Meeting No. 23

ISR paper:

Message from the editor:

 Re: LZ15192 
    First measurement of proton's charge form factor at very low Q 2 with
    initial state radiation
    by M. Mihovilovi\v{c}, A. B. Weber, P. Achenbach, et al.

 Dear Dr. Merkel,

 The above manuscript has been reviewed by our referees. A critique
 drawn from the reports appears below. On this basis, we judge that
 while the work probably warrants publication in some form, it does not
 meet the Physical Review Letters criteria of impact, innovation, and
 interest.

 The paper, with revision as appropriate, might be suitable for
 publication in Physical Review. If you submit the paper to Physical
 Review, the editors of that journal will make the decision on
 publication of the paper, and may seek further review; however, our
 complete file will be available.

 If you submit this manuscript or a revision of it to Physical Review,
 be sure to respond to all referee comments and cite the code number
 assigned to the paper to facilitate transfer of the records.

 If you feel that you can overcome or refute the criticism, you may
 resubmit to Physical Review Letters. Please accompany any resubmittal
 by a summary of the changes made and a brief response to all
 recommendations and criticisms.

 Yours sincerely,

 Kevin Dusling
 Associate Editor
 Physical Review Letters
 Email: prl@aps.org
 http://journals.aps.org/prl/

 IMPORTANT: Editorial "Review Changes"
 http://journals.aps.org/prl/edannounce/PhysRevLett.111.180001

 ----------------------------------------------------------------------
 Report of Referee A -- LZ15192/Mihovilovic
 ----------------------------------------------------------------------

 The letter reports on precise measurements of the coincidence elastic
 scattering H(e,e'p) process. In particular, the experiment measures
 the radiative tail of the process, where a real photon is generated.
 The experiment is innovative -- for decades, the radiative tail and
 our uncertainty of the tail has been a curse for experimentalists. The
 effort to derive knowledge of the proton structure from the tail is
 certainly welcome.

 However the authors themselves conclude that the experiment is unable
 to distinguish between the CODATA and muonic hydrogen radii. That
 marginalizes the physics impact of this work.

 I would recommend that the paper be expanded substantially, and
 published in a PRC article. In particular, before this paper can be
 published it ought to include quantifiable description and analysis of
 the contribution from nitrogen and oxygen elastic scattering to these
 spectra. Reference 15 also states that the Havar foils gave
 substantial contributions, and mentions a "havar line". The
 contribution does need quantified, but Havar is a complicated mix of
 Co, Cr, Ni, W, Mo, Mn and even a little Fe. This experiment has
 wonderful resolution and very precise systematics -- does the Havar
 really contribute only a single line, or many?

 At low beam currents and energies, the spectrometer A had to be used
 for relative luminosity measurements. What was the uncertainty? The
 absolute luminosity was left to float (amounting to a .1 - .5%
 correction). It needs to be explained why that is not a single
 absolute number for all 3 beam energies.

 The authors and the lab are known for their incredibly precise work.
 Publishing how this was all done would be of great service. For
 example, Figure 3 seems to show a systematic upward shift of this
 experiment compared to the world average. Is that correct? Does it
 indicate a normalization issue?

 Finally, a brief discussion early in the paper as to how the
 experimenter can distinguish Born-i,f from BH-i,f would also be
 helpful to the reader (especially if this is to go to a more general
 audience).

 ----------------------------------------------------------------------
 Report of Referee B -- LZ15192/Mihovilovic
 ----------------------------------------------------------------------

 Excellent paper, and well written. The topic of the proton charge
 radius is very topical, and the disagreement of the electronic
 (electron scattering and hydrogen Lamb shift) charge radius
 measurements with the muonic measurements is a complete mystery right
 now. The ISR technique provides a technique to go to very low Q2,
 which is necessary to accurately measure the proton RMS radius. I
 believe this paper will be of broad interest to the community. For the
 uninitiated reader, it would be useful to quote in the paper the
 radius determinations from the muonic hydrogen Lamb shift
 measurements, so the reader understands the level of the discrepancy.
 This paper has not provided the definitive measurement of the proton
 RMS charge radius through the ISR technique. However, it has certainly
 shown a possible path forward to reconcile the electron and muon
 measurements, or point to a problem with the muon Lamb shift
 measurements. I strongly encourage PRL to publish this paper.

My answers to the comments would be:

Dear Editor.

Thank you for your answer. We examined the two reviews carefully. We can not agree with assessment of the the first referee on our result being marginal, which then led to your negative critique of the paper.

The measurement of the proton charge form-factor gives us the insight into the structure of the proton. How well we can describe the nucleon depends on the range and the precision of the available form-factor data. The accessible range on both (low Q**2 and high Q**2) sides is limited by the available experimental equipment.

In the content of the problem of the proton radius, the data at very low Q**2 have became interesting, because they are most important for the determination of the radius. Unfortunately, by using conventional experimental approaches and the existing equipment, the data could be collected only at Q^2 higher than 0.004GeV**2.

In our paper we report on a innovative new experimental approach, which, using the same experimental equipment, which offers access to the proton charge form factor at even smaller Q**2. We performed the pilot experiment of this type, demonstrated that approach works and for the first time established that the existing theoretical description of the radiative processes, which constitute the radiative tail, are understood to a sub-percent level. Although these processes have impact on all the experiments that study electron included reactions, no one before us has yet investigated how precise the available description of these processes really is. As an example: these corrections influence all the existing form-factor measurements in a level of 10-20%. The high level of consistency between the theory and the experiment that we observed in the regions even more than 200MeV away from the elastic line, demonstrates that radiative results indeed could be trusted and reinsures previous results. Our efforts in showing this and the importance of these finding has been recognised by both referees.

After establishing, that the approach works, we also provided new measurement of the form factors with a competitive precision in the Q**2 range between 0.001 and 0.004, where no prior measurements existed. We believe, that these new data should be of interest to the whole community.

For the purpose of the completeness, the obtained form factors were then fitted with a model and proton radius was extracted. As written in the paper and emphasised by the referees, the obtained value is not capable of distinguishing between the CODATA value and the muonic result. However, it should be recognised that, due to the small lever arm that data have on the radius in the vicinity of G(0) = 1, it is very difficult to obtain a precise enough value for the proton radius, using only low Q**2 data, even if the data have sub-percent precision like ours or even better. Furthermore, even if the radius would be determined with a good enough precision, it could be systematically off due to the missing higher Q**2 data, that determine higher momenta of the fit. Hence, to get a reliable estimate of the radius the data we provided need to be combined with the existing world collection. Our low Q**2 data, will improve the absolute normalisation of the fit, while the higher Q**2 data will fix the first (radius) and higher momenta of the fit (please see I.Sick 2017). However, we believe that this last step is outside the scope of this paper, whose goal was to provide independent determination of the form-factor in the yet unmeasured regime.

We would also like to answer specific questions on the analysis:

1.) Referee I was concerned about the background originating from the Havar foils. Havar is indeed a complex mixture of different metal elements. Since all these elements are very heavy, they have tiny recoil corrections. Hence, their elastic lines can no longer be distinguished, in spite of the 1E-4 relative momentum resolution of the spectrometers. However, the elastic lines can be accompanied by the peaks of the transition states, where a particular element in a Havar is excited to a higher state. In the experiment we collected empty cell data for all the kinematics in order to estimate such backgrounds. As shown in Ref. 15 such excited states from Fe, Co and Cr could be observed in spectrometer A, which was set to a reasonably high Q**2 setting. However, the Spectrometer B, employed for the measurement of the cross-section, was always positioned at the very low Q**2 setting. There, the cross-section for the transition states is suppressed and in the spectra we detected no measurable presence of these excited states. Therefore, we could estimate the backgrounds to the hydrogen spectra originating from the walls by using a simulation, that considers only elastic states. The simulation does not consider Havar as a one peak, but considers all the elements separately. Furthermore, the simulation was preferred before the empty-cell data, because it properly considers the energy-losses in the Hydrogen between the walls, which was not present during the empty-cell measurements.

2.) Referee I also asked about the uncertainty of the luminosity: As summarised in the section about the systematic uncertainties, was the uncertainty of the relative luminosity determination estimated to 0.17%. For each energy setting was the data taking divided into 8-10 runs. Hence, the relative uncertainty cold be determined by comparing the ratios of the number of events in Spectrometer A with the number of events in B for each of the runs.

For the absolute calibration of the luminosity a reference point would be required. One option would be an elastic point. Since these data (elastic points) were omitted from the analysis, we decided to normalise the data to the average of all the points, as shown in Fig. 2. This served us as a first guesses of the normalisations for each energy setting. The final absolute normalisations were then determined from the fit. Since the (normalisation) measurements were taken at different energies and different angles, different cuts were applied to the Spec. A data to isolate good events, and since these data also depend on the applied FF model, it is right to expect that the tree normalisation factors are different. The small deviations of the final normalisation factors from one (ranging from 0.1% to 0.5%) only demonstrate, that our first guess with the average were already very good.

3.) Regarding the potential normalisation issue, that referee 1 observed in the data: When looking at the data in Fig. 3 one indeed may get a feeling that the low Q**2 points have a larger slope, which then flattens at around Q**2 =4E-3. Similarly one can see a dip at Q**2 = 0.0018, consisting of two points from different kinematics (and different beam energies). Therefore, we investigated, what kind of modifications in our analysis would be necessary, to bring either the two points at 0.0018GeV**2 up or the overshooting points down. We realised, that the modifications to the backgrounds would be unrealistically big. The points we show, are robust. Since we have no model, that could explain the fluctuation, we decided to fit the data through the middle.

4.) Regarding the question on how one can distinguish Born-i,f from the BH-i,f: The new experimental approach that we are presenting in the paper relies on the idea of exploiting the BH-i (shown info Fig. 1) to reach the information on the proton form factors. All other diagrams represent unwanted “background”. The largest contribution comes from the BH-f. The contributions of the Born diagrams (Born-i, f) are much smaller due to 1/m**2 suppression factor.

However, in the experiment, where we detect only final electron, we can not distinguish between any of these processes. We are sensitive only to the square of their combined amplitudes. Our analysis is therefore based on the comparison of the data to the simulation that includes as many different diagrams (processes) as possible. To confirm, that approach works, we first investigated the high Q**2 part of the data, where form factors have been measured before. Once we know, that approach works, we determine the form factors by finding parameterisation of the form factors in the simulation, that matches the cross section data best. As explained in the paper, at this point we assume, that the whole difference between the data and the simulation (see Fig. 2) is due to the difference between the simulated and the true FF. All other effects are considered in the estimation of the systematic uncertainty.

In conclusion, we believe, that our manuscript reports on a innovative new experimental approach based on the initial state radiation. With the results shown, we demonstrated for the first time, that radiative corrections are understood to a sub-percent level. This result influences the whole electron-scattering community but is interesting also for the field of particle physics, which uses similar approach to study e+e- reactions. The paper also provides new measurements of the proton charge form factor in a region that is not accessible with conventional methods. The obtained values have a statistical uncertainty comparable with past experiments, but with a systematic uncertainty superior to any previous elastic measurement. These successes were clearly recognised by both referees.

The only criticism we could recognise in the response from the referees, is related to obtained value of the proton charge radius and our inability to distinguish between the CODATA value and the muonic-hydrogen result. Therefore we would like to emphasise one more time, that this is not due to inaccuracy of our results, but due to the fact that all the measurements are close to one. Using only our data gives us only little strength to determine the slope of the form factor. Furthermore, a precise fit would be difficult even if the data would have an unprecedented precision. To determine the radius, our data should be combined with the data at higher Q**2, which is outside the scope of this work. Hence, we believe that the limited precision of the radius, should not be the reason of the negative evaluation of our paper. Therefore, we would like to ask you to reconsider your decision on renouncing the publication of our manuscript in the Phys. Rev. Lett.

New scattering chamber and new entrance flange:

We should discuss how to build the scattering chamber and He-filled snout for the next experiment. The old tech. drawings should help:

z.-nr. 1086 - 1 anschlußkammer 1 spektrometer b.tif

z.-nr. 1087 - 1 anschlusskammer 2 3 spektr b.tif

z.-nr. 1109 - 1 streukammer.tif

New experiment:

We are planning to experiments in February. The one-weekend experiment and the full 12C(e,e'p) experiment. I have designed a new target ladder, that can support 9 targets. For the full beamtime we need: 12C-stack, Pb, Al2O3, 12C-thin. This leaves 5 places for our experiment. I would suggest 4 Lupolen targets of different thicknesses and one additional carbon target with a larger thickness.

1.)

The experiment is motivated by the discrepency we see between the simulated and measured height of the tail. The addvantage of the Lupolen is, that the peak is broad and that we have the theory under control. I would measure at 495MeV. There othe optics works best and the field is stable! The Foerster probe works there. The problem is, that the spectra are contaminated with Carbon excited states. However it seems that we have this under control. 2.)

3.)

4.)

Thoughts of Jan Friedrich

I discussed with Jan Friedrich this problem last year and he suggested, that this effect could be replaced by e-e- scattering. With this we should be able to reproduce the tail.

Optics NIM paper

I am working on the NIM paper. I started generating plots for the paper. Here are the plots I created so far. I have gridfinder plot, the importance plot of various matrix elements and first achieved resolution. I would like to use this NIM also to explain the problem with the NMR in spectrometer B.