• Aucun résultat trouvé

STRUCTURAL MODEL OF FLUORINATED AMORPHOUS-SILICON (a-Si : F)

N/A
N/A
Protected

Academic year: 2021

Partager "STRUCTURAL MODEL OF FLUORINATED AMORPHOUS-SILICON (a-Si : F)"

Copied!
5
0
0

Texte intégral

(1)

HAL Id: jpa-00220900

https://hal.archives-ouvertes.fr/jpa-00220900

Submitted on 1 Jan 1981

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

STRUCTURAL MODEL OF FLUORINATED AMORPHOUS-SILICON (a-Si : F)

H. Matsumura, K. Sakai, Y. Kawakyu, S. Furukawa

To cite this version:

H. Matsumura, K. Sakai, Y. Kawakyu, S. Furukawa. STRUCTURAL MODEL OF FLUORINATED AMORPHOUS-SILICON (a-Si : F). Journal de Physique Colloques, 1981, 42 (C4), pp.C4-209-C4-212.

�10.1051/jphyscol:1981443�. �jpa-00220900�

(2)

JOURNAL DE PHYSIQUE

CoZZoque C4, suppZ6ment au nOIO, Tome 42, octobre 1981

STRUCTURAL MODEL OF F L U O R I N A T E D AMORPHOUS-SI L I C O N ( a - ~ i : F)

H.

Matsumura,

K.

Sakai, Y. Kawakyu and

S .

Furukawa

Department of Applied Electronics, Tokyo I n s t i t u t e of Technology, Nagatsuda,, Midori-ku, Yokohama 227, Japan

Abstract: A structural model for fluorinated amorphous-silicon containing no hydrogen (a-Si:F) is presented based on the experimental results of transmission electron microscopic measurement, infrared absorption and Rutherford backscattering measurements and their changes due to chemical etching. It is concluded that aSi:F consists of many amorphous-silicon pains of about 40

A

in size, grain boundaries are filled by SiF; molecular gas and that the enlargement of grain size is a key factor to improve properties of a-Si:F such as photo- conductivity.

Among various types of amorphous-silicon alloys (a-Si), purely fluorinated a- Si containing no hydrogen (a-Si:F) has unique advantage for its heat-resistant proper ty. That is, the a-Si:F, which is produced mainly by the sputtering of silicon in gas mixture of Ar and SiF4 , is heat-resistant up to 60o0c (I), while other a-Si such as hydrogenated one (a-Si:H) changes its properties at a temperature as low as 350°C (2). It has been demonstrated (3,4) that the conductivity

anal

p-n type of a-Si:F can be controlled by the dope of boron or phosphorus as well as a-Si:H. A thin film transistor made from a-Si: F operates almost similarly to that made from a-Si

:

H ( 5 ) .

Eowever, a-Si:F has a defficiency to be overcome for its application to photo-voltaic devices at &he present moment. The photo-conductivity of it is lower than that of a- Si:H by about 4 orders of magnitude. Therefore, it is very useful to investigate a way to improve the photo-conductivity of a-Si:F for wider application of it.

Knights et a1. (6) suggested based on their morphorogical study that the proper- ties of a-Si could be explained by its micro-structure. Thus, as a first step to im- prove the photo-conductivity of a-Si:F, we studied its micro-structure and speculated a most possible structural model for it. And as a special unique technique for study, we observed the change in micrographs of transmission electron microscope

(TEM)

due to anisotropic chemical etching, and compared it with similar change in infrared ab- sorption spectra. And we also determined the positional relation among atomic species in a-Si:F by observing the change of Rutherford backscattering

(RBS)

spectra of a-Si:F due to etching.

Fromthese experiments, we have found as follows: 1) The a-Si:F film consists of many small grains of a-Si net-work. For instance, when fluorine content in the film is about 15 atomic

$,

the grain size is about 40

.%

or niore. 2) The a-Si:F contains considerable amounts of Si-F4 bonds as well as Si-F2 bonds, and 3) the Si-F4 bonds (=SiF4 molecules) are mainly lscalized at the grain boundaries together with Ar atoms used as sputtering gas. Based on these results, finally, we present a most possible structural model for a-Si:F. The model suaests that the enlargement of grain size is a key factor for improvement of properties such as photo-conductivity.

Experiment and Results

The a-Si:F film was deposited by usingthe diode type radio frequency sputtering system. An intrinsic crystalline Si of the resistivity as high as 1000 ncm was used

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1981443

(3)

JOURi'JAL DE PHYSIQUE

Fig.1 TEM micrographs f o r samples deposited Fig.2 TEM micrographs f o r samples a) a t R(SiF4)= a) 0 %, b) 3 % and c) 6 before etching and b) a f t e r

% with thickness of 0.07 pm. etching by 10 % KOH solution.

a s a s p u t t e r i n g t a r g e t , and gas mixture of SiFq and A r a s t h e s p u t t e r i n g gas. The s u b s t r a t e temperature Ts during deposition was 350°c, t h e pressure of s p u t t e r i n g gas Pgas= 4X Torr, t h e s p u t t e r i n g power 200 W and t h e deposition r a t e 2 t o

3

a/sec.

The r a t i o of p a r t i a l pressure of SiF t o t h e t o t a l pressure of gas mixture, denoted by R(SiF4 ) here, was varied from 0

#

t o 6

%.

According t o our previous work ( 3,4)

,

t h e F content i n a-Si:F Eould be proportionally controlled from 0 t o 15 atomic

%

by t h i s v a r i a t i o n of R(SiFt+ ), and t h e e l e c t r o n i c conduction mechanism of t h e a-Si:F changed from t h e v a r i a b l e range hopping type t o t h e a c t i v a t e d type a s R(SiF4) reached t o 6

%.

Figure 1 , a )

,

b ) and c ) show t h e TFM micrographs of a-Si:F f i l m s which were de- posited onto carbon coated micro-grids a t R(SiFq)=O, 3 and 6

%

respectively. The ' c h i c h e s s of f i l m s wasn.07 v q ~ and t h e accelerated voltage of e l e c t r o n s was 200 kV i n t h i s case. This f i g u r e c l e a r l y demonstrates t h a t a s R(biF4) increases, t h a t is, t h e F content increases, g r a i n s t r u c t u r e becomes obvious, and t h a t t h e a-Si:F f i l m deposited a t R(SiF4)=6

%

c o n s i s t s of both many gray-colored g r a i n s of about 40

8

i n s i z e and long dark-colored boundaries surrounding them. This implies t h a t atomic species and/or d e n s i t i e s a r e l o c a l i z e d i n t h e a-Si:F f i l m of R(SiF4)=6

%.

It is well known t h a t t h e brightness in TEM micrographs is very roughly pro- portional t o both t h e r a t i o of atomic number t o t h e mass number, Z/M, and density of atomic species ( 7 ) . Additionally, according t o our RBS measurement, t h e a S i : F f i l m contains only F, S i and A r atoms (3,4). Thus, i f t h e s e atomic s p e c i e s a r e localized and a l s o i f t h e density of them is t h e same each o t h e r , S i atoms (Z/hf=0.5) must cause a s l i g h t l y b r i g h t e r p a t t e r n than F ( Z / ~ = 0 . 4 7 ) and A r atoms ( ~ / ~ 1 = 0 . 4 5 ) . That is, i f we may speculate t h a t t h e gray-colored g r a i n s correspond t o a - S i g r a i n s and t h a t F and Ar atoms a r e a t t h e dark-colored boundaries, it would appear reasonable t o obtain t h e TEM micrograph p a t t e r n a s shown i n Figure 1 , c ) .

A s a s p e c i a l unique technique t o v e r i f y t h e v a l i d i t y of t h i s speculation and t o determine t h e micro-structure of a-Si:F, we observed t h e change of TEM micrographs due t o anisotropic chemical etching of samples, and compared it with s i m i l a r changes i n i n f r a r e d absorption spectra and RBS s p e c t r a . From t h e speculation above, we had expected t h a t some p a r t s of t h e g r a i n surface would be etched off a t f i r s t by aniso- t r o p i c etching, then atomic species which were pinched between g r a i n s would be released. So t h a t , we had thought t h a t t h e atomic species a t t h e g r a i n boundaries could be detected by comparing t h e change of TEM micrographs due t o etching with changes i n i n f r a r e d absorption and RBS s p e c t r a . Here, a s t h e anisotropic etchant, KOH was chosen, s i n c e t h e surface of a-Si g r a i n s could be etched without intermediate oxidation processes by using it.

Figure 2,a) and b ) show t h e TEM micrographs before and a f t e r etching of sample produced a t R(SiF4) =6

%

respectively. The etching was c a r r i e d out a t o'C by 10

%

KOH s o l u t i o n , and t h e i n i t i a l thickness of a-Si:F film, 0,2

u m ,

was reduced t o about007 urnby t h i s etching. Despite of t h e same thichness a s one shown i n F i g . l , c ) , t h e g r a i n s t r u c t u r e is not obvious i n F i g . 2 , b ) . This appears t o imply t h a t t h e p a r t i c u l a r types of atomic species were dissolved i n KOH s o l u t i o n during etching.

Figure 3 shows t h e change of i n f r a r e d absorption spectra due t o etching f o r t h e sample a l s o produced a t R(SiF4)=6

%.

The r e s u l t s of assignment f o r each absorption

(4)

-.-.

.-.--.

%,

,,.'.--...----.-,

thick film

'

-

*"'---

C

In . -

g 5

-

I-

zf

or S i - 5

a-SiF -Ted at:

R(SiF,)= 6 %

T5=3K) Ethed ty:

P =b~lO-~Torr lO% WH at 2O0C

,285

w I I

,

I

Fig.3 Change of i n f r a r e d absorption Fig.4 Change of F t o S i contents and A r spectra due t o etching. t o S i contents due t o etching.

peak i n t h e spectrum ( 8 , 9 ) a r e indicated by arrows. This assignment was a l s o confirm- ed by our recent gas evolution experiment. I n t h i s f i g u r e , t h e etching time of sample i n 10

%

KOH s o l u t i o n a t 20 OC is taken a s a variable parz.~eker. The i n i t i a l thickness, 0.4 urn, was reduced t o 0.2 vm a f t e r t h e etching f o r 90 seconds. A spectrum f o r 0.2 u m thicksample which was not etched is a l s o shown f o r comparison. The f i g u r e c l e a r l y

demonstrates t h a t Si-F4 bonds a r e more l i k e l y t o be etched than other bond configu- r a t i o n . And Figs.2 and 3 apparently suggest t h a t g r a i n boundaries a r e i n i t i a l l y f i l l e d by a t l e a s t Si-Fq bonds, but t h a t t h e Si-F4 bonds were dissolved i n etchant.

Next, we measured t h e change of F and A r contents i n a-Si:F f i l m due t o etching.

The contents of atomic species i n t h e f i l m were measured by RBS method using 2.8-Mev helium ions. Figure 4 shows t h e r e s u l t s f o r samples having t h r e e d i f f e r e n t i n i t i a l thicknesses 0.15, 0.20 and 0.25 urn. I n t h i s f i g u r e , t h e r a t i o of F t o S i contents and A r t o S i contents, denoted by NF/NSi and N&% respectively, a r e described a s a function of r e l a t i v e etched thickness, but e values of N F / ~ S i and N ~ a r e nor- N ~ ~ malized by ones before etching. The s i m i l a r r e s u l t s f o r t h e sample deposited a t R(SiF4)=3 %, i n which t h e g r a i n s t r u c t u r e is not obvious a s shown i n Fig.1, a r e shown together f o r comparison. This f i g u r e demonstrates t h a t t h e reduction of F and A r contents is c l e a r only f o r t h e sample having apparent g r a i n s t r u c t u r e . This f i g u r e a l s o shows another important f e a t u r e s f o r reduction of F and A r contents. That is, t h e reduction r a t e of F content is almost t h e same a s t h a t of A r content. This i n d i c a t e s t h a t A r atoms a r e localized a t t h e position near t o Si-F4 bonds. The local- i z a t i o n of A r atoms near t o Si-F4 bonds can be a l s o confirmed from t h e gas evolution experiment ( l o ) , since t h e evolution of SiF4 gas and A r gas were observed a t t h e same temperature when samples were heated. And these Figs.2, 3 and 4 appear t o confirm t h e v a l i d i t y of our speculation, t h a t is, gray-colored g r a i n s in TEM micrographs corre- spond t o a-Si g r a i n s and A r and F atoms mainly e x i s t a t t h e dark-colored boundaries.

S t r u c t u r a l Model and Discussions

Now, we can imagine a s t r u c t u r a l mgdel f o r a-Si:F f i l m deposited a t ~ ( S i F 4 ) = 6 $.

The f i l m c o n s i s t s of a-Si g r a i n s of 40 A i n s i z e , and t h e grain boundaries a r e f i l l e d by Si-F4 and A r atoms. The f i l m contains Si-F2 bonds a s shown i n Fig. 3. Since t h e existence of Si-F2 bonds tends t o block t h e f u r t h e r growth of a-Si g r a i n , t h e Si-F2 bonds must be concentrated a t t h e surface of a-Si g r a i n s . Thus, we can draw a most possible s t r u c t u r a l model a s shown i n Fig.5. I n t h e f i g u r e , open c i r c l e s , closed squares and closed c i r c l e s r e f e r t o S i , F and A r atoms respectively, and t h e l i n e s between atoms correspond t o s i n g l e bonds.

(5)

C4-2 12 JOURNAL DE PHYSIQUE

Here, finally, we speculate the reason for low photo-conduc- tivity of a-Si:F film based on the structural model. That is, as easily envisaged, the grain struc- ture may act as a block for electron transport. The grain sur- face may have electronic defects due to dangling bonds. And photo- carriers which are generated inside a-Si grains may be trapped at the grain surface. This must be the reason for low photo-conduc- tivity of a-Si:F film.

The spin density of our a-Si:F samples deposited at R(Sick )=6

$

was estimated to be

ROiFi, = 6

- 6

X

10 ' cm-' from ESR measurement

.F OSi @ A r

by Shimizu (11). Supposing that

Fig.5 Structural model for a-Si:F film the value of spin density is equal deposited at R(SiF4)=6

%.

to the dangling bond density at the grain surface, and also supposing that the grain is a sphere or a column of radius r=20

%, a

very rough but suggestive relation between r (8) and the dangling bond density ( ~ m - ~

)

can be derived easily as 1.2 Xl0 19/r for r much larger than Si bond len&h. That is, if we

can

increase the grain size by one order, the dangling bond density is lowered by one order

md,

reaches the equivalent value for glow-discharged a-Si:H. And this discussion is completely consistent

witfn

the fact that the radius r of columns in the glow-discharged a-Si:H appears to in- crease to the order of 100 8 as the conditions for deposition was improved as Knights et al. (6) reported and that the dangling bond density drops to the value lower than our a-Si:F by about one order. That is, the fundamental idea of this model might be applicable even to the glow-discharged a-Si:H.

Conclusions

From above experimental results

ancl discussions, we can conclude as follows:

1) As R(SiF4

)

increases, that is, the F content in a-Si: F increases, grain structure becomes obvious.

2) The Si-F4 bonds, that is, SiF4 molecules, exist at the grain boundaries together with Ar atoms.

3) From these results, a most possible structural model for a-Si:F deposited at R(SiF4)=6 % is presented. The model suggests that the enlargement of grain size is a key factor to improve the properties of a-Si:F such as photo-conductivity.

The authors would like to express their gratitude to Professor Shimizu at University of Kanazawa for his ESR measurement. This work is partially supported by the 1980 Grant-in-Aid for Special Project Research (2) from the Ministry of Education of Japan.

References

1) MATSUMlTRA H. et al., Appl. Phys. Letts., 36 (1980) 439 2). McMILLAN J.A. et al., J. Appl. Phys., 50 n979) 5238

3) MATSUMlTRA H. et al. , Proc 15th Int Conf on Physics of Semiconductors. 1980, p1253 4) MATSUMURA

H .

et al.,

J.

Appl. Phys., 52 (1981) 291

5

)

MATSUMURA H . ,

KANAMORI M. and

FURUKAW~S . , (unpublished) 6) KNIGHTS J.C. et al., Appl. Phys. Letts., 35 (1979) 244

7 )

REIMER

L., Z. Angw. Phys.,

9 (1957) 34

8) LEY L. et al., Proc 15th Int Conf on Physics of Semiconductors, Kyoto 1980, p1241 9) SHIMADA T. et al.,

Jpn.

J. Appl. Phys., 19 (1980) L265

10)

MATSLMUFUl

H. et al. , J. Appl. Phys. (to be published)

11) SHIMIZU T., private communicationand UEDA S. et al., Proc of this Conference.

Références