EFFECT OF THE ADDITION OF DIFFERENT METAL OXIDES IN LEAD BOROSILICATE GLASSES ON THE ELECTRICAL CHARACTERISTICS OF SbSn COMPOSITION-BASED THICK-FILM RESISTORS

The paper reports the effect of addition of metal oxides in lead borosilicate glasses on electrical characteristics of SbSn alloy-based thick-film resistors. The Sb and Sn powder (1" by weight) is taken in two quartz tubes separately, vacuum sealed at 10 -Torr and heated in a resistive furnace .at 430C and 630C respectively. The conventional glass [1] is modified by using different dopants like tungsten oxide, cobalt oxide, lithium oxide, titanium dioxide, venadium pentoxide, chromium oxide, nickel oxide and manganese dioxide. The resistive pastes are formulated with both powders, 5% glass and conventional organic binder. The firing temperature is optimized for eight glasses. The sheet resistivity varies from 1600 /I--1 to 40 O/I-q, with negative temperature co-efficient of resistance varying from 2000 ppm/C to 800 ppm/C respectively. Material characterization is carried out using the XRD technique. Aging studies of resistors at room temperature over the period of two months indicate that these resistors stabilize within 15-20 days.


INTRODUCTION
Commercially available thick film pastes are costly because of the use of Ruthenium oxide and Ruthenates.In order to get low cost resistive pastes, S.H. Bhide [1] et al and M.R. Kadam [2] et al tried to formulate SnO2 pastes.Kattimani  [3] et al have reported the effect of loading of an SbSn alloy in SnO2.In all these low-cost pastes, conventional borosilicate glass is used.The TCR is of the order of few thousands ppm/C.
In the present paper, a study of the effect of addition of metal oxides in lead borosilicate glasses on electrical characteristics (TCR and p) of SbSn alloy-based thick-film resistors is reported.

EXPERIMENTAL
The Sb and Sn powders are taken in 1" 1 proportion by weight and ball milled for 24 hours to get a homogeneous mixture in an acetone medium.The powder is taken in two quartz tubes separately.The tubes are vacuum sealed at 10 -5 Torr and heated in a resistive furnace to 430C and at 630C, respectively, for 3 hours.After heating the quartz tube is taken out and broken carefully to get the powder, which is ball milled for 24 hours to get a particle size of about 4-5 microns.The SbSn powders at 430C (I) and 630C (II) are used throughout the experimentation.
The SbSn resistor paste is formulated by having 95 % of SbSn powder and 5% glass as the solid phase and ethyl cellulose and butyl carbitol acetate (BCA) as an organic vehicle used as a temporary binder.The paste is prepared keeping the solid-to- liquid ratio 70:30 by weight.The liquid phase consists of 8% ethyl cellulose and 92% BCA.
The roperties of the glass control the film properties to a great extent.Here, lead oxide-based borosilicate glass is selected.The composition [4] of the glass is 70% PbO, 18% .SiO_, 9% A1203 and 3% B203 by weight.The ingredient oxides used are chemically 99.9% pure.The component oxides are mixed by wet milling in a ball mill using acetone as a medium.Mixing is carried out for nearly one hour to get uniform mixing of all oxides.The cake is dried and transferred into the platinum or nickel crucible and heated with a mixture of cooking LPG gas + O,.
The molten oxide mixture is quickly poured into distilled water in a process called fritting.The glass lumps and frits are powdered in an agate morter pestle and filtered using a fine mesh (140 mesh) to get fine powder.The powder is ball milled in a stainless steel jar with stainless steel balls for 48 hours.
Eleven different glasses are made in addition to the conventional glass as shown in Table I.Fourteen pastes with glasses of set I and 10 pastes from set II are formulated and their electrical characteristics are studied.The thick film resistors are fabricated using the standard technique [2].The resistors are fired between 450  (% composition)  to 700C to get good adhesion and stable resistance.At 650C the TCR and p is minimum and the adhesion is good as tested by adhesion tape.Therefore, optimized firing temperature is 650C.
All the resistive pastes are found to have good printability and thixotropic properties.The resistances are measured and the sheet resistivity of each resistor is calculated [3].TCR measurement is carried out in the range of 30C to 130C.The temperatures are measured using Cromel Alumel thermocouple with the help of temperature controller.Material characterization is done for all the samples by XRD in the range of 20 20 to 100 , keeping all parameters constant.The peaks observed in X-ray spectra of SbSn I and II powders with various glasses are iden- tified for possible chemical composition with standard X-ray diffraction data [5] (d values agreeing with _+0.003A).Aging characteristics are studied over the period of two months at room temperature.

RESULTS
For optimization of firing temperature (Fig. 1 and 2), the pastes having the com- position of 95% SbSn powder II (630C) and 5% glasses are used.Here, 95 % SbSn powder II is doped with metal oxide WO3, TiO, Cr203, NiO and MnO2.The resistors are fired at 450C.Since the adhesion is found to be poor in this case, firing is carried out at 500C, 550C, 600C, 650C and 700C.At 650C, good adhesion is obtained as tested with Scotch (tixo tape).Also ps and TCR are min- imum, so a firing temperature of 650C is opted throughout the experiment.Graph of TCR and p versus firing temperature are given in Fig. 1 and 2. After printing and firing at 650C, the p and TCR of the 24 samples are measured as given in Table II.
Aging studies for about 2 months (Fig. 3) at room temperature indicate that the resistors stabilize within 15-20 days.

DISCUSSION OF RESULTS
It has been reported [3] that the firing temperature of 650C for the powder II with G gives better adhesion, though not minimum I and TCR.The composition of the paste was 25% SnO2, 70% SbSn, and 5% glass, and the firing range was 500- 750C.In the present case the optimized temperature is 650C.
It has also been reported [1] that the optimum firing temperature is 550C, for 90% SnO2 doped by 5% Sb and 5% glass in the range of 450C to 600C.M.R. Kadam et al have reported [2] the optimum firing temperatures of 725C, for the pastes having composition of 45% SnCI2, 45% SnO2, 5% Sb, and 5% glass in the range of 550-775C.
In the above two cases, functional material has reacted with the glass, which is obvious from the peaks of PbSnO2 obtained in XRD.In the present case, SbSn and glass are the only functional materials.In conclusion, it could be said that  functional material plays an important role in determining the optimized firing temperature even if the same glass is used.
Table III shows that glass is playing an important role in determining the p, and TCR of the thick film resistors, though glass is not reacting with the functional material.When glasses are doped with 5% metal oxide as a replacement of PbO, which is the base material for the glass, it affects ps and TCR.For SbSn powder no.II, the trend of Ps and TCR for the same glasses is towards the higher side.
This can be explained with the help of XRD patterns of the pastes in set I. For powder II, conduction is dominated by SnO.One of the reasons for having the lower resistivity for first SbSn powder is the decreasing trend of Sb205 intensities.
Zero intensities in XRD patterns can be attributed to the distructive interference [6].
The possible chemical reactions for the SbSn-based pastes can be explained on the basis of retention of carbon in the thick film resistors.The retention of carbon is observed in XRD patterns, which is not due to the improper firing.The liquid phase of the temporary binder EC and BCA are the long chain compounds.When they are heated up to 650C, the carbon is retained, which can also be revealed by the thermogravimetric analysis of the pastes.The TGA shows the actual loss to be less than the expected loss [7].The XRD of the alloys show the dominant phase of SbO (100%) intensity.The retention of carbon supports the reduction of SnO2 when subjected to the firing cycle.SnO2 + C 650C)SnO + CO ' It is found from the XRD results that during the firing process, part of SbSn gets decomposed into Sb and Sn.Part of Sb gets converted into Sb203 and Sb205 and Sn gets converted into SnO and SnO2.The possible reaction mechanism during the firing cycle can be written as Dominant phases of SnO are shown by the XRD's of SbSn I. SnO2 is reduced to SnO as explained above.In case of powder II, dominant peaks are of SbxOy.In the process of firing, Sb oxides (Sb203 and Sb205) are formed.For SbSn powder I, Sb205 is showing a sharp decrease with increasing conductivity, which is not that sharp for powder II.This may be the probable reason for increased conductivity for powder I in set I against powder II.Conduction in SbSn powder II is dominated by SnO.For SbSn powder I, conduction is governed by free Sb and SnO (Figure 4 and 5).
In conclusion, one can say that even if dopant metal oxides are of a lower proportion (0.25%) in the total solid, it affects both the resistivity and TCR of the pastes.
Aging characteristics (Fig. 3) show that resistors having a low value of sheet resistivity stabilize very soon.The percentage of aging ranges from 8 to 11% in the case of SbSn powder no.I.In the case of powder no.II, resistors having a low value of sheet resistivity stabilize within 10 days.The resistors having a high value of sheet resistivity stabilize within 22 days.

TABLE Composition
The percentage of aging in this case XRD intensities of SbSn powder [I and II].SbSn POWDER ! (IStSET OF SAMPLE) SbSn POWDER I IzndsET OFSAMPLE)