Comments to NIM_paper_part2_v1.pdf dated 19/6/2024 20:49 ======================================================== Jan Timmermans 18 November 2024 Apologies for being a bit late after my promisse in last week's meeting. But the number of proposed corrections is quite large. line 1: performance of -> particle identification with ->Done line 15: for a 1 m track -> per 1 m of track length ->Done line 17: proposal for change of "the template fit" is "successive hit distance fit" or some other name, but I do not insist -> added "using the successive distances between the hits" line 22: "resolution", what is exactly meant by this? -> For the abstract "resolution" is fine IMO, in the paper details are given. line 23: using circular tracks -> using full circular tracks (as all tracks in B=1 T data are circular) -> not all circles are "full" contained in the detector lines 24/25: not clear (yet) what you mean with 1) "incident angle"? At which point of the full circular track? -> phi in the xy plane of the circle and orientation wrt the detector fixed 2) "within an uncertainty of 16 um"? Is this the single point resolution? in the ILD experiment -> it is the uncertainty on the resolution: maximum deviation of the constant value as a function of phi. line 26: of -> for ->Done line 27: in ILD -> in the proposed ILD experiment at a future ILC e+e- collider ->Done line 36: of 55x55 -> 256 x 256 pixels of 55 x 55 ->Done line 37: ionisation vs. ionizing (in line 51): make consistent spelling ->Done lines 43-44: paper [3]. The paper -> paper [3], which ->Done line 46: REMOVE, as I propose the following sub-sections to sections with the appropriate numbering ->Done line 53: in the ILD experiment -> in the proposed ILD experiment at ILC ->Done line 55: 2.1 -> 2. ->Done It seems you have (not yet) defined the quantity "dE/dx" as a measure of the energy loss of the beam particle (i.e. the number of track hits per chip?) -> Added "Particles can be identified by their characteristic energy loss per unit of track length $dE/dx$ loss and/or the number of primary clusters $dN/dx$ produced a long the track. In a GridPix detector one can measure the number of hits produced along the track and their relative distance." lines 58-64: Are there any (further) selections/cuts on these (combined) long tracks, i.e. (e.g.) angle, curvature, ...? -> "In order to measure the track performance of $dE/dx$ or $dN/dx$, a track selection was applied selecting tracks crossing the central chips - defined in ref \cite{Beuzekom:2024part1}. The individual chips were calibrated to give the same mean number of hits per chip." line 59: hits is -> hits per chip were ->Done line 60: are -> were ->Done line 63: to the track -> to a track from several events ->Done line 67-68: what is "consecutive pixels" here exactly: a) associated to and along the track? or... b) inside a box of e.g. 5x4 or 5x 10? line 68: pixels -> pixels along the track -> Added two lines "Both methods project the hits along the track in the $xy$ plane. This gives a distribution of hits as a function of the distance along the track in pixels. The first method rejects large multi-electron clusters with more than in total 6 hits in 5 consecutive pixel bins .." line 71: what is "minimum" here? projected along the track? -> the distances between all hits can be calculated but only the minimum is used. Text is now: "The second method exploits the distribution of the minimum distance between consecutive hits in the $xy$ precision plane." line 76: by the -> by the traversing beam -> Done line 81: At lower distance -> At lower distance, -> Done line 83: per 1 m track -> per 1 m of track length -> Done to distance -> to the distance -> Done formula (1): slope d_{xy} -> slope_d_{xy} -> ?? change notation? ADD here a figure with an example distance distribution for a single track and template fit superimposed -> We cannot show an individual fit ... likelihood page 4, Table 1: 1) make columns left-aligned 2) line "Method": remove resolution ( 2 times) 3) remove the line with the two single "%" symbols 4) line "1" and line "2": remove these method numbers 5) add the "%" symbols after the result numbers -> Done line 88: provides -> provide ->Done lines 92-93: should it be LINEAR? -> No one can correct for non-linearities; if it is linear no correction has to be applied Figure 1: 1) make (a bit) larger, i.e. use full horizontal space -> Done 2) inside figure: make larger axes numbers, titles and captions 3) right plot axis title: the fitted slope is given in units of [1/pixels] -> 2-3 I don't know. In a journal you can zoom. Figures are large and separate. This is NEW unit and not described in the text. Ofcourse, if you accept suggestion to make a figure with an example distance distribution, that can be explained there that its horizontal axis is in 'pixels' and not in um -> The unit of the fitted slope is inverse (pixel) distance, as is clear from the formula in Eq.~\ref{eq:expo}. 4) in caption: and MIP 1 m long -> and for a MIP of 1 m long -> Done line 114: 2.2 -> 3. -> Done line 116: threshold -> threshold (TOT) lines 119-121, i.e. "The track selection ...performance." -> -> Rephrased as: "The track selection is the same as the one that was described in section 2. However, there is not a real description of the track selection there -> The track selection is the same as in section 2. It is described in paper + section 2 "central chips". lines 121-122: the run numbers are meaningless (they are also not mentioned in the part I paper. Just call them '6 GeV' and '5 GeV' runs and maybe add the number of analysed 'events' or better 'tracks' -> If one wants to reproduce the results run numbers are relevant to quote. line 123: for values -> in the interval -> Done line 126: time over threshold -> TOT (and also in rest of paper) -> Done line 128: half -> half (in x) -> Done line 130: "by the total run time" - do you take the duty cycle of the machine into account? As these are different for the 5 and 6 GeV runs. -> No it is just the data taking time. So it is NOT the max instantanuous rate. lines 130-131: 6934 and 6935 -> at 5 GeV -> Added 5 GeV/c lines 133-134: remove "due to the high statistics' -> Done Table 2: 1) title: time OF threshold -> TOT -> Done 2) in 2nd line of table, put units in [..] -> Well: now it is clear these are then units; 10^3 s (now) OR [ 10^3 s] or 10^3 [s] 3) as mentioned above, replace run number by beam energy -> See answer/discussion abopve 4) number of triggers for line '6935' must be similar to the previous line -> Corrected: indeed 7.34 k trigger are in run 6934 (in stead of 73.4) 5) should mention somewhere that our actual beam trigger rate is HIGHER because of the inefficiency of our 2-fold scintillator coincidence OR correct the numbers for this -> The exact values of instantanuous beam rate and trigger efficiencies, don't go in the raw hit rate of the chips (that is measured trigger less). Figure 2: 1) what is the definition of "event" here? I see only green and blue hits -> An event is defined by the trigger number (so all hits after the trigger till the next trigger are show). Indeed I used in the histogram toa = t_trigger - t hit. Instead of toa = t_hit t_trigger. Plot should be remade. 2) in caption: run 6969 event 2 in a -> a 5 GeV beam energy run at also : xy plane in color -> xy plane; in colour, -> Rephrased as: "An event display for run 6969 event 2 taken at a 5 GeV/c beam momentum in a $B$ = 1 T field. The hits are shown in the $xy$ plane in colour the time of arrival is shown." line 148: 2.3 -> 4 lines 150-151: in run 6969 with -> at 6 GeV beam momentum and -> "An example event in run 6969 is shown in figure" line 151: remove 'field' -> Still I give the information in the caption: -> "An event display for run 6969 event 2 taken at a 5 GeV/c beam momentum in a $B$ = 1 T field" line 152: energetic particles ???? ALL are very low energy w.r.t. the beam energy so say here the order of magnitude of the momentum of these (half) circles -> Rephrased: "A large variety of hit patterns can be observed: large radii (open) circles, smaller size radius circles from low momentum particles, curlers and more confined bursts." -> Added to text "A track with a momentum of 1 MeV/c will have a typical radius of 121 pixels." IF possible: add an artificial circle that just fits between top and bottom and give the momentum of this example track. Ofcourse that can also be mentioned with one line in the main text -> Done with one line see above line 157: run 6969 -> a run at was -> were line 157 and following AND figure 3: in re-reading this paragraph carefully now, I have quite a bit of remarks/questions to the content and the figure 3: first of all, please make the line thickness of the histograms MUCH thinner -> Will be done as you remark further on in line 170, a radius of 50 pixels around the burst center corresponds to an area of 7854 pixels. So requiring >100 hits in such an area means you look at an occupancy from only ~1%. So that is not (yet) relly a "burst" -> That is correct. In the core of the distribution one may have a large occupancy. -> That is why I give characteristics for > 100 > 200 and > 400 hits in a burst. Is there a "burst" present in figure 2? e.g. at x,y ~ 600,850 ? If so, would be nice to add to the figure a small "arrow" pointing to that "burst" and add few words in figure caption -> I don't think it helps ... bursts with more than 100 hits? Figure 3: 1) reduce histograms line thickness (a lot) 2) reduce the number of hor. axis scale numbers, e.g. 0, 10, 20, etc. and 0, 40, 80 etc. 3) hor. axes labels: radius 90 -> radius90 and time 90 -> time90 -> 1)-3) Will be done 4) in caption: for run 6969 event 2 in a -> for a 5 GeV beam energy run at -> Changed line 169: might stay -> may remain -> Done line 170: run 6969 -> the run used in figure 3 -> Changed lines 172-176: 1) I wonder whether it would be better to move this to the next section? -> NO this why burst finding - using the characterics show in this paper - IS relevant for tracking. Also: what do you mean by "One could" in line 174? Is it "done" or "not done"? -> The pattern recognition for ILD should be changed (so One could) -> Rephrased: "In the pattern recognition one could run a burst finding algorithm and down weight in the track fit the hits associated to bursts." line 177: 2.4 Resolution study -> 5. Track resolution as function of track angle -> Done (I proposed in an earlier meeting to leave this section out: I am still in favour of this, but do not insist) line 180: incident angle -> incident angle of the track, -> Done line 181: is minimal if the incident track angle is parallel to the strip -> is best for track angles parallel to the pad strip -> Done line 182: remove "This is due ...layout." -> Done line 185: at non-zero B field ALL tracks are circular (unless they interact). So better say "full circle" tracks ? (2 times) The needs more text: "For circles, the incident $\phi$ angle of the track depends on the position of the individual hits on the circle in the $xy$ plane. The range of $\phi$ angles depends on the radius. For radii smaller than 500 pixels a large $\phi$ range can be probed. Using the residuals in the $xy$ plane, it is possible to measure the resolution of the hits as a function of the incident $\phi$ angle of the track." (by the way, this study could have been done also with "beam" tracks at different angles of the detector) -> That is correct. BUT ... in that case we need may to be able to rotate the detector up to +/- pi/2 ... and that is not trivial line 190: 1) again, what is the definition of "event" here? -> See above defined by the trigger 2) circles large clusters -> circles, large clusters -> Done 3) what is meant by "down weighted"? Should be more explicit. Then you can end sentence and start new sentence with "Hits within 15 pixels ... ->Will add a sentence lines 191-192: 1) this is all very ARBITRARY 2) 4 pixels -> 4 pixels (220 um) -> No it matches the approximate resolution -> Rephrase: "In the circle fit, the resolution in $xy$ was estimated to be about 4 pixels and in $z$ it was 1 mm." line 194: dof -> d.o.f. ( 2 times) -> Done line 196: 0.4 MeV/c -> 0.8 MeV/c (please check my number!) -> I think 0.4 is OK ... A track with a momentum of 1 MeV/c will have a typical radius of 121 pixels. line 197: pi/4 to 7pi/4 corresponds to 0.785 to 5.498 but in figure 4 the plotted range is from ~ 0.5 to 5.7 -> That is right should be: -> "The $phi$ hits below $\pi/8$ and above $15\pi/8$ were removed." line 200: which "resolution" do you refer to? Finally, the resolution in $xy$ was extracted - using a Gaussian fit to the track residuals in the range of $\pm 2 \sigma$ around the centre. line 201: I just don't understand this statement, because if the above mentioned "resolution" is single point xy resolution, then I don't see why it should depend on the radius? Rephrased this: " Because the resolution depends on the radius (i.e. the momentum) and small radii span a large phi range, the data was re-weighted as a function of the circle radius." line 201: data was -> data were lines 202-203: strictly speaking this sentence is un-understandable, "by using a Gaussian fit". A fit to what? I assume it is again the fit of the residuals to the expression that was given in the part I paper, but is not obvious here -> Indeed. Rephrased "Finally, the resolution in $xy$ was extracted - using a Gaussian fit to the track residuals in the range of $\pm 2 \sigma$ around the centre." line 203: be consistent with the spelling of centre vs center -> Done line 210: change title to: 6. Projected particle identification performance for a Pixel-TPC in the proposed experiment ILD at a future ILC Also: I think it would be good to have this section following the (renumbered) section 2. line 211-212: remove this first sentence -> Done line 212: ILD -> ILD [4] -> Done -> "The TPC of the proposed ILD detector \cite{Behnke:2013}" line 213 (middle) - 215: make this as first sentence in this section 6, but modifying the first part: The electron resolution from the test beam -> The particle ID for electrons in the test beam -> Redone as proposed: " The particle identification performance of electrons in the test beam for momenta of 5-6 GeV/c was measured to 2.9\% for the template fit and 3.6\% for the $dE/dx$ truncation method at $B$ = 1 T for 1 m long tracks with 60\% coverage. The TPC of the proposed ILD detector \cite{Behnke:2013} has an inner radius of 329 mm, an outer radius of 1770 mm and a half length of 2350 mm. In the ILD TPC this will correspond to an expected electron resolution of 2.4\% (fit) and 3\% (truncation) at polar a angles of $\theta =\pi/2$ ($\cos\theta=0$) and a track length ($\rm tlength_0$) of 1441 mm." page 10, Figure 4: 1) vertical axis lable: sigma xy -> \sigma_xy 2) caption line 2: a) curved -> curve -> Done b) Eq. 3 -> Eq. (3) -> NOT Done paper has Eq. everywhere line 215: for a 1 m long -> for 1 m long -> Done lines 216-217: I am afraid that the tracklength at 90 degr. is shorter than 1441; there will be space needed for 'services' on the inside (both on inner and outer) -> Could be but this a calculation lines 219-220: you say "similar or worse" but this is NOT true. If we run T2K at driftvoltage of 280 V/cm (maximum of drift velocity) the transverse diffusion is not optimal, but at B=3.5 T, the diffusion constant is ~ 30_40 um/sqrt(cm) But maybe you mean in line 219 "the actual amount of transverse diffusion" in the beam test and in ILD? Please improve wording. -> A more accurate sentence would be "Depending on the B field value (3.5 T - foreseen at ILC - or e.g. 2 T), the resolution in the transverse plane after drifting to the endplate will be similar or worse than the test beam situation. The resolution for a GridPix TPC in ILD will therefore lie between 2.4 and 3\% at $\cos\theta=0$." -> I took out the sentence because it discusses something complex: the impact of the xy resolution ("diffusion process") on the dEdx performance... Basically, the range we give is correct. line 220, 221 and 225: the resolution -> the PID resolution -> Done equation (4): \sigma = -> \sigma_i = -> Done line 225-226: largest track length .... 0.85. -> largest track length, which corresponds to cos\theta = 0.85 in ILD. -> Done line 227: for the energy loss -> of the energy loss -> Done line 228: were used [6] -> were used as given in [6] -> Done page 11, Figure 5: 1) vertical axis label: (sigma) -> [$\sigma$)] -> Still to do 2) caption: line 1: projected separation -> projected PID separation line 2: line corresponds -> lines correspond line 2: electron resolution -> electron PID resolution -> Done line 230: The separation -> the PID separations line 230: w.r.t. pions is -> w.r.t. pions are line 235: do NOT start a new line -> All Done line 249: 1 m track -> 1 m long track -> Done line 251: impressive. It is currently, the best -> impressive, and is currently the best line 259: circular -> full circle -> Rephrased "using reconstructed circle tracks." line 268: CEPC collider a Pixel TPC -> proposed CEPC collider in China, a Pixel TPC (and ADD a reference) line 271: at the EIC -> at the Electron Ion Collider (EIC) at Brookhaven National Lab (and ADD reference)a line 271: DRD1 collaboration (add reference) DRD1 Extended R&D Proposal, Colaleo, Anna, https://cds.cern.ch/record/2885937 -> I don't agree that we can quote meaning references for these items. The are NO official publications: - CEPC no publihsed detector TDR (CDR is old) - EIC on published detector TDR (status "white paper") - DRD1 has no publication either .... ยง