Comments from Colorado (Kevin Stenson)

This is the institional report from University of Colorado Boulder, completed by Kevin Stenson. The comments are divided below into Type B (physics related, figure related, or major text changes) and Type A (editorial changes).

Type B:

L53: You state that CT-PPS detects protons scattered at angles corresponding to 97% or less of the original momentum. There must also be a limit on the other side. Shouldn't this also be included? This also appears in L265-268.

Changed L53 to "scattered at small angles and carrying between about 84 to 97\% of the incoming beam momentum. We did not repeat this in L265-268, since the Standard Model l+l- signal is expected to be negligible at the lower limit of xi, so the upper limit is what determines how much of the cross section is visible.

L66-67: It would be useful to include more information about these detectors. What is the physical size? What is the pitch? Is the readout analog or digital? Are the detectors capable of resolving hits from individual bunch crossings? What is the hit efficiency? What is the acceptance? Which side of the beamline is covered (just the inside, I believe)?

The strip detectors themselves have been used in many previous TOTEM analyses and there are more details in reference 7. However we have added the following to the description, excerpted from that reference:

"In total each RP contains 512 individual strips, with a pitch of 66 $\mu$m. The hit efficiency per plane is estimated to be $>97 \%$ before irradiation of the sensors. The signal from the silicon detectors is contained within one 25 ns bunch crossing of the LHC. The data are read out using a digital VFAT chip~\cite{Kaplon:2005ce}, and recorded through the standard CMS data acquisition system."

@article{Kaplon:2005ce, author = ``Kaplon, J. and Dabrowski, W.'', title = ``{Fast CMOS binary front end for silicon strip detectors at LHC experiments}'', journal = ``IEEE Trans. Nucl. Sci.'', volume = ``52'', year = ``2005'', pages = ``2713-2720'', doi = ``10.1109/TNS.2005.862826'', SLACcitation = ``%%CITATION = IETNA,52,2713;%%'' } %ENDCOLOR

L56-57, L75-78: Why discuss the timing detectors if they are not used? I would suggest removing these sentences. L78-9: Why discuss a future configuration that is not relevant to the analysis?

Part of the motivation for this paper is to describe the alignment and optics calibration procedures used for forward protons in low-beta/high-luminosity conditions. These will be relevant for any type of detector used in CT-PPS, including the timing and 3D pixels. A large amount of data has already been collected with the timing+pixel detectors in 2017, so we think it would be sub-optimal to repeat everything for the analyses that will use those detectors.

There needs to be some reordering of the paper. L83-84 and Figure 3 talk about "track impact points", which is not defined and there is no explanation about why we would care about this. This section should be moved to Section 4.2 after the proton track reconstruction is described.

Done for now, but this means we need a forward reference to "track impact points" when discussing the optics corrections. TBD if this is acceptable

Figure 3 caption: Need to clarify "but slightly reduces the low x acceptance". Can you write "but slightly reduces the acceptance for x<6\unitmm x < 6 \unit m m " or something similar? It may be useful to change "at 15σ 15 σ from the beam" to "at 15σ 15 σ from the beam in the x x direction" to make it clear that this is why there are no events with x smaller than ~4mm.

Changed, merging this and other comments: "Example of track impact point distribution (in arbitrary units) measured in RP 210F, sector 45, at 15~$\sigma$ from the beam in the $x$ direction. The beam center is at $x=y=0$. The track selection includes a matching requirement with RP 210N, which results in a more pure sample, but slightly reduces the acceptance for low values of the position $x$." We haven't added exact values, since the effect will not be identical for the two arms, and only one is shown as an example.

L133-137: I do not understand this procedure at all.

Tried to completely rewrite this: "The symmetry of the track impact point distribution in $y$ is exploited to determine the vertical position of the beam with respect to the sensors. This is possible since, unlike the horizontal case, the maximum $y$ values of the distribution are visible in both the $y > 0$ and $y < 0$ directions. In practice a mild linear dependence of the mean $y$ position on $x$ is observed, consistent with a non-zero value of the vertical dispersion, $D_{y}$. This effect is accounted for by performing a linear fit of the $y$ coordinate of the maximum of the track impact point distribution, as a function of $x$. The fitted function is then extrapolated to $x = 0$ (horizontal beam position) to determine the alignment in $y$. Here again, the resulting uncertainty is of order 150~$\mu$m, arising from the fit extrapolation uncertainty, the horizontal alignment uncertainty, and the uncertainty from varying the fit range."

L140-142: You mention "track segments measured in the RPs" but this has not been described. Are these "track segments" the same as the "tracks" in Section 4.2? If so, then please forward reference Section 4.2. If not, please describe briefly what a track segment is.

It seems the "segments" is no longer really needed here, so it is removed for now. Previously there was an expanded section on the alignment, that discussed both tracks fitted in a single RP, and tracks reconstructed in all RPs of one arm. In the current analysis the reconstruction is only used RP by RP.%ENCOLOR%

L151: In reading this, I was surprised to find no dependence on Theta^*_x. This is mentioned in L178-179 but that is pretty far away. It may be useful to mention here that the dependence on Theta^*_x is subleading.

Added: "The dependence on the horizontal scattering angle enters only as a sub-leading term in the transport matrix."

L166-167: I don't think relegating the entire proton track reconstruction to a thesis is appropriate. I would suggest expanding this section. I have no idea what "linear patterns" means? It is also not clear what is meant by "track impact points".

We have slightly reworded this as below. We have tried to clarify that "linear patterns" simply means a pattern of hits along the z-axis consistent with a straight line, since there is no magnetic field to bend the tracks in this region. "Track impact points" are already defined in the same sentence where the phrase is first used, as the track position evaluated at the center of the RP along the z axis. We're open to adding more details from Ref. [8], if there are other specific things that are unclear.

"Since there is no significant magnetic field in the region of the CT-PPS RPs, real tracks follow a linear trajectory through the silicon strip detectors. In each RP, track reconstruction therefor starts with a search for linear patterns along $z$ among the hits detected in the 10 planes, as described in Chapter 3 of Ref.~\cite{Kaspar:2011eva}. The search is performed independently in each of the two strip orientations (with angles of $+45^{\circ}$ and $-45^{\circ}$ with respect to the bottom of the RP); hits in at least 3 out of 5 planes are required. If only one pattern is found in both orientations, the patterns can be uniquely associated and a track fitted, yielding a ``track impact point" evaluated at the center of the RP along $z$. Figure~\ref{fig:hit_distribution} shows a typical distribution of the track impact points (i.e. the points where the reconstructed tracks intersect a detector plane) in a RP at 15~$\sigma$. When there is more than one pattern in any strip orientation, a unique association is not possible and no track is reconstructed. The inefficiency due to multiple tracks depends on the pileup, and ranges between 15 and 40\% over the course of the 2016 data taking period."$

L175: You write "track impact points" here. Is there really more than one track impact point in a single RP for a given proton? I would think there is only one so it would be "track impact point".

Yes, removed the "s"

L184: Need to specify what "high" means.

Changed to "For values of xi \simeq􏰄 0.04, the leading uncertainty comes from the dispersion"

Section 5: You should move L207-209 to Section 5 as they help define the Data set and also L209 is not about Central variables, which is the subsection it is in.

We would prefer to leave this - the trigger is based only on central (muon) variables, and is the most basic step of the event selection. It's true that the HLT is used to define the Primary Dataset used in the analysis, but this is internal CMS jargon and not what general readers would necessarily understand as being related to "Data sets and Monte Carlo samples"

L207-208: Need to specify if the pT requirement is on each lepton or on the dilepton system. Also, should indicate if there is a requirement that the two leptons have opposite charge or not.

Changed to "Events were selected online by requiring the presence of at least two muon candidates of any charge, each with transverse momentum $\pt>38$~GeV, or at least two electron candidates of any charge, each with $\pt>33$~GeV. No requirement on the proton was imposed online."

L210-214: Should rewrite to make clearer. I assume the two leptons must be of the same flavor. Is this requirement placed before or after the selection of the two highest pT leptons. Is the opposite charge requirement placed after selecting the two highest pT leptons?

Modified to read: "Offline, the tracks of the two highest-$\pt$ lepton candidates of the same flavor in the event are fitted to a common vertex. The vertex position from the fit is required to be consistent with that of a collision ($|z| <15$~cm), with a $\chi^{2} < 10$ (probability greated than $0.16\%$ for 1 degree of freedom). The lepton candidates are further required to each have $\pt > 50~\GeV$, and to pass the standard CMS quality criteria~\cite{Chatrchyan:2012xi,Khachatryan:2015hwa}. In the final stage of the analysis the leptons are required to also have opposite charge."

L219: Not exactly clear what the "veto distance" is. Is it the distance between the dilepton vertex and the distance of closest approach of the track in question?

Yes, exactly - given that other reviewers did not seem to comment on this, we prefer to keep the current version that describes this in 2 words instead of 16.

L227: I would suggest changing "looser" to "smaller". If it is not smaller then I think you need to reevaluate your logic.

OK, yes it is smaller, changed.

Also, why did you not reweight the MC to reflect the actual primary vertex distribution?

Simply reweighting based on the z distribution of our signal vertex does not help, since all of the simulated pileup vertices in the event will still follow the wrong distribution, giving the wrong pileup density, and therefore the wrong probability for a nearby track to be mis-associated to the signal vertex. We would either need to reproduce the minbias MC with a different vertex distribution before the pileup mixing step, or somehow reweight based on the true distance between the signal vertex and all other vertices in the event. The former isn't practical for the large central DY MC production, and the latter would (to our knowledge) need more information than is available in the AODSIM.

L233-248: I do not understand the point of this paragraph. Is this a result? Is this needed for the analysis? If so, it would be good to explain why this is relevant and how it will be applied to the analysis. A better explanation of "rapidity gap survival probabilities" would be helpful.

It's not strictly needed for the analysis since we don't quote a cross section, but the ARC/RC had several concerns about the agreement between data and the out of the box MC in the central region. So as requested we performed several studies on the effect of modifying the MC with more recent theoretical survival probability predictions, and included a summary of these in the paper. As explained starting on Line 240, it is mainly there to explain a possible reason for the tendency of the MC to over-estimate the data, particularly for lepton pairs at forward rapidities, as shown in the Figure.

L235-237: This is unclear. Is it correct to say that all MC samples are first normalized to the luminosity and then the LPAIR samples are further normalized by multiplying by the rapidity gap survival probabilities? If so, then change "For the \textsc{LPAIR}" to "In addition, for the \textsc{LPAIR}"

Yes it's correct, changed as suggested

L236: Looking at reference 26, I see numbers of 0.89, 0.76, and 0.18 rather than 0.89, 0.76, and 0.13. Please check and correct the paper or analysis if necessary.

The proton dissociation value in Table 1 of reference 26 is calculated separately for "incoherent" and "evolution" terms. The values we quote in the paper were suggested by the authors (L. Harland-Lang) as the best estimate to apply for the physical events produced by the Monte Carlo generator. We have added a sentence explicitly saying this We considered including a "priv. communication" citation for this, but did not following advice from the ARC.

L251-252: It would be useful to give some idea of where this formula comes from. Also, can't you simply change η(l+) η ( l + ) to |η(l+)| | η ( l + ) | and remove L252?

It's just energy-momentum conservation in the limit that the proton p_z >> p_x, p_y. We can expand on this if needed. As mentioned on L252 there are 2 possible (different) solutions corresponding to the two protons, so we don't think it's possible to drop the +/- and just use the absolute value of |eta| (which would only give one solution).

L263: I don't know what is meant by "observed in inclusive proton distributions"

We mean we take the sample of all events passing our dilepton trigger, which is dominated by accidental pileup coincidences, that yield a very large sample of protons from pileup interactions in the same BX. We have modified the sentence to read: "...observed in an inclusive sample of dilepton triggered events, with no selection to enhance $\gamma\gamma$ production, is used.". Other suggestions are welcome.

L265-268: Why is there a difference between Sector 45 and Sector 56? Should this information be added to the paper?

The optics/dispersion were significantly different between beam 1 and beam 2 in 2016, leading to different acceptance on the two arms. We've added a sentence saying this explicitly.

L292-3: Does "simulated LPAIR events" mean "simulated double dissociation LPAIR events" here? Or is it all LPAIR events? Please ensure the text is clarified.

Yes - changed as suggested

L334-335: There should be an explanation for why a 100% uncertainty is appropriate.

We have clarified that there are no direct measurements of this process, hence the large uncertainty

L354-355: It is not clear to me what is meant by this sentence.

Specified - "The invariant masses and rapidities of the candidate $\mu^{+}\mu^{-}}$ events are consistent..."

L382-383: It may be useful to know what the overlap is between these two events and the two highest mass events (which may also be background)

There is no overlap - a sentence has been added: "The vertical component of the scattering angle for the two highest-mass $e^{+}e^{-}$ events is compatible with zero."

L386-7: It would be useful to specify how many events there were

Modified to read: "Of the 20 total events selected, 13 have a track in both the near and far RPs. In these events, the two independent $\xi$ measurements agree within $4\%$."

L391-393: It seems like it would be useful to show a figure with the acoplanarity distribution with the fits to the elastic and single dissociation contributions.

We have these figures are in the AN - we did not include these in the paper because we have not fully evaluated the fit systematics that would be needed to present these as a physics result. Based on discussions with ARC/RC and input from other reviewers, we removed the discussion of this fit.

Results: It seems that the analysis does not try to compare the total amount of single proton dissociation events with whatever is expected from LPAIR simulations or some other theoretical calculation. Is there any reason for this? It would require some absolute normalization that may add some additional complexity. Is that why it wasn't done?

Yes, the main issue (mentioned on L74, L192, and L371) is that the strips suffered heavy radiation damage throughout 2016, mainly in the region closest to the LHC beam. They then partly recovered during periods with no beam, and after raising the high voltage. This leads to a low and strongly time-dependent efficiency variation at low x, and therefore low xi. There are also other pileup-related inefficiencies related to the extra tracks veto in the central tracker, and the reconstruction of multiple protons in the RPs. For BSM searches the low xi region may be cut away to work in the region away from the radiation damage, but in our case we tried to go as low as possible in xi to keep some statistics for the SM process. We have done some internal checks that the number of dimuon events for xi>0.05 is close to the expectations after including the largest inefficiencies; following input from the ARC and other reviewers we have added text saying this, without giving a quantitative cross section, since not all of the systematics and subheading corrections have been fully evaluated

Figure 2: Much of the text is quite small and nearly impossible to read when printed.

Tentative cleanup of the figure, increased labels sizes, removed sizings for unused pots, ...

Figure 5: The Greek symbol in the figure does not look much like what is in the text.

It appears to us to be a xi. We can investigate with the creator of the plot if it's possible to use a different font.

Figure 7: I read the last sentence in the caption but I still do not understand why the y-range is different. If there are events with acoplanarity below the minimum y value, are they simply not plotted? It seems like it would be easier to make the y-ranges the same and remove the last sentence of the caption.

This was the result of a long discussion during the PAG review. Now it's replotted with the same y-axis range (within resolution of both channels). The last sentence of the caption has been removed.

Figure 7: Most of the text is too small. To save space, you could plot the legend just once.

Changed font size, legend now plotted just once.

Figure 8: It may be better to have a logarithmic y axis for the plots versus mass

We've tried this, but find it mainly emphasizes the DY background, while obscuring the relative contribution of the different elastic and dissociation gamma-gamma processes. We therefore propose to keep the linear version in the paper, but will add the log version to the AN

Figure 8: It would be better to have the ratio plots directly below the main plot with no gap and only one set of x-axis values and labels. This saves space.


Figure 8: For the top plots, the bins are 50 GeV wide and go from 0 to 800 GeV. This means there is a bin from 100-150 GeV. However, there is a cut at 110 GeV so it is really just showing the events from 110-150 GeV. This doesn't seem desirable. Perhaps it would be better to use 55 GeV bins from 0 to 770 GeV to avoid this.


Figure 8: The y-axis label should be "Events / X" even though there are no units.


Figures 8,9,10,11: The minus signs are too small as they are hyphens. One way to get a properly sized minus sign in root is to replace "-" with "{#font[122]{\55}}"


Figures 10,11: The top text of the plots should include the luminosity.


Figure 11: I'm not sure you need the text "LHC Run-II, pre-TS2"

This label has been suppressed from the figure, as these technical details are already mentioned in section 5.


Type A:

You need to standardize on whether to include the speed-of-light or not. I see that masses are given as GeV /c^2 but momenta are given as GeV.

We've switched back to natural units everywhere, pending any objections from the ARC+RC

In CMS papers particle symbols (like p for proton, e for electron) are not italicized. You should use the appropriate pen names.

pen names implemented

I don't think writing "Si" instead of "silicon" is useful. Also, I don't think "strips" should ever be used as a noun. I think it should be replaced by "strip detector"

Don't really understand this comment - "strips" is certainly a plural noun. However, we have replaced "Si" with "silicon" everywhere. We have also modified the text to use "strip detector" whenever we refer to the detector as a whole, but keep "strips" for the detector description, where we discuss the properties of the individual silcon strips.

When writing something with the symbol sigma like 15sigma and 4.6sigma, you sometimes have a space between the number and the symbol (usually when referring to how close the RP is to the beam) and sometimes there is no space (usually when referring to significance of the observation). Please be consistent.

We have tried to make this uniform (with a space before the sigma).

I have always seen the symbol for rapidity as "y y ". You seem to be using a capital Y or an Upsilon.

Changed to "y" everywhere.

The CMS pub guidelines frown on the use of "toy" arguing it is jargon.

Hm, "Monte Carlo" is probably also too jargon-y, and just "simulation" may be misleading since elsewhere we use this to refer to the full GEANT4 simulation. We have changed to "fast parametric simulation" following other CWR comments.

Abstract: Should remove the sentence "The spectrometer and its operation are described, along with the data and the background estimation" as it is too detailed and uninformative.

The abstract has been modified following other comments - the sentence is no longer there.

Abstract: Should remove the sentence "They also demonstrate that CT-PPS performs as expected" as it does not really add anything

We're not sure this is obvious, given that Roman Pot detectors have never been operated in high luminosity hadron colliders before. For ARC/RC/LE

L2: Suggest changing "The present paper" to "This paper"


L6: Remove "at"


L16: Change "single proton tagged" to "single tagged proton"


L19: Change "rightmost" to "right"


L19: Should you change "proton" to "particle" since we don't actually know what type of particle we are detecting? It could be a pion for example.

OK, changed to: "The right diagram is considered background if a proton from the diffractive dissociation is detected, or if a particle from another interaction in the same bunch crossing (pileup) or from beam-induced background is wrongly associated to the dilepton system."

L21: Change "dimuon" to "dilepton"


L27: Suggest "at the Fermilab Tevatron and CERN LHC, but"

OK, changed to avoid confusion with other Tevatrons and LHCs

L33: "_The_ CT-PPS"

Changed to "The CT-PPS detectors"

L34: "of the LHC"

Yes - changed

L74: Change "underwent" to "suffered"


Figure 2 caaption: "along with the timing RPs not used here."


Figure 2 caption: Replace the 2nd-to-last sentence with "The red (blue) arrow indicates the outgoing (incoming) beam."


L87: Change "as well as" to "and"


L117: Serial comma after "shifts"


L119: Remove "in principle"


L122-123: The way this sentence is worded, it sounds like the difference between the physics fills alignment and the alignment fill alignment is that the former is data-driven (implying the other is not). Suggest simply removing "; it is data-drizen for both horizontal and vertical alignment" but you may have another solution.

OK, done, and merged "a different procedure is thus used" into the preceding sentence.

L126: Serial comma after "currents"

Serial comma added

L196-199: It would be better to remove the word "signal" and combine these two sentences into one sentence. For example "Simulated samples of elastic L201:... ..Twiki makes a mess of the symbols in the rest of this sentence... see the rest of the comment in CDS

Change "Consequently, only" to "Only"

"Consequently" removed

L212: Change "P =" to "probability greater than"


L213: Change "required to have" to "required to each have" (assuming this is accurate)

Correct - changed

L214: Change "tracks" to "leptons"

Right - changed

L250: Suggest "proton track in CT-PPS are"

Changed as suggested

L254: Add "LPAIR" in front of "simulated events" (assuming that is correct)


L259: Add "Results from" in front of "LPAIR" and change "indicates" to "indicate"

Changed as suggested

L276: Change "dissociative backgrounds" to "double dissociative backgrounds" or "double proton dissociative backgrounds"

Changed - "double dissociative backgrounds"

L291: Could change "More in detail, the" to "The"

Changed as suggested

L319-323: I would suggest rewording this to start with the total and then follow with the numbers inside 2sigma. This is more natural and directly matches the numbers in the table.

Changed as suggested

Tables 1-4: Change the font size to be the same as the paper, combine into 1 or 2 tables total, and put the caption above the table. Also, switch the column order of 2sigma and full and the explanation in the caption. You could also consider putting the Tables at the end of the section and including the systematic uncertainties.

Also, remove all lines that define the outside border.

Done except for the font size, since this would conflict with previous complaints that the tables are too small when the font size is the same as the text. For simplicity, we've left the tables in the current position, with the statistical errors only.

L333: Change "dimuon" to "dilepton"


L350: Remove "(cf. Section 7)" as it is not needed


L361: Remove "discussed in Section 7" as it is not needed


L389-390: "assumed to be correlated"


L390: "between the two"


L404: "over the expected background"


The following references have issues: 4,7,13,14,15,18,19,21,24,26,27,28,29. Here is what needs to be fixed:

  • Volume letters should be part of the title and not part of the volume number
  • There should not be ending page numbers
  • There should not be issue numbers


Edit | Attach | Watch | Print version | History: r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r1 - 2018-01-31 - JonathanHollar
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Sandbox All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2023 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback