MAGNETIC FIELD DEPENDENCE OF THE PHOTO- AND DARK CONDUCTIVITY IN DOPED a-Si:H FILMS

HAL Id: jpa-00220884

https://hal.archives-ouvertes.fr/jpa-00220884

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.

MAGNETIC FIELD DEPENDENCE OF THE PHOTO- AND DARK CONDUCTIVITY IN DOPED

a-Si:H FILMS

D. Weller, H. Mell, L. Schweitzer, J. Stuke

To cite this version:

D. Weller, H. Mell, L. Schweitzer, J. Stuke. MAGNETIC FIELD DEPENDENCE OF THE PHOTO-

AND DARK CONDUCTIVITY IN DOPED a-Si:H FILMS. Journal de Physique Colloques, 1981, 42

(C4), pp.C4-143-C4-146. �10.1051/jphyscol:1981428�. �jpa-00220884�

JOURNAL DE PHYSIQUE

CoZZoque C4, suppZ6ment au n O I O , Tome 42, octobre 1981 page C 4 - 1 4 3

MAGNETIC F I E L D DEPENDENCE OF THE PHOTO- AND DARK CONDUCTIVITY I N DOPED

a-Si:H

FILMS

D. Weller, H. Mell, L. Schweitzer and J. Stuke Fachbereich Physik, Universitat Marburg, F. R. G.

Abstract.- We have studied the influence of magnetic field on the photo- conductivity uph and dark conductivity ud of glow discharge deposited a-Si:H films doped with phosphorus or boron up to a nominal content of 7 %. The relative change of uph is of the order of to depending on temperature and on the dopant concentration. Its dependence on magnetic field strength B varies significantly with the position of the Fermi level.

In highly doped material ( B ~ H ~ / S ~ H ~ > and P H ~ / S ~ H * > lob2)

,

we observe - for the first time in this material

-

magnetic field induced changes of ad This magnetoconductivity effect is attributed to a second current path in localized band tail states close to the Fermi level.

Introduction.- Recently we found that the photoconductivity opt, of glow discharge deposited a-Si:H films changes by as much as 2 % if a magnetic field B c l O kGauss is applied1'*. The dependence of Aap~/uph on B is anomalous and strikingly similar to the magnetoresistance effect previously observed in the variable range hopping re- gime of evaporated a-Si and a - ~ e ~ ' ~ . Both effects have been successfully interpreted in terms of a model by Movaghar and schweitzerlf 5 1 6 , in which the recombination rate or hopping rate, respectively, depend on the spin relaxation rate which in its turn is modified by the applied field.

Our previous studies of aph (B) were restricted to undopedl or lightly doped2 a-Si:H films. Here we report on an extension of these investigations to boron and phosphorus contents up to nominally 7 %. In the course of this study we also found that the dark conductivity ud is influenced by the magnetic field if the dopant con- tent exceeds a certain limit for boron and for phosphorus).

Experimental.- The a-Si:H films were deposited onto fused quartz substrates mounted at the grounded anode of a capacitively coupled glow discharge system. SiH4 gas was premixed with PH3 or B2H6 and diluted at a ratio of 1:l with Ar. Pressure during de- position was 0.6 mbar, deposition rate % 3 R/s and substrate temperature 280°c. ad and up: were measured by means of two evaporated Cr electrodes, 0.5 mm apart. To determine the dependence of both quantities on the magnetic field strength B we used a sawtooth type field generated by chopping the magnet current at a frequency of

1.5 Hz. The resulting change of the specimen current was amplified and monitored with a storage oscilloscope. Photoconductivity measurements were made with illumina- tion by a tungsten-iodine lamp at an incident power of % 50 mw/cm2. In order to avoid changes of o during optical exposure (Staebler-Wronski effect7) the samples were illuminated

8 :

^% 12 hours prior to the measurements. Some checks indicated that this treatment did not significantly change the dependence of Aaph/uph on B and T.

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1981428

JOURNAL DE PHYSIQUE

Fig. 1: The relative change of the photo- conductivity, A u ~ ~ ! u ~ ~ , as a func;

tion of magnetic freld B for vari- ous temperatures. The doping ratio is a) [PH~]/[s~H~] = and b) [B~H~]/[s~H~] =

Results.- Fig. 1 compares the in- ) fluence of magnetic field B on the

photoconductivity oph in two a-Si:H films doped with nominally the same content of phosphorus or bo- ron. Plotted is the relative change of aph as a function of B for vari- ous temperatures T. Other properties

(e.g. thermopower and temperature dependence of aph) clearly indicate that the charge carriers are elec- trons in Fig. la and holes in Fig.

lb. The behaviour seen in Fig. la is very similar to that previously re- ported for undoped a-Si:H which usu- ally is also n-type. Both specimens in Fig. 1 show complicated, anoma- lous dependencies which are similar

,

at low temperatures ( T L 120 K in Fig. la) but markedly different above 150 K. In order to ease the discussion we regard the experimen- tal curves as being composed of a positive component el and a negative component e2 as indicated by the dashed curves in Fig. la (T = 100 K)

.

Obviously in Fig. la the component el decreases much stronger with in- creasing temperature than el, where-

10 as in Fig. lb el and e2 decrease at

ceeds a certain limit ( [ B ~ H ~ /[SiH4l=

Fig. 2: Same dependence as in Fig. 1 for the

lo-3

and [PH~]/[s~H~] = 10-2) also doping ratio a) [PH~]/[s~H

1

= the dark conductivity ad of our and b) [B~H~]/[s~H~] = a-Si:H films is affected by the mag-

netic field B. Two typical examples of the dependence of Aad/ad on B and on the temperature T are shown in Fig. 3. The dopant content is 3 % for both the phosphorus doped film (Fig. 3a) and the boron doped film (Fig. 3b). It is interesting to note that the curves in Fig. 3 are very similar to those in Fig. 2 but opposite in sign. The magnitude of the components el and e2 determined at ~ = 1 5 0 K is plotted in Fig. 4a as a function of the dopant con- tent. Both components seem to saturate at high doping levels and steeply decay to- wards low doping ratios.

a more similar rate.

The behaviour of n-type films

8-

is appreciably different from that in Fig. la if the pH3 content lies in the range 3.10-~ to This is demonstrated by the curves in Fig.

2a measured for a sample which was deposited at a gas doping ratio [PH~]/[s~H~] =

loa3.

Here the nega- tive component e2 is small compared to el, the latter being of similar magnitude as in Fig. la. In contrast to this strong change for n-type films the behaviour of p-type speci- mens does not significantly vary in the doping range investigated (cf.

Figs. lb and 2b).

6- v-

z

^4-

0"

0 2 4 6 8 1 0 0 2 L 6 8 1 0

WkG) B (kG) When the dopant content ex-

'1

t? 2 - Q

-2 -

a-Si: H

[ P H ~ ] / [ s ~ H ~ ] = ~ o - ~ '

0 - y

f

a

- & ? . , . , . I . , , , . , . , . , . , . ( B ~ H ~ ] / [ s I H ~ ] = I O "

Discussion.- Before explaining the

3 7 ---,

-

experimental results we briefly sum-

tures T. The doping ratio is a)

Ao ^- 2

[PH~]/[S~H~] = 3 -

lo-'

and b) [B~H~]/

-

phq,? ( I - b 2 ) ^H -

[siH4] = 3 .

lo-'

o d v2 2 2 (1)

ph H +V

2 . 1 ~

0 -

0

where HI is the internal anisotropy field due to dipolar and hyperfine interactions and H, is the external field B, both ex- pressed in frequency units (1 ~ 2 2 . 8 . 1 0 ~ Hz).

The quantity b is a measure of the g-tensor anisotropy which is assumed to be propor- tional to the g-shift 6 9 (b2 = 3/10 6g2), _v is the frequency of fluctuations of the ani- sotropy fields and d is the dissociation or thermal release rate. For simvlicitv we have

marize our previously published model

a-Si: H

[ P H ~ ] / [ s I H ~ ] = ~

for magnetic field dependent recom- bination' 1 6. We assume that, at least for part of the recombination

/ processes, the rate limiting step is a transition of an electron (or hole)

1 - I

^a-Si:H

⁽

used here the approximation v->> d >;w.

lo-,

1

^a ~ = l w K r f ~ of the Relation experimental (1) results: explains two i) Aoph/uph features can

d

be either positive or negative depending on

- a

whether H1/v>b or vice versa, and ii) it

saturates for Ho >v. Two other important

lo4 features of the experimental curves are the

simultaneous occurance of positive and ne- gative contributions (el,e2) and the anoma-

lo.' K).3 '04 la' lous linear or sublinear dependence on B at

[S,Hsl/lSiHd

[PHJ/[SIH~I

small magnetic fields. They become plausible

Fig. 4: Dependence of a)

1

A U ~ / U ~

1

and b) if one takes Into account that, in a-Si:H ud both measured at T = 150 K on films, the quantities HI, b,

v

and d will the doping ratios [B~H~]/[S~H~] have broad distributions, so that the ex- and [PH~]/[s~H~] perimental curves will be a complicated

superposition of positive and negative ef- fects of the type described by eq. (1). Therefore, eq. (1) can only be used for a qualitative discussion of the experimental data.

Q from a singly occupied localized

= _-

^-1- state (e.9. a tail state) to another

s

one which already carries a spin

g -2 - (e-g. a neutral dangling bond). If

a

-3 - the spin flip or relaxation rate W,

The above model of spin dependent transitions was first used to explain the magnetic field induced changes of od in evaporated a-Si and a4e5. In these mate- rials an appreciable contribution to od is due to hopping processes in the localized states near the Fermi level. If the final state of a hopping transition is doubly

-I -

-5

in both centers, is smaller than the recombination rate R, the latter depends on the spin configuration in

0 the pair of initial states. Of par-

ticular importance are those pairs

6 , . , . I . , . , . , , . . . , . , . . . having triplet character (parallel

O lo O spins). They can only recombine if

B(kG) B (kG) one of the two spins flips before the

pair dissociates. The dependence of W Fig. 3: The relative change of the dark con- on the applied field B leads to a

ductivity Aod/ud, as a function of relative change of the photoconduc- maqnetic field B for various tempera- tivity6

C4- 146 JOURNAL DE PHYSIQUE

occupied, the rate of this transition, rT, is affected in the same way as the recom- bination rate of triplet pairs,

%.

Since ad increases with increasing r~ whereas aph decreases with increasing %, the changes of ad and oph must be opposite. This is just what we observe for highly doped a-Si:H films (cf. Figs. 2 and 3).

In weakly doped n-type films the negative component e2 of Aaph/uph becomes stronger compared with the positive component el when the temperature Increases (Fig. la). This behaviour indicates that the distribution of the fluctuation fre- quency

v

changes with temperature such that the centroid shifts to larger values.

For p-type films, on the other hand, this shift seems to be less pronounced, since the ratio of e2 and el varies only little with temperature. The reason for this dif- ferent behaviour is not yet understood.

The most striking result of the present study is the strong decrease of the negative component e2 when the phosphorus content exceeds 'L 3 - 1 0 - ~ (cf. Figs. la and 2a). We attribute this effect to a change of the predominant recombination channel.

At the lower doping levels the recombination rate presumably is limited by a tran- sition from a localized tail state to a singly occupied dangling bond state. At the higher doping ratios

( 2

3-10-4) the F e m i level EF lies in the band tail above the dangling bond levels, so that the recombination rate is determined by a transition from a localized state in the upper part of the tail to another one near EF which is singly occupied. This latter assignment is supported by recent ESR data, which indi- cate that the g-tensor anisotropy of electrons in the conduction band tail, and therewith the quantity b in eq. (I), is smaller than that of the dangling bond elec- trons. As a consequence, the negative component e2 should be smaller if the final state of the transition is a doubly occupied tail state instead of a dangling bond state, in agreement with our experimental results (cf. Figs. la and 2a). For p-type films we do not observe a significant change in the dependence of Aa h / ~ p h on B with the dopant content, indicating that there 1s no marked change

05

the anisotropy of the g-tensor.

In undoped and weakly doped a-Si:H films EF lies in an energy range where the density of states g(E) is so sraall that the charge transport in these states does not yield a noticeable contribution to ad. This situation changes when, due to doping, the Fermi level EF is shifted into a region of larger g(E). The results in Fig. 4a indicate that, for boron doped films, hopping near E contributes signifi- cantly to ad (at T = 150 K ) if the dopant content exceeds

l o - ' .

In phosphorus doped films this occurs only at an appreciably higher doping ratio ( > This result reflects the large asymmetry in the effective density of gap states indicated by field effect data8. We believe that the magnetic field dependent processes are hops from singly occupied tail states to other singly occupied tail states. The striking similarity of Aad/od with -Auph/a in highly doped films (cf. Figs. 2 and 3) sug- gests that such inter-tail .transi~?ons also determine the recombination rate in these films.

MELL H., MOVAGHAR B., SCHWEITZER L., phys.stat.So1. (b)

3

(1978) 531 MELL H., MOVAGHAR B., J. Nan-Cryst. Solids 3 5 + 3 6 (1980) 761

MELL H., STUKE J., J. Non-Cryst. Solids

4

(1970) 304

MELL H., Proc. 5th Int. Conf. Amorphous and Liquid Semiconductors, Taylor and Francis, London 1974, p. 203

MOVAGHAR B., SCHWEITZER L., J. Phys. C

11

(1978) 125

MOVAGHAR B. , RIES B.

,

SCHWEITZER L., Phil. Mag. B

41

(1980) 141 STAEBLER D.L., WRONSKI C.R., J. Appl. PhyS.

2

(1980) 3262

MADAN A., LeCOMBER P.G., SPEAR W.E., J. N0n-Cryst. Solids

20

(1976) 239 DERSCH H., STUKE J., BEICHLER J., phyS.Stat.Sol. (b) (in press)