thermal performance of the petal prototype · picture of the tested petal prototype built at desy...

Anja Beck DESY Summer School 2018

Thermal Performance of the Petal Prototype

Anja Beck, TU Dortmund, GermanyDESY Summer Student Programme 2018

ATLAS Experiment

Supervisor: Claire David

September 5, 2018

Abstract

With the LHC luminosity upgrade, requirements regarding heat resistance for the detector parts aggravate sig-nificantly. To ensure stable measurements and avoid thermal runaway during data taking, assessing the thermalperformance of each of the detector parts is crucial. This report treats the preparations and measurements con-cerning the thermal behaviour of a prototype of a wedge-shaped local support unit, called petal, for the ATLASinner tracking detector during the Summer Student Programme 2018 at DESY.This includes determination of the emissivity of the sprayed coating for the infrared measurements, studies onthe theoretical conversion between emitted power and temperature, exploring the camera software, as well as firstresults of the measurements with the petal.

Acknowledgements

I am very grateful towards Claire David and Jan-Hendrik Arling for creating an enjoyable and comfortable at-mosphere in the lab with no fear of stupid questions. Also, I want to thank the team of engineers for their helpand cheerful appearances, in particular Frauke Poblotzki for proofreading. Additionally, I want to thank SergioDiez-Cornell for his advice whenever needed. And a last big thank you goes to the entire DESY ATLAS group thatmade the office a happy and sociable place to work and have BBQs at.

Contents

1 Introduction 21.1 Petal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.2 Cooling System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 Infrared Theory 32.1 Emissivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.2 Conversion from Temperature to Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.3 Comparing Manual and Camera Computations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

3 Emissivity Measurements 63.1 Set-Up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

4 Understanding the Camera Software 9

5 Preparing the Petal 105.1 Last Pre-Study on Gluing Pt100s and Paint Emissivity on Si . . . . . . . . . . . . . . . . . . . . . . . 105.2 Spraying the Petal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

6 First Results 14

7 Summary and Outlook 17

Thermal Performance of the Petal Prototype 1


1 Introduction

The high-luminosity upgrade for the LHC planned to start in 2026 imposes new challenges on detectors. An enhancedluminosity is related to the need of more particle detecting channels as well as elevated radiation levels damaging theelectronics, both of which constitute heat development. To avoid a thermal runaway and ensure reliable measurements,the electronics and most importantly the Silicon sensors must be held at a controlled temperature, which requires aspecific cooling system for all detector components.

1.1 Petal

The summer student project documented in this report was related to the future inner tracking detector (ITk) of theATLAS experiment, specifically to the thermal performance of the end-cap local structures. Figure 1 shows the entireATLAS detector with the current inner tracker, composed of the Pixel Detector, SCT Tracker and TRT Tracker, fromthe closest to the furthest of the interaction point. This will be entirely replaced by the ITk tracker, whose schematicscan be seen on Figure 2. The detector parts studied at DESY are the so-called petals that are assembled in discsperpendicularly around the beampipe, resembling the petals of a flower as shown in figure 3. Figure 4 displays apicture of the tested petal prototype built at DESY to simulate the thermal heat load. The prototype consists of acooling loop embedded in two carbon-fiber facesheets, blank Silicon glued on each side, and dummy electronics gluedon the sensors.

Figure 1: Schematics of the ATLAS detector

(a) Current ATLAS inner tracker.(b) Planned new tracker ITk for High-Luminosity LHC.

Figure 2: Inner tracking detector of ATLAS before and after the planned upgrade.



Figure 3: Schematics of the assembly of thepetals in the ATLAS inner tracking detec-tor endcap.

Figure 4: Picture of the tested petal prototype.

1.2 Cooling System

Figure 5 shows a prototype of the bare cooling loop. The cooling system uses the phase change enthalpy. Liquid CO2

is pumped into the cooling loop where some of it evaporates if exposed to heat. This evaporation requires energy(enthalpy of evaporation), which is taken from the heat source.

Figure 5: Picture of a bare cooling loop.

2 Infrared Theory

To assess the thermal performance of the petal, we measure the emitted radiation in the infrared (IR) spectrum. Toproperly evaluate the data measured with the IR camera, we need to understand the behaviour of IR radiation andcamera software. This section gives an overview over these topics.

2.1 Emissivity

Every body emits electromagnetic radiation depending on its temperature. Light in the IR spectrum behaves identicalto the more intuitive visible light. This means that surfaces can emit, absorb, and reflect IR radiation. Being purelyinterested in the emitted power, we need to minimize reflection in the IR region. The emissivity ε describes the abilityof a surface to emit IR radiation. The emissivity of a real surface is defined as the ratio of its radiant exitance Mreal

and the radiant exitance of a black body at the same temperature MBB:

ε =M

MBBl (1)

As the black body is an ideal emitter, the emissivity is between 0 and 1. In terms of reflection, an emissivity of 0corresponds to total reflection, whereas 1 corresponds to no reflection. So, to reduce the non-wanted IR radiations, wecover the petal with a high emissivity coating. The determination of the exact emissivity value for the chosen coating,a sprayed black paint, is described in section 3.



2.2 Conversion from Temperature to Power

To fully comprehend and also for being able to check the camera data, the derivation of a relation between the emittedpower and temperature is crucial. As we are trying to approach an ideal black body using the high emissivity paint,Planck’s law for black body radiation,

p(λ, T ) =2πhc2

λ5

1

exp (hc/(λkBT )) − 1, (2)

can be a good start. The variables are the wavelength λ and the surface temperature T . The IR camera measuresradiation over a range of wavelengths, so we need to integrate this over said range and obtain

F (T ) =

∫ λmax

λmin

C1

λ5

1

exp (C2/(λT )) − 1dλ . (3)

The constants are C1 = 2πhc2 and C2 = hckB

, with the Planck constant h, the speed of light in vacuum c, and theBoltzmann constant kB. As the air around the system is not at absolute zero and thus emits IR radiation (referred toas ambient), the total incoming IR radiation is given by

P (T ) = εF (T )︸︷︷︸emission

+ (1 − ε)F (Tamb.)︸︷︷︸reflection of ambient

. (4)

Note that we neglect the transmissivity here.

2.3 Comparing Manual and Camera Computations

In the following, we test the accuracy of equation 4 by varying different variables.

Different Wavelength Ranges

We are not entirely sure which range of wavelengths is the most relevant for our set-up. Therefore, we tried differentintegration bounds, the result of which is displayed in figure 6. Seemingly, the smaller range approximates the cameravalues better but still not satisfyingly.

Different Emissivities

We are also interested in how the emissivity influences the conversion. Figure 7 shows P (T ) for different emissivityvalues. A striking observation: varying the emissivity between 0.90 and 1.00 for P = 60 W/m2 leads to a temperatureuncertainty of almost 10 ◦C.

Fit with Global Scaling Factor

We also tried fitting the equation to the camera values using a global scaling parameter:

P (T ) = τ [εF (T ) + (1 − ε)F (Tamb)] . (5)

In figure 8, we see the fit obtained with the built in function curve fit of the python package scipy.oPtimize. Looking atthe plot, one intuitively remarks that this does not seem to be the best fit. Interestingly, this computational deviationis not reflected by extremely high uncertainties:

τ1.00 = 0.821 ± 0.018 ,

τ0.95 = 0.856 ± 0.021 ,

τ0.90 = 0.896 ± 0.024 .

Even if not conclusive, through these studies, we could gain more intuition on the physical entity of radiated powerand emissivity’s quantitative impact on temperatures.



20 15 10 5 0 5 10

60

80

100

120

140

P in

W /

m2

Computed power 8 - 13 mIR camera data 8 - 13 mComputed power 7.5 - 14.0 mIR camera data 7.5 - 14.0 m

20 15 10 5 0 5 10T in C

1.1

1.2

1.3

1.4

1.5

Ratio

Com

p/Ca

m

8 - 13 m7.5 - 14.0 m

Figure 6: Testing with different wavelength ranges.

30 20 10 0 10

40

50

60

70

80

90

100

110

120

P in

W /

m2

comp, eps = 1.00cam, eps = 1.00comp, eps = 0.95cam, eps = 0.95comp, eps = 0.90cam, eps = 0.90

30 25 20 15 10 5 0 5 10T in C

1.2

1.4

1.6

Ratio

Com

p/Ca

m

1.000.950.90

Figure 7: Test with different emissivities.



30 20 10 0 10

40

50

60

70

80

90

100

P in

W /

m2

comp, eps = 1.00, tau = 0.822camcomp, eps = 0.95, tau = 0.856camcomp, eps = 0.90, tau = 0.896cam

30 25 20 15 10 5 0 5 10T in C

1.0

1.2

1.4

Ratio

Com

p/Ca

m

1.000.950.90

Figure 8: Fitting the function to the camera values using a global scaling factor.

3 Emissivity Measurements

This section treats the determination of the emissivity of the paint. To do so, we need the paint to be next to a surfaceof known emissivity. We use high emissivity tape with εT = 0.95. In terms of equation (4), we can then write

PP = εPF (TP) + (1 − εP)F (Tamb) for the paint,

PT = εTF (TT) + (1 − εT)F (Tamb) for the tape.

Assuming that both areas have the same real temperature because of their proximity and therefore emit the sameamount of IR radiation, we set PP = PT. After rearranging, we obtain an equation for the emissivity of the paint

εP = εTF (TT) − F (Tamb)

F (TP) − F (Tamb). (6)

In the following, we first describe the used set-up for measuring the emissivity of the paint and afterwards analyse theobtained data.

3.1 Set-Up

As the petal will later be used at temperatures below 0 ◦C, the thermal performance tests need to be adapted to lowtemperatures, too. Therefore, we measure the emissivity of the paint on the cold side of a Peltier element. Actually, weuse two Peltier elements next to each other covered by aluminium plates on both sides. For simplicity, we shall refer tothis combination of Peltier elements as ’the Peltier’. Figure 9 shows its cold side with the tape and paint. Additionally,we measure the surface temperature with contact temperature sensors (Pt100), which we use as reference to the IRmeasurements. The IR data is the main temperature measurement that also serves to calculate the emissivity.The Peltier sits in a closed cardboard box having merely a hole for the camera lens. The purpose thereof is to shieldthe measurement from outside radiations and thus reflections. The additional dry air flushing ensures a low humid-ity to avoid ice formation. Furthermore to secure an optimal cooling of the Peltier, we cool its warm side using a chiller.

To gather the needed data, we run a current through the Peltier and wait some minutes until the Pt100s and thevoltage stabilise. Then we note the Pt100 values as well as the temperature and radiant power computed by the



Figure 9: Peltier element used for measuring the emissivity of the paint. We see the cold side with a tapedand a painted area. There are also for thermocouples (type: Pt100) which are named according to theirposition (e.g. TL means tape left). Additionally there is an IR camera measurement area next to each ofthe thermocouples.

software for the four measurement areas. Additionally, we measure the ambient temperature and relative humidity inthe box. We stop the measurement when the Peltier stops cooling efficiently (no stable minimum) or when ice formson the Peltier.

We perform three sets of measurements, each time trying to lower the relative moisture in the box. Using humidityabsorbing beads works well. Covering the inside walls of the cardboard box with acrylic glas to reduce the surface oforganic material does not lead to any improvement.

3.2 Results

Figure 10 shows the results for the emissivity of the paint for all three runs. The emissivity values are quite stablewithin each of the measurement runs but vary a lot among them. We have no explanation for this yet.

20 15 10 5 0 5temperature in C (PT100)

0.910

0.915

0.920

0.925

0.930

emiss

ivity

emissivity of the paint

min. humidity: 8.0min. humidity: 7.0min. humidity: 6.0Average

Figure 10: Paint emissivity values for the three different measurement runs with their absolute average.



Figure 11 displays the difference in temperature between the Pt100s and the corresponding IR measurement pointtaken with the respective emissivity value. For the tape values, this is mostly within the acceptable range of 1 ◦Cdeviation. For the paint values the deviation reaches more than 4 ◦C. Given the good result for the tape, we suspectthe paint emissivity value of 0.920 to be wrong.

20 15 10 5 0 5 10 15T in C (PT100)

1

0

1

2

3

4

T in

CTL, h = 8.0TR, h = 8.0PL, h = 8.0PR, h = 8.0

TL, h = 7.0TR, h = 7.0PL, h = 7.0PR, h = 7.0

TL, h = 6.0TR, h = 6.0PR, h = 6.0

Figure 11: Temperature difference ∆T for all available data points in the three different measurementruns. For the tape the emissivity is set to 0.95, for the paint it is 0.92.



4 Understanding the Camera Software

When reaching the camera sensor, the IR photons induce a current. The software interprets this signal and displaysthe corresponding temperature. According to the user’s manual, these computations are based on equation (4) (seesection 2.2). As described in section 2.2, we are not yet able to reproduce the camera data with it.

To gain confidence in the software nevertheless and intuition for crucial variables in infrared measurements, we con-ducted studies using the camera software. Figure 12 shows the result. To obtain it, we took one of the thermogramstaken during the emissivity measurements described in section 3 and chose one measurement point (paint right). Wethen manually set different ambient temperatures and varied for each of them the emissivity, keeping all other variablesconstant. In brief, the plot shows the dependence of the object temperature on ambient temperature and emissivity.

0.800 0.825 0.850 0.875 0.900 0.925 0.950 0.975 1.000emissivity

30.0

27.5

25.0

22.5

20.0

17.5

15.0

12.5

10.0

obje

ct te

mpe

ratu

re in

C

Tamb = 05CTamb = 10CTamb = 15CTamb = 20CTamb = 25CTamb = 30CTamb = 35C

Figure 12: Temperature on the tape for different emissivities and ambient temperatures computed by thesoftware.

Figure 13 shows the same plot using the same thermogram. As pointed out by the red dotted line, the emissivity forthis measurement with real Tamb = 21.9 ◦C and real surface temperature Tpt100 = −11.54 ◦C should roughly be 0.96.This is close but not identical to the manufacturer value of 0.95. Thus, we can trust the measurement reasonably butthere is either still some aspect of the measurement we do not understand or a small error in the software we need todetermine.

0.94 0.95 0.96 0.97 0.98 0.99emissivity

16

15

14

13

12

11

10

obje

ct te

mpe

ratu

re in

C -11.54

temperature on Tape for real T_amb = 21.9

Tamb = 05CTamb = 10CTamb = 15CTamb = 20CTamb = 25CTamb = 30CTamb = 35Cmanufacturer: eps = 0.95software: eps = 0.96thermocouple

Figure 13: Temperature on the tape for different emissivities and ambient temperatures computed by thesoftware. The ambient temperature was 21.9 ◦C, leading to an emissivity of 0.96 as illustrated with thered line and dot.



We can also use this kind of plot to evaluate the impact of the uncertainty in the emissivity value (see section 3) on thetemperature measurement. Figure 14a shows the impact of ambient temperature and emissivity on the temperatureon the paint. Assuming Tamb = 20 ◦C, the emissivity range and the following temperature range is being highlightedby the orange dotted lines. More precisely, the emissivity range of 0.905 ≤ ε ≤ 0.930 determined in section 3 leadsto a temperature uncertainty of just above 1 ◦C at a real surface temperature of roughly Tpt100 = −10 ◦C (see 14a).For colder temperatures around Tpt100 = −20 ◦C, the following temperature uncertainty rises to 2 ◦C (see 14b). Toaccomplish the uncertainty threshold of 1 ◦C on the temperature measurement in all temperature ranges, we need torestrict the emissivity uncertainty to around 1 %.

0.89 0.90 0.91 0.92 0.93 0.94 0.95 0.96emissivity

20

18

16

14

12

obje

ct te

mpe

ratu

re in

C

-14.7

-13.4

-11.44

temperature on Paint for real T_amb = 21.9

Tamb = 05CTamb = 10CTamb = 15CTamb = 20CTamb = 25CTamb = 30CTamb = 35Cthermocouple

(a) Tpt100 = −11.44 ◦C.

0.89 0.90 0.91 0.92 0.93 0.94 0.95 0.96emissivity

34

32

30

28

26

24

22

20

obje

ct te

mpe

ratu

re in

C

-26.5

-24.5

-21.42

temperature on Paint for real T_amb = 21.0

Tamb = 05CTamb = 10CTamb = 15CTamb = 20CTamb = 25CTamb = 30CTamb = 35Cthermocouple

(b) Tpt100 = −21.42 ◦C.

Figure 14: Object temperature on the paint for different emissivities and ambient temperatures computedby the software, including a display of the effect of an uncertainty in the emissivity on the temperaturefor an ambient temperature of 20 ◦C.

5 Preparing the Petal

This section treats the preparation of the petal in the first two subsections and eventually the first results of thethermal performance tests done in late August 2018 in the third subsection.

5.1 Last Pre-Study on Gluing Pt100s and Paint Emissivity on Si

As outlined in section 2.1, the petal needs to be covered with a high emissivity coating. The chosen paint was examinedin section 3. Before the tests could start, we decided to do a last pre-study to check if the paint when sprayed ontoblank silicon (Si) has the same emissivity as on aluminium (Al) and to check if gluing the Pt100s onto Si is as reliableas clamping. The set-up for this test is similar to the one described in 3.1: the same Peltier element sitting in thesame cardboard box. The only modifications are on the surface of the Peltier as shown in figure 15.

Reliability of Glued Pt100s

Figure 16 compares the temperature measurement of three glued Pt100s and one clamped Pt100 on the Si plate.Oneof the glued Pt100s shows a large deviation whereas the two remaining glued Pt100s are rather close to the clampedone. We consider gluing to be a trustworthy measurement overall.

Emissivity of the Paint on Si

To obtain an emissivity value for the paint on Si, we use again equation 6, comparing the Si measurement areas withthe rightmost tape area. Additionally, we compute again emissivities for the paint on aluminium, comparing eachpaint area with the closest tape area. Figure 17 shows the resulting emissivities. Notably, the values are relativelystable over the entire explored temperature range (≈ −20 ◦C to 15 ◦C) and the values for the paint on aluminium arein the range determined in section 3. We expect to have the same emissivity independent of the underlying material.Unfortunately, the emissivities of the paint on Si are higher than this range. We suspect reflections of the clamps, theglue, and the shiny uncoated part of the Si plate to possibly have caused this result. The grouping of the values forthe three left and right Si measurement areas supports this argument.



Figure 15: Cold side of the Peltier to test the emissivity of the paint on Si and the reliability of gluedPt100s. The Si plate (yellow contour) has two parts: one unpainted side with three glued and one clampedPt100, and one painted side with six IR measurement areas. The green and blue squares represent the IRmeasurement areas on the tape and paint.

time

8

7

6

5

4

3

tem

pera

ture

in C

SiGlued1SiGlued2SiGlued3SiClamped

Figure 16: Temperatures as measured by the three glued and the one clamped Pt100s on the Si.

20 15 10 5 0 5 10 15camera temperature in C

0.90

0.91

0.92

0.93

0.94

0.95

0.96

emiss

ivity

PaintLeftPaintCentrePaintRight

SiTopLeftSiMiddleLeftSiBottomLeft

SiTopRightSiMiddleRightSiBottomRight

avg paint on Alavg paint on Si

Figure 17: Emissivity values for the paint on the Peltier (aluminium) and paint on the Si plate.



5.2 Spraying the Petal

Intuitively, one would spray the paint onto the petal by laying it on a horizontal surface and spraying from above. Thiswould lead the paint to distribute evenly due to gravitational forces and give a sleek layer of paint. However, numberone priority in the spraying process is preserving the wirebonds and the electronics in general. Having a big amountof liquid paint on the petal seems to be too risky which is why we choose to apply the paint from a distance whilethe petal is in a horizontal position. This also prevents dripping of drops from the bottle opening onto the petal. Infigure 18, we see the petal before and after the spraying procedure. Figure 19 shows a close-up of the petal. Althoughtests on dummy silicon sensors yielded to a smooth matt surface, the prototype exhibits a rough, sandy surface. Itcould be linked to a scattering effect on the spray with the wirebonds, creating small bubbles of paint staying on thesilicon sensors. So far, we have no reason to believe that this interferes with the accuracy of the measurements, as thecamera is relatively far away from the petal (more than one meter).

If a surface is recorded with an IR camera facing the surface perpendicularly, the surface reflects the heat emitted bythe camera sensor. This effect is called Narcissus effect. After spraying, we test the petal for this effect, as it is morepractical and possibly more accurate, due to more symmetric focusing, to record the petal without an angle betweenit and the camera lens. Fortunately, we do not observe any reflections in the thermogram.

(a) before (b) after

Figure 18: Petal before and after applying the paint.

Figure 19: Close up view of the paint structure.



6 First Results

This section treats the first thermal performance tests. This includes a description of the data taking and a firstglimpse at some results.

The measurement takes place in a dark box made of acrylic glass, which shields the experiment from visible andinfrared radiation. We connect the petal to TRACI, a device providing the required saturated liquid CO2 for thecooling loop within the petal. We manually set pressure values and wait for the pressure to reach a steady state.Then, we turn on the electronic modules on the petal in various combinations and record thermograms with the IRcamera. To understand possible gravitational effects caused by the different orientations of the petals later in theendcap, we repeat this routine for various orientations. Figure 20 shows the thermograms of two orientations, referredto as horizontal and vertical.

(a) horizontal (b) vertical

Figure 20: Thermograms of the petal with the front side electronics fully powered at a setpoint of 31 bar.In 20a, the petal was in a horizontal position exactly as it is displayed in the thermogram. In 20b, the petalwas rotated 90 degrees counter clockwise, since the camera was moved in the same way, the thermogramlooks similar to the horizontal orientation.

A first step in analysing the thermograms is looking at the temperature along the cooling loop. Therefore, we take animage on which the cooling loop is clearly visible. We use the measurement option in IRBIS and draw measurementlines perpendicular to the cooling loop onto the thermogram. By taking the lowest temperature along each of theselines, we presume to obtain the actual temperature of the cooling loop (that reaches the surface of the petal). Thedistances between the chosen lines is 10 pixels on the straight parts of the loop. We model the curved parts by eyetrying to ensure consistent distances.

Figure 21 shows the temperature along the cooling loop for the horizontal and vertical orientation and differentsetpoints. The route starts on the upper pipe in the thermograms and ends at the lower pipe, which follows the CO2

flow. We include the parts of the pipe that are outside of the petal. To give some markers, we plot the bent pipeparts as black lines. So for example, curve3 and curve4 together represent the 180 degree turn on the right side ofthe thermograms, with a short straight part in between. Remarkably, we can spot each of the electronics (ASICs,powerboard) as peaks in the temperature distribution. Further analysis and interpretation of these results will bedone, but for now we observe that the temperature is stable along the pipe and that the cooling of the electronicsscales nicely with the CO2 set to lower temperatures.For a comparison of the two orientations see figure 22. Notably, even though the temperatures for the horizontalorientation are roughly 1 ◦C higher all along the pipe, the temperature distributions show identical behaviour.



curv

e1

curv

e2

curv

e3cu

rve4

curv

e5

curv

e6

15

10

5

0

5

10

15

20

T in

C

setpoint = 20 barsetpoint = 23 bar




(a) horizontal

curv

e1

curv

e2

curv

e3cu

rve4

curv

e5

curv

e6

0

5

10

15

20

T in

C


setpoint = 40 bar setpoint = 45 bar setpoint = 51 bar

(b) vertical

Figure 21: Temperature along the cooling loop for the horizontal and vertical orientation and differentsetpoints.



0

5

10

15

20

T in

C

horizontal, setpoint = 31 barvertical, setpoint = 31 barhorizontal, setpoint = 35 barvertical, setpoint = 35 barhorizontal, setpoint = 40 bar

vertical, setpoint = 40 barhorizontal, setpoint = 45 barvertical, setpoint = 45 barhorizontal, setpoint = 51 barvertical, setpoint = 51 bar

curv

e1

curv

e2

curv

e3cu

rve4

curv

e5

curv

e6

2.5

0.0

2.5

5.0

T in

C

setpoint = 31 bar setpoint = 35 bar setpoint = 40 bar setpoint = 45 bar setpoint = 51 bar

Figure 22: Top: Temperature distribution along the cooling loop for different setpoints for both, horizontal(dashed) and vertical (dotted), orientations. Bottom: Temperature difference ∆T (Thor. − Tver.) along thecooling loop.



7 Summary and Outlook

As mentioned in the introduction, the project documented in this report evolves around the high-luminosity upgradeof the LHC and the necessary enhanced thermal resistance of the detectors. More specifically, this report treats IRmeasurements of a thermomechanical prototype related to the upgrade of the inner tracking detector in the ATLASexperiment. Following the rather detailed explanation in the foregoing sections, we give a summary of our results here.

We did a variety of studies on the theoretical conversion between emitted IR radiation and object temperature. Sofar, the results are little satisfying. But, we are talking to experts in the community and other teams involved in thethermal assessment of the future tracker.

We determined the emissivity of the paint in a range between 0.905 and 0.930. This result is coherent with ourexpectations but it needs to be narrowed down to achieve a temperature uncertainty of less than 1 ◦C. Given that theemissivity calculation is based on the conversion equation between radiant power and temperature, further explorationin this direction needs to be done.

The camera software seems to overall give good results with some error, e.g. it overestimates the known emissivity ofthe tape by 1 %.

The petal was fully prepared for the measurements. After testing the trustworthiness of glueing the Pt100s onto theSi, we successfully installed eight Pt100s on the petal. Additionally, we sprayed the petal black. And even though thedried coating looks sandy, it successfully removes most of the unwanted reflected IR radiations.

The measurements of the petal with turned-on electronics have started and so far we observe that the temperaturealong the cooling loop is stable (at least for the examined cases). The ongoing measurements need to be analysed andevaluated. The final results are planned to be presented at the preliminary design report (PDR) in the beginning ofOctober.


thermal performance of the petal prototype · picture of the tested petal prototype built at desy...

Documents