J Mater Sci: Mater Electron (2010) 21:861–867 DOI 10.1007/s10854-009-0008-z ‘Lead Free’ thick film thermistors: a study of variation in glass frit concentration and organics composition Shweta Jagtap • Sunit Rane • Suresh Gosavi Dinesh Amalnerkar • Received: 17 August 2009 / Accepted: 27 October 2009 / Published online: 10 November 2009 Ó Springer Science+Business Media, LLC 2009 Abstract There is a need to develop environment friendly electronic materials due to worldwide constraint for the use of hazardous chemicals such as lead, cadmium which are the main constituents of the electronic components. Considering such need, the work on lead free thick film thermistors has been initiated. Thick film materials have been proved to possess economical processing and functional advantage over other technologies and quick turnaround production of hybrid microcircuits. Thick film thermistors are different from the conventional thermistors (bulk) not only by the preparation process but also in composition, transport properties etc. In this paper, we have specifically focused on the influence of glass frit content and organic composition on the properties of lead free spinel based NTC thermistors. We noted that the glass frit concentration is responsible for the change in physical as well as electrical properties of the thick film thermistor. However, the type of organic vehicle (i.e., composition) did not show any adverse effect on microstructure and electrical properties of the thermistor. S. Jagtap  S. Rane (&)  D. Amalnerkar Thick Film Materials Laboratory, Centre for Materials for Electronics Technology, Panchawati, Off. Dr. Bhabha Road, Pune 411008, India e-mail: sunitrane@yahoo.com S. Jagtap Department of Electronic Science, University of Pune, Pune 411007, India S. Gosavi Department of Physics, University of Pune, Pune 411007, India e-mail: swg@physics.unipune.ernet.in 1 Introduction A negative temperature co-efficient (NTC) thermistor is a ‘thermally sensitive resistor’ whose primary function is to exhibit decrease in resistance with increase in temperature. Normally, NTC thermistor materials were prepared by reaction of binary or ternary system of transition metal oxides such as Mn, Co, Ni, Cu, Fe etc. with spinel structure of the general formula AB2O4 [1–5], which is the first choice of many manufacturers owing to their high thermistor constant [6]. For these NTC materials, the dependence of specific resistance versus temperature is described by, q ¼ q0  expðb=T Þ where, q0 is the resistivity at infinite temperature. Resistivity at infinite temperature is determined by the total number of ‘B’ lattice sites that can take part in the hopping process, T is the temperature and b is the thermistor constant which is defined as the ratio between the activation energy for electrical conduction and the Boltzmann constant [7–9]. There are different technologies used for the preparation of thermistors such as bulk, thin film and thick film. Thick films have been used to manufacture hybrid integrated circuits, networks and components for all segments of electronics industry such as military, data processing, medical, automotive, telecommunications and consumer. Thick film thermistors are different from the conventional thermistors not only by the preparation process but also in composition, transport properties. Thick film thermistors are dispersions of spinel powders along with glassy binder in organic fluid vehicles. Thick film thermistors are normally prepared in the chip form due to ease of fabrications, processing and cost effective technique; widely used for measuring the average temperature of a surface and model 123 862 plates for evolution of thermal conductions of fluids. Thick film thermistor paste primarily consists of three constituent’s namely semiconducting oxide powder, glass frit, and organic vehicle. The important role of glass frit in thick film pastes is to promote sintering of the semiconducting oxide powder during firing and enable binding of the metal film to the substrate. Similarly an organic vehicle which is a temporary binder disperses the functional material and binder component to impart the desired rheological properties to the paste. However, thick film paste is a complex non-equilibrium system and the constituent as well as percentage of both organic vehicle and glass frit plays an important role in deciding the physical, electrical and microstructural properties of the film. Similarly processing conditions also have great effect on deciding the above mentioned properties. Normally, these thick film pastes are screen-printed in desired pattern on ceramic substrates, then dried and fired at high temperature. During firing, complex reactions occur between the paste constituents which sometimes lead to new crystallization of glassy matrix [10]. Considering the concerns of environmental problems associated with the use of hazardous chemicals such as lead, cadmium (which are the main constituents of the thick film thermistors), we have initiated the work on ‘lead free’ thick film NTC thermistors. In this paper, we present the influence of glass frit content and organic composition on the properties of lead free spinel based NTC thermistors. 2 Experimental Mn1.85Ni0.8Co0.35O4 thermistor composition was prepared by ceramic technique using oxides of Mn, Co, Ni. The precursor oxide powder mixture was sintered at 1000 °C for 4 h at a ramp rate of 10 °C/min. The detailed powder synthesis and processing parameters have been elaborated in our earlier communication [11]. Thick film pastes were formulated by mixing of spinel powder, glass frit and an organic vehicle. Thick film thermistor prepared using the spinel powder shows very high resistance and non-linear electrical properties. Therefore, to achieve the lower resistance value and linear behavior, normally precious metals such as silver, copper, ruthenium dioxide are being added. In this work, we have added 42% RuO2, as a conducting media (out of the total functional phase of the paste) to lower down the resistance value. This specific percentage of RuO2 was optimized based on our previous experimentation. Here, two types of pastes were formulated: (1) by variation in glass frit content and (2) by variation in the organic vehicle composition. The glass frit is prepared indigenously and the composition is based on oxides of calcium- barium-alumino borosilicate. 123 J Mater Sci: Mater Electron (2010) 21:861–867 2.1 Variation in lead free glass content Thick film thermistor pastes were prepared by dispersing the spinel thermistor powder material (300–500 nm size), RuO2 powder and lead free glass frit in an organic vehicle. The organic vehicle which is denoted as vehicle ‘A’ is a mixture of ethyl cellulose and 2,2-butoxy ethoxy ethyl acetate. The ratio of inorganic phase and organic phase was kept at 70:30. Different sets of pastes were prepared by changing the glass frit content by keeping all other constituents and parameters of the paste constant. The details of samples prepared and the data related to these paste formulations is given in Table 1. 2.2 Change in organic vehicle composition In this part of experiment, three sets of thermistor pastes were prepared using three types of organic vehicle. The vehicle compositions prepared by changing its ingredients were denoted as vehicle ‘A’, vehicle ‘AA’ and vehicle ‘B’. Generally, in thick film pastes, the organic part consists of ethyl cellulose (8%) and the solvent (92%). However, the formulation of organic vehicles used in thick films embodies a great deal of experience and expertise, and is usually proprietary information of the paste manufacturers. In the present work, Vehicle ‘A’ is a mixture of ethyl cellulose and 2,2-butoxy ethoxy ethyl acetate, vehicle ‘B’ is a slight modification of vehicle ‘A’ in which the amount of 2,2-butoxy ethoxy ethyl acetate increased by 2% by reducing the equivalent amount of ethyl cellulose where as vehicle AA is the mixture of ethyl cellulose and solvents such as butyl cellosolve, butyl carbitol acetate and b-terpinol. Here, the inorganic phase i.e. functional phase and glass frit was kept constant and only the composition (i.e. type) of organic vehicle was changed. Here, it must be noted that all the paste formulation conditions were kept similar to that of Sect. 2.1 except variation in the composition of organic vehicle. The details of samples prepared are given in Table 2. In both the cases, we have formulated the pastes at 5 g scale, therefore the viscosity of our pastes was subjective Table 1 Details of samples prepared by changing the glass frit content Sample code Inorganic phase Functional phase (Spinel material ? RuO2 (%) Glass (%) Organic vehicle TA1C5A 95 5 A TA1C10A TA1C15A 90 85 10 15 A A J Mater Sci: Mater Electron (2010) 21:861–867 863 Table 2 Details of thick film paste compositions prepared by changing the organic vehicles Inorganic phase (70 wt. %) NiCo2O4 RuO2 TA1C15A Organic vehicle (wt. 30%) Functional phase Glass frit TA1C10A 90 10 A TA1C10B 90 10 B TA1C10AA 90 10 AA Intensity (Arb units) Sample code CoMn2O4 NiMn2O4 and defined considering the parameters viz. screen printability, thixotropy, spreading, drying conditions. Screenprinted planar thick film thermistor patterns of size 2 mm2 were produced on pre-fired lead free silver electrodes on 96% alumina substrate. After screen-printing, the films were dried under IR lamp for 10–15 min and then the films were fired at 850 °C at dwell of 10 min in BTU furnace with 60 min firing profile. The thickness of the fired thermistor films was measured by using Talysurf thickness profiler and it was found to be 25 ± 2 lm. X-ray diffraction (XRD) data of the films were taken on an automated powder X-ray diffractometer system (Model- AXS D8 Advance, Bruker) in the 2h range of 20–80° with step size of 0.1° using a copper target. The structural analysis along with detection of impurity phases has been carried out using the XRD data. The microstructure/surface morphology of the fired thermistor films was observed under scanning electron microscope (SEM) with EDAX attachment (JEOL make). The electrical resistance of thermistor films was measured in the range from 25–300 °C using a 4‘ digit digital multimeter (Model- U1252A, Agilent make). TA1C10A TA1C5A 20 30 40 50 60 70 80 2θ Fig. 1 X-ray diffractogram of the fired thermistor thick films with different glass percentage percentage in the thick films. The important observations are as follows: a. The presence of spinel phases such as NiMn2O4 (average intensity 25%), CoMn2O4 (average intensity 20%) and NiCo2O4 (average intensity 13%), was detected in sample containing 5% glass frit (TA1C5A). Besides, the significant presence of RuO2 phase with the average peak intensity of 50% was observed. b. The significant presence of spinel phase such as NiMn2O4 (average intensity 50%) along with considerable presence of RuO2 (average intensity 43%) was observed in sample containing 10% glass frit (TA1C10A). Also feeble presence of spinel phases such as CoMn2O4 (average intensity 5%) and NiCo2O4 (average intensity 8%) has been noted. c. Sample containing 15% glass frit (TA1C15A) has dominant presence of spinel phase NiMn2O4 (average peak intensity 55%) with appreciable presence of RuO2 (average intensity 24%). Also feeble presence of spinel phases such as CoMn2O4 (average intensity 11%) and NiCo2O4 (average intensity 9%) was observed. 3 Results and discussion 3.1 Influence of glass frit concentration Figure 1 shows the X-ray diffractogram of thermistor films containing different glass percentage. Variation in the average peak intensity corresponding to different phases and d values is given in Table 3. The variations in peak intensity of the phases are co-related with the glass The microstructure of the fired thick film thermistors was observed under scanning electron microscope (SEM) Table 3 Data corresponding to d values and variation in average peak intensity towards different spinel phases Sample code/‘d’ value (Ao) NiMn2O4 CoMn2O4 NiCo2O4 RuO2 2.53 2.09 1.46 Avg. (%) 2.48 2.71 1.60 Avg. (%) 2.44 1.43 1.57 Avg. (%) 2.55 3.18 1.68 Avg. (%) TA1C5A – 50 26 25 61 – – 20 – – 40 13 52 67 36 50 TA1C10A 100 25 30 50 – – 13 5 – 12 12 8 – 90 40 43 TA1C15A 100 50 14 55 26 – 5 11 – 12 14 9 – 38 33 24 123 864 J Mater Sci: Mater Electron (2010) 21:861–867 Table 4 Data on the electrical parameters such as sheet resistance, thermistor behavior thermistor constant Resistance (ohm) (K ohm) with the energy dispersive X-ray analysis (EDAX) attachments. The SEM photograph of the representative sample containing 15% (TA1C15A) is shown in Fig. 2. It is seen that the conducting grains are well dispersed in the glassy matrix. The EDAX of the fired films reveals that as expected there was significant presence of the conducting phase i.e., RuO2 in samples containing 5 and 10% glass frit (TA1C5A and TA1C10A) samples with the mass percent of 46.67 and 42.72, respectively whereas the sample with 15% glass frit (TA1C15A), the mass percentage of RuO2 was 28. The resistance of the film samples was measured as a function of temperature. Fig. 3 shows the change of resistance with respect to change in the temperature. It is seen from the figure that the samples with 5% glass frit shows PTC (ohm) Sheet resistance (X/h) Thermistor behavior b (K) TA1C5A 60 PTC/NTC -600/370 TA1C10A 1K NTC 1258 TA1C15A 450 K NTC 3930 nature up to 165 °C and then it switches to NTC behavior. However, the samples with 10 and 15% glass frit show only NTC behavior. The resistance decreases exponentially with respect to increase in temperature in case of sample with 15% glass frit where as the sample containing 10% glass frit shows quasi-linear nature between the resistance and temperature. The data on the electrical parameters such as sheet resistance, thermistor constant, thermistor behavior are given in Table 4. The thermistor constant (b) was calculated by the following equation [12–14]: Fig. 2 SEM photograph corresponding to fired thick film thermistor (sample TA1C15A) Fig. 3 Behavior of thermistor as a function of temperature Sample code b25300 ¼ lnðR25 =R300 Þ ð1=T25 Þ  ð1=T300 Þ where R25 and R300 are the resistances measured at 25 °C and 300 °C, respectively. 3.2 Influence of organic vehicle composition As mentioned in the text above, the effect of organic vehicle composition on the thermistor properties was also studied. In addition to the glass frit, the organics used for the formulation of thick film pastes also plays an important role especially in deciding the microstructure of the final film. Organic vehicle is a generic term describing a blend 500 400 300 200 100 0 NTC 1000 800 600 400 200 0 85 80 75 70 65 60 55 50 TA1C10A NTC TA1C5A PTC 0 50 NTC 100 150 200 Temperature ( °C ) 123 TA1C15A 250 300 J Mater Sci: Mater Electron (2010) 21:861–867 of some volatile solvent(s) and polymers or resins which are needed to provide an homogeneous suspension of the particles of the functional materials(s) and a rheology suitable for printing of the film configuration; hence the vehicle is a temporary sacrificial ingredient, which is removed completely in the process during which the microstructure of the deposits is formed. The composition of the organic vehicle is responsible for, or contributes at determining the shelf life of the paste, its drying rate on the screen, the change of printability with ambient temperature, the resolution of line, their cosmetic appearance and some electrical properties of the fired films [15]. Rane et.al [16] reported that the chemical compositions of the organics affect the physical properties of the lead free glasses and stated that composition of organic vehicle is responsible the devitrification of glass. The formation of small single crystals of RuO2 due to the reactions between ruthenate and the organic vehicle when the ruthenate based thick film resistors fired at 500 °C reported by Hrovat et al. [17]. They also interpreted that the after complete burning of organic vehicle the metallic ruthenium turns to RuO2 and at again at higher firing temperature formed the pyrochlore ruthenate. The backscattered SEM images corresponding to thermistor film with variation of organic vehicle composition are shown in the Fig. 4. The SEM image reveals more or less denser and smoother surface with connected grain boundaries in all the cases in addition to some pores that are observed on the surface. This is normal tendency in thick films as the organics involved in the thick film paste formulation burns out during firing 865 which leads to evolution of trapped gases. The oxide particles are grown due to the firing of the film which gets homogeneously dispersed in the glassy matrix. However, we did not note any adverse effect as observed by the above researchers [16, 17] on the thick film thermistor due to the composition of organic vehicles. The surface morphology of the thick film thermistors prepared using vehicle ‘A’ and vehicle ‘B’ seems to be almost similar (Fig. 4i, ii), which is expected, since the constituents of the organic vehicles are same except there is slight variation in their composition. However, the cavities observed on the surface of the thick film thermistors prepared using vehicle ‘AA’ (Fig. 4iii) are greater is size as compared to that of other two cases. This may be due to the different wetting conditions of dispersed solid particles by the vehicle since the polarity as well as the mechanical strength of the vehicle might be different as different solvents are used. The degradation of polymer and flow decomposition products during binder burnout represents a coupled heat transfer, mass transfer and reaction kinetics problem. The generation of unequal pressures of gas phase decomposition products within the film leads to internal stresses within the ‘green’ body of ceramics is a problem of continuous network since the concentration of binder is changing spatially and temporarily during the heating cycle. Thus, the material properties are also changing the overall coupling between the reaction transport and stress effects becomes more complex [18]. The similar phenomenon may also exist in our films. Fig. 4 SEM photographs of the thick film thermistors with different organic vehicles: (i) TA1C10A, (ii) TA1C10B and (iii) TA1C10AA 123 866 J Mater Sci: Mater Electron (2010) 21:861–867 Table 5 Data on the electrical parameters such as sheet resistance, thermistor behavior and thermistor constant corresponding to thick film thermistors with different organic vehicle Sample code Sheet resistance (KX/h) ± 10% Thermistor behavior b (K) TA1C10A 1 NTC 1258 TA1C10B 1 NTC 1265 TA1C10AA 1 NTC 1270 The values of sheet resistance thermistor constant and thermistor behaviour of the film samples are mentioned in Table 5. It is seen from the table that the sheet resistance of all the films is 1 kX/h. From this table, it was noted that sheet resistance and thermistor constant of all the film samples remain constant with variation in composition of organic vehicle even though there is a small variation noted in the microstructure. When the resistance of the film samples was measured as a function of temperature, it showed the quasi-linear nature (i.e., same as TA1C10A sample, which is shown in Fig. 3). Therefore, we can say that thermistor films do not show any adverse result for these variable organic compositions. 4 Conclusion Our results show that the amount (%) of glass frit in the thick film paste plays an important role in deciding the sheet resistance and hence the thermistor constant of the films. The variations in glass percentage were reflected in XRD, SEM/EDAX results as well on the electrical properties (sheet resistance and thermistor constant) of the thermistors. The composition that contains 5% glass frit (TA1C5A) shows dominance of RuO2 phase with the considerable presence of the spinel phases such as NiMn2O4, CoMn2O4, NiCo2O4 and a sheet resistance (60X/h) with dual behavior i.e., PTC-NTC. In other words, the composition, initially exhibits PTC nature and then it translates into NTC nature. The sample, TA1C10A (10% glass frit) shows almost equal presence of NiMn2O4 and RuO2 phases with small appearance of other spinel phases such as CoMn2O4 and NiCo2O4 with NTC behavior and sheet resistance value 1 KX/h. There is dominance of spinel phase in case of TA1C15A composition which also shows dominance of NiMn2O4 phase with markable presence of RuO2, and hence the sheet resistance is higher (450 KX/h) as compared to that of the samples TA1C5A and TA1C10A. The change in organic vehicle composition did not show any affect on sheet resistance as well as thermistor constant. Also the microstructure corresponding to samples with different organic vehicle show more or less denser and 123 smoother surface morphology with connected grain boundaries in all the cases. This confirms that the microstructural properties of our thick film thermistors are not sensitive to the organic composition. Acknowledgments The work was supported by the Department of Information Technology, Ministry of Communication and Information Technology, Govt. of India through the sponsored project. The authors are grateful to Dr. U. P. Phadke, Dr. Krishnakumar, Dr. S. Chatterjee, Department of Information Technology, Ministry of Communication and Information Technology, New Delhi for their active support related to the project. References 1. O. Shpotyuk, A. Kovalskiy, O. Mrooz, L. Shpotyuk, V. Pechnyo, S. Volkov, Technological modification of spinel based CuxNi1-x-y Co2Mn2-yO4 ceramics. J. Eur. Ceram. Soc. 21, 2067–2070 (2001) 2. R. Metz, Electrical properties NTC thermistors made of manganite ceramics of general spinel structure. J. Mat. Sci. 35, 4705– 4711 (2000) 3. Y. Abe, T. Meguro, T. Yokoyama, T. Morita, J. Tatami, K. Komeya, Electrical properties of sintered bodies composed of a monophase cubic spinel structure Mn(1.5-0.5x)Co(1?0.5x)Ni0.5O4 (0 B x B 1). J. Ceram. Process. Res. 4, 140–144 (2003) 4. J.L. Martin De Vidales, P. Garcia-Chain, R.M. Rojas, E. Vila, O. Garcia-Martinez, Preparation and characterization of spinel type Mn-Ni-Co-O negative temperature co-efficient ceramic thermistors. J. Mat. Sci. 33, 1491–1496 (1998) 5. Z. Wang, C. Zhao, P. Yang, A. Winnubst, C. Chen, X-ray diffraction and infrared spectra studies of FexMn2.34-xNi0.66O4 (0 \ x\1) NTC ceramics. J. Eur. Ceram. Soc. 26, 2833–2837 (2006) 6. R. Kamat, G. Naik, Thermistors- in search of new applications manufacturers cultivate advanced NTC techniques. Sens. Rev. 22, 334–340 (2002) 7. K. Park, D. Bang, Electrical properties of Ni-Mn-Co-(Fe) oxide thick film NTC thermistors prepared by screen printing. J Mat. Sci.: Mat. Electron. 14, 81–87 (2003) 8. S. Kanade, V. Puri, Composition dependent resistivity of thick film Ni1-xCoxMn2O4 (0 B x B 1) NTC thermistors. Mater. Lett. 60, 1428–1431 (2006) 9. N. Vittal, G. Shrinivasan, C.R. Aiyer, R.N. Karekar, Correlation between X-ray diffraction studies and conductivity dependence of Ag loading in thick film thermistors. J. Appl. Phys. 68, 1940– 1943 (1990) 10. Arima, H.: Thick film thermistors and RTDs. In: Prudenziati, M., (ed.) Handbook of thick film sensors, vol. 1, pp. 127–150, Elsevier, (1994) 11. S. Jagtap, S. Rane, S. Gosavi, D. Amalnerkar, Preparation, characterization and electrical properties of spinel type environment friendly thick film NTC thermistors. J. Eur. Ceram. Soc. 28, 2501–2507 (2008) 12. Macklen, E.D.: Thermistors, Electrochemical Publications Ltd. (1979) 13. S. Jagtap, S. Rane, S. Gosavi, D. Amalnerkar, Ruthenium dioxide doped manganite-based NTC thermistors for low-resistance applications. Microelectron. International 25, 19–23 (2009) 14. Jagtap, S., Rane, S., Gosavi, S., Amalnerkar, D.: Low temperature synthesis and characterization of NTC powder and its ‘lead free’ thick film thermistors. Microelectron. Eng. (in press) doi:10.1016/ j.mee.2009.05.026, (2009) J Mater Sci: Mater Electron (2010) 21:861–867 15. Prudenziati, M.: Paste, inks and slurries. In: Prudenziati, M., (ed.) Handbook of thick film sensors, vol. 1, 1994, pp. 113–124, Thick Film Sensors, Elsevier, (1994) 16. Rane, S., Prudenziati, M., Morten, B.: Organic vehicle effects on devitrification of lead free glasses used in thick film technology. Proceeding in emerging microelectronics & interconnection technologies (EMIT-2004) IMAPS, Bangalore, India, 27–28 Nov 2004, 25–29, (2004) 867 17. M. Hrovat, Z. Samardzija, J. Holc, D. Belavic, The development of microstructural and electrical characteristics in some thick film resistors during firing. J. Mat. Sci. 37, 2331–2339 (2002) 18. B. Peters, S.J. Lombardo, Optimization of multi-layer ceramic capacitor geometry for maximum yield during binder burnout. J. Mat. Sci. Mat. Electron. 12, 403–409 (2001) 123