Date: Thu, 06 Jul 2006 13:02:35 -0500 To: bccmeteorites@ispwest.com From: Randy Korotev <korotev@wustl.edu> Dear Mr. DeRusse: 1) I have no knowledge of the debate you mention on rockhounders.com. I have never visited the site. I was not the person who engaged in the cowardly debate. I despise tactics of the type you describe. 2) I did, however, remove your link, and for that I am sorry. What I describe below is not a justification, just an explanation. I did not know until last week that there was a lunar meteorite entry in Wikipedia. A colleague called my intention to it and suggested that I might want to fix the errors. I did not know until that time that users could make such changes. Without reading any of the guidelines, I just jumped in an started editing. That was a mistake. (This type of behavior is also a constant source of irritation to my wife.) I assumed that I was editing material written by Wikipedia staff. When I came to your link, I clicked on it, didn't look at the URL, arrogantly said, "that's not the best info on this subject," and deleted it. The thought process lasted half a second. It wasn't until sometime afterward that I discovered the log file, saw what I had done, and had an "oh, crap" reaction. I decided that if you wanted to fix it, you would, and you'd let me know. That's what happened! I'm truly sorry; I shouldn't have done it. Sincerely, Randy Korotev

Mr. DeRusse thanked me for my apology. I responded with the following message.

Randy Korotev
14 July 2006

---

Looking at the Data

last updated: July 19, 2006

What evidence is there that BCC9601 is a lunar meteorite? Until the last few days (I first wrote this on July 19, 2006), I had not closely examined the information and data presented by Mr. DeRusse on his website. This essay will focus on two pieces of data he presents - the mineralogy of the rock as determined by XRD (X-ray diffraction) analysis and the chemical composition of the rock as determined by EDS (energy dispersive X-ray spectroscopy).

Mineralogy

There are only 4 minerals that make up 98-99% of the crystalline material of the lunar crust and all lunar meteorites: plagioclase feldspar (mainly, anorthite, the high-Ca, low-Na variety of plagioclase), pyroxene (orthopyroxene, clinopyroxene, low-Ca pyroxene), olivine, and ilmenite. There are many kinds of minor minerals, but plagioclase, pyroxene, olivine, and ilmenite, plus glass formed from melting of these minerals, account for 98-99% of the mass of most lunar rocks (breccias and basalts) and regolith. (The next most abundant phase is probably iron-nickel metal from meteorites that strike the Moon.)

[Aside: Pulverizing a lunar rock on Earth does not yield lunar regolith. "Regolith is the term for the layer or mantle of fragmental and unconsolidated rock material, whether residual or transported and of highly varied character, that nearly everywhere forms the surface" (reference 4, p 84). Lunar regolith was produced and modified by natural processes - meteorite impacts and various "space weathering" processes. Powdered rock is not regolith.]

On 18 July 2006 I found the following information on Mr. DeRusse's website. It is from a letter and report from K/T GeoServices of Argyle, TX, dated June 25, 2002. The report is for a sample with the "Sample ID" of "Lunar dust." I presume here that this is a pulverized sample of BCC9601. The details of the report are below:

I stress two points about these data:

1) XRD is a great technique for identifying and quantifying the relative abundance of minerals. It is rarely used by meteorite scientists as a means of determining whether or not a rock is a meteorite, however, because we prefer to use petrographic techniques (a thin section examined with a petrographic microscope and electron microprobe) which provide additional information, namely, the composition of the minerals and their textures (how the minerals are shaped and assembled into a rock, which in turn tells us how the rock formed).

Bottom Line - If it were my purpose to determine whether the mineralogy of a rock was consistent with that of a lunar rock, I would have picked a different technique, one that is less ambiguous and provides more information. XRD analysis, for example, cannot determine the existence of glass, which is often an important constituent of lunar meteorites. (K/T GeoServices state in their report: "XRD methods can quantify crystalline material only." Glass, by definition, is not crystalline.) It also cannot, for example, determine the composition of plagioclase.

2) XRD data can never prove that a rock is a meteorite because all of the major minerals in meteorites (except iron-nickel metal, which is invisible to XRD) also occur in terrestrial rocks. In other words, the presence of plagioclase, pyroxene, olivine, and ilmenite is consistent with a lunar meteorite, but their presence does not prove that a rock is a lunar meteorite because these 4 minerals are also common constituents of Earth rocks. XRD can, however, provide strong evidence that a rock is not a meteorite if it shows the presence of minerals that do not occur in meteorites.

In that regard, the analysis above sets off several warning flags. Quartz is exceedingly rare in meteorites, though some lunar basalts contain a much as a few of percent quartz. [See the discussion below and Tables A5.22, A5.23, A5.24 on pages 5-207 and 5-208 of reference 5.] K-feldspar is also rare in meteorites and lunar rocks. Phlogopite, a mica, has never been observed in any kind of meteorite. Micas contain water as part of their crystal structure, and there is no water on the Moon. No mica or other hydrated mineral has ever been found in an Apollo sample. Quartz, K-feldspar, and phlogopite are all common in terrestrial minerals, however.

Bottom Line -Taken at face value, there is nothing about the XRD analysis that suggests to me that the rock might be a lunar rock. So far, it appears to be a rather mundane terrestrial rock of the granite clan.

Chemical Composition

On his website Mr. DeRusse provides a report from AMIA Laboratories on the chemical composition of two samples of BCC9601 ("Black" and "Gray") and several other samples. The data were obtained by EDS analysis (apparently, July 15, 1999). The analytical report, which I took from Mr. DeRusse's web site, is reproduced below:

EDS analysis provides data for those elements that geochemists and geologists call the "major elements," that is, those elements that make up the bulk of the rock and have concentrations of about 1% or more. It provides no information about minor and trace elements. The suite of elements determined by EDS is, however, sufficient for a geochemist to state either "this is not a meteorite" or "this rock might be a meteorite." Put another way, the major-element composition is never sufficient to positively demonstrate that a rock is a meteorite, but it can prove that a rock's composition is inconsistent with that of any known meteorite. ("Necessary but not sufficient.") Again, EDS is not used much by meteoriticists because it is a semiquantitative (approximate, not exact) technique. There are better (more exact, less ambiguous) techniques which, of course, are more expensive to do.

Two things about the data jump out immediately as suspicious:

(1) The sum of the concentrations of the listed elements is exactly 100.0% for each sample.

(2) There is no column for Mg (magnesium)! All meteorites contain Mg. Where's the Mg?

Sum of Elements and Oxygen - I'll discuss the first issue first. EDS cannot accurately determine the absolute concentration of an element, it can only determine relative abundances, that is, ratios like Na/Si, Al/Si, etc. It is common when reporting EDS data to adjust (normalize) the data so that the sum of all elements measured is 100%. The fact that the sums are all exactly 100% indicates to me that this normalization was done - and that's OK if it's understood that when all elements actually present are considered, the sum would be greater than 100%. The report acknowledges that "Carbon and oxygen were present in all samples," but no data are reported. Put another way, the report only includes data for the metallic elements and silicon.

In an igneous rock, the concentration of O (oxygen) is about 50%. So, I know that the actual concentrations of all the elements in the table are actually about half the values given. Geologist prefer to measure major elements with techniques like X-ray fluorescence spectroscopy and wave-length-dispersive X-ray spectroscopy because these techniques give the absolute concentrations of elements, not just ratios. Neither of these techniques measures oxygen, however. So, the standard procedure in geology and geochemistry is to measure, say, the absolute concentration of each element, but report the concentration as the oxide. AMIA labs did not do it this way because they probably mainly analyze metals that do not contain much oxygen. Igneous rocks consist of silicates and oxides. For silicate rocks, if one takes all of the major elements and reports the concentration values as percent oxide (SiO₂, Al₂O₃, FeO, etc.), not percent metal, the sum-of oxides comes out to about 100% because oxygen is the only major anion in silicate rocks. In fact, if the sum-of-oxides is not in the 98-101% range, then we usually assume that "something is wrong." Either the analysis is bad or the rock contains a significant concentration of some element not measured (typically, H in water or C in carbonate).

In order for me to compare the results for BCC9601 above to compiled data for lunar meteorites, I have to convert the BCC data to the oxide equivalent, which is easy to do (see conversion chart), and renormalize the sum-of-oxides to 100%. In the figures below, I plot the oxide values obtained in this way. By lunar convention, I report all Fe (iron) as FeO. For analysis of Earth rocks, it's typical to see all Fe reported as Fe₂O₃. There is no Fe(III) on the Moon, however; all Fe is Fe(0) and Fe(II). On Earth all Fe is Fe(II) and Fe(III). All of this means, for example, that the ~10% Fe reported for BCC96001 (table above; metallic elements only) is equivalent to ~7% FeO (whole rock, including oxygen).

Regarding Mg - Geochemically, a curious aspect of the data is that there is no concentration column for Mg (magnesium), a major element in all stony meteorites. The absence of Mg concentration values means either that concentrations of Mg were not determined or that an attempt to determine Mg was made but the concentrations were below detection limits, presumably about 1%. I assume that the latter explanation is correct on the basis of the X-ray spectra provided by Mr. DeRusse, one of which is reproduced below:

If Mg occurs in a rock at concentrations of 1% or more, the peak lies between that for Na and Al (because Mg lies between Na and Al in the periodic table). There is no Mg peak in the spectrum, yet there is a tiny peak for Ti, which is present at a concentration of little less than 1%. Therefore I assume that the Mg concentration was not reported because it was below the detection limit. For the purpose of the normalization I discussed above, I assume that BCC96001 contains 1% MgO.

So (finally) - Below are some plots that compare the composition of BCC96001 (two points, one for the black and one for the gray sample) with those for all lunar meteorites for which I have data. (Each green square represents a different stone.)

	Lunar meteorites plot along a linear trend on this plot because nearly all the Al2O3 is in plagioclase and all the FeO is in the other three minerals, pyroxene, olivine, and ilmenite. The only way a lunar rock can plot to the left of the trend is if it contains some other mineral, like quartz, that contains neither element. [I go into more technical detail about this correlation in references (2), (3), and (4).] No lunar meteorite contains more than 1% quartz or any silica polymorph. Many terrestrial igneous rocks, like granites, contain quartz. These minerals drag the composition off the lunar trend toward the origin of this plot.
	The ratio of Al2O3 to CaO is rather constant in lunar rocks because plagioclase is the only important carrier of both elements, except that in basalts some CaO is carried by clinopyroxene. BCC9601 is significantly depleted in CaO relative to all lunar meteorites and is depleted in Al2O3 compared to meteorites from the feldspathic highlands of the Moon.
	A characteristic feature of lunar rocks is that they are depleted in alkali elements like Na (sodium) and K (potassium, below) by about a factor of 10 compared to terrestrial igneous rocks. This is why alkali feldspar is rare on the Moon and lunar plagioclase is very poor in the albite (sodium) component. In the lunar highlands, plagioclase is typically 96% anorthite and 4% albite. (If you've go no sodium, you can't make albite.) Given the high whole-rock concentration of Na2O, I predict that the plagioclase in BCC96001 is much more albitic than is lunar plagioclase.
	The bottom three plots also show how much richer in silica (SiO2) BCC9601 is compared to lunar rocks. The excess silica is carried by the quartz, which is very rare in lunar meteorites. With 71% SiO2, BCC is typical of terrestrial siliceous volcanic rocks and some granites and granodiorites. BCC is rich in K2O because it has a lot of K-feldspar. Lunar meteorites contain very little K feldspar because the Moon contains very little potassium.
	Among all meteorites, the feldspathic lunar meteorites have the lowest concentrations of MgO. However, BCC9601 contains even less Mg than the most Mg-poor lunar meteorite. The analytical report for BCC9601 did not provide data for MgO because the concentration was below detection limits. On this plot, I assume the concentration is 1% MgO. If it were as great as, say, 5% (as with K2O), it would have been detected and a value would have been reported.

It is self evident from these graphs that BCA9601 does not compositionally resemble any lunar meteorite.

So, What Is It? - Rock names like "basalt" and granite" are usually based on mineralogy, texture, chemical composition, and inferred mode of formation. Because of its high silica (SiO₂) concentration, the composition of BCC is most similar to terrestrial andesites, granites, granodiorites, and rhyolites . I am not a terrestrial geochemist, I don't know all the terrestrial rock types, and I don't have a large database of compositions of terrestrial rocks at my disposal. In the plots above, I plot data for high-silica geochemical reference standards (Govindaraju, 1994) from around the world. The composition of BCC doesn't match perfectly with any of them, but BCC is clearly more like a terrestrial granitic or siliceous volcanic rock than it is like any lunar meteorite. If there's a terrestrial geochemist or geologist out there who has a better idea, please let me know.

Lunar "Granite" "What's that green 'X' in the plots?" some of you are asking. There is a rare rock type on the Moon that consists of K-feldspar (alias, potash feldspar), silica minerals (quartz, cristobalite, trydimite), and glass that is sometimes called granite and sometimes called felsite. All samples are small and they vary considerably in composition. The green X represents the average composition of all the lunar granite/felsite data that I have in my database. Yes, BCC9601 more closely resembles a lunar granite in composition than it does any lunar meteorite, none of which are granites. But, consider the following information, all of which is in the form of annotated quotations from reference 5 (p. 5-33 and 5-141) regarding lunar granite and quartz: "Small fragments consisting largely of silica and potash feldspar, referred to as granites and felsites, have long been recognized in highlands rocks ... All are fine-grained by terrestrial standards ... Only 2 fragments greater than 1 g have been recognized ... Many small fragments are brecciated or even glassy and remelted." [Note: BCC is big.] "granites/felsites have higher modal silica and potash feldspar [than quartz monzodiorites], and commonly little pyroxene or plagioclase feldspar ..." [Note: In contrast, BCC contains 50% plagioclase + pyroxene.] "The largest lunar granite clast yet found, from Apollo 14 breccia 14321, weighs 1.8 g and contains 40 vol % quartz ..." [Note: BCC is much bigger, but it contains about the same proportion of quartz.] "Consistent with the general absence of hydrous minerals on the Moon, the lunar granites do not contain mica or amphibole, as do granites on Earth." [Note: I think this is why some petrologists reject the term "granite" in favor of "felsite" for these rocks. BCC contains 4% phlogopite, a mica.] "The most common texture in granites/felsites is that of a granophyric (sensu lato) intergrowth of a silica phase and potash feldspar ... Accessory minerals include pigeonite, augite-ferroaugite, fayalite, ilmenite, zircon, and phosphates; silica-potash-rich glass is common ..." [Note: I'm no petrologist, but to me this description of the texture is inconsistent with that of the photomicrographs of BCC9601.] "The most common silica mineral in mare and KREEP basalts is not quartz, but cristobalite, which can constitute up to 5 vol % in some basalts. " [Note: The XRD results for BCC indicated quartz. One important advantage of XRD is that it can, in fact, distinguish between quartz and its polymorph, cristobalite.]

 

So, BCA9601 does not have the characteristics of a lunar granite. Although BCC is compositionally more like a lunar granite than any lunar meteorite, the similarity is a necessary consequence of the fact that both BCC and lunar granites are dominated by silica minerals and K feldspar. Rocks dominated by quartz and K feldspar are common on Earth and rare on the Moon. Thus, the weak compositional similarity does not really offer any support to the hypothesis that BCC9601 is of lunar origin.

Bottom Line - Now that I've taken the time to examine the mineralogical and compositional evidence provided by Mr. DeRusse, I conclude that none of the evidence even hints to me that BCC9601 might be a lunar rock. In total, the evidence strongly argues that it is not a lunar rock. The mineralogy and composition are perfectly consistent with terrestrial origin. I know that siliceous volcanic rocks occur in west Texas, but I don't know enough about Texas geology to know how close to Austin they occur. The rock looks like it's been rounded by abrasion in a river, so it may well have come from some distance away.

What It Would Take to Convince Me That I'm Wrong - If it were demonstrated that BCC9601 contains those radionuclides that can only be naturally produced by exposure to cosmic-rays while a rock is in space (¹⁰Be, ¹⁴C, ²⁶Al, ⁴¹Ca, ³⁶Cl), then I would admit that it is a meteorite. The presence of comogenic radionuclides is always the ultimate test for "Is it a meteorite?" If the cosmic-ray exposure history were consistent with that of other lunar meteorites, then I'd admit that BCC9601 was a lunar meteorite.

Bottom Bottom Line - If someone wants to sell you a 5-legged dog, but the dog only has 4 legs, it's probably just a 4-legged dog.

Randy Korotev
19 July 2006

See also:
How Do We Know That It's a Rock from the Moon?
Chemical Classification of Lunar Meteorites

References

(1) Govindaraju K. (1994) 1994 compilation of working values and sample description for 383 geostandards. Geostandards Newsletter, v 18, p 1-158.

(2) Korotev R. L. (2005) Lunar geochemistry as told by lunar meteorites. Chemie der Erde, v 65, p 297–346.

(3) Korotev R. L., Jolliff B. L., Zeigler R. A., Gillis J. J., and Haskin L. A. (2003) Feldspathic lunar meteorites and their implications for compositional remote sensing of the lunar surface and the composition of the lunar crust, Geochimimica et Cosmochimica Acta, v 67, p 4895–4923.

(4) Lucey P., Korotev R. L., Gillis J. J., Taylor L. A., Lawrence D., Elphic R., Feldman B., Hood L. L., Hunten D., Mendillo M., Noble S., Papike J. J., and Reedy R. C. (2006) Chapter 2. Understanding the lunar surface and space-moon interactions. In New Views of the Moon, pp. 83–219. Reviews in Mineralogy and Geochemistry, Volume 60 (Jolliff, B. L. M. A. Wieczorek, C. K. Shearer, and C. R. Neal, eds.). Mineralogical Society of America, Washington, DC.

(5) Papike J. J., Ryder G., and Shearer C. K. (1998) Chapter 5. Lunar Samples. In Reviews in Mineralogy, Vol. 36, Planetary Materials (ed. J. J. Papike), pp. 5-1–5-234, Mineralogical Society of America, Washington.