Exp Tech

DOI 10.1007/s40799-017-0183-4

Experimental Investigation of Shot Peening on Case

Hardened SS2506 Gear Steel
E. Nordin1,2 · B. Alfredsson2

Received: 29 October 2015 / Accepted: 17 April 2017

© The Author(s) 2017. This article is an open access publication

Abstract Shot peening is a manufacturing process com- Introduction

monly used to increase fatigue life in components for the
automotive and aircraft industry. In this paper the effect of Shot peening is a manufacturing process where cold work-
shot peening is described for a case hardened gear steel. ing of the surface is achieved by a propelled stream of
For gears there are three main factors from shot peening spherical shots, often called media. The process is mainly
that influence fatigue life: residual stresses, microstructure used to improve fatigue properties in mechanical parts by
and surface roughness. The paper describes an experimental introducing compressive stresses in the surface layer. The
series where these parameters were measured for common fatigue performance of gears can be drastically improved by
industrial shot peening settings. The aim was to show how the shot peening operation which is used in several indus-
several different measurement techniques and results corre- trial sectors [1]. In the aircraft industry it is also used to
spond to each other for certain shot peening parameters. The form metal panels, called peen forming [2]. In the medical
aim was also to gather experimental results that can be used implant industry it is not only used to increase fatigue and
for verification of shot peening simulations. To simplify fretting fatigue resistance but also to increase the physiolog-
measurements and decrease variation, flat steel plates were ical acceptance by adjusting the surface roughness [3, 4].
used as targets. Residual stress, full width at half maximum Other areas are inhibiting stress corrosion cracking; close
(FWHM), retained austenite, surface roughness/topology, pores in castings to prevent leaking; create surface rough-
hardness and Barkhausen noise were measured and related ness that improve adhesion of paint; adjust surface profile
to microstructural changes. The mean indentation diameter for enhanced oil retention of seals [5]; creating surface
was measured for individual shots at low coverage of each topologies that improve fatigue life of tools, gears and cams
intensity which was used to determine the average velocity [6]. There are three important properties influenced by shot
of the media. The mean diameter and hardness of the shot peening when fatigue is considered: residual stresses, work
peening media was also determined. hardening and surface topography [7, 8]. Schulze [9] gives
a comprehensive overview of the properties that shot peen-
Keywords Shot peening · Residual stress · Retained ing influences and how different manufacturing parameters
austenite · Surface roughness · Vickers hardness · affect the result.
Barkhausen noise · EBSD There are three main parameters that specify the shot
peening: intensity, coverage and media size. The media size
is a measure of the diameter of the shots, coverage is the
ratio of indented area to the total area of the target and inten-
 E. Nordin sity is a measure of the media’s ability to create indentations
[email protected] [10]. Intensity is quantified by using standardized Almen
1
strips. The procedure is explained in detail by Kirk [10] and
Scania CV AB, Södertälje, Sweden
in SAE J443 [11].
2 Department of Solid Mechanics, KTH Royal Institute Kobayashi et al. [12] did some original investigations
of Technology, Stockholm, Sweden on shot peening by dropping large bearing balls onto steel
 Exp Tech

plates and measuring the residual stresses in the plate sur- can be heated to decrease its flow stress. Harada et al. [17]
face. They compared the results with static indentations by found an optimum working temperature at 300 ◦ C; above
loads that created the same indentation diameter as the drop 300 ◦ C the maximum residual stress decreased.
tests did. The dynamic impacts resulted in a tensile residual The compressive stresses near the surface are balanced
surface stress while the static indent had zero stress at the by tensile stresses deeper inside the material. Menig et al.
centre of the indent. However, by creating more dynamic [18] used neutron diffraction to experimentally show that
indentations around the first indent the tensile surface stress the tensile stresses were evenly distributed inside a sample
changed to a compressive stress. which had been shot peened on both sides.
Wohlfahrt [13] compares a collection of shot peening The measure full width at half maximum (FWHM) is the
experiments from different authors from 1956 up to 1983 width of the X-ray diffraction peak at half of its maximum
and discusses the dependence of the residual stress pro- value. It is obtained simultaneous to the residual stress mea-
files on the relation between media and target hardness. surements when using X-ray diffraction. Larger diffraction
Wohlfahrt divided the creation of residual stresses in two peak widths generally indicate larger amounts of cold work
processes: direct plastic elongation of the surface layer and [19–21]. Zinn and Scholtes [22] noticed such behaviour
plastic deformation at the maximum shear stress depth due for annealed 42CrMo4 steel. However, when the steel was
to the Hertzian pressure. He found that target hardness that quenched and tempered to a hardness of 500–520 HV a
is much lower than the media hardness tends to deform the decrease in FWHM occurred from shot peening. Schulze [9]
surface layer. This gives high surface roughness and the explains this decrease by hard materials having very high
maximum compressive residual stress occurs at or near the dislocation densities and that the shot peening led to dis-
surface. Medium target hardness which is about the same location rearrangements with energetically more favourable
as the media hardness gives lower surface roughness and a structures.
distinct maximum compressive stress under the surface. If Hasegawa et al. [23] studied the temperature change of
the target is harder than the media, then there will be small the surface of the shot peened target. The temperature was
amounts of direct surface plastic deformation and low sur- measured with an infra-red thermograph. A case hardened
face roughness. The maximum compressive stress that can steel with hardness 700 HV showed no practical tempera-
be attained beneath the surface can be large, both absolutely ture change when shot peened while an annealed steel with
and relatively compared to the residual stress in the surface. hardness 220 HV displayed a 40 ◦ C rise. A much higher
Residual stresses are mostly measured with X-ray abrasive wear of the soft steel was suggested as the cause of
diffraction or hole drilling techniques but there exist other the temperature change.
methods, like relaxation procedures [14]. Wang et al. [15] Prevey and Cammett [24] investigated the effect of cov-
shot peened several different steels and aluminium alloys erage on AISI 4340 and conclude that both the residual
with cast steel shots of 44–48 HRC hardness. They found stress profile and cold work were saturated at only 20%
linear relations between the maximum achievable compres- coverage. The bending fatigue life with R = 0.1 was also
sive residual stresses and the materials yield and ultimate highest for the 20% coverage and was lower for higher
tensile strengths. They also quantified the depth of the coverages. This is in conflict with recommendations to
compressive residual stress maximum as a function of have at least 100% coverage. Prevey et al. explain this by
the peening intensity and the materials yield and tensile suggesting that previous investigations of fatigue life usu-
strength. Ahmad and Crouch [16] shot peened a case car- ally used fully reversed bending which can cause reversed
burized gear steel, SAE 8620, and concluded that a higher yielding with loss of compressive residual stresses and the
shot hardness typically produces higher compressive resid- fatigue improvement seen is mostly due to work hardening.
ual stresses, dual peening increases the magnitude of the Torres and Voorwald [25] also investigated AISI 4340. They
stresses and by removing the oxidized upper surface layer varied the intensity while keeping the coverage constant at
the surface compressive residual stress is increased. Harada 200%. An optimum fatigue life for rotating bending was
et al. [17] measured compressive residual stresses up to achieved for a medium intensity.
2000 MPa in a case carburized steel SCM420 target when Ando et al. [26] investigated how retained austenite con-
cemented carbide shots of hardness 1405 HV were used. If tent and peening under pre-stress influences the induced
regular cast steel shots with 490 HV hardness were used compressive residual stresses. They concluded that strain-
then compressive residual stresses of about 1000 MPa were induced martensite that is transformed from retained austen-
measured. The Vickers hardness of the workpiece was also ite during shot peening does not influence the surface
increased to about 880 HV for the cast steel shots and to and maximum residual compressive stress. Ishikura et al.
1050 HV for cemented carbide shots. Very high hardness [27] studied the case carburizing steel SCM420H with dif-
shots do however have the drawback that they easily frac- ferent hardnesses and microstructures. Shot media with
ture. To increase the service life of the media the workpiece hardness 950 HV was used. The results indicated that the
Exp Tech

amount of retained austenite does not greatly influence peening also decreased the amount of retained austenite
the maximum residual stress. The maximum achievable from 10% to less than 1%. Marconi remarks that the phase
compressive residual stress is linearly dependent on target transformation of retained austenite, which is non-magnetic,
hardness as long as the media hardness is higher than the to the magnetic martensite will increase the Barkhausen
target hardness. Tsuji et al. [28] reached the conclusion that noise amplitude, thus counteracting the decrease the com-
the achievable compressive residual stress is about 60% of pressive stress should give.
the yield stress. The increased compressive residual stress A relatively new method of shot peening is called
when retained austenite was transformed was attributed to “severe shot peening”. The aim is to produce a very fine
the increased yield stress. On the other hand, Van Wijk et al. grained (nanocrystallized) surface material. The shot peen-
[29] studied a shot peened carbonitrided steel and noted that ing is done with high coverage and/or high intensity.
when the shot media was softer than the target, 100% of the Bagherifard et al. [34] used an intensity of 0.18 mm Almen
residual stress can be explained by retained austenite trans- C and a coverage of 1500%. The average grain size was
formation. When the shot was harder than the target the measured to 50 nm and the fatigue life improved with 246%.
contribution from plasticity and phase transformation to the Hassani-Gangaraj et al. [35] used high coverage to create
resulting compressive residual stress was about 50% each. a fine grained structure that more easily diffuse nitrogen,
Barkhausen noise is a phenomenon that occurs when a decreasing the cost of gas nitriding.
ferromagnetic material is subjected to a varying external Zhang et al. [36] compares traditional shot peening and
magnetic field. The magnetic moments of atoms organize micro-shot peening with respect to the scuffing resistance.
into magnetic domains with constant magnetic field. If the They conclude that shot peening increases the scuffing
magnetic field is varied, each separate magnetic domain will capacity and micro-shot peening can improve the scuffing
change its field suddenly. This creates a stepwise increase load up to 3 times. Micro-shot peening can also decrease the
in the magnetic response of the material which is called surface roughness compared to traditional shot peening.
the Barkhausen effect. The stepwise magnetic change is The purpose of this study was to experimentally deter-
influenced by the microstructural state of the material [30]. mine the effects of shot peening at industrial conditions on
Suominen and Tiitto [31] investigated three different types parameters important for rolling contact fatigue in a com-
of steels that were shot peened. The results show that the mon case hardened gear steel. Apart from the parameters
change of Barkhausen noise and the FWHM does not only that influence fatigue, parameters that are important for the
depend on the hardness before shot peening but also on the simulation of shot peening were included in the investiga-
specific types of steel. A martensitic precipitation harden- tion. In particular coverage, intensity and media size were
ing stainless steel, 15-5PH, with 45.3 HRC hardness and varied within normal process specifications: coverage was
2.27◦ FWHM before shot peening, showed a decrease in 100% – 400%, intensity was 0.2 – 0.5 mm Almen A and the
Barkhausen noise and an increase in FWHM, both indicat- media was cut wire of 0.7 and 0.35 mm nominal diameters.
ing work hardening. AISI 4330M with a hardness of 43 Double shot peening was also investigated. The specimens
HRC and 3.05◦ FWHM had small changes in FWHM and an were then first shot peened with the larger media and in
increasing Barkhausen noise. AISI 4340M with hardness of a second operation shot peened with the smaller media.
57.5 HRC and 5.45◦ FWHM showed a decrease in FWHM The investigated parameters include: residual stress profile,
and an increase in Barkhausen noise which indicated work retained austenite content, work hardening, surface rough-
softening. Moorthy et al. [32] investigated surface and sub- ness and Barkhausen noise. By measuring the parameters on
surface stresses in a case carburized steel with Barkhausen the same or identically shot peened specimens the goal was
noise. For elastic applied stresses the Barkhausen noise to produce a unique set of data for relating, comparing and
increased for tensile stresses and decreased for compres- verifying different properties against each other. Series with
sive stresses. If the applied stress plastically deformed the low coverage were included in the work for later verification
specimens and the Barkhausen noise was measured after of shot peening models and to determine the media velocity.
unloading, then a more complex behaviour was found. The
material just beneath the case might plastically deform Experimental Set-Up
while the case, with its higher yield stress, might still be
elastically deformed. With different frequency ranges the Test Specimen
behaviour at different depths could in theory be detected
but such an analysis may require further research. Marconi The test specimens used in this work are shown in Fig. 1.
et al. [33] used Barkhausen noise to monitor shot peen- They are 30 × 30 mm square plates with a thickness of 10
ing on Pelton wheels. The shot peened specimens showed mm. They were manufactured from gear blanks made of the
an increased Barkhausen noise although compressive resid- Swedish standard SS 92506 steel, see Table 1 for the mate-
ual stresses were introduced by the shot peening. The shot rial specification. This steel is approximately equal to DIN
 Exp Tech

Fig. 1 Drawing of the test

specimen (a) and photo of a
plate after case hardening (b)

(a) Drawing (b) Plate

21NiCrMo2 or SAE 8620 [37]. The surface intended to be shot peening. This allowed the analysis to focus on the shot
shot peened was ground in soft state to a surface roughness peening effect on the measured properties.
of about Ra = 0.5 µm. The plates were then heat treated
in a pusher furnace. The process time was 702 minutes for Shot Peening
carburizing and 117 minutes for the diffusion step. The tem-
perature and carbon potential was 925 ◦ C and 0.80 wt.% The compressed air shot peening equipment is shown in
for the carburizing step and 850 ◦ C and 0.98 wt.% for the Fig. 3. It was a dual pressure pot machine which means
diffusion step. The plates were then oil quenched for 5 min- that the compressed air and shot peening media was mixed
utes in 130 ◦ C and after that in 60 ◦ C for 5 minutes. Finally in a pressurized tank and flowed together in hoses to the
they were tempered in 200 ◦ C for 72 minutes. The tem- nozzles. The shot flow was controlled by a magnetic shot
pering gave a near surface hardness of 716 HV1 and the flow control [7]. The nozzles were venturi bore type with a
case depth, defined as the depth where the hardness is 550 throat diameter of 7.9 mm, an exit diameter of 12 mm and
HV1, became 1.7 mm. Figure 2 shows a cross section view approximately 80 mm distance between throat and exit. The
below the surface of the microstructure in the shot peened nozzles were positioned 15 cm above the test specimens and
case. The surface is located to the left in the figure. The with an impact angle of 90◦ . This gave a shot peened area
martensite plates are light coloured and 0.5–3 µm wide, of about 20 – 25 mm in diameter at the test specimens. To
this corresponds well with the results by de Diego-Calderón effectively cover the 30 mm wide specimens, two nozzles
et al. [38]. The retained austenite inclusions are visible with a slight offset were used, see Fig. 3. Six test speci-
as relatively round, dark structures with less than 1 µm mens and two Almen strip holders were mounted on a round
width. The fatigue properties for this material are reported disc with radius 19 cm. During shot peening the disc rotated
by Preston [39] for rotating bending and supplemented with 10 rpm. Note that each plate was only under the shot
with pulsating torsion results by Alfredsson and Olsson blast a certain amount during the rotation of the disc. The
[40]. The monotonic tensile properties of the case and core shot peening times reported in this paper were the times the
can be found in Alfredsson and Olsson [41]. One hundred machine was operating. The effective time each plate was
plates were manufactured and all were measured with a shot peened was much shorter, approximately 1/30 of the
Barkhausen probe after hardening. Five of the plates showed total time the machine was running. For the specification of
a larger variation in Barkhausen noise over the plate width the shot peening this is however not important since it is the
compared to the rest of the plates. These were discarded to coverage that is the controlling parameter.
minimize the influence of non-shot peening effects and so The media was cut wire, spherical conditioned to grade
that all plates would have a uniform condition before the G3 and hardness 58 – 63 HRC, according to AMS 2431/8

Table 1 Chemical composition of SS 92506

C Si Mn P S Cr Ni Mo Cu Al N
% % % % % % % % % % %

0.17– max 0.65– max 0.020– 0.35– 0.40– 0.15– max max 0.005–
0.23 0.40 0.95 0.025 0.040 0.70 0.70 0.25 0.35 0.050 0.015
Exp Tech

Fig. 4 Cut wire shot peening media with nominal diameter 0.7 mm.
Fig. 2 Cross section view of microstructure close to the surface after This media had been used in the shot peening machine for a long time
heat treatment and medium shot peening: light areas are marten-
site plates and dark smooth rounded spots are inclusions of retained
austenite. Scale is 2 µm
sieved and had fresh media added for a long time in the shot
peening machine.
The hardness of the 0.7 mm media was measured by
[42]. Two nominal sizes were used, 0.7 mm and 0.35 mm. embedding the shots in a mounting holder. A flat cross
Figure 4 shows the 0.7 mm media. As can be seen, size, section surface was then ground and polished to approxi-
shape and roundness differed between each shot. The 0.7 mately half the shots diameter. Vickers hardness was mea-
mm media had been used in the machine for a long time sured in the middle of 50 shots giving a mean hardness of
while the 0.35 mm was a fresh refill. More than 30 shots 742 HV1 and a standard deviation of 50 HV1. Compared to
of the media were measured with a calliper giving a mean the specification 58 – 63 HRC, which can be translated to
diameter of 0.84 mm and a standard deviation of 0.1 mm 653 – 775 HV, the measured mean was in the upper area and
for the 0.7 mm media. For the 0.35 mm media the mean the lower specification limit was reached with 2 standard
diameter was 0.36 mm and the standard deviation 0.02 mm. deviations.
According to [43], S280 has a mean diameter of 0.84 mm. Shot peening machine parameters and the intensity mea-
Fitting a normal distribution to the allowed cumulative per- surements are shown in Table 2. Media is the nominal diam-
centage for S280 [44], gives a mean of 0.81 mm and a eter of the shot peening balls, Pressure is the air pressure
standard deviation of 0.12 mm. It is therefore likely that the used to propel the shots, Mass flow is the total mass of the
difference in mean and the relatively large standard devia- balls leaving the nozzle, Duration is the total machine time
tion of the 0.7 mm media compared to the 0.35 mm media each shot peening of an Almen strip took and Arc height is
is due to the fact that the 0.7 mm media has been circulated, the corresponding measured bending height of the Almen
strip. The arc heights was measured on an Almen Gage TSP-3
[45] and a 2PF curve [46] was fitted to the Duration and
Arc height values. The intensity for the specific media,

Table 2 Shot peening intensity measurements

Media Pressure Mass flow Duration Arc heights Intensity

[mm] [bar] [kg/min] [s] [mm A] [mm A]

0.7 1.0 7.5 40, 80, 0.207, 0.238, 0.22

160, 320 0.251, 0.252
0.7 1.6 7.5 20, 40, 0.294, 0.355, 0.34
80, 160 0.369, 0.400
0.7 3.0 7.5 20, 40, 0.498, 0.510, 0.49
80, 160 0.564, 0.601
0.35 2.8 7.5 20, 40, 0.182, 0.221, 0.24
80, 160 0.256, 0.268
Fig. 3 Test specimens in the compressed air shot peening machine
 Exp Tech

pressure and mass flow is defined as the saturation point and i3 were 3 different series with increasing shot peening
which is the point on the curve when a 100% increase intensity. The last three variants were double shot peened,
in duration gives a 10% increase in arc height [11]. An meaning that they were first shot peened with 200% cover-
example from the second row in Table 2 is shown in Fig. 5. age of the 0.7 mm media with intensities i1, i2 or i3 and then
Coverage was checked for the 0.35 media by painting shot peened with 100% coverage with the 0.35 mm media
plates with a fluorescent marker pen, PeenScan Pen [47], with intensity iD.
which shines bright in ultra-violet light. When the shot
peening balls hit the surface the paint is worn off and a Residual Stress and Retained Austenite
visual indication of the coverage can be seen. Figure 6 shows
plates shot peened with six different times (the brightness The residual stress and retained austenite measurements
and contrast has been increased from the original picture to were made with an Xstress 3000 G3 X-ray diffractome-
better show in print). The plate with the shortest shot peen- ter [48]. Stresses were measured in three directions on the
ing time, 15 s, has a lot of fluorescent paint left. As the shot plates with 45◦ between each direction. Since the plates and
peening time increase more and more of the paint is worn the shot peening had no preferred direction all directions
of. From these tests a rough estimate of the time to 98% gave similar values and an average was made. A collima-
(full) coverage was found at 55 s. For the 0.7 mm media tor with 2 mm diameter was used for all measurements.
the time to reach 98% coverage was known by experience Electropolishing was used to remove material from the shot
of the operator of the shot peening machine to be 165 s. peened surface so that measurements could be made at
It was therefore not measured in the same way as the iD different depths. Only a circular area with a diameter of
media. Even though intensity will influence the coverage the approximately 8–9 mm was removed. The residual stresses
estimate of full coverage was so rough that the same time were not corrected for relaxation because of this. Each depth
was used for all three intensities. It can here be noted that removed by the electropolishing was measured by a digital
since the mass flow of both media sizes were the same, 7.5 dial indicator. There is no precise way to electropolish to
kg/min, the number of shots per minute for the 0.35 mm size an exact depth. Therefore, the depth was determined indi-
was much larger than of the 0.7 mm size. Even though the vidually for each residual stress measurement. At the same
indentation diameter of the 0.35 mm media was smaller than time as the residual stress was measured the width of the
of the 0.7 mm media this meant that the time to reach full diffraction peaks was also recorded, which is called full
coverage was expected to be shorter for the smaller media width at half maximum (FWHM). The X-ray diffraction
size. measurement was made using Cr-Kα radiation and the lat-
In practise the time to reach 98% is called full cover- tice plane {211} (2 156.4◦ ). The tube had a voltage of 30
age (called 100% here after) and multiples of that time is kV and the current were 7 mA. Stresses were determined
denoted as 200%, 300% and so on. In total there were 12 using sin2  method, 9 tilt angles () from −40◦ to +40◦
shot peening variants, also counting plates that were not shot with tilt oscillation of ± 5◦ and cross correlation technique.
peened, denoted case hardened (CH). Table 3 lists all the The exposure time at each position was 20 seconds. The
shot peening variants. Note in particular that series i1, i2 elastic constants used were E = 211 GPa and ν = 0.3. In
materials with dual phases that have different yield strength
the microstresses can differ [49]. The residual stresses mea-
sured on one lattice plane are therefore not macroscopic but
they are usually assumed to be close. Although it can be
considered industry standard to measure residual stresses
with X-ray diffraction on shot peened specimens some
caution should be considered when comparing with other
sources.
The retained austenite was measured using Cr-Kα radia-
tion with a tube voltage of 30 kV and current of 7 mA. The
ferrite phase was measured on planes {211} (156.4◦ ) and
{200} (106.1◦ ) with an exposure time of 20 seconds. The
austenite phase was measured on planes {220} (130◦ ) and
{200} (80◦ ) with an exposure time of 40 seconds. The peak
for the ferrite phase was fitted with Split Pearson VII and
the background with a linear fit. The peak for the austenite
phase was fitted with a Gauss function and the background
Fig. 5 Calculating the intensity for the second row in Table 2 with a parabolic function.
Exp Tech

Fig. 6 Measurement of
coverage for the 0.35 mm media
with PeenScan pen. The number
beneath each plate is the number
of seconds each plate had been
shot peened

15 s 30 s 45 s 60 s 75 s 90 s

X-ray diffraction measurement uncertainty shows the max estimated uncertainty range for the depth
measurements.
The measurement of residual stresses included several com-
plex experimental and analytical sources of uncertainties. Surface Measurements
These can broadly be categorized into instrument, specimen,
operator and elastic constants related. The elastic constants Surface roughness was measured with a Mitutoyo SJ-301
E and ν are preferable determined by measurement but if surface roughness tester. The instrument is a mobile, sty-
that is not done the operator has to decide what values to lus type, roughness tester with a separate probe that uses
use. In this case 211 GPa and 0.3 was used for Young’s mod- a lateral skid type pick-up. Using an adjustable arm on a
ulus and Poisson’s ratio. The uncertainty in the calculated magnetic base the probe can be positioned for stable and
stress due to uncertainty in elastic constants was assumed aligned measurements. The measurements were made trans-
to be 5%. The instrument related uncertainties were esti- verse to the grinding marks. Each variant of shot peening
mated during the analysis as the goodness of fit of the sin2  was measured on 5 individual plates and an average was
plot. The methodology to estimate the combined uncer- calculated.
tainty was performed according to [50]. The uncertainty The surface topography was measured with confocal
in depth depends on the repeatability of the experimental microscope PLu 2300 from Sensofar [51]. An area of
set up, the accuracy of the indicator, the flatness of the 665 × 659 µm was measured with 2 µm spacing between
etching and the location of the diffraction beam. The accu- points. In height direction, planes were scanned with 0.2 µm
racy of the digital indicator used for the depth measurement apart. The measured areas were then evaluated and plotted
was 3 µm. The repeatability of the carriage used to hold in the analysis software SensoMap 4.1 [51].
the specimen was 5 µm. The flatness of the etching was
evaluated within an circle with 2 mm diameter which cor- Hardness
responds to the size of the X-ray diffraction beam. The
height difference due to the etching process was assumed to Vickers hardness was measured with a Matsuzawa MXT
increase with increasing etch depth and level out to a con- 30 hardness tester equipped with a digital camera and the
stant value of 35 µm at depths deeper than 0.3 mm. The Buehler Omnimet MHT analysing software. Measurements
combined standard uncertainty was then expanded to corre- were done on polished cross-sections of the plates to obtain
spond to 95% confidence using a multiplication factor of 2. the hardness as function of depth below the shot peened sur-
To improve the visibility of the curves individual error bars face. A force of 100 gf was used which gave an indent size
were not plotted for the depth. A depth legend in each figure around 15 µm. This is considerable larger than the typical
size of the martensite and the retained austenite structures
which are less than 1 µm in the current microstructure, see
Table 3 Shot peening variants Fig. 2. Each micro-hardness indent will cover a number
of martensite grains and in most cases a retained austen-
Label Media Intensity Coverage ite inclusion. Therefore, the micro-hardness measurements
[mm] [mm A] [%] can be seen as good averages of the local hardness and the
i1 0.7 0.22 200 values will be comparable to macro-hardness values. Also,
i2 0.7 0.34 100, 200, 400
according to Parrish [52] the austenitic grain size does not
i3 0.7 0.49 200
influence the hardness value to any great extent.
iD 0.35 0.24 100, 200, 400
For the Vickers hardness measurement the scatter was
estimated based on the non-shot peened plate and a repe-
i1+iD 200 + 100
tition measurement on plate i2-200%. The estimated scat-
i2+iD 200 + 100
ter was ± 15 HV0.1 and it was assumed equal for all
i3+iD 200 + 100
indents. The scatter in depth measurement was estimated to
 Exp Tech

0.01 mm based on the position accuracy of the equipment for a plate shot peened with intensity i2 and 200% coverage
and measurement procedure. after heat treatment. The inverse pole figure shows a more
even or limited coloured distribution of grain orientations
Barkhausen Noise Analysis from the surface and down to a depth of around 40 µm com-
pared to the plate in Fig. 7(b) that had not been shot peened.
Barkhausen noise measurements were made with a The phase map shows fewer fcc-structures from the surface
camshaft probe and a MicroScan 500-2 unit from and down to a depth of about 80 μm.
StressTech [48], using the Rollscan mode. The gain for
the channel used was set to 50, the magnitude to 50 and
the mode to RH 125 Hz. The magnetization frequency of Results
125 Hz reaches a couple of mm down into the material.
The pick-up frequency was 2.5 MHz which records sig- Residual Stresses
nals around 40–100 µm deep [48]. These settings were
chosen based on experience to give measurement values in Figure 8 shows the residual stresses for the non-shot peened
an appropriate range for the plates. The Barkhausen noise specimens (CH) and the four different intensities i1, i2, i3
probe was moved over the plates at a certain rate while and iD, all with 200% coverage. It is clear from Fig. 8 that
recording the signal. The mean value was recorded for each the maximum compressive stress around −1200 MPa was
plate. Each variant was measured on 5 individual plates and independent of intensity and media size. When the three dif-
the mean value of these was used as the result for each ferent intensities i1, i2 and i3 from Table 3 are compared in
shot peened variant. The actual numerical values have no Fig. 8 it was concluded that the compressive stresses pene-
physical meaning. It is only the changes between different trate deeper with higher intensity. The intensity iD, for the
variants that are interesting. The Barkhausen noise value for media with small shot diameter, and i1, for the large diam-
the non-shot peened plates was therefore used as a refer- eter, were almost the same. Therefore, it was also noted
ence value and the percent of change from this value was that a larger shot diameter create deeper compressive stress
calculated for each shot peened variant. profiles for the same Almen intensity than a smaller shot.
The non-shot peened (CH) residual stresses around –100
EBSD Characterization of the Microstructure MPa are slightly lower than other researchers have reported,
e.g. Epp and Zoch [53] or Dengo et al. [54]. The reason for
Electron backscatter diffraction (EBSD) was used to char- this is partly that the case depth of 1.7 mm comprise about
acterize the microstructure on a cross section of a case a 1/3 of the plates thickness of total 10 mm and partly that
hardened plate and a plate which had been hardened and the quenching rate is more even due to the lower mass and
shot peened with intensity i2 and 200% coverage. The sam- different geometry than on a gear tooth.
ples were analyzed using a JEOL 7800F SEM fitted with Figure 9 shows the residual stresses for the i2 and iD
a Bruker e-FlashHD EBSD detector. The software used for intensities with 100, 200 and 400% coverage. The natu-
the measurements and the post-processing of the data was ral scatter in the residual stress measurements and in the
Bruker Quantax. The measurements were done with a tilt depth of material removal for the measurements was larger
angle of 70◦ , a working distance around 20 mm and an than the difference between different coverages. This indi-
acceleration voltage of 15 kV. The step size was 80 nm in cates that the shot peening could already have saturated
Fig. 2 and 250 nm in Fig. 7. the residual stresses for the present definition of 100%
Figure 7(a), (b) shows the EBSD measurement on a cross coverage.
section of a case hardened plate as a phase map in (a) and Figure 10 shows the residual stress results for double shot
as an inverse pole figure in (b). In the phase map BCC- peening according to Table 3. Compared to the single shot
structures are coloured blue and FCC-structures coloured peened results for i1, i2 and i3 in Fig. 8 it was seen that the
red. In this case it means that martensite is blue and retained maximum stresses had moved closer to the surface, which
austenite is red. The colour scheme for the inverse pole was expected from a contact mechanics view. The com-
figure is red, green and blue for Miller indices [001], [101] pressive stresses at depth deeper than the maximum stress
and [111], respectively. It can be seen that the micro struc- decrease slightly due to equilibrium.
ture consist mainly of martensite. The retained austenite is A comment should also be made on the estimated scat-
homogeneously dispersed between the martensite. The pole ter bar in Figs. 8, 9, 10, 11, 12, 13, 14 and 15 for the
figure in Fig. 7(b) show a relatively wide spread through all depth position of individual measurements made with X-
Miller indices, which suggest a random distribution of mate- ray. The scatter bar is valid for residual stresses, FWHM
rial grain orientation after case hardening. Figure 7(c), (d) and retained austenite content. This relatively wide scat-
shows the corresponding phase map and inverse pole figure ter bar was estimated for positions from the peak values at
Exp Tech

Fig. 7 EBSD result of case

carburized and hardened plate in
(a) and (b). Case hardened and
shot peened with intensity i2
and 200% coverage in (c) and
(d). Length bar is 20 µm

(a) Phase map (b) Inverse pole figure

(c) Phase map (d) Inverse pole figure

0.05–0.1 mm and below the surface. Closer to the surface Thus, there is an uncertainty in the depth positions that
the scatter decreased from the bar value and was zero at should be remembered but the actual positions in the figures
the surface. The scatter band should also be interpreted as are probably fairly correct.
a worst case scenario for individual data points. The most
probably scenario for each data point is however that its Full Width at Half Maximum (FWHM)
position is relatively correctly measured, i.e. close to the
centre in the scatter bar, which is confirmed by the reason- Figure 11 shows FWHM measurement for the non-shot
ably smooth profiles found for each series in Figs. 8–15. peened specimen (CH) and the four different intensities.
The non-shot peened specimen has a constant peak width of

Fig. 8 Comparison of residual stresses for case hardened (CH) and

case hardened followed by shot peening with the four different inten- Fig. 9 Effect on residual stresses from different amount of coverage
sities and 200% coverage, see Table 3 for i2 and iD
 Exp Tech

orientation to a depth of around 40 µm compared to the

plate in Fig. 7(b) that had not been shot peened. Less vari-
ation in grain orientation should give more homogeneous
residual stresses at a micro scale between different grains.
The reduced variation can explain the decrease in FWHM
value in the shot peened affected depth below the surface in
Fig. 11. Also, since the retained austenite has mostly trans-
formed to martensite in this area this will also reduce the
variation in residual stresses at micro scale, again suggesting
the reduced FWHM value.
A noticeable difference between the larger media with
intensities i1, i2 and i3 and the small media iD is that the
larger media has resulted in a sharp drop of the peak width at
the surface while the smaller media, iD, did not. The higher
peak width for the small media at the surface is repeated
Fig. 10 Residual stresses created by the double shot peening. The dif- for the double shot peened specimens in Fig. 12. A possible
ferent intensities i1, i2 and i3 had 200% coverage and iD had 100%
cause for this behaviour can be that iD has a relatively high
coverage
intensity compared to i1, i2 and i3. This means the iD–serie
will give more plastic deformation at and near to the surface.
about 6.3◦ , except at the surface where the peak width drops Figure 13 shows FWHM values for the different amounts
to 5.2◦ . A cause for this deviation can be seen in Fig. 7(a). of coverage for the intensities i2 and iD. The FWHM val-
The top 5 µm surface layer contains a higher ratio of red ues for the iD intensity is not dependent on the coverage
FCC-structures than the bulk material. The red structures but for the i2 intensity there is a notable change between
display a banded texture from the surface indicative of grain 100 and 200% coverage. From 200% to 400% coverage the
boundary oxides. The typical procedure is that oxygen has surface FWHM value for intensity i2 change but not below
diffused into the grain boundaries and created oxides with the surface. This is in agreement with the previous remark
alloys elements such as Si, Mn or Cr. This reduces the that iD have a greater relative plastic deformation and thus
amount of alloy elements in the surrounding bulk material the curves suggest saturated change in micro-stresses and
and lowers its martensitic start temperature. The top surface retained austenite transformation. The i2 intensity need a
layer will therefore have a more bainitic structure which will higher coverage to reach a similar saturation.
have a different FWHM response than martensite.
For the differently shot peened series the affected FWHM Retained Austenite
depth was consistent with the residual stress depth in Fig. 8
and the phase transformed depth in Fig. 7(c). The inverse The retained austenite content was measured with X-ray
pole Figure 7(d) shows a more even distribution of grain diffraction. Figure 14 shows the retained austenite in the

Fig. 11 Comparison of FWHM for case hardened (CH) and the four Fig. 12 FWHM created by the double shot peening. The different
different intensities and media sizes in Table 3 for 200% coverage intensities i1, i2 and i3 had 200% coverage and iD had 100% coverage
Exp Tech

Fig. 13 Effect of coverage on FWHM for i2 and iD Fig. 15 Retained austenite for different amount of coverage for
intensities i2 and iD
four series that had been single shot peened with differ-
ent intensities and 200% coverage. The figure includes some
0.35 mm media, see Figs. 14 and 15 which is in agreement
individual measurements on double shot peened specimens.
with residual stresses in Fig. 13 and FWHM in Fig. 13.
The influence of coverage on iD and i2 is illustrated in
However, in contrast to those results there is no significant
Fig. 15. The 200% coverage series includes all depths
phase transformation difference between the series with dif-
whereas the 100 and 400% series is recorded with some
ferent intensities i1, i2 and i3 in Fig. 14. This indicates that
individual measurements. Table 3 summarizes the shot
the transformation of retained austenite occurs even at low
peening details of the different series.
intensities. The EBSD phase map pictures in Fig. 7 can only
Overall, the retained austenite measurements agree with
give a qualitative illustration but it can confirm that retained
the residual stress and FWHM measurements. They also
austenite has transformed to a depth of about 0.1 mm. The
confirm the EBSD pictures of retained austenite islands
double shot peening points in Fig. 14 show some further
before and after shot peening. At the surface almost full
phase transformation close to the surface, which agree with
phase transformation has been attained, which is notewor-
the increased residual stress results close to the surface in
thy since full stress and strain induced phase transformation
Fig. 10. Thus, the second shot peening stage does give a
is much harder to attain in compression than in tension [55].
small effect on the near surface results for retained austen-
The 0.7 mm media was effective to a larger depth than the
ite and residual stresses. In the same way as for residual
stresses and FWHM, the different coverages in Fig. 15 did
not significantly affect the remaining amount of retained
austenite.

Hardness

Figure 16 compares Vickers hardness profiles for a non-

shot peened plate (CH) and the four intensities (i1, i2, i3
and iD), all with 200% coverage. There are some devia-
tions in the individual micro-hardness measurements, as is
illustrated in the error legend, but the trends are clear both
within each series and between the series for the different
intensities. Close to the surface the three different inten-
sities i1, i2 and i3 and large media resulted in relatively
comparable hardness values at 870 HV. Beneath the surface
and down to about 0.2 mm, a higher intensity increased the
Fig. 14 Retained austenite for the four different intensities and 200%
hardness more than a lower intensity. The small 0.35 mm
coverage, the non shot peened specimen CH (as case hardened) and media differed by displaying the highest surface hardness
double shot peening but lower hardness increase below the surface. All these
 Exp Tech

is concluded that plastic deformation hardening causes the

increased hardness from 100% to 200% coverage. This is
supported by the much higher velocity of the media for the
iD intensity (76 m/s) compared to the velocity for the i2
intensity (34 m/s), see Table 5.
Figure 18 shows the HV0.1 measurements for the double
shot peened specimens. Compared to the single shot peened
specimens in Fig. 16 there might be a slight increase in hard-
ness at the surface. As with residual stresses, FWHM and
retained austenite content the effect of double shot peening
is limited to near surface changes.

Surface Roughness

When surface roughness was measured several different

Fig. 16 Vickers hardness for the four intensities in Table 3 with 200% parameters like maximum height of profile Rz, total height
coverage of profile Rt, skewness of profile Rsk or the kurtosis of the
profile Rku were evaluated but the arithmetic mean devi-
relative differences in hardness changes are quite reason- ation Ra showed the clearest trends. Figure 19 shows the
able given the trends in phase transformations to the harder surface roughness in the form of the Ra parameter for the
martensite phase, increased compressive residual stresses different intensities. The unpeened plates had an Ra around
and work hardening by plastic deformation that the shot 0.5 µm. When increasing the intensity for the 0.7 mm media
peening indents introduced into the plates. from i1 to i2 and i3 the surface roughness also increased.
Figure 17 shows the effect of different coverages for This was because a higher intensity means a higher ball
the i2 and iD intensities. The i2 intensity seems not to be velocity and therefore deeper impacts. The smaller media
influenced by the coverage at the surface but shows an 0.35 mm had an intensity iD close to i1 but because the
increasing hardness for increasing coverage at about 0.1 balls were so small they must have a much higher velocity
mm beneath the surface. The iD intensity on the other hand to reach the same intensity. The balls therefore made deeper
shows a large increase in surface hardness between 100 and impacts and gave the largest surface roughness among the
200% coverage but no practical change for 400%. These tested variants. The last three rightmost columns in Fig. 19
results indicate that the surface hardness had saturated at show results for double shot peened plates. The plates were
100% coverage for the i2 intensity but had not saturated first shot peened with i1, i2 or i3 at 200% and then shot
for 100% coverage for iD, i.e. the small media. Since the peened with iD at 100%. Compared to the single shot
FWHM results in Fig. 13 and the retained austenite results peened plates Ra increase slightly but not as much as to
in Fig. 15 both shows a saturation at 100% coverage it reach the roughness of single shot peened iD. This was
likely caused by the deformation hardening of the surface

Fig. 17 Vickers hardness for iD and i2 intensities and different

coverage Fig. 18 Vickers hardness for the double shot peened specimens
Exp Tech

before shot peening. Observe that the scales in height and

width direction are different. Compared to the shot peened
surfaces of i1 with 200% coverage in Fig. 21(c) or i3 at
200% coverage in Fig. 21(e) it was noted that the shot peen-
ing completely removed the grinding marks. The height
of the surface roughness is however of the same order.
Figure 21(b), (d) and (f) show the surface after double shot
peening. As a reminder, the i1, i2 and i3 intensities were
with 200% coverage and then 100% coverage of intensity iD
was added to the surfaces. Compared to the plates without
the added double shot peening in Figs. 21(c), (e) and 22(c),
the double shot peening added the same higher frequency of
peaks as in Fig. 22(b).
Comparison of different coverages is shown in Fig. 22 for
Fig. 19 Mean Ra for plates shot peened with different intensities
intensities i2 and iD. The figures show the same tendencies
that surface roughness decreases with increasing coverage
for i2 and that iD has a maximum at 200% coverage, see
and residual stresses introduced in the surface by the first
also Ra in Fig. 20.
stage of shot peening.
Figure 20 shows Ra as the coverage was increased.
Mean Indentation Diameter, Depth and Volume
For the i2 intensity, Ra decreased slightly with increas-
ing coverage. This seemed reasonable since the first balls
Figure 23 shows indentations in the surfaces for a very low
would create deeper impacts. As the surface hardened and
coverage, around 20%. More than 30 indentation diameters
the residual stress increased with more impacts, the later
and volumes were measured for each intensity in Table 3
impacts would not go as deep and should therefore flatten
and the mean and standard deviation were estimated. The
earlier peaks. This argument does, not quite seem to hold for
diameter was estimated by fitting a circle to the most distinct
the small media, iD, where the surface roughness increases
edges. However, since the indents were seldom perfectly
from 100% coverage while it shows a slight decrease from
round the circle edge would often cover a larger or smaller
200% to 400% coverage. The small media size compared to
local radius than the real indent area. The volume of each
the grinding marks might be one cause for this result.
indentation was measured by extracting each indent in the
SensoMap software and evaluating the volume from the
Surface Topography
top of the indent down to the bottom. Because of the rela-
tively high peaks and valleys in the surface roughness of the
Figures 21 and 22 present images of the surface topography.
plates the mean void volume of the unpeened surface was
Figure 21 focuses on the effects of varied intensity whereas
subtracted from the indent volume to give the material vol-
Fig. 22 illustrates how coverage adjusts the topography.
ume that had been displaced by the impacts. The depth was
Figure 21(a) shows a sample of the ground plate surface
evaluated as the deepest value found in the extracted vol-
ume for each indent. Table 4 summarizes the results. It was
interesting to note that the standard deviations of the mean
diameters and indentation depth, for the three different
intensities with media size 0.7 mm, are about the same.

Media Velocity

The media velocity was not measured in the shot peening

machine and this is not usually done or needed for produc-
tion purposes. However, for simulating shot peening in an
FEM program the velocities are needed. One way to deter-
mine the media velocity is to compare the indents in Fig. 23
from shot peening at low coverage with indents made at
known velocities. The determination procedure for media
velocity is described by Nordin and Alfredsson [56]. Sin-
Fig. 20 Mean Ra for different amounts of coverage gle particles were shot against a plate (same manufacturing
 Exp Tech

Fig. 21 Surface topography of a

ground plate before shot peening
in (a). Double shot peening
topography is shown in (b), (d)
and (f). Topography for i1 and
i3, 200%, is shown in (c) and (e)

(a) No SP (b) i1 + iD

(c) i1 200% (d) i2 + iD

(e) i3 200% (f) i3 + iD

batch and 0.7 mm media as in this work) and the veloc- shows the evolution of relative Barkhausen noise with
ity was measured with photodiodes. Confocal microscope increased shot peening time. The measurements were from
measurements were made of each single indent which could the coverage check presented in Fig. 6 and the iD values in
then be compared to the low coverage plates, illustrated in Table 6. The iD values are at 55, 110 and 220 s and were
Fig. 23. Table 5 shows the Almen A intensity, the corre- the averages over 5 specimens. The other points were only
sponding mean indent area and estimated velocity for the measured on one specimen. The scatter was estimated from
present shot peening. the standard deviation of the 5 repeated coverages.

Barkhausen Noise
Discussion
Table 6 summarises the Barkhausen noise results for each
intensity and coverage. The values increase both with increas- The residual stresses measured in the plates have a max-
ing intensity and increasing coverage for the single shot peened imum around −1200 MPa. Ahmad and Crouch [16] used
specimens. The double shot peened variants on the other a very similar steel, SAE 8620, and reached about −1400
hand show quite stable results, with around 240% noise MPa for a similar media hardness. The hardness of those
increase from the non-shot peened specimens. Figure 24 targets was about 61 HRC which corresponds to 722 HV,
Exp Tech

Fig. 22 Comparison of different

coverage for intensity i2 and iD

(a) i2 100% (b) iD 100%

(c) i2 200% (d) iD 200%

(e) i2 400% (f) iD 400%

quite close to the hardness on the plates in this paper (716 a harder shot media of 820 HV. It should however be noted
HV1). Adding double shot peening Ahmad and Crouch [16] that the measured hardness will depend on both the plas-
reached even higher maximum compressive stress, around –1500 tic deformation induced by the shot peening, causing either
MPa, something that was not seen in this work. Ishikura work hardening or work softening, and on the transforma-
et al. [27] reached around –1600 MPa for a case hardened tion of retained austenite into martensite. In review articles,
steel of 718 HV but with a much harder media (950 HV). Altenberger [57] and Vöhringer [58] mention that the hard-
Hasegawa et al. [23] reached around –1500 MPa in a target ness measured on shot peened specimens is increased by
plate of 700 HV and with 820 HV media. Thus, the residual compressive residual stresses. The work hardening of the
stresses measured in this investigation were slightly lower specimen is therefore misinterpreted to be greater than what
than what can be attained and there is a potential to reach it is. As a more reliable assessment of the work harden-
higher compressive stresses with a harder media. ing Altenberger [57] and Vöhringer [58] recommend the
Vickers hardness at about 40 µm depth changed from FWHM value. However, Carlsson and Larsson [59] refer
around 730 HV0.1 to around 870 HV0.1 by the shot peen- to several earlier sources and also show by numerical sim-
ing, an increase of about 140 HV0.1. This was very similar ulations that residual stresses only have a small effect on
to the results by Hasegawa et al. [23] even though they used hardness values from a sharp indenter. The decrease of
 Exp Tech

Fig. 23 Surface topography of

the ground plates with low
coverage, around 20%, but with
the different intensities in Table 3

(a) iD (b) i1

(c) i2 (d) i3

the FWHM values can be explained by the more homoge- Figures 8 and 9 display increases in compressive residual
neous orientation of the martensite as shown in Fig. 7(d). stress to depths of about 0.1 mm for the 0.35 mm media
It is therefore concluded that the hardness has increased and to 0.15–0.20 mm for the 0.7 mm media. Based on the
due to both work hardening and transformation of retained same penetration depth for residual stress increase, FWHM
austenite. change and retained austenite transformation it was con-
The profiles of hardness, retained austenite and FWHM cluded that the shot peening penetration depth was 100–150
suggested a shot peening penetration depth; a depth where µm for the 0.35 mm media and intensity iD. For the 0.7 mm
material properties were changed due to work hardening media with intensities i1, i2 and i3 it was 150–200 µm. The
and phase transformation. The FWHM curves in Figs. 11 conclusion was supported by the hardness measurements
and 13 and retained austenite curves in Figs. 14 and 15 which showed a slightly deeper depth with increased hard-
give penetration depths at about 0.1–0.15 mm for the 0.35 ness. However, in this respect it should be noted that the
mm media and 0.15–0.20 mm for the 0.7 mm media. The hardness is a more direct and straight forward measurement
hardness profiles in Figs. 16 and 17 for shot peened spec- of the mechanical properties. It is more sensitive to changes
imens display changes from the case hardened specimen and therefore showed a slightly deeper affected depth.
at slightly deeper positions, down to 0.15 mm for the The surface roughness values showed similar results as
0.35 mm media and 0.20–0.25 mm for the 0.7 mm media. reported by other researchers [57]. Ogawa et al. [60] reached

Table 4 Measured indentation

diameters, indentation depth, Label Mean diam. Indent depth Indent vol. Displaced vol.
indentation volume and [μm] [μm] [μm3 ] [μm3 ]
displaced material volume for
each intensity i1 128 ± 22 5.1 ± 0.8 38000 ± 16000 18000 ± 11000
i2 151 ± 23 5.5 ± 0.7 63000 ± 24000 34000 ± 17000
i3 175 ± 21 6.3 ± 0.9 110000 ± 32000 66000 ± 24000
iD 108 ± 15 5.6 ± 0.9 28000 ± 10000 14000 ± 8000
Exp Tech

Table 5 Measured intensity, indentation area and corresponding

velocity

Label Media Intensity Area Velocity

[mm] [mm A] [μm2 ] [m/s]

i1 0.7 0.22 15523 24

i2 0.7 0.34 22623 34
i3 0.7 0.49 33976 49
iD 0.35 0.24 10506 76

Rmax (peak to peak) surface roughness of 8 µm for tar-

get hardness around 750 HV and shot hardness of 700 HV.
Hashimoto and Hoyashita [61] observed a smoothing of the
roughness peaks but the Rmax value was almost unchanged
Fig. 24 Change in Barkhausen noise amplitude for the 0.35 mm
when shot peening a ground surface of 3 µm Rmax. A media for increasing shot peening time
shot ball of 600 HV hardness was however used on a tar-
get with 720–750 HV hardness which may have been the
cause as explained by Wohlfahrt [13]. Figure 22 shows that Conclusions
the grinding marks in Fig. 21(a) were practically removed
by shot peening. From a fatigue point of view this should Shot peening a case hardened SS2506 gear steel with the
be advantageous by decreasing the number of positions that current machines, methods and best practise gave the fol-
can initiate fatigue due to notch factor sensitivity. lowing conclusions. The conclusions assume that full cover-
Barkhausen noise results in this work showed similar age has been reached and is valid for the current media and
behaviour as results in the literature [31, 32, 62] when hard intensity range.
steels were plastically deformed. The Barkhausen signal • The maximum compressive stress was independent of
increased for increasing intensity and coverage. Just as the intensity.
decrease in FWHM values can be explained by the increased • The depth of the maximum compressive stress
homogeneity of grain orientations the Barkhausen signal increased with increasing intensity but was independent
should increase with the larger magnetic domains. Figure 24 of coverage.
suggests that changes in Barkhausen noise amplitude could • Adding a double shot peening operation with the
be an alternative method to measure coverage than by smaller media onto a surface that had already been
optical or fluorescent paint methods. There was a marked peened with a larger media increased the residual
knee in Barkhausen noise at the point that was judged as stresses by approximately 10% from the surface and
100% coverage by the PeenScan method. This can be use- down to 25 µm below the surface.
ful where optical inspection is difficult, for example inside • Shot peening effectively changed the topography and
holes. However, more experiments with different kinds of removed the grinding marks.
materials must be evaluated to confirm this behaviour. • The surface roughness increased with increasing inten-
sity for a given coverage.
• For a given intensity shot peening initially increase
Table 6 Change of Barkhausen noise amplitude in percent relative to the surface roughness. After a certain coverage had
non-shot peened specimens
been reached, the roughness decreased with increasing
Single shot peened Double shot peened coverage.
• At low coverage, individual indents could be identified
Label 100% 200% 400% Label 200%+100% and measured. The size of the individual indentations
increased with increasing intensity but the standard
i1 − 180 − i1+iD 240
deviation of the indentation size remained the same.
i2 160 210 250 i2+iD 240
• Low coverage indents are useful to save the shot peening
i3 − 270 300 i3+iD 240
impact on the target. The media velocity of the produc-
iD 220 260 290
tion equipment can be estimated if a correlation function
 Exp Tech

exist or can the attained between velocity and single 9. Schulze V (2002) Characteristics of surface layers produced by
impacts for the media and target combination at hand. shot peening. In: Wagner L (ed) International conference on shot
• peening 8 (ICSP-8), Munich, pp 145–160
Vickers hardness showed the largest hardness increase
10. Kirk D (2007) Peening intensity curves. The Shot Peener
at the surface of the target. 21(3):24–30
• Residual stress, FWHM and retained austenite are all 11. (2003) SAE Standard J443 - Procedures for using standard shot
affected to the same depth and hardness to a slightly peening test strip. http://www.sae.org/
deeper depth. 12. Kobayashi M, Matsui T, Murakami Y (1998) Mechanism of cre-
ation of compressive residual stress by shot peening. Int J Fatigue
• The Barkhausen noise relation against peening time 20(5):351–357
showed a knee at the estimated full coverage time. 13. Wohlfahrt H (1984) The influence of peening conditions on the
• The more homogeneous grain orientation after shot resulting distribution of residual stress. In: Balcar GP, Fuchs
peening can explain both the decrease in FWHM values H (eds) Second international conference on shot peening (ICSP-
2), Chicago, pp 316–331
and the increase in Barkhausen signal. 14. Totten G, Howes M, Inoue T (eds) (2002) Handbook of residual
stress and deformation of steel. ASM International
Acknowledgments The authors would like to thank Scania CV AB 15. Wang S, Li Y, Yao M, Wang R (1998) Compressive residual stress
for supporting this work and its employees for answering many ques- introduced by shot peening. J Mater Process Technol 73(1-3):64–
tions. The shot peening was made by Anders Hedberg at Arboga 73
Ytstruktur [63]. Measurements of residual stress and retained austenite 16. Ahmad A, Crouch ED (2002) Dual shot peening to maximize
were made by Thomas Hammarlund and Anna Lepeshkina at Scanias beneficial residual stresses in carburized steels. The Shot Peener
Material department. EBSD measurement were made by Peter 16(2):7–10
Hedström at Materials Science and Engineering, KTH Royal Institute 17. Harada Y, Fukaura K, Haga S (2007) Influence of microshot peen-
of Technology. ing on surface layer characteristics of structural steel. J Mater
Process Technol 191(1–3):297–301. Advances in Materials and
Processing Technologies, July 30th - August 3rd 2006, Las Vegas,
Open Access This article is distributed under the terms of the Nevada
Creative Commons Attribution 4.0 International License (http:// 18. Menig R, Pintschovius L, Schulze V, Vȯhringer O (2001) Depth
creativecommons.org/licenses/by/4.0/), which permits unrestricted profiles of macro residual stresses in thin shot peened steel plates
use, distribution, and reproduction in any medium, provided you give determined by x-ray and neutron diffraction. Scripta Materialia
appropriate credit to the original author(s) and the source, provide a 45(8):977–983
link to the Creative Commons license, and indicate if changes were 19. (2016) The measurement of subsurface residual stress and cold
made. work distributions in nickel base alloys. http://www.lambdatechs.
com/documents/204.pdf
20. Bagherifard S, Pariente IF, Guagliano M, Bandini M (2011)
Fatigue analysis of shot peened smooth and notched steel speci-
References mens. In: Champaigne J (ed) 11th International conference on shot
peening (ICSP-11), South Bend, pp 93–97
1. Breuer D, Matlock B (2011) Adaptation of shot peen parame- 21. Ludian T, Yi S, Wollmann M, Wagner L (2008) Analyzing the
ters for gear geometry. In: Champaigne J (ed) 11th International near-surface deformation gradients due to shot peening and ball-
conference on shot peening (ICSP-11), South Bend, pp 447– burnishing of alpha brass and stainless steel. In: Tosha K (ed)
452 International conference on shot peening 10 (ICSP-10), Tokyo
2. Gariėpy A, Larose S, Perron C, Bocher P, Lėvesque M (2013) On 22. Zinn W, Scholtes B (2005) Influence of shot velocity and shot
the effect of the orientation of sheet rolling direction in shot peen size on almen intensity and residual stress depth distributions. In:
forming. J Mater Process Technol 213(6):926–938 Kirk D (ed) International conference on shot peening 9 (ICSP-9),
3. Bagherifard S, Hickey DJ, Luca AC, Malheiro VN, Markaki AE, Oxford
Guagliano M, Webster T (2015) The influence of nanostruc- 23. Hasegawa N, Hattori K, Watanabe Y, Fukuyama K (1999) The
tured features on bacterial adhesion and bone cell functions on temperature change on surface of workpiece during peening. In:
severely shot peened 316l stainless steel. Biomaterials 73:185– Nakonieczny A (ed) International conference on shot peening 7
197 (ICSP-7), Warsaw, pp 54–59
4. Gorth D, Webster T, Derda J, Champaigne J (2011) Decreased 24. Prevey PS, Cammett JT (2002) The effect of shot peening cover-
bacteria functions on shot peened titanium. In: Champaigne J (ed) age on residual stress, cold work and fatigue in a Ni-Cr-Mo low
11th International conference on shot peening (ICSP-11), South alloy steel. In: Wagner L (ed) International conference on shot
Bend, pp 387–391 peening 8 (ICSP-8), Munich, pp 295–304
5. Lu X (2011) Shot peening - a tribological approach. In: Champaigne 25. Torres MA, Voorwald HJ (2002) An evaluation of shot peening,
J (ed) 11th International conference on shot peening (ICSP-11), residual stress and stress relaxation on the fatigue life of AISI 4340
South Bend, pp 435–439 steel. Int J Fatigue 24(8):877–886
6. SAE International (2001) Warrendale, PA. SAE manual on shot 26. Ando K, Eto H, Matsui K (2002) Influence of retained austenite,
peening strain-induced martensite and bending stress upon shot peening-
7. Baiker S (ed) (2006) Shot peening - A dynamic application and its induced residual compressive. In: Wagner L (ed) International
future. MFN publishing house, Wetzikon conference on shot peening 8 (ICSP-8), Munich, pp 191–199
8. Llaneza V, Belzunce FJ (2015) Study of the effects produced by 27. Ishikura R, Kano T, Kobayashi Y, Ujihashi S, Okumura K (2008)
shot peening on the surface of quenched and tempered steels: Influence of workpiece conditions on residual stress induced by
roughness, residual stresses and work hardening. Appl Surf Sci shot peening. In: Tosha K (ed) International conference on shot
356:475–485 peening 10 (ICSP-10), Tokyo, pp 244–249
Exp Tech

28. Tsuji T, Kobayashi Y, Ishikura R, Inoue K (2011) Influences 45. (2016). Almen Gage TSP-3. http://www.electronics-inc.com/
of mechanical properties and retained austenite content on shot- products/almen-gages/2-almen-gage/
peening characteristics. In: Champaigne J (ed) International con- 46. (2016). Saturation curve solver instruction manual. http://www.
ference on shot peening 11 (ICSP-11), South Bend shotpeener.com/tools/solver.php
29. Van Wijk S, Francois M, Sura E, Frabolot M (2011) Retained 47. (2015). Peenscan pen. http://www.metalimprovement.com/
austenite and residual stress evolution in carbonitrided shot- peenscan pen.php
peened steel. In: Scardi P, Ricardo C (eds) Materials science forum 48. (2016). Stresstech. www.stresstech.fi
681, Riva del Gardax, pp 374–380 49. Suzuki K, Shobu T (2011) Residual stresses in austenitic stainless
30. Kleber X, Pirfo Barroso S (2010) Investigation of shot-peened steel due to high strain rate. In: Scardi P, Ricardo C (eds) Materials
austenitic stainless steel 304l by means of magnetic barkhausen science forum 681, Riva del Garda, pp 278–283
noise. Mater Sci Eng.: A 527(2122):6046–6052 50. Fitzpatrick ME, Fry AT, Holdway P, Kandil FA, Shackleton J,
31. Suominen L, Tiitto K (1993) Evaluation of quality in shot peened Suominen L (2005) Determination of residual stresses by x-ray
components with barkhausen noise. In: Ericson T, Pluvinage G, diffraction. Measurement Good Practice Guide No. 52. 2005-09-
Castex L (eds) IITT-International, Detroit 01
32. Moorthy V, Shaw BA, Hopkins P (2006) Surface and subsur- 51. (2016). Sensofar. http://www.sensofar.com/
face stress evaluation in case-carburised steel using high and low 52. Parrish G (1999) Carburizing: microstructures and properties.
frequency magnetic barkhausen emission measurements. J Magn ASM International, p 104
Magn Mater 299(2):362–375 53. Epp J, Zoch HW (2014) Comparison of water jet peening and
33. Marconi P, Lauro M, Bozzolo W (2002) Shot peening on pel- laser shock peening with shot peening for the improvement of
ton wheels: Methods of control and results. In: Wagner L (ed) fatigue properties of casehardened steel gears. In: Wagner L (ed)
International conference on shot peening 8 (ICSP-8), Munich International conference on shot peening 12 (ICSP-12), Goslar,
34. Bagherifard S, Fernandez-Pariente I, Ghelichi R, Guagliano pp 214–219
M (2013) Fatigue behavior of notched steel specimens with 54. Dengo C, Meneghetti G, Dabal M (2015) Experimental analysis of
nanocrystallized surface obtained by severe shot peening. Mater bending fatigue strength of plain and notched case-hardened gear
Des 45:497–503 steels. Int J Fatigue 80:145–161
35. Hassani-Gangaraj SM, Moridi A, Guagliano M, Ghidini A (2014) 55. Hazar S, Alfredsson B, Lai J (2016) Mechanical modeling of cou-
Nitriding duration reduction without sacrificing mechanical char- pled plasticity and phase transformation effects in a martensitic
acteristics and fatigue behavior: the beneficial effect of surface high strength bearing steel. Submitted for publication
nano-crystallization by prior severe shot peening. Mater Des 56. Nordin E, Alfredsson B (2016) Measuring shot peening media
55:492–498 velocity by indent size comparison. J Mater Process Technol
36. Zhang J, Li W, Wang H, Song Q, Liantao L, Wang W, Liu Z 235:143–148
(2016) A comparison of the effects of traditional shot peening and 57. Altenberger I (2002) Alternative mechanical surface treatments:
micro-shot peening on the scuffing resistance of carburized and microstructures, residual stresses and fatigue behavior. In: Wagner
quenched gear steel. Wear 368369:253–257 L (ed) International conference on shot peening 8 (ICSP-8),
37. (2016). Material conversion table. http://mdmetric.com/tech/ Munich
InternationalMaterialGradeComparisonTable.pdf 58. Vöhringer O (1987) Changes in the state of the material by shot
38. de Diego-Calderon I, Rodriguez-Calvillo P, Lara A, Molina- peening. In: Wohlfahrt H (ed) International conference on shot
Aldareguia JM, Petrov RH, De Knijf D, Sabirov I (2015) Effect peening 3 (ICSP-3), Garmisch-Partenkirchen
of microstructure on fatigue behavior of advanced high strength 59. Carlsson S, Larsson P (2001) On the determination of resid-
steels produced by quenching and partitioning and the role of ual stress and strain fields by sharp indentation testing.: Part I:
retained austenite. Mater Sci Eng A 641:215–224 Theoretical and numerical analysis. Acta Mater. 49(12):2179–
39. Preston S (1991) Bending fatigue strength of carburising steel ss 2191
2506. Mater Sci Technol 7(2):105–110 60. Ogawa K, Yamada H, Saruki K, Yokoi M, Inuzuka M (1990) Influ-
40. Alfredsson B, Olsson M (2001) Applying multiaxial fatigue ence of residual stress on fatigue of carburized and shot peened
criteria to standing contact fatigue. Int J Fatigue 23(6):533– notched specimens. In: Iida K (ed) International conference on
548 shot peening 4 (ICSP-4), Tokyo
41. Alfredsson B, Olsson M (1999) Standing contact fatigue. Fatigue 61. Hashimoto M, Hoyashita S (1996) Improvement of surface
Fract Eng Mater Struct 22(3):225–237 durability of case-carburized and hardened gear by shot
42. SAE Aerospace (2007) AMS2431/8 - Peening media (AWCH) - peening and barrelling processes. In: Champaigne J (ed)
conditioned carbon steel cut wire shot, high hardness http://www. International conference on shot peening 6 (ICSP-6), San
sae.org/ Francisco
43. (2015). Shotpeener - Cast steel shot and equivalent cut wire shot 62. Kleber X, Vincent A (2004) On the role of residual inter-
size chart. http://www.shotpeener.com/learning/csvscwsizes.pdf nal stresses and dislocations on barkhausen noise in plastically
44. (2005). SAE Standard J444 - Cast shot and grit size specifications deformed steel. NDT&E Int 37(6):439–445
for peening and cleaning. http://www.sae.org/ 63. (2016). Arboga ytstruktur. http://www.cwst.se