Origin of unusual Ag diffusion profiles in CdTe

1

J. Lehnert, M. Deicher, K. Johnston, Th. Wichert, H. WolfExperimentalphysik, Universität des Saarlandes, 66123 Saarbrücken,Germany

The ISOLDE CollaborationCERN, Geneva, Switzerland

ISOLDE Workshop and Users meeting 2013

Support by BMBF, contract 05KK7TS2

Outline

2

Diffusion

Uphill diffusion of Ag in CdTe

Influence of metal layers on Ag diffusion

Influence of oxidized crystal surfaces

Summary

3

Radiotracer technique

aCdTeØ = 6 mmd = 800 µm

Implantation

CdTethermal

treatment

550 … 900 K60 keV

30 nm

111Ag+

Resolution > 1 µm T1/2 > 1 h

Mechanicalsectioning

3

4

Diffusion in solids

Monotonously decreasing profiles

( ) X

XJ X D

x

( )X

J Xt x

24( , )

2

xDtNC x t e

Dt

for

0 200 400 600 800

1011

1012

1013

1014

conc

entra

tion

(cm

-3

)

depth (µm)

111AgCdTe

)()0,( xNtxC

4

Diffusion of 111Ag in CdTe

5(H. Wolf et al., Phys. Rev. Lett. 94 (2005) 125901)

Monotonously decreasing profile

Uphill diffusion profile

0 200 400 600 8001011

1012

1013

1014

0 200 4001011

1012

1013

1014

Tdiff = 800 Ktdiff = 60 minCd pressure

depth (µm)

depth (µm)

con

cent

ratio

n (c

m

-3

) Tdiff = 570 Ktdiff = 30 minArgon

5

6

ModelDeviation from stoichiometry

Initial state of the crystal Te rich

Highly mobile Cd interstitialsHighly mobile Ag dopant (interstitial)

Source of Cd interstitials during diffusion

DC < 0

Cd Cd

i Cd[ΔC]=[Cd ]-[V ]

Ag dopants images the profile of the intrinsic defects

0 200 400 600 8001011

1012

1013

1014

depth (µm)

DC < 0

DC < 0DC > 0 DC > 0

DC > 0 DC > 0

7

Alternative sources for Cdi defects?

Metal layers

Oxidized crystal surfaces

8

Diffusion without Cu layer at 500 KDiffusion with Cu layer at 500 K

DC < 0

111Ag in CdTe

Metal layers can initialize uphill diffusion

(H. Wolf et al., Appl. Phys. Lett. 100 (2012) 171915 )

Influence of metal layers (Cu)

without metal layerwith metal layer

30 nm Cu layer

9

metal layer CdTe

Cdi source

CdTemetal layer metal-Tealloy

550 K

Model

Ag dopant images the profile of deviation of stoichiometry again?

ΔC < 0

CdTe

metal layer

ΔC > 0

CdTe

10

Can Ag image a previously prepared deviation of stoichiometry?

ΔC < 0

CdTe

Metal layer

2nd Annealing 550 K 60 min

Concentrationmeasurement

Creating the deviation of stoichiometry before implantation of Ag

1st Annealing 550 K 30 min

ΔC > 0

CdTe

ΔC < 0

Implantation111Ag

CdTe

ΔC < 0ΔC > 0

ΔC < 0

EtchingHCl

CdTe

ΔC > 0

Ar atmosphere

Ar atmosphere

11

Cd source is created by Te-metal alloy

Concentration of the dopant images the profile of deviation from stoichiometry again

Can we create a Cd source at the surface without evaporated metals?

Experimental result

ΔC < 0

CdTe

metal layer

ΔC > 0

CdTe

550 K 60 minArgon

12

Influence of the condition of the crystal surface

As delivered samples show uphill diffusion

Br-methanol, KOH etching suppresses uphill diffusion

Surface condition has significant effect on Ag diffusion

as delivered etched (Br-methanol, KOH)

550 K 30 minArgon

570 K 30 minArgon

13

CdTe CdTeTe oxide

RT+air

Cdi source

Natural oxidation at room temperature and air 1)

TEM measurements: TeO2 at the surface 2)

No evidence of Cd at the surface

1) T. B. Wu, J. S. Chen, C. D. Chiang, Y. M. Pang, and S. J. Yang, J. Appl. Phys. 71, 5212 (1992)

2) F. A. Ponce, R. Sinclair, and R. H. Bube, Appl. Phys. Lett. 39, 951 (1981)

Mechanism

Cd source is created by natural oxidation process

Can we create a Cd source by oxidizing etching?

14

Experimental idea

Oxidation creates a Cd source

ΔC < 0

etched CdTe

Etching HCl/H2O2 CdTe

ΔC < 0

CdTe

ΔC < 0

Implantation

Annealing 550 K 30 min

ΔC<0ΔC>0

Concentration measurement

550 K 30 minArgon

550 K 30 minArgonas delivered

15

- Uphill diffusion is initialized by a Cd source- Ag dopant images the profile of the deviation from stoichiometry- Cd source can be provided by:

Summary

CdTeTe oxide

Cdi source

Cd Cd

DC <0DC >0 DC >0

external Cd pressure

Cdi source

CdTemetal layer metal-Tealloy

evaporated metal layer natural and forced oxidation

Outlook Verifying of the results by diffusion experiments AND observation of

the state of the surface (XPS or similar method)

16

Thanks to

Uni Freiburg:M. FiederleA. Fauler

Uni Bonn:R. Vianden

CERN:ISOLDE Collaboration

BMBF for financial support

Uni Prague:R. GrillE. Belas