Emission channeling with short-lived isotopes (EC-SLI)

of acceptor dopants in nitride semiconductors

U. Wahl1, V. Augustyns2, J.G. Correia1, A. Costa1, E. David Bosne1,

T.A.L. Lima2, G. Lippertz2, L.M.C. Pereira2, M.R. da Silva3,

K. Temst2, and A. Vantomme2

(Emission Channeling with Short-Lived Isotopes,

the EC-SLI collaboration)

1) Centro de Ciências e Tecnologias Nucleares, Instituto Superior Técnico,

Universidade de Lisboa, Portugal

2) KU Leuven, Instituut voor Kern- en Stralingsfysica, Belgium

3) Centro de Física Nuclear da Universidade de Lisboa, Portugal

Spokespersons: U. Wahl and L.M.C. Pereira

uses

Proposal INTC-P-489

Motivation: Mg in nitride semiconductors

• Nitrides are base material e.g. for white LEDs, blue lasers, power devices,voltage transformers

• Mg is the only technologically relevant p-type dopant in GaN

• Other group II impurities, e.g. Be, Ca were also tried in GaN but did not succeed as acceptors

• To be electrically active, Mg acceptors (group II) should occupy substitutional Ga (group III) sites

• Mg doping works but not without problems

Nobel prize in Physics 2014

I. Akasaki, H. Amano, S. Nakamura

Doping problems in GaN: limitations at high [Mg]

• Doping limitations: once [Mg] surpasses ~1019-1020 cm3, further introduction of Mg does not lead to an increase in the hole concentration. Reasons?

• Passivation of MgGa acceptors by H

• Compensation by native defects: VN+

• Solubility limit: formation of Mg3N2 precipitates.

• Formation of Mgi interstitials (double donors) also predicted by theory, but previously never experimentally observed!

• EC-SLI experiments have given first clear evidence for the existence of Mg interstitials and its coupling to the Fermi level amphoteric nature of Mg

G. Miceli, A. Pasquarello,

PRB 93 (2016) 165207

Emission channeling: basic principle

33 cm2 Si pad detector

2222 pixels of 1.31.3 mm2

currently: max. ~ 5 kHz

b emission channeling patterns from 27Mg (t1/2=9.5 min) in p-GaN:Mg

fit results:

200°C: ~72% 27Mg aligned with c-axis , ~33% on interstitial sites

0.20 pA implantation current, total fluence 1.11011 cm2

-2

-1

0

1

2

-1 0 1 2 3

(01-

10)

(11-20)

(11-20

)

(01-

10)

1.42 - 1.50

1.33 - 1.42

1.25 - 1.33

1.16 - 1.25

1.07 - 1.16

0.99 - 1.07

0.90 - 0.99

0.82 - 0.90

-1 0 1 2 3

-2

-1

0

1

2

(11-20

)

(11-20)

(b) best fit

1.42 - 1.50

1.33 - 1.42

1.25 - 1.33

1.16 - 1.25

1.07 - 1.16

0.99 - 1.07

0.90 - 0.99

0.82 - 0.90

-1 0 1 2 3

-2

-1

0

1

2

(01-10)

(01-

10)

(11-

20) (11-20)

(d) 100% Mgi(c) 100% SGa

1.59 - 1.70

1.48 - 1.59

1.37 - 1.48

1.26 - 1.37

1.15 - 1.26

1.04 - 1.15

0.93 - 1.04

0.82 - 0.93

-1 0 1 2 3

-2

-1

0

1

2

(01-10)

(11-20)

(a) experiment

1.34 - 1.43

1.26 - 1.34

1.17 - 1.26

1.08 - 1.17

1.00 - 1.08

0.91 - 1.00

0.82 - 0.91

0.74 - 0.82

U. Wahl et al,

PRL 118-8 (2017)

Lattice location precision of 27Mg in p-GaN:Mg

• Location of interstitial Mgi in between HA and HAB site

(+0.60±0.14) Å from ideal O sites, but only 13% Mgi at RT

-1.0 -0.5 0.0 0.5 1.0

0.97

0.98

0.99

1.00

HAHB O HABHAB

(c) [-2113]

distance from O site along c-axis [Å]

20°C

300°C

450°C

0.80

0.85

0.90

0.95

1.00

(b) [-1101]

2 r

elat

ive

to s

ubst

ituti

onal

sit

es o

nly 0.70

0.80

0.90

1.00

exp

(d)

(a) [-1102]

fit

(e)

exp

(f)

fit

(g)

exp

(h)

fit

(i)

-1.5

-1.0

-0.5

0.0

0.5

-1.0 -0.5 0.0 0.5 1.0

(a)

(0110)

experiment

-1.0 -0.5 0.0 0.5 1.0

1.93 - 2.10

1.77 - 1.93

1.60 - 1.77

1.44 - 1.60

1.27 - 1.44

1.11 - 1.27

0.94 - 1.11

0.77 - 0.94

simulation SGa + Mgi sites

(e) [0001]

-1.5

-1.0

-0.5

0.0

0.5

(202

1)

(b)

(1120)

2.11 - 2.30

1.92 - 2.11

1.74 - 1.92

1.55 - 1.74

1.36 - 1.55

1.17 - 1.36

0.99 - 1.17

0.80 - 0.99

(f) [1102]

-1.0

-0.5

0.0

0.5

1.0

(c)

(1011

)

(1120)

(d)

2.03 - 2.20

1.86 - 2.03

1.69 - 1.86

1.52 - 1.69

1.35 - 1.52

1.18 - 1.35

1.01 - 1.18

0.84 - 1.01

(g) [1101]

[deg]-1.0 -0.5 0.0 0.5 1.0-1.5

-1.0

-0.5

0.0

0.5

(0110)

(1120)

-1.0 -0.5 0.0 0.5 1.0

1.63 - 1.75

1.52 - 1.63

1.41 - 1.52

1.29 - 1.41

1.18 - 1.29

1.07 - 1.18

0.95 - 1.07

0.84 - 0.95

(h) [2113]

High angular resolution

(small solid angle)

1.5-3 pA

1.51012 cm2 per pattern

U. Wahl et al,

PRL 118-8 (2017)

B

b emission channeling patterns from 11Be (t1/2=13.8 s) in nid-GaN

RT, fit results

(not yet corrected for background):

~21% 11Be on SGa sites,

~30% near interstitial O sites

majority of 11Be on interstitial sites!

-1

0

1-1.0 -0.5 0.0 0.5

(0110)

experiment

-1.0 -0.5 0.0 0.5

1.06 - 1.08

1.04 - 1.06

1.02 - 1.04

1.01 - 1.02

0.99 - 1.01

0.97 - 0.99

0.96 - 0.97

0.94 - 0.96

simulation 30% Bei+21% SGa sites

[0001]

-1

0

1

fat1810a

(2021)

(1120)

1.10 - 1.13

1.08 - 1.10

1.06 - 1.08

1.04 - 1.06

1.02 - 1.04

1.00 - 1.02

0.98 - 1.00

0.96 - 0.98

[1102]

0

1

(110

2)

(1011)

(1120)

1.07 - 1.08

1.05 - 1.07

1.04 - 1.05

1.02 - 1.04

1.01 - 1.02

0.99 - 1.01

0.98 - 0.99

0.96 - 0.98

[1101]

[deg]-1 0 1

0

1

(41-50)

(41-50)(0110)

(41-50)(2

112

)

(0110)

(1120)

-1 0 1

1.016 - 1.022

1.010 - 1.016

1.004 - 1.010

0.998 - 1.004

0.992 - 0.998

0.986 - 0.992

0.980 - 0.986

0.975 - 0.980

[2113]

Lattice location precision of 11Be in nid-GaN

• Location of interstitial Bei in between HB and HAB site

(0.66±0.13) Å from ideal O sites.

• In comparison to Mg, Be sits on the opposite side from O

High angular resolution

(small solid angle)

0.15 pA

4.71011 cm2 per pattern

-1.0 -0.5 0.0 0.5 1.0

0.85

0.90

0.95

1.00

HAHB O HABHAB

[-2113]

distance from O site along c-axis [Å]

0.85

0.90

0.95

1.00

[-1101]

2 r

elat

ive

to s

ub

stit

uti

onal

sit

es o

nly

0.80

0.85

0.90

0.95

average: HB->HAB 0.367(104)

i.e. -0.664(0.129) Å from O -> HAB

[-1102]

Interstitial 27Mg in different doping types of GaN

Mgi (mobile) + VGa MgGa

fi=fi0 N n0t exp[EM/kBT]

with attempt frequency

n0=21013 s1

des

tro

yin

g

p-t

yp

e ch

arac

ter

• Interstitial Mgi enhanced in p-GaN and suppressed in n-GaN.

• Site change of 27Mg from interstitial to substitutional Ga sites as

function of implantation temperature allows to estimate

activation energy for migration of Mgi as EM 1.32.0 eV.U. Wahl et al,

PRL 118, 8 (2017)

0 100 200 300 400 500 600 700 8000

5

10

15

20

25

30

35

Arrhenius step models:

N=1 EM

=2.0 eV

N=105 E

M=1.3 eV

implantation current [pA]: 0.20 0.27 1.4-3.0

nid-GaN

p-GaN:Mg

GaN:Mg

n-GaN:Si

inte

rsti

tial

27M

g [

%]

implantation temperature T [°C]

Comparison: site changes of interstitial Li, Be, Na, Mg in nitrides

Material Intersti-

tial ion

Ionic

radius (Å)

Temperature

TOS (°C)

Migration energy experim.

EM,exp (eV) [Ref]

Migration energy theoretical

EM,theo (eV) [Ref]

GaN Li+ 0.59 ~ 427 ~1.7 [Dalmer JAP 1998,

Ronning JAP 2000]

1.4 () 1.55 (||)

[Bernardini PRB 2000]

Be2+ 0.27 ? ( > 20)

[this work]

1.2 () 2.9 (||)

[Van de Walle PRB 2001]

Na+ 0.99 ~ 900 2.2 3.4 [Ronning JAP 2000

PhD thesis L.Amorim 2016]

Mg2+ 0.57 ~ 400 1.3 2.0

[this work]

0.15 () 0.68 (||) [Harafuji

PSSC 03, Harafuji JJAPL 04]

AlN Li+ 0.59 ~ 427 ~1.7

[Ronning JAP 2000]

Na+ 0.99 600-900 2.0 2.6 [Ronning JAP 2000,

PhD thesis L. Amorim 2016]

Mg2+ 0.57 300-400 1.1 1.7

[Amorim APL 2013]

• Interstitial substitutional site changes observed for light alkalis and

alkaline earths, but not for Ca and Sr [B. De Vries et al, JAP (2006)]

• Interstitial migration energy is correlated with ionic radius of diffusing ion

• Migration energies tend to be lower in AlN

Proposed further experiments:

• 27Mg:

• Detailed mapping of site change Mgi MgGa in all doping types: does site change temperature depend on doping type?

If yes, EM(Mgi) or [VGa] depend on doping type

• Fluence dependence of [Mgi] at various implantation temperatures:

Introduction rates of compensating defects by implantation.

• GaN:Mg: can its p-type activation be observed around 400-500°C?

• Lattice location of 27Mg in hydrogenated GaN samples

New insights in electrical activation mechanisms of Mg

• 11Be:

• So far only results in nid-GaN at RT. Site changes Bei BeGa/Al of 11Be in nid-GaN, p-GaN:Mg and AlN as function of implantation temperature

Experimental insights why Be is not an active dopant + EM(Bei)

Beam request

isotope yield [atoms/mC] target – ion source shifts [8h]

27Mg 1107 Ti-W - RILIS Mg 16

11Be 6106 UCx-W or Ta-W -

RILIS Be

8

Total 24

• We welcome sharing beam times with other users.

• However, for 27Mg runs use of Ti-W targets is a must, since e.g. SiC or UCx targets suffer from 27Al and 27Na contamination.

EC-SLI Publications + theses:

• Detailed list available on INDICO

• 2012-2016: 19 papers published

• 2017: 1 paper accepted, 3 submitted

• 2012-2016: 3 PhD theses + 2 Masters theses completed

• 2017: 4 PhD theses ongoing