Damage buildup and recovery in III–V compound

semiconductors at low temperatures

A. Turos

a,b,*

, A. Stonert

a

, L. Nowicki

a

, R. Ratajczak

a

,

E. Wendler

c

, W. Wesch

c

a

Soltan Institute of Nuclear Studies, 05-400 S´wierk/Otwock, Poland

b

Institute of Electronic Materials Technology, 01-919 Warsaw, ul. Wolczynska 133, Poland

c

Friedrich Schiller University, 07743 Jena, Max-Wien-Platz 1, Germany

Available online 1 August 2005

Abstract

Results are presented of the RBS/channeling study of the structural defect behavior in ion bombarded

In

x

Ga

1

x

As

y

P

1

y

(0 6 x, y 6 1) compounds at temperatures ranging from 15 K (LT) to 300 K (RT). Experiments con-

sisted of implantation with different ions to fluences ranging from 4

· 10

13

to 5

· 10

15

at./cm

2

at different temperatures

followed by in situ RBS/channeling measurements. Successive measurements of LT implanted samples were performed
during warming up to RT.

Broad recovery stage beginning at 100 K for all compounds was revealed. It was attributed to the defect mobility in

the group III sublattice. Steep damage buildup up to amorphisation with increasing ion dose was observed. The defect
production efficiency is much higher at LT than at RT. The consequences of defect mobility at RT for ion implanted
semiconductor structures are discussed.
2005 Elsevier B.V. All rights reserved.

PACS: 61.80.Jh; 68.55.Jk; 68.55.Ln; 61.10.Nz

Keywords: III–V semiconductor compounds; Defects; Ion implantation; RBS/channeling

1. Introduction

Modification of semiconductor properties by

ion implantation is a well-established technologi-
cal process. For structures based on compound
semiconductors it is used for introducing active
doping, compositional mixing of quantum wells

0168-583X/$ - see front matter

2005 Elsevier B.V. All rights reserved.

doi:10.1016/j.nimb.2005.06.097

*

Corresponding author. Address: Institute of Electronic

Materials Technology, 01-919 Warsaw, ul. Wolczynska 133,
Poland. Tel.: [phone]; fax: [phone].

E-mail address:

[e-mail]

(A. Turos).

Nuclear Instruments and Methods in Physics Research B 240 (2005) 105–110

www.elsevier.com/locate/nimb

and formation of isolation regions in electronic de-
vices. Point defects and their complexes determine
optical and electrical properties of semiconduc-
tors, diffusion of impurities as well as the recovery
of crystalline lattice after ion bombardment and
subsequent annealing.

Broad recovery stage at low temperatures exists

for III–V semiconductor compounds

[1–5]

. It is

located between 100 K and 400 K and is attributed
to the recombination or reconfiguration of a
variety of defects with different activation energy.
Thus, the investigation of thermally activated
processes can be decisive for identification of de-
fects. On the other hand defect mobility at RT
can lead to important effect transformation after
RT implantation and subsequent storage. Hence,
structure and distribution of radiation defects
are of great scientific and technological interest
and their reproducible control is crucial for elec-
trical and structural properties of implanted
materials.

In this paper we review the results of the study

of defect buildup and recovery in arsenide and
phosphide semiconductor compounds after ion

implantation at temperatures below 50 K and sub-
sequent warming up to RT. This process was mon-
itored by in situ RBS/channeling measurements.
Since the main objective of this workwas to eluci-
date the properties of point defects and their com-
plexes light ion bombardment was applied. Light
ions produce principally diluted binary collision
cascades and are well suited for the study of simple
defects.

2. Experimental

In

x

Ga

1

x

As

y

P

1

y

(0 6 x, y 6 1) binary, ternary

and quaternary compound semiconductors were
studied. Epitaxial layers of these compounds were
grown using the MOCVD technique in the Aix-
tron AIX200RD reactor at the Institute of Elec-
tronic Materials Technology, Warsaw on

h1 0 0i

semi-insulating GaAs and InP substrates. These
were: InP, In

0.53

Ga

0.47

As, In

0.82

Ga

0.18

As

0.52

P

0.48

.

Layers of such compositions are lattice matched
to InP substrates and consequently they are not
strained due to the pseoudomorphic growth.

Energy (keV)

600

800

1000

1200

Yield

0

500

1000

1500

2000

Random

10E12

4E12

2E12

1E12

Virgin

Fig. 1.

h1 0 0i aligned spectra measured in situ for In

x

Ga

1

x

As

y

P

1

y

(x = 0.82, y = 0.52) epitaxial layer before and after implantation

to different fluences of 150 keV N ions at 15 K. The spectra are labeled with fluences in 10

13

at./cm

2

.

106

A. Turos et al. / Nucl. Instr. and Meth. in Phys. Res. B 240 (2005) 105–110

Fluence (1x10

13

at/cm

2

)

0

10

20

30

40

50

60

70

80

Amorphous fraction

0.0

0.5

1.0

InGaAsP - 15 K
InGaAs -15 K
InP - 15 K
InGaAsP - RT

Fig. 2. Damage buildup in different semiconductor compounds upon N-ion implantation at 15 K. Also shown is the similar curve for
InGaAsP measured after RT implantation.

T

a

(K)

0

100

200

300

N

d

(15K)/N

d

(T

a

)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

InP
GaAs
InGaAs
InGaAsP

After
24 hr

Fig. 3. Defect recovery in various semiconductor compounds during warming from 15 K to up 295 K. The solid lines are drawn to
guide the eye.

A. Turos et al. / Nucl. Instr. and Meth. in Phys. Res. B 240 (2005) 105–110

107

Experiments were carried out using the Romeo

and Julia two beam facility at Institute of Solid
State Physics, FSU Jena. The beam delivered by
the ion implanter Romeo were used to produce de-
fects, while the tandem accelerator Julia equipped
with a 3-axis goniometer was applied for in situ
ion-channeling (RBS/c) measurements. The exper-
iments were carried out following two schemes:

(i) Epitaxial layers were implanted to increasing

fluences of 150 keV N

+

ions at two tempe-

ratures (15 K–50 K) and RT until amorphi-
sation was attained; damage buildup was
monitored by in situ RBS/c measurements.
In order to avoid sample heating upon ion
implantation beam current density was kept
below 1 lA/cm

2

.

(ii) Implanted at low temperatures epitaxial

layers were analyzed in situ by RBS/c with
1.4 MeV

4

He ions and stepwise warmed up

to RT. At each selected temperature the sam-

ple was stored for 10 min and then cooled
down to 15 K where aligned RBS/c spectra
were measured. Subsequently, the sample
was warmed to the next preset temperature
and the whole procedure was repeated.

The evaluation of channeling data measured at

different temperatures requires consideration of
several usually neglected factors, like changes of
thermal vibration amplitudes of crystal atoms
and defect production or removal by the analyzing
beam. The Monte Carlo computer code McChasy
described in detail elsewhere

[5]

was used for the

purpose.

3. Results

Fig. 1

shows the

h1 0 0i aligned spectra for a In

x

-

Ga

1

x

As

y

P

1

y

(x = 0.82, y = 0.52), epitaxial layer

taken at 15 K prior to and after ion implantation

Channel number

400

500

600

700

800

Y

ield

0

1000

2000

3000

Random
Virgin

1E13 As/cm

2

as implanted

1E13 As/cm

2

after 2 months at RT

1E12 As/cm

2

as implanted

1E12 As/cm

2

after 2 months at RT

Fig. 4. Random and

h1 0 0i aligned RBS spectra for InP single crystal prior and after 1.2 MeV As-ion implantation to different fluences

and after prolonged storage at RT.

108

A. Turos et al. / Nucl. Instr. and Meth. in Phys. Res. B 240 (2005) 105–110

to different fluences of 150 keV N ions. Typical
damage peakthat forms upon ion implantation
is in this case composed of four peaks each corre-
sponding to a different sublattice. Damage profiles
for each sublattice were calculated by fitting simu-
lated spectra to the experimental ones yielding the
damage buildup curve shown in

Fig. 2

. There is a

small difference in defect production efficiency at
LT between different In concentration containing
compounds. The damage ingrowth is much slower
for RT ion implantation. Here again the difference
between the studied compounds is rather small.
Temperature dependence of defect production effi-
ciency is a strong indication that important defect
transformations occur at temperatures below RT.

Fig. 3

shows damage recovery curve for samples

implanted at LT. For all studied compounds
broad recovery stage begins at approximately
100 K and extends above RT. Although our exper-
imental setup does not allow direct measurements
above RT the prolonged storage at RT after
warming up of a sample implanted at LT leads

to the further reduction of defect content. For
comparison similar recovery curve for GaAs is
also plotted in

Fig. 3

.

Defect mobility at RT has profound conse-

quences on properties of ion implanted semicon-
ductor

structures.

Fig.

4

shows

channeling

spectra for InP single crystal implanted in random
direction at RT to different fluences of 1.2 MeV As
ions. After prolonged storage at RT important
damage reduction was clearly visible. Damage
depth profiles calculated from these spectra and
corresponding lattice strain profiles determined
by high resolution X-ray diffraction (HRXRD)

[6]

are shown in

Fig. 5

. One notes a strong corre-

lation between RBS/channeling and HRXRD
data.

4. Discussion and conclusions

Despite considerable workin recent years the

defect structure development in ion-irradiated

Depth (nm)

0

200

400

600

800

1000

-∆

a/a (ppm)

0

200

400

600

800

1000

Defect Concentration (at.%)

0

20

40

60

80

100

1E13 As/cm

2

XRD as implanted

1E13 As/cm

2

RBS/c as implanted

1E13 As/cm

2

RBS/c after 6 months at RT

1E13 As/cm

2

XRD after 6 months at RT

Fig. 5. Damage depth distribution for InP single crystal implanted with 1.2 MeV As ions to different fluences and the corresponding
strain profile for as implanted samples and after prolonged storage at RT. Note that the strain is negative, i.e. the implanted layer is
under tensile stress.

A. Turos et al. / Nucl. Instr. and Meth. in Phys. Res. B 240 (2005) 105–110

109

III–V semiconductor compounds is not well
understood. Although ion bombardment processes
in InP have been studied quite extensively over the
last decade, there remains a lackof understanding
of the fundamental mechanism of dynamic anneal-
ing that controls the damage buildup, amorphisa-
tion and annealing in other In-based compounds.
There is a general agreement that high concentra-
tions of Frenkel pairs can be frozen in during low
temperature irradiations and that defects in the
two sublattices react independently. Those that an-
neal below room temperature are related to the
group III sublattice whereas the group V intersti-
tials become mobile at approximately 500 K. Pos-
itron annihilation spectroscopy confirmed that the
large annealing stage between 100 K and 300 K is
due to the mobility of In vacancies and their sub-
sequent recombination with In interstitials

[7]

.

Recombination with P interstitials is also possible
leading to the formation of antisite defects. The
fact that this occurs in a quite large temperature
range suggests that a variety of trapping centers
of both kinds of interstitials is involved. Defects
that survive well above room temperature belong
principally to the group V sublattice.

There is only a small difference in damage build-

up in for all studied compounds. Akano et al.

[8]

observed that InP can be more easily amorphised
than InGaAs upon O-ion bombardment at 80 K.
At the present it is not clear whether this difference
can be attributed to the ion itself or to their less
accurate method of spectra analysis. Damage

recovery at low temperatures reveals much impor-
tant differences: the fastest recovery was observed
for InGaAsP whereas it is slowest in InP. Conse-
quently the residual damage at RT amounts to
40% of the initial one as compared to 18% for
InGaAs. This is apparently related to the kind
of formed defects which are unknown at the
moment.

The great mobility of group III interstitials has

important consequences in the practice. Shelf stor-
age of ion implanted semiconductor structures,
even the processed ones, can produce important
changes of their properties. Defect presence influ-
ences directly the conductivity and defect induced
strain modifies the bandgap.

References

[1] S.T. Picraux, Radiat. Effects 17 (1973) 261.
[2] D. Stievenard, X. Boddaert, J.C. Bourgoin, H.J. von

Bardeleben, Phys. Rev. B 41 (1990) 5271.

[3] A. Stonert, A. Turos, B. Breeger, E. Wendler, W. Wesch,

Mod. Phys. Lett. B 15 (2001) 1437.

[4] K. Karsten, P. Ehrhart, Phys. Rev. B 51 (1995) 10508.
[5] L. Nowicki, A. Turos, R. Ratajczak, A. Stonert, F. Garrido,

these Proceedings,

doi:10.1016/j.nimb.2005.06.129

.

[6] A. Turos, J. Gaca, M. Wojcik, L. Nowicki, R. Ratajczak, R.

Groetzschel, F. Eichhorn, N. Schell, Nucl. Instr. and Meth.
B 219–220 (2004) 618.

[7] M. To¨rnquist, J. Nissila, F. Kiessling, C. Corbel, P.

Hautoja¨rvi, Mater. Sci. Forum 143–147 (1994) 347.

[8] U.G. Akano, I.V. Mitchell, F.R. Shepherd, C.J. Miner,

Nucl. Instr. and Meth. B 106 (2004) 618.

110

A. Turos et al. / Nucl. Instr. and Meth. in Phys. Res. B 240 (2005) 105–110