Text preview for : TECHBRIEF7.pdf part of HP TECHBRIEF7 HP Publikacje TECHBRIEF7.pdf
Back to : TECHBRIEF7.pdf | Home
High-Frequency, Over-Temperature
Measurements
Recent advances in thermal probing stability of your system should be
equipment have dramatically greater than -60 dB or 0.1%. You
reduced the cost of obtaining tem- need to perform a new calibration
perature-dependent S-parameter whenever the stability reaches
data, while increasing throughput -40 dB or 1%.
capabilities. These advances make These factors affect calibration
thermal device modeling and stability:
process characterization both practi-
l Instrument variations over time.
cal and affordable. The extreme
l Probe placement errors resulting
temperature range of interest
from the wafer chucks thermal
(-65o C to +2OO" C), however,
expansion.
increases the complexity of the cali-
l Changes in the calibration stan-
bration and measurement consider-
dards with temperature.
ations.
l Electrical drift, due to probe and
cable temperature variations.
General Extraction Method To ensure measurement integrity,
The most common extraction check calibration stability frequently.
methods for temperature and bias- As the chuck changes tempera-
dependent device modeling utilize Figure 1. Cascade's MicroChamber found on
ture, it expands and contracts. The
o
cold-FET, pinched-FET, and hot- both manual and semi-automatic probing sys-
total chuck expansion (from -65 C
FET S-parameter measurements tems.
to +2OO" C) is about 9 mils. This
performed under bias conditions. affects probe overtravel. At each
Recent studies suggest that such Thermal Probing temperature, you need to adjust the
measurements uniquely determine overtravel of the probe tips. In addi-
Essential features of a thermal prob-
the device ECPs (equivalent circuit tion, the wafer diameter changes
ing system include:
parameters), unlike brute force with temperature, resulting in small
optimization methods. The ex- l Dry, frost-free probing environ- spacing changes between devices.
tracted ECPs can be fit to a linear ment with fast purge time. Cascade's Summit-series of semi-
temperature dependence: l Dark, EMI-isolated enclosure that automatic thermal probe stations
reduces measurement uncertainty. include control software that auto-
ECP(Vg, Vds, T) = ECP l Fast, accurate, stable wafer heating matically compensates for these
(Vg, Vds, To) [1 + B(V g, Vds) and cooling. changes. This minimizes the impact
(T -To)] l Limited electrical drift of calibra- on measurement accuracy.
tion standards, probes, and cables
Where To = 300K and B is the over temperature.
Compensation capability for wafer
Calibration Methods and Stability
temperature coefficient for a given l
ECP, usually expressed in parts per expansion (TCE). You can use the LRM, LRRM,
thousand. SOLT, and TRL calibration meth-
Cascade Microtech's patented
ods for over-temperature measure-
MicroChamberTM provides all of
ments. Although often ignored, the
these features.
implementation of custom calibra-
tion software greatly affects calibra-
Calibrating and Measuring tion accuracy. For example, you
must consider variations in the
Calibration stability and measure-
dielectric constant of the substrate
ment accuracy over temperature are
material, via hole integrity, or
the most important, and yet the
changes in the load (match) stan-
most often neglected, thermal prob-
CASCADE
TM
dard with temperature. To avoid
ing issues. The initial calibration
; these difficult issues, use Cascade's
calibration software and ISS. LRM
and LRRM are the preferred cali-
bration techniques. The primary
consideration with LRM or LRRM
standards is the sensitivity of the
50-ohm load standard to tem-
perature variations. The thermally
isolated auxiliary stages on the
Summit 10600-series probe station
minimize the temperature variation
of this load, by reducing the tem- 25x
perature excursions of the calibra-
tion elements. A main chuck tem-
perature range of -65o C to +200o
C corresponds to an auxiliary stage 0.500 12.50 FREQ-GHZ 26.50 '
temperature range of about +5" C
to +65" C. This smaller temperature Figure 2. Calibration stability at 200o C for a 25o C calibration.
excursion reduces the total variation
in the load resistance to about calibration stability is unacceptable. potentially high gain at low
0.7%. LRM standards fabricated Note that the probe continues to temperatures, some devices may
on-wafer have greater variations, change during the first 15 minutes. oscillate. Take precautions to
perhaps as large as 3%. For optimal results, therefore, allow ensure stability.
Changes in the probes and cables a 15-minute probe stabilization
greatly impact calibration and mea- period at each temperature.
surement accuracy. With a +200o C General Extraction Results
chuck temperature, the associated Considerable activity is taking
coaxial connector temperature is Measurement Issues
place in the area of temperature-
about +75" C. Since the probes Along with probe temperature, dependent device modeling.
directly contact the substrate, the there are measurement considera- Although most work focuses on
coax connector and coax-to-copla- tions. These include RF power GaAs MESFET technology,
nar transition experience significant levels, resistive loss in the bias PHEMT and HBT modeling
thermal changes. Experiments show lines, and device stability. At low efforts continue to increase. Figure
that these changes are large enough temperatures, the device gain 3 shows typical ECP temperature
to require new calibrations at each increases considerably. To ensure coefficients for a 0.5 x 200 pm
temperature. linear operation over the entire MESFET device.
Figure 2 illustrates the changes temperature range, adjust the RF
observed in calibration stability for power levels to the device.
a room temperature calibration after Finally, you must consider
Summary
the chuck changes to +200 o C. The device stability. Due to the Points to remember are:
l Room temperature calibrations are
not valid for over-temperature
measurements.
l Stabilize probes and cables for at
least 15 minutes before perform-
ing a calibration.
l Check calibration stability fre-
quently.
l Implementation of custom calibra-
tion standards requires careful con-
sideration.
Ids gm Cgs Cgd Rds tau Ri Rs Rd Cb ft
Figure 3. Temperature coefficients for 0.5 x 150 mm MESFET.
Cascade Microtech, Inc., 14255 SW Brigadoon Ct., Beaverton, Oregon 97005, USA
Tel: (503) 626-8245 Fax: (503) 626-6023 Telex: WU62072315
Japan: 03-3320-6410; UK: 0295-721216; France: 13-956-81-31; Germany: 89-460-2071
:
Copyright 0 1994 Cascade Microtech, Inc.
CASCADE
TM
Cascade Microtech and MicroChamber are trademarks of Cascade Microtech, Inc.
m MicroChamber manufactured under U.S. Patent No. 5,266,899
Advanced Microelectronic Probing Solutions
TECHBRIEF7-0694