‘IVISSIEJ RETURNING MATERIALS:
P1ace in book drop to
LIBRARIES remove this checkout from
-_ your record. FINES will

 

 

 

be charged if book is
returned after the date
stamped below.

 

 

 

 

LASER (ZKW, CONTINUOUS WAVE CO LASER) MELIING

2

AND ALLOYING OF STEEL WITH CHROMIUM

By

Narendra B. Dahotre

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY

Department of Metallurgy, Mechanics and Materials Science

1987

ABSTRACT

LASER (2KW, CONTINUOUS WAVE CO2 LASER) MELTINC

AND ALLOYING OF STEEL WITH CHROMIUM

By

Narendra B. Dahotre

Laser surface alloying (LSA), a process of growing interest
for local surface modification, relies upon a suitable composition and
microstructure for satisfactory on-the-job performance. The LSA
technique was used to form in situ Fe-Cr-C alloys on A181 1018 steel
substrate. In this process Cr powder of different particle sizes were
mixed together to get optimum packing density and deposited onto the

substrate surface. The surface was then melted by using a 2kw CW 002

laser.

The processing conditions were related to solute (Cr) content,
microstructural refinement of the laser alloyed zones as well as the
heat affected zone (HA2). Similarly the existence and the nature of
defects such as cracks and porosity in alloyed region are reported.
Solute content was of principal concern as it is the single most
important factor affecting the properties of laser surface alloys. . The

effects of varying the laser power, beam diameter, traverse speed and

With deep respect and unflinching love

To

Latika, my mother and Bapurao, my father,

who taught me great qualities like

endurance and commitment

and who have been ideal parents to me throughout my life.

iii

ACKNOWLEDGMENTS

I would like to express my appreciation and gratitude to
Professor Kali Mukherjee, my major advisor, for his constant support,
encouragement, patience, guidance and friendship, and especially

invaluable assistance in the preperation of this thesis.

Dr. K. N. Subramanian, Dr. Dahsin Liu and Dr. W. Pratt, my
thesis committee members, for their support, suggestions and

understanding.

Dr. Rohan Abeyaratne, for willing to stay in my thesis
committee even after moving to MIT and teaching me new subjects sudh as
Continuum Mechanics and Theory of Elasticity. I deeply respect him as a

good teacher and above all a thorough gentleman.

Department of Metallurgy, Mechanics and Materials Science at
Michigan State University for providing me teaching assistantship during
my studies and Division of Engineering Research at MSU for partially

supporting my research project.

Finally, Anita, who has been an ideal wife and who had to bore
so much and my brothers Diwakar, Raj endra and Abhay for their love,
encouragement and support as well as their generous attitude towards
sharing all happy and mostly discouraging, frustrating moments during

the course of my research.

iv

LIST OF CONTENTS

Page

LIST OF TABLES ................................................. v

LIST OF FIGURES ................................................. V1

I. INTRODUCTION ................................................. 1
II. LITERATURE SURVEY

A. 2.1. An Overview of Laser ................................ 4

2.2. Solid State Lasers .................................. 4

2.3. Gas Lasers .......................................... 7

2.4. Dye Lasers .......................................... 8

2.5. Semiconductor Lasers ................................ 8

2.6. Chemical Lasers ..................................... 9

2.7. Industrial Lasers and Their Applications ............ 9

2.7.1. Ruby Laser ................................... 11

2.7.2. CO2 Laser ................................... 15

B. 2.8. Laser Process Variables ............................. 20

2.8.1. Absorptivity ................................. 20

a. Absorption of Laser Radiation by Metals ... 20

b. Absorption of Laser Radiation by

Semiconductors and Insulators ............. 26

c. Methods to Change Absorptivity ............ 30

2.8.2. Laser Beam Characteristics ................... 32

a. Laser Beam Diameter ....................... 32

b. Laser Beam Power .......................... 35

2.8.3. Traverse Speed ............................... 36

C. 2.9. Laser Surface Alloying .............................. 38

V

vl

2.9.1. Scanning Methods for Different Laser Sources . 39
2.9.2. Methods of Deposition of Alloying Elements ... 42
2.9.3. Heat-, Fluid-, and Mass-Transfer in Laser
Processing ................................... 43
2.9.4. Thermophysical Properties .................... 51
2.9.5. Phase-Diagram Consideration .................. 55
2.9.6. Surface Alloying: Ferrous-Based Metals ....... 60
2.9.7. Surface Alloying: Non-Ferrous Metals ......... 66
2.9.8. Laser Treatment of Metal Silicides and
Semiconductors ............................... 70
III. EXPERIMENTAL.PROCEDURE ..................................... 75

IV. RESULTS AND DISCUSSION

4.1. A Simplified Theoretical Model of Relationships among

Laser Surface Alloying (LSA) Process Parameters ......... 80
4.2. Topographical Features .................................. 93
4.3. Microstructure, Hardness and Diffusion .................. 109

4.4. Structural Analysis of Laser Surface Alloyed (LSA)
Region. ................................................. 121

‘V. CONCLUSIONS .................................................. 144

VI. APPENDICES

(A) Spatial Profile of a Laser Beam ......................... 146
(B) Martensitic Transformation .............................. 148
(C) Crystallography of Martensite ........................... 150

'VII. REFERENCES ................................................. 154

‘LIST OF TABLES

Page
The common laser types and methods of excitation. ........... 10
Laser fabrication processes. ................................ 13
Values of emissivity for various metals at laser
wavelengths. ................................................ 23
Binary and ternary systems that have been laser surface
alloyed. .................................................... 56
Typical results on Ag-Ni system, both thin film surface
alloying and surface melting of implants (ref. 110). ........ 58
Transformations observed in pulsed laser treatment of
Fe-B system. ................................................ 65
Trapping and solubilities in silicon (ref. 117). ............ 71
The numerical values of thermophysical constants of material
and process parameters used in this study. .................. 88
Maximum height of powder layer (hp) that can be melted all
through with specific value of specific value of specific
energy (E) - P / (V.d). ..................................... 92

Figure

10.

11.

12.

13.

14.

LIST OF FIGURES

Page

Three energy levels available to atoms or ions in a laser
material, transition from the metastable state to the
ground state results in the emission of laser light. ....... 6
Power density required for different laser processes. ...... 12
A ruby laser. (a) with helical flash tube, (b) with
parallel laser rod and flash tube. ......................... 14
Schematic diagram of a conventional discharge COzzszfie
laser. ..................................................... 14
Two methods of exciting a C02 gas laser.
(a) Discharge applied to €02:N2:He mixture directly.
(b) Only N2 excited in discharge followed by resonant

energy transfer to CO, downstream. ..................... l7
Vibrational energy level diagram for the C02 and N2
molecules. ................................................. l9
Wavelength dependence of e, n and K for Ti at 300K. ........ 22
Temperature dependence of e at 10.6pm for several metals. .. 22
Absorptivity of 304 stainless steel at 10.6pm as a
function of temperature and surface condition (ref. 13). ... 25
The reflectivity and thermal coupling coefficient of
A181 1045 steel at 10.6pm (ref. 19). ....................... 25
Absorption coefficient a, for several insulators at
wavelengths between 100 and 700nm. ......................... 28
Infrared transmission of some insulators including losses
due to reflection. ......................................... 29
Absorption versus scan speed for four different coatings
(reference 23). ............................................ 31

Beam manipulation technique used for laser.
(a) segmented mirror (ref. 23), (b) two axis vibrator system
(ref. 23), (c) torric mirrors of treating (ref. 30). ....... 34

viii

15.

l6.

l7.

18.

19.

20.

21.

22.
23.

24.

25.

26.

27.

28.
29.
30.

31.

ix

P/(Db.V)1/2 versus depth of hardening (ref. 33). ........... 37

Normalized welding speed vs. normalized laser power for
welds in stainless steel and Al made with AVCO HPL-IO
laser (ref. 31). ........................................... 37

Schematic of laser processing systems.
(a) CW C02 laser, (b) pulsed or Q-switched laser. .......... 40

Calculated melt depth vs. irradiation time based on one
dimensional computer heat-flow model from ref. 115. ........ 44

Temperature of laser heated Al as function of time and at
various depths; Tm is melting temperature (ref. 67). ....... 46

Surface alloy produced by laser alloying Au films on Ni

(reference 40). ............................................ 48
Velocity field of flows in liquid caused by inhomogeneous
heating of liquid surface by laser beam (ref. 76). ......... 50
Schematic of surface alloying. ............................. 52
Vapor pressure of elements as function of temperature

(ref. 116). ................................................ 53
Equilibrium phase diagrams, from reference 82. ............. 57

Cr concentration profile on cross-section of melt trail
(250nm deep) for Cr on Fe scanned with CW C02 laser

-1
(10 about lmW cm , tpabout 20ms) (ref. 89). ............... 61

Schematic cross section of compositionally modulated Fe-B
film before and after laser melting by 30ps laser pulse
with Gaussian intensity profile (ref. 98). ................. 64

Liquid-phase growth of epitaxial silicides using pulsed
laser.
(a)-(c) deposited metal layer.

(d)-(f) reacted silicide layer. ............................ 73
Microstructure of A181 1018 steel before laser treatment. .. 76
Packing factor vs. wtt of coarse powder in mixture. ........ 77

Schematic illustration of laser beam, substrate geomertry
and coordinate system. ..................................... 79

Schematic of the conditions in a sample before and after
laser surface alloying. .................................... 82

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

Relation between height of powder layer (hp) and
concentration of alloying element (w x 100). ............... 89

Relation between traverse speed (v) and concentration
of alloying element Cr, (w x 100). ......................... 90

Relation between laser beam diameter (d) and
concentration of alloying element Cr, (w x 100). ........... 91

Relation between specific energy (E) and concentration
(wt.%) of alloying element. ................................ 94

A surface layer of Cr powder mixture of different
particle sizes. ............................................ 96

An optical micrograph showing Cr powder layer thickness
in cross-section. .......................................... 96

Top view (KY-plane) of a laser surface alloyed AISI 1018

steel with the distance between center to center of two
successive passes equal to: (a) 2mm, (b) 1mm and

(c) 0.5mm. ................................................. 97

The appearance of surface of the sample in which laser
passes were made in orthogonal directions (with center to
center distance between successive passes equal to 0.5mm). . 98

A SEM micrograph illustrating concentration. of porosity
and cracks in the area where two successive laser passes
overlap. ................................................... 100

A SEM micrograph showing the development of cracks in the
region where two successive laser passes overlap. .......... 100

Schematic of temperature distribution around heat source.

(a) and (b) are the representation in the form of a hill,

(c) and (d) are the representation in topographic

projection. The temperature distribution in (a) and (c)

is in a plate treated with higher traverse speed where as

the temperature distribution in (b) and (d) is in a plate
treated with comparatively slower traverse speed. .......... 102

A SEM micrograph of the solidified structures of laser
surface alloyed steels with Cr. ............................ 106

Variation of dendrite arm spacing (Q) against cooling
rate (R) in commercial steels containing from 0.1 to
0.9%C (ref. 127). .......................................... 108

Cross-sectional views of laser alloyed A181 1018 steel

samples after surface grinding.

Center to center distance between successive passes is

(a) 2mm, (b) 1mm .......................................... 110
(d) 0.5mm, (d) 0.5mm in both orthogonal directions ........ 111

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

Microstructure in the region between two successive laser
passes spaced at about 2mm apart. ..........................

An optical micrograph showing the cellular dendritic
grains along traverse direction in fusion zone. ............

An optical micrograph illustrating the cellular dendritic
grains inside the laser tracks orthogonal to eachother.

Microhardness (Vickers, 900gm load) profile of Cr surface
alloyed AISI 1018 steel, across the surface as a function
of distance. Center to center distance between two
successive laser passes was: 1mm. ..........................

Microhardness (Vickers, 900gm load) profile on cross-
sectional surface of Cr surface alloyed region, as a
function of distance. Center to center distance between
two successive laser passes was: 0.5mm in both orthogonal
directions. ................................................

Electron microprobe traces taken across the surface alloy
in a sample with center to center distance between two
successive laser passes equal to: (a) 2mm, (b) 1mm,

(c) 0.5mm, and (d) 0.5mm in both orthogonal directions.

A TEM micrograph of the laser surface alloyed region
showing three phase structure. .............................

TEM micrographs of laser alloyed AISI 1018 steel showing
martensitic structure. (a) a bright field image, (b) an
SAD pattern from (a), (c) a schematic diagram of the SAD
with [001] zone axis. ......................................

TEM micrographs of laser alloyed AISI 1018 steel showing
martensitic structure. (a) a bright field image, (b) an
SAD pattern from (a), (c) a schematic diagram of the SAD
with [111] zone axis. ......................................

TEM micrographs of laser surfce alloyed AISI 1018 steel
showing the M23C. type carbides. (a) a bright field image,
(b) an SAD pattern from (a), (c) a schematic diagram of the
SAD. .......................................................
The STEM X-ray micro-chemical analysis of the M23C, type
carbide precipitate showing that precipitate is rich in
Cr content. ................................................

High resolution transmission electron micrographs of laser
alloyed region showing M23C. type carbide precipitate. .....

X-ray diffractometry data from laser surface alloyed
AISI 1018 steel. (a) as-laser-treated, ....................
(b) after lightly polishing the surface of surface in (a).

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

TEM micrograph of laser surface alloyed AISI 1018 steel

showing carbide precipitates (this treatment was carried

out without binder while depositing Cr powder on the

substrate). ................................................ 132

The STEM X-ray micro-chemical analysis of laser surface
alloyed region in AISI 1018 steel. ......................... 134

A color photograph corresponding to energy dispersive x-ray
maping, showing the distribution of Cr in matrix and carbide
precipitate. The color code bars from right to left indicate
the increasing order of concentration. ..................... 135

A color photograph corresponding to energy dispersive x-ray
maping, showing the distribution of Fe in matrix and carbide
precipitate. The color code bars from right to left indicate
the increasing order of concentration. ..................... 136

A color photograph corresponding to energy dispersive x-ray
maping, showing the distribution of Cr and Fe together in
matrix and carbide precipitate. Green color represents

element Fe and red color represents element Cr. ............ 137

A cross-sectional diagram for Fe-Cr-C system containing
0.1%C (ref. 138). .......................................... 139
Some low-order TEan cavity modes for cylindrical symmetry.

The dark areas represent areas of enhanced output, although
the distribution of intensity within these regions is not

uniform. ................................................... 146
Isothermal transformation diagram for 0.06%C-12.5%Cr (wt.%)
stainless steel (ref.l49). ................................. 148
Bain correspondence in the transformation of austenite 1,

to martensite a'. 0: Fe atom; X: positions available for

C atom. .................................................... 151

Lattice constants of tetragonal martensite and austenite
in quenched carbon steels (ref. 152). ...................... 152

Atomic arrangement in austenite 7. 0: Fe atom;
X: positions available for C atom. ......................... 152

INRODUCTION

Laser surface alloying (LSA) is a unique process designed to
produce different alloy compositions and microstructural features on the
surface of a substrate material, thereby improving its resistance to
wear, corrosion, fatigue and impact. Laser surface alloying process
consists of applying a thin coating of alloying elements to the surface
of a substrate and then scanning a high power laser beam to melt and
fuse the coating into the substrate to form an alloy. Some times the
alloying elements are delivered in the form of powders into the laser

generated melt pool and the powders are also melted by the laser.

The production of surface alloying by laser is influenced by
a number of variables and the interaction of these variables. The
absorption of laser energy, together with thermal properties of the
materials, controls the thermal history of the alloy melt such as
temperature profile in the alloyed zone, the time for which material is
molten and solidification rates. The overall composition and
microstructure of the alloyed material is determined by the degree of
mixing due to convection and diffusion and cooling rate during liquid-
to-solid and solid-to-solid state transformation. For rapidly cooled
alloys, many investigators (130, 134) have reported generally fine
microstructures, increased solid solubilities of alloy elements,
non-equilibrium crystalline and amorphous phases and achievements of

high point defect concentrations. Laser surface alloyed materials

expected to demonstrate broadly the same features because of the

inherent rapid solidification.

The phase transformation in Fe-Cr-C ternary and Fe-Cr-Mn-C
quaternary systems has been studied by many investigators (130-134, 142)
using different techniques such as rapid solidification, laser cladding,
electron beam rapid quenching and laser surface alloying. The
microstructures observed range from complex non-equilibrium to
equilibrium phases such as ferrite, carbides, martensites and some
retained austenite. The set of microstructures obtained in ferrous-
based systems during above processes is correlated to different cooling

rates, alloying elements and their compositions.

A study of LSA reveals that the influence of the magnitude of
process variables is not well known. Very little guidance is available
to choose a set of operating conditions with an end use in mind. In the
absence of a quantitative analysis, there is a need for development of
relationships of laser process parameters that may be used to
characterize the LSA process. These relationships would reduce the many
variables in LSA process into useful forms. These parameters would

facilitate experimentation and could be used for process control as

well .

The present investigation is focused on determining the
relationship between process variables and composition of laser surface
alloyed material. Interaction between process variables are discussed.

In the present study : attention is also provided to the microstructural

changes occuring in AISI 1018 steel alloyed with chromium by laser. The
microstructures of the samples were characterized by optical microscopy,

scanning electron microscopy (SEM), and transmission electron microscopy

(TEM).

II. LITERATURE SURVEY

A. 2.1. An Overview of laser

The word laser is an acronym for Light Amplification by the
Stimulated Emission of Radiation. There are many types of lasers; most
of which emit infrared or visible light. The light produced is
coherent, compared to regular light which is incoherent. Within
coherent light, the individual light rays have the same wavelength, or
color; and they are in step with each other. The light does not spread
out appreciably as it travels great distances, because the concentration
of the energy is at a very sharply defined point. Before coherent
radiation can be produced, a combination of population inversion,
stimulated emission and amplification must occur. The process needed

for completing each of these steps depends on the type of laser.

2.2. Solid State Lasers

The first type of laser which was developed is the solid state

laser. Inside of a resonator, there is a host material, which has
laser material dispersed throughout it. The ends of the resonator are
covered with mirrors. At one end, the mirror is totally reflecting or

nontransmitting. The other end has a partially transmitting silver

coating .

Optical radiation or pumping by a flash tube causes the laser

material to become excited while the host material remains transparent

to the exciting radiation and the laser light is produced. Only a
small part of the output from a flash tube is absorbed by the laser
materials. The rest is dissipated as heat (1). This absorbed energy
allows a transition to a higher energy state for some of the laser
material. When the photons decay from this high level, most will fall
to an intermediate or metastable level. The rest of the photons return
to the ground state as fluorescence. This stimulates the emission of
additional photons, with the same wavelength, energy and phase, from the
lasing medium. As more photons are stimulated, the majority of the
ions soon become displaced into the metastable state. This process is
called a population inversion. Since the atoms prefer to reside in the
ground state, they will try to return to it. Laser emission begins
when a random ion falls to the ground level and emits a photon with the
emission wavelength. When this photon strikes other metastable ions,
they emit similar photons (2) (see Figure 1). As more photons are
emitted, they begin to travel along the solid state rod. The stream of
photons becomes coherent as it travels, additional photons are also
emitted. Eventually, when the radiation has become intense enough, it
will burst from the partially transmitting end of the tube as a single

pulse of coherent light.

There are a variety of solid state lasers available. The
most common is the ruby pulsed laser (see section 2.7.1). It has a
three level transition process similar to the one just described. Four
level transition lasers are also possible. Other examples of solid
state mediums include; the neodymium doped yittrium-aluminum garnet or
NszAG, which is pulse operated; and neodymium in glass, which is a

continuous wave laser. The continuous or mean pulse power output is

 

 

HIGH —
ENERGY / 1* METASTABLE
LEVEL LEVEL -
/ '
LASER
LIGHT
\ RADIATED
oaouno
LEVEL on
sun: \‘

 

Figure 1. Three energy levels available to atoms or ions in a laser
material, transition from the metastable state to the ground

state results in the emission of laser light(ref3 2).

7

governed largely by the dissipation of the thermal energy in the
crystal. This is a function of the thermal properties of the metal,
the length and surface area of the lasing medium, and the wide range of

wavelengths (3) .

2.3. Gas Lasers

There are a large number of gases which are able to exhibit
laser action. This type of laser is called a gas laser. Some of the

more common types are 002, argon, helium-neon, and helium-cadmium. Gas

lasers may have a continuously flowing or contained lasing medium.
Again, the medium needs to be excited to a higher energy state in order
to emit photons of energy. The pumping source is normally an
electrical discharge through electrodes inside of the laser cavity.
The atoms become excited after collisions with other atoms or electrons
accelerated by the electric field. Excitation is also possible with
electron beams, chemical transitions, and transitions induced by sudden
changes in pressure through expansion of the gases (4). Gas lasers can
be used as continuous wave (CW) or pulse wave (PW) lasers; but the CW
mode is the one most often used. The actual lasing sequence will
depend on the gas being used, although it will be similar to that of the

solid state laser (The CO2 laser will be described in more detail in

section 2.7.2).

2.4. Dye Lasers

Dye lasers are possible due to the naturally efficient
fluorescence of many organic compounds. Variable outputs over a wide
range of wavelengths can be obtained at high powers. CW-argon and PW-
nitrogen lasers are often used to excite the dye lasers. Flash lamps
may also be used. Again, the stimulated emission is the necessary

effect needed to produce a coherent laser light.

2 . 5 . Semiconductor Lasers

Semiconductor materials can also exhibit laser properties when
an electrical field is applied. The laser may consist of one material
or a combination of materials, which give an output at an intermediate
wavelength. The gallium-arsenide(GaAs) laser is the most commonly

studied because its lasing action is obtained at room temperature (5).

Electrodes are attached to opposite sides of the semiconductor
material and a continuous or pulsed energy supply is used to excite it.
The emission of light will occur from the faces of plane perpendicular
to the electrodes. The output of semiconductor lasers extends over a
wide range of wavelengths, since the energy transitions are poorly
defined because electron transitions are between energy bands and not
discrete energy levels (6). Because excitation due to pulses of current
results in temperature fluctuations, a continuous operation can only be

possible if a very efficient cooling process is used to control them.

2 . 6 . Chemical Lasers

Chemical lasers depend on chemical transitions to obtain
emission. The actual transition is between gas molecules. The
chemical laser may be used as a portable high power laser since there is
no need for external energy supplies. The pumping energy is due to the
reaction of the gases. The gases may be produced by low power
dissociation from less reactive compounds with an auxiliary power supply
(7) . The actual laser transition is due to a sudden expansion of the

excited gas. This is very similar to the C02 laser when it is pulsed.

CW emission is also possible in chemical lasers. The most common

chemical reactions used are: CO, DF-CO, at 10.6pm and 1101, HP and DF

over a range of 2.7-4.0pm wavelengths. The common laser types and

methods of excitation are grouped in Table l.

2.7. Industrial lasers and their Applications

There are many different industrial applications for lasers,
many of which are still being researched. 0f the lasers currently

available, only 002, neodymium and argon lasers have sufficient output

power to do more than evaporate thin film and up to now the cost of
argon ion lasers has precluded their use for fabrication. Different
laser fabrication processes can be done by varying the output power
density. This may be achieved by changing the focus of high power
lasers. The range of power density necessary for such processes as

melting, welding, drilling, surface glazing, surface cladding and

10

Table 1. Common laser types and methods of excitation.

Normal Operating

Range (pm)

Normal Method of

Excitation

Solid state

Gas (a) monoatomic
(b) molecular
(c) metal vapor

Liquid organic dye

Semiconductor

-Chemical

0.6-3
0.2-3
2-100
0.2-0.6
0.3-1.2
0.3-31
2-100

Optical

Glow and arc discharge
Glow discharge

Glow discharge

Flash tube and laser
Electric field

Chemical reaction

11

surface alloying can be seen in Figure 2 and a listing of various laser
fabrication processes is done in Table 2. Laser application do not
have to be just in industrial areas. Other possible uses include:

biomedical, communications, printing and scientific areas of study.

2.7.1. Ruby Laser

The ruby laser is the most commonly used solid state laser.

A cylinder made of synthetic ruby, (A120, doped with 0.5%Cr), acts as

the lasing medium. It is the chromium which provides the laser action.
A flash tube is used to excite the lasing medium. It may be in the
form of either a helical tube wrapped around the crystal, or a rod

placed parallel to the ruby (see Figure 3).

The flash tube emitsan intense flash of light with a wavelength
of 5.6m. A small amount of this light (approximately 4%), will be
absorbed by the laser. This will begin the lasing sequence described

in section 2.2. The final coherent wavelength produced is 5.6m

Because the ruby laser has only a three level transition
process, there are long periods of time between population inversions.
This is an intrinsic property for every three level solid state laser.
It is possible to overcome this problem by using a solid state lasing
medium containing four levels which are capable of participating in the
laser action (8). The photons capable of producing laser light will be
produced between the second and third energy levels. Since this

population inversion will be able to occur at faster rates, there is

12

l

Breakdown in
air

l

107
l

Breakdown in
smoke

 

l

Drilling,
cutting.
and
machining
metals

Power density (W/mmzi
8m
l

 

10‘

 

Gas jet
3 metal cutting.
10 - . '
non-metal cutting,
and welding

 

 

to2

Figure 2. Power density required for different laser processes.

13

Table 2. Laser fabrication processes.

Line and Metals, plastics, glass, ceramics,

profile cutting ceramics, textiles, composites,

wood on
Welding Metals, plastics, composites Pulsed,CW
Scribing Ceramics Pulsed,CW
Drilling Metals, ceramics, plastics Pulsed
Melting Refractory materials, metals,

non-metals CW
Machining operation Metals, plastics, ceramics Pulsed,CW

Heat treatment Metals CW

14

(a)

 

 

(b)

 

 

Figure 3. A ruby laser. (a) with helical flash tube. (b) with parallel
laser rod and flash tube.

COszme "a" cozluzm.
lnld 0W“ Outlet

4

 

 

 

 

Flexible P”.
Gold-Cum [silo-s . Penis" Yr mi'
“m" "M mm" MW"
Stainless [Mel Miner
Element

Figure 4. Schematic diagram of a conventional discharge C02: N2: He

laser.

15

less wasted energy. Possible four level transition mediums include:

uranium or samarium in calcium fluoride crystals.

The ruby solid state laser is operated in a PW mode. This is
necessary because of the length of time required between pulses of
energy. The laser characteristics are also quite dependent on the rod
temperature (9) . The lasing medium generates a great amount of heat as
it is excited. This heat must be dissipated by an effective cooling
process between pulses. The PW mode is also used because it is
difficult to find a continuous source of optical excitation energy.
Because of these reasons, CW mode lasers have been developed. They are
used for laser operations which cannot be efficiently performed with a

PW mode laser .

2.7.2. 00 Laser

2

The C0, laser is the most widely used gas laser. The
conventional CW CO, laser is nothing more than a water cooled tube with

mirrors on both ends through which the laser mixture is flowed and
electrically excited (10) (see Figure 4). There are two methods for

exciting a C0, laser. The first method applies the discharge energy
directly to the continuously flowing CO,:N, :He mixture. In order to

achieve the highest output power from this method, a ratio 0.8:l:7 of
the gases must be used. The output power increases steadily with the
flow rate of the gas. This is due to the enhanced removal rate of the

dissociation products such as CO and 60,, and the direct cooling of the

16

discharge. The second method consists of separate inlets for the N2
and C02 gas. In this case, only the N2 is excited by the discharge.
The N2 molecules will later transfer the energy to the C02 as it

collides with it downstream (see Figure 5). In either case, the

discharge current will determine the rate at which the 002 molecules in

the laser tube are pumped by direct electron excitation or by collision

with excited N2 molecules.

The basic C02 laser process again involves exciting enough of

the lasing medium to a higher energy or vibrational level to cause a

population inversion. The electrical discharge excites the N2 molecules

in order to establish a population in energy level 1, (or vibrational

level v - 1). This energy is dissipated by means of a resonant energy

0
transfer to a second level 2, (vibrational 00 l), which is the upper

laser level of C02. This is accomplished through superelastic

collisions with ground state CO, molecules, causing them to be excited

into the higher level. Soon this level will have an excess of CO2
molecules compared to the lower energy levels. This is the required
population inversion. The decay from level 2 to a lower energy state 3

o

(vibrational 10 0) , results in lasing at 10.6pm. If the CO, molecules
0

fall instead to an even lower energy state 3' , (vibrational 02 0), a

9.6m laser energy is produced. After lasing, the co2 molecules must

be allowed to return to their ground state. If this is not

accomplished, a build up at level 3 or 3' will eliminate the possibility

17

 

 

 

 

 

 

 

 

 

 

 

(0}

Figure 5. Two methods of exciting a CO2 gas laser.
(a) Discharge applied to C0,:N2:Me mixture directly.

(b) Only N2 excited in discharge followed by resonant energy
transfer to CO2 downstream.

18

of continued population inversion. The helium in the discharge helps

to quiet the excitation in the 602 molecules through collisions,

allowing them to return to their ground state (see Figure 6). The
excitation and deexcitation rates are controlled by rediative or non-
radiative decay rates, which are intrinsic properties of the laser ion

and the host material.

The resulting laser radiation, which can be continuously
produced is highly monochromatic and exhibits high spatial coherence
and high temporal coherence correlation. The spatial coherence results
in high radiance from the laser sources. Once produced, the laser
energy is allowed to leave the discharge section of the tube through the
partially transmitting mirror as a coherent beam. The other end of the
discharge tube is covered by a totally reflecting mirror. This mirror

helps to direct the laser energy to the opposite end of the tube.

There are many uses and advantages to the C02 CW laser. The

light emitted is coherent, directional and has a high brightness.
Because of this, it can be used for many material fabrication processes.
The monochromaticity of the laser light allows it to be used for atomic
and molecular spectroscopy. Its temporal coherence allows
interferometric measurements and nonlinear optics to be performed. The

greatest advantage of the C0, laser is its ability to perform in either

a PW or CW mode. This increases the number and variety of processes it

can be used for.

 

 

19

 

Elem. l . .
Vfinflaul
2349 __ . [my lrsnsler m . m
2331‘ "2 iv s "M '
1 (2)
lQQu
laser 9.69
has
”as h- i”: q
' o
\ . (3 )
\ I
\ I
\ I
667 - L17 01'0: "U
l
l
I
l
l
L__ i
“2 Ground (:02 Ground ital:
sue an; m: ‘9
Figure 6. Vibrational energy level diagram for the CO2 and N2

molecules .

20

B. 2.8. Laser Process Variables

Well-thought-out statistically designed experiments are needed
to establish a relationship between process parameters in laser
processing. This is the only way to understand the interaction between
the process parameters and reproducibility (11). The effectiveness of
any laser material process depends on the manipulation of various
variables. These variables are grouped into two major catagories:
independent process variables and dependent process variables. The
major independent process variables for laser treatment include incident
laser beam power, incident laser beam diameter, absorptivity of the
coating and the substrate and traverse speed across the substrate
surface. Themophysical properties of the substrate also play an
important role. The dependent variables are considered to be depth of
hardness, geometry of the heat affected zone and microstructure and

metallurgical properties of the laser treated material.
2.8.1. Absorptivity
a. Absorption of Laser Radiation by Metals
The efficiency of laser treatment depends on the absorption of
light emery by the workpiece. Any heat transfer calculation for laser
processing is based on the absorbed emery. For an opaque solid, the

fraction of incident radiation absorbed is;

e - 1 - a, 2.8.1

21

where e is absorptivity and R0 is the reflectivity at normal

incidence. R0 and 6 can be calculated from measurements of optical
constants or the complex refractive index. For a complex refractive
index;

111 - n - iK 2.8.2

the reflectivity at normal incidence is;

2 2 2 2
R0 - [ (n - 1) + K 1 / [ (m + 1) + K 1 2.8.3

The absorptivity is then;

s - (4.n) / [ (m + 1)2 + K2 ] 2.8.4

In general, n and K for metallic materials are functions of
wavelength and temperature. The variation in n and K with wavelength
and corresponding changes in e for Ti at 3001( are shown in Figure 7. It
is apparent that both It and K are relatively slowly varying functions of
A over the range 0.4 < A < 1.0m and e is large in this range. At
longer wavelengths, n and K both increase rapidly with A and 8 decreases

to a small fraction of its value at shorter wavelengths.

Absorptivities for metals at wavelengths characteristic of Ar+,

ruby, NszAG and CO, lasers are summarized in Table 3. The temperature

dependence of e for some metals at 10.61am is shown in Figure 8. Data

122

 

 

 

 

/.0 l 20
TITANIUM
In,“
b
2
z t k
's' a. 5 ~ " ’ o
a
O
u
m
1
/ n
0.0 I

 

0’ [.0 I 0.0
WAVEL ENGTH (F In)

Figure 7. Wavelength dependence of e, n and K for Ti at 300K.

 

0-’2 '- IO. 6pm

0 ,/ MILD STEEL

AISOI'VIVITV

 

0.82 M

500 I000 1500
TEMPERATWE ('CI

 

 

 

Figure 8. Temperature dependence of c at 10.6pm for several metals.

23

Table 3. Values of absorptivity for various metals at laser wavelengths.

Metal Ar Ruby Nd-YAG CO2
(500 mm) (700 mm) (1000 mm) (10 pm)
Aluminum 0.09 0.11 0.08 0.019
Copper 0.56 0.17 0.10 0.015
Gold 0.58 0.07 -- 0.017

Iridium 0.36 0.30 0.22 ---
Iron 0.68 0.64 i -- 0.035
Lead 0.38 0.35 0.16 0.045
Molybdenum 0.48 0.48 0.40 0.027
Nickel 0.40 0.32 0.26 0.030
Niobium 0.58 0.50 0.32 0.036
Platinum 0.21 0.15 0.11 0.036

Rhenium 0.47 0.44 0.28 ---
Silver 0.05 0.04 0.04 0.014
Tantalum 0.65 0.50 0.18 0.044
Tin 0.20 0.18 0.19 0.034
Titanium 0.48 0.45 0.42 0.080
Tungsten 0.55 0.50 0.41 0.026
Zinc -- -- 0.16 0.027

24

for this relationship was calculated from the temperature dependent
absorptivity from the expression given by Duley (10). However, values of
6 given apply only for clean metal surfaces heated in vacuum. In most
practical applications of laser heating this assumption will not be
valid because of oxide formation or the presence of surface
contamination. When this is the case, values of e in the infrared can

be increased substantially.

Duley et a1. (12) have investigated the effect of oxidation
on e (10.6pm) for several metals heated in air. Data obtained by
Wieting and de Rosa (13) on the absorptivity, e, of type 304 stainless
steel at 10.6mm are shown in Figure 9. Figure 10 shows the effect of a
1 minute oxidation in air on e (at 10.6pm). It is evident that
e (10.6pm) increases substantially when surface oxide is present. Since
the growth of an oxide layer is time dependent at a particular

temperature, is is also a function of time.

The above mentioned data were obtained under low power

irradiation with a CW CO, laser. A variety of dynamic effects can change

8 at high irradiation levels. These effects have been extensively

investigated by many researchers (14-20). These experiments show that
a substantial increase in the coupling coefficient for laser radiation
at the target occurs when the intensity is sufficient to initiate a
breakdown plasma in the focal region. This plasma absorbs some
fraction of the incident laser radiation and transfers the resulting
energy to the target via hydrodynamic expansion. While the emery
coupled from the laser to the target can be increased by this mechanism,

it is deposited over a larger area than the normal beam focus. As such

25

 

 

 

 

 

 

 

 

' I I I I I ] 1 I
1‘ r- .4
1' -l
‘C ’- cl
E .l
‘ G
o
2 _
’ d
L'
) —l
E
S ‘1
2 on _ 0 use: e -l
‘ ,. A as - escslveo ‘
310 83C
use C] ‘
1 J 1 l 1 l 1 l L l
0 200 C“ m m 1”

 

TEMPERATURE 1°C!

Figure 9. Absorptivity of 304 stainless steel at 10.6mm as a function
' of temperature and surface condition (ref. 13).

PEAK lmsm lw cm'zl

 

10" lo' 10'
1’ l 1
MS! 106 Steel
assume”
. ll!) ,_ ethereal Coupling Codlidelll

E
E
2 ~16
g E
5 a
x
.8
10.4 g
I
3
u
i

 

 

 

 

mm! mm u not!

Figure 10. The reflectivity and thermal coupling coefficient of
AISI 1045 steel at 10.6pm (ref. 19).

26

it may inhibit thermal processes (drilling, material removal) at the
focal point. An example of the increase in coupling coefficient that

occurs with increasing intensity is shown in Figure 10.

While 5 may be important in the initial stages of the heating
of metallic targets with laser radiation, it is unimportant in many
practical laser heating applications. The importance of e is
diminished when material removal has proceeded to the point where a
cavity or keyhole has formed in the workpiece. In this case, the

cavity acts as a blackbody absorber with e effectively equal to unity

(21).
b. Absorption of Laser Radiation by Semiconductor and
Insulators
Absorption of light by insulating) materials is a strong
function of wavelength. In the infrared, absorption arises from

vibrational modes of the crystal lattice or in organic solid by

intermolecular vibrations. Absorption coefficients a ”102-102cm” are
typical within these bands. In the visible, absorption may occur due
to impurities (e.g. transition metal ions, crystal defect centers etc.)
or due to the tail of strong ultraviolet absorption bands. Absorption
can also occur due to discrete electronic transitions in molecular

crystals (e.g. many organic solids). Absorption coefficients are

8 s -1
typically 10 -10 on within absorption bands.

27

Figure 11 shows absorption coefficients, a, for several
refractory materials in the visible and ultraviolet. a can be related

to the transmission of a sheet of thickness t via;

-a.t

I / I, - e 2.8.5

where I0 is the incident intensity and I is the transmitted intensity.

A useful measure of the thickness required for significant attenuation

of incident radiation is given by;

L-a 2.8.6

8 -1

where L is the attnuation length. A strong absorber has a - 10 cm and
-e 1 -1

L - 10 on while a relatively weak absorber has a - 10 cm and,

-1
L - 10 cm. The relation between a and refractive index is given by,
a - (4.11.10 / A 2.8.7

where K is the imaginary term in the complex refractive index and, A is

the wavelength of incident light. In the visible region, nominally

15 -1
transparent materials will typically have R - 10 or a - 10 cm . Data
on the transmission of several insulating materials in the infrared are

shown in Figure 12.

Many semiconducting materials are opaque in the visible

28

 

  

 

 

 

I o 5
F03 04
I 0'5 —
Sic
ION [MEANT
:‘ M’0
9
's
0
\
o
10‘ ..
SIC
’03 1 1 1 1 1
700 600 500 ‘00 300' 200

(00
Alma)

Figure 11 .

Absorption coefficient a, for several insulators at
wavelengths between 100 and 700mm.

29

 

 

 

 

 

 

 

 

 

 

I00! I 1 * I
g X GOA: 0.5mm
3 sax - \ ‘ "
e,
S
E

510,
[0mm
1 1 1
I 5 IO I5 20

WAVELENGTH (pm)

Figure 12. Infrared transmission of some insulators including losses

due to reflection.

30

(A s 1 pm) and relatively transparent in the infrared. This is due to
the influence of band-to-band absorption in the visible. In the
infrared, absorption is primarily due to free carrier absorption and
transitions to impurity levels. This combination of factors means that
most semiconductor annealing is done with Nd:YAG or other short
wavelength lasers. Most materials with 81-0 bonds are relatively
transparent in the visible but absorb strongly at or near 10pm. Thus,

processing of quartz, glass and silicate minerals is best accomplished

with the 002 laser. Unlike metals where radiation is absorbed in the

skin depth at the surface, absorption in insulators and most

semiconductors occurs over the attenuation length L, which can be much

-4
larger than typical skin depths. In the infrared L z 10 cm and thus,

in many instances heating must be considered to be a volume effect.

c. Methods to Change Absorptivity

The absorptivity of material can be increased by several
methods prior to laser treatment. The workpiece surface is commonly
prepared by applying uniform absorbent coatings. Most commonly used
absorbent coatings include colloidal graphite, manganese phosphate ,zinc-
phosphate and black paint. A mixture of sodium and potassium silicate
is also known to produce very high absorptivity. Nonetheless, exact
absorptivity of any of these coatings is a matter of debate. Arata(22)
quoted 50-90% absorptivity for phosphate coatings without melting the
substrate and dependency on interaction time, whereas Trafford et a1.
(23) quoted a value around 70-80! for no-melt situations depending on

the traverse speed, shown in Figure 13. For colloidal graphite coating

31

1'00

 

l
v Monqonese

. Phosphate

"Aline Phosphate
OColloidol Grophlle
...O Poinl II

(D
O
.200
. '3le
J

'(D
O

l

q
0
l

at
O
l

Absorptivity , .%

Surface
.5). Melt

10

No
Melt 1
' 30

 

 

8'//////. /////,

08

Spéed, mm/s

Figure 13. Absorption versus scan speed for four different coatings
(reference 23).

32

and paint coating Trafford et al. reported absorptivity in excess of
80%, where as Courtney et a1. (24) reported a value around 60%. Data
illustrated in Figure 13 were obtained by calorimetric measurements
using 0.4% carbon steel substrate. Courtney et a1. determined their
absorptivity by comparing experimental isotherm with heat transfer model
(24). The major problem encountered due to above surface coating is
the contamination of the solidified region at the surface from these
elements. These problems can be avoided in some laser processes by
making surface rough. The roughness on surface can be achieved by

methods such as sand blasting, producing scratches by sand paper etc.

Anyway, it is a general consensus that these methods provide
absorptivity anywhere from 50-80%. Of course coating thickness,
coarseness and adherence to the substrate will always influence the
absorptivity. Besides, in the case of enhansing absorption by coating,
absorptivity is not the only issue, heat transfer between the coating
and the substrate is also a big concern. This problem is yet to be

addressed by any researcher.
2.8.2. Laser Beam Characteristics
a. Laser Beam Diameter
This is one of the most important variables because it
determines the power density and, thus, coverage rate. However, it is

very difficult to measure for high power laser beams. This is partly

due to the definition of what is to be measured. A Gaussian beam (see

33

Appendix A) diameter may be defined as the diameter where the power has
2
dropped to l/e or l/e of the central value. The beam diameter defined

on the basis of l/e2 of the central value contains more than 80% of the
total power whereas the power contained for l/e beam definition is
slightly over 60% (1). Many techniques have been employed to measure
the beam diameter. Single isotherm contouring techniques such as
charring paper and drilling acrylic or metal plates, suffer from the
fact that the particular isotherm they plot is both power and exposure
time dependent. One of the better methods for the measurement of beam

diameter is the photon drag detector (25-29) .

Unlike laser welding or cutting, a wider beam with uniform
intensity distribution is preferred for laser heat treatment and laser
surface alloying as opposed to a tightly focused high-peak power low

order-mode Gaussian beam (TEMOO). This is because the uniform intensity

distribution generates uniform case depth. Covering a large area using
overlapped zone due to back tempering. One has to take great care to
avoid this. Alternatively, a broader beam can cover larger areas with

relatively little risk.

There are different methods of beam manipulation that can be
used to obtain a broad beam with uniform intensity distribution. Figure
14 shows some of the methods which are used to produce suitable beam
patterns at the workpiece. The beam integrator consists of a focusing
mirror with several segments of individual mirrors focused on the same
focal plane. The greatest advantage of this system is that it can start

with a beam of any spatial distribution and convert it into a uniform

Focusing Miser
Consists ot
Individually

Pr eoliqned
Segments

Each Segment
Reflects Only
Port of Beam

    
  
  
    
   

 

incoming Laser Beam

Superimposed
Segments of
Loser Beom

Treated Area

a
..

' Foaming
Oscillating ‘ Mirror
Mirror “

l
‘1 Laser Beam
\
i 7“‘ 45-
\ Minor
§§1g‘ \
Pmnted O c1llot'1nq Minor
Areo
Works-etc
b

 

 

Figure 14. Beam manipulation technique used for laser.

(a) segmented mirror (ref. 23), (b) two axis vibrator system
(ref. 23), (c) torric mirrors of treating (ref. 30).

35

intensity beam of the required size. The rastering of a finely focused
beam to cover a larger area (Figure 14b) is another technique. Here two
mirrors are vibrated at a very high frequency to create the required
pattern. Both these techniques ultimately produce a rectangular beam
pattern with uniform intensity. However, for heat treatment of
cylidrical bodies a beam of appropriate shape and size must be moved

along a spiral path. In this method, it is very difficult to avoid
overlapping. A better alternative is to use a doughnut-shaped TEM:r beam

with toric mirrors as shown in Figure 14c (30). This eliminates

overlapping problem and produces a uniform case depth.

b. Laser Beam Power

The depth of penetration during laser treatment is directly
related to the power density of the laser beam and is a function of
incident beam power and beam diameter. Exactly at the focal point of
the focusing lens, incident beam diameter is minimum possible by that
lens and hence power density is maximum at that location for fixed power
level. As the sample surface moves away from the focal point, towards
the lens or away from it, incident beam diameter increases
proportionately. Hence energy density coupling with material have a
variable magnitude depending upon the position of the surface of

interaction.

Generally, for constant beam diameter penetration increases
with the increasing power. Locke et a1. (31) and Baardsen et a1. (32)

report that penetration increases almost linearly with incident beam

36

power. Generally Courtney and Steen (33) concluded when laser hardening
En8 steel (0.36%C, 0.8%Mn, 22%Si, 0.25%Ni, 0.5%Cu, 0.08%Cr, 0.015%P and

0.2%S) that the depth of hardening is closely correlated with the

parameter P/(Db.V)]'/2 and depth of hardness is shown in Figure 15.

2.8.3. Traverse Speed

Traverse speed determines the time. It is inversely
proportional to the depth of hardening as shown in Figure 15. Several
parametric studies of penetration welding with laser sources have been
described (31.34.35). Analysis of these data suggests that the relation
between penetration depth a, laser power P, and welding speed V, can be

expressed as;

1 2 _,
a - fi-P / .v 7 2.8.8

where ,6 and 7 are constants that depend on the laser source/focusing

system and the material welded. This relationship was varified by Locke
and Belle (31) for 304 stainless steel and by Ball and Banas (36) for
1010 rimmed steel. Plots of normalized weld speed vs. normalized laser
power such as that derived from theory developed by several researchers
(37-39) show that penetration welding utilized laser power efficiently
with overall efficiencies of 20-80% being typical. Greater efficiency
tends to be obtained at high welding speeds, since at large V, less
incident power is lost by conduction in the workpiece. Figure 16

summarizes some data obtained by Locke and Bella (31).

Figure 16.

 

secl
o
8 as

W' y-uwz
o
5

 

 

 

 

Depth (mi

Figure 15. P/(Db.V)1/2 versus depth of hardening (ref.

 

 

 

 

33) .
'00 f m I I l l T I l
O l 1.11
l- ETA'NLESS ‘
x 9.1 u
o 12 W i STEEL d
A t w ‘
V 9.6 EU AL
0 it 11' /
_ <7 10 5.851: /’
STAINLESS
STEEL
' v - vats VELOCITY K13 ‘
i- d 8 6E“ DIAMETER / .4
e - meant. DIFF. O
,. r- routes 0 4
V0 1 . tmcxuess 00
e o- 1151.1 1611'. o /
K. THERMAL WW"?
0 ‘53
1 l" ol- ....
0 fl
[ Al 1
0.1 1 1 61 1 I 1 1 1 1
1 10 100
.9.
10k

Normalized welding speed vs. normalized laser power for
welds in stainless steel and Al made with AVCO HPL-lO

laser (ref. 31).

38

C. 2.9. Laser Surface Alloying

In many military and civilian applications, surface
degradation by corrosion or erosion is a limiting factor in the
effective service life. Often, the surface of a hardware stands between
a hostile environment and the interior of the material. For metallic
materials, even pure water and oxygen are hostile environments. Such
chemically hostile environments give rise to corrosion, stress corrosion
cracking or corrosion fatigue in metallic components. Mechanically
hostile environments such as impingements of dust particles, abrasive
forces against hard surfaces such as sand, gravel or boulders give rise
to surface wear, surface cracks and surface fatigue etc. These chemical
and mechanical degradations of metallic surfaces can be minimized by
various means such as ,

* organic coatings or painting.

* designing the entire hardware with a corrosion and erosion

resistant material.

* cladding, metallizing or surface alloying.
While coating or painting can provide a protection (atleast temporarily)
against corrosion, they have very little effect, if any, in improving
mechanical properties of the surface. The second option, i.e. ,
designing the entire structure with a high strength and corrosion
resistant metal or alloy, is economically and strategically unrealistic
in most cases. Thus improvements of the surface layer properties by
cladding or metallizing is a more realistic approach. It is in this
area that laser surface alloying has the most potential. It is also a

very promising field for large scale commercial applications.

39

Laser alloying is a material processing method which utilizes
the high power density available from focused laser sources to melt
metal coatings and a portion of the underlying substrate. Since the
melting occurs in a very short time and only at the surface, the bulk of
the material remains cool, thus serving as an infinite heat sink. Large
temperature gradients exist across the boundary between the melted
surface region and the underlying solid substrate. This results in

rapid self-quenching and resolidification. Quench rates as great as

1011Ks'1 and concomitant resolidification velocities of 20 ms.1 have

already been achieved. What makes laser surface alloying both
attractive and interesting is the wide variety of chemical and
microstructural states that can be retained because of the rapid quench
from the liquid phase. These include chemical profiles where the
alloyed element is highly concentrated near the atomic surface and
decreases in concentration over shallow depths (hundreds of nanometers),
and uniform profiles where the concentration is the same throughout the
entire melted region. The types of microstructures observed include
extended solid solutions, metastable crystalline phases and metallic

glasses.

2.9.1. Scanning Methods for Different Laser Sources

Schematic diagrams of two laser processing systems are given
in Figure 17. The system shown in Figure 17b is typical for pulsed or
Q-switched lasers. In the continuous or 'CW' mode the focused laser
beam (spot size 1:!) will produce a melt trail of width w, when sample is

traversed with velocity V,under the laser beam. The interaction or

4O

‘ dell ‘

l
l
l
1 l

l, cuLaeceo View ->
i

®\X~.\\k\ W\§\\\\\\XW\\\§

 

 

 

 

 

 

 
   
   
  
 

(3)
L668 .,__fl
h—‘ f " ”””” 1
1 1 ' Lmean 111001271011
1 : not:
‘ 1
L ; LENS ENLARGEO View—1
(6) sun: new

 

Figure 17. Schematic of laser processing systems.
(a) CW CO2 laser, (b) pulsed or Q-switched laser.

41

dwell time is normally computed from the quotient d/V. These are
generally between tens of microsecond and hundreds of milliseconds. In
order to produce area coverage from individual melt trails, subsequent
trails under the focused beam are offset by a relatively slow
translation in the X- or Y-direction of the table. For a fixed beam
diameter and transverse velocity, the degree of overlapping indicated
by the hatching in the enlarged views of the right hand side of Figure
17a, will depend on the choice of table velocity. Variations on this
schematic include the use of focusing mirrors (see article 2.7.2) and

the use of inert gas shield.

In Figure 17b, a schematic representation of a typical laser
processing system for a pulsed or Q-switched source is given. These
laser sources produce a train of laser 'pulses' characterized by the
pulse duration and pulse repetition rate. For the Q-switched lasers
these pulse durations (considered as interaction time) range from tens
to hundreds of nanoseconds, while for the conventionally pulsed lasers
these time range from microseconds to tens of milliseconds. Repetition
rates range from a fraction of a hertz to tens of kilohertz. As
indicated in the enlarged view, each individual focused laser event
produces a melt spot and these melt spots must be raster scanned on the

surface if area coverage larger than the spot itself is desired.

In the case of the CW laser the area per unit time scanned is
dependent on the effective spot size, the degree of overlap, and the
relative transverse velocity. For the pulsed or Q-switched lasers it is
the repetition rate together with the spot size and overlap factor which

determine the area per unit time scanned.

42

2.9.2. Methods of Deposition of Alloying Elements

Deposition methods may be broadly classified as predeposition
(put down on the substrate in a separate step before the laser
treatment) or codeposition (injected into the melt at the time of laser
treatment). The very thin films (less than 500mm) used when surface
alloying with the shallow melting Q-switched laser pulses are invariably
put down with predeposition methods such as vacuum evaporation (40-42)
sputtering (43) or ion implantation. Thicker predeposited films are
more commonly electroplated (44,45), sprayed (46-48) or rolled (49) onto
the substrates. The ideal predeposited film would be of uniform
thickness, be low in porosity, have good adherence, possess a clean
substrate/film interface, and have an optically clean surface. Most of
LSA investigators report alloying-element losses during the laser
mixing. It is clear that many of these reported losses are not simple
evaporation but are caused by shortcoming of the deposited film.
However, to date, there has been only one (50) published investigation

directly comparing different deposition methods for LSA processing.

Codeposition is a attractive because it implies single step
processing. Particle injection directly into the melt trail produced by

CW-CO, lasers is the most popular codeposition method (51-53). Wire

injection is another technique proposed as a codeposition method (54) .
The codeposition methods offer the potential advantage of real time
control over the most important processing variable, i.e. alloying
element supply. The particle or wire feed rate may be changed during

laser melting either by design or in other processing variables.

43

Particle and wire dimensions are such that it is unlikely that either

method would prove successful with any laser source other than CW-COz.

2.9.3. Heat-, Fluid-, and Mass-Transfer in Laser Processing.

In order to understand the intermixing of alloying element
and the substrate in the transient liquid, many investigators have
suggested several models of heat and fluid flow. For the case of CW
laser sources and metals heat flow has been modelled by Mehrabian et a1.
(55-59), and others (60-63). There are few investigators (64,65) who
did model the motion of the solid/liquid interface or the rapid
resolidification phenomenon. There has been considerable modelling of
laser-induced melting, resolidification, and solute diffusion in Si
because of interest in pulsed laser annealing and semiconductor
processing (66) . However, such detailed modelling for pulsed heating

has only just recently started to appear in the case of metals (67,68) .

Figure 18 illustrates results from a one dimensional heatflow
model. The laser spot diameter is assumed to be large compared with the
melt depth and it is further assumed that there are realistic
assumptions for short laser pulses and for the faster scan speed of the
CW laser sources. In this figure the melt depth is plotted versus the
total time for three metals (Al, Fe and Ni) and several absorbed heat
fluxes. Note, this is a log-log scale and that approximate time range
capabilities for typical classes of material processing lasers are shown
on the upper time-abscissa. The maximum melt depth corresponds to an

absorbed heat flux and irradiation time such that the molten surface

44

PULSEO SOLID STATE

 

 

 

 

 

 

  
     

 

 

 

 

 

Q-SWHCHEO souo sure: cw on PULSEO coz
10" ---- ---§-.- 1 Wm. --.-W-f - --.-- 1-..- <1
Assoeeeo near FLUX-win" '9 3
> __ __ _ _ “i a 3 ‘
.03 i __ i F. 9 dry .1
i ........... Ni ' I / 1
E i 9 N // 1
1 2 l ., /
r. '0 r , .3
O- E 9 1
a I o ‘
g i : l I 4
I
1- 10 .9. ' . 1
.J E- . i : ‘
u . ‘ . ‘
I I 1 l
I ' O
1
i E/ : I . 1‘
: ' I 1
i I I : 1
10" - -..-“.1 11-111 1-”. -
' ‘ ’ ° 4 '3 -2 -1
10' lo7 106 103 l0 l0 IO lO

Figure 18. Calculated melt depth vs. irradiation time based on one
dimensional computer heat-flow model from ref. 115.

45

reaches the vapourization temperature. In this instance the heat flux

is removed .

This model suggests that the large range of melt depths
theoretically accessible as a function of irradiation time and laser
fluence. However, since pulsed or Q-switched laser sources have fixed
or rather limited ranges of pulse lengths, and similarly CW lasers have
practical lower limits on the dwell time, there exists no single laser
source that can be used to cover the entire range of melt depths. Note,
for example, that a 10ms Q-switched laser pulse will have (vapourization
limited) maximum melt depths of about 260nm and 900mm in Ni and Al,

respectively. In contrast, a CW-C02 laser with a dwell time of lms

(focused spot diameter of 0.5mm and a transverse velocity of 0.5ms-1)
will have maximum melt depths of about 50 and 200mm, respectively. The
submicrometer melt depths would prove useful if one were trying to
retain high alloying-element concentrations from a very thin deposited
film. Where deeper intermixing is required it is clear that the pulsed
or continuous lasers must be utilized. The theoretical results
presented in Figure 18 may also be used to avoid some experimental
pitfalls. It would be futile, for example, to attempt to laser surface
alloy a 2m electrodeposited Ni layer into an underlying Fe substrate
with a lOOns Q-switched laser pulse. A melt depth of 2m in Ni is not
accessible (without significant vaporization) for irradiation times of

100ns .

In Figure 19, temperature-time plots, resulting from a one
dimensional heat-flow model by Dona dalle Rose and Miotell (67) are

presented for Al melted by a 30ns Q-switched laser pulse. The

46

 

 

 

 

 

 

2500 .— 1
a
2000 -- 1
9 .
a! JSFVTI
315nm ..
t3 ‘500 665nm
tr 1365mm
w
o.
2
ul .
*- looo ‘
1hJ a b
- - - -N. -- -- .
500 I 1
o 1 J 1 1
100

 

Figure 19. Temperature of laser heated Al as function of time and at
various depths; Tm is melting temperature (ref. 67).

47

calculated temperature-time history at four subsurface depths is given:
i.e. 35, 315, 665 and 136Snm. Using the calculated temperature
transient at 35nm the general response of the metal surface upon
exposure to very short high-fluence laser pulses can be better
understood. The solid phase is heated to the melting point in a time
much smaller than the pulse length. The temperature of the surface
liquid rises even faster than the solid did because of its lower thermal
conductivity. The maximum temperature is reached in about 20ns. The
liquid cools back to the melting temperature and resolidification

occurs. The total time in the liquid state is about lOOns.

It would be expected that the solute depth profile would be:
concentration peaked at the surface, sharp fall off over a depth of
order of the diffusion length, a low (near-zero value) out to the melt
depth, and a zero value beyond the melt depth. Such a profile appears
in Figure 20 for the case of Au surface alloyed into Ni following
irradiation by a Q-switched laser source (40). Such profiles have been
successfully reproduced with straightforward models incorporating little
more than known liquid state diffusion coefficients and heat-flow
estimates of the melt time (69). Secondary effects such as a threshold
for convective mixing (70-73) and thermal-gradient drift (74,75) have

also been observed.

The case of relatively thin films and deeper melting (pulsed
or continuous laser sources) is illustrated by the Au-in-Ni depth

profiles (40) in Figure 20 for a melt time > so»; (CW-CO, laser). From

the calculations summarized in the curves in Figure 18 it can be seen

that, even for very fast scanning speeds and thus short dwell times, the

48

 

 

   
  
   

. .. ////
LASERM' ‘Aufi Ni

40

 

 

 

30 -
I C02

C>!U-C>-Nd:¥AGt
2O

 

Au CONCENTRAIION,at.-°I.

 

o 200 ' 400 600
DEPTH ,nm

Figure 20. Surface alloy produced by laser alloying Au films on Ni
(reference 40).

49

melt depth and melt time will be much greater than those for the Q-
switched lasers. The greately increased melt time allows diffusion of
the Au over much greater depths. Over the depth range analysed the Au

profile in Figure 20 is completely flat.

For the case of the multikilowatt CW-CO, laser, the scanning

speeds are relatively slow and the melting very deep with dwell times in
the millisecond to fractions of a second regime and melt depths on the
scale of millimeters. Even at these much longer melt times though the
solute diffusion length lags the melt depth, and simple diffusive mixing
cannot explain the experimentally observed homogeneous solute
distributions. Gladush et a1. (76) have extended the model of fluid
flow set up by inhomogeneous melting to include both center-to-side and
front-to-back motions. Figure 21 comprises two figures from their work
and indicates the direction of fluid motion predicted as a result of
inhomogeneous heating of a liquid at its free surface by laser
radiation. Flow patterns like these have been experimentally observed
using high speed cinematography of the surface of metals and in cross-
sectional views of transparent low-melting-point waxes (77). Chande and
Hazumder (62) have further extended the convection model to the specific

case of solute mass transfer in LSA with multikilowatt cw-co, lasers.

In their model the convection flow is responsible for dispersion
throughout the puddle volume of solute rich pockets of melt and ordinary
diffusion is resposible for providing short range homogenization outward

from these solute rich pockets.

5C)

  
   

 
  

v...\
eoeee ,.
00 O. s
O 000
000
9.0.9.

 
  

    
 
   
  
 
 

     
      

(a)

 

 

Velocity field of flows in liquid caused by inhomogeneous

Figure 21.
heating of liquid surface by laser beam (ref. 76).

51

2 . 9 . 4 . Thermophysical Properties

Among the several constraints that influence or dictate LSA
process, is the thermophysical properties of the material. In this

section the thermophysical constraints are examined. Figure 22, is the

o

schematic of surface alloying. The symbols D, T, P and R stand for the
thermal diffusivity, melting point, vapour pressure, and normal spectral
reflectance of either the substrate (B) or the film (A). The focused

laser irradiation is characterized by the irradiation time tp, the

incident fluence Io, and the optical spot size (in the case of a

Gaussian spatial distribution defined by the l/e2 points). It is
important to point out that in most instances, both for theoretical
models and in actual experiments, the effective spot size is much larger
than the melt depth. Typical effective spot-size/melt-depth ratios for
the Q-switched ruby and frequency-doubled Nd:YAG are 10000:1 and 100:1,
respectively. Even in the case of millisecond pulsed lasers and
relatively slow scanning speeds with continuous lasers the ratio is

often at least 10:1.

When considering an A-B combination for LSA some cosideration
needs to be given to the relative melting points and vapor pressures of
the elements involved. Figure 23 presents vapour-pressure data for
fourteen elements and, for comparision, the melting points of five
elements frequently used as substrates. One could easily imagine the
futility of attempting to laser surface alloy Zn (low melting point;
high vapor pressure) into Mo (high melting point). Evaporative loss (of

Mn) has been reported (46) in case of CW-CO, laser alloying of

52

\

£30. 4,
:3..- _

.wcahoaHm ooouusn we oaumaonom .NN shaman

 

maahao

 

  

flammmmawxe fizz/5545515 c at as

 

«35.02»
6F...

53

.AoHH .mouv

 

ouzusuodaou mo acuuucdm ms mucoEoHo mo ouaaaoud uone> .mu ouzmwm
. 0.. mantauazw»
250.. act-oz v0.3.8
9.. z a _z 64
a e a a e
000

000v OOQOOONOONN 898m. 003 888: 000. com

 

_ _

 

  

 

O.

0.

Sum .
O.

 

 

no: ‘ed

54

Cr-C-Mn mixtures on AISI 1018 steel and for Pb, Cr, and Sb in the case
of Q-Nd:glass laser alloying of Pb (42), Cr(7l) and Sb(78) films on A1.
Alloying element loss is not the only detrimental effect that can result
from high vapor pressure species. The formation of bubbles (originating
from either the substrate or coating) and their trapping within the
resolidified near surface region can occur. For example, Bergmann and
Mordike (79) have characterized, as a function of P content (0.02-

2.0%) , the formation of bubbles resulting from Gil-CO: laser melting of

cast Fe-alloys. Even at the lowest P content, bubbles are formed and
trapped within the resolidified surface region. This suggests that such

alloys might prove unsatisfactory as a substrate choice for LSA.

Reflectance is one of the most important parameters in
coupling laser energy into the metal (this aspect has been discussed in
detail in article 2.7.1) . The rate of heat propagation in transient-
state processes is determined by the thermal diffusivity. Knowledge of

the thermal diffusivity and the irradiation time allow one to estimate a
thermal diffusion length from the expression 1 - (2.D.tp)1/2. In laser

alloying with a configuration like that shown schematically in Figure
22, the effective thermal diffusivity will be some weighted average
dependent on the values for the film and substrate elements, and the
relative thickness of the film. Since the melt depth is some fraction
of the thermal diffusion length the thermal diffusivity can play an
important role in affecting the final melt depth and thus the resulting

alloying profile .

55

2 . 9 . S . Phase—Diagram Consideration

The themodynamic constraints which control the surface
alloying are examined qualitatively in this section. Table 4 is a
compilation from the published literature of the binary and ternary
systems that have been explored by LSA. They have been grouped into
three categories: mutual solubility in both liquid and solid state,
liquid-state solubility but limited or non-existent solid-state

solubility and systems with liquid state immiscibility.

Systems within group 1 exhibit complete solubility and would
appear to be the simplest system to work with. A binary system typical
of this group would be Pd-Ni (phase diagram in Figure 24a).
Substitutional surface alloys have been made in this system by laser

mixing of thin films of Pd on single crystal Ni with both CW-CO, and

Q-Nd:YAG lasers (80,81). Systems from this group can be useful in
experiments designed to study diffusion and kinetic phenomena free from

thermodynamic constraints .

A film-substrate combination offering one of the worst
thermodynamic obstacles is that of Ag-Ni whose equilibriml diagram is
shown in Figure 24b, and which is typical of systems in group 3. There
is extensive liquid-state immiscibility and very limited solid
solubility. Table 5 is a summary of results obtained for thin film Ag
deposits on, and Ag implantation in single crystal (llO)Ni. Only for
the very thin Ag films can substitutional solid solutions of Ag in Ni be
produced. The peak concentrations are several atomic per cent and are

very much larger than the equilibrium room temperature values. However,

£56

Table 4. Binary and ternary systems that have been laser surface alloyed

l. Mutually soluble:

Cr-Fe Au-Pd W-V V-Fe Zr-Ti Au-Ag-Pd

2. Liquid solubility-limited or non-existent solid solubility:

Cu-Ag Si-Al Cr-Cu Ni-Nb Zr-Ni Rh-Si Co-W Cr-Al
Sn-Al C-Fe Au-Ni Co-Si Au-Sn Cu-Zr Cu-Al Zn-Al
Mo-Fe Eu-Ni Nb-Ti Pd-Ti Mo-Al Zr-Al Nb-Fe Hf-Ni
Ni-Si Pt-Ti Ni-Al Ni-Be Ni-Fe Sn-Ni Pd-Si Sn-Ti
Sb-Al W-Cr W-Fe Ta-Ni Pt-Si Zr-V Co-Cu Zr-Fe
Al-Nb Ni-Cr-Cu
. Liquid and solid imiscibility:
Cd-Al Pb-Cu Au-Ru Pb-Al Pb-Fe Cu-W Cu-Mo Ag-Ni

57

.600

 

 

(a)

 

 

 

TWO MELTS

--."

 

TEMPERATURE, 0c
"r"“-iq..--.

g

 

 

“gal-ngooonoooooo

 

1 (b)

 

l700

 

I Two MELTS ‘

|

”00;:

 

 

1 (c)

 

 

Figure 24.

Equilibrium phase diagrams, from reference 82.

Table 5. Typical results on Ag-Ni system, both thin film surface

alloying and surface melting of implants (ref. 110).

cw co,
Ag on Ni (110)
1 nm ----

20 nm Sustitutional
deep, 2-phase
surface peak

200 nm 2-phase
Ag * Ni (110)
10 -2
10 cm --~-
11 -2
10 cm ----

LASER
Q-NszAG fd-Q-NszAG
Substitutional
------ solid solution
Substitutional ------

solid solution

2-phase

Remains

substitutional

Ag precipitates .........

form some Ag(<500)

remains substitutional

59

with increasing Ag thickness the point is reached where, although two
liquids are formed and some intermixing occurs, a single-phase
homogeneous liquid does not form. Further insight comes from
experiments on the Q-Nd:YAG melting of Ag-implanted Ni. In this case
the implanted alloys already possess a lateral homogeneous Ag in Ni

distribution. In fact, the implants are by and large substitutional.

From these experiments following observation can be made. If
the total film and melted substrate will not interdiffuse in the liquid
state there is no chance of producing a single-phase alloyed region.
If, i. the film is thin enough, ii. the melt temperature high enough,
and iii. the time in the melt (at high temperatures) long enough a
single-phase liquid may result. The extent and shape of the liquid-
miscibility gap will dictate whether or not a single-phase liquid region
is traversed during the temperature excursion. Even if a homogeneous
liquid is produced rejection of the solute ahead of the resolidification
interface may occur, leading to possible nucleation of a solute
concentration in the liquid are important factors in determining whether

or not second-phase nucleation and solute transport will occur.

The binary systems exhibiting the intermediate constraint
i.e. group 2, those with a single phase above the liquidus but two or
more phase regimes below it, are the systems receiving greatest
attention. It is this group where LSA offers the greatest opportunity
for development of new metal alloys. Conventional metal forming
techniques, particularly those developed for producing glassy metals,

can achieve very high cooling rates, they produce samples of severely

60

limited physical dimensions. Laser scanning may offer the best way to

produce these surfaces over a relatively large area.

2.9.6. Surface Alloying: Ferrous-Based Hetals

The majority of the investigators (87-97) have chosen to use
a low carbon steel substrate though some experiments have been performed
on technically pure Fe. Pre-irradiation coating techniques have
included rolled foil, powder, paste, electroplating, flame spraying, and

sputtering .

Malian (89) has investigated the compositional uniformity of

electroplated Cr alloyed into high-purity Fe as a function of (IV-002

laser incident power density. Figure 25 represents the Cr-concentration
depth profile. This particular melt stripe may be physically described
as semicircular and has a uniform 38% Cr over the entire 250pm melt
depth. Molian and Wood have also reported the results of
microstructural and investigations on Or surface alloys produced with

CW-CO, laser on both low carbon (0.280) steel (92,93) and high purity Fe

(90,91). The laser surface alloyed (melt stripe) microstructure
consisted of austenitic needles in a dislocated ferritic matrix. The
austenitic needles contain a high density of twins which do not extend
across the ferrite grain boundaries and follow the commonly observed

Kurdj umove - Sachs orientation relationship .

Moore et a1. (87) have described the surface alloying of Cr

into low carbon (AISI 1018) steel. Or was electroplated to thickness of

 

 

 

 

 

 

5C) T 1
:2
é 40L. . MELT DEPTH
. .___—_'-——+ —
2 . N;
2
'<130r- _.
c:
p.
2
m
2
0
U .
O ‘0 a
(b)
l l
C) 1CK) 2CK) 3CK)
DEPTH,um

Figure 25. Cr concentration profile on cross-section of melt trail.
(250pm deep) for Cr on Fe scanned with CH CO, laser

-1
“(lo about law cm , tpabout 20ms) (ref. 89).

62

15pm and melting was accomplished with CW-C02 laser over a wide range of

both tranverse speeds and incident power densities. Since one of the
main interests in LSA of Fe based substrates is the production of a
stainless steel surface equivalent, these investigators have sought to
characterize the electrochemical behavior of laser surface alloyed
steels. Laser surface alloyed Cr surface alloys subjected to anodic

polarization scans in deaerated 0.1M N280, passivate. With increasing

Cr content both the critical current density for passivation and the
current density within the passive region decrease. In this medium the
1018 steel (without the surface alloy) undergoes metal dissolution at

all anodic potentials.

Another area of ferrous-based surface alloying is the making
of amorphous surface layers. Continuous ribbons of metallic glass
produced via commercial roller methods are contained in size. In LSA
process the intimate contact between the surface melt and the
selfquenching bulk lying underneath the melt will result in a highly
efficient heat-transfer coefficient. Ultrahigh quenching rates are
accessible through the use of nanosecond and picosecond laser pulse
melting. For laser alloying methods to be practical the metallic glass
must not only be produced as a result of an individual laser event, but
the glassy region thus produced must remain amorphous throughout

subsequent heat cycling resulting from the need for overlap.

The potential for forming unique amorphous surface alloys is
illustrated in the recent work of Lin and Spaepen (98). They have laser

alloyed modulated Fe-Fe,B films with 30ps laser pulses from mode locked

63

Nd:YAG laser, see Figure 26. It is interesting to note that the melt
time for 30ps laser pulse is estimated, and has been experimentally

measured to be of the order of lns. For liquid state diffusivities of

,4 2 ,1
10 cm s that means a diffusion length of about 3nm, assuming only
simple diffusion as responsible for intermixing of layers. With this

in mind the predeposited modulated Fe/FeSB/Fe/FesB ------- film had

thickness of 1.6-3.5nm. The total thickness of the modulated structure
was about 100nm. The films were deposited on 1pm thick Al, which in

turn was on a bulk Cu substrate, thus efficient quenching was insured.

The quench rates are of the order of 1012K;1 or greater. Fe-B
alloys between 5 and 24 at.-%B were all found to be amorphous following
laser alloying. The more conventional splat quenching techniques can
make amorphous Fe-B only if they contain 12-28 at.-%B, thus the much
higher quenching rate of the picosecond laser melting has allowed

amorphization at much lower B concentration.

Borodina et a1. (99) have studied the crystalline -v amorphous

transition in binary Fe-B system. In their work both CW--CO2 and TEA-CO2
lasers produced insufficient selfquenching to make amorphous Faun”

from crystalline starting material. In work of Inal et al. irradiation

of thin Penn” glassy foils with Q-ruby laser irradiation resulted in

crystallization effects varying from nucleation of aFe clusters to

continuous crystallization of aFe-Fezn. In examining the various

irradiation times and transformations given in Table 6, it is somewhat

surprising that a binary system like Fe-B, which can readily be made

s ,1 '
amorphous by the 'bulk' splat-quenching methods (cooling rates<10 Ks ),

61$

30 picosecond laser pulse

Ll

original modulated film

 

 

 

 

 

 

 

homogeneous. Fe-B glass Fe B
5. 2m- .. =;-:;=:::=:-:-,,’/ 3
\\\\\\\\\\\\\\\\ '1'2'"..-,'-: 3; , .. III/I /
2 “mm“ 2. .. .. ,,,,,,; '\ Fe
' SI Al film
i \\\\\\\\\\\ \
\\ ‘\Cu.substrate
\ \ \ \ \ \ \ \ \ \ \\
[A A:
" 100 pm I'

Figure 26. Schematic cross section of compositionally modulated Fe-B
film before and after laser melting by 30ps laser pulse with
Gaussian intensity profile (ref. 98).

(55

Table 6. Transformations observed in pulsed laser treatment of Fe-B

system.
Laser Ref. Irradiation Transformation
time observed
CW-CO, 99 200 ms Crystalline ~ crystalline
TEA-CO2 99 < 1 ps crystalline * crystalline
Q-ruby 143 25 ns Amorphous 4 crystalline

mi-Nd:YAC 98 30 ps Crystalline 4 amorphous

66

could not be made amorphous (< lps TEA-C02) or be retained as amorphous

(25ns Q-ruby) for these very short laser pulse irradiation times. One

would certainly expect the cooling rates for these pulse lengths to be

a -1
in excess of 10 Ks . The explanation may lie in the fact that in the
crystalline sample experiments the melt is intimately associated with

crystalline seeds.

In order to make LSA an important means of fabricating glassy
surfaces, the substrate may be selected based on structural or
electrical properties, while the glassy surface is prepared to meet
perhaps corrosion or oxidation resistance. For example, Bergmann and
Mordike (47,48) point out that epitaxial growth is common in many Fe-
based alloys and is to be avoided when considering the composition of
the crystallizing surface melt and substrate structure. If because of
the presence of the substrate, nucleation is not necessary for the melt,
then amorphization will be that much more difficult to achieve even with
the larger heat-transfer coefficient, thermal gradients, and quenching
rates. Thus the B in a Fe-Cr-C-B surface melt serves to constrict the
1-loop and the time needed for heterogeneous nucleation becomes the

relevant period within amorphization must compete.

2.9.7. Surface Alloying: lon-Ferrous—lletal

Most of the LSA studies has been carried out on the
technically important elements (or alloys of) Al, Cu, Ni, and Ti. Among
these there are relative advantages and disadvantages from the

standpoint of ease of laser processing, thermophysical constants and

67

post-treatment analysis. For example, high purity Cu, and Al have very
high thermal conductivities and thermal diffusivities. In combination
with high reflectances this makes them more difficult to laser surface
melt reproducibly, particularly in the case of overlapping melt puddles
or stripes. 0n the other hand, these larger conductivities and
diffusivities will result in greater temperature gradients across the
melt/solid interface, more efficient heat quenching, and higher regrowth
velocities. The lower conductivities and diffusivities of Ni and Ti
make them much easier to laser process. For example, they can easily be

surface melted with CW-CO, (10.6pm) laser radiation, while Cu and Al can

not. Titanium tends to absorb oxygen and expected to be the most

sensitive to laser processing in air.

If sufficiently high cooling rates are achieved then both
epitaxial regrowth and crystalline nucleation from the melt may be
supressed and a metallic glass may result. Continuous, pulsed (ms),
Q-switched (ns), and mode-locked (ps) lasers have been used with the
intent of retaining the amorphous state on cooling of melt puddles.
Both multiphase crystalline substrates and multilayered thin films have

been used to provide melts with requisite compositions.

Lin and co-workers (100,101) have reported using the
picosecond output of a mode-locked Nd:YAG laser to make amorphous thin
film multilayer in the Ni-Nb (100) and Flo-Ni, Ho-Co, and Nb-Co binary
systems (101). The initial multilayer structures are similar to that in
the Fe-B case shown in Figure 26 where the individual film thickness are
less than the mixing length (diffusion distance for metal in liquid

metal for t - lns). Total deposited film thickness were about 100nm

68

which corresponds to the estimated melt depth for the pulse length,

laser fluences and sample geometry used.

In the Ni-Nb system (100) the glass forming range was
determined to cover the range 23-82 at.-%Ni with supersaturated fee and
bcc solid solutions outside of this glass forming region. The authors
point out that the glass formation range in the Ni-Nb system when
produced by conventional splat-cooling method (102) is only 40-70
at.-%Ni. In the case of the Mo-Ni system (101) reported studies were
performed at three compositions 30, 50, and 60 at.-%Ni. In all
instances amorphization took place as a result of laser alloying and
melt quenching of the homogeneous liquid. Note that Lin et al. (101)
found in the case of the 30 at.-%Ni sample same bcc microcrystallites at
the edge of the melt spot. They point out that the laser fluence is
spatially Gaussian, as shown in Figure 26, and may produce only a
partially melted region around the perimeter of the spot, thus providing
a site for epitaxial growth to occur. In closing this discussion on
laser-melt quenching of metallic glasses it is pointed out that care
must be taken fully to characterize the near-surface region before
reaching definitive conclusions on the transformations that have taken
place. One example from the literature will suffice to illustrate this
point. It was reported that amorphous Al laser surface melted with a
lSns Q-ruby laser pulse. It was also concluded (104, 105) that this
amorphous Al layer significantly improved electrochemical corrosion
behavior of laser treated Al surface. But shortly after that several
groups (106, 107) including the original author (108) reported further
studies on Q-ruby laser melting of Al, which by utilization of other

characterization techniques showed unambiguously that a 150nm thick

69

amorphous Al layer could not be forming on pure A1, and that it was

likely the observed amorphous phase was impurity stabilized.

Surface alloy formed in the non-ferrous metals have also
shown to be corrosion resistant. An example is recent work on the Pd-Ti
system. Commercially available bulk alloys of Ti containing fractions
of a per cent Pd are widely used by chemical processing industries. Now
protection against the reducing-acid environments is required only at
the metal/corrosive interface, rather than throughout bulk, hence
surface alloying appears reasonable approach, particularly to save cost
in reduced consumption of Pd. Surface alloying of Ti with Pd has been
accomplished by thermal solid-state diffusion of thin Pd-electroplates
(109,110) and by laser alloying (111). Laser surface alloying of Pd in
T1 offers several advantages over thermal diffusion. The thermal
diffusion (solid state) suffers from concentration irregularities (109)
associated with block or grain boundaries. The rapid diffusion and
intermixing associated with liquid-state laser melting has been shown,
generally to produce homogeneous surface alloy regions. The long times
at elevated temperatures associated with furnace diffusion require an
inert atmosphere to protect Ti from embrittlement by oxygen, nitrogen or
hydrogen. The time elevated temperatures in Q-switched laser processing
can be accomplished in air. A furnace treatment takes the bulk as well
as the surface up to elevated temperature. Thus substrate alloys that
might be affected by such a thermal treatment would be excluded as
possible choices. Short time laser processing will not affect the bulk
material and thereby there is no restriction in the choice of

substrates.

70

2.9.8. Laser Treatment of Metal Silicides and Semiconductors

The laser irradiation of semiconductors has been the subject
of much research over the past ten years (66). It is now possible using
a variety of pulsed lasers and irradiation conditions, to control the
velocity of the solidification front over the range of 1-20ms-1.
Several novel aspects of crystal growth have been observed in this very
high velocity regime. Impurities in the melt can be engulfed by the
solidifying interface and trapped in the solid at concentrations far in

excess of equilibrium solid solubilities. Table 7 gives the maximum

concentration of some impurities that can be traped at velocities in the

range 3-5ms-1. The maximum concentrations of impurities are determined
by either thermodynamic considerations or interfacial breakdown due to
the phenomenon of constitutional supercooling. This phenomenon of
kinetic trapping implies that interfacial segregation coefficients are

also enhanced as shown in Table 7.

These are few significant differences between pulsed melting
of Si and of metal ingeneral. Metals are solidified at high
velocities, large densities of defects result. However, below the
amorphization velocity in Si, perfect epitaxial regrowth results with no
extended defects. This difference between the covalent semiconductor
and metals originates in different yield strengths of the materials.
Enhancements of solid solubilities of impurities at the solidifying
interface are much greater in Si or GaAs than in most metals. The
semiconductors start from a much lower impurity content and enhancements
should be proportionately greater. Fundamentally, the most important

difference between elemental semiconductors Si and Ge and elemental

71

Table 7. Trapping and solubilities in silicon (ref. 117).

Dopant

Dopant concentration ratio;

solid/liquid

-1
At very slow At 4 ms

solidification solidification

Maximum concentration of

dopant in solid

-1
At very slow At 4 ms

solidification solidification

As
Sb
Ga
In

Bi

K0 x'
0.80 1.0
0.35 1.0
0.30 1.0
0.023 0.7
0.008 0.2
0.0004 0.15
0.0007 0.4

caucucm's) Cmuoucm's)
1500 6000
70 2000
45 450
0 8 150

72

metals is that it has been possible to quench amorphous Si and Ge from

the melt .

There have been two sorts of experiments on semiconductors:
i) the alloying of semiconductor film with substrates, such as Ge on Si,
and ii) the alloying of metal films with Si and GaAs. The reason for
alloying Ge on Si, has been to form an epitaxial-semiconductor
overlayer. The alloying of metal films on semiconductors is considered
below. The formation of metal contacts is a crucial part of
semiconductor technology. The contacts can either be rectifying or
non-rectifying. Such contacts ultimately control the function of a
solid-state device and represent one of the most important areas of thin

film metallurgy .

The silicide contacts are widely used in integrated circuit

technology. Thin metal films such as Pt are deposited on Si and reacted

at temperatures in the range 500-10000C to form uniform silicide layer.
The silicides are good metal conductors and make ideal contact
materials. Poate et al. (112) have conducted extensive research on
deposited films of Pt, Ni or Pd in the liquid phase to form silicide
layers. It was observed that the silicide layers were not single phase
but instead consisted of cells with silicide rich well (see Figure 27a-
c). The cells resulted from the breakdown of solidifying interface.
The interface moves through a region of rapidly changing melt
composition. Segregation at the interface results in instabilities and
the lateral rejection of solute to form the cell walls, this is

constitutional supercooling.

73

SER
POLYCRYSTALUNE Ht M550 u a: sutuocs

   

(e) (b) (1:)

mean:
REACTEO sass, ”-550 uses

(NON-WV)
V7

§é

  
 

  

 

 

(d)

Figure 27.

Liquid-phase growth of epitaxial silicides using pulsed
laser.

(a)-(c) deposited metal layer
(d)-(f) reacted silicide layer

74

One way of overcoming this cell formation is shown
schematically in Figure 27d-f. The film is reacted in the solid phase
at low temperatures to form a single phase layer. The silicide layer is
then melted just to the interface. Solidification ensues through a
region of equilibrium phase concentration and single phase material
results. The rationale behind this process is to grow epitaxial

silicides. Silicides such as N1812 and CoSi2 are cubic in structure and

have lattice parameters within 1% of Si. It is possible to grow single

crystal CoSi2 and N1812 in the solid phase using ultra high vacuum

deposition and reaction techniques. These sohpisticated reaction
techniques are necessary to prevent impurities accumulating at grain
boundaries or interfaces and thus inhibiting epitaxial growth. Epitaxy
can be achieved by use of the pulse melting techniques (113) without the
need for ultra high vacuum techniques. The liquid phase regrowth
process is much less sensitive to the single crystal metal layers. They

can be removed by a subsequent anneal in a furnace.

III. EXPERIMENTAL PROCEDURE

AISI 1018 steel (0.15-0.20 wt.%C, 0.6-0.9 wt.%Mn, 0.04 wt.%P,
and 0.05 wt. 158) samples in the shape of a rectangular slab (2.5cm X

1.25cm X l.25cm)were used for this study. The specimens were annealed

0
at 450 C for about 1 hour. The specimens were mechanically polished and

then etched in reagent nital (3m1 HNO, and 100ml ethyl alcohol) to

reveal the prior microstructure (Figure 28). Chromium powder
(electrolytic grade with 99.8% purity) was used for alloying purpose.
In order to maximize the powder packing density, powder of average
particle sizes 40pm and 60m in the ratio of 3:1 by weight were mixed

together. The packing factor obtained this way was approximately 0.30
(see Figure 29). A slurry of chromium powder in a suitable organic
binder was made and uniform layer was applied onto the specimen surface.
The organic binder was allowed to set before laser irradiation. During
laser irradiation, organic binder evaporates and powder particles are

free to infiltrate into the base metal.

The specimens were mounted on an electrically driven X-Y

table and irradiated with a continuous wave C0, laser beam of 10.6pm

wavelength at a power level of 1500 to 2000 watts (schematic is shown in
Figure 17a). Argon gas was directed on the sample coaxial with laser
beam to avoid surface contamination. The beam was focused at the
surface of the specimen to obtain spot size of 0.2m. The specimen
travel speed employed through the study was 8 cm/min. Several laser

tracks in one direction parallel to each other were made. The

75

76

 

Figure 28. Microstructure of AISI 1018 steel before laser treatment.

77

.0u50xfis ta u003oa emueoo we was

.e> Houumu mewxoem

.mm madman

E: E menace mace so a

n

_

 

 

an

d

nu

an

U3

oo—

 

 

 

ou.o

54.0

mu.o

on.o

¢~.o

. HOME-I SNIXDVJ

78

intertrack spacing was maintained as 2mm, 1mm, and 0.5mm respectively.
Also one group of samples was treated with two sets of parellel tracks
at right angle to each other. For this case, in both directions the

center to center distance between successive passes was 0.5mm.

For optical microscope observation of laser melted region, an

etching reagent consisting of 60ml H20, 40ml HC1, and 10ml HNO, was

used. The microhardness measurements were done on a KENTRON Micro
Tester with a 900gm load. Concentration of elements in the alloyed
surfaces were determined by using a ARL (Applied Research Laboratories,
Inc.) electron microprobe analyser and a Vacuum Generator's 118-501 field
emission scanning transmission electron microscope (FE-STEM) . The FE-
STEM was operated at 100 kV. Microstructural investigations of laser
alloyed surfaces were carried out by optical, Hitachi S415A scanning
electron microscope (SEM) and Hitachi H-800 transmission electron

microscope (TEM). The TEM was operated at 200 kV.

The samples were prepared for TEM and FE-STEM observations
from laser alloyed region by cutting out 0.3m wafers parallel to the
treated surface (XY-plane, see Figure 30) on Buehler diamond cutting
wheel machine. After mechanically polishing to a thickness of 0.15mm,
3mm discs were punched from the wafers. Finally the 3mm discs were

thinned to final observation in a solution containing 950ml acetic acid

0
and 50ml perchloric acid with an applied voltage of 52.5 volts at 25 C

using a twin jet polisher (Tenupol-Z, Struers, Inc.).

79

LASER
BEAM

 

 

Figure 30. Schematic illustration of laser beam, sub

strate geomertry
and coordinate system.

IV. RESULTS AND DISCUSSION

4.1. A Simplified Theoretical Model of Relationships Among Laser

Surface Alloying (ISA) Process Parameters .

Laser Surface Alloying (LSA) being a complex event,
considerable time and experimental efforts would be required to analyse
it thoroughly. The principal variables in LSA are the optical
properties of substrate (absorptivity), the laser beam power, beam
diameter, traverse speed and the height of powder layer of alloying
element. A mathematical model offers a powerful alternative to a purely
experimental approach in analysing this complex problem. A model may be
used to check physical concept of the process, estimate unmeasurable

parameters and to study the effects of varying parameters.

There are several methods used to deposit or deliver alloying
elements onto the surface of material, such as: vacuum vapor deposition,
electroplating, powder coating with an organic binder, thin foil or wire
feeding, ion implantation, chemical vapor deposition, powder injection
in the metal pool etc. For a given material-laser system, the method of
deposition or delivery of alloying elements can produce very different
and products. Laser energy coupling is strongly modified by the
reflectance of the surface. Thus, a layer of powder particles, for
example, behaves quite differently compared with a smooth vapor
deposited or plated layer of the same element. For large surfaces,
powder application is most economical and amenable to automatic process

control. If powder is applied, then powder packing density can be

80

81

maximized by different particle size blending. Powder packing density
and layer thickness can be used to estimate melt depth for a desired
alloy composition. For example: if the desired alloy composition at the
end of LSA process is S-(w x 100)% P, where S is the base metal, P is
the alloying element and w is the weight fraction of alloying element P,
then the conditions before and after LSA are shown schematically in

Figure 31. In this figure, hs is the melt depth in base metal, hp is
the height of powder layer, ps is the density of pure bulk element S, pp
is the density of powder layer with a as a packing factor and a.pp < 1.

Also in the same figure, d is the width of laser track, 1 is the length
of material to be melted. The traverse speed is v, and time required to
make a laser track of length l with this speed is t. Hence the mass of

powder and base material melted during LSA is expressed as following:

mass of powder melted (mp) - a.pp.hp.d.v.t 4.1.1

mass of base metal melted (m8) - ps.h8.d.v.t 4.1.2
s i n c e ,

w - (a.pp.hp.d.v.t) / [(a.pp.hp.d.v.t) + (ps.hs.d.v.t)] 4.1.3
1. . - . .h . .h + .h 4.1.4

e v (a pp p) / [(0 pp p) (ps s)1

the expression for the melt depth (ha) in base metal is given by;

hs - hp.[(l/w) - 1].(pp / p8).a 4.1.5

It is clear from above expression that the melt depth of base metal and

ultimately the final composition of the alloyed region is dependent on

82

.wc«%oaae commune momma
menus use ouomon madame e a“ ecoauancoo 0;» mo causaonom .Hm shaman

 

 

1_))
\\

 

 

m «83 x 5 - a

 

 

 

 

 

 

 

 

 

 

 

83

the height of powder layer along with the packing factor of powder
particles. These derivations were carried out with the assumption
that, there is no interaction of heat from previous pass/passes and the
last one (this is valid for a single track experiment) and there is no
loss of powder due to evaporation. If the assumptions like these are

not made, the problem becomes very complex.

With the above assumptions, this model is extended to
establish the relationship among several other LSA process parameters

for a Gaussian laser beam operating in TEM“ mode. In order to reduce

the complexity of the problem, some additional assumptions were made:
1) all the properties of the liquid and solid metal are constant,
independent of temperature, 2) there are no convective heat losses to
surrounding, 3) the temperature developed in the process is S boiling

point temperature of the elements.

As the substrate is covered with a powder layer, the laser
beam interacts with the powder layer first, and then heat is transferred
through this molten layer to the substrate. Conceptually, the whole LSA
process can be divided into two successive steps: powder melting and
substrate melting. The laser energy (Q) absorbed by the system can be

separated into two parts correspondingly;

Q .. QP + Qs 4.1.5
where, Qp - energy required for melting powder

Qs - energy required for melting base metal

84

The individual expressions for terms involved in above equation are

given by;

Q - P' .t.A

T
_ 0 p P
Qp mp.fir CP.dT + mp.AHf

f

T
_ 5 8
Q3 ms'fTr CP.dT + ms.AH

where, P'is the laser beam power.

A is the absorptivity of powder layer.

4.1.6

4.1.7

4.1.8

C5, C3 are the specific heat at constant pressure, of powder and

base metal respectively.

AME, AM :are the enthalpy (per unit mass) of fusion of the

coating metal (powder) and the base metal respectively.

TI. is the temperature of the system before LSA process

To is the boiling point temperature of the coating metal

(powder) .

T3 is the boiling point temperature of the base metal.

After using equations 4.1.6, 4.1.7 and 4.1.8 in equation 4.1.5, it

transforms to;

r T
, _ 0 p s p s s
P t.A [ mp. fir CP.dT + mp.AHf] + [ ms. fir CP.dT + ms.AHf ]

4.1.9

851

Again the values of Inp and ms obtained in equations 4.1.1 and 4.1.2

respectively are substituted in equation 4.1.9 which further gives;

T

. _ 0 p p
P t.A (a.pp.hp.d.v.t) . [ fTr CP.dT + AHf]
T
6 s s
+ (ps.hs.d.v.t) . [ [Tr CP.dT + AHf 1 4.1.10

The equation 4.1.10 is rearranged to convenient form as follows:

T
P'/ (v.d) - (a.pp.hp / A) . [ fTo c§.dr + AH? 1
r

T

+ (ps.hs / A) . [ ff: G;.dT + an; 1 4.1.11

At this juncture, a very important LSA process parameter called specific

energy(E) is defined. It is very useful in characterizing a LSA process.
E - P'/ (v.d) 4.1.12

Further the equation 4.1.11 can be expressed either in terms of hp or hs

by using the relationship obtained in equation 4.1.4. Hence;

T
[P'/ (v.d)].(l/hp) - (a.pp/A).[ ff' cg.dr + An; 1
r

T
6 s s _
+ (a.pp/A).[ If: CP.dT + AHf ].(1/w 1) 4.1.13

for a given powder layer and base metal system, 0:, pp, A, CB 0;, AH?

and AH: are constants and therefore, equation 4.1.13 can be written as:

[P'/ (v.d)] . (1/hp)- M + N.[(l/w) - 1] 4.1.14

where M and N are constants and values of them are given as;

T
_ 0 p p
M (a.pp / A) . [ IT: CP.dT + AHf 1 4.1.15
Tfls s
N - (a.pp / A) . 1 fTr CP.dT + AHf 1 4.1.16

Similarly, the heat balance equation in terms of hs can be represented

as following;

[P'/ (v.d)] . (l/hs) - M'.[w / (1 - w)] + N' 4.1.17

where M' and N' are constants for given powder layer and base metal

system and they are given as;

r
M' - (ps / A) . 1 15’ c§.dr + An; 1 4.1.18
1'

4.1.19

T
N' - (pa/A) . [pr c;.dr+An‘f‘1
r

87

Equations 4.1.14 and 4.1.17 are very useful equations which
give access to LSA process parameters. One can choose the combination
of values of beam power, beam diameter, traverse speed, height of powder
layer or the depth of melt in base metal to be alloyed according to the
required final concentration of alloying element or vice-verse. This
selection of values of parameters can be done prior to LSA process. By
using equations 4.1.14 and 4.1.17 and the values of constants quoted in
Table 8, several useful relationships are plotted for chromium-steel
system. Figure 32 shows the variation in concentration of
alloying element Cr (wt.%) against the variation in height of powder

layer (hp) at constant specific energy (E). It is obvious from the

Figure 32, that for constant specific energy and predetermined melt
depth, the concentration (wt.%) of alloying element increases as
increase in the height of powder layer. This is because more and more

amount of alloying element is added to the same volume of base metal.

The variation of concentration of alloying element Cr (Wt.%)
against the variation in traverse speed (v) with fixed laser power

(P') and fixed height of powder layer (hp) for constant beam diameter

at the surface is represented in Figure 33. When the traverse speed
decreases, the duration of interaction of laser beam with material
increases. Due to longer duration of interaction, the base metal is
melted to larger volume and at the same time the fixed volume of powder
is being mixed with larger and larger volume of base metal thereby
decreasing the concentration of alloying element. Similarly, the
representation of variation of concentration of alloying element Cr
(WtJ) against the variation in laser beam diameter at the surface with

fixed laser beam power (P') and fixed traverse speed (v) for constant

138

Table 8. The numerical values of thermophysical constants of material

and process parameters used in this study (ref. 118).

Fe Cr
Specific heat at constant pressure (Cp) 10.0 9.4
,1 -1
cal.deg .mole
Boiling point temperature (Tb) 3070 2690
o
C
Enthalpy of fusion (AHf) 3.3 5.0
Kcal / mole
Bulk density (p) 7.87 7.19
3
gm / cm
Absorptivity (A) 0.5

Packing factor (a) 0.3

(39

Wt.-% ALLOYING ELEMENT c:,(v x 100)

 

 

 

 

25 ‘10 5 2.5
_ l I I 7
20— --‘0.05
E - P7(V.d), (Watt-Sec.)/mm2
for (V.d) " 1 ’1
16’— [’1
'---— for (v.d) - 2 7’
. 0’
‘3'
I,’ /
H g, I
0 I, bW,’
a 12 "' [I ’l/
” “5’ 1"?"
I; /// // Q ://$1$—001 E
fin. ” I], / Q. ’ ’/ c.
/ / / ./ o
\ ,.__ / I I /.’00 a.
1—4 8 ’l I, // //’ X. 0 ,c
v ,” // ’x/ // Q, ‘1‘)
I, // ’/ /// g " X50°
I,’ /// l/ I/ Y: x150
. ' ,’ I I/ s ’ __
lb- ////:/:/ 200° Q.25
// /’/:’/ E ’
////;’ "10.5
/¢$’
,2’9’ —1.0
0 ” J I I l
0 10 20 30 4O
[(I/V) - 1]

Figure 32. Relation between height of powder layer (hp) and

concentration of alloying element (w x 100).

9C)

Wt.-% ALLOYING ELEMENT Cr, (0 x 100)
30 10 5.0 3.0

 

 

 

 

 

 

 

 

200 10
F 0
'2
160 -
5. 15 .5.
° i
a:
U)
‘0 a:
" .:
(0 o
o t:
P .-
>< a
’_\ 25 :>
~31 80:
,.
v
50
1100
200
o ,
0 8 16 24 32

[(I/V) - 1]

Figure 33. Relation between traverse speed (v) and concentration of
alloying element Cr, (w x 100).

91

Wt.—% ALLOYING ELEMENITk;(w x 100)

 

 

 

 

 

30 10 10 30
8° 1 ' l l l
— fi=2kw
v= 25 inches/Min. “'0-15
60—-
..
I .—
E
240— “—0.25 .
x 2
A E
n In
a — .
20 ~—-- -—0 5
—- —l.0
' l l 1 —’2°5
0..
0 8 16 24 32

[um - 11,

Figure 34. Relation between laser beam diameter (d) and concentration
of alloying element Cr, (w x 100).

92

Table 9. Maximum height of powder layer (hp) that can be melted all

through with specific value of specific energy (E) - P/(v.d)

E h
P
2

[ (Kwatt.sec) / mm ] mm
2.00 0.70
1.75 0.62
1.50 0.50
1.25 0.44
1.00 0.35
0.75 0.26
0.50 0.18

93

height of the powder layer (hp) is done in Figure 34. As the laser beam

diameter increases, the value of specific energy also decreases which in
turn melts lesser volume of the base metal while same volume of powder
is mixed with smaller and smaller volume of base metal. This ultimately

results in increasing the concentration of the alloying element.

Finally the relationship between specific energy (E) and
concentration (weight%) of alloying element is represented in Figure 35.
Again for fixed height of powder layer, the concentration of alloying
element decreases against the increase in specific energy. Higher the
specific energy deeper is the penetration of heat into base metal at
fixed height of powder layer and this leads to distribution of same

volume of powder into larger volume of base metal.

In order to make LSA process effective, the first requirement
is to melt at least complete layer of powder through its height. Hence
there is maximum height of the powder layer that can be melted all
through with specific value of specific energy (E) - P'/ v.d which has
to be calculated before deciding upon the values of other LSA process
parameters. Table 9 quotes the values of specific energy and associated
maximum height of powder layer that can be melted all through. These

calculations are based on the equation 4.1.7.

4 . 2 . Topographical Features

An optical micrograph of AISI 1018 steel, after annealing and

before laser treatment, is seen in Figure 28. The microstructure

95

consists of an uniform grains of pearlite and ferrite. A surface layer
of chromium powder ( mixture of different particle sizes as mentioned
earlier) is shown in Figure 36. From micrographs such as this, an
analysis of packing density of powder and pores can be made. Note that
pores facilitate absorption of the laser beam and consequently, alloying
does not require additional application of an energy absorption coating.
The chromium powder layer thickness is illustrated in a SEM micrograph
in Figure 37. This thickness was about 0.3mm. SEM micrographs in
Figures 38a, b, and c show evidence for the rippled topograph in a laser
track which has been the object of considerable attention by other
investigators (38, 119). As these intertrack distances are smaller,
more surface area is melted. It is evident from Figure 38c that for
0.5mm center to center distance between two successive tracks, the
melted regions nearly overlap and a complete surface melting occurs.
The appearence of the surface of the sample, in which the laser tracks
were made in two perpendicular directions (0.5mm center to center
distance) is seen in Figure 39. The ripple marks correspond to the
second set of tracks. In this treatment also, the successive passes are
overlapped in both directions ensuring complete and thorough melting of

the sample .

From the figures 38c and 39, it is quite clear that porosity
is developed in the solidified region. Presence of cracks in this
solidified region is also a predominant feature as seen from these
micrographs. Porosity and cracks are especially concentrated in the
area where two successive laser passes overlap (see Figure 40). This
area basically is the heat affected zone (HA2) next to fusion line for

both successive passes. This narrow HA2 is the most vulnerable part of

96

 

Figure 36. A surface layer of Cr powder mixture of different particle
sizes.

. 7",”
171(11):; _‘"3l JU*‘3HH

 

Figure 37. An optical micrograph showing Cr powder layer thickness in
cross-section.

 

Figure 38. Top view (KY-plane) of a laser surface alloyed AISI 1018
steel with the distance between center to center of two
successive passes equal to: (a) 2mm. (b) 1mm and (c) 0.5mm.

98

,:9

ENSURE! Z‘SK 18*5NH t

 

Figure 39. The appearance of surface of the sample in which laser
passes were made in orthogonal directions (with center to
center distance between successive passes equal to 0.5mm).

99

area surrounding fusion zone (120). The micro-distribution of alloy
elements take part in this region (121) . It has also been concluded
earlier (122) that there is a very high vacancy concentration and
vacancy clusters in this region. High concentration of vacancies and
subsequent coalescing could provide sites for initiation of cracks in

HA2.

The SEM micrograph in Figures 40 and 41 show the cracks in
the region where two cosecutive laser passes overlap. The cracks are
parallel to the direction of motion of workpiece. The cracks of similar

nature were also observed in W2 steel treated by using 2KW, CW C02 laser

(120). The formation of cracks and the nature of cracks depend
basically on the distribution of heat in the material during laser
treatment. Rosenthal (123) has developed the mathematical theory of
heat distribution during a interaction of a moving point heat source
with material (similarly the laser heat source used in present study can

be considered as a point heat source).

During the development of mathematical theory for the
distribution of heat, Rosenthal assumed following things: (a) the
physical coefficients of the material are constants, (b) heat exchanges
(losses) through the surface to surrounding atmosphere are neglected in
regard to the heat flow in the piece itself. Under these assumptions,
the differential equation of heat at a point is written as follows

(123):

2 2 2 2 2 2
a r/ax + a r/ay + a T/az - 21.(aT/ac) 4.2.1

10¢)

\

f
8683 :31; 1855111"

 

Figure 40. A SEM micrograph illustrating concenntration of porosity and
cracks in the area where two successive laser passes overlap

 

Figure 41. A SEM micrograph showing the development of cracks in the
region where two successive laser passes overlap.

101

where x, y, and z are the coordinates of the point; T is the temperature

of the point, 1/2A is the themal diffusivity and t is the time.

When this kind of treatment (interaction of point source with
material) is performed over a sufficient length, a state is soon created
in the piece, which is called quasi-stationary. To define this state an
observer is assumed to be stationed at the source of heat during
treatment. Then if the quasi-stationary state is reached in the piece,
the observer will notice no change in the temperature distribution
around the heat source. To look at it in another way, let the
temperature distribution around the heat source be represented by a
hill, the isotherms being the level lines of the hill (lines in Figure
42a, b), then in a quasi-stationary state of heat transfer, the hill
will move as a rigid body on the surface of the plate without undergoing
any modification either in size or shape. As a consequence, the traces
left on the surface by different isotherms will become straight lines

parallel to the direction of the welding.

To account for the establishment of the quasi-stationary
state in equation 4.2.1, the origin of coordinates is transfered from
the plate to the source (Figure 30). To this end, if y-axis lies in the
direction of piece motion, y must be replaced by a new coordinateJ,

such that ,

i - y-v.t 4.2.2

where v is the traverse speed. Now equation 4.2.1 becomes thus;

1(12

 

Figure 42.

‘

Schematic of temperature distribution around heat source.
(a) and (b) are the representation in the form of a hill,
(c) and (d) are the representation in topographic
projection. The temperature distribution shown in (a) and
(c) is in a plate treated with higher traverse speed where
as the temperature distribution in (b) and (d) is in a plate
treated with comparatively slower traverse speed.

103

2 2 2 2 2 2
a T/ax + a 1/81 + a T/az - -21.v.(aT/aw) + 21.(aT/ac) 4.2.3

But according to earlier assumption, the temperature undergoes no change
with time, in new system of coordinates, attached to the heat source,
thus the derivative with respect to time vanishes and equation 4.2.3

becomes:

2 2 2 2 2 2
a T/ax + a 1/31 + a T/az - -2A.v.(aT/aW) 4.2.4

This is the differential equation of the quasi-stationary state of point

heat source treatment.

The solution of the equation 4.2.4 for the temperature
distribution in a semi-infinite plate is obtained after taking into
account the following boundary conditions.

(1) since the source is a point source, the heat flux through the

surface of the cylinder 211.r.g (g-thickness of the plate and r-radius of

a circle drawn around the heat source and r-(x2+y2+zz)1/2) must tend to
the value of the total heat Q, delivered to the plate as the radius r,
tends to zero,

-2.lI.r.g.K.(aT/8r) -+ Q, as r —-) 0
where K is the thermal conductivity.

(2) there can be no change in temperature with respect to the
thickness,

dT/dz - 0

(3) temperature of the plate remains unchanged at a very great

distance from the source,

104

T‘—+To, as r—‘FQ
where, To is the initial temperature.

with these conditions, the analytical solution given by Rosenthal (123)
to to such a steady state, three dimensional heat flow equation is as

follows:

T - To - [Qp/(ZII.K)].exp(-A.v.y).[K°.(A.v.r)/g] 4.2.5

where K0 is the Bessel function of second and zero order. The value of

this function can be found for each value of A.v.r in the tables of

Bessel functions.

If the physical state of the plate would undergo no change in
the immediate vicinity of the heat source, then equation 4.2.5 would
give T-o, for r-O, which is impossible. A closer approximation would
then be necessary taking into account the finite size of the heat
source. This indicates that solution applies only outside of the fused

zone i.e. , only below T-Tf ('1‘f is the temperature of fusion) and that

the more distant the considered point from the heat source, the better

is the approximation .

The graphical representation for the temperature distribution
around heat source, based on equation 4.2.5 is done in Figure 42. A
family of isotherms are developed around heat source. The rise of
temperature in front of the heat source is steeper than the fall of
temperature behind source. The points on the workpiece passing through

maximum temperature at the same instant are located on a line (locus),

105

curve 11, (Figure 42c and d), which is curved backward. This is due to
finite speed of heat flow in metals, which delays the occurence of the
maximum temperature in volume elements parallel to the direction of
workpiece motion. The more distant the volume elements from the heat
source the greater the delay. As a consequence steep thermal gradient
is established and hence thermal stresses are developed in the material
which are expected to give rise to cracks parallel to the direction of
motion of the workpiece. For the same material with identical geometry
the speed of scanning affects mostly the shape of isotherms. The higher
the speed, the more elongated are the isotherms and the more pronounced

the lag of the curve (Figure 42c and (1).

Because of the experimental difficulties involved in
measuring the temperature and cooling rates in laser surface alloying
process (due to high energy density and short interaction time) several
heat flow models based on theoretical concepts were addressed (57, 61,
63, 115). An alternative technique for determination of the
solidification rates, is to measure the dimensions of microstructural
features such as grain size, dendrite or cell spacing etc. Figure 43 is
the representative SEM micrograph of a surface treated with two
orthogonal sets of laser passes. The microstructure is cellular
dendritic in nature with two distinct regions of different dendritic
cell sizes. The region on left is the one melted twice by overlapping
orthogonal passes, where as the region on right is a portion between two
parallel passes of laser. The relatively larger size of the dendrite
cell indicates the cooling rate in the region on left in Figure 43 was

slower as compared to the cooling rate in the region on right.

1(J6

25K 68*3Hn

 

Figure 43. A SEM micrograph of the solidified structures of laser
surface alloyed steels with Cr.

107

Several investigators (124-127) have shown the dependency of
dendrite cell size on the cooling rate. The general empirical
relationship between dendrite cell size and cooling rate for a wide

variety of alloys is expressed as follows (128):

e - b.R'“ 4.2.6

where, CD is the dendrite cell size, R is the solidification rate, b is

the constant whose value depends on the diffusion coefficient of solute
in liquid and also on initial composition of the alloy and n is another
constant whose value is generally very close to 1/2. Hence it is clear
from equation 4.2.6 and from the experimental data on dendrite cell size
in ferrous alloys (Figure 44) obtained by Suzuki et al (127) that
dendrite cell size depends very much on cooling rate. Thus the relative
larger size of the dendrite cell indicates that the cooling rate in the
region on left in Figure 43 was slower as compared to the cooling rate

in the region on right.

The other important factor controlling dendrite cell size is
the alloy element content. Assuming equilibrium at the two liquid-solid
interfaces, a diffusion couple is established in the liquid between the

two dendrite cell arms such that (128):
n
J ~ ~pL.DL.(c:-CL)/¢ . 4.2.7

where, j is the concentration gradient, DLis the diffusion coefficient,

Ctand C2 are the weight fractions of solute in the liquid in equilibrium

1()8

 

 

 

 

 

 

 

 

 

 

 

a 1000 __

2 : e

8 - ° .

g 500 2‘

g .- 0. ° 3. le

6 200 ' L _ .‘ :t

d . Q4 ’ I».

Q " \I‘

m

z 100 )- , .

5 : . e K
1 C

E 50 v 0 \

E h-

S 30

g 20 11 A 111 1 1 111

 

 

 

 

 

 

 

 

 

 

 

1 2 3 5 10 2030 50 SW 200 500 KID
AVERAGE COOLING RATE ('C/min)

Figure 44. Variation of dendrite arm spacing (0) against cooling rate
(R) in commercial steels containing from 0.1 to 0.9% carbon.

(ref. 149).

109

with arms of radius m and n respectively. Hence for a selected
solidification rate, the cell size will decrease as the concentration
gradient increases (124). As the region on left in Figure 43, has been
melted twice, it results in redistribution and dilution of the solute.
This variation on dilution produces localized variation in the
solidification temperature (129) in accordance with the phase

equilibrium.

4.3. Microstructure, Hardness and Diffusion

Typical cross-sectional views of laser alloyed specimens
after surface grinding (xz plane in Figure 30) are shown in Figures 45a,
b, c,and d. As the depth to which these specimens were ground was not
the same, it is not possible to compare the melt depths in them. But
these cross-sectional views represent the microstructural changes which
occured during laser processing. The reagent nital was used to reveal
these microstructures. It is clear from these microstructures that
nital was unable to etch the chromium rich laser alloyed region. The
region surrounding the fusion zone, in all cases, is essentially
martensitic in nature. A comparison of structures in Figure 45a to d
also shows the spread of this newly formed martensite region around the
fusion zone deep into base metal. The band width of this solid state
transformation around fusion zone in above four laser treatments (Figure
45a-d) measured to be 1.3mm, 1.5mm, 2.9mm, and 3.8mm respectively.
These measurements give idea about heat penetration into the substrate
as well as the level of temperature developed at different locations for

different kind of transformations. Ashby and Esterling (130) have

Figure 45.

110

 

:4, . :j '3‘. . __ . .1
"RN .z. "6"» ,-
4? v1.9 C ’. LI" ,.', ’1‘". 16:1;7’ h". 1.1., a g
:‘1 ' - ’4; ’ A. spa-$473.14.; .1 :500pm
' 1 ' in: v. ‘1...) -- (4 _ , 1.
a " ’ “ '1 ‘43.; 3;’~'.\FV'~"1"ELZV::‘€.J_ 1 l

 

Cross-sectional views of laser alloyed AISI 1018 steel
samples after surface grinding.
Center to center distance between successive passes is

(8) 2mm, (b) 1mm

111

 

Figure 45. Cross-sectional views of laser alloyed AISI 1018 steel
samples after surface grinding.

Center to center distance between successive passes is
(c) 0.5mm. (d) 0.5mm in both orthogonal directions

‘1)

C

.1-

.5:

'1
c- t;

.3151:

'1}

LL.
0

,.
JAYS "
-r
C
a T'
.73)

I"

f ,

f

H1

112

derived the equation for temperature developed at a point below the

center of the beam. In this case a laser beam of power P' and radius rb

is tracked in Y direction with velocity v across the surface of a solid

(Figure 30). The temperature below the center of the beam is given by;

1(z t) - To + {(A.P'/v) / 2.H.A.[t.(t+to)]1/2).exp-[(z+zo)2/(2t/A)]

4.3.1

where A is the thermal conductivity, To the substrate temperature, A the
absorptivity of the surface and t is the time. The constant to measures

the time for heat to diffuse over a distance equal to the beam radius,

to - r: / (2/1) 4.3.2

and the length 20 measures the distance over which heat can diffuse

during the beam interaction time. The above equation is established for
only one laser pass (similar to the conditions followed in experiment in
Figure 45a). Calculations based upon the above equations show the

temperature developed within this solid state transformation band is

o o
somwhere between 960 C to 1150 C.

Calculations based upon equation 4.3.1 can not be carried out
for conditions followed in experiments in Figures 45b to d. The small
intertrack spacing as well as overlapped tracks result in an interaction
of heat energy from successive passes. This interaction produces deeper

heat penetration and higher degree of temperature than would be

°IJL

"..71‘

s.-‘ OJ

113

predicted by equation 4.3.1. Hence, measured values of width of solid
state transformation zones increase from treatment in Figure 45b to
Figure 45d (1.3mm, 1.5mm,- 2.9mm, and 3.8mm respectively, as mentioned
earlier). As the intertrack spacing decreases, the gap between
martensite region around neighboring fusion zones is bridged (Figure
45b) . In case of very close laser tracks, in addition to martensite
region, the fusion zones are also bridged (Figure 45c). Finally, for
the two sets of overlapping orthogonal passes, particularly, produce two
distinct uniform layers of alloyed zones followed by a solid state
transformed martensite region (Figure 45d). The martensite structure in
all of the above cases except that shown in Figure 45d is columnar in
nature. The disappearance of columnar structure in Figure 45d is
attributed to tempering of fusion zone caused by successive overlapping

passes in orthogonal directions.

The heterogeneity of microstructure in the form of bands of
martensite observed in heat affected zone (HA2) in some specimens (see
Figures 45b, c, and d) may result from local variations in composition.
Microstructural heterogeneities of various forms have previously been
reported in laser treated materials (131-133). Such features have been
interpreted in terms of factors such as compositional differences due to
incomplete solution of phases, dependence of crystal growth direction on

thermal gradients, and differences in cooling rate.

The views of an alloyed surface after surface grinding are
shown in Figures 46, 47, and 48 (in XY plane in Figure 30). The sample
in Figure 46 was etched with nital. It delineates the microstructure in

the region between two successive passes at about 2mm apart. The

114

 

Figure 46. Microstructure in the region between two successive laser
passes spaced at about 2mm apart.

 

Figure 47. An optical micrograph showing the cellular dendritic grains
along traverse direction in fusion zone.

115

 

Figure 48. An optical micrograph illustrating the cellular dendritic
grains inside the laser tracks orthogonal to eachother.

116

martensitic structure next to each fusion zone separated by the region
of pearlite in ferrite matrix is indicative of substantial variation of
cooling rate in this region. As mentioned earlier reagent nital does
not reveal the metallurgical microstructure of alloyed region, hence

samples were etched with a reagent consisting of 60ml H20, 40ml HCl, and
10ml HNOs. The optical micrographs in Figures 47 and 48 are the

microstructures in laser alloyed region on surface revealed by the above
mentioned reagent. Figure 47 exhibits cellular structure developed
along traverse direction (along the trajectories of heat flow). The
formation of a given morphology of solidification structure is

determined by Gs/R ratio (G8 - thermal gradient present in the solid at

the solid-liquid interface, R - growth velocity of the solid-liquid
interface) during solidification. Cellular growth structures. form in

preference to dendritic structures if Gs/R ratio is high. It appears

that the growth rate is comparatively higher in this region. Figure 48
also illustrates the cellular grain structure inside the laser tracks
orthogonal to each other. The region where the two tracks are
overlapped does not reveal the microstructure clearly. Possibly this

region may have developed a highly corrosion resistant phase.

The structural heterogeneities observed in both solidified
and solid state transformed (HA2) regions are the reasons of very large
fluctuations in the values of microhardness across the surface alloy
(Figure 49). This variation is due to the presence of very fine duplex
microstructure consisting of metastable phases and carbide precipitates
in alloyed region. It is also evident from Figure 49 that the highest

value of hardness in each fused region drops successively from the final

117

.EEH "no: common momma

e>aemeoo=e can decays; commando mousse cu mouceo .ooceueuv

mo couuocsu e as commune ecu mmomoe .Hooue maoa HmH< vohoaae
commune no mo adduced Apnea ewoom .nuexoa>v emocpuesouo«: .oe shaman

€55 woztma
nxm “We. a!” AHN 0;. O

 

_ _ _ _ _ _ _ _ .

 

 

 

mean
“we:

 

l

“we.

Amzesm ex sz uoammpm mos use zo

 

.o

Aver

Auop

O 0
D N
N N
zww/Ba' (a 006) suaxom ‘ssauouvu

 

 

Oo;w

118

track to the first track. The drop in hardness value is due to the
tempering action by the heat supplied from following laser tracks. The
extent of tempering and hence the level of the highest value of hardness
in a track depends upon the number of following passes. The uniform
hardness (DPN 230) was obtained on the surface alloyed by two sets of
orthogonal laser passes. Figure 50 is the hardness traverse on cross-
sectional surface (xz plane in Figure 30). The hardness value drops
across the laser melted/HA2 interface and finally reaches to the value

of hardness (DPN 135) of the base metal.

Finally the relative diffusion profile of laser alloying
elements as a function of intertrack distance was studied. Figure 51
shows a set of electron microprobe traces across the surface alloy (in
XY plane in Figure 30) . This figure shows that average chromium content
in laser melted region is uniform. As the intertrack spacing decreases,
more amount of chromium diffuses into the HAZ surrounding fusion zone
(Figures 51a and b). This solid state diffusion, deep into heat
affected region, is due to the interaction of heat flow along the
surface from successive laser passes. When laser tracks are laid almost
overlapping each other in one direction, the chromium content in the
surface is about 40 wt.‘ (Figure 51c). But for the same powder
thickness, laser power and traverse speed, if sample is treated for two
sets of orthogonal laser tracks, almost overlapping in each direction
then chromium content in surface layer drops by about 700 (to about 12
th) as seen in Figure 51d. Due to two sets of multiple laser passes,
the base metal is melted to greater depth and also heat diffuses to deep
below surface thereby distributing same amount of chromium over larger

volume of base metal.

119

.edouuoeuun
accomocuuo soon a“ aEm.o News evened momma o>uem0003e o3u
awesome eoceumun ueucoo cu mouceo .ooceumun mo coauocam
e we .coamem nehoaae commune no mo commune denounces
-mmouo do cauwoua Apnea awooo .euexo«>v meecnuenouowx

¢=Evzeamo
m; . 0., m6

 

_. i a._ _

 

£23m Ne ad
2383 smegma E 84.38 may. no

 

 

 

.on euawum

0.0 H
anew

O

1 N
3

S

I.., omp MS
.A.

l 3
H
I1okpu
S
1 \67
m

x

Ru

.ll

w

Cram
3

f'l

i
.1. ., .~
' 1..--

1...... .

. 1..-...1- ' I ' I
IC—LAsea Tame—>1 .. k-f—taseaxaacKT—f‘l

 

  
 
 
  

 

 

 

1 so . - . .
2
3
a: 40
I
u
I! 20
a;
3’
o
0 200 400 1800 2000 2200 2200
DISTANCE 1:10" (cm)
100 . .
- -if;t-.=..(b)
. ,.._.. .. . . .... ';
; i I . 2 .. - ' - - 1J18
-.Lu ..-... .--...g -.l— — o. .. . a. : ; . . . :
°°.':;.:T.|—+ ... 1'".
. F—LASER TRACK—4 l“—.us“,"“*9"—‘I:'
: 60 , , I. . ‘ ' -' ' . _ .. , .-.
2 . .- .
g o
c 40
I
0
'fi 20
~
3
o
0 200 400 800 1000 1200 1400 1600
DISTANCE x 10“(cu)
(C)
1 X
g i
9 2
5 3
* 5
s

 

L—tml veeea——:—aeeee.uaea—-1 L—eeeee seaea

eons-ea I 10" (so) masses 1 es“ (en

Figure 51. Electron microprobe traces taken across the surface alloy in
a sample with center to center distance between two
successive laser passes equal to: (a) 2mm, (b) 1mm,

(c) 0.5mm, and (d) 0.5mm in both orthogonal directions.

121

4.4. Structural Analysis of Laser Surface Alloyed (LSA) Region

The phase transformation in Fe-Cr-C ternary and Fe-Cr-Mn-C
quaternary systems has been studied by many investigators (134-139) by
using different techniques such as rapid solidification process (RSP) ,
electron beam rapid quenching (EBRQ) process and laser cladding (LC)
process (140, 141). They observed basic microstructures consisting of
ferrite, martensite, fine dispersion of different carbides and some
retained austenite. The transformation product obtained by this
technique depends mainly upon heat and mass transfer, fluid flow and

cooling rates.

Transmission electron microscopic investigation of laser
surface alloyed region indicated that the microstructure consists of
three microsructurally distinct regions,elongated, blocky and matrix,
as shown in Figure 52. TEM micrographs shown in Figure 53a illustrate
that the elongated microstructure predominantly consists of lath
martensite. The substructure of the lath martensite consists of a high
density of dislocations, some of which form cell structure, a
substructure similar to that of cold worked iron. The laths within a
packet of martensite were separated by low angle boundaries. Careful
analysis of lath martensite structure indicated the presence of at least
two types of lath morphologies: convergent laths, and regular parallel
laths. Regular parallel laths were most frequently observed with
individual laths having planar boundaries (Figure 53a). A convergent
lath is occasionally observed within each martensite packet (Figure
54a). Messler et al. (136) reported that at slow quench rates, the

martensite laths are less parallel, thicker and more lenticular but that

122

 

Figure 52. A TEM micrograph of the laser surface alloyed region showing
three phase structure.

123

 

(b)

. ._§_. (c)
110 no

020 011 020
A18=1.41, ZONE AXIS [001]

 

Figure 53. TEM micrographs of laser alloyed AISI 1018 steel showing
martensitic structure. (a) a bright field image, (b) an SAD
pattern from (a), (c) a schematic diagram of the SAD with
[001] zone axis.

124

 

101° 60
110 A ,/
8
011
(b) . (c)
017 ,

_._ no
10]

 

A/3=1, ZONE AXIs['1'11]

Figure 54. TEM micrographs of laser alloyed AISI 1018 steel showing
martensitic structure. (a) a bright field image, (b) an SAD
pattern from (a), (c) a schematic diagram of the SAD with
[111] zone axis.

125

the degree of alignment perfection among parallel laths increased as the
quench rate increased. Kelly and Nutting (142) attributed the change in

martensite morphology and substructure to stacking fault energy and Ms
temperature. Addition of chromium to a steel lowers the Ms temperature

and also lowers the stacking fault energy thereby producing martensite

with dislocation substructure instead of twinned one.

The crystal structure of lath martensite is usually bcc in
low-carbon steels (143). This is probably because of segregation of
carbon atoms to dislocations during quenching. In the present LSA
processed samples, on the basis of diffraction analysis, as seen in
Figures 53b and c (with [001] zone axis) and Figures 54b and c (with
[111] zone axis, the martensite structure was bcc (see Appendix C). The

lattice parameter of bee-martensite calculated from this analysis was

0

found to be 3.15 A. Iron-base martensites that contain only
substitutional alloying elements (Fe-Ni, Fe-Ni-Cr for example) usually
have bcc crystal structures because the substitutional alloying elements

are distributed at random on the lattice sites (143).

In addition to this martensite structure, LSA region also
contains block type precipitates as shown in Figure 55a. These
precipitates are generally irregular in shape and their size varied from
0.31am to 1.6m. Such large differences appeared due to differences in
alloy composition and cooling rates. These precipitates have been

identified as 142,0, type with fcc crystal structure as suggested in the

literature (144). This analysis was based on selected area diffraction

(SAD) pattern obtained for one of the precipitates and is shown in

126

(a)

 

 

2120 412 o/
08.4 ’ .
A 45
(b) ,
840 3&0 (C)

111—20 084 41—20
A/B=1.414, ZONE AXIS [00?]

 

Figure 55. TEM micrographs of laser surfce alloyed AISI 1018 steel
showing the M230. type carbides. (a) a bright field image,

(b) an SAD pattern from (a), (c) a schematic diagram of the
SAD.

127

Figures 55b and c. The corresponding STEM X-ray micro-chemical analysis
(Figure 56) from these precipitates is essentially rich in Cr content
with partial substitution of Cr by other metals like Fe. Therefore, the

carbide precipitates had the (Cr,Fe)23C5 type structure (14h). These

precipitates contain a high density defects. These defects appear to be
closely spaced stacking faults. The high resolution bright field TEM of
the carbide precipitates shows a regular array of parallel lattice
fringes (Figure 57). The lattice spacing between fringes is

approximately 35m. Similar kind of fringe pattern was observed in M703
(hexagonal structure) (140) and MGC (fcc) (141) carbide precipitates

obtained in laser clad Fe-Cr-Mn-C alloy.

The parallel investigations of laser surface alloyed region
were carried out by other methods to support above observations.
Figure 58 shows the results from X-ray diffractometry analysis of a
region of AISI 1018 steel alloyed with Cr by overlapping laser tracks.
The reflections obtained from the as-laser-treated surface (Figure 58a)

were indexed in terms of oxides [(Cr,Fe)2O,, Fe30‘], carbides
[(Cr,Fe),,C,, (Cr,Fe).,C,] and Cr-ferrite. The formation of oxides may

well be due to particularly strong melt-enviroment interactions. The
carbides were formed by the interactions of carbon from base metal as
well as to some extent from the organic binder used to hold Cr-powder on
the top of the base metal. At this juncture it is very difficult to
quantify the volume fraction of carbon contributed by binder to this
process. But it is cetainly true that carbides do form in absence of
binder due to carbon from base metal. This is clear from the TEM

micrograph (Figure 59) obtained from a laser surface alloyed AISI 1018

1213

 

3000. - canton murmurs Cr
2500 “

2000 ‘

1500 '

COUNTS

.. Fe
1000 Gr

500 -

 

Fe

 

 

 

 

Figure 56. The STEM X-ray micro-chemical analysis of the M,,C. type

carbide precipitate showing that precipitate is rich in
Cr content.

129

 

Figure 57. High resolution transmission electron micrographs of laser
alloyed region showing MZSC. type carbide precipitate.

i f
f
a?‘
..1 .4 .
N ‘ I
. _ _- f
\ v
‘ .r
-- . .

 

 

1:30

 

 

_o
01
(001) caha-1'19)
(IIE) ”gees
__ o
(zzv) 9953(35'19)- 5"
(I'vz) E:>L('a.i‘:t:>)--' -
' E L a 0
(Z V0) 0 (3:1 13)—' "'""‘m
(Hz) 903(ags13) J
_o
\D
(ow) ”095*
(zoo) 130—» ‘
_o
P‘
(W9) 9o£z(a.-I‘13)_.
__,o
OD
(211) 1920—»
A
cu
V
o
‘— 0‘

 

AilSNBiNI

 

DiFFRACTION ANGLE, 29

(a) as-laser-treated

X-ray diffractometry data from laser surface alloyed

AISI 1018 steel.

Figure 58.

man-

 

 

 

‘131

Amvca mowuusm mo oommusm ecu mafinmuaoa haunwaa Hound nay

.Heoum waoa HmH<
cohoaae mueMRSm momma aouw sump %uus80uoeumm«p hou.x .mm ouswum

om .msoz< zepoéaaa

.i!

on 3 on .8 on 8 o...

_ . _ _ égrisyéyééfiési

 

 

 

‘12—.

(ZOO) 3""
(ZII) .0——~
(ZII) D—o

 

(no).

 

 

 

 

(IIZ) 803mg ‘19)...
(920) 93£Z(ag":ro)—'

(on) 9:)“(95 ‘13) —"

3v

(€€€)'(Sll) 9953(93'19)

 

 

 

AllSNHlNI

132

 

Figure 59. TEM micrograph of laser surface alloyed A181 1018 steel
showing carbide precipitates (this treatment was carried out
without binder while depositing Cr powder on the substrate).

133

sample with chromium. In this case the Cr-powder was deposited on
surface without any binder and the laser treatment was carried out with

same values of parameters but without coaxial argon gas flow.

X-ray diffractometric analysis of the laser treated surface
after lightly polishing (as 70pm), showed ferrite (a), martensite (a'),

carbide [(Cr,Fe),,C.] and some traces of oxide [(Cr,Fe),O,]
(Figure 58b). The disappearence of other kind of carbides, (Cr,Fe).,Cs ,

is attributed to cooling rates present in this region. At the surface,
extremely rapid solidification process and higher content of Cr and C

introduces nonequilibrium phase such as: (Cr,Fe).,C,. However, in above

both cases a few reflections remained unaccounted for.

‘The quantitative observations of '1 distribution of elements in
matrix and in carbide were done simultaneously by STEM X—ray
micro-chemical analysis. The results are shown in Figure 60. The
relative distribution of Cr and Fe in these regions show that carbides
are very rich in Cr in contrast to base metal. Further these results
were made more convincing by a technique called as an image analysis.
The STEM image of laser alloyed surface (Figure 60) was assigned with
different pseudocolors for different elements and several images were
obtained on CRT screen. These images show distribution of elements Cr
(Figure 61) and Fe (Figure 62) independently and Cr-Fe together as a
composite (Figure 63). These images show that Cr is uniformly
distributed in matrix except in the region surrounding a carbide. This

region is totally depleted of Cr.

134

SENS: 1 "HG! 80068 InlaTEEL Oifilfl_flfl3

 

Figure 60. The STEM x-ray micro-chemical analysis of laser surface
alloyed region in AISI 1018 steel.

135

 

Figure 61. A color photograph corresponding to energy dispersive x-ray
maping, showing the distribution of Cr in matrix and carbide
precipitate. The color code bars from right to left
indicate the increasing order of concentration.

Figure 62.

136

 

A color photograph corresponding to energy dispersive x-ray
maping, showing the distribution of Fe in matrix and carbide
precipitate. The color code bars from right to left
indicate the increasing order of concentration.

137

RED-CR GREEN-FE

 

Figure 63. A color photograph corresponding to energy dispersive x-ray

maping, showing the distribution of Cr and Fe together in
matrix and carbide precipitate. Green color represents
element Fe and red color represents element Cr.

‘n .4.
I.
.\~
, I
..
.
- '-

\l

\.,

 

 

138

The general microstructure survey of the laser surface

alloyed region showed a fine uniform distribution of M236. type

precipitate in martensite and ferrite matrix. 0n the basis of
microstructural analysis and ternary Fe-Cr-C system phase diagram
(Figure 64), possible phase transformation sequence could be presented

as follows:

L--)L+a—-)(a+a')I+L—-O(a+a')1+(a+H2,C.)II

Where L is the liquid. It appears from the above sequence that the
first transformation products (I) were ferrite a and martensite 0',

followed by the formation of M230. carbide along with ferrite. This

sequence differs from the conventional austenitized and “quenched
microstructure of dislocated lath martensite with fine interlath films
of retained austenite. Such changes in the as-quenched microstructure
are often found in rapid solidification process (ESP) and are a direct
consequence of the high rate of solidification and subsequent solid

state cooling (145)associated with laser alloying.

Figure 64 is a binary pseudo-phase diagram derived from a
three-dimensional ternary Fe-Cr-C diagram by taking cross-section at a
carbon content of 0.1 weight percent. This pseudo-binary diagram is a
close approximation of the alloy studied, and thus this binary section
is helpful in understanding the solidification and cooling reactions.
However, caution must be exercised since the Figure 64 is an equilibrium
phase diagram where as LSA process is dominated by non-equilibrium

conditions. On cooling from the liquid, an Fe-12§Cr-0.18§C (all wt.§)

"s

1:39

 

0.10% CARBON

 

 

 

 

 

 

 

 

 

 

 

 

I,” a + (C!) + “sacs ((31) + fiasco
400 I 1 l I I‘
0. 20 40' so

WEIGHT PERCENTAGE cunounm

Figure 64. A cross-sectional diagram for Fe-Cr-C system containing
0.l§C (ref. 138).

140

alloy would first pass through the a + L, two phase region, with

possibly some high temperature ferrite formation. Once below about

13000C, the alloy would be in the 1-phase field until about 900 06.

Below this temperature, ferrite becomes stable once again (due to Cr,
which is a ferrite stabilizer) in association with martensitic
transformation of the unstable 'y-phase. It is well known that rapid
solidification process can produce large undercoolings below liquidus
temperature, often large enough for hypercooling to occur (146).
Hypercooling is the condition that exists when the undercooling is far
enough below the solidus temperature to ensure that any temperature
increase during recalescence will not bring the solid/liquid interface
temperature up to that of the solidus. When this occurs, solidification
takes place without diffusion. Considering the Fe-Cr-C system, it is
likely that an undercooling would be sufficient for hypercooling to
occur, resulting in the formation of austenite. Finally, due to
extremely high rate of cooling this austenite further transforms to
martensite (see Appendix C). Thus, in this process ferrite and
martensite are the products of inherent rapid rate of quenching. In our
present study, no interlath austenite has been detected by either TEM or
X-ray diffractometry studies. One possible reason for this may be that

the martensite structure is extremely fine and thus any austenite film

0
present (between martensite laths) is s 100 A thick, and therefore,

difficult to detect .

Extension of solid solubility of Cr and C in ternary system

(Fe-Cr-C) depends upon the cooling rate. The cooling rate in LSA

5 e 0
process is very close to that of RSP i.e. 10 to 10 C/sec. At this

141

cooling rate, the above mentioned phases (ferrite and martensite) are
unable to reject substantial amounts of solute atoms such as Cr and C.
Therefore, the remaining excess amount of Cr and C would be rejected by
the non-equilibrium phases into the remaining melt. Now this remaining
melt contained supersaturated elements such as Cr and C which probably
formed a deep eutectoid. Again because of high cooling rates, fcc

carbide, MNC. type, would be formed which is an equilibrium carbide

precipitate (Figure 55). This explanation is based upon the assumption

that MNC, carbides are formed directly from the melt and was not

prevented by rapid quenching. Similar explanation was stated by Molian

et al. (147) in formation of Mac. carbides, during phase transformation

studies of laser processed Fe-30%Cr-0.2%C (all wt.%) surface alloy. But
at this juncture, there is no other way to confirm this explanation.

The another possible explanation for formation of u,,c, carbide is that

it could be through a solid 'state transformation. In this LSA process,
several overlapping passes were laid on the material. The temperature
developed at laser beam-metal interaction is sufficiently high to raise
the temperature of surrounding region to several hundreds of degrees.
The solidified and alloyed region due to previous pass/passes undergoes
a tempering treatment. In low carbon-alloy steel cementites and
carbides formed during solidification are dissolved during tempering and
a sequence of alloy carbides are formed (143). Similarly, in present

LSA process, MNC. type carbides may have been formed from cementites

and other type of carbides, due to tempering. The formation of alloy
carbides in this manner is sometimes referred to as secondary hardening

(143).

142

In the present study, the rate of cooling is of the same order
of magnitude as for RSP but the final microstructural products were
different. They consisted of carbide precipitate in ferrite and
martensite matrix, which is different from the microstructures reported
by others (135, 141, and 145-147) in Fe-base alloys. This is probably
due to different alloying elements and difference in their amount.

Singh and Mazumder (141) observed a/ M.C / x(chi) / M7C, as

solidification products in laser clad 43.38Fe-508Cr-5.l%Mn-0.6%C (all
wt.%). The occurence of austenite or martensite was not observed
because of high Cr content (Cr > 40%) in ferrite matrix which supressed
the formation of any austenite. Rayment and Thomas (137) noticed
ferrite and martensite as end products in rapid solidification followed
by electron beam melting of Fe-3%Cr-2%Mn-0.5%Mo-0.3%C (all wt.§). Mn
is an austenite forming element and thus it is surprizing that no
austenite was retained. On the other hand, with increasing the content
up to lOwt %Cr in ternary Fe-Cr-C and quaternary Fe-Cr-Mn-C system, a

duplex microstructure of ferrite and martensite, along with Mac type

carbide precipitates were reported for rapidly solidified alloys by
Malian et al. (147). With further increase of Cr content up to about 20

WA, the basic microstructure of the ternary alloy consisted of ferrite

e o
with martensite for a cooling rate of about 10 C/sec. (147) . Further
raising Cr content to 30 wt.‘ in the same ternary alloy, Malian et a1.

noticed ferrite and Mac. carbide as the final products of rapid

5 o -
solidification. For a cooling rate of 10 C/sec., Inoue et al. (135)
reported a single phase austenite microstructure in an Fe-l7tCr-1.5&C
(all th) steel. In general, the volume fraction of austenite phase

decreased with increasing Cr content in the ternary Fe-Cr-C system.

143

From this, it is possible to argue that the existence or formation of
austenite phase in Fe-Cr-C and Fe-Cr-C-Mn systems, depends not only on

composition of the alloy but depends upon the cooling rates.

V . CONCIDSIORS

The present study of laser surface alloying of AISI 1018

steel with chromium leads to the following conclusions:

An increase in the laser power increased the depth of laser
alloyed zone and hence it produces a laser surface alloy with lower
alloy content.

An increase in the traverse speed decreased the depth of laser
alloyed zone which in turn increased the content of alloying element in
the laser surface alloy.

An increase in the laser beam diameter (at the surface) also
decreased the depth of laser alloyed zone which in turn increased the
content of alloying element in the laser surface alloy.

An increase in the height of chromium powder layer onto the top of
the substrate surface increased the content of alloying element in the
laser surface alloy.

The defects such as porosity and cracks are produced during laser
surface alloying. The cracks in the heat affected region are more or
less parellel to the direction of the motion of the workpiece.

A laser surface alloyed Fe-Cr-C alloy produced a very fine grain

microstructure of ferrite, lath martensite and complex carbide (M23C6)

precipitation. The substructure of martensite was dislocated and
crystal structure of martensite was bcc.

The increase in solid solubility and high cooling rate produced
first martensite along with high temperature ferrite followed by the

formation of 1423C6 (fcc) carbide precipitates. The M2366 carbide

144

145

precipitates were uniformly distributed in the matrix which had a high

content of chromium.

APPENDIX.A

Spatial Profile of a Laser Beam.

00 ’0

O

   

a!
(m

0
x

02

Figure 65. Some low-order TEMum cavity modes for cylindrical symmetry.

The dark areas represent areas of enhanced output,although
the distribution of intensity within these regions is not
uniform.
The spatial profile of a laser beam is determined by the
geometry of the laser cavity. The shape of the laser cavity in a
direction transverse to the optical axis dectates the boundary
conditions for the wave equation which determines which configurations
of the electromagnetic field will be allowed in the cavity. When this
cross section is symmetrical, as in a cylindrically or rectangularly

shaped resonators, the spatial profile of the permitted transverse

electromagnetic modes is taken from the theory of waveguides.

1416

147

Particular modes are labeled TEan where m and n are the number of nodes

in two orthogonal directions. Both cylindrical and rectangular

geometries have the TEM” (Gaussian) mode as the transverse mode of
highest symmetry. Other higher order modes are labeled TEM“, TEM“,
TEM” and so on. Some of these modes for cylindrically shaped resonator

cavities are shown in Figure 65. Since, in a Gaussian mode a spatially
uniform incident laser fluence is available, the lasers with this kind
of mode structures are widely used in surface treatments. In order to
inhibit operation on high order modes the diameter of the laser cavity

can be reduced until operation on TEM” or the fundamental mode is

produced. The critical diameter of the resonant laser cavity for

initiation of the TEM“ mode is a function of the cavity length.

Operation in the fundamental mode provides true diffraction limited beam
divergence. Higher order mode outputs are accompanied by larger beam
divergence. Operation in high order mode may also be accompanied by
instabilities in laser output as oscillation shifts from one transverse

mode to another with similar gain.

APPENDIX.B

Martensitic Transformation.

If a sample of steel is cooled with a very high cooling rate
from the temperature in a austenite 7, phase region, then it transforms
into entirely different phase in major quantity along with very little
amount of conventional phases. The new non-equilibrium phase is called
as martensite. This transformation is diffusionless and there is no
change in chemical composition. The transformation proceeds only during
cooling and ceases if cooling is interupted. Therefore, the
transformation depends only upon the decrease in temperature and is
independent of time. A transformation of this type is said to be
athermal. The temperature of the start of martensite formation is known

as the Ms temperature and that of the end of martensite formation as the

Mf temperature.

TIME
| a I Mom I h |
12 4 81530602 4 81530602 4 8153060

 

sun an
rxm 1&0

:5 12m 648 U

“" g

310x: sma_

p-

g 3‘:
an «as:

g a.

U

h an sm"
«n zoo
zoo lbs-quenched 3.7 RC 93

Figure 66. Isothermal transformation diagram for 0.06‘C-12.5%Cr (wt.%)
stainless steel (ref. 149).

1418

149

Figure 66 is a isothermal transformation diagram for
Fe-12.5%Cr-0.06%C (all wt.%) alloy. This isothermal diagram is a close
approximation of the alloy studied (Fe-12.5%Cr-0.18%C, all wt.%) and
hence it is useful in understanding the cooling reaction. The cooling
rate at the tip of the nose of the transformation start curve (in Figure
66) is called as the critical cooling rate for that alloy. For a any
cooling rate higher than the critical cooling rate, a steel will be

transformed into martensitic phase in major amount. In our present

laser surface alloying process the cooling rate attained is £105-

0
106 C/sec. This cooling rate is much higher than the critical cooling
rate for the alloy studied. Hence the end product consists of

martensite in substantial amount.

APPENDIX.C

Crystallography of Martensite .

The crystal structure of martensite a' , obtained by quenching
the austenite 1, phase in steels has a bcc/bct (body centered cubic/body
centered tetragonal) lattice. In 1924 Bain (149) demonstrated how the
bet lattice could be obtained from the fcc structure with the minimum of
atomic movement, and the minimum of strain in the parent lattice. As
shown in Figure 67, an elongated unit cell of the bee structure can be
drawn within two fcc cells. Transformation to a bet unit cell is
achieved by contracting a axis little and expanding c axis markedly.
This finding has been confirmed by many researchers, all of whom
reported the same results for the lattice parameters as those shown in
Figure 68 (151-153). The axial ratio c/a increases with the carbon

content. This relation is given by the equation (154,155);

c/a - 1.000 + 0.045 (wt.%C)

The volume of the unit cell also increases linearly with increasing
carbon content. This suggests that carbon atoms are in interstitial
sites in the iron lattice. The sites are 1/2 1/2 0 and/or the

equivalent sites, as shown in Figure 69.

It is interesting fact that the bain deformation involves the
absolute minimum of atomic movements in generating the bcc/bct from the
fcc lattice. Examination of Figure 67 shows that the Bain deformation

results in the following correspondence of crystal planes and

150

1:51

 

.aouo
0 wow canoawo>e macaUHmon ”x “sous oh no ..8 ouamcouuma
on .r ouucoumsm mo acuumauoumsmuu ozu cw oocopcommouuoo amen .mo ouswah

 

 

 

 

 

 

 

 

 

Figure 68.

1552

07 of austenite (3)

sand 0 01 martensite (X)

 

Axial ratio 470 at martensile

l.
C (*1 °/a)

Lattice constants of tetragonal martensite and austenite in
quenched carbon steels (ref. 152).

 

O

 

1;:
L: A.»

Figure 69. Atomic arrangement in austenite 1. 0: Fe atom; X: positions

available for C atom.

 

153

directions:{1ll}1 planes are approximately parellel to (011)“, planes
and that the relative direction can vary between <101>1//<lll>a, (the
Kurdjmov-Sachs relation) and <1T0>1//<101>a' (the Nishiyama-Wesserman

relation).

On adding special elements, such as Ni, Cr, or Mn to steels
containing C or N, we obtain tetragonal martensites, as in plain carbon
or nitrogen steels. Although the lattice parameters a and c change with
the size of the added special element, the axial ratio c/a depends only
on the carbon or nitrogen (156). This fact can be understood from the
fact that the tetragonality is due to the ordered arrangement of the C

or N atoms.

The martensite in substitutional alloys, that do not form an
ordered lattice is likely to be cubic, as in pure iron. Even if these
alloys contain interstitial atoms, the martensite is cubic as long as
the interstitial content is small. This is why no data are shown in
Figure 68 for carbon contents less than 0.25%. There are two
possibilities in this case: one is that the axial ratio is so close to
unity that the tetragonality can not be detected; the other, that the

martensite can be cubic as long as the carbon content is small.

10.

ll.

12.

l3.

14.

15.

16.

VII.REFERENCES

J. E. Harry, Industrial Lasers and Their Applications, McGraw-
Hill, London, England, (1974), p. 8.

J. M. Carroll, The Story of the Laser, E. P. Dutton and Co. Inc.,
New York, (1970), p. 31.

J. E. Harry, Industrial Lasers and Their Applications, McGraw-
Hill, London, England, (1974), p. 77.

J. E. Harry, Industrial Lasers and Their Applications, McGraw-
Hill, London, England, (1974), p. 82.

J. E. Harry, Industrial Lasers and Their Applications, McGraw-
Hill, London, England, (1974), p. 88.

J. E. Harry, Industrial Lasers and Their Applications, McGraw-
Hill, London, England, (1974), p. 89.

J. E. Harry, Industrial Lasers and Their Applications, McGraw-
Hill, London, England, (1974), p. 87.

J. E. Harry, Industrial Lasers and Their Applications, McGraw-
Hill, London, England, (1974), p. 86.

M. Ross, ed.: Laser Applications, vol. 1, Academic Press, New
York, (1971), p. 215.

W. W. Duley, Co2 Lasers: Effects and Applications, Academic
Press, New York, (1976), p. 15.

T. Chande and J. Mazumder, Metall. Trans. 8, 14B, (1983), p. 181-
190.

W. W. Duley, D. J. Semple, J. P. Morency, and M. Gravel, Opt.
Laser Technol., 11, (1979), p. 281.

T. J. Wieting and J. L. De Rosa, J. Appl. Phys., 50, (1979),
. 1071.

P
M. I. Arzuov, A. I. Barchukov, F. V. Bunkin, V. I. Runov, and A.
M. Prokhorov, Sov. J. Quantum Electron., 5, (1975), p. 523.

8. Marcus, J. E. Lowder and D. L. Mooney, J. Appl. Phys., 47,
(1976), p. 2966.

M. E. Maher and R. 8. Hall, J. Appl. Phys., 49, (1978), p. 2254.

1154

17.
18.
19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

1555

E. Sturmer and M. van Allmen, J. Appl. Phys., 49, (1978),p. 5648.

J. E. Robin, J. Appl. Phys., 49, (1978), p. 5306.

U

M. Roessler and V. G. Gregson, Appl. Opt., (1978), p. 992.

. A. McKay and J. T. Schriempf, Appl. Phys. Lett., 35, (1979),
. 433.

S’UL.

. M. Steen and M. Eboo, Met. Const., 11, (1970), p. 332.

Y. Arata, "Basic Characteristics of Large Output High Energy
Density Heat Source”, Paper No. 2, in Proc. lst Joint US/Japan
Int. Laser Processing Conf., Laser Institute of America, Toledo,
Ohio, (1981).

D. N. H. Trafford, T. Bell, J. H. P. C. Megaw and A. S. Branden,
Heat Treatment'79, The Metal Society, London, (1979), p. 33-38.

J. Mazumder and W. M. Steen, J. Appl. Phys., 51(3), (1980),
p. 941

A. F. Gibson, M. H. Kimmit, and A. C. Walker, Appl. Phys., Lett.,
17, (1970), p. 75-77.

A. F. Gibson and A. C. Walker, J. Phys. C., Solid State Physics,
4, (1971). p. 2209-2219.

G. C. Lim and W. M. Steen, Optics and Laser Technology, June
1982, p. 149-153.

C. C. Lim and W. M“ Steen, J. Phys. E: Sci. Instrum., 17, (1984),
p. 999-1007.

C. Courtney and W. M. Steen, Appl. Phys., 17, (1978), p. 303-307.
0. A. Sandven, Laser Applications in Materials Processing, SPIE,
Society of Photo-Instrumentation Engineers, Washington, D. C.,
(1979), p. 198.

E. V. Locke and R. A. Hells, IEEE J. Quantum Electronics, vol.
QE-lO, No. 2, (1974), p. 179-185.

E. L. Baardsen, D. J. Schmatz and R. E. Bisaro, Welding J., 52,
April (1973), p. 227-229.

C. Courtney and W. M. Steen, paper No. 2, in Proc. Int. Conf. on
Advances in Surface Coating Technology, British Welding
Institute, England, (1978), p. 219-232.

R. C. Crafer, Weld. Inst. Res. Bull., 17, (1976), p. 28.

a. A. Willgoss, J. a. P. Megaw and J. u. Clark,0pt.Laser
Technol., April (1979), p. 73.

W. C. Ball and C. M. Banas, Soc. Auto. Eng., Report 740863,
(1974).

37.

38.

39.
40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.
55.
56.

1£56

. T. Swift-Hook and A. E. F. Gick, Weld. J., 52, (1973),
. 4925.

D
P
T. R. Anthony and H. E. Cline, J. Appl. Phys., 48, (1977),
p. 3888-3894.

J. Mazumder and W. M. Steen, J. Appl. Phys., 51, (1980), p. 941.
C. W. Draper et al., Appl. Surf. Sci., 7, (1981), p. 276-280.

Yu. A. Bykovskii, V. S. Kulikauskas, V. N. Nevolin, and I. D.
Khabelashivili, Zh. Tekh. Fiz., 52(1), (1982), p. 61-63.

A. K. Jain et al., Radiat. Eff., 63, (1982), p. 175-181.

C. W. Draper et al. in 'Laser and Electron Beam Interactions with
Solids', eds.: B. R. Appleton and G. K. Celler, North-Holland,
Amsterdam, (1982), p. 413-418.

P. G. Moor and L. S. Weinmann, in 'Laser Application in Materials
Processing', SPIE Proc., Society of Photo-Optical Instrumentation
Engineers, Bellingham, Wash., vol. 198, (1979), p. 120-125.

S. M. Copley et al., in 'Rapidly Quenched Metals III' ed.: B.
Cantor, The Metals Society, London, vol. 1, (1978), p. 147-150.

D. S. Gnanamuthu, Opt. Eng., 19, (1980), p. 783-792.
. W. Bergmann and B. L. Mordike, Z. Metallkd., 71, (1980),
. 658-665.
. 863-869.

H
P
H. W. Bergmann and B. L. Mordike, J. Mater. Sci., 16, (1981),
P
V. S. Kovalenko, V. I. Volgin, Fiz. Khim. Obrab. Mater., 3,
(1978), p. 28-30.

V. S. Kovalenko, V. I. Volgin and V. V. Mixhailov, Tekh.
Organizatsiya Proizvod., 3, (1977), p. 50-52.

J. D. Ayers, R. J. Schaefer and W. P. Robey, J. Met., 33(8),
(1981), p. 19-23.

J. D. Ayers and T. R. Tucker, Thin Solid Films, 73, (1980),

p. 201-207.

C. Sepold, in Proc. Int. Cong. on 'Surface Technology 1981', ed.:
C. Rohrbach, Dusseldorf, VDI-Verlag GmbH, (1981), p. 323-330.

J. H. P. C. Megaw, Surfacing J., 11, (1980), p. 6-11.

R. Mehrabian, Int. Met. Rev., 27, (1982), p. 185-208.

S. C. Hsu, S. Kou and R. Mehrabian, Metall. Trans., 9B, (1978),

. 221-229.

'6

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

157’

. C. Hsu, S. Kou and R. Mehrabian, Metall. Trans., 11B, (1980),
. 29-38. .

. Kou, S. C. Hsu and R. Mehrabian, Metall. Trans., 12B, (1981),
. 33-45.
. 411-413.

S
P
S
P
J. A. Sekhar and R. Mehrabian, Metall. Trans., 12B, (1981),
P
T. Chande and J. Mazumder, Metall. Trans., 14B, (1983), p. 181-
1

. E. Cline and T. R. Anthony, J. Appl. Phys., 48, (1977),
. 3895-3900.

. 41-43.

. Mazumder and W. M. Steen, J. Appl. Phys., 51, (1980),

H
P
T. Chande and J. Mazumder, Appl. Phys. Lett., 41, (1982),
P
J
p. 941-947.

M. Bertolitti and C. Sibila, IEEE J. Quantum Electronics, vol.
QE-17, No. 10, October 1981, p. 1980-1989.

J. S. Goela and C. B. Paliwal, Int. Comm. Heat Mass Transfer,
13, (1986), p. 121-125.

J. M. Poate and J. W. Mayer, in 'Laser Annealing of
Semiconductors', New York, Academic Press, (1982).

L. F. Dona dalle Rose and A. Miotello, Radiat. Eff., 53, (1980),
p. 7-18.

A. K. Jain, V. N. Kulkarni and D. K. Sood, Appl. Phys., 25,
(1981), p. 127-133.

P. P. Pronko et al., in 'Laser and Electron Beam Solid
Interactions and Materials Processing', ed.: J. F. Gibbons et
al., North-Holland, New York, (1981), p. 599-606.

C. Battaglin et al., in 'Lasers and Electron Beam Solid
Interactions and Materials Processing', ed.: J. F. Gibbons et
al., North-Holland, New York, (1981), p. 615-621.

A. K. Jain, V. M. Kulkarni and D. K. Sood, Thin Solid Films, 86,
(1981). p0 1‘9.

. Sawchyn and C. W. Draper, Appl. Surf. Sci., 18, (1984),
. 86-105.

Battaglin et al., J. Appl. Phys., 53, (1982), p. 3224-3230.

Miotello and L. F. Dona dalle Rose, Phys. Lett., 87A, (1982),
317-320.

133’ C) -o F1

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

89.
90.

91.

15!!

L. F. Dona dalle Rose and A. Miotello, in 'Laser and Electron
Beam Interactions with Solids', eds.: B. R. Appleton and G. K.
Celler, Nroth-Holland, Amsterdam, (1982), p. 425-430.

C. G. Gladush, L. S. Krasitskaya, E. B. Levchenko and A. L.
Chernyakov, Kvantovaya Elektron, 9, (1982), p. 660-667. (Sov. J.
Quantum Electron., 12, (1982), p. 408-412).

0. Esquivel, S. M. Copley and M. Bass, "Convective Processes in
Continuously Melted Zones”, unpublished work, Dept. of Materials
Science, USC, Los Angeles, CA 90007.

A. K. Jain et al., Nucl. Instrum. Methods, 168, (1980),
p. 275-282.

H. W. Bergmann and B. L. Mordike, Z. Werkstofftech, 14, (1983),
p. 228-237.

C. W. Draper et al. in ' Laser and Electron Beam Processing of
Materials', eds.: C. W. White and P. S. Peercy, Academic Press,
New York, (1980), p. 721-727.

C. W. Draper and J. M. Poate, in 'Surface Modifications and
Alloying', eds.: J. M. Poate et al., Plenum Press, New York,
(1983), p. 385-404.

M. Hansen and K. Anderko, 'Costitution of Binary Alloys', McGraw
Hill, New York, (1958).

L. Buene et al., Appl. Phys. Lett., 37 (9180). p. 385-387.

C. W. Draper, L. Buene, J. M. Poate and D. C. Jacobson, Appl.
Opt., 20, (1981), p. 1730-1732.

L. Buene et al., in 'Laser and Electron Beam Solid Interactions
and Materials Processing', eds.: J. F. Gibbons et al., North
Holland, New York, (1981), p. 583-590.

D. M. Follstaedt, in 'Laser and Electron Beam Interactions with
Solids', eds.: B. R. Appleton and G. K. Celler, North Holland,
Amsterdam, (1982), p. 377-388.

P. G. Moore and E. McCafferty, J. Electrochem Soc., 128, (1981),
p. 1391-1393.

J. B. Lumsden, D. S. Gnanamuthu and R. J. Moores, in 'Corrosion
of Metals Processed by Directed Energy Beams' eds.: C. R. Clayton
and C. M. Preece, Metallurgical Society of AIME, Warrendale, PA,
(1982). p. 129-134.

P. A. Malian, Scr. Metall., 16, (1982), p. 65-68.

P. A. Malian and W. E. Wood, Scr. Metall., 16, (1982),
. 1301-1303.

P
P. A. Malian and W. E. Wood, J. Mater. Sci., 18, (1983),
p. 2555-2562.

 

 

92.

93.

94.

95.

96.

97.

98.

99.

100.

101.

102.

103.
104.
105.
106.

107.
108.

109.

110.

111.

A.

"U “U"! 'U’U

1£59

. A. Malian and W. E. Wood, Mater. Sci. Eng., 56, (1982),
. 271-277.

. A. Malian and W. E. Wood, J. Mater. Sci., 18, (1983),
2563-2571.

. A. Malian, Mater, Sci. Eng., 58, (1983), p. 175-180.

V. Belotskii, V. S. Kovalenko, V. F. Volgin and V. I.

Pshenichnyi, Fiz. Khim. Obrab. Mater., 3, (1977), p. 24-27.

C

. Sepold, in 'Gas-flow and Chemical Lasers', ed.: J. F.

Wendt, Hemisphere Publishing Corp., Washington D.C., (1978),

P
L

. 561-569.

. Weinmann, J. Devault and P. Moore, in 'Applications of Lasers

in Materials Processing', ed.: E. A. Metzbower, American Society
for Metals, Metals Park, Ohio, (1979), p. 245-257.

C
P
G
p.
C
P
C
6

. J. Lin and F. Spaepen, Appl. Phys. Lett., 41, (1982),
. 721-723.

. G. Borodina et al., Dokl. Akad. Nauk SSSR, 259, (1981),
826-829 (Sov. Phys. Dokl., 26, (1981), p. 761-763).

. J. Lin and F. Spaepen, Proc. Mater. Res. Soc., 28, (1984), 28,
. 75-80.

. J. Lin, F. Spaepen and D. Turnbull, J. Non-Crystallgr. 801.,
1-62, (1984), p. 712-767.

R. C. Ruhl, B. C. Giessen, M. Cohen and N. J.Grant, Acta Metall.,

1

P

P

P

D
4

F.

M.
p.

N.

5 (1967), p. 1693-1702.

. Mazzoldi et al., Phys. Rev. Lett., 44, (1980), p. 88-91.

. L. Bonora et al., Electrochem Acta, 25, (1980), p. 1497-1499.
. L. Bonora et al., Thin Solid Films, 81, (1981), p. 339-345.

. M. Follstaedt, P. S. Peercy and W. R. Wampler, Phys.Rev.Lett.,
6, (1981), p. 1250.

Pleiter and K. G. Prasad, Phys. Lett., 84A, (l981),p. 345-348.

Stern and H. Wissenberg, J. Electrochem. Soc., 106, (1959),
759-764.

D. Tomashov, G. P. Chernova and T. A. Fedoseeva, Prat. Met.

(USSR), 13, (1977), p. 134-138.

N

. D. Tomashov, G. P. Chernova and T. A. Fedoseeva, Corrosion

(Houston), 36, (1980), p. 201-207.

C

. W. Draper et al., Thin Solid Films, 75, (1981), p. 237-240.

 

 

112.

113.

114.

115.

116.

117.

118.

119.

120.

121.

122.

123.

124.

125.

126.

127.

128.

129.

130.

16C)

J. M. Poate, H. J. Leamy, T. T. Sheng and G. K. Celler, Appl.
Phys. Lett., 33, (1978), p. 918-920.

R. T. Tung, J. M. Gibson, D. C. Jacobson and J. M. Poate, Appl.
Phys. Lett., 43, (1983), p. 476-478.

C. W. White, S. R. Wilson, B. R. Appleton and F. W. Young Jr., J.
Appl. Phys., 51, (1980), p. 738-749.

S. C. Hsu, S. Chakravorty and R. Mehrabian, Metall. Trans., 9B,
(1978), p. 221-229.

A. N. Nesmeyanov, 'Vapor Pressure of Elements', Academic Press,
New York, (1963).

M. Van Allmen and K. A. Affolter, Proc. Mater. Res. Soc., 28,
(1984), p. 81-85.

 

 

O. Kubaschewski, E. LL. Evans and C. B. Alcock, Metallurgical
Thermochemistry, Fourth Edition, Pergamon Press, New York, N. Y.,
(1967), p. 303-447.

S. L. Narasimhan, S. M. Copley, E. M. Vanstryland and M. Bass,
Metall. Trans., 10A, (1979), p. 654-655.

 

N. B. Dahotre, A. Hunter and K. Mukherjee, J. Mater. Sci., 22,
(1987), p. 403-406.

B. S. Kasatkin and A. M. Tsaryuk, Automatic Welding, 2, (1965),
p. 1-7.

K. Mukherjee, T. H. Kim and W. T. Walter, in 'Lasers in
Metallurgy', eds.: K. Mukherjee and J.Mazumder, The Metallurgical
Society of AIME, Warrendale, PA, (1981), p. 137-150.

D. Rosenthal, Welding J., 20, (1941), p. 220s-234s.

R. E. Spear, G. R. Gardner, Trans. AFS, 71, (1963), p. 209.

T. F. Bower, H. D. Brady and M. C. Flemings, Trans. AIME, 236,
(1966), p. 624.

M. C. Flemings, D. R. Poirer, R. V. Barone and H. D. Brody,
J. Iron Steel Inst., 208, (1970), p. 371.

A. Suzuki, T. Suzuki, Y. Nagaoka and Y. Iawata, Nippon Kingaku
Gakkai Shuho, 32, (1968) also Brutcher Transl., 7804, (1969.

M. C. Flemings, Solidification Processing, McGraw Hill Inc., New
York, (1974), p. 146-154.

W. F. Savage, E. F. Nippes and J. S. Erickson, Welding J., 55,
(1976). p. 213s-22ls.

M. F. Ashby and K. E. Esterling, Acta Metall., 32(11), (1984),
p. 1935-1948.

131.

132.

133.

134.

135.

136.

137.

138.

139.

140.

141.

142.

143.

144.

145.

146.

147.

161

Y. W. Kim, P. R. Strutt and H. Nowotny, Metall. Trans., 10A,
(1979), p. 881.

K. Lipscombe, W. M. Steen and D. R. P. West, in 'Rapid
Solidification Processing: Principles and Technologies II', eds.:
R. Mehrabian et al., Claitor's Publishing Division, Baton Rouge,
LA, (1980), p. 189.

S. M. Copley, D. G. Beck, 0. Esquivel and M. Bass, in 'Lasers in
Metallurgy', eds.: K. Mukherjee and J. Mazumder, Metallurgical
Society of AIME, New York, (1981), p. 11.

G. Christodoulou, A. Walker, W. M. Steen and D. R. F. West,
Metals Technology, 10, (1983), p. 215-223.

A. Inoue, T. Iwadochi, T. Minemura and T. Masumato, Trans. J. I.
M., 22, (1981), p. 197.

R. W. Messler, G. S. Ansell and V. I. Lizunov, Trans. ASME, 62,
(1981), p. 213.

J. J. Rayment and G. Thomas, in 'Rapidly Solidified Amorphous and
Crystalline Alloys', eds.: B. H. Kear, B. C. Giessen and M.Cohen,
Proceedings of the Materials Research Society, North Holland
Publishing Co., New York, 8, (1981), p. 511.

M. Carbuciccho and G. Palombarini, J. Mat. Sci., 15, (1985),
p. 75-82.
P

. A. Malian, J. Mat. Sci., 20, (1985), p. 2903-2912.

J. Singh and J. Mazumder, Materials Science & Technology, 2, July
1986, p. 709-713.

J. Singh and J. Mazumder, Metall. Trans., 18A, (1987),p. 313-322.

P. M. Kelly and J. Nutting, J. Iron Steel Institute, (1961),
p. 199.

Metals Handbook, Metallography, Structures and Phase Diagrams,
American Society for Metals, Metals Park, Ohio, vol. 8, (1973),
p. 202-204.

K. W. Andrews, D. J. Dyson and S. R. Keown, Interpretation of
Electron Diffraction Patterns, Plenum Press, New York, (1967),
p. 197.

J. J. Rayment and B. Cantor, Metall. Trans, 12A, (1981), p. 1557.

M. Cohen, B. H. Kear and R. Mehrabian, in Proceedings of 2nd Int.
Conf. on Rapid Solidification Processing, Reston, Virginia, March
1980, Claitors Publishing Division, Baton Rouge, LA, p. 1.

P. A. Malian, P. J. Wang, K. H. Khan and W. E. Wood, in the
Materials Research Society Proceedings of 'Rapidly Solididified
Amorphous and Crystalline Alloys', eds.: B. H. Kear, B. C.

 

 

 

148.

149.

150.

151.

152.

153.

154.

155.

156.

1152

Giessen and M. Cohen, North Holland Publishing Co., vol. 8,
(1981). P. 511.

O. T. Inal, L. E. Murr and F. G. Yost, Mater. Sci. Eng., 51,
(1981). p. 101-110.

K. G. Brickner, Stainless Steels for Room and Cryogenic
Temperatures in Selection of Stainless Steels, American Society
for Metals, Metals Park, Ohio, (1968), p. 1-29.

E. C. Bain, Trans. AIME. 70, (1924), p.25.

W. B. Pearson, Handbook of Lattice Spacing and Structure of
Metals and Alloys, Pergamon, Oxford, (1958), p. 921.

K. Honda and Z. Nishiyama, Sci. Rep., Tohoku Univ., 21 (1932),
p. 299. ’

C. S. Roberts, Trans. AIME, 197, (1953), p. 203.

C. S. Roberts, B. L. Averbach and M. Cohen, Trans. ASM, 45,
(1953), p. 576.

P. G. Winchell and M. Cohen, Trans. ASM, 55, (9162), p. 347.

Z. Nishiyama, Sci. Rep., Tohoku Univ., 24, 128(1935), p. 517.