1. UK Calibration Activities
Lee Thompson
University of Sheffield
3rd Hyper-Kamiokande EU Open Meeting
CERN
27th -28th April 2015

2. Calibration topics PMT gain calibration PMT timing calibration
Water attenuation length measurement
Water scattering measurement
Simulating the Cerenkov ring
Need to understand:
Coverage
Redundancy
Intensity
Wavelength
Light profile

3. Proposed systems (UK) Pulsed light sources
Short duration light pulses from LEDs
Light coupled into optical fibres (possibly via lens and standard fibre)
Graded index multimode fibre from pulser to injection point
Fibre ends inject light directly into the detector (via a diffuser)
Illuminate multiple PMTs on other side of a tank
Continuous low pulse rate operation during data taking
No risk to the detector (mechanical, radioactivity)
Electronics (which may require intervention) is easily accessible
Pseudo-muon light source
Create a Cerenkov light cone at a fixed angle
Mimic expected Cerenkov spectrum
No use of radioactive source to generate Cherenkov light

4. Calibration requirements
PMT amplitude and timing
monitoring of t0 of the pulse, width of the pulse
diffuser profile - needs to be uniform in terms of: time, width, intensity, angular occupancy (a single PMT should be hit by >4 diffusers)
dynamic range (to permit single and multiple photoelectron measurements)
single wavelengths
Scattering in water
multiple wavelengths down to ~300nm
narrow beams of order  1 degree (no diffuser?)
monitoring of intensity

5. Calibration requirements
Attenuation in water
multiple wavelengths down to ~300nm
intensity monitor
fine control over light level
wider beam than for scattering - or order 30 degrees (needs simulation studies), uniform in intensity and angle
good pulse-to-pulse reproducibility

6. Pulsed light sources: Why LEDs?
Pros
LEDs are cheap (few $/£/€ each), low cost driver electronics
Exist in the wavelength range of interest to us
Possible to cover a wide spectral response with multiple LEDs
Spectral emission in a narrow wavelength range (e.g nm)
Can be pulsed at few nanoseconds pulse width
Number of photons per pulse at fibre end from 103 to 105
Significant UK expertise in using LEDs for calibration systems (ANTARES, SNO+)
In principle relatively easy to couple to optical fibres to deliver light to a remote location

7. Pulsed light sources: Why LEDs?
Cons (LEDs)
LEDs aren’t designed to be pulsed, no guarantee that a particular LED will deliver ns pulses – need to identify suitable models
Variations within a batch of nominally identical LEDs
Higher drive currents required to illuminate compared to laser diodes
Coupling of fibre to LEDs leads to large optical power losses, need to improve on current method for small core diameters
Cons (Laser diodes)
high cost of laser diodes with individually optimised coupling into graded index fibre
laser mode hopping – need to strictly control temperature and current to avoid this, leads to amplitude variations  
coherent light, possible unwanted interference patterns in light output, non uniformity within light cone 
fast laser modulation relies on a bias current – may lead to unitended light emission between pulses.

8. Pulsed light sources: flow diagram

9. Pulsed light sources: illumination mode

10. Pulsed light sources: illumination mode
Features
Front end uses programmable FPGA which controls multiple LED driver (pulser) circuits on one board
LEDs will be coupled to graded-index multi-mode optical fibre, low dispersion (100ps per 100m) and low attenuation. Large core step index fibre would lead to unacceptable 14 nanosecond optical pulse width over 100 m.
Fibre to be coupled to a suitable diffuser at the detector end for PMT and water attenuation measurements, different or no diffuser for scattering measurements
Aims
Inject of order photons per flash
Minimum optical pulse width (few nanoseconds)
Minimize losses between LED and diffuser

11. LED, fibre and couplings
LED will be directly coupled to the graded index fibre either
via a drilled hole in the LED plastic lens
via a step index fibre and lens
First option enables the fibre to be placed close to the diode emission region
Graded-index multimode fibre small diameter (60 micrometer) presents a challenge for efficient coupling, various coupling mechanisms under consideration

12. Pulsed light sources: monitoring modes
Propose to monitor the light flashes at 2 different points in the chain:
At the LED
Will have a small SiPMT coupled directly to the LED to measure the characteristics of each individual light pulse.
At the diffuser
Directly coupled return fibre of the same type as delivers the light to monitor the average optical pulse shape and round trip time.
Both systems will constantly monitor light pulse characteristics, e.g. amplitude, timing, pulse-to-pulse reproducibility, longevity, variation with environmental parameters,
Known and stable pulse characteristics essential, e.g. to measure water properties including attenuation and scattering

13. Pulsed light sources: monitoring at the LED

14. Pulsed light sources: monitoring at the diffuser

15. LED pulser designs: status
Three different LED pulser designs under consideration based on MOSFETs and BJT (bipolar junction transistors):
Quad MOSFETs in a H-bridge arrangement with on chip driver circuits
Quad OpAmps – may struggle to deliver sufficient current
Zener diode
Monitoring circuit also prototyped
White Rabbit tests underway

16. LED pulser designs: status
The baseline design is a quad-MOSFET circuit. A dual MOSFET driver in SNO+ gave < 1 ns at 104 photons per pulse into the fibre
Investigating discrete transistors (BJT) circuits as alternative
Several prototype LED pulsers now constructed and under test

17. LED couplings: status
Have checked l460nm light over 100m of graded index multimode fibre.
Series of tests to scan the light profile across a standard SM LED (that can be pulsed) in progress
Using a 3D scanning stage with 0.01mm positioning accuracy and Thorlabs optical meter
300 micron beam spot close to device
These tests will help to guide the final decision on which coupling method between the LED and graded index fibre

18. Pseudo-muon light source
Objective is to inject a Cerenkov-like cone of light into the detector
Simple application of refraction across a boundary of different refractive indices
Can be achieved using a short, narrow transparent (acrylic) tube along with a light source which produces almost parallel light
Different muon momenta can be achieved by using different lengths
In the limit of q1 -> zero and n1 -> b (≈1) then q2 -> qCerenkov
Independent of ni and wavelength
2.5m line source simulates 500MeV muon
Vertical scale ≠ Horizontal scale

19. Pseudo-muon source: next steps
Optical simulation of the source to drive the design
Identification of a suitable selection of LEDs that, when their output light is convoluted, will give a good approximation to a Cerenkov spectrum
Tests of a prototype in a large water tank in the UK to evaluate angular profile and optical output
Ultimately deploy in 1 kton prototype

20. Simulation work: scattering calibration
Fibres mounted on PMT support structure with collimated (~few degree opening angle) beams fired across detector
Fast pulses (~2ns, LED or laser) at 4-5 different wavelengths to probe λ4 Rayleigh scattering dependence.
Operate at low intensity so in-beam PMTs mostly in single pe regime.

21. Simulation work: scattering calibration
Analyse through summation of hits over run
Use timing and spatial cuts to select analysis regions dominated by direct, reflected and scattered hits.
Scale scattering in MC to match data scattered hits
Use in-beam hits for normalisation

22. Simulation work: PMT timing calibration
Propose similar approach to SNO+
SNO used centrally deployed isotropic laserball
SNO+ will use embedded fibres, ~30 degree opening angle, sharp LED pulses
To account for different delays for each LED driver relative to trigger, ensure each PMT sees light from ≥ 2 fibres. (more for redundancy)
Cycle through all fibres at high rate
Scan LED intensities to sample whole charge range.

23. Simulation work: PMT timing calibration
Plot PMT hit times vs charge for each channel
Linear offset = PMT cable delay
Low charge turn up = time walk effect

24. Timelines 2015 2016 2017 LED PULSER BOARDS, TESTING
LED PULSER BOARDS, TESTING
Board Prototyping
Pulse characterisation with LEDs
Final board choice and tests
LED – FIBRE COUPLING, DIFFUSERS
Optimise LED-fibre coupling
Diffuser performance and choice
Characterisation of light pulses in water
PSEUDO-MUON SOURCE
Optical modelling and simulations
Design and build prototype
Water tank tests
SIMULATIONS

25. Summary and next steps
UK activities in this area have started in earnest
A large number of UK universities involved:
Liverpool, Sheffield : LED pulsers, characterisation, testing
Warwick: LED-fibre couplings
Lancaster: DAQ-calibration interface
Imperial, Edinburgh, Liverpool, Sheffield: Muon source
QMUL: simulations
Initial simulation studies underway
First prototype LED pulsers now constructed and under evaluation
LED-fibre coupling schemes under study