Sha W., Malinov S. Titanium Alloys: Modelling of Microstructure, Properties and Applications

Подождите немного. Документ загружается.

Differential scanning calorimetry and property measurements 85

material’s ductility, is the opposite, namely the elongation increases, with

higher temperature of ageing (Fig. 4.16). Obviously, the ageing temperature

is a very significant parameter for control of the microstructure and properties

in β-titanium alloys. By altering this temperature, the desirable combination

of strength and ductility can be achieved.

100 µm

4.13

Microstructure of β21s alloy after water quenching and ageing at

480 °C for 8 hours.

Aged Water Ageing at Ageing at Ageing at

(starting state) quenched 480 °C 540 °C 595 °C

Vickers hardness (HV)

370

360

350

340

330

320

310

300

290

280

270

4.14

Vickers hardness of β21s alloy with different processing

conditions.



Titanium alloys: modelling of microstructure86

4.3 Effects of hydrogen penetration

Titanium alloys are susceptible to hydrogen, which causes embrittlement,

leading to the deterioration of the properties of the alloys. For a Ti-42Al-

11Nb (at.%) alloy, the yield strength increases with increasing amount of

hydride, but the ultimate tensile strength, ductility and fracture toughness

decrease. Therefore, the amount of hydrogen that a titanium alloy can absorb

during service is a major measure of the ability of the alloy to retain good

properties. Much effort has been made to quantify the hydrogen susceptibility

of titanium alloys. This section describes the effect of hydrogen.

Confusion exists concerning the form of hydrogen after entering titanium

alloys. It was suggested that a hydride with the structure (TiAl)H

might

form. Samples of Ti-49.9Al (at.%) were ground, polished and cathodically

480 540 595

Ageing temperature (°C)

Tensile strength (MPa)

1350

1300

1250

1200

1150

1100

1050

1000

4.15

Tensile strength of β21s alloy after ageing at different

temperatures.

480 540 595

Ageing temperature (°C)

Elongation (%)

4.16

Elongation of β21s alloy after ageing at different temperatures.



Differential scanning calorimetry and property measurements 87

charged in a 1 mol/l NaOH and 250 mg/l As

solution at room temperature

with a current density of 300 A/m

. X-ray diffraction showed that for a

sample with a single γ phase before charging, a hydride phase was formed

after 4 hours of charging. The hydride phase was determined to be (TiAl)H

0.5

with tetragonal lattice parameters a = 0.450 nm and c = 0.327 nm (c/a =

0.727).

The surface of a γ and α

two-phase titanium aluminide sample was

covered with a black layer after prolonged cathodic hydrogen charging.

Analytical transmission electron microscopy showed that the surface layer

was a hydride based on (TiAl)H

. It was determined that the hydride had a

tetragonal crystal structure with lattice parameters a = 0.452 nm and c =

0.326 nm (c/a = 0.721), and it was present up to 25 µm in depth from the

surface of the sample after charging for 2 hours. The charging condition was

a 5% H

solution, and current density equal to 5000 A/m

. The weight

change of the sample with increasing charging time was also measured. The

weight decreased with increasing charging time after 1 hour.

Hydrogen charging of Ti-42Al, Ti-45Al and Ti-50Al induced crack formation

after a short charging time, while additional charging produced pits within

the γ phase in the γ and α

two-phase coexisting grains. However, no damage

was seen in the α

phase or equiaxed single α grains.

Cast Ti-48Al-2Cr and wrought Ti-46.5Al-4(Cr,Nb,Ta,B) gamma titanium

aluminides were cathodically charged with hydrogen for various times to

study possible hydrogen trap sites and the nature of the hydride being formed

(Sundaram et al., 2000). The charging solution was 1

N H

solution with

1 g/l thiourea. The charging time varied between 1 and 24 hours and the

current density was 1 A/m

. No visible change was noted on the charged

surface after 8 hours as compared with the uncharged surface. However,

after 16 hours, small black particles, presumably hydrides, were observed on

the surface. After 24 hours, a black hydride layer covered the surface. After

X-ray diffraction patterns were obtained, Sundaram et al. (2000) stated that

the result proved to be similar to the previous findings.

Sundaram et al. (2000) determined that a hydride was formed when hydrogen

entered their samples. However, this does not discount the possibility that

some hydrogen may occupy the interstitial sites in the alloys. Both could

take place simultaneously.

The following summarises the further investigation into the hydrogen

susceptibility of β21s and Ti-46Al-1.9Cr-3Nb alloys, by measuring the amount

of hydrogen penetrating into the sample within a specific charging time.

Establishing a rate of penetration can give an indication of the ability of the

alloy to resist hydrogen.

Immediately after charging, some of the samples were placed in a solution

of glycerine for 48 hours at room temperature. This technique had been used

before, and the quantity of hydrogen released during 48 hours at room



Titanium alloys: modelling of microstructure88

temperature was used as a measure for the quantity of hydrogen absorbed

during permeation. After such experiments, no hydrogen was released and

collected after 48 hours. The hydrogen remained within the samples.

The average Rockwell hardness numbers before and after charging were

exactly the same. Some of the samples cracked, however, on using the Rockwell

hardness indentation after charging.

The effects of hydrogen on the samples could be clearly seen after the

experiment, especially after severe charging, such as 4 hours at 85 °C in

2.8% H

at 250 A/m

. The samples had suffered extensive corrosion

while the remainder was covered with a black layer. Sundaram et al. (2000)

reported similar damage.

The additional peaks present in the X-ray diffraction patterns for charged

samples indicate the presence of a hydride (Fig. 4.17). The plane index

values corresponding to the peaks in the diffraction patterns were determined

by comparison with data by Sundaram et al. (2000), as the hydride peaks

established here are similar to XRD data of hydrides available from these

authors. The intensity of the peaks is summarised in Table 4.3. The fluctuation

in the total integrated intensities of the peaks in the XRD patterns is, most

likely, due to the inevitable positional variation in the experimental set-up

for different samples.

4.17

X-ray diffraction patterns for Ti-46Al-1.9Cr-3Nb alloy. (a)

Uncharged; (b) charged for 1 hour at room temperature, 250 A/m

5% H

solution; (c) charged for 4 hours at 85 °C, 250 A/m

, 2.8%

solution; (d) charged for 2 hours at room temperature,

5000 A/m

, 5% H

solution.

Intensity (c.p.s.)

1000

750

500

250

20 25 30 35 40 45 50 55 60 65 70

2θ (°)

(a)

(b)

(c)

(d)

001γ

110γ

111γ

201α

002γ

200γ

201γ

220γ

101H

111H

121H

102H



Differential scanning calorimetry and property measurements 89

There is a small amount of α

phase present in the uncharged Ti-46Al-

1.9Cr-3Nb alloy. However, the small amount of α

phase disappears with

charging (see Fig. 4.17).

Ti-46Al-1.9Cr-3Nb alloy subject to cathodic charging has shown cracking.

The XRD pattern for the β21s alloy charged in the solution of H

with

0.2 g/l As

produces no additional peak in the 2

range between 20 and

70°, when compared with the XRD pattern for the uncharged alloy (Fig.

4.18). No hydride has formed. Comparison of the XRD patterns for uncharged

alloy shows the possible heterogeneity of the material, concerning the shift

of peaks, although there might be a contribution from experimental factors.

Both intensity and 2

can be altered, depending on the level at which the

sample is placed within the XRD apparatus. Therefore, if all peaks move in

the one direction, or if all peaks drop in intensity, it is possible that this is due

to samples being placed at different levels during the XRD experiments.

However, if the intensity of one peak increases while that of another peak

decreases in relation to the uncharged XRD pattern of the sample in question,

this is not because of an experimental factor. There is an increase in the

amount of β phase while the amount of α phase decreases.

In spite of the absence of hydride, the surface of β21s still displays damage

caused by hydrogen charging, and this could be seen even with the naked

eye. Charging for 2 hours at room temperature, 5000 A/m

in 5% H

solution produced extensive cracking, while the surfaces of the samples

charged for 2 hours at 85 °C and 250 or 5000 A/m

were blackened when

compared to their original appearance.

Table 4.3

X-ray diffraction peak relative height of titanium aluminide after

hydrogen charging

Phase

hkl

Uncharged 5% H

, 2.8% H

, 5% H

0.2 g/l As

, 1 g/l thiourea, 0.2 g/l As

250 A/m

, 250 A/m

, 5000 A/m

1 h, room 4 h, 85 °C 2 h, room

temperature temperature

γ 001 11 12 – 17

110 8 12 22 11

111 100 100 100 100

002 26 26 25 21

200 25 26 35 46

201 4 5 13 8

220 18 20 25 24

201 13 –––

(TiAl)H

101 –– 40 16

111 – 26 85 30

121 –– 18 8

102 –– 11 4



Titanium alloys: modelling of microstructure90

Table 4.4 shows the hydrogen analysis results.

It has already been established that a hydride phase based on (TiAl)H

forms in Ti-46Al-1.9Cr-3Nb alloy during cathodic charging, but no hydride

forms in β21s during cathodic charging. Table 4.4 shows that hydrogen does

enter β21s, and this leads to the conclusion that the hydrogen occupies the

interstitial sites in β21s.

From Table 4.4, the amount of hydrogen present before charging is

approximately 40 times greater in β21s than Ti-46Al-1.9Cr-3Nb alloy. However,

4.18

X-ray diffraction patterns for β21s alloy. (a) Uncharged; (b)

charged for 2 hours at room temperature, 5000 A/m

, 5% H

solution; (c) charged for 2 hours at 85 °C, 5000 A/m

, 5% H

solution.

Intensity (c.p.s.)

20000

18000

16000

14000

12000

10000

8000

6000

4000

2000

34 35 36 37 38 39 40 41 42 43 44 45

2θ (°)

(a)

(b)

(c)

100α

110β

101α

Table 4.4

Rates of hydrogen penetration

Material Current Temperature Time Hydrogen Hydrogen Rate of

density (°C) (h) (ppm) entered penetration

(A/m

) (ppm) (ppm/h )

Ti-46Al- Uncharged ––24 ––

1.9Cr-3Nb 250 room 1 99 75 75

250

(a)

85 4 638

(b)

614 154

5000 room 2 643 619 310

β21s Uncharged ––963 ––

250 85 2 1132 169 85

5000 85 2 1293 330 165

(a)

The charging solution was 2.8% H

. A 5% H

solution was used for all

other samples.

(b)

This corresponds to 2.5 at.% of hydrogen.



Differential scanning calorimetry and property measurements 91

calculating the rates of hydrogen penetration for each of the samples has

demonstrated that β21s is less susceptible to hydrogen penetration. Both an

increase in the current density and an increase in the temperature of the

solution cause an increase in the rate of hydrogen penetration.

In summary, the small amount of α

phase in the Ti-46Al-1.9Cr-3Nb

alloy becomes undetectable after hydrogen charging. A hydride based on

(TiAl)H

, which has tetragonal lattice parameters of a = 0.452 nm and c =

0.326 nm (c/a = 0.721), forms in Ti-46Al-1.9Cr-3Nb alloy after cathodic

charging. Hydrogen charging induces crack formation on both Ti-46Al-1.9Cr-

3Nb alloy and β21s. No hydride is formed in β21s by cathodic charging. The

hydrogen enters the alloy and occupies the interstitial sites between the

atoms. During charging, there is an increase in the amount of the β phase and

a decrease in the amount of the α phase on the β21s surface. β21s is less

susceptible to hydrogen penetration than Ti-46Al-1.9Cr-3Nb alloy. The ability

of β21s to resist hydrogen penetration, when compared with the ability of Ti-

46Al-1.9Cr-3Nb alloy, is approximately greater by a factor of two.

4.4 References

Ohnuma I, Fujita Y, Mitsui H, Ishikawa K, Kainuma R and Ishida K (2000), ‘Phase

equilibria in the Ti–Al binary system’, Acta Mater, 48 (12), 3113–23.

Sundaram P A, Quadakkers W J and Singheiser L (2000), ‘Hydrogen effusion in cathodically

charged gamma titanium aluminides’, J Alloys Compd, 298 (1–2), 274–78.



Thermodynamic modelling

Abstract: Thermodynamic calculations using Thermo-Calc are used to

quantify the phase fraction and element partition in a number of

conventional titanium alloys and γ-TiAl based alloys, chosen from all of the

titanium alloy groups and including both well investigated alloys and those

that are relatively new. Calculations of the phase constitution and element

distribution in these alloys show a good agreement with the experimental

measurement. In addition, Thermo-Calc calculation can help identify the

existence of some phases that are not readily observed experimentally. It can

be used as a guide in alloy design.

Key words: titanium aluminides based on TiAl, phase transformation,

precipitates, phase stability, thermodynamics.

5.1 Introduction

The thermodynamics modelling is usually based on the Gibbs energy

calculation. Two most commonly used calculation packages are Thermo-

Calc (Andersson et al., 2002) and MTDATA (Davies et al., 2002). These

packages have been successfully applied for thermodynamic calculations of

different materials systems (Dore et al., 2000; Ekroth et al., 2000; Eskin,

2002; Gorsse and Shiflet, 2002; Guo and Sha, 2000; Jarvis et al., 2000; Sha,

2000; Zackrisson et al., 2000) (Chapters 3 and 6).

Thermodynamic calculation can supplement experimental characterisation,

and allows the prediction of phase type, and fraction and element distribution

in different phases. Thermo-Calc is a computer package developed particularly

for thermodynamic calculations of multi-component equilibrium as a function

of pressure, temperature and the combined effect of alloying elements, using

a databank of assessed thermodynamic data and models for the phases in the

system that are as good an approximation of the nature as possible. It employs

rigorous thermodynamic expressions and numerical methods of minimising

the chemical energy of the system, so that interpolation between the available

experimental data can be made. Good agreement has been obtained in the

past between the calculated phase compositions and experimental

measurements.

In this chapter, thermodynamic calculations of the phase equilibria in

different titanium alloys will be shown, including Ti-6Al-4V (Ti 6-4), Ti-

6Al-2Sn-4Zr-2Mo (Ti 6-2-4-2), Ti-6Al-2Sn-4Zr-6Mo (Ti 6-2-4-6), Ti-8Al-

1Mo-1V (Ti 8-1-1), Ti-5.8Al-4Sn-3.5Zr-0.7Nb-0.5Mo-0.35Si (IMI 834), Ti-

6Al-7Nb (IMI 367), Ti-10V-2Fe-3Al (Ti 10-2-3) and TIMETAL β21s. These

alloys are chosen from all of the titanium alloy groups and include both well-



Titanium alloys: modelling of microstructure96

investigated alloys and those which are relatively new. The calculations are

performed for the actual alloy composition taking into account all major

alloying elements. Further thermodynamic calculation procedures are given

in Chapters 3 and 6. The results from the thermodynamic calculations and

their coupling with the models for simulation of the microstructure evolution

in titanium alloys are discussed in several chapters in this part of the book.

5.2 Conventional titanium alloys

In conventional titanium alloys, there are different equilibrium amounts of α

and β phases as well as equilibrium compositions of both phases at different

temperatures. We shall calculate the phase equilibria for different titanium

alloys in wide temperature ranges. The calculated data are compared with

experimentally obtained data (Chapter 6).

5.2.1 Ti-6Al-4V (Ti 6-4)

The calculated amounts of the α and the β phases for the Ti-6Al-4V alloy are

given in Fig. 5.1a, showing increased equilibrium amount of the α phase

when the temperature decreases. The tendency is in agreement with the

behaviour of the α → β phase transformation in titanium alloys.

Further, we calculate the equilibrium chemical compositions of both α

and β phases as functions of the temperature (Fig. 5.1b). Such calculation is

important for further modelling work because it gives information for the

diffusion redistributions between the two phases during the course of β to α

transformation.

5.2.2 Ti-6Al-2Sn-4Zr-2Mo (Ti 6-2-4-2)

The calculation for the Ti-6Al-2Sn-4Zr-2Mo alloy also shows increased

amounts of the α phase and decreased amounts of the β phase with temperature

decrease (Fig. 5.2a). Compared to the Ti-6Al-4V alloy, the α phase in the Ti-

6Al-2Sn-4Zr-2Mo alloy is more stable in the entire temperature range. The

amount of the α phase at lower temperature for this alloy approaches 100%,

while in the Ti-6Al-4V alloy, there is some amount of the β phase remaining.

On the other hand, for the Ti-6Al-4V alloy, the phase composition of 50%α

+ 50%β is at about 905 °C, while in the Ti-6Al-2Sn-4Zr-2Mo alloy, the

50%α + 50%β phase composition is at around 955 °C. This is why the

Ti-6Al-2Sn-4Zr-2Mo alloy is considered as a high-temperature titanium

alloy, and can work at higher temperatures as compared to the Ti-6Al-4V

alloy.

The calculations of the equilibrium concentrations of the alloying elements

in the α and the β phases for this alloy suggest that the β to α transformation



Thermodynamic modelling 97

at cooling is in conditions of diffusional redistribution and enrichment of

molybdenum in the β phase (Fig. 5.2b).

The Ti-6Al-2Sn-4Zr-6Mo alloy is similar to the Ti-6Al-2Sn-4Zr-2Mo alloy

but with a higher amount of molybdenum, and is classified as an α + β alloy.

The thermodynamics of the two alloys is similar.

Alpha

Beta

Phase amounts (%)

100

750 800 850 900 950 1000

(°C)

(a)

750 800 850 900 950 1000

(°C)

(b)

Content (wt.%)

Al in alpha

V in alpha

Al in beta

V in beta

5.1

Calculated equilibria versus temperature in the Ti-6Al-4V alloy:

(a) equilibrium amounts of the α and the β phases; (b) equilibrium

concentrations of aluminium and vanadium in the α and the β

phases. The calculations are for composition of Al = 6 wt.%; V = 4

wt.%; Fe = 0.2 wt.%; O = 0.15 wt.%; N = 0.03 wt.%; and C = 0.01

wt.%.

