3D-RID The application of micro- machining to radiation imaging detectors

3D-RIDVal O’SheaComo 9/10/’03 3D-RID -- a detector development project funded by the European Commission Partners:Applied Scintillation Technologies (UK) Czech Technical University (CZ) Metorex Oy (FI) Mid Sweden University (SE) Royal Institute of Technology(SE) Surface Technology Systems(UK) University of Freiburg(DE) University of Glasgow(UK)

3D-RIDVal O’SheaComo 9/10/’03 Outline of talk: Motivation for 3d -- what is it? Simulation Processing -- micromachining Devices Conclusion

3D-RIDVal O’SheaComo 9/10/’03 3D type detectors X-ray photon Legend: BC = Back contact ROIC = Read out circuit 1. Scintillator2. Scintillator 3.Neutron absorber 4.Directpn-junction light guiding to CCDpn -junction detection pn -junction detection detection p p E E n depth Si CCD CsI Si pitch X-rayphoton Si ROIC CsI BC Neutrons Si ROIC LiF BC ROIC Particles + - BC low doped Si N N P P Silicon or GaAs ROIC BC + - depth pitch

3D-RIDVal O’SheaComo 9/10/’03 The electrodes are cylindrical and are biased to create an electrical field that sweeps the charge carriers horizontally through the bulk. The electrons and holes are then collected at oppositely biased electrodes. Traversing ionising radiation creates electron-hole pairs in proportion to the energy deposited in the material. The free charges then drift through the device under the influence of an applied electric field. Proposed by S.Parker et al, Nucl. Instr. And Meth. A 395 pp (1997). Equal detector thickness W 2D >>W 3D ionising radiation E h + e - E -ve +ve n n + p + -ve W2DW2D x 2D 3D 3D Geometry

3D-RIDVal O’SheaComo 9/10/’03 ISE Simulation Electric Potential Distribution 50  m pitch 300  m thick silicon doping: 1x10 19 /cm 3 10  m pore diameter 50 x 90  m pitch 300  m thick silicon

3D-RIDVal O’SheaComo 9/10/’03 ISE Simulation Electric field Distribution 50V bias

3D-RIDVal O’SheaComo 9/10/’03 ISE Simulation

3D-RIDVal O’SheaComo 9/10/’03 2D MEDICI software package  Si material parameters.  Schottky barrier height= 0.65eV.  SRH statistics +impact ionisation.  No charge trapping or surface current.  Sidewall damage caused by reactive ion etching, N D =10 17 atm./cm 3. Simulated cell Defect distribution MediciSimulation

3D-RIDVal O’SheaComo 9/10/’03 3D- Detector Structure Unit cell Pixel n+n+ p+p+ 50  m 35  m p+p+ p+p+ n+n+ n+n+ 50  m

3D-RIDVal O’SheaComo 9/10/’03 Simulated Depletion Voltages Si N D [cm -3 ] Pixel size [  m] V depl. [V]

3D-RIDVal O’SheaComo 9/10/’03 –Connection from the readout & detector die combination to the communication / utility circuitry Basic image cell : vertical hybridisation of detector, read-out electronics & Interface circuitry & connector 3D-Stacked Imaging Tiles Courtesy Eric Beyne IMEC

3D-RIDVal O’SheaComo 9/10/’03 Tiling of 3D-image stacks Tiling on frame : –flexible design, –high accuracy alignment Courtesy Eric Beyne IMEC

3D-RIDVal O’SheaComo 9/10/’03 Charge Sharing and Photon Counting Medipix2 Schematic Courtesy Xavi Llopart and the Medipix Collaboration

3D-RIDVal O’SheaComo 9/10/’03 Conventional ICP Source

3D-RIDVal O’SheaComo 9/10/’03 Process Results - Si 10  m diameter pore 50  m pitch 8  m thick PR mask <1% exposed area Results: 182µm depth 40:1 selectivity to PR 2.0µm/min

3D-RIDVal O’SheaComo 9/10/’03 Process Results - Si 10  m diameter hole 50  m pitch 7-8  m thick remant PR mask <1% exposed area SF 6 / C 4 F 8 Process ASE-HRM Source

3D-RIDVal O’SheaComo 9/10/’03 Process Results - Advanced GaAs Etch (AGE) Process in ICP Cl 2 / SiCl 4 + CH 4 Switched Process Profile Angle =89.6°, Etch Rate = 1.25  m/min, Selectivity=6.7:1, Depth = 37  m

3D-RIDVal O’SheaComo 9/10/’03 Process Results - GaAs Scallops visible in walls Anisotropic profile

3D-RIDVal O’SheaComo 9/10/’03 TOPS:The Strathclyde Electron and Terahertz to Optical Pulse Source Strathclyde University.. 3mJ laser pulses. 40fs pulse duration. 1kHz pulse repetition rate. 400nm wavelength. Ti:Sapphire laser Femtosecond laser  Cold processing.  Low shockwave damage.  Tapering.  Repeatability.  Surface debris. Advantages: Disadvantages: Laser drilling

3D-RIDVal O’SheaComo 9/10/’03  diameter : entrance hole:10  m. exit hole: 6  m.  depth: 200  m. Laser drilling

3D-RIDVal O’SheaComo 9/10/’03 Electrochemical Etching Chemical reactions between Si atoms, HF solution and “h +” (positive charge carriers) HF concentration Current density Applied bias Light intensity Temperature Silicon properties (type, orientation, resistivity) Setup - Parameters

3D-RIDVal O’SheaComo 9/10/’03 Macropore formation: Principle SCR (electric field) in silicon h + directed towards the pore tips  Anisotropic etching Passivation of the pore walls is due to the SCR (Lehmann’s model)  Importance of the silicon resistivity ! Condition for stable pore formation  J at the pore tip = J ps  Possibility to control the diameter

3D-RIDVal O’SheaComo 9/10/’03 Preprocessing for electrochemical etching Front side: Inverted pyramids  focus the current lines (h + ) at the top of the pyramids Back side: Formation of an Al grid  uniform back side contact (especially for high resistivity sample) Al Si Photolithography + KOH etching Al deposition + photolithography + Al etching

3D-RIDVal O’SheaComo 9/10/’03 Formation of pores by EE: Large pores Spacing = um; Depth = um; Wall thickness  4 um;Active area > 80%.

3D-RIDVal O’SheaComo 9/10/’03 Images of the samples D32 and D41 (  = 2-5 k  cm) Demonstrators - Thindrill Depth 200  m dia: Depth 360  m dia: 10-16

3D-RIDVal O’SheaComo 9/10/’03 Formation of pores by EE: Thindrill

3D-RIDVal O’SheaComo 9/10/’03 Large pitch (45 or 30 µm)  High resistivity (  = 2-5 k  cm) Depth: 380 µm, wall: 4 µm, Pitch: 30 µm, aspect ratio  100! Large pores - Sloped Walls Depth: 180 µm, pitch: 30 µm, diameter from 6 to 18 µm. A linear function was applied to the current. i (mA) depth 26 3

3D-RIDVal O’SheaComo 9/10/’03 Diffusion 1: 1150ºC, 1h45’ Profile along A A 5 µm AFM SSRM Thickness at the pore bottoms: 3  m. Thickness on a planar wafer (SIMS): 6  m. Transport of boron down to the pore bottom may be limited. Formation of pn junctions: Results from diffusion

3D-RIDVal O’SheaComo 9/10/’03 Diffusion 1: SIMS profiles at different positions along the pore depth: - No B in the substrate (profiles c, g). Walls fully doped. Formation of pn junctions: Results from diffusion

3D-RIDVal O’SheaComo 9/10/’03 Diffusion 2: 1050ºC, 1h10’. SIMS profiles at different positions along the depth: - [B] in pores  [B] in a planar sample; no significant variation along pore depth. - Boron atmosphere in the pores maybe more uniform at 1050ºC than at 1150ºC. - Boron layers on each side of the walls. Formation of pn junctions: Results from diffusion

3D-RIDVal O’SheaComo 9/10/’03 SCM at a pore bottom of a DRIE matrix after deposition: typical signature of a pn junction SCM AFM A Profile along A PN Junction Formation: Results from LPCVD

3D-RIDVal O’SheaComo 9/10/’03 Why: concept of x-ray imaging detector below Formation of pn junctions on walls of pores Process steps: 1. Formation of pore arrays 2. Formation of pn junctions in pore walls 3. Filling pores with a scintillator, CsI(Tl) 4. Contacts and bump-bonding to the ROC

3D-RIDVal O’SheaComo 9/10/’03 Making Electrodes Metal evaporation: Ti (33nm) Pd (33nm) Au (150nm) Tracks of Al (150nm) Wire bonding (25µm wire)

3D-RIDVal O’SheaComo 9/10/’03 3D GaAs Performance 241 Am Readout by DASH-E (P.Sellar – RAL) Bias 10V Run at –30 C as there is no leakage current compensation

3D-RIDVal O’SheaComo 9/10/’03 Conclusion Much more to do Si diode arrays 256X256 and 55 um pitch are being made Charge integrating chip for this array is designed -- Analogue Medipix Testing of and characterisation of bumped assemblies Improvement of filling uniformity – other fillings