Obsidian Tools: Newsletter Challenge (12/04/07)

The simplest approach would be to determine the source of the obsidian. Rocks that are all of a certain type can be distinguished by the geographic locations of the strata they come from. For instance, sandstone from my area in Oklahoma, in the Arkansas River basin, has characteristics that distinguishes it from sandstone taken from other river basins.

Once our explorers locate the strata the obsidian came from, they can then look for clues that might tell them when the obsidian was mined.

This is just a preliminary guess on my part, based on a faint memory of a geology class I took about a dozen years ago.

Otherwise, they will probably look at other aspects of the tools.
Do the tools have any special markings or symbols on them that might indicate which group made them? Do the tools match the style of a known group?
I'm not sure about the chemical makeup of obsidian, but I wonder if it might contain other radioactive isotopes that can be dated in a manner similar to Carbon-14.
If the dates of the earthquakes is known, those dates can be used to set limits on the date of the tools.

How far did the earthquakes move the tools from the site of the settlement? I assume the settlement was at the edge of the lake, and if our heroes found the tools close to the middle of the lake bed, they must have beenmoved there by eathquake motion. If the site of the original settlement can be found, the movement of the tools could be traced back to a specific erathquake, or series of quakes.

It's always puzzled me when I see 'experimental archaeologists' - they demonstrate something like flint-napping. Well what happens to the tools they demonstrate making ? I could walk along the next day and think the discarded product was the 'real deal'. I can't see yet how a piece of worked volcanic glass can be dated without some additional evidence. This is of course assuming we are supposed to figure the date at which the tools were made, rather than the age of the rock itself.

For those interested ; http://www.natural-pathways.co.uk/flint-knapping.htm

and http://en.wikipedia.org/wiki/Obsidian_use_in_Mesoamerica

I have a friend who specializes in repairing broken obsidian and flint tools and yes dating the tool by analyzing the material is going to give you the age the obsidian was formed, not the age of the tool. Dating the tools is left to evaluating the shape and design of the tool itself or dating things found around it. Different era's of peoples shaped the tools differently, so this will give you an era not the exact time the tool was made.

In repairing tools he will take a broken piece and using various resins will build back the missing areas of the piece, reshape and paint it to match the rest. In this he has to have a thorough understanding of the different era's of tools to fill in the missing parts. In the circles of collectors they always keep track of fully intact pieces and those which have been repaired.

Nice info Rick. In the second link I posted it mentions something called 'Hydration dating'. I'm not sure from the description that even this could give a meaningful answer. It dates how long the piece of rock has been exposed to air/moisture. That's not exactly the same as knowing when the tool was formed from a chunk of rock. If I was wandering around thinking of making an Obsidian knife or something, I'd look for a conveniently shaped chunk that was simply lying around. A convenient splinter of such stuff may not even require too much working. I appreciate that the average Neolithic type tool would take a lot of work and skill, but the question is a little vague. At present, my best guess is the Hydration dating method, though as you mention style of design is often used in archaeological dating. Surrounding contemporary artefact's may also give clues.

Yes you are right about a small piece. If you have watched the history channel you have probably seen tools being made. Usually you find a shard which is flat and about the right thickness and start breaking off pieces to form the tool. So the hydration method provides zero information about when the tool shaped.

It is usually reasonably straightforward to identify some surfaces that were definitely created during the knapping process - so that would be the least of your problems

I also did a little home work on wikipedia, and discovered that obsidian tools are of volcanic glass origin. They are of a crystal structure that changes over a period of time, but this structure brake down is influenced by water, and as the tools where found in an old lake bed, one would think they had been subjected at some time to water, making it more difficult to determine there age. So I think that calculating there age is a combination of the possible time exposed to water (a history of the area), plus the natural brake down of the Crystal structure.

Regards JD.

Wikipedia gives information on Photoluminescence Dating as applied to sediments, but this not practical here. The usual way to determine age in the absence of sediments to measure the hydration profile. That can be very precise** when the environmental conditions are reasonably known. In this case, the temperature history will probably be reasonably well-estimated from the general geology of the region, but the water content of the surroundings will be somewhat uncertain due to the shifts in the strata; nevertheless, reasonable estimates of both the age of the artefact and the time-evolution of the water-content of the surroundings can often be made based on the detailed distribution of the water-content near the surface.
**Complexities such as the unknown concentration&strain-dependence of diffusion rate mean that there are difficulties with direct measurement of age using this technique; however, ages can often be quite well determined by direct comparison with like materials found in locations where other aging methods are available.

Other methods based on relaxation of the sub-surface stresses induced by working the material have also been used. These appear to have lost popularity of late, so I have been unable to find a web reference.

Because the tools were found on an ancient lake bed, you wouldn't neccessarily know if they were submerged at any time. In approx 850years, the area went from a cedar, cypress, pine and oak forest to a cactus, yucca, and agave desert. That kind of huge climatic swing would make it hard to estimate the fickian diffusion rate of water accurately. From what i've read, ODDSIMS (obsidian diffusion dating by secondary Ionization mass spectrometry) gets an accurate depth of hydrogen diffusion that returns an age that is more in line with C-14 dating, even if the sample is found on a riverbed.

Not my field - but my casual readings suggest that there has been some disagreement about whether the hydrogen signature is due to water diffusion or to ionic hydrogen diffusion**. I had understood that the usual source for ionic hydrogen is liquid water, so that would make it even more sensitive to the immediate environment than diffusion of moisture.

**Not having seen adequate evidence either way, I somewhat lean to the view that the origin of the hydrogen signature is water diffusion, because the non-oxygen-enhanced results I have seen are all from from "high-temperature acceleration" experiments***. However, although the relative levels**** are quite low and so subject to discussion, the naturally-aged samples appear to have some evidence of increased oxygen associated with the hydrogen. That might indicate that the reason for the better results from the SIMS measurements could be down to the higher sensitivity, and consequent ability to measure at greater depths where the concentration dependency on time was more rapid than the dependency on surface concentration (the dependence on surface concentration is linear, whereas the dependency on age approximates an error function, which would vary extremely rapidly with time at a suitable depth).
***I would also note that the apparent lack of increase in indiffused oxygen could also be due to a complementary outdiffusion of other oxides. One exploitation of this has process been the creation of high-refractive-index surface layers in LiNbo3 after heating - due to outdiffusion of Lithium oxides
****Glasses being primarily oxides, the initial concentration of oxygen is rather high...

I'll agree with you on that-my post was more a question than a statement. I will also add that volcanic silicates (obsidian is anywhere from 60-80%) are mainly acidic in nature, and any moisture in the air >20% relative humidity would allow H diffusion easily, also SIMS is a very accurate analytical tool. But; are the rates of diffusion predictable? I have no idea. The reading I've done on ODDSIMS and hydration dating imply that the limiting factor for both is an unstable climate. Although the basin teotihuacan is in has been fairly constant climate-wise since ~800CE, the previous millenium wasn't - that is my only problem with hydration and/or ODDSIMS. My choice chemistry-wise is K-Ar dating. It is the most predictable dating system outside of c14. The only problem is that its accuracy starts at >100,000 years. I'll keep pitching ideas over the plate (wickets), you swing at the one you like.

I fear they are all somewhat suspect. Long term, I believe that relief of work stress (particularly open to misinterpretation on CR4?) could be promising for age as such, because it has modest environmental dependence - but it will need a much better data-base than exists at present.

I agree your reservation about hydration - the best that appears possible is to compare the results between samples of similar porosity when the environmental history of some of them is reasonably well understood. So, once you have a starting point, you can estimate the age. But in order to minimise the sensitivity to humidity variations you need to work at depths where the sensitivity to time becomes much greater than that due to surface humidity - and that means the concentrations will be rather low. At that point you are competing with the fluctuation in the intrinsic moisture concentration during the formation of the glass; the saving grace here is that the obsidian will already be very old by the time the tool was made, so the fluctuations from the original formation will have evened out over the distances of interest. However, if you rely solely on hydration, there would appear to be nothing that you can do to compensate for temperature variations over the period; what is needed is a range of different measurement types, each with different dependency on temperature.

I think that this area of forensic archaeology currently has more potential than achievement. We would be looking at quite large error-bars.
I imagine that the official answer will most likely use subsurface hydration measurements - and this would be fair enough, as that seems corresponds to present practice

i think that the challenge poser picked teotihuacan specifically because the absence of any abundance of anecdotal records. There wasn't much specific written evidence other than trade records from other areas and civilizations. The first millenium CE and first 2 BCE relied mostly on the written word and civil styles (pottery, jewelry etc) for chronologies. Save that, I know we can date the actual obsidian, but not when some future mexican mined it and turned it into a tool. So hydration dating with its limitations seems to be the time tested best method. Unless there is a "Acme obsidian date machine 3000" out there.

I could not find any obsidian hydration rates. I found info about sample collection, perparation, and size, but not rates. How fast does this rock absorb moisture? Lets say the hydration "rind" is 10 microns. Is that on a scale of weeks, years, decades, centuries, or millenia?

There are so many ways for this method to be wrong. I think the best you could do is compare it to other known samples hydration rates and come up with an "older than this, but younger than that" relative time. (As long as the obsidian you found is not opaque.)

The early papers looked at a rim that was associated with more complex materials modification due to high-levels of hydration and related interactions. That layer would have been in the order of 30-um thick, but the environmental dependencies here meant that dates obtained in this way would be pretty dubious unless you had good knowledge of the environmental conditions.

The effects at deeper levels are somewhat less sensitive to the global environment (SFIK, this is largely because of the effect of the boundary found in the early work). However, I couldn't find diffusion rates*** on the web either (even paid-for, as the abstracts were insufficient to work out whether the data would be there); however, it must be available somewhere in the literature**.

Because of the changes near the surface, you are unlikely to get far with optical means in any case; as indicated by cuznmonkey, SIMS**** is definitely the preferred technique. Yes, it's destructive in principle, but the hole it needs to drill is only microns in diameter.

Fyz

**"Get thee to a library hence" etc.
***The closest I found was a statement that these were "variable" depending on the source of the original material.
****In principle, SIMS can have nm depth discrimination, so the depth discrimination will be limited primarily by the quality of the surface

I'll guess that the migration of other elements into the obsidian would start after the tool was made and would form a layer, or detectable ring if the obsidian was sliced in half. I do not know the rates of migration or which element would be the best to research, but known migration rates could be compared... if there is such a thing. If it's "real" it should date back to about 500 A.D. I think.

pick a random date. When someone challenges that date let them prove you wrong. Once they do, you have your date!

pick a random date

How about May 12th. It's a nice date, ..... not too hot ... not too cold....

I was going with the same theme. thow out a date and let STLengineer tell you its wrong! :)

In some cases an anthropologist or lithics expert can date a point to a particular culture based on the knapping techniques used to produce the point. The point morphology may also indicate the cultural origin.

Bill Morrow

This is where my thoughts were headed. I am not familiar with all the chemical analyses etc that are available. However my old friend, Dr. Bartolino taught of the visual identifiers. ie; stylizations.

What about the sediments or biological growth toward the edges of the tool. Couldn't something be done to these to at least figure out when he dropped it in the lake?

Volcanic glass is often aged by Thermoluminescence. The radioactive decay of potassium rich soil (lake bed) doses the obsidian and the glow curve when the material is heated to 5000C dates it accurately.

Sounds great except wouldn't you be dating the potassium in the lake bed not necessarially the tool?

No, the K39,k40, and k41 isotopes dose the obsidian when it is buried, and volcanic glass, because of its high silica in combination with undeveloped chrystal structure gives off a characteristic glow when it is heated (5000C in a N atmosphere). The glow will give an intensity regression which will correspond to the amount of time the obsidian was in contact with the potassium. Further testing can pinpoint the volcanic area the obsidian came from, and even date the eruption as long as Uranium, Thorium or Rhubidium is present. The USGD uses this method.

That will only give a latest-possible date, based on the time it might have been in contact with the most potassium-rich local soils. If the tool spent unknown times in sandy soils, peaty soils, etc. (as implied in the question) that wouldn't work too well.

Surely even early man was more careful than that with his tool(s)?

The entire city of teoticuaha was built beside a sourceless lake. While the civilization was in its zenith, the major industry was agriculture and trade. As the lake dried up, the area would have been very rich in potassium. Also, I think the changing moisture makeup of the area - especially in the centuries after the fall of the city would have made accurate hydration studies difficult even with ODDSIMS.

Won't almost anything "glow" at 5000°C? or did I miss something important?

Clue me in please?

http://geology.cr.usgs.gov/capabilities/gronemtrac/geochron/thermo/tech.html

This technique has been tweaked to register the date of the volcanic eruption, as well as get a unique fingerprint by the combination of elements (U, Th, Rh) present.

I can see the idea of the dosing but you still need to know something about the source isotope. All isotopes have a half life where the radiation decreases over time. So to know what the age of the tool you need to know the radiation level of the potassium. Also it would only lead you to the date the tool was introduced to the potassium which could be 100's of years after it was formed.

See post #29

Obsidian Hydration Dating. From the time the obsidian was tooled to present the surface absorbs moisture. The rate of absorption gives a time line.

Yes the obsidian obsorbs moisture but it is also absorbing moisture before it is tooled. So unless you tool it from the middle of a gigantic rock, which is highly unlikely based on the technology of the period, the hydration method is inconclusive.

As I understand it, the depth to which the moisture will diffuse after knapping is quite small, so you the knapping does not need to remove a great thickness of material for the variation of the post-knap diffusion to become readily distinguishable from any pre-existing background moisture.

Of course, I'm not expert in this area

Fyz

Check the Patent date(s). This should give the approximate, earliest date for the tools. If they have no date, and they say, "KORG" on them; they would be approximately 72,000 years old. If neither of these methods work out, Obsidian Hydration Analysis may be in order. I hate this though, as it is slightly damaging to the specimin.

Use Obsidian Hydration Dating. IS Frey

Just a thought. Wouldn't a lake bed site at Teotihuacan be volcanic in nature and hence acidic. Would that have some bearing?

Here is what I think will do the trick:

1. Use X-ray diffraction, scanning electron microscopy, and electron microprobe analysis. These techniques will enable to determine the tools' chemical composition as well as the crystal structure of the tools' minerals.
2. Use Rock-magnetic dating methods to establish the age of the artifacts collected at the Teotihuacan site and compare them to the soil history of a nearby site that will contain a more-complete record of sediment deposits through time. Some ash layers in the Teotihuacan area can be directly dated by the 40Ar/39Ar method, and others, by virtue of their distinctive chemical composition, can be matched to correlative dated layers outside the Teotihuacan area.
3. Tool-making techniques, procedures, raw materials and patterns of discard and production debris left behind on the landscape from the tool-using activities are readily apparent and found in the archaeological records which correlate with evolutionary changes.
4. Factoring in other known geological events, such as the natural movement of the Earth's magnetic poles over time, piece together a detailed age sequence for the archeological levels.

The answer suggests that a "compositional line profile (I assume that is versus depth) can be utilised to determine the age of the obsidian glass (tool)". However, given the unknown environmental history in the present case, that is at variance with the drift of the reference, according to which can should be heavily qualified to read "can in principle at some time in the future".

Specifically, the summary in the linked reference indicates the difficulties that remain to be overcome before this technique can be regarded as fully reliable under the stated conditions: "may establish appropriate mathematical expressions, which will (should that be 'would'?) make the method essentially intrinsic and independent of environmental factors, since they (the environmental factors) are embedded within the diffusion equations".
The included references do not make light of the problem of determining diffusion rates either - some of them addressing the difficulties due to the diffusivity changing with sample history, and others into the issues of variability between samples.

Nevertheless, this appears to be the least inaccurate physical dating method available at present - though in this case I suspect comparison with known tools from the same culture is likely to provide better information.

I wonder if this inaccuracy is really true? Lots of you think it will be inaccurate, but I would postulated that the stone was formed many millions of years ago ( Cretaceous age, max 140 odd million years), but it was probably only formed into a tool inside the last 1 million years, probably a lot less....could even have been only 10,000 years ago for example.

Therefore the length of time the stone was just stone was infinitely longer than it was as a formed tool, so processes that measure the water loss, could possibly more accurate than we are giving them credit for.....

....just a thought from a complete archaeological amateur!!

According to the cited reference, the specialists in the field seem to believe the inaccuracy is true - it appears that only outsiders believe that the method is currently capable of providing aging data where environmental conditions are unknown.

However, it does appear that the specialists believe that in principle obsidian dating could become accurate enough to be useful without knowing the environment.

But when the reference given by the questioner to support his solution states that the techniques are not yet capable of providing the quality of information implied, it makes sense to take notice.
The remainder of my previous comments were just to indicate where to look. Admitting that I've not achieved my objective, I'm resorting to a list; the problems identified as requiring work to achieve dating without a priori knowledge of the environment include:
. Obsidian is of variable composition, and both the linear and non-linear diffusion properties vary with composition. (The best known variations are initial water content and porosity, but variations in soluble salt content can also contribute).
. Both H2O and H+ diffusion are thought to contribute, but oxide diffusion and the high oxygen content of the basic glass makes discrimination difficult.
. The diffusion properties of obsidian glasses change according to the moisture levels in the glass. Part of this is driven by strains and electro-potentials, so it is not sufficient just to know the local concentration.
N.B. As you surmise, the interior of a given piece of glass will be very uniform before the tool is made. But, as the references, the above points (and even my earlier comment) clarify, initial nonunifrmity is not the major problem.

You may recognise similarities with the diffusion of dopants into silicon. But there are major differences: the diffusions of interest in silicon can be performed in sensibly short times; and monocrystalline silicon would be much more straightforward to model theoretically than a multiconstituent glass.
Even so, the silicon analogy is illuminating - consider a silicon surface that is exposed to a well-behaved dopant at a constant surface concentration; the diffusion rate of the dopant will depend on temperature, but the profile when the 50% level has reached a given depth will be independent of the temperature history. That means that for this simple case there is no way to discriminate between times at different temperatures - and therefore no way to determine exposure time unless you already know the temperature history. A worst-case view would be that the same should apply to obsidian dating - and so far as I can judge, the dating methods so far assume that the diffusion rate is within certain narrow bounds determined by the site.
However, there is potentially more information (as with As into silicon, for example), but even if we knew everything about the diffusion properties of the obsidian we would have two variables other than time (surface concentration and temperature) to unravel. Doing that without assumptions would require very accurate data, even if we had exact models of the diffusion process under all conditions**.

That's probably far too long...

**I did once see some results from reverse-engineering adds-on to a silicon process. In the end indistinguishable performance was achieved - but with very different time-temperature profiles. The irony was that the original version of the process was apt to produce defects in the silicon with only very small deviations from the specified profiles, whereas the reverse-engineered version turned out to be quite robust.

#2 was a tad extreme. Hydration dating from sites like Teotihuacan shouldn't be much less accurate than +/- 30%, so you could tell if it was a modern experiment (should that read fake?)

Fyz