Inspection
deposition processes are terminated by time, it
is the deposition rate variation which is
responsible for film thickness variations. The run
to run variation in deposition rate was therefore
used as a basis for comparison purposes.
During the test, which consisted of 50
consecutive deposition runs, the cylinder of
dilute silane was purposely changed, in order to
simulate a real life manufacturing event. Typically
this would require a process re-qualification
since the exact percentage of silane in the gas
Issue VI 2009
mixture can be somewhat variable. In this
square4
instance we proceeded with the test with no
.com
additional qualification runs added.
The upper section of the graph in figure 5
(in red) shows the deposition rate variation that
was observed over the course of 50 wafers. Substantial time is saved, improving overall tool Fig4: Variation of
There is a gradual shift of 2-3% in the mean uptime and availability, ultimately improving the depositied film
deposition rate from run to run, with an obvious COO of the equipment. thickness over time
drop of approximately 2% that occurred when In a development environment, process
oasiasemiconductor
the silane cylinder was changed. For all 50 runs optimization can take place at a faster pace,
.eur
the mean deposition rate was 97.6nm/min with a since real time deposition rates can be obtained
www
standard deviation of 2.02nm/min, or 2.07%. in situ, without the need for constant process
The lower section of figure 5 (in blue) shows adjustments and off line measurements. In the
27
the film thickness variation that was measured Plasma-Therm laboratory this non intrusive
for the 50 wafers. Note that the thickness technique has been used during development of
reported is a single point measurement at the deposition processes for silicon dioxide, silicon
centre of the wafer since this coincides with OEI oxynitride and silicon nitride films with a wide
reflectance measurement point. This is range of refractive indices and deposition rates.
representative of the average wafer thickness An incidental advantage of this technique is
since the within wafer uniformity is good that the OEI configuration also works effectively
(approximately 0.2% sigma over 49 points). For as a conventional OES setup. This is useful in a
all 50 runs the mean deposited film thickness deposition system when used during periodic
was 298.2nm with a standard deviation of only plasma cleans. The endpoint (the point at which
1.1nm, or 0.37%. Further, the total range of all deposits have been removed) can be Fig 5: Deposition rate
deposited thickness was <4nm or +/- 0.65%. detected and the process terminated, thus and film thickness
This data shows there is an obvious avoiding excessive time being spent as is typical variation run to run
advantage to controlling film thickness based on using a time-based process.
a variable time process controlled by endpoint
versus a fixed time process. In this instance an
improvement in wafer to wafer film thickness
repeatability by a factor of 4-5 is realized.
Part of this improvement is a result of the
process shift which occurs after a gas cylinder
change, which would not normally be as
significant if process re-qualification had taken
place. Even if only the data from the first 30 runs
is considered (prior to the gas cylinder change)
the improvement is still significant: a standard
deviation of 0.36% (thickness variation) versus
0.9% (rate variation), or approximately a factor
of 3.
The lack of need for process re-qualification
and deposition rate measurement is significant.
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44