‘ V ¢flak 3m.

. infimww. .w

. 1.: .= .u. t

,M 3?»...W mm a.
if“... an

.. a...

A.
.3.
,

       

    

 

m.T.e5tdvz€z
2.. ,fm. muvmwwm ‘ J. ,
e 3“th k1:

 

(WV.

                     
          

 

      

   

   

 

 

 

.J‘tvd .¢J {I}: (in, l v A, u
. V3.1. ‘ 3.3““: {e
,. :33: 5.5.1.;
1.. Z). fhulvzrrl ,3: Vi .
111.31.. um}... ufihumak . k
.. g... “£9333
u lu’ufitlc ‘96. V 4 . A
. .. v ‘3. «flu -xi u:
I...” ;. $2.311: :1 J.“
.r c 1 Iv“) ‘ ll
... .5. 13.2? ,

21.51:... 'v4

        

.1.
“.41.. a ”turn .
.éwwnfihkune: flu

 

1.1!. x r
1, f .n
’1}; All"'
3mm»... .,
1V . (l?

 

01.51‘99. 7). it :sz
.vfttxil‘fil Ir...

.4 .5?! in 1'!
5163!...

 

        
                    

.1 \I

u“
figwu
, Ifwznrun in”

l -
,&J m

 

 

 

 

 

  
 

 

 

. ‘ ‘ ‘ . T. 7 u . l. . . . A . . ‘
‘ ‘ . . v V. a
v a . . , . . . . , ‘ ‘
y A . . . V q A n . y . y .. ‘ . . . . ‘ rr . , ‘ ‘ . .
, . ‘ V l . «1101.)! ‘
. y in: mi ‘ . . . u . . . . a . . , , . . .. , . . . - t ‘ u.u.....u.a...n.l..}ht..w!.nn.u.huuufltp
. .2, I . 1!. A! , v r v ‘ t . : 1 n ‘ . (v n.
V . a 4 ,1 I y .l n. w I 1 W n v v:
. ‘ ‘ v ‘ , . ‘ ‘ .. HI 13,590,;
‘ .FL. I A ‘ A J LEI “w... ”thuarrfl.“ » fl ‘f-lhd. J v
1 ‘ . 5 r n. 3. . n ‘ oA u z a Ell” .1 . .r‘. l5 .I. (c l ‘
V .1 H I. — Sud. th ‘ is.“ ,- . .3.“ .77 2.“. 1d,. pr..lu.u|uv.l..n1.fl\n, .¢.:..:.uu.m..- 1L ifvfiwflrfiulflm ..: .3 c . . 1123;? a 5.. In: .1. ‘ .1» .f
. v ' ' 1, l...l. > I v' . . I .l . . . r . r

meats illlllHllHlllllillilHlilHliIlllllillililisollilllsillillllll

31293 01559

 

LIBRARY

Michigan State
University

 

 

 

This is to certify that the

thesis entitled ‘ - ‘ C
HYDROQEN CHARGHQG‘ AN]: THE. EFFECT or LAIHcI‘Di-.

C—HfiKGE‘p 54129106.th m3 A

'sz‘ncocikte - KEiNf-‘oacizl AchtNeM ccmPcsvcES

presented by

has been accepted towards fulfillment
of the requirements for

Mo. :5 i; Q. ~( degree in “(LEC- vLcds SC-\L enga—

 

 

   

Major professor

Date (6 9

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

PLACE ll RETURN BOXtoremovethiecheckoutfl'om yourrecord.
TO AVOID FINES return on or More dete due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU leAn Alfinnetive ActichEquel Opportunity lnetituion

 

HYDROGEN CHARGING AND INTERNAL HYDROGEN EFFECTS ON
IN TERFACIAL AND FRACTURE
PROPERTIES OF METAL MATRIX COMPOSITES
By

Chidambara Ramalingam Arunachalam

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requrements
for the degree of

MASTER OF SCIENCE

Department of Materials Science and Mechanics

1994

James P. Lucas

ABSTRACT
HYDROGEN CHARGING AND THE EFFECT OF CATHODIC

CHARGED HYDROGEN IN A
PARTICULATE - REINFORCED ALUMINUM COMPOSITES

By

Chidambara Ramalingam Arunachalam

This investigation was conducted to assess the influence of hydro-
gen on interfacial and bulk fracture properties of Al/SiC composite.
Cathodic charging was used to introduce hydrogen into the samples. As a
reference hydrogen was introduced into the matrix material Al-6061-T6.
To determine the hydrogen concentration in Al/SiC composite a thermal
desorption technique was used. In spite of sustained charging periods.
the evidence suggested that high hydrogen conCentration existed at the
surface. Corrosion products and extensive surface cracking were
observed on the surface of the composite materials. Fracture surface
morphologt showed insignificant difference between the charged and the
uncharged samples indicating the lack of internal hydrogen penetration
into the specimen. Due to hydrogen-induced surface corrosion of charged
samples. final ligament fracture was brittle in nature as opposed to
uncharged samples which exhibited ductile behavior. In this study
emphasis is given to the role of surface hydrogen absorption on failure in

Al/SiC composites.

ACKNOWLEDGEMENTS

First and foremost I would like to thank my adviser Professor
James P. Lucas for all his encouragement, help and support, without
which this work would not have been possible. I would also like to thank
my committee members Professor Melissa J. Crimp of the Department of
Materials Science and Mechanics and Professor Dennis J. Miller of the
Department of Chemical Engineering for their help and suggestion from
time to time. Thanks are also due to Professor K. Mukherjee and all other
faculty, staff and students of the Department of Materials Science and
Mechanics for their support. Last but not the least I would like to thank
my guru, parents and brother for their support.

TABLE OF CONTENTS

List of Tables
List of figures
1. Preface
2. Introduction
3. Background
3.1 Hydrogen in Metals (w.r.t Al)
3.1.1 Solubility
3.1.2 Diifusivity
3.1.3 Occupying Sites
3.1.4 Trapping

3.1.4.1 Vacancies

3.1.4.2 Dislocations

3.1.4.3 Grain Boundaries
3.1.4.4 Precipitates

3.1.4.5 Particle- Matrix Interface

3.2 Hydrogen Charging

3.3 Cathodic Hydrogen in Metals
3.3.1 Pure Aluminum
3.3.2 Aluminum alloys

3.4 Hydrogen Transport by Dislocation
3.4.1 Theory
3.4.2 Experimental Evidence

3.5 Thermal Desorption for Qualitative
Measurement of Hydrogen.
3.4.1 Theory
3.4.2 Basic Operation

iv

5

16

19
19
2O

21
21
23

25
25
25

6.

Experimental Procedure
4.1 Materials '

4.2 Cathodic Charging of Hydrogen
4.3 Desorption Experiments
4.4 Characterization of Al-SiC Composite
and Decohesion experiment
4.4. 1 Mounting the Tensile Samples
4.4.2 Polishing
4.4.3 Microstructurai Characterization
4.5 Mechanical Testing
Results and Discussion

5.1 Cathodic charging of hydrogen

5.2 Characterization of All 810 Composite and
Decohesion Experiment

5.3 Fracture Surface Morphology
5.4 Thermal Desorption Experiment
5.5 Mechanical Testing

Conclusions and Future Work

BIBLIOGRAPHY

30
3O

30
35
36
36
37
37
37
4O

4O

45
50
57
65
78

81

LIST OF TABLES

Table 3.1 Solubility data for hydrogen in solid aluminum
Table 3.2 Diffusivity data of hydrogen in aluminum

Table 3.3 Thermal conductivity of various gases relative to air
Table 4.1 Material used in this study

Table 4.2 Sample dimensions used corresponding to various
tests are tabulated.

Table 5.1 Sample dimensions, charging conditions and material
used are shown for various desorption curves
obtained.

Table 5.2 Flexural test results obtained are shown for both
the charged and the uncharged samples of Al/SiC.

Table 5.3 Tensile test results obtained are shown for both the

charged and the uncharged samples. Corresponding
sample i.d and material used are also tabulated.

vi

26

31

38

59

67

68

Figure 3.1

Figure 3.2

Figure 4.1

Figure 4.2

Figure 5.0

Figure 5.1

Figure 5.1

Figure 5.2

Figure 5.3

Figure 5.4

Figure 5.5

Figure 5.6

LIST OF FIGURES

Cathodic charging apparatus

Schematic diagram illustrating the working principle
of pulse chernisorb 2700

a). Shematic diagram illustrating the cathodic
charging set-up. b). Photographs shows the cathodic
charging set-up.

Design of the three point bend test fixture.
a). Macrograph shows the crack formed in steel
charged with hydrogen. b) Macrograph indicates the

absence of crack in the uncharged steel.

a). Hydrogen-induced corroded surface layer of
Al/SiC Composites.

b) Acquired X-ray spectrum for the corroded
surface layer.

Microstructure of Al/SiC composite a)700X b)2.20KX.

a) Uniform precipitates of Mngi in the bulk matrix.
b) Micrograph shows cracked SiC particles.

Shows void growth associated with the
fractured particles a) 10.0KX b) 1.10KX.

a). Micrograph showing transparticle fracture,

slip bands and microvoid coalescence of the matrix
[charged 61 hours]. b). Fracture surface shows
extensive p71astic deformation as striations on the
inside of the void [charged 6% hours].

Shows bottom slip band of figure 5.6 b at higher
magnification.

vii

17

27

34

39

41

43

44

47

48

49

53

54

Figure 5.7 a). The fracture surface shows a cluster of SiC at higher
magnification (no hydrogen). b). Extensive plastic
deformation was observed as striations are prevalent
on the inside of the void walls as a result of intense
slip bands intersection (no hydrogen). 55

Figure 5.8 Fracture surface morphology shows a bunch of SiC
particles. microvoid coalescence of the matrix and void
growth around the particle for both the charged
(for 19 hours) and the uncharged samples [b]. 56

Figure 5.9 Desorption signal obtained by heating the empty quartz
tube to 723 K. 60

Figure 5.10 Compares the desorption curves obtained for
charged and uncharged steel samples. 61

Figure 5.11 Desorption curve obtained for steel charged with
hydrogen [Sample i.d: D51]. 62

Figure 5.12 Compares the desorption curves obtained for
charged and uncharged Al-6061-T6 samples. 63

Figure 5.13 Compares the desorption curves obtained for
charged and uncharged Al/SiC (Vf = 10%) samples. 64

Figure 5.14 Effect of Surface Hydrogen on the Percentage Final
Failure Strain for Al-6061-T6 and Al/SiC Composites
(Tested in the Transverse Direction). 69

Figure 5.15 Stress-Strain curves obtained for Sample Fl
[Al/Sic (V; = 20%)]. 70

Figure 5.16 Stress-Strain curves obtained for Sample F2
[Al/SiC Wf = 20%)]. 71

Figure 5.17 Stress-Strain curves obtained for Sample 8- 1
[Al-6061-T6]. 72

Figure 5.18 Stress-Strain curves obtained for Sample B-2
[Al-6061-T6]. 73

Figure 5.19 Stress-Strain curves obtained for Sample B-3
[Al-6061-T6). 74

Figure 5.20 Compares the Stress-Strain curves obtained for

viii

charged and uncharged samples of Al/SiC (Vf = 10%)
tested in transverse direction. 75

Figure 5.21 Stress-Strain curve obtained for Sample C5
[Al/SiC (Vf =10%)] tested in transverse direction. 76

Figure 5.22 Compares the Stress-Strain curves obtained for
charged and uncharged samples of Al/SiC [Vf =20%)
tested in transverse direction. 77

Chapter 1
miss:

The interaction of materials with their environment is a subject that
has been studied for long time. It is imperative to study the behavior of
material in any environment before they can be used in that particular
environment. Hydrogen is present in many environments, hence it is

important to study the interaction of hydrogen with materials.

Interaction between hydrogen and many materials generally result
in the degradation of the material's properties. In the case of metals deg-
radation can occur with very small quantities of hydrogen, on the order of
atomic parts per million (at. ppm.). As material scientists are developing
new materials it is essential to study the behavior of these materials in
various environment before they can be widely used. Despite the fact that
research has been done for numerous years on this sub) ect, many
aspects of internal hydrogen-induced effects are not yet clearly under-
stood and controversy exists still.

Chapter 2
Introduction

Most of the work in the field of hydrogen has focussed on mono-
lithic materials such as iron and titanium alloys and to a lesser extent.
aluminum alloys. In contrast. metal matrix composites have received
much less attention.

Because of high specific stiffness and high specific strength of
metal matrix composites. the application of these materials has
increased, significantly displacing their monolithic counterparts.
Although much research has been done on mechanical properties of
metal matrix composites obtained under benign test conditions. very little
mechanical properties data are found for these materials tested under
severe environment condition.

One of the objectives of this study is to choose a suitable technique
to charge hydrogen and to design a charging cell. The technique used to
charge hydrogen in this study was cathodic charging. This study was
conducted to assess the influence of hydrogen on interfacial and bulk
properties of Al/SiC composite. Both 20% and 10% volume fraction SiC
particulates was used. As a reference, hydrogen was introduced into the
matrix material Al-6061-TB. Tensile coupons of composite material (both
charged and uncharged) were strained to the same extent, polished using
a diamond compound on a lapping wheel. and observed for void nucle-
ation at the interface using a Scanning Electron Microscope.

In order to qualitatively measure the amount of hydrogen charged.
a thermal desorption technique was used to get the desorption curve for
steel, Al-6061-T6 and Al/SiC composite. Desorption curves obtained for

2

3
hydrogen charged samples were compared to those of the as-received

sample.

Several important conclusions were made from the above study.
Although sustained charging times were used. there was significant evi-
dence that absorbed hydrogen resided mostly on the surface. Cathodic
charging of hydrogen results in a corrosion product on the surface of the
specimen. An extensive fracture morphologr study was conducted. In
general. fracture surface morphologies showed insignificant differences
between the charged and the uncharged samples. This is to be expected
considering the lack of internal hydrogen penetration into the specimen.
The surface hydrogen effect was manifested only by the unique fracture
behavior of the final ligament failure during mechanical testing. The final
ligament in the charged samples showed a brittle behavior as opposed to
that in the uncharged samples. which exhibited a ductile behavior. For
the matrix material (Al-6061-T6). the differences in strain between the
charged and the uncharged samples were found to range froml to 2%.
This due to the combined effect of surface hydrogen and corroded surface

layer.

3
hydrogen charged samples were compared to those of the as-received

sample.

Several important conclusions were made from the above study.
Although sustained charging times were used. there was significant evi-
dence that absorbed hydrogen resided mostly on the surface. Cathodic
charging of hydrogen results in a corrosion product on the surface of the
specimen. An extensive fracture morphology study was conducted. In
general. fracture surface morphologies showed insignificant differences
'between the charged and the uncharged samples. This is to be expected
considering the lack of internal hydrogen penetration into the specimen.
The surface hydrogen effect was manifested only by the unique fracture
behavior of the final ligament failure during mechanical testing. The final
ligament in the charged samples showed a brittle behavior as opposed to
that in the uncharged samples. which exhibited a ductile behavior. For
the matrix material (Al-6061-T6). the differences in strain between the
charged and the uncharged samples were found to range froml to 2%.
This due to the combined effect of surface hydrogen and corroded surface

layer.

Cha ter 3

Bac und
3.1 Hydrogen in Metals (w.r.t All

For nearly a century. it has been known that hydrogen embrittles
metals. Early investigation clearly indicates that hydrogen easily embrit-
tles more open body centered cubic material like iron and steel. From a
casting standpoint. unsoundness in casting, mechanical weakness gener-
ated by hydrogen and blistering in wrought products has been noted [1].
Dissolution of hydrogen can lead to embrittlement by formation of brittle
matrix hydrides. Though much research has been done regarding hydro-
gen induced porosity they don’t give much information about diffusion
and solubility of hydrogen [1]. It has also been found that aluminum
alloys are susceptible to stress corrosion cracking, which is often attrib-
uted to hydrogen embrittlement. even though gaseous hydrogen does not

cause any significant embrittlement [2].

The methods that are generally used to introduce hydrogen into the
metal are cathodic and gas phase charging techniques. Despite the fact
that aluminum has a low hydrogen lattice diffusivity. deep bulk embrittle-
ment has been observed. Diffusivity of hydrogen depends on the trapping
propensity at defects in engineered materials. Dislocations have been
proposed as one such trap sites to many metals including aluminum.
Many studies indicate strong interaction between hydrogen and disloca-

tions but they are not very conclusive [3.4].

3.1.1 Solubility
Pure Aluminum:

Numerous studies have been conducted into the equilibrium solu-
bility of hydrogen in pure aluminum [5). Solubility of hydrogen in alumi—
num is very low. No methods have been used to measure solubility of
hydrogen in aluminum are gas extraction technique and permeation tech-

nique.

The gas extraction technique involved successively condensing out
the various gas species in progressively colder traps. finally leaving only
hydrogen gas. If pressure of this gas is measured along with its known
volume. the amount of hydrogen gas originally in the sample can easily be
calculated. This technique has been used in numerous studies Eborall
and Ransley [6]. Ransley and Neufeld [7], Eichenaeur and Pebler [8).
Eicheneauer and coworkers [9]. The rate that the hydrogen is outgassed
from the sample is indicative of the diffusivity of hydrogen in the sample.
The same treatment has often been used to measure specifically the diffu—

sivity of hydrogen in aluminum which is discussed in the later section.

Hashimoto and Kino [5] used permeation technique to measure
solubility of hydrogen in aluminum. Though they were specifically inter-
ested in hydrogen diffusion in aluminum at high temperatures. they were
able to obtain the solubility of hydrogen in aluminum using a mass spec-
trometer in an ultra high vacuum chamber. the current through the spec-
trometer was measured. Hydrogen permeation through the specimen was
detected by the Hf ion current and recorded on a recorder. The equilib-

rium concentration (solubility) was obtained from the stationary ion cur-

rent. 1, given by

C: LI/kAD

where

A - Exposed area

L - Sample thickness

k- Constant (k=No/ 2k)

No- Number of lattice sites per unit volume

2k- Permeation rate corresponding to the hydrogen ion current of

1A.

Although several works have measured solubility of hydrogen in alumi-
num. the result do not generally agree [5]. Solubility data for hydrogen in
solid aluminum are shown in Table 3. 1.

Table 3.1 Solubility data for hydrogen in solid aluminum

 

 

 

 

 

 

 

 

Co (atm'm) BF (kcal/mol) Reference

1.48 X 10'4 9.5 Ransley and Neufeld

2.7 X 10'2 19.4 Eichenauer and Pebler

2.2 X 10’3 13.9 Eichenauer, Hattenbach
and Pebler

3.1 X 10'3 15.1 Eichenauer

3.4 x 10'3 15.5 Edwards and Eichenaucr

2.3 X 10'0 23.1 Hashimoto and Kino

 

 

 

 

 

c = C0 «JP exp (-EF/kT)

3.1.2 Diffusivity

The diffusion of hydrogen in metals has been measured using vari-
ous techniques like desorption technique and permeation technique.
However. a large scatter in data indicates how difficult it is to determine
the diffusivity of hydrogen in metals. A scatter in the measurements may
be attributed to the trapping of hydrogen at lattice imperfections such as
impurities. dislocations. grain boundaries. etc., and to the influence of
different surface conditions. A large degree of discrepancy in H-diffusivity
data obtained for metals is due to the difference in surface conditions.
The various surface conditions that can affect the diffusivity of hydrogen
in metals are

1. Sample pretreatment can lead to differing states of oxide layer.
In case of aluminum the oxide layer acts as a barrier to diffusion of hydro-
gen. Thus the diffusion of hydrogen into the sample can be increased by
minimizing the thickness of the oxide layer.

2. The difference in final surface preparation. The diffusivity of
hydrogen into the sample depends upon the method of final surface prep-
aration. The various methods that can be used are [10]

a) Vacuum annealing after heavy cold deformation

b) Mechanical abrasion using emery or diamond paste

c) Electrochemical polishing
Mechanical polishing results in large concentration of voids and disloca-
tions which leads to enhanced solubility [10]. Structure resulting from
heavy cold deformation cannot be completely annihilated by vacuum

annealing. This results in difference in concentration of semi-microscopic

8
voids and micro—cracks which leads to difference in H-diffusivity [10].

In desorption technique. the procedure generally used consists of
assuming a diffusivity and calculating an out-gassing profile. This
method is generally limited to high temperature range where desorption
rate can be measured. In permeation technique [5] the diffusion of hydro-
gen in aluminum was measured using a quadrupole mass spectrometer
and an ultra-high vacuum system. In their study [5] hydrogen was intro-
duced into specimen by two methods such as gas phase charging and
electrochemical charging. Hydrogen permeation through the specimen
was detected by the hydrogen ion current and recorded on a recorder.
The diffusion of hydrogen in aluminum has been measured at high tem-
perature and room temperature by permeation method. Diffusion data of

hydrogen in aluminum are shown in Table 3.2.

It should be noted here that with exception of few most of experi-
ments were conducted using desorption technique and at high tempera-
ture. K.Papp [11] have shown that activation energy and pre-exponential
factor of the hydrogen diffusion in high purity aluminum are smaller than
those of the less pure. The effect of impurities results in a decreases of

the diffusivity D. at the range of temperature investigated.

As mentioned earlier there have been few measurements made in
lower temperature range. A major problem in measuring the flux of
hydrogen through aluminum is that of low diffusion of hydrogen in alumi-
num. To overcome this problem a new technique making use of a com-

posite specimen incorporating a very thin layer of aluminum deposited on

9

Table 3.2 Diffusivity data of hydrogen in aluminum

 

 

 

 

 

 

 

 

 

 

2
Material Method D95: / Q KJ/mol Literature
99.99% Al Desorption
633-873 K 1.1 x10‘5 41 Eichenauer
[9].
Al 99 Desorption
773-873 K 1.2 x10‘5 61 Andrew
843-903 K 2.0 x 1045 50 Matsuo
99.8% Al Desorption 25 x10-2
723-863 K 90 K.Papp and
E.Kovacs-
Csetenyi
[1 1].
99.999% Al Desorption
723-863 K 1.9 x10‘5 40 K.Papp and
E.Kovacs-
Csetenyi
[1 1].
99.995% Al Desorption
723-923 K 1.01 x10'5 47.69 Outlaw.
Peterson
and Schmid
99.999% Al Permeation
573-673 K 2.1 x10‘5 59.83 Hashimoto
and Kino
(1983) [5].
99.995% Al Permeation
285-328 K 9.2 x10-5 55.23 Ishikawa
and McLel-
Ian [12].

 

 

 

 

 

D= Do exp (-Q/RT)

 

10
a thicker substrate of a fast diffusing. high hydrogen solubility media.

The sandwiches consisted of layers of palladium. iron and aluminum and
were formed by vapor deposition onto the iron substrate. To avoid inter-
metallic formation between aluminum and palladium the experiments
were conducted at room temperature. The measured diffusivities were

between 10‘15 and 10’16 cm2 per second.

T. Ishikawa [12] employed an electrochemical technique. They

' measured diffusivity of hydrogen in aluminum in the temperature range
of 285-328 K. The room temperature measurements showed good agree-
ment with several of high temperature data. It suggested that hydrogen
diffusion in solid aluminum is controlled by single mechanism from 300 K
up to the melting temperature. Since the vacancy concentration at 300 K
was negligible. good agreement between room temperature and high tem-
perature data indicated that vacancy could not play an important role in
the diffusion process. Since there was no deviation from Arrhenius lin-
earity observed in this experiment, it implies that trapping was not signif-
icant. It should be noted here that this behavior could have occurred

because there were not enough trapping sites in the sample used.

Like solubility. diffusivity of hydrogen depend on the trapping pro-
pensity at defects in engineering materials. These traps include disloca-
tion cores. grain boundaries. lattice substitutional sites and particle-
matrix interfaces. The strength of traps is represented by the binding
energy. The most potent trap site are those at particle-matrix interface
(e.g: particle matrix Fe3C interface with trap energy about 84 KJ/mol).
Extending this one would expect that binding energy of Al/SiC interface

11
in metal matrix composites to be high. Though trap sites will decrease

the diffusion of hydrogen in the material at low and ambient temperature
the local hydrogen concentration can be increased in the sample signifi-

cantly higher than in the pure lattice.

3.1.3 Occupying Sites
The two interstitial sites that hydrogen can occupy are octahedral
and tetrahedral sites. There is controversy concerning on which site does

hydrogen reside. exactly.

Eichenaur [9] measured solubility of deuterium in 99.99% alumi-
num between 673 and 903 K. He also calculated the solubility of deute-
rium in aluminum assuming octahedral site occupancy. Both these
calculations agreed well leading him to conclude that hydrogen and deu-
terium occupied Ectahedral sites in aluminum.

Positive muons are regarded as light isotope of hydrogen and used
to investigate several properties of hydrogen in metals. Manninen and
Beck [13] used muons to investigate properties of hydrogen in metal.
They point out that numerically. the tetrahedralsite has been found to be
the most stable position for the proton. in agreement with muon results in
doped samples [14] and with theoretical calculations [15]. In view of the
fact that some theoretical calculations using various potential modeling
give very similar energies for the tetrahedral and octahedral sites (within
0.1 e.v). The result concerning the 'priority‘ of the tetrahedral site may
not be conclusive. They conclude that though the energy minimum is at

the tetrahedral site. the barrier is extremely low so that energetics seem to

12
allow an almost free motion of the proton along the cubic diagonal.

Larson and Norskov [15] used a jellium approximation and calcu-
lated energy potential for the proton. They found that energy minimum
occurred at the tetrahedral site rather than octahedral site.

3.1.4 Trapping

Hydrogen often preferentially occupies particular lattice sites; trap-
ping sites. Several types of sites have been found to trap hydrogen.
These traps include dislocation cores. grain boundaries. vacancies. voids.
lattice substitutional sites and particle-matrix interfaces. Trapping would
be more efficient at lower temperature and has been proposed to explain

deviations from Arrhenius linearity in many metal systems [16].

3.1.4.1 Vacancies

The possibility of hydrogen trapping at a vacancy in aluminum was
first suggested by Popovic and Scott [17]. They used non-linear theory to
predict that the binding energy between a vacancy in aluminum and a
hydrogen to be 1.23 eV.

In most subsequent investigations. a spherical solid model was
used [18, 19.20]. The results obtained within spherical solid models indi-
cate a preference for octahedral interstitial site contrary to preference of

vacancy as suggested by Popovic.

Hashimoto and Kino [5] measured diffusivities of hydrogen through
pure aluminum foils. They obtained the migration energy of hydrogen dif-

13
fusion at high temperature as 0.61-0.62 eV. which was near to the migra-

tion energy of a vacancy. The value for D0 being 0.21-0.26 cm2/ sec was
somewhat larger than that of a vacancy. 0.02-0. l cm2/sec. which was
estimated from the concentration of thermal equilibrium vacancies. They
suggested that these facts might mean that the diffusion mechanisms are
different between high temperature and room temperature. Dissolved
hydrogen probably migrates freely through the interstitial sites at room
temperature where the hydrogen concentration is much higher than that
of the vacancies. On the other hand. at high temperatures where vacancy
concentration is large enough to couple with almost all of the dissolved
hydrogen. it may diffuse together with a vacancy due to larger binding
energ between hydrogen and the vacancy [5]. Such a diffusion mecha-
nism would give an explanation of the measured Do being somewhat
larger than that of a vacancy because the existence of hydrogen might

enhance the lattice vibration around a vacancy [5].

Larson and Norkov [15] calculated the trapping energy of hydrogen
to be 1 eV. Their conclusion was that hydrogen is trapped near the tetra-
hedral site next to a vacancy with a binding energ of about leV.

On the other hand. experiments conducted by Ishikawa and
Mc.Lellan [12] suggest that hydrogen diffusion in solid aluminum occurs
by single mechanism from 300 K up to the melting temperature. The dif-
fusivity measurements failed to show any deviation from linear Arrhenius
relationship over temperature range of 285-328 K and other measure-
ments at high temperature range. From the above observation. they con-

clude that trapping at lattice vacancies was unlikely.

. 14
3.1.4.2 Dislocations

lf hydrogen should be transported by dislocations then at first
hydrogen must be trapped at the dislocations. Later. by motion of these

traps. hydrogen transport occurs.

Zurek and others [21] proved that dislocation can act as trap by
their internal friction experiments. They were able to identify a very
strong peak as a result of hydrogen substitution atom pairing, and. a sec-
ond. less pronounced peak resulting from hydrogen dislocation interac-
tion for 7075-T6 aluminum with deuterium. They later used SIMS to
obtain deuterium concentration profiles. Fatigue and cold rolled samples
showed a high concentration of deuterium at the surfaces and significant
concentration level of deuterium extending into the material much deeper
than the observed oxide layer. This observation suggests trapping of
hydrogen at defects induced by these treatments. notably dislocations.

3.1.4.3 Grain Boundaries

Edward [22] discusses the phenomenon of reversible hydrogen
trapping at grain boundaries in super pure aluminum. They measured
the hydrogen solubility in 99.99% aluminum between 673 and 903K.
They found that there was decline in hydrogen solubility with that of grain
growth. The observed decline in solubility should in this case be related
to the reduction in grain boundary area accompanying the grain growth.
By measuring change in grain boundary area and change in solubility,
the grain boundary solubility was calculated as

LIB-b) = 1.15x1o-6 exp [(-50640 J /mol I-I)] (mol H Atm'l/2 / m2 g.b)

(RT)

 

15
Edward and coworkers calculated the binding energy between

hydrogen and the grain boundary to be as 34 kcal/mol.
The work of Hardwick [23] showed that the extent of brittle fracture
depend on local hydrogen accumulation at grain boundaries and extent of

dislocation transport of hydrogen.

3.1.4.4 Precipitates

Precipitates both in matrix and grain boundaries can act as traps
for hydrogen. Schmans [24] studied the role of primary interrnetallic par-
ticles situated at grain boundaries on the effect of hydrogen penetration.
Schmans observed that grain boundary precipitates in Al-Zn-Mg alloys
can getter hydrogen away from the boundary; this is very beneficial
because hydrogen trapped at specific precipitates would not reduce the
cohesive strength of the entire boundary. He also reported Man2 precip-
itates show a distinct propensity for promoting hydrogen bubble nucle-
ation. Bubbles were seen to form on grain boundaries that did not
contain precipitates but these bubbles were transient and therefore would
not provide beneficial trapping that the grain boundary precipitates pro-
vided. Thus they concluded that the hydrogen bubbles show affinity for
both grain boundary precipitates in overaged microstructures and for pri-
mary interrnetallic phases.

Christodoulou and Flower [25] reported similar observation in
hydrogen embrittlement of Al-6%-Zn-3%-Mg alloy exposed to water at ele-
vated temperature. Nguyen and coworkers [26] observed that in 7075
alloy there is clear tendency for decreasing susceptibility with increasing
size of both matrix and grain boundary precipitates. They concluded that

16
size of matrix precipitates is the most important factor in determining

susceptibility to hydrogen embrittlement. while grain boundary precipi-
tates also play role. but to a lesser extent.
In general one can conclude that interrnetallic precipitates. both in
matrix and grain boundaries are very potent trapping sites for hydrogen.
3.1.4.5 Particle- Matrix Interface

Like interrnetallic precipitates particle-matrix can also serve as
traps. Asaoko [27] calculated the binding energy of particle-matrix Fe3C
interface and found it to be very strong of about 84KJ/mol. The binding
energy for dislocation cores and grain boundaries are also relatively
strong of about 20-60KJ/mol.
3.2 Hydrogen Charging
Methods for hydrogen charging include

1. Cathodic charging.

2. Cathodic charging during tensile testing [Straining electrode test)

3. Gas phase charging technique.
Cathodic charging is the most often used method to charge hydrogen.
Cathodic charging is more effective than other charging methods because
of

1. High fugacity of hydrogen

2. The effect of A1203 layer as a barrier for diffusion of hydrogen is
minimized in this method. A1203 being an amphoteric oxide reacts with
the electrolyte used for charging and hence decreasing its effectiveness as
a barrier for diffusion of hydrogen.
Two types of circuit can be used for cathodic charging of hydrogen: cur-
rent and voltage control. Figure 3.1 shows a general cathodic charging

apparatus. Constant voltage source or constant current source can be

17
used depending upon the type of circuit selected. In case of a constant

current source a voltmeter can be used to measure the voltage between
the sample and the anode. Generally an acidic electrolyte with a recombi

nation poison is used for cathodic charging of hydrogen.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 3.1 Cathodic charging apparatus.

A. Sam 1e. cathode.

B. Ano e.

C. Voltmeter.

D. Power source: constant voltage source or constant current source.
E. Ammeter.

18
First let us examine what a hydrogen recombination poison does. In a

typical electrochemical set-up hydrogen ions are attracted to the metal
surface where they are neutralized. The resultant hydrogen atoms can
either be absorbed by the metal or recombine to form molecular hydro-

gen. which bubbles off the surface. This process can be represented as

1/2 H2 (adsorbed) —> “2 H2 (bubbles)

A:

- O
H+(solution) + e > H (adsorbed)

\

O
H (absorbed)

 

A hydrogen recombination poison slows the rate of formation of H2
(adsorbed) from Hoiadsorbed) by reducing the recombination rate constant.
K. thus slowing the recombination step leads to absorption of more
hydrogen atoms [28]. A number of elements. particularly from groups VA
and VIA in the periodic chart. have been identified as hydrogen recombi—
nation poisons in metallurgical studies of hydrogen embrittlement. The
most commonly used recombination poisons are A5203 and C52. In this
study constant voltage source of — 1.5V was applied to the cathode (sam-
ple). As soon as the voltage is applied. electron is supplied by the anode
(platinum wire) and the reduction of H“ to H2 occurs at the cathode. This

molecular hydrogen bubbles off the surface.

Cathodic charging during straining is generally used to investigate
the possibility of dislocation transport of hydrogen. The experiment

19
involves cathodic charging. similar to that described earlier. while the

sample is simultaneously strained in tension.

Gas phase charging is done by using Sieverts apparatus. The
quantity of hydrogen charged depends on pressure, temperature and time
used for charging. The amount of H introduced can be determined by the

pressure change in the system [29).

3.3 Cathodic Hydrogen in Metals

Steel and Titanium are the most extensively studied metals for
hydrogen embrittlement. Both of these materials are quite sensitive to
hydrogen. Beta 1 and VT-15 titanium alloys are more sensitive to hydro-
gen when compared to Beta III titanium alloys. For Beta III titanium
alloys the percent elongation of tensile specimens is appreciably
decreased only by high H contents. much higher than those necessary to
embrittle to the same extent other B Ti alloys: Beta I (l‘i- 13V- 11Cr-3Al)
and VT-15 (Ti-7Mo-11Cr-3Al). similarly charged and pulled at the same
rate [30]. Andler and coworkers concluded that hydrogen redistribution
(results in H2 transport) occurs under the influence of a stress gradient in
B-phase titanium alloys. This phenomenon was not observed in a-phase
titanium alloys [31]. The fact that stress induced hydrogen redistribution
was not observed in (it-phase titanium alloys is due to the low hydrogen
diffusion coefficient in a-phase titanium at room temperature. Steel

being a open structure is easily susceptible to hydrogen embrittlement.

3.3.1 Pure Aluminum
Watson. Shen and Meshii [32] reported the effects of cathodic
charging on pure aluminum. They reported that cathodic charging

20
decreased the ductility, as measured by strain to failure. All their sam-

ples showed a fully ductile fracture modes. They also reported that

cathodic charging caused slight hardening in tensile tests.

3.3.2 Aluminum alloys
The most extensively studied aluminum alloy has been 7075. 7050

and 2014 aluminum alloy have also been studied.

Gest and Troiano [2] was the first to study the effect of hydrogen on
aluminum alloy. They used high strength aluminum alloy 7071 T651.
The embrittlement was found to depend on strain rate and test tempera-
ture. The embrittlement was most severe at lower strain rate. They also
confirmed the reversiblity of hydrogen induced damage. The internal fric-
tion experiment obtained a very weak peak at 170K signifying hydrogen-

dislocation interaction.

Tehari [33] studied embrittlement of 7075 aluminum alloy in differ-
ent aging conditions over a range of four orders of magnitude in strain
rate. They found that the dependence of embrittlement on strain rate was
not constant but rather went through a maximum. The strain rate for
maximum embrittlement was different for different aging conditions.
They also showed that the recovery at low strain rates was not due to dif-
fusional losses of hydrogen from the sample: they proposed that diffu-
sional escape from detrimental trap sites to more innocuous lattice sites
was responsible for the recovery. This explanation was confirmed by the
observation that holding the samples in vacuum for 48 hours between

charging and testing did not affect the embrittlement [34].

21
Albrecht. Thompson and Bernstein [35] studied the role of micro-

structure in hydrogen assisted fracture of 707 5 aluminum. For all micro-
structures (underaged. peakaged and overaged). a temperature dependent
loss of ductility due to hydrogen was found. The ductility loss was a func-
tion of microstructurefbeing highest for the underaged and lowest for the
overaged material. which suggested a correlation between hydrogen
embrittlement and slip planarity. Generally the effect has been explained
that overaging increases the number of incoherent precipitates which
cannot be sheared by passing dislocations. Underaged microstructure
are dominated by coherent shareable precipitates. This explanation has

often been used to imply dislocation transport mechanism.

It is important here to mention one aspect of cathodic charging phe-
nomena in aluminum alloys. Albrecht [3] found that most but not all of
the embrittlement could be recovered by machining away the surface
layer after charging. They suggested that total effect of hydrogen charging
must be separated into two smaller effects. One due to a corroded sur-
face layer. and the other associated with embrittling effect of hydrogen
retained in interior of the specimen.

Almost all research on aluminum alloy has been confined to 7075
and 2124 alloy. Very few research has been done on 6061 alloy. the
material used in this particular study.

3.4 Hydrogen Transport by Dislocation
3.4.1 Theory
Diffusivity of hydrogen by bulk means is very low. Dislocation

transport has been proposed to account for the higher diffusivities seen in

22
experiments.

The nature of dislocation is that it attracts hydrogen. Although.
there is no longer doubt that dislocation sweeping in result deep hydrogen
penetration. deeper than by lattice diffusion. The question still remains
whether dislocation sweeping can result in significant local enrichment.
or supersaturation, of hydrogen at critical failure sites in the material.
such as the interface of internal traps. Traps may be precipitate particle
interface. grain boundaries and inclusions or internal voids in the mate—
rial whose interfaces are often preferential site for microvoid or microc-
rack initiation. It should be noted that microvoid or microcrack can be
initiated when a critical hydrogen concentration is exceeded in local

regions.

The theoretical basis for quantitatively evaluating any possible
supersaturation or enrichment of hydrogen at specific sites in the mate-
rial as a result of dislocation sweep in of hydrogen was originally laid by
two kinetic models.

In one model. by Johnson and Hirth [3]. they assume an equilib-
rium between trapped population and the matrix. If the sink for disloca-
tion-transported atomic hydrogen is a void. then recombination to the
molecular state would be expected upon discharge. Hydrogen emission
and diffusion away from the void must then be preceded by molecular
dissociation. The recombination and dissociation rates are sufficiently
high that local equilibrium is established. and is believed to be the most
probable situation. Even in the case of annihilation of hydrogen-laden
dislocation in a trap free matrix they claim that equilibrium situation will

23
result. With a equilibrium situation. most of hydrogen leaves the trap

and very little supersaturation occurs.

The other model. by Tein [4], contends that internal traps would
strip hydrogen off the moving dislocation because of the binding emery of
the traps. This would result in supersaturation of hydrogen at the trap-
matrix interfaces. Enrichment would occur if the arrival rate of hydrogen
to traps by dislocation sweeping were greater than diffusion of hydrogen
to surrounding matrix. The model predicted that significant hydrogen

build-up can occur at the interface.

Hirth and Dienes [36] on the rate of equilibration of diffusional
traps in solids considered the rate at which equilibrium boundary condi-
tion is attained at an internal diffusion source. The source is a trap site
with a binding emery for the diffusion species relative to the bulk lattice.
The case of a given initial amount in the trap and the case where the dif-
fusion species is supplied to the trap at a constant rate are considered.
Two types of traps are also considered

1. Short range trap where the binding emery is released in one
jump.

2. Long range trap where the binding emery is released in more
than one jump.
In their model they concluded that long range trap sites required more
time to equilibrate than short range trap sites.
3.4.2 Experimental Evidence

Frankel and Latanison [37] in their experiment observed direct evi-

dence for dislocation transport of hydrogen in nickel single crystal using

24
the electrochemical permeation modified to allow for sirnultameous defor-

mation. The hydrogen flux increased in the easy glide region of deforma-
tion even after accounting for the effects of decrease in specimen
thickness. At fast strain rate the only explanation for this is dislocation
transport. However. it should be noted that polycrystals did not show sig-
nificant dislocation transport.

Albrecht [38] used straining electrode test to study 7075 alumi-
num. This test combines straining and cathodic charging. They con-
cluded that combined staining and charging enhances hydrogen
embrittlement of am equiaxed HP 7075 alloy. relative to prechargimg alloy.
This observation was consistent across a wide spectrum of aluminum
alloys such as 2124 [39] and 7050. In many alloys a microstructure
favoring slip planarity has been correlated to high hydrogen susceptibil-
ity. generally significant embrittlement is observed in the underaged
microstructure. Albrecht [38] found that in HP 7075 for a properly ori-
ented grain structure where significant number of grain boundaries are
subjected to normal stresses. the hydrogen induced fracture is primarily
imtergranular. provided the local concentration is high enough. They con-
cluded from these findings that dislocation transport of hydrogen in alu-
minum is not only possible. but it is kinetically preferred to volume
diffusion. Furthermore. it enhances embrittlement far in excess of that

possible by static charging of hydrogen alone.

25
3.5 Thermal Desorption for Qualitative Measurement of Hydrogen.

3.5.1 Theory

Chemisorption instrument is generally used by chemist for measur-
ing the density of active sites on the surface of a catalyst. It is also used
to measure the total surface area of catalyst or any powdered solid. Here.
the instrument is used to measure the amount of hydrogen charged.
Though the results obtained are qualitative. they give a very good idea
about the amount of hydrogen charged. By calibrating the instrument
one will be able to use the instrument to measure the amount of hydrogen

charged.

3.5.2 Basic Operation

Schematic diagram of the instrument is shown in Figure 3.2. A
mixture of He/szgas can be used as carrier gas but in this experiment
only nitrogen gas was used. This was done to increase the sensitivity for

desorption of hydrogen.

The Pulse Chemisorb 2700 has a Thermal conductivity detector
(Wheatstome bridge network) which is capable of detecting the difference
in thermal conductivity of gas that enters the Quartz tube and the gas
that leaves the tube. Nitrogen gas was selected as a carrier gas because.
nitrogen and hydrogen have good difference in thermal conductivity so
that even the presence of hydrogen in small amounts can be detected.
Table 3.3 gives thermal conductivity of various gases relative to air. One
can see from the table that conductivity of Nitrogen and oxygen are

almost the same. This implies that even if there is evolution of oxygen

26
from the A1203 layer the set-up is not quite sensitive to oxygen.

Table 3.3 Thermal conductivity of various gases relative to air

 

 

 

 

 

Name Chemical Conductivity
Formula (Relative to air)
= = =
Hydrogen H2 7 .07
Helium He 5.84
Nitrogen N 2 1.00
Oxygen 02 1.02
Water vapor
H20 0.67

 

 

 

 

 

27

 

 

HC/N2 . [J

Reference I

Detector J—p

 

 

 

 

Thermal Conductivity detector (I-lot wire filament)
[Wheatstome Bridge]

 

 

 

 

KJ

Quartz tube

Figure 3.2 Schematic diagram illustrating the working principle of pulse
Chemisorb 27 00

,- 28
A heating mantle was used to heat the quartz tube containing the sam-

ples. It is essential here to understand how the Thermal conductivity
detector (Wheatstone bridge network) is capable of detecting the differ-
ence in thermal conductivity of gas that enters the Quartz tube and the
gas that leaves the tube. The Wheatstone bridge network consists of a hot
wire filament and the filament is powered by a carefully controlled con-
stant current. When a gas is flown over the hot filarnemt the temperature
of the filament changes. It should be noted here that the resistance of the
filament changes with temperature. Thus the change in resistance of the
filament is controlled by the thermal conductivity of the gas. Hence the
resistance of the filament on the way of incoming gas and the resistance
of the filament on the way of outgoing gas is going to be different. This
difference in resistance of the filaments manifest itself as difference in
voltage. In conclusion. the difference in thermal conductivity of gas that
enters the Quartz tube and the gas that leaves the tube is given in the
output signal as difference in voltage.

There is a display meter which indicates detected output signal.
When the analysis gas is more conducting than the carrier gas. depress-
ing the + REL condition push-button below the meter results in a
positive display. Using a graphic plotter the detected output signal was

recorded with respect to time.

Using the above technique desorption curve (qualitative) was
obtained for

1. Steel

2. Al-6061-T6

3. Al/SiC (Vf = 10%).

29
Steel was first used to test the sensitivity of the instrument for

hydrogen desorption. Desorption curves of as received sample and hydro-
gen charged sample were obtained so that they can be compared. The
samples were heated to a temperature of about 718 K. Rate of heating
was constant for all the samples. The temperature selected was high
enough for hydrogen to desorb. Desorption curves [as received and
hydrogen charged] for A1-6061-T6 and Al/SiC (Vf = 10%) were also
obtained.

Chapter 4

4.0 Experimental Procedure
4. 1 Materials

Three materials have been used in this study. They are steel. Al-
6061-T6 and Al/SiC composite. The SiC particulate composite was sup-
plied by Dural. Two different volume fractions of SiC were considered 10
and 20 pct. The matrix material used is 6061 and is processed by stir
casting. The details of the materials are shown in Table 4. 1.
4.2 Cathodic Charging of Hydrogen

To study the effect of internal hydrogen on any properties of materi-
als. hydrogen must be introduced into the material electrochemically
[cathodic charging] or thermodynamically [gas-phase charging] depending
on the type and condition of the material charged. Our aim is to charge
hydrogen in aluminum composites where the presence of passive A1203
layer hinders the diffusion of hydrogen into the material. This problem is
more pronounced in gas-phase charging than in cathodic charging.
Hence cathodic charging is selected as a method to charge hydrogen into
the desired material.

Cathodic charging was accomplished using an acidic electrolyte.
For steel acidic electrolyte was made of 5% by volume sulfuric acid solu-
tion plus 0.25ng liter A5203 as a recombination poison [28]. For A1-
606 1 -T6 and Al/SiC composite materials acidic electrolyte was made of
HCl [pl-ll] plus 0.25gm/liter A5203 as a recombination poison. The spec-
imens were cathodically charged for desired hours under an applied con-
stant potential of - 1 500mV Vs standard calomel electrode. The current
measured between sample and the anode was between 0.1 to 0.4 A. The

current usually started at lower current and increased slowly over the

30

31

Table 4.1 Material used in this study

 

 

 

 

 

 

 

 

 

 

 

Reinforc ”56°13
Supplier Matrix AHOY. -ement size Process Product
composrtron Vol% 111 form
microns
1. Dural 6061 Al-lMg- 20 SiC 25 Stir cast Rect
0.6Si-0.28 0.9 x 2.9 ”
2. Dural 6061 Al-lMg- 10 SiC 25 Stir cast Rect
0.6Si-0.28 0.9 x 2.9”
3. Steel
4. Al-606 1 -T6

 

 

32
duration of the charging. the increase was likely due to a change in the

pH of the solution.

The charging cell contained an anode which consisted of four plas-
tic posts. suspended from a plastic disk. A platinum wire anode was
wound outside the post in a single helix. The plastic disk was placed atop
a glass beaker which contained the charging solution. The specimen
which was the cathode. was suspended from a hole in the center of the
plastic disk. The specimen was then in the center of the platinum anode
helix. The entire arrangement is shown in Figure 4.1. In the set-up a
three electrode cell arrangement is used. The potentiostat is operated in
the comtr.E mode. the instrument acts as a potentiostat, causing the
potential selected by the applied potential/ current controls to be main-
tained between the working and reference electrodes by varying the volt-
age applied to the counter electrode. The counter (auxillary) electrode can
be any chosen electrode desired. because its electrochemical properties
do not affect the behavior of the working electrode (sample): it is usually
chosen to be an electrode that does not produce substances by electroly-
sis that will reach the working electrode surface and cause interfering
reactions there. Hence in this set-up platinum is chosen as the auxillary
(counter) electrode. The electrometer probe monitors the reference elec-
trode potential. It should be understood here that working electrode
. (sample) is maintained at either virtual or hard ground (zero volts) by the
instrument. That is to say. although reference is frequently made to a
potential applied to the working electrode. in actual fact the working elec-
trode (sample) is maintained at zero volts. and an Opposing potential is
applied to the counter electrode (platinum) and detected by the reference

electrode.

33

 

 

 

 

 

 

 

 

 

 

 

 

Potentio stat
Aux . W. E . Ref .
/’ 1
Electrometer Probe
CE Reference
["1 } E] El 1’8 electrode
(9 __...“Lucitc” post
1 ‘\

 

 

>§EJ :1: __>21iterbeaker
/

 

 

 

 

 

 

 

 

 

./1

 

>Pt wire cage

 

(a)

 

Figure 4.1 a). Schematic diagram illustrating the cathodic charging set-
up . b) Photographs shows the cathodic charging set-up.

35
Frequently some parts of sample were painted with etch-resistant lac-

quer. Since etch resistant lacquer is a mom-conducting lacquer the H+
ions will not go to that surface for reduction. An example of this would be
the tensile sample where the non-gauge section were painted with lac-
quer. This would increase the availability of H+ ions on the gauge section.
Hence the sample is charged with hydrogen more effectively. After charg—
ing. before desorption. the samples were rinsed in acetone. Prior to
charging the specimens were polished with 600 grit paper so as to
decrease the thickness of passive A1203 layer and specimens were also

cleaned with acetone.

After charging. the samples were removed. rinsed in acetone and
them immediately they were loaded in MTS system for tensile testing (or
any other test performed) or taken in the quartz tube to perform the des-
orption tests.

4.3 Desorption experiments
As mentioned earlier Chemisorption instrument was used to get
desorption curves for both charged and uncharged samples. Small bars

of samples of

Length : 1.85”
Width : 0.046”
Thickness : 0.037”

were used. For both charged and uncharged samples. the samples were
polished with 600 grit paper so as to minimize the thickness of A1203
layer. The samples were also cleaned with methanol.

The following procedure was followed to obtain the desorption

36
curves. Flow controller was switched ‘on’ and the N2 cylinder was

opened. Regulator pressure was set to off. Now 2700 was switched on.
The system was purged with N2 to remove any air that might be in the
system. The detector output was allowed to stabilize for few hours. Now
the sample was loaded in the quartz tube. Nitrogen flow rate was
adjusted to 5 (10 cm3/mim). The set-up was allowed to stabilize for 15 to
20 minutes. Peak area was selected to X10 so as to have high sensitivity.
Temperature was adjusted to 723K. Chart recorder was turned on to
record the desorption curve obtained. Temperature. detector signal out-

put amd corresponding time were also recorded.

4.4 Characterization of Al-SiC Composite and Decohesion Experi-
ment

To study the composite material as such and to study the interfa-
cial decohesion both charged and uncharged samples were strained to the
same extent. polished using diamond compound on a lapping wheel and

studied using a scanning electron microscope.

4.4. 1 Mounting the Tensile Samples

After straining to the same extent both charged and uncharged
samples were mounted on epoxy resin. The samples were mounted on
epoxy resin to facilitate polishing. The sample to be mounted was placed
on an aluminum foil and was sprayed with mold release. This facilitate
the removal of the sample from the epoxy mount after polishing. A ring
made of Lucite was placed surrounding the tensile sample. this served as
a mould wall. Epoxy resin and hardener were mixed thoroughly in the

ratio of 1:1 and poured in the ring before it got cured. Time taken to cure

37
to solid state was 9 to 12 hours.

4.4.2 Polishing

The epoxy mounted sample was polished using diamond paste on a
lapping wheel for several hours. It should be noted here that the metal-
lography preparation of particle reinforced metal matrix composites is a
difficult process. This is because of the extreme difference in hardness of
the SIC and Al-matrix material. The samples were first polished with 6
micron diamond paste for two hours. then the samples were polished for
several hours using 1 micron diamond paste on nylon cloth. Compound
thinner was also used. After polishing the samples were rinsed using

methanol and then taken out from the epoxy mount.

4.4.3 Microstructural Characterization

The polished samples were coated with gold. The gold coated sam-
ples (both charged and uncharged) were studied using Scanning electron
microscope for decohesion of interrnetallic particles. decohesion of rein-

forced particles and general microstructure characterization.

4.5 Mechanical Testing

All mechanical tests were performed using Instron or using MTS T” servo-
hydraulic test system. A three point bend test fixture was designed so
that samples of very small dimensions can be tested (Figure 4.2). To get
accurate strain values an extensometer was used during tensile test.
Tests were performed at room temperature at a crosshead speed of 0.05

inch/ min. After testing the fracture morpholoy was studied using scam-

ning electron microscope. Table 4.2 gives the detail about sample dimen-

38

sions used for both tensile and flexural tests.

Table 4.2 Sample dimensions used corresponding to various tests are tab-

ulated.

 

 

 

 

 

 

 

 

 

Type of Sample ID Nominal Sample Material
testrn g dimensrons used
Flexural test 81 and 8—2— Depth = 0.0566” Steel
Width = 0.4559”
Span = 0.6409”
Flexural test F1 and F2 Depth = 0.064” Al/SiC
Width = 0.455” (Vf = 20%)
Span = 0.640”
Tensile test B-1,B-2 Gauge length: 0.607” A1-6061-T6
and B-3 Mdth = 0.5”
Thickness = 0.088”
Tensile test C1,C2,C3 Gauge length: 0.607” Al/SiC
and C4 Width = 0.5” (Vf = 10%)
Thickness = 0.068”
Tensile test C5 Gauge length: 0.607” Al/SiC
VVrdth = 0.5” (Vf = 10%)
Thickness = 0.062”
Tensile test C6.C7,C8, Gauge length: 0.75” Al/SiC
C9 and C10 Mdth = 0.5” (Vf = 20%)
Thickness = 0.065”

 

 

 

 

 

39

 

 

 

 

 

 

 

 

 

 

 

Figure 4.2 Design of the three point bend test fixture.

Chapter 5

Results and Discussion

5.1 Cathodic Charging of Hydrogen

To make sure that the charging setup worked properly. steel sam-
ples were charged with hydrogen and flexural test were performed imme-
diately using a three point bend test. The results obtained were
compared with the uncharged samples of the same dimension. An indica-
tion that charging process is sound is recognized by the presence of
hydrogen bubbles around the sample. Two steel samples were simulta-
neously heated to austenizing temperature range and then quenched in
water. This process was done to make the steel more brittle. One of the
steel sample was charged with hydrogen and flexural test was performed
immediately. The result obtained was compared with the uncharged steel
sample. Flexural strain and flexural strength were dramatically reduced
for steel charged with hydrogen indicating that the charging setup works
properly. Uncharged steel (no heat treatment) was ductile that it did not
even crack as opposed to that of the charged steel. Figure 5.0 shows the
crack formed in the steel charged with hydrogen.

Steel was selected because steel being a more open structure is
more susceptible to hydrogen embrittlement when compared to that of
aluminum. Desorption curves and mechanical test results obtained
clearly shows that hydrogen can be charged easily into steel when com-
pared to aluminum. The presence of the A1203 layer in case of aluminum
acts as a barrier for hydrogen diffusion further reducing the amount of
hydrogen charged.

40

41

 

(b)

Figure 5.0 a). Macrograph shows the crack formed in steel charged with
hydrogen (Magnification 10X). b) Macrograph indicates the absence of
crack in the uncharged steel (Magnification 10X).

- 42
Cathodic charging resulted in a corroded surface layer. Cathodic

charged surface of Al/SiC composite was examined using a scanning elec-
tron microscope. The micrograph obtained (Figure 5. 1) clearly shows the
hydrogen-induced corroded surface layer. The spectrum acquired using a
emery dispersive analysis of X-ray (EDAX) indicated the presence of alu-
minum and chlorine in the corroded surface layer (Figure 5.2). Figure 5.1
shows that the corrosive attack follows mainly the grain boundaries. there
was also some indication of transgramular cracking. The corroded sur-
face layer was removed by polishimg indicating the thickness of corroded
surface layer is small. Albrecht and coworkers measured the thickness of
corroded surface layer to be around 0. 15mm [40]. The result obtained
shows that the surface of Al/SiC composite is corroded as a result of
cathodic charging similar to that of the results obtained by Albrecht and
coworkers [40] in high strength aluminum alloys. Typical of cathodic
charging is extensive cracking of corrosion products on the surface of the

composite material.

43

 

Figure 5.1 a) Hydrogen -induced corroded surface layer of Al/SiC
composites.

44

 

 

c>

 

 

\\

2.56 KeV

Figure 5.1 b) Acquired X-ray spectrum for the corroded surface
layer.

45
5.2 Characterization of A1/SiC Composite and Decohesion Experi-

ment

Al/SiC composite consists of matrix metal aluminum and the reinforced
particles of SiC. The addition of SiC to aluminum and it alloys confers
substantial improvements in specific modulus and strength. However.
these improved properties are associated with decrease in ductility and
toughness compared with aluminum alloy. The decrease in ductility is
evident by comparing the stress-strain curves obtained for Al-6061-T6
with that of stress-strain curves obtained for Al/SiC composites. Figure
5.2 shows the microstructure of Al/SiC composite. The microstructure of
Al/SiC composite also exhibited uniform precipitates of MgZSi in the bulk
matrix as shown in Figure 5.3 a. It should be noted here that SiC parti-

cles are angular where as MgZSi particles are circular and small.

The failure mechanism observed for the above composite material
was similar to the failure mechanism proposed by an investigation con-
ducted by Lucas and others [41]. First the SiC particulates fail by
transparticle fracture. Figure 5.3 b shows cracked SiC particles. Frac-
turing of particulate is the primary fracture processes and may be consid-
ered as stress controlled process. Figure 5.4 clearly shows such voids
formed. Failure of the matrix ligament is a strain controlled process and
occurs by microvoid coalescence. The strain controlled process is influ-

enced to a great extent by void nucleation associated with particulates.

Decohesion experiment was performed to study how hydrogen

influence the decohesion at particle matrix interface. Tensile coupons of

46
composite material (both charged and uncharged) were strained to the

same extent. polished using diamond compound in a lapping wheel. and
observed for void nucleation at the interface using scanning electron
microscope. Polished sample did not give rise to a flat surface between
the reinforcement and matrix hindering the proper analysis of decohesion
at the particle / matrix interface. No clear decohesion was observed at the
particle matrix interface for both the charged and the uncharged samples.
This indicates there is good bond strength between SiC and aluminum
matrix. Florn and Arsenault [42] calculated the interfacial bond strength
and inferred from their calculation that debonding of SiC particulates
from the matrix (Al-6061-T6) to be a rare event. Interface decohesion will
be a rare event if the bond strength is very high when compared to the
yield strength of the matrix. No interfacial decohesion was observed by
Roebuck [43] in Al-6061/SiC composite material.

Lucas and coworkers [41] clearly show the interface decohesion for
A201 /SiC metal matrix composite. Thus debonding of SiC from the
matrix depends on the type of matrix material used. Polishing was done
to remove the corroded surface of the charged sample so that microstruc-
ture can be viewed and the influence of hydrogen on the decohesion
behavior of the particles from the matrix can be studied. It should be
noted here that by polishing one is also removing the mostly heavily
charged material. Figure 5.4 shows the ductile fracture occurs by void
nucleation at SiC particulates and also at the matrix imperfections

(Mgzsi. inclusions).

47

 

a baa—camps .
a m o wOQEoo 2 0 pg v :2 N m ousmrm

 

 

,n .._- fl-.—-—_————

 

48

6203.53 9m toxemia 96% £3quwa 3 .02me x39 65 E “mum: .8 9852985 Each—GD H mm 259m

3. 3

3m. Won "Em,

 

49

NM 2; 3 X: 0.3 .6 $65qu 0836qu 65 55 “382683 539% 20> 96am Tm pun—ME
3. E

 

50
5.3 Fracture Surface Morphology

Fracture surface morpholoy was studied for

l. 10% SiC charged with hydrogen for 6% hours.
2. 10% SiC charged with hydrogen for 12 hours.
3. 20% src charged with hydrogen for 6 g hours.
4. 20% SiC charged with hydrogen for 19 hours.

Fracture surface morpholoy showed the same mode as reported by
Lucas and others [41]. It should be noted here that though Lucas and
others studied fracture toughness of the composite using a chevron notch
short bar specimen the same fracture phenomenon applies to failure of
the composite under tension. The fracture morpholoy is characterized
primarily by deep voids centered around the SiC particles. Ductile failure
occur on the matrix ligaments of the fracture process zone by microvoid
coalescence in the matrix. Failure in the matrix ligament is a strain con-
trolled process. The ligament between the cracked particles ahead of the
main crack failed mostly by shear and microvoid coalescence in the
matrix. Fracture of SiC occurred in a multifaceted manner. SiC fracture
in a multifaceted manner because of the presence of pre—existing flaws in

the particulates.

Figure 5.5 shows transparticle fracture of SiC and shows the pres-
ence of large deep void around the SiC particles. The above micrograph
also shows microvoid coalescence of the matrix. Fracture mechanism is of
the one explained by Ritchie [44]. Failure has been modeled as microvoid
coalescence using a stress modified strain criterion originally proposed by
McClintock [45] and adopted by Mackenzie and coworkers [46]. The dim-

51
ple rupture evident is initiated by void nucleation and growth at the SiC

reinforcement ahead of the crack tip. The subsequent growth of such
voids becomes limited by a plastic shear instability between neighboring
voids and the main crack tip. the linkage constitutes crack growth. The
instability is caused by localization of strain between large voids and
occurs by a fine scale coalescence of microvoids nucleated at constituent
phase inclusions in the matrix ligament. Such ductile failure of the
matrix can be said as strain induced. in that macrocrack growth by link-
ing of void occurs when critical strain is exceeded ahead of the crack tip.
At first. McClintock [45] pointed out that the void growth during ductile
fracture is also strong function of the ratio of mean to equivalent flow
stress. It also requires the failure criterion for ductile fracture should
involve critical strain to be reached over a minimum amount of material
rather than at a single point ahead of the crack tip. Accordingly. Macken-
zie and others [46] have proposed a criterion for ductile failure where a
critical fracture strain (2;) is exceeded over a characteristic distance (10')
ahead of the crack tip. It is important to note that critical strain is a func-

tion of stress state.

Figure 5.5 also shows the presence of extensive plastic deformation
on the inside of the void wall as a result of intense slip band intersection
(charged for 6 % hours). Figure 5.6 shows a closer look at the slip band.
Figure 5.7 shows the fracture surface of the uncharged Al/SiC composite.
The micrograph shows a cluster of SiC and presence of slip bands on the
inside of the void walls. In general. fracture surface morpholoy showed
insignificant differences between the charged and the uncharged sample.

This is to be expected considering the lack of internal hydrogen penetra-

52

tion into the specimen.

The fracture surface micrograph for both the charged and the

uncharged sample shows good bond strength between the SiC particles
and the aluminum matrix. No fractured interfaces were observed
between A1 and SiC. Fracture surface showed insignificant difference
between the charged and the uncharged samples. It is important here to
compare this result with that of the result obtained by You and coworkers
[47]. They studied the effect of hydrogen in P/M 2124-20 pct SiCp com-
posites and found that void formation and decohesion was observed at
the interface. Discrepancy in results is due to
1. You and coworkers [47] used dynamic charging (straining electrode
test) of hydrogen. This involves both straining and simultaneous charg-
ing. Straimimg causes dislocation sweeping and dislocation sweeping
results in deep hydrogen penetration. deeper than by lattice diffusion.
2. No fractured interfaces was observed even for the uncharged sample in
case of Al 6061/SiC composite material whereas some instances frac-
tured interface was observed in Al 2124/SiC composites. This indicates
that bond strength between SiC and Al-6061 is good.

Fracture surfaces of charged samples (6 to 12 hours) indicated lack
of internal hydrogen penetration into the specimen. To make sure of this
observation sustained charging time of 19 hours was used to charge All
SiC (Vf = 20%). Figure 5.8 compares the charged and uncharged samples
of Al/SiC (Vf = 20%). The fracture morpholoy again indicated insignifi-
cant differences confirming the lack of internal hydrogen penetration.
Fracture morpholoy was also analyzed at various position of the fracture

surface.

.350: mm www.883— 20>
2.: Co 622.: 65 :o 9.83me mm 6395866 oamma 9,6586 96% schism 33095 .3 6.50: mm commune.
KENS on: no someone—moo 2665:: can manna mam dugout 205.8%:65 MERE? annuwouoaz .3 m.m Pam—rm

3

Wrap. «x, axon

I ‘

 

 

 

 

 

 

 

 

 

Goawwom Coaumom

54

6638in 55m
E ed .
n m m Esau he came mam Eaton 96
em ch 8%
E

 

55

Anemone»: 05 8568535 woman 3? own—BS
Lo :52 m we was» 20> 05 no 2:2: 65 so 3:61.85 v.8 mac—32.3 ma outcome we? nouns—Cave cams—a
2,6:me .3 Acumen»: 65 connogcwwfi upswE um 9w .«o .8526 w macaw 835m 8368“ BE. .8 En ousmfi

 

 

 

 

 

 

 

 

 

O
GO—uumom douumcnu

56

 

._£ 8358 now8nocs 05 E8 3.52.— 9 8b commune 65 Son .8.“ 268m o5 655.8 539% 26> p8
5588 65 .6 ooGoomoROo 9068:: £265.82“ 9m .8 52:5 8 965m awe—0:985 coat—.5 838$ Qm 0.53m
8w

 

 

 

 

 

O
sou—mom

57
5.4 Thermal Desorption Experiment

Desorption curves were obtained for steel. Al-6061-T6 and Al/SiC
composites. Figure 5.10 shows that the amount of desorption signal
obtained for steel charged with hydrogen is one hundred times stronger
than that of the uncharged steel. This indicates that the Chemisorb 2700
is pretty sensitive for desorption of hydrogen. Figure 5.9 shows the des-
orption signal obtained by heating the empty quartz tube which can be
used as a base reference. Figure 5.12 indicate that the amount of desorp-
tion signal obtained for aluminum charged with hydrogen is only slightly
higher than the uncharged aluminum. This behavior can be explained by
comparing the diffusivity of aluminum (closed packed structure) and steel
(open structure).

Ishikawa and Mclellan [48] calculated the diffusivity of hydrogen in
99.99% aluminum at room temperature by permeation method. On the
basis of their experiments they calculated diffusivity

D = 9.2x10'5 exp - (965Tl—‘43) m2/S.

Volkl and Alfeld [43] found the diffusivity of hydrogen in iron to be

D= 2.0 x 10 '7

exp 4%- ) mZ/S.

The two important results that are obvious from the desorption
curves are

1. Quantity of desorption signal obtained for Al-6061-T6 and Al/
SiC is very low when compared to steel.

2. Desorption signal obtained for Al-6061-T6 and Al/SiC (charged

58
ones) indicate the absence of internal hydrogen in the charged samples.

It is important here to discuss the reason behind the lack of inter-
nal hydrogen penetration. As discussed in the background diffusivity of
hydrogen in aluminum by bulk means is very low. The various trap sites
like dislocation cores. grain boundaries. lattice substitutional sites and
particle-matrix interface will further decrease the diffusion of hydrogen
into the material. Although the trap sites decrease diffusion of hydrogen
in the material the local hydrogen concentration can be increased in the
sample significantly higher than pure lattice. The major mechanism by
which hydrogen can be transported deeper into the material is by disloca-

tion transport.

Albrecht [35] used straining electrode test to study 707 5 aluminum
alloy. They concluded that combined straining and cathodic charging
enhance embrittlement far in excess of that possible by cathodic charging
alone. From their findings they also concluded that dislocation transport
of hydrogen is kinetically preferred to volume diffusion. This observation
was consistent across a wide spectrum of aluminum alloys such as 2124
[39] and 7050 alloy. In static charging where there is no straining hydro-
gen diffuses by bulk means rather than by dislocation transport. The
lack of internal hydrogen penetration is attributed to the lack of effective
dislocation transport in static charging of hydrogen.

Figure 5.13 compares the desorption curves obtained for the
charged and the uncharged samples of Al/SiC (V; = 10%). Desorption sig-
nal of the charged sample is very high initially (for first ten minutes) and

59

then drops to the level of uncharged samples. The desorption curve of the

uncharged is very low initially and then increases after ten minutes and

again drops. This indicates that the hydrogen is charged only on the sur-

face of the sample. Table 5.1 gives all information of the material used

and the corresponding sample identification.

Table 5.1 Sample dimensions. charging conditions and material used are
shown for various desorption curves obtained.

 

 

 

 

 

 

 

 

 

Thickness = 0.037”

 

Uncharged

 

Sample ID Sample Dimensions Hydrogem/ No H2 ”43:21:61]
= : L =:_—
Figure 5.10 Length = 1.85” Charged (12 Steel
Width = 0.046” hours) 8:
Thickness = 0.037” Uncharged
DSl (Figure Length = 1.85” Charged (12 Steel
5.1 1) Width = 0.046” hours)
Thickness = 0.037”
DS2 (Figure Length = 1.85” Charged (12 Al-6061-T6
5.12) Width = 0.046” hours) &
Thickness = 0.037” Uncharged
D83 (Figure Length = 1.85” Charged ( 12 Al/SiC
5.13) Width = 0.046” hours) & (Vf = 10%)

 

 

 

 

 

 

0.08 I I 7 I I I I I
0.06 ) -
8
.E
59.
n
o
a
Sr .
'5 0 04 '-
C
.9
9-
8
o
D O
0.02 - "
0O 1 0 20 3O 4O 50 60 70 80 90

T1me in minutes

Figure 5.9 Desorption signal obtained by heating the empty quatrz tube to
723 K.

Desorption curve for steel charged with hydrogen

6]

 

0|
1

 

 

.5
I

 

(O
I

 

Desorption signal obtained

[0
1

 

 

fi

I

I

Desorption cum obtained tor 31mm hydrogen)

 

012 f

?

Desorption signal obtained
0
3

0.02 r

 

 

 

 

50

0C!

100 150

Turn. in «unites

.1

250

 

 

0 20 40

60
Time in minutes

100

120

Figure 5.10 Compares the desorption curves obtained for charged and

uncharged steel samples.

62

Desorption curve for steel charged with hydrogen

 

 

 

 

 

 

8 I I I I I I I l I
7’“ .
6b -
B
.5
£95- 4
.0
O
'76
6
.m4- .
C
.9
§3~ .
U)
8
2’ -(
Sample i.d : 081
1- .
O _ l

 

0 2O 4O 60 80 100 120 140 160 180 200
Time in minutes

Figure 5. 1 1 Desorption curve obtained for steel charged with hydrogen
[Sample i.d: DSI].

0.25

0.2

Desorption signal obtained

0.05

63

Desorption curve for Al-6061-T6

 

I

 

 

I T I I I I I I

 

I

-'- charged with hydrogen
- No hydrogen
Sample i.d : D82

 

l l l l l

10 20 30 40 50 60 70 80
Time in minutes

Figure 5. 12 Compares the desorption curves obtained for charged and
uncharged Al-6061-TB samples.

 

90

Desorption curve for AVSiC (volume fraction: 10%)
0.4 I I I I l

 

I

0.35 “ r

o
in .0
or to

Desorption signal obtained
0
N

 

 

 

 

 

 

0.15
—X
0.1 -
'1 -'- charged with hydrogen
0'05 I. - No hydrogen '
5.. Sample i.d : 083
i" 1 i 4 4 1
o 20 40 so so 100 120

Time in minutes

Figure 5. 13 Compares the desorption curves obtained for charged and
uncharged Al/SiC “’1‘ = 10%) samples.

65
5.5 Mechanical Testing

To study the effect of hydrogen in Al/SiC composites. tensile tests and
flexural tests were performed. The results obtained are shown in Table

5.2 and Table 5.3. The stress-strain curves obtained are shown in

1. Figure 5.15 to figure 5.16 (flexural tests).
2. Figure 5.17 to figure 5.23 (tensile tests).

It is clearly evident from the stress-strain curves obtained for Al/SiC
composites. the hydrogen effect was manifested only in the final ligament
fracture of the samples. The ligaments in the charged samples with a cor-
roded surface showed a brittle behavior as opposed to that in the
uncharged samples which exhibited ductile behavior. Since only the final
ligament fracture behavior was altered it clearly shows the lack of internal
hydrogen in the composite samples. Effect of surface hydrogen on the
final failure strain% for Al-6061-T6 and Al/SiC composites (tested in

transverse direction) is shown in figure 5.14.

Ductility of SiC/A1 composites. as measured by strain to failure is
reduced with increasing reinforcement content. Figure 5.14 shows the
effect of reinforcement content on the behavior of composite material. For
uncharged A1/SiC Wf = 20%) the average final strain is 8.1%. approxi-
mately 2% less than that of Al/SiC (Vf = 10%). Composite specimens with
low reinforcement contents of 10% volume fraction SiC mostly showed
(macroscopically) 45 degree chisel point type of fracture across the width
of the specimen. At reinforcement contents (~ 20 volume percent) a 45

degree chisel formed at one edge and extended about half-way through

66
the width of the specimen but then became flat.

This is similar to observation made by McDanel (49] showing the
relation between reinforcement content and fracture modes. By compar—
ing Figure 5.20 and Figure 5.22 one can see that the ultimate strength
increased with increasing reinforcement content. Considering the small
thickness of samples used it is understandable that there is only small
increase in ultimate strength. By carefully examining the two figures one
can also see that as the reinforcement content is increased. the stress-
straim curves enter the plastic flow region at a higher flow stress. David
McDanel [49] indicates that the strength increase was probably caused by
closer packing of the reinforcement and smaller interparticle spacing in
the matrix. Although the stress-strain curves for Al-6061-T6 were repro-
ducible. for Al/SiC composite a broad range of strain was obtained. This

is due to the following reasons

1. Unavoidable differences in the surface preparation.

2. Given the small thickness of the samples the difference can be
attributed to the difference in distribution of reinforcement parti
-cles (SiC).

3. Unavoidable differences in the quantity of hydrogen charged.

The results obtained indicate that the difference between the charged and
uncharged samples (both for Al-6061-T6 and AI/SiC) is due to hydrogen-

induced surface corrosion phenomenon.

67

Table 5.2 Flexural test results obtained are shown for both the charged
and the uncharged samples.

 

 

 

 

 

 

 

 

 

 

 

Hydrogen/ Final .
Sample ID No fracture Pemgss Maggral
hydrogen strain% 11
F1 Uncharged 5.0 308 Al/SiC
(Vf = 20%)
F1 _. Charged 4.75 238 Al/SiC
(Vf = 20%)
F2 Uncharged 5. 1 340 Al/SiC
(Vf = 20%)
F2 Charged 4.4 209 Al/SiC
(Vf '—" 20%)

 

 

68

Table 5.3 Tensile test results obtained are shown for both the charged
and the uncharged samples. Corresponding sample i.d and material used
are also tabulated.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Hydrogen] Final .
Sample ID No 3:3 fracture Msézral
hydrogen strain%

B-l Uncharged 49.62 17.06 Al-6061-T6
B-l Charged 49.87 15.86 A1-6061-T6
B-2 Uncharged 49.62 17.67 Al-6061-T6
B-2 Charged 49.62 16.10 Al-6061-T6
B-3 Uncharged . 49.62 17.3 A1-6061-T6
B-3 Charged 49.38 16.3 Al-6061-T6
C1 Charged 17.93 7.28 Al/SiC

(Vf = 10%)
C2 Uncharged 17.44 9.39 Al/SiC

(Vf = 10%)
C3 Charged 18.4 10.68 Al/SiC

(Vf = 10%)
C4 Uncharged 18.4 11.98 Al/SiC

(V,~ = 10%)
C5 Uncharged 16.46 8.53 Al/SiC

(Vf = 10%)
C6 Charged 18.18 5.94 Al/SiC

(Vf = 20%)
C7 Uncharged 18.42 7.67 Al/SiC

(Vf = 20%)
C8 Uncharged 18.91 8.62 Al/SiC

(Vf = 20%)
C9 Charged 18.18 7.41 Al/SiC

(Vf = 20%)
C10 Charged 18.42 7.5 Al/SiC

(Vf = 20%)

 

 

 

 

 

 

Percentag. Final Failure Strain

69

 

I Uncharged (lensile lest)

 

““ Charged (Tensile lest)
I Uncharged (Flexural lest)

 

:3; Charged (Flexural lest)

 

 

 

 

 

 

 

Al- Al/SiC Al/ 51C
0%1- (V .f = (V .f =
16 10%) 20%)

Figure 5.14 Effect of Surface Hydrogen on the Percentage Final Failure Strain
for Al - 6061 - T6 and AI/SiC Composites (Tested in the Transverse Direc-
tion)

7O

Stress-Strain curve for AI/SiC (Volume fraction = 20%)

 

 

 

 

350 I I I I j I I I I
300 - ~
250 - ~
Ezoo- .
E.
«n
3
(‘5 150 - .1
100 - -"- Charged with hydrogen (6hrs) -
- No hydrogen
5° ' Sample i.d : F1 -
. J v 1 I l l l l l l
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

% Strain

Figure 5.15 Flexural Stress-Strain curves obtained for Sample F1 [Al/SiC
(Vf = 2096)].

 

7i

Stress-Strain curve for Al/SiC (Volume fraction = 20%)

 

350 r u u 1 u

250

 

 

 

 

F5200 .
2.
3
2
(,3 150 ~
100 ~
-'- Charged with hydrogen(6hrs)
50 I - No hydrogen q
Sample i.d : F2
. l l l L
O 1 2 3 4 5 6

% Strain

Figure 5.16 Flexural Stress-Strain curves obtained for Sample F2 [Al/SiC
(Vf = 20%)].

Stress[ks11

72

Stress-Strain curve for Al-6061-T6

 

 

 

 

 

25 - -

20 q

15 . . ..
- - Charged With hydrogen (12hrs)

10' - No hydrogen ‘
Sample i.d :8-1 .

O l l l l A 1 l l l
O 2 4 6 8 1O 12 14 16 18 20
% Strain

Figure 5.17 Stress-Strain curves obtained for Sample B-l [Al-6061116].

 

73

Stress-Strain curve for Al-6061-T6

 

 

50

I

45

I

35

30"

Stress[ksi]
N N
O 01

.a
U1

 

.5
C

-r- charged with hydrogen (12hrs) ‘
- No hydrogen

Sample i.d : B-2

 

l l l l l l 1 l

 

2 4 6 8 1O 12 14 16 18
o/o Strain

Figure 5.18 Stress-Strain curves obtained for Sample B-2 lAl-6061-T6].

Stress[ksi]

 

 

74

Stress-Strain curve for Al-6061-T6

 

 

 

50 --------------- a, .. . . ,
45 - -
4O - -
35 - l -
30 / T
15 [If] -‘- Charged with hydrogen (12hrs) a
" - No hydrogen
1 o- J
Sample i.d : B-3
0 l 4 l l l
0 8 1o 12 14 16 1e
% Strain

Figure 5.19 Stress-Strain curves obtained for Sample B-3 [Al-6061118].

Stress-Strain curve for Al-SiC(volume fraction-10%)

 

20

16"

.a
.5
I

.5
N
I

 

C1

Stress[ksi]

.a
o
I

*7
ch

 

6 — No hydrogen

4 _ -'- Charged with hydrogen (12hrs) .

 

 

 

% Strain

Figure 5.20 Compares the Stress-Strain curves obtained for charged and
uncharged samples of All 810 (V,- = 10%) tested in transverse direction.

Stress[ksi]

18

76

Stress-Strain curve for Al-SiC(volume traction=10°/o)

 

 

I I I I j I I I

 

- No hydrogen

 

6
Sample i.d : C-5
4 l L i l L J l
0 1 2 3 4 5 6 7 8

Figure 5.21 Stress-Strain curve obtained for Sample CS [Al/SiC (V1-
=lO%)] tested in transverse direction.

 

Stress[ksi]

77

Stress-Strain curve for Al-SiC(volume iraction=20%)

 

I

 

 

 

'10

c1

Charged with hydrogen (18hrs)
.. No hydrogen

 

l

2 3 4 5 6 7 8 9
% Strain

-
p
-

Figure 5.22 Compares the Stress-Strain curves obtained for charged and
uncharged samples of All SiC (Vf =20%) tested in transverse direction.

Chapter 6
Conclusions and Future work

This study has shown several features of the interaction between hydro-
gen and Al-6061-‘I‘6 and All SiC composites.

1. A cathodic charging cell was designed so that hydrogen can be
charged into any material desired.

2. A thermal desorption technique was used to determine the
hydrogen concentration in steel. Al-6061 -T6 and Al/SiC composite mate-
rials. The desorption curves obtained indicate that high hydrogen con-
centration existed at the surface even for prolonged charging times (12
hours). This clearly indicates that cathodic hydrogen interaction with A1-
606 1 -T6 and Al/SiC composites is of a surface phenomenon rather than
that of the bulk.

3. Cathodic charging of hydrogen results in corrosion of the sur-
face layer of Al/SiC composites.

4. Extensive fracture morphology study was conducted. In gen-
eral, fracture surface morphology of Al/SiC showed insignificant differ-
ence between the charged and the uncharged samples. This signifies the
lack of internal hydrogen penetration into the specimen. The fracture sur—
face morphology study also gives a clear picture of the mechanism of frac-
ture of Al/SiC composite materials.

5. The surface hydrogen effect was manifested only by the unique

78

. 79
fracture behavior of the final ligament failure during mechanical testing.

This is clearly evident from the stress—strain curves obtained for Al/SiC
composites (both flexural and tensile tests). Due to hydrogen- induced
surface corrosion of charged samples, final ligament fracture was brittle
in nature as opposed to uncharged samples which exhibited ductile

behavior.

6. For Al-6061-T6 the difference in percentage strain to failure
between the charged and the uncharged samples ranged from 5 to 8%.

7. General microstructure characterization of Al/SiC composite
was conducted using scanning electron microscope. The various micro-
graphs obtained shows that fracture process of Al/SiC occurred by

3) Failure of the reinforcement by transparticle fracture.

b) Ductile failure in the interparticle ligament by void coalescence.

The following recommendations for future work are made as a

result of this investigation.

1. The experiments suggest that surface was playing an important
role in absorption and desorption of hydrogen. The experiments indicate
the effect of cathodic charged hydrogen on Al/SiC and Al-6061-1‘6 as a
surface phenomenon. It is imperative to characterize and standardize the
nature of the surface. The surface can be standardized by removal of the
oxide layer by sputtering or by electrochemical techniques. The surface
can be modified systematically and the effect of surface modification on

the behavior of the material to cathodic charged hydrogen can be studied.

80

Surface analytical equipment can be used to characterize the surface.

2. High temperature methods of charging hydrogen can be tried to
study the effect of temperature in hydrogen absorption in a composite

material.

3. A method to homogenize the sample charged with hydrogen
should be found out. This can help to understand the behavior of the

interface of composite material in presence of internal hydrogen.

4. Embedded Atom model can be used to study the behavior of

interface of composite material to internal hydrogen.

81

BIBLIOGRAPHY

l D. E. J. Talbot, “Effects of Hydrogen in Aluminum, Magnesium, Copper
and their Alloys“, International Metallurgical Reviews, Vol.20, 197 5,
p. 166.

2 R. J. Gest and A. R. Troiano, “Stress Corrosion and Hydrogen Embrit
tlement in an Aluminum Alloy”, Corrosion-NACE, Vol.30, No.8, 1974.
p.274.

3 H. H. Johnson and J. P. Hirth. “Internal Hydrogen Supersaturation
Produced by Dislocation Transport”, Met trans.. Vol.7A, 1976, p. 1543.

4 J. K. Tein, A. W. Thompson, 1. M. Bernstein and R. J. Richards,
“Hydrogen Transport by Dislocations”, Met. Trans., Vol.7A, 1976, p.821.

5 E. Hashimoto and T. Kino, “Hydrogen Diffusion in Aluminum at High
Temperature”, Met. Phy. Vol.13, 1983, p.115.

6 R. Eborall and C. E. Ransley, “The Reaction of an Aluminum -Magne-
slum Alloy with Water Vapor and the Adsorption of Hydrogen”,J our Inst
Met., Vol 71, 1945, p.525

7 C. E. Ransley and H. Neufeld, “The Solubility of Hydrogen in Liquid and
Solid Aluminum“, Jour Inst Met., Vol 49, 1948/ 1949, p.599.

8 W. Von Eichenauer and A. Pebler, “Measurements of the Diffusion Coef-
ficients and the Solubility of Hydrogen in Aluminum and Copper”. Zeit
Mett., Vol.48, 1957, p.373.

9 W. Von Eichenauer, K. Hattenbach and A. Pebler, “The Solubility of
Hydrogen in Solid and Liquid Aluminum“, Zeit Mett., Vol 52, (1961),
p.682.

10 K. Kikuchi and R. B. McLellan, “The Solubility and Diffusivity of
Hydrogen in Well-Annealed and Deformed Iron“, Acta Met., Vol.31, 1983,
p.961.

11 K. Papp and E. Kovacs-Cseteny, “Diffusion of Hydrogen in High Purity
Aluminum”, Scripta et., Vol.15, 1981, p.161.

82

12 T. Ishikawa and R. B. Mcllellan, “The Difusivity of Hydrogen in Alumi—
num”. Acta Mettall., Vol.34, 1986, p. 1091.

13 G. Solt, M. Manninen and H. Beck, “Lattice Relaxation Around Inter—
stitial Hydrogen in Aluminum”. Jour Phys F: Met Phy., Vol.13. 1983.
p. 1379.

14 Ritcher, “Nuclear and Electron Resonance Spectroscopes Applied to
Materials Science”, ed E. N. Kaufman and G. K. Shenoy (Newyork: North-
Holland).

15 D. S. Larsen and J. K. Norskov, “Calculated Energies and Geometries
for Hydrogen Impurities in A1 and Mg”, J. Phys F: Metal Phys., Vol.9 No.10
1979, p.1975.

16 J. Vokl and G. Alefeld, “Diffusion of Hydrogen in Metals”. Applied Top-
ics in Applied Physics : Hydrogen in Metals I. G. Alefeld and J. Volkl, eds..
Springler Verlag, Berlin, p.321.

17 Popovic and M. J. Scott, “Nonlinear, Self-Consistent Theory of Proton
Screening in metals Applied to Hydrogen in Aluminum and Magnesium”.
Phys Rev Lett.. Vol.33, 1974, p. 1 164.

18 Manninen. M and Nieminen R. M J. Phys. F: Met. Phys., Vol.9. 1979.
p.1333.

19 Kahn L. M, Perrot. F and Rasolt. M, Phys. Rev. B. Vol.17, 1980,
p.1592.

20 Perrot. F and Rasolt. M, Phys. Rev. B., Vol.17, 1980, p.1592.

21 A. K. Zurek, H. L. Marcus, J. N. Cecil and R. Powers, “SIMS Study of
Deuterium Trapping in Ion Implanted Aluminum Alloys”, Met Trans.,
Vol.1 1A, 1980, p.1920.

22 NR. Edwards and W. Von Eichenauer, “Reversible Hydrogen Trapping
at Grain Boundaries in Super pure Aluminum”, Scripta Met., Vol.14,
1980, p.971.

23 D. A. Hardwick, A. W. Thompson and I. M. Bernstein, “The Effect of
Copper Content and Microstructure on the Hydrogen Embrittlement of A1-
6Zn-2Mg Alloys”, Met Trans., Vol. 14A, 1983. p.2517.

24 G. M. Schamans. “Hydrogen Bubbles in Embrittled Al-Zn-Mg Alloys”.
Jour Mat Sci., Vol.13, 1978, p.27.

83

25 L. Cristoloudou and H. M. Flower, “Hydrogen Embrittlement and
Trapping in A1-60/o Zn-3% Mg”. Acta Met., Vol.28, 1980, p.481.

26 D. Nguyen, A. W. Thompson and I. M. Bernstein, “Microstructural
Effects on Hydrogen Embrittlement in a High Purity 7075 Aluminum
Alloy”, Acta Met., Vol.35, 1987, p.2417.

27 Asaoka, T. Dagbert, C. Aucouturier and J. Galland, J ., Scripta Met..
Vol.1 1A. 1977, p.467.

28 B. J. Berkowitz, J. J. Burton, C. R. Helms and R. S. Polizzotti, “Hydro
gen Dissociation Poisons and Hydrogen Embrittlement”, Scripta Meta,
Vol. 10, 1976, p.871.

29 N. E. Patan and 0. Buck, “The Effect of Hydrogen and Temperature on
the Strength and Modulus of Beta Phase Ti Alloys“, Effect of Hydrogen on
Behavior of Materials, Edt. Anthony W. Thompson and I. M. Bernstein.

30 B. Criqui, J. P. Fidelle and A. Clauss, “Effects of Internal and External
Hydrogen on Mechanical Properties of Beta III Titanium Alloy Sheet”,
Effect of Hydrogen on Behavior of Materials, Edt. Anthony W. Thompson
and I. M. Bernstein.

31 P. N. Adler, R. L. Schulte, E. J. Schneid, E. A. Kamykowski and F. J.
Kuehne. “Stress Induced Hydrogen Redistribution in Commercial Tita
nium Alloys”, Met. Transc., Vol.1 1A, 1980, p.1617.

32 J. W. Watson, Y. Z. Shen and M. Meshi, “Effect of Cathodic Charging
on the Mechanical Properties of Aluminum”, Met Trans., 1988.

33 M. Tehari, J. Albrecht, I. M. Bernstein and A. W. Thompson, “Strain
Rate Effects on Hydrogen Embrittlement of 7075 Aluminum", Scripta
Met., Vol.13, 1979, p.871.

34 M. Mueller, 1. M. Bernstein and A. W. Thompson, “Recovery Behavior
of Hydrogen Charged 7075-T6 Aluminum”. Scripta Met., Vol.17, 1983.
p. 1039.

35 J. Albrecht, I. M. Bernstein and A. W. Thompson, “Evidence for Dislo—
cation Transport of Hydrogen in Aluminum”, Met Trans., Vol.13A, 1982,
p.8 1 1.

36 J. P. Hirth and G. J. Dienes, “On the Rate of Equilibration of Diffu
sional Traps in Solids”, Acta Met., Vol.30, 1982, p.2061.

84

37 G. Frankel and R. M. Latinson, “Hydrogen Transport during Deforma—
tion in Nickel: part II. Nickel Single Crystals”, Met Trans., Vol. 17, 1986 B.
p.869.

38 J. Albrecht, A. W. Thompson and I. M. Bernstein, “The Role of Micro-
structure in Hydrogen Assisted Fracture of 7075 Aluminum“, Met Trans..
Vol.10A. p.1759.

39 D. A. Hardwick, M. Tehari, A. W. Thompson and I. M. Bernstein,
“Hydrogen Embrittlement in a 2000-Series Aluminum Alloy”, Met Trans.,
Vol.13A. 1982, p.235.

40 J. Albrecht, B. J. McTiernan, I. M. Bernstein and A. W. Thompson, “
Hydrogen Embrittlement in a High-Strength Aluminum Alloy”, Scripta
Metal, Vol.11, 1977, p.893.

41 J. P. Lucas, P. K. Liaw, J. J> Stephens and J. Nunes, “Effects of Par-
ticulate Fracture in Determining Fracture Toughness of Metal Matrix
Composites”,Morris E. Fine Sysposium, TMS Publication. 1991, p. 171.

42 Y. Flom and R. J. Arsenault, “Interfacial Bond Strength in an Alumi-
num Alloy 6061-SiC Composite”, Materials Science and Engineering,
Vol.77, 1986, p.191.

43 B. Roebick, “Fractography of a SiC Particulate Reinforced Aluminum
Metal Matrix Composites”, Journal of Materials Science Letters, Vol.6,
1987, p. 1 138.

44 R. O. Ritche, W. L. Server and R. A. Wullart, “Critical Fracture Stress
and Fracture Strain Models for the Prediction of Lower and Upper Shelf
Toughness in Nuclear Pressure Vessel Steels”, Met. Trans., Vol.10A. 1979.
p.1557.

45 F. A. Mc. Clintock, Fracture, An Advanced Treatise, H. Leibowitz, ed..
Vol.3, p.47, Academic Press. NY, 1971.

46 A. C. Mackensie, J. W. Hancock and D. K. Brown, Eng. Fract. Mech.,
Vol.9. 1977, p.167.

47 C. Y. You, M. Dolalr, A. W. Thompson and I. M. Bernstein, “Micrstruc-
ture Property Relationships and Hydrogen Effects in a Particulate-Rein-
forced Aluminum Composite”, Met. Transc., Vol.22A, 1991, p.2445.

48 T. Ishikawa and R. B. McLellan,“ The Diffusivity of Hydrogen in Alumi-
num”, Acta. Metall, Vol.34, No.6, 1986, p.1091.

85

49 David. L. McDanels, “Analysis of Stress-Strain, Fracture and Ductility
Behavior of Aluminum Matrix Composites Containing Discontinuous Sili
con Carbide Reinforcement“, Met Trans., Vol.16A, 1985, p. 1 105

"Ililllilll1111111111“