d Semiannual Phase Report No. 1 LEAD TELLURIDE BONDING AND SEGMENTATION STUDY Covering period August 1, 1967 - January 31, 1968 Contract No. NAS 5-9149 Prepared by Tyco laboratories, Inc. Bear Hill Waltham, Massachusetts 02154 for , National Aeronautics and Space Administration Goddard Space Flight Center Greenbelt, Maryland 20771 Tyco Laboratories, Inc. Bear Hill Waltham, Massachusetts 02154 Lead Telluride Bonding and Segmentation Study Semiannual Phase Report No. 1 Covering period August 1, 1967 - January 31, 1968 Contract No, NAS 5-9149 H. Bates F. Wald M. Weinstein for National Aeronautics and Space Administration Goddard Space Flight Center Greenbelt, Maryland 2077 1 SUMMARY I. Constitutional Studies A. Tin-Telluride-Metal-Systems Accurate lattice constant determinations on SnTe as in- ? corporated in SnTe-Metal alloys revealed changes in the original lattice constant of the starting Sn. 492Te,508 composition. " From these findings it has to be concluded that PbTe-SnTe thermoelectric alloys in contact with metals might change their composition and therefore their electrical properties over long periods of time. Since such conditions could be important for the mechanism of aging of tungsten contacts to 3P material, a more thorough in- vestigation of similar changes in the W-PbTe-SnTe system is indicated. B. Silicon -Germanium -Metal-A lloys Investigations of the interaction of Si-Ge solid solution alloys withTi, Zr, V, Nb, Ta, Cr, Mo, W, Mn, ReandCohaverevealed two basic types of constitution. The first is one in which equilibria between Si-Ge solid solutions and solid s olutions of metal silicides and germanides exist. The elements Ti, Zr, Ta, and Co apparently belong to this category. The second type of system produces a metal silicide in equilibrium with either pure Ge or with the Si-Ge solid solution. Elements behaving in this fashion are W, V, Nb, Cr, Mo, Mn, and Re. It was also found that phosphorous present as a dopant in n-type Si-Ge was concentrated in the M&i2 formed by addition of Mo. This is taken to indicate that metal phosphides (and presumably borides in the p-type material) can form and apparently dissolve in the metal silicide. 11. Segmented Si -Ge -PbTe Thermocouples Measurements of the efficiency of segmented couples have shown efficiencies of approximately 9% at hot junction temperatures of 800" and above, and power outputs of 2.25 to 2. 8 watts per couple. i 111. Pore Migration Quantitative metallography has shown evidence for fairly extensive migration of pores in pressed-and-sintered Pb. $h 6Te thermoelements after operation for 3800 hours in a temperature gradient from 510" to 50°C. IV. Life Testing A. Isothermal Testing Tungsten bonds to 3P elements have been tested to 6000 hours at 525°C. One half of the bonds remained intact; those measured 2 showed an increase to 95 fl - cm contact resistivity from an 2 initial value of 16 pi2 - cm . B. Gradient -Testing of W -Bonded 3N -3P Couples Six of fourteen bonded couples tested to 800 hours at a hot junction temperature of approximately 520°C survived approximately 35 unscheduled thermal cycles with unchanged total resistance. C. Gradient -Testing of W -Bonded 3P Elements Post -tes t analysis of gradient tested elements showed decreased resistivity and Seebeck voltage at hot ends of elements and precipitation of extensive amounts of an unidentified phase at intermediate and high temperatures. ii CONTENTS Page No, SUMMARY i LIST OF ILLUSTRATIONS V LIST OF TABLES viii I. INTRODUCTION 1 11, CONSTITUTIONA L STUDIES 3 A. Systems Containing Tin Telluride 3 1. Introduction 3 2. Experimental Results 5 3. Discussion and Conclusions 9 B. Systems Containing Silicon-Germanium 9 1. Introduction 9 2. Experiments and Results 11 a. Electron Microprobe Analysis 11 b. Phase Equilibrium Studies 12 c. Conclusions 13 C. Investi ations on the Tun sten Diffusion Bond 14 Using Ecanning Electron hicroscopy 111. SEGMENTED Si -Ge-PbTe THERMOCOUPLES 19 A. Introduction 19 B. Measurements 21 C. Results 26 iii CONTENTS (cont. ) Page No. IV. PORE MIGRATION IN SINTERED PbTe-SnTe 29 THERMOELECTRIC ELEMENTS A. Introduction 29 B. Experimental Procedures 31 C. Results 34 V. LIFE TESTING 41 A. Isothermal Tests 41 B. Couple Life Testing 44 1. PbTe Couple Test No. 1 47 2. PbTe Couple Test No. 2 52 C. Gradient Life Test IV 54 1. Results 54 2. Post-Test Analysis of Elements 60 a. Room Temperature Resistivity 60 b. Seebeck Voltage 62 e. Microstructural 62 VI. REFERENCES 71 iv LIST OF ILLUSTRATIONS Fig. No. Page No. 1 Scanning electron micrograph of p-type (3P) tungsten 15 diffusion bonded specimen. Magnification: 720 X 2 Scanning electron micrograph of p-type (3P) tungsten 16 diffusionbonded specimen. Magnification: 2900 X 3 Scanning electron micrograph of p -type (3P) tungsten 17 diffusion bonded specimen. Magnification: 6200 X 4 Scanning electron micrograph of p-type (3P) tungsten 18 diffusion bonded specimen. Magnification: 14, 500 X 5 Plan view and cross section of Si-Ge-PbTe couple for 20 testing and efficiency measurements. 6 Theoretical efficiency of PbTe, Si -Ge, and segmented 22 Si -Ge -PbTe. 7 Schematic diagram of apparatus for measurement of 23 conversion efficiency. 8 Calculated and n-e asured current -voltage -power 27 relationships for Si -Ge -PbTe operating between 800" and 55°C. 9 Calculated migration rate as a function of temperature 30 for four pore sizes. 10 Schematic curves illustrating the distribution of pores 33 over the len th of an element held in a gradient at (a) zero time (b5 after initial migration (c) steady-state dis - tribution after extended operation. 11 Distribution of pores as a function of length in as- 35 bonded elements. Includes pores from 10 -35 microns in diameter. 12 Distribution of pores as a function of length in gradient- 36 tested elements, held 3850 hours at TH = 510°, TC = 60°C, hot end at left. 13 Frequency distribution of pores at opposite ends of 37 as -bonded elements; includes pores from 10-55 microns in diameter. V LIST OF ILLUSTRATIONS (cont. ) Fig, No. Page No, 14 Frequency distribution of pores at hot end cold ends 38 c of gradient-tested elements; includes pores from 10 - 55 microns in diameter. 15 Macrophotograph of fractured elements and their 45 electrodes. Elements isothermally tested 6000 hours at 525°C; total number of temperature cycles approximately seven. 16 Results of life test of fourteen W-bonded 3N-3P couples. 50 Broken lines represent period of uncontrolled thermal cycling. 17 Life test data on total resistance, Seebeck voltage, and 55 AT for three unbonded 3P elements; average TH = 510°C. 18 Life test data on total resistance, Seebeck voltage and 56 AT for three W-bonded 3P elements; average TH = 510°C. 19 Life test data on Seebeck voltage and AT for single and 57 doubly bonded 3P elements; average TH = 510°C. 20 Variation of room temperature resistivity with distance 61 from hot end for gradient-tested unbonded and W-bonded 3P elements; average TH = 510°C for 3850 hours. 21 Examples of gross cracking of gradient -life tested 64 elements .. 14X 22 Examples of microcracking found generally in 65 gradient -tested elements. 23 (a) Manganase oxide particles near hot junction of 67 gradient -tested 3P element, plus unknown second phase-medium gray. 150X (b) Manganese oxide particles and unknown phase- 67 same as above. 750X 24 (a) Transition area into band of precipitation of un- 68 known phase in gradient-tested elements. 70X (b) Particles of unknown phqse precipitated in 68 gradient-tested 3P elements. 750X 25 Area within unknown precipitate zone showing poorly 69 defined morphology of particles. 750X vi LIST OF TABLES Table No. Page No. I L tti e Parameters of Various Compositions of Tin 6 Telluride I1 Lattice Parameters of Tin Te lluride in Alloys with 8 Meta 1s I11 Measured Performance of Segmented Si -Ge -PbTe 28 IV Actual Median Diameters of Pore Size Categories 40 Measured V Contact and Element Resistances of W-Bonded 3P 42 Elements Tested 6000 hours at 525°C VI Summary of Average Properties of W-Bonded 3P 43 Elements Tested 6000 hours at 525°C VI1 Initial Properties of W-Bonded 3P Elements To Be 46 . Tested Isothermally VI11 Couple Life Test I - Initial Operating Parameters 48 IX Initial Room Temperature Resistance of PbTe Couples - 53 Life Test I1 X Seebeck Voltage of Gradient Life -Tested Elements 63 vii I. INTRODUCTION The widespread use of thermoelectric power generation has been anticipated for some years as the solution to a number of specialized power -supply problems. However, the application of thermoelectrics has been hindered by a number of major materials problems. These problems can be divided into those associated with (1) the physical characteristics, (2) the chemical behavior, and (3) the low conversion efficiency of thermoelectric materials. The predominantly covalent nature of most thermoelectric alloys results in materials which are generally weak and brittle, In addition, PbTe alloys have a high thermal expansion coefficient which leads to susceptibility to thermal shock and thermal stress cracking. The mismatch in expansion coefficient between PbTe and metals also causes a fundamental physical incompatability which must be dealt with in contacting these materials at the hot side. The vapor pressure of PbTe alloys precludes operation in vacuum or requires encapsulation. Porosity in sintered PbTe not only contributes to their mechanical instability, but may also pose a substantial long -term hazard to the integrity of metallurigical bonds by migration of the pores in the temperature gradient. The high temperature mechanical properties of PbTe alloys are very poorly defined, and there is almost a complete lack of under- standing as to what role they may have in generator design.