[b1_ana] b1 higher twist

[Dustin Keller]

Hi b1,
Axx is the very best possible tensor polarized observable that we can 
extract being it is available without negative polarization.  But just a 
reminder about the uncertainty, the range of systematic uncertainty in Axx 
would be around the green point (middle point) in the plot I showed.  The 
uncertainty in the observable we are discussing now would be a combination 
of the uncertainty in Axx and the unpolarized cross section.  This is true 
for both statistical and systematic.  So if we are able to find additional 
points of interest at say elastic or something else then we may as well 
use b1 as additional motivation.  Otherwise the overall concern has not 
changed.

dustin




On Fri, 19 Apr 2013, Oscar Rondon-Aramayo wrote:

> Hi Simonetta,
>
> Thank you for checking that Jaffe's formulas refer to the electron beam. It
> would be more difficult to measure things with the field along the q vector.
>
> We don't need to do two separate measurements (in the sense of measuring
> first one cross section and then the other). We can have two target cells in
> the beam at the same time, one polarized, the other unpolarized. They would
> essentially have the same acceptance. Corrections for the residual
> difference in acceptances can be investigated following the method used 25
> years ago by the EMC experiment that found the proton spin crisis.
>
> The EMC had to take data on two targets of opposite polarizations, because
> they could not flip the helicity of their muon beam. Nevertheless, they
> managed to control the systematic effects of their acceptances, etc. A
> discussion of their method is posted on my b1 page
> http://twist.phys.virginia.edu/~or/b1/EMC-piegaia_thesis.pdf
>
> The method I suggested is based on the EMC approach,
> http://twist.phys.virginia.edu/~or/b1/b1_method.pdf
>
> with the difference  being one cup polarized, the other not, and alternating
> them during the run, just like the EMC did. Narbe has already done
> simulations that show we can cleanly separate the events from each cup, see
> http://twist.phys.virginia.edu/~or/b1/zbeam2.eps
> (He has many more examples, I'll send links to his plots folder later)
>
> So, my view is that I think the method can achieve the necessary statistical
> precision, and I'm not yet convinced that the systematics cannot be
> controlled to the corresponding precision.
>
> For example, in the outline of the method I proposed, the \delta pf that is
> introduced before eq. (10) is a relative fraction, should really be dpf/pf.
> So in eq. (11), the contribution of the dilution factor is suppressed by the
> relative error on the packing fraction, which is about 4% for current
> experiments (SANE, RSS) and can be made smaller by using targets shaped as
> disks, instead of irregular fragments, etc. For an error in f itself ~
> 0.05*f (RSS achieved 0.047*f), and f=0.3, we have a systematic error on b1
> due to f and pf of  0.05*0.3*0.04 = 6E-4. Remember that f and pf are proxies
> for the target thickness and unpolarized contributions.
>
> The same kind of relative errors are involved in the propagation of the
> differences in charge and acceptance. We need to estimate them
> conservatively, following the EMC approach, before reaching conclusions.
>
> Finally, there is the method of taking the ratio sigma_pol/sigma_unpol
> discussed in Anklin's and Boeglin's proposal, which could be an alternative
> to the EMC method, and could be directly applied to the data taken
> simultaneously on polarized+unpolarized cups, as I propose.
>
> Cheers,
>
> Oscar
>
> On Thu, 18 Apr 2013 18:21:54 -0400
>  "Simonetta  Liuti" <sl4y at cms.mail.virginia.edu> wrote:
>> Hi All,
>>
>> even if higher twists are negligible, I am skeptical of the Jaffe et al.
>> formulae for extracting  b_1 (Chapter 6).
>>
>> In fact, I checked that the observable is OK in the sense that they are
>> referring to the electron beam (and not the virtual photon).
>>
>> However, I expect systematic errors to be very big. The "polarized cross
>> sections" observables can be obtained with two separate measurements. A
>> polarized spin 1 target one, and an unpolarized one. The two measurements
>> have to be carried out separately, just to be clear. And this is going to
>> impact the systematics.
>> This is exactly why Hermes came up with A_zz which only involves target
>> polarization.
>> I do not see the advantage of going back to Section 6 of Jaffe et al.,
>> while one could do something new and exciting  measuring deuteron DVCS
>> asymmetries.
>>
>> However...please let me know what you think.
>> Simonetta
>>
>> On Thu, 18 Apr 2013 16:06:07 -0400
>> "O. A. Rondon" <or at virginia.edu> wrote:
>>> Hi,
>>>
>>> There is a paper by Hoodbhoy, Jaffe and Sather, which says that higher
>>> twist (twist-4) effects on b1 at Q^2 = 1 GeV^2 are only 5%. So they
>>> would seem negligible at the kinematics of our proposal or HERMES. See
>>>
>>> http://dx.doi.org/10.1103/PhysRevD.43.3071
>>>
>>> Cheers,
>>>
>>> Oscar
>>> _______________________________________________
>>> b1_ana mailing list
>>> b1_ana at jlab.org
>>> https://mailman.jlab.org/mailman/listinfo/b1_ana
>>
>> **************************************************************
>> **************************************************************
>> Prof. Simonetta Liuti                     telephone (434) 982-2087
>> Department of Physics               FAX       (434) 924-4576
>> University of Virginia              home      (434) 973 9593
>> 382 McCormick Rd.
>> PO Box 400714
>> Charlottesville, VA 22904-4714
>>
>
> _______________________________________________
> b1_ana mailing list
> b1_ana at jlab.org
> https://mailman.jlab.org/mailman/listinfo/b1_ana
>