I've checked the purity of the gamma-rich trigger in run 9 pp as a function of eTtrg for "exclusive" gamma-rich triggers and "inclusive" gamma-rich triggers. The "exclusive" and "inclusive" monikers indicate:

• exclusive gamma-rich: TSP = (0.2, 0.6)
• inclusive gamma-rich: TSP = anything (though I set an upper limit of 100 in practice)

The trigger eT range was restricted to 9 - 20 GeV, and the tracks used in the calculation were required to have pTtrk = (1, 20) GeV/c and zTtrk = pTtrk / eTtrg = (0.3, 0.6). For convenience, I've plotted R rather than 1 - R below:

Where the solid lines indicate the statistical error on the value of R averaged over all eTtrg. Here 'D' indicates the per-trigger yield of Near-Side (NS) charged hadrons after subtracting an event-wise pedestal. The pedestal is determined by fitting 2 gaussians and a constant to the delta-phi correlations (the constant term being the pedestal). The near-side here is defined to be delta-phi = +- 0.63 rad. The delta-phi and zTtrk correlations for eTtrg = 9 - 20 GeV are presented below. The correlations for the individual bins are attached.

Note that in the eTtrg = (17, 20) bin, the value of R is greater than 1 (~2 for both exclusive and inclusive). This is due to low statistics in this bin, and these values might be better determined by using a less conservative hot tower list.

Note: I forgot to set the y-axis of the delta-phi correlations to a fixed range, so the two y-axes on the plot are not the same.

Update [02.01.2018]: I've attached the hot tower list used in this calculation.