4:40 pm Wednesday Afternoon, November 11, Room: J1
Vacuum Characterisation of Magnetron-Sputtered Amorphous
Carbon Films for the Eradication of Electron Cloud Effects in
Particle Accelerators
A. Ashraf, P. Chiggiato, P. Costa Pinto, M. Taborelli, Ch. Yin Vallgren, I.
Wevers
CERN, 1211 Geneva 23, Switzerland
G. Debut, R. Kersevan
ESRF, Grenoble, France
Outlook:
• Electron Clouds
• Magnetron Sputtered Carbon Films
• Vacuum Characterization
• Influence of Coating Parameters (Discharge Gas Pressure)
• Conclusions
1
Introduction: the LHC is back
CERN is preparing the Large Hadron Collider for a restart in a few days.
The beam energy will be progressively increased to the nominal value of 7 TeV
with a luminosity of 1034 cm-2s-1. (Luminosity=rate of p-p collisions at the intersection
points)
A higher LHC luminosity (1035 cm-2s-1) is required in about 8 to 10 years, when
an additional rate of collisions will be needed to reduce the statistical errors.
The improvement will be obtained by focusing the beam and increasing the
number of protons in a bunch.
The main limitation to the LHC bunch intensity lies in its injection chain, in
particular the SPS. Amongst the limiting factors, electron cloud is one of the
most severe.
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
2
Electron Clouds in High-Intensity Particle Accelerators
Electron clouds in beam pipes are generated by electron multipacting on the wall of the
vacuum chamber.
+
Electron cloud effects:
• Transverse emittance blow-up (bunch expansion).
• Dynamic pressure rise (electron stimulated desorption).
• Septum magnet sparking.
• Beam losses.
The electron cloud mechanism is eradicated whenever the maximum secondary
electron yield (dmax) of the beam pipe wall is lower than a well defined threshold.
For the LHC beam in the main ring and its injector chain:
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
dmax < 1.3
3
Carbon Films for the Eradication of Electron Clouds
For as cleaned st. steel, copper and aluminum dmax>2.
Lower values can be obtained by high temperature bakeout and high electron
bombardment doses (>10-3 C mm-2).
Ti-Zr-V film coating have dmax≈1.1 after activation at temperature higher than 180°C
(24h). But they cannot be applied to the SPS because the SPS magnet vacuum
chambers are not bakeable.
Graphite is a very interesting candidate…but it needs to be deposited onto the vacuum
chamber walls.
SEY of graphite
dmax
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
dmax
N. Rey Whetten, J. Appl. Phy. 34(1963)771
4
Carbon Films for the Eradication of Electron Clouds
Magnetron sputtered carbon films in cylindrical configuration.
-U
Graphite rod
Standard coating parameters (magnetic field 150 G)
Sampl
e
Discharge
gas
ID
100
mm
Sputtered
C atoms
Discharge Voltage
Current
Power
Max. sub. T
Pressure
Dep. Time
Thickness
gas
[V]
[A]
[W]
[°C]
[Torr]
[h]
[nm]
Ne
822
0.57
466.3
160
1.8E-02
22
580
1.2
+
B
dmax= 0.95
1.0
SEY
Vacuum chamber
The low dmax of graphite is preserved
0.8
0.6
5
0.4
0
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
400
800
1200
Primary Electron Energy [eV]
1600
5
Carbon Films for the Eradication of Electron Clouds
Advantages of magnetron-sputtered C films:
• They do not need any in situ bakeout to attain the low dmax .
• Their dmax is lower than that of activated TiZrV and scrubbed surfaces.
• Multiple exposures to air do not increase the dmax if the samples are correctly stored.
• Good adhesion, no loose dust C particles.
• Resistive behavior: major impact on the impedance can be excluded.
10 μm
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
0.2 μm
6
Vacuum characterisation of magnetron sputtered C films
To be implemented in the vacuum system of particle accelerators, magnetron
sputtered C films need to be thoroughly characterized. The following characteristics are
to be measured:
1. Water vapor outgassing rate for unbaked samples (important for the SPS on which
the pumping speed cannot be increased)
2. Outgassing rates of the main gases released after bakeout (necessary for baked
accelerators, for example the damping rings of CLIC)
3. Electron stimulated desorption yields.
4. Photon stimulated desorption yields.
A 50 cm long, 10 cm diameter, stainless steel vacuum chamber was used for the first
tests.
A 2-m long, 6 cm diameter, double-wall for water cooling, stainless steel vacuum
chamber was used for the PSD measurement.
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
7
10
-7
10
-8
10
-9
-2
Outgassing Rate [Torr l s cm ]
Water Vapor Outgassing
-1
Conductance C= 2.88 l s-1 for H2O
Sample
TMP SN2= 250
ls-1
Penning gauge
10
-10
10
-11
Carbon coated
Bare stainless steel
10
0
10
1
10
2
Pumping Time [h]
The water vapor outgassing rate is higher than that of uncoated stainless steel by a factor
of 20 after 100 h of pumping.
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
8
Thermal Outgassing of Baked C Films
The sample is baked at 150°C (24h)
BA gauge
RGA
Sample
Sample
Main chamber
NEG strip pumped
Q
C
Penning
gauge
H2
LN2 trap
Exhaust

tacc
Sc is the conductance of the orifice
IRGA the RGA signal
RGA the calibration factor
tacc the accumulation time
Mass 2 RGA signal [A]
TMP station
260 l/s
Sc   I RGA   RGAdt
10
-10
10
-11
Start opening
variable leak valve
10
-12
10
-13
1200
1300
1400
1500
1600
1700
Relative Time [s]
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
9
-4
1 10
-4
8 10
-5
H2
Quantity of gas accumulated [Torr l]
1 10
2
Quantity of accumulated H [Torr l]
Thermal Outgassing of Baked C Films
6 10
-5
4 10
-5
2 10
-5
10
-3
10
-4
10
-5
0
10
20
30
40
50
60
2
CO
2
10
-6
10
-7
CH
10
-8
10
-9
70
Ne
-10
0
5 10
4
Compared to uncoated stainless steel:
H2: about 5 times lower
CO2: at least a factor of 5 higher
The outgassing of the discharge gas (Ne) is not an issue
1 10
5
2 10
5
2 10
5
3 10
5
Accumulation Time [s]
Accumulation Time [h]
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
CO
4
10
0
H
Outgassing Rate
Gas
H2
CH4
CO
CO2
Ne
[Torr l s-1 cm-2]
3 x 10-13
2 x 10-16
2 x 10-15
7 x 10-15
1 x 10-16
10
Electron Stimulated Desorption
The system is baked at 300°C (24h); the sample is measured unbaked and after 2 h heating at some
selected temperatures
Electron energy: 500 eV --Bombarding current: 1 mA--Estimated bombarded area: 200 cm2
Measurement taken after 100 s of bombardment
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
11
Electron Stimulated Desorption
10
-1
H
Bare St. Steel
2
-2
st
C coating (1 cycle)
10
-3
10
-4
10
-5
nd
C coating (2 cycle)
0
50
100
150
200
250
10
Desorption Yield [molecules/electron]
Desorption Yield [molecules/electron]
10
300
-1
CO
10
10
-3
10
-4
10
-5
CO has the highest ESD yield,
followed by H2 and CO2.
nd
C coating (2 cycle)
0
50
10
Desorption Yield [molecules/electron]
Desorption Yield [molecules/electron]
10
4
-2
Bare St. Steel
10
-3
st
C coating (1 cycle)
10
-4
10
-5
nd
C coating (2 cycle)
0
50
100
150
200
250
100
150
200
250
300
For heating temperature higher
than 120°C, the ESD yields of
the C coated sample are lower
than that of bare stainless
steel.
Heating Temperature (2h) [°C]
-1
CH
Bare St. Steel
-2
Heating Temperature (2h) [°C]
10
st
C coating (1 cycle)
300
Heating
Temperature
(2h) [°C]11, 2009
Paolo Chiggiato,
AVS 56th,
San Jose, November
-1
CO
10
-2
10
-3
10
-4
10
-5
2
st
C coating (1 cycle)
Bare St. Steel
CO and CO2: C coated and
uncoated samples have a
similar behavior.
nd
C coating (2 cycle)
0
50
100
150
H2 and CH4: when heated, the
carbon coated sample is at
least 10 times better than the
uncoated.
200
250
Heating Temperature (2h) [°C]
300
12
Desorption Yield [molecules / electron]
ESD yields for 24 h heating at 250°C
10
-1
10
-2
Venting in air: 24 h
Heating Cycle : 250°C for 24 h
H St. Steel
2
10
-3
CH St. Steel
4
10
-4
H a-C
2
10
-5
CH a-C
4
10
-6
0
1
2
3
4
5
Number of air venting / heating cycles
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
13
Desorption Yield [molecules / electron]
ESD yields for 24 h heating at 250°C
10
-1
10
-2
Venting in air: 24 h
Heating Cycle : 250°C for 24 h
CO St. Steel
10
-3
10
-4
CO St. Steel
2
CO a-C
CO a-C
2
10
-5
10
-6
0
1
2
3
4
5
Number of air venting / heating cycles
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
14
Photon Stimulated Desorption
Gate Valve
Conductance
SIP
SIP
Critical Energy
20.5 KeV
Angular acceptance
4.234 mrad
Photon Flux (E>10eV)
2.94x1015 photons (s mA)-1
Beam Energy
6 GeV
Typical Beam Current
185 mA
Angle of incidence = 25 mrad
Angle of incidence = 25 mrad
Diaphragm
SP
PG BAG
PG
Water or liquid nitrogen
cooling circuit
NEG-coated chamber
Front-End Slits
and Absorber
C coated chamber
QMA
BAG PG
Turbomolecular
pumping group
PG : Penning Gauge
BAG : Bayard-Alpert Gauge
QMA : Quadrupole Masss Analyser
SIP : Sputter Ion Pump
SP : Sublimation Pump
The system is bakes at 300°C (24h). The
sample is not baked.
31
The sample is separated from the rest of the
system by a gate valve (at the diaphragm
position, not pictured in the drawing); it is
pumped by an auxiliary TMP during the
bakeout of the system.
At the end of the bakeout, the gate valve is
opened.
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
15
Photon Stimulated Desorption
The photon desorption yield of the unbaked C coated sample is lower
than that of uncoated stainless steel.
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
16
Photon Stimulated Desorption
CO and CO2 are the two leading
gases; on the contrary, for
stainless steel, H2 is the main
desorbed gas.
In progress: measurement of
baked carbon coated samples.
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
17
Role of the Discharge Gas Pressure
10
-7
-2
Outgassing Rate [Torr l s cm ]
-2 in particular water
Is there any chance to reduce even further the outgassing
P =1.8rates,
x 10 Torr
Ne
-3
-8
vapor outgassing of unbaked
the coating parameters?
P =3.9 xby
10changing
Torr
10 samples,
-1
Ne
Decreasing the discharge
gas pressure by a factor of 4.6
-9
10Sampl
e
ID
100
-10
10 mm
Discharge Voltage
Current
Power
Max. sub. T
Pressure
Dep. Time
Thickness
gas
[V]
[A]
[W]
[°C]
[Torr]
[h]
[nm]
Ne
830
0.31
257.3
140
3.9E-03
23
715
Uncoated st. steel
10
-11
10
-1
10
0
10
1
10
2
Pumping Time [h]
Modification of the outgassing rates.
Unbaked samples: Water vapor outgassing is reduced by a factor of 10
Baked samples (150°C, 24h):
• H2, CO and CO2 : no significant variation (less that a factor of 3 change).
• Ne: 2 order of magnitude higher (from 10-16 to 10-14 Torr l s-1 cm-2)
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
18
Conclusions
1. Magnetron sputtered carbon films are an effective solution to eradicate electron clouds in high
intensity particle accelerators. They can be implemented both in bakeable and unbakeable beam
pipes of existing and future accelerators.
2. The thermal and stimulated outgassing features are in general better than those of uncoated
stainless steel, except for the water vapor outgassing of unbaked samples.
3. The latter records a reduction of one order of magnitude by decreasing the discharge gas
pressure by a factor of about 5…at the detriment of the Ne outgassing rate.
Ongoing activities
1. Optimization of the coating parameters.
2. Study of the implementation of such coatings in real accelerators, with the final objective of
coating the whole SPS ring (8 Km, more than 1000 vacuum chambers).
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
19
Backup slides
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
20
Backup slides: LHC injector chain
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
21
Backup slides: Why a lower discharge gas pressure reduces
water outgassing?
L. G. Jacobsohn and F. L. Freire, Jr. J. Vac. Sci. Technol. A 17.5., Sep/Oct 1999, p. 2841
‘…a decrease of the network interconnectivity occurs for increasing plasma pressure
depositions.’
Lower pressure means:
Higher mean free path for C and energetic Ne atoms
higher energy of impingement onto the growing film
higher density
lower porosity
lower water intake.
Paolo Chiggiato, AVS 56th, San Jose, November 11, 2009
22