Centrality dependence of low-momentum direct-photon production in $\mathrm{Au}+\mathrm{Au}$ collisions at $\sqrt{s_{\mathit{N}N}}=200 \mathrm{GeV}$

The PHENIX experiment at RHIC has measured the centrality dependence of the direct photon yield from $\mathrm{Au}+\mathrm{Au}$ collisions at $\sqrt{s_{{}_{\mathit{N}N}}}=200 \mathrm{GeV}$ down to $p_T=0.4 \mathrm{GeV}/c$. Photons are detected via photon conversions to $e^{+}{e}^{-}$ pairs and an improved technique is applied that minimizes the systematic uncertainties that usually limit direct photon measurements, in particular at low $p_T$. We find an excess of direct photons above the $N_{\mathrm{coll}}$-scaled yield measured in $p+p$ collisions. This excess yield is well described by an exponential distribution with an inverse slope of about $240\mathrm{MeV}/c$ in the $p_T$ range $0.6–2.0 \mathrm{GeV}/c$. While the shape of the $p_T$ distribution is independent of centrality within the experimental uncertainties, the yield increases rapidly with increasing centrality, scaling approximately with $N_{\mathrm{part}}^{\alpha }$, where $\alpha =1.38\pm 0.03\left(\mathrm{stat}\right)\pm 0.07\left(\mathrm{syst}\right)$.

Figure 1

Histograms of the $e^{+}{e}^{-}$ pair invariant-mass distribution from data. (a) The distribution of masses calculated with the normal reconstruction algorithm (vtx). (b) The distribution of masses calculated with the alternate track model assumption (HBD).

Figure 2

A view of the cut space used for the conversion photon identification. The mass as calculated under the standard reconstruction algorithm (vtx) is plotted on the horizontal axis, while the mass as calculated under the alternate track model (HBD) is plotted on the vertical axis. The dotted (red) box indicates the region used to identify photon conversions.

Figure 3

Histograms of the $e^{+}{e}^{-}\gamma$ invariant-mass distributions for two different $p_T^{ee}$ bins. The left column (a) and (c) displays the mass for $0.4<p_T^{ee}<0.6 \mathrm{GeV}/c$; the right column (b) and (d) displays the mass for $1.8<p_T^{ee}<2.0\mathrm{GeV}/c$. The top row (a) and (b) shows the $e^{+}{e}^{-}\gamma$ invariant-mass foreground distribution in blue, with the normalized background distribution from the mixed events in red. The bottom row (c) and (d) shows the isolated pion peak after subtraction of the normalized background. The masses are calculated from the HBD readout plane origin assumption on the electron tracks. The centrality bin is 0%–20%.

Figure 4

Average conditional acceptance $\langle {\varepsilon }_{\gamma }f\rangle$ to detect a photon from a ${\pi }^0$ decay in the EMCal, if the other photon converted in the HBD electronics and was reconstructed as an $e^{+}{e}^{-}$ pair. The abscissa gives $p_T^{ee}$, the $p_T$ of the $e^{+}{e}^{-}$ pair. The $p_T$ cut of 0.6 (2007) and $0.4 \mathrm{GeV}/c$ (2004) is on the photon detected in the EMCal. For the $p_T$ cut of $0.6 \mathrm{GeV}/c$ we show the results for two methods, a full MC simulation (points) and a fast MC simulation (histogram). For the $p_T$ cut of 0.4 GeV, the fast MC simulation is shown as a dashed histogram.

Figure 5

(a) Ratio $R_{\gamma }$ as a function of photon $p_T$ from the 2007 (red open square) and from the 2010 data sets (blue closed circle) in minimum-bias $\mathrm{Au}+\mathrm{Au}$ collisions. Statistical uncertainties are dominated by the ${\pi }^0$ yield extraction. They are plotted as vertical lines. All other systematic uncertainties are added in quadrature and shown as filled boxes. (b) $R_{\gamma }$ in the combined 2007+2010 measurement.

Figure 6

Ratio $R_{\gamma }$ as a function of photon $p_T$ for the combined 2007 and 2010 data sets in centrality bins 0%–20%, 20%–40%, 40%–60%, and 60%–92%. Statistical uncertainties plotted as vertical lines are dominated by the ${\pi }^0$ yield extraction. All other systematic uncertainties are added in quadrature and shown as filled boxes. (a) and (b) We also show earlier results from Ref. [2], obtained by extrapolating virtual photons to zero mass.

Figure 7

Direct photon $p_T$ spectra for minimum-bias $\mathrm{Au}+\mathrm{Au}$ collisions from this measurement (solid symbols) and $\mathrm{Au}+\mathrm{Au}$ and $p+p$ collisions (open symbols). (Open circles and up-triangles) Low $p_T$ spectrum obtained with virtual photons in $p+p$ and $\mathrm{Au}+\mathrm{Au}$ [2]; (open squares and down-triangles) spectrum of real photons, measured in the EMCal in $p+p$. Open squares are 2003 data [35]; open down-triangles are 2006 data [36]. (Open stars) Spectrum with real photons, measured in the EMCal in $\mathrm{Au}+\mathrm{Au}$ in 2004 [37]. The dashed line is a fit to the combined set of $p+p$ data, extrapolated below $1 \mathrm{GeV}/c$, and the solid line the $p+p$ fit scaled with the number of minimum-bias $\mathrm{Au}+\mathrm{Au}$ collisions. Bands around lines denote $1\sigma$ uncertainty intervals in the parametrizations of the $p+p$ data and the uncertainty in $N_{\mathrm{coll}}$, added in quadrature.

Figure 8

Direct photon $p_T$ spectra in centrality bins 0%–20%, 20%–40%, 40%–60%, and 60%–92%. Widths of filled boxes indicate bin widths in this analysis. The green bands show a $N_{\mathrm{coll}}$-scaled modified power-law fit to the PHENIX $p+p$ data and its extrapolation below $1 \mathrm{GeV}/c$; cf. Fig. 7. One-sided errors denote $1\sigma$ upper limits; other uncertainties are as in Fig. 7.

Figure 9

Direct photon $p_T$ spectra after subtraction of the $N_{\mathrm{coll}}$ scaled $p+p$ contribution in centrality bins 0%–20%, 20%–40%, 40%–60%, and 60%–92%. Uncertainties are plotted as in Fig. 8. Dashed lines are fits to an exponential function in the range $0.6\mathrm{GeV}/c<p_T<2.0\mathrm{GeV}/c$.

Figure 10

Integrated thermal photon yields as a function of $N_{\mathrm{part}}$ for different lower $p_T$ integration limits. The dashed lines are independent fits to a power law.