UDK 621,3:(53+54+621 +66)(05)(497.1 )=00 ISSN Strokovno društvo za mikroelektroniko elektronske sestavne dele in materiale Časopis za mikroelektroniko, elektronske sestavne dele in materiale Časopis za mikroelektroniku, elektronske sastavne dijelove i materijale Journal of Microelectronics, Electronic Components and Materials INFORMACIJE MIDEM, LETNIK 26, ŠT. 2(78), LJUBLJANA, junij 1996 ISKRA KONDENZATORJI SEMIČ 1951 - 1996 INFORMACIJE 1996 INFORMACIJE MIDEM LETNIK 26, ŠT. 2(78), LJUBLJANA, JUNIJ 1996 INFORMACIJE MIDEM GODINA 26, BR. 2(78), LJUBLJANA, JUN 1996 INFORMACIJE MIDEM VOLUME 26, NO, 2(78), LJUBLJANA, JUNE 1996 Izdaja trimesečno (marec, junij, september, december) Strokovno društvo za mikroelektroniko, elektronske sestavne dele in materiale. Izdaja tromjesečno (mart, jun, septembar, decembar) Stručno društvo za mikroelektroniku, elektronske sastavne dijetove i materiale. Published quarterly (march, june, september, december) by Society for Microelectronics, Electronic Components and Materials - MIDEM. Glavni in odgovorni urednik Glavni i odgovorni urednik Editor in Chief Tehnični urednik Tehnički urednik Executive Editor Uredniški odbor Redakcioni odbor Executive Editorial Board Časopisni svet Izdavački savet International Advisory Board Naslov uredništva Adresa redakcije Headquarters Mag. Iztok Šorli, dipl.ing., MIKROIKS d.o.o., Ljubljana Mag. Iztok Šorli, dipl. ing. Doc. dr. Rudi Babič, dipl.ing., Fakulteta za elektrotehniko, računalništvo in informatiko Maribor Dr.Rudi Ročak, dipl.ing., MIKROIKS d.o.o., Ljubljana mag.Milan Siokan, dipl.ing., MIDEM, Ljubljana Zlatko Bele, dipl.ing., MIKROIKS d.o.o., Ljubljana Dr. Wolfgang Pribyl, SIEMENS EZM, Villach, Austria mag. Meta Limpel, dipl.ing., MIDEM, Ljubljana Miloš Kogovšek, dipl.ing., Iskra INDOKd.o.o., Ljubljana Dr. Marija Kosec, dipl. ing., Inštitut Jožef Štefan, Ljubljana Prof. dr. Slavko Amon, dipl.ing., Fakulteta za elektrotehniko, Ljubljana, PREDSEDNIK - PRESIDENT Prof. dr. Cor Claeys, IMEC, Leuven Dr. Jean-Marie Haussonne, C.N.E.T. Centre LAB, Lannion Dr. Marko Hrovat, dipl.ing., Inštitut Jožef Stefan, Ljubljana Prof. dr. Zvonko Fazarinc, dipl.ing., CIS, Stanford University, Stanford, USA Prof.dr.Drago Kolar, dipl.ing., Inštitut Jožef Stefan, Ljubljana t RNDr. DrSc. Radomir Kužel, Charles University, Prague Dr. Giorgio Randone, ITALTEL S.l.T. spa, Milano Prof.dr. Stane Pejovnik, dipl.ing., Kemijski inštitut Boris Kidrič, Ljubljana Prof. dr. Giovanni Soncini, University of Trento, Trento Prof.dr. Janez Trontelj, dipl.ing., Fakulteta za elektrotehniko, Ljubljana Dr. Anton Zalar, dipl.ing., ITPO, Ljubljana Dr. Peter Weissglas, Swedish Institute of Microelectronics, Stockholm Uredništvo Informacije MIDEM Elektrotehniška zveza Slovenije Dunajska 10, 1000 Ljubljana, Slovenija (0)61 -316 886 Letna naročnina znaša 12.000,00 SIT, cena posamezne številke je 3000,00 SIT. Člani in sponzorji MIDEM prejemajo Informacije MIDEM brezplačno. Godišnja pretplata iznosi 12.000,00 SIT, cijena pojedinog broja je 3000,00 SIT. Članovi i sponzori MIDEM primaju Informacije MIDEM besplatno. Annual subscription rate is DEM 200, separate issue is DEM 50. MIDEM members and Society sponsors receive Informacije MIDEM for free. Znanstveni svet za tehnične vede I je podal pozitivno mnenje o časopisu kot znanstveno strokovni reviji za mikroelektroniko, elektronske sestavne dele in materiale. Izdajo revije sofinanci rajo Ministrstvo za znanost in tehnologijo in sponzorji društva. Scientific Council for Technical Sciences of Slovene Ministry of Science and Technology has recognized Informacije MIDEM as scientific Journal for microelectronics, electronic components and materials. Publishing of the Journal is financed by Slovene Ministry of Science and Technology and by Society sponsors. Znanstveno strokovne prispevke objavljene v Informacijah MIDEM zajemamo v: * domačo bazo podatkov ISKRA SAIDC-el, kakor tudi * v tujo bazo podatkov INSPEC Prispevke iz revije zajema ISI® v naslednje svoje produkte: Sci Search®, Research Alert® in Materials Science Citation Index™ Scientific and professional papers published in Informacije MIDEM are assessed into: * domestic data base ISKRA SAIDC-el and "foreign data base INSPEC The Journal Is indexed by ISI® for Sci Search®, Research Alert® and Material Science Citation Index™ Po mnenju Ministrstva za informiranje št.23/300-92 šteje glasilo informacije MIDEM med proizvode informativnega značaja, za katere se piačuje davek od prometa proizvodov po stopnji 5 %. Grafična priprava in tisk BIRO M, Ljubljana Grafička priprema i štampa Printed by Naklada 1000 izvodov Tiraž 1000 primjeraka Circulation 1000 issues UDK621,3:(53+54+621+66), ISSN0352-9045 Informacije MIDEM 26(1996)2,Ljubljana Uvodnik nove predsednice društva 78 New MIDEM President Editorial ZNANSTVENO STROKOVNI PRISPEVKI PROFESSIONAL SCIENTIFIC PAPERS A. Macher, K. Reichmann, 0. Fruhvvirth, K. Gatterer, G.W. Herzog: Primerjava NTC materialov s perovskitno strukturo z NTC materiali s spinelno strukturo za uporabo pri povišanih temperaturah 79 A. Macher, K. Reichmann, 0. Fruhwirth, K. Gatterer, G.W. Herzog: Perovskite Versus Spinel Type NTC Materials for Application at Elevated Temperatures J. Maček, M. Marinšek: Priprava disperzij nikelj/cirkonijev dioksid z gel-precipitacijo nikljovega hidroksida in hidratiziranega cirkonijevega oksida: vpliv reakcijskih pogojev na karakteristike 86 J. Maček, M. Marinšek: The Preparation of Nickel/Zirconia Dispersions from Nickel Hydroxide/Hydrous Zirconium Oxide Gel-Precipitate Precursors: Influence of the Reaction Conditions on the Characteristics B. Saje, B. Reinsch, S. Kobe-Beseničar, D. Kolar, I.R. Harris: Nitriranje Sm2Fei7 zlitine z dodatkom tantala 94 B. Saje, B. Reinsch, S. Kobe-Beseničar, D. Kolar, I.R. Harris: Nitrogenation of Sm2Fei7 Alloy with Ta Addition B.Cvikl: O nizkofrekvenčni C-U odvisnosti Ag/n-Si(111) Schottky-jevih diod, nanešenih po metodi curka ioniziranih skupkov, CIS 97 B.Cvikl: On Low Frequency C-U Relationship of the Ionized Cluster Beam, ICB, Deposited Ag/n-Si(111) Schottky Diodes K. Korošec, A. Vesenjak, B, Jarc, M. Šolar, R. Babič: Izvedba nerekurzivnega digitalnega sita s standardnimi integriranimi komponentami v modoficirani obliki porazdeljene aritmetike 107 K. Korošec, A. Vesenjak, B. Jarc, M. Solar, R. Babič: The FIR Digital Filter Realization with Standard Integrated Circuits in the Modified Distributed Arithmetic Structure UPORABA PLAZME V ELEKTRONIKI APPLICATION OF PLASMA IN ELECTRONICS I. Šorli, W. Petasch, B. Kegel, H. Schmid, G. Liebel, W. Ries: Procesi v plazmi. II. del: Uporaba v elektroniki 113 I. Šorli, W. Petasch, B. Kegel, H. Schmid, G. Liebel, W. Ries: Plasma Processes. Part II.: Applications in Electronics PREDSTAVLJAMO PODJETJE Z NASLOVNICE REPRESENT OF COMPANY FROM FRONT PAGE Iskra tovarna kondenzatorjev in opreme Semič, Slovenija 121 Iskra Capacitor Factory, Semič, Slovenia MIDEM IN NJEGOVI ČLANI MIDEM SOCIETY AND ITS MEMBERS Občni zbor društva MIDEM 122 General Assembly of MIDEM Society A. Zalar: Ustanovitev novega instituta 126 A. Zalar: Foundation of a new Institute KONFERENCE, POSVETOVANJA, SEMINARJI, POROČILA CONFERENCES, COLLOQUYUMS, SEMINARS, REPORTS D. Belavič: Delavnica ISHM/NATO 1996 127 D. Belavič: ISHM/NATO Workshop 1996 M. Kosec: Delavnica COST 514: Feroelektrične tanke plasti 128 M. Kosec: Workshop COST 514: Ferroelectric Thin Films S. Novak: Poročilo s simpozija 129 S. Novak: Report from Symposium D. Križaj: Poročilo s konference ISPSD'96 130 D. Križaj: ISPSD'96 Conference Report VESTI 131 NEWS KOLEDAR PRIREDITEV 135 CALENDAR OF EVENTS MIDEM prijavnica 137 MIDEM Registration Form Slika na naslovnici: Proizvodi Iskre Kondenzatorji Semič Frontpage: Products of Iskra Capacitor Factory Semič VSEBINA CONTENT 1 Spoštovane članice, spoštovani člani društva MIDEM Hvala za zaupanje, ker ste me izvolili za predsednico. Zavedam se, da obdržati to, kar društvo MIDEM je, ne bo lahko. Društvo izdaja revijo zavidljive kvalitete. Koliko pa je revij v Sloveniji, iz katerih ISI zajema članke za baze podatkov? Podobno velja za letno strokovno konferenco MIEL-SD. Letos je slaba polovica prijavljenih referatov iz tujine. Med avtorji so zelo ugledna imena. Visok strokovni nivo revije in konference je potrebno obdržati, verjamem pa, da ga je mogoče tudi zvišati. Računam na vas, na vašo strokovno moč in temu ustrezne prispevke. Dovolite mi, da k sodelovanju posebej povabim članice in člane iz gospodarstva. Poleg strokovnih člankov ste dobrodošli z novicami iz razvoja in proizvodnje, z vašimi uspehi in problemi. Ob koncu mi ne zamerite še dveh trivialnih vprašanj. Smo kot del tehnične inteligence zadovoljni s svojo vlogo v družbi? Nas je kdo, kdaj, kot skupino strokovnjakov vprašal za kakšno mnenje v zvezi z razvojnimi in drugimi gospodarskimi odločitvami? Če nas že ne sprašujejo, pa se sami kdaj oglasimo v javnosti. "MIDEM je trdno društvo" je dejal dosedanji predsednik dr. Rudi Ročak ob koncu svojega mandata. Ostanimo to še naprej. Rudiju pa hvala, ker je veliko pripomogel k tej trdnosti. Predsednica društva MIDEM dr. Marija Kosec, dipl. ing. UDK621,3:(53+54+621 +66), ISSN0352-9045 Informacije MIDEM 26(1996)2, str. 79-85 PEROVSKITE VERSUS SPINEL TYPE NTC MATERIALS FOR APPLICATION AT ELEVATED TEMPERATURES A. Macher, K. Reichmann, O. Fruhwirth, K. Gatterer*, G.W. Herzog Institut für chemische Technologie anorganischer Stoffe, ^Institut für Physikalische und Theoretische Chemie, Technische Universität Graz, Österreich Key words: NTC, Negative Temperature Coefficients, spinel type materials, perovskite type materials, material conductivity, polarons, Impendance spectroscopy, Impedance spectra, relaxation phenomena, polaron jump frequency, probability of polaron hopping Abstract: Commercial NTC's made of spinel type NiMn2C>4 materials cannot be used at elevated temperatures because of different reasons, e.g. too low resistivity and activation energy, poor thermal stability. Searching for proper materials we investigated Al and Ti doped LaCoC>3 with the composition LaMexCoi-x03 ranging from x = 0,1 to 0,7. In contrast to NiMnžCU the DC conductivity of the perovskite type material can be tuned by Ti doping within a wide range of applicability. The conduction mechanism was studied with impedance spectroscopy and leads to the conclusion, that both kinds of materials conduct via small polaron hopping. Relaxation phenomena are observed in the frequency range 1 to 1000 kHz and Interpreted with theories of Holstein, Sewell and Appel. To explain the difference of the DC response upon doping of spinel and perovskite type materials the conventional probability factor had to be modified. Perovskitni in spinelni NTC za uporabo pri visokih temperaturah Ključne besede: NTC koeficienti temperaturni negativni, materiali tipa spinel, materiali tipa perovskite, prevodnost materiala, polaroni, spektroskopija impedančna, spektri impedančni, pojav relaksacije, frekvenca preskoka polarona, verjetnost preskoka polarona Povzetek: Komercialni NTC upori na osnovi splnela NiMn2C>4 iz več razlogov niso uporabni pri visokih temperaturah. Imajo prenizko upornost, previsoko aktivacijsko energijo In so toplotno slabo obstojni. Z namenom dobiti primernejše materiale smo raziskovali prevodnost perovskitov na osnovi LaCoC>3 s sestavo LaMexCoi-xC>3, kjer je Me: TI, Al in 0.1 < x< 0.7. V nasprotju s spinelnimi materiali na osnovi NiMn2C>4 se da z dodatkom Ti prevodnost perovskitov na osnovi LaCoC>3 spreminjati v širokem območju. Mehanizem prevajanja, ki smo ga študirali z impedančno spektroskopijo, se da pri obeh tipih materialov pojasniti s takoimenovanlm "small polaron hopping". Relaksacijske pojave v frekvenčnem območju 1 do 1000 kHz smo razložili s Holstein, Sewell, Appel teorijo. Pri razlagi razlike v prevajanju dopiranih spinelov in perovskitov je bilo potrebno sprementi standardni faktor verjetnosti. INTRODUCTION NTC materials are usually based on inverse spinels of the type Nii-xMn2+x04. In air they are thermally stable only in the range x = - 0,2 to 1 above 700°C /1/. Therefore, conventionally sintered materials are always composed of quenched high temperature phases reflecting consequently the technology of processing. The temperature dependent inversion between cations on tetrahedral and octrahedral sites is said to be connected with a disproportion of Mn3+ ions, and leads to rather complicated cation distributions /2, 3, 4/. The inversion phenomenon is also a main reason for aging problems, which can be overcome by stabilizing bivalent ions on tetrahedral sites. For instance, a sufficient improvement can be obtained by substitution with Zn2+ ions/5, 6, 7/. The application of sintered materials demands a stable DC conductivity with respect to temperature changes, in many cases, at high enough temperatures the DC conductivity displays an exponential temperature dependence, which can be fitted well by a = a0exp kT (1) a specific conductivity ao preexponential factor Ea activation energy k Boltzmann's constant absolute temperature T However, at lower temperatures the DC conductivity deviates from this exponential behaviour in a rather intrinsic manner, i.e. different processing steps or even doping are not very effective. Up to now little attention has been paid to this low temperature deviations, although they are already observable in the undoped materials like NiMn2Û4 and CoMn204 /8, 9/. In general, manganite materials cannot be used for applications at elevated temperatures because of too low resistivity, too small activation energy and poor 79 Informacije MIDEM 26(1996)2, str. 79-85 A. Macher, K. Reichmann, O. Fruhwirth, K. Gatterer, G.W. Herzog: Perovskite Versus Spinel Type NTC Materials for Application at... thermal stability. Aiming at proper materials, we investigated the perovskite system LaMexCoi-x03 by substituting Al3+ and Ti4+ for Co3+ ions. Ti and Mg doping was first studied by Ramadass et al. /10/. They reported, that in contrast to the above mentioned manganites, the DC conductivity of the perovskite LaCo03 can be modified within a wide range. In this work we compare DC and AC conductivities of spinel and perovskite type materials and apply small polaron models fortheir interpretation. We anticipated, that both types of materials conduct via a similar process of small polaron hopping. EXPERIMENTAL Powders of different spinel and perovskite type materials were produced by coprecipitation of hydroxides from chloride solutions at about pH 11 and calcination at 900°C for 30 minutes. They were pressed into pellets and sintered between 1100 and 1200°C in air for several hours. Electrical contacts were made with a silver paste (E 4031 Demetron) fired at 750°C. Characterization of powders and pellets were made by XRD, REM and titration of the Co3+ content. Experimental details are described elsewhere /5,6/. DC measurements within the temperature range -30 to +400°C were carried out potentiostatically using platinum clamps. For AC measurements we used the HP impedance analyser 4192A, operative in the frequency range 5 Hz to 13 MHZ. DC CONDUCTIVITY conductivity, Ti and Al doping of LaCo03 leads to a dramatic decrease of conductivity at x >0,1. The effect is unexpectedly high (fig. 4) and demands for a reasonable explanation. Many conductivity curves deviate at low temperatures from the exponential dependence. 0,0015 0,0030 0,0025 m [K-1: 0,0030 0.0036 Fig. 2: DC conductivity of Al doped cobaltites Typical examples of DC conductivities for both kinds of materials are shown in figs.1 to 3. Whereas Zn doping of NiMn204 has rather little influence on the Zr^jNiMn,^ ZnNift6Mn1404 Ni0.fleMn232°4 & Fig. 1: —i—i—i—i—i—■—i—i—i—<—i—■-0.0015 0,0020 00025 0,0030 0,0035 0iCD40 0.00*5 1/T [K"1] DC conductivity of Zn doped manganites The simplest explanation for the observed exponential conductivity is based on the idea, that charges are hopping over an energy barrier (activation energy) located between neighbouring cations of different valency, similar to the activated motion of ions in solids (jump model, /11 f). A theory developed by Holstein /12/ describes the charge transfer as a phonon assisted hopping of (localized) electrons i.e. polarons, leading to deviations from pure exponential behaviour below the Debye temperature. For the diffusion coefficient of activated electrons in a linear chain Holstein obtains two solutions, one due to the diffusion of localized states at higher temperatures (D|0c with NTC character), and one due to the motion in a band at very low temperatures (Dband with PTC character): D,„c = a22rtv0| 2 n exp h2v„ Jy 2Ycsch(hv0 / 2kT) 'hvr -2 Y tan hi l 4kT (2) 80 A. Macher, K. Reichmann, O. Fruhwirth, K. Gatterer, G.W. Herzog: Perovskite Versus Spinel Type NTC Materials for Application at... Informacije MIDEM 26(1996)2, str. 79-85 Dband=a 2tw0 2Ycsch(hv0 /2kT) (3) exp -2Ycschf-^ UkT vo vibrational frequency Y = Eb/hvo a jump distance Eb polaron binding energy J overlap integral hvo vibrational energy LaTijCo^O. E ¿J -4 Fig. 4: Composition dependence of DC conductivity at 400 K ne2a2 kT c(1 - c)v0 exp kT (5) At this stage one can roughly estimate the DC conductivity of manganites and compare it with the experimental value. Use of the data vo = 1013 Hz, N = 2,7x1028 m"3, c = 0,5, a = 3A, Ea = 0,35 eV, e= electron charge and T = 300 K gives s = 4,9 x 10"4 Scm"1. The measured value is about 5,4x10"4 Scm"1. a.0015 Ofxm 0,CCQ5 m [K'1] 0,0036 Fig. 3: DC conductivity of Ti doped cobaltites In the high temperature case with hvo < < kT the exponent tanh(hvo/4kT) can be replaced by its argument. With n for the polaron density, EA=Eb/2 for the activation energy, J = hvo and the Nernst relation o = (ne2/kT)D one obtains for the DC conductivity ne2a2 kT Pv0exp -E. 4kT (4) The factor P contains all preexponential D|0c terms except a2vo. In the classical adiabatic limit of hopping P equals 1. Assuming that n is approximately given by c(1-c)N ion pairs of different valency (N density of cations, c fraction of donor ions, 1-c fraction of acceptor ions) one obtains aoot 1/t [k'1] Fig. 5a: Schematic representation of Holstein formulae (2) and (5) 81 Informacije MIDEM 26(1996)2, str. 79-85 A. Macher, K. Reichmann, O. Fruhwirth, K. Gatterer, G.W. Herzog: Perovskite Versus Spinel Type NTC Materials for Application at... regions within the ceramic material. Thus we tried to find an atomistic interpretation. Fig. 5b: Holstein formula (3) on the basis of manganite data As shown in fig. 5a the Holstein formulae (2) and (5) seem to be a good approach for explaining not only the intrinsic low- but also the exponential high-temperature regimes. IMPEDANCE SPECTRA To prove the overall applicability of the small polaron model, we searched for a possibility to measure independently the polaron jump frequency. From impedance spectra we found relaxation phenomena between 1 kHz and 1 MHz. As can be seen from the dispersion curves in figs. 6 and 7 the relaxation effects of both perovskite and spinel type materials are quite similar. In the region around 1 MHz an electrical dispersion is observed, which is caused by the common oxide properties. The relaxation effects can be simulated electrically by Randies' equivalent circuits, which are sometimes used in electrochemistry. The circuits consist of a series combination of a resistor (Ri) and a parallel condensor (Ci) - resistor (R2): g„ 1 / (R1 +R2) + (l/R1)(cot)2 t = - 1 + (cot) R, R, +R2 r2c1 (6) Although such equivalent circuits yield a good data fit for the complex properties, it is not possible to assign the circuit elements to conductive and capacitive -Spinel, Zn0JMg03Ni0eMn12O4 Perovskite, LaTi03Co07O3 lg v [Hz] Fig. 6: Comparison of impedance spectra of spinel and perovskite type material Fig. 7: Impedance spectra of Ti doped cobaltites Sewell /13/ calculated the complex conductivity of small polarons from first basic principles. In a Debye-like way Appel /14/ applied the two-site model for hopping charges (equivalent to a dipolar flip-flop process) and obtained a frequency dependence, which equals Sewell'sand Randies'formulae: 82 A. Macher, K. Reichmann, O. Fruhwirth, K. Gatterer, G.W. Herzog: Perovskite Versus Spinel Type NTC Materials for Application at... Informacije MIDEM 26(1996)2, str. 79-85 Re(cr) = G ne2a2 1 co x kT t 1 + co2t2 + On (7) Thus from the inflection point of Gp or Z dispersion curves (indicated by arrows in figs. 6 and 7) the relaxation time x = 1 /cd can be obtained and interpreted as the reciprocal mean jump frequency given by c(1-c)v. The temperature dependence of the DC conductivity and of the mean jump frequency are equal: the real part is given by independently diffusing charges. At frequencies higher than the mean jump frequency the dipoles cause an additive part to the conductivity by localized relaxation. The dispersion curves are broader by about 25% than predicted by the applied formulas, probably due to the strong coupling of hopping charges to the ionic environment, i.e. during the hopping process neighbouring ions are rearranging almost immediately. Observed and expected mean jump frequencies agree satisfactorily. 1 1 - = — exp t Tn kT (8) The independently measured data are compared in fig. 8. 1/to.ranges from 1012 to 1013 Hz and has the expected order of magnitude. 00B30 0.0032 a0034 Q0CB6 0.0038 0.0CM0 O.OX2 O.OOM 1/t [tc1] FIG. 8: Comparison of temperature dependence of DC conductivity and mean jump frequency of In doped manganite Besides the temperature shift of the impedance dispersion curves a similar shift with composition x is observed, that is also directly connected with DC conductivities (fig. 9). Both shifts give strong evidence, that relaxation sets in when the AC frequency v exceeds the mean jump frequency 1/x. From the experimental results of DC and AC conductivities and the applied dynamic models a phenomenologi-cal picture of hopping charges in NTC materials can be drawn. They conduct via polarons, that are able to diffuse through the lattice. At low enough frequencies Fig. 9: Composition dependence of mean jumping frequency compared with composition dependence of the conductivity o of LaTixCoi-xC>3 HOPPING IN SPINELS A more detailed interpretation taking into account the different atomistic structure of the materials with regard to application still remains to be discussed. Because of the inverse spinel structure of NiMn204 (degree of inversion 70-90% 121) in connection with disproportion, the octahedral cation chains contain Mn3+ in a matrix of Mn4+ ions, corresponding to an electron majority. Undoped NiMn204 shows a negative Seebeck coefficient /15/. However, the Seebeck coefficient does not depend exclusively on the excess charge density but also on the relative mobilities of the hopping charges. So for undoped stochiometric materials with equal positive and negative charge densities, the sign of the Seebeck coefficient is determined by the ratio of mobilities. This seems to be the case with undoped LaCo03 /16/. To understand the different conductivity response upon doping we start the discussion from a more or less 83 Informacije MIDEM 26(1996)2, str. 79-85 A. Macher, K. Reichmann, O. Fruhwirth, K. Gatterer, G.W. Herzog: Perovskite Versus Spinel Type NTC Materials for Application at... hypothetical intrinsic situation. In this situation the densities are given by the equilibrium of disproportion, the corresponding energies we estimate to be about 1 eV. Now, doping of spinels in the range 0,1 < x< 0,5 causes no appreciable change as observed e.g. in ZnNiMnC>4 or CuxNii-xMn204 or even in the non-stoichiometric compound NixMn2-x04/17, 18, 19/. The reason, why spinels doping in the range x = 0 to about 0,5 is not very effective seems to be simply given by the native high degree of polaronic disorder provoked by inversion. Observed conductivity changes range within one order of magnitude, the activation energies remain approximately constant. The weak density dependence of conductivity is usually approximated by the jump probability product c(1-c). HOPPING IN PEROVSKITES A quite different situation of polaronic disorder is found in perovskites. Doping with donors or acceptors below x = 0,1 causes a quite normal i.e. linear conductivity increase with x (see Th4+ and Sr2+ doping in LaCo03, /16/). The same holds for Ti4+ doping, but above the limit x = 0,1 there is an exponential-like decrease with increasing x. To explain the crucial exponential-like decrease (fig.4) we assume, that the probability factor c(1-c) is inadequate and has to be improved by statistical methods. There are two statistical approaches, a random liquid model leading to a probability given by cexp(-c) /20,21/ and a solid state model leading to a binominal distribution probability /22, 23/, which itself is the precursor of the Gaussian function exp(-c2); For small c values the first model would give on expansion the product c(1 -c). According to the binominal distribution, the probability of finding no or one acceptor ion in the second nearest shell around a donor ion or vice versa, is given in table 1. For high x values a pronounced decrease of donor-acceptor pairs (conducting polarons) is calculated. Both models supply probability functions with an initial increase and a following exponential-like decrease of conducting Co3+/Co2+ or blocking Ti4+/Co2+ and AI3+/Co3+ pairs. As no irregularities in the susceptibility behaviour are observed /10/, we exclude super-exchange hopping via oxygen ions to one of the 6 nearest neighbours. Hence a hopping process to one of the 12 second nearest neighbours over a distance of about 6 Â remains to be operative. This is in contrast to hopping in spinels over distances of about 3 Â to one of the 4 octahedral neighbours. CONCLUSIONS Comparing the electrical properties of doped manga-nites and cobaltites, one can state that small polaron hopping is operative in both kinds of materials. However, there are essential differences upon doping concentration, arising from the different structures. In the perovskite case there is no inversion provoking high native disorder, so that doping is much more effective than in the spinel case. Hence for a quantitative description of the hopping probability one needs a better statistical approach. Because of the exponential decrease of DC conductivity with increasing Ti doping, the higher activation energies and the much better thermal stability, perovskites could be used for NTC applications up to 500°C. AI doped cobaltites do not meet all of these requirements. LITERATURE /1 / Y.V. Golikov, S. Tubin, D. Bamburov, V. Balakirew, "High Tech Ceramics", ed.by P. Vincenzini, Elsevier Sei. Publ., Amsterdam (1987) 237 /2/ B. Boucher, R. Buhl, M. Perrin, Acta Cryst,, B25 (1969) 2326 /3/ E.G. Larson, R.J. Arnott, D.G. Wickham, J. Phys. Chem. Solids, 23 (1962)1771 /4/ V. Brabers, J. Terhell, Phys. Stat. Sol., 69 (1982) 325 /5/ O. Fruhwirth, A. Macher, K. Reichmann, H.G. Schuster, Third Euro-Ceramics Vol.2, ed. by P. Duran, J.F. Fernandez, Faenza, (1993) 395 /6/ A. Macher, Thesis, Technische Universität Graz (1994) /7/ A. Feltz, A. Seidel, Z. Anorg. Allg. Chem. 608 (1992) 166 /8/ K.V. Reichmann, G.W. Herzog, Ceramica acta, 5-6 (1992) 85 /9/ K.V. Reichmann, Thesis, Technische Universität Graz (1992) /10/N. Ramadass, J. Gopalakrishnan, M.V.C. Sastri, J. Less-Common Metals 65 (1979) 129 /11/1. Bunget, M. Popescu, Physics of Solid Dielectrics in Mat. Sei. Monographs, Elsevier, Amsterdam (1984) /12/ T. Holstein, Annals of Physics, 8 (1959) 343 /13/ G.L. Sewell, Phys. Rev., 129 2 (1963) 597 /14/ J. Appel, Solid State Phys., 21 (1968) 193 /15/ K.V. Reichmann, DiplomaThesis, Technische Universität Graz (1991) /16/ P. Gerthsen, K.H. Härdtl, Z. Naturf., 17a (1969) 514 /17/ J. Töpfer, A. Feltz, Solid State Ionics, 59 (1993) 249 /18/ B. Gillot, R. Legros, R. Metz, A. Rousset, Solid State Ionics, 51 (1992)7 /19/ S.T. Kshirsagar, J. Soc. Jap. 27 (1969) 1164 /20/1. Fritsch-Faules, LR. Faulkner, J. Electroanal. Chem., 263 (1989) 237 /21/ M. Inokuti, F. Hirayama, J.Chem.Phys. 43 (1965) 1978 Table 1: Composition of second shell x no acceptor [%] one acceptor [%] 0.01 88.64 10.74 0.05 54.04 34.13 0.10 28.24 37.66 0.20 6.87 20.62 0.30 1.38 7.12 0.40 0.22 1.74 0.50 0.02 0.29 84 A, Macher, K. Reichmann, O. Fruhwirth, K. Gatterer, G.W. Herzog: Perovskite Versus Spinel Type NTC Materials for Application at... /22/T. Luxbacher, H.P. Fritzer, K. Gatterer, C.D. Flint, J. Appl. Spectr., 62 (1995) 26 /23/ T. Luxbacher, H.P. Fritzer, R. Sabry-Grant, C.D. Flint, Chem. Phys. Lett., 241 (1995) 103 ACKNOWLEDGEMENT The authors are indebted to SIEMENS+MATSUSHITA COMPONENTS, Deutschlandsberg, especially to H.G. Schuster for the preparation of some materials, and the FORSCHUNGSFÖRDERUNGSFONDS DER GEWERBLICHEN WIRTSCHAFT, Austria, for stimulation and financial support of this work. Informacije MIDEM 26(1996)2, str. 79-85 A. Macher, K. Reichmann, O. Fruhwirth, G.W. Herzog Institut für chemische Technologie anorganischer Stoffe, Technische Universität Graz, Österreich K. Gatterer, Institut für Physikalische und Theoretische Chemie, Technische Universität Graz, Österreich Stremayrgasse 16/111 A-8010 Graz, Austria tel. +43 316 873 82 85 fax +43 316 837 619 Prispelo (Arrived): 10.5.1996 Sprejeto (Accepted): 18.6.1996 85 Informacije MIDEM 26(1996)2, Ljubljana UDK 621,3:(53 + 54+621 +66), ISSN0352-9045 THE PREPARATION OF NICKEL/ZIRCONIA DISPERSIONS FROM NICKEL HYDROXIDE/HYDROUS ZIRCONIUM OXIDE GEL-PRECIPITATE PRECURSORS: INFLUENCE OF THE REACTION CONDITIONS ON THE CHARACTERISTICS Jadran Maček and Marjan Marinšek Faculty of Chemistry and Chemical Technology University of Ljubljana, Ljubljana, Slovenia Key words: composite materials, microstructure properties, Zr Zlrconla gels, hydrolysis, solvent influences, crystallization, thermal analysis, dispersions of nickel, Zr Zirconium oxide, SOFC, Solid Oxide Fuel Cells, TPR, Temperature Programmed Reduction Abstract: Dispersions of nickel in a zirconia ceramic matrix were prepared by the gel-precipitation method from a methanol solution and subsequent thermal treatment (drying, calcination and TPR). Substituting methanol for water and using gaseous ammonia for initiation of gelation provides a reaction medium in which the system of hydrolysis reactions and above all condensation reactions can be controlled to a large degree. A study is made of the influence of the reaction conditions, temperature and final pH of the reaction mixture on the composition and characteristics of the composite materials. Well defined dispersions of nickel in zirconia matrix could be obtained in this way. Priprava disperzij nikelj/cirkonijev dioksid z gel-precipitacijo nikljevega hidroksida in hidratiziranega cirkonijevega oksida: vpliv reakcijskih pogojev na karakteristike Ključne besede: materiali sestavljeni, lastnosti mikrostrukturne, Zr geli cirkonijevi, hidroliza, vplivi topil, kristalizacija, analize termične, Ni disperzija niklja, Zr dioksid cirkonijev, SOFC celice ¡zgorevalne za okside trdne, TPR redukcija temperaturna programirana Povzetek: Disperzije niklja v keramični matrici cirkonijevega dioksida so bile pripravljene z uporabo gel-precipitacijske metode iz metanolnih raztopin in kasnejše termične obdelave (sušenje, kalcinacija in TPR). Zamenjava vode z metanolom kot reakcijskim medijem in uporaba plinastega amoniaka za sprožtev hidroliznih in kondenzacijskih reakcij zagotovi reakcijski medij, v katerem lahko v večji meri kontroliramo potek hidroliznih in predvsem kondenzacijskih reakcij. Namen prispevka je študij vpliva reakcijskih pogojev (temperature in končne pH vrednosti reakcijske mešanice) na karakteristike končnih kompozitnih materialov. Z uporabo gel-precipitacijske metode lahko pripravimo homogene disperzije niklja v matrici cirkonijevega dioksida. Introduction The reaction medium has a significant effect on the course of the gel-precipitation and the properties of the end product. Although accepted as a standard reaction medium, water restricts the reaction conditions to its physico-chemical properties. Solvation reactions can be modified substantially if water is replaced by other in particular organic media such as methanol. This alcohol has a lower dielectric constant and dipole moment than water (eMeOH25°C = 32.6, £H2025°C = 78.5, /JMeOH = 1.70, PH20 = 1.84), so that the influence of the electrostatic potential in methanol is consequently greater /1,2/. The dipole moment of the solvent or reaction medium determines the range of the influence of the individual sol particles on the neighboring particles. In this way the electrostatic double layer of the sol and so the sol coagulation process are affected. One of the major problems concerning the sol-gel processes is the reproduction and the reliability of the results obtained. Thus, the study of hydrolysis and condensation reactions can be very helpful to set-up the appropriate reaction conditions in preparing composite materials with well defined final microstructural and morphological properties. Zirconium oxide and its solid solutions are materials of current scientific and technological interest. They find application in various fields of materials science such as high technology ceramics and ionic conductors /3/. Recently it became clear that non-equilibrium and metastable phases, prepared by decomposing hydrous zirconium oxide at temperatures below 1000 K, might also be of great interest, for instance, to find new func- 86 J. Macek, M. Marinsek: The Preparation of Nickel/Zirconia Dispersions from Nicke! Hydroxide/Hydrous Zirconium Oxide Gel-Precipitate ... Informacije MIDEM 26(1996)2, str. 86-93 tional materials in heterogeneous catalysis /4/. A composite such as nickel - zirconia can be prepared in several ways. The most used and reported process for the preparation of nickel dispersion in zirconia matrix is the subsequent deposition of nickel on the already formed zirconia /5-7/. The degree of homogeneity of such composites can be enhanced by the simultaneous gel precipitation of both precursors needed for composite formation /8/. Suitability of such processes for the preparation of nickel dispersions in zirconia matrix was the objective of our work. We also investigated the effect of the Ni2+ ions on some solid-state properties of zirconia, such as crystallization and thermal behavior. These materials could be used for further preparation of solid oxide fuel cell (SOFC) anodes. Experimental By the gel-coprecipitation method zirconium and nickel were precipitated from the water or methanol solutions. The starting solutions of the metal chlorides were prepared by dissolving 38 g of NiCl2-6H20 tetrachloride (Kemika Zagreb, p.a.) and the corresponding amount of ZrCU (Fluka, assay >98%) in 400 ml of water or methanol, to which a twofold stoichiometric excess of water necessary for the reaction was added. A hydrolysis reaction was initiated by the introduction of gaseous ammonia (flow rate 3.88 lir1) through a glass frite into the solution of metal chlorides. This solution was vigorously agitated by a pitched-blade turbine 2000 revsmin"1). The introduction of ammonia initiates hydrolysis of the reaction mixture and precipitation of the hydrated zirconia and nickel hydroxide. The product was filtered and washed with distilled water until no reaction on chloride ions was observed (AgN03 test). In cases of a final pH of 7, the chloride ions cannot be washed out completely. The precipitate was dried for six hours at 120°C. The dried sample was milled in a ball mill and calcined for two hours at 500°C in air flow (18 llr1). After calcination, temperature programmed reduction (TPR) in a dynamic atmosphere of 4 vol% hydrogen and 96 vol% argon with a flow rate of 18 lh-1 was used for the reduction of nickel. A tube furnace, a heating rate of 5 Kmin-1, a final temperature of 500°C and thermostating for two hours at this temperature were used. The samples were cooled down in the same atmosphere. The amount of nickel in the samples was determined by the volumetric method and by atomic absorption spectroscopy using a Perkin Elmer Zeeman 5100 apparatus. The particle size distribution of the precipitates was determined by laser beam diffraction on a Fritsch Ana-lysette 22 apparatus. Auger Electron Spectroscopy (AES) depth profiling was performed on a PHI SAM 545 A analyser using electron static beam of primary electrons (3 keV, 0.5 mA and 40 (im in diameter). The sample was etched with Ar+ ions at an incident angle of 47° and etching rate of 2 nmmin-1. The thermal properties of the samples and the temperature of the crystal structure transformations were determined using a Netsch 409 STA thermoanalyser. Scanning electron microscopy (Jeol T-300 microscope) and specific sur- face area determination by the BET method and Strohlein area meter were used for further characterization of the samples. The thermally treated samples were also characterized by X-ray powder diffraction using a Philips PW-1710 instrument (30 mA, 40 kV in Cu-Ka radiation) and a Guinier de Wolff camera. The PDF CD-ROM database, sets 1-42, was used for the identification of samples. Results and Discussion Nickel dispersions in a ceramic matrix can be prepared either by separate formation of nickel and ceramic powders and their subsequent homogenization, by impregnation of zirconia powder by nickel, or by coprecipita- tion of nickel and ceramic precursors followed by appropriate thermal treatment and reduction. Our research was focused on the formation of nickel dispersions in a zirconia matrix by gel-coprecipitation. This method is attractive inasmuch as it reduces the number of operations required and yields a product with a high degree of homogeneity. A nonaqueous solvent, namely methanol, was used for the experiments besides water, in order to control the precipitation and gel formation reactions better. A major difficulty in coprecipitation reactions of binary cation systems is the difference in precipitation rates of precipitation of both cations especially when they differ appreciably, as is the case of nickel and zirconium cations. In addition to the discrepancy in hydrolysis rates, the polymerization reactions of the two species also differ. The course of coprecipitation of nickel hydroxide and hydrated zirconium oxide was followed by atomic absorption spectroscopy (AAS). The results of progress of hydrolysis reaction followed by AAS are summarised in Table 1. Precipitation of nickel hydroxide slightly lags behind that of hydrated zirconium oxide. The pH of the suspension rises very rapidly within the range of pH values at which precipitation of both gels is most intense, so that the two gel-precipitates overlap in their formation. Although, by terminating the flow of ammonia into the reaction mixture the reactions did not stop completely but nevertheless the results of AAS measurements show that precipitation of zirconia overtakes the precipitation of nickel hydroxide. The concentrations of the two cations change very little up to pH value 3. Between pH 3 and 4, condensation reactions take place and formation of the gel-precipitate of hydrated zirconium oxide occurs, /9,11,13,14,17/ while polymerization should be complete at pH 9 /16/. Nickel precipitates at pH>7 as Ni(OH)2. With an excess of the precipitating reagent a soluble hexaamino complex of nickel is formed([Ni(NH3)6]Cl2), which lowers the nickel content in the product /12/. In all cases the highest nickel content in the ceramic matrix was obtained at pH 8. At higher pH values the results obtained in methanol differ greatly from those obtained in water. The solubility of the nickel hexaamino complex is appreciably lower 87 Informacije MIDEM 26(1996)2, str. 86-93 J. Macek, M. Marinsek: The Preparation of Nickel/Zirconia Dispersions from Nicke! Hydroxide/Hydrous Zirconium Oxide Gel-Precipitate ... in this reaction medium than in water, so that at higher pH values the nickel concentration in the reaction medium decreases further and the nickel content of the product correspondingly increases. Washing the product in water can redissolve part of the nickel thus precipitated. It was expected from the start of experimentation that completely homogenous products could not be obtained by coprecipitation processes. The precipitation of hydrated zirconia begins at lower pH values and furthermore takes place at different rates as the precipitation of nickel hydroxide. In these precipitation experiments hydrated zirconium oxide and nickel hydroxide are obtained. Because the zirconium precipitation begins earlier, the nuclei and also the centers of the precipitated particles should be composed to a larger extend of hydrated zirconium oxide and the outer layers of nickel hydroxide. Between these two extremes a zone of intermediate composition should exist. However, the Inhomogeneity of dominance of zirconium or nickel phase are limited to the microscopic level as it was confirmed by the Auger and SEM analysis (Figures 1 and 2). Table 1: Results of precipitation reactions as followed by MS 60.00 Sample PH Zr4+ [gl-1] Ni2+ [gl-1] A1 1.0 12.50 17.41 A2 4.0 10.50 17.21 A3 5.0 <100 ppm 17.20 A4 6.0 <100 ppm 16.09 A5 8.0 <100 ppm 11.55 A6 10.0 <100 ppm 13.94 B1 1.0 12.50 17.63 B2 3.0 12.25 17.61 B3 4.0 <100 ppm 17.58 B4 6.0 <100 ppm 15.26 B5 7.0 <100 ppm 8.07 B6 8.0 <100 ppm 0.45 B7 10.0 <100 ppm <100 ppm r A samples were prepared from aqueous solution ' B samples were prepared from methanol solution with an addition of twofold stoichiometric amount of water needed for the hydrolysis reaction The etching of the pressed tablets of the coprecipitated product during the Auger analysis reveals rather uniform composition in both cations although the etching depth was 120 nm that is deeper than the average particle diameter. The deviation from the average con- -Zr ; -Ni -0 20 30 40 Sputter time (min) Fig. 1: AES composition - depth profile of C10ZrN83.ie sample (36.68 wt% of Ni and 63.32 wt% of zirconia) centration is for zirconium and nickel at most 2,36 at. %. The reason for mentioned results could be, that in gel precipitation reactions we should not adhere to the standard theories of precipitate formation. The gel of hydrated zirconia is a very voluminous 3D network with ample space inside the structure for the remaining reaction mixture. Nickel hydroxide precipitates inside this 3D framework as well as on the already formed surface area of the hydrated zirconia giving spatially very uniform dispersions. Upon syneresis and drying of the gel this fine dispersion of nickel is preserved. SEM micrographs of the C10ZrN8s.44 sample with 40.18 vol% nickel in ZrÛ2 matrix after heat treatment, Fig. 2a: SEM surface of C10ZrN8s 44 sample (40.18% of Ni) 88 J. Macek, M. Marinsek: The Preparation of Nickel/Zirconia Dispersions from Nicke! Hydroxide/Hydrous Zirconium Oxide Gel-Precipitate ... Informacije MIDEM 26(1996)2, str. 86-93 reduction and sintering show the homogeneity or heterogeneity of nickel dispersion in Ni-ZrC>2 composite materials at the microscopic level. Figure 2 shows a fracture in the C10ZrN8s.44 sample tablet (Fig. 2a) and a fracture in the etched sample (Fig. 2b). Both images show the nickel dispersion in the ceramic matrix to be relatively homogeneous, although there are microscopic areas where one or the other phase is dominant. Fig. 2b: Surface of C10ZrN8sA4 sample after etching with HCI The particle size and distribution affect the possible application of these cermets since they influence also the specific surface, porosity, density, etc. of the fired products. The characteristics of the intermediate gel-precipitate can be changed to a degree by the reaction conditions and therefore the influence of reaction conditions, i.e. pH and temperature and the mixing rate, on the size of the precipitates was studied (Table 2). The mixing rate was throughout all experiments kept constant in order to minimise its influence on the particle size by breaking the gel into smaller particles. Higher pH values of the reaction mixture and higher temperatures lead to precipitates of lower mean particle size. The pH of the reaction medium can influence the reaction and the product through changes in reaction kinetics and by changing the charge of the particles in the solution. Up to pH 7 the reactions proceed in a similar way in all cases. Hydrolysis that is highly sensitive to the pH of the reaction medium is assumed to be completed before this pH value is attained. The stages subsequent to the hydrolysis reaction, i.e. gelation, aging and possibly agglomeration of the precipitate, influence particle size even more strongly. Being faster in alkaline media, the aging process in particular favors smaller agglomerate sizes that agrees with the results. Like the mean particle size, the standard deviation falls as temperature and pH increase, leading to a narrower distribution of particle sizes. Table 2: Nickel content, mean agglomerate size and standard deviation, as a function of temperature and pH of the reaction mixture Sample Gel-precipitation Composite Temperature (°C) Final pH Ni content (wt%) d (um) ct (|im) A10ZrN7 10 7 9.71 27.99 14.68 A10ZrN8 10 8 27.57 23.71 12.80 A10ZrN9 10 9 24.94 20.44 12.45 A10ZrN10 10 10 18.58 17.10 9.43 A20ZrN7 20 7 9.51 22.98 12.03 A20ZrN8 20 8 26.84 18.10 9.59 A20ZrN9 20 9 24.22 16.25 8.42 A20ZrN10 20 10 18.30 13.16 7.64 A30ZrN7 30 7 9.09 20.66 10.30 A30ZrN8 30 8 23.38 16.48 8.92 A30ZrN9 30 9 20.67 11.99 7.47 A30ZrN10 30 10 16.93 9.21 6.58 B10ZrN7 10 7 24.34 23.38 10.56 B10ZrN7.7 10 7.5 25.29 19.52 9.71 B10ZrN8 10 8 31.52 17.74 8.54 B10ZrN8.5 10 8.5 30.63 16.71 9.32 B10ZrN9 10 9 29.55 16.11 9.13 B10ZrN10 10 10 28.00 12.16 6.87 * A series samples were prepared from aqueous solution * B series samples were prepared from methanol solution with an addition of twofold stoichiometric amount of water needed for the hydrolysis reaction The mean agglomerate size of the gel-precipitation obtained in methanol is greater than in an aqueous medium under the same conditions. This accords with the mentioned interaction of the charged sol species in the solvent with a lower dielectric constant and smaller dipole moment. Since the gelation-precipitation system is bimodal, not only the major constituent, i.e. the zirconia precursors, but also the presence of nickel hydroxide in the product has a large influence on the 89 Informacije MIDEM 26(1996)2, str. 86-93 J. Macek, M. Marinsek: The Preparation of Nickel/Zirconia Dispersions from Nicke! Hydroxide/Hydrous Zirconium Oxide Gel-Precipitate ... course of the reaction (table 3). A higher initial Ni2+/Zr4+ ratio influences the amount of nickel in the precipitate-as well as the mean particle size, which diminishes with higher initial ratio of metals. Table 3: Nickel content in the ceramic matrix, mean particle size of the precipitate, standard deviation and specific surface area as a function of the initial Nf+/Zr4+ molar ratio Sample Initial Composite Ni2/ Zr4+ ratio Ni (wt%) d (urn) o (nm) Spec. I surf. I area (m g ) C10ZrN8o 0 0 19.69 8.79 71 B10ZrN8 2.04 3.16 31.52 36.68 17.74 8.54 55 51 C10ZrN83.i6 17.20 8.40 C10ZrN84.94 4.94 8.44 41.82 50.76 15.71 9.32 48 C10ZrN8s.44 12.23 5.07 47 C10ZrN827.22 27.22 75.98 10.26 7.11 33 C10ZrN8~ J oo 99.13 i 3.48 1.46 19 all samples were prepared from methanol solutions at 1CPC with an addition of twofold stehiometric amount of water, final pH value of 8 and calcination in air at 500°C for two hours followed by TPR to 500°C The relation between various parameters of the system under observation is given in Figures 3 and 4 according to mathematical correlation of the results. Correlation equations were derived for the samples obtained from methanol solutions at a constant pH value of 8 or at a constant initial Ni2+/Zr4+ molar ratio of 2.04. Fig. 3: Average particle size profile of precipitated composites as a function of final pH In the reaction mixture and initial molar ratio Ni2+/Zr4+ Fig. 4: Chemical composition of finally prepared composites as a function of final pH in the reaction mixture and initial molar ratio Ni2+/Zr4+ 0 Fig. 5: 10 20 30 40 20 50 60 70 X-ray diffractions of samples; a) dry C10ZrN8sA4, b) thernally treated C10ZrN8sA4 and c) thernally treated C10ZrN8o 90 J. Macek, M. Marinsek: The Preparation of Nickel/Zirconia Dispersions from Nicke! Hydroxide/Hydrous Zirconium Oxide Gel-Precipitate ... Informacije MIDEM 26(1996)2, str. 86-93 The structure of the components was determined from the X-ray powder diffraction data (Figure 5). The dried samples appeared to be amorphous to X-ray. Their diffraction patterns exhibited only a broad band, around 2© = 30.5° characteristic for amorphous ZrC>2 and the reflections characteristic for NiO. After thermal treatment (calcination and TPR at 500°C) the undoped ZrC>2 (C10ZrN8o sample) reveals both, the formation of monoclinic and tetragonal modifications. The formation of a thermodynamically metastable tetragonal structure has been reported in the calcination of amorphous hydrated zirconium gels /10,18-20/. Dollimore /21/ has suggested a scheme for the transformation from the amorphous to the crystalline phase over an intermediate metastable tetragonal phase which at increasing temperatures slowly changes into a stable monoclinic modification. The formation of the intermediate structure is a consequence of the stabilization of the metastable structure in the first phase of crystallization by lower surface energy /15,16,23,24/. X-ray difractograms of the C10ZrN88.44 sample indicate the course of crystallization of the gel-precipitate of hydrated zirconium oxide and nickel hydroxide. Gel dried at 120°C does not display the orderly structure of Zr02- The position of the diffraction lines corresponds to nickel hydroxide, but the high background and broad peaks indicate that this hydroxide is not well crystallized. After calcination and TPR the structure of the sample is rearranged. The addition of nickel strongly affected the relative amount of monoclinic and tetragonal zirconia, causing a sharp increase in the tetragonal fraction and a decrease in the monoclinic fraction. The most intense monoclinic line (d=3.16 A, 20=28.24°) is no longer visible. According to M. Valigi et.all /5/ the increasement of methastable tetragonal fraction is the consequence of an interaction of a limited fraction of nickel with the hydrous zirconia surface during the experimental set-up. The diffraction lines at 2.034 and 1.762 respectively correspond to the face centered metal nickel (FCC) formed by TPR. However, more detailed descriptions of the crystalline phase formation during thermal treatment of mixed nickel-zirconium gels prepared by the gel-precipitation method can be found elsewhere /22/. The processes occurring during thermal treatment of the samples were monitored by TG/DTA analysis. The TG/DTA curves of the pure Ni(OH)2 (C10ZrN8<„), C10ZrN88.44 sample (50 wt% Ni and 50 wt% Zr02 in the end product) and the pure zirconium gel (C10ZrN8o sample) are shown in Figure 6. The DTA curve of the C10ZrN8oo sample (pure nickel hydroxide precipitate) displays two endothermic peaks at temperatures of 120.1 °C and 279°C respectively. The first endothermic peak, which is accompanied by a considerable mass loss, is the result of the removal of physically bonded water, while the endothermic peak at 279°C results from the dehydration of the nickel (II) hydroxide into amorphous nickel (II) oxide. Similar results have been reported by Godali and Livingston /25/. According to the literature the NiO structure develops gradually in the temperature region between 250°C and 815°C /25,26/, which could be assigned to the unpronounced broad exothermic peak that follows dehydration of the nickel hydroxide. In the DTA experiments for nickel free hydrous zirconium oxide (C10ZrN80 sample) two thermal effects were detected: i) a broad endothermic effect in the range approximately 50-300°C (peak temperature 147°C), due to to the evaporation of methanol and water from the surface of gels, and the elimination of physically bonded water trapped in pores of partially dried gel, and ii) a very sharp exothermic peak at approximately 444°C due to the transition from an amorphous to the crystalline structure (the so-called "glow phenomenon" /19/). More detailed dehydration studies of zirconia gels prepared by the gel-precipitation method and the influence of water partial pressure in the drying atmosphere can be found elsewhere /27/. The TG/DTA analysis of the zirconium and nickel gel-precipitate mixture (the C10ZrN8s.44 sample) shows that the TG and DTA curves are not simple combinations of the hydrated zirconium oxide and nickel hydroxide curves. The drying process proceeds over a broad temperature interval, as in the case of the C10ZrN8„o sample. Two endothermic effects, attributed to the elimination of physically bonded water and the dehydration of Ni(OH)2, and a broad unpronounced exothermic effect of zirconia crystallization and NiO structure formation were detected. Likewise, the endothermic peak of the dehydration of the nickel $ o -a 0 200 400 600 800 Temperature (°C) ig. 6: TG (—), DTA (—) analysis of samples: a) C10ZrN8Q44, b) C10ZrN8oo and c) C10ZrN8o 91 Informacije MIDEM 26(1996)2, str. 86-93 J. Macek, M. Marinsek: The Preparation of Nickel/Zirconia Dispersions from Nicke! Hydroxide/Hydrous Zirconium Oxide Gel-Precipitate ... hydroxide occurs at 300°C rather than at 279°C there is no clear exothermic peak of the zirconium oxide crystalline arrangement at 444°C, which may be attributable to the masking of this process by that of the formation of nickel oxide in the same temperature region (Figure 6). The exothermic peak was affected by the presence of nickel and was shifted to higher temperature and broadened as the nickel content increased. Similar results were reported by M. Valigi et. all /5/ who attributed such a behavior of the drying gel to the nickel interaction with hydrous zirconia matrix. He suggested, that because of the nickel interaction a fraction of the zirconia may not be able to crystallize or its crystallization rate is too slow that the process escapes detection. Conclusions Processes for preparation of nickel dispersions in zirconia matrix by the gel coprecipitation method were studied in nonaqueous media. The use of methanol as a solvent facilitated a controlled and stepwise hydrolysis of zirconium tetrachloride precursor that cannot be otherwise obtained in aqueous medium. The improvement to the so far described processes is the substitution of aqueous ammonia solution, used for the pH correction, for the gaseous ammonia. Thus the water needed for the hydrolysis and ammonia are introduced separately that enables a wider choice of experimental conditions and due to the lower volume concentration of gaseous ammonia and its better dispersion in the reaction mixture also less pronounced local supersatu-rations. Although the stability of zirconium and nickel cations in the methanol solution toward pH of the medium and the rates of depletion of the reaction mixture on these cations differ the products show rather high degree of homogeneity. In order to solve the problems concerning the reliability of the results obtained the relation between various parameters of the system was proposed. The addition of nickel to the composite strongly affect the course of crystallization of the composite. Higher imputes of nickel cause a sharp increase in the tetragonal fraction and a decrease in the monoclinic fraction. The crystallization of the composites is attributed by the exothermic effect detectable by DTA analysis. The exothermic peak of crystallization was shifted to higher temperature and broadened as the nickel content increased. Literature /1/ R.D. Nelson, Handbook of Powder Technology, Vol. 7, Dispersing Powders in Liquids, Ed. J.C. Williams, T. Allen, Elsevier Science Publishers, Amsterdam, 1988, pp. 137-150. /2/ C.J. Brinker, G.W. Sherer, Sol-Gel Processing, Academic Press, Inc., San Diego, 1990, pp. 36-37 and 127-130. 131 C.H. Steele, in High Conductivity Solid Ionic Conductors, Recent Trends and Application, Ed. T. Takahashi, World Scientific, Singapore, 1989, p. 402. /4/ A. Cimino, D. Cordischi, S. De Rossi, G. Ferraris, D. Gazzoli, V. Indovina, G. Minelli, M. Occhiuzzi and M. Valigi, Studies on Chromia/Zirconia Catalysts I. Preparation and Characterization of the System, J. Catal., 127, 1991, p. 774. /5/ M. Valigi, D. Gazzoli, R. Dragone, M. Gherardi and G. Minelli, Nickel Oxide-Zirconium Oxide: Ni2+ Incorporation and its Influence on the Phase Transition and Sintering of Zirconia, J. Mater. Sei., 5 (1), 1995, pp. 183-89. /6/ J. Grosßmann, K. Rose and D. Sporn, Processing and Physical Properties of Sol-Gel Derived Nanostructured Ni-Zr02 Cermets, in Proc. of the 4th International Conference on Electronic Ceramics & Applications, Electroceramics IV, (R. Waser, S. Hoffmann, D. Bonnenberg, Ch. Hoffmann), Augustinus Buchhandlung, Aachen, Germany, 1994, pp. 1319-22. /7/ Z. Ogumi, T. loroi, Y. Uchimoto, Z. Takehara, T. Ogawaand K. Toyama, Novel Method for Preparing Nickel/YSZ Cermet by a Vapour-Phase Process, J. Am. Ceram. Soc., 78 (3), 1995, pp. 593-98. /81 P. Cousin and R.A. Ross, Preparation of Mixed Oxides: a Review, Materials Science and Engineering, A130, 1990, pp. 119-125. /9/J. Livage, M. Henry, C. Sanchez, Sol-Gel Processing of Transition Metal Oxides, Prog. Solid. St. Chem., 18,1988, pp. 259-286. /10/ H.Th. Rijnten, Physical and Chemical Aspects of Adsorbents and Catalysts, Ed. B.G. Linsen, Academic Press, London, 1970, pp. 315-372. /11/ F.G.R. Gimblett, Inorganic Polymer Chemistry, Butterworths & Co., London, 1963, pp. 106-108, 291. /12/D. Nicholls, Pergamon Texts in Inorganic Chemistry, The Chemistry of Iron, Cobalt and Nickel, 24, Ed. I.C. Bailar Jr., Pergamon Press, Oxford, 1975, pp. 1109-1161. /13/W.L. Jolly, The Synthesis and Characterisation of Inorganic Compounds, Prentice-Hall, Inc., London, 1970, p. 69. /14/ L.M. Zaitsev, and G.S. Bochkarev, Formation of O Bridges in Zr Compounds, Russ. J. Inorg. Chem. (English Trans.), 7, 1962, p. 409. /15/ F.A. Cotton and G. Wilkinson, Advanced Inorganic Chemistry, 5th Ed., John Willey & Sons, New York, 1988, pp. 744,779-780. /16/A.F. Wells, Structural Inorganic Chemistry, Fourth Edition, Clarendon Press, Oxford, 1975, pp. 448-9. /17/A.N. Ermakov, I.N. Marov and V.K. Balyaeva, Preparation of Aqueous Solutions of Zr Oxychloride, Russ. J. Inorg. Chem. (English Transl.), 8,1963, p. 845. /18/J. Livage, K. Doi and C. Mazieres, Nature and Thermal Evolution of Amorphous Hydrated Zirconium Oxide, J. Am. Ceram. Soc., 51 (6), 1968, pp. 349-353. /19/ A.A. Rahman, Applications of Thermal Analysis in Surface Chemical Investigations of Zirconia Gels, Thermochim. Acta, 85, 1985, pp. 3-13. /20/ A. Clearfield, Crystalline Hydrous Zr02, Inorg. Chem., 3,1964, p. 146. /21/ D. Dollimore, A. Dyer, G.A. Galmen and C.A.C. Kang, Proc. 2nd European Symp.Thermal. Anal., ed. D. Dollimate, Heyden 1981, p. 387. /22/J. Macek and M. Marinäek, A Study of Nickel Zirconia Composite Materials Prepared by Gel-Precipitation Method in Nonaqueous Media, Fizika A, 4 (2), Zagreb 1995, pp. 413-22. /23/ R.C. Garvie, High Temperature Oxides, Ed. M.A. Alper, Academic Press, N. Y., 1970, pp. 118-164. 92 J. Macek, M. Marinsek: The Preparation of Nickel/Zirconia Dispersions from Nicke! Hydroxide/Hydrous Zirconium Oxide Gel-Precipitate ... Informacije MIDEM 26(1996)2, str. 86-93 /24/ T.A. Wheat, Preparation and Characterization of Lime-Stabilized Zirconia, J. Can. Ceram. Soc., 42,1973, pp. 11-18. /25/ A.M. Godalla and T.W. Livingston, Thermal Behavior of Oxides and Hydroxides of Iron and Nickel, Thermochlm. Acta, 145, 1989, pp. 1-9. 26/ C. Duval, Inorganic Thermogravimetric Analysis, Elsevier Publishing Company, Amsterdam, 1953, p. 224. /27/ M. Marinšek, B. Novosel and J. Maček, Dehydration of Zirconia-Gels Followed by Thermal Analysis, Proc. 23rd International Conference on Microelectronics, MIEL'95 and 31st Symposium on Devices and Materials, SD'95 (I. Šorli, B. Kren, M. Limpel), Terme Čatež, Slovenia, 1995, pp. 289-94. prof. dr.Jadran Maček Marjan Marinšek, dipl. ing. Fakulteta za kemijo in kemijsko tehnologijo, Katedra za anorgansko kemijsko tehnologijo, Aškerčeva 5, 1000 Ljubljana, Slovenija Tel.: +386 61 176 05 19 Fax: +386 61 125 82 20 Prispelo (Arrived): 02.04.96 Sprejeto (Accepted): 07.05.96 93 Informacije MIDEM 26(1996)2, Ljubljana UDK621,3:(53+54+621 +66), ISSN0352-9045 NITROGENATION OF SmaFeiy ALLOY WITH Ta ADDITION B. Sajea,b, B. Reinschc, S. Kobe-Beseničarb, D. Kolarb, I.R. Harrisd a Magneti Ljubljana d.d., Ljubljana, Slovenia bJožef Stefan Institute, Ljubljana, Slovenia c Max Planck Institute for Metals Research, PML, Stuttgart, Germany*, dSchool of Metallurgy and Materials, University of Birmingham, United Kingdom Key words: permanent magnets, nitrogenation, nitrides, Ta, Tantalum, Sm-Fe-Ta alloys, Sm-Fe-Ta powders, TPA ThermoPlezie Analyzers, SEM, Scanning Electron Microscopy, XRD, X-Ray Diffraction, TMA, Thermomagnetic Analysis Abstract: The nitrogenation behaviour of a Sm-Fe-Ta based alloy which can be used for the preparation of Sm-Fe-N based permanent magnets has been described. Diffusion experiments on thin polished plates provided the nitrogenation processing parameters. Thermomagnetic analysis of partially and fully nitrided powders showed that the required nitrogenation times are somewhat lower than the calculated values which was attributed to the powder condition. Nitriranje Sm2Fei7 zlitine z dodatkom tantala Ključne besede: magneti trajni, nitriranje, nitridi, Ta tantal, Sm-Fe-Ta zlitine, Sm-Fe-Ta prahovi, TPA analizatorji termopiezo, SEM mikroskopija elektronska skenirna, XRD uklon Rentgen žarkov, TMA analiza termomagnetna Povzetek: Opisan je postopek nitriranja zlitine Sm2Fe-i7 z dodatkom tantala, ki je primerna za izdelavo trajnih magnetov na osnovi Sm-Fe-N. Procesne parametre smo določili s pomočjo difuzijskih eksperimentov. Za te eksperimente smo uporabili tanke polirane ploščice. Termomagnetna analiza delno in v celoti nitriranih prahov je pokazala, da so časi potrebni za nitriranje nekoliko krajši od izračunanih. To pripisujemo morfologiji prahov. Introduction and experimental work Permanent magnets based on the Srri2Fe-i7N3-§ (8=0.3) interstitial ternary phase are considered to be an attractive proposition for bonded magnets/1/. Unfortunately the binary Srri2Fei7 phase is formed through a peritectic reaction between primarily crystallised iron and Sm-rich liquid. Free iron especially, unless removed by a subsequent isothermal homogenisation treatment, reduces the coercivity of the subsequent nitride when used for permanent magnets. Known methods for creating an alloy without free iron are either high tempera-ture-long term annealing or addition of up to 5 at.% of Nb /2/ or Ta /3/. There is much theoretical and experimental evidence of the nitrogenation of as-cast and homogenised alloys, but the diffusion parameters such as activation energy (Ea) and preexponential frequency factor (D0) appear quite inconsistent. The activation energy for nitrogenation in pure nitrogen ranges from 66 to 133 kJ/mole and frequency factor (D0) from 1 .02*10~6to 1.95*10-10m2/s /4-7/. There are no data for Ta modified alloys. Therefore it was the aim of this work to study comparatively the nitrogenation of as cast and annealed standard and Ta modified alloy to obtain diffusion parameters which would help to predict optimal processing parameters. * present address: Robert BOSCH GmbH, FV/FLW, POB 106050, D-70049 Stuttgart The nitrogenation was carried out on induction melted Srri2Fei7 and Sm2Fei6Tai alloys in pure nitrogen. The stoichiometric composition of cast Sm2Fei7 was additionally homogenised for one week at 1100°C in argon to obtain nearly single phase material. The approximate nitrogenation temperature was determined by means of athermopiezic analyser (TPA). Alloys were nitrided in 1 bar of pure nitrogen at temperatures from 350 to 550°C for different times (from 1 to 16 h) and examined with optical (Zeiss) and scanning electron microscopy (SEM Jeol EPMA 840 A). From the depth of the nitrogenated layer, activation energy and frequency factor were calculated. The Srri2Fei6Tai alloy was also milled in a ball mill to study the nitrogenation behaviour of powder. Powder was nitrogenated for different times at 1 bar of pure nitrogen at 450 °C. The nitrogenated powder was then characterised by means of scanning electron microscopy (SEM, Jeol EPMA 840 A), X-ray diffraction (XRD Philips, Cu Ka source, step scan mode, step 0.02°, time/step 10 s, Ag as a standard) and thermomagnetic analysis (TMA, Manics DSM 8, horizontal Faraday principle, Hext =100). Results and discussion The results of the diffusion experiments are shown on Fig.1. The square of the average depth of the nitrided layer (and therefore the nitrogen diffusion) in the stoichiometric alloy is slightly larger than that in the Ta 94 B. Saje, B. Reinsch, S. Kobe-Beseničar, D. Kolar, I.R. Harris: Nitrogenation of Sm2Fe17 Alloy with Ta Addition Informacije MIDEM 26(1996)2, str. 94-96 modified alloy. From the measurement the activation energy for Srri2Fei7 alloy was determined to be 82.32±8.97 kJ/mole with frequency factor of 1.7*10"10 m2/s and activation energy for Sm-Fe-Ta alloy 92.82 ±11.96 kJ/mole with frequency factor of 5.3*10"10 m2/s. From the data obtained it was possible to calculate that sufficient nitrogenation time for spherical particles of 10 (im diameter would be around 10 hours, according to the equation published in /7/. Fig. 1: Square of the average nitrogen layer depth vs. time at 72.3 K For the nitrogenation and magnetic properties measurements the milled powder was used. The average particle size of the powder was about 10 |j.m (as determined with Cilas Alcatel Laser particle sizer) but its irregular morphology has to be noted as well as the broad particle size distribution. Due to these features slightly different nitrogenation behaviour was anticipated as predicted in theoretical modelling /7/. XRD diffraction of the Sm-Fe-Ta powder nitrided for 10 hours showed characteristic peak shifts due to lattice expansion of the Sm2Fei7 phase (Fig. 2a) when compared with the XRD trace of the non nitrided Sm-Fe powder (Fig 2b). The TaFe2 phase didn't change upon nitrogenation and there is no free Fe detectable in the Ta modified non nitrided alloy (see Fig 2c). There was also a small peak attributed to free Fe observed in the nitrided powder (Fig. 2a). Since the average nitrogenation temperature was too low to induce the overall decomposition of the nitride this was attributed to the combined effect of the decomposition of the Sm2Fei7Nx into SmN and secondary Fe during nitrogenation due to surface effects as reported in /2/ and possible presence of the remanent primary Fe from the cast material. Several features are apparent from the thermomagnetic scans of fully, partially and non nitrided Sm2Fei7 alloy which are shown on Fig. 3. Only the magnetisation curves of non nitrided and the alloy nitrided for 6 hours exhibit one Curie point corresponding to Sm2Fei7 and Sm2Fei7N3-§ phase respectively. The other traces exhibit contribution from both the nitrided shell and the core. The Curie point of the Srri2Fei7N3-5 shell (470°C) remains virtually unchanged irrespective of the nitrogenation times. This shows, that, even after a short nitrogenation time, a layer of nitrogen saturated shell is formed. The Tc of the core increases with nitrogenation time over the range of 100°C. This shift may be caused by the expansion of the core, due to the strain caused by the volume expansion of the nitrogenated shell. Another possibility is that it is due to a combination of this factor together with the presence of regions of intermediate nitrogen concentration, as shown in the Sm-Fe-Nb system /8/. The Sm-Fe-Ta powder appeared to be fully nitrided even after 6 hours of nitrogenation (for the given processing parameters) which is not in agreement with calculations in which temperature, time, N2 pressure and average particle size were used as the defined values. This difference was therefore attributed to the state of the milled powder i.e. irregular morphology, possible cracks or even anisotropic diffusion through the lattice, geometrical factor, particle size distribution and the surface condition which were omitted from the calculation. Fig. 2: XRD traces of a) nitrided Sm-Fe-Ta alloy, b) as-cast Sm2FeM alloy, and c) as-cast Sm-Fe-Ta alloy. 0 100 200 300 400 500 600 700 800 Temperatura (b, of Ag/n-Si(111) Schottky diode when deposited by conventional UHV vacuum deposition methods as derived from l-V measurements /3/, is reported to be 0.78 eV. However, if deposited by ionised cluster beam, ICB, deposition method /1/, depending on the acceleration voltage, Ua, the apparent potential barrier height of the above structure, seemingly could be tailored to vanish. For our purpose it is to be noted, that by the proper choice of Ua, the silver ions possess enough kinetic energy to penetrate the Si substrate. Since the accelerated Ag+ ions may, for large enough Ua, penetrate the Si wafer a few nm in depth, they might strongly contribute to an additional doping density /4/ within this region. In this respect an attempt /4/ has just um Fig. 2. An example of the room temperature measured reverse l-U characteristics of ionized cluster beam deposited Ag/n-Si(111) Schottky diodes forAg+ ions acceleration voltages Ua=0, 300 V (these measurements on the drawing coincide) and 1kV, is presented /6/. The circles, squares and the triangles are the calculated values, derived for the case of thermionic-diffusion theory of the majority charge carriers transport taking the Schottky barrier height as a parameter, as described in ref. /6/. 98 Informacije M1DEM 26(1996)2, str. 97-106 B. Cvikl: On Low Frequency C-U Relationship of the Ionized Cluster Beam, ICB, Deposited Ag/n-Si(111) Schottky Diodes tively behaves as a linear element, however in the forward direction the conditions, due to the experimental difficulties, are presently stil! not yet determined in details. In the latter work /6/ it was shown that the distinct features of the reverse l-U temperature dependent measurements (i.e. the appearance of the "knee bend" as well as the large differences of the reverse saturation current densities) of these ICB deposited Schottky junctions could be sufficiently well described in terms of the thermionic-diffusion theory of the charge carrier transport. In the computation the existence of the distinct Ag+ donor additionally enriched region in n-Si sample extending from the Ag/Si junction up to the Ag+ ions maximum penetration length, L, (which is a function of Ua) has been assumed. The doping effect of oxygen, carbon, sulfur and other surface impurities, which are for Ua nonzero also present within Si wafer, have been neglected. The results obtained seem to imply the fact, 0.007 - 0.006 - 0.005- 0.004 - Us vS" 0.003 - "6 0.002 - 0.001 - 0.000 - -0.001 - Ua=300V T=297K 2 Ur [V] UJV] Fig. 3. Two examples of low frequency, v =2 kHz (the amplitude of the ac measuring signal was set to 10 mV), room temperature measurements of the C-U relationship of ionized cluster beam deposited Ag/n-Si(111) Schottky diodes for Ag+ ions acceleration voltages of Ua=300 V and 1 kV, are exhibited /13/. The lines are guides to the eyes only. that by the ICB deposition method it might be possible directly to modulate the Schottky barrier height. The measured low frequency capacitance-voltage, C -U, relationship exhibit similar peculiarities, fig. 3. in the sense that for low values of silver ions acceleration voltages the diagram C"2 versus applied voltage U (fig. 3) exhibit well defined linear relationship as predicted by the theory /3/, however, say, for acceleration voltages Ua > 2 kV, it may soon become nonlinear. It is a well established fact /7/, that the Si wafer, following the surface preparation, is under normal laboratory atmospheric conditions almost immediately covered by extremely stable (for the period up to one year) dielectric thin film consisting of about 0.7 nm thick native oxide layer on top of which about 0.2 nm thick organic contamination layer is in general also present. Obviously, the extent of the Ag+ enriched n-Si regions is most conveniently expressed in terms of the penetration lengths, L, which is accessible to Ag+ ions (at a given value of Ua) within the n-Si wafer. This distances have been correspondingly calculated (including the 1.2 nm thick oxide layer on the Si surface) and have been found to be L=2.4 nm for Ua=300 V, and L=4.0 nm for Ua= 1 kV /8/. As it is well known /3/ the impurity silver atoms incorporated within the Si substrate can act either as donors (Ed (Ag) in Si = 0.370 eV above Ev) or as V"VW —Vr, Fig. 4. Schematic drawing (not to scale) of forward biased ionized cluster beam deposited Ag/n-Si(111) metal/semiconductor junction. The region within the semiconductor, up to the abrupt plane at depth L, contains additional (besides electrically activated shallow phosphorous dopant level denoted as dots) deep lying Ag donor and acceptor impurities, which are assumed to be homogeneously distributed. L denotes the maximum penetration range, of silver ions for the given value of the external acceleration voltage, Ua. Within the region 0'dv (5> from which it follows, 'dV .dx -ftpi(V)dV (6) and the minus sign is taken on account of potential V decrease with an increasing depth x. At this point it has to be mentioned that at x=L, due to the discontinuity of the excess charge density the electric potential V, while itself continuous, possesses the discontinuous derivatives. Similarly, the integration of the eq. (4) within the range L < x < W, results in dV dx JvL Pi(V)dV (7) since due to potential being flat at the end of the depletion region (dV/dx)x=w = 0 and once again the minus sign is taken on account of potential V decrease with an increasing distance x. From the fig. 4. it is evident that the redistribution of space-charge due to the deep Impurities present is expected to occur within the region 0 < x < L for small applied forward and reverse voltages. Specifically, for large forward U, all deep lying Ag+ donors will be electrically neutral and the Ag" acceptor states will be completely populated resulting in an effective decrease of the net positive space charge within this region. And conversely, for larger reverse applied voltages U the acceptor states are empty and the excess positive space charge pi (x) within the same region exceeds that of the normal conditions as provided by the phosphorous doping in n-Si substrate. Obviously, for each value of the silver ions concentrations NAg within this enriched region there exists an interval of the applied external voltage to which monotonically corresponds the space charge density variations pi (x) for 0 < x < L. For this reason the electron potential energy qV(x), for concentrations of electrically activated silver atoms Nag large enough, may considerably vary for different values of the applied voltage and it is just this variation which presumably might be responsible for the observed C-V behavior of ICB deposited Schottky diodes. In order to compare the experimental l-V characteristics to the predicted ones based upon the model depicted in fig. 4., the detailed numerical evaluation of the electron potential V(x), for 0 < x < W, at each value of the applied external voltage U is required. This, however, is a subject of a future publication /13/. The general characteristics of the differential capaci-tance per unit area, C, of the model diode defined as, C = 9Q, (8) can be for the above case discussed in the sufficient particulars without adhering to such the detailed numerical procedures. In the eq. (8), the term 3Qd represents the infinitesimal change in the total charge per unit area due to all uncompensated electrically active impurities in the depletion region, which is occurring due to the 3Vd, the infinitesimal change in the so called diffusion potential Vd. The diffusion potential, as well known, 101 B. Cvikl: On Low Frequency C-U Relationship of the Informacije MIDEM 26(1996)2, str. 97-106___Ionized Cluster Beam, ICB, Deposited Ag/n-Si(111) Schottky Diodes is for an ordinary Schottky diode defined as the difference between the potential at the junction metal/semiconductor interface (i.e. x=0) and the potential evaluated at the edge of the depletion region, x=W, thus Vd = V(x=0) - V(x=W). In what follows we adhere to the approximation (termed the low frequency approximation) according to which occupations of donor or acceptor energy levels instantaneously respond to the capacitance measuring test small signal (alternating) voltage which is superimposed on an externally applied direct voltage bias. Viewing fig. 4., the low frequency depletion-layer capacitance C per unit area can be evaluated as a resultant capacitance of two capacitors, Ci, and C2, connected in series. In order to see this one introduces at x=L two connected, infinitely closely spaced, conducting plates. On account of induced charges appearing on each plate, the charges are equal in magnitude but of opposite signs. Consequently, one may immediately generalize, dV dx rdv ) . = 9l x-L Idx. L~ es dV" ) _ Ql . dx, (14) (15) The excess charge densities for the two regions remains to be defined. In the lowest approximation one writes, noting that the excess charge density per unit area of eq. (12), Q1, is directly related to pi of the eq. (7) simply as Q1 = pi/S, and similarly for Q2, where the net excess charge density in the region (1) is brought about by the electrically activated phosphorous shallow donors (of density Np+) in conjunction with the deep lying Ag (donor as well acceptor) impurity atoms of densities NAg+ and NAg", p1 = q(Np+ + Na9+ - NAg") (16) C, 3q, 3Vdi ¡ = 1,2 where, and similarly Vd1 = V(x = 0)-V(x = L) Vd2=V(x = L)-V(x = W), (9) (10) (11] respectively. From here onwards the following abbreviations; V(x=0) = V0 and V(x=L) = Vl will be adhered to. The internal electric field in the semiconductor is oriented in the negative direction of x-axis from the edge of the depletion region (where V(x=W) = constant and consequently E=0) towards the metal/semiconductor interface junction. Applying the Gauss's law, £DdS = p, first on the surface of the cylinder placed between x=0 and x=L, and secondly on the surface of the one placed at x=L and x=W one obtains, E„ - E, Q, p2 = q(Np+ - Ne~(x=L)). (17) Here, Ne"(x= L) denotes the density of the displaced free carriers (i.e. electrons) at the edge of the enriched region, x=L /4/, but the rough estimate shows that this contribution can be, if one is not being interested in the C-V frequency dependence, in the first approximation, neglected. Generally speaking, this contribution also varies (on account of the electrical activation or deactivation of deep lying silver donors and acceptors) with the change of the applied external bias DC voltage. For the exact calculations, the shape of the conduction band in the vicinity around x=L has to be calculated in details. The probability, w(E), for donor or acceptor to be electrically activated is given by /3/, w(E„ n; Nd 1 + gde 1 -(Ed-EF) (18) (12) where, Ed is the donor energy level (as defined with respect to the conduction, Ec, or alternatively valence energy band, Ev, respectively), Ef the Fermi energy in the semiconductor and gd is the ground state degeneracy of the donor impurity level, equal to gd=2. Obviously the factor (1-w(Ed)) is the probability, that the donor is not electrically activated. Similarly, the probability that the acceptor level is electrically active is provided by the expression, where Q1 denotes the excess charge density per unit (junction) area within the first, i.e. Ag enriched, region. Similarly, one obtains Q2 L _ yielding the two expressions to be evaluated as, (13) w(ea) = n: 1 1 + gae "(Ea-EF) (19) where ga, the ground state degeneracy factor is equal to ga=4, for acceptor levels of energy Ea. As schematically indicated on fig. 4, due to the metal/semiconductor junction potential barrier the elec- 102 Informacije M1DEM 26(1996)2, str. 97-106 B. Cvikl: On Low Frequency C-U Relationship of the Ionized Cluster Beam, ICB, Deposited Ag/n-Si(111) Schottky Diodes tronic potential V(x), or equivalent^ the conduction band Ec, exhibits relatively strong curvature in the interval 0 b) „Pzd b2 = 9d,P e Pq(u-ib) ; 9a,Ag 1 -pq(U-<&b) —Pza (26) (27) (28) In the transformations above, the following relation has been used, Vdi =V0-Vl = (t>b - U - Vl, fig. 4, where <(>b, is the Schottky barrier height. Likewise, the evaluation of the second integral yields, l2(Vd2) = !^p2(V)dV = j0Vd2p2(s + Vw)ds = qNdP and b4 is given by, 1 , 1 + b4e~PqVd2 Vri„ + — In--- d2 Pq 1 + b4 b4 = gd,p e e (29) (30) In the derivation the relation, Vd2 = Vl-Vw = Vl-(U + where = (kT/q)ln(Nc/Nd), and Nc is the effective density of states in the conduction band has been employed. Finally, performing the indicated operations as suggested by the eq. (9), noting that Qi = Qi(Vdi, Vd2, Vd), where the diffusion potentials are defined by the eqs. (10), (11), consequently 3Q, 3Q, 3Q, 3Q, 3Q< t 1 - - + —- = —1--1 = etc. av, dvr„ avn avri the following capacitance's per unit area are derived: These expressions are to be evaluated using the eqs. (16) to (22). After the appropriate transformation of the integral limits, noting the relevant integrals can be easily evaluated, the results are given by, li(Vd1) = v°pi(v)dV = j^1 p,(s + VL)ds: ivL q N d,Ag V, VH, 1 , 1 + b.e^1 -In 1 (3q 1 + b, 1 , 1 + bPePqVd1 •— "--- Pq 1 + b, N a,Ag 1 . 1 + b3ePqVd1 yM - — In— 3----- Pq 1 + b, (25) d.Ag N. a.Ag c, = q. 1 + b^"1 1 + b2epqVd1 ! + b3iPqVd1 Vi,(vdi) + i2(vd2) and similarly, Nri I- j u c 2 = q]h.u*> Na,Ag and Nd.Ag, In the lowest approximation (neglecting the Ag impurities in Si substrate), pi(V) = p2(V) = p(V), and one therefore obtains from the eq. (23), (34) Qi = V2^7[VC p(v)dv - K p(v)dv and likewise, Q2=^2ea^p(Y}W, the edge of the depletion region (q£, = kTln(Nc/Nd) and Nc is the effective density of states in the conduction band). Consequently, combining eqs. (36) and (37) the usual linear C"2 versus U relationship is obtained, in accordance with the C-U measurements for Ua = 0 and 300 V samples, as indicated on fig. 3., respectively. Some relevant examples of the low frequency C-U relationships, calculated for the various values of the deep level impurity concentrations, taking the potential, Vl, as an independent parameter, are explicitly exhibited on figs. 5-7. U [V] Fig. 5. The calculated values of the resulting capacitance, 1/C2, based on the eq. (33) of the text, versus the external applied DC bias, U, is presented for various values of the parameters. The lines shown are calculated for the following values of parameters expressed in terms of the constant No=1016 cm"3: line a) Vl=0.45 V, Nd,Ag=10 N0, A/d,p=7 No, Na,Ag=40 N0; line b) Vl=0.65 V, A/d,Ag=70 No, Na,p=1 N0, /Va,Ag=40 N0; line c) Vl=0.65 V, Nd,Ag=0, Nd,p=1 N0, A/a,Ag=0- The latter curve represents the usual case of the homogeneously doped metal/Si semiconductor junction. For the discussion of C-U relationship at large values of Ag ions acceleration voltages, one has to adhere to the full numerical analysis, originating from the exact solution of the Poisson equation /13/. Nevertheless, the simplified model, as depicted on fig. 4., of ICB deposited Ag/Si diodes might possibly provide a direction towards understanding of the findings presented in ref. /6/, where an unusual feature was reported. Namely, applying the usual l-V evaluation procedure /14/ on the raw experimental data, it was found, quite contrary to the expectations, that the ideal- 104 B. Cvikl: On Low Frequency C-U Relationship of the Informacije MIDEM 26(1996)2, str. 97-106_Ionized Cluster Beam, ICB, Deposited Ag/n-Si(111) Schottky Diodes o Fig. 6. The calculated values of the resulting capacitance, 1/C2, based on the eq. (33) of the text, versus the external applied DC bias, U, is presented for various values of the parameters. The lines shown are calculated for the following values of parameters, expressed in terms of the constant A/0= 701 cm'3: line a) Vl=0.45 V, A/d,Ag=0, A/d,p=7 N0, A/a,Ag=0; (the curve represents the usual case of the homogeneously doped metal/Si semiconductor junction); line b) Vl=0.45 V, A/d,Ag=7 No, A/d,P= 1 No, N a,Ag=40 N0; line c) V\_=0.45 V, A/d,Ag= 10 No, Nd,P= 1 N0, /Va,Ag= 100 No- Fig. 7. The calculated values of the resulting capacitance, 11C2, based on the eq. (33) of the text, versus the external applied DC bias, U, is presented for various values of the parameters. The lines shown are calculated for the following values of parameters, expressed in terms of the constant No=10 cm"3: line a) Vl=0.45 V, Nd,Ag=0.3 No, Nd,P=1 N0, Na,Ag=90 N0; line b) Vl=0.45 V, A/d,Ag=0, A/d,p= 1 No, Na,Ag=90 N0; line c) Vl=0.45 V, A/d,Ag=2 No, Nd,P=1 N0, Na,Ag=60 N0. Note that the shape of the calculated cun/es exhibits different curvatures, depending upon the values of parameters. ity factor, n, as well as the series resistance, Rs, are bias dependent. The results of an alternative, rather involved and more general, analysis /15/ seem to support the above findings. One notes, that the results of ref. /6/, are based upon the assumption, that the n-Si enriched region is characterized by the constant density of electrically activated silver donors throughout the whole interval of bias investigated. Now, according to the model of fig. 4, this assumption is certainly no longer valid; for forward biases, due to the increasing density of activated acceptor silver impurities accompanied by the corresponding density decrease of the Ag+ donors, the net excess space charge within 0 < x < L interval ought to be bias dependent and at all times smaller than if only shallow donor P impurities would have been present. Consequently, the series resistance ought to be bias dependent all the way up to its given upper limit above which the net space charge ought to remain constant. Similar characteristic feature, but in the reversed order, is expected to exists for the case of an applied reverse bias. This behavior would be expected to appear always, providing the enriched range exceeds certain minimal Lmin and/or certain minimal Ag impurities densities, which are yet to be determined and thoroughly investigated. Strictly speaking, the presented ICB Ag/n-Si structures consequently may not be truly considered to be a typical representative of an ordinary Schottky metal/semiconductor diodes. The measured strong temperature l-U dependency /4/ of the ICB deposited Ag/n-Si(111) Schottky diodes might now be possible, on the basis of the proposed model exhibited on fig. 4., to assign primarily to the effect of the strong temperature dependent activation of the additional Ag donors and acceptors rather than (in conjunction with the variation of P donors activation) to the direct 1/kT variation itself. There exists another very important feature which deserves a comment. Namely, as reported in ref. /4/ and /6/ the effective Schottky barrier height presumably decreases with an increased silver ions acceleration voltage Ua in clear disagreement with the DIGS model prediction of the strong Fermi level pining at the metal/semiconductor interface for all cases of strong disorders of a semiconductor interface region. Consequently, as DIGS model is generally considered to work well in clear cases of interface disorders, and as an ICB deposition method is certainly expected to produce just such an effect, the question how the effective Schottky barrier is related to the true Schottky barrier height is an important question which still requires to be answered. 4. Conclusions The low frequency capacitance- voltage, C-U, relationship of ionized cluster beam, ICB, deposited Ag/n-Si(111) Schottky diodes for nonzero acceleration voltage of the silver ions, has been investigated on the basis of the postulated model of the semiconductor energy bands. In the derivation the fundamental assumptions made is the homogenous, spatially limited distribution, laying within the semiconductor energy gap, of the bias voltage dependent activated deep Ag 105 Informacije M1DEM 26(1996)2, str. 97-106 B. Cvikl: On Low Frequency C-U Relationship of the Ionized Cluster Beam, ICB, Deposited Ag/n-Si(111) Schottky Diodes donor and acceptor impurities energy levels. These are (in addition to the shallow phosphorous donor level throughout the Si sample) spatially confined within the Si region extending up to the maximal silver ions penetration length L, which is a function of the silver acceleration voltage Ua. It is argued, that it is only for small values of L and/or small Ag impurity concentrations, within the described impurity additionally enriched semiconductor region, that the C-U relation is expected to exhibit the linear relationship, in accordance with the measurements. The previously observed bias dependence of the semiconductor series resistance, at constant temperature, as well as the strong temperature dependency of the previously reported l-U measurements, are being phe-nomenologlcally explained in terms of the proposed model. The effects of the additional, within the Si energy gap placed impurity levels arising on account of surface atoms, like O, C, N and traces of S, to be in the process of the Ag impact also transferred into the semiconductor, have been neglected. These impurities may contribute to the additional space charge within the Ag enriched n-Si region, but very likely their importance, as traps, ought to be taken into account if one is concerned with C-V frequency dependency. However, their possible effect, as microcluster formations, on the Schottky barrier height formation, in a sense as first proposed by Freeouf and Woodall /16/, is not to be neglected. It is argued, that the relationship of the disordered (due to Ag and surface atoms) enriched semiconductor layer (at small penetration lengths L), on the Fermi level pinning and consequently on the DIGS implications concerning the metal/semiconductor junction, might be most conveniently investigated by carefully controlled and suitably designed ICB experiments in UHV conditions. 5. References /1/ T. Takagi, Vacuum 36, (1986) 27. /2/ W. L. Brown, M.F. Jarrold, R.L.McEachern, M. Sosnowski, G. Takaoka, H. Usui and I. Yamada, Nucl. Instr. and Meth. in Phys. Research, B59, (1991) 182. /3/ S.M. Sze in Physics of Semiconductor Devices, 2nd Edition, John Wiley & Sons, New York (1981). /4/ B. Cvikl and T. Mrdjen, Fizika, A4, (1995) 2, 403. /5/ F. A. Padovanl and R. Stratton, Solid-State Electron. 9, (1966) 695, see also C. R. Crowell and V. L. Rldeout, Solid-State Electron., 12, (1969) 89. The discussion of their results Is also presented in the reference /12/. /6/ B. Cvikl, Zs. J. Horvath, T. Mrden, 23rd International Conference on Microelectronics, MIEL'95 and 31st Simposium in Devices and Materials, SD'95, Proceedings, p. 391 -396. September 27.-29, 1995, Terme Čatež, Slovenia. /7/T. Takahagi, I. Nagal, A. Ishitanl, H. Kuroda, Y. Nagasawa, J. Appl. Phys. 64 (1988) 3516. /8/J. F. Ziegler, J. P. Biersack and U. Littmark in The Stopping and Range of Ions in Solids, Pergamon Press, New York, (1985). /9/ H. Hasegawa and Hideo Ohno, J. Vac. Sci. Technol., B 4, (1986) 1130, see also K. Koyanagi, S. Kasai and H, Hasegawa, Jpn. J. Appl. Phys., 32, (1993) 502. /10/ R. R. Senechal and J. Baslnski, J, Appl. Phys., 19, (1968) 3723. /11/ G. I. Roberts and C. R. Crowell, J. Appl. Phys., 41, (1970) 1767. /12/ E.H. Rhoderic and R.H. Williams In Metal-Semiconductor Contacts, 2nd Edition, Clarendon Press, Oxford, (1988). /13/ T. Mrdjen, B. Cvikl and D. Korošak, to be published. /14/ S.K. Cheung and N.W. Cheung, Appl. Phys. Lett., 49, (1986) 85. /15/ D. Donoval, M. Barus and M. Zdimal, Solid-State Electronics, 34, (1991) 1365. /16/ J. L. Freeouf, J. Woodall, J. Appl. Phys. Letts, 39, (1981) 727. Acknowledgment The endeavors of Mess. M. Ko'elj, T. Mrden, F. Moškon, E. Krištof and D. Korošak, at the Division of Reactor Physics, in the course of very involved ionized cluster beam experiments and data treatment, as well as to Professor A. Levstek and M. Sc. C. Filiplc at the Condensed Matter Physics Division of the "J. Stefan" Institute, for having made the equipment available and for helping us in C-V measurements, are all greatly appreciated. The acknowledgment is also due to M. Sc. R. Jecl, Faculty of Engineering, for numerical evaluations serving as a basis for some of the arguments presented in this presentation. Dr. Bruno Cvikl Fakulteta za gradbeništvo, Univerza v Mariboru, Smetanova 17, 2000 Maribor, ali Institut "J. Stefan", Univerza v Ljubljani, Jamova 39,1000 Ljubljana Tel.: +386 (0)61 188 54 50 Fax: (0)61 374-919; Email: bruno.cvikl@ijs.si Prispelo (Arrived): 31.5.1996 Sprejeto (Accepted): 18.6.1996 106 UDK621,3:(53+54+621 +66), ISSN0352-9045 Informacije MIDEM 26(1996)2, Ljubljana IZVEDBA NEREKURZIVNEGA DIGITALNEGA SITA S STANDARDNIMI INTEGRIRANIMI KOMPONENTAMI V MODIFICIRANI OBLIKI PORAZDELJENE ARITMETIKE K. Korošec, A. Vesenjak, B. Jarc, M. Solar, R. Babič Fakulteta za elektrotehniko, računalništvo in informatiko, Univerza v Mariboru Maribor, Slovenija Ključne besede: DSP procesiranje signalov digitalno, filtri digitalni, filtri nerekurzivni, FIR filtri s trajanjem omejenim impulza odzivnega, filtri digitalni s trajanjem omejenim Impulza odzivnega, izvedbe praktične, aritmetika porazdeljena, koeficienti v aritmetiki porazdeljeni, deli sestavni integrirani standardni, CMOS vezja EPROM pomnilniki, H CMOS vezja hitra, frekvence vzorčenja Povzetek: V članku je opisana izvedba univerzalne strukture nerekurzivnega digitalnega sita s 15 koeficienti v porazdeljeni aritmetiki s standardnimi integriranimi komponentami. Pri tem smo uporabili novi modificirani postopek izračunavanja delnih vsot koeficientov, ki omogoča povečanje dinamičnega območja izhodnega signala vsaj za 6 dB, obenem pa dosežemo tudi zmanjšanje aparaturne kompleksnosti strukture. Pri uporabi standardnih hitrih CMOS vezij in nezahtevnega EPROM pomnilnika za shranjevanje vnaprej izračunanih delnih vsot koeficientov smo pri 12-bitni kvantizaciji vhodnega signala dosegli frekvenco vzorčenja fv=143 kHz. S simulacijskimi rezultati smo ilustrirali povečanje dinamičnega območja izhodnega signala po predlaganem modificiranem postopku izvedbe porazdeljene aritmetike, prikazana pa je tudi primerjava med teoretičnimi, simulacijskimi in merilnimi rezultati frekvenčnih karakteristik nizkoprepustnih in visokoprepustnlh digitalnih sit, ki smo jih načrtali s programskim paketom DF-PAK. The FIR Digital Filter Realization with Standard Integrated Circuits in the Modified Distributed Arithmetic Structure Keywords: DSP, digital signal processing, digital filters, nonrecursive filters, finite impulse response digital filters, practical implementations, distributed arithmetic, coefficients in distributed arithmetic, standard integrated components, CMOS circuits, EPROM, H CMOS circuits, sample frequencies Abstract: In this article the hardware realization of 15 tap general FIR digital filter structure with standard integrated circuits for the implementation of digital filters with arbitrary frequency responses is presented. A new modified distributed arithmetic structure is used for the computation of output signal sequence with the purpose to increase the dynamic range of output signal. Apart from concentrated arithmetic mechanization where the output sample calculation is determined with the sum of products of two vectors, the distributed arithmetic technique permits this calculation only with summing and shifting operations of the precalculated partial sums of the coefficients. When the bipolar input signal is converted into unipolar form the offset binary format of input samples is obtained. In this way a new approach to the modified partial products calculations is necessary. Then the output sample calculation is also distinguished from the classical distributed arithmetic technique in one significant step that in the last calculation step no subtraction is needed. Our digital filter is realized in the structure with 12 bits analog to digital and digital to analog conversion for input and output signals, 14 bit memory register length for partial sums of coefficients presentation in look up table and with 16 bits register length of arithmetic unit. With standard H CMOS circuits and EPROM memory with access time of 200 ns the sample frequency of 143 kHz is obtained. As this structure is also suitable for FPGA implementation, higher sampling frequencies are expected. In comparison with classical distributed arithmetic structure the increasing of 6 to 8 dB of the dynamic range of output signal is obtained. This increasing depends slightly on the number of digital filter coefficients and on parameters of predetermined frequency characteristics. Filter coefficients are designed with digital filter design software DF-PAK. The frequency responses with theoretical, simulated and measured results for low pass and high pass digital filter implementations are also presented. 1. Uvod Pri aparaturni izvedbi digitalnih sit, je pomembna izbira realizacijske oblike. Osnovni kriteriji, ki vplivajo na izbiro, so: - dobro ujemanje med teoretičnimi in želenimi oziroma zahtevanimi lastnostimi, - čim manjši vpliv kvanti-zacije na spremembo frekvenčnih karakteristik in velikost šumnega signala na izhodu, - mala aparaturna kompleksnost ter - velika hitrost delovanja. Porazdeljena aritmetika ali ROM akumulator struktura /1, 2/ predstavlja določen prispevek pri izbiri izvedbene oblike. Posebej je zanimiva zaradi manjše in nezahtevne aparaturne kompleksnosti, tako da postaja aktualna za izvedbo tudi s programabilnimi polji logičnih vrat. S porazdeljeno artimetiko je označen postopek izračunavanja skalarnega produkta dveh vektorjev na elementarnem bitnem nivoju brez uporabe običajnih množilnikov, tako da ga lahko s pridom izkoristimo pri izvedbi nerekurzivnih digitalnih sit. Nerekurzivna digitalna sita so zaradi dodelanih postopkov načrtovanja, 107 Informacije MIDEM 26(1996)2, str. 107-112 K. Korošec, A. Vesenjak, B. Jarc, M. Šolar, R. Babič: Izvedba nekurzivnega digitalnega sita s standardnimi... svoje enostavnosti in predvsem linearnega faznega pomika zelo zanimiva za področje digitalne obdelave signalov. Za nerekurzivno digitalno sito zapišemo odziv y(k) v obliki konvolucijske enačbe y(k) = "i h(n) x (k - n), n=0 (1.1) ki ima tudi obliko skalarnega produkta dveh vektorjev y = hT • x. S h(n), n = 0,1,..., N-1 je označena končna sekvenca N koeficientov impulznega odziva, ki obenem določa vektor koeficientov nerekurzivnega digitalnega sita h, z x(k-n) pa je označena sekvenca časovnih odtipkov vhodnega signala, ki določajo komponente vektorja x. 2. Porazdeljena aritmetika in modificirani izračun delnih vsot koeficientov Če so vrednosti vhodnega signala x(k) omejene na intervalu [-1, +1], jih lahko zapišemo tudi v dvojiški binarni obliki s končno dolžino besede Bx bitov x(k-n) = -b0(k-n)+ £b,(k-n)-2- , n = 0,1,...N-1. množenja z 2"'. Za seštevanje uporabimo seštevalnike, ki so dovolj enostavna in hitra vezja, za odštevanje je potrebno posebej tvoriti dvojiški komplement, množenje z 2"' pa se preprosto izvaja s pomikom vsebine v aritmetični enoti za i-bitov in z zakasnilnimi elementi T. Zaradi pridobitve na hitrosti odziva vse možne delne vsote koeficientov običajno izračunamo vnaprej in jih zapišemo v pomnilnik vrste ROM, njihove vrednosti pa pri izračunu izhodnih vrednosti sproti določa naslovni vektor vhodnih spremenljivk, ki ga definiramo z vhodnim poljem. Na sliki 2.1 je prikazana bločna shema osnovne strukture nerekurzivnega digitalnega sita v porazdeljeni aritmetiki. arilmclična enota x(k) t'¡(10 h,(k-l) ROM 2N -lokacij t>,(k-N+l) x(k-l) vhodno polje B,\ bitov x(k-N+l) Z y(k) p —N izhod y Slika 2,1: Osnovna struktura nerekurzivnega digitalnega sita v porazdeljeni aritmetiki (2.1) Z bi(k-n) so označene binarne spremenljivke, ki zavzemajo vrednosti 0 ali 1, pri tem je bo(k-n) najbolj utežni bit, ki predstavlja predznak, bBx-i(k-n) pa je najmanj utežni bit. S povezavo obeh izrazov po enačbah 1.1 in 2.1 dobimo osnovo za izračun trenutne vrednosti izhodnega signala za nerekurzivno digitalno sito po principu porazdeljene aritmetike /2/. y(k) = -'Ih(n) ■ b0(k - n) + i"1 lh(npt(k - n)2"' n=0 ¡=1 n=0 Cez v,(k)= Xh(n)bi(k--n) n=0 (2.2) ;2.3) označimo delne vsote koeficientov, ki predstavljajo vmesni korak pri računanju y(k), dobimo poenostavljeno obliko zapisa izračunavanja izhodnih vrednosti v obliki y(k) = -v0(k)+ Sv,(k)2-. ¡=1 (2. 4) Pri uporabi konstantnih koeficientov digitalnega sita so delne vsote odvisne le od nabora N binarnih spremenljivk bi(k-n). Za določitev trenutne izhodne vrednosti potrebujemo le postopek seštevanja (odštevanja) in V tabeli 2.1 so prikazane delne vsote koeficientov vi(k) v odvisnosti od naslovnega vektorja brez upoštevanja potrebnega normiranja. Tabela 2.1: Delne vsote koeficientov pri osnovni strukturi porazdeljene aritmetike naslovni vektor Vi (k) 0 0...000 0 1 0...001 ho 2 0...010 hi 3 0...011 ho+hi 4 0...100 h2 gN-1 1...111 hN-1 +hN-2+... + ho Ker raste število delnih vsot koeficientov eksponentov s številom koeficientov sita N, potrebujemo pomnilnik z 2n pomnilniškimi lokacijami. Pri današnjem stanju tehnologije je neekonomično izvajati sita z N>21 koeficienti impulznega odziva. V našem primeru smo izbrali kompromisno rešitev z N = 15, pri čemer potrebujemo ROM s kapaciteto 32 K x 16 bitov. Pri dvojiškem zapisu vhodnega signala je v vhodnem polju z 1000 ... 0 zapisana najmanjša (negativna) vred- 108 K. Korošec, A. Vesenjak, B. Jarc, M. Šolar, R. Babič: Izvedba nekurzivnega digitalnega sita s standardnimi..._Informacije MIDEM 26(1996)2, str. 107-112 nost, z 0111 ... 1 pa največja (pozitivna) binarna vrednost. Modificirano obliko porazdeljene aritmetike dobimo, če pred vpisom v vhodno polje vhodni bipolarni signal pretvorimo v unipolarno obliko x (k -n) = ^(k -n)2~' n = 0,1,...N-1. (2.5) vhod n Tedaj preide izraz za izračunavanje izhodnega signala po enačbi 2.4. v y(k) = ym(k) = 8Xvim(k)-2-', (2.6) i=0 pri čemer so z vim(k) označene modificirane vrednosti delnih vsot koeficientov /3/. S primerjavo izrazov pod enačbama 2.4 in 2.6 vidimo, da v slednjem primeru ne potrebujemo več posebnega postopka odštevanja v aritmetični enoti. S tem se zmanjša kompleksnost prikazane strukture na sliki 2.1, obenem pa se še izognemo možnosti prelivanja vmesnih rezultatov preko normirane vrednosti ena v aritmetični enoti pred končnim odštevanjem vrednosti vo(k) . Zaradi spremenjenih vrednosti vhodnega signala v vhodnem polju je potrebno delne vsote koeficientov simetrirati in normirati. Postopek simetriranja je odvisen od vrste uporabljene frekvenčne karakteristike sita in ga izvedemo za vsako sito posebej. Modificirane delne vsote koeficientov Vmi(k) so prikazane v tabeli 2.2. Tabela 2.2: Delne vsote koeficientov pri modificirani obliki porazdeljene aritmetike. naslovni vektor vmi (k) 0 0...000 1/2(-ho-hi-h2-...-hN-i) 1 0...001, 1/2(+h0-hi-h2-...-hN-i) 2 0...010 1/2(-ho+hi-h2-...-hN-i) 3 0...011 1/2(+h0+hi-h2-...-hN-i) 4 0...100 1/2(-ho-hi + h2-...-hN-i) 2N-1 1...111 1/2(+ho+hi + h2+... + hN-i) Novo modificirano strukturo porazdeljene aritmetike pri izvedi nerekurzivnih digitalnih sit kaže bločna shema na sliki 2.2. Na vhodni strani je sicer potreben dodatni pretvornik bipolarnega signala v unipolarno obliko, zato pa se izognemo postopku odštevanja delnih vsot koeficientov, ki jih določa naslovni vektor najbolj utežnih bitov v vhodnem polju pri izračunu vsakokratne trenutne izhodne vrednosti. SI. 2.2: Digitalno sito v modificirani obliki porazdeljene aritmetike Pri skrbnem simetriranju in normiranju modificiranih delnih vsot koeficientov dosežemo povečanje dinamičnega območja izhodnega signala za najmanj 6 dB. Primerjavo simulacijskih rezultatov frekvenčnih karakteristik nizkoprepustnega sita s petnajstimi koeficienti po klasičnem in modificiranem postopku porazdeljene aritmetike kaže slika 2.1. 0 0.1 0.2 0.3 0.4 0.5 f/fv Slika 2.3: Primerjava frekvenčnih odzivov nizkoprepustnega sita s petnajstimi koeficienti po klasičnem in modificiranem postopku Izhodni signal iz nove strukture je tudi unipolarne oblike, vendar imamo tudi na izhodu na voljo več enostavnih postopkov za njegovo spremembo v običajno bipolarno obliko. 3. Izvedba Digitalno sito v modificirani porazdeljeni aritmetiki smo izvedli v laboratorijski obliki s standardnimi integriranimi komponentami. Zasnova vezja je univerzalna, tako da omogoča realizacijo nerekurzivnih digitalnih sit s 15 koeficienti s poljubno frekvenčno karakteristiko. Za pretvorbo analognega signala v digitalno obliko smo izbrali 12-bitni A/D pretvornik MAX 122, ki omogoča vzorčenje signala z maksimalno vrednostjo fVmax = 333 kHz = 1 /T. Pretvorbo binarnega vhodnega signala x(k) 109 Informacije MIDEM 26(1996)2, str. 107-112 K. Korošec, A. Vesenjak, B. Jarc, M. Šolar, R. Babič: Izvedba nekurzivnega digitalnega sita s standardnimi... v unipolarno obliko smo izvedli z invertiranjem najbolj utežnega bita. Tako pretvorjen signal vodimo preko 12-bitnega vzporedno-zaporednega pomičnega registra v zaporedni pomični register vhodnega polja z velikostjo 180 bitov. Sestavili smo ga iz CMOS gradnikov HCT 164. Vsebina vhodnega polja v trenutku t = kT določa toliko različnih naslovnih vektorjev kot je število bitov vhodne binarne besede Bx = 12 . Za usklajeno delovanje skrbi krmilno vezje s kvarčnim oscilatorjem fo = 2 MHz, ki generira vse potrebne časovne signale: - osnovne urine impulze ui, - inverti-rane urine impulze U2, - impulze za startanje analogno-digitalne pretvorbe U3, - za čitanje vrednosti A/D pretvornika v vzporedno zaporedni pretvorik U4, - za krmiljenje vhodnega polja U5, - za brisanje stare vrednosti iz zadrževalnih celic aritmetike U6 ter - za krmiljenje izhodnih zadrževalnikov U7. Časovne poteke signalov iz krmilnega vezja prikazuje slika 3.1. Osrednji del strukture je pomnilnik za shranjevanje delnih vsot koeficientov. Uporabili smo dve EPROM vezji 27256 saj potrebujemo pomnilnik s kapaciteto 215 x 14 bitov. Za zapis delnih vsot koeficientov smo uporabili 14 bitov. Vsebina EPROM pomnilnika določa obliko frekvenčne odvisnosti izhodnega signala. Programski paket BRUMEC /3/ omogoča natačni izračun delnih vsot koeficientov, 14-bitno kvantizacijo pa smo opravili pred vpisom v pomnilnik. Aritmetično logična enota je sestavljena iz seštevalnega in zadrževalnega vezja. Zaradi zmanjšanja vplivov pogreška kvantizacije aritmetične enote na izhodni signal smo uporabili 16-bitno strukturo. Osnovni gradniki so 4-bitni seštevalniki 4008. Vezje izvaja postopek seštevanja in deljenja z dve. To je izvedeno med posameznimi seštevanji s pomikom vsebine v desno, postopek krmili negirana osnovna ura U2- Po dvanajstih seštevanjih je zaključen izračun trenutne izhodne vrednosti y(k). Na izhodu seštevalnika se pojavi 16-bitni Slika 3.1; Časovni poteki signalov iz krmilnega vezja Tabela 3.1: Koeficienti nizkoprepustnega in visokopre-pustnega digitalnega sita h(n) nizkoprepustno sito fP=0,1 fv; fz=0,2 fv visoko prepustno sito fz=0,2 fvi fp=0,3 fv h(0) = h(14) 0.132637E-01 0.264990E-01 h(1)=h(13) -0.227501 E-01 -0.133865E-05 h(2) = h(12) -0.447545E-01 -0.440890E-01 h(3) = h(11) -0.380495E-01 -0.202039E-06 h(4) = h(10) 0.271117E-01 0.934248E-01 h(5) = h(9) 0.141917E+00 -0.518952E-06 h(6) = h(8) 0.254379E+00 -0.313903E+00 h(7) 0.301295E+00 0.500000E+00 podatek. Ker smo na izhodu uporabili 12-bitni pretvornik digitalnega v analogni signal MAX 507 smo uporabili le dvanajst najbolj uteženih bitov bi5, bi4,...b4. Vezalno shemo celotne strukture prikazuje slika 3.2. Pri izbrani frekvenci urinih impulzov smo ob dvanajst bitni kvantizaciji vhodnega signala dosegli pri izvedeni strukturi frekvenco vzorčenja fv = 143 kHz. Z uporabo vezij večje stopnje integracije ali s programabilnimi polji logičnih vrat bi zaradi zmanjšanja vplivov ožičevanja elementov, zlahka to vrednost frekvence še povečali. Izvedeno univerzalno strukturo digitalnega sita za 15 koeficientov v modificirani obliki porazdeljene aritmetike smo uporabili za prikaz vezja z nizkoprepustno in visok-oprepustno frekvenčno karakteristiko. Za spremembo frekvenčne odvisnosti je potrebno le zamenjati EPROM pomnilnike z vpisanimi delnimi vsotami koeficientov. Digitalni siti smo načrtali s programskim paketom DF-PAK /4/. Uporabljene koeficiente obeh sit, ki so osnova za izračun delnih vsot koeficientov, podaja tabela 3.1. Na slikah 3.3 in 3.4 so prikazane teoretične, simulacijske in izmerjene frekvenčne karakteristike obeh sit. Pri simu-lacijskih rezultatih smo zajeli vse vplive kvantizacije /5/. Na vhodu in izhodu sita smo dodali A/D in D/A pretvornik, meritve frekvenčnih karakteristik pa smo opravili z generatorjem spremenljive frekvence na vhodu in meritvijo temenske vrednosti izhodnega signala z merilnikom vršne vrednosti. Pri izbranem nizko-prepustnem situ načrtanem po optimalnem postopku s prepustno frekvenco fp = 0,1 fv in zaporno frekvenco fz = 0,2 fv smo dosegli slabljenje A= - 28 dB. Pri realizaciji visokoprepustnega sita z zaporno frekvenco fz= 0,2 fv in prepustno frekvenco fp = 0,3 fv pa smo dosegli slabljenje A= - 30 dB. S primerjavo teoretičnih, simu-lacijskih in izmerjenih frekvenčnih odzivov vidimo, da smo dosegli dokaj dobro ujemanje frekvenčnih potekov, odstopanje vrednosti slabljenja v zapornem frekvenčnem področju pa je bilo v obeh primerih manjše od približno AA = 2 dB. 110 K. Korošec, A. Vesenjak, B. Jarc, M, Šolar, R. Babič: Izvedba nekurzivnega digitalnega sita s standardnimi Informacije MIDEM 26(1996)2, str. 107-112 111 K. Korošec, A. Vesenjak, B. Jarc, M. Šolar, R. Babič; Informacije MIDEM 26(1996)2, str. 107-112_Izvedba nekurzivnega digitalnega sita s standardnimi... H(f/fv) i 0.8 0.6 0.4 0.2 ^ % i ♦ meritev \\ teoretična kar. \\ — simulncijska kar. \\ \\ \ 0 0.1 0.2 0.3 0.4 0.5 f/IV Slika 3.3: Frekvenčna karakteristika nizkoprepustnega sita H((7fv) 1 0.8 0.6 0.4 0.2 / / ♦ meritev • teoretična knr. ! / - .simulncijska kar. - 0.1 0.2 0.3 i/IS 0.4 0.5 Slika 3.4: Frekvenčna karakteristika visokoprepust-nega sita 4. Zaključek Prikazali smo izvedbo univerzalne strukture nerekur-zivnega digitalnega sita s standardnimi integriranimi komponentami za realizacijo sit s poljubno frekvenčno odvisnostjo. Pri tem smo uporabili modificirano obliko porazdeljene aritmetike, s katero smo pri zmanjšani aparaturni kompleksnosti strukture dosegli povečanje dinamike izhodnega signala od 6 do 8 dB. Povečanje dinamike je odvisno od števila koeficientov digitalnega sita in od vrste frekvenčne odvisnosti. Pri isti kompleksnosti aritmetične enote dosežemo tako praktično dva krat manjši vpliv kvantizacije aritmetične enote na izhodni signal. S simulacijskimi frekvenčnimi poteki nizko prepustnega sita s 15 koeficienti smo ilustrirali povečanje dinamike izhodnega signala in s tem tudi ojačenja digitalnega sita v prepustnem frekvenčnem področju. dosegli frekvenco vzorčenja fv = 143 kHz. Frekvenca vzorčenja je sorazmerna številu bitov za zapis vhodnega signala. Njeno povečanje bi pri izbranih komponentah dosegli že s skrbnejšo izdelavo ožičenja vezja, s katerim bi zmanjšali vplive parazitnih kapacitivnosti. Praktično pa bomo dosegli boljše rezultate z uporabo programabilnih polj logičnih vrat saj predstavlja opisana struktura digitalnega sita le osnovo za izvedbo digitalnega sita z integiranimi komponentami večje stopnje integracije (FPGA elementi). V rezultatih smo prikazali še dve frekvenčni odvisnosti. Pri realizaciji nizko prepustnega sita s prepustno frekvenco fp = 0,2 fv in zaporno frekvenenco fz = 0,3 fv smo dosegli slabljenje A = - 28 dB. S preprogramira-njem EPROM pomnilnika z vpisanimi delnimi vsotami koeficientov pa smo realizirali še visoko prepustno sito z zaporno frekvenco fz = 0,2 fv in prepustno frekvenco fp = 0,3 fv ter dosegli slabljenje A = - 30 dB. V obeh primerih smo dosegli v splošnem dobro ujemanje frekvenčnih potekov med teoretičnimi, simulacijskimi in izmerjenimi odzivi, zaradi več vplivov pa je le prišlo do odstopanja vrednosti slabljenja v zapornem frekvenčnem področju AA = 2 dB. 5. Literatura /1/ B. Liu, A. Peled. A New Hardware Realization of Digital Filters. IEEE Trans, on A. S. S. P., Vol. ASSP 22, pp. 456-462, Dec 1974 /2/ Stenley A. White. Applications of distributed arithmetic to digital signal processing: A tutorial review. IEEE ASSP Magazine, pages 4-19, Jul. 1989 /3/ M. Brumec, Izvedba digitalnega sita 14 stopnje v porazdeljeni aritmetiki, diplomsko delo, TF Maribor, ERI, Maribor 1993 /4/ F.J. Taylor, T. Stouraitis, Digital Filter Design Software for IBM PC, Marcel Dekker Inc., New York, 1987 /5/ R. Babič, M. Solar, B. Stiglic, Analiza vplivov kvantizacije pri izvedbi digitalnih sit v porazdeljeni aritmetiki. Zbornik prve elektrotehniške in računalniške konference ERK'92, strani 13-16, Portorož, Slovenija, 1992. Kari Korošec, dipl. inž., doc. dr. Rudolf Babič, dipl. inž., doc. dr. Mitja Solar, dipl. inž., Bojan Jarc, dipl. inž., Anton Vesenjak, inž., Univerza v Mariboru, Fakulteta za elektrotehniko, računalništvo in informatiko 2000 Maribor, Smetanova 17 tel.: +386 62 25 461 fax: +386 62 211 178 Z uporabo nezahtevnih standardih integriranih komponent smo pri 12-bitni kvantizaciji vhodnega signala Prispelo (Arrived): 28.5.1996 Sprejeto (Accepted): 18.6.1996 112 Informacije MIDEM 26(1996)2, Ljubljana UPORABA PLAZME V ELEKTRONIKI APPLICATION OF PLASMA IN ELECTRONICS PLASMA PROCESSES PART II : APPLICATIONS IN ELECTRONICS I. Sorli*, W. Petasch, B. Kegel, H. Schmid, G. Liebel, W. Ries *MIKROIKS d.o.o., Ljubljana, Slovenia Technics Plasma GmbH, Kirchheim, Germany 1.0 INTRODUCTION Plasma is obtained by producing a discharge in gases or gas mixtures under vacuum through the application of high frequency alternating voltage. The gas in the chamber is brought to an excited (ionized) state. As well, active radicals and UV radiation are released. Electrons and UV light, resulting from the recombination processes are essential for maintaining the plasma. These components are the actual energy carriers, which are ultimately responsible for the production of chemically active radicals. This highly active process gas is capable of reacting with the surface of the material to be treated even at low temperatures. During the process fresh gas is continuously fed into the chamber. The reaction products are evacuated by the vacuum pump. Plasma excitation via microwaves (2.45 GHz) has proved especially effective, since the efficiency of the gas discharge increases considerably with increasing frequency but still requiring very low electrical power. This results in strong, intensive ionization and production of radicals and thus a more cost effective process. Today's microwave excitation technology makes it possible to use the low pressure plasma processes economically in industrial mass production in either continuous or batch systems using large process chambers. Small bulk parts, as well as large components can be effectively cleaned and activated. Very important issue of low pressure plasma is its penetrability. The gas enters the smallest crevices, making it possible to process three - dimensional parts with complex geometries. Another very important fact is that plasma processes are environmentally friendly and as such are alternatives to CFC cleaning processes. Thus, main advantages of low-pressure plasma technology are: • dry process • energy saving through low power consumption • inexpensive supplies, cost - effective gases • switch - off chemistry: the process stops immediately when the power is turned off, no disposal of waste • cleaner, safer workplace, simple operation • high penetration power into narrow spaces - an advantage in degreasing or activating parts with complex shapes 8 constant process conditions, good reproducibility • meets or exceeds air emission standards • parts are absolutely dry after treatment 2.0 SUPER FINE CLEANING WITH LIQUID PHASE PRECLEANING AND SUBSEQUENT PLASMA TREATMENT 2.1 GENERAL Cleaning and degreasing is a widely used industrial procedure. Stricter environmental requirements and recently discovered facts regarding the effect of previously used cleaning and drying agents on atmosphere chemistry make a radical review of the conventional cleaning technologies necessary. The plasma process is especially well suited for removing organic contaminants and residual films (such as greases, oils, waxes or solvents) when the films are very thin and super clean surfaces are required. A very important characteristic of low pressure plasma is its penetrating power. The gas penetrates into small pores that are difficultt or impossible for liquids to access. Thus, even parts with complex shapes can be easily processed with plasma (cutouts, small radii, bore holes, slots). The penetration capability allows the plasma to reach even cracks with micrometric dimensions. Since most cleaning problems are concerned with the removal of organic substances from the surface, oxygen is used in most cases as the reactive gas. Oxygen is" so one of the most important process gases in the treatment of almost all types of materials, In the case of polymer activation which is discussed in Part III, best results are achieved with oxygen for the most common polymers. 113 Informacije MIDEM 26(1996)2, Ljubljana Mechanical abrasion through particle bombardment (sputtering effect) has a subordinate role in the type of equipment used for this purpose. The effect of oxygen is enhanced by the admixture of small amounts (five to ten percent) of argon or helium. Slightly preheating the parts to between 80 and 100°C, whenever the material to be treated allows, is also helpful. Since the process gas immediately returns to its original state after the gas discharge has been shut off, possible residues capable of causing long term corrosion represent no problem. Plasma cleaning means that organic impurities are removed by chemically transforming them in volatile products CO, CO2 and H2O, figure 1. In short, the process is a dry one, no submerging of the parts into a liquid takes place and energy cost for a separate drying phase can be saved. Thus, when oxygen is used the cleaning effect consists of oxidative conversion of the organic contaminants. 2.2 ENVIRONMENTAL PLUSES TO PLASMA The plasma process does not use acids, lyes, solvents or liquid halogenated hydrocarbons. The reaction with the process gas takes place in hermetically sealed chambers. For operation, this means no disposal problems and clean workplace conditions, since no poisonous gases or vapors are produced and no hazardous liquids are handled. A plasma system can be easily operated by semi - skilled personnel. The process is simple and does not require special education. Waste gas emissions satisfy and exceed TA clean air standards. There is no fire or explosion hazard. 2.3 DOWNSIDE TO PLASMA Critical examination discloses the following limitations of this process: 1. Intensive contamination is always irregularly distributed. Thus, the treatment period is long. 2. Photochemical reactions within the surface layer may lead to a crosslinking reaction of the layer material which results in a reduction in the etching rate. 3. Inorganic components (machininng chips, debris and other particles) cannot be removed. They tend to form non-volatile oxides or salts remaining on the surfaces. A solution is provided by a precleaning step, in which the inorganic components are removed and the processing agent amounts are reduced. This leaves a relatively uniform residual layer, which can be removed by means of plasma in a few minutes. 2.4 PRECLEANING WITH DIFFERENT MEDIA As discussed above, precleaning may be required depending on the initial conditions. The medium used for preclean, the material to be cleaned and the chemical composition of the contamination have an influence on the results of the subsequent plasma cleaning. In the following example, solvent in the form of glycol ether and water, are used as cleaning agents. The part must be dry prior to plasma fine cleaning. Plant design with the respective peripherals (distilling unit, water treatment and purification) are shown in figure 2. Examples and Results of Cleaning Procedures Application: Degreasing of metallic parts Case Description: Cleaning of brass, aluminum and steel parts prior to assembly Cleaning requirements: Minimum grease level (low residue carbon) Former cleaning process: Aqueous and CFC Alternative solution: Aqueous precleaning and final plasma cleaning. Results showing residual carbon (mg/cm2) left after cleaning are presented in table 1. TABLE 1 Material Aqueous Cleaning Aqueous + Tri Aqueous + Plasma X12CrNiS188 29.3 6.8 2.7 CuZn39Pb3 20.7 5.3 3.5 AlCuMgPb 19 7.5 3.7 114 Informacije MIDEM 26(1996)2, Ljubljana Additionally, the good cleaning results are confirmed by quantitative measurements of the adhesion of the cleaned parts, table 2. Test setup: copper tube with 3 mm diameter, glued into an aluminum block, both suitably cleaned; glue used: Gupalon 30, setting for four hours at 120°C TABLE 2 Parts: Injection molded and stamped parts Material: Non-ferrous metals, silver, plastic Contamination: stripping agents, stamping and bending oils, flux Cleaning agent: Glycol ether (Zestron LP) Water: demineralized, continuously recirculating Cleaning requirements: visually spotless, greaseless, low electrical resistance, imprintability Capacity/cycle time: process - dependent Batches: 40 x 50 x 20 cm perforated baskets Floorspace required: approx. 40 m2 Installed power: 65 kW Figure 3: Combination wet/plasma cleaning of metal and plastic parts for the automotive industry Cleaning method Tear Strength (N) Aqueous 531 Aqueous + Tri 642 Aqueous + Plasma 785 Combination solvent/water and plasma treatment STATION 1 Cleaning STATION 2 Rinse _MW STATION 4 \ PLASMA Flo'f eguk to _MW STATION 4| PLASMA A xX Gas: Gas; Flo' v ■eguk ñ Gasi Ú Gas! Ion exchange colun i 1st stage : Cleaning - dipping with u!trasound( glycolether) 2nd stage : Cleaning - rinsing ( DI water) 3rd stage : Drying in a forced-air oven, at 70 deg.C 4th stage : Oxygen plasma cleaning Recirculation of the cleaning agent ( glycol ether ) through a particle filter, purification of rinsing water via photo-oxidation of the entrained glycol ether with hydrogen peroxide added, continuous monitoring via oxidation - reduction potential and conductivity measurement after ion exchange, temperature control in furnace, totally enclosed with central exhaust system, automatic handling, automatic process control. Other process versions are possible ( e.g. plasma only, aqueous precleaning and plasma, precleaning only, etc.) Figure 2: Combination of solvent/water and plasma treatment 115 Informacije MIDEM 26(1996)2, Ljubljana 3.0 APPLICATION OF PLASMA TECHNOLOGY IN ELECTRONIC PACKAGING Prohibition of CFC's and, in addition increasing surface quality requirements make new solutions for electronic board cleaning necessary. As well, plasma applications of great interest are metal finishing prior to fluxless soldering, device cleaning for reliable wire bonding and surface treatment for adhesive bonding. 3.1 WIRE BONDING ON CRITICAL SURFACES The yield of plasma treatment is demonstrated by comparative experiments with and without plasma pre-treat-ment. Parameters of practical significance such as the strength of wire bonds are used as evaluation criteria. In particular, for bonding wires to printed circuit board substrates, bonding to the connection metal platings is a frequent source of problems. In addition to the inho-mogenity of the circuit board's matrix structure, the quality of the bond pad surface is a frequent cause of failures. Preceding process steps such as soldering and adhesive bonding cause organic contamination, which may precipitate on the bond pads, e.g. in the chip - on - board (COB) technology. With the introduction of a plasma process, prior to bonding wet chemical cleaning can be completely omitted. Wire bonds are characterized using strength tests such as the pull test according to MIL-STD 883D. Then a metallographic ground cross section of a wire bond is prepared in order to make the joint area visible. Mechanical Strength Test: Pull Test The destructive pull test (MIL - STD 883D, Method 2011) is a standard test for determining the strength and reliability of wire bonds. The wire bridge is pulled with a hook under the effect of a continuously increasing pulling force until rupture. The maximum force is measured, making sure that the individual measurements are reproducible. When evaluating the pull test, one should basically distinguish between different rupture characteristics or failure modes. We distinguish between the failure modes of bond detachment from metallization and wire rupture in the deformation zone. The different qualities of substrate surfaces can affect bond detachment, but also influence the introduction of ultrasonic energy in the bond, which leads to different kinds of wire deformation at the bond with the resulting effects on wire strength in the heel area. In figure 4 possible failure modes in the pull test are shown. A, B: Failure in bond at interface between wire and metallization (bond detachment), 1st or 2nd bond E, F: Wire rupture at the bond heel, 1 st and 2nd bond H: Bond failure: no bond was made Well defined contamination of the bond surfaces Laboratory tests were performed with well defined contamination on bond surfaces. Surface: Chemically deposited Ni/Au on FR4 substrate. Bonding process: Ultrasonic wedge bonding with 25 |j.m AISÎ1 wire. In order to investigate the effectiveness of a plasma cleaning process, a contamination was applied which represents a real - life contamination actually occuring during electronic manufacturing. The surface contamination and its effect on bondability was tested using a reflow soldering process. The soldering paste on the substrate was melted. The flux component constituted the bond pad contamination. Prior to bonding, the substrates were subjected to plasma cleaning. Then over 100 bonded wire edges were placed on each substrate and tested with the destructive pull test. Plasma cleaning Two different plasma processes were used with the samples contaminated with the flux residue of the reflow soldering. The O2 plasma has an oxidizing effect on the organic deposits; the Purigon (trade name of Linde AG) plasma also has a reducing effect on the metal oxides due to the H2 component. Results and discussion Table 3 shows the results of the individual test series. The average force for all failure modes (bond strength) is given with the standard deviation a, as well as the distribution by individual failure modes, including the corresponding average force. Test series a) shows the succesful optimization in relation to the initial conditions. Bond detachment is minimized and the average force of 7.2 cN is significantly larger than the suggested value of 4.0 cN. With contaminated surfaces (series b)) the average rupture force 116 Informacije MIDEM 26(1996)2, Ljubljana TABLE 3 Test series F cN G cN Distribution (%) Average force (cN) for different failure modes Remarks A B E F H a) initial state 7.2 0.8 8.0 6.1 1.0 6.1 66.0 7.2 25.0 7.6 0 Bond optimization: A, B min. b) contaminated 5.5 2.9 24.8 4.8 6.0 6.8 35.0 7.5 15.4 7.5 17.1 Bonding results significantly inferior d1) O2 plasma 6.9 1.8 0.7 7.0 2.2 6.3 81.0 7.3 10.9 7.4 5.1 Bonding results almost as good as a). Failure modes A, B drastically reduced d2) Purigon plasma 6.2 2.2 19.6 5.7 6.5 5.2 44.2 7.3 21.0 7.1 8.7 Improvement but clearly inferior to O2 plasma c) contaminated - with parameter optimization 6.0 2.0 38.5 4.5 0.7 9,2 28.1 7.3 29.6 7.2 3,0 No substantial improvement. Effect of contamination cannot be eliminated by parameter optimization e) O2 plasma - with parameter optimization 7.6 0.8 0 5.5 6.8 3.7 8.1 90.8 7.7 0 Slight improvement compared to d1 drops down to 5.5 cN, while the standard deviation increases dramatically to 2.9 cN. Bond detachment increases drastically, and bonding errors are also observed. This negative effect of the contamination can be eliminated by plasma cleaning in an O2 plasma. In particular, bond detachment types A and B disappear nearly completely. Purigon plasma is not as effective as O2 plasma, because the organic deposits only respond to oxidizing action. The main component of Purigon, hydrogen, cannot be active, but it reduces the partial pressure of oxygen, For test series c) we attempted to compensate for surface contamination by optimizing the bonding parameters. Since bond detachment A cannot be reduced below the 38.5% limits, the procedure was not successful. Thus the effect of contamination cannot be eliminated by optimization only. In the case of O2 plasma cleaning, however, the subsequent optimization results in a slight improvement of the bond strength (series e)). 3.2 ENHANCE HYBRID RELIABILITY THROUGH PLASMA CLEANING Poor wire bonding is the primary cause of failure in hybrid integrated circuits. To create a successful wire bond, strong intermetallic contact between pads on a hybrid and the bond wires must be achieved. This is possible only when the two surfaces are brought into close contact. Once forced together, interatomic forces create a bond. Wire or pad contamination, however, hinders this process. To remove organic and inorganic contaminants and form strong, low failure bonds, hybrid components should be plasma cleaned prior to wire bonding. To plasma clean hybrids, plasma of inert gases or a mixture of inert/reactive gases - such as Ar, Ar/02, Ar/N2 - chemically removes molecular layers of contamination. Argon mechanically dislodges contaminants, while Ar/02 is a mechanical dislodge and chemical oxidation process. Contaminants can form on hybrids and ICs during assembly. For example, organic contaminants due to poor rinsing after wet chemical photoresist, strip or etch appear on semiconductor die. Hybrids assembled with an epoxy die attach may suffer from resin bleed. When organic resin runs over the adjacent surface of the ceramic and conductors through capillary action, wire bond strength can be reduced. Another contaminant is outgasses produced during epoxy cure. Although modern epoxies are classified as solventless, they contain low molecular weight diluents, which control epoxy viscosity and rheology. By formulation, diluents bind chemically with the epoxy as it polymerizes. Nonetheless, even with an exhaust system in operation, some diluents can be outgassed during epoxy cure and can build up on the hybrid and impede bonding. The dicing process for wafer segmentation is another source of contaminants. Water used to cool the dicing diamond saw blade may have impurities, catalyze contaminants from previous process steps, or result in bond pad corrosion. In addition, the aluminum metallization from which most semiconductor bond pads are made 117 Informacije MIDEM 26(1996)2, Ljubljana readily oxidizes, if this oxidation layer is too thick, bond pad reliability can suffer. The hybrid assembly environment also inroduces contamination from hand oils used by machine operators, oil fumes and particulates in the air, and more. Containers for hybrid substrates, waffle packs for ICs, and other storage materials can be contamination sources. Careful handling and processing during hybrid and IC assembly can only minimize contamination; therefore, assembly techniques must incorporate a cleaning step in the process flow. Removing epoxy resin bleed is a good method to test plasma cleaning effectiveness - if resin bleed can be eliminated, any nonvisible contaminants present also can be removed. Innitially only Ar gas was used to clean the contamination, but long processing times required were unsuitable for a production environment. A 98% Ar / 2% O2 mixture provided a sufficiently fast process time, although there is always fear that oxygen might discolor the silver filled die attach epoxy. In figure 5a) and 5b) a comparison between AES spectra of an uncleaned and , as; 'lig'-Ei sr«6e.a9žv.8Q9 i)ftT=e.e8 ier9s s? 3is-sf v r i : ti^ i Rr» ' ! I kf i 1 ; . ' • -f—c-,------ • KINETIC EHLRCY, F¥ Figure 5: AES - Scanning Auger microanalysis of a bond pad on a component a) before plasma cleaning b) after plasma cleaning plasma cleaned bond pad (plasma condition: 98%Ar/2%02, 500 W, 15 min, 0.75 torr) demonstrates a sharp reduction in the metallization's carbon peak after plasma clean which also means a reduction in organic contaminant level. Wire bond test results Test results shown in table 4 clearly show the efficacy of plasma cleaning. Samples thermosonically bonded on a semiautomatic gold bonder with 6 to 8 g tensile strength and a 3 to 5 percent elongation wire boasted stronger bonds than did hybrids that had not been plasma cleaned. Bond pull strength increased 13 to 25 % with a 13 to 17 % reduction in the standard deviation. TABLE 4 DESTRUCT PULL TEST SAMPLES SAMPLE I Before Cleaning After Plasma Cleaning X = 5.3 g X = 6.65 g S = 1.89 g S = 1.57 g Failure Mechanism Failure Mechanism 8 - Bond lifts 7 - Neck downs 10 - Neck downs 5 - Wire breaks SAMPLE II Before Cleaning After Plasma Cleaning X = 6.78 g X = 6.65 g S = 1.31 g S = 1.57 g Failure Mechanism Failure Mechanism 11 - Bond lifts 23 - Neck downs 1 - Wire break 31 - Neck downs 5 - Wire breaks NON-DESTRUCTIVE WIRE BOND PULL TEST Product 1 Product 2 Uncleaned Plasma Cleaned Total bonds tested - 28.050 ■Total bonds tested -18.305 Total bonds falied - 317 Total bonds falied -12 % Failure 1.13% % Failure 0.066% Plasma Cleaned Total bonds tested -11.826 Total bonds falied -15 % Failure 0.13% A more significant result was a shift in the failure mechanism of the bond when pulled to destruction - failures shifted to neck down and wire breaks versus bond lifts. The cleaner components permitted a reduction in ultrasonic power levels on the bonder and less rework of parts coming off the bonder due to missed bonds. 118 Informacije MIDEM 26(1996)2, Ljubljana Conclusion Plasma cleaning of hybrids and ICs can increase bond pull forces and reduce standard deviations of destructive pull tests, decrease ultrasonic power levels, and wide the effective bond window. In addition, life tests and non-destructive bond pull tests demonstrate an improvement in long - term reliability. 4.0 APPLICATION OF PLASMA TECHNOLOGY IN PCB PRODUCTION 4.1 PLASMA DESMEARING AND ETCHBACK OF MULTILAYER PRINTED CIRCUIT BOARDS Technics Plasma offers manufacturers of multilayer printed circuit boards a reliable, clean, easy - to - use production tool for removing drill smear from either flex or rigid boards. 110 mg 100 Ç0 Coupon Weight Loss Distribution Weight [oss of epoxy laminate coupons after plasma treatment depending on position in process chamber Average value \ » 4 720°C 110 100 90 80 70 Position in electrode cage Temperature Distribution Process c.nd The degree and uniformity of desmearing and etch-back can be simulated by measuring trie weight loss of toe uncovered epoxy laminates. Tte graph shows the variations in the weight loss in (he individual chamber locations. These values can be used to estimate the uniformity of etch-back on the inside of the hole. Mean values of 5 test points per circuit board. Ihe temperature is measured with the he/p of measuring bands. It rises 0.5-1.5'C/min during the process according to the generator power set. The temperature constantly shown by the process controller is measured by a resistance thermometer. Process Chamber Temperature Figure 6: a) etch uniformity versus laminate position in the chamber b) process temperature versus laminate position in the chamber heat cycle only) gas stabilization time low temp, set point' (80°C for example) high temp, set point (110°C for example) RF-power TVpical sequence PC-5000 Figure 7: Typical process sequence for drill smear removal The drill smear found inside drilled holes can be safely removed in an Oxygen - Freon plasma. When desired the system can also be used to perform controlled etch back in multilayer polyimide and epoxy glass boards. The uniformity across one board, from board to board and batch to batch is usurpassed due to patented Planartube electrodes which are driven by low frequency RF generators. Typical performance results are shown in figure 6 where etch rate and laminate temperature uniformity are displayed. As well, in figure 7 typical process sequence is displayed. Figure 8: Technics Plasma desmear system 4.2 FLUXFREE SOLDERING WITH PLASMA PRETREATMENT Introduction During the wave soldering process the active and passive components are connected with the circuit carrier of the PCB by a collective soldering procedure and thus made an operative flat pack group. The jointning partners are in general supplied in an unsolderable status, covering layers impede the wetting procedures in the solder melting, figure 9. organic layers SnPb Cu6Sr>5 Cu3Sn Cu covering layers impair the soldering Figure 9: Covering layers impair the soldering 119 Informacije MIDEM 26(1996)2, Ljubljana These layers are organic, they cover as part of industry atmospheres all surfaces and have the even more impeding effect the longer the molecule chains of the hydrocarbons are. Further layers are metal oxides originated by reactions with oxygen. Before the solder process can be started a preparation is necessary to ensure the jointing ability. Usually flux is used. Therefore everybody associates wave soldering with flux. A procedure without flux, however, shows differently. The active ingredient in modern, low solid content flux is an organic acid. Heat supply effects the transformation from metal oxide to metal complexes respectively to metal. Due to the unknown supply status of components and PCBs (quality and thickness of the covering layers are usually not known) one has to work with a surplus of flux so that the chemical process may be incomplete and not wanted residues remain on the flat pack groups. These residues may effect the long term reliability by decreasing the surface resistance respectively by migration. In principle these processes apply to all flux variations no matter whether they are conventional colophony formations, low solid content types or formic acid in protective atmosphere solder systems. The solder process is executed under the protective atmosphere wherease the concentration of the remaining oxygen may be higher than in the conventional protective atmosphere systems. The gas consumption runs at 14 - 15 m3/h. The flat pack groups leave the system completely clean and free of any residues. Perspective: Fluxfree reflow soldering under plasma 1. Printed circuit board 2. Assembly 3. Plasma pre-treatment 4. Reflow soldering under protective atmosphere ^«pprex. 100 um 3nPö (StPAO, OPT1PAD) component ~J idhMlv* cnnipmioiit component Principle of fluxfree soldering SnO, SnC>2, PbO form a porous monolith with a very high melting point (1000°C). The surface failure spots are bare after the plasma treatment, since oxygen plasma ashes away all organic contaminants as well as activates the surface increasing its wettability. Now, an appropriate vaporized process material (purified water) is applied to the surface. Condensate precipitates in the fine openings of the oxide layer. When the SnPb layer prepared this way contacts the solder wave the temperature rises fast (500 K/s) and leads to the quick volume expansion of the precipitated condensate amounts. The volume increase (1000 : 1) lifts the oxide layers, fragments are washed away by the moving solder wave and the solder connection Is formed by the copper - tin diffusion at the oxide free areas, figure 10. Solder connections generated by such a new method have of course to undergo a line of tests in order to prove their quality and long term reliability. All visual, structural, electrical, physical and burn in tests turned out to be successful. There are no restriction in the range of components that can be processed. All available component shapes are suitable for the plasma method. System Concept A piling device collects pallets (5 off per batch) from the feeder and passes them on to the receptacle drawer of the plasma chamber, figure 11. The chamber closes and evacuates. After completion of the plasma treatment the pallets are extended, separated again, and by means of band conveyor they run into the preheating, used only for tempering. Then the condensation process is effected and the conveyance to the dual wave. a) 1. Initial status • organrc layers SnPö H^iffir-m-5l3lli2aiion 2. Plasma effect 3. Condensation FR4 ocsar ramami 4. Soldering a) blast-off ol oxides b) lluxfree T^iiir BE - SnPb - melaüüaüon -Cu- llai pack group b) Figure 10: a) Principle of fluxfree soldering b) Detailed mechanism behind fluxfree soldering 120 Informacije MIDEM 26(1996)2, Ljubljana c dtrtc ) lir t? t !t > /2/ R.Buck, ENHANCE HYBRID RELIABILITY THROUGH PLASMA CLEANING, Hybrid Circuit Technology, December 1988 13/ H.Schmid, Super Fine Cleaning with Liquid Phase Precleaning and Subsequent Plasma Treatment, PC, October 1995 /4/ Technics Plasma GmbH, Application Reports Figure 11: Fluxfree soldering system concept For more information about Technics Plasma systems and their applications, please call: 5.0 LITERATURE /1/ S.Wolf, R.N.Tauber, SILICON PROCESSING FOR THE VLSI ERA, Volumel : Process Technology, Lattice Press 1986, ISBN 0-961672-3-7 MIKROIKS d.o.o., Mr. Iztok Sorli Dunajska 5, 1000 Ljubljana, Slovenia tel. +386 (0)61 312 898 fax.+386 (0)61 319 170 PREDSTAVLJAMO PODJETJE Z NASLOVNICE REPRESENT OF COMPANY FROM FRONT PAGE ISKRA Kondenzatorji Industrija kondenzatorjev in opreme d.d. Iskra Kondenzatorji d.d. praznuje v letošnjem letu 45 let delovanja. Skromni začetki proizvodnje navitih kondenzatorjev na sedanji lokaciji v letu 1951 so bili nadaljevanje razvojnega dela na institutu IEV v Ljubljani, ki je bil ustanovljen z namenom, da razvija elektronske komponente za nastajajočo slovensko elektronsko industrijo. V začetku je bila proizvodnja prilagojena možnostim in zahtevam domače industrije. Kmalu se je pojavila zahteva za širjenje programa proizvodov za nove aplikacije. Že po 10-tih letih delovanja je proizvodnja obvladovala širok program kondenzatorjev za različne namene uporabe, tako v elektroniki, kakor tudi na področju elektroenergetike. Proizvodnja je bila delno velikoserijska in pa maloserijska in prilagojena posameznim kupcem. V 70-tih letih se je tovarna začela intenzivno usmerjati v prodajo preko meja domovine. Izvoz je postal izziv, še bolj pa potreba za nadaljnji razvoj. Prodaja, predvsem v industrijsko visoko razvite države, je zahtevala absolutno izpolnjevanje zahtev kakovosti in solidnega poslovanja. Raziskave in aplikativni razvoj je bilo potrebno usmeriti v produkte, ki so obetali ekonomično proizvodnjo in konkurenčnost na trgu. Razvoj je sledil trendom v svetu s široko paleto proizvodov za uporabo na področju elektronskih naprav, naprav elektroenergetike in odprave radiofrekvenčnih motenj. Večina proizvodov mora ustrezati mednarodnim in celi vrsti nacionalnim predpisom in standardom tako v Evropi kakor v ZDA in Kanadi. Ustreznost tem predpisom in standardom mora biti potrjena z atesti, ki so obvezni za plasma proizvodov na tehnično in komercialno izredno zahtevnem tržišču. Zahteve na področju razvoja, proizvodnje, trženja in na področju ostalih segmentov poslovanja so privedle do spoznanj, daje možno zagotavljati zahtevano kakovost le s certificiranim sistemom ISO 9001, kar je bilo tudi realizirano v letu 1995. Danes firma Iskra Kondenzatorji samostojno obvladuje vse poslovne aktivnosti. Preko 90% proizvodov proda na tujih trgih. Kupci in uporabniki proizvodov Iskre Kondenzatorji so svetovno znani proizvajalci elektronskih naprav, računalnikov, bele tehnike, električnih orodij, elektroenergetskih sistemov, električnih strojev itd. Tovarna predstavlja pomembnega proizvajalca kondenzatorjev v svetovnen merilu, tako po količini, kakor po kvaliteti. Za uveljavitev v svetovnem merilu je bilo potrebno veliko strokovnega dela, ki je vseskozi potekalo v sklopu tovarne ob sodelovanju z domačimi strokovnjaki in znanstveno-raziskovalnimi institucijami. V razvoj proizvodov, ki obsega ob bazičnem predvsem aplikativni razvoj, je bilo vloženo ogromno interdisciplinarnega znanja iz vseh področij naravoslovnih ved, še posebej, ker firma ob osnovnem programu obvladuje tehnologijo proizvodnje praktično vseh sestavnih delov. 121 Informacije MIDEM 26(1996)2, Ljubljana Ob osnovnem programu proizvodnje vseh vrst kondenzatorjev, katerih glavni sestavni del predstavlja dielektrik iz termoplastičnih materialov (poliester, polipropilen) in specialnega kondenzatorskega papirja, pa je firma preko celotnega obdobja delovanja razvijala proizvodno opremo za proizvodnjo kondenzatorjev in sestavnih delov. V zadnjem obdobju je na tem področju dosežen poseben viden uspeh, ko je večina zahtevnih strojev in naprav v proizvodnji izdelek lastnih strokovnjakov. Tak koncept lastnega razvoja izdelkov in proizvodnje opreme daje tovarni popolno tehnološko neodvisnost brez tujih licenc in omejitev na trgu. Tovarna iskra Kondenzatorji d.d. Semič ob svoji 45. letnici prodaja proizvode v 40. državah, ustvari letno 60 mio DEM prihodka in ob proizvodnji tudi večino sestavnih delov, nudi zaposlitev preko 1300 delavcem. Načrti za bodočnost so optimistični in predvidevajo nadaljnji razvoj proizvodnje in izdelkov, prilagojenih aplikacijam v najzahtevnejših napravah na vseh področjih elektronske in elektro industrije. Proizvodni program obsega naslednje proizvode: - kondenzatorji za elektroniko - enosmerni-metalizi-rani, folijski poliester, polipropilen, - kondenzatorji za visoko impulzne obremenitve, - kondenzatorji za elektroniko - miniaturni, - kondenzatorji in filtri za odpravo radiofrekvenčnih motenj razreda X1, X1Y, - kondenzatorji za odpravo radiofrekvenčnih motenj -samoozdravljivi - X1 in X2, - kondenzatorji za pogon elektromotorjev, - kondenzatorji za kompenzacijo fluorescenčnih svetilk, - kondenzatorji za kompenzacijo jalove energije na nizki in srednji napetosti, - avtomatske naprave za kompenzacijo jalove energije, - specialni kondenzatorji za elektroenergetiko in druga področja uporabe, - stroji in merilna oprema za proizvodnjo kondenzatorjev, - manipulatorji in naprave za avtomatizacijo industrijskih procesov, - orodja za predelavo plastike in preoblikovanje kovin, - specialna oprema, orodja in stroji po zahtevah naročnika. ISKRA KONDENZATORJI d. d. Vrtača 1, 8333 Semič Slovenija tel. 068/67-709 fax. 068/67-110 MIDEM IN NJEGOVI ČLANI MIDEM SOCIETY AND ITS MEMBERS OBČNI ZBOR DRUŠTVA MIDEM Spoštovani člani društva MIDEM! Kot prilogo zapisniku občnega zbora društva MIDEM podajamo poročilo dosedanjega predsednika dr. R. Ro-čaka o delu v preteklem mandatnem obdobju, oz. okvirni plan dela društva za naslednje triletno obdobje, ki ga je pripravila novoizvoljena predsednica dr. M. Kosec in je podan kot uvodnik v tej številki Informacij MIDEM. ZAPISNIK občnega zbora društva MIDEM, kije bil dne 15.05.1996 ob 18. uri v diplomski sobi Fakultete za elektrotehniko, Tržaška 25, Ljubljana. Dnevni red: 1. Otvoritev občnega zbora 2. Izvolitev organov občnega zbora 3. Poročilo predsednika 4. Poročilo člana Izvršilnega odbora, zadolženega za finance 5. Poročila ostalih organov društva 6. Diskusija po poročilih 7. Razrešitev organov društva 8. Predlog kandidacijske komisije za izbiro novih organov društva 9. Izvolitev novih organov društva 10. Smernice delovanja društva za naslednje triletno obdobje 11.Razno Prisotni: S. Amon, M. Kosec, M. Komac, F. Čuk, F. Jan, D. Belavič, I. Pompe, T. Mrdjen, M. Slokan, R. Ročak, I. Šorli in M. Limpel. 122 Informacije MIDEM 26(1996)2, Ljubljana ad 1. ad 4. Ob 18. uri je predsednik društva MIDEM, R. Ročak predlagal, da prestavimo začetek občnega zbora za pol ure, da bi dosegli kvorum. Čeprav ob 18:30 kvorum ni bil dosežen, smo začeli z občnim zborom. Predsednik društva R. Ročak je otvoril občni zbor in predlagal izvolitev delovnega predsedstva. ad 2. Za delovnega predsednika je bil soglasno izvoljen I. Šorli, za zapisnikarja M. Limpel, za overovatelja M. Slokan in S. Amon. V komisijo za sklepe so bili izvoljeni: M. Kosec, I. Potjipe in D. Belavič, v volilno komisijo pa: M. Limpel, F. Čuk, F. Jan in T. Mrdjen. ad 3. Poročilo predsednika R. Ročaka je priloženo. Nekaj odlomkov iz diskusije: - potrebno je izboljšati stike med organi društva in članstvom, - ponovno navezati stike z nekdanjimi člani iz BiH, Srbije in Črne gore, - pridobiti študente v vrste članov, - osnovati posebno študentsko sekcijo; za pridobivanje študentov je potrebno angažirati profesorje, - problematičen je majhen vpis na tehniške fakultete v Ljubljani in v Mariboru, - društvo naj vpliva na študijske programe, - organizacija inženirskega programa - pri tem naj sodeluje Splošno združenje za elektrotehniko, - v študij tehniških smeri je potrebno vključiti osnove ekonomije, - potrebno je vzgajati kadre, ki bodo sposobni prenašati raziskovalne dosežke v industrijo, - društvo lahko posreduje med iskalci zaposlitve in podjetji, - društvo naj sodeluje pri izdelavi študijskih in raziskovalnih programov, pri tem naj se vključijo tudi ambasadorji znanosti, ki so člani našega društva: Z. Fazarinc, D. Kolar in S. Pejovnik. - rast zaposlovanja je možna na dva načina: ali bodo tujci pokupili naša podjetja in z njimi začeli nov razvojni ciklus, ali z rastjo domačih novonastalih privatnih podjetij, kar bo počasnejše, trajalo bo 5 do 10 let, - društva morajo biti organizirana civilna družba, delovanje društev mora dobiti v družbi večji odmev. Sklep št.: 1 - Sprejme se poročilo predsednika. : Sklep št.: 2 - Akcijo za včlanitev študentov-dodiplomcev sprožimo. Organiziramo sestanek s profesorji in jim predstavimo dejavnost društva. To nalogo bi lahko opravil novi predsednik. Poročilo člana IO, zadolženega za finance (poročilo je priloženo zapisniku). Sklep št.: 3 - Finančno poročilo se sprejme. ad 5. Poročilo Nadzornega odbora je prebral njegov predsednik (poročilo je priloženo zapisniku) Poročilo Častnega razsodišča je prebral delovni predsednik (poročilo je priloženo zapisniku ). ad 6. Ker društvu kronično primanjkuje finančnih sredstev, je potrebno intenzivno iskati sponzorje, tudi med novimi malimi podjetji. Sklep št.: 4 - Poročila organov društva se sprejmejo. ad 7. Občni zbor je soglasno razrešil vse organe društva. ad 8. Kandidacijska komisija je predlagala kandidate za nove organe društva, kandidatna lista je priložena zapisniku. ad 9. Volilna komisija je pregledala glasovalne listke in preštela oddane glasove prisotnih in tistih, ki so glasovali po pošti. Vseh veljavnih glasovnic je bilo 24. Posamezni kandidati so dobili naslednje število glasov: dr. Marija Kosec, kand. za predsednico: 23. Za člane IO je kandidiralo 21 članov. Izvoljeni kandidati so dobili naslednje število glasov: prof. dr. S. Amon: 24 dr. R. Ročak: 24 mag. I. Šorli: 21 mag. M.Limpel: 22 dr. R. Babič: 21 F. Jan: 22 dr. M. Komac: 22 mag. M. Kramberger: 17 I. Pompe: 20 prof. L. Trontelj: 21 dr. W. Pribyl: 19 ker je doseglo 16 glasov kar 5 kandidatov, smo z javnim glasovanjem izbrali štiri preostale člane IO. Izvoljeni so bili: J.Štefanič: prof. dr. G. Soncini: 11 glasov 11 glasov 123 Informacije MIDEM 26(1996)2, Ljubljana dr. I. Krivka: 11 glasov in mag. G. Lipnjak 11 glasov Za Nadzorni odbor so kandidirali 4 člani, izvoljeni so bili: mag. F. Čuk: 20 glasov E. Pirtovšek: 20 glasov in mag. S. Šolar: 19 glasov. Za Častno razsodišče so kandidirali 4 kandidati; izvoljeni so bili: dr. B. Lavrenčič: 23 glasov mag. M. Slokan: 20 glasov in dr. M. Gliha: 15 glasov. ad 10. Smernice za nadaljno delo društva je podala novoizvoljena predsednica, njen govor je priložen zapisniku. POROČILO PREDSEDNIKA DRUŠTVA MIDEM ZA OBDOBJE OD 1.10.1992 DO 15.5.1996 Od zadnjega volilnega občnega zbora v Portorožu, dne 1.10.1992 do danes, je preteklo sedem in pol mesecev več, kot pa to predvidevajo pravila našega društva. Razlog temu ni pretirana želja za oblastjo, ki bi ga mogoče lahko kdo pripisal sedanjim organom društva, temveč je nastala zaradi spremembe termina skupščine in načina volitev, ki smo si ga prvič v zgodovini društva zastavili. Upam, da oba proceduralna "prekrška" nista tako velika, da bi kdo iz tega naredil problem. Takrat, ko smo imeli skupščino v Portorožu, so po Zagrebu padale granate in bilo je precej jasno, da bodo nastopili težki časi za naše egzemplarno, rekel bi skoraj neverjetno, sodelovanje članov iz nekdanje Jugoslavije. Takrat smo poslali poziv našim članom, da povzdignejo glas proti noriji in možni moriji. Pred tem smo že zastavili nov program dela društva, nova pravila, ki so na tej skupščini bila tudi sprejeta. Ta pravila so omogočila tudi formalno včlanitev strokovnjakov, ki priznavajo naša pravila in so iz kateregakoli konca sveta. S tem smo tudi formalno zastavili popolno internacionalizacijo društva. Pravila so bila objavljena v Informacijah MIDEM 4/92, registracija društva z novimi pravili pa je bila potrjena 7.10.1993. Ob sprejemu slovenskega zakona o društvih lahko ugotavljamo, da so naša pravila v popolnem skladu z zakonom, čeprav predlagam, da se med sklepe današnje skupčine uvrsti možnost formalnih popravkov členov, ki bi jih bilo potrebno iz zakonskih razlogov preformulirati. Delovanje društva je karakterizirano predvsem z aktivnostjo članov na konferencah društva in objavljanju strokovnih prispevkov v reviji. Preden dam nekaj statističnih podatkov o tem, naj navedem še nekaj podatkov o delovanju organov društva. Izvršilni odbor in časopisni svet sta se sestajala enkrat letno, ožji sekratariat na formalnih sejah devetkrat, veliko več pa na neformalnih delovnih sestankih, posebej ob sestankih uredniškega odbora, ki se je sestal petintridesetkrat. INFORMACIJE MIDEM Vsa leta smo uspeli redno izdati vse štiri predvidene številke. Le pri ta četrti se je vedno malo časovno zatikalo. INSPEC je nadaljeval redno zajemanje podatkov iz časopisa, uspel pa nam je tudi "veliki met", da nas je ISI - Institute for Scientific Information izbral za zajemanje podatkov v tri svoje podatkovne baze: SciSearch, Research Alert in Materials Science Citation Index. Nekaj statističnih podatkov je v priloženi tabeli, ki jo je pripravil odgovorni in glavni urednik Iztok Šorli. Iz tabele se lahko vidi, da je slovenskih prispevkov v zadnjih dveh letih 67%, v slovenščini pa 33%. To izkazuje uspešno realizacijo treh naših ciljev: da si pridobimo neslovenske avtorje, da slovenski avtorji pišejo v svetu razumljivem jeziku, da pa kljub temu še nadalje omogočamo gojenje in napredovanje slovenske strokovne terminologije. Časopisni svet petnajstih priznanih strokovnjakov in članov ima 8 neslovencev, Uredniški odbor pa so sestavljali Slovenci in en Hrvat. Informacije MIDEM Pregled člankov, 1992 - 1995 1992 1993 1994 1995 j domači 24 21 15 15 tuji 1 7 10 9 slovenščina 14 7 8 8 angleščina 11 21 17 16 SKUPAJ 25 28 25 24 domači, % 96 75 60 62 tuji, % 4 25 40 38 slovenščina, % 56 25 32 33 angleščina, % 44 75 68 67 Verjetno se ne moremo ne strinjati s sklepom, da sta ta dva odbora v preteklem obdobju delovala izredno dobro, zato bi se vsem njunim članom javno zahvalili, predvsem pa Iztoku Šorliju, ki je bil motor in nosilec vseh dejavnosti. KONFERENCE MIEL - SD Uspešno sta bili vsako leto organizirani združeni posvetovanji o mikroelektroniki MIEL in sestavnih delih SD. Statistični podatki o referatih in udeležencih so v priloženi tabeli. Lahko se vidi trend zadrževanja skupnega števila referatov nad 50, vendar s tendenco zmanjševanja števila referatov iz Slovenije in povečevanja referatov iz drugih držav. Vse konference so bile izredno dobro organizirane, za kar smo kot društvo zmeraj dobili pohvale samih udeležencev, v prijetnih krajih. Naj samo spomnim: 1992 v Portorožu 1993 na Bledu 124 Informacije MIDEM 26(1996)2, Ljubljana Konferenca MIEL-SD Pregled udeležbe in referatov, 1992 - 1996 MIEL-SD'92 Portorož MIEL-SD'93 Bled MIEL-SD'94 Rogla MIEL-SD'95 Terme Čatež MIEL-SD'96 Nova Gorica Skupaj referatov 64 55 53 53 64 • domači 46 42 41 36 40 • tuji 18 13 12 16 24 Skupaj udeležencev - 61 68 73 * • domači - 49 57 58 * • tuji - 12 11 15 * Število držav -referati 9 7 6 9 11 Število držav -udeleženci - 8 6 7 * 1994 na Rogli 1995 v Čatežu 1996 pa bo v Novi Gorici. DRUGE DEJAVNOSTI Društvo v preteklem obdobju ni oraganiziralo nobenih seminarjev, šol, ali kaj podobnega, če izvzamemo javno prezentacijo na stojnici Ministrstva za znanost Republike Slovenije ob razstavi Sodobna elektronika. ČLANSTVO Koliko nas je po tem zgodovinskem premetavanju v preteklem obdobju? Uradno 402, iz petnajstih držav. Statistika je v prilogi. Pri tem niso šteti eventuelni člani iz bivših jugoslovanskih republik Srbije in Črne Gore, katerih članstvo smo ob embargu svetovne organizacije tudi mi zamrznili, ni članov iz Bosne in Hercegovine ter Makedonije, ki so se v vojni vihri razpršili, izgubili iz poštnih evidenc ali pa so bili ubiti. Predlagam sklep, da se strpno poskusi ponovno navezati stike z njimi in jih ponovno voditi v evidenci, saj kriterij o neplačanju članarine kot razlog izstopa ni vtem primeru sprejemljiv. Edini član, ki je bil na svojo eksplicitno željo izbrisan iz članstva že leta 1992, je bil Ljutica Pešič iz Beograda. Zahvaljujem se vsem za podporo med mojim "vladanjem". Upam, daje bilo aktivno, uspešno in ne preveč svojeglavo. Zahvaljujem se vsem aktivnim članom, zahvaljujem se vsem podjetjem in njihovim direktorjem, ki so sponzorirali bodisi časopis, bodisi konference. Posebna zahvala gre tudi ministru za znanost Republike Slovenije za to obdobje, gospodu Radu Bohincu za njegovo izredno osebno podporo in podporo ministrstva, ki ga je vodil. Lahko pa se že zahvalimo tudi sedanjemu ministru gospodu Andreju Umeku, ki nadaljuje z isto politiko podpore. Sicer pa naj končam Pregled članov društva MIDEM, maj 1996 država fizična oseba pravna oseba SLOVENIJA 274 36 ZDR. DRŽAVE 2 AVSTRIJA 4 NEMČIJA 4 HRVAŠKA 92 2 ANGLIJA 2 MAKEDONIJA 9 ITALIJA 5 FRANCIJA 2 ČEŠKA 1 ŠVICA 1 UKRAJINA 1 BELGIJA 1 ŠVEDSKA 1 DANSKA 1 BOLGARIJA 1 SKUPAJ 402 38 SLOVENIJA 274 36 TUJINA (15 držav) 128 2 SKUPAJ 402 38 s postulatom: ni težko podpirati nekaj, kar je trdno. To pa naše društvo zagotovo je. V Ljubljani, 15.5.1996 dr. Rudolf Ročak 125 Informacije MIDEM 26(1996)2, Ljubljana USTANOVITEV NOVEGA INSTITUTA Inštitut za tehnologijo površin in optoelektroniko (ITPO), ki ima status zavoda, deluje od prvega decembra leta 1995. Ustanovljen je bil v času lastninskega preoblikovanja Inštituta za elektroniko in vakuumsko tehniko, p.o. (IEVT), v sodelovanju z Ministrstvom za znanost in tehnologijo (MZT), ustanovitelj pa je Tehnološko-razvo-jni sklad Republike Slovenije. ITPO zaposluje štirinajst ljudi, od tega devet raziskovalcev, ki imajo na Teslovi 30 v Ljubljani na razpolago laboratorijske in druge prostore v izmeri nekaj več kot 500 m2. Verjetno ne bo odveč kratka razlaga, zakaj je do ustanovitve ITPO sploh prišlo. Ministrstvo za znanost in tehnologijo je že leta 1991 imenovalo štiričlansko komisijo, ki je analizirala kronično slabo stanje IEVT in predlagalo organizacijske spremembe, ki naj bi omogočile avtonomnost manjšinskega raziskovalnega dela IEVT in transparentnost porabe sredstev namenjenih raziskovalnemu in razvojnemu delu, ki so se tako ali drugače prelivala v večji, proizvodni del IEVT. Žal pa je bilo samo priporočilo MZT prešibko, da bi prišlo do korenite reorganizacije IEVT že v navedenem obdobju, in sledilo je nadaljnje triletno slabšanje razmer in močno osipanje raziskovalnega kadra IEVT, od približno petinštirideset raziskovalcev v letu 1992 na manj kot dvajset ob koncu leta 1995. Ustanovitev ITPO je bila torej nujni izhod v sili za skupino raziskovalcev, ki so na svojem področju bili že do tedaj dokaj uspešni, pa tudi trdno vpeti v slovensko in mednarodno raziskovalno sfero ter industrijo. V okviru nove raziskovalne organizacije bomo še povečali aktivnost in kvaliteto dela na specializiranih raziskovalnih področjih. Inštitut za tehnologijo površin in optoelektroniko kot raziskovalni zavod opravlja temeljne, razvojne in aplikativne raziskave na področju naravoslovja in tehnologij. Njegova osnovna dejavnost je na področju preiskav in tehnologij površin trdnih snovi in tankih plasti, vakuumske optoelektronike, tehnike visokega in ultravisokega vakuuma, vakuumskih tehnologij, tehnike plazme, razvoja specialnih elektronk in optoelektronskih komponent. Teme petih mladih raziskovalcev nakazujejo smeri bodočega razvoja ITPO. Teme treh doktorandov so vakuumska ploskovna izolacija, interakcija vodikove plazme s površinami trdnih snovi ter preiskave površin z rentgensko fotoelektronsko spektroskopijo (XPS = ESCA), dva pa pripravljata magisterij s področja luminiscentnih materialov ter postopkov analize reflek-tometrskih merilnih rezultatov. Z Laboratorijem za analizo površin in tankih plasti smo vključeni v Nacionalni center za mikrostrukturno in površinsko analizo, v katerem sta še Laboratorij za mikrostrukturno analizo Odseka za keramiko ter Laboratorij za elektronsko mikroskopijo Odseka za fiziko trdne snovi z Inštituta Jožef Štefan. Laboratorij na ITPO je specializiran za preiskavo površin trdnih snovi (AES, SAM, SEM) in tankih plasti ter kompozitnih materialov in njihovih faznih mej (TFA). Opravljamo tudi mikroanalizo kovinskih, steklenih in keramičnih materialov (EMPA, EDX, WDX). ITPO dobro sodeluje z najpomembnejšimi slovenskimi tehničnimi inštituti, z univerzama v Ljubljani in Mariboru ter s slovensko industrijo, na primer s Fotono, Cryorefom, Iskro in drugimi. Sodelavci ITPO imamo vzpostavljeno dobro bilateralno sodelovanje s priznanimi tujimi institucijami in v okviru mednarodnih projektov v Evropi in ZDA, kar nam omogoča dostop do raziskovalne opreme, ki je v Sloveniji še nimamo, in do najnovejših informacij, pomembnih za naša raziskovalna področja. Bolj pomembna kot naštevanje tujih institucij so področja dela, na katerih sodelujemo z njimi. Ta so preiskava reakcij na faznih mejah tankih plasti (MPI Stuttgart, DLR, Köln), preiskava večplastnih struktur iz superprevodnih tankih plasti in kovinskih oksidov (FZ, ITP, Karlsruhe), preiskava reakcij v trdni fazi (Müfi, Budimpešta), optimizacija profilne analize tankih plasti (PHI, München, PHI, Minnesota), sodelovanje pri izgradnji žarkovnih linij na sinhrotronih (Elettra, Trst, FZ, Karlsruhe, COPERNICUS), tehnike plazme in obdelava površin materialov (Univerza Bratislava, CEEPUS) ter v zadnjem času ionska implantacija (IAEA, Dunaj). Področje, na katerem ITPO deluje, je v zadnjih letih zapustilo več raziskovalcev, zato je ena glavnih nalog ITPO vzgoja novih kadrov za lastne potrebe in kasneje tudi za druge institucije. Področje preiskave površin zastopamo tudi pri rednem in podiplomskem študiju na obeh slovenskih univerzah. Poskrbeti bomo morali tudi za obnovo infrastrukturne opreme za področja, na katerih delamo, pri čemer pričakujemo sodelovanje z MTZ in z vsemi zainteresiranimi, ki tovrstne preiskave neobhodno potrebujejo pri svojem raziskovalnem delu ali v industriji. Nekateri sodelavci ITPO aktivno sodelujemo v Društvu za vakuumsko tehniko Slovenije in v Mednarodni zvezi za vakuumsko znanost, tehniko in aplikacije, kot tudi v uredniških odborih domačih strokovnih časopisov in v tujih recenzijskih odborih. A. Zalar Sodelavci inštituta za optoelektroniko tehnologijo površin in 126 Informacije MIDEM 26(1996)2, Ljubljana KONFERENCE, POSVETOVANJA, SEMINARJI, POROČILA CONFERENCES, COLLOQUYUMS, SEMINARS, REPORTS ISHM/NATO Workshop 1996 (poročilo s konference) NATO Advanced Workshop and Exhibition on Microelectronic Interconnections and Microassembly Udeležil sem se konference oz. delavnice za povezovanje v hibridni mikroelektroniki, ki sta jo sponzorirala NATO in ISHM (International Society for Hybrid Microelectronics). Konferenca je bila v dneh od 18. do 21. maja v Pragi. Na konferenci s približno 45 udeleženci iz 21 držav je bilo predstavljenih 32 referatov v petih sekcijah: 1. Trendi v pakiranju in povezovanju 2. Spajkanje in Flip Chip tehnologija 3. Žično bondiranje in TAB 4. Multi Chip Moduli 5. Povezovanje z debeloplastno tehnologijo V poročilu bom na kratko opisal vsebino nekaterih zanimivejših predavanj, na razpolago pa je zbornik povzetkov. Zbornik referatov bo izšel predvidoma v prvi polovici naslednjega leta. Registrirani udeleženci 1. Združene države Amerike 9 2. Češka 7 3. Slovaška 3 4. Belgija 2 5. Finska 2 6. Italija 2 7. Japonska 2 8. Madžarska 2 9. Nizozemska 2 10. Poljska 2 11. Švedska 2 12. Ukrajina 2 13. Velika Britanija 2 14. Bolgarija 15. Danska 16. Francija 17. Irska 18. Kanada 19. Norveška 20. Slovenija 21. Španija Skupaj: Mikroelektronika v Skandinavskih državah Soren Noerlyng (Micronsult, Danska) je predstavil aktivnosti na področju hibridne mikroelektronike v štirih Skandinavskih državah: Danska, Norveška, Švedska In Finska. Posebej je osvetlil, dogajanja na sledečih povezovalnih tehnologijah: MCM (Multi Čhip Modules), Flip-chip, TAB (Tape Automated Bonding), COB (Chip On Board), in PTF (Polimer Thick Film). Na področju raziskav so raziskovalne institucije in proizvodne firme povezane v skupne projekte omenjenih štirih držav in/ali v skupne projekte Evropske Unije. Miniaturizacija v elektroniki za široko potrošnjo Co Van Veen (Philips, Nizozemska) je predstavil zahtevo po miniaturizaciji tudi v elektroniki namenjeni za široko potrošnjo. Primerjal je tri tehnologije COB (Chip On Board), TAB (Tape Automated Bonding) in Flip-chip. Slednji tehnologij je dal prednost pred prvima dvema. Tehnologiji COB očita zahtevno čiščenje zaradi kombinacije tehnologij spajkanja in žičnega bondiranja. Tehnologiji TAB pa očita drago proizvodno opremo. Medtem ko tehnologija Flip-chip v zadnjem času dosega ponoven razvoj na področju zmanjševanja dimenzij in tehnoloških procesov za izdelavo in uporabo Flip-chip komponent. Primerjava plastične in hermetične inkapsulacije Nihal Sinnadurai (TWI, England) je predstavil obsežno študijo zanesljivosti elektronskih komponent (10.000.000 enot) v telefonskih centralah na 260 lokacijah v Veliki Britaniji in Indiji. Zasledovali so klimatske in ostale pogoje okolice ter analizirali katere komponente in zakaj odpovedujejo. Na osnovi rezultatov so določili pospeševalne faktorje degradacije komponent. Poleg tega so ugotovili, da plastično inkapsulirane elektronske komponente celo manj odpovedujejo kakor her-metično inkapsulirane komponente (?!). Pogoji okolice, ki so vplivali na odpovedi: 1. Temperatura 55% 2. Vibracije 20% 3. Vlaga 19% 4. Prah 6% Komponente, ki so odpovedovale: 1 1. Transistorji 36% 1 2. Integrirana vezja 32% 3. Diode 15% 4. Upori 9% 47 5. Hibridna vezja 8% 127 Informacije MIDEM 26(1996)2, Ljubljana Vzroki odpovedi: 1. Električna preobremenitev 44% 2. Nezanesljiva komponenta 40% 3. Vzrok odpovedi ni najden 10% 4. Napačna komponenta 2% 5. Ostalo 4% Flip-chip tehnologija Nekaj predavateljev (Bill Brox, IVF, Švedska; Peter Bodo, IMC, Švedska; Carlo Cognetti, SGS Thomson, Italija; I. Suni, VTT Electronics, Finska) je izrazito povdar-ilo perspektivnost Flip-chip tehnologije za povezovanje in pritrjevanje integriranih vezij. Po njihovih raziskavah s Flip-chip tehnologijo dosežejo večjo gostoto funkcij na hibridnem vezju in višjo delovno frekvenco. To dosežejo zaradi zmanjševanja dimenzij. Dimenzije zmanjšujejo, tako da razporedijo priključke oziroma kroglice po površini (Ball Grid Array), razmak priključkov zmanjšujejo proti dimenziji 80 ¡¿m, debelino Flip-chipa pa proti 700 |im. Poleg tega pa povečajo zanesljivost, saj se z uporabo Flip-chipa izognejo enemu niviju povezav (spajkanje oziroma žično bondi-ranje). Cenovna primerjava je tudi na strani Flip-chip tehnologije, saj bi se pri masovni produkciji Flip-chip komponente pocenile do 7 krat (?!). Po trditvah enega izmed avtorjev so v firmi Erikson mnenja, da če bi prej poznali vse probleme, ki jih imajo z SMD tehnologijo, bi šli direktno na Flip-chip tehnologijo. Drugi avtor pa zatrjuje, da je SGS pripravljen za masovno produkcijo Flip-chip komponent. Poleg tega ocenjuje, da bodo te komponente cenejše od SMD verzije. Kasneje je to stališče omilil z izjavo, da bo vse odvisno od povpraševanja. Multi Chip Module V sekciji za multi chip module je Illyefalvi-Vitez (Tehnična univerza v Budimpešti, Madžarska) predstavil mednarodni projekt za cenene multichip module z udeležbo Velike Britanije, Belgije, Madžarske, Romunije in Slovenije. G. Harsanyi (University Park Campus, Miami, Florida) je obravnaval zanesljivost večplastnih MCM z vidika elek-tromigracije kovinskih ionov. Trdi, da temperatura žganja močno vpliva na intenzivnost nastajanja dendri-tov v večplastnih MCM. Računalniške simulacije Kar nekaj predavateljev je predstavilo računalniške simulacije in/ali matematično modeliranje temperaturnih razmer na mikroelektronskem debeloplastnem vezju, mehanskih lastnosti različnih spojev, ter mehanske napetosti v komponentah, substratu in spojih. V sekciji "Povezovanje z debeloplastno tehnologijo" je bil predstavljen tudi naš referat, ki obravnava posebnosti, ki jih mora upoštevati projektant pri izdelavi senzorjev in pretvornikov, katerih pretvorniški del je izdelan s hibridno mikroelektronsko tehnologijo. Naslednja konferenca oz. delavnica (ISHM/NATO Workshop 1997) naj bi bila organizirana maja 1997 v Sloveniji. Darko Belavič HIPOT-HYB, Šentjernej RO HV Ljubljana Delavnica COST 514: Feroelektrične tanke plasti Za uvod: feroelektrične tanke plasti na siliciju so potencialni kandidati za vrsto novih mikroelektronskih, op-toelektronskih in mikromehanskih komponent. Najpogosteje se omenjajo spominski elementi, kjer imajo omenjene plasti bodisi aktivno bodisi pasivno funkcijo, valovodi, preklopniki, modulatorji, mikrosenzorji in ma-kroaktuatorji. Zato je to, kakšnih 10 let staro raziskovalno področje, ki združuje raziskovalce in inženirje s področja materialov in mikroelektronike, v zadnjih letih deležno izjemne pozornosti. Pred tremi leti se je v Evropi začel izvajati koordiniran projekt COST 514, Feroelektrične tanke plasti. Njegov namen je intenzivirati in povezati raziskave materialov, tehnologij in potencialnih aplikacij in na ta način hitreje rešiti tehnološke probleme, ki so upočasnili vpeljavo feroelektričnih tankih plasti v komercialne produkte. Danes v tem projektu, v katerega je vključena tudi skupina iz Odseka za keramiko IJS, sodelujejo vse pomembne evropske raziskovalne institucije (19), ki delajo na tem področju. Projekt povezuje pet podpro-jektov: Zanesljivost feroelektričnih tankih plasti, Procesiranje tankih plasti z laserji, Feroelektrični materiali za kondezatorje v neposrednem kontaktu s silicijem, Razvoj plasti za SAW aplikacije in Materiali in tehnologije za optoelektroniko. V okviru projekta je bila že tretjič zapovrstjo organizirana delavnica, katere namen je bil predstaviti delo in rezultate raziskovalnih skupin ter skupnega projekta. Letos je drugega in tretjega marca to delavnico organziral Inštitut za materiale iz Madrida. Poleg sodelavcev projekta COST 514 so na njej sodelovali še gostje iz Amerike, gost iz Siemensa iz Munchna in gost iz Instituta za fiziko iz Rige. Posebej velja omeniti predavanje prof. Kingona iz North Carolina University, ki je predstavil vsebino raziskav pa tudi finančni angažma Japonske in Amerike na področju raziskav feroelektričnih spominskih elementov. Tu predvidevajo največjo proizvodnjo in finančni uspeh feroelektričnih tankih plasti. Po njegovem mnenju se je na tem področju zgodilo tisto, kar so raziskovalci že nekaj let pričakovali (in upali): resničen prodor feroelek- 128 Informacije MIDEM 26(1996)2, Ljubljana tričnih tankih plasti v industrijske raziskave. Predavatelj je naštel vrsto industrijskih gigantov, ki so se za to odločili. To je potrdil tudi dr. VVersing iz Siemensa. Sicer pa zelo na kratko nekaj strokovnih zaključkov z delavnice. Raziskave kažejo, da se med materiali največkrat pojavlja trdna raztopina na osnovi Pb(Zr,Ti)C>3-Za mikromehanske aplikacije zaenkrat sploh nima pravega tekmeca, medtem ko za pasivne elemente v spominskih celicah (DRAM) iščejo druge materiale. Med procesnimi tehnologijami se vsaj Evropa največ ukvarja s sol-gelom, manj pa z naprševanjem in nanašanjem iz parne faze (MOCVD). Uveljavlja se tudi nanašanje plasti z laserji (PLD), kjer se, poleg ostalega, intenzivno dela na modifikaciji opreme in postopka, tako da bi metoda omogočala nanašanje plasti na velike površine. Napredovali so tudi postopki pri nadaljnjem procesiranju elementov, ki so v veliki meri prenešeni iz polprevodnih tehnologij. Kot sem omenila, v COST 514 sodeluje tudi skupina iz Odseka za keramiko Instituta "Jožef Štefan". Ukvarjamo se s sol-gel postopkom nanašanja plasti in strukturno karakterizacijo plasti. V projekt smo se vključili tudi zato, ker verjamemo, da je uporabnost in zahtevnost proizvodnje teh elementov tako raznolika, da je tu lahko mesto tudi za slovenska podjetja. Če kogarkoli strokovno in podjetniško to področje podrobneje zanima, smo tu z dodatnimi informacijami. dr. Marija Kosec Odsek za keramiko Institut "Jožef Štefan" Jamova 39, 1001 Ljubljana Tel.: 1773-368 Faks: 1263-126 Elektronska pošta: Marija.Kosec @ijs.si TRIBOLOGY - SOLVING FRICTION AND WEAR" poročilo s simpozija Od 9. do 11. januarja 1996 je v Esslingenu (Nemčija) potekal 10. mednarodni simpozij z naslovom "Tribology - Solving Friction and Wear" Problems. Simpozija sem se udeležila s prispevkom "Comparison of the Fretting Wear in 100Cr6/100Cr6, Si3N4/Si3N4 and Si3N4/100Cr6 Contacts in Lubricated and Dry Conditions". Delo je nastalo v sodelavi s Centrom za tribologijo in tehnično diagnostiko, zato je del stroškov potovanja krila Fakulteta za strojništvo. Srečanje je potekalo na Tehnični akademiji Esslingen, ki je bila tudi glavni organizator. Udeležilo se ga je preko 800 ljudi, predstavljenih je bilo okoli 300 polurnih govornih prispevkov. Predavanja, ki jih je spremljala razstava raziskovalne opreme, so potekala v sedmih vzporednih sekcijah. Precejšen poudarek je bil na obrabi raznih strojnih delov, orodij za obdelavo in na metodah zmanjševanja trenja in obrabe. Najbolj obiskane so bile vsekakor sekcije, ki so obravnavale mazanje. Govora je bilo o različnih vrstah maziv in dodatkov, pri čemer je bilo največ zanimanja za nova -biorazgradljiva olja. Odpadna olja namreč predstavljajo veliko obremenitev za okolje, zato v zadnjem času intenzivno poskušajo mineralna in druga olja zamenjati z razgradljivimi, ki pa zaenkrat še ne kažejo povsem zadovoljivih rezultatov. Simpozija se je udeležila tudi relativno velika skupina predstavnikov različnih podjetij za proizvodnjo ali prodajo maziv iz Slovenije. Nekaj sekcij je bilo posvečenih materialom, pri čemer so glavni predstavniki obrabno odpornih materialov različni keramični materiali. Tako se v firmah, ki se sicer ukvarjajo z mazivi (npr. Lubrizol, ki je eden najmočnejših proizvajalcev olj in aditivov), vedno bolj intenzivno posvečajo raziskavam materialov in ustreznih maziv za keramične elemente, saj vedno več kovinskih delov (tudi v boljših avtomobilih) zamenjujejo s keramičnimi. Glavna predstavnika sta Si3N4 in SiC, katerih prednosti so predvsem nizek koeficient trenja in majhna obraba na dolge roke ter odpornost na povišane temperature, predvsem pa manjša občutljivost na eventuelno pomanjkanje maziva v primerjavi s kovinami. Predstavljena je bila vrsta zanimivih prispevkov o obnašanju različnih keramičnih materialov pri različnih triboloških pogojih. Glavno sporočilo, ki ga je poslušalec lahko razbral iz predavanj, je bilo, da so tribološke lastnosti keramike zelo odvisne od pogojev uporabe. Glavni parameter, ki določa vrsto in intenzivnost obrabe, je temperatura na stiku, pri čemer se lahko visoka temperatura (preko 1000°C) pojavlja zelo lokalno. Pri teh pogojih pa kljub pregovorno visoki kemijski stabilnosti keramičnih materialov prihaja do tribokemijske reakcije. V primeru AI2O3 prihaja do hidratacije, pri Zr02 pride lahko do fazne transformacije, značilnost neoksidne keramike pa je velika odvisnost od mazalnega sredstva. V prisotnosti vodnih molekul nastajajo oksidi, katerih lastnosti zavisijo od vrste gibanja elementov. V sekciji o tribologiji kostnih nadomestkov v človeškem telesu je bil med drugimi predstavljen prispevek, ki je opisoval nenavaden pristop skupine raziskovalcev iz ZDA in Nemčije. Izhajali so iz dejstva, da je idealen tribološki sistem katerikoli sklep v človeškem telesu, zlasti pa kolčni, ki mora dolga leta prenašati relativno velike obremenitve. Do obrabe v sklepu pride le v redkih primerih, ki jih večinoma povzroči kakšna bolezen. To pomeni, da je - dokler ne pride do spremembe v sestavi mazalne tekočine v sklepu, mazanje idealno in vredno posnemanja vsaj v primeru vgradnje umetnih kolkov. Predstavljena je bila tudi študija možne povezave med obrabo sklepov in osteoartritisom. V mnogih primerih so namreč v obolelih sklepih našli obrabne delce, ki bi lahko povzročili razvoj bolezni. Rezultati opazarjajo na pomembnost pravilne izbire materiala za kostne nado- 129 Informacije MIDEM 26(1996)2, Ljubljana mestke in njihove geometrije v tribološkem kontaktu. Predstavljeni so bili rezultati primerjave obrabne odpornosti različnih biokompatibilnih materialov pri pogojih, ki simulirajo realne. Splošni vtis, ki sem ga dobila na srečanju, je bil, da je kljub zadržkom pri uporabi keramičnih materialov v konstrukcijskih delih, ki so sledili začetnim pretirano optimističnim napovedim, keramika material, ki v strojnih delih sicer počasi, vendar uspešno zamenjuje klasične materile. Ovira za hitrejše uvajanje je še vedno skeptika uporabnikov (konstrukterjev) in zaenkrat slabo obvladovanje primernih mazalnih sredstev. Vendar pa kaže, da se keramični materiali vendarle vztrajno uveljavljajo v strojnih elementih. Saša Novak Odsek za keramiko Institut "Jožef Štefan" Jamova 39, 1001 Ljubljana Tel.: 1773-368 Faks: 1263-126 POROČILO S KONFERENCE ISPSD'96 (The 8th international symposium on power semiconductor devices and IC's) Od 20. do 23. maja je bila na Maui, Hawaii, ZDA, konferenca oz. simpozij o močnostnih elementih in integriranih vezjih. Konferenca zajema zelo ozko a obenem pomembno vejo polprevodniških elementov, ki jih potrebujemo praktično v vsaki električni napravi za regulacijo, krmiljenje in zaščito elektronskih delov. Na konferenci se zberejo vsi pomembnejši raziskovalci iz tega področja, predvsem pa je to konferenca, kjer se zberejo vsi pomembnejši proizvajalci elektronskih komponent. Zato je konferenca še posebno zanimiva in zelo koristna, saj je industrija tista, ki narekuje razvoj novih elementov in določa prihodnje trende razvoja. Največ udeležencev konference je prišlo iz ZDA, na drugem mestu je bila Japonska, nato pa Evropa. Izkazalo se je, da Japonci vodijo predvsem na področju diskretnih elementov, američani in evropejci pa na področju integracije le teh v močnostna integrirana vezja oziroma tako imenovane pametne močnostne elemente (smart power devices). Sam sem predstavil delo z naslovom "Diffused spiral junction termination structure: modeling and realization", soavtorja Prof. dr. Slavko Amon (Laboratory za elektronske elemente, Fakulteta za elektrotehniko, Ljubljana) in dr. Georges Charitat (LAAS/CNRS, Toulouse, Francija). Predstavili smo rezultate nove zak-Ijučitve planarnih spojev, ki omogoča zelo visoke prebojne napetosti in je obenem zelo robustna in enostavna za izdelavo. Delo je bilo zelo dobro sprejeto, posebno glede na dejstvo, da smo pokazali tudi eksperimentalne rezultate struktur, ki so bile procesirane v našem laboratoriju. To je posebno velik uspeh glede na relativno skromno opremo našega laboratorija, ki se ne more primerjati z raziskovalnimi možnostmi velikih polprevodniških firm kot so Motorola, SGS-Thomson, International Rectifier, Texas Instruments, Daimler-Benz, Mitsubishi Electric, Philips, Toshiba, Fuji Electric, Hitachi, itd. Glede na močno udeležbo velikih industrijskih podjetij, ki so prikazali najnovejše raziskovalne dosežke, je bilo iz akademskih raziskovalnih organizacij le 17% vseh predstavljenih del. Konferenca je pokazala, da se akademske organizacije lahko enakomerno kosajo z industrijskimi le z zelo domiselnimi novimi koncepti in raziskavami novih struktur, ki bodo morda vgrajeni v prihodnje generacije polprevodniških elementov. Več o smereh razvoja visoko-napetostnih in močnostnih elementov bom poskusil pripraviti v posebnem prispevku za MIDEM, kogar pa zanimajo članki predstavljeni na konferenci, tako na papirju kot na CD disku (zgoščenki), se lahko oglasi pri avtorju članka. Aloha, dr. Dejan Križaj Laboratorij za elektronske elemente Fakulteta za elektrotehniko, Ljubljana VSE ČLANE DRUŠTVA MIDEM VLJUDNO PROSIMO, DA PORAVNAJO ČLANARINO ZA LETI 1995 IN 1 996. ALL MIDEM MEMBERS ARE KINDLY ASKED TO PAY MEMBERSHIP FEE FOR 1 995 AND 1996. HVALA LEPA! THANK YOU VERY MUCH! 130 Informacije MIDEM 26(1996)2, Ljubljana VESTI - NEWS News from AMS The year 1995 - the most successful to date in the history of the company! For a number of reasons the year 1995 was an exceptional one for Austria Mikro Systeme International AG: 1) The business development of the AMS Group with extraordinary high growth rates of sales, profit and cash earnings: The sales increased by 71% from 1,107 MATS in 1994 to 1,887 MATS in 1995. This represents a significant outperformance compared with most competitors. The profit on ordinary activities increased by 82% from 147 MATS in 1994 to 267 MATS, including the 26 MATS losses of the two new consolidated subsidiaries. This represents a return on sales of 14%. The cash earnings with a more than 60% growth of 431 MATS reached a new record. Key Figures 1995 AMS AG AMS Group1' Sales (MATS) 1,698 1,887 Sales growth absolute (MATS) + 592 + 780 Sales growth in % + 54% + 71% Profit on ordinary activities (MATS) 292 267 Profit growth + 99% + 82% Return on sales 17% 14% Cash earnings (MATS) 406 431 Cash earnings in % of sales 24% 23% Total assets (MATS) 2,653 3,782 Equity (MATS) 1,527 2,166 Equity ratio 58% 57% 1) Consolidation of SAMES for 9 months, of Thesys for 2 months. The ÔVFA result and the cash earnings according to ÔVFA will be published on April 24, 1996. 2) The majority participations: In July 1995 Austria Mikro Systeme acquired 51 % of SAMES in South Africa - ASIC specialist for solid state electronic current metering, security and identification applications - and in October 1995 51.25% of THESYS in Germany - ASIC specialist for multimedia, communications, industrial and automotive electronics - inline with the longterm company strategy to expand the customer base and to improve the market position as a result of the increased product and process offerings of the partnerships. Both subsidiaries, though having achieved a high increase in sales and having improved their earnings in the past business year are as planned in a loss situation. 3) The capital increase: The nearly complete exercise of the acquisition rights on the new shares by the existing shareholders documents the confidence of the employees and shareholders in the longterm strategy of the company. The proceeds from the rights issue of around 733 MATS were used partially for financing the participations; the remainder will be used for the further expansion at company headquarters in Unterpremstatten to finance the further growth of Austria Mikro Systeme. 4) The future potential created: With the establishment of the AMS Group approximately 1,500 highly qualified and motivated employees at three manufacturing locations as well as in the worldwide design centres and in the sales offices are available to assist customers more efficiently and to better utilize the possibilities of the international ASIC market. Outlook for 1996: The situation on the international semiconductor market at the beginning of 1996 has been characterized by a general decrease in demand. The AMS Group will not be able to disengage completely from this worldwide trend. Thus, the pace of sales and earnings development will not be sustained. However, we expect that there will be no fundamental change in the midterm growth outlook for the relevant markets which the AMS Group serve. The First Quarter 1996 of the AMS Group The AMS Group reports the results for the first quarter 1996 (in MATS): AMS Group,1 - 3'96 e Order Entry 405 » Net Sales 572 O Backlog 870 O Employees 1,526 o Capital Expenditure 153 • Profit before Taxes 38 (after minority interests) 131 Informacije MIDEM 26(1996)2, Ljubljana The AMS Group with its members Thesys, Austria Mikro Systeme and Sames is a company group that specializes in the design and production of ASICs (application specific integrated circuits) for the market segments communications, multimedia, automotive and industrial electronics and meets the customer challenges with state-of-the-art technology and know how. The AMS Group, formed in 1995, reported net sales of 1,887 million ATS for the year 1995 (after consolidation of Sames for 9 months and Thesys for 2 months). Since the beginning of the year main emphasis was placed on capacity competences and the AMS Group was optimized by the simultaneous use of the synergies in design, mask lithography, production, test and quality. Hence, the prerequisites were established to strengthen the better basis for the future in the changed environment of the general sluggishness of the economy, especially in the field of semiconductors. The AMS Group will not be able to disengage completely from this worldwide trend. Thus, the pace of sales and earnings development will not be sustained. However, we expect that there will be no fundamental change in the midterm growth outlook for the relevant markets which the AMS Group serve. AMS Schloß Premstätten A-8141 Unterpremstätten Austria Fax: +43 (03136) 52 501, 53 650 Tel.: +43 (03136) 500 Email: info @ams.co.at http: //www.ams.co.at News from "Solid State Technology" AMD: $3 billion for MPU fab and design center in Dresden As anticipated, Advanced Micro Devices has announced a ten-year plan to invest $3 billion in a microprocessor center in Dresden, Germany. The plan has been approved by the company's board of directors and is now subject to final approval by the German state of Saxony, the Federal Republic of Germany, and the European Economic Community. Both Saxony and the German federal government will provide substantial financial assistance through grant allowances and loans. The new facility will be AMD's first wafer fab in Europe. Chairman and CEO Jerry Sanders noted that Germany is the biggest market in Europe and cited the highly skilled workforce as one reason for choosing the location. About 1400 people will be employed at the new center. Over the next five years, the company will spend $1.5 billion to construct an 87,000-square-meter plant, to be named Fab 30. It will include about 9000 square meters of cleanroom space, and have a capacity of up to 6000, 200-mm wafers/week. Initial production, commencing by the end of 1998, will be at 0.25 micron, with later migration to 0.18 micron planned. Groundbreaking will take place before the end of 1996, said Sanders. The initial investment phase also includes a design center which will begin operations about two years after groundbreaking. Siemens recently opened its own DRAM fab in Dresden, where 16-Mbit and 64-Mbit devices will be produced. The former East German city is also home to Zentrum Mikroelektronik Dresden, a state-controlled producer of ASICs, fast SRAMs, and other devices with annual sales of about DM 40 million ($29 million). AT&T to expand Spanish wafer fab AT&T Microelectronics recently celebrated the 10th anniversary of its fab in Madrid, Spain, with the word that it will invest an additional $145 million to add capacity and extend the facility's capability to 0.35 micron during 1996. ASICs, FPGAs, and DSPs utilizing the new technology will start coming off the line in 1997; wafer output is expected to grow to 5000/week by the middle of 1996. IMEC to handle ADEQUAT+ work The Belgian IMEC research institute has been selected to coordinate advanced development work on 0.25- and 0.18-micron CMOS processes under the 15-month ADEQUAT+ project, which is funded by the Esprit Framework IV program. The $3 million effort will develop interconnect processing steps and modules for 0.25-mi-cron CMOS by the end of this year; these back-end modules will then be combined with front-end capabilities developed in the just-completed ADEQUAT-2 project. Concept testing and patterning feasibility for 0.18-micron front-end modules is expected to be complete by the end of this year, and a lithography process should be ready a year later. Front-end modules are to be prepared by 1998, and back-end by 1999. Partners include the Dutch DIMES research center, the German Fraunhofer Institutes, the French GRESSI center, and chipmakers GEC-Plessey Philips, Siemens, and SGS-Thomson. SGS-Thomson plans new fabs SGS-Thomson will begin work on two new 200-mm wafer fabs in 1996, one in Italy and one in another country according to European press reports. Company executives said they expect to invest between $800 million and $1 billion in newfacilities in 1996, about the same as 1995, and will also spend between $50 million and $100 million on eight factory upgrades. A formal announcement of the firstfab is expected shortly, said the company. Strong mask growth seen for European market The European reticle/mask market will grow 10.2% in 1996, according to US firm The Information Network, Williamsburg, VA. The market is expected to reach $134 132 Informacije MIDEM 26(1996)2, Ljubljana million this year, up from $122 million in 1995. Growth in the mask market will be driven by IC growth. The strong mask market showing in Europe is attributed to heavy investments by major semiconductor manufacturers Philips, SGS-Thomson, and Siemens; growth rates in excess of 90% among smaller European IC companies; Europe's lead in telecom and electronic components in automobiles; and strong consumption of masks by US companies in Europe. Low-k dielectrics featured at 1996 DUMIC conference As the world moves closer to using low dielectric constant (k) polymers for inter-metal dielectric (IMD) applications, engineers are being forced to consider all of the different materials properties they will need. This search was the focus of many of the 33 presentations at the second annual Dielectrics for VLSI/ULSI Multilevel Interconnection Conference (DUMIC), held in February in Santa Clara, California. Some 500 attendees, up from 400 last year, heard reports on new materials and reexaminations of well-known dielectrics. Researchers from Fujitsu presented a paper entitled, "Fluorocarbon Films Deposited by PECVD with High Thermal Resistance and Low Dielectric Constant". This work is the first time that low-k polymer films have been deposited by a process in which the reaction occurs on the wafer surface, and it opens up the possibility of entirely new processes for IMD. Films were deposited into 0.65-jim-deep quarter-|im gaps, with excellent results. Using C2H2 and C4F8 precursors, a k lower than 2.4 was achieved with thermal decomposition and glass transition temperatures higher than 400°C. A last-minute substitution on the program was an exciting proof-of-concept paper on 2.4-k polymer foams. Since air has a k of 1.0, If some fraction of the polymer is converted into a gas, then the overall material would see a fractional reduction in k. Researchers at IBM-AI-maden have worked out a two-step foam process where molecular phase segregation first produces a primary phase with a high glass transition temperature and a second phase that is burnt off during the second low-temperature step. There are two significant benefits of this two-step process. First, the burn-off can be done after metallization, which is advantageous for maintaining topography in vias. Second, there is very tight distribution of pore sizes between 10 and 100 nm. Vacuum-deposited Parylene was the subject of two presentations. Specialty Coating Systems, of Indianapolis, IN, presented data showing excellent film qualities, including 2.28 k. Unfortunately, researchers from Rensselaer Polytechnic Institute presented data showing that it is difficult to achieve deposition rates above 200 A/minute. The best conditions used a 500 V/cm electric field that reduced atomic oxygen incorporation and improved the deposition rate by 2 to 3.5 times. The dielectric constant is not the only material property that is important for an IMD material. Just as the need for lower-k materials is driven by increasing interconnert speeds, there is a corresponding need for improved thermal conductivity to compensate for metal Joule heating. Researchers at Texas Instruments compared the characteristics of polyimide, methylsilsesquioxane spin-on polymer (SOP), and inorganic porous hydrogen silsesquioxane (HSQ) using ASTM electromigration test structures. The results quantify the 3 to 4 times reduction in thermal conductivity seen in these new low-k materials (see table). For a given current density, the reduced thermal conductivity will lead to a larger temperature Increase, resulting in shorter electromigration lifetimes. Results of Tl study Dielectric film Thermal Conductivity (mW/cm °C) PECVD 11.5 HDP-CVD oxide 12.0 Polymide 2.4 SOP 2.4 HSQ 3.7 The Tl researchers showed through both simulation and experimental results that in a multilevel metal system, the inter-level dielectric (between metal layers) plays a more important role for thermal conduction than intra-level dielectrics (between metal lines within a given layer). Their report suggests that low-k materials be used only at the intra-level, where they will have the greatest impact for RC reduction and will not significantly reduce the thermal conductivity of the system. This opinion was echoed by DUMIC luncheon speaker Betsy Weitzman, senior staff scientist in the Materials Research and Strategic Technologies organization within Motorola's Semiconductor Product sector. The process flow to create such a composite dielectric is identical to that currently used for spin-on glass (SOG) gap-fill, though the driver has changed from mechanical to electrical performance. Expect to start hearing about this process flow coming to a fab near you. 1-Gbit DRAM designs, exotic memories, highlight 1996 ISSCC Attendance at the 43rd IEEE International Solid State Circuits Conference, held February 8-10 in San Francisco, jumped to 2800 from last year's 2200, with many sessions (including those concerning low-voltage devices and video cameras on a chip) filled to overflowing. Following are some of the highlights of the presented papers. 8 This is the second year that 1-Gbit DRAM designs were presented. Mitsubishi has used x-ray lithography to define high-k barium-strontiumtitanate (BST) 0.29-square-micron gates, with 3 poly,1 tungsten, 2 copper, and 1 aluminum interconnect layers. Sam- 133 Informacije MIDEM 26(1996)2, Ljubljana sung uses KrF optical lithography to define mid-k tantalumpentoxide 0.334-square-micron gates, with 3 poly,1 titanium-silicide, 2 tungsten, and 2 aluminum layers. • Hitachi has developed an 8 x 8-bit "single-electron-memory" cell array using 3-nm-thick nanocrystalline silicon structures to control individual electrons by the Coulomb blockade effect. Data-line current changes were detected with the addition ("write") and the removal ("erase") of each of five separate electrons. Eventually, circuits based on this nonvolatile technology could achieve Giga- to Tera-bit densities. Write/erase times would be faster than conventional flash memories simply because the number of electrons moved would be dramatically lower. • IBM researchers presented a 0,5-micron, triple-met-al,1-Mbit, 100-MHz DRAM for cache memory that is intended for use as the second chip in a new microprocessor multichip module being designed by IBM. C4 flip-chip interconnects will handle approximately 1000 interconnects between chip and substrate. The device will also be available in a single-chip package compatible with the Intel Pentium bus. • NEC showed a prototype 60 nano-second 1 -Mbit nonvolatile ferroelectric memory chip. The 2000-A-thick strontium-barium-tantalate (SBT) storage cell is formed by a sol-gel process, and has platinum contacts. The single 34.72 square-|j.m transistor and single 1-fim aluminum CMOS process yields a 90.9-sq-mm chip. Presenter Hiroki Koike of NEC said that second-generation chips could be in production in two to three years. • Researchers from the Georgia Institute of Technology showed novel process and circuit designs to produce "adiabatic" MOS chips for gigascaie integration. These energy-recovery logic circuits are called adiabatic because they rely on the relationship between entropy and heat in a closed system as described by the second law of thermodynamics. ARPA and the SRC sponsored this work, which could lead to high-speed chips that would consume much less power. • Stanford's Center for Integrated Systems has developed a BiCMOS active substrate probe card fabricated using silicon MEMS process techlnology. Air pressure deflects micron-scale polyimide membranes to bring up to 1000 tungsten probe tips - in sets of two - into contact with test pads. Low-voltage current is forced between the two tips on each pad to break the oxide laver, allowing multiple probes without excess physical damage. Active circuitry on the card can be located just 1 cm from the probe tips to improve timing accuracy. Članstvo v Eurolab Slovenija Sredi leta 1992 je iniciativna skupina sedmih predstavnikov preskusnih laboratorijev, Urada za standardizacijo in meroslovje (USM) in Zveze inženirjev in tehnikov Slovenije (ZITS) ustanovila Sekcijo preskusnih laboratorijev pri ZITS, ki predstavlja nacionalno vejo Evropske organizacije za preskušanje (EUROLAB) in je bila na generalni skupščini Eurolab januarja 1993 sprejeta v članstvo kot opazovalka. Slovenija kot opazovalka ni imela nikakršnih obveznosti, s tem pa tudi ne možnosti aktivnega delovanja. V Eurolab je Slovenijo zastopal direktor USM, pri čemer je USM, kolikor je bilo mogoče, tudi populariziral delo Eurolaba z objavami v Sporočilih in s spodbujanjem udeležbe na prireditvah Eurolaba, kot je bila npr. organizacija udeležbe na Simpoziju Eurolab aprila 1994. Na zadnji skupščini, januarja 1996, pa je bil status Slovenije spremenjen v pridruženo članico. S tem se odpirajo povsem nove možnosti aktivnega sodelovanja, pa tudi seveda obveznost plačevanja članarine. Zato je napočil trenutek, da se Eurolab v Sloveniji vzpostavi kot širše interesno združenje preskusnih, analitskih in kali-bracijskih laboratorijev v Sloveniji. O namenu in pred- nostih, ki jih prinaša članstvo v takšnem združenju, preberite v Izhodiščih za delovanje slovenskega združenja preskusnih in kaiibracijskih laboratorijev, Eurolab Slovenija. Če vas zanima sodelovanje pod navedenimi pogoji, odgovorite, prosimo, na vprašanja iz Vprašalnika za včlanitev v Eurolab Slovenija, ki je objavljen na zadnjih straneh Sporočil, in ga izpolnjenega vrnite na Urad za standardizacijo in meroslovje, z oznako "EUROLAB Slovenija". V novembru 1996 (predvidoma 7. novembra 1996) nameravamo organizirati skupščino vseh laboratorijev, ki se 'elijo včlaniti v Eurolab. Skupaj s skupščino bo potekal tudi seminar z naslovom: Zagotavljanje kakovosti v laboratorijih, ki bo obravnaval praktične vidike uvajanja sistema kakovosti v preskusne, analitske in kalibracijske laboratorije. Zoran Svetik, predsednik pripravljalnega odbora EUROLAB Slovenije, Slovenski institut za kakovost in meroslovje, Tržaška c. 2, Ljubljana 134 Informacije MIDEM 26(1996)2, Ljubljana KOLEDAR PRIREDITEV 1996 JULY 03.07.-05.07.1996 8th INTERNATIONAL CONFERENCE ON METERING AND TARIFFS FOR ENERGY SUPLY Brighton, UK INFO.: + 44 171 240 8830 08.07.-10.07.1996 2nd IEEE INTERNATIONAL ON-LINE TESTING WORKSHOP Biarritz - Saint-Jean-de-Luz,FRANCE Info.: + 33 76 57 46 19 21.07.-24.07.1996 29th ANNUAL CONVENTION "MICROSTRUCTURE:KEY TO ADVANCES IN MATERIALS" Pittsburgh, Pennsylvania, USA Info.: + 1 412 476 5883 AUGUST 05.08.-09.08.1996 9th INTERNATIONAL CONFERENCE ON MOLECULAR BEAM EPITAXY Malibu,CA,USA Info.: + 1 805 492 7072 31.08.-01.09.1996 7th INTERNATIONAL WORKSHOP ON COMPUTER ANIMATION AND SIMULATION Poitiers,FRANCE Info.: + 41 21 693 52 46 SEPTEMBER 02.09.-05.09.1996 22nd EUROMICRO CONFERENCE Prague, Czech Republic Info.: +44 1232 245 133 05.09.-06.09.1996 5th INTERNATIONAL WORKSHOP ON TIME VARYING IMAGE PROCESSING AND MOVING OBJECT RECOGNITION Florence, ITALY Info.: + 39 55 4796279 11.09.-13.09.1996 RTP'96 THERMAL PROCESSING CONFERENCE Boise, Idaho, USA Info.: +1 512 310 2884 25.09.-27.09.1996 2nd INTERNATIONAL CONFERENCE ON SATELLITE COMMUNICATIONS Moscow, RUSSIA Info.: + 7 95 203 4985 25.09.-27.09.1996 9th INTERNATIONAL SYMPOSIUM ON SYSTEM SYNTHESIS La Jolla,CA,USA Info.: +1 909 787 4710 25.09.-27.09.1996 INTERNATIONAL WORKSHOP ON THERMAL INVESTIGATIONS OF IC's AND MICROSTRUCTURES Budapest, HUNGARY Info.: E mail: Bernard.Courtois imag.fr 25.09.-27.09.1996 24th INTERNATIONAL CONFERENCE ON MICROELECTRONICS MIEL'96 AND 32nd SYMPOSIUM ON DEVICES AND MATERIALS, SD'96 Nova Gorica, SLOVENIJA Info.: + 386 61 312 898 135 Informacije MIDEM 26(1996)2, Ljubljana DRUŠTVO MIDEM IN KONFERENCA MIEL-SD NA INTERNETU Dragi člani društva in bralci revije! Ob pomoči Dejana Križaja s Fakultete za elektrotehniko, Ljubljana, laboratorij za elektronske elemente, smo koncem meseca aprila 1996 postavili dve MIDEM strani na INTERNETU in sicer: 1. Predstavitev društva MIDEM in revije "Informacije MIDEM" na naslovu http://pollux.fer.um-l j-si/MlEL/MIDEM.htm 2. Predstavitev konference MIEL-SD'96 na naslovu http://pollux.fer.uni-lj-si/MIEL/miel96.htm 3. Elektronsko pošto lahko pošiljate na naslov. lztok.Sorli@guest.arnes.si Pri vpisu pazite na velike in male črke!! MIDEM SOCIETY AND MIEL-SD CONFERENCE ON INTERNET Dear readers and Society members! With the help of Mr. Dejan Križaj, Faculty for Electrical Engineering, Ljubljana, Laboratory for Electron Devices, MIDEM Society has, since end of April 1996, two pages on INTERNET: 1. Presentation of MIDEM Society and Journal "Informacije MIDEM", address http://pollux.fer.um-lj-si/MIEL/MIDEM.htm 2. Presentation of the Conference MIEL-SD'96, address http://pollux.fer.uni-lj-si/MIEL/miel96.htm 3. Email can be sent to: lztok.Sorli@guest.arnes.si When inputting the address, please type lower and upper case letters as indicated. 136