T. Mastoridis

Synchronous Phase Shift at LHC

The electron cloud in vacuum pipes of accelerators of positively charged particle beams causes a beam energy loss which could be estimated from the synchronous phase. Measurements done with beams of 75 ns, 50 ns, and 25 ns bunch spacing in the LHC for some fills in 2010 and 2011 show that the average energy loss depends on the total beam intensity in the ring. Later measurements during the scrubbing run with 50 ns beams show the reduction of the electron cloud due to scrubbing. Finally, measurements of the individual bunch phase give us information about the electron cloud build-up inside the batch and from batch to batch.

Running the RF at higher energy and intensity

The improvements done to the RF parameters and hardware in 2011 are reviewed. Then the upgrades planned for 2012 are presented: further reduction of capture losses with the longitudinal damper, batch by batch blow-up at injection and modification of the controlled blow-up to preserve bunch profile. Operation at higher energy is readily possible with the present RF power, and does not degrade longitudinal stability thanks to the controlled longitudinal emittance growth during the ramp. For operation with higher beam current, the observations in 2011 indicate that there is no single bunch instability issue with up to 3 10 11 p per bunch. With the large gain of the RF feedback and One-Turn feedback, the cavity impedance at the fundamental will not be a limitation for ultimate intensity (1.7 10 11 p per bunch) with 25 ns spacing. The klystron power (300 kW RF at saturation) is sufficient for 25 ns operation with nominal intensity (2808 bunches per beam, 1.1 10 11 p per bunch). An RF roadmap for going beyond will be outlined: it calls for an upgrade of the LLRF only and should allow for operation with ultimate beam intensity (25 ns spacing, 2808 bunches, 1.7 10 11 p per bunch) after Long Shutdown one. NEW FEATURES 2011 VERSUS 2010 The statistics on RF faults has been presented at the Evian workshop [1] and will not be repeated here. Increased capture voltage At extraction, the SPS RF is 7 MV (200 MHz). The bunch has 1.5 ns length * (4t), and 4.5 10 -4 energy spread E/E (2 E), resulting in a 4tE emittance of 0.5 eVs † . In 2011 the LHC capture voltage was increased from 3.5 MV (2010) to 6 MV. The bucket area was increased from 0.9 eVs in 2010, to 1.2 eVs. The bucket half height is now 9.6 10 -4 E/E. With 7 MV at 200 MHz, the SPS bucket area at extraction is 3 eVs but the longitudinal distribution is limited to a much smaller region: controlled longitudinal blow-up is applied during the SPS ramp to keep the beam stable [2]. The blow-up is turned off near the moment when the voltage program corresponds to a bucket area of 1.05 eVs only. The voltage is then raised adiabatically to 7 MV for bunch shortening before transfer to the LHC. We can therefore * We quote the 4length of a Gaussian bunch having the same Full Width at Half Maximum as the measured bunch. † At CERN it is customary to quote the longitudinal emittance as † At CERN it is customary to quote the longitudinal emittance as 4tE. Note that, for a Gaussian distribution, and small filling factor, 95% of the particles are within a 6tE area. The 4tE area contains 86.5% of the particles. limit the SPS bunch to a 1.05 eVs contour in a stationary 7 MV bucket (at 200 MHz). Figure 1 shows the situation: the 1.05 eVs contour falls almost entirely within the LHC bucket. Assuming a Gaussian distribution for the SPS bunch, truncated at the 1.05 eVs contour, the calculated loss is 0.02%. Figure 1: Longitudinal phase space at injection: We assume a Gaussian distribution for the SPS bunch and display contours corresponding to steps of 5% in integrated intensity. The Gaussian is truncated at the 1.05 eVs contour (yellow). In figure 2 we introduce a small injection error (100 ps and 10 -4 p/p). This results in a small portion of the bunch falling outside the LHC bucket (calculated 0.4 % loss with the truncated Gaussian model). In 2011 we have observed 0.5% loss from injection to start ramp. Figure 2: Longitudinal phase space at injection with a small error: 100 ps and 10 -4 p/p. A consequence of the voltage mismatch (matched voltage is around 2.5 MV) is the bunch length reduction after capture (from 1.5 ns to 1.1 ns). We could take advantage of the large available bucket to blow up the longitudinal emittance after each injection and restore the 1.5 ns length (batch-by-batch blowup). With 1.5 ns and 6 MV, we get 0.83 eVs (4Et) emittance. We could increase it further by capturing with 8 MV as for the Lead ions, leading to 0.97 eVs. This is planned for 2012. Larger voltage in physics In 2011, controlled longitudinal emittance blow-up was applied in the eleven minutes long ramp, keeping the bunch length around 1.2 ns, while the RF voltage was increased linearly from 6 MV to 12 MV. (In 2010 we used 8 MV only in physics). The 12 MV provide a larger longitudinal emittance, thereby reducing the transverse emittance growth due to Intra Beam Scattering. At the beginning of the 3.5 TeV flat top we now have 2 eVs longitudinal emittance in a 4.7 eVs bucket (1.5 eVs in a 3.8 eVs bucket in 2010).