[Clascomment] OPT-IN: Cross sections for the exclusive photon electroproduction on the proton and Generalized Parton Distributions

[Hyon-Suk JO]

Hi Stepan,

Thanks for your comments.

The revised version of the paper can be found here:
https://www.jlab.org/Hall-B/shifts/admin/paper_reviews/2015/dvcs_cross_sections_prl_draft_v4-5413120-2015-03-31-v4.pdf

> - page 2, line 105, the CLAS calorimeter polar angular range for photons is 8 to 45, it became 15-45 due to IC. The sentence should be corrected, to clarify that

On lines 104-108, we replaced "The polar-angle acceptance
of CLAS for photons (~15-45 deg) was extended down to 5 deg
with the addition of a specially designed electromagnetic
calorimeter ("inner calorimeter", IC [13]), with full
azimuthal coverage." with "A specially designed
electromagnetic calorimeter ("inner calorimeter", IC [13])
was added to the CLAS detector and allowed the detection
of photons for polar angles from about 5 deg to 16 deg,
with full azimuthal coverage". We will simply omit the
polar-angle coverage of the EC as we discarded the photons
detected in the EC from our analysis.

> - page 2, line 144, sentence "In order ... we chose to minimize the size of our bins, ...". Can you give the value of minimized bin size the resolution of the detector for that variable

As the unpolarized cross section varies strongly as a
function of xB, we defined most of our bins in xB with a
size of 0.03 or 0.04. Our average resolution in xB is 0.005
while our average resolution in Q2 is 0.035 GeV^2.

> - page 3, line 184, "... of 12.3%, meant to compensate for various effects that 185 are not well reproduced by the simulations". This means it either comes from efficiency calculation or radiative effects. Both effects will be very different for elastic events. is not it. How elastic with very different kinematics can be used for this arbitrary correction

We agree with you that this correction is probably the
weakest point of this analysis. However, it was agreed
within the e1-dvcs group to adopt it consistently in all
the cross section publications [see for instance
I. Bedlinskiy et al., Phys. Rev. C 90, 025205 (2014)]. This
factor accounts for problems in the efficiency evaluation,
and not in radiative corrections (as the elastic factor
was computed from the ratio of the measured _radiated_
elastic cross section compared to Cole Smith's calculation
of its theoretical value). So, efficiency problems for the
double-arm elastic channel should be common to the "ep" part
of the DVCS channel. It is true that the kinematics of the
two reactions are different, and that there is not much
overlap for either of the two particles. Nonetheless, it
was the only well-known reaction that we could use to cross
check and correct our normalization. This factor includes
the efficiency of the SC counters which was estimated to
be around around 95%, as well as other efficiency factors
which are not accounted for in the Monte Carlo, such as
trigger efficiency. We quote a big systematic on this
factor, among the biggest ones that we have, to account
for the limitations of this correction. On lines 183-188,
we replaced "In addition, we applied a global
renormalization factor of 12.3%, meant to compensate for
various effects that are not well reproduced by the
simulations. This factor was determined from the analysis
of the elastic scattering ep->e'p', comparing the
experimental cross section to the well-known theoretical
one." with "In addition, we applied a global
renormalization factor of 12.3%, determined from the
analysis of the elastic scattering ep->e'p', by comparing
the experimental cross section to the well-known
theoretical one. This factor compensates for various
kinematic-independent inefficiencies, not well reproduced
by the simulations.".

> - one of interesting results from Hall-A measurement was Q2 independence of the Im-part of the Compton FF, that was attributed to twist-2 dominance. It seems we have much more data to say something with respect to this important question. Why we are not giving Q2 dependence? Is it because there is not enough statistics in any of x-bins to have few Q2 points?

Because of the small size of our bins in xB, we had to
adopt larger bins in theta_e (for enough statistics), only
leading to two Q2 values for each bin in xB. This is indeed
the reason why we could not extract a Q2 dependence similar
to what was done in the Hall A paper.

Cheers,
Hyon-Suk