[Halld-physics] eta etaprime comments

[zihlmann]

Since I did not know where to send the comments about the eta-etaprime

analysis report I will miss use this list. see attachment.


-------------- next part --------------
Comments on Sigma eta and eta-prime
----------------------------------------

line 41: missing energy cut -1.6GeV < ME < 1.8GeV, sounds very large. How
         does the spectrum look like. the missing mass cut is much narrower.
     
         Nominal analysis Cuts:
         There is no mention of UnUsedEnergy. Is there a cut on that quantity?
         If not how does it look like? Depending on the event topology and
	 type of the final state particles this may be quite differently
	 looking.


line 51: This last sentence is kind of confusing. The events are considered
         to be exclusive, so there should always be a proton and a pi+ a pi-
	 and a pi0. Maybe the way you call it is different, like two positive
	 charged tracks, one negative charged track and two neutral particles,
	 but it requires to be "exclusive", meaning the kinematic fitter is
	 fitting for an exclusive configuration. Here again comes the question
	 about the unused energy cut.

figure 2.1: the cut line (red) is only used as an assignment of the positive
            charged track to be a proton candidate, right?

line 56: vertex cut 50-78, I assume 50 is a type? should be 52 as shown in
         fig. 2.2?

Unused energy cut? this is important for all neutral final states, apart
                   from the recoil proton. p+3pi0 and p+2g final states on
		   the eta.

From the text it is not completely clear if the pi0 and the eta masses
are constrained in the fitter or not. And if figure 2.3 shows the measured
photon quantities?

3.3 Phase shift: The phase is determined from a fit to the phi asymmetry
                 in the omega mass region and then fixed for the asymmetry
		 fits of the actual data on eta and etaprime.
                 0) is this a fair thing to do? text needs some good arguments.
                 1) are the values compatible with the rho analysis for example?
                 2) the difference between 3.96deg and 138.74deg is defined
		    by (3.96+180.)-45 = 138.96 which is close, by 0.22deg. Is
		    this within the statistical error of the fit parameters?

Side band correction: This correction only works if the asymmetry in the
side band is the same as under the peak. Maybe for the omega peak there
is enough statistics to determined the asymmetry in several mass bins to
try to support this assumption.

3.4.2 and eq. 1 : the definition of the asymmetry as (PERP-PARA)/SUM
                  should have a reference. That will help to fix the sign
		  convention.

caption 3.8 do not condemn the photon beam polarization into a foot note!


fig. 3.9 
and following
plots:        1) the SB asymmetries in the second row look very skewed in
                 the individual plots as function of phi, and only look ok
		 in the asymmetry, why?
	      2) for the SB the summed average is not flat at all.
	      3) the summed t-average is flat, how flat? does a flat horizontal
	         line fit do a good job as a fit?
	      4) similar questions my hold for the individual t-bin data.
	 

3pi0 final state. again the unused energy cut is a question? Also there may
be several combinations that make a 3pi0 final state with 6 photons. What
is done with regards to uniqueness for an event?

-t distribution: is there a significant difference between the bin centers
                 and the mean -t value in the bins?


figure 3.19 is confusing. It is stated that this channel is very clean and
has basically no background. However there is a background (green) shown
with an f value of 5.7%. Why is this background so large and
why it is the functional form concave? There are no side bands so
how is a background  determined?

The phase is fixed and the same as in the pi+pi-pi0 case. 3pi0 might have
quite a different event topology an detector acceptance. This could have
and effect on the phase. Are there any studies done with regards to this?

2gamma final state: again all neutral, so this may be sensitive to the
unused energy cut.

The asymmetry under the eta peak is assumed to be the same as the omega.
since there is a gamma missing from the omega this event topology may
alter the fixed phase assumed/used in the fits. This may require some
further tests to minimize any bias.

The corrections in terms of asymmetry are quite large and with some
significant variations with -t. In particular the 3rd t-bin, the
background has very large asymmetry(opposite sign). It may not be fair
but the error bars after the background subtraction do not look any
bigger than before subtraction.

before going to the eta-prime results here would be a good place to
compare the results with what has been published by GlueX.


eta-prime to pi+pi-eta:
a) which eta decay channel is used? all the previous?
b) what is the cut on the inv. mass of the eta?
c) was the eta mass constrained in the kin. fit?
d) the eta to pi+pi-pi0 decay channel has also a pi+ and a pi-
   is there a significant double counting possible?
   what about event uniqueness? 

e) is there sufficient statistic for the pi+pi-eta system to have
   separate asymmetries for the 3 different eta decay modes ?

line 259: what are these potential deviations from the standard model?

last paragraph on page44 with regards to the limited statistics currently
available: What may also be important is that this ration is of order
1%. Note that in some of the eta decay modes the background corrections
are of the same order or even significantly larger. This may pose an inherent
problem to control even with much more statistics.

systematic test plots 4.2, 4.3, 4.5 ...: It would be useful to present this
data as tables with numbers for comparison since the plots alone with a rather
large vertical scale does not allow any judgment.


Systematic studies on the cut dependence are done separately for the 0/90 and
45/135 data sets. What does a comparison between these two data sets on
the effect of different cut changes look like. Are they similar, the same,
or completely different. Some numerical tables may be useful for comparison.

fig. 4.16 to 4.19: The uncertainties are coming from the fit results. These
are much larger for the individual t-bins than for the integrated result
indicated with the point at t=0. This is not simple statistics anymore. If
one would average the four values of the t bins together using the error bars
as weight what would the resulting value and error look like?

line 355: Why can the phase offset not depend on the decay channel? The detector
acceptances may be different and could potentially introduce an offset that
may be different.