surface treatment, computer methods and experimental ... › secure › elibrary › papers ›...

Fatigue strength of ion-nitrided T1-6A1-4V alloy

in high cycle region

Shin-ichi Nishida & Nobusuke Hattori

Faculty of Science & Engineering, Saga University, Honjo-machi 1,

Saga, 840, Japan

E-mail: [email protected]

Abstract

Though various kinds of surface treatment methods were applied to improve thewear resistant properties of Ti-6Al-4V alloy, it has been reported that the brittlesurface layer of this alloy deteriorated their fatigue properties. Furthermore,when this alloy is nitrided by gas and salt bath, the fatigue strength of nitridedone decreases as compared with that of non-nitrided one. On the other hand,as the ion-nitriding method is treated in clean environment, the surface of thematerial could become free from hydrogen and oxygen. There are few reportsabout the fatigue strength of ion-nitrided titanium alloys.

Fatigue tests have been accomplished by rotating bending fatigue testingmachine using a partial shallow notch specimen after different temperatures ofheat treatment. In addition, the specimen surface has been successivelyobserved by replica method during tests and it seems clear that the fatiguestrength is improved by removing the chemical compound surface layer, becausethis chemical compound layer deteriorates the fatigue strength of nitridedspecimen. From the above results, the surface layer would be requested tohave the equivalent toughness of the matrix to improve its fatigue strengthaccording to the view point of practical use.

1 Introduction

Generally speaking, titanium alloys have been widely used in the aerospaceindustries because of their superior properties such as high specific strength andhigh corrosion resistance etc.. In addition, not only high fatigue resistance but

Transactions on Engineering Sciences vol 17, © 1997 WIT Press, www.witpress.com, ISSN 1743-3533

200 Surface Treatment, Computer Methods and Experimental Measurements

also high wear resistance are essentially required for the many kinds of criticaluse of structures. Though these kinds of alloys especially show the low wearresistance, various kinds of surface hardening methods had been applied toimprove these problems. When this material is nitrided by gas or salt bath; e.g.Morita et al [1,2], Tokagi et al [3,4], the abrasion resistance increasesremarkably. On the other hand, as the material absorbs hydrogen and oxygen,the toughness of materials is deteriorated. Furthermore, it has been reportedthat fatigue strength of nitrided materials decreases as compared with that ofnon-nitrided one. The authors have considered that the ion-nitriding processwould be one of the important methods to accomplish the above requirements.As the ion-nitriding process is treated in clean atmosphere, the surface ofmaterial is free from hydrogen and oxygen. However, there are few reportsabout the fatigue properties of ion-nitrided titanium alloy [5,6,7,8,9].

In this study, the fatigue properties, especially, the fatigue strength, fatiguecrack initiation and propagation behavior have been investigated using thespecimen ion-nitrided Ti-6Al-4V alloy which is one of the most representativetitanium alloys.

2 Experimental procedure

The material used in this test is the ion-nitrided Ti-6Al-4V alloy. Tables 1 and2 list the chemical composition and the heat-treatment conditions of specimens,respectively.

Figure 1 shows the shape and dimensions of specimens. Though the authorshad started the experiments using the specimen of type A, they have facedtroubles to remove the specimen's warp during ion-nitriding and have changedfrom the specimen of type A to type B afterward. That is, the specimen of typeB is smaller than type A, type B is less affected than type A about thespecimen's warp during ion-nitriding process.

After machining , all of the specimens are polished, and annealed in vacuumat the temperature of 600 °C for 30 minutes afterwards. Therefore, ion-nitriding process has been adopted to the titanium specimen. The fatigue testshave been performed by both Ono-type (capacity; 98 N-m, 3,400 rpm) andsmall-Ono-type (capacity; 14.7 N-m, 3,000 rpm) rotating bending fatiguetesting machines. Each specimen used in fatigue test has the partial shallownotch, which does not affect for its fatigue strength at all, in order to limit thefatigue damaged part. Fatigue crack initiation and propagation behavior arecontinuously observed using replica method at the partial shallow notch.

3 Result and discussion

Figure 2 shows an example of micro-structure of the longitudinal section of thespecimen nitrided at 780°C for 1 hour (T8N-1). In this case, the chemical

compound layer of titanium and nitrogen is approximately 2/zm in the thickness.


Surface Treatment, Computer Methods and Experimental Measurements 201

Although it is clear that the surface hardness is larger than that of matrix, itwas difficult to measure the hardness in the surface layer itself for the thicknessof a few micron meter. The hardness of this part was evaluated by an othermethod. That is, Figure 3 shows the relation between surface hardness andindentation depth. The indentation depth becomes deeper according to largerload weight. As shown in this figure, the surface hardness of the specimen,T8N-1, has exceeded over Hv 1000. In addition, as ion-nitriding temperaturebecomes lower, the surface hardness becomes also lower.

Figure 4 shows the effect of heat treatment conditions on fatigue strength.When considering the fatigue limit is the limit stress in which the specimen canstand for the stress amplitude by 10^ cycles, the fatigue limit of the specimen,T8N-1, is smaller than that of the specimen, TNo, by 20 MPa. On the otherhand, fatigue limit of the specimen, T8N-1, is larger than that of the specimen,T8V-1, by 80 MPa. Namely, though the fatigue limit of ion-nitrided specimen(T8N-1) is slightly smaller than that of the specimen (T8V-1) which is onlyannealed, that of ion-nitrided specimen becomes considerably larger than thatof the specimen which is treated with the same heat treatment history.

Figure 5 shows the effect of ion-nitriding hours on fatigue strength.According to this figure, the fatigue limit of the specimen, T8N-1, becomeshigher than that of the specimen, T8N-5, by 30MPa. That is, the fatigue limit isdeteriorated, when ion-nitriding hours becomes larger. Though the ion-nitriding temperature changes from 780 °C (T8N-5) to 680 °C (T7N-5) bykeeping constant for another condition, there is no effect on the fatigue strength.These data is smaller than that of the specimen nitrided by 1 hour (T8N-1).The ion-nitriding temperature changes 580°C (T6N-5) and 480 °C (T5N-5).The same tendency as before would be concluded about the specimens T6N andT5N etc..

Figure 6 shows the relation between the fatigue limit and heat treatmenttemperature including ion-nitriding temperature. When the temperaturebecomes lower, the fatigue limit becomes higher. Though the fatigue limit ofthe ion-nitrided specimen, T8N-5, is higher than that of the vacuum annealedspecimen in the same temperature, T8V-1, by SOMPa, that of the ion-nitiridedspecimen, T8N-5, becomes increased by removing the nitrided surface layer bySOMPa and this value is smaller than that of the no heat treatment specimen,TNo, by 20MPa. This is considered that the fatigue limit of Ti-6Al-4Vspecimen will be deteriorated by lOOMPa under heat treatment in thetemperature of 780 °C and increased by SOMPa under ion-nitriding. When thesurface brittle layer was removed after ion-nitriding, its fatigue limit alsobecomes increased by SOMPa (see Figures 5 and 6). Even if the surface brittlelayer was removed, nitrogen diffusion layer exists at the surface of the specimenand this diffusion layer raises the fatigue limit of the specimen by SOMPa.Static tensile tests had been already performed using the ion-nitrided specimen,T8N-1, and annealed one in vacuum, T8V-1, for the purpose of investigatingthe effect of surface layer on crack initiation.



Figure 7 shows the successive observation results of surface state duringstatic tensile test, where £ , means the total strain ( elastic strain + plastic

strain) [6,7]. The black parts are a phase and white parts here and there are

/3 phase in the matrix. Figures 7 show the ion-nit ride d specimen (T8N-1) andthe annealed specimen in vacuum, T8V-1, in the same heat treatment condition.In Figure 7, when the total strain becomes about 0.7%, which is within elasticrange, micro cracks have already initiated at a phase in perpendicular toloading direction. The length and number of cracks become increase accordingto increasing strain.

From the above results, even if there exists the very thin surface layer, i.e.,about 2fJ.m in the thickness, micro-cracks initiate at less than 0.7% of the totalstrain and initiate in the perpendicular to the loading direction.

Figure 8 shows the representative successive observation results obtained byreplica method for the fatigue crack initiation behavior of the specimen, T8N-1,under the stress amplitude of 450MPa. Fatigue cracks initiate at 4X10* cycles,

and the cracks propagate and reach to be a final fracture at 5.34 X10* cycles.That is, as the fatigue cracks initiate at about 75% of fatigue life ratio, thespecimen breaks as soon as the fatigue cracks initiate. However, fatigue cracksnot always initiate from the partial notch, because of severe limitation of the slipsystem in this material. Therefore, it is not certain whether this crack in Figure8 would be the main crack for fracture or not.

Figure 9 shows the representative successive observation results obtained byreplica method for the fatigue crack initiation behavior of the specimen, T7N-5,under the stress amplitude of 400MPa. From this figure, the fatigue crackinitiates at the small surface defect, which was made during ion-nit ridingprocess by 1X10* cycles. The fatigue cracks initiate at about 10% of the

fatigue life ratio, they propagate and reach to be a final fracture by 1.33 XIO^cycles. When comparing Figures 8 and 9, there is much difference about thecycle ratio for fatigue crack initiation. The cycle ratio for fatigue crackinitiation becomes rather small, when fatigue cracks initiate at some surfacedefects. While, it becomes larger as shown in Figure 10, when fatigue cracksinitiate in a phase itself.

In addition, according to the observation results of surface states of thespecimens, T7N-5 and T8N-5, subjected to the stress amplitude of fatigue limitby 1X10^ cycles. It is found that the non-propagating micro-cracks are notexisted in these specimens. The fatigue strength of nitrided specimen becomesslightly smaller than that of non-nitrided one and becomes considerably largerthan that of the specimen with the same heat treatment history as the above.

Conclusion

Fatigue tests had been accomplished by rotating bending fatigue testing machineusing a partial shallow notch of ion-nitrided Ti-6Al-4V alloys after the different



temperatures of heat treatment. In addition, the specimen's surface has beensuccessively observed by replica method during tests.

The main results obtained in this study are as follows:(1) There exists a chemical compounds layer, which consists of Ti and nitrogenof 2 fJL m in the thickness, and the diffusion layer of about 10 IJL m. The

micro-Vickers hardness number of the surface becomes over 1,000 .(2) The fatigue strength of ion-nitrided specimens are inferior to those of non-ion-nit ride d specimens. On the other hand, that of ion-nitrided specimens aresuperior to those of annealed specimens in the same temperature condition asthe nitriding.(3) Though the fatigue limit of the specimens nitrided at 780 °0 for 5 hours

roughly becomes the same as that of the specimen nitrided at 680 °C for 5 hours,

the former's fatigue strength is inferior to that of the specimen nitrided at 780 °Cfor 1 hour.(4) As the fatigue strength becomes improved by removing the chemicalcompound surface layer, this chemical compound layer deteriorates the fatiguestrength of nitrided specimen. From the above results, the surface layer wouldbe requested to have the equivalent toughness to the matrix for improving itsfatigue strength from the view point of practical use.

Reference

1. Morita T., Shimizu M. and Kawasaki K., JSME (in Japanese), FatigueProperty of Nitrided Ti-6Al-4V Alloy, Vol.56, No.529, pp!915-1919,1991.

2. Morita T., Shimizu M. and Kawasaki K., JSME (in Japanese), The Effectof Surface-Hardened Layer Produced by Gas-Nitriding on the FatigueStrength of Pure Titanium, Vol.59, No.563, pp!650-1655, 1993.

3. Tokaji K,Ogawa T. and Shibata H., Science of machine (in Japanese),The Effect of Nitriding on Fatigue Properties of Pure Titanium andTitanium Alloy (1), Vol.46, No.8, pp852-857, 1994.

4. Tokaji K.,Ogawa T. and Shibata H., Science of machine (in Japanese),The Effect of Nitriding on Fatigue Properties of Pure Titanium andTitanium Alloy (2), Vol.46, No.9, pp944-948,1994.

5. Nishida S., Hattori N. and Watanabe K., Effect of lon-nitriding onFatigue strength of Ti-6Al-4V Alloy, Proc. The 22nd Symposium onFatigue, SMSJ (in Japanese), pp221-224, 1994.

6. Nishida S., Hattori N. and Nisitani H., Tensile Fracture of Ti-6Al-4VAlloy, TTz Pr rmf G/7 ME (m Japa/zga?), No.958-2, pp38-40, 1995.

7. Nishida S., Hattori N. and Nisitani H., Tensile & Fatigue Properties oflon-nitrided Ti-6Al-4V Alloy, Preprint of Japan Congress of Materials



9.

f, pp7-8, 1995.Nishida S., Hattori N. and Watanabe K., Fatigue Properties of lon-nitrided Ti-6Al-4V Alloy, Proc. of Asian Pacific Conf. for Fracture andSfrfTzgf/z 96, pp447-452, 1996.Nishida S., Hattori N., Fatigue Strength of Ion Nitrided Ti-6Al-4V Alloy,Proc. 74th JSME Fall Annual Meeting (in Japanese),pp223-224, 1996.

Table 1 Chemical composition. massAl V Fe O N H6.1 4.2 0.15 0.14 0.011 0.010 0.0043

Table 2 Heat treatment condition.SymbolsTNo

T8N-1

T8V-1

T8N-5

T7N-5

T6N-5

T6NS-5

T6V-5

T5NS-5

T5V-5

Kinds of heat treatmentNo heat treatment

780°CXlhr lon-nitriding

780°CXlhr Vacuum annealing*

780°C X5hrs lon-nitriding

680*C X5hrs lon-nitriding

580°CX5hrs lon-nitriding


580°CX5hrs Vacuum annealing*


480°C X5hrs Vacuum annealing*

Type of specimenA

A

A

B

B

B

B

B

B

B

RemarkX

CD

#

• e

•+

f

<e>

K

AT6NS-5 and T5NS-5 were nitrided with steel plate.* ; No lon-nitriding© : Nitrided surface layer removed



-©.

80

2

10

0•©.

\

G

30

210

^

10

^

/

iI

80

(a) Type A

ci\L r\ -/^f_LJ-s-r

20

60

oo

~ ,

(b)TypeBFigure 1: Shape and dimensions of specimens.

Nitrided surface layer

Matrix(Ti-6Al-4V)

4 » Axial direction 10 JJL m

Figure 2: Microstructure of nitrided Ti-6Al-4V alloy.



1400

1200

1000

800

600

400

200

0

The number in parenthesis meansloading weight; unit N

tO. 049'

*\

0) T8N-1

4- T6N-5

« T6NS-5

* T6V-5

X T5NS-5

A T5V-5

0 0.5 1 1.5 2 2.5 3 3.5 4

Depth of Vickers hardness indentation mark, h //m

Figure 3: Micro Vickers hardness distribution of nitrided Ti-6Al-4V alloy.

CO ~rf\f\A 700

«3lO cnnDUO

Q)

r> cs\n-+-> 500

(0 /inn400(A0)i %•*•* *5f\r\oo 300

CD ^_ ^ %#

»*

0 i

0

* CD®w »

* TNo

(D T8N-1

® T8V-1

^

^

rp

-->

1x10^ 1x10* 1x10*

Number of cycles to failure,

1x10*

Figure 4: S-N curves.



« 700

0)-o

a.coCOc/>Q>-«->CO

600

500

400

300

• T7N-5CD T8N-1• T8N-5© T8N-5 Surface layer ren

; • !AN • !^f ^ATN ^

; #0*"% Q^

: W 0* •

;

fioved

ii

! ^

2V

1x10* 1x10* 1x10* 1x10*

Number of cycles to failure, Nf

Figure 5: S-N curves.

* 5000_

* 450\o

*? 400E•r—

o 350o>

300

x No••• Ion

: M. * T^»! ®<6>A Vac: ! : © Sur

,

i

* •

; i

heat t-nitri-nitriuum anface 1

©_

#

«

, 1 . , . . . , , , .

reatmentding (T-N)ding (T-NS) -nealingayer removed

i

300 400 500 600 700 800 900 1000

Heat treatment temperature,*C

Figure 6: Relationship between fatigue limit and heat treatment temperature.



T8N-1 (see Table 3)

Figure 7: Successive surface observation of crack initiation [6],

(7 ,= 450 MPa, Nf = 5.34X10" cycles

Figure 8: Successive observation of fatigue crack initiation (T8N-1).

^

N=0

' "^" %' , ~ -" ' ,\-\ * >,Y

,-" '/ y%

# - ;-.'*"•'>-

2x10"

: \. \C(C',S

.. \: " *-;-' '; -i*y -** "y ' " '

$

3x10^

#

4- 10"

\ .. '

. •;•% %\4 ^ \ \ " -1- :^ •- ^^.i ^ '- ^

f J' Jv:L-'\

5AlO"'cvcks

,= 400 MPa, N, = 133X10^ cycles < ^ Axial direction 50 // m

Figure 9: Successive observation of fatigue crack initiation (T7N-5).


surface treatment, computer methods and experimental ... › secure › elibrary › papers ›...

Documents