Today's Progress 15. June. 2006
Study of (K^-, X^0) spectrum with T0-NT (1)
Study of the (stoppedK^-,gamma) TOF spectrum with Kmu2-determined correction function
First, we examine the availability of Kmu2-determined correction functions. For gamma ray,
NT-detected-time, Tnt, is defined by
Tnt = (Tu+Td)/2 -f(sqrt(phu*phd))*g(xi) - h(eta_0)+offset
,where eta_0 is now z-center of the NT segment, and xi is determined by Tu-Td.
All functions, offset are kept as those obtained from K^+, at this moment.
The time-residual is then defined by
delta T = Tnt - Tt0 - TOFt0->stop -Ltof/c
where Tt0 is kept as that obtained by T0-PA analysis for (stopK^-,X) (run129-134).
The 1/beta resolution plot for gamma-like events. The NT slewing correctiorn functions/offsets are just kept
as those obtained by Kmu2 decay evets of stopped K^+.
The comparison of the time-residuals after Kmu2-determined slewing correction between Kmu2-originated muon and gamma-ray for NCsegment 1(L-row1-layer1).
The comparison of the correlations of sqrt(phu*phd) and time-residual after Kmu2-determined slewing correction between Kmu2-originated muon and gamma-ray for NCsegment 1(L-row1-layer1).
The comparison of the profiles of correlations of sqrt(phu*phd) and time-residual after Kmu2-determined slewing correction between Kmu2-originated muon(black) and gamma-ray(red) for NCsegment 1(L-row1-layer1).
Depending on the segment, substantial "residual correlation" and the deviation of peak position exist for gamma-ray events. After removing that, we obtain substantial improvement of 1/beta distribution for "gamma-ray", as presented below.
The arm-by-arm dependence of the gaussian sigma for gamma-like events on the light output. Left and right pannels are for L/R arms, respectively, and black/red lines are from method 2/3 described in the text, respectively.
The arm-by-arm dependence of the gaussian center for gamma-like events on the light output. The conventions for L/R:color are as for the previous figure.
We, however, need to be more carefull - because the best correction function for "gamma-ray" is not necessarily the best correction function for "neutron", as we have already experienced in the comparison between "muon" and "gamma-ray". In order to study the most relevant correction method, we now examine the following three ways, namely,
just keep the correction (method 1),
invoke the additional offset to tune the residual center to be 0 (method 2), and
invoke the additional correction function to eliminate the "residual correlation" (method 3),
for following two event sets,
gamma,
neutron from stop Sigma^+.
For gamma, expected ordering of peak width is of course 0>1>2, and scale of peak center deviation is 0>1=2 at least when we do not classify the 1/beta by light output. However, the performance for gamma does not ensure the performace for neutron, hence all 3 are applied for monochromatic neutron events, and compared, although, the available energy region is very limited. Note that the stopped Sigma^+ events are picked up from Kst-charged-triggered event set to enlarge the available statistics by factor 5 over. Hence, vertex position used is BLC-PDC one.
The comparison of 3 correction functions (method1->black, method2->red, method3->green) for neutral particles. 1/beta region of 1(gamma-like:top panel)/5.2(neutron from stopped Sigma^+:bottom panel) for L(top) and R(bottom) arms are presented, respectively. Event selections are applied to the charged particle detected on the opposite arm - if the charged particle detected on the opposite arm is proton, the neutral particles detected on the original arm are dropped from the plot.
The improvement by method 3 is clear for gamma, while we cannot discriminate superior method for low-energy neutron. It should be noted that the original gamma-ray event set comes from Kst-VTC-NT-triggered events, hence ~80 % of gamma events are different from those used to tune the offset/correction function (at least, the vertex definition is different even for the events which are included in both of Kst-charged-triggered and Kst-VTC-NT sets.) After run-by-run Stop K selection has been established, we will come back to the issue, because in-flight contamination makes 1/beta distribution broader, which makes the discrimination of the methods more difficult.