Please pardon my lengthy response

Actually, in direct bonding applications, there is generally no need to use
a "fixture" to hold the two wafers together.  Of course, this does depend on
the makeup of the substrates being bonded.

In the case of silicon, there is something that can be performed prior to
the placement into the 1000 deg C furnace called a "prebond".  Basically,
this prebond is the fixing of two mirror polished (roughness less than 0.5
nm for best results) substrates to each other via  Van der Waals forces.
This can be performed with either hydrophobic or hydrophilic surfaces,
although the hydrophilic case provides the strongest prebond, with a bond
strength of up to 150 mJ/m2 (by contrast, to fully annealed directly bonded
silicon wafers have a bond strength of 2 J/m2).  This prebond can be
observed by merely removing two prime wafers from their cassette, and
pressing the polished sides together (although the bond strength is much
weaker).  This is not stiction, as stiction affects will typically still
allow shear forces to easily move one wafer across the other.  If the
prebond has occurred, the wafers will not be able to slide against one
another, but instead remain fixed in place (thus, maintaining any kind of
alignment that has been performed).

Several cleaning methods can be used to increase the surface energy of the
substrates, to increase the strength of the prebond.  The purpose of these
cleaning methods is to create a surface terminated in hydroxyl (-OH) groups,
which will promote hydrogen bonding between the two substrates.  RCA clean,
Piranha, megasonically activated DI rinse, reverse RCA, and boiling nitric
acid have all been used to increase bond strength.  It is very important
that no HF dip process is used at any point prior to forming the pre-bond,
as this will strip away the hydrophilic surface we are attempting to
achieve. In our studies, the boiling nitric produced the strongest prebond,
followed very closely by the reversed RCA process.  Because of safety
issues, we prefer the revered RCA process (listed below).  In addition to
wet chemical activation methods, some dry methods can be used, such as
oxygen plasma, or UV/Ozone cleaning.  A final DI flush is recommended just
prior to bonding, regardless of the method used for cleaning.

Once the surfaces have been activated, they can be bonded by simply placing
one substrate face up on a table, placing the other face down directly over
(a thin layer of air will prevent direct contact between them, assuming the
second is laid down directly parallel to the first), and forcing them into
contact at one point.  The bond front that begins at that point will extend
to the full interface in a few seconds.  This bond front can be observed
using any sort of full wafer IR inspection set up.  Please note that, since
this method relies on the surfaces of the wafers coming into direct contact,
ant particle contamination will seriously hinder the prebond, creating voids
up to 4 orders of magnitude greater than the size of the particle.  For
instance, a 1 um particle can cause a 1 cm wide void along bond interface.
Because of the annealing temperature (above 1000 deg C), any air trapped in
these voids may cause delamination, as the overall pressure in the void will
increase from 1 atm to approximately 5 atm.

For the reason mentioned above, please note that substrates can only be
bonded in atmosphere if they contain no cavities (except through holes).  In
order to bond wafers to create sealed cavities, the bond must be performed
in some level of vacuum (typically, 200 mbar absolute would be sufficient).


Once the prebond is performed, assuming there is no trapped air, the wafers
will remain fully fixed in place during the anneal process.  Once its done,
the anneal can be performed by simply placing the wafer stack into a furnace
in the same manner as you would place a single wafer into the furncae.  No
clamping fixtures are necessary.

Reverse RCA clean (RCA2 then RCA1 no HF dip)

        RCA2 (SC-2)
        HCl (37%): H2O2 (30%):H2O in a volume ratio of 1:1:6 at 700C for 5
minutes
        DI water rinse

        RCA1 (SC-1)
        NH4OH (29%):H2O2 (30%):H2O in a volume ratio of 1:1:5 at 700C for 5
minutes
        DI water rinse

Best Regards,

Chad Brubaker



EV Group-Technology, Tel: (602) 437 9492 x 111, Fax: (602) 437 9435
E-mail: [email protected] , Web:
www.EVGroup.com



-----Original Message-----
From: shay kaplan [mailto:[email protected]]
Sent: Monday, December 10, 2001 7:46 AM
To: [email protected]
Subject: Re: [mems-talk] fixture for bonding two wafers at 1000 deg C


try two quartz (fused silica) photomasks with the chrome stripped off.

shay

[email protected] wrote:

> Sir/Madam,
>          In the abscense of a proper bonding equipment; Can anyone tell
> me , "How two wafers can be held together in a furnace/oven", to perform
> a "Direct bonding process".
>    To be more precise, what kind of fixture materials can be used for
> temperatures in the range of 1000 deg.C.  If anyone knows any company that
> provides these kind of fixtures , please forward the information.
>
> Regards,
> javeed shaikh mohammed.
> _______________________________________________
> [email protected] mailing list: to unsubscribe or change your list
> options, visit http://mail.mems-exchange.org/mailman/listinfo/mems-talk
> Hosted by the MEMS Exchange, providers of MEMS processing services.
> Visit us at http://www.mems-exchange.org/
_______________________________________________
[email protected] mailing list: to unsubscribe or change your list
options, visit http://mail.mems-exchange.org/mailman/listinfo/mems-talk
Hosted by the MEMS Exchange, providers of MEMS processing services.
Visit us at http://www.mems-exchange.org/

[demime 0.98e removed an attachment of type application/octet-stream which had a
name of Chad W Brubaker (E-mail).vcf]