reconstruction

Reconstructing the requires that we find a soft , which is combined with the candidate. Guided by the studies of the decays $\dzpiz$ and $\dppiz$ in dpi0, we reconstruct using the constraining method [63] which performs a fit that incorporates uncertainties in the positions as well as the energies of each photon shower in the electromagnetic calorimeters.³³ Again, guided by our earlier studies we choose the following cuts:

• IE-IE only
• are assumed to come from the primary vertex
• Minimum number of blocks per IE cluster: 2
• with both clusters of size 2 are rejected
• Mass of within 25 of the nominal mass
• $\chisq(\pizero) < 2.0$
• $E_{\text{cluster}} > 1.0~\gev$
• Energy in the center block, $E_{\text{center}} > 0.5~\gev$
• Fraction of energy in the center block, $f_{\text{center}} > 0.35$

Some of these requirements warrant additional explanation. First, although IE-OE give a visible signal in $\dstar \to \dmeson \pizero$, the mass resolution and signal-to-noise are considerably degraded in comparison to IE-IE . Including these would dramatically increase $\lc \pizero$ background without adding much signal. This precludes their use in reconstructing the . Second, we use a loose ($\pm 25$ ) mass cut since soft have much worse resolution than with higher momentum. (A cut of 15 is typical for hard .) Finally, the ``standard'' reconstruction requires IE clusters with 4 or more blocks. However, many soft don't satisfy this requirement. These departures from the standard reconstruction algorithm are explored more fully in dpi0.

We also arbitrate based on the returned from the constraint routine. If two (or more) have a shared photon, we select the with the lowest . The performance of this arbitration on $\dstar \to \dmeson \pizero$ is shown in dpi0.

Combining the candidates with candidates, we obtain the $\lc \pizero$ mass difference distributions shown in sc_mass:scp_signal.

 

Figure 6.4: Mass difference distributions for . The left histogram is included for illustrative purposes only; wider binning makes the more visible. The lower histograms in each plot are from the sidebands in sc_mass:scp_lc scaled appropriately.

sc_mass:scp_signal