On 11/28/06, Peter Kammel <
kammel@npl.uiuc.edu> wrote:
Dear Sacha,
Thanks for the careful drafting of this plan. Sorry for my belated
reply, too much going on...
I have a few small comments/questions:
i) The assumption that A is independent of P_D in your paper could
be compromised if you have an additional source (small leak or
trap on separator material), which produces deuterium at a small
level. This would not directly show up in the calibration which
is performed at higher concentration.
But I believe such a term would lead to a higher observed P_Df
in your procedure, meaning that the extracted P_Df is an upper
limit.
If you will estimate the rest deuterium on the surface after good pumping you will see that the deuterium quantity is negligible. It is the fixed volume of deuterium which is much smaller then in the initial gas. Linear approximation has its own error, but it is connected with separator itself. Oleg promised to discuss with us this question.
ii) Concerning your figure, it was not clear to me how uniform the
pressure/temperature is in the reboiler. When you say 0.499 mbar
at the middle plane, does that mean that is had different
pressures below and above midplane ?
It seems to me that you could also define a {A V_R}_i from
your measurements to check how constant the V_R really is.
0.499 mbar is the differential pressure between lower part of the reboiler and the upper part. This is direct hydrostatic pressure and the height of the liquid hydrogen. Instability of this pressure is instability of the level and the quantity of the hydrogen (in the liquid form) in the reboiler. I used average value of the differential pressure (from the experiment) and Gauss half width for this pressure (experimental value too).
iii) Your final error evaluation is conservative, a quadratic
addition would lead to a smaller error.
All my errors are systematic (not statistical). In this case I HAVE TO have simple sum (not quatratc).
iv) A significant uncertainty in your error estimation comes
from the AMS measurement. We are all extremely impressed by
the 0.2 ppm uncertainty quoted by the ETH group, but hope
to get more details how they achieve this, in particular
regarding the sample preparation.
This is the main error. Perhaps possible to minimize it. But it does not depend from us.
v) Finally, as an aside, I was wondering whether we can check the system
by analyzing for tritium. I remember that AMS is extremely sensitive
to tritium. I am sure Oleg knows whether there is enough tritium
in hydrogen to be sensitive?
This is question to Oleg. But when we clean water from heavy water we kill TO2 at all! To my mind we do not have sensible quantity of tritium and impossible to make QUANTITIVE measurements.
Apart from this technical remarks, let me emphasize that the deuterium
calibration is of key importance.
- It will eliminate an important systematics in the final MuCap result.
- Trace deuterium detection (with pdm fusion) was a main item in our
CRDF proposal. I am perfectly willing to give up my pdm idea, if
we can provide <0.1 ppm deuterium diagnostics with your proposed
method.
- It is a beautiful method, pushing the current technological limits.
So we should definitely perform this task.
With best regards
Peter