Quality Control of Silicon Pixel Wafers for the CMS Phase-1 Pixel Upgrade

General information

• Contact persons: Kamuran Dilsiz, Leonard Spiegel
• Author list: Kamuran Dilsiz, Suleyman Durgut, Kai Yi, Leonard Spiegel
• Reviewers: Marko Dragicevic, Danek Kotlinski, Anders Ryd (chair), Katja Klein (ex-officio)
• Published as CMS Note-2019/002
• Published on 5th of December 2019 by the Turkish Journal of Physics as Turk. J. Phys. (2019) 43: 541-550

First version of the draft

• First version of draft, as received on 11th of March 2019 pdf

Comments to the first version

• Comments by Katja, from 2nd of April 2019: pdf
• Comments by Anders. I excluded comments that overlapped below with Katja, so the list is not that long

Abstract.

The first sentence that ends with "that would perform equally well in the radiation environment of the inner tracker" is a but confusing to me. What exactly do you mean by perform equally well in the radiation environment? If this is an upgrade you would expect better performance? The goal of the phase one upgrade was to improve the readout bandwidth - but preserve the sensor performance?

300fb^{-1} - should have a space (or ~ in latex) between the value and units. **This goes for many places in the paper, please check**

Did Sintef actually design the sensors?

Line 9 and 10 'sub-detector' is used with two different meanings. First the tracer is a sub-detector, in the second case it has two sub-detectors. I would use the first meaning.

Line 23 very hard to see from Fig 2 that the upgraded detector has one more layer...

Line 20 Here and elsewhere 2x8 should be consistently written $2\times 8$.

Line 20 One ROC does not server 66560 pixels! This is for the whole module.

Line 33 Change 'old' to 'originial'?

Figure 5 Would it be better to have a Circuit diagram that shows the connections?

Table 1 '04' to '4' and '02' to '2'

Line 96 First sentence is a repeat of what is said earlier. Remove?

Line 101 "In this paper" sounds like something you have in an intro. Remove?

Line 128 - 131 Move to conclusions?

Line 131 What do you mean by "....as each sensor had to be separated to use in the detector"?

Fig. 7 For wafer 324 why does the blue (sintef) line end at ~50 V? The voltage range in this figure is different than all others. Also here you use lowercase letters in Sintef - elsewhere it is all capital.

Table 3 'good' in subscript should not be in italic.

Line 123 You say all 8 sensors on 30 wafers is 239 sensors. This means that it was not actually all 8. Can you just say 239 sensors on 30 wafers?

* Comments from Danek

General comments.

1) The past and present tense is mixed. It should be consistence ether all in past or in present tense.

2) One should avoid "first person" and use "third person" so "was done" instead of "we did".

Abstract

As far as I know the vendor has not designed the sensor but only fabricated it.

Please remove "designed" from the sentence "Because of this, ...".

1. Introduction

Line 5: the "giant camera" part of the sentence is childish. One can just skip it and say "With a .....the detector records particle events ....."

Line 12: I think one should first mention precise hits per track and only later maybe talk about jets and long-lived particles. So "The pixel system of the CMS detector provides up to three precise hits per track, which allows ..."

Line 15: I would say "The silicon tracker is located close to the LHC beams with the innermost pixel layer located 3 cm from the interaction point."

Line 26: replace "identical" with "similar".

Line 30: um2, the square is missing

Line 30: one does not read channels, so say "... to readout the signal, the pixels were ..."

Line 31: replace "after metalizing pads" with "using metalized pads"

Line 33: delete "had an efficiency more than 99% and" it is irrelevant here.

Line 35: modify "LHC Phase 1 upgrade", LHC did not have a phase 1 upgrade, the pixels had it. say "Following the LHC long shutdown ..."

Line 37: delete "detector"

Line 39: replace "present" with "the"

Line 40: replace "in addition to" with "and in addition allows operation up to ..."

Line 46: replace "we checked" with "was checked".

Line 48: replace "did not advance to" by "was not used for"

2. Production Wafers and Experimental Test Setup

Line 50: delete "essentially"

Line 51: replace "the more" with "free"

Line 52: delete "In detector operation"

Line 54: replace "At full depletion or higher bias voltages" with "At bias voltages corresponding to the full sensor depletion"

Line 57: "... the SENSOR bulk"

Line 59: replace "... and we checked the capability of these wafers " with "... and re-tested at the ..."

Line 63: say "TEST structures."

Line 64: say "The FNAL wafer testing setup used a Summit ..."

Line 66: replace "as provided by the" by "in the"

Line 69: what are "ammeters"?

Line 69,70: the tense should be changed from present to past.

Line 74: the sentence "In practice..." is not easy to understand, rewrite please.

3.Quality Control and the Production Sensor Wafers

Line 94: Replace the sentence "To define ..." with just "SINTEF used a similar but not identical definition."

Line 95: do not use "We", so say "All sensors and diodes were tested for 5 or 10 wafers in each shipment"

Line 99: "almost THE same"

Line 101: "we show plots for sensors on" so delete "THE" and replace "in" with "on".

LINE 103: delete "THE" in "the sensors"

Line 106: delete "a little"

Line 108-109: rewrite this sentence, maybe something like "However, the SINTEF data for these sensors do not show any bad behaviour."

Line 109-110: replace the sentence "Figure 8 ..." by "Figure 8 shows in sensors that were found good at FNAL but were labelled bad according to the SINTEF data."

Line 113: replace "may" with "might"

Line 115: " ... the measurements WERE taken MIGHT HAVE BEEN different"

Line 118: replace "an" with "the" and delete "since they are consistent with out criteria"

Line 119: "Because of some DIFFERENCES in the two setups, such as ..."

Line 120: replace "data" with "measurements"

Line 121: replace "by our rules according to FNAL" by "from the FNAL ..." replace "data" with "tests"

Line 122: replace present tense with past, so "FNAL results show 208 good sensors ..."

Line 123: replace "over" by "In"

Line 128: delete "enough", replace "use" by "be used"

Line 129: replace "data about whether a sensor was good or not by "tests"

Line 130: avoid "we", so "After the tests were finished the production wafers were sent to dicing and bump bonding." no need to mention RTI and "separated prior to" is just confusing.

Answers to comments to first version

• Answers_For_Danek_Comments.pdf: 2nd version of the draft and answers for comments.</verbatim>

• Answers_For_Anders_Comments.pdf: 2nd version of the draft and answers for comments.</verbatim>

• Answers_For_Katjas_Comments.pdf: 2nd version of the draft and answers for comments.</verbatim>

Second version of the draft

• Second version of draft, as received on 15th of April 2019 FPix_LS_12Apr19.pdf

Comments to the second version

• Comments on second draft by Anders: Sensor_Test_Paper.pdf</verbatim>

• 2nd round of comments from Danek: fpix_sensors.txt</verbatim>

• Comments by Katja: pptx pdf

• Comments by Marko

Abstract

L2: “fresh sensors” —> new, unirradiated sensors

L2: “would perform better performance” - too many “perform” but what does “perform better” actually mean? I would say: “..that show equal or better performance under Run 3 conditions.”

L5: “Because of this, the same vendor .. ” I don’t understand why surviving Run 3`requires you to use the same vendor. You could say that the new sensors use similar technology and layout and because of tight timing and good experience you choose to use the same foundry again.

Introduction

L11 Only here I realised that you describe the “Phase 0” detector. You may want to mention that: “Before the Phase-1 upgrade, the pixel system …

L21 “…was completed in early 2017 after an …”

L26 it is not clear where and what the FPix is. A picture would help.

L28: I would argue that the smaller ROCS are bump bonded to the sensor, eg: “Two rows of 8 ROCs are bump bonded to each sensor connecting 66 560 pixels assembly, were each pixel has a size of 150 x 100 µm.”

L29 If there are other pixel sizes you might want to mention where and why

L34-35 remove brackets

L35 what long shutdown, you mention this here for the first time

Figure 2: You should make clear that in the left picture the upper/lower are upgraded/original and in the right its left/right old/new. Especially for the right picture the caption is misleading.

L43 “tested by the Scientific and Industrial Research [8] (the foundry) organization in Norway.” sound very odd. SINTEF is an Norwegian acronym (Stiftelsen for industriell og teknisk forskning) you can spell it out here once then use SINTEF.

Production Wafers and ...

L53 you have spelled out FPIX and used the abbreviation before. Please use FPix here as well.

L63 visible

L67 why do you mention needles in brackets here. I would use either needles or tips throughout.

L68 You say the “layout was programmed”, where you mean the movement of the tables was automatised. If then you say you moved manually anyway its not clear why you mention the automation in the first place.

Figure 4: The picture looks distorted …

Quality Control of Production ...

L86: I assume you were not only staring at the data and then selecting sensors randomly. If you had some election criteria (highest currents, odd IV shapes, barely within specs) please mention it.

L89: Not sure what the - means? I would simply mention the acceptance criteria, its clear that they should be identical for the manufacturer and us.

L94 “subsequent material” sounds odd. I assume you refer to the next section or the remainder of the paper …

L97: Say: “Fig. 7 shows the IV curves …”

I am still missing an explanation of the biasing scheme used in these sensors and its potential influence on the measurements. It might be that some of the fluctuations in the lower voltage part of the IV curves are due to that: pixels getting properly biased only after sufficient bias voltage creates a punch through connection to these pixels.

Tables 3-4 Do these long tables really add any crucial information? They are tiresome to read ...

Summary: isn’t a rate of 5% bad sensors found among sensors that where supposed to be all good rather alarming? You have to expect the same number of false positives among the sensors not tested by FNAL meaning that many modules where built with sensors that where not good? Was this of no consequence to the project?

Answers to comments to second version

• Answers for Anders comments
• Answers for Danek's comments
• Answers for Marko's comments
• Answers for Katja's comments

Third version

• 3rd version of the draft, received on 13th of May: pdf

Comments to the third version

• FPix_3rd_version_LS_Ryd.pdf: Comments on third version (Ryd)</verbatim>
• Comments by Katja to 3rd version (25th of May): pdf pptx

Answers to comments to third version

• Comments_by_Anders_4th.pdf: Answers_to_Anders_Comments.

• Comments_by_Danek_4th.pdf: Answers_to_Danek's_Comments.

• Comments_by_Katja_4th.pdf: Answers_to_Katja's_Comments.

Fourth version

• 4th version of the draft, received on 29th of May: FPix_4th_version.pdf

Comments to the fourth version

• Comments by Katja to 4th version, 31st of May pdf pptx
• FPix_4th_version_Ryd.pdf: Comments on the 4th version of the paper</verbatim>
• FPix_4th_version_Marko.pdf: Comments by Marko to 4th version</verbatim>

Answers to comments to fourth version

• Comments_by_Anders_5th.pdf:Answers_to_Anders_Comments.

• Comments_by_Marko_5th_Version.pdf: Answers_to_Marko's_Comments.

• Comments_by_Katja_5th_Version.pdf: Answers_to_Katja's_Comments.

Fifth version

• 5th version of the draft, received on 7th of June: FPix_5th_version.pdf