Re: [asa] Denver RATE Conference (Thousands...Not Billions)_Part 2

From: Steven M Smith <>
Date: Wed Oct 17 2007 - 12:12:41 EDT

Kirk Bertsche wrote on 10/15/2007 11:32:28 PM:
> ...
> I recently posted some general comments about radiocarbon and
> contamination. Since then I have gone through the written RATE
> radiocarbon claims in detail, read most of the references, and have
> spoken with the expert who measured the radiocarbon samples for
> them.
> ...

This information is excellent. Thank you, Kirk!

I also have obtained Baumgardner's list of 90 "AMS measurements on samples
conventionally deemed older than 100,000 years" (Table 1) and have been
checking the cited references. I have made it through about half of the
articles. Unlike Kirk, I do not have a background in radiocarbon research
and am on a steep learning curve. But I thought I would share a few
things I have learned so far with the list.

1) Early expectations of the Accelerator Mass Spectrometry (AMS)
technology were that it might be able to measure Carbon-14 (C-14) dates up
to 100,000 years old and should allow measurements of much tinier samples
than the conventional beta counting method. The reality is best summed up
by the title of a journal article in _Nuclear Instruments and Methods in
Physics Research B_ (NIM-B), "Early expectations of AMS: Greater ages and
tiny fractions. One failure? - One success" (1987, v.29, p. 97-99). The
AMS may obtain marginally older dates than the conventional beta counting
method but its biggest success is in the ability to process samples much
smaller than 500 mg. During the conference, Baumgardner made a big deal
out of the fact that AMS technology could potentially obtain C-14 dates
back to 100,000 years (100 ka) and suggested that this "failure" was due
to intrinsic C-14 in every bit of fossil carbon. The article cited above
states, "The age limit depends on true or apparent ion counts, which may
originate in sample preparation, in the accelerating system itself, or
from ions other than 14C which are not adequately suppressed."

2) A large number of the reports of C-14 in Carbon "dead" samples that
Baumgardner cites come from articles that describe the methodology &
instrumental set-up of new AMS facilities. Note some of the following
titles: "Accelerator mass spectrometry of 14C at Nagoya University",
"Accelerator radiocarbon dating at SFU", "AMS 14C measurements and
preparative techniques at NIES-TERRA", "AMS at the TTT-3 tandem
accelerator in Naples", "AMS sample handling in Groningen", "An assessment
of laboratory contamination at the IsoTrace radiocarbon facility", etc.
(That is just the articles with titles that start with the letter A.) As
I read these articles, my weird sense of humor kicks in and I am reminded
of a bunch of kids at school comparing their cars or stereos. "My tandem
AMS can get a date as old as 85 ka on a blank aluminum sample holder with
the ion gate closed." "Yeah, well my Tandetron AMS get >90 ka!" These
are articles about experiments that are testing the absolute limits of
individual laboratory facilities, not experiments designed to measure
"intrinsic" or "in-situ" C-14. For some reason, Baumgardner did not list
C-14 dates of 85,000 years for blank aluminum sample holders in his Table

3) As Kirk points out, many of Baumgardner's cited samples also come from
studies of sources of laboratory C-14 contamination in the sample
preparation or instrumental analyses. The two samples in Baumgardner's
list that Kirk noted are prime examples that Baumgardner is ignoring this
fact. Each sample was listed twice: First with a C-14 measurement on a
"untouched" fraction, and a second measurement made after "graphitization"
to measure the component of introduced contamination. The article I
mentioned above under note 1 lists "Geological graphite -- powdered and
encapsulated under argon: 69,030 +/- 1,700. Same sample -- except
prepared in air: 63,800 +/- 1,100. An article by D. L. Kirner, R. E.
Taylor and J. R. Southon (Reduction in backgrounds of microsamples for AMS
14C dating, Radiocarbon, 37, 697-704, 1995) summarizes a lot of work on
sources and levels of C-14 contamination in AMS work. I reproduce their
Table 1 below that gives 4 different general sources of background C-14 --
each with 2 or 3 types of possible contamination (9 types of contamination
total). Baumgardner ignored 8 of these sources of C-14 contamination,
implying that his measurements were all due to 1 category in this table --
"Non-14C-"dead" sample: sample erroneously assumed to contain no 14C."

TABLE 1. Potential Sources of Background in AMS 14C
                 Measurements Using Catalytically Condensed
                 Graphitic Carbon Targets

A. Machine background: 14C detected when graphite in the
     sample holder does not contain 14C.
    1. Detector anomaly: 14C pulse registered in detector
         circuitry when no 14C ion is present.
    2. Ion identification anomaly: particle of same 14C
         mass/energy ratio as 14C but which is not 14C reaches
    3. Beam-line contamination: 14C internally derived from
         an accelerator beam line component reaches detector.

B. Combustion/acidification background: 14C introduced
     during production of CO2 from sample.
    1. Materials contamination: 14C introduced from materials
         in combustion/acidification tube.
    2. Tube contamination: 14C introduced from combustion

C. Graphitization background: 14C introduced during the
     graphitization process.
    1. Materials contamination: 14C introduced from materials
          in reaction tube.
    2. Tube contamination: 14C introduced from reaction tube.

D. Pseudo 14C-"dead" sample background: 14C is present
     in a sample material that should not contain 14C because
     of its geologic age.
    1. Non-14C-"dead" sample: sample erroneously assumed
          to contain no 14C.
    2. Contaminated 14C-"dead" sample: 14C introduced into
          material that contains no 14C.

4. Radiocarbon measurements are different than other radioisotopic dating
methods. Most methods (K-Ar for example) measure the amount of
radioactive parent material (K-40) and the amount of resultant daughter
isotope (Ar-40) left in the sample. The ratio of K-40/Ar-40 is used to
determine the age of the sample. Hence a little bit of Ar-40
contamination (the most common contamination problem in the method)
primarily affects the ages of very young samples. Thus K-Ar dating is
seldom used, or used with greatest caution, on samples less than 1 million
years old (i.e. 5,000 years +/- 1 million years is a problem; 246 million
years +/- 1 million years is not). However, radiocarbon instruments do
not measure the parent (C-14) to daughter (N-14) ratio, they measure total
C-14 present and compare it to modern carbon (either C-14/C or % modern
carbon: pMC); therefore a small amount of C-14 contamination becomes
critical only for the oldest samples (i.e. 49.25 +/- 0.01 pMC [~5,000
years] is not much of a problem; 0.01 +/- 0.01 pMC [~70,000 years] is a

It is interesting to me that the more I read the actual literature
Baumgardner cites, the less concern I have with his small amounts of C-14
in carbon "dead" materials issue.

(Disclaimer: Opinions expressed herein are my own and are not to be
attributed to my employer ... or anyone else.)
 Steven M. Smith, Geologist, U.S. Geological Survey
 Box 25046, M.S. 973, DFC, Denver, CO 80225
 Office: (303)236-1192, Fax: (303)236-3200
 -USGS Nat'l Geochem. Database NURE HSSR Web Site-

To unsubscribe, send a message to with
"unsubscribe asa" (no quotes) as the body of the message.
Received on Wed Oct 17 12:14:05 2007

This archive was generated by hypermail 2.1.8 : Wed Oct 17 2007 - 12:14:05 EDT