LIMITATIONS ON THE USEFULNESS OF METALLIC THIN FILM SEMICONDUCTORS FOR PHONON DETECTION

HAL Id: jpa-00221629

https://hal.archives-ouvertes.fr/jpa-00221629

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.

LIMITATIONS ON THE USEFULNESS OF

METALLIC THIN FILM SEMICONDUCTORS FOR PHONON DETECTION

S. Rogers, C. Shaw, H. Wiederick

To cite this version:

S. Rogers, C. Shaw, H. Wiederick. LIMITATIONS ON THE USEFULNESS OF METALLIC THIN

FILM SEMICONDUCTORS FOR PHONON DETECTION. Journal de Physique Colloques, 1981,

42 (C6), pp.C6-317-C6-319. �10.1051/jphyscol:1981692�. �jpa-00221629�

JOURNAL DE PHYSIQUE

CoZZoque 66, suppldment au n022, Tome 42, dScembre 1981

LIMITATIONS ON THE USEFULNESS OF METALLIC THIN FILM SEMICONDUCTORS FOR PHONON DETECTION

S.J. Rogers, C . J . Shaw and H.D. ~ i e d e r i c k *

Phys<es Laboratory, University of Kent a t Canterbury, Canterbury, Kent, England

* R. M. C., Kingston, Cmzada

Abstract.- Ultra-thin metallic films behave essentially as semiconductors.

W e ave explored the use of N i - C r devices in t h i s regime f o r phonon pulse det:ction. The evaporated films, which were typically 2. 20 B. thick, were deposited between layers of SiO, on a sapphire substrate. Suitable films had room temperature resistances per squarg i n the range 10 - 13 KQ and these values increased t o lo5 - lo6 sz ^{a t 1 K.} By using very short (2Su) conduction paths i n parallel, device resistances were reduced t o % 100 Q a t ~ O K . W e have tested devices down t o O.loK and two serious problems limit t h e i r usefulness: the low temperature resistances are quite non-ohmic for small currents; the thermal time constants a r e surprisingly long.

Metallic films a few atomic dimensions i n thickness behave essentially as semi- conductors('); t h e i r equivalent specific r e s i s t i v i t y is high and the resistance variation with temperature, T, can be approximately represented by R = l$, exp(Tofr), where To i s some characteristic temperature. The conduction mechanism is predomin- antly tunnelling between the island structures which a r e characteristic of such films. The aperiodic potential of these structures serves t o localize the electrons, and t h i s leads t o i n f i n i t e r e s i s t i v i t y a t OOK('). Systems showing such localization effects have been of considerable recent interest. The thin film r e s i s t i v i t y is i n general not much affected by magnetic f i e l d s , and t h i s suggests that i t night serve to detect phonon pulses i n experiments requiring such fields. Almost any non- superconducting metal could be used for a bolometer, but, following Griffin and

~ o c h e l ( ~ ) we have concentrated our work on 80/20 N i - C r alloy films which can be successfully deposited a t room temperature.

The films studied, which were typically s 20 8 i n average thickness, were deposited by evaporation (at a rate

4 Qsec) between layers of SiO, on a sapphire substrate, through which phonon pulses could be transmitted. The Si4, layers were deposited a t r a t e s i n the range 20 - 100 8/sec by the evaporation of SiO under a pressure < 10-5 t o r r . The f i r s t oxide layer on the substrate ^{( %} 1OOO 1) provided a chemically inert well characterised surface for the Ni-Cr, and the second protective oxide layer (*3000 1) wwas deposited on top of the Ni-Cr film withuut breaking the vacuum. The variation with temperature of the resistance of the N i - C r films depends c r i t i c a l l y upon t h e i r thickness, and semiconductor behaviour i s observed over a thickness range of only a few 1 (4). The room temperature r e s i s t i v i t y of our films provided a sensitive guide t o t h e i r low temperature properties; films with

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1981692

C6-3 18 JOURNAL DE PHYSIQUE

resistivities in the range 10 - ^{13 KQ} per square were most suitable for use as bolometers at 1°K.

The resistance values for suitable films changed little in cooling to 77OK, though in some cases there was an initial decrease of up to lo%, but the resistances typically increased to lo5 - ^{lo6 52} per square at 1°K. Such values are inconveniently large if the bolometer circuit is to have an electrical time constant short compared with the time scale of typical phonon pulses, and in practical devices an interdig- ital comb geometry was adopted in which conduction was across Ni-Cr strips % 25 p

wide. The device resistance values were lo2 - lo3

!;''I times less than for a square film. Somewhat sur- prisingly, the change in geometry markedly affected

sg the resistance characteristics of the films. In

narrow geometries the equivalent specific resisti-

2. 5 ¹

- vity per square was increased as compared with

I- V) -

V)

that for a square monitor film. Fig. 1 shows this

Id 102 1 o3

a CONDUCTION PATH LENGTH (u) effect for a series of simultaneously deposited films of differing conduction path length; the Fig. ^Effect of film geometry enchancement factor here is the ratio of the

on resistivity measured resistance to that calculated from geo- metry and the resistance of a square film. Reduc- ing the conduction path length also changed the film thichess needed for the desired resistance variation at 1 ' ~ ; it was therefore necessary to use thinner films for the interdigital devices.

Fig. 2 shows the resistance variation in the temperature range 0.1 - 10'~ for two such devices in which the conduction path length was 29 11.

The difference in the two characteristics is a measure of the difficulty experienced in making devices to a given specification. In each case, above 1°K the resistance variation is essentially exponential, but there is an unexpected plateau at lower temperatures. Although the measured resistances in this plateau region varied little with power dissipation, which was a 1 ~ ~ l l W for these measurements, it is possible that the apparent almost constant resistance arises because of non-ohmic effects. We note, however, that in

l l l l r l I 8 t I I I M

106

- -

TEMPERATURE

this temperature range the resistance of some thin films has been observed to show a logarithmic temp- Fig. 2 Resistance of twodevices erature and power dependence (5). The resistivity

(conduction path = "IJ) plateau severely limits the devices1 useful

105

c: -

-

3 0 4 z 2 9 z

LL

103

l o 2 . 0.1

'% -

-

-

- I ^{- - -}

- ₈ - - -

b -

/ ;

- - z 1

'\i

1

W-O-O-. .,

- 't -

- \ -

- *\ - -

' . ¹

-

1 f i l l l u ~

1

1 .o

_{I I}

0.8 - ^... 4 . 2 0 ~

-

-

-

0 40 80 120 160

VOLTAGE (mV)

temperature range for phonon detection. Even at higher temperatures it is difficult to make practi- cal use of the potential sensitivity because the film resistances are highly non-ohmic for quite small power inputs. For phonon pulse detection, the device must be fed with a biassing current, but typically a current of only 50 markedly reduced the sensitivity. This will be seen in Fig. 3 which shows the I-V characteristics of an interdigital device at three temperatures; the differential resistance varies little with temper- ature except for small biassing currents.

Fig. 3 I-V characteristics

for an interdigital - In developing these devices it was thought that device non-ohmic effects would arise when the voltage

between metallic islands was

kTo/e, and that with To values of a few degrees, if the film granularity was of the same scale as the film thickness (20 - 40 I), it would be possible to use biassing voltages s 100 mV.

This accords with the observations of Griffin and Mochel that some of their thermo- meters were essentially ohmic up to power levels a lo-'+ W. The allowable dissi- pation levels for the bias power in our devices were lo4 times lower than this.

The usefulness of these devices is further limited by their surprisingly long thermal time constants (s 20 The thermal origin of these time constants was confirmed in an experiment with a superconducting A% bolometer in series with a Ni-Cr device on the same sapphire substrate. The superconducting film responded rapidly (< 0.1 vs) to the phonon flux and gave an undifferentiated signal, but the Ni-Cr film gave an integrated signal, of the opposite polarity, from which the original phonon pulse profile could be obtained by electrical differentiation.

Although their sensitivity is poor, we have made some use of the Ni-Cr devices in experiments on supekconducting Aa, single crystals. It is not clear that they offer any real advantage over carbon films which also function as integrating dev- ices and are similarly insensitive to magnetic fields.

References

(1) Neugebauer, C.A. and Webb, M.B., J. Appl. Phys. 2, ^{74 (1962)}

(2) Adkins, C.J., Phil. Mag. 36, 1285 (1977)

(3) Griffin, E.E. and Mochel, J.M., Rev. Sci. Instrum. 5, 1265 (1974)

(4) Dynes, R.C., Garno, J.P. and Rowell, J.M., Phys. Rev. Lett. 40, 479 (1978)

(5) Dolan, G.J. and Osheroff, D.D., Phy. Rev. Lett. 43, 721 (1979). -