1. Photomultipliers: Eyes for your Experiment Example: photomultipliers  today: mature and versatile technology  vast range of applications  … and still gets improved by innovations Edinburgh, 23.07.2003 Stephan Eisenhardt University of Edinburgh extension of our senses new knowledge in fundamental processes new technology understand tools on fundamental level correct interpretation of results

2. 2 Edinburgh, 23.07.2003Stephan Eisenhardt o Photo emission from photo cathode Q.E. = N p.e. /N photons o Electron collection –Focusing optics –optimise efficiency –minimise transient time spread o Secondary emission from dynodes: è Electric potential è Electron multiplication dynode gain: g(E) = 3…50 o Total gain:G =  g i e.g. 10 dynodes with g=4  G = 4 10  10 6 Basic Principle Schematic of Photomultiplier Tube (Philips Photonic) Equi-potentials and trajectories in a fast input system

3. 3 Edinburgh, 23.07.2003Stephan Eisenhardt A Brief History o 1887: photoelectric effect discovered by Hertz o 1902: first report on a secondary emissive surface by Austin et al. o 1905: Einstein: "Photoemission is a process in which photons are converted into free electrons." o 1913: Elster and Geiter produced a photoelectric tube o 1929: Koller and Campbell discovered compound photocathode (Ag-O-Cs; so-called S-1) o 1935: Iams et al. produced a triode photomultiplier tube (a photocathode combined with a single-stage dynode) o 1936: Zworykin et al. developed a photomultiplier tube having multiple dynode stages using an electric and a magnetic field o 1939: Zworykin and Rajchman developed an electrostatic-focusing type photomultiplier tube o 1949 & 1956: Morton improved photomultiplier tube structure è commercial phase, but still many improvements to come

4. 4 Edinburgh, 23.07.2003Stephan Eisenhardt Extension of Vision o Spectral sensitivity: 400<S( )<750nm110<S( )<1600nm o Time resolution: ~50ms50ps…10ns o Spatial resolution: ~100 lines/mm (2500dpi) 2mm…50cm  Intensity range: O (10 16  /mm 2 s) (daylight) 1  … ~10 8  /mm 2 s single photon sensitivity (~1mA anode current for 1’’ tube) after adaptation  Life time: O (10 5 C/mm 2 70yrs) O (10C) for semi-transparent cathode o o Human eye: o o Photomultiplier:

5. 5 Edinburgh, 23.07.2003Stephan Eisenhardt PC vacuum glass Photoemission e-e-  PC semitransparent photocathode opaque photocathode  e-e- PC substrate o 2-step process: –photo ionisation –escape of electron into vacuum o o Multi-reflection/interference due to high refractive index bialkali: n(  = 442 nm) = 2.7 o o Q.E. difficult to measure   often only an effective detection efficiency determined: + +internal reflection from a metallic surface + +collection of the photoelectrons + +electronic theshold…

6. 6 Edinburgh, 23.07.2003Stephan Eisenhardt E e [eV] h [eV] Spectral Response e-e-  : E = h o o Photo electric effect – –h > E g + E A + dE/dx – –E e = h - W - dE/dx o o Spectral response: – –Quantum efficiency – –Cathode sensitivity o In the band model : Alkali Photocathode

7. 7 Edinburgh, 23.07.2003Stephan Eisenhardt Alkali Photocathodes Q.E. (Philips Photonic) S( ) [mA/W] [nm] semitransparent photocathodes transmittance of window materials LiF [nm] T [%] Bialkali : SbK 2 Cs, SbRbCs Multialkali : SbNa 2 KCs Solar blind : CsTe (cut by quartz window) human sensitivity (arb. units) human eye (shape only)

8. 8 Edinburgh, 23.07.2003Stephan Eisenhardt 0.4 absorption coefficient  [cm -1 ] [nm] h [eV]   4 · 10 5 cm -1   25 nm A = 1/    1.5 · 10 5 cm -1   60 nm Photocathode Thickness Blue light is stronger absorped than red light! o semi-transparent cathodes è best compromise for the thickness of the PC:  photon absorption length A (E ph )  electron escape length E (E e ) o o Q.E. of thick cathode: red response  blue response  o o Q.E. of thin cathode: blue response  red response 

9. 9 Edinburgh, 23.07.2003Stephan Eisenhardt Alkali Photocathode Production o evaporation of metals in high vacuum –< 10 -7 mbar –< 10 -9 mbar H 2 O partial pressure –no other contaminants (CO, C x H y...) èbakeout of process chamber (>150 o C) and substrate (>300 o C) o condensation of vapour and chemical reaction on entrance window è relatively simple technique d Sb  5 nm Q.E.  3 - 5 % Q.E.  15 - 20 % o o Example: SbK 2 Cs simplified sequential bialkali process

10. 10 Edinburgh, 23.07.2003Stephan Eisenhardt Phototube Fabrication o internal: Indium seal Hot glass seal o o external: CERN/LHCb evaporation and encapsulation facility

11. 11 Edinburgh, 23.07.2003Stephan Eisenhardt Semiconductor Photocathodes o high Q.E. and spectral width o negative electron affinity o  complex production Q.E. [%] [nm] (Hamamatsu) molecular beam epitaxial growth 10 -10 mbar a) grow PC on crystalline substrate b) create interface layer c) fuse to entrance window d) etch substrate away a)b)c)d)

12. 12 Edinburgh, 23.07.2003Stephan Eisenhardt secondary e - VV dynodes Secondary Emission o alloy of alkali or earth-alkali and noble metal o alkaline metal oxidises  insulating coating –Large gain –Stability for large currents –Low thermal noise o Statistical process: Poisson distribution –Spread (RMS) –Largest at 1 st and 2 nd dynode o typically 10 … 14 dynode stages o linearity limits: –pulse:space charge –DC:photo current << bleeder current bleeder chain scheme of external circuit for dynode potentials electron multiplication

13. 13 Edinburgh, 23.07.2003Stephan Eisenhardt venetian blind box and grid linear focusingcircular cage Classic Dynodes o o compact o o fast time response o o best photoelectron collection efficiency o o good uniformity o o excellent linearity o o good time resolution o o fast time response o simple design o good for large PC  sensitive to Earth B-field (30-60  T)! no spatial resolution o o uniformity:  o o photoelectron collection efficiency: 

14. 14 Edinburgh, 23.07.2003Stephan Eisenhardt o linear focusing B||z B||x B||y  50  T  0.5 Gauss B||x  200  T 2 Gauss o o venetian blind Sensitivity to Magnetic Fields z x y o o Earth field: 30-60  T è è requires  -metal shielding

15. 15 Edinburgh, 23.07.2003Stephan Eisenhardt Modern Dynodes foil microchannel plate proximity mesh o o excellent linearity o o excellent time resolution: transit time spread: 50ps o o B-field immunity up to 1.2T B-field o o spatial resolutionvia segmented anode o o good e - collection efficiency o o excellent time characteristics o o stable gain o o …up to 20mT o o low cross-talk pixels metal channel o o good uniformity o o poor e - collection efficiency!  

16. 16 Edinburgh, 23.07.2003Stephan Eisenhardt segmented anode Multianode PMT o Position sensitive PMT –8x8 metal channel dynode chains in one vacuum envelope (26x26 mm 2 ) –segmented anode: 2x2 mm 2 –active area fraction: 48% o UV glass window o Bialkali photo cathode –QE = 22…25% at = 380 nm o Gain – 3. 10 5 at 800 V o Uniformity, Crosstalk –much improved o Applications: –medical imaging –HERA-B, LHCb: Ring Imaging Cherenkov counters MaPMT window & electron focusing pixel active area Relative distance [0.1 mm] pulse charge p.e. probability

17. 17 Edinburgh, 23.07.2003Stephan Eisenhardt PMT Characteristics o Fluctuations –number of secondary electrons –Poisson distribution o Saturation –space charge –large photon current o Non-linearity –at high gains o Stability –drift, temperature dependency –fatigue effects èMonitoring o Sensitive to magnetic fields –Earth: 30-60  T èrequires  -metal shielding o o single photon events to oscilloscope (50  ) 5mV 1ns charge integration  pulse height spectrum MaPMT

18. 18 Edinburgh, 23.07.2003Stephan Eisenhardt pulse amplitude 1 photoelectron 2,3 photoelectrons fit to data MaPMT   first dynode gain:  = 4 – –broad signal distributions – –Poisson shape for n pe =1,2 – –significant signal loss below threshold cut Pulse Height Spectrum  first dynode gain:  = 25 –clear separation of n  signals –Gaussian shape –Poisson for  probability pulse amplitude (Gilmore) o o Photo conversion possible at 1st dynode – –Gain reduced by  1 – –explains data between noise and 1 p.e. o o contradiction when HV is changed!! – –this data is contaminated with incompletely sampled charge pulses

19. 19 Edinburgh, 23.07.2003Stephan Eisenhardt Conclusion o know your tools o don’t fool yourself with immature conclusions o always cross-check as far as possible o o photomultipliers are a mature and versatile technology o o suitable for a vast range of applications in light detection: – –Spectral sensitivity: 110<S( )<1600nm – –Time resolution: down to 50ps – –Intensity range: 1  … anode current O (1mA) – –Spatial resolution:down to 2mm with low cross-talk – –Magnetic field:up to 1.2T o o development goes on… flat panel PMT: next generation MaPMT

20. 20 Edinburgh, 23.07.2003Stephan Eisenhardt Super-Kamiokande: 20’’ tubes