Ultrasonic Characterization of Surface Modified Layers
BRET J. ELKIND, MOSHE ROSEN, and HAYDN N.G. WADLEY
Nondestructive techniques are required for the in-process characterization of rapidly solidified and
surface modified layers to fulfill the role of sensors in emerging intelligent materials processing
technologies. In steels, where surface modification via directed high energy sources is being investigated for surface hardening, it has been found that a difference exists in the Rayleigh wave
velocity of martensite and pearlite. The difference in velocity can be used to characterize the hardness
of a surface modified layer on a pearlite substrate. By varying the Rayleigh wave frequency (and thus
the depth of wave penetration) and measuring velocity dispersion, it has also been possible to
determine nondestructively the depth of modified surface layers on both AISI 1053 and 1044 steels
produced by electron beam melting.
I.
INTRODUCTION
RAPID solidification processing (RSP) refers to solidification processing where the cooling rate exceeds 102 K 9 s -~.
The emergence of this innovative approach to materials
processing has resulted in the ability to produce novel
microstructures which bestow improved properties upon
materials processed by these techniques (Figure 1). These
high technology materials possess microstructures with
enhanced chemical homogeneity, extended solid solubility,
microcrystallinity, refined precipitate distributions, and
metastable phases, some of which are amorphous. This has
resulted in significant improvements in corrosion, erosion,
wear resistance, and improved strength.
One of the methods for the production of rapidly solidified microstructures is that of surface modification by directed high energy beams (lasers and electron beams). These
sources of high intensity heat are capable of rapidly melting
(e.g., in less than a few milliseconds) a thin surface layer
before heat has appreciably diffused to the substrate. This
results in a very steep temperature gradient normal to the
surface, and thus a mechanism for rapid quenching of the
melt (by the substrate) upon removal of the heat source. By
rapidly scanning the heat source over the sample surface,
cooling rates in the 103 to 108 K 9 s -L range are achieved and
solidification velocities of up to 1 to 10 m 9 s -L induced. The
application of this process to near net shape components
eliminates the need for appreciable further processing and
thus can optimize process productivity.
Two broad classes of RSP surface microstructures are
currently being investigated. One is concerned with the formation of amorphous surface layers on crystalline metal
substrates. The second is the formation of modified crystalline surface layers with or without chemical modification.
Chemical modification with this technique can be achieved
by incorporating alloying powders into the melt prior to its
solidification.
The formation of amorphous surface layers or the layer
glazing process, as it is sometimes referred to, has been
demonstrated to be technically feasible, but its widespread
application is limited by two problems: the difficulty of
producing deep (>20 p~m) amorphous layers and the elimiBRET J. ELKIND and MOSHE ROSEN are with the Materials Science
Department, The Johns Hopkins University, Baltimore, MD 21218.
HAYDN N. G. WADLEY is with National Bureau of Standards, Gaithersburg, MD 20899.
Manuscript submitted August 15, 1985.
METALLURGICALTRANSACTIONS A
nation of cracking. The former problem arises either from
the reduction of temperature gradient for deeply melted layers, which causes a reduction of solidification velocity to a
value below the threshold for glass formation,2 or from the
occurrence of amorphous to crystalline transitions in the
heat affected zones of previously vitrified material during
sequential melting passes in adjacent material. The cracking
problem, on the other hand, arises because of the large
tensile residual stresses that are produced by solidification
shrinkage strains.
Microcrystalline modified surface layers may possess enhanced resistance to erosion, wear, and corrosion. These
beneficial attributes are important in a diversity of high
technology applications where rrfaterial performance has
been the primary limitation. If this material's processing
approach is to be utilized in practice, it is important to
control spatial nonuniformities (of microstructure and chemistry), porosity, cracking, depth of modification, and surface topology. Microstructural and chemical heterogeneity
are believed to be caused by uncontrolled variations in process variables--principally temperature gradient, solidification rate, cooling rate, and by convective mixing in the
melt. Microstructural inhomogeneities may also be caused
by the annealing of already processed material by subsequent melt passes in adjacent material.
Clearly, if the potential of surface modification with directed high energy sources is to be fully realized, techniques
(or sensors) need to be developed to measure, in process,
key process variables (those controlling microstructure) and
to detect cracking and other deleterious conditions. With
such a capability it may then be possible to use automated
process control 3 as a means for reliably producing uniformly
high quality material. Previous research4 has shown the potential of acoustic emission for detecting cracking during
surface modification. The work to be reported here is concerned with development of an ultrasonic technique to characterize the depth of surface modified microcrystalline layers and their hardness.
The basis of the technique stems from the difference in
ultrasonic velocity between the layer and substrate. In this
study, the technique has been applied to the characterization
of a surface modified layer on a steel substrate. The phenomenon responsible for the technique (the difference in
ultrasonic velocity of substrate and layer), it should be
noted, is actually more pronounced if the surface layer is
amorphous, because the elastic constants, and thus ultrasonic velocity, of the substrate (crystalline) and the layer
U.S. GOVERNMENTWORK
NOT PROTECTED BY U.S. COPYRIGHT
VOLUME 18A, MARCH 1987--473
Conventional
Microstructures [
J
r I
"~
! Composition
Refined
I and Process ~l
Microstructures]
Coarse Dendrites,
Eutectics and Other
Microconstituents
De:ent"l
lOCi
Fine Dendrites,
EutecUca and Other
Microconstituents
[
Extended Solid Solutions I
Microcrystalline Structures
Metastable Crystalline Phases ~ v 7
Amorphous So ds
Increasing
Homogeneity
Microstructures
Increasing
Cooling Rate
(K/s)
Fig. 1 - - Microstructural consequences of rapid solidification. Note: These classifications are very approximate
and depend on compositional and process variables as well as on the cooling rate.'
(amorphous) are more different. Consequently, the microcrystalline layer study is in some respects a worst case test
of the approach.
When applied to steels, electron beam surface modification has been observed to cause substantial improvements in
hardness due to the formation of martensite/bainite and
carbide precipitate refinement. 5 Systematic investigations
have shown that the hardness and depth of modification are
controlled by scanning rate. 6'7's Thus, variation of the scan
velocity could provide a convenient potential method for the
control of hardness and melt depth, provided they can be
measured in-process. The use of ultrasonic surface acoustic
waves (Rayleigh waves) has been investigated here for the
characterization of the depth and hardness of electron beam
modified surface layers.
Rayleigh waves are ultrasonic waves that propagate at the
surface of materials. The particle displacements in the wavefront prescribe elliptical motion involving both shear and
longitudinal components, the amplitudes of these motions
decreasing exponentially with depth below the surface. The
rate of decrease in amplitude is, however, dependent upon
wavelength; at a depth of approximately one wavelength,
the particle amplitude is only about 20 pct (and the particle
energy is only 4 pct) of that at the surface. The velocity of
Rayleigh waves in a homogeneous, isotropic, half-space is
dependent only upon the density and elastic constants and is
independent of wavelength (or frequency). In inhomogeneous materials with weakly depth dependent elastic
properties, however, the Rayleigh wave velocity varies with
wavelength (or frequency) (i.e., with depth of penetration)
because waves of different frequency sample a different net
modulus. Thus, the measurement of velocity dispersion
could potentially provide a rapid, nondestructive method for
selectively probing sub-surface properties. 9,~~
II.
Several plate samples, 7.6 cm x 2.5 cm x 0.6 cm in
size, were austenitized at 1025 ~ for 45 minutes. One sample was subsequently quenched in ice-brine to produce a
sample of bulk martensite. The remaining samples were
slowly (furnace) cooled to 650 ~ isothermally transformed for 15 minutes, and then air-cooled to room temperature. This resulted in samples with a bulk pearlitic
microstructure.
After completion of preliminary ultrasonic and microstructure characterizations, each of the pearlitic samples was
subjected to an identical electron beam surface modification
process. A 25 KV accelerating voltage was used to accelerate a 0.25 cm diameter, 9 mA electron beam onto the specimen surface. The beam power, calculated by multiplying the
beam current by the accelerating voltage, was 225 watts.
This beam was scanned over the sample length at the relatively slow rate of 0.3 cm 9 s -~. After each scan, the beam
was translated 0.0254 cm across the width of the specimen,
and the scan repeated so as to modify the entire surface.
The time for a single scan was 7.5 cm - 0.3 cm 9 s -~ =
25 seconds and for the entire surface modification approximately 40 minutes. Thus, substantial tempering of prior
modified material is anticipated to have occurred during
subsequent passes.
B. Specimen Characterization
Following heat treatment, the samples were characterized to provide baseline hardness, density, longitudinal and
shear ultrasonic velocity, and metallographic data of bulk
(pearlitic) and surface layer microstructures.
C. Rayleigh Wave Measurements
In the present investigation, Rayleigh wave speed measurements were conveniently made by means of the mode
conversion device shown in Figure 2. Longitudinal waves
were generated by a piezoelectric transducer attached to a
2024 aluminum wedge. The wedge made contact with the
sample over an area of 2.54 cm x 0.07 cm with a contact
angle of about 60 deg. This resulted in the conversion of
longitudinal elastic displacements to Rayleigh waves that
EXPERIMENTAL
A. Specimen Preparation
Tests were, in the main, performed on AISI 1053 steel,
the composition of which is given in Table I. An additional
AISI 1044 alloy was used to confirm the approach.
Table I.
Element
Wt pct
C
0.528
474--VOLUME 18A, MARCH 1987
Mn
0.71
Composition of AISI 1053 Steel Samples
P
S
0.012
0.022
Si
0.17
Cr
0.039
Ni
0.034
Mo
0.011
Co
0.009
Cu
0.051
METALLURGICALTRANSACTIONS A
Rayleigh Wave Mode Conversion Device
Sliding Mount
t-
II
II
II
~_._.l~_~J
Transmitting
/
Transducer ~
)
II
II Receiving
/I ~ T r a n s d u c e r
[ . ~ )
Longitudinal~ ~ / . . , . . A l u m i n u m
~ . ~ ~"
Waves
~
W.dges ~
~ff/lllllllllllllllll~'_
Surface Mod fed
Layer I
~ _.~_____
_
Non-treatedSubstrato
lr
T
Fig, 2 - - Schematic diagram of the mode conversion technique for generation and detection of Rayleigh surface waves on surface modified samples.
propagated across the surface to the second wedge. Mode
conversion at the second wedge then occurred, and the resuiting longitudinal waves propagated through the wedge
and were detected, again, piezoelectrically. The insertion
loss* at 2.25 MHz was 43 dB and compares favorably with
*The insertion loss was calculated from 20 • log~o (V/Vo) where Vo is
the voltage applied to the transmitting transducer and the V, that measured
across the receiving transducer.
other methods of generation (see page 22 of Reference 11).
By varying the resonance frequency of the piezoelectric
transducer pair, it was possible to vary the frequency of the
Rayleigh waves between 2.2 and 15.5 MHz. The lower
frequency limit was imposed by the need for the wedge
contact width (700/zm) to be greater than or equal to onehalf of the Rayleigh wavelength (1400/xm at 2.1 MHz
in steel with a Rayleigh wave speed of - 3 m m 9 -I) if
unacceptable insertion losses were to be avoided. The high
frequency limit was determined by the ultrasonic attenuation
characteristics of the material.
The Rayleigh velocity was obtained by dividing the
wedge separation by the time-of-flight (TOF) of a Rayleigh
wave pulse between the wedge pairs. The TOF's for both the
entire path (wedges and sample) and through the wedges
alone were measured separately using a pulse-echo technique. The TOF of the Rayleigh pulse was then found by
subtraction. The accuracy of a TOF measurement was four
parts in 104 while that of measuring the wedge separation
was 1 part in 103. Thus, the accuracy of the Rayleigh wave
velocity was approximately +--5 m - s -~ (for a velocity of
3000 m 9 s- ~).
III.
RESULTS
A. Homogeneous Samples
The results of preliminary characterization of homogeneous pearlitic and martensitic microstructure specimens are
shown in Table II. Noteworthy, in Table II, is the obserTable II. Measured Physical Properties
of Homogeneous AISI 1053 Samples
Physical Quantity
Density
Shear wave velocity
Longitudinal wave
velocity
Rockwell macrohardness
Pearlite
7796 Kg 9 m -3
3239 m 9 s-~
6011 m 9 s -~
Martensite
7779 Kg 9 m 3
3167 m 9 s5945 m 9 s -~
Rs = 91.5
Rc = 57.0
METALLURGICALTRANSACTIONS A
vation that both the longitudinal and shear wave velocities
in pearlite are greater than those of martensite, an observation that has been reported by others.~2
For isotropic linear elastic materials, the density and
wavespeeds can be used to determine Young's modulus, the
bulk and shear moduli, and Poisson's ratio.9 Furthermore, it
is possible to predict the Rayleigh wave velocity, VR, using
the relation:
VR=
( 0 . 8 7 + 1~222u) • Vs
1+
where v is Poisson's ratio and Vs, the shear wave velocity.
These deduced quantities for homogeneous material are
given in Table III.
B. Microstructure of Surface Modified Material
The surface modified samples were sectioned (perpendicular to the electron beam pass direction), polished and
etched in 2 pct nital to reveal the microstructures produced
and to facilitate microhardness measurements. Figure 3
shows, at low magnification, a view of a region about midway across the width of AISI 1053 modified material. In
Figure 4 a similar region is shown together with a microhardness depth profile. Starting at the bottom, it can be seen
that the substrate had a pearlite microstructure with a grain
size of - 2 5 / x m (see Figure 5 for a more detailed view). As
one progresses vertically toward the surface (upward in
Figures 3 and 4), the peak temperature experienced during
modification increases. This at first results in dissolution of
cementite, then a transformation to austenite, and finally
melting during the heating part of the thermal cycle. Rapid
cooling from these states then produces a wide spectrum of
microstructures that are vertically separated. Subsequent
melt passes further complicate the microstructures due to
tempering in the heat affected zone.
The first modified microstructure encountered above the
substrate is a transition zone approximately 300/zm thick
(Figure 4), composed of a mixture of partially dissolved
lamellar cementite, ferrite, and spheroidal carbides. This
transition zone corresponds to a layer that did not exceed the
eutectoid temperature during heating. Thus, only carbide
dissolution and no transformation to austenite occurred on
heating. Therefore, upon subsequent cooling, the carbon
supersaturated ferrite underwent a fine scale precipitation
resulting in a distribution of spheroidal carbides. The presence of these carbides and a higher dissolved carbon content
in the ferrite evidently increased the hardness (Figure 4),
since no martensite or bainite was evident.
Above the transition zone a layer of acicular microstructure was observed (Figure 4). This corresponds to a
region where the eutectoid temperature was exceeded and
Table III, Deduced Physical Properties
of Homogeneous AISI 1053 Samples
Physical Quantity
Young's modulus (E)
Bulk modulus (K)
Shear modulus (G)
Poisson's ratio (v)
Rayleigh wave
velocity (VR)
Pearlite
211.9 GPa
172.7 GPa
81.78 GPa
0.295
0.3002 m 9 s -j
Martensite
203.1 GPa
170.9 GPa
78.00 GPa
0.301
2939 m 9 s -~
VOLUME 18A, MARCH 1987--475
Fig. 3--A section of surface modified AISI 1053 steel. The plane of the micrograph is perpendicular to the direction of electron beam motion (2 pet
nital etch).
austenite had nucleated. The dimensions of the acicular
features increased with vertical distance from the transition
zone consistent with an increase in prior austenite grain size
due to more extensive (but still solid state) heating. This
increase in grain size probably accounts for the diminishing
hardness as one approaches the top of the layer. Detailed
examination of this region (Figure 6(b)) indicates that
the dominant microstructure is either a lower bainite, or
(possibly) a tempered martensite.
A distinctive microstructure transition occurs at the upper
end of the acicular zone as one encounters the region that
underwent melting. This zone of melting was - 3 3 0 / z m in
depth. It is shown at higher magnification in Figure 6(a).
The microstructure is acicular consisting in the main of
upper bainite, but it also contained significant volumes of
ferrite and, in places, small dark etching regions reminiscent
of very fine (unresolvable) pearlite. At lower magnification
(Figure 3) alternate light (low hardness) and dark etching
bands are observed. The dark areas correspond to dense
distributions of very fine spheroidal carbides and are the
heat affected zones of subsequent melt passes.
Microhardness measurements were made on the surface
modified AISI 1053 alloy and are summarized in Table IV.
The low hardness (Re = 22 to 29) for the melt layer is
consistent with the microstructure observed. Harder layers
would presumably have been obtained by using lower heat
flux, faster scan speed, and reducing the overlap of successive melt passes. However, even for the last pass (not subsequently tempered), the hardness of the melt layer was still
at most only Rc = 39.
Table IV.
To test the method developed on the AISI 1053 alloy,
additional Rayleigh wave measurements were made on a
similarly treated AISI 1044 alloy whose microstructure is
shown in Figure 7. The depth melted was - 4 5 0 / x m , the
acicular (but unmelted) layer was 125/zm thick, and transition zone thickness approximately 200/zm.
C. Rayleigh Wave Measurements
Figure 8 shows a plot of the measured Rayleigh wave
velocity as a function of wavelength (penetration depth) for
the surface modified layer and the unmodified pearlitic substrate as obtained from the unmodified reverse side of the
sample. It can be seen that the substrate exhibits a constant
velocity independent of wavelength with an average value of
3001 - 1 m 9s -1 in excellent agreement with the predicted
value (in Table III) from bulk wave measurements. The
absence of dispersion from a homogeneous sample indicates
that the experimental approach does not introduce any significant velocity measurement errors.
The velocity measured on the surface modified layer exhibits significant dispersion. At high frequencies (shallow
penetration depths) a limiting velocity of 2960 --- 2 m 9 sis achieved. This represents the velocity of the layer and is
- 2 0 m " s -l greater than the value predicted from bulk
wave measurements on a homogeneous martensite sample.
This difference is most probably due to the bainite microstructure of the layers since bainites are well known to have
elastic properties intermediate between those of pearlite and
martensite. ~2 We see in Figure 8 that as the wavelength
Properties of Surface Modified AISI 1053 Alloy
Region
Thickness
Microstructure
Melt layer
Acicular layer
Transition zone
Substrate
0.330 mm
0.435 mm
0.300 mm
5.2 mm
tempered upper bainite
bainite
pearlite/spheroidal carbides
pearlite
476--VOLUME 18A,MARCH1987
Hardness
R,. = 22 to 29
R,. = 27 to 30
R,. = 22 to 29
RB = 94.0
METALLURGICALTRANSACTIONSA
Surface
1
I
!
1
0.1
0.2
?
I
0.3
0.4
Melt Layer
(
0.5
E
E
"" 0.6
r
r
Acicular
Layer
9
0.7
.jl
_o
0
m
0.8
r
O
c
C~ 0.9
l
1.0
Transition
Zone
1.1
F
1.2
1.3
I
1.4
I
150
Substrate
r ] Low Values for
Light Etching Regions
I
I
200
250
300
Vickers Hardness
I
350
Fig. 4 - - A narrow region of the sample shown in Fig. 3 at a higher magnification revealing the vertical
separation of microstructure (2 pct nital etch) together with the microhardness variation with depth.
increased the velocity began to approach the value of pearlite consistent with an enhanced contribution of the substrate
to the elastic constants sampled by the wave.
The Rayleigh wave velocity for the AISI 1044 steel is
shown in Figure 9. The velocity for the substrate is different
from that of the AISI 1053 steel, but a similar dispersion is
observed on the surface modified layer. Interestingly, the
penetration depth at which dispersion becomes detectable is
less than that in the AISI 1053 steel.
IV.
DISCUSSION
Rayleigh waves propagating over the surface of electronbeam modified surfaces sample only the modified (mar-
METALLURGICAL TRANSACTIONS A
tensitic) layer at high frequencies (since the wave does not
penetrate the layer) and some combination of the layer and
substrate (pearlitic) at lower frequencies. In the high frequency region, the velocity is found to be approximately
frequency independent, consistent with wave propagation in
an approximately uniform, nondispersive medium. Then the
limiting velocity can be viewed as a characteristic of the
layer microstructure and not its depth. The velocity for
AISI 1053 surface layers is intermediate between that anticipated for martensite and pearlite, and is consistent with
the observed bainite state. There thus appears considerable
merit to the use of this asymptotic value to characterize
layer hardness.
Rayleigh waves with greater penetration depths yielded
velocities that were intermediate between those of the sub-
VOLUME 18A, MARCH 1987--477
Fig. 5 - - A typical region of the substrate microstructure revealing its
coarse pearlitic nature (2 pct nital etch).
strate and the layer. Provided the substrate is much thicker
than the layer, then when the wavelength (penetration) is
very much greater than the layer thickness, the velocity
approaches the Rayleigh velocity for pearlite because contributions from the layer become increasingly small. This
limit was not reached in these samples due to their finite
thickness, but sufficient data for a depth determination were
nevertheless obtained. If we had continued to longer wavelengths, the limiting velocity achieved would have been that
of a plate (Lamb) wave which has a different velocity to that
of the surface wave. This consideration, however, would
be of import only if one were attempting to characterize the
substrate hardness beneath an overlayer.
To determine the layer thickness, the region of interest is
the intermediate velocity region, where the velocity results
from the simultaneous sampling of several different microstructures. For example, in AISI 1053 at a frequency of
4.2 MHz (penetration depth = 705/~m), the Rayleigh velocity was 2960 m 9s-~; however, at 2.2 MHz (penetration
depth = 1357 /xm), the apparent Rayleigh velocity was
2984 m" s -j. The velocity thus depends upon both the
characteristic velocities of the layer and substrate and the
layer depth.
(a)
(b)
Fig. 6--Typical microstructure present in (a) the melted and (b) the acicular layers of AISI 1053 steel (2 pct nital etch).
Fig. 7 - - A section of surface modified AISI 1044 steel. The plane of the micrograph is perpendicular to the direction of electron beam motion (2 pct
nital etch).
478-- VOLUME 18A, MARCH 1987
METALLURGICAL TRANSACTIONS A
..-_--
,,
XL= [V--Vs]
3000
Substrate
2990
E
2980
o
>
2970 r
Surface Modified
2960 [ AIS1-1053 Steel ]
2950 I
0
I
I
200
I
I
I
I
I
I
I
400
600
800
1000
Penetration Depth IX] (/~m)
I
I
1200
I
I
1400
Fig. 8--Variation of the Rayleigh wave velocity with wavelength (penetration depth) for AISI 1053 steel.
2965
2960
lr.t--~
i
Substrate
2950
m
2940
u
>
Jr
O~
-$
2930
Substituting data from Figure 8 into this expression yields
a value for a modified layer depth of 600 to 800/zm, the
variability depending upon the precise value of wavelength
taken. Metallography for AISI 1053 indicated that the combined acicular and melt depth was --850/xm, in reasonable
agreement to that deduced bearing in mind the simplicity of
a law of mixtures model.
A more complete analysis incorporating exponential
weighting of the Rayleigh wave particle displacement might
give a more exact inversion, and thus deeper insight, into
the functional behavior of the modulus with depth. However, for the present purpose, this simple approach appears
sufficient, at least for the AISI 1053 alloy. Indeed, an even
simpler depth determination criterion such as the penetration
depth at which the velocity curve increases 2 m 9 s-1 above
the short wavelength velocity limit predicts the metallographically deduced depths of hardening for both AISI
1053 and 1044 alloys to within ---25/xm.
The Rayleigh wave technique has proved, therefore, to be
an accurate and reproducible technique for determining the
depth of hardening of surface modified steels. Since it is
nondestructive, and methodologies which would not interfere with processing are emerging, it can potentially be used
for on-line quality control purposes. If account is taken of
the effect of sample temperature fluctuations on the velocity,
it could serve the role of a process control sensor for feedback control of depth of modification and hardness, particularly if emerging noncontact measurement methodologies
are utilized.
Surface M o d i f i e d / ~ ~
m
2920
XA
LVL -- Vs_l
|-!
V.
[ AISI-1044 Steel ]
2910
I
I
l
1
I
I
I
I
200
300
400
500
600
700
800
900
Penetration Depth IX] (/~m)
Fig. 9--Variation of the Rayleigh wavespeed with wavelength (penetration depth) for AISI 1044 steel.
The inverse problem of deducing the layer properties
from velocity dispersion data is complex and is the object of
ongoing study. Here, we use a very simple analysis that
gives results of reasonable accuracy. For the case where the
wave penetrates the surface modified layer, we may, to a
first approximation, assume the apparent velocity to be determined by the layer thickness: wavelength ratio according
to a law of mixtures:
v =x
• VL + (1 --X) • Vs
where Vz = the velocity of the layer
vs = the velocity of the substrate
x = XL/A, the ratio of layer thickness to wavelength
(n.b. we define (x = 1 for A --< XL)
In the long wavelength limit, x tends to zero and the
apparent velocity is vs, while for A < xL, the velocity becomes that of the layer, VL.
TO determine the layer depth, we rearrange the equation
above:
METALLURGICALTRANSACTIONSA
SUMMARY
1. Samples of pearlitic plain carbon steels were subjected to
an electron-beam glazing process resulting in a (casehardened) microstructurally modified surface layer,
2. Hardness and ultrasonic velocity data indicated that the
modified surface layers possessed a higher hardness but
a lower velocity than the substrate materials. This decrease in velocity may be a good indication of the level
of hardness attained in the modified layer.
3. Dispersion of the Rayleigh wave velocity can be utilized
to make a nondestructive estimate of the depth of the
microstructurally modified surface layer.
ACKNOWLEDGMENTS
We are grateful to Dr. R. Schaefer for his assistance and
advice with electron beam surface modification, to C. Brady
for help with metallography, and to the Defense Advanced
Research Projects Agency for funding of this work under
DARPA Order Number 4275 (Major S. Wax, Program
Monitor).
REFERENCES
1. M. Cohen, B. H. Kear, and R. Mehrabian: Rapid Solidification Processing: Principles and Technologies H, M. Cohen, B.H. Kear, and
R. Mehrabian, eds., Claitor's Publishing Division, Baton Rouge, LA,
1980, p. 1.
VOLUME 18A, MARCH 1987--479
2. W.J. Boettinger, E S. Biancaniello, G.M. Kalonji, and J. W. Cahn:
Rapid Solidification Processing: Principles and Technologies H,
M. Cohen, B. H. Kear, and R. Mehrabian, eds., Claitor's Publishing
Division, Baton Rouge, LA, 1980, p. 50.
3. R. Mehrabian and H. N.G. Wadley: Journal of Metals, February
1985, p. 51.
4. R.B. Clough, H. N. G. Wadley, and R. Mehrabian: Lasers in Materials Processing, E.A. Metzbower, ed., ASM International, Metals
Park, OH, 1983, p. 37.
5. B.G. Lewis, D. A. Gilbert, and P. R. Strutt: Rapid Solidification Processing: Principles and Technologies H, M. Cohen, B. H. Kear, and
R. Mehrabian, eds., Claitor's Publishing Division, Baton Rouge, LA,
480--VOLUME 18A, MARCH 1987
1980, p. 221.
J.E. Jenkings: Thin Solid Films, 1981, vol. 68, p. 341.
J.E. Jenkings: Thin Solid Films, 1981, vol. 68, p. 343.
J. Masumber: Journal of Metals, 1983, p. 18.
G.J. Curtis: Ultrasonic Testing, J. Szilard, ed., Wiley, New York,
NY, 1982, p. 297.
10. W.L. Plant: Elastic Waves in the Earth, Elsevier Scientific Publ. Co.,
New York, NY, 1979, p. 102.
11. I.A. Viktorov: Rayleigh andLamb Waves, Plenum Press, New York,
NY, 1967.
12. N. Grayeli, D. B. llic, E Stanke, C. H. Chou, and T. C. Shyne: Proc.
1979 Ultrasonic Symposium, IEEE, 1979, p. 273.
6.
7.
8.
9.
METALLURGICAL TRANSACTIONS A