Collimated Beam Projector status

LSST-France, LPSC, Grenoble, june 2023

09/06/2023

T. Souverin, J. Neveu, M. Betoule, S. Bongard, S. Brownsberger, J. Cohen Tanugi, S. Dagoret Campagne, P. Fagrelius, F. Feinstein, P. Ingraham, C. Juramy, L. Le Guillou, A. Le Van Suu, P. E. Blanc, F. Hazenberg, E. Nuss, B. Plez, E. Sepulveda, K. Sommer, C. Stubbs, N. Regnault, E. Urbach

• Introduction

2

Introduction: what is a CBP ?

It is composed of two parts:

​

• a tunable monochromatic light
• an optical device able to form a parallel beam

3

The goal is to mimic a monochromatic star of known flux, to calibrate the response of a telescope and its filters.

​

CBP, for Collimated Beam Projector, is a device able to shoot:

• a known quantity of photons
• at a known wavelength
• and in a parallel beam.

​

​

​

Introduction: what is a CBP ?

4

#photons

at λ

#ADU

Telescope response:

R(λ) = #ADU / #photons

Introduction: what is a CBP ?

5

#photons

at λ

#ADU

Telescope response:

R(λ) = #ADU / #photons

Setup device

6

CBP optics

CBP response measurement

StarDICE response measurement

Laser

Integrating sphere w/ monitoring instruments

How do we measure our responses ?

7

CBP response R_CBP [𝛾.C⁻¹]

StarDice response R_SD [ADU.𝛾⁻¹]

• Q_solar : solar cell charges [C]
• Q_phot : photodiode charges [C]
• Q_ccd : stardice charges [ADU]
• 𝜖_solar : solar cell quantum efficiency [C.𝛾⁻¹]
• e = 1.6x10⁻¹⁹ [C]

Setup device

8

R_CBP

CBP response measurement

StarDICE response measurement

Q_phot

Q_phot

Q_ccd

II. Instruments

9

Integrating sphere

10

Two instruments in the integrating sphere, to monitor the input light:

• a spectrograph to monitor the laser wavelength
• a photodiode to monitor the flux quantity

• Spectrograph

11

Spectrograph wavelength calibration

• Acquisition of Hg-Ar spectra before and after CBP run
• Apply line detection

​

​

• Fit 3rd order polynomial using SNR>20 lines to map detected and tabulated wavelengths, with uncertainties
• Rescale uncertainties to get reduced chi2~1

12

Wavelength calibration total uncertainties

13

• Total uncertainties globally below 0.1nm, dominated by calib uncertainties

Angstrom level

b. Photodiode

14

Monitoring photodiode

15

• Photodiode plugged to the integrating sphere and connected to a electrometer
• Monitor the total charge collected in the photodiode Q_phot in Coulomb

c. Solar Cell

16

CBP output with Solar Cell

17

• Large solar cell calibrated with a NIST photodiode
• Measure the photons at the output of the CBP

CBP output with Solar Cell

18

• Large solar cell calibrated with a NIST photodiode

d. StarDice telescope

19

StarDice telescope

20

• Find spot position on camera
• Aperture photometry with dark subtraction and ghost contamination evaluation
• 3 pinholes: 75um, 2mm and 5mm

⇒ Measure Q_CCD the photons collected by the camera in ADU

III. Measurements

21

Plan

Measurements in different conditions to evaluate systematics and make pupil stitching :

• Spectrograph calibration
• CBP response :
• Solar Cell measurement ; 5mm pinhole
• Long and short distance (~16cm difference) ; 5mm pinhole
• Cap on the CBP to measure ambient light
• StarDice response :
• Same position ; every camera filter ; 75µm, 2mm, 5mm pinhole
• 8 positions on the mirror ; 75µm pinhole (“pupil stitching”)
• 4 positions on different quadrants but same radius
• 4 positions at different radius but same quadrant
• (4x4) positions on the CCD ; 75µm pinhole

​

​

​

​

22

Solar cell

Solar cell

a. CBP response

23

CBP response error budget

​

​

• statistical uncertainties below 0.1%
• solar cell calibration at 0.1%
• main systematics comes from scatter light (diffused light on mirror irregularities, surfaces, etc…)

​

24

CBP transmission, 5mm

25

Solar Cell measurement ; 5mm pinhole

b. StarDICE response

26

StarDice response, 5mm

27

Image for 5mm pinhole for light at 841nm

• Statistical precision around 0.4% for [400 - 1000] nm

StarDice filters transmission, 75µm

28

Filter leakages

​

​

​

Detection of out-of-band leakages below 0.1% level

​

29

Filter edges

30

Visible blue-shift of the filter edges when going to high incident angles

StarDICE grating transmission, 75µm

31

Grating → disperse light to observe absorbing rays

• Uncertainty around 0.3% for 1st order in [400 - 1000] nm range

Summary of the laser CBP

• Proof of concept: 5 per-mil calibration of the CBP itself
• main systematics can be under control in future versions
• precise wavelength calibration and resolution
• measurement of filter leakages below 1e-3
• Last task to do before paper publication: write the optical model to synthesize the full pupil transmission
• We gain a huge expertise on using a CBP for per-mil telescope throughput measurement
• transpose it to the qualification of the Rubin CBP
• build a Travelling CBP to measure telescopes on-site, like AuxTel, ZTF, Subaru, etc…

32

IV. RubinCBP @ Tucson

33

Rubin CBP

34

Solar cell

Counterweight

Integrating sphere

CBP telescope

Laser fiber

Photodiode

Rubin CBP masks

• Multiple pin-holes to shine different parts of the filters and focal plane at the same time => constellation of artificial stars

​

• Masks can be rotated, made by Harvard or Tucson mechanical workshops

​

• Still needs to decide how the constellation must look like :
• a constellation of stars
• with one star per amplifier/CCD
• not too small
• not too big

35

First data set

We are now able to pilot all Rubin CBP instruments and make tests to calibrate them.

36

Simulation of masks

Real ray tracing simulation !

37

Simulation of 50µm pin-holes

38

50μm pinhole fills ~⅓ of an LSST amplifier

CBP+LSST magnification factor is ~16

(24000 rays traced in 3.5s)

V. Travelling CBP @ LPNHE

39

Travelling CBP

Travelling version to work in dome conditions to calibrate AuxTel, ZTF, Subaru,... throughputs.

Replace the class 4 laser by a conventional intense lamp + (double)-monochromator

Under construction with Kélian Sommer (LUPM)

40

Xenon lamp

mono

chromator

CBP telescope

June 2022

Travelling CBP

Travelling version to work in dome conditions to calibrate AuxTel, ZTF, Subaru,... throughputs.

Replace the class 4 laser by a conventional intense lamp + (double)-monochromator

Under construction with Kélian Sommer (LUPM)

41

Laser-Driven Light Source

mono

chromator

June 2023

Thanks for your attention

42

IV. Major corrections

43

a. 532nm contamination correction

44

Spectrograph wavelength calibration

45

• Load the parameters, make the mapping and propagate detection uncertainties plus calibration uncertainties
• contribution of 0.03nm systematic uncertainty from calibration in the visible range, more outside

​

532nm contribution : extraction

46

• ⍺(λ) = Q_SP,532/Q_SP(λ) in the [532 - 644] nm range

​

• Within the [560 - 644] nm range, Q_SP,532 and Q_SP(λ) are well separated so we can make a good estimation of ⍺

​

• Below 560 nm, the shape of the psf in the spectrograph induces a superposition between the two peaks

​

⇒ We fit the values of ⍺(λ) between [560-644] nm for all the runs at a given QSW and we extrapolate in the range of [532-560] nm

532nm contribution : extraction

47

• ⍺(λ) = Q_SP,532/Q_SP(λ) in the [532 - 644] nm range

​

• Within the [560 - 644] nm range, Q_SP,532 and Q_SP(λ) are well separated so we can make a good estimation of ⍺

​

• Below 560 nm, the shape of the psf in the spectrograph induces a superposition between the two peaks

​

⇒ We fit the values of ⍺(λ) between [560-644] nm for all the runs at a given QSW and we extrapolate in the range of [532-560] nm

532nm correction : application

48

532nm contribution : g filter demonstration

49

g filter : cut after ~560nm

​

→ we don’t see the main wavelength light, but only the 532nm contribution

532nm contribution : g filter demonstration

50

b. Ghost correction

51

Ghost photometry

52

Ghost correction

• Φ₀ : Main spot flux
• Φ_G : Ghost flux
• ​
• 75µm photometry : Φ_tot = Φ₀
• 5mm photometry : Φ_tot = Φ₀ + Φ_G

We consider that the contribution of the ghost is a function of lambda f(λ) :

Φ_G = f(λ) x Φ₀

We can deduce the main spot contribution for the 5mm pinhole :

Φ₀ = Φ_tot/(1+f(λ))

53

Ghost photometry : looking for the masks

Produces a stack of all the similar datas

→ Create a mask for the main spot and the ghost

54

Ghost photometry : fitting positions

55

• Find the barycenter of the main spot and mask it
• Fit the best position for the ghost with a gaussian filter step by step

Ghost photometry : fitting positions

56

Ghost mask

Data with main spot masked

Fit with a high sigma → reduces the sigma and fit again → until sigma=1

Ghost photometry : background subtraction

57

Find ghost position and draw the vertical symmetric according to the main spot position

→ calculate the mean of the symmetric photometry and subtract it

Ghost photometry : same mirror positions

58

Ghost photometry : different radius positions

59

Ghost photometry : spline with all data (except radius 1)

60

Ghost photometry : spline with all data (except radius 1)

61

Ghost photometry : spline with all data (except radius 1)

62

Ghost photometry : IR oscillations

63

Ghost photometry : IR oscillations

64

c. Intercalibration 5mm/75µm

65

Intercalibration 5mm/75µm : goals

66

The pinholes are not the same when we shoot in the Solar Cell (5mm) or in the StarDice telescope (75µm)

• See if the ratio between the 5mm and 75µm pinhole is flat or not
• Understand the ghost contribution in the 5mm case
• Correct the ghost contribution thanks to the analysis with the 75µm

Growth curve

67

Why this growth after 250 pixels radius even when there is nearly no ghost ?

Growth curve : log scale

68

Why this growth after 250 pixels radius even when there is nearly no ghost ?

Growth curve : beyond 250 pixels radius

69

Same image but in logarithm and with a vmax value to see the contrast

​

3 visible elements :

• The border of the pinhole are faint
• The diffusion over the diaphragm
• The ghost in the left

⇒ Present at all wavelengths, so why is it higher in IR ?

Growth curve : evolution of the 5mm hole

70

Ratio 75µm/5mm

71

Ratio 75µm/5mm decrease in the IR, it can be either :

• 75µm decreases → not what we observe
• 5mm increases → what we observe

Ratio 75µm/5mm

72

Radius

73

Rubin CBP masks

Numbers to have in mind:

• CBP focal length is 625mm, primary mirror diameter is 330mm
• LSST + CBP magnification factor is 16 (LSST focal is 8.4m x 1.2 = 10.1m):
• 100um on mask = 1600um on LSSTCam = 160 pixels = ~⅓ amplifier
• max size to have a spot that fit an amplifier + annulus to estimate dark is 150um
• mask holder inner part is 25.4mm wide
• LSST full focal plane projected on mask is 600mm/16 = 38mm
• N_CCD = 189, N_amplifiers = 8 N_CCD = 1512 => constellation of N_pinholes = 1512 stars
• Effective pinhole diameter : D_eff = sqrt(N_pinholes) D_{1 pinhole}
• 1 star per amplifier : D_eff = sqrt(1512)x150um = 5.8mm
• 1 star per CCD : D_eff = sqrt(189)x150um = 2.1mm