—

5 t
fimfirfitfi
.5 .. ...mw.: .. 3:56.52. .
z .fi..Mnrfl.. av flak
Vޤ . .

. .
y . . ”pa—j}
z... - . . . . . . . 3;...
p . u “7 . . . . , . . , . ’.v¢U.,lnI.
l .. . . . . .. . e:

I» . v

.n. 2....”
1 x... t
‘1 i .
i.

 

. . o.
(as)! n...
ma»..- 3.5.9111.

$3.... . , , . . .

. . ..
.. . Inks” .4]

 

 

Thfias

Imnmllllllllllilll‘ilfimmmi

3 1293 01770 0687

 

LIBRARY
Michigan State
§ University I

i

O

 

 

This is to certify that the

thesis entitled

A Preliminary Study of Hot Corrosion by Sodium Chloride on

A Titanium-Aluminum-Chromium-Niobium Alloy

presented by

Gregary Thomas Dowling

has been accepted towards fulfillment
of the requirements for

Master's degree in Materials Science

 

 

(flumfggufib

Major professor

 

Date 7/WM /99y
O

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

PLACE IN RETURN BOX to remove this chedtout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1M ammonium-m4

#— ’ ’ " —— ’—_‘*'-—” _—*‘ ‘—

A PRELIMINARY STUDY OF HOT CORROSION BY SODIUM CHLORIDE ON
A TITANIUM-ALUMINUM-CHROMIUM—NIOBIUM ALLOY
By

Gregary Thomas Dowling

A THESIS

Submitted to
Michigan State University
In partial fulfillment of the requirements
For the degree of

MASTER OF SCIENCE
Department of Materials Science and Mechanics

1999

ABSTRACT
A PRELIMINARY STUDY OF HOT CORROSION BY SODIUM CHLORIDE ON
A TITANIUM-ALUMINUM-CHROMIUM—NIOBIUM ALLOY
By

Gregary Thomas Dowling

In an effort to examine the potential of salt to damage a TiAl alloy (nominal
composition in at%, Ti-48Al-2Cr-2Nb), in conditions relevant to heat engines, the alloy
was subjected to hot corrosion by application of simulated sea salt by aqueous means.
Bare ground specimens were sprayed with seawater solution, dried, and exposed in an air
furnace at 760°C for times between 0.1 and 20 ks. Obvious attack occurred in as little as
0.3ks in regions where salt was deposited. In such regions, oxide grth was accelerated
and non-adhesive. Regions below accelerated growth were pitted heterogeneously in a
manner that indicated the alpha-2 phase was preferentially attacked, particularly in
lamellar regions. Mechanical strength was measured by four-point bending testing. A
reduction of fracture strength was observed only for specimens with the longest exposure.
Pre-oxidized specimens were exposed to a salt fog and then a heat cycle. Accelerated
oxide grth was observed afier numerous salt application and heat cycles for short pre-
oxidation times (3.6 ks) but not for longer pre-oxidation times (360 ks). Intermittent
weight gain was measured for short pre-oxidation times (3.6 ks). EDS was used to
analyze elemental concentrations on surfaces of all specimens. DTA analysis of salt/TiAl
mixtures in N2 indicate that endothermic reactions occur around 700°C. X-ray diffraction

was used to determine the products of the above reactions.

To My Wife

Deborah

iii

ACKNOWLEDGEMENTS

I would like to thank my wife, Deborah, and the rest of my family members for
their love and support. I would also like to thank my advisor, Dr. T.R. Bieler, and my
committee members, Dr. M. J. Crimp and Dr. J. Lucas, for advice and editorial
comments, and Dr. R. Schalek for ESEM work.

This study was partially supported by Air Force Office of Scientific Research

PRET Program F 49620-95-1 -0359.

iv

Table of Contents

CHAPTERllNTRODUCTION

CHAPTERZ : REVIEW OF RELATED LITERATURE... ..

..viii

.IX

.1

..4

2.1 GENERAL CHARACTERISTICS OF TITANIUM ALUMINIDES ................. 4

2.1.1 General AspectsofAluminides.....................................
2.1.2 Characteristics ofTigAlBased Alloys.......................... . .. ..
2.1.3 Characteristics OfTiAlBased Alloys...
2.1.4Applications............................................

2.2 GENERAL CORROSION OF METALLIC MATERIALS...

2..21TypesofCorrosion.... ..
2. 2. 2 Environmentally Assisted COrIOsion..
2. 2. 3 Mechanical effects of Corrosion...

2.3 OXIDATION OF TITANIUM ALUMINIDES... ..

2.3.1 Generaleidation ofTiAlAlloys.....
2.3.2 ThermodynamicConsiderations.....................................................
2.3.3 Elementallnfluences
2.3.4Prevention/Protection................. .

2.4 HOTCORROSION OF MATERIALS.....................

2 4.1 Fundamentals of Hot Corrosion“ .
2. 4. 2 Hot Corrosion of Nickel Based Super Alloys
2. 4 3 Hot Corrosion Due to Sodium Chloride...

2.5 STRATEGY FOR INVESTIGATING HOT CORROSION... .. .. ..

2.5.1 Hot Corrosion Summary
2.5.2 Research Approach..................................

....xo'O

'oo'xi'mA

\O

11

11

12
l6
17

18

........18
.. ...21
2.4.4 LaboratoryTesting..........

.24

.24

.24
.25

CHAPTER3:EXPERMENTALPROCEDURE..........................................
3.1.1 Salt Application, F our—Point Bend and Pre-oxidation Specimens .........

3.2 SALTAPPLICATION....................................
3.3 FOUR-POINT BEND.............. . .. .....
3.4 PRE-OXIDATION
3.6 DIFFERENTIAL THERMAL ANALYSIS......
37X-RAYDIFFRACTION
4.1 INTRODUCTION........................................
4.2 SALTDISTRIBUTION...........................................

4.3 SALT ATTACK ON BARE GROUND SPECIMENS... ..

28

28

....... 28

28

29

.32

.32

.33

34

.35

.35

..35

.38

4.4 EFFECT OF SALT ATTACK ON FOUR-POINT BENDING STRENGTH ...... 60

4.5 SALT APPLICATION ON PRE-OXIDIZED SPECIMENS... ..
4.5.1 Interrupted WeightGain............................

4.6 REACTIONS ANDPRODUCTS
CHAPTERS:ANALYSISANDDISCUSSION...........................................

5.3 STRENGTH OF EXPOSED BARE GROUND SPECIMENS ...............

vi

.64

.64

.69

.69

.69

.74

74

.75

....... 77

CHAPTER 6 : CONCLUSIONS AND RECOMMENDATIONS... ..

6.1 CONCLUSIONS............................... 81

6.2 RECOMMENDATIONS FOR FURTURE RESEARCH... ..

vii

5.4 EFFECTS OF PRE-OXIDATION... .. ....78

...81

...81

List of Tables

Table 1. Selected Properties OfTi3Al and TiAl alloys
Table 2. Thermodynamic Data for Free Energy Change Calculations ..................
Table 3. Four-point bend results on specimens with salt application. ..
Table 4. Four-point bend results on specimens without salt application
Table 5. Products identified in 90wt% TiAl filings: 10wt% salt DTA powder... . .. .. ..

Table 6. Products identified in 10wt% TiAl filings:90wt% salt DTA powder... . . ....

viii

.6

........ 15

61

.62

.71

.71

List of Figures

Figure l. TiAlBinary Phase Diagram
Figure2.EllinghamDiagramofSomeCommonOxides........................
Figure 3.Electrochemical Model
Figure4.Nebulizer Set-upto Apply SaltFogto Specimens..............................
FigureS.TestFixtureforFour-PointBending............................
Figure 6. Naturallyexposedsaltdistn'bution.......................... .
Figure 7.Nebulizersaltdistribution................... ..
Figure 8. Saltparticle size distribution...............................
Figure 9.RatioofEDSintensity for(a)Na,(b)Cl
Figure 10.RatioofEDSintensityfor(a)Al,(b)O...........................................
Figure11.RatioofEDSintensityfor(a)Cr,(b)Nb..........................................
Figure 12. EDS scan ofelectropolished specimen — “Base Metal”... ..
Figure 13. 0.15ks specimen-generalarea

Figure 14. 0.15ks specimen-salt clumpnearapit..............................

Figure 15. EDS scan ofa surface pit on 0.15ks Specimen (Figure 14)...
Figure 16. 0.36ks specimen -acce1erated oxide growth...... ..

Figure 17. 0.36ks specimen-exfoliated salt deposit...

..13

...20

....30

...31

36

.37

..38

.39

.40

.41

.42

.44

.45

.46

47

.48

Figure 18. EDS scan of the metal at edge of exfoliated salt deposit (region 3) ............ .49

Figure 19. EDS scan of the metal/salt interface by exfoliated salt deposit (region 2). . .. .50

Figure 20. 0.36ks specimen - light and heavy salt deposits and resulting oxide ............ 51

Figure 21. Surface morphology of specimen exposed for 1.08ks - general oxide .......... 52

ix

Figure 22.
Figure 23.
Figure 24.
Figure 25.
Figure 26.
Figure 27.
Figure 28.
Figure 29.
Figure 30.
Figure 31.
Figure 32.
Figure 33.
Figure 34.
Figure 35.

Figure 36.

Figure 37

Figure 38.
Figure 39.

Figure 40.

Surface morphology of specimen exposed for 1.08ks - fracture in oxide... ..
Surface morphology of specimen exposed for 1.08ks - damaged area...
EDS scan of an attacked lamella region (or; phase) in Figure 23. . ..
Specimen exposed for 19.44ks - exfoliated oxide...

Specimen exposed for 19.44ks - damaged area under oxide... ....
EDS scan of an attacked lamella region (or; phase) of Figure 25...

Effect of salt exposure on fracture strength .............................................

Fracture strength of specimens that had no slippage and not at a pin .......

0.36ks specimen fracture surface-oxidizedpit..............................
19.44ks specimen fracture surface—remnant ofspalled oxide............
Pre—oxidation specimens after 1St Cycle-3.6ks specimen................
Pre-oxidation specimens after 10 Cycles-3.6ks specimen.............
Pre-oxidation specimens after 1St Cycle-360ks specimen......................
Plottedweightgainresults..........................
Weightgainphotographs................
DTAofcastingfilingsandsalt.....................................................
DTAofmechanicallyalloyedTIA] andsalt
XRD scan Of90wt% TiAl filings:10wt% salt DTA powder...

XRD scan of10wt% TiAl filings:90wt% salt DTA powder.

.53

54

55

57

.58

59

...62

....62

.63

63

.65

66

.67

68

.68

..70

.70

72

73

CHAPTER 1

INTRODUCTION

Titanium-aluminides are candidate materials for high temperature engine
applications. These alloys possess low density, high melting temperatures, good elevated
temperature strength and modulus retention, environmental resistance, and good creep
properties [1 -4].

High temperature components suitable for titanium aluminides are jet engine
turbine blades [5, 6], turbocharger turbine components [7], pistons and valves [8]. The
proposed operating temperature range for jet engine components using titanium
aluminides is from 600°C to 900°C (possibly higher). Turbocharger components are
expected to operate at temperatures of 600°C to 700°C for diesel engines, and 850°C and
above for gasoline engines [7]. Titanium almninides proposed for use in these
applications range in composition from Ti-23 - 24at°/o Al and Ti-46 - 52at% Al (Ti3Al -
(12 phase and TiAl - 7 phase including alloys with combinations of these two phases).
Composites utilizing silicon carbide (SiC) or alumina (A1203) fibers in a titanium
aluminide matrix [9] have also been considered for the above applications.

Titanium aluminides’ resistance to oxidation and hot corrosion has limited
their use at high temperature, and their very low ductility at room temperature and
fracture toughness is also a concern. Oxidation resistance of these alloys increases as the
percentage of aluminum in the alloy is increased [51] and is influenced by additional
elements [10]. Oxidation of titanium aluminides in pure oxygen leads to formation of a

continuous alumina (y—AIZO3) scale which provides protection to about 1000°C, whereas

oxidation in air leads to formation of a scale consisting of alumina and titania (rutile —
TiOz) and limits protection to about 800°C [33].

There has been very little study about hot corrosion of titanium aluminides. Hot
corrosion of metal alloys is defined as the phenomenon of accelerated high-temperature
oxidation in a gaseous atmosphere of a material whose surface is coated with a thin film
of fused salt. Early studies of this phenomenon were performed to examine the
accelerated corrosion of metals in aircraft gas turbine engines. This corrosion was caused
by a thin layer of sodium sulfate (NazSO4) that form on the airfoil material from a
reaction between sulfur in the fuel and NaCl (ingested during combustion in the form of
seawater aerosol) [l 1]. Sodium chloride can also act directly with engine components
when condensation of seawater spray (or mist) lands on engine components while
engines are not running at airports near seacoasts. Studies [45,46] have shown that hot
corrosion of nickel-based superalloys induces accelerated oxidation and can cause
catastrophic oxidation. However, coating nickel-based superalloys with a layer that
forms a protective oxide increases the creep rate, reduced rupture life and rupture
ductility [33, 39, 41].

Since titanium aluminides are targeted for internal and external applications in
turbine engines, hot corrosion should be a concern. Titanium and titanium alloys can be
pitted and experience stress corrosion cracking (SCC) when exposed to environments
containing chloride salts. Handling and cleaning of parts can also lead to cracking caused
by hot corrosion due to salt residues left from fingerprints and chlorinated cleaning
solutions [13]. Yao and Merek [15] observed increased weight gain, pitting and grain

boundary attack after subjecting a TiAl-based alloy to hot corrosion by NaCl.

This study will be concerned with hot corrosion by NaCl on the alloy Ti-48Al-
2Cr-2Nb. Since little research to date has focused on hot corrosion as it applies to
titanium aluminides, and for the sake of academic convenience, the research path
followed had many short branches which had to be evaluated and assessed as neatly as
possible. The study examined many exploratory routes that have provided guidance for
further research on the hot corrosion. The need to investigate hot corrosion reactions and
their products required use of several analytical tools (i.e., EDS, XRD, DTA, etc.)

The motivation used to provide a focus for this study was to consider the scenario
of short inter-island flights having limited duration where salt water vapor has an
opportunity to condense on engine parts during down time. From this scenario, exposure
times were derived and used in the experimental procedure which includes: analysis of
changes in oxide growth, determining if hot corrosion affected the alloy’s strength, and

investigating if a pre-oxidation layer is protective.

CHAPTER 2

REVIEW OF RELATED LITERATURE

2.1 GENERAL CHARACTERISTICS OF TITANIUM ALUMINIDES
2.1.1 General Aspects of Aluminides

Aluminides are important classes of alloys that have gained much attention
because of their unique properties. They possess excellent elevated temperature strength,
resistance to oxidation and corrosion, have relatively low densities, and high melting
points. These properties result from the fact that aluminides (generally) form ordered
intermetallic alloys. The use of most intermetallic alloys are limited by their fracture
resistance (although microalloying can enhance it), creep resistance, ambient temperature
ductility and fabricability [18]. Some of the most important aluminides are Fe3Al (D03),
FeAl (BZ), NiAl (BZ), NI3AI (L12), TiAl (L10) and Ti3Al (D019).

Iron aluminides, Fe3Al and FeAl, have excellent oxidation and corrosion
resistance when exposed to hostile environments. Their main drawback is low fracture
resistance especially in ambient air where embrittlement occurs due to water vapor.
However, when tested in an oxygen environment their ductility is quite high. These
aluminides also offer low material costs.

The nickel aluminide Ni3Al possess good oxidation resistance and excellent
strength, in fact, it is the most important strengthening constituents in nickel-based
supperalloys. Ni3Al has low fracture resistance due to embrittlement by water vapor.
The nickel aluminide NiAl has excellent oxidation resistance at elevated temperatures

and has been used as a coating for hot components in hostile environments. NiAl,

however, lacks fracture resistance at ambient temperatures, as well as strength and creep
resistance at elevated temperatures.

The titanium aluminides, TiAl and Ti3Al, have low densities, high retention of
strength, and acceptable creep resistance at elevated temperatures. The Ti3Al alloys have
low fracture resistance, structural instabilities, and cracking problems when exposed to
hostile environments under cyclic loads. TiAl alloys possess greater strength at elevated
temperatures, better oxidation resistance and a lower density than the Ti 3A1 alloys.

2.1.2 Characteristics of TI3AI Based Alloys

Titanium aluminides based on Ti3Al have the D019 (Strukturbericht
symbol) or hP8 (Pearson symbol) crystal structure, which is an ordered hexagonal
structure. Ti3Al is commonly referred to as the alpha-2 phase on the binary TiAl binary
phase diagram, Figure l, and has a wide composition range from 22 at% to 39 at%Al.
The most common Ti3Al-base alloys contain from 8 to 18 at%Nb which improves a
majority of mechanical properties; these materials form two phase alloys (alpha-2 +
beta/82). Micro-alloying with Mo, Ta and Cr can provide strengthening, oxidation
resistance is enhanced by Mo and Ta, and creep resistance by Mo. The compositions of a
few of the more important TigAl-base alloys (from an engineering standpoint) are Ti-
24Al-11Nb, Ti-25Al-10Nb-3V-1Mo, Ti-25Al-17Nb, Ti-25Al-l7Nb-1Mo and Ti-24.5Al-
8Nb-2Ta-2Mo. Table 1 list some important properties of Ti3Al-base alloys [1]. Ti3Al has
been shown to exhibit limited ductility up to 900°C with an apparent ductile-brittle
transition temperature occurring above 600°C. Fracture is entirely by cleavage up to
600°C, above 600°C there are signs of plasticity but the main mode of fracture is still

cleavage up to 900°C [2].

Atomic Percent Aluminum

 

 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1--.. 1 1‘0 2} 33 49 5‘0 60 lo 39 9‘0 100
1700 . L
670°C
1500. \ L
1500-
1400" (BT1) !
5 130°. ,_- . :0} ifs‘ri
4.4 1200- I, TiAl '0’}..- \j
o I Soot.“
1100- ,' I ,0. :3: i
- .' ' t
1000- T13R25: f. if
-= 82°C .0 : I
°° (“it i i i: +— TiA13 - '
s s s
100- I 5 s 665:sz ‘ 60 . 5°C
600‘ i 5 II ‘rm —=.---
: 1 a 2 2 arms (All)
500 T . .1 j .‘ we” . , . 1"
0 10 20 30 40 50 so 70 so 90 1150
Ti Weight Percent Aluminum A1
Figure l. Ti-Al Binary Phase Diagram [19]
'_l‘_able 1. Selected Properties of Ti3A1 grid TiAl alloys [1, 3, 20]
Property TisAl-base alloys TiAl-base alloys
Structure D019 Ll0
Density (g/cm3) 4.1 — 4.7 3.7 — 3.9
Modulus (GPa) 110 - 145 160 — 180
Yield Strength (MPa) 700 - 990 400 - 650
Tensile Strength (War) 800 — 1140 450 - 800
Ductility (% at RT) 2 - 10 l — 4
Ductility (% at HT) 10 - 20 10 — 60
Fracture Toughness (MPamm) 13 — 3O 10 - 20
CHE) Limit (°C) 760 1000
Thermal Expansion (lO°/°C) 10 12
Oxidation (°C) 650 900

2.1.3 Characteristics of TiAl Based Alloys

Titanium aluminides based on TiAl have the Llo (Strukturbericht symbol) or tP4
(Pearson symbol) crystal structure, which is slightly tetragonal distortion of face centered
cubic structure with alternating layers of Ti and Al along the c axis. TiAl is commonly
referred to as the gamma phase on the TiAl binary phase diagram, Figure 1, and has a
composition range from 44 at% to 52 at% Aluminum. TiAl-based alloys can be divided
into single phase (7) alloys and two-phase (y + a2) alloys. The (7 + 012)/ 7 phase boundary
composition lies at about 49 at% aluminum at lOOO°C and varies depending on the type
and amount of other alloying elements added [3]. The single-phase alloys contain from
50 at% to 52 at% aluminum and at least one element such as W, Nb, or Ta [4]. The two-
phase alloys contain from 44 at.%-49at.% aluminum and also two or more alloying
additions from the group V, Mn, Cr, Nb, Ta, W, Mo, Si, C, B, N, P, Se, Te, Ni, Mo and
Fe. Small additions of Nb, Mo, T182 and Ta can improve strength. Oxidation and creep
resistance is enhanced by additions of Nb, Si, Mo and W [22], and Cr, V, Mn and Mo
can be added to improve ductility.

Several studies [3, 5, 8, 21] have compared the elevated-temperature properties of
the single phase (y) and the two-phase (y + (12) alloys. The lamellar structure of the two-
phase (7 + a2) alloys gives better ductility and creep resistance than the single phase (7)
alloys. A finer lamellar structure within the duplex microstructure has been shown to
increase room temperature ductility. Table 1 list some important properties of TiAl-base

alloys [1].

2.1.4 Applications

Titanium aluminides have been studied for use in aerospace, automotive, and
power generation applications. Titanium aluminide matrix composites are also being
considered for these applications because they have properties which surpass monolithic
titanium aluminide [6, 9]

Titanium-based alloys have played an important role in the aerospace industry
since the introduction of J57 Turbojet rotors in 1957 [3]. Aerospace research on titanium
aluminides has increased due to Government programs. The Integrated High
Performance Turbine Engine Technology Initiative (IHPTET) focused on developing
TiAl-based engine components. Commercial and military turbine engine manufacturers
have proposed using TiAl for turbine engine components [5, 20]. The National
Aerospace Plane program (NASP) studied monolithic TiAl-based alloys and/or
composites for a high-temperature, lightweight honeycomb or truss core airframe panel
structure.

Not only aerospace applications are of interest, but also automotive applications
such as turbocharger motors, pistons, and valves stimulate studies on TiAl [8]. These
automotive components have been produced from ceramic materials such as Si3N4 or
SiC, which have lower densities than their metallic counterparts. Low weight is
especially important for turbo rotors in that it reduces their inertia and hence minimizes
turbo-lag [7]. Superalloys were previously used but their density of about 8g/cm3 is
much higher than the lightweight ceramics that have a density of about 3.2g/cm3. TiAl-

based alloys are attractive for this application because they have a density comparable to

ceramics, easier fabrication methods compared to ceramics, and better fracture toughness
than ceramics.

2.2 GENERAL CORROSION OF METALLIC MATERIALS

2.2.1 Types of Corrosion

Corrosion can be classified as wet or dry. Wet corrosion involves a liquid, i.e.,
aqueous solutions or electrolytes. Dry corrosion occurs without the presence of a liquid
and usually at high temperatures (above the dew point of the environment). An example
of wet corrosion is corrosion of steel by water and an example of dry corrosion is the
oxidation of titanium at 500 °C.

The forms that corrosion manifests itself are important for detection and
prevention. There are eight forms of corrosion which may be unique to a certain situation
but they are related to one another, these are: (1) uniform or general attack; (2) galvanic,
or two metal corrosion; (3) crevice corrosion; (4) pitting; (5) intergranular; (6) selective
leaching, or parting; (7) erosion corrosion; (8) stress corrosion [23].

2.2.2 Environmentally Assisted Corrosion

Environmental factors such as temperature, concentration and velocity of the
corrosive medium can cause and/or accelerate corrosion. The purity of a metal (or alloy
composition) will also affect its corrosion resistance. Surface texture (distribution of
grain orientations), grain size and the presence (or lack) of inclusions can influence
corrosion resistance of metals and alloys. Marine environments are especially corrosive
to metallic materials; this is a major concern since about 3/4 of the earth’s surface is

covered by oceans.

To illustrate the effect of metal purity, consider a nickel aluminide [26] when
exposed to solutions of 0.5M NaOH (pH 13.1), 0.5M NaCl (pH 6.5), 0.5M H2S04 (PH
0.7) and HNO3 (pH 0.6) it had electrochemical behavior similar tO nickel which means
nickel aluminide could be substituted for nickel and that purity has little affect in those
environments.

Another example of purity involves a stainless steel pipe used in cooling systems
utilizing seawater. Steel (or iron) will experience pitting and crevice corrosion due to
chlorine (from seawater and hypochlorite added to prevent bio-fouling) and operating
temperatures of the system. An ELI (extra low interstitial) superferritic stainless steel
containing large amounts of Mo and Cr along with small percentages of Ni and Ti will
provides excellent corrosion resistance at temperatures greater than 50°C and good
repassivation up to 50°C [24].

As mentioned above, a materials texture can influence its corrosion resistance. A
study by Gregory and Brokmeier [25] shows that the stress corrosion cracking (SCC)
sensitivity is qualitatively dependent on the loading direction and crystallographic
orientation. This study was performed on a Ti-6Al-4V alloy in aqueous NaCl solution,
and in the presence of a crack.

Variable concentrations (differing pH levels) of the corrosive environment may
alter corrosion resistance also. Given three different solutions of 3.5% NaCl with pH’s of
3, 7 and 10, polarization tests on Ti3Al and Fe3Al show that the breakdown potential is

independent of pH but the corrosion potential becomes more positive with lower pH [27].

10

2.2.3 Mechanical Effect

Mechanical properties such as ductility, fracture strength, yield strength and
toughness can be diminished due to exposure to corrosive environments.

In certain cases the stress and the environment are favorable for hydrogen
embrittlement (HE) to occur. This is especially true for some aluminides. The ductility
was reduced and the fracture mode changed from ductile transgranular to brittle
intergranular during SSC testing of a Ni3Al where HE occurred [28]. Also, Ti3Al was
very susceptible to HE with a cathodic potential beyond — 1000 mVSCE using the slow
strain rate technique (SSRT). This alloy exhibited lower tensile strength, lower ductility,
surface cracking during plastic loading with transgranular and branched cracking [27].

High temperature oxidation of TiAl interrnetallics can cause embrittlement by
dissolved oxygen and presumably a loss of tensile properties. In general, sub-surface (or-
case) embrittlement is more of a problem with az-TigAl based alloys than with v—TiAl
based alloys. McKee and Huang [42] observed a-case formation on v-TiAl alloys that
were oxidation prone such as Ti-48Al-2Cr, after a 850 °C - 900 °C exposure, and with
more resistant alloys such as Ti-48Al-2Cr-2Nb, after a 1000 °C exposure.

2.3 OXIDATION OF TITANIUM ALUMINIDES
2.3.1 General Oxidation of TiAl Alloys

The oxide layer(s) which form during high temperature oxidation of TiAl alloys
vary depending on the gaseous environment surrounding the sample. An oxygen
environment can produce a continuous, protective alumina layer on a binary TiAl alloy

up to lOOO°C [29], however, exposure in air at 1000 °C forms a scale consisting of TiOz

and A1203.

ll

 

TiAl alloys that contain a third or fourth alloying addition always form a mixed
scale of T102 and A1203 regardless if oxidized in air or oxygen. The scale formation
during exposure to O; is the following [10]: at the base metal/scale interface is a layer
consisting of T102 and A1203 with small pores distributed throughout the layer. The outer
layer consists mainly of large TiOz grains, and the region in-between these layers is rich
in A1203 with large voids. The scale is not considered to be continuous because of its
porous nature, but provides a barrier and is considered “protective”. In contrast, a
specimen exposed to an air environment at an elevated temperature, introduces nitrogen
into scale growth. The presence of nitrogen, and temperatures above 800 °C, leads to
formation of intermixed TiN and A1203 on the alloy oxide interface that interferes with
the development of the alumina-rich layer [31], this may account for the development of
a non-protective oxide layer.

2.3.2 Thermodynamic Considerations

In a two-phase (y + a2) TiAl alloy, the oxygen pressure of Al/A1203 is several
orders of magnitudes below that of Ti/TiOz in the TiAl/TiAl3 region, so, A1203 is more
stable. But, moving from the TiAl/TiAl3 (Al-rich) region to the Ti3Al/TiAl (Ti-rich)
region, the oxygen pressures of Al/A1203 and Ti/TiOz increases and decrease,
respectively, allowing T102 to be the dominate oxide [32]. To protect TiAl alloys up to
temperatures in excess of 1200°C it is necessary to develop a continuous A1203 oxide
layer [33]. The challenge is how to suppress the formation of TiOz and allow just the

formation of A1203. Reference to the Ellingham diagram for oxides, Figure 2, shows that

G” for the formation of A1203 is more negative than G” for the formation of T102.

12

_-.
'—'—

O
i

Ac°-ar1np0,, kilojoulos

 

 

M20 ratio
0 CO/CO2 ratio

IT] I T I I
110'“ 10"2 10'” 10'" 10,"7 10" 10"

 

I I I
10" 10" 10"

I

I
10"
10"

10"1

10"

 

10’;

 

10'2

 

-s
10 ‘4

 

-s
10 .4

 

u.
10‘

 

 

—300

-500

10"

 

 

 

 

 

4!.

30: + 0,- l=2/3cn1l3,

 

 

 

 

 

4/5Nb .+ O = 2/

5.1.1 [

 

 

Nb

 

i
+ o, = 1/2Nbl,0

 

IT

 

~700

—eoo

-..,OL

>r‘i/

-1000

-1100

—1200

 

/

 

 

 

 

-\3IYI|:E',

 

 

A

 

 

/
{A
/;

 

 

 

 

1+0;

m+q=mq

 

 

 

 

 

 

 

= T102

 

 

 

I

 

5 I “I
4/3Al + o, = 2/3A1

33

 

 

 

2’

 

 

T'—

M melting point of metal

 

 

'fl melting point of mid]

 

 

10'"

\1

 

10'”

 

 

l l

 

 

Ii

 

 

 

 

 

 

 

 

 

 

 

 

I

 

0

200 400
”43"“)

CO/CO2 ratio

0.0

I Temperatu re°c

1030

1'00

1200
0-0

1400
10"

 

1600
10"

\10'“

- 103

- 1o;

141

10:

to;

 

 

-:
A0,
-2
10,

10:

1—i

 

 

H,IH20 ratio

10“

 

 

Figure 2. Ellingham Diagram of Some Common Oxides [47].

13

However, given an expected minimum service temperature of 650° and an oxygen partial
pressure of 0.21 atm., it is expected that both oxides will form and in fact they do.
Consider the following reactions taking place during the oxidation of a binary
TiAl alloy,
2A1“) + 3/2 0203) = A1203(,) (1)
and
TItS) + 02111) : T1028) (11)
The corresponding free energy changes in joules for the above equations are,
(7" a) = 4676000 + 320T (111)
5" a.) = -910000 + 173T (IV)
and at 650°C, 119°”) = ~1381kJ and A_(_}°(m = -750kJ. Equations (III) and (N) can be
expressed as,
AG° ——~ ~RTan = ..RT1n[aA1203 / (an)2 (p02 )3’2] (V)
and
AG" = -RTan “ ‘RTlnIaTiOZ / (an) (1302)] (V1)
Assuming that the activities for A1203 and T102 are unity and the oxidation occurs in air
where p02 is 0.21 atm., then equations (V) and (VI) can be rewritten respectively as
AG° = -RTln[(aA1)'2 (p02)'3/2] = 2RTln(aA1) + 3/2 RTln(0.21) (VII)
and
AG° = -RTln[(aTi) (p02)]'l = RTln(an) + RTln (0/21) (VIII)
By using these equations, the activities of Al and Ti can be found. These activities will
change depending on the relative amounts of Al and T1 in the alloy and upon addition of

other micro-alloying elements.

14

Niobium is commonly added to TiAl alloys to enhance oxidation resistance and may
form Nb205 [37] or Nb204. The reaction to form each niobium oxide and the corresponding

free energy expressions are given below.

Nb + 02 = 1/2Nb204 (IX)
AG = -798500 — 20.25TlogT + 243T J/mol (X)

4/5Nb + 02 = 2/5Nb205 (XI)
AG = -763600 —— 27TlogT + 263.6T J/mol (XII)

At 650°C, free energy change to form Nb204 is -1259kJ/mol and for Nb205 is -l485kJ/mol.
Thermodynamic calculations between 320 °C and 950 °C show that aluminum and

titanium can react with NaCl and oxygen to produce NaA102 and Na2T103, and other
products [15]. The following reactions are possible below 800 °C,

2NaCl + Al + 404 = 2NaA102 + C12 (XIII)
and

4NaCl + T1 + 302 = 2Na2T10 + 2Cl2 (XIV)
(the melting point of NaCl is 801°C.) Pertinent thermodynamic data that can be used to find
the free energy change for these reactions is listed in Table 2, the last two values are for
constant pressure heat capacity in equation form cp = a + bT.

Table 2.Thermodyn_amic Data for Free Energy Change Calculagons I35]

 

 

 

 

 

 

 

 

 

 

 

 

 

Element 01' AHzgg A8293 a b
Compound kJ/mol J/(molXK)

NaCl -411.2 72.13 12.07

Al 28.29 4.946 2.963
02 20.49 7.168 0.998
NaAlO2 -1 135 70.71 17.52

E 30.69 5.250 2.520
_Na2T103 -1591 121.7 30.03

£12 ~ 8.760 0.270

 

15

 

The change in free energies for Equations (XHI) and (XIV) can be calculated using the

following equation

 

AC
AG 2 [AH298 + I2; AdeT] -- T[ASZ98 +12; T” d’]‘] (XV)

Using the same temperature of 650°C (923K) like the above oxidation example, the free
energy change for reactions (XIII) and (XIV) at 650°C are —1519kJ/mol and -
1782kJ/mol, respectively.

The change in free energy for reactions (1) and (II) at 650°C, AGO) = -1397kJ/mol
and AG“) = -757kJ/mol. The free energies of formation for Nb204 , Nb205, NaAlO2 and
Na2T103 are all lower than those for T102, but close to that for A1203. Given that Nb is a
micro-alloying element in the TiAl alloy and that Na is present in the form of NaCl, from
say seawater, all the oxides listed previous can occur if their activities are favorable.

2.3.3 Elemental Influences

The following elements have been found to increase the oxidation resistance of
TiAl alloys: Nb, W, C1, Mo, Si, Zr, Hf, Co, Te, Se and P [10,16,36,37,38]; the first five
yield the best results. Microeadditions such as Pd, Pt, Cu, V, Y, Cr and Mn tend to be
detrimental to oxidation resistance [10,16,37], whereas Fe, Ni, Co, Au, Ag, Sn, Y, Zr, Hf
and Ta have little effect on oxidation resistance [10,16,3 8].

There are four general rules for selecting an element that might be beneficial to
oxidation resistance [10]. The first is the valence-control rule (VCR or the Wagner-
Hauffe rule) which involves micro-alloying to slow (or halt) the diffusion of oxygen.

Since T102 grows mainly by diffusion of oxygen via oxygen vacancies, reducing oxygen

16

vacancies in T102 will enhance the growth of A1203. A study by Schutze and Hald [16]
used Cl' ions to interact with the T102 lattice to reduce the number of oxygen vacancies
and thus reduce diffusion rates in T102 limiting its growth.

Wagner’s scaling model is another guideline which establishes the criteria to form
a protective A1203 layer on the surface as opposed to forming internal platelets of A1203

and is given by the following expression,

 

N )I—flg—N DOV“ 05 XVI
I“ 3 0 D.11V()x ( )

where NA] is the aluminum content; No is oxygen solubility in the substrate; DA) and D0
are diffusivities of aluminum and oxygen in the substrate, respectively; VM and V0 are
volume of TiAl and A1203 per mole of Al in the alloy, respectively; and g is a conversion
factor for the transition from internal to external oxidation. Most of the parameters for
TiAl have not yet been established.

A third guideline for selection of elements that will enhance oxidation resistance
is choosing an element that will itself form an oxide barrier layer; to date Si, Nb and Mo
form such oxides, and these elements have been used in TiAl based alloys. The last
guideline involves modifying the initial oxide layer such as occurs with small additions
of Zr and Hf. Ion implantation using one of the following has also been investigated: Nb,
V, Ta, W, Cr, A], Y and Ce [10,40]. The implantation of Nb, Ta, W and Al were
beneficial to oxidation resistance, whereas, Y, Ce, Cr, and V-implantation were not.

2.3.4 Prevention/Protection
In general, metallic materials can be protected against high-temperature oxidation

only by the formation of scales consisting of either Cr203, A1203, or 3102, because only

17

these oxides have a sufficiently slow parabolic growth rate [32]. To protect y-TiAl alloys
a continuous A1203 (or) scale must be formed. However, the engineering alloys that are
based on y-TiAl develop scales that contain mixtures Of A1203 and T102 when oxidized.
The problem is even more complex when oxidation occurs in air so that nitrogen reacts
with T1 to form a TiN layer close to the metal substrate [33].

Three coating methods have been used to protect y-TiAl alloys: MCrAlY (M =
N1, Fe and Co), aluminizing, and silicides/ceramics [33]. Aluminizing has been used to
protect metals, especially nickel-based alloys for turbine applications [39]. More
recently, a fluidized bed of powdered ceramics consisting of mostly A1203 (60 mass%),
and C1203, Nb205 or W03 (40 mass%) were used to protect y-TiAl alloys and Inconel
713C [41]. This process was successful in protecting both types of alloys, and the W03
ceramic results gave the best result.
2.4 HOT CORROSION OF MATERIALS
2.4.1 Fundamentals of Hot Corrosion

Hot corrosion is accelerated oxidation that occurs at elevated temperatures caused
by salt deposits on metal surfaces. Early studies of this phenomenon were due to failures
of boiler tubing and severe attack of gas turbine airfoil materials. Turbine engines were
found to have hot corrosion due to Na2S04 (and/or NaV03), where sodium comes from
ingesting seawater aerosol and sulfur (or vanadium) comes from the fossil fuels. Hot
corrosion is classified as Type I or Type 11 [11,43]; Type I describes high temperature hot
corrosion (HTHC) 1n the range of 825°C - 950°C where the condensed salt liquifies (the

melting point of Na2S04 is 884°C); Type II or low-temperature hot corrosion (LTHC)

l8

occurs well below the melting point of Na2SO4 and is characterized by non-uniform
attack such as pitting.

Sodium sulfate (Na2SO4) is called an oxyanion-fused salt that exhibits an
acid/base chemistry with Na20 as the basic component and S03 the acidic component.
This type of salt also acts as an ionic conducting electrolyte. Consequently, hot corrosion
is electrochemical in nature, and it involves fluxing of the protective metal oxides as
either acidic or basic solutes in the fused salt [43]. Bomstein and DeCrescente [44]
performed studies with sulfur coatings and Na2SO4 on Ni-Co superalloys and concluded
that indeed it was Na20 which caused hot corrosion to occur and not the sulfur
components. This implies that Cl' ions are not important in hot corrosion conditions.

A model describing the basic electrochemical mechanism that occurs during hot
corrosion is shown in Figure 3. A cathodic reduction of $03 (as 82072”) or molecular
oxygen occurs at different interfaces, the location depends on the availability of ions from
the salt and the alloy. Figure 3(a) shows a salt film with a low concentration of transition
metal ions and a high permeability of 803 or 02; reduction of the oxidant species takes
place at oxide/salt interface, where electrons are supplied directly from the metal
oxidation reaction. The oxide/salt interface is then the most basic area within the salt
film, and either acidic or basic solutes of the oxide film can be formed there. In the case
of Figure 3(b), there is a high concentration and diffusivities of M” and M3+ , and
reduction of the oxidant species can occur at the salt/gas interface. This can occur if
electron charge is carried through the salt either by the counter diffiision of two
differently charged transition metal ions, or as electronic conduction in the film resulting

from electrons hopping between the transition metals. The salt/gas interface is then the

19

most basic area within the salt film, and either acidic or basic solutes of the oxide film
can be formed there. Therefore, the reduction reaction governs the gradient in basicity
across the salt film and it can be applied to corrosion of multi-component alloys by
various salts [l l]. The dissolution of the oxide layer and precipitated of the oxide within
the salt film continues in either case, but the reforming of the oxide layer may or may not
occur depending on the permeability of 02. The author did not mention relative thickness
of the oxide or salt film. Since dissolutioned oxide is reprecipitated within the salt film,

its relative thickness should increase.

 

 

 

 

 

 

 

Thin thin salt - thin alt
Tim 5
oxide film oxide film
Metal film Metal film [1912.]
Mo K . MO .
02 Gradient >Gradrent
or $03 102(3) [M31 02(8)
(as s2 03')
1“"qu £3- \ 2 EM or basic
2e- 801d or basic M—o-Mz*+ A (11 11.111
Oxidation 112* dissolution .2°.‘ 2+ 3:0 °°
Reaction _' 0f M0 Ondation --" 0 2+
Reaction 2M +1f202®
- 2M?+du02f
1,20 +23-—-o-02’ 3+ + re 611011
[rcduition reaction 2M + 2°""2M2 reaction
locally basic] locally basic]
(3) (b)

 

 

 

Figure 3. Electrochemical Model [1 l]

20

2.4.2 Hot Corrosion of Nickel Based Super Alloys

Hot corrosion (HC) of nickel-base alloys is accelerated oxidation caused mainly
by the presence of sodium sulfate (Na2SO4), other salts such as sodium metavanadate
(NaVO3) and sodium chloride (NaCl) can also cause HC in these alloys.

Nickel-base alloys exhibit two types of HC, either the alloy has oxidation induced
by Na2SO4 with limited corrosion similar to situations without Na2SO4 present or the
oxidation induced by Na2S04 is catastrophic. Both Types of HC require a liquid phase
present on the surface of the alloy. These reactions need not occur at or above the
melting point of Na2SO4 (883 °C) since the oxides formed by the alloy combine with
Na2SO4 to form lower temperature melts.

The hot corrosion of pure nickel and Na2SO4 has been studied [45]. Ions within
Na2SO4 and nickel oxide (N10) react forming a porous nickel oxide. This reaction is not
self-sustaining, it is limited by the amount of Na2SO4 present. The activities of the ions
to react in the Na2SO4 and N10 mix is only favorable during the initial stage of oxidation
when conditions are also favorable for the formation of nickel sulfate.

Nickel-based supperalloys can react differently than pure nickel depending on
alloying additions. Three common nickel-based alloys, Ni-Al, Ni-Cr and Ni-Cr-Al, were
studied for HC by Na2SO4 [45]. These were chosen since they develop N10, Cr203 and
A1203 as oxides. Observations from the Ni-Al alloys show that Na2S04 accelerates
oxidation initially (increased weight gain) and Na2S04 reacted with A1203 forming
NaAlO2, but that the oxidation kinetics were basically unchanged. The Ni-Cr alloys
differ from the Ni-Al alloys in that their oxidation weight gain with Na2SO4 applied was

less than their oxidation weight gain without Na2S04 applied. Also, Ni-Cr alloys have

21

Cr203 layer which is protective for an initial time period, and then sulfur from Na2S04
enters the alloy through the oxide layer, sulfur and Cr203 react forming chromium
sulfides below the surface of the Cr203 scale. The Ni-Cr~Al alloys, with high amounts of
Cr, had oxidation weight gains similar to Ni-Cr alloys, whereas the low Cr alloys
displayed oxidation weight gains similar to the Ni-Al alloys. The Ni-Cr-Al alloys with
high Cr form an initial Cr203 scale next to the metal interface, then a A1203 layer starts to
form in—between the external Na2SO4 deposit and the Cr203 scale. Sulfur from the
Na2SO4 diffuses through the porous A1203 layer and reacts with Cr203 producing
chromium sulfides. After an incubation period, the A1203 layer becomes continuous and
protective stopping further diffusion of sulfur. The Ni-Cr-Al alloys with low Cr do not
form a continuous protective A1203 scale due to sulfur from Na2SO4 reacting with Cr at
the metal interface, this produces chromium sulfides and preventing the formation of the
necessary chromate which is a precursor to continuous alumina formation.

A superalloy containing nickel and cobalt has been studied for HC by sodium
sulfate (Na2SO4), sodium metavanadate (NaV03) and sodium chloride (NaCl) [46], this
alloy was also subjected to tensile loading during HC. The hot corrosion stress testing
using these salts lowers the rupture live of Ni-Co superalloys. However, unstressed
specimens have a decreased corrosion rate after time due to the formation of protective
nickel vanadates. Consequently, the kinetics of HC differ depending on whether the
material is stressed.

2.4.3 Hot Corrosion Due to Sodium Chloride
Hot corrosion due to NaCl on a Ni-Co alloy was mentioned in the above study

[46]. The NaCl in the molten salt mixture decreased the solution’s viscosity allowing

22

easier decomposition of Na2804 which lead to formation of oxychlorides. Several
reactions between Ni, N10, Cr203, Cr and C12 then result in corrosion at grain boundaries.

Sodium chloride can also cause HC in austenitic and ferritic stainless steels by
reactions between Cr203, Cr and C12 [48]. Hot corrosion resistance in the ferritic steels,
can be increased by additions of aluminum, and austenitic steels can be improved by
additions of silicon. The author noted that chlorine was detected at grain boundaries
while sodium was not, but did not infer that chlorine acts as the corrosion accelerant for
these steels. Another study by Fujikawa and Maruyama [49] states that austenitic
stainless steels that are exposed to HC by NaCl above 500°C have a greater depth of
corrosion. Also additions of Mo are beneficial for corrosion resistance but Cr is only
beneficial above the melting point of NaCl (801°C).

A recent study performed by Yao and Marek [15] on a TiAl alloy (T1-47Al-2Nb-
2Mn-7 Vol. % TiB2) revealed that deposited NaCl accelerates corrosion. They observed
weight gain after high temperature exposure, and the NaCl and oxide surface morphology
was examined by SEM. In their study, NaCl vapor deposited crystals with an
approximate size of 6.51m x 311m were uniformly distributed on the specimen
(approximately 2.3 x 104mg mm"2 of NaCl ). The specimen was heated in air to a
temperature of 760°C (below the 801°C melting point of NaCl) and observed afier 0.36ks
and 0.72ks. After the first 0.36ks (six minutes) there was nodular growth on and around
the salt crystals. Afier 0.72ks (six minutes later) the grth had accelerated to cover the
crystal and a majority of the surrounding area (about twice the crystal’s size). These
nodules were attributed to formation of Ti and Al oxides. The scale that formed was non-

protective.

23

2.4.4 Laboratory Testing

A salt fog test is an accelerated corrosion test in which specimens are exposed to a
fine mist of a solution usually containing sodium chloride but sometimes modified with
other chemicals (Ref. ASTM BI 17) [14]. Burner rigs are also used where the specimen
is placed within the exhaust flow which may have a solution of salts injected into the
flow. A typical test cycle consists of 3 minutes at 845°C (1550°F), idling; 2 minutes at
1010°C (1850°F), takeoff; then removal from the flame for 2 minutes, engine shutdown
[12]. Another test is a Dean test in which Na2S04 is heated to above about 1100°C to
establish an appropriate vapor pressure and the sample is maintained at a temperature
below the dew point where condensation occurs [50].

Experimenters have used various methods for applying salt to specimens for
example: specimens are dipped in a solution of NaCl [49], soaked in NaCl (26%) for 5
min [48], sprayed with a solution of Na2SO4 (60 wt%), NaVO3 (30 wt%) and NaCl (10
wt%) onto heated specimens at 150°C [46], coated the specimen with 0.5mycm2 of
Na2S04 by spraying the solution on a heated specimens at 150°C [45]. They have also
tried different heat cycles; heating for 2 hours at 450-850 °C then air cool [49], heating
for 2 hours 450-750 °C then air for cool 5 minutes —— repeat for 50 cycle [48], heat for
100 hours at 900 — 975°C then air cool [46].

2.5 STRATEGY FOR INVESTIGATING HOT CORROSION
2.5.1 Hot Corrosion Summary

Hot corrosion (HC) is electrochemical in nature and involves an oxyanion-fused

salt that exhibits an acid/base chemistry. The salt sets up a flux of ions which uses a

metal’s (or alloy’s) oxide(s) as an electrolyte.

24

The hot corrosion of nickel-based superalloys has been studied and well
documented. Sodium sulfate (Na2SO4) is the salt that has the most corrosion potential for
HC of nickel-based superalloys. Sulfur ions diffuse through an alloy’s oxide(s) to form
either nickel or chromium sulfides. Other salts that are known to cause HC in nickel-
based superalloys are sodium metavanadate (NaVOg) and sodium chloride (NaCl).

Stainless steels can experience HC by NaCl, it has been extensively studied for
this common material. Chloride ions diffuse through the steel’s oxide and react with
chromium carbides, this causes pitting and intergranular attack. The chromate oxide,
which protects stainless steels, can also be attacked by chloride ions.

Titanium aluminides have had limited HC studies performed on them. Two
recent studies focused on the HC of titanium aluminides by NaCl [15] and by a sulfur
environment [5 2]. It was theorized that sodium can react to form compounds with
oxygen and aluminum or titanium. Sodium chloride was found to accelerate a porous
non-protective oxide growth in air but it has no effect on the alloy in an oxygen-free
environment. The HC by a sulfur environment caused the TiAl alloy to develop a non-
protective oxide scale that spalled ofi‘. The study also determined that sulfide precipitates
formed at the interface of the oxide and the metal.

2.5.2 Research Approach

This study will investigate the hot corrosion of the intermetallic TiAl (nominal
composition in at% Ti-48Al-2Cr-2Nb) by sodium chloride (NaCl). The scenario used to
motivate the investigation is to consider aircraft engines operating in frequent short

flights of about 1.8ks (30 min.) in a marine setting such as the Hawaiian Islands.

25

The NaCl used was a simulated seawater solution (ref. ASTM D 1141—90)
deposited by airborne application (atomized in a fog). The method is similar to ASTM
Bl 17 [14], which was chosen due to its ease of reproducibility in the laboratory,
convenience and expense. Bare ground and pre-oxidized TiAl specimens were coated
with a thin layer of salt water in a manner that simulates conditions for engines at rest in
coastal airports and similar to conditions experienced in coastal regions [30]. Specimens
were heated in an air environment to 760°C for various times, then quickly air-cooled.
The times represent start-up and immediate shutdown of an engine; start-up, short flight
and then shutdown in repetition. These times were chosen because there is no set
duration in the literature for HC testing, so experimenters have more or less developed
their own procedures. More developed procedures exist for burner rig testing [12].

The compositions of the surface and subsurface layers were measured using EDS.
This was used to determine if sodium or chloride was present (or not present) similar to
findings in previous studies [15,46,48,49]. Flexural (four—point bend) testing for strength
degradation was performed after the heat cycles along with surface morphology
examination by SEM. The method and fixtures used for the bend test are similar to
ASTM C 1161 (flexural strength of advanced ceramics at ambient temperatures). This
was chosen because of the brittle nature of ordered intermetallic materials. An
interrupted weight gain experiment was run to verify accelerated oxide growth, like those
previously performed on this alloy [33]. Differential Thermal Analysis (DTA) on
powdered metal and simulated sea salt at temperatures from RT to 900°C, and Powder X-
ray Diffraction (XRD) was performed to analyze the many possible reactions and

reaction products that might occur. Due to the use of different micro-alloying elements in

26

TiAl alloy, that can alter the activities of the constituents of these alloys, and depending

on the exposure environment, many different reactions occur and products result [15, 52].

27

CHAPTER 3
EXPERIMENTAL PROCEDURE

3.1 MATERIALS

The material used for this study came fiom an investment cast plate of
composition Ti-47.9Al-2Cr-2Nb (PRET material). The plate was cast at Howmet Corp.
(Whitehall, MI.) and was heat treated at 1093°C for 5h, hot isostatically pressed (HIPed) at
120501: for 4h and heat treated at 1205°c for 2 additional hours followed by rapid cooling.
Also, a sample of mechanically alloyed TiAl powder was used for part of the differential
thermo—analysis testing.
3.1.1 Salt Application, Four-Point Bend and Pro-oxidation Specimens

Specimens utilized for the salt application, four—point bend testing, pre-oxidation
study and the weight gain study were cut using electrodischarge machining (EDM). The
specimens were ground using a series of SiC grinding papers from 240 to 600 grit to remove
the EDM heat affected area (resulting dimensions were approximately 34.0mm x 4.0mm x
1.3mm) and then ultrasonically cleaned in acetone.
3.2 SALT APPLICATION

Several specimens measuring approximately 37.0mm x 7.0mm x 4.3mm had
simulated seawater solution (Aquarium Systems Instant Ocean®) of approximately 3.5 %
salinity with pH = 9.0 applied by means of a fine mist. Three methods were explored to
apply the salt water, using a spray bottle on specimens inclined at an angle of about 60°,
or creating a fog using either a nebulizer (OMRON Corp. Air Nebulizer Model NE-C08
CompAir®), or a 1 meter tall standing pipe with an air stream blowing bubbles in salt water

at the bottom of the column. One sample was exposed on a residential balcony on the NW

28

side of Maui, at a distance of about 50 yards from the coast, to determine how salt was
deposited in natural conditions. The nebulizer method was finally determined to produce
a salt distribution most similar to that occurring by coastal regions (similar to the salt
distribution on the Maui exposed specimen). The spray bottle method was used on the
first group of specimens, but all subsequent groups of specimens had salt applied by the
nebulizer method. The nebulizer was filled with 3 ml of simulated seawater for a run time
of about five minutes. The nebulizer output was connected to a 9.5mm inner diameter
and 1.5m long coiled polymer hose with a coiled structure, which in turn was connected
to a polymer funnel as illustrated in Figure 4. The hose prevented the larger droplets
from reaching the specimen, and the funnel directed the flow evenly to the specimens and
restricted stray air currents. The specimens were then allowed to air dry for 24 hours.
Morphologies of the deposited salt crystals were viewed using a light microscope to
measure the size distribution and area of coverage.
3.3 FOUR-POINT BEND

Specimens were heated in a Lindberg furnace in an air environment. The
specimens were placed on a layer of alumina powder on a copper plate. A control
thermocouple was connected to the copper plate. Groups of specimens were exposed for
0.15 ks, 0.36 ks, 1.08 ks, 4.4 ks and 19.44 ks at 760°C. A control group of specimens,
which did not have any salt application, was exposed for times of 0.0 ks, 0.36 ks, 1.08 ks
and 19.44 ks at 760°C. The specimens were removed from the furnace and air-cooled on

a refractory brick.

29

Specimens
under

I UNIS] ' 1A1 Pump

 

Figure 4. Nebulizer Set-up to Apply Salt Fog to Specimens.

30

Pin

Clcvice
Joint

I Upper CompressionPlate I

 

 

 

 

 

l LowerCornpressionPlatc l

 

Figure 5. Test Fixture for Four-Point Bending

31

An environmental scanning electron microscope (ESEM) Electroscanm model 2020 was
used to view and analyze the resulting surface morphology of the heated samples to
examine how hot corrosion occurred. The environment within the ESEM was water
vapor at a pressure of 2.5 to 5 torr. Energy Dispersion Spectroscopy (EDS) was used to
analyze certain areas of interest while viewing the specimens using the ESEM.

Afier ESEM analysis the specimens were subjected to a four-point bend test. The
four-point bend tests were performed using an lnstron 4206 tensile/compression tester
and the fixture shown in Figure 5. The specimens were loaded by downward movement
of the crosshead applied at a rate of 0.1 mm per minute. Morphologies of the fracture
surfaces were studied using a JEOL 6400 SEM.

3.4 PRE-OXIDATION

A set of specimens was pre-oxidized in the furnace prior to application of a salt
fog and 760°C exposure. Half of the specimens were pre-oxidized for 3.6 ks and the other
half for 360ks; all were pre-oxidized in an air environment. Salt was applied to all
specimens followed by a 1.08ks, 760°C exposure. The salt application and heat cycle
were performed ten times on the 3.6ks specimens. Some specimens had salt applied
every cycle while others had salt applied every other cycle. The 360ks specimens were
exposed to just a single salt and heat cycle. Afier each cycle the specimens were viewed
using a light microscope and ESEM, and if there seemed to be a change from one cycle to
another photos were taken and EDS performed.

3.5 WEIGHT GAIN
Another set of specimens was pre-oxidized in the furnace prior to application of a

salt fog and 760°C exposure. These specimens were extra pieces from the casting. The

32

specimens were ground using a series of SiC grinding papers from 240 to 600 grit to remove
the EDM heat affected area and then ultrasonically cleaned in acetone. All specimens were
pro-oxidized for 3.6 ks (l h) in an air environment. After the pre-oxidation, one
specimen (specimen S) had salt applied to all of its surface area, two specimens
(specimens H1 and H2) had salt applied to half of their surface area and the other half of
their surface area was masked using a Kimwipes® paper tissue to prevent salt from
reaching the surface, and one specimen (specimen NS) was used as a control that had no
salt applied to its surface area at all. The surface areas were approximately 6.10cm2 for
specimens 8 and NS, and 4.84cm2 for specimens H1 and H2. Specimens were run through
ten cycles of salt application and heat exposure, except for the control specimen that was
only exposed to the heat cycle. Initially and after each cycle the weight of each
specimen was taken using a Denver Instrument Co., Model A-200D scale with a
sensitivity of +/- 0.0001 g and recorded. 35-mm photographs of the resulting surface
appearance were also taken.
3.6 DIFFERENTIAL THERMAL ANALYSIS

A piece of the casting from which previous samples were taken was used to
obtain a fine powder for Differential Thermal Analysis (DTA). The casting was held in a
vice and a steel file was used to obtain filings. Afier a significant amount of filings were
Obtain a magnet was used to separate the steel filings from the casting filings. This was
accomplished by moving the magnet directly below a thin piece of paper, on which both
types of filings were lying. This process was done as many times as possible until it was

seen that the magnet attracted no more steel filings.

33

Two different mixtures of simulated sea salt and casting filings (or the
mechanically alloyed TiAl powdered metal) were prepared for DTA testing: one mixture
had a weight ratio of 9:1 and the other had a weight ratio of 1:9, simulated sea salt to
casting filings, respectively. The amounts for the mixtures were measured using a
Denver Instrument Co., Model A-200D scale with a sensitivity of +/- 0.0001g. The
constituents for each mixture were mechanically combined using an alumina mortar and
pestle for a period of 7 to 10 minutes until the mixture “looked” homogeneous. The
mixtures were then placed in a glass vessel and sealed with a polymer lid (creating a
reasonably air tight seal.)

DTA runs were performed using TA Instniments 910 Differential Scanning
Calorimeter equipped with a 1600 DTA test fixture along with Thermal Analysis 2200
data acquisition system and software. Trials were run with sample sizes of approximately
30 mg from room temperature to 900°C at a ramp rate of 20°C/minute with a 100
ml/minute nitrogen gas purge. The standard used was 99.999°/o pure aluminum.

3.7 X-RAY DIF FRACTION

Powder samples from the DTA castings were used for X-ray diffraction analysis.
The powder was attached to the surface of a 2.5mm X 76mm X 1.5mm soda-lime glass
slide using petroleum jelly in a 5mm X 15mm area. A Scintag model XDS 2000
diffractometer and associated software was used. The system was set to scan from 20 to
90 degrees for the 9:1 salt to powder ratio and from 5 to 90 degrees for the 1:9 salt to

powder ratio at a scan rate of 0.5 degrees per minute.

34

CHAPTER 4
RESULTS

4.1 INTRODUCTION

The results obtained from the various experimental methods employed clearly
indicates that exposure to salt followed by elevated temperatures will cause hot corrosion
of the alloy Ti-48Al-2Cr—2Nb. There is evidence of reactions occurring between bare
ground specimens and salt, but pre-oxidation appears to prevent these reactions from
happening. Exposure to salt and elevated temperatures for long times (about 20ks) only
slightly degrades the fracture strength of this alloy, but the statistical scatter Of the data
can be attributed to surface or interior microstructural flaws present as suggested from
fracture surface micrographs.
4.2 SALT DISTRIBUTION

Salt crystals deposited by means of water vapor were either small singular crystals
with a sharp rectangular shape or a cluster of rectangular crystals with a roughly circular
shape. The salt particles had a size range from about 1 to 30 microns and covered up to
15% of a specimen’s area. The small crystals often formed in grooves from the grinding.
The spray bottle method caused a few large water spots, which deposited salt cluster with
sizes close to 150 microns. Figure 6 shows the salt particle distribution that occurred due
to natural exposure near the seacoast and Figure 7 shows the salt particle distribution that
occurred by using the nebulizer method. The naturally exposed specimen had the
smallest area coverage, but the size distribution was most similar to the method that used
the nebulizer. Figure 8 shows the salt distributions by nebulizer, natural exposure, and

4in.-dia. standing tube (specimen placed two feet above air agitated seawater solution).

35

...-4

....
Ike»!

.... ... a. n
. 4.1.. ..
. 10%.

2

I

3

OUT}.

.5

wins“.

H.

.

it

b u». v...)

V

11““.
n ... a

... A
.. ...?flv .
Jrfiikww v
. ....

an.

 

stribution.

Figure 6. Naturally exposed salt di

36

 

 

izer salt distribution.

Nebul

7

igure

F

37

 

J

  

 

 

 

 

 

0.4 1- Salt Distribution by Various Methods .3
: —o—srANOING ruse :
>~ 0.3 — -o-— NEBULIZER ..
g 1 —E’:1-— HAWAII i
0' t 1
0.2 T .1
E 2]] -
0.1 E -
I ,/ .\

0 I r .L l . . . L

0 5 1 25 30

6in ion) 2°
Figure 8. Salt particle size distribution

4.3 SALT ATTACK ON BARE GROUND SPECIMENS

The general features of salt attack were similar for the spray bottle and the
nebulizer specimens, (excepting the large clumps of salt on the spray bottle specimen), so
images of attack are shown together. An environmental scanning electron microscope
(ESEM) along with energy dispersion spectroscopy (EDS) were used to determine if hot
corrosion -by salt altered the composition, surface and interfacial morphologies of the
alloy. A summary of EDS results measured on different features is presented in Figure 9
through Figure 11. The results are displayed as a ratio of the element of interest counts to
total titanium counts obtained at a particular area, i.e., all scans are standardized to total
titanimn counts for that particular scan. EDS was performed on an electropolished
specimen, which was not exposed to a heat cycle or salt fog, to obtain a basis for the rest
of the measurements taken. This will be referred to as the “base metal” and is indicated

on the vertical axis in Figure 9 through Figure 1 1.

38

 

 

 

 

 

 

 

  

 

 

 

0.2 L I ~I T TTFIIF IflfI—Ifi'I—ITIT—H—fl-I—fi'mT—T'TI'T—T'fi If 1 T7 I171] fl 7 I I IIII
Area Mode Black-1 st Run . EDS-Relative -
l
I [3 Metal Red—2nd Run Sodium Counts I
0.15 if?) Oxide metal a, lamella r
e ' edge of salt €199 25 and I
Fi . 16md ' -
E ” Me jag? 17) salt/metal Fig. 26) __ i
‘0’ — Metal “—7 interface
a: 01 f. . , (Fig. 16and 7
0“? r 0““ / Fig. 18) / ..
g I . interface / / / .,
8 l . (’1 ’1’ l. /" lamella
005 I— pitted area I / V region I/ r‘
\ ... "I i " ,/ and i i I
I base metal I. .' 7 Fl . 23) “
i (Fig 11) Li 9 j
0 : W14.Lil_i__._ 1H; i__1_1..l__1!... ML.__- 1-.L1_I;- 1 1 1 1 1 lliI _..-.i. -1.-- 1__J_L.l.
(a) 1 10 100 1000 10‘ 105
Time(s)
0.35 l T I I ITI'; —#_ F'IT’TLT‘TTIIT'TT” "—‘7' MIMI—"VII? I I I I I l Err I I I I I III
” Area Mode . '
ELI Metal 0 EDS-Relative
0.3 rm Oxide ' Chlorine Counts
A C Sag Mode sent/metal interface
EOZSE‘. Metal mdmetalatedge -
0 . . Oxide £59.11; Fig. 17,
2°: 0 2 , ”i “me I / lamella
& Black-1st Run pit
g 0.15 1394an Run a (Fig 25 1
o " pitted aea /'/ Fig 26) I
0 0.1 {— (Fig. 13 and Fig. 14) 7” /
" lanella /
-' l [:1 — "99'0" , . / -
0.05 D888 metal {-.-} a .. (:19. g, " / -
(Fig 11) I .--- " '9 ) :
0 L3 1 l 1 lLll 1 l l-.1--ld_l_Ll_...____i._Ln 4-14-11. 4 L 1 l 1 1 ill I ,l_-_1_.-.1_-1_1._Li.l
(b) 1 10 100 1000 10‘ 105
Time(s)

Figure 9. Ratio of EDS intensity for (a) Na, (b) Cl.

39

.0
oo

 

 

 

 

 

 

 

 

 

 

7 ba‘Ze—hgigi l’l‘T‘T‘ ’7' 1 "" "I 'I’ l ‘T‘T‘I‘ »:;T"”’ T. ‘V‘ITTTW‘ T‘i'TI T‘TI’v‘WfiT '* I 'T"i 7 T?
O 7 ‘ (Fig. 11) I I edge ofsalt _;.
Area Mode U r/ (Fig. 16. Flg. 17) I
0-5 I LI Metal salt/metal  
c: : 1:1 Oxide interface . 1
E 0.5 : Spot Mode (Fig. 16, . ED.S'R9|auV9 ;
3; - . Metal Fig, 13) Aluminum Counts 1
o i i
“E 0.4 r o Oxide I
E : ' “ Interface I 9 I" i
:3 0-3 E— Black-1 st Run _ / . lamella ”I
8 » Red-2nd Run pitted area P", 7
02 L (Fig. 13, Fig. 14) m (Flg. 25, , g
:. /,/"‘1-—J Fig. 26) I
0.1 I» lamella region / I I —:
(Fig. 22, Fig. 23) (I) 5
o I 1 1-11. 11 . 111 1__1_ 1 1 1 1111 1 1 11 1 1L1_I.___.L_1- 1 1 1 1111 1 ,-1_._1.._1_1_1_11:

(a) 1 10 100 1000 10‘ 1o5

Time(s)

o 14 I '— ' 1 r r T 1 i I] "‘"T."“ r""“1‘7_'—*TI#"T r T 1—’rr T fiTW—TTT r I r 1 iT—r‘T:
[A [Am-91°- . EDS-Relative 1
I” U Metal Black-1 st Run Oxygen C aunts:
0.12 O Oxide Red-2nd Run -.
A 0 1 I SM Mode . Q
5 ° ; I Metal ~;
‘6’ C. . Oxide - .— 1 I
'3': 0-08 » >< Interface 1W3 :
at? - " ed .. , 0 lamella _
1 pitt area 1.? region 1
g 0.06 :7 (Fig. 13. Fig. 14) ..... .1 {(Fig. 22. 6
o I metal at ,1. (.... 4‘/ Hg 23) lamella 7.
0 0 04 1- edge of salt //_,../ "/ L. 1 pit. -.
' L (Fig. 16, Fig. 17) / [---] _, ~:— (Flg- 25. 3
L salt/metal fig. 26) f.
0.02 [:1 base metal interface . 4
1 (F19. 11) (Fig. 16, Fig. 18) I _ :
0 1.- ._.._1-.- 111-1 111-1111.-.; - 1 1 .111LL1 1 1 111 111 1 1 1 1 1 111 _-....1 1-1-1.1111

(b) 1 10 100 1000 10‘ 1o5

Tirne(s)

Figure 10. Ratio of EDS intensity for (a) A], (b) O.

40

0.12

 

 

 

 

  

 

 

 

 

"“"“T‘-—r"1—‘TT?TTT“‘EAV * i ’fi f rm i r' ""‘T‘""'T‘“T_I‘T_TTII """" r YT—‘W—T‘T‘T'T—WT
’, W EDS - Relative i
0 1 :_ Bi Metal Black_1st Run Chromium Counts;
' - U Oxide Red—2nd Run a
’E‘ . §29—‘ ”Ode salt/metal . lamella —
h 0.08 f 3 Metal ,. I inteiface it I:
Q r 0 Oxide ' f (Fig. 16, Fig. 18) y 9F. 25 -
'g 0 06 I ‘ 'nte'face I / ‘éigg'zeih
g H . /. ,. -
_, pitted area fi—~._____ lamella
‘g‘ . (Fig. 13, Fig. 14) . L1 region ~
8 0'04; “I“ i (Fig. 22 I "Q
whase metal metal at _.__. . . Fig. 23) f _ -i
0.02 H(Flg. 11) edgeOfsalt ,, ,ng , i
~ (Fig. 16, Fig. 17)
o i _,_ .1 -. i--... l_.l_i.L LL______ L__ L._l_ 1.4,: 11L. -..; __l_L l 4;.- 14% .-._1__ ...LA_L. LL_1.JJ_L___.__L_L_L_LLLJ
(a) 1 10 100 1000 10‘ 1o5
Time(s)
1 ———' 7" 'I ‘u—r-ThT—TT‘TT ‘u-‘T‘V' I" If If VIII—IT "7 ’fiF—IMT’T—TT‘TI 7' ”—1.. I I IT TI II I I I I I I T.
. WE EDS-Relative
i L] Meta. 0 Niobium Counts
0-8 ‘ O Oxide ”’
A 5. mm _
.. . 4
S _ . Metal salt/metal inten‘ace _
E 0-6 Oxude and metal at edge
.0 _ X interface (Fig. 16, Fig. 17, Fig. 18) ..-.
fi lamella
m i Black-1 st Run pit
E 0-4 ‘ Red-2nd Run “:59 25 H
3 * . ’ 'i
8 * pitted area Hg 26) -j, ’1
O 2 ”’ (Fig. 13, Fig. 14) - . lamella / *‘
- " U _ region ,, ” A
- I ,1 . , 4
_ / (FIQ 22. 7“ -
. . , ' I
O L.L_L_.L_.J_J_.L_Ll_i_- _ - .L__..L L ; :1 '11 1 A L 1 l H ,, 1 J J. 1 l 1111 z_.__L__L__1_.;_L;L_L
(b) 1 10 100 1000 10‘ 1o5
Time(s)

Figure 11. Ratio of EDS intensity for (a) Cr, (b) Nb.

4]

Counts

 

190001
18000
17000
16000
15000-4
14000
13000
12000
11000
1

 

 

 

Nb

LILLJ UM“

 

 

 

 

 

Ti

chr

 

 

 

IIIII‘ITIIITIITIlIIj1ITIIIIIII‘III—I}TIIIIIITI

0 1 2 3 4

5 6

8
Energy (keV)

Figure 12. EDS scan of electropolished specimens -— “Base Metal”.

42

Images were taken at particular areas of interest are shown in Figure 13 through
Figure 26. Sample EDS scans for a few of these areas immediately follow the
photograph. The data obtained from the EDS scans was used to create the plots shown in
Figure 9 through Figure 11. Figure 12 is the EDS scan of the base metal, the specimen
used was electropolished and was stored in an open-air environment (this may account
for the detection of nitrogen and oxygen).

Figure 13 and Figure 14 depicts the general surface of the O. 15 ks specimens. Salt
crystals have spread out forming tree-like structures indicated by the arrow in Figure 13.
EDS reveals small concentrations of Na and C1 in these areas. In another area with a salt
clump adjacent to a pit, there are larger Na, Cl and 0 concentrations, and lower A], Cr,
Nb concentrations compared to the base metal (Figure 14). The EDS scan of this pitted
area is shown in Figure 15.

The surface morphologies of the 0.36ks specimens show various stages of oxide
growth. Figure 16 shows grth of oxide that follows the tree-like structure seen on the
0.15ks specimens. Four distinct regions can be identified in Figure 17. The topography
rises from region 1 where a salt clump chipped off to the thin oxide layer of region 4.
Region 1 is where the salt deposit was originally and then got knocked off. Region 2 is
the interface region between the salt and the metal (Figure 19 is the EDS scan). The metal
is apparent in region 3 (Figure 18 is the EDS scan). A thin oxide layer is identified in
region 4. Around the salt particles on these specimens, there is an increase in the Na and
Cl concentrations, the 0 concentration stayed about the same, and there was a decrease in
the Al and slight decreases in Nb and Cr concentrations. The obviously attacked areas on

the 0.36 ks specimens had a decrease in Al concentration, and increases in the Cr, Cl and

43

I
”it!
fwe :‘1'Q1
“tr-”Ifnfinblf

l ‘ V

{1,175.19 0W“ ‘W

£32339 ‘g

 

Figure 13. 0.15ks specimen - general area.

 

Figure 14. 0.15ks specimen - salt clump near a pit.

45

 

23000
22000

 

 

 

O

 

 

Nb

Cl

Cl

 

 

 

Cr
Cr

 

 

III—rle—IIIIIIIIITIIIIIIIIIIIIIIIIIII

0 1 2

3

4

5

B 7
Energy (keV)

Figure 15. EDS scan of a surface pit on 0.15ks specimen (Figure 14).

46

'F‘ I
3:; fur"

t

,1 -< e-.

...“
.9. -msefl,,..,‘..,, ..

a {'31

 

Figure 16. 0.36ks specimen - accelerated oxide growth.

47

‘1

 

Figure 17. 0.36ks specimen — exfoliated salt deposit.

48

 

15000—4
14000
13000
120004
11000

 

 

 

 

 

 

 

 

 

 

 

Ti

CrCr

 

 

rirrlrrrrlfi—Iri—lrrlrlrfirlrriilrirTerIIIrI—rTI—

0 1 2 3 4

5 e 7 a 9
Enemy (Rev)

Figure 18. EDS scan of the metal at edge of exfoliated salt deposit (region 3).

49

 

19000 m
180001
170001
16m
1 5000-
14000
13mm

110001
10000—«

8000}
7000
60001

40004
3000« Ti
20001 Nb

1000‘ L/LJl /\ Cl I c' c,

 

 

 

 

 

 

 

 

 

Energy (keV)

Figure 19. EDS scan of the metal/salt interface by exfoliated salt deposit (region 2).

50

 

Figure 20. 0.36ks specimen - light and heavy salt deposits and resulting oxide.

51

 

Figure 21. Surface morphology of specimen exposed for 1.08ks - general oxide.

52

 

Figure 22. Surface morphology of specimen exposed for 1.08ks - fracture in oxide.

53

,«r

. "may
i

 

Figure 23. Surface morphology of specimen exposed for 1.08ks - damaged area.

54

 

140004
13000-
12000
110N‘

 

 

 

 

 

 

 

 

_-
c w—T 1W h A

IIITTIITIIIFIII—rIIII IIIIIITjIIIITI‘fiIIIITIITI

 

0 1 2 3 4 5 a 7 e 9
Enemy (keV)

Figure 24. EDS scan of an attacked lamella region (012 phase) in Figure 23.

55

Nb, and about the same concentration of 0. Figure 20 shows an oxide area where
the salt deposition is light, area A, and other areas where the salt deposition was heavier,
area B. The area where salt deposition is heavier is where the oxide grth is greater.
Figure 21 shows the oxide scale that developed on the specimens exposed for 1.08 ks.
The oxide has low concentrations of Al, Na and Cl.

Where the oxide layer is fractured in Figure 22, there is a decrease in Al
concentration, and O, Na, and Cl was also detected For the damaged region on the 1.08
ks specimens in Figure 23, there is a decrease in Al concentration, an increase in Nb and
Cr concentrations; also 0 and Cl were detected. The EDS of the damaged region in
Figure 23 is shown in Figure 24. The a; phase (Ti3Al) is being preferentially attacked as
evident from the erosion of lamallar microstructure. The 1.08ks specimens developed an
oxide that spalled off, due to thermal shock, when the specimen was remove from the
furnace, this did not occur on any of the other specimens.

Figure 25 shows the oxidized area on the specimen exposed for 19.44 ks. There
are increased concentrations of Na, 0, and a trace of Nb at the oxide surface. There was
a decrease in Al concentration and no C1 or Cr at the oxide surface. At the interface of
the oxide layer and metal, the Al concentration is lower. Figure 26 shows a damaged
area, which was under the exfoliated oxide. The EDS scan (Figure 27) shows that the
area has no concentration of Na or Cl, decreased Al concentration, and increased
concentrations of Cr, Nb and 0. Once again, the erosion of lamallar microstructure

reveals or; attack. All the above EDS values are in comparison to the base metal.

56

“raw ..

Irv - 1.1m ‘ A

 

Figure 25. Specimen exposed for 19.44ks — exfoliated oxide.

57

m

.. ‘
2"“.

if!

1 .-.,

sir

 

Figure 26 Specimen exposed for 19.44ks - damaged area under oxide.

58

 

 

 

Figure 27. EDS scan of an attacked lamella region (a; phase) of Figure 25 .

59

Ti

 

 

 

 

 

 

 

 

 

4.4 EFFECT OF SALT ATTACK ON FOUR-POINT BENDING STRENGTH

The four-point bend test results are given in Tables 3 for specimens that were
exposed to salt and the elevated temperature cycle, and Table 4 gives results for
specimens exposed to the elevated temperature cycle only. Figure 28 shows the plotted
data for all specimens. From these results, it is apparent that a consistent degradation was
observed for the longest time, whereas the fracture strength was variable and higher for
the shorter exposures. This particular alloy has shown to have yield strengths close to 400
MPa which is comparable to 400 — 630 MPa yield strengths in the literature [3]. The
variability in the fracture strengths may be attributed to uneven loading conditions, i.e.,
the test may have simulated a three point bend test in lieu of a four point bend test. This
may be inferred due to the fact that most fractures occurred at or near the contact point
between the upper pin and the specimen. Loading errors due to fixture slippage resulted
in tests being restarted. Figure 29 shows the data for specimens that did not slip or
fracture at a pin during the test, and shows a more accurate trend in decreasing strengths
with increasing time. A schematic of the casting showing the locations from where the
specimens were cut is shown in the inset of Figure 28.

Fracture initiation sites were sought in one specimen from each exposure time.
There was little distinction that could be found between the specimens; all had multiple
initiation sites along the tensile surface. Figure 30 Illustrates a region where the surface
was not flat, possibly due to oxidation in a pitted area (arrow). However, most of the
exposed surface near the fracture had no adherent oxide. Some residual oxide is evident

at the arrow in Figure 31.

60

Table 3. Four-point bend results on specimens with salt application

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Time MOR Length of Comments on Fracture Location

ksec MPa pieces, mm

0.15 480 25, 9 within 1 mm of pin, slight curvature in longer piece,
fixture slipped-test restarted, top section of slab

0.15 410 21.5, 12.5 2.5 mm toward center from pin, fixture slipped-test
restarted, top section of slab

0.15 516 8, 9.5, 16 at both pins

0.15 469 AVERAGE VALUE

0.36 474 21.5, 12.5 toward center, z2.5 mm from pin, fixture slipped-test
restarted, top section of slab

0.36 616 235,105 at pin, fixture slipped-test restarted, top section of slab

0.36 444 26, 8 z2 mm fromjin

0.36 409* 23.3, 11.3 Fixture slipped twice before fracture within 1 mm of pin,
*largest stress occurred on second loading, bottom section
of slab

0.36 486 AVERAGE VALUE

1.08 571 25, 9 within 1 mm of pin, top section of slab

1.08 577 23, 11 within 1 mm of pin, top section of slab
accidentally preloaded and bent, fractured z1.5 mm from

1.08 142 25.5, 8.5 pin, (not counted in avg.)

1.08 541 16.9, 17.9 Fracture near center of specimen, bottom section of slab

1.08 563 AVERAGE VALUE

4.4 492 11.3, 23.5 within 1 mm of pin, bottom section of slab

19.44 433 19.5, 14.5 in center, fixture slipped—test restarted, top section of slab

19.44 431 23, 11 within 1 mm of pin, fixture slipped-test restarted, top
section of slab

19.44 423 24.5, 9.5 within 0.5 mm ofpin

19.44 507 12.7, 22.1 within 2 mm of pin, bottom section of slab

19.44 449 AVERAGE VALUE

 

 

 

 

Table 4. F our-point bend results on specimens without salt application

 

 

 

 

 

 

Time MOR Length of Comments on test and fiacture location

ksec MPa Pieces, mm

0.0 523(1) Ran test twice and could not fracture specimen due to
480(2) repeated slipping of the fixture, bottom section of slab

0.36 565 24.18, 10.66 Fracture near inner pin, bottom section of slab

1.08 163(1) 24.13, 10.64 Fracture near inner pin, fixture slipped twice before
111(2) fracture accurring on third run, bottom section of slab
450(3)

19.44 529 24.42, 10.33 Fracture near inner pin, bottom section of slab

 

 

 

 

 

Note: The number in parentheses shown in Table 4 refers to the loading run.

61

 

Modulus of Rupture , MPa

no

650

O)
O
O

0'1
01
O

500

450

Modulus of Rupture , MPa

400

 

 

 

 

 

 

F1
top]

{-10%

bottoml
Ill
iiliiiliiiliriliLiiLmilri
20 40 60 80 100 120
(Exposure time at 760°C in air, 8) 1' 2
= 0.5-2.5 mm from pin, ap= at pin, bp=both pins
nc = near center, _s = fixture slipped, restarted
specimens from top (italics) bottom of casting

 

 

 

 

 

 

Figure 28. Effect of salt exposure on fracture strength

_"’I*1 TIF“1‘T‘1'T‘1F‘R,
Caps 4—point bending of TiAl
7 O MOR,nosalt i
i .np 9m 0 MOR with salt
1% . . avg MORwsalt

slip ’

npi

np

 

 

 

 

 

 

 

 

 

iii‘irrriii'iiii'fiifiii
; Best 4-point bend data of TiAl
)— _i
F MOR, Salt F
. . g3; np 0 MOR, NoSalt
~ MOR, A\g

 

 

 

40 60 80 1 00 1201 2
(Exposure time at 760°C in air, 3) '

20

140

Figure 29. Fracture strength of specimens that had no slippage and not at a pin.

62

 

Figure 30. 0.36ks specimen fracture surface - oxidized pit.

 

Figure 31. 19.44ks specimen fracture surface - remnant of spalled oxide.

63

4.5 SALT APPLICATION ON PRE-OXIDIZED SPECIMENS

The 3.6ks pre-oxidized specimens developed a discontinuous oxide layer. The
oxide continued to grow after each cycle of salt and heat exposure and could be readily
observed using light microscopy. Analysis by EDS revealed the presents of Na but no
Cl, this trend continued for all cycles. The 3.6ks specimens developed surface
morphologies similar to that shown in the upper right-hand comer of Figure 21. There
were no obvious signs of salt attack unlike that which was observed on the bare
specimens. Figures 32 and 33 show optical micrographs of the 3.6ks pre-oxidized
specimens, the darker areas are oxide growth.

The 360ks pre-oxidized specimens developed a continuous oxide layer. The
surface morphology of the oxide layer did not change afier salt and heat cycling. There
were particles observed on the oxide surface by ESEM, and the particles contained traces
of Mg, Ca and C] which are all constituents of the sea salt. Figure 34 depicts to general
oxide covering on the 360ks pre-oxidized specimens, the particle shown on the surface of
the oxide is a possible contaminant from the furnace.

4.5.1 Interrupted Weight Gain

Weight gain results for 3.6ks pre—oxidized specimens are shown in Figures 35.
The slope is positive and reproducible for specimens that had salt applied and is zero for
the specimen with no salt application. A photograph of the specimens after ten cycles is
shown in Figure 36. It is obvious that NaCl stimulates oxide growth by the dark

discoloration of the surfaces where the salt was applied in Figure 36.

64

‘1 1,14.

_ _ . f.» l J
.c- ' . 14,33 ~ 1.“ ‘1. *‘tb, k,“
mm» c» um» , -...
‘ a ' . . ..- , ’

‘4‘ . .

 

 

Figure 32. Pre-oxidation specimens after 1st Cycle - 3.6ks specimen

65

 

 

Figure 33. Pre-oxidation specimens afier 10 Cycles - 3.6ks specimen.

66

 

 

360ks specimen

Cycle

imens afier 1"

Figure 34. Pre-oxidation spec

67

 

 

 

 

 

 

   
   
 

 

2.5 i l 1 1 r { 1 1‘
’, +Salt —r—— Half_2 1
N" 2 r . Half_1 No Salt ,
E r — : I i
5, 15 L Salt applied on Ti-48AI-2Cr-2Nb .3, ,l’ q
3 Y; 1.08 ks 760 C exposures 1
{B i l
1 . 3
g i l E 11
.E 1 ‘ I .1
5 0.5 .3
E ' i
i ’ ;
'5’ o fili/ ' I . j
i , . ‘ {
‘t ‘ i i i I 1
-1 . , , ‘
° 1 2 3 4 5 6 7 8 9 1o
Cycle Number
Figure 35 . Plotted weight gain results.

  

 

I

SALT. l-IALF,

  

   

    

F 110 "ALT“

Figure 36. Weight gain photographs.

68

4.6 REACTIONS AND PRODUCTS
4.6.1 DTA

Differential thermal analysis of the casting filings and salt mixtures indicate that
reactions occur at about 615 0C, 675 °C, 780 °C and 810 °C. These results are shown in
Figure 37 where the reactions are labeled with an R All the reactions are endothermic
(DTA convention is that endothermic responses are indicated by the downward peak
[54]) and start to occur at the label. The standard used was 99.999% pure Al, which has a
melting point of 660 °C. The melting of the standard appears to be masked by reactions
occurring in the mixtures. Similar results were obtained with the mechanically alloyed
TiAl powder except that a reaction occurred at about 650 °C and between 720 0C to 730
°C as shown in Figure 38.
4.6.2 XRD

The x-ray diffraction performed on the 90wt% TiAl filings: 10wt% salt identified
many products resulting from the heating of the mixture during DTA. Table 5 lists all
products that may have formed, the 1St through 4m 20 angles are the diffraction angles
identified by a peak finder program that match an element or compound. The plotted x-
ray scan is shown in Figure 39 and the labeled peaks correspond with the numbers in
brackets shown in Table 5. XRD on the 10wt% TiAl filings:90wt°/o salt identified two

products, salt and y-TiAl as shown in Table 6, and Figure 40 shows the plotth x-ray scan

where salt is labeled S and y-TiAl is labeled y.

69

1.5

 

1' I T l 1

Ti-48-2-2 and Salt in N2

     

 

1
4
g 0.5
a
(D
O.
E o
(D
i— - .2 —+— 1OTIAI-1
_ —u— 10TiAl-2
~0-5 ” ‘ —— 90TiAI-1
_ .,,L_ 90TIAI-2
_1 .7 7”, .fi. ...iL ,1. 4.. ,1 _L.1, .1, 1.7.1._E "1.“, m-_u_...1.,._u ..... L ..1 ...1
600 650 700 750 800 850

Temperature Difference
O
.m .

.<'=
01

Temperature(degnees Celcius)

Figure 37. DTA of casting filings and salt.

 

 

 
     

"""" “j’Maehaiiicauyaiayea TiAl y
R and Salt in: N2 ‘ R , 7

    

—+—— 90TiAl-1
+ 90TIAI-2
“—47-— 10TiAl-1
-—"—— 10TIAl-2

 

2.... ., i 2 I L _1.. A, ‘- . .4. l J .1L_L_.r., 1 .71.. 1.. 1 1.. 4 .__1
650 700 750 ,
Temperature (degrees Celcrus)

Figure 38. DTA of mechanically alloyed TiAl and salt

70

Table 5. Products identified by XRD on 90wt% TiAl filings: 10wt% salt DTA powder

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Species strongest 2nd 3rd 4th
20 peak strongest strongest strongest
20 peak 20 peak 20 peak
TiO; (212) 27.474 [1] 36.170 [5] 54.257 [23] 68.973 [29]
TixCryOOH (154) 27.474 [1] 36.290[6] 35.847 [4]
Ti3Al (91) 54.257 [23] 41.264 [11] 72.182 [31]
Nb4.62N2.l4 (60) 36.170 [5] 38.74319]
n-Cr4Ti4O (56) 41.264 [11] 43.452 [14] 42.261 [12]
NbN0_6Oo.2 (52) 41 .264 [11] 35.847[4]
AlTi3 (51) 41.264 [11] 72.182 [31]
szN (47) 35.847 [4] 33.845 [2] 38.410 [8]
NbTi4 (40) 54.748 [25] 39.127 [10] 42.261112]
Cr(39) 39.127 [10] 44.253 [15] 48.53 [18]
NazOz (36) 35.115 [3] 50.872 [21]
TiAl (30) 38.743 [9] 45.581 [16]
TiO (30) 43.152 [13] 43.452 [14]

 

 

 

 

 

 

Note: The number in square brackets is the peak reference number in order of increasing

20 and the number in parenthesis is the total intensity of all peaks identified for the

species determined from the Scintag software peak finder (Figure 37).

Table 6. Products identified by XRD on 90wt% TiAl filipgs:10wt°/o salt DTA powder

 

 

 

 

Species strongest 20 2nd strongest 20 3rd strongest 20 4th strongest 20
peak peak peak peak

TiAl 21.20 45.68 65.72

NaCl 31.689 45.448 56.523 27.374

 

 

 

 

 

71

 

 

1000

800

600

A Intensity

00

200

 

O

 

T T 1 l rT TITIT YYTITTIT [TlrrrTfirlIl

90wt% TiAl Filings
10wt% Salt

 

 

 

 

 

 

ill—LLJJILLIJIILIIJ J_.Lllll_llllllllll

 

 

 

20

30 40 50 60 70 80
2 0

Figure 39. XRD scan of 90wt% TiAl filings: 10wt% salt DTA powder.

72

90

my

Intens

 

 

 

 

 

 

SOOOLT 1v r1 ... r1 rr .. 111 .. 1‘ ., .1 ‘11 r1 .. :r_
. s -
L -
2500 :' 10wt% TiAlCasfing Filings L
1500E- i
l 1
1000;— 1
iv 1
. S .
500 S S -
+— I 'Y :
0 b.1114 . l .11 .1 ll ....11 l l r 1 l 1.1
20 30 40 50 60 7o 80 90
2 9

Figure 40. XRD scan of 90wt% TiAl filings: 10wt% salt DTA powder.

73

CHAPTER 5
ANALYSIS AND DISCUSSION
5.1 INTRODUCTION

Reactions between the alloy, Ti-48Al-2Cr-2Nb, and NaCl, when exposed to high
temperatures, appear to be detrimental to the alloy by preferential attack of the a; phase
if it is not pre-oxidized. It is apparent that the reaction kinetics between the bare metal
and NaCl occur at a much greater rate than the reaction kinetics between the alloy’s
natural oxide and NaCl.

Two forms of oxidation were observed, a smooth surfaced and dense form that
increased in thickness with time, and heterogeneous eruption of needle and plate shaped
crystals the grew quickly from a point on the smooth oxide surface. The presence of salt
appeared to stimulate the growth of the latter type, which is evident in Figures 20 and 32,
and is clearly not a protective oxide. From ref. [16], the latter type of oxide is fast growth
of TiOz. At temperatures greater that 700°C, grth of TiO; is faster than A1203. At
900°C in an air, an alloy with activities of 1 for Ti and Al, A1203 is slightly more stable
than TiOz. Changes in activities to either species due to the introduction of another
species (i.e., Cr, Nb, etc.) can alter this stability. However, in experiments, A1203
formation is often observed to occur in the beginning of the oxidation process, leading to
depletion in Al and a Ti enrichment [16]. This may account for the low EDS values for
aluminum, where the concentration is low at the metal and oxide surfaces as seen in

Figure 9a.

74

5.2 ELEMENTAL TRENDS

EDS measurements were used to determine elemental trends. Each EDS scan was
standardized to the total titanium counts detected during that scan. There were high
concentrations of Na detected on the oxide surface, and low concentrations at metal
surfaces and interfaces (Figure 8a). The surface Na concentration increased with time.
The vapor pressure of sodium at 760°C is about 230 torr. (0.3 atm.), so Na evaporation is
possible. Sodium ions are known to transport through alumina, and the sodium may also
react with alumina to form NaxAlOy [15]. Exposure to NaCl at 760 °C may lead to the
formation of Na5A104, in a manner similar to the equations in the introduction. However,
x-ray diffraction experiments indicate the presence of Na202 (Figure 37 and Table 5) that
was formed by a reaction(s) between Na and oxides that were present on the casting. The
probability of Nazoz formation is low, because there are only two characteristic peaks of
low intensity and the low availability of oxygen. There is a higher probability that TiOz
forms since it is the more stable oxide. The availability of oxygen in the system is
limited because the DTA powder used was heated in a nitrogen atmosphere (100ml per
minute purge was used). Oxygen may have come from either previous formed oxides on
the casting surface or interstitial oxygen.

Chlorine concentrations are high in and around pitted areas, and surrounding oxides,
especially at early times (the 0.15 and 0.36ks specimens in Figure 8b). Chlorine ions can
act as a catalyst for oxide growth on bare metal specimens [16] since chlorine tends to
stay on the metal surface until sufficient oxide has developed. At longer exposure times
the Cl concentrations decrease toward zero at the surface (Figure 8a) and increases at the

interfaces and oxide layers. This could happen due to the higher vapor pressure of Cl that

75

allows it to escape, or a dilution effect as the oxide grows. X-ray diffraction analysis of
the 10wt% salt:90wt°/o TiAl powder mixture detected no chlorine containing compounds
present, the only chlorine containing compound detected in the 90wt% salt:10wt% TiAl
powder mixture was NaCl.

The oxygen concentration increases as exposure time increases (Figure 9b), which is
expected since titanium and aluminum oxides layers grow thicker with time. The
observed relative decrease in aluminum concentrations at the oxide surface is consistent
since TiO; is known to grow faster and occupy the outer oxide layer [10]. However, the
decreased aluminum concentration at the metal interface implies that either aluminum
atoms evaporate, or that alumina forms in locations that are not as easily detected by EDS
(e.g. in the oxide close to the metal surface, [10]

There are higher concentrations of Cr and Nb at and around damaged or pitted areas
(Figure 10), where much greater amounts of Ti and Al were consumed to form oxides.
Nb is added to TiAl alloys to enhance their oxidation resistance by formation of szOs
within the TiOz/Alzog/szos oxide layer [37]. The exposure to salt precluded the
formation of szOs oxide, but salt may have catalyzed the formation of NbTi4, szN and
NbNxOy (Table 5). Cr additions are detrimental to oxidation resistance and are only
added to improve ductility. The exposure to hot corrosion by NaCl may have decreased
the ductility due to the formation of Cr4Ti4O (Table 5).

Nitrogen was not detected in any of the EDS scans. Rakowski detected TiN in the
oxide interfacial region of specimens exposed at 800 — 900 °C in air and atmospheres
containing various amounts of nitrogen [31], where the presence of TiN interferes with

the formation of a protective oxide layer. The volume fraction of TiN was not reported,

76

so if it is small, it may not be detectable. This alloy will form TiN within its oxide layers
afier long time exposure (9000 hours) at 704 °C in air [33]. XRD detected nitrogen
compounds (Table 5) where the nitrogen reacted with niobium in lieu of titanium.
5.3 STRENGTH OF EXPOSED SPECIMENS

There is no significant effect of accelerated oxidation on the fracture strength due
to salt and heat exposure for short times on bare ground specimens. The presence of
multiple initiation sites indicates that flaws due to local attack were below the critical
crack length. The variability in the fracture strengths may be attributed to variations in
local orientation conditions in regions of maximum stress and/or imperfect loading
conditions. Only the longest exposure time showed a decrease in the scatter in strength,
but these fracture strengths were still above the yield stress [3] and above the lowest
values obtained in specimens with short or no exposure (Table 3, Table 4 and Figure 27).

The variation in strength of the specimens can also be correlated to imperfections
in the loading conditions or specimen dimensions. Imperfect loading led to conditions
more similar to a three-point bend test than a four-point bend test. The three-point bend
test condition may be inferred due to the fact that the specimens with the highest strength
fractured at or near the contact point between the upper pins and the specimen. During
this test condition a smaller volume of material was subjected to the maximum stress, and
the probability of a flaw or a microstructural condition that stimulates fracture initiation
is lower. The lower fracture strengths correlate with fracture positions farther from the
pin locations, this is typical for the four point bend test condition where a larger volume
of material was subjected to a more uniform stress state. During this test condition the

probability of a flaw or a microstructural condition that stimulates fracture initiation is

77

 

higher which implies that damage due to salt attack did not create surface damage severe
enough to become critical in the regions between the pins.

The lack of the oxide on the fracture surfaces may be attributed to either the effect
of salt in making the oxide non-adherent, or spallation due to the release of strain energy
during fracture. Also, there is a thermal expansion mismatch between the oxide and alloy
that could promote de—bonding at the interface. The oxide may have spalled off or de-
bonded prior to fracture (as in the case of the 1.08ks specimens). The oxide that
developed on specimens exposed to salt appears to enhance the strength up to exposure
times of 1.08ks, but thermal shock may have caused the oxide to spall in the conditions of
this experiment (Figure 29), so thermal shock may degrade the oxide strength with longer
exposure times. An investigation using slower cooling rates is necessary to evaluate
whether the degraded strength is due to thermal shock or the nature of the accelerated
oxide growth.

It is apparent that pitting and preferential attack of (12 occurs on bare ground surfaces
of the alloy during short exposure times in the presence of salt, but this damage is not
severe. However, the oxides stimulated by salt are not adherent. Since the oxides layers
that form during hot corrosion by NaCl are non-protective (i.e., not continuous), there is a
possibility for hydrogen embrittlement or oxygen absorption (on-case embrittlement) to
occur and degrade the alloys strength [1 , 34, 42]. It is not clear whether the non-adherent
oxides will fall ofi in engine conditions, taking the salt with it and subsequently allowing

a more protective oxide to develop before further exposure to salt occurs.

78

5.4 EFFECTS OF PRE-OXIDATION

Pre-oxidizing is beneficial to limit or prevent surface damage caused by hot
corrosion by NaCl on this alloy. Pre-oxidation prevented spalling upon removal from the
furnace, and the surface oxides were not easily removed with a pick. The presence of
sodium but not chlorine in EDS measurements suggests that the Cl evaporates. Short
pre-oxidation times result in a discontinuous oxide consisting of thin, if not transparent
oxide, and regions where platelets of TiO; emanate from a point region to form a porous
oxide on the surface. With increasing time these more porous oxide features cover a
larger fraction of the specimen.

Weight gain by this oxidation mechanism is accelerated with the application of
salt, as shown in Figures 33 and 34. These results agree with similar weight gain results
obtained by Yao and Marek [15] on a Ti-47Al-2Nb-2Mn + 7vol% TiBz alloy that had
NaCl-induced hot corrosion. Their study shows that salt increases the slope of initial
oxidation kinetics. Afier an incubation period the oxidation kinetics of the salt exposed
specimens switch from linear to parabolic similar to other TiAl alloys which have not
been exposed to salt [10, 16, 31, 33, 37]. Thus salt apparently stimulates formation of
rapid growing titanium oxide evident in the periphery of Figure 32. For the longer pre-
oxidation time of 360ks, NaCl has no obvious effect on the oxide layer, which had the
fast growing porous TiO; oxide layer on the surface. It is clear from these experiments
that salt stimulates the formation of these porous oxides that cause rapid weight gain from
a pre-existing smooth and dense oxide. However, such oxides also form on specimens

that have no salt exposure.

79

The direct application of salt to bare TiAl alloys causes growth of a non-
protective scale, but it is not known whether the application of salt on an oxidized surface
will degrade a protective layer. Thus, further study is needed to quantify the effect that
salt has on the grth rate of the porous oxide on an existing oxidized surface, and the

relationship that this rapid growth has on the metal/oxide interface.

80

 

CHAPTER 6
CONCLUSIONS AND RECOMMENDATIONS

6.1 CONCLUSIONS

A method to apply salt water to specimens that is similar to natural conditions near
sea coasts was developed and used to apply salt to bare ground and pre-oxidized
specimens of Ti-47.9Al-2Cr-2Nb alloy. After exposure in a 760°C air furnace for times
between 0.15 - 20 ks, the salt stimulated rapid growth of non-protective porous oxides on
specimens with bare ground surfaces. Pitting that attacked the 0:2 (Ti3Al) phase
preferentially was evident in regions where the non-protective oxide was spalled off.
However, such pits had no apparent effect on fracture initiation in 4-point bending.

Pre-oxidizing the surface prevented the spalling of the oxide observed in the bare
ground specimens. From EDS measurements, Cl was most ofien found in the metal-
oxide interface when salt was directly applied to bare ground specimens, and Na was
most often found on the surface of the oxide on both the bare ground and pre—oxidized
specimens.
6.2 RECOMMENDATIONS FOR FUTURE RESEARCH

Further investigations are needed to determine if a non-adherent oxide will fall off in
engine operating conditions, taking the salt ions with it, and if a protective and adherent
oxide will grow in its place. Also, the longer term effects of accelerated oxide growth
due to salt need to be examined to determine whether condensation of water containing
sea salt will have any detrimental effect on the long term stability of protective oxides.

The oxide layer that forms during hot corrosion consists of many different products.

These products might form at different areas within the oxide layer. Exposure time might

81

influence the location of the products within the oxide layer. Results from this study
indicate that fracture strength (modulus of rupture) reaches maximum values at about
1.08ks (18 min), and that reaction products between salt and the alloy could reside
within the oxide layer (Table 5). Previous oxidation studies (performed in air or 02) on
titanium aluminides [31, 33, 37] have identified Nb205 and TiN as reaction products
present in the oxide layers. It would be beneficial to determine if there is a correlation
between the reaction products, their location in the oxide layer, and fracture strength. X-
ray photoelectron spectroscopy (XPS) may be utilized to perform this identification
process by examining a cross-section of the oxide layer, and determining if bonding
within the oxide changes with exposure time and position.

Another research path that might elucidate strength degradation caused by hot
corrosion is to perform in-situ testing of a stressed sample within an ESEM that provides
an atmosphere of appropriate gases. Advantages to performing this type of testing are:
the growth kinetics of the oxide can be viewed and recorded; its possible to determine a
rupture stress at exposure temperatures; to see how surface damage could affect elevated
temperature deformation phenomena such as creep crack growth, by monitoring stress
relaxation or strain with constant loading conditions. A main disadvantage is the
difficulty controlling the temperature of the specimen without causing too much localized
heating or burning out the heating device. Such in-situ testing was attempted but
problems were encountered with heater controller and the integrity of the heating
element. The limited space within the tensile test fixture of the ESEM and the size of the

specimens presents a challenge to designing the heating device.

82

An investigation using slower cooling rates could determine if there is a correlation
between fracture strength and thermal shock or the nature of the oxide growth. Also,
longer cyclic studies on specimens with thicker pre-oxidation oxide layers can elucidate
if thicker oxides are more protective, this will be evident if thicker pre-oxidation layers
gain less weight than thinner pre-oxidation layers. Longer cyclic studies will determine
the rate kinetics of accelerated oxidation, i.e., if the behavior turns parabolic, cubic or
continues to be linear. The alloy studied by Yao and Marek [15] switched from a linear
behavior to a parabolic behavior similar to other TiAl alloys which had no salt exposure
[10,16, 31, 33, 37].

A very important aspect when studying hot corrosion is to simulate, as closely as
possible, conditions within an engine. Burner rigs can accomplish this but they are not
simple or economic, especially in an academic setting. A proposed way of simulating
engine conditions is heating an exposed specimen within an alumina vessel, while
controlling a flow of engine gases through the vessel. The main problem with this

approach is handling the engine gases, which can be toxic.

83

 

10.

ll.

12.

BIBLIOGRAPHY

Kim, Y-W. and F roes, F.H. “Physical Metallurgy of Titanium Aluminides.” High
Temperature Aluminides and Intermetallics. The Minerals, Metals & Materials
Society (1990): 465-492.

Lipsitt, H.A., Shechtrnan, D. and Schafiik, RE. “The Deformation and Fracture of
Ti3Al at Elevated Temperatures.” Metallurgical Transactions A. 11A (1980): 1369-
1375.

Kim, Y-W. “Intennetallic Alloys Based on Gamma Titanium Aluminide.” JOM. July
(1989): 24-30.

Kim, Y-W and Dirniduk, D.M. “Progress in the Understanding of Gamma Titanium
Aluminides.” JOM. August ( 1991 ): 40-47.

Austin, CM. and Kelly, TJ. “Development and Implementation Status of Cast Gamma
Titanium Aluminide.” Structural Intermetallics. The Minerals, Metals & Materials
Society (1993): 143-150.

Larsen, J .M., Williams, K.A., Balsone, SJ. and Stucks, M.A. “Titanium Aluminides
For Aerospace Applications.” High Temperature A luminides and Intermetallics. The
Minerals, Metals & Materials Society (1990): 521-555.

Nishiyama, Y., Niyashita, T., Susumu, 1. and Noda, T. “Development of Titanium
Aluminide Turbocharger Rotors.” High Temperature Aluminides and Intermetallics.
The Minerals, Metals & Materials Society (1990): 557-584.

Yamaguchi, M. “High Temperature Intermetallics — With Particular Emphasis on
TiAl.” Materials Science and T echnology Vol. 8 April (1992): 299-307.

Misra, A.K. “Reaction of Ti and Ti-Al Alloys With Alumina.” Metallurgical
Transactions A. 22A (1991): 715-721.

Tani guchi, S. and Shibata, T.. "Influence of Additional Elements on the Oxidation
Behavior of TiAl." Intermetallics. 4 (1996): 885-893.

Rapp, R.A., “Chemistry and Electrochemistry of the Hot Corrosion of Metals”,
Corrosion. Vol. 42 No. 10 (1986): 568-577.

Failure Analysis and Prevention,4th ed., ASM International, Metals Park, Ohio, 1992,
pp.280.

84

 

l3.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

ASM. Metals Handbook Desk Edition. 7‘h ed., Metals Park, Ohio:ASM
International, pp. 9010, 1992.

ASM. Metals Handbook Desk Edition. 7‘h ed., Metals Park, Ohio:ASM
International, pp. 1033, 1992.

Yao, Z. and Marek, M., “NaCl-Induced Hot Corrosion of a Titanium Aluminide
Alloy.” Mater. Sci. & Eng. A192l193 (1995): 994-1000.

Schutze, M. and Hald, M., “Improvement of the Oxidation Resistance of TiAl Alloys
by Using The Chlorine Effect.” Materials Division, Karl-Winnacker-Institut der
DECHEMA e. V., D-60061 Frankfurt am Main.

S. Tani guchi, Materials and Corrosion 48 ( 1997): 1-9.

Liu, C.T., “Recent Advances in Ordered Intermetallics.” Materials Research Society
Symposium Proceedings, Vol 288 (1993): 3-19.

ASM. Alloy Phase Diagrams. ASM Handbook Vol. 3., Metals Park, Ohio:ASM
International, pp. 2054, 1992.

Wright, P.K., “On Ductility and Toughness Requirements for Intermetallics in
Aircraft Engines.” Structural Intermetallics, Materials Research Society (1993): 885—
893.

Yamaguchi, M. and Inui, H., “TiAl Compounds For Structural Applications.”
Structural Intermetallics, Materials Research Society (1993): 127—142.

Seo, D.Y., Bieler, T.R., An, S.U. and Larson, D.E., “Changes in Microstructure
During Primary Creep of a Ti-47Al-2Nb-1Mn-0.5W-0.5Mo-0.28i Alloy.”
Metallurgical Transactions A. 29A (1998): 89-98.

Fontana, M.G. Corrosion Engineering. New York: McGraw-Hill, 1986.

Moretti, G., Quartarone, G., Tassan, A. and Zingales, A., “Pitting Corrosion Behavior
of Superferritic Stainless Steel in Waters Containing Chloride.” Werkstofie and
Korrosion. 44 (1993): 24-30.

Gregory, J .K. and Brokmeier, H.G., “The Relationship Between Crystallographic
Texture and Salt Water cracking Susceptibility in Ti-6Al-4V.” Mater. Sci. & Eng.
A203 (1995): 365-372.

Ricker, R.E., Hall, DE. and Fink, J .L., “The Effect of Aqueous Environements on the

Fracture Behavior of Ductile Nickel Aluminide.” Scr. Meta/l. Mater. 24 (1990) 291-
296.

85

 

 

27. Bavarian, B., Hartutouni, S. and Zamanzadeh, M., “Evaluation of the
Environmentally assisted Cracking of Aluminide Interrnetallic Compounds.” Mater.
Sci. & Eng. A153 (1992): 613-618.

28. Ricker, R.E., “Origins of the Aqueous Corrosion and Stress Corrosion Cracking
Behavior of Ductile Nickel Aluminide.” Mater. Sci. & Eng. A198 (1995): 231-238.

29. Meier, G.H., Pettit, RS, and Hu, 8., “Oxidation Behavior of Titanium Aluminides.”
J. De Physique I V, Colloque C9 (1993): 395.

30. Cole, 1.8. And Ganther, W., “A Preliminary Investigation into Airborne Salinity
Adjacent to and Within the Envelope of Australian Houses”, Construction and
Building Materials Vol. 10 No. 3 (1996): 203-207.

31. Rakowski, J .M., Pettit, F.S., Meier, G.H., Dettenwanger, F., Schumann, E., Ruhle,

M., “The Effect of Nitrogen on the Oxidation of y-TiAl.” Scripta Metal]. Vol. 33 No.
6, (1995): 997-1003.

32. Rahmel, A. and Spencer, P.J., “Thermodynamic Aspects of TiAl and TiSiz Oxidation:
The Al-Ti-O and Si-Ti-O Phase Diagrams.” Oxidation of Metals Vol. 15 No. 1/2,
(1991): 53—68.

33. Brady, M.P., Brindley, W.J., Smialek, J .L. and Locci, I.E., “The Oxidation and
Protection of Gamma Titanium Aluminides.” JOM. August (1996): 46-50.

34. Du, H.L., Datta, P.K., Lewis, DB. and Bumell-Gray, 1.8., “Air Oxidation Behaviour of
Ti-6Al-4V Alloy Between 650°C and 850°C.” Corrosion Science Vol. 36 No. 4, (1994):
631-642

35. CRC. Handbook of Chemistry and Physics. 79th ed., Boca Raton, FL:CRC Press
Inc., 1998.

36. Tsuyama, S., Mitao, S. and Minakawa, K., “Alloy Modification of y-Base Titanium
Aluminides for Improved Oxidation Resistance, Creep Strength and Fracture
Toughness.” Mater. Sci. & Eng. A153 (1992): 451-456.

37. Kim, B.G., Kim, GM. and Kim, C.J., “Oxidation Behavior of TiAl-X (X=Cr, V, Si,

Mo, Nb) Intermetallics at Elevated Temperatures.” Scripta Metall. Vol. 33 No. 7,
(1995): 1117-1125.

38. Partridge, A. and Winstone, M.R., “Oxidation Behavior of Highly Alloyed y-TiAl
Alloys.” Structural Intermetallics, Materials Research Society (1997): 405-410.

39. Streiff, R. and Poize, S., “Oxidation of Aluminide Coatings on Unalloyed Titanium.”
High Temperature Corrosion, ed. Rapp, R.A. Houston, TX: NACE (1983): 591—597.

86

 

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

Zhang, Y.G., Li, X.Y., Chen, C.Q., Zhang, X.J., Zhang, TH. and Zhang, H.X. “The
Influence of Ion Implantation Upon High Temperature Oxidation Behavior of Ti-
483t.%A1 Alloy.” Structural Intermetallics, Materials Research Society (1997): 353-
360.

Kawaura, H., Nishino, K. and Saito, T., “New Surface Treatment Using a Fluidized
Bed Furnace For Improving Oxidation Resistance Of TiAl-Based Alloys.” Structural
Intermetallics, Materials Research Society (1997): 377—3 83.

McKee, D.W. and Huang, SC, The Oxidation Behavior of Gamma-Titanium
Aluminide Alloys Under Thermal Cycling Conditions.” Corrosion Science, Vol. 33
No. 12 (1992): 1899-1914.

Rapp, R. A. and Zhang, Y.S., “Hot Corrosion of Materials: Fundamental Studies.”
JOM. August (1994): 47-55.

Bomstein, NS. and DeCrescente, MA, “The Role Of Sodium and Sulfur in the
Accelerated Oxidation Phenomena - Sulfidation.” Corrosion, Vol. 26, No. 7 (1970):
209-214.

Goebel, J .A., Pettit, F8. and Goward, G.W., “Mechanisms for the Hot Corrosion of
Nickel-Base Alloys.” Metallurgical Transactions, Vol. 4 (1973): 261—278.

Deb, D., Ramakrishna Iyer, S. and Radhakrishnan, V.M., “Assessment of High
Temperature Performance of a Cast Nickel Base Superalloy in Corrosive
Environment.” Scripta Materialia, Vol. 35 N0. 8 (1996): 947—952.

Gaskel, D.R., Introduction to Metallurgical Thermodynamics. 2nd ed. New York:
Hemisphere Publishing Corporation, 1981.

Hiramatsu, N., Uematsu, Y., Tanaka, T. and Kinugasa, M., “Effect of Alloying
Elements On NaCl-Induced Hot Corrosion of Stainless Steels.” Mater. Sci. & Eng.
A120 (1989): 319-328.

Fujikawa, H. and Maruyama, N., “Corrosion Behavior of Austenitic Stainless Steels
in the High Chloride-Containing Environment.” Mater. Sci. & Eng. A120 (1989):
301-306.

Saunders, SR]. and Nicholis, J .R., “Hot-Salt Corrosion Testing — An International
Intercomparison.” Materials at High Temperatures, Vol. 13, No. 3 (1995): 115—120.

Chaze, AM. and Coddet, C., “Influence of Aluminum on The Oxidation Of

Titanium Between 550 and 750 °C.” Journal of the Least Common Metals, 157
(1990):55—70.

87

52. Du, H.L., Datta, H.L., and Hwang, S.K., “High Temperature Corrosion Behavior of
Ti—46.6A1-1.4Mn-2Mo in Environments of Low Oxygen and High Sulfur Potentials
at 700 °C and 900 °C.” Materials Science Forum, Vol. 251-254 (1997): 219—226.

53. Saitoh, Y., Asakawa, K. and Mino, K., “Embrittlement of Titanium Aluminides By
Slight Oxidation.” Aspects of High Temperature Deformation and Fracture in
Crystalline Materials Proceedings of the 7th JIM International Symposium (1994):
583—5 89.

54. Brown, Michael E., Introduction to Thermal Analysis Techniques and Applications.
New Yorszhapman and Hill, 1988.

88