Heinrich J.G., Aldinger F. (Eds.) Ceramic Materials and Components for Engines

Подождите немного. Документ загружается.

This Page Intentionally Left Blank

INTERNALLY COOLED MONOLITHIC

SILICON NITRIDE AEROSPACE COMPONENTS

Jonathan

Best*, James

Cawley*, Ramakrishna T, Bhatt** and Dennis

Fox***

(*)

Materials Science and Engineering, Case Western Reserve University

10900 Euclid Avenue, Cleveland, Ohio 44106

(**)Army Research Laboratory Vehicle Tech. Dir., 21000 Brookpark Rd.

Cleveland, Ohio 44135

(***)

NASA Glenn Research Center, 21000 Brookpark Rd. Cleveland, Ohio 44135

ABSTRACT

set of rapid prototyping (RP) processes have been

combined with gelcasting to make ceramic aerospace

components that contain internal cooling geometry.

mold and core combination is made using a MM6Pro

(Sanders Prototyping, Inc.) and SLA-250/40

(3Dsystems, Inc.). The MM6Pro produces cores from

ProtoBuildTM wax that are dissolved in room

temperature ethanol following gelcasting. The

SLA-

250/40 yields epoxy/acrylate reusable molds. Parts

produced by this method include two types of

specimens containing a high density of thin long cooling

channels, thin-walled cylinders and plates, as well as a

model hollow airfoil shape that can be used for burner

rig evaluation of coatings. Both uncoated and mullite-

coated hollow airfoils has been tested in a Mach

0.3

burner rig with cooling air demonstrating internal

cooling and comfirming the effectiveness of mullite

coatings.

INTRODUCTION

The efficiency of a gas turbine engine

increased

by raising the turbine inlet temperature and the

maximum temperature is ultimately limited by material

properties. In order to approach service near

90%

their melting point, superalloy blades and vanes have

become elegantly engineered’. In part this is

microstructural, for example, the progression from

equiaxed to directionally solidified to single crystal and

the gamma gamma-prime strengthening. But of equal

importance is the development of coring technologies

that permit internal cooling and the application of

environmentally resistant coatings. The former allows

the surface temperature of the component to kept to a

maximum that is significantly below the ambient

whereas the latter mitigates the effects of oxidation and

other corrosion processes as well as providing a thermal

barrier.

Much of the work on producing ceramic materials

has focused of the development of material properties

such as resistance to thermal shock and oxidation, high

temperature strength and stiffness, and creep resistance.

And the results have been singularly successful.

particular, both commercial and research grades silicon

nitrides (usually alloys containing rare earth oxides

combination with aluminosilicates) are available which

can tolerate sudden quenches

excess of 1000°C, have

strengths greater than

800

MPa, and which tolerate

service operating temperatures up to 13OOOC. In-situ

reinforced silicon nitride formulations have been

developed with fracture toughnesses

excess of

MPadm.

The combination of such properties are

immediately suggestive of application in gas turbine

(and rocket) applications’. Furthermore, the

effectiveness of internal cooling has been demonstrated

by Tsuchiya et al’ who have shown that the

reasonable to for the surface temperature of silicon

nitride components in a

150OOC

environment to be kept

below 1300°C through the use of internal air cooling,

which require modest flows.

Engineering the surface is suggested by work that

has been done demonstrating the use of plasma-sprayed

mullite to produce thermal and environmental barrier

coatings on silicon carbide and silicon nitride4.

The work reported herein represents an effort to

develop a process strategy to prototype silicon nitride

parts capable of internal cooling that is robust and easy

implement. Conceptually, the process is quite

straightforward: molds and cores are produced by two

different commercial RP processes and assembled; a

gelcasting slurry is prepared and forced into the cavity

under low pressure; after gelling, the part is liberated

through a combination of dissolution to remove the

cores and diassembly to remove the reusable molds.

The resultant wet part is dried and fired via conventional

means. Details of the process are given elsewhere’.

Gelcasting is a process developed at Oak Ridge

National Laboratory6 and it is particularly amenable to

use with soft tooling, such as that produced by polymer-

based RP processes because only very low pressures are

employed. In fact, Jamalabad et al. suggested the use

gelcasting in fugitive molds sometime ago’. The

approach taken

this work differs

that a combination

of RP methods is used, the design of the molding system

is different, and that a process for setting the gel,

immersion

a poly(ethy1ene glycol)

(PEG),

is used

incorporated.

The particular implementation of gelcasting

used

this work uses water-soluble organic materials.

Two monomers are employed: methacrylamide (MAM)

which polymerizes as a linear chain and

N,N’-

463

methylenebisacrylamide (MBAM) which is a

crosslinker. Just prior to casting an initiator,

N,N,N’,N’-

tetramethylethylenediamine

(TEMED), and

a catalyst, ammonium peroxydisulfate (APS), are mixed

into the slurry to commence the gelation process.

Standard stereolithography was used to

produce reusable molds. A SLA250/40 was used to

build epoxy/acrylate parts that survived molding and

subsequent immersion in ethanol and PEG without

noticeable damage

distortion. Dozens of casts were

made using the same epoxy/acrylate parts.

The Sanders Prototyping process was used to

produce cores and mold lines which were dissolved,

after casting, by soaking in room temperature ethanol

for several hours. A ModelMaker MM6Pro was used.

The Sanders process employs two materials to !kilt a

part. The actual part is built out of a “green wax”

(ProtoBuildTM) and the surrounding supports are built

out of a “red wax” (ProtoSupportTM). The waxes are

melted and each layer of the part is deposited using two

separate ink jets to spray droplets. A mechanical

grinder levels off each layer to the correct thickness.

These process steps continue until the entire part is

built. Upon completion the red supporting wax can be

selectively removed by heating to 40-50°C in a VS-0

bath. Beyond the fact that the green wax can be

dissolved with ethanol at room temperature, this method

is particularly suited to the production of cores

small

characteristic dimension

requiring fine details.

All CAD work was carried out using Rhino3D

(Robert McNeel and Associates) a NURBS package that

runs on PentiumTM class PCs.

PROCEDURE

Three types of parts were built. The first was a

reverse-engineered model airfoil shape based on

archival superalloy parts left over from the original

thermal barrier coating development program at NASA

Lewis Research Center (now NASA Glenn Research

Laboratory). The part allows internal cooling and was

readily coated with plasma sprayed mullite.

The other

two were demonstration parts to permit the a critical

evaluation of the possibility

for using thin cooling

channels of long length within the walls

thin-walled

plates and cylindrical sections.

In initial experiments, molds were entirely

epoxy/acrylate and a mold release was used to effect

separation of the gelled part from the mold. Although

often successful, distortion due to demolding stresses

was enough of a nuisance that soluble mold linings were

developed and used with great success. Mold linings

were made by the exact same process as the cores and

were simultaneously removed.

The net result is that

the space filled by the slurry was entirely bounded by

surfaces made up of the Protobuild material. This was

true for all of the parts discussed.

Figure 1 shows the CAD representation of the

molding system used to make the thin-walled internally

cooled hollow cylinder. Similar mold systems are used

for other components. In practice, the joints of the

assembled mold were sealed with a soft patching wax

(Kindt-Collins) and held together with teflon tape as

needed. After the casting has gelled, the entire mold is

immersed in ethanol to dissolve the wax components of

the molds. Leachable paths are designed such that all

the wax parts are connected to the external surface. The

wax linings and cores dissolve away

a few hours.

The part is then submerged

PEG for drying.

the

gating structure would restrict drying shrinkage creating

undesirable stresses within the part, it is trimmed from

the part prior to drying. After the part is dried, typical

ceramic firing procedures are employed. The process

steps are illustrated in Fig.

using a different mold/core

set to illustrate each.

FABRICATED PARTS

The results

the molding process as well

represented by the results shown in Fig. 3 for the

cylinder. The cast part contains some flashing that must

be manually removed, but the overall shape and size is

accurate and the high density of small diameter holes

produced by the cores are excellent. The pathway of the

cooling holes is visible in the fired piece because of the

locally thin walls. Very similar results were obtained

with the flat plate.

The simulated airfoil specimen is documented

Figs. 4 and

Figure 4 illustrates the results

the

reverse engineering process and Fig.

shows the

resultant silicon nitride version, which was an excellent

geometrical match for the original casting.

Mach 0.3 jet fuel burner rig testing was

conducted on superalloy, uncoated silicon nitride and

mullite-coated silicon nitride specimens. A typical test

condition is illustrated in the photograph appearing

Fig. 6. Quantitatively results indicate that cooling air

produced the expected comparable reductions

surface

temperature, up to 1 30°C for flowrates of approximately

1.5 lit/s.

CONCLUSIONS

The fabrication of internally cooled monolithic

silicon nitride aerospace components has been

demonstrated. Gelcasting coupled with rapid

prototyping technology allows for the production

these components which range from thin-walled shapes

with cooling channels to simpler hollow shapes. The

internal geometry of these parts is made possible by wax

cores produced by the Sanders Prototyping process,

which can be dissolved from wet gelcast parts with

ethanol. This technique brings the same flexibility

shape forming associated with metal casting to the

realm of ceramics processing. A hollow silicon nitride

airfoil produced with this process has been tested with

cooling air

a Mach 0.3 burner rig. Cooling air

reduced the surface temperature up to 130°C below the

uncooled condition

1 185OC.

464

ACKNOWLEDGEMENTS

Liu, CWRU, was of enormous help in general

ceramic processing and rapid prototyping. R. Babuder,

CWRU/NASA, was similarly invaluable in his help

with binder burnout and firing. Mike Cuy, NASA,

assisted with the burner rig testing. The work was

funded under NASNCWRU Cooperative Agreement on

Ceramic Processing Grant

NAS3-404.

REFERENCES

Sims, “A History of Superalloys for

Superalloy Metallurgists,” pg. 399 in

Superalloys

1984,

edited by M. Gell, C.

Kortovich, R. H.

Bricknell, W. B. Kent, and

Radavich, AIME,

1984.

Pollinger, J.P., Progress in fabrication of silicon

nitride structural components for turbomachinery

applications. ASME Paper 96-GT-347.

Tsuchiya, T., Furuse, Y., Yoshino,

S.,

Chikami,

R.,

Tsukagoshi,

K.,

Mori, M., Development of air-

cooled ceramic nozzles for a power generating gas

turbine. ASME Paper 95-GT-105.

K. N. Lee and

A. Miller, “Development and

Environmental Durability of Mullite and

MulliteNSZ Dual Layer Coatings for Sic and

Si3N4 Ceramics,”

Surface and Coatings

Technology,

86-87

[

1-31 142-8 (1996).

Best, J. D. Cawley,

Bhatt, and D. Fox,

“Fabrication and Testing of Internally Cooled

Monolithic Silicon Nitride Aerospace

Components,

Am. Ceram. Soc.,

submitted.

0.0.

Omatete, M.

Janney,

D. Nunn,

“Gelcasting: from Laboratory Development toward

Industrial Production.

Journal of the European

Ceramic Society

(1997) 407-413.

Jamalabad,

V.R.,

Whalen, P.J., Pollinger, J.,

“Gelcast molding with rapid prototyped fugitive

molds,”

pp. 71-78 in

Proceedings of the Solid

Freeform Fabrication Symposium,

ed.

D.L.

Bourell, J.J. Beaman, H.L. Marcus,

Crawford, J. W. Barlow. The University of Texas

at Austin, Austin, Texas, 1996.

465

Figure

shows the

CAD

rendition of the assembly of the thin-walled internally cooled cylinder mold system.

The individual parts can be seen in the center overview from right to left: inner gating, inner lining, core half,

outer lining half, outer gating half. The design feature are highlighted numerically:

registration features,

slurry casting tube,

bottom filling tray,

flow path slats in core,

overflow space,

gate removal is

accomplished by cutting along this plane,

uppermost of three slots for dissolution

outer lining, and

leachable ethanol path for core dissolution (square cutouts are obsolete flow paths).

466

Figure

shows the processing steps of the soluble core gelcasting process: a) filling the mold, b)

dissolving ProtoBuild mold components in ethanol, and c) the final green part separated from, but

still

within

the

enoxv/

acrvlate

mold.

Figure

shows various views of the thin-walled cooled cylinder: a) untrimmed green state, b) trimmed

green state, and c) fired component. Diameter of coin is

19.05

mm.

Figure

Reverse engineering of a internally cooled model airfoil test specimen. a) Deduced CAD

file from measurements on original casting.

and c) Two views of the original casting and a

stereolithography model made

for

verification purposes.

467

Figure

Two views of the green hollow simulated airfoil.

Diameter of coin is

19.05

mm.

Figure

Silicon nitride airfoil in Mach

0.3

jet

fuel burner flame with.3 lids of cooling air.

Superalloy casting and mullite-coated airfoils

also were tested.

468

PREPARATION AND HIGH TEMPERATURE STRENGTH

G&A1209/MgO COMPOSITES

M. Shimada,

Sakamoto and

Yamane

Institute

for

Advanced Materials Processing, Tohoku University

Sendai 980-8577, Japan

ABSTRACT

Gd4Al~OdMg0 composite (MgO:

~01%) was

prepared from the mixed powders of Gd203,

A1203

and

MgO at

1600

for

h in air. The relative density of

the composite was 99.1%. The X-ray powder diffraction

pattern of the composite was identified by the mixture of

Gd4A1209 and MgO.

other peaks were observed. The

composite had a uniform microstructure with grain sizes

2 pm for Gd4AlZOYand of about 4 pm for MgO.

bending strength of 400 MPa measured at 1100

cooling from 1400

was

1.6

times higher than that

measured at

1100

on heating. This could be explai-

ned by the transformation toughening mechanism of the

high temperature phase Gd4Al2aat

110

cooling.

INTRODUCTION

Recently, there has been a great effort to overcome

the intrinsically brittle nature of ceramics in order to

apply ceramic materials to engineering purposes. Since

the fracture toughness of ceramic materials is not high

enough, their use under high stress is limited. It is highly

desirable to improve the mechanical properties of ce-

ramic materials. Transformation toughening is one

effective approach

improving fracture toughness and

fracture strength of brittle ceramics. Transformation

toughening requires the stress-induced martensitic trans-

formation of particles, as well

known

in tetragonal zir-

conia ceramics (1). The main purpose of our study

find new oxide ceramics with stress induced phase trans-

formation.

Rare-earth aluminate,

%A1209

rare-earth ele-

ment) is monoclinic with space group P2Jc at room

temperature

(2).

Yamane et.

Al.

(3) reported that

YdA11Q showed a reversible thermal phase transforma-

tion from low temperature monoclinic

high tempera-

ture monoclinic phases at 1377

"C.

The results of high-

temperature X-ray powder diffraction showed that the

unit cell volume of the high temperature phase Y4A1209

measured at 1400

on cooling from 1450

was

0.5%

smaller than that of the low-temperature phase at the

same temperature of 1400

on heating from room

temperature (3). This volume change was in accordance

with the results of dilatometry. The phase transformation

of YdA1109 was concluded

be diffusionless by com-

parison

the crystal structures of the high- and low-

temperature phases determined by high-temperature

neutron diffraction (4). Gd~A1209 showed a reversible

thermal phase transformation and thermal hysteresis at

around

1100

is about 200

lower than that of

y4AlZa.

The present study attempts to stabilize the high-

temperature phase of Gd4A1209 by preparing composites

with MgO and examines the effect of phase transformation

toughening by measuring the high temperature fracture

strength.

EXPERIMENTAL

Powders of Gd203,

AlZ03

and MgO were used as

starting materials. These powders were weighed in the

appropriate proportion

G&A1209 and

vol%

Gd4Al209/85 vol% MgO and mixed in a plastic vessel

for 24 h by a wet ball milling with ethyl alcohol and

alumina balls. After drying at

for

h in air, the

mixed powders were uniaxially pressed at 30 MPa, and

then isostatically pressed at 200 MPa

form pellets

(5x30~50 mm). The compact pellets were sintered at

1600

for

h in air. The sintered pellets were cut into

rectangular coupons 2x4~20

mm.

The crystal structure

was studied by X-ray diffraction

(XRD).

The mic-

rostructure was observed by scanning electron micros-

copy (SEM). The thermal behaviour of Gd4A1109 and

Gd4AlZO&ig0 composites was characterized from

room temperature to 1500

by high temperature

XRD

and thermal mechanical analysis (TMA). The bending

strength of Gd4A120&lg0 composites was measured

from room temperature

1400

by a three-point

bending test with a crosshead sped of

0.5

mm/min and a

span length of

mm.

RESULTS

AND

DISCUSSION

The X-ray powder diffraction pattern of

Gd4A120&ig0 composite

(85

vol% MgO) was identi-

fied by the mixtures of Gd4A12OY and MgO phases.

other peaks were observed. The Scanning electron mi-

crograph of polished surface of composite is shown in

Fig.

this figure, with area is Gd4Al2a and black

area

MgO. As seen in this figure, the composite shows

a uniform microstructure with grain sizes of

2 pm

for Gd4A12O9 and of about

for MgO. The rela-

tive density of composite was

99.l%,

which indicates

that composite was well densified by heating at

1600

for

in air.

The results of high temperature XRD showed that the

lattice parameters

Gd4Al2O9shrank

0.2

0.3%

along

469

the c axis at the phase transformation fiom the low-

temperature phase. Since the

XRD

patterns of high-

temperature phase was similar with those of low-

temperature phase, it is considered that the crystal struc-

ture of high-temperature phase belongs to a monoclinic

structure. Anisotropy of thermal expansion was ob-

served below the phase transition temperature. The av-

erage thermal expansion of the

axis was about twice as

large as that of the

axis.

Fig 1 Polished surface of G&A120&lg0 composite

Temperature dependence of linear thermal expansion

MgO, Gd4A1209/Mg0 composite is shown in Fig. 2.

large thermal hysteresis was found in Gd4A1209. From

the results of linear thermal expansion of Gd4A1209, the

volume contraction started at 1100 "C and finished at

1230 "C on heating process, and on cooling process the

volume expansion started at 1000

and finished at

880

"C. The thermal linear expansion

the composite was

between those

G&A1209 and MgO. No thermal hys-

teresis like Gd4A1209 was found in composites

Gd4A1209/Mg0 due to the small amount of Gd4A1209

content.

500

1000

1500

Temperature

('C)

Fig. 2 Temperature dependence of linear thermal

Expansion of MgO, Gd4A1209 and

Gd.+41209/

MgO composite

500

400

300

200

loot

200

400

600

800

1000

1200

1400

Temperature

("C)

Fig. 3 Temperature dependence of bending strength

of Gd4A120&lg0 composite

The temperature dependence of bending strength

measured for the Gd4A1209/Mg0 composites is shown

in Fig. 3. The bending strengths were 190 MPa at room-

temperature and about 280 MPa at 700 "C on both heat-

ing and cooling processes. At 1100 "C, a bending

strength of 230 MPa was measured on heating, and a

higher strength of 400 MPa was measured on cooling.

The bending strength measured at

100 "C on coo-

ling was 1.6 times as much as the bending strength

measured at the same temperature

heating. This im-

provement could be explained by the transformation

toughening mechanism of the stabilized high-

temperature phase

G&A1209 at this temperature.

ACKNOWLEDGMENTS

This work is partly supported by NED0 under the

Synergy Ceramics Project of The Industrial Science and

Technology Frontier Program promoted by AIST, MITI,

Japan.

REFERENCES

(1) R. C. Garvie, R.

Hannink and R. T. Pascoe, Ceramic

Steel?, Nature 258 (1975) 703-704

(2)

J. W.

Reed and A. B. Chase, The Unit Cell and Space

Group of 2Y203.A1203, Acta Crystallogr. 15 (1962)

812

H. Yamane, M. Omori, A.

Okubo

and

Hirai, High-

Temperature Phase Transition of Y4A120g, J. Amer.

Ceram. SOC. 76 (1993) 2382-2384

H.Yamane,

Shimada and B. A. Hunter, High-

Temperature Neutron Diffraction Study of Y4A1209,

J. Solid State Chem. 141 (1998) 466-474

470

CHARACTERIZATION

IN SITU SIC-BN COMPOSITES

Guo-Jun Zhang* and Tatsuki Ohji

Synergy Ceramics Laboratory, National Industrial Research Institute

Nagoya,

Nagoya, Aichi

463-8687,

Japan

ABSTRACT

Sic-BN composites were prepared by in

situ reaction using Si,N4, B4C and

reactants. Adding Sic powder into the

reactants controlled the content of BN in the

composite. For comparison, monolithic Sic

and SIC-BN composites with the same phase

compositions were produced by conventional

process. The in situ process was advantageous

for obtaining better composites with fine grain

size, homogeneously and isotropically

distributed microstructures. The elastic

modulus decreased markedly with increasing

BN content and obeyed well both to the

exponential and polynomial rules. The bending

strength increased slightly with increasing BN

content up to 15 vol%, and then gradually

decreased; the decrease was accelerated above

~01%. The results were explained by a

percolation network formation of the BN

phase. The fracture toughness decreased with

increasing BN content. It can be excepted that

these in situ Sic-BN composites have

excellent mechanical strain tolerance and

thermal shock resistance.

INTRODUCTION

Silicon carbide (Sic) is a potential candi-

date material for structural applications at high

temperatures. In many cases of application

ceramic components are assembled with other

components made of metals to form a hybrid

structural part. The failure of such hybrid part

often occurs in the ceramic components at the

connecting section due to the mismatch of

thermal expansion coeffi-cients and elastic

strains in different materials. To avoid these

fractures, it is required that the elastic modulus

of Sic ceramic material is reduced. BN

ceramics, which shows very low elastic

modulus among ceramic materials, have been

widely used as the second phase to reduce the

elastic modulus and improve the thermal shock

resistance of monolithic ceramic materials,

such as in Si3N4-BN [l-31 and SIC-BN [4-61

composites. Unfortunately, the bending

strength of BN-containing composites

generally decreased with increasing the BN

content compared to the monolithic ceramics.

One of the main reasons is that the poor

sinteribility of BN affects densification

behavior of BN composite. The other is that

because BN is weak, BN agglomerates or large

BN particles/platelets may act as fracture flaws.

This means that BN dispersoids with fine

particle size and homogeneous distribution are

the key factors to obtain high strength compo-

sites. Reactions are also used to synthesize

BN-based composites. For example, Coblenz

and Lewis

[l]

used the in situ reactions of

Si3N4

B203, A1N

B203 or Si3N4

A1N

B,O, to produce SO2-BN, A1203-BN and

mullite-BN composites, respectively.

Kusunose et a1

[3]

prepared Si3N4/BN

nanocomposites by in situ formation of nano-

sized h-BN coated on a-Si3N4.particles with a

mixture of boric acid and urea in hydrogen gas.

In addition, in the Si-B-C-N system, Riedel et

[7,8]

developed a silicoboron carbonitride

ceramic material, which is stable to 2000

OC,

from polymeric boron-containing

polysilylcarbidi-imide precursor.

By in situ process, one can generally

obtain composites with finer and more homo-

geneous microstructures, higher chemical and

microstructural stability at high temperature,

and better mechanical properties than the

conventionally processed ones. Recently the

authors proposed the following reaction

synthesize SIC-BN composite:

The weight percents of Sic and BN in the

obtained composite according to reaction (1)

are 54.79

and 45.21

respectively, and the

volume percents are 46.29

and 53.71

respectively. By adding Sic powder into the

reactants, the content of Sic in the composite

can be adjusted. Here the preparation of the in

situ

composites will be firstly introduced, then

the mechanical properties of the obtained

composites will be characterized.

Si3N4

B4C

3SiC

4BN..

. .

EXPERIMENTAL PROCEDURE

The starting powders were Si3N4 (E-10

grade, Ube Industries, Japan), B4C (F1 grade,