Previous Up Next

6  Neutrino Beam Design

The geographic location of BNL on one side of the continent allows us to send beams to a variety of distances including very long baselines of 2000 km or more. This is shown in Figure 38. The distances from BNL to Lead, SD (Homestake), and WIPP in NM are 2540 and 2880 km, respectively. The respective dip angles are 11.5, and 13.0 degrees. The difficulty of building the beam and the cost increases with the dip angle but all these angles and distances are feasible.

Our conceptual design for a beam to Homestake is shown in Figures 39 and 40. It can be adapted to any far location in the Western direction. Our design addresses a number of issues. At BNL we are constrained to keep the beam line above the water table which is at a shallow depth (∼ 10 m) on Long Island. Therefore the beam has to be constructed on a hill that is built with the appropriate 11.5 degree slope. Fortunately, it is relatively easy, and inexpensive to build such hills on Long Island because of the flat, sandy geology. It is important to keep the height of the hill low so that the costs are not dominated by its construction. The proton beam must be elevated to a target station on top of the hill. The cost of the hill can be lowered by bending the proton beam upwards as quickly as possible. We have, however, chosen the design and the bend angle used for the RHIC injection lines in our proposal because the RHIC injection lines have well known costs.

The proposed fast-extracted proton beam line in the U-line tunnel will be a spur off the line feeding RHIC. It will turn almost due west, a few hundred meters before the horn-target building. In addition to its 90 degree bend, the extracted proton beam will be bent upward through 13.76 degrees and then down by 15 degrees to strike the proton target. The downward 11.30 degree angle of the 200 meter meson decay region will then be aimed at the 4850 feet level of the Homestake laboratory. This will require the construction of a 54 meter hill to support the target-horn building, so as to avoid any penetration of the water table. At its midpoint (about Lake Michigan) the center of the neutrino beam will be roughly 120 km below the Earth's surface.

Figure 38: Possibilities for very long baselines from BNL. The distances from BNL to Lead (Homestake), and WIPP are 2540, and 2880 km, respectively.


Figure 39: The beam line for sending a neutrino beam to Homestake mine, South Dakota. This same beam line can be adapted for any far location in the Western direction.


Figure 40: Elevation view of the neutrino beam line to Homestake, South Dakota. For a nearer location a much smaller hill can be constructed. In this beam we assume a decay tunnel length of 200 m. For a shorter tunnel the cost of the hill will reduce as shown in Table 5.

6.1  Optimization of the wide band spectrum


Figure 41: The design of the horn focusing system used for the E734 experiment adapted from the E889 proposal.


Figure 42: Wide band horn focussed neutrino spectrum for 28 GeV protons on a copper target. The spectrum is approximately the same if Super-Invar is used as target material. Spectra of neutrinos are calculated at various angles with respect to the 200 m decay tunnel axis at the AGS and at a distance of 1 km from the target.

For this report we have attempted to optimize the beam for the Homestake distance (2540 km). However, our optimization process could be applied to any distance. As already explained, the ideal beam for Homestake will be a broadband beam that covers ∼0.5 GeV to ∼7.0 GeV range. The νµ→ νe process through Δ m212 (solar oscillations) will generate a sizable effects at the lowest energies. The energy range 1-3 GeV will be important for the detection of CP violation. The energy region 3-5 GeV contains the first matter enhanced (for neutrinos with regular mass hierarchy) νµ→ νe oscillation maximum. Recall that the highest energies are important for establishing the existence of νµ→ νe signature because this region is free from the neutral current π0 background and should have very good efficiency for the signal. Lastly, the energy region 6-7 GeV is important for the νµ disappearance measurement.

To obtain a broad band neutrino spectrum we have adapted the standard scheme of multiple parabolic horns, each one focuses a different pion momentum region. The difficulty with this approach is that the lowest energy pions we need to capture and focus are approximately 1-2 GeV and come from a long target. Figures 41 and 43 shows the design of the target and horn geometry for a conventional wide band neutrino beam, similar to that used in previous experiments at BNL, such as E734. The E734 design used a water cooled 1.5 interaction length copper target. The calculated energy distributions of a νµ beam produced by 28 GeV protons is shown in Figure 42 [21]. The 0∘ calculation has been shown consistent with neutrino beam data [29]. A copper target will not survive the 1 MW intensity proton beam that we propose. Therefore, both new materials and new focusing geometries must be considered.

We discuss the target in much more detail in a later section. The two main issues in the target design are the target material and the space available for cooling. If a dense material, such as Super-Invar, is used then the spectrum will be approximately the same as shown in Figure 42. A better approach is to use graphite as the target material and modify the horn geometry to allow for a longer target (Figure 43). The result of these modifications is shown in Figure 44. The electron neutrino contamination is shown on the same scale in Figure 45. We have used a 1.5 interaction length graphite target. As shown in the figures, the flux resulting from a graphite target is considerably higher in the 3.5 to 8 GeV region. There is no significant change in the ratio of electron type neutrinos to muon type neutrinos between a graphite and a copper target. We have used the neutrino flux from Figures 44 and 45 for the calculation of event rates and backgrounds in the rest of this report.

Figure 43: The horn geometry in the GEANT simulation. The vertical and horizontal scales are in the ratio of 1 to 13. The beam is incident from the left.

There is a large (∼ 50%) model dependent uncertainty on the neutrino flux at high energies (>4 GeV). In particular the hadron production model in MARS gives lower flux than in GEANT [30]. This uncertainty will most likely be resolved by new experiments [31, 32] in the near future.

Figure 44: Wide band horn focussed muon neutrino spectrum for 28 GeV protons on a graphite target. The spectra of neutrinos are calculated at various angles with respect to the 200 m decay tunnel axis and at a distance of 1 km from the target.


Figure 45: Wide band horn focussed electron neutrino spectrum for 28 GeV protons on a graphite target. Spectra of νe are calculated at various angles with respect to the 200 m decay tunnel axis and at a distance of 1 km from the target.

Further work on the optimization of this spectrum for the very long baseline experiment is ongoing. Further optimization focuses on enlarging the horns to accept more lower energy pions so that the flux near 0.5 GeV can be enhanced, using an evacuated or helium filled decay tunnel, and as using the hadron hose [33] to capture more higher energy particles.

6.2  Target Station

To use the 1 MW proton driver proposed for BNL, serious consideration must be given to the target selection. It is desirable to choose a solid target for generating a high intensity neutrino beam. For pion production with high power proton beams, target integrity becomes an important issue. Up to now, the production of secondary particles has been limited to proton beams with average beam power on the order of 100 to 200 kW. We now have to consider a target which can survive a 1 MW (or greater) average proton beam power. For a 28 GeV proton beam, 1 MW beam power implies 2.23× 1014 proton/sec. For a rep-rate of 2.5 Hz we then must consider nearly 100 TP per spill. A number of options have been considered and investigated both in terms of the material selection as well as the feasibility of target configuration. In evaluating the target choices the following concerns are being addressed: Findings of a number of recent studies [24], including experimental results from AGS Experiment E951 [34], on target issues for the muon collider/neutrino factory project are taken into consideration in this effort.

Figures 47 and 48 show the spectra of π+ and π- that are produced from a 2-interaction length target for various materials. For a conventional neutrino beam, the useful part of the pion spectrum is in the energy region above 2 GeV. For this reason, high-Z targets are no longer advantageous and low-Z targets are preferred.

In addition to maximizing the flux, the target/horn configuration must survive the thermal shock induced by the beam and the high current. Specifically, the target scheme must (a) ensure the removal of the deposited beam energy within the 400 ms period and (b) survive the thermally induced elastodynamic stresses that are expected to be comparable to the mechanical strength of most common materials. Similar concerns are valid for the horn, itself, which will be subjected to rapid heating and, as a result, high levels of thermal stress that will propagate in its volume. In order to satisfy the first requirement, several cooling scenarios are being investigated such as edge-cooling, forced helium cooling in the space between the target and the horn, and radiation cooling. All of these schemes present challenges stemming from integration with the horn in a limited space. To satisfy the second requirement, materials must be selected such that they can withstand and attenuate the thermal shock and be radiation resistant. To address this, low-Z carbon based materials such as graphite and carbon-carbon composites are being considered. These materials, while they have a lot of promise, present some challenges. Figure 46 shows the target mounted in the first horn. Also the helium cooling system for the target and the water cooling manifold for the horn are shown.

Figure 46: Sketch of the first horn with the graphite target mounted. The target is cooled by helium. The horn is cooled by spraying water on the conducting surface.

Two different forms of carbon, ATJ graphite and a carbon-carbon composite are considered as candidate target materials. These two types have been exposed to the AGS beam in the E951 experiment[34]. The carbon-carbon composite is a 3-D woven material that exhibits extremely low thermal expansion below 1000oC and responds like graphite above that. Preliminary studies on the feasibility of using carbon-based targets for this neutrino beam have been conducted. Specifically, utilizing the energy deposition estimates from MARS for 1 mm and 2 mm RMS beam spots (corresponding to 3 mm and 6 mm radii of target), the thermal shock response and the survivability potential of the target were studied. The total energy deposited on the target (and which needs to be removed between pulses) is 5.1 kJ for the 1mm spot and 7.3 kJ for the 2mm spot.

Since the 1 mm RMS beam spot is the most serious case, it is examined in detail. For the 100 TP beam the peak energy density is of the order of 720 J/gram. This is expected to lead to instantaneous temperature increases of ∼ 1000∘C. A detailed finite-element analysis that involves both the horn and the target needs to be performed so the heat removal of the system can be optimized and, most importantly, so the thermal shock stresses can be computed. A material with a small thermal expansion should experience smaller thermal stresses. However, carbon-carbon composite materials exhibit an increasing thermal expansion at higher temperatures. This behavior of the material needs to be examined further. If the high temperature performance of this material is not satisfactory, a larger beam spot size could be used. From energy density considerations, a 2 mm rms beam spot would have a peak temperature rise per pulse that is less than a third of the 1 mm rms case. This would ensure that the material will be well within the safe zone. Cooling of the front-end is achieved by maintaining the temperature at the surface of the first 4 cm to 27oC.

Figure 47: The number of π+ per incident proton is shown as a function of its momentum for carbon, copper and mercury targets. The target is two interactions lengths long for each material.


Figure 48: The number of π- per incident proton is shown as a function of its momentum for carbon, copper and mercury targets. The target is two interactions lengths long for each material.

We examine the optimal geometry for high-energy pion production utilizing a carbon target. In Figure 49 we see the result of varying the radius of a 1.5 interaction length (60 cm) long carbon target as we varied the proton beam radius. For this analysis the target radius was constrained to 3 times the proton beam rms radius. We note that although the total secondary pion production increases with radius, the desired high-energy portion of the production spectra is enhanced with smaller beam spot sizes. In Figure 50 we fix the beam/target radius at (2mm/6mm) and find that the production of 7-9 GeV pions increases with target length up to about 80 cm (2 interaction lengths) and then remains essentially constant up to 2 m.

We now explore the impact of bringing 100 TP protons/spill onto a carbon target. For this analysis we utilize MARS to calculate the energy deposition due to the hadronic showering within the target. We examine the two cases of 3 mm and 6 mm radius targets in Figure 51. We note the peak energy deposition density occurs near the entrance of the target and has the respective values of 700 and 200 J/g. As a figure of merit, 300 J/g is considered the danger regime where metal targets suffer damage due to the propagation of thermal generated pressure waves through the material. There is, however, evidence that carbon can withstand energy depositions in this regime. The best evidence to date comes from experience in the NUMI target development program. The NUMI carbon target is designed to expect 390 J/g peak energy deposition. A NUMI target test, performed in 1999, utilized a specially focussed beam to produce energy depositions in the range of 400 to 1100 J/g without any external evidence of target breakup.

Figure 49: The ratio of the numbers secondaries to the number of primaries is shown as a function of RMS beam radius. The target radius is assumed to be three times the RMS beam radius and the target length is 1.5 interaction lengths.


Figure 50: The ratio of the number of secondaries to the number of primaries is shown as a function of the target length for a target radius of 6 mm and a RMS beam size of 2 mm.


Figure 51: The energy deposition is shown as a function of target axial position for a 28 GeV 100 TP beam.

The secondary particle shower resulting from the interaction of primary protons with the low-Z target will add to the transient heat load of the horn. This shower will be less significant for low-Z targets than for high-Z targets. However, its effect will be examined, and added to the electric resistance heat load estimated above.

There will be activation of the target and horn structure due to secondary and primary particles. This activation will be primarily due to spallation products and neutrons generated in the secondary shower. The survival of the primary target in the radiation field needs to be examined. This can only be carried out experimentally using a prototypic proton beam on samples of the appropriate target material. The change in physical properties including, thermal expansion coefficient, elastic modulus, and yield strength, need to be examined as a function of proton fluence.

In the current option the target is an 80-cm long cylindrical rod with 12 mm diameter sizes. The 12 mm diameter target is chosen to intercept 100 TP, 2 mm rms proton beam. With this beam size, the total energy deposited as heat in the target is 7.3 kJ with peak temperature rise of about 280∘C. Heat will be removed from the target through forced convection of helium through the outside surface. This is a good solution for the 1 MW beam power.

6.3  Cost of the neutrino beam


Table 5: Preliminary direct cost (FY02$M) of building the neutrino beam with 200 meter decay tunnel. These costs do not include EDIA contingency, and overhead.


Item basis cost
Proton transport RHIC injector $11.8 M
Target/horn E889 $3 M
Installation/Beam Dump New $2.6 M
Decay Tunnel E889 $0.4 M
Conventional const. (hill) New $8 M
Conventional const. (other) E889 $9 M
Total   $35 M
A preliminary estimate of the direct costs without burdens is shown in Table 5. The costs are based on the the RHIC injector work, as well as the E889 proposal and the neutrino factory study. The conventional construction costs are dominated by the size of the hill which is approximately proportional to the third power of the decay tunnel length. In our cost estimate we assume that we will bury the beam dump underground to reduce the height of the hill. It is assumed that the target station shielding can be retrieved from existing resources. We have also estimated the cost assuming a 200 m long decay tunnel. The spectra shown in Figure 42 are based on this 200 m tunnel filled with air.






Previous Up Next