In August 2000 the first phase of operation (Run I) of HERA, DESY’s electron/positron-proton collider, came to a successful conclusion after the machine reached a luminosity of 2 x 1031 cm^-2 s^-1, surpassing its original design luminosity by 30%. The total luminosity delivered between 1992 and 2000 to the colliding-beam experiments H1 and ZEUS amounted to about 190 pb^-1, and electron and positron beams with a longitudinal spin polarization of up to 60% were routinely delivered to the HERMES experiment, which uses a gas target. In addition, proton-nucleus interaction rates as high as 5-20 MHz were provided for the HERA-B experiment. The run enabled the four experiments to publish a large number of results on the strong and electroweak interactions as well as physics beyond the Standard Model.

The objective of the second phase of the HERA program, Run II, was to operate with a greater luminosity, about four times higher than the design luminosity of Run I. The upgrade began in 2001 and proved challenging in several respects. By October 2004 the collider had completed a year of successful running with positrons and could be switched to electron-proton operation, for the first time since 1999.

The upgrade challenges
In order to provide the greater luminosity required for Run II, the interaction regions of the colliding-beam experiments had to be rebuilt to reduce the beam cross-section at the collision points by a factor of three, to values of 112 µm x 30 µm. In addition, the interaction regions of H1 and ZEUS were to be equipped with pairs of spin rotators to allow for longitudinally spin-polarized lepton beams in collisions with protons.

The requirements of the upgrade represented a challenging engineering project. Strong focusing magnets had to be fitted inside the existing detectors – a task made very difficult by the small apertures involved, the limited available space and insufficient access to support points. In addition, new technologies had to be developed in order to improve the interaction regions still further, which had already been optimized during HERA Run I. The technical novelties developed for the upgrade include large-aperture superconducting combined-function magnets with small outer dimensions, which are supported inside the narrow aperture of the colliding beam-detectors. This means that a beam separation closer to the experiments is necessary to achieve stronger focusing.

Moreover, inside the strong magnetic fields of the superconducting magnets, the lepton beam emits high-power synchrotron radiation. This requires a sophisticated vacuum system to handle the large power loads and to provide at the same time the excellent vacuum pressure of 0.1 nanotorr needed around the detectors for tolerable background conditions.
The upgraded components were installed during a shutdown from September 2000 to July 2001, which was followed by a period of technical commissioning and commissioning with beam in the autumn of 2001. The high luminosity that can be achieved in the upgraded configuration was demonstrated soon after the accelerator was restarted. In October 2001, a specific luminosity of L_spec= 1.8 x 10³⁰ mA^-2 cm^-2 s^-1 was reached with a small number of bunches, a value about two-and-a-half times greater than the ones achieved before the upgrade.

Fighting background problems
For the collider experiments, H1 and ZEUS, however, it was a different story. It turned out that the backgrounds they saw were larger than expected and this prevented turning on the tracking detectors in H1 and ZEUS. There was even a risk of damaging some detector components close to the beam. This led to considerable joint efforts between the accelerator and experimental groups to explore and to understand the reasons for the high background and to develop appropriate countermeasures. These efforts included detailed Monte Carlo simulations of the background conditions, which were benchmarked with accelerator experiments. This process required a considerable amount of accelerator study time. The results and conclusions were discussed during an international workshop in July 2002 and the improvement program was presented to an international review committee in January 2003.

This thorough analysis led to the conclusion that the backgrounds generated by protons lost in the interaction-region beam correlated with the poor initial vacuum conditions in the new system in the presence of the positron beam. The vacuum recovery was also slowed down by considerable thermal desorption of synchrotron radiation masks inside the beam pipe close to the experiments. This was due to higher-order mode heating at injection energy when the bunches are short. In addition, in the spring of 2002 it became apparent that the ZEUS detector was also being hit by scattered synchrotron radiation. This was caused by a problem with a mask that was actually designed to shield the detector against it.
These problems limited the intensity during running in 2002, and this in turn allowed only slow recovery and conditioning of the vacuum. However, by the end of 2002 a significant improvement in the vacuum at the interaction region and a corresponding reduction of the proton-induced background had been achieved. This indicated that tolerable background conditions with full beam currents would be possible after further conditioning.

At the same time, the more intricate operational procedures of the upgraded accelerators were consolidated. These include global and local orbit stabilization systems with active feedback, which control the beam orbit to 0.1 mm during injection, acceleration, low-beta squeezing, tuning and luminosity running. High longitudinal spin polarization of the positron beam was tuned up and measured for the first time simultaneously at all three interactions points during a test run in February 2003. HERA was then able to report the achievement of a world first: the collision of a longitudinally spin-polarized positron beam with high-energy protons.

Before this achievement could be exploited in the physics program, however, the shutdown period from March to July 2003, which was needed to complete the experimental detector upgrades, was used to improve the synchrotron-radiation masks. The shape of the masks was changed to reduce higher-order mode losses of the beam, the cooling of the masks was improved, and the problem with the mask inside the ZEUS detector was resolved. Furthermore the pumping of the beam pipe inside the H1 detector and in a long beam-pipe section inside one of the magnets was improved to speed up the vacuum conditioning. These measures all achieved the desired effects: the vacuum system recovered quite quickly after the shutdown, the higher-order mode heating was reduced considerably and – most importantly – the problem with scattered synchrotron radiation in the ZEUS detector was completely solved.

Back to high luminosity

High-luminosity operation with protons and positrons started after vacuum conditioning with beam in October 2003. However, beam intensities in November and December were limited by new rules on radiation safety, which required an upgrade of the active machine-protection system. This was accomplished by the end of December 2003. Then, from January 2004, the HERA beam currents were increased steadily and the operating currents previously achieved in 2000, of around 100 mA protons together with 48 mA positrons, were reached. From January to June 2004, the HERA luminosity was increased from 1.6 x 10³¹ cm^-2 s^-1 to 3.8 x 10³¹ cm^-2 s^-1, which is twice the value achieved in 2000. At the same time, the longitudinal positron spin polarization was tuned to values up to 50%. By August 2004, a total integrated polarized positron-proton luminosity of 92 pb^-1 had been delivered to the collider experiments. As a result, all three HERA experiments – H1, ZEUS and HERMES – have successfully taken data in 2004, with interesting first results presented in August 2004 at the International Conference on High Energy Physics in Beijing.

HERA’s luminosity upgrade is nearly complete, and we are now looking at increasing the luminosity again by another 50%. This requires further increasing the beam intensities, and better control of the beam parameters and the specific luminosity. An improvement program to achieve this goal during 2005 is under way.

The proton background conditions for the experiments steadily improved during the 2004 run. In February 2004, the ZEUS experiment reported excellent background conditions together with large luminosity, and the proton-induced backgrounds in H1 have been demonstrated to be tolerable up to the highest beam intensities. Unfortunately, a number of vacuum leaks in the interaction regions due to a weakness in the design of a flange connection temporarily led to larger vacuum pressure there, resulting in poor background conditions. During a shutdown in August and September 2004, which was required to perform the annual safety tests and some detector repair work, the interaction-region vacuum system was improved further.

After this shutdown, HERA resumed operation with protons and electrons, rather than positrons, for the first time since 1999. To maximize the integrated luminosity of HERA over the coming years, a program is under way to improve the availability of the components and the overall operational reliability. In addition, a longitudinal broadband damper system is being developed to control coupled bunch instabilities. This will help to control the proton bunch length and will provide a minimum effective transverse beam size so as to maximize luminosity.

The present plan is to continue the electron run until mid-2006, then switch back to positrons and complete the HERA data-taking by mid-2007. The three experiments are ready and eagerly awaiting a large harvest of HERA II data.