1. Kin Yip For STAR Collaboration Brookhaven National Laboratory Physics with Tagged Forward Protons at RHIC May 30 - June 4, 2013, Rehovot/Eilat, Israel Mainly : Elastic pp scattering program at RHIC Results (A N, r 5, & A NN /A SS ) Central Exclusive Production STAR

2. RHIC-Spin Accelerator Complex June 2, 2013Kin Yip 2 STAR PHENIX AGS LINAC BOOSTER Pol. Proton Source Spin Rotators 15% Snake Siberian Snakes 200 MeV polarimeter AGS quasi-elastic polarimeter RF Dipoles RHIC pC “CNI” polarimeters RHIC absolute pH polarimeter Siberian Snakes AGS pC “CNI” polarimeter 5% Snake  * ~ 21 m for pp2pp/STAR in 2009 Former location of pp2pp

3. June 2, 2013 Kin Yip 3 Physics with Tagged Forward Protons p + p  p + X + p Double Pomeron Exchange (DPE) diffractive X= particles, glueballs Discovery Physics p + p  p + p elastic Single Diffraction Dissociation (SDD) QCD color singlet exchange: C  1(Pomeron), C  1 azimuthal rapidity This talk

4. Helicity amplitudes for spin ½ ½  ½ ½ June 2, 2013Kin Yip4 Elastic Scattering

5.  A N and Nuclear Coulomb Interaction June 2, 2013Kin Yip 5 In the absence of hadronic spin-flip contributions, A N is exactly calculable. [ Kopeliovich & Lapidus, Sov. J. Nucl. Phys. 19, 114 (1974).] N.H. Buttimore, et al., Phys. Rev. D 59 (1999) 114010. Our data reach

6. E704@FNAL  s = 19.4 GeV PRD48(93)3026 pp2pp@RHIC  s = 200 GeV PLB632(06)167 HJet@RHIC PRD79(09)094014 no hadronic spin-flip no hadronic spin-flip with hadronic spin-flip with hadronic spin-flip HJet@RHIC PRD79(09)094014 no hadronic spin-flip no hadronic spin-flip Previous A N measurements in the CNI region June 2, 2013 Kin Yip 6 no hadronic spin-flip no hadronic spin-flip

7. Implementation at RHIC  Detectors June 2, 2013 Kin Yip 7 Vertical (-58 m) Horizontal (-55 m) Horizontal (55 m) Vertical (58 m) IP (STAR) Silicon pitch is ~100  m

8. The most significant matrix elements are L eff ’s, so that approximately : x D   x * y D   y * Scattered protons have very small transverse momentum and travel with the beam through the accelerator magnets Roman Pots (RP) get very close to the beam without breaking accelerator vacuum (~ 12  ) Excellent detector efficiencies (plane eff. > 99%) allow us to have clean data. RP positions have been optimized such that with proper machine optics (parallel-to-point focusing), the positions of the protons at the RP’s depend almost exclusively on the scattering angles Beam transport equations relate measured position at the detector to scattering angles : a 11 a 13 a 14 a 21 a 22 a 23 a 24 a 31 a 32 a 33 a 41 a 42 a 43 a 44 xxyyxxyy x*  x * y*  y * D IP Roman pots and transport June 2, 2013 Kin Yip8

9. Selection cuts … Collinearity June 2, 2013 Kin Yip 9 Various data quality and fiducial cuts to select elastic protons and eliminate backgrounds  ~21 million events (all spin combinations) One of the most important cuts is collinearity for protons on both sides of IP A typical distribution of  x =  [  x (West)   x (East)] and  y =  [  y (West)-  y (East)] for a run is shown here. The width (   58  rad ) here is consistent with the beam divergence. The background-to-signal ratio under the Gaussian distributions in 3  is  0.4%.

10. Calculation of single-spin asymmetry A N June 2, 2013 Kin Yip 10 Square root formula: don’t need external normalization, acceptance asymmetry and luminosity asymmetry cancel out We have all bunch polarization combinations: , , ,   can build various asymmetries where Beam polarization*: P B = 0.604±0.026 and P Y = 0.618±0.028 (P B +P Y ) = 1.224 ± 0.038, (P B – P Y ) = –0.016 ± 0.038 = 0.013(P B +P Y ) and there is an additional global error ~ 4.4% on (P B +P Y ). *Averaged for our fills from the official Run’09 CNI polarimeter results http://www4.rcf.bnl.gov/~cnipol/pubdocs/Run09Offline/ Both beams polarized – half of the statistics, but effect ~ (P B +P Y ) Opposite relative polarization – effect ~ (P B –P Y ) should be close to 0 – systematic check and  is the azimuthal angle.

11. A N fits in 5 bins June 2, 2013 Kin Yip 11 False Asymmetry Published: Phys. Lett. B 719 (2013) 62–69

12. Results on A N and r 5 June 2, 2013Kin Yip 12 1  2  3  no hadronic spin-flip no hadronic spin-flip Our fit STAR Published: Phys. Lett. B 719 (2013) 62–69

13. r5r5 June 2, 2013 Kin Yip 13 Phys. Lett. B 719 (2013) 62–69 Phys. Lett. B632 (2006) 167- 172

14. Cannot use square root formula – have to rely on normalized counts K +/– = N +/– / L +/– where N = # elastic events; L=normalization factor Double spin effects are seen but very small 0.003<−t (GeV) 2 <0.035 STAR PRELIMINARY Large systematic shift of 0-line is possible due to normalization STAR PRELIMINARY Both A NN and A SS are very small ~10  3 (except for the lowest t-range where larger systematic shifts may occur). Further normalization studies use INDEPENDENT ratios instead of luminosity and we have considered using STAR detectors such as beam-beam counters (BBC), vertex position detector, zero degree calorimeter and even Wall Current Monitor in RHIC. A NN and A SS June 2, 2013 Kin Yip 14 A NN A SS

15. Careful study of several STAR subsystems for extracting of normalization counts showed that BBC is nearly free of double spin effects and can be used for normalization. Analysis of 3 processes in BBC with sensitivity to various physics allows to estimate the uncertainty in  (A NN +A SS )/2 at the level of  8  10  4. June 2, 2013 Kin Yip 15 Accuracy of our determination on A NN / A SS Projection on  (A NN +A SS )/2

16. In the double Pomeron exchange process, each proton “emits” a Pomeron and the two Pomerons interact producing a massive system M X where M X =      gg(glueballs)?,  c (  b ), qq(jets), H(Higgs boson), The massive system could form resonances. We expect that, due to the constraints provided by the double Pomeron interaction, glueballs, hybrids, and other states coupling preferentially to gluons, will be produced with much reduced backgrounds compared to standard hadronic production processes. For each proton vertex one has t four-momentum transfer  p/p M X = invariant mass 16 Central Production in DPE pp MxMx RHIC June 2, 2013Kin Yip

17. June 2, 2013 Kin Yip17 Invariant mass of centrally produced π  π  pp + 2 charged non-collinear tracks p T miss < 0.02 GeV Δθ > 0.15 mrad (avoid cosmic) │dE/dx –(dE/dx) π │  3σ Distribution is similar to AFS@ISR Dominated by low invariant mass pairs  1 GeV Characteristic sharp drop ~ 1 GeV may be due to     rescattering or f 0 (980) interference with the S-wave background { Eg. Acta Phys. Polon. B42:2013 (2011) [arXiv:1011.0960 [hep-ph] }arXiv:1011.0960

18. June 2, 2013 18 Phase II: Phase-II* (intermediate step) and Phase-II allow running with standard conditions in the STAR expt. (no special runs needed)  thus large data samples. Phase II Kin Yip

19. To summarize: >20 million good elastic events recorded in 5 days of data taking with RP’s in 2009 at  s=200 GeV and special machine optics  *=21 m. Published the most precise measurement of A N at  s = 200 GeV (Phys. Lett. B 719 (2013) 62–69) Double-spin asymmetries A NN and A SS and CEP results planned to be submitted for publication in near future. Planning a Phase II which allows us to study diffractive physics (Central Production, Single Diffraction Dissociation and its spin dependence) and exotic physics etc. The physics program with tagged forward proton at STAR helps explore physics potential and discovery possibilities at RHIC. June 2, 2013 Kin Yip 19

20. June 2, 2013 Kin Yip 20 Backup

21. Roman Pots (RP) @ STAR in 2009 Vertical and Horizontal RP setup for a complete  coverage June 2, 2013 Kin Yip 21

22. June 2, 2013Kin Yip 22 Silicon Detector Performance in the 2009 run Only 5 dead/noisy strips per ~ 14000 active strips (active area limited by acceptance) Overall plane efficiency > 99%. Excellent detector efficiencies allow us to have clean data. After excluding (3) edge strips and hot/dead strips

23. Data was taken in 5 days ~July 2009 >70 million triggers Elastically triggered ( ~33 million events) Outside sequencer (DAQ) reset time window Valid hit/strip with ADC  pedestal_per_channel + 5  A Cluster: ≤5 valid consecutive hits with ADC sum separated from the pedestal (depending on size) Clusters in A/C and B/D strips are within < 200  m (2 strips) A track on each side (a track is formed by at least a cluster on each Roman Pot) Collinearity (between the scattering angles in the East and West) Timing-vertex cut Fiducial cuts and getting rid of hotspots in the RP’s nearest to the beam (~tail of the beam)  ~21 million events (before considering the spin combinations) June 2, 2013Kin Yip 23 2009 run and main Selection Cuts

24. Finalizing the analysis … We have focussed our efforts to figure out the best and most accurate transport for the RHIC configuration that we have used during the 2009 run. Cross-checked the transport and the off-axis effects by established software such as MADX and Turtle (to arrive at the same result). Determined (eg.) the signs of angles and magnet strengths related to our experiment with the help of accelerator physicists and dedicated new/old beam experiments. Determining the best way to determine the scattering angle from positions in the RP’s (using MC simulation to tell us the accuracies). We have spent a lot of time in alignment by detector position surveys in the RHIC tunnel and using the (over-)constraints from the elastic data to better align the detector geometry. A lot of systematic checks have been done. Eg.: June 2, 2013 Kin Yip 24

25. June 2, 2013Kin Yip 25 A check to show that we understand the optics of our system: We compare the slope of the straight line fit of the angle  (RP) vs the coordinates(RP) obtained from the data to the slopes of the fits in Turtle simulations when we vary the quad. strength. For example : Conservatively assigned an systematic error of 0.5% on magnetic strength  1.4% in - t. P. Pile, et al., in: Proceedings of IPAC12, 2012, p. 1131. One example of systematic checks :

26. June 2, 2013 Kin Yip26