An Investigation into the Tribological Properties of Thermally-Oxidized 6Al-4V

bistredingdongMechanics

Oct 31, 2013 (3 years and 9 months ago)

206 views




An Investigation into the Tribological Properties of Thermally
-
Oxidized 6Al
-
4V
Titanium
Coupled

with
Select

Engineering
Polymer
s

by

Matthew E. Lessard

An Engineering Project
Submitted to the Graduate

Faculty of Rensselaer Polytechnic Institute

i
n

Partial

Fulfillment of

the

Requirements for the degree of

MASTER OF ENGINEERING IN MECHANICAL ENGINEERING













Approved
:


__________
_______________________________

Dr.
Sudhangshu

Bose
,
Thesis Adviser





Rensselaer Polytechnic Institute

Hartford, Connecti
cut

April, 2011



2




























© Copyright
201
1

by

Matthew E. Lessard

All Rights Reserved



3

CONTENTS

LIST OF TABLES
................................
................................
................................
..............
4

LIST
OF FIGURES

................................
................................
................................
............
5

ACKNOWLEDGMENT

................................
................................
................................
....
6

ABSTRACT

................................
................................
................................
.......................
7

1.

INTRODUCTION

................................
................................
................................
........
8

1.1

Polymer Composite Bearing Materials

................................
.............................
10

1.2

The Titanium
-
Polymer Tribosystem

................................
................................
.
12

2.

THEORETICAL BACKGROUND
................................
................................
............
16

2.1

The Hydrogen
-
Assisted Wear Mechanism

................................
.......................
16

2.2

Thermal Oxidation (TO) Treatment of T
itanium

................................
..............
21

3.

METHODOLOGY / APPROACH

................................
................................
.............
26

3.1

Thermal Oxidation (TO) Surface Engineering
................................
..................
27

3.2

Polymeric Test Materials

................................
................................
..................
31

3.3

Wear Testing (Pin
-
on
-
Disk Method)
................................
................................
.
33

4.

PIN ON DISK TRIBOMETER

................................
................................
..................
36

4.1

Design and Configuration

................................
................................
.................
37

4.2

Data Acquisition System (DAQ)
................................
................................
.......
44

5.

EXPERIMENTAL RESULTS

................................
................................
...................
47

5.1

Test Disks (Thermally
-
Oxidized Ti64)

................................
.............................
47

5.2

Alpha Case Analysis

................................
................................
.........................
51

5.3

Polymer Test Pins
................................
................................
..............................
54

5.4

Wear Testing Results

................................
................................
........................
55

6.

DISCUSSION AND CONCLUSIONS

................................
................................
......
72

7.

REFERENCES

................................
................................
................................
...........
75




4

LIST OF
FIGURES

Figure 1.

SEM Image of Wear Scar on the Test Surface of a 6Al
-
4V T
itanium Test Disk
(Subjected to Dynamic Pin
-
on
-
Disk Wear Testing) [7]

................................
.............
13

Figure 2.

SEM image of
Beach Marks

on a Worn Ti64 Disk Surface [2]

.......................
14

Figure 3.

SEM Image Showing Severe Wear of a Ti64 Test Surface by Third
-
Body Wear
Mechanism [25]

................................
................................
................................
..........
15

Figure 4.

Illustration of Proposed Hydrogen
-
Assisted Wear Mechanism for the
T
ribosystem UHMW/Ti6Al4V [8]

................................
................................
.............
18

Figure 5.

Chart Comparing Wear Rates (as a function of Incubation Time) for Various
Metal
-
Polymer Tribosystems [7]

................................
................................
................
19

Figure 6.

The Unit Cell of Rutile TiO
2

(Titanium atoms are grey and oxygen atoms are
red). [29]

22

Figure 7.

Cross
-
Sectional Micrograph (a) Showing Alpha Case Layer Detail

and Plot
(b) showing Knoop Microhardness for a Thermally
-
Oxidized Ti64 Sample [27]

.....
23

Figure 8.

Alpha Case Depth as a Function of Heat
-
Treatment Temperature for
Commercially
-
Pure (CP) Titani
um, Held at Temperature for 1
-
2 Hrs [4]

................
24

Figure 9.

Microhardness
-
Depth Profile for Two 6Al
-
4V Titanium Articles Subjected to
Heat Treatment in an Air Atmosphere at 850°C for Different Lengt
hs of Time [10]

25

Figure 10.

A Comparison Showing the Variation in Case Thickness for Furnace
-
Cooled
samples and Air
-
Cooled samples as a Function of Treatment Time (A) and
Temperature (B). F
igures (C) and (D) Show SEM Image of Oxide Layer Cross
-
Section for Air
-
Cooled Sample and Furnace
-
Cooled Sample (respectively) [20]

.....
29

Figure 11.

The Basic Chemical Structure of the Acetal Copo
lymer Chain [16]

..............
32

Figure 12.

The General Chemical Structure of Epoxy [17]
................................
..............
33

Figure 13.

General Configuration of the Pin on Di
sk Wear Test

................................
.....
33

Figure 14.

Pin on Disk Tribometer

................................
................................
...................
36

Figure 15.

Pin on Disk Tribometer Cross
-
Section View

................................
..................
38

Figure 16.

Pin on Disk Load Arm Component

................................
................................
.
40

Figure 17.

Load Arm Structural Analysis at Max Anticipated Operating Load
Conditions (Spring F
S
=180
lbf, F
D
=24 lbf)

................................
................................
42

Figure 18.

Test Rig Proximity (Speed) Sensor

................................
................................
.
43




5

LIST OF
TABLES




6

ACKNOWLEDGMENT

I would like to thank professors

Sudhangshu

Bose

and
Ernesto
Gutierrez
-
Miravete

for
their guidance and support
while developing and compiling this
work
. I would also like
to thank my mentors
and colleagues
at Kamatics Corporation for their
direction and
assistance
, especially
Mat Mormino, Mark Broding and Tom Rutled
ge
. I owe many
thanks to my
parents
who
have

always supported and encouraged me during
trying
times.
And above all
, I would like to thank my
wife Tiffany
for
her
love, support and
patience while I was earning my degree.



7

ABSTRACT

This project examines the

tribological properties
of
the
titanium
-
polymer
tribosystem
and evaluates the potential improvements that can be achieved by employing
titanium
Thermal Oxidation

(TO)
surface engineering techniques
.

The specific wear
rates realized when coupling titanium

alloys with various engineered plastics
(polyethylene, thermosets, etc.)
in dynamic operation are often found to be significantly
higher than those experienced between the same polymeric materials and
other
metallic
materials
possessing
comparable
interfa
ce
characteristics (finish, hardness, etc.). It is
believed that
these relatively poor wear characteristics are directly linked to a hy
drogen
-
embrittlement phenomenon
associated with
the titanium substrate
; one

which
results
from
the breakdown of certain
reactive groups found in
many
engineering

polymers.
By
subjecting titanium alloys to
specific thermal process treatments

in an oxygen
-
containing
atmosphere
, it
is possible to develop a dense
oxygen
-
diffused
case on
the surface of the
titanium alloy.
Base
d on historical testing performed by other researchers
,
it is believed
that
this oxygen diffused layer should be effective in preventing the inward diffusion of
hydrogen
polymer
byproducts into the titanium bulk
,

thereby
reducing the
specific
wear
rates
si
gnificantly
in th
ese

system
s
.


The ultimate goal of this project is to evaluate the tribological performance of
6Al
-
4V titanium
as
related

to
three specific engineering polymers, and also the
improvements that can be achieved by applying different types
of thermal oxidation
treatments to the titanium interface surface.



8

1.

INTRODUCTION

As a structural material, Titanium and its alloys have proven their utility in a variety
of industries throughout modern design; none more so than in the aerospace industry.
T
he relatively light weight of titanium, coupled with its
exceptional
mechanical
properties
(high strength)
and
inherent
corrosion
-
resistan
ce

make it invaluable as a
structural material in
an industry where weight
-
savings comes at a premium
. Despite its
me
chanical merits however, the relatively poor tribological properties
(wear and friction

characteristics
)
of titanium tend to limit its application in systems with the potential for
dynamic contact. This is
certainly a
disadvantage

when employing
t
itanium
alloys in
kinematic linkages and
dynamic load
bearing applications
,

to the point that significant
research and design efforts have been dedicated in recent years to developing coatings
and
surface engineering
processes to mitigate wear and reduce friction
in
systems

containing
these materials
.



Historically, the majority of these titanium surface engineering efforts have focused
on the use of
hard coatings

and similar technologies to establish a wear
-
resistant barrier
between titanium components. Hard co
atings rely on the deposition of certain materials
(i.e. chrome oxide, tungsten carbide,
etc.)
on interface surfaces using various
application
techniques
(
thermal spraying,
chemical

vapor deposition
, physical

vapor deposition
)
that
yield a
very hard and
ab
rasion
-
resistant superficial coating.

The use of hard interface
coatings is an effective practice
in

reducing titanium component wear, however,
most
are
still
limited
to

relatively low bearing pressures

in the absence of boundary
lubrication
. When
dynami
c
loads increase beyond threshold
contact
pressures, friction
values often
rise to
unacceptable levels
, leading to excessive wear and premature failure
.


This need for titanium interface coatings

and
materials capable of dynamic
operation under
higher

co
ntact
-
pressures often leads designers to consider the use of
polymeric bearing materials (
engineer
ed

plastics
)

for
those
applications where surface
coating technologies fall short
.
T
he
inherent low
-
friction characteristics
and controlled
wear
-
rates
that c
an be achieved by coupling
polymer composite
s

with controlled mating
surfaces
make
s

them
ideal for dry sliding
bearing
applications

that operate at low speeds.


9

M
etal
-
polymer tribosystem
s
(dynamic material couples)
are typically designed with
specific mater
ial
s
that yield very low frictional forces and also keep the rate of material
wear
(volume loss)
very low. These tribosystems are designed such that all material
wear occurs in the polymer material, effectively making the plastic component a
s
acrificial e
lement in the assembly and saving the cost of replacing
mating
metallic
components.



There are a number of metallic and ceramic materials that present a suitable mating
surface for bearing polymers, most of which are characterized by a relatively high
mat
erial hardness and fine surface finishes.
Historically,
it is
these characteristics
that
have demonstrated the greatest impact on specific wear rates at high contact pressures,
however the composition of the mating substrate also plays an important role i
n the life
of the bearing system.
High
-
hardness, corrosion
-
resistant steels are the most common
mating
materials
for engineered plastics;
with a
carefully co
ntrolled mating surface
finish, very
low, uniform
polymer
wear rates
can be achieved in addition t
o a low
-
friction interface
. Similar
(low) polymer
wear rates can be expected
when mated
with
other materials such as nickel
-
based
and cobalt
-
based
super alloys

that maintain
similar
levels of
hardness
and
surface
finish requirements
.



Interestingly howe
ver, when mating these same polymer materials against titanium
alloys possessing the same
general material properties (similar
hardness levels and
surface finish requirements
)

as the
more common
stainless steels, the resultant specific
wear rates are often

found to be significantly higher

for the same operating conditions
.
In addition to the increase in specific polymer wear rates
, these
dynamic
titanium
-
polymer
systems also
often
result in actual wear (volume loss) of the titanium
counter
sur
f
ace
; a condit
ion that is not realized with other mating materials
. Given the
significant difference in relative material hardness between titanium alloys and the
mating plastic material
s
, this wear phenomenon
proves quite puzzling.



As discussed later in this paper,

it is believed that this
unusual

wear phenomenon is
a direct result of a
hydrogen
-
embrittlement

effect
developed

in the titanium
substrate



10

during operation.


It is t
he
breakdown of certain
reactive

polymer
groups
found in
a
number of engineered
plastic
s

t
hat
ultimately
drives this reaction and the ensuing wear
distress caused in the titanium interface surface.

Historical research
[7. 8]
has shown
that this
wear mechanism is prevalent in the 6Al
-
4V titanium / UHMW polyethylene

dynamic tribosystem and also
that the application of certain surface engineering
processes
(thermal oxidation)
can
significantly

reduce the wear of titanium surfaces

in
this system
.

It is known that this same wear mechanism
also
exists in other titanium
-
polymer tribosystems and this
project will attempt to
characterize the
wear and
friction

properties of three such systems. Additionally, this project also attempts to evaluate the
improvements that can be achieved by employing various sur


1.1

Polymer

Composite Bearing Materials

The
compo
site

bearing polymer

designation

encompasses
a wide range of
engineered plastics that are
specifically
designed to function in dynamic systems with
demanding operating and environmental conditions.
Polymer bulk matrices are
modified by the addition of
str
engthening and self
-
lubricating ‘fillers’
to


in a bulk polymer matrix is an effective method of reducing relative friction and
specific wear in metal
-
polymer
and ceramic
-
polymer
tribosystems. When
polymer
composites
are coupled with
metallic
materials p
ossessing specific surface
characteristics (high superficial hardness, fine surface finish, etc.), are often realized.





The utility of engineering plastics is realized where applications require structural
materials that possess a level of chemical resi
stance or reduced component weight that
cannot be achieved by metals or ceramics. It is the tribological properties of these
materials, however that makes them invaluable in applications where dynamic contact
exists between machine components
.



11


Significant

advances have been made in recent years in the field of polymer
tribology, enabling the use of these materials in systems with very demanding operating
conditions. Of particular interest is the usefulness of polymer materials as solid
lubricants in slidin
g bearing applications where boundary lubrication is not feasible or
cost
-
effective. In modern tribology, only a few select materials have been identified as
possessing the unique molecular properties necessary to make for an effective bearing
polymer.

……L
EAD IN REQD

To make for a useful tribological material, significant efforts have been made to
increase the wear resistance of PTFE, leading to the development of a number of PTFE
composite materials. These composites can typically be classified according t
o two basic
categories which describe the general method of modification [
XX
]; Bulk
-
modified
composites and Interface
-
modified composites. Bulk
-
modified composites employ
‘hard’ filler materials such as ceramics, metals or synthetic fibers directly in the
bulk
self
-
lubricating material (i.e.
Polytetrafluoroethylene

or
PTFE). The function of these
‘hard’ filler materials is essentially to strengthen the polymer matrix, increasing its load
-
carrying capacity and wear
-
resistance. Some of the more common filler
materials used to
increase bulk material strength are glass, aramid and carbon (all usually in short fiber
form).

On the other end of the composites spectrum, interface
-
modified composites employ
a softer self
-
lubricating material (i.e. PTFE, graphite, mol
ybdenum disulfide) as a ‘filler’
in a harder polymer matrix. The PTFE fillers (also commonly in short fiber form)
provide the composite with the desired low friction properties, while the strength and
toughness of the bulk matrix (some common polymers are
epoxy, phenolic and PEEK),
is augmented by additional fillers. Alumina (aluminum oxide) and titanium oxide are
frequently employed as strengthening fillers in interface
-
modified polymer composites
and their particle size is central to the level of wear res
istance that can be achieved using
these fillers. Recent studies [2] have shown that using fillers with particle size on the
nano
-
scale (termed nanocomposites) can effectively reduce wear rates by one and in



12

some cases, two orders of magnitude when compare
d to micro
-
sized filler particles. This
is also true of bulk
-
modified composites.


The application of

engineered plastics and self
-
lubricating polymers as a barrier
against wear
distress (and
ensuing fretting

fatigue failures)
in structural components has
proven an effective technology in the aerospace industry; most notably in bearing
applications. By interfacing ferrous materials possessing specific surface characteristics
(high material hardness / fine surface finishes) with engineered polymer composite
s, it is
possible to design assemblies that are capable of operating at very high
bearing contact

pressures while maintaining very low, controlled wear rates and friction coefficients.
The high strength
-
weight characteristics of titanium alloys would
ther
efore
make them
an ideal candidate for such applications
, however

the wear characteristics of this material
are often prohibitive.



1.2

The Titanium
-
Polymer Tribosystem


In most metal
-
polymer tribosystems, the relationship between
material
wear and
hardness

agrees with the standard model for wear of materials
that was

proposed by
J
ohn
F.

Archard
[
1
8
]

in the early 1950’s

(often referred to as the Archard Equation)
.
According to Archard,
for

most isotropic material
s,
the rate of wear in a material is
found to

be
inversely proportional

to the hardness of that material.


That is to say in a simplified model that, as the hardness of the material is increased,
the specific wear rate of the material
will decrease
. This is typically also true of the
wear rate
s

o
f
the mating polymer materials in the

same

tribosystems.
As the hardness of
the metallic material in this couple is increased, the specific wear rate of the polymer
material is most often also found to decrease.
This relationship
has been

effectively
Where:

v

=
Material v
olume loss

k

=
Wear coefficient

W

=
Normal load

H

=
Material Hardness




13

com
prehensively verified

by
W. Wieleba

[19] through his work
using
PTFE composites

and
various steel mating surfaces.


The
titanium
-
polymer

tribo
system

however
,

appears to be an exception
of sorts to
the
standard wear model that
was
proposed by Archard. Num
erous studies have been
conducted
[
2
,
3
,
7, 8
]
investigating the tribological behavior of titanium coupled with
several

different
engineering polymer
s

using varying operating conditions and
researchers consistently witness the same type of wear phenomenon
regardless of
operating parameters (load, speed, etc.). In most every case, the specific wear rates
developed in the polymer material
s

were
much
higher than
expected
and
more notably,
the
(
much harder
)

titanium countersurface
s

also displayed
significant

w
ear
. In general,
the increase in the wear rate of the plastic material could possibly be explained
as strictly
being
a function

of the dynamics between the titanium microstructure and
that of the
polymeric material, however the wear mechanism developed in

the titanium substrate is
much more difficult to explain.


Figure 1.

SEM Image of Wear Scar on the Test Surface of a 6Al
-
4V Titanium Test
Disk (Subjected to Dynamic Pin
-
on
-
Disk Wear Testing) [
7
]


The relative hardness (and
correspondingly,

the strength)

of allo
ys such as 6Al
-
4V
(the most common

of the
titanium alloy
s
) is two and in some cases, three orders of
magnitude higher than the engineered plastic
s with which they are mated
. When
compared against ultra high molecular weight polyethylene (UHMW
/Ti64 is a
tr
ibosystem that is commonly found in the medical industry
),
the relative hardness of
6Al
-
4V

titanium
is nearly
700 times

greater

[
8
]
. Given
this
large disparity

in material


14

hardness, one would
tend to
expect that the
softer
polymeric material would exhibit

all
of the component wear in this tribosystem
,

however
experimental testing has
demonstrated that
during dynamic operation under load,
the titanium component
consistently
displays significant material wear. For instance, the rate of titanium
material wea
r
generated during
dynamic
operation against UHMW
polyethylene plastic
is quite often comparable
to
the
titanium
wear rates
that develop
when
running against
significantly harder
tool steels and ceramic

materials using the same operating
conditions

[
3]
.


A
n examination of the actual wear scars that develop
on
titanium
surfaces during
testing
against these engineered plastics
suggest that there are some abnormal wear
mechanisms at work behind this strange wear phenomenon; characteristics that are not
typical
ly found in steel
-
polymer tribosystems. This is evidenced by the presence of
certain wear pattern characteristics that
indicate
a type of micro
-
fretting fatigue failure
mechanism.
One
such characteristic that
is
frequently discovered post
-
test
is the
pre
sence of
‘beach marks’
or superficial micro
-
cracks oriented transversely across the
wear zone (perpendicular to the sliding test direction). This micro
-
cracking pattern
indicates a lack of
material
toughness at the
surface of the material which translates

to
micro
-
crack propagation and liberated wear particles.


Figure 2.

SEM image of
Beach Marks

on a
W
orn
Ti64 Disk
S
urface
[
2]




15

Another titanium surface failure mechanism that often manifests itself during
testing at elevated contact pressures is the presence of ‘we
ar craters’ or localized
liberation of larger titanium wear particles in a scarred wear field. This wear
characteristic is sometimes difficult to detect however as the liberated particles often
become trapped in the wear zone and a type of third
-
body wear

mechanism is generated.
Liberated wear particles trapped in the wear path will sometimes become embedded in
the softer mating counterface
, increasing the depth of axial wear scars during additional
testing
through
a plo
wing

action.


Figure 3.

SEM Image Showing Se
vere Wear of a Ti64 Test Surface by Third
-
Body
Wear Mechanism [25]



These are
just

two distinct characteristics that are frequently identified in the
wear scar of a titanium
surface
following dynamic t
esting against certain polymer

countersurfaces. There

are a number of
other
unique
wear
characteristics that
also
develop in these tribosystems
,
many of which
are influenced by operating conditions
such as speed, contact pressure,
temperature,
etc. An experienced metallurgist could
certainly
theorize on the

root cause for each individual wear characteristic, h
owever, it is
not the goal of this project to
develop

a taxonomy

of titanium wear mechanisms
. R
ather
,
this work hopes to identify a
common link between these failure modes and
to test a
surface enginee
ring process that has the potential to significantly reduce wear in these
systems.



16

2.

THEORETICAL BACKGROUND


2.1

The
Hydrogen
-
Assisted

Wear

Mechanism

One theory to explain the
strange wear phenomenon that exists in titanium
-
polymer tribosystems was proposed by
t
hree researchers
in 2001
during an experimental
investigation of the wear performance of 6Al
-
4V titanium coupled with UHMW
polyethylene
[
8
]
.
Their work originally initiated
in 1999 [
7
]
as a study of various
titanium surface engineering processes that
migh
t

be used to reduce the
wear rate of
UHMW polyethylene when operated against titanium materials.

This is a common
material combination in medical total joint replacement prosthesis applications and
premature wear of the
plastic component in these systems

is a constant concern. Using
standard pin
-
on
-
disk wear testing methods, H. Dong and his associates were able to
successfully demonstrate significant gains in the friction and wear properties of UHMW
coupled with Ti64 by introducing different coatings and

surface engineering techniques
;
specifically, diamond
-
like coatings

(DLC)
, thermal
-
oxidation (TO) and oxygen di
ffusion
(OD) surface treatments
.


During the course of their testing however, these researchers encountered an
unexpected wear phenomenon

which
they initially found difficult to explain. To
establish a
control
baseline for comparison

against each of the individual samples
, wear
testing was first performed
using
an
un
-
treated

6Al
-
4V titanium

test disk (
dynamically
tested against a
UHMW
polyethylen
e test pin). After completing this baseline test, they
discovered that the titanium test surface had sustained significant wear distress in the
form of deep, wide grooving (scarring) while the much softer
plastic
test pin
maintained

a relative
ly

low wear
rate.
During subsequent testing with the
modified test samples
(
DLC, TO and OD
)
, none of the surface
-
modified titanium components exhibited any
wear.


Initially, these researchers attempted to

explain
this
wear
phenomenon
as
characterized by a three
-
body
abrasive wear mechanism, perpetuated by a
poor

ionic


17

attraction between the titanium substrate and the thin oxide layer that forms naturally
on
titanium
surfaces that are exposed to air
. They suggested that this natural surface oxide
layer (very thin, typ
ically 5
-
10 nm thick)
became
liberated during operation against the
UHMW test pin and particles were subsequently trapped at the test in
terface in the
polymer material,
generating
the
abrasive wear mechanism. Following an additional two
years of wear test
ing with Ti64
-
UHMW material couples,
this team
found new evidence
to support
a
different type of
theory
for this wear condition;
hydrogen
-
assisted

wear
.


The hydrogen
-
assisted wear mechanism
appears to
originally
have been
proposed
by a team of Russian
researchers in the
late
19
7
0’s
[
13
]
,
and then with the first English
publication in 1993 by A. L. Zaitsev [
11
] while examining
the trib
ological performance
of
several
polymeric materials

coupled with
tungsten carbide
-
cobalt coatings
.
During
testing, Zaits
ev noticed a trend wherein hard alloys
(primarily tungsten and cobalt
-
based
materials)
experienced higher than expected wear rates when coupled in dynamic contact
with polymers such as

polycaproamide (PCA),
phenol formaldehyde

polymer (PF) and
EDC
e
poxy co
mpound
. All of these materials

are found to
contain reactive
chemical
groups such as amide, ether and hydroxyl groups

in their polymer structure.


The theory of hydrogen
-
assisted

wear
that
was proposed by Zaitsev and
then
advanced considerably by
Dong [
8
] attempts to explain this
abnormal

metal
-
polymer
wear
relationship
as
being
a function
of the reactive chain groups

found in these
polymers
. During
the
dry sliding
action
between
the
se

material
s

and the
ir

metal
lic

counterparts
under load, a significant
amount of dynamic energy is dissipated in the
form of heat. Dong and Zaitsev suggest
ed

that this intense localized heat, coupled with
the plastic def
ormation and strains that occur

at the test interface are enough to
cause
a
breakdown of the
hydrogen
-
cont
aining
amide

and
hydroxyl groups

found in the
se

engineered plastics
.
During continued dynamic operation, the high heat retained at the
test interface aids the inward diffusion of
this
excess hydrogen into the
(titanium) bulk
material substrate.





18

Titani
um
is a material that is
known to be susceptible to
hydrogen
-
embrittlement
,
a phenomenon
that reduces the toughness
(general resistance to cracking) and ductility
of a material
. The natural propensity of hydrogen atoms for titanium atoms is
also
a
driving

function in this relationship; the hydrogen atoms tend to work their way into
the
titanium bulk and
collect

at
grain boundaries. As these hydrogen atoms accumulate at
these locations and subsequently
develop into
titanium hydrides, their growth causes
su
bsurface pitting and microcracks in the titanium bulk

material
. Subsequently, the
interfacial shear forces generated by the sliding action of the plastic test material across
the outer surface of the titanium test article subsurface fracture of the titani
um material
(liberation of wear debris

particles
).

Figure
4

provides a graphical illustration of the
hydrogen
-
assisted

wear mechanism as proposed by these researchers.



Figure 4.

Illustration of Proposed Hydrogen
-
Assisted Wear Mechanism for the
Tribosystem UH
MW/Ti6Al4V [
8
]


In an effort to
validate

this theory,
Dr.
Hanshan Dong and his associates
[
8
]
conducted a series of dyna
mic
pin
-
on
-
disk
wear tests using the Ti64/
UHMW material
couple
, and set to work characterizing the
associated
wear mechanisms by analyzi
ng the
morphology of
both
the
titanium test
surfaces
and
UHMW plastic countersurfaces
.
Testing was conducted using standard
unmodified

6Al
-
4V mill product (hot
-
rolled, mill
Hydrogen Diffusion

Hydride Formation

Hydride Growth and
Surface Cracking

Third Body (Abrasive) Wear
Mechanism

-

Wear Part
icles
Embedded in Mating Surface



19

annealed bar) to establish the base wear relationship with UHMW and testing was al
so
conducted with Ti64
samples
that had been subjected to a unique thermal oxidation
process
. Additionally, annealed 316L stainless steel samples were subjected to the same
wear test so that a comparison could be made between the Ti
tanium
/UHMW couple to
t
hat with a stainless steel product
of
lower
material
hardness.
All testing was conducted
using water
-
lubricated operating conditions
(simulating periprosthetic applications)
and
Figure
5

below provides a summary of the
time that
was
required to generate a

certain
level of wear in each sample (termed
I
ncubation
T
ime
;

a

longer
incubation time
indicates
a lower material wear rate).

Correlating incubation time with specific wear (of the
metallic mating surfaces), it can be seen that the wear rate of the
as
-
po
lished

titanium
test surface was significantly higher than the 316L test surface (approx 15x
-
20x higher)
even though the hardness of the stainless material was about 2/3 that of the titanium test
material


Figure 5.

Chart Comparing Wear Rates (as a function of Incu
bation Time) for
Various Metal
-
Polymer Tribosystems [7]


The
se results

seem to agree well with the results obtained by other researchers

performing similar wear tests with titanium alloys
; excessive wear damage sustained by
the titanium test surface during

dynamic operation against a much softer engineered
plastic

material
. After
completing these tests
,
Dong (et al.) performed an
extensive
a
nalysis
for both
the
un
-
modified and thermally
-
oxidized
samples

using
transmission


20

electron microscopy (TEM)
to inspe
ct
the test
surface morphology
.

G
low discharge
spectroscopy (GDS)
was also used to analyze the depth profile
of elements
in the wear
track post
-
test.


During the course of their analysis, these researchers discovered
that
a significant
concentration of t
it
anium hydrides (
TiH
X
)

were present
in
the
outer
titanium oxide layer
and more importantly
,

in
the α
-
phase
titanium
bulk



hydrides that did not exist prior to
testing
. The
formation
of
titanium hydrides
below

the direct interface layer indicates
that the
diffusion of hydrogen
products
into the titanium bulk could well have
contributed to the deep abrasive wear
that was
experienced during
these
test
s
. This
condition was
detected
in the
un
-
modified

titanium
test
sample
s

as well as the
thermally
-
oxidized and

oxygen
-
diffused
test
sample
s
.
However, t
he percent composition of
TiH
X

found in the α
-
Ti
bulk
of the
thermally
-
oxidized samples
was
significantly lower

than
the percentage that was detected
in the
un
-
modified
titanium sample
s
. This stands to
reason as testing has
demonstrated
that
the
naturally
-
occurring
outer
titanium
surfa
ce
oxide
layer
serves as a
fairly
effective barrier against hydrogen diffusion
(reduces
the
diffusivity of hydrogen into titanium)
[1
2
].



It was b
ased on these findings

that Dr. Hanshan Dong

(et al.)
proposed
the theory of
hydrogen
-
assisted

wear
as pote
ntially being the driving function behind this
abnormal

titanium
-
polymer wear relationship.

The presence of subsurface
titanium
hydrides

in
these tested articles appears to be a strong indicator that hydrogen has a significant
influence on the wear of tit
anium alloys by softer plastic materials.
T
his theory was
established using wear results
generated
with the 6Al
-
4V titanium

UHMW polyethylene
tribosystem,
however
it also pertains to
tribosystems involving titanium
coupled with
other engineered plastics c
ontaining certain polymer chain groups.






21

2.2

Thermal Oxidation
(TO) Treatment
of Titanium

Inherently, t
itanium materials
have excellent corrosion
-
resistant properties and are
not easily compromised by attack from
acidic media
. These protective properties ar
e
afforded
by
the presence of
a

thin

oxide layer that forms naturally
on
the surface of
titanium
materials
over time as
they are
exposed to
an
oxygen
-
containing atmosphere
(air). This oxide layer is typically very
stable and
adherent
,

but also
is commonly

very
thin. After
long
-
time
exposure
of newly
-
machined titanium to a typical
air

atmosphere
,
this outer oxide layer
will only grow
to be about 3
-
4 nm
(.1
μ
-
in.)
thick [
7
]. Many
researchers believe that the poor wear relationships exhibited by titanium tribosystems is
primarily attributed to
the removal of this thin oxide layer by the dynamic interaction
(relative sliding motion) with a mating surface

and
a
lso
,

that a more robust boundary
layer
applied to
this interface could help to improve this condition.


Dr. Hanshan Dong
demonstrated
in
his research
efforts
[
7
,
8
]

that
the application
of

t
hermal
o
xidation

treatments to titanium
-
base alloys can
significan
tly

improve the
tribological
properties of
these materials when
dynamically
coupled with
UHMW
polyethylene
.
Th
e

wear phenomenon
that affects this material couple
however, is
also
known to occur in other titanium
-
polymer tribosystems

[
5
]
; systems involving

engineered plastics

containing
similar reactive
polymer
end groups. Based on the theory
of Hydrogen
-
Assisted wear, the application of titanium TO treatments should also be
effective in mitigating wear in these dynamic systems.

(
…and the development of
di
fferent titanium TO treatments is the focus of this project…)


The thermal oxidation
process
is
performed
much like the name implies, titanium
articles are subjected to
a
heat treatment
process
in an oxidizing atmosphere. The
titanium
responds to this env
ironment by
developing
an oxygen
-
rich, cas
e hardened
outer surface layer which is
often referred to as
alpha case
.
This
alpha case
layer
generally
consists of two
modified regions

or layers
:



-




22

-

Oxide Surface Layer

-

The outermost layer is
typically very t
hin (usually on the order of a few
micrometers in thickness) and consists primarily
of
TiO
2

(titanium dioxide). This usually
manifests as
Rutile,

the most compact and stable
of the TiO
2

polymorphs (the other two being
Anatase

and

Brookite
) and

its presenc
e on the
surface of the titanium article significantly
improves corrosion resistance of the titanium
article. The relative hardness of this outermost
layer is also much higher than that of the titanium
bulk (TiO
2

being a ceramic in its basic form).


Figure 6.


Th
e Unit Cell of
Rutile TiO
2

(Titanium atoms are gre
y

and oxygen atoms are red). [
29
]



-

Oxygen
-
Diffused Zone



During the heat treatment process, oxygen also tends to
diffuse
downward
into the bulk titanium
, forming an oxygen
-
diffused (TiO) region
direct
ly below the outer oxide layer.
This TiO region
is comprised primarily of
α
-
phase
(hexagonal close
-
packed)
titanium
microstructure due to the alpha
-
stabilizing nature of oxygen in a titanium matrix (hence the designation
alpha
-
case
).

The solid
-
solution h
ardening effect of this oxygen diffusion process is also
considerable. A microhardness check performed on a thermally
-
oxidized titanium
samples will often show an increase in relative hardness of
around
2
-
3 times that of
the titanium bulk material.

Depen
ding on thermal atmosphere conditions, this
oxygen
-
diffused region can extend anywhere from 2
-
300+
μ
m (.00001
-
.012 in.)
down into the titanium.



23


Figure 7.

Cross
-
Sectional Micrograph (a) Showing Alpha Case Layer Detail and

Plot (b) showing Knoop Microhardness for a Thermally
-
Oxidized Ti64 Sample
[
27
]


The
physical
properties

of th
is

alpha case layer
,
as well as
the respective
depth of
penetration and
microstructural
composition are all highly influenced by the thermal and
atmospheric controls employed during thermal oxidation treatments

and a wide range of
results can be achieved by
varying

these parameters.
For

example, the relationship
between heat
-
treatment temperature and resultant depth of penetration do not relate
linearly; as heat treatment temperature is increased, an exponential increase in the actual
depth of alpha case is realized.
This is a function
of the effect that temperature has on
the
diffusivity of oxygen in
to

titanium and is also influenced by the
b
eta
-
transus

temperature of the material. The
b
eta
-
transus is the temperature at which the
microstructure of titanium converts from an
a
lpha
-
phase
(
h
exagonal
c
lose
-
p
acked)
microstructure to a
b
eta
-
phase (
b
ody
-
c
entered
c
ubic) microstructure. The diffusivity of
oxygen into a fully
b
eta material at temperature is higher
than in an
a
lpha microstructure
.
The beta
-
transus for commercially pure titanium i
s around 880 °C (1615 °F) and is
different for the various alloys. Figure
8

shows this relationship between depth of case
and heat
-
treatment temperature.


The majority of thermal oxidation treatments are conducted at a
heat
-
treatment
temperature
below the

beta
-
transus for that
specific
alloy to reduce the impact on the
general mechanical properties of the titanium. Processing titanium alloys for extended


24

lengths of time above the beta transus can sometimes cause excessive precipitation of
alloying element
s out of the solid solution, reducing the strength and toughness of the
material. Additionally, crossing the beta transus temperature initiates the transformation
process of the titanium matrix from alpha to beta
phase
and the
transition back to alpha
pha
se (cooling) has the potential to result in an undesirable grain structure or other
anomalies such as retained beta / beta fleck (conditions that are known to affect material
toughness).


Figure 8.

Alpha Case Depth as a Function of Heat
-
Treatment Temperature for
Commercially
-
Pure (CP) Titanium
, Held at Temperature for 1
-
2 Hrs

[
4
]


The hardness
-
depth
profile in
the oxygen
-
diffused zone
of the alpha case layer
is
primarily attributed
the relative density of oxygen atoms in the titanium solid solution
(increasing de
nsity correlates to increasing
hardness). In addition to the heat
-
treatment
temperature, the actual time held at temperature, the oxygen content of the furnace
atmosphere and the method of cooling all have an effect on the resulting hardness
gradient in t
his region.

With respect to the dwell time (time held at temperature),
Figure
9

shows a
hardness
-
depth
profile
for two 6Al
-
4V titanium samples that have been
subjected to a heat treatment process at 850 °C in an air atmosphere for different lengths
of ti
me.



25


Figure 9.

Microhardness
-
Depth Profile for Two 6Al
-
4V Titanium Articles Subjected
to Heat Treatment in
an
Air
Atmosphere
at
850°C for Different Lengths of Time
[
10
]


Significant work has been accomplished in recent years
[XX
-
XX GROUP]
to
increase the understand
ing of titanium thermal oxidation treatments and the
tribological
improvements that can be achieved by employing these surface engineering techniques
in specific operating environments. It is clear from the wide range of results obtained
though experiment
al testing that the

however, the accurate selection of process parameters will make or break



titanium interface surfaces. A wide range of properties have been…

interface surfaces by carefully controlling the …..XXXXXX



26

3.

METHODOLOGY / APPROACH


The theory

of ‘hydrogen
-
assisted’ wear that was proposed by Dr. Hanshan Dong
does much to explain the wear phenomemon behind
the titanium
-
polymer tribosystem
and his work with
UHMW
[
8
]

offers some very conclusive findings. However, little
-
t
o
-
no research has been pu
blished involving other titanium
-
polymer tribosystems [
5
] that
have
also
been known to develop the
se

abnormal wear patterns.
Two engineered
polymers
that have also been known to cause excessive damage to mating titanium
surfaces during dynamic interaction

are acetal
copolymer
and epoxy compound. The
latter is frequently employed as a bearing polymer in a composite form (combined with
various solution
-
strengthening and friction
-
reducing fillers such as PTFE)

while
acetal
copolymer

is often used in its un
-
m
odified
form
as a bearing plastic.


It is the goal of this project to experimentally evaluate the tribological
improvements that can be achieved with the Ti64
-
acetal and Ti64

epoxy tribosystems
by employing various thermal oxidation techniques.
Using an A
STM standard dynamic
wear testing approach, this project will attempt to characterize the wear and friction
properties of these polymer materials in addition to UHMW polyethylene (basis for
comparison against historical research) as operated against
titani
um samples
that have
been
subjected to four
unique
thermal oxidation processes
. To establish a control
baseline for comparison, each of these polymer samples will also be tested against

an
unmodified 6Al
-
4V titanium test sample.


Standard metallurgical ch
aracterization techniques are used to determine the
properties of the alpha case product
developed during
each thermal oxidation process
. It
should be noted, however that the equipment required to detect the presence of titanium
hydrides in
the tested tit
anium surface are not available and this project is strictly






27

3.1

Thermal Oxidation
(TO)
Surface Engineering

The range of case properties and titanium microstructures that can be developed by
varying the control parameters applied
in

a thermal oxidation trea
tment
is

extensive.
A
few studies published in recent years [
X,X
] have attempted to characterize the wear
characteristics of
an array
of
different
titanium TO
-
treated samples,
however these
studies were conducted
using a 6Al
-
4V substrate and different ste
el / ceramic mating
materials. The titanium
-
polymer tribosystem is inherently different and it is difficult to
say whether the
tribological
improvements witnessed using different steel and ceramic
countersurfaces
would translate over to polymer tribosyste
ms. Each system is unique

and should really be considered on a case
-
by
-
case basis.


The ultimate goal of this project is to establish a baseline characteristic wear
relationship for three different titanium
-
polymer tribosystems and then to evaluate the
improvements in wear and friction that
might
be achieved by applying titanium thermal
oxidation treatments. Four different thermal oxidation treatments
were selected for this
project based on a number of criteria. In each case, consideration was provided

to:

-

Case microhardness profile

-

Depth of penetration

-

Composition of the outermost
boundary
surface layer

-

Adhesion between
surface
layer and oxide layer

-

Resultant grain structure of the Ti bulk

-

Ratio of
boundary
surface layer
thickness to OD zone


These ca
se properties were all considered while developing the thermal control
parameters for each process while also keeping an eye on the impact that each has from
a manufacturing ‘producibility’ standpoint. For example, parameters such as excessive
time held a
t temperature could potentially pose a problem if these TO treatments were
applied in a full
-
scale production environment.






28

Thermal Oxidation Treatment
-

Case #1

Heat Treat Temperature:


850

°
C

(1
562

°F)

Furnace Atmosphere:



Air

Time at Temperature:



2

Hrs

Cooling:





Air cooled


Case #1 would be considered the most straightforward TO treatment solution from a
manufacturing standpoint; short time at temperature, no furnace atmospheric controls
and air cooling. This treatment
has shown favorable result
s in the thermal oxidation
efforts performed by Dr. Jun Qu and Dr. Peter
Blau [
21
,
23, 27
].

Using this process,
they successfully demonstrated that the wear resistance of Ti64 could be increased
significantly in different steel
-
titanium systems using both

lubricated and un
-
lubricated
operating conditions.


Thermal Oxidation Treatment
-

Case #2


Heat Treat Temperature:


850 °C (1
562

°F)


Furnace Atmosphere:



Air


Time at Temperature:



5 Hrs


Cooling:




Furnace cooled


Case #2 uses the same heat treatment

temperature and atmosphere that was used in Case
1 but extends the oxidation time and more importantly, changes the method of cooling.
Research has shown [
20
] that cooling the thermally
-
oxidized samples in a furnace as
opposed to an air
-
cool method can s
ignificantly improve the mechanical properties of
the case layer as well as the depth of penetration. The strength of adhesion between the
oxide layer and the substrate is also improved significantly which most often translates
to improved wear resistance
.



29



Figure 10.


A Comparison Showing the Variatio
n in Case Thickness for Furnace
-
Cooled samples and Air
-
Cooled samples as a Function of Treatment Time (A) and
Temperature (B).
Figures
(C) and (D)
Show
SEM Image of Oxide Layer Cross
-
Section for Air
-
Cooled Sample a
nd Furnace
-
Cooled Sample (respectively)
[
20
]


Thermal Oxidation Treatment
-

Case #3

This is a 2
-
stage thermal treatment; the parts are initially treated in a standard air
atmosphere for the stated time and then in a vacuum for a given time, followed by
fur
nace cooling (under vacuum).
Number
(1)

is standard air atmosphere,
number
(2)
indicates
vacuum atmosphere


(1)

Heat Treat Temperature:



850 °C (1
562

°F)


(1)

Furnace Atmosphere:




Air


(1)

Time at Temperature:




25 minutes


(2)

Heat Treat Temperatu
re:



850 °C (1
562

°F)


(2)

Furnace Atmosphere:




Vacuum


(2)

Time at Temperature:




20 Hrs


(2)

Cooling:





Furnace cool under vacuum


Case #3 represents a very different type of thermal oxidation treatment as compared
against the other
‘traditional
’ TO
processes evaluated in this project.
This thermal cycle


30

employs a
unique
2
-
stage process
that
is designed to develop
a very deep and adherent
oxygen
-
diffused case
layer
while minimizing the resultant outer surface TiO
2

oxide
layer. The first stage o
f this process (850 °C / air) initiates the formation of the TiO
2

outer surface oxide layer much like the other processes tested in this project
. H
owever,
by subsequently subjecting the oxidized samples to continued heat treatment in a
vacuum atmosphere (
second stage), a different type of diffusion mechanism is initiated
which forces the oxygen deeper into the substrate. The low partial
-
pressure of oxygen in
the vacuum atmosphere most likely causes
the following reactions:

Ti(s)

+ O
2
(g)

= TiO
2
(s)

(1)

TiO
2
(s)

+ T
i(s)

=
2O (dissolved in Ti) (2)

which helps to explain the dissolution of the surface oxide layer (
that was
generated
during the first stage) and the resultant increase in absorption of oxyg
en into the
titanium.


Using this

unique thermal treatme
nt process, researchers have successfully
demonstrated [
10
] that
oxygen
-
diffused
case
depths
of
300+
μ
m (.012”) can be achieved
(as compared to around 100
-
120
μ
m using ‘traditional’ TO treatments). Based on the
theory of hydrogen
-
assisted wear in titanium
-
polymer tribosystems, it is believed that
the depth of the oxygen
-
diffused layer could directly be related
to the wear resistance of
the sample
.


Thermal Oxidation Treatment
-

Case #4


Heat Treat Temperature:


650

°C (1202 °F)


Furnace Atmosphere:



Air


Time at Temperature:



48 Hrs


Cooling:




Furnace cooled


The goal of thermal oxidation case #4
was to focus more on the tribological properties of
the outer TiO
2

oxide scale,
while
placing less emphasis on the depth

and hardness of the
TiO oxygen
-
diffuse
d zone. This treatment takes place at a much lower temperature (650
°C) than the other three treatments and accordingly, requires a much longer time at
temperature to develop the desired case properties. Experimental
research
performed by


31

S. Kumar et al.

[
30]

using 6Al
-
4V titanium
suggests
that a dense TiO
2

layer of
exceptional superficial hardness (higher than reported for most other TO treatments) is
possible. Kumar’s research however, focused on the microstructural and
electrochemical properties of va
rious TO treatments and less on tribological properties
(no wear performance data reported).


3.2

Polymeric Test
Materials

As noted above,
the focus of this project is
on the tribological properties of surface
-
modified titanium
coupled with
three different
e
ngineered plastic

materials commonly
used in tribological applications:

-

Ultra High Molecular Weight (UHMW) Polyethylene

-

Acetal Copolymer Resin

-

Epoxy Compound


In their engineered forms, each of these polymer materials
can be
successfully
employed
as a low
-
friction barrier against wear between metallic components, however
each is also known to exhibit a

strange
wear relationship with titanium alloys (discussed
in Section 1.2). It has been proposed
[
8
]
that this titanium wear phenomenon
(hard
alloys damaged
by operation against soft polymers)
is directly related to a hydrogen
-
embrittlement effect in the titanium that is caused by the breakdown of certain reactive
groups found in
various

polymer materials; the so
-
called
hydrogen
-
assisted

wear
mechanism
.


To da
te,
the
only
tribosystem to be tested and analyzed for metallurgical evidence to
directly support this theory is the UHMW
-
Ti64 tribosystem. Neither acetal copolymer
or epoxy compound have been subjected to the
regimen of wear tests and subsequent
analysis

that UHMW has seen,
however
both

of these materials contain

polymer chain
groups
that were identified by researcher H. Dong [
8
] and A. L. Zaitsev [
11
] as
potentially being drivers in the hydrogen
-
assisted wear mechanism.
Dong and Zaitsev
have suggested t
hat the destruction of
amide
,
ether

or
hydroxyl

groups found in the


32

polymer material ultimately results in
the reduction of
toughness
in the titanium
material.


Acetal
c
opolymer
is
a linear (unbranching) polymer that is formed by the
copolymerization of tr
ioxane with small amounts of comonomer. The resulting
oxyethylene groups (
ether

group family) are formed by random carbon
-
carbon links
throughout the polymer chain (Figure
11
). These C
-
C bonds are the link that provide
s

this polymer with a
relative
ly sta
ble
chain structure. However, once subjected to
degradative conditions such as high strain, excessive heat (etc.), th
e
chain will
depolymerize

(breakdown)

or ‘unzip’


releasing H
-
C
-
H alkanes
until it hits a C
-
C bond
[
16
]
.
The mechanics of th
e

depolymeri
zation process associated with acetal copolymer chain
therefore appear to meet the requirements proposed in the hydrogen
-
assisted wear model.


Figure 11.


The Basic
Chemical
Structure of the Acetal Copolymer Chain
[
16
]



Epoxy is a thermosetting resin that is forme
d by the copolymerization of two
primary components, an epoxide chain (epoxy resin component) and a polyamine
monomer ‘hardener’ component. The polymerization or ‘curing’ process of epoxy
is
characterized by
extensive polymer cross
-
linking and the end res
ult is a fairly complex
chain structure.

There are quite a few epoxy resins in use today, all with varying chain
structures evolved through different cross
-
linking systems. All of these compounds
however, appear to have a common trait, hydroxide groups e
xist in the base polymer
chain, connected by a single carbon
-
oxygen bond.



33


Figure 12.


The General Chemical Structure of Epoxy
[
17
]


The cross
-
linking action inherent to the curing process of epoxy compound is
effective in impeding the depolymerization of most ep
oxy chains, however the
hydrogen
-
containing hydroxide groups are typically released during this process [
14
,

17
]. This again fits with the hydrogen
-
assisted wear model as proposed by Dong and
Zaitsev with respect to the titanium
-
epoxy tribosystem.


3.3

Wear
Testing (Pin
-
on
-
Disk Method)

One of the most widely accepted methods for evaluating
the
wear
and friction
properties
of
different
material

couples
is the
Pin
-
on
-
Disk

test.

Just as the name implies,
this test method consists of a test ‘disk’ component and
a test ‘pin’
, oriented such that
the axis of the pin is oriented perpendicular to the flat surface of the disk (and off
-
center). With a
constant
load applied to the pin (along its axis), the
wear
test is then
conducted by rotating the disk about
its
cente
r

(while the pin
remains static
). The test
continues until the pin has completed a
pre
-
defined

number of cycles or ‘distance
traveled’ in relative motion.





Figure 13.

General Configuration of the Pin on Disk Wear Test

Test Disk

(Unmodified 6Al
-
4V Titanium)

(
Thermally
-
Oxidized Titanium)

Test Pin

(UHMW Polyethylene)

(Acetal Copolymer)

(Epoxy Compound)

Wear Track

(Volume Loss)

Disk is rotated about its center
causing circular wear path

Load applied to test pin

(Axial direction)

V(s)

F
N



34


As the test disk cycles through revolutio
ns and the test pin continues to travel over
the same circular path under load, the test samples (
either the
pin,
the
disk or both based
on
the
materials bein
g tested) will accumulate wear which is reflected in the form of lost
material (
volume loss). Aft
er the test is completed (total required distance traveled

is
achieved
), both the test disk and the test pin are measured to quantify the amount of
material that has worn away. In terms of the test disk component, the
total
wear can be
calculated by
physi
cally
measuring the width and depth of the resultant
wear scar
.
Often, a profilometer is required to measure the depth due to the rough, non
-
uniform
profile at
its root
. After testing, wear scars are also commonly very shallow and difficult
to record usi
ng standard measurement tools (drop gauge, etc.). Lost volume in the test
pin component is more easily calculated using standard measurement tools (calipers,
micrometer).

Alternatively, the lost material volume in each completed test sample can be
verifie
d by first cleaning, and then weighing the samples using precision scales. The
rate of wear in each test component is then typically reported as a function of distance
traveled and effective pressure. This is commonly referred to as the
Specific Wear Rat
e

for that material

or the
Wear Coefficient
.

According to Bhushan [
1
]

the wear coefficient
(K) is calculated as:


With wear rates commonly specified in units of (mm
3
/N

m) or (in
3
/lbf

ft).


This
Pin
-
on
-
Disk test
method was
determined to be
the most suitab
le
approach
for
this project based on a number of
criteria
:

Time
Required to Test


The standard testing parameters associated with POD
tests allows for quick turnover. Also, the relatively simple configuration of
the test samples means less time required

for manufacture and setups are very
quick.



35

Cost


The overall size of the required test samples is small; geometry is simple
and easy to manufacture.

Reliability


The straightforward configuration of this test
limits the potential for
error caused by e
xternal and system
-
related variables

Similarity


This also appears to be the method that was selected for each test
case in all of the research papers published on titanium thermal oxidation.
Maintaining similar operating parameters allows for direct com
parison of test
results.


It should also be noted that a Pin
-
on
-
Disk tribometer
(test apparatus)
was not
readily available and it was necessary to design and manufacture
one prior to testing
.
The ability to manufacture a tribometer in a short
amount of
ti
me that would provide
reliable and repeatable results with data acquisition capabilities also factored heavily in
the test selection process. The
design of this test apparatus is discussed in detail in
Section 4.




36

4.

PIN ON DISK TRIBOMETER

The wear testing
e
lement

of this project was central to verifying and validating the
impact that titanium
TO
treatments have on the tribological properties of
t
i
tanium
-
polymer tribosystems.
As noted in Section 3.3, the ASTM standard
Pin
-
on
-
Disk

test
method was
identified
a
s the most appropriate
approach to
characterize
relative wear and
friction properties
, however a Pin on Disk testing apparatus was not readily available.
A
number of
commercial
Tribometer systems were evaluated, however all were found to
be cost prohibiti
ve (systems such as the Falex ISC cost $40K+)
.

I
t was therefore necessary to
design and manufacture
a

cost
-
effective
Pin
-
on
-
Disk
testing apparatus from scratch; one
that was capable of
producing
the desired dynamic
interface and also
provided the
data a
cquisition capabilities necessary to
record
dynamic
friction
data.
Figure
14

d
epicts the test rig which was completely designed,
manufactured and tested during the first six weeks of this project.

The following
sections
briefly describe the design approa
ch,
features and capabilities of this Pin on
Disk testing apparatus that was ultimately used t
o conduct all of the wear testing
outlined in this paper.





Figure 14.

Pin on Disk Tribometer




37

4.1

D
esign and Configuration

The apparatus was originally designed using the AS
TM standard for Pin on Disk
testing
(ASTM G99)
as a guideline for functional requirements.
This specification is
fairly broad concerning the actual configuration of the apparatus, constraining only the
system level requirements (i.e. specific pin
-
to
-
disk
interface and alignment requirements,
relative motion, etc.). The actual method of test specimen constraint, load application

and power transmission are all left to the discretion of the designer so initial efforts
were
focused on
a cost
-
effective solutio
n that would provide consistent and reliable results.

As directly specified in
ASTM G99 Section 3.1,

the Pin on Disk
wear
test consists
of two test specimens
;

a pin with a radius
e
d
or flat
tip

which
is positioned perpendicular
to the other

specimen
, usuall
y a flat circular disk. The test machine causes either the disk
specimen or the pin specimen to revolve about the disk center. In either case, the sliding
path is a circle on the disk surface. The plane of the disk may be oriented either
horizontally or ve
rtically. The pin specimen is pressed against the disk at a specified
load
,

often
by means of an arm or lever
.


Load is then applied to the pin specimen
(normal to the face of the disk) and maintained throughout the test.

The test apparatus designed for
this project is configured with a load arm / disk
spindle
-
type configuration. The test disk component is fastened directly to
a cylindrical

spindle


using a shoulder screw to ensure accurate

alignment and concentricity between
t
he test disk and the spind
le. The spindle detail is
machined from 17
-
4 PH stainless steel
(condition H1150) and is
supported and constrained in X, Y and

Z directions by two
ABEC1 sealed ball bearings. These bearing
s

provide the supporting reaction force
against thrust loads appl
ied normal to the disk test surface and also allow rotation of the
spindle/disk assembly about the axis of the spindle
. Ball bearings were selected over
roller thrust bearings for this application as they introduce less frictional drag force into
the syst
em (the calculated
increase in effective
drag force
is
less than 1% of
the actual
test

force in the drag force


within the limits of the strain gauge measurement error).
Future testing with this apparatus could potentially incorporate fluid contaminants
so
sealed bearing units were selected (also effective in preventing the ingress of pin wear
debris).



38

The entire test disk/spindle assembly is driven
by
a DC gearmotor
, mounted
horizontally
to
the bottom of the upper mounting plate. A 90° miter gear

syste
m is used
to
transmit the power directly from the gearmotor to the spindle which provides the
relative motion between the (rotating) test disk and the static test pin (loaded from
above).
The miter gear mounted on the
output
shaft of the ge
a
rmotor is manu
factured
from case
-
hardened alloy steel while the mating gear (mounted to the lower shouldered
end of the spindle) is manufactured from a brass alloy. The assembly was originally
designed with a small wet sump setup below the
miter gears but upon operatio
n, it was
discovered that the higher operating speeds would cause the lubricating oil to migrate
out of the sump along the shaft of the gearmotor. It was decided that conventional shaft
seals would add to much drag to the shaft of the gearmotor so the sum
p was eventually
removed and a higher viscosity gear oil was used to lubricate the gear interface
(
intermittent

application).


Figure 15.


Pin on Disk Tribometer Cross
-
Section View




39

The
Test

Pin

specimens

are contained in an
U
pper
L
oad
A
rm

assembly
,

consisting
of a
number of components which are all illustrated in
Figure XX
. In
i
tially, the test pins
are installed in a ’pin holder’ component (blind hole, slip fit installation)

and retained
with a set screw

mounted transversely
. The pin holder is pinned at two locati
ons in a
four
-
bar type linkage configuration which is used to ensure constant vertical alignment
of the test pin. The
pin
holder is pinned at the uppermost location to a
control arm

assembl
7
y
, consisting of five individual components. The control arm ass
embly utilizes
a RH/LH threaded rod with a threaded clevis pin at one end (mates with the pin holder)
and a female spherical bearing at the opposite end. By rotating the threaded rod, the
operator changes the effective length of the control arm assembly,
which in turn changes
the angle of incidence between the pin holder and the mating disk surface. Once the
operator has dialed in the correct length of the control arm and confirmed absolute
vertical alignment of the pin holder, the length is locked down u
sing LH/RH jam nuts.
This configuration guarantees that as the pin holder rotates about the pinned center of the
Load Arm

component (to which it is also attached), the test pin will always remain
vertical and normal to the test surface.

During the initial

design phase, this use of a four
-
bar type control arm linkage was
deemed critical to the performance of the apparatus provided the materials
that were
being tested. Most often, the materials being characterized with Pin on Disk tests are
metal/metal or m
etal/ceramic couples and the volume loss of the pin specimen is
typically very low. In this test case, the hardness disparity between the titanium disks
and the polymer test pins is significant so the potential for wear in the
softer
polymer pin
is high.

As the height of the test pin changes throughout the test due to wear, the angle
of incidence between the pin and the disk test surface must remain zero or the loading
profile at the test interface will consistently change (uneven bearing pressure gradien
t).



The
Load Arm

component is
central to the operation of the entire test assembly.
The load arm is pinned at two locations
, the outermost end is machined
with
a
close
-
tolerance
clevis/thru hole configuration
to allow installation of the pin holder and

re
tention with a spring pin. A
transverse
thru hole
located
approx.
1.625” from the
opposite end acts as the center of articulation for the load arm (main
Pivot Pin

installed
through this hole).
A larger vertical thru hole is located outboard of the pi
vot pin


40

location; this hole allows sufficient clearance for a threaded ‘T’ fastener which centers
through a compression spring.



Figure 16.


Pin on Disk Load Arm Component


While
evaluating

all of

the commercially available Pin
-
on
-
Disk tribometer
solutions
on the
market, it was noticed that in all cases, the normal force test load
limit
was
relatively low (
max load
ranging 10N to 60N). Given the range of polymer products that
will potentially be evaluated using this rig in the future, it was desired that this
rig
be
capable of applying F
N

test forces ranging 1
-
5
0+ lbf (220+ N). This capability would
allow for testing
of

a standard .125” (flat end) pin
at effective bearing pressures up to
and exceeding 4000 psi. It was this requirement that ultimately led to the s
election of a
compression spring
-
driven load design. Other load application
methods
that were
considered during the design phase were a calibrated (slung) deadweight configuration
and also the use of hydraulic/pneumatic actuators, however the spring desig
n
offered

the
best combination of
adaptability,
control

and reliability at a significantly reduced cost.
This test assembly uses a floating ‘seat’ component made of 7075
-
T6511 aluminum
upon which the compression spring rests. A T
-
bolt runs through the ce
nter of the
compression spring (cut and ground ends). As the operator tightens the upper flange nut,
the T bolt compresses the load spring which in turn imparts an upward force on the end
of the load arm component. The upward force on the load arm create
s the F
N

test load
on the test pin which is installed at the opposite end of the load arm. A close tolerance


41

RC
fit between the
Pivot Pin and the Load arm ensures that the test pin remains static
relative to the rotating test disk. Dry film lubricants ar
e used to mitigate wear between
the Pivot Pin
(manufactured from 17
-
4PH stainless steel)
and the Load Arm while a
laminate wear pad (thermoset
ting resin
/PTFE/Polyester fabric) is bonded to the upper
Load Plate to prevent premature wear at the
contact
inter
face
when the
Flange Nut

is
tightened
.


The inverse mechanical advantage of the load arm (4.5:1 distance from fulcrum to
test pin and applied load respectively)
i
s used to ensure a consistent loading profile
throughout the wear regime
. As the pin wears
, t
he vertical drop has little e
ffect on the
magnitude of the applied load (
full pin sample wear results in normal force load
reduction of less than 1.1%
)
. This
extended

arm geometry also provides
increased
precision in the
measurement of the Normal
f
orce (F
N
) and Drag
f
orce (F
D
). The Load
Arm is outfitted with four strain gauges, configured in two individual half
-
Wheatstone
bridge circuits to monitor the normal force applied to the test pin and also the drag force
resulting from the frictional forces betwee
n the pin and the rotating test disk.
The half
-
bridge arrangement was used to offset thermal effects on the strain gages which are also
temperature
-
matched to the aluminum arm substrate
(
load arm
manufactured from 7075
-
T651
). The ‘necked
-
down’ geometry o
f the long center section of the load arm was
designed specifically to
counter
the
maximum
design
operating loads while
also
allowing

sufficient deflection
to give accurate and precise strain readings.



42


Figure 17.


Load Arm Structural Analysis at Max Anticipated Op
erating Load


Conditions (Spring F
S
=180 lbf, F
D
=24 lbf)


The spindle speed input which is used to calculate the relative test speed and the
total distance travelled is acquired as an analog input from an inductive