D0 SMT Group Meeting Minutes: 2/19/98
(Minutes by Krish Gounder)
Present: The usual suspects from D0 SMT Group, Mont & Harry
and the new D0 Collaborators from Kansas State (Noel and Ron).
Micron Visit Report: (Uli/Ron)
Uli and Ron talked in detail about their trip to Micron recently.
Here is a detailed report of the trip from them in their own words:
2/15/97
The meeting began with a summary of the status of D0 detector processing
315 D0 stereo detectors and 151 D0 wedge detectors are in process. This
includes 200 at the start stage.
D0 Stereo Detectors
*The first side processed typically has fewer shorts. D0 asked that the
p-side be processed first since we expect to bias the detector with the
full depletion voltage on the p-side.
*Two types of silicon are being used:
* 111 Orientation at 5-10 Kohm-cm. This increases to 6-9 Kohm after
processing resulting in a depletion voltage of ~20V.
* 100 Orientation 3-6 Kohm-cm which will provide higher depletion voltage
*The current yield is close to 50%. (The 99.5% is considered an impossible
spec.) The detector leakage currents are getting better and tend to
decrease after cutting.
*Micron would like to know:
* D0's choice of silicon
* The side to process first.
* The quality of the polyimide coating is improving. The old polyimide
is streaky and patchy. The detectors were cleaned using
amine/water/amine. The water step is now being left out with improved
results.
*Details of oxide processing:
* diffusion implant
* 2400 A capacitive oxide
* Implant through oxide
* Remove oxide
* Reoxidize at 1000 degrees (2400-2600 angstroms)
* polysilicon for resistor imaging
* anneal at 900 dg for 1/2 hour
* PECVD oxide deposition (3100 A)
* weak p-etch on surface (remove 250 A)
Discussion of Nitride dielectric- MPI uses 100 A of silicon nitride grown at
800 deg. CSEM uses a plasma deposited nitride which would foul the CVD tanks
at Micron.
D0 Double metal detector
H1 devices have ~98% good via holes. Shorts are due to voids in the 4 micron
PCVD layer. Metal uniformity is crucial in 6". Need the new sputterer
which was ordered to achieve uniform metal thickness.
- - Micron expects to finish H1 double metal 3/15
- - Yield of DM process now >50%
Neil Greenwood discussed techniques to make reliable vias in the second metal.
A titanium mask is used to make the holes visible. Micron is working on a
process to planarize the structures using spin-on photoresist and an etch-back
technique.
Micron will make corrections to the DSDM masks by the end of next week. Masks
are expected to arrive at Micron by mid-March. The first set of 6" wafers
will include some with just double metalization and a set of single sided
double metal devices to develop the process. The single sided detectors can
also serve as a back-up solution if the DSDM process has problems.
The bulk Si material will be 100 orientation, high resistivity (6kohm-cm) to
result in depletion voltages of 27-35 V. 200 wafers have been purchased.
Passivation techniques will have to be developed (no passivation on H1
detectors). Polyimide may not work due to the high temperature cure. Silox
should work but may result in thinner than usual coat (less than 0.5 micro
meters) on thick oxide.
There was a production problem in Dec. when a HEPA filter ruptured. This has
caused increased numbers of shorts in all of the AC coupled detectors open
in that time.
Schedule Discussion
Micron presented a combined production schedule for CDF and D0. The schedule
assumes that a new sputtering system comes into operation at then end of
March. The system is necessary to increase metalization throughput and to
provide adequate metalization thickness. The schedule also assumes that
12 new people will be added to the Micron production staff (from the current
9 to 21). Seven new hires have been authorized. Micron also does not have
the capacity to fully probe all detectors during peak production. Several
alternatives were discussed:
1) Micron will purchase an additional automatic probe station
2) Some fraction of the detectors (~50%) are probed by D0 and CDF
3) Micron increases the throughput of the probing system
D0 and CDF requested some revisions to the schedule to provide joint CDF/D0
goals and a number of delivery milestones. The final revision of the schedule
is included as an appendix
In order to maintain the schedule, Fermilab has to accept prototypes and give
go-ahead for production on the following schedule:
* CDF/ISL end 3/98
* CDF/SVXII end 4/98
* D0 F-wedge end 3/98
* D0 90 deg end 9/98
Micron will monitor the probing and mask rate and provide weekly reports
of the number of wafer ar each process stage.
Detector Test Results: (Mike/Krish)
Mike presented a summary of the test results on 3 double-sided
detectors tested at Oklahoma. All three detectors have depletion voltages
higher than 20 V and leakage currents less than a few micro Amps.
Most capacitor breakdown voltages are higher than 180 V. But when the
polarity is reversed, they seem to breakdown consistently at about
120-30 V for all channels retested! The polysilicon resistances are in
the rabge of 1-3 M Ohms except one (~8 M Ohms)! The coupling capacitances
measured appeared much lower than the nominal value of 90-110 pF. After
some discussion, it was pinned down to a slight calculational error. These
tests will be repeated for verification. Ron suggested a special meeting
to discuss the various production test details the next day.
Krish reported the final results from the capacitance breakdown
testing at UC, Riverside. In addition to those reported at the January
Meeting, two more double sided detectors were bonded at Fermilab and
then were tested for breakdown. The results confirmed the earlier results.
There was almost no evidence of capacitor breakdown for voltages less
than 180 V for all channels tested (about 1000) in the case of
double-sided detectors. For 28 channels, the testing was done
before and after bonding. No observable systematic shifts in currents
were observed.
Burn in Results: (Evgeny/Cecilia)
Evgeny and Cecilia reported about their efforts with burn in tests
briefly. While Evgeny explained the some of the details of tests he
designing, Cecilia showed plots depicting pedestal levels, gain and
signal/noise ratio. She also described briefly about the CDF plans and
future plans for D0. Please see the meeting transparencies for more
details.
Proposal for CMS fabrication at SiDet: (Petros)
Due to time limitations, Petros presented a proposal to join the
efforts of CMS pixel detector group very briefly. He also handed out
a copy of written report for such a possible effort.