RHX Dating Update

My collaborators and I mailed out some reports today about Fired Clay Ceramic Rehydroxylation Dating! I am pleased to say that we've completed our work that was supported by a small grant from the National Science Foundation's Division of Behavioral and Cognitive Sciences (Award #1112327) and an Fellowship Award to Patrick Bowen from the DeVlieg Foundation.

I know many people are very excited about RHX dating (and also very skeptical).  I was very excited about the proposed technique when it was published in 2009. This technique proposes to use the clock-like, nano-scale process by which water molecules bond with clay mineral crystals.  By taking an old ceramic sherd, heating it to remove any mass caused by humidity (atmospheric water), measuring it's mass, then firing the sample at a higher temperature to drive off all the rehydrated (adhered) water and rehydroxylated (chemically bonded) water, one can measure the sample's mass without any water present.  After that, carefully tracking the mass of the sample as it quickly starts to reabsorb water from the air allows you to generate an equation that models the time past, the water mass gained, and the rate at which this occurs.  So long as you can match the temperature in the room to the average lifetime temperature of the object, a bit of math lets the lab technician calculate how long it took for the sample to reach the weight at which it was discovered by archaeologists. All my posts about RHX dating are all here.

In 2010, we applied for an NSF grant to study this process and see if we could replicate the findings of the UK researchers that had proposed it.  We did win that grant, but we did some background work and tried to replicate their study. We published our results in 2011.  With that publication under our belt, we reapplied to NSF for more funding to upgrade our lab equipment to match the quality of that being used by the UK team.  At first, we were rejected, but then the NSF found a bit of money that allowed us to improve our instruments.  Using that grant, we purchased a new Citizen CM11 Microbalance along with a Coy Humidity Control Glove Box. This microbalance allows us to measure 0.001 mg of variation in mass, fully two orders of resolution and sensitivity higher than the balance that was available at Michigan Tech before the NSF award! It's was a lesser piece of equipment than we'd hoped to purchase, but it has permitted us to take our experiments to the next level of research quality.

In November-December 2011, We tested the station to determine the stability of humidity and temperature within the box as well as the stability of microbalance while it is reading in the box chamber environment. Between December 2011 and August 2012, Jarek and Patrick collected and analyzed the mass-gain data corresponding to rehydration/rehydroxylation of Davenport sherds and some brick samples from Houghton, Michigan.  After running tests on intact sherds, we also ran some pulverized samples. Jarek and Patrick worked over the data to examine the influence of relative humidity on mass gain and time curves.  From that analysis, we just submitted an article for review and publication. This one deals with the question of which model (expressed in an equation) best describes the samples behavior as they rehydrate and rehydroxylate over time. We also raised some questions about the micropore structure of the sherds and how that influences the model.

Figure 2 is from our report to NSF. It shows the fractional mass versus time obtained for the Davenport pottery sample under constant temperature and humidity (plot on the left hand side). 

Figure 2. Compiled fitting results for intact Davenport sherd in both linear time (left column) and time^1/n with n= 3.77 (right column). Results were obtained at 22^oC and 20%RH.

                  

                   

                  The right-hand plot in Figure 2 also shows the resulting theoretical fit using the following empirical equation describing the mass gain (m) (or fractional mass gain):

where t is the time, β is the term representing physically bonded water, α is the rehydration kinetics term, γ is the rehydroxylation rate, and n is the rehydroxylation exponent.  

Agreement between experimental data and theoretical model is remarkable!  

We found problems that present the researchers with challenges to making RHX a standard archaeometric tool for dating fired clay ceramic artifacts.  I will write more about those in another post, but I can't say too much until our article is published. I'm also excited to report that our initial success led to a new award from NSF (BCS-1219540)!

Our new collaboration will allow our team at Michigan Technological University to coordinate an experiment involving five different international teams of faculty and students (Michigan Technological University, Arizona State University, California State University-Long Beach, Tel Aviv University, and The University of New England) working on new RHX experiments, advised by a team of UK scientists, many of whom initially proposed this technique, and continue to work on it's refinement (Universities of Manchester, Edinburgh, and Bradford).  We will send a set of blind samples for testing to five labs including materials scientists and archaeologists from seven different universities in four countries. 

I am very excited!

News update: the good, the bad, and the ugly.

The good news:

Now that the academic semester is almost at an end, I will spend a bit of time on the project preparing some artifacts for a new exhibition!  The Church of Jesus Christ of Latter-day Saints' Church History Museum is going to install a version of Potters of the Gathering in their lobby.  The exhibit will open May 7th.  Kirk Henrichsen is using pottery to remind people about nineteenth century foodways and domestic life, to explore the ideal of self-sufficiency in Latter-day Saint communities, and to show people how important archaeology is as a tool to increase our understandings of the past.

Mr. Henrichsen has gathered together some new material for this show.  He will be using most of the important objects from the original exhibit- including the Thompson Collection, the Utah State Parks Collection (including the Deseret Pottery artifacts), and the Utah Pottery Project collection.  He has also found some other cool items to include, such as a Deseret Agricultural and Manufacturing Association Medal from the Territorial Fair, which will be displayed with some award winning pots!  It should be a great show and will be the first time most of this material has been exhibited to the public in Salt Lake City.

The museum has invited me to give a talk at the opening and I am excited to see the new installation.

The bad news:

Both organizations declined the proposals I'd written seeking funds for more work.  One of the grants would have supported additional fieldwork. The other sought funds to allow more experiments to develop and refine ceramic rehydroxylation dating (RHX dating).  No matter how many times it happens, reviewers rejections always sting a little, even when accompanied by encouragement to "revise and resubmit."

Either way, the lack of support funds means that we will box up the artifacts and put more of the analysis on hold for the summer.  I'll be co-teaching the 2010 Industrial Heritage and Archaeology Field School this summer at the site of the Cliff Mine in Keweenaw County in Michigan's Upper Peninsula.

I'll continue to collaborate with my friends in Materials Science.  Since Patrick Bowen was awarded two grants for undergraduate research, we will move as far ahead as we can investigating reydroxylation.  I think he will work with Jarek Drelich to characterize the clay and ceramic used by the Davenports and try to figure out the significance of the different water-absorbtion processes at work in ceramics.

We'll reflect on the reviewers' comments for a few days and then meet to plan the strategy for the rest of the summer and the fall term.

Congratulations to Patrick Bowen

I am pleased to congratulate Patrick Bowen, Michigan Technological University undergraduate student in Materials Science and Engineering.  Mr. Bowen won two major fellowship competitions and he was awarded funds for his ongoing collaborations studying rehydroxylation dating of archaeological ceramics.  He was awarded one of Michigan Tech's Summer Research Experience for Undergraduates (SURF) Fellowships as well as a Michigan Space Consortium Undergraduate Research Fellowship.

Patrick has been an important part of the interdisciplinary research team that Jarek Drelich and I put together to assess this newly published dating technique.  Mr. Bowen, Helen Ranck, and Jessica Beck, three MTU undergraduates in the Departments of Materials Science and Social Science, did exceptional work designing tests to assess the usefulness of RHX Dating for the Utah Pottery Project.  Patrick is presenting the preliminary results of their analysis at MTU's 2010 Undergraduate Research Expo.

The SURF and MSC Fellowship programs are both very competitive and I am proud of Mr. Bowen for his successful applications!

We are waiting to hear decisions on the proposal I wrote to support more field research as well as a National Science Foundation proposal to support more research into Rehydroxylation Dating.

I'm keeping my fingers crossed on our outstanding proposals!

Rehydroxylation Dating-- 10-21-09 update

Our collaborating student researchers are making steady progress in their efforts to replicate the ceramic rehydroxylation dating technique published by Moira Wilson and her colleagues over the summer. If you wish to read all my posts about rehydroxylation, click here.

Helen Ranck, Patrick Bowen, and Jessica Beck have been working on different parts of the problem and they've learned a great deal so far.

Here is one of Helen's graphs that describes the mass gain of one of her test samples:

She and Patrick have been trying to find out the best way to keep the sherds at a constant temperature and atmosphere while the fragment absorbs water. Patrick has discovered some key changes in practice that have really helped reduce the variation in calculated dates, bringing the projections closer to our expectations. Jarek Drelich, associate professor in Michigan Tech's Materials Science and Engineering department, has been very helpful working with them both.

Jessica Beck has finished some of her testing. She calculated the porosity of a group of the sherds using a methodology outlined by the National Institute of Standards and Technology (NIST). Jessica discovered that the Davenports' earthenware ranges from 10%-15% porous, and that both the median and the mode will be around 13% or 14%. I'm looking forward to her final conclusions and her estimates on firing temperature as revealed through her other testing!

Rehydroxylation Dating-- testing a new tool!

While we were digging in the field this summer, a team of seven materials scientists (led by Moira Wilson) from the University of Manchester and the University of Edinburgh and Ian M. Betts, an archaeologist with the Museum of London, published a paper in the Proceedings of the Royal Society A called "Dating fired-clay ceramics using long-term power law rehydroxylation kenetics."

Three Michigan Technological University students, Helen Ranck, Patrick Bowen, and Jessica Beck, are trying to replicate the technique.

Materials scientists and engineers have known that ceramic minerals slowly reabsorb water from the environment after they are fired. Dr. Wilson's team discovered that the rate at which environmental water recombines with clay minerals as hydroxyls is governed by a kinetic law at the nano-scale. They found that the rate was influenced by temperature, but was not changed by the quantity of water present in the environment.

Scientists could actually measure the rate of water mass gain for any given ceramic fragment by heating a sample in a kiln and then waiting and measuring the increase in mass as the clay molecules slowly recombined with environmental moisture at a known temperature. Lab researchers can then calculate the date of firing with these known measurements:

1. This determined rate of water mass gain.

2. The mass of the sample after excavation (when it contained all the re-bonded water).

3. The mass of the sample after the test firing (mass of the ceramic fragment minus the molecularly recombined water mass that it had absorbed since it was fired).

4. The average temperature through time the sample experienced since firing in the depositional environment.

Of course, including the +/- error, universal in archeometric dating.

If this technique works as well as the authors assert, it will add another powerful tool to archaeological techniques around the world. It will also revolutionize the Utah Pottery Project. Remember that one of the main goals of our archaeological research is to reconstruct the learning process through which these individual potters or potting groups, such as the Davenport family, adapted their skills and knowledge to Utah's new environment and raw materials.

One of the biggest problems we've had is that we can not use the ceramic fragments themselves to date the features full of broken kiln wasters, such as Andy's excavation of the cellar feature this summer. We rely upon stratigraphic clues (that waster pile is older than X, but younger than Y) or other artifacts found in the deposits, like stylish imported ceramic fragments, which can be dated. The features full of industrial wasters only rarely also contain other clues, however.

There are some other archaeometric techniques archaeologists use to date ceramics directly, such as Archaeomagnetic Dating and Optically Stimulated Luminescence. These techniques are useful also, but are either very specific to only work on kiln foundations (archaeomagnetic) or expensive and require expertise we don't have (OSL or TL). Either way, most of those techniques work better when applied to the distant past, and not the nineteenth century.

Historical Archaeology is like ethnoarchaeology in many ways. Given that we are studying people and sites in the historic period, we know a great deal more about accurate temperatures than archaeologists studying deep antiquity. I have the advantage of knowing the year, and even the month, during which some of these ceramics were fired. I hope we can test and refine the technique to higher levels of precision.

If this technique works as described, we will have an inexpensive tool that will allow us to build direct chronologies from the waster fragments. We will be able to sort the Davenports' waste into categories and know which ones reflect the steep learning curve from the 1850s and which pieces have clues about ongoing improvement and the training of the next generation of potters!

Moira Wilson corresponded with me over the summer while we were digging, and I appreciate her encouragement to consider application of her team's work. I am excited to be working with this puzzle with Jaroslaw W. Drelich, an associate professor in MTU's Department of Materials Science and Engineering. The original article appeared here. Useful discussions and interesting commentary about rehydroxylation dating appeared here, here, and particularly here.

Maps, part 2: Archaeometry

People often ask me if I can tell where a pot was made by studying it closely and reading the clues of form, fabric, style, and glaze.  I still have to answer them, "Not yet, not yet."  We are just starting the detailed study of each pottery shop. 

In addition, the problem is really pretty complicated. Right now there seem to be some elements of style that can be attributed to individuals, but we must have caution.  A decorative pattern or shape of form may appear made by one potter in one location, but could be imitated or shared by another potter somewhere else.  Take this example of a decorative pattern which archaeological research showed to be made by Frederick Petersen in Salt Lake City.

 This pattern appears on two antique pots we will exhibit in our show at the Iron Mission  Museum this summer.  Can we conclude that these pots were made by Frederick Petersen?  Not yet.  Even if we assume that all the other potters in Utah refrained from crassly or simply copying each other's decorations and styles, consider these possibilities:

1. Frederick Petersen learned to make pots as an apprentice of his uncle, Niels Jensen.  Frederick, along with Mr. Jensen and his two other apprentices, Jacob Hansen and Frederick Hansen, all immigrated to the Salt Lake Valley in 1852.  The vessel forms the three young apprentices made would have been informed by common elements of style derived from their Danish heritage and their common experience under their master's tutelage.  As the other apprentices grew up and established their own shops, they later worked in Brigham City and Hyrum, Utah. Could they not have shared some common decorative strategies?

2. While not as important in this specific case, we know that clay workers moved around between potteries.  Different potters collaborated at different times.  This means that one person could pick up a decorative technique, vessel form, or glaze type while collaborating and then carry those techniques forward to later work.

3. Of course, we also have many examples of parent-to-child transition of craft practices.  Some families established 'craft dynasties' with multiple generations of potters.  The Roberts family serves as the best example of this pattern, since the generations of this family spread out over the entire state, from Vernal to Panguitch.

So style remains a problematic tool right now for attributing vessels to specific makers.  There will be several examples of these puzzles in the show, including the Petersen example above.  Only a few potters stamped their ware.  Of those those that used maker's marks like on the previous post, they only stamped a fraction of what they made.  

After further study and excavation, we do expect to be able to attribute specific styles to specific individuals or at least pottery shops, but we have another tool to approach this problem: archaeometry.

Archaeometry is an area of scholarship that involves measuring the characteristics of archaeological artifacts, generally using materials science.  In this case, we are using two well-established tools to examine Utah-made pottery, Instrumental Neutron Activation Analysis (INAA) and Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS).  The details of the technique aren't important to review here, but those curious can find more detail at sites like the Archaeometry Lab of University of Missouri's Research Reactor (INAA here and LA-ICP-MS here).

Both techniques involve measuring the quantity of different isotopes in an object and statically  comparing the ratios of those isotopes.  Each technique measures 33 or 34 different elements and then the lab staff subject the data to multivariate statistics. In the case of INAA for example, we pound small fragments ceramic kiln wasters from each site into a powder, seal them in vials, and drop them into the core of a nuclear reactor (pictured below).  After irradiating them, the samples are put into a detector that counts the unique decay of each isotope in the ceramic fabric.  These counts are very accurate and are measured in parts-per-million (ppm) or parts-per-billion (ppb).

 

When comparing all that data using computer statistical programs, the lab researchers plot each sample onto a graph in multi-dimensional space, one axis for each isotope.  It is easy to imagine the three-dimensional version.  I put a picture below of a bi-plot using Aluminum and Sodium.  This graph appeared in an article my colleagues and I published in Historical Archaeology in 2007.  As the analyst adds data about more isotopes, they add dimensions and the data forms into 'clouds' where samples with similar ratios of isotopes tend to group together.

The analysts produce these multi-dimensional clouds that result in something like a statistical 'fingerprint' for each pottery.  When doing this type of study, one hopes that the clouds don't overlap very much so that some specific isotopes, usually the rare ones, can be used to argue that a sample goes with one group or another.  If things work out well, when one begins to compare unknown samples (such as an unmarked antique pot) to the known samples (from archaeologically excavated kiln wasters), one can make a statistical argument like, "We are 90% confident that Pot X was made at Pottery Y."

We've been fortunate because our data shows very clear patterns, and with the caution that our sample size is small for multivariate statistics, the results are very promising.  We have completed two studies where we compared archaeological fragments from consumption sites in  Utah and Nevada and the results have been very provocative.  As we map these patterns, we can consider a set of sites like the Muddy Mission, near modern-day Las Vegas.  Most of the ceramic fragments we studied matched those we knew were made in Salt Lake City and another set we think had been made in Provo.  But some of the ceramics that Muddy Mission residents used had been made by Thomas Davenport in Parowan, for example.  Other pieces matched J. J. Hansen's pottery from Hyrum, Utah, 500 miles away on the other end of the Mormon Domain.  (This study is forthcoming in a book called Archaeologies of the American West, edited by Margaret Purser and Mark Warner).

We are very excited by these results.  We expect that as we add more samples to the database, we'll be able to start mapping the extent of distribution of each potteries products in time and space.  In addition, we will start trying to sort out the different routes through which people circulated pottery in the economy: direct market exchange from stores, peddling, neighboring, family gift giving, tithing and poverty relief, and even population migration. 

This is a long term goal for our studies.  It will help us to really understand how individuals and families like the Davenport family managed to make their businesses succeed over their lives in Utah.