Subject: carbon photo absorption coefficient
From: Martin Zombeck 
Date: Fri, 25 Oct 1996 10:47:11 -0400
Sender: mvz@sisko.harvard.edu

During the calibration of the HRC-I subassembly we uncovered an
error in the entry for the mass absorption coefficient
for carbon at carbon k alpha in the Henke, Gullikson, and Davis 1993
tables (Atomic Data and Nuclear Data Tables v. 54, 181-342 (1993).
This error is significant for anyone modeling the transmission of
organic thin films or using organic gases (methane, propane, etc.)
in proportional counters in order to make absolute qe measurements.

What follows is e-mail correspondence between me and Eric Gullikson
discussing our result.

To: EMGullikson@lbl.gov
Subject: carbon mass attenuation coefficients
Date: Tue, 15 Oct 1996 09:55:32 -0400
From: Martin Zombeck 

We have been attempting to derive the effective path length of a gas
proportional counter by the following method:

        1. measure the count rate as a function of fill pressure
using several incident energies and several fill gases.

	2. monitor possible variations in source flux with a
 separate gas filled proportional counter

	3. calculate the gas density using the fill pressure, the
measured temperature of the gas, and the gas composition

	4. calculate the mass absorption of the fill gas using
the gas composition and the mass absorption coefficients of the
elemental constituents (from Henke, et al., 1993)

	5. Derive the thickness t from a best fit to the data using
exponential absorption as a model.

We control the gas pressure in both counters to within 1 Torr,
monitor the temperature, and read the pressure with both Hg
manometers and calibrated pressure transducers. We are also using
certified fill gases.

Here's what we obtain:

     Mn K-alpha (Fe55 source), P-10:               6.28 cm

     Al K-alpha (well filtered e-impact), P-10     6.22 cm	

     C K-alpha (well filtered e-impact), Methane   7.96 cm !!!

Physical measurement of offset counter window to back of
counter (body diameter is 2") is 6.35 cm.

      We will make a further measurement at Al K-alpha with Methane.


Is it possible that (u/p) for carbon at carbon K is in error?

Your 1993 (and website) value is 1.96 x 10^3 cm^2 g^-1.

Your 1982 value is 2.35 x 10^3 cm^2 g^-1 (20% higher than 1993)

If we use our derived path length to derive a value for (u/p) for
carbon, we obtain

             2.54 x 10^3 , within 8% of your 1982 value.

We have made the typical idiot checks: two idiots independently doing the
calculations. We have also cross-checked our calculations with your
website and get agreement when we use the same set of coefficients,
path lengths, gases, energies, pressures, and temperatures.

What's up?

Thanks.


Martin Zombeck
------------------------------------------------------------------------

Date: Sat, 19 Oct 96 18:30:51 PDT
From: egull@grace.lbl.gov (Eric Gullikson )
To: mvz@sisko.harvard.edu
Subject: carbon mass attenuation coefficients

Dear Martin,

Thank you for your email last week.  It sounds like you guys have done some
very careful measurements.  I have checked the published data that was used
in generating the last set of tables (1993) and found the following:

Energy, mu, reference
 277. 2106. 6
 277. 2206. 11
 277. 1720. 18
 277. 2300. 87
 277. 2808. 120
 277. 2535. 128
 277. 2051. 143
 277. 1930. 144
 277. 3700. 234

The last value is derived from relectance measurements and it is not too
surprising that there are large errors associated with it.  If you just
take the average of the remaining eight values you get 2207 cm^2/gm.  However,
you can see the spread in the measurements and your value of 2540 is certainly
within reason.  In fact, reference 128 which gives a value quite close to
yours appears to be one of the more thorough works.  As for our
value, it does appear to be too low compared to the average measured  
value.  In part
this may be due to an error in our transcribing the measurement from
reference 144.

Thank you very much for letting me know about this.  I hope that when you're
done you will consider publishing your results so that we can refer to them in
the future.

Best Regards,
Eric

 *************************************************************************
 Eric M. Gullikson
 Center for X-Ray Optics, 2-400		 Phone: (510) 486-6646
 Lawrence Berkeley Laboratory		 FAX:   (510) 486-4550
 1 Cyclotron Road			 email: EMGullikson@lbl.gov
 Berkeley, California 94720
 ********************************************************************

References

Henke, B. L., Lee, P., Tanaka, T. J., Shimbukuro, R. L. and Fujikawa, B. K. 1982, "Low energy X-ray interaction coefficients: photoabsorption, scattering and reflection", Atomic Data and Nuclear Data Tables, 27, 1-144.

Henke, B. L., Gullikson, E. M. and Davis, J. C. 1993, "X-ray interactions: photoabsorption, scattering, transmission, and reflection at E=50-30000 eV, Z=1-92", Atomic Data and Nuclear Data Tables, 54, 181-342.