Subject: Re: draft of the Combined Monopole Analysis paper: read/correct it
From: Erik Katsavounidis (Erik.Katsavounidis@lngs.infn.it)
Date: Fri Sep 08 2000 - 09:16:18 EDT
Hi Laura, Ivan, Fausto,
I only now got a chance to go through the proposed paper on "a combined
analysis technique for the search for fast magnetic monopoles with the
MACRO detector", so for whatever they worth here are some comments,
although past due.
(1) section 3.2 (p.4/5): I'm afraid there has been a misunderstanding with
what regards to the scintillator calibrations presented in this section.
It is quoted that:
"Here we will concentrate ... the raw *attenuated* ADC data
(i.e. those which measure the PMT pulse after an attenuation of a factor10).
The calibration of the response of each single ADC is performed by means of
a nitrogen laser ..."
As far as I know, the publicly released scintillator *energy* calibrations
make NO USE OF THE LASER DATA. Unless you have analyzed the laser data on
your own, all energy calibrations for data following the 6-month RUN
calibrate the NON-ATTENUATED ADC using MUON DATA ONLY. The only LASER-based
calibration of the ATTENUATED ADC that I know of has been a special LASER
RUN we took in June 1993 and Rong-Zhi Liu analyzed for his thesis. This
work of his DID involve LASER-based calibration of the ATTENUATED ERP ADC.
However, this work never made it to the calibration database and as far as
I know never became public (maybe you got a hold of RongZhi's constants?)
It was though included in the 1997 MACRO published paper on searches for
fast monopoles where the ERP-based result was also reported.
The calibration procedure though described there is NOT the one used in
today's ERP data. The present ERP calibrations calibrate the NON ATTENUATED
ADC with cosmic ray muons and then fix the linear regime of the attenuated
ADC through requesting consistency with the non-attenuated one.
This makes obvious that the statement at the end of the third paragraph of
page 5 regarding the limit of ERP based energy reconstruction (few GeV) and
the accompaning figure 4 (from the 1997 fast monopole MACRO paper) do not
apply if you have used the standard distribution of ERP calibration
constants.
(2) section 3.2 (p.5), item #1 of the enumerated section: a cut on the TDC
high/low consistency within 3ns is there described. I would like to point
out that this might be dangerous a priory. Mostly because the timewalks
that enter in this time reconstruction may be highly unreliable and the
fact that the dynamic range of the signals you expect in this search is
enormous. This cut is used in the neutrino analysis being a ToF measurement
and cuts radioactivity background, which is not the case in the fast
monopole search. I am sure if you relax this cut completely, it will have
no effect in the remaining events.
Within the same scintillator analysis path, I was wondering if there are
other ERP detail cuts like number of ERP boxes etc that are not listed
here (I somehow recall in a previous memo by Ivan on his original analysis
a cut like that.)
(3) section 3.2 (end of p.5 beginning of p.6): a connection is made here between
the average PMT pulse width (50ns) to the fact that saturation effects will
start for pulses from particles traveling in a counter for more than 135ns.
This is not obvious to me-- would you like to elaborate? The instrinsic PMT
response is rather irrelevant for pulses lasting more than this.
(4) section 3.2 (top of p.6),derivation of the 470MeV energy deposition of a
5x10^-3 beta monopole. If x30Imin is the rate of energy deposition then for
20cm I come up with something of the order of 1GeV-- am I missing something
here (energy loss vs. scintillation light- I believe we are always talking
about scintillation light here)?
(5) section 4 (p.9/10) describes the analysis scheme: could you please
provide --if easy-- the effect of each cut in the number of remaining
events? I think it is useful for the reader to appreciate what each
cut removes at each step.
(6) section 5.3 (p.11) derivation of minimum E through eqn 4: a 1.8 MeV/(gr/cm2)
value is used for Imin which is then multiplied by the scintillator density.
I guess this is another mystery number for MACRO. In the ERP energy
calibration code the muon peak is fixed at 1.8 MeV/cm (notice the
difference in units with what you are using). Now, being this a rather
critical issue in our analyses, over the last few months, Charlie Peck at
Caltech has performed a detailed derivation of the energy loss of CR muons
in MACRO using as realistic counter folding as possible based on all-time
ERP calibrations. The result will appear in the ERP energy calibration
section of the technical papers (1.8 MeV/cm is your number though.)
(7) section 5.2 (p. 11) The ERP efficiency is defined here as:
"... the fraction of (ERP) boxes with e(box)>=95% ..." where e(box) is
the fraction of tracks hitting a given ERP box over the ones expected.
An accompanying plot (fig 10, p.24) shows these numbers for a sample
RUN/SM. In the figure caption of this plot, it is stated that for this
sample RUN, 3 counters do not fulfill the 95% efficiency requirement
while if I read the plot correctly I see only 2 (?).
Going back to the definition of the global ERP efficiency, I am a bit
concerned about being overly conservative. If I go to figure 11, p.25
I see this global ERP efficiency staying for most of the time at around
92%, which is low IMHO. This is roughly the *single* box ERP efficiency
to single tracks. Your analysis requires at least one in order to
process an event. If we assume that all tracks intercept at least 2 or
3 ERP boxes or which only one is needed in order to be analyzed, the
global ERP efficiency for this analysis that I would expect is O(99%).
Of course, not that it really matters when setting upper flux limits-
you have every right to be overly conservative.
(8) section 5.3 (end of p.11): "an energy deposit of 235 MeV corresponds to
less than 1000 counts of the ERP attenuated ADC"-- this is clearly
position along the counter dependent... Out of curiosity I took ~30 days
of MACRO running and asked for all clean single muons passing within
1 meter of a PMT that resulted in an attenuated ERP ADC value greater than
1000 counts: this was around 30 (I'm counting histogram entries directly
on a plot).
(9) section 6 (end of p.13 beginning of p.14). It is stated that:
"As a further check the measured values of dL/dx and beta of surviving
candidates have been compared with the expected monopole light yield as
a function of its velocity. For all the events dL/dx is well below
expectations"
I think this is a very important statement and it probably merits inclusion
of the corresponding plot showing this in the paper.
(10)section 6.1 (p.14) presents the final flux upper limit. It is there
stated that "by taking into account all the cuts required by the
analysis, the acceptance turns out to be ~3565m2sr". I am a bit confused
on how to interpret this number. Does it refer to a specific detector
configuration at a specific point in time? The running from 1992 to 1996 had
all sorts of different detector configurations and any sort of
acceptance average or integration wouldn't make much sense to me.
Morever, I was wondering if your acceptance calculation has folded
the CR39 acceptance too. Given that CR39 is part of the analysis and
the final event selectino/rejection one may argue that it should have
it folded (although I understand that this may end up to a philosophical
discussion.) At any rate, my toy acceptance for full MACRO when I insert
the CR39 acceptance yields a smaller number for the one you quoted above,
but then again I do not know how to interpret your number above.
(11)figure 3 (p.17) shows "Energy loss distribution for a sample of single
muons, as reconstructed followed the method discussed in the text"
Well, this needs to be clarified a little bit. I think we still have
to decide if the correct x-axis is MeV/(gr/cm2) or MeV/cm. You opted
to plot vs. MeV/(gr/cm2) and quote a 1.84 as the mean which I think
contradicts with the standard calibration procedure. At any rate, I think
more than average and rms shown on a plot like that, the landau fit+
parameters shown on it would be very instructive to the reader.
Talk to you later,
--Erik
This archive was generated by hypermail 2a24 : Fri Sep 08 2000 - 09:16:24 EDT