13_2005_Nowicki_Damage_Buildup.pdf

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

www.elsevier.com/locate/nimb

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 ´ 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 e9ciency is much higher at LT than at RT. The consequences of defect mobility at RT for ion implanted

semiconductor structures are discussed.

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

* Corresponding author. Address: Institute of Electronic

Materials Technology, 01-919 Warsaw, ul. Wolczynska 133,

Poland. Tel.: +48 603 092223; fax: +48 8645496.

E-mail address: turos_a@itme.edu.pl (A. Turos).

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

doi:10.1016/j.nimb.2005.06.097

106

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

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] .Itis

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 h 100 i

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.

2000

Random

1500

10E12

1000

4E12

2E12

500

1E12

Virgin

600

800

1000

1200

Energy (keV)

Fig. 1. h 100 i 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 .

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

107

1.0

0.5

In GaAsP - 15 K

InGaAs -15 K

InP - 15 K

InGaAsP - RT

0.0

Fluence (1x10 13 at/cm 2 )

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.

1.2

1.0

0.8

0.6

0.4

InP

GaAs

InGaAs

InGaAsP

After

24 hr

0.2

0.0

100

200

300

T a (K)

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.

108

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

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:

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.

(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-

3. Results

Fig. 1 shows the h 100 i 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

3000

Random

Virgin

1E13 As/cm 2 as implanted

1E13 As/cm 2 a f ter 2 months at RT

1E12 As/cm 2 as implanted

1E12 As/cm 2 after 2 months at RT

2000

1000

400

500

600

700

800

Channel number

Fig. 4. Random and h 100 i 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.

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

109

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 e9ciency 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 e9-

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

100

1E13 As/cm 2 XRD as implanted

1E13 As/cm 2 RBS/c as implanted

1E 1 3 As/cm 2 R BS /c after 6 months at RT

1E 1 3 As/cm 2 XRD after 6 months at RT

1000

800

600

400

200

400

600

800

1000

Depth (nm)

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.

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: