Fabrication of crystals for channeling at INFN

1
Fabrication of crystals for channeling at INFN
Vincenzo Guidi
Sensors and Semiconductors lab University of
Ferrara and INFN - Italy
Fermilab, December 6-7, 2007
2
Outlook

• Sample preparation
• Treatment of surfaces
• Surface characterization
• Participation to H8RD22 experiment(CERN, IHEP,
  INFN, JINR, PNPI)
• Interest in collimation experiment at FNAL

3
Dicing of crystals I
Si samples for channeling are achieved from
wafers or ingots. Proper sizing of the crystal
to the wanted application involve dicing and
other operations that partly damage the surface
of the crystal
4
Dicing of crystals I
Si samples for channeling are achieved from
wafers or ingots. Proper sizing of the crystal
to the wanted application involve dicing and
other operations that partly damage the surface
of the crystal
Mosaicity is being generated at surfacewhich is
often of the order of the critical angle of
impinging particles
5
Dicing of crystals II
Crystals were diced from a commercially available
wafer using a fine-grit blade to minimize the
mechanical damage during the cut.
Diamond grain size 4-6 ?m Density 62 Dicing
speed 0.5 mm / min
Diamond blade
Dicing machine
6
Isotropic chemical etching
Planar etching erodes the surface of a crystal
layer by layer.

• HNO3 oxides Si to SiO2
• HF erodes SiO2
• CH3COOH as a solvent

Planar etching erodes the surface of a Si crystal
layer by layer HF, HNO3, CH3COOH (2155)
7
RBS - channeling mode
Rutherford backscattering under channeling mode
is a powerful method to investigate the structure
of a surface
Channeled particles penetrate deeper into the
crystal and are more difficult to emerge out of
the surface, resulting in a decrease of the
current
8
RBS spectra
gt 12??m
Chemical etching I
APL 87 (2005) 094102
9
AFM analysis
10
Roughness analysis
As cut
We observed roughness increase with chemical
etching
CE1
11
Etching vs. polishing
Measurements taken in an external beam line at
IHEP with 70 GeV protons
Mechanically polished surface
Chemically etched surface
RSI 73 (2002) 3170
12
Experiment H8RD22
1-mrad-accuracy goniometer

• Multistrip detectors
• (10/30 mm spatial resolution)
• AGILE type
• AMS type

• External line H8 of the SPS
• 400 GeV/c protons
• lt 8 mrad divergence

Spokesman Dr. Walter Scandale
13
Volume reflection
Unchanneled particle
Uo
d
Channeled particle
Volume-reflected particle
Volume-captured particle
Volume reflection was predicted by Taratin and
Vorobiov In 1988
Channeled particle
14
Observation of VR

1. Primary beam
2. Channeling
3. Dechanneling
4. Volume reflection
5. Volume capture
6. Primary beam

PRL 98 (2007) 154801
Measurements with sample ST4 (now here at FNAL)

• Deflection occurred at about 13.5 mrad
• Fraction of diverted particles larger that 97
• Large acceptance

15
Demand for lower roughness
An application of crystals would be the
collimation of beam halo in next generation of
hadron machines (e.g. the LHC)
Particles in the halo drift outwards at the rate
of 2 nm per turn. Since the tune is not integer,
the particles will hit the crystal every 10-20
turns and thereby the first impact parameter of
the particles onto the crystal will be in the
range of 100 nm (curtesy of V. Previtali and R.
Assmann)
It demands a crystal with a roughness lower than
100 nm on the lateral faces of the crystal
16
Rivisiting chemical etching
Formation of a morphologically non homogeneous
surface may be prevented by an unusually large
concentration of HNO3
Oxidation of Si by HNO3 preferentially begins at
the defects forming nitride oxide, which locally
catalyzes further the oxidation reaction in the
neighboring area and finally results in formation
of a crater.
An excess of HNO3 strengthens the oxidation,
which involves the whole surface nearly at
uniform rate. Properly short timing for reaction
leads to defect removal leaving a smooth surface
17
AFM analysis
As cut
New CE
Old CE
New CE Old CE
Standard roughness (235) nm (13520) nm
Significant decrease in roughness was observed
with the new CE
18
RBS spectra
4He
Low signal of backscattered particles means good
crystalline quality
p
Chemical etching II
APL 91 (2007) 061908
19
Structure of the surface
The surface ?min is defined as the ratio of the
RBS yield under channelling alignment and random
condition extrapolated at the surface channel.
The higher the degree of crystalline order in the
lattice the lower the surface ?min due to the
reduction of dechanneling from the defects in the
crystal.
probe as cut CE1 CE2
4He 16 4.1 2.2
p 18 4 2
Experimental ?min for a Si 110 sample with
probe particles of 2 MeV. An ideally perfect
surface yields ?min 2
20
Anisotropic etching
Anisotropic etching is a feasible way to realize
highly flat damage free surfaces of crystals
entirely by wet chemical methods
Etch rate on different silicon planes for KOH
20 at 40 C
(100) (110) (111)
7.1 µm/h 10.7 µm/h Negligible
nanometric roughness was achieved
21
Photolythography
a) Starting material (110) silicon wafer
b) LPCVD deposition of silicon nitride thin layer
c) Silicon nitride patterning
d) Etching of Si in KOH solution, silicon
nitride acts as masking layer
e) Silicon strips release
f) Removal of silicon nitride
22
VR vs. bending radius
Inefficiency vs. curvature radius
Deflection angle vs. curvature radius
23
Multistrip crystal
2mm
2mm
Frame
Cross section
Silicon strips
Thickness 0.5mm
70mm
24
Multiple VR with multistrip

• We were able to align up to 4 strips
• With 3 strips, deflection angle is 40.5 µrad
• Reflection efficiency is 93

25
Deformation of the frame
Anticlastic bending in a strip
Because of the frame, global deformation of the
crystal occurs
Ideal alignment condition
26
Ultra-thin frames

• Realization of ultra-thin frames
• Verification of independence of rotation of each
  strip

27
Interest in collimation project at FNAL

• Provision of crystals (yes)
• Characterization of crystals (yes)
• Refurbishing of previously used crystals, e.g.
  O-shaped crystal (yes)
• Participation in experimental runs (yes)
• Detectors (to be determined)
• Other issues (e.g. goniometer requested)