Canale L.C.F., Mesquita R.A., Totten G.E. Failure Analysis of Heat Treated Steel Components

Подождите немного. Документ загружается.

(Fig. 71). The fracture surface of crack G had the

characteristic ratchet marks or shear ridges that

indicate multiple initiation sites along the outer

edge of the bolt. The shape and appearance of

the exposed crack was indicative of fatigue.

Bolt H (Fig. 72) showed no crack indications,

using ultrasonic inspection or magnetic particle

inspection. This bolt showed signiﬁcant cir-

cumferential scoring around the periphery of

the shank, at approximately the midpoint of the

shank. More general wear was seen on the shank

than bolt B. Because no crack indication was

seen by NDT, no additional work was performed

on this bolt.

Fig. 67 Ultrasonic signal responses for each bolt inspected

386 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_351-393.pdf/Chap_11/ 18/8/2008 3:48PM Plate # 0 pg 386

Fig. 68 Ultraviolet light photographs of magnetic particle indications for each bolt inspected. NDT, nondestructive testing

Case Studies of Steel Component Failures in Aerospace Applications / 387

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_351-393.pdf/Chap_11/ 18/8/2008 3:48PM Plate # 0 pg 387

Bolt I (Fig. 72) showed scoring along the

circumference of the shank. In this bolt, crack

indications were found by NDT in both the

shank and the head. The exposed shank crack is

shown in Fig. 73. The shank fracture surface of

crack I had the characteristic ratchet marks or

shear ridges that indicates multiple initiation

sites along the outer edge of the bolt. The shape

and appearance of the exposed crack was indi-

cative of fatigue. The crack indication in the

head-to-shank ﬁllet radius of bolt I was opened

in a similar manner as the other bolts, and the

fracture surface is shown in Fig. 74. The head-

to-shank ﬁllet radius fracture surface had the

characteristic ratchet marks or shear ridges that

indicate multiple initiation sites along the outer

edge of the shank ﬁllet radius (Ref 9). The shape

and appearance of the exposed crack was indi-

cative of fatigue.

Fig. 69 As-received condition of bolts B, C, and G

Fig. 70 Fracture surface of bolt C with fracture schematic

Fig. 71 Fracture surface of bolt G with fracture schematic

388 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_351-393.pdf/Chap_11/ 18/8/2008 3:49PM Plate # 0 pg 388

The as-received condition of bolt A is shown

in Fig. 75. A photograph of the fracture surface

of the failed bolt is shown in Fig. 76. The

Fig. 72 As-received condition of bolts H and I

Fig. 73

Fracture surface of the exposed crack in the shank of

bolt I

Fig. 74

Fracture surface of the exposed crack in the head

ﬁllet radius of bolt I

Fig. 75 As-received condition of the failed bolt

Fig. 76 Fracture surface of the failed bolt

Case Studies of Steel Component Failures in Aerospace Applications / 389

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_351-393.pdf/Chap_11/ 18/8/2008 3:49PM Plate # 0 pg 389

fracture surface shows ratchet marks that are

characteristic of fatigue origins. Multiple fatigue

origins, located to one side of the bolt, were

observed. A large fatigue zone was apparent.

The ﬁnal fracture zone was badly damaged. The

bolt showed evidence of machining marks on the

sides of the shank at the site of crack initiation

(Fig. 77). Opposite the initiation site, on the side

of the shank, the surface appeared to be polished,

Fig. 77

Machining marks and polishing evident on the shank

of the failed bolt

Fig. 78

SEM micrograph of representative multiple fatigue

origins associated with machining marks on exposed

cracks (100 mm)

Fig. 79

SEM micrograph of the initiation site of the failed

bolt, showing cracking beginning at the localized

surface wear (1 mm)

Fig. 80

SEM micrograph showing initiation of cracking on

the shank of bolt I at localized surface wear (400 mm)

Fig. 81

SEM micrograph of representative fatigue striations

found on the bolt fracture surfaces (2 mm)

390 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_351-393.pdf/Chap_11/ 18/8/2008 3:49PM Plate # 0 pg 390

with more evidence of machining marks

(Fig. 77).

The fracture surfaces on bolts A, C, G, and I

were examined using the SEM. The fracture

surfaces of the exposed cracks in the bolts were

characterized by multiple fatigue origins, initi-

ating along machining marks (Fig. 78). In the

fractured bolt, initiation of cracking began at a

region of localized wear and scoring on the

shank of the bolt (Fig. 79). This was also evident

in the initiation of cracking in the shank on bolt I

(Fig. 80). Striations, characteristic of fatigue,

were seen on the fracture surface of bolt A

(Fig. 81). Machining marks found in the head-

to-shank ﬁllet radius were evidence that the

head-to-shank ﬁllet radius was machined

(Fig. 82). This is unacceptable according to the

design speciﬁcation.

A metallographic specimen was prepared

using standard techniques. This specimen was

used to examine the microstructure and micro-

hardness of bolt A. A photographic montage of

the metallographic specimen is shown in Fig. 83.

The tempered martensite microstructure of

bolt A was typical for quenched and tempered

low-carbon steel (Fig. 84). Bolt A failed in

the shank; in addition, a crack approximately

0.15 mm (0.006 in.) long was found in the head-

to-ﬁllet radius (Fig. 85). Neither magnetic par-

ticle nor ultrasonic NDT detected this crack.

The failed bolt was submitted to a micro-

hardness survey to determine the extent of de-

carburization. The microhardness survey was

taken with a Knoop indenter, with a 500 g load.

Indentations were taken at 0.0318 mm

Fig. 82

SEM micrograph of machining marks at the head-

to-shank ﬁllet radius of bolt I (1 mm)

Fig. 83

Microstructure at bolt A fracture through section 1-1.

The overall microstructure is shown as well as the

location of additional micrographs. There is no evidence of a

cold-formed head. Etched with 2% nital (667 mm)

Fig. 84

Micrograph of representative quenched and tem-

pered martensite found in the submitted bolts.

Etched with 2% nital (25 mm)

Fig. 85

Micrograph of the crack found in the head-to-shank

ﬁllet radius of the failed bolt. Etched with 2% nital

(25 mm)

Case Studies of Steel Component Failures in Aerospace Applications / 391

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_351-393.pdf/Chap_11/ 18/8/2008 3:49PM Plate # 0 pg 391

(0.00125 in.) intervals until the hardness read-

ings began to stabilize. The point at which the

hardness stabilized was called the core hardness.

For the failed bolt, the core hardness was 404

KHN. In accordance with the speciﬁcation, the

point at which the hardness is within 20 KHN of

the core hardness is deﬁned as the depth of

decarburization. For the failed bolt, this was at

0.064 mm (0.0025 in.) deep, meeting the spe-

ciﬁcation requirements. The hardness indentions

are shown in Fig. 86. The results of the micro-

hardness survey indicated that the surface

hardness (Knoop), converted to Rockwell C, is

26 HRC. The Knoop core hardness, converted to

Rockwell C, is 40 HRC. While the partial de-

carburization of the bolt satisﬁes process speci-

ﬁcation requirements, the steep gradient of the

hardness proﬁle and the surface softness mea-

sured indicated that there would be degradation

in the potential fatigue life of the bolts.

Determination of the composition of the

failed bolt was performed by arc-spark spec-

trometry. The carbon content of the steel was

determined by Leco carbon analyzer. Results are

shown in Table 6. The chemical results indicate

that this bolt does not meet the composition

requirements of Table 2 for any of the speciﬁed

bolt alloys.

The hardness of each bolt was measured using

standard procedures. The calibration of the

hardness tester was veriﬁed, using two calibra-

tion test blocks (54.1+1.0 HRC and 34.6+1.0

HRC). The measured hardnesses on the test

blocks were found to be within the range of the

hardness test block calibration. A section of

the bolt was taken by cutting perpendicular to

the long axis of the bolt, in the threaded region.

The hardness of each bolt was determined by

taking a hardness measurement on the end of the

cut section. Results are shown in Table 7.

Based on this investigation, it was conclu-

ded that:

 Ultrasonic inspection does not have the

sensitivity to detect cracks of the size and

geometry seen in the bolts.

 The fractured bolt failed by fatigue, initiat-

ing at localized wear on the bolt shank. The

soft surface hardness from partial decarbur-

ization aggravated the fatigue failure.

 The exposed cracks in bolts C, G, and I

initiated at machining marks at the head-to-

shank ﬁllet radius, and cracking propagated

by fatigue. Evidence of machining at this

radius was found. Bolt I had a second

initiation site on the shank.

 The composition of the failed bolt did not

meet the chemical requirements of S105

bolts. However, the chemistry of the bolts is

not thought to have contributed to the failure

of bolt A.

 The bolts were machined instead of cold

formed. This results in lower fatigue strength

and proof load.

REFERENCES

1. G.E. Dieter, Mechanical Metallurgy, McGraw-

Hill, Inc., 1986

2. Met allography and Microstructures, Vol 9,

Metals Handbook, 9th ed., American Society

for Metals, 1985

3. D.A. Ryder et al., General Practice in Failure

Analysis, Failure Analysis and Prevention,

Fig. 86

Micrograph of Knoop hardness (500 g load) survey of

failed bolt (100 mm)

Table 6 Measured chemistry of the failed bolt A

Element

Percent,

wt%

C 0.42

Si 0.32

Mn 0.82

P 0.010

S 0.014

Cr 0.49

Mo 0.15

Ni 0.45

Table 7 Measured hardness of submitted bolts

Bolt

Hardness,

Rockwell C

C 43.4

B 36.4

H 41.6

G 40.4

I 39.2

Failed bolt A 36.9

392 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_351-393.pdf/Chap_11/ 18/8/2008 3:49PM Plate # 0 pg 392

Vol 11, Metals Handbook, 9th ed., American

Society for Metals, 1986

4. B.E. Wilde, Stress-Corrosion Cracking,

Failure Analysis and Prevention, Vol 11,

Metals Handbook, 9th ed., American Society

for Metals, 1986

5. W.J. Jensen, Failures of Mechanical Fast-

eners, Failure Analysis and Prevention, Vol

11, Metals Handbook, 9th ed., American

Society for Metals, 1986

6. Threaded Steel Fasteners, Failure Ana-

lysis and Prevention, Vol 11, Metals

Handbook, 9th ed., American Society for

Metals, 1986

7. A.G. Glover et al., Failures of Weldments,

Failure Analysis and Prevention, Vol 11,

Metals Handbook, 9th ed., American Society

for Metals, 1986

8. E. Alban, Failures of Gears, Failure An alysis

and Prevention, Vol 11, Metals Handbook,

9th ed., American Society for Metals, 1986

9. R.J. Parrington, Fractography of Metals and

Plastics, Pract. Fail. Anal. Vol 2 (No. 5),

2002, p16

Case Studies of Steel Component Failures in Aerospace Applications / 393

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_351-393.pdf/Chap_11/ 18/8/2008 3:49PM Plate # 0 pg 393

Failure Analysis of Powder Metal

Steel Components

S. Ashok and Sundar Sriram, Sundram Fasteners Ltd.

POWDER METALLURGY (PM) technol-

ogy provi des a cost-effective method of produ-

cing near-net shape products, especially when

a large number of the same or similar products

are required. While the initial powder used

is expensive compared to wrought steels, the

manufacturing process can avoid machining

steps and provide more uniform composition

control. Total cost of producing PM parts can be

less expensive than conventional metalshaping

(casting, forging) in many cases. It was ﬁrst

adopted by the auto industry. Since this initial

inception, other industries, such as the refri-

gerator, lawnmower, and hand tools industries,

have also extensively adopted PM parts (Ref 1).

Manufacture of powder metal parts involves

pressing of metal powder in a die, sintering

the green compact so that metallurgical bonding

takes place, sizing or coining for densiﬁcation

and dimensional correction, followed by ﬁnish-

ing operations. The ﬁnishing operations include

machining, case and through hardening, steam

treatment, and so on (Ref 1).

This chapter reviews failure aspects of struc-

tural ferrous PM parts, which form the bulk of

the PM industry. Focus is on conventional PM

technology of parts in the density range of 6 to

7.2 g/cc. This constitutes 90% of the parts pro-

duced by PM technology.

The PM part undergoes multiple heat treat-

ments, some of which are unique to PM while

others are similar to that used by conventional

manufacturing processes, such as forging and

casting. These include:

 Powder annealing

 Sintering

 Case hardening

 Carbonitriding

 Case carburizing

 Through hardening

 Induction hardening

 Steam treatment

This chapter brieﬂy introduces the processing

steps involved in PM (Fig. 1). These heat treat-

ments and the PM process steps are essential

to understanding failure analysis of PM parts.

The methods used for analyzing the failures are

then discussed. Methods of failure analysis itself

have recently become more systematic in terms

of prevention and ensuring more robust process

and high-quality products. Some case studies

are given that illustrate different failures and the

methods of prevention of these failures.

Powder Metallurgy Process

Powder Production. Solid-state reduction

of iron ore and atomization of molten iron are

the most common methods of producing iron

powders for structural parts. The powder thus

produced should be favorable for compaction

such that:

 It ﬁlls the die cavity repeatably, which is

measured in terms of the apparent density

and ﬂow rate of the powder.

 It can be compressed to the desired density

with the lowest pressure, which is measured

in terms of compressibility.

 It has sufﬁcient strength at that density so

that the green part can be handled, which is

measured in terms of its green strength.

The ﬁlling characteristic of a powder is

measured by its apparent density and its

ﬂow rate. The compressibility of a powder is a

function of the particle morphology, processing

history, and purity. Water-atomized iron pow-

ders can be compressed to highe r densities

compared to sponge iron powder. The higher

the impurity content (typically oxides), the

poorer the compressibility. As a process, all iron

powders are general ly annealed in a reduc-

ing atmosphere. Special high-compressibility

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_395-415.pdf/Chap_12/ 18/8/2008 3:51PM Plate # 0 pg 395

Failure Analysis of Heat Treated Steel Components

L.C.F. Canale, R.A. Mesquita, and G.E. Totten, editors, p 395-415

DOI: 10.1361/faht2008p395

www.asminternational.org

powders are produced by double-annealing

operations.

Blending. Alloying elements are typically

mixed with the iron powder by blending in

a double-cone blender. Elemental addition of

alloying elements, that is, graphite, copper, or

nickel, has its drawback, because these elements

tend to segregate in the product as the alloying

element content increases. A recent develop-

ment in this regard is the introduction of diffu-

sion-bonded iron powder and binder- treated

premixes. These processes are proprietary. For

example, the Distaloy process is diffusion

bonding of nickel, molybdenum, and copper to

iron powders developed by Hoganas AB, Swe-

den. The Ancorbond and Starmix methods are

binder treated premixs to reduce graphite-

lubricant segregation. In the case of diffusion-

bonded iron powders, the alloying elements,

typically copper, nickel, and molybdenum, are

mixed with iron, and the mix is annealed at

a temperature of approximately 850 to 920



As a result, bonding of the alloying elements

to the iron particle takes place, and this pre-

vents segregation of alloying elements in the

products. The same principle is used to prevent

carbon-lubricant segregation. In this case, the

binder, which is added during powder mixing,

is sprayed on the iron-carbon mix. This binder

ensures that each iron particle has a coating

of graphite and thus minimizes segregation.

Case studies in this chapter illustrate this

mechanism.

Compaction. The most widely used method

of compaction is axial pressing of loose powder.

Initially, the die cavity is ﬁlled with the loose

powder from a hopper by gravity; the density is

typically the apparent density of the iron pow-

der, whi ch is 2.4 to 3 g/cc. When compaction

commences, that is, the punches move toward

each other, densiﬁcation takes place by particle

rearrangement and plastic deformation. Thin

sections of individual particles bend or break,

cold welding and interlocking of n eighboring

grains takes place, and the voids are ﬁlled by the

material, which becomes squeezed into them.

The density of 5.5 g/cc is the lowest density

of iron powder parts ﬁt for practical handling in

the green stage. On attaining this density, a rapid

increase in pressure is observed when the part is

compacted to densities between 6 and 7.2 g/cc.

Compacting in the higher-density range gradu-

ally increases the rate of plastic ﬂow of the

metal. This results in the following:

 Increased work hardening of powder parti-

cles

 Increased friction against the die walls

Structural ferrous powder metal parts typi-

cally have sections of varying thickness. With

respect to heat treatment, it is very important that

all sections in the parts have uniform density

throughout. Low sectional densities result in

through hardening and wide variation in the case

depth within the PM part. Various methods of

compaction have been developed to produce

parts with minimal difference in density. These

include:

 Double-action compaction

 Spring ﬂoating die compaction

 Compaction by withdrawal process

 Multilevel compaction using multiple mov-

ing p unches in special multiplaten presses

Fig. 1 The powder metallurgy process

396 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_395-415.pdf/Chap_12/ 18/8/2008 3:52PM Plate # 0 pg 396