THE UNIVERSITY OF MICHIGAN 4404-1-M FINAL MEMORANDUM REPORT A STUDY OF THE PHYSICAL AND ELECTRICAL PROPERTIES OF CHONDRITE METEORITES by J. B. Howe and E. F. Knott January 1962 Contract No. 1002 THIS DOCUMENT CONTAINS INFORMATION PROPRIETARY TO THE CONDUCTRON CORPORATION Conductron Corporation 343 S. Main Street Ann Arbor, Michigan

THE UNIVERSITY OF MICHIGAN 4404-1-M Patent Rights All Patent Rights, and all inventions, improvements, or discoveries (whether or not patentable) resulting from the research work carried out in the performance of Contract No. 1002 between the Regents of The University of Michigan and the Conductron Corporation are and shall be the sole property of the Conductron Corporation. Proprietary Data The study, analysis, test data, and material contained in this report is Proprietary Data of the Conductron Corporation and shall not be duplicated, used, or disclosed in whole or in part for any purpose to other than those personnel of The University of Michigan directly engaged in the work being carried out under Contract No. 1002 without the prior written approval of the Conductron Corporation. ii, i,,,,

THE UNIVERSITY OF MICHIGAN 4404-1 -M ACKNOWLEDGEMENTS Most of the electrical measurements were done in the Radiation Laboratory by J. Kimball, O. Sandus and one of the authors (E. F. K. ). Metallographic and petrographic studies were performed in the University of Michigan's metallography laboratory by D. Goodell and one of the authors (J. B. H.). D. Goodell, in addition, devised special instrumentation and techniques for use in the X-ray diffraction analysis and was responsible for most of the photographic studies. Important guidance, direction and assistance was given by Professor A. Olte of the Electrical Engineering Department and Radiation Laboratory, and by Dr. M. Miller and Dr. D. Ragone of the Conductron Corporation, Ann Arbor, Michigan. Valuable advice was also given by Professor W. Kelly of the Geology Department and by Professor H. Urey of the University of California at LaJolla....iii

THE UNIVERSITY OF MICHIGAN 4404-1-M TABLE OF CONTENTS EI 3INTRODUCTION 1 II ELECTRICAL INVESTIGATIONS 6 III MICROSCOPIC AND ANALYTICAL INVESTIGATIONS 37 IV CONCLUSIONS 62 APPENDIX I List of Meteorites Studied 65 REFERENCES 67 iv

THE UNIVERSITY OF MICHIGAN 4404-1-M "h — Such as it is, it promises the prize Of the one world complete in any size That I am like to compass, fool or wise." Robert Frost, "A Star in a Stoneboat" INTRODUCTION The non-linear electrical properties of chondrite meteorites were discovered during previous work by the University of Michigan Radiation Laboratory under NASA Research Grant NsG-4-59. The purpose of the work reported here has been to investigate these properties more thoroughly in bulk samples and to identify those materials, in the complex structure of the meteorites, which are responsible for their electrical properties. 1. 1 Description of Chondrite Meteorites The chondrites are stone meteorties which contain spheroidal grains (chondrules) of crystalline or glassy sillicates ranging in size from a small fraction of a millimeter to several millimeters. These grains may comprise practically the entire volume of the meteorite, but more commonly they are dispersed in a groundmass consisting of fragments of silicate crystals and broken chondrules. The chondrules often exhibit a "rayed" or "fan" crystal structure with the rays originating from a point on the surface. ---------— 1 -- ------— ii —------

THE UNIVERSITY OF MICHIGAN 4404-1-M Some chondrites are so friable that the chondrules can be easily removed from the groundmass. Others are quite hard with a compact structure, probably as a result of a metamorphic process which has caused partial recrystallization; under the microscope the chondrules are then seen only as ghost structures with indistinct boundaries. The silicate minerals in the chondrules and groundmass are most commonly olivine, (Mg, Fe, Mn)2 SiO4, and the orthohombic pyroxenes enstatite, bronzite, and hypersthene. Enstatite has the composition MgSiO3; in bronzite and hypersthene the magnesium is replaced in part by iron with small amounts of aluminum and calcium. Grains of nickel-iron and troilite (FeS) are common constituents of most chondrites and may make up nearly a quarter of the volume. The grains occur both in the chondrules and in the matrix and have irregular outlines, their shapes being determined mainly by the shapes of the surrounding silicates. The largest grains may be several millimeters across and all sizes can be seen down to the limit of microscope resolution. Chromite, (Fe, Mg) (Cr, Al, Fe)2 04, is another conspicuous mineral, although it is typically only about a tenth as abundant as nickel-iron and troilite. It most commonly occurs as irregular grains similar in size and shape to the mediumsized nickel-iron and troilite grains, and may in some cases be magnetite, FeO34, rather than chomite. Some chromite grains may have a rather high nickel content. 2

THE UNIVERSITY OF MICHIGAN 4404-1 -M Some chondritic meteorites are interlaced with dark veins, which may be several millimeters wide. These consist of glassy or microcrystalline silicates, usually containing globules and veinlets of metal and sulphide. The fusion crusts which form on the surface of these meteorites during their passage through the atmosphere are typically about one millimeter thick. The crusts have a thin outer zone of glass with many minute inclusions of an opaque mineral which is probably magnetite, and an inner zone wherein the silicates have been partially melted to form glass. Metal and sulphide grains in this zone have also melted and supplied material for the fine network of veins which pervade the inner zone, filling cracks in silicate and chromite grains. Most chondrites contain hydrated oxides of iron and nickel as replacements of metal and sulphide grains and as veins. These are more numerous in "finds" than in observed falls, and it seems likely that they are entirely of terrestrial origin. The oxide veins have formed either as crack fillings or as replacements of metal or sulphide veins which had formed previously as crack fillings, and they range in width from sub-microscopic up to several tenths of a millimeter. The larger ones may be several centimeters in length, with much complicated branching. 1. 2 Classification of Chondrites Meteorites have been classified according to several systems, most of these divide the stone meteorites into two large sub-groups: the chondrites and the -3 - -

THE UNIVERSITY OF MICHIGAN 4404-1-M achondrites. Further classification is made difficult by the smallness of the sampling, both in numbers and in size of individuals, and by the lack of knowledge concerning the genesis of meteorites, since any workable classification system contains implicitly certain assumptions as to the origin of and relationships between the objects being classified. The most commonly-used system is Prior's modification L1 of the Tschermak-Brezina system, which divides the chondrites into three groups according to the iron content of the magnesian silicate minerals, a as enstatite-chondrites, bronzite-chondrites, and hypersthene-chondrites. Each of these groups may be further subdivided according to color, as white, intermediate, gray, or black; and structure, as crystalline (compact mass, chondrules break with matrix), spherical (friable mass, chondrules do not break with matrix), brecciated (containing conspicuous large angular fragments) and veined (containing conspicuous dark veins). A special, rare class is that of the carbonaceous chondrites. 1. 3 Sources of Samples Most of the meteorites studied during the work reported here were purchased from the American Meteorite Laboratory, Denver, Colorado, or from the American Meteorite Museum, Sedona, Arizona. Several samples were obtained on loan from the Perry Collection through the generous permission of the University of Michigan Geology Department, and several others were kindly loaned by Professor H. C. Urey -----------------------— 4 ----------------

THE UNIVERSITY OF MICHIGAN -. 4404-1 -M of the University of California at LaJolla. A complete list of the meteorites studied is given in Appendix I. 1.4 Summary of Investigation The electrical investigation began as a continuation of the studies of the properties of large slabs with dimensions on the order of centimeters. Variations of potential were mapped with current passing through the slabs, and conductivity of the slabs was determined as a function field intensity and temperature. A few measurements were made on a microscopic scale using small probes. For one sample, the Hall Effect was measured. Samples of many of the meteorites were studied under the microscope and by other means in order to identify their constituents and to obtain an understanding of their structure, both as a means of elucidating their electrical behavior and as a first step in answering some of the many questions concerning their histories. The samples were examined microscopically in the form of polished sections and also as thin sections. A few areas of some samples were studied with an electron beam microanalyzer. X-ray diffraction analyses were obtained for some of the constituent minerals. ------------------- 5 i

--- THE UNIVERSITY OF MICHIGAN 4404-1-M II ELECTRICAL INVESTIGATIONS 2.1 Conductivity Versus Field Intensity and Temperature in Bulk Samples General Discussion For relatively large meteorite samples the conductivity has been found to be a power function of field intensity and absolute temperature. This has been observe in every sample whose conductivity was high enough to enable measurements to be made with the available instrumentation. The scale of the inhomogeneities in these samples is of the order of tens or hundreds of microns or less, and the "large" samples studied have dimensions on the order of tens of millimeters. Since both field intensity and temperature are involved it is obviously important that both these variables be accurately known, if not controlled. This presented the greatest experimental problem due to the fact that the temperature of a sample depended not only on the ambient conditions but also on the amount of power dissipated in the sample during the experiment. Two different approaches were followed: one technique limited the amount of power dissipation by using a capacitor discharge circuit with a small time constant and fairly small amounts of charge; the other technique used the opposite approach and expanded the time scale of the measurement while attempting to control the sample temperature by controlling the ambient temperature. Each of these approaches had certain advantages and disadvantages. ---------------- 6

THE UNIVERSITY OF MICHIGAN 4404-1-M 2.2 Sample Preparation Samples for all the "bulk" electrical measurements were prepared in the same way, and many of the samples were used in more than one type of experiment. Rectangular slabs were sawed from the meteorites, with dimensions which depended on the size of the specimen. The lengths ranged approximately from half a centimeter to five centimeters, widths from half a centimeter to four centimeters, and thicknesses were approximately half a centimeter. The faces were ground flat, and opposite faces were made parallel within 0. 001 inch. The slabs were dried by heating them at 2000C. for twenty hours, and were stored between experiments in dessicant jars. For some of the experiments silver paint was applied to two opposite faces on each slab to act as an electrical contact, but for the measurements in controlled temperature environments the slabs were simply clamped between parallel plates in a sample holder fabricated from Invar. The surfaces in contact with the sample were gold-plated. 2. 3 Capacitor-Discharge Measurements The first series of measurements was made under a previous contract [2]. The results will be reviewed here because they have in general been substantiated by later work and because they provide a good view of the relative behavior of a number of samples under similar experimental conditions. Values of capacitance ------— 7 —-----------

THE UNIVERSITY OF MICHIGAN 4404-1-M were chosen to give a time constant of 10 milli-seconds or less. Results were recorded by means of an oscilloscope and Polaroid camera. Eleven samples were used, all from the group purchased from the American Meteorite Laboratory. (See Appendix I. ) Two of the samples, Hugoton 280. 335 and Norcatur 525.18, had such low conductivities that their characteristics could not be measured. The results are given in Figures 1 and 2. (Note the difference in scale between the two figures. ) The curves are identified by the American Meteorite Laboratory numbers of the specimens, which are listed in Table I. TABLE I AML No. Meteorite Name 92. 1277 Plainview 92.1280 0 92. 1281, 92.1290 176. 34 Harrisonville 234K Roy 405.60 Ladder Creek 464.181 La Lande 670.14 Calliham ----- --------- 8 _ —-------

60- 92.1280(b) 92. 1277 92. 127 Note: Measurements were made for current passing in the direction of the c-axis, 92. 1281 except where noted. II92.1281 f 50 / Curves for samples 92.1290(b) and 234 K are practically coincident. I 40 Z V~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 176. 34 FIGoRE 1. TYPICAL CHONDRITE CHARAC5.TEITC 30 CD / P j co~ 20 10 -64181 CT) -4 92. 129 0(b) 0~~~~~~~~~~~~~~~~~~ 0 2 4 6 8 10 12 14 16 18 r Field Intensity, thousands of volts per meter FIGURE 1. TYPICAL CHONDRITE CHARACTERISTICS

600 Note: Measurements were made for current passing in the direction of the c-axis, except where noted. 4 400 z heI C c - U2 / 6 70.14 1 300 0 1 2 3 4 5 6 7 Field Intensity, thousands of volts per meter FIGURE 2. CHARACTERISTICS OF THE LALANDE AND CALLIHAM CHONDRITE FIGURE 2. CHARACTERISTICS OF THE LALANDE AND CALLIHAM CHONDRITE

THE UNIVERSITY OF MICHIGAN 4404-1-M Most of this data was obtained with the current flow parallel to the longest dimension of the slabs, except that in three cases the current flow was made parallel to the intermediate dimension, as indicated by a (b) after the identification number. One sample was measured in both directions with quite different results. The slab of Plainview 92. 1281 contained a transverse crack, and after measurements were made on the whole slab it was sawed in half perpendicular to the long dimension and the measurements were repeated. The conductivities of the two halves and of the uncut sample were all of the same order of magnitude, but the non-linearity of the half containing the crack was greater than that of the whole slab. Since the cracked half actually showed the highest current density of the three pieces, for the higher values of field intensity (and higher values than the whole slab for all field intensities) it would seem that the influence of the crack is at least obscure. The conclusions from this series of experiments were as follows: All samples which could be measured exhibited the non-linear behavior, some more than others. The degree of non-linearity was felt to be related to the number of cracks and veins in the sample. There was some concern that some or all of the apparent non-linearity might be due to temperature rise in the samples, even though the power input was very small. Under the present contract, therefore, a series of measurements was made on two samples (Plainview 92.1281 and Potter 476. 75) using progressively 11

THE UNIVERSITY OF MICHIGAN 4404-1-M smaller values of capacitance. The smallest time constants for which data could be obtained were, for Plainview, about 100 microseconds and for Potter, about 36 microseconds; the upper and lower extreme values of time constant differed by a factor of at least 100. 2. 4 Measurements with Controlled Temperature The sample temperature was controlled by two means. For one series of experiments an oil bath was used; for another series an oven was used, first with an air atmosphere and later with a nitrogen atmosphere. 2. 4o 1 Oven with Air Atmosphere Experimental Set-up The first series of measurements was made in an oven with no attempt to control the atmosphere. The Temco oven was capable of providing temperatures up to 1100 C., and the temperature was sensed by a chromel-alumel thermocouple junction placed in contact with one of the gold-plated electrodes of the sample holder. The thermocouple leads were brought out to the "XI input of a MoselyX-Y recorder. The majority of the measurements were made using a constant-current source consisting of a regulated power supply with a high resistance in series with the sample. The voltage across the sample was impressed on the "Y" input of the recorder. ------------------------ 12 -— _................a

THE UNIVERSITY OF MICHIGAN 4404-1-M Procedure and Results It was found that thermal gradients in the oven during the warming-up period made the data unreliable. Therefore, each run was made by first bringing the oven rapidly up to the maximum desired temperature and then allowing it to cool slowly with the heater turned off. Data was recorded during the cooling period, which lasted about eight hours. Measurements were made on several different slabs of Leedy and Plainview. One of the preliminary runs on a Plainview sample showed that after heating to 500~C. the conductivity suffered a permanent change. Two fresh slabs of Plainview were then each taken to successively higher temperatures in two series of runs, the last run in each case going to about 3600C. After the temperature went above approximately 130Co., there was a progressive and irreversible increase in conductivities; the curve for each run was smooth except for an inflection at about 1300C. which persisted until the final run to 3600. The reason for this was not apparent. The possibility of errors in temperature measurement or of changes in the characteristics of the contacts between the sample and the electrodes and of oxidation effects in the sample and the electrodes could not be discounted. (The gold-plated electrodes became tarnished during these runs. ) It was therefore decided to proceed with experiments which would eliminate these possible sources of error. 13

THE UNIVERSITY OF MICHIGAN 4404-1-M 2.4.2 Oil Bath Experimental Set-Up A Pyrex vessel 30 inches in diameter and 30 inches high was filled with five gallons of transformer oil; heating elements and a cooling coil were mounted near the bottom and an electrically-driven stirrer provided continuous oil circulation. Temperature was measured by a mercury-in-glass thermometer, the bulb of which was about 10 cm from the sample. The sample holder was of the same type used in the oven experiments. Voltage from a regulated power supply was impressed on the sample and also on the "XI input of the X-Y recorder; a voltage proportional to the current through the sample was impressed on the "Y" input. With this arrangement it was felt that the sample and the oil were isothermal within 1~C., provided that power dissipation in the sample was less than four or five watts (depending on the size of the sample). Temperatures were limited by the characteristics of the oil to a maximum of about 150~C. Procedure and Results Runs were made at ten-degree intervals of temperature, the voltage being increased smoothly and rapidly from zero to the maximum permissible value by manually rotating the power supply control knob. (The upper limit of voltage was determined by the allowable power dissipation in the sample, see above.) For every 14

--.THE UNIVERSITY OF MICHIGAN 4404-1-M sample, this gave a family of smooth curves of voltage versus current with temperature as a parameter. The record of the runs on Plainview 92. 1281 is reproduced as an example in Figure 3. Figure 4 is a replot of this data in terms of current versus temperature with voltage as a parameter. By plotting this data on log-log paper it becomes apparent that the current is a simple power function of absolute temperature for constant voltage, and a single equation in T and V was found to provide a good fit to the entire family of curves: -1-7 3 70 9 2.41 n I = (0. 453 x 10 17)T V + 90) 00 where I = current, amperes V = volts T = temperature, 0K Taking into account the overall dimensions of the sample, it is possible to express the behavior in terms of average current density and voltage gradient. ||= (2. 89 x 10-17) T3. 70 [(IEI + 46.9)2o41 1050o (2) where j = amperes/cm2 E = volts/cm Similar results were obtained for a slab of Calliham, yielding data for which the following provided a good fit: 15 -

30 il15~ /140 rna 20 /, 110 I ~~~~~~~~~~~~~~~~~~~~~~~~~i~~~~~~~////~ 00 /8 90, 38'~/ y180 r 3s3~ 70 ~ ~ SA i'v~~~~l p]N, VIE W 92. 1 2 8 volt's ~ ~ ~ ~ ~ ~ ~ ~ 60t 34 ~" _q 4o 7 /40 F'G. R" VS,AG t SAMPLE INy OIL 1281D CURRENT BS VOAAT 200 A) V$VLAGE WVITHTEPRTlEASAAAME:

THE UNIVERSITY OF MICHIGAN 4404-1-M 22 200 200 FIGURE 4. PLAINVIEW 92.1281D EXPERIMENTAL DATA 20 " OF FIG. 3 REPLOTTED WITH VOLTAGE AS A PARAMETER. 18 -- 16 175 14 12 150 0 -- / s/ 9/ v, volts 10 ~~~~~2 L ==C-2~-~~~~~~~~~~12 100 0 20 40 60 80 100 120 140 160 t, ~C 273 293 313 333 353 373 393 413 433 T, ~K 17....._

THE UNIVERSITY OF MICHIGAN 4404-1-M I = 10-13 T4.0 2.88 (V + 127)1 67 9020J Not all the samples tested gave similar results, however, and for some it was not possible to separate the voltage and temperature dependencies. For all these meteorites there was no apparent change in characteristics within this temperature range, and curves obtained for ascending and decending temperatures were very nearly identical. However, a sample of Potter 476. 75 underwent a permanent change at about 370 Ko The effect of this is shown in Figure 5. After the change the current apparently became a linear function of temperature. 2.4.3 Oven with Nitrogen Atmosphere Experimental Set-up The oil bath experiments provided a good environment for the samples, but were subject to an upper temperature limit of 150~C. Therefore, an enclosure was constructed for the oven so that a non-corrosive atmosphere could be provided for experiments at higher temperatures. This enclosure was made from a pneumatic water tank 16 inches in diameter and 24 inches long which was cut apart and provided with flanges, inlet and exhaust ports, and pressurized insulated connectors for the oven power and test leads. The thermocouple junction was placed in contact with the sample and temperature was monitored with a potentiometer, the X-Y 18

THE UNIVERSITY OF MICHIGAN 4404-1-M 150 FIGURE 5. POTTER 476. 75B CURRENT VS TEMPERATUR (IN OIL BATH) SHOWING PERMANENT CHANGE IN CHARACTERISTICS AT APPROXIMATELY 370K 30 100 ma A 20 10 0- I - I I 280 300 320 340 360 380 400 420 ~K --- 19 - ----

THE UNIVERSITY OF MICHIGAN 4404-1-M recorder being used for voltage and current measurement as in the oil bath experiments. Procedure and Results The enclosure was evacuated after installation of a sample, and then filled with dry nitrogen at a pressure slightly above atmospheric. Data was taken during the times of increasing temperature as well as the times of decreasing temperature. However, there was still some uncertainty as to the accuracy of the temperature measurement, or rather, as to how representative of the actual sample temperature the measured value was. No apparent oxidation effects were observed on either the samples or the electrodes even for the highest temperatures obtained, approximately 6500K. For the lower-temperature portions of the runs on samples which had been used for the oil bath experiments, the results agreed quite closely with those obtained in the oil bath. As the temperature was increased the samples generally underwent a permanent change, each in a different manner, although one, Holly, did not. A sample of Potter 476. 75 (not the same sample used in the oil bath) underwent a permanent change at about 340~K, (see Figure 6a. ) Succeeding runs to higher temperatures caused no further changes; (See Figure 6b and c). The sample of which was used for the oil bath experiment underwent a change at approximately 20 -

C- vUNIV E R SITY O I O-l- F M'ICHIGAN 4404-i -M 10 3 ^GURE POT 30T40 450 ol 300 volts 30 icreasig temperature a decreasing temperature..5~ ~~c~T-" 200 (a) r), l 100 0 300 I00 O450 oK 10 300 volts 5 (b) Run 2 200!577 -n n n n 100 300 0 350 400 55o oK 10 - 300 volts 200 0 R L~C~- (C) Run 3.,100 0 o 0

THE UNIVERSITY OF MICHIGAN 4404-1-M 3700K although the form of the curves during the initial heating was different. (See Figure 5.) 2. 5 Measurement of Hall Effect Experimental Set-up The Hall Effect was measured on a slab of Plainview, 7.193 x 3. 401 x 0. 252 cm in size. Current and Hall voltage leads of No. 28 copper wire were attached with silver print joints approximately 1 cm long, and the magnetic field was provided by a 2000-gauss permanent magnet with pole pieces about 4 cm in diameter. A 90 volt battery was used to provide a potential of 39. 5 volts through the use of a dropping resistor. The Hall voltage was recorded on the Moseley X-Y recorder driven by a General Radio DC Amplifier/Electrometer. Shielding was provided by a metal box enclosing both the sample and the magnet. Procedure and Results Three sets of measurements were made using three different combinations of current and field directions, each set consisting of at least three repeated measurements. In each case the value was determined by taking the average value graphically from a 100-second recording of the Hall voltage. Due to the inhomogeneity of the sample a voltage was measured in the absence of a magnetic field, and ------------------------- 22_ _ _-_-_-_ —_-_-_-_-_-_-*-_ —

THE UNIVERSITY OF MICHIGAN 4404-1-M therefore, the true Hall voltage was taken as the difference between the average values with and without the field. Repeatability of the results was not good, but the polarity at least was consistent and indicated that if the material was homogeneous the dominant carriers are positive. The inhomogeneity of the sample, however, makes it impossible to make any definite statement as to the sign of the carriers. The average value of the Hall voltage was found to be approximately 0. 19 volt, giving a Hall coefficient of 0. 5 cm /coulomb. This compares to values for germanium ranging from 1000 to 150, 000 cm3/ coulomb. The difference may be explained by the fact that the conducting portion of the meteorite occupies a very small portion of the sample volume, probably less than one percent. 2.6 Potential Maps These measurements provided a look at the electrical properties of meteorites on a scale midway between that of the bulk measurements described in the preceding section and that of the microscopic probe experiments to be described in a later section. Experimental Set-up and Procedure Slabs of several meteorites were used for these studies. The slabs were cut in the same manner as were those for the bulk conductivity measurements, but --------------------- 3 --------------------

THE UNIVERSITY OF MICHIGAN 4404-1-M their size was somewhat greater. Silver paint was applied to opposite ends and the sample was held flat on a board between spring-loaded contact strips. The surface was explored with a steel probe which formed one of the points of a pantograph; thus the positions at which readings were taken were automatically plotted. The applied voltage was adjusted to give a current flow of a few milliamperes through the sample, except that for samples with low bulk conductivity a maximum of 100 volts was applied. Both current and voltage were read in the measuring circuit between the probe and one end of the sample, using a Hewlett-Packard 410B voltmeter (100 megohms/volt) and a Leeds and Northrup microammeter. The circuit is shown in Figure 7, where the dashed line represents the boundary of the sample. In this greatly simplified schematic the total end-to-end resistance of the sample is represented by two resistances in series, R1 and R2. Their junction, assumed to be physically close to the surface, is coupled to the point at which the probe makes contact by a third resistance, Rc. The potential, E, at the junction is desired. Rg is the resistance of the meter, and Rs is a shunting resistance approximately equal to Rg. Each measurement was made with and without the shunting resistance, thus enabling the value of both E and Rc to be calculated. c Results The potential map for a sample of Ladder Creek is presented in Figure 8 as a typical example; some samples gave simpler, others, more complicated, patterns. 24

THE UNIVERSITY OF MICHIGAN 4404-1-M,, _____ Metallic grain R1 -I BI Probe I R E R 2 25 |L1 — ^s Rs FIGURE 7. MEASUREMENT CIRCUIT FOR OBTAINING POTENTIAL MAPS........... 25...._

. THE UNIVERSITY OF MICHIGAN - 4404-1 -M 99 85 97 99 91 87 99 86 99 92 66 99 92 100 volts applied -— 84 83 66 I =.94 ma. 80 77 \ 90 / 50 42 27 67 53 6, / 52 57 _ 58'158 _58 26 2 24 28 25 22 12 20 22 26 17 13 16 8 6 Potential Map (volts) Under scored values were obtained indirectly. FIGURE 8. POTENTIAL MAP OF SAMPLE 405. 60, LADDER CREEK, KANSAS.......,,,, 26

THE UNIVERSITY OF MICHIGAN 4404-1-M Applied voltage was 100 volts, with a current of 0. 94 ma. The figures represent potentials measured with the probe in contact with metal or sulphide grains at the respective locations. No potential could be measured on the silicate matrix, and some metal or sulphide grains showed zero potential. For samples with visible cracks the equipotential lines tended to be oriented parallel to the cracks. 2. 7 Electric Probe Studies The work on bulk conductivity and potential mapping had led to the hypothesis that the "grap" oxide veins in the meteorites were responsible for their behavior as non-linear conductors. It was therefore important to determine if possible the electrical characteristics of individual veins, and a series of attempts was made to design experiments using small probes and micromanipulators under the microscope. These experiments did not get beyond the exploratory stage, but a few interesting facts were discovered. At first two probes were used, but it was determined that erroneous results were being obtained and the last experiments were carried out with four probes. 2. 8 Two Probe Studies Experimental Set-up Probes were made from steel divider points sharpened on fine abrasive paper to a radius of approximately 0. 05 mm. They were held by a pair of 27

THE UNIVERSITY OF MICHIGAN 4404-1 -M micromanipulators and could thus be easily positioned within the field of view of a binocular microscope having a maximum magnification of 60x. Three types of electrical instrumentation were employed. Preliminary explorations of the sample surface were made using a VTVM ohmeter. A few measurements were made with the capacitor discharge and oscilloscope circuit used for the bulk conductivity experiments, and other measurements employed a regulated power supply and the Moseley X-Y recorder. Most of the measurements were made on thin slices, with a thickness of about 0, 015 inch, in order to insure that the segments of oxide vein material to be studied were as nearly two-dimensional as possible and not connected electrically with the other veins in the sample. Procedure and Results The binocular microscope with oblique illumination was ideal for observing the probes, but was not very good for identifying the constituents of the meteorites. A homemade vertical illuminator improved matters somewhat, but it wa)s still quite difficult to see the smaller veins and grains, and to distinguish between metal and sulphide grains. However, this was not a major handicap in the preliminary work. After visually locating an area of interest, the probes were lowered into contact with the surface. The pressure was not measured, but it could be estimated 28

THE UNIVERSITY- OF MICHIGAN 4404-1-M and made reasonably uniform by observing the amount of movement as the probes were lowered further after their first contact, thus causing the tip to slide a short distance along the surface. (The probes were mounted in their holders at an angle of about 45~ to the horizontal.) Probe spacings from a few millimeters to a few centimeters were used, with potentials of from zero to 150 volts. When the X-Y recorder was used the voltage was varied manually with the power supply control. With this instrumentation it was necessary that the resistance of the path between the probes be approximately one megohm or less. One of the most important discoveries made during these first probe experiments was that the oxide vein material varied widely in conductivity; many large veins had resistances too high to be measured even with the probes spaced less than a millimeter apart. It was therefore necessary to explore with the probes and ohmeter to find measurable regions. Overheating in the probe contact zones was sometimes a problem, and a few areas of metal and oxide were vaporized in the course of the experiments. Measurements were made on oxide veins, on individual oxide replacement grains, on individual metal and sulphide grains, and between pairs of metal and/or sulphide grains. Two chunks of terrestrial hydrated iron oxide (limonite, replacement of FeS2) and a piece of Fe203 were also probed, but their conductivities were so low that measurements could not be made. 29

THE UNIVERSITY OF MICHIGAN 4404-1 -M The meteorites studied were Calliham, Harrisonville, Hugoton, Plainview, and Potter. All oxide vein material which could be measured showed non-linear characteristics. There seemed to be some correlation between conductivity and the appearance of the vein in oblique or dark-field illumination; the conducting material was dark and the non-conducting material was lighter reddish. Sometimes conducting and non-conducting material occurred side-by-side in the same vein. The silicate matrix was non-conducting. Some pairs of grains were connected electrically and others were not, confirming the results of the potential mapping work. Those that were connected showed non-linear characteristics. The validity of the two-probe measurements was questioned on the grounds that spurious effects might be introduced at the contacts. Measurements on a carbon resistor confirmed these suspicions. Two-probe measurements on a flat ground surface of the resistor gave resistances which were too high and which showed an apparent negative temperature coefficient and a very slight spurious non-linearity in the potential-versus-current curve. When the current was carried by the resistor leads rather than by the probes, valid results were obtained. It was therefore decided to abandon the two-probe measurements and to develop a fourprobe technique. 30

THE UNIVERSITY OF MICHIGAN 4404-1 -M 2. 7 Four-Probe Studies Experimental Set-up The outer, current-carrying probes were the same as the ones used for the two-probe experiments; that is, they were steel divider points ground to a radius of approximately 0. 05 mm. Divider points were also used for the inner probes, but the tips of these were etched to a radius of approximately 0. 02 mm; this was the finest tip size for this material which could be easily manipulated without danger of frequent breakage. A metallographic microscope was used with a magnification of 40 x. A combination of vertical and oblique illumination gave excellent visibility for both the probes and the sample surface. All work was done on well-polished samples of approximately 0. 015-in. thickness. Both inner probes, and one of the outer probes, were held by micromanipulators. The second outer probe was attached to the frame of the microscope by means of an insulating support; the microscope's mechanical stage permitted the sample to be positioned under this probe. Power for the outer probes was supplied by a 22 / 45-volt heavy-duty "B" battery across a 20 K-ohm potentiometer. The maximum potential used was about 25 volts. Various combinations of meters were used to measure inner probe voltage and outer probe voltage and/or current, including a Hewlett-Packard Model 425 A 31

THE UNIVERSITY OF MICHIGAN - 4404-1-M micro-voltmeter/micro-microammeter, a Hewlett-Packard Model 410B VTVM, a Knight VTVM, and several panel-type microammeters and milliammeters. The measurements were never carried beyond the exploratory stage and no great degree of accuracy is claimed for the results. Most of the time spent on these experiments was employed in trying to discover the source of the very low-frequency noise which was picked up by the inner probes. This noise had peak values of the order of several millivolts and was present even when the outer probes were entirely removed from the sample. It was very erratic in its occurrence, but its immediate cause appeared to be mechanical vibration. Finally it was discovered that if the sample was repolished immediately before being placed on the microscope stage the noise was eliminated. Results After the noise probelm had been solved, three sets of measurements were made. The results of these will now be presented briefly. The measurements were made on two different slices of Calliham, designate as Slides No. 24 and 25. These slides each contained a section of the same large oxide vein; the slices had been sawed from opposite edges of a slab about 2 cm. wide and 0. 5 cm. thick which was traversed by the vein. (Figure 9 shows a portion of this vein in Slide 25. ) --—. —_-........ —. no32

THE UNIVERSITY OF MICHIGAN 4404-1-M FIGURE 9: OXIDE VEIN IN CALLIHAM SIDE NO. 25, SHOWING TWO OUTER PROBES. ONLY ONE INNER PROBE IS BEING USED, AND IT IS NEAR ONE OF THE OUTER PROBES. (X 40). -------..,0 5 1-1E | l 1|S' "';\ -;1D!'';................................... I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~........ - 2E1t

THE UNIVERSITY OF MICHIGAN 4404-1-M On Slide 24, the measurements were made on a small vein branching from the large vein, with one outer probe on the large vein near the junction. Three different runs were made, with the probe positions unchanged between the first two and changed slightly for the third. The outer probes were about 3 mm. apart and the inner probes about 0. 5 mm. apart and halfway between the outer probes. During the first run the outer probe and inner probe voltages were measured yielding a nonlinear relationship with a uniform drift of values during four scans of the range. For the second run the outer probe current and inner probe voltage were measured, yielding a linear relationship. For the last run, outer probe current, outer probe voltage, and inner probe voltage were measured, yielding non-linear relationships between each pair of parameters. On Slide 25, all four probes were first placed on the large vein, spaced uniformly about 1 mm. apart. One run was made, measuring outer probe current, outer probe voltage, and inner probe voltage. Results are shown in Figure 10. Here the relationship between outer probe and inner probe voltage was linear. Measurements were repeatable within less than one percent over a period of several minutes, and reversing the polarity at the battery caused no change in values. Finally the two outer probes were positioned about 3 mm apart on the large vein in Slide 25 and one of the inner probes was used to explore the intervening area of the vein, the voltage on the outer probes remaining at a constant 10. 0 volts. The ---------------------- 4 ---------------------

THE UNIVERSITY OF MICHIGAN 4404-1-M 70 m 60 3.0 0 I 50 2.5 5 o 40 2.00 < o 30 - 1.5 > 0 10 0.5 o i — o Outer Probe Voltage - volts 70 60 C / 0 40 0 p 20 20 o 10 0 0 0.50 1.00 1.50 2.00 2.50 2.75 Inner Probe Voltage - volts FIGURE 10. RESULTS OF FOUR-PROBE ELECTRICAL MEASUREMENTS ON CALLIHAM SLIDE NO. 24. 35

THE UNIVERSITY OF MICHIGAN 4404-1 -M outer probe current increased slowly over a period of two hours from 36 to 50 microamps, the power being on about 20 percent of the time. Reversing the battery polarity had no effect on the values. The voltage between the inner probe and one of the outer probes was measured, and the results were, to say the least, difficult to interpret; the vein apparently being as inhomogeneous electrically as it was visually. A large part of the area of the vein, especially the central zone, gave voltage readings near zero in relation to either probe. In general, the pattern of potentials in the vein was extremely complex, and since the size of the probe was large compared to the scale of visible inhomogeneity it was not possible to correlate the voltage readings with, for instance, the different shades of gray within the vein. Figure 9 is a typical example showing the inner probe near the positive probe in a spot which gave a reading of -9. 2 volts relative to the positive probe. 36

THE UNIVERSITY OF MICHIGAN 4404-1-M III MICROSCOPIC AND ANALYTICAL INVESTIGATIONS 3.1 Techniques of Sample Preparation and Study Since much of the time spent on this phase of the work was devoted to the development of techniques, a brief description of these techniques will be given here 3. 1.1 Slices and Thin Sections Some of the samples were polished and studied in the form of sawed slabs or in the form of chunks with one side or corner sawed off to give a flat surface. For most of the samples, however, slices 0. 020 in. thick were cut from the edges of slabs. This was done partly to conserve material, partly to provide more or less two-dimensional sections of veins, and partly to provide a good starting point for the preparation of true thin sections for study in transmitted light. The slices were prepared by the Electrical Engineering Shop, using a diamond-impregnated disc on a magnetic chuck grinder. Some of these slices were mounted in cylindrical mounts of polyester resin or bakelite and many of them were mounted on microscope slides using "Lake-side 70" cement. Special techniques had to be developed for polishing the more friable samples, although for the compact meteorites standard metallographic polishing 37

THE UNIVERSITY OF MICHIGAN 4404-1-M methods could be used. Even for compact specimens these standard methods, using cloth-covered metal laps with diamond paste abrasive, had the disadvantage of producing some relief on the surface due to the greater toughness of the metal and sulphide grains, but the method is quick and easy and the results are good enough for most purposes. When the standard polishing methods are used for friable samples the small silicate crystals are plucked from the surface. For some of the less friable samples good results were obtained simply by using glass plates instead of clothcovered laps and polishing by hand with two or three successively finer grades of abrasive powder in water suspension, but for the most friable specimens even this technique resulted in severe plucking. Urey and Mayeda (Ref. 3) used a technique of plastic vacuum impregnation for their studies of the metal particles in chondrites. We have studied some of the samples prepared by them and it appears that this method prevents wholesale large-scale plucking but does not succeed in holding the smallest silicate and metal grains in place. We developed a method of surface impregnation using either "Lake-side 70" cement dissolved in alcohol or polyester resin. This method holds almost all the surface grains in place but does not result in very deep penetration, so that care is required during the preliminary polishing with 600-grit abrasive to avoid wearing away the impregnated layer. With proper care and the use of glass 38

THE UNIVERSITY OF MICHIGAN 4404-1-M plates and hand polishing it is possible to achieve very good results, but the process is rather tedious and time-consuming and was used on only a few samples. The final polish was always obtained by the very brief use of a cloth-covered metal lap with alumina abrasive, the lap having been specially prepared with a hard-textured cloth of the type generally used for diamond abrasive and without the use of a soft backing layer. In the preparation of thin sections for observation in transmitted light, the slices were first polished on one side and then flipped over, being sure that the polished surface was parallel with the slide. (The cement is softened by heating to 140 F. ) Rough grinding to a thickness of about 150 microns and polishing was done on a glass plate. Extreme care is required to maintain a uniform thickness over the whole sample. Most of the microscopic examination of the samples was done on polished opaque sections under vertical illumination. In this type of light the silicates appear as various shades of dark gray with grain boundaries mostly invisible. Nickel-iron grains are a brilliant white, but since the samples were not etched, it is not possible to differentiate between the two metallic phases kamacite and taenite. The sulphide grains are a brilliant yellowish white. Under polarized light FeS2 appears black at all orientations while FeS grains vary from gray to black as the stage is rotated, making it possible to determine the size of individual crystals. The 39

THE UNIVERSITY OF MICHIGAN 4404-1-M hydrated iron oxides appear as various shades of light gray and the chromite and/or magnetite grains are a uniform light gray. When dark field illumination is used on polished sections the metal and silicate areas and the chromite/magnetite grains appear black, while the hydrated oxides vary from light orange-red to dark red. The silicate crystals are white to light orange, and areas of glass or recrystallized glass can be distinguished by their finer compact texture and sometimes by their color since they usually appear nearly black (probably due to their content of very small sulphide or metal particles). Thin sections examined with transmitted polarized light show the structure of the silicate minerals and glass areas, and the minerals may be identified by petrographic techniques. The metals, sulphides, and magnetite of course are black, and the chromite is also black except that very thin edges may show a reddish transluscence when the analyzing prism is not used. The hydrated oxides are transluscent to opaque with reddish-orange hues. Our thin sections were made with a highly-polished surface and without cover glasses; they appeared the same as ordinary thick polished sections in vertical illumination providing there was no reflecting surface below them. This was a valuable feature with the Leitz "Panphot" microscope, which could be used with either reflected or transmitted light by flipping a mirror in the illuminating system. Observation of thin sections in vertical polarized light gave a particularly revealing view of their structures. 40

- THE UNIVERSITY OF MICHIGAN 4404-1-M 3.1.2 Sulphur Prints This technique provides a simple method of identifying and locating sulphides on polished surfaces. Ordinary photographic paper is soaked in dilute sulphuric acid and the sample is pressed against the wet paper for a few seconds. Whereever a sulphide is present a dark spot appears on the paper. This technique was used occasionally during the early part of the program. It has the disadvantage of being destructive, and its use was discontinued in our work, since the sulphides could be identified visually. 3.1.3 Electron Beam Microanalyzer We had two sessions of two days each with an electron beam microanalyzer (EBM) belonging to the Advanced Metals Research Corporation of Somerville, Massachusetts. This instrument operates by bombarding the sample surface with a beam of electrons approximately one micron in diameter. The x-rays emitted by the bombarded area are diffracted by a mica crystal and the diffracted beam is sensed by a scintillation counter. Elements with atomic weights below that of sodium cannot be detected, but for the other elements concentration can be obtained to within a few weight percent using approximate corrections and to within a few tenths of a percent using refined correction calculations. The EBM can be used in two basic modes. Either the beam can be left on............... 41.......

THE UNIVERSITY OF MICHIGAN 4404-1-M one spot and the diffracted x-ray beam scanned for all the elements or the detection equipment can be peaked on one wavelength and the sample moved around under the electron beam to determine the concentration of a given element in any area relative to a standard for that element which is mounted beside the sample. Most of our time with the instrument was spent using the latter mode of operation. 3.1.4 X-Ray Diffraction Analysis X-ray diffraction analysis of powder camera records was used to identify some of the minerals. Samples were obtained at first by sawing small segments of material from the edges of 0. 015-inch slices mounted on glass slides. The saw was a special instrument fabricated primarily for use in isolating sections of oxide vein for electrical work. The sawing technique proved to be very laborious and was used only once. However it proved possible to obtain good samples from areas the size of the larger metallic grains by using a fine-tipped dental burr. 3.1. 5 Photography Many photographs of the samples were taken with a standard metallographic camera, at magnifications of from 70 x to 250x. Special camera equipment was set up to make macrophotographs of entire samples at approximately 5x from which enlarged prints were made with a scale of 10:1. The primary purpose of the photographs was to help in the work with the EBM, but they were also useful for all other phases of the study..X —-----..... 42

THE UNIVERSITY OF MICHIGAN 4404-1-M 3.2 Results of Investigations 3.2.1 Hydrated Oxides When examined under vertical illumination practically all the meteorites we have studied are seen to contain a network of gray veins. These are the surface expressions of thin interlacing sheets of oxide which fill cracks in the silicates and to a lesser extent in the metal and sulphide grains. (See Figure 11.) These veins range in size from several tens of microns down to the limit of vision. There are also gray grains which represent replacement of metal or sulphide grains by oxide; examples of all stages in the replacement process can often be found in one sample. (Figure 14) These oxides are present in all our samples of "finds"; they occur also in most of the "falls", but always less well developed both in extent and in size. The vein walls are typically sharp and fresh-looking with no apparent reaction between the vein material and the walls; opposite walls would fit together if the vein were removed, although there are some exceptions to this. In meteorites with large and numerous oxide veins there is usually some void space in both the veins and the grains, as if the process of filling had not been completed. The vein material appears inhomogeneous in all types of light. This is best observed in vertical illumination where one can see a complex banding of various shades of gray in all but the smallest veins. The banding is more or less parallel to the vein or grain boundaries and represents a cross section of colloform 43

THE UNIVERSITY OF MICHIGAN 4404-1-M FIGURE 11: TYPICAL HYDRATED OXIDE VEINS IN CALLIHAM. LIGHT GRAIN NEAR BOTTOM OF PHOTO IS TROILITE. A LARGE TROILITE GRAIN OCCUPIES MOST OF UPPER RIGHT PORTION OF PHOTO BUT APPEARS SOMEWHAT DARKER DUE TO LIGHTING. (I 250).......- 44.. —, -

THE UNIVERSITY OF MICHIGAN 4404-1-M structure which results from layer-by-layer deposition of material on the walls. Exceptions to this are found in Calliham and will be discussed below. The appearance of the oxide veins and grains is closely similar to what is observed on polished sections of samples taken from near-surface oxidized zones of terrestrial sulphide ore bodies, and the material in the meteorite veins has been shown by our x-ray diffraction analysis and by our EBM work to be essentially the same as the material in these terrestrial occurrences. The hydrated iron oxides are loosely termed limonite and were formerly thought to be amorphous. They actually occur as sub-microscopic crystals of several different minerals, the most common of which are goethite and lepidocrocite (gamma and alpha Fe203 H20 respectively); lines for both these minerals have been identified in the x-ray films. The EBM analysis of the oxides showed that in most cases they contain about five percent nickel, and rarely up to ten percent or more. Some spots contained up to about five percent of silicon and sulphur. The nickel content seemed to correlate with the color of the oxide, tne darker areas being richer in nickel. Where the electron beam remained for a considerable length of time on one spot, there developed on the oxide a network of the fine cracks suggestive of dehydration. It was mentioned above that our samples of Calliham contain unusual structures in the oxide veins. These are rhythmic transverse bands of lighter and darker oxide material which occur in some areas of the smaller veins. (Figure 12) 45

THE UNIVERSITY OF MICHIGAN 4404-1 -M FIGURE 12: HYDRATED OXIDE VEIN IN CALLIHAM WITH TRANSVERSE BANDING. SMALLER UNBANDED GRAINS CUT THE SILICATES AND THE LARGE TROILITE GRAIN. DARK MARKINGS ON THE TROILITE ARE PITS CAUSED BY POLISHING (X 250)........ —-- ~46

THE UNIVERSITY OF MICHIGAN - 4404-1-M The bands are always very nearly perpendicular to the vein walls. Spacing of the bands is roughly equal to the vein width. (Of course the apparent width of any vein on a polished surface depends on the angle at which the surface cuts the vein.) There is some tendency for the banding to occur in the vicinity of sharp bends in the vein. The individual bands are too small to be reliably distinguished by the EB beam, but a scan along one of them showed a rhythmic variation in nickel and iron content with nickel varying from about 2% to about 5% and iron varying inversely from about 58% to about 65%. There are several instances in various meteorites where an oxide vein changes abruptly into a sulphide vein and one large vein in Potter has large remnant patches of sulphide. It is possible that many if not all oxide veins are actually replacements of sulphide (or metal) veins by oxide. The oxide grains do in many cases show "ghost" structures of the material replaced; the best examples of this are in Kingfisher where metal crystal boundaries are sometimes preserved as ghosts in the oxide. This may be the explanation of the banded oxide veins in Calliham. The hydrated oxides are generally assumed to be of terrestrial origin, that is they are the result of corrosion of the meteorite after it landed on the earth. This is partly an assumption; however, it is supported by the fact that these oxides are clearly more abundant in "finds". It would be very surprising if hydrated oxide 47

THE UNIVERSITY OF MICHIGAN 4404-1-M had not formed in "finds". Proof that some of the oxide veins are of terrestrial origin has been observed in a very few instances where the veins are present in the fusion crust. The opposite case, of oxide veins apparently destroyed during the formation of the fusion crust, has not been observed. The puzzling fact remains that some meteorites which were picked up a few days or even hours after they were observed to fall have extensive (though small) oxide veins and incipient oxide replacements of grains. It is difficult to see how this could have happened on a museum shelf. 3.2.2 Metal and Sulphide The nickel-iron and sulphide grains were not studied in detail. No etching was done, so that the relationships between the two possible nickel-iron phases (taenite and kamacite) could not be seen. However, certain information can be obtained from the shape of the grains, from the crystal sizes in the sulphide grains, from inclusions in the grains, and from metal and sulphide veins. The shape of these grains suggests unequivocably that they are interstitial to the silicates; that is, they were the last material to solidify, although they may originally have been present as solid grains in a breccia or angular "sand" of silicates and metallic particles. They are highly irregular in outline and the finer details of their shape appear to be determined by the shape of the adjacent silicate grains. On well-polished surfaces one can often see very fine veins of metal or 48

_... THE UNIVERSITY OF MICHIGAN... 4404-1 -M sulphide extending from a grain into cracks in the silicates (Figure 13), sometimes giving an overall "shredded" appearance to the borders of the metallic grains (Figure 14). In a few chondrites the metallic grains contain fragments of silicate crystals, usually somewhat rounded. This can be unmistakably seen in Hessle (Figure 15); of course, most occurrences of "islands" of silicate in metallic grains are due to the accidental plane of cutting in the sample. (A different situation is seen in the achondrite Abee where the nickel-iron grains contain phenocrysts of silicate which grew in the molten metal.) Urey and Mayeda (3] found that the metal grains in chondrites contain taenite, kamacite, and plessite (a fine mixture of taenite and kamacite) in various combinations suggesting that the grains cannot have been molten after they arrived at their present locations within the meteorites. This contradicts the findings set forth above, and the two viewpoints are difficult to reconcile on the basis of the work we have done. Urey and Mayeda's hypothesis supposes that the chondrites are random mixtures of metallic and silicate materials, a view which is opposed by most authorities on the basis of evidence from a large number of analyses which suggests that most meteorites are in chemical equilibrium and that there is a rather smooth variation in oxidation state of the iron and in nickel content of the nickel-iron among all chondrites. (See, for instance, Ringwood [4].) The sulphide grains have in general the same mode of occurrence as the nickel-iron and in different meteorites may be more plentiful or less plentiful4 49

THE UNIVERSITY OF MICHIGAN 4404-1-F FIGURE 13: AREA IN HARRISONVILLE REMOTE FROM FUSION CRUST AND DARK VEINS SHOWING FINE VEINS EXTENDING FROM GRAINS. CLEAR WHITE AREA IS NICKEL-IRON. PITTED WHITE IS TROILITE. GRAY GRAIN AT LOWER RIGHT CORNER IS CHROMITE. MOST VEINS HAVE BEEN REPLACED BY HYDRATED OXIDES. (X 250). 5 50 -.......... xxx.........:: ~; FIGURE 13: AREA IN HARRISONVILLE REMOTE FROM FUSIONj::~:~: CRUST AND DARK VEINS SHOWING FINE VEINS EXTENDINGii FROM GRAINS. CLEAR WHITE AREA IS NICKEL-IRON.~~~~~~~~~~~~~~~~,~~jiijiiiliii PITTED WHITE IS TROILITE. GRAY GRAIN AT LOWER:~ ~Biiiliiiii~i:~ RIGHT CORNER ISCHROMITE. MOSTVEINS HAVE BEE REPLACDBYHYDATEDOIDES. X250)

THE UNIVERSITY OF MICHIGAN 4404-1-M FIGURE 14: AREA IN PLAINVIEW REMOTE FROM FUSION CRUST AND DARK VEINS SHOWING "SHREDDED" GRAIN BOUNDARIES: THE TWO LARGE WHITE AREAS ARE NICKEL-IRON. ALL SMALLER WHITE AREAS ARE TROILITE. THERE ARE NUMEROUS HYDRATED OXIDE VEINS, AND THE NICKEL-IRON IS PARTIALLY REPLACED BY HYDRATED OXIDE. (X 250). -----------------------— ll 51- - ------------

THE UNIVERSITY OF MICHIGAN 4404-1 -M.......4. FIGURE 15: LARGE NICKEL-IRON GRAIN IN HESSLE CONTAINING FRAGMENTS OF SILICATE CRYSTALS. (X 250). -------------— 52 —------------

THE UNIVERSITY OF MICHIGAN 4404-1-M than the metal. In some meteorites even the large grains are mostly monocrystalline (e.g., Calliham) while in others (e.g., Harrisonville) they are polycrystalline. It is rather common to see irregularly-shaped areas in a sample surface which are rich in small sulphide grains, as if a large grain had been partially dispersed, In some samples (e.g., Kingfisher) the large metal grains are often surrounded by a region of small sulphide grains. The sulphide is nominally troilite (FeS), but is actually (Fe, Ni)S. The nickel content as determined by the EBM is generally about 1 or less, although in a few cases it was somewhat higher and in one case was about 25%. In Calliham the EBM identified one grain of pyrite (FeS2); two nearby grains appear by their shape and optical properties to be the same. These are euhedral crystals and are in contact with troilite. The nickel content of the pyrite grain was about 0. 5 % and the adjacent troilite had about 5% nickel. 3.2.3 Chromite This mineral occurs in inrregular grains similar to the metals and sulphides but without shredded boundaries. It is more rarely seen as euhedral crystals within metal grains. It occurs rather commonly as minute octahedra in small areas of glass (or devitrified glass). The octahedra generally occur in linear groups and are often associated with a normal large chromite grain which appears to have been partially melted. (See Figure 16)...........- 53.....................53.

THE UNIVERSITY OF MICHIGAN 4404-1-M W X\''.R, m k ~~~!i l i I::::: 0: 0 i:: v:: FIGURE 16: AREA IN MODOC WITH CHROMITE OCTAHEDRA IN DIVITRIFIED GLASS. DARK GRAY IS GLASS, MEDIUM GRAY IS SILICATE CRYSTALS, AND LIGHT GRAY IS CHROMITE. WHITE AREAS ARE NICKEL-IRON. THE CHROMITE OCTAHEDRA MAY BE DERIVED FROM THE LARGE FRACTURED CHROMITE GRAIN NEAR THE CENTER OF THE PHOTO. (X 250). - - 54 -

THE UNIVERSITY OF MICHIGAN 4404-1-M Chromite grains or octahedra were examined in four spots on Calliham and Modoc with the EBM. Chromium content was from about 12% to about 30%. One spot showed about 20% nickel, but the others had less than 1 %. Aluminum was searched for in one large grain and in an adjacent crystal and was found in a concentration of about 5 %. 3.2.4 Dark Veins The dark veins which give "veined chondrites" their name appear dark to the naked eye and are opaque in thin sections. Under the microscope with vertical illumination they do not appear markedly different from the silicate "host rock" of the meteorite except that they usually contain many globular, spheroidal, or lacy inclusions of nickeliferous troilite and less commonly of nickel-iron. Figures 17 and 18 show typical examples. In dark field illumination these veins appear distinctly darker than the host rock, being composed of glass or submicroscopic crystals. Boundaries with the host rock are distinct but smoothly rounded, and where crystals from the host rock have been entrained in the vein material their corners are rounded. The sulphide and metal in these veins has probably been entrained from grains in the host rock. Where a sulphide or metal grain forms part of a vein boundary the grain appears shredded with lacy extensions entering the vein material, and commonly there is evidence for flow with filaments of the grain material parallel to the vein walls. 55

THE UNIVERSITY OF MICHIGAN 4404-1-M FIGURE 17: DARK VEIN IN HARRISONVILLE. THE SPHEROIDAL GLOBULES AND THE VEINS ARE TROILITE. BOUNDARY OF VEIN IS AT LOWER RIGHT. (X 250). -- ---..... —----- u 56

THE UNIVERSITY OF MICHIGAN 4404-1-M FIGURE 18: DARK VEIN IN FOREST CITY. LARGE NICKEL-IRON GRAIN FORMS LEFT BOUNDARY OF VEIN. THE SPHEROIDAL GLOBULES AND THE LARGE GRAIN AT BOTTOM CENTER APPEAR TO BE MIXTURES OF NICKEL-IRON AND TROILITE. (X 250). 57

THE UNIVERSITY OF MICHIGAN 4404-1-M The globules in the veins are mostly of sulphide with metal globules appearing only in close proximity to metal grains in the boundaries. The globules range in size from a few microns down to the limit of vision and the smaller ones are nearly spherical. There is a definite tendency for the globules to be larger in the central portion of a vein probably due to the fact that the central portion was the last to solidify and the globules had more time to coalesce. Small veinlets of sulphide often occur, generally as off shoots from globules or as connections between several globules. The material in the veinlets appears to be continuous with that in the globules and no doubt they are the result of cracking of the glassy vein before the sulphide solidified. Where a veinlet transverses a globule the two halves of the globule are sometimes displaced relative to each other. The larger sulphide globules are very finely polycrystalline; it is not possible to see whether or not the smaller ones are monocrystalline. In veins where there are both metal and sulphide globules, some globules appear to have a eutectic structure of metal and sulphide; this was not confirmed with the EBM but the structure is very fine. The bulk of the dark vein material in Harrisonville was shown by the EBM to be very rich in silicon (30-40%) with about 10% iron, 5% calcium, and traces of nickel and aluminum. In the two areas for which element scans were performed no magnesium or sodium were found. 58

THE UNIVERSITY OF MICHIGAN 4404-1-M The dark veins were probably formed in the meteorite parent bodies. They do not seem to be continuous with the fusion crusts and it is unlikely that such large and numerous cracks could have formed and been filled during a meteorite's flight through the earth's atmosphere without wholesale disintegration. 3.2.5 Glassy Areas The small glassy areas found in most chondrites have been mentioned above as being the home of small chromite octahedra. (See Figure 16.) This material is probably in most cases a cryptocrystalline devitrified glass. In vertical illumination it appears slightly darker than the crystalline silicates and occurs interstitially to them, or occasionally as lamellae between partings of large crystals. In dark field illumination it appears distinctly lighter than the crystalline silicates; in thin section its glassy or cryptocrystalline nature is apparent. This material may in part be the so-called maskelynite or vitreous plagloclase mentioned in the literature. The EBM results on a few spots suggested the presence of higher aluminum, silicon, and calcium content and lower iron content than in the matrix crystals. The visually-observed structural relationships, and particularly the occurrence with included chromite octahedra from adjacent remelted chromite grains, indicate that the "glassy" areas may be the result of intense local heating followed by rather rapid cooling. It might be useful to investigate the possibility that shock waves are the responsible agents. 59

THE UNIVERSITY OF MICHIGAN 4404-1 -M 3.2.6 Fusion Crusts Most of our samples of observed falls and some of the "finds" include portions of the fusion crust, which in all cases looks very similar. An example is shown in Figure 19. The crust is quite thin, approximately 0. 5 to 1 mm, and appears to the naked eye as a black selvedge on the polished section. When observe under the microscope in vertical illumination it is seen to consist of two zones, The outer zone, which may be missing in some areas, is less than 0. 5 mm thick and is a region of ablation and flow consisting of glass with many minute gray inclusions which may be magnetite; they are too small for their shapes to be resolvable, and too small also for identification with the EBM. The inner zone has been altered by partial melting of the constituents, with areas of glass and many veinlets of sulphide and metal filling a network of cracks in the crystals and glass. Chromite grains are often cut by these veinlets and evidently the chromite did not melt until it was reached by the ablation zone. There are very rare occurrences of metal or sulphide globules in the glass of the inner zone. 60

THE UNIVERSITY OF MICHIGAN 4404-1-M FIGURE 19: FUSION CRUST ON MODOC. THE BLACK RIM ALONG THE OUTER EDGE OF THE CRUST AT RIGHT IS CAUSED BY ROUNDING OF THE EDGE DURING POLISHING. OUTER ZONE OF CRUST IS HERE ABOUT THE SAME WIDTH AS THIS BLACK RIM. THE WHITE GRAIN NEAR BOTTOM EDGE OF PHOTO IS SULPHIDE; OTHERS ARE NICKEL-IRON. THE VEINS ARE SULPHIDE. THE OUTER EDGE OF THE LARGEST GRAIN HAS A RIM OF SULPHIDE (NOT EASILY SEEN IN THE PHOTO) WHICH IS PARTIALLY REPLACED BY LIGHT GRAY HYDRATED OXIDE. (X250). 61..

THE UNIVERSITY OF MICHIGAN 4404-1-M IV CONCLUSIONS 4.1 Electrical Investigations All measurements of conductivity versus field intensity and temperature with bulk samples showed non-linear relationships. Two meteorites had such low conductivities that they could not be measured with the available equipment. For temperatures below approximately 150 C the conductivity of most samples was found to be a power function of absolute temperature for constant voltage and of voltage for constant temperature. At temperatures above approximately 150~ C the behavior was complicated for most samples by the appearance of reversible and irreversible changes in the form of the relationships, the cause of which was not determined. Probe measurements with a resolution on the order of millimeters showed that the samples were electrically very inhomogeneous and that the conductivity was probably due to the network of fine veins. Two probe and four-probe measurements on polished 0. 015-inch thick samples with a resolution on the order of tens of microns showed that the veins themselves were electrically inhomogeneous; the non-linear relationship between ----------------- 62

THE UNIVERSITY OF MICHIGAN 4404-1-M conductivity and field intensity was usually observed for those regions which had measurable conductivities. 4. 2 Microscopic and Analytical Investigations Techniques of preparing polished sections and polished thin sections of friable meteorites were perfected. Powder samples from microscopic grains were obtained by a drilling technique. Analysis of the samples was carried out by x-ray diffraction analysis, by means of the optical microscope, and by the use of an electron beam microanalyzer. The veins which are responsible for the electrical conductivity of the meteorites were found to be nickeliferous hydrated iron oxides, probably of terrestrial origin. Evidence was found that the metal and sulphide grains in chondrites have in most cases been molten since the meteorites achieved their present configuration, and that perhaps many of the oxide veins are replacements of metal and sulphide veins. The large dark veins in veined chondrites, which contain numerous spheroidal globules of sulphide, and less commonly of metal, are thought to be the result of invasions of the meteorite by hot liquid silicate with resultant melting and entrainment of local metallic grains; a process which probably occurred in the meteorite parent body rather than during the flight through the earth's atmosphere. Most chondrites have more or less numerous small areas of glass or of 63

THE UNIVERSITY OF MICHIGAN 4404-1-M cryptocrystalline devitrified glass which sometimes contains minute crystals of chromite. The chromite appears to have been derived from adjacent large grains by remelting and subsequent rapid cooling, and the glass may have a similar origin. It is suggested that shock waves may have been the responsible mechanism. 64

THE UNIVERSITY OF MICHIGAN 4404-1-M APPENDIX I LIST OF METEORITES STUDIED A. Samples purchased from American Meteorite Museum, Sedona, Arizona, and American Meteorite Laboratory, Denver, Colorado. Name AML No. Class Date Found Beenham 414.71 Intermediate chondrite 1937 T. 82 t TT.82,i.172 i t it t 23 it.. 223 Brady 384.3 Chondrite 1937 Briscoe Co. 477.4 Gray chondrite 1940 Calliham 670.14 Chondrite Recog. 1958 Harrisonville 176.34 Veined gray chondrite 1933 Holly 386.7 Chondrite 1937 Hugoton 280.48 Brecciated crystalline chondrite 1935,I.335 T TT IT,t Kelly 398.25 White to gray brecciated chondrite Kingfisher 601.25 Black chondrite 1950 Ladder Creek 405. 60 Spherical crystalline chondrite 1937 La Lande 464.181 Spherical crystalline chondrite 1933 Leedy - Chondrite Fell 1943 Morland 282.1 Crystalline chondrite 1935 Norcatur 525.16 Veined intermediate chondrite Recog. 1948.18 T T ft iTT 94.24 ft tt Plainview 92.1277 Fragments of spherical 1917.1280 Chondrite embedded in it.. 1281 veined intermediate chondrite.ti,.1290 t i,t __-(?) i,,, Potter 476.30 Brecciated veined gray chondrite 1941 IT.75 t T T..75 Roy 234 K Spherical crystalline chondrite 1933 Taiban 541.7 Black chondrite 1934 I —------------------ - 65 --

THE UNIVERSITY OF MICHIGAN 4404-1-M B. Samples loaned by University of Michigan Geology Department from Perry Collection Name P.C. No. Class Date of Fall Alfianello 117 Intermediate hypersthene chondrite 1883 Beardsley 23A Chondrite 1929 Forest City 11 Brecciated spherical bronzite chondrite 1890 Hessle 110 Spherical bronzite chondrite 1869 Linwood 190E Iron, coarsest octahedrite Found 1940-41 Modoc 73 Veined white hypersthene chondrite 1905 Pultusk 2301 Veined gray bronzite chondrite 1868 Richardton 100.41 Veined spherical bronzite chondrite 1918 Rose City 16 F Brecciated black bronzite chondrite 1921 C. Samples loaned by Professor H. C. Urey, University of California, LaJolla. Name Class Date Abee Achondrite Fell 1952 Allegan Chondrite Fell 1899 Carcote Crystalline bronzite chondrite Found 1888 Densmore Chondrite Found 1879 Estacado Crystalline bronzite chondrite Found 1883 Gilgoin Station Crystalline bronzite chondrite Found 1889 Mocs Veined white hypersthene chondrite Fell 1882 Modoc Veined white hypersthene chondrite Fell 1905 Tulia Veined crystalline chondrite Found 1924 66

THE UNIVERSITY OF MICHIGAN 4404-1-M REFERENCES 1. Prior, G.T. 1923. "Catalogue of Meteorites" British Museum (Natural History), Ondon. 2. Knott, E. F. 2 Feb. 1961. "Conductivity Studies of Stone Chondrites", U. of M. Rad. Lab. Memo. No. 02914-523-M. 3. Urey, H. C. and Mayeda, T. 1959. "The Metallic Particles of Some Chondrites", Geochim. et. Cosmochim. Acta, Vol. 17, pp 113-124. 4. Ringwood, A.E. July 1961. "Chemical and Genetic Relationships Among Meteorites", Geochim. et Cosmochim, Acta, Vol. 24, pp 159-197. 67 -