Characterization of solutionizing behavior in VT14 titanium alloy using ultrasonic velocity and attenuation measurements

Characterization of solutionizing behavior in VT14 titanium alloyusing ultrasonic velocity and attenuation measurements

Anish Kumar a, T. Jayakumar a,*, Baldev Raj a, K.K. Ray b

a Metallurgy and Materials Group, Indira Gandhi Center for Atomic Research, Kalpakkam 603102, Indiab Indian Institute of Technology, Kharagpur 721302, India

Received 11 December 2002; received in revised form 6 March 2003

Materials Science and Engineering A360 (2003) 58�/64

www.elsevier.com/locate/msea

Abstract

VT14 titanium alloy (Ti�/4.5Al�/3Mo�/1V) was subjected to a series of heat treatments consisting of solutionizing for 1 h at the

selected temperatures in range of 923�/1323 K at an interval of 50 K, followed by water quenching. Hardness and optical microscopy

results are correlated with ultrasonic longitudinal and shear wave velocities and attenuation in these specimens. Ultrasonic velocities

and hardness decrease with solution annealing temperature (SAT) in the 923�/1123 K range. Beyond 1123 up to 1223 K, they

increase slightly. Beyond 1223 K, ultrasonic velocities become constant, whereas hardness increases up to 1323 K. Ultrasonic

attenuation exhibits an opposite behavior to velocity and hardness. Further, for the first time, authors have shown that ultrasonic

velocity can be used to identify the b-transus temperature in this alloy. Because of non-monotonous variation of velocity and

attenuation with solutionizing temperature, it was not possible to identify the SAT using any one of these parameters. Hence, a new

parameter, ratio of normalized differential of ultrasonic attenuation to normalized differential of ultrasonic velocity (RNDAV) has

been used, which is found to increase monotonously with SAT and hence enabling unambiguous characterization of SAT in solution

annealed VT14 alloy.

# 2003 Elsevier B.V. All rights reserved.

Keywords: Ultrasonic velocity; Ultrasonic attenuation; VT 14 titanium alloy; Microstructural characterization

1. Introduction

Titanium alloys, by virtue of their excellent specific

strength and modulus and better intermediate tempera-

ture strength, are the most preferable structural materi-

als for aerospace applications. Further, due to their

excellent corrosion resistance and good compatibility

with human organs, titanium alloys are also widely used

for human implants. Solution annealed and tempered

a�/b titanium alloys possess better mechanical proper-

ties, such as yield, tensile and fatigue strength, than a�/bannealed alloys [1]. The solution annealing temperature

(SAT) plays an important role, as it decides the volume

fraction of primary a and b phases and volume fraction

of the alloying elements in different phases [2]. The

amount of b stabilizing elements in b phase governs the

stability of the phase upon rapid cooling to room

temperature and hence decides the product phase.

Further, if solutionizing is carried out above b-transus

temperature, the alloy looses its ductility due to the

substantial increase in grain size. While elastic properties

of most of the structural materials differ very marginally

with heat treatments, titanium alloys can exhibit varia-

tions as much as 10% [2]. As propagation of ultrasonic

wave depends upon the elastic properties of the material,

ultrasonic velocity can be a very good parameter for

characterization of heat treatments and corresponding

microstructure in titanium alloys.

Ultrasonic parameters, such as velocity and attenua-

tion, have been correlated with the microstructural

features evolved during heat treatments in ferritic steels

[3,4], superalloy [5], aluminum alloy [6] and many other

materials. Ultrasonic techniques have been used for

determination of yield strength [7], fracture toughness

[8], grain size [3], volume fraction of second phases [3,9],

* Corresponding author.

E-mail addresses: [email protected] (A. Kumar),

[email protected] (T. Jayakumar), [email protected] (B. Raj),

[email protected] (K.K. Ray).

0921-5093/03/$ - see front matter # 2003 Elsevier B.V. All rights reserved.

doi:10.1016/S0921-5093(03)00495-7

mailto:[email protected]




etc. Recently, Anish Kumar et al. [3] have used the

ultrasonic velocity and attenuation measurements for

characterization of the microstructures obtained by

various solution annealing treatments in modified9Cr�/1Mo ferritic steel. They showed that ultrasonic

velocity is useful for the determination of volume

fraction of ferrite and martensite and for identifying

the Ac1 and Ac3 critical temperatures, whereas, ultra-

sonic attenuation could be well correlated with the prior

austenitic grain size.

In the present work, ultrasonic velocity and attenua-

tion measurements are correlated with the microstruc-ture generated by solutionizing treatment in VT14

titanium alloy. For the first time, authors have shown

that ultrasonic velocities can be used for identifying the

b transus temperature in this alloy. As it was not

possible to determine the SAT using either ultrasonic

velocity or attenuation alone due to their non-mono-

tonous variation with temperature, a new parameter has

been identified, which is a function of both velocity andattenuation. This parameter, defined as the ratio of

normalized differential of ultrasonic attenuation to

normalized differential of ultrasonic velocity (RNDAV),

is found to increase monotonously with SAT, hence can

be used for nondestructive identification of solution heat

treatment of VT14 alloy.

2. Experimental

2.1. Heat treatment and specimen preparation

Various specimens of VT14 titanium alloy (Ti�/4.5Al�/

3Mo�/1V) of dimensions 20�/20�/12 mm were pre-

pared from a b-heat-treated (1323 K/1 h) and water

quenched disk of diameter 100 and 12 mm thickness.

These specimens were subjected to a series of heattreatments consisting of solutionizing for 1 h at the

selected temperatures in the range of 923�/1323 K at an

interval of 50 K, followed by water quenching. Metallo-

graphic examination was carried out to reveal the

microstructures in different specimens. The etchant

used is Kroller’s reagent [10]. The hardness of these

specimens was also measured using Vicker’s hardness

tester at a test load of 10 kg. Surface grinding of thesespecimens was carried out to obtain the specimens of

10.59/0.3 mm thickness with plane parallelism to an

accuracy of better than 9/3 mm.

2.2. Ultrasonic measurements

Fig. 1 shows the schematic of the experimental setup

used for ultrasonic measurements. 100 MHz broad bandpulser-receiver (M/s. Accutron, USA) was used to give

the electrical pulse to the transducer for generating the

ultrasonic waves. The transducers consist of piezo-

electric crystals, which vibrate upon the application of

electrical pulse across the thickness and generate the

mechanical waves. For the longitudinal wave transdu-

cer, the piezoelectric crystal is cut perpendicular to the

electric axis (X-cut crystals) and hence when the

electrical pulse is applied in the thickness direction, the

crystal vibrates in the thickness direction (i.e. perpendi-

cular to the specimen surface) and generates the long-

itudinal wave in the specimen coupled to the transducer.

Whereas for the shear wave transducer, piezoelectric

crystal is cut perpendicular to the mechanical axis (Y-cut

crystal) and hence when the electrical pulse is applied in

the thickness direction, the crystal vibrates in the width

(i.e. parallel to the specimen surface) direction and thus

generates the shear wave in the specimen coupled to the

transducer. The schematic of the propagation of long-

itudinal and shear waves are as shown in Fig. 2. In the

case of longitudinal wave, the particle vibration is in the

direction of wave motion, whereas, in the case of shear

wave, the particle vibration is perpendicular to the

direction of wave motion.

A normal beam delay line piezoelectric transducer of

15 MHz frequency (PDG152), supplied by M/s. NDT

Systems Inc., California, was used for longitudinal wave

velocity and attenuation measurements and shear wave

velocity measurements were carried out using 5 MHz

Fig. 1. Schematic of the experimental setup for ultrasonic measure-

ments.

Fig. 2. Schematic of the particle vibration and wave motion during

propagation of (a) longitudinal and (b) shear waves.

A. Kumar et al. / Materials Science and Engineering A360 (2003) 58�/64 59

piezoelectric normal beam shear wave transducer

(V155), supplied by M/s. Panametrics Inc, USA. The

transducer was coupled with the specimen using ultra-

sonic couplant ZG-F (M/s. Krautkramer, Germany) for

longitudinal waves and honey (M/s. Panametrics, USA)

for shear waves. The reflected ultrasonic waves from the

specimen were picked up by the same transducer and

ultrasonic waves were converted to electrical signal. This

electrical signal was acquired by the receiver at optimum

gain and damping settings, and the rf signal was fed to

the 500 MHz digitizing oscilloscope (Tektronix

TDS524). The signal was digitized at 500 MHz and

averaged for about 50 signals. The gated backwall

echoes (2048 ns duration) from the oscilloscope were

transferred to the personal computer with the help of

General Purpose Interface Bus (GPIB) interfacing and

software developed in LabVIEW. Specific software was

developed in LabVIEW 3.1.1 for the acquisition of

ultrasonic signals and on-line determination of ultra-

sonic velocity and attenuation using cross correlation

and frequency spectrum based methodologies, respec-

tively. Fig. 3 shows the front panel of the software

developed for ultrasonic velocity and attenuation mea-

surements. Before starting the experiment, the thickness

of the specimen and the difference in the number of the

backwall echoes are entered in the front panel. If the

first and second backwall echoes are considered then

this difference in the number of the backwall echoes is 1.

Cross correlation technique has been used for the

precise velocity measurements. Various steps involved in

this methodology are as follows: (a) acquisition of two

digitized echoes from the oscilloscope, (b) cross correla-tion of the echoes to find the approximate time delay, (c)

application of an interpolation method for accurate time

delay and (d) calculation of velocity measurement using

time delay and thickness of the specimen along with the

following equation;

V�2 � Difference of Echo � Thickness

Time delay(1)

The accuracy in time of flight measurement is better

than 9/0.3 ns and in turn the accuracies in the

measurement of ultrasonic longitudinal and shearwave velocities for thickness of the specimens in range

of 10.59/0.3 mm are better than 9/2.5 and 9/1.5 m s�1,

respectively. The use of cross correlation may lead to

errors in the case of multiple peaks, pulse expansion and

jitters. To ensure that there is no error in the velocity

measurements, after the calculation of the time of flight,

the backwall echoes are given the calculated time delay

to check whether they exactly overlap or not (Fig. 3).However, in the present case there appears to be no

pulse expansion, jitter and appreciable change in the

signal, the backwall echoes overlapped well in all the

measurements.

For attenuation measurements, frequency spectrum

based ultrasonic attenuation measurement method was

Fig. 3. Front panel of the software for ultrasonic velocity and attenuation measurements.

A. Kumar et al. / Materials Science and Engineering A360 (2003) 58�/6460

used. The peak value in the autopower spectrum of the

first backwall echo was detected and the 6 dB drop point

was calculated. The frequency values corresponding to

the 6 dB drop level were calculated for both sides of the

peak value. All the data points lying in between these

two 6 dB drop points were summed for the first (S1) and

second (S2) backwall echoes. Ratio of S1 to S2 is the

ratio of the areas of autopower spectra of the two

echoes, which in fact gives the ratio of energy content of

the two backwall echoes in this frequency band. Then

the ultrasonic attenuation coefficient was calculated by

the following formula:

Attenuation Coefficient (dB mm�1)

�20 log

S1

S2

2 � thickness (mm)(2)

The maximum scatter in the attenuation coefficient

measurement was less than 9/0.02 dB mm�1.

3. Results and discussion

Fig. 4 shows micrographs of specimens solutionized at

1073, 1123, 1173 and 1223 K. The amount of primary adecreases with increase in the heat treatment tempera-

ture up to 1173 K. Whereas, the specimen solutionized

at 1223 K shows the presence of a? marteniste only,

indicating that the b transus (Ac3) temperature of VT-14

alloy lies between 1173 and 1223 K. In any a�/btitanium alloy, volume fraction of b phase increases

with increase in temperature up to b transus tempera-

ture. But as the amount of b stabilizing elements is fixed,

the volume fraction of b-stabilizing elements in b phase

decreases with increase in temperature. Such b phase

composition is unstable at room temperatures at which

it decomposes into a lesser volume fraction of b and

secondary a. The same occurs during cooling to room

temperature, if enough time is given for diffusion to take

place. If the alloy is quenched to room temperature

(diffusion is not allowed), the b phase becomes unstable

and either remains as unstable b, or transforms to soft

aƒ (orthorhombic) or hard a? (hexagonal closed packed)

martensite, depending upon the SAT [2]. The unstable bhas been reported to be a soft phase having the lowest

moduli and highest damping capacity among all the

phases present [2,3].

Figs. 5 and 6 show, respectively, the variation in the

longitudinal and shear ultrasonic velocities and in

attenuation with SAT for solution annealed and

quenched specimens. Fig. 6 also shows the variation in

hardness with SAT. Ultrasonic velocities and hardness

decrease with SAT upto about 1123 K followed by a

gradual increase upto 1223 K. Beyond 1223 K, the

ultrasonic velocities become almost constant, however,

hardness increases continuously. The initial decrease in

ultrasonic velocities and hardness with increase in SAT

is attributed to the increased amount of b phase having

lesser stabilizing elements [2]. This decreases the elastic

modulus of the b phase [2] and hence of the alloy soaked

Fig. 4. Micrographs of the specimens solution annealed at (a) 1073 K,

(b) 1123 K, (c) 1173 K and (d) 1223 K in air for 1 h followed by water

quenching.

Fig. 5. Variation in ultrasonic velocities with SAT in the specimens

that were solution heat treated followed by water quenched.


at higher temperatures. The minimum in ultrasonic

velocities at about 1123 K is attributed to the formation

of maximum amount of soft unstable b phase having the

lowest modulus. The increase in the velocities and

hardness for SAT beyond 1123 K is attributed to the

formation of hard a? martensite during quenching

instead of unstable b phase. Beyond 1223 K, ultrasonic

velocities become constant due to formation of 100% a?martensite with similar modulus at all heat treatment

temperatures above 1223 K. This is in agreement with

the metallographic studies, which exhibited the presence

of only a? martensite in the specimen solutionized at

1223 K (Fig. 4d). The total variation in ultrasonic

longitudinal and shear wave velocities in the present

study are �/1.76 and �/6.9%, respectively. The larger

variation and better accuracy (due to lower velocity and

hence higher transit time) of shear wave velocity

compared with longitudinal wave velocity measure-

ments make shear wave velocity a more reliable para-

meter for microstructural characterization in VT14

alloy.

The variation in ultrasonic attenuation with SAT

exhibits opposite behavior to that of the velocities and

hardness. It increases with SAT to a peak at about 1123

K followed by continuous decrease up to about 1223 K.

Beyond 1223 K, ultrasonic attenuation remains almost

constant up to 1323 K. The maximum in the attenuation

at about 1123 K is again attributed to the formation of

maximum amount of soft unstable b phase having the

highest damping. The decrease in attenuation beyond

the maximum is attributed to the formation of hard a?martensite instead of unstable b phase.

The present study reveals that ultrasonic velocities

and attenuation in VT14 alloy are affected by the SAT.

However, because of non-monotonous variation of

ultrasonic velocity and attenuation with SAT (Figs. 5

and 6), any value of velocity or attenuation in an

unknown specimen may be correlated with solutionizing

at either a temperature below the peak or with a

temperature above the peak (about 1123 K). Identifica-

tion of SAT is difficult; when using either velocity or

attenuation parameter alone. For example, if a specimen

exhibits a shear wave velocity of 2940 m s�1 (line A�/B

in Fig. 5), it can be correlated with SAT of either 1200 or

1100 K (Fig. 5). Similarly, the attenuation plot (Fig. 6)

exhibits that the specimen is either solutionized at 1200

or at 1000 K. Hence for exact identification of SAT,

some other ultrasonic parameter is required to be

identified, whose variation is monotonous with tem-

perature.

Ultrasonic velocity is dependent upon the modulus

and density of the phases present, whereas ultrasonic

attenuation depends upon the damping and scattering of

the waves. Also as different phases affect the modulus,

density, scattering power, and damping power to

different extents, they will have different effects on

ultrasonic velocity and attenuation. Hence the correla-

tion between ultrasonic velocity and attenuation should

be different for different phases affecting these two

ultrasonic parameters differently. Based on this ap-

proach, Fig. 7 shows the variation in ultrasonic velo-

cities with ultrasonic attenuation. Examination of this

figure shows that ultrasonic velocities decrease more

rapidly upto the peak values (specimens solutionized

upto 1123 K), as compared with that beyond the peak

values. This means that a parameter, which accounts for

the variations in both ultrasonic attenuation and

velocity, can be used to identify whether the specimen

is solutionized at a temperature above or below the peak

(i.e. 1123 K). As the shear velocity is a better parameter

than longitudinal wave velocity, all further analyses

were carried out for shear wave velocity only. The ratio

of the normalized differential between ultrasonic at-

tenuation corresponding to the heat treatment tempera-

ture and the lowest treatment temperature (923 K) to the

normalized differential in ultrasonic shear wave velocity

relative to that at the lowest temperature (923 K) is used

as a new parameter named as RNDAV. RNDAV for

Fig. 6. Variation in ultrasonic attenuation and hardness with SAT in

the specimens that were solution heat treated followed by water

quenched.

Fig. 7. Variation in ultrasonic velocities with attenuation for speci-

mens solution annealed at different temperatures.


each specimen has been calculated as follows:

RNDAVu�(au � a923 K)=a923 K

(Vu � V923 K)=V923 K

where au and Vu are attenuation and shear wave

velocity for the specimen solutionized at temperature u

and a923 K (0.21 dB mm�1) and V923 K (3115 m s�1) are

attenuation and shear wave velocity for the specimen

solutionized at 923 K.

Table 1 and Fig. 8 show the variation in RNDAVwith SAT. The negative values of RNDAV are due to

the fact that the variations in ultrasonic velocity and

attenuation are always in opposite directions to each

other for all the specimens studied (Table 1). The

horizontal line XY at RNDAV:/�/43 corresponds

specimen solutionized at 1123 K (peak in attenuation

and velocity). Since the RNDAV increases monoto-

nously with SAT, it can be used to determine the SATunambiguously.

In the present study, the use of RNDAV has been

demonstrated for the identification of SAT for solution

annealed followed by water quenched VT14 alloys. The

use of RNDAV for other heat treatment conditions and

in different alloy systems is also being explored. The

specimen corresponding to the extreme values of velo-

city and attenuation can be taken as reference in thoseconditions.

4. Conclusion

The present study reveals that ultrasonic velocity and

attenuation measurements can be used for non-destruc-

tive characterization of solutionizing behavior in Ti�/

4.5Al�/3Mo�/1V (VT 14 alloy). For the first time, it

has been shown that ultrasonic velocities can be used in

identifying the b transus temperature in titanium alloys.

It is also shown that ultrasonic shear wave velocity is a

better parameter than longitudinal velocity for micro-structural characterization in VT14 alloy. Further, it is

found that neither of ultrasonic velocity nor attenuation

alone can be used for the determination of SAT due to

their non-monotonous variation with SAT. However,

the combined use of velocity and attenuation in terms of

a new parameter defined as RNDAV is found to vary

monotonously with SAT and hence is useful for

unambiguous identification of the SAT in the solutionannealed VT14 alloy.

Acknowledgements

We are thankful to Dr S.L. Mannan, Associate

Director, Materials Development Group and Mr P.

Kalyanasundaram, Head, Division for PIE and NDT

Development, Indira Gandhi Center for Atomic Re-search for their cooperation. Authors are also thankful

to Mr M. Narayana Rao, Nuclear Fuel Complex,

Hyderabad, for providing the Ti�/4.5Al�/3Mo�/1V alloy.

Table 1

Variation in ultrasonic shear wave velocity, ultrasonic attenuation and RNDAV with SAT

SAT (K) Ultrasonic shear wave velocity (V), (m s�1) Ultrasonic attenuation (a), (dB mm�1) /

au � a923 K

a923 K

/

Vu � V923 K

V923 K

RNDAV

923 3115 0.2068 0 0 �/

973 3096 0.2888 0.39652 �/0.0061 �/65.01

1023 3055 0.4043 0.95503 �/0.01926 �/49.58

1073 2998 0.56 1.70793 �/0.03756 �/45.47

1123 2900 0.835 3.03772 �/0.06902 �/44.01

1173 2923 0.46 1.22437 �/0.06164 �/19.86

1223 2954 0.3811 0.84284 �/0.05169 �/16.31

1273 2956 0.3607 0.7442 �/0.05104 �/14.58

1323 2957 0.308 0.48936 �/0.05072 �/9.65

Fig. 8. Variation in RNDAV with SAT.


References

[1] T. Ahmed, Mater. Sci. Eng. A243 (1998) 206.

[2] Y.T. Lee, G. Welsch, Mater. Sci. Eng. A128 (1990) 77.

[3] A. Kumar, K. Laha, T. Jayakumar, K.B.S. Rao, B. Raj, Metall.

Mater. Trans. A 33A (2001) 1617.

[4] A. Kumar, B.K. Choudhary, T. Jayakumar, K.B.S. Rao, B. Raj,

Mater. Sci. Technol. 19 (2003) 637.

[5] A. Kumar, V. Shankar, T. Jayakumar, K.B.S. Rao, B. Raj, Phil.

Mag. A 82 (13) (2002) 2529.

[6] M. Rosen, L. Ives, S. Ridder, F. Biancaniello, R. Mehrabian,

Mater. Sci. Eng. 74 (1985) 1.

[7] A. Kumar, T. Jayakumar, B. Raj, Phil. Mag. A 80 (11) (2000)

2469.

[8] A. Vary, Mater. Eval. 46 (1988) 642.

[9] T.R.G. Kutty, K.N. Chandrasekharan, J.P. Pannakal, S.K.

Ghoshal, P.K. De, NDT. E. Int. 20 (1987) 359.

[10] J. Newby (Ed.), Metals Handbook: Metallography and Micro-

structures, vol. 9, Kathleen Mills, American Society for Metals,

9th edition (1989).


Characterization of solutionizing behavior in VT14 titanium alloy using ultrasonic velocity and attenuation measurements

Documents