Several
techniques exist for estimating the
potential lifetime of
a planar diffusion source. One method is to periodically dope
a silicon wafer with a source held in a diffusion furnace
and observe
how the resulting sheet resistivity varies with time. Figure
3 shows the sheet resistivity obtained on silicon wafers doped
for 30 minutes with GS-139 BoronPlus sources that were held
at 975°C and 1025°C. Little change is observed during
the 700-800 hours of test time. Similar data are shown for
GS-245 sources.
A
second method of estimating the source’s lifetime is
to measure the amount of weight loss at a use temperature
as the B2O3
evolves. When the source no longer loses weight, the evolution
of B2O3
has ended. A weight loss curve plotted in Figure 4 for the
GS-139 BoronPlus source also indicates that hundreds of hours
of use are available at 975°C.
The
typical weight loss curve in Figure 4 shows that the B2O3
evolution rate from the BoronPlus sources gradually decreases
with continued use time. In fact, the linear relationship
obtained when the data are plotted versus the square root
of use time (Figure 5) indicates that the B2O3
is actually evolving through a diffusion-controlled process.
This means that as B2O3 evolves from
the surface of the source, a B2O3 concentration
gradient will develop from its interior. This gradient causes
additional B2O3 to diffuse to the surface
replenishing the supply of B2O3 needed
for continued evolution. This “reservoir” of B2O3
contained within the diffusion source is sufficient for many
hours of use. Holding the B2O3 within
the source instead of on its surface is also directly responsible
for decreased moisture absorption from the room air between
runs.
The actual
lifetime of BoronPlus sources used in typical plant production
environments depends upon many factors such as temperature
of use, care in handling, the device being manufactured, the
sensitivity of the process to the gradual decreasing evolution
rate, etc. Typical use times exhibiting acceptable results
of up to 500 hours near 1000°C and up to 150 hours near
1100°C have been reported.
Doping
Properties of BoronPlus Sources
A
typical set of sheet resistivity versus deposition time curves
for temperatures ranging from 850°C to 1150°C are
plotted in Figure 6. Each user should determine similar sets
of curves that are characteristic of their diffusion process,
since these will depend somewhat on such parameters as the
type of BoronPlus source being used, furnace recovery and
heating/cooling rates, gas flow rates, etc.
When
the various processing conditions are optimized, uniformities
of 2% across the silicon, 3% across the boat, and 4% run-to-run
should be attainable. Although these uniformities are generally
considered to be typical of the planar diffusion system, most
processors have been able to significantly improve over them.
Several
techniques exist for estimating the
potential lifetime of
a planar diffusion source. One method is to periodically dope
a silicon wafer with a source held in a diffusion furnace
and observe
how the resulting sheet resistivity varies with time. Figure
3 shows the sheet resistivity obtained on silicon wafers doped
for 30 minutes with GS-139 BoronPlus sources that were held
at 975°C and 1025°C. Little change is observed during
the 700-800 hours of test time. Similar data are shown for
GS-245 sources.
A
second method of estimating the source’s lifetime is
to measure the amount of weight loss at a use temperature
as the B2O3
evolves. When the source no longer loses weight, the evolution
of B2O3
has ended. A weight loss curve plotted in Figure 4 for the
GS-139 BoronPlus source also indicates that hundreds of hours
of use are available at 975°C.
The
typical weight loss curve in Figure 4 shows that the B2O3
evolution rate from the BoronPlus sources gradually decreases
with continued use time. In fact, the linear relationship
obtained when the data are plotted versus the square root
of use time (Figure 5) indicates that the B2O3
is actually evolving through a diffusion-controlled process.
This means that as B2O3 evolves from
the surface of the source, a B2O3 concentration
gradient will develop from its interior. This gradient causes
additional B2O3 to diffuse to the surface
replenishing the supply of B2O3 needed
for continued evolution. This “reservoir” of B2O3
contained within the diffusion source is sufficient for many
hours of use. Holding the B2O3 within
the source instead of on its surface is also directly responsible
for decreased moisture absorption from the room air between
runs. 
A
typical set of sheet resistivity versus deposition time curves
for temperatures ranging from 850°C to 1150°C are
plotted in Figure 6. Each user should determine similar sets
of curves that are characteristic of their diffusion process,
since these will depend somewhat on such parameters as the
type of BoronPlus source being used, furnace recovery and
heating/cooling rates, gas flow rates, etc.
When
the various processing conditions are optimized, uniformities
of 2% across the silicon, 3% across the boat, and 4% run-to-run
should be attainable. Although these uniformities are generally
considered to be typical of the planar diffusion system, most
processors have been able to significantly improve over them.