Vladislav Zakharov Monday February 10 th , 2020 2 Background & - - PowerPoint PPT Presentation

Vladislav Zakharov

Monday February 10th, 2020

1

Background & Motivation:

u

Time Projection Chamber (TPC)

Ø A type of detector Ø A type of capacitor

Vlad

2

u

Outer & Inner mandrel construction in our lab

u

To be used in sPHENIX

u

Can be used in Electron Ion Collider (EIC)

Collider Experiments

Vlad

3

u

Charged particles, ions or leptops (𝑓"), hit neutral atoms in fix target.

u

They hit each other in beam-beam

u

Soon, ion-to-leptop in beam-beam EIC =)

Overview of sPHENIX at RHIC

4

Vlad

u

Multiple different detectors, in layers on top of each other, are needed to measure all the results.

u

Tracking:

Ø High precision (and high cost) pixilated silicon

detectors

Ø TPC: measures tracks from charged particles with

the help of a 𝐶-field

u

Energy Deposition:

Ø Electro-Magnetic Calorimeter: measures energy

“showers” from electrons & photons

Ø Hadronic Calorimeter: Energy from hadrons

u

Other detectors

Ø Scintillators e.g. RICH, µ-detector, etc.

Vlad

5

ATLAS Lego model at WIS J

TPC’s Operating Principle

• 1. Anode & Cathode separated by a dielectric fluid (usually gas; unless you’re looking for 𝜉).
• 2. Particle traverses the gas, ionizing it
• 3. Uniform 𝐹-field drifts the resulting charges
• 4. Anode is segmented to see the track left by the particle

Vlad

6

≈ 2 . 1 1 m ≈1.6m

Detection Stage

u

Unique interleaving “Zig-Zag” pads

u

Maximize charge sharing through:

Ø max incursion of neighboring pads Ø Minimal tip-to-tip spacing

u

Over a decade work minimizing Differential Non-Linearity (DNL) - measures deviation from expected results across pads

Vlad

7

Charge clouds collected on multiple vs. a

• n single pad

𝑌~ ∑ 𝑟*𝑦* ∑ 𝑟* 𝜀- ≈ 𝑇𝑗𝑕𝑜𝑏𝑚 𝑂𝑝𝑗𝑡𝑓 𝜀- ≈ 𝑋 12 W is width

• f the pad

Amplification Stage

u

1 electron doesn’t have enough charge to overcome electronics noise

u

Need to use gain:

Ø Gas Electron Multiplier (GEM) Ø Micro-Megas (µM) Ø Multi-Wire Proportional Chambers (MWPC)

u

Create large local 𝐹-field that accelerate the incoming electrons. The high-energy 𝑓" then hits the nearby neutral gas molecules and forces them to release multiple 𝑓".

u

With a high enough 𝐹-field, or several stages to cascade, the resulting electron cloud can be reliably detected

Vlad

8

Gas Electron Multiplier (GEM)

9

u

About 2,000 gain. Quad-Stack pioneering by ALICE

u

ΔV = top to bottom of single foil, ΔV = between two GEMs

Ø ΔV and ΔV are comparable at ≈200—400V, but distance

Δd≈2—4mm while Δd≈40—60µm!

Ø 𝐹;<*=> = .4 kV/cm, 𝐹><?@A=B< ≈ 1s kV/cm, 𝐹CDEB ≈ 10s kV/cm

Vlad

Pad Plane ΔV1 ΔV2 ΔV3 ΔV4 ΔV1 ΔV2 ΔV3 ΔV4

Real GEM photos

Vlad

10

IBF & Space Charge (SC)

u

SC is the enemy of resolution

u

⃗ 𝑤*D@ ;<*=> = 𝐿𝐹 (large K {Ne}, large 𝐹 = 400𝑊/𝑑𝑛)

u

Detector performance limited by the fluctuations in deflections since SC is not continuous on average

u

Minimize C: Bias Operating Point of Micro pattern Gas Detector (MPGD) for low IBF (such as was done by ALICE), Passive IBF shielding (topic for today’s talk)

11

𝜍 𝑠, 𝑨 ∝ UD@*V?>*D@ ∗ XYE>*ZE*[*>\ ∗]?>B

^_`a b"

c deafgh i j

<k Primary IBF K (mobility) of Ne

Vlad

0% IBF

Radius [m] z [m]

10

1% IBF

Radius [m] z [m]

100 ρ(r,Z) [fC/𝑑𝑛n]

u At 2,000 gain & only 1% IBF

, 20 ions are drift and only 1 is primary.

This is 95% of the Space Charge!

Electron vs. Ion Transport in a Gas

u

Battling SC requires distinguishing between 𝑓" and ion transport

u

Both obey the Langevin Equation for transport:

𝑛 𝑒 ⃗ 𝑤 𝑒𝑢 = 𝑟𝐹 + 𝑟 ⃗ 𝑤×𝐶 − 𝜆 ⃗ 𝑤

u

Full characterization is VERY COMPLEX requiring calculations & measurements

u

Nonetheless, we can direct our calculations using simplified considerations

u

The basic “Langevin Distinctions” between 𝑓" and ions are:

u Opposite q: Design Forward-Backward Asymmetry into electric fields u Different ⃗

𝑤: Typically opposite in direction, different in magnitude… Use 𝐶 to our advantage

u

It is possible to design structures that utilize all these differences to:

Ø Minimize the amount of ions coming from the avalanche and reaching the main drift volume Ø Retaining high 𝑓" transport to the avalanche zone

12

Vlad

Edrift Etransfer Ehole core core halo halo halo halo core core

Forward-Backward Asymmetry

u

The classic GEM picture with 𝐹><?@A=B< > 𝐹;<*=>

u

Only a fraction of the transfer field lines originate in the drift volume

u

Effective transparency difference for forward-backward:

Ø Driving characteristic is the field ratio: vgwxayzew v{w_zg Ø Most electrons get through (and avalanche), while many Ions are blocked

electrons forward ions backward

13

Vlad Garfield & Magboltz simulation of charge dynamics of 2 e- arriving in a GEM hole. e- paths are yellow, ion paths are red. Green spots at ionization locations.

Bohmer et al. – SC Effects in an Ungated GEM-based TPC

GEM Quad-Stack Data

u Fundamental tradeoff of IBF efficacy vs. Energy Resolution:

u Gain biased toward last GEM(s) [nearest pads] è Low IBF u Gain biased away from first GEM(s), coupled to gas è Gain

fluctuations… decreased resolution

14 2014-03-03 TDR for the Upgrade of the ALICE TPC

Energy Resolution

Vlad

Odd & even GEMs are:

1) Aligned but vary in pitch 2) Rotated with respect to

each other This reduces the chances of an ion from the pad plane floating to the gas volume

1 2 3 4

u Nothing beats µM for Field Ratio u Most extreme by lowering 𝐹*@;Y[>*D@

Ø Mid GEM lowers the induction field for the v|e}`~

vx|`•e

concept, but eats 𝑓"

Ø Top GEM provides some gain to compensate for 𝑓" loss in Mid GEM

15

Vlad

• S. Aiola et al – Combination of dual-GEM and μM as gain elements for a TPC

Hybrid: Dual-GEM and microMegas (µM)

µM zoom’ed in

• V. Manuel et al – A Radiation Imaging Detector

Made by Post processing a Standard CMOS Chip

Data from µMegas

Vlad

16

Raw

Better… but still competition: IBF vs Resol.

Measurements of IBF vs. field ratio for a 1,500 lpi (lines per inch)

• micromesh. Done with an intense (10mA-10keV) X-ray gun.

Colas P. et al - IBF in the Micromegas TPC for the Future Linear Collider

• S. Aiola et al – Combination of dual-GEM and μM as gain elements for a TPC

IBF reduction without 𝑓" Resolution Loss?

u

In any multi-stage gain structure, a low gain stage makes irreducible contributions to gain fluctuations.

u

The first (early) stage(s) of 4G and 2G-µM must have low gain since they are coupled strongly to the gas.

u

Nonetheless, the field ratio principle (large

vgwxayzew v{w_zg

è low IBF) applies even without gain.

u

Therefore, a passive structure generating a field ratio can lower IBF with little or no loss in energy resolution.

17

Vlad

Passive Mesh Calculations/Simulations

u

Full Garfield transport calculations

u

Drift Field is fixed to sPHENIX (400 V/cm)

u

Transfer Field is scanned: Ed, 2Ed, 3Ed, 4Ed, 5Ed, 6Ed (from sublime to ridiculous)

u

Magnetic field is scanned (relevant for low ET) 0T , 0.5T , 1.0T , 1.5T , 2.0T , 2.5T

u

Ideal result would be 100% 𝑓" transparency and 100% ion-blocking Ideal Ideal 5X better 3X better 2X better (than 25% at a ratio of 1)

18

Vlad

Among the Best Studied: Passive Meshes

u

Transfer field 2-3x Edrift (reasonable)

u

IBF improvement factors 2-3x (excellent)

u

𝑓" Transmission 90-98% “Etched” Mesh A simple mesh should lighten the burden and improve performance on any 4G or 2G-µMEGAS

• structure. However, an improvement of only 2-

3x in ion blocking would mean that IBF is still the dominant source of SC

19

Vlad

Bi-Polar Gating Grid

u

Blocks ions by collecting them on negative wires… but blocks 𝑓" with positive wires

u

Active gating: Since ions & 𝑓" have different drift velocities and hence different drift times, turn the voltages off to allow 𝑓" to pass and then turn them back on to collect the ions.

Ø Creates dead time while waiting for ions to be collected. Potentially huge data loss in high

luminosity experiments. Vlad

20

Vlad

21

The 𝑓" are coming But ions are drifting back from the gain stage We still need the signal from 𝑓"

What about the Magnetic Field Term?

u

Negligible for SLOW ions… not negligible for 𝑓"

Ø In sPHENIX: 𝑤;<*=> ≈ 80

⁄

‚ƒ @A, B = 1.4 Tesla

u

Traditionally, one attempts to zero this term to avoid extra distortions by making 𝐹 ∥ 𝐶

u

Nonetheless, one can make a localized ⃗ 𝑤×𝐶 kick that only 𝑓" feel

v

This concept is discussed in detail in Blum’s book

𝑛 𝑒 ⃗ 𝑤 𝑒𝑢 = 𝑟𝐹 + 𝑟 ⃗ 𝑤×𝐶 − 𝜆 ⃗ 𝑤

22

Vlad

Question: Can the magnetic field aid electrons in passing through an otherwise closed gate?

𝐶 = 0

Introduction of Magnetic Field:

u

Magnetic Field brings electrons through.

u

Ions remain blocked.

23

Vlad

Simulations of the Bi-Polar Wires

24

Vlad

Ne:CF4 (90:10), 𝐹><?@A=B<= 600V/cm, 𝐹;<*=>= 300V/cm, Wire pitch = 1mm

• e- transparency not perfect (70%)
• But all the ions are still blocked

WIS Data

Vlad

25

Vlad

26

Amplification wires

(most positive item)

Mylar ±ΔV Bi-polar

grid wires

Mesh Pad

Fe-55

“field” +

• +

+ + + + + +

2.3 cm 0.8 0.6 0.3 0.0

𝐹;<*=>=(mylar-grid)/1.5cm 𝐹><?@A=(grid-mesh)/0.2cm

WIS Set-Up

uX-ray source on top ->

ionization in any region instead of just drift volume

uMWPC used for amplification

• > Mesh needed to help

terminate the field

Primary Push is ALONG the wire!

u

… 𝐹× … 𝐶 near the wire is along the wire.

u

Once the electron picks up a velocity along the wire, only then does it move transverse to the wire…miss the wire…get transmitted…save the day.

u

Can we tolerate or compensate the distortion along the wire?

27

Vlad

Distortion as Differential Non-Linearity

u

Electron Displacements from ideal trajectory are cyclic.

u

The cycle repeats with the same period as the wires.

u

Cyclical shifts from ideal positions are known as Differential Non-Linearity.

Large + Push Large - Push

Dx QUESTION: Can we define a specially distorted pad shape to compensate the “DNL” in electron positions introduced by the Bi-Polar grid?

28

Vlad

u

Years of effort have gone into minimizing or eliminating DNL from zig-zag pad response.

How to Anti-Distort the electrons?

u We know how to define a zig-zag shape with minimal DNL:

Ø Maximum “incursion” of neighboring pads (>95%) Ø Minimal tip-to-tip spacing (< spot size of avalanche)

u The following procedure defines the anti-distortion zig-zag shape:

• 1. Match the wire pitch to the pad pitch.
• 2. Generate electrons at positions that SHOULD intercept the gaps between Zig-Zag.
• 3. Propagate electrons through all distortions.
• 4. Match the actual pad gaps to the determined electron landing spots.

u Shorthand: Design that each electron lands on the same pad number as

without distortion!

Ø Normally, distortions are accounted for later in the data analysis stage. But now

it’ll be done in the collection stage, thus freeing up computing power and code- writing time for other analysis tasks.

29

Vlad

Distortion Examples

u Bipolar wires on top of pads (one example, multiple arrangements possible)

30

Vlad

Original Pad Edges

Initial 𝑓" positions

Distorted Pad Edges

Final 𝑓" positions

31

Vlad

Colors to guide the eye

Distorted Pad Shape

32

Vlad

Summary

u

IBF is, by far, the main contribution to SC

u

Quad-GEM & dual-GEM + µM are able to reduce IBF

u

Passive wire mesh can reduce IBF

u

Passive bi-polar wires can reduce IBF and we might be able to account for position resolution by creating pre DNL-distorted pads

Thank you

u Current sPHENIX design meets all of our goals, but we

will still study this as a possibility for improvement

IBF & Space Charge (SC)

u

Ionization ∝ 𝑎‡: Use Low-Z Primary Gas (Ne)

u

Multiplicity: # of particles from collision (nature)

u

Rate: beam-crossing interactions (we control, ≈100kHz)

u

z = distance from CM

u

ρ greater at smaller r since it’s closest to beam collision, and particles spread ∝

b <k

u

ρ greater at smaller z since it sees more ions as they drift to the CM

Vlad

33

Backup