BO-Related Defects : Overcoming Bulk Lifetime Degradation in Crystalline Si by Regeneration

%2UHODWHG'HIHFWV2YHUFRPLQJ%XON/LIHWLPH'HJUDGDWLRQ LQFU\VWDOOLQH6LE\5HJHQHUDWLRQ

*LVR+DKQ

^D

6YHQMD:LONLQJ

^E

DQG$[HO+HUJXWK

^F

8QLYHUVLW\RI.RQVWDQ]'HSDUWPHQWRI3K\VLFV.RQVWDQ]*HUPDQ\

DJLVRKDKQ#XQLNRQVWDQ]GH^EVYHQMDZLONLQJ#XQLNRQVWDQ]GH^FD[HOKHUJXWK#XQLNRQVWDQ]GH

.H\ZRUGVFU\VWDOOLQH6LVRODUFHOOERURQR[\JHQGHIHFWUHJHQHUDWLRQK\GURJHQ

$EVWUDFW%RURQR[\JHQUHODWHGGHIHFWVIRUPHGXQGHUZRUNLQJFRQGLWLRQVRIDF6LVRODUFHOOFDQEH D VKRZVWRSSHU IRU QHZ FHOO FRQFHSWV HQDEOLQJ KLJKHU FRQYHUVLRQ HIILFLHQFLHV 7KH UHFRPELQDWLRQ DFWLYLW\RIWKHVHGHIHFWVFDQEHUHGXFHGWRQHJOLJLEOHYDOXHVE\DUHJHQHUDWLRQSURFHVVXQGHUHOHYDWHG WHPSHUDWXUHV DQG LQ WKH SUHVHQFH RI H[FHVV FKDUJH FDUULHUV LQ WKH 6L EXON ,W LV VKRZQ WKDW WKLV SURFHVVDOVRUHOLHVRQWKHSUHVHQFHRI+LQWKHF6LEXON5HJHQHUDWLRQNLQHWLFVFDQEHVSHGXSE\

KLJKHUWHPSHUDWXUHVKLJKHUFRQFHQWUDWLRQVRIH[FHVVFKDUJHFDUULHUVDQGKLJKHU+FRQFHQWUDWLRQLQ WKH F6L EXON %XW FDUH KDV WR EH WDNHQ WR DYRLG D GHVWDELOL]DWLRQ UHDFWLRQ WDNLQJ SODFH DW KLJKHU WHPSHUDWXUH UHVXOWLQJ LQ WKH %2UHODWHG GHIHFWV EHLQJ DJDLQ SUHVHQW LQ WKH UHFRPELQDWLRQDFWLYH VWDWH $ VWDWH PRGHO ZLWK WKH FRUUHVSRQGLQJ UHDFWLRQ UDWHV EHWZHHQ WKH GLIIHUHQW GHIHFWV VWDWHV GHVFULEHV WKH H[SHULPHQWDO ILQGLQJV DQG FDQ EH XVHG IRU SUHGLFWLRQV RI DQ RSWLPL]HGUHJHQHUDWLRQ SURFHVV

,QWURGXFWLRQ

,QGLFDWLRQV WRZDUGV D ERURQR[\JHQ%2UHODWHG GHJUDGDWLRQ ZHUH ILUVW SXEOLVKHG LQ >@

7KLVHIIHFWFDQVHYHUHO\OLPLWPLQRULW\FKDUJHFDUULHUEXONOLIHWLPHIJELQ2DQG%ULFKFU\VWDOOLQH6L F6LPDWHULDOVOLNHSW\SHPRQRFU\VWDOOLQH&]RFKUDOVNL&]JURZQ6LXSRQLOOXPLQDWLRQ,WVFDOHV OLQHDUO\ ZLWK % FRQFHQWUDWLRQ >%@>@DQGDOPRVWTXDGUDWLFDOO\ZLWKLQWHUVWLWLDO2FRQFHQWUDWLRQ

>2L@ >@ 'HSHQGLQJ RQ >%@ DQG >2L@ D VRFDOOHGIXQGDPHQWDO OLIHWLPH OLPLW FDQ EH FDOFXODWHG >@

UHVXOWLQJLQYDOXHVRIIJEaȝVIRUȍFP&]PDWHULDOZLWKW\SLFDO>2L@RIÂFP7KLV UHGXFWLRQLQIJEVHYHUHO\DIIHFWVVWDEOHFRQYHUVLRQHIILFLHQF\ȘRIVWDWHRIWKHDUWVRODUFHOOVDQGOLPLWV ȘWRYDOXHVaIRUSW\SH&]6LVRODUFHOOVZLWKIXOO$OEDFNVXUIDFHILHOG%6)WHFKQRORJ\

&XUUHQWO\DQHZJHQHUDWLRQRIF6LVRODUFHOODUFKLWHFWXUHLVHQWHULQJLQGXVWULDOPDVVSURGXFWLRQ WKHSDVVLYDWHGHPLWWHUDQGUHDUFHOO3(5&FRQFHSW,QWKHVHVRODUFHOOVUHFRPELQDWLRQDWWKHUHDU VLGHLVVLJQLILFDQWO\UHGXFHGE\XVHRIDVXUIDFHSDVVLYDWLQJGLHOHFWULFOD\HUVWDFNDQGRQO\ORFDO PHWDOFRQWDFWVWRWKHEDVHRIWKHVRODUFHOO:LWKWKLVWHFKQRORJ\KLJKHUHIILFLHQFLHV!FDQEH UHDFKHG LPPHGLDWHO\ DIWHU FHOO IDEULFDWLRQ RQ ODUJH DUHD SW\SH &]6L %XW XSRQ LOOXPLQDWLRQ WKH IRUPDWLRQRI%2UHODWHGGHIHFWVOLPLWVȘWRYDOXHVRQȍFPPDWHULDO7KHUHIRUHVWUDWHJLHVWR DYRLGWKHVHYHUHGHJUDGDWLRQGXHWR%2UHODWHGGHIHFWVDUHQHFHVVDU\WRH[SORLWWKHIXOOSRWHQWLDORI

%GRSHGSW\SH&]6LVRODUFHOOV

&RQYHQWLRQDOPHDVXUHVWDNHQWRGHDOZLWK%2UHODWHGGHJUDGDWLRQDUH

x UHGXFLQJ >2L@ LQ &]6L HJ E\ PDNLQJ XVH RI PDJQHWLF ILHOGV GXULQJ FU\VWDO JURZWK UHVXOWLQJLQKLJKHUFRVWV

x UHSODFH%E\*DWDNLQJLQWRDFFRXQWWKDWWKHVHJUHJDWLRQFRHIILFLHQWRI*DLV©UHVXOWLQJ LQLQKRPRJHQHRXVGRSLQJFRQFHQWUDWLRQDORQJWKHLQJRWKHLJKW

x DYRLGXVHRISW\SHZDIHUVDQGXVHHJ3GRSHGQW\SHZDIHUVLQVWHDGJHQHUDOO\OHDGLQJ WRPRUHFRPSOH[VRODUFHOOSURFHVVLQJDQGWKHUHIRUHKLJKHUFRVWV

x UHGXFH>%@E\XVLQJȍFPPDWHULDOLQFUHDVLQJRKPLFORVVHVHVSHFLDOO\IRU3(5&W\SH VRODUFHOOVGXHWRWKHODWHUDOPRYHPHQWRIPDMRULW\FDUULHUVWRWKHORFDOEDFNFRQWDFW

$VDOOPHDVXUHVPHQWLRQHGDERYHKDYHGLVDGYDQWDJHVQRLQGXVWULDOVROXWLRQKDVEHHQIRXQG\HW

Konstanzer Online-Publikations-System (KOPS) URL: http://nbn-resolving.de/urn:nbn:de:bsz:352-0-322951 Erschienen in: Solid State Phenomena ; 242 (2016). - S. 80-89

https://dx.doi.org/10.4028/www.scientific.net/SSP.242.80

5HJHQHUDWLRQRI%2UHODWHG'HIHFWV

7KHGHJUDGDWLRQRI%2UHODWHGGHIHFWVXQGHULOOXPLQDWLRQVKRZVWHPSHUDWXUHGHSHQGHQWNLQHWLFV IROORZLQJ WKH $UUKHQLXV ODZ DQG FDQQRW EH VSHG XS ZLWK VWURQJHU LOOXPLQDWLRQ OHYHOV DV NLQHWLFV VDWXUDWHVDWLOOXPLQDWLRQOHYHOVRIaVXQV>@7KHGHJUDGDWLRQLVWRWDOO\UHYHUVLEOHE\DQQHDOLQJ LQWKHGDUNDWaƒ&IRUVHYHUDOPLQXWHV%XWXSRQLOOXPLQDWLRQGHJUDGDWLRQWRZDUGVWKHGHJUDGHG VWDWHVHWVLQDJDLQ

,Q+HUJXWKHWDOGLVFRYHUHGDWKLUGVWDWHRIWKH%2UHODWHGGHIHFWZKLFKVKRZVQHJOLJLEOH UHFRPELQDWLRQDFWLYLW\OLNHWKHDQQHDOHGVWDWHEXWLVVWDEOHXSRQIXUWKHULOOXPLQDWLRQ)LJ7KLV VWDWH FDQ EH UHDFKHG IURP WKH GHJUDGHG VWDWH DW HOHYDWHG WHPSHUDWXUHV HJ ƒ& DQG LQ WKH SUHVHQFHRIH[FHVVFKDUJHFDUULHUVǻQLQWKHF6LEXON>@5HJHQHUDWLRQNLQHWLFVDOVRIROORZVWKH

$UUKHQLXVODZDQGLVDGGLWLRQDOO\GHSHQGHQWRQǻQZLWKKLJKHUǻQVSHHGLQJXSUHJHQHUDWLRQ

)LJ VWDWH PRGHO RI WKH %2UHODWHG GHIHFW ZLWK UHVSHFWLYH UHDFWLRQV EHWZHHQ GLIIHUHQW VWDWHV DIWHU>@

$VUHJHQHUDWLRQNLQHWLFVLVGHSHQGHQWRQǻQUHJHQHUDWLRQWLPHFRQVWDQWVRQVRODUFHOOOHYHOYDU\

GHSHQGLQJRQELDVLQJFRQGLWLRQV+LJKHVWǻQLVUHDFKHGDWRSHQFLUFXLWYROWDJH9RFZKHUHDVǻQLV UHGXFHG DW PD[LPXP SRZHU SRLQW PSS DQG KDV D PLQLPXP DW VKRUW FLUFXLW FXUUHQW GHQVLW\ MVF FRQGLWLRQ7KLVPHDQVWKDWUHJHQHUDWLRQLVIDVWHUDW9RFFRPSDUHGWRPSSRUMVFFRQGLWLRQ>@2Q WKHRWKHUKDQGDEHWWHUVXUIDFHSDVVLYDWLRQDWWKHUHDUIRUWKH3(5&WHFKQRORJ\OHDGVWRKLJKHUǻQLQ WKH FHOO FRPSDUHG WR IXOO $O %6) WHFKQRORJ\ 8QGHU WKH DVVXPSWLRQ RI NHHSLQJ WKH IURQW VLGH WHFKQRORJ\DQGEXONGLIIXVLRQOHQJWKFRQVWDQWWKLVVKRXOGOHDGWRIDVWHUUHJHQHUDWLRQIRU3(5&W\SH VRODUFHOOV>@

7KH,PSDFWRI+\GURJHQIRU5HJHQHUDWLRQ.LQHWLFV

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

,QWKHIROORZLQJDVHULHVRIH[SHULPHQWVZDVFDUULHGRXWWRSURYHWKLVK\SRWKHVLV7KHHIIHFWRI EDUULHUOD\HUVEHWZHHQ+ULFK3(&9'6L1[+DQGF6LDQGZDVFKHFNHGE\LQVHUWLQJ$O2OD\HUV RIYDU\LQJWKLFNQHVVGHSRVLWHGE\DWRPLFOD\HUGHSRVLWLRQDQGDOORZLQJRQO\YHU\VORZUHJHQHUDWLRQ ZKHQXVHGZLWKRXWD6L1[+FDSSLQJOD\HU>@:KLOHDQQPWKLFN$O2OD\HUGRHVQRWVORZ

82

down the regeneration process after a firing step, a 16 nm thick layer does. This can be explained by the reduced H diffusion from the SiNx:H through the thicker Ah0₃ layer. It was also observed, that a deposited PECVD SiNx:H without following firing step to release large ammmts ofH into the c-Si bulk does not lead to a significant regeneration effect [14].

For time resolved measurements during regeneration on lifetime samples, effective lifetime 'teff

(including surface recombination as well) is normally measured directly after an annealing step without any BO-related defects in the degraded state (Fig. 1) and then several times during regeneration. Assuming Shockley-Read-Hall recombination to be the dominant process in the bulk, the measured 'teff can be transformed into a n01malized defect concentration N* (t) via

• 1 1

N (t)= - -- - - -

reff (t) Teff,annealed

(1)

because lifetime of a certain SRH recombination cha1111el scales inversely to the defect density. A fit of the decaying N* (t) during regeneration yields a characteristic regeneration time constant to or a regeneration rate 1/t{) via

(2)

Iii' 0.008 Iii' _0.012

:J.. :J..

.._

l ..

SiNx:H only .._

... ;s

0.010 AI203 (8 nm) I SiNx:H

-

?

^0.006

•

^~

AI203 (16 nm) I SiNx:H

..

0.008

z • ^z

c _{all fired} c

0 0 0.006

~ _... ^{0.004 :;::.} co _...

-

^c _~

-

^{c 0.004}

0.002 ~

c c _0.002

0 0

(,) (,)

6 SiNx:H not fired

• SiNx:H fired

-

^(,)

-

^{(,) 0.000}

~ 0.000 ~

Q) 0 200 400 600 800 ^Q)

0 0 0 25 50 75 100 125

Regeneration time [min] Regeneration time [min]

Fig. 2: Effect of ALD Ah0₃ layers of varying thickness as H diffusion bani.er during firing on regeneration kinetics (left). Effect of firing of a PECVD SiNx:H layer on regeneration kinetics (right). For both investigations n01malized BO-related defect concentration after anneal (first data point) and dmi.ng subsequent regeneration at 130°C ar1d 0.6 snns illumination is given. (after [14])

In addition, another method for incorporation of H in the c-Si bulk was applied. Samples treated in a microwave-induced remote H plasma at temperatures <250°C for incorporation of H led also to a significant regeneration, even without an additional firing step which in this case is not needed to diffuse H into the c-Si bulk [15].

Impact of Hydrogen Concentration. If H is a prerequisite for regeneration to occur, then the question arises whether regeneration kinetics can be sped up with increasing [H]. To check this, again a seri.es of experiments was conducted by varying the PECVD SiNx:H firing step in a conventional fast firing belt furnace [ 16, 17]. Here a higher peak fui.ng temperature leads to more H being released from the SiNx:H layer into the c-Si bulk and a faster regeneration. In addition the belt speed was var·ied, resulting in different temperature ramps dmi.ng cool-down. For higher peak firing temperature, a fast cool-down speeds up the regeneration process, while for low peak firi.ng

temperatures <700°C a fast cool-down does not accelerate regeneration (Fig. 3). This can be explained by a quenching effect of H: if temperatme is high, H can effuse out of c-Si during cool- down faster than additional H can enter the c-Si bulk from the SiNx:H source. This suggests that [H]

in the c-Si bulk is another imp01iant parameter aprui from temperature and M to describe and understand regeneration kinetics.

c

0.16

..._ E

..-

~

.$ ~

0.12

c 0.08

:;::; 0

~ ~ 0.04

a.>

0>

a.>

a::: 0.00

0.006

fast

0.004 0.002

600 650 700 750 800 850 900 Peak sample temperature [⁰C]

0.16

c

~ E 0.12

a.>

-

_{~ 0.08}

Peak Temp. [•C]

- •- 850 +/-30

T

^fast

- · - 800 +I-30 _..,_ 750 +/-20

- •- 650 +/- 15 medium

- T- 700 +/-15 ~

c !

i

^{0.04 / /}

a3 •

^{slow ./ _______.. _ _}

~

0> •

·~

^t

~ 0.00 '--,-~-.--~-.---~-....--~-...-'

20 30 40 50 60

Average cooling rate 700°C-

ssooc

[Kis]

Fig. 3: Dependence of regeneration kinetics on peak firing temperature and belt speed (left). For peak firing temperatures <700°C, higher cooling rates do not lead to faster regeneration anymore.

(after [17])

Description of Defect Kinetics via Reaction Rates

A vety instmctive model for explaining the observed kinetics during regeneration was published in 2010 based on the occUlTing transitions between the different states of the BO-related defect (Fig. 1) [18]. A set of lineru· differential equations describes the occupation of a respective staten; and the reaction towards or out of this specific state via reaction rates k;j with iij. The respective reaction constants can be influenced by pru·ameters like temperature, M , or [H]. An example for the change in occupation of the degraded state B ns can therefore be given as

(3)

With this model, the behavior of recombination activity can be simulated. Under the assumptions that only reaction constants kAB (degradation) and ksc (regeneration) have values >0 and kAB =10 ksc, the resulting occupation of states in the course of time is shown in Fig. 4 (left).

Struiing after an annealing step (with all defects being in state A), the degraded state B is populated.

Once there are defects available in state B, the regeneration reaction sets in and state C gets populated. As both states A and C show negligible recombination activity, the lifetime measurement will follow the line representing the sum of defects being in state A and state C. h1 Fig. 4 (right) a lifetime measurement in the course of time during regeneration is shown (1 slm illumination at vmying temperatures). The fit according to the model described above leads to excellent matching and therefore proves that the 3-state model can describe the situation vety well.

84

100 80

;R :::...

c 60

-

^...

I

I ...

, ^...

, ^...

I ...

:+:; 0

ro 40

a. :J

8

20

0 ^{0 I I I I I I I}

\~

^{• '. • '-•. ..-:· • /· ;· ~*' /}

0 10

... -·'p

--- -· -·

/

.,...,.·"' _{- -nA(t)}

; '

.

_----n

8{t) ... ; '

>· _nc(t)

. ... ^{. .....}

/ ... _...

... _{... -·-·-}

nA + nc ... ...

... ...

-

^...

_{--- -}

20 30 40

Time [a.u.]

150~~---.~-.---.-~~-~

135 120

~ 105

C/)

2; 90

l-' ~

75 60

45 0 15 30 45 Time [min]

c 140°C

..

120oc

v 100oc 60 75 90

Fig. 4: Occupation of different states of the BO-related defect (Fig. 1) according to the model described in the text (left). The measmed minority canier lifetime values during regeneration of an annealed sample (1 sm1 at temperatmes given in the graph) can be fitted vety well according to the model assuming that both states A and C have negligible recombination activity (right). (after [18])

Effect of fast and slow Regeneration on long term Equilibrium of Defect States. All fom reactions shown in Fig. 1 are thetmally activated and of Anhenius type, describing the probability of conversion p conv from one state to another

(4)

with characteristic trial frequency Vchar, activation energy Ea, Boltzmann constant kB and temperatme T. While anneal and destabilization reaction rate depend only on T, the degradation reaction rate is influenced also by Lill (generated, e.g., by varying illumination intensity) up to illumination levels satmating for values >0.01 SllllS [6]. This mem1s that for fixed values ofT and illumination levels

>0.01 Sllll, degradation, anneal and destabilization reaction rates show fixed values, whereas regeneration reaction rate depends also strongly on specific values of Lill (adjusted, e.g., by illumination level) and [H]. Table 1 gives an overview of values for Ea and reaction rates for T = 200°C.

Table 1: Activation energies and reaction rates at 200°C and 1 Sllll illumination for reactions shown in Fig. 1. (values from [19-22])

Reaction path Ea reVl Time constants rminl

Annealing 1.3 0.25

Degradation 0.4 0.03

Regeneration 1.0 vatying Destabilization 1.25 60

As all fom reactions shown in Fig. 1 ru·e active in parallel, the ratio of the time constants determines the long te1m equilibrium values of occupancy for the three defect states. For a possible industrial application, a fast and complete regeneration process is desired. As regeneration is sped up by higher T, it is interesting to investigate whether T can be increased independently of the specific sample with still all defects ending up in the regenerated state in long te1m equilibrium, or whether there is a peak T that should not be exceeded. At higher T the destabilization reaction struts to compete with regeneration, therefore the ratio between the time constants of regeneration and

destabilization becomes impmtant. If this ratio is 1/100, almost all defects (99%) can reach the regenerated state in long tenn equilibrium (Fig. 5, left). If this ratio is only 1/2, then only ~60% of the defects end up in the regenerated state, and also the annealed state stays populated (Fig. 5, right). As the only difference between the two scenru·ios shown in Fig. 5 is the time constant of regeneration (0.04 min vs. 33 min, other time constants fi:om Tab. 1), it becomes clear that for a complete regeneration at high T (200°C or above) the srunple has to have a shmt regeneration time constant.

~ 1.0 -,--~..__

-

^<1l

-

^C/)

t>

2 <1>

u

0

annealed

0.5

regenerated

2 1.0, ---C/)

1!! C/)

t>

~ <1>

~ 0.5

0

~

annealed

c

<1l Q.

::J

(.) (.) 0.0+--=---~

0

1 E-4 1 E-3 0.01 0.1 1 Time [min]

annealed

10 100 1000

Fig. 5: Occupancy of BO-related defects with time under illmnination (1 sun) at 200°C, struting after annealing. For fast regenerating samples (assumed regeneration time constant: 0.04 min) all defects end up in the regenerated C in long term equilibrium (left), while a regeneration time constant of 33 min leads to incomplete regeneration with a fraction of defects ending up in the (unstable) annealed state A (right). (after [19])

High Speed Regeneration. The regeneration time constant at a given T can be influenced by Llli and [H]. This means that aprut from adjusting illmnination to speed up regeneration, the main parruneter detennining whether regeneration is complete at given T and Llli is the amount of H in the c-Si bulk present in the fmm needed for the regeneration reaction to occm [14]. The combination of chosen Llli and distribution of H therefore detetmines the peak T allowed for reaching complete regeneration.

Fig. 6 visualizes what this means in practice by plotting the completeness of regeneration in long tenn equilibrium in dependency of regeneration time constants with process temperatme as parameter. For regeneration temperatme of 200°C it can be concluded that a regeneration time constant <1 min is needed to end up with almost all defects in the regenerated state C, whereas a time constant of ~50 min will lead only to 50% of the defects in the regenerated state. At 160°C, a time constant of 50 min will lead to almost complete regeneration, and a time constant of 50 min at 130°C to complete regeneration.

86

...

(])

...

m 100

C/)

"0

2

~ 80

me...

' -(]) c c .Q 60

(])+-'

O>ro

(]) ' -

'- .3

40 ..._ ro

0 C/)

>- '- 20

(.) (]) CQ: m ro

a.

(.) ::3

0

1 min

\ \

\

. _.

. - - - ,•

- 130°C \

- - 160°C •

• • • 200°C - · -240°C - · · •280°C

\

\

\

.

\

.

\

'

_\

50 min

' ... . ... . . . ....

^_

^'

^,

(.)

0 10"² 10"¹ 10° 10¹ 10² 10³ 10⁴ 10⁵ 10⁶ 10⁷ Regeneration time constant [s)

Fig. 6: Dependency of completeness of regeneration in long term equilibrimn on regeneration time constants with T as parameter. (after [19])

For fast regenerating samples, regenerating temperature can be increased (with also illumination intensity being increased), so that regeneration time constants in the range of 1 scan be reached [14, 23, 24). Generation of Llli is also possible via laser illumination leading to regeneration time constants <1 s [26].

Model of Regeneration

A general model explaining the observed dependency of regeneration kinetics on [H] has been developed [14, 15]. At temperatures <100°C, His normally bound to impurities in c-Si (e.g. dopants or other impurities/defects) or to H forming molecular H. Assuming that at least pa1t of the H is bmmd in a configuration so that dissociation is enhanced tmder carTier injection, e.g., in BH-pairs with a dissociation energy being reduced to 1.1 eV under injection conditions [14], this might explain the observed necessity for carTier injection for regeneration to occur and the activation energy for the regeneration reaction of 1.0 eV (Table 1). Slow cool-down rates after peak firing temperature can result in less overall [H] in the c-Si bulk and to H being botmd to other defects than B [14, 25], showing a more stable configuration with higher dissociation energy. This applies also to temperatm·e steps ar·otmd 100-400°C being canied out after cool-down as shown in [14, 25].

Therefore, the temperature history after peak firing (introduction of H into the c-Si bulk) determines the regeneration rate at given regeneration temperature and &1.

Depending on Llli, the charge state of H can be changed in p-type c-Si material. E.g., upon illmnination part of the normally positively charged

W

can be present in the neutral state H^0. The importance of the H char·ge state for mobility of H in c-Si and the possible bom1ding to and therefore passivation of defects like, e.g., the BO-related defect was highlighted recently by several authors [24, 26, 27). ill combination with the model presented in the paragraph above, this might explain the observed need for hydrogen, temperature ar1d Llli for regeneration to occur.

Within the model also the observed dependency of regeneration rate on the amount of dopants, with higher dopant density leading to lower regeneration rates [28], can be explained at least qualitatively. The more dopants present in the c-Si bulk, the more H is trapped at these defect sites and the slower is the trap limited diffusion of H towards the BO-related defect. This behavior was also observed in compensated c-Si, where for compensation of the present P more B (and possibly also Ga) is present, significantly increasing the overall p-type dopant concentration and slowing down regeneration rate despite of the same net doping concentration [19, 28-30). A longer regeneration time constant then allows only lower regeneration temperatures at constant Llli for complete regeneration (see above) or leads to incomplete regeneration as observed experimentally [19).

6XPPDU\

7KHUHJHQHUDWLRQSURFHVVRI%2UHODWHGGHIHFWVOHDGVWRDVWDWHRIWKHGHIHFWWKDWVKRZVQHJOLJLEOH UHFRPELQDWLRQDFWLYLW\DQGQRGHWHFWDEOHGHJUDGDWLRQRIOLIHWLPHXSRQIXUWKHULOOXPLQDWLRQDWURRP WHPSHUDWXUH7KLVVWDWHFDQEHUHDFKHGE\DSSO\LQJPRGHUDWHWHPSHUDWXUHVLQWKHSUHVHQFHRIH[FHVV FKDUJHFDUULHUVDQG+LQWKHF6LVDPSOH

7KH QHFHVVLW\ RI + LQ F6L EHLQJ UHOHDVHG IURP HJ 6L1[+ OD\HUV GXULQJ D KLJK WHPSHUDWXUH ILULQJVWHSIRUDIDVWUHJHQHUDWLRQWRRFFXUFRXOGEHVKRZQE\DVHULHVRIH[SHULPHQWV,WFRXOGDOVR EH VKRZQ WKDW D KLJKHU FRQFHQWUDWLRQ RI + LQ WKH F6L EXON OHDGV WR KLJKHU UHJHQHUDWLRQ UDWHV SURYLGHG WKDW + LV SUHVHQW LQ D IRUP WKDW DOORZV D KLJK PRELOLW\ RI DWRPLF + XQGHU UHJHQHUDWLRQ FRQGLWLRQV

$GHVFULSWLRQRIUHDFWLRQVEHWZHHQWKHGLIIHUHQWVWDWHVRIWKH%2UHODWHGGHIHFWYLDUDWHHTXDWLRQV JLYHV LPSRUWDQW LQVLJKWV LQ WKH NLQHWLFV RI WKH GHIHFW WUDQVLWLRQV XQGHU VSHFLILF H[SHULPHQWDO FRQGLWLRQV DQG DOORZV SUHGLFWLRQV IRU GHVLJQLQJ D SURFHVV WKDW OHDGV WR DOPRVW FRPSOHWH UHJHQHUDWLRQRIDWKH%2UHODWHGGHIHFWV$SDUWIURPUHJHQHUDWLRQWHPSHUDWXUHǻQDQG>+@LQWKHF 6LEXONRIWKHVDPSOHWREHUHJHQHUDWHGSOD\DFUXFLDOUROH

)DVW UHJHQHUDWLQJ VDPSOHV FDQ EH FRPSOHWHO\ UHJHQHUDWHG ZLWKLQ VHFRQGV SURYLGHG WKDW ǻQ LV DGDSWHGDWKLJKUHJHQHUDWLRQWHPSHUDWXUHDQG+LVSUHVHQWLQVXIILFLHQWFRQFHQWUDWLRQDQGVXLWDEOH IRUP

$ PRGHO GHVFULELQJ WKH UHJHQHUDWLRQ SURFHVV ZDV SUHVHQWHG DQG LV EDVHG RQ + EHLQJ ERXQG WR D GHIHFWWKDWDOORZVGLVVRFLDWLRQDWUHJHQHUDWLRQWHPSHUDWXUHDQGSUHVHQFHRIH[FHVVFKDUJHFDUULHUV

$QH[DPSOHLV+EHLQJERXQGWR%ZLWKDFWLYDWLRQHQHUJ\RIH9XQGHULOOXPLQDWLRQFORVHWRWKH RQHREVHUYHGIRUUHJHQHUDWLRQH90RUHWUDSVIRU+LQWKHF6LEXONOHDGWRORZHUPRELOLW\RI +GXULQJUHJHQHUDWLRQDQGWKHUHIRUHVORZGRZQWKHUHJHQHUDWLRQSURFHVV7KLVOHDGVWRLQFRPSOHWH UHJHQHUDWLRQLIUHJHQHUDWLRQWHPSHUDWXUHLVFKRVHQWRRKLJK

$FNQRZOHGJHPHQWV

3DUWRIWKLVZRUNZDVIXQGHGE\WKH*HUPDQ%08XQGHUFRQWUDFWVDQG$7KH FRQWHQWLVWKHUHVSRQVLELOLW\RIWKHDXWKRUV

5HIHUHQFHV

>@ +)LVFKHU:3VFKXQGHU,QYHVWLJDWLRQRISKRWRQDQGWKHUPDOLQGXFHGFKDQJHVLQVLOLFRQVRODU FHOOVLQ3URF^WK,(((396&3DOR$OWR86$SS

>@ 6:*OXQ]65HLQ::DUWD-.QREORFK::HWWOLQJ'HJUDGDWLRQRIFDUULHUOLIHWLPHLQ

&]VLOLFRQVRODUFHOOV6RO(Q0DW6RO&HOOV

>@ -6FKPLGW$&XHYDV(OHFWURQLFSURSHUWLHVRIOLJKWLQGXFHGUHFRPELQDWLRQFHQWHUVLQERURQ GRSHG&]RFKUDOVNLVLOLFRQ-$SSO3K\V

>@ 6 5HLQ 5 )DOVWHU 6: *OXQ] 6'LH]4XDQWLWDWLYHFRUUHODWLRQRIWKHPHWDVWDEOHGHIHFWLQ

&]VLOLFRQZLWKGLIIHUHQWLPSXULWLHVLQ3URF^UG:&3(&2VDND-DSDQSS

>@ . %RWKH 5 6LQWRQ - 6FKPLGW )XQGDPHQWDO ERURQR[\JHQUHODWHG FDUULHU OLIHWLPH OLPLW LQ PRQRDQGPXOWLFU\VWDOOLQHVLOLFRQ3URJ3KRWRYROW5HV$SSO

>@ -6FKPLGW.%RWKH6WUXFWXUHDQGWUDQVIRUPDWLRQRIWKHPHWDVWDEOHERURQDQGR[\JHQUHODWHG GHIHFWFHQWHULQFU\VWDOOLQHVLOLFRQ3K\V5HY%

>@ $+HUJXWK*6FKXEHUW0.DHV*+DKQ$QHZDSSURDFKWRSUHYHQWWKHQHJDWLYHLPSDFWRI WKHPHWDVWDEOHGHIHFWLQERURQGRSHG&]VLOLFRQVRODUFHOOV,Q3URF^WK:&3(&:DLNRORD86$

SS

>@ $ +HUJXWK * 6FKXEHUW 0 .DHV * +DKQ $YRLGLQJ ERURQR[\JHQ UHODWHG GHJUDGDWLRQ LQ KLJKO\ERURQGRSHG&]VLOLFRQ,Q3URF^VW(&396(&'UHVGHQ*HUPDQ\SS

>@ $+HUJXWK0.DHV*6FKXEHUW*+DKQ)XUWKHULQYHVWLJDWLRQVRQWKHDYRLGDQFHRIERURQ R[\JHQUHODWHGGHJUDGDWLRQE\PHDQVRIUHJHQHUDWLRQ,Q3URF^QG(&396(&0LODQ,WDO\

SS

>@$ +HUJXWK * 6FKXEHUW 0 .DHV * +DKQ ,QYHVWLJDWLRQVRQWKHORQJWLPHEHKDYLRURIWKH PHWDVWDEOHERURQ±R[\JHQFRPSOH[LQFU\VWDOOLQHVLOLFRQ3URJU3KRWRYROW5HV$SSO

>@$ +HUJXWK 5 +RUEHOW 6 :LONLQJ 5 -RE * +DKQ &RPSDULVRQ RI %2 5HJHQHUDWLRQ G\QDPLFVLQ3(5&DQG$O%6)VRODUFHOOVVXEPLWWHGWR(QHUJ\3URFHGLD

>@.$0Q]HU+\GURJHQDWHGVLOLFRQQLWULGHIRUUHJHQHUDWLRQRI/LJKW,QGXFHG'HJUDGDWLRQ,Q 3URF^WK(8396(&+DPEXUJ*HUPDQ\SS

>@* .UXJHO : :RONH - *HLONHU 6 5HLQ 53UHX,PSDFWRIK\GURJHQFRQFHQWUDWLRQRQWKH UHJHQHUDWLRQRI/LJKW,QGXFHG'HJUDGDWLRQ(Q3URF

>@6:LONLQJ&%HFNK6(EHUW$+HUJXWK*+DKQ,QIOXHQFHRIERXQGK\GURJHQVWDWHVRQ

%2UHJHQHUDWLRQ NLQHWLFV DQG FRQVHTXHQFHV IRU KLJKVSHHG UHJHQHUDWLRQ SURFHVVHV 6RO (Q 0DW 6RO&HOOV

>@6:LONLQJ$+HUJXWK*+DKQ,QIOXHQFHRIK\GURJHQRQWKHUHJHQHUDWLRQRIERURQR[\JHQ UHODWHGGHIHFWVLQFU\VWDOOLQHVLOLFRQ-$SSO3K\V

>@6 :LONLQJ 6 (EHUW $ +HUJXWK * +DKQ ,QIOXHQFH RI VKRUW KLJK WHPSHUDWXUH VWHSV RQ WKH UHJHQHUDWLRQRIERURQR[\JHQUHODWHGGHIHFWVLQ3URF^WK(8396(&3DULV)UDQFHSS

>@6:LONLQJ6(EHUW$+HUJXWK*+DKQ,QIOXHQFHRIK\GURJHQHIIXVLRQIURPK\GURJHQDWHG VLOLFRQ QLWULGH OD\HUV RQ WKH UHJHQHUDWLRQ RI ERURQR[\JHQ UHODWHG GHIHFWV LQ FU\VWDOOLQH VLOLFRQ -

$SSO3K\V

>@$+HUJXWK*+DKQ.LQHWLFVRIWKHERURQR[\JHQUHODWHGGHIHFWLQWKHRU\DQGH[SHULPHQW-

$SSO3K\V

>@6 :LONLQJ 0 )RUVWHU $ +HUJXWK*+DKQ)URPVLPXODWLRQWRH[SHULPHQWXQGHUVWDQGLQJ

%2UHJHQHUDWLRQNLQHWLFVVXEPLWWHGWR6RO(Q0DW6RO&HOOV

>@-6FKPLGW.%RWKH6WUXFWXUHDQGWUDQVIRUPDWLRQRIWKHPHWDVWDEOHERURQDQGR[\JHQUHODWHG GHIHFWFHQWHULQFU\VWDOOLQHVLOLFRQ3K\V5HY%

>@65HLQ75HKUO::DUWD6:*OXQ]*:LOOHNH(OHFWULFDODQGWKHUPDOSURSHUWLHVRIWKH PHWDVWDEOHGHIHFWLQERURQGRSHG&]RFKUDOVNLVLOLFRQ&]6LLQ3URF^WK(&396(&0XQLFK

*HUPDQ\SS

>@6:*OXQ](6FKlIIHU65HLQ.%RWKH-6FKPLGW$QDO\VLVRIWKHGHIHFWDFWLYDWLRQLQ&]

VLOLFRQ E\ WHPSHUDWXUH GHSHQGHQW ELDVLQGXFHG GHJUDGDWLRQ RI VRODU FHOOV LQ 3URF ^UG :&3(&

2VDND-DSDQSS

>@6:LONLQJ-(QJHOKDUGW6(EHUW&%HFNK$+HUJXWK*+DKQ+LJKVSHHGUHJHQHUDWLRQRI

%2GHIHFWVLPSURYLQJORQJWHUPVRODUFHOOSHUIRUPDQFHZLWKLQVHFRQGV,Q3URF^WK(8396(&

$PVWHUGDP1HWKHUODQGV

88

>@%-+DOODP3*+DPHU6:DQJ/6RQJ11DPSDOOL0'$EERWW&&KDQ'/X$0 :HQKDP / 0DL 1 %RURMHYLF $ /L ' &KHQ 0< .LP $ $]PL 6 :HQKDP $GYDQFHG K\GURJHQDWLRQRIGLVORFDWLRQFOXVWHUVDQGERURQR[\JHQGHIHFWVLQVLOLFRQVRODUFHOOVVXEPLWWHGWR (Q3URFHGLD

>@6$0F4XDLG0-%LQQV5&1HZPDQ(&/LJKWRZOHUV-%&OHJJ6ROXELOLW\RIK\GURJHQ LQVLOLFRQDWƒ&$SSO3K\V/HWW

>@3 +DPHU % +DOODP 6 :HQKDP 0 $EERWW 0DQLSXODWLRQ RI K\GURJHQ FKDUJH VWDWHV IRU SDVVLYDWLRQRISW\SHZDIHUVLQSKRWRYROWDLFV,(((-3KRWRYROW

>@&6XQ)(5RXJLHX[DQG'0DFGRQDOG$XQLILHGDSSURDFKWRPRGHOOLQJWKHFKDUJHVWDWHRI PRQDWRPLFK\GURJHQDQGRWKHUGHIHFWVLQFU\VWDOOLQHVLOLFRQ-$SSO3K\V

>@6'XERLV1(QMDOEHUW-3*DUDQGHW50RQQD-.UDLHP/LJKW,QGXFHG'HJUDGDWLRQDQG 5HJHQHUDWLRQLQFRPSHQVDWHGXSJUDGHGPHWDOOXUJLFDOVLOLFRQLQ3URF^UG(8396(&9DOHQFLD 6SDLQSS

>@%/LP$/LX'0DFGRQDOG.%RWKHDQG-6FKPLGW,PSDFWRIGRSDQWFRPSHQVDWLRQRQ WKHGHDFWLYDWLRQRIERURQR[\JHQUHFRPELQDWLRQFHQWHUVLQFU\VWDOOLQHVLOLFRQ$SSO3K\V/HWW

>@5 6RQGHQD $ +ROW $. 6RLODQG (OHFWULFDO SURSHUWLHV RI FRPSHQVDWHG Q DQG SW\SH PRQRFU\VWDOOLQHVLOLFRQLQ3URF^WK(8396(&+DPEXUJ*HUPDQ\SS