semi合集-English.pdf - 第7589页

SEMI MF2166-0304 © SEMI 2004 2 2.2 Three m ethods are describ ed to cover various experim ental arrangem ents (see Related Inform ation 1 for a discussi on on the t wo different types of corona). 2.2.1 Method A — For sys…

100%1 / 7923
SEMI MF2166-0304 © SEMI 2004 1
SEMI MF2166-0304
PRACTICES FOR MONITORING NON-CONTACT DIELECTRIC
CHARACTERIZATION SYSTEMS THROUGH USE OF SPECIAL
REFERENCE WAFERS
This standard was technically approved by the Global Silicon Wafer Committee and is the direct
responsibility of the North American Silicon Wafer Committee. Current edition approved for publication by
the North American Regional Standards Committee on October 16, 2003. Initially available at
www.semi.org February 2004; to be published March 2004. Originally published by ASTM International as
ASTM F 2166-02. Last previous edition SEMI MF2166-02.
1 Purpose
1.1 Non-contact dielectric characterization systems
(NCDCS) that supply surface charge by means of a
corona instead of an electrode are frequently used to
determine the oxide characteristics of MOS device
structures. These instruments can measure a number of
characteristic parameters, including:
1.1.1 initial surface voltage (V
surf
or V
cpd
),
1.1.2 flatband voltage (V
fb
),
1.1.3 effective charge (Q
eff
),
1.1.4 density of interface traps (D
it
),
1.1.5 mobile charge (Q
m
), and
1.1.6 electrical dielectric thickness (T
ox
).
1.2 To determine that the NCDCS is performing
correctly, it is necessary to test the tester. This practice
provides procedures for testing these systems with the
use of special reference wafers with specified
characteristics.
1.3 The procedures in these practices are designed to
ensure that the NCDCS being evaluated provides
reliable data for selected device-related parameters.
1.4 Flatband voltage is an important parameter in the
fabrication of MOS devices. Its value is dependent on
the charge structure within and on the dielectric as well
as at the silicon-dielectric (Si-SiO
2
) interface. These
charges consist of interface and bulk trapped and fixed
charge, mobile charge distributed within the dielectric,
and charge residing on the top of the dielectric.
Although the total charge in and on the oxide
determines flatband voltage, the charge near the silicon
has the greatest effect.
1.4.1 Most device applications require that the mobile
charge be minimized to provide stable threshold
voltage. Stability of the charges at elevated
temperatures is a measure of mobile charge density
within the dielectric. The mobile charge changes
location under temperature and electric field stress.
The mobile charge density can be calculated from the
resulting change in flatband voltage.
1.4.2 Equipment used to produce dielectrics can
become contaminated in various ways. When such
equipment becomes contaminated, it usually introduces
an increased level of mobile ions into the dielectric.
1.4.3 Measuring V
surf
or V
cpd
, the contact potential
between the probe and the semiconductor, with a
Kelvin or Monroe probe allows a quantitative
measurement of the charges associated with the
semiconductor surface and the dielectric, independent
of top (contact) electrode variations as in MOS C-V
measurements. A special reference wafer has the
mobile ions confined to the top half of the dielectric. It
is therefore a good check of the Kelvin or Monroe
probe, which is most sensitive to charges near the top of
the dielectric.
1.5 The procedures in this practice can be used to
ascertain that the dielectric characterization system(s)
meet quality control requirements. They can be used as
a training tool for tester applications.
1.6 The results obtained by these measurements can
also be used to expose and troubleshoot tester problems
including wafer handling and software issues.
2 Scope
2.1 These practices describe the use of wafers with
special electrical and physical characteristics for
controlling and monitoring performance of non-contact
dielectric characterization systems (NCDCS) that
employ corona. They apply in general to dielectrics
with charge structure that can be measured by the
characterization system. Their applicability to
dielectrics of less than 3-nm electrical oxide thickness
depends on the capability of the NCDCS. A thickness
correlation between the reference wafer(s) and thin-
dielectric monitor wafers must be established.
2.1.1 The reference wafers must be designed to check
the repeatability and functionality of all the parts and
subsystems of the characterization system. Thermally
oxidized wafers are not sufficient for this purpose
because of parametric drift due to temperature-bias
stressing and exposure to the atmosphere.
SEMI MF2166-0304 © SEMI 2004 2
2.2 Three methods are described to cover various
experimental arrangements (see Related Information 1
for a discussion on the two different types of corona).
2.2.1 Method A — For systems that employ point
source corona,
2.2.2 Method B — For systems that employ line source
corona and can be programmed to obtain a full set of
parameters, and
2.2.3 Method C — For systems that employ line source
corona but can only determine mobile ion density.
2.3 For each of the three methods, the total testing time
can be reduced by using a short-cycle-time
modification.
2.4 Special reference wafers are used to determine
selected performance characteristics of these dielectric
characterization systems, including initial surface
voltage (V
surf
or V
cpd
), flatband voltage (V
fb
), effective
charge (Q
eff
), density of interface traps (D
it
), mobile
charge (Q
m
), and electrical dielectric thickness (T
ox
).
2.5 The parameters determined relate to the properties
of the NCDCS subsystems in the following way:
2.5.1 V
fb
and Q
eff
data are used to check the surface
photovoltage (SPV) probe, SPV strobe and corona
biasing subsystems, by monitoring behavioral changes
through accumulation and around the flat band
condition.
2.5.2 D
it
data are used to check the same subsystems as
in Section 2.5.1 , by monitoring behavioral changes
through the depletion region of the Q
bias
vs SPV curve.
2.5.3 Q
m
data are used to check the temperature-stress
subsystem, the corona subsystem, and the Kelvin or
Monroe probe.
2.5.4 T
ox
data are used to check the corona stepping
and voltage bias and the Kelvin or Monroe probe air-
gap setting.
NOTICE: This standard does not purport to address
safety issues, if any, associated with its use. It is the
responsibility of the user of this standard to establish
appropriate safety and health practices and determine
the applicability of regulatory or other limitations prior
to use.
3 Limitations
3.1 The presence of light on the wafer surface during
the measurements can affect the results obtained and
should be controlled according to the system
manufacturer’s recommendations.
3.2 Relative humidity in the measurement environment
above 60% may affect measurement values.
3.3 Stray electromagnetic fields external to the
NCDCS may affect measurement results and should be
limited according to the system manufacturer’s
recommendations.
3.4 Wafer handling should be done in accordance with
accepted clean room practices, because surface
contamination can affect the measurement results.
3.5 Change in the characteristics of the special
reference wafers due to handling and other effects may
affect measurement results. The special reference
wafers must be checked at least once every six months,
to assure consistent performance.
4 Referenced Standards
4.1 SEMI Standards
SEMI M20 — Specification for Establishing a Wafer
Coordinate System
SEMI MF1241 — Terminology of Silicon Technology
NOTICE: Unless otherwise indicated, all documents
cited shall be the latest published versions.
5 Terminology
5.1 General terms related to testing of semiconductor
materials may be found in SEMI MF1241.
5.2 Definitions
5.2.1 accumulation condition — the region of the
capacitance-voltage (C-V) curve for which a 5-V
increment toward a more negative voltage for p-type
material, or toward a more positive voltage for n-type
material, results in a change of less than 1% in the
maximum capacitance, C
max
.
5.2.2 bias temperature stress, (BTS) — a voltage
applied to the top surface of a dielectric on a wafer
whose temperature is maintained at a defined level for a
specified time.
5.2.3 contact potential difference, (V
cpd
) — the
potential measured by a Kelvin or Monroe probe that
results from the potential difference between two
metals or between a semiconductor and a metal, due to
their difference in work function. Also called surface
voltage (V
surf
); see Section 5.2.16
.
5.2.4 corona temperature stress, (CTS) — a BTS in
which bias voltage is supplied by a corona discharge of
known charge and voltage applied to the top surface of
a dielectric on a wafer whose temperature is maintained
at a defined level for a specified time.
5.2.5 density of interface traps, (D
it
) — the calculated
interface charge per unit area that accounts for the
difference in gradient between the measured and
SEMI MF2166-0304 © SEMI 2004 3
theoretical Q
bias
vs. SPV curve or equivalent “Q-Q”
plot, in the band gap region.
5.2.6 depletion condition — the condition that exists in
the semiconductor surface region when the free carrier
density is insufficient to neutralize the fixed dopant
charge density of donors and acceptors.
5.2.7 dielectric (oxide) electrical thickness, (T
ox
) — the
silicon dioxide equivalent thickness (when the silicon
dioxide dielectric constant is assumed) calculated from
electrical measurement of Q
bias
vs V
surf
(or V
cpd
) for the
capacitor.
5.2.8 effective charge, (Q
eff
) — the oxide charge/cm
2
in
the capacitor that gives the measured flatband voltage.
5.2.9 flatband condition — the point at which an
external applied voltage causes there to be no internal
potential difference across the MOS structure.
5.2.9.1 Discussion — In the flatband condition, the
SPV is zero.
5.2.10 flatband voltage, (V
fb
) —- the applied voltage
necessary to produce the flatband condition.
5.2.11 inversion condition — the condition in which a
minority carrier layer is formed at the semiconductor
surface separated from the bulk by a depletion region.
5.2.12 Kelvin probe — a reference electrode that
vibrates perpendicular to the wafer surface and
generates an AC signal by varying the electrode-wafer
capacitance.
5.2.13 mobile charge density, (Q
m
) — the calculated
charge/cm
2
that moves in the oxide under temperature
and electrical field stress, creating shifts in SPV, V
fb
,
and V
cpd
(or V
surf
).
5.2.14 Monroe probe — a reference electrode that is
stationary and generates an AC signal by the horizontal
vibration of a grounded fork which shields the probe
from the wafer.
5.2.15 surface photovoltage, (SPV) — the change of
the electrostatic potential of the silicon surface caused
by illumination.
5.2.16 surface voltage, (V
surf
) — the potential measured
by the Kelvin or Monroe probe that results from the
potential difference between two metals or between a
semiconductor and a metal, due to their difference in
work function. Also called contact potential difference
(V
cpd
); see Section 5.2.3 .
6 Summary of Practices
6.1 Method A
6.1.1 A special reference wafer is loaded into the
NCDCS.
6.1.2 A series of tests, which must include CTS, is
executed to position mobile ions into a consistent pre-
test condition. All CTS times must be at least 4
minutes to allow full movement of the mobile ions.
6.1.3 Three additional identical test sequences are
executed to collect data to evaluate performance of the
instrument.
NOTE 1: The bias, temperature and time used for the last
CTS prior to carrying out the above tests should be employed
for these test sequences.
6.1.4 Performance of the NCDCS and its selected
components is checked by measuring the following
characteristics: initial surface voltage (V
surf
), flatband
voltage (V
fb
), effective charge (Q
eff
), density of interface
traps (D
it
), mobile charge (Q
m
), and electrical dielectric
thickness (T
ox
).
6.1.5 Measurements are made at three points on the
surface of the wafer and the mean, standard deviation
and relative standard deviation of the results obtained in
the three identical sequences (Section 6.1.3 ) are
computed and recorded.
6.1.6 The results are compared with the values
supplied with the special reference wafer to evaluate the
NCDCS performance.
6.1.7 If it is desired to check the system over its full
range of operation, this procedure is repeated using two
or more additional special reference wafers with
parameters that cover the typical parameter ranges in
Section 7.4 .
6.1.8 Short-cycle Time Modification of Method A —
The short-cycle time modification is performed in the
same manner except that the measurement of Q
m
is
carried out first, in order to assure consistent alignment
of the mobile ions prior to the other measurements.
The Q
m
measurement is followed by the measurement
of the other parameters. No repetitions are required; the
test sequence in this modification is only run once.
6.2 Method B
6.2.1 First any remaining residual charge on a special
reference wafer is neutralized by immersing it in DI
water.
6.2.2 Then the special reference wafer is loaded into
the NSCDS and the surface voltage (contact potential
difference) is zeroed to within ± 0.3 V.
6.2.3 Then the steps in Sections 6.1.2 –6.1.4 and 6.1.6
are followed.
6.2.4 If it is desired to check the system over its full
range of operation, this procedure is repeated using two
or more additional special reference wafers with