DER Materials Quaterly Progress Report Jan 1 Mar 31, 2002 HR230 19513

User Manual: HR230

Open the PDF directly: View PDF .
Page Count: 115

Download
Open PDF In Browser	View PDF

DER MATERIALS QUARTERLY
PROGRESS REPORT
For the Period
January 1, 2002 to March 31, 2002
Prepared by:
David P. Stinton, Manager, and
Roxanne A. Raschke
DER Materials Research
Oak Ridge National Laboratory
For:
Department of Energy
Office of Power Technologies
Energy Efficiency and Renewable Energy

DER MATERIALS QUARTERLY PROGRESS REPORT
January—March 2002

TABLE OF CONTENTS
Introduction
RECUPERATORS
Advanced Alloys for High Temperature Recuperators
P. J. Maziasz, B. A. Pint, R. W. Swindeman, K. L. More, and M. L. Santella
Oak Ridge National Laboratory, Oak Ridge, Tenneessee
Recuperator Alloys – Composition Optimization for Corrosion Resistance
R. Peraldi and B. A. Pint
Oak Ridge National Laboratory, Oak Ridge, Tennessee
Recuperator Materials Testing and Evaluation
E. Lara-Curzio
Oak Ridge National Laboratory, Oak Ridge, Tennessee
CERAMIC RELIABILITY FOR MICROTURBINE HOT-SECTION COMPONENTS
Reliability Evaluation of Microturbine Components
H-T Lin, M. K. Ferber, and P. F. Becher
Oak Ridge National Laboratory, Oak Ridge, Tennessee
Long-Term Testing in Water Vapor Environments
M. K. Ferber and H-T Lin
Oak Ridge National Laboratory, Oak Ridge, Tennessee
Reliability Analysis of Microturbine Components
S. F. Duffy, E. H. Baker and J. L. Palko
Connecticut Reserve Technologies, LLC
NDE Technology Development for Microturbines
W. A. Ellingson, E. R. Koehl, A. Parikh, and J. Stainbrook
Argonne National Laboratory, Argonne, Illinois
CHARACTERIZATION OF ADVANCED CERAMICS FOR INDUSTRIAL GAS
TURBINE/MICROTURBINE APPLICATIONS
Oxidation/Corrosion Characterization of Monolithic Si3 N4 and EBCs
K. L. More and P. F. Tortorelli
Oak Ridge National Laboratory

Mechanical Characterization of Monolithic Silicon Nitride Si3 N4
R. R. Wills, S. Hilton, and S. Goodrich
University of Dayton Research Institute, Dayton, Ohio

Microstructural Characterization of CFCCs and Protective Coatings
K. L. More
Oak Ridge National Laboratory, Oak Ridge, Tennessee
High-Temperature Environmental Effects on Ceramic Materials
P. F. Tortorelli, K. L. More, and L. R. Walker
Oak Ridge National Laboratory, Oak Ridge, Tennessee
High Speed Burner Rig Development
B. Schenk and G. Schroering
Honeywell Engines, Systems & Services
DEVELOPMENT OF HIGH-TEMPERATURE COATINGS
Environmental Protection Systems for Ceramics in Microturbines and Industrial Gas Turbine
Applications, Part A: Conversion Coatings
S. D. Nunn and R. A. Lowden
Oak Ridge National Laboratory, Oak Ridge, Tennessee
Environmental Protection Systems for Ceramics in Microturbines and Industrial Gas Turbine
Applications, Part B: Slurry Coatings and Surface Alloying
B. L. Armstrong, K. M. Cooley, M. P. Brady, H-T Lin, and J. A. Haynes
Oak Ridge National Laboratory, Oak Ridge, Tennessee
High-Temperature Diffusion Barriers for Ni-Base Superalloys
B. A. Pint, J. A. Haynes, K. L. More, and I. G. Wright
Oak Ridge National Laboratory, Oak Ridge, Tennessee
POWER ELECTRONICS
Development of High-Efficiency Carbon Foam Heat Sinks for Microturbine Power Electronics
J. W. Klett, A. D. McMillan, N. C. Gallego
Oak Ridge National Laboratory, Oak Ridge, Tennessee
MATERIALS FOR ADVANCED RECIPROCATING ENGINES
Development of an Ultra Lean Burn Natural Gas Engine
R. M. Wagner, T. J. Theiss, J. H. Whealton, J. M. Storey, and J. B. Andriulli
Oak Ridge National Laboratory, Oak Ridge, Tennessee

Advanced Materials for Exhaust Components of Recripracating Engines
P. J. Maziasz
Oak Ridge National Laboratory, Oak Ridge, Tennessee

RECUPERATORS

Advanced Alloys for High Temperature Recuperators
P. J. Maziasz, B. A. Pint, R. W. Swindeman, K. L. Moore, and M. L. Santella
Metals and Ceramics Division
Oak Ridge National Laboratory
P.O. Box 2008, Oak Ridge, TN 37831-6115
Phone: (865) 574-5082, E-mail: maziaszpj@ornl.gov

Objective
The main objective of this program is to work with commercial materials suppliers (foil and thin
sheet) and recuperator manufacturers to develop the improved or advanced materials they need
for their near-term and longer-term microturbine applications. The near term portion of this
program will identify or develop low-cost stainless steels and alloys (no more than 1.5 times the
cost of commercial 347 stainless steel) with more creep- and corrosion-resistance for better
performance and durability at about 704oC (1300oF). The longer-term portion of this program
will identify or develop stainless alloys or materials that can perform reliably at 760oC (1400oF)
and higher. Near-term success should have an immediate impact on current microturbines,
particularly those using alternate or more corrosive fuels. Longer-term solutions will benefit
advanced microturbines which are being designed today to operate at higher temperatures.
Currently, the specific compact recuperator technologies being addressed in this program
included primary surface recuperators (PSR) and the brazed plate and fin recuperators (PFR).
Other types of advanced recuperator technology can be included in the future. ORNL will
collaborate with materials suppliers to test the properties (tensile, creep, corrosion) of candidate
foils related to component manufacturability and in-service performance, while the OEM’s
obtain sufficient and appropriate material to test actual recuperator manufacturing processes,
and make recuperator components for evaluation and testing. There is synergy and
coordination between this particular materials program and other ORNL programs that are: a)
evaluating the corrosion resistance of recuperator foils at 650-800oC in 10% water vapor, and
b) developing an engine-based advanced recuperator materials and component test facility.
Highlights
Materials for use to about 704oC (1300oF)
The most promising materials for this goal range from standard 347 stainless steel with modified
processing for more creep-resistance, to compositionally modified 347 stainless steels with
modified processing with significantly better creep-resistance and aging resistance. This group
could also include stainless steels with more Cr and Ni than standard 347 steel. This quarter,
direct evaluation of fresh and engine-tested recuperator components began. Such data will
enable ORNL to define the optimum performance upgrade solution for each microturbine
manufacturer. This quarter, another ORNL modified 347 steel (#4) was tested and lasted about

575 h during creep at 750oC and 100 MPa, with about the same creep rate found in alloy 625.
New oxidation/corrosion data on these ORNL modified 347 steels at 800oC and 10% water
vapor showed significantly better behavior than standard 347 stainless steel up to 1000 h. New
data on weldability showed the new ORNL modified 347 steels to be as weldable as standard,
commercial 347 steel.
Materials for use at 760oC (1400oF) or higher
The most promising materials with significantly better creep-resistance and corrosion-resistance
in this temperature range at 3-4 times the cost of 347 stainless steel are HR120 (Haynes) or
modified alloy 803 (Special Metals). These Fe-based austenitic stainless alloys have about
25%Cr and 35%Ni. Alloys with even higher performance would include HR230, alloy 740, or
alloy 625, or HR214 at 5-10 times the cost of 347 steel, and these are mainly Ni-based alloys
with significant amounts of Cr, Co, Mo, W or Al added. This quarter, creep-rupture screening
at 750oC and 100 MPa was completed, with variations in processing to vary grain size all
showing relatively similar good creep behavior. Fine-grained HR120 lasted almost 900 h and
had about 30% rupture ductility, but also a higher creep rate. Discussions about producing
commercial foils of these alloys began this quarter. Weld screening of HR120 and the modified
803 alloy relative to standard, commercial 347 steel also began this quarter. Oxidation
screening of both HR120 and the modified 803 alloy showed excellent oxidation resistance after
1000 h at 800oC.
Technical Progress
Current natural gas microturbines in service today have recuperators made of 347 stainless steel
and generally do not exceed 650-675oC. While higher efficiency, larger turbines sizes and
alternate fuels all require more performance as well as higher reliability from the recuperator,
economic factors limit more costly solutions for recuperators, which are already about 30% of
the total system first cost. This recuperator materials program is driven by the need to provide
microturbine OEMs with the most affordable materials that upgrade the performance of their
specific recuperator technologies. This ORNL program has two essential parts: (a) analysis and
evaluation of engine-tested recuperator components to exactly define the performance and
manufacturing details related to improved recuperators for each OEM, and (b) commercialscale trials of processing and alloy compositional modifications of type 347 stainless steel and of
alloy 120, and possibly of modified alloy 803, with benchmarking to establish the performance
relative to standard, commercial 347 stainless steel foils and sheets for comparison. There is
close coordination with another ORNL program that has converted a Capstone 60 kW
microturbine into an advanced recuperator test facility.
Recuperator Component Analysis
Several different microturbine OEMs have provided pieces of fresh and engine-tested PFR and
PSR recuperators made from standard 347 stainless steel for analysis and testing last quarter,
and sectioning and characterization began this quarter. These components are manufactured

from coils of standard, commercial 347 steel that range from 3-4 mil foil up to 10 mil sheet, and
include the additional manufacturing steps of welding and/or brazing. ORNL is characterizing the
engine-tested components and measuring the changes relative to fresh (as-manufactured)
components. In addition, samples of some fresh recuperator components began lab testing in
air + 10% water vapor at several different temperatures to help understand and interpret the
engine-tested component behavior. This engine-exposed 347 stainless component data will
define the performance and/or life increases that can be achieved by substituting the advanced
materials. Detailed component analysis is also important feedback to the OEM’s, and will be
used to help define the best way to use advanced materials to make recuperators capable of
better performance at higher temperatures for each different microturbine OEM.
Selection and Commercial Scale-Up of Advanced Recuperator Materials:
a) Materials for use to about 704oC (1300oF)
A summary of the room-temperature tensile properties of 4 mil foils of the advanced alloys
down-selected from the previous ORNL advanced alloy screening program are given in Table
1, including several more modified 347 steels tested this quarter (ORNL modified 347 #3 and
#4), together with cost relative to standard, commercial 347 stainless steel. Foils of the
standard stainless steels and alloys were produced by lab-scale processing methods at ORNL
from commercial plate or tubing stock. The ORNL developmental modified 347 stainless steels
were melted as 10-15 lb ingots, and then hot-rolled into plate and processed into foil at ORNL.
Creep-rupture properties for these foils testing in air at 750oC and 100 MPa (high-stress creep
for accelerated testing) are given in Fig. 1. Last quarter, a new 20Cr-15Ni austenitic stainless
steel (ORNL mod 3) was produced and tested, and this quarter another modified 347 (19Cr12.5Ni, ORNL mod 4) was also produced and tested. The new ORNL 347 mod 4 showed
exceptional creep-strength (as good as or better than alloy 625 in terms of primary creep and
secondary creep rate), and lasted about 575 h and ruptured with almost 3% elongation. This
quarter, several more very similar modified heats of 347 were made with small adjustments in
Nb and C content and other minor impurities to boost the rupture ductility without reducing the
creep strength, and they will be creep-tested next quarter.
Table 1. Tensile Properties of Advanced Heat-Resistant Austenitic Stainless Steel and Alloy Foils Tested at
Room Temperature
Alloy

Yield Strength (MPa)

Materials for use to about 704oC (1300oF)
347 HFG steel
283
347 standard steel
173
ORNL mod. 347 #1
281
ORNL mod. 347 #2
232
ORNL mod. 347 #3
259
ORNL mod. 347 #4
285

Materials for use at 760oC (1400oF) or higher

Total Elongation (%)

50
47
33
23
36
29

Approximate
Normalized Cost
1.0
1.0
1.0
1.0
1.0
1.0

alloy 803 (standard)
modified alloy 803 (1A)
alloy 120 (standard)

242
294
415

39
30
34

3
3
3.5

(0.004 inch thick foils processed at ORNL at different conditions tailored to each particular steel or alloy)

12
creep at 750°C 100 MPa

347HFG SS
(foil from tubing)

CREEP STRRAIN (%)

standard
347 SS foil

ORNL mod 3

ORNL mod 4

2
ORNL mod 2

0
0

100

200

300

400

500

600

TIME (h)

Figure 1. Plot of creep-strain versus time for three new ORNL modified laboratory heats of
type 347 stainless steel (17-18Cr, 10-13Ni, ORNL mod 2 and 4, and 20Cr-15Ni, ORNL
mod 3) tested in air at 750oC. For comparison, foil from standard commercial 347 stainless
steel, and from foil produced by splitting, flattening and rolling commercial 347H tubing
(Sumitomo, H – high carbon, FG – fine grained), both with similar ORNL lab-scale foil
processing, are also included.
This quarter, ORNL began oxidation studies of these new modified 347 steels together with
other advanced alloy, including the 120 and 803 austenitic alloys. Tests to 1000 h were
completed at 750oC and 800oC with 10% water vapor this quarter, and new testing began at
650 and 700oC. Specimens were also made from portions of fresh 347 stainless steel
recuperator components and put into the oxidation test at 700oC. Results of the testing at
800oC are shown in Fig. 2. These are greatly accelerated tests for the purpose of screening the
alloys and ranking them relative to each other, not tests to qualifying alloys for use at these
conditions. Clearly, the standard commercial 347 stainless steels suffer rapid and catastrophic

corrosion attack, but the new ORNL modified 347 steels all show much better corrosion
resistance, and fall closer to the behavior of NF709 (Fe-20Cr-25Ni-Nb) and alloy 625 for this
limited testing time. Longer time testing at 700 and 750oC will be more relevant to extended
microturbine service at about 700oC, but clearly, the ORNL modified 347 steels that show
improved creep resistance also show improved corrosion resistance relative to standard
commercial 347 stainless steel. Such behavior is consistent with at least higher levels of Cr and
Ni in the modified 347 steels, as well as other new elements added to stabilize the austenite
phase at high temperatures.
2.0
Std. 347 (A-L)
347HFG

1.8

800°C
air + 10vol.%H 2O

1.6

347 Mod - 2

1.4

347 Mod - 4
347 Mod - 3

1.2
1.0
0.8
0.6
0.4

NF709
Alloy 625

0.2
0.0

Haynes 120

-0.2

803 mod A

-0.4
0

200

400

600

800

1000

Time (h) in 100h cycles at 800°C
Figure 2. Oxidation testing of foil coupons in air + 10% water vapor at 800oC, with cycling to
room temperature every 100h for weight measurements. Foils of both commercial stainless
steels, alloys and superalloys and ORNL developmental steels were lab-scale processed at
ORNL and were also used to make tensile/creep specimens. All foils were made from plate
stock, except for 347 HFG and NF709, which were made from split and flattened boiler
tubing. All foils are in the solution-annealed condition.
This quarter, new testing began on both the weldability and the thermal aging resistance of all the
candidate advanced recuperator stainless steels and alloys. Weldability was investigated
because welding is a critical manufacturing step in making any recuperator component, and

there is the general perception that more stable austenitic stainless steels and alloys (Cr and Ni
equivalents of alloying elements, with Cr and elements that behave like it promoting ferrite
stability, and Ni and elements that behave like it promoting austenite stability) are more difficult
to weld. Weldability was tested by using gas-tungsten-arc (argon shielded) welding to produce
autogeneous, partial penetration welds on all of the thicker plate stock (0.1 to 0.03 inches thick)
of the commercial, or ORNL developmental steels and alloys used to produce foils. This is a
much more severe test than welding foils. The data so far indicates that all of the ORNL
modified 347 steels are as weldable as standard, commercial 347 steel (Allegheny-Ludlum),
with no hot or cold-cracking.
Cross-sections of these welds will be examined
metallographically next quarter, as well as some testing of other welding methods (i.e. laser
welding).
This quarter, thermal aging began of foil tensile specimens at 700 and 750oC for 2500 h to
measure the room-temperature ductility remaining after aging. 300 series stainless steels are
prone to significant ductility loss after aging at 650oC and above, and cold cracking is a concern
for recuperator components that must endure cycling or thermal shocks due to rapid heating or
cooling. This testing will be another measure of the improvements of the modified 347 stainless
steels or the advanced austenitic alloys in terms of retaining ductility by resisting the effects of
aging.
Discussions with microturbine OEMs and 347 stainless steel foil producers continued this
quarter to define their interest in selecting one or two of the ORNL modified 347 steels for
commercial scale-up and foil production, as well as modifying the processing of standard 347
stainless steels to improve its performance.
b) Materials for use at 760oC (1400oF) or higher
The advanced austenitic alloys with more Cr and Ni than 347 stainless steel represent a
substantial upgrade in high temperature strength and creep-resistance relative to standard 347
stainless steels, particularly with appropriately modified foil processing. Standard alloy 803
(Special Metals, Fe-25Cr-35Ni-Ti-Nb) has very good creep-rupture ductility but much less
creep-resistance than standard 347 stainless steel, as shown in Fig. 3. The modified 803 alloys
were developed in a joint project between ORNL and Special Metals to improve its creep
resistance, and several different processing conditions (1A and 1A.1, varying primarily grain
size) still produce consistent improvements in creep-strength and rupture ductility relative to
standard 347 stainless steel. Better fracture ductility and better creep strength will make this
alloy have better performance and reliability for recuperator applications. Standard HR120
(Haynes International, Fe-25Cr-35Ni-Nb-N) alloy processed into foil at ORNL (A) has
similarly good creep behavior. A second sample of HR120 alloy was processed differently (B,
much shorter annealing times to refine the grain size, similar to the modified 803 alloy), and it
lasted about 900 h and ruptured with about 30% ductility after creep testing at 750oC and 100
MPa. While all of these alloys show good rupture lives and ductility, their creep rates are
considerably higher than the ORNL modified 347 stainless steels (alloy 120 and modified 803

reach 2-3% creep strain in about 200 h), which exhibit creep resistance similar to the Ni-based
superalloy 625. Total creep-strain is also a consideration and constraint in most recuperator
designs.

50
foil creep
750°C 100 MPa

standard 803

CREEP STRAIN (%)

standard 120 (B)

30
modified 803
(1A)
(1A.1)

standard 347SS
10

standard 120 (A)

0
0

200

400

600

800

1000

TIME (h)

Figure 3. Plot of creep strain (%) versus time for creep-rupture tests at 750oC and 100 MPa
run in air for the commercial advanced stainless alloys processed into 0.004 inch thick foils using
laboratory-scale processing methods at ORNL, and down-selected for commercial scale-up.
All alloys are in the solution-annealed (SA) condition.
The advanced alloys were also evaluated for oxidation/corrosion resistance at 750 and 800oC,
and data at 800oC is also shown in Fig. 2. Both HR120 and modified 803 show good
resistance to oxidation mass gain, consistent with higher Cr contents. Their high-temperature
oxidation/corrosion resistance, even for more corrosive alternate fuels, should be far superior to
347 stainless steel. They also should have similar manufacturing characteristics relative to 347
steel, particularly the modified 803 (Table 1). Their corrosion resistance and creep strength
indicate capability of being used at 760oC or above relative to standard 347 stainless steel
without having to increase section thickness to compensate for increased temperature.

This quarter, tensile specimens of the advanced HR120 and modified 803 were also included in
the thermal aging experiments at 700 and 750oC for 2500 h to determine their resistance to
ductility loss at room temperature. Plate stock of both alloys was also included in the welding
tests, and results will be evaluated next quarter.
In summary, ORNL began a new advanced recuperator materials technology program in FY
2002 to complete lab-scale studies and then to scale them up to commercially produced foil and
sheet products that microturbine OEMs can then use to make upgraded recuperator
components with more performance and reliability at higher temperatures. The goal is
manufacturing on a trial basis and engine-testing. This project expands and applies its unique
and systematic data base on mechanical behavior and oxidation/corrosion effects of stainless
steel and alloy foils, focused on addressing and meeting the needs of the U.S. microturbines
industry for improved recuperators.
Status of Milestones
New Program
FY2002 – Develop stainless steel near type 347 stainless steel composition and cost with
maximum performance in foil form at 700-750oC (complete by June 2002) – on schedule.
Industry Interactions
Discussions continued with the microturbine OEMs Ingersoll-Rand Energy Systems and
Capstone Turbines during this quarter. Ingersoll-Rand shipped engine-tested and fresh
recuperator components to ORNL for analysis and evaluation, and Capstone agreed to provide
some standard, commercial 347 stainless steel foils for baseline properties testing. Discussion
about advanced alloy commercial foil production continued with Elgiloy, Haynes International
and Special Metals. Elgiloy does produce standard HR120 foil, and provided some to ORNL
this quarter. Since the modified 803 alloy is developmental and has only been produced in small
quantities, discussions continued this quarter to obtain a larger heat of the new modified alloy
803 from Special Metals for Elgiloy to process into foil. Discussions with General Motors
Global Alternative Propulsion Center, who have needs for heat-exchanger/reformer technology
for fuel-cell applications that are similar to advanced microturbine recuperators, continued this
quarter as well. More detailed interactions and discussions began this quarter with United
Technology Research Center to define and describe their microturbine recuperator interests and
needs.
Problems Encountered
None.
Publications/Presentations

A presentation on Advanced Alloys for High Temperature Recuperators was made by
P. J. Maziasz, B. A. Pint, R. W. Swindeman, K. L. More and E Lara-Curzio at the DOE/DER
Microturbines and Industrial Gas Turbines Peer Review, held March 12-14, 2002 in Fairfax,
VA.

Recuperator Alloys – Composition Optimization for Corrosion Resistance
R. Peraldi and B. A. Pint
Metals and Ceramics Division
Oak Ridge National Laboratory
P.O. Box 2008, Oak Ridge, TN 37831-6156
Phone:(865) 576-2897, E-mail: pintba@ornl.gov
Objective
In order to provide a clear, fundamental understanding of alloy composition effects on corrosion
resistance of stainless steel components used in recuperators, the oxidation behavior of model alloys is
being studied. The first goal of this study was to determine the critical Cr content and the effect of Ni
concentration required to minimize the accelerated corrosion attack caused by water vapor. Other
factors to be investigated include the effects of temperature, alloy grain size, phase composition and
minor alloy additions. The composition and microstructure effects also will provide data for lifeprediction models and may suggest a mechanistic explanation for the effect of water vapor on the
oxidation of steels. This information will be used to select cost-effective alloys for higher temperature
recuperators.
Highlights
Comparing the performance of austenitic and ferritic alloys, the fine-grained austenitic alloys generally
performed better at lower temperatures, 650-700°C, while the coarser-grained ferritic alloys performed
better at 800°C. Using Cr diffusivity data in ferritic and austenitic alloys from the literature, the
superior performance of ferritic alloys at 800°C can be explained because of the faster Cr diffusivity in
ferritic alloys, compared to austenitic alloys. However, the Cr diffusivities are similar at the lower
temperatures. This suggests that the addition of Ni plays an important role in improving the corrosion
resistance at 650-700°C.
Technical Progress
Experimental Procedure
As outlined in previous reports, several model austenitic and ferritic cast alloys were vacuum annealed
for 4h at 1100°C prior to testing. Rolled alloys were prepared from these cast materials. Pieces of cast
material were hot forged and hot rolled at 1100°C to 2.5mm thickness, followed by cold rolling to
1.25mm. Annealing was done for rolled materials under Ar + 4%H2 for 2 minutes at 900 and 1000°C
for ferritic and austenitic alloys, respectively. Oxidation specimens were cut from the rolled plate
materials with dimensions (1x12x18mm) to produce a similar surface/volume ratio (about 2.3) as the
cast coupons. All specimens were polished to 600-grit SiC paper and cleaned in acetone and methanol.
From light microscopy of etched specimens (Figure 1), characterization of the average grain size was
calculated as ≈150µm for cast alloys, and ≈100µm and ≈10µm for rolled ferritic and austenitic
specimens, respectively. All of the cast and rolled alloys were oxidized in air + 10vol.% water vapor
with 1h cycles between 650 and 800°C.

Austenitic: Fe-18Cr-30Ni

Ferritic: Fe-16Cr

Cast

Rolled
100µm
Figure 1: Light microscopy of etched austenitic Fe-18Cr-30Ni (left column) and ferritic Fe-16Cr (right
column) for cast and rolled samples after annealing.
Volume and grain boundary Cr diffusivity.
Comparison of Cr diffusivity in austenitic and ferritic alloys was made from diffusion data measured
by Tokei et al. [1]. In Figure 2, effective diffusion coefficients (dashed lines), taking into account both
volume diffusion and grain-boundary diffusion with an average grain size of 150µm corresponding to
the cast alloys, are shown and compared to the volume diffusion coefficient (solid lines). In the
temperature range of 597-805°C, the average grain size of the cast alloys was too large to strongly
enhance the Cr diffusivity in both ferritic and austenitic alloys, Figure 2.
Austenitic vs. ferritic.
Figure 2 also includes the different oxidation performances observed in air + 10% water vapor: either a
thin protective oxide scale (POS) was observed or accelerated attack (AA) occurred which sometimes
included spallation. Increasing the temperature delayed the breakaway observed on ferritic alloys
whereas it reduced the protective oxide growth stage for austenitic alloys. In Figure 2, white points
represent the good oxidation performance as they can maintain a protective oxide scale after 100x1h
cycles and black points represent accelerated attack. At 800°C, the Cr diffusivity was calculated to be
more than one order of magnitude higher in the ferritic than in austenitic alloys (Figure 2) providing a
plausible explanation for the better corrosion resistance of ferritic alloys at higher temperatures.
Between 600 and 650°C, the calculated Cr diffusion was similar for ferritic and austenitic alloys.
Without an advantage in Cr diffusivity, the superior performance of the austenitic alloys at lower
temperatures, suggests some beneficial role of Ni in improving resistance to water vapor.
Cast vs. rolled.
If ferritic alloys were more oxidation resistant than austenitic alloys because of their higher Cr
diffusivity, then, in the same way, a fine alloy grain size would increase the Cr diffusivity and
therefore the corrosion resistance of the alloy. Figure 3 is based on a similar calculation as was used
for Figure 2, but for rolled materials, taking into account a grain size of 100µm for austenitic and
10µm for ferritic alloys. Previous experimental results have shown that rolled austenitic alloys delayed
the AA or AAS compared to cast specimens. The higher Cr diffusivity in fine-grained austenitic alloys
explains these results, Figure 3. For ferritic alloys, rolled and cast specimens had a similar behavior as
both alloys have a similar grain size and thus a similar Cr diffusivity.

-1

8.5

10000/T (K )
10
10.5

9.5

11.5

1E-14

: POS after 100x1h cycles
: AA before 100x1h cycles

1E-15

CAST
FER.
Dvol (solid)
Deff (dashed) g=150µm
for Fer. Fe-12Cr

CAST
AUS.

Dvol (solid)
Deff (dashed) g=150µm
for Fer. Fe-9Cr

-1

Dvol and Deff (m .s )

1E-16

1E-17

1E-18

Dvol (solid)
Deff (dashed) g=150µm
for Aus. Fe-18Cr-8Ni
Dvol (solid)
Deff (dashed) g=150µm
for Aus. Fe-20Cr-32Ni

1E-19

1E-20
850

800

750
Temperature (°C)700

650

600

Figure 2: Arrhenius plot of Cr coefficient diffusion in cast austenitic and ferritic alloys calculated from
Tökei et al. data [1] for bulk diffusion (solid lines) and bulk + grain boundary diffusion with an
average grain size of g = 150 µm (dashed lines). Upper and lower areas correspond to the Cr
diffusivity in cast ferritic and austenitic alloys, respectively.

-1

8.5

9.5

10000/T (K )
10
10.5

11.5

1E-14

: POS after 100x1h cycles
: AA before 100x1h cycles

CAST+ROLLED
FER.

1E-15

ROLLED
AUS.
CAST
AUS.

-1

Deff (m .s )

1E-16

Results given
for austenitic
alloys only

1E-17

1E-18

1E-19

1E-20

850

800

750
700
650
Temperature (°C)

600

Figure 3: Arrhenius plot of Cr coefficient diffusion calculated from Tökei et al. data [1] for bulk +
grain boundary diffusion with an average grain size of g = 150 µm for cast ferritic and austenitic (solid
lines), g = 100µm for rolled ferritic (dashed lines) and g = 10µm for rolled austenitic (dashed lines).
For ferritic alloys, Cr diffusivity are superimposed for cast and rolled materials.

Status of Milestones
Open literature publication of experimental results. (August, 2002)
Industry Interactions
Discussed materials needs with Bowman Power Systems and André Mom from the Dutch Gas Turbine
Association
Problems Encountered
None.
Publications/Presentations
Oral presentation at the TMS Annual Meeting, Seattle, USA, (February, 17-21st 2002).
"Effect of chromium and nickel contents on high temperature oxidation of stainless steels in mixed air
and water vapor.", R. Peraldi and B. A. Pint.
Publication submitted to Materials at High Temperatures:
"Effect of chromium and nickel contents on the oxidation behavior of ferritic and austenitic model
alloys in air with water vapor.", R. Peraldi and B. A. Pint.
Reference
[1]: Tökei Zs., K. Hennesen, H. Viefhaus and H. J. Grabke, Materials Science Technology, Vol 16
(2000), pp 1129-1138.

Recuperator Materials Testing and Evaluation
E. Lara-Curzio
Metals and Ceramics Division
Oak Ridge National Laboratory
P.O. Box 2008, Oak Ridge, TN 37831-6069
Phone: (865) 574-1749, E-mail: laracurzioe@ornl.gov

Objective
The objective of this sub-task is to screen and evaluate candidate materials for the next
generation of advanced microturbine recuperators.
To attain this objective, a
commercially-available microturbine was acquired and in coordination and collaboration
with its manufacturer, it was modified to operate at recuperator inlet temperatures as high
as 843°C. The durability of candidate recuperator materials will be determined by
placing test specimens at a location upstream of the recuperator, followed by
determination of the evolution of the material's physical and mechanical properties as a
function of time of exposure. During exposure tests inside the microturbine, it will be
possible to subject test specimens to various levels of mechanical stress by using a
specially-designed sample holder and pressurized air. The selection of materials to be
evaluated in the modified microturbine will be made in coordination and collaboration
with other tasks of this program and with manufacturers of microturbines and
recuperators.
Highlights
A procedure was developed, in collaboration with personnel from the Materials Joining
& NDT Group at ORNL, to laser-weld thin (plain or corrugated) metallic foils to a
sample holder for evaluation inside a microturbine. The sample holder will enable the
application of mechanical stresses to the metallic foil test specimens through internal
pressurization. The critical parameters for laser-welding thin metallic foils were found
by trial and error and the final procedure is capable of producing leak-tight structures.
Technical progress
During this reporting period significant progress was made towards the operational
readiness of the microturbine test facility. Although original plans included the indoor
installation of this test facility, these plans had to be modified as a result of restrictions
placed by various organizations within Oak Ridge National Laboratory. After several
environmental and safety reviews, the test facility is finally in progress of becoming
operational. Figure 1 is a photograph of the modified 60kW Capstone microturbine and
its gas compressor in their current outdoor location behind the High Temperature
Materials Laboratory at ORNL. Gas and air lines have been installed, and electrical
connections are being finalized. Solid-state power electronics in the microturbine will

allow it to operate in grid-connect mode and hence, it will supply electrical power to
ORNL.
Progress has also been made in the preparation of test specimens and the validation of the
sample holder design. In collaboration with Dr. Michael Santella of the Materials Joining
& NDT Group at ORNL, a study was initiated to determine the optimum parameters for
laser-welding foil test specimens to sample holders. To date, flat and stamped foils of
347 stainless steel and different Haynes alloys (120, 214, 230) have been laser-welded to
sample holders fabricated with the same alloy (to avoid stresses that would result from a
possible mismatch in thermal expansion behavior). Figure 2, which is a photograph of a
sample holder, shows flat and corrugated foils at locations 1 and 3 of the sample holder,
respectively. Visible at locations 2 and 4 are holes that allow the placement of
thermocouples next to test specimens to monitor their temperature during the test.
Another role of the holes is to allow the passage of air to pressurize the cavity that is
formed when a foil test specimen is welded to the sample holder. Dimensions for the
sample holder are shown in Figure 3.

Figure 1. Microturbine test facility. On the left is a100 psi gas compressor. On the right
is the modified 60kW Capstone microturbine. Next to the microturbine is an exhaust
noise suppressor.

Figure 2. Photograph of sample holder for evaluation of foil test specimens. Note plain
and corrugated foil test specimens at locations 1 and 3 of the sample holder.

0.75

1.00
1.825

0.60

0.45
0.70

0.25

4.70

0.10
0.70

0.70

Ø 0.20, through hole, 4 plc s.

0.175

sect ion v iew A-A'

Figure 3. Dimensions of sample holder.
When the sample holder is pressurized, the foil test specimens will be subjected to a
multiaxial state of stress that will be dominated by tangential (hoop) stresses. Figure 4
shows the tangential stress distribution obtained by the finite-element method in a section
of the sample holder with a foil test specimen. The results presented in Figure 4
correspond to an axisymmetric elastic analysis with symmetric boundary conditions at
the bottom of the model and unitary pressure in the cavity formed by the sample holder
and the foil. Work is currently under way to extend this analysis by including plasticity
and creep deformation. With respect to the results in Figure 4 note that the maximum
tangential stress occurs in the unsupported section of the foil and at the location where

the foil test specimen is welded to the sample holder. The magnitude of the maximum
tangential stress in the unsupported section of the foil agrees well with the value
predicted by a simple strength of materials analysis for a pressurized thin-walled vessel.
In this case, the magnitude of the maximum tangential stress is given by the following
relation:
σθ = P

r
t

(1)

where P is the magnitude of the pressure inside the vessel, r the radius of the cylinder and
t the thickness of the wall. For the analysis in Figure 4, the ratio (r/t) is approximately
equal to 10.

Figure 4. Stress distribution in foil specimen when subjected to internal pressurization.
Status of Milestones
FY 2001 - To acquire a commercially available microturbine, and initiate a testing
campaign using a modified recuperator. Met.
Problems Encountered
None.
Publications/Presentations
None.

CERAMIC RELIABILITY FOR
MICROTURBINE HOT-SECTION COMPONENTS

Reliability Evaluation of Microturbine Components
H-T Lin, M. K. Ferber, and P. F. Becher
Metals and Ceramics Division
Oak Ridge National Laboratory
P.O. Box 2008, Oak Ridge, TN 37831-6068
Phone: (865) 576-8857, E-mail: linh@ornl.gov

Objective
Evaluate and document the long-term mechanical properties of very small specimens
machined from ceramic components (e.g., blades, nozzles, vanes, and rotors) in as
processed and after engine testing at ambient and elevated temperatures under various
controlled environments.
This work will allow microturbine companies to verify
mechanical properties of components and apply the generated database in advanced
design and lifetime prediction methodologies. The work also provides a critical insight
into how the microturbine environments influence the microstructure and chemistry, thus
mechanical performance, of materials.
Highlights
A meeting was held between HT Lin (ORNL) and Vimal Pujari (Saint-Gobain) during
Cocoa Beach Meeting to discuss the mechanical testing plan for Norton NT154 silicon
nitride ceramic. A preliminary testing matrix and the number of billets needed were
outlined. Mechanical testing for bend bars with the as-processed and as-machined
surface will be carried out. In addition, mechanical testing for other potential silicon
nitride materials developed by Saint-Gobain was also discussed. The test matrix
proposed for NT154 will be refined based on the operating conditions provided by
microturbine companies when the materials are ready for testing.
Technical Progress
Characterization and mechanical testing for Rolls-Royce Allison uncoated SN282 and
coated AS800 silicon nitride vanes was completed during this reporting period. Analysis
of x-ray diffraction and SEM indicated that the microstructure and intergranular phase
composition were stable with time. However, both uncoated SN282 and plasma-sprayedoxide (PSO) coated AS800 vanes did experience significant material recession. The
recession rate appeared to be controlled by temperature and to a minor extent upon the
sintering additives. Results also indicated that the experimental Honeywell PSO EBC
employed in the present engine tests did not provide an adequate protection (diffusion
barrier) from gas turbine environments. Thus, the PSO coated AS800 vanes exhibited
similar recession rate and materials degradation to those uncoated AS800 vanes evaluated
in the Phase I engine test. Finally the strength measured for biaxial disks machined from
the uncoated SN282 airfoil surfaces did not change significantly with time. On the other
hand, the strength of samples machined from PSO coated AS800 airfoil surfaces

decreased with an increase in exposure time due to the formation of a very rough surface
and/or subsurface damage zone. These observations indicate that the lifetime of the
silicon nitride vanes is mostly controlled by dimensional and not mechanical
considerations. Therefore, alternative EBC systems need to be explored and developed to
inhibit material recession and ensure a long-term stability and performance in gas turbine
environments.
Dynamic fatigue tests on commercially available siliconized silicon carbide (Si-SiC)
materials were continued during this reporting period. The objective of dynamic fatigue
tests is to generate mechanical database for microturbine companies for their component
lifetime modeling tasks. Two commercial Si-SiC materials, i.e., CoorsTek SCRB210 and
Schunk, were evaluated under this subtask. Dynamic fatigue tests were conducted in
four-point bending using 20mm/40mm, α-SiC, semi-articulating fixtures at 1150°C in
air. Stressing rates of 30 and 0.003 MPa/s were employed to evaluate the fatigue effect.
Load was continuously measured as a function of time, and flexure strength was
calculated using ASTM C1161. Mechanical results showed that both CoorsTeck and
Schunk Si-SiC materials exhibited little or no strength degradation when tested at
1150°C, as shown in Table 1 and Figs. 1-4). Results also showed that the Weibull
modulus obtained was not sensitive to the test temperature. On the other hand, the
strength obtained at 0.003 MPa/s for CoorsTek Si-SiC was about 20% higher than that
obtained at 30 MPa/s, while the strength of Schunk Si-SiC was not influenced by the
stressing rate at 1150°C. The increased strength observed for CoorsTek Si-SiC could be
attributed to the softening of Si. The dynamic fatigue results showed that the Schunk SiSiC exhibited a higher dynamic fatigue exponent (N = 702) than that obtained for the
CoorsTek Si-SiC (51) at 1150°C, suggestive of a better slow crack growth (fatigue)
resistance, as shown in Figs. 5 and 6. The difference in characteristic strength and
Weibull modulus between CoorsTek and Schunk Si-SiC material could arise from the
difference in matrix SiC grain size and Si content.
Table 1. Summary of uncensored Weibull strength distributions for CoorsTek and Schunk Si-SiC materials
(longitudinally machined) per ASTM C1161.

Material

# of
Spmns.
Tested

Stressing
Rate
(MPa/s)

Temp.
(°C)

Uncens.
Weibull
Modulus

CoorsTek

13.03

CoorsTek

0.003

10.38

Schunk

8.4

Schunk

0.003

7.68

CoorsTek

1150

16.64

CoorsTek

0.003

1150

15.05

Schunk

1150

10.52

Schunk

0.003

1150

11.37

± 95%
Uncens.
Weibull
Modulus
9.13,
17.54
7.07,
14.46
5.76,
11.60
5.21,
10.69
11.58,
21.83
10.47,
20.48
7.35,
14.24
7.86,
15.56

Uncens.
Chrctstic
Strength
(MPa)
356
329
295
274
336
403
290
294

± 95%
Uncens.
Chrctstic
Strength
(MPa)
343,
369
313,
344
278,
312
257,
291
326,
346
389,
415
277,
304
281,
306

CoorsTek Si-SiC (SCRB 210)
Uncensored Flexure Strength Distribution
1150°C - 30 MPa/s - Longitudinally Machined
2

99.9
99.0

Max. Lik. Fit
Strength Data

90.0

ln ln ( 1 / ( 1 - P f ) )

50.0

m = 16.4 (11.58, 21.83)
σ = 336 MPa (326, 346)

-1

20.0

n = 20 specimens

-2

10.0

-3

5.0

-4

2.0
ASTM C1161-B
20/40mm fixture
Test dσ/dt = 30 MPa/s

-5

1.0

95% Confidence
Bands Shown

Probability of Failure, P f , (%)

0.5

-6
100

200

300

400

500

600

700

Failure Stress (MPa)

Figure 1. Strength distribution of CoorsTek Si-SiC tested at 1150°C and at 30 MPa/s.

CoorsTek Si-SiC (SCRB 210)
Uncensored Flexure Strength Distribution
1150°C - 0.003 MPa/s - Longitudinally Machined
2

99.9
99.0

Max. Lik. Fit
90.0

Strength Data

ln ln ( 1 / ( 1 - P f ) )

50.0

-1

-2

m = 15.05 (10.47, 20.48)
σ = 403 MPa (389, 415)

20.0

10.0

n = 20 specimens
-3

5.0

-4

2.0

-5

-6
100

1.0

ASTM C1161-B
20/40mm fixture
Test dσ/dt = 30 MPa/s

95% Confidence
Bands Shown

200

300

400

Probability of Failure, P f , (%)

0.5

500

600

700

Failure Stress (MPa)

Figure 2. Strength distribution of CoorsTek Si-SiC tested at 1150°C and at 0.003 MPa/s.

Schunk Si-SiC
Uncensored Flexure Strength Distribution
1150°C - 30 MPa/s - Longitudinally Machined
2

99.9
99.0

Max. Lik. Fit
1
Strength Data

90.0

ln ln ( 1 / ( 1 - Pf ) )

m = 10.52 (7.35, 14.24)
σ = 290 MPa (277, 304)

50.0

-1

n = 20 specimens
20.0

-2
10.0

-3

5.0

-4

2.0
ASTM C1161-B
20/40mm fixture
Test dσ/dt = 30 MPa/s

95% Confidence
Bands Shown

-5

1.0

Probability of Failure, Pf , (%)

0.5

-6
100

200

300

400

500

600

700

Failure Stress (MPa)

Figure 3. Strength distribution of Schunk Si-SiC tested at 1150°C and at 30 MPa/s.
Schunk Si-SiC
Uncensored Flexure Strength Distribution
1150°C - 0.003 MPa/s - Longitudinally Machined
2

99.9

Max. Lik. Fit

99.0

Strength Data

90.0

))

m = 11.37 (7.86, 15.56)
σ = 294 MPa (281, 306)

50.0

-1

n = 20 specimens
20.0

-2
10.0

-3

5.0

-4

2.0

95% Confidence
Bands Shown

-5

1.0

ASTM C1161-B
20/40mm fixture
Test dσ/dt = 30 MPa/s

Probability of Failure, P f , (%)

0.5

-6
100

200

300

400

500

600

700

Failure Stress (MPa)

Figure 4. Strength distribution of Schunk Si-SiC tested at 1150°C and at 0.003 MPa/s.

600

CoorsTek SCRB210
@ 1150°C

Fracture Strength (MPa)

500
400

300

N ~ -51

200

100
0.001

0.01

0.1

100

Stressing Rate (1/s)

Figure 5. Dynamic fatigue exponent of CoorsTek Si-SiC tested at 1150°C.

600

at 1150°C
in air

Fracture Strength (MPa)

500
400

300

N ~ -702
200

100
0.001

0.01

0.1

100

Stressing Rate (1/s)

Figure 6. Dynamic fatigue exponent of Schunk Si-SiC tested at 1150°C.
The SASiC MOR bars with and without EBC, which was developed by UTRC, were
received from Greg Ojard at UTRC for mechanical property evaluation. The purpose of

this study is to provide an understanding on the effect of EBC system on the mechanical
response of the substrates. Preliminary test results at room temperature will be reported
in the next quarterly report.
Status of Milestones
Complete characterization of microstructure and mechanical properties for UTRC silicon
nitride microturbine components by September 2002. On schedule.
Complete the database generation of commercial Si-SiC materials for microturbine
companies’ lifetime modeling task by June 2002. On schedule.
Industry Interactions
Communication with John Holowczak at UTRC to discuss the delivery status of the
Kyocera SN281 and SN282 microturbine rotors and machining status of the co-processed
billets.
Communication with Greg Ojard at UTRC to discuss the mechanical testing status of
SASiC bend bars with and without UTRC EBC, and also the test matrix of Kyocera
SN282 silicon nitride.
Communication with Josh Kimmel and Mark van Roode at Solar Turbines on the
mechanical testing results of SN88 silicon nitride bend bars with EBC.
Communication with David Bath at Kyocera on the test matrix for SN282 silicon nitride
manufactured with various green machining parameters.
Communication with Vimal Pujari at Saint-Gobain to discuss the mechanical testing of
Norton NT154 silicon nitride materials.
Communication with Russ Yeckley at Kennametal to discuss the mechanical testing of
α/β SiAlON materials.
Problems Encountered
None.
Publications
H. T. Lin and M. K. Ferber, “Mechanical Reliability Evaluation of Silicon Nitride
Ceramic Components After Exposure in Industrial Gas Turbines,” to be published in a
special issue of Journal of European Ceramic Society for the United Engineering
Foundation International Conference on "Structural Ceramics and Ceramic Composites
for High-Temperature Application," October 7-12, 2001, Seville, Spain.

Long-Term Testing in Water Vapor Environments
M. K. Ferber, and H-T Lin
Metals and Ceramics Division
Oak Ridge National Laboratory
P.O. Box 2008, Oak Ridge, TN 37831-6068
Phone: (865) 576-0818, E-mail: ferbermk@ornl.gov

Objective
The objective of this project is to develop test facilities for evaluating the influence of high-pressure and
high-temperature water vapor upon the long-term mechanical behavior of monolithic ceramics having
environmental barrier coatings.
Highlights
The prototype environmental containment system, which was modified to provide for more uniform
introduction of steam into the gage section of the button-head tensile specimen, was evaluated.
Preliminary tests of an NT154 silicon nitride (manufactured by Saint-Gobain Ceramics & Plastics, Inc.,
Northboro, MA) indicated that this approach resulted in both localized recession and enhanced
oxidation
Technical Progress
The effort this period focused on the design and fabrication of hardware for injecting steam into the hotzone of existing creep/stress rupture test machines. Based upon previous work, the most effective
method involves using either SiC or alumina tubes to inject steam onto the gage section of the tensile
specimen (see Figure 1). Based upon preliminary testing, this approach has been extremely effective in
simulating recession and enhanced oxidation. For example, Figure 2 illustrates the NT154 specimen
after 500h of exposure in a water vapor saturated environment at 1 atm total pressure and 1200°C.
The loss of material in the vicinity of the injection tube is indicated by the white line. Scanning electron
microscopy (SEM) of the surface in the SiC injection tube contact region (Figure 3) revealed localized
attack of the silicon nitride grains. No silica was detected in this region.

Approach 3

Hollow
Ceramic
Tube
Button-head
specimen

Grips

SiC or Alumina
Tube

Tensile
Specimen

SS Water Lines Wrapped
With Heating Tape

Water Pump

Figure 1: Direct Steam Injection System.

NT154 Tensile
Specimen

Injection
Region

Recessed
Area

Figure 2: NT154 Tensile Specimen after 500 h Exposure to Saturated Water Vapor Environment at
1200°C.

Figure 3: Surface of NT154 Specimen Surface in SiC Tube Contact Region.

Water Vapor Out

Water Vapor In

Gage Section

Figure 4: Schematic Illustration of the Water Vapor Flow for the Direct Injection System.
The reason for the relatively high recession may be a result of the higher than expected velocity of the
water vapor resulting from the confinement of gas in the small volume of the injection tube and forced

flow through the narrow gap that separates the injection tube from the tensile specimen (Figure 4). For
example, in the NT154 test, water at 25°C was introduced into the heated injection system at a rate of
0.1 cc/s. Based upon the density difference between steam at temperature and the water, this rate
increases to 673 cc/s. The linear velocity in the injection tube is estimated to be 34 m/s. An even higher
velocity is expected as the water vapor is forced to flow through the narrow gap between the gage
surface and injection tube.
One can use this information in conjunction with the NASA-Glenn volatility model [1] to calculate the
loss of material expected for the NT154 tensile test. For a water vapor pressure was 1 atm, a
temperature of 1200°C, and a gas velocity of 34 m/s, the recession is estimated to be 49 µm which is
about a factor of 3 lower that the measured value of 126 µm.
Status of Milestones
Milestone 1: Design, fabricate, and evaluate steam containment system for existing creep-stress rupture
rigs and issue letter report (April 1, 2001/Delayed 12 months/Completed April 10, 2002).
Milestone 2: Conduct preliminary environmental stability tests on uncoated SN282 and issue letter
report (July, 2002/On schedule/ Delayed 4 months).
Milestone 3: Modify 4 test frames to accommodate direct steam injection system (March 2002).
Industry Interactions
Held several discussions with Jim Kesseli of Ingersoll-Rand Energy Systems concerning material
availability and database access.
Held discussions with Curt Johnson and Reza Sarrafi-Nour of GE concerning database and future
materials testing.
Held extensive discussions with Greg Ojard of UTRC concerning implementation of an ongoing study
for EBC development for silicon nitride ceramics.
Problems Encountered
None.
Publications/Presentations
A presentation on ceramics was made by Matt K. Ferber at the Peer Review of the Microturbine and
Industrial Gas Turbine Programs, held on March 12-14, 2002 in Fairfax, VA.
Reference
J. L. Smialek, R. C. Robinson, E. J. Opila, D. S. Fox, and N. Jacobson, “SiC and Si3N4 Recession due
to SiO 2 Scale Volatility Under Combustor Conditions, “ Advanced Composite Materials, Vol. 8, No.
1, pp. 33-45, (1999).

Reliability Analysis of Microturbine Components
S. F. Duffy, E. H. Baker, and J. L. Palko
Connecticut Reserve Technologies, LLC
2997 Sussex Ct., Stow, OH 44224
Phone: (216) 687-3874, E-mail: sduffy@crtechnologies.com

Objective
The work under this contract is separated into four tasks as outlined below:
Task 1 - Connecticut Reserve Technologies, LLC (CRT) will interface WeibPar, an in-house Weibull
parameter estimation software package, with the reliability modules of NASA GRC’s CARES/Life.
Task 2 - CRT will provide development and design expertise in support of the companies involved in
the DoE Microturbine project. CRT will aid design engineers in order to assess the life of micro-turbine
ceramic components. This will help to establish life cycle costs. These predictions will be based on the
finite element results as well as the material specific information generated from the laboratory tests. The
CARES/Life programs will provide the estimates for the component life.
Task 3 – Transient Reliability Analysis will be added to NASA’s existing CARES/Life ceramic
reliability algorithm software. Work will entail coding the algorithm within the Fortran module and
integrating the Fortran modules with a Windows GUI.
Task 4 –Develop a ceramics material reliability database graphical user interface (GUI) for Windows
PC computers. Cooperate with ORNL, GRC and UDRI personnel to specify formats and capabilities
of a ceramic specimen database for fast-fracture, fatigue, creep, and recession properties of materials.
The database should have the ability to integrate with CARES/Life and CARES/Creep life prediction
codes.
Highlights
Working with Matt Ferber and Ingersoll Rand to produce a graphical tool to analyze strength at
temperature databases relative to a specific structural component.
Technical Progress
Task 1 – The revised scope for this task has been stated above. CRT has a verbal agreement with
ORNL to substitute WeibPar as the Weibull parameter estimation software portion of this task. The
necessary modifications to the software have been outlined. It has been determined that WeibPar will
act as the controlling application. WeibPar will therefore execute CARES/Life as a background
function. All user input and output to CARES/Life will controlled by WeibPar.

Task 2 –CRT provided Ingersoll Rand with the ability to assess baseline failure data at temperature
using the pooling techniques provided by the WeibPar algorithm. A spreadsheet was developed that
utilizes estimates of the Weibull material scale parameter (as a function of temperature) and the Weibull
modulus as well as the size scaling relationship inherent to Weibull analysis. The result is a characteristic
strength versus temperature graph where the user can input the characteristic volume or characteristic
area of a component. The resulting curve adjusts up or down based on the geometry and load applied
to the structural component being analyzed. An example was provided using AS800 material data
provided by Matt Ferber at ORNL. This approach provides industry with a graphical screening tool
where the design engineer can quickly assess a particular material relative to a specific component.
Knowing the maximum stress and use temperature for a given component, the design engineer can see
graphically whether the material is worth considering for the design application. This feature will be
incorporated in the WeibPar algorithm under the next phase of this contract.
Task 3 – Batdorf reliability routines have been added to the transient reliability module of CARES/Life.
Example problems are being created to test and demonstrate the transient capabilities. An updated
input file structure is being reviewed. Once the input file structure is fixed the Windows GUI will be
updated accordingly.
Task 4 – Discussions with Matt Ferber at ORNL has led to structuring material data based on use.
The three use categories will be: (1) Structural Ceramics, (2) Thermal Ceramics, and (3) Electrical
Ceramics. Each use classification has material properties specific to a design application. This will
simplify the data storage format.
Status of Milestones
Milestones are on schedule.
Industry Interactions
See Task 2 progress note above.
Problems Encountered
None.
Publications/Presentations
None.

NDE Technology Development for Microturbines
W. A. Ellingson, E. R. Koehl, A. Parikh, and J. Stainbrook
Argonne National Laboratory
9700 South Cass Avenue
Argonne, IL 60439
Phone: (630) 252-5068, E-mail: ellingson@ornl.gov

Objective
The objective of this project is development of nondestructive evaluation/characterization (NDE/C) technologies
for: (1) evaluating low-cost monolithic ceramics for hot section components of microturbines or industrial gas
turbines, (2) evaluating environmental barrier coatings (EBCs) for monolithic ceramics or ceramic matrix
composites, and (3) evaluating other materials which are part of the technology to advance DER technologies
such as ceramic -metal joints. The project is directly coupled to other OPT-DER projects focused on materials
developments directed towards low-cost, high volume monolithic ceramics, environmental barrier coating
systems and related technologies such as ceramic -metal joining.
Highlights
There are two highlights this period. First, we have set up the new 6-axis articulated robot arm so that NDE
methods can be applied to complex shaped vanes and blades. Second, we received the new 80 µm pixel,
CMOS-electronic based X-ray detector which will allow very high spatial resolution for small flaw detection.
Technical progress
Technical work this period focused on 3 areas: (1) getting the robot arm set up and software developed so it will
accept CAD files of turbine vanes and blades, (2) further developments towards high speed X-ray imaging for
low-cost monolithic ceramic materials, and (3) continued work on NDE for EBCs on SiC/SiC composites.
1. Robot Arm
The articulated robot arm is being set up to allow NDE of complicated geometries such as vanes and blades. The
major tasks include: path programming the robot, coordinating data acquisition and robot path, computer
connection between the robot and a personal computer. Further tasks will include studying resulting data as a
function of the part geometry mapping the damage/data onto the 3D part surface.
Robot System:
The system includes ABB’s 6 axis articulated robot arm model IRB 140 (M2000), controlled by ABB S4C+
controller with a teach pendant. The path programs for the robot are developed using the offline programming
and simulation package RobotSTUDIO from ABB. The Robot Arm is mechanically mounted on an optical
worktable. The software allows virtual robot simulation without actually running the commands/programs on
the real robot.
Software and Interfacing:
The software used to control the robot includes RobotSTUDIO, an offline programming and simulation software
from ABB. The Data Acquisition (DAQ) software being developed is written in LabVIEW so that there is a
“user friendly” interface. The procedure we follow is to do simulation of the actual system in RobotSTUDIO,
then export the program to the Robot controller’s language (RAPID) and port that into the robot’s
programmable logic controller (PLC). Depending on the task to be accomplished, minor modifications

(add/delete routines, change velocity, etc.) can be added to the robot program in the PC. In our case we interface
the robot program to the DAQ program for NDE data acquisition. We must insert a line of code to send
signal/data to the serial port of the controller at certain desired positions (start and end of scan lines). This is
accomplished because the controller is connected directly to the DAQ computer via a RS-232 serial link cable.
The structure of the DAQ program is listed below in Table 1.
Table 1. Procedure for Robot
SET the port address for SERIAL PORT
SET the port address for POWER METER
CHECK CONTINUOUSLY for signal from SERIAL PORT
CONTINUOUSLY receive data from the POWER METER
WAIT for START OF LINE signal from robot controller
START WRITING DATA TO FILE
IF (SIGNAL FROM CONTROLLER = 0)
STOP WRITING DATA TO FILE
ELSE
CONTINUE WRITING DATA TO FILE
END PROGRAM AFTER SCANNING IS DONE
Obtaining Laser Scatter NDE Data:
In order for the robot to be coupled to the NDE laser scatter device to study delams and thickness of EBCs on
vane and blades, several software modifications were needed for the DAQ software.
For example, at the end we get a laser-scatter data file containing the detected laser power and then we need to
re-set the scan line. For the first test, to test our software, we produced a simple 50 mm square plate (see Fig. 1),
and we decided to take 100 data points per line. As part of this we must set the speed of the robot arm to match
our requirement of 100 data points. The speed and data points are governed by the response of the power meter.
Considering that the spot size of laser beam ˜ 1-2 mm, the distance to be scanned ˜ 50 mm; number of data
points we want in 50 mm is 100, the max readout speed of the power meter is 100 Hz, meaning that in 1 second
it can receive 100 data points (maximum). Thus, to receive 100 data points (in 50 mm) takes 1 second. This
implies we could set the speed of robot to be 50 mm/s (maximum speed). We also need to take into account the
data transfer rates of signal to and from PC/Robot controller.
There are several steps involved when using the robot. These include:
1) Mount the sample on the robot
2) Import the CAD file into RobotSTUDIO (RS).
3) Position and manipulate the robot to match actual orientations.
4) Define tool path and work object in RobotSTUDIO.
5) Make target points on test object by intersecting the surface of interest with a imaginary plane
6) Orient test object surface to keep normal to laser beam
7) Make a path from these targets by chaining them together
8) Repeat steps 5-7 by changing the position of intersecting plane to get a number of paths (scan lines).
9) Export the program from RobotSTUDIO.
10) Edit the program file to insert commands to interrupt data writing to file when it’s the end of a line and
restart it at the start of a new line.
11) Transfer the robot program to the programmable logic controller (PLC) via FTP from the computer.
12) Start the DAQ program then start the robot control program

As noted earlier, in order to study the robot motion and attempt to acquire our first NDE laser scatter data set,
we had a simple 50 mm square flat plate made with 1.6 mm holes drilled at various locations. Figure 1 shows a
diagram of the plate.

1.6 mm holes
(5 locations)

Fig. 1. Robot arm – NDE data acquisition test plate, 50 mm x 50 mm x 12.5 mm
Part of the laser back-scatter NDE effort for EBCs on vanes and blades, concerns the way the laser light is
incident upon the sample. Two configurations have been studied: (a) normal incidence, and (b) oblique
incidence. For the robot arm to be used, it must be controlled in such a way that the local surface normal is
always known. Figure 2 shows schematically this concept. This shows angle-of-incidence scattering but
normal incidence with normal reflectance can also be achieved.

θ
Incident
laser beam
Imaginary
surface normal
and angle

Scattered
laser beam

Fig. 2. Schematic diagram of the laser-scatter concept and local surface normal.
The example shown is for angle -of-incidence scattering.
It was noted earlier that for the robot scan data to be acquired, the data are only acquired in one direction. That
currently is left to right but can be reversed. Figure 3 shows schematically this idea.

(a)

(b)

Start new line
scan. Send “1” to
DAQ program

End of scan line
send “0” to DAQ
program

Fig. 3. Schematic diagram showing the left-justified data acquisition scan path from the robot.
(a) Schematic of path on object; (b) detail of start-stop of line scan.
2. NDE development for on-line low-cost monolithics
Work this period focused on two areas: (1) initial installation and check out of the new 80 µm CMOS-based
electronic X-ray detector, and (2) continued evaluation of the spatial resolution capabilities of the large 40 cm
by 40 cm amorphous silicon flat panel detector being used for full scale 3D imaging.
80 µm linear detector
This period we received the new 80-µm square pixel detector from Envision. We interfaced the detector to new
software to allow initial images to be acquired. Figure 4 shows a schematic of the set up using lead letters,
ANL, for the first image. Shown in Fig. 5 is the resulting first projection image.

Fig. 4. Schematic diagram showing location of lead letters used
for first CMOs X-ray image acquisition.

Fig. 5. Schematic diagram and resulting X-ray image taken of section of the letter.
Large Flat Panel for Direct Digital Imaging
We continued this period to evaluate our large flat panel detector in order to evaluate the spatial resolution
capability. We began the effort to experimentally determine the detail-detection diagram for the 3D CT scanner.
Figure 6 below shows the resulting defect-detection diagram. Note this is for existing 400 µm square pixel
detector and the new 200 µm square pixel detector is expected to arrive in July. A new defect-detection diagram
will then be generated.

Defect Detection Diagram
50000
45000

178 mm Pins
178 mm Void

Grayscale Difference

40000

112 mm Pins
112 mm Void

67 mm Pins
67 mm Void

35000
30000
25000
20000
15000
10000
5000
0
0

0.2

0.4

0.6

0.8

1.2

Defect Diameter (mm)

Fig. 6. Defect-detection Diagram for large area (40 cm by 40 cm) flat panel detector
with 400 µm pixels and a 200 µm square X-ray source size.
In order to establish the defect-detection diagram, we used three cylindrical discs of different
diameters: 67 mm, 112 mm and 178 mm. These were made of gelcast AS800 and were cut from
sections of an unbladed rotor. These sections were each about one cm thick. In each disc we drilled
three holes using an ultrasonic drilling machine. The three holes were 390 µm, 600 µm and 1.1 mm in
diameter. X-ray acquisition tests were conducted with nothing in the drilled holes as well as with
stainless steel inserts in the holes. We used two reconstruction algorithms for the reconstructed images
from which the measurements were made. One algorithm is called a “parallel” beam code and the other
is called a “cone” beam code. These each provide different attributes. Shown below in Fig. 7 is one of
the discs used for the defect-detection diagram. It is the middle disc that is 112 mm in diameter. Also
shown in Fig. 7 are two of the X-ray computed tomography images for each of the reconstruction
algorithms. By using the X-ray CT images as the data sets, a line profile across the features of interest
was drawn and the full width-half max (FWHM) values were used to obtain the feature sizes. This
kind of analysis was also used to establish the gray scale value. The difference of the gray scale value
at this peak was subtracted from the “air” or background value and this difference was then plotted
against known feature size as shown in Fig. 7.

Section D
(middle)

Conebea
m
(b)

(a)

112 mm

(c)

Fig. 7. One of the AS800 discs used for x-ray computed tomography studies for defect-detection
diagram. (a) Optical photomicrograph of the disc showing the ultrasonically drilled holes as well as the
existence of a natural crack, (b) X-ray computed cross section using the cone beam algorithm,
(c) X-ray computed tomography cross section using the parallel beam algorithm
2. NDE development for EBC coated monolithics and Composites
We received additional samples this period of EBC monolithics and the results will be reported next
period. We also conducted tests on EBC-CMC. These will be reported next period.
Status of milestones
Milestones are on schedule
Industry/National Lab Interactions
Discussions took place this period with a variety of institutions involved with the efforts of this project.
Discussions were held with Matt Ferber of ORNL relative to the EBC coated vanes from Rolls
Royce/Allison
Problems encountered
None.

Publications/Presentations
Two papers were written and submitted to the 46th ASME Gas Turbine and Aeroengine Technical
Congress to be held June 3-6 in Amsterdam, The Netherlands.
Trips/Meetings
W. A. Ellingson attended the 26th annual International Conference on Advanced ceramics and
Composites sponsored by the American Ceramic Society January 13-18, 2002, in Cocoa Beach, FL.
W. A. Ellingson attended the 26th annual Conference on Composites, materials and Structures
sponsored by the United States Advanced Ceramics Association, USACA, The National Institute of
Ceramic Engineers, NICE, as well as the US Department of Defense and NASA held January 28-31,
2002, in Cocoa Beach, FL.
W. A. Ellingson participated in the International Conference on Advances in Life Assessment and
Optimization of Fossil Power Plants sponsored by the EPRI held March 11-13, 2002, in Orlando, FL.

CHARACTERIZATION OF ADVANCED
CERAMICS FOR INDUSTRIAL GAS TURBINE/
MICROTURBINE APPLICATIONS

Oxidation/Corrosion Characterization of Monolithic Si3N4 and EBCs
K. L. More and P. F. Tortorelli
Metals and Ceramics Division
Oak Ridge National Laboratory
P. O. Box, Oak Ridge, Tennessee 37831-6064
Phone: (865) 574-7788, E-mail: morekl1@ornl.gov

Objectives
Characterization and corrosion analyses of Si3 N4 materials provided to ORNL as part of the HotSection Materials/Component Development Program
Exposures of candidate Si3 N4 materials to high water-vapor pressures (in Keiser Rig) to simulate
high-temperature, high-pressure environmental effects associated with microturbines
Evaluate the reliability of environmental barrier coatings (EBCs) on silicon nitrides for selected
microturbine applications
Highlights
Residual strength measurements were made on three different Honeywell Si3 N4 materials
(uncoated) after exposure in the Keiser Rig for 500, 1000, 1500, and 2000h at 1315°C and 3% or
20% H2 O. After exposure for 2000h at 3% H2 O, there was a loss of ~20% in strength for AS800
Si3 N4 and no measurable strength loss for AS950 Si3 N4 . As expected, the loss of strength for
AS800 Si3 N4 exposed for 2000h at 20% H2 O was greater than that observed after 2000h at 3%
H2 O. After 2000h at 20% H2 O, the AS800 material lost ~33% of its strength. The AS950
appears to be performing somewhat better than the AS800. However, 2000h data at 20% H2 O
for AS950 is not yet available for direct comparison with the AS800 data.
Technical Progress
A furnace system that provides a high-temperature, high-pressure, low-flow-velocity (< 0.1 fps)
mixed-gas environment (ORNL’s Keiser Rig) serves as a means to conduct first-stage
evaluations of Si3 N4 and Si3 N4 +EBC coupons provided by Honeywell Ceramic Components
(HCC) in support of hot section materials/component development for microturbines. A
summary of the specimen runs conducted in the Keiser Rig for this project was given in a
previous quarterly report (DER Quarterly Progress Report for July 1, 2001 – September 30,
2001). Since this report was published, four more runs have been conducted in the Keiser Rig at
1315°C completing 2000h exposure at 3% and 20% H2 O for all three types of Si3 N4 being
evaluated in this study (slipcast AS800, gelcast AS800, and slipcast AS950). Several HCC
experimental environmental barrier coatings (EBCs) have also been exposed in the Keiser Rig
(DER Quarterly Report for October 1, 2001 – December 31, 2001).

During this quarter, extensive mechanical property evaluations were conducted on uncoated
Si3 N4 after exposure to 3% and 20% H2 O at 1315°C for 500, 1000, 1500, and 2000h in the
Keiser Rig. The starting Si3 N4 specimens were 2.54 cm X 5.08 cm plates with one as-processed
surface and one machined surface. The plates are hung from an alumina rod in the Keiser Rig.
Enough plates of each of the three Si3 N4 materials were exposed in the Keiser Rig at 1315°C to
finish with three plates of each type of Si3 N4 exposed at either 3% or 20% H2 O for each of four
times of 500, 1000, 1500, 2000h. Four-point bend specimens were machined from each plate
after exposure. Three four-point bend specimens were machined from each Si3 N4 plate for a
total of nine four-point bend test specimens at each operating condition and time. The asprocessed surface was always on the tensile face during mechanical testing since as-processed
surfaces have been shown to have lower oxidation rates than machined surfaces.
A summary of the strength results for all three Si3 N4 materials exposed at 1315°C and 3% H2 O
are shown in Figure 1.
Strength loss for each of the AS800 materials exposed was
approximately the same after 2000h, 16-17%, whereas no measurable loss was observed after
2000h exposure at 1315°C for the AS950 Si3 N4 in the Keiser Rig. When the three Si3 N4
materials were exposed in the Keiser Rig at 1315°C and 20% H2 O, significant strength losses
was observed for all the Si3 N4 materials after exposure. As shown in Figure 2, which
summarizes the strength data for Si3 N4 exposed at 1315°C and 20% H2 O, the gelcast AS800 lost
~33% of its strength after 2000h. The strengths of slipcast AS800 and slipcast AS950 exposed
for 2000h at 20% H2 O have not been measured yet, however, after exposure for only 1500h, the
strength of the slipcast AS800 decreased by ~35% and the strength of the AS950 decreased by
~20%. At both water-vapor pressures, AS950 had the least amount of strength reduction of the
three different Si3 N4 materials and of the two AS800 compositions, the gelcast material appears
to be superior to the slipcast material.

800
0 hours

700

500 h
1000 h

600

1500 h
2000 h

500
400
300
200
100
0
AS800 Gelcast (AP)

AS800 Slip Cast (AP)

AS950 Slip Cast (AP)

Figure 1. Summary of Si3 N4 materials exposed for 0-2000h at 1315°C and 3% H2 O in Keiser
Rig.

800
0 hours
500 h

700

1000 h
1500 h

600

2000 h

500
400
300
200
100
0
AS800 Gelcast (AP)

AS800 Slip Cast (AP)

AS950 Slip Cast (AP)

Figure 2. Summary of Si3 N4 materials exposed for 0-2000h at 1315°C and 20% H2 O in Keiser
Rig.
Status of Milestones
08/2001

Complete 2000 h exposures on three different Honeywell Si3 N4 materials in
ORNL’s Keiser Rig, characterize microstructural changes, and determine material
recession rates. Report results.
Milestone is complete. Results were reported/presented at Honeywell Engines &
Systems on March 20, 2002.

Industry Interactions
Attended Workshop on Microturbine Application, January 17-18, 2002, College Park, MD.
Discussed ongoing collaborative work with several microturbine industries.
Visited Honeywell Engines & Systems, Phoenix, AZ on March 20, 2002 to discuss current status
of silicon nitride microturbine materials work (described herein). Presented results on Keiser
Rig exposures of coated and uncoated silicon nitride.
Honeywell Ceramic Components is supplying the silicon nitrides (with and without coatings) for
exposures in the Keiser Rig and characterization at ORNL. Collaborations (meetings, conference
calls, e-mails) with Honeywell Ceramic Components, Honeywell Corporate Technology, and
Honeywell Engines & Systems are ongoing.
Problems Encountered
None.

Publications/Presentations
K. L. More and P. F. Tortorelli, “Current Status of Si3 N4 Exposures at ORNL,” presented during
visit to Honeywell Engines & Systems, Phoenix, AZ, March 20, 2002.

Mechanical Characterization of Monolithic Silicon Si3N4
R. R. Wills, S. Hilton, and S. Goodrich
University of Dayton Research Institute,
300 College Park, KL-165, Dayton, OH 45469-0162
Phone: (937) 229-4341, E-mail: roger.wills@udri.udayton.edu

Objective
The objective of this project is to work closely with microturbine materials suppliers to characterize
monolithic ceramics and provide the data obtained to microturbine manufacturers via the Web site
database as well as user-friendly software which will allow perspective users to readily compare
different silicon nitrides. This project consists of the following four tasks:
•
•
•
•

Task 1: Evaluate Strength and Slow Crack Growth of New Materials
Task 2: Modify Six Existing Creep Frames to Allow Introduction of Water Vapor
Task 3: Evaluate the Effects of Water Vapor Upon Honeywell’s Silicon Nitride Ceramics
Task 4: Develop “User Friendly” Software for Searching Existing Mechanical Properties
Database

Task 1 is motivated by material needs of both Ingersoll- Rand (IR) Energy Systems and UTRC. The
ceramic materials being considered by IR Energy Systems include Kyocera’s SN235 and SN237 for
which the required mechanical property data are somewhat limited. In the case of the UTRC
microturbine, Si-SiC is a prime candidate for the combustor. The goal of Task 2 is to modify existing
facilities to evaluate the effects of water vapor upon the strength and creep rupture behavior of both
uncoated and coated Si3N4 ceramics. In Task 3, the facilities developed in Task 2 will be used to
evaluate the mechanical behavior of specimens with and without an environmental barrier coating
(EBC).
Highlights
Strength, creep and slow crack growth data were generated for a siliconized silicon carbide (Coorstek
SC-2) this reporting period. This material is of interest to UTRC for their microturbine project.
Technical Progress
Rectangular billets and rods of a commercial siliconized silicon carbide were procured. In siliconized
(or reaction sintered) SiC, SiC and C mixtures are first formed into shaped compact. This compact is
then exposed at high temperatures to molten Si, which infiltrates into the pore structure and reacts with
C to form SiC. The resulting material consists of the original SiC grains, newly formed SiC, and residual
Si.
Four-point flexure bars having approximate dimensions of 3 by 4 by 50 mm were machined from the

billets. Fast fracture at room temperature is complete. Dynamic stressing rate data collected so far at
1100°C, 1200°C and 1300°C is presented below together with Weibull plots.
Dynamic Fatigue Strength as a Function of Stressing Rate
CoorsTek SC-2
April 2, 2002

Test
Temperature
(°C)
20
1100
1200
1300
20
1100
1200
1300

Stressing
Rate
(MPa/s)
30
30
30
30
0.003
0.003
0.003
0.003

Specimens
Tested
15
15
15
15
0
12
15
6

Flexural
Strength
(MPa)
415
386
432
402
300
285
254

Four Point Flexure Test Results at 20°C
CoorsTek; SC-2

April 2, 2002

Specimen
Number
16
17
18
19
20
21
22
23
24
25
26
27
28
29

Stressing Rate
(MPa/s)
30
30
30
30
30
30
30
30
30
30
30
30
30
30

Load
(lbs)
117
111
120
112
122
102
115
117
112
116
105
113
117
116

Flexural Strength
(MPa)
431
407
440
410
446
376
423
427
409
425
386
415
428
426

Standard
Deviation
(MPa)
22
8
21
31
15
22
11

Weibull
Modulus
25
59
25
16
17
-

103

376

Four Point Flexure Test Results AT 1100°C
CoorsTek; SC-2

April 2, 2002

Specimen
Number
41
42
43
45
46
47
48
50
51
52
53
79
80
82
83
100
101
102
103
104
105
106
107
108
109
110
111

Stressing Rate
(MPa/s)
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
0.003
0.003
0.003
0.003
0.003
0.003
0.003
0.003
0.003
0.003
0.003
0.003

Load
(lbs)
105
105
100
108
106
104
108
107
104
106
108
106
104
103
107
79
84
84
73
87
86
83
85
84
86
78
80

Flexural Strength
(MPa)
384
387
366
393
387
379
397
391
379
389
397
389
381
377
392
289
303
308
265
316
313
303
309
309
312
287
292

Four Point Flexure Test Results AT 1200°C
CoorsTek; SC-2

April 2, 2002

Specimen
Number
34
35
36
54
55
56
57
58
59
60
61
62
63
64
65
89
90
91
92
93
95
96
97
98
99
115
116
118
119
120

Stressing Rate
(MPa/s)
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
0.003
0.003
0.003
0.003
0.003
0.003
0.003
0.003
0.003
0.003
0.003
0.003
0.003
0.003
0.003

Load
(lbs)
118
106
122
118
119
126
111
126
114
112
125
121
117
121
120
84
75
73
80
80
81
74
78
77
76
85
83
62
87
80

Flexural Strength
(MPa)
432
389
444
432
435
461
405
462
416
409
458
442
426
439
436
305
272
265
290
288
294
270
284
282
280
310
302
226
317
294

Four Point Flexure Test Results AT 1300°C
CoorsTek; SC-2

April 2, 2002

Specimen
Number
37
39
66
67
68
69
70
71
72
73
74
75
76
77
78
84
85
86
87
88
121

Stressing Rate
(MPa/s)
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
0.003
0.003
0.003
0.003
0.003
0.003

Load
(lbs)
115
97
119
117
102
120
116
101
105
121
115
105
112
107
97
70
75
65
69
68
71

Flexural Strength
(MPa)
421
354
436
427
371
438
424
372
385
444
419
382
409
393
355
253
272
238
252
248
259