Engineering Failure Analysis 7 (2000) 377384

www.elsevier.com/locate/engfailanal

Premature fatigue failure of a spring due to quench cracks

B. Ravi Kumar*, D.K. Bhattacharya, Swapan K. Das, Sandip Ghosh
Chowdhury
Materials Characterization Division, National Metallurgical Laboratory, Jamshedpur 831 007, India
Received 29 October 1999; accepted 7 November 1999

Abstract
The premature fatigue failure of a compression spring during service has been analysed. Optical and scanning
electron microscopy, hardness and tensile testing and X-ray diraction techniques were employed. Quench cracks,
presumably due to improper quenching, were observed across the cross-section of the spring. It is established
theoretically that the diameter of the spring rod was smaller than the critical diameter for this composition of steel.
X-ray diraction showed the presence of a small amount of retained austenite (average grain size of 2 mm). 7 2000
Elsevier Science Ltd. All rights reserved.
Keywords: Spring; Quench cracks; Fatigue failure; Machinery failure

1. Introduction
Compression springs are open wound with varying space between the coils and are provided with
plain, plain and ground, squared, or squared and ground ends. When springs are designed to work at
maximum stress, allowance should be made for the eect on stress of the specied type of end. The
number of active coils varies depending on the shape of the ends which inuences the load bearing
capacity of the spring. A solid end becomes inactive as the spring approaches the fully compressed
condition. Chemical composition, mechanical properties, surface quality, availability and cost are the
principal factors to be considered in selecting steel for springs. Both carbon and alloy steels are used
extensively. The change in the value of G (modulus of rigidity) resulting from a change in material must
be compensated for by changes in wire diameter, mean diameter, or number of active coils. Springs
made from dierent steels must have precisely the same spring rate. Therefore, the number of active
coils, the wire diameter or the mean diameter must dier.
* Corresponding author. Tel.: +91-657-426-0916; fax: +91-657-426-527.
E-mail address: [email protected] (B. Ravi Kumar).
1350-6307/00/$ - see front matter 7 2000 Elsevier Science Ltd. All rights reserved.
PII: S 1 3 5 0 - 6 3 0 7 ( 9 9 ) 0 0 0 4 8 - 5

378

B. Ravi Kumar et al. / Engineering Failure Analysis 7 (2000) 377384

For dynamic application, in which fatigue strength is an important factor, the performance of spring
materials depends on surface quality. As the initiation and growth of fatigue cracks is strongly aected
by surface quality, freedom from surface imperfections is of paramount importance in applications of
highly stressed springs for shock and fatigue loading.
Generally, steel springs are used in the quenched and tempered condition which gives optimum
strength and toughness. The microstructure and strength after the heat treatment process depend on the
cooling rate obtained during quenching; complete transformation to martensite is desired after cooling.
If the diameter of the rod is less than the critical diameter DC ), only martensite will result; whereas the
material will not completely harden due to martensite transformation if the diameter is more than the
DC : The tendency of a steel to become hardened due to martensitic transformation is known as
hardenability. This is a strong function of section size. Thus, the section size of the material will have to
be chosen properly to ensure a fully martensitic structure throughout the entire cross-section. No
retained austenite or ferrite or pearlite should be present in the microstructure. Ferrite and pearlite
reduces strength, whereas retained austenite leads to embrittlement [14].
Failure of springs generally occurs due to fatigue. The crack can initiate from the surface or at a
depth below the surface depending on the materials processing conditions. During quenching, if the
quenching rate is not chosen properly, quench crack/distortion may take place leading to tensile residual
stress on the surface. The presence of retained austenite is undesirable. It is also found that iron
sulphides present in steel along the grain boundaries felicitate the propagation of cracks generated
during hardening [5]. Shot peening of springs is a common practice to improve fatigue strength by
prestressing the surface in compression. However, excessive surface roughening during peening with
coarse shot lessens the benets of peening.
In the present study, a spring used in a B&W make of Ball and Race type coal pulverizer was
supplied by one of the manufacturers for failure analysis. The spring has a single coil and was used in
the coal pulverizer to maintain continuous pressure on the top grinding ring as shown in Fig. 1. The
diameter of the spring rod was 2 inches. Because of the nature of the feed in the pulverizer, the spring
experienced a spectrum of loading i.e. the spring was stressed under fatigue loading. The general
experience of the user was that these springs failed within a very short period of service life presumably
due to fatigue crack growth [6]. In the present paper, the causes of premature failure of the spring have
been evaluated.

Fig. 1. Assembly of the coal pulveriser with the spring.

B. Ravi Kumar et al. / Engineering Failure Analysis 7 (2000) 377384

379

2. Experimental
Small chips from the steel spring were taken out to analyze the chemistry of the spring material. The
chemical analysis was carried out by atomic absorption spectrometry as well as conventional techniques
and it was found to be C: 0.77, Si: 0.11, Mn: 1.72, Cr: 1.0, V: 0.07; S and P were within the limits. The
material is to be used in the quenched and tempered condition and suitable for spring manufacturing.
Samples were cut from areas near the fracture surface and away from it for metallographic studies.
Surfaces prepared by standard metallography practice were examined by optical and scanning electron
microscopes under unetched as well as etched conditions for inclusion detection and analysis. Elemental
analysis of the inclusions was carried out by energy dispersion X-ray analysis (EDX) in the Scanning
Electron Microscope (SEM). The SEM was also used to examine the fracture surface. Hardness was
measured across the cross-section of the rod with 30 kg load in a Vickers hardness tester. Tensile and
Charpy impact tests were carried out to ascertain the mechanical properties of the material. X-ray
diraction analysis was carried out for detecting the presence of retained austenite. Co Ka radiation was
used for this purpose. To support the XRD studies a special etchant was used to highlight the presence
of retained austenite in metallographically prepared samples.
3. Results
3.1. Microstructure
The microstructure of the etched specimens showed a uniform tempered martensite throughout the
cross-section of the steel bar (see Fig. 2). This type of microstructure is expected from this grade of steel
[7]. It appears that the material has good hardenability as evident from the homogeneity of the
microstructure. Further, cracks were observed to extend from the periphery to the centre of the sample
(see Fig. 3). These cracks, typical of quench cracks, result from an improper quenching treatment.
Under the SEM, some more micro cracks in the specimens near the failed surface as well as away from
the failed surface were observed (see Fig. 4). A few oxide inclusions along with slag inclusions

Fig. 2. Micrograph showing the tempered martensitic microstructure.

380

B. Ravi Kumar et al. / Engineering Failure Analysis 7 (2000) 377384

Fig. 3. Cracks across the cross-section of the spring rod.

containing Ca, Si, Mg, Al and S were also identied and the amount of those inclusions are within the
normal limits [7].
3.2. Fractography
A small piece containing the fracture surface was cut for examination by SEM. No signicant
information on the fracture surface could be obtained since the fracture surface seems to have
experienced rubbing action in a few areas. However, it could be observed from Fig. 5 that the fracture
initiated from an edge of the fracture surface (i.e., at a region on the curved surface of the spring bar).
The corresponding point is marked with an arrow in the photograph (see Fig. 6). The presence of
radiating beach marks around this region also conrmed that the cracks propagated under fatigue
loading and led to premature failure.

Fig. 4. Microcracks observed in the specimen both near and away from the fracture surface.

B. Ravi Kumar et al. / Engineering Failure Analysis 7 (2000) 377384

381

Fig. 5. SEM photograph of the fracture surface.

3.3. Mechanical properties

The macrohardness of the material was found to be 0480 VHN (equivalent to Rc 48). It was uniform
throughout the cross-section of the bar. The uniform hardness indicates homogeneity of the
microstructure and good hardenability.
The Charpy impact toughness of the spring was found to be 13 J. The is half of the specied value
(27.4 J) for this grade of steel in an annealed condition [7]. The low impact toughness is due to the
presence of the microcracks observed earlier.
Tensile tests of the spring material were performed on round specimens with gauge length along the
axis of the bar. The yield strength is 01200 MPa, UTS is 1450 MPa and the elongation is 1% (average
of three measurements).

Fig. 6. Close-up of the fracture surface.

382

B. Ravi Kumar et al. / Engineering Failure Analysis 7 (2000) 377384

Fig. 7. X-ray diractogram showing the g (austenite) peak.

3.4. X-ray diraction analysis

The X-ray diractogram of the specimen is shown in Fig. 7. The diraction pattern very clearly
indicates the (111) austenite peak in the steel. Quantication of the austenite was not possible with this
reection due to overlapping of this (111) line with the (110) line of the martensite. However, other lines
such as (200) which is normally used for quantitative analysis with Co Ka radiation was also detected.
The procedure for quantitative estimation of retained austenite is given in detail by Cullity [8].
Quantitative estimation using this line showed approximately 5% retained austenite. Further, a special
etching solution (40 g NaOH, 8 g Cr2O3 and 72 cm3 distilled water) [9] is used to highlight the austenite

Fig. 8. Presence of g (austenite) phase (white spot) in the tempered martensitic microstructure.

B. Ravi Kumar et al. / Engineering Failure Analysis 7 (2000) 377384

383

phase in the microstructure, which could be distinctly visualised as the white constituent within the dark
matrix (see Fig. 8). This conrms the presence of retained austenite in steel.

4. Discussion
The uniformity in the microstructure shows good hardenability of the material. This is further
conrmed by uniform macrohardness (VHN 480). There were radial quenching cracks across the cross
section from the surface to core of the sample.
During quenching, transformation of the spring material to martensite takes place after the surface
temperature decreases to the martensite start Ms temperature. The Ms temperature for this grade of
steel can be evaluated from the following expression [10]:
Ms 8C 561 474 C 33 Mn 17 Ni 17 Cr 21 Mo :

The elements within the parentheses denote wt% in the chemical composition. Ms evaluated from this
equation was 1228C.
If the section size varies, the cooling rate varies from the surface to the centre. There are two factors
which are used to describe the hardenability of the material with respect to section size. These are the
ideal critical diameter (DI) and critical diameter (DC) for a particular quenching medium. After
quenching in an ideal medium H 1), the diameter at which a minimum 50% martensite forms is
known as the ideal critical diameter (DI). The diameter at which a minimum of 50% martensite forms
after quenching in any medium is known as the critical diameter (DC). The exact spring diameter to be
used in practice should be comparable to the critical diameter (DC) to avoid generation of excessive
stresses.
The ideal critical diameter can be derived from the equation [11]:
DI D0I

X


fi X i :

D0I is a constant determined by the carbon content and the grain size and is available in the literature
for various carbon contents and grain sizes. Xi is the percent of alloying element, i; fi is the multiplying
factor due to element i. These are also given in ref. [11]. Thus, for the above material with a grain size
of ASTM No. 9, D0I will be 0.7 and the multiplying factors are fMn 3:2, fSi 1:1 and fCr 2:75: Thus,
DI will be evaluated to be 5 inches. The corresponding DC for water quenching with H 3 will be 4.5
inches. From the above analysis, it is found that the critical diameter is quite high compared to the
actual diameter of the spring (2 inches).
The X-ray diraction line prole and the optical micrograph showed the evidence of the retained
austenite 05% in the hardened sample even after tempering. Normally, the simultaneous occurrence
of quench cracks and retained austenite is not expected. As per the carbon level in this material, the
martensite morphology will be plate type. Retained austenite will be present in the material in the
interplate region. However, this can not be established truly from the photomicrograph (see Fig. 8). The
presence of retained austenite indicates that the austenite has a good stability against transformation.
Also, it is evident that austenite is present in the form of small islands (average size of 2 mm). The
chemical analysis of the material indicates that it has quite high Mn, which imparts stability to the
austenite during quenching.

384

B. Ravi Kumar et al. / Engineering Failure Analysis 7 (2000) 377384

5. Conclusions
The present failure is due to an improper hardening and tempering process. The presence of quench
cracks was the main cause of the premature failure of the spring.
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]

Gavrilyuk VG, Meshkov YuYa. Metal Sci Heat Treatment 1967;7:547.

Norstrom LA, Westerberg T, Ohrberg N. Scand J Metall 1980;9(2):7982.
Kozyrev GV, Toporov GV. Metal Sci Heat Treatment 1973;15(12):1064.
Sarikaya M, Jhingon AK, Thomas G. Metall Trans A 1983;14A(6):112133.
Mukhopadhyay NK, Das SK, Ravikumar B, Ranganath VR, Ghosh Chowdhury S. Engg Fail Anal 1997;4(3):16170.
Failure analysis of spring of a coal pulverizer. Report NML/MTC/1.33a/97, 1997.
ASM Handbook, vol. 1. 10th ed. OH, USA: ASM Metals Park, 1990.
Cullity BD. In: Elements of X-ray diraction. Reading, MA: Addison-Wesley, 1977. p. 411.
Marder AR, Benscoter AO. Metallography 1982;15:7385.
Pickering FB. In: Physical metallurgy and the design of steels. London: Applied Science, 1983. p. 229.
De Renata AF, Doane DF. Metal Progress 1971;65:100.