Difference between revisions of "03/31/2020"

From GlueXWiki
Jump to: navigation, search
m
 
(10 intermediate revisions by the same user not shown)
Line 9: Line 9:
 
***[https://halldweb.jlab.org/wiki/index.php/File:PPerpSum_bggen_30980.png Transverse momentum distribution for hadronic events]
 
***[https://halldweb.jlab.org/wiki/index.php/File:PPerpSum_bggen_30980.png Transverse momentum distribution for hadronic events]
 
***[https://halldweb.jlab.org/wiki/index.php/File:PPerpSum_gen_compton_simple_30980_2.png Transverse momentum distribution for Compton events]
 
***[https://halldweb.jlab.org/wiki/index.php/File:PPerpSum_gen_compton_simple_30980_2.png Transverse momentum distribution for Compton events]
***[https://halldweb.jlab.org/wiki/index.php/File:Pprot_bggen_30980.png Proton momentum distribution]
+
***[https://halldweb.jlab.org/wiki/index.php/File:Pprot_bggen_30980.png Proton total momentum distribution]
 
**It will take more work to get the correct beam spectrum and background for all the generators.
 
**It will take more work to get the correct beam spectrum and background for all the generators.
 
**The efficiency is now also determined for the sum of the transverse momentum of the final state.
 
**The efficiency is now also determined for the sum of the transverse momentum of the final state.
Line 17: Line 17:
 
*SŠ: Coded the unpolarized and polarized Bethe-Heitler formulae, see [https://halldweb.jlab.org/wiki/images/3/36/Bh.pdf his report]  
 
*SŠ: Coded the unpolarized and polarized Bethe-Heitler formulae, see [https://halldweb.jlab.org/wiki/images/3/36/Bh.pdf his report]  
 
**AD's comments on Simon's document:
 
**AD's comments on Simon's document:
***BH asymmetries seem very small. For GDH, we are looking at asymmetries at the 3% level (see e.g. Helbing review's Fig. 38: Δσ 10 μb, with A=Δσ/(2σ₀). It seems the BH asymmetry is lower by significantly more an order of magnitude. Further, Mark simulation indicate a 10% trigger efficiency for BH, so if it is true, Δσ is further suppressed by an order of magnitude.  This points toward the possibility that we can ignore the BH for the proposal.
+
***BH asymmetries seem very small. For GDH, we are looking at asymmetries at the 3% level (see e.g. Helbing review's Fig. 38: Δσ 10 μb, with A=Δσ/(2σ₀). This points toward the possibility that we can ignore the BH for the proposal.
 
***Regarding the structure functions for the asymmetry, the atomic form factors should not be needed since the atoms are not polarized.  
 
***Regarding the structure functions for the asymmetry, the atomic form factors should not be needed since the atoms are not polarized.  
 
***For g<sub>1</sub>  and g<sub>2</sub>, probably only their values at very small-x that are truly relevant. If so, one can use a simple Regge parameterization of g<sub>1</sub>, see e.g. arXiv:1808.03202. For g<sub>2</sub>, we could just assume g<sub>2</sub><sup>ww</sup>(x,Q²)=-g<sub>1</sub>(x,Q²)+ ∫<sub>x</sub>¹g₁(y,Q²)/y dy.
 
***For g<sub>1</sub>  and g<sub>2</sub>, probably only their values at very small-x that are truly relevant. If so, one can use a simple Regge parameterization of g<sub>1</sub>, see e.g. arXiv:1808.03202. For g<sub>2</sub>, we could just assume g<sub>2</sub><sup>ww</sup>(x,Q²)=-g<sub>1</sub>(x,Q²)+ ∫<sub>x</sub>¹g₁(y,Q²)/y dy.
 
***For question #2, the angle coverage, we are planing to use the Compton Calorimeter, which cover down to 0.2°. (Note: we do not have the ComCal in the simulation yet).
 
***For question #2, the angle coverage, we are planing to use the Compton Calorimeter, which cover down to 0.2°. (Note: we do not have the ComCal in the simulation yet).
 +
 +
*AD:
 +
**Found out how to write Simon's initial properly: Š
 +
**Discussed with Richard Jones and Sasha Somov how easy it would be running in Hall D at 0.5 pass. The bottomline is: very difficult because of the smaller Lorentz boost and the long distance between the radiator and the collimator. Here are the details:
 +
***Richard Jones: It is going to be difficult to tag that beam with our tagger, for two reasons.
 +
***#only a very small fraction of the photon beam (rough estimate of 0.03%)  will make it through the collimator, while the tagger has to count the entire beam. This means that accidentals will be high even at very modest photon beam intensity at the target.
 +
***#the height of the "stripe" of post-bremsstrahlung electrons at the tagger focal plane will be an order of magnitude larger. This will lead to substantial scraping on the poles of the tagger magnet, and only a fraction of them actually being detected by the tagger. This needs to be simulated, but it might be that the quadrupole can be used to help mitigate this. Those electrons that are detected in the tagging counters without scattering from the poles will have an enhanced probability of having their photons pass the collimator, which may partially offset issue #1.
 +
****Bottom line is that it may work but not easily. The biggest irreducible issue will be the substantial scattering from the poles of the tagger by post-bremsstrahlung electrons. If the showers from these interactions dominate the tagger rate, it will be difficult to use the tagger at all for this measurement. If that is the case, can you make a meaningful measurement without tagging?
 +
****Why not consider an alternative: run at the full energy, and tag the low-energy part of the spectrum. You need to build a custom low-energy tagger, which might be a diamond strip detector like they used in the Compton in hall c. You could place it in the vacuum near the exit edge of the tagger magnet on a movable ladder, and sweep it around to cover different parts of the low-energy spectrum. These photons would be fully collimated, and very clean because the electron beam current would be turned way down to keep the low-energy flux at a level compatible with tagging. This sounds like something I would be interested in working on, provided that the physics of the low-energy part of the measurement justifies the effort.
 +
***Abstract of AD email exchanges with Richard:
 +
****AD: Regarding 1, why does so little beam go through the collimator, and how does it compare with the usual situation? Is that because of larger multiple scattering at low energy? Since we plan to use the thin Al. radiator which is about 20 times thiner than a standard diamond, then the multiple scattering should be only twice worst (~10/sqrt(20)). Is that enough to degrade significantly the transmission? Probably I am missing the reason for the low transmission. 
 +
*****RJ:  It is not due to multiple scattering, the effect is the same for a 1e-15 radlen radiator. It is the bremstrahlung cone that scales with m/E.
 +
*****AD: Ah, it's the boost! Then it's a factor ~10^2 lower transmission compared to the nominal energy, is that right?
 +
*****RJ:It is a factor 10 in the cone angle, so the transmission factor goes down by a factor 100, because of it scales like area.
 +
****AD: Regarding 2, I should have mentioned that we plan to run with both low current (50nA) and the thin Al radiator, which makes a factor ~300 less in photon flux/post-bremsstrahlung electrons. Is this making point 2) not a problem?
 +
*****RJ:No, lowering the rate does not help at all, if 99.97% of the hits in the tagger are background not coming with a collimated photon tagging is hopeless, right?
 +
*****AD: Right. Now that I understand point 1), I get this one too. Thank you.
 +
****AD: We thought about having a lower energy tagger as you suggest, especially with the offer from Lund to re-use their small tagger (which seems to have now disappeared). But we thought it was more trouble than it was worth: These low energy data have already been measured at MAINZ, so for us, taking these data again is just a matter of having an overlap with a very well measured domain. ****AD:It seems this is too much trouble for the benefit. This would have been nice because
 +
****# it offers a good check of our absolute calibration and polarimetries and
 +
****# It allows to fully cover the energy span to form the GDH integral.
 +
****RJ: I still think you might consider mounting a diamond strip detector on a movable ladder and making a low-energy tagger. If you built it, I can imagine a people in the N* community might think of things to do with it. Imagine a low energy widely variable flux tagged beam with 100% polarization, and very large (factor >30) coherent enhancement.
 +
****AD: If we decide to build a low energy tagger we don't need to discuss it especially in the proposal: since it will operate concurrently no additional special beam time is required. One setback of this option for the GDH proposal (beside having to build/commission a new detector) is that it provides very low circular beam polarization, about 5% in the region of interest.
 +
****RJ: Yes, if it is circular polarization you need then generating low-energy photons from high-energy electrons is not what you want.
 +
***Sacha Somov:
 +
****You can easily simulate this run condition using  gxtwist.
 +
****The problem at small beam energies is the bremsstrahlung angle (m/E), as you have already discussed (I believe that the  quad would not help). The rate of accidentals in the tagger has to be considered.
 +
****The magnet acts as a collimator (the gap is 3 cm); we see this effect  at the endpoint energy.  A photon, which bremsstrahlung electron  interacts wit the magnet pole, will likely not make it through the  collimator, but you'll get some background.
 +
****You may consider to slightly increase the collimator diameter (you don't need linear polarization (?)).  If there is a problem with the AC hole size, you can run with the beam profiler.
 +
****All in all, I believe that this would not be easy to do, and will  require some studies/optimizations.
 +
***AD reply to Sasah: Thank you Sasha. Yes, it looks like it would be more work and time than it is worth. Also, such run would be invasive to the other halls, so it's probably not worth pursuing.  A bigger collimator hole would certainly help and not using the Active Collimator should not be any issue since I don't think we care about the photon beam profile (we are interested only in circular polarization). But event though, it would not be enough to help with the factor 100 loss in transmission.
 +
 +
*JS: Justin reminded us of a consideration for the GlueX endorsement: [http://www.gluex.org/GlueX/Bylaws.html by-law D.1] All proponents of such proposed experiments must be members of the GlueX Collaboration, or have a plan approved by the EG and CB for joining the Collaboration and participating in ongoing GlueX experiments. Justin believe this is required at the endorsement stage, so we should consider how the authors listed on our proposal would consider joining the collaboration or have a plan for that at the time of the endorsement vote. As a consequence, AD removed for now the non-GlueX authors from the proposal (except Simon Širca and C. Keith) and will ask them about such commitment. Simon will write a statement of commitment if the GlueX in-house review of the proposal is successful.

Latest revision as of 17:17, 31 March 2020

Present: M.D., A.D., S.Š, J.S.

  • General:
  • SŠ: Coded the unpolarized and polarized Bethe-Heitler formulae, see his report
    • AD's comments on Simon's document:
      • BH asymmetries seem very small. For GDH, we are looking at asymmetries at the 3% level (see e.g. Helbing review's Fig. 38: Δσ 10 μb, with A=Δσ/(2σ₀). This points toward the possibility that we can ignore the BH for the proposal.
      • Regarding the structure functions for the asymmetry, the atomic form factors should not be needed since the atoms are not polarized.
      • For g1 and g2, probably only their values at very small-x that are truly relevant. If so, one can use a simple Regge parameterization of g1, see e.g. arXiv:1808.03202. For g2, we could just assume g2ww(x,Q²)=-g1(x,Q²)+ ∫x¹g₁(y,Q²)/y dy.
      • For question #2, the angle coverage, we are planing to use the Compton Calorimeter, which cover down to 0.2°. (Note: we do not have the ComCal in the simulation yet).
  • AD:
    • Found out how to write Simon's initial properly: Š
    • Discussed with Richard Jones and Sasha Somov how easy it would be running in Hall D at 0.5 pass. The bottomline is: very difficult because of the smaller Lorentz boost and the long distance between the radiator and the collimator. Here are the details:
      • Richard Jones: It is going to be difficult to tag that beam with our tagger, for two reasons.
        1. only a very small fraction of the photon beam (rough estimate of 0.03%) will make it through the collimator, while the tagger has to count the entire beam. This means that accidentals will be high even at very modest photon beam intensity at the target.
        2. the height of the "stripe" of post-bremsstrahlung electrons at the tagger focal plane will be an order of magnitude larger. This will lead to substantial scraping on the poles of the tagger magnet, and only a fraction of them actually being detected by the tagger. This needs to be simulated, but it might be that the quadrupole can be used to help mitigate this. Those electrons that are detected in the tagging counters without scattering from the poles will have an enhanced probability of having their photons pass the collimator, which may partially offset issue #1.
        • Bottom line is that it may work but not easily. The biggest irreducible issue will be the substantial scattering from the poles of the tagger by post-bremsstrahlung electrons. If the showers from these interactions dominate the tagger rate, it will be difficult to use the tagger at all for this measurement. If that is the case, can you make a meaningful measurement without tagging?
        • Why not consider an alternative: run at the full energy, and tag the low-energy part of the spectrum. You need to build a custom low-energy tagger, which might be a diamond strip detector like they used in the Compton in hall c. You could place it in the vacuum near the exit edge of the tagger magnet on a movable ladder, and sweep it around to cover different parts of the low-energy spectrum. These photons would be fully collimated, and very clean because the electron beam current would be turned way down to keep the low-energy flux at a level compatible with tagging. This sounds like something I would be interested in working on, provided that the physics of the low-energy part of the measurement justifies the effort.
      • Abstract of AD email exchanges with Richard:
        • AD: Regarding 1, why does so little beam go through the collimator, and how does it compare with the usual situation? Is that because of larger multiple scattering at low energy? Since we plan to use the thin Al. radiator which is about 20 times thiner than a standard diamond, then the multiple scattering should be only twice worst (~10/sqrt(20)). Is that enough to degrade significantly the transmission? Probably I am missing the reason for the low transmission.
          • RJ: It is not due to multiple scattering, the effect is the same for a 1e-15 radlen radiator. It is the bremstrahlung cone that scales with m/E.
          • AD: Ah, it's the boost! Then it's a factor ~10^2 lower transmission compared to the nominal energy, is that right?
          • RJ:It is a factor 10 in the cone angle, so the transmission factor goes down by a factor 100, because of it scales like area.
        • AD: Regarding 2, I should have mentioned that we plan to run with both low current (50nA) and the thin Al radiator, which makes a factor ~300 less in photon flux/post-bremsstrahlung electrons. Is this making point 2) not a problem?
          • RJ:No, lowering the rate does not help at all, if 99.97% of the hits in the tagger are background not coming with a collimated photon tagging is hopeless, right?
          • AD: Right. Now that I understand point 1), I get this one too. Thank you.
        • AD: We thought about having a lower energy tagger as you suggest, especially with the offer from Lund to re-use their small tagger (which seems to have now disappeared). But we thought it was more trouble than it was worth: These low energy data have already been measured at MAINZ, so for us, taking these data again is just a matter of having an overlap with a very well measured domain. ****AD:It seems this is too much trouble for the benefit. This would have been nice because
          1. it offers a good check of our absolute calibration and polarimetries and
          2. It allows to fully cover the energy span to form the GDH integral.
        • RJ: I still think you might consider mounting a diamond strip detector on a movable ladder and making a low-energy tagger. If you built it, I can imagine a people in the N* community might think of things to do with it. Imagine a low energy widely variable flux tagged beam with 100% polarization, and very large (factor >30) coherent enhancement.
        • AD: If we decide to build a low energy tagger we don't need to discuss it especially in the proposal: since it will operate concurrently no additional special beam time is required. One setback of this option for the GDH proposal (beside having to build/commission a new detector) is that it provides very low circular beam polarization, about 5% in the region of interest.
        • RJ: Yes, if it is circular polarization you need then generating low-energy photons from high-energy electrons is not what you want.
      • Sacha Somov:
        • You can easily simulate this run condition using gxtwist.
        • The problem at small beam energies is the bremsstrahlung angle (m/E), as you have already discussed (I believe that the quad would not help). The rate of accidentals in the tagger has to be considered.
        • The magnet acts as a collimator (the gap is 3 cm); we see this effect at the endpoint energy. A photon, which bremsstrahlung electron interacts wit the magnet pole, will likely not make it through the collimator, but you'll get some background.
        • You may consider to slightly increase the collimator diameter (you don't need linear polarization (?)). If there is a problem with the AC hole size, you can run with the beam profiler.
        • All in all, I believe that this would not be easy to do, and will require some studies/optimizations.
      • AD reply to Sasah: Thank you Sasha. Yes, it looks like it would be more work and time than it is worth. Also, such run would be invasive to the other halls, so it's probably not worth pursuing. A bigger collimator hole would certainly help and not using the Active Collimator should not be any issue since I don't think we care about the photon beam profile (we are interested only in circular polarization). But event though, it would not be enough to help with the factor 100 loss in transmission.
  • JS: Justin reminded us of a consideration for the GlueX endorsement: by-law D.1 All proponents of such proposed experiments must be members of the GlueX Collaboration, or have a plan approved by the EG and CB for joining the Collaboration and participating in ongoing GlueX experiments. Justin believe this is required at the endorsement stage, so we should consider how the authors listed on our proposal would consider joining the collaboration or have a plan for that at the time of the endorsement vote. As a consequence, AD removed for now the non-GlueX authors from the proposal (except Simon Širca and C. Keith) and will ask them about such commitment. Simon will write a statement of commitment if the GlueX in-house review of the proposal is successful.