Thermal measuring and testing – Temperature measurement – In spaced noncontact relationship to specimen
Reexamination Certificate
1998-03-18
2001-01-30
Hirshfeld, Andrew H. (Department: 2859)
Thermal measuring and testing
Temperature measurement
In spaced noncontact relationship to specimen
C374S002000, C374S110000, C374S126000
Reexamination Certificate
active
06179466
ABSTRACT:
BACKGROUND OF THE INVENTION
The invention relates to techniques for making improved noncontact temperature measurements of a semiconductor substrate by correcting measurements of substrate temperature and by compensating for the emissivity sensitivity variations across the surface of the substrate.
In many semiconductor device manufacturing processes, the required high levels of device performance, yield, and process repeatability can only be achieved if the temperature of a substrate (e.g., a semiconductor wafer) is tightly controlled during processing. To achieve that level of control, it is often necessary to measure the substrate temperature in real time and in situ, so that any unexpected temperature variations can be immediately detected and corrected for.
For example, consider rapid thermal processing (RTP), which is used for several different fabrication processes, including rapid thermal annealing (RTA), rapid thermal cleaning (RTC), rapid thermal chemical vapor deposition (RTCVD), rapid thermal oxidation (RTO), and rapid thermal nitridation (RTN). In the particular application of CMOS gate dielectric formation by RTO or RTN, thickness, growth temperature, and uniformity of the gate dielectrics are critical parameters that influence the overall device performance and fabrication yield. Currently, CMOS devices are being made with dielectric layers that are only 60-80 Å thick and for which thickness uniformity must be held within ±2 Å. This level of uniformity requires that temperature variations across the substrate during high temperature processing cannot exceed a few ° C.
The wafer itself often cannot tolerate even small temperature differentials during high temperature processing. If the temperature difference is allowed to rise above 1-2° C./cm at 1200° C., the resulting stress is likely to cause slip in the silicon crystal. The resulting slip planes will destroy any devices through which they pass. To achieve that level of temperature uniformity, reliable real-time, multi-point temperature measurements for closed-loop temperature control are necessary.
Optical pyrometry is being widely used for measuring temperatures in RTP systems. Pyrometry exploits a general property of objects, namely, that objects emit radiation with a particular spectral content and intensity that is characteristic of their temperature. Thus, by measuring the emitted radiation, the objects temperature can be determined. A pyrometer measures the emitted radiation intensity and performs the appropriate conversion to obtain temperature (T). The relationship between spectral emitted intensity and temperature depends on the spectral emissivity of the substrate and the ideal blackbody radiation-temperature relationship, given by Planck's law:
I
b
(
λ
,
T
)
=
2
C
1
λ
5
(
ⅇ
c
2
λ
T
-
1
)
(
1
)
where C
1
and C
2
are known constants, &lgr; is the radiation wavelength of interest, and T is the substrate temperature measured in ° K. According to an approximation known as Wein's distribution law, this expression can be rewritten as follows:
I
b
(
λ
,
T
)
≈
K
(
λ
)
·
ⅇ
-
C
2
λ
T
(
2
)
where K(&lgr;)=2C
1
/&lgr;
5
. This is a good approximation for temperatures below about 2700° C.
The spectral emissivity &egr;(&lgr;,T) of an object is the ratio of its emitted spectral intensity I(&lgr;,T) to that of a blackbody at the same temperature I
b
(&lgr;,T). That is,
ε
=
I
(
λ
,
T
)
I
b
(
λ
,
T
)
.
(
3
)
Since C
1
and C
2
are known constants, under ideal conditions, the temperature of the wafer can be accurately determined if &egr;(&lgr;,T) is known.
As was described previously, variations in temperature across the surface of a substrate of more than a degree or two may result in damage to the substrate and undesirable process variations. One method for monitoring the temperature at various localized regions of the substrate includes the use of a plurality of temperature probes (pyrometers or the like). In these multi-probe systems, temperature readings from the various probes can be used for real-time control of the heating element in the RTP of substrates.
However, despite its widespread use in the semiconductor industry, optical pyrometry still suffers from limitations due to an inability to accurately measure the emissivity of the substrate. Moreover, even if the emissivity of the substrate is known at a given temperature, it may change with temperature. The changes are usually not accurately measurable and thus they introduce an unknown error into the temperature measurements. Errors on the order of 10° C. or more are not uncommon.
The spectral emissivity of a substrate depends on many factors, including the characteristics of the wafer itself (e.g. temperature, surface roughness, doping level of various impurities, material composition and thickness of surface layers) and the process history of the wafer. Another related term is the effective emissivity of an object. The effective emissivity is the ratio of the measured spectral intensity emitted by the object to that of a blackbody at the same temperature. The effective emissivity of an object differs from the spectral emissivity for the same object in that the effective emissivity takes into account the environment in which the object resides. The effective emissivity of a substrate may be influenced by the characteristics of the process chamber in which the substrate is positioned. Therefore, a priori estimation of substrate emissivity cannot provide a general purpose pyrometric temperature measurement capability.
In addition, the environment at each probe in a multi-probe system is unique. A pyrometric probe positioned in one of these unique environments may exhibit sensitivity to substrates having particular emissivity characteristics, introducing an error component in the temperature reading. Across the surface of the substrate, one or more of the probes may exhibit differing sensitivities to the emissivity of the substrate (hereinafter referred to as emissivity sensitivity across the surface of the substrate). Substrates having a generally low emissivity level may exhibit large variations in emissivity sensitivity over the surface of the substrate. Accordingly, a multi-probe temperature measurement system which does not account for variations in the emissivity sensitivity across the surface of the substrate may produce less than optimal results.
Systems which only seek to compensate for emissivity errors by a singular approximation of emissivity for the entire substrate will lead to acceptable results; however, room for improvement exists.
SUMMARY
In one aspect, the invention provides a method for correcting temperature probe readings in a thermal processing chamber for heating a substrate and includes calibrating the thermal processing chamber to derive emissivity sensitivity factors for each probe, forming a reflecting cavity on one side of the substrate and heating the substrate to a process temperature. Using a first probe, a second probe and at least a third probe energy may be sampled from the reflecting cavity. The first probe may have associated therewith a first effective reflectivity and the second probe may have associated therewith a second effective reflectivity, wherein the first and second effective reflectivities are different. The sampled energy from the first second and third probes may produce first, second and third temperature indications, respectively. From the first and second temperature indications, an effective emissivity for the substrate may be derived along with a corrected temperature reading for the first probe. The corrected temperature reading may be derived from the sum of the first temperature indication and an adjustment temperature. The corrected temperature reading for the first probe may be a more accurate indicator of an actual temperature of the substrate in the environment of the first probe than are uncorrected
Hunter Aaron
Knoot Peter
Mershon Jason
Miner Gary E.
Peuse Bruce W.
Applied Materials Inc.
Fish & Richardson
Hirshfeld Andrew H.
LandOfFree
Method and apparatus for measuring substrate temperatures does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for measuring substrate temperatures, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for measuring substrate temperatures will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2461622