by Alberica Toia. Published: 13 October 2012

“It's the last frontier -you would say.
We're still here dancing on the edge of the world”
(Lawrence Ferlinghetti)



On September 12th we arrived at P2 very early in the morning, waiting for news and updates from the accelerator team. We were meant to collect data from a short pilot-run, few hours of data taking that should allow the machine to optimize the setup of the injection and collision of the proton and lead beams and the experiments to be ready to interpret and understand the physics message they were carrying.

Similar experiments have been carried out at RHIC, with d+Au collisions. The main difference with LHC is that RHIC has two independent rings and hence the magnetic fields that bend the colliding particles and stabilize the beams are independent. In contrast LHC is based on a two-in-one magnet design with identical bending fields on the central orbits of the two beams. This two-in-one magnet design of the LHC (as seen in Figure 1) fixes the relation between the momenta of beams in the two rings. However, LHC can adjust independently the radio frequencies of the p and Pb beams as it uses different RF systems in each ring. The RF systems are capable of operating independently at the different frequencies required during injection and ramping. This feature of LHC also allows keeping both particle types on stable central orbits inside their respective rings during injection and acceleration.

Before the test run we were all nervous since no one was sure about the success of the test. There were many things that could go wrong and we had to be very cautious.

Early in the morning, the control room at P2 was full of people who were trying to set up the detectors and ensure a smooth run of p-Pb collisions. We reviewed the status of detectors, triggers, plans and procedures.



Diagram showing the cross sectors of the bending magnets of RHIC (left) and LHC (right)


The trigger menu had been discussed and we went through every detail just a few hours before the beginning of the data-taking. We seemed to get some noise in the SPD, seen in the dry run (i.e. runs taken for test, without beam) which would lead to the exclusion of the SPD from the trigger configuration. In addition, we reviewed the procedures for calibrating the detectors, reconstructing the data, simulating in Monte Carlo with the same configuration as in reality. Everything had to be fast and efficient for a successful run!

Once we had “stable beams” at the LHC we switched on a minimum number of detectors, mainly those that are needed for the early analyses.

We had agreed for a minimum number of events (500k). After we collected those, we switched on all other ALICE detectors. To get these data we spent almost four hours in this configuration before changing the colliding position by half a meter up- or down-stream. This special configuration allows to exploit the rapidity coverage of our detector, and to understand our systematic uncertainties.

-13:50: LHC starts the first attempt of a fill for proton and lead ions into the ring. Shortly afterwards the accelerator team has to deal with some problems with the control of the radio frequencies of the two beams and change the setup of the machine.

-16:00 LHC starts injecting; this time a problem appears in the communication between the LHC and the SPS and the lead-ion beam is not properly injected

- 18:30: It seems that the LHC is ready to fill again with ions. Unfortunately, the ion beam is lost during the ramp-up because of new collimator settings that did not work. The accelerator team just had to go back to the original ones we had planned to use. As LHC accelerates two different beams of protons and lead ions various dynamic issues have to be taken into account.

- 22:30: We get the next fill; ramp reaches its full energy around 23:00 with 15 bunches for each beam. It seems to be right this time! Finally, around midnight we get the first collisions!

- 00:50: LHC makes a beam loss monitoring using 2 out of the 15 bunches that were injected. The resulting “loss maps” are used to check and characterize the beam and the setting of the collimators.

- 1:26: stable beams are declared. We collect data with the settings we had agreed upon. At least we had two long golden runs. At dawn we get collisions on a displaced vertex, and we collect few more hours of data.



After 24 hours, around 9:35 in the morning the LHC dumps the beams. This marks the end of a successful test both for the capabilities of LHC but also for the physics program of ALICE.

We have collected more data than we hoped for. The detectors ran smoothly and not a single run was interrupted by a detector failure. There seems to be some more background activity than what we had foreseen, considering the low beam intensity.



p + Pb ion collision recorded by the +ALICE Experiment on 13 September 2012 at a center of mass energy per colliding nucleon-nucleon pair of 5.02 TeV. Such p + Pb collisions serve as an important reference for studying the novel hot, dense matter produced in Pb + Pb collisions at the LHC.


We have 2 runs taken in ADJUST: a test run of ~1 h with closed gates, and a test run ~0.5 h with open gates (25 ns), taken with our minimum bias CINT1 trigger. Moreover we have two golden runs taken with STABLE BEAMS with the slightly more selective CINT7 trigger. During these runs we collected 550k events with the minimum defined set of detectors, and 1.2M with all the detectors.Finally, we also performed four runs with displaced vertex (+/- 50 cm). They are important to test our detector settings. In fact collisions happen at a given time with respect to the beam clock. For these tests the beam clock phase did not change. This means that the collisions happen earlier or later by 1.67 ns at a different position. Therefore we have to adjust our coincidence timing. The total collected statistics gives us ~ 1.7 M min. bias events which is useful for studying the physics performance (measuring dN/d?, R_pA, strangeness, etc). We all agreed to go through the reconstruction right away, using the first calibration done online while at the same time we were going through the standard procedure for offline calibration.

Proton-nucleus (pA) collisions have long been recognized as a crucial component of the physics programme with nuclear beams at high energies and their role as references for understanding the nucleus-nucleus data and elucidating the partonic structure of matter. Experience from previous heavy-ion programs shows that p-A baseline is essential for the interpretation of some main discoveries (J/? suppressionm jet quenching) and hence it is important to get data from these collisions at LHC.



One of the first proton-lead collision events seen by ALICE (13.9.2012). This is a reconstructed event from the High Level Trigger (HLT), showing tracks from the Inner Tracking System (ITS) and the Time Projection Chamber (TPC)


The data collected during the test should be sufficient to study the multiplicity, which is strongly interconnected with the centrality. The p+A collisions provide a unique opportunity for characterizing the nuclear dependence of parton distribution functions and understanding the final-state medium effects such as collective phenomena that are present in the QGP. Data from the p-A test run are also crucial for deciding the triggering that will be used in January, when the intensity will be higher. We can try to understand how the trend of momentum spectra is formed and the factors that affect its shape. We also need some basic particle composition of this new type of collisions in order to be ready for the full p-A run that will take place in January.

The recomissioning will start on 07/01 and the data taking period is scheduled from the 18/01 until the 10/02. It has been almost one week later and we're still here, dancing at the edge of the world. It's the last frontier.