Minutes of the Booster Commissioning Working Group held on 30th of July 2009

Present: A. Lombardi, G. Bellodi, M. Vretenar, D. Nisbet, T. Zickler, R. Scrivens, U. Raich, B. Puccio (first part), O. Aberle, Y. Kadi, C. Carli, T. Hermanns, K. Hanke, B. Mikulec.

Agenda:

1. Communications
2. Follow-up of open actions
3. Transverse emittance measurement after Linac4 (T. Hermanns)
4. Upgrade of the LBS line (T. Hermanns)
5. AOB

1. Communications

The last minutes have been approved.

B. Mikulec is preparing a document on interlocks, which are necessary to safely operate Linac4 and the injection of the beam into the PS Booster. A first version will possibly be distributed for reading and discussion by the end of the week.

On Wednesday, 29. July 2009, a meeting dedicated to the BLMs for Linac4 took place. The key message of this discussion is that the specifications of these devices for the entire PS complex will be defined and filled into a document. For that reason discussions with BI will be started.

2. Follow-up of open actions

No open action could be closed.

3. Transverse emittance measurement after Linac4

B. Mikulec and T. Hermanns presented the results of a study, which tested the possibility to measure the transverse emittance values at the exit of Linac4 (dump line) and close to the PSB injection point (LBE line) via the Three-Monitor-Method (090730_TransverseEmittanceMeasurement.pdf). From the theoretical point of view the emittance values can be reconstructed by a linear formalism when the beam spot sizes at three positions and the transfer matrices between these locations are known. This implies that non-linear effects and first and foremost the influence of space charges can be neglected. Technically, the programme PATH simulates the beam size “measurements”, while the transfer matrices are extracted from a Trace3d-simulation of the measurements lines.

The beam profile monitors are optimally installed at positions where the phase space ellipses rotate by 60 degrees in phase space. This implies that the first monitor is located before, the second monitor close to and the third monitor behind a minimum of the beta-function (beam waist). In the present simulations all angles in both transverse planes lie between about 54 and 59 degrees.

The experimental requirements are established by beam manipulations with two quadrupoles. In the dump line the beam waist is generated after the first bending magnet in the transfer line but before the Linac4 beam dump. For that reason the quadrupole right after the Linac4 exit can be used, and only a second additional quadrupole of the transfer line style must be inserted into the so far empty line behind the bending magnet.

The beam sizes at these three positions are at the order of a few millimetres for the two outer positions and a few 100 micrometers for the monitor in the middle. As the beam size at this mid-position needs a special measuring device to determine such a small diameter a summarizing overview of an informal meeting with experts from BI taken place earlier in July was given (https://twiki.cern.ch/twiki/bin/view/SPL/Minutes13July2009). The main conclusion of this meeting is that a resolution of 10 micrometers could be obtained if scintillating screens with appropriate readout devices (high resolution and shielded against radiation) are used.

Then, the results of the reconstructed emittance are compared to the simulated ones. It is concluded that for the horizontal and for the vertical planes the deviation is only of the order of a few percent provided the stability of the method for the chosen configuration.

This point is assessed in several ways. At first, the beam current is reduced from 65 mA to 20 mA and 40 mA, respectively. Having determined new transfer matrices the emittance values are again compared between simulation and reconstruction. It is worthwhile to mention that the beam size at the mid-positions is shrinking further, so that high-resolution measurements become even more important. Then, assuming a measurement precision of 10 micrometers, it can be observed that the effect of less space charge lets the error decrease (in the best case for the horizontal emittance and 20 mA the error becomes less than 1%). In contrast, if only a resolution of 100 micrometers can be achieved the deviation can be as large as approximately 50% because for the smallest beam size a maximum rounding error is propagated to the error on the emittance measurement. At last, it is cross-checked if the emittance for I = 20 mA (40 mA) can be reconstructed if the transfer matrices for the nominal current of 65 mA are used. The error becomes several ten percents in the worst case. Therefore, per current interval a dedicated set of transfer matrices must be provided.

In the next step, it was verified that the Three-Monitor-Method gives similar results if other input parameters are used. For that reason, the same principle is applied to the beam at the entrance to the LBE-line. Thus, three beam profile monitors must be added to this line. From the technical drawings sufficient empty spaces can be identified, while the two required quadrupoles are already available. At last, a beam dump must be installed into the concrete wall separating the LBE-line from the PSB area. All other elements of the line are left untouched expect for a steerer dipole after the second quadrupole which must be moved about 300 mm downstream to create sufficient space for the first monitor tank. According to R. Scrivens this movement does not influence the operation of the line with heavy ions.

With this setup the beam spot sizes are comparable to the ones in the dump line. The deviation between simulated and reconstructed emittance values are less than 1% if no large rounding error is present.

For the LBE-line measurements systematic error studies were performed such that (a) the measured beam size error is varied by +/-50 micrometer (corresponding to a 10% error for the measurement at the second monitor), (b) the field-gradients of the quadrupoles are varied by +/-1%, and (c) the monitor positions are shifted by +/-5 mm.

The systematic errors for (b) and (c) are only of the order of 1% while the low precision of the beam size measurement of the mid-position is directly propagated to the error on the emittance measurement. In general, the systematic errors are controllable and the method gives stable results even if the experimental situation is fluctuating. However a high resolution for the monitors is desirable to keep the error at an acceptable level of about 10%.

In summary, it can be mentioned that the method works well for emittance measurements for both lines, and the most important systematic errors are low. Nevertheless, a high-resolution measurement of the beam size close to the beam waist is desirable to keep the reconstruction error acceptably low. Especially for the LBE-line, the Three-Monitor-Method is favoured because the existing kicker magnets and their power supplies, as well as the two slits need not to be upgraded. Only small modifications of the line become necessary by inserting new monitors and moving a steerer dipole. What concerns the suitability of the quadrupoles for the Three-Monitor-Method T. Hermanns will check with T. Zickler if they can tolerate the settings required for the new operational mode.

A beam dump at the end of the line must be installed in any chosen scenario not depending on the method how the transverse emittance values are measured. Its dimensions (collimator and core) must be adapted to the beam sizes. As soon as the design of the lines is closer to finalization, B. Mikulec and T. Hermanns will provide the dedicated beam parameters to Y. Kadi.

Discussion on Transverse Emitance Measurements

In order to assess further the stability of the method it was proposed that not only the emittance, but also the Twiss-parameters should be reconstructed. The latter can serve as new input parameters, so that it can be tested if the solution is iteratively converging towards a stable value. In addition, it was asked what the largest acceptable error on the knowledge of the transfer matrices is. Furthermore, R. Scrivens asked about the error on the input Twiss-parameters. It was suspected that they are related to the matching parameters between Linac4 and transfer line as well as transfer line and PS Booster. B. Mikulec and T. Hermanns will continue their systematic error study including these three aspects.

Then, it was asked by U. Raich what the influence of electron stripping on the measurement is. The electron could further interact in the scintillating material so that the beam size is smeared.

R. Scrivens mentioned furthermore, that the beam halo could possibly increase the beam size measurements. It is not sure that a uniform background description can be obtained which could be subtracted from measurement data.

Hence, it was generally questioned if the resolution of 10 micrometers is actually achievable.

Several options were proposed to become more independent on the required resolution. A. Lombadi suggested, to use either more than the minimum number of three monitors, or to perform a second measurement with the available three monitors but changed quadrupole settings. These additional measurements should then be included in the emittance reconstruction. In addition, she reminded of a method that has been reported on the CARE workshop 2008 (http://adweb.desy.de/mdi/CARE/Bad_Kreuznach/ABI_workshop_2008.html) on an alternative solution to reconstruct the emittance by a Monte Carlo method. T. Hermanns has agreed to check the suitability of these propositions for the present lines.

At last, it was discussed to use the LBE-line as dump line for transfer line commissioning rather than the LBS line. The requirement of a proper beam dump is essential but it seems to be easier to use the LBE-line for that purpose. The alternative would be to use the adjacent LBS-line. But there the beam dump must be installed in several meters height. A common agreement on this aspect could be noted, so that this became the future beam dump hypothesis for transfer line commissioning.

4. Upgrade of the LBS line

T. Hermanns presented an update on the studies for the energy measurement with the LBS line close to the injection point of the Linac4 beam into the PS Booster (090730_UpgradeLBSLine.pdf). The key question is which elements of the line must be upgraded to get it operational with a 160 MeV-Hminus beam.

The final plot shows the correlation between energy and vertical position for each particle. This must be shaped such that the central energy value and the energy spread can be derived. In this study the line is simulated using 42610 ions in PATH. A slit selects a beam slice, which is analyzed by a spectrometer magnet and detected by a SEM grid. So a few – mainly geometrical – assumptions on the slit, the spectrometer magnet, and the SEM grid are implemented as listed in the introductory part of the presentation. In the beginning, it is confirmed that the measurement is not biased when the slit selects a slice of the beam. The correlation factor of energy versus vertical position at the entrance of the slit is only 0.004. So still the entire energy distribution of the beam is represented by this sub-sample. C. Carli asked if the vertical dispersion is negligible, too, which was confirmed by A. Lomardi who said that the dispersion is compensated for in the transfer line.

For the LBS-line seven different scenarios for the spectrometer magnet design and the posi-tion of the SEM-grid are generated using the Trace3d-simulation code. Scenario 1 leaves the present line unchanged, while the other six scenarios base on the principle that the SEM-grid is located at the minimum of the vertical beta-function. This minimizes the vertical beam spread due to its natural width by the evolution of the beta-function. Only the extent of the beam due to the energy sorting in the bending magnet should significantly contribute to the vertical width. These six scenarios differ mainly in the layout of the spectrometer magnet (edge angles, radius, bending angle) and as a consequence in the drift lengths to and from the magnet.

One criterion to assess the quality of a scenario is the ratio of the beam spot size between a beam with a certain energy spread and a monochromatic beam. For that reason the beam size at the SEM-grid position is determined not only for a hypothetical beam with negligible en-ergy spread, but also for beams with 75 keV, 150 keV and 300 keV (nominal) energy spread. It can be observed per scenario that the ratio approximately scales with the same factor as the energy spread increases. For the nominal spread value it is larger by one order of magnitude, if these line layouts are considered where the SEM-grid is located in the minimum of the beta-function (all but scenario 1). Here the correlation factor between energy and vertical position is greater than 0.99, while the slope of the distribution is of the order of 70 keV/mm.

In order to determine the central beam energy from the position measurement the continuous distribution is discretized simulating a SEM-grid signal. The central value is determined by the maximum of a fitted polynomial of second order to the SEM-grid-simulated data. Having shifted the energy is shifted by up to +/-1 MeV (interval for energy ramping) a linear calibration curve of energy versus vertical position is derived. Furthermore, it is estimated that a time-differentiated readout of about 1 MHz seems to be feasible which copes with the requirements of a time-resolved energy measurement during energy ramping.

The main results of this study are summarized in the following list.

• The total energy interval, which is covered by a single SEM-grid wire, is about 50 to 75 keV (scenario dependent), which is less than half a RMS of the energy distribution. So the energy spread can theoretically be resolved as well.
• The bending angle of the spectrometer magnet, the length between the exit of the magnet, and the SEM-grid wire clearance can influence the SEM-grid energy resolution. For a distance between adjacent wires of 1 mm the resolution lies between approximately 80 keV and 110 keV again depending on the scenario. That is maybe not sufficiently low but these figures are still lower than the RMS of the energy distribution.
• Increasing/decreasing the vertical slit aperture diminishes/enhances the beam spot size ratio by approximately the same factor. But an increase of the slit width above 1 mm seems to be not an option to keep the beam spot size ratio acceptably high.
• Increasing the particle absorption rate, i.e. reducing the current without closing the slit, does not significantly affect the beam spot size ratio. That means the blow-up of the beam due to space charge effects is sufficiently small.

At the end it can be concluded which of the element must be upgraded in order to prepare the LBS line for a 160 MeV beam. The bending magnet BHZ 40 can be kept, as it can tolerate currents up to 210 A, while for the LBS(LBE) lines only 111(172) A are needed. The vertical slit aperture should be 1 mm or even less, while the longitudinal dimension must be large enough to absorb all particles hitting the material. For the spectrometer magnet no precise specifications can be given so far as they strongly depend on the design. But in general it can be calculated that a field quality dB/B=5*10-4 should be achieved. The aperture will be in the order of several millimetres (vertical) and a few centimetres (horizontal). For the SEM grid a size of +/-5 mm must be provided to measure the full energy spread. If furthermore energy shifts up to +/-1 MeV are allowed for the central value of the distribution is shifted within +/-15 mm (worst case from all scenarios). But obviously, the wire spacing must be 0.5 to 1 mm to get a sufficiently high resolution.

So far no conclusions can be drawn what concerns the selection or preference on a certain scenario. As soon as the question of the dimension of a beam dump is solved further constraints on the available space might be imposed so that some scenarios can potentially be rejected or preferred. It was noted that a close collaboration between Y. Kadi and T. Hermanns is necessary to make progress on that topic.

Discussion

On the aspect of time-differentiated SEM-grid readout, it is required that the frequency is at the order of 1 MHz because according to C. Carli, the energy ramping is performed from minimum (-1 MeV) to maximum (+1 MeV) within 10 microseconds, even if it could still increase. U. Raich pointed out that presently only the readout with a frequency of 250 kHz is available, but 1 MHz might be possible.

Y. Kadi mentioned that a slit of such a low aperture but such a large length is possible to construct but a very challenging object because it is rather a collimator than a slit. So C. Carli reminded of a scenario where due to beam manipulation a minimum of the beta-function on the SEM-grid is created. If this technique can be applied the construction of the above-described slit is not necessary anymore.

Then, it was proposed to build a pulsed spectrometer magnet with power supply because they could be possibly cheaper. This discussion is postponed until more details on the design and operation mode become available.

K. Hanke asked for adding the power supplies to the upgrade list, and R. Scrivens complemented the list by asking for a B-field measurement device.

4. AOB

There has been no other business.

-- BettinaMikulec - 30 Jul 2009

• 090730_TransverseEmittanceMeasurement.pdf: Presentation on transverse emittance measurements of the Linac4 beam (dump line and LBE line) with the Three-Monitor-Method

• 090730_UpgradeLBSLine.pdf: Upgrade on LBS Line%META:FILEATTACHMENT{name="090730_TransverseEmittanceMeasurement.pdf" attachment="090730_TransverseEmittanceMeasurement.pdf" attr="" comment="Presentation on transverse emittance measurements of the Linac4 beam (dump line and LBE line) with the Three-Monitor-Method" date="1249460783" path="090730_TransverseEmittanceMeasurement.pdf" size="7277278" stream="090730_TransverseEmittanceMeasurement.pdf" user="Main.ThomasHermanns" version="1"}%