New Mexico Tech

Sample Description

Sample T1,2 contained two tiny rod or lamellar-shaped pieces, primarily gray-black in color with some brown-white deposits at several locations on the surface. Sample T1,2-A is approximately 4.4 mm long and 0.6 mm in diameter. Sample T1,2-B is approximately 5.75 mm long and 0.6 mm in diameter (See micrograph in Fig. 1). These two samples were observed to be rather strongly magnetized along their long axes.
Sample T3 also contained two small pieces. T3-A is long (approx. 4.1 mm) and thin with an irregular geometry (see Fig. 2), while T3-B is essentially equiaxed with diameter of about 3 mm. T3-A has many interesting features visible under a 4X microscope. A yellow-white flake-like substance is attached to the surface. In the neck region of the sample there appears to be a high concentration of small, reflective aggregates with a copper-gold color. T3-B has several yellow-brown spots on the surface. T3-A was observed to be magnetized, while T3-B was not.
All samples were found to be non-conducting when the probes of a DMM were held to the surface.
Physical Properties

Density

Bulk densities of the samples were measured using an immersion technique (based on Archimedes’ principle). By accurately measuring the sample mass, both dry and when submerged in a liquid of known density, one can calculate the bulk density of an unknown. In this case, toluene (rho = 0.862 g/cm^3) was used rather than water to avoid any possible hydrolysis or oxidation reactions with the sample. A Sartorius precision balance was used for measurements to the nearest 0.00001 g. Mass and density values are tabulated below:
 
 
 

T1,2-A T1,2-B T3-A T3-B
m (g) 0.01070 0.01395 0.00657 0.00404
rho (g/cm^3) 5.62 5.81 4.95 2.70

Despite the small size of the samples, repeatable values were obtained and are believed to be accurate. The difference between T1,2-A and -B is probably within experimental error, and thus 5.7 g/cm^3 is taken as the bulk density of T1,2. The difference between these and T3-A is considered to be significant, suggesting some difference in material. Sample T3-B has a density roughly half that of the others.
 

Hardness

  Two samples were chosen for hardness tests, T1,2-B and T3-B. About 1/3 of the length of T1,2-B was broken off and warm mounted in a slug for grinding and polishing utilizing standard metallographic techniques. The polished surface ofthe core material was highly reflecting and white-gray in color, clearly metallic in nature. While grinding through sample T3-B, a dull, black color is all that was observed, indicating the lack of metallic core. This is in agreement with the very low density measured. Once the sample was polished to a very flat and smooth finish, a Vickers Hardness number was obtained from a Leco Tester using a diamond tip micro-indenter. Five indentations were made for each sample. The size of the resulting indentations were measured under a light microscope, and averaged. This average value, along with the known applied load were used to come up with the Vicker’s Hardness number. The hardness values are tabulated below.
 
 
Vickers Knoop Moh’s comparison
T1,2-B 821 821 7 SiO_2
T3-B 136 136 3 CaCO_3 

Vickers corresponds to the Knoop hardness scale. A corresponding Moh’s hardness (1-10 scale) is given for comparison. A huge difference in hardness is observed. T1,2-B is seen to be very hard, as Quartz or a very hard (carburized) tool steel. T3-B is seen to be relatively soft, as Calcite.

SEM and Chemical Analyses

  Due to small sample size, chemical analysis was limited to a qualitative analysis using X-ray Energy-Dispersive Spectroscopy (EDS)* as an auxiliary unit on a scanning electron microscope (SEM)**. The two samples T1,2-B and T3-A (shown in Figs. 1 & 2) were chosen for analysis. Since the samples were found to have an insulating outer shell, it was necessary to deposit a very thin layer of carbon (few angstroms thick) to avoid charging during imaging. Figure 3 gives the EDS analyses for various locations on sample T1,2-B. Major amounts of iron, phosphorous, and calcium, and minor amounts of chlorine were detected. The spectra shown were all obtained using a collection time of one minute. As there is virtually no difference in these spectra, it is concluded that the composition of the cladding material is uniform over the entire sample. Figure 1 (b) shows a higher mag shot of the cladding material near the central portion of the sample. The obvious cracking may be indicative of differential thermal expansion/contraction of cladding and core materials. A calcium dot map suggested an even distribution of calcium over the sampling area. A dot map taken from the protruding area on the lower left side of the sample also suggests no real preferential segregation of calcium (Figure 1(A)). Figure 1(C) shows some of the features of the surface in the divot area on the right side of the sample. This surface is a bit rougher and devoid of the clean craze-like cracks seen in Figure 1(B).
 

Both the microstructure and composition of sample T3-A varies tremendously across the sample. I identified primarily three different locations: the bulk region near the tip, the flaky deposit bottom right, and the neck region (Figure 2). The higher mag shot of the bulk tip region reveals a microstructure like that of sample T1,2-B; crazing-like cracks across a relatively smooth surface. The bulk tip region was found to be similar chemically as well, with major constituents Fe, Ca, and P, and minor amounts of Cl (Figure 4). However minor amounts of copper, and aluminum were also detected. The dark, bulk region near the center of the sample was found to be nearly identical.
As the difference in back-scattered intensity suggests, the "flaky" region has a very different composition and phase. It was found to contain a lesser amount of iron, with major quantities of silicon, phosphorous, molybdenum, chlorine, sodium, calcium and a trace of copper. The material may also contain a number of elements in between, since the peaks are broad and overlapping. Its microstructure seen in the higher mag shot is very interesting, the flake-like nature perhaps indicative of a layered silicate sheet structure.
 

The "neck" region appears to be highly metallic. EDS reveals a host of metals in addition to the major iron: copper, aluminum, tin, and nickel. The "balls" seen in the higher mag shot correspond to the gold-copper colored aggregates seen under the optical stereoscope at low mag. Individual EDS scans taken both directly on the "ball" and in the "rough" area at the center of the picture were essentially the same as Figure 4(A).

Phase Identification

X-ray Diffraction Experiments
 

Attempts at obtaining an x-ray diffraction (XRD) pattern from the samples using our Philips diffractometer were unsuccessful. Detection of any reflections at all from such small samples (without grinding to a powder) requires special instrumentation and conditions. Therefore, the samples were taken to an X-ray facility equiped with a state-of-the-art Siemens D-5000 diffractometer. The samples were mounted on "Zero-background" quartz slides, and scanned from 5 to 90 degrees 2 theta at an extremely slow speed over a period of 10 hours using Cu K-alpha radiation. Both samples T1,2-A and -B were mounted side-by-side to maximize sample area. Sample T3-A was mounted and scanned alone. As seen from the print-out of the raw data in Figure 5, the two samples are not much different with respect to detected crystalline phases. Both T1,2 and T3 contain fairly well-defined peaks at about 21.2 degrees, 23.6 degrees, 28.1 degrees, and 31.4 degrees 2 theta. T1,2 however contains two additional peaks at about 32.2 degrees and 53.0 degrees which are significant. The broad hump at about 16o and the large rise in intensity at low angles indicates a significant quantity of amorphous phase. No reflections were detected above 60 degrees 2 theta. Figures 6 and 7 show expanded views of the important range of angles with much of the background noise removed. The lattice spacing (d) is called out for each observed peak.
 

Search-match procedures were then conducted on these patterns to try and identify the specific phases present. A combination of the traditional hand search (Hanawalt Method) and automated search-match software (JADE 3.0) was used. The search was complicated by two factors:

1. the extremely low signal-to-noise ratio due to the small sample size, and
2. the presence of multiple phases in each sample Despite these difficulties,
a reasonable match was found for a mixture of three or four phases (see Appendix):

1. Anapaite, Ca_2Fe(PO_4)_24H_2O
2. Goethite, FeO(OH)
3. Phosphorus oxide
4. gamma-Iron phosphide, FeP_4

Iron phosphide may account for the two rather diffuse peaks observed in T1,2 but absent in T3.Since the effective penetration depth of the x-rays is likely to be on the order of 25 micrometer, most of the reflected signal is due to the cladding material rather than the core, which explains why there is no strong iron peak.

Metallography
 

In an attempt to learn more about the iron, or iron-alloy core of these samples, traditional metallography using an optical microscope was performed. The ground and polished cross-section of sample T1,2-B was etched using "Nital" (HNO_3 & methanol). The fact that nital etched the sample very quickly affirms the presence of an iron-rich alloy. Etching revealed a very fine (too fine to produce a good photograph) maze-like pattern of light and dark regions, reminiscent of a slowly-cooled eutectic composition. Although the microstructure did not reveal a "classic" Pearlite structure, the system is presumed to be iron- carbon, with the dark phase being perhaps cementite (Fe_3C) in a matrix of ferrite (alpha-Fe). A high percentage of finely dispersed carbon may account for the very high hardness (VH=821) as reported earlier. Ferrite is favored over austenite (gamma-Fe) due to the fact that the sample core is apparently ferromagnetic.

Summary
 

Sample T1,2 can be described as needle or lamellar in shape, with a predominantly iron core and a non-conducting, dark gray-black coating. This coating or surface layer material has Fe, Ca, P, Cl and very possibly some lighter elements (i.e. C, O) as its constituents. The phase analysis via x-ray diffraction was not absolutely conclusive due to the extremely small sample size, however the best fit to the obtained pattern suggests Anapaite, Ca_2Fe(PO_4_)2H_2O, Goethite, FeO(OH), iron phosphide, FeP_4, and phosphorus oxide, P_2O_5, as likely phases. The microstructure of the core (polished and etched) as observed under an optical microscope resembles an iron rich alloy with large amounts of carbon, probably in the form of iron carbide. The iron is likely to be alpha-Fe with a body-centered-cubic packing (bcc structure) since the samples are magnetized. The hardness of this core material is very high, in the neighborhood of high carbon tool steels.
 

Sample T3-A is a very complex mixture of materials. While the inner core is presumed to be similar to T1,2, the outer portion is comprised of a combination of many different elements and phases, depending upon the location. A majority of the cladding is the same as T1,2. However, a flake-like substance deposited on a portion of the sample is made up of Fe, Si, P, Mo, Na, and Ca. This may be some complex silicate mineral. The "neck" region of this sample may actually give a representation of the core metallic constituents: Fe, Cu, Ni, Al, & Sn. This alloy may have been oxidized for lack of protective phosphide coating. Sample T3-B was apparently a "chunk" of the amorphous/mineral cladding material with no metallic core, as evidenced by the very low density and lack of magnetization.
 

Returned Samples
 

Samples T1,2-A, and T3-A are returned in full to NIDS. Approximately 2/3 of sample T1,2-B is returned. None of T3-B is returned. These samples were ground and polished for hardness and microstructure analyses.

* Tracor Northern
** Hitachi