this talk reviews radiation detection physics with emphasis on high resolution gamma ray detectors using semiconducors

1. SEMICONDUCTOR RADIATION DETECTORS Eugene E. Haller University of California and LBNL, Berkeley, CA 94720 Fourth Summer School in Radiation Detection and Measurements, July 24-30, 2011 Munich, Germany

3. "Nuclear Radiation Detector Materials," eds. E.E. Haller, H.W. Kraner and W.A. Higginbotham, Mat. Res. Soc. Proc. Vol. 16 (North Holland, 1983).

4. "Detector Materials: Germanium and Silicon," E.E. Haller, IEEE Trans. Nucl. Sci. NS-29, No. 3, 1109-18 (1982).

6. Semiconductor Properties

7. Interaction of Radiation with Semiconductors

8. P-N-Junctions

9. Detector Materials: Ge, Si, (GaAs)

11. 7/19/11 E. E. Haller 5 J. Q. Philippot, IEEE NS 17(3), 446 (1970)

13. resistive detectors,

14. gas ionization chambers,

15. proportional counters, etc.f g

16. 7/19/11 E. E. Haller 7 The Ionization Chamber:Operating Principle The field lines of a charge in volume V end on the electrodes and "influence" an electric charge. The relative quantities of charge influenced on each electrode depends only on the geometry. When the charge moves (in an externally applied field), the relative quantities of influenced charge change, forming a displacement current. When the charge stops moving, the displacement current stops flowing. Charge may stop flowing because it gets "trapped" at a defect center or gets "collected" at a contact. In a typical p-n junction, charge ideally flows to one of the electrodes. In a resistive detector, charge arriving at one electrode may lead to the injection of a new charge at the opposite electrode. We define the "photoconductive" gain: Note: For a high resolution detector, a fixed, constant gain is required (for p-n junctions: G = 1.00; though there are ways to compensate incomplete charge collection) = average distance traveled by a free charge L = inter electrode distance

17. 7/19/11 E. E. Haller 8 SEMICONDUCTOR PROPERTIES Diamond Crystal Lattice: [100] Direction

18. 7/19/11 E. E. Haller 9 [110] Direction

19. 7/19/11 E. E. Haller 10 [111] Direction

20. 7/19/11 E. E. Haller 11 Silicon Real space structure, [100] projection (The sign = represents the two electrons in each bond.) Electronic band structure (CB = conduction band, VB = valence band)

21. 7/19/11 E. E. Haller 12 Doped Silicon Donor doping (e.g., phosphorus on Si sites) provides extra electrons which can move into the conduction band or compensate an acceptor; Acceptor doping (e.g., aluminum on Si sites) creates holes which can move into the valence band or compensate a donor

22. 7/19/11 E. E. Haller 13 Radiation detector performance is affected by temperature in several ways. Two important ones are: A. Generation of free electrons and holes by thermal ionization across the bandgap. This affects the bias current I (also called "leakage" or "dark" current) through the detector. This current is a source of electrical noise, in its simplest form "shot" noise ( df= 2eIdf; e = charge of the electron, f = frequency) B. Trapping and de-trapping of the free charge carriers at defect or doping related energy levels in the bandgap. Temperature Effects

23. 7/19/11 E. E. Haller 14 A. Free Carrier Concentration At finite temperatures a very small number of bonds in a crystal is broken. For each broken bond there exists a mobile electron and a mobile hole. The concentration of these electrons (n) and holes (p) is called intrinsic carrier concentration ni: ni2= NcNv exp (-Egap/kBT)= np (Law of mass action) Nc = conduction band effective density of states Nv = valence band effective density of states kB = Boltzmann's constant = 8.65 × 10-5eV/K Egap = energy gap (Si: 1.1 eV, Ge: 0.7 eV, GaAs = 1.42 eV, CdTe: 1.47 eV, Hgl2: 2.13 eV, AlSb: 1.58 eV) ni (Si, 300K) = 2 x 1010 cm-3 ni (Ge, 300K) = 3 x 1013 cm-3 ni (GaAs, 300K) = 107 cm-3

24. 7/19/11 E. E. Haller 15 Thermal Carrier Generation The thermal generation of electrons and holes gives rise to the reverse (also called "leakage" or "dark") current. This current constitutes a source of electronic noise. Consequences • Because of the small bandgap (Eg=0.7 eV), Ge diodes must be cooled in order to achieve sufficiently low reverse currents. • Si diodes (Eg=1.2 eV) have to be cooled only for high resolution applications. • CdTe, CdZnTe, GaAs, AlSb, ThBr, PbOand Hgl2 should not require cooling.

25. 7/19/11 E. E. Haller 16 Arrhenius plot (log conc. versus 1/T) of the intrinsic carrier concentration of the major semiconductors

26. 7/19/11 E. E. Haller 17 B. Charge Trapping Electrons and holes are also created in ionization events caused by radiation. In order to be detected they have to traverse the detector crystal all the way to their respective electrodes. On the way they encounter shallow and deep impurities, some of which may be charged. Trapping and release of charge from deep impurities is described as follows: emission rate e (per level) of a carrier to the nearest band: e =  <v> Nband exp (-E/kBT) with = carrier capture cross-section (cm2) <v> = average thermal velocity = (3kT/m*)1/2 Nband= effective density of states = 2 (2πm*kT/h2)3/2 kB = 8.65 × 10-5eV/K E = binding energy (depth!) of a particular level

27. 7/19/11 E. E. Haller 18 B. Charge Trapping (cont.) Shallow dopant levels in Si (E ~ 45 meV) or Ge (E ~ 10 meV) do not trap charge for any significant length of time at temperatures above 77 K (liquid nitrogen). Their binding energy is too small for trapping. However, already very small concentrations of deep levels (E > 100 meV) trap charge effectively and do not release it within the collection time. The fluctuations in this charge loss lead to asymmetric peaks in the energy spectrum.

28. 7/19/11 E. E. Haller 19 Deep Levels <—> Trapping 1: trapping of an electron by a deep level: 2: recombination of the trapped electron with a free hole or 3: detrapping after a time t.

29. 7/19/11 E. E. Haller 20 Electric Charge Transport At low electric fields the carrier velocity is proportional to the electric field: The mobility µ rises with decreasing temperature because the phonon density drops: At high electric fields velocity saturationoccurs. The carriers emit phonons* at a rapidly increasing rate and the velocity becomes almost constant. (See figures 11-2 on pages 341 and 342 in G. Knoll’s textbook). It is interesting to note that the saturation velocity for most semiconductors lies around 107cm/s. _______________________________ * Lattice vibrations (quantized E = hw)

30. 7/19/11 E. E. Haller 21 The Asymmetric Planar N+-P-Junction (parallel plate capacitor) Ionized positively charged donors Ionized negatively charged acceptors

31. 7/19/11 E. E. Haller 22 The ionized shallow impurity levels constitute a space charge e |NA - ND|. Poisson's equation relates the potential  to the charge: 2 = -e |NA - ND| /o  = dielectric constant, o = permittivity of vacuum In one dimension: ∂2/∂x2 = -e |NA - ND| /o integrating twice leads to:  = d2 |NA - ND| =d2|NA - ND| C CSi = 7.72  10-8Vcm CGe = 5.64  10-8Vcm with: d = width of the depletion layer Planar N+-P-Junction

32. 7/19/11 E. E. Haller 23 Planar N+-P-Junction For Ge junctions we find: V = d2 | NA - ND | x 5.64 x 10-8 (Vcm) and for Si: V = d2 | NA - ND | x 7.72 x 10-8 (Vcm) with V = applied voltage (V) d = depletion depth (cm) | NA - ND | = net-impurity concentration (cm-3) EXAMPLE: Planar P-I-N Detector d = 2 cm, V = 3000V What is | NA - ND | for full depletion? Ge: 1.33 x 1010 cm-3 Si: 9.7 x 109 cm-3 } These are extremely small concentrations compared to ~ 4 x 1022 Ge or Si per cm3!

33. 7/19/11 E. E. Haller 24 Ultra-pure Ge Crystal Growth Ultra-purity (~ 1 residual impurity in 1012 cm-3 Ge atoms!) requires highly specialized materials: e.g. Ultra pure silica for the crucible, ultra pure graphite for RF coupling, Ultra pure H2 atmosphere and ultrapure chemicals.

34. 7/19/11 E. E. Haller 25 Ge seed An ultra-pure Germanium single crystal is being “pulled” from a melt contained in a silica crucible at 936oC. The atmosphere is pure Hydrogen. Heat is supplied by the water cooled radiofrequency (RF) coil surrounding the silica envelope. This bulk crystal growth technique carries the name of it’s inventor, “Jan Czochralski.” Ge single crystal Ge melt Silica crucible RF coil

35. 7/19/11 E. E. Haller 26 The Floating Zone (FZ) crystal growth process is used for ultra pure silicon up to 6” in diameter. No crucible is used. The ambient is typically nitrogen but vacuum and argon have been used.

36. 7/19/11 E. E. Haller 27 Poly Si feed rod One turn RF coil Si single crystal Silicon Floating Zone (FZ) crystal with 10 cm diameter is being pulled. The single turn RF heating coil creates the liquid “floating zone” between the lower part (single crystal) and the upper polycrystalline section.

37. 7/19/11 E. E. Haller 28 Determining Ultra-Purity Electrical conductivity is an accurate measure of the net-impurity concentration. The intrinsic conduction can be “turned-off” through cooling to liquid nitrogen temperature (77K).

38. 7/19/11 E. E. Haller 29 Dopant Impurity Profiles Typical impurity profile of an ultra-pure Ge crystal. The acceptor aluminum (Al, dashed line) does not segregate in silica grown Ge, leading to a constant concentration. The donor phosphorus segregates (P, dotted line). In our particular example, the phosphorus concentration equals the aluminum concentration at 80% of the melt frozen and exceeds it beyond. In the Al dominated part, the crystal is p-type, in the P dominated type, the crystal is n-type. All ultra-pure Ge single crystals contain about 1014cm-3 H, O and C, inert impurities. Special thermal treatment can lead to the formation of shallow acceptors [A(Si,H) and A(C,H)], and donors (D(O,H). These trigonaldopants are not thermally stable and disappear at ~160oC.

39. E. E. Haller 30 Charge Collection in Real Time in a 1 mm thick Si p-n-Diode Courtesy J. Fink et al., NIM A 560, 435-443 (2006) n+ P+ α α n-Si + − min E-field max 7/19/11

40. E. E. Haller 31 Charge Collection in Real Time in a 1 mm thick Si p-n-Diode Courtesy J. Fink et al., NIM A 560, 435-443 (2006) α α n-Si + − (b) min E-field max 7/19/11

41. 7/19/11 E. E. Haller 32 The n+-i-p+ Diode Detector or

42. 7/19/11 E. E. Haller 33 Free Carrier Velocity as a Function of the Electric Field

43. 7/19/11 E. E. Haller 34 High-Purity Ge Detector Structures Planar high-purity germanium (hpGe) radiation detector After G. S. Hubbard, MRS Proc. Vol. 16, 161 (1983) Coaxial structure hpGe radiation detectors. Device (a) has standard electrode geometry (SEGe); detector (b) has been fabricated with reverse electrode geometry (REGe)

44. 7/19/11 E. E. Haller 35 Position sensitive orthogonal-strip hp Ge detector with 64 strip electrodes on each side fabricated using amorphous-Ge contacts. Strip pitch is 1.25 mm. The active volume of the detector is 9 cm X 9 cm X 2 cm. (Courtesy P.N. Luke, LBNL)

45. Is this the end of the story ? No, because ultra-pure Ge detectors offer the highest resolution gamma ray spectroscopy only when they are cooled. This requirement causes serious problems in certain applications especially in security surveying situations. There is a need for room temperature medium resolution detectors. Many groups have tested many materials but only a very small number look promising. What do we need: - High Z material which can stop gamma rays efficiently - A bandgap larger than ~1.7 eV so no cooling is required - A bandgap < 2.5 eV to keep the number of electron-hole pairs large for good resolution - High purity to minimize the trap concentration - large µτ (mu-tau) products for electron and holes (µ = mobility, τ = free carrier lifetime) 7/19/11 36 E. E. Haller

46. 7/19/11 E. E. Haller 37 Properties of Compound Semiconductors [Courtesy M.R. Squillanteet al., MRS Proc. Vol. 302, 326 (1993)]

47. Lattice Const. and Bandgapsof Important Semiconductors 7/19/11 38 E. E. Haller

48. What do we find: - there is no material matching all the above conditions - there are a few materials with at least one free carrier exhibiting good µτ (mu-tau) products The challenge: - Obtain good energy resolution with single carrier collection - Form a High Resistivity Detector instead of a p-n junction 7/19/11 39 E. E. Haller

49. E. E. Haller 40 Why do we want high resistivity? No Radiation, No Signal: I=0 - + High Bias In presence of radiation, strong signal: I + - High Bias 7/19/11 - +

50. 41 The “Right Kind” of Resistivity These should be high! These should be low! Three options for realizing low free carrier concentration: Achieve extremely high purity Materials: Only FZ- Si and high purity Ge 2) Deep level compensation Materials: GaAs,CZT and most other resistive material candidates 3) Shallow level compensation Materials:Ge:Ga,Li; TlBr(?) 7/19/11 E. E. Haller

51. 42 The “Right Kind” of Resistivity The Fermi level must be in the middle of the band gap for the highest resistivity possible at room temperature. The 3 ways to do this: 3. Shallow Level Compensation* 2. Deep Level Compensation 1. Ultra Purity CB EF VB 7/19/11 E. E. Haller * The idea of shallow level compensation or passivation is attractive because deep levels frequently negatively affect carrier mobilities. This mechanism provides a way to achieve low carrier concentrations without affecting mobilities. However, its realization with Li drifting of p-type Ge or Si is thermally unstable. This lead to the development of ultra-pure Ge

53. Energy Resolution of ~1%* obtained, close to that of Ge.

55. Pushing the energy resolution limit while simplifying electronics

57. Higher Density than CZT: 7.56 g/cm3 vs 5.29 g/cm3

58. ~Higher Resistivity than CZT: 1010-1011 ohm-cm

59. Band Gap: 2.68 eV, Indirect; 3.04 eV, Direct

60. Transport Properties comparable to CZTElectronic transport properties dramatically improved in last decade by RMD! TlBr detector, courtesy of RMD E. E. Haller 7/19/11

61. 7/19/11 E. E. Haller 45 The “Frisch Grid”: An Old Idea with Great Relevance and Promise • Gas ionization chambers suffer from the same problem as many semiconductor detectors:  of one charge species (ions) is significantly lower than of the other charge species (electrons). • Solution: the "Frisch Grid" (FG)

62. 7/19/11 E. E. Haller 46 Reincarnation of the Frisch Grid as two Sets of Interdigitated Contacts A and B P.N. Luke, Appl. Phys. Lett. 65(22) 2884 (1994)

63. 7/19/11 E. E. Haller 47 Cd0.8Zn0.2Te (5 x 5 x 5 mm3) Courtesy P.N. Luke, LBNL

64. 7/19/11 E. E. Haller 48 Courtesy P.N. Luke, LBNL

65. 7/19/11 E. E. Haller 49 Sub 1% Resolution CZT Detector

66. 7/19/11 E. E. Haller 50 CZT CPG vs Ge CZT coplanar-grid detector – 1 cm3 (cooled to -30 °C) Peak/Compton=16.2 Ge detector – 3 cm3 Peak/Compton=11.4 Courtesy P.N. Luke, LBNL

67. 7/19/11 E. E. Haller 51 Conclusions • The ideal semiconductor material for all radiation detection applications does not exist but it can be approached closely. • Certain material property requirements and application requirements are incompatible (e.g. bandgap: large for low leakage current but small for small energy per e/h pair). • Si & Ge are the high resolution spectrometer materials. Detectors exhibit excellent stability, good efficiency, timing, etc. • Thin epitaxial films (100-150 µm) of high-purity GaAs (|NA-ND|< 1012 cm-3) have been grown by the LPE technique and early spectrometer results look promising. In contrast, semi-insulating (SI) GaAs contains very large concentrations of deep traps which make the material highly resistive and lead to extreme charge trapping.

68. 7/19/11 E. E. Haller 52 Conclusions, cont. • CdTe, CdZnTe, ThBr& HgI2 are room temperature materials. Low energy X rays can be detected with good resolution. Medium and high energy photons ( rays) still pose problems. Trapping and poor hole transport seem to be fundamental and/or related to material inhomogeneitiesand defects. • Single-polarity charge sensing using coplanar electrodes (an analog of the Frisch grid in gas proportional counters) looks very promising for semiconductors with good collection of at least one type of charge carrier (typically electrons; see: P.N. Luke, Appl. Phys. Lett. 65, 2884 (1994) and more recent publications).

69. 7/19/11 E. E. Haller 53 Thanks for Your Attention