Electrical transport properties of the semiconducting layer compounds GaS and GaSe

Barrier energies and I-V curves of several contact metals on n-type GaS are studied. Thedependenee of harrier energy on the work function of the metal can be interpreled by means ofchemica! interaction with the GaS. INTRODUCTIONTHE INVESTIGATION of eJectrical transportproperties of the high ohmic (108-lOSfl cm)semiconductors GaS and GaSe is hamperedbecause of difficulties in controlling the con-tact properties. Contacts used in literatureon GaS and GaSe are: indium contacts onGaS and GaSe, checked.for ohmic behaviour[1]; evaporated Au on GaSe and GaTe[3],[4]; soldered Cu on GaSe[5] and alloyedln-Hg contacts on GaS[6]. Receritly, Kurtinand Measurface harriers on p-type GàSe by a photo-response method.We determined the barrier enèrgies of thecontacts on n-type GaS by measuring thephotoresponse iil a backwan technique. Inaddition, current-voltage (1-V) curves of thecontacts were obtained by means of voltageprobes (in darkness). All the measurementswere carried out at room temperature. Toavoid differences in bulk properties of thecrystals we took them all from one ampulla(the GaS platelets have been grown byvapour transport [8]).Together with the metal contacts to beinvestigated gold contacts were evaporatedwithout breaking the vacuum, thus havinga reference on each sample. *Present address: Philips GloeilampenfabriekenEindhoven. Netherlands.597We found that in our case on GaS as wellas in that of Kurtin and Mead on GaSe therelation between the harrier energies and themetal work functions shows a dependeneeon whether the heat of formation of the con-sidered metal-sulphur (selenium) compoundis greater or smaller than that of galliumsulphide (selenide). SAMPLE PREPARATIONWithout any surface treatments the cry-stals, thin platelets, were mount~d in a maskon a small furnace in a high vacuum plant. Theevaporation of the contact material occurredfrom Mo and W boats in a vacuum betterthan to-s Torr, argon atmosphere. Duringthe first moments of evaporation a shutterwas used to intercept surface. impurities fromthe metàl. Because of the better sticking onthe surface of the sample, its temperature wasraised to 300°C during evaporation. We alsoevaporated on cold (30°C) samples forpurposes of comparison.For the contact configuration see insetin Fig. I. EXPERIMENTSThe high resistances of the samplesneccessitated the use of electrometers forthe voltage and the current measurements(Keithley 61 OB and 417 respectively). Thelight souree was a quartz halogen lamp .598A. H. M. KIPPERMAN and H. F. VAN LEIDEN/ Fig. I. Curreni-voltage curves of several contacts on 11·type GaS: forward scales are expanded flve times.lnset: contiguration of the contacts on each sample.inc!uding gold as the reference. Subscripts: c. ·cold": andh. 'hot". combined with narrowband interferencefilter ( Balzers 820). I ntensity was calibratedwith a Si-photodiode (Siemens BPY I I) in ashort circuit arrangement.n-TypeGaSThe 1-V curves of the main contacts arepresented in Fig. I. The strong blockingcontacts as Au, Ag and Cu constantly showedftuctuations and also current creep duringseveral hours after the reverse voltage wasapplied.Figure I shows that the temperature duringthe evaporation of the contacts is veryimportant for their behaviour. Especially thereverse currents of the base metals Al. Mgand Sb appear to be. different for 'hot' and'cold' evaporated contacts. while the samecurves were measured for . Au contacts. Anearly ohmic 1-V curve is obtained in thecase of 'hot' evaporated Sb contacts.Barrier energies were measured by thephotoresponse method[9]. At a constantnumber of incident photons the curves ofthe square root of the photocurrents vs.pboton energy .are straight lines. The interceptof the hv-axis corresponds to the barrierenergy ~8 i.e. the energy dilference betweenthe Fermi level and the GaS conduction bandat the interface.Owing to the fact that the main contactswere of the same material. the photocurrentsopposed each other. Therefore a voltage wasapplied across the sample so that the reversecontact determined the. photocurrent. Byswitching the polarity the barrier energy ofthe other contact was achieved. The dark-eurrent consequent upon this voltage wascompensated. At the end of our experimentswe measured a Ag contact in combinationwith an ohmic contact (Sb) with and withoutthe voltage applied. In both cases we foundthè same barrier energy. therefore we concludethat barrier lowering due to image force is ofno importance.In Fig. 2 we have plotted the barrier energies . · metat work flJI'(:t;on!lm (eViFig. 2. Banier energies of contacts on 11-GaS and p-GaSevs. work functions of the contact metals: the work func-tion of tungsten is also indicated. Subscripts: c. ·cold':andh. 'hof. METAL SURFACE BARRIERS599 on n-GaS vs. the work functions of themetals cPm [I 01: together : with the harrierenergies on p-GaSe, measured by KurtinandMead. p-Type GaSPreliminary measurements of 'hot' contactson p-GaS indicated a behaviour opposite tothat ofthe n-type samples.Nearly linear 1-V curves showed Au,Ag and Cu contacts. The voltage drop acrosseach of the contacts is less than a few percent of the applied voltage. Sb contactsshowed a blocking non-linear 1-V curve. DISCUSSIONThe crystals had been grown as thin plate-lets, about I 0 p.m in thickness and had aperfectly clean surface. When they wereexposed to air only a very ~mail amount ofimpurities adhered because of the presenceof a Van der Waais bond on the surface.Apart from this, cleaving of the samplesprior to deposition is not recommended,because of the large number of surface statesthat will arise. This statement is supportedby the investigations of Turner and Roderiekon metal-silicon Schottky harriers [ 11] andthe electron microscope pictures of cleavedGaS and GaSe samples publisbed by Basinskiet a/.[12]Therefore, the surfaces of our samplesWere heither etched nor cleaved prior todeposition ofthe contact metal.Comparison of the curves of the gold re-ference contacts did not indicate the presenceof surface leakage.Some remarks should he made about the1-V curves in Fig. I. The forward curveshave been influenced by the ohmic voltagedrop of the bulk between m~ün contact andvoltage probe.As to the reverse direction we tried to findout if the current increase nüght be causedby tunnelling through the harrier. EspeciallySbh shows a nearly ohmic behaviour, incontrast with Sbc which was evaporated onthe samesurface ofthe sample without break-ing the vacuum. According to Zalm[l3] thelog I vs. v-112 curve will be linear for highvoltages if tunnelling is the main transportmechanism. So we plotted the reverse currentsfor several contact metals in Fig. 3 and cameto the condusion that tunnelling is likely. lH0-5 1}6 0-7 1}8____. V1f.l (\blts) Fig. 3.. Reverse currènt vs. voltage-112 curves of severalcpntacts on n-type GaS to show tunnelling according toZalm[ 13]. Subscripts: c, 'cold': and h, 'hot'. lt was impossible to determine the satura-tion owing to the rapid onset of the tunnelling.The main contacts on each sample provedto have the sameharrier energies; moreover,the harrier energies of all reference contacts(gold) were equal. This indicated that if thereare any differences in surface conditions, theydo not influence the harrier height and so wemay compare the results ohtained on severalsamples.Consirlering the curves of Fig. 2 there arethree remarkable points: there is a sharp kink in hoth the GaS andGaSecurve;the harrier energies of 'hot' contacts of the A. H.M. KIPPERMAN and H. F. VAN LEIDEN base metals on GaS are as good as fixed atthe Ga vahie;the slopes of the curves of p-GaSe andn-GaS are equal in the same intervals of tPm·We wiJl explain these points considering thechemica! interaction of the contact roetal andthe sulphur (selenium) atoms in the GaS(OaSe) Jattice. c.Comparing the heats of formation[l4] ofGaS (OaSe) and the sulphide (selenide)compounds ofthe contact metals we find threeregions of reactivity:I. Au. Ag. Cu, Pt and Pd wiJl not reactwith the S or Se because of the muchlower heats of formation of these com-pounds compared with GaS or OaSe.2. Sn and Sb may react; the heats of forma-tion have about the sarne values as GaSorGaSe.3. AI. In, Mg. Ca. Li and Cs wiJl react withthe S or Se and consequently· leavebebind a very thin layer of Ga at theinterface.·The thin Ga layer will change the effectivework frinction of the contact metal. When,in the 'hot' case, more Ga is formed by thereaction, the barrier· energy is equal to that ofthe Ga contacts.The crystal surfaces of GaS and OaSe areequal, viz. a closed packed layer S (or Se) anda Van der Waals bond to the next S (or Se).Moreover. the chemica! nature of S and Se isnearly the sarne; so equal slopes . may beexpected.The ·slope of right hand parts of the curvesin Fig. 2 is 0·8 both for GaS and for OaSe.In terms of uniform density of surface statesaccording to Cowley and Sze[l5] this means2x 1012 states (eV-1 cm-2), while Kurtin andMead have calculated 7x 1012• lt should bepointed out, however, that the latter plottedthe barrier energy vs. · the electronegativityofthe roetal (Pauling units).A remark should be made on the deviatingpoints in Pt and Pd in Fig. 2. Because Pt andPd are difficult to evaporate, we suppose thata smal! amount of the tunggten filament isevaporated too, thus lowering the workfunctions of the Pt and Pd contacts. Acknow/edgement-The authors wish to thank ProfessorDr. F. van der Maesen for bis constant interestand inanyhelpful discussious.· REFERENCESL BUBE R. H. and LINO E. L., Phys. Rev. 119, 1535(1960).2. BUBE R. H. and LINO E. L., Phys. Rev. ll5. 1159(1959).3. FISCHER G. and BREBNER J. L..J. Phys. Chem.Solids 23, 1363 (1962).4. LEUNG P.C.. ANDERMAN G. and SPITZERW. G .. J. Phys. Chem. Solids rl. 849 ( 1966).5. GUSEINOV G. D. and RASULOV A. J.. Phys.Status Solidi 18.911 (1966).6. KIPPERMAN A. H.M. and Van der LEEDENG. A., Solid St. Commun. 6, 657 (1968).7. KURTIN S. and Mead C. A .. J. Phys. Chem. Solids29, 1856(1968).8. LIETH R. M. A., Van der HEIJDEN C. W. M. andVan KESSEL J. W. M.. J. cryst. Growth 5. 251(1969).9. FOWLER R. H., Phys. Rev. 38. 45 ( 1931 ).10. MICHAELSON H. B .. J. appl. Phys. 21. 536(1950).11. TURNER M.J.and RODERICK E. H .. Solid. St.Electron. ll, 291 ( 1968).12. BASINSKI Z. S.• DOVE D. B. and MOOSER E ..H.P.A.,34, 373(1961).13. ZALM P.,Phi/. Res. Rep. U, 423 (1956).14. KUBASCHEWSKI 0., EVANS E. L. L. andALCOCK C. B.. Metallurgica/ Thermochemistry.Pergamon Press. Oxford ( 196 7).15. COWLEY A. M. and SZE S. M .• J. appl. Phys. 36,3212 (1965).