Publiez

S'identifier

S'inscrire

Uniaxial stress dependence of the « EL2 » and « EL3 » deep levels in bulk GaAs

de C. Merlet (Auteur)

publié par

physics

s'abonner

Domain: Physics
We present uniaxial stress measurements performed on the activation energies of both the deep EL2 center in GaAs (820 meV), often called « O » center, and the EL3 electronic trap (600 meV). The EL2 level is found to separate from the conduction band at a rate of (1.1 + 0.6) meV/kbar for uniaxial stress in the (100) direction and (0.8 ± 0.5) meV/kbar for uniaxial stress in the (111) direction. Both values are consistent with the pressure coefficient reported under hydrostatic conditions. The second level (EL3) exhibits more important stress dependences : (1.8 ± 0.3) meV/kbar under (100) and (2.8 ± 0.4) meV/kbar under (111) compressions, respectively.
lire la suite replier
Télécharger
 ⁄   

Partager

1517
Uniaxial
stress
dependence
of
the
«
EL2 »
and
«
EL3 »
deep
levels
in
bulk
GaAs
G.
Bastide,
G.
Sagnes
and
C.
Merlet
Centre
d’Etudes
d’Electronique
des
Solides
(*),
U.S.T.L.
Place
E.-Bataillon,
34060
Montpellier
Cedex,
France
(Reçu
le
20
mars
1980,
révisé
le
16
juin
1980,
accepté
le
23
juin
1980)
Résumé.
2014 Nous
présentons
des
résultats
expérimentaux
concernant
les
variations
de
l’énergie
d’activation
de
2
centres
profonds
dans
l’AsGa
sous
contrainte
uniaxe.
Les
2
niveaux
étudiés
sont
le
centre
EL2
(820
meV),
souvent
désigné
dans
la
littérature
par
la lettre
«
O
»,
et
le
centre
EL3
(600
meV).
L’expérience
montre
que
l’énergie
d’activation
de
ces
2
centres,
mesurée
par
rapport
au
minimum
0 3 9 3 c
de
la
bande
de
conduction,
croit
avec
la
contrainte
uniaxe.
Pour
le
centre
EL2
la croissance
est
de
1,1
±
0,6
meV/kbar
pour
une
contrainte
uniaxe
appliquée
dans
la
direction
(100) et
de
(0,8
±
0,5)
meV/kbar
pour
une
contrainte
appliquée
dans
la
direction
(111).
Ces
valeurs
sont
en
accord
avec
les
résultats
publiés
concernant
l’effet
d’une
pression
hydrostatique.
Le
niveau
EL3
présente
une
anisotropie
plus
grande
que
le
niveau
EL2
puisque
0394E
=
1,8 ± 0,3
meV/kbar
en
compression
(100)
et
2,8 ±
0,4
meV/kbar
en
compression
(111).
Abstract.
2014
We
present
uniaxial
stress
measurements
performed
on
the
activation
energies
of
both
the
deep
EL2
center
in
GaAs
(820
meV),
often
called
«
O »
center,
and
the
EL3
electronic
trap
(600
meV).
The
EL2
level
is
found
to
separate
from
the
conduction
band
at
a
rate
of
(1.1
+
0.6)
meV/kbar
for
uniaxial
stress
in
the
(100)
direction
and
(0.8
±
0.5)
meV/kbar
for
uniaxial
stress
in
the
(111)
direction.
Both
values
are
consistent
with
the
pressure
coefficient
reported
under
hydrostatic
conditions.
The
second
level
(EL3)
exhibits
more
important
stress
dependences :
(1.8
±
0.3)
meV/kbar
under
(100)
and
(2.8
±
0.4)
meV/kbar
under
(111)
compressions,
respectively.
Revue
Phys.
Appl.
15
(1980)
1517-1520
OCTOBRE
198
Classification
Physics
Abstracts
71.55Fr
-
73.30Fy -
71.70Ej
Introduction.
- In
cubic
compounds,
the
asso-
ciation
of shallow
impurity
states
with
satellite
minima
of
the
conduction
band
has
long
been
established
[1-3].
In
this
case,
the
corresponding
energy
levels
closely
follow
a
given
minimum
of
the
conduction
band
versus
external
perturbations,
like,
for
instance,
hydrostatic
pressure
[1,
2]
or
a
change
in
temperature
[3].
Concerning
deeper
impurity
states,
the
situation
is
more
complex
[4].
The
general
assumption
in
this
case
is
that
they
correspond
with
a
close
admixture
of
r,
L
and
X
contributions ;
the
shorter
the
extent
of
the
potential
in
real
space,
the
larger
the
admixture
in
k-space.
Depending
on
the
amount
of
admixture,
one
expects
to
find
pressure
coefficients
in
GaAs
for
the
thermal
ionisation
energy
of
impurity
ranging
between
the
limiting
values :
zero
(for
pure 03931
like
state)
and
14.1
meV/kbar
(for
X,-like
state).
In
good
agreement
with
these
qualitative
arguments,
recent
data
concern-
ing
electron
irradiation
in
GaAs
[5]
report
for
different
levels,
lying
at
0.18,
0.31
and
0.71
eV
below
the
minimum
of
the
conduction
band,
the
pressure
coeffi-
cients :
8.8,
14
and
11
meV/kbar
respectively.
Even
more
interesting
are
the
data
collected
[6]
for
the
electronic
trap
located
210
meV
below
the
bottom
of
the
conduction
band
in
GaAs.
lt
shifts
with
a
slope
11.6
meV/kbar
under
hydrostatic
pressure
but
exhibits
shear
dependences :
4.0
meV/kbar
under
(111)
stress
and
3.5
meV/kbar
under
(100)
stress.
The
difference
between
the
two
uniaxial
results
is
too
large
to
be
associated
with
experimental
uncertainties
(according
to
the
authors
of
Ref.
[6])
and
confirms
the
predomi-
nent
X-like
character
of
this
bound
state.
Concerning
now
the
deep
oxygen
«
EL2 »
center
which
is
most
commonly
found
in
GaAs
(El
=
820
meV),
two
reports
of
the
associated
pres-
sure
dependence
have
been
given
[7,
8].
Both
concern
hydrostatic
conditions,
but
White et
al.
[7],
using
a
photocapacitance
technique
report
(1.2
±
1 ) meV/kbar
while
Zylberstein
et
al.
[8],
using
a
transient
capaci-
tance
technique,
report
(3.8 ±
0.3)
meV/kbar.
The
purpose
of
this
paper
is
twofold.
First
obtain
Article published online by
EDP Sciences
and available at
http://dx.doi.org/10.1051/rphysap:0198000150100151700
1518
an
independent
determination
of
the
pressure
coeffi-
cient
of
the
« EL2 »
level
through
uniaxial
stress
experiments
and
eventually
collect
informations
about
the
symmetry
properties
of
the
associated
wave
func-
tions.
Second,
extend
these
experiments
to
a
second
level
with
a
comparable
value
of the
activation
energy.
We
have
chosen
the
level
identified
as
«
EL3 »
in
the
work
of
Martin et
al.
[9].
Together
with
the
«
EL2
»
center,
this
level
exists
in
bulk
materials
and
both
.release
their
electrons
near
room
temperature.
1.
Expérimentât.
-
We
measure
the
uniaxial
stress
dependence
of
the
thermal
activation
energy
for
deep
trap
electrons
using
a
transient
capacitance
technique
[10].
The
Schottky
barrier
is
made
from
gold
evapo-
rated
on
the
(110)
face
of
a
small
parallelepiped
(1
x
1
x
10
mm’)
of bulk
GaAs.
All
samples
are
cut
with
their
long
dimension
along
the
(100)
or
the
( 111 )
crystallographic
axis.
The
stress
apparatus
has
been
already
described
[11]
and
permits
application
of
large
uniaxial
compressions
under
satisfying
condi-
tions
of
homogeneity.
The
experimental
process
is
as
follows.
First
we
select
a
value
of
the
uniaxial
stress,
the
diode
is
then
biased
in
the
forward
direction
and
the
traps
filled
up.
Next
the
diode
is
reversely
biased,
the
traps
release
their
electrons
and
the
transient
signal
is
measured
on
a
Boonton
72
BD
capacitance
meter
followed
by
a
HP
3437
digital
voltmeter :
each
experiment
corres-
ponds
with
1000
sampling
points
delayed
by
5
ms
on
a
single
transient
characteristic.
The
data
are
then
stored
in
the
memory
of
a
HP
9825
calculator
for
fur-
ther
numerical
analysis.
During
an
experiment,
the
temperature
has
to
be
adjusted
in
such
a
way
that
the
total
capacitance
transient
signal
for
a
given
trap
is
about
5
seconds.
The
typical
values
were
327
K
for
EL2
level
and
293 K
for
EL3.
Since
the
emission
rate
depends
both
on
the
activation
energy
and
the
temperature
in
an
exponen-
tial
way,
this
one
had
to
be
kept
constant
within
0.05
K
through
a
complete
experimental
run.
This
constitutes
our
major
source
of uncertainty.
2.
Results
and
discussion.
- Two
series
of
crystals
have
been
used.
Their
electron
concentrations,
obtain-
ed
from
the
plot
of
1/C2
=
j.(V),
were :
2.5
x
1015
cm-3
and
1.3
x
1016
cm-3 .
Preliminary
investigations
of
the
samples
by
classical
DLTS
gave
for
the
zero
stress
emission
rate :
-
crystal
1
:
en
=
3.8
x
106
T2
exp( -
0.60/KT)
(EL3
level)
2013
crystal
2 :
en
=
4
x
10’
T2 exp( -
0.82/KT) (EL2 level).
These
values
are
in
good
agreement
with
previously
published
data
[8,
10].
Concerning
the
capacitance
transient
signal,
for
the
EL3
level
crystal
1,
it
was
found
to
be
purely
exponential
and
well
described
by
equation :
0394C
= a
+
b exp -
t/03C4n
(1)
where
r.
=
1/en.
A
least
mean
square
fit
procedure
(through
the
1000
experimental
points
collected
for
a
single
run)
allows
the
determination
of
the
3
indepen-
dant
parameters
a,
b and
0 3 C 4 n .
Next,
a
plot
of log
z"
versus
uniaxial
stress
gives
the
change
in
ionisation
energy.
The
high
sensitivity
of
the
method
arises
from
the
exponential
law :
en
=
1/03C4n ~
T2
exp -
(El/kT) .
Provided
the
temperature
is
kept
constant
during
the
experiment,
changes
in
0 3 9 4 0 3 C 4 n / 0 3 C 4 n
give
MI/kT.
The
typical
value
of
the
sensitivity
for
our
apparatus
in
such
that
a
3
%
relative
change
in
’tn
could be
detected.
This
corresponds with
AE
=
0.5
meV
at
about
300
K.
Concerning
EL3,
typical
results
are
shown
in
figure
1
for
(100)
and
(111)
directions,
respectively.
We
find
slopes :
(100)
stress. :
dEIdp
=
(1.8
±
0.3)
meV/kbar
(111)
stress :
dEIdp
=
(2.8
±
0.4)
meV/kbar .
Fig.
1 .
-
Plot
of
Ln
103
Tn
versus
uniaxial
stress
for
the
« EL3
»
level.
1519
In
both
cases,
increasing
the
stress,
one
increases
the
ionization
energy
but
the
rate
of
variation
is
30
%
less
for
uniaxial
stress
applied
in
the
(100)
direction
when
compared
with
the
( 111 )
direction.
For
convenience,
we
list
in
table
1
the
stress
depen-
dence
of
both X
and
L
minima
of
the
conduction
band
with
respect
to
the
F,,, central
minimum
in
GaAs.
All
data
are
collected
from
the
recent
review
of
Aspnes
Table
I.
-
Theoretical
stress
dependences,
with
respect
to the
r
minimum
of
the
conduction
band,
for
pure
X
1-like
or
L1-like
impurity
levels.
All
data
are
collected
from
the
recent
review
of
ASPNES
et
al.
[6].
Stress
X,-like
L 1 -like
configuration
impurity
impurity
Hydro
(a)
High
energy
stress-split
components.
They
are
associated
with
a
triplet
for
(111)
stress
and
a
doublet
for
(100)
stress.
(b)
Center
of
gravity
associated
with
an
unresolved
multiplet.
(c)
Low
energy
components
associated
with
a
single
for
both
(111)
and
(100)
stress.
et
al.
[6].
Under
uniaxial
stress,
we
give
different
slopes
associated
with :
a)
the
high
energy
stress-split
component,
b)
the
center
of
gravity
of
the
multiplet
and
c)
the
low
energy
component.
Since
we
are
involved
in
thermal
activation
experiments
where
the
Boltzmann
factor
is
determinant,
we
favor
the
first
value
unless
the
splitting
energy
is
much
less
than
kT.
In
this
case,
the
experimental
anisotropy
is
consistent
with
an
admixture
of
X-like
wave
functions,
in
the
electronic
states
of
trapped
electrons
around
the
impurity
site.
Consider
now
the
deeper
«
EL3 »
level.
First
the
observed
transient
capacitance
signal
does
not
follow
a
single
exponential law
as
for
the
EL3
center.
Here
it
has
to
be
corrected
for
a
small
linear
contribution
and
a
good
fit
is
obtained
by
setting :
This
additional
linear
term
for
the
EL2
level
has
been
also
observed
by
R.
H.
Wallis
et
al.
[5]
but
its
physical
origin
is
not
clear.
It
may
probably
result
from
the
high
value
of
the
volume
density
of
the
EL2
level
or
from
the
contribution
to
the
release
of
elec-
trons
of
another
deep
level
with
a
smaller
emission
rate.
A
comparison
is
shown
in
figure
2
between
the
experimental
data
and
the
curve
«
a »
computed,
according
to
equation
(1).
When
equation
2
is
used,
Fig.
2.
-
Transient
capacitance
signal
of
the
«
EL2 »
level
approxi-
mated
by :
a)
0394C
= a
+ b
exp(-
tJi"),
calculated
curve :
tn
=
1
110
ms ;
b)
0394C
= a
+ b
+
C
exp( -
t / 0 3 C 4 n )
calculated
curve
0 3 C 4 n
=
790
ms.
Fig.
3.
-
Plot
of
Ln
103 Tn
versus
uniaxial
stress
for
the
«
EL2
»
.
level.
a)
(111)
stress;
b)
(100)
stress.
the
computed
curve
« b »
is
indistinguishable
from
the
experimental
one.
From
a
series
of
data,
collected
at
different
stress
magnitudes
in
(111)
and
(110)
directions,
we
deduce
the
change
in
trap
ionization
energy
versus
applied
pressure.
The
results
are
shown
in
figure
3.
We
find :
Neglecting
all
possible
shear
contributions,
the
corresponding
hydrostatic
pressure
coefficient
is
(3
+
1.5)
meV/kbar.
lt
supports
the
experimental
1520
value :
(3.8 +
0.3)
obtained
by
the
authors
of
refe-
rence
[8].
Acknowledgments.
-
The
authors
are
indebted
to
Dr.
L.
Hollan
(LEP)
for providing
the
crystals
used
in
these
experiments.
Thanks
are
due
to
Dr.
Housin
M.
for
his
technical
assistance.
We
had
stimulating
discussions
with
Dr.
A.
Mircea,
Pr.
M.
Rouzeyre
and
Dr.
J.
Camassel
who
is
greatly
acknowledged
for
his
contribution
to
the
final
redaction
of
this
manuscript.
References
[1]
PAUL,
W.,
Proced.
9th
Int.
Conf.
on
the
Physics
of
semi-
conductors,
Moscou
(1968)
p.
16.
[2]
KOMCZYKOWSKI,
M.,
POROWSKI,
S.
and
CHROBOCZEK,
J.,
Proc.
11th
Int.
Conf.
Physics
of
Semiconductors,
Warsaw
(1972)
p.
1050.
[3]
CAMASSEL,
J.,
AUVERGNE,
D.
and
MATHIEU,
H.,
J.
Physique
Colloq.
38
(1974)
C3-67.
[4]
BASSANI,
F.,
LKDONLSI,
G.
and
PRESIOZI,
B.,
Rep.
Prog.
Phys.
37
(1974)
1099.
[5]
WALLIS,
R.
H.,
ZYLBERSZTEJN,
A.,
BESSON,
J.
M.
and
BOUR-
GOING
J.,
Scd
Lund
Conference
Deep
level
impurities
in
Semicond.
Ste
Maxime
(1979).
[6]
ASPNES,
D.
E.
and
CARDONA,
M.,
Phys.
Rev.
B
17
(1978)
741.
[7]
WHITE,
A.
M.,
PORTEOUS,
P.,
SHERMAN,
W.
F.
and
STADT-
MULLER,
A.
A.,
J.
Phys.
C
10
(1977)
L-473.
[8]
ZYLBERSZTEIN,
A.,
WALLIS,
K.
H.
and
BESSON,
J.
M.,
Appl.
Phys.
Lett.
32
(1978)
764.
[9]
MARTIN,
G.
M.,
MITONNEAU,
A.
and
MIRCEA,
A.,
Electron.
Letters
13
(1977)
191.
[10]
LANG,
D.
V.
and
LOGAN,
R.
A.,
J.
Electron.
Mater.
4
(1975)
1053.
[11]
MERLE,
P.,
PASCUAL,
J.,
CAMASSEL,
J.
and
MATHIEU,
H.,
Phys.
Rev.
B
B
21
(1980)
1617.

Chargement...

Signaler un abus

  • 0 vote(s)

    0

  • 13 lecture(s)
  • 0 commentaire(s)
  • 0 téléchargement(s)
Publié le : 29/06/2012
Langue : Français
Nombre de pages : 4
Type de la publication : Rapports et thèses
Thème : Savoirs >

Science de la nature

Source : Revue de Physique Appliquée

17/1000 caractères maximum.

Suivez YouScribe

 

Ajout de cette lecture à votre activité Facebook

Vos amis seront au courant que vous avez lu ce document.

D'accord
Ne pas ajouter
Configurer cette fonctionnalité