Aging of NTC ceramics investigated by magnetic measurements

Abstract

Negative temperature coefficient (NTC) ceramics or thermistors are widely used for temperature sensing and control. These ceramics are mostly based on mixed oxides of Mn, Ni, Fe, Co and Cu which crystallize in a spinel structure. A main problem with these materials is aging. This is a drift in resistivity with time which occurs at elevated temperatures, i.e., 150 ◦C. A number of models have been reported to explain the aging phenonema. Zurbruchen and Case [1] report that besides contact degradation, the degree of inversion (between tetrahedral and octahedra sites in the spinel) changes upon aging. This is concluded from changes in the crystal structure. We have observed similar results in the system Mn2−x In0.1Nix O4 [2]. Feltz [3] reports two models for aging of NTC ceramics. The first is an oxygen uptake or release, the second is also the redistribution of the cations over the sublattices. Mossbauer measurements on Mn3−x Fex O4 (x = 0.58 and 1.05) before and after aging have shown that migration occurs of Fe3+ from the A-site to the B-site [4]. Recently, we reported on aging in the series Mn2.46−x Ni0.54Fex O4 [5]. It was shown that aging is strongly dependent on the iron content. For samples without iron the aging (1R/R) is lower than 1% (after 1000 h at 150 ◦C). When iron is introduced the aging strongly increases to 20% for x = 0.75. However, at higher iron contents the aging decreases again to <1% for x = 1.05. It is also shown that aging can be suppressed when the ceramics, after sintering, are cooled in a controlled atmosphere to prevent oxidation or reduction of the ceramics. During this stoichiometric cooling the partial oxygen pressure is regulated as described in [6]. In this letter the aging of Fe containing nickel manganites is investigated by magnetic permeability measurements for samples which, after sintering, are cooled in air and samples which are stoichiometrically cooled. Two samples have been investigated: Mn1.71Ni0.54Fe0.75O4 and Mn1.41Ni0.54Fe1.05O4. Samples have been prepared from mixtures of Mn2O3, NiO and Fe2O3 in the appropriate ratios. These powders have been ballmilled on a roll-bank and calcined at 800 ◦C for two hours. The calcined powders have been milled again and thereafter been granulated using polyvinylalcohol. From the granulated powders rings or pellets have been pressed. From these rings and pellets a first series has been sintered in air for 6 h at 1250 ◦C and consequently cooled in air. The second series of samples was sintered in air and cooled using a controlled atmosphere (O2/N2) to prevent oxidation of the ceramics. The set-up used for these experiments has been described in [7]. X-ray powder diffraction revealed that all samples are single phase cubic spinels. Aging of the samples has been carried out at 150 ◦C for 1000 h. The results of the aging tests after 1000 h on the resistivity measured at T = 25 ◦C are (see [5]): • Mn1.41Ni0.54Fe1.05O4,aircooled:1R/R0= 13.5% • Mn1.41Ni0.54Fe1.05O4, stoichiometric cooled: 1R/R0= 1.3% • Mn1.71Ni0.54Fe0.75O4, air cooled: 1R/R0= 1.5% • Mn1.71Ni0.54Fe0.75O4, stoichiometric cooled: 1R/R0= 1.0%