nice looking V-grooves.
Another trick is to use a pre-etch. That is, do a anistropic etch to create a
V-groove pit and use the sides of the pit to align the mask. For best results
you need two pits far part on the wafer to get good alignment. But, this adds an
extra litho/etch/anisotropic etch step to the process.
I found taht in my work, simply buying good wafers (i.e. Xstal plane to flat
alignment), then aligning to the major flat worked well for most projects.
"[...]

Thank you all,
Robert Antaki


This is my original request:


To:   [email protected]

Subject:  alignment to crystal planes

Hello to all,

I am looking for different approches to align precisely a photolithography mask
to the crystallographic planes of silicon. The reason is simple, I will do an
anisotropic wet etch with the transfered patterns and require them to be well
aligned to the crystal structure for the etch to proceed properly.

I work with standard 150mm {100} silicon wafers, but comments from other fields
are welcome.

Regards,

Robert Antaki
Implantation engineer,
Mitel S.C.C.
Bromont, Qc, Canada