Engineering Failure Analysis 17 (2010) 380–386

Contents lists available at ScienceDirect

Engineering Failure Analysis

journal homepage: www.elsevier.com/locate/engfailanal

Study of the failure of one machining tool

A.M. Irisarri, E. Silveira *
INASMET – Tecnalia, Mikeletegi Pasealekua 2, 2009 San Sebastian, Spain

a r t i c l e i n f o a b s t r a c t

Article history: This paper presents the study of the failure of one tool that was broken in two pieces induc-
Received 1 April 2009 ing a serious damage in the component that was being machined. The main difficulty for
Received in revised form 23 June 2009 this analysis came up against the need of keeping the broken tool in the as-received con-
Accepted 21 August 2009
dition due to legal requirements. Due to this situation it was not possible obtaining tensile
Available online 28 August 2009
test specimens to determine the strength and ductility of the material or metallographic
samples which could reveal its microstructure. Consequently, the study was restricted to
Keywords:
observing the fracture surfaces in the scanning electron microscope and identifying by
Non-destructive analysis
Notch
X-ray energy dispersive spectrometry, the nature of the substances which could have
Scanning electron microscopy played a role in the failure process. This analysis allowed identifying the failure origin
Intergranular fracture was sited on a notch induced on the periphery of the tool in a zone where a section change
and a marked stress concentration existed. This notch was covered by a dark substance
whose composition was near the same than that of the blueing coating, indicating that it
was already opened when this surface treatment was applied. These results, leaded to a
conclusion blaming the machining operation previous to the blueing coating for the failure.
Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Failure analysis is a process performed in order to determine the root causes or factors that led to an undesired loss of
functionality. During the investigation the analyst must collect, examine and evaluate all the available data in order to deter-
mine its plausible origins and the sequence of the events which leaded to the failure. The immediate objective is to find the
root causes of the failure in order to obtain the compensation for the induced damage even if the really important is to pre-
vent similar failures in new components. Usually these failure analyses involve obtaining samples for chemical analysis,
mechanical tests and metallographic studies. It is not necessary to remark that this process has a destructive nature and
the original as-received condition is lost.
However, in some cases the failure analysis is restricted to identifying its origins without destroying the failed compo-
nent, making it more difficult determining the root causes of the failure. This situation was found in the failure analysis
of a machining tool whose fracture induced a serious damage in the component that was being machined. The high cost
of this component induced a legal process to determine the root causes of the failure and who is responsible for it and must
pay the requested compensation but precluded destroying the part for obtaining samples even when obtaining previously
evidence of its as-received condition. Consequently, the failure analysis had to be performed using a non-destructive meth-
odology analysis, keeping the broken tool without modifying it.
The present paper describes the failure analysis of one machining tool carried out by non-destructive methods and the
provisional conclusions reached after this study.

* Corresponding author.
E-mail addresses: esilveir@inasmet.es, elenasilveira78@hotmail.com (E. Silveira).

1350-6307/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engfailanal.2009.08.006
 A.M. Irisarri, E. Silveira / Engineering Failure Analysis 17 (2010) 380–386 381

2. Experimental procedure

The first step in the analysis consisted in obtaining the maximum information on the previous history of the failed tool.
Unfortunately, the only data available was the known fact that the tool failed suddenly, damaging the costly component
which was being machined. Fig. 1 presents a general view of the two pieces in which was broken the machining tool, as
it was received in the laboratory. No information about the steel grade used for manufacturing the tool or the heat or surface
treatments was given.
It is convenient remarking that the failure analysis must possess a fully non-destructive character. Consequently, the vi-
sual examination, which constitutes always an important stage in these studies, is even more transcendental in this case
when no metallographic samples can be obtained. Not only the fracture surfaces but also the whole tool was examined both
by the naked eye and using a small stereoscopic microscope for detecting the finest facets. The limitation of no destroying
the tool precluded cutting it but, fortunately, one of the two broken pieces was small enough to allow introducing it into the
vacuum chamber of the scanning electron microscope, carrying out its fractographic analysis. Moreover, the X-ray energy
dispersive spectrometer incorporated to this microscope facilitated identifying the nature of the phases which were present
on the fracture surface.
In order to overcome the limitation arisen from the impossibility of obtaining the samples from the tool an alternative via,
based on the use of metallographic replicas for revealing the microstructure of the steel, was considered. Actually, this meth-
odology possesses a near no destructive character as the induced damage is just that due to the polishing and kindly etching
of a small area of the periphery. However, this possibility had to be also rejected as the periphery of the tool was covered by a
coating, which seems to correspond to blueing by its morphology, which precluded gaining access to the base steel without
eliminating it from a reduced area, action that was not permitted.

3. Results and discussion

Visual examination of the fracture surfaces pointed towards a fracture of brittle morphology, in good agreement with the
sudden and premature failure. Both, the naked eye examination and that more detailed, performed with the help of the ste-
reoscopic microscope, revealed the presence of a marked notch, darker than the remaining fracture surface, sited on the left
hand side of Fig. 2. The morphology of this notch points towards the machining of the section change of the tool existing at
that point, with an accentuated neck, clearly visible in Fig. 3, which presents a lateral view of the two halves of the tool. From
this origin the crack grew circumferentially until the sudden failure of the tool was induced when the unbroken ligament
was not able for withstanding the loads applied on it. Fig. 4 presents schematically the whole failure process while in figure,
can be observed the necking of the tool in this zone.
Fractographic analysis in the scanning electron microscope of the smallest piece of the broken tool yielded some very
interesting data for knowing not only the fracture origin but also the mechanism responsible for its progression. It was ob-
served that the dark notch was covered by a cracked substance deposited on it. Fig. 5 presents a micrograph obtained in this
area. X-ray energy dispersive spectrometry led to the spectrum shown in Fig. 6, where significant peaks of iron, oxygen and
carbon are observed. Considering this results it looks logical identifying the layer that covers the notch as iron oxide. The
origin of the peak of carbon is later discussed.

Fig. 1. General view of the two pieces of the failed tool.

382 A.M. Irisarri, E. Silveira / Engineering Failure Analysis 17 (2010) 380–386

Fig. 2. Upper view of one of the fracture surfaces.

Fig. 3. Lateral view of the failed tool. Indication of the failure origin.

Fig. 4. Schema of the failure process.

The shade of this notch is very similar to the colour found in the periphery of the tool where a surface coating was applied.
The morphology of this coating points towards blueing as the selected treatment. Blueing is a process in which steel is par-
tially protected against rust by an electrochemical conversion coating which results from the oxidizing chemical reaction of
iron on the surface selectively forming magnetite. This black iron oxide provides protection against corrosion after been trea-
ted with a water-displacing oil to reduce wetting and galvanic corrosion [1]. Without entering into details of this process it
must be indicated that there are two different blueing methods; acid and alkaline. The first one, although giving a layer offer-
ing better aspect and durability, constitutes a long time process. This is the reason for usually selecting the alkaline one.
The aspect of the layer deposited on the notch was very similar to that of the surface coating. To check if this similitude
arose from the same origin the X-ray energy dispersive spectrum of the periphery of the tool was obtained. Fig. 7 exhibits
 A.M. Irisarri, E. Silveira / Engineering Failure Analysis 17 (2010) 380–386 383

Fig. 5. Layer deposited on the notch where the origin of the failure was sited.

Fig. 6. X-ray energy dispersive spectrum of the deposited layer.

Fig. 7. X-ray energy dispersive spectrum of the coating applied on the periphery.

this spectrum. Comparison between Figs. 6 and 7 reveals their great similitude, just differing in the higher peaks of oxygen
and carbon present in the last one. The higher peak of oxygen in the analysis of the periphery can be justified by the delib-
erate application of blueing in this zone while in the notch this element was introduced accidentally. A plausible explanation
for the peak of carbon attributed it to the water-displacing oil applied in the blueing process or to some lubricant used in the
machining operation. Consequently, it looks very highly probable that the notch was already present when the blueing pro-
cess was carried out, being oxidized by the coating solution.
384 A.M. Irisarri, E. Silveira / Engineering Failure Analysis 17 (2010) 380–386

Fig. 8. End of the notch present on the fracture surface.

This result supports the hypothesis that the notch was generated during a deficient machining operation of the neck of
the tool, and was already present before the blueing treatment was applied. Even when most of the notch was covered by
this dark layer making difficult the observation of the fractographic facets existing below it, scanning electron microscope
examination of the end of the notch allowed detecting some features that were identified as the damage tracks left by
the cutter. Fig. 8 presents a micrograph giving an clear example of these tracks.
Once the origin of the failure was identified it rested determining the mechanism which promoted its propagation. A new
thorough examination of the fracture surface of the tool revealed a markedly intergranular morphology as it is shown in
Fig. 9. Various failure mechanisms that induce an intergranular fracture could be claimed; intergranular corrosion, hydrogen
embrittlement, quenching cracking, or temper embrittlement are among them. X-ray spectrometry allowed detecting the
presence of some chromium carbides, usually found in the tool steels and the presence of other particles whose spectrum
is presented in Fig. 10. Significant peaks of chlorine, sodium and oxygen and other smaller of potassium and sulphur are ob-
served in the last ones. Iron peak did not come from the own particles but from the steel sited below them that was excited
by the beam.
Firstly, it must be considered the possibility that the particles did not participate actively in the fracture process but they
were deposited on the crack lips once opened. This means that the failure was induced by a mechanism foreign for the par-
ticles and afterwards the fracture surface was contaminated by them. It must be against this possibility that particles of this
nature can be hardly present in the machine-shop environment where the failure was induced. Nevertheless, the cutting flu-
ids used for lubricating the interface between the cutting edge and the part could contain the so called extreme pressure (EP)
additives containing organic sulphur, chlorine or phosphorus compounds, which may chemically react with the metal sur-
face under high pressure condition. The presence of these substances in the cutting oil could explain the presence of some of
these elements on the fracture surface but it seems more logical considering that they were introduced during the machining
of the own tool and no during the latter service. This hypothesis is supported by the fact that the sodium or potassium are
not in the composition of these cutting oils but were detected in significant amounts on the fracture surface.

Fig. 9. Intergranular fracture in the crack growth area.

 A.M. Irisarri, E. Silveira / Engineering Failure Analysis 17 (2010) 380–386 385

Fig. 10. X-ray energy dispersive spectrum of some of the particles present on the intergranular fracture surface.

However, these two elements were present in the blueing environment where a solution of potassium nitrate, sodium
hydroxide, and water heated to the boiling point. If the origin of these elements was this one it would indicate that the crack
was already opened when the blueing process was applied. This fact rejects the possibility that the intergranular failure was
induced during the operation of the tool. Consequently, the damage induced in the tool before it began working was more
serious than that represented by the notch that was identified as the origin of the failure. The various possibilities are now
discussed.
One explanation would associate it to hydrogen embrittlement that was induced during a pickling operation which was
performed for removing rust previously to the blueing. In the pickling of steel the level of hydrogen absorption is strongly
affected by both the bath temperature and the nature of the acid being sulphuric the more damaging, followed by hydrochlo-
ric [2]. No information about the actual route followed for manufacturing of the tool was given and so the existence of a pick-
ling process cannot be confirmed. This shortness of information and the impossibility of destroying the broken parts for
obtaining metallographic samples which could give important additional data about the damage induced in the steel con-
stitute two decisive drawbacks for determining the failure process, precluding obtaining definitive conclusions.
The possibilities of cracking induced during the quenching and tempering of the steel must be also considered. Quenching
cracks are induced by the volume changes which occur when the steel is cooled. Stresses are set up which may cause the
metal either to distort or to crack when the ductility is insufficient for plastic flow to occur. Such cracks may appear some
time after the quenching or in the early stages of tempering. They usually found in components of irregular section and when
sharp angles or stress concentration points are present in the design. Preventing this quenching cracking hinders the in-ser-
vice failure due to fatigue emanating from a hidden quenching crack [2]. These quenching cracks lead to an intergranular
fracture similar to that observed in the present study but the geometry of the tool does not include so marked section
changes to justify the cracking. Consequently, even if cannot be fully rejected this possibility it does not look very probable.
Many alloy steels when tempered in the critical range progressively embrittled in an intergranular way. This failure is
associated with the segregation of certain elements to the grain boundaries, which reduces the intergranular cohesion of
iron. Recently, a failure associated to a cementite layer at the grain boundaries formed during the temper of the steel and
promoted by phosphorus segregation has been claimed [3]. A technique suggested for evaluating the action of this temper
embrittlement consisted in impact breaking a notched specimen inside the vacuum chamber of the Auger spectroscope and
checking the segregation of elements to the grain boundaries. This methodology is known to have allowed a precise and reli-
able determination of concentrations of segregating specimens and the interactions between the alloying elements and the
impurities. Unfortunately, the impossibility of destroying the failed tool precludes obtaining this impact specimen.
Due to the repeatedly commented impossibility of destroying the failed tool, it was not possible to reach a definitive con-
clusion about the mechanism which was responsible of the crack propagation. In any case, discarding the in-service failure
by a corrosive environment, which looks hardly probable, the other three possible mechanisms (hydrogen embrittlement,
quenching cracking and temper embrittlement) suppose that the crack has already grown when the blueing was applied.
That means that not only the crack was nucleated before this treatment was applied but also it was partially grown during
the manufacturing operations previous to the coating process.

4. Conclusions

Even if the impossibility of destroying the failed tool and obtaining specimens for mechanical testing or metallographic
samples represents a serious limitation the following conclusions can be reached. Some of them (those about the origin of
the failure and it is root cause) can be considered definitive while other ones are just formulated with a provisional character
until a deeper analysis of the broken component can be performed.
386 A.M. Irisarri, E. Silveira / Engineering Failure Analysis 17 (2010) 380–386

(a) Failure of the tool was initiated in a notch generated by the cutter during machining of a zone where a neck exists.
(b) From this origin fracture progressed circumferentially until the moment when the ligament which remained unbroken
was not large enough for withstanding the applied load and an unexpected sudden fracture of the tool in two pieces
happened.
(c) Considering that the above mentioned notch was covered by a layer of the same nature than that formed on the
periphery of the tool after the blueing process indicates that the notch was already present when this surface treat-
ment was applied.
(d) Once this initial notch was generated, fracture grew exhibiting a clearly intergranular morphology which indicates
that the grain boundaries were the weakest points of the material.
(e) Three failure mechanisms could be blamed for this intergranular fracture (hydrogen embrittlement, quenching crack-
ing and temper embrittlement). It would be necessary a deeper analysis involving destroying in order to achieve a
definitive conclusion on this point.

References

[1] Budinski KG. Overview of surface engineering and wear. In: Effect of surface coatings and treatments on wear. ASTM STP, vol. 1278; 1996. p. 4-21.
[2] Kumar BR, Bhattacharya DK, Das SK, Chowdhury SG. Premature fatigue failure of a spring due to quench cracks. Eng Fail Anal 2000;6(7):377–84.
[3] Godec M, Mandrino DJ, Jenko M. Investigation of the fracture of a car’s drive shaft. Eng Fail Anal 2009;16(4):1252–61.