Dear Yuma,

I have read the draft and could not find any major issues in it. Indeed, I read the draft is perfect.

I write a couple comments just for your considerations.

L118: 6 / 32 = 187ns. The meaning of the symbol ÷ varies depending on the language.

L137: How did you obtain the distributions shown in Fig.4? You may want to add more information about the data, collision data or MC. You may also want to add the way to distinguish the signal hits from off-time background hits.

L220: What is the time window for the (time-) integration of the radiation dose? The normalization area of the radiation dose is not clear to me; is the listed value, 70 krad, normalized to the area of one L3 sensor?

Best,
Takeo

Dear Uematsu-san,
Please find below some language/style comments. I thought the structure was good.
All the best,
Jim

Line 57: collision -> collisions
Line 58: Delete ‘The’
Line 58: consists of -> consists of the following components:
Line 58: dumping -> damping
Line 61: mass-resonance -> resonance
Line 62: high statistics -> large samples
Line 63: none of the hyphens are required for layer-$n$ here
Line 64: 4-GeV -> 4 GeV
Line 64: for the -> for
Line 65: measurement -> measurements
Line 65: integrated -> an integrated
Line 68: rewording of last sentence: The data accumulated before July 2021 corresponds to an integrated luminosity of 213 fb$^{-1}$.
Line 70 and 71: no hyphens
Line 74: Detailed ..appear -> A detailed description of the SVD appears
Line 85: in case -> in the case of
Line 85: Ks decay -> Ks mesons, which decay
Line 87: energy deposit dE/dx -> ionisation energy deposits. [dE/dx is not used again]
Line 89: no need for capitalisation in defining acronyms, e.g., Double -> double
Line 90: $X_0$ -> of a radiation length
Line 91: readout Aluminium -> aluminium readout
Line 98: layer 456 -> layer 4, 5, and 6 (twice)
Line 98: forward/slanted region -> forward region, which is slanted.
Line 102: here and elsewhere use a ~ between value and unit to ensure they appear on the same line so in this instance 60~V.
Line 103: , 20~V -> and 20~V
Line 103: , and -> ;
Table 1 caption: taking -> taken
Line 107: for CMS -> for the CMS
Line 107: APV25 -> The APV25
Line 107: for over -> for a dose up to
Line 109: APV25 -> the APV25
Line 112: each sampled information in the -> each sample’s information in an
Line 114: comma before in
Line 115: to operate -> operation
Line 117: successive six -> six successive
Line 119: which accepts -> to accommodate
Line 123: APV25 -> the APV25
Line 135: the smaller -> smaller
Line 136: the smaller. To reduce -> reduced. To minimize
Line 138: the single -> a single
Line 138: read out the signals -> read out of the signals is
Figure 2 caption line 1: The plot explains the sampling -> Sampling
Line 140: 700W->700~W
Line 149: was gone after cable reconnection -> which was remediated by reconnecting a cable
Line 151: stably -> continuously
Line 154: losses -> is lost
Line 156: SNR -> signal-to-noise ratio
Line 163: well agrees -> agrees
Figure 3 caption: Layer -> layer (twice)
Figure 3: space before opening parentheses e.g. green (blue)
Line 168: the hadrons event data -> hadron-event data
Line 170: of SVD -> of the SVD
Figure 4 caption: The plot shows an example -> Example
Line 187 and 189: samples -> sample’s
Line 217: damages->damage
Line 255: SVD -> The SVD
Line 257: in higher luminosity -> during higher luminosity running
Reference general: no comma before et al., volume numbers should be in bold, same format for journal abbrevation, i.e., [3] and [4].
Line 271: (2010). -> (2010), (also [3])
Line 290: babar -> BaBar
Figures 6 and 7 caption: The effect -> Effect
Figure 6 caption: forward region -> forward region (Micron) [to match legend]

Hello Yuma,

very nice paper, I do not have any major comment. Most of my comments were already pointed out by Jim, so I just drop them.
The remaining ones are reported below.

-- main comments:

line 84-90: from the point of view of the importance, I suggest to move the standalone tracking & PID capabilities before the Ks reconstruction.

line 122: the 6-sample time window is effectively (6-1)/ 32 MHz = 156 ns and not 6/32 MHz = 187 ns because the last sample closes the interval.

line 199: the first of the subset of 3 samples depends on the relative shift and on the trigger bin, so the selected 3 samples are not in fixed positions. What is fixed is the latency of the first of the 3 samples with respect to the trigger arrival. Maybe you can change it into "by selecting consecutive three samples at a fixed latency with respect to the L1 trigger signal"

lines 207-210: we probably want to keep it as it is, but the loosening by a factor 2 on the occupancy is optimistic

-- minor comments (text typos/suggestions)

line 42: operating -> operated? Jim did not pointed this out, so I maybe wrong.

line 107: can consider adding "approximately" in "approximately 224,000 strips"

line 150: problems -> problem (again, Jim did not pointed this out, so I maybe wrong.)

line 172: "in candidate hadronic event"?

-- Figures & Tables

Figure 1: layer3 and PXD have the same color, but I guess you can't change the picture.

Table 1: "Table of the dimensions for the three sensor types"

Figure 2: "Example of sampling..."

Figure 4 caption: the black distribution represents background, but not all background is off-time. I suggest to remove "off-time".

Figure 5: you can consider to write explicitly that it's the 3-sample divided by the 6-sample efficiency in the caption.

Dear Uematsu-san and Giulia,

Just following up on the language queries. First 'operating' is correct as the SVD is still working inshallah. When it no longer is we should move to operated.

Second, this is a grey area and both are in common usage as they are describing hypothetical situations: a single problem or many problems that didn't happen. However, re-reading the sentence I actually think this clause is redundant as it conveys the same information as 'smoothly and reliably', i.e., a major problem would prevent you using these adjectives, so in the interests of concise prose it could be dropped.

All the best,

Jim

Dear Yuma,

thanks a lot for the very nice and clear paper.

Below my comments that you might consider to improve a bit the text in some points.
It's a long list but most of them are really minor things.

There were a few things that are more relevant related to the :
- resolution plot esplanation
- limit on rad damage for svd sensors worth to be added


ciao,

Giuliana



Fig 1
Make it bigger if possible
In this picture SVD is in RED and PXD in light blue, but with the colored box for the different sensor types, especially Layer 3 is NOT really well visible. Can you use a different color for the BOX of L3 sensor NOT similar to PXD?
Caption: change it adding this …. In the upper half of the VXD the locations of the three types of SVD DSSDs are indicated by boxes in three colors:…

Line 77
Remove. Beside the VXD … starts with
Diamond sensors [3], used to monitor the radiation dose and for the beam abort system, are mounted on the IP beam pipe and the bellows pipes outside of the VXD…

Line 79
Remove : “They measure the dose rates in these locations.”

Line 80: à The diamond’s measured doses are used to estimate the dose in the SVD.

Line 81: à The diamond system also send

Line 86: à … limiting the PXD readout data volume

Line 87: critical à crucial

Line 93 :
Remove this sentence from here, not clear at all. Same info added later in the text.
The aluminum readout strips are AC-coupled to every other n/p-side strips (electrodes) on the n-type substrate over the silicon oxide layer.

You want to say 2 things here:
-the strips are AC coupled to what??? You can mention this later in line 109 when you start to talk about the APV25
-there is a floating strip. And this can go simply in the caption of table 1 as I proposed below.

Line 99: mention that small and large sensors are rectangular
“small” rectangular sensors in layer 3, “large” rectangular sensors in the barrel region of layers 4, 5, and 6, …..

Line 101 blue box à other color you used for Layer 3 sensor if you change the fig 1 as proposed.



Line 102: Not only dimensions but also pitches etc are summarized . Change to:
The main characteristics of these three types of sensors are summarized in Tab. 1.

Table 1
Mention in the caption that there is a floating strip in between 2 readout strip
“….All sensors have one intermediate floating strip between two readout strips. “

Line 109.
Add a sentence for AC coupled strips.
Sensor strips are AC coupled to the front-end ASIC, the APV25 [5] which was originally developed for the CMS silicon tracker.

Line 122:
.. offers a large enough time window

Line 141 : change a bit to make it clear:

APV25s chips are mounted on a single side of the sensor and readout of the signals from the opposite side is performed via wrapped flexible printed circuits.


Line 144 :
The chips are cooled by a bi-phase –20â—¦ C CO2 evaporative cooling system.

Line 155: continuously à stably

Line 158:
the sentence is not clear. the reason of the charge loss on n side is not that the inter-strip capacitance of the floating strips with large pitch is significantly smaller than on p side….The effect is due to the presence of the floating strip, but the interstrip capacitance in p and n side are not that different. The loss originates if these interstrip capacitance becomes comparable with the capacitance of the strip to the back side, that for small pitch is negligible. This last one scales with the pitch/2 (strip implant pitch) so the capacitance to the back is a factor 3 higher on n side than on p side. When the capacitance to the back side become larger (as in n side) and no longer negligible w.r.t. the interstrip capacitance we start to see the loss.

This is too much to be explained in the text but the sentence should be corrected a bit:

On the v/n-side, 10–30% of the collected charge is lost compared to MIP due to the smaller inter-strip capacitance of the floating strips with larger strip pitches than the u/p-side.
à On the v/n-side, 10–30% of the collected charge is lost compared to signal collected on the u/p side, due to the presence of the floating strip combined with the large pitch on the n side.



Line 164: insert some commas.
The cluster position resolution is estimated from the residual between the cluster position and the track position, not biased by the target cluster, after subtracting the effect of the track extrapolation error.

Line 160
On the resolution the comment on the plots is not completely correct especially for L3.

The behaviour of the resolution vs incident angle is: a minimum for 2 strip clusters (when the the projection of the track length on the strip is = pitch , since we have the floating), but after this angle it would also start to deteriorate.

Now as written in the SVD paper (page 171 of V1 ) I report also below, for the very small pitch of u side the minimum resolution is expected at very small angle: it is 4-7 degree for L3-L456. So I would expect the minimum close to normal incidence and then some deterioration of the resolution, which is not very visible, since I think we see still some effect of the track estrapolation error not well subtracted that : increase a bit the resolution also at normal incidence, and diluted a bit the behaviour with angle.

So my point here is that the “main” issue in our current plot for resolution especially in L3 u side is NOT that we don’t observe a reduction of the resolution of with angle in L3 Uside, as it’s written now, but probably that we can still improve the U side resolution ….


FROM SVD PAPER
“ The cluster position resolution as a function of the track incident angle is shown in figure for data and simulated events. The observed resolution has the expected shape, showing a minimum at the incident angle for which the projection of the track along the direction perpendicular to the strips on the detector plane corresponds to two strip pitches. Given the various sensor pitches with one floating strip, the minimum is expected at 4 (7) degrees for the u/P side and at 14 (21) degrees on the v/N side, for layer 3 (4, 5, 6). Considering the various sensor pitches with the floating strip (table 1), the expected digital resolution for perpendicular track is 7 (11) ðœ‡m on u/P side and 23 (35) ðœ‡m on v/N side for layer 3 (4,5,6), similar to the measured resolutions.”


My proposal would be something like this:

The observed resolution has the expected shape, showing a minimum at the incident angle for which the projection of the track along the direction perpendicular to the strips on the detector plane corresponds to two strip pitches.
Given the various sensor pitches with one floating strip, the minimum is expected at 14 (21) degrees on the v/n side and at 4 (7) degrees for the u/p side, respectively for layer 3 (4, 5, 6).
The resolution for normal incident angle is also in good agreement with the expected digital resolution, that is 23 (35) ðœ‡m on v/n side, 7 (11) ðœ‡m on u/p side, respectively for layer 3 (4, 5, 6).
Still some studies are ongoing to improve the analysis for the cluster resolution especially for the layer 3 U side, were at normal incidence a slightly higher resolution is measured (9 um) compared to the expectations

Line 217:
After the mentioned dose/yr and eq. n fluence I would mention the “limit” of SVD w.r.t integrated damage (before you mentioned the one on occupancy from tracking) to give
the full picture. Then you would say that although we have some safety margin ~ 2-3 on detector limit w.r.t the BG estrapolation (also considering the integrated radiation damage after 10 yr of operation) we are exploring the VXD upgrade to increase this margin considering the large uncertainty on BG estrapolation ..

I would add here a sentence:

The corresponding integrated dose, using the data/MC rescaled BG estrapolation, is about 0.2 Mrad/smy, and the equivalent 1MeV neutron fluence is about 5 × 1011 neq/cm2/smy (smy: Snowmass Year = 10^7 sec).
Considering the radiation hardness of the SVD sensors, about 10 Mrad and about 10^13 neq/cm2 , based on the past experience of similar DSSD sensors used in the BaBar silicon vertex tracker (quote the BaBar Nim Paper) ,
we expect to be able to safely operate the SVD even for 10 years at high luminosity, with some safety margin w.r.t BG extrapolation of about a factor 2-3.

Line 219: add some text
This uncertainty, together with the relative small safety factor 2-3 between the BG estrapolation and the detector limits, motivates the VXD upgrade which improves the tolerance of the hit
rates and the radiation damage, and the technology assessment is ongoing for
multiple sensor options.

Line 233:
In the first two and a half years of operation the integrated radiation dose
in the layer-3 mid-plane sensors, which are the most exposed in the SVD,
is estimated to be 70 krad. The estimation is based on the measured dose by the diamonds on the BP exploiting the measured correlation between the SVD occupancy and the diamonds dose.

Line 242:
à The evolution of the noise with the integrated dose is shown in Fig. 7.

Line 254
Fixed oxide charges on sensor surface increase non-linearly, enlarging inter-strip capacitance…..
à Fixed oxide charges on sensor surface increases with dose, with some saturation expected at around 100 krad, enlarging also non-linearly the inter-strip capacitance, also expected to saturate with dose. The noise saturation is already observed on the v/n-
side and also starts to be seen on the u/p-side.
(remove last sentence This behavior agrees with the increase of fixed oxide charges.)

Line 253:
From this measurement full depletion voltages consistent with measurements performed before the installation on the bare sensors were obtained, ranging from 20 to 60V, and so far no…

Dear Yuma,

sorry to bother you but Robin spotted 2 small typos in the autor list and produced the new one on confluence: "vertex2021_pasd12_corrected.tex"
Please use that in the version to submit.

Thank you,

Stefano