1
????????? ???????? ? ??????? ? ????????????

• ??????? ??????
• ???????? ?????? ???????? ???? ???,????????????,
• (in cooperation with Göttingen University)

2
??????? ???????? ?? ?????? ???????? ??????????
????????????????, ?? ? ???????? ??????????
????????? ??????????, ??????? ?? ?????? ? ???
????? ????? ????????? ????????????????
?? 10 ??? ???????????? ????????? ??????? ???????
????? ??? ? 20 ??? ? ?????????? ?????
???????????????!
?????? ????????? ??????? ???????? ? ???????? (??
95) ?? Si ??????? ????? ????? ? ???????, ??
?????????? ?, ? ????????, ????????? ???????????
????????? ??????? ? ??? ?? 24 ? ? ??????? (?? 50
???) ?????? ?????? ??????? Si- a-SiH
?????????, ? ????????, ???????? ??? ?? 30
????? ???????? ????????????? ????????? ???????
??? ????????? 15 ???.
3
??????? ?? ???????????? ? ??????? ????-?????????
? ???????? (? ???. ????)
4
???????? ?????? ??????? ???? ????????-????
????????????? ?????????? ?????????
????????? ????????? ?????????
(???????_??????)/ ?????????????
5
? ?????, ????? 40-50 ??? ????????? ??????? ??????
???????? ?????????? ??????? ??? ?????
6
??? ??????? ????????? ????????? ????????? ?? Si ?

• ????????????? ????? ???????? ???????- ?????
  (?????? EG-Si gt ????? ??????? SG-Si ?? ??????
  ??? ?????????? ??? ??? ????? ??????? UMG-Si -
  ?????????? ???????????????? ???????.)
• ????? ??????? ?????? ???????????? ???????
  (?????? C-Si gt poly-Si, ?????? ??????????
  ?????????? ????? ? ???? ?????? ???? ??? ?????)
• Casting, EFG, silicon sheets from powder
  (Fraunhofer), silicon on substrate, nano-Si films
  .

7
???????? ???????? ????????? ????? ?????
????-???????????? ?????????? ? ??????? ????-Si
???????
????? ?????????? ????????.
????? ?????????? ???????????? ??? ??????? ??????
???? ????? LDe(Dete)1/2 gt2d (LDe
-???????????? ????? ??????????) ??? d100
??????, te gt 12 ms ??? d200 ??????, te gt 48 ms
8
??? ???????????? ????? ????? ?????????? ?????????
? ??????? ????-????????

• te ???????????? ????????-???????? ?????????????
  ?? ???????? ???????, ?????????????, ?????? ?????
• ????????????? ? ?????? ???????????? ?????????
  ??????? ?????????? ???????? (Fe, Cr, Au, Ti,
  Ni.) ? ?? ??????????? ? ??????? ????????? ?
  ?????????
• ????????????-????????? ?????????? ?????????????
• ???????????? ? ?????? ? ? ???????? ?????.

LBIC - ????????????? ???????? ?????????? ????????
(T300K) ????? ??? ??????????? (??? ???? ??? ???
??????? ???????????). ??? ??? ????? ??????????,
??? ????????? ?????? ??? ????????? ?????????????
?? ??????????? ? ?? ??????? ?? ??????????.
9
??? ?? ??? ????? ? ??????????? ? Si ?
? ???????, ? Si ??????????? ??????????
???????????? ???????????? ??????????? ??? Tgt
600oC ???????????? ??????????? ??????
(dissociated 0o and 60o of the glide set).
10
Shockley partial dislocations are reconstructed
and produce only relatively shallow 1D bands EDe
and EDh, split from conduction and valence bands
by the dislocation strain.
However, they can have deep localised electronic
states due to 1. core defects (reconstruction
solitons, jogs) 2. Impurity atoms
Core view in the glide (111) plane of 30o (left)
and 90o (right) Shockley partial dislocations in
their unreconstructed (A) and reconstructed (B)
states. Reconstruction energy gain is of about
0.5 eV/bond
11
e-h recombination at dislocation
Transitions between 1D-bands (1) give small
recombination rate, i,e, recombination at clean
dislocation is small. But, recombination is
drastically enhanced by presence of even a small
concentration of deep levels defects at
dislocation.

• Capture of free electrons and holes to 1D-bands.
• Their motion along dislocation
• Capture from the 1D bands to deep states of
  defects and recombination.

12
Theoretical calculationsV.Kveder, M.Kittler,
W.Schroter, Phys.Rev.B 63, 115208 (2001)

• Coulomb band-bending, repulsive for majority
  carriers but attractive for minority carriers
  Strong amplification of deep-level
  recombination by 1D bands (at low NDD)
• At very large NDD total recombination rate can
  be smaller than for the same number of metal
  atoms randomly distributed in a bulk (charge and
  diffusion limitation)

13
e-h recombination at dislocation experiments
(EBIC, LBIC)
It was found by many authors that recombination
rate at dislocation vary strongly from sample to
sample and depends strongly on contamination
level. Perfect agreement with theory. At high
excitation recombination rate decreases strongly
due to decrease of dislocation charge
14
Why dislocations collect impurities ?

• Segregation in the strain field around
  dislocations (EB0.4-0.8eV, R1-2 nm)
• Binding to the dislocation core (EB up to 2-3eV
  !, Rlt0.5nm)
• Preferential nucleation of metal silicide
  particles at dislocation when metal impurities
  gets supersaturated during cooling. (Because
  barier is lower and concentration is higher).
• Preferential growth of metal particles near
  dislocation which works as a sink for V and Si-I.
  

• Combinations of all these 4 factors are possible!
  
• The electronic properties of impurities at
  dislocation may differ from those in bulk!
• Obviously, we need more research to clarify the
  situation for different impurities !

15
Interaction with strain field.
The dominant term of elastic interaction comes
from the difference in size of Si atom (r0r0
0.117 nm ) and impurity atom (rimp) DE(x,y)
P(r0-rimp)3 In average it is of the order of
0.3-0.7eV (depending on impurity type). At
Tgt500oC it cannot give large number of atoms at
dislocation. But, it is long distance
interaction and can enhance strongly kinetic of
incorporation of impurities to the
chemically-bonded state in dislocation core and,
also, can enhance precipitation kinetic.
16
Chemical interaction with core
According to calculations, some impurities have
large binding energies in dislocation cores.
Usually, some energy barrier should be overcome,
which is different for different places (RD,
jogs) The electronic properties of impurities at
dislocation core can strongly differ from those
in a sample bulk. Usually, several configurations
of atom in a core with different properties are
possible
Example Au at dislocations N.Fujita,
A.T.Blumenau, R.Jones, Soberg, P.R.Briddon
Strong increase of Au solubility in dislocated Si
was experimentally confirmed by NAA A.Rodriguez,
H.Bracht, I.Yonenaga, J. Appl. Phys., 95, 7842,
(2004)
17
??? ??????? ???????????? ????????? ???????? ?
???????? ???????????? ????????? ????????? !

• ?????????????? - ????????? ??????? ??????
  ?????????? ???????? ? ?????? ??? ??? ?? ??????
• - segregation gettering - ?? ???? ???????
  ????????????? ??????? ? ???????. ?? ???????
  ??????????? ?????? ?????? ????????. (???????? PDG
  ??? ???????????? p-n-???????? ? AlG ???
  ???????????? ????????? p-p ????????)
• - precipitation gettering - ?????????? ???
  ??????????? ?????? ?????? ???????? ?? ????
  ????????????????? ????? ???????????? ???? ???????
  ? ??????? ??????? (????????, BDG ??? ????????????
  ???? ?????? ??? ?????????????).
• 2. ?????????? - ??????????? ????????????
  ???????? ???????? ? ???????????? ??????????
  ????????? ?? ???? ?? ??????? ? ??????? ?????????
  ? ????????? (????????, ? ?????????)
• ????????? ?????? ??????? ? ??????????? ?????
  ????, ??????? ??????????? ? ????????????.
  ????????, b-FeSi2 (?lt930oC) ?????????????
  ???????????, ? a-Fe2Si5 (?gt930oC)- ?????? ?????
  ??????? ? ???????????? (??????????
  ?????-???????)

18
??????????? ????????? ????????????? ???????
????????? ???????????? !

• ??????????? ??????
• ??????? ?? ????? ??????? T ????? ????????
  ??????? ? ?????????? ? ?????????? ???????????.
• ???????? ????? ???????? ????? ??????? ??????
  ????????????????? ? ????? ? ??????????? ????????
  (??????) ?? ??? ??????? ?????? ??? ????????!

????? ????????? ??????? ????????? ?????
??????????? ?????. gt ??????? ????????????
???????????? ?????????????. ??? ????? ???? ?????
??? ???????? ?????????.. So, systematic
investigation of interaction of different
impurities with dislocations and their electronic
properties is nessesary Still in progress
19
??? Si ??? ?????????? PDG ? AlG ??? ????????
?????? ?????? ? ??????? ???????????????
???????????? ??????????. (??. ????????
V.Kveder, W.Schroeter, A.Sattler, M.Seibt,
Materials ScienceEngineering B71, 175-181
(2000))
?????????????? ??????????, ? ????????, ?? ????
??????????? ?????????? ???? ???????-??????
??????????? ??????? Si-I ???????? ? ??????????
???????? substitutional ?????? ???????, ? ????
???????, ? ????????? ????????? ???????????? ?
? ?????????? ???????????? segregation factor S
20
??? ???????? ? ?????? ??????? ???????????

1. ?????????? ???????? ??????? ??? ???????? ?
   ??????????? ??????????? ??????. ??? ????????
   ???????? PDG, ?????? ?? ??????????? ???????????
   ???????? ????????, ? ?.?. ??? ???????? ?
   ????????? ??????????? ???????????? ??????
   gettering simulator
2. ?????????????? ????????? ?????? ? ????????????.
   ??? ???????? ? ???????? ? ????????????? ???????
   ??????? ?????????????. ??? ????????? ? gettering
   simulator ???? ????? ??????? ?????. ???? ????
   ?????? ?????? ??? Au ? Cu ???????????? ???????
   ??????? ? ?????????? ????????.
3. ??????????? ??????????? (nucleation barier) ???
   ??????????? ???????????? ?? ?????????? ???????
   ??????, ??? ? ??????? ?????????. ??? ??????????
   ????????????? ??????? ??????? precipitation
   gettering ? ?????? ?? ??????????? ?????????
   segregation gettering. ????? ???????? ?
   gettering simulator ????????? ?????????????????
   ??????????? nucleation barier ?????????
   ???????????? ??????
4. ??? ?????????? ???????? ???????????? ?????????
   ?????? ??????????? ?????????? ??? ?????? ?
   ???????????? ???????? ?? ???????????
   ???????????? ??????? ??????? ? ??????????
   ????????.

21
?????? ??? ???????? ???????????????
????????????, ?????? ?? ????????? ???
??????????????? ????? ? ???????? ????????
??????????!
DLTS-measurements show that combination of
gettering and hydrogen plasma passivation reduces
strongly concentration of all deep levels at
dislocations (from 3.106cm-1 to below 7.104cm-1)
V. Kveder, M. Badylevich, E.
Steinman, A. Izotov, M. Seibt and W. Schroeter,
Aplied Phys. Letters, 84(12), 2106-2108 (2004)
22
Do clean dislocations affect the solar cell
efficiency ?
EBIC before and after PDG (Kittler et al)
Spectra of photo-current for p-n-structure with
very high dislocation density after
AlGPDGH-plasma. Note that Clean dislocations
can be also used to make a Si-LED for 1.5m !
Kveder et al, phys. stat. sol. (a) 202, 901
(2005) .
23
Example (1) Ni at dislocation
Evolution of dislocation DLTS spectra of the same
dislocations depending on their pre-history
(n-FZ-Si, NDIS105cm-2) The C-line seems to be
some impurity at dislocation (maybe NiS?) V.
Kveder, V. Orlov, M. Khorosheva, M.Seibt, Solid
State Phenomena 131-133, 175-181 (2008)
24
No theory for Ni at dislocation yet Experiment
(TEM) - in presence of dislocations all Ni
precipitates seems to be at dislocations
Is Ni at dislocation a members of the DLTS
C-defect family ?
25
Example 2 Au at dislocations Theory (N.
Fujita, A. T. Blumenau, R. Jones, S. ?berg, P. R.
Briddon)
Atomic structure and binding energies
30o glide, Eb1.71eV 90o glide,
Eb2.13 eV
26
Au at dislocations Experiments.

1. There is a lot of Au at dislocation. Dislocations
   become pined by Au atoms and we can not unlock
   and move them anymore by stress ! Strong increase
   of Au solubility in dislocated Si was also
   experimentally confirmed by NAA A.Rodriguez,
   H.Bracht, I.Yonenaga, J. Appl. Phys., 95, 7842,
   (2004)
2. But we do not see the DLTS of Au at dislocations
   (for some reasons). We continue working to
   understand the situation

27
Summary and Conclusion

• Clean dislocations in Si are mainly
  reconstructed and not very active in
  recombination.
• Strong interaction of dislocations with
  transition metal impurities is a key to
  understand the role of dislocations in poly-Si.
• The important role of dislocations as
  heterogeneous nucleation sites for silicide
  precipitates is beyond doubt. This is important
  for gettering of metal impurities from dislocated
  silicon such as typical multicrystalline
  materials for photovoltaics.
• Minority carrier recombination at dislocations
  is consistently described by the concerted action
  of shallow 1D bands, which are an intrinsic
  property of dislocations, and deep levels at or
  close to the dislocation core which seemingly are
  related to a large extent to impurities.
• There are still many open questions ! A lot of
  work for scientists

28
Thank you for your attention!
29
Spin-dependent reactions of impurities at
dislocations Oxygen
Theoretical calculations (R.Jones team) usually
gives several possible configurations of impurity
at dislocation core
Magnetic field (with or without MW field (EPR))
can change the spin configuration
(singlet-triplet) in thermally excited state and
thus stimulate the change in defect configuration
or its dissociation. (spin-dependent chemical
reactions).
30
The concentration and state of impurity (for
example Oxygen in CZ-Si) at dislocation can be
monitored by measurements of Unlocking stress,
nessesary to make dislocation mobile again.
Exposition of dislocations decorated by oxygen in
static magnetic field at room T results in a
strong reduction of pinning effect of oxygen
accumulated at dislocation due to change in its
configuration stimulated by spin dependent
reactions (see also I.Yonenaga, K.Takahashi,
Journal of Physics Conference Series 51, 407
(2006))
M.V.Badylevich, Yu.L.Iunin., V.V.Kveder,
V.I.Orlov, Yu.A.Osipyan, JETP, 97, 601-605
(2003)
31
Heating of spin-system in EPR conditions make
opposite effect compared with static magnetic
field.
30 minutes at ?295?, ?(9.5???)200 ???. Before
EPR treatment dislocations were contaminated by
oxygen atoms at ?600??
Unlocking stress measured at 600oC decreases
after 30min exposition in magnetic field B
(proportionally to B), but increases after
keeping sample in EPR conditions Two different
spin-dependent reactions occur in the same time
32
the D1, D2 luminescence

• D1 dominates at high T - it corresponds to
  deeper (and more populated) states
• D1 is NOT a phonon replica of D2 (shown also by
  M.Kittler)
• Transitions between 2 levels one is coupled to
  conduction, another to valence band
• Radiation recombination at D1D2 is much faster
  than Ev-Ec no emission at 1.1ev !

33
Dependence on excitation level

• At low T tendency to saturation
• Saturation for deeper states occurs at lower
  excitation level
• Can we make a laser with that !?

34
From optical absorption we can estimate the
concentration ND of D1-D2 centres
Using Einstein relation between the rates of
spontaneous (1/ tR) and stimulated transition one
can easily deduce for ND ND(8pn2 e tR/c2)
?a(n) dn We have already estimated tR as
(2-3)10-7sec. It gives us ND(3-5)1014cm-3
for Ndis(3-5)108cm-2 So, distance between
D1D2 centres along dislocation is of about
10nm. Fits to idea of jogs, constrictions etc. at
dislocation
35
The nature of D1, D2 is still unknown !
Dislocation jogs, segments of Lomer dislocations,
multi-vacancy and/or self-interstitial clusters
trapped in the core ??
Puzzle why we do not see the D1-centers in DLTS ?
Spectra of dislocation luminescence in silicon
samples with different dislocation densities
ND. The spectra are normalized to show the same
integral intensity.